Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Impacts of urban hillslope development and agriculture on hydrology and water quality in the Chilliwack… MacDonald, Jennifer Rae 2005

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2005-0081.pdf [ 42.49MB ]
Metadata
JSON: 831-1.0091819.json
JSON-LD: 831-1.0091819-ld.json
RDF/XML (Pretty): 831-1.0091819-rdf.xml
RDF/JSON: 831-1.0091819-rdf.json
Turtle: 831-1.0091819-turtle.txt
N-Triples: 831-1.0091819-rdf-ntriples.txt
Original Record: 831-1.0091819-source.json
Full Text
831-1.0091819-fulltext.txt
Citation
831-1.0091819.ris

Full Text

IMPACTS OF URBAN HILLSLOPE DEVELOPMENT AND AGRICULTURE ON HYDROLOGY AND WATER QUALITY IN THE CHILLIWACK CREEK WATERSHED, BRITISH COLUMBIA by JENNIFER RAE MACDONALD B.Sc.(Env.), The University of Guelph, 2000 A THESIS SUBMnTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Resource Management and Environmental Studies) THE UNIVERSITY OF BRITISH COLUMBIA February 2005 © Jennifer Rae MacDonald, 2005 ABSTRACT The Lower Fraser Valley (LFV) has one of the most rapidly growing urban populations in Canada, and as a result water pollution problems associated with non-point source (NPS) pollution from urban expansion and agricultural intensification are increasing rapidly in this region. At the same time, the increasing demand for housing combined with the protection of agricultural land in the valley has pushed development onto the hillslopes. The transition from natural forest cover to impervious surfaces alters the hydrologic system, and increases the rate and volume of stormwater runoff that reaches the receiving watercourses. Due to the sensitivity of hillslope environments, and because upland activities may have damaging consequences downstream, development on these hillslopes presents many unique challenges for stormwater management. This research project uses a watershed approach to examine the impacts of land use (agriculture and urban development) on hydrological processes and surface water quality in a mixed land use setting. In Chilliwack, forest land on the hillslope is being converted into urban developments, and plans are under way to house up to 50,000 people on the hillslopes in Chilliwack over the next 25 years. The impact of this conversion on hydrology and water quality was examined in streams draining recently completed urban development (up to 2000 houses) by comparing the results with streams originating from undisturbed forested land. Using samples collected at twenty stations, a baseline was established for water quality and trace metals in sediments for various sub-watersheds in the study area. These results indicate that the lowland agricultural activities are the major source of NPS pollution in the watershed. Nutrient levels are elevated during the wet season, and many of the agricultural tributaries show evidence of eutrophication in the summer season. Trace metals associated with agricultural operations (Cu, Fe, Mn, Cd and Zn) were also elevated in the sediment of agricultural streams. Spatially, ammonia, orthophosphate and trace metals increased in the downstream direction along Interception Ditch (a large agricultural drainage ditch) indicating the effects may be cumulative. Results from the hillslope urban sites indicated that the hydrologic impacts of the development are the most important at this stage. Peak runoff was shown to be up to 1416% higher and lag times were up to 30 hours shorter in the suburban hillslope catchment (26% TIA) than for the forested catchment (4% TIA). While the impact on water and sediment quality was minimal, concentrations of orthophosphate, dissolved magnesium and potassium did show significantly elevated concentrations compared to the forested tributaries. ii Currently, the City of Chilliwack is experimenting with a number of innovative stormwater management designs (e.g. on-site detention ponds, infiltration galleries) in attempts to infiltrate much of the stormwater into the soil in these hillslope developments, before it reaches the streams. It is suggested that incorporating these low impact designs and source control methods may be more effective at mitigating the impacts of development than conventional stormwater management systems. iii T A B L E OF CONTENTS Abstract ii Table of Contents iv List of Tables viii List of Figures x Abbreviations xiv Acknowledgements xvi 1 INTRODUCTION 1 1.1 Study Goals and Objectives 2 2 LITERATURE REVIEW 3 2.1 Agricultural Non-Point Source Pollution 3 2.2 Impacts of Urbanization 4 2.2.1 Hydrological Characteristics of Stormwater Runoff. 4 2.2.2 Urban Runoff Water Quality 6 2.2.3 Ecological Impacts of Urbanization in the Lower Fraser Valley 7 2.2.4 Imperviousness as an Environmental Indicator 7 2.3 Stormwater Management 7 2.3.1 History of Stormwater Management 7 2.3.2 Low Impact Development (LID) and Source Control Measures 8 2.3.3 Effectiveness of Source Control Methods 11 3 T H E CHILLIWACK C R E E K WATERSHED: DESCRIPTION OF T H E STUDY A R E A 13 3.1 Physical Setting 13 3.1.1 Topography and Slope 15 3.1.2 Geology and Soils 15 3.1.3 Climate 17 3.2 Watercourse Characteristics 20 3.2.1 Stream and Drainage Network 20 3.2.2 Fish Habitat 24 3.2.3 Flooding Issues and Natural Hazards 25 3.3 Groundwater Resources 25 3.4 Human Activity / Land Use 27 3.4.1 Population Trends and Spatial Distribution 27 3.5 Hillslope Development: Issues and Concerns 29 3.5.1 Stormwater Management Innovations in the City of Chilliwack 30 3.5.2 Demonstration Projects 32 4 METHODOLOGY 33 4.1 Sampling Methodology 33 4.2 Field Methods 37 4.2.1 Surface Water 37 4.2.2 Sediments 37 4.2.3 Precipitation and Streamflow Data 38 iv 4.3 Laboratory Analysis 39 4.3.1 Surface Water Samples 39 4.3.1.1 Nutrients (Nitrate, Ammonia and Orthophosphate) 39 4.3.1.2 Dissolved Elements in Water 39 4.3.2 Sediment Samples 40 4.3.2.1 Physical Properties 40 4.3.2.2 Total Carbon and Nitrogen 41 4.3.2.3 Phosphorus 41 4.3.2.4 Trace Elements in Sediments 41 4.4 Quality Analysis and Quality Control 42 4.4.1 Site Variability 42 4.4.2 Method Accuracy of Water and Sediment Analysis 42 4.4.3 Method Precision for Water and Sediment Analysis 42 4.5 Geographic Information System (GIS) Methodology 43 4.5.1 GIS Database Creation 43 4.5.2 Contributing Area and Buffer Area Delineation 44 4.6 Water and Sediment Quality: Data Analysis Methods 46 4.6.1 Statistical Analysis of Water Quality Data 46 4.6.2 Statistical Analysis of Sediment Quality Data 47 4.6.3 Relationships between Land Use, Sediment, and Water Quality 47 4.7 Water Quantity: Data Analysis 48 4.7.1 Delineation of Storm Events 48 4.7.2 Calculating Storm Response Characteristics 48 5 L A N D USE IN THE CHILLIWACK C R E E K WATERSHED 51 5.1 Current Land Use 51 5.2 Trends in Land Use 54 5.3 Impervious Surface Area 57 6 CLIMATE AND HYDROLOGY 59 6.1 Climatic Characteristics 59 6.1.1 Temperature 59 6.1.2 Precipitation 61 6.1.2.1 Temporal Trends 61 6.1.2.2 Seasonal Variations 61 6.1.2.3 Spatial Variability 63 6.1.2.4 Distribution of Daily Precipitation 65 6.2 Streamflow Distribution 67 6.2.1 Variability in Streamflow 67 6.3 Hydrologic Response to Storm Events 71 6.3.1 Distribution of Storm Events 71 6.3.2 Storm Response Characteristics 73 6.3.2.1 Between Catchment Comparisons 74 6.3.3 Hydrograph Comparison of Individual Storm Events 75 6.4 Data/Analysis Considerations 82 7 WATER AND SEDIMENT QUALITY 83 7.1 Sediment and Water Quality Indicators 84 7.1.1 Nutrients 84 7.1.1.1 Nitrogen (Nitrate, Nitrite and Ammonia) 84 7.1.1.2 Phosphorus (Orthophosphate) 85 7.1.2 General Water Chemistry 87 7.1.2.1 pH. 87 v 7.1.2.2 Temperature 87 7.1.2.3 Specific Conductivity 87 7.1.2.4 Dissolved Oxygen 88 7.1.3 Major Ions in Water 88 7.1.4 Metals in Water and Sediment 89 7.1.4.1 Sources of Metals in Urban and Agricultural Areas 89 7.1.4.2 Chemistry and Fate of Trace Metals in the Aquatic Environment 90 7.2 Water Quality Results 93 7.2.1 Spatial and Temporal Trends in Water Quality Parameters 93 7.2.1.1 Variations in Ammonia-N 97 7.2.1.2 Variations in Nitrate-N 99 7.2.1.3 Variations in Orthophosphate-P 101 7.2.1.4 Variations in Specific Conductivity 103 7.2.1.5 Variations in pH 104 7.2.1.6 Variations in Temperature 106 7.2.1.7 Variations in Dissolved Oxygen 106 7.2.1.8 Variations in Dissolved Elements (Major Ions and Trace Metals) in Water 109 7.2.2 Comparison between Land Use Categories: Water Quality Parameters 117 7.2.3 Variations with Discharge 125 7.2.3.1 Influence of Storm Events on Water Chemistry 125 7.3 Results for Sediment Parameters 129 7.3.1 Sediment Properties: Particle Size 129 7.3.2 Variations in % Carbon and % Nitrogen 130 7.3.3 Metals in Sediment 131 7.3.3.1 Temporal Trends 131 7.3.3.2 Spatial Trends 131 Manganese 135 Lead 135 Iron, Copper, Zinc and Cadmium 137 Magnesium, Potassium, and Sodium 142 7.3.4 Phosphorus in Sediment 146 7.3.4.1 Bio-Available Phosphorus 146 7.3.4.2 Total Phosphorus 146 7.3.5 Comparison between Agriculture, Urban and Forest Land Uses: Sediment Parameters 148 7.4 Comparison of Water and Sediment Quality to Provincial and Federal Guidelines During Baseflow and Stormflow Conditions 149 7.4.1 Water Quality Compared to Provincial Water Quality Guidelines 149 7.4.2 Sediment Quality Compared to Federal Sediment Quality Guidelines 152 8 L A N D USE AND W A T E R INTERACTIONS 154 8.1 Correlations with Agricultural Land Uses 157 8.2 Correlations with Natural Land Cover 159 8.3 Correlations with Urban Land Uses 161 8.4 Comparison Between Land Use Components 163 8.5 Correlations with Impervious Surface Area 163 9 DISCUSSION 164 9.1 Climate and Hydrology 164 9.1.1 Climate Variability 164 9.1.2 Distribution of Storm Events 165 9.1.3 Overview of Storm Response for the Different Sub-Watersheds 165 9.1.3.1 Forested Systems: Parsons Brook and Elkview Creek. 165 9.1.3.2 Urban Systems: Teskey Creek and Lefferson Creek 166 vi 9.1.3.3 Agricultural Systems: Semiault Creek 166 9.1.3.4 Groundwater Influenced System: Luckakuck Creek 166 9.1.3.5 Downstream (Mixed) Systems: Bailey Ditch and Chilliwack Creek. 167 9.1.4 Comparison of the Hydrologic Effects between Forested and Urban Sub-Catchments 167 9.1.4.1 Lag Time 167 9.1.4.2 Peak Runoff Rate 168 9.2 Water and Sediment Quality 168 9.2.1 Impacts of Agricultural Land Uses 168 9.2.1.1 Nutrients in Water and Sediment 168 9.2.1.2 Dissolved Oxygen (DO) 7 70 9.2.1.3 Specific Conductivity 170 9.2.1.4 Temperature 170 9.2.1.5 pH 171 9.2.1.6 Metals in Water and Sediment 171 9.2.2 Impacts of Urban Land Uses 173 9.2.2.1 Intensive Urban (Ml 9) 173 9.2.2.2 Sub-Urban Residential Hillslope Development: Comparison with Forested Control Area 173 9.2.3 Conditions in the Forested Area 174 9.2.4 Cumulative Effects (Downstream Trends) 175 9.2.5 Influence of Storm Events on Water Quality 176 9.3 Relationships between Land Use and Water and Sediment Quality 178 10 SUMMARY AND CONCLUSIONS 180 10.1 Land Use 180 10.2 Climate and Hydrology 180 10.3 Water and Sediment Quality 181 10.4 Relationships between Land Use and Water and Sediment Quality 182 11 RECOMMENDATIONS 183 References 187 Appendices 198 Appendix A: Land Use Information 198 Appendix B: Climate and Hydrology Data 209 Appendix C: Water Quality Data 237 Appendix D: Sediment Data 272 Appendix E: Quality Analysis 287 Appendix F: Land Use Correlations 293 vii LIST OF TABLES Table 2.1 Contaminant Removal Ranges in Percent for Several LID Practices 11 Table 3.1 Major Watercourse Characteristics within the Chilliwack Creek Watershed 23 Table 3.2 Summary Information for the Sardis Vedder and Rosedale Aquifers 26 Table 3.3 Summary of Low Impact Design Demonstration Projects in the City of Chilliwack 32 Table 4.1 Watercourse Classification for Sampling Stations 36 Table 4.2 Summary of Water Quality and Sediment Parameters Measured on Each Sampling Date .... 38 Table 4.3 Data Sources for the GIS Database 43 Table 5.1 List and Description of Land Use Categories 52 Table 6.1 Definition of Minor, Intermediate, and Major Storm Classes in Relation to Total Rainfall (mm) and Peak 15-minute intensity (mm/hr) 71 Table 6.2 Distribution of Storm Events in Three Storm Classes (Minor, Intermediate, Major) at Two Sites (Promontory and Marble Hill) for the Wet and Dry season 72 Table 6.3 Summary of "Average" Storm Characteristics for a Three Year Period (2.001 to 2003) 73 Table 6.4 Summary of Storm Response Variables for the Various Hydrometric Stations in the Chilliwack Creek Watershed 73 Table 6.5 Rainfall Summaries for the Six Selected Storm Events Based on Data from the Promontory Tipping Bucket Rain Gauge 76 Table 6.6 Lag Time and Peak Runoff Rate for Each Catchment for Two Minor Storm Events 76 Table 6.7 Lag Time and Peak Runoff Rate for Each Catchment for Three Intermediate Storm Events. 77 Table 6.8 Lag Time and Peak Runoff Rate for Each Catchment for a Major Storm Event 77 Table 6.9 Differences in Lag Time, Peak Runoff and Total Runoff between the Teskey (Urban) and Parsons (Forested) Catchments, for Three Storm Classes 80 Table 7.1 Overall Surface Water Chemistry of the Chilliwack Creek Watershed, and Comparison with Natural Background Levels and Other Studies in the Lower Fraser Valley (LFV) 94 Table 7.2 Major Ions and Trace Metals in Water for the Chilliwack Creek Watershed, and Comparison with Natural Background Levels and Other Studies in the Lower Fraser Valley (LFV) 95 Table 7.3 Overview of Mann-Whitney Comparison between Land Use Categories for Water Quality Parameters (Wet and Dry Seasons) 117 Table 7.4 Overview of Mann-Whitney Comparisons Between Land Use Categories for Major Ions and Trace Metals in Water (Seasons Combined) 117 viii Table 7.5 Overview of Mann-Whitney Comparison Between Wet and Dry Season Values for Water Quality Parameters 118 Table 7.6 Overview of Total Metal Concentrations in Sediments Showing a Significant Difference Between October 2002 and July 2003 Sediment Sampling Sets 131 Table 7.7 Metal Concentrations in Sediment for the Chilliwack Creek Watershed, and Comparison with Natural Background Levels and Other Studies in the Lower Fraser Valley (LFV) 133 Table 7.8 Overview of Mann-Whitney Comparisons between Land Use Categories for Trace Elements in Sediment 148 Table 7.9 Sampling Stations within the Chilliwack Creek Watershed Exceeding B.C. Water Quality Guidelines for the Protection of Aquatic Life during the Wet and Dry Seasons 150 Table 7.10 Sampling Stations within the Chilliwack Creek Watershed Exceeding Canadian Sediment Quality Guidelines for the Protection of Aquatic Life 153 Table 8.1 Characteristics of Land Use within Each Contributing Area 155 Table 8.2 Spearman's Rank Correlation Coefficients for Independent Contributing Areas and 100 m Buffers: Agricultural Land Use Indices versus Water and Sediment Quality Parameters.... 158 Table 8.3 Spearman's Rank Correlation Coefficients for Independent Contributing Areas and 100 m Buffers: Natural Land Cover Indices versus Water and Sediment Quality Parameters 160 Table 8.4 Spearman's Rank Correlation Coefficients for Independent Contributing Areas and 100 m Buffers: Urban Land Use Indices versus Water and Sediment Quality Parameters 162 Table 8.5 Overview of Spearman's Ranks Correlations (p<0.05) Between Land Use and Various Water and Sediment Quality Parameters 163 ix LIST OF FIGURES Figure 3.1 Location of the Chilliwack Creek Watershed within the Lower Fraser Valley, British Columbia 14 Figure 3.2 Topography of the Chilliwack Creek Watershed 16 Figure 3.3 Surficial Geology for the Chilliwack Creek Watershed: Dominant Texture, Parent Material and Drainage Class 18 Figure 3.4 Soils of the Chilliwack Creek Watershed 19 Figure 3.5 Stream Network for the Chilliwack Creek Watershed 21 Figure 3.6 Predicted Population Trends for Various Regions in the City of Chilliwack, 1998-2026... 28 Figure 3.7 Chilliwack's Integrated Stormwater Management Strategy for Managing a Complete Spectrum of Rainfall Events 31 Figure 4.1 Water and Sediment Sampling Stations, Chilliwack Creek Watershed 2002-2003 34 Figure 4.2 Percent of Land Use Activity within each Contributing Area 36 Figure 4.3 Map Showing the Contributing Areas for the Sampling Stations in the Chilliwack Creek Watershed 45 Figure 4.4 Graphical Representation of Storm Response Characteristics 49 Figure 5.1 Land Use in the Chilliwack Creek Watershed Study Area, 2002 53 Figure 5.2 Housing Development in the Chilliwack Urban Corridor by 'Type' of Dwelling 54 Figure 5.3 Land Use Changes (1995-2002) in the Chilliwack Creek Watershed 55 Figure 5.4 Land Use Changes 1995-2000: Promontory Development Region of the Chilliwack Creek Watershed Study Area 56 Figure 5.5 2003 Dwelling and Populations Estimates for Promontory 57 Figure 5.6 Percent Total Impervious Area by Total Upstream Contributing Area 58 Figure 6.1 Monthly Mean Maximum and Minimum Temperatures for the 1971 -2000 Normal Period, Chilliwack Climate Station 60 Figure 6.2 Mean Annual Temperature (°C) and 11-Year Running Means for the Chilliwack Climate Station 60 Figure 6.3 Total Annual Precipitation and 11-Year Average for the Chilliwack Climate Station 61 Figure 6.4 Mean Monthly Precipitation for the 1971-2000 Normal Period, Chilliwack Climate Station 62 x Figure 6.5 Total Monthly Precipitation for the Study Period (May 2002 to November 2003), Chilliwack Climate Station 62 Figure 6.6 May 2002 - November 2003 Cumulative Precipitation; and Average Cumulative Precipitation. Chilliwack Climate Station 63 Figure 6.7 Daily Precipitation at the Various Monitoring Stations throughout the Watershed for the Study Period (May 2002 to November 2003) 64 Figure 6.8 Distribution of the Number of Annual Rainfall Events (top graph); and the Distribution of the Annual Rainfall Volume (bottom graph) 66 Figure 6.9 Location of Streamflow Measurements within the Chilliwack Creek, Elk Creek and Ryder Creek Watersheds 68 Figure 6.10 2000-2003 Total Daily Runoff (mm) for the Various Sub-Catchments in the Chilliwack Creek Watershed 69 Figure 6.11 Lag Time: Boxplots for the Different Sub-Catchments within the Chilliwack Creek Watershed 74 Figure 6.12 Peak Runoff Rate : Boxplots for the Different Sub-Catchments 75 Figure 6.13 Hydrograph for a Minor (12-Nov-01) and Intermediate Storm Event (18-Mar-01) 78 Figure 6.14 Hydrograph for a Major Storm Event (17-Jul-02) 79 Figure 6.15 Comparison of Teskey (urban) and Parsons (forested) Hydrographs for Two Storm Events 81 Figure 7.1 Location of Sampling Stations along the Interception Ditch and Chilliwack Creek Mainstems, and Selected Tributaries 96 Figure 7.2 Spatial and Seasonal Variations in Streamwater Ammonia-N in the Chilliwack Creek Watershed 98 Figure 7.3 Spatial and Seasonal Variations in Streamwater Nitrate-N in the Chilliwack Creek Watershed 100 Figure 7.4 Spatial and Seasonal Variations in Streamwater Orthophosphate-P in the Chilliwack Creek Watershed 102 Figure 7.5 Spatial and Seasonal Variability in Streamwater Conductivity in the Chilliwack Creek Watershed 105 Figure 7.6 Spatial and Seasonal Variations in Dissolved Oxygen Concentrations in the Chilliwack Creek Watershed 108 Figure 7.7 Spatial and Temporal Variations in Dissolved Calcium (Ca) in the Chilliwack Creek Watershed I l l Figure 7.8 Spatial and Temporal Variations in Dissolved Potassium (K) in the Chilliwack Creek Watershed 112 xi Figure 7.9 Spatial and Temporal Variations in Dissolved Magnesium (Mg) in the Chilliwack Creek Watershed 113 Figure 7.10 Spatial and Temporal Variations in Dissolved Sodium (Na) in the Chilliwack Creek Watershed '. 114 Figure 7.11 Spatial and Temporal Variations in Dissolved Iron (Fe) in the Chilliwack Creek Watershed 115 Figure 7.12 Spatial and Temporal Variations in Dissolved Manganese (Mn) in the Chilliwack Creek Watershed 116 Figure 7.13 Nitrate-N Comparisons by Land Use Category 119 Figure 7.14 Ammonia-N Comparison By Land Use Category 120 Figure 7.15 Specific Conductivity Comparisons by Land Use Category 121 Figure 7.16 pH Comparisons by Land Use Category 122 Figure 7.17 Major Ions (K, Mg, Na, Ca) Comparison by Land Use Category 123 Figure 7.18 Dissolved Trace Metals (Fe, Mn) Comparison by Land Use Category 124 Figure 7.19 Storm Hydrograph for the 16-Oct-2003 Storm Event 126 Figure 7.20 Response of Zinc (Zn), Manganese (Mn) and Nitrate During the 16-October-2003 Storm Event 127 Figure 7.21 Response of Aluminum (Al) and Iron (Fe) During the 16-October-2003 Storm Event.... 128 Figure 7.22 Percentages of Sand, Silt and Clay at Each Sampling Station 129 Figure 7.23 %Carbon and %Nitrogen for July 2003 Sediment Samples 130 Figure 7.24 Spatial and Temporal Variations of Total Manganese (Mn) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 136 Figure 7.25 Spatial and Temporal Variations of Total Iron (Fe) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 138 Figure 7.26 Spatial and Temporal Variations of Total Copper (Cu) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 139 Figure 7.27 Spatial and Temporal Variations of Total Cadmium (Cd) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 140 Figure 7.28 Spatial and Temporal Variations of Total Zinc (Zn) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 141 Figure 7.29 Spatial and Temporal Variations of Total Magnesium (Mg) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 143 xii Figure 7.30 Spatial and Temporal Variations of Total Potassium (K) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 144 Figure 7.31 Spatial and Temporal Variations of Total Sodium (Na) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 145 Figure 7.32 Total P and Bio-Available (orthophosphate) Concentrations in Sediment, 147 Figure 7.33 Spatial Trends in Total Phosphorus Concentration in Streambed Sediments along Interception Ditch and Chilliwack Creek 147 Figure 8.1 Chilliwack Creek Watershed Land Use 100 m Land Use Buffers 156 xiii ABBREVIATIONS Places and Institutions B.C. British Columbia BCMELP British Colombia Ministry of Environment, Lands and Parks DFO Department of Fisheries and Oceans EC Environment Canada EPA Environmental Protection Agency FRAP Fraser River Action Plan FVRD Fraser Valley Regional District GVRD Greater Vancouver Regional District LFV Lower Fraser Valley MWLAP Ministry of Water Land and Air Protection Chemical Symbols and Formulas A l Aluminum Ca Calcium Cd Cadmium Co Cobalt Cr Chromium Cu Copper Fe Iron K Potassium Mg Magnesium Mn Manganese Na Sodium Ni Nickel P Phosphorus Pb Lead Zn Zinc XIV Other Abbreviations ALR Agricultural Land Reserve BAP bio-available phosphorus BMP Best Management Practices CV coefficient of variation DPS degree of phosphorus saturation DO dissolved oxygen EIA effective impermeable area GIS Geographic Information System HMP hexametaphosphate HOF hortonian overland flow ICP-AES Inductively Coupled Plasma - Atomic Emission Spectrometry IQR interquartile range ISQG interim sediment quality guidelines IR infrared LID low impact development M A C maximum acceptable concentration M A R mean annual daily rainfall MIT minimum interevent time M M T methylcyclopentadienyl manganese tricarbonyl NPS non-point source OCP Official Community Plan PEL probable effect level SEL severe effect level SOF saturated overland flow TEL tetraethyl lead TIA total impervious area TIN triangular irregular network TP total phosphorus W B M Water Balance Model YSI Yellow Springs Instrument X V ACKNOWLEDGEMENTS I am grateful for the financial support provided by the Natural Sciences and Engineering Research Council of Canada (NSERC) post graduate scholarship, and the Canadian Water Network National Centre of Excellence. I would like to thank my supervisor, Hans Schreier, for his continual guidance, support, and encouragement throughout this project. I am fortunate to have had a supervisor with such remarkable patience and enthusiasm. Thank you for having confidence in me, even when mine wavered. I am also very grateful for the contribution of the members of my supervisory committee, Ken Hall, Dan Moore and Les Lavkulich. I also extend my thanks to the City of Chilliwack for providing invaluable data for this project. In particular, my appreciation goes to Dipak Basu and Peter O'Byrne for taking time from their busy schedules to share their personal expertise and their knowledge of the area. In addition, I am indebted to the many students and staff at RMES who have helped me during my time at UBC. Special thanks to Julia Brydon, lone Smith, Sandra Brown, Gina Bestbier, Kevin Li, Jamie Ross, and Michelle Revesz for their invaluable assistance in the lab and field, and for their constant moral support. I would also like thank Carol Dyck and Keren Ferguson who were my lifelines in the water and soil quality lab. Thanks to my amazing friends who were there for me through the highs and especially the lows. I would especially like to thank Kristina (and Kirby) for our runs and daily chats; Marcus and Seungsoo for their humour when I needed it; my friends from afar - Anna, Noreen, and Kristy - for your never-ending friendship; Aunt Wilma for giving me a home away from home. And of course Geoff, for being unbelievably patient, supportive, caring and understanding throughout this whole process. Thanks for all the laughs, hugs, chocolate, dinners - and for constantly reminding me to take some time to enjoy life. You were always by my side when I needed you - even when it wasn't the most pleasant place to be. Thank you all for your friendship and support, even when I went 'a bit' crazy -1 am lucky to have found such good friends. To Mom, Dad and Juli who gave me tremendous encouragement in everything I have done. Thank you for your constant love and support, and for always believing in me so I could believe in myself. You were there for me - always. xvi 1 INTRODUCTION Our water resources are greatly affected by human activities. Alteration of the land surface for a variety of uses has induced changes to natural processes, modified water pathways and led to the deterioration of watercourses throughout the world (Peters et al., 1997). The Lower Fraser Valley (LFV) in British Columbia has some of the most productive agricultural land in Canada. It also has a rapidly growing urban population. As a result, water pollution problems associated with non-point source pollution (NPS) from urban expansion and agricultural intensification are increasing rapidly in this region. In addition, as impervious surfaces replace natural land cover, the changes in the hydrologic system increase stormwater runoff. In order to properly manage and protect aquatic systems it is important to have a clear understanding of the impacts of this land use intensification on water resources at the watershed level. Land use trends and ongoing concerns in the Chilliwack Creek watershed are representative of what is happening in the LFV: urban expansion into the forested hillslopes, agricultural intensification in the lowland and aquatic degradation due to NPS pollution and cumulative effects. The land use distribution of the Chilliwack Creek watershed provides an excellent opportunity to investigate water management issues in a mixed land use setting. Although watersheds are increasingly managed under the concept of multiple uses, studies concerning the effect of land use on water quality, even in mixed land use watersheds, generally focus on investigating the impact of only a single type of land use. Research on the interaction between, or the cumulative effects of multiple land use activities, distributed in both time and space, is rare (Sidle and Hornbeck, 1991). Evidence is increasing that the combined effects of several land use activities may have more devastating effects on water quality than the impact of individual land uses (Sidle and Sharpley, 1991; MacDonald, 2000). Furthermore, upland activities may interact with natural processes and/or other land use effects and have damaging consequences downstream. With Chilliwack's population expected to nearly double by 2025 and the protection of agricultural land in the valley, there is a demand to convert the forested upland area into residential housing. While the negative impacts of watershed urbanization on streams have been well documented, there have been very few investigations on the specific case of hillslope urbanization in the LFV. Hillslopes are sensitive environments and there is a greater potential for downstream impacts such as flooding, and an increased risk of slope instability when they are developed. For this reason, conventional stormwater management practices, which are designed to remove runoff from impervious surfaces as efficiently as possible and which deliver stormwater to receiving waters much faster and in greater volumes than natural conditions, may not be adequate to mitigate the effects of development in hillslope areas. 1 Currently, the City of Chilliwack is experimenting with a number of innovative stormwater management designs (e.g. detention ponds, on-site infiltration galleries) in an attempt to infiltrate much of the stormwater into the soils in these hillslope developments before it reaches the streams. The collection of baseline data during pre-development and in the early stages of urbanization will allow for comparison with data collected in later years to evaluate the effectiveness of this new stormwater management policy, and suggest modifications if necessary. 1.1 Study Goals and Objectives The research project will use a watershed approach to examine impacts of land use (agriculture and urban hillslope development) on hydrological processes and surface water quality in a mixed land use setting. Forest land on the hillslope is being converted into suburban residential developments, and plans are under way to house up to 50,000 people on the hillslopes in Chilliwack over the next 25 years. As development of the hillslopes progresses, it will become increasingly important to better understand the hydrological effects and pollutant loadings of these urban hillslope systems, and in turn how best to mitigate these negative impacts. To provide information how hillslope developments in the LFV will impact hydrology and water quality, this study compares data collected in streams draining a new urban hillslope development (up to 2000 houses) with data collected in streams originating from undisturbed forested land. At the same time, the impact from these two upland land uses on the agricultural lowland will be investigated. The specific objectives of the study are: 1) To establish baseline information for the watershed in terms of streamflow, land use, sediment and water quality, which is an essential pre-requisite for future impact assessment of the continuing urban hillslope development; 2) To investigate how the recent hillslope urbanization has altered streamflow response to different storm events, using the forested hillslope as a control; 3) To determine seasonal and spatial variability in water and sediment quality in the watershed, and compare the surface water and sediment quality of small streams draining undisturbed upland forest, recently competed residential developments in the same physical setting, and agricultural streams in the lowland; 4) To investigate how cumulative impacts of the different land use activities propagate along the mainstem downstream, and to determine whether the upstream urbanization is affecting downstream water quality; 5) To compare the use of buffers versus contributing areas for examining the relationships between land use indices and water/sediment quality in a mixed land use setting. 2 2 LITERATURE REVIEW Trends in the Lower Fraser Valley (LFV), and elsewhere in North America, suggest that agricultural activities are intensifying and that urban areas are continuing to expand to support the constantly growing populations. As the landscape is altered new stresses are place on the aquatic systems from both contaminant inputs and hydrologic changes. Both agricultural and urban areas have been recognized as important sources of runoff and non-point source (NPS) pollution (Leopold, 1968; Choe et al., 2002; Sharpley et al., 1994). Runoff from agricultural land was identified as the primary cause of water quality problems in over 40% of surveyed rivers in the United States (EPA, 2002a). Contaminated stormwater runoff is recognized as a leading source of water quality problems in urban settings; however the hydrologic impacts from urbanization can often be more harmful than the pollutants it carries (BCMELP, 1992). This chapter provides some background information on the issues of agricultural and urban NPS pollution, as well as the hydrologic changes that result from urbanization. 2.1 Agricultural Non-Point Source Pollution In Canada, modern agricultural practices have been developed to produce a higher yield from a smaller land base resulting from greater inputs of fertilizer and pesticides. Furthermore, over the years there has been a steady increase in the livestock population coupled with a decrease in the number of livestock farms (Statistics Canada, 2002; Smith, 2004). For instance, in the LFV the number of dairy cows has increased by 70% over the last 10 years, and the area now has the largest number of dairy cows per farm in Canada. The increased number of chickens per farm has been particularly dramatic, with a 52% increase between 1996 and 2001 (Schreier et al., 2004). In areas of intensive agricultural production, the higher inputs of fertilizer and manure applied to the land often exceed the crop requirements and the ability of the soil to assimilate it (Zebarth et al., 1999; Chadwick and Chen, 2002; Schreier et al., 2004). This is particularly problematic in areas where the spreading of manure is used as a means for disposal. Inadequate storage capacities often result in manure applications at times when there is low crop demand, for example in the late fall when high rainfall exacerbates the risk of surface water contamination. A recent study of the LFV region analyzed nutrient dynamics in agricultural systems in order to determine areas of excess nutrient applications. It was determined that 65% of the areas in the LFV had surplus nitrogen in excess of 100 kg/ha/year and phosphorus surplus applications were in excess of 50 kg/ha/year. (Schreier et al., 2004). This condition can result in an increase in nutrient/contaminant loss in runoff that may then contribute to eutrophication and contamination of the receiving waterways. 3 The transfer of pollutants from fields to surface water may occur as direct runoff, or by infiltration through the root zone and discharge as seepage of subsurface flow. The extent to which pollutants are transferred from fields to surface and groundwater is influenced by chemical speciation, availability to crops, soil properties, and factors controlling hydrologic processes (topography, drainage characteristics, climate) in addition to land management practices such as manure and fertilizer application rates, timing and method of application, and the time interval between applications (Muhammetoglu et al., 2002; Gburek et a l , 2000; Sharpley et al., 1994). For example, if manure is spread over the soil surface rather than incorporated through tillage, particularly if heavy rainfall occurs within a few days of application, the risk of surface and groundwater contamination increases (Daniel et al., 1994; Giting et al., 1998). Soil-bound pollutants (e.g. ammonia, phosphorus, trace metals) are generally lost through surface runoff. Higher losses occur in areas with a reduced soil infiltration capacity, steeper topography, increased concentrations at the soil surface, and a limited riparian zone. For more soluble contaminants (such as nitrate) the factors which enhance water pollution include high fertilizer or manure application rates, cropping systems with low uptake efficiency, and tile drainage systems (Nielsen et a l , 1982; Zebarth et al., 1999). This study focuses on nutrients and trace metal contamination; however there are a number of other contaminants that can contribute to agricultural non-point source pollution of waterways such as soil/sediment particles, pathogens, pesticides, fertilizers, hormones and antibiotic residues. A more detailed discussion of the chemistry, potential sources and the environmental impacts of nutrient and trace metals in aquatic systems is found in Chapter 7. 2.2 Impacts of Urbanization 2.2.1 Hydrological Characteristics of Stormwater Runoff In undeveloped areas, such as forested hillslopes, rainwater is stored in surface depressions or absorbed by soil and vegetation though various environmental processes including evapotranspitration (ET) and infiltration. Hewlett (1982) states that the majority of precipitation (up to 70% in temperate regions) leaves as evaporation. Precipitation that infiltrates the soil surface travels through the soil as either shallow subsurface flow or deep seepage that replenishes groundwater (Ziemer and Lisle, 1998). This groundwater storage ultimately maintains baseflow during dry periods. Because infiltration capacities of natural systems are generally high due to the high organic matter content and the activity of microorganisms which create an open soil structure (Dunne and Leopold, 1978), subsurface flow accounts for nearly all the water that is delivered to the stream channel (Harr, 1977). This water moves very slowly and only those parts of the catchment located near the stream itself generally contribute to stormflow (Booth, 2000). During a rainfall event, the water table rises and soils can become saturated 4 increasing the area contributing to rapid stormflow. In these saturated areas subsurface flow emerges as return flow; this and any additional rainfall runs off as saturated overland flow (SOF). Zones of SOF generally occupy small areas which expand during wet periods - a phenomenon known as the variable source area concept of storm runoff (Dunne and Leopold, 1978; Zeimer and Lisle, 1998; Booth, 2000). These areas generally occur where shallow subsurface flow converges in topographic depressions or accumulates in areas of decreasing hillslope gradient (e.g. in valleys near the stream channel). The size of these saturated areas influences the amount of stormflow that is generated. Hortonian overland flow (HOF), which occurs when rainfall falls on the land surface more rapidly than the soil can absorb it causing the excess rainfall to run over the land surface, is another mechanism by which stormflow can occur. Water from HOF flows above ground at substantially higher rates than subsurface flow; however, because infiltration rates generally exceed rainfall intensities in the Pacific coastal region, HOF does not generally occur in undeveloped forest areas (Zeimer and Lisle, 1998). When the landscape is altered, the type and magnitude of runoff processes are changed. As urban development progresses, the catchment surface undergoes a transformation from pervious to impervious as vegetation is cleared, soil is compacted and the land is graded, and impervious buildings and streets are constructed. These changes reduce interception, evapotranspiration, and the water storage capacity of the land (Savini and Kammerer, 1961; Burges et al., 1998; Konrad and Booth, 2002). The major change influencing runoff processes results from covering parts of the land surface with impervious surfaces (e.g. roofs, sidewalk, streets, and parking lots). This reduces the infiltration capacity of these areas to zero and converts what was once subsurface flow directly to HOF, accelerating stormwater runoff to ditches and streams following storm events - even for the short, low intensity storms which would generally not produce runoff under natural conditions. Consequently, the precipitation that falls reaches the stream within a few minutes, instead of what had been a delay of hours to days. This, in turn, increases the severity of flooding - with increases in the volume of runoff and the magnitude of peak flow (Carter, 1961; Leopold, 1968; Arnold and Gibbons, 1996; Konrad and Booth, 2002). While these effects are seen during both small, frequently occurring events and large, infrequent events, they are generally most pronounced during the smaller events (Schueler, 1994). These effects of urbanization are further exacerbated by the higher efficiency of water transport to stream channels provided by stormwater drainage pipes (Walsh, 2000; Arnold and Gibbons; 1996). Flooding is particularly problematic where urbanization is occurring in upland areas as the generation of stormwater can significantly alter the flow regime in the entire watercourse downstream. The decreased infiltration also reduces groundwater recharge, which not only threatens water supplies but reduces the groundwater available to supply baseflow - thus lowering low flows during dry weather. (Arnold and Gibbons, 1996, Dunne and Leopold, 1978). 5 The hydrologic consequences of urban development have deleterious effects on the receiving waterways, and often lead to wider, straighter channels and scouring of the stream bed as the stream tries to deal with the additional flow (Schueler, 1992; Booth, 1990). The enhanced runoff and more frequent flooding also cause erosion from construction sites and of stream banks and channels (Arnold and Gibbons, 1996). Furthermore, the prevalence of large woody debris, which has long been recognized as a key factor in creating complex channel conditions and habitat diversity, has been shown to decline with urbanization (Stephens et al., 2003; Homer, 1998). These factors, in turn, reduce the diversity and availability of in-stream habitat - as pool and riffle sequences and overhead cover are lost, and the streambed is covered by a uniform blanket of eroded sand and silt from the sediment loaded runoff (Schueler, 1992; Arnold and Gibbons, 1996; Walsh, 2000). Engineering responses to flooding (e.g. stream diversions, channelization, damming, culverts and piping) further destroy stream beds and related habitat (Arnold and Gibbons, 1996). Finally, the reduction in tree cover surrounding the stream results in water temperature fluctuations, which will stress fish (Arnold and Gibbons, 1996). 2.2.2 Urban Runoff Water Quality The urbanization-induced hydrological changes also have the potential to cause significant water quality problems in the local receiving waters. Stormwater runoff exhibits degraded water quality as it picks up pollutants that have accumulated on impervious surfaces, which are then transported directly to streams. Moreover, the decrease in infiltration reduces filtration of pollutants through soil and the uptake by plants, thereby further increasing the pollutant load entering the receiving waters. The highest pollutant loads are usually seen during the initial periods of stormwater runoff - a phenomenon known as the 'first flush' (Lee et al., 2002). These pollutants pose a toxicity risk to aquatic organisms, particularly in the summer when low flows have been substantially reduced due to reduced groundwater recharge. Constituents in urban stormwater runoff include suspended solids, bacteria, heavy metals, nutrients, pathogens, pesticides, organic matter, oils and grease; and they are derived from various sources. Paved areas such as highways and streets in urban areas are considered "stormwater intensive" land uses since they are highly impervious, and accumulate pollutants from vehicular activity (e.g. tire and break wear, vehicle emissions) (Marsalek et al., 1999). About a third of all pervious areas (e.g. lawns, parks) in the urban landscape receive high rates of irrigation, fertilizers and insecticide applications (Schueler, 1995). Because these pervious areas are frequently interlaced with impervious surfaces these pollutants often migrate to the impervious areas, which increase their potential to end up in nearby waterways. Lawn fertilizer, road dirt, soils, leaf fall, grass clipping, animal wastes, and detergents were identified as the primary sources of phosphorus in urban and suburban settings (Washbusch et al., 1999). Erosion from construction sites is a significant source of sediment pollution in streams. On a unit area basis, construction sites export sediment at 20 to 1000 times the rate of other land uses (CWP, 2000). 6 2.2.3 Ecological Impacts of Urbanization in the Lower Fraser Valley Both the hydrologic disruption and degradation of water quality associated with urban development can have significant ecological impacts. Urban sprawl has resulted in habitat loss and a decline in fish population in many of the small streams and wetlands in the LFV, which are critical spawning and rearing habitat for several salmonid species (BCMELP, 2000; Stephens et al., 2003; Slaney, 1996). In the LFV, 71% of streams are considered threatened or endangered, and a further 15% have been lost altogether as a result of urban growth (BCMELP, 2000; Stephens et al., 2003). 2.2.4 Imperviousness as an Environmental Indicator Numerous studies have shown a link between impervious surface area and the degradation of aquatic systems, with strong correlations found between hydrology, loadings from NPS pollution, thermal pollution, habitat structure, and biological integrity and diversity (Schueler, 1992; Booth et al., 1993; Schueler, 1994; Arnold and Gibbons, 1996). Consequently, catchment imperviousness is often used as an environmental indicator of aquatic system degradation in urban areas. Schueler (1994) reviewed the various studies that related imperviousness to changes in aquatic systems and concluded that "this research, conducted in many geographic areas, concentrating on many different variables, and employing widely different methods, has yielded a surprisingly similar conclusion - stream degradation occurs at relatively low levels of imperviousness (-10%)". While stream degradation begins at 10% imperviousness, it becomes completely degraded above the 30% threshold (Schueler, 1994; Arnold and Gibbons, 1996). A study in the Puget Sound region of Washington State found that water quality impacts were less important than hydrological or riparian zones changes in the degradation of stream health. It was determined that the impacts of poor water quality and concentrations of metals in sediments did not show significant impact on aquatic biological communities until above 50% total impervious surface area. 2.3 Stormwater Management 2.3.1 History of Stormwater Management Proper management of stormwater runoff is a key component of protecting property, water quality and aquatic ecosystems. Traditionally, stormwater management has been achieved through engineered drainage systems in which a series of gutters, drains and storm sewers collect rainwater from roads and transport it though pipes into nearby streams. The main goal is to remove runoff as quickly as possible from developed areas to prevent on-site flooding. In the early 1970s, the use of stormwater detention ponds was introduced as an additional method of detaining water in order to reduce peak flows in receiving waters. Detention ponds consist of a storage area into which stormwater runoff is directed and 7 then released gradually through a constricted outlet pipe. Detention ponds also have the additional benefit of providing some treatment of contaminated stormwater runoff through adsorption of contaminants to sediments, sedimentation, plant uptake and microbial processes (Pettersson, 1998). However, neither of these approaches fully prevents aquatic degradation or flooding risks. For example, while detention ponds slow down the water and reduce peak runoff rates, they do not reduce the total runoff volume. Instead, the total volume is spread out over a longer period of time, which can result in erosive streamflow over longer periods of time (Stephens et al., 2002). As stormwater runoff in the Lower Fraser Valley becomes more problematic as urban areas expand onto the surrounding hillslopes, municipalities are starting to change the way in which they approach stormwater management. Instead of piping water directly to streams, Chilliwack and other municipalities in the Greater Vancouver Regional District (GVRD) are making an effort to incorporate low impact development (LID) practices and source control alternatives to mitigate effects of stormwater runoff. These low impact and source control alternatives are discussed in the next section. 2.3.2 Low Impact Development (LID) and Source Control Measures Catchment imperviousness and the design of drainage infrastructure are the primary determinants of the quantity and quality of urban stormwater runoff delivered to receiving streams. Low impact development is a new approach to land planning which uses certain source control technologies and design practices to ensure that a site's post development hydrologic functions mimic those in its pre-development state (Nataluk and Dooley, 2003). Some of the basic principles include: 1) preserving the natural evapotranspiration capacity through conservation, landscaping and green roofs; 2) using designs that limit the creation of impervious areas; 3) incorporating source control strategies that preserve natural infiltration capacity by infiltrating rainfall near the source; and 4) re-using rainwater for irrigation and for indoor uses. As previously mentioned, the amount of impervious surface area in a catchment has been linked to flooding and stream degradation. Therefore, limiting the impervious coverage can reduce runoff and partially mitigate these problems. There are many strategies which could be used to reduce the amount of impervious surface areas when designing new residential developments. Stone et al. (2004) suggest that the most effective approach to reducing the area of residential impervious surfaces is to decrease lot size. For example, it was shown that, a reduction in the average lot size of new development from 4000 to 2000 m 2, a reduction in the frontage from 21 to 15 m, and a reduction in the front yard setback from 12 to 8 m would reduce total parcel impervious area by approximately 30%. Designing residential areas with narrower streets and street designs which reduce the size and number of intersections would also help 8 decrease the volume of stormwater runoff as streets are a significant portion of the impervious surface within residential subdivisions (Stone, 2004). It has been estimated that the elimination of parking on one side of the street can reduce stormwater runoff by 25 percent (CWP, 1998). Other practices include shared driveways, use of alternative/pervious pavements, and center islands in cul-de-sacs. Permeable pavement is an alternative to the common asphalt pavement. There are several types of porous paving materials, through which up to 80% of intercepted stormwater can infiltrate (Brattebo and Booth, 2003). These pavements tend to function better in low traffic areas such as parking lots, driveways and sidewalks since they are easily clogged, impeding their performance. Maintenance requires annual high powered vacuuming of the area to remove sediments (Nataluk and Dooley, 2003). However, implementing these urban design practices that reduce impervious coverage is not enough to protect downstream watercourses, since low levels of impervious coverage (10%) can cause significant damage (CH2MHill , 2002b). Source control measures provide a means to further reduce the runoff volume from impervious surfaces, as well as improve the quality of stormwater before it reaches the stream. These options can be implemented on a small site scale and are generally designed to capture and infiltrate small storm events and the first portion of larger storms on site, thereby reducing the volume of overland runoff. In terms of water quality, the first flush of pollutants that gets washed off the impervious surfaces at the beginning of rainfall events will be filtered and receive some treatment as they infiltrate into the ground. Some of the structural and non-structural options that can be implemented on a small site scale are described below. Lawns and landscaped areas have reduced infiltration capacity as the surface soils layers are often removed and heavily compacted, and replaced by a thin layer (often less than 50 mm) of imported topsoil. Runoff from these pervious areas can be virtually eliminated by providing 300 mm layer of landscaped absorbent soil, even where the hydrologic conductivity of the underlying soil is low (CH2MHill, 2002b). Re-directing runoff from impervious surfaces to areas where it can infiltrate (generally near the source) is another method by which stormwater runoff can be reduced. The effectiveness of the method will vary significantly depending on the type of surface over which the runoff is dispersed (CH2MHill , 2002b), and can be enhanced by creating infiltration facilities that are designed to retain runoff and provide time for water to infiltrate: • Infiltration galleries/trenches or soak-away pits are excavated areas filled with aggregate material to hold water until it can infiltrate into the ground. They are generally designed to retain the first flush and have been shown to be effective at pollutant removal and recharging groundwater tables. Maintenance is important to avoid clogging and groundwater contamination can be a concern if 9 proper studies are not undertaken prior to implementation (Brydon, 2004; Shammaa et a l , 2001). Exfiltration galleries/trenches function in a similar manner except that once the water has filtered though the soil media it is collected in an underlying drain system and conveyed to the stormwater system (Nataluk and Dooley, 2003). • Bioretention areas are shallow depressions that are filled with soil and vegetation to promote infiltration, evapotranspiration, filtration and uptake of nutrients and other pollutants (EPA, 2000). The planting soils used contain some clay which adsorbs pollutants such as hydrocarbons, heavy metals and nutrients. Often an organic layer is included to promote the degradation of petroleum based pollutants through the action of microorganisms. The nutrients and metals will eventually lower the cation exchange capacity of the soils (its ability to absorb pollutant particles through ion attraction), and consequently, the soil needs to be replaced after 5 to 10 years (Nataluk and Dooley, 2003). Infiltration galleries would be slightly more effective than a bioretention facility (of the same size) due to the higher storage capacity of gravel over absorbent soil. The effectiveness of these two systems could be enhanced by placing an infiltration chamber under the gallery or designing for surface ponding in a bioretention area, as both increase storage capacity (CH2MHill, 2002b). Additional opportunities to manage stormwater exist as the water is conveyed to the stream. Grass swales are vegetated channels designed to convey water away from streets and structures and provide an effective replacement for the tradition curb and gutter system in residential subdivisions. They generally function as a mechanism to slow runoff and as filtration/infiltration tools. Often they are used to convey water to a subsequent infiltration or bioretention area and function as a pre-treatment mechanism that filters sediment from stormwater. In general, grass channels are most effective when flow depth is shallow and slow (EPA, 2000). Periodic mowing and removal of sediment are their main maintenance requirements. The performance of swales is dependent on not only channel length, but also longitudinal slope and the use of dams to slow flows and allow for greater infiltration (EPA, 2000). Grass filter strips can be used as pre-treatment devices to intercept stormwater and remove sediment before water enters infiltration devices. They are most effective on minimal slopes (<2%) and under shallow flow conditions (Nataluk and Dooley, 2003). Stormwater runoff can also be avoided by redirecting rooftop runoff (that would normally be conveyed into gutter and storm sewers) onto vegetated areas such as grass swales, bioretention areas and French drains (excavated pits with aggregate stone). Alternatively, rainwater can be harvested in rain barrels or 10 cisterns and later used in irrigation of lawns. This would have the added benefit of reducing municipal water consumption (Fergusson, 1998). These smaller scale infiltration techniques can be implemented in combination, and in series, with stormwater detention ponds in order to maximize the opportunities to mitigate stormwater runoff issues. In addition, there are several non-structural Best Management Practices (BMP) that can increase the effectiveness of stormwater management. These include reducing the use of contaminants (e.g. phosphate detergents, de-icing agents, and fertilizers), street sweeping, and maintenance of structural BMPs. 2.3.3 Effectiveness of Source Control Methods The effectiveness of these source controls varies with their design, with precipitation patterns, and with soil type, among other factors (CH2MHill, 2002b). Currently, limited research has been conducted on the effectiveness of LID in retaining predevelopment hydrology and reducing pollutant loadings cause by stormwater runoff on developed sites. Still there are a few studies that have tried to analyze the effectiveness of various LID practices based on runoff and pollutant removal capabilities. Table 2.1 presents the contaminant removal efficiencies for various LID practices for several parameters. These results were originally presented in reports by Urbonas (2000) and EPA (2000), each of which draws data from various studies. Table 2.1 Contaminant Removal Ranges in Percent for Several LID Practices Type of LID TSS TP TN NH 4 + N0 3 Zn Pb Cu Fe Mn Porous Pavement 80-95 65 75-85 n/a n/a 98 80 n/a n/a n/a Grass lined swale 20-40 0-15 0-10 n/a n/a 0-20 n/a n/a n/a n/a Grass Buffer Strip 10-20 0-15 0-15 n/a n/a 0-20 n/a n/a n/a n/a Infiltration Basin 0-98 0-75 0-70 n/a n/a 0-99 0-99 n/a n/a n/a Biorentention systems n/a 16-87 49-75 49-75 15-26 64-98 70-97 43-97 n/a n/a Asphalt w/ grass swale n/a (-94) 42 45 44 46 59 23 52 40 Permeable Pavement 91 3 9 85 66 75 85 81 92 92 w/grass swale 11 Overall, the removal rates for metals were greater than the removal rates for nutrients. However, it can also be seen from Table 2.1 that there were wide ranges in the reported percent removals for metals. This is due to the variation in site conditions (terrain slopes, soil stability, detention time, incoming pollutant loads, soil condition, geology, local climate) and site-specific design details (Urbonas, 2000). Despite this, when properly designed for site local conditions it is very likely that these LID designs will remove pollutants from stormwater to some degree. Less information was found on the effectiveness in terms of runoff. Sabouring (1999) showed that total runoff volumes from grassed swales were 6-30% less than conventional systems. Another study showed that a parking lot with swales and permeable pavement had 80-90% less runoff than basins without swales, and 60-80% less runoff than basins with concrete or asphalt and swales, for rainfall events less than 2 cm. There were fewer differences between pavement types during larger storms, but basins with swales still showed about 40% less runoff compared to basins without swales (Kuo et al., 1999). The most appropriate source control options and design features for any given development (or re-development) site must be evaluated based on site specific conditions, such as soil type, land use type, rainfall, and groundwater characteristics (CH2MHill, 2002b). However, there appears to be limited scientific research on what variables affect the efficiency of the different designs, which LIDs function most effectively under what conditions, and how their performance will vary over time. The GVRD recently completed a study that attempted to answer some of these questions (CH2MHill, 2002b). To do this, they developed the Water Balance Model (WBM), an "interactive model that can simulate the performance of impervious controls, absorbent landscaping, infiltration facilities, green roofs and rainwater harvesting under various development scenarios" (Stephens et al., 2003). Since then the model has been enhanced to make it more user-friendly with the goal that it can be used as a decision support tool to evaluate land use planning decisions and their ability to meet stormwater management objectives at both the individual development site and watershed scale (Stephens et al., 2003). 12 3 T H E CHILLIWACK C R E E K WATERSHED : DESCRIPTION OF THE STUDY A R E A This research project was conducted in the Chilliwack Creek watershed, located in the Lower Fraser Valley, British Columbia (B.C.). This watershed was chosen for the following reasons: it encompasses several different land uses, each of which has the potential to negatively impact water quality and alter hydrology; there had been no previous comprehensive survey of surface water quality within the watershed (what minimal sampling that had been done did not include the upland area where development is beginning); the issues in the watershed are representative of what is happening in the LFV (i.e. urban expansion on the hillslopes, agricultural intensification in the lowland, and aquatic degradation due to NPS pollution and cumulative effects); and finally, the distribution of agriculture, forest and urban land uses in the watershed allows for an assessment of upland-lowland interactions and cumulative effects. The watershed has a number of tributary streams that are dominated by each of the three land uses (urban, forest, and agriculture); the residential urban developments and forests dominate the upland portion of the watershed, while agriculture is dominant in the lowland. Water quality and the hydrologic response of the suburbanized section of the hillslope is used as an indication of what is likely to be experienced by the forested portion of the hillslope if it is developed using conventional designs. This chapter will give a description of the study area with respect to its physical setting and climatic conditions, soils and geology, surface water and groundwater resources, and land use. An overview of the development that is currently underway and planned for the hillslopes in the Chilliwack area is also given. Finally, the innovative stormwater management approach Chilliwack is taking in order to address the issues and concerns of this hillslope development is outlined. 3.1 Physical Setting The Chilliwack Creek watershed is located in the City of Chilliwack, approximately 100 km east of Vancouver. It is situated at the eastern end of the Fraser Valley Regional District (FVRD) with the Fraser River to the north and the Cascade mountains to the south. Figure 3.1 shows the study area with respect to the entire Chilliwack Creek watershed. Although the study area itself does not cover the entire Chilliwack Creek watershed, for simplicity it will be referred to as such throughout this thesis. The study area has a catchment area of 55.4 km 2 representing 65% of the entire watershed (84.7 km2). Chilliwack Creek originates in Sardis and flows north through Chilliwack, eventually entering the Fraser River east of Chilliwack Mountain. 13 Chilliwack Creek Watershed Figure 3.1 Location of the Chilliwack Creek Watershed within the Lower Fraser Valley, British Columbia 14 This study focuses primarily on a sub-catchment of the Chilliwack Creek watershed known as Interception Ditch (Figure 3.1). This sub-catchment can be divided into three somewhat distinct areas based on land use. The flat lands at the base of the hillslope are primarily agricultural, while the hillslope has both a forested area and a recently constructed urban area. The western end of the hillslope currently has a small residential development (Promontory) which is continuing to be developed, while the remaining hillslope is forested but is slated to be developed in the future. Water from all of these areas eventually drains into Interception Ditch, a large constructed drainage ditch, which flows into Chilliwack Creek. 3.1.1 Topography and Slope The Chilliwack Creek watershed is comprised of the low-lying valley of the Fraser River floodplain (~10 masl), and the slopes of Mt. Thom (-480 masl) and Lookout Ridge (706 masl) as well as Ryder Lake Uplands (338 masl). The hillslope area has an average slope of 14°. However, some portions of the watershed are steeper with slopes of up to 62°. The transition zone between the hillslope and lowland valley is generally abrupt. This change in slope drastically changes the flow regime as fast flowing water from the hillsides is suddenly contained in the flat slower flowing ditches of the lowland. The lower velocities in the ditches result in lower flow capacities, and cause frequent flooding during heavier rainfall events. The topography of the watershed is depicted in Figure 3.2. 3.1.2 Geology and Soils The surficial geology of the Chilliwack area is described and mapped by Armstrong (1980). The soils and surficial geology in the area can be divided into four regions: • The hillsides (eastern section) are underlain by Mesozoic and upper Paleozoic sedimentary and metamorphic bedrock from the Pre-Tertiary era. In the eastern portion of the study area, this bedrock is overlain by thin (less than 2 m thick), medium-textured, glacial and eolian sediments. The Lonzo Creek soil series covers most of this region of the hillslope. • The Promontory region of the hillslope is underlain by late Pleistocene Sumas drift (till and glacial-fluvial deposits) and is locally capped by up to several meters of loess. The depth to bedrock in this area commonly exceeds 30 m. Marble Hil l , Abbotsford, and Ryder soils dominate this section of the hillslope. 15 Figure 3.2 Topography of the Chilliwack Creek Watershed • The floodplain consists of alluvial and floodplain sediments underlain by quaternary sand and gravel deposits that extend to depths of over 400 m (Monahan and Levson, 2003) Some of these floodplain deposits are derived from the Fraser River, and some are from the Chilliwack River during the post-glacial period. The floodplain soils derived from these deposits consist of laterally and vertically accreted silt loam and silty clay loams. Towards Chilliwack Creek medium-textured deposits from local streams become important, forming different soils: Lickman, Bates, and McElvee. • The lowlands at the base of hillslope near Promontory (southwestern section) consists of gravel deposits from the alluvial fan where the Chilliwack/Vedder River enters the Fraser Lowland have prograded over older deposits in the Fraser River valley (Dakin, 1994). These deposits are over 35 m thick at the mountain front (Dakin, 1994). Sardis and Hopedale soils dominate this portion of the watershed. A number of areas in the watershed, predominantly the area adjacent to the toe of the slope and sections of the floodplain, have experienced regular flooding in the past. In these areas, bog, swamp and shallow lake deposits (lowland peat, organic silt loam, silty clay loam) cover older floodplain deposits. These deposits form organic soils - primarily Annis soil and Banford muck. A map of the surficial materials defined by geologic origin (parent material), dominant texture and drainage is shown in Figure 3.3. Figure 3.4 is a map depicting the distribution of soils types in the watershed. Table A.2 describes the some characteristics of the major soils in the study area (Appendix A). 3.1.3 Climate The Chilliwack Creek watershed experiences a temperate climate, with cool dry summers and mild wet winters. In the winter, the Pacific westerlies bring in moist air and low pressure systems from the coast. As a result, about three quarters of the annual precipitation for the Lower Fraser Valley falls between October and March, and almost all of this falls as rain (Swain et al., 1997). During the summer, the presence of a high pressure system off the coast results in mostly clear skies and low rainfall. Summer storms are usually brief and intense. The months of July and August typically have the lowest amount of rainfall, and greatest evapotranspiration (Swain et al., 1997). As a result, little rainfall contributes to streamflow or replenishes groundwater during these months. The climate of the Chilliwack area is discussed further in Chapter 6. 17 1 0.5 0 1 2 Dominant Texture - Parent material Drainage Class I T-I V///// No data (urbanized) Sandy Loam- Colluvlum Sandy Loam- Fluvial Loamy Sand - Fluvial Silty Loam - Fluvial Silty Loam - Eolian Silty Loam - Eolian over Fluvial Glacial Silty Loam - E otian over Till ] Silty Clay Loam - Fluvial Organic Organic Fen R- Rapldety drained W-Well drained M - Moderately drained I - Imperfectly drained P - Poorly drained V- Very poorly drained Source: City of Chilliwack Figure 3.3 Surficial Geology for the Chilliwack Creek Watershed: Dominant Texture, Parent Material and Drainage Class 18 Soil Name No data • H Bates EH Gibsons iSS Lonzo Creek 1 1 Prest (shallow) N/A (urbanized) ?m Cannell • • Gravel Pit r~n Monroe 1 ' 1 Poignant HI Abbotsford n Calkins Henderson CZ3 McElvee ( S 3 Pelly Annis • i Elk Iv-y?;! Hopedale lllllll Marble Hill MB Pelly (shallow) wm Arnold ESI Fairfield 1 I Isar 7fA Matsqui 33 Recent Alluvium Blackburn Grevell \r^3 Lickman M i Niven F T ^ Ryder EB Banford B2B Grigg ea Uckman (shallow) MB Prest S§3 Sardis Source: City of Chilliwack Figure 3.4 Soils of the Chilliwack Creek Watershed 19 3.2 Watercourse Characteristics 3.2.1 Stream and Drainage Network The stream network for the Chilliwack Creek watershed is shown in Figure 3.5. As Chilliwack Creek flows from Sardis through Chilliwack towards the Fraser River it is joined by a number of major tributaries: Interception Ditch, Luckakuck Creek, Semiault Creek, and Atchelitz Creek. A l l but Atchelitz Creek are included in the study area. Characteristics and potential pollution sources for each major tributary are summarized in Table 3.1 Chilliwack Creek and Luckakuck Creek are former channels of the Chilliwack/Vedder River (Rood and Hamilton, 1995). Most of the substrate is silt, except near areas of upwelling where gravels are found (FRAP, 1999). The flows of these two creeks, as well as flows in Atchelitz Creek, result in part from groundwater inflow and seepage (Rood and Hamilton, 1995). Together Chilliwack and Atchelitz Creeks form a wetland area of approximately 145 ha. The wetland area is classified as having approximately 70% stream water and 30 % floodplain marsh (FRAP, 1999). Semiault Creek originates in the agricultural region east of Chilliwack Creek, and its flow consists predominantly of surface runoff from the adjacent fields. Interception Ditch is a large constructed drainage ditch which receives water from the hillslope area and also drains part of the agricultural area adjacent to the hillslope. Surface water extraction for irrigation and industrial licenses affects flow in all of these streams. Rood and Hamilton (1995) suggest that over forty percent of the summer low flows in Chilliwack and Luckakuck Creeks are consumed by water demands. Summer low flow problems have also been reported in Atchelitz Creek. Recent summer water use has been rated as more excessive than average, and as a result the groundwater table has been dropping in the vicinity of this watercourse (FRAP, 1999). In addition, it has been stated that irrigation withdrawals on Semiault Creek consume 100% of the naturalized summer 7-day mean low flow (FRAP, 1999). There have been significant impacts to streams within the Chilliwack Creek watershed as a result of population growth and intensification of human activity (both urbanization and agricultural activities). Most of the natural vegetation and riparian zones have been removed, and the streams and ditches have been channelized along various reaches to support agriculture and control flooding. In addition, water quality degradation has resulted from agricultural runoff and urban stormwater runoff and from erosion. 20 Interception Ditch Sub-watershed 1 0.5 2 Kilometers Roads Watercourses / V Streams and ditches /S/ Semiault Creek / \ / Benchley Creek / \ / Teskey Creek / \ / Chilliwack Creek / " S / Interception Ditch / \ / Bkview Creek /\S Parsons Brook Luckakuck Creek /S/ Armstrong Ditch / \ / Lefferson Creek / N / " Walker Creek /\y Atchelitz Creek / V Bailey Ditch Source: City of Chilliwack Figure 3.5 Stream Network for the Chilliwack Creek Watershed 21 The catchment area of Luckakuck creek has experienced considerable urbanization, so that at present approximately 2 1 % of the sub-watershed is effectively an impermeable area (EIA). Both industrial and residential developments are encroaching on the creek and riparian vegetation has been removed from all but a few reaches (Rood and Hamilton, 1995). There is also industrial development along the lower reaches of Chilliwack Creek and Atchelitz Creek; these streams receive stormwater from a number of bulk petroleum facilities and cooling effluent from various food processing plants. Industrial activities on Atchelitz creek also include a canning factory, a food processing plant, and a sawmill (FRAP, 1999). Agricultural activities are also affecting the watercourses in various portions of the watershed. The upper reaches of Atchelitz Creek, the mid-sections of Chilliwack Creek, and the entire length of Semiault Creek and Interception Ditch are the most severely affected areas. Bank erosion and sedimentation in ditches and channels, low flows, water quality issues (low DO, high ammonia and coliform levels), removal of riparian vegetation are ongoing concerns in these areas. Hillslope streams are characterized by relatively high gradients, with relatively good quality water in the undisturbed forests. Most of these streams eventually drain into Interception Ditch. Further development of the hillslopes will greatly increase the EIA of the upstream system, potentially altering the flow regime of the system. Depending on stormwater management practices, this may have severe consequences for the downstream lowland agricultural land. 22 Table 3.1 Major Watercourse Characteristics within the Chilliwack Creek Watershed Chilliwack Creek Semiault Creek Luckakuck Creek Atchelitz Creek Interception Ditch Status ( F R A P 1997) Endangered Endangered Endangered Endangered Not surveyed Trend ( F R A P 1999) Declining • Riparian zone loss • channelized • riparian zone loss • riparian zone loss Not surveyed Watercourse Environmental Issues ( F R A P 1997) Possible Pollution Sources • channelized • significant water diversions • significant water quality problems • urbanization has significantly altered stream basin significant water quality problems significant water diversions urbanization has significantly altered stream basin • E I A > 10% • significant water quality problems • urbanization has significantly altered stream basin significant water quality problems adjacent agriculture urban development industrial discharges adjacent agriculture adjacent agriculture urban development industrial discharges • adjacent agriculture • industrial discharges • adjacent agriculture • Bailey landfill Water Licence Demand (Rood and Hamilton 1995) Domestic (gal/day) 19,300 16,000 500 0 Irrigation (ac-ft) 1,357 293 317 263 Industrial Use (gal/day) 1,517,043 4,723 1,500,500 10,000 3.2.2 Fish Habitat Chilliwack Creek and its tributaries support populations of a variety offish species including: coho salmon {Oncorhynchus kisutch) and chum salmon {Oncorhynchus keta); steelhead trout {Oncorhynchus mykiss), rainbow trout and cutthroat trout {Onchohynchus clarki clarki). Non-salmonid species include various carp, sturgeons, sculpins, suckers, sticklebacks, and calico bass. The rare and endangered salish sucker (Catostomus sp.) has also been recorded in the system (FRAP, 1999). The salmonid species spawn in the upper reaches of Luckakuck, Atchelitz and parts of Chilliwack Creek. Semiault Creek and Interception Ditch support no spawning; and fish in these tributaries consist primarily of coarse fish species. This thesis does not focus on fish; however land use activities in the watershed can affect fish habitat and productivity. As most of the lowland streams flow through agricultural areas, many of the constraints are related to agriculture. The primary concerns include: 1) decreased stream flows due to loss of recharge areas and the number of water withdrawal licenses on these stream; 2) decreased spawning habitat due to sedimentation and siltation resulting from bank erosion; and 3) the removal of fish habitat elements (gravel substrate, riparian areas, large organic debris etc.) due to regular clearing of waterways (FRAP, 1999). Notable water quality problems have also been evident. Low dissolved oxygen, high phosphorus and ammonia values, high water temperature and fecal coliform counts have been recorded. In addition, the Chilliwack pump station causes fish passage problems. Table 3.1 summarizes some of the pressures and concerns that the Department of Fisheries and Oceans (DFO) has noted for the various watercourses. Ongoing industrial and residential development is also increasing the risk of altering the hydrology and degrading water quality from stormwater runoff, contaminated discharges, and riparian vegetation removal. Riparian habitat provides important benefits to fish populations. Many of the streams draining the hillsides and uplands contribute important nutrients into fish bearing streams. Protecting riparian areas in the upland headwaters is therefore essential in minimizing potential environmental disruption caused by development. 24 3.2.3 Flooding Issues and Natural Hazards The portion of the watershed located in the lowland is subject to frequent flooding resulting from heavy rainfall or snowmelt (spring freshet) from the Fraser River, and from winter storm and flood flows from the Chilliwack River/Vedder Canal basin. The 200-year flood zone covers most of the valley to the north, and some of the lands to the south of the Trans-Canada Highway, and is susceptible to flooding. The area has experienced major flooding as recently as 1984. Since then a system of dikes was built to protect adjacent land from the Chilliwack/Vedder River and the Fraser River from flooding. In addition to the dyking system, the Chilliwack Pump Station located at the end of Chilliwack Creek pumps out water during high flows. The security against flooding provided by the dykes is being reduced as gravel originating from upstream is transported and deposited in the confined channel. As the beds of the Fraser or Chilliwack rivers rises (aggrades), the water surface level also rises for a given flow; and as a result, the level of flood protection afforded by the dykes along the river is reduced. It is known that in some places along the Fraser River, the dykes are now insufficiently high to assure protection against the water levels for which the dyke system was designed (i.e. the 1894 flood record) (UMA, 2001; Church et al., 2001). As a result, gravel (minimum of 685,000 cubic meters) is being removed on an almost annual basis to minimize the risk of flooding. The hillside and uplands areas may also be subject to natural hazards such as flooding, erosion and instability, particularly i f urban development alters stream flow and increases erosion potential. Erosion of upland areas creates serious downstream consequences, such as local flooding, impediment to farmland drainage, and potential loss of property on the hillsides. 3.3 Groundwater Resources Groundwater is the principal source of drinking water in the City of Chilliwack. Two major aquifers have been identified in the Chilliwack Regional District: Sardis-Vedder and Rosedale (Table 3.2). The Sardis-Vedder aquifer is a shallow (~10 m), unconfined aquifer that covers an area of 25 km 2 , and is up to 60 m thick in places (Dakin, 1994). The aquifer is comprised of a sandy, gravel alluvial fan formed by the Chilliwack River where the river exits the Cascade Mountains, and becomes known as the Vedder River (Rood and Hamilton, 1995). It is capped with a thin, relatively permeable layer of sand and silty sand. 25 Flow in the aquifer is relatively fast (3380 L/s or 107,000,000 m3/yr) and occurs in a radial direction within the fan (IRE, 2001). A portion of this water is extracted by pumping wells and a portion discharges as springs. These springs contribute to the headwaters of a number of streams including Luckakuck Creek, Chilliwack Creek, and Atchelitz Creek. It is likely that flows in many of the smaller tributaries are maintained by groundwater recharge during the late summer (Hamilton and Rood, 1995). Recharge for the aquifer comes from both infiltration of precipitation (12,352,810 m3/yr) and leakage from the perched bed of the Vedder River (Dakin 1994; IRE, 2001). The Sardis-Vedder aquifer is the most important source of drinking water for the City of Chilliwack with five wells producing more than half of the drinking water for the community. The aquifer is very productive with sufficient capacity to accommodate further population growth. However, because the Sardis-Vedder aquifer is very shallow, unconfined and overlain by extremely permeable soil, the aquifer is highly vulnerable to contamination. In 1997, the City of Chilliwack implemented a groundwater protection plan to ensure that no significant land use change occurs in the aquifer recharge area and within the capture zones of the well. Chilliwack's secondary supply source, the Rosedale aquifer, was abandoned in 1997 due to high iron and manganese levels (City of Chilliwack, 2004). The high organic content and low flushing rates in this aquifer are the likely causes of the high dissolved iron and manganese concentrations. This aquifer was formed by Fraser River deposits and underlying glacio-fluvial sediments (Dakin, 1994). Table 3.2 Summary Information for the Sardis Vedder and Rosedale Aquifers (based on Dakin, 1994) Aquifer Area (km2) Average Thickness (m) Estimated Recharge (106 m3/yr) Annual Abstraction (106 m3) Recharge Source Sardis-Vedder 25 25 15 8 • Precipitation • significant recharge from Chilliwack/Vedder River Rosedale 28 40 10 2 • Precipitation • surface water recharge from Fraser River 26 3.4 Human Activity / Land Use The City of Chilliwack has both an urban sector and a substantial rural presence. A large portion of the watershed is comprised of the flat Fraser River floodplain and much of the land base surrounding the urban communities is part of the B.C. Agricultural Land Reserve (ALR). Traditionally agriculture has been the main economic activity within the watershed. Farms (primarily dairy/beef or forage crops) surround the stream channel and most of the lower agricultural area of the basin has been ditched and removed of its riparian vegetation. Currently, the headwaters and upper portions of the watershed area are primarily forested with new residential developments being built. The study area itself reflects this mixed land use setting, and supports a number of different land uses in various sections of the catchment. Current land use, its spatial distribution and the changes in land use since 1995 are discussed in more detail in Chapter 5. 3.4.1 Population Trends and Spatial Distribution The Lower Fraser Valley (LFV) is one of the fastest growing regions in Canada. The population in City of Chilliwack has grown from 41,471 residents in 1981 to 62,927 residents in 2001 (City of Chilliwack 2004), at an average annual growth rate of about 2.5%. It is estimated that by about 2010 the population will reach 85,000 (Official Community Plan (OCP) target). Currently, over eighty-two percent of the population lives in urban communities or surburban neighbourhoods, and the balance in the rural hillsides and farming areas (City of Chilliwack, 2004) (Figure 3.6). Population growth is guided by the city's Official Community Plan, produced in 1998. Because urbanization in the valley floor is restricted by the presence of the A L R lands, essentially establishing an urban containment boundary, the OCP has directed future growth towards two main areas in planning for this 85,000 population milestone: 1) densification and infilling of the existing urban corridor; and 2) development of selected hillside and upland areas (namely Promontory, Chilliwack Mountain and Eastern Hillsides). • Promontory: Promontory is a hillside community situated near the southern end of Sardis-Vedder with a current population of about 2,800. It is still in an active phase of development, and upon completion is expected to hold a population of 7,000. Over the next ten years it is expected to accommodate just under nine percent of the growth in the district. • Chilliwack Mountain: Another suburban hillside community, located west of Chilliwack Proper on Chilliwack Mountain. Currently the area consists of rural homes, and much of the hillside remains under forest cover. However, the present population of about 1,000 is expected to triple by 2026. 27 Development Areas: ^ C h i l l i w a c k Mountain rtflHEastern Hillsides r^pchllliwack Proper r^pRyder Lake Uplands dpSardis d f Rosedale Promontory ^ P G ^ 6 1 1 1 3 3 1 6 # Yarrow ^ ^ F r a s e r River roads arc ys^ChiHiack Regional District 1:145.000 Source: District of Chilliwack Official Community Plan (1998) 60 I Exisiting Population (1998) I Projected Population (-2026) Chilliwack Proper Sardis-Vedder Promontory Chilliwack Rural Communities E. Upland Area Mountain (Lowlands) Figure 3.6 Predicted Population Trends for Various Regions in the City of Chilliwack, 1998-2026 28 • Eastern Uplands Area: This area includes both Ryder Lake and the Eastern Hillsides. The Eastern Hillsides is a 1320 ha area adjoining the eastern boundary of the municipality. Suburban subdivisions in this area currently hold 900 residents, but according to the development plan for the area it could have a built-out capacity of 8,000. Overall, about sixteen percent of the city's population growth over the next ten years is expected to be in this area. Ryder uplands is likely to be the last area slated for development. At present the plan is to maintain it as a long-term community development reserve, with a low development density. However, once the Eastern Hillsides has reached its development capacity this area will likely be developed to accommodate a suburban residential community in order to meet housing demands. Long term plans call for five development areas, 1500 ha of developed land and a population of about 60,000 (City of Chilliwack, 2004). 3.5 Hillslope Development: Issues and Concerns The hillside and upland area slated for development is a sensitive environment, and may be subject to natural hazards such as flooding, erosion and slope instability. In addition, the area is located upstream of rich agricultural lowlands, which may pose liability issues for the municipality. According to B.C. drainage law, the municipality is liable for any runoff and flooding impacts resulting from urbanization. Consequently, there are a number of issues and concerns that need to be addressed if development of the hillsides is to proceed (CH2MHill, 2002): • Any increase in impervious surface on the hillslope will increase flow volume and velocity of stormwater to the low lying farm land. This could result in flooding along the natural and constructed drainage system, or could aggravate existing flooding problems both on-site or downstream; • Increased flow rates and volumes could destabilize the existing balance of the natural geomorphic drainage systems (e.g. increase stream bank erosion) resulting in bank or slope failures, destruction of habitat, downstream sedimentation leading to a decreased channel capacity which could impede farmland drainage, and smothering of spawning beds; • Alteration of the natural topography and native vegetation could result in unstable soil conditions in slopes/embankments, or may increase water temperature; • Alteration of the groundwater interflow could adversely change downstream baseflows and/or impair existing water rights; and • Rapid and direct transport of contaminants associated with urban activities may degrade water quality in the streams. 29 3.5.1 Stormwater Management Innovations in the City of Chilliwack With the pressure to develop on the steeper hillsides, the potential for downstream impacts and the resulting liability issues, the municipality has moved away from tradition stormwater management practices and has instead adopted an innovative integrated master drainage plan that requires land developers in the uplands to protect the natural water balance. In 2002, the City of Chilliwack completed their "Policy and Design Criteria Manual for Surface Water Management" which provides guidance on how to design on-site drainage systems that reduce runoff volume at the source. This manual uses concepts from, and was developed as a case study to the B.C. guidebook on stormwater management {Stormwater Planning: A Guidebook for BC). As a result, Chilliwack has become recognized as a leader within British Columbia in promoting and implementing changes in the philosophy of, approach to and standards for stormwater management (Stephan and Pringle, 2004). Conventional stormwater management practices are limited in that they generally focus only on controlling peak flows (by piping rainwater runoff to streams) during the larger infrequent storm events, while they neglect to manage runoff volumes from smaller events (CH2MHill , 2002a). However, many of the impacts to the aquatic system are dominantly controlled by the cumulative effects of smaller storm events rather than by rare, high magnitude storm events (McClintock et al., 1995). Chilliwack's innovative stormwater management approach is to manage for the complete spectrum of rainfall events, from the small frequent events to the extreme events (see Figure 3.7). The overall objective is to use a combination of source control and traditional stormwater management practices to decrease the volume of runoff that flows to the stream, thereby creating a situation that approximates the water balance of a naturally vegetated watershed (CH2MHill, 2002a): • Rainfall Capture (retention/runoff volume reduction): The small frequently occurring rainfall events, which account for the bulk of the total rainfall volume, are to be captured and infiltrated (or re-used) at the source (on building lots and within road right-of-ways). The goal is to control runoff volume so that the watershed behaves as though it has less than ten percent impervious surface area. • Runoff Control (detention/runoff rate reduction): The intermediate events are to be detained and released to watercourses or drainage systems at a rate that approximates the natural forested condition (~1 L/s/ha). • Flood Risk Management (conveyance): Extreme events (e.g. 100-year rainfall event) are to be safely conveyed to downstream watercourses. 30 RAINFALL DISTRIBUTION PATTERN Characterized from rainfall data compiled from short and long term climate stations < 30 mm precipitation (< 50% MAR*) 30 to 60 mm precipitation (50% MAR* to MAR*) > 60 mm precipitation (> MAR*) PERFORMANCE TARGETS Established for managing a complete spectrum of rainfall events RAINFALL CAPTURE RUNOFF CONTROL FLOOD RISK (Reduce Runoff Volume) (Runoff Rate Control) MANAGEMENT JX Capture the first 30 mm of Detain the next 30 mm of Ensure that stormwater rainfall per day at the source rainfall per day, and release infrastructure can safely and restore it to natural it to sewers or streams at the convey storms greater than hydrologic pathways (i.e. natural interflow rate. 60 mm. infiltration, evapo-transpiration, or reuse). DESIGN CRITERIA Performance Targets have been translated into 'design criteria' for application at the site level RAINFALL CAPTURE RUNOFF CONTROL FLOOD RISK MGMT. (Retention Facilities') (Detention Facilities) (Conveyance Facilities) n • Capture 300 m3 of rainfall • Provide an additional • Provide 'escape routes' for per hectare of impervious 300 m3 of detention extreme storms surface area storage per hectare of • Ensure that these routes • Infiltrate at the natural impervious surface area are hydraulically adequate infiltration rate of the local • Release to storm sewers or and physically adequate soils, and/or: streams at a rate of 1 L/s • Reuse within the per hectare development site MAR = mean annual daily rainfall Figure 3.7 Chilliwack's Integrated Stormwater Management Strategy for Managing a Complete Spectrum of Rainfall Events (Source: CH2MHH1,2002a) 31 3.5.2 Demonstration Projects There are now a number of innovative development projects within the Chilliwack Creek watershed where source control and low impact design technologies have been applied. Source control strategies for future development may include absorbent landscaping, infiltration facilities and extended detention, preservation of significant natural areas, green roofs, rainwater capture/reuse, and/or pervious cascading swale channel systems on the hillside. Low impact design technologies used may include smaller lot sizes, narrower roads, and elimination of curbs, gutters, storm drain and sidewalks. Table 3.3 Summary of Low Impact Design Demonstration Projects in the City of Chilliwack Project Location Source Control/Low impact development technology ustrial/ imercial Stream International Chilliwack - exfiltration gallery adjacent to stream - stormwater retention/detention trench •a e = o - u Chevron Gas Bar Sardis - exfiltration gallery under pavement and landscaping Fetterly Place Promontory storm sewer with partial exfiltration trench - soakaway pit on each lot - detention pond at outlet Note: conventional sub-division (i.e. curb and gutter) (A Byrant Place Eastern Hillsides - full exfiltration trench - on lot soakaway pits - detention pond - no curb and gutter Developmei Russel Heights Promontory (multi-family) - runoff from lots is directed to a large green area for infiltration - road runoff is directed to an exfiltration gallery that will provide detention for medium storms (large storms will bypass the infiltration gallery) Residential 1 Peach Road subdivisions - runoff goes into yards and then drains to a surface swale for infiltration - water runs down the road to an infiltration gallery - detention pond is being build to handle the largest storms - small lots, narrow roads - no storm drain Suncor Developments - road runoff flows to an infiltration gallery - narrow roads, no sidewalks - no curb and gutter West Point (Copper Ridge) - exfiltration gallery adjacent to small hillside stream Edward Street Apartments - exfiltration gallery under pavement The city plans to install monitoring stations at some of the residential housing projects to monitor both runoff volume and water quality (temperature, dissolved oxygen, turbidity, and pH). Practical experience and performance from these demonstration projects will enable constant improvement to land development and rainwater management practices. 32 4 METHODOLOGY This project used a variety of data sources in its analysis: surface water quality data obtained from the analysis of grab samples, sediment quality data obtained from the analysis of two sets of streambed sediment samples, continuous streamflow and precipitation data recorded from existing hydrometric stations and tipping buckets set up by the City of Chilliwack, and land use information created using a Geographical Information System (GIS). The following sections describe in more detail the methods used for data collection and analysis. 4.1 Sampling Methodology As previously mentioned, the Chilliwack Creek watershed was chosen for this study because it has sufficient tributary streams that are dominated by three different land uses (urban, forest, and agriculture). Urban development and forests dominate the hillslopes, while agriculture is dominant in the lowland which allows for an assessment of upland-lowland interactions and cumulative effects. A total of twenty sampling sites were selected for this study (Figure 4.1). Six of these sites (G l , U3, U4, M10, F13 and F14) were selected to be adjacent to the existing streamflow gauges in order to investigate the relationship between water quality and water quantity. The remaining sampling locations were chosen to ensure adequate distribution throughout the stream network, and to make sure there was sufficient representation from areas with different land use activities - namely, agriculture, urban and forest. During the preliminary survey on 13-May-2002 only fourteen stations were investigated. A site located downstream along Chilliwack Creek was added later that month as a potential indicator of the combined overall water quality of the stream network before it entered the Fraser River. In addition, five sites were also subsequently added to ensure that there would be sufficient data from each of the urban, forest and agricultural land use categories for statistical analysis. 33 Samp Hit? Stations (lay 0 Agriculture © Forest O Urban O Spring Fed • Mxed Monitoring Stations -fr Rain gauge Tipping buc ket rain gauge A Hjriiometric stations / V Streams Contour lines (100m) Chilliwack Creek Watershed Figure 4.1 Water and Sediment Sampling Stations, Chilliwack Creek Watershed 2002-2003 34 Each sampling site was classified into one of five land use categories: agriculture, urban, forest, mixed, or spring-fed. Note that the 'urban' category for sampling stations refers to hillslope tributaries that drain the new residential development of Promontory; and that any streams draining urban lowland areas are included in the 'mixed' category. Classification was determined based on the dominant land use immediately upstream of the stations, and is summarized in Table 4.1. For clarification, the land use classification of each station is indicated by the letter preceding the sampling station number (A = agriculture, U = urban, F = forest, G = spring fed, M = mixed). In total, six sites were classified as urban influenced, six as agriculturally influenced sites, three sites were considered forested, and four sites were classified as mixed land use. From Figure 4.2 it is obvious that the percentage of urban, forest, and agricultural land varies not only between land use categories, but between sampling stations of the same land use category. Within the agriculture category, stations A2, A16 and A18 are the most intensively used agricultural areas. Station A l 1 is located directly adjacent to a tree farm operation, and was included in the agriculture land use group despite having less than twenty percent of its contributing area under agriculture. It should also be noted that if we consider the entire area upstream of a sampling site both stations A17 and A18 receive water from the sub-urban residential area of the hillslope; and consequently, are not solely influenced by agriculture. Urban stations U5 and U3 have a larger percentage of their drainage area covered by urban activities than the other urban sampling stations (U4, U7, U8 and U6). The three forested sites (F12, F13, and F14) were relatively undeveloped and, consequently, were chosen as control sites to represent streamwater quality before any form of contamination would be introduced to the stream by land use activities. Any agricultural activity contributing to these stations is predominantly small hobby farms. Station G l , located on Luckakuck Creek, is spring fed and was therefore used as a measure of the general groundwater chemistry in the area. 35 Table 4.1 Watercourse Classification for Sampling Stations (Sampling stations are arranged upstream to downstream for each watercourse). SAMPLING STATION W A T E R C O U R S E L A N D USE M19 Chilliwack Creek Mixed M20 Chilliwack Creek Mixed A15 Interception Ditch Agriculture A16 Interception Ditch Agriculture A18 Interception Ditch Agriculture F13 Elkview Creek Forest F14 Parson's Brook Forest F12 Parson's Brook Forest A l l Armstrong Ditch Agriculture M9 Teskey Way Ditch Mixed A17 Teskey Way Ditch Agriculture M10 Bailey Ditch Mixed 117 Lefferson Creek Urban U4 Lefferson Creek Urban U5 Teskey/Thorton Creek Urban U3 Teskey/Thorton Creek Urban U8 Walker Creek Urban U6 Benchley Creek Urban A2 Semiault Creek Agriculture G l Luckakuck Creek Spring fed B Agriculture • Natural • Urban • Under Development estimated to have ~ 10% imperviousness during sampling period (2002-2003) Figure 4.2 Percent of Land Use Activity within each Contributing Area (based on 2002 land use map) 36 4.2 Field Methods 4.2.1 Surface Water The field sampling took place over the period of one hydrologic cycle to include both wet and dry conditions. Streamwater grab samples were taken on nine dates (approximately monthly to bi-monthly) between May 2002 and July 2003, primarily during baseflow conditions. In addition, samples were collected during a storm event on 16-Oct-2003. Over this storm event, a series of four samples were taken (approximately every 75 minutes) at four different sampling stations (A2, U3, F12 and A18) within the catchment. Station M20 was sampled once near the end of the storm sampling. Grab samples were taken in acid washed polyethylene bottles, and stored in a cooler with ice until analysis. Nitrate-N + nitrite-N (N0 3" -N + N0 2 " -N), orthophosphate-P (P0 4 3" - P), and ammonia-N (NH3- N) were then analyzed in the lab within 24 hours of being sampled. On three of the eight sampling dates (27-May-2002; l-May-2003; and 9-July-2003), part of each grab sample was separated and later analyzed for various trace elements: aluminum (Al), calcium (Ca), cadmium (Cd), cobalt (Co), chromium (Cr), copper (Cu), iron (Fe), potassium (K), magnesium (Mg), manganese (Mn), sodium (Na), nickel (Ni), phosphorus (P), lead (Pb) and zinc (Zn). Preparation and laboratory techniques are described in Section 4.3.1. Parameters that were measured in situ include specific conductivity and water temperature using a Yellow Springs Instrument Co. (YSI) Model #30M/50 FI meter, and dissolved oxygen (DO) using a YSI Model #58 meter. pH was analyzed in the lab using a Beckmann 44 pH meter. Due to problems with broken probes, data for all these parameters could not be collected on every sampling date. Table 4.2 shows the parameters measured on each sampling date. 4.2.2 Sediments Samples of stream bed sediments were collected twice during the dry season, on 10-Oct-2002 and 9-July-2003. The sampling was done twice in order to observe any variation between years, and because the first sampling did not include four stations (A2, M10, M19 and M20). At each site, sediment samples were collected from the surface sediment layer using a 4 m pole with an aluminum pot attached at one end. Each sample was then place in a plastic bag, and refrigerated in the lab until analysis for trace elements, particle size, orthophosphate and degree of phosphorus saturation (DPS) was completed (see section 4.4.2 for details on the laboratory analysis). 37 Table 4.2 Summary of Water Quality and Sediment Parameters Measured on Each Sampling Date Sampling Date P A R A M E T E R A N A L Y T I C A L o © • o <=> o o © o • o E Q U I P M E N T 6 s •May • s < Oct-l -Nov -Dec--Mar •May 1 "3 1-9 1 r~ i-H «s o i—1 t-l <s m o l-H O © M Ml R: IlllliS Nitrate-N Lachat X X X X X X X X Ammonia-N Lachat X X X X X X X X Orthophosphate-P Lachat X X X Specific Conductivity Probe (DM) X X X X X X X X X PH Probe (DM) X X X X X X X X Temperature Probe (DM) X X X 3 X X X X Dissolved Oxygen Probe (DM) X X X 3 X X Trace Elements (Dissolved) ICP X X X SEDIME.VI: Trace Elements (Total) ICP X 5 X Particle-Size X Orthophosphate Lachat X Total Carbon LECO X 5 X Total Nitrogen L E C O X 5 X No data for site A l 1 (stream dry) 2 Preliminary sampling date (no data collected for stations U5, U7, F12, A15, A17, M20) 3 No data for stations A2, M10, F13, A15, A16, A17, A18, A19 4 Ammonium oxalate extractable elements 5 No sampling for stations A2, M10, Ml9 and M20 6Lachat - Lachat QuickChem FIA+ 8000 Flow Injection Analyzer, ICP - Vista Pro CCD Simultaneous Inductively Coupled Plasma - Atomic Emission Spectrometry, LECO - LECO CNS-2000 furnace 7 DM - direct measurement 4.2.3 Precipitation and Streamflow Data Flow data from a number of hydrometric stations throughout the watershed was obtained from the City of Chilliwack. The dates of operation for each hydrometric station can be found in Table B.l (Appendix B), and the locations of these stations are shown in Figure 4.1. Each streamflow sampling site has an automated ISCO 4150 Flow Logger, which collects data at 15 minute intervals. The sensor on the meter uses Doppler technology to measure the average velocity of the flow in the stream, while an integral pressure transducer measures the liquid depth to determine the flow area. The flow logger then calculates the flow rate by multiplying the area of flow in the stream by its average velocity (ISCO, 2004). 38 Table B.l also gives the channel characteristics and type of conversion used in the flow calculation for each station. Precipitation data were obtained from one of two sources. Table B.2 shows the record period and sampling frequencies for the different rain gauges (Appendix B). Daily data were obtained from an Environment Canada climate station (station # 1101530) rain gauge which is located at the Chilliwack Airport in the lowland valley. In addition, the City of Chilliwack operates two tipping buckets (Promontory and Marble Hill) which are located on the hillslope; these tipping buckets collect continuous data at five minute intervals. Unfortunately, the data records from the tipping buckets are incomplete and sporadic; consequently, any analysis performed with the Promontory and Marble Hil l data sets should be interpreted with caution. Due to the major gaps in the data, use of upland precipitation records has been restricted to investigating storm events where the data exists, and no effort has been made to determine annual or monthly rates at these stations. 4.3 Laboratory Analysis 4.3.1 Surface Water Samples 4.3.1.1 Nutrients (Nitrate, Ammonia and Orthophosphate) In the lab on the day after sampling, the water samples were filtered through Whatman #42 (2.5 urn) ashless paper to remove any particulate matter. Dissolved nutrients (N03~ + NO2" - N, PO4-P3" and NH3-N) were analyzed on a Lachat QuickChem 8000 Flow Injection Analyzer. Total ammonia as nitrogen (NH3-N), nitrate as nitrogen (N03"-N), and orthophosphate as phosphorus (P04"3-P) were analyzed the day after sampling in the UBC Soils Department laboratory. Prior to analysis, the water samples were filtered using Whatman #42 ashless paper to remove any particulate matter. Dissolved nutrients were analyzed on a Lachat QuickChem 8000 Flow Injection Analyzer using method # 12-107-04-1-B for N0 3"-N (detection limit 0.10 mg/L), method #10-107-06-2-A for NH 4 +-N (detection limit 0.10 mg/L), method #10-115-01-1-A for PO4 (detection limit 0.02 mg/L). A brief description of these methods can be found in Appendix C. 4.3.1.2 Dissolved Elements in Water Samples used in the analysis of dissolved elements were filtered through Whatman #42 ashless paper, and concentrated trace metal grade H N 0 3 (~ 0.5 mL H N 0 3 per 50 mL sample) was then added as a preservative. Each sample was analyzed in the UBC Soils Department laboratory using a Varian Vista Pro (RadialTorch) CCD Simultaneous Inductively Coupled Plasma - Atomic Emission Spectrometry (ICP-AES). Table C.2 in Appendix C lists the calculated detection limits for each of the elements analyzed. 39 4.3.2 Sediment Samples Al l sediment samples were wet-sieved with distilled water using Tyler stainless steel sieves. A random sub-sample of the original sampled sediment was sieved using No. 230 mesh size to obtain the < 63u.m fraction. This fraction represents the clay and silt sized particles which tend to be the greatest accumulators of metals. The < 63|Jm fraction was then placed in a glass beaker and dried at approximately 75°C until the sediment was completely dry. The dried samples were then gently ground using a mortar and pestle (so as not to destroy the structure of the particles and to create additional surface area). The samples were then stored in plastic containers until analysis. Similarly, a separate sub-sample of the original sediment was sieved using No. 9 mesh to separate the fraction smaller than 2 mm. This was also dried at 75°C, ground using a mortal and pestle and stored in plastic containers until analysis. 4.3.2.1 Physical Properties Textural analysis was performed on the < 2mm dried sediment fraction using a simplified method for particle size determination discussed by Kettler et al. (2001). The rapid method outlined in this paper, which does not take into account the presence of particulate organic matter, was used because it was assumed that geological sediments have not had time to develop or accumulate an influential amount of organic material. In this method, 45 mL of sodium - hexametaphosphate (HMP)' ((NaP04)6) at an aqueous concentration of 3% by weight was added to 15 grams of the dried sediment sample in a 100 mL Berkman centrifuge tube. The centrifuge tube was then placed on a reciprocating shaker for two hours. The purpose of the sodium in HMP is to complex any Ca 2 + in solution and to replace Ca 2 + with Na + on the sediment particle. Sodium has a larger hydrated radius than calcium; this results in the dispersal of individual soil particles and the breakdown of soil aggregates (Kettler, 2001). Na + may also dissolve any organic matter binding particles (Lavkulich, pers. com. 2003). The phosphorus in the HMP binds to both A l and Fe (which are both major binding agents in soils) and precipitates them out of the solution. After dispersion, the sediment slurry was sieved through a standard 0.053 mm mesh (no. 270) sieve to separate the sand faction. The slurry was sieved until the liquid passing through the sieve appeared clear. The particles that did not pass through the sieve (the sand fraction) were collected and rinsed into a weigh boat and dried at 55°C to a constant weight. The percentage of sand is then calculated based on its fraction of the original mass. The solution and particles (silt and clay) that passed through the 0.0053 mm mesh were collected in glass beakers. The solution was then stirred with a glass stirring rod to ensure suspension of all sediment particles and left undisturbed for three hours. During this time the silt particles settled out leaving the clay particles in suspension. After three hours the solution of suspended clay particles was decanted and discarded. The settled silt fraction was rinsed into a weigh boat and dried at 55°C to constant weight. The percentage of 1 Calgon (detergent grade HMP) was used 40 silt was calculated based on its fraction of the original mass. The percentage of clay was calculated as the difference of 100 percent minus the sum of the percent sand and percent silt. 4.3.2.2 Total Carbon and Nitrogen The dried < 2 mm sediment fraction was passed through a 1 mm mesh (no. 16) sieve. Carbon and nitrogen content in the sediments were determined using a LECO CNS-2000 furnace. Approximately 0.5 g of sediment was placed in the sample holder. In the combustion chamber the furnace heat (1350°C ) and oxygen gas cause the sample to combust; which converts any elemental carbon and nitrogen into CO2, N 2 and NO x gas. These gases are then passed through the infrared (IR) cells to determine the carbon content, and a thermal conductivity cell to determine the nitrogen content. 4.3.2.3 Phosphorus Bioavailable phosphorus (measured as orthophosphate-P) was also analyzed. "Bioavailable P was determined using the Bray-Kurtz P-l method (0.025N HC1 + 0.03N NH4F). 20ml of extractant was added to 2.00 g of sediment and shaken for 5 minutes. The samples were then filtered through Whatman #42 filter paper and analyzed using a Lachat 8000-series QuikchemAE FIA colorimeter that employs the molybdate blue method (Murphy and Riley, 1962) to determine orthophosphate concentrations" (Li, 2003). 4.3.2.4 Trace Elements in Sediments Analysis for total trace metal concentrations in sediment was performed on the < 0.63 p:m dried sediment fraction. The U.S. Environmental Protection Agency (EPA) 200.2 metal digestion method was used for the digestion (Smoley, 1992). In the method, 1.00 g of well mixed <0.63[Jm dried sediment fraction was placed in a 125 mL glass beaker. An aqua regia solution (4 mL of 1:1 nitric acid (FINO3) and 10 mL of 1:4 hydrochloric acid (HQ)) was added to the sample using a calibrated pipette. Metals were then extracted from the sediment samples by covering the beaker with a watch glass and refluxing the sample in the dilute acid mixture for about 1 hour in an oven at approximately 80°C. The digested samples were cooled and then filtered through Whatman #42 filter paper. The filtered solution was then quantitatively transferred to a 100 mL volumetric flask and diluted to volume with deionized water. The digested sediment samples were analyzed on a Varian Vista Pro (Radial Torch) CCD Simultaneous ICP-AES. The sediments were analyzed for the following elements: A l , Ca, Cd, Co, Cr, Cu, Fe, K, Mg, Mn, Na, N i , P, Pb, and Zn. Final concentrations found in the sediments are presented as mg/kg dry weight for the silt/clay fraction. Table D.l located in Appendix D lists the detection limits for the analysis. 41 4.4 Quality Analysis and Quality Control 4.4.1 Site Variability To obtain an estimate of intra-site sampling variability, water samples were collected in triplicate at one or two stations selected at random on two of the eight water sampling days. The coefficient of variation (CV) for each set of triplicates was calculated, and the overall average CV was calculated as a measure of the intra-site variability for each parameter. Raw data and the results for replicate analysis are given in Appendix E. A l l parameters had an average CV below 8%, with the exception of ammonia-N. The exceptionally high (48.90%) CV value for ammonia may be due to sample contamination or natural variability. Within site variability was not measured for sediments. 4.4.2 Method Accuracy of Water and Sediment Analysis The accuracy of the water sample analysis was determined by measuring a series of standards (samples with known concentrations) every ten samples analyzed on the Lachat and ICP-AES during each sampling run. The accuracy of the digestion technique for trace elements was determined using measurements of trace element concentrations in certified reference sediments. MESS-1 marine reference sediment was analyzed twice with the October sampling set, while Priority PollutnT™/CLP (Lot No. DO35-540) reference sediment was analyzed once with the July 2003 sampling set. Only certified values for Cr, Cu, Ni and Zn were available for the MESS-1 sediments. Results indicate that sediment sampling recoveries for all elements were within performance limits for July 2003 analysis. For the October 2002 sediments, chromium, nickel, copper and zinc were all outside the 8% error range, with Cr (67.2%) and Ni (32.7%) having very poor recoveries. See Appendix E for a complete list of results. 4.4.3 Method Precision for Water and Sediment Analysis Replicate samples were sub-sampled from the dried and sieved sediment samples, and analyzed to provide an indication of the precision of the various analytical methods used. Precision was measured by calculating the percent difference between duplicate values, and then calculating the overall average percent difference for each parameter. Raw data and precision results for trace elements, orthophosphate (Bray-Kurtz P-l), and particle size can be found in Appendix E. In July 2003, copper and iron showed the highest variability in analytical results with an average percent difference between replicates of 17.3% and 12.2% for copper and iron, respectively. The high percent difference calculated for copper and iron for these data was primarily due to the presence of a high duplicate at site M9, which gave a percent difference of 33.4 and 24.2, respectively. Still, the high variability in duplicate analysis for these two metals suggests a lower confidence in Fe and Cu results. The 42 average percent difference of all other elements ranged from 0.63% for zinc to 5.91% for sodium in July 2003, and from 2.35% for phosphorus to 8.61% for sodium in October 2002. In general, the October 2002 sediments showed a higher variability in the analytical results for most elements, with the exception of copper and iron. Method precision results for textural analysis showed that there was higher variability in calculating the silt and clay fractions than the sand fraction. 4.5 Geographic Information System (GIS) Methodology 4.5.1 GIS Database Creation A GIS was used to document the biophysical resources in the watershed. For this study, ArcGIS® Desktop 8.2 (a product of ESRI Inc.) was used to aggregate and synthesize this database, and then to analyze spatial trends. The digital layers of the different watershed characteristics entered into the GIS database were obtained from a variety sources, shown in Table 4.3 below. The stream network, soils, roads, orthophotos, and a TIN (triangulated irregular network) showing the topography were obtained from the City of Chilliwack in digital form. Separate coverages were created for a number of other features, such as the location of sampling stations, hydrometric stations and precipitation monitoring sites in the watershed. Locations of these stations were approximated, using the orthophoto, road and watercourse layers as a guide. Table 4.3 Data Sources for the GIS Database Layer Source Stream network Digital, City of Chilliwack Roads Digital, City of Chilliwack Topography (TIN) Digital, City of Chilliwack Watershed/Contributing Areas Manually digitized from TIN, contours and drainage pipes/outlets Drainage pipes, outlets et cetera Digital, City of Chilliwack Soils Digital, City of Chilliwack Land Use(1995, 2000) Manually digitized from orthophotos Land Use(2002) Manually digitized from orthophotos and digital map provided by the City of Chilliwack Sampling sites Manually digitized Hydrometric stations Digital, City of Chilliwack Orthophotos (1995) Triathlong Inc. (digital) Orthophotos (2000) Digital, City of Chilliwack 43 Land use maps were based on the orthophotos and the land use map provided by the City of Chilliwack. The 1995 and 2000 land use information was digitized from 1:20,000 scale color orthophotos of the respective years. Current land use information was taken from a digital land use map for 2002 provided by the City of Chilliwack. However, due to inconsistencies in the polygons (i.e. overlaps, gaps) a new 2002 land use map was created. To do this, the 2000 land use map was altered using land use information from the 2002 digital map file. See Table 5.1 for a summary of the land use categories used to characterize the watershed and Figure 5.1 for a map of the current (2002) land use. Information on soils and surficial materials specific to the watershed was based on a 1:25,000 digital soil map provided by the City of Chilliwack. This map was originally digitized from 1962 Soil Survey Map of the Chilliwack Area (Comar, Sprout and Kelley, 1962). For simplicity, each soil unit was split into surface and subsurface layers and defined by geologic origin (parent material), dominant surface and subsurface texture, drainage and perviousness. The parent materials in the watershed included organic material, colluvial, eolian or fluvial deposits, and indicate the way by which the soil was carried, sorted and deposited. The surface texture was defined as the texture of the topmost discrete soil unit with a thickness of at least 10 cm. The subsurface layer was defined as the major textural class below the surface layer. Appendix A shows the classification categories for the soil and surficial material attribute data. 4.5.2 Contributing Area and Buffer Area Delineation In order to examine the relationship between land use and water quality, the contributing area for each sampling station was delineated. On the hillslope, delineation was based primarily on topography; however topographical features, such as larger roads where ditches can redirect runoff from the adjacent land, were taken into account. In the urbanized section of the hillslope storm drains can alter the natural drainage pattern by collecting water from the land and releasing it to the stream at specific points. As a result runoff does not always follow the topography of the catchment. In order to overcome this difficulty the location of drainage pipes and outfall locations (obtained from a digital file provided by the City of Chilliwack) were also used in determining the contributing areas in this part of the hillslope. In the lowland agricultural area, the topography is relatively flat, and the Triangular Irregular Network (TIN) was not always detailed enough to clearly delineate the drainage areas based on topography. When topographical data was not clear, the midpoints between watercourses were used to determine the contributing areas. A map of the contributing areas is shown in Figure 4.3, and Table F.l lists the receiving watercourse and the size of the drainage area for each of the contributing areas. 44 Figure 4.3 Map Showing the Contributing Areas for the Sampling Stations in the Chilliwack Creek Watershed No contributing area was delineated for station G l because it receives a significant portion of its water from the Sardis-Vedder aquifer. Groundwater movement does not necessarily follow topography since it is subject to the hydraulic properties of the aquifer, input to (recharge) and outflow from (discharge) the aquifer system, and the geological factors that may block or create flow paths. Consequently, the boundaries of surface water and groundwater drainage areas do not always coincide, and delineating contributing areas for the aquifer requires an understanding of the aquifer properties that is beyond the scope of this thesis. Water quality results from the sampling stations were also correlated with land use within a buffer zone surrounding the streams. The buffer tool in ArcMap® was used to create three different sized buffers (50 m, 100 m and 200 m) around the stream network. A map of the land use within the 100 m buffer is shown in Figure 8.1. Using the geoprocessing wizard in ArcMap® , land use information was extracted from the various land use coverages, and summarized by contributing area, buffer zones and the watershed as a whole. Both the total area upstream of the sampling station (cumulative area), as well as the area between sampling stations (independent area) were calculated for each station (see Appendix A). 45 4.6 Water and Sediment Quality: Data Analysis Methods The SPSS for Windows 12 software program was used for all statistical data analysis. Based on normality tests it was found that the water quality and sediment data were skewed (i.e. non-normally distributed) for most parameters. In addition, for some land use categories, the sample size was very small. For these reasons, non-parametric statistical techniques were used for all data analysis. These methods are less powerful than parametric procedures; however, they can be used when the assumptions required for parametric statistics are not met (Townend, 2002). 4.6.1 Statistical Analysis of Water Quality Data Water quality results were divided into 'wet season' (November 2002 to early May 2003) and 'dry season' (May to August 2002, and July 2003) based on differences in flow and precipitation. In total, there were four sampling dates for each of the two seasons. For each parameter an average value was calculated for both the wet and dry season and used to determine seasonal trends. The Mann-Whitney U test was used to determine if there were significant differences between the wet and dry season data. This test is the non-parametric alternative to the unpaired t-test, and tests whether two groups have the same median. Data for dissolved ions were only measured on three sampling dates, and a Kruskal-Wallis test (the non-parametric equivalent to the analysis of variance (ANOVA)) confirmed that the medians of the different dates were statistically similar (ct>0.05). Therefore, results for dissolved ions were not separated into wet and dry season, and no seasonal analyses were performed for the dissolved ions. Graphs and box plots were created in order to visualize spatial and seasonal trends. In these graphs, the wet and dry season averages are plotted for stations along the Interception Ditch mainstem to the mouth of Chilliwack Creek. Boxplots have the advantage of displaying the full range of data without requiring the parametric assumption of normality. In addition, they are useful in determining critical values at specific locations or on specific dates, which may be masked by the use of wet and dry season means. The boxplots for each variable are shown in Appendix C for each site over the sampling dates, for each sampling date by site. Boxplots by land use category are also shown. Boxplots represent the range from the lower bound of the second quartile to upper bound of the third quartile (a distance describes as the 'interquartile range' (IQR)), with the line between them marking the median. Data within a distance of 1.5 times the IQR of either the bound is noted by the extended whiskers. Measurements beyond more than 1.5 times the IQR from the median are considered outliers and are denoted by individual data circles "o". Values greater than 3 times the IQR are considered extreme values (represented by an asterisk "*"). For each parameter, water quality data were grouped by land use category (previously defined in Section 4.2), and a Mann-Whitney U test was used to make pair-wise comparisons (for wet season and dry 46 seasons separately and for both seasons combined) between the three dominant land use categories (agriculture, urban, and forest). When making multiple comparisons, a correction is needed to adjust for the increased chance of making a Type 1 Error (concluding that there is a significant difference between the populations of two groups when really there is not) in at least one case. Therefore, the Bonferroni method was use to adjust the level at which a result was considered 'significantly different'. This method divides the significance level that would normally be used (a = 0.05) by the number of possible pairwise comparisons (3) for a new significance level of a = 0.017. 4.6.2 Statistical Analysis of Sediment Quality Data A similar approach to that used for the analysis of water quality data was taken for the analysis of the sediment data. Boxplots were created for each parameter by land use categories (see Appendix D). The downstream graphs plotted the data for the two sampling dates along the mainstem and its tributaries. Sediment data was tested (by land use category and all data combined) using a Wilcoxon Signed Rank test (ot=0.05) to determine if the levels of metals had changed significantly between the two sampling dates. Stations A2, M10, M l 9 and M20 were omitted from this part of the analysis because they were not sampled in October. The Wilcoxon Signed Rank test was chosen over the Mann-Whitney U test because the samples were assumed to be dependent due to the fact that comparisons were being over time. Mann-Whitney U tests (a=0.0167 using the Bonferroni adjustment) were used to make pair-wise comparisons between land use categories (for both dates individually and for the pooled data set). 4.6.3 Relationships between Land Use, Sediment, and Water Quality As previously described, land use indices were calculated for each contributing area and within three different width buffer zones (50 m, 100 m and 200 m). A Spearman's Rank correlation test (1-tailed, a=0.05) was used to determine the relationship between these land use indices and the various water and sediment quality parameters. This test measures both linear and non-linear correlations and can be used even for small sample sizes (Helsel and Hirsch, 1992). A strong positive or negative correlation suggests a relationship between two variables, but does not determine whether changes in one of the variables actually cause changes in the other. This cause-and-effect relationship needs to be established separately (Townend, 2002). Correlations were carried out separately for each season, as well as for the combined data. Results were considered significant at p <0.05. 47 4.7 Water Quantity: Data Analysis Individual storm events between 2001 and 2003 were delineated for both the Marble Hil l and Promontory precipitation data. A number of rainfall characteristics were then calculated for each storm event: total precipitation (in mm); duration (in hrs); peak intensity (5 min, 15 min, 30 min, 1 hour, 3 hour and 6 hour) (in mm/hr); and antecedent dry period (in hrs). See Tables B.6 and B.7 in Appendix B. Storm classes were created based on total rainfall and peak 15-minute intensity (see Table 6.1 for classification), and the percentage of storms in each class was then calculated. For each storm event, the lag time (in hrs) from peak rainfall to peak runoff, and the peak runoff rate (in mm/hr) were calculated at each of the flow monitoring stations. Only storms with a total rainfall greater than 10 mm were used in this part of the analysis since it was assumed that runoff from smaller storms would be insignificant. The Wilcoxon Signed Rank test was then used to determine if statistically significant differences (a=0.05) existed between the different catchments for each storm response characteristics. Chilliwack and Semiault were omitted from this analysis. Boxplots were also created using all storms events at each station. It should be noted that, although they are shown in the same graph, the storm events represented in the boxplots are not identical for each station. Finally, hydrographs were created for six individual storm events to visualize the differences in the response of the different catchments. These events are described in section 6.3.2.2. 4.7.1 Delineation of Storm Events Continuous rainfall data (Promontory and Marble Hill) for a three-year period (2001-2003) was separated into a record of individual storm events, using a minimum interevent time1 (MIT) of 6 hours. Any period with less than 0.5 mm of rainfall over a 1 hour period was considered to be insignificant. The time series of 15 minute precipitation data from the Promontory tipping bucket was separated into 93, 66 and 51 independent events for 2001, 2002 and 2003 respectively; while the data from the Marble Hil l tipping bucket was differentiated into 99, 69 and 52 individual events for the same three years. 4.7.2 Calculating Storm Response Characteristics In this study, peak runoff rate calculations were based on the total streamflow volumes. Since baseflow was not separated and subtracted from the total streamflow volume to derive direct runoff, peak runoff 1 Minimum interevent time (MIT) is defined such that "rainfall pulses separated by a time less than this value are considered part of the same event" (Bedient and Huber, 1992) 48 rate may be more accurately considered to be peak discharge per unit area. In small, urbanized streams baseflow is often neglected because it represents such a small fraction of the total flow. However, in natural streams and larger rivers baseflow may be a significant fraction of streamflow due to the contribution along banks from the water table (Bedient and Huber, 1992). The equations used to calculate the lag time and peak runoff rate are defined below. A graphical depiction of the input variables is shown in Figure 4.4 Lag Time (in hrs) was defined as the time from peak rainfall to peak runoff: Lag Time (LT) = t(Qp) - t(Pp) Peak Runoff Rate (in mm/hr) was defined as the peak discharge per unit area. Peak Runoff Rate (R p)= Q p / A note: multiply value in m/s by (3600* 1000) to get a value in mm/hr (runoff rate) multiply value in m W m 2 by 105 to get a value in mV'/km 2 (peak discharge per unit area) • Time (t) j Figure 4.4 Graphical Representation of Storm Response Characteristics 49 The results of this study are discussed in the next four chapters. First, the current land use and trends over the last 8 years are discussed in Chapter 5. Chapter 6 focuses on the climatic conditions in the watershed and the spatial and seasonal variability in streamflow, as well as the extent to which the Promontory development has contributed to changes in streamflow. Next, the overall surface water and sediment quality is discussed in Chapter 7, with a focus on spatial and temporal trends. Because water and sediment samples were taken from three distinct land use areas, analysis was done to determine if the water and sediment quality from the different areas had any distinct chemical signatures, and to determine the impact of the new hillslope development. The next chapter (Chapter 8) uses the GIS database and water quality and sediment data to examine the interactions between water quality, sediment quality and land use. A brief discussion and summary of the results are found in Chapter 9. Conclusions and management recommendations are found in Chapters 10 and 11. 50 5 LAND USE IN THE CHILLIWACK C R E E K WATERSHED Land use in the study area was characterized for 1995, 2000 and 2002 with the aid of GIS. Table 5.1 describes the land use categories that were used to characterize the watershed. Note that because a more detailed land use map was already available there are additional subcategories for the 2002 land use map. Once incorporated into the GIS database, the GIS program was used to make quantitative comparisons in land use changes. This chapter will discuss the current land use in the watershed, and identify trends in land use change over the past seven years. 5.1 Current Land Use Figure 5.1a shows the various land use activities in the study area for the year 2002. The study area contains a mixture of different land use activities that are not evenly distributed throughout the watershed (see Figure 5.1). The eastern and upper portions of the hillslope are the least developed regions in the watershed; rural residences and hobby farms are scattered throughout the primarily forested hillslopes. Promontory, a new low-density residential development, is located on the lower western portion of the hillslope. This development accounts for 13% of the 'residential' land use in the study area. The lowland area covers approximately 75% (4154 ha) of the land base in the watershed. About 72% (2990 ha) of the lowland valley is part of the Agricultural Land Reserve (ALR), and is some of the most productive agricultural lands in the province. Soil capability for agriculture is frequently Class 2 and 3 in the valley, and upland soil capability is generally Class 4, 5 and 6 (City of Chilliwack, 1998). Dairy farming/beef cattle production and crop farming are the most common agricultural land uses in the watershed. While the data collected did not distinguish between the type of crop cover (all crops were grouped under 'arable'), based on observations during field excursions it was obvious that corn and forage are the most commonly grown crops. In the lowland agricultural area, cattle and arable land account for 49 and 47% of the agricultural practices, respectively. There is also a small presence of poultry operations, as well as greenhouse and horticultural developments in this region of the watershed. This agricultural land base confines the urban centers of Chilliwack and Sardis to the eastern sections of the Chilliwack watershed. Overall, urban land use covers about fifteen percent of the land base in the study area. Of this, 72% is found in the urban-corridor, 16% in the lowland agricultural area, 11% in the Promontory development, and less than 1% on the remaining hillslope (Ryder Uplands) (see Table A.4 in Appendix 5). Tables A.4 and A.5 in Appendix A summarize the land use in the watershed. Although the data are not discussed by individual contributing areas, this data can also be found in Appendix A. 51 Table 5.1 List and Description of Land Use Categories (^indicates additional sub-categories used for 2002 land use map only) L A N D USE L A N D USE SUBCATEGORY DESCRIPTION Arable* Land cultivated under crops, grain, fruits/berries, or pasture. Livestock* Cattle Land used for cattle (dairy or beef) Poultry Land used for poultry production. AGRICULTURE Horticulture* Land used to produce flowers or trees, excluding greenhouses. Greenhouse* Includes all areas under greenhouse production. Hobby Farm* Includes field area of hobby farms. Unused Agricultural Land* Agricultural land that is not currently under cultivation. Transportation Includes all major roads (highways), and railways. Smaller roads were split down the middle and included with adjacent land uses. Residential Low Density* High Density* I/D Low Density* URBAN I/D Medium Density* Industrial/Commercial (I/D) and Institutional (Inst.) I/D High Density* Other* Includes non-residential, non-commercial/industrial activities such as the airport and municipal dump. Inst. Med/Low Density* Inst. High Density* Rural Residential (Estate) Includes any rural house on the hillslopes, as well as residential homes in the agricultural lowlands (including any adjacent land not being used for agriculture). Recreation Includes grass areas, playing fields, and parks not included under 'wilderness parks'. GREENSPACE Open Space1 Shrubs Large areas of shrub vegetation, including vegetation along streams. Minor landscaping was not included. Unused Open Space Areas of land that are unused and that are not forested. Clearcuts Areas that have been logged/cleared for non-development purposes. May be partially re-vegetated. Forest Wilderness Parks A l l parks that are primarily forested/natural vegetation. Forest Includes all forested land on the hillslope, as well as larger clusters of trees. UNDER D E V E L O P M E N T Cleared for Development Includes areas that have been cleared of vegetation for development purposes, but have yet to be developed. W A T E R Water Areas under water (lakes, ponds, large rivers). Small creeks, streams and ditches are not included. 'The open space category was created to represent all non-agricultural clearing, and consists of any open space that is covered in grass or low lying shrubs, including parks and playing field, unused open space, as well as areas that have been clearcut. Rural residential land use was also classified as 'open space' because most of the land area associated is usually grassy open space, and because it was assumed that the inputs would different than from either urban or agricultural land uses. b) Four distinct regions in the watershed Urban Corridor Promontory Ryder Uplands Hillslope ALR L a n d U s e C a t e g o r i e s Agriculture: Urk.m: 4 ^ Arable Residential- low density Cattle ^ ^ R e s i d e n t i a l - med/high density Poultry ^ ) Industrial/Commercial - low density <Z25 Ho rticu Itu re/G re enh ou se 4 ^ Industrial/Commercial - medium density C3> Hobby Farm Industrial/Commercial - high density CZ> Unsued agricultural land 4^ Industrial/Commercial - other Institutional - low/medium density Natural: Institutional - high density 4 ^ Wilderness parks Transportation 4 ^ Forest <Z^> Rural Residential Other: C Z D Recreational Under Development CZ^> Shrubs O Water C U D Unused open space <CH> Clearcut Figure 5.1 Land Use in the Chilliwack Creek Watershed Study Area, 2002 5.2 Trends in Land Use Land use activity, spatially illustrated for 1995, 2000 and 2002 in Figure 5.3, has changed only marginally when the entire watershed is considered (Table A.5). The two largest land uses (agricultural and forest) both decreased slightly over the seven year period (by 4% and 5%, respectively). Urban residential and rural residential land uses have increased by 27% and 5%, respectively. In general, the presence of the ALR has ensured that land in lowland valley remains agricultural in nature, and consequently, there has been very little change in land use activities in this region. However, a few small residential areas have been built in the lowland (e.g. at the confluence of Interception Ditch and Chilliwack Creek). While the area of agricultural land has not increased, due to the high quality of the valley agricultural land and the proximity to a large urban population and food processing plants, agricultural activities in the regions are becoming increasingly intensive. Agricultural intensification results in greater inputs of nutrients, pesticides and other contaminants, and a greater potential for ground and surface water contamination. Small areas of undeveloped land in the urban corridor are continually being developed, particularly in Sardis and on Little Chilliwack Mountain. Past trends show that there has been a gradual shift from predominantly single family homes to a greater mix of single and multiple family housing; and future estimates suggest that this trend will continue (Figure 5.2). T | , 1998 2003 -2026* 1998 2003 -2026* B Single Detached • Townhouse • Apartments B Other •Based on OCP Scenario of 85,000 population threshold. Figure 5.2 Housing Development in the Chilliwack Urban Corridor by 'Type' of Dwelling (City of Chilliwack, 2003) 54 Figure 5.3 Land Use Changes (1995-2002) in the Chilliwack Creek Watershed The most substantial land use changes in the watershed were in the Promontory hillslope development area (see Figure 5.3). In 1995, the suburban community in the Promontory region was in the initial stages of development, with 18.4 ha of existing residential area and 47.7 ha of land that had been cleared for development purposes. By 2002, most of the land that had visibly been cleared for development was residential housing or unused open space. Overall, the land base dedicated to residential homes had increased by a factor of 4.3 to 78.2 ha, while the area of unused open space increased by a factor of 2.5 (or 33.9 ha). Figure 5.4 shows a detailed breakdown of these changes. Agric. Res. Ind./Com. Open Rural Rec. Unused Forest Under Space Res. OS Dev. Figure 5.4 Land Use Changes 1995-2000: Promontory Development Region of the Chilliwack Creek Watershed Study Area Currently, most of the housing developments in the Promontory development are large, single family dwellings (Figure 5.5). In addition, the subdivisions were generally built with wide roads and driveways, resulting in a significant increase in the impervious surface area. Urban development on the hillslope has been rapid and is expected to affect the area through increased contaminant inputs and re-engineering of the hillslope tributaries. Interception Ditch and the downstream agricultural area are particularly at risk as increases in the effective impermeable area in the upstream systems could potentially alter their flow regime depending on stormwater management practices. 56 41 4% 208 20% • Single Detached • Duplex • Townhouse • Apartment 19 2% Total Number of Dwelling: 1,017 Total Population: 2,821 Average Household Size: 2.49 Source: City of Chilliwack (2003) Figure 5.5 2003 Dwelling and Populations Estimates for Promontory 5.3 Impervious Surface Area Impervious surface area is a useful indicator for water management of urban areas because it is easy to measure. Numerous studies have shown a link between impervious surface area and the degradation of aquatic systems, with strong correlations found between hydrology, loadings from NPS pollution, biological integrity (Schueler, 1992; Booth et al., 1993; Schueler, 1994; Arnold and Gibbons, 1996). In this study, the total impervious surface area (TIA) for each contributing area was determined indirectly from a land use/land cover map. The calculation involved three steps: 1) calculating the amount of each land use category within the contributing area, 2) multiplying each total area within each category by an 'imperviousness factor' typical for that category (see Appendix A for the imperviousness factors used), and 3) summing the results for all land use categories to get an overall TIA value. The accuracy of the results will be influenced by the selection of appropriate imperviousness factors and accuracy and scale of land use mapping. The total impervious area of each contributing area is summarized in Table A.7 and illustrated in Figure 5.6. Within the watershed the %TIA per contributing area ranges from about 2 to 40 percent, with the lowest values predominantly in the agricultural area and forested sections of the hillslopes. The contributing areas for Teskey Creek and Chilliwack Creek have %TIA near 30 %, the value at which streams become completely degraded (Besbier et al., 2000; Arnold and Gibbons, 1996). Another four contributing areas (A 18, U4, M19 and U5) have values above 10% TIA (the threshold above which stream degradation has been shown to begin). The %TIA within buffer zones was generally similar, 57 except at station A18. While this station had over 10% TIA in the contributing area, only 4 % of the buffer zone is impervious surface area. It should also be noted that %TIA for stations M10 and A17, which are below the hillslope development, are considerable greater when the entire cumulative area above the station is considered. Hillslope Urban • Contributing Area • 100 m buffer O l O CD 0 0 i -2 < < < < < Figure 5.6 Percent Total Impervious Area by Total Upstream Contributing Area 58 6 CLIMATE AND HYDROLOGY The hydrology of an area reflects the interactions between climate, physiographic factors, geology and vegetation, as well as human activity (Moore, 1991). Water inputs to the surface are determined by the climate (e.g. amount, intensity and distribution of rainfall), while the other factors control the subsequent partitioning of inputs into overland flow, soil moisture, groundwater and streamflow (Moore, 1991). Modifications of land surface (particularly during urbanization) can produce changes to the hydrologic characteristics of the land surface and modify pathways and rates of water flow. As a watershed is developed, its surfaces are made less pervious and natural channels straightened or hardened (lined with concrete); as a result there is typically more rapid runoff leading to higher peak discharge and total runoff volumes (Leopold, 1968; Anderson, 1968; Dunne and Leopold, 1978). Ultimately, these hydrologic impacts associated with urban development have serious adverse effects on the environment including: channel erosion and widening, loss of groundwater recharge, decreasing ecological diversity of the aquatic community, and downstream flooding (Booth, 1990; Konrad, 2000; Schueler, 1992; Schueler, 1994). Our knowledge of hydrologic response is strongest in homogenous catchments, yet many resource management problems are focused on heterogeneous drainage basins where land surface characteristics are changing over time (Zhang and Smith, 2003). The continuing urban development of the hillslopes in the Chilliwack Creek watershed is especially important in terms of stormwater management problems, and it is therefore critical to understand and manage the hydrologic impacts of this drastic change in land use. This chapter will discuss both the climatic factors and trends that may influence streamflow in the Chilliwack Creek watershed, as well as the spatial and seasonal variability in streamflow. Of particular interest is the extent to which land use changes taking place on the hillslopes are contributing to changes in streamflow. 6.1 Climatic Characteristics 6.1.1 Temperature Temperature data collected at the Chilliwack regional airport (Environment Canada climate station # 1101530) were obtained from the National Climate Data and Information Archives operated by Environment Canada (EC). Figure 6.1 shows the minimum, mean and maximum normal monthly temperatures for the 1971-2001 period. The mean annual temperature is 10.5°C at the Chilliwack regional airport (11 masl). January is the coldest month with mean a temperature of 2.3°C July and August are the warmest months with mean temperatures of 18.5°C and 18.4°C, respectively. 59 40 35 30 o o 25 0) 20 i _ £ 15 a 10 E a> H 5 0 -5 -10 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 6.1 Monthly Mean Maximum and Minimum Temperatures for the 1971-2000 Normal Period, Chilliwack Climate Station (EC, 2003) There also appears to be a warming trend between 1911 and 2001 of about 2.1°C (Figure 6.2). Moore (1991) found a similar increase in temperature at the Agassiz climate station from about 1970 on. Without further analysis, it is not possible to say whether this shift is climate related (e.g. global warming) or whether it is due to other influences. 13 i 12 —•— Mean Annual Temperature —•— 11 year moving average Figure 6.2 Mean Annual Temperature (°C) and 11-Year Running Means for the Chilliwack Climate Station (EC, 2003) 60 6.1.2 Precipitation 6.1.2.1 Temporal Trends As shown in Figure 6.3 below, precipitation exhibits substantial inter-annual variability; and as a result, it is difficult to define trends in the data with any certainty. However, although not statistically verified there appears to be a slight increase in precipitation as well as year to year variability since the early 1 9 0 0 s -3000 i 1 500 0 Ii i i i i i i i i i i i i i i i i i i ii i i i i i i i i i i i i i i i i i i i i i i i i i i ii ii i i i i i i ii i i i i i i i n i i i i i i i i i i i i i i . i i i i i i i i ! ^ - < 0 ^ " C O T - C O T — C O * — CO * - CO * ~ CO T— iO f— fO r-i - i - c v i c N c o c o - t - ^ r m m c o c D r ^ r ^ c o c o c n c n o C f t 0 5 0 5 0 > 0 5 C f t C D O > 0 5 0 5 0 > C y ) 0 0 > 0 0 0 ^ 0 ^ 0 w- * - « - T- •*- r- *- f -r~ -e- •*- * - *- f- If m~ w- ©| -•—Total Annual Precipitation —•— 11 year moving average Figure 6.3 Total Annual Precipitation and 11-Year Average for the Chilliwack Climate Station (EC, 2003) 6.1.2.2 Seasonal Variations Monthly precipitation averaged over a thirty-year period (1971-2000) is shown in Figure 6.4. The 30-year mean annual rainfall at the Chilliwack airport was 1787.8 mm, with a maximum of 271.8 mm in December and a minimum of 54.3 mm in July. An average of 72% of the precipitation at this station falls between the months of October and March; however, this trend was not observed during the sampling period itself (May 2002 to July 2003). The summer and fall of 2002 was an exceptionally dry period, which was followed by an extremely wet period. These extremes can be seen in the cumulative precipitation graph in Figure 6.5. In this figure the flatter sections represent periods with little precipitation, and the steeper sections show the wetter periods. October, in particular, had much lower precipitation than the 30-year average for that month. For this reason the dry period included October sampling. Total precipitation for August 2002 (22.5 mm), July 2003 (22.0 mm) and August 2003 (9.9 mm) were also much lower than the 30-year mean. Conversely, October 2003 (364.7 mm) and November 2003 (278.7 mm) had considerably more precipitation than the corresponding 30-year averages (Figure 6.6). Overall, the study period was a period of great variability and extremes. 61 • Snowfall (mm water eq.) • Rainfall (mm) Figure 6.4 Mean Monthly Precipitation for the 1971-2000 Normal Period, Chilliwack Climate Station (EC, 2003) 900 Figure 6.5 Total Monthly Precipitation for the Study Period (May 2002 to November 2003), Chilliwack Climate Station (EC, 2003) 62 2250 Figure 6.6 May 2002 - November 2003 Cumulative Precipitation; and Average Cumulative Precipitation. Chilliwack Climate Station (EC, 2003) 6.1.2.3 Spatial Variability Figure 6.7 shows daily rainfall hyetographs for the different stations throughout the watershed over the study period. Using the Wilcoxon sign test, significant differences were found between the three stations. Unexpectedly, precipitation recorded at the Chilliwack climate station was significantly greater (p<0.05) than the precipitation recorded at both upland stations (Promontory and Marble Hill). The difference may attributed be to the fact that tipping buckets are located on the north facing slope, to an undercatch of the tipping buckets or to actual variability in storms. Precipitation at Marble Hil l was found to be slightly greater (p<0.05) than precipitation at the Promontory station. 63 E E c o ro o. o OJ k-Q. >-60 50 40 30 20 O 2 10 Marble Hill Tipping Bucket u * * i i i 4 | L L i , 4 Water Sampling Date ; Water and Sediment Sampling Date 4-f No data * Refer to Figure 6.9 for location of precipitation monitoring stations C\J CM CM C\J CVJ C M C M C O C O C O O C O c O C O £2 cn o co 9 9 CO u o cu Z Q -> 2 2 2 roo.ro 5 < 2 9 9 CD a R > co z 60 E 5 0 E, C 40 o * CL Promontory Tipping Bucket 30 a. >> 20 '5 Q S 10 o u I I I I I I I I l f l E c N C N c ^ c N C N ^ C N i S w n n n c g m c o c o c o c o 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 « -8 | £ S1 § = ? 6 8 Z Q 2 1 O Z Figure 6.7 Daily Precipitation at the Various Monitoring Stations throughout the Watershed for the Study Period (May 2002 to November 2003) 64 6.1.2.4 Distribution of Daily Precipitation The City of Chilliwack's performance targets for stormwater retention, detention and conveyance are based on the regional mean annual daily rainfall (MAR), which has been calculated to be 60 mm (CH2MHill, 2002a). As previously mentioned daily rainfall that is less than 50% M A R (< 30 mm) is captured on site; the next 30 mm is detained and released at the natural interflow rate; and storms greater than the M A R (> 60 mm) are conveyed directly to the streams (Figure 3.7). Daily precipitation data from 2000, 2001 and 2003 were used to calculate the average annual distribution of daily precipitation relative to the three categories described above (< 30 mm, 30-60 mm, and > 60 mm). The percent of the total annual volume for each category was also calculated (Figure 6.8). Precipitation data for 2002 were excluded from the analysis because the data for December were missing at the Chilliwack climate station. Note that values were calculated for all three rain gauges; however, Promontory and Marble Hil l values come from incomplete data sets. Still, it is obvious that under Chilliwack's new stormwater management plan most of the total rainfall is to be retained at the source and relatively little rainfall is to be conveyed to the outlet of a development site. Similar results were found in the study by CH2MHil l (2002a). 65 180.0 158.0 128.3 • • • H i <30 7.0 1.7 2.0 30-60 Rainfall Event* Size B Chilliwack a Promontory • Marble Hill 2-7 0.0 0.3 >60 * rainfall event defined as total daily rainfall depth ** Average of2000, 2001 and 2003 rainfall data 82 •85-' ' 9 m m m <30 B Chilliwack E Promontory • Marble Hill 17 1« 13 30-60 Rainfall Event* Size >60 Figure 6.8 Distribution of the Number of Annual Rainfall Events (top graph); and the Distribution of the Annual Rainfall Volume (bottom graph) 66 6.2 Streamflow Distribution Continuous streamflow measurements have been made at eighteen stations throughout the Chilliwack regional district, at varying times since about 1997 (Figure 6.9). In addition, a baseflow discharge survey was performed in October 2002 at various points along Interception Ditch, Semiault Creek, Chilliwack Creek, Atchelitz Creek and Luckakuck Creek (O'Byrne, 2002). This study focuses on the eight hydrometric stations that are within the study area itself. Each of these eight hydrometric stations has a drainage area of less than 3 km 2, with the exception of the Chilliwack Creek station (55.4 km2) and the Semiault Creek station (10.2 km2). Land use properties vary between the catchments. Teskey and Lefferson have experienced urbanization over the past eight years. These catchments contain impervious surfaces distributed throughout the primarily residential developments. Some of the more recent development areas were developed using low impact design technologies; however this study was done when conventional design practices were being used. Parsons and Elkview drain relatively undeveloped catchments and are used as control stations for the study. Bailey station lies at the base of the hillslope and receives water from both the urbanized upstream catchments as well as from a less developed area of the hillslope. Luckakuck Creek is spring fed, but stormwater outlets from the Sardis area are discharged into the stream above the hydrometric station. The Semiault Creek station drains the large section of the lowland agricultural area. The final station lies downstream of all these areas towards the mouth of the Chilliwack Creek. 6.2.1 Variability in Streamflow Daily discharge varies greatly over the year. In general, the winter portion of the hydrograph is characterized by a succession of storm peaks corresponding to the passage of frequent storms, superimposed on a high base flow. Conversely, summer base flow is significantly lower. Discharge from the base flow survey carried out in October 2002 is shown in Appendix B (Table B.3). Base flow values range from 0.007 m3/s at station C3 to 2.880 mVs at station CI (both along Chilliwack Creek). For all watercourses, base flow increases in the downstream direction with one exception: the furthermost downstream station along Semiault Creek (SI) is almost ten times lower (0.063 m3/s) than station S2 upstream (0.582 m3/s). Water withdrawls for irrigation during the summer months is likely a major cause of the lower baseflow observed downstream. During the study period, maximum daily flows range from 0.048 m3/s at Lefferson to 4.193 m3/s at the Chilliwack station (Table B.4 in Appendix B). The variation in daily discharge per unit catchment area for the different sub-catchments is shown in Figure 6.10 and summarized in Table B.5 (Appendix B). It was not possible to calculate or compare monthly or annual discharge for the different stations because the data series was incomplete and had an unequal distribution of gaps between the different data sets. 67 C O I U Kilometers 1 SiSwStSQ^SMM 4 k m i M i g i M i l U 5 B u k y P t d t g f t M t M . O M M X m u l "31531 204.71 43646 I n * * * * t i * 9 B n l > . CMi««ao»*iaY»m. M m Ml am OSES U M t l l K a t o 11.3003 Hxrmbi 30, S » t E K o t a X.20M McUtMlOmk r«hn«y8.36CBt> DK»W 31.3003 9 M O M I I A r t 17,1 IT. EOT O f f e r 313033 10 acmtgmt 11 H a u l 13 i f c C n A a f c — i W f h 13 M M 14 M f N d l 13 g B g H f e 16 A d x M s C n A — B j r a r i W W l i HMt> t > « W 51,i(B3 18. m u D K B l x 8,3003 MylO.M0OlaMtyai.X03 17 S a a taut Owk 9 Rui c«m M 18 j g g S j j t | 130 S wl 161600 ram W r a c y 6 . X (B« i b t c » W 3l.3003~ t w y l 3 t t i t t . t f r f b j 3 1 . X 0 3 D . f b * X . X O l K D K B b l 31.3003 $ Base Flow Survey Stations A Hydrometric Stations / % / Streams / \ / Roads L - A Chilliwack Creek Watershed C 3 C3 Elk Creek Watershed Ryder Creek Watershed 1) 10 m aslvsas used lor a l I ovfand stations 2) Incomplete data over t ie record period, at most stations Figure 6.9 Location of Streamflow Measurements within the Chilliwack Creek, Elk Creek and Ryder Creek Watersheds Luckakuck Creek (GW) Bailey Ditch (M) —i o to 00 CS O S 2 CO <N Q E 35 30 25 1 20 15 10 5 0 Study Period Semmihault Creek (A) 2000 2001 2002 o. < £ « 5-O "3 < O 3 £ 4> E? 03 U 2 c/l O S 2 Q £ OS 2 e 30 25 15 c 2 10 00 03 U 2 S 2 c 2000 ! 2001 ; 2002 2003 r J L u 35 30 25 20 ] Q E 15 < ^ 10 5 Q. < a < 3 $ - 0 -> < Chilliwack Creek (M) 2000 2001 2002 3 £ o i l CL < Figure 6.10 (cont.) 2000-2003 Total Daily Runoff (mm) for the Various Sub-Catchments in the Chilliwack Creek Watershed 6.3 Hydrologic Response to Storm Events A stream's response to a rainfall event is influenced by many factors, including the intensity and duration of the storm, the topography of the basin, land use and land cover properties, and the hydrologic conditions preceding the storm event (e.g. antecedent soil moisture). In this section the storms are first classified based on the precipitation data. The hydrologic response of the Chilliwack Creek basin is then examined, with particular emphasis placed on contrasting the response of the suburbanized and the forested sub-catchments. 6.3.1 Distribution of Storm Events The gaps in the precipitation record at both Marble Hil l and Promontory stations preclude the determination of complete annual distributions. Instead, the cumulative frequency distribution graphs were created for each of the storm variables (shown in Figure B.8, Appendix B). These graphs represent storms that fall within the available data from the combined 2001-2003 record period. The cumulative distributions were generally similar for Marble Hill and Promontory. Three storm classes were then defined (minor, intermediate and major) in relation to total rainfall and peak 15-minute intensity (Table 6.1). Table 6.2 presents the seasonal distribution of storm classes at both tipping bucket sites. Table 6.1 Definition of Minor, Intermediate, and Major Storm Classes in Relation to Total Rainfall (mm) and Peak 15-minute intensity (mm/hr) Peak 15 minute Rainfall Intensity (mm/hr) <5 I 5-10 I >10 Total Rainfall (mm) <10 Minor 10-30 Intermediate >30 Major 71 Table 6.2 Distribution of Storm Events in Three Storm Classes (Minor, Intermediate, Major) at Two Sites (Promontory and Marble Hill) for the Wet and Dry season SITE Y E A R T O T A L M I N O R INTERMEDIATE M A J O R % D A T A R E C O R D MISSING WET DRY WET DRY WET DRY WET DRY WET DRY PROMONTORY 2001 57.0% (53) 43.0% (40) 50.5% (47) 34.4% (32) 6.5% (6) 7.5% (7) 0.0% (0) 1.1% (1) 17.3% 24.2% PROMONTORY 2002 65.2% (43) 34.8% (23) 60.6% (40) 33.3% (22) 4.5% (3) 0.0% (0) 0.0% (0) 1.5% (1) 11.2% 37.4% PROMONTORY 2003 94.1% . IS ) 5.9% (3) 82.4% (42) 5.9% (<> 11.8% 101 0.0% (0) 0.0% (0) 0.0% (0) 27.7% 62.6% PROMONTORY Overall average <>8.6% (114) 31.4"., (AC) (.1 4% (12'). 27.1% 157) 7.1",, (15) 3.3% (7) 0.0% 10) 1.0%;. 12) 18.7% 41.1% MARBLE HILL 2001 62.6% (62) 56.5% (39) 75.0% (140) 37.4% (37) 43.5% (30) 25.0% (80) 56.6% (56) 31.3% (31) 6.1% (6) 6.1% (6) 0.0% (0) 0.0% (0) 11.2% 26.8% MARBLE HILL 2002 44.9% (31) 34.8% (24) 10.1% (7) 7.2% (5) 1.4% (1) 1.4% (1) 45.4% 25.2% MARBLE HILL 2003 65.4% (121) 21.2% (66) 9.6% (18) 3.8% (13) 0.0% (1) 0.0% (1) 47.0% 62.6% MARBLE HILL Overall , 6 ^ 6 % average ] (| -10) 36.4% '"'(80)'-''" 55 (•"•. 30 0% (121) . (66) 8 2% (18) 5 9% (13) 0 5% - (1) 0 .5% ; • (i) 34.5% 38.2% Note: The top number represents the percentage of all storms belonging to a specific class; the bottom number in parenthesis is the total number of storms belonging to a specific class. On average, 68.6 % of all storms occurred in the wet period whereas 31.4 % occurred during the dry season at the Promontory site. The majority (84.9%) of storms were minor events, while major events were infrequent (1.0 % of all storms). Over the three year record period, only two major events were captured. However, the Chilliwack climate station recorded 5 days with a total rainfall accumulation over 30 mm during the period of missing data at Promontory (Table B.9 in Appendix B), suggesting that a number of major rainfall events may not have been captured by the tipping bucket gauges. The storm class distribution within each season varied slightly between the two sites. However it is suspected that this variability is, at least in part if not primarily, due to the unequal distribution in data gaps between the two sites. The percent of record missing for both sites is shown in the last two columns of Table 6.2. Table 6.3 gives the median and range of storm characteristics for each class, based on three years of data (2001 to 2003) from the Promontory tipping bucket rain gauge. A more detailed breakdown and additional statistics are given in Appendix B (Table B.10). 72 Table 6.3 Summary of "Average" Storm Characteristics for a Three Year Period (2001 to 2003). Based on data from the Promontory tipping bucket. Total Peak Intensity (mm/hr) Duration Ant. Dry Period (hrs) N Precip. (mm) 5 min 15 min 60 min (hrs) M I N O R Median 3.3 3.60 2.00 2.00 4.38 31.88 187 Range 0.4-40.6 1.2-40.8' 1.0-14.0 1.0-4.0 0.3-34.5 6.3-653.0 I N T E R M E D I A T E Median 17.3 8.40 7.00 4.50 15.00 47.75 22 Range 10.8-80.2 6.0-79.2 5.0-27.0 3.0-8.0 2.8-46.5 8.5-215.3 M A J O R Median 49.2 28.80 20.00 14.00 22.50 173.00 2 Range 44.6-53.7 20.4-37.2 16.0-24.0 9.0-19.0 6.3-38.8 117.0-229.0 6.3.2 Storm Response Characteristics Storms with a total rainfall accumulation greater than 10 mm were investigated in further detail. For each event, a number of storm response variables were calculated. The magnitude of the event is represented by peak discharge. Response time is represented by the lag to peak value, which was computed as the time difference between the peak discharge and peak rainfall. The storm response variables for individual storm events are listed in Appendix B, and summarized for the eight hydrometric stations in Table 6.4 below. Table 6.4 Summary of Storm Response Variables for the Various Hydrometric Stations in the Chilliwack Creek Watershed Station Peak Discharge (m3/s) Peak Runoff Rate (mm/hr) Lag Time (hrs) Parsons Median 0.110 0.194 14.00 (F) Range 0.029-0.241 0.052-0.423 2.28-31.75 Elkview Median 0.134 0.223 14.00 (F) Range 0.041-0.732 0.067-1.217 0.25-31.25 Lefferson Median 0.024 0.152 1.25 (U) Range 0.009-0.288 0.059-1.803 0.00-12.25 Teskey Median 0.421 0.92 0.38 (U) Range 0.076-1.724 0.165-3.731 0.00-7.00 Luckakuck Median 0.220 0.267 2.75 (GW) Range 0.129-0.681 0.156-0.827 0.25-13.00 Bailey Median 0.385 0.317 6.25 (M) Range 0.031-1.058 0.025-0.872 1.00-38.25 Semiault Median 3.052 0.710 8.88 (A) Range 0.808-5.162 0.187-1.196 5.25-47.25 Chilliwack Median 3.479 0.806 9.00 (M) Range 0.905-4.816 0.180-1.617 4.00-17.75 73 6.3.2.1 Between Catchment Comparisons The storm response variables were compared across six of the sub-catchments using a Wilcoxon sign test. Chilliwack and Semiault were omitted from this analysis. The results of the significance tests are provided in Appendix B (Table B.19). The Teskey catchment showed significantly shorter lag times than all the other catchments, while both the forested catchments (Parsons and Elkview) had significantly longer lag times than other catchments. It is interesting to note that the Luckakuck catchment showed significantly shorter lag times than all stations except Teskey and Lefferson. Trends for lag times at the different catchments can be seen in Figure 6.11. 50-40 H Forest Hillslope Urban 12 30-•c. CD E ro ™ 20-10 O H - T T T X X —I 1 1 1 1 1 1 1 Parsons I Teskey I Luckakuck | Semiault I Elkview Lefferson Bailey Chilliwack Figure 6.11 Lag Time: Boxplots for the Different Sub-Catchments within the Chilliwack Creek Watershed Peak runoff rates were highest at the Teskey catchment. Lefferson had significantly lower peak runoff rates than all catchments with the exception of Elkview and Parsons. A l l other stations had statistically similar peak runoff rates, with one exception: Bailey had significantly lower peak runoff rates than Elkview. Boxplots comparing peak runoff at the various catchments are shown in Figure 6.12. 74 4H £ 3-TO or o c or co CD D. O H Parsons I Teskey I Luckakuck | Semiault I Elkview Lefferson Bailey Chilliwack Figure 6.12 Peak Runoff Rate: Boxplots for the Different Sub-Catchments within the Chilliwack Creek Watershed In summary, the two forested catchments (Parsons and Elkview) did not show any significant differences from each other; while the two urban stations (Lefferson and Teskey) were significantly different for all the variables. Teskey (the most urbanized catchment) had a shorter lag time, and higher peak discharge and peak runoff rate than Parsons (the least developed catchment). Significant differences (p<0.05) were also found between Elkview (forested) and Teskey (U) for all variables. 6.3.3 Hydrograph Comparison of Individual Storm Events A subset of six storm events is used to further investigate the response of the various sub-catchments. The storms were selected to represent a range of rainfall conditions (Table 6.5). In total, two minor, three intermediate, and one major storm were selected. The following section will outline the characteristics of these storm events and the response of each catchment. A complete table of response variables for each storm event is given in Appendix B. The hydrographs not presented in this section can also be found in Appendix B. 75 Table 6.5 Rainfall Summaries for the Six Selected Storm Events Based on Data from the Promontory Tipping Bucket Rain Gauge . . „ . . . Fraction of Storm Total ~ . , Ant. „ . Peak Intensities „ . , „ .. . , „ . _ ^ Total _ Peak . ,. „ Rainfall exceeding rainfall rates _ . Duration _ . Dry . (mm/hr) _ , If .1 Storm E v e n t (hrs) , R a l " Period , R a l " of (mm/hr)': Class (mm) (mm) 5 15 60 %> %> %> %> min min min 2.5 5 10 25 -8Mar-01 19.00 18.1 60.25 1.3 7.2 5.2 3.6 7.9 1.3 0 0 Intermediate 26-Oct-01 25.50 32.5 25.5 0.9 4.8 4.0 3.3 8.8 0 0 0 Minor 3-Nov-Ol 30.50 40.6 14.75 1.1 6.0 4.4 3.4 11.5 0 0 0 Minor 8-Dec-01 14.75 22.1 30.50 0.9 18.0 6.8 3.2 18.6 1.7 0 0 Intermediate 7-Jun-02 6.25 53.7 229.0 5.8 37.2 24.0 18.9 72.0 56.0 40.0 Major 10-Mar-03 17.00 18.9 10.25 0.9 7.2 5.6 4.0 14.7 0 0 0 Intermediate ' Fractions based on 15 minute rainfall intensity Minor Storm Events: October 26, 2001 and November 13, 2001 Of the 6 storm events, the two minor storms had the longest rainfall durations and 15-minute rainfall intensities below 4.5 mm/hr. Total rainfall accumulation for both storms is on the upper range for minor storm events. The 26-October-2001 storm produced rainfall accumulation of 32.5 mm during a 25.5 hour period, while the 13-November-2001 storm event produced a slightly higher (40.6 mm) rainfall accumulation over a 30.5 hour period. A summary of values for a few key response variables is shown in Table 6.6, and the hydrograph for each storm event is shown in Figure 6.13. Table 6.6 Lag Time and Peak Runoff Rate for Each Catchment for Two Minor Storm Events Hillslope Stations Downstream Stations Parsons Elkview Teskey Lefferson Luckakuck Semiault Bailey Chilliwack Land Use: Forest Forest Urban Urban Spring-Fed Agriculture Mixed Mixed 26-Oct-01 Storm Event Peak Runoff Rate ^ (mm/hr) 0.58 3.73 0.22 0.25 n/a 0.36 n/a Lag Time (hrs) n/a 19.25 6.75 6.75 13.00 n/a 38.25 n/a 13-Nov-Ol Storm Event Peak Runoff Rate ^ (mm/hr) 1.22 1.88 0.31 0.39 n/a 0.50 n/a Lag Time (hrs) n/a 10.75 0.00 10.00 0.25 n/a 10.75 n/a Intermediate Storm Events: March 18, 2001; December 8, 2001 and January 25, 2003 The intermediate storms had the lowest total rainfall accumulations of the six storm events at 18.1 mm, 22.1 mm and 18.95 mm. Antecedent dry conditions ranged from short (10.25 hrs) for the 18-March-2003 event to relatively long (60.25 hrs) for the 18-Mar-2001 event. However, it should be noted that the antecedent condition prior to the small event (<10 mm) which preceded the March event was 189 hrs. A summary of the response variables is presented in Table 6.7. The hydrograph for the 18-Mar-01 storm event is shown in Figure 6.13. 76 Table 6.7 Lag Time and Peak Runoff Rate for Each Catchment for Three Intermediate Storm Events Hillslope Stations Downstream Stations Parsons Elkview Teskey Lefferson Luckakuck Semiault Bailey Chilliwack Land Use: Forest Forest Urban Urban Spring-Fed Agriculture Mixed Mixed 18-Mar-01 Storm Event Peak Runoff Rate . (mm/hr) 0 1 8 0.20 n/a 0.34 n/a n/a 0.38 n/a Lag Time (hrs) 14.00 16.00 n/a n/a n/a 3.75 n/a 8-Dec-01 Storm Event Peak Runoff Rate ^ (mm/hr) 0.74 1.44 0.41 0.51 n/a 0.59 n/a Lag Time (hrs) n/a 3.00 1.50 0.00 0.25 n/a 6.25 n/a 10-Mar-03 Storm Event Peak Runoff Rate ^ (mm/hr) 0.19 0.75 n/a 0.39 0.75 n/a 1.62 Lag Time (hrs) n/a 2.75 1.25 n/a 3.25 17.25 n/a 15.25 Major Storm Event: June 17, 2002 The minor and intermediate storms described above were the product of modest rainfall accumulation over medium to long duration during the wet season. In contrast, the 17-June-2002 storm is a short duration, high intensity rainfall event during the dry season. This shortest storm event produced the highest rainfall accumulation at 53.7 mm delivered over a period of 6.25 hours. Peak rainfall rates exceeded 37 mm/hr at the 5 minute interval, and forty percent of the total rainfall fell at rates exceeding 25 mm/hr at the 15-minute interval. This summer storm also had the longest antecedent dry period at 229 hours (almost four times longer than any of the other storms described). The 17-June-2002 event was the largest storm in the 2001-2003 Promontory record; however the missing portion of the record includes two storms with daily precipitation totals greater than 60 mm (110.3 mm on 16-October-2003, and 73.8 mm on 20-October-2003) as measured at the Chilliwack climate station. The storm response for the various catchments are summarized in Table 6.8, and shown in Figure 6.14. It is interesting to note that for this event Lefferson has a much higher peak runoff and total runoff than Teskey, which is opposite of the usual trends. Table 6.8 Lag Time and Peak Runoff Rate for Each Catchment for a Major Storm Event Hillslope Stations Downstream Stations Parsons Elkview Teskey Lefferson Luckakuck Semiault Bailey Chilliwack Land Use: Forest Forest Urban Urban Spring-Fed Agriculture Mixed Mixed 26-Oct-OI Storm Event Peak Runoff Rate „ _ . (mm/hr) 0 2 4 0.21 0.38 1.80 0.71 n/a 0.71 n/a Lag Time (hrs) 12.75 2.00 0.50 1.00 2.75 n/a 9.75 n/a 77 Figure 6.13 Hydrograph for a Minor (12-Nov-01) and Intermediate Storm Event (18-Mar-01) 78 Figure 6.14 Hydrograph for a Major Storm Event (17-Jul-02) 6.3.4 Comparison between Forest and Urban Hillslope Catchments Land use on the hillslope has been changing over the last 10 years. Because the hydrologic data record is only a few years long it was not possible to detect the effects of changing land use within the same catchment. Instead a comparison of the urbanized section of the hillslope and the forested area was used to examine the potential effects of urbanization. The Parsons and Teskey catchments were chosen because they were determined to be the best examples of a forested and developed hillside catchment, respectively. The differences between Parsons and Teskey catchments for lag time and peak runoff rate are shown in Table 6.9 for each storm class separately and for all storms combined. 79 Table 6.9 Differences in Lag Time, Peak Runoff and Total Runoff between the Teskey (Urban) and Parsons (Forested) Catchments, for Three Storm Classes Minor Intermediate Major A l l Storms N = 5 6 1 13 Lag Time (hrs)1 Max diff. 31.75 19.5 31.75 Min diff. 13.5 8.5 12.25 8.5 Mean diff. 19.65 13.29 15.85 Peak Max diff. 1214% 1519% 1519% Runoff Min diff. 106% 116% 159% 106% Rate2 Mean diff. 411% 524% 447% 1 Values represent how much longer the response time is at Parsons than at Teskey (in hours) 2 Percentage represents the value at Teskey expressed as a percent of the value at Parsons. Note that a value greater than 100% indicates that the value at Teskey is greater than Parsons; and a percentage less than 100% indicates that the value at Parsons is greater. These results indicate that: a) Lag time was much longer at Parsons than at Teskey (between 8.5 and 31.75 hours longer over all storm events); and b) Peak runoff rates at the Teskey station were higher than at the Parsons station - up to 15 times greater during an intermediate storm event. Not all events were significantly greater; peak runoff rates showed less than a 16% increase for some minor and intermediate events. The results for one major event were included in Table 6.9; however, it should be noted that there is some uncertainty as to the validity of these results. Unexpectedly, peak runoff rate showed relatively little difference between Teskey and Parsons for this storm event. Lefferson (the other urban station, which usually has the lowest peak runoff rate) had a much higher peak runoff rate than both Teskey and Parsons. Peak runoff values at Lefferson were 4.7 and 5.7 times greater than at Teskey (respectively), and 7.5 times and 5.0 times greater than at Parsons according to the available data. Hydrograph Comparison: Figure 6.15 below shows graphically the difference between the storm response at Teskey and the storm response at Parsons for two different storm events: an intermediate storm event on 18-Mar-2001 and a major storm event on 17-Jun-2002. The top graph highlights the response typical of an urbanized catchment, and the bottom graph highlights the response typical of a forested catchment. 80 Catchment Area (ha) CA TIA (%) 100m buffer Parsons Forested 204.77 26.3 29.0 Teskey Urban 64.52 4.3 4.2 Annotation: The top graph highlights the 'typical' characteristics of an urbanized catchment. The bottom graph highlights the 'typical' characteristics of a forested catchment. Figure 6.15 Comparison of Teskey (urban) and Parsons (forested) Hydrographs for Two Storm Events 81 6.4 Data/Analysis Considerations In using this approach to identify the hydrological response of the different catchments (and ultimately the hydrologic consequences of urbanization of the hillslope) there are a number of concerns in the data and analysis that must be mentioned: 1) the length of the hydrological records is relatively short; 2) there is an unequal distribution of gaps between the different stations, and therefore it is difficult to compare estimates and results may be unreliable; 3) the natural variability of hydrological systems is generally high; 4) land use impacts are compounded by the complexity of hillslope processes, and potentially, by climate variability (as a result, differences in streamflow cannot be attributed solely to land use, but may also reflect differences in geology, topography, storm patterns etc); and 5) rainfall may not be uniformly distributed over the catchment. While this is a somewhat crude method, the analysis does provide a first approximation of the magnitude of storm events for the different catchments in the watershed. 82 7 WATER AND SEDIMENT QUALITY For this study, measurements of ambient water quality, both in the water column and in the sediments, were used to define the status of water quality in the watershed, to identify and quantify trends in water quality and to evaluate overall impacts of various sources of pollution. Specific parameters and streamwater constituents were used as indicators' to provide a representative picture of water quality. The indicators themselves may not negatively affect the aquatic environment; they may, however, suggest the presence of harmful constituents or potential for future degradation (Hayman, 2000). The major polluting sources expected from the different land use activities within the watershed determined the choice of parameters. Table C. 1 lists the chosen parameters and the land uses that may impact the parameters (Appendix C). It should be noted that this study focuses on chemical indicators, which by themselves cannot fully answer questions about the ecological response to a pollutant. Often by the time chemical concentrations reach a detectable level at the basin scale substantial insult to the ecosystem may have already occurred (Cairns, 1993). Also, knowledge of chemical concentration is not always representative of biological availability. Biological indicators, on the other hand, are continually exposed to the effects of various stressors and are therefore able to integrate the indirect and interactive effects of many different stressors over time (Chapman and Kimstach, 1996). As a result, they can identify impairments of aquatic life from unknown or unregulated chemicals and non-chemical impacts that chemical monitoring is unlikely to reveal. The following section briefly discusses each variable measured with respect to its origins, possible sources and its behavior and transformation in the aquatic system. Next, this chapter examines the current water quality conditions of the watercourses in the Chilliwack Creek watershed by looking at the spatial and temporal trends in the collected water quality and sediment data, and by comparing the collected data to water quality and sediment guidelines. The data from the sampling stations were grouped into three categories according to the dominant land use in the area immediately upstream of the sampling sites (agriculture, forest and urban as outlined in Table 4.1); statistical methods were then used to compare these groups. The focus of this chapter is to determine if surface water quality is different in the agricultural, urban and forested areas of the watershed. However, the specific relationship between land use and water quality is not emphasized here since it is explored in depth in a subsequent chapter. 1 Indicators are characteristics of the environment that, when measured, provide information about the current status of specific pollutants in the ecosystem, suggest potential for future degradation and identify responses to both anthropogenic and other stresses (Cairns et al., 1993). 83 7.1 Sediment and Water Quality Indicators 7.1.1 Nutrients When present in excess, nitrogen and phosphorus are the nutrients of most concern for water quality. Nutrient enrichment, particularly phosphorus since it is generally the limiting nutrient for plant productivity, contributes to excessive growth of aquatic vegetation and algae - a process known as eutrophication (Correll, 1998). This excess growth can ultimately lead to depressed oxygen levels when this plant material decomposes. In addition, certain blue-green algae associated with eutrophic waters form toxins, which can pose a health risk to both livestock and humans (Sharpley et al., 1994). Accelerated eutrophication causes a general deterioration of water quality and often limits the use of surface water for drinking, industry, recreation and fisheries purposes (Hayman, 2000). This section focuses on the sources, fate and chemistry of these nutrients in a watercourse. 7.1.1.1 Nitrogen (Nitrate, Nitrite and Ammonia) Nitrogen is an important nutrient for many aquatic organisms, particularly aquatic plants, as it is a vital component of amino acids and proteins (Hatch et al., 2002). It occurs in many forms in the environment and takes part in many biochemical reactions. However, when present in excess of local requirements, nitrogen can lead to pollution of watercourses. In the context of water pollution, nitrogen occurs in three forms that are of known concern: ammonia (NH 3 , which dissolves to form NHV"), nitrite (NCV) and nitrate (N03") (Waite, 1984). Most nitrogen enters the aquatic system naturally from the atmosphere and the soil through a process called nitrogen fixation. Nitrogen fixation is primarily performed by algae and various nitrogen fixing bacteria. Nitrogen can also enter the aquatic system due to ammonification, a process where bacteria break down organic matter to create ammonia. Consequently, increased concentrations of ammonia could be a useful indicator of organic pollution from sewage sludge or manure runoff. Ammonia can be found in two forms in water: as un-ionized ammonia NH 3 , or as the ammonium ion NH4 1". At a neutral pH typical of most natural waters the latter predominates. However, high temperatures and higher pH (pH>8) can cause a shift from the ammonium ion to the more toxic un-ionized ammonia form (Burt et al.,1993; Sharpley et a l , 1994). Ammonia concentrations greater than 2.5 mg/L have been shown to be toxic to many aquatic organisms (Eghball and Billey, 1999). Ammonia (NH3) is also the predominant form of nitrogen under anoxic conditions, such as those found within sediments. When sufficient oxygen is available, ammonia that enters the watercourse is rapidly oxidized to nitrate by the activities of microorganisms. This process, known as nitrification, occurs in two steps: ammonia is 84 first converted to the intermediate form nitrite, which is then oxidized to nitrate. While nitrite is quite toxic to aquatic life, the usually rapid oxidization of nitrite to nitrate prevents high concentrations from accumulating in aquatic systems. However, under conditions of high temperature and poor aeration ammonia oxidation exceeds nitrite oxidation, and nitrite can accumulate (Hatch et al., 2002). Under anaerobic conditions nitrate can be reduced to nitrite, and in most cases further reduced to nitrogen gas (N2)- This process is called denitrification, and it is how nitrogen is lost from the system (Burt et al., 1993). Nitrification of ammonia to nitrate is a key process which mobilizes nitrogen and promotes its loss from agricultural fields to watercourses (Hatch et al., 2002). Nitrate is relatively stable, very soluble and because of its negative charge does not become fixed on clay or organic matter. It therefore remains highly mobile and is the major source of nitrogen pollution to watercourses. On the other hand, ammonia is positively charged and tends to be relatively immobile. However, in soils subject to erosion it can occasionally be removed and reach watercourses in overland flow. Although nitrogen occurs naturally in unpolluted waters, it is generally found in low concentrations (0.07 to 2.3 mg/L) (Stednick, 1991). Concentrations of nitrogen in streams are increased through human sources such as municipal sewage discharge, leaching or runoff of inorganic nitrate fertilizers through soil in suburban and rural areas, agricultural runoff of manure or inorganic fertilizers, and leaking of septic tanks. Nitrate pollution has been considered a hazard to human health for many years for two reasons. First, concentrations exceeding 10 mg/L were associated with methomeglobemia (blue-baby syndrome), and secondly, the ingestion of large amounts of nitrate was thought by some to cause stomach cancer (Addiscott et al., 1991). However, recent medical research has questioned some of the reasoning behind this established view (Addiscott et al., 1999; Hatch et al., 2002). 7.1.1.2 Phosphorus (Orthophosphate) Phosphorus (P) is an essential nutrient for algae and plants, yet it is also one of the scarcest elements available in both terrestrial and aquatic ecosystems in terms of its demand (Leinweber et al., 2002). While phosphorus is present in numerous different forms, only orthophosphate (P043") can be assimilated by bacteria, algae and plants (Correll, 1998). Dissolved orthophosphate occurs in one of three forms, depending on the pH; H 2 P0 4 " and HP0 4 2 " are the inorganic anions that predominate in the normal pH range of natural waters, while H 3 P0 4 is more abundant in acidic environments (Waite, 1984). 85 Phosphorus in soils and water originates naturally from the weathering of phosphate bearing rock and decomposition of organic matter (Chapman and Kimstach, 1996). Anthropogenic sources that contribute to elevated P levels include domestic wastewaters (particularly those containing detergents) and runoff from agricultural lands on which manure or inorganic fertilizers have been applied. A major cause for excessive phosphorus concentrations on farmland is excessive manure and fertilizer application. The nitrogen to phosphorus ratio of manure (2:1 to 6:1) is typically lower than what crops require (7:1 to 11:1) (Gburek et al., 2000); however, most farmers apply manure to their fields based on nitrogen demands. Phosphorus is delivered to aquatic systems from the land surface in both dissolved and particulate forms. Since P is strongly adsorbed into the soil profile, runoff and erosion are the main mechanisms by which this P is transported (Parry, 1998; Sharpley et al., 1994; Correll, 1998). In cultivated soils losses of particulate P can be extremely high and constitute most of the P transferred in runoff (60-90%), whereas runoff from grass or forest lands carries little sediment and is, therefore, dominated by the dissolved form (Daniel et al., 1998; Sharpley et al., 1992). Once in the watercourse, particulate phosphorus input into watercourses may be deposited in the bottom sediments or it may release P and organic P into solution, and eventually hydrolyzed to orthophosphate (Waite, 1984; Daniel et al., 1998). Orthophosphate can follow one of two pathways: formation of insoluble metal-complexes and biological uptake. The phosphate can complex with aqueous cations, particularly iron, aluminum and calcium, to form insoluble molecules which precipitate out into the sediments (Waite, 1984). These phosphates-metal complexes are no longer available for plant uptake. Consequently, the concentration of available phosphorous in the aquatic system is in part a function of the factors (primarily pH and dissolved oxygen concentration) affecting the solubility of these metal complexes. For instance, under anoxic conditions Fe 3 + is reduced to Fe 2 +and phosphate is released back into the water column. Orthophosphate can also be taken up by plants where it is converted into organic phosphorus (as polyphosphates). Eventually, when the plant dies this immobilized P in the plant tissues is release back into solution when bacteria re-hydrolyze the polyphosphate back to orthophosphate (Waite, 1984) In most natural waters orthophosphate levels range from 0.005 to 0.02 mg/L (Chapman and Kimstach, 1996). Phosphorus itself is not toxic at concentrations found in natural waters (Campbell and Edwards, 2001; Doljilo and Best, 1993) and consequently, surface impacts of phosphorus loads do not pose a risk to human health but do create environmental and aesthetic concerns. Hence, water quality guidelines are set in to prevent eutrophication rather than direct phosphorus toxicity. For British Columbia, the drinking water quality guideline has been set at 0.01 mg/L total P. No guideline has been set for the protection of aquatic life in streams. 86 7.1.2 General Water Chemistry 7.1.2.1 pH pH is a "determining factor in almost every natural (chemical or biological) process", and as a result it is an important variable to measure in water quality assessments (Stednick, 1991). The pH range of most natural waters is between 6.5 and 8.5, and is primarily dependent on the concentration of carbonates and carbon dioxide in the water body. As a result, pH levels may be higher during the daytime, particularly in eutrophic waters when aquatic plants are actively removing carbon dioxide from the water through photosynthetic activity (Doljilo and Best, 1993). The geology of the catchment and the soil types in the drainage area also influence pH. For example, drainage waters from forests are usually more acidic because of the presence of humic and fulvic acids (Doljilo and Best, 1993). pH also affects the equilibrium between soluble and solid species of trace elements. In general, trace elements tend to become more soluble at lower pHs - and consequently, lowering the pH levels can allow the release of toxic metals that may otherwise be attached to the sediment and unavailable to water system (Chapman and Kimstach, 1996). Once these metals are mobilized they are potentially more available for uptake by aquatic life. 7.1.2.2 Temperature Temperature influences almost every biological, chemical and physical process in a waterbody, and consequently, the concentration of many water quality constituents (Chapman and Kimstach, 1996). As temperature increases, the solubility of gases (N 2, 0 2 , C0 2 ) in water decreases. Also, higher temperatures generally increase the rate of biochemical processes involved in metabolism, growth and reproduction. (Doljilo and Best, 1993). In warmer waters, increased respiration can lead to increases in oxygen consumption and increased decomposition of organic matter. This increased metabolic oxygen demand in conjunction with the reduced solubility of oxygen in water at higher temperature can cause an oxygen deficit which can be harmful to aquatic life. Furthermore, temperature affects the solubility and toxicity of many chemical compounds (such as pesticides and trace metals) and therefore can influence the effects of pollutants on aquatic life. Streamwater temperature may also be an indication of groundwater influence, particularly in the dry season when groundwater makes a greater contribution to streamflow. 7.1.2.3 Specific Conductivity Specific conductivity is a measure of the water body's ability to conduct an electric current. The current is conducted in solution by the movement of ions. In natural waters, the ions in solution (essentially, Na+, Mg 2 + , Ca 2 + , K+, CI", S0 4 2", H C 0 3 ' and C0 3 2") are formed predominantly by the dissociation of inorganic compounds, mostly mineral salts. Organic compounds dissolve very little and consequently contribute 87 little to conductivity. Specific conductivity has no significant health implications. It is, however, sensitive to the amount of salts dissolved in water and can be used as a convenient, rapid method of estimating total dissolved solids and salinity in a watercourse. These dissolved solids and salts can have implications for domestic and agricultural water use (Stedick, 1991; Doljilo and Best, 1993). Conductivity in streams is affected by the geology of the area through which the water flows. For example, streams that run through granite bedrock will have a much lower conductivity than those that flow through limestone, shale or through clay soils. Higher conductivity readings can also result from pollution sources such as urban or agricultural runoff. Other factors that can affect conductivity include temperature and discharge; warmer water and low flow conditions generally contribute to conductivity readings. 7.1.2.4 Dissolved Oxygen Dissolved oxygen (DO) content is another vital parameter of any water body because it is essential for most forms of life. Dissolved oxygen concentrations below 5 mg/L may adversely affect the function and survival of biological communities, and concentrations below 2 mg/L can lead to fish mortality. The solubility of heavy metals is also influenced by the dissolved oxygen content of the water body -solubility of most metals decreases under low oxygen conditions. A number of factors may influence the concentration of DO in water, including turbulence, biological activity and temperature. Water gains oxygen from the atmosphere. Therefore, physical movement, such as rapids, promotes dissolving of oxygen in water. Secondly, respiration by aquatic plants, and the decomposition of organic matter consume oxygen from the water. Consequently, water discharges high in organic matter and nutrients can lead to decreases in the dissolved oxygen content as a result of the increased microbial activity. Accordingly, DO is a good indicator of the degree of pollution by organic matter. The solubility of oxygen also decreases as temperature increases. Variation in DO can occur seasonally or daily in relation to temperature and biological respiration (related to photosynthesis and the decomposition of organic matter). Both an increase in temperature and increase in respiration will lead to decreases in DO concentrations. 7.1.3 Major Ions in Water While waters contain a vast array of chemical constituents, a relatively small number of elements make up the majority of the species found in most natural waters. The major cations are the alkaline and alkali earth metals, which exist largely as free ions (Ca 2 +, Mg 2 + , Na + , and K+). The major anions (CI", HC0 3 " and S04 2") were not measured in this study. Concentrations of major ions are naturally very variable in surface waters, and mainly depend upon the natural geology of the catchment and the factors influencing 88 the weathering process of rocks and soils (Doljido and Best, 1993; Chapman and Kimstach, 1996). Potassium (K) is generally found in low concentrations in natural waters since potassium bearing minerals (e.g. microcline K A L S i 3 0 8 , leucite K A L S i 2 0 6 , silvine KC1, kainite KC lMgS0 4 -3H 2 0, carnallite KClMgCl 2 -6H 2 0, glacerite K 3NA(S0 4 ) 2 ) are relatively resistant to weathering (Doljido and Best, 1996). However, because potassium is an essential element for plant growth potassium salts are widely used as fertilizer; and consequently can enter watercourses with agricultural runoff. Sodium (Na) minerals (e.g. halite NaCl, thenardite Na2S04, albite NaSi3Og) tend to be highly soluble, and consequently, the natural level of sodium in water is considerably higher than potassium. Increased concentrations may arise from the use of salts on roads and from sewage and industrial effluents. Because elevated sodium in soil can degrade soil structure and thereby restrict water movement, sodium is commonly measured where water is to be used for irrigation (Chapman and Kimstach, 1996). Magnesium (Mg) is present in ferromagnesium minerals and some carbonate rocks (e.g. dolomite CaMg(C0 3) 2 , and magnesite MgC0 3 ) , while calcium (Ca) is readily dissolved from rocks rich in Ca minerals - particularly carbonates (calcite CaC0 3 ] dolomite CaMg(C0 3) 2 , limestone) and sulphates (e.g. gypsum CaSCy2H 20). Calcium concentrations can fall when Ca carbonate precipitates due to increased water temperatures, photosynthetic activity or loss of C 0 2 in the system (Chapman and Kimstach, 1996). Together, calcium (Ca) and magnesium (Mg) salts are responsible for the hardness of water, and when heated they form insoluble scales in water heaters and coders reducing their efficiency. The amount of hardness in the water can also modify the toxicity of some cations due to competition between the toxic cation and the Ca and Mg ions at the exchange sites in the aquatic organisms. 7.1.4 Metals in Water and Sediment 7.1.4.1 Sources of Metals in Urban and Agricultural Areas Trace metals are naturally present in freshwater systems from erosion and the weathering of rocks and soils. However, with the increasing urbanization and agricultural intensification over the last few decades metals introduced into the aquatic environment are increasingly coming from anthropogenic sources such as direct discharge of effluent from domestic or industrial activities, wet and dry atmospheric deposition, and mining (Chapman and Kimstach, 1996). In many areas, recent metal accumulation has been occurring at a much faster than the historical rate. This section focuses on sources of metal pollution in agricultural runoff and urban stormwater. In agricultural watersheds the primary sources of metals are fertilizers and manure. Some trace metals are essential micronutrients for animals and are frequently added to livestock and poultry feed as growth promoters and antibiotics (Smith, 2004; McBride and Spiers, 2001; Nicholson et al., 1999). Zinc (as zinc oxide or zinc sulphate) and copper (as copper sulphate) are the most common feed supplements. As a 89 result, manure can contain high levels of copper and zinc, as well as other metals such as iron, manganese and lead. Zinc sulphate is also used as a plant fertilizer, zinc chloride is used as a pesticide and organic zinc compounds are used as fungicides (Ohnesorge and Wilhelm, 1991; Smith, 2004). In general, livestock and poultry manure contribute most to the input of Cu and Zn, whereas fertilizers are the dominant source of Cd in agricultural (both arable and livestock) systems (de Vries, 2002). Atmospheric deposition is generally a larger contributor of Pb than agricultural sources, although it is used in pesticides and can be found in livestock manure (deVries, 2002). Stormwater runoff mobilizes large quantities of contaminants from the urban environment (Characklis and Wiesner, 1997; Davis et a l , 2001), including trace metals (Choe et al., 2002). A number of studies have found various levels of metals in runoff from urban areas, particularly highway runoff (Wu et al., 1998, Sansalone and Buchberger, 1997, Marsalek et al., 1999). Common metals associated with urban runoff include: copper, zinc, iron, lead manganese, nickel, arsenic and chromium. Generally the levels follow the order: Zn (20-5000 ug/L) > Cu ~ Pb (5-200 ug/L) > Cd (<12 ug/L) (Davis et al., 2001). In urban areas, trace metals are introduced into the environment through construction materials and chemicals; however motor vehicles are recognized to be the primary source of most metals (Gibb et al., 1991; Davis et al., 2001; Sansalone and Buchberger, 1997). Transport related sources of metals include gasoline (Pb, Mn), diesel fuel (Cd), exhaust emissions (Pb, Ni), lubricating oils (Pb, N i , Zn), grease (Zn, Pb), tire wear (Cd, Zn), asphalt paving wear (Ni), break lining wear (Cu) (Legret and Pagotto, 1999; McCallum, 1995). Due to the significant reduction in the use of tetraethyl lead (TEL) as a gasoline additive (15 mg//L since 1989 instead of 40 mg/L) there has been a decrease in the lead content in surface waters (Legret and Pagotto, 1999). The natural levels in gasoline (-10 mg/L) may still contribute lead (Lee and Jones-Lee, 1993). This reduction of Pb based gasoline additives was associated with an increase in the use of methylcyclopentadienyl manganese tricarbonyl (MMT) as an alternative octane enhancer (Egyed and Wook, 1996). Consequently, M M T fuel additive is suggested to be a potential source of Mn accumulation in urban streams (Mielke et al., 2002). Other sources of metals include wires and pipes (Cu, Zn), the corrosion of iron and steels products (Fe), and lead based paints (Pb). Galvanized roofs have been shown to be a significant source of Zn. 7.1.4.2 Chemistry and Fate of Trace Metals in the Aquatic Environment Elements in aquatic systems may be present in many forms such as dissolved ions, dissolved organic or inorganic complexes, precipitated as metal oxides, hydroxides, carbonates and sulphides, or adsorbed onto clays and humic materials. In general, only a small proportion of the total metal load in aquatic systems is actually found in the dissolved fraction, while bottom sediments often become contaminated with 90 metals. This is because sediments have a remarkable ability to remove metal ions from the water column through various processes (e.g. ions exchange, precipitation, chelation), collectively known as sorption. Clay minerals, hydroxides and oxides (particularly those of Fe, A l and Mn), as well as particulate organic matter (POM) are the most important constituents in these sorption processes. Humic substances contain a large amount of hydroxyl and carboxylic functional groups, which can act as cation binding sites when they dissociate in water to release protons. Similarly, the hydroxyl groups on the edges of phyllosilicate clays can donate protons to aqueous solution in return for metal uptake. Metal adsorption onto these sites is pH dependent (Evans, 1999). At the broken edges of clay particles, negatively charged oxygen atoms provide additional binding sites. Clay minerals may also have a permanent structural charge (as a result of structural imperfections caused by isomorphic substitution or non-ideal occupancy in the octahedral sheets) which can act as additional sites for ion-exchange (Evans, 1999). Under oxic conditions, the surfaces of oxide and hydroxide minerals can donate protons to aqueous solution in return for metal uptake. Under reducing conditions, the adsorption mechanism of oxides reverses, resulting in re-mobilization of sorbed metals (Stumm and Morgan, 1970). Consequently, increasing the organic matter and clay content in sediments provides more sites for metal adsorption thus reducing the amount of metal that is found in the dissolved fraction. Texture has been cited by many investigators as the most important factor affecting trace metal concentrations. The fine-textured sediment, which contains an appreciable amount of organic matter, Fe and Mn oxides, and clay minerals, tends to accumulate higher concentrations of metals due to the higher adsorptive capacity of these constituents (Fergusson, 1990; Salomons and Forstner, 1984). The speciation of trace metals in natural waters has an important influence on their mobility/transport, chemical reactivity (behavior), and biological availability and toxicity. Research has shown that the biological availability and behavior of trace metals in aquatic systems is closely related to chemical forms rather than the total metal concentration. Small dissolved metal species and free metal ions, in particular, are believed to be most easily absorbed in biota from the water column (Pagenkopf, 1983; Gundersen and Steinnes, 2001), and therefore most likely to result in adverse effects. In most unpolluted waters, concentrations of free metals are low due to the abundance of natural ligands, adsoption sites and near neutral condition; however, changes in water chemistry and metal loadings can significantly increase the free metal ion concentration. Factors that can influence the proportion of metals in each fraction include the physical properties of sediment (texture, organic matter content, type of clay present), chemical conditions (pH, water hardness, and redox conditions), metal concentration and the presence of other ligands (Evans, 1999; Salomons and Forstner, 1991; Gundersen and Steinnes, 2001). 91 In general, pH seems to have the greatest effect on the solubility of metals, with a greater metal retention in sediment and lower solubility of metal cations as pH increases. This occurs because as pH decreases, cations compete with extra hydrogen (H+) and aluminum (Al 3 +) for positions on exchange sites (Forstner and Salmons, 1991). The extent of adsorption of metals typically goes from near zero to near 100% over a relatively small pH range (this 'adsorption edge' is dependent on the type of metal). As a result, a small shift in the pH of surface water can cause a sharp increase or decrease in the dissolved metal levels (Salomons and Forstner, 1984). Changes in redox (oxidation/reduction) conditions in aquatic systems can influence the availability of metals by changing the oxidation state of the metal ion and its solubility (Evans, 1999). Iron and manganese oxides are particularly susceptible to oxidation and reduction reactions; with the solubility of Fe and Mn increasing under reducing conditions. The amount of metal in sediment may decrease if the concentration of anions in solution (e.g. chloride, biocarbonate, nitrate) is increased, through the addition of salts for example, as the binding sites in the sediment constituents will have to compete with the anions in solution for the metallic cation. (Salomons and Forstner, 1984). Sediment dynamics will have an influence on the degree of variability within river systems. 'Hot spots' are created where a combination of pollutant loading and sediment dynamics produce sites with high impact potential (Rhoads and Cahill, 1999). In general, concentrations in urban streams generally decline as distance from a point source (e.g. stormwater outfall) increases. The highest concentrations in the downstream reaches generally occur in areas that promote the accumulation of fine sediment (e.g. regions of low velocity such as vegetated area (Rhoads and Cahill, 1999). 92 7.2 Water Quality Results The analysis of the water quality results is presented in this section. As previously mentioned, surface water samples from twenty stations throughout the watershed were collected between May 2002 and July 2003 as well as during a storm event in October 2003 (see Figure 4.1 for the location of the sampling stations). Analytical results for streamwater (nitrate-N, orthophosphate-P, ammonia-N, major ion and trace element concentrations) as well as field measurements of streamwater specific conductivity, pH, dissolved oxygen, temperature are presented in Appendix C, and summarized in Tables 7.1 and 7.2 for the complete data set and for sites grouped by land use categories (agriculture, urban, forest). Natural background levels in streamwater, and results from other studies in the Lower Fraser Valley (LFV) are also included in these tables for comparison. 7.2.1 Spatial and Temporal Trends in Water Quality Parameters The spatial and seasonal variability in each of the water quality parameters is addressed using series of graphs. In these graphs, the wet and dry season averages are plotted for stations from the headwaters of the Interception Ditch mainstem to the mouth of Chilliwack Creek in the upstream to downstream direction (represented by the solid lines). The Chilliwack Creek mainstem (Ml9 to M20) is also shown (represented by the dashed lines). It should also be noted that in these graphs the first station along the 'Interception Ditch mainstem' is a forested tributary station (Parsons Brook, F13); the next three stations are along the ditch itself, and the final station is at the mouth of Chilliwack Creek. For the Chilliwack Creek mainstem, the upstream station is the most intensive urban sampling site in this study (Ml9). Values for selected forested, urban and agricultural tributaries are shown in a separate graph; arrows on the mainstem graph indicate where a tributary enters the stream. Figure 7.1 shows the location of the sampling stations on the two mainstems (Interception Ditch and Chilliwack Creek) and the stations along the tributaries. The locations of a few key points of confluence are also shown on the map: 1) the point where the streams draining the 'urban hillslope' (UH) enter Interception Ditch (between station A16 and A18); 2) the point of confluence for the forested hillslope tributary (FH) with Interception Ditch (between station A15 and A16); and 3) the point where Interception Ditch enters Chilliwack Creek. 93 Table 7.1 Overall Surface Water Chemistry of the Chilliwack Creek Watershed, and Comparison with Natural Background Levels and Other Studies in the Lower Fraser Valley (LFV) p H Dissolved Sp. Cond. Temperature Nitrate-N Ammonia-N Ortho-P Oxygen (mg/L) (uS/cm) CC) (mg/L) (mg/L) (mg/L) 2002-2003, Chilliwack Creek Streamwater Sampling Mean 7.4 9.2 245 11.9 0.75 0.27 0.09 ALL SITES Range 6.0-9.0 3.9-15.2 97-550 4.6-22.6 bd-5.68 0.05-5.23 bd-0 .16 _ Mean 7.27 9.2 275 12.6 0.75 0.43 0.11 AGRICULTURAL SITES — Range 7.0-9.0 6.0-12.6 130-550 6.1-22.6 bd-5.68 0.05-4.23 bd-0 .16 r, „ Mean 7.6 9.5 213 11.2 0.51 0.13 0.08 URBAN SITES — Range 7.0-8.0 3.9-12.8 97-373 4.6-16.6 0.06-1.35 bd-0.71 bd-0.13 „ Mean 7.6 9.7 217 11.9 0.64 0.16 0.06 FOREST SITES — Range 7.0-8.0 4.5-12.6 110-322 5.0-18.7 0.23-1.11 bd-0 .60 bd-0.08 Background Concentrations in Streamwater Natural levels in freshwaters 6.0-8.5b bd-18.4 a typically > 10 10-1000b 0.002-6.6a <0.001-0.490a typically <0.1 0.005-0.020b Aggasiz (control site)e 5.8 (5.2-6.0) 10.7 (5.7-17.4) 25 (10-45) 8.2 (2.7-15.9) 1.500 (0.006-0.498) 0.063 (0.002-0.241) 0.009 (0.003-0.014) FORESTED CONTROL Vedder Mountain 2003-2004 7.7 10.5 111 7.7 0.013 Al l bd SITES IN L F V STUDIES (Sumas Watershed)0 7.3-7.9 4.4-14.0 70-148 2.6-14.8 bd-0.410 Upper Elk Creek (Chilliwack hillslope region 1998-2001)d 10.2 6.2-14.5 10.8 2.5-20.2 0.008 0.005-0.504 Streamwater Concentrations, mean and range (Impacted watersheds in the Lower Fraser Valley) AGRICULTURAL Sumas River (Abbotsford)0 7.4 6.6-8.7 7.9 2.7-12.3 277 186-358 10.3 3.9-23.9 2.26 0.46-4.92 0.230 bd-1.260 0.0350 bd-0.120 WATERSHEDS - •• Lower Elk Creek 12.0 6.8 0.006 (Chilliwack 1998-2001)d 6.5-13.2 0.8-13.9 0.005-0.440 a C C R E M (1987) cited in Berka (1996) * bd = below detection, taken as detection limit for analysis b Chapman and Kimstach (1996) c Smith (2004) d Schreier et al. (2004) eAddah(2002) Table 7.2 Major Ions and Trace Metals in Water for the Chilliwack Creek Watershed, and Comparison with Natural Background Levels and Other Studies in the Lower Fraser Valley (LFV) (mg/L) Fe M n C a K M g Na 2002-2003, Chilliwack Creek Streamwater Sampling ALL SITES Mean 0.369 0.073 28.0 1.26 5.36 6.79 Range b d - 1.685 bd-0.380 13.7-53.1 bd-2.71 1.90-9.65 2.92-13.90 AGRICULTURAL Sn Mean 0.796 0.134 30.3 1.18 5.06 6.56 FES Range 0.210-1.685 bd - 0.252 16.2-53.1 bd-2.42 2.60-7.90 3.64-9.00 URBAN SITES Mean 0.095 0.017 24.2 1.30 6.70 7.16 Range bd-0.330 bd-0.058 15.6-29.5 0.78-2.23 5.34-9.65 22.93-13.90 FORESTSITES Mean 0.087 0.018 20.8 0.70 2.68 7.68 Range bd-0.185 bd-0.042 13.7-27.1 bd -0.87 1.90-3.22 4.20-9.98 Background Concentrations in Streamwater Natural ranges of (dissolved) concentrations in streamwater0 0.055e 0.006e 0.06-210 0.1-6.3 0.05-80 0.06-350 Global average (MCNC) d 8.0 1.0 2.4 3.7 PRISTINE STREAMS DRAINING COMMON ROCK TYPE F Granite 0.78 0.3 0.38 2.0 Sandstone 1.8 0.82 0.75 1.2 Shale 8.1 0.78 2.9 2.4 Carbonate 51 0.51 7.8 0.8 FORESTED CONTROL SITES IN L F V STUDIES Vedder Mtn. 2003-2004 (Sumas Watershed)8 All bd All bd 14.38 9.3-193-.4 All bd 3.1 2.2-4.1 Lower Elk Creek (Chilliwack hillslope region 2000, means)b 0.133 (spring) 0.053 (fall) 0.005(spring) 0.036(fall) Streamwater Concentrations, mean and range (Impacted watersheds in the Lower Fraser Valley) AGRICULTURAL WATERSHEDS Sumas River (Abbotsford)a 0.355 0.05-1.14 0.052 bd-0.161 17.1 7.7-25.7 2.39 0.07-5.80 19.66 12.6-28.1 Hope Slough (Chilliwack 2000, means) b 0.696 (spring) 0.717 (fall) 0.146(spring) 0.116(fall) *bd - below detection a Smith (2004) b Schreier et al. (2004) c Chapman and Kimstach (1996) after avergaes from a survey of 250 pristine streams in France (Meybeck, 1986) and from 75 sites world-wide (Meybeck 1987); d M C N C (most common natural concentrations) corresponding to median value obtained in 60 major rivers (Meybeck 1979); c Yeats and Bewers (1982) cited in Salomons and Forstner (1984) rMeybeck and Helmer (1989) cited in Chapman and Kimstach (1992); based on 75 unpolluted rivers in monolithological watersheds (rock type proportion close to the estimated global proportion of Meybeck (1987) Point of Confluence: Semiault Creek (agricultural tributary) enters Chilliwack Creek Point of Confluence: Urban Hillslope Tributaries Lh\i Point of Confluence: Interception Ditch enters Chilliwack Creek • Point of Confluence: Forested Hillslope Tributaries Sampling Stations • not graphed O Mainstem • Chilliwack Creek (upstream station) 9 Tributary Stations Sampled Streams f \ f Agriculture / % / Mixed / \ / Urban / \ / Forest Spring Fed Unsampled Streams and Ditches Roads 60 m contours Q3 Chilliwack Creek Watershed 1 0.5 0 1 Kilometers Figure 7.1 Location of Sampling Stations along the Interception Ditch and Chilliwack Creek Mainstems, and Selected Tributaries 7.2.1.1 Variations in Ammonia-N Throughout the 2002-2003 sampling season, ammonia (NH4+-N) concentrations ranged from below detection (0.10 mg/L) to 5.23 mg/L. The highest value was measured in Bailey Ditch (M10) on 12-Dec-02, and exceeded the B. C. Approved Water Quality Guidelines for the protection of aquatic life. On this date, ammonia concentrations at stations A16 (3.41 mg/L), A17 (4.23 mg/L) and A18 (1.77 mg/L) also exceeded the maximum permissible total concentration for the protection of aquatic life. Furthermore, in July 2003 the ammonia concentration at station A18 was very near the average 30-day concentration guideline for ammonia-N (<0.15 mg/L at pH = 8.8 and temperature = 22°C). Ammonia concentrations varied spatially along the Interception Ditch and Chilliwack Creek, as shown in Figure 7.2. Concentrations are low at the upstream station of Chilliwack Creek with a significant increase in ammonia-N concentration when agricultural tributaries (Semiault Creek and Interception Ditch) enter the mainstem. Interception Ditch shows a similar downstream trend with low concentrations in the (forested) headwaters and an increase in the downstream direction as agricultural influence intensifies. This increase was most pronounced between stations A16 and A18, where Teskey Way Ditch (a ditch with both urban and agricultural influence) drains into the mainstem. Concentrations at M10 and A17 (Teskey Way Ditch) are high suggesting that that this portion of the watershed may be contributing ammonia-N to Interception Ditch. These two stations are located downstream of the urban hillslope tributaries as well as some agricultural land. It should be noted that station M10 may also receive water from the Bailey landfill site, but from the sampling sites in this study there is no way to differentiate i f the high levels at M10 are due to upstream agricultural activities or leakage from the landfill site. An exception to the spatial pattern described above occurred on 12-Dec-02 when concentrations peaked at station A16 (3.41 mg/L). The high concentrations detected on 12-Dec-2002 at station A16, A17, A18 and M10 likely resulted from runoff caused by a moderate rainfall (>20 mm) immediately prior to and during the sampling. A l l other concentrations measured during the sampling period were below 0.900 mg/L. This result suggests that there is a high potential for rapid and higher losses of nitrogen to occur during rainfall events. Figure 7.2 shows the wet and dry season trends excluding the 12-Dec-02 sampling data. 97 1. E, c o TO k_ C V o c o o c o E E < 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 Forested (FH) Urban Hillslope (UH) Agriculture GW I • II F12-F14 U6 U8 U5 U3 U7 U4 A11 A2 G1 Tributaries F13 A15 A16 A18 M19 M20 1 km Interception Ditch Mainstem to Mouth of Chilliwack Creek (by distance downstream) Legend: A Grey triangles represent dry season means. — Solid line represents Interception Ditch Mainstem • Black squares represent wet season means (excl. 12-Dec-02) Dashed line represents Chilliwack Creek FH= forested hillslope tributaries (Elkview Creek); U H = urban hillslope tributaries Notes: Error bars represent the range of measured values. Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.2 Spatial and Seasonal Variations in Streamwater Ammonia-N in the Chilliwack Creek Watershed 98 Seasonally, ammonia-N concentration in the mid and lower sections of Chilliwack Creek and Interception Ditch were higher in the wet season than in the summer season. In agricultural areas, surface runoff and erosion are the major sources of ammonia to streams. In the wet season, surface runoff increases and biological uptake decreases because plants are not growing; this leads to increased inputs to streams. The opposite occurs in the dry season because there is less runoff and because plants and organisms are actively absorbing ammonia. In addition, the warmer water temperatures during the dry season would encourage nitrification (conversion of ammonia to nitrate). Tributaries that were less influenced by temperature and algae and vegetation growth, namely shaded hillslope tributaries, showed minimal seasonal variation. 7.2.1.2 Variations in Nitrate-N During the sampling period, the mean nitrate-N concentration was 0.75 mg/L. Values below detection limit were measured in the middle section of Interception Ditch (A 15 and A16) during the dry season (October and July, respectively). The highest values were consistently recorded in Semiault Creek (A2) during the wet season. In general nitrate concentrations in the watershed are low; and all values were below BC Water Quality Guidelines for drinking water (10 mg/L) and aquatic habitat (200 mg/L). However it should be noted that, during the wet season, Semiault Creek (intensive agriculture) and Chilliwack Creek had nitrate-N concentrations that were above the 3 mg/L level that is indicative that the stream is impacted by anthropogenic influence (Schreier, pers. comm., 2004). The spatial and seasonal variability in nitrate-N during low flow (dry) and high flow (wet) conditions for Interception Ditch, Chilliwack Creek and selected tributaries is illustrated in Figure 7.3. Nitrate-N concentrations along Interception Ditch varied only slightly between stations, remaining consistently below 1.43 mg/L along the length of the stream. The trends described were most pronounced on the December sampling date, which occurred just after a storm event when discharge was still increasing. Chilliwack Creek showed the opposite trend with nitrate-N concentrations decreasing downstream. It should be noted that concentrations at the mouth of Chilliwack Creek were, on average, double the concentrations measured in Interception Ditch. Semiault Creek, which flows through the main agricultural area in the watershed, had high nitrate concentrations, and empties into Chilliwack Creek prior to station M20. However, due to the consistently high nitrate-N concentrations recorded at the upstream station (Ml9), it is uncertain how much this agricultural area is influencing nitrate levels in Chilliwack Creek. 99 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Forested (FH) Urban Hillslope (UH) 5.68 Agriculture GW O) E c o c <u o c o o T F12-F14 U6 U8 U5 U3 U7- U4 A11 A2 G1 Tributaries E c o o c o o z d> F13 A15 A16 A18 M19 Interception Ditch Mainstem to Mouth of Chilliwack Creek (by distance downstream) 1 km Legend: A Grey triangles represent dry season means. • Black squares represent wet season means Solid line represents Interception Ditch Mainstem Dashed line represents Chilliwack Creek FH= forested hillslope tributaries (Elkview Creek); U H = urban hillslope tributaries Notes: Error bars represent the range of measured values. Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.3 Spatial and Seasonal Variations in Streamwater Nitrate-N in the Chilliwack Creek Watershed 100 Wet season values were higher and more variable than dry season values throughout the watershed, reflecting the effects of higher stormwater runoff volume and lower biological uptake. In the wet season, the lowest levels of nitrate were consistently found in tributaries with minimal to no agricultural activity, whereas in the dry season the lowest nitrate levels were often found in streams influenced by agriculture. Seasonal variation was greatest for the agricultural sites, and smallest for the urban hillslope streams. This is particularly evident at station A2, the most intensively agricultural sampling site, which exhibited some of the highest nitrate-N values in the watershed in the wet season, but also had some of the lowest values observed in the dry season. One exception to this seasonal trend was seen at the upper station along Chilliwack Creek (Ml9). At this site, nitrate-N concentrations were higher during the dry season. This station maintained relatively high levels of nitrate-N during both the wet and the dry season (2.36 mg/L and 2.98 mg/L, respectively). There are a number of stormwater outfalls that drain into Chilliwack Creek above this sampling site, and it is assumed that these contribute significant amounts of nitrate year round, overwhelming the effects of other contributing factors. 7.2.1.3 Variations in Orthophosphate-P Three sample sets (12-Dec-2002, 3-Mar-2003 and l-May-2003) were not used because of problems with the laboratory analytical technique. Of the remaining sampling sets, almost all values for 27-May-2002 and 9-July-2003 were below detection. Values for 12-Nov-2002 (wet season) and for 21-Aug-2002 and 10-Oct-2002 (dry season) are plotted in Figure 7.4. Overall, orthophosphate-P levels ranged from below detection (0.02 mg/L) to 0.16 mg/L. The highest value was detected at station A17 in August 2002. The spring-fed station (Gl) and the forested control stations (F12, F13, and F14) generally had the lowest orthophosphate-P concentrations in the watershed, suggesting anthropogenic sources may be responsible for higher phosphate concentrations in the rest of the watershed. In contrast, August and/or October orthophosphate levels in Semiault Creek, in the lower sections of both Interception Ditch and Chilliwack Creek, and in some of the urban hillslope tributaries were above levels cited as critical for causing accelerated eutrophication in lakes (0.1 mg/L) (Schreier, pers. comm., 1994). There is no water quality guideline for phosphorus in streams. 101 ^0 .20 |»0.18 | 0.16 | 0.14 8 0.12 c 0 0.10 1 0.08 ** n •=0.06 w 2 0.04 a 5 0.02 r o 0.00 Urban Hillslope (UH) A— —A • — G — — O o o GW Q 0 o F12- F14 U6 U8 U5 U3 U7 U4 A11 A2 G1 Tributaries _ 0.2 ^ 0.18 c" 0.16 o g 0.14 j § 0.12 <S 0.1 i 0.08 I 0.06 Q. S o 0.04 Q. J 0.02 o o F13 A15 A16 A18 M19 M20 Interception Ditch Mainstem to Mouth of Chilliwack Creek 1 km (by distance downstream) Legend: • Grey triangles represent 21-Aug-03 concentrations • Black squares represent 10-Oct-02 concentrations O Orange circles represent 12-Nov-02 concentrations Solid line represents Interception Ditch Mainstem Dashed line represents Chilliwack Creek FH= forested hillslope tributaries (Elkview Creek); U H = urban hillslope tributaries Notes: Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.4 Spatial and Seasonal Variations in Streamwater Orthophosphate-P in the Chilliwack Creek Watershed 102 In contrast to nitrogen, orthophosphate-P values were consistently lower in November (wet season date) than in October/August (dry season dates) in Chilliwack Creek and all tributaries, regardless of land use. This is the opposite of the seasonal trends found in other studies in the LFV (Schendel, 2001; Wemick, 1996; Berka, 1994). This difference may be attributed to dilution as November samples were taken after a high rainfall period. Dilution was noted for other water quality parameters (e.g. specific conductivity) on this date. Concentrations were generally low in the headwaters and increased along Interception Ditch before leveling off towards the mouth of Chilliwack Creek. In August and November there is a pronounced increase between A16 and A18. A different spatial trend was observed along Interception Ditch in October, where the most pronounced increase occurred in the lower reaches (i.e. the headwaters to A16). Along Chilliwack Creek, concentrations increased in the downstream direction as agricultural tributaries (Interception Ditch and Semiault Creek) enter the stream and residential activities increase. 7.2.1.4 Variations in Specific Conductivity Specific conductivity measurements ranged from 97 to 550 /_/S/cm. The highest values were consistently measured in Semiault Creek (A2) and Bailey Ditch (M10). The lowest value was measured at station U3 (Teskey Creek) on 12-Nov-02. Luckakuck Creek (Gl) also had very low values over the entire sampling period suggesting that the groundwater has a lower specific conductivity than surface water. While the difference is minimal, some seasonal variability can be observed in most tributaries and in the lower sections of Interception Ditch (Figure 7.5). As expected, winter conductivity readings are lower because there is more water in the stream, which leads to dilution. This dilution effect is most evident in the Nov-2002 measurements taken under very high flow conditions. It is interesting to note that while Dec-2002 readings were also taken under high flow conditions, conductivity values are high. This could be attributed to the time the sampling took place during the storm event; December samples were probably taken when contaminants were still being flushed into the stream before dilution occurred. These higher conductivity values further emphasize the idea that sediment and manure are entering the stream during runoff events. Streams that were less influenced by land use had similar conductivity values year round. Specific conductivity increased along Interception Ditch with progression downstream. These increases are probably due to the introduction of salts to the watercourse from agricultural sources (such as inorganic fertilizers and animal manures). Nitrate-N is relatively constant along Interception Ditch. However, both ammonia-N and orthophosphate-P increase in concentration with progression downstream. 103 There was little to no variation along Chilliwack Creek, and concentrations were similar to those measured in Interception Ditch suggesting contributions from both urban and agricultural activities in this stream. Differences in conductivity between the forested tributaries suggest that the geology through which Elkview Creek (F13) and Parsons Brook (F12, F14) flows may be different from one another. Elkview Creek has a very low conductivity (values are lower than the most other streams in the watershed year round) due to the fact that the water draining into Elkview Creek flows over quartz-rich metamorphic rocks which do not readily dissolve and therefore contribute less dissolved ions to the drainage water. Since shale, which contains more soluble ions, dominates much of the upland area drained by Parson's Brook, water in this stream is expected to have a higher conductivity. 7.2.1.5 Variations in pH Measured pH values ranged from 6.3 to 8.8, with a mean value of 7.4. The highest pH was measured at station A16 in July 2003. Benchley Creek (U6) and Walker Creek (U8) also had mean pH values above 7.8. The lowest pH was consistently measured at the spring station (Gl) indicating that the groundwater generally had a lower pH than surface water. This lower extreme value is the only value that did not meet B.C and Canadian water quality guidelines for the protection of aquatic life (6.5-9.0). Seasonally, pH values were higher during the summer season in the lower and mid sections of Interception Ditch, in Semiault Creek and in Teskey Way Ditch (A 17). pH in streams usually decreases with increasing temperature, which is opposite to the seasonal variation measured along Interception Ditch. This is possibly due to liming of the lowland agricultural fields, or to the fact that these stations have more aquatic plants in the summer period, which consume C 0 2 during photosynthesis, causing an increase in pH. With the exception of a drop between A16 and A18, pH along Interception Ditch and Chilliwack Creek mainstems showed little variation. 104 E o S5 3__ > O 3 •D C o o u 600 500 400 300 200 00 Forested (FH) Urban Hillslope (UH) Agriculture GW A I F12-F14 U6 U8 U5 U3 U7- U4 A11 A2 G1 Tributaries 600 -g* 500 o CO f 400 > •a c o o o 8 a 300 200 100 F13 A15 A16 A18 M19 M20 Interception Ditch Mainstem to Mouth of Chilliwack Creek (by distance downstream) Legend: A Grey triangles represent dry season means. ^ — Solid line represents Interception Ditch Mainstem • Black squares represent wet season means — Dashed line represents Chilliwack Creek FH= forested hillslope tributaries (Elkview Creek); U H = urban hillslope tributaries Notes: Error bars represent the range of measured values. Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.5 Spatial and Seasonal Variability in Streamwater Conductivity in the Chilliwack Creek Watershed 105 7.2.1.6 Variations in Temperature Water temperatures of the hillslope streams ranged from 4.6°C to 6.6°C in March to maximum temperatures between 13.7°C and 18.7°C in August. One exception to this occurs at station F12 which shows a high December temperature of 16.9°C. This result was considered anomalous and discarded from further analysis. Lowland stations had a higher temperature range (with minimum March temperature between 6.1°C and 10.2°C for all stations, to a maximum temperature of 22.6°C at station A16 in July). Canopy cover and ambient temperature largely governed the variability in stream temperature within the watershed. Both the urban and forested hillslope tributaries were shaded by canopy, whereas the agricultural watercourses had little to no vegetated riparian zones that could offer shade cover. Luckakuck Creek is the exception because it is spring fed. This stream was one of the coolest streams in the summer and one of the warmest streams in the winter reflecting groundwater inputs to the stream. During both the summer and winter, water temperatures along Interception Ditch increase from station F13 towards A16. In the summer temperatures decrease slightly towards station A18, while in the winter they continue to increase over this stretch. One possible explanation is that the stream may be influenced in part by groundwater between these stations. 7.2.1.7 Variations in Dissolved Oxygen Due to equipment problems dissolved oxygen (DO) was only measured on five of the eight sampling dates (Table C.10 in Appendix C). Both DO concentrations and percent saturation (standard saturation concentration tables) were considered. Percent saturation was included because it corrects for temperature effects, and therefore gives a better description of the actual oxygen demand in the water at a specific site. Figure 7.6 shows the fluctuations in the wet and dry season mean DO concentrations. However, it should be noted that 3-Mar-2003 was the only wet season sampling date for which DO was measured, and that this date is not be representative of the entire wet season. Dissolved oxygen levels were quite variable throughout the sampling period ranging from concentrations of 3.5 mg/L (39% DO saturation) at station U5 to supersaturated levels of 15.2 mg/L (>150% DO saturation) at station M9, in August. Station F14 is the only site that consistently had values above saturation, with the exception of one low value of 47% recorded in August 2002. Over the sampling period, two stations (U5 and F14) had concentrations below the 5.0 mg/L guideline for adult and juvenile salmonids. With the exception of station A17, every station was below the 9 mg/L guideline for salmonid embryos at least once during the sampling season. 106 Seasonal differences in DO concentrations were seen in the headwaters of both Chilliwack Creek and Interception Ditch, and in most hillslope tributaries. The higher wet season values at these locations can be attributed to turbulence (higher flows causing increased riffle action) and cooler temperatures in the stream, which promote the dissolving of oxygen in water. Lower biological activity and decomposition rates, combined with increased flushing of organic matter from the system also characterize the wet season; and both of these factors reduce the oxygen demand within the aquatic system (Addah, 2002). While the difference is minimal, the opposite trend was observed in the lower reaches of the Chilliwack Creek and Interception Ditch mainstem. These locations are located in the lowland areas where streamflow is slower. In addition, the abundance of aquatic vegetation present at these sites would release oxygen to the water during photosynthesis; this phenomenon would be emphasized in the dry season when productivity is at its peak. Both DO concentration and % D O saturation showed similar spatial patterns along the Interception Ditch and Chilliwack Creek mainstems. DO concentrations in the headwaters were somewhat higher than concentrations downstream year-round, with one exception. A peak, which is more pronounced in the summer, occurs at station A16 along Interception Ditch. Note also that at its mouth, Chilliwack Creek consistently had the lowest DO levels during both the wet (5.9 mg/L, 53%) and dry season (6.9 mg/L, 69%). 107 E. c 0) >. X O > O to to 5 1 g c 0) s? o ? o (0 (0 18.0 16.0 14.0 12.0 10.0 8.0 6.0 4.0 2.0 0.0 Fo rested (FH) Urban Hillslope (UH) Agriculture GW 1 1 I I 1 1 ill i i i i . — • > -TI I T i \ i • H i 1 *•, \ i i • • ! i i i i i i i - _L ; i i i i i i i 1 -i i i i i i i • i i i i i F12-F14 U6 U8 U5 U3 U7 U4 A11 A2 G1 Tributaries 1 F13 A15 A16 A18 M19 M20 Interception Ditch Mainstem to Mouth of Chilliwack Creek km (by distance downstream) Legend: A Grey triangles represent dry season means. • Black squares represent wet season means Solid line represents Interception Ditch Mainstem Dashed line represents Chilliwack Creek FH= forested hillslope tributaries (Elkview Creek); U H = urban hillslope tributaries Notes: Error bars represent the range of measured values. Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.6 Spatial and Seasonal Variations in Dissolved Oxygen Concentrations in the Chilliwack Creek Watershed 108 7.2.1.8 Variations in Dissolved Elements (Major Ions and Trace Metals) in Water Dissolved streamwater concentrations of 6 elements were considered for all sites: calcium (Ca); iron (Fe); potassium (K), magnesium (Mg), manganese (Mn), and sodium (Na). In addition to these, a further 8 elements (Cd, Cr, Co, Cu, Ni , P, Pb, and Zn) were analyzed but were below their respective detection limits for all sites on all sampling dates. Aluminum was also consistently found below detection limit (0.05 ppm) except in Armstrong Ditch ( A l l , 0.42 ppm) and in Interception Ditch (A 16, 0.06 ppm) on 3-May-2003. Due to the lack of data, these elements are not considered further. A tabular summary of the data is provided in Appendix C, and Table 7.2 shows a statistical summary of the data obtained from the analysis of major ions and trace metals in streamwater for the watershed. Based on the graphs in Figures 7.7 through 7.12, no overall temporal trend is evident for any of the 6 elements. As previously mentioned, a Kruskal-Wallis test showed that for each element there was no significant difference between the median concentrations of the three sampling sets (dates). Spatial Variability of Major Ions (Ca, Mg, Na, K) in Water The lowest concentrations of Ca, K and Mg in the watershed were consistently detected in Elkview Creek (F13). Parson's Brook (F12, F14), Armstrong Ditch (A l 1) and the upper reaches of Interception Ditch (A 15) also had low K and Mg levels. The highest Ca levels were found in Semiault Creek (A2), while Bailey Ditch (M10) and the upper station on Teskey Creek (U5) showed some of the highest Mg levels. Station M19 consistently had the highest K concentrations measured in the watershed, with relatively high concentrations of Caand Mg also detected at this site. Bailey Ditch (M10), Teskey Way Ditch (A 17) and Chilliwack Creek (M20) also showed some of the highest concentrations for all major ions, except Na. Sodium behaved slightly differently than the other major ions. For example, Parson's Brook, which had low concentrations of most ions, showed some of the highest sodium concentrations in the watershed. Upper Thorton/Teskey Creek (U5) had the highest Na levels, while Walker Creek (U8), Benchley Creek (U6), Luckakuck Creek (Gl), and the lower reaches of Lefferson Creek (U7) had the lowest levels of Na. Finally, little to no variability can be seen in Luckakuck Creek (Gl) for any of the major ions, and concentrations at this site are generally low and comparable to concentrations detected at the forested control stations (F12, F13, and F14). 109 Spatial trends with progression downstream were also evident (Figure 7.7 to 7.10). A l l cations showed a definite increase in concentration along the Interception Ditch mainstem from its headwaters to where it joins Chilliwack Creek. The area of Interception Ditch with the most substantial increase occurs downstream of station A16, where Teskey Creek enters the Ditch. (Note that streams draining into Teskey Creek flow through residential developments, agricultural land, and past the Bailey landfill site). The high ion concentrations in Teskey Way Ditch (21.45 ppm, 7.44 ppm and 2.00 ppm, for Ca, Mg, and Kat station A17, respectively) may partially explain the substantial increase along this stretch of Interception Ditch. Concentrations in Chilliwack Creek were equal (Na, Mg) or greater (Ca, K) than concentrations in lower reaches Interception Ditch. Spatially, Ca concentrations decreased and Mg concentrations increased along Chilliwack Creek. Spatial Variations in Trace Metals fFe, Mn) in Water Manganese (Mn) concentrations ranged from below detection (0.005 ppm) to 0.38 ppm, with a mean value of 0.07 ppm. The highest Mn concentration was measured in Bailey Ditch (M10) in May 2003. In addition, Semiault Creek (A2), Chilliwack Creek (M20), and the downstream section of Interception Ditch (A 18, A16) all had concentrations above 0.10 mg/L on at least two of the three sampling dates. Iron (Fe) concentrations ranged from below detection (0.05 ppm) to 1.69 ppm, with a mean value of 0.37 ppm. As with Mn levels, Bailey Ditch (M10), Armstrong Ditch (A l l ) , and the downstream section of Interception Ditch (A 18, A16) had relatively high Fe levels. Semiault Creek (A2), which is intensively agricultural, showed the highest Mn concentrations, but had low Fe concentrations. Spatially, concentrations of Fe and Mn were low in the headwaters and increased along Interception Ditch; however, the trend from Interception Ditch to the mouth of Chilliwack Creek differed. Manganese concentrations are greater at the mouth of Chilliwack Creek than concentrations measured along Interception Ditch, while iron concentrations in Chilliwack Creek remain much lower than those in Interception Ditch. Concentrations of the two metals also increase along Chilliwack Creek which may be due to contributions of the agricultural tributaries that enter the stream. Fe and Mn concentrations in the hillslope tributaries and in Luckakuck Creek (Gl) were consistently low (concentrations are nearly always below or very near detection limit). These spatial variations of Fe and Mn are shown in Figures 7.11 and 7.12, respectively. 110 60 50 Urban Hillslope (UH) s I ^ £ 40 U d> 30 20 10 O o • • F12— F14 U6 U8 U5 U3 U7 U4 A11 A2 G1 Tributaries 60 7 1 km F13 A15 A16 A18 M19 M20 Interception Ditch Mainstem to Mouth of Chilliwack Creek (by distance downstream) Legend: • Grey triangles represent 27-May-02 concentrations • Black squares represent 3-May-02 concentrations O Orange circles represent 9-Jul-03 concentrations Solid line represents Interception Ditch Mainstem Dashed line represents Chilliwack Creek FH= forested hillslope tributaries (Elkview Creek); UH = urban hillslope tributaries Notes: Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.7 Spatial and Temporal Variations in Dissolved Calcium (Ca) in the Chilliwack Creek Watershed 111 F12-F14 U6 Tributaries F13 A15 A16 A18 M19 M20 1 km Interception Ditch Mainstem to Mouth of Chilliwack Creek i (by distance downstream) Legend: • Grey triangles represent 27-May-02 concentrations — Solid line represents Interception Ditch Mainstem • Black squares represent 3-May-02 concentrations — Dashed line represents Chilliwack Creek O Orange circles represent 9-Jul-03 concentrations FH= forested hillslope tributaries (Elkview Creek); U H = urban hillslope tributaries Notes: Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.8 Spatial and Temporal Variations in Dissolved Potassium (K) in the Chilliwack Creek Watershed 112 12 c o 10 c <u c o = c Q. o o. o r c a> •- i c « O) V •o c o — > -oreste Urban Hillslope (UH) Agriculture GW O 4 • ft o o k u F12-F14 U6 U8 U5 U3 U7 U4 A11 A2 G1 Tributaries 1 km F13 A15 A16 A18 M19 M20 Interception Ditch Mainstem to Mouth of Chilliwack Creek (by distance downstream) Legend: A Grey triangles represent 27-May-02 concentrations ^ — Solid line represents Interception Ditch Mainstem • Black squares represent 3-May-02 concentrations — Dashed line represents Chilliwack Creek O Orange circles represent 9-Jul-03 concentrations FH= forested hillslope tributaries (Elkview Creek); U H = urban hillslope tributaries Notes: Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.9 Spatial and Temporal Variations in Dissolved Magnesium (Mg) in the Chilliwack Creek Watershed 113 c o c E I a O b. O « H •g I o n (0 Q 16 14 12 10 F12—F14 U6 U8 U5 U3 U7 U4 A11 A2 G1 Tributaries 16 1 k m Interception Ditch Mainstem to Mouth of Chilliwack Creek —I (by distance downstream) Legend: O Orange circles represent 12-Nov-02 concentrations Solid line represents Interception Ditch Mainstem • Black squares represent 10-Oct-02 concentrations - - - Dashed line represents Chilliwack Creek FH= forested hillslope tributaries (Elkview Creek); U H = urban hillslope tributaries Notes: Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.10 Spatial and Temporal Variations in Dissolved Sodium (Na) in the Chilliwack Creek Watershed 114 1.8 1.6 c o '•S — 1-4 2 E 1 £ 1.2 c r o 2 1.0 o « o E 0.8 8 - 0.4 w 0.2 0.0 Urban Hillslope (UH) F12— F14 U6 U8 U5 U3 U7 U4 A11 A2 G1 Tributaries Legend: A Grey triangles represent 27-May-02 concentrations — Solid line represents Interception Ditch Mainstem • Black squares represent 3-May-02 concentrations — Dashed line represents Chilliwack Creek O Orange circles represent 9-Jul-03 concentrations FH= forested hillslope tributaries (Elkview Creek); U H = urban hillslope tributaries Notes: Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.11 Spatial and Temporal Variations in Dissolved Iron (Fe) in the Chilliwack Creek Watershed 115 c g re _^ +-> c © o c o O a V) a> c re o> c re -o > o (0 in 0.30 0.25 E S 0.20 ^ a> | 0.15 E re I °-10 0.05 0.00 Forested (FH) Urban Hillslope (UH) Agriculture GW m — — o o F12— F14 U6 U8 U5 U3 U7 U4 A11 A2 G1 Tributaries Legend: • Grey triangles represent 27-May-02 concentrations ^ — Solid line represents Interception Ditch Mainstem • Black squares represent 3-May-02 concentrations — Dashed line represents Chilliwack Creek O Orange circles represent 9-Jul-03 concentrations FH= forested hillslope tributaries (Elkview Creek); U H = urban hillslope tributaries Notes: Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.12 Spatial and Temporal Variations in Dissolved Manganese (Mn) in the Chilliwack Creek Watershed 116 7.2.2 Comparison between Land Use Categories: Water Quality Parameters In this section the differences in water chemistry of streams draining the three dominant types of land uses are compared. An examination of the differences between sites grouped by land use will help distinguish the relative contributions of the different land uses for the various contaminants. Specific emphasis is placed on the difference between streams draining the upland urban area and the forested hillslopes since this would provide an idea as to what happens to aquatic systems as the hillslope is developed (i.e. converted from forest to urban). An overview of the differences in water chemistry between the three land uses is presented in Tables 7.3 and 7.4 below. For each parameter, a Mann-Whitney U test was used to make pair-wise comparison between land use categories for wet and dry season data separately, and for both seasons combined. Due to their small sample sizes (n=3), dissolved elements were omitted from the wet and dry season analysis but were included in the statistical analysis of the combined seasonal data. Table 7.3 Overview of Mann-Whitney Comparison between Land Use Categories for Water Quality Parameters (Wet and Dry Seasons) NH/-N N0 3-N PO4 3-P pH Temp Sp. Cond. DO Dry Season Agriculture vs Forest A>F* F>A A>F Agriculture vs Urban A>U* U>A* A>U* A>U* A>U Forest vs Urban U>F Wet Season Agriculture vs Forest A>F A>F* A<F F>A* Agriculture vs Urban A>U A>U U>A A>U A>U Forest vs Urban F>U U>F* A l l values significant at the a=0.0T67 level unless otherwise noted by '*' * Significant at the a=0.05 level Table 7.4 Overview of Mann-Whitney Comparisons Between Land Use Categories for Major Ions and Trace Metals in Water (Seasons Combined) Ca K Mg Na Fe Mn Agriculture vs Forest A>F* A>F* A>F A>F A>F Seasons Combined Agriculture vs Urban U>A* A>U A>U Forest vs Urban U>F U>F Al l values significant at the a=0.0167 level unless otherwise noted by '*' * Significant at the a=0.05 level 117 Seasonal trends may differ depending on the land use of the surrounding area due to differences in environmental conditions and processes. Table 7.5 summarizes the seasonal differences in water chemistry for individual parameters. Table 7.5 Overview of Mann-Whitney Comparison Between Wet and Dry Season Values for Water Quality Parameters NFC-N NO/-N PO43-P pH Temp Sp. Cond. DO Agriculture X X 0 0 Urban X o o X Forest X 0 0 X *Symbology for statistically significant differences (a = 0.05): ' X ' = wet > dry; 'O' = wet < dry Time series graphs were created in order to visualize the trends in water quality parameters for the different land use categories over the sampling period. The data series in these graphs represent trends for agriculture, urban, and forest, based on the mean for each land use category. Values for stations M l 9 and G l were also plotted on this graph to represent water quality of a more intensive urban stream and a spring-fed stream, respectively. These graphs do not show the variability within land use categories; however deviation from the observed trend is noted when it was thought to be important. It should be emphasized that any differences found between the land use categories are not necessarily the result of different inputs from the various land activities. Other abiotic and biotic factors that may influence water quality at a given site include: 1) flow characteristics of the stream (Arheimer et al., 1996); 2) the mechanism of water flow from the land (i.e. the proportion of surface runoff and groundwater, flow paths, and type of runoff event) (Chapman and Kimstach, 1996; Jordan et al., 1997); 3) internal river processes including biological and chemical transformations (Arheimer and Liden, 2000); 4) geology of the drainage area; 5) topography (Herpe and Troch, 2000) and 5) mixing of water from different tributaries with different water quality (Chapman and Kimstach, 1996). Nitrate-N: Within the dry season (summer), nitrate-N concentrations in the agricultural streams were significantly lower than the nitrate-N concentrations of the forested tributaries (p<0.017) and urban tributaries (p<0.05). Agricultural watercourses tend to have more aquatic vegetation and algae growth in the summer; therefore, any nitrate that is on the land or that runs off into nearby watercourses is actively being taken up by the plants and aquatic biota that are growing at this time of year. In addition, the action of nitrifying bacteria is increased at higher temperatures (Heathwaite, 1993); thus, in warmer waters, 118 typical of unshaded agricultural streams in the dry summer season, nitrification (conversion of ammonia to nitrate) is increased. In the wet season, the urban hillslope streams had the lowest nitrate-N concentrations of the three main land use categories. The agricultural area had the highest nitrate-N levels, particularly on sampling dates that occurred during or immediately following a rain event (12-Nov-02 and 12-Dec-02); however, agriculture and forest concentrations were not statistically different. The lack of significant difference could be attributed to natural variability, differences in site conditions or possible unknown sources. Other studies in the LFV have shown that small hobby farms and leakage from septic tanks in rural areas can contribute nitrate to surface waters (Wernick, 1996; Cook, 1994). However, a more detailed study would have to be done to confirm this possibility. The more intensive urban station (Ml9) along Chilliwack Creek had high nitrate-N levels during most of the year. As previously mentioned, storm outfalls are thought to contribute nitrate to the stream year round. The low values in November and December are likely a dilution effect from high stormwater runoff volumes. Nitrate-N concentrations at the spring-fed station were consistently around 1 mg/L throughout the year, below the 3 mg/L that is considered indicative of impacted groundwater (Schreier et al., 1996). Agriculture —•—Urban - A — F o r e s t —©—Spring Fed intensive Urban Figure 7.13 Nitrate-N Comparisons by Land Use Category 119 Ammonia-N: Agricultural ammonia-N concentrations were significantly higher than either the urban or forested ammonia-N concentrations in the wet season. The largest differences between agricultural and hillslope areas occur on dates where significant rainfall occurred (12-Nov-02 and 12-Dec-02). Armstrong Ditch ( A l l ) and the furthest upstream station along Interception Ditch (A15) are exceptions. These two agricultural stations had concentrations comparable to those measured in forest and urban hillslope tributaries. Stormwater outfalls at station M19 do not appear to be contributing ammonia-N to the stream. Unlike nitrate-N, ammonia concentrations in Luckakuck Creek were not constant year round. The difference in the pattern seen for the two nutrients suggests that ammonia levels in Luckakuck Creek may be influenced by land use, while nitrate-N is not. CO co CM o =3 < C\J CO co CO o o o o CO May* "5 a May* 7 co CO i— • Agriculture -Urban • Forest - Spring Fed Intensive Urban Figure 7.14 Ammonia-N Comparison By Land Use Category Orthophosphate: Agricultural and urban sampling sites had significantly higher orthophosphate-P levels than the forested sites, both in November (p<0.05) and for the combined August/October data (p<0.017) Still, it appears that agricultural sampling station had higher concentrations than the urban sampling stations, particularly in the wet season. Station A15 was the only agricultural site that had similar concentrations to any urban or forest sites. 120 Specific Conductivity: The agricultural streams had significantly higher wet season specific conductivity values compared to the urban streams. This implies that agricultural activities in the watershed are a larger contributor of inorganic salts to the watercourses than the urban activities on the hillslope. This is supported by the fact that no significant difference was found between urban and forest. It is suspected that the urban area is not yet large enough to see significant impacts from the introduction of salts typical in urban runoff (such as road salt and fertilizers). Specific conductivity of both the urban and forest areas are primarily influenced by the geology of the hillslope area. The influence typical of a more intensified urban area can be seen in the higher conductivity values observed at station M19 during the dry season when dilution was not an issue. Some variability among the agricultural stations did exist. The major agriculturally influenced station A2 (Semiault Creek) consistently had some of the highest specific conductivity values during both the wet and the dry season, as did station A17. In contrast, Armstrong Ditch station ( A l l ) had much lower conductivity values, particularly in the dry season. The fact that this station receives much of its water from the hillslope area may explain the lower values observed at the site. Figure 7.15 Specific Conductivity Comparisons by Land Use Category 121 Dissolved Oxygen: During the wet season, both DO concentration and % D O saturation were significantly greater in the forest area when compared to the agricultural area. The higher runoff velocity on the hillslopes creates turbulence, which promotes the dissolving of oxygen from the air to the stream. Therefore, the difference in topography between the forest tributaries and lowland agricultural streams may partially explain the higher DO concentrations of the forested tributaries. Since the urban land use is also on the hillslope it is reasonable that there was no significant difference between forest and urban sites. No significant difference was found during the dry season, when runoff is at its lowest. p H : Schendel's (2001) study in the adjacent (Elk Creek) watershed showed soils in the area were generally acidic (3.23 to 6.25), but that the forested hillslope soils were more acidic than the soils sampled in the lowland agricultural area. However, this does not translate to the watercourses. During the wet season, pH values were significantly lower in the agricultural streams than in the hillslope tributaries (forest and urban). No significant difference was found between the urban and forested hillslope tributaries. The higher acidity of the agricultural watercourses is probably due to the addition of chemical fertilizers and manure inputs during runoff events. During the dry season, all land use categories had statistically similar (p>0.017) pH values -•—Agricul ture -m— Urban - A - F o r e s t —*— Spring Fed Intensive Urban Figure 7.16 pH Comparisons by Land Use Category 122 Major Ions in Water (Ca, Mg, K, Na): When comparing concentrations for major ions, it can be observed that calcium is the major ion with the broadest range and highest mean concentration (13.7-53.1, 28.0 mg/L). Overall, the following sequence, going from least to most abundant, can be established: K<Mg<Na«Ca when all sites are considered. When separated by land use the sequence is the same, although the relative difference in concentration differs. For example, for forested sites, Mg concentrations are much lower than Na concentrations (Table 7.2, Figure 7.17). 30.3 24.2 20.8 21.6 38.7 Agriculture Urban Forest Spring-Fed Intensive (G1) Urban (M19) H Potassium • Magnesium B Sodium • Calcium Figure 7.17 Major Ions (K, Mg, Na, Ca) Comparison by Land Use Category The magnesium and potassium cations appear to be strongly influenced by land use. Overall, tributaries on the forested hillslope had significantly lower (p<0.017) concentrations of Mg and K. These results are similar to the spatial trends found in soils of Elk Creek (an adjacent watershed). Schendel's (2001) study of Elk Creek soils found that Ca, Mg, and K concentrations were significantly lower in the upland areas than the lowland areas, while Na was not found to exhibit any spatial trends. Differences between land use could be due to fertilizer application to urban lawns and agricultural fields. It is interesting to note that at the 0.05 level magnesium concentrations of the urban streams are also significantly greater than concentrations in agricultural streams. It is uncertain what the source of these elevated levels is, but they could be a result of construction activities in this portion of the watershed. While concentrations of K and Mg at the agricultural sites were generally higher than concentration detected at the forest sites, some variability did exist. Agriculture stations that are closer to the hillslope (A15 and A l l ) had concentrations similar to forested sites, whereas downstream stations along Interception Ditch (A16, A18), station A17, and sites which do not receive drainage waters directly from the hillslopes (A2), had much higher concentrations. 123 Trace Metals in Water (Fe, Mn): For trace metals, iron concentrations were higher than manganese concentrations. The difference between the two is significantly larger for agriculture sites (0.66 mg/L), than for forest (0.07 mg/L) or urban (0.08 mg/L) sites. 0.80 mg/L Agriculture Urban Forest Spring-Fed Intensive (G1) Urban (M19) Figure 7.18 Dissolved Trace Metals (Fe, Mn) Comparison by Land Use Category Both iron and manganese concentrations were the highest for the agricultural land use category (Table 7.4). Metals such as Mn and Fe are added to animal feed as nutrient supplements, and as a consequence animal manures can contain high concentrations of these metals and surface water can be at risk of metals leaching from fields receiving livestock manure. The occurrence of iron and manganese in aqueous solution is also largely dependent on environmental conditions, particularly conditions that influence oxidation and reduction reactions (Chapman and Kimstach, 1996). The lower pH in the agricultural streams could promote the release of iron from sediments. 124 7.2.3 Variations with Discharge A Spearman's rank correlation was used to determine the relationship between water quality indicators and streamflow. This was done separately for each monitoring station that was adjacent to a hydrometric station; however no significant relationships were found. It is thought that the lack of data, particularly during high flow, most likely inhibited determining any relationships between water quality parameters. 7.2.3.1 Influence of Storm Events on Water Chemistry Water samples were collected at two agricultural sites (A2, A18), one urban site (U3), and one forested site (F14) during a storm event on 16-Oct-2003, and analyzed for nitrate and dissolved trace elements. Samples were collected four times at each station, approximately every 75 minutes. In addition, one sample was also taken at the mouth of Chilliwack Creek (station M20) near the end of the sampling period. Considering that the hydrologic response of urban watersheds is typically fast, and that runoff frequently occurs in surges (which would be accompanied by pulses in contaminants to the streams), this is considered somewhat 'sparse' sampling, particularly if loads are to be determined (Macdonald et al., 1997). Although samples were analyzed for the presence of a number of metals, only A l , Mn, Ca, Fe, K, Mg and Na were consistently present in concentrations above detection at all stations (Table C.13 in Appendix C). In contrast to baseflow sampling, A l was also above detection for all sites, and Zn was above detection at station A2 (Semiault Creek). The sampled event was the first large storm (>100 mm) of the winter rainy season following an unusually long dry season. Unfortunately, the sampling period only covered the initial stages of the storm; the majority of the runoff occurred after sampling ended (Figure 7.19). In addition there is no precipitation data from the tipping buckets, and only two of the hydrometric stations (Semiault and Teskey) were operational at the time. The 'urban' sampling set, taken at station U3, is located in the same spot as the Teskey flow station. There were no hydrometric stations directly adjacent to either of the 'agricultural' sampling stations. The hydrometric station along Semiault Creek is located approximately 3.3 km downstream of the sampling station A2 (and therefore is probably not entirely representative of flow at point of sampling). Over the sampled period of the storm, the Teskey Creek hydrograph shows a minor peak shortly before the main storm runoff began; flow at the Semiault station only begins to increase near the end of sampled period. Despite missing the major storm peak, when the water quality data are plotted with the progression of the storm event there is evidence that the storm runoff is already influencing water chemistry (Figure 7.20 and 7.21). 125 0.035 0 .030 -ro c 0.025 -c 0.020 -a> ^ w a> l- 0.015 -CB a> 0.010 -n j= u </> 0.005 -5 0.000 12 16 2 0 24 28 32 36 4 0 4 4 Hours beginning 16-Oct-2003 12:00 AM 0.5C </> 0.00 Figure 7.19 Storm Hydrograph for the 16-Oct-2003 Storm Event Nitrate and metals linked to agricultural activities, particularly animal manure (Mn, Zn), showed significant increases in concentration at the agricultural stations during the storm event. In Semiault Creek (A2), nitrate, manganese and zinc showed particularly strong increasing trends reaching maximum concentrations of 7.20 mg/L, 0.35 mg/L and 0.06 mg/L, respectively. Station A l 8 showed a similar trend (but with lower concentrations) for Mn and nitrate. Overall, concentrations were greater than the concentrations recorded during monthly (baseflow) sampling, and it is thought that concentrations would continue to increase throughout the storm until dilution started. Fe and A l data trends observed at the urban station (U3) and the downstream agricultural station (A18) were remarkably similar. Samples from both stations showed an increase in concentration until 5 hours (±1 hour) after the onset of the storm, followed by a decline to initial levels. This corresponds to about a 3 hour lag time (±1 hour) between peak runoff and peak concentrations. Peak concentrations of 0.75 mg/L Fe and 0.78 mg/L A l were measured at the urban site, and higher peak concentrations of 1.24 mg/L Fe and 1.12 mg/L A l were recorded at the downstream agricultural station A18. At station A2 dissolved aluminum and iron concentrations are lower and continually increase over the course of the sampling to maximum concentrations of 0.41 mg/L Fe and 0.43 mg/L A l . As expected, concentrations at the forest station remained low with minimal change over the storm event. 126 Legend: -»-Forest(F14-Elkview) —ir- Apiculture (A2 - S emiault C te ek) — U rban (Station U 3 - Te skey) Agic\Jture(A18-IctefceplicBiDitch X Cumulatrre (M20 - ChiDsyack) Hours beginning 16-Oct-2003 12:00 PM Figure 7.20 Response of Zinc (Zn), Manganese (Mn) and Nitrate During the 16-October-2003 Storm Event 127 Legend: -"-Forest(F 14-Elkview) —*— Agiculture (A2 - Semimjt Creet^  # Urban(StationU3 - Teskey) - -'— Agicuture (A 18 -1 nterc eption Ditch * C umulatr/e (M20 - Cbiliwack) Hours beginning 16-Oct-2003 12:00 PM Figure 7.21 Response of Aluminum (Al) and Iron (Fe) During the 16-October-2003 Storm Event 128 7.3 Results for Sediment Parameters Streambed sediments were collected in October 2002 and again in July 2003, and analyzed for carbon and nitrogen content, and for trace elements. The July sediment sample set was further analyzed for bio-available phosphorus (as orthophosphate) and texture. Results for all these properties are shown in Appendix D. 7.3.1 Sediment Properties: Particle Size The smaller sediment size fractions have a relatively higher adsorptive capacity, and thus, it is expected that this finer fraction (generally O.063 mm) will have a higher concentration of metals. This silt and clay component also has been shown to have a greater potential to become re-suspended. Consequently, this fraction is important in understanding site contamination. Figure 7.22 shows the percentages of sand, silt and clay at each site on the July-2003 sampling date. Figure 7.22 Percentages of Sand, Silt and Clay at Each Sampling Station Overall, particle sizes were quite variable throughout the watershed. The clay fraction ranged from 1.3 to 17.3%, the silt fraction from 0.9 to 66.1%, and sand fraction from 19.3 to 97.7%. While the sand, silt and clay content of the sediment did not show any significant differences between the three major land use categories (forest, urban, agriculture), sediments sampled from upland streams appeared to have a higher sand content. Topography is one of the main factors controlling sediment transport at the catchment scale; the steeper grade of the hillslope area through which these stream flows is steep enough to move finer sediments downstream even under summer flow conditions. Two exceptions are Elkview Creek (F13) and upper Teskey Creek (U5) which showed lower percent sand (65.8% and 41.8%, respectively) 129 than the other hillslope stations. Sampling station F13 is at the transition between the hillslopes and the flat agricultural land; as a result, water at this point slows rather abruptly and finer sediments will begin to settle out. In contrast, significant clay-silt fractions (<0.053 mm) were found in most lowland streams (Semiault Creek (49.7%), Armstrong Ditch (42.5%), Interception Ditch (21.3 to 50.7%), and the mouth of Chilliwack Creek (50.1%)). The largest percent fines (<0.053 mm, clay + silt) was found in Bailey Ditch (M10) with 80.7%. Sampling was done in the low flow period when water in these lowland streams and agricultural ditches tends to be stagnant, which allows the fine particles to settle out. Station M19 and A17 are exceptions, and had the highest percentages of sand (>95%) in the sediment collected in the watershed. 7.3.2 Variations in % Carbon and % Nitrogen Percent carbon and percent nitrogen in the sediment were used as a measure of the organic matter content. Percent total carbon (%C) varied from 0.35% at station U3 to 7.04% at station G l . Nitrogen content ranged from 0.03% to 0.51%, and was correlated strongly with percent carbon (r = 0.955). As with carbon, the highest % N was measure in Luckakuck Creek (Gl). Carbon and nitrogen showed similar trends. In general, the carbon and nitrogen content in the sediment collected from the lowland agricultural area did appear elevated compared to that of the sediment collected from the upland areas. This is reflected in the Mann-Whitney tests for comparison between land use categories for % N . • % Carbon in Sediment • % Nitrogen in Sediment (/10) Figure 7.23 % Carbon and % Nitrogen for July 2003 Sediment Samples 130 7.3.3 Metals in Sediment Sediment samples were taken at the end of the dry season (extended period of low flow conditions), and therefore the total trace metal concentration in sediments was used as an indicator of accumulated metal contamination over the previous year. The initial chemical analysis for sediments focused on the total concentration of twenty-two elements (Al, As, Ba, B, Ca, Cd, Cr, Co, Cu, Fe, K, Mg, Mn, Mo, Na, N i , P, Pb, Se, Si, Sr and Zn) at each sampling site. A complete tabular summary of the sediments metal data is provided in Appendix D. Of these, concentrations of four elements (As, B, Mo, Se) were consistently below their respective detection limit and will not be considered further. Phosphorus is discussed separately in a subsequent section. 7.3.3.1 Temporal Trends An overview of the differences in metals between the two sampling dates is shown in Table 7.6 Sediment samples taken in October 2002 generally had higher metal concentrations than samples taken in July 2003, for both the agricultural and urban land use categories. However, these differences may simply reflect the natural variability in trace metals due to the physical and chemical properties of the sediment, variability in streamflow, and in-stream conditions. The urban area is in a development stage, and therefore soils and land in the area are continually being disturbed as new developments are built. Therefore, it is also possible that while low flow sampling is stable in the undisturbed forested hillslope, the development of the urban area on the hillslope may be impacting trace metal concentrations. Table 7.6 Overview of Total Metal Concentrations in Sediments Showing a Significant Difference Between October 2002 and July 2003 Sediment Sampling Sets Land Use Elements showing a significant difference at a = 0.05 Agriculture* A l , Cr, Cd, Cu, Fe, K, Mg, Ni , Zn, Co Urban A l , Cd, Co, Cr, Fe, K, Mg, Na, N i , P, Ca Forest None Combined Data* A l , Cr, Cu, Fe, K, Mg, Ni , Zn, Co, Na, Mn, Ca excludes station A2 since samples were only collected in July 2003 7.3.3.2 Spatial Trends The range and mean of trace metal concentrations measured in the streambed sediment of Chilliwack Creek and its tributaries are presented in Table 7.7. These data are separated by land use, and compared to background concentrations and concentrations of these elements found in sediments for other studies in the Lower Fraser Valley. Because concentrations of metals in sediment may be enriched through natural processes (such as weathering), and consequently, influenced by the composition of local soils and 131 geology, information on background concentrations can help in determining the extent to which human activities have contributed to the concentrations of sediment associated metals. As with water quality data, results are presented graphically to visualize changes in metal concentrations from the upstream to downstream direction for the Interception Ditch and Chilliwack Creek mainstems and their tributaries. Both the October 2002 and July 2003 samples are shown. Most metals are not above natural background concentrations when compared to the data for Vancouver region sediments (Table 7.7), and match values measured at reference sites from other studies in the LFV. However, if we compare the data to the reference sites from Smith (2004) located on Vedder mountain, Cd, Mn, Zn, Fe, Mg, Ca, Na, and K all showed higher concentrations than those measured at this reference site. This thesis will focus on the results of iron (Fe), copper (Cu), zinc (Zn), cadmium (Cd), manganese (Mn), magnesium (Mg), sodium (Na) and potassium (K), as these elements exhibited the most interesting spatial variation throughout the watershed. 132 Table 7.7 Metal Concentrations in Sediment for the Chilliwack Creek Watershed, and Comparison with Natural Background Levels and Other Studies in the Lower Fraser Valley (LFV) 2002-2003, Chilliwack Creek Streambed Sediment Sampling (mg/kg)1 A l Fe M g C a Na K Si Sr Ba ALL SITES Mean 11250 33898 5370 4373 171 509 1355 32.3 Range 6927-17014 15328-89488 2204-22565 2568-9900 89-309 232-1176 904-2082 20.2-51.8 137 59 - 294 AGRICULTURAL SITES Mean 12009 43825 5512 4270 208 550 1366 32.1 Range 8188-15245 22027-89488 2854-8387 3362-6525 130-303 232-1006 1037-2082 23.9-43.4 166 109-294 URBAN SITES Mean 10859 22153 4083 3991 134 459 1300 28.4 Range 6927-15964 15328-30322 2204-6207 2568-5356 89-209 347-655 1079-1836 20.241.3 117 91-145 FOREST SITES Mean 12642 30671 5289 4111 123 331 1451 40.5 Range 9517-17014 2228344339 3090-8106 2861-5222 100-156 246-460 936-1900 28.7-50.5 121 79-185 Background Concentrations Western US sediments (<63um fraction)" Mean sediments 7.2% 4.1% 1.4% 60% 0.6% 2.0% 24.5% 230 460 Upper Illinois R. Basin low order streams median (<63 /urn fraction)0 2.9% NTS 92G Vancouver map sheet (<177 fim fraction), mean and range' 2.02% (0.4-10.5)% FORESTED CONTROL SITES IN LFV STUDIES Vedder M m 2004 (Sumas Watershed)0 12195 20964 10296 5929 181 202 616 Streambed Sediment Concentrations (Impacted watersheds in the LFV) Sumas River 9514 48693 63297 4354 551.4 487.8 634 AGRICULTURAL (Abbotsfordf (5238-12318) (39866-73763) (8628-126150) (2967-6028) (231-911) (359-725) (427-878) WATERSHEDS Agassiz/Harrison Hot 5% (1.9-8.7)% Springs'1 , mean and range RURAL RESIDENTIAL/ Salmon River (Langley), Aug 19918 mean and range 6400 48700 7540 12189 14636 32.2 AGRICULTURAL WATERSHED (45600-84400) (33100-114000) (4930-11800) (6830-16800) (9560-19600) (16.0-79.8) URBAN WATERSHED Burnette River1 (Vancouver), median and range 2.0% (596248651) 3424 (247-8087) Combest (1991) cited in Cook (1994) b Wedepohl (1968) cited in Salomons and Forstner (1984) c Colman and Sazolon (1992) cited in Cook (1994) d BCMOEMPR (1990) cited in Cook (1994) e unpublished data, I. Smith (pers. comm..) 8 Cook (1994) h Addah (2002), dry season data only, omitting spring and control stations 'McCallum(1995) ' all concentration are in mg/kg dry weight, unless otherwise noted Table 7.7 (cont.) Metal Concentrations in Sediments for the Chilliwack Creek Watershed, and Comparisons with Natural Background Levels and Other Studies in the Lower Fraser Valley (LFV) 2002-2003, Chilliwack Creek Streambed Sediment Sampling (mg/kg)1 C d Co C r C u M n Ni P Pb Z n ALL SITES Mean 4.0 12.7 29.2 38.9 763 27.9 1150 45.1 120 Range 2.5-11.2 10.0-27.7 18.4-64.9 17.9-80.8 142-1805 16.2-45.1 500-5110 25.8-76.2 54-265 AGRICULTURAL SITES Mean 4.9 13.6 28.6 40.9 591 28.2 1426 149 Range 2.9-11.2 10.0-27.7 20.1-35.9 29.0-56.6 226-1318 17.9-34.5 958-2884 84-218 URBAN SITES Mean 2.7 11.3 27.9 26.8 952 27.3 716 81 Range 2.5-3.2 10.0-14.4 18.4-35.7 17.9-33.5 446-1656 16.2-37.2 620-849 54-119 FORESTSITES Mean 3.8 15.6 27.5 44.2 838 31.7 702 114 Range 2.5-5.1 10.0-24.0 20.7-34.6 24.9-63.1 359-1446 20.2-45.1 500-862 65-154 Background Concentrations Western US sediments (<63u.m fraction)3 20-210 0-110 9-52 49-510 Mean sedimentsb 0.17 14 72 45 770 52 670 19 95 Upper Illinois R. Basin low order streams median (<63 nm fraction)0 56 23 26 27 100 N T S 92G Vancouver map sheet 8 44 26 322 7 7 48 (<177 nm fraction)d (1-32) (12-515) (2-415) (53-2100) (1-165) (1-140) (10-1000) FORESTED V e d d e r M t n 2 0 0 4 CONTROL SITES IN , „ . . . ^ , , . E T „ , , (Sumas Watershed) L F V STUDIES V ' 15.7 51.2 61.1 443 33.6 620 Bd 47.4 Streambed Sediment Concentrations (Impacted watersheds in the L F V ) Sumas River 38.2 111.9 28.5 1770 571.9 1539.7 25.5 81.1 AGRICULTURAL (Abbotsford)e (11.2-69.3) (39.6-186) (17.1-45.2) (261-7232) (80.2-1148) (312-3296) (Bd-29.6) (53-109) WATERSHEDS Agassiz/Harrison Hot Springs11, mean and range 54 (30-148) 442 (189-1296) 54 (30-149) 145 (0-737) RURAL RESIDENTIAL/ Salmon River (Langley???)8 mean and range 26.6 124 43.1 177 27.2 166 39.1 170 A GRICUL TURAL A TERSHED (21.1-30.9) (115-140) (26.9-84.0) (53-684) (16.2-43.3) (64-373) (10.8-67.4) (92-344) URBAN WATERSHED Burnette River1 (Vancouver), median and range (<3-17) 28 (5-141) 56 (25-279) 807 (194-3402) 14 (<6 - 52) 63 (22-407) 143 (60-391) a Combest (1991) cited in Cook (1994) E Cook (1994) b Wedepohl (1968) cited in Salomons and Forstner (1984) h Addah (2002), dry season data only, omitting spring and control stations c Colman and Sazolon (1992) cited in Cook (1994) 1 McCallum (1995) d BCMOEMPR (1990) cited in Cook (1994) ' all concentration are in mg/kg dry weight, unless otherwise noted e unpublished data, I. Smith (pers. comm..) Manganese Manganese concentrations in the watershed ranged from 142 to 1805 mg/kg, with a mean of 763 mg/kg. The lowest Mn concentrations were measured in Luckakuck Creek (G l , mean 227 mg/kg), Armstrong Ditch ( A l l , mean 304 mg/kg), Bailey Ditch (M10, mean 343 mg/kg) and along Interception Ditch. The highest Mn concentration (1805 mg/kg) was recorded at station M l 9 . This station is located below a number of outfalls which drain water from an area that is more intensively urbanized than the Promontory area. The increase in impermeable surface area and traffic intensity of this urban area could be responsible for these higher Mn concentrations (as manganese based fuel additives get flushed into the stream during storm events). A number of stations in the Promontory region also showed high Mn levels including Benchley Creek (U6, mean 1138 mg/kg), and the lower station along Teskey Creek (U3, mean 1199 mg/kg) and Lefferson Creek (U4, mean 1232 mg/kgj. Spatially, Mn in sediment was higher in the headwaters (Oct sampling set only) than along Interception Ditch (see Figure 7.24). Concentrations increased along Interception Ditch itself (with a peak between station A16 and A18 where low dissolved Mn concentrations were recorded in water samples on the same date); concentrations then increase past A18 towards the mouth of Chilliwack Creek. High Mn concentrations were recorded at the upstream (intensive urban) station along Chilliwack Creek, and decreased at the mouth (M20). It is interesting to note the sites with high Mn in sediment generally had lower Mn in water, and vice versa. Lead Lead concentrations at most sites were below the detection limit (<25 mg/kg), with a few exceptions. Notably, the highest lead concentrations were recorded in Luckakuck Creek (Gl) with a mean of 73.1 mg/kg. Sampling stations along Chilliwack Creek (Ml9, M20) also showed higher lead levels in sediments, with concentrations of 56.5 mg/kg and 50.2 mg/kg, respectively. Concentrations above detection were also recorded in Teskey Way Ditch at station M9, in the lower reaches of Teskey Creek (U3), and in Parson's Brook (F12, F14) in July 2003; however, these levels were below the interim sediment quality guideline for lead (ISQG) (35 mg/kg). 135 20 18 .£ ~ 16 _ o S •2 •&> 14 S 1 c 2 0) o 1 x O (0 o i C 0) 2 | 2 o o ~ H Forested (FH) 12 10 Urban Hillslope (UH) Agriculture GW F12-F14 U6 U8 U5 U3 U7 U4 A11 A2 G1 Tributaries F13 A15 A16 A18 M19 M20 Interception Ditch Mainstem to Mouth of Chilliwack Creek 11 km I (by distance downstream) Legend: • Black squares represent 10-Oct-02 samples A Grey triangles represent 09-Jul-02 samples. Solid line represents Interception Ditch Mainstem Dashed line represents Chilliwack Creek FH= forested hillslope tributaries (Elkview Creek); U H = urban hillslope tributaries Notes: Error bars represent the range of measured values. Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.24 Spatial and Temporal Variations of Total Manganese (Mn) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 136 Iron, Copper, Zinc and Cadmium The lowest concentrations of these metals were found in the urban hillslope tributaries (Benchley Creek, Walker Creek, Teskey Creek, and Lefferson Creek), Parsons Brook (F13), Armstrong Ditch 1 (A l l ) , and in the upstream section of Interception Ditch (A 15). Metal concentrations at the spring-fed station were also generally low, with the exception of Cu (42.5 mg/kg), which had values comparable to the agricultural stations. Apart from iron, which is found in high concentrations in local rocks and soils, concentrations of these metals in these tributaries were consistently below sediment quality guidelines. Iron (Fe), copper (Cu), zinc (Zn) and to a lesser extent cadmium (Cd) are metals typically associated with agricultural activities. It is not surprising, therefore, that the high concentrations of these metals were found in streams associated with agricultural activities: Semiault Creek (A2), Chilliwack Creek (M19, M20), Bailey Ditch (M10), and station A17 had high levels of all four metals. Fe and Cd, in particular, seem to be associated with agricultural activities, with the highest concentrations of Fe (89 488 mg/kg) and Cd (11.2 mg/kg) recorded in Semiault Creek (the most intensive agricultural station). When plotted in an upstream to downstream direction, as shown in Figures 7.25 to 7.28, these four metals showed similar trends. Concentrations along the mainstem generally increased from station F13 to the mouth of Chilliwack Creek (M20), with a drop exhibited between station A16 and A18 for the July 2003 sampling date. The drop is not likely an error since it is found for all four metals, but instead may be due to variability in physical or chemical properties of the sediment or differences in the site conditions (e.g. reducing conditions). For example, there was a peak in the DO and pH of streamwater at station A16 on the July 2003 sampling date. The increasing downstream trend suggests a potential cumulative impact of agriculture on these metals in sediment. Concentration of Fe and Cu also increased along Chilliwack Creek reaching concentrations of 882449 mg/kg (Fe) and 80.8 mg/kg (Cu) at the mouth, while Zn and Cd were high along the entire length of the stream. The difference in trends along Chilliwack Creek suggest that agricultural tributaries are the main contributor of Fe and Cu to Chilliwack Creek, while urban activities are the primarily source of Zn and Cd. ' A l l showed a very high Zn concentration in October, but a much lower value was observed in July. Without further sampling it is uncertain whether the high concentration was due to error, or a due to a source that was not present in July 2003 137 100 I? o "&> "« £ is o c o © o O l -C X O (0 o £ & I 11 £ co 80 60 40 20 100 90 80 0 o) 1 ^ 7 0 | S 60 g * 50 o & * § 40 - £ 2 =5 30 o o> H CO c 20 10 Forested (FH) Urban Hillslope (UH) Agriculture GW : • i X F12-F14 U6 U8 U5 U3 U7 U4 A11 A2 G1 Tributaries F13 A15 A16 A18 M19 M20 Interception Ditch Mainstem to Mouth of Chilliwack Creek ^ 1 k m j (by distance downstream) Legend: • Black squares represent 10-Oct-02 samples A Grey triangles represent 09-Jul-02 samples. Solid line represents Interception Ditch Mainstem Dashed line represents Chilliwack Creek FH= forested hillslope tributaries (Elkview Creek); UH = urban hillslope tributaries Notes: Arrows indicate the approximate location of tributaries draining into the mainstem Figure 7.25 Spatial and Temporal Variations of Total Iron (Fe) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 138 100 90 Forested (FH) F12- F14 U6 U8 U5 U3 U7 Tributaries U4 A11 G1 F13 A15 A16 A18 M19 M20 Interception Ditch Mainstem to Mouth of Chilliwack Creek | 1JEJ | (by distance downstream) Legend: • Black squares represent 10-Oct-02 samples • Grey triangles represent 09-Jul-02 samples. Solid line represents Interception Ditch Mainstem Dashed line represents Chilliwack Creek FH= forested hillslope tributaries (Elkview Creek); U H = urban hillslope tributaries Notes: Arrows indicate the approximate location of tributaries draining into the mainstem Figure 7.26 Spatial and Temporal Variations of Total Copper (Cu) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 139 O o) ss S E o >-u w i | 75 «0 13 .E 12 10 i Forested (FH) Urban Hillslope (UH) Agriculture GW F12— F14 U6 U8 U5 U3 U7 U4 A11 A2 G1 Tributaries F13 A15 A16 A18 M19 M20 Interception Ditch Mainstem to Mouth of Chilliwack Creek 1 k m (by distance downstream) Legend: • Black squares represent 10-Oct-02 samples A Grey triangles represent 09-Jul-02 samples. Solid line represents Interception Ditch Mainstem Dashed line represents Chilliwack Creek FH= forested hillslope tributaries (Elkview Creek); UH = urban hillslope tributaries Notes: Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.27 Spatial and Temporal Variations of Total Cadmium (Cd) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 140 30 F12-F14 U6 U8 U5 U3 U7 U4 A11 A2 G1 Tributaries F13 A15 A16 A18 M19 M20 Interception Ditch Mainstem to Mouth of Chilliwack Creek 1 km (by distance downstream) Legend: • Black squares represent 10-Oct-02 samples • Grey triangles represent 09-Jul-02 samples. Solid line represents Interception Ditch Mainstem Dashed line represents Chilliwack Creek FH= forested hillslope tributaries (Elkview Creek); U H = urban hillslope tributaries Notes: Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.28 Spatial and Temporal Variations of Total Zinc (Zn) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 141 Magnesium, Potassium, and Sodium The highest sediment concentrations of these three metals (K: 1176 mg/kg, Mg: 22565 mg/kg, Na: 309 mg/kg) were measured in Luckakuck Creek (Gl) in October 2002. However, July concentrations were 1.9 (Na) to 5.7 (Mg) times lower than the October concentrations at this station. Throughout the rest of the watershed Mg ranged from 2204 to 8386 mg/kg, K from 232 to 1006 mg/kg, and Na from 89 to 303 mg/kg. Magnesium concentrations were slightly higher in upper Parsons (F12) and the lower reaches of Interception Ditch and Chilliwack Creek (M20), while potassium concentrations were higher at stations M l 9 , A17 and M10. High Na concentrations were recorded in Armstrong Ditch (A l l ) , Teskey Way Ditch (A 17), and upper reaches of Chilliwack Creek (M l 9). Spatially, concentrations for all three elements were generally low in the headwaters and increased in the downstream direction along Interception Ditch, and then remained constant or decreased slightly towards the mouth of Chilliwack Creek (Figures 7.29 through 7.31). Slight variations to this trend were observed in October including a sharper increase in Na between A15 and A16, and a slight drop in Mg levels at A16. K and Na concentrations decreased from M l 9 to M20 along Chilliwack Creek. 142 Forested (FH) — F12-F14 U6 U8 U5 U3 U7 U4 A11 A2 G1 Tributaries 25 T g> 20 F13 A15 A16 A18 M19 M20 Interception Ditch Mainstem to Mouth of Chilliwack Creek 11 km | (by distance downstream) Legend: • Black squares represent 10-Oct-02 samples — Solid line represents Interception Ditch Mainstem • Grey triangles represent 09-Jul-02 samples. — Dashed line represents Chilliwack Creek FH= forested hillslope tributaries (Elkview Creek); UH = urban hillslope tributaries Notes: Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.29 Spatial and Temporal Variations of Total Magnesium (Mg) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 143 F12-F14 U6 U8 U5 U3 U7 U4 A11 A2 G1 Tributaries F13 A15 A16 A18 M19 M20 Interception Ditch Mainstem to Mouth of Chilliwack Creek i 1 k m i (by distance downstream) Legend: • Black squares represent 10-Oct-02 samples — Solid line represents Interception Ditch Mainstem • Grey triangles represent 09-Jul-02 samples. — Dashed line represents Chilliwack Creek FH= forested hillslope tributaries (Elkview Creek); U H = urban hillslope tributaries Notes: Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.30 Spatial and Temporal Variations of Total Potassium (K) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 144 40 35 c § & 3 0 I I 25 § ° o S O c | I 10 5 0 40 35 c O) 30 o E M "B> entra E o 25 u c o CO 20 o c CO z 0 ) E 15 To *-* 1 o H </> c 10 Urban Hillslope (UH) Agriculture GW • • • A A a M ok • & A- A F12-F14 U6 U8 U5 U3 U7 U4 Tributaries A11 A2 G1 F13 A15 A16 A18 M19 M20 Interception Ditch Mainstem to Mouth of Chilliwack Creek 11 km | (by distance downstream) Legend: • Black squares represent 10-Oct-02 samples • Grey triangles represent 09-Jul-02 samples. Solid line represents Interception Ditch Mainstem Dashed line represents Chilliwack Creek FH= forested hillslope tributaries (Elkview Creek); U H = urban hillslope tributaries Notes: Arrows indicate the approximate location of tributaries draining into the mainstem. Figure 7.31 Spatial and Temporal Variations of Total Sodium (Na) Concentrations in Streambed Sediments in the Chilliwack Creek Watershed 145 7.3.4 Phosphorus in Sediment Sediments were analyzed for two different forms of phosphorus: bio-available phosphorus (BAP) measured as orthophosphate-P and total phosphorus (TP). The analytical results are listed in Appendix D. 7.3.4.1 Bio-Available Phosphorus Bio-available phosphorus (or orthophosphate) concentrations ranged from below detection limit (for 7 of 20 sites) to 29.9 ppm in Luckakuck Creek. From Figure 7.33 it appears that sediment samples from streams in the upland residential area (U3-U8) had slightly higher BAP concentrations. However, a series of Mann-Whitney tests did not show any statistically significant differences between agriculture, forest and urban land use types. A number of non-urban streams also showed higher BAP concentrations, including Teskey Way Ditch (A17: 28.5 ppm), Parson's Brook (F14: 24.3 ppm; F12: 14.0 ppm), Luckakuck Creek (G l : 29.9 ppm), Armstrong Ditch (A l 1: 12.6 ppm) and Chilliwack Creek (M19: 27.2 ppm). 7.3.4.2 Total Phosphorus Total phosphorus (TP) ranged from 500 mg/kg to 5110 mg/kg throughout the watershed, with a mean of 1150 mg/kg. The hillslope tributaries (both forest and urban) generally had low TP concentrations, while concentrations in the lowland agricultural ditches were higher (particularly in Semiault Creek and at station A16 along Interception Ditch). Spatially, TP concentrations increased from the headwaters of Interception Ditch to the mouth of Chilliwack Creek, with a drop between A16 and A18 (after confluence with urban hillslope tributaries). While concentrations in the upper reaches of Chilliwack Creek (Ml 9) are comparable to concentrations along Interception Ditch (reflecting that some phosphorus is originating from urban sources), TP increases in the downstream directions as tributaries draining the large agricultural area in the lowland (e.g. Semiault Creek and Interception Ditch) enter the Chilliwack Creek. These trends are shown in Figure 7.33 below. When land use types were compared, TP in the sediments from the agricultural lowland streams were significantly higher than from hillslope tributaries (both forest and urban). This was expected as agriculture is known to contribute phosphorus to watercourses. 146 Forest Urban Hillslo CD CD E, o o o X Q_ *-> o 0 ) 0 0 0) i - CM i - 5 If) CO 00 l - h- CM 1- 1- T - T- T - < < < < < < CO CM s ' t in n oo ID 3 D 3 3 => =3 35 30 25 20 15 10 5 0 E Q. Q. CU +-> CO sz Q. CO o JC a o .c •c o Figure 732 Total P and Bio-Available (orthophosphate) Concentrations in Sediment, July 2003 Sampling 60 TH ~ 5 0 CD O O) * E CD h_ C 0) u c o O o o 40 to 30 c u ^ E 7 5 - | 20 o o h- (0 c - 10 Legend: • Black squares represent 10-Oct-02 samples • Grey triangles represent 09-Jul-02 samples. — Dashed line represents Chilliwack Creek — Solid line represents Interception Ditch Mainstem FH= forested hillslope tributaries (Elkview Creek); UH = urban hillslope tributaries Notes: Arrows indicate the approximate location of tributaries draining into the mainstem. F13 A15 A16 A18 M19 M20 Interception Ditch Mainstem to Mouth of Chilliwack Creek 1 km (by distance downstream) Figure 7.33 Spatial Trends in Total Phosphorus Concentration in Streambed Sediments along Interception Ditch and Chilliwack Creek 147 7.3.5 Comparison between Agriculture, Urban and Forest Land Uses: Sediment Parameters For each element, the Oct-2002 and July-2003 data sets were pooled and a Mann-Whitney U test was used for pair-wise comparisons to determine if differences existed between urban, forest and agriculture land use categories. An overview of the results is summarized in Table 7.8. Table 7.8 Overview of Mann-Whitney Comparisons between Land Use Categories for Trace Elements in Sediment Parameters showing a significant difference at a=0.05: Combined 2002/2003 July 2003 Agriculture > Forest Na*, P*, % N N a , P Agriculture < Forest - -Agriculture > Urban Cd*, Fe*, P*, Zn*, Cu*, Mg*, Na*, % N Cd*, Fe*, P*, Zn*, C o , C u , M g , Na , % N Agriculture < Urban M n M n Urban > Forest Cd* C d , Fe Urban < Forest K -* indicates significant difference at a=0.0T76 level (Bonferroni adjustment) Combined 2002/2003: n (forest) = 6 ; n (agriculture) = 10 ; n (urban) = 12 July 2003: n (forest) = 3 ; n (agriculture) = 6; n (urban) = 6 Zinc (Zn), iron (Fe), copper (Cu), cadmium (Cd), magnesium (Mg), sodium (Na) and phosphorus (P) all showed significantly higher levels in agriculture compared to the urban catchments. Interestingly, Na, P and % N were the only parameters that had lower levels at the forest sites compared to the agricultural sites. Also worthy of noting is the fact that with there were minimal significant differences between urban and forest land use categories. Higher concentrations of Cd and Fe and lower K concentrations were found at the forested sites compared to the urban sites. Nickel (Ni), calcium (Ca), cobalt (Co), chromium (Cr), and aluminum (Al) showed no significant differences between land use categories. Boxplots by land use category are shown in Appendix D. 148 7.4 Comparison of Water and Sediment Quality to Provincial and Federal Guidelines During Baseflow and Stormflow Conditions Comparing the streamwater and sediment results to the various water quality and sediment criteria provides some indication of the health of the aquatic system. Overall, the water quality throughout the watershed appears to be moderately good. According to water quality and sediment guidelines Semiault Creek and Interception Ditch are the most degraded watercourses, while Luckakuck Creek and the hillslope tributaries have the best conditions for aquatic life. However, it is important to note that provincial and federal guidelines generally do not take into account the potential for cumulative impacts of contaminants (Addah, 2002); and sediment criteria do not account for the confounding effects of the physiochemical attributes of the sediment (such as particle size, organic matter content, chemical species and complexes) or for metal bioavailability which may change the potential for toxic effects at a specific site (McCallum, 1995; CCME, 2001). As a result, concentrations below acceptable levels may still be having an impact on the watercourse 7.4.1 Water Quality Compared to Provincial Water Quality Guidelines Water quality results are compared with BC Water Quality Guidelines for the Protection of Aquatic Life in Table 7.9. While wet and dry season means for nitrate, ammonia, pH and iron are below the provincial guidelines at all sampling stations within the watershed, about half the stations do not meet the guidelines for dissolved oxygen or temperature. Furthermore, a large number of stations have at least one value throughout the sampling season that does not meet guideline values. This suggests that critical levels may be occurring during different times of the year, at certain locations. A l was consistently below detection limit, except within Armstrong Ditch in May 2002. On this date, a value of 0.41 mg/L was recorded, which is above the B.C. water quality guideline (max 0.1 mg/L dissolved A l at pH>6.5). 149 Table 7.9 Sampling Stations within the Chilliwack Creek Watershed Exceeding B.C. Water Quality Guidelines for the Protection of Aquatic Life during the Wet and Dry Seasons (MWLAP, 1998) Water Quality Parameter BC Water Quality Guidelines1 Stations Exceeding Water Quality Guidelines Nitrate - N ^ 40 mg/L (avg) 200 mg/L (max) None 10 mg/L (max)4 None Orthophosphate No criteria Ammonia (Total) 1.07-27.0 mg/L (max)3 (depends on pH and temperature) Dry season: None Wet season: M10, A16, A17, A18 Dissolved 5.0 mg/L (inst. min) (adult/juvenile life stages) Dry season: U5, F14 Wet season: None Oxygen 9.0 mg/L (inst. min) (buried embryo/alevin life stages) Dry season: A l l stations below criteria except A17 Wet season: G l , A2, U3, U4, U5, A18, M20 pH 6.5-9.0 Dry and Wet season: G l Temperature 12°C (incubation maximum for fall and spring) Dry season: A l l stations Wet season: A l l stations above criteria except U8 19°C (max daily temperature) Dry season: A18, A16 Wet season: None Sp. Conductivity No criteria Calcium (Total)2 4 mg/L, high sensitivity to acid inputs 4-8 mg/L, moderate sensitivity > 8 mg/L, low sensitivity Low sensitivity for all stations. Manganese (Total) 0.7-1.9 mg/L (depends on CaC0 3 ) None Iron (Total) 2 0.3 mg/L (max) Dry season: A2, M9, M10, A l 1, A15, A l , A17, A18, M20 Wet season: A2, U5, M9, M10, A l l , A15, A l , A17, A18, M20 Sodium (Total) No criteria Potassium (Total) No criteria Magnesium (Total) No criteria Refers to Guideline for the Protection of Aquatic Life unless otherwise noted 2 Working Guideline 3 Values determined using rages of pH and temperatures observed within the Chilliwack Creek watershed. 4 Drinking water guideline 150 Ammonia concentrations exceeded the maximum permissible total concentration on one sampling date (12-Dec-2002) for a number of sites: M10 (Bailey Ditch), A16 and 18 (Interception Ditch), and A17. Exceedence on this date was likely due to runoff caused by a large storm event occurring just prior to sampling. While the sampling scheme did not capture it, it is possible that ammonia concentrations may have exceeded guidelines during other intense runoff events which promote loss of nitrogen to streams, particularly in agricultural areas where nitrogen is applied to fields and is more available to be lost to steams. Lower concentrations of ammonia may also be toxic depending on how long they are maintained. In July 2003, the ammonia concentration at station A18 was above the average 30-day guideline of ammonia-N (>0.15 mg/L at pH = 8.8 and temperature = 22°C). Because samples were only taken once a month, it cannot be determined whether concentrations were maintained around this level for a 30-day period. However, temperatures at this site (and other sites along Interception Ditch and Chilliwack Creek) increase substantially in the summer, and pH values are usually above 7.5. It is, therefore, feasible that the water would remain toxic to aquatic life during the summer months. A l l stations were out of compliance with the provincial water quality guidelines for temperature at least once over the sampling period. Overall, almost all watercourses in the Chilliwack Creek watershed are limited in their ability to support the incubation of salmonid embryos. Luckakuck Creek is the only watercourse that is able to consistently maintain temperatures below the 12°C guideline for incubation. This is likely due to influence of the cooler groundwater inputs in the stream during the summer season. During the wet season, Armstrong Ditch, Parson's Brook, station A15, as well as all the urban hillslope tributaries were also able to maintain temperatures below the 12°C guideline. Overall, Interception Ditch appears to be an area of relatively high temperatures, in particular, stations A18 and A16, where temperatures above the daily maximum (19°C) were recorded during the summer. As a result, based on the high temperatures recorded along Interception Ditch it is unlikely that it would be able to support a viable fish population. With respect to dissolved oxygen, station A17 is the only site for which DO levels did not fall below the 9.0 mg/L minimum requirement of DO for buried embryo development at any time during the sampling period. Wet season data are restricted to one sampling date (03-March-2003), and therefore it is difficult to make conclusions as to what is happening over the wet period. However, based on the available data there is the potential that a number of streams may be able to support embryonic fish development at this time. The downstream site along Chilliwack Creek (M20) is the most impacted by an oxygen deficit. Two sites did not meet the minimum DO level required for adult and juvenile fish: Elkview Creek (F14) and Teskey Creek (U5). Overall, DO levels appear to be more critical during the dry season then the wet season. 151 The spring-fed station is the only sampling station that did not meet provincial for pH. pH at this station dropped to 6.3 on two occasions. It is assumed that the lower value at this station is likely the result of the influence of geology on the water chemistry, rather than the result of land use activities. In general, agriculturally influenced streams - Semiault Creek (A2), Armstrong Ditch (A l l ) , Interception Ditch (A 15, A16, A18), Chilliwack Creek (M20), Bailey Ditch (M10) and station A17 had iron concentrations above the BC Water Quality Guideline (0.3 mg/kg) on at least two of the three sampling dates. Values above the criteria were also detected at stations M9 and U5. COMPLIANCE DURING STORMFLOW CONDITIONS: At least one storm sample for all stations (except M20) exceeded the guidelines for dissolved A l (max 0.1 mg/L at pH>6.5) and total Fe (0.3 mg/L). Zinc was only above detection at station A2. Concentrations of the latest sample during the storm event (0.05 mg/L) exceeded water quality guidelines for 'total' Zn (0.04 mg/L at hardness of 100 mg/L CaC0 3 ) . Nitrate concentrations did not exceed any guidelines for water quality; however concentrations were near the 10 mg/L guideline for drinking water quality at station A2, and it is expected that concentrations would continue to increase over the course of the storm and exceed this limit. 7.4.2 Sediment Quality Compared to Federal Sediment Quality Guidelines For sediments, the comparison with Canadian Sediment Quality Guidelines for the Protection of Aquatic Life reveals a number of stream stations exceeding the different criteria for levels of cadmium (Cd), chromium (Cr), copper (Cu), iron (Fe), nickel (Ni), lead (Pb) and zinc (Zn) (Table 7.10). N i , Cu, Fe and Cd concentrations were above the lowest effect level (LEL) for all stations, including the forested control stations. This suggests that the natural background levels of these metals may be high, rather than an influence of the surrounding land use on sediment toxicity. Similar to high streamwater iron concentrations observed at the agriculturally influenced station, sediment iron concentrations were above the severe effect level (SEL) for stations A2 (Semiault Creek), A l 1 (Armstrong Ditch), U5 (lower Teskey Creek), M9 (Teskey Way Ditch), A16 and A18 (Interception Ditch), F12 (Parson's Brook) and M20 (Chilliwack Creek). Cadmium concentrations above the probable effect level (PEL) and zinc concentrations above the ISQG were found at most of these same sites. Cd and Zn concentrations above these levels were also found at M10 and A17. It is thought that agriculture may be leading to Fe and Cd toxicity in these sediments. Chilliwack Creek maintained some of the highest metal concentrations overall, particularly station M l 9 , which lies directly below an outfall draining a relatively dense urban area. While it is not possible to determine if the metal concentrations were impacting the aquatic system 152 without further testing, this site is the most likely to be impacted as it has some of the highest concentrations of Pb, Cd, Zn, Mn, and Cu recorded in the watershed. This site is also the only site for which chromium exceeded the sediment quality guideline. The aquatic environment of Semiault Creek is also likely to be impacted, with Cd, Fe, Zn, Cu above sediment guidelines as well as high concentrations of Co. Table 7.10 Sampling Stations within the Chilliwack Creek Watershed Exceeding Canadian Sediment Quality Guidelines for the Protection of Aquatic Life (CCME, 2003) Canadian Sediment Sediment Quality Parameter Quality Guidelines (mg/kg)u Stations Above Canadian Sediment Quality Guidelines Cadmium 0.6 (TEL) All stations above guideline 3.5 (PEL) M9, F12, A15, A16, A17, A18, A2, M10, M19, M20 Chromium 37 (ISQG) M19 90 (PEL) None Copper 35.7 (ISQG) All stations above guideline except U3, U4, U6, U7, U8, F13 197 (PEL) None Iron 2100 (LEL) All stations above guideline 4380 (SEL) A2, U5, M9, A l 1, F12, A16, A18, M20 Nickel 16 (LEL) All stations above guideline 75 (SEL) None Lead 31 (LEL) G1,M19, M20 250(SEL) None Zinc 123 (ISQG) A2, M9, A l 1, F12, F14, A15, A16, A17, A18, M19, M20 315 (PEL) None No guidelines are;aVailable for Al,Ba,.Ca, Co, K, Mg, Na, P, Si, Si 153 8 LAND USE AND W A T E R INTERACTIONS The various land use types and their changing pattern affect both the hydrologic regime and water quality of the adjacent watercourses. This chapter discusses the relationship between the different land uses and both the water and sediment quality. To examine these relationships, the watershed was divided into sub-watersheds (contributing areas) by delineating the area draining each sampling station, as described in chapter 4. A map of these contributing areas in the watershed is shown in Figure 4.3. Next, the proportion of land with a given land use (referred to hereafter as 'land indices') within each individual (independent) contributing area was calculated (as a percent of the total contributing area) using GIS. In addition, because water quality at a given station may be influenced by land uses farther upstream than the immediate contributing area, land use indices were also calculated for the cumulative contributing area. This cumulative contributing area comprises the total watershed area upstream of a given sampling stations (that is, all the contributing areas upstream of the water sampling station). Furthermore, because areas closer to the stream may have a greater influence on water quality, land indices were also calculated for buffer zones of three different widths (50 m, 100 m and 200 m) parallel to the stream channel. By using buffers as opposed to contributing areas, only the land use directly adjacent to the stream is related to water quality, minimizing the incorporation of irrelevant or less influential land uses. Other studies that have attempted to relate land use indices to water quality parameters have found that, in agricultural areas, the buffer zone technique gave better results than the use of contributing areas (Addah, 2002). A Spearman's Rank correlation test performed on land use indices and water (and sediment) quality parameters revealed significant correlation between several parameters. These results are presented below for urban, natural and agricultural land uses separately. A full summary of the results is shown in Appendix F. Note that a positive correlation coefficient indicates that as the percent of land use increases, the concentration of the water (or sediment) parameter also increases. Conversely, a negative correlation indicates that as the percent of land use increases, the concentration of the parameter decreases. Overall, there were minimal differences between relationships found using the independent versus the cumulative land indices. There were also minimal differences between the three different sized contributing areas. Results from correlations using the independent contributing areas and 100 m buffers correlations are discussed here. A summary of selected land use indices for each (independent) contributing area and 100 m buffer is shown in Table 8.1, and a map of the 100 m land use buffers is provided in Figure 8.1. 154 Table 8.1 Characteristics of Land Use within Each Contributing Area % L a n d Use within Contributing Areas Watercourse Station Total agric. Arable Cattle Open Space Forest Ind./ C o m . Res. High density urban Low density urban A15 31.4 10.3 4.3 19.8 48.7 0.0 0.0 0.0 0.0 Interception Ditch A16 79.1 22.7 39.4 8.4 12.5 0.0 0.0 0.0 0.0 A18 54.3 51.0 0.4 22.6 9.9 1.7 11.5 5.4 7.7 Teskey Way Ditch A17 61.8 40.3 20.8 22.6 14.1 0.0 1.5 0.0 1.5 Semiault Creek A 2 69.9 26.7 42.1 14.8 15.1 0.2 0.0 0.0 0.0 Armstrong Ditch A l l 14.8 10.8 2.4 17.8 67.4 0.0 0.0 0.0 0.0 Elkview Creek F12 19.0 10.89 0.00 22.64 56.31 0.09 1.39 0.00 1.39 F14 8.88 0.00 0.06 10.24 80.04 0.84 0.00 0.00 0.00 Parsons Brook F13 14.1 6.54 1.84 17.70 68.25 0.00 0.00 0.00 0.00 Teskey Creek U3 5.2 2.7 1.6 21.6 23.4 1.1 48.7 0.0 48.7 U5 2.9 1.2 0.0 33.9 33.8 3.7 25.4 9.5 23.2 Lefferson Creek U4 3.3 0.0 0.0 80.9 7.5 0.0 8.3 0.0 8.3 U7 8.0 7.5 0.5 30.0 57.8 0.0 4.2 2.7 1.5 Benchley Creek U6 14.6 14.6 0.0 16.1 65.2 0.0 4.1 0.0 4.1 Walker Creek U8 0.0 0.0 0.0 21.6 78.4 0.0 0.0 0.0 0.0 Teskey Way Ditch M 9 18.3 17.2 0.0 44.1 37.6 0.0 0.0 0.0 0.0 Bailey Ditch M10 28.2 13.9 0.0 25.5 34.5 7.1 4.8 1.4 10.4 Chilliwack Creek M19 27.9 27.1 0.8 24.1 7.3 8.6 31.5 1.2 39.1 M20 72.8 32.3 38.7 4.9 0.7 6.4 14.3 7.6 13.6 155 o Sampling stations Land Use Categories • Arable • Cattle I I Poultry H I Horticulture • Hobby Farms I I Unused Agricultural Land • Residential - Low density • Residential - High Density • Rural Residential I | Industrial/Commercial • Forest I I Open Space 1 0.5 0 2 km Figure 8.1 Chilliwack Creek Watershed Land Use 100 m Land Use Buffers 8.1 Correlations with Agricultural Land Uses Agricultural land use was subdivided into different activity types: arable, cattle, poultry, horticulture (tree farms), greenhouse, unused (fallow), and hobby farms for the land use analysis (Table 5.1). The 'total agriculture' category represents the sum of all these agricultural land uses. A summary of results for arable, cattle and total agricultural operation are presented in Table 8.2, and the complete results are provided in Appendix F. Only a few relationships were found between water and sediment results for other agricultural activities (Appendix F). The results showed that nutrients (nitrate, ammonia, orthophosphate), conductivity, temperature, dissolved Fe and Mn, were positively correlated with total agricultural land and percent arable land. The relationships were generally consistent between the wet and dry season, with the exception of nitrate (for which no relationship was seen in the dry season). pH was the only water quality parameter that had a negative relationship, and this was only significant in the dry season. The percent land dedicated to cattle (%cattle) generally showed these same relationships in the wet season (with the exception of orthophosphate where no relationship was seen), but in the dry season only the relationship with pH and with ammonia were significant. It is interesting to note that for wet season nitrate, the relationship was stronger for %cattle and total agriculture than for %arable. Overall, the relationships were slightly stronger in the wet season and there was relatively little difference between the relationships found using contributing areas versus 100 m buffer zones. A number of metals in sediments (Cd, Co, Cr, Cu, Fe, N i , Zn, and P) were consistently correlated with %total agricultural land and %arable land. Of these Cd, Cu, Fe, Zn and P concentrations had the strongest relationships (r > 0.6). Weaker relationships with K, Ca, Mg, and Na were also found for both %total agricultural and %arable land, with a few exceptions. Percent cattle showed fewer significant relationships with metals in sediment, and no relationship with K, Mg or Ca. It is interesting to note, that in contrast to total P, the bio-available P in sediment (orthophosphate) was negatively correlated to total agricultural land. As with the water quality correlations, there was minimal difference between the results using contributing areas and results using 100 m buffer zones. 157 Table 8.2 Spearman's Rank Correlation Coefficients for Independent Contributing Areas and 100 m Buffers: Agricultural Land Use Indices versus Water and Sediment Quality Parameters A G R I C U L T U R E muepenaeni Area Correlations Total Arable Cattle CA : Buffer CA \ Buffer CA : Buffer Ammonia-N wet 0.69 0.71 0.56 0.63 0.47 0.45 dry 0.50 0.68 0.50 0.61 0.39 0.35 Nitrate-N wet 0.55 0.50 0.33 0.39 0.54 0.60 dry -0.25 -0.15 -0.17 -0.21 -0.26 -0.02 Orthophosphate-P wet 0.54 0.54 0.61 0.59 0.19 0.16 Oi H dry Hi •< Specific wet 0.74 0.61 0.67 0.61 0.40 0.45 < Conductivity dry 0.46 0.38 0.46 0.33 0.24 0.38 DO wet -0.18 -0.31 -0.10 -0.21 -0.25 -0.28 < dry -0.20 -0.23 -0.18 -0.16 -0.18 -0.13 RQl pH wet -0.04 -0.23 -0.22 -0.29 -0.03 0.04 ATE dry -0.69 -0.73 -0.66 -0.74 -0.51 -0.48 Temperature wet 0.83 0.77 0.87 0.80 0.42 0.34 dry 0.61 0.57 0.46 0.52 0.25 0.20 Iron (Fe) 0.64 0.76 0.55 0.72 0.46 0.43 Manganese (Mn) 0.73 0.81 0.60 0.71 0.56 0.53 Calcium (Ca) 0.35 0.23 0.50 0.33 0.19 0.22 Cadmium (Cd) 0.84 0.65 0.71 0.70 0.51 0.54 Cobalt (Co) 0.62 0.42 0.46 0.42 0.43 0.55 Chromium (Cr) 0.47 0 . 3 3 0.51 0.42 0 . 1 6 METERS Copper (Cu) 0.70 0.50 0.67 0.57 0 . 3 2 0.43 METERS Iron (Fe) 0.83 0.67 0.65 0.68 0.48 0.51 METERS Nickel (Ni) 0.49 0.31 0.49 0 . 3 2 0.11 0,l< ) Zinc (Zn) 0.68 0.53 0.61 0.60 0.49 0.62 PH H Phosphorus(P) 0.81 0.61 0.71 0.68 0.59 0.58 EDIMEN Calcium (Ca) 0.40 0 . 2 3 0.53 0 . 3 3 n 93 0 .''>.'* EDIMEN Potassium (K) 0 .3 / 0 3 3 0.57 0.42 - 0 . 1 0 -0.01 Magnesium (Me) 0.61 0.53 0.59 0.55 0.12 0,20 Sodium (Na) 0.63 0.62 0.63 0.71 0.64 0.70 % Nitrogen 0.42 0.50 0 . 2 3 0 31 0.49 0.46 Orthophosphate -0.48 - 0 . 3 0 -0.43 •"Li, \ -0,05 * Values in bold indicate significant correlations at a =0.05 for a one-tailed test 158 8.2 Correlations with Natural Land Cover As shown in Table 5.1, the 'natural' land use category is subdivided into forest and open space. Open space encompasses all non-agricultural clearings (e.g. parks, playing fields, clearings) and was designed to represent land that was not forested but was also not being used for agriculture. This was done because it was assumed that this land base would have different inputs than agricultural fields or forests. Rural residential land (which includes both the building and surrounding land) was also included in this category since most of the land is cleared space. A summary of the results for 'total natural', forest and open space categories is presented in Table 8.3. Complete correlations results are located in Appendix F. Correlations between both the percent total natural area (%natural) and percent forested land (%forest) with water and sediment parameters were generally similar using contributing areas or 100 m buffers. Temperature, conductivity and ammonia were negatively correlated with %forest and %natural land use in both the wet and dry season. Negative relationships were also seen with other nutrients (nitrate and orthophosphate) and dissolved oxygen (DO) in the wet season, and with pH in the dry season. A l l major ions (Ca, Mg, Na, K) and trace metals (Fe, Mn) were negatively correlated with these two land indices. %open space only showed significant correlations with conductivity (wet and dry), pH (dry), and temperature (wet), but these were only seen using the 100 m buffer method. Negative relationships were also found between metals (Cd, Co, Cu, Fe, N i , Cr, and P) in sediments and %natural land use. Similar to relationships with agriculture, the strongest correlations were found with Cd, Cu, Fe, Zn and P (r > 0.6). Negative correlations with Ca, K, Mg and Na were also found. Similar correlations were found for %forest with the exception of Co, Cr, N i , Ca, and Mg; however, relationships with Co and Ca are seen when the cumulative index is used (Appendix F). It is interesting to note that for %natural and %forest a few more significant correlations were found using the contributing area method, while significant correlations with %open space were only found using the 100 m buffer. 159 Table 8.3 Spearman's Rank Correlation Coefficients for Independent Contributing Areas and 100 m Buffers: Natural Land Cover Indices versus Water and Sediment Quality Parameters Cumulative Area NATURAL H < > < c Pi u H Total Forest Open Space CA Buffer CA Buffer CA Buffer Ammonia-N wet -0.46 -0.44 -0.29 -0.28 -0.24 -0.02 dry -0.52 -0.66 -0.42 -0.59 -0.22 -0.01 Nitrate-N wet -0.48 -0.52 -0.41 -0.35 -0.41 -0.39 dry -0.22 -0.22 -0.37 -0.18 0.34 -0.03 Orthophosphate-P wet -0.45 -0.40 -0.45 -0.41 -0.10 -0.14 dry Specific wet -0.62 -0.51 -0.64 -0.44 -0.08 -0.15 Conductivity dry -0.81 -0.67 -0.86 -0.58 0.10 -0.26 DO wet 0.53 0.61 0.63 0.66 -0.11 -0.02 dry 0.39 0.50 0.20 0.34 0.06 0.08 pH wet 0,23 0.52 0.31 0.60 -0.12 -0.09 dry 0.61 0.74 0.51 0.66 0.22 0.30 Temperature wet -0.58 -0.57 -0.46 -0.45 -0,29 -0.30 dry -0.37 -0.28 -0.32 -0.28 0,03 0,12 Iron (Fe) -0.36 -0.47 -0.25 -0.36 -0.11 0.03 Manganese (Mn) -0.53 -0.61 -0.44 -0.52 -0.22 -0.09 Calcium (Ca) -0.55 -0.51 -0.74 -0.58 0.15 -0.14 Potassium (K) -0.57 -0.49 -0.74 -0.47 0,28 -0.18 Magnesium (Mg) -0.43 -0.36 -0.62 -0.44 0,33 0.05 Sodium (Na) -0.32 -0.23 -0.52 -0.28 0.48 0.28 Cadmium (Cd) -0.70 -0.62 -0.60 -0.51 -0 ,32 -0.45 Cobalt (Co) -0.68 -0.52 -0.57 - 0 . 3 7 -0.46 Chromium (Cr) -0.46 -0.40 -0 .30 -0.56 Copper (Cu) -0.69 -0.52 -0.52 ••'U. k.d -0.53 Iron (Fe) -0.62 -0.56 -0.51 -0.46 -0.43 Manganese (Mn) -0,10 0.01 - 0 . 0 3 _Q 31 Nickel (Ni) -0.59 -0.40 •0.32 - 0 . 1 0 -0.40 -0.73 Zinc (Zn) -0.72 -0.61 -0.64 -0.47 -0.49 Phosphorus (P) -0.59 -0.54 -0.57 -0.51 Calcium (Ca) -0.57 od -0.50 -0.52 Potassium (K) -0.37 n %c\ -0.37 Magnesium (Mg) -0.44 -0 .22 -u. i i Sodium (Na) -0.64 -0.70 -0.53 -0.54 % Nitrogen -0,31 -0.39 0 t2T> - 0 . 3 3 -0,2.1 * Values in bold indicate significant correlations at a =0.05 for a one-tailed test 160 8.3 Correlations with Urban Land Uses Urban land use was divided in two ways. First, urban categories were grouped based on the type of urban use (industrial/commercial versus residential). Second, total urban land use was grouped based on the density of use (high density versus low density). A summary of the correlation results is shown in Table 8.4, and the complete results of the correlations test are provided in Appendix F. Urban land use activities are concentrated in two regions of the watershed: the urban centers of Chilliwack and Sardis (part of which are within the contributing areas for stations M l 9 and M20), and the urban hillslope development of Promontory. In both areas, the percentage of low intensity urban is much larger than high intensity urban within the individual contributing areas, and the percentage of residential is much larger than industrial land uses (Table 8.1). The correlations presented did not include stations M19 and M20 (which are more intensively urbanized). This was done in order to determine if any relationships could be found with the residential hillslope development. The percent of both low density (%low density) land use, and percent residential (%residential) showed similar results to each other and to the results found for total urban land use (%urban): positive correlations with nitrate in the dry season, negative correlations with ammonia in the wet season, and positive correlations with dissolved K and Mg. In sediment, Cd, Fe and Ca showed negative relationships with %residential land use. A number of (weak) correlations were found using the 100 m buffer that were not found using the contributing area methods. These were: positive correlations with pH in the dry season, and negative correlations with wet season ammonia, dissolved Fe and Mn, and sediment-bound Co, Cu and Zn. Correlation results for the percent high density (%high density) and commercial/industrial (%C/I) are slightly different, and there were significantly more correlations between %C/I and water quality parameters. Relationships that were not found with %low density or %residential include negative correlations with dissolved sodium and dry season DO, and positive correlations with dry season ammonia and both orthophosphate and conductivity in the wet season With the exception of a weak positive relationship with copper, no associations with sediment-bound metals were found. 161 Table 8.4 Spearman's Rank Correlation Coefficients for Independent Contributing Areas and 100 m Buffers: Urban Land Use Indices versus Water and Sediment Quality Parameters Urban Independent Area Correlations Total Commercial/Ind. Residential High Density Low Density CA Buffer CA Buffer CA Buffer CA Buffer CA Buffer wet -0.38 0.39 0.20 -0.44 0.33 0.20 -0.07 -0.44 Ammonia-N dry -0,11 0.50 0.46 0.14 0.42 0.17 0.18 wet -0,15 ij. ci/. 0 .20 •0.28 -0,24 -0.18 -0.11 -0,22 -0.24 iNitrate-lN dry 0.55 0.37 0,26 ' 0.31 0.47 0.45 •0.05 0 06 0.57 0.47 wet -0.20 0.42 0.24 0.10 -0,28- 0.39 0,10 0,13 -0.28 [ETERS Orthophosphate-P dry [ETERS wet -0.23 0.48 0 25 , ' • -0,32 0,19 0 04 -0 72 PARAM Specific Conductivity dry 0.47 0.19 0.69 0.56 f) fi--: 0.26 0.39 0.08 PARAM r A / \ wet -0.48 -0.32 -0.44 -0.59 ...() y:') -0.41 -0.39 -0.19 > 1J(J dry -0,09 O.OS -u .3 / -0.24 -0 61 -0.56 -0,20 „ I j wet -0.01 0,06 -0.28 L -0.32 0.1 1 -0.46 -O.OS 0.12 < pH dry 0.22 0.42 -0,29 -0.15 0.51 : 20 0.15 0.51 o • wet -0,18 -0.36 0.29 -0,06 .', ! 0 0 .03 U H Temperature dry -0.05 •0.36 0.32 0,02 -0.37 •0.02 •< - Iron (Fe) -0,16 -0.43 0.24 0 ,14 -0.46 ". -0.04 -0.46 Manganese (Mn) -0.12 0.38 0 ,30 -0.44 -0.08 -0.45 Calcium (Ca) 0.35 ' 0,36 0.38 CO 2 3 -0,02 0,25 Potassium (K) 0.79 0.53 0.51 0.57 0.69 0.47 0.57 0.54 0.70 0.46 Magnesium (Mg) 0.68 0.58 0 36 0.64 0.58 0.45 0.57 0.64 0.57 0.43 Sodium (Na) 0.49 0.30 0.56 0.48 0.30 0.11 3,2 0.34 0 2 VI as Cadmium (Cd) -0,35 O i l -0.39 -0.55 -0,19 -0.29 -0.3? u Cobalt (Co) •0.23 0.30 0,07 0 31 -0.39 -0.34 -0,29 a -S Copper (Cu) -0.33 0.38 -0.04 ••G .20 -0.42 •0.16 -0.41 a> Iron (Fe) -0.35 0.13 -0.41 -0.55 -0.37 < -0. Nickel (Ni) 0.19 -0.17 0 . 2 " -0.51 0.14 -0.47 H Zinc (Zn) -0.03 0.03 -0.09 -0,14 -0.05 u Phosphorus (P) ••0.35 0,26 0.01 0,2? -0.48 -0.42 -0.50 -0.2.4 5 Calcium (Ca) -0.37 0.18 0 ,09 -0.40 -0.55 -0.07 -0.39 </) Potassium (K) -0.06 0.04 -0.13 n '? -0 11 -0.41 -0.48 -0.02 -0.46 * Values in bold indicate significant correlations a=0.05 for a one-tailed test 8.4 Comparison Between Land Use Components A subset of the results was chosen for comparison, and is shown in Table 8.5. The land indices were chosen because they were most representative of the comparisons throughout this thesis. Overall, the correlation results support the findings in the previous chapter. Agricultural land uses were associated with increases in both nutrients and trace metals (in water and sediment). The impacts of residential land use on water and sediment quality were less pronounced. The only parameters that showed relationships unique to /Presidential land use were dissolved magnesium and potassium. Forest cover was generally associated with lower concentrations of nutrients and metals. Table 8.5 Overview of Spearman's Ranks Correlations (p<0.05) Between Land Use and Various Water and Sediment Quality Parameters Agriculture Urban Forest (%total agriculture) (%residential) (%forest) Nutrients in Wet season N H 4 , N 0 3 , P 0 4 (- P0 4 ) Water Dry season N H 4 N 0 3 ( -NH 4 , N 0 3 ) Physical Water Wet season cond., temp. DO, (- cond., temp.) Parameters Dry season cond., temp., (-pH) pH (-cond, temp) Ions and Water (dissolved) Fe, M n K , M g (- Fe, Mn, Ca, K , Mg) Metals Sediment (total) Cd, Co, Cr, Cu, Fe, Zn, P, Ni , Ca, Mg , Na (-Cd, Fe, P) (- Cd, Cu, Fe, Zn, P, K, Na) 8.5 Correlations with Impervious Surface Area Percent total impervious surface area (%TIA) did not correlate well with water or sediment quality parameters. In the wet season, specific conductivity and dissolved Mn, Ca, K and Mg were all positively correlated with %TIA. A negative relationship with dissolved oxygen was the only relationship found in the dry season. 163 9 DISCUSSION The objectives of this project were to provide information on the current status of the watershed in terms of hydrology as well as water and sediment quality, to determine the impact that the new hillslope development is having on the hydrology and water quality of the hillslope and lowland agricultural area, and to investigate the links between both surface water and sediment quality and land use. Changes in land use were identified, water and sediment samples were collected and analyzed, precipitation and flow data were examined, and comparisons were made between the three primary land uses (agriculture, urban, and forest). 9.1 Climate and Hydrology 9.1.1 Climate Variability Climate variability is an important parameter affecting the hydrologic regime of a watershed; and precipitation is the key climate variable of concern to stormwater management (Watt et al., 2003). Precipitation during the sampling period was characterized by extremes - a long dry period followed by a wet season with record rainfalls. October 2002 had significantly less precipitation (24.0 mm) than the 30-year average (167.3 mm). In contrast, October 2003 was extremely wet with a total monthly precipitation of 364.7 mm. Of this, 254.6 mm fell over a six day period (and 100.3 mm fell during the storm event on 16-Oct-03). This is consistent with the increasing climatic variability noted in the literature. Precipitation data for southwestern British Columbia suggest that precipitation frequency, intensity and duration are changing compared to the mid 20 t h century, and climate change has been implicated as the primary contributor to these observed trends (Stephens et al., 2002a). Environment Canada models project increasing fall and winter precipitation, decreasing late spring-early summer precipitation, and more intense rainstorms (Stephens et al., 2002a). Other studies have also reported that the frequency of heavy and extreme precipitation events is increasing (Houghton et al., 2001). In terms of stormwater management, the increased seasonal rainfall and more frequent heavy rainfall events mean that there will be more runoff to manage in the future. Typically, stormwater management infrastructures are designed to convey a particular historical rainfall pattern; however assumptions about climate are generally static and based on limited historical data (Watt et al., 2003). If rainfall inputs increase as a result of climate change, the designs that worked in the past may not be adequate in the future. For example, storage volumes for detention ponds designed to reduce peak outflows to predevelopment conditions are based on a specific design rainfall event. As precipitation increases the storage volume required would be larger for the same design rainfall. 164 9.1.2 Distribution of Storm Events An examination of the precipitation records for Chilliwack show that the majority (88.6%) of the storm events recorded were minor events, and that total daily rainfall was usually below 30 mm. Total daily rainfall exceeded 60 mm on 1 of 68 days (over the three year period that was measured) and, on average, accounted for about 13% of the total annual rainfall volume. In contrast, 70% of the total annual rainfall volume over this same period was generated by events of less than 30 mm. The implication for stormwater management, based on these data, would be that strategies which incorporate low impact designs and source control methods (which are designed to capture and infiltrate precipitation from these small events on-site) may be more effective at mitigating the impacts of development than conventional stormwater management systems (which are generally designed to control peak runoff rate for a few large storm events; and are often not designed to mitigate for runoff rate and increases in runoff volume from the smaller, frequently occurring storms). If Chilliwack's new stormwater management plan is effective, over 75% of the rainfall will be captured and detained on site, greatly reducing the rate and volume of runoff reaching nearby streams. 9.1.3 Overview of Storm Response for the Different Sub-Watersheds The Chilliwack Creek watershed has a range of land use properties, which results in a heterogeneous mix of hydrologic response properties for the various sub-catchments. This section will discuss the general response of the individual sub-catchments based on the hydrographs and storm characteristic results presented in Chapter 6. 9.1.3.1 Forested Systems: Parsons Brook and Elkview Creek The drainage area for both Elkview Creek and Parsons Brook is mostly rural and undeveloped (forested) land. Therefore, it is reasonable that the hydrographs exhibit a longer time to peak discharge followed by a slow recession as is generally seen in the hydrographs for the Parsons. While Elkview Creek showed significantly longer lag times than the other catchments, a number of the hydrographs for individual storm events exhibited a much shorter time peak than Parsons. A possible explanation for the shorter lag time is that a road runs directly adjacent to Elkview Creek. Direct runoff from the impervious road surface would decrease the travel time to the stream, while precipitation falling in the forested parts of the catchment would infiltrate and flow much slower as subsurface stormflow, and continue to reach the stream much later causing the slower recession. In summary, the response at Parsons Brook is more typical of a forested catchment while the response at Elkview Creek may be influenced in part by a small section of impervious surface directly adjacent to the stream. 165 9.1.3.2 Urban Systems: Teskey Creek and Lefferson Creek Storm flow peaks for both of the semi-urbanized catchments closely mimic precipitation patterns, and display a shorter time to peak than the other stations. In addition, the peak runoff rate at Teskey Creek (which has a higher proportion of impervious area) is markedly higher than most of the other stations. This suggests that much of the rainfall is reaching the stream as direct surface runoff from impervious surface areas in the Promontory development. In contrast, peak runoff rate at Lefferson is generally lower than most stations. Although this catchment is currently being developed, the area of impervious surface is lower than it is for Teskey Creek. In addition, the gauge itself is on a buffered area of the stream with steep slopes on either side. Consequently, the lower storm runoff rates at Lefferson may be because the area developed (i.e. the impervious surface area) is still insufficient to increase surface runoff. If this is the case, the shorter lag time is possibly the result of overland flow down the steep slopes at either side of the stream gauge; and not due to urbanization. One exception was observed during the major storm event on 17-Jul-02 storm event. 9.1.3.3 Agricultural Systems: Semiault Creek Peak runoff rates were generally high, with moderate response times. The median peak runoff rate (0.710 mm/hr) of this agricultural catchment is up to 3.7 times the median peak runoff rates in the undeveloped catchments. Agricultural drainage systems (like those found in the Semiault Creek catchment) have been shown to increase outflow from fields by 5 to 20% depending on the system, site conditions and soils (Schreier et al., 2002; Ritter and Shirmohanmmadi, 2001). Agricultural practices have also been shown to reduce the infiltration capacity of the soil which would increase surface runoff. This is the result of a number of factors, such as soil compaction, sealing of the soil surface by sediment laden runoff, and the reduction of organic matter content that maintains a soil texture that is conductive to infiltration (Brady and Weil, 1996). 9.1.3.4 Groundwater Influenced System: Luckakuck Creek Flow at the Luckakuck Creek station can be characterized as having: 1) a slightly higher baseflow ; 2) a relatively short lag time; and 3) a somewhat muted/stable response overall. This stream originates as a spring and is influenced by groundwater from the Sardis-Vedder aquifer, which contributes to baseflow in the stream. The soils above the aquifer are very permeable and consequently less surface runoff is expected in this area. In addition, groundwater accretion resulting from a particular storm is normally released over an extended period of time (Viessman and Lewis, 1996). Consequently, the more stable response observed is reasonable. Part of the watershed surrounding the stream above this station is urbanized; yet it does not seem to be increasing peak flows significantly. The short lag time, however, may be influenced by the stormwater outfall above this station. 166 9.1.3.5 Downstream (Mixed) Systems: Bailey Ditch and Chilliwack Creek The timing and magnitude of the streamflow from the upstream tributaries play a role in determining the hydrologic response at the downstream stations. For example, the Chilliwack Creek hydrograph shows both a rapid and extended response. The station itself lies in an urban section of the stream, and the initial rapid response is likely from immediate stormwater runoff from impervious surfaces near the station. As water from the upper portions of the watershed reach the station the stream continues to respond. The blips in the hydrograph (see Figure B.l) following rainfall are likely due to contributions from different portions of the watershed. Bailey station, which lies downstream of the Promontory development peaks slightly after the urban catchments and has a more gradual response. 9.1.4 Comparison of the Hydrologic Effects between Forested and Urban Sub-Catchments Urbanization, with the accompanying loss of vegetation, replacement of soil with impervious surfaces, and routing of stormwater runoff directly to stream channels, has a significant impact on many of the processes that control streamflow (McCuen, 1998). A number of major effects of urbanization on hydrologic processes have been identified in the literature: 1) a higher proportion of precipitation appears as surface runoff (i.e. increased runoff volumes and higher runoff/rainfall ratios); 2) the catchment response time to precipitation is accelerated and the lag time between precipitation and runoff is decreased; 3) peak flow magnitudes are generally increased; and 4) low flow is typically decreased due to reduced contributions from groundwater storage (Rose and Peters, 2001). The Promontory development on the hillslope is still relatively small, and the impacts are not as large as would be observed in a more intensive urban catchment. Still, some of the effects listed above are evident. 9.1.4.1 Lag Time Catchment response time appears to be related to its land use/land cover properties. The median lag time of the suburbanized catchments, Teskey (64.5 ha) and Lefferson (166.4 ha), are extremely short, and more than 12 hours faster than the lag time of the forested catchments, Parsons (204.8 ha) and Elkview (216.2 ha). In addition, variability for the urban catchments is minimal compared to the variability observed in the forested catchments (Figure 6.11). A possible explanation for the low variability within the urbanized catchments is that impervious surfaces reduce the infiltration capacity to zero so that more runoff occurs as overland flow. When this happens, some of the factors influencing the runoff process (e.g. antecedent soil moisture) have a lesser influence on the timing of storm flow. In contrast, in undeveloped catchments the timing (and magnitude) of streamflow, is more dependent on storm and catchment characteristics. The part of the drainage basin that is contributing rapid storm runoff to the channel tends to expand through an entire storm season, making any changes in stream flow more intense and the lag time shorter for similar-sized storm occurring later in the wet season (Hewlett and Hibbert, 167 1967; Booth, 2000). For example, a summer storm with a longer antecedent dry period would be more likely to have a longer lag time because the soil will be able to absorb more precipitation, and more water will flow to the stream as subsurface flow and less as rapid overland flow. 9.1.4.2 Peak Runoff Rate The box plots shown in Figure 6.13 and the sign test results indicate that peak runoff rates for the analyzed storm events at the Teskey station (0.165-3.731 mm) were up to 15 times higher than at the undeveloped hillslope stations (e.g. Parsons and Elkview). Higher peak flows and higher volumes of water discharged to streams during storm events are common consequences of urban development, and have been noted in literature for many years (Anderson, 1968; Leopold, 1968; Carter, 1961). It is interesting to note, however, that for some storm events there was a minimal difference, and in some instances total runoff volume was higher at Parsons (forested stream) than at Teskey. On closer inspection, these events occurred soon after another storm event while runoff was still increasing. As a result, the response seen in the forested catchment is a combination of the response of the previous storms and the actual event being measured, resulting in a much higher streamflow. Response in the more urban catchments is immediate, and therefore not influenced by the previous event. Conversely, the largest differences between the forest and urban catchments were observed for intermediate sized events (e.g. l-Jun-01) following periods with relatively little rain, when forests have a larger storage capacity for water. During wetter periods, the soils in undeveloped basins become saturated and additional rainfall is converted to runoff as much as it does in an urban basin (Konrad, 2003). 9.2 Water and Sediment Quality 9.2.1 Impacts of Agricultural Land Uses Water and sediment sampling revealed a number of impacts that appear to be related to agricultural activities. These include: elevated nutrient concentrations in lowland watercourses, evidence of eutrophication, depressed oxygen levels in streams and ditches, and the enrichment of streambed sediment with trace metals associated with animal feeds and fertilizers. 9.2.1.1 Nutrients in Water and Sediment Nutrient results indicate that most of the streams in the lowland agricultural area have levels above those of the forested control sites for nitrate in the wet season, and for orthophosphate and ammonia throughout the year. While none of the samples exceeded the B.C. Water Quality Guidelines, Semiault Creek consistently had the highest levels of nitrate and reached a concentration of 5.68 mg/L in December 2002. Chilliwack Creek was the only other waterway where the nitrate concentration rose above the 3 mg/L level that is indicative of impact by anthropogenic activities (Schreier, pers comm., 2004). For 168 ammonia, the B.C. Water Quality Guidelines are based on the risk to aquatic fish, and are set for both continuous exposure (30-day average) to ammonia and for maximum acceptable concentrations (MAC). These guidelines vary with pH and temperature, which affect both the toxicity of ammonia and the form in which it occurs (Mueller and Helsel, 1999; Nordin and Pommen, 1986). Concentrations at the downstream station along Interception Ditch (A 18) were above the 30-day average guideline (>0.15 mg/L at pH = 8.8 and temperature = 22°C) on 9-July-2003. While samples were not taken over a 30 day period, it does not seem implausible that average concentrations may be sustained above this level for the duration of the month. Ammonia concentrations also exceeded the M A C at a number of stations in December 2002: Bailey Ditch (M10, 5.233 mg/L), Teskey Way Ditch (A17, 4.23), and Interception Ditch (A 16, 3.41 mg/L, A18, 1.77 mg/L). Semiault Creek and the downstream regions of Interception Ditch and Chilliwack Creek experienced seasonal trends for both ammonia and nitrate. In general, concentrations were found to be higher and more variable in the wet season, with the highest concentration recorded during winter and fall runoff events (e.g. December 2002, October 2003). This further supports the assumption that nutrient influx to the streams is primarily from overland runoff and that there is the potential for higher nitrogen loads to be transported to the stream during winter rainfall events. While manure and fertilizers are applied to agricultural fields in the summer months there is less runoff and higher biological uptake of nutrients both by crop and aquatic plants. This may account for the lower nitrate and ammonia levels the agricultural waterways at this time of the year. The fact that ammonia and nitrate concentrations in Bailey Ditch are relatively high year-round suggests that this portion of the watershed may be contributing nitrogen to Interception Ditch. The contributing area for Bailey Ditch includes the Bailey Landfill site, but the site is also downstream of a number of agricultural operations. Waterways surrounded by agricultural operations also had significantly higher levels of orthophosphate than the forested reference sites. There are no provincial or federal guidelines for phosphate in streams. However, to control eutrophication the EPA recommends that total phosphorus should not exceed 0.1 mg/L in streams (Mueller and Helsel, 1999). Concentrations of orthophosphate exceeded 0.1 mg/L in Semiault Creek, Bailey Ditch and in the downstream sections of Interception Ditch and Chilliwack Creek in August 2002 and October 2002. In Semiault Creek, and in the lower reaches of Interception Ditch and Chilliwack Creek, the total P concentration in sediments was also greater than in the forested tributaries; however the BAP was lower. While bound to sediments, phosphorus is not available to plants or organisms, and does not pose an immediate threat of eutrophication. However, if physical conditions 169 (such as pH, temperature, DO) change to favor the dissolution of the phosphate complexes, the sediment-bound phosphorus may be re-released to the water column. Where concentrations of these nutrients are high, they warrant concerns about toxicity to fish and accelerated eutrophication (which can lead to decreased oxygen levels in the water). Eutrophication (excess algal growth and plant proliferation) can reduce the potential use of water for recreation, industry, and drinking purposes. Furthermore, the conversion of ammonium to nitrate in streams will remove oxygen from water and adversely affect fish populations in the streams. 9.2.1.2 Dissolved Oxygen (DO) Although DO levels were below the provincial water quality guidelines in several agricultural streams during the summer months, DO levels were not found to be significantly different than the forested sites. Semiault Creek and the lowest regions of Interception Ditch had DO levels below 9 mg/L (the minimum requirement for buried embryo/alevin life stages) on all sampling dates except July 2003. The downstream region of Chilliwack Creek was the most impacted by oxygen deficit, with concentrations below 6.2 mg/L on three of the four sampling dates. The flow in these streams is relatively stagnant, which combined with higher summer temperatures and nutrient concentrations likely contributed to the low DO levels at these sites. 9.2.1.3 Specific Conductivity B.C. Working Water Quality Guidelines indicate that specific conductivity should not exceed 700 uS/cm for drinking water, and 700 to 5000 pS/cm for irrigation purposes (depending on soil and crops) (MWLAP, 1998). These levels were not exceeded at any station throughout the sampling period. However Semiault Creek consistently had the highest levels in the watershed reaching levels of 550 uS/cm in December 2002. Other streams surrounded by agricultural activities (Interception Ditch, Bailey Ditch and Teskey Way Ditch) also had higher conductivity values on this date. This is likely an indication that sediment and manure are entering the stream during runoff events. 9.2.1.4 Temperature Spot measurements of temperature (i.e. sampling once per station on each sampling date) are limited in their usefulness since diurnal fluctuations cannot be determined. In addition, the data on temperature do not allow for the determination of abrupt changes in stream temperature which have been shown to be harmful to some aquatic species. Still, Interception Ditch (A 16 and A18) and Semiault Creek have relatively high temperatures in the summer months, which is likely the result of a combination of the 170 stagnant nature of these watercourses and the minimal shade cover. Increased temperatures reduced the solubility of oxygen, which combined with the elevated metabolic oxygen demand, may impact many fish species. In addition, temperature influences other parameters (such as pH) and can therefore influence the solubility of other chemical species and their effect on aquatic life (MWLAP, 1998). 9.2.7.5 pH pH levels in the watershed are within provincial guidelines (6.5-9.0). However, the results indicate that many of the agricultural streams have significantly lower wet season pH values in comparison to the forested tributaries. Manure and fertilizer tend to be acidic, and consequently, runoff from agricultural fields and manure storage areas could be the cause of the lower pH values noted in the receiving waterways (Sharpley et al., 1998; Smith, 1994). This lowering of pH is of concern because the additional hydrogen ions compete with metal cations for positions on sediment exchange sites. As a result, more metals are found in the water (bioavailable fraction) which increases the risk of toxicity to aquatic plants and organisms. 9.2.1.6 Metals in Water and Sediment The following elements are of interest in agricultural watercourses because they are frequently found in manure and fertilizers and once in waterways they may present toxicity problems for aquatic biota: Cu, Zn, Cr, Mn, Cd, and Fe (McBride and Spiers, 2001; Nicholson et al., 1999; deVries, 2003; Here and Tessies, 1996). Concentrations of sediment-bound Cu, Zn, Fe and Cd in agricultural streams were not found to be significantly different than in the forested streams; this is likely due to the high concentrations observed in Elkview Creek. However, concentrations of these metals do appear to be elevated, and are generally higher than the rest of the hillslope tributaries. Significantly higher levels of dissolved iron and manganese were found in the agricultural streams in comparison to both the urban and forested tributaries. Zinc was not found to be above detection limit (>0.01mg/L) for any of the water samples during monthly sampling. However, during first major runoff event after the summer months (October 2003 storm samples) Zn levels rose slightly above 0.03 mg/L (the Canadian Water Quality Guideline for the Protection of Aquatic Life) in Semiault Creek. In sediment, Zn concentration exceeded the ISGQ (123 mg/L) at all agricultural sites. Water sampled did not have detectable levels (> 0.05 mg/L) of dissolved Cu; however sediment results indicated that concentrations exceeded the lowest effect level (LEL) for Cu in all watercourses. 171 In the agricultural sites, the highest concentrations (> 40 ppm) were found in the lower reaches of Interception Ditch. Cu and Zn are of greater concern when found together since their toxicity increase when present together in the aquatic environment (Anderson, 1988). Fe and Mn are often found in association in nature, and natural background concentrations are generally high in the region. In agricultural areas of the watershed Fe concentrations were elevated in both water and sediments, while high Mn concentrations were only detected in water. The likely sources are agricultural runoff, and release from Fe and Mn bearing minerals. The mean Fe levels in water at the agricultural sites (0.80 mg/L) was over 8 times greater than mean levels for the forest (0.10) and urban (0.09) sites. The B. C. Water Quality Guidelines for the Protection of Aquatic Life for Fe in water is 0.03 mg/L, and Fe concentrations in all agricultural watercourses exceeded this level on all sampling dates with one exception; Semiault Creek had a level of 0.21 mg/L in July 2003. On this date the highest Fe concentrations in sediments were recorded at this site (89488 ppm). High dissolved Fe levels (>1.6 mg/L), as well as sediment 4}ound Fe levels greater than the severe effect level (SEL) (4380 ppm) were noted in the lower reaches of Interception Ditch. Peak Mn concentrations (>0.2 mg/L) were also measured in Semiault Creek and the lower sections of Interception Ditch. It is interesting to note that Semiault Creek, which had the highest concentrations of sediment4Dound Fe and dissolved Mn, had the lowest dissolved Fe concentrations. The occurrence of these metals in water and sediment are influenced by pH, dissolved oxygen, and redox potential. As dissolved oxygen and redox potential decrease, Fe and Mn oxides become soluble, but Mn oxides are more easily dissolved than Fe oxides (Singh and Steinnes, 1994). Manganese oxides are capable of oxidizing Fe 2 +, and therefore M n 2 + is found in solution before dissolved Fe 2 + as the redox potential progressively decreases (Stumm and Morgan, 1970). Cadmium has cumulative and highly toxic effects in all chemical forms (EPA, 2004). Water samples did not have detectable levels of Cd. Sediment samples, however, had concentrations above the severe effect level at all agricultural stations. Cd levels in Semiault Creek were double the concentrations at most of the other agricultural sites. The lower reaches of Interception Ditch also had elevated concentrations of Cd. The likely sources are manure and phosphate fertilizers applied to adjacent agricultural fields. Zinc and copper (which are also found in higher concentrations in these areas) are known to increase cadmium's toxicity. It is encouraging that most metals were not found in detectable concentrations in the water column. However, the high concentrations in sediment could pose a risk to aquatic biota should the metals be released into the dissolved state. 172 9.2.2 Impacts of Urban Land Uses 9.2.2.1 Intensive Urban (M19) While this study focused primarily on the potential impact of the new suburban residential hillside development (Promontory), station M l 9 provides an idea of the impact from the more intensive urban activities in the lowland area. Station M19 is located on the upper reaches of Chilliwack Creek below a number of stormwater outfalls that drain a relatively dense urban area. Results obtained from this station show elevated streamwater concentrations of nitrate, magnesium, calcium and the highest potassium concentrations in the watershed. Stormwater discharges arc likely also responsible for the relatively high specific conductivity values and low pH values measured at this station. The most significant impact at this station appears to be elevated metal concentrations in the sediments. Concentration of Ca, Cr, and Mn were the highest measured in the watershed. Pb, Fe, Cu, Cd, Na, K, Ni and Zn also showed elevated concentrations compared to other stations within the watershed. A number of metals had concentrations that were above their respective LEL (Cu, Pb, Fe, Ni), PEL (Cd) or interim sediment quality guideline (Zn). This indicates possible toxic impacts to aquatic organisms, and significant degradation of the aquatic ecosystem in this area. However, due to the lack of sampling stations along this creek, it is not certain how localized the effects of stormwater discharges to the watercourse are (i.e. it is not known how far downstream these conditions persist). Overall, metal concentrations at this station are within the range found by McCallum (1995) in a stormwater study conducted in the Brunette watershed (a heavily urbanized watershed in the LFV). 9.2.2.2 Sub-Urban Residential Hillslope Development: Comparison with Forested Control Area Comparison of water and sediment conditions between forest and urban tributaries on the hillslope provides some indication as to whether the new hillslope development (Promontory) is impacting the water quality of the hillslope tributaries. Streamwater orthophosphate concentrations, dissolved Mg and K concentrations, and K levels in bed sediments were significantly higher in the residential sections of the hillslope. Potassium chloride is a major component in fertilizers, which may be the cause of higher potassium levels in the hillslope streams. Lawn fertilizers, animal wastes, grass clipping, and household detergents have all been identified as sources of phosphorus in residential areas (Washbusch et al., 1999). The potential source of magnesium is less clear; however, Mg is found in construction materials (e.g. cement) and may reach streams with stormwater runoff, or it may be released when soil is disturbed during construction activities. Enrichment of sediments with trace metals was not apparent in the urbanized section of the hillslope. This was unexpected since urban areas have often been associated with non-point source pollution from impervious surfaces, particularly with metals and phosphates (Washbusch et al., 1999; Characklis and 173 Weisner, 1997). The Promontory development is still relatively small and primarily residential, and it is thought that stream impacts associated with urban activities and urban stormwater runoff may not yet be sufficiently evident. Still, a difference in the trace metal content between the 2002 and 2003 sediment samples was evident in the developed section of the hillslope, suggesting that some changes may be occurring in the area as development proceeds. Construction sites are not thought to be important sources of metal contamination, unless the soil is already contaminated (EPA, 2002b). Since much of the Promontory area is still being developed, it may have been more useful to have measured a pollutant associated with the construction phase of urban development, such as sediment. It has been shown that sediment runoff rates from construction sites are typically 10 to 20 times greater than those of agricultural land, and 1000 to 2000 times greater than forested areas (EPA, 2004). 9.2.3 Conditions in the Forested Area Water and sediment quality-for Elkview Creek (F13) and Parsons Brook (F12, F14) were generally reflective of what was expected at the control stations, with a few exceptions. The forest hillslope tributaries exhibited the lowest concentrations of streamwater orthophosphate, dissolved ions (Ca, K, Mg), most trace metals and showed the highest wet season DO concentrations. Relatively low concentrations of dissolved metals (Fe, Mn) were also found. The lack of a significant difference between wet season nitrate concentrations of the forest and agricultural sites might reflect a high natural contribution leached from forest soils, contribution from small hobby farms located on the hillslope, and/or leakage of septic systems from rural residential areas. Geological sources are likely responsible for much of the Co, Ca, A l , N i and Cr observed in the watershed since little spatial variability is observed. None of these elements were found in concentrations above the detection limit for water, and concentrations of Co, Cr and Ca in the streambed sediments were generally below the average background levels in B.C. (see Table 7.7). However, concentrations of N i were above the lowest effect level (16 ppm) at all sites throughout the watershed. Parsons Brook (F13) had some of the lowest concentrations in sediments for most metals. In contrast, the upper station on Elkview Creek (F12) had the highest concentrations of A l , Co, Ni , Ca in the watershed (with the exception of Chilliwack Creek), as well as elevated concentrations of Cr, Zn, Cd, Cu and Mn. Cu levels were higher in Elkview Creek (> 60 ppm at F14) than in any of the agricultural streams, but all other hillslope sites had levels near average levels (26 ppm) in the Vancouver region (see Table 7.7). Elkview Creek was also one of the few sites that had concentrations of Pb that were above detection in July 2003. Cd, Zn and Pb are often found in association in natural environments (MWLAP, 1998), which may indicate that the elevated 174 concentrations of Cd and Zn at this site are at least in part from geological materials. The high concentrations of all these metals recorded at this site may be attributed to higher rates of weathering of the exposed bedrock and shale above this site. Differences in the geology of the area draining Elkview Creek and Parsons Brook account for some of the variations in water quality between the two streams. Elkview Creek generally had the lower conductivity values and concentrations of dissolved ions (Ca, K and Mg) in the watershed, and higher Na concentrations than Parsons. However, these differences were relatively small compared to the more land use impacted regions in the watershed. 9.2.4 Cumulative Effects (Downstream Trends) Rarely does a single land use dominate a drainage basin; rather the effect of land use throughout the watershed is a conglomeration of all the individual land uses in the watershed. In the Chilliwack Creek watershed, the different land uses are distributed so that forested areas are located on the upland hillslopes, agricultural is below in the valley bottom, and the urban centers are located the furthest downstream. The new residential development of Promontory is also located in the hillslope with tributaries draining into the agricultural reaches. As a result, differences in the characteristics of the landscape are often difficult to separate from those of the land use itself. In addition, it is difficult to separate the effects of progressive changes in stream processes from land use effects when land use patterns change along the stream (Grove, 2001). Still, downstream trends along Interception Ditch and Chilliwack Creek give some indication of the spatial cumulative effects in the watershed. Distinctive downstream trends of nutrients and various elements (dissolved and sediment-bound) were observed in the watershed. In water, concentrations of ammonia, orthophosphate, dissolved elements (Ca, K, Mg, Fe, Mn), conductivity values and temperature were generally low in the headwater and increased along Interception Ditch with progression downstream. Dissolved oxygen showed the opposite trend, decreasing downstream. Sediments concentrations of Fe, Cu, Cd, Zn, K, Na (which are all found in manure and fertilizers) increased along Interception Ditch. The spatial patterns are thought to reflect the cumulative effects of agricultural activities. It is possible that Elkview Creek (which had high concentrations of these metals in sediment and drains into Interception Ditch) is contributing to the higher concentrations in the lower reaches of Interception Ditch. However, the fact that Ni , Co, Cr, and A l (which are not associated with agricultural but were also found in higher concentrations in Elkview Creek) did not show increases along Interception Ditch suggests that agricultural operations are likely responsible for the observed increasing trends. 175 Determining the impact of the Promontory development on water and sediment quality in the lowland is complicated by the fact that the variables measured (particularly trace metals and phosphate) have both urban and agricultural sources. However, since concentrations in the urban area were generally below background concentrations, this area of the watershed is not likely contributing to pollution of the lowland streams. As the hillslope development continues to expand, it will likely begin to impact water and sediment quality downstream. Measuring a parameter unique to urban activities (e.g. hydrocarbons) would help to distinguish the urban impact in the agricultural system. The impact of both agricultural and urban activities can be seen in the high metal concentrations in sediments, extremely low DO levels and high specific conductivity measured in the lower reaches of Chilliwack Creek (site M20). Agricultural tributaries seem to be the primary source of the Fe and Cd found at this site since high levels are not seen at the intensive urban site upstream. Most of the iron at this site is held up in sediment. In contrast, Mn and Pb (which are not found in high concentrations in the sediments collected in agricultural streams) are elevated in the upper section of Chilliwack Creek suggesting urban activities are their main source throughout the stream. A combination of inputs from agricultural tributaries and urban activities likely contributed to the high Cu, Zn and phosphorus levels measured at station M20. 9.2.5 Influence of Storm Events on Water Quality The relative contribution of contaminants from storm events versus those resulting from background flows is an important consideration in the development of stormwater management strategies. Characklis and Wiesner (1997) showed that "even unremarkable storm events can contribute the equivalent of weeks or even months of background contaminant loading in a 24 hour period". In agricultural areas, large quantities of manure are applied to the fields in the fall to ensure sufficient winter storage. At this time of the year, conditions (e.g. greater rainfall, minimal vegetation, and exposed soil) promote erosion and large amounts of manure and associated contaminants (metals, nutrients) are transferred to the stream with surface runoff (particularly during the first major runoff event of the wet season). While this study could not look at storm loadings in detail, data was collected during the first major runoff event after the dry summer period. Data from monthly grab samples showed that concentrations were generally higher and more variable during high flow periods (wet season). This was most obvious on 12-Dec-02 when sampling occurred during the initial stages of a runoff event Spikes in nitrate, ammonia and conductivity were seen in Bailey Ditch (which receives water from urban and agricultural land, and may also be influenced by 176 runoff from the Bailey landfill site), along Interception Ditch and Semiault Creek. It was previously suggested that these high concentrations were likely attributable to runoff from agricultural activities. Additional sampling from the initial stages of a large storm event in October 2003 provided further evidence of the impacts of storms events. Nitrate and metals (Mn, Zn) often associated with agricultural activities (particularly livestock manure) showed significant increases in concentration at the agricultural stations during the storm event. A l and Fe also showed variations with the storm progression. It is thought that the source may not be anthropogenic, but rather natural (e.g. soil components). Both of these metals are found in high concentrations in local soils; therefore flushing out of mobile pools of weathered elements held in the soil may be responsible for the increases at these sites. The difference in response between stations may be due to differences in flow/catchment characteristics upstream of the site. Concentrations of both A l and Fe reached a maximum at the 5 hour (±1 hr) point, about 3 hours after peak runoff flow. Another possible source of the A l and Fe is re-suspension of bed sediment, or soil erosion (which may account for the longer than expected lag time). While it is acknowledged that these observations come from a limited number of samples, the results suggest that storm water runoff may be an important mechanism in transporting nutrients and dissolved metals to streams in the lowland agricultural area, and to a lesser extent in the recent urban development on the hillslope. It should be noted that a weakness of this study, and of many water quality investigations, is the use of concentration values which do not take into account the influence of discharge on water quality. Associating discharge with water quality variables is useful for a number of reasons. Firstly, investigating the relationship between flow and a particular water quality parameter can give an idea as to the origin of the parameter (Albek, 2003). For example, surface runoff generally carries soils and associated metals and contaminants, subsurface runoff leaches DOC and nutrients from the soils, and groundwater often provides most of the major ions associated with weathering (Chapman and Kimstach, 1996). Secondly, knowing the hydrologic flow at a sampling station would allow for estimates of contaminant loadings for specific water quality variables to be made. However, a major part of contaminant transport to streams takes place during high flow events (Sanden et al., 1997). Not only are contaminants transported with surface runoff during a precipitation event, studies have shown that some contaminants can be re-suspended during storm events. Consequently, water quality data collected on a monthly basis are inadequate for the assessment or modeling of many water quality problems as storm event samples are underrepresented or missed. 177 9.3 Relationships between Land Use and Water and Sediment Quality The use of contributing areas and 100 m buffer zones gave similar results when used to correlate land uses with water and sediment quality. This was somewhat unexpected since other studies have found that using buffer zones gave better relationships than the use of contributing areas (Addah, 2002), particularly in agricultural areas where topography is generally flat and does not encourage runoff. It may be that in the Chilliwack watershed, the drainage system (including the smaller agricultural tributaries and ditches) are very effective at collecting runoff and transporting it to Interception Ditch or Semiault Creek. If this is the case, land outside the 100 m buffer region would have an impact on water quality. Correlation results showed that agricultural land use activities significantly increase the nutrient concentrations in the watercourses. The strongest relationships were found with ammonia in the wet season, suggesting contaminated runoff as a major source of nutrients to streams. The relationship with nitrate was weaker and only found in the wet season. Other studies in the LFV suggest that the use of contributing areas may be inappropriate for determining relationships with nitrate (Berka, 1996; Wernick, 1996). It is thought that nitrate values are less influenced by nearby land uses due to the nitrification (conversion of ammonia to nitrate). Percent total agricultural land and percent total arable land were the best indicators of water pollution from agricultural activities. In addition to nutrients, these indices were positively correlated to conductivity in the wet season, and negatively correlated with pH in the dry season. In terms of metals, both were also good indicators of higher levels of Cd, Co, Cr, Cu, Fe, Zn, P, Ni in sediment, and of Fe and Mn in water. The strongest correlations were found with Cd, Cu, Fe, Zn and P which are frequently added to livestock feed as growth promoters, and as a result are also found in manures (Nicholson et al., 1999; McBribe and Spiers, 2001; and Sharpley et al., 1998). As expected, correlations with %forest showed that an increase in forested area within a contributing area resulted in water quality conditions reflective of the control stations. The lower anthropogenic inputs of nutrients and metals typical of forested regions are reflected in the negative correlations with orthophosphate, conductivity and most metals (in both water and sediment) associated with agriculture. Since most forested areas occurred on the steeper hillslopes, it is not surprising that an increase in forest is associated with higher DO levels and lower temperatures where flows are generally higher and more turbulent and canopy cover typically shades the tributaries. It is interesting to note that during the wet season, no relationship was found with nitrate or ammonia. This is consistent with the previous finding that there was no significant difference between nitrate-N concentrations in agriculture versus forest sites; it was suggested that septic systems from rural residences or hobby farm activities may be contributing nitrate to the stream. 178 Only a few relationships were found between urban land use indices and the water/sediment parameters. There was a similar lack of significant relationships found between %TIA and the water/sediment parameters. These results are inconsistent with findings from other studies which generally show increases in metals, and to a lesser extent nutrients, from impermeable surface areas and urban centers. However, the major urban centers were excluded from the analysis and consequently, the hillslope development of Promontory made up the majority of the urban land included in the correlations. This residential area is relatively small compared to other urban areas, and thus contaminant inputs are likely not as apparent. In addition, the collected water and sediment data (previously discussed) suggest that agriculture is a much larger contributor of nutrients and trace metals to the watercourse at this point. Since many of the contaminants associated with urban activities are seen in elevated concentrations in the agricultural part of the watershed, even if the Promontory development was contributing metals or nutrients to the streams, the correlations would likely be masked by the higher levels observed in agriculture. Dissolved Mg and K were the only parameters that showed relationships that were unique to the %residential land. There are a number of limitations inherent in the assumptions made when using contributing areas to evaluate land use-water quality interactions. First, in using the contributing area method the spatial variability in pollutant loading to streams from the different land use activities within a contributing area is considered to be random (i.e. no effort is made to distinguish the distance of a land use from the sampling stations, or its relative importance). However, pollutant loads to the stream from different land use activities may be correlated over space. For example, no distinction is made between land uses located near the stream or sampling station and land use activities located further away, which would likely have a lesser impact on water quality at the station. It would be interesting to look at the possibility of incorporating geostatistical methods (which could quantify this spatial correlation) into the technique. Unfortunately, this would prove difficult due to complexity of the system. There are numerous factors affecting the transport of the contaminants to the stream, and the fate of contaminants in the stream, and each contaminant would behave differently. A second assumption is that all runoff and contaminants originate within the contributing area itself. However, the fact that water flows downstream implies that if the distance between sampling locations is not great enough, the observations will not be independent. Other factors which could influence the accuracy of the associations between variables include: errors in the mapping of the land use, and in the delineation of contributing areas. Finally, the percent of land use does not give any measure of the intensity of the land use activities. A number of studies have shown that land use indices which reflect the intensity (such as stocking density or rate of fertilizer application) are better indicators than the type and area of land use. 179 10 SUMMARY AND CONCLUSIONS 10.1 Land Use The Chilliwack Creek watershed has a number of land uses distributed in four relatively distinct areas. Agricultural is the predominant land use in the lowland valley, and is dominated by dairy farms and horticultural activities. The agricultural land reserve (ALR) limits the urban expansion of the urban centres of Chilliwack and Sardis, where seventy six percent of the population currently lives. This has led to a gradual shift from predominantly single family homes to a greater mix of single and multiple family housing. However, urban infilling and densification are expected to reach a maximum in the near future. As a result, the most substantial changes in land use in the watershed took place in the Promontory development area of the hillslope. Hillslope communities are expected to absorb approximately 37% of the regions growth over the next four to seven years. At present, the drastic shift from forest to low density residential sub-divisions has resulted in an increase in impervious surface area, which has altered the hydrology of these catchments. The %TIA (total impervious surface area) for Teskey Creek is currently 26% (near 30%, the value at which streams become significantly degraded), and the %TIA for Lefferson Creek is just below 10% (the value where stream degradation has been shown to begin). As development of the hillslope continues, the natural equilibrium of the existing watercourses on both the hillside area and of the receiving ditch system in the lowland areas will be affected. The increased imperviousness and hydraulic efficiency of streamflow in these urban hillslope tributaries will cause an increase in stormwater discharge which, depending on stormwater management practices, could increase stream bank erosion and cause flooding downstream. 10.2 Climate and Hydrology The analysis of rainfall distribution over a 3 year period revealed that on average less than 2% of storm events (based on daily rainfall) exceeded 60 mm, and 70% of the total annual rainfall volume was generated by events of less than 30 mm. This suggests that stormwater management strategies which incorporate source control methods (such as on-site infiltration) to deal with runoff volume of smaller events would be more effective at mitigating the impacts of development than conventional stormwater management systems. Under Chilliwack's new stormwater management plan over 75% of the rainfall would be captured and detained on site, which if effective, would greatly reduce the rate and volume of runoff reaching nearby streams. The comparison of storm response characteristics between the urban and forested sub-catchments suggests that the Promontory development has altered the hydrologic response of the Teskey sub-catchment. Results indicate that under the conventional development practices used in the Promontory 180 development to this point, the peak runoff rate is up to 1416% higher and the lag time is up to 31 hours shorter at Teskey (26%TIA) than at Parsons (forested, 4%TIA). For the intermediate events, the peak runoff rate is 4.2 times greater and the lag time is 13.3 hours shorter at Teskey, on average. While the Promontory development on the hillslope is still relatively small, it is already beginning to impact the hydrology of the system, even for minor (low rainfall and intensity) events that are the majority of the storms. 10.3 Water and Sediment Quality Lowland agricultural activities and the intensive urban centers were found to be the major source of NPS pollution in the watershed. Agricultural streams had significantly higher concentrations of ammonia and orthophosphate year round than forested tributaries, as well as elevated nitrate levels in the wet season. Nutrient input to the streams seemed to be greatest during winter rainfall events, which flush manure and fertilizer residues from fields into the adjacent waterways. Ammonia concentrations exceeded provincial guidelines directly after a large event in December, and nitrate levels reached 7.6 mg/L during the first large rainstorm in the fall 2003. Evidence of eutrophication is most pronounced in the more intensively agricultural streams which had higher nutrient concentrations - namely Semiault Creek (A2) and the lower reaches of Interception Ditch (A 16, A18). Minimum dissolved oxygen concentrations recorded at these stations are below the 9.0 mg/L guideline for salmonid embryos during the summer months. The highest concentrations of Fe, Cu, Zn, Na, P and Cd in sediments, and Fe and Mn in water, were generally found in the agricultural area. Sodium, which is associated with agricultural manures, was also found in higher concentrations in the sediments of agricultural waterways. Semiault Creek, which had the most intensively used agricultural land base, is the most degraded agricultural watercourse with Cd, Fe, Zn and Cu sediment concentrations above their lowest effect level, and high concentrations of Co. The worst overall sediment quality was seen in upper reaches of Chilliwack Creek, likely due to runoff from nearby impermeable surface areas. The upstream station contributes high concentrations of Mn, Cr, Pb, Fe, Cu, Cd, Ni , K and Zn that accumulate in the sediment and may adversely affect habitat and biota should the metals undergo changes to the bioavailable state. Concentrations of copper, zinc, iron, nickel, cadmium and zinc all exceeded the provincial guidelines for sediments. Distinctive spatial patterns of specific conductivity, DO, nutrients (ammonia and orthophosphate), dissolved elements (Ca,K, Mg, Fe, and Mn) in water, and Fe, Cu, Cd, Zn, K, P and Na in sediments were observed in the watershed. Concentrations of DO were generally high in the headwaters and decreased along Interception Ditch with progression downstream; all other parameters showed the opposite trend, 181 increasing downstream. These spatial patterns are thought to reflect the cumulative effects of agricultural activities. Poor water and sediment quality in the lower regions of Chilliwack Creek (site M20) reflect the cumulative effects of upstream urban and agricultural activities. Station M20 showed extremely low DO levels (<6.2 mg/L) and high specific conductivity for most of the year, as well as high levels of most metals and phosphorus in sediments. Agricultural tributaries appear to be the primary source of Fe and Cd, while urban activities are the likely sources of Mn and Pb. At this point, the impact of the Promontory development on water quality appears to be minimal. Tributaries draining the development area did show significantly higher concentrations of orthophosphate and potassium (dissolved and sediment-bound) as compared to the forested tributaries, suggesting the residential lawns and activities may be contributing fertilizers and household detergents to the streams. Dissolved magnesium concentrations were also elevated in this area, possibly due to construction activities. A geologic source is likely responsible for much of the Co, A l , N i and Cr observed in the watershed. Elkview Creek had the highest concentrations of these elements in the watershed (with the exception of Chilliwack Creek), as well as elevated concentrations of Cr, Zn, Cd, Cu and Mn, which were attributed to higher rates of weathering over the visibly exposed bedrock above this site. 10.4 Relationships between Land Use and Water and Sediment Quality Land use indicators were calculated for each sampling site by delineating contributing areas and 100 m buffers around the sites using a GIS. The use of contributing areas and 100 m buffer zones was found to yield similar results when used to correlate land uses with water and sediment quality Percent total agricultural and percent arable land were the best indicators of water quality degradation from agricultural activities as they were significantly correlated to high levels of nutrients (N03"-N, NFi4+-N, PO43*), higher specific conductivity, high concentrations of dissolved Fe and Mn and most ions, high levels of sediment-bound Cd, Cu, Fe, Zn and P, and to low pH values. The percent of forest cover was the best indicator of water and sediment quality, having the opposite correlations to percent total agricultural. With the exception of increased dissolved K and Mg, residential areas were not found to influence water or sediment quality. It is thought that this is due to the small size of the hillslope development, and the larger agricultural inputs. 182 11 RECOMMENDATIONS Stormwater management will become increasingly important in Chilliwack and other areas in the Lower Fraser Valley as development continues to expand into the surrounding hillslopes. Increased climatic variability and surface modifications (land use changes) are the key issues to consider when dealing with runoff management issues. Hillslope developments present additional challenges as slope stability and increased stormwater runoff pose a threat to downstream areas. As can be seen in this study, water and sediment quality impacts of the Promontory development are minimal to this point. However the hydrology in the system has already been affected. By incorporating low impact and source control methods in the design of any new development further impacts of future developments can be eliminated (or at least reduced), thereby, freeing up resources in the future to address the remediation of existing pollution concern in the agricultural area. Below are some recommendations that would be useful in minimizing the impact of hillslope development in the future, and improve the current state agricultural NPS pollution in the watershed. Incorporate LID and source control methods into the design of new developments: Traditional stormwater management approaches (such as curb and gutter) are effective at eliminating on-site flooding by quickly conveying runoff to a BMP (e.g. detention pond) or stream. However, they often result in an increase in runoff volume, downstream flooding, and provide a mechanism for further degradation of receiving waters (erosion, water quality and habitat degradation). Source control and low impact development (LID) approaches should be incorporated in the overall management strategy to help achieve stormwater and pollution reduction. These principles are based on controlling stormwater at the source using a combination of a number of integrated micro-scale infiltration, retention and detention areas that are distributed throughout the development site, and a reduction in impervious surfaces (Coffman, 2000). There are many strategies which could be used to reduce the amount of impervious surface areas when designing new residential developments including smaller lot sizes, narrower streets, the use of alternative pavements (e.g. porous paving materials) in driveways and sidewalks, and alternative street designs to the traditional grid patterns. Where possible runoff should be directed to pervious areas (e.g. grass swales, biorentention areas, and infiltration trenches) in order to disconnect the impervious surfaces from the streams. These functional landscapes/pervious areas allow stormwater to infiltrate into the underlying soil promoting pollutant treatment (through adsorption, filtration and sedimentation), groundwater recharge and runoff volume reduction (through infiltration). 183 Encourage residents to manage their stormwater: Rainwater harvesting from rooftop runoff for later use in watering lawns and gardens can help reduce runoff flow to surface waters. This practice could be encouraged by offering rain barrels to residents at a subsidized price. Preserve environmentally sensitive features and buffer strips: Hillslopes are sensitive environments, and factors such as drainage conditions, slope stability and riparian areas are even more important to consider in development planning. A complete inventory of these sensitive natural environmental features should be conducted so that these areas can be protected when designing the development. There are a number of small streams in the upland that drain into the valley. Riparian buffers should be maintained along all these streams to minimize downstream flooding, erosion and deleterious effects on the aquatic habitat. Wide riparian buffers will also help infiltrate stormwater and filter pollutants before they reach the stream, and allow the incorporation of detention ponds and wetland within the buffer zone. Consider pollutant removal in the design of detention ponds: Stormwater ponds are generally constructed to control flooding; however they have been shown to be a useful tool in removing some of the pollutants in stormwater runoff before they are flushed into the natural watercourse (Brydon, 2004; Bartone et al., 1999; Kennedy and Mayer, 2002). Thus, when detention ponds are built they should incorporate features known to enhance contaminant removal in the ponds, such as vegetation, wetland soil types, and continuous baseflow. Proper maintenance and cleaning to remove contaminants will also be important in their long term management. Consider climatic variability in design of the stormwater management plan: The increase in seasonal rainfall and more frequent heavy rainfall events (that has been predicted in other studies) should be considered when dealing with stormwater management issues. Stormwater management practices should be designed with future rainfall patterns in mind, and not based on historical data to ensure that they can adequate deal with the higher runoff volumes in the future. Facilitate adoption of LID through public education and developer incentives: There are a number of challenges in implementing LID including the risk associated with the performance uncertainty of these new development practices. Homeowners are often concerned that without conventional controls, such as curb and gutters, they will need to deal with issues such as basements flooding and property damage. Furthermore, many people view the reduction of street width or 184 construction of detention ponds as undesirable and unsafe. Consequently developers are worried about market acceptance, higher costs and liability issues. In Chilliwack development costs for the LID residential projects are approximately $800 higher per lot than for conventional systems (due to requirements for redundant stormwater facilities in case the LID facilities don't perform as expected) (PSAT, 2003). Providing incentives for developers in conjunction with public education on stormwater issues and LID would help overcome some of these obstacles. Use Chilliwack as a case-study for evaluating LID practices: Low impact development (LID), which incorporates source control methods, is a relatively new concept in stormwater management and not yet widely implemented. Chilliwack is currently in the process of experimenting with these techniques in some of their newer hillslope developments projects, and as a result, could be used as a case-study. Hydrologic, water and sediment quality data from this study provides a baseline for pre-development and for the early stages of development using conventional practices in the watershed. Comparing adjacent development sites, one built using traditional stormwater control measures and the other using LID practices, would provide a needed assessment of the effectiveness of these LID methods in retaining pre-development hydrology and as a mechanism for preventing or reducing pollutant in stormwater runoff from development sites. Currently, limited research has been conducted on the various LID practices - and consequently, there is very little scientific data available for making decisions about which Best Management Practices (BMP) function most effectively under what conditions, or what variables directly affect the efficiency of the different designs. In addition, the effects of infiltration methods on slope stability should be investigated. Encourage infilling and densification in Chilliwack and Sardis: While LID practices and infiltration methods will likely help minimize some of the environmental impacts of hillslope development, as development continues in the area the risk of potential contaminant and hydrologic impacts increase. Infilling and densification in the low elevation urban center of Chilliwack and Sardis should be encouraged in order to reduce urban encroachment on hillslopes. Implement BMPs for sedimentation and erosion during the construction-phase: There will be significant amount of construction activity in the upland portion of the watershed as development continues. It is recommended that best management policies be implemented prior to any land clearing in order to limit the impact to the watercourse and aquatic system, particularly from sedimentation. These policies should include a detailed sediment and erosion control plan and monitoring during the course of clearing to ensure that the plan is properly implemented, as well as long-term monitoring for disturbed sites until green-up is established. More information can be obtained by 185 referring to Best Management Practices Guide for Stormwater, Appendix H: Construction Site Erosion and Sediment Control Guide (GVSDD, 1999). Implementation of agricultural waste and nutrient management practices: The use of best management practices should be encouraged to reduce agricultural NPS pollution. In particular, manure and fertilizer application should be managed to ensure that nutrients and trace metals do no build up within the soil. This could include: relocating excess manure, reduced applications of inorganic fertilizers, improved manure handling (i.e. better timing of manure applications) and storage, and improved feeding strategies. 186 REFERENCES Addah, J. (2002). The Impact o f Agricultural Land Uses on Water and Sediment Quality in the Agassiz/Harrison Hot Springs Watershed. M . S c . Thesis, B . C . Department of C i v i l Engineering, University of British Columbia. Addiscott, T . M . (1999). Nitrate and Health. In: Wilson, W . S . , Bal l , A . S . and R . H . Hinton (eds). Managing Risks of Nitrates to Humans and the Environment. Royal Society o f Chemistry: Cambridge, pp. 247-249. Albek, E . (2003). Estimation of Point and Diffuse Contaminant Loads to Streams by Non-Parametric Regression Analysis o f Monitoring Data. Water, Air, and Soil