The National Stormwater Quality Database (NSQD), Version 4.02

Transcription

1 R. Pitt, A. Maestre, and J. Clary February 7, 208 The National Stormwater Quality Database (NSQD), Version 4.02 Background The National Stormwater Quality Database (NSQD) is an urban stormwater runoff characterization database developed under the direction of Dr. Robert Pitt, P.E., of the University of Alabama and the Center for Watershed Protection under support from the U.S. Environmental Protection Agency. Originally released in 200, followed by several updates by Dr. Pitt and Dr. Alexander Maestre (also while at the University of Alabama), it has recently moved to a new long-term home as a companion project to the International Stormwater BMP Database. The NSQD is being maintained as a separate stand-alone database, serving as an important resource for municipal stormwater managers and researchers who are seeking urban runoff characterization data. The NSQD can be searched for water quality data based on land use, state, and EPA Rain Zone, along with several other criteria. The NSQD can be downloaded from the website, and a new web application has been developed to accommodate custom data extraction queries. The NSQD Version 4.02 (last updated January 205) can be downloaded in two formats containing the same information ( NSQD Version 4.02 Excel Spreadsheet (original format) NSQD Version 4.02 Access Database (new format) In 200, the University of Alabama and the Center for Watershed Protection were awarded a U.S. Environmental Protection Agency, Office of Water 04(b)3 grant to collect and evaluate stormwater data from a portion of the NPDES (National Pollutant Discharge Elimination System) MS4 (municipal separate storm sewer system) stormwater permit holders. In 2008, the NSQD was updated with additional data under continued 04(b)3 support from the EPA. These stormwater quality data and site descriptions were collected and reviewed to describe the characteristics of national stormwater quality, to provide guidance for future sampling needs, and to enhance local stormwater management activities in areas having limited data. The monitoring data collected over nearly a ten-year period from more than 200 municipalities throughout the country have a great potential in characterizing the quality of stormwater runoff and comparing it against historical benchmarks. Prior use of the NSQD data include statistical analyses of the data and providing recommendations for improving the quality and management value of future NPDES monitoring efforts. The current version 4.02 contains the results of about 9,00 urban runoff events and more than 00 different constituents (but many of the obscure constituents have few data). The data are distributed between six main urban land uses: 49%, 20% commercial, 3%, 6%, 6%, and 6% open. Description of the NSQD, Version 4.02 Table lists the constituents included in version 4.02 of the NSQD, and the number of observations (in some cases, most of the observations are below the reported detection limits).

2 Table. Constituents and Numbers of Observations Included in NSQD, version 4.02 (having at least 50 observations) Total events: 9,30 Precipitation depth: 5,72 Runoff depth: 2,59 Hardness:,670 Alkalinity: 525 ph: 3,253 Temperature:,25 TDS, 4,58 Conductivity:,57 Chloride: 869 Total solids: 00 Total suspended solids: 7,73 Turbidity: 936 BOD5: 5,227 COD: 5,290 DO: 92 Fecal coliforms: 2,223 Fecal streptococcus:,37 Total coliforms: 282 Total nitrogen:,23 Total Kjeldahl N: 7,044 Total organic N: 66 Ammonia: 3,020 Nitrate N:,028 Nitrite N: 74 Nitrite + nitrate: 5,748 Total phosphorus: 8,09 Filtered P: 4,05 Ortho phosphate: 746 Filtered ortho P: 244 Total antimony:,584 Filtered antimony: 64 Total arsenic: 2,44 Filtered arsenic: 770 Total barium: 582 Total beryllium:,509 Filtered beryllium: 578 Total cadmium: 4,05 Filtered cadmium: 96 Total chromium: 2,328 Filtered chromium: 82 Total copper: 5,95 Filtered copper:,002 Cyanide:,338 Total iron: 608 Filtered iron: 556 Total lead: 363 (before 984) Total lead: 5,032 (since 984) Filtered lead:,06 (since 84) Total mercury:,702 Filtered mercury: 706 Total nickel: 2,64 Filtered nickel: 807 Total selenium:,737 Filtered selenium: 682 Total silver:,880 Filtered silver: 766 Total thallium:,423 Filtered thallium: 653 Total zinc: 6,638 Filtered zinc: 984 Table. Constituents and Numbers of Observations Included in NSQD, version 4.02 (having at least 50 observations) (continued) Oil and grease: 2,330,2-Dichloropropane: 22* Total petroleum hydrocarbons: 295 Trans-,3-Dichloropropene: 50* Acrolein: 464*,3-Dichloropropylene: 42* Acrylonitrile: 205* Ethyl benzene: 575 Benzene: 23 Methyl bromide: 207* Bromoform: 89* Methyl chloride: 32 Carbon tetrachloride: 89* Methylene chloride: 457 Chlorobenzene: 23*,,2,2-Tetachloroethane: 222* Chlorodibrimo methane: 89* Tetrachloroethylene: 99 Chloroethane: 23* Trichloroflourormethane: 56* Chloroethylvinylether: 624 Toluene: 573 Chloroform: 499,2-Transdichloroethylene: 82* Dichlorobro methane: 6,,-Trichloroethane: 226,-Dichloroethane: 258*,,2-Trichloroethane: 222*,2-Dichloroethane: 247 Trichloroethylene: 83*,-Dichloroethylene: 7* Vinyl chloride: 222* * All, or almost all, non-detected (about 3 organic compounds had the highest detection frequency) 2

3 The NSQD contains data from about 600 outfall sampling locations, with a median of 0 samples per site (maximum 5). Most of the data in version 4 was obtained from Phase I NPDES municipal monitoring programs, along with several other sources (USGS, BMP database outfall data before controls, special state and academic research programs, NURP, etc.). More than 700 new storms were added to this most recent version of the database since version 3. The International Stormwater BMP Database, hosts to the NSQD, now provides data extraction and analysis tools to assist the user. Most of the effort associated with the NSQD during its development included QA/QC processes to verify the data, especially for unusual conditions. Figure illustrates the sources of the data within EPA rainfall zones (not the same as EPA administrative regions). Figure. Distributions of NSQD data sources per EPA rainfall zones. Maestre (many publications) was one of the primary researchers who helped develop and evaluate the NSQD. He investigated first-flush factors, examined distributions of the values and explored various data substitution options for handling non-detectable values, amongst many other interesting stormwater and data analysis issues. Figure 2 is an example of an Xbar-S chart he prepared for each constituent to identify sites within a land use group that did not appear to fall within expected data distributions. He the reviewed those sites, including aerial photographs. 3

4 .000 Median TSS mg/l UC L=00 _ X=52.5 LC L= Site Sample StDev (Log) Site 29 Figure 2. Xbar-S chart of Total Suspended Solids in commercial land use sites indicating unusually high values at some locations (Maestre) UC L=0.594 _ S=0.39 LC L=0.89 Figure 3 is an example from one of the questionable sites that had high suspended solids concentrations. This photograph shows evidence of sediment tracking from an adjacent sand and gravel operation affecting an adjacent monitored commercial site. Figure 3. Aerial photograph of a monitored commercial area having unusually high TSS concentrations showing proximity to quarry operation. 4

5 It is well know that stormwater characteristics vary considerably. Geographical area and land use have been identified as important factors affecting base flow and stormwater runoff quality, for example. Many graduate students and other stormwater researchers have used earlier versions of the NSQD to explore potential relationships of factors that may affect stormwater quality. A selection of these efforts are listed in the references. As an example, Bochis (200) during her PhD research using an earlier version of the NSQD calculated the main factors and interactions affecting stormwater quality, as shown on Table 2. Yellow and green cells note statistically significant relationships (p<0.05); yellow should be used for predictive purposes as they contain the highest order interaction terms. Table 2. Main Factors and Interactions Affecting Stormwater Quality (Bochis 200) Constituent EPA Rain Land Season Zone Use (LU) (SN) (EPA) LU*SN LU*EPA SN*EPA LU*EPA*SN TSS mg/l < < < <0.000 BOD 5 mg/l < < < COD mg/l < < < TP mg/l < < < <0.000 NO 2 + NO 3 mg/l < < < TKN mg/l < <0.000 < Cu mg/l < < < Pb mg/l < < < Zn mg/l < < < <0.000 Land use is a consistent factor affecting the observed variation in stormwater quality, while location (EPA rain zones) are also very important, and season alone had little effect. An example use of these data to calibrate WinSLAMM for different regions in the US is shown below (Pitt 20). Figure 4 is a map showing the NSQD version 3 data locations separated into six major geographical locations. Table 3 shows the land use and location distribution of 4 locations that had detailed development characteristics available to supplement the NSQD water quality data. 5

6 Figure 4. NSQD data locations organized into six geographical areas used to calibrate WinSLAMM. Table 3. Number of Locations by Geographical Area and Land Use for WinSLAMM Calibration Commercial Industrial Institutional Open Resid. Freeways/ Space Highways Central East Coast Great Lakes (the USGS/DNR files) Northwest Southeast Southwest Total by Land Use Total by Region These areas were described in the WinSLAMM model using the best available site data. The long-term average modeled concentrations were compared to the observed concentrations for each location. Example comparison scatterplots are shown in Figure 5. Obviously, some constituents had better matches that others, but the lines of equal values were all significant. It should be noted that all models require calibration and verification. The NSQD is a good place to start for preliminary analyses, but additional locally collected information is necessary for the greatest reliability. 6

7 Figure 5. WinSLAMM modeled and observed concentrations for selected NSQD locations. Data Summary The following is a brief analysis using data from NSQD ver to determine large-scale land use significant differences and associations. Only data from single land use locations were used in these analyses, limiting these analyses to about 5,200 runoff events overall. The remaining data are from mixed land use areas and were therefore not used in this example analysis. Generally, the minimum number of observations per land use per constituent selected for these analyses was about 50 (if fewer, they were not examined). Some of the constituents also had relatively large amounts of non-detected values. The non-detected values were substituted with constant values for these graphical and robust statistical analyses. A selection of eighteen major stormwater constituents were examined having suitable amounts of data, including: ph, SS, COD, fecal coliforms, TKN, NH 3, NO 2+NO 3, P, filtered P, Cd, Cr, Cu, filtered Cu, Pb, Ni, Zn, filtered Zn, and oil and grease. The data for all NSQD constituents were downloaded, and data from the mixed land use locations were removed. The remaining data were then sorted into the six major land use categories (commercial,,,,, and open ) and the numbers of observations (and non-detectable values) determined. Non-detectable values were substituted with appropriate constant values to not distort the data distributions and plots. Statistical summaries of the data for these six land uses (and for the total data set with all land uses combined) were then examined. Minitab (version 6) was then used to create multiple probability distributions (all in log-normal, except for ph). SigmaPlot (version 3) was then used to prepare grouped box and whisker plots of the land uses and overall data sets, along with nonparametric Kruskal-Wallis One Way Analysis of Variance on Ranks (most 7

8 of the data were not normally distributed, even after log transformations due to the large amounts of information). The results of these plots and statistical tests were then used to identify candidate groupings of the land uses. The plots and analyses were then repeated for the candidate groups. Some further data groupings were needed for a few of the constituents until the Kruskal-Wallis tests indicated that all of the remaining data groups were significantly different (p <0.05). These data are summarized on the following pages (grouped box-and-whisker plots along with numbers of observations, the coefficient of variation (standard deviation divided by the mean), the median concentration, and the percentage of all observations that had not-detected concentrations. Values of non-detected occurrences larger than about 5% are high-lighted in yellow. These indicate potential distortions of analytical results (such as the COV values). The box plots clearly illustrate the calculated significance of the different data groupings. The Appendix presents these results for the individual land uses and for the final combined land use groups. In most cases, had the highest concentrations and open areas had the lowest concentrations, as expected. Industrial data were often combined with the high concentration data groups and and data were often combined with the lower concentration data groups. In many cases, the land use data sets associated with the lower concentrations had many nondetected observations, as expected, and the open and areas sometimes had few sampled events. The COV values (and the plots) indicate typically expected large variations in the stormwater concentrations in these groups. However, the very large number of data available for most data groupings increases the confidence of the calculated medians (and data ranges), and the confidence calculations of the comparison tests. Most of the confidence results (p values) are much smaller than the typically accepted value of 0.05 (most are <0.00), also indicating significant power in the comparison results. 8

9 NSQD ver 4 Suspended Solids by Land Use 2 e+5 0 e+4 e+3 ph 8 6 TSS (mg/l) e+2 e+ e+0 4 e- 2 ph - open 2 3 : Open Space 2: Commercial, Freeways, and Industrial 3: Institutional and Residential ph - commercial, freeway, and ph - and count COV median % ND e-2 2 : Freeways and Industrial 2: Commercial, Institutional, Residential, and Open Space SS - commercial,, SS - freeway and, and open count 967 3,905 COV median (mg/l) % ND

10 0000 e+8 e e+6 COD (mg./l) 00 0 FC (#/00mL) e+5 e+4 e+3 e+2 e : Freeways 2: Commercial, Industrial, and Residential 3: Institutional 4: Open Space COD - commercial,, and COD - open COD - COD - count 30 2, COV median (mg/l) % ND e+0 e- 2 3 : Residential 2: Commercial and Open Space 3: Freeways, Industrial, and Institutional FC - commercial and open FC - freeway,, and FC - count COV median (#/00mL) 9,000 3,000 2,000 % ND

11 TKN (mg/l) 0 Ammonia (mg/l) : Freeways and Residential 2: Commercial, Industrial, and Institutional 3: Open Space TKN - freeway and TKN - commercial,, and TKN - open count 2,290 2, COV median (mg/l) % ND : Freeways 2: Commercial, Industrial, Institutional, and Residential 3: Open Space NH3 - commercial,,, and NH3 - open NH3 - count 88,82 2 COV median (mg/l) nd (0.) % ND

12 00 0 NItrite plus Nitrate (mg/l) : Freeways 2: Commercial, Industrial, Open Space, and Residential 3: Institutional NO2NO3 - NO2NO3 - commercial,, open, and NO2NO3 - count 5 3, COV median (mg/l) % ND

13 00 00 Total Phosphorus (mg/l) Filtered Phosphorus (mg/l) : Residential 2: Commercial, Freeways, and Industrial 3: Institutional and Open Space TP - commercial, freeway, and TP and open TP - count 2,277 2, COV median (mg/l) % ND : Residential 2: Commercial, Freeways, and Institutional 3: Industrial and Open Space filt P - commercial, freeway, and filt P - and open filt P - count, COV median (mg/l) % ND

14 Cadmium (ug/l) 0 Chromium (ug/l) : Freeways 2: Commercial, Industrial, Institutional, and Residential 3: Open Space Cd - commercial,,, and Cd - Cd - open count 90 2,56 52 COV median (μg/l) 0.69 nd (0.25) nd (0.25) % ND : Freeways and Industrial 2: Commercial 3: Institutional, Open Space, and Residential Cr - freeway and Cr - commercial Cr -, open, and count COV median (μg/l) nd (2) % ND

15 Copper (ug/l) 00 0 Filtered Cu (ug/l) : Freeways 2: Commercial, Industrial, and Residential 3: Institutional and Open Space Cu - commercial,, and Cu - Cu - and open count COV median (μg/l) % ND : Freeways 2: Industrial and Institutional 3: Commercial and Residential 4: Open Space filt Cu - and filt Cu - commercial and filt Cu - filt Cu - open count COV median (μg/l) nd (2) % ND

16 Lead (ug/l) 0 Nickel (ug/l) : Freeways 2: Commercial, Industrial, and Residential 3: Institutional and Open Space Pb - commercial,, and Pb - and open Pb - count 232 2, COV median (μg/l) % ND : Freeway and Industrial 2: Commercial, Institutional, and Residential 3: Open Space Ni - freeway and Ni - commercial,, and Ni - open count COV median (μg/l) 6.0 nd (.0) nd (.0) % ND

17 e+5 e+5 e+4 e+4 Zinc (ug/l) e+3 e+2 e+ Filtered Zinc (ug/l) e+3 e+2 e+0 e+ e- 2 3 : Industrial, Freeways, and Commercial 2: Institutional and Residential 3: Open Space Zn - commercial, freeway, and Zn - and Zn - open count,804 2, COV median (μg/l) % ND e : Industrial, Commercial, and Freeways 2: Institutional and Residential 3: Open Space filt Zn -, commercial, and freeway filt Zn - and filt Zn - open count COV median (μg/l) nd (0) % ND

18 Oil and Grease (mg/l) : Freeways 2: Commercial, Industrial, and Residential 3: Open Space O&G - commercial,, and O&G - O&G - open count 50,462 9 COV median (mg/l) nd () % ND

19 References Bochis, Elena-Celina. MSCE thesis. Magnitude of Impervious Surfaces in Urban Areas, Department of Civil, Construction, and Environmental Engineering, the University of Alabama pdf Bochis, Elena-Celina. Ph.D. dissertation. Characteristics of Urban Development and Associated Stormwater Quality, Department of Civil, Construction, and Environmental Engineering, the University of Alabama tion_dec_200.pdf Center for Watershed Protection and R. Pitt. Monitoring to Demonstrate Environmental Results: Guidance to Develop Local Stormwater Monitoring Studies using Six Example Study Designs. U.S. Environmental Protection Agency, Office of Water and Wastewater. EPA Cooperative Agreement CP Washington, D.C., 76 pages Hyche, H., R. Pitt, A. Maestre Extensions to the National Stormwater Quality Database (NSQD). Conference CD Water Environment Federation Technical Exposition and Conference, San Diego, CA, October 2007 Hyche, Hunter. MSCE thesis. Extensions to the National Stormwater Quality Database, Department of Civil, Construction, and Environmental Engineering, the University of Alabama Maestre, A., Pitt, R. E., and Derek Williamson. Nonparametric statistical tests comparing first flush with composite samples from the NPDES Phase municipal stormwater monitoring data. Stormwater and Urban Water Systems Modeling. In: Models and Applications to Urban Water Systems, Vol. 2 (edited by W. James). CHI. Guelph, Ontario, pp Maestre, Alex. Ph.D. dissertation. Stormwater Characteristics as Described in the National Stormwater Quality Database, Department of Civil, Construction, and Environmental Engineering, the University of Alabama Maestre, A., R. Pitt, S.R. Durrans, and S. Chakraborti. Stormwater quality descriptions using the three parameter lognormal distribution. Effective Modeling of Urban Water Systems, Monograph 3. (edited by W. James, K.N. Irvine, E.A. McBean, and R.E. Pitt). CHI. Guelph, Ontario, pp Maestre, A. and R. Pitt. The National Stormwater Quality Database, Version., A Compilation and Analysis of NPDES Stormwater Monitoring Information. U.S. EPA, Office of Water, Washington, D.C. (final draft report) August Maestre, A. and R. Pitt. Identification of significant factors affecting stormwater quality using the National Stormwater Quality Database. In: Stormwater and Urban Water Systems Modeling, Monograph 4. (edited by W. James, K.N. Irvine, E.A. McBean, and R.E. Pitt). CHI. Guelph, Ontario, pp Maestre, A. and R. Pitt. Stormwater databases: NURP, USGS, International BMP Database, and NSQD. Chapter 20 in: Contemporary Modeling of Urban Water Systems, ISBN , Monograph 5. (edited by W. James, E.A. McBean, R.E. Pitt, and S.J. Wright). CHI. Guelph, Ontario. pp Pitt, R. A. Maestre, and R. Morquecho. Evaluation of NPDES Phase Municipal Stormwater Monitoring Data. Published proceedings of the National Conference on Urban Storm Water: Enhancing Programs at the Local Level. February 7-20, Sponsored by the US EPA and Chicago Botanic Garden. Chicago

20 Pitt, R., A. Maestre and R. Morquecho. Compilation and review of nationwide MS4 stormwater quality data. 76 th Annual Water Environment Federation Technical Exposition and Conference. Los Angeles, CA. Oct (conference CD-ROM). Pitt, R. E., A. Maestre, R. Morquecho, and Derek Williamson. Collection and examination of a municipal separate storm sewer system database. Stormwater and Urban Water Systems Modeling. In: Models and Applications to Urban Water Systems, Vol. 2 (edited by W. James). CHI. Guelph, Ontario, pp Pitt, R., A. Maestre, and R. Morquecho. Nationwide MS4 stormwater phase database. Watershed 2004, Water Environment Foundation. Dearborn, MI. July 4, (conference CD-ROM). Pitt, R., A. Maestre, and R. Morquecho. Stormwater characteristics as contained in the nationwide MS4 stormwater phase database. Water World and Environmental Resources Conference 2004, Environmental and Water Resources Institute of the American Society of Civil Engineers, Salt Lake City, Utah. July 27 August, (conference CD-ROM). Pitt, R. and A. Maestre. The National Stormwater Quality Database. Stormwater Management Research Symposium. University of Central Florida, Stormwater Management Academy. Orlando, FL, Oct. 2, (conference CD-ROM). Pitt, R. and A. Maestre. Stormwater quality as described in the National Stormwater Quality Database. 0th International Conference on Urban Drainage, Conference CD. IWA Copenhagen, Denmark. August Pitt, R., A. Maestre, H. Hyche, and N. Togawa. The updated National Stormwater Quality Database (NSQD), Version 3." Conference CD Water Environment Federation Technical Exposition and Conference, Chicago, IL, October Pitt, R. Regional WinSLAMM Calibrations using Standard Land Use Characteristics and Pollutant Sources. Data report prepared for Geosyntec, Cadmus, and EPA in support of new stormwater regulations. April modeling_examples/standard_land_use_file_descriptions_final_april_8_20.pdf 20

21 Appendix: Land Use Differences for NSQD ver Data ph summary stats: ph - commercial ph - ph - ph - ph - open ph - count mean stdev COV median min max # ND % ND all ph NSQD ver 4 ph by Land Use ph : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All data combined 2

22 Kruskal-Wallis One Way Analysis of Variance on Ranks P Values (yellow comparisons are significant) P values ph - commercial ph - ph - ph - ph - open ph - ph - commercial X <0.00 <0.00 ph - X <0.00 <0.00 ph - X 0.09 <0.00 <0.00 ph X <0.00 ph - open <0.00 <0.00 <0.00 <0.00 X <0.00 ph - <0.00 <0.00 <0.00 <0.00 X summary stats: ph - open ph - com free ph - institut res indus count mean stdev COV median min max # ND 0 2 % ND ph : Open Space 2: Commercial, Freeways, and Industrial 3: Institutional and Residential 22

23 All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Q P Ranks phopen vs phinstututres <0.00 phopen vs phcomfreindus <0.00 phcomfreindus vs phinstututres <0.00 Industrial combined with commercial and freeway areas Institutional combined with areas Suspended Solids summary stats: SS - commercial SS - SS - SS - SS - open SS - all TSS (mg/l) count, ,364 4,87 mean stdev COV median min max 2,385 4,800 2,490 4,380 8,728 4,68 4,800 # ND % ND

24 NSQD TSS by Land Use e+5 e+4 e+3 TSS (mg/l) e+2 e+ e+0 e- e : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All data combined All Pairwise Multiple Comparison Procedures (Dunn's Method) : P values SS - commercial SS - SS - SS - SS - open SS - SS - commercial X 0.05 <0.00 SS X SS - <0.00 X 0.8 <0.00 <0.00 SS X 0.33 SS - open < X 0.63 SS < X 24

25 summary stats: SS - free indus SS - com instit res open count 967 3,905 mean stdev COV median min 0. max 4,800 8,728 # ND 6 32 % ND NSQD ver 4 Suspended Solids by Land Use e+5 e+4 e+3 TSS (mg/l) e+2 e+ e+0 e- e-2 2 : Freeways and Industrial 2: Commercial, Institutional, Residential, and Open Space All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Ranks Q P P<0.050 SSFreeIndus vs SSComInsResOp <0.00 Yes Freeway and combined Commercial,,, and open combined 25

26 COD summary COD - COD - COD - COD - COD - open COD - all COD stats: commercial count mean stdev COV median min max 635, ,674,674 # ND % ND and open non-detects >5% NSQD ver 4 COD COS (mg/l) : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All areas combined 26

27 Kruskal-Wallis One Way Analysis of Variance on Ranks P values P values COD - commercial COD - COD - COD - (many NDs) COD - open (many NDs) COD - COD - commercial X <0.00 < COD - X 0.37 <0.00 < COD X <0.00 < COD - (many NDs) <0.00 <0.00 <0.00 X < COD - open (many NDs) <0.00 <0.00 <0.00 <0.00 X <0.00 COD < <0.00 X summary stats: COD - COD - com COD - COD - open indus res count 30 2, mean stdev COV median min max,03, # ND % ND and open areas have many nondetected values 27

28 COD (mg./l) : Freeways 2: Commercial, Industrial, and Residential 3: Institutional 4: Open Space All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Ranks Q P CODFree vs CODOpen <0.00 CODFree vs CODIstit <0.00 CODFree vs CODComIndusRes CODComIndusRes vs CODOpen <0.00 CODComIndusRes vs CODIstit <0.00 CODIstit vs CODOpen <0.00 Industrial combined with commercial and areas 28

29 Fecal Coliforms summary stats: FC - FC - FC - FC - FC - open FC - FC - all commercial count ,447 mean 27,893 8,553 44,756 63,837 8,290 55,5 stdev 7,526 22,79 262,640 37, ,25 282,90 COV median 4,200 2,000,980 2,50 9,000 3,700 min ,600 0 max 60,000 60,000 3,600,000 4,300 2,800,000 5,230,000 5,230,000 # ND (both < and >) % ND NSQD ver 4 Fecal Coliforms by Land Use e+8 e+7 e+6 FC (#/00 ml) e+5 e+4 e+3 e+2 e+ e+0 e : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All areas combined 29

30 Kruskal-Wallis One Way Analysis of Variance on Ranks P values P values FC - commercial FC - FC - FC - (too few data) FC - open FC - FC - commercial X FC X <0.00 FC X <0.00 FC - (too few data) X FC - open X FC <0.00 < X summary stats: FC - FC - com open FC - free indus instit count mean 8,290 35,920 39,27 stdev 365,25 62, ,522 COV median 9,000 3,000 2,000 min 4 2 max 5,230,000 2,800,000 3,600,000 # ND (both < and >) % ND

31 e+8 e+7 e+6 FC (#/00mL) e+5 e+4 e+3 e+2 e+ e+0 e- 2 3 : Residential 2: Commercial and Open Space 3: Freeways, Industrial, and Institutional All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Ranks Q P FCRes vs FCFreeINdusInstit <0.00 FCRes vs FCComOpen <0.00 Industrial combined with and (few data) Commercial and open combined Total Kjeldahl Nitrogen summary TKN - TKN - TKN - TKN - TKN - open TKN - TKN - all stats: commercial count mean stdev COV median min max # ND % ND

32 NSQD Ver 4 TKN by Land Use TKN (mg/l) : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All Areas Combined Kruskal-Wallis One Way Analysis of Variance on Ranks P values P values TKN - commercial TKN - TKN - TKN - TKN - open TKN - TKN - commercial X <0.00 <0.00 <0.00 < TKN - <0.00 X <0.00 <0.00 <0.00 <0.00 TKN - <0.00 <0.00 < TKN - <0.00 <0.00 X X <0.00 <0.00 TKN - open <0.00 <0.00 <0.00 <0.00 X <0.00 TKN <0.00 <0.00 <0.00 <0.00 X 32

33 summary stats: TKN - free res TKN - com indus instit TKN - open count 2,290 2, mean stdev COV median min max # ND % ND TKN (mg/l) : Freeways and Residential 2: Commercial, Industrial, and Institutional 3: Open Space All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Ranks Q P TKNFreRes vs TKNOpen <0.00 TKNFreRes vs TKNCoIndusInstit <0.00 TKNCoIndusIns vs TKNOpen <

34 Ammonia summary stats: NH3 - commercial NH3 - NH3 - NH3 - NH3 - open NH3 - NH3 - all count mean stdev COV median min max # ND % ND Many non-detects NSQD ver 4 Ammonia by Land Use 00 0 Ammonia (mg/l) : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All Areas Combined 34

35 Kruskal-Wallis One Way Analysis of Variance on Ranks P values P values NH3 - commercial NH3 - NH3 - NH3 - NH3 - open NH3 - NH3 - commercial X < < NH3 - <0.00 X <0.00 <0.00 <0.00 <0.00 NH3 - <0.00 X <0.00 NH < X NH3 - open <0.00 <0.00 < X <0.00 NH < <0.00 X P values NH3 - commercial NH3 - NH3 - NH3 - NH3 - open NH3 - NH3 - commercial X < < NH3 - <0.00 X <0.00 <0.00 <0.00 <0.00 NH3 - <0.00 X <0.00 NH < X NH3 - open <0.00 <0.00 < X <0.00 NH < <0.00 X summary stats: NH3 - NH3 - com ind NH3 - open inst res count 88,82 2 mean stdev COV median min max # ND % ND

36 00 0 Ammonia (mg/l) : Freeways 2: Commercial, Industrial, Institutional, and Residential 3: Open Space All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Q P Ranks NH3Free vs NH3Open <0.00 NH3Free vs NH3CoIndInsRes <0.00 NH3CoIndInsRes vs NH3Open <0.00 Industrial combined with commercial,, and Nitrites plus Nitrates summary stats: NO2NO3 - commercial NO2NO3 - NO2NO3 - NO2NO3 - NO2NO3 - open NO2NO3 - count mean stdev COV median min max # ND % ND N02+NO3 - all 36

37 NSQD ver 4 Nitrite plus Nitrate by Land Use 00 0 Nitrite plus Nitrate (mg/l) : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All Areas Combined Kruskal-Wallis One Way Analysis of Variance on Ranks P values NO2NO3 - commercial NO2NO3 - NO2NO3 - NO2NO3 - NO2NO3 - open NO2NO3 - P values NO2NO3 - commercial X <0.00 < NO2NO3 - <0.00 X <0.00 <0.00 <0.00 <0.00 NO2NO3 - <0.00 X <0.00 NO2NO3 - <0.00 <0.00 <0.00 X <0.00 <0.00 NO2NO3 - open 0.82 <0.00 <0.00 X 0.9 NO2NO3 - <0.00 < X 37

38 summary stats: NO2NO3 - NO2NO3 - com ind open res NO2NO3 - count 5 3, mean stdev COV median min max # ND 6 % ND NItrite plus Nitrate (mg/l) : Freeways 2: Commercial, Industrial, Open Space, and Residential 3: Institutional All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Ranks Q P NO2NO3Free vs NO2NO3Instit <0.00 NO2NO3Free vs NO2NO3CoIdOpRe <0.00 NO2NO3CoIdOpRe vs NO2NO3Instit <0.00 Industrial combined with commercial, open, and 38

39 Total Phosphorus summary TP - TP - TP - TP - TP - open TP - TP - all stats: commercial count, ,277 4,929 mean stdev COV median min max # ND % ND NSQD ver 4 Total Phosphorus by Land Use 00 0 Total Phosphorus (mg/l) : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All Areas Combined 39

40 Kruskal-Wallis One Way Analysis of Variance on Ranks P values P values TP - commercial TP - TP - TP - TP - open TP - TP - commercial X < <0.00 <0.00 <0.00 TP - <0.00 X 0.4 <0.00 < TP X <0.00 <0.00 <0.00 TP - <0.00 <0.00 <0.00 X <0.00 TP - open <0.00 <0.00 <0.00 X <0.00 TP - < <0.00 <0.00 <0.00 X summary stats: TP - TP - com free indus TP - instit open count 2,277 2, mean stdev COV median min max # ND % ND Total Phosphorus (mg/l) : Residential 2: Commercial, Freeways, and Industrial 3: Institutional and Open Space 40

41 All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Ranks Q P TPRes vs TPInstitOpen <0.00 TPRes vs TPComFreInd <0.00 TPComFreInd vs TPInstitOpen <0.00 Commercial, and areas combined Institutional and open areas combined Filtered Phosphorus summary stats: filt P - commercial filt P - filt P - filt P - filt P - open filt P - filt P - all count ,244 2,657 mean stdev COV median min max # ND % ND

42 NSQD ver 4 Filterable Phosphorus by Land Use 00 Filterable Phosphorus (mg/l) : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All Areas Combined Kruskal-Wallis One Way Analysis of Variance on Ranks P values P values filt P - commercial filt P - filt P - filt P - filt P - open filt P - filt P - commercial X <0.00 filt P - X filt P X <0.00 filt P X filt P - open X <0.00 filt P - < < <0.00 X summary stats: filt P - filt P - com free filt P - indus open instit count, mean stdev COV median min max # ND % ND

43 00 Filtered Phosphorus (mg/l) : Residential 2: Commercial, Freeways, and Institutional 3: Industrial and Open Space All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Ranks Q P filtpres vs filtpindusopen <0.00 filtpres vs filtpcomfreinst <0.00 filtpcomfrein vs filtpindusope <0.00 Industrial combined with open commercial,, and combined Cadmium summary Cd - Cd - Cd - Cd - Cd - open Cd - Cd - all stats: commercial count mean stdev COV median min max # ND % ND

44 NSQD ver 4 Cadmium by Land Use Cadmium (ug/l) : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All Areas Combined Kruskal-Wallis One Way Analysis of Variance on Ranks P values P values Cd - commercial Cd - Cd - Cd - open Cd - Cd - commercial X < Cd X <0.00 <0.00 Cd - X < Cd - open <0.00 <0.00 <0.00 X Cd < X 44

45 summary stats: Cd - Cd - com indus Cd - open inst res count 90 2,56 52 mean stdev COV median min max # ND 35, % ND Many non-detectable values Cadmium (ug/l) : Freeways 2: Commercial, Industrial, Institutional, and Residential 3: Open Space All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Ranks Q P CdFree vs CdOpen <0.00 CdFree vs CdCoIndsInstRes <0.00 CdCoIndsInstRes vs CdOpen <0.00 Commercial,,, and areas combined 45

46 Chromium summary stats: Cr - commercial Cr - Cr - Cr - Cr - open Cr - Cr - all count mean stdev COV median min max # ND % ND NSQD ver 4 Chromium by Land Use Chromium (ug/l) : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All Areas Combined 46

47 Kruskal-Wallis One Way Analysis of Variance on Ranks P Values P values Cr - commercial Cr - Cr - Cr - Cr - open Cr - Cr - commercial X <0.00 < Cr - <0.00 X <0.00 <0.00 <0.00 Cr - < X <0.00 <0.00 <0.00 Cr <0.00 <0.00 X Cr - open <0.00 <0.00 X Cr <0.00 <0.00 X summary stats: Cr - free indus Cr - commercial Cr - instit open res count mean stdev COV median min max # ND % ND Many non-detectable values 47

48 Chromium (ug/l) : Freeways and Industrial 2: Commercial 3: Institutional, Open Space, and Residential All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Q P Ranks CrFreeIndus vs CrInstitOpenRes <0.00 CrFreeIndus vs CrCom <0.00 CrCom vs CrInstitOpenRes <0.00 Copper summary stats: Cu - commercial Cu - Cu - Cu - Cu - open Cu - Cu - all count ,654 3,794 mean stdev COV median min max , ,360 # ND % ND

49 NSQD ver 4 Copper by Land Use Copper (ug/l) : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All Areas Combined Kruskal-Wallis One Way Analysis of Variance on Ranks P values P values Cu - commercial Cu - Cu - Cu - Cu - open Cu - Cu - commercial X <0.00 <0.00 < Cu - <0.00 X <0.00 <0.00 <0.00 <0.00 Cu - <0.00 X <0.00 < Cu - <0.00 <0.00 <0.00 X <0.00 Cu - open <0.00 <0.00 <0.00 X <0.00 Cu < <0.00 <0.00 X 49

50 summary stats: Cu - Cu - com indus res Cu - instit open count mean stdev COV median min max # ND % ND Copper (ug/l) : Freeways 2: Commercial, Industrial, and Residential 3: Institutional and Open Space All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Ranks Q P CuFree vs CuInstitOpen <0.00 CuFree vs CuComIndusRes <0.00 CuComIndusRes vs CuInstitOpen <0.00 Industrial combined with commercial and Institutional and open combined 50

51 Filtered Copper summary stats: filt Cu - filt Cu - filt Cu - filt Cu - filt Cu - filt Cu - Cu - all commercial open count mean stdev COV median min max # ND % ND NSQD ver 4 Filtered Copper by Land Use Filtered Copper (ug/l) : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All Areas Combined 5

52 Kruskal-Wallis One Way Analysis of Variance on Ranks P values P values filt Cu - commercial filt Cu - filt Cu - filt Cu - filt Cu - filt Cu - commercial X <0.00 filt Cu - <0.00 X <0.00 filt Cu X 0.34 filt Cu X 0.02 filt Cu - < X summary stats: filt Cu - filt Cu - indus instit filt Cu - com res filt Cu - open count mean stdev COV median min max # ND % ND Many non-detectable values 52

53 Filtered Cu (ug/l) : Freeways 2: Industrial and Institutional 3: Commercial and Residential 4: Open Space All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Ranks Q P filtcufree vs filtcuopen <0.00 filtcufree vs filtcucomres <0.00 filtcufree vs filtcuindusin filtcuindusin vs filtcuopen <0.00 filtcuindusin vs filtcucomres filtcucomres vs filtcuopen <0.00 combined with institution commercial combined with 53

54 Lead (since 984) summary stats: Pb - Pb - Pb - Pb - Pb - open Pb - Pb - all commercial count ,345 3,38 mean stdev COV median min max , ,200 # ND % ND NSQD ver 4 Lead (since 984) by Land Use Lead Uug/L) : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All Areas Combined 54

55 Kruskal-Wallis One Way Analysis of Variance on Ranks P values P values Pb - commercial Pb - Pb - Pb - Pb - Pb - commercial X <0.00 <0.00 <0.00 Pb - <0.00 X <0.00 <0.00 <0.00 Pb - <0.00 X <0.00 <0.00 Pb - <0.00 <0.00 <0.00 X <0.00 Pb - <0.00 <0.00 <0.00 <0.00 X summary stats: Pb - Pb - com ind Pb - instit open res count 232 2, mean stdev COV median min max 660, # ND % ND Lead (ug/l) : Freeways 2: Commercial, Industrial, and Residential 3: Institutional and Open Space 55

56 All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Ranks Q P PbFree vs PbINstitOpen <0.00 PbFree vs PbComIndusRes <0.00 PbComIndusRes vs PbINstitOpen <0.00 Industrial combined with commercial and areas Institutional and open areas combined Nickel summary Ni - Ni - Ni - Ni - Ni - open Ni - NI - all stats: commercial count ,485 mean stdev COV median min max # ND % ND Many non-detectable values 56

57 NSQD ver 4 Nickel by Land Use Nickel (ug/l) : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All Areas Combined Not enough data for reliable Kruskal-Wallis analyses. summary stats: Ni - free indus Ni - com instit res Ni - open count mean stdev COV median 6.0 min.0 max # ND % ND

58 Nickel (ug/l) : Freeway and Industrial 2: Commercial, Institutional, and Residential 3: Open Space All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Ranks Q P NiFreIndus vs NiOpen <0.00 NiFreIndus vs NiComInstitRes <0.00 NiComInstitRes vs NiOpen Freeways and combined Commercial,, and combined Zinc summary stats: Zn - commercial Zn - Zn - Zn - Zn - open Zn - ZN - all count ,789 4,062 mean stdev COV median min max 3,050 2,00 8,00, ,700 4,700 # ND % ND

59 NSQD ver 4 Zinc by Land Use e+5 e+4 e+3 Zinc (ug/l) e+2 e+ e+0 e : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All Areas Combined Kruskal-Wallis One Way Analysis of Variance on Ranks P Values P values Zn - commercial Zn - Zn - Zn - Zn - Zn - commercial X <0.00 <0.00 Zn X <0.00 <0.00 Zn X <0.00 <0.00 Zn - <0.00 <0.00 <0.00 X Zn - <0.00 <0.00 <0.00 X 59

60 summary stats: Zn - com free indus Zn - instit res Zn - open count,804 2, mean stdev COV median min max 8,00 4, # ND % ND e+5 e+4 e+3 Zinc (ug/l) e+2 e+ e+0 e- 2 3 : Industrial, Freeways, and Commercial 2: Institutional and Residential 3: Open Space All Pairwise Multiple Comparison Procedures (Dunn's Method) Comparison Diff of Ranks Q P ZnIndusFreCom vs ZnOpen <0.00 ZnIndusFreCom vs ZnInstitRes <0.00 ZnInstitRes vs ZnOpen <0.00 Industrial combined with and commercial areas Institutional and areas combined 60

61 Filtered Zinc summary stats: filt Zn - commercial filt Zn - filt Zn - filt Zn - filt Zn - open filt Zn - filt Zn - all count mean stdev COV median min max,000,766 7, ,020 7,300 # ND % ND NSQD ver 4 Filtered Zinc by Land Use e+5 e+4 Filtered Zinc (ug/l) e+3 e+2 e+ e : Commercial 2: Freeways 3: Industrial 4: Institutional 5: Open Space 6: Residential 7: All Areas Combined 6

62 Kruskal-Wallis One Way Analysis of Variance on Ranks P Values P values filt Zn - commercial filt Zn - filt Zn - filt Zn - filt Zn - filt Zn - commercial X filt Zn - X 0.56 <0.00 filt Zn X <0.00 filt Zn X 0.28 filt Zn <0.00 < X summary stats: filt Zn - indus com fre filt Zn - instit res filt Zn - open count mean stdev COV median min max 7,300, # ND 6 86 % ND e+5 e+4 Filtered Zinc (ug/l) e+3 e+2 e+ e : Industrial, Commercial, and Freeways 2: Institutional and Residential 3: Open Space 62