UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Imperviousness and trace metals in stream sediments : urbanisation in the Lower Fraser Valley between… Iwata, Oh 2007

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-ubc_2007-0442.pdf [ 25.52MB ]
Metadata
JSON: 831-1.0101111.json
JSON-LD: 831-1.0101111-ld.json
RDF/XML (Pretty): 831-1.0101111-rdf.xml
RDF/JSON: 831-1.0101111-rdf.json
Turtle: 831-1.0101111-turtle.txt
N-Triples: 831-1.0101111-rdf-ntriples.txt
Original Record: 831-1.0101111-source.json
Full Text
831-1.0101111-fulltext.txt
Citation
831-1.0101111.ris

Full Text

IMPERVIOUSNESS A N D TRACE M E T A L S IN S T R E A M SEDIMENTS: URBANISATION IN THE LOWER FRASER V A L L E Y B E T W E E N 1973-2006 by Oh Iwata B . S c , The University of Tokyo, 2005 A THESIS S U B M I T T E D I N P A R T I A L F U L F I L M E N T OF T H E R E Q U I R E M E N T S F O R T H E D E G R E E OF M A S T E R OF S C I E N C E in T H E F A C U L T Y OF G R A D U A T E S T U D I E S (Resource Management and Environmental Studies) T H E U N I V E R S I T Y OF B R I T I S H C O L U M B I A August 2007 © O h Iwata 2007 Abstract Impervious surfaces have been used as a good indicator of urban impact on aquatic health because urban runoff impacts the biodiversity of aquatic organisms. The impact of urbanisation on metal concentrations in sediments was examined in 28 sub-watersheds in the Lower Fraser Valley. GIS techniques were used to determine rate of urbanisation in 1973 and 2006 and the metal concentrations in sediments in 2006 were then compared with historic data collected at the same sites. In 2006 the highest imperviousness was observed in the Alouette and N / W Vancouver sub-watersheds (50-70% and 40-60%). A l l sub-watersheds experienced increases in imperviousness, (A l-30%o), which led to increases in trace metals concentrations in sediments (e.g., Cu , Pb, Zn , N i ) . Z n and N i were the most sensitive metals related to forest loss, increases imperviousness, and traffic, whose impacts were most evident within a 100 m riparian buffer zone. The % imperviousness forms a curvilinear relationship with aquatic biodiversity indices, with a threshold of 10-15 % of imperviousness in the early stages of urbanisation. A n opposite curvilinear relationship exists between % imperviousness and several trace metals. The rate of increase in metals tends to be significantly higher in the early stages of urbanisation but flattens out once high level of imperviousness was achieved, possibly due to sediment saturation. Several factors account for significantly higher metal levels in N / W Vancouver than in Alouette, two sub-watershed with similar imperviousness: 1) higher traffic density, 2) enhanced storm sewer system connections, and narrower vegetative stripes within a buffer zone, 3) abundant phosphorus, and 4) higher organic matter. Imperviousness is not the direct cause of stream degradation, but is considered the conveyance system that brings pollutants directly into the stream. Complex interactions between these and other factors all influenced the high metal levels, and traffic was likely the principal source. K E Y W O R D S : Imperviousness, trace metal, stream sediment, land use, automobile traffic, zinc, riparian buffer zone, Lower Fraser Valley, GIS i i Table of Contents Abstract i i Table of Contents i i i List o f Tables v i List o f Figures ix Acknowledgements x i i 1 Introduction 1 1.1 Urbanisation and Population Growth in Lower Fraser Val ley 1 1.2 Land Use Activities & Pollution of Streams 1 1.3 Sediment Contamination and Land Use Quantification 2 2 Literature Review 4 2.1 Metal Pollution Problem in Urban Stormwater Runoff 4 2.1.1 S ource of metals 4 2.1.2 Typical metal concentrations in water and sediments 5 2.1.3 Seasonal and storm event effects (first flush) 9 2.2 Imperviousness as an Indicator 11 2.2.1 Total imperviousness and effective imperviousness 11 2.2.2 How imperviousness is measured 12 2.3 Interactions between Imperviousness and Water Quality and Stream Health 13 2.3.1 Imperviousness and aquatic biota 13 2.3.2 Imperviousness and water quality 14 2.3.3 Imperviousness and sediments 14 2.4 Stormwater Management 17 3 Methods 20 3.1 Study Area 20 3.2 Land Variable 23 3.2.1 Topography and mineral texture 23 3.2.2 Land use 23 3.2.3 Land cover 24 3.2.4 Buffer zones 24 3.2.5 Road density 24 3.2.6 Historical changes in land variables 25 3.3 Sediment Quality 25 3.3.1 Field methods 25 3.3.2 Sample preparation and analysis 26 3.3.3 Analytical techniques '. 27 3.3.4 Sediment quality data in 1973 29 3.3.5 Sediment quality guidelines 30 3.3.6 Historical changes in sediment quality data 30 i i i 3.4 Statistical Methods 31 3.4.1 Factor analysis: land variables 31 3.4.2 Factor analysis: macro and micro elements 32 3.4.3 Cluster analysis by sediment quality data 32 3.4.4 Correlation between land variables and sediment quality data 33 3.4.5 Correlation on the historical changes 33 3.4.6 Mann-Whitney U test 33 4 Results and Analyses 34 4.1 Land Variables 34 4.1.1 Topography and mineral texture 34 4.1.2 Land use 37 4.1.3 Land cover 44 4.1.4 Land use in a 100 m buffer zone 46 4.1.5 Road density 50 4.1.6 Historical changes in land variables 50 4.1.7 Factor analysis on land variables 53 4.1.8 Summary of the section 1 57 4.2 Sediment Analysis 58 4.2.1 Aqua regia and loss on ignition 58 4.2.2 Historical changes in sediment quality 63 4.2.3 Factor analysis on sediment quality 65 4.2.4 Cluster analysis by sediment quality data 67 4.2.5 Summary of section 2 69 4.3 Land Use-Sediment Interaction 69 4.3.1 Linking imperviousness with metal concentrations 69 4.3.2 Relationships between Imperviousness vs. metal concentrations in the 2006 sediments in the Alouette and N / W Vancouver sub-watersheds 72 4.3.3 Correlations between land variables and sediment quality data 76 4.3.4 Mann-Whitney U test 80 4.3.5 Summary of the section 3 84 5 Discussion 86 5.1 Land use patterns and changes 86 5.2 Sediment metal levels and changes 86 5.3 Relationships between land use and metals 88 5.3.1 Imperviousness and zinc as indicators of urbanisation 88 5.3.2 Trace metals from atmospheric sources 89 5.3.3 Effectiveness of a buffer zone 89 5.3.4 Multivariate relationships between land use and metals 90 5.3.5 Curvilinear relationships between imperviousness and metal increases ..91 5.3.6 Imperviousness as the best indicator? - Alouette vs. N / W Vancouver 92 5.3.7 Cause and effects of stream degradation - biotic integrity and metals 95 5.3.8 Bioavailability of sediment metals 97 5.3.9 Analytical accuracy of the laboratory procedures in the past 97 iv 6 Conclusions 99 Bibliography 101 Appendices I l l A . Chemistry of Sediment Quality I l l B . Methodology for sediment quality analysis. 124 C. Aerial Images 128 D . Land Use •. 137 E . Land cover , 150 F. A 100 m Buffer zone 156 G . Road Density 160 H . Historical Changes in Land Use/cover 161 I. Factor Analysis on Land Variables 162 J. Analysis for Sediments and Water Samples : 164 K . Factor Analysis on Sediment Quality Data 169 L . Cluster Analysis on Sediment Quality Data 170 M . Mann-Whitney U test among the clustered sub-watersheds 171 v List of Tables Table 2-1 Table 2-2 Table 2-3 Table 2-4 Table 2-5 Table 2-6 Table 2-7 Table 2-8 Table 2-9 Table 2-10 Table 2-11 Table 2-12 Table 2-13 Table 3-1 Table 3-2 Table 4-1 Table 4-2 Table 4-3 Products and uses of trace metals. Source: Moore and Ramamoorthy (1984); Minton (2005); W H O (2007) 4 Streambed total metal concentrations (median) in the Brunette River and Stil l Creek sub-basins in 1993. Bo ld numbers mean the significant difference between the basins at the 5% probability level. Units are mg/kg except for Fe (%) (dry weight). Adapted from McCa l lum (1995) 6 Trace metal concentrations in streambed sediments in various studies. Units are mg/kg except for Fe and A l (%) (dry weight) 7 Element concentrations in the sediments in agricultural watersheds. A l l units are mg/kg except for iron (%). N / A is "not available". Modified from Smith (2004). 7 Concentrations of trace metals in surface runoff from recent studies. Units are pg/L, and median values are shown 8 Trace metals in precipitation collected from the Brunette Watershed. These data are from 36 weekly samples collected between Jan. 31 - Dec.5, 1997. Adapted from Hal l etal . 1999 9 Presumed relationship between imperviousness and land use. Modified from Dinicola (1990); Booth and Jackson (1997) 11 Percentage of imperviousness for various land cover classes. Modified from Brabecetal . (2002) 12 Trace metal concentrations in surface water in relation to urban land use types. Units of trace metals are ug/L, and those of land use are % 14 Percent land use in the Brunette River Watershed (total area). Adapted from McCa l lum (1995) 15 Percent land cover in the Brunette Watershed (total area). Adapted from McCa l lum (1995) 15 Trace metal concentrations in the sediments in the Brunette River (median values, n=32). Bold numbers mean the values of 1993 are significantly different from those of 1973 at the 5% probability level. Adapted from McCa l lum (1995) 15 The best management practices in stormwater management. Modified from Zandbergen et al. (2000) 18 Elements of focus in 1973 and 2006 27 Land variables used for factor analysis. Colluvium/ti l l means colluvium/glacial till-geology, and glaciomarine means glaciomarine deposits-geology. Road-density was used for only 2006 data 32 Topography and mineral texture in studied sub-water sheds 37 Factor loadings of varimax rotation and rotation sums of squared loadings for 1973 land variables. Correlations higher than 0.5 (as an absolute value) are shown, and bold presents correlation>0.7. (B) denotes the land use/ cover within a 100 m buffer zone. Communality means the proportion of each variable explained by the extracted factors '. 55 Factor loadings of varimax rotation and rotation sums of squared loadings for 2006 land variables. Correlations higher than 0.5 (as an absolute value) are shown, and bold means correlation>0.7. (B) denotes the land use within a 100 m buffer zone. Communality means the proportion of each variable explained by the extracted factors 56 vi Table 4-4 Comparison of sediment quality between this study and Brydon (2004) at the same sites (v5 and v6) 58 Table 4-5 Duplicates data of laboratory analysis of sediment quality for selected sub-watersheds (a2, k4 and t3). N A of C d shows the data were below the detection limit 60 Table 4-6 Component loadings and rotation sums of squared loadings for sediment quality data in 1973. Only correlations greater than r= 0.5 are shown, and bold means a correlation r=>0.7. Communality means the proportion of each variable explained by the extracted factors 66 Table 4-7 Factor loadings with varimax rotation for sediment quality data in,2006. Only correlations greater than 0.5 (as an absolute value) are shown, and bold presents a correlation>0.7. P - phosphorus, and O M - organic matter. Communality means the proportion of each variable explained by the extracted factors 67 Table 4-8 Clustering sub-watersheds according to factor scores of sediment quality data in 2006 68 Table 4-9 Spearman's rank correlation between land variables and sediment quality data in 1973. (B) denotes "a 100 m buffer zone". Only significant correlations are shown (p < 0.05), and bold presents that correlation is significant at the 0.01 level 77 Table 4-10 Spearman's rank correlation between selected land variables and sediment quality in 2006. (B) denotes "a 100 m buffer zone". Only significant correlations are shown (p < 0.05), and bold presents that correlation is significant at the 0.01 level. 78 Table 4-11 Spearman's rank correlation for historical change data. (B) denotes "a 100 m buffer zone". Only significant correlations are shown (p < 0.05), and bold presents that correlation is significant at the 0.01 level 80 Table 4-12 The clusters of the sub-watersheds for Mann-Whitney U tests (2006 data) 81 Table 4-13 Median values of each cluster. Parenthesis denotes land use in a buffer zone 81 Table 4-14 Median values of each cluster. Units are mg/kg dry weight 82 Table A . 1 Primary minerals and their characteristics. Modified from Egger (2007) 112 Table A . 2 Classification of metals and donor atoms of complexing sites by their hard and soft characters. Modified from Buffle and de Vitre (1993) 114 Table A . 3 Range of concentrations of ligands or complexing sites in natural fresh waters. = S-OH refers to inorganic solid surface sites. - C O O H and - O H refer to total concentrations of carboxyl and phenolic sites in natural organic matter. N o r g , S o r g refer to total concentrations of organic nitrogen and sulfur. Modified from Buffle (1984) 114 Table A . 4 Solubility product of hydroxides, phosphates and sulfides (Lindsay 1979) 122 Table A . 5 Solubility product of fluorides, iodates and oxalates (Lindsay 1979) 122 Table A . 6 Stability constants of organic-metal compounds (Furia 1972; Smith et al. 2001; George Eby Research 2006) 123 Table A . 7 Lowest readable limits of I C P - A E S and average blank concentrations in this study. 124 Table A . 8 Comparison of the and <63 um <177um sediment particle fractions of the Brunette Watershed sediments. Adapted from McCa l lum (1995) 125 Table A . 9 Median concentrations differences and significance of ranked t-test for comparison of aqua regia versus nitric/perchloric acid digestions (n=14), by using the Brunette Watershed sediments. Adapted from McCa l lum (1995) 126 Vll Table A . 10 Median concentrations differnces and significance of ranked t-test for comparison of atomic absorption (flame A A ) and I C P - A E S detection techniques (n=12), by using the Brunette Watershed sediments. Adapted from McCa l lum (1995) 126 Table A . 11 Sediment quality guidelines in Canada and the United States 127 Table A . 12 Land use proportion of the each of the watersheds in 1973 (%) 148 Table A . 13 Land use proportion of the each of the watersheds in 2006 (%) 149 Table A . 14 Land cover proportion of the each of the watersheds in 1973 (%) 154 Table A . 15 Land cover proportion of the each of the watersheds in 2006 (%) 155 Table A . 16 Proportion of land use in a 100 m buffer zone in 1973 (%) 156 Table A . 17 Proportion of land use in a 100 m buffer zone in 2006 (%) 157 Table A . 18 Proportion of land cover in a 100 m buffer zone in 1973 (%) 158 Table A . 19 Proportion of land cover in a 100 m buffer zone in 2006 (%) 159 Table A . 20 Road Density in the sub-watersheds in 2006 (km/km 2) 160 Table A . 21 Historical Changes in Land Use/cover (%) 161 Table A . 22 Factor component loadings of promax rotation on land variables in 1973 162 Table A . 23 factor component loadings of promax rotation on the data 163 Table A . 24 Results of chroma after loss on ignition for sediments and nitrate for surface water. N A represents that samples that samples were not collected 164 Table A . 25 Sediment quality data in 1973 165 Table A . 26 Sediment quality data in 2006 in this study 166 Table A . 27 Historical changes in sediment quality from 1973 to 2006 (% change from 1973). 168 Table A . 28 factor component loadings of promax rotation on the data 169 v i i i List of Figures Figure 2-1 Metal concentrations in streambed sediments, and land use in the Puget Sound Lowland (PSL) in Washington State. FR, D V and H W Y denote forest, developed and highway, respectively. Concentrations are dry weight. Modified from Brandenberger et al. (2003) 6 Figure 2-2 Relationship between watershed imperviousness and benthic-biological integrity in the Puget Sound Lowland (PSL) streams. Modified from Booth et al. (2004). 13 Figure 2-3 Element concentrations (dry weight) in streambed sediments in stormwater facilities in Lower Fraser Valley. Left bar represents sediments at the inlet, and right bar represents outlet ones. SD denotes 52Ave/221A Street Detention Facility (Langley). Griffin denotes Griffin Park Biofiltration Pond (North Vancouver). H W Y denotes Westview Interchange Detention Pond (North Vancouver). Tempe denotes Tempe Heights Pond (North Vancouver). Oakalla denotes Oakalla Biofiltration System (Burnaby). Parenthesis means imperviousness in a collective drainage. Modified from Brydon (2004) 16 Figure 2-4 Imperviousness vs. metal concentrations in urbanised watersheds in Alaska. Spearman's rank correlation is shown (r s). Modified from Ourso and Frenzel (2003) 17 Figure 3-1 The location of the five study area (Alouette, Hatzic, Kanaka, Thornhill, N / W Vancouver) 22 Figure 4-1 Alouette sub-watersheds 35 Figure 4-2 Hatzic sub-watersheds 35 Figure 4-3 Kanaka sub-watersheds 35 Figure 4-4 Thornhill sub-watersheds 35 Figure 4-5 W Vancouver drainage areas 36 Figure 4-6 N / W Vancouver drainage areas 36 Figure 4-7 A n aerial photo of the sub-watersheds a l , a2 and a3 in Alouette in 1973 39 Figure 4-8 A n orthophoto of the sub-watersheds a l , a2 and a3 in Alouette in 2006 39 Figure 4-9 A n aerial photo of the sub-watershed a4 in Alouette in 1973 40 Figure 4-10 A n orthophoto of the sub-watershed a4 in Alouette in 2006 40 Figure 4-11 Land use in the sub-watersheds a l , a2 and a3 in Alouette in 1973 41 Figure 4-12 Land use in the sub-watersheds a l , a2 and a3 in Alouette in 2006 41 Figure 4-13 Land use in the sub-watershed a4 in Alouette in 1973 42 Figure 4-14 Land use in the sub-watershed a4 in Alouette in 2006 42 Figure 4-15 Land use proportion in the whole sub-watersheds in 1973 43 Figure 4-16 Land use proportion in the whole sub-watersheds in 2006 43 Figure 4-17 Land cover in the sub-watersheds a l , a2 and a3 in Alouette in 2006 44 Figure 4-18 Land cover in the sub-watershed a4 in Alouette in 2006 45 Figure 4-19 Land cover proportion in the whole sub-watersheds in 1973 45 Figure 4-20 Land cover proportion in the whole sub-watersheds in 2006 46 Figure 4-21 Land use within a 100 m buffer zone in the sub-watershed a4 in Alouette in 2006. 47 Figure 4-22 Land use proportion within a 100 m buffer zone in 1973 48 Figure 4-23 Land use proportion within a 100 m buffer zone in 2006 48 Figure 4-24 Land cover proportion within a 100 m buffer zone in 1973 49 Figure 4-25 Land cover proportion within a 100 m buffer zone in 2006 49 Figure 4-26 Road density and imperviousness in studied sub-watersheds in 2006 50 ix Figure 4-27 Historical changes in land use in the whole sub-watershed (1973-2006). Note that the sub-watersheds at the headwaters are shown, yet h4 is shown instead of h3 (and so forth) 51 Figure 4-28 Historical changes in land use within a 100 m buffer zone (1973-2006) 52 Figure 4-29 Historical changes in land cover in the whole sub-watershed (1973-2006) 52 Figure 4-30 Historical changes in land cover within a 100 m buffer zone (1973-2006) 53 Figure 4-31 Eigenvalues of the principal component factor analysis for land use in 1973, and those of the Broken stick model 54 Figure 4-32 Comparison of the sediment quality data of this study with those of McCa l lum (1995) and Brydon (2004). Study was done at the same sites (v5 and v6) in Brydon (2004) (denoted as v5-JB and v6-JB) as well as SD, Tempe and Oakalla in 2006. Sediment quality study was also done in the Brunette River sub-watershed in 1993 (Still C . and Brunette) 59 Figure 4-33 Metal concentrations in the sediments in 1973 and 2006, and threshold effect level (TEL) and probable effect level (PEL) of the guidelines in Canada and U.S . 62 Figure 4-34 Historical changes in sediment concentrations 64 Figure 4-35 Dendrogram of the Ward method of cluster analysis. Factor scores of 2006 sediment quality data were used 68 Figure 4-36 Historical changes of imperviousness and metal concentrations between 1973-2006. Arrows represent the changes from 1973 to 2006. " O " denotes data in 2006, and "+" denotes that in 1973 70 Figure 4-37 Imperviousness vs. element concentrations in 2006 (Cu Pb Z n Cd). A symbol "a" denotes Alouette sub-watersheds, and "v" denotes N / W Vancouver sub-watersheds 73 Figure 4-38 Bar charts and Mann-Whitney U tests for land variables among the clustered sub-watersheds (2006). The median and range for each cluster are presented. The unit of road density is km/km 2 82 Figure 4-39 Bar charts and Mann-Whitney U tests for metals in sediments among the clustered sub-watersheds (2006). The median and range for each cluster are presented 82 Figure 4-40 Summary of Mann-Whitney U tests for 2006 data. Results of significant levels at 0.05 level are shown. 1= imperviousness; F=forest-cover; R=road-density; Mc=coarse mineral; Mf=fine mineral. Parenthesis denotes land cover in a buffer zone 83 Figure 5-1 A schematic diagram of the relationships between imperviousness and biotic integrity/ metal levels 96 Figure A . 1 The Watersheds h l -2 (1973-above; 2006-below) 128 Figure A . 2 The Watersheds h3-7 (1973-above; 2006-below) 129 Figure A . 3 The Watershed k l (1973-above; 2006-below) 130 Figure A . 4 The Watersheds k2, k4-6 (1973-above; 2006-below) 131 Figure A . 5 The Watersheds k3 (1973-above; 2006-below) 132 Figure A . 6 The Watershed t l (1973-above; 2006-below) 133 Figure A . 7 The Watersheds t2-7 (1973-above; 2006-below) 134 Figure A . 8 The Drainages v3-4 in 2006 135 Figure A . 9 The Drainages v5-6 in 2006 136 Figure A . 10 The watersheds h l -2 (1973-above; 2006-below) 137 Figure A . 11 The watersheds h3-6 in 1973 138 Figure A . 12 The watersheds h3-7 in 2006 139 Figure A . 13 The watershed k l (1973-above; 2006-below) 140 Figure A . 14 The watersheds k2, k4-6 in 1973 141 Figure A . 15 The watersheds k2, k4-6 in 2006 142 Figure A . 16 The watershed k3 (1973-above; 2006-below) 143 Figure A . 17 The watershed t l (1973-above; 2006-below) 144 Figure A . 18 The watersheds t2-7 (1973-above; 2006-below) 145 Figure A . 19 The drainages v3-4 in 2006 146 Figure A . 20 The drainage v5 in 2006 147 Figure A . 21 The watershed k l in 2006. Only land cover within urban land use (rural-urban, dense-urban, commercial-industrial) is delineated 150 Figure A . 22 The watershed k2 in 2006. Only land cover within urban land use (rural-urban, dense-urban, commercial-industrial) is delineated 151 Figure A . 23 The drainage v5 in 2006. Only impervious-surfaces are delineated 152 Figure A . 24 The drainages v5-6 in 2006. Only impervious-surfaces are delineated 153 Figure A . 25 Eigenvalues of principal component factor analysis for land variables in 2006 and those of broken-stick model 163 Figure A . 26 Eigenvalues of principal component analysis for sediment quality data in 2006 and those of broken-stick model 169 Figure A . 27 Cluster analysis by using factor scores (factor 1, 2 and 3) of 2006 sediment quality data with the Complete Linkage method 170 Figure A . 28 Mann-Whitney U test for each variable among the clustered sub-watersheds in 2006. F W : forested watersheds (hl-h3, t l , k3), R W : rural watersheds (h4, k l - k 2 , k5, t2-t3, t6), U W : urban watersheds (al-a4, k4), V D : Vancouver drainages (v3-v6). The median and range for each cluster are presented 171 XI Acknowledgements I wish to give special appreciation to my supervisor, Dr. Hans Schreier, for all his guidance and help throughout this research. His generous and tremendous encouragement was invaluable in carrying out the whole study. I also greatly thank my research committee members, Dr. Les Lavkulich and Dr. Ken Hal l , for their supervisions as to laboratory procedures and completion of this thesis. I received numerous supports from Richard Boase, at the District of North Vancover, for map and data sources and background information for study areas, and thus I am greatly thankful for his help. M y appreciation goes to many students and staff at the Institute for Resources, Environment and Sustainability, who provided me with numerous help and encouragement, with special thanks to Trudy Naugler, Stephanie Lepsoe, Regina Bestbier, Dr. Sandra Brown, Jennifer MacDonald, Mandeep Purewal, Tashi Tsering, Al ice Cohen and Julie Wilson. I also wish to thank Carol Dyck for her help in the soil laboratory at U B C . I am also thankful to numerous friends at the Green College, U B C , who have always encouraged me, with special thanks to Gaku Ishimura. I would also like to thank Heiwa Nakajima Foundation (Tokyo), which made this research possible through its financial support. The research was supported by the Canadian Watershed Network " Non-point Source Pollution Assessment Project. Finally, I wish to give special appreciation to my parents, who have always trusted my challenge abroad, as well as genuine friends in Japan who have always given me spiritual encouragement. x i i 1 Introduction 1.1 Urbanisation and Population Growth in Lower Fraser Valley Urbanisation has been progressing rapidly over the past decades in the Lower Fraser Valley ( L F V ) , which is the centre of human activities for residential, commercial, industrial and agricultural uses in British Columbia (Dorcey and Griggs 1991; Boyle et al. 1997; Berka et al. 2001). Population has increased and it has expanded to the outskirts of the Greater Vancouver Regional District ( G V R D ) (Hall and Schreier 1996; Tomalty 2002; B C S T A T S 2007). Due to the protection of the agricultural land in the lower part of the L F V most urban development is occurring on steep slopes on the north shore of the Fraser River. The cities of North Vancouver and Coquitlam, located to the north and east side of Vancouver have experienced very rapid population increases (Zandbergen et al. 2000; Alexander and Tomalty 2002). Population increase and residential development are expected to continue in coming decades (Hall and Schreier 1996; G V R D 2004). 1.2 Land Use Activities & Pollution of Streams Urban expansion has resulted in intensified land use activities in all land use sectors (agricultural, commercial/industrial and residential) (Hall and Schreier 1996; Boyle et al. 1997). Most land use activities create non-point sources of pollution (NPS) and these are more difficult to track than the point sources (Cabe and Herriges 1992; Carpenter et al. 1998; Zandbergen et al. 2000). Intensive agriculture, forest harvesting and urban stormwater are the primary sources of N P S pollution, and a number of studies has been conducted to determine land use impacts on aquatic health in the L F V . The influence of agricultural activities has been examined for nutrients originating from fertilizers and manure applications (Vizcarra et al. 1997; Berka et al. 2001; Schreier and Brown 2004), livestock increase (Schreier et al. 2004; Smith 2004). and pesticide and trace metals (Szeto and Price 1991; Smith 2004). Studies would also suggest concerns about bacterial 1 pathogens and antibiotics (Halling-Sorensen et al. 2000; Smith et al. 2002). Linkages have been developed between agricultural land use and stream contamination (Smith 2004) and there is considerable evidence of groundwater contamination (Szeto et al. 1994; Wohl 1996). The impacts of urbanisation on aquatic health have been mainly studied in the Brunette River watershed in the G V R D (McCallum 1995; Zandbergen 1998). Although the pollutants from factories, septic systems, and sewage plants are recognized as point sources, many of these have been regulated (Dorcey and Griggs 1991). NPS pollutants originated from gasoline, oil/grease, paints, excrements of pets, wear of tires and brakes and car emissions which introduce heavy metals, nutrients, pathogens and organic compounds to surface waters (Hall et al. 1976; Hall et al. 1999; Schindler et al. 2006), via stormwater running from impervious surfaces (Zandbergen et al. 2000; Houston 2004). 1.3 Sediment Contamination and Land Use Quantification Heavy metals in water and sediments have been discussed in numerous studies, as they have long been recognized as pollutants originating from urban activities (Hall et al. 1976; Birch et al. 1996; Kelly et al. 1996; Hall et al. 1999; WHO 2007). The health offish and invertebrates have been addressed as the indicators of cumulative impacts of various metals and other pollutants from land use activities (Karr 1981; May et al. 1996; Schreier et al. 1997; Zandbergen et al. 2000). However, identification of aquatic organisms is a laborious task, and they do not tell which pollutant is the actual cause of the environmental degradation (Hickey and Clements 1998; US EPA 2006). Moreover, stream quality is characterized by a complex mixture of numerous factors, e.g., topography, channel morphology, geology, and vegetation distribution, as well as urbanisation (Forman and Alexander 1998; Facchinelli et al. 2001; Jarvie et al. 2002). Individual factors need investigation i f one wants to clarify actual causes and effects of stream degradation. Due to the locally specific factors, it can be misleading to build a linkage between urbanisation and sediment quality in spatially unrelated watersheds. This issue can be addressed by temporal comparison of land use in the same watershed. Several studies have tried to quantifying land use 2 activities (McCal lum 1995; Hal l et al. 1999; Preciado and L i 2006), but detailed land use/cover delineation in urban areas in L F V in relation to sediment quality, is limited. This is also true of historical land use, and the data between two different historic sampling events. The overall objective of this study is to build a linkage between urbanisation and streambed sediment quality, by examining land use and sediment quality at two different times over a 33 year period. The impacts w i l l be analysed, examining trace metal changes in sediments in relation to land use changes in the watersheds. This study w i l l investigate complicated chemical and environmental systems, identify a simple procedure to monitor stream health, and make recommendations on how to protect it. Specific aims of this study are as follows; Study aims: 1) Quantify urbanisation: Determine the extent and changes in urbanisation in L F V between 1973 and 2006, using historic aerial photos, orthophotos and GIS. Land use, land cover classes and % of imperviousness were the key variables used in this analysis, 2) Determine sediment quality: quantify the current and historical changes in streambed sediment quality in L F V , using several trace metals, 3) Linking land use with sediment quality: investigate the relationships between urbanisation and trace metals in sediments with a focus on urbanisation, 4) Identify the key land use factors that affect sediment quality: Determine which of the land use and metals in sediments show the best relationships and how these relationships change with increasing imperviousness. 3 2 Literature Review 2.1 Metal Pollution Problem in Urban Stormwater Runoff 2.1.1 Source of metals Trace metals originate from point and non point sources in an urban environment, as well as from geological sources. A wide range of metals are used in human activities and eventually find their way into local streams primarily associated with sediments. Table 2-1 shows the sources and uses of trace metals. Table 2-1 Products and uses of trace metals. Source: Moore and Ramamoorthy (1984); Minton (2005); W H O (2007). Metal Copper (Cu) Lead(Pb) Cadmium (Cd) Product and use Properties (e.g., malleability, ductility, conductivity, corrosion resistance, alloying qualities) - use in the electrical, construction, plumbing and automotive industries. The largest single user of Cu is the electrical industry accounting for > 50%. Heat exchangers, bus bars, magnet wire and windings in motors, and generators. Discharge of mine tailings, fertilizers and municipal and industrial sewage. Piping, building materials, solders, paint, ammunition and casting in the past. Storage batteries, metal products, chemicals, and pigments In more recent times. Antiknock agent in gasoline (tetraethyllead): the production of non-leaded gasoline for use in automobiles with emission control devices has sharply reduced it in gasoline. Gasoline additives and batteries accounted for 75% of the total lead consumption in the US in 1975. Anti corrosive and highway traffic safety paints. Until 1970s, automobile exhaust accounted for about 50% of the total inorganic lead absorbed by humans. Most commonly found associated with Zn in carbonate and sulfide ores, and is obtained as a by product in the refining of other metals. Electroplating, pigments, plastic stabilizers, batteries. Galvanizing iron, steel products and alloys. Corrosion resistant coating which is finished with electroplated metal coating. They are used in Zinc (Zn) construction, automobile, building industries for roofing, heating, ventilation, door panels. Ni-Zn butteries in automobile. ZnS in pigment. 4 Table 2-1 Continued. Metallurgical: production of ferroalloys. Resistance to corrosion and oxidation. Stainless steel and heat-resistant steels in petrochemical Chromium processing, furnace, cutlery. ^ C r ^ Refractory: refractory bricks, mortars, furnaces. Chemical: pigments, dyes in textile industry. Coinage, electroplating and alloys. Corrosion resistance (high strength and durability), electrical conductivity, alloy. Stainless steel and plating. Nickel (Ni) Transit and railway car manufacture. Ni-Cr alloys have long been used in heating elements of domestic stoves. In the manufacture of iron and steel alloys and manganese compounds and as an ingredient in various products, such as batteries, glass and fireworks, an oxidant for cleaning and bleaching. Manganese (Mn) A n organic manganese compound, methylcyclopentadienyl manganese tricarbonyl (MMT): an octane-enhancing agent in unleaded petrol. Compounds used in fertilizers and fungicides and as livestock feeding supplements. Chlorine and caustic soda production, pesticides, eiectricai apparatus. Mercury Measuring devices (thermometers, barometers, manometers), electrical (Hg) conductor and coolant. 2.1.2 Typical metal concentrations in water and sediments Trace metals in stream sediments Highway runoff is reported to provide high heavy metal contributions to streambed sediments in the Puget Sound Lowland in Washington State in the U.S. (Brandenberger et al. 2003). Cr was the exception because the level was uniformly high regardless of land use. Cu, Pb and Zn levels were slightly higher in urban developed sites than in forested sites, and considerably higher at highway sites (Figure 2-1). 5 150 O) CL 3 o O 100 50 I Chromium (Cr) Copper (Cu) I Lead (Pb) • Zinc (Zn) • • • • i i i ; 600 400 B E 200 N 0 £ r j £ 0 £ 0 1 0 £ > > > > > > L L L L U - L L L L Q Q Q O G G Land use X X X I Figure 2-1 Metal concentrations in streambed sediments, and land use in the Puget Sound Lowland (PSL) in Washington State. FR, D V and H W Y denote forest, developed and highway, respectively. Concentrations are dry weight. Modified from Brandenberger et al. (2003). Two urbanised sub-basins in the Brunette Watershed in L F V , were compared in 1995 (the Stil l Creek and the Brunette River sub-catchment areas) (McCal lum 1995). Metal concentrations significantly increased between 1973-1993, except for the decrease in Pb (Table 2-2). A significant enrichment of Cu , Pb and Zn were measured in Still Creek compared to the Brunette River. Traffic density resulting from runoff loadings rather than residential cover was considered the major cause of metal contamination (Sutherland 2000). Table 2-2 Streambed total metal concentrations (median) in the Brunette River and Still Creek sub-basins in 1993. Bold numbers mean the significant difference between the basins at the 5% probability level. Units are mg/kg except for Fe (%) (dry weight). Adapted from McCallum (1995). Element Still creek sub basin Brunette R. sub-basin Cu 130 51 Pb 130 55 Zn 251 128 Cd 141 103 Ni 17 12 Fe 2.1 2.1 Mn 576 807 Organic matter 6.2 5 6 A s an example, mean concentrations of aquatic sediments in the Sydney Harbour estuary are reported to be 102 mg/kg for C u , 202 mg/kg for Pb and 902 mg/kg for Z n adjacent to industrialized areas, where as sediments in highly urbanised sub-catchments had mean values of 22 mg/kg for Cu , 72 mg/kg for Pb and Z n (Birch et al. 2000). Results of studies in the urban environment are presented in Table 2-3 (Heiny and Tate 1997; Birch et al. 2000; Enguix Gonzalez et al. 2000; Sutherland 2000; Shea 2003). Concentrations in the agricultural area were studied in L F V to track potential use of trace metals as growth promoters (Table 2-4) (Smith 2004). Table 2-3 Trace metal concentrations in streambed sediments in various studies. Units are mg/kg except for Fe and A l (%) (dry weight). Land use Cu Pb Zn Cd Co Cr Ni Fe% Al% Mn Place Source Urban 33 76 198 The Piedmont Province, Georgia Shea 2003 Undergoing urbanization 36 37 159 The Piedmont Province, Georgia Shea 2003 Urban 175- 10-200- 320- Manoa stream, Sutherland 210 75 300 340 Hawaii 2000 Paramatta River Birch et al. 2000 Highly urban 90 223 983 18 32 4 577 catchments, Australia Sewage and industrial wastewaters 43 69 234 32 2.0 336 The Guadaira River in Southwest Spain Enguix Gonzalez et al. 2000 Urban-agriculture mixed 46 44 190 0.77 49 3.2 6.3 880 The South Platte River, Colorado Heiny and Tate 1997 Table 2-4 Element concentrations in the sediments in agricultural watersheds. All units are mg/kg except for iron (%). N/A is "not available". Modified from Smith (2004). Land use Al Zn Cr Co Cu Fe% Mn Ni Pb P Agriculture 0.69 75 149 48 22 4.2 1140 799 N/A 1097 Forest 0.91 41 31 20 53 1.6 386 28 N/A 597 7 Trace metals in surface waters Trace metals in surface water usually have much lower concentrations than those found in sediments, as sediments are usually negatively charged, are a sink for metal accumulation. Runoff from a paved highway showed that concentrations in water were 75 p,g/L of C u and 550 pg/L of Z n (as event mean concentrations), whereas a sampling station which drained the watershed from a mixed land use had 30 pg/L C u and 70 u.g/L Z n in the Brunette Watershed in L F V (Hall et al. 1999b). Table 2-5 presents the metal contaminations of surface runoff form recent studies (Mesuere and Fish 1989; Mesuere and Fish 1989; Mosley and Peake 2001; Rose et al. 2001; Buffleben et al. 2002). Table 2-5 Concentrations of trace metals in surface runoff from recent studies. Units are iig/L, and median values are shown. Place Characterist ics T ime C u Pb Zn C d C r Ni Fe Source Bal lona Creek, Santa Monica Bay, C A Residential land use > 60% Dec, 1997 Mar, 1998 24 19 19 41 0.5 1.2 2.2 6.7 6.5 7.8 Buffleben et al. 2002 Atlanta metropolitan region, Georg ia Runoff from parking lot May-Oct 1998 6.0 15.4 163 4.2 Rose et al. 2001 Kaikorai stream, New Zealand Urban runoff from rainfall events Base flow Sep-Dec, 1996 3.2 4.9 46 674 Mosley and Peake 2001 Storm flow Sep-Dec, 21.9 34 233 1833 1996 Runoff from Aug 1987-May 88 Portland, Oregon suburban parking lot into a pond ' 29 1.2 3 Mesuere and Fish 1989 Trace metals from atmospheric sources Hal l et al. 1999 studies atmospherically transported contaminants as the other major pathway of trace metals to waterways. Contaminants originating from both within the basin and outside the basin can deposit on impervious surfaces. It reports the concentrations of metals in precipitation, out of the assessment of which was conducted near Burnaby Lake (Belzer et al. 1997) (Table 2-6). Cu , Zn , N i and M n deposition are reported to be typical values in urban areas, and Pb is low 8 due to removal from gasoline. These values are lower than the concentrations in surface runoff by one to two orders (Table 2-5), yet metal accumulation from precipitation appears to contribute considerably to runoff loadings. Brewer and Belzer (2001) also indicates the atmospheric deposition as a significant portion of the total metal load to the Burnaby Lake watershed, and metal concentrations in the urban environment were one to three orders of magnitude higher than those measured in a rural location 100 km away from Burnaby Lake. Table 2-6 Trace metals in precipitation collected from the Brunette Watershed. These data are from 36 weekly samples collected between Jan. 31 -Dec.5,1997. Adapted from Hall et al. 1999.. Element Mean concentration (ug/L) Cu 1.03 Pb 0.50 Zn 15.2 Cd 0.38 Cr 0.51 Ni 0.60 Fe 21.6 Mn 1.88 Al 15.0 Mg 52.4 Ca 247.0 Na 338.0 2.1.3 Seasonal and storm event effects (first flush) First flush Metals deposited over the dry season are mobilized in the dissolved phase during the first storm of the wet season (Seasonal first flush) (Buffleben et al. 2002). Due to the rapid mobilization of particles deposited during the dry period by runoff, metal levels in runoff are usually much higher during these storm events than during prolonged wet periods (Quek and Forster 1993). High levels of C u and Pb have been found in first flush, after the water has sat in the drain pipes overnight (Hall and Schreier 1996). Concentrations increased during the initial hydrograph rise, and reach their maximum level prior to peak discharge (Rose et al. 2001). The quality of first flush storm runoff is comparable to that of raw sewage (Christensen et al. 1978). Anthropogenic 9 constituents are generated mainly from traffic-related activities (Sansalone and Buchberger 1997). Peaks in contaminant concentrations occur during high flow rates when transportation of highway contaminants is most efficient (Hoffman et al. 1985). Highway runoff were considered the source of over 50% of the annual contaminant loads of Pb and Z n entering a river in a recent studies (Harrison and Wilson 1985). A first flush effect is clearly seen for the dissolved (< 0.45 urn) components, but the particle-associated metals may show a much more complex temporal variation that are related to storm intensity and the flushing of large-grained sediment through the drainage system (Harrison and Wilson 1985). While Rose et al. (2001) reports that a large proportion of the Z n is likely to be adsorbed and transported on surfaces of the suspended sediment, C u and C d are mainly in dissolved form while Pb, Fe and A l are mainly particulate-bound (Sansalone and Buchberger 1997). Seasonal event Recent studies in the L F V demonstrate that the summer stormwater runoff had higher concentrations of contaminants than the winter storms, for Cu , Pb, Z n and M n . The summer rainfall events tend to be more infrequent and are more intense on average than the winter rainfall events. Thus, contaminants have a longer time to accumulate on impervious surfaces and are flushed into streams when the rainfall intensity is high (Hall et al. 1999b). Diluted input from steady winter rains can induce low metal levels in the water column. Extended dry periods in the summer allow the watershed to accumulate metals which are then released in runoff during fall rain events. A significant increase in dissolved metals occurred in the fall in the receiving water, probably from a combination of high-metal runoff (Mesuere and Fish 1989). Overall loads of contaminants can show the opposite trend to this summer-winter relationship. Runoff into streams can be divided into wet-weather flow (WWF) and dry weather flow (DWF) . The W W F is defined as the surface runoff produced by precipitation events, whereas the D W F is the flow in dry seasons resulting from groundwater inflow, automobile washing, and other residential and commercial water uses. The W W F discharges from urban storm drains into receiving waters have long been recognized as a significant source of contaminant loads and cause of water-quality degradation. Under dry weather in a Los Angeles, California, W W F accounted for about 60% of annual nickel loadings, 70% for lead, 80% for Cu , and 50% for Cr. 10 While the loadings in wet season were greater, dry season loadings also significantly contributed to the loadings (McPherson et al. 2005). 2.2 Imperviousness as an Indicator 2.2.1 Total imperviousness and effective imperviousness Impervious surfaces are areas where water does not infiltrate into the soil (e.g., roads, rooftops, sidewalks, patios, parking lots), and hence imperviousness (% impervious areas within a watershed) has been used as an indicator of the impact of urbanisation on streams. Impervious surfaces have significant impacts on the aquatic ecosystems, since it changes the hydrological pattern in urbanised watershed from those in undisturbed watershed. Various contaminants in urban environment are rapidly conveyed from impervious surfaces to streams, resulting in stream pollution (Arnold Jr and Gibbons 1996; Zandbergen et al. 2000). The two most common measures of imperviousness are total impervious area (%TIA) and effective impervious area (%EIA). T I A includes all impervious surfaces in the watershed, whereas E I A is calculated by impervious surfaces which are directly connected to the surface drainage systems (Alley and Veenhuis 1983; Beyerlein 1996; Booth and Jackson 1997; Brabec et al. 2002). E I A is estimated to be 0.15 times the T I A in highly urbanised watersheds in Denver, Colorado, and most studies estimated the E I A based on T I A percentages (Table 2-7) (Brabec et al. 2002). However, there is no universal agreement as to the ratio of E I A / T I A . Table 2-7 Presumed relationship between imperviousness and land use. Modified from Dinicola (1990); Booth and Jackson (1997). Land use %TIA %EIA Low density residential 1 unit per 2-5 acres 10 4 Medium density residential 1 unite per acre 20 10 Suburban density 4 unites acre 35 24 High density multi-family or 8+ units per acre 60 48 Commercial and industrial 90 86 11 2.2.2 How imperviousness is measured Several methods exist to calculate watershed imperviousness. Conventionally aerial photos have been used because manual interpretation and digitizing are most trustworthy in categorizing land surfaces (Anderson et al. 1996). The introduction of orthophotos has improved the accuracy of ' measuring imperviousness from aerial image interpretation (White and Greer 2006). Digitizing aerial images on GIS is probably the most reliable method to identify a large range of land uses. However, this method consumes a significant amount of time/labor. Because exact determination of imperviousness is laborious, estimates can be made via land use data. Some studies estimate imperviousness in proportion to each land use type (Table 2-8). The traffic component often exceeds the rooftop component (Schueler 1994; M a y et al. 1998). Table 2-8 Percentage of imperviousness for various land cover classes. Modified from Brabec et al. (2002) . Land cover class Notes Mean Range Single-family residential <0.25 acre lots 39 30--49 0.25-0.5 acre lots 26 22--31 0.5-1.0 acre Its 15 13--16 Includes multi-family residential 30 22--44 Multiple-family residential 66 53--64 Commercial 88 66--98 81 52--90 Industrial 60 40 11--57 Open 5 1~ 14 Land use classification based on interpretations of satellite imagery and remote sensing has been available for some time. Landsat and I K O N O S imagery has been used to classify land surfaces for impervious surfaces, forest fragmentation and urban sprawl (Sleavin et al. 2000; Hurd et al. 2001; Civco et al. 2002). Automatic categorization is found to be a powerful, valuable way to identify land information, decreasing the tasks for digitizing (Goetz 2003). A concern is the inaccuracy of automatic color identification in images, which is particularly the case for older aerial photos. Recent increases in satellite image resolution and development of interpretative methods have resulted in improved classifications (Sawaya et al. 2003). 12 2.3 Interactions between Imperviousness and Water Quality and Stream Health 2.3.1 Imperviousness and aquatic biota Aquatic health is quantified through species indices such as the index of biotic integrity (IBI) (Karr 1981; Schueler 1994; M a y et al. 1998; Morley and Karr 2002). Benthic invertebrates (e.g., species belonging to Ephemeroptera, Plecoptera, Thrichoptera) or fishes are used to quantify biodiversity in a stream. The indices reveal overall aquatic health, integrating the cumulative effects of various stressors. Stream degradation exponentially proceeds as imperviousness rises (Figure 2-2) (May et al. 1996; Booth and Jackson 1997; M a y et al. 1998; Booth et al. 2004; Snyder et al. 2005). 5 0 - • 4 0 " \ m 3 0 -m m 2 0 -1 o -• • • - • • 0 1 1 0 2 0 T o ta I Ii 4 0 ii p e rv io u s 6 0 A r e a (% ) 8 0 Figure 2-2 Relationship between watershed imperviousness and benthic-biological integrity in the Puget Sound Lowland (PSL) streams. Modified from Booth et al. (2004). IBI, species richness, and other biodiversity indices are found to exponentially decrease at the early stage of urbanisation with a threshold of 10-15% of imperviousness, above which benthic invertebrate diversity is seriously degraded; namely the relationship is curvilinear. This indicates the existence of a threshold of stream degradation at the early stage, and minor changes in urbanisation could result in major changes in stream condition (Wang et al. 2001). A riparian buffer zone is found to be the key feature to help regulate environmental conditions in stream ecosystems (Naiman et al. 1993; May et al. 1998). A n artificially modified riparian zone 13 is uniformly structured and anthropogenic impacts on aquatic health appear severe (Booth et al. 2004). Sensitive-area ordinances, now in effect in most local municipalities in British Columbia, typically require a riparian forest cover of 30 m width (Wenger 1999; Castelle and Johnson 2000). However, a variety of human influences fragments the riparian corridors, with road crossings often having the most damaging impacts (Barton et al. 1985). 2.3.2 Imperviousness and water quality Studies are rare on the relationship between imperviousness and trace metal concentrations in water, because trace metal levels in water are very low and analytically challenging (Smith 2004). A few studies of the relationship are shown in Table 2-8 (Sabin et al. 2005; Stein and Ackerman 2007). Table 2-9 Trace metal concentrations in surface water in relation to urban land use types. Units of trace metals are ug/L, and those of land use are % . Place Land use Cu Pb Zn Cr Ni Fe Source Imperviousness Los Angeles 80% 5.9-37 1.2-1.6 32-320 2.1-20 2.1-8.5 "OdDlM c l dl. 2005 Los Angeles Commerci al(1) Densely residential (2) Industrial (3) Highly urbanize d area (1 Low density residential + 2 + 3) Stein et al. LA River 8 7 10 25 30 25 2 122 3 3 288 2007 Coyote Creek 13 7 14 34 41 5.8 2 57 0.3 1 469 San Gabriel River 19 7 13 39 52 26 3 213 0.4 9 571 San Jose Creek 11 4 15 30 41 8 2 117 0.7 5 1911 Walnut Creek 8 4 5 17 31 13 3 73 0 1 558 Ballona 16 22 7 45 36 19 4 79 2 5 515 2.3.3 Imperviousness and sediments McCal lum (1995) and Hal l et al. (1999a) investigated streambed sediment quality in the Brunette Watershed in Burnaby in the L F V in 1973 and 1993. The watershed is dominated by residential/industrial use, and imperviousness was determined to be about 40% (Table 2-10 and 2-11). A l l metals (except N i and Fe) increased between 1973-1993 (Table 2-12). The slight increase in imperviousness (34 to 41%) led to significant increases in Cu , Pb, Z n and M n levels. 14 Table 2-10 Percent land use in the Brunette River Watershed (total area). Adapted from McCallum (1995) Land use 1973 1993 Change Residential 40.8 45.7 4.9 Industrial 11.9 13.2 1.3 Commercial 3.6 4.1 0.5 Institutional 6.6 6.4 -0.3 Transportational 2.7 2.7 0 Agricultural 1.4 o. -1.4 Open space 32.9 28 -5 Table 2-11 Percent land cover in the Brunette Watershed (total area). Adapted from McCallum (1995) Land cover 1973 1993 Permeable 66 59 Impermeable 34 41 Effective impermeable 26 32 Table 2-12 Trace metal concentrations in the sediments in the Brunette River (median values, n=32). Bold numbers mean the values of 1993 are significantly different from those of 1973 at the 5% probability level. Adapted from McCallum (1995) Element Concentrations (mg/kg, % in Fe) 1973 1993 Cu 30.9 56 Pb 96.9 63 Zn 98.6 143 Ni 15.1 14 Fe 2.5 2.0 Mn 349.4 807 Brydon (2004) examined metal concentrations in streambed sediments in stormwater detention systems in the L F V (Figure 2-3). A l l drainages were located in highly urbanised areas, and the imperviousness ranged from 50-100%. Substantially higher levels of Cu , Pb and Z n (500-600, 200, more than 900 mg/kg, respectively) were observed in sediment trap facilities along 15 highways. A c i d extractable metals (i.e., Cu, Pb, Zn) saw increased concentrations when imperviousness increased. SD(50) Griffin HWY Tempe Oakalla (60) (100) (65) (55) SD(50) Griffin HWY Tempe Oakalla (60) (100) (65) (55) LL < o E SD(50) Griffin HWY Tempe Oakalla (60) (100) (65) (55) SD(50) Griffin HWY Tempe Oakalla (60) (100) (65) (55) SD(50) Griffin HWY Tempe Oakalla (60) (100) (65) (55) Figure 2-3 Element concentrations (dry weight) in streambed sediments in stormwater facilities in Lower Fraser Valley. Left bar represents sediments at the inlet, and right bar represents outlet ones. SD denotes 52Ave/221A Street Detention Facility (Langley). Griffin denotes Griffin Park Biofiltration Pond (North Vancouver). H W Y denotes Westview Interchange Detention Pond (North Vancouver). Tempe denotes Tempe Heights Pond (North Vancouver). Oakalla denotes Oakalla Biofiltration System (Burnaby). Parenthesis means imperviousness in a collective drainage. Modified from Brydon (2004). 16 Sediment quality was significantly correlated with percent impervious area in an Alaska study, with Z n having the highest correlation (r s = 0.87) in Anchorage (Figure 2-4) (Ourso and Frenzel 2003). Storm drains and roads were reported to be influential factors on the stream degradation. o E o 10 20 30 40 % impervious surface Pb (rs=0.65) 50 o o oo o o (0 • X o I 5 o o tM "A Zn A A 10 20 30 40 % impervious surface Zn (rs=0.87) 50 E . o O) o CM 10 20 30 40 % impervious surface Cd (rs=0.66) 50 10 20 30 40 % impervious surface Ni (rs=0.65) 50 Figure 2-4 Imperviousness vs. metal concentrations in urbanised watersheds in Alaska. Spearman's rank correlation is shown (rs). Modified from Ourso and Frenzel (2003). 2.4 Storm water Management During rainfall events, stormwater flushes the contaminants e.g., metals, nutrients, organic compounds, pathogens, sediments, into receiving waters. Such impacts are exacerbated by traditional drainage systems and end-of-pipe solutions as this design concentrates contaminants (Marsalek and Chocat 2002) and thus, increasing the risk of environmental degradation (US E P A 2000; Gafiield et al. 2003). Traditional drainage systems and treating drinking water require 17 infrastructure investments and are financially demanding. Increased concerns have led to the introduction of the best management practices (BMPs) . The purpose of B M P s is to mitigate the adverse impacts of urbanisation and to address the issues of urban runoff, including flood protection, water supply management, groundwater quality, wastewater management and receiving water quality (Marsalek and Chocat 2002; Gaffield et al. 2003). Stormwater mitigation measures, such as wider riparian vegetation, infiltration basins, detention ponds or constructed wetlands significantly mitigate the impacts of stormwater runoff (Houston 2004). The examples of B M P s are shown in Table 2-13. Table 2-13 The best management practices in stormwater management. Modified from Zandbergen et al. (2000). Structural = = = ^ ^ = ^ ^ = = = = = = = = = ^ _ = Detention p o n d s Storing and infiltrating stormwater into permeab le sur faces . Vegetat ive c o v e r is preferable to e n h a n c e infiltration. Per iodical removal of the sed iments is n e c e s s a r y to avoid overflow and c logging of inlet/outlet. Cons t ruc ted Acce lera t ing infiltration a n d evapotranspirat ion. P roper introduction of plants wet lands for select ive contaminants is n e c e s s a r y . A s h a p e of the facilities and topography are important to e n s u r e effective s torage for both p o n d s a n d wet lands. Vegetat ive o p e n Conveying/ t reat ing stromwater runoff on site. C h a n n e l s differ accord ing to c h a n n e l s in the width, depth, surficial materials/vegetat ion (ditches, g r a s s channe l , dry/wet street right of way swales) . P roper soi ls s u c h a s s a n d y loam are preferable. C u l - d e - s a c s R e d u c i n g a r e a s of drive w a y s and increasing larger perv ious s u r f a c e s with landscap ing . Dr iveways with porous p a v e m e n t s (gravel /sand) a n d a s h a r e d dr iveway are s u g g e s t e d . M a n a g i n g C o n d u c t i n g roofwater to infiltration facility with c o a r s e material /vegetative Roofwater sur faces , or to cistern for gardening use . Establ ishing green roofs. Cluster ing hous ing Al lowing more permeab le a reas . units Parking lots with Min imized s p a c e , porous pavement , lots sur rounded by bioretention B M P s facilities (e.g., s a n d filter/filter stripe with c o a r s e materials, or road s ide swales) . Ripar ian Mitigating stormwater impacts, f looding, a n d ensur ing basef low. D iverse vegetat ive buffers channe l morphology (e.g., meander ing) with more than 3 0 m buffers on both ' s i d e s of the river is preferable. Nonstructural _ _ = = = = = = = = = = = ^ = = ^ ^ _ ^ ^ _ = = = M a n a g e m e n t P lanning a n d des ign ing B M P s , with periodical m a n a g e m e n t . Regulat ion/educat ion for the public for genera l u s e of structural B M P s . Street s w e e p i n g , u s e of alternate materials in veh ic les , regulating fuel addit ives. 18 These techniques can moderate the hydrological regime and enable contaminants to be detained. Also soil micro-organisms play a key role in modifying contaminants and wetland can act as filters. In this way contaminants can be retained at the source. However, long term management of these detention facilities needs to be considered. Treatment of metal enriched sediments in detention systems is one of the long term challenges facing urban stormwater managers. 19 3 Methods 3.1 Study Area The five study areas are located on the north side of the Fraser River in the Lower Fraser Val ley ( L F V ) . Streambed sediments were analysed in 1973 (Westwater Research Centre 1973) in four of the selected watersheds: Alouette, Hatzic, Kanaka, and Thornhill watersheds. A fifth watershed was added in West and North Vancouver (denoted as N / W Vancouver in this study) to represent a well established and densely populated urban environment (Figure 3-1). The watershed boundary was derived using a 20 m by 20 m digital elevation model ( D E M ) that determines flow accumulation and watershed function in ArcGIS. This boundary was corrected using the 1 0 m contour map of Terrain Resource Information Mapping (TRIM) data from the Ministry of Sustainable Resource Management at the Government of British Columbia (2000) (http://srmwww.gov.bc.ca/gis/arctrim.html). T R I M was also used to develop vector data sets (i.e., road and streams) of the watersheds. The Alouette watershed is located approximately 40 km east of Vancouver. The headwaters of the watershed are located on forested mountain slopes. The lower portion of the watershed is located in the urbanised area of Maple Ridge. The sediment samples were collected in the highly urbanised part of the watershed (site a l to a3), and at site a4 in the rural portion of the watershed. The Hatzic watershed is located approximately 65 km east of Vancouver. The urban portion of the watershed is in Mission, the central part of the watershed is dominated by agriculture and the headwaters are on steep mountain slopes. Sediments were collected on the eastside of the watershed (site h i and site h2) where land use is mostly forest with some logging activities, and at the southwest side of the watershed where urbanisation is near the headwater of the watershed (from site h3 to site h7). The Kanaka watershed is located southeast of the Alouette watershed, the eastern suburb of Maple Ridge. Five low density urban watersheds (site k l , k2, k4, k5 and k6), and one forested watershed was chosen (site k3) were chosen for analysis. 20 The Thornhill watershed is located between Maple Ridge and Mission. A forested watershed was chosen for site t l , and the others are low density urban areas in a rural environment (from t2 to t7). The N / W Vancouver sites were located at the north shore of Burrard Inlet. The area is urbanised on steep slopes of the mountain. Site v l and v2 were located in the pond in the Ambleside Park at West Vancouver, which is a sport park facing the Burrard Inlet. Site v3 and v4 were located in small streams that receive stormwater from urbanised hillslopes. Site v5 was located at the inlet of the Mosquito Creek, a constructed wetland which receives stormwater from an urban area above. Site v6 was located at the stormwater detention pond near the highway, where the stormwater from the highway is directed to the pond. 21 T r V ^ y British CoiumMa ; 0 5 10 20 A Figure 3-1 The location of the five study area (Alouette, Hatzic, Kanaka, Thornhill, N/W Vancouver). 22 3.2 Land Variable Spatial information was obtained from an evaluation of the land use/cover based on historic aerial photos and recent ortho-photos. ArcGIS 9.1, ESRI software of Geographic Information Systems (GIS), was used for the analyses. 3.2.1 Topography and mineral texture Mean and a range of elevation, and mean slope were calculated from digital elevation model ( D E M ) for each watershed. In each watershed the gradient of the streams was calculated for a 200 m sections upstream of the sampling station. The soil survey for the L F V (Luttmerding 1980) was used to determine the mineral texture of the parent material of each watershed. 3.2.2 Land use Aerial photos and orthophotos (aerial photos which were spatially corrected, to the N A D 83 coordinate system) were analysed for delineating land use using the ArcGIS software. Aerial photos from March 1971 were used to determine the urban density in the early 1970's (source: Map library, Department of Geography at the University of British Columbia). The photos were scanned and geo-referenced on ArcGIS . Orthophotos of March 2004 provided by D F O were used to determine the most recent land use (note: land use data are referred to as the data in 1973 and 2006 when sediments were sampled). The photos were digitized and analysed with ArcGIS and the following land use categories were selected: • Forest • Rural-urban • Dense-urban • Commercial-industrial • Grass-agriculture • Others 23 Grass and agriculture were merged together because agricultural areas were mostly small farms and small. Areal percentage of each land use category in a watershed was calculated. Land use in N / W Vancouver was quantified only in 2006 since no sediment quality data in 1970s was available in the areas. 3.2.3 Land cover The land cover classification included the following categories: • Impervious-surface: surfaces which do not allow water to penetrate into the soil (e.g., asphalt, concrete, parking lots, roofs). • Forest-cover: forests and individual trees. • Grass-cover: grassy cover, exposed soils and agricultural land which allows rainfall infiltration. The use and cover were delineated in all watersheds, and land cover ratio was determined. This ratio was extrapolated to land use in the other watersheds to estimate land cover. Watersheds a l -a4 and v5 were digitized for all land cover categories in the watershed. Watersheds k l - k 2 and v3-v4 were digitized mainly for impervious-surface. B y using an average percentage of land cover in each land use category, land cover in all watersheds in both 1973 and 2006 was estimated. 3.2.4 Buffer zones Land use and land cover within 100 m buffer zones of the streams were calculated. A hundred metre buffer zones along both sides of the streams were generated on ArcGIS , and the percentage of land use/cover in both 1973 and 2006 were calculated in the same manner as was the land use in the whole watershed. Buffer zones were created along storm sewers in watersheds v3-v6. 3.2.5 Road density The road density was calculated as length/area (km/km 2), by using road vector data of T R I M . In the very dense urban areas the roads were digitized. 24 3.2.6 Historical changes in land variables Historical changes in all land use/ land cover and road density were calculated for all watersheds. Topography and geology are assumed to be the same. 3.3 Sediment Quality 3.3.1 Field methods Streambed sediment sampling Stream-bed sediments were collected at the following sites; • Four sub-watershed sites in the Alouette watershed (al-a4) • Seven sub-watershed sites in the Hatzic watershed (hl-h7) • Six sub-watershed sites in the Kanaka watershed (kl-k6) • Seven sub-watershed sites in the Thornhill watershed (tl-t7) • Six sub-watershed sites in the Vancouver area ( v l -v6) With the exception of the N / W Vancouver sites the twenty-four sediment samples were collected at the same locations as those sampled in 1973 (Westwater Research Centre 1973) with the following exceptions. • (i) site k3 was sampled approximately 200 meters upstream from Hal l ' s location because the stream could not be found in bushes. • (ii) site h7 was added downstream of h6. The sample collection took place on August 2, 2006 for sites hl -h7 and tl-t7, August 3, 2006, for the sites al-a4 and k l - k 6 , and August 17, 2006 for sites v l - v 6 . One separate stream-bed sediment grab sample per site was collected using an aluminum pot affixed to a 2 m wooden pole. Each sample was first passed through a stainless steel (S.S.) 2 mm sieve prior to placing in high wet-strength plastic bags. Streambed sediments were directly placed without being sieved at the site t6 and v6, since there was no water in the streams. 25 Samples were transported back to the laboratory and stored up to 3 weeks at 4 °C in a refrigerator prior to sample preparation. Stream water sampling Stream water was sampled at the same sites, except for the site t6 and v6 where there was no water in the streams. Samples were collected in 125 ml plastic bottles and transported back to the laboratory and stored at 4 °C in a refrigerator prior to sample analysis. Temperature and specific conductance Stream temperature and specific conductivity were measured at the same sites, by using Y S I 30 Conductivity, Salinity, & Temperature Meter. 3.3.2 Sample preparation and analysis Stream sediments The samples were wet sieved and passed through S.S. 180 u,m sieve then S.S. 63 urn sieve by using distilled water to get the fraction of silt and clay in the sediments. The samples of the site t6 and v6 were passed through S.S. 2 mm sieve before being passed though 180 urn one, because the samples were not sieved at the sites. The 63-pm-size fraction is usually analysed for trace elements, because 63 um is the boundary of silt/sand fractions, and clay and silt maintain electrostatic negative charges of sediments the most (Shelton and Capel 1994; Birch et al. 1996; Heiny and Tate 1997; Schreier et al. 1997; DiVenere 2006). Sediments passing through the sieve were dried at 105 °C in an oven for 24 hours and then disaggregated. These samples were placed in the plastic bottles until further analyses. 26 Stream water Water samples were filtered to remove solid materials in the water and nutrients were analysed at the Soi l Science laboratory, at U B C prior to nutrient analysis. 3.3.3 Analytical techniques Acid digestions (Aqua regia) The sediments were digested using the aqua regia methods in the Soil Science laboratory at the University of British Columbia (UBC) . This is a standard E P A method (US E P A 1992). Elements of focus are shown in Table 3-1. The elements in 1973 are heavy metals, while major cations, phosphorus and organic matter contents were added in 2006, to identify more detailed characteristics of sediment quality. Table 3-1 Elements of focus in 1973 and 2006. 1973 2006 C o C o C u C u F e F e M n M n Ni Ni P b P b Z n Z n C d C r Al C a K M g N a P h o s p h o r u s O r g a n i c Matter 27 0.5 g of the dried sediments were weighed and placed in 250 ml Phillips beakers. Four ml of 1:1 nitric acid (FINO3) and ten ml of 1:4 hydrochloric acid (HC1) were added in the beakers. Each of the beakers was covered with a watch glass. The samples were digested in the beakers on a hot plate and gently refluxed for 30 minutes. Temperature of the samples was kept 85-90 °C. After cooling, the samples were filtered through Whatman #42 filter paper into 100 ml volumetric flasks, and the digests were made up to 100 ml with de-ionized water. ICP atomic emission spectroscopy ICP-Atomic Emission Spectroscopy ( ICP-AES) was used to analyse the sediment samples. Water and soil samples were treated with HC1 acids and heated. The digests were then analysed for trace metals by an atomic emission optical spectroscopic technique (US E P A 2004). The Varian Vista-Pro C C D Simultaneous ICP-OES was used for the analysis (Tyler 1995). The detection limits together are provided in Appendix B . Duplicates were made for the sediments of three sub-watersheds (a2, k4 and t3), to assess the precision of sediment analysis. Accuracy was not determined in this study, as the historic samples were not available for re-analysis and the analytical instrumentation and extraction techniques in 2006 have changes since that time. Instead, results of sediment quality were compared with existing studies. Loss on ignition (LOT) and color Loss on ignition (LOI) was determined to provide a measure of the organic content in the sediment samples (Heiri et al. 2001). Five grams of sediments were weighed in crusibles, and placed in an oven at 105 °C overnight. The dried samples were weighed and the color of the samples was recorded referring to Munsell Soil Color Charts (Munsell Color 1975). The samples were heated to 500 °C for 3 hours to burn off the organic material, and the samples were weighed again to calculate L O I on the basis of sample dry weight. The color of the samples after the ignition was recorded referring to the Munsell Charts. 28 Nitrate (NO3) analysis Nitrate levels in water was detected by using L A C H A T QuikChem FIA+ 8000 series (US E P A 2002). 3.3.4 Sediment quality data in 1973 The following metal data was available for the 1973 sediments analysed by (Westwater Research Centre 1973): Co, Cu , Fe, M n , Pb and Zn. For trace metal extraction from the sediments, the samples were dried in an oven at 110°C for 48 hours and then disaggregated in a mortar prior to sieving through a 177 pm mesh nylon screen. A one gram sample of the sieved material was digested with 10 ml of a nitric perchloric acid mixture (a 4:1 mixture of concentrated HNO3 annd 70 percent HCIO4). The samples were refluxed for 1 hour, evaporated to dryness until no white fumes evolved, cooled and diluted with 5 ml o f 6 M HCI and 15 ml distilled water prior to aspiration into an atomic absorption spectrophotometer (Techtron or P .E. 303) against standards in 1.5M HCI . The technique was used to determine Fe, M n , Cu , Zn , Pb, N i and Co with a detection limit of 0.1 mg/kg dry weight. Cr was determined by emission spectrography. The sieved sample was ignited at 550°C for three hours and 0.1 g of the ash mixed with an equal weight of graphite containing an indium standard. The samples were analysed by a DC-arc spectrographic procedure (K. Flectcher, personal communication). The detection limit was 25 mg/kg dry weight. The notable differences of sediment analysis in 1973 from that of this study are (1) the < 177 pm particle size fraction was analysed instead of 63 pm, (2) sediments were dry sieved (not wet sieving), (3) the digestion method was nitric/perchloric acid (NP) instead o f aqua regia ( A R ) , and (4) the atomic absorption spectrophotometer (flame A A ) technique was used instead of the ICP-A E S technique. McCa l lum (1995) compares these differences by using the Brunette Watershed sediments (Appendix B) . The combination of nitric and perchloric acid is the strongest extraction method. The I C P - A E S technique measures metals more accurately than flame A A . In the study of the Brunette Watershed, sediment Cu , Pb, Z n and N i concentrations differed by less than 10% between N P and A R digestion methods. The fact that concentrations derived from two techniques were 29 linearly correlated allows direct comparisons between data collected using the different detection techniques (McCal lum 1995). Fine fraction tended to have higher concentrations than coarse one, A R tended to produce lower concentrations than N P for Pb, Z n and M g , and I C P - A E S produced lower concentrations than flame A A for Cu , Pb, Cr and N i . This fact may compensate the differences of analytical techniques of this study from those in 1973; this study uses A R and I C P - A E S with finer sediments, while 1973's study used N P and flame A A with coarser sediments. 3.3.5 Sediment quality guidelines Canada has two freshwater sediment quality guidelines; interim sediment quality guidelines (ISQGs) and probable effect level (PEL) (Canadian Council of Ministers of the Environment 2002). A n equivalent of ISQGs in the United States Geological Survey (USGS) is threshold effect level (TEL) (USGS 1999), and ISQGs were referred to as T E L in this paper. T E L and P E L are flexible interpretive tools for evaluating the toxicological significance of sediment chemistry data, as well as for prioritizing actions and management decisions. Chemical concentrations below T E L are not expected to be associated with any adverse biological effects. Concentrations above P E L are expected to be frequently associated with adverse biological effects. Chemical concentrations between T E L and P E L represent the range in which effects are occasionally observed. The use of these two values is a practical means of characterizing sites as having a minimal, potential, or significant toxicological concern and further investigations are recommended (Environment Canada 2004). Sediment quality guidelines, T E L and P E L of U S G S are also used to interpret any impacts of sediment trace metals (USGS 1999) (Appendix B) . 3.3.6 Historical changes in sediment quality data Historical changes in selected heavy metals (Co, Cu , Fe, M n , Pb and Zn) between 1973 and 2006 were calculated for the same sampling sites. 30 3.4 Statistical Methods A statistical package, SPSS 15.0, was used for the analyses (SPSS Inc. 2007). 3.4.1 Factor analysis: land variables Factor analysis determines the contribution to the total variance for each of the variables. Principal component factor analysis begins with the principal component analysis (PCA) . The objective of P C A is to take a number of variables and find combinations of these to produce new indices (principal components) that are uncorrelated in order of their importance, and that describe the variation in the data. Eigenvalues are the variance of the principal components, and the first principal component has the biggest eigenvalues. Component loadings are the correlation between original variables and principal components (Manly 2004). Factor analysis uses the first few principal components. It rotates the axes of new indices identified, to find simpler structure of the loadings. Each factor is composed of selective variables, and therefore it is possible to find similarities among variables and characteristics of the factors (Manly 2004). Factor analysis was applied to land variables in both 1973 and 2006 to identify the similarity among the variables and find the most important variables. Variables used are presented in Table 3-2. The number of factors extracted was determined by using eigenvalues of factors before rotation of axes (principal component factor analysis). Factors whose eigenvalues are more than those of the Broken-stick model were retained. This method assumes that i f sum of the eigenvalues is divided randomly amongst the various components, then the expected distribution of the eigenvalues follows a broken-stick distribution. Observed eigenvalues are considered interpretable i f they exceed eigenvalues generated by the broken-stick model. The eigenvalues of the model are calculated as: p bk= S 1/J 31 where p is number of variables and bk is the size of the eigenvalue for the k-th component under the broken-stick model (Jackson 1993). Varimax rotation was used as it is the most common method for an orthogonal rotation. Promax rotation which is also a common oblique rotation method was also used to compare the results (Manly 2004). Table 3-2 Land variables used for factor analysis. Colluvium/till means colluvium/glacial till— geology, and glaciomarine means glaciomarine deposits-geology. Road-density was used for only 2006 data. T o p o g r a p h y / g e o l o g y L a n d u s e / c o v e r L a n d u s e / c o v e r in buffer z o n e s C o a r s e - m i n e r a l Forest Fores t F ine-minera l Grass-agr icu l ture Grass-agr icu l ture Stream-gradient Rura l -urban Rura l -urban S l o p e D e n s e - u r b a n D e n s e - u r b a n Elevat ion (range) Commercia l - industr ia l Commercia l - industr ia l C a n o p y - c o v e r C a n o p y - c o v e r G r a s s - c o v e r G r a s s - c o v e r Impervious-surface Impervious-surface Road-dens i ty (only in 2006) Road-dens i ty (only in 2006) 3.4.2 Factor analysis: macro and micro elements Factor analysis was also applied to the data of macro and micro elements in both 1973 and 2006 to identify the similarity among the elements and the most important elements, and find underlying factors behind them. This was conducted in the same manner as factor analysis for land variables. 3.4.3 Cluster analysis by sediment quality data Cluster analysis was used to aggregate the observations of data to find clusters in which selected variables are similar (Manly 2004). This study aimed at finding similar groups of sub-watersheds in terms of sediment quality. Grouped sub-watersheds were further used to analyse the impacts of land use/cover on sediment quality in a surrogate, multivariate manner. The factor scores of sediment quality data were used as the variables to cluster. Complete Linkage and Ward methods 32 (hierarchical cluster analyses) were applied as they are one of the most common methods (Punj and Stewart 1983). A dendrogram of each result was produced. 3.4.4 Correlation between land variables and sediment quality data The plots of imperviousness vs. concentrations of each element were created to find the relations between the two, and the similarities and differences among the elements. Statistical package R ( C R A N 2007) was used to create the plots. Data of both land variables and sediment quality were not normally distributed, and therefore non-parametric statistics were applied. Spearman's rank correlation coefficients were calculated between selective land variables and sediment quality data for both 1973 and 2006 data. 3.4.5 Correlation on the historical changes Spearman's rank correlation coefficients were calculated between the historical changes in land variables and concentrations in metals. This is to see i f the increase in concentrations were caused by the urbanisation during a long time of period (33 years). 3.4.6 Mann-Whitney U test Mann-Whitney U test was applied to the selected variables among the clustered watersheds, to detect the cause of the differences in sediment quality. Stations located at the most upstream site of the watershed were chosen, eliminating the ones which were spatially related. Bar charts were created to see the range and median of the variables within the clusters. 33 4 Results and Analyses 4.1 Land Variables 4.1.1 Topography and mineral texture Figures 4-1, 4-2, 4-3, 4-4, 4-5 and 4-6 show the location of the sub-watersheds examined in this study, namely Alouette, Hatzic, Kanaka, Thornhill and N / W Vancouver, respectively. Table 4-1 provides the topographic and mineralogical (texture) characteristics of all sub-watersheds. The sub-watersheds under investigation are underlain mostly by granitic geology with glacial and glacio-fluvial depositions on top (Mackintosh and Gardner 1966; Luttmerding and Sprout 1968). This is responsible for highly variable textures and mineral composition in the different sub-watersheds. Two textures classes were dominant: (1) coarse texture being composed of glacial ti l l /colluvium materials and gravels, and (2) fine textured deposits dominated by clay minerals. A l l Alouette sub-watersheds have fine textured deposits while the Hatzic sub-watersheds are dominated by coarse textured materials. The dominant surface materials in the Kanaka sub-watersheds were fine textured deposits. The sub-watershed k l has an equal mixture of coarse and fine materials while the Thornhill sub-watersheds consist of a mixture of a 25-75% of coarse and fine textured materials. N o mapped data were available for the N / W Vancouver sub-watersheds (v3-v6) but the parent material is mostly of glacial marine origin with alluvium and marine material in the low part of the watersheds. 34 35 Figure 4-5 N/W Vancouver drainage areas. Figure 4-6 N/W Vancouver drainage areas. 36 Table 4-1 Topography and mineral texture in studied sub-watersheds. Watershed Area (km2) Elevation Min Max Range (m) (m) (m) Slope Mean (°) (m) Stream gradient 200 m upstream (%) Mineral Texture (%area) Coarse Fine a1 0.47 19 50 31 40 2.0 2.0 0 100 a2 0.53 15 51 36 35 2.5 2.0 0 100 a3 0.51 17 40 23 33 2.2 2.0 0 100 a4 0.90 50 106 56 83 3.1 0.6 0 100 hi 0.79 16 874 857 541 23.6 1.8 100 0 h2 1.09 41 920 878 638 21.2 31.1 100 0 h.3 1.33 136 245 109 162 5.8 2.8 84 14 h4 3.17 129 245 116 154 4.5 1.0 66 17 h5 6.12 69 291 222 150 5.7 0.6 81 14 h6 7.26 18 291 273 141 6.4 1.0 100 0 h7 7.55 12 291 279 138 6.5 4.8 100 0 k1 3.18 79 336 257 129 5.1 0.0 53 42 k2 0.72 16 98 82 59 3.6 0.9 0 100 k3-2004 1.20 201 497 297 313 7.4 4.5 100 0 k3-1971 1.29 184 497 313 305 7.4 5.0 100 0 k4 1.28 45 105 60 85 2.3 3.0 0 79 k5 2.15 17 105 88 79 3.2 3.0 0 100 k6 2.74 5 105 100 70 3.7 0.0 0 98 t1 8.15 182 872 690 463 8.9 0.2 100 0 t2 1.70 103 190 86 159 3.5 5.4 51 49 t3 1.07 109 276 168 156 7.1 3.3 73 27 t4 1.55 99 276 178 149 6.9 1.7 60 40 t5 2.87 57 276 219 127 6.2 0.1 43 57 to 1.74 102 282 181 166 4.8 5.2 57 43 t7 3.11 47 282 236 139 4.5 4.0 36 64 v3 0.21 11 129 118 63 11.0 3.5 100 0 v4 0.01 9 16 7 12 5.2 1.2 100 0 v5 0.11 90 101 11 95 1.8 2.2 100 0 v6 0.06 77 119 42 92 3.7 1.0 100 0 4.1.2 Land use It was difficult to determine the drainage areas of sites v l - v 2 because the urban storm sewer pipe network has not clearly been mapped by the municipality. However, the land use of the drainage areas of these two sub-watersheds is similar to sub-watersheds v3 and v4, and as a result similar levels of imperviousness were determined for sub-watersheds v l and v2. Figure 4-7 and 4-8 show an aerial photo in 1973 and orthophoto in 2006 of the sub-watersheds al-a3, while Figure 4-9 and 4-10 present those of a4. Land use in the sub-watersheds al-a3 is 37 presented in Figure 4.-11 and 4-12 for 1973 and 2006, respectively. Figure 4-13 and 4-14 are for a4 in 1973 and 2006, respectively. Appendix C shows aerial images of all the other areas. Land use, and generalized and detailed land use values in all the other sub-watersheds are also provided in Appendix D. The percentage of the different land uses for each sub-watershed in 1973 and 2006 is summarised in Figure 4-15 and 4-16. In 1973, the Alouette sub-watersheds (al-a3) were already highly urbanised, while those in the Kanaka Creek-were only slightly urbanised. The other sub-watersheds were mostly rurally/urban fringe areas or were partially covered by forests. In 2006, the sub-watersheds in the Alouette and the N/W Vancouver showed the highest level of urbanisation. Those in the Kanaka and Hatzic sub-watersheds were moderately urbanised, while those in the Thornhill sub-watersheds had a low level of urbanisation. The sub-watersheds hl-2 and t l were mostly covered by intact forests and are used as reference sites. 38 Figure 4-7 An aerial photo of the sub-watersheds a l , a2 and a3 in Alouette in 1973. Figure 4-8 An orthophoto of the sub-watersheds a l , a2 and a3 in Alouette in 2006, 39 Figure 4-9 An aerial photo of the sub-watershed a4 in Alouette in 1973. Figure 4-10 An orthophoto of the sub-watershed a4 in Alouette in 2006. 40 — Watershed boundary Stream • Sampling site Land use 1973 VZA Forest Figure 4-11 Land use in the sub-watersheds a l , a2 and a3 in Alouette in 1973. — Watershed boundary Stream • Sampling site Land use 2006 WA Forest Figure 4-12 Land use in the sub-watersheds a l , a2 and a3 in Alouette in 2006. 41 — Watershed boundary Stream • Sampling site Land use 1973 WA Forest m i Grass-agriculture • Rural -urban H Dense-urban • I Commercial- Industrial • Others 0 150 300 Figure 4-13 Land use in the sub-watershed a4 in Alouette in 1973. — Watershed boundary • • Stream • Sampling site Land use 2006 V2A Forest EH Grass-agriculture • Rural-urban H i Dense-urban • i Commercial- Industrial O Others r u ~ u 0 150 300 Figure 4-14 Land use in the sub-watershed a4 in Alouette in 2006. 42 • ~ * s - ~ * ^ ^ * - * - - s z £ Z S Z x : ^ ra . * ro re ro Watershed 3 Forest • Dense-urban Grass-agriculture • Rural-urban 8 Commercial-industrial • Others Figure 4-15 Land use proportion in the whole sub-watersheds in 1973. 100% 80% 60% 40% 20% 0% i t C M T - T - C O I > « - « ) T - C « J ' ^ , « O r « - l O C O I O < N I ^ - C O ^ - « ) U ) ^ ' T - C S C O C O ^ - U > « ) SZ_£+J^+'+J^4-'+-'SZS:-,-'+JJZ^ mszsz^m^t re re re > > > > Watershed B Forest Grass-agriculture • Rural-urban ** Dense-urban BB Commercial-industrial • Others Figure 4-16 Land use proportion in the whole sub-watersheds in 2006. 43 4.1.3 Land cover Figure 4-17 and 4-18 show land cover in al-a4 sub-watersheds in 2006. Land cover proportions in 1973 and 2006 for each sub-watershed are summarised in Figure 4-19 and 4-20, respectively. Land cover in the other watersheds was digitised and the data are presented in Appendix E . In 1973, the Alouette sub-watersheds (al-a3) had a high proportion of imperviousness ranging from 32 to 45%. A l l the other sub-watersheds had low proportion of imperviousness, less than 10%, except for k4-k6 (12%, 11% and 10%, respectively). B y 2006, the Alouette sub-watersheds (al-a3) showed an increased proportion of imperviousness ranging from 50 to 73%. N / W Vancouver sub-watersheds had the highest proportion of imperviousness with the minimum of 38% and maximum of 87% (v3-v6). Nine sub-watersheds had 10-20%) imperviousness (a4, h3-h7, k4-k6), and the other twelve sub-watersheds had imperviousness of less than 10% (hl-h2, k l - k 3 , tl-t7). —Watershed boundary Stream Figure 4-17 Land cover in the sub-watersheds al, a2 and a3 in Alouette in 2006. 44 — Watershed boundary Stream • Sampling site Land cover 2006 Figure 4-18 Land cover in the sub-watershed a4 in Alouette in 2006. re.*.*.* re re re Watershed s Forest-cover Grass-cover • Impervious-surface • Others Figure 4-19 Land cover proportion in the whole sub-watersheds in 1973. 4 5 C N T - T - C O I ^ O T - C N ^ C O I ^ U ^ C O I O C N T r c O T f t O U I ^ T - C N C O C O T l - L n C O £ R + J ^ + J + < I I * J * - £ £ + J + J I . j ( ; C 8 £ £ J > £ ^ ^ (0 (0 CO > > > > Watershed S Forest-cover Grass-cover • Impervious-surface • Others Figure 4-20 Land cover proportion in the whole sub-watersheds in 2006. 4.1.4 Land use in a 100 m buffer zone Figure 4-21 shows the land use within a 100 m buffer zone along both sides of the streams in sub-watershed a4. The different land uses within a 100 m buffer zone is shown in Figure 4-22 and 4-23, for 1973 and 2006 respectively. Figure 4-24 and 4-25 are for land cover proportions within a 100 m buffer zone for 1973 and 2006. The urban land use class consisted of "rural-urban", "dense-urban" and "commercial-industrial" categories. Detailed values of land use and land cover within a 100 m buffer zone for 1973 and 2006 are presented in Appendix F. In 1973, the proportion of urban land use in the sub-watersheds a2-a3 was 45-50%, whereas the other sub-watersheds had a lower proportion with less than 20%. The proportion of urban land use increased from 1973 to 2006, with nine sub-watersheds exceeded 30% of urban land use in 2006. The proportion of imperviousness in 1973 was generally small, with the maximum of about 20%> in the sub-watersheds a2-a3. Imperviousness increased significantly from 1973 to 2006 especially for a l (from 6% to 31%). In the sub-watersheds a2-a3 it increased from 20-25% to 35 40%. N / W Vancouver sub-watersheds had very a high proportion of imperviousness (about 35 85%). • 100 m buffer Land use 2006 Forest E H Grass-agriculture • Rural-urban Dense-urban Figure 4-21 Land use within a 100 m buffer zone in the sub-watershed a4 in Alouette in 2006. 47 £ £ J £ * ' * ' * ' * ' * , ^ * ' - , - , x : £ £ £ * re .* ro re re Watershed 3 Forest Grass-agriculture • Urban • Others Figure 4-22 Land use proportion within a 100 m buffer zone in 1973. 100% 80% 60% 40% 20% Q % !B B i B 58 B B ' B 8 i B B , B B i a i B i B i B i g i a H B B i g i g i a B i H . B • £ * ' £ * ' * ' £ : £ * ' * ' £ £ j £ £ j i ' * : ' + J j i J i re re re re > > > > Watershed 3 Forest Grass-agriculture • Urban • Others Figure 4-23 Land use proportion within a 100 m buffer zone in 2006. 48 100% 80% 60% 40% 20% 0% N T - r t T - « ^ ( O S T - N l O ( O I O « ^ ( N ^ ( O I O * T - « » > N szsz*+J+J~+J*'**J+Jszszszszx! ro re re re Watershed B Forest-cover Grass-cover • Impervious-surface • Others Figure 4-24 Land cover proportion within a 100 ni buffer zone in 1973. sz*Jsz*J*Jszsz*J*Jszsz^sz^*i*^^^ to ^  ^  re re re > > > > Watershed B Forest-cover Grass-cover • Impervious-surface • Others Figure 4-25 Land cover proportion within a 100 m buffer zone in 2006. 49 4.1.5 Road density Figure 4-26 presents the road density (km/km 2) in the sub-watersheds, together with their imperviousness. Detailed values for both the sub-watershed and a 100 m buffer zone are provided in Appendix G . The N / W Vancouver sub-watersheds have the higher road density (20-2 2 30 km/km ) followed by the Alouette sub-watersheds (about 15 km/km ). in c o T3 "O o 35 30 25 20 15 10 I Road density Impervious-surface innnlliiiiniWllli 100 80 60 40 .2 < M T - T - C O r - t D T - C > J ^ - l D r ~ I O O l O C M ^ - < » > ^ - < O I O ^ - T - « M O C O ^ - l O t D Watershed Figure 4-26 Road density and imperviousness in studied sub-watersheds in 2006. o in m 3 O E Q. E 20 -4.1.6 Historical changes in land variables Historical changes in land use The historical changes in land use between 1973 and 2006 are shown in Figure 4-27, first for the whole sub-watershed, and in Figure 4-28 for a 100 m buffer zone. Among the spatially connected sub-watersheds (e.g., k4-k6), only the sub-watersheds at the headwaters are displayed. The sub-watershed h4 represents the typical land use among h3-h7 and was used as a representative example. Detailed historic change values are presented in Appendix H . "Forest" decreased and urban land use increased both in the whole sub-watershed and within a 100 m buffer zone for most sub-watersheds. In al-a3, "commercial-industrial" highly increased in the 50 whole sub-watershed, and "dense-urban" category increased significantly within a 100 m buffer zone. Most sub-watersheds saw a 10-20% urbanisation, and some were in the 20-30% range. For land cover, most sub-watersheds saw a decrease in "forest-cover", and increases in "grass-cover" and "impervious-surface" both in the whole sub-watershed and within a 100 m buffer zone (Figure 4-29 and 4-30). Detailed values are provided in Appendix H . Only the Alouette sub-watersheds saw a decrease in "grass-cover". In contrast, grass cover increased significantly in the Thornhill watersheds. "Impervious-surfaces" increased most in the Alouette watershed. Only sub-watersheds al-a3 saw an increase in "commercial-industrial" land use over the last thirty years, whereas "rural-urban" land use increased in most other sub-watersheds. This is consistent for land use within a 100 m buffer zone. A n increase in "grass-cover" was dominant in rurally urban sub-watersheds in the Hatzic, Kanaka and Thornhill. Imperviousness has conspicuously increased in the sub-watersheds al-a3. 60% 40% §> 20% c re .c " 0% o -20% -40% • Forest Grass-agriculture • Rural-urban M Dense-urban • Commercial-industrial 4 hi h2 I , h4 k1 I K2 V% |<4 t1 |t Tjt Watershed Figure 4-27 Historical changes in land use in the whole sub-watershed (1973-2006). Note that the sub-watersheds at the headwaters are shown, yet h4 is shown instead of h3 (and so forth). 51 -40% J Watershed Figure 4-28 Historical changes in land use within a 100 in buffer zone (1973-2006). -40% J Watershed Figure 4-29 Historical changes in land cover in the whole sub-watershed (1973-2006). 52 30% 20% • 10% TO | 0% O -^10% -20% -30% I Forest-cover Grass-cover I Impervious-surface i l l I . l l I |1 |2 |3 |4 hi h2 |i4 |1 |2 |3 *4 t1 |2 t3 |6 Watershed Figure 4-30 Historical changes in land cover within a 100 m buffer zone (1973-2006). 4.1.7 Factor analysis on land variables Land variables in 1973 The results of factor analysis for the 1973 data are presented in Figure 4-31 (the Broken-stick model was used) and in Table 4-2 (factor loadings of varimax rotation). Comparing the eigenvalues of factors before rotation with those of the Broken-stick model, the first two factors contain most of the information. However, the third eigenvalue is also close to that of the Broken-stick model, and hence three factors were extracted in the subsequently performed rotations. The results with the varimax rotation and promax rotation were very similar (Appendix I), and the former method was applied since it is the most common method of factor analysis and achieved a simple structure. 53 -m- Land use 1973 Broken Stick Figure 4-31 Eigenvalues of the principal component factor analysis for land use in 1973, and those of the Broken stick model. Looking at factor loadings, the "impervious-surface" and the "dense-urban" category both in the whole sub-watershed and within a 100 m buffer zone have the highest positive correlation with factor 1. "Forest" (land use) and "forest-cover" (land cover) in the whole sub-watershed have a moderate negative correlation with factor 1. "Grass-agriculture" and "grass-cover" both in the whole sub-watershed and within a 100 m buffer zone are negatively correlated with factor 2, and "forest" (land use) and "forest-cover" (land cover) both in the whole sub-watershed and within a 100 m buffer zone are negatively correlated with factor 2. "Coarse-textured material" has a high negative correlation with factor 3, whereas "fine-textured mineral" has a high positive correlation with factor 3 as well as the "rural-urban" category both in the whole sub-watershed and within a 100 m buffer zone. The "dense-urban" and "impervious-surface" categories are the most important factor 1 loadings (positive). "Grass-agriculture" and "grass-cover" represent the factor 2 loadings (positive) as well as "forest (B)" (negative). "Coarse-and fine textured materials" represent the factor 3. 54 Table 4-2 Factor loadings of varimax rotation and rotation sums of squared loadings for 1973 land variables. Correlations higher than 0.5 (as an absolute value) are shown, and bold presents correlation>0.7. (B) denotes the land use/ cover within a 100 m buffer zone. Communality means the proportion of each variable explained by the extracted factors. FactoM Factor2 Factor3 Communality Elevation -0.70 0.75 Slope -0.68 0.69 Stream-gradient 0.15 Coarse-mineral -0.87 0.94 Fine-mineral 0.87 0.93 Forest -0.66 -0.61 0.98 Forest (B) -0.82 0.96 Grass-agriculture 0.87 0.88 Grass-agriculture (B) 0.97 0.97 Rural-urban 0.82 0.86 Rural-urban (B) 0.91 0.89 Dense-urban 0.96 0.96 Dense-urban (B) 0.93 0.88 Commercial-industrial 0.89 0.80 Commercial-industrial (B) 0.23 Forest-cover -0.72 -0.55 0.98 Forest-cover (B) -0.76 0.95 Grass-cover 0.79 0.95 Grass-cover (B) 0.92 0.95 Impervious-surface 0.94 0.98 Impervious-surface (B) 0.90 0.92 Rotation Sums of Squared Loadings %Variance explained Cumulative contribution% 6.4 6.1 5.1 30 29 25 30 59 84 Land variables in 2006 The same factor analysis was performed on the land variables in 2006. A s before, three factors were extracted using the varimax and promax rotation. Appendix I shows the graph of eigenvalues before rotation and factor loadings with promax rotation, respectively. Looking at factor loadings (Table 4-3), "forest-cover" both in the whole sub-watershed and within a 100 m buffer zone is positively correlated with factor 1, whereas "impervious-surface" and "road-density" both in the whole sub-watersheds and within a 100 m buffer zone are negatively correlated with factor 1. "Grass-cover" both in the whole sub-watershed and within a 100 m buffer zone is positively correlated with factor 2, as well as "grass-agriculture" and "rural-urban". "Coarse-mineral" is positively correlated with factor 3, as well as topographical variables. "Fine-mineral" is negatively correlated with factor 3. From the loadings, "forest-cover" and 55 "impervious-surface" are the key variables loading on factor 1 while "grass-cover" was the key variable loading on factor 2, and "coarse/fine-textured material" was the key variable on the factor 3 loadings. Factor analysis on land variables achieves distinct categorization of the variables for both 1973 and 2006 data. Land variables within a 100 m buffer zone have similar component loadings as those in the whole sub-watershed, but the correlations are stronger for the whole sub-watershed with factor 1 in 1973. For the 2006 data, the loadings of a 100 m buffer zone are mostly similar to those in the whole sub-watershed. Table 4-3 Factor loadings of varimax rotation and rotation sums of squared loadings for 2006 land variables. Correlations higher than 0.5 (as an absolute value) are shown, and bold means correlation>0.7. (B) denotes the land use within a 100 m buffer zone. Communality means the proportion of each variable explained by the extracted factors. Factorl Factor2 Factor3 Communality Elevatoin -0.59 -0.52 0.55 0.91 Slope 0.80 Stream-gradient 0.37 Coarse-mineral 0.96 0.94 Fine-mineral -0.97 0.95 Forest -0.87 0.94 Forest (B) -0.88 0.95 Grass 0.79 0.83 Grass (B) 0.85 0.84 Rural-urban -0.53 0.70 0.86 Rural-urban (B) 0.57 Dense-urban 0.70 0.50 Dense-urban (B) 0.72 0.54 Commercial-industrial 0.73 0.61 Commercial-industrial (B) 0.68 0.52 Forest-cover -0.91 0.93 Forest-cover (B) -0.93 0.95 Grass-cover 0.87 0.81 Grass-cover (B) 0.92 0.86 Impervious-surface 0.93 0.93 Impervious-surface (B) 0.96 0.96 Road-density ' 0.97 0.97 Road-density (B) 0.92 0.89 Rotation Sums of Squared Loadings 10.2 5.0 3.2 %Variance explained 44 22 14 Cumulative contribution% 44 66 80 56 4.1.8 Summary of the section 1 In 1973, the Alouette sub-watersheds (al-a3) were already highly urbanised, with a high proportion of imperviousness ranging from 32 to 45%, while those in the Kanaka Creek were only slightly urbanised (imperviousness of about 10%>). The other sub-watersheds were mostly rurally/urban fringe areas or were partially covered by forests. In 2006, the sub-watersheds in the Alouette and the N / W Vancouver showed the highest level of urbanisation (an increased proportion of imperviousness ranging between 50-73% and 38-87%>). The other sub-watersheds were mostly rural (10-20%) imperviousness) and some were still covered with intact forests. Land use/ cover within a 100 m buffer zone was mostly proportional to those within the whole sub-watershed, and the imperviousness in al-a3 was 25-40% in 2006. A 100 m buffer zone covered the drainages of N / W Vancouver. Road density was calculated for the 2006 data, and the N / W Vancouver sub-watersheds have the higher road density (20-30 km/km ) followed by the Alouette sub-watersheds (about 15 km/km ). Between 1973 and 2006 most sub-watersheds saw an increased of 10-20% in the amount of urbanisation, and some were in the 20-30% range. Most sub-watersheds saw a decrease in "forest-cover", and increases in "impervious-surface". Imperviousness has conspicuously increased in the sub-watersheds al-a3 (10-30%). Factor analysis using the 1973 land variables identified the "dense-urban" and "impervious-surface" to be the most important factor 1 loadings; "grass-agriculture" and "grass-cover" to be the factor 2 loadings; "coarse/fine-mineral" to be the factor 3 loadings. In 2006, "forest-cover" and "impervious-surface" were the key variables loading on factor 1 while "grass-cover" was the key variable loading on factor 2, and "coarse/fine-mineral" was the key variable on the factor 3 loadings. The loadings of a 100 m buffer zone are mostly similar to those in the whole sub-watershed both in 1973 and 2006. 57 4.2 Sediment Analysis 4.2.1 Aqua regia and loss on ignition Accuracy and precision The aim of aqua regia was to analyse the 2006 sediment samples and compare them with samples collected in 1973. This is difficult because the historic samples were not available for re-analysis and the analytical instrumentation and extraction techniques have changes since that time. The same digestion technique was used as in other L F V sediment studies carried out over the past 15 years a t ,UBC (McCal lum 1995; Brydon 2004) (Figure 4-32). Although the accuracy of sediment quality results was not determined in this study, the data were compared to these previous studies (Table 4-4). Na , N i and Cr levels in this study were higher than those referenced by (Brydon 2004). Table 4-4 Comparison of sediment quality between this study and Brydon (2004) at the same sites (v5 and v6) Element v5 v6 This Brydon This Brydon study 2004 study 2004 Cu 300.7 278.6 540.7 642.1 Pb 353.7 237.4 132.7 218.3 Zn 813.1 666.8 712.4 1072.6 Cr 88.5 38.2 142.1 81.3 Ni 47.2 21.7 39.5 26.0 Fe 4.54 2.55 3.29 2.04 Mn 802.7 578.3 328.9 274.6 Al 2.04 1.25 1.09 1.02 P 2093.0 1663.4 1027.1 1193.7 K 889.5 703.6 791.2 869.6 Mg 4527.8 2901.1 4826.2 3845.9 C a 12152.0 8657.0 5952.8 5253.7 Na 971.9 363.3 881.7 397.3 58 I 3 Cu • I -H973 12006 2004 1993 o o CM J£ E . o « i a i i £ £ j : j < x > > = > > = > u j a = U j S IA (O C J2 ~ 1= > > " o w i m Zn Hi . I l l ^ ,i ' s s = « 1 S g 00 Figure 4-32 Comparison of the sediment quality data of this study with those of McCallum (1995) and Brydon (2004). Study was done at the same sites (v5 and v6) in Brydon (2004) (denoted as v5-JB and v6-JB) as well as SD, Tempe and Oakalla in 2006. Sediment quality study was also done in the Brunette River sub-watershed in 1993 (Still C. and Brunette). Sample blanks were all below detection limit of the ICP-AES instrument, indicating no contamination occurred during laboratory procedures. A l l sediment quality data were above detection limits of ICP-AES except for Cd. Average blank levels are shown in Appendix B together with lowest detection limits. The duplicates data of the sub-watersheds a2, k4 and t3 were shown in Table 4-5. % difference was calculated by the difference between the duplicates divided by the bigger value. Average of % difference was also calculated. Duplicates presented no large variability among the samples 59 except for C d (as the concentration of C d is very trace). This indicates that no significant deviations existed in laboratory analysis. Table 4-5 Duplicates data of laboratory analysis of sediment quality for selected sub-watersheds (a2, k4 and t3). NA of Cd shows the data were below the detection limit. Watershed a2 k4 t3 a2 k4 t3 a2 k4 t3 a2 k4 t3 Element Cu Pb Zn Cd % difference 3.6 16.0 11.1 11.1 15.0 28.8 7.9 13.5 5.4 73.9 NA NA Average of % difference 10.2 18.3 8.9 73.9 Element Cr Ni Co % difference 8.9 32.5 7.4 11.7 33.0 8.3 9.7 9.5 0.3 Average of % difference 16.2 17.6 6.5 Element Fe Mn Al P % difference 11.3 9.4 1.4 10.5 6.9 1.6 12.2 12.2 1.4 5.9 5.4 3.8 Average of % difference 7.4 6.3 8.6 5.0 Element K Mg Ca Na % difference 12.0 14.9 1.3 9.9 8.4 1.6 12.8 18.7 1.6 1.9 18.0 9.9 Average of % difference 9.4 6.6 11.0 9.9 Elements in the sediments The results of aqua regia and I C P - A E S and loss on ignition (LOI) are shown in Figure 4-33 for the elements together with the sediment quality guideline in Canada and the United States (USGS). Chroma after L O I and nitrate levels in surface water were examined, but did not provide comprehensive results (Appendix J). Canadian sediment guidelines: • Cu : Sediments from sites v2-v6 greatly exceeded P E L . The sites al-a3 in 1973 and 2006 and v l in 2006 were slightly above T E L . • Pb: Sediments from sites v2-v6 in 2006, and a l , a2 and t6 in 1973 largely exceeded P E L . The other sites were mostly below T E L . • Zn: Sediments from site a2, v2-v6 in 2006 were above P E L , and al-a3 were above T E L in 1973. • Cd : Sediments from sites a l in 2006 and v l - v 6 exceeded T E L but no site exceeded P E L . 60 • Cr: Sediments from sites h4, v4 and v6 were above P E L , and al-a4, v3 and v5 were above T E L . In terms of U S G S guidelines, • Fe: Sediments from all sites did not exceed T E L . • A l : Sediment from sites al-a4 slightly exceeded T E L but none of the sites exceeded P E L . • N i : Sediments from all Alouette sites (al-a3, in both 1973 and 2006) and N / W Vancouver sites (v3-6) exceeded P E L , as well as k4 and t3 in 2006 and t l in 1973. • M n : Sediments from sites a2, a3 and h4 in 2006 exceeded P E L . They were also above T E L in 1973, and many of the sites were above T E L levels in 1973 and 2006. P E L were greatly exceeded in the N / W Vancouver sites for Cu , Pb and Zn, whereas those in other sub-watersheds including the urbanised Alouette sub-watersheds were mostly below T E L or between T E L and P E L , although Z n levels at several Alouette sites slightly exceeded P E L . Cr levels also exceeded P E L at a few sites, yet the exceedances were relatively small compared to Cu , Pb and Zn. Organic matter contents at the sites in the Thornhill and the N / W Vancouver sub-watersheds were higher (2-4%) than those in the Alouette, the Kanaka and the Hatzic sub-watersheds (0.5-2%o) (Figure 4-33, f). Detailed sediment quality data for 1973 and 2006 are presented in Appendix J. 61 O) o J; 01 E o o CM Cu-1973 • Cu-2006 TEL (Canada) — PEL (Canada) ro to ro mszszjz^c^c^c^*^-^ > > > > (a) Cu o o Pb-1973 • Pb-2006 TEL (Canada) — PEL (Canada) ro ro ro ro si Catchment (b) Pb 0)0 E "~ Zn-1973 • Zn-2006 TEL (Canada) — PEL (Canada) I j I j _ . | u _ ~ ~ (c) Zn at E • Cd TEL (Canada) — PEL (Canada) I l l (d) Cd 160 120 • Cr TEL (Canada) — PEL (Canada) 40 30 -&20 10 Co-1973 I Co-2006 ro ra ro ra sz sz sz Figure 4-33 Metal concentrations in the sediments in 1973 and 2006, and threshold effect level (TEL) and probable effect level (PEL) of the guidelines in Canada and U.S. 62 60 JJ40 "a E 20 Ni-1973 • Ni-2006 TEL (USA) — PEL (USA) 11 (a) Ni m m CM O CN Fe-1973 I Fe-2006 TEL (USA) •PEL (USA) l l I I I I I I l I | I I I I I I A3 VO (0 10 (b) Fe IAI TEL (USA) — PEL (USA) fflfflffllB£££^^^*'*i*J>>>> « Q A i * o o o E Mn-1973 • Mn-2006 TEL (USA) — PEL (USA) III (d) Mn (e) Phosphorus (Q Organic matter (LOI) Figure 4-33 Continued. 4.2.2 Historical changes in sediment quality Concentrations increased at most sites for all heavy metals (Co, Cu, Fe, Mn, Ni and Zn) except for Pb between 1973 and 2006 (Figure 4-34). Pb concentrations greatly decreased at sites al, a2. Sites t l , and t6 saw decreases in most metals except for Cu and Pb. Detailed values of historical changes are shown in Appendix J. Caution is needed in these interpretations due to the differences in analytical procedures between 1973 and 2006, and therefore only the large differences in metal concentrations should be considered as indications of increased contamination. 35 30 o,25 l> 20 !« ° 10 5 0 lllllll Mil a1 a2 a3 a4 hi h2 h4 k1 k3 k4 tl t3 t6 Watershed (a) Cu o o 04 Ol J2 CL O O CO a1 a2 a3 a4 hi h2 h4 k1 k3 k4 t1 t3 t6 Watershed (b) Pb 200 150 100 O) E c 50 N 0 -50 -III • III II a1 a2 a3 a4 hi h2 h4 k1 k3 k4 t1 t3 t6 Watershed (c) Zn 15 10 D) 5 K Ol „ E 0 o O -5 -10 -15 h l.llll.ll i1 a2 a3 a4 hi h2 h4 k1 k3 k4 t1 t3 t6 Watershed (d) Co 30 20 oi 10 oi E 0 10 20 30 a1 a2 a3 a4 hi h2 h4 k1 k3 k4 t1 t3 t6 Watershed (e)Ni 4 3 2 ^ 1 LL 0 -1 -2 h2 a1 a2 a3 a4 hi h4 k1 k3 k4 t1 t3 t6 Watershed (0 Fe Figure 4-34 Historical changes in sediment concentrations. 64 o o o C N S o E o l l l l - . l l l l . l l a1 a2 a3 a4 hi h2 h4 k1 k3 k4 t1 t3 t6 o o o Watershed (s)Mn Figure 4-34 Continued. 4.2.3 Factor analysis on sediment quality Sediment quality in 1973 A factor analysis with varimax rotation was performed using all metals from the 1973 data. Component loadings and rotation sums of squared loadings are shown in Table 4-6. Co and Mn showed the highest correlations with factor 1, and Fe and Ni were also highly correlated with factor 1. Pb showed the highest correlation with factor 2, followed by Cu and Zn. Thus factor 1 contains transitional trace metals (Co and Ni) and oxide-forming metals (Fe and Mn), whereas factor 2 has soft acid metals (Cu, Pb and Zn). 65 Table 4-6 Component loadings and rotation sums of squared loadings for sediment quality data in 1973. Only correlations greater than r= 0.5 are shown, and bold means a correlation r=>0.7. Communality means the proportion of each variable explained by the extracted factors. F a c t o r l Factor2 C o m m u n a l i t y C u 0.60 0.78 0.97 P b 0.89 0.99 Z n 0.65 0.70 0.91 C o 0.84 0.96 Ni 0.74 0.50 0.91 F e 0.79 0.99 M n 0.91 0.94 Rotat ion S u m s of S q u a r e d L o a d i n g s o .o z . o % V a r i a n c e exp la ined 0.50 0.35 C u m u l a t i v e contribution 0.50 0.85 Sediment quality in 2006 A factor analysis was performed for the 2006 sediment quality data. The first three eigenvalues were greater than those of the Broken-stick model before rotation (Appendix K ) , and thus three factors were extracted by varimax rotation (Table 4-7). Appendix K shows the factor loadings with promax rotation. Highest loadings on factor 1 were Cu , Pb, Zn , Cd , and phosphorus. Cr and Ca are also moderately correlated with factor 1. Co is strongly correlated with factor 2, and Fe, M n and A l are moderately correlated with factor 2. N a has a moderate negative correlation with factor 2. K and M g are strongly correlated with factor 3. 66 Table 4-7 Factor loadings with varimax rotation for sediment quality data in 2006. Only correlations greater than 0.5 (as an absolute value) are shown, and bold presents a correlation>0.7. P - phosphorus, and O M - organic matter. Communality means the proportion of each variable explained by the extracted factors. F a c t o r l Factor2 Factor3 C o m m u n a l i t y C u 0.92 0.92 P b 0.93 0.90 Z n 0.97 0.93 C d 0.94 0.90 C o 0.86 0.96 C r 0.70 0.54 Ni 0.64 0.84 F e 0.68 0.51 M n 0.67 0.50 Al 0.66 0.70 N a -0.70 0.62 0.88 M g 0.97 0.97 K 0.96 0.96 C a 0.69 0.68 P 0.79 0.67 O M -0.58 0.47 Rotat ion S u m s of S q u a r e d L o a d i n g s % V a r i a n c e exp la ined Cumulative contribution% 5.7 3.5 3.0 36 22 19 36 58 77 4.2.4 Cluster analysis by sediment quality data Clustering of watersheds using the 1973 metal data Watersheds were classified by cluster analysis using the factor scores of the sediment quality data in 1973. However, this did not provide meaningful clustering of the sub-watersheds. Clustering of watersheds using the 2006 metal data Figure 4-35 shows the dendrogram with the Ward method, clustered by factor scores of the three factors of the 2006 sediment data. Table 4-8 presents the clustering of the sub-watersheds. Using the Complete Linkage method gave the same results (Appendix L ) . 67 R e s c a l a d D i s t a n c e C l u s t e r Combine C A S E L a b e l Num k2 13 k5 16 t 3 20 t l 24 k l 12 h5 9 he 10 h4 S t4 21 t 5 22 t2 19 t6 23 h7 11 h i 5 h2 6 h3 1 k3 14 t l 18 a4 4 k4 15 k6 XI a l 1 a3 3 a2 2 v3 25 v6 28 26 v5 21 10 15 —+-20 25 --+ J J Figure 4-35 Dendrogram of the Ward method of cluster analysis. Factor scores of 2006 sediment quality data were used. Table 4-8 Clustering sub-watersheds according to factor scores of sediment quality data in 2006. E l e m e n t s S u b - w a t e r s h e d C l u s t e r l h1-h3, t1, k3 Gluster2 h4-h7, k1-k2, k5, t2-t7 Cluster3 a1-a4 , k4, k6 Cluster4 v3-v6 These clusters are widely different from one another in terms of sediment quality. Looking at sub-watersheds clusters, cluster 1 contains only intact forested sub-watersheds; cluster 2 contains 68 rural sub-watersheds; cluster 3 contains urbanised sub-watersheds; and cluster 4 contains the N / W Vancouver sub-watersheds which were heavily urbanised. 4.2.5 Summary of section 2 According to the result of aqua regia and I C P - A E S , P E L were greatly exceeded in the N / W Vancouver sites for Cu , Pb and Zn, whereas those in other sub-watersheds including the urbanised Alouette sub-watersheds were mostly below T E L or between T E L and P E L , although Z n levels at several Alouette sites slightly exceeded P E L . Concentrations increased at most sites for all heavy metals (Co, Cu , Fe, M n , N i and Zn) except for Pb between 1973 and 2006, but the levels of increases varied widely. A factor analysis on the sediment quality data in 1973 identified that factor 1 contains transitional trace metals (Co and N i ) and oxide-forming metals (Fe and Mn) , whereas factor 2 has soft acid metals (Cu, Pb and Zn). A factor analysis performed for the 2006 sediment quality data identified soft acid metals (i.e., Cu , Pb, Zn, Cd , phosphorus) as the key variables of factor 1, metals forming oxides (i.e., Fe, M n , A l ) as the key variables of factor 2, and M g and K to be of factor 3. Cluster analysis of the 2006 sediment data using the Ward method and the factor scores of the first three factors classified the sub-watersheds into four groups: cluster 1 (intact forested); cluster 2 (rural sub-water sheds); cluster 3 (urbanised); and cluster 4 (the N / W Vancouver sub-watersheds). 4.3 Land Use-Sediment Interaction 4.3.1 Linking imperviousness with metal concentrations Historical changes in metal concentrations from 1973 to 2006 were plotted against the changes in imperviousness from 1973 to 2006 (Figure 4-36), for Cu , Pb, Zn , Co , N i , Fe, and M n . Sediment quality in forested sub-watersheds in Hatzic and Thornhill fluctuated without a notable increase in imperviousness. Sub-watersheds in rurally urban environment (i.e., with imperviousness < 20%) saw a comparatively large increase in metal concentrations with an increase in imperviousness, for all metals except Pb. On the other hand, an increases in metal concentrations 69 in the Alouette sub-watersheds (with imperviousness > 30%) were less dramatic than increases in metal in rural watersheds with similar increases in imperviousness. Z n was the exception. o OO to o -I Imperviousness < 20% O 2000s + 1970s 10 — i — 15 l o o 00 o 5 O 2000s Imperviousness > 30% +• 1970s 30 40 60 — i — 70 l o Ziinpervious surface o CO to e Imperviousness < 20% 0 2000s + 1970s o eo o o CM o o O 2000s + 1970s Imperviousness > 30% 10 15 20 30 40 l o ~ 60 70 80 % impervious surface Figure 4-36 Historical changes of imperviousness and metal concentrations between 1973-2006. Arrows represent the changes from 1973 to 2006. " O " denotes data in 2006, and "+" denotes that in 1973. 70 o tn CM o • CN to » -r 5 ~ 0 2000s 4- 1970s Imperviousness < 20% to CO o CO CN o CM o 0 2000s 4- 1970s 0 4° Imperviousness > 30% Y 10 15 20 30 40 limpervious surface 50 60 70 80 CN o . CN Imperviousness < 20% O 2000s 4- 1970s 10 15 20 m CO LA CM o CM O 2000s Imperviousness > 30% +. 9^703 30 40 50 To Ximpervious surface 80 o to G O .— CM O 2000s 4- 1970s Imperviousness < 20% T-0 o cn o to o 0 2000s 4- 1970s 10 15 20 Iimpervious surface Imperviousness > 30% 30 40 50 60 70 80 Figure 4-36 Continued. 71 CO <0 Imperviousness < 20% 0 2000s 4- 1970s Imperviousness > 30% 0 o 0 2000s •I- 1970s 10 15 20 30 40 50 60 70 80 {impervious surface g 2 0 Imperviousness < 20% 4 O / f /ft"/** / 'jl/f/f ° 2000s 1970s o Imperviousness > 30% O 2000s + 1970s 10 15 20 30 40 Ximpervious surface i 1 1-50 60 70 80 Figure 4-36 Continued. 4.3.2 Relationships between Imperviousness vs. metal concentrations in the 2006 sediments in the Alouette and N/W Vancouver sub-watersheds Metal concentrations vs. imperviousness were plotted for the 2006 data for the Alouette and N / W Vancouver sites (Figure 4-37). Concentrations of Cu , Pb, Zn , Cd , Cr, Ca, Na , Phosphorus and organic matter were higher in N / W Vancouver than those in Alouette except for a few cases (i.e., v6 of C a and phosphorus, and v5 of organic matter). Concentrations of A l , M n , Co, N i , K and M g in Alouette sites were greater than in N / W Vancouver. It should be noted that Cu , Pb and Z n levels in N / W Vancouver were above P E L , whereas those in Alouette were above/below T E L (only the Z n level in a2 is slightly above P E L ) . 72 Cu v3 v4 v6 v5 a 3 a 1 a2 40 50 60 70 80 Impervious sur face {%) Pb v3 v5 v4 a3 a1 a2 v6 40 50 60 70 80 Impervious sur face (%) Zn v3 J a 3 a l a2 v6 40 50 60 70 80 Impervious sur face (X) Cd - ' j v 5 v4 v6 v3 j al a3 a2 40 50 60 70 80 Impervious sur face {%) Figure 4-37 Imperviousness vs. element concentrations in 2006 (Cu Pb Zn Cd). A symbol " a " denotes Alouette sub-watersheds, and "v " denotes N/W Vancouver sub-watersheds. 73 Fe 50 60 70 80 Impervious surface (%) Al 40 50 60 70 80 Impervious surface (%) Hn al • al Vit] v4 v6 v3 40 50 60 70 80 Impervious surface {%) Co I a 2 a 3 a l ; v5 v4 v6 40 50 60 70 80 Impervious surface (%) Cr Ni o vfi O a2 CO o CM - LA a3 bo o ] al O v5 y4 \ bo E LA v5 o v4 CO v3 o ^ a1 a'3 • •{ a2 i o v3 v6 i 40 i 50 i 60 i i 70 80 40 50 60 1 1 70 80 Impervious surface (30 Impervious surface (%) Figure 4-37 Continued. 74 Ca v3i v5i v4 a3 a I a2 v6 40 50 60 70 80 Impervious surface (%) \ a2 a3 \ al V3: v5 v 4 v6 40 50 60 70 80 Impervious surface (%) v3! Mg a"3 v5 a1 V4 a2 v6 40 50 60 70 80 Impervious surface (X) Na v5 v 4 v6 v3 a2 a3 !- -al i 40 50 60 70 80 Impervious surface (X) Figure 4-37 Continued. 75 Phosphorus Organic matter v5 2.0 2.5 v3 v4 v6 v4 v3 - " 1 al I a2 v6 | al a2 a3 O v * 3 T r 40 50 60 70 80 Impervious surface (X) ~i 1 1 r 40 50 60 70 80 Impervious surface (X) Figure 4-37 Continued. 4.3.3 Correlations between land variables and sediment quality data Spearman's rank correlations for 1973 data The results of the Spearman's rank correlation between land use variables and metals in 1973 sediments are provided in Table 4-9 (only significant correlations [p < 0.05] are presented). C u , Pb, Zn , and N i have high positive correlations with "impervious-surface" and "dense-urban" sites. Co showed the strongest correlations with these variables. They have high negative correlations with "forest" and "forest-cover", as well as high negative correlations with topographical variables. 76 Table 4-9 Spearman's rank correlation between land variables and sediment quality data in 1973. (B) denotes "a 100 m buffer zone". Only significant correlations are shown (p < 0.05), and bold presents that correlation is significant at the 0.01 level. Cu Pb Zn C o Ni Fe Mn Elevation -0.51 -0.42 -0.42 -0.43 -0.44 8 8- a -Slope -0.57 -0.54 -0.50 -0.50 Stream-gradient Text Coarse-mineral -0.49 -0.47 0.47 Text Fine-mineral 0.45 0.49 0.42 0.51 Forest -0.65 -0.58 -0.57 -0.63 -0.63 -0.47 Forest (B) -0.62 -0.43 -0.51 -0.57 -0.48 -0.46 Grass-agriculture 1 use Grass-agriculture (B) 1 use Rural-urban 0.42 Lane Rural-urban (B) 0.42 -0.48 Lane Dense-urban 0.73 0.62 0.63 0.64 0.75 0.52 Dense-urban (B) 0.54 0.56 0.50 0.51 0.58 Commercial-industrial 0.50 0.44 0.44 0.49 Commercial-industrial (B) Forest-cover -0.66 -0.58 -0.57 -0.64 -0.64 -0.47 cu > Forest-cover (B) -0.65 -0.47 -0.53 -0.59 -0.52 -0.48 o o Grass-cover 0.52 0.46 0.42 0.51 0.45 0.44 Grass-cover (B) 0.55 0.44 0.45 CD _ l 1 m pervious-surface 0.68 0.51 0.57 0.64 0.64 0.46 Impervious-surface (B) 0.60 0.56 0.43 0.52 0.56 Spearman's rank correlations for 2006 data The result of Spearman's rank correlation for the 2006 data is provided in Table 4-10 (only significant correlations [p < 0.05] are presented). C u , Pb and Z n had strong positive correlations with "impervious-surface" and "road-density". They also had high positive correlations with "dense-urban" sites. Z n had the strongest positive correlation with these variables. These metals had strong negative correlations with "forest" and "forest-cover". These metals had high negative correlations with topographical variables. N i had strong positive correlations with "dense-urban", "impervious-surface" and "road-density". It had high-strong negative correlations with "forest", "forest-cover", "elevation" and "slope". M g and K had a strong positive correlation with "fine-textured materials", as opposed to a strong negative correlation with "coarse-textured materials". Phosphorus had high positive correlations with "dense-urban", "impervious-surface" and "road-density" categories. Strong negative 77 correlations were found between phosphorus and "forest" and "forest-cover". Organic matter had no significant correlation with land variables. Table 4-10 Spearman's rank correlation between selected land variables and sediment quality in 2006. (B) denotes "a 100 m buffer zone". Only significant correlations are shown (p < 0.05), and bold presents that correlation is significant at the 0.01 level. Cu Pb Zn C d Fe Mn Al C o Elevation -0.56 -0.44 -0.73 -0.40 -0.58 2 9- C l Slope -0.41 -0.57 -0.39 -0.81 ro Stream-gradient Text ure Coarse-mineral -0.55 -0.76 Text ure Fine-mineral 0.40 0.60 0.79 Forest -0.48 -0.71 -0.85 -0.58 -0.41 Forest (B) -0.50 -0.80 -0.75 -0.59 Grass-agriculture -0.66 -0.47 0.49 1 use Grass-agriculture (B) -0.61 -0.38 0.40 1 use Rurally-urban -0.65 -0.43 -0.53 Lane Rurally-urban (B) -0.39 -0.53 Lane Densely-urban 0.48 0.52 0.74 0.38 0.52 0.53 Densely-urban (B) 0.62 0.55 0.72 0.43 0.52 0.47 Commercial-industrial 0.47 0.57 0.70 0.49 0.49 0.46 Commercial-industrial (B) 0.45 0.59 0.72 0.39 0.52 0.40 Forest-cover -0.48 -0.68 -0.84 -0.56 -0.44 Forest-cover (B) -0.56 -0.78 -0.81 -0.57 cu > Grass-cover 0.46 o o Grass-cover (B) 0.43 T3 1 m pervious-surface 0.57 0.67 0.88 0.54 0.39 0.48 ro Impervious-surface (B) 0.75 0.68 0.85 0.55 0.42 Road-density 0.64 0.78 0.86 0.57 0.40 Road-density (B) 0.70 0.64 0.72 0.42 0.42 78 Table 4-10 Continued. Cr Ni Na Mg K C a P O M Elevation -0.79 -0.86 -0.47 -0.46 -0.44 9 9- a -Slope -0.55 -0.78 -0.63 -0.62 Stream-gradient Text ure Coarse-mineral -0.52 0.47 -0.89 -0.87 Text ure Fine-mineral 0.56 -0.39 0.92 0.86 Forest -0.84 -0.66 -0.38 -0.51 -0.73 Forest (B) -0.82 -0.51 -0.39 -0.52 -0.75 Grass-agriculture -0.38 1 use Grass-agriculture (B) -0.39 1 use Rurally-urban -0.49 -0.49 Lane Rurally-urban (B) -0.39 -0.49 Lane Densely-urban 0.63 0.68 0.46 0.64 Densely-urban (B) 0.58 0.73 0.41 0.60 Commercial-industrial 0.60 0.58 0.43 0.56 Commercial-industrial (B) 0.64 0.56 0.41 0.68 Forest-cover -0.83 -0.69 -0.45 -0.68 Forest-cover (B) -0.84 -0.57 -0.39 -0.51 -0.75 > Grass-cover o o Grass-cover (B) 0.39 0.42 T3 Impervious-surface 0.85 0.76 0.40 0.66 CO _ l Impervious-surface (B) 0.73 0.70 0.37 0.63 Road-density 0.81 0.67 0.40 0.45 0.73 Road-density (B) 0.63 0.64 0.42 0.61 Spearman's rank correlations for historical change data Table 4-11 provides the result of the Spearman's rank correlations between land use changes and historical change in metal data. Most metals do not have significant correlations with land variables, except for Pb and Zn. Pb is correlated with "grass-cover", and Z n is with "impervious-surface". 79 Table 4-11 Spearman's rank correlation for historical change data. (B) denotes "a 100 m buffer zone". Only significant correlations are shown (p < 0.05), and bold presents that correlation is significant at the 0.01 level. C u Pb Zn C o Ni Fe Mn Forest Forest (B) Grass-agriculture <o Grass-agriculture (B) 0.54 3 Rural-urban 0.49 0.42 c Rural-urban (B) 0.43 0.44 —1 Dense-urban Dense-urban (B) -0.51 Commercial-industrial 0.53 Commercial-industrial (B) 0.47 Forest-cover -0.48 g> Forest-cover (B) 8 Grass-cover 0.57 Grass-cover (B) 0.59 J3 Impervious-surface 0.53 Impervious-surface (B) 0.60 0.51 4.3.4 Mann-Whitney U test The Mann-Whitney U significance test was used to examine the possible causes of metal concentrations in the sediments. Sub-watersheds were grouped according to the results of the cluster analysis (Table 4-12). The tests were applied to the sub-watersheds at the headwaters, so that each cluster should not contain spatially related sub-water sheds. The sub-watersheds in clusters 1,2,3 and 4 were identified as F W (forested sub-watershed), R W (rural sub-watershed), U W (urbanised sub-watersheds) and V D ( N / W Vancouver drainages), respectively. The median and range values of land variables for each cluster are shown in Table 4-13 and Figure 4-38. Only selected variables are presented, considering the factor analyses and correlations of both land variables and sediment quality. "Impervious-surface", "road-density" and "forest-cover" were selected because of their high correlations with metals. Material textures were also selected since they had significant correlations with metals. Cu , Pb and Z n were selected among soft acid metals because they are the most common metals in discussing sediment contamination. N i was selected among transitional metals since it had the highest correlations with land variables. M n 80 was chosen because its hydroxide forming capability. K was selected among other major cations as it is an important element in many minerals and is also used for fertilizer. P was selected because of its high correlations with most variables. The median and range for each cluster is shown in Table 4-14, and Figure 4-39. The variables that showed significant differences between the clusters differences are presented in Figure 4-40. The tests for the other elements are shown in Appendix M . Table 4-12 The clusters of the sub-watersheds for Mann-Whitney U tests (2006 data). E l e m e n t s S u b - w a t e r s h e d C l u s t e r l (FW) h1-h3, t1, k3 Cluster2 ( R W ) h4, k1-k2, k5, t2-t3, t6 C luster3 (UW) a1-a4 , k4 Cluster4 (VD) v3-v6 Table 4-13 Median values of each cluster. Parenthesis denotes land use in a buffer zone. Cluster Forest -cover % (Forest -cover) % Imper-v ious % (Imper-vious) % R o a d k m / k m 2 (Road) k m / k m 2 Minera -c o a r s e % Mineral -fine % F W 89 86 2 2 1.4 0.9 100 0 R W 59 59 9 9 2.7 3.7 53 43 U W 34 47 35 24 8.6 5.3 0 100 V D 18 18 55 55 30.6 30.6 100 0 81 100H 80H I o 60i a s p 4 0 i 20H The Whole Watershed Forest cover Buffer Zones Forest cover 100H 80 60 40 20H The Whole Watershed -aft! Buffer Zones • F W • R W • u w • V D 1 r Impervious Road density Impervious Road density iooH e 80 to 60 • 40H 20 I surface surface 0-1 u r ^ Coarse-mineral I Fine-mineral Figure 4-38 Bar charts and Mann-Whitney U tests for land variables among the clustered sub-watersheds (2006). The median and range for each cluster are presented. The unit of road density is km/km 2 Table 4-14 Median values of each cluster. Units are mg/kg dry weight. Cluster Cu Pb Zn Ni Mn K P FW 34 10 69 16 474 780 935 RW 30 10 112 33 909 1376 848 UW 47 15 243 48 1036 2175 988 VD 522 347 859 42 414 840 1377 Figure 4-39 Bar charts and Mann-Whitney U tests for metals in sediments among the clustered sub-watersheds (2006). The median and range for each cluster are presented. 82 (a) Land variables Cluster (smaller values) FW RW UW VD F _ W F M c F (F) Mc F(F) CU ra R is w CO 1 Mf F F (F) Mf Huster (la Huster (la 1 R (1) (R) Mf IR(I) FMf u V D 1 R (1) (R) 1 R (1) (R) Mc R (1) (R) Mc (b) Sediment quality Cluster (smaller values) FW RW UW VD F _ W 0)" 1 ra R i w CO Zn Ni Mn K Mn K luster (la Zn Ni Mn K Cu Zn Ni K O V D Cu Pb Zn Ni P Cu Pb Zn Ni P Cu P b Z n P Figure 4-40 Summary of Mann-Whitney U tests for 2006 data. Results of significant levels at 0.05 level are shown. 1= imperviousness; F=forest-cover; R=road-density; Mc=coarse mineral; Mf=fine mineral. Parenthesis denotes land cover in a buffer zone. F W / R W matrix: R W had greater imperviousness and less forest cover than F W (about A <10% and A -30% in median, respectively). The urbanisation led to significant increases in Z n and N i levels (A 50 and A 15 mg/kg, respectively). Only Z n showed a significant increase among soft acid metals (i.e., Cu, Pb, Zn). F W / U W matrix: in addition to above land use variables (imperviousness and forest cover), U W had greater road density than F W (A 30%, A -50%, A 7 km/km 2 , respectively), but likewise only Z n and N i were significantly different (A 170 and A 30 mg/kg, respectively). Only Z n showed a significant increase among soft acid metals (i.e., C u , Pb, Zn). 83 F W 7 V D matrix: there was no apparent mineralogical difference. V D had greater imperviousness, road density and less forest cover than F W (A 50%, A 30 km/km 2 and A -70%>, respectively), Cu , Pb, Zn, N i and phosphorus were significantly greater in V D (A 485, A 335, A 790, A 25 and A 440 mg/kg, respectively). However, there was no significant difference for K and M n . R W / U W matrix: the matrix was not influenced by mineral texture. U W had greater imperviousness, road density and less forest cover than R W (A 15%, A 6 km/km 2 and A -25%, respectively). A s a result, Cu , Z n and N i were significantly greater in U W than in R W (A 17, A 130 and A 15 mg/kg, respectively). R W / V D matrix: V D had greater imperviousness, road density and less forest cover (A 45%, A 28 km/km and A -40%, respectively). Consequently, C u , Pb, Zn , N i and phosphorus were significantly greater in V D (A 490, A 335, A 750, A <10 and A 530 mg/kg, respectively). U W / V D matrix: although imperviousness was not significantly different between the clusters, V D had greater imperviousness within a 100 m buffer zone, road density and less forest cover than U W (A 30%, A 25 km/km 2 and A -26%, respectively). A s a result, Cu , Pb, Z n and phosphorus were by far higher in V D (A 475, A 330, A 615 and A 390 mg/kg, respectively). In all matrices where "fine textured materials" areas were greater with smaller "coarse textured materials" areas, both M n and K had greater levels. 4.3.5 Summary of the section 3 The rate of increase in metals between 1973-2006 was generally higher in rural sub-watersheds (i.e., with imperviousness < 20%) than in those more heavily urbanised sub-watersheds. This suggests that these relationships are likely curvilinear with higher increases in the early stages of urbanisation. When comparing the sub-watersheds with high level of urbanisation (2006 data for N / W Vancouver and Alouette), it was found that metal concentrations of soft acid metals (Cu, Pb, Zn , Cd), phosphorus and organic matter were higher in N / W Vancouver than those in Alouette, while concentrations of M n , N i and K were greater in the Alouette sub-watersheds than in N / W 84 Vancouver. Cu , Pb and Z n levels in N / W Vancouver were above P E L , whereas those in Alouette were above/below P E L . This is likely attributed to differences in traffic density. Road density was used as a surrogate indicator because no traffic density data were available to do the comparison. In both 1973 and 2006, Cu , Pb, Zn , and N i have high positive correlations with urban variables such as "impervious-surface", and negative correlations with forest variables and topographical variables. M g and K had strong correlations with the texture of the surface material. Historic change data only showed a few correlations. Mann-Whitney U tests were applied to the clustered sub-watersheds (FW, R W , U W and V D ) . Increases in imperviousness and decreases in forest cover generally led to increases in trace metals. Z n and N i were the most sensitive to these land use/ cover changes; A < 1 0 % in imperviousness increased Z n and N i , followed by C u with A 15%. Metal levels in V D were always greater than the other clusters. Imperviousness in V D was not significantly different from that in U W , but imperviousness within a 100 m buffer zone, road density and forest cover differed. In all matrices where "fine textured mineral" areas were greater, and "coarse textured mineral" areas were smaller, both M n and K had higher levels. 85 5 Discussion 5.1 Land use patterns and changes Imperviousness values found in 2006 in the Alouette sub-watersheds (al-a3) (50-73%), and in the N / W Vancouver sub-watersheds (about 38-61%) are typical for multi-family residential density (Dinicola 1990; Booth and Jackson 1997; Brabec et al. 2002). These two areas showed the highest rate of urbanisation while the remaining sub-watersheds in 2006 showed relatively small increases in imperviousness. Although urbanisation has progressed in most sub-watersheds between 1973-2006, the greatest increases in imperviousness occurred in the Alouette sub-watersheds (A 10-30%). This increase is greater than the changes that occurred in the Brunette Watershed between 1973-1993 (A 5-10%) (McCal lum 1995) (Figure 4-29, and Table 2-9 and 2-10). The changes in the remaining sub-watersheds since 1973 were relatively small (A 1-10%). Most sub-watersheds experienced a loss of forest cover, but the changes were highly variable. The greatest changes in forest cover occurred in the Thornhill sub-watersheds (A 20-25% in t3 and t4), while the other sub-watersheds were in the A 10-20% range. Percent of forest-cover and imperviousness within a 100 m riparian buffer zone were found to be the best land use indicators. The Alouette sub-watersheds saw a 12-25% increase in imperviousness between 1973-2006, while the changes in the remaining sub-watersheds ranged between 1-9%. 5.2 Sediment metal levels and changes Urbanisation is strongly related to increase in occurrence of soft acid metals (e.g., Cu , Pb, Zn), and transitional metals such as N i . C u is the most common metal widely used in all aspects of urban activities (e.g., electrical, plumbing and automotive industries) (Moore and Ramamoorthy 1984). Pb has long been used for piping, solders and batteries. Pb was added to gasoline until 86 1975 when a program of removal was implemented. The wear of tires is a significant source of Zn, and brake linings and salting also contribute to high Z n levels to impervious surfaces (Christensen and Guinn 1979; Oberts 1986; Thomson et al. 1997). N i is used for corrosion resistance in cutlery and stainless steel, alloyed with Cr (Moore and Ramamoorthy 1984). N i and Cr are believed to be less toxic compared to soft acid metals. Hence Z n and C u are likely to be the metals that are the most concern of traffic intensity. Some studies use factor analysis to identify the similarity in sediment metals. Riba et al. (2002) categolises 1) C u and Fe, 2) Pb, Z n and Cd , 3) Fe and M n , Pekey et al. (2005) categolises 1) M n , Cr, Cu , N i and Fe, 2) C d Co, Pb and Zn, 3) M g and A l , and Enguix Gonzalez et al. (2000) does Fe and M n according to factor loadings. Factor analysis of this study was able to achieve superior differentiation of metal characteristics. The analyses were able to separate soft acid metals (Cu, Pb, Z n and Cd), from metals forming oxides (Fe, M n and A l ) , and basic cations (Mg and K ) . The patterns of the metal levels can partially be explained according to the differences in metal characteristics. Chemical bonding of the soft acid metals with ligands is very strong, and thus these metals can be very conservative and insoluble in the aquatic environment. Oxides and basic cations are closely associated with minerals and their weathering products. The behaviour of transitional metals varies widely due to bivalent association with both soft and hard donor atoms (Appendix A ) (Buffle and de Vitre 1993; Sparks 1995). The N / W Vancouver sediments in 2006 had the highest levels of soft acid metals (Cu, Pb and Zn) (about 300-540, 130-490 and 710-1610 mg/kg, respectively), whereas those in Alouette ( a l -a3) were considerably lower (55-70, 20-50 and 240-370 mg/kg, respectively). On the other hand, N i levels were similar in both places (about 35-60 mg/kg). Soft acid metal levels in N / W Vancouver are considerably higher than those in highway sites in the Puget Sound Lowland (PSL) in Washington State (60-80, 30-75, 200-500 mg/kg, respectively), where climate condition is similar (Brandenberger et al. 2003). They are similar to the Alouette levels, as well as other studies (30-90, 35-75, 155-300 mg/kg, respectively) (Heiny and Tate 1997; Enguix Gonzalez et al. 2000; Sutherland 2000; Shea 2003) (Georgia, Hawaii , Spain and Colorado, respectively). Pb and Z n levels in N / W Vancouver are similar to those in highly urbanised area in Australia (Birch et al. 2000). Metal levels in the Alouette sediments were similar to those in the Brunette Watersheds in the L F V (McCal lum 1995) (Table 2-11), and levels in N / W Vancouver, were consistent with sediments measured from stormwater detention systems in Greater Vancouver (Brydon 2004) (Figure 2-3). These indicate that sediments in Alouette and Brunette had typical 87 metal levels in urbanised watersheds, and those in N / W Vancouver presented contamination at an extraordinary level. Historical increases in the metals varied widely (Figure 4-34). The greatest changes of C u occurred in k3, t2 and tl (A 20-30 mg/kg), with much lower values in the Alouette sub-watersheds (about A 10 mg/kg). The changes in N i levels were variable with rural sub-watersheds (mainly in Thornhill) showing greatest changes (A 20-30 mg/kg), while the Alouette sub-watersheds increases were significantly smaller (A 4-11 mg/kg). It is surprising that the greatest increases did not occur in the urbanised sub-watersheds, indicating the complexity of sediment-metal interactions. Changes in Z n levels were the greatest in Alouette (al-a4) and Hatzic (h4) (A 80-175 mg/kg). Sharp decrease of Pb in a l , a2 and t6 (A -490, -60, -130 mg/kg, respectively) is ascribed to the replacement of lead additive in gasoline by organic derivative of manganese ( M M T ) through 1980s in Canada (Loranger and Zayed 1994; Pott and Turpin 1996). Accordingly, M n increased in most sub-watersheds and some (a2, a3 and h4) exceeded the P E L in the United Stats (Figure 4-33 and 4-36). 5.3 Relationships between land use and metals 5.3.1 Imperviousness and zinc as indicators of urbanisation Correlation results clearly confirm the strong association of Cu , Pb, Zn , Cr and N i with forest loss and increased urban activities (Table 4-7 and 4-8). The strongest correlations were found between imperviousness and a number of metals. Impervious surfaces provide a surface that facilitates transport to the aquatic environment. Roofs of houses and buildings can be a source of metals i f the runoff flows through galvanized pipes. The greatest accumulation of contaminants occurs on traffic roads which also generate hydrocarbon residuals. Surface runoff flushes the contaminants from roads into the conveyance system and into streams (Arnold and Gibbons 1996; Booth et al. 2002). The correlations between Z n with urban land variables were the highest followed by Cr and N i . This suggests that these are the best sediment quality indicators for measuring the aquatic impacts of urbanisation. Both imperviousness and Z n were strongly 88 associated with road density, (a surrogate indicator of traffic), and automobile traffic is likely generating most of the toxic metals which get deposited on roads and are flushed into urban streams. Strong negative correlations of these toxic metals with topographical variables are understandable, since urbanisation is limited in mountainous areas. M g and K had the strongest correlations with the texture of the surface material, inferring that mineral weathering leads to release of these elements. The sub-watersheds underlain by finer textured materials are dominated by weathering of primary minerals into secondary minerals, and in the process that basic cations such as K + and M g 2 + are released and become associated with particulates by cation exchange (Buffle and de Vitre 1993; Sparks 1995). Strong correlations of phosphorus with urban land use/ cover (positive) and forest variables (negative) are explained as follows; urban land uses introduce additional anthropogenic nutrients through lawn fertilizers, pet waste and septic tank effluent etc. to water bodies (US E P A 1990; Carpenter et al. 1998; Tufford et al. 1998; Brett et al. 2005). In addition, a loss of forest cover minimizes retention, uptake and recycling of nutrients generated (Wahl et al. 1997; Abelho 2001; Brett et al. 2005). 5.3.2 Trace metals from atmospheric sources Trace metals from atmospheric sources can be also a significant contribution to sub-watersheds (Section 2.1.2, Trace metals from atmospheric sources), yet this was not investigated in this study as it is considerably difficult to differentiate how much of the metals in stream sediments originated from atmospheric sources inside/outside of a sub-watershed. However, impervious surfaces especially roads are the depositional places from atmosphere, and precipitation flushes off contaminants on the surfaces. Hence impervious surfaces still work as a surrogate indicator of contaminants from both outside and within a watershed. 5.3.3 Effectiveness of a buffer zone Correlations of land use/ cover with sediment quality within a 100 m buffer zone were not necessarily higher than those in the whole sub-watershed, but both correlations were similar (Table 4-7 and 4-8). This is reasonable considering that the factor loadings with the land use variables gave the same results (Table 4-2 and 4-3). Land use/ cover within a buffer zone generally has a greater influence on stream quality than land use in the whole watershed (Naiman 89 et al. 1993; Wenger 1999; Berka et al. 2001; Houston 2004). This study suggests that these indicators are useful regardless i f they are measured within a 100 m buffer zone or within the whole watershed. However, the efforts to determine these variables in a quantitative manner are much less in the buffer zone than classifying the whole watershed for land use activities. 5.3.4 Multivariate relationships between land use and metals The summary of the Mann-Whitney U test presents multivariate association between metals and land variables (Section 4.3.4), confirming the correlation results. Generally, imperviousness increase and forest loss led to the increases in trace metals, and Z n and N i were the most sensitive to the land use changes. A 10% increase in imperviousness significantly increased Z n and N i , followed by C u with A 15% of imperviousness. The associations of Z n with traffic and N i with residential activities may account for the sensitiveness of these metals. It should be noted that A 30% of imperviousness and A -50% of forest did not lead to significant increases in C u and Pb in ( F W / U W matrix), but because of the small number of samples, the statistical analysis was limited. The evaluation of land use within the buffer zone proved to be more valuable. Although imperviousness in V D did not differ from that in U W , imperviousness in a 100 m buffer zone, road density and forest cover differed. Consequently, imperviousness is indeed a partial indicator of urbanisation and its impacts on aquatic health. It is an indirect indicator that also should include traffic activities, especially within a buffer zone. The metals that are of the greatest concern are the soft acid metals (Cu, Pb and Zn). Fine textured minerals are dominated by secondary clay minerals, which underwent mineral weathering processes from primary minerals. During the processes minerals lose major cations such as K and M g from their crystal structures (Ure and Berrow 1982). This could result in greater release of K in the sub-watersheds underlain by finer minerals. Sediments composed of weathered minerals also contain higher levels of oxides consisting of Fe and M n (Ure and Berrow 1982; Sparks 1995), which should account for the greater M n levels in the clusters having greater areas of fine minerals. Notably, although M n increase stood out in most sub-watersheds attributed to M n in M M T , urbanised areas (VD) had lower M n levels than in more rural areas (RW). M n levels are associated with not only an industrial factor but a mineralogical effect. 90 5.3.5 Curvilinear relationships between imperviousness and metal increases Based on the changes in metals between 1973 and 2006 it is evident that they form a curvilinear relationships with imperviousness (e.g., Cu , N i ) . The change of slope is much higher in the early stages of urbanisation, whereas in areas of high imperviousness the increases in metals were less evident. This suggest that the impacts of urbanisation on metals in sediments start very early in the process of urbanisation. It is readily possible that the adsorption capacity of sediments decreases as the sediments get saturated (Di Toro et al. 1986). It can be speculated that the Alouette sediments were saturated by abundant metals, and could no longer adsorb additional elements. A t that point the metals could exist in a dissolved phase or be associated with the suspended portion of the sediments, but this assumption needs to be evaluated more closely in order to clarify sediment-water interactions. This is challenging due to the difference in time scales between sediments and water, and the problem of collecting suspended sediments during storm events and analytical problems associated with measuring low metal concentrations in water (Minton 2005). The reasons why measurements of metals in suspended sediments and in water were not pursued in this study were two fold. First, simple indicators are needed i f we hope to arrive at a practical monitoring program for measuring urban impact and secondly it can be argued that new sediments are constantly deposited in the stream and as a result bed sediment analysis in late summer provide a good indicator of stream health (Walling and Amos 1999; McBride 2001; McKee et al. 2003). However, the sub-watersheds with flat topography are expected to provide fewer sediments than those with steep slopes. The Alouette sub-watersheds are relatively flat and new sediment supply can be small, resulting in the earlier saturation of the sediments. The adsorption capacity of sediments is highly dependent on sorptive media in sediments, which vary widely: (hydroxide surfaces, ion exchange within clay minerals, organic matter or a metal-ligand complex ) (Tessier et al. 1979; Landner 1987; Fergusson 1990). A significant fraction of metals is often associated with organic matter due to its strong adsorption capacity (Davis and Leckie 1978); cation exchange capacity of organic matter is usually higher than that of clay minerals (Sparks 1995). Organic matter content was not high in the Alouette sediments (Figure 4-33, f and 4-37), while abundant Fe /Mn oxides in the Alouette streams were not possibly sufficient to adsorb all increased metals. 91 Z n was the only metal that showed a consistent linear relationship with imperviousness, with no conspicuous decrease in the slope of a historical change relationship (Figure 4-36). It is speculated that adsorption capacity for Z n in the Alouette sediments was relatively high. 5.3.6 Imperviousness as the best indicator? - Alouette vs. N/W Vancouver The comparison between Alouette and N / W Vancouver adds complexity to metal-sediment interactions (Figure 4-37). Metal levels in N / W Vancouver were by far greater than P E L , while those in Alouette did not exceed the P E L as much. The question arises why a similar extent of urbanisation can produce significantly different metal levels for soft acid metals. Several possible factors can account for this phenomenon. First, traffic was different between the two areas. Second, effective imperviousness could also be different between the two. Third, phosphorus could affect the solubility of metal compounds. Fourth, an association with organic matters could influence the transport of metal compounds. Road-density and traffic The N / W sub-watersheds had higher road-density than the Alouette ones (Table 4-11 and Figure 4-38). Z n and C u generation from the wear and corrosion of car tires and bodies is associated with road surfaces. A number of studies found high concentrations of soft acid metals in road dusts, a car tire (Revitt et al. 1990) and storm sewer sediments (Charlesworth and Lees 1999). Metal levels are found to be elevated in gully pots indicating that the materials from the road surfaces accumulate in stormwater system (Sutherland 2000; Birch and Scollen 2003). Thus high concentrations of Z n and C u in sediments are most likely to reflect their sources of vehicle based activity, and its impacts seem to be greater than those of residential land uses. It is suggested that greater traffic density was the partial reason for elevated trace metal levels in N / W Vancouver. Effective imperviousness and a buffer zone Effective imperviousness, namely impervious surfaces directly connected to the surface drainage systems, truly have impacts on streams. If runoff from impervious surfaces is infiltrated into soils then the effect on streams is minimal (Alley and Veenhuis 1983; Beyerlein 1996; Booth and Jackson 1997; Brabec et al. 2002). The ratio of E I A (effective impervious area) vs. T I A (total 92 impervious area) varies widely (Dinicola 1990; Booth and Jackson 1997), and therefore T I A does not necessarily represent the actual extent of urbanisation impacts on a watershed. A systematically planned urban drainage in which stormwater is directed to soils and vegetation can potentially minimize the E I A and transport of contaminants to streams. Riparian vegetation within a buffer zone can decrease the E I A , which is crucial in protecting stream quality from stormwater runoff (Cooper et al. 1987; Naiman et al. 1993; Osborne and Kovacic 1993; Lee et al. 2000). Riparian buffers can be responsible for removing trace metals, by retention, adsorption and infiltration by vegetation and soils (Dillaha and Inamdar 1996; Lowrance 1998; Lee et al. 2000). This is effective only when stormwater conveyances do not short circuit hydrologic pathways by bypassing riparian buffers (Paul and Meyer 2001; Roy et al. 2006). The forest cover in the Alouette buffer zone was between 40-60% whereas that in N / W Vancouver was between 0-25% (more storm sewers) (Figure 4-38). The imperviousness in the buffer zone in the Alouette sub-watershed was smaller (median ~ 20%) than in its whole sub-watershed (median ~ 35%), whereas that in N / W Vancouver was much higher (median ~ 55%). In the Alouette buffer zone some 20-30 m of vegetation was maintained directly adjacent to the stream (Figure 4-17) and this would allow more filtration of stormwater runoff, whereas storm sewers dominated the drainage in N / W Vancouver. A narrower vegetative buffer perhaps led to the greater fresh sediment entry into the stream in N / W Vancouver than in the Alouette system. Phosphates Phosphorus is widely used for residential lawns as a component fertilizer (McDonald 2000; Bennett et al. 2005), and vast residential areas in N / W Vancouver are considered to produce high phosphorus loads. The sorption of metals from the liquid phase onto the solid phase is the most important chemical processes affecting the behavior of metals in soils (Forstner 1979; Evans 1989; Gong and Donahoe 1997). Cations such as C a can form insoluble phosphates when combining with phosphorus. Solubility products (pK s p ) of the phosphates differ according to each cation (Appendix A ) . The p K s p value is greater for soft acid metals than other metals. Phosphorus was likely to form insoluble phosphates with abundant existing soft acid metals, that precipitated in the streambed (Hubbard and Lowrance 1994; Naiman and Decamps 1997) in 93 N / W Vancouver, while both phosphorus and trace metal levels were lower in the Alouette sediments (Figure 4-39). This hypothesis can be challenged that clay minerals and Fe /Mn oxides may provide stronger adsorption capacity for C u , Pb and Z n (Abd-Elfattah and Wada 1981). Close relationships have been found between amounts of adsorbed phosphate and soil contents of A l and Fe, which is also dependent on p H (Borggaard et al. 2004). Fe, M n and A l as well as K and M g levels in Alouette were relatively higher than those in N / W Vancouver, although phosphorus was lower in Aloette. This adds the complexity of element-sediment interactions on solid materials. Speciation study is needed to evidence in which fractions toxic elements existed. Organic matter Organic matter content in N / W Vancouver was comparatively higher than that in Alouette (median), although the values were not statistically significant (Figure 4-37). Organic matter is believed to be one of the main controls of heavy metal sorption (Gong and Donahoe 1997), forming compounds with metal ions (chelation) and therefore having significant impacts on retention and mobility of trace metals (Appendix A ) . The stability and solubility of organic matter are highly variable according to its size, salient p H and type (i.e., non humic substances, humin, humic acids and fulvic acids). Chelated E D T A (common detergent for residential use) is very stable in an aquatic environment (Moore and Ramamoorthy 1984; Petrovic et al. 1999). Some solution metals can be preferentially bound to insoluble organic matter (Bolt and Van Riemsdijk 1987). C u sorption capacity of soil components is generally in the following order: organic matter > Fe/ M n oxides > > > clay minerals (Alloway 1995; Sparks 1995; Gong and Donahoe 1997). C u 2 + , P b 2 + and Z n 2 + can form stronger complexes with humic acids than other metals (Stevenson 1976), becoming stable and insoluble in streams. The N / W Vancouver sediments had the high concentrations of organic matter and trace metals, indicating the stable/ insoluble form of metal organic compounds was likely to be one of the controls in N / W Vancouver. 94 Imperviousness alone is not a cause These suggest that imperviousness itself is not an actual contaminant source but is a non-adsorbent depositional place which facilitates the transport of the contaminants. Complex interactions between other factors all influenced the high metal levels in N / W Vancouver, and traffic was likely the principal cause. 5.3.7 Cause and effects of stream degradation - biotic integrity and metals Most ecological indices such as an index of biotic integrity (IBI) represent the cumulative impacts of urbanisation on a watershed scale, integrating the effects of a wide range of contaminants. Benthic invertebrates are likely to be more sensitive than fishes, due to their sedentary l iving environment (Karr 1981; Schueler 1994). Aquatic health degradation as expressed by species diversity appears to be exponential with a threshold of 10-15% (Figure 2-2) (May et al. 1998; Zandbergen et al. 2000; Booth et al. 2004). This clearly indicates the existence of a threshold of stream degradation at the early stage of urbanisation. The increase in metals in sediments can be explained as a possible cause of rapid stream degradation that has an impact on the aquatic biota. The metal-imperviousness relationships showed similar curvilinear responses but is in the opposite direction of the IBI-impervious relationship. This is particularly evident for C u and N i (Figure 4-36). The rising level of metals during urbanisation is likely the critical period affecting aquatic biota (Figure 5-1). A curvilinear relation was not identified for Zn , and it is speculated that the adsorption and saturation capacity for Z n was much higher. 95 I 1 1 1 0 10 20 30 40 50 % imperviousness Figure 5-1 A schematic diagram of the relationships between imperviousness and biotic integrity/ metal levels. It should be noted that contamination of the Alouette sediments was not at an excessive level when examining individual metals (Figure 4-33). Adverse effects appear when a metal level is higher than P E L according to the guidelines (Environment Canada 2004), but Cu , Pb and Cr levels in Alouette were below P E L levels, and Z n levels were just below and above P E L in 2006 (Figure 4-33). Less toxic N i levels exceeded P E L . Nonetheless, most urban watersheds of existing studies present stream degradation at an early stage of urbanisation (Karr 1981; Schueler 1994; M a y et al. 1998; Morley and Karr 2002). Sediment quality guidelines are determined by laboratory procedures by using the mortality o f selected aquatic organisms against individual contaminants. Hence the impacts of combinations of contaminants on aquatic species cannot be determined easily. Cumulative effects of the contaminants can be measured only in an additive but not in a synergistic or antagonistic manner. For instance, C u is more toxic i f Z n is present. The presence of C a and M g in water reduces the toxicity of metals to fish (Schreier et al. 1997; Zandbergen et al. 2000). Alouette sediments can generate these combined cumulative impacts of metals, although individual metals did not greatly exceed the critical levels. 96 5.3.8 Bioavailability of sediment metals Total metal concentrations do not necessarily reveal the immediate adverse effects on aquatic ecosystems. First, it is challenging to identify how much of the total metals are bound to sediments in bioavailable form. Speciation of trace metals in sediments has long been examined by using sequential extraction methods, and these techniques include up to five extractants: cation exchange, Carbonate-dissolving, acidic reducing, extractants that release organic- and sulfide-bound metals, and strong acidic for dissolving silicates or minerals (residual). Those methods suffer from nonselectivity of the extractants and trace element redistribution among phases during extraction (Shan and Chen 1993), and thus their status as useful analytical tools is controversial (Kheboian and Bauer 1987). In addition, bioavailability itself differs depending on the ambient environment and each aquatic species. Fe/ M n oxides and organic matter contents are the main controls of heavy metal sorption to sediments (Davis 1984; Tessier et al. 1985; Young and Harvey 1992; Gong and Donahoe 1997), being less bioavailable to aquatic organisms (Tessier et al. 1979; Landner 1987; Fergusson 1990; Young and Harvey 1991). However, metals may be released in the aquatic environment due to biodegradation of organic matter. Graphically the N / W Vancouver sediments were found to contain high an organic matter content (Figure 4-33), whereas the Alouette sediments contain more Fe and M n , than N / W Vancouver. Further investigations is needed to decide which areas contained more bioavailable metals. 5.3.9 Analytical accuracy of the laboratory procedures in the past Comparing historic sediment analysis conducted in 1973 (Westwater Research Centre 1973) with those from 2006 is always problematic because instrumentation and analysis techniques and accuracy change. This might affect the variability of historical change data in this study, particularly in a few cases showing decreases in some metals (e.g., Co, N i , Mn) in undisturbed sub-watersheds covered by forests (control sites) (Figure 4-36). On the other hand, Z n in those sub-watersheds showed little historical difference, and C u showed considerable increases with little increase in imperviousness. These results imply that there is no consistent variation across the different metals in those forested sub-watersheds. This could be attributed to inaccuracy of the analytical procedure. 97 In this study only large differences between 1973 and 2006 were considered to be significant. However, the fact that most of the sub-watersheds showed a significant increase with increased imperviousness provided an indication that the quality of the sediments has been affected. 98 6 Conclusions The impact of urbanisation on metal concentrations in sediments was examined in 28 sub-watersheds in the Lower Fraser Valley. GIS techniques were used to determine change in urbanisation betweenl973 and 2006 and the metal concentrations in sediments in 2006 were then compared with historic data collected at the same sites. In 2006 the highest imperviousness was observed in the Alouette and N / W Vancouver sub-watersheds (50-70% and 40-60%>). A l l sub-watersheds experienced increases in urbanisation but the greatest increases in imperviousness between 1973 and 2006 occurred in the Alouette sub-watersheds (A 10-30%). Percent of forest-cover and imperviousness within a 100 m riparian buffer zone were found to be the best land use indicators. Urbanisation is strongly related to increases in trace metals (e.g., Cu , Pb, Zn , N i ) . Z n and C u are likely to be the metals that are most indicative of traffic intensity. The N / W Vancouver sediments in 2006 showed the highest levels of soft acid metals, whereas those in Alouette (al-a3) were considerably lower. It is surprising that the greatest increases in C u and N i did not occur in the urbanised Alouette sub-watersheds, indicating the complexity of sediment-metal interactions. Automobile traffic is likely generating most of the toxic metals, especially Zn, which get deposited on roads and are flushed by rainfall runoff into urban streams. Mineral weathering and cation exchange processes led to release of basic cations such as K and M g . Anthropogenic nutrient inputs from residential areas are the major source of phosphorus increase in streams. Land use indicators are recommended to be measured within a 100 m buffer zone considering the efforts to determine these variables in a quantitative manner. Generally, increases in imperviousness and losses of forest cover led to the increases in trace metals, and Z n and N i were the most sensitive metals related to these changes. Imperviousness is an indirect indicator of urbanisation and i f combined with traffic density is a highly reliable indicator of aquatic health deterioration in urban streams. Several studies have shown that the % imperviousness forms a curvilinear relationship with aquatic biodiversity indices, and a threshold of 10-15 % of imperviousness is considered to be critical level above which benthic invertebrate diversity is seriously degraded. Imperviousness is not the direct cause of habitat deterioration but is considered the conveyance system that brings 99 contaminants directly into the stream. Urban land use activities and traffic generate the contaminants that are the cause of habitat impairment. In this study it was shown that an inverse curvilinear relationship exists between impervious surfaces and several trace metals. The rate of increase in metals in sediments was clearly more rapid in the early stages of urbanisation and tends to flatten out once high level of imperviousness was achieved. This could be related to sediment saturation. The Alouette and N / W Vancouver sediments were possibly saturated by abundant metals, and could no longer adsorb additional elements. However, further investigations in water and suspended sediments are necessary to clarify the complicated sediment-water interactions. Organic matter, which has strong adsorption capacity, was not high in the Alouette sediments, while abundant Fe/ M n oxides in the Alouette streams were not possibly sufficient to adsorb all increased metals. Nonetheless, we hope to arrive at a practical indicators for measuring urban impacts, and bed sediment analysis in late summer provide a good indicator of stream health as new sediments are constantly deposited in the stream. Metal levels in N / W Vancouver were by far greater than P E L , while those in Alouette did not exceed the P E L as much, although they had similar imperviousness. Several possible factors can account for this phenomenon. First, high concentrations of Z n and C u in sediments are most likely linked to transportation related activities. Second, vegetation in the Alouette buffer zone was maintained directly adjacent to the stream and this would allow more filtration of stormwater runoff, whereas storm sewers and a narrower vegetative buffer in N / W Vancouver led to the greater entry of fresh sediment directly into the stream, which enhanced the metal levels. Third, phosphorus was likely to form insoluble phosphates with abundant existing soft acid metals, that precipitated in the streambed in N / W Vancouver. Fourth, organic matter contents in N / W Vancouver were comparatively higher than that in Alouette, and they form chelated compounds with metal ions which have significant impacts on retention and mobility of trace metals. Complex interactions between these and other factors all influenced the high metal levels in N / W Vancouver, and traffic was likely the principal cause. 100 Bibliography Abd-Elfattah, A . and Wada, K . 1981. "Adsorption of Lead, Copper, Zinc, Cobalt, and Cadmium by Soils that Differ in Cation-Exchange Materials" Journal of Soil Science, 32 (2): 271-283. Abelho, M . 2001. "From Litterfall to Breakdown in Streams: A Review" The Scientific World Journal, 1: 656-680. Alexander, D . and Tomalty, R. 2002. "Smart Growth and Sustainable Development: Challenges, Solutions and Policy Directions" Local Environment, 7 (4): 397-409. Al ley , W . M . and Veenhuis, J.E. 1983. "Effective Impervious Area in Urban Runoff Modeling" Journal of Hydraulic Engineering, 109 (2): 313-319. Alloway, B.J . , 1995, Heavy Metals in Soils, Springer, 384 p. Anderson, G . L . , Everitt, J .H. , Escobar, D .E . , Spencer, N . R . , Andrascik, R .J . 1996. "Mapping Leafy Spurge (Euphorbia Esula) Infestations using Aerial Photography and Geographic Information Systems" Geocarto International, 11 (1): 81-9. Arnold Jr, C L . and Gibbons, C.J . 1996. "Impervious Surface Coverage: The Emergence of a Key Environmental Indicator." Journal of the American Planning Association, 62 (2): 243-258. Barton, D.R. , Taylor, W . D . , Biette, R . M . 1985. "Dimensions of Riparian Buffer Strips Required to Maintain Trout Habitat in Southern Ontario Streams" North American Journal of Fisheries Management, 5 (3a): 364-378. B C S T A T S . 2007. "Greater Vancouver Regional District." Community Facts. http://www.bcstats.gov.bc.ca/data/dd/facsheet/CF170.pdf Belzer, W.P. , Evans, C , Poon, A . , Snyder, G . 1997. "Wet Concentrations and Deposition Rates of Organic Compounds and Inorganic Substances in Ambient A i r Surrounding Burnaby Lake, B C for 1995." Environment Canada.Vancouver, B C . D O E F R A P . Bennett, E . M . , Carpenter, S.R., Clayton, M . K . 2005. "Soil Phosphorus Variability: Scale-Dependence in an Urbanizing Agricultural Landscape" Landscape Ecology, 20 (4): 389-400. Berka, C , Schreier, H . , Hal l , K . 2001. "Linking Water Quality with Agricultural Intensification in a Rural Watershed" Water, A i r , & Soil Pollution, 127 (1): 389-401. http://www.springerlink.com/content/xw7715251025w737/ Beyerlein, D . 1996. "Effective Impervious Area: The Real Enemy." Proceedings of the Impervious Surface Reduction Conference, City of Olympia, Washington. Birch, G.F. and Scollen, A . 2003. "Heavy Metals in Road Dust, Gul ly Pots and Parkland Soils in a Highly Urbanized Sub-Catchment of Port Jackson, Australia" Australian Journal of Soil Research, 41 (7): 1329-1342. Birch, G.F. , Evenden, D . , Teutsch, M . E . 1996. "Dominance of Point Source in Heavy Metal Distributions in Sediments of A Major Sydney Estuary (Australia)" Environmental Geology, 28 (4): 169-174. Birch, G.F. , Robertson, E . , Taylor, S.E., McConchie, D . M . 2000. "The use of Sediments to Detect Human Impact on the Fluvial System" Environmental Geology, 39 (9): 1015-1028. Bolt, G . H . and V a n Riemsdijk, W . H . 1987. "Surface Chemical Processes in Soil ." In Aquatic Surface Chemistry: Chemical Processes at the Particle-Water Interface. John Wiley and Sons, N e w York, pp 127-164,. Booth, D . B . and Jackson, C R . 1997. "Urbanization of Aquatic Systems: Degradation Thresholds, Stormwater Detection, and the Limits of Mitigation" Journal of the American Water Resources Association, 33 (5): 1077-1090. Booth, D . B . , Hartley, D . , Jackson, R. 2002. "Forest Cover, Impervious-Surface Area, and the Mitigation of Stormwater Impacts" Journal of the American Water Resources Association, 38 (3): 835-845. 101 Booth, D . B . , Karr, J.R., Schauman, S., Konrad, C P . , Morley, S.A., Larson, M . G . , Burger, S.J. 2004. "Reviving Urban Streams: Land use, Hydrology, Biology, and Human Behavior" Journal of the American Water Resources Association, 40 (5): 1351-1364. Borggaard, O .K. , Szilas, C , Gimsing, A . L . , Rasmussen, L . H . 2004. "Estimation of Soil Phosphate Adsorption Capacity by Means of a Pedotransfer Function" Geoderma, 118: 55-61. Boyle, C . A . , Lavkulich, L . , Schreier, H . , Kiss , E . 1997. "Changes in Land Cover and Subsequent Effects on Lower Fraser BasinEcosystems from 1827 to 1990" Environmental Management, 21 (2): 185-196. Brabec, E . , Schulte, S., Richards, P . L . 2002. "Impervious Surfaces and Water Quality: A Review of Current Literature and its Implications for Watershed Planning" Journal of Planning Literature, 16 (4): 499-514. Brandenberger, J . M . , May, C .W. , Kohn, N . P . , Bingler, L .S . , Sequim, W . 2003. "Potential Stormwater Impacts on Sediment Quality in Urbanizing Clal lam County Streams." Battelle Pacific Northwest Division. Prepared for Clallam County, Department of Community Development, Environment Quality Division. http://www.clallam.net/streamkeepers/assets/applets/Clallam_Stormwater_Sediment_2003 .p df Brett, M . T . , Arhonditsis, G .B . , Mueller, S.E., Hartley, D . M . , Frodge, J.D., Funke, D . E . 2005. "Non-Point-Source Impacts on Stream Nutrient Concentrations Along a Forest to Urban Gradient" Environmental Management, 35 (3): 330-342. Brewer, R. and Belzer, W . 2001. "Assessment of Metal Concentrations in Atmospheric Particles from Burnaby Lake, British Columbia, Canada" Atmospheric Environment, 35 (30): 5223-5233. Brydon, J. 2004. "The Effectiveness of Stormwater Ponds in Contaminant Removal from Urban Stormwater Runoff in the Lower Fraser Valley, B . C . " A thesis submitted in partial fulfilment of the requirements for the Degree of Master of Science in the Faculty of Graduate Studies, Resource Management and Environmental Studies, the University of British Columbia. Buffle, J. 1984. "Natural Organic Matter and Metal-Organic Interactions in Aquatic Systems" Metal Ions in Biological Systems, 18 Buffle, J., and de Vitre, R., 1993, Chemical and Biological Regulation of Aquatic Systems, C R C Press Buffleben, M . S . , Zayeed, K . , Kimbrough, D . , Stenstrom, M . K . , Suffet, I .H. 2002. "Evaluation of Urban Non-Point Source Runoff of Hazardous Metals Entering Santa Monica Bay, California" Water Science & Technology, 45 (9): 263-268. Cabe, R. and Herriges, J .A. 1992. "The Regulation of Non-Point-Source Pollution Under Imperfect and Asymmetric Information" Journal of Environmental Economics and Management, 22 (2): 134-146. Canadian Council of Ministers of the Environment. 2002. "Canadian Sediment Quality Guidelines for the Protection of Aquatic Life." http ://www.ccme. ca/ assets/pdf/sedqgsummarytable .pdf Carpenter, S.R., Caraco, N . F . , Correll, D . L . , Howarth, R .W. , Sharpley, A . N . , Smith, V . H . 1998. "Nonpoint Pollution of Surface Waters with Phosphorus and Nitrogen" Ecological Applications, 8 (3): 559-568. Carpi, A . 2007. "Chemical Bonding" accessed in Jan, 2007. http://www. visionlearning.com/library/module_viewer.php ?mid=55&l=&c3= Castelle, A . J . and Johnson, A . W . 2000. "Riparian Vegetation Effectiveness." National Council for A i r and Stream Improvement Inc. 102 Charlesworth, S . M . and Lees, J .A. 1999. "The Distribution of Heavy Metals in Deposited Urban Dusts and Sediments, Coventry, England" Environmental Geochemistry and Health, 21 (2): 97-115. Christensen, E.R. and Guinn, V . P . 1979. "Zinc from Automobile Tires in Urban R u n o f f Journal of the Environmental Engineering Division, 105 (1) Christensen, E.R. , Scherfig, J., Koide, M . 1978. "Metals from Urban Runoff in Dated Sediments of a very Shallow Estuary" Environmental Science & Technology, 12 (10): 1168-1173. Civco, D . L . , Hurd, J.D., Wilson, E . H . , Arnold, C . L . , Prisloe, M . P . 2002. "Quantifying and Describing Urbanizing Landscapes in the Northeast United States" Photogrammetric Engineering and Remote Sensing, 68 (10): 1083-1090. Clark, J. 2007. "Chemguide" accessed in March, 2007. http://www.chemguide.co.uk/ Comprehensive R Archive Network ( C R A N ) 2007. "The R Project for Statistical Computing" accessed in Apr/2, 2007. http://www.r-project.org/index.html Cooper, J.R., Gi l l iam, J.W., Daniels, R . B . , Robarge, W.P . 1987. "Riparian Areas as Filters for Agricultural Sediment" Soil Science Society of America Journal SSSJD 4, 51 (2) Davis, J .A. 1984. "Complexation of Trace Metals by Adsorbed Natural Organic Matter" Geochimica Et Cosmochimica Acta, 48 (4): 679-691. Davis, J .A. and Leckie, J.O. 1978. "Effect of Adsorbed Complexing Ligands on Trace Metal Uptake by Hydrous Oxides" Environmental Science & Technology, 12 (12): 1309-1315. D i Toro, D . M . , Mahony, J.D., Kirchgraber, P.R., O'Byrne, A . L . , Pasquale, L . R . , Picci r i l l i , D . C . 1986. "Effects of Nonreversibility, Particle Concentration, and Ionic Strength on Heavy-Metal Sorption" Environmental Science & Technology, 20 (1): 55-61. Dillaha, T . A . and Inamdar, S.P. 1996. "Buffer Zones as Sediment Traps Or Sources." Samara Publishing Limited, Samara House, Tresaith, Cardigan SA43 2JG ( U K ) . Dinicola, R.S. , 1990, Characterization and Simulation of Rainfall-runoff Relations for Headwater Basins in Western K i n g and Snohomish Counties, Washington, Department of the Interior, U S Geological Survey; Books and Open-File Reports Section Distributor Distributions in Sediments of A Major Sydney Estuary (Australia)" Environmental Geology, 28 (4): 169-174. DiVenere, V . J . 2006. "Stream Processes, Earth and Environmental Sciences, Earth Science Notes, Columbia University Summer Session 2006" accessed in Nov. , 2006. http ://www. Columbia. edu/~vj d 1 / streams_basic .htm Dorcey, A . H . J . , and Griggs, J.R., 1991, Water in Sustainable Development: Exploring Our Common Future in the Fraser River Basin: Vancouver, B . C . , Canada, Westwater Research Centre, Faculty of Graduate Studies, the University of British Columbia Egger, A . E . 2007. "Minerals III: the Silicates" accessed in Jan, 2007. http://www.visionlearning.com/library/module_viewer.php?mid=140&l=&c3= Enguix Gonzalez, A . , Ternero Rodriguez, M . , Jimenez Sa, J.C., Fernandez Espinosa, A . J . , Barragan de la Rosa F.J. 2000. "Assessment of Metals in Sediments in a Tributary of Guadalquiver River (Spain). Heavy Metal Partitioning and Relation between the Water and Sediment System" Water, A i r , & Soil Pollution, 121 (1): 11-29. Environment Canada 2004. "Canadian Sediment Quality Guidelines (CSeQGs)" accessed in Apr/02, 2007. http://wvvvv.ec.gc.ca/ceqg-rcqe/English/Ceqg/Sedimentydefault.cfm EPA/903/R-03/003" accessed in Apr, 2007. http://www.epa.gov/maia/html/llstream.htm Evans, L . J . 1989. "Chemistry of Metal Retention by Soils" Environmental Science & Technology, 23 (9): 1046-1056. Facchinelli, A . , Sacchi, E . , Mallen, L . 2001. "Multivariate Statistical and GIS-Based Approach to Identify Heavy Metal Sources in Soils" Environmental Pollution, 114 (3): 313-324. 103 Fergusson, J .E. 1990. The Heavy Metals: Chemistry, Environmental Impact and Health Effects., Pergamon Press, 382 p. Forman, R.T.T. and Alexander, L . E . 1998. "Roads and their Major Ecological Effects" Annual Review of Ecology and Systematics, 29 Forstner, U . 1979. "Metal Transfer between Solid and Aqueous Phases" Metal Pollution in the Aquatic Environment,: 197-270. Furia, T.E. , 1972, Chapter 6 - Sequestrants in Foods, Handbook of food additives, 2nd ed., C R C Press Gaffield, S.J., Goo, R . L . , Richards, L . A . , Jackson, R.J . 2003. "Public Health Effects of Inadequately Managed Stormwater R u n o f f American Journal of Public Health, 93 (9): 1527-1533. George Eby Research 2006. "Stability Constants (Log K l ) of various Metal Chelates" accessed in 2007 Mar, 2007. http://george-eby-research.com/html/stability_constants.html Goetz , . 2003. " I K O N O S Imagery for Resource Management: Tree Cover, Impervious Surfaces, and Riparian Buffer Analyses in the Mid-Atlantic Region." Remote sensing of environment. Gong, C. and Donahoe, R.J . 1997. " A n Experimental Study of Heavy Metal Attenuation and Mobil i ty in Sandy Loam Soils" Applied Geochemistry, 12 (3): 243-254. Greater Vancouver Regional District ( G V R D ) 2004. "Population Projection 2006-2031" accessed in Apr/10, 2007. http://www.gvrd.bc.ca/growth/keyfacts.htm Hal l , K . J . and Schreier, H . 1996. "Urbanization and Agricultural Intensification in the Lower Fraser River Valley: Impacts on Water use and Quality" GeoJournal, 40 (1): 135-146. Hal l , K . J . , Kiffney, P., Macdonald, R., McCal lum, D . , Larkin, G . , Richardson, J., Schrier, H . , Smith, J., Zandbergen, P., Keen, P. 1999. "Non-Point Source Contamination in the Urban Environment of Greater Vancouver: A Case Study of the Brunette River Watershed" Health of the Fraser River Aquatic Ecosystem, 2: 109-134. Hal l , K . J . , Kiffney, P., Macdonald, R., McCal lum, D . , Larkin, G . , Richardson, J., Schrier, H . , Smith, J., Zandbergen, P., Keen, P. 1999a. "Non-Point Source Contamination in the Urban Environment of Greater Vancouver: A Case Study of the Brunette River Watershed" Health of the Fraser River Aquatic Ecosystem, 2: 109-134. Hal l , K . J . , Kiffney, P., Macdonald, R., McCal lum, D . , Larkin, G . , Richardson, J., Schrier, H . , Smith, J., Zandbergen, P., Keen, P. 1999b. "Non-Point Source Contamination in the Urban Environment of Greater Vancouver: A Case Study of the Brunette River Watershed" Health of the Fraser River Aquatic Ecosystem, 2: 109-134. Ha l l , K . J . , Yesaki, I., Chan, J. 1976. "Trace Metals and Chlorinated Hydrocarbons in the Sediments of a Metropolitan." Watershed. Technical Report N o . 10, Westwater Research Centre, The University of British Columbia, Vancouver, B . C . Halling-Sorensen, B . , Lutzhoft, H . C . , Andersen, H.R. , Ingerslev, F. 2000. "Environmental Risk Assessment of Antibiotics: Comparison of Mecil l inam, Trimethoprim and Ciprofloxacin" The Journal of Antimicrobial Chemotherapy, 46 Suppl 1: 53-8; discussion 63-5. Harrison, R . M . and Wilson, S.J. 1985. "Chemical Composition of Highway Drainage Waters; I. Major Ions and Selected Trace Metals" Science of the Total Environment, 43 (1-2) Heiny, J.S. and Tate, C M . 1997. "Concentration, Distribution, and Comparison of Selected Trace Elements in Bed Sediment and Fish Tissue in the South Platte River Basin, U S A , 1992-1993" Archives of Environmental Contamination and Toxicology, 32 (3): 246-259. Heiri , O., Lotter, A . F . , Lemcke, G . 2001. "Loss on Ignition as a Method for Estimating Organic and Carbonate Content in Sediments: Reproducibility and Comparability of Results" Journal of Paleolimnology, 25 (1): 101-110. 104 Hickey, C . W . and Clements, W . H . 1998. "Effects of Heavy Metals on Benthic Macroinvertebrate Communities in N e w Zealand Streams" Environmental Toxicology and Chemistry, 17 (11): 2338-2346. Hoffman, E.J . , Latimer, J.S., Hunt, C D . , M i l l s , G . L . , Quinn, J .G. 1985. "Stormwater Runoff from Highways" Water, A i r , & Soil Pollution, 25 (4): 349-364. Holtan, H . , Kamp-Nielsen, L . , Stuanes, A . O . 1988. "Phosphorus in Soi l , Water and Sediment: A n Overview" Hydrobiologia, 170 (1): 19-34. Houston, J. 2004. "Land use, Riparian Buffers and the Effects of Urbanization on Stormwater Runoff in the Hoy Creek Watershed, Coquitlam, B . C . " A thesis submitted in partial fulfilment of the requirements for the Degree of Master of Science in the Faculty of Graduate Studies, Department of Soil Science, the University of British Columbia. Hubbard, R . K . and Lowrance, R .R. 1994. "Riparian Forest Buffer System Research at the Coastal Plain Experiment Station, Tifton, G A " Water, A i r , & Soil Pollution, 77 (3): 409-432. Hurd, J.D., Wilson, E . H . , Lammey, S.G., Civco, D . L . 2001. "Characterization of Forest Fragmentation and Urban Sprawl using Time Sequential Landsat Imagery." Proceeding of A S P R S 2001 Annual Convention, St. Louis, M O , A p r i l 23-27. Jackson, D . A . 1993. "Stopping Rules in Principal Components Analysis: A Comparison of Heuristical and Statistical Approaches" Ecology, 74 (8): 2204-2214. http://www.jstor.Org/view/00129658/di960355/96p00052/0 Jarvie, H.P. , Oguchi, T., Neal, C . 2002. "Exploring the Linkages between River Water Chemistry and Watershed Characteristics using GlS-Based Catchment and Locality Analyses" Regional Environmental Change, 3 (1): 36-50. Karr, J.R. 1981. "Assessment of Biotic Integrity using Fish Communities" Fisheries, 6 (6): 21-27. Kel ly , J., Thornton, I., Simpson, P.R. 1996. "Urban Geochemistry: A Study of the Influence of Anthropogenic Activi ty on the Heavy Metal Content of Soils in Traditionally Industrial and Non-Industrial Areas of Britain" Applied Geochemistry, 11 (1): 363-370. Kheboian, C. and Bauer, C F . 1987. "Accuracy of Selective Extraction Procedures for Metal Speciation in Model Aquatic Sediments" Analytical Chemistry, 59 (10): 1417-1423. Landner, L . 1987. "Speciation of Metals in Water, Sediment and Soil Systems" Speciation of Metals in Water, Sediment and Soil Systems: Proceedings of an International Workshop, Sunne, October 15—16, 1986.Editor: Lars Landner, Lecture Notes in Earth Sciences, http://adsabs.harvard.edu/abs/1987LNES...! 1 L Lee, K . H . , Isenhart, T . M . , Schultz, R . C , Mickelson, S.K. 2000. "Multispecies Riparian Buffers Trap Sediment and Nutrients during Rainfall Simulations" Journal of Environmental Quality, 29 (4): 1200-1205. Lindsay, W . L . 1979. Chemical Equilibria in Soi l s . : New York, John Wiley, 449 p. Loranger, S. and Zayed, J. 1994. "Manganese and Lead Concentrations in Ambient A i r and Emission Rates from Unleaded and Leaded Gasoline between 1981 and 1992 in Canada: A Comparative Study" Atmospheric Environment, 28 (9): 1645-1651. Lowrance, R. 1998. "Riparian Forest Ecosystems as Filters for Nonpoint-Source Pollution." Successes, Limitations and Frontiers in Ecosystem Science, Springer-Verlag, N e w York, U S A . Luttmerding, H . A . 1980. "Soils of the Langley-Vancouver Map Area." R A B Bulletin 18, Report N o . 15, British Columbia Soil Survey, Volume 1, Soil Map Mosaics and Legend, Lower Fraser Valley. Luttmerding, H . A . and Sprout, P . N . 1968. "Soil Survey of Miss ion Area." Preliminary Report N o . 9 of the Lower Fraser Valley Soil Survey, British Columbia Department of Agriculture, Kelowna, B . C . http://sis2.agr.gc.ca/cansis/publications/bc/bc9_pre/bc9_pre_report.pdf 105 Mackintosh, E .E . and Gardner, E . H . 1966. " A Mineralogical and Chemical Study of Lower Fraser Valley Al luv ia l Sediments" Canadian Journal of Soil Science, 46 (1): 37-45. Manly, B .F . J . , 2004, Multivariate Statistical Methods: a primer (Third ed.), Chapman & H a l l / C R C Marsalek, J. and Chocat, B . 2002. "International Report: Stormwater Management" Water Science & Technology, 46 (6): 1-17. May, C , Horner, R.R. , Karr, J.R., Mar, B . W . , Welch, E . B . 1998. "The Cumulative Effects of Urbanization on Small Streams in the Puget Sound Lowland Ecoregion" Proceedings of the Puget Sound Research, May, C .W. , Horner, R.R. , Karr, J.R., Mar, B . W . , Welch, E . B . 1996. "Assessment of Cumulative Effects of Urbanization on Small Streams in the Puget Sound Lowland Ecoregion." Urban Streams Conference, November 15-17 1996, Areata, C A . http://scholar.google.com/url?sa=U&q=http://vvvvw.psat. wa.gov/Publications/98_proceeding s/pdfs/1 a_may.pdf McBride, M . , 2001, Spatial effects of urbanization on physical conditions in Puget Sound Lowland streams: University of Washington, Ph.D. dissertation, McCal lum, D . 1995. " A n Examination of Trace Metal Concentration and Land use in an Urban Watershed." A thesis submitted in partial fulfilment of the requirements for the Degree of Master of Applied Science in the Faculty of Graduate Studies, Department of C i v i l Envineering, the University of British Columbia McDonald, D . K . 2000. "Ecologically Sound Lawn Care for the Pacific Northwest" Crop Science, 45: 546-552. McKee , L . , Leatherbarrow, J., Pearce, S., Davis, J. 2003. " A Review of Urban Runoff Processes in the Bay Area." A Report Prepared for the Sources Pathways and Loadings Workgroup ( S P L W G ) , San Francisco Estuary Regional Monitoring Program for Trace Substances (RMP) . http ://www.sfei.org/rmp/reports/splwg/Urban_runoff_literature~000.pdf McPherson, T . N . , Burian, S.J., Stenstrom, M . K . , Turin, H.J . , Brown, M . J . , Suffet, I .H. 2005. "Trace Metal Pollutant Load in Urban Runoff from a Southern California Watershed" Journal of Environmental Engineering, 131: 1073. Mesuere, K . and Fish, W . 1989. "Behavior of Runoff-Derived Metals in a Detention Pond System" Water, A i r , & Soil Pollution, 47 (1): 125-138. Minton, G.R. , 2005, Stormwater Treatment: Biological , Chemical, and Engineering Principles (Second ed.): Seattle, W A , Resource Planning Associates, 472 p. Moore, J.W., and Ramamoorthy, S., 1984, Heavy Metals in Natural Waters, Springer, 268 p. Morley, S.A. and Karr, J.R. 2002. "Assessing and Restoring the Health of Urban Streams in the Puget Sound Basin" Conservation Biology, 16 (6): 1498-1509. Mosley, L . M . and Peake, B . M . 2001. "Partitioning of Metals(Fe, Pb, Cu , Zn) in Urban Run-Off from the Kaikorai Valley, Dunedin, N e w Zealand" N e w Zealand Journal of Marine and Freshwater Research, 35 (3): 615-624. Munsell Color, 1975, Munsell Soil Color Charts, Munsell Color, Macbeth Divis ion of Kollmorgan Corp Naiman, R.J . and Decamps, H . 1997. "The Ecology of Interfaces: Riparian Zones" Annual Review of Ecology and Systematics, 28: 621-658. Naiman, R.J . , Decamps, H . , Pollock, M . 1993. "The Role of Riparian Corridors in Maintaining Regional Biodiversity" Ecological Applications, 3 (2): 209-212. Oberts, G . L . 1986. "Pollutants Associated with Sand and Salt Applied to Roads in Minnesota" Water Resources Bulletin, 22 (3): 479-484. Osborne, L . L . and Kovacic, D . A . 1993. "Riparian Vegetated Buffer Strips in Water-Quality Restoration and Stream Management" Freshwater Biology.Oxford, 29 (2): 243-258. 106 Ourso, R.T. and Frenzel, S.A. 2003. "Identification of Linear and Threshold Responses in Streams Along a Gradient of Urbanization in Anchorage, Alaska*" Hydrobiologia, 501 (1): 117-131. Paul, M . J . and Meyer, J .L. 2001. "Streams in the Urban Landscape" Annual Review of Ecology and Systematics, 32: 333-365. Pekey, H . , Bakoglu, M . , Pekey, B . 2005. "Sources of Heavy Metals in the Western Bay of Izmit Surface Sediments" International Journal of Environmental & Analytical Chemistry, 85 (14): 1025-1036. Petrovic, M . , Kastelan-macan, M . , Horvat, A . J . M . 1999. "Interactive Sorption of Metal Ions and Humic Acids Onto Mineral Particles" Water, A i r , & Soil Pollution, 111 (1): 41-56. Pott, U . and Turpin, D . H . 1996. "Changes in Atmospheric Trace Element Deposition in the Fraser Valley, B . C . , Canada from 1960 to 1993 Measured by Moss Monitoring with Isothecium Stoloniferum" Canadian Journal of Botany/Revue Canadien De Botanique, 74 (8): 1345-1353. Preciado, H.F . and L i , L . Y . 2006. "Evaluation of Metal Loadings and Bioavailability in A i r , Water and Soil Along Two Highways of British Columbia, Canada" Water, A i r , & Soil Pollution, 172 (1): 81-108. Punj, G . and Stewart, D . W . 1983. "Cluster Analysis in Marketing Research: Review and Suggestions for Application" Journal of Marketing Research, 20 (2): 134-148. http://links.jstor.org/sici?sici=0022-2437(198305)20%3A2%3C134%3ACAIMRR%3E2.0.CO%3B2-C Quek, U . and Forster, J. 1993. "Trace Metals in Roof Runof f Water, A i r , & Soil Pollution, 68 (3): 373-389. Revitt, D . M . , Hamilton, R.S. , Warren, R.S . 1990. "The Transport of Heavy Metals within a Small Urban Catchment" The Science of the Total Environment, 93: 359-373. Riba, I., DelVal ls , T .A . , Forja, J . M . , Gomez-Parra, A . 2002. "Evaluating the Heavy Metal Contamination in Sediments from the Guadalquivir Estuary After the Aznalcollar Min ing Spil l (SW Spain): A Multivariate Analysis Approach" Environmental Monitoring and Assessment, 77 (2): 191-207. Rose, S., Crean, M . , Sheheen, D . , Ghazi, A . 2001. "Comparative Zinc Dynamics in Atlanta Metropolitan Region Stream and Street R u n o f f Environmental Geology, 40 (8): 983-992. Roy, A . H . , Freeman, M . C . , Freeman, B . J . , Wenger, S.J., Ensign, W . E . , Meyer, J .L. 2006, "Importance of Riparian Forests in Urban Catchments Contingent on Sediment and Hydrologic Regimes" Environmental Management, 37 (4): 523-539. Sabin, L . D . , L i m , J .H. , Stolzenbach, K . D . , Schiff, K . C . 2005. "Contribution of Trace Metals from Atmospheric Deposition to Stormwater Runoff in a Small Impervious Urban Catchment" Water Research, 39: 3929-3937. Sansalone, J.J. and Buchberger, S.G. 1997. "Partitioning and First Flush of Metals in Urban Roadway Storm Water" Journal of Environmental Engineering, 123 (2): 134-143. Sawaya, K . E . , Olmanson, L . G . , Heinert, N . J . , Brezonik, P .L . , Bauer, M . E . 2003. "Extending Satellite Remote Sensing to Local Scales: Land and Water Resource Monitoring using High-Resolution Imagery" Remote Sensing of Environment, 88 (1): 144-156. Schindler, D .W. , Di l lon , P.J., Schreier, H . 2006. " A Review of Anthropogenic Sources of Nitrogen and their Effects on Canadian Aquatic Ecosystems" Biogeochemistry, 79 (1): 25-44. Schreier, H . and Brown, S. 2004. "Multiscale Approaches to Watershed Management: Land-use Impacts on Nutrient and Sediment Dynamics." Scales in Hydrology and Water Management, I A H S Press, : 61-75. 107 Schreier, H . , Bestbier, R., Brown, S., Hal l , K . 2002. "Agricultural Watershed Management." A multi-media C D - R O M textbook for a graduate level course for distance learning, Institute of Resources, Environment and Sustainability, University of British Columbia. Schreier, H . , Bestbier, R., Derksen, G . 2004. " A Quantitative Assessment of Agricultural Intensification and Associated Waste-Management Challenges in the Lower Fraser Val ley" 2003 Georgia Basin/Puget Sound Resarch Conference Proceedings, Feb., Schreier, H . , Ha l l , K . , Brown, S., Tamagi, W. , Lavkulich, L . M . 1997. "Integrated Watershed Management." A n electronic multi-media textbook (700 pages) for internet graduate study courses. C D - R O M @ IRE, U B C , Distributed learning, Continuing studies, U B C . Schueler, T.R. 1994. "The Importance of Imperviousness" Watershed Protection Techniques, 1 (3): 100-111. Shan, X . and Chen, B . 1993. "Evaluation of Sequential Extraction for Speciation of Trace Metals in Model Soil Containing Natural Minerals and Humic A c i d " Analytical Chemistry, 65 (6): 802-807. Shea, J .A. 2003. "Trace Metal Concentrations in Stream Sediments from Urban and Less-Urbanized Watersheds in the Piedmont Province of Georgia" Geological Society of America, Seattle Annual Meeting, Abstracts with Programs, 35 (6): p. 240. http://gsa.confex.com/gsa/2003AM/finalprogram/abstract_59851 .htm Shelton, L . R . and Capel, P .D. 1994. "Guidelines for Collecting, Processing Samples of Stream Bed Sediment for Analysis of Trace Elements and Organic Contaminants for the National Water-Quality Assessment Program." U . S . Geological Survey, Eearth Science Information Center, Open-File Reports 94-458, 20 pp. Sleavin, W.J . , Prisloe, S., Giannotti, L . , Civco, D . L . 2000. "Measuring Impervious Surfaces for Non-Point Source Pollution Modeling." Proceedings, 2000 A S P R S Annual Conference.May 22-26, Washington, D C . Smith, D . L . , Harris, A . D . , Johnson, J .A. , Silbergeld, E . K . , Morris Jr, J .G. 2002. "Animal Antibiotic use has an Early but Important Impact on the Emergence of Antibiotic Resistance in Human Commensal Bacteria" Proceedings of the National Academy of Sciences, 99 (9): 6434-6439. Smith, I. 2004. "Cumulative Effects of Agricultural Intensification on Nutrient and Trace Metal Pollution in the Sumas River Watershed, Abbotsford, B . C . " A thesis submitted in partial fulfilment of the requirements for the Degree of Master of Science in the Faculty of Graduate Studies, Resource Management and Environmental Studies, the University of British Columbia. Smith, R . M . , Martell, A . E . , Motekaitis, R.J . 2001. "NIST Critically Selected Stability Constants of Metal Complexes: Version 6.0." N I S T Standard Reference Database 46, U . S . Department of Commerce, National Institute of Standards and Technology (NIST), Standard Reference Data Program,. Snyder, M . N . , Goetz, S.J., Wright, R . K . 2005. "Stream Health Rankings Predicted by Satellite Derived Land Cover Metrics" Journal of the American Water Resources Association, 41 (3): 659-677. Sparks, D . L . , 1995, Environmental Soil Chemistry (1st ed.), Academic Press SPSS Inc. 2007. "SPSS 15.0 for Windows®." http://www.spss.com/spss/whats_new_modules.htm Stein, E . D . and Ackerman, D . 2007. "Dry Weather Water Quality Loadings in Ar id , Urban Watersheds of the Los Angeles Basin, California, U S A " Journal of the American Water Resources Association, 43 (2): 398-413. Stevenson, F.J . 1976. "Stability Constants of Cu2+ , Pb2+ , and Cd2+ Complexes with Humic Acids" Soil Science Society of America Journal, 40: 665-672. 108 Sutherland, R . A . 2000. "Bed Sediment-Associated Trace Metals in an Urban Stream, Oahu, Hawaii" Environmental Geology, 39 (6): 611-627. Szeto, S .Y. and Price, P . M . 1991. "Persistence of Pesticide Residues in Mineral and Organic Soils in the Fraser Valley of British Columbia" Journal of Agricultural and Food Chemistry, 39 (9): 1679-1684. Szeto, S.Y., Grove, G. , Liebscher, H . , H i i , B . , Zebarth, B . J . 1994. "Nonpoint-Source Groundwater Contamination by 1, 2, 2-Trichloropropane, A Trace Impurity in Soil Fumigant Formulations" Journal of Environmental Quality, 23 (6): 1367-1370. Tessier, A . , Campbell, P .G .C . , Bisson, M . 1979. "Sequential Extraction Procedure for the Speciation of Particulate Trace Metals" Analytical Chemistry, 51 (7): 844-851. http://pubs.acs.Org/cgi-bin/abstract.cgi/ancham/l 979/5 l/i07/f-pdf/f_ac50043a017.pdf?sessid=600613 Tessier, A . , Rapin, F., Carignan, R. 1985. "Trace Metals in Oxic Lake Sediments: Possible Adsorption Onto Iron Oxyhydroxides" Geochimica Et Cosmochimica Acta, 49 (1): 183-194. Thomson, N . R . , McBean, E . A . , Snodgrass, W. , Monstrenko, L B . 1997. "Highway Stormwater Runoff Quality: Development of Surrogate Parameter Relationships" Water, A i r , & Soil Pollution, 94 (3): 307-347. Tomalty, R. 2002. "Growth Management in the Vancouver Region" Local Environment, 7 (4): 431-445. Traina, S.J. and Laperche, V . 1999. "Contaminant Bioavailability in Soils, Sediments, and Aquatic Environments" Proceedings of the National Academy of Sciences of the United States of America, 96 (7): 3365-3371. http://www.pnas.Org/cgi/content/abstract/96/7/3365 Tufford, D . L . , McKel la r Jr, H . N . , Hussey, J.R. 1998. "In-Stream Nonpoint Source Nutrient Prediction with Land-use Proximity and Seasonality" Journal of Environmental Quality, 27 (1): 100-110. Tyler, G . 1995. ' T C P - M S , Or I C P - A E S and A A S 7 - A Comparison" Spectroscopy Europe, 7 (1): 14-22. http://formation-concours.univ-lyonl.fr/annales/annales_2004/2004_a_asi_sciencemateriaux3_laroch.pdf United States Environmental Protection Agency (US E P A ) 2006. "Developing Biological Indicators: Lessons Learned from Mid-Atlantic Streams. United States Environmental Protection Agency (US E P A ) , 1992, Methods for the determination of metals in environmental samples: supplement I, Environmental Monitoring Systems Laboratory (Cincinnati, Ohio).; United States Environmental Protection Agency. Office of Research and Development United States Environmental Protection Agency (US E P A ) . 1990. "National Water Quality Inventory: 1988 Report to Congress." Office of the Water Program Operations, Water Planning Division, Washington D C , U S A . United States Environmental Protection Agency (US E P A ) . 2000. "National Water Quality Inventory." Report EPA-841-R-02-001. United States Environmental Protection Agency (US E P A ) . 2002. "Standard Operating Procedure, for Nitrate - Nitrite in Lake Water (QuikChemFIA+8000 Method)." LG203 Revision. http://www.epa.gov/glnpo/monitoring/procedures/sop2007/Ch2/LG203_050314.pdf United States Environmental Protection Agency (US E P A ) . 2004. "Mult i -Media, Mul t i -Concentration, Inorganic Analytical Service for Superfund (ILM05.3)." United States Environmental Protection Agency, Office of Solid Waste and Emergency Response, O S W E R Document 9240.1-43FS, E P A Publication 540-F-04-001. http://www.epa.gov/superfund/programs/clp/download/ilm/ilm53fs.pdf 109 United States Geological Survey (USGS) 1999. "Sediment Toxicity Testing Methods and Data Interpretation: Sediment Effect Concentration Database" accessed in Dec/15, 2006. http://vvww.cerc.usgs.gov/pubs/sedtox/sec.htm Ure, A . M . and Berrow, M . L . 1982. "The Elemental Constituents of Soils" Environmental Chemistry, 2: 94-204. Vizcarra, A . T . , Lo , K . V . , Lavkulich, L . M . 1997. "Nitrogen Balance in the Lower Fraser River Basin of British Columbia" Environmental Management, 21 (2): 269-282. Wahl , M . , McKel la r , H . , Will iams, T. 1997. "Patterns of Nutrient Loading in Forested and Urbanized Coastal Streams" Journal of Experimental Marine Biology and Ecology, 213 (1): 111-131. Walling, D . E . and Amos, C M . 1999. "Source, Storage and Mobilisation of® Ne Sediment in a Chalk Stream System" Hydrological Processes, 13: 323-340. Wang, L . , Lyons, J., Kanehl, P., Bannerman, R. 2001. "Impacts of Urbanization on Stream Habitat and Fish Across Multiple Spatial Scales" Environmental Management, 28 (2): 255-266. Wenger, S., 1999, A Review of the Scientific Literature on Riparian Buffer Width, Extent and Vegetation, Office of Public Service & Outreach; University of Georgia; Institute of Ecology Westwater Research Centre. 1973. "Trace Metals in Stream Sediments in Lower Fraser Valley." Unpublished data collected by Westwater Research Centre at the University of British Columbia. White, M . D . and Greer, K . A . 2006. "The Effects of Watershed Urbanization on the Stream Hydrology and Riparian Vegetation of Los Penasquitos Creek, California" Landscape and Urban Planning, 74 (2): 125-138. Wohl , J. 1996. " A Literature Review of the Economics of Manure Management Options in the Lower Fraser Valley." Fraser River Act ion Plan, D O E F R A P 1996-15, Prepared for B C Ministry of Environment, Lands and Parks, Environment Canada, and B C Ministry of Agriculture, Fisheries. http://wvVw.rem.sfu.ca/FRAP/9615.pdf World Health Organization (WHO) 2007. "Chemical Hazards in Drinking-Water" accessed in Feb, 25, 2007. http://www.who.int/water_sanitation_health/dwq/chemicals/en/index.html#A Young, L . B . and Harvey, H . H . 1991. "Metal Concentrations in Chironomids in Relation to the Geochemical Characteristics of Surficial Sediments" Archives of Environmental Contamination and Toxicology, 21 (2): 202-211. Young, L . B . and Harvey, H . H . 1992. "The Relative Importance of Manganese and Iron Oxides and Organic Matter in the Sorption of Trace Metals by Surficial Lake Sediments" Geochimica Et Cosmochimica Acta, 56 (3): 1175-1186. Zandbergen, P. 1998. "Urban Watershed Ecological Risk Assessment using GIS: A Case Study of the Brunette River Watershed in British Columbia, Canada." Journal of hazardous materials. Zandbergen, P., Schreier, H . , Bestbier, G . , Hal l , K . , Brown, S. 2000. "Urban Watershed Management." A multi-media C D - R O M textbook for a graduate level, internet-distributed, learning course (with video, voice, images, GIS, graphics and text), C D - R O M @ I R E . 110 Appendices A. Chemistry of Sediment Quality This section is provided to present the background chemistry to understand origin, solubility, mobility and attenuation of cations within the hydrologic cycle. Chemical bonding In ionic bonding, electrons are completely transferred from one atom to another. Either losing or gaining negatively charged electrons, forming ions, and attracted to each other. This bonding dissolves easily in water. In covalent bonding, elements w i l l share electrons in an effort to f i l l their valence shells. In coordinate covalent bonding, the bonding pair comes from only one of the atoms called the donor atom. The other atom, the acceptor atom, simply accepts the sharing responsibilities. In metallic bonding, each atom loses its valence electrons to become a positive ion. This bonding is the electrostatic attraction between the positively charged ions and negatively charged electrons - "an array of positive ions in a sea of electrons". A s a whole, the strength of the bonding follows the order of, covalent bonding > coordinate covalent > metallic > ionic (Carpi 2007). Primary minerals Minerals are a natural inorganic compound with definite physical, chemical, and crystalline properties. Primary minerals are not altered chemically since its deposition and crystallization from molten lava, and can be classified according to the differences in the shape of tetrahedral arrangement (silica bonds to four surrounding oxygens) (Table A - l ) (Egger 2007). I l l Table A. 1 Primary minerals and their characteristics. Modified from Egger (2007). Olivine M g z + and F e z + are octahedral ly coord inated by 0 a t o m s linked with S i 0 4 tetrahedra. Py roxenes , amph ibo les Sil icate an ions polymerize; they share an oxygen a tom with a neighbor ing tetrahedron. T h e c h a i n s bond to cat ions i.e., F e 2 + , M g 2 + , and C a 2 + to ba lance the negat ive charge . M i c a s (muscovite, biotite) Every tetrahedron s h a r e s three of its o x y g e n ions with neighboring tetrahedra, shee ts are formed. V a n der W a a l s force works between the sheets . Quartz , fe ldspar E a c h tetrahedron s h a r e s all of its oxygen a t o m s with adjacent tetrahedra, a very strong 3 -d imensional f ramework of S i - 0 b o n d s is formed. Quar tz is pure S i 0 2 . In the fe ldspars , o n e or two out of every four S i 4 + ions is rep laced by an A l 3 + , creating a c h a r g e imba lance that must be so lved through the p r e s e n c e of additional cat ions: K + , N a + , and C a 2 + . - K- fe ldspars: K + cat ion, or alkali fe ldspar - P lag ioc lase fe ldspars: N a + a n d C a 2 + Secondary minerals Secondary mineral is one resulting from the weathering of a primary mineral; either by an alteration in the structure or from re-precipitation of the products of weathering (dissolution) of a primary mineral. They occur primarily in the clay fraction but also located in silt. Phyllosilicates Clay minerals are called phyllosilicates. Phyllosilicates are assemblages of tetrahedral and octahedral sheets, in which one silica is surrounded by six oxygens. Isomorphous substitution plays an important role in forming clay minerals; the atom can be substituted by the one which has similar cationic radius, which determines the coordination number of the ions. This results in the permanent negative charges of clay minerals. In the tetrahedral sheet, A l 3 + substitutes for S i 4 + " In the octahedral sheet, Fe 2 + , Fe 3 + , M n 2 + , M g 2 + , N i 2 + , Z n 2 + , or C u 2 + can substitute for A l 3 + . Another component of the negative charge in soil is due to pH dependent charges. This is caused by protonation and deprotonation of functional groups on inorganic soil minerals (clay minerals, amorphous materials, oxides etc.) and soil organic matter. Cation exchange capacities (CEC) of soil minerals mean the net negative charges which can be balanced by positive charge in the 112 form of exchangeable cations. A major component of a C E C is attributable to the secondary clay minerals and soil organic matter (Sparks 1995). Oxides and hydroxides Major cations constantly leach from minerals throughout weathering, and oxygen reacts with the residual components of aluminum, iron and manganese, forming oxides and hydroxides. They are highly insoluble, and have high specific surface areas and reactivity, which leads to adsorbing cations on their surfaces. Most of them are amorphous, and exist as coatings on phyllosilicates and humic substances, and as mixed gels. Gibbsite [Al(OH)3] and bohemite ( A l O O H ) are important for aluminum. Goethite (FeOOH) and hematite (Fe20s) are important for iron (Sparks 1995). Cations and ligands Hard cations are alkali metals (Na + and K + ) and alkaline earth metals ( M g 2 + and C a 2 + ) that interact via electrostatic, ionic reactions (group I). Soft cations are, for instance, C u + , Z n 2 + , C d 2 + , Pb react to form metallic and covalent bonds (group III). Transitional metals form complexes of intermediate strength (group II) (Figure A - l ) . The hard donor atoms such as F and O prefer hard metal ions while the soft donor atoms (e.g., I, S) prefer softer metal ions. Ligands can be divided into (1) simple inorganic ligands, X , major anions - their donor atom is oxygen and they prefer hard metals; (2) hard sites of natural organic matter referred to as L H - carboxyl and phenolic sites; (3) soft sites of soil organic matter, denoted as Ls, N and S containing sites (Figure A-2) . Group I metals prefer hard ligands, but weak complexes. Group III metals have greater affinity for soft sites Ls than for hard sites L H or X ligands (Sparks 1995). Bivalent group II metals ( M n 2 + 2T" 2"T" 2"f" * Fe Co N i C u ) have affinity for both hard and soft sites and thus compete with, • Group I metals for LH sites. - 1 are less strongly bound, but at higher concentrations. • Group III metals for Ls sites - III are at lower concentrations, but more strongly bound. Hydroxides or hydrous oxides, such as Fe(III), Al(III) and Mn(IV) , have very reactive surfaces, fixing and adsorbing other metals on their surfaces (Buffle and de Vitre 1993). 113 Table A . 2 Classification of metals and donor atoms of complexing sites by their hard and soft characters. Modified from Buffle and de Vitre (1993). Metal G r o u p s Pre fe rence of Meta ls for L igand A t o m s Hard Metal Ions: 1 Hard donor a t o m s Soft donor a t o m s (H + ) , L f , N a \ K + , B e 2 + , M g 2 + , C a 2 + , S r 2 + , F e 3 + , A l 3 + , S c T F > 0 > N ~ C l > Br > 1 > S Transit ion Metal Ions: II C r 2 * , M n 2 + , F e 2 + , C o 2 + , N i 2 + , C u 2 + Soft Metal Ions: III A g + , A u + , T i + , C u + , Z n 2 + , C d 2 + , H g 2 + , P b 2 + , S n 2 + F < 0 < N ~ C l < Br <1 < S Table A. 3 Range of concentrations of ligands or complexing sites in natural fresh waters. =S-OH refers to inorganic solid surface sites. - C O O H and - O H refer to total concentrations of carboxyl and phenolic sites in natural organic matter. N o r g , S o r g refer to total concentrations of organic nitrogen and sulfur. Modified from Buffle (1984). Simple l igands X Cl", C 0 3 2 ' , S 0 4 2 " , S - O H , F", P 0 4 3 " S i tes of organic complexants L H Ls - C O O H , - O H Norg. S 0 rg C a a n d ' M g have their principal function in stabilizing or modifying the conformation of biological structures (the cellular "skeleton"). Group II metals provide the majority of metal co-enzymes, frequently as a result of their redox properties. Group III metals are very toxic because their complexes with Ls sites are often more stable than those formed by group II metals. They can therefore replace group II metals and thereby block catalytic action or modify vital structures. Soft sites in the water system are proteins and aquagenic fulvic acids that have a strong tendency to adsorb on sedimenting inorganic particles. They can be even more strongly blocked by reacting with either S(II) to form insoluble metal sulfides or with sediment organic matter that has a high content of S-containing sites (Buffle and de Vitre 1993). 114 Fundamentals of the Elements Alkali metal and Alkaline Earth Metals • Alkali metal A l k a l i metals (Na: sodium, K : potassium) are found in the alkali feldspars and micas. Their existance is controlled by the weathering rate of K-feldspars, affected by p H , and A l 3 + and S i 4 + concentrations. In feldspar, SiC>4 and A I O 4 tetrahedra have cavities which hold K + , N a + and C a 2 + to keep electroneutrality. K + released by weathering is adsorbed by soil colloids and clays, or form K-mineral such as illite. N o chelates or insoluble salts of K exist in soils nor is K fixed in organic combination. (Ure and Berrow 1982). • Alkaline earth metal M g 2 + has a coordination number of six and thus can replace A l 3 + in octahedra. It is an important constituent of a large number of common rock forming silicates (i.e., olivines, pyroxenes, 2_|_ amphiboles, micas) and of sediments and dolomite (Ure and Berrow 1982). C a is the fifth element in the order of abundance. It occurs in various forms, and in sedimentary rocks such as in limestones or dolomites. Metals Forming Oxides/Hydroxides • Aluminum Aluminum is the most abundant metallic element and constitutes about 8% of the Earth's crust. A l 3 + has a coordination number of 4 or 6, and thus exists both in tetrahedra and octahedra as replacing S i 4 + . Oxides and hydroxides (e.g., gibbsite [Al(OH)3], bohemite [ A l O O H ] , corundum [AI2O3]) after rocks being weathered are insoluble, but extremely soluble in low p H (Sparks 1995). It can form complexes with various organic compounds (e.g., humic or fulvic acids) and inorganic ligands (e.g. fluoride, chloride, and sulfate), most but not all of which are soluble. In pure water, aluminum has a minimum solubility in the p H range 5.5-6.0; concentrations of total dissolved aluminum increase at higher and lower p H values ( W H O 2007). This is because A l is amphoteric. 115 • Iron F e 2 + and F e 3 + have a coordination number of 6, replacing A l 3 + in octahedra which are linked to aluminosilicate tetrahedra, and form ferromagnesian minerals (e.g., olivine, pyroxenes and amphiboles). In contrast, the framework-type silicates such as quartz and feldspars contain only trance amount of iron (Sparks 1995). After rock weathering, insoluble iron oxides (e.g., goethite [FeOOH] and hematite [Fe2C>3]) commonly exist as a coating of phyllosilicates, and attract cations onto their high surface areas. Iron also readily form carbonates, phosphates and sulfides.(Ure and Berrow 1982; W H O 2007). • Manganese The most environmentally and biologically important manganese compounds are those that contain M n 2 + , M n 4 + or M n 7 + , and M n oxides occur as coatings on soil partcles, adsorbing cations. In surface waters, M n occurs in both dissolved and suspended forms, depending on such factors as p H , anions present and redox potential. For instance, as redox potential of M n 3 + and M n 4 + is very high, their oxides can oxidize C r 3 + to C r 6 + at p H 5 because the range of pe values for 94-reduction of manganese is higher than that for chromium reduction. M n predominates in most water at p H 4-7, but more highly oxidized forms may occur at higher p H values (Sparks 1995; W H O 2007). Transitional Metal Ions • Chromium • • 3+ • The common use of chromium is as an electroplated protective coating against corrosion. Cr is one of the least toxic of the trace elements, considering its oversupply and essentiality. C r 6 + is strongly oxidizing and very toxic. (Moore and Ramamoorthy 1984). Cr(VI) can easily be reduced to Cr(III) by organic matter. In water, Cr(III) is a positive ion that forms hydroxides and complexes, and is adsorbed at relatively high p H values. In general, Cr(VI) salts are more soluble than those of Cr(III), making Cr(VI) relatively mobile ( W H O 2007). Cobalt The common oxidation states of cobalt are +2 or +3. The distribution of cobalt after release by weathering is dependent upon the type of clay being formed and on the formation of iron and Fe and M n oxide phases. Cobalt released by weathering is strongly accumulated in clay minerals, Fe 116 oxides and M n oxides. Cobalt does not seem to be strongly associated with organic matter in soils and, where secondary M n oxides are a significant phase, these appear to be the most active in adsorbing cobalt. Cobalt is adsorbed much more strongly on to M n than on to Fe oxides (Ure and Berrow 1982). • Nickel A n oxidation state of +2 is the most common. It is classified as a borderline element between hard and soft acid, and abundant as oxides, carbonates, silicates, and sulfides. Chloride, nitrate, and sulfate are soluble in water, whereas carbonates and hydroxides are far less soluble and sulfides and oxides are practically insoluble in water (Moore and Ramamoorthy 1984; W H O 2007). Nicke l has one of the least toxic priority of the heavy metals, and is essential at trace levels for human health. Acute toxicity may arises from competitive interaction and interference with five major essential elements, Ca, Co, Cu , Fe and Z n (Moore and Ramamoorthy 1984). Soft Metal Ions • Copper The oxidation state of copper complexes are +1, +2, +3, although Cu(II) (cupric) is the most common. C u + (cuprous) is a typical soft acid (Moore and Ramamoorthy 1984). Copper strongly adsorbs to clay materials in a pH-dependent fashion, and adsorption is increased by the presence of particulate organic materials (Barceloux, 1999; Landner & Lindestrom, 1999). Free copper ions are chelated by humic acids and polyvalent organic anions ( W H O 2007). Copper is highly toxic to most freshwater aquatic species (Moore and Ramamoorthy 1984). • Zinc Zn-carbonate and sulfide are the important compounds. Sulfide weathering gives rise to relatively high concentrations of dissolved zinc because of the solubility of zinc sulfate. The released zinc is generally adsorbed by clays and secondary oxides. Clay content, iron oxides, organic matter, and p H are also important in controlling the amount of zinc (Ure and Berrow 1982; Moore and Ramamoorthy 1984). Stability of Zn-organic complexes is enhanced by the presence of N and S donor atoms in the ligand. Acute toxicity of zinc to freshwater invertebrates is relatively low. A s with other metals, increasing water hardness decreases toxicity to 117 invertebrates. Zinc is an essential element. Z n deficiency leads to delayed healing, suppression of enzymatic activity, and immune response (Moore and Ramamoorthy 1984). • Lead Lead is a member of the group IV elements (C, S i , Ge, Sn and Pb) of the periodic classification. Lead is truly metallic compared to carbon and silicate. Lead has stable +2 and +4 oxidation states, and form organo derivatives, such as tetraethyllead - an antiknock agent in gasoline (Moore and Ramamoorthy 1984). Lead being easily adsorbed on clay minerals and organic substances. Lead tends to become firmly fixed in soils particularly those rich in organic chelating groups. The inorganic salts of lead are generally highly insoluble and lead can therefore be precipitated as sulfate, carbonate, or phosphate (Ure and Berrow 1982). • Cadmium Cadmium is the second member of the Group l i b triad (Zn, Cd , Hg) in the periodic table, and the stable state is Cd(l l ) . Moderate covalency in bonds leads to high affinity for sulfhydryl groups, increasing l ipid solubility, bioaccumulation, and toxicity. C d seems to displace Z n in many vital enzymatic reactions, causing disruption or cessation of the activities of organisms. The most significant example of cadmium poisoning in humans, itai-itai disease in Japan form the 1940's to I960'. Patients suffering from the disease showed signs of osteomalacia in bones and calcification and pyelonephritis in kidneys. This resulted in skeletal deformation and renal dysfunction (Moore and Ramamoorthy 1984). Phosphorus Erosion and surface runoff introduce phosphorus (P) to a stream. Particulate P adsorbed on colloidal surfaces can be 75-90% of total P in runoff as it is extremely insoluble. Common inorganic source of P is apatite. Particulate P can be organic P from agricultural, residential and industrial sources, phosphates, clays, and oxides. Complex or condensed phosphates (polyphosphates) which are mainly manmade for use (e.g., detergents, material from water treatments) are discharged with domestic and industrial wastewaters, as well as generated by all l iving organisms (dissolved inorganic form; H2PO4", HPO4" ) (Holtan et al. 1988; Schreier et al. 2002). 118 Soil Organic Matter Soil organic matter (SOM) improves soil structure, water holding capacity, aeration, and aggregation. It is a source of macro and micro nutrients, especially for carbon. Its high specific surface and C E C lead to sorption of elements and pollutants. S O M can be divided into humic substances (HS) and non humic substances. HS is defined as "a general category of naturally occurring, biogenic, heterogeneous organic substances that can generally be characterized as being yellow to black in color, and have high molecular weights". HS can be divided into humin, humic acid (HA) and fulvic acid (FA) according to their size (humin > H A > F A ) . The negative charge of HS is primarily derived from ionization of acidic functional groups, e.g., carboxyls, phenolic O H , alcoholic O H , ether. The surface areas and C E C of S O M are higher than those of clay minerals. The point zero charge (pzc), defined as the p H at which the colloidal particle has no net charge, is low, about 3, S O M is negatively charged at p H values greater than 3. A s p H increases, the degree of negative charge increases due to the deprotonation or dissociation of H from functional groups. S O M forms the complexation of metal ions affecting the retention and mobility of metal contaminants in soils and waters. If two or more organic functional groups (e.g., carboxylate) coordinate the metal ion, forming an internal ring structure, chelation, a form of complexation, occurs. The major complexing sites are carboxyl and phenolic groups. Formed complexes are stable and enhance transport of toxic metals in waters. S O M complexes can also occur with clay minerals that are coated with metal oxides such as A l and Fe oxides (Sparks 1995). Sorption Sorption is at the interface between the solid surface and the solution (Adsorption is one of them). Adsorption determines the quantity of plant nutrients, metals, pesticides, and other organic chemicals that are retained on solid surfaces (Sparks 1995). Sorption occurs by next five mechanisms (Minton 2005); • Sorption to Fe, A l , and M n oxides • Sorption to exchangeable sites of clays and organic matter • Chelation and complexation in organic matter • Precipitation with carbonates and sulfides 119 • Uptake by plants, biofilms, and soil organisms Surface functional groups are chemically reactive molecular unit bound into the structure of a solid at its periphery. Organic sites are carboxyl, carbonyl, phenolic etc., whereas inorganic sites have siloxane surface associated with the plane of oxygen atoms bound to the silica tetrahedral layer, at the edges of inorganic minerals. The surface functional groups can be protonated or deprotonated by adsorption of H and O H , as below, S-OH + H o S - O H 2 S-OH <=> S-0 + H Surface complexing occurs when the interactions of surface functional group with an ion creates a stable molecular entity. Physical forces work as van der Waals forces forming electrostatic outer-sphere complexes (ion exchange), whereas chemical forces work by forming inner sphere complexation (i.e., a ligand exchange mechanism, covalent bonding, hydrogen bonding). In outersphere complexes, water is present between surface functional group and the ion. Complexes through electrostatic and coulombic interactions are weak. In inner sphere complexes, metals are directly bound to oxygen. This is due to covalent or ionic bonding, and metals are absorbed within the structure, thus forming strong complexes (Sparks 1995). In acidic soils, aluminum, iron and manganese form oxide complexes on the soil particles. Other metals e.g., Zn , N i and C d sorb to these complexes. Complete sorption occurs above a p H 6 for Pb and Cu , but 7 for Z n and Cd . Generally metals prefer to sorb to organics (Minton 2005). It should be noted that soft acids tend to form covalent bonding, and thus can be adsorbed in low p H (Sparks 1995). Solubility Products Ions in the soil solution can form a number of species due to hydrolysis, complexation, and redox reactions. Solubility is the equilibrium relationship of ions with respect to solid and liquid phases, affected by p H (Sparks 1995). For a given solid, M x L y , a general dissolution reaction is: 120 M x L y ( s ) » H 2 ° xM y + (aq) + yL x"(aq) (1) where M y + (aq) and L x (aq) are the aqueous metal and ligand ions M and L , respectively. A n equilibrium constant for this reaction is defined as: K = [ M y + ] x [ L x - ] y / [ M x L y ] (2) where [ ] denote activities. A solubility product K s p (equilibrium constant) is then defined as: K s p = K / [ M x L y ] (3) If the solid M x L y is in its standard state then [ M x L y ] becomes unity and K s p becomes: K s p = [ M y + ] x [ L x " ] y (4) For a fixed activity of L x ~ , the solid with the smallest numerical value of K s p w i l l support the smallest equilibrium activity of M y + . Values of - log K s p ( pK s p ) are shown for hydroxides, phosphates and sulfides in Table A-4 and fluorides, iodates and oxalates are in Table A - 5 . Hydroxides, phosphates, sulfides are the least soluble, and ligands of soft bases (I, F etc.) and oxalate are more soluble (Sparks 1995; Traina and Laperche 1999) 121 Table A . 4 Solubility product of hydroxides, phosphates and sulfides (Lindsay 1979). Hydrox ide P h o s p h a t e Su l f ide A l u m i n i u m A I ( O H ) 3 3 3 - 3 4 A I P 0 4 19 -21 C a l c i u m C a ( O H ) 2 6 C a 3 ( P 0 4 ) 2 2 6 - 3 3 C a d m i u m C d ( O H ) 2 1 4 - 1 5 C d 3 ( P 0 4 ) 2 33 C d S 2 7 - 2 8 Cobalt( l l ) C o ( O H ) 2 15 C o 3 ( P 0 4 ) 2 35 C o S 2 1 - 2 6 Cobalt( l l l ) C o ( O H ) 3 44 Chromium( l l ) C r ( O H ) 2 16 Chromium( l l l ) C r ( O H ) 3 31 Copper( l ) C u 2 S 48 Copper( l l ) C u ( O H ) 2 20 C u 3 ( P 0 4 ) 2 37 C u S 37 Iron(ll) F e ( O H ) 2 1 6 - 1 7 F e S 19 Iron(lll) F e ( O H ) 3 3 8 - 3 9 F e P 0 4 22 M a g n e s i u m M g ( O H ) 2 1 1 - 1 2 M g 3 ( P 0 4 ) 2 2 4 - 2 5 Manganese ( l l ) M n ( O H ) 2 9 - 1 3 M n S 1 1 - 1 4 Nickel(ll) N i ( O H ) 2 1 5 - 1 6 N i 3 ( P 0 4 ) 2 32 N i S 1 9 - 2 5 Lead(l l ) P b ( O H ) 2 5 - 2 0 P b S 2 8 - 2 9 Z i n c Z n ( O H ) 2 17 Z n 3 ( P 0 4 ) 2 33 Z n S 2 3 - 2 5 Table A . 5 Solubility product of fluorides, iodates and oxalates (Lindsay 1979). Fluor ide lodate O x a l a t e C a l c i u m C a F 2 9-11 C a ( l 0 3 ) 2 6 - 7 C a C 2 0 4 9 C a d m i u m C d F 2 3 Cd(IQ 3) 2 8 C d C 2 0 4 8 Copper( l l ) C u C 2 0 4 10 Iron(ll) F e F 2 6 M a g n e s i u m M g F 2 8-11 M g C 2 0 4 5 - 7 Manganese( l l ) 8 M n ( l 0 3 ) 2 7 Nickel(ll) 2 N i ( l 0 3 ) 2 5 Lead(ll) P b F 2 P b ( l 0 3 ) 2 13 P b C 2 0 4 9 Z i n c Z n F Z n ( l 0 3 ) 2 6 Z n C 2 0 4 8 The toxicity of metal M is directly proportional to the activity of the free metal ion; M y + is the most toxic form of M . Therefore, the least toxic solid form of M w i l l have the smallest aqueous equilibrium activity of M y + . Analogously, the most toxic solid w i l l be that which supports the largest aqueous equilibrium activity of M y + . The solubility product serves as a relative constraint on the reactivity and potential toxicity of metals (Traina and Laperche 1999). 122 Stability Constants The Stability Constant, K s t ab, is the equilibrium constant for the equilibrium that exists between a transition metal ion surrounded by water molecule ligands and the same transition metal ion surrounded by ligands of another kind in a ligand displacement reaction. It is common to write an overall expression for the overall ligand displacement reaction. Given the ammonia solution to a solution containing hexaaquacopper(II) ions, overall equilibrium reaction is, [ C u ( H 2 0 ) 6 ] 2 + ( a q ) + 4 N H 3 ( a q ) o [ C u ( N H 3 ) 4 ( H 2 0 ) 2 ] 2 + ( a q ) + 4 H 2 0 ( I ) A n d equilibrium constant is, [[Cu(NH 3) 4(H 20) 2] 2 a + q ) ] K stab . 2+ , r i i [[Cu(H 20) 6] ( a q ) ] [ N H 3 ( a p ) ] The value is often expressed as log K s t ab. The larger the value of log K s t ab, the more stabilising are the ligands of the transition metal ion (Clark 2007). Log K s t ab o f some organic-metal compounds are shown in Table A - 6 . It should be noted that E D T A compounds have higher values as they have more functional groups, whereas the values of oxalic acid compounds are low as they have fewer functional groups. Fulvic acids also have fewer functional groups and thus metal compounds with these organics are relatively soluble. Table A. 6 Stability constants of organic-metal compounds (Furia 1972; Smith et al. 2001; George Eby Research 2006). Organic ligands Al(lll) Ca Co(ll) Cu Fe(ll) Fe(lll) Mg Mn Ni Zn Pb Acetic acid C H 3 C O O H - 0.5 2.2 - - : 0.5 - 0.7 1.0 Citric acid CeH807 -11.7 3.5 4.4 6.1 3.2 11.9 2.8 3.2 4.8 4.5 4.1 EDTA C , 0 H 1 6 N 2 O 8 16.1 10.7 16.2 18.8 14.3 25.7 8.7 13.6 18.6 16.5 18.0 NTA C 6 H 9 N 0 6 >10 6.4 10.6 12.7 8.8 15.9 5.4 7.4 11.3 10.5 Oxalic acid H2C2O4 7.3 3.0 4.7 6.3 >4.7 9.4 2.6 3.9 5.2 4.9 Salicylic acid C 6 H,(OH)C0 2 H 14.1 - 6.7 10.6 6.6 16.4 4.7 2.7 7.0 6.9 123 B. Methodology for sediment quality analysis Table A . 7 Lowest readable limits of I C P - A E S and average blank concentrations in this study. Element Lowest Readable Limits Average Blanks in this study Co 0.1 -0.0041 Cu 0.1 0.0210 Fe 0.1 0.3886 Mn 0.01 0.0038 . Ni 0.1 0.0007 Pb 0.25 -0.0018 Zn 0.025 0.0039 Cd 0.1 -0.0042 Cr 0.05 0.0008 Al 0.2 0.3554 Ca 0.1 0.1289 K 0.5 -0.0628 Mg 0.05 0.0768 Na 1 0.1544 P 0.25 0.0181 * Source: Tyler (1995). * All the units are mg/kg. 124 Table A . 8 Comparison of the and <63 um <177um sediment particle fractions of the Brunette Watershed sediments. Adapted from McCa l lum (1995). S a m p l e ID F e % Fine C o a r s e C u mg/kg F ine C o a r s e P b mg/kg F ine C o a r s e Z n mg/kg F ine C o a r s e R 4 2.8 2.4 166 104 286 214 366 258 105 4.5 3.3 67 37 72 42 176 116 119 2.9 2.1 145 58 188 66 266 110 120 3 2.3 129 61 162 77 296 153 126 3.4 3.4 99 57 126 78 284 178 128 4.8 4.8 133 85 154 98 197 134 S a m p l e ID C d mg/kg F ine C o a r s e Ni mg/kg F ine C o a r s e M n mg/kg F ine C o a r s e C r mg/kg F ine C o a r s e R 4 2.5 2 33 27 335 275 53 47 105 1 0.5 20 13 1555 1045 36 22 119 1 0.5 22 12 535 245 47 26 120 1.5 0.5 22 12 2250 1105 32 21 126 2.5 1.5 26 19 1715 1070 38 37 128 1 0.5 21 16 1345 860 35 38 * S a m p l e s are d igested with a q u a regia a n d a n a l y s e d by I C P - A E S . * F ine: < 63 p m . C o a r s e : <177pm. 125 Table A . 9 Median concentrations differences and significance of ranked t-test for comparison of aqua regia versus nitric/perchloric acid digestions (n=14), by using the Brunette Watershed sediments. Adapted from McCa l lum (1995). Element Med ian percent differece, Nitric A c i d minus A q u a R e g i a p va lue F e -15 0.00 M g 8 0.02 M n -3 0.06 P b 8 0.01 C u 5 0.18 Z n 2 0.02 Ni -6 0.09 * p va lues indicate the probability that the m e a n s are identical. * F l a m e A A detection technique w a s u s e d . Table A . 10 Median concentrations differnces and significance of ranked t-test for comparison of atomic absorption (flame A A ) and I C P - A E S detection techniques (n=12), by using the Brunette Watershed sediments. Adapted from McCa l lum (1995). M e d i a n percent difference, ICP minus < Element f lame A A p va lue M n -4 0.12 P b -55 0 C u -59 0 Z n 0 0.48 Ni -19 0 C r -43 0 * p va lues indicate the probability that the m e a n s are identical. 126 Table A . 11 Sediment quality guidelines in Canada and the United States. Element C a n a d a U S I S Q G P E L T E L P E L C u 35.7 197 28.0 101.2 P b 35 91.3 37.2 81.7 Z n 123 315 98.1 544.0 C d 0.6 3.5 0.6 3.2 C r 37.3 90 36.3 119.4 F e N /A N /A 18.84 24.76 M n N /A N /A 631.3 1184.8 Ni N /A N /A 19.5 32.8 Al N /A N /A 2.55 5.96 A s N /A N /A 10.8 48.4 * S o u r c e : U S G S (1999) a n d C a n a d i a n C o u n c i l of Ministers of the Env i ronment (2002) * All units are mg/kg (dry weight), except for F e and A l (%). T h e I S Q G s (Interim freshwater sed iment quality guidel ines) in C a n a d a are the equivalent of T E L s (threshold effect levels) in the United States. P E L s - probable effect levels. 127 C. Aerial Images Figure A . 1 The Watersheds h i -2 (1973-above; 2006-below). 128 Figure A . 2 The Watersheds h3-7 (1973-above; 2006-below). 129 ure A . 3 The Watershed k l (1973-above; 2006-below). 130 Figure A . 4 The Watersheds k2, k4-6 (1973-above; 2006-below). 131 132 Figure A. 6 The Watershed tl (1973-above; 2006-below). 133 134 Figure A . 8 The Drainages v3-4 in 2006. 135 Figure A . 9 The Drainages v5-6 in 2006. 136 D. Land Use Watershed boundary Sampling site Stream 100 m buffer 1 Meters 0 375 750 1,500 N A Land use 1973 Forest Grass-agriculture Rural-urban Dense-urban Others Figure A . 11 The watersheds h3-6 in 1973. 138 Watershed boundary Sampling site Stream 0 375 750 IMeters 1,500 N A I I 100 m buffer L a n d u s e 2006 Forest Grass-agriculture Rural-urban Dense-urban Others Figure A . 12 The watersheds h3-7 in 2006. 139 Watershed boundary Stream • Sampling site I 1100m buffer Land use 1973 Forest Watershed boundary Stream • Sampling site I 1100m buffer Land use 2006 Forest Figure A . 13 The watershed k l (1973-above; 2006-below). 140 Figure A . 14 The watersheds k2, k4-6 in 1973. 141 Figure A. 15 The watersheds k2, k4-6 in 2006. 142 Watershed boundary Stream • Sampling site _ 100m buffer Land use 1973 Forest Grass-agriculture Rural-urban H Dense-urban Commercial-industrial Others IMeters 1,000 Watershed boundary Stream • Sampling site I I 100m buffer Land use 2006 Forest Grass-agriculture Rural-urban | Dense-urban Commercial-industrial Others IMeters 1,000 Figure A . 16 The watershed k3 (1973-above; 2006-below). 143 Watershed boundary Sampling site Stream I I 100 m buffer Land use 1973 Forest Grass-agriculture Rural-urban I Dense-urban Commercial-industrial Others l Meters 375 750 1,500 Watershed boundary Sampling site Stream I I 100 m buffer Land use 2006 Forest Grass-agriculture Rural-urban M Dense-urban Commercial-industrial Others 1 Meters 375 750 1,500 Figure A . 17 The watershed t l (1973-above; 2006-below). 144 — — Vtetershed boundary # Sampling site Vtetershed boundary # Sampling site Figure A . 18 The watersheds t2-7 (1973-above; 2006-below). 145 Watershed boundary • Sampling site Storm sewer Land use 2006 Forest Grass-agriculture Figure A. 19 The drainages v3-4 in 2006. 146 Watershed boundary • Sampling site Storm sewer Land use 2006 N A I I I I I I Meters 0 125 250 500 Figure A . 20 The drainage v5 in 2006. 147 Table A . 12 Land use proportion of the each of the watersheds in 1973 (%). W a t e r s h e d Forest G r a s s - Rura l -urban D e n s e - C o m m e r c i a l -Others Urban=(1)+ agriculture (1) urban (2) industrial (3) (2)+(3) a1 12.1 27.5 11.7 40.3 8.3 0.1 60.4 a2 14.7 8.3 9.3 47.4 20.2 0.1 76.9 a 3 18.4 7.2 3.0 64.4 7.1 0.0 74.5 ' a4 64 .3 16.2 19.3 0.1 0.0 0.1 19.4 h i 98.5 1.1 0.0 0.0 0.0 0.4 0.0 h2 100.0 0.0 0.0 0.0 0.0 0.0 0.0 h3 64.2 22.0 11.1 0.0 0.0 2.7 11.1 h4 53.0 34.7 11.1 0.0 0.0 1.2 11.1 h5 53.7 35.2 9.9 0.0 0.0 1.2 9.9 h6 54.8 34.2 9.7 0.0 0.0 1.3 9.7 k1 78.1 12.4 5.4 0.0 0.7 3.5 6.0 k2 52.7 29.6 15.8 0.0 0.0 1.9 15.8 k3 99.0 0.2 0.8 0.0 0.0 0.0 0.8 k4 52.9 16.1 18.8 12.2 0.0 0.0 31.0 k5 51.9 18.8 21.8 7.4 0.0 0.1 29.2 k6 52.5 20.3 20 .7 5.8 0.0 0.6 26.6 t1 89.5 0.0 0.1 0.0 0.0 10.4 0.1 t2 79.9 9.5 9.5 0.0 0.0 1.1 9.5 t3 86.1 10.6 1.3 0.0 0.0 2.0 1.3 t4 78.8 15.5 2.8 0.0 0.0 2.9 2.8 t5 66.2 22.2 9.5 0.0 0.0 2.1 9.5 t6 78.6 11.5 6.5 0.0 0.0 3.4 6.5 t7 70.9 21.1 6.1 0.0 0.0 1.9 6.1 Table A . 13 Land use proportion of the each of the watersheds in 2006 (%). W a t e r s h e d Forest G r a s s -agriculture Rura l -urban (1) D e n s e -urban (2) C o m m e r c i a l -industrial (3) Others Urban=(1)+(2) +(3) a1 8.3 6.8 5.5 41 .9 36.9 0.5 84.4 a2 7.8 5.6 6.4 17.9 61.8 0.5 86.1 a 3 10.1 1.9 5.2 63.8 18.5 0.5 87.5 a4 42.4 25.1 25.4 2.3 4.2 0.6 31.9 h i 95.4 3.6 0.5 0.0 0.0 0.4 0.5 h2 98.8 0.0 0.0 0.0 0.0 1.2 0.0 h3 41.2 23.5 35.1 0.0 0.0 0.2 35.1 h4 29.2 34.1 25 .7 5.2 5.0 0.9 35.9 h5 38.1 32.0 24.0 2.7 2.6 0.7 29.3 h6 39.6 32.5 22.8 2.3 2.2 0.5 27.3 h7 39.6 32.2 22 .0 3.4 2.2 0.5 27.7 k1 52.4 23.4 18.1 0.0 3.1 3.0 21.2 k2 45.4 20.8 29 .0 1.1 0.7 3.0 30.8 k3 86.2 0.9 12.9 0.0 0.0 0.0 12.9 k4 38.7 16.1 21 .3 24.0 0.0 0.0 45.3 k5 39.7 17.3 22 .7 17.7 2.5 0.2 42.8 k6 43 .3 17.4 23.1 13.9 2.0 0.4 38.9 t1 92.8 4.9 2.1 0.0 0.0 0.2 2.1 t2 65.0 9.4 24.7 0.0 0.0 0.9 24.7 t3 52.9 17.2 28.8 0.0 0.0 1.1 28.8 t4 49 .5 24.0 25.7 0.0 0.0 0.8 25.7 t5 46 .3 25.3 27 .9 0.0 0.0 0.6 27.9 t6 61.2 15.0 21.1 0.0 0.0 2.7 21.1 t7 61.7 18.6 18.2 0.0 0.0 1.5 18.2 • v 3 8.1 1.7 0.0 90.6 0.0 0.0 90.6 v 4 7.8 , 0.0 0.0 47.1 44.5 0.6 91.6 v 5 1.7 0.0 0.0 93.9 4.5 0.0 98.3 v 6 0.0 0.0 0.0 0.0 100.0 0.0 100.0 E. Land cover Watershed boundary Stream • Sampling site Figure A . 21 The watershed k l in 2006. Only land cover within urban land use (rural-urban, dense-urban, commercial-industrial) is delineated. 150 I I I I I I Meters 0 250 500 1,000 Figure A. 22 The watershed k2 in 2006. Only land cover within urban land use (rural-urban, dense-urban, commercial-industrial) is delineated. 151 Figure A . 23 The drainage v5 in 2006. Only impervious-surfaces are delineated. 152 Figure A . 24 The drainages v5-6 in 2006. Only impervious-surfaces are delineated. 153 Table A . 14 Land cover proportion of the each of the watersheds in 1973 (%). W a t e r s h e d Forest -c o v e r G r a s s -c o v e r Impervious-sur face O t h e r s a1 25.9 42 .0 31.9 0.1 a2 24.1 31.5 44 .6 -0.2 a 3 28.0 32.0 40 .5 -0.5 a4 68.2 24.8 6.5 0.5 h i 93.0 5.0 1.3 0.7 h2 94.2 4.3 1.3 0.2 h3 68.2 24.3 4.5 3.0 h4 60 .9 32.7 4.7 1.7 h5 ' 61.4 32.4 4.4 1.7 h6 62.2 31.7 4.3 1.8 k1 77.7 15.1 3.5 3.7 k2 60.1 31.7 5.8 2.3 k3 93.4 4.8 1.5 0.3 k4 58.9 28.4 12.4 0.2 k5 58.6 30.2 10.9 0.4 k6 59.2 30.1 9.8 0.9 t1 84.3 3.9 1.5 10.3 t2 79.4 15.3 3.9 1.4 t3 84.1 11.8 1.8 2.3 t4 78.7 15.8 2.3 3.2 t5 69.8 23.6 4.1 2.5 te 78.1 15.0 3.2 3.7 t7 73.3 21.2 3.2 2.3 Table A . 15 Land cover proportion of the each of the watersheds in 2006 (%). W a t e r s h e d Forest -c o v e r G r a s s -c o v e r Impervious-sur face Others a1 12.8 33.2 53.9 0.0 a2 12.5 14.6 73.0 0.0 a3 21.1 28.7 50.2 0.0 a4 52.1 36.2 11.7 0.0 h i 90.9 7.0 1.5 0.7 h2 93.0 4.3 1.3 1.4 h3 51.2 37.6 10.6 0.6 h4 41.7 41.8 15.2 1.3 h5 48.9 38.6 11.4 1.1 h6 50.2 38.2 10.6 1.0 h7 50.1 38.0 11.0 1.0 k1 61.2 30.3 8.6 0.0 k2 57.1 35.2 7.7 0.0 k3 83.7 11.5 4.6 0.2 k4 47.4 33.5 19.0 0.1 k5 48.2 33.1 18.3 0.4 k6 51.2 32.1 16.1 0.6 t1 89.0 8.6 1.9 0.5 t2 68.1 23.0 7.7 1.1 t3 59.4 30.3 8.9 1.4 t4 57.5 33.2 8.2 1.2 t5 55.1 35.1 8.8 1.0 t6 65.3 24.8 6.9 3.0 t7 66.2 25.7 6.2 1.8 v 3 22.1 39.5 38.5 0.0 v 4 14.9 23.8 61.3 0.0 v 5 25.7 25.9 48.4 0.0 v 6 0.0 12.9 87.1 0.0 155 F. A 100 m Buffer zone Table A . 16 Proportion of land use in a 100 m buffer zone in 1973 (%). Watershed Forest Grass-agriculture Rural-urban (D Dense-urban (2) Commercial-industrial (3) Others Urban=(1)+ (2)+(3) a1 42.4 43.9 13.1 0.0 0.7 0.0 13.8 a2 49.8 3.7 14.0 32.4 0.0 0.2 46.4 a3 41.3 8.4 8.6 41.8 0.0 -0.1 50.4 a4 66.5 13.2 20.4 0.0 0.0 0.0 20.4 hi 97.7 1.8 0.0 0.0 0.0 0.5 0.0 h2 100.0 0.0 0.0 0.0 0.0 0.0 0.0 h3 61.3 35.4 2.6 0.0 0.0 0.7 2.6 h4 39.3 54.4 6.1 0.0 0.0 0.2 6.1 h5 48.8 46.4 4.3 0.0 0.0 0.4 4.3 h6 56.1 39.7 3.6 0.0 0.0 0.6 3.6 k1 77.2 12.2 6.1 0.0 0.0 4.4. 6.1 k2 69.0 12.0 14.7 0.0 0.0 4.3 14.7 k3 96.7 0.4 2.9 0.0 0.0 0.1 2.9 k4 56.3 25.2 13.5 5.0 0.0 0.0 18.5 k5 62.5 19.8 15.7 1.9 0.0 0.1 17.6 k6 66.5 18.6 13.5 1.3 0.0 0.1 14.8 t1 90.3 0.0 0.3 0.0 0.0 9.4 0.3 t2 68.1 17.6 14.2 0.0 0.0 0.0 14.2 t3 84.3 15.7 0.0 0.0 0.0 0.0 0.0 t4 82.1 10.6 2.2 0.0 0.0 5.1 2.2 t5 76.1 12.5 7.6 0.0 0.0 3.8 7.6 t6 73.3 7.6 9.5 0.0 0.0 9.5 9.5 t7 78.6 9.8 5.8 0.0 0.0 5.8 5.8 156 Table A . 17 Proportion of land use in a 100 m buffer zone in 2006 (%). Watershed Forest Grass-agriculture Rural-urban (D Dense-urban (2) Commercial-industrial (3) Others Urban=(1)+(2) +(3) a1 31.4 20.2 4.0 29.4 14.7 0.3 48.1 a2 28.9 4.0 9.0 43.9 13.9 0.3 66.8 a3 26.1 0.0 6.7 66.8 0.5 0.0 73.9 a4 48.9 18.4 17.8 6.0 8.8 0.0 32.7 hi 92.8 5.9 0.9 0.0 0.0 0.5 0.9 h2 98.6 0.0 0.0 0.0 0.0 1.4 0.0 h3 30.0 45.5 24.1 0.0 0.0 0.4 24.1 h4 18.0 59.7 19.9 0.0 2.3 0.1 22.2 h5 38.1 42.7 17.7 0.0 0.8 0.7 18.5 h6 47.2 37.5 14.1 0.0 0.6 0.5 14.8 h7 48.7 36.1 13.8 0.3 0.6 0.5 14.7 k1 59.7 15.8 24.0 0.0 0.0 0.5 24.0 k2 47.1 20.5 32.3 0.0 0.0 0.1 32.3 k3 65.2 0.0 34.8 0.0 0.0 0.0 34.8 k4 50.1 10.1 19.8 20.0 0.0 0.0 39.8 k5 59.8 7.9 22.9 7.4 1.1 0.9 31.4 k6 66.7 6.9 19.7 5.2 0.8 0.7 25.7 t1 88.8 6.1 4.3 0.0 0.0 0.8 4.3 t2 42.2 31.5 26.3 0.0 0.0 0.0 26.3 t3 86.0 14.0 0.0 0.0 0.0 0.0 0.0 t4 68.1 21.7 10.2 0.0 0.0 0.0 10.2 t5 67.2 16.0 16.8 0.0 0.0 0.0 16.8 t6 55.3 8.6 27.1 0.0 0.0 9.0 27.1 t7 70.1 7.8 16.7 0.0 0.0 5.4 16.7 v3 8.1 1.7 0.0 90.6 0.0 0.0 90.6 v4 7.8 0.0 0.0 47.1 44.5 0.6 91.6 v5 1.7 0.0 0.0 93.9 4.5 0.0 98.3 v6 0.0 0.0 0.0 0.0 100.0 0.0 100.0 157 Table A . 18 Proportion of land cover in a 100 m buffer zone in 1973 (%). W a t e r s h e d Forest -c o v e r G r a s s -c o v e r Impervious-sur face Others a1 53.6 39.9 5.9 0.6 a2 54.4 24.3 21.2 0.1 a 3 47 .9 27.7 24.6 -0.2 a4 69 .7 23.4 6.7 0.3 h i 92.5 5.5 1.3 0.8 h2 94.1 4.3 1.3 0.3 h3 67.3 28.9 2.5 1.2 h4 52.2 43.1 3.7 1.0 h5 58.8 37.0 3.1 1.1 h6 63.8 32.2 2.8 1.2 k1 77.0 15.3 3.1 4.6 k2 70.7 19.5 5.3 4.5 k3 91.6 6.0 2.0 0.4 k4 62.6 29.4 7.6 0.4 k5 67.1 25.9 6.5 0.5 k6 70.1 23.8 5.7 0.5 t1 85.1 4.0 1.6 9.3 t2 71.3 23.1 5.2 0.4 t3 83.5 14.6 1.5 0.4 t4 80.5 12.2 2.1 5.3 t5 76.2 16.3 3.5 4.0 t6 72.7 13.8 4.0 9.5 t7 77.6 13.5 3.1 5.9 158 Table A . 19 Proportion of land cover in a 100 m buffer zone in 2006 (%). W a t e r s h e d Forest -c o v e r G r a s s -c o v e r Impervious-sur face O t h e r s a1 40.6 28.7 30.7 0.0 a2 37.2 23.2 39.6 0.0 a 3 35.9 27.8 36.3 0.0 a4 53.8 33.0 13.2 0.0 h i 89.0 8.6 1.6 0.8 h2 92.9 4.2 1.3 1.6 h3 44.3 46 .5 8.1 1.0 h4 36.1 53.8 9.2 0.9 h5 50.2 41.4 7.2 1.3 h6 56.7 36.2 6.0 1.1 h7 57.7 35.2 6.0 1.0 k1 64.0 26.6 9.3 0.0 k2 56.9 35.1 8.0 0.0 k3 67.6 22.1 10.1 0.2 k4 55.8 27.5 16.5 0.1 k5 63.4 23.6 11.9 1.1 k6 68.8 20.6 9.7 0.8 t1 85.9 10.5 2.5 1.1 t2 52.6 38.4 8.5 0.5 t3 84.6 13.5 1.5 0.3 t4 71.6 23.8 4.2 0.4 t5 70.4 23.4 5.8 0.4 t6 59.2 23.4 8.5 8.9 t7 71.0 17.7 5.8 5.5 v 3 22.1 39.5 38.5 0.0 v 4 14.9 23.8 61.3 0.0 v 5 25.7 25 .9 48.4 0.0 v 6 0.0 12.9 87.1 0.0 159 G. Road Density Table A . 20 Road Density in the sub-watersheds in 2006 (km/km 2). W a t e r s h e d R o a d density Road-dens i ty in buffer z o n e s a1 11.4 3.9 a2 14.6 11.4 a 3 14.6 9.4 a4 3.3 1.8 hi 0.1 0.0 h2 0.9 0.0 h3 3.5 0.9 h4 4.2 3.7 h5 3.5 2.3 h6 3.7 2.5 h7 3.8 2.7 k1 3.5 1.1 k2 2.7 4.0 k3 3.0 6.8 k4 5.8 6.2 k5 5.3 5.4 k6 4.6 4.4 t1 1.4 1.7 t2 2.7 3.9 t3 1.9 0.0 t4 2.2 1.6 t5 2.1 1.3 te 2.7 3.6 t7 2.6 2.7 v 3 22.3 22.3 v 4 30.8 30.8 v 5 30.5 30.5 v 6 31.8 31.8 160 H. Historical Changes in Land Use/cover Table A. 21 Historical Changes in Land Use/cover (%). Land use Land cover Watershed Forest Grass- Rural- Dense- Commercial-Others Urban=(1)+ Forest- Grass- Impervious-agriculture urban (1) urban (2) industrial (3) (2)+(3) cover cover surface a1 -3.7 -20.7 -6.2 1.6 28.6 0.4 24.0 -13.0 -8.8 22.0 a2 -6.9 -2.7 -3.0 -29.4 41.6 0.5 9.2 -11.7 -16.9 28.4 a3 -8.3 -5.3 2.1 -0.5 11.4 0.7 13.0 -6.9 -3.3 9.7 a4 -21.9 8.9 6.1 2.3 4.2 0.5 12.5 -16.1 11.4 5.2 hi -3.1 2.5 0.5 0.0 0.0 0.0 0.5 -2.1 1.9 0.2 h.2 -1.2 0.0 0.0 0.0 0.0 1.2 0.0 -1.1 -0.1 0.0 h3 -23.0 1.5 24.0 0.0 0.0 -2.5 24.0 -17.0 13.3 6.1 h4 -23.8 -0.7 14.6 5.2 5.0 -0.3 24.8 -19.1 9.1 10.5 h5 -15.6 -3.2 14.1 2.7 2.6 -0.6 19.4 -12.6 6.2 7.0 h6 -15.1 -1.7 13.1 2.3 2.2 -0.7 17.6 -12.0 6.5 6.3 k1 -25.7 11.1 12.7 0.0 2.5 -0.5 15.2 -16.5 15.2 5.1 k2 -7.3 -8.8 13.2 1.1 0.7 1.1 15.0 -3.1 3.5 1.9 k3 -12.8 0.8 12.1 0.0 0.0 -0.1 12.1 -9.7 6.7 3.1 k4 -14.3 0.0 2.5 11.8 0.0 0.0 14.3 -11.5 5.1 6.6 k5 -12.2 -1.5 0.9 10.2 2.5 0.2 13.6 -10.4 2.9 7.4 k6 -9.2 -2.9 2.4 8.0 2.0 -0.2 12.4 -8.0 2.1 6.2 t1 .. 3.3 4.9 2.0 0.0 0.0 -10.2. 2.0 4.8 4.7 0.4 t2 -14.9 -0.1 15.2 0.0 0.0 -0.2 15.2 -11.3 7.7 3.9 t3 -33.3 6.6 27.6 0.0 0.0 -0.9 27.6 -24.6 18.4 7.1 t4 -29.3 8.6 22.9 0.0 0.0 -2.1 22.9 -21.2 17.4 5.9 t5 -20.0 3.0 18.4 0.0 0.0 -1.5 18.4 -14.7 11.5 4.7 t6 -17.4 3.5 14.6 0.0 0.0 -0.7 14.6 -12.9 9.8 3.7 t7 -9.2 -2.5 12.1 0.0 0.0 -0.4 12.1 -7.1 4.5 3.0 I. Factor Analysis on Land Variables Table A . 22 Factor component loadings of promax rotation on land variables in 1973. F a c t o r l Factor2 Factor3 C o m m u n a l i t y C o a r s e - m i n e r a l -0.80 0.54 0.94 F ine-minera l 0.78 -0.55 0.93 S t r e a m gradient 0.15 S l o p e -0.75 0.69 E leva t ion -0.79 0.75 Forest -0.97 0.98 Grass-agr icu l ture -0.78 0.88 Rural ly -urban 0.61 -0.49 -0.50 0.86 Dense ly -u rban 0.72 0.62 0.96 Commerc ia l - indust r ia l 0.61 0.62 0.80 C a n o p y - c o v e r -0.96 0.98 G r a s s - c o v e r 0.92 0.95 1 m perv ious-sur face 0.80 0.56 0.98 Forest (B) -0.91 0.96 Grass-agr icu l ture (B) -0.74 0.97 Rura l ly -urban (B) 0.70 -0.62 0.89 D e n s e l y - u r b a n (B) 0.57 0.73 0.88 Commerc ia l - indust r ia l (B) 0.23 C a n o p y - c o v e r (B) -0.93 0.95 G r a s s - c o v e r (B) 0.79 0.95 Imperv ious-sur face (B) 0.73 0.62 0.92 Rotation Sums of Squared 6.4 6.1 5.1 Loadings Contr ibut ion% 30 29 25 C u m u l a t i v e contr ibut ion% 30 59 84 * L o a d i n g s with only greater than 0.5 (as an absolute value) are s h o w n . Bold m e a n s c o r r e l a t i o n s . 7 . (B) denotes the land u s e in buffer z o n e s (and s o forth for the further document ) . 162 Figure A . 25 Eigenvalues o f principal component factor analysis for land variables in 2006 and those of broken-stick model. Table A . 23 factor component loadings o f promax rotation on the data. Factorl Factor2 Factor3 Communality Coarse-mineral 0.96 0.94 Fine-mineral -0.97 0.95 Stream-gradient 0.37 Slope 0.63 0.80 Elevation -0.59 -0.52 0.55 0.91 Forest -0.87 0.94 Grass-agriculture 0.79 0.83 Rurally-urban -0.53 0.70 0.86 Densely-urban 0.70 0.50 Commercial-industrial 0.73 0.61 Canopy-cover -0.91 0.93 Grass-cover 0.87 0.81 Impervious-surface 0.93 0.93 Road-density 0.97 0.97 Forest (B) -0.88 0.95 Grass-agriculture (B) 0.85 0.84 Rurally-urban (B) 0.57 Densely-urban (B) 0.72 0.54 Commercial-industrial (B) 0.68 0.52 Canopy-cover (B) -0.93 0.95 Grass-cover (B) 0.92 0.86 Impervious-surface (B) 0.96 0.96 Road-density (B) 0.92 -0.16 0.16 0.89 Rotation Sums of Squared Loadings Contribution% Cumulative contribution% 10.2 5.0 3.2 44 22 14 44 66 80 * Only correlation s greater than 0.5 (as an absolute value) are shown, and bold means a correlations.7. 163 J. Analysis for Sediments and Water Samples Table A . 24 Results of chroma after loss on ignition for sediments and nitrate for surface water. N A represents that samples that samples were not collected. W a t e r s h e d 105 C (1-6) 500 C (1-6) N 0 3 ( m g / L ) a1 2 6 0.52 a2 2 6 0.64 a3 2 6 0.57 a4 2 6 0.26 h i 2 6 0.23 h2 2 6 0.18 h3 2 4 0.29 h4 3 4 0.25 h5 3 4 0.26 h6 3 6 0.50 h7 3 6 0.47 k1 2 6 0.22 k2 2 4 0.27 k3 4 6 0.50 k4 2 6 0.62 k5 2 6 0.39 k6 2 6 0.21 t1 4 8 0.10 t2 2 6 0.98 t3 2 6 0.45 t4 1 6 0.40 t5 2 6 0.36 te 2 6 N A t7 2 6 0.63 v1 2 4 0.03 v2 2 2 0.09 v3 2 4 0.26 v4 3 4 0.94 v5 2 6 0.78 v6 1 4 N A * C h r o m a is a consecut ive categorizat ion from 1 to 6; the bigger the n u m b e r is, the redder the s a m p l e is. 164 Table A . 25 Sediment quality data in 1973 . W a t e r s h e d C o C u F e M n Ni P b Z n a1 18.4 60.3 3.47 683.8 40 .5 518.1 175.4 a2 28.5 53.6 4 .17 902.6 49.1 104.8 245.6 a 3 25.0 46.9 4 .17 875.2 50.6 25 .7 161.4 a4 16.1 30.5 0.71 436.4 38.8 7.2 70.2 h i 4.7 12.6 2.26 307.7 13.8 10.2 44.4 h2 15.6 25.3 4 .37 615.4 28.9 8.2 71.1 h3 13.2 21.1 5.10 574.4 25.5 6.8 80.0 h4 10.1 16.9 2.12 290.9 7.3 11.1 35.1 h5 15.6 33.7 6.56 779.5 36.2 11.4 80.0 h6 13.2 21.1 4 .37 471.8 25.5 7.2 71.1 k1 9.6 20.3 1.63 203.6 22.7 6.5 42.1 k2 10.7 20.3 1.70 436.4 23.5 10.4 49.1 k3 5.7 13.5 1.98 261.8 6.3 12.8 35.1 k4 10.7 27.1 1.63 349.1 29.7 15.4 70.2 k5 20.1 30.5 2.12 610 .9 39.7 16.8 98.2 k6 12.9 23.7 1.91 349.1 28.1 13.5 70.2 t1 19.3 25.3 5.83 615.4 33.6 2.7 71.1 t2 8.2 16.3 2.86 283.5 2.5 6.6 44.4 t3 7.0 16.3 2.86 243.0 11.6 7.4 44.4 t4 7.8 8.2 2.14 364.6 13.2 6.6 35.6 t5 9.5 16.3 2.14 162.0 26.1 7.8 44.4 t6 9.0 20.4 2.86 243.0 19.2 143.0 62.2 t7 9.9 12.2 2.86 222.8 15.4 12.4 35.6 * S o u r c e : Westwate r R e s e a r c h Cent re (1973). * Units are mg /kg except for F e (%). Table A . 26 Sediment quality data in 2006 in this study. W a t e r s h e d C o C u F e M n Ni P b Z n C d C r a1 25.4 70.0 4.52 1097.8 50.5 32.1 297.4 1.1 53.5 a2 33.5 66.9 5.93 1320.7 60.5 47.1 369.0 0.5 61.4 a 3 1 24.6 55.2 4 .93 1385.6 54.3 20 .0 240 .5 0.4 47 .9 a4 22.4 39.6 3.98 784.8 44.8 8.5 245.6 0.0 47.1 h i 7.5 22.8 3.61 372.4 17.9 8.7 59.0 0.3 25.7 h2 10.3 34.6 2.67 474 .5 15.7 5.8 68.9 0.0 22.7 h3 12.5 17.4 2.98 615 .9 24.8 17.2 78.0 0.3 58.3 h4 18.4 28.7 3.64 1882.4 32.7 16.1 186.9 0.2 141.0 h5 16.0 23.9 3.69 2194.2 30.2 12.2 133.6 0.0 38.9 h6 16.0 35.2 3.62 1817.6 30.4 25.6 121.2 0.0 37.2 h7 15.2 25.5 3.71 1261.8 31.3 26.0 105.3 0.3 44.6 k1 15.8 35.6 3.17 806.7 31.6 10.4 112.4 0.3 36.9 k2 19.2 30.3 3.10 1092.1 36.0 5.9 87.7 0.0 38.2 k3 11.7 42.8 4.42 964.3 12.7 . 17.1 92.5 0.1 24.7 k4 21.7 37.2 3.91 974.8 44.8 8.0 135.9 0.0 47.8 k5 17.3 33.0 3.43 953.4 36.0 10.3 130.7 0.0 39.7 k6 20.5 32.6 3.58 782.7 40.8 10.6 123.8 0.0 44 .9 t1 9.6 34.3 5.78 235.3 8.3 9.6 56.3 0.0 15.4 t2 13.2 44 .3 3.35 709.0 32.9 18.8 102.2 0.0 43.4 t3 17.4 30.5 3.51 908.6 35.8 8.3 110.4 0.2 41.2 t4 14.1 22.7 2.76 383.4 34.2 6.2 81.5 , 0.1 39.0 t5 14.4 29.2 2.98 953.7 30.3 9.4 114.1 0.0 36.6 t6 16.5 29.2 3.53 848.4 31.7 11.2 133.7 0.1 37.0 t7 18.0 32.6 3.52 857.3 42.2 7.9 95.0 0.0 42.4 v1 7.3 183.7 1.91 173.5 27.2 78.6 277.2 1.1 40.4 v 2 11.1 358.0 2.67 242.4 56.1 337.0 520.4 1.3 57.1 v 3 10.7 503.8 3.02 258.0 37.0 341.2 904.8 1.9 72.4 v 4 13.0 621.5 4.95 498 .3 45.4 487.8 1612.1 2.4 96.3 v 5 18.5 300.7 4.54 802.7 47.2 353.7 813.1 3.0 88.5 v 6 10.8 540.7 3.29 328.9 39.5 132.7 712.4 2.3 142.1 * Units are mg /kg except for F e (%). 166 Table A.23 Sediment quality data in 2006 in this study (Continued). W a t e r s h e d A l C a K M g N a P O M a1 2.89 5294.1 2191.2 10848.8 652.7 1093.9 1.2 a2 3.72 6467 .7 3275.5 15449.6 747.6 1115.8 1.1 a 3 3.15 5551.3 2815.3 13020.1 709.9 878.0 1.0 a4 2.87 4862 .7 2159.5 10508.9 534.9 997.8 1.1 h i 1.81 8590.6 1119.6 4089.2 494.0 935.1 0.5 h2 2.63 9340.6 877.1 5042.4 815.7 671 .5 0.6 h3 1.99 6272.0 535.0 4340.1 478.0 936.7 0.9 h4 2.50 5242.9 801.9 5803.0 711.6 1008.5 0.6 h5 2.00 6279.6 823.2 5515.9 813.0 1077.7 0.7 h6 2.04 6027 .3 876.2 5445.5 1040.3 1028.4 0.5 h7 2.09 4301.2 687.6 5018.9 555.2 1131.5 2.0 k1 2.84 4673 .5 1375.8 7939.7 591.1 688 .5 1.6 k2 1.94 4938 .5 1608.1 8137.4 534.4 777.9 0.8 k3 2.19 4586 .3 780.1 4135.6 1026.9 1022.5 0.6 k4 2.87 4419.0 1922.9 10356.2 612.4 799.1 1.0 k5 2.43 4421 .5 1495.4 7928.7 543.6 854.2 1.2 k6 2.75 4420.2 1724.0 9299.1 543.7 .978 .0 1.2 t1 3.35 2740 .3 303.1 2431.1 555.4 543.1 3.6 t2 2.57 3527.6 1052.4 6277.9 581.0 985.2 2.0 t3 2.69 4076 .7 1379.8 8060.5 668.6 774.7 1.7 t4 2.40 3915.6 1086.2 7224.4 562.0 696.1 1.8 t5 2.60 4230.2 1081.6 6535.0 630 .7 777.9 1.8 t6 2.62 3330.9 1107.7 6679.2 670.1 848.4 2.4 t7 2.18 4769.4 1517.5 8688.7 620.2 773.0 1.1 v1 0.92 23358.0 5235.2 14610.6 84173.3 1402.9 3.3 v 2 1.93 6339.3 3076.4 10968.0 40866 .7 1367.2 3.9 v 3 1.30 26642.4 748.6 5390.1 782.7 1195.0 2.6 v 4 1.46 25874 .5 954.7 5281.7 982.4 1559.5 2.6 v 5 2.04 12152.0 889.5 4527.8 971.9 2093.0 0.9 v 6 1.09 5952.8 791.2 4826.2 881.7 1027.1 2.1 * Units are mg/kg except for A l (%). 167 Table A. 27 Historical changes in sediment quality from 1973 to 2006 (% change from 1973). Watershed Co Cu Fe Mn Ni Pb Zn a1 38 16 30 61 25 -94 70 a2 18 25 42 46 23 -55 50 a3 -2 18 18 58 7 -22 49 a4 40 30 463 80 16 19 250 hi 59 80 60 21 30 -15 33 h2 -34 37 -39 -23 -46 -30 -3 h3 -6 -17 -42 7 -3 152 -2 h4 81 69 71 547 347 45 433 h5 2 -29 -44 181 -17 7 67 h6 21 67 -17 285 19 254 70 k1 65 75 95 296 39 62 167 k2 79 49 82 150 53 -43 78 k3 104 216 123 268 101 34 164 k4 103 37 140 179 51 -48 94 k5 -14 8 62 56 -9 -39 33 k6 58 38 87 124 45 -21 76 t1 -50 36 -1 -62 -75 254 -21 t2 60 172 17 150 1241 183 130 t3 149 87 23 274 210 12 148 t4 81 179 29 5 158 -7 129 t5 52 79 39 489 16 21 157 t6 83 43 23 249 66 -92 115 t7 82 167 23 285 175 -36 167 168 K. Factor Analysis on Sediment Quality Data 8 T 0 "I 1 r— —r— —, , PC1 PC2 PC 3 PC4 PC5 PC6 Figure A. 26 Eigenvalues of principal component analysis for sediment quality data in 2006 and those of broken-stick model. Table A. 28 factor component loadings of promax rotation on the data. F a c t o r l Factor2 Factor3 C o m m u n a l i t y C o 0.86 0.96 C u 0.92 0.92 F e 0.68 0.51 M n 0.67 0.50 Ni 0.64 0.84 P b 0.93 0.90 Z n 0.97 0.93 C d 0.94 0.90 C r 0.70 0.54 Al 0.66 0.70 C a 0.69 0.68 K 0.96 0.96 M g 0.97 0.97 N a -0.70 0.62 0.88 P 0.79 0.67 O M -0.58 0.47 Rotation Sums of tZ 7 i e Squared Loadings O.t o.o o.U Cont r ibu t ion% 36 22 19 Cumulative contribution% 36 58 77 * Only correlation s greater than 0.5 (as an absolute value) are s h o w n , a n d bold m e a n s a correlation>0.7. P - phosphorus , a n d O M - organic matter. 169 L. Cluster Analysis on Sediment Quality Data Dendrogram u s i n g Comple te L i n k a g e R e s c a l e d D i s t a n c e C l u s t e r Combine C A S E 0 5 10 15 20 L a b e l Num + h h h +-25 k5 16 t3 20 k2 13 t 7 24 k l 12 t6 23 t2 19 t 5 22 t4 21 h5 9 h6 10 h7 11 h4 8 h3 7 k3 14 h i 5 h2 6 t l k 18 v3 25 v6 28 v4 26 v5 27 a4 4 k4 15 k6 17 a l 1 a3 3 a2 2 J J Figure A . 27 Cluster analysis by using factor scores (factor 1, 2 and 3) of 2006 sediment quality data with the Complete Linkage method. 170 M . Mann-Whitney U test among the clustered sub-watersheds Figure A . 28 Mann-Whitney U test for each variable among the clustered sub-watersheds in 2006. F W : forested watersheds (hl-h3, t l , k3), R W : rural watersheds (h.4, k l - k 2 , k5, t2-t3, t6), U W : urban watersheds (al-a4, k4), V D : Vancouver drainages (v3-v6). The median and range for each cluster are presented. 171 4-1-• F W •RW |uw Fe g f o c o t (d) Hydroxide forming elements Al Mn Cluster (smaller values) FW RW UW VD V w \ g R 5, w & Mn \ Mn Al (0 r u 0) v v Mn FeAl Mn Al o VD "Bi E +1 • F W | U W •RW BVD 61 Ca Mg (e) Basic cations o ^ , <N E i i Na Cluster (smaller values) FW RW UW VD F •~- W in , v <D g R S W O) Mfif K Mg K (0 r U a w in " MfifK Mg K Mg K O VD NaCa Na Ca o Win r E _ • FW •RW |UW lVD (j) Chemical ligands Cluster (smaller values) OM FW RW UW VD « w <u \ § R S W cn \ ra r u a w .... ... - -H VD P P p P denotes phosphorus. ure A . 28 Mann-Whitney U test for each variable among the clustered sub-watersheds in 2006 (Continued). 172 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0101111/manifest

Comment

Related Items