Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Relationship between agricultural land use and surface water quality using a GIS: Sumas River Watershed,… Berka, Caroline Svatava 1996

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

Item Metadata


831-ubc_1996-0351.pdf [ 9.89MB ]
JSON: 831-1.0086974.json
JSON-LD: 831-1.0086974-ld.json
RDF/XML (Pretty): 831-1.0086974-rdf.xml
RDF/JSON: 831-1.0086974-rdf.json
Turtle: 831-1.0086974-turtle.txt
N-Triples: 831-1.0086974-rdf-ntriples.txt
Original Record: 831-1.0086974-source.json
Full Text

Full Text

RELATIONSHIPS BETWEEN AGRICULTURAL LAND USE AND SURFACE WATER QUALITY USING A GIS: SUMAS RIVER WATERSHED, ABBOTSFORD, B.C. by CAROLINE S VATAVA BERKA B.Sc.Eng., Queen's University, Kingston, Ontario, 1991 A THESIS SUBMITTED IN PARTIAL F U L F I L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF MASTER OF SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES Resource Management and Environmental Studies We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A June 1996 © Caroline Svatava Berka, 1996 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada Date June ZH ; 1*1*11* DE-6 (2/88) ABSTRACT The Sumas Prairie is one of the most intensively used agricultural floodplains in Canada. Dairy farmers are the traditional occupants of the floodplain, but the past 20 years have seen the development of turf and vegetable farms, and large hog, chicken and turkey operations. Nutrient management and related water contamination have been recognized as major issues over the past decade, but due to the non-point nature of the pollution it has been difficult to analyse the contributing sources and to mitigate the impacts. The Sumas River watershed was thus investigated as an illustration of how land use activity affects water quality with a focus on non-point source pollution from agriculture. A Geographic Information System (GIS) was used to integrate resource data for the watershed, which included surficial geology, soils, current and historic land use, agricultural intensification and population growth. River sediments and water quality were analysed in seasonal, spatial and historical contexts. GIS overlay techniques were used to summarize land use activities within the drainage areas to sampling points, or "contributing areas". Indices of land use activities were developed within the contributing areas and correlated to the water quality parameters to identify significant relationships. Examples of land use indices included nitrogen loadings over contributing areas and animal stocking densities. Zinc concentrations in river sediment were elevated from those measured twenty years ago and are attributed to agricultural sources while high chromium and nickel concentrations occur from natural sources. The nutrient concentrations and fecal coliform counts in stream water increased dramatically in the rainy season. Manure, particularly when spread in the wet season due to lack of winter storage, is likely entering the stream via runoff. Dissolved ii oxygen levels were low in this same period, and on a site specific basis year round. One tributary, Marshall Creek, was found to have elevated nitrate levels in the summer with the suspected source being contaminated groundwater from the neighbouring Abbotsford aquifer. Animal stocking densities and surplus nitrogen loadings were found to be high, as compared to values found in the literature. Significant relationships were identified between surplus nitrogen applied to farm land, amount of clay soil texture by area, and ammonia-N concentrations in the wet season. Similarly, these two land indices were negatively correlated with dissolved oxygen levels in both the wet and dry seasons. Nitrate-N concentrations were positively correlated to amount of clay and organic soils in the contributing area, but negatively correlated to the amount of sandy texture. The results indicate agricultural best management practices need to be more aggressively pursued in the watershed, with regard to amount of manure and time of application. Areas with higher nitrogen loadings coincided with areas of water quality degradation. Techniques developed in this research can be used to evaluate the impact of non-point source pollution from agriculture on stream water quality in a quantitative manner and provide watershed managers with a tool to address non-point source pollution. The densification of animals and farms on the floodplain, emergency responses due to frequent flooding, and the impact of contaminated groundwater on the stream, are issues that should be given renewed attention. iii TABLE OF CONTENTS Abstract ii Table of Contents iv List of Tables viii List of Figures ix List of Abbreviations and Symbols xi Acknowledgements xii 1.0 Introduction 1 1.1 Background Issues 1 1.2 Objectives 4 1.3 General Methodology 4 1.3.1 GIS Database 6 1.3.2 Land Resource Data 6 1.3.3 Overall Land Use 7 1.3.4 Land Indices 7 1.3.5 Water Resource Data 9 1.3.6 Water Quality Sampling, Analysis and Trends 9 1.3.7 Correlation Investigation and Significant Relationships. . . .10 1.4 The Watershed Study Unit 11 1.5 Planning and Community Initiatives 12 2.0 Land Characteristics and Use in the Sumas River Watershed 14 2.1 Physical Setting 14 2.2 Land Use 14 2.3 Trends in Land Use 19 2.4 Trends in Agricultural Land Use 22 2.5 Measures of Agricultural Activity 24 2.5.1 Livestock Densities 25 iv 2.5.2 Surplus Nitrogen 26 2.5.3 Livestock Densities and Surplus Nitrogen in the Sumas Watershed 28 3.0 Water Resources in the Sumas River Watershed 34 3.1 Climate 34 3.1.1 Precipitation 34 3.2 River and Drainage Network 37 3.2.1 General 37 3.2.2 Annual Flow Regime 40 3.2.3 Flooding 41 3.2.4 Groundwater Resources 44 3.2.5 Discharge Permits 46 3.3 Fisheries and Wildlife 49 4.0 Water Quality Investigation 53 4.1 Water Quality Indicators 53 4.1.1 Trace Metals in River Sediments 53 4.1.2 Surface Water Indicators 55 4.2 Sampling Methodology 60 4.2.1 Sediments 60 4.2.2 Surface Water 60 4.3 Laboratory Analysis 62 4.3.1 River Bed Sediment Analysis 62 4.3.2 Water Constituent Analysis 64 4.3.3 Quality Analysis and Quality Control 66 4.4 Spatial and Temporal Trends Shown by Water Quality Indicators 68 4.4.1 Trace Metals in River Sediments 68 Spatial Trends 70 v Temporal Trends 75 4.4.2 Surface Water Constituents 78 Seasonal Trends 83 Spatial Trends 86 Historical Trends 90 Variation with Discharge 92 5.0 Land Use - Water Quality Relationships 94 5.1 Development of Indices 97 5.2 Relationships Between Indices by Enumeration Area and Water Quality 100 5.3 Relationships Between Indices by Contributing Area and Water Quality 103 5.3.1 Canadian Portion of Watershed 103 5.3.2 Whole Watershed 107 6.0 Summary and Recommendations 113 6.1 Land Use Summary 114 6.2 Water Quality Summary 115 6.3 Summary of Relationships Between Water Quality and Land Use 116 6.4 Recommendations 118 7.0 References 121 APPENDICES 127 Appendix A: Within Site Replication Results and Comparison of Chemex and 1993/94 Digestion Methods 127 Appendix B: Accuracy and Precision Calculations 128 vi Appendix C: Sediment Sampling Results 129 Appendix D. Water Sampling Results 130 Appendix E: Box Plots of Water Quality Results by Station and by Sampling Date 137 Appendix F: Wet Season/Dry Season Water Quality Results 146 Appendix G: Historical Wet Season Water Quality 152 Appendix H : Scatterplots of Historical Water Quality vs. Discharge 155 Appendix I: Nitrogen Model Methodology by Brisbin (1995) 157 Appendix J: Summary of Nitrogen Balance for the Canadian Portion of the Sumas Watershed 160 Appendix K: Variables Used in the Spearman Rank Correlation 161 Appendix L: Indice Values by Contributing Area 164 Appendix M : Spearman Rank Correlation Coefficients for Canadian Portion of Sumas Watershed 168 vii LIST OF TABLES Table 2.1 Land Use in the Canadian and U. S. Portions of the Sumas Watershed 19 Table 2.2 Growth in Farm Numbers 22 Table 4.1 Variability in Water and Sediment Site Replications 67 Table 4.2 Accuracy Measurements for Water Analysis 68 Table 4.3 Comparison of Measured Trace Metal Concentrations with Example Background Concentrations and Preliminary Guidelines and Criteria 69 Table 4.4 Tests of 1974 and 1993/94 Trace Metal Concentrations in Sediments . . . .78 Table 4.5 Comparison of Surface Water Chemistry with Natural and Criteria Levels 80 Table 5.1 Spearman Rank Correlation Coefficients for Canadian Portion Analysis 104 Table 5.2 Spearman Rank Correlation Coefficients of Whole Watershed Analysis 108 viii LIST OF FIGURES Figure 1.1 Location of the Sumas River Watershed in the Lower Mainland Region 2 Figure 1.2 Thesis Framework for Studying Non-point Source Pollution 5 Figure 2.1 Topography of the Canadian Portion of the Sumas Watershed 15 Figure 2.2 Surficial Geology of the Sumas Watershed 16 Figure 2.3 1988 Land Use in the Canadian Portion of the Sumas Watershed 18 Figure 2.4 Land Use Changes 1954-1988 - Canadian Portion of the Sumas Watershed 20 Figure 2.5 Growth in Number of Dwellings from 1981 -1991 21 Figure 2.6 Agriculture Census Totals, 1986 and 1991 23 Figure 2.7 Frequency of Animal Densities by Farm in the Canadian Portion of the Sumas Watershed 30 Figure 2.8 Surplus Nitrogen Loading Estimates for the Canadian Sumas Watershed/Region 32 Figure 3.1 1994 Daily Precipitation and 1984-1994 Average Daily Precipitation at the Abbotsford Airport 35 Figure 3.2 1994 Cumulative Precipitation and 1984-1994 Average Cumulative Precipitation 36 Figure 3.3 The Sumas River System 38 Figure 3.4 Average Annual Hydrograph and 1994/1995 Hydrograph. Sumas River at Huntingdon border crossing 42 Figure 4.1 Water and Sediment Sampling Stations 61 Figure 4.2 Stream Sediment Trace Metal Concentrations 71 Figure 4.3 Scatter Plot of Ni vs. Zn Concentrations in Sediments 74 ix Figure 4.4 Comparison of 1993 and 1994 Sediment Sampling Results 76 Figure 4.5 Comparison of 1974 and 1993/94 Metal Concentrations in Sediments . . . .79 Figure 4.6 Box Plots of Ortho-P by Sampling Date and Dissolved Oxygen by Station 82 Figure 4.7 Water Quality Plots: Chloride and Ortho-P, Nitrate-N and Ammonia-N 84 Figure 4.8 Comparison of Nitrate-N Behaviour in Marshall Creek and the Sumas Mainstem 88 Figure 4.9 Mstorical Wet Season Values of Nitrate-N in the Sumas Mainstem 91 Figure 5.1 Animal Density by Enumeration Area 95 Figure 5.2 Sampling Stations and Contributing Areas in the Sumas Watershed 96 Figure 5.3 Surplus Nitrogen Loading by Contributing Area 99 Figure 5.4 Trends of Water Quality Values and Animal Numbers in Three E As . . . .102 Figure 5.5 Ammonia, Dissolved Oxygen, Surplus N , and Soils: System of Relationships in Canada 105 Figure 5.6 Nitrates, Soils and Land Use: System of Relationships in Canada 106 Figure 5.7 Ammonia, Dissolved Oxygen, Surplus N , and Soils: System of Relationships in the Whole Watershed 109 Figure 5.8 Nitrates and Soils: System of Relationships in the Whole Watershed . . .110 Figure 5.9 Cumulative Indices and Nitrate-N Values. Upstream to Downstream on the Sumas Mainstem 112 x LIST OF ABBREVIATIONS A N D SYMBOLS AES Atmospheric and Environmental Services A U E Animal Unit Equivalent A L R Agricultural Land Reserve C C M E Canadian Council of Ministers of the Environment (formerly CCREM) C C R E M Canadian Council of Resource and Environment Ministers DO Dissolved Oxygen (mg/L) E A Enumeration Area ESP Environmental Sustainability Parameter (IRC, 1994) GIS Geographic Information System GPS Geographical Positioning System ha hectares (-2.47 acres) rs Spearman rank correlation coefficient TRIM Terrain Resource Inventory Management U B C University of British Columbia U S D A United States Department of Agriculture U.S. E P A United States Environmental Protection Agency WMS Waste Management Survey (by IRC Integrated Resource Consultants) xi ACKNOWLEDGEMENTS This project would not have been possible without funding from the following: 1) Tri-Council (National Sciences and Research Council, Medical Research Council, Social Sciences and Humanities Research Council), Green Plan Eco-Research Project; 2) Environment Canada and Department of Fisheries and Oceans, Fraser River Action Plan; 3) T.P. (Pat) Harkness of the Ministry of Attorney General, Provincial Emergency Program (Chilliwack); and, 4) Ministry of Environment, BC Environmental Research Scholarship. I thank all of these agencies for their support. In addition, the following have provided valuable data, reports or personal expertise for which I am very grateful: Frank Wright, Dyking, Drainage & Irrigation Superintendent, City of Abbotsford; Orlando Schmidt, Dairy Producer's Conservation Group; Chuck Timblin, Whatcom County Conservation District; Whatcom County Planning and Engineering Departments; Fish and Wildlife and Environmental Protection staff of the Ministry of Environment, Lands and Parks; and Klohn-Crippen Consultants Ltd. I would like to thank my supervisor, Hans Schreier, for his endless support and enthusiasm. His creativity and patience throughout this work have assisted me enormously in facing many challenges. My sincere appreciation is also given to the other members of my thesis committee: Ken Hall, Les Lavkulich and Eric Bonham, for their precious time and guidance. Thank you to Sandra Brown, Yao Cui, Alice Kenney, and Wayne Tamagi for being so helpful (and patient!) with the GIS and hypertext work for this project. I would also like to thank the Soils laboratory and Environmental Engineering laboratory staff, especially Jane Huang and Amy Que, for their help conducting the laboratory analyses. My R M E S colleagues I thank for their shared knowledge, understanding, humour, and friendship. To my mother, who has provided love and support in so many ways -1 can never thank you enough! Finally, to my best friend and companion, Graham - thank you for giving me my required daily doses of your wisdom, love, and laughter. xii 1.0 Introduction Water pollution from intensive agriculture is now recognized as a problem of global proportion (Owens, 1994) but, due to the non-point nature of the pollution, it has been difficult to analyse the contributing sources and to mitigate the impacts. The aim of this thesis is to explore relationships between land use activity and water quality in the Sumas River watershed in Abbotsford, B.C. , with an emphasis on agricultural non-point source pollution. 1.1 Background Issues The Sumas River originates in Whatcom County, Washington State, and joins the Fraser River east of Abbotsford, British Columbia (Figure 1.1). The watershed comprises a very flat floodplain surrounded by steep mountain slopes. This floodplain, known on the Canadian side as the "Sumas Prairie", is considered one of the most productive agricultural areas in Canada and is part of the B.C. Agricultural Land Reserve (ALR), which protects farmland from urbanization. As a result, high population growth within the City of Abbotsford is steadily urbanizing the mountain slopes of the watershed. This urbanization can impact water quality in terms of urban inputs of non-point source pollution, and water quantity in terms of the hydrological effects of clearing the land and increasing the impervious area. Flooding of the farmland is a frequent phenomenon, varying from local minor floods to major flooding from the Nooksack River in Washington when it overflows. Flooding has serious ramifications to health, environment, economics and emergency procedures due to the high number of animals in the floodplain, and the presence of an important national 1 2 transportation corridor, the TransCanada Highway. Sedimentation within the stream system is also a major issue in the watershed, influenced by flooding and wind erosion in the Prairie. It is of particular concern in the headwaters, where a natural landslide contributes very fine asbestos sediment with high concentrations of trace metals to the stream (Schreier, 1987). The chemical and drainage properties of the asbestos rich sediments provide very poor conditions for vegetative growth. The agricultural activity in the Sumas River watershed is economically important to the Lower Fraser Basin. The intensive agriculture has developed into a $250 million investment (District of Abbotsford, June 1993), and gross farm revenues in 1991 were greater than $68 million, while expenses were greater than $53.5 million (IRC, 1994). The Prairie produces 17% of all dairy products in British Columbia and is also largely devoted to vegetable production (District of Abbotsford, June 1993). Recent intensification from rapid increases in poultry and swine production has produced livestock densities amongst the highest in Canada. The intensity of agricultural activity in both the Canadian and U.S. portions of the watershed have created water quality problems and degraded fish habitat in various reaches of the Sumas River system (IRC, 1994, Puget Sound Water Quality Authority, 1990). Agricultural activity has also been named responsible for the contamination of groundwater in the Abbotsford aquifer (Liebscher et al., 1992), a portion of which is included in the western area of the watershed study boundary. The above issues prompted the Sumas Sustainability Study, spearheaded by the District of Abbotsford (now the City of Abbotsford, having amalgamated with Matsqui) and begun in 1993. The Study brought together the City of Abbotsford, the Ministry of 3 Environment, farmer/producer groups, Whatcom County, and other interest groups and stakeholders to address some of the above issues. Flooding concerns received particular attention. The Westwater Research Centre, UBC, was able to contribute research regarding water quality aspects of sustainability to the Study through the Fraser Basin Ecosystem Study, funded by the Tri-Council Secretariat Eco-Research Program. Thus this thesis plays a dual role. It is one of the watershed case studies investigated in the Fraser Basin Ecosystem Study, which explores general sustainability issues in the Lower Fraser Basin, and it is also part of Westwater's contribution to the Sumas Sustainability Study, which is a locally driven, action-oriented initiative. 1.2 Objectives The research documented in this thesis was undertaken with the following objectives: 1) Quantify current land use activities and intensity. Where possible make comparisons to historic land use and identify trends. 2) Document the current status and historic changes in water quality and trace metals in sediments. 3) Compute a terrestrial nitrogen balance. 4) Relate land use to water quality using a Geographic Information System (GIS). 1.3 General Methodology The general approach used to study non-point source pollution and pursue the objectives of this study can be represented by the flowchart in Figure 1.2. The following Figure 1.2 Thesis Framework for Studying Non-point Source Pollution GIS DATABASE Land resource data compilation f Computation of overall land use trends 1 f Development of land indices on a contributing area basis (e.g., surplus N loading, animal density, soil and land use types) Water resource data compilation I Water quality sampling (over one year period) Spatial and temporal analysis of results Summarize water quality dynamics with wet season / dry season trends Correlation investigation between land indices and wet season / dry season water quality values Identification of significant relationships and implications to the watershed 5 paragraphs describe the components of this flowchart, and collectively describe the study methodology. 1.3.1 GIS Database A GIS is a computer database and graphics system which can perform spatial analytical functions with geographically referenced data (Burrough, 1986). The GIS TerraSoft© is extensively used in this thesis to integrate the land resource data, and analyse and display the spatial pattern of land use activity. Any entity with a position in space can be represented, along with its descriptive attributes, in digital form in a GIS. If resource information is not available digitally (i.e., computer files with positioning coordinates and attribute information related to those coordinates), then hard copy maps or photos can be digitized using computer graphics hardware and the GIS software. The strength of using a GIS in non-point source pollution investigations lies in its spatial analytical capabilities. The main spatial function employed during this research is the "overlay". This allows one to investigate a variety of questions regarding the interactions of the resource data, limited mainly by the availability and quality of the data, and the creativity of the investigator. As simple examples, which farms and how many farms fall within a subdrainage area, or the total area of different surficial soil types within a census enumeration area, can both be determined. In this way, land resource data can be quantified and integrated in various spatial contexts. 1.3.2 Land Resource Data The provincial government's TRIM (Terrain Resource Inventory Management) maps, at a 1:20 000 scale, were used as a base map for the GIS. Digitally formatted soil survey maps for Canada (Luttmerding, 1980) and the U.S. (USDA, 1992) were input and the coding 6 generalized according to surface and subsurface texture, parent material, and drainage properties. Land use data was digitized into the GIS from aerial photos for different time periods. Farm locations were obtained from a combination of aerial photos, orthophotos, GPS (Geographical Positioning System) data from a waste management survey (IRC, 1994), and windshield surveys. Farm attribute information, such as size, number and type of animals, and types and hectares of crops, was also obtained from the IRC waste management survey. Population and agricultural census data were obtained for the enumeration areas covering the watershed from Statistics Canada, along with a geographical key for the enumeration areas which were digitized into the GIS. For the U.S. portion of the watershed, the digital base map showing stream and road topology was obtained from the Whatcom County Planning Department. Land use data was also provided by this department, in the form of a database containing land parcel centroid coordinates with associated parcel size and land use code attributes. Information on agricultural activity, including crop acreages in the watershed, dairy farm locations and approximate herd sizes was provided by the Whatcom County Conservation District. 1.3.3 Overall Land Use The land resource data compiled above was summarized and analyzed for the whole watershed using a combination of database queries, GIS overlays, and the generation of various plots or graphs. Where possible, historical trends were also examined. The treatment of the land resource information is discussed in more detail in Chapter 2. 1.3.4 Land Indices The land resource data summarized for the whole watershed was then quantified for 7 each area draining to a water sampling station, called "contributing area" in this thesis. Some investigation of resource data by enumeration area was also explored, but this proved to be limited in its application to water quality relationships, as described in Section 5.2. Quantifiable measures, which were labelled "land indices", included hectares of soil texture, parent material, and drainage types within a contributing area, and hectares of land use types within a contributing area. Measures specific to agricultural activity included animal density and surplus nitrogen application to the land. Again, the land indices by contributing area were obtained by GIS overlays with the soils, land use or farm location maps with the contributing area boundaries. The development of land indices is discussed further in Section 5.1. One index which was given particular attention in the research work was the amount of surplus nitrogen loading to the land. This index value was considered the measure which would most directly impact the nutrient content of the streamwater, and thus have the potential of representing agricultural inputs to the stream. To signify the importance of this index, computing a terrestrial nitrogen balance was included as an objective of this thesis. The nitrogen balance uses a mass balance approach which considers sources and sinks of nitrogen to compute a surplus value applied to the land. This surplus is theoretically available to enter groundwater through leaching or surface water via runoff. The model used to compute the mass balance includes a calculation of: manure production of nitrogen by animals minus management, application and volatilization losses; inorganic fertilizer application minus crop requirements; and atmospheric deposition and denitrification losses. A more detailed description of the mass balance calculations is given in Section 2.5.2. 8 1.3.5 Water Resource Data Water resource data was collected to describe the biophysical setting of the water quality investigation. Knowledge of the hydrology, history, and other characteristics of the water resource allow for the interpretation of the water chemistry results in the proper context. The water resource data was not used directly in the GIS analysis or correlation investigation, but rather as a background knowledge base to assist in determining the implications of the results. Information collected includes: flow and precipitation data, reports on hydrology and flooding, fisheries data, reports on groundwater in the Abbotsford aquifer, and waste discharge permits. Chapter 3 describes the sources of information and provides summary highlights. 1.3.6 Water Quality Sampling, Analysis and Trends Grab samples of river bed sediment were taken at 23 sites in August 1993 and August 1994. Trace metal concentrations for copper, chromium, nickel and zinc in the sediment were determined for the two sampling times. Differences from year to year were computed, and results plotted in the upstream to downstream direction to determine spatial trends. The results were also compared to a baseline dataset recorded in 1974 by Westwater Research Centre, U B C , by statistically testing for significant changes in concentration levels over the 20 year time period. Water quality sampling was conducted at 16 stations, including one control station, over one annual cycle. Dissolved oxygen, pH, temperature, conductivity, chloride, nitrate-N, ammonia-N, orthophosphate and dissolved organic carbon were all measured. Samples for fecal coliform analysis were collected on three of the eight sampling dates. The results were 9 analysed in seasonal, spatial and historical contexts, the latter by using comparisons with historical water quality data. The spatial analysis was accomplished by a visual comparison of the data for each site as represented by a box plot, and also by plotting the data in an upstream to downstream direction. By separating the sampling dates into "dry" and "wet" season categories, and averaging the results for each site by these seasons, a clear picture of the seasonal dynamics was shown. The averaging of the data into wet/dry seasons served to dampen the "noise" of the water quality data, allowing for clearer and stronger correlations with the land indices. Chapter 4 encompasses the whole of the water quality investigation, including the methods and results of streamwater constituents and sediment trace metal analyses. 1.3.7 Correlation Investigation and Significant Relationships The measures and analyses of resource data for land and water are finally brought together in a correlation analysis between land indices and water quality values, in order to determine significant relationships. A Spearman rank correlation coefficient matrix was generated to identify the strength of relationships among wet season water quality averages, dry season water quality averages, and the land indices. Strong relationships imply not only that the measured land characteristic or activity has an influence on the water quality, but that the land index, which is often more easily and cheaply measured than the water chemistry, may be a good environmental indicator for water quality. Chapter 5 describes the correlation investigation and significant relationships found. Chapter 6 summarizes the conclusions drawn from Chapter 5 and preceding chapters, and provides recommendations, or implications, to the watershed. 10 1.4 The Watershed Study Unit This thesis methodology uses the watershed as a study unit. It assumes that stream water quality is influenced mainly by the properties of the land defined in area by the natural surface drainage boundaries of the stream. It also assumes that water quality at a point is influenced mainly by the subdrainage, or contributing area, to that point. These are reasonable assumptions, as many biological phenomena and human activities are water dependent, and surface water drainage boundaries are comparatively easily delineated. However, boundaries and properties of influence are not necessarily simply defined. Firstly, while watershed boundaries are defined by topography, data pertaining to the watershed is commonly collected according to political boundaries. This often forces generalizations, assumptions, and inherent inconsistencies in summarizations for contributing areas, but is an unavoidable reality and is rarely ameliorated with the selection of some other basis for a study unit. The Sumas watershed includes jurisdictional areas from two nations (Canada and the U.S.), three municipal districts (Matsqui and Abbotsford- now amalgamated to the City of Abbotsford, and Chilliwack), one county (Whatcom), and one Regional District (Central Fraser Valley). Most of the land area on the Canadian side falls within the City of Abbotsford. Consequently, much of the research focused on the City of Abbotsford portion of the watershed, particularly on the Sumas Prairie floodplain. Another watershed study limitation is that water quality can very well be impacted by forces which extend beyond the watershed boundary, such as air pollution, and groundwater pollution as discovered in the water quality investigation of Chapter 4. Despite the above recognized limitations, the spatial pattern of measured water quality parameters in the 11 watershed context is useful in interpreting the relationships between land use and water quality (Cook, 1994), especially when investigating non-point source pollution. 1.5 Planning and Community Initiatives The non-point source pollution investigation should recognize the related goals, programs and initiatives of the Sumas watershed community, although only the Canadian Sumas community is described here due to data availability. The City of Abbotsford must harmonize the background issues described earlier with the objectives listed in their Official Community Plan (District of Abbotsford, 1993). These objectives include: the diversification and promotion of economic activity; "the protection, conservation and maintenance of lands that are environmentally sensitive or subject to hazardous conditions by limiting development in order to reduce high damage costs"; the provision of adequate supply of housing types but the minimization of potential conflicts between housing and other land uses; and, the preservation of agricultural land and the promotion of agricultural industry. Some of these objectives are addressed in the previously mentioned Sumas Sustainability Study. Another program, which works toward the control of soil erosion, is led by the Sumas Prairie Soil Conservation Group. This group consists of farmers, advisors, and administrators from the public sector, who are responsible for initiatives which not only preserve productive soils, but protect the aquatic environment, as efforts to control soil erosion also tend to control the extent of pollution in agricultural runoff (Owens, 1994). Other groups which work towards addressing environmental issues and commodity group concerns include the Sustainable Poultry Farming Group, the Hog Producers Sustainable Farming Group, the Dairy 12 Producers' Conservation Group, and the Sumas Prairie Dyking and Drainage Committee (Schmidt, O., pers. comm., 1995). Soil erosion and the amount of manure applied are recognized problems by farmers of the Sumas Prairie, and irrigation and drainage, including flood mitigation, understandably receive high levels of attention and funding. It appears that water quality in the Sumas River system is not perceived to be a significant problem by the farmers; there has been no concern expressed in the Sumas Sustainability Study meetings regarding the quality of irrigation water for crops nor for animal watering. The issue of water quality in the Sumas River as yet receives low priority. There are many community groups, however, that work or recreate on the river system and may be more concerned with water quality issues. These groups, such as the: no-motors club, the waterski club, the walking society, the rowing club, Ducks Unlimited, the Rod and Gun Club, the historic society, and bird watchers (Wright, F., pers. comm., 1995) may perform a variety of habitat/ecosystem enhancement works, and may help to change the level of priority currently given to water quality issues. It is hoped that the relationships discovered during this research will add to the knowledge and facilitate decisions regarding the abatement of agricultural non-point source pollution, for the Sumas Prairie farmers, other community groups, and government agencies alike. 13 2.0 Land Characteristics and Use in the Sumas River Watershed 2.1 Physical Setting The Sumas River watershed comprises a long, flat-lying valley and the steep slopes of the Sumas and Vedder mountains, which sandwich the valley to the northwest and southeast respectively. The valley averages 5 km in width and extends approximately 35 km from the Fraser River in the north to the Nooksack River, Washington, in the south. The two mountains reach elevations of approximately 800 m, while the valley bottom remains close to mean sea level (NHC and Hamilton, 1994; Klohn Leonoff, 1989). The topography of the watershed is depicted in Figure 2.1. The Sumas valley was an arm of the sea during much of the Quaternary period. The sea filled the valley with 300 m or more of marine deposits which was later topped with less than 5 m of post-glacial lacustrine deposits from Sumas Lake (Armstrong, 1983, cited in N H C and Hamilton, 1994). Figure 2.2 shows the predominance of sand and loam in the surficial geology (Luttmerding, 1980; USD A, 1992). Noticeable features in this figure include the sandy lake bottom in the northeast portion of the valley, and the gravel deposits of the Abbotsford aquifer in the west. 2.2 Land Use A large portion of the Sumas watershed is a floodplain created by the Sumas river itself and affecting neighbouring river systems. Human land use in the Sumas watershed is historically dominated by dairy farming and pastureland use in the floodplain, with some harvesting of forested areas in the surrounding mountainsides. Vegetable production is also a 14 •o <u £2 0) I CO CO E CO <D O o o_ c CO T3 CO c CO O CD o >» . c CL 2 o> o CL CM 3 D) 15 Figure 2.2 Surficial Geology of the Sumas Watershed. Source: Generalized from the 1:25 000 soils map (Luttmerding, 1980) and digital maps provided by Whatcom County Planning Dept. (USDA, 1992). 16 primary activity on the prairie, particularly in the more sandy areas of the floodplain. The last few years have seen a considerable rise in sod, hog and poultry production. The majority of the low lying land falls within the British Columbia Agricultural Land Reserve (ALR), and is therefore is not zoned for urban or industrial use. Economic and demographic forces shape and define land use surrounding the ALR, as well as modify activity within the ALR, resulting in an ever-changing, dynamic watershed. This dynamism was largely evident during the study period, most notably expressed through subdivision activity on the slopes of Sumas mountain, industrial construction on the Sumas highway corridor, and new animal housing units on the Sumas prairie. Various "snapshots" of land use information for different years were compiled from census data and aerial photographs. Generalized land use categories, were used in the interpretation of aerial photos, available for the years 1954, 1963, 1979 and 1988 at a 1:10 000 scale. Land use polygons were digitized from the aerial photos into the GIS using the API90 Analytical Plotter. Following the land use classification scheme used by the Ministry of Agriculture (Sawicki and Runka, 1986), the following generalized categories were used: 1) Agriculture; 2) Forest, which included areas being harvested; 3) None perceived, which were areas where no obvious activity was discernable and likely included lands kept for speculation; 4) Park/Recreation/Wetlands, which included wildlife parks, 5) Residential, which included the urban areas and agricultural communities with high residential densities, and 6) Commercial/Industrial, which included major transportation corridors, but did not include commercial cultivation of forest resources. The 1988 land use map is shown in Figure 2.3. 17 T3 CD w I— E CO CD £ c .2 o ~ c -2 g 0) £ s c •— CD o ^ o CO O co J> CD £ co c o — o CD O V) _ T3 ^ !« oo s-CD O t - CO CO CNJ CD ZJ CD 2r o CD CD -*—• CO O CD W ID •o c co O o a I I S 5 3 -x "S •g 8 * a) a ° CO *= C E £ o « - 8 p c 0 18 Current land use information for Whatcom County was provided by the Whatcom County Planning Department. This consisted of a database of land parcel centroids with x,y coordinates, a land use code, and acreage for each land parcel. Table 2.1 presents the total hectares and percentage of the Canadian and U.S. portions of the watershed for each category. The Canada/US. border coincidentally divides the watershed into two roughly equal areas, and these areas have a very similar distribution of land use categories. Table 2.1 Land Use in the Canadian and U.S. Portions of the Sumas Watershed Category Area in Canada (ha) % of Canadian Portion Area in U.S. (ha) % of U.S. Portion Agriculture 9851 59 9354 56 Forest 4463 27 5289 32 None perceived 494 3 548 3 Wetlands/Park/Recreatio 234 1 153 1 Residential 892 5 1143 7 Industrial/Commercial 550 3 297 2 Not coded 110 1 877 5 Total 16594 100 17661 100 2.3 Trends in Land Use Land use maps for the Canadian portion of the watershed were also prepared for the years 1954, 1963, and 1979 on the GIS. Figure 2.4 shows the area of each land use category from one year to the next. Note that agriculture and forest areas are reduced by a factor of ten for illustration purposes. The land areas under agriculture and forest, the two largest land uses in the watershed, 19 20 change very little relative to the original areas. Residential and industrial/commercial areas increase by factors of 4 and 6.5 respectively, but still remain a relatively small portion of the whole watershed, as indicated in the 1988 land use map, Figure 2.3. While illustrating the encroachment of urbanization into forest and agriculture, the preceding examination of land use categories does not indicate how land activity is changing within the categories. Population census data from 1981, 1986, and 1991, is shown below in terms of number of dwellings, to show increasing density of land use in the watershed. The census enumeration areas (EAs) were divided into three groups: urban, rural agriculture, and forest/suburban, which included the EAs which were previously forested but now contain new subdivisions. The highest growth rates are in this latter group, but more surprisingly, the rural agriculture group also exhibits fairly high growth rates. New subdivisions may be included in this group of EAs as well. The rate of growth in dwellings is high across all groups, indicating overall intensification of land use across the Canadian portion of the watershed. Figure 2.5 Growth in Number of Dwellings from 1981 -1991. Source: PCCensus. 8000 - . 6000 •5 4000 2000 1981 1986 Census Year 1991 Dwelling Area Categories Urban Rural Agriculture Forest/Suburban 21 2.4 Trends in Agricultural Land Use This thesis focusses on agricultural non-point source pollution, hence emphasis was placed on documenting agricultural land use trends. Using the series of aerial photos as a basis, large farms were digitized into the GIS, their approximate locations were identified by a symbol, and the year of the aerial photo in which the farm first appeared was noted. Windshield surveys supplied the locations of farms appearing after 1988. In this manner, the number of new farms added between years, and the rate of increase in farm numbers, were determined. Table 2.2 summarizes the results. Although the rate of increase in number of large farms is low, the steady increase over the years on a constant land base signifies an increased density of agricultural activity. Table 2.2 Growth in Farm Numbers Year Farm Count % increase/year 1954 224 -1963 233 0.4 1979 248 0.4 1988 271 1.0 1994 283 0.7 Intensification is more dramatically illustrated by looking at changes between 1986 and 1991 Agriculture Census data, as presented in Figure 2.6. (Note the scale factors for land area, cattle, pig and poultry numbers.) Again, the land base remains constant while number of farms slightly increase between these years. However, while the number of cattle remains 22 Figure 2.6 Agriculture Census Totals, 1986 and 1991. The Census enumeration areas cover the Canadian portion of the Sumas Watershed only. 600 —i agriculture land area (x100 ha) cattle (x100) livestock/poultry farms farms pigs (x100) poultry (x1000) 400 — _Q E z 200 0 1986 total 1991 total 23 relatively constant, the number of pigs increase by about 50% and the number of poultry by about 75%, in just five years. The windshield surveys conducted during the study period verified that new housing units for hog and poultry were being constructed, suggesting that numbers of these animal types are still on the rise. Although it has been estimated that broiler operations export 30-40% of the manure produced (Brisbin, 1995), the remainder of the manure, together with that from other types of poultry operations, the hog operations, and the dairy farms, is applied to the land. The soils and vegetation of this constant land base have a limited capacity to assimilate the increasing amounts of manure, resulting in a greater risk of contamination of the Sumas water resources. 2.5 Measures of Agricultural Activity The problem of manure management is neither newly identified nor unique to the Sumas watershed. Many European countries have been struggling for years with the problem of too much manure, not enough land, and contaminated water resources. Technical and regulatory solutions, and the efforts put toward them, vary widely. No panacea has yet been developed which would effectively handle all economic, environmental and population growth issues associated with this non-point source problem. While this thesis confirms the existence of intense agricultural activity in the Sumas watershed, the main objective is to examine how the problem is related to water quality in the Sumas River. Measures of agricultural activity, including quantity of excess manure applied to the land, are needed for this examination. 24 2.5.1 Livestock Densities One measure of agricultural intensity is livestock density, which is an animals-to-land area ratio. Animals are sometimes converted to animal unit equivalents (AUE), based on the amount of waste they generate or the pollution potential of the waste. The practice of converting animals to animal units, the conversion formula that is used, and the criteria used for regulation of manure application, varies between countries (Anderson et al., 1990). This thesis borrows conversion factors from the Ontario Agricultural Code of Practice (MOAF, 1976) which equates 1 dairy cow (plus calf) to 1 horse, 4 sheep, 125 laying hens, 1000 broiler chickens, etc. based on the amount of nitrogen in their manure. This type of conversion is useful when there are several types of animal operations common to a region, as in the Sumas watershed. A measure of what is a reasonable density based on nitrogen content of the manure is required as a guideline when evaluating the densities calculated for the Sumas watershed. Denmark regulates the amount of nitrogen per hectare by restricting densities to 2.3 dairy cow units per hectare and 1.7 pig units per hectare (Anderson et al., 1990). Ontario suggests a range depending on soil types and operation size, with more forgiving densities for larger operations and operations on clay or loam. Their values range from 2.5 to 3.7 AUE/ha for small operations on clay/loam and sand respectively, and from 3.1 to 4.7 AUE/ha for large operations (MOAF, 1976). For the purposes of this thesis, 2.5 A U E per hectare is considered the value at which nitrogen application rates may be a valid concern. It is recognized that higher values may be reasonable on some soil types, but a conservative approach must be considered in light of the demonstrated growth in animal numbers occurring in the watershed. 25 2.5.2 Surplus Nitrogen The abundance of manure in agricultural areas of the Lower Fraser Valley of British Columbia has not gone unnoticed by government agencies. An agricultural waste management steering committee has been established including representatives from the B C Ministry of Environment, Lands and Parks, the BC Federation of Agriculture, the B C Ministry of Agriculture, Fisheries and Food, Agriculture and Agri-Food Canada, and Westwater Research Centre (TJBC). This group has recently guided a study on agricultural nutrient modelling in the Lower Fraser Valley (Brisbin, 1995). The nutrient model for nitrogen calculates the amount of surplus nitrogen being applied to the land using a mass balance approach. Sources of nitrogen, including inorganic fertilizers, atmospheric deposition, and livestock nutrient production, are reduced by crop uptake, volatilization and management losses, and the result is a surplus (or deficit) applied to the soil in kg N/ha/year. The model further estimates denitrification rates in the soil and the final losses of nitrogen to the atmosphere, surface water and groundwater. For a more complete description of the model, see Brisbin (1995); an excerpt of the report with the model methodology is given in Appendix I. Brisbin's model has been adopted by Wernick (1996) for use in determining nitrogen flows in the Salmon River watershed, Langley, B.C. The methodology of this model can be described by considering a single farm. Firstly, manure production on the farm is calculated by multiplying the nitrogen production rates in kg/year for each type of animal manure by the number of each animal type on the farm. It is assumed that 30% of the broiler manure is exported in this calculation. Losses of nitrogen to the air, land and surface water via 26 volatilization, infiltration and runoff, are estimated using manure management factors for each animal type developed by Brisbin (1995). Similar factors are used to estimate the total mass of manure nitrogen which will be applied to the land. This is added to the net crop requirements, which is the difference between the nitrogen applied as inorganic fertilizer and the estimated uptake of the crops on the farm. The crop uptake is based on nitrogen uptake rates for different crop types used in Brisbin (1995). Finally, an amount of nitrogen is added to account for atmospheric input, and an amount is subtracted to account for some denitrification of the manure nitrogen. The former is assumed to be the sum of an estimated background deposition of 9 kg/ha/year and a 30% return of the volatilized nitrogen calculated in the management losses. The denitrification loss is assumed to be 10% of the net manure applied. The final result of adding and subtracting all the sources and sinks of nitrogen is the mass of surplus nitrogen produced by the farm each year. This is converted to a loading by dividing by the area of the farm, or by the area reported under crops, which gives an even higher loading rate. In this thesis, the surplus nitrogen and the two loading rates were calculated by farm, by contributing areas to sampling stations, and for the watershed overall. A summary of the overall nitrogen balance calculated for the Canadian portion of the Sumas watershed using this model is outlined in Appendix J. What is deemed "excessive" surplus nitrogen loading is a controversial and complicated issue, as some losses of nitrogen from soils through nutrient cycles is normal and expected. One way this issue is approached in Brisbin (1995) is by computing a rough potential dilution of the nitrogen in water in comparison to the 10 mg/L drinking water criterion for nitrate-N. This would be the concentration if 100 kg N/ha is diluted in 1 metre of 27 water, which could be 1 metre of rainfall (a depth easily reached over one year in the Lower Fraser Valley), or one metre of groundwater recharge. Due to this reasoning, 100 kg surplus N/ha/year is the value used in this thesis to gauge "excessiveness". 2.5.3 Livestock Densities and Surplus Nitrogen in the Sumas Watershed Animal numbers for individual farms in the Sumas Prairie were obtained from a Waste Management Survey (WMS) conducted by Integrated Resource Consultants (IRC) in the winter of 1993/1994, under contract to the B.C. Ministry of Environment and Environment Canada. This survey also provided crop information and the geographical location of 132 unique farm locations as determined by a global positioning system (GPS) unit. Farms which did not participate in the WMS (an additional 156 farms) were identified on the aerial photos and field checked to determine the type of operation. These farms were then given average values of size and animal numbers as calculated from the WMS farms by operation type. If the farms surveyed included all the largest farms, then the average values applied to the other farms may be slightly high. However, because the surplus nitrogen loading calculated is roughly the same magnitude as those calculated for the Abbotsford area by Brisbin, and by using the 1991 Census data (see Appenix J), it is believed that the resulting A U E densities and nitrogen balances calculated are reasonable estimates for the watershed. Chapter 5 will also illustrate that smaller areas within the watershed can have much higher densities and surplus loading rates than indicated by the overall figures. Using the census data, the AUE/farmed ha was calculated to be 3 AUE/farmed ha. Using W M S data for the watershed, plus the WMS averages, the livestock density was calculated to be 2.3 AUE/farmed ha. In the Whatcom County portion of the Sumas 28 watershed, the density was estimated as 1.6 ATJE/farmed ha overall. According to the latter two estimates, the densities are not too high, if the value of 2.5 AUE/ha is used as the criterion (see Section 2.5.1). This may well be the case if all the animal manure is equitably spread over all agricultural land within the watershed. However, this is not practically the case. Farmers in general spread on their own land, on land that they rent, and very occasionally on their neighbours' land by special agreement (IRC, 1994 and C. Timblin, pers. comm.). In almost all cases the land is nearby, as it is often impractical and/or uneconomical to transport the manure. Thus some farmed lands in the watershed may receive much less manure than other lands. To illustrate the fact, animal densities were calculated on a farm by farm basis, using the WMS data. Of the WMS farms, 126 were reportedly animal operations. Figure 2.7 shows the number of farms that fall into A U E density ranges from 0 to 5 and above AUE/ha (ND=Not enough Data). About 50% of the farms have densities below, while about 40%) have densities above 2.5 AUE/ha. Moreover, at least 40% of the farms below the gauge value of 2.5 are in the 2-2.5 range, which means that many of these farms will approach more critical density levels if animal numbers continue to increase. In the Whatcom County portion of the watershed, the overall AUE/farmed ha was estimated to be 1.6. Again, the same argument as above applies. Although farm densities could not be calculated with the data available, the Whatcom County Conservation District was able to provide herd sizes and acreages for 14 of the 65 dairy farms (Timblin, C , pers. comm.). The livestock densities calculated for these farms ranged from 2.6 to 7.8 dairy cows per ha, with an average of 5.4 cows per hectare. Animal densities tend to be higher on U.S. farms because they are not restricted by a quota system as are Canadian farms. 29 Figure 2.7 Frequency of Animal Densities by Farm in the Canadian Portion of the Sumas Watershed. Source: Waste Management Survey (IRC, 1994). ND= no data. 40 — i Animal Unit Equivalents/ha Herd sizes of 600 to 1000 are common, while in the Canadian Sumas watershed, dairy farms of this size are still relatively rare. However, competition with US production, to which the Canadian Sumas dairy farmers are sensitive, is steadily driving up Canadian herd sizes. Up to 700 head are now found in several barns on the Sumas Prairie (Wright, F., pers. comm.). The nitrogen balance summary in Appendix J shows the surplus nitrogen loading for the Sumas area, in Canada, computed using different data sources. These values are represented by the bar graph in Figure 2.8. The first bar represents the loading computed using the data from the WMS farms and the averages of these farms applied to the missing farms identified in the aerial photographs. For the second bar, WMS averages were not used. Instead, the GIS was queried to determine which WMS farms fell within each agricultural 30 census enumeration area. The total surplus nitrogen produced by the farms within one enumeration area was subtracted from the total surplus nitrogen estimated for that enumeration area. The difference was then distributed evenly over the watershed agricultural area within the enumeration area which was not accounted for by the WMS farms. This procedure was repeated for all the enumeration areas which overlapped with the watershed, and the summation produced a final nitrogen surplus for the watershed which theoretically included the farms missing from the WMS database. The third bar uses only census data from all the enumeration areas which overlap with the watershed, and represents the area covered by the enumeration areas which is much larger than the Canadian portion of the watershed. The final bar represents surplus nitrogen values computed by Brisbin for large and small farms added together. The Abbotsford study area in Brisbin's report includes, but covers a greater area than, the Canadian Sumas watershed. These four different values are shown to illustrate the effort carried out in this thesis work to compute quantities in different ways, using available data sources, as a method of double-checking results. In the complicated and inaccurate task of computing nitrogen budgets, accurate and precise values, although desirable, are not necessary to identify a trend. In this case, all four estimates are in roughly the same range, and all are near or above the 100 kg surplus N/ha gauge value discussed in Section 2.5.2. The overall surplus nitrogen loading calculated for the Whatcom County portion of the watershed is 68 kg/ha. As discussed in the case of A U E densities, this value is likely not indicative of the loadings on more localized areas. The impacts to the environment and human and animal health from excess nitrogen are well documented in the literature. The impacts to surface water are dependent not only on 31 surplus nitrogen loadings, but on timing of application, amount and intensity of rainfall, soil properties and various management factors, including tillage, crop residue, and cropping systems (Owens, 1994). Figure 2.8 Surplus Nitrogen Loading Estimates for the Canadian Sumas Watershed/Region. 200 Z to Q_ 1— Z5 CO 100 D a t a S o u r c e / M o d e l W M S farms + a v e r a g e s W M S farms + C e n s u s Agricultural C e n s u s Brisbin (1995) • I by farmed area by cropped area Projects carried out by the Sumas Soil Conservation Society with various Producer Groups in the Sumas watershed, and the studies guided by the Agricultural Waste Management Steering Committee, attempt to address problems due to runoff and excess nitrogen loadings. The people involved, many of whom work on the land and in the watershed daily, understand the problems well. However, due to the diffuse nature of non-point source pollution, economic pressures, public resistance to regulations, and the difficulty of regulation enforcement, the 32 solutions to the problems remain elusive. By looking at water quality and its relationships with some quantifiable agricultural land use indices, it is hoped that the work presented here will add to the knowledge base and help to target solutions. 33 3.0 Water Resources in the Sumas River Watershed 3.1 Climate The weather in the Sumas River watershed is dominated by low pressure systems, particularly in the winter, bringing heavy rains and flooding. Snow and freezing temperatures are occasionally brought by polar air in the winter. Clear skies, warm temperatures and low rainfall predominate in July and August when high pressure systems are more common. During this time, soil moisture deficiencies often develop and irrigation is required to promote high crop yields (ESL and Webb, 1987; Halstead, 1986). 3.1.1 Precipitation Approximately 75% of precipitation in the Lower Fraser Mainland falls between the months of October to March (Halstead, 1986). This wet period is evident in the hyetographs and cumulative precipitation graphs in Figures 3.1 and 3.2, which show average precipitation data for a 10 year period, and precipitation data for the study period of this thesis (March 1994 to February 1995). Water sampling dates are shown with the latter data. The total precipitation for 1994 was approximately equivalent to the 10 year average of about 1500 mm, yet the summer of 1994 was dryer than the 10 year average. The steep sections on the cumulative precipitation graphs illustrate the wet season period, while the flatter sections show the dry season. Water samples representative of both seasons were collected, and, as is illustrated in the study period hyetograph, samples were collected during a March 1994 and a February 1995 storm. 34 Figure 3.1 1994 Daily Precipitation and 1984-1994 Average Daily Precipitation Abbotsford Airport, AES gauge. 40 20 -A 0 -J 1984 -1994 average daily precipitation 40 E E, c g TS -•—• "cL ' o CL 20 ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ $ ^ g xy xy xy xy ,cy xy xy xy xy xy xy xy xy xy xy 0 n \ $ ^ \ ^ ^\ oco\ ^ \ # \ # \ n q \ n n \ Nq>° q N \ ^ ^ Date 35 Figure 3.2 1994 Cumulative Precipitation and 1984-1994 Average Cumulative Precipitation. Abbotsford Airport, A E S gauge. 3.2 River and Drainage Network 3.2.1 General The stream network for the Sumas River watershed is shown in Figure 3.3. The total drainage area of the Sumas River system is approximately 34 255 ha, roughly half of which exists in Canada, and half in Whatcom County, U.S. A large lake of over 8000 ha once occupied the Sumas floodplain, accepting flows from Saar Creek and the Sumas and Vedder Rivers (US Army COE, 1993). The lake was reclaimed in 1925 so that the land could be farmed. The Sumas Lake Reclamation Project included the "construction of the Vedder Canal, the Sumas Pump Station...the diversion of the Sumas River along with four creeks, and the momentous task of the construction of over 40 km of flood protection dykes" (District of Abbotsford, June 1993). The Barrowtown Pump Station, including an earthfill dam and upgraded dyke, replaced the Sumas Pump Station in 1984 to provide improved flood protection, drainage and irrigation. The old lake bottom is now drained by the Sumas Lake Canal and pumped into the Sumas River downstream of the dam. A network of ditches convey runoff from within the dyked area, including runoff from Vedder Mountain and the town of Yarrow via Stewart Slough, to the Sumas Lake Canal. The ditches are an important source of irrigation water during the dry season, with water licences having been granted to a large number of land owners (Harris, 1990). The irrigation and industrial licences in the Sumas River system can demand a very large part of summer low flows (NHC and Hamilton, 1994). Sedimentation in the ditches and channels is an ongoing problem requiring annual maintenance, including cleaning and deepening, to ensure adequate drainage and irrigation performance. A major 37 Figure 3.3 The Sumas River System dredging project for removal of substrate in the lower Sumas River (downstream of the Trans Canada crossing) was undertaken in 1987 to increase the capacity of the channel and minimize winter flooding problems (ESL and Webb, 1987). Normal flows of the Sumas River and its tributaries drain by gravity to the Fraser River through a floodbox in the dam beside the Barrowtown Pump Station. Flow through the floodbox is controlled by the electro-hydraulic operation of a steel gate (KPA, 1987). Two irrigation inlets from the Sumas River provide water to the Sumas Prairie. Both surface and sub-irrigation methods are used in the Prairie. The reclaimed lake, system of dykes, and other facilities of this complex hydraulic system function similarly to the polders commonly farmed in the Netherlands. Since Dutch settlers had moved into the area before the draining of the lake, this transfer of technology is not surprising. The Sumas River mainstem is dyked along its south bank to the confluence with Saar Creek (the dyke continues along Saar Creek and then Arnold Slough) to protect the low-lying area to the east from flooding. This downstream portion of the Sumas River mainstem is characterized as slough-like, with approximately a 0.02 percent gradient. High water temperatures are known to occur in the summer months due to the low water velocities. The channel is wide, typically 100 m across with sections reaching 250 m, and the substrate is predominantly silt. From Saar Creek to Vye Road, the channel becomes narrower (40 to 60 m) and meanders through farm land with a more visible current. Small gravel is the predominant bed material, and the water is generally clearer with more abundant tree cover. Although the gradient is still approximately 0.02 percent, the backwater effect from the Barrowtown Dam (see Section 3.2.2) is diminished here resulting in more evident stream 39 flow. Between Vye Road and the U.S. border, the channel gradient increases to 0.06 percent, and the width is generally less than 45 m. The bed material varies from fines in the glide areas to small gravel in the pools and riffles (ESL and Webb, 1987). The Sumas River mainstem, Saar Creek and Arnold Creek all originate in Whatcom County, U.S. There the Sumas mainstem is meandering with low stream gradients and joined by numerous tributaries and creeks. Of particular significance is Swift Creek which carries very fine sediment from a natural landslide in its headwaters to the Sumas River (Schreier, 1986). Consequently, the mainstem river bed is filled with this sediment below the confluence of Swift Creek (ESL and Webb, 1987). The principal tributary to the Sumas River in Whatcom County is Johnson Creek, which joins the Sumas River near the town of Sumas. Marshall (or Lonzo) Creek is the main tributary within Canada on the north side of the Sumas River. This creek originates from groundwater springs flowing from a ridge (ESL and Webb, 1987) near the western boundary of the watershed, or approximately the eastern edge of the Abbotsford aquifer. In the summer months, the downstream sections of Marshall Creek are also subject to backwatering from operation of the Barrowtown Pump Station. 3.2.2 Annual Flow Regime Around mid-May, or at the start of the Fraser freshet, the floodbox gates at the Barrowtown Pump Station are closed and water from the Sumas River is pumped to the downstream side of the dam. This prevents Fraser flood backwater from entering the Sumas River (US Army COE, 1993). The gates remain closed to provide irrigation water for farm land in the Sumas Prairie until September 15, the official start of the salmon spawning period. While the gates are closed, backwatering from the dam results in almost no visible flows in the 40 Sumas River from the Barrowtown Pump Station to several kilometres upstream of Hougen Park (ESL and Webb, 1987). Over the winter, the floodbox gates are normally left open and the water level in the Sumas River is generally below an elevation of +3 metres. However, if levels in the Fraser or Vedder Canal rise to an elevation of +3.5 metres within a few hours, then the floodbox gates are closed to prevent flooding of the Sumas River (KPA, 1987). The only current flow gauge on the Sumas River system is one maintained by the Water Survey of Canada at the Sumas River border crossing near Huntingdon. Figure 3.4 shows the average annual hydrograph determined from average daily flow rates measured from 1952 to 1994 at this gauge. Also shown is the hydrograph for the 1994-95 period of sampling. Again, water sampling was conducted during low flow in the dry season and also during high flow events in the wet season. Flow during the summer of 1994 was generally lower than the 40 year average. The average of daily flows over the 40 year period varied from approximately 1 to 8 m3/s. 3.2.3 Flooding Flooding of the Sumas Prairie is a frequent occurrence and has a wide range of consequences. Not only are there major flood damage costs, but also major economic and logistical ramifications due to potential highway closures and emergency evacuations of many thousands of animals. There are also potential human and environmental health consequences due to animal mortalities and pollution from animal waste facilities. Records from 1876 to before the Sumas Lake was drained, indicate that the Sumas Prairie was flooded on four occasions from the Fraser River, and three times from a Nooksack River overflow from the south. The construction of the pump station and dykes have since prevented flooding from 41 42 the Fraser River. However, these facilities were not designed to handle additional floodwaters from Nooksack River overflows. Flood protection from these facilities relies on the assumption that the volume of floodwater from the Sumas drainage is relatively small compared to backflooding from the Vedder and Fraser Rivers (Cook, 1995; US Army COE, 1993). The Nooksack overflow occurs because the Sumas River shares a portion of its watershed with that of the Nooksack River due to the flat topography between them. As a result, the western portion of the Sumas Prairie, which is not protected by dykes, has been inundated to varying extents by Nooksack overflows on 12 separate occasions since the lake was drained. Four of these floods (two in 1990) have occurred within the last 10 years. The frequency of overflows appears to be increasing, the suspected cause being the aggredation of sediment in the Nooksack channel due to the cessation of gravel mining in this river twenty years ago. Flooding of the West Sumas Prairie is often exacerbated by: 1) backwater in the Sumas River when it is constrained at the floodgates due to high Fraser and Vedder water levels; and/or 2) local storm and snowmelt events which may themselves cause localized flooding in the Prairie. In addition, it has been estimated that a major overflow could cause an avulsion (a permanent change in the direction of river flow) with catastrophic consequences. An International Task Force has been in place since 1991 to investigate the flooding issues of the Sumas Prairie and recommend various solutions. Whatcom County also initiated a Comprehensive Flood Hazard Management Plan, due for completion in 1996, which includes Nooksack overflow mitigation considerations. (Sellars et al., 1991; Cook, 1995; US Army 43 COE, 1993; District of Abbotsford, June 1993; Klohn Leonoff, 1989, 1991 and 1993; Task Force, 1991 and 1994.) 3.2.4 Groundwater Resources The Sumas Prairie is unusual in that it is a rural agricultural area serviced with drinking water by the city system. This water comes from the City of Abbotsford's wells located on the Abbotsford aquifer (ERM, 1992), the eastern portion of which lies in the watershed study area (see Figure 3.3). A steep escarpment extending north-south parallel to Sumas Way shows the visible extent of the glaciofluvial sand and gravel deposit, called the Sumas Drift, which comprises the Abbotsford aquifer. The eastern parts of the aquifer are also known to contain glacial till and clay components. Half of this 200 km 2 aquifer lies in Canada, and half in Washington State (Liebscher et al., 1992). Flow in the aquifer is mainly in a southerly direction into the Nooksack River system in Washington. However, groundwater flow from the aquifer is in several directions, including significant discharges to the east into Marshall (Lonzo) Creek, formerly via a series of large springs (BC M O E L P and EC, 1994). Halstead (1986) estimated the total discharge to all springs, prior to the development of high yield wells, to be 8.3 M m3/yr. The annual recharge of the aquifer is estimated to be 26.8 M m3/yr (BC M O E L P and EC, 1994). This is a very important aquifer, supplying industrial and municipal drinking water to the City of Abbotsford (approximately 4 M m3/yr); domestic and irrigation water to farms situated on the aquifer (approximately 3 M m3/yr); and water supply to the Fraser Valley Trout Hatchery (approximately 3.7 M m3/yr), located below the aquifer escarpment (BC M O E L P and EC, 1994). The aquifer is of significance to the Sumas River system in several 44 ways. Firstly, it is an example of agricultural land use practices causing nitrate and pesticide contamination of a water resource. In one region of the aquifer, 60% of the water samples collected had nitrate-N concentrations which exceeded the 10 mg/L acceptable maximum concentration for drinking water. Contamination by nitrate also signals the potential for contamination by other pollutants (Liebscher et al., 1992). Secondly, the Sumas Prairie depends on the Abbotsford aquifer for its drinking water. Because Marshall Creek is fed by springs from this aquifer, the water quality of the Sumas River system is affected by water quality conditions within the aquifer. Furthermore, there has been a steady decline since 1982 of water levels in nearby observation wells noted by the Fraser Valley Trout Hatchery (BC M O E L P and EC, 1994). If in fact the aquifer is being "mined", this may result in a reduction of flow in Marshall Creek, and thus a reduction in dilution potential for contaminants entering the stream. Because ground and surface water interactions are complex, both quantitatively and qualitatively, it is difficult to postulate further on the influence of the Abbotsford aquifer on the Sumas system without the appropriate monitoring and modelling studies. Other groundwater resources in the Sumas River watershed include groundwater which occurs on Sumas Mountain in sand and gravel formations, or in fractured, fissured or weathered bedrock aquifers. However, wells in these locations are generally shallow with low yields. There are also shallow and relatively low yield wells at the base of Sumas Mountain where significant sand and gravel deposits exist as old beach deposits from the previous lake, or in alluvial fans of major creeks from Sumas Mountain (ERM, 1992). The land in the areas of these other groundwater resources are not intensively farmed. However, the potential of 45 groundwater development beneath the Sumas Prairie has been considered. It is believed that beneath the 240 m or so of silt and clay lie substantial sand and gravel deposits from a pre-glacial and/or glacial Fraser River route. Deep gas wells have found sand and gravel deposits at over 300 m, but the water at this depth was found to be slightly brackish. A deep test/production well would be required to properly determine the potential to produce drinking water (ERM, 1992). If groundwater production were to be pursued in this area, impacts from land activities would likely be minimal and difficult to detect, due to the overlying layer of silt and clay and the extreme depth of the sand and gravel deposits. 3.2.5 Discharge Permits Water quality investigations for non-point source pollution should not neglect known point sources of potential contamination. The permit database for the Lower Mainland, maintained by the B.C. Ministry of Environment, Lands and Parks, Environmental Protection Branch, was reviewed to determine all permitted discharges into the Sumas River System. Three active effluent permits were identified, and one cancelled permit was also found. These permits, each unique in their terms and conditions, are briefly described below. In Whatcom County, the major point source is the sewage treatment plant for the town of Sumas. The cancelled permit (permit number PE-8618) belonged to Shell Canada Products Ltd. for discharging "from a petroleum products bulk marketing Marshall Creek via a drainage ditch". The permit limited effluent flow and stipulated concentration limits of total extractable hydrocarbons. Grab sampling and reporting of results was required. However, according to M O E L P staff, this permit was cancelled because of a change in permitting regulations. Under the new regulations, the site was not required to have a permit, 46 although the practices remained the same. Although there was likely good reason for the regulation changes, this example serves as a reminder that the potential for pollution exists from sites not regulated by governments, whether it be from industrial, agricultural, residential or other sites, and of the non-point or point source nature. Fraser Valley Milk Producers has a permit (permit number PE-4608) to discharge cooling waters from an evaporated milk plant located near the Trans Canada/Sumas Way interchange, to an unnamed tributary of Marshall (Lonzo) Creek. The quantity and temperature are limited to 2300 m3/day and 23 °C respectively, and a report of monthly measurements of these characteristics is provided to the Ministry every year. No other water quality characteristics are reported. More parameters were required to be monitored by Coaspac Meat Ltd., which discharged slaughterhouse effluent to a field which contains a ditch tributary to the Marshall (Lonzo) Creek system. Biochemical oxygen demand (BOD), ammonia-nitrogen, pH, and temperature were measured in the ditch once per year upstream and dowstream of the field site. In the five years of data available at the Ministry office, only ammonia levels in the ditch showed the tendency to increase after flowing through the field. These increases, and overall values, were fairly low. The lowest upstream value measured was <0.1 mg/L, while the highest downstream value of ammonia was 3.5 mg/L. These values are within the range measured throughout the Sumas system during this thesis work, and generally within criteria levels, both of which are given in section 4.4.2. The maximum measured discharge of effluent to the field was 0.3 m3/day and the monthly average was 4.5 m3/month. While the runoff from this field may have represented a point source of pollution, it was believed that the 47 unalarming values together with the discovery that this plant apparently shut down in April 1994, just one month after this study's sampling was begun, likely meant that the plant's activity did not influence this study to a significant degree. The major point discharges to the Sumas River system are from the Fraser Valley Trout Hatchery (permit number PE-1726), which discharges to Marshall (Lonzo) Creek, and the town of Sumas wastewater treatment system, located approximately 300 m upstream of the border along the Sumas mainstem. The Trout Hatchery obtains its water from a well tapping the Abbotsford aquifer, and discharges approximately 7000-16000 m3/day to a lagoon which flows into Marshall (Lonzo) Creek. Samples are taken once every three months of the well water and at the discharge to Marshall (Lonzo) Creek. A wide spectrum of parameters are measured. Generally, the hatchery does not have problems meeting any of the permit level requirements. The interesting exception is the permit level for nitrate-nitrogen, which is 10 mg/L. During the period of investigation for this thesis, the hatchery was in the process of having the permit amended because the water supplied from the well itself exceeded the nitrate levels permitted to be discharged by the hatchery. In this respect, nitrate contamination of the Abbotsford aquifer is decidedly impacting the water quality of Marshall (Lonzo) Creek, since the hatchery pumps water out of the aquifer and drains it to the Creek. The lagoon, however, may provide some treatment of this water. Water quality problems in the Sumas mainstem may cause one to suspect the town of Sumas wastewater treatment plant, which discharges to the Sumas River approximately 300 m south of the border. However, the facility is monitored daily to ensure that adequate treatment of the sewage is occurring and the levels meet the criteria of the Washington State 48 Department of Ecology (DOE). Water quality problems recorded in the Sumas River, including low dissolved oxygen levels and high fecal coliform levels, are attributed to non-point pollution from agricultural practices. Before the implementation of Best Management Practices (BMP's) for controlling non-point agricultural pollution in Johnson Creek, the D O E monitored the water quality in this creek in 1980/81 and 1988/89. They found conditions similar or worse than in the Sumas mainstem downstream of the treatment plant ( K C M , 1990). 3.3 Fisheries and Wildlife Inventories of fish habitat and wildlife resources provide useful information regarding the health of an ecosystem. Physical habitat for salmonid fish resources was rated low to moderate for the Sumas reach of the Barrowtown Pump Station to Saar Creek. The quality of habitat improves as one moves further upstream, with the 2-3 km reach just downstream of the international border possessing the highest quality habitat in the Canadian portion of the Sumas River. Saar Creek is considered to have good fish habitat in the U.S. upper reaches, with gravel substrate, a high gradient, and overhanging vegetation. However, during the summer, the Canadian portion of this creek has low levels of dissolved oxygen and high water temperatures, creating very poor rearing habitat for salmonids. Arnold Creek also has poor habitat throughout its stretch for similar reasons, and also due to a silty stream substrate (ESL and Webb, 1987). Marshall Creek has been identified and targeted as having a high potential for enhancement work in the Canadian portion of the system (Klassen et al., 1995). The substrates are predominantly fines mixed with small gravel, there is abundant vegetative cover, 49 and the water generally has low temperatures and turbidity, and adequate levels of dissolved oxygen. The Fraser Valley Trout Hatchery is located in the headwaters of Marshall Creek and has released steelhead and cutthroat trout into the creek since 1978 (ESL and Webb, 1987). The highest quality fish habitat on the Sumas system overall occurs in the U.S. headwaters, particularly in the many tributary creeks. Problems are encountered, as in Canada, with cattle entering streams, manure storage practices, fish blockages, channelization of creeks, and loss of riparian vegetation. The Washington Department of Fisheries has focused enhancement work on the tributary streams, particularly Johnson Creek, as the Sumas mainstem has been filled by sediment from Swift Creek and is subject to yearly flooding (ESL and Webb, 1987). The species present in the Sumas River system include coho, chum and pink salmon, steelhead and cutthroat trout, and non-salmonids such as sturgeons, carp, lampreys, whitefish, sculpins and stickleback. The chum and pink salmon spawn in the lower reaches of the Sumas mainstem while the coho spawn mainly in the U.S. headwaters and the steelhead spawn in the upper reaches. Stewart Slough and the upper reaches of Marshall Creek are also popular spawning grounds for coho, chum, steelhead and cutthroat trout. Arnold Slough supports no spawning, and a little rearing outside the summer months. Rearing habitat during the summer months is in general limited to the upper tributaries of the system due to high temperatures and low levels of dissolved oxygen. Migration and spawning activities for the salmon species generally begin in October and may last until January (DFO, 1995). The timing of migration and spawning is particularly important considering the seasonal variation in water quality as 50 discussed in Section 4.4.2. It has also been noted that coho migrations occur as distinct runs which generally occur in conjunction with significant rainfall events (ESL and Webb, 1987). Agricultural land use occurs along 91% of the stream length of the Sumas River. Many of the constraints to fish production listed in the Department of Fisheries and Oceans' database are therefore related to agriculture. These include siltation from farmlands and erosion from cattle, and the effect on water chemistry from pesticides and nutrients from agricultural runoff. Fish kills due to agricultural runoff have been reported in tributaries to the Sumas Lake Canal (DFO, 1995). Other possible constraints include inadequate fish passage at the Barrowtown Pump Station, dredging of bed sediment, and siltation and pollution in Marshall Creek due to highway construction and industrial development. The industrial and residential development within the Marshall Creek watershed also increases the effective impervious area, which may alter the hydrologic regime (Klassen et al., 1995). The Sumas River mainstem is described as "heavily angled" (DFO, 1995) and listed fourth in priority for sea-run cutthroat enhancement in a study of the lower mainland and Sechelt Peninsula (De Leeuw, 1981). This study attributed limitations on fish production to lack of quality rearing habitat and inadequate adult escapement. Improvements to the Barrowtown Pump Station by 1984 have decreased the fish mortality through the pumps (District of Abbotsford, June 1993), yet rearing habitat continues to be reported as poor in much of the river system due to slough-like flows and degraded water quality conditions (ESL and Webb, 1987). Wildlife resources within the Sumas River watershed have experienced dramatic habitat changes over this century. At the turn of the century, millions of ducks and geese 51 enjoyed the Sumas Lake and over 8000 ha of marginal land and sloughs. The drained and farmed Sumas lowlands of today have a reduced capacity to attract and hold waterfowl (ESL and Webb, 1987). Bird species which do frequent the remaining river and marsh areas include eagles and marsh hawks, migrating ducks and geese, and migrating and over-wintering swans. Turkey vultures and grouse nest on Sumas mountain. Sumas mountain also supports wildlife populations of blacktail deer, black bear, coyote, bobcat, racoon, and, unique to British Columbia, the mountain beaver. Cougar have also been sighted in the past (Teskey, 1990) and muskrat and wild mink are abundant along the Sumas ditches and canals (ESL and Webb, 1987). Urban development on Sumas mountain signifies a permanent loss of habitat to most of the wildlife species. In addition, the use of developed areas by many wildlife species is often incompatible with human use of the area. All types of development, forestry, agriculture or urban, has the potential to degrade the aquatic habitat for both wildlife resources and fisheries. The degradation can be due to: streambank vegetation removal; instream works; alteration of the hydrological regime due to removal of trees and vegetative cover in the watershed; erosion due to increased flood flows as part of the altered hydrological regime; siltation from construction works; urban stormwater pollution; and, contamination from septic systems, animal manure, pesticides and fertilizers (Teskey, 1990). As fish productivity and wildlife use are among the most responsive and strongest indications of the health of an ecosystem, it is essential that distress signals given by these important natural resources are heeded in watershed management decisions. 52 4.0 Water Quality Investigation 4.1 Water Quality Indicators The choice of water quality indicators used to characterize a water resource can be based on whether it is used for sustaining aquatic life and fisheries, drinking water, recreation and health, and/or irrigation and livestock watering. To completely characterize the water resource and its implications to all of these functions, with all the possible measures of biota, sediment and the water column, would be costly and impracticable. Therefore, easily measured indicators are often chosen which will reflect the general environmental condition and target suspected anthropogenic stresses on the aquatic resource. The availability of historic data for comparison also influences the choice of indicators. This section describes what indicators were measured, typical sources, and why they are of interest. 4.1.1 Trace Metals in River Sediments Some metals are required in trace concentrations by living organisms for normal physiological function and the regulation of many biochemical processes (Chapman, 1992). However, most trace metals are of concern when they reach higher concentrations because of their potential to become toxic, to become bioavailable and bioaccumulate within organisms, and because they do not degrade. The toxicity of metals in solution depends upon many factors such as their degree of oxidation and speciation, as well as their total concentrations. Although different trace metals behave very differently in accumulation and transport mechanisms (Moldan and Cerny, 1994), the conditions which generally cause a release of metal ions into solution are low pH (acidification) and low redox potential (anoxic or reducing conditions). 53 Because sediments accumulate and act as a sink for trace metals, they are a common medium to assess metal pollution in aquatic environments (de Groot, 1982). Their assessment is also important to protect aquatic ecosystems, as many benthic and epibenthic organisms may be exposed to chemicals through their contact with bed sediments (CCME, 1995). The U.S. E P A has identified arsenic, cadmium, copper, chromium, mercury, nickel, lead and zinc as the eight top priority metals of environmental concern (Chapman, 1992). Of the eight metals listed, cadmium, copper, chromium, nickel, lead and zinc were measured in the sediments collected from the Sumas riverbed. These metals may enter the aquatic environment through weathering and erosion of natural geologic components within the river catchment, and by inputs from human activities. The largest anthropogenic sources are sewage, industrial wastewater and mining discharges, and atmospheric deposition resulting from smelting and the burning of fossil fuels. Anthropogenic sources of chromium, lead, nickel and zinc are mainly from industrial activities, such as metal plating or cement manufacturing in the case of chromium. Mining, smelting and combustion of fossil fuels are sources of lead and nickel, and the manufacturing of some foods is also a source of nickel. Zinc, iron, and steel production, wood combustion and waste incineration are all potential sources of zinc. Zinc is also a required nutrient in animal feed (CCREM, 1987). Accumulation from diffuse sources, such as street runoff (including wear materials from autobodies and tires as well as exhaust products), fertilizers, sludges, pesticides, and animal feed, may also cause significant metal enrichment of soils and sediment. Sutton et al. (1983) found that supplemental CuS0 4 in swine diets increased Cu levels in manure spread on the soil. Increased levels of Cu were measured in the top 31cm of the soil. In the Sumas 54 River watershed, a natural landslide of serpentinitic material is known to contribute sediment containing high levels of chromium and nickel (Schreier, 1987). Cadmium and copper may both potentially be introduced through agricultural activities, and zinc may be selectively added to certain crops and animal rations as a micronutrient. There are no apparent sources of lead, particularly with the decline in the use of leaded gasoline. It has been stressed that sediment quality must be evaluated in conjunction with natural background concentrations of substances, and that natural levels themselves may have adverse biological effects. A detailed regional assessment of sediment quality, including intensive sampling at a number of uncontaminated sites has been suggested to determine ambient conditions and the contribution of natural processes (CCME, 1995). In a review of natural background levels of metals in rocks and sediments, including the Vancouver, B.C. area and Western U.S. sediments (see Table 4.2 in Section 4.4.1), and a review of the sources of natural variability in sediment analysis, including effects of particle size distribution and organic content, Cook (1994) states that the concentration of trace metals in sediments is limited in its usefulness as an early warning indicator of anthropogenic stress. A very high degree of enrichment is needed to indicate human influence beyond the high natural variability in trace metal concentrations, and suitable background levels are difficult to obtain. In spite of the above, sediments can be evaluated to determine spatial trends, and to prioritize and focus potential future research activities. Results can also be compared with historically collected data to identify long term trends. 4.1.2 Surface Water Indicators Several parameters were chosen in this study to provide an indication of surface water 55 quality. These included conductivity, pH, temperature, dissolved oxygen, chloride, orthophosphate, ammonia, nitrate, organic carbon and fecal coliforms. Conductivity, or specific conductance, is directly related to the concentration of total dissolved solids and major ions in the water, and thus is used as a surrogate measurement for this concentration. The pH is a master variable that influences all biological and chemical processes within a water body. The pH itself is influenced by industrial effluents and atmospheric deposition, and by photosynthesis and respiration cycles. Temperature affects biological activity in a water body which in turn affects the water chemistry. Temperature and pH affect the toxicity of other subtances, such as ammonia, and high temperatures cause a decrease in dissolved oxygen concentrations. Dissolved oxygen (DO) is easily measured and essential to all forms of aquatic life. The concentration of dissolved oxygen in surface water is a general indication of the degree of pollution by degradable organic matter (Chapman, 1992). The chloride ion is not toxic to humans but high concentrations can make water unpalatable for drinking or unfit for livestock watering, cause corrosion in metal pipes and kills many types of plants (Stednick, 1991; Chapman, 1992). Major sources of chloride are weathering of igneous rocks and sedimentary salt deposits, atmospheric dispersion of sea salts, volcanic gases and hot springs (Hem, 1985; Chapman, 1992). Higher concentrations in some areas are caused by salt water intrusions, industrial and sewage effluents, salting of roads, and irrigation drainage. Potassium chloride is used intensively in fertilizers and also in the manufacturing of insecticides (CCREM, 1987). The circulation of chloride ions in the hydrological cycle is mainly by physical processes, and thus chloride is often used to calculate water balances or as a tracer to indicate human or animal pollution. 56 Phosphorus is an essential nutrient for living organisms, and is often the limiting nutrient which controls primary productivity in freshwater aquatic ecosystems (Waite, 1984). Phosphorus naturally enters aquatic systems through the weathering of phosphorus bearing rocks and the decomposition of organic material. Anthropogenic sources are mainly domestic wastewaters, industrial effluents, and fertilizer runoff. Phosphorus (P) exists in surface waters as both dissolved and particulate species, organic and inorganic. The species measured directly in this study is the dissolved, inorganic orthophosphate (ortho-P) because it is the form of P which is bioavailable. Because ortho-P is readily taken up by plants, its concentrations are usually low in freshwaters. High concentrations indicate the presence of pollution and is responsible for eutrophication (Chapman, 1992; Sharpley et al., 1994). Unpolluted waters contain small amounts of ammoniacal nitrogen (ammonia-N), and higher concentrations usually indicate organic pollution from domestic sewage, industrial waste, or agricultural runoff. Commercial fertilizers contain highly soluble ammonia and ammonium salts, and transport via the atmosphere or irrigation waters occurs when concentrations exceed the immediate plant requirements (CCREM, 1987). Natural seasonal fluctuations of ammonia concentrations occur with the death and decay of phytoplankton and bacteria. This is very pronounced in nutrient rich waters (Chapman, 1992). Ammoniacal nitrogen, also described as total ammonia, exists in two forms: as the unionized molecule (NH 3) and as the ammonium ion (NH 4 +). The equilibrium between the two are determined principally by pH and temperature. High pH and high temperatures favours the unionized form, which is appreciably more toxic. There is a pronounced changeover from ammonium to ammonia as the pH rises from 7.0 to 8.0 (Ellis, 1989). Tables 57 have been developed giving undissociated ammonia levels for varying total ammonia concentrations, pH levels, and temperatures for the Fraser River (Drinnan and Clark, 1980), and giving maximum criteria levels for total ammonia based on pH and temperature (BCMOELP, 1994). This study measures total ammonia-N in the Sumas River, and uses the latter reference table to compare with provincial criteria. The nitrification of ammonia leads to the oxidized nitrite form of nitrogen, and further oxidation produces the nitrate ion. Because the first oxidation is the rate limiting step of the reaction, nitrite is rarely found in appreciable concentrations in surface waters (Ellis, 1989). In anaerobic environments, nitrate may be biochemically reduced to nitrite and eventually to nitrogen gas. This process is called denitrification. The laboratory procedure used in this study measured total nitrite-N plus nitrate-N; however; the concentrations found were assumed to be solely in the nitrate form, and are expressed as such. Nitrate is a highly mobile ion because it is chemically unreactive in dilute aqueous solutions and its common salts are soluble in water. This makes it a good early warning indicator of contamination, as it is commonly introduced into the environment through anthropogenic sources such as municipal and industrial wastewaters, septic tanks, and feedlot discharges. The leaching of inorganic nitrate fertilizers through soils in suburban and rural areas is also known to contribute nitrate to streamwater, with concentrations generally higher in the spring and early summer months (CCREM, 1987). The use of nitrogen fertilizers and the discharge of wastewaters from intensive indoor rearing of livestock can be the most significant sources of nitrate in regions with intensive agriculture (Chapman, 1992). The actual concentrations depend on a variety of factors, including time of ploughing, soils, 58 fertilizer application rates, proportion and quality of groundwater input versus runoff input in the stream, and biological transformations. Nitrate is a concern in drinking water, as concentrations exceeding 10 mg L" 1 are known to cause methaemoglobinaemia (or "blue baby syndrome"or infant cyanosis), a potentially fatal condition. Carcinogenic compounds are also suspected of being formed from nitrates (Ellis, 1989). Excess applications of nitrogen affect the health of soils and waters also. When there is a lack of plant uptake or microbial immobilization, application of excess nitrogen enhances direct leaching and the nitrogen mineralization capacity of the soil, releasing nitrate from soil organic matter as well as fertilizer. In the long term, these factors may contribute to the depletion of soil fertility, increased soil acidity, and acidification and eutrophication of surface waters (Moldan and Cerny, 1994). Organic carbon can act as a surrogate measure for biochemical oxygen demand (BOD), although the degree of pollution measured is less than the latter because of the exclusion of the oxygen consuming reactions of other elements. BOD itself is a measure of the oxygen required by a water sample for aerobic micro-organisms to oxidize organic matter to a stable, inorganic form. The measurement of organic carbon is a reliable, quicker method which provides an approximation to the oxygen demand and degree of pollution (Ellis, 1989). Fecal coliforms are the microbiological indicators normally measured to indicate the presence of pathogens in water, and therefore they indicate a risk to users for drinking, food preparation, irrigation, and recreation. Waters are contaminated by careless spraying of manures, runoff, and use of water by domestic livestock and wildlife. Any presence of these organisms indicate recent fecal contamination from warm-blooded animals. Fecal streptococci 59 are measured to provide information on the nature of the source, as there are more fecal streptococci in warm-blooded animals other than humans (Stednick, 1991). 4.2 Sampling Methodology 4.2.1 Sediments Grab samples of surface stream sediments were taken during low flow conditions on September 3, 1993 and August 15, 1994 at the locations shown in Figure 4.1. In 1993, the locations were based on the locations sampled in 1974 by Hall (unpublished data), in order that historical comparisons could be made. In 1994, several stations were dropped to reduce the laboratory cost and effort while maintaining even coverage of the Canadian portion of the watershed, and some water quality sampling locations were added to obtain continuity between the water and sediment sampling. Altogether, 31 unique sites were sampled; 23 common sites, 29 (including 2 in Whatcom County) in 1993, and 25 (including 3 in Whatcom County) in 1994. In both years, the streambed sediments were collected using an aluminum pot attached to a 2.5 m wooden pole. At each site, the top several centimetres of sediment were collected. The samples were stored in an ice-filled cooler in plastic bags and then deep frozen in the laboratory until the time of analysis. Replicate samples were collected at 3 of the 25 sites (stations 116, 502, and 503). These replicates provide a means of calculating within site variability of results. 4.2.2 Surface Water Surface water grab samples were taken monthly or bi-monthly, from March 1994 to 60 Figure 4.1 Water and Sediment Sampling Stations. Sediment samples were not taken at stations 8 and 14. February, 1995 at the locations shown on Figure 4.1. Station 15, located on Kilgard Creek in a forested area on Sumas Mountain, was selected as a control station as little anthropogenic impact was expected in this relatively undeveloped area. On four of the eight sampling days, 3 replicate samples were taken at one station (a different station for each day) in order to obtain a measure of within site sampling and laboratory analysis variability. The samples were stored on ice until analysis the following morning. On three of the sampling days, separate samples were collected in brown glass bottles and sent directly to EVS Environmental Consultants for bacteriological testing. In situ field measurements, including pH, conductivity, dissolved oxygen and temperature, were taken at the time of water sampling. The pH meter, which also measured temperature, was an Orion model 420A. The conductivity meter was a Yellow Springs Instrument Model 33, and the dissolved oxygen meter a Yellow Springs Instrument Model 57. 4.3 Laboratory Analysis 4.3.1 River Bed Sediment Analysis The sediment samples were wet-sieved using distilled water to obtain the <63 //m fraction into acid-washed 1 L glass beakers. The separation of the <63 /u,m fraction includes the clay/silt particles which tend to be the greatest metal accumulators, and is a common practice in sediment studies (Cook, 1994). The <63 /urn fraction was placed in glass beakers and dried at 105 °C until all the water had evaporated. The dried samples were then disaggregated using an agate mortar and pestle, and stored in plastic containers. Approximately 2 g of each sample was weighed out into a 62 crucible. Ten percent of the sample set was replicated so that variability of the analysis method could be calculated based on the results of same-sample measurements. Two certified standard reference materials, MESS-1 and BCSS-1 (National Research Council, Chemistry Standards Program) were included in the sample set so that the accuracy of the analysis method could be determined. The samples were then ignited in a furnace at 850 °C for 6 hours and reweighed, to remove and calculate the organic matter content respectively. The samples were digested using hydrofluoric acid (HF) in a closed vessel (Page et al., 1982). The ratio used was 0.2 g sample to 6 mL HF. Prior to adding the HF (48%), 1 mL of Aqua regia (3 mL HC1 to 1 mL H N 0 3 ) was added to the sample and shaken to decompose any carbonates present and disperse the sample. The samples with HF were then placed on a mechanical shaker for approximately 8 hours. Deionized water (10 mL) and 2 g of boric acid (H3BO3) were then added to the solution. The samples were brought up to 50 mL in volumetric flasks and given to the laboratory technician in the U B C Soils Department for total metal analysis on the inductively coupled plasma spectrometer (Simultaneous ICP-AES Jarrell Ash). The analytical method used for the 1974 collection of samples is described in Hall et al., (1976). The main differences are: 1) the samples were dry sieved for the <177 ptm fraction, 2) the samples were digested with nitric perchloric acid, 3) heat was used to aid the digestion, and 4) the metal concentrations were detected using flame atomic adsorption (AA) spectrophotometry. Colman and Sanzolone (1992) compared effects of dry sieving versus wet seiving, and found the latter to produce higher metal concentrations. The differences, however, were on average less than 10%, which is small relative to the within site variability 63 of this study given in Table 4.1 of section 4.3.3. Although heat was not used in the digestion step of this study, the digestion is nevertheless considered total since there was no indication of undissolved sample after the mechanical shaking (Page et al., 1982). McCallum (1995) carried out a study to compare flame A A and ICP-AES detection techniques. The ICP produced 55 and 59% lower values for lead and copper respectively, and 19% lower values for nickel. There was no significant difference in zinc results. Because an increase in metal concentrations from 1974 to 1994 is the trend of concern, the lower ICP values serve to ensure that the test is conservative, and will possibly offset the increased concentration effect of wet sieving and smaller particle size analysis. To further complement the tests carried out by McCallum (1995), the <180 /um fraction of the sediments collected from the replicate sites were sent to Chemex Laboratories Ltd., North Vancouver, with instructions to digest them as closely as possible to the method described in Hall et al. (1976). Changes to the procedure were desired by Chemex Laboratories. Instead of the nitric perchloric acid mixture of 4:1 concentrated H N 0 3 and 70% HC10 4 , and 30 mL final volume, a 10% HC1 concentration was used and brought up to a 50 mL final volume. Chemex used an ICP for the metal detection, and their results for the replicate site samples are provided in Appendix A. 4.3.2 Water Constituent Analyses Nitrate-N, ammonia-N, chloride and orthophosphate were analysed the day after sampling on a Lachat X Y Z QuikchemAE autoanalyser in the U B C Soils Department laboratory. If the samples were visibly turbid, they were filtered through 41 Whatman ashless paper before analysis. Methods and standards followed the appropriate QuikChem Method 64 No. as outlined by the manufacturer (Lachat Instruments, 1990). Dissolved nitrate+nitrite-N was analyzed using Method No. 12-107-04-1-B. In this method, nitrate and nitrite is passed through a copperized cadmium column which reduces all the nitrate to nitrite. This is then diazotized with sulfanamide and coupled with N-(l-naphthyl)ethylinediamine dihydrochloride. A dye is produced which is read at 520 nm and determines the concentration of nitrate+nitrite-N. Because nitrite is readily oxidized to nitrate in most aquatic environments, the resulting concentrations are assumed to consist of only nitrate-N, or N 0 3 - N , throughout this thesis. Method No. 12-115-01-1-A was used to determine dissolved orthophosphate concentrations. This ion is also colorimetrically determined, but at 660 nm following a reaction with ammonium molybdate and antimony potassium tartrate, under acidic conditions, and then a reduction with ascorbic acid (LaChat Instruments, 1990). Dissolved ammonia concentrations were measured using Method No. 10-107-06-2-D. The samples are digested in sulfuric acid and then the ammonia is converted to the ammonium ion using a mercuric oxide catalyst. A concentrated buffer is added to raise the pH to a known basic level which converts the ion to ammonia. The sample is then heated with salicylate and hypochlorite and colorimetrically determined at 660 nm (LaChat Instruments, 1990). Method 10-117-07-1-A was used to determine chloride concentrations. Mercuric thiocyanate is reacted with the chloride, displacing the thiocyanate. This reacts with aqueous iron (III) to produce hexacyanoferrate (III). The resultant absorbance of the compound is 65 measured at 480 nm to determine the concentration of dissolved chloride (LaChat Instruments, 1990). Samples for dissolved organic carbon analysis were filtered through 41 Whatman ashless filter paper and kept frozen until analysis. Dissolved organic carbon was then analyzed using a Shimadzu (TOC-500) Total Organic Carbon Analyzer by the U B C Civil Environmental Engineering laboratory. The concentration of dissolved organic carbon was calculated as the difference between total dissolved carbon and dissolved inorganic carbon, which are measured by the analyzer. E V S Environmental Consultants performed the microbial analysis using the membrane filtration method, according to procedures described in "Standard Methods for the Examination of Water and Wastewater, 18th ed., 1992, APHA". 4.3.3 Quality Analysis and Quality Control Variability within a site, or caused by the analytical methodology, was determined by using the coefficient of variance (CV) for the site replications. The results for the site replications are given in Appendix A for sediments and Appendix D for water, and are summarized by average and maximum coefficients of variance in Table 4.1 below. The results of cadmium and lead in sediments are not included in this analysis for reasons described in Section 4.4.1. In addition, for sediments, measurements of trace metal concentrations in certified standard reference materials provide an accuracy range for analytical results, and measurements of sample replications in the lab provide for analytical variability. See Appendix B for the details of these results. The percentage deviation from the certified values 66 of the standard references ranged from 0% for nickel to 81% for copper. Copper introduced the highest deviation from certified values; the average deviation excluding the copper results Table 4.1 Variability in Water and Sediment Site Replications Water (mg L') Nitrate-N Ortho-P Chloride Ammonia-N Organic Carbon Average coefficient of variance, CV (%) 5 14 5 14 62 Maximum CV (%) 9 25 7 25 126 Sediment (mg kg1) Cr Cu Ni Zn Average CV (%) 17 29 19 9 Maximum CV (%) 19 34 25 13 was 10%. Copper also showed the highest variability in analytical results. The average percentage difference between sample replicates for copper was 34%, while the averages of all the other metals ranged from 6% for zinc to 10% for chromium. The recovery for copper in the sediment analysis was not satisfactory. Although the exact cause was not investigated, potential sources of error include interference or matrix effects due to the unusual asbestos material present. Because of the low confidence in the copper results, indications of copper trends are not stressed or elaborated upon for the remainder of this thesis. Accuracy for water analysis was determined by the Soils Department laboratory by measuring samples with known concentrations, or standards, with the autoanalyser for each sampling run. The results obtained from the laboratory are given in Table 4.2 below. 67 Table 4.2 Accuracy Measurements for Water Analysis Constituent range of % difference between known and measured (n) average difference per record (%) standard deviation chloride -15.8 to 13.1 (6) 0.86 9.5 ammonia-N -89.8 to 11 (9) -7.57 31.1 nitrate-N -25.0 to 9.5 (6) -7.25 11.9 Ortho-P -10.0 to 36.7 (9) 9.23 15.6 4.4 Spatial and Temporal Trends Shown by Water Quality Indicators Factors influencing the quality of water at a given sampling station include: proportion of surface runoff and groundwater, reactions within the river system governed by internal processes, mixing of water from tributaries of different quality (more apparent in heterogeneous basins), and inputs of pollutants (Chapman, 1992). These factors manifest themselves by spatial and temporal trends shown in the sampling results. The statistical measures and plots presented in the following sections and respective appendices were calculated using the SPSS for Windows Release 6.1.2. software package. 4.4.1 Trace Metals in River Sediments The range and median of trace metal concentrations measured in the riverbed sediment of the Sumas River and its tributaries is presented in Table 4.3. These are compared with examples of background concentrations of these trace elements found in sediments for other studies, and with preliminary guidelines and criteria for sediment quality compiled by Hall (1992). The trace metal concentrations for cadmium and lead were below the detection limits of 0.2 mg kg"1 and 3 mg kg"1 respectively, as calculated by McCallum (1995) for the same ICP analyser, and are thus not included in this table. Although McCallum (1995) used a dilution ratio about 1/6 of that used in this study, his detection 68 Table 4.3 Comparison of Measured Trace Metal Concentrations with Example Background Concentrations and Preliminary Guidelines and Criteria Element (mg kg"1) Cr Cu Ni Zn 1993,1994 Sumas Sediment Sampling Median Range 120 119 45-371 51-353 41 38 12-117 8-79 149 171 42-1930 54-1886 154 133 29-300 32-276 Background Concentrations* world surficial continental rocks0 71 River Suspended Sediments: Mackenzie River 8.5 Yukon River 115 world average 100 world rivers0 120 32 42 416 100 50 49 22 90 80 127 126 350 240 Streambed Sediments. Rhine River 47 Illinois River, median 56 (90th percentile) (74) Western U.S., range 20-210 Vancouver region, mean and 48 range 10-1000 51 23 (35) 0-110 26 2-415 46 26 (35) 7 1-165 115 100 (241) 49-510 48 10-1000 Guidelines and Criteria11 USEPA for Great Lakes Harbours: non-polluted <25 moderately polluted 25-75 heavily polluted >75 OntarioProvincial Guidelines and British Columbitf1 Criteria: lowest effect 26 severe effect 110 Guidelines for dredged material 25 Wisconsin Criteria: 100 <25 25-50 >50 <20 20-25 >50 <90 90-200 >200 16 110 25 100 16 75 25 T66" 120 820 100 100 a. compiled by Cook (1994) from various sources, unless otherwise noted b. compiled by Hall (1992) from various sources, unless otherwise noted c. compiled from various sources in Chapman (1992) d. BCMOELP(1994) 69 limits are sufficiently low to assume that even if the detection limits for this study were approximately 6 times greater (i.e., a near linear relationship of matrix effects), the concentrations of cadmium and lead in the Sumas River sediments would still be too low to discern spatial trends and are not sufficiently elevated to cause concern. The trace metal concentrations, although higher than the preliminary guidelines presented, are not above the natural background range given for Vancouver region sediments. The exception is nickel, whose median falls within natural levels but whose range includes very high concentrations. However, the source is known to be a natural geologic deposit exposed by a landslide (Schreier, 1987). The range in natural levels found and the differences in criteria emphasize the importance of evaluating trace metal enrichment on a site specific basis, with a knowledge of the local geology. Spatial Trends A table of results for each site is given in Appendix C. When plotted in an upstream-downstream direction as shown in Figure 4.2, the nickel and chromium results show very similar trends. Both metals show low concentrations (station 506) prior to the confluence with Swift Creek (station 501). These upstream values and trends are consistent with studies that researched the stream sediment effects from the landslide in Whatcom County (Schreier, 1987; Schreier et al., 1987; Schreier, 1986). The zinc values are relatively consistent from upstream to downstream, with an anomaly at station 116 for the 1993 sampling; this is assumed to be an outlier due to sampling or analytical error. The copper values show no apparent trend and are less consistent from 1993 to 1994. However, the values are consistently within a natural range of 10 to 50 mg kg"1. Both zinc and copper exhibit peaks 70 Figure 4.2 Stream Sediment Trace Metal Concentrations 506 501 116 115 114 502 134 503 123 128 1 35 mainstem (upstream -> downstream) Sumas Lake Canal Station j — 1993 506 501 116 115 114 502 134 503 123 128 135 mainstem (upstream -> downstream) Sumas Lake Canal Station 71 Figure 4.2 (continued) Stream Sediment Trace Metal Concentrations 2000 -i , J Swift Creek 1000 H 2000 - i CT> CD E, c o -•—• c Q) O c o O 1000 4. * 1—l—I I I I I t... +.. o O) co r^cyjf) <M?>S i O S J o c o c o ^ i j > c p f « . r -cv co co ro cv cv ro Tributary Stations heavily polluted US EPA 506 501 116 115 114 mainstem (upstream -> downstream) 502 1 34 503 Station 123 128 135 Sumas Lake Canal 72 at stations 113 and 122. These do not appear to be in error since the peaks are shown both in 1993 and in 1994. The peaks may be due to road runoff, or from a point source such as a nearby house or farm. More sampling would be required to determine the possible natural or anthropogenic sources. The four metals can be divided into two behavioural categories: 1) landslide influenced, comprising nickel and chromium and exhibiting a strong downstream trend, and 2) non-landslide influenced, comprising copper and zinc which have no apparent natural source but exhibit similar spatial patterns nonetheless. This can be seen in scatter plots of N i vs. Zn or one metal vs. the other (Figure 4.3). In all the scatter plots, similar patterns emerged of the clusters of sites which have high nickel and chromium values with low zinc and copper values, or high zinc and copper with low nickel and chromium values. The highest Ni/Cr, lowest Cu/Zn cluster occurred at Swift Creek and its confluence with the Sumas River. The influence of the landslide is shown further downstream by the next two clusters, comprising stations 114, 115, and 116 for one cluster, and 502 and 506 for the other. It is interesting to note that, not only do these clusters represent sites with the highest Ni and Cr sediment concentrations, but these sites also happen to have among the lowest concentrations of Cu and Zn. This is because the parent serpentinic material from the landslide has lower concentrations of Cu and Zn than other soils in the area (Schreier, 1987). On the opposite side, the cluster which consistently showed high values of Cu and Zn and low values of N i and Cr was comprised mainly of stations 133, 113, 122 and 121. These sites represent sediments in an irrigation ditch, thus suggesting agricultural inputs, at the west 73 •sr co en a> CD CD to k-0) w _D 0 lo CL to 1 CO ui -*—» c CD E T 3 CD CO c o CO I— c CD o c o O c c N N -o g a Z O o o O "O a: § <- o CD £ -£ CD 8 CD CO £ co CD i— O ) i i . c o CO 8 8 cz N (6>|/6lu) uonejjuaouoo i n 74 end of No.4 Rd., Marshall Creek at Angus Campbell Rd. mentioned earlier, and upper Arnold Creek, respectively. Temporal Trends Figure 4.4 is a box plot comparing the results of the 1993 and 1994 sampling for the 23 common sites only. The box plot shows the median, 25th and 75th percentiles, and the largest and smallest observed values that are not outliers or extremes by the extended lines from the box. The apparent similarity of results from 1993 to 1994 are confirmed by the non-parametric Mann-Whitney U test for two independent samples. The two sample sets were found to be not significantly different. This confirms that low flow sediment sampling is very stable, and is a useful basis of comparison for long term trends. Having determined that the trace metal concentrations of the 1993 and 1994 sediment samples are not significantly different, the two sample sets were combined into one and tested, using the Mann-Whitney U test, against the data set from 1974 (Hall, unpublished data). The concentrations of copper, nickel and zinc were found to be significantly higher in 1993/94 than in 1974 ( <0.0T). Chromium concentrations were not available from 1974. The comparison of these data sets is questionable due to the differences in analytical methods as described in Section 4.3.1. The results of the digestion of the <180 /um fraction of the replicate samples by Chemex Laboratories (Appendix A) differ greatly from the <63 jum fraction in two of the three replicate sites. This is likely due to the effect of particle size distribution. The results for the site with much finer sediment (station 116 at the border) differed by less than 12% in either direction for all metals. The results for the other two 75 c to to o CD q= 2 c CO CD > CO iS o co c: T 3 w s i CO CD II If to O CO - c o ^ O 0) DO CO g cS CO CD CD - t (V X CD g l C L co E {2 CO J IK O to ^ 0 - C D E <o CO -g CD C L E co 8.5 CO "D C co Co ® co r- £ O) n c CO C CD T - CO CD o £ I C L O CD o CM -9 CO CD CD C L C 2 E Z ^ O £ ^ co O ) CO O ) CO (6>J/6LU) UOIJBJJU90UOQ CD i_ =5 CD 76 replicate samples, both taken at parks near sandy beaches, tended to be on average 40% lower in metal concentrations for the <180 pirn fraction of sediments. A sensitivity test was performed by testing the 1974 data against the 1993/94 data reduced by 30, 40 and 50%, again using the Mann-Whitney U test for two independent samples. Because there is uncertainty associated with the different digestion methods between the two periods, and other sources of error such as within site variability and instrumental error, it is difficult to ascertain whether the increase in concentrations are real. In general, the nitric perchloric digestion is more rigorous than the HF digestion, which would result in lower concentrations being measured with the HF digestion for the 1994/95 sediments. The larger fraction size measured in 1974 (i.e., <177 jum fraction vs. <63 fxm fraction in 1994/95) should bias the results in the other direction, or lower results for the 1974 sediments, since metals tend to associate with the smaller particles in sediments. However, these statements are dependent on many factors including the actual resultant matrix of the solution measured and the organic composition of the sediments. Quantifying the total error would be onerous and difficult as the older digestion techniques are no longer readily available. The sensitivity analysis provides a simple comparison of results assuming different levels of error, and if a significant difference is found when the error is assumed to be large (i.e., 50%) then the change in concentration is believed, in this thesis, to be real. The results of the sensitivity analysis are shown in Table 4.4 below. 77 Table 4.4 Tests of 1974 and 1993/94 Trace Metal Concentrations in Sediments 1993/94 data reduced by: 30 40 50 MetaU Cu n.s. n.s. § Ni * n.s. n.s. Zn * * * n.s. = data sets not significantly different * = 1993/94 concentrations significantly higher (aO.Ol) § = 1974 data set significantly higher (aO.01) There is no significant difference between the two sample sets for copper, until the 1994 data set is reduced by 50%, which makes the 1974 concentrations significantly higher. Nickel is significantly higher in 1993/94 when the data is reduced by 30%, but further reductions show no significant difference. The significance maintained at a 30% reduction may indicate that the landslide material has continued to travel downstream from the headwaters. Although the enriched nickel sediment appears to have stayed within the main channel, comparison of nickel levels in the Sumas irrigation waterways or tributary streams with nickel levels in other lower Fraser streams (Hall et al., 1976) indicates that the whole Sumas watershed is enriched with nickel. Zinc remains significantly higher in 1993/94 with even a 50% reduction, leading to the postulation that there has been significant enrichment of zinc in the Sumas streambed sediments over the past twenty years. Figure 4.5 shows the change in concentrations, without reductions, between the two time periods. 4.4.2 Surface Water Constituents Results and field measurements of the water quality sampling are given in Appendix D. These are summarized in Table 4.5 by the median of all sites, and compared with natural background levels and British Columbia water quality criteria. The B.C. criteria are based on 78 j5 03 co cz CO CD E o j | to b 2 o o ra CL C X T3 O 3 CQ o X to QJ c w o ® | | T3 > CD M _ CO O C CD — CD C CO CO c o '-£3 CD 2 - c C "O CD C O CO O co" O a> J8 1 a3 CD 8 k * ~ T3 " D C c CO Co N - CM CO ^ " h - CO CO CO CO CO o £ E z j S I o o o o o o CO o o CM o o (6>|/DUl) U0IJBJJU80U0Q LO CD i— =3 CD 79 Dissolved Organic Carbon (mg L'1 ) o 00 i — ( N 0.01-26 i ^ O S o r- ^ °. -6 O C 0.005-0.02b Ammonia-N, Total (mgL"1 ) Os CN O <i -A o c O.001-0.49 typicallyO. 1 1?? a e in o\ * . ^ os § g ^ CN - a & <N 1 1 s—' oj CN CN 4+ ++ i in o r~- r--p- — 0 - in o o o — —; —: 8 H-I i oo SO r--' -o — e 0.002-6.6 o CN © o 200 (max) < 40 (ave) o o o Chloride (mgL"1 ) 13.6 1.2- 53.5 1-10 00 o in CN 100-700 Conductivity (uS cm"1 ) 212 35-400 .0 o o o 1 o 700-5000 1400-4200 Dissolved Oxygen (mgL"1 ) 8 1.5-13.9 nd-18.4 typically>10 >3-ll (salmonid)* >3-6.5 (non-salmonid) CN pH (log scale) 7.6 6.5- 9.4 6.0-8.5b 6.5-8.5 site specific 5.0-9.0 5.0-9.5 5.0-9.0 Indicator Measured (all sites) median range Natural levels in freshwatersd world river average Criteria" Drinking Water Aquatic Life (freshwater) Irrigation* Livestock * Recreation S3 .3 m a, T 3 a n N C3 t q S 4> o •a •a 00 T3 •a a S3 O •a s -a -a <u w ir-es T3 53 3 CB „ _ a s s o o ^ <u <u &> « Crt o s S 4— ++• T3 c 1) T3 O C 1) OS W 2 S 2 os to r-os <J oo -a 2) J3 5 U u 2 o II II II 80 the Canadian Water Quality Guidelines (CCREM, 1987), and so only the provincial criteria are given since they are largely the same. No criteria levels have been proposed for ortho-P nor for organic carbon. Fecal coliform results will be discussed later on in this section. The medians of the other indicators generally fall below criteria levels, and within natural ranges. However, the use of medians or averages masks the occurrence of more critical levels at specific locations or during different times of the year. This is suggested by the range of values seen, but is more properly addressed by the box plots in Appendix E. These box plots show the range of values of each indicator for each site over the sampling dates, and for each sampling date over all sites. Examples are provided for ortho-P by date and dissolved oxygen by station in Figure 4.6. These figures illustrate well how trends can be observed for parameters both by time of year and by site. The orthophosphate values are greater in high flow winter periods and dissolved oxygen values tend to be consistently lower at stations 9, 10, and 16. In addition, spatial and seasonal variability may be more clearly shown by the series of figures in Appendix F. In these figures, averages of sampling dates taken during the high flow period of November through March, denoted as "wet season", and averages taken during the low flow period of June to August, denoted as "dry season", are plotted for the Sumas mainstem in an upstream to downstream direction. The average values for the tributaries are shown as inset plots. Careful examination of both of these series of plots allows for a characterization of the seasonal and spatial trends shown by each indicator. 81 Figure 4.6 Box Plots of Ortho-P by Sampling Date and Dissolved Oxygen by Station. Outliers and extremes are labelled by site or date, and shown by circles and stars respectively. Ortho-P value for October 3, 1994 on Saar Ck. falls outside of axis range. N = 13 16 16 16 15 15 16 02-MAR-1994 29-JUN-1994 24-AUG-1994 24-NOV-1994 09-MAY-1994 26-JUL-1994 03-OCT-1994 01-FEB-1995 16 14 DATE g 12 § 10 O ) ! 8 (U 6 > o to 4 to H OOSMftY-1994 C0&JWY-1994 T . T T r i p u? I T 023JUN-1994 I 024-AUG-1994 O01-FEB-1995 | S^ vS^  — I — OB-CCEJ394 O03-CCT-1994 N= 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10 11 12 13 14 15 16 STATION 82 Seasonal Trends The most noticeable trends in both the box plots and the wet season/dry season plots are the consistently higher values of ortho-P, nitrate-N and ammonia-N across all sites in the wet season, as shown in Figure 4.7. During this period, the values of conductivity and pH are consistently lower. The anomalies of these trends are: • Nitrate-N levels in Marshall Creek (stations 13 and 5). High levels occur in the dry season and lower levels in the wet season. Marshall Creek will be discussed further in the Spatial Trends section. • The higher pH levels in the Sumas headwaters near the Swift Creek confluence. The pH values at the Sumas headwaters are dominated by the asbestos sediments in Swift Creek, known to be alkaline in nature (Schreier, 1987). Chloride values exhibit no discernable seasonal trend as shown in Figure 4.7, although the largest range and highest median value as seen on the box plot in Appendix E occurs on February 1, 1995, which was the sampling day of highest discharge when many stations were flooded. Fecal coliform and fecal streptococci counts are much greater on this high flow sampling day across all sites. The values are well above the Canadian Guidelines of a maximium of 200 FC/100 mL for recreational use (CCREM, 1987) or the same value designated by the province for irrigation water used on vegetables/fruit which is eaten raw (IRC, 1994). The 200 FC/100 mL criteria level in both cases applies to the geometric mean of at least 5 samples, which cannot be calculated with the limited measurements taken in this study, but the criteria is given for comparison purposes. In the past, the fecal coliform to fecal streptococci ratio was used as an indicator of the nature of the fecal source, with a low 83 Figure 4.7 Water Quality Plots: Chloride and Ortho-P E <D 20 ;g _o sz O 10 40 -I 20 H 30 - i i i i i i i i i i i i i i 14 1011 Tributary Stations 13 5 15 Control - A — Wet Season (Nov-Mar) - ® — Dry Season (Jun-Aug) F T i—r i—i—r i—r 16 2 3 4 12 6 Sumas mainstem (upstream --> downstream) Sampling Station g 8 Sumas Lake Canal 0.8 n 16 2 3 4 12 6 7 9 8 Sumas mainstem (upstream -> downstream) Sumas Lake Canal Sampling Station 84 Figure 4.7 (continued) Water Quality Plots: Nitrate-N and Ammonia-N 1 16 2 3 4 12 6 7 9 8 Sumas mainstem (upstream -> downstream) Sumas Lake Canal Sampling Station 16 2 3 4 12 6 7 9 8 Sumas mainstem (upstream —> downstream) Sumas Lake Canal Sampling Station 85 ratio indicating a non-human animal source. The use of a ratio, however, is now considered "highly questionable, if not inaccurate" (CCREM, 1987) due to many factors, including differential die-off rates between these two groups. The isolation of the enterococcal species, within the fecal streptococcal group, is now considered more useful in the determination of type, source and degree of fecal contamination. The average concentrations of dissolved oxygen are relatively constant throughout the year. However, very low levels occur at particular sites at certain times of the year, as discussed below. Temperature predictably increases during the summer months, yet it should be noted that temperature values at some sites during these months reach values far above those desired for fisheries purposes (BCMOELP, 1994). Spatial Trends Chloride, ortho-P, nitrate-N, and conductivity all show a tendency during the dry season to increase in concentration from the headwaters to station 3 (at the US/Canada border) followed by a decrease as one moves further downstream. Without knowing the flows at each station, and modelling the physical and biochemical processes diluting or transforming these constituents, it is difficult to surmise the cause of these trends. Other notable spatial trends are indicator-specific. On the three sampling dates for microbes, fecal coliform counts for stations 3 and 4 on the Sumas mainstem are all slightly higher than the 200 FC/100 mL guideline. Values at station 10 on Arnold Slough and station 13 on Marshall Creek are consistently and considerably higher than the guideline, with values typically around 1000 FC/100 mL. Marshall Creek (stations 13 and 5) has consistently the highest nitrate-N values, on several instances approaching the 10 mg L" 1 criteria level. And, 86 unlike all other stations, the nitrate-N values are higher during the dry season than in the wet season, as seen in Figure 4.8. It is believed that this behaviour is caused by significant inputs of groundwater from the contaminated Abbotsford aquifer, both from natural springs and the trout hatchery discharge. In the summer these inputs comprise a greater proportion of streamflow, whereas runoff dilutes these inputs in the winter. Arnold Slough, represented by stations 10 and 11, is also a remarkable tributary. It consistently shows the highest levels of ammonia-N and the lowest pH and dissolved oxygen levels. Fortunately, lower pH pushes the equilibrium of ammoniacal nitrogen towards the less toxic ammonium ion form. However, the pH values are still generally above 7.0, and the temperatures in Arnold Slough can increase substantially in the summer, which would drive the equilibrium in the opposite direction. At pH and temperature values encountered during this sampling survey, the ammonia levels in Arnold Slough are generally below criteria levels, but if higher levels such as that measured at station 10 on July 26, are more chronic than could be detected in this sampling scheme, then the water would be toxic to freshwater aquatic life. This is based on the average 30-day concentration criteria of total ammonia nitrogen (BCMOELP, 1994) for given pH and temperature values. On July 26, 1994 the temperature and pH values at station 10 were 19°C and pH 7.2 respectively, giving an average 30-day concentration criteria of 1.32 mg L " 1 , while the level measured on this day was 1.64 mg L" 1 of ammonia-N. Whether the concentrations are maintained around this level over a 30 day period can not be ascertained in this study, but the prospect does not seem infeasible. Similar arguments apply to other stations which exhibited high ammonia-N levels, particularly in the Sumas Lake Canal. 87 Figure 4.8 Comparison of Nitrate-N Behaviour in Marshall Creek and the Sumas Mainstem. Concentrations are highest and greater in Marshall Creek in the dry season, unlike the remainder of the system which has higher values in the wet season. 88 Dissolved oxygen levels in Arnold Slough are perhaps a greater concern. The median DO values are around 5 mg L" 1 with extreme lows of around 2 mg L" 1 reached in June and November. Continued sampling by Schreier (pers. com.) through 1995, and the sampling conducted by IRC (1994) confirms the low values, particularly in the fall season. Unfortunately, the time of lowest values coincides with the time of migration and spawning of many salmonid species (FREMP, 1990). Provincial objectives for dissolved oxygen levels, although not given for Arnold Slough, are that any discrete sample taken from the Sumas River, Marshall Creek, or Saar Creek, should not be below 11 mg/L during the embryo and larval stages of salmonids, and not below 8 mg/L during other life stages (BCMOELP, 1995). The results in the Sumas River system indicate that various tributaries and reaches have difficulty in meeting this objective. Saar Creek deserves mention due to the results measured at station 14 on October 3, 1994. On this day, unusually high values of ortho-P (2.07 mg L' 1 ) and ammonia-N (4.029 mg L"1) were detected. Both of these indicators were at values well above the values measured at any other station over the whole sampling period. The lowest DO level (1.5 mg L"1) of the sampling period was also measured at Saar Creek on this day. Chloride values were also considerably higher than any other station that day. Fecal coliforms were regrettably not sampled on this day. The elevated levels of so many indicators in concert suggests that sample contamination or analytical error is unlikely. It appears that this anomaly is due to an isolated point source contribution to the stream, possibly from an animal in the stream, as Saar Creek on all other sampling days did not stand out as an area of concern. 89 Finally, the influence of agriculture on water quality can be illustrated by observing the results at the control station on Sumas mountain (station 15) and on Swift Creek (station 1), both of which have little or no agricultural activity occurring in their catchments. The plots in Appendices E and F show consistently lower values at stations 1 and 15 for ammonia-N, conductivity, nitrate-N, and orthophosphate, and higher levels of dissolved oxygen, than at all the other stations. Station 15 also had lower chloride values than other stations throughout the sampling period. Historical Trends Historical water quality data from 1970 to the present was compiled from a variety of sources, including the provincial S E A M and older EQUIS databases, the B C M O E L P Fisheries and Wildlife files for the Sumas River, the Environment Canada Envirodat database, and previous data published on the Sumas River (Whelan et al., 1986; Schreier, 1986 and 1987; E S L and Webb, 1987; Hall et al., 1974; Benedict et al., 1973). Historical wet season values (November through March) for all the measurements taken on the Sumas mainstem were plotted for each water quality indicator (see Appendix G). Wet season values were examined since that was the period of concern identified by this study's sampling results. Unfortunately, the scarcity of data, large data gaps, and a wide spread of values for almost all the parameters made it difficult to discern any trends. Also, using only the wet season data does not nearly account for the variation in concentration with discharge. Nitrate-N (see Figure 4.9) showed the clearest trend of an increasing spread of data values, particularly since 1990. Despite the poor relationships indicated in Figure 4.9 and the plots of Appendix G, the direction of the trends which one would expect with the gradual deterioration of water 90 CD CO SZ CO r -O CD • o *E co > CD S o CO CL CO CD O . £ «o CO CD CO '4= g S cu . S CO co J2> CO CO CO CD 3 CO CD CD C O Z > £ -co CO o Z .E ° -Q CO n CD 2 IS C O o sz to - ~ co to CD — CO o "CD -j= CO o 9> -d CO CO S S e 111 CN CO o d i i c r to o r < LU >-CD lO CO CM en CD L l 91 quality (increasing chloride and nutrient levels, decreasing dissolved oxygen levels and pH) are shown. These trends may be indicative of the increasing types and intensities of land use activities occurring in the watershed. Variation with Discharge One weakness in the water quality analysis undertaken in this thesis is the lack of consideration given to discharge and its influence on water quality. During flood periods, water entering a stream has different origins (surface and subsurface runoff, and groundwater) which produces marked variations in water quality (Chapman, 1992). Some of the variation due to discharge was removed with the separation of data into wet and dry season categories. This helped to identify seasonal variation in the 1994-95 sampling period, yet was not sufficient to separate the discharge effects when searching for historical trends. The influence of discharge on concentration is not simply a dilution effect, but is related to sheet erosion and bed remobilization, and the flushing of soil constituents (Chapman, 1992). These in turn are influenced by rainfall intensity and duration, and rainfall patterns prior to the sampling dates. The latter factors influence the quality and quantity of runoff into the stream. Furthermore, when time is included with the relationship between sediment transport and discharge, a hysteris loop is often observed (Chapman, 1992). This means that there may be more than one value of concentration for the same flow magnitude, creating a large spread of values in the scatter diagram. Despite these suspected complications, a simple least squares regression curve was attempted for each indicator using historical flow measured at the border hydrometric station, and the historical water quality data at the nearest stations (3 and 4). The scatterplots with these regressions are presented in Appendix H. The strongest 92 relationship with discharge is shown with ammonia-N and ortho-P concentrations, both of which increase exponentially with increased discharge. Determining the relationship of water constituent concentrations with discharge is inhibited by the lack of historical data, especially results from high flow sampling. Also, the effect of water quality changes over time are excluded in this search for a relationship, as discharge was excluded from the search for long-term trends presented above. Ideally, all water quality data should have a flow associated with them, measured at the time and place of sampling, so that mass loadings can be used to determine relationships instead of concentrations. This type of flow data was not available for the historical data, nor was flow measured at each station as part of the sampling strategy for this thesis, due to a lack of resources and equipment. This seriously limits the interpretation of the data and the potential for water quality modelling. Nevertheless, certain trends, as described in previous sections in this chapter, can be identified. 93 5.0 Land Use - Water Quality Relationships Chapters 2 and 4 presented a picture of the Sumas watershed, including: 1) the types of land use activities and land use trends; 2) spatial, seasonal variability, and historical changes in water quality; and 3) spatial and historical changes in sediment trace metal values. This chapter explores the relationships between land use and water quality characteristics using enumeration areas (EAs) and contributing areas to sampling points as the bases of comparison. Figure 5.1 shows the enumeration area boundaries in relation to the stream network and watershed boundary, and Figure 5.2 shows the delineation of contributing areas to the sampling stations. It is apparent from this latter figure that, due to the human control of drainage in this agroecosystem as described in Chapter 3, the contributing areas appear "unnatural", their boundaries being characterized by straight lines. The absence of topographical relief, together with the action of humans controlling and at times even reversing the natural drainage patterns, made the delineation of contributing areas a considerable challenge in the prairie portion of the watershed. Nevertheless, the overriding drainage pattern to the sampling stations is as shown, and was generally confirmed by Frank Wright, the Superintendent of Dyking, Drainage and Irrigation for the City of Abbotsford. The relationships explored in this chapter were limited by the availability of the data, and the applicability of the data to be used as indicators on the spatial area basis. Only current land and water characteristics as measured during the study period were statistically tested. Also, analysis of the Canadian portion of the Sumas Watershed, where more detailed information is available, is separated from the analysis of the watershed as a whole, where estimations due to differing information formats and level of detail may compromise the 94 95 Figure 5.2 Sampling Stations and Contributing Areas in the Sumas Watershed. results. Because of this separation of information types and availability, some summaries for contributing areas were unavoidably defined or divided by the international border. For the whole watershed analysis the divided areas for contributing area #10 (Arnold Creek) were added together and summarized as one, which resulted in different values for this contributing area than in the Canadian portion analysis. 5.1 Development of Indices Land indices are simply the characteristics described in Chapter 2, namely the amount of land use types, measures of agricultural intensity such as surplus nitrogen loading and animal densities, and the coverage of the surficial geology, based on properties such as texture class, drainage capability and parent material. The characteristics were quantified and summarized on a spatial basis within the watershed to produce land indices that could be related to water quality values at a sampling point. Depending on the nature of the data, the calculations were performed using either database queries or GIS spatial analysis functions, or a combination of both. For example, in the case of surplus nitrogen loading by contributing areas, the nitrogen balance results for each farm as mentioned in Section 2.5.2, were input to the farm database, which is linked to the GIS graphics component, i.e., each farm's geographic location is linked to values calculated for that farm. The overlay function of the GIS was used to determine which farms fell into which contributing areas, allowing the total surplus nitrogen loading for each contributing area to be calculated. Similarly, AUEs for each contributing area were summed. Both nitrogen loading and AUEs were divided by total farm and/or crop land area within the 97 contributing area to express these indices as comparable rates or densities. A major assumption in these calculations is that if a farm building falls within the boundaries of a contributing area, then all values associated with that farm, i.e., land area, crops and animals, also fall within the contributing area. However, because the contributing areas are relatively large in comparison to farm size, the error produced by this assumption should be fairly small. The GIS enables visual presentation of the indices, as shown in Figure 5.3 which illustrates surplus nitrogen loading per farmed hectare by contributing area. For the agricultural contributing areas, the values for surplus nitrogen loading ranged from 57 to 332 kg N/farmed hectare, and 7 out of the 11 contributing areas had surplus nitrogen loadings greater than 100 kg N/farmed hectare. The values for A U E density ranged from 0.4 to 4.5 AUE/farmed hectare and 3 contributing areas had densities near or over the 2.5 AUE/ha average standard. The breakdown of a watershed into smaller contributing areas and the quantification of land use activity by these smaller areas illustrates that while the overall watershed may appear to support the intensity of land use activity, more localized areas can be undergoing much higher stresses and demands, due to the unequal distribution of the activities. As can be seen in Figure 5.3, the highest loading occurs in a contributing area bordering Arnold Slough, where low dissolved oxygen and high ammonia-N values were measured in the stream. Information on agricultural activity in the U.S. portion of the watershed was obtained from the Whatcom County Conservation District, who provided approximate locations of all (65 in total) dairy farms in the watershed, typical herd sizes (from 150 to 1000 head), and a summary of crop acreage for the Sumas and Saar Creek watersheds. The crop acreages were 98 Figure 5.3 Surplus Nitrogen Loading by Contributing Area. apportioned to the contributing areas based on the amount of agricultural land within the contributing area as calculated from the land use information provided by the Whatcom County Planning Department. Locations of other animal operations were not available, although there are reportedly very few (Timblin, C , pers. comm.). Agriculture census data from the USD A for 1992 was obtained, as a check, for the town of Sumas by zip code summary (available on the Internet). Although geographical location is unknown for this data, it confirms the dominance of dairy production in the area listing only 25 small (<50 head) beef, 2 small (<50 head) hog operations, and a few small broiler and layer operations. If these operations do exist within the watershed, they are unaccounted for in the calculations of animal stocking density and surplus nitrogen. The results for nitrogen loading ranged from a deficit of -29 kg N/agricultural ha in the headwaters area to 214 kg N/agricultural ha in the Arnold Creek contributing area. Similarly, the A U E density ranged from 0.2 to 3.7 AUE/agricultural ha. The GIS overlay function was also used to calculate total and percentage area of land use types and properties of the surficial geology by contributing area. A complete list of the indices used for correlation calculations is given in Appendix K, and the values of the indices for each contributing area are given in Appendix L. 5.2 Relationships Between Indices by Enumeration Area and Water Quality Enumeration areas are delineated on the basis of political boundaries and population densities; they have no meaningful relationship to the natural course of water (or even human-controlled drainage) to a particular point in a river. Nevertheless, land use information is most 100 readily available in census format, so some initial exploration of this data is a useful first step towards identifying trends. Figure 5.1 shows the major agricultural enumeration areas shaded by livestock density range. As explained in Chapter 2, approximately 2.5 AUEs per ha signifies a density which may be reaching the absorptive capacity of the soil (Anderson et al., 1990). The EAs having densities above this value are shaded in orange and pink in Figure 5.1. Due to the small number of agricultural EAs, and the difficulty of choosing a sampling station representative of the EA's drainage area, statistical analysis of the indices and water quality values is not meaningful. Relationships were thus identified using graphical and visual techniques. Water quality sampling stations 4, 8, and 11 were chosen to be paired with EAs 201, 117, and 118 respectively (as identified in Figure 5.1) for the comparison of land use and water quality. Figures 5.4 a) and b) illustrate the trends identified. The water quality values in these figures are wet season averages for each sampling site. There is an apparent relationship between cattle numbers and ammonia or orthophosphate in the stream, and between livestock densities or pig numbers, and nitrate levels in the stream. These trends may be a result of different manure handling strategies, or different pathways that nitrate, phosphorus and ammonia enter the aquatic environment. One further complication of using census data is the combining of data in two separate EAs by Statistics Canada when few farms exist in one of them. This, understandably, is to protect the privacy of individual farms, but renders a less accurate spatial distribution of the data. In particular, the data for the large E A on the north side of the Sumas River, where relatively few farms exist, is combined with that of E A 117. For this reason, the area on the 101 Figure 5.4 Trends of Water Quality Values and Animal Numbers in Three EAs 15000 pigs 9 A U E (x 0.001) nitrate-N 3.00 5000 4&201 8&117 11&118 a) Sampling Site & Enumeration Area 2.00 o 1.00 0.00 cattle orthophosphate (mg/L) ammonia-N (mg/L) r- 1.00 8000 - i E _CD *S CO O ° 4000 CD .o E Z 0.50 u c o O -0.00 4&201 8 & 117 1U118 b) Sampling Site & Enumeration Area 102 north side of the river was not included in the density calculation, and the area of E A 117 was determined using the GIS rather than the lumped census figure. The combining of data in rural areas also serves to decrease the potential number of cases in a statistical analysis. Agriculture census data is useful to identify historical and overall land use activity over the watershed, and to corroborate data compiled from other sources, i.e., aerial photographs and the WMS data. With a judicious selection of water sampling stations, tentative relationships to water quality are also shown. However, because of the limitations of using census data for identifying spatial relationships, particularly with the help of statistics, further exploration of relationships based on E A boundaries was abandoned. 5.3 Relationships Between Indices by Contributing Area and Water Quality 5.3.1 Canadian Portion of Watershed The degree of association between the indices calculated for the contributing areas, and wet and dry season averages for selected water quality parameters, were examined using non-parametric Spearman rank correlation coefficients (rs). The resultant correlation coefficients which were greater than 0.5 and had significance levels less than 0.05, are presented in Appendix M . A selection of these relationships, which have interesting implications, are highlighted in Table 5.1 below. Appendix M includes tables showing the significant relationships between water quality variables themselves, and between land indices which pertain to land use activity and the indices based on soil properties. Examination of the correlation matrices lead to the following observations: 1) Dissolved oxygen levels are negatively correlated, and wet season ammonia levels 103 Q O BO w 5? so d SO d CN SO d d so d £1 SO d •a § £ CB So " O £ SI as r-d as oo SO d en so f -d so d 55 u 03 4=1 iz eB § b a g , U CB 1> to so d m so d in f-d d m d as r-d <n SO d 00 d m oo CN r-d cn d CO oo d CN 00 5> so d SI SO d 00 d £1 so d o d §1 so d sS IT) X © CB a a o •a T3 CD g o cu J3 u o o. i CJ CU O O B o to V c B - H J3 a 3 'S O 2 s •R m H cj o .2 •c 9 1/3 cu H u W C4 . a 104 are positively correlated, to surplus nitrogen application rates and amount of finer textured soils within a contributing area. This can be represented by the contextual diagram below. Surplus nitrogen application rates are also positively correlated to amount of finer textured soils, which may mean either that one of these land indices' relationship to the water quality is due to the other, or that the two indices compound each other in affecting water quality. The latter is believed to be the case, as stronger relationships are found in the wet season when greater runoff would occur on fine-textured soils. Figure 5.5 Ammonia, Dissolved Oxygen, Surplus N, and Soils: System of Relationships in Canada Surplus N/ha + ^ + Ammonia-N Dissolved Oxygen Clay/Loam 2) There is a positive relationship between nitrate values and the amount of organic soils, or soils with very poor drainage, in a contributing area. Both of these variables are also positively correlated with the area of land with "no perceived activity", a land use tending to occur in the more industrial areas, and bordering the Abbotsford Aquifer. It is believed therefore that these relationships are strongly influenced by the contribution of contaminated water from the Abbotsford Aquifer. However, it is possible that the organic soils are 105 providing a source of nitrogen which is mineralized and then nitrified to nitrate. Nitrate values are also positively correlated to the percentage of silt in a contributing area, and negatively correlated to the percentage of sand in the area. These two soil variables are not correlated to the "no perceived activity" land use type. From these observations, it can be postulated that nitrates are higher from areas with poor drainage and/or organic material, and lower nitrate values occur in the stream water when the area has sandy texture. Nitrates in the sandy areas may be infiltrating into the subsurface and undergoing denitrification. This may be possible in sandy soils, which are normally considered well aerated, if the surface and subsurface irrigation produce an anaerobic environment, and applied manure acts as a carbon source to the denitrifying bacteria. Alternatively, the nitrates may be entering the stream outside the contributing area due to the various influences on groundwater flows from extensive pumping and drainage. The diagram below summarizes the system of relationships found for nitrate levels. Figure 5.6 Nitrates, Soils and Land Use: System of Relationships in Canada Sand Organic soils/ very poor drainage Nitrate-N + Silt No perceived land use activity 106 Strong relationships with other water quality variables and the land indices were not identified. Because phosphorus tends to be associated with sediment, sampling for total phosphorus over many storm events, with frequent sampling intervals (and thus a modified statistical analysis) may yield relationships with land use. Unfortunately, this is a very labour, equipment and time intensive process to carry out. Other relationships which one might expect to surface may not be presented because of a failure to meet the chosen relationship criteria. This does not mean that they did not nearly meet the criteria. In the statistical analysis of contributing areas in the Canadian portion of the Sumas watershed, no relationships or patterns emerged which conflicted with common sense or contradicted each other. 5.3.2 Whole Watershed Using the soils, land use, and agricultural data collected for Whatcom County, and the water quality results from stations 2 and 16 in Whatcom County and stations 3 and 14 near the border, the non-parametric correlation computation was repeated in a "whole watershed" context. The Spearman rank correlation coefficients for the whole watershed analysis are given in Table 5.2. In general, the relationships found were similar to those of the Canadian portion analysis, but the values of the coefficients were lower. This is likely the result of two major factors. Firstly, the sampling sites in Whatcom County had a poor spatial distribution since the focus was on conditions in the headwaters and monitoring the water quality crossing the border. Secondly, no field verification was conducted in the United States for agricultural activity, and as the census data suggests, some animal operations were missed. Conductivity, specifically in the wet season, is correlated to indices of agricultural 107 4 O Q 8 3 § 60 tn, 5? EH O Q in PI VO OS © VO d m 3 m § 55 i) St, »> O £ vo d 00 m Q VO d 3 VO 00 m ea cn m m CTv VO VO in C v VO O o ^ Si Q 8 in 8 VO in 1 31 VO in d IT) X © •8 d 3 ? g o 1 .8 C3 •5 H .s a .8 51 o o a & J3 SI 1) J3 C 01 8S 108 intensity in the whole watershed analysis. This further confirms the wet season phenomena of contaminated runoff, as more fertilizer and manure entering the stream would increase the salt concentration. Interestingly, the amount of clay in the contributing area as a relevant index disappears, and in its place the amount of gravel appears instead, showing the opposite relationships to ammonia-N, dissolved oxygen and surplus nitrogen as did the amount of clay. Thus Figure 5.5 is transformed to Figure 5.7. Again, the relationships appear in the wet season, suggesting that less contaminated runoff occurs in the coarser textured contributing areas during this time. It should be noted that changes in relationships with soils from the Canadian analysis to the whole watershed analysis are likely due to the different soil classification system used by the United States. Although the same basis for categorization of the soil properties was attempted using the soil descriptions, some inconsistencies were inevitable. Figure 5.7 Ammonia, Dissolved Oxygen, Surplus N, and Soils: System of Relationships in the Whole Watershed Surplus N/ha Gravel With respect to nitrates, the relationship with "no perceived land use" disappears, while the 109 relationships with organics, sand and finer texture material remain. This is represented in the figure below. Figure 5.8 Nitrates and Soils: System of Relationships in the Whole Watershed Sand Nitrate-N Organic soils/ poor drainage Clay The conceptual diagrams for both the Canadian and the whole watershed analysis demonstrate an important issue in the study of non-point source pollution. This is that the influence of land use activity on non-point source pollution cannot be separated from the inherent features of the land, i.e., soils. Although the statistical analysis assumes indices are independent, in reality humans concentrate certain activities over certain soil types; this is necessary for agricultural productivity. In the Sumas Prairie, vegetable production is concentrated on the sandier soils of the old lake bed, resulting in the animal operations tending to develop on the less desirable, finer soils. Although finer textured soils have greater adsorption capacities, the poorer drainage of these soils appears to contribute to more surface runoff. Unfortunately, the combination of animal waste on these finer soils exacerbates the problem of agricultural pollution to the stream during the wet season. In both the Canadian and the whole watershed analysis, nitrate levels in the stream do not appear to be related to the indices of agricultural intensity. In the water quality 110 investigation, nitrate was the only constituent measured that clearly increased in concentration from upstream to downstream (see Figure 4.8). This corresponds with the cumulative increase in the indice values if the contributing areas are added together from upstream to downstream, as illustrated in Figure 5.9. While the method of independently testing contributing areas, regardless of land activity upstream, identifies relationships of land use with ammonia and dissolved oxygen, the method appears to be inappropriate for relationships with nitrate. There may be a sufficient lag time as ammonia-N is nitrified to nitrate-N in the stream, so that nitrate values are less influenced by nearby land use, and more influenced by the cumulative effects of intense activity in the watershed as a whole. Examination of the system dynamics using water quality modelling techniques would be useful in testing this hypothesis. I l l Figure 5.9 Cumulative Indices and Nitrate-N Values. Upstream to Downstream on the Sumas Mainstem. AUE = Animal Unit Equivalents. a) Nitrate-N and Cumulative Surplus N/ha 3 1 6 2 MaWem SteMon 6 7 • Surplus N/ha -Nitrate-N (wet season) • Nitrate-N (dry season) 2.0 "to c 3 . 0 LU D -6.5 0.0 b) Nitrate-N and Cumulative AUE density 3 Mainstem Stal&n 6 2.5 2 1.5 + 1 0.5 0 AUE/ha — Nitrate-N (wet season) - - Nitrate-N (dry season) 112 6.0 Summary and Recommendations Problems related to agricultural waste management in the Sumas River watershed consist of an inadequate land base for manure disposal, low dissolved oxygen and high ammonia levels, and the potential for fish kill conditions. Despite the introduction of Environmental Guidelines for various producer groups (BCMOAFF, 1992 and 1993) and the 1992 Code of Agricultural Practice for Waste Management, IRC (1994) found that farms are operating on average at only 60% of the recommended environmental sustainability level, which was determined from an environmental sustainability parameter (ESP) based on the Guidelines and the Code (Palmer and Rising, 1996). The standard of practices were found to vary widely from farm to farm, as represented by the ESP, with manure storage being identified as one of the most critical factors for environmental sustainability. This thesis aimed to explore relationships between land use activity and water quality in the Sumas River watershed with an emphasis on agricultural non-point source pollution. Links between agricultural intensity, soil characteristics, and water quality were identified. In the process of investigating these links, spatial and temporal trends in land use and water and sediment quality were documented, and a potential watershed management tool was developed through the use of a GIS. Throughout the thesis work, trends and links were considered in context of the biophysical factors which characterize the Sumas River watershed, including flooding recurrences, controlled hydraulics, fisheries resources, and the contamination from a natural landslide. Social and economic factors were not considered, and these have potentially serious and broad implications to water quality and quantity 113 management. This thesis therefore contributes a small but significant piece of the Sumas Sustainability Study, the start of an evolving watershed management puzzle. 6.1 Land Use Summary Compilation and analysis, with the help of a GIS, of aerial photos, census data, Waste Management Survey (IRC, 1994) data, and this study's field work data, has revealed the following land use trends: 1) Residential and industrial/commercial encroachment is occurring onto other land use types, such as forested or lands under speculation; 2) The agricultural land base remains constant, but agricultural intensification is apparent since farm and dwelling numbers continue to grow. The 1954 aerial photos of the Canadian portion of the Sumas watershed revealed 224 farms; 283 farms were identified in the 1994 aerial photos. Within the Canadian agricultural land base, cattle numbers remained constant between 1986 and 1991, while pig and poultry numbers increased by 50 and 75% respectively. The combination of increasing animal densities and flooding problems results in an increasing risk of economic and environmental damage, in terms of evacuation procedures, animal mortalities, and the spread of wastes and pathogenic organisms. The overall animal density for the Canadian portion of the watershed was 2.3 animal unit equivalents (AUE)/farmed ha. On a contributing area basis, the density ranged from 0.4 to 4.5 AUE/farmed ha for the Canadian side and 0.2 to 3.7 AUE/farmed ha on the U.S. side. When calculated on a farm basis, 44% of the farms that participated in the Waste Management Survey had densities above the 2.5 AUE/ha average standard. Annual surplus nitrogen loadings were found to range from 123 to 151 kg N/farmed ha for the overall watershed, 114 compared to the benchmark figure of 100 kg N/ha, and from a deficit of -29 to a surplus of 332 kg N/farmed ha on a contributing area basis. 6.2 Water Quality Summary Sampling of stream sediment for trace metals revealed a general enrichment since the 1970's of zinc levels throughout the watershed, with many areas having concentrations above the lowest effect level given by British Columbia and Ontario criteria. At local sites, including one on Arnold Creek and another on Marshall Creek, the sediment zinc concentrations were well above the U.S. E P A designated "heavily polluted" level for the Great Lakes Harbours. Potential anthropogenic sources of zinc include automobile traffic, and fertilizer and animal feed supplements. The nickel and chromium contamination of sediments due to the landslide in the headwaters was apparent in the results. A dramatic decrease in these metal concentrations with distance downstream of the landslide was measured, but the levels remained well above natural levels even at the lowest station. Overall nickel and chromium levels increased since the 1970's, likely due to the gradual movement of landslide sediment downstream over time. Further distribution of this contaminated sediment within the watershed will likely occur with each successive flood. Nutrient and fecal coliform levels in the water column were much higher in the wet season than in the dry season, and highest on the sampling day of greatest discharge. This is likely from runoff carrying manure which is spread in the wet season due to lack of winter storage. Winter flood waters could potentially carry very high loads of agricultural 115 contaminants. Nitrate-N levels exhibited an increase in the downstream direction; ammonia may be nitrifying to nitrate as it is carried in the often slow-moving stream. The highest nitrate-N levels were recorded in Marshall Creek. This creek differed from the rest of the system in that nitrate-N levels were highest in the dry season. Groundwater input, which is proportionately greater in the dry season, is the suspected source. Arnold Slough was identified as having the poorest water quality within the system. High ammonia and temperature levels, and low dissolved oxygen levels create a hazardous environment for fish, particularly in the fall season which coincides with the migration of spawning salmon. Fecal coliform levels were found to be highest in Arnold Slough, Marshall Creek and Saar Creek. Although a 5 sample/30 day geometric mean could not be calculated, the individual coliform counts were often above the 200 FC/100 mL limit recommended for water recreation and irrigation of vegetables/fruit eaten raw (CCREM, 1987 and B C M O E L P , 1994). Finally, collection of historic data since 1970 from other sources allowed a tentative illustration of the trends toward an increasing spread of data values during the wet season for nitrate-N, ortho-P, ammonia-N and chloride. This data also illustrated the relationship of increased levels of ammonia-N and ortho-P with increased discharge. 6.3 Summary of Relationships Between Water Quality and Land Use The correlation of land use indicators with surface water quality on a contributing area basis resulted in several significant relationships. Ammonia-N correlated positively with surplus nitrogen loading and with the amount of clay as a surficial texture in the contributing 116 area. Dissolved oxygen levels had a negative correlation with these same two indicators. These relationships were strongest using the wet season water quality data. It therefore appears that the more surplus nitrogen applied to the land, the lower will be the dissolved oxygen levels, and the higher the ammonia-N levels, in the stream. These negative impacts are compounded in the wet season and if the surplus nitrogen is applied to very fine surface textures, both factors equating to increased runoff conditions. Nitrate-N levels did not correlate with any of the agricultural activity indicators. It did, however, correlate positively with the amount of organic soils and clay surficial texture, and negatively with the amount of sandy texture in the contributing area. Mineralization of organic nitrogen in soils may be another source of nitrates to the stream. Nitrates in the field may be infiltrating into sandy areas and undergoing denitrification and/or appearing in the system further downstream of the sampling point, which would offer another explanation for the increasing nitrate levels in the downstream direction. An interesting point emerging from these relationships is the implication to certain traditional agricultural waste management philosophies applicable to most other agricultural watersheds. Agricultural applications are usually considered more of a risk on coarse textured soils due to groundwater quality concerns. However, groundwater quality beneath the Sumas Prairie is a relative non-issue, since farms are serviced with municipal water. It is therefore on the fine textured soils, which serve to protect the groundwater, where agricultural applications pose the greatest threat to the water resource of concern, the surface water. 117 6.4 Recommendations The above summaries lead to several possible recommendations, addressed to individuals or agencies able to pursue research within the watershed, or to decision-makers and managers which influence the activities within the watershed. 1. Investigation of high zinc levels. This study did not pursue the identification of the potential sources of high zinc concentrations in the sediment. Possible anthropogenic sources should be researched. Further sediment and soil sampling is required to help determine both the source and whether the trend of increasing levels continues. 2. Investigation of groundwater inputs to Marshall Creek. Groundwater contribution from the contaminated Abbotsford aquifer is indicated in this study. However, what and how much of the potential contaminants from the aquifer is not known. An investigation of the groundwater hydrology in this vicinity and the discharge of the trout hatchery, and their effects on water quality and quantity in the stream, should be conducted to determine if this phenomena/practice should be of concern. 3. Agricultural waste management. The use of best management practices should continue to be encouraged to reduce agricultural pollution, including nutrient inputs and fecal coliform counts. Over the long term, the best management practices for the Sumas River watershed would ideally include measures to decrease the potential for the spread of contaminants during floods, considering the frequency of their occurrence. This may involve the modification of manure and animal storage facilities. Better manure management during the fall rainy season needs to be more aggressively encouraged. This critical period results from the unfortunate combination of a) farmers 118 needing to empty their manure storage facilities in preparation for the winter, b) an increase in volume and intensity of rainfall, c) oxygen-consuming die-off of summer algal blooms (themselves a result of nutrient enrichment), and d) the migration of spawning salmon. The Arnold Slough subwatershed should be particularly targeted for the implementation of best management practices. 4. Continued monitoring and consideration of land use trends, agricultural intensification, and flooding impacts. Increasing industrial and residential areas will generate water contaminants of their own, and will gradually change the nature of water quantity and quality as it is currently documented in the watershed. The increasing animal densities, farm and dwelling numbers should be monitored to see if these trends continue. Land management decisions regarding intensification should be made in light of the capacity of the land to accept the wastes generated or alternatives available to deal with the wastes, and in light of the flooding risk to the Prairie. Increasing animal densities have major implications to flood emergency response plans, mitigation, and environmental health. 5. Incorporation of site specifics into land management policies. The relationships identified between land use and water quality illustrate the importance of the influence of soil types. Blanket policies which do not take into consideration these details in the context of other watershed characteristics may prove to be overly restrictive, ineffective and inefficient. 6. Development of water quality/watershed management model. To aid further research investigations and facilitate land management decisions, a water quality/watershed management model needs to be developed. This would include water sampling in conjunction 119 with flow measurement to develop and calibrate a water quality model. The GIS database built in this study could be further developed to provide input to the water quality model and explore management scenarios. The ESP developed by IRC (1994) could be used as a more comprehensive management indicator to relate to water quality than the animal densities and surplus nitrogen loading used in this study. The quantification of the ESP on a contributing area basis and its correlation with water quality values may lend more understanding to land use - water quality relationships. Future sampling schemes should have a greater sampling density on the U.S. side than employed in this study, and also should increase the sampling density in identified problem subwatersheds (e.g., Arnold Creek watershed). 7. Consideration of social and economic factors. It is impossible to make good water management decisions in isolation from social and economic influences. Although they are not addressed in this study, the watershed management model will naturally need to incorporate these complicating factors. For example, community goals and desires, urbanization and industrialization pressures, behaviour of commodity markets, free trade impacts and agricultural competition, and prices of animal feed and other agricultural inputs, all influence the dynamics of land use and the resulting water quality in the watershed. 120 7.0 References Anderson, G.D., De Bossu, A.E. , and Kuch, P.J., 1990. Control of Agricultural Pollution by Regulation. In: Agriculture and Water Quality, International Perspectives. (Eds.) Braden, J.B., and Lovejoy, S.B., Lynne Rienner Publishers, Boulder and London. B C M O A F F , 1992. Environmental Guidelines for Poultry Producers in British Columbia. British Columbia Ministry of Agriculture, Fisheries and Food, Abbotsford, B.C. B C M O A F F , 1993. Environmental Guidelines for Dairy Producers in British Columbia. British Columbia Ministry of Agriculture, Fisheries and Food, Abbotsford, B.C. B C M O E L P , 1994. Approved and Working Criteria for Water Quality - 1994. Water Quality Branch, Environment Protection Department, Ministry of Environment Lands and Parks. B C M O E L P and EC, 1994. Groundwater Resources of British Columbia. Published by B.C. Environment and Environment Canada. B C M O E L P , 1995. Water Quality Assessment and Objectives for the Fraser River from Hope to Sturgeon and Roberts Banks, First Update, Technical Appendix. Water Quality Branch and Environmental Protection Division, Ministry of Environment Lands and Parks. Benedict, A . H . , Hall, K.J . , and Koch, F.A., 1973. A Preliminary Water Quality Survey of the Lower Fraser River System. Westwater Research Centre, University of British Columbia. Brisbin, P.E., 1995. Agricultural Nutrient Management in the Lower Fraser Valley. In: Agricultural Inventory for the Lower Fraser Valley. Part of the Management of Agricultural Wastes in the Lower Fraser Valley Program. Draft document prepared for B.C. Ministry of Environment, Lands and Parks. Burrough, P. A. , 1986. Principles of Geographical Information Systems for Land Resources Assessment. New York, Oxford University Press. C C M E (Canadian Council of Ministers of the Environment), 1993. Protocols for Deriving Water Quality Guidelines for the Protection of Agricultural Water Uses. Eco-Health Branch, Ecosystem Science and Evaluation Directorate, Environment Canada, Ottawa, Ont. C C M E , 1995. Protocol for the Derivation of Canadian Sediment Quality Guidelines for the Protection of Aquatic Life. Report C C M E EPC-98E, March 1995, Ottawa, Ont. 121 C C R E M (Canadian Council of Resource and Environment Ministers), 1987. Canadian Water Quality Guidelines. Task Force on Water Quality Guidelines, Ottawa, Ont. Chapman, D. (ed.), 1992. Water Quality Assessments: A Guide to the Use of Biota, Sediments and Water in Environmental Monitoring. Chapman and Hall Ltd. London, U K . Colman, J. A. , and Sanzolone, R.F., 1992. "Geochemical Characterization of Streambed Sediment in the Upper Illinois River Basin". Water Resources Bulletin, American Water Resources Association, Vol.28, no.5, pp.933-950. Cook, K . E . , 1994. An Evaluation of Water Quality and Land Use in the Salmon River Watershed, Langley, B.C., Using GIS Techniques. M.Sc. thesis, Department of Soil Science, University of British Columbia, Vancouver, B.C. Cook, K. , 1995. "Historical Floods in the Sumas Prairie, British Columbia". Prepared for Dr. Hans Schreier, Resource Managment and Environmental Studies, University of British Columbia. November 1995. de Groot, A.J. , 1982. "Standardization of Methods of Analysis for Heavy Metals in Sediments." Hydrobiologia 92, pp. 689-695. De Leeuw, A.D. , and Stuart, K . M , 1981. Small Stream Enhancement Possibilities for Sea-Run Cutthroat Trout in the Lower Mainland and Sechelt Peninsula. Volume 4 of 4: Enhancement Options. Fish and Wildlife Branch, Ministry of Environment. DFO, 1995. Stream Information Summary (SIS) report on the Sumas River and its Tributaries. Obtained from SIS database from Habitat Inventory Program, Department of Fisheries and Oceans, February 1995. District of Abbotsford, June 1993. "A Guide to the Barrowtown Pump Station and the Sumas Prairie Floodplain", District of Abbotsford, June 1993. District of Abbotsford, 1993. Official Community Plan. Drinnan, R.W., and Clark, M.J.R., 1980. Fraser River Estuary Study, Water Quality, Water Chemistry, 1970-1978. Victoria, B.C. Ellis, K . V. , 1989. Surface Water Pollution and its Control. Macmillan Press Ltd. E R M , 1992. District of Abbotsford Water Study. Prepared for District of Abbotsford. 122 E S L and Webb, 1987. Feasibility Study Sumas River Drainage Abbotsford, B.C. Volume 3, Part A and B . Fish and Wildlife Resources. Prepared for B.C. Ministry of Agriculture and Fisheries, Agricultural Engineering Branch. FREMP, 1990. Agricultural Runoff Contamination in the Fraser River Estuary. Waste Management Activity Program Discussion Paper, Fraser River Estuary Management Plan, December 1990. Hall, K.J . , Koch, F.A., and Yesaki, I., 1974. Further Investigations into Water Quality Conditions in the Lower Fraser River System. Westwater Research Centre, University of British Columbia. Hall, K . J., Yesaki, I., and Chan, J., 1976. Trace Metals and Chlorinated Hydrocarbons in the Sediments of a Metropolitan Watershed. Technical Report No. 10, Westwater Research Centre, University of British Columbia, Vancouver, 74 p. Hall, K.J . , 1992. Limnology of Okanagan Lake, B.C., Related to the ProposedKelowna STP Outfall. Prepared for Dayton and Knight Ltd., Westwater Research Centre, University of British Columbia. Halstead, E.C., 1986. Groundwater Supply - Fraser Lowland, British Columbia. Environment Canada, Inland Waters Directorate, National Hydrology Research Institute, Saskatoon, Saskatchewan. NHRI Paper No. 26. Harris, H.G. , 1990. "Computer Modelling of B.C.'s Sumas Prairie Drainage." In: Innovations in River Basin Management, Canadian Water Resources Association Conference Proceedings 1990, Penticton, B.C. Hem, J.D., 1985. "Study and Interpretation of the Chemical Characteristics of Natural Water." U.S. Geological Survey Water Supply Paper 2254. Washington, D.C., 263 p. IRC, 1994. Agricultural Land Use Survey in the Sumas River Watershed - Summary Report. IRC Integrated Resource Consultants Inc. Prepared for B.C. Ministry of Environment, Environment Canada and Department of Fisheries and Oceans. K C M , 1990. Lower Nooksack River Comprehensive Flood Hazard Management Plan. Draft. Prepared for Whatcom County. Klassen, H . , Barraclough, J.B., and Mascarenhas, M . , 1995. Strategic Review of Fisheries Resources of Urban Streams in the Lower Fraser Valley. Draft report, Fraser River Action Plan. 123 Klohn Leonoff, 1989. Engineering Studies for Floodplain Management Plan. Prepared for the District of Abbotsford. Klohn Leonoff, 1991. Flooding of West Sumas Prairie November 9-12, 1990. Prepared for B.C. Environment, Water Managment. Klohn Leonoff, 1993. Nooksack River Avulsion Study. Prepared for B.C. Environment, Water Management. K P A , 1987. Feasibility Study Sumas River Drainage Abbotsford, B.C. Volume 1, Engineering Assessment. Prepared for B.C. Ministry of Agriculture and Fisheries, Agricultural Engineering Branch. Lachat Instruments, 1990. Methods Manual for the QuikChem Automated Ion Analyzer. Milwaukee, Wisconsin. Lamb, J.C., 1985. Water Quality and its Control. Wiley, 385 p. Liebscher, H. , Hii , B. , and McNaughton, D., 1992. Nitrates and Pesticides in the Abbotsford Aquifer, Southwestern British Columbia. Environment Canada. Luttmerding, H.A., 1980. Soils of the Langley-Vancouver Map Area. R A B Bulletin No.18, B.C. Soil Survey Report No. 15, B C M E , Kelowna. McCallum, D., 1995. An Examination of Trace Metal Contamination and Land Use in an Urban Watershed. University of British Columbia. M O A F , 1976. Agricultural Code, of Practice, Ministry of Agriculture and Food, Ministry of the Environment, Ministry of Housing, Ontario. Moldan, B . and Cerny, J. (eds.), 1994. Biogeochemistry of Small Catchments: A Tool for Environmental Research. John Wiley and Sons Ltd. Nagpal, N .K. and Pommen, L.W., 1994. Approved and Working Criteria for Water Quality -1994. Prepared for the Ministry of Environment, Lands and Parks, Water Quality Branch, Environmental Protection Department, B.C. N H C (Northwest Hydraulics Consultants) and Hamilton, R., 1994. Hydrology and Water Use for Salmon Streams in the Chilliwack/Lower Fraser Habitat Management Area, British Columbia. Draft Report prepared for Fraser River Action Plan, Department of Fisheries & Oceans. 124 Owens, L .B . , 1994. "Impacts of Soil N Management on the Quality of Surface and Subsurface Water." In: Soil Processes and Water Quality. (Eds) Lai, R. and Stewart, B .A. Lewis Publishers. Page, A . L . , Miller, R.H. , Keeney, D R . (eds), 1982. "1-5 Method for Digestion with Hydrofluoric Acid in a Closed Vessel." Methods of Soil Analysis: Part 2, Chemical and Microbiological Properties. 2nd Edition. American Society of Agronomy, Inc., Soil Science Society of America, Inc., Wisconsin, USA. Palmer, M . and Rising, N . , 1996. "The Development of an Environmental Sustainability Parameter for Agriculture." Canadian Water Resources Journal, Vol.21, No. 1, pp. 13-25. Puget Sound Water Quality Authority, 1990. Water in Whatcom County: A Watershed Guidebook. 2nd Edition. Sawicki, J. and G.Runka, 1986. Land Use Classification for British Columbia. Prepared for the Ministry of Agriculture and Food, Soils Branch, and the Ministry of Environment, Surveys and Resources Mapping Branch. M O E Manual 8. Victoria, B .C. Schmidt, O., 1995, personal communication, response to questionnaire, June 2, 1995. Dairy Producers' Conservation Group Project Coordinator. Schreier, H . , 1986. Trace Metals and Asbestos Fibres in the Bed-Sediments of the Lower Sumas River - A Reconnaissance Survey. Prepared for B.C. Ministry of Agriculture and Food, Engineering Branch, Abbotsford, B.C. Department of Soil Science and Westwater Research Centre. Schreier, H . , 1987. "Asbestos Fibres Introduce Trace Metals into Streamwater and Sediments", Environmental Pollution 43 , pp. 229-242. Schreier, H . , Northcote, T . G , and Hall, K. , 1987. "Trace Metals in Fish Exposed to Asbestos Rich Sediments." Water, Air, and Soil Pollution, 35 (1987) pp.279-291. Sellars, C D . , Montgomery, R.A., Wright, D.F., 1991. The Impact of Overflow from the Nooksack River, Washington, on Flood Levels in West Sumas Prairie, British Columbia. 1991 Annual Conference of the Canadian Society for Civil Engineering. Sharpley, A, Chapra, S.C., Wedepohl, R., Sims, J.T., Daniel, T.C. and Reddy, K.R. , 1994. "Managing Agricultural Phosphorus for the Protection of Surface Waters: Issues and Options". Journal of Environmental Quality, 23: 437-451. 125 Stednick J.D., 1991. Wildland Water Quality Sampling and Analysis. Academic Press, Inc., San Diego, California, 217 p. Sutton, A . L . , Nelson, D.W., Mayrose, V.B. , Kelly, D.T., 1983. "Effect of Copper Levels in Swine Manure on Corn and Soil." Journal of Environmental Quality, Vol . 12, no. 2, pp. 198-203. Task Force, 1991. Preliminary Report on Nooksack River Trans-Boundary Flooding. Nooksack River International Task Force, December 1991. Task Force, 1994. Status Report No. 2. Nooksack River International Task Force. May 1994. Teskey, J., 1990. Fish and Wildlife Managment Submission: Sumas Mountain Local Resource Use Plan and Crown Land Plan. Fish and Wildlife Management, Ministry of Environment. Timblin, Chuck, personal communication, February 22, 1996. Whatcom County Conservation District. U.S. Army COE (Corps of Engineers), 1993. Reconnaissance Report, Nooksack and Sumas Basins, Washington, Flood Damage Reduction Study. USD A, 1992. Soil Survey of Whatcom County Area, Washington. United States Department of Agriculture and Soil Conservation Service. Waite, T.D., 1984. Principles of Water Quality. Academic Press, Inc. Wernick, B. , 1996. Land Use and Water Quality Dynamics on the Urban-Rural Fringe: A GIS Evaluation of the Salmon River Watershed, Langley, B.C. University of British Columbia. Wehlen, M . , et al., 1986. Biophysical Reconnaissance and Late Summer Juvenile Salmonid Utilization of Selected Lower Fraser Tributaries, Sto:Lo Nation Society. Wright, Frank, personal communication, response to questionnaire, May 9, 1995. Dyking Drainage & Irrigation Superintendent, City of Abbotsford. 126 CO O -*—< 0) c o —' co Q> O ) Q g> C O O J O ) T— T 3 C ro x a> E CD .c O >*— o c o CO c ca O . E o O •a c CO CO co (V or c g ca o a. CD or a> in c < x T 3 C CD Q . < •5 i CU o l 0) > t I: a) o |T3 j= 5* I4 <o p O .SI E 0) a E ra (0 oo CM CN CO CM CO CO CM CO CO O) CO CM 00 OO s CD CO CO 00 CM CM CM 00 CO 00 3 o CO o CO CO CM CM CM CM CM 00 in CM CO CO CO I CO CO CO 3 in 00 CO o 5 I CM I 00 co co s1 SI Q co o m CM > CO CO cu 'l o -Q CD —I X cu E CU _c O >> -O. •=  ca E CO CO V o CO vfc C O •B «= CD •f3 •o ro T5 TO "So o 3* 3 O 5= t -00 cu o c cu "O o c cu T3 X CU E CU JO. O CU o 127 Appendix B: Accuracy and Precision Calculations Standard Reference Cr ppm (mg/kg) Cu ppm (mg/kg) Ni ppm (mg/kg) Zn ppm (mg/kg) MESS-1 measured: certified: %error outside range: 51.19 71 +/-11 15 3.95 25.1 +/- 3.8 81 25.29 29.5 +/- 2.7 7 158.63 191 +/-17 9 BCSS-1 measured: certified: %error outside range: Chemex: 90.18 123 +/-14 17 52 4.09 18.5 +/-2J 74 14 54.53 55.3 +/- 3.6 0 49 92.99 119+/-12 13 106 REPLICATES 1114 1114-R %diff 276.96 257.58 7 30.75 24.88 19 1086.63 1085.94 0 106.50 107.72 1 1122 1122-R %diff 131.94 127.74 3 111.11 123.59 11 228.32 230.13 1 274.46 284.01 3 1503 1503-R %diff 178.92 101.45 54.83 10.97 80 206.55 112.28 46 135.40 95.64 29 2501 2501-R %dlff 348.26 357.03 3 16.22 27.92 72 1862.43 1910.15 3 63.38 67.85 7 2503D 2503D-R %diff 97.86 100.30 2 10.54 9.17 13 102.49 97.32 5 90.17 89.58 1 2506 2506-R %diff 222.43 235.51 6 49.41 73.21 48 492.85 532.76 8 179.15 182.45 2 788C 788C-R %diff 18.27 18.94 4 6.35 4.83 24 14.96 13.36 11 51.66 48.49 6 792B 792B-R %dlff 38.71 32.71 15 -6.74 -6.83 1 21.92 20.65 6 75.25 73.70 2 AVE %diff std. dev 10.4 13.9 33.6 29.6 9.8 14.9 6.5 9.5 AVE %diff (excluding std. dev sample 1503) 5.8 4.6 27.0 24.7 4.7 3.9 3.2 2.5 128 Appendix C: Sediment Sampling Results, September 1993 and August 1994 Note: Cd and Pb results were below detection limits 1993 1 1994 Station Cr ppm Cu ppm Ni ppm Zn ppm i Station Cr ppm Cu ppm Ni ppm Zn ppm (mg/kg) (mg/kg) (mg/kg) (mg/kg) 1 (mg/kg) (mg/kg) (mg/kg) (mg/kg) 1 1112 101.50 31.68 86.08 167.48 1 2112 92.78 33.75 78.68 158.18 1113 162.16 79.08 163.49 338.55 | i 2113 115.35 68.63 114.46 275.82 1114 1114-R 276.96 257.58 30.75 24.88 1086.63 1085.94 106.50 107.72 2114 249.32 44.21 1011.41 121.94 1115 248.16 27.68 946.40 113.27 1116 321.40 36.92 1334.87 299.69 2116 2116B 2116C 2116D 318.37 235.32 225.03 211.32 24.97 18.73 10.49 16.01 1225.99 921.76 831.54 684.84 87.63 76.47 80.60 92.93 1 1 1 1 1118 120.38 44.20 148.83 120.31 2118 149.42 47.20 176.00 133.25 1119 153.84 39.31 190.19 131.72 % 1120 136.82 48.60 173.53 115.79 2120 165.33 46.86 188.51 124.92 1121 117.33 83.31 215.85 154.05 i 2121 160.93 79.38 236.11 187.65 1122 1122-R 131.94 127.74 111.11 123.59 228.32 230.13 274.46 284.01 i 1 2122 50.56 32.53 81.29 244.16 1123 97.46 45.21 130.99 198.58 2123 134.96 59.25 187.30 149.22 1125 123.93 11.75 147.34 87.85 2125 126.98 33.36 144.57 125.20 1126 78.73 34.99 104.65 113.34 1127 84.76 40.92 150.30 178.91 1128 65.03 48.25 143.80 144.03 1129 72.82 26.53 133.84 68.18 1 2129 118.47 38.48 204.99 140.52 1130 76.28 49.16 132.88 193.66 1 2130 102.45 44.63 153.36 198.60 1131 52.44 20.99 75.11 110.10 1 & 1133 89.00 42.86 142.50 187.13 2133 98.84 53.78 180.47 230.69 1134 150.62 31.40 249.61 127.94 1 2134 134.21 22.01 170.97 105.98 1135 108.81 40.64 112.01 144.64 i 2135 117.35 26.54 120.81 124.89 1136 107.71 56.87 68.62 214.08 1 2136 90.17 42.45 53.92 123.06 1145 74.37 53.65 66.93 207.46 1 2145 81.87 52.56 69.51 191.06 1146 44.97 45.81 42.18 200.86 i 2146 80.28 37.41 58.82 211.56 1500 367.70 24.01 1929.59 28.78 i i 2500 338.32 7.73 1809.03 32.29 1501 370.86 12.10 1827.30 45.32 2501 2501-R 348.26 357.03 16.22 27.92 1862.43 1910.15 63.38 67.85 1502 175.52 37.53 391.52 123.47 2502 2502B 2502C 2502D 182.49 141.55 133.42 120.72 41.55 30.66 22.39 20.70 429.01 322.96 288.44 281.18 131.05 107.17 101.66 99.41 1503 1503-R 178.92 101.45 54.83 10.97 206.55 112.28 135.40 95.64 I 2503 2503B 2503C 2503D 2503D-R 167.03 130.92 96.55 97.86 100.30 33.46 13.32 6.21 10.54 9.17 171.20 121.07 93.74 102.49 97.32 118.14 93.43 84.57 90.17 89.58 1504 106.21 32.10 110.30 187.60 2504 95.85 42.78 129.36 159.70 2505 123.56 32.61 47.87 175.99 1 2506 2506-R 222.43 235.51 49.41 73.21 492.85 532.76 179.15 182.45 129 Appendix D: Water Sampling Results pH March 2 May 9 June 23 June 29 July 26 Aug 24 Oct 3 Nov 24 Feb 1, 95 03-02 05-09 06-23 06-29 07-26 08-24 10-03 11-24 02-01 Station Median Max Min 1 8.5 8.5 8.5 8.1 9.4 8.3 8.4 9.1 9.2 8.5 9.4 8.1 2 7.7 7.6 8.0 8.0 7.8 7.6 8.0 7.9 8.2 8.0 8.2 7.6 3 7.3 7.5 8.2 7.8 7.9 7.8 7.2 7.8 7.9 7.8 8.2 7.2 4 7.3 7.9 8.0 8.1 8.2 7.9 7.8 7.8 7.6 7.8 8.2 7.3 5 6.7 7.2 - 8.0 7.5 7.3 7.3 7.5 6.9 7.3 8.0 6.7 6 6.8 7.4 7.4 8.0 8.1 8.0 7.6 7.8 7.1 7.7 8.1 6.8 7 7.0 8.3 7.6 8.1 9.1 8.0 7.8 7.7 7.1 7.9 9.1 7.0 8 6.9 8.5 6.9 8.0 7.8 7.4 7.5 7.4 7.0 7.5 8.5 6.9 9 6.8 7.7 - 7.7 7.5 7.4 7.4 7.2 6.9 7.4 7.7 6.8 10 6.9 6.6 7.3 7.5 7.2 7.3 7.2 8.4 6.8 7.3 8.4 6.6 11 6.9 6.5 7.4 7.4 7.9 8.6 7.4 7.3 6.9 7.4 8.6 6.5 12 7.2 7.6 7.1 8.0 8.4 8.2 7.8 7.6 7.1 7.7 8.4 7.1 13 6.8 7.5 7.2 7.8 7.8 7.5 7.7 7.4 6.6 7.5 7.8 6.6 14 - 7.1 7.6 7.5 7.3 7.3 6.8 7.5 7.1 7.3 7.6 6.8 15 - 6.9 7.9 7.7 7.9 7.5 7.7 9.1 7.2 7.7 9.1 6.9 16 - 6.9 7.1 7.3 7.1 7.2 6.9 7.0 7.0 7.0 7.3 6.9 Median 6.9 7.5 7.5 7.9 7.9 7.5 7.6 7.7 7.1 7.6 Max 8.5 8.5 8.5 8.1 9.4 8.6 8.4 9.1 9.2 9.4 Min 6.7 6.5 6.9 7.3 7.1 7.2 6.8 7.0 6.6 6.5 Conduc ivity (micromhos/cm) March 2 May 9 June 23 June 2 July 26 Aug 24 Oct 3 Nov 24 Feb 1, 95 03-02 05-09 06-23 06-29 07-26 08-24 10-03 11-24 02-01 Station Median Max Min 1 130 400 160 185 290 250 170 125 90 170 400 90 2 166 240 245 240 290 230 165 180 107 230 290 107 3 180 260 250 260 320 250 170 185 111 250 320 111 4 182 265 250 270 350 270 170 190 122 250 350 122 5 183 245 245 240 280 235 140 215 91 235 280 91 6 140 265 225 275 300 265 175 162 95 225 300 95 7 140 260 230 265 270 215 170 175 132 215 270 132 8 155 205 220 210 260 200 150 190 92 200 260 92 9 198 175 180 210 295 260 160 200 86 198 295 86 10 192 260 250 295 315 280 180 232 91 250 320 91 11 203 290 245 270 335 270 - 205 35 258 335 35 12 160 290 245 270 325 300 - 175 130 258 325 130 13 126 200 212 235 240 230 - 210 120 211 240 120 14 - 130 108 142 205 295 - 65 37 130 295 37 15 - 100 242 135 130 110 100 110 75 110 242 75 16 - 240 200 192 200 230 120 152 98 200 240 98 Median 166 253 236 240 290 250 168 183 94 212 Max 203 400 250 295 350 300 180 232 132 400 Min 126 100 108 135 130 110 100 65 35 35 130 Appendix D: Water Sampling Results Dissolved Oxygen (mg/L) March 2 May 9 June 2 June 29 July 26 Aug 24 Oct 3 Nov 24 Feb 1, 95 03-02 05-09 06-23 06-29 07-26 08-24 10-03 11-24 02-01 Station Median Max Min 1 - 10.0 10.2 9.5 9.7 11.5 10.5 13.9 11.4 10.4 13.9 9.5 2 - 8.2 8.4 9.2 9.2 9.7 9.9 9.7 8.6 9.2 9.9 8.2 3 - 9.1 7.4 8.0 8.0 9.0 9.8 10.3 8.6 8.8 10.3 7.4 4 - 9.6 8.0 8.9 8.1 9.4 10.2 10.2 9.0 9.2 10.2 8.0 5 - 8.6 6.0 9.2 5.6 6.6 9.5 6.9 7.4 7.2 9.5 5.6 6 - 9.5 5.8 9.0 7.2 8.9 9.8 11.9 9.0 9.0 11.9 5.8 7 - 12.2 8.2 10.4 8.0 9.3 10.2 11.9 9.4 9.8 12.2 8.0 8 - 12.0 7.8 10.0 8.8 7.9 7.8 6.6 6.7 7.9 12.0 6.6 9 - 11.4 5.8 8.4 5.1 4.6 6.5 6.0 7.2 6.3 11.4 4.6 10 - 6.4 2.8 4.1 4.0 5.2 6.0 2.7 6.0 4.7 6.4 2.7 11 - 4.5 2.5 5.4 5.6 13.8 12.4 2.1 9.5 5.5 13.8 2.1 12 - 9.0 6.0 8.3 6.9 11.2 8.0 10.4 8.6 8.5 11.2 6.0 13 - 9.7 9.2 9.3 9.4 8.7 6.8 10.2 8.3 9.3 10.2 6.8 14 - 9.5 8.3 8.2 6.5 6.8 1.5 10.9 10.5 8.3 10.9 1.5 15 - 9.4 9.8 9.2 8.7 8.8 9.5 12.0 10.8 9.5 12.0 8.7 16 - 5.5 3.2 3.8 3.3 4.0 9.2 4.5 7.5 4.3 9.2 3.2 Median 9.5 7.6 9.0 7.6 8.9 9.5 10.2 8.6 8.0 Max 12.2 10.2 10.4 9.7 13.8 12.4 13.9 11.4 13.9 Min 4.5 2.5 3.8 3.3 4.0 1.5 2.1 6.0 1.5 Temperature (cleg C) March 2 May 9 June 23 June 2 July 26 Aug 24 Oct 3 Nov 24 Feb 1, 95 03-02 05-09 06-23 06-29 07-26 08-24 10-03 11-24 02-01 Station Median Max Min 1 - 16.0 14.0 16.0 17.0 12.7 6.4 3.3 8.1 14.0 17.0 3.3 2 - 13.0 13.0 13.5 13.5 12.3 9.5 4.9 8.5 12.3 13.5 4.9 3 - 16.0 17.0 17.0 17.0 15.0 11.0 4.9 8.7 16.0 17.2 4.9 4 - 17.0 17.5 18.0 20.0 17.3 12.0 4.6 8.8 17.3 20.0 4.6 5 - 18.0 20.0 18.5 20.0 16.1 13.0 7.3 9.3 18.0 20.0 7.3 6 - 19.5 19.5 20.0 24.0 22.0 15.0 5.4 8.6 19.5 24.0 5.4 7 - 21.0 20.0 20.5 25.0 23.0 13.0 4.7 8.6 20.5 25.0 4.7 8 - 21.0 20.5 20.0 25.0 22.3 14.4 6.7 8.8 20.3 25.0 6.7 9 - 20.0 19.0 19.0 24.0 20.9 13.4 6.4 8.7 19.0 24.0 6.4 10 - 18.0 16.5 17.0 19.0 18.3 11.5 6.8 8.8 17.0 20.4 6.8 11 - 20.0 20.0 20.0 26.0 22.2 13.5 6.1 8.0 20.0 26.0 6.1 12 - 21.0 20.0 20.0 25.0 22.1 13.2 5.4 8.6 20.0 25.0 5.4 13 - 12.0 12.5 14.5 14.0 14.6 12.2 8.5 8.9 12.5 14.6 8.5 14 - 17.0 14.5 16.0 19.0 16.8 12.5 4.4 7.8 15.3 19.0 4.4 15 - 14.0 13.0 15.0 15.5 13.6 11.0 3.7 7.6 13.6 15.5 3.7 16 - 12.0 13.0 - 17.5 11.8 7.2 5.0 8.6 11.9 17.5 5.0 Median 17.5 17.3 18.0 19.5 17.1 12.4 5.2 8.6 16.7 Max 21.0 20.5 20.5 26.0 23.0 15.0 8.5 9.3 26.0 Min 12.0 12.5 13.5 13.5 11.8 6.4 3.3 7.6 3.3 131 Appendix D: Water Sampling Results Orthophosphate (mg/L) March 2 May 9 June 29 July 26 Aug 24 Oct 3 Nov 24 Feb 1, 95 03-02 05-09 06-29 07-26 08-24 10-03 11-24 02-01 Station Median Max Min 1 0.030 0.013 0.016 0.012 0.528 0.039 0.040 0.017 0.024 0.528 0.012 2 0.062 0.030 0.027 0.038 0.033 0.057 0.100 0.360 0.048 0.360 0.027 3 0.236 0.056 0.068 0.098 0.098 0.088 0.142 0.372 0.098 0.372 0.056 4 0.244 0.042 0.080 0.090 0.097 0.087 0.143 0.120 0.094 0.244 0.042 5 0.514 0.046 0.031 0.015 0.018 0.037 0.135 0.412 0.042 0.514 0.015 6 0.302 0.049 0.053 0.022 0.044 0.088 0.135 0.732 0.071 0.732 0.022 7 0.227 0.021 0.030 0.030 0.045 0.102 0.152 0.191 0.074 0.227 0.021 8 0.237 0.000 0.010 0.028 0.011 0.088 0.183 0.389 0.058 0.389 0.000 9 0.266 0.003 0.023 0.030 0.050 0.098 0.236 0.293 0.074 0.293 0.003 10 0.383 0.084 0.068 0.066 0.104 0.120 0.140 0.807 0.112 0.807 0.066 11 0.441 0.046 0.139 0.062 0.048 0.099 0.169 0.447 0.119 0.447 0.046 12 0.246 0.063 0.090 0.064 0.069 0.084 0.118 0.354 0.087 0.354 0.063 13 0.113 0.028 0.044 0.049 0.060 0.054 0.069 0.442 0.057 0.442 0.028 14 - 0.039 0.052 0.114 0.087 2.074 0.051 0.376 0.087 2.074 0.039 15 - 0.011 0.025 0.029 0.026 -1 0.044 -1 0.025 0.044 0.000 16 - 0.027 0.032 0.042 0.035 0.099 0.125 0.900 0.042 0.900 0.027 Median 0.244 0.035 0.038 0.040 0.049 0.088 0.135 0.374 0.069 Max 0.514 0.084 0.139 0.114 0.528 2.074 0.236 0.900 2.074 Min 0.030 0.000 0.010 0.012 0.011 0.000 0.040 0.000 0.000 Chloride (mg/L) March 2 May 9 June 2 July 26 Aug 24 Oct 3 Nov 24 Feb 1, 95 03-02 05-09 06-29 07-26 08-24 10-03 11-24 02-01 Station Median Max Min 1 7.674 11.452 10.643 16.214 16.894 21.542 15.421 15.146 15.284 21.542 7.674 2 8.852 14.120 14.104 14.207 11.383 13.002 14.694 25.535 14.112 25.535 8.852 3 9.319 18.699 17.910 19.700 15.157 14.836 14.961 24.921 16.534 24.921 9.319 4 9.559 18.626 19.196 20.373 15.053 14.451 16.375 19.998 17.501 20.373 9.559 5 13.518 17.682 12.664 11.988 10.865 9.816 16.765 13.923 13.091 17.682 9.816 6 8.250 17.249 15.887 13.660 12.098 13.775 14.236 28.896 14.006 28.896 8.250 7 8.294 16.958 16.762 11.737 9.244 13.765 13.949 35.382 13.857 35.382 8.294 8 8.912 9.882 9.234 14.040 9.919 11.449 11.748 16.732 10.684 16.732 8.912 9 11.022 6.867 10.966 16.739 13.197 11.376 13.169 12.749 12.063 16.739 6.867 10 11.780 17.106 18.355 19.174 12.703 13.188 16.707 18.117 16.907 19.174 11.780 11 12.555 17.276 17.556 20.205 11.114 13.105 14.374 16.159 15.267 20.205 11.114 12 7.797 17.756 17.389 19.612 15.480 13.663 13.704 40.729 16.435 40.729 7.797 13 12.705 11.096 12.094 10.380 10.167 11.262 10.936 27.751 11.179 27.751 10.167 14 - 9.064 8.152 53.461 14.914 30.542 3.833 20.497 14.914 53.461 3.833 15 - 3.317 1.238 3.964 4.915 7.290 3.425 6.479 3.964 7.290 1.238 16 - 12.032 11.555 10.589 9.972 11.486 11.825 28.150 11.555 28.150 9.972 Median 9.319 15.539 13.384 15.211 11.741 13.147 14.093 20.248 13.584 Max 13.518 18.699 19.196 53.461 16.894 30.542 16.765 40.729 53.461 Min 7.674 3.317 1.238 3.964 4.915 7.290 3.425 6.479 1.238 132 Appendix D: Water Sampling Results Ammonia-N (mq ID March 2 May 9 June 2 July 26 Aug 24 Oct 3 Nov 24 Feb 1,95 03-02 05-09 06-29 07-26 08-24 10-03 11-24 02-01 Station Median Max Min 1 0.051 0.029 0.036 -1 0.003 0.027 0.035 0.006 0.028 0.051 0.000 2 0.209 0.003 0.009 -1 0.010 0.006 0.031 1.022 0.010 1.022 0.000 3 0.588 0.022 0.049 0.018 0.055 0.025 0.102 0.996 0.052 0.996 0.018 4 0.656 0.021 0.057 -1 0.037 0.020 0.091 0.641 0.047 0.656 0.000 5 2.379 0.043 0.025 -1 0.098 0.013 0.039 0.703 0.041 2.379 0.000 6 1.213 0.128 0.029 -1 0.037 0.125 0.196 1.009 0.127 1.213 0.000 7 0.650 0.020 0.024 0.019 0.017 0.148 0.207 0.456 0.086 0.650 0.017 8 0.927 0.000 -1 0.015 0.010 0.703 0.572 0.904 0.294 0.927 0.000 9 1.118 0.000 0.009 -1 -1 0.760 1.032 0.673 0.341 1.118 0.000 10 0.566 0.634 0.600 1.641 0.729 1.545 0.862 1.178 0.796 1.641 0.566 11 0.843 0.752 0.341 -1 0.010 0.702 1.049 0.994 0.727 1.049 0.000 12 0.509 0.153 0.103 0.048 0.036 0.243 0.234 0.664 0.194 0.664 0.036 13 0.385 0.026 0.082 -1 0.164 -1 0.104 0.652 0.093 0.652 0.000 14 - 0.268 0.086 0.105 0.285 4.029 0.054 0.655 0.268 4.029 0.054 15 - 0.037 0.042 -1 0.023 0.085 0.006 0.149 0.037 0.149 0.000 16 - 0.040 0.038 -1 0.018 0.206 -1 2.570 0.038 2.570 0.000 Median 0.650 0.033 0.040 0.000 0.030 0.148 0.103 0.688 0.199 Max 2.379 0.752 0.600 1.641 0.729 4.029 1.049 2.570 4.029 Min 0.051 0.000 0.000 0.000 0.000 0.000 0.000 0.006 0.000 Nitrate- M (mg/L) March 2 May 9 June 29 July 26 Aug 24 Oct 3 Nov 24 Feb 1,95 03-02 05-09 06-29 07-26 08-24 10-03 11-24 02-01 Station Median Max Min 1 0.599 0.121 0.231 0.198 -1 0.392 0.631 0.378 0.305 0.631 0.000 2 0.789 0.324 0.48 0.578 0.411 0.510 0.690 0.619 0.544 0.789 0.324 3 3.105 2.048 1.629 1.422 1.037 1.597 2.985 2.084 1.839 3.105 1.037 4 2.834 1.869 1.456 1.253 0.851 1.324 3.044 2.073 1.662 3.044 0.851 5 3.132 4.364 3.463 3.867 3.520 2.638 4.366 2.117 3.492 4.366 2.117 6 2.384 1.715 1.469 0.394 0.703 1.369 2.569 1.982 1.592 2.569 0.394 7 2.581 1.526 1.236 0.165 -1 1.198 3.360 2.700 1.381 3.360 0.000 8 1.425 0.374 0.189 0.224 0.111 0.449 1.233 1.166 0.412 1.425 0.111 9 1.288 0.438 0.296 0.227 0.117 0.391 0.682 1.248 0.414 1.288 0.117 10 4.102 0.324 0.209 0.344 0.311 0.285 2.775 2.169 0.334 4.102 0.209 11 4.586 0.236 0.406 0.224 0.097 0.256 1.817 1.358 0.331 4.586 0.097 12 2.736 1.105 1.148 0.746 0.427 0.936 2.678 2.572 1.126 2.736 0.427 13 1.882 7.606 5.159 7.090 3.972 4.642 6.170 2.472 4.901 7.606 1.882 14 - 0.683 0.377 -1 -1 -1 1.666 1.017 0.377 1.666 0.000 15 - 0.073 0.203 0.352 0.137 0.278 0.136 0.803 0.203 0.803 0.073 16 - 0.714 0.872 0.951 0.720 0.743 0.948 0.675 0.743 0.951 0.675 Median 2.581 0.699 0.676 0.373 0.361 0.627 2.193 1.670 1.228 Max 4.586 7.606 5.159 7.090 3.972 4.642 6.170 2.700 7.606 Min 0.599 0.073 0.189 0.000 0.000 0.000 0.136 0.378 0.000 133 o o > o (A CO £ ZJ CO CD CU o> .c Q . E CD V) to X T3 C CU O -Q . < 8 CO CM O o 3 o E c o a mm n U *[ > • o in ut 5 o : 69 d • CM • CM O i " * c CM i 3 CM O CD t- fM 8 9 8 CO . CM • O i l CM 8 oo II CO I  3 o i l CO 111 co co' 1° In co \9 oo • iri i 134 Appendix D: Water Sampling Results Fecal Coliform and Fecal Streptococci Levels in the Sumas River (CFU/100mL) July 26, 199 4 August 24, 1994 February 1, 1995 F.coliforms streptococci F.coliforms streptococci F.coliforms streptococci 120 250 620 790 <20 <20 42 32 65 14 1100 800 330 330 360 39 700 1100 350 80 249 16 800 500 150 30 1040 25 <10 9400 86 80 140 <10 4700 4300 450 <10 10 10 1300 800 100 <100 50 <10 2900 3400 20 <10 100 <10 <10 4300 690 700 1530 90 800 2600 100 <100 180 80 500 900 230 90 170 40 1600 1300 680 380 1230 1440 1200 1100 430 610 610 500 170 800 130000 300 1320 90 6 2 41 91 40 30 300 800 135 CO 0) Q . E ro CO i_ (D » CO CO c o CD O Q . or CO Q) C c o Q X T 3 C (D Q . Q . < DOC 1 99.7 11.2 10.7 40.5 51.2 126 o o o iri co' co r- io co ^ - M T - o) m CO oo cb oi CM r— io io CM oi ^ T - to oi co CO CM L O CD m oi co ^ ^ ~ CM CD CD CM O Q 19.6 22.2 19.1 co r--. co d T— CM o o o ^ oi r-~ oi oo iri ( N * r qoo CO S O ^ " co oo ^ r-' <° it CM iq "t (•) CO CO co m m co d T~ oo CM CO TDC CO Tt CO oi n oi T— CO CN 00 1^ CO d d 0 0 co m 0 o o 01 d CM T - CM CM co to 00 d CM 00 CM O)i CO 01 in co' T— CM CM CM T— O CO CO T " CM in cq N s m oi oi T-; oq o od T~ 00 CO CM 00 Nitrate-N 1.469 1.539 1.496 1.501 0.035 2 CM CO r- CM s m co CN o» o> CO CO CO CO CO CO 3.804 0.185 5 o r~- T - T -i*» -«fr oo -r- CM CJ> O CD s iri CO 6.360 0.599 9 00 1*- 00 in o o CO TJ- CO 1.387 0.023 2 m o> Ammonia-N 0.029 0.034 0.047 0.037 0.009 25 TJ- co m ( D I D O * T— T™ T— T— d o d o 0.146 0.027 19 CD LO LO O O O CD T- t- 1- O d o d o ' 0.103 0.005 5 0.994 0.987 1.102 0.951 1.009 0.065 m CM Chloride 15.887 15.888 15.781 15.852 0.061 0 N W Tt Ol CO CM i - O ) T - rt N N d oi oi cd 9.376 0.574 6 CD CM O LO n T - i - it o> CO o> 00 d c\i o d 11.251 0.709 6 14.859 16.713 16.905 16.159 1.130 m r-Ortho-P co m oo m co m o o o o d d 0.059 0.006 10 o m LO CM co 10 m oo o o o o d o d o 0.063 0.013 20 O ) • « -CO LO 00 o> o o o o d d d d 0.075 0.019 25 r-~ CM •* •* co •*•«»;•>*; * CD d ci c> 0.450 0.008 2 m T - CM Station CO co CO ^— IDate June 29 average CV(%) Aug 24 average CV(%) Nov 24 average CV (%) Feb 1, 95 average CV (%) Average CV (%) Maximum CV (%) 136 Appendix E: Box Plots of Water Quality Results by Station and by Sampling Date Ammonia-N 4.5' 4.0' 3.5' —J 1 3.0' z 2.5' co 2.0' J 1.5' 1.0' .5' 0.0, « S O C T - 1 9 9 4 *02-MVc-1994 -M01-F O0»IWW"19a4i 5 6 7 7 7 7 7 6 5 7 7 7 6 7 6 5 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10 11 12 13 14 15 16 Station 4.5 4.0 • * 3.5 g 3.0 Z 2.5 | 2.0i 1.5 1.0 .5 0.0 N *14 *5.0 *16 15 16 14 1 14 14 15 02-MAR-1994 29-JUN-1994 24-AUG-1994 24-NOV-1994 09-MAY-1994 26-JUL-1994 03-OCT-1994 01-FEB-1995 Date 137 Appendix E: Box Plots of Water Quality Results by Station and by Sampling Date Dissolved Oxygen 16 14 OC8-MW-1994 CC&M»Y-1994 02-MUG-1994 O01-FEB-198S O03CC^tB94 OCBOCT-1994 N= 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10 11 12 13 14 15 16 Station 16 16 16 16 16 16 16 09-MAY-1994 29-JUN-1994 24-AUG-1994 01-FEB-1995 23-JUN-1994 26-JUL-1994 03-OCT-1994 Date 138 Appendix E: Box Plots of Water Quality Results by Station and by Sampling Date Chloride *01-FEB-1995 O01-FEB-1996 •K01-FEB-1996 •K01-FEB-1995 *01-FEBJ395 026JW6"!WT O0&NKY-1984 -K01-FE N= 8 8 8 8 8 8 8 8 8 8 7 8 8 6 6 7 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10 11 12 13 14 15 16 Station 50 *14 O1.0 OlS 814 16 15 N= 13 16 15 15 16 16 02-MAR-1994 29-JUN-1994 24-AUG-1994 24-NOV-1994 09-MAY-1994 26-JUL-1994 03-OCT-1994 01-FEB-1995 Date 139 Appendix E: Box Plots of Water Quality Results by Station and by Sampling Date Conductivity 500 400 E o o E 300 > 200^ o O 100 1 I IUL 19 026JU M994 OOI+fcB-1995 X *23JI O01-FEB-1995 N= 9 9 9 9 9 9 9 9 9 9 8 8 8 7 8 8 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10 11 12 13 14 15 16 Station 500 200 16 12 16 16 N= 13 16 16 16 16 02-MAR-1994 23-JUN-1994 26-JUL-1994 03-OCT-1994 01-FEB-1995 09-MAY-1994 29-JUN-1994 24-AUG-1994 24-NOV-1994 Date 140 Appendix E: Box Plots of Water Quality Results by Station and by Sampling Date Dissolved Organic Carbon 120 100 80 F 60 E, tz O -O l_ (0 O o "c <0 "O CD > O 00 00 b 40 20 0 •K290UN-1994 X I I 024-AUG-1995 I 029JUN-1994 N = 6 6 6 6 6 6 5 5 6 6 6 5 6 6 6 6 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10 11 12 13 14 15 16 Station N = 16 15 16 16 14 16 09-MAY-1994 26-JUL-1994 24-NOV-1994 29-JUN-1994 24-AUG-1994 01-FEB-1995 Date 141 Appendix E: Box Plots of Water Quality Results by Station and by Sampling Date Nitrate-N 8 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10 11 12 13 14 15 16 Station N= 13 16 16 14 10 15 16 16 02-MAR-1994 29-JUN-1994 24-AUG-1994 24-NOV-1994 09-MAY-1994 26-JUL-1994 03-OCT-1994 01-FEB-1995 Date 142 Appendix E: Box Plots of Water Quality Results by Station and by Sampling Date pH 1 0 . 0 , _ 9.5 J 6.0 I N= 9 9 9 9 8 9 9 9 8 9 9 9 9 8 8 8 1.00 3.00 5.00 7.00 9.00 11.00 13.00 15.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 Station 10.0-9.5' 9.0' 8.5' —r— —|— —T— 8 ' 0 , | |—I—| p ^ ^ j [ I —I—| I 7.5' m m m m t l t t m m m m m . [——| _ | 7.0' mmmmm ' ' | —I— 6.5' — J — — 6.0 I . _ a a a a 02-MAR-1994 23-JUN-1994 26-JUL-1994 03-OCT-1994 01-FEB-1995 09-MAY-1994 29-JUN-1994 24-AUG-1994 24-NOV-1994 Date 143 Appendix E: Box Plots of Water Quality Results by Station and by Sampling Date Temperature 30 r — O N= 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 7 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10 11 12 13 14 15 16 Station N= 16 16 15 16 16 16 16 16 09-MAY-1994 29-JUN-1994 24-AUG-1994 24-NOV-1994 23-JUN-1994 26-JUL-1994 03-OCT-1994 01-FEB-1995 Date 144 Appendix E: Box Plots of Water Quality Results by Station and by Sampling Date Orthophosphate 1.0-.9-.8-.7-.6-.5" .4' .3-.2-.1 • 0.0 -K01-FE O01-FEB-1995 O01-FEB-1995 •K01-FEl9lfikBEB'1995 OQ2-W«-tS64 -H01-FEB-1995 N= 7 8 8 8 8 8 8 8 8 8 8 8 8 7 5 7 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10 11 12 13 14 15 16 Station N= 13 02-MAR-1994 29-JUN-1994 24-AUG-1994 24-NOV-1994 09-MAY-1994 26-JUL-1994 03-OCT-1994 01-FEB-1995 Date 145 Appendix F: Wet Season/Dry Season Water Quality Results 20 n 10 H A a • I I I I I I I - 2 0 - , £ c Q) CD £ 1 0 A T3 CD _> O CO CO 14 1011 Tributary Stations I I I I I I 13 5 15 Control > 14,11 5 1 1 -A * -- A t — Wet Season (Nov-Mar) — Dry Season (Jun-Aug) T T i—r i i i fy* i 16 2 3 4 12 6 Sumas mainstem (upstream ~> downstream) 9 8 Sumas Lake Canal Sampling Station O CD CD 2 , CD 3 20 c5 CD Q. E 40 -I 20 H 40 - i A A 14 1011 13 5 15 Tributary Stations Control 14,11 I A A -T T i—r 12 T 6 16 2 3 4 Sumas mainstem (upstream --> downstream) Sampling Station Wet Season (Nov-Mar) Dry Season (Jun-Aug) 1 1 9 8 Sumas Lake Canal 146 Appendix F: Wet Season/Dry Season Water Quality Results 400 -200 - • 0-9 • 1 1 1 1 1 1 1 1 1 1 1 1 1 1 14 1011 13 5 15 Tributary Stations 14,11 Control ^ 3 0 0 - | CO O E 200 -I o T 3 C o O 100 Wet Season (Nov-Mar) Dry Season (Jun-Aug) I 4 TT 1—r t—1—r 1—r 16 2 3 4 12 6 7 Sumas mainstem (upstream -> downstream) 9 8 Sumas Lake Canal Sampling Station 10 n 8 H 10 - 1 i . 8 6 —+ 14 1011 13 5 15 Tributary Stations Control t— Wet Season (Nov-Mar) $— Dry Season (Jun-Aug) i r 1—r n—1—r 12 6 7 16 2 3 4 Sumas mainstem (upstream —> downstream) Sampling Station 1 1— 9 8 Sumas Lake Canal 147 Appendix F: Wet Season/Dry Season Water Quality Results E, CD 20 ;g _o O 10 40 n 20 H 30 - i I I I I I I I I I I I I 1 14 1011 13 5 Tributary Stations - m 15 Control ' a ! - 5 » T - i 1—i—i—i—r 16 2 3 4 12 6 7 Sumas mainstem (upstream -> downstream) Sampling Station t— Wet Season (Nov-Mar) i— Dry Season (Jun-Aug) it— T T 9 8 Sumas Lake Canal 0.8 - i E, o TS JO CL CO O x: CL O sz •c O 0.4 - \ 0.0 0.8 -t 0.4 H 0.0 / • •-• 14 1011 13 5 15 Tributary Stations Control TT i—r 12 T T — A — Wet Season (Nov-Mar) - @ — Dry Season (Jun-Aug) 16 2 3 4 Sumas mainstem (upstream --> downstream) Sampling Station g 8 Sumas Lake Canal 148 Appendix F: Wet Season/Dry Season Water Quality Results 8 -I 4 H T—r 4 - i CD E Z 2 - | T I I 14 1011 Tributary Stations I I I I I I T 13 5 15 Control —Ar— Wet Season (Nov-Mar) - ® — Dry Season (Jun-Aug) T T i—r i—i—r i—r 16 2 3 4 12 6 Sumas mainstem (upstream —> downstream) 9 8 Sumas Lake Canal Sampling Station 2 -I 1 H 4 - i CD E 'c o E E < 2 H / * e-a 14 1011 13 5 Tributary Stations 15 Control — A Y — Wet Season (Nov-Mar) - ® — Dry Season (Jun-Aug) 16 2 3 4 12 Sumas mainstem (upstream --> downstream) 9 8 Sumas Lake Canal Sampling Station 149 Appendix F: Wet Season/Dry Season Water Quality Results 1*80 c o . Q CO O •I 40 CO •a o > o CO CO b 80 n 40 H H 2 *H8 • l l l l l l l l l l l l 14 1011 Tributary Stations 13 5 15 Control Wet Season (Nov-Mar) Dry Season (Jun-Aug) A A F T i—r i—i—r i—r 16 2 3 4 12 6 7 Sumas mainstem (upstream -> downstream) g 8 Sumas Lake Canal Sampling Station 40 -I 20 H ^-v 40 -1 _J c o k— CO o o 'c co O) o c •o CD > o CO CO Q 20 H 14 1011 13 5 15 Tributary Stations Control 14, it - A — Wet Season (Nov-Mar) - ® — Dry Season (Jun-Aug) A * F T i—r 16 2 3 4 Sumas mainstem (upstream --> downstream) ~i—i—r 12 6 7 1 1 9 8 Sumas Lake Canal Sampling Station 150 Appendix F: 3 Sampling Dates, Fecal Coliform Results 2000 - | 1000 H m t \ 4 l l "t l — l — l T l I — l T I 6000 - i 4000 H E o o D LL O V) E & 2000 -I o O 15 o CD 14 10 11 Tributary Stations 13 5 14,11 Sampling Date —A-— July 26, 1994 —<fj&— August 24, 1994 — + — February 1, 1995 1—r 16 2 3 4 12 6 7 Sumas mainstem (upstream --> downstream) Sampling Station g 8 Sumas Lake Canal E o o D O o o o o o +-» Q. CD (fi "co o CD 10000 5000 H 6000 - i U_ 4000 H 2000 H 0 T + / 14 10 11 Tributary Stations 13 5 14,11 Sampling Date :— July 26, 1994 —i$— August 24, 1994 — + — February 1, 1995 16 2 3 4 12 6 7 9 8 Sumas mainstem (upstream --> downstream) Sumas Lake Canal Sampling Station 151 Appendix G: Historical Wet Season Water Quality. All stations on the mainstem are plotted. Wet Season = November through March. Historical Wet Season Values Orthophosphate Sumas mainstem 1.0 •ft) E, .8 CO to .6 Q . 8 .4 x: 8- .2 j§ 0.0 1970 Rsq= 0.1978 1980 1990 2000 YEAR Historical Wet Season Values PH Sumas mainstem x Q . 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 a • mm ! • i : - r — " H i — 1970 1980 1990 2000 Rsq = 0.0190 YEAR Historical Wet Season Values Nitrate-N Sumas mainstem 6' Z 1. 0J 1970 1980 1990 2000 Rsq =0.4092 YEAR 152 Appendix G: Historical Wet Season Water Quality. All stations on the mainstem plotted. Wet Season = November through March. Historical Wet Season Values Ammonia-N Sumas mainstem Rsq= 0.0928 1970 1980 1990 2000 YEAR Historical Wet Season Values Dissolved Oxygen _ Sumas mainstem 14 Rsq = 0.0290 1970 1980 1990 2000 YEAR Historical Wet Season Values Conductivity Sumas mainstem 1970 Rsq = 0.0042 1980 1990 2000 YEAR 153 Appendix G: Historical Wet Season Water Quality. All stations on the mainstem are plotted. Wet Season = November through March. Historical Wet Season Values Chloride Sumas mainstem 50' 5" 40' 30' cu 1 »• 5 io> P f L i ^ I T T H -o l _ 1970 • I 1980 1990 YEAR 2000 Rsq =0.1824 154 Appendix H: Scatterplots of Historical Water Quality vs. Discharge. All seasons at stations 3 and 4 only are plotted. Discharge is as measured at station 3 (U.S. border). 1.2 ^ 1.0 "S> E 8 ro 0 ' c o E 4 E < .2 0.0 y • • y y y y y y y • • y « • " y y II y y B y y ^ y" • • 10 20 Flow (m3/s) 30 Rsq = 0.5605 30 E, <D O 10 10 20 Flow (m3/s) 30 Rsq = 0.1702 6.0 5.0 O) 4.0 E, Z 3.0 i 0) i= 2.0 1.0 0.0 B B i" " • BjB b y B N S »N \ ' \ \ \ \ \ 10 20 30 Rsq = 0.1742 Flow (m3/s) 155 Appendix H: Scatterplots of Historical Water Quality vs. Discharge. All seasons at stations 3 and 4 only are plotted. Discharge is as measured at station 3 (U.S. border). 400 E o co o sz E 3 300 200 o 3 T3 C o O 100 • • • m. • * — * • • • 10 20 30 Rsq = 0.2380 Flow (m3/s) •4 | .3 s CO co o sz C L SZ .1-O o.o. rf» % ^  v 10 Flow (m3/s) 20 Rsq = 0.5632 156 Appendix I Nitrogen Model Methodology by Brisbin (1995) The following description of the nitrogen model (see figure at end of this appendix) is adopted from Brisbin (1995): The model relies on a number of input variables: livestock inventory; unit livestock nutrient production; manure management practices and associated nutrient losses; agricultural land base inventory; unit crop nutrient uptake; inorganic fertilizer use; and soil-atmosphere nitrogen exchange factors. The model utilizes these variables in a series of calculations to estimate the losses of nutrients to the atmosphere, surface water and groundwater. The calculations are described by the following steps: 1) Unit livestock nutrient production estimates are applied to livestock inventory values to generate total manure nutrient production by livestock type and commodity. 2) Nutrient loss factors (the percentage of the nutrient "lost" during a particular component of the manure management process, to be applied to the total amount of the nutrient entering that component of the system) for various manure management system components are prorated by the distribution of the management system components to generate composite loss factors for each commodity group. The composite loss factors are then applied to the total manure nutrient production for each commodity group to generate estimates of nutrient losses which occur at different steps of the manure management process and a net application of manure to land. The net application to land includes that "applied" by livestock on pasture. Losses during the manure management process include losses to the atmosphere, losses to surface water, losses to groundwater and export. The model estimates losses which occur during housing and collection, from yard areas and pasture, from storage and during land application. Export losses refer to nutrients which are utilized in such a manner that they are not applied to the agricultural land of the area to which the model is being applied. 3) Unit crop uptake and inorganic fertilizer application values are applied to the land base inventory and a value for crop nutrient uptake minus inorganic fertilizer application is calculated. 4) An estimate of the soil-atmosphere nitrogen exchange is made utilizing estimates of a 157 background net input to soil plus estimates of a return flow from agricultural activities which is calculated as a percentage of the total losses to the atmosphere during the manure management process (denitrification losses are not used in this calculation). 5) The values for total manure nutrient production (amount excreted) and manure management losses are combined with the crop - inorganic fertilizer application balance and the soil-atmosphere balance to generate an estimate of the surplus (or deficit) applied to the soil. A surplus value does not necessarily mean that excessive amounts of nutrients are being applied to the soil. The term "surplus" in this case means only that the nutrients produced in manure less manure management losses plus inorganic fertilizer applications plus net input from the atmosphere exceed crop nutrient uptake. "Losses" of nutrients from soils is part of the various nutrient cycles and cannot be eliminated from agricultural systems; "no surplus" is simply not attainable. Surplus applications, as defined in this study, must be interpreted as excessive or not and this interpretation must consider their ultimate destination (surface water or groundwater) and the sensitivity of that destination to nutrient loading. 6) Surplus applications to the land are then partitioned into losses to denitrification, surface water and groundwater using various soil release factors. Denitrification losses are calculated as a percentage of the "net manure application". "Deep losses" (nutrients which move below the rooting zone) are calculated as the surplus application less denitrification losses and are split between losses to surface water and losses to groundwater. The model has the ability to estimate the exchange of nutrients within the soils (releases from organic matter within the soil and immobilization within the soil). However, in this study it was assumed that the net of this exchange would be zero; the amount released from the soil equals the amount immobilized by the soil. When the model is applied to a particular geographic area it is being assumed that the rate of mobilization (mineralization) within that geographic area equals the rate of immobilization within the area; this does not imply that the two rates are equal for all locations within that area, only that they are equal over the entire area. Nutrient exchanges in the soil can be very significant over the short term; however over a long period of time a soil system will tend to approach an equilibrium where annual rates of immobilization equal rates of mobilization. 158 X T 3 C CD Q . C L < ^ Q . •"g 2 £ « « £ Q 2 E '— CD o E » X . t i f V U J z CO cu II 1 CO Soi Los o c *•* + CD c CD E s CD O) "3 CD - z Man C O o o T 3 O SZ a) Co O C D CD .5= C D - Q O W o o 0 4= 1 "S I o T 3 CD u c c o o c CD CL E o P £ 1 \ O ' 5 c r CD cm c T3 CD CD O (J) •i= CD Z Z + = + + o o c c CD CD > > c c CL CL g 2 o o CD <= £ CD 3 O) C P CO £z o cT o o to cz CD CD > > Li CD posh ctor LUO CD L L O 12 o T5 CO 00 CO o {= CD E E CD CD ste TO C CO CO 2 159 E CD (A 0) D) n k_ > A CD o c ro ra 03 S" o 'E ra c >. 73 C CO -Q 0 1 E 10 co ro •n in 3) Q. E 3 J5 co it. ."8 C CD co c T3 „ € i 8.(2 co 5 ^ (0 io 0) T-Si £ I a. a. ra •a cu CO P 8 fen O . Z ra-° i a> co CM 8 9 CO is 5 ra E i # 1 c CB g -M c ,o CO 1-8 -2. | | 3 C CO IB CM O z co I co 1 3 CD CO •ia It 8 2: I CO CO CO c .o E CO CO 8-T3 c c CD 8-CO CD •s a 13 CD CD u c jo IB -Q c o CO 3 CO c CD O CD 1 3 c CB CD O c >> c o CO CO 8 1 3 CD CO 3 ra 3 O IB O .<a 160 Appendix K Variables Used in the Spearman Rank Correlation (94 variables and 11 cases written) Variable Name Description CA Contributing Area Number, not used in correlation ID only Water Quality PHWET Wet Season pH PHDRY Dry Season pH CONDWET Wet Season Conductivity CONDDRY Dry Season Conductivity DOWET Wet Season Dissolved Oxygen DODRY Dry Season Dissolved Oxygen ORTHOPWE Wet Season Ortho-P ORTHOPDR Dry Season Ortho-P CLWET Wet Season Chloride CLDRY Dry Season Chloride AMMONWET Wet Season Ammonia-N AMMONDRY Dry Season Ammonia-N NITWET Wet Season Nitrate-N NITDRY Dry Season Nitrate-N Land Activity Intensity EA_N EA_N_GIS EA_N_FM EA_N_CRP WMS_N WMS_N_GI WMS_N_FM WMS_N_CR AUETOT AUEDENS PLTRYTOT DCOWTOT Total Surplus Nitrogen, calculated using WMS and Census data Surplus nitrogen per total hectares Surplus nitrogen per farmed hectares Surplus nitrogen per cropped hectares Total surplus nitrogen, calculated using WMS data and averages Surplus nitrogen per total hectares Surplus nitrogen per farmed hectares Surplus nitrogen per cropped hectares Total Animal Unit Equivalents Animal Unit Equivalents per farmed hectare Total number of poultry Total number of dairy cows 161 PIGTOT Total number of pigs NO FARMS Total number of farms Land Use Types LU1TOT Total hectares agriculture LU2T0T Total hectares forest LU3TOT Total hectares none perceived land use LU4TOT Total hectares park/wetlands/recreation LU5TOT Total hectares residential LU6TOT Total hectares industrial/commercial LU1PRCNT % area agriculture LU2PRCNT % area forest LU3PRCNT % area none perceived land use LU4PRCNT % area park/wetlands/recreation LU5PRCNT % area residential LU6PRCNT % area industrial/commercial Surficial Soil Texture (approx., first 25 cm) SURF1TOT Total hectares organic SURF2TOT Total hectares clay SURF3TOT Total hectares loam SURF4TOT Total hectares silt SURF5TOT Total hectares sand SURF6TOT Total hectares gravel SURF1 PRC % area organic SURF2PRC % area clay SURF3PRC % area loam SURF4PRC % area silt SURF5PRC % area sand SURF6PRC % area gravel Subsurface Soil Texture (approx. below 25 cm) SUB1TOT Total hectares organic SUB2TOT Total hectares clay SUB3TOT Total hectares loam SUB4TOT Total hectares silt SUB5TOT Total hectares sand SUB6TOT Total hectares gravel SUB1PRCN % area organic SUB2PRCN % area clay 162 SUB3PRCN SUB4PRCN SUB5PRCN SUB6PRCN % area loam % area silt % area sand % area gravel Soil Drainage Capability DRN1TOT DRN2TOT DRN3TOT DRN4TOT DRN5TOT DRN6TOT DRN1PRCN DRN2PRCN DRN3PRCN DRN4PRCN DRN5PRCN DRN6PRCN Total hectares excessive Total hectares well Total hectares moderately well Total hectares imperfect Total hectares poor Total hectares very poor % area excessive % area well % area moderately well % area imperfect % area poor % area very poor Parent Material PM2TOT PM3TOT PM4TOT PM5TOT PM6TOT PM7TOT PM10TOT PM2PRCNT PM3PRCNT PM4PRCNT PM5PRCNT PM6PRCNT PM7PRCNT PM10PRCN Total hectares colluvium Total hectares outwash Total hectares alluvium Total hectares glacio-fluvial Total hectares lacustrine Total hectares glacial till Total hectares organic % area colluvium % area outwash % area alluvium % area glacio-fluvial % area lacustrine % area glacial till % area organic 163 E 8 CM 8 CM CM O CO s o o O O O h-CN O -*-> LU < 0 Z 001 E E 5 CO CO CO CO E a i— • o. IS1 I si CO '51! IS] 5 -5f m o CD CM CO CO CD I C M 05 o o CM ISI in oo cnt 1 0.021 0.04 0.01 0.03 0.00 0.021 0.201 0.18 Q. CO ZJ cnt I 0.00 0.06 0.00 0.05 0.03 0.01 0.33 [Iu5pr •*—» c p 0.00 0.00 0.13 0.00 0.00 0.001 0.01 |lu4pt •cnt | 0.08 0.04 0.00 0.01 0.02 0.02 0.02 0.03 0.091 a CO _3 -*—' c p 0.30 0.89 0.69 0.32 0.29 0.03 0.82 Ilu2pr I lulprcnt 1 0.91 I 0.56 I 0.09 I 0.10 I 0.90 j I 0.63| 0.681 0.991 0.95| 0.331 0.001 |lu6tot I 7.48 j I 67.321 CO co' I 27.4| 19.641 CM CO 11.6| |lu5tot I 0.92 j I 110.321 I 0.16| I 32.081 CO m I 9.521 556.641 •*-» o _3 1 5.24| CN I 57.921 I 55.881 1 4.52| 1.32| 21.121 |lu3tot r-co I 80.32] I 2.921 I 6.041 I 5.241 I 96.44| I 11.121 36.48| 156.36| lu2tot I 545.56| I 796.41 I 308.44I I 24.881 I 1862.4] 185.641 49.28| 54.441 lultot I 437.28 j 1025.16 78.72 42.8 I 1084.121 I 3708.84I 1 437.161 186.12 1208.04 549.2 0.12| |CA m CO CO a> o CM co in |CA c o ajo uv t CO c o Q I O § o d 1° § CM d CN CJ> CO CN CD CO CO 15 CN CN 8 oo oo CN O CM 164 to CN CN CM CN CN CN CO CO CO CO CN 165 3 -Q C O o >. CO cv cu o 'a c X TJ C CU Q_ < 166 cu sz cu -*—' CO a) o co c +-» .Q -* ' c o O >> .Q CO cu 3 CO > cu o TO c X TO C cu C L C L < 2 5 3 3 8 <0 ro u <d < E « o < O c ro O CD CO < £ 01 ro lor ro , - I « "a t> cu m CL < > J O CU I < i> ro ] o <" 2 z a- < 5 ±i i O a) O ) < < * CO LH " O < § .ro ro </> XJ CL Q ->- CL 5 ° oo ro CL TD co E CO CM S ; cu E c -*—' c cu 1 . . CO 10 = 0) CO . 2 CD T3 C > in co 1 jg" oo ro Z $ <D ro o c S CO o •£ ^ o> cu b oo cu 3 CO cu « cu ro oo cu T3 ro CO c: oo oo ro c CD > CD CO £ «> CU -£ . . 0 ro T3 >> CO II <2 7 3 Q E <o o c c oo cu ?> 00 CU ^ I .« it o cu II o to in o d v CL » r N CO, 168 o . o < 0_ co d 2 O CO Q. JS o £ . cu < CL E TZ CO 8.2 E P to eg ; O co A. 8 jo ™ O JS o CO E 2 a « O CO ' CO O <u 1 « O o 1^ CD d 3 CO CD d d in oo d in co d d co d oo d CD d CN d CN N-d 00 0 5 , o CO in d CD d m d CO d ft c o CO S JS °> £ 8 "5 o to s « « t u H=- O 4 ¥ v > c: CL IQ tu "D 'C o sz O 169 <= T5* O Q) o .E D- c c o ro ro g ro < 3 .£> C c o O >» -Q «/> <D O O CO 00 > ro 3 a ro CD a. CO Q) ro a) a. CL < g ) CO O o cu co CD CO s ro CO o ro > ro LL ro 6 "ro I ° (0 -*-• ' ro 2 CO > =j ro 3 ° o o Q. . cu > >r CO O CU Q . CO £> ro a) o > *-( O d CD O <D d CD d CNI d Pi d d co CD d 1^ CO d SS1 00 d CO d c OB <D^  7 3 CU > 0 1 lb |Q <u . 170 E ™ 31 E CO e O D> O .£ g =§ c o ra O d) » CU 0) CO CL o £ CO J2 o o> o K =5 CD C w J3  5 5 u aS |o " p # O B I * co as E ™ b # o s o as O 2 CL 2-. (0 _ . ' CL <| !Li_* 171 is l i CL * |CL S a) re 8 :o as a |C3 B ICQ SS ICQ 3 ICQ as co B . CL CL a-:s 1 o li . z z # S i CO c CL 31 » as 5 m • » S CL < I 172 CO •2 § <= & o w ~ CO ja a> 2 < o co 0 £ I 1 2 | <= <? ra o 1 >• ra - ° 0 in o. a W .2 X I c I/) — CL O IE CO ro fl> » o or TJ ra a> o n c ^ co J3 o # Io a E 2 A g U — 3 # o # O B E Q_ 9> m • E • o 173 CD 7 3 •c o sz O Q > "•8 7 3 C o O Q > 3 7 3 C o • o : Q_ O Q o Is s 174 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items