<%BANNER%>

Granulometry, Partitioning and Speciation of Urban Source Area Snow Pollutants Generated from Vehicular Transportation a...

Permanent Link: http://ufdc.ufl.edu/UFE0041129/00001

Material Information

Title: Granulometry, Partitioning and Speciation of Urban Source Area Snow Pollutants Generated from Vehicular Transportation and Infrastructure
Physical Description: 1 online resource (184 p.)
Language: english
Creator: Magill, Natalie
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: gamma, hazen, lake, particle, partitioning, snowmelt, speciation
Environmental Engineering Sciences -- Dissertations, Academic -- UF
Genre: Environmental Engineering Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This study focuses on pollutants found in urban snowmelt. Stormwater studies and regulations have become fairly common in the US while snowmelt studies are rare. The purpose of this study is to show how urban snow can be a source of pollution in the environment. While snow from the Lake Tahoe watershed was studied, the methodology can be utilized in other urban areas with significant snowfall and traffic. The importance to the Lake Tahoe region is to qualify an additional source of solids, metals, and nutrients into the lake. This study is important to northern US cities and other cold-climate cities as it is a good source of data in a field with few extensive studies.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Natalie Magill.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Sansalone, John.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0041129:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041129/00001

Material Information

Title: Granulometry, Partitioning and Speciation of Urban Source Area Snow Pollutants Generated from Vehicular Transportation and Infrastructure
Physical Description: 1 online resource (184 p.)
Language: english
Creator: Magill, Natalie
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: gamma, hazen, lake, particle, partitioning, snowmelt, speciation
Environmental Engineering Sciences -- Dissertations, Academic -- UF
Genre: Environmental Engineering Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This study focuses on pollutants found in urban snowmelt. Stormwater studies and regulations have become fairly common in the US while snowmelt studies are rare. The purpose of this study is to show how urban snow can be a source of pollution in the environment. While snow from the Lake Tahoe watershed was studied, the methodology can be utilized in other urban areas with significant snowfall and traffic. The importance to the Lake Tahoe region is to qualify an additional source of solids, metals, and nutrients into the lake. This study is important to northern US cities and other cold-climate cities as it is a good source of data in a field with few extensive studies.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Natalie Magill.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Sansalone, John.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0041129:00001


This item has the following downloads:


Full Text

PAGE 1

1 GRANULOMETRY, PARTITIONING AND SPECIATION OF URBAN SOURCE AREA SNOW POLLUTANTS GENERATED FROM VEHICULAR TRANSPORTATION AND INFRASTRUCTURE By NATALIE MAGILL A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Natalie R Magill

PAGE 3

3 To everyone who aspired to accomplish something, without knowing the enormous amount of work required, and still went through with it after they found out.

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank my parents and sisters for always believing in me and for always being there. For his continued support, I woul d like to thank Adam Winberry ; I could not have done it without him. I would also like to thank Steve and Karen Winberry for continuously motivating me. I would like to thank Dr. John Sansalone for mentoring me throughout my academic career. I would also like to thank my committee members for their reviews and advice: Dr. Joseph Delfino, Dr. Paul Chadik, and Dr. Willie Harris. Appreciation is also extended to Dr. Ben Koopman for the VBA knowledge, which made easier work of my data analysis. For their continu ous help with brainstorming and problem solving, I would like to thank Dr. Jong Yeop Kim and Dr. Srikanth Pathapati. For their assistance in the lab, I would like to thank Gwen Ryskamp and Megan Baxter, without them I might still be titrating. Speci al thanks to Saurabh Raje and Ruben Kertesz for assistance with SWMM and GIS. I would also like to thank my previous grad students for teaching me everything they could: Brian EngChong Voon, Chris Dean, Dr. Chad Cristina, Aimee (Blazier) Richardson, Dr. Hong Lin, and Erin (Krielow) Lahr. I would also like to thank my fifth grade teacher, Mrs. Whitworth, for the environmental spark that led me here.

PAGE 5

5 TABLE OF CONTENTS ACKNOWLEDGMENTS ...............................................................................................................4 page LIST OF TABLES ...........................................................................................................................8 LIST OF FIGURES .......................................................................................................................10 ABSTRACT ...................................................................................................................................15 1 GLOBAL INTRODUCTION .................................................................................................17 2 NONCOLLIODAL GRANULOMETRY OF URBAN SNOW SNOWMELT GENERATED FROM TRAFFIC LOADINGS IN THE LAKE TAHOE BASIN ................23 Introduction .............................................................................................................................23 Objectives ...............................................................................................................................25 Background .............................................................................................................................26 Methodology and Laboratory Analysis ..................................................................................27 Sampling Locations .........................................................................................................27 Sampling Metho d ............................................................................................................28 Suspended, Settleable, and Sediment Fractions ..............................................................28 Particle Size Distribution (PSD) Analysis .......................................................................29 Particle Density of PM ....................................................................................................30 Specific Surface Area and Total Surface Area ................................................................30 Surface Charge and Point of Zero Charge (PZC) ............................................................31 Settling Velocity ..............................................................................................................32 Treatability Study ............................................................................................................33 Results .....................................................................................................................................35 Relative Mass of Suspend ed, Settleable and Sediment Particles ....................................35 Particle Density ...............................................................................................................35 PSD Analysis ...................................................................................................................36 SSA and Total SA ...........................................................................................................36 Surface Charge and PZC .................................................................................................38 Settling V elocity ..............................................................................................................38 Treatability Study ............................................................................................................38 Conclusions .............................................................................................................................39 Discussion ...............................................................................................................................41 Nomenclature ..........................................................................................................................42 3 DISTRIBUTION OF PA RTICULATE BOUND METALS IN SOURCE AREA SNOW IN THE LAKE TAHOE WATERSHED ...................................................................55 Introduction .............................................................................................................................55 Objectives ...............................................................................................................................58 Methodology and Laboratory Analysis ..................................................................................58

PAGE 6

6 Sampling Locations .........................................................................................................58 Snow Sampling ................................................................................................................59 Particle Size Distributions (PSD) ....................................................................................60 Specific Surface Area and Total Surface Area ................................................................60 Metals Analysis ...............................................................................................................61 Data Analysis ...................................................................................................................61 Modeling Basin Treatability for PM based Metals .........................................................63 Results .....................................................................................................................................64 Discussion ...............................................................................................................................68 Summary and Conclusi ons .....................................................................................................69 Nomenclature ..........................................................................................................................71 4 PARTITIONING AND SPECIATION OF METALS IN URBAN SNOWMELT EXPOSED TO TRAFFIC ACTIVITIES IN THE LAKE TAHOE WATERSHED ..............83 Introduction .............................................................................................................................83 Objectives ...............................................................................................................................85 Background .............................................................................................................................85 Methodology ...........................................................................................................................88 Sampling Locations .........................................................................................................88 Snow Sampling ................................................................................................................89 Laboratory Analysi s ........................................................................................................90 Water chemistry parameters .....................................................................................90 Sample digestion for metal partitioning ...................................................................90 Ion analysis ...............................................................................................................91 Speciation analysis ...................................................................................................91 Results .....................................................................................................................................92 Conclusions .............................................................................................................................96 Discussion ...............................................................................................................................97 N omenclature ..........................................................................................................................98 5 SPECIATION OF NUTRIENTS AND ANIONS IN URBAN SNOWMELT EXPOSED TO TRAFFIC ACTIVITIES IN THE LAKE TAHOE WATERSHED ...............................112 Introduction ...........................................................................................................................112 Background ...........................................................................................................................114 Objectives .............................................................................................................................116 Methodology and Laboratory Analysis ................................................................................117 Sampling Locations .......................................................................................................117 Snow Sampling and Preparation ...................................................................................117 Laboratory Analysis ......................................................................................................118 Water quality parameters .......................................................................................118 PSD an alysis ...........................................................................................................119 Metals analysis .......................................................................................................119 Ion analysis .............................................................................................................119 Speciation analysis .................................................................................................120 Results ...................................................................................................................................120

PAGE 7

7 Conclusions ...........................................................................................................................125 Discussion .............................................................................................................................127 Nomenclature ........................................................................................................................128 6 SETTLING INDUCED CHANGES TO THE GRANULOMETERY OF PARTICULATE MATTER IN SNOW IMPACTED BY TRAFFIC IN THE LAKE TAHOE WATERSHED .......................................................................................................145 Introduction ...........................................................................................................................145 Objectives .............................................................................................................................146 Background ...........................................................................................................................147 Methodology and Laboratory Analysis ................................................................................148 Sampling Locations .......................................................................................................148 Snow Sampling and Preparation ...................................................................................149 Sample Analysis ............................................................................................................149 Water quality analysis ............................................................................................149 Composit e samples .................................................................................................150 Downstream samples ..............................................................................................150 Particle size analysis ...............................................................................................151 Modeling of PSDs .........................................................................................................151 Treatability Study ..........................................................................................................152 Results ...................................................................................................................................153 Water Quality Samples ..................................................................................................153 Composite Samples .......................................................................................................153 Watershed Samples .......................................................................................................154 Treatability Study ..........................................................................................................154 Conclusions ...........................................................................................................................155 Discussion .............................................................................................................................156 Nomenclature ........................................................................................................................157 7 GLOBAL CONCLUSIONS .................................................................................................170 LIST OF REFERENCES .............................................................................................................174 BIOGRAPHICAL SKETCH .......................................................................................................184

PAGE 8

8 LIST OF TABLES Table page 11 Percent of material passing a specified sieve size required for de icing sands in Nevada as determined by ASTM. ......................................................................................22 12 Table of Grain size classifications from ASTM. ...............................................................22 21 Location of snow sampling sites. ......................................................................................45 22 Total PM dry mass and aqueous volume along with relative sediment, settleable, and suspended particle mass. ....................................................................................................45 23 Geometric mean (dg), central tendency ( 50), skewness (Sk), kurtosis (KG), and unifoI). The explanations of the parameters are shown below. ............................46 24 Summary of gamma distribution parameters for individual sites and the median PSD. The first column indicates values for the mass PSD plot (Figure 2 4 top). The mass balance error (MBE) is also shown for the mass PSDs. The second column (PSD by SA) indicates values for the SA PSD plot (Figure 24 bottom). ........................................47 31 Summary of sites, locations, and sampling dates. .............................................................73 32 Summary of total metal mass (mg per 1.0 kg of PM) and total PM (MT) (units of g) measured for each site and the mean of all sites. The total liquid volume of the melted sample is shown in the VT column ( units of L). ....................................................73 33 Measured vs. modeled parameters from the gamma distribution. Cumulative mass (g) distribution was summed from l argest to smallest particle diameter. The dXm values represent the diameter at which X percent by mass is greater. ...............................74 41 Summary of sites, locations, and sampling dates. ...........................................................100 42 Summary of water chemistry parameters in snowmelt samples. .....................................101 51 NPDES effluent discharge limits. (USEPA 2003). NPDES permit numbers CAG616001, CAG6161002, CAG 6160032. ..................................................................130 52 Summary of sites, locations, and sampling dates. ...........................................................131 53 Summary of MINTEQ input data for snowmelt samples. The median charge balance for the snowmelt samples is 4.98%. (pH = log{H+}) ....................................................132 54 Water quality, water chemistry, and solids summary for the snowmelt samples. ...........133 55 Summar y of MINTEQ inputs for control snow samples. The median charge balance for the control snow samples is 48.46%. (pH = log{H+}) .............................................134

PAGE 9

9 56 Water quality, water chemistry, and solids summary for control snow samples. ............135 57 The percent d ifference in value between the MINTEQ model run at sea level and S Lake Tahoe elevation for the snowmelt samples. An increase in percentage indicates an increase in median value of snowmelt at elevation. ....................................................136 58 The percent difference between the MINTEQ model run at sea level and S. Lake Tahoe elevation for the control snow samples for the carbonate system species. A positive percen tage value indicates a higher average concentration at S. Lake Tahoe elevation as compared to sea level. ..................................................................................136 59 The equilbrium constants (pKa) values of the ammonia, carbonate, and phosphate species as a function of temperature. (Benjamin 2002, Goldberg et al. 2002, Plummer and Busenberg 1982, Stumm and Morgan 1996) .............................................136 61 Summary of sites, locations, and sampling dates. Sites represented with Site #s were sampled as snow, while the remaining 3 were sampled as snowmelt. ............................159 62 Summary of the unsettled (0 minutes), batch settled (60 minutes), and percent reduction for SSC, VSSC, and turbidity. Results represent 12 observations for 9 sites. .................................................................................................................................160 63 Mean, median, maximum, and minimum values for particulate matter (PM) indices (d10, d50, d90) with associated redu ction percentages. (Samples collected as snow.) .......161 64 Values of SSC, VSSC, turbidity, parametric indexes (d10, d50, d90), and modeling parameters ( ) with increased settling times. (Samples collected as snowmelt). ........162

PAGE 10

10 LIST OF FIGURES Figure page 21 Particle density as a function of particle diameter. ............................................................47 22 Particle size distributions for all eleven sites are shown in the top plot; values shown are normalized to 1000g. Parameters listed are for a transpose cumulative gamma distribution used to model the cumulative particle size distribution. The middle plot shows incremental specific surface area. The bottom plots show incremental total surface area and cumulative percent, values shown are normalized to 1000g. .................48 23 A comparison of calculated and measured SSA values for snow particulate matter. Measured values are the average and standard deviation of all SSA data presented; calculated values assume constant specific gravity, spherical particles, and solid particles. (Sansalone et al 1998). ......................................................................................49 24 Image of 4 grains from the 4750 < d< 2000 mm 1 103B sample. The paperclip in the lower part of the image has a diameter of 1 mm. The samples shown represent A) a smooth grain with minimal surface area, B) a grain with debris that would contribute to surface area, C) a porous grain with a large surface area, and D) a quartz grain with moderate surface area. ...........................................................................50 25 Surface charge as a function of pH for a 425 m (left) and a 75 m (right) particle, the average and standard deviation of the point of zero charge is also shown. .................50 26 Point of zero charge by site. Plot (a) shows the 3 samples from Site 1, plot (b) shows the 3 samples from Site 2, plot (c) shows the 2 samples from site 4, plot (d) shows the remaining Sites 3,5 and 6. ............................................................................................51 27 Measured and modeled settling velocities. ........................................................................51 28 (Top) Average temperature data. Average high and low data from Western Regional Climate Center (WRCC 2002), mean temperature from (WCI, 2008). (Bottom) Influent and effluent flow rates for the sedimentation basin. ............................................52 29 The mass and number based removal efficiency of the basin for PM greater than 25 m in diameter given the 2008 precipitation and temperature conditions. .......................52 210 The effect of flow rate on the number based removal efficiency of the basin for PM with diameters greater than 1 m. .....................................................................................53 211 Range of influent and effluent number based PSDs. .........................................................53 212 EDS (Energy dispersive spectrometer) analysis on 3 particle diameters. The EDS shown for the 2000 mm particle diameter is the average of 3 separate grains. The EDS for the 425 and 63 um particle diameters are bulk EDS plots and were taken over multiple grains with areas of 27 mm2 and 4.3 mm2 respectively. .............................54

PAGE 11

11 31 The snowmelt basin in South Lake Tahoe, CA and shown in white. The surrounding watershed that contributes runoff to the basin is shown in grey. Coordinates at centroid of the watershed are: 38 54 55 N and 119 58 51 W. ................................75 32 Total metal Mass measured for selected metals normalized to 1000 g of total PM on the left. PSD of PM for all sites and percent finer by mass of the median values is shown on the right. .............................................................................................................76 33 Distribution of metal mass as a function of particle diameter for selected metals (Al, Fe, As, Cr). Results are illustrated on a surf ace area (column 1) and a mass basis (column 2) with measured cumulative distribution modeled with a cumulative gamma distribution. (TM = Total Mass of Metal Species) ...............................................77 34 Distribution of metal mass as a function of particle diameter for selected metals (Cu, Cd, Pb, Zn). Results are illustrated on a surface area (column 1) and mass basis (column 2) with measured cumulative distribution modeled with a cumulative gamma distribution. (TM = Total Mass of Metal Species) ...............................................78 35 Summary of gamma distribution parameters ( ), all coefficient of determinations (r2) exceed 0.94. The median values shown (white circles) are the result of taking the median value of incremental metal mass data for all eleven sites, then modeling the incremental metal mass with a g amma distribution. ....................................................79 36 The incremental precipitation for the watershed, cumulative precipitation for the watershed, and cumulative watershed yield into the basin. The period of simulation was from 1Jan08 through 31July08. ..................................................................................80 37 Measured temperature values fo r the modeled year. Hourly 2008 data is shown along with daily mean temperature values for both 2008 and a twenty year mean for 19892008. The period of simulation was from 1Jan08 through 31July08. .....................80 38 The top left graph shows influent flow rates over the modeled period, effluent flow rates are shown in the top right graph. Surface overflow rates are shown in the center figures, with the time series on the left and the frequency distribution on the right. The stage/storage curves for the sedimentation basins are shown in the bottom plots. The period of simulation was from 1Jan08 through 31July08. ...............................81 39 Total effluent metal and PM mass. Values assume no partitioning of metals into the aqueous phase. Note the difference in scale where total metal mass is i llustrated in g and total mass in kg. Mass concentrations are the summation of PM (d > 25m) and across the entire simulation with a total effluent volume of 59825 m3. The period of simulation was from 1Jan08 through 31July08. .................................................82 310 The median and range of influent and effluent PSDs. .......................................................82 41 Location of sampling sites within the Lake Tahoe watershed. ........................................106

PAGE 12

12 42 Non parametric plot of dissolved and particulate bound metals. Total control snow values are shown in the dissolved metals plot as an open circle. Applicable regulations for total dissolved metals are shown in the right most plot. California Toxic Rule values, CMC (Criterion maximum concentration) and CCC (Criterion continuous concentration), are shown for Cd, Cu, Fe, Pb, and Zn. The NPDES permitted effluent limit for discharge to a collection system or surface water for Ni is also shown. .......................................................................................................................107 43 Dissolved fractions (fd) for all sites. ...............................................................................107 44 Kd (equilibrium partition coefficient) values for all sites. ...............................................108 45 Speciation results for Aluminum (top), Cadmium (center), and Chromium (bottom) for 1, 5, 10, and 25C. Snowmelt samples are shown as box nwhisker plots while control samples are shown as a scatter plot. ....................................................................109 46 Speciation results for Copper (top), Iron (center), and Lead (bottom) for 1, 5, 10 and 25C. Snowmelt samples are shown as box nwhisker plots while control samples are shown as a scatter plot. ..............................................................................................110 47 Speciation results for Manganese (top), Nickel (center), and Zinc (bottom) for 1, 5, 10 and 25C. Snowmelt samples are shown as box nwhisker plots while control samples are shown as a scatter plot. .................................................................................111 51 The locations of the six sampling sites. ...........................................................................137 52 Parametric plot of dissolved parameters: anion concentrations in mg/L of the respective anion, DOC in units of mg/L as C, alkalinity in units of mg/L as CaCO3 and pH on the right axis. Snowmelt samples are shown in box nwhisker format while control snow samples are shown as a scatter plot. .................................................137 53 SSC (total sample), TSS (supernatant sample), TDS, and turbidity for both snowmelt and control snow samples. ...............................................................................................138 54 Total, suspended, and dissolved COD the snowmelt samples, the total COD value for the control snow is also shown. The median COD dissolved fraction, fd, equals 0.05. ..138 55 Dissolved alkalinity and hardness components for snowmelt and control snow samples. ............................................................................................................................139 56 Particle size distribution for 20032005 snowmelt samples. ...........................................139 57 Speciation results for Ammonia ( top), Nitrate (center), and Phosphate (bottom), for 1, 5, 10, and 25C. Snowmelt samples are shown as box nwhisker plots, while control samples are shown as a scatter plot. ....................................................................140

PAGE 13

13 58 Speciation results for Carbonates (top), Chlorides (center), and Sulfate (bottom) for 1, 5, 10, and 25C. Snowmelt samples are shown as box nwhisker plots while control samples are shown as a scatter pl ot. ....................................................................141 59 The percent difference of temperature values with respect to the 1C value for the most dominant species. Both snowmelt and control snow samples are shown. .............142 510 Evolution of gaseous CO2 and NH3 from the samples at sea level. Snowmelt samples (n=17) are shown as box and whisker plots while control snow (n=5) samples are shown as a scatter plot. Values are shown in atmospheres and mmol/L units (calculated using ideal gas law, R = 0.08206 L atm/mol K) ..................................143 511 The frequency distribution of measured pH values in snowmelt samples. The mean, standard deviation, and count are shown in the plot. The statistics for the pH of the most affected by pH are also shown here along with the snowmelt pH distribution using equilibrium constants at 25C. The pH frequency distribution is shown in the background of the following plots for comparison purposes. ..........................................144 512 The mean and one standard deviation of the snowmelt (n=17, 8.2570.623) and control snow (n=5, 5.1380.543) samples compared to the final pH values seen if the pH was allowed to change in MINTEQ. ..........................................................................144 61 The snow sampling locations are shown in part A and the inset shown in part B illustrates the snowmelt sampling locations. The distance of flow from the upstream US HWY 50 source area snowmelt location to the downstream storm sewer sampling location is 690 m. The distance of flow from the downstream storm sewer to the outfall sampling location at Lake Tahoe is 760 m. ................................................163 62 SSC, VSSC and turbidity for source area snow (sampled as snow). Time 0 represents the unsettled snowmelt sample and settled results are determined after 60 minutes of quiescent batch settling. .................................................................................164 63 Mean and standard deviation of all nine unsettled (t = 0, T0) and settled (t = 60 min., T60) snow sample PSDs (sampled as snow). PSDs are modeled with a cumulative gamma distribution; the parameters for the model are shown. Measured and modeled (Newtons Law) settling velocity at 5C is shown in the right plot. Settling velocity for discrete PM sizes are compared to model results utili zing Newtons Law. .................................................................................................................164 64 Incremental PSDs and incremental concentrations of PM for all nine source area sites. (Sampled as snow.) .................................................................................................165 66 The role of snowmelt sample holding time on the gamma distribution parameters ( ) and the d50m index. There is no stastically significant difference ( = 0.05) between the slopes of the trend lines and zero indicating no significant coagulation or flocculation during snowmelt holding time. ....................................................................166

PAGE 14

14 68 The change in and d50M with increased settling time for the source area pavement snowmelt and storm sewer snowmelt samples. (Samples collected as snowmelt.) ........................................................................................................................168 69 Influent and effluent PSDs for snow source areas (sampled as snow), the downstream (gutter snowmelt flow, sampled as snowmelt), and storm sewe r outfall to Lake Tahoe (sampled as snowmelt) are shown in the left column. Effluent PSDs calculated by utilizing Newtons law (5C, specific gravity = 2.60) to determine settling velocity. The role of Type I settling utilizing surface overflow rate to determine effluent concentration is shown in the right column (based on a surface area of 1,000 m2) for influent PSDs and concentrations at each location. .......................169

PAGE 15

15 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy GRANULOMETRY, PARTITIONING AND SPECIATION OF URBAN SOURCE AREA SNOW P OLLUTANTS GENERATED FROM VEHICULAR TRANSPORTATION AND INFRASTRUCTURE By Natalie Magill December 2009 Chair: John Sansalone Major: Environmental Engineering Sciences Vehicular transportation systems and impervious pavements are sources of contaminants such as particulate matter (PM), metals, nutrients, petroleum products, deicing and anti skid agents, litter, as well as anthropogenic chemicals, litter and gross solids. In this study, snow impacted by pavement and vehicular transport within the Lake Tahoe watershed was examined. PM was analyzed as a function of particle diameter and particle size distributions (PSDs) were modeled as cumulative gamma distributions. The sediment fraction (dp >75m) dominated the PSD, for which the gravimetric mean fract ion represented 92% of the entire gradation. Results indicate Al (15 g/kg) and Fe (4.2 g/kg) are the highest PM based concentrations; Cd (0.18 mg/kg) and As (4.9 mg/kg) are the lowest. When modeling particulate bound metal PSDs with a gamma distribution no statistical difference in site means was found for the scaling parameter, indicating that the distribution of the metal mass for the PSDs did not significantly vary between metals. Results indicate that Type I sedimentation is capable of separating the sediment fraction (> 75 m) and majority of metal mass. Metal partitioning between particulate bound and dissolved fractions indicated that metal elements examined are predominately bound to PM (as suspended PM) with median particulate bound fractions (fp) greater than 0.80. Speciation

PAGE 16

16 indicates that divalent ionic concentrations decrease with increasing temperature while carbonates and metal carbonate complexes increase with increasing temperature. Median and mean values for total nitrogen and phos phorus do not exceed the permit limits, though the maximum values for the data set do exceed the limits. Suspended snowmelt PM is dispersive and does not demonstrate Type II (flocculant) settling behavior up to 96 hours even with PM concentrations exceedin g 10,000 mg/L. A treatability study of the entire gradation of PM indicates that surface loading rates below 6.3x104 mm/s are required to treat suspended snowmelt PM to levels required as a filter influent. While the basin is effective at separation of coarser PM based metals, results illustrate that extended sedimentation will not reduce PM, PSDs and turbidity to discharge levels for Lake Tahoe without secondary treatment; for example a combination of coagulationflocculation followed by filtration.

PAGE 17

17 C HAPTER 1 GLOBAL INTRODUCTION Lake Tahoe is an oligotrophic lake located on the border of California and Nevada that has shown decreasing clarity and increasing eutrophication since 1967 (Jassby et al. 1995). The decreased clarity is due to a combination o f mineral inputs from the watershed and phytoplankton (and phytoplanktonderived materials) (Jassby et al. 1999). Within a 20 year period, the lake shifted from nitrogen to phosphorus limited with respect to algal growth (Goldman and Hoffman 1975, Goldman et al. 1993, Hatch et al. 2001). NPDES (National Pollutant Discharge Elimination System) discharge limits to Lake Tahoe currently exist for suspended solids, turbidity, total iron, total phosphorus, and total nitrogen (USEPA 2003). In the Lake Tahoe are a, six locations along highway US 50 were sampled. Highway US50 follows the Lake Tahoe shoreline through both California and Nevada. A portion or all of these sites were sampled over five winter seasons occurring in January 2003, December 2003, December 2004, December 2005, and December 2008. These sites represent samples from two different states which use different winter maintenance practices across different land uses: roadside pavement samples, parking lot samples, and a local snow dump site. In the city of South Lake Tahoe, CA, snow maintenance practices include the collection and transportation of urban snow to a local storage basin. The snow remains in this storage area and allowed to melt, the melt water flows through a series of two sedimentation basins followed by infiltration into the subgrade (Sansalone et al. 2003a). Winter maintenance practices along US Highway 50 in Nevada include the application of de icing salts and de icing sands. Deicing salts are obtained through state bids and ar e required to be solid sodium chloride with a maximum size of 0.5 inches and 95% pure sodium chloride. The rock salt must be free from dirt and other foreign material. Deicing salts are tested for total

PAGE 18

18 phosphorus, total nitrogen, TKN, iron, moisture content, and percent of NaCl. Winning bidders are required to deliver the material to the requested stockpile location. De icing sands are also obtained through state bids. Requirements for de icing sands include a durability index or hardness of greater t han 75, loss of abrasion of less than 33%, and total phosphorus content of less than 10 parts per million (ppm). The particle diameter requirements are listed below in Table 11. This information can be found in the bids documents for the state of Nevada and is located in the supporting documents. A few terms utilized throughout the following chapters include particle, sand, and particle size distribution. Within this study, a particle is defined as a body of finite mass and structure; a solid phase mater ial for which there are no cohesive forces between another solid phase material under dry conditions. A particle can be a mineral grain of a rock fragment of common rock forming minerals and is typically described in terms of size and shape. Size definitions used here are from American Society for Testing and Methods (ASTM) definitions. For example, sand can be defined either chemically or physically. To avoid confusion sand is utilized to describe a size classification (i.e. sand size) within this do cument. These definitions are described in more detail in Table 1 2. Particle size distributions used primarily in Chapters 2, 3, and 6 are defined as a gravimetric distribution as a function of particle diameter. Sources of particulate matter include atmospheric deposition, vehicular deposition, and applied grit and sand (Oberts et al. 2000). Compared to stormwater suspended PM concentrations, corresponding values found in snowmelt can be several orders of magnitude higher (Sansalone and Buchberger 1996, Sansalone et al. 2003b). The highest concentrations in snowmelt are seen during rain onsnow events (Westerlund et al. 2003). The quality of urban snowmelt is dependent on highly variable factors such as snow residence time, vehicular traffic

PAGE 19

19 loadings, and snow maintenance practices and results in the need for a comprehensive physical and chemical characterization of the PM. Sources of metals on roadways include wet and dry deposition, traffic activities, and deposition of deicing salts, leaching from infrastructure, and highway construction or repair (Field and Sullivan 2003, Sansalone and Glenn 2002). E xample s of metal deposition from traffic activities include cadmium from brakes and tires; copper from brakes and tires; lead from brakes, tires, fuels and oils, and de icing salts; and zinc from brakes, tire s, and frame & body (Huber et al. 2001, Sansalone and B uchberger 1996, Glenn and Sansalone 2002). Acidic snowmelt has been shown to decrease pH and increase concentrations of dissolved metals in receiving water bodies (Servos et al. 1987). Roadway runoff may contain high heavy metal loadings that can negative ly impact receiving waterbodies, aquatic life, and the food chain (Glenn and Sansalone 2002). These metals partition between dissolved and particulate phases and the dissolved metals can speciate into various complexes. Speciation is the first step in de termining the toxicity or bioavailability of species or the negative effects on biota (Morrison 1989). The mobility of metals in the aquatic environment is also affected by maintenance practices that include the application of deicing salts containing sodium chloride (Nelson et al. 2006). Deicing salts have also been shown to contain metal elements such as zinc, copper, nickel, arsenic, and cyanide (Goldman and Hoffman 1975, Sansalone and Glenn 2002). Dissolved metals are often difficult to remove; adsor ptive filtration and other advanced treatment methods are typically required (Sansalone and Teng 2005). Passive infiltration systems have been shown effective in rural areas, but inefficient in urban areas where metals tend to sorb to PM (Oberts 2003, Sansalone and Buchberger 1996).

PAGE 20

20 Snowmelt nutrient inputs into Lake Tahoe are of a significant public concern to prevent further eutrophication of the lake. C oncentrations of nitrate and ammonia in snow can represent a significant amount of the input bioavail able nitrogen to the soil during generation of snowmelt and the relative concentrations will vary depending on meteorological conditions and biological acti vity within the snowpack (Jones 1999). The shift from nitrogen to phosphorus limitation in Lake Tahoe is credited to atmospheric deposition of nitrogen, causing an increased importance on reducing phosphorus inputs into the lake (Hatch et al. 2001). The use of deicing salts can result in both structural and ecological issues, aiding in the corrosion of infrastructure and negatively impacting roadside vegetation and receiving waterbodies (Amrhein et al. 1992). The quality of urban snowmelt is highly variable; it is a function of residence time, traffic activity, and geography. Samples taken from the same general area show significant differences in the amount of PM contained within the snowpack and a large range of PSDs. Sedimentation methods have been shown to be extremely effective in reducing solids loadings, both as suspended solids concentration (SSC) and turbidity. Both turbidity and suspended solids have been regulated for discharge into a water body in the Lake Tahoe area with NPDES permits (USEPA 2003). Turbidity poses aesthetic, filterability, and disinfection concerns and the allowable limit is 20 NTU (Sawyer et al. 2003, USEPA 2003). Suspended solids pose similar concerns, along with a toxicity threat due to the particulate bound metals, and are regulated at 50 mg/L for discharge into a local water body (Viklander et al. 2003, Sansalone and Buchberger 1996, USEPA 2003). Snowmelt is a vital source of water for the western United States and its economy, supplying agriculture, industry and domestic supply; it also supplies water to forests for growth and waterways for fish and wildlife (Feth et al. 1964). While snowmelt runoff can exhibit some

PAGE 21

21 of the highest pollutant loadings, it is also identified as one of the most difficult non point sources of pollution to manage (Oberts 2003) Current winter maintenance practices include plowing roadw ay snow into snowbanks and the application of deicing or antiskid compounds. Common treatment methods for snow include allowing the snow to melt in place, collecting and transporting the snow to a locality and allowing it to melt, or collecting and dumpin g into local water bodies depending on local pollution control regulations. Additional treatment of snowmelt is hindered due to ice formation, inhibited biological activity, frozen soils, and chemically altered water quality behavior due to the increased c oncentrations of sodium chloride (Oberts 2003). This is exacerbated by the fact that early melts generally have high soluble content, while either late melts or rain onsnow events result in high solids content (Oberts 2003). Treatment of high chloride waters is unrealistic with the exception of dilution, so the only possibility to reduce chloride levels in snowmelt would be to limit the amount of chloride applied to roadways (Oberts 2003). Treatment of snowmelt with detention or filtering is problemat ic because there is a risk of the snowmelt in the basin or filter freezing (Oberts 2003). Any treatment option requires careful characterization of the physical and chemical nature of urban snowmelt. Nomenclature ASTM American Society for Testing and Materials NPDES National Pollutant Discharge Elimination System PM Particulate matter ppm parts per million PSD Particle size distribution SSC Suspended solids concentration TKN Total Kjeldahl Nitrogen USEPA United States Environmental Protection Agency

PAGE 22

22 Table 1 1. Percent of material passing a specified sieve size required for de icing sands in Nevada as determined by ASTM. Sieve Size Particle Diameter (mm) Percentage Passing % #4 4.76 93 100 #8 2.38 40 80 #16 1.19 15 60 #50 0.300 0 4 #100 0.150 0 4 #200 0.075 0 2.5 Table 1 2. Table of Grain size classifications from ASTM. Size Classification Upper Particle Size (mm) Lower Particle Size (mm) Boulders 1000 1000 Cobbles 300 300 Gravel 75 75 Coarse Sand 4.75 2 Medium Sand 2 0.425 Fine Sand 0.425 0.075 Silt 0.075 0.005 Clay 0.005 0.001

PAGE 23

23 CHAPTER 2 NONCOLLIODAL GRANULOMETRY OF URBAN SNOW SNOWMELT GENERATED FROM TRAFFIC LOADINGS IN THE LAKE TAHOE BASIN Introduction Anthropogenic activities such as traffic generate PM and pollutants in the built environment (Sansalone et al. 2003b). This anthropogenic material can be incorporated into fallen snow, for example on a pavement surface, or entrapped in plowed snow that ma y form a physical barrier alongside a roadway. Typical snow maintenance practices include plowing snow into banks along the roadside and application of traction or anti skid material such as sand or grit and/or de icing salts (Glenn and Sansalone 2002, Oberts 2003). Snow banks function effectively as an accumulator of PM generated by urban traffic activities over an extended period of time (Dean et al. 2005, Sansalone et al. 2003a, Sansalone et al. 2003b, Westerlund and Viklander 2006). The accumulation of PM and pollutants, the accumulative or plowed geometry of the snow as well as the residence time of the pollutant and PM in the snow influence the partitioning of pollutants to or from PM (Sharma and McBean 2002). Sources of PM include atmospheric dep osition, dry deposition from traffic sources, biogenic sources and applied grit and sand (Oberts et al. 2000). The relative contributions are dependent on land use; in urban areas loadings are correlated to distance from the source area, for example roadw ays or industries, as well as the residence time of the snow bank (White et al. 1995). With respect to vehicular deposition 4050% of the PM are generated from pavement wear, 2030% from tire wear, 15% from vehicle part abrasion and less than 3% of total particles is generated from outside the right of way as atmospheric deposition, on a gravimetric basis (Sansalone et al. 2003b). With respect to land use, snow from residential areas is typically less polluted than snow from commercial areas (Zinger and D elisle 1988) which is less pollutant than roadway snow. Some studies have shown that the extent of snow, snow bank, and soil pollution along

PAGE 24

24 highways is typically restricted to a 10 m wide band from the pavement edge, while other studies have demonstrated that the combination of snow snowmelt and rainfall runoff can result in a much wider distribution of soil pollution transitioning to background or surrounding levels (Novotny et al. 1998, Teng et al. 2002). Snowmelt suspended PM concentrations can be several orders of magnitude higher when compared to stormwater values (Sansalone and Buchberger 1996, Sansalone et al. 2003b). For instance, a case study in northern Sweden reported particle number density (PND) for snowmelt at values eight times higher than in rainfallrunoff (Westerlund and Viklander 2006). While snow snowmelt can exhibit high pollutant loadings, this is also one of the most difficult challenging non point sources of pollution to manage, in particular in the solid phase (Oberts 2003). Studies of snow snowmelt PM have focused on a variety of PM characteristics (Kim and Sansalone 2008). For example, a Cincinnati study examined the accretion and elution of PM in snow during the snowmelt process (Sansalone et al. 2003b). Studies have also dem onstrated the potential of clarification for the treatment of snowmelt PSDs from a snowmelt detention basin (Sansalone et al. 2003a). Other researchers focused on atmospheric pollutant deposition in an urban snowpack (Sharma and McBean 2002) or compared P ND in snowmelt and rainfall roadway runoff (Westerlund and Viklander 2006). Other researchers focus on the toxicity of snow samples as a function of distance from roadways (White et al. 1995). Researchers have studies the effect of de icing compounds on the chemicals in snowmelt. For example, a number of researchers have identified the presence of cyanide in snowmelt due to salt anti caking agents (ferrocyanide) (Novotny et al 1998, Sansalone and Glenn 2002). The effect of the type of de icing materials and length of winter conditions on snow chemistry was studied by comparing two Swedish cities with similar traffic loads (Reinosdotter and Viklander 2005). Swedish

PAGE 25

25 municipalities have tried snow separation strategies to minimize the negative effects of s now dumping in receiving water bodies. For example, heavily polluted snow is transported to a locality holding facility, essentially a detention basin which is operated to all snow melting and sedimentation thereby minimizing negative environmental effects (Reinosdotter and Viklander 2006). The c haracteristics of this snow in terms of turbidity, suspended solids, and 26 other index parameters w ere analyzed in Montreal, where snow dumping is a favored method of removal, but only a temporary alteration in w ater chemistry was identified (Zinger and Delisle 1988). Studies on the effect of snowmelt on the receiving waters during periods of base flow and snowmelt have included measurements of total dissolved solids (TDS) and PM as total suspended solids (TSS) a s a function of flow rate, and have shown that PM and nutrient transport is correlated to variations in the snowmelt hydrograph (Anderson et al. 2000). The impetus of this study was to provide a comprehensive examination of urban source area PM from a maj or transportation land use traversing the Lake Tahoe Basin for the purpose of provide granulometric parameters for in situ UOP designs loaded by such source areas. Objectives There are seven objectives of this study focused on measuring and modeling the gr anulometric parameters for transportation source areas in the Lake Tahoe Basin. The first objective was to examine PSDs of transportation source area urban snow PM and model the PSDs. The second objective was to determine particle density across the PSD. The third objective was to measure the specific (SSA) and total surface area (SA) distributions across the PSD. The fourth objective was to measure the surface charge and total surface charge distributions across the PSD. The fifth objective was to det ermine the fractions of suspended, settleable, and sediment for each source area site and sampling season. Given the mean temperature values of the region, the sixth objective was to measure the settling velocity of PM

PAGE 26

26 across the PSD at water temperatures below 10C, examine the influence of water temperature on the settling velocity of this PM and compare to modeled settling velocities. The final objective was to utilize measured settling velocities in a treatability study, to determine removal efficiencies of the basin under real flow and PM conditions. Background Sedimentation is one of the oldest and still the most common method of separat ing PM from liquids. Assuming Type I settling of PM and using a force balance on a particle settling at equilibriu m the settling velocity of a particle can be modeled with Newtons Law, applicable for all laminar, transitional and turbulent settling regimes. For Reynold s numbers less than approximately five ( a laminar settling regime ) Stokes Law can be derived from Newtons Law with a constant drag coefficient, Cd. Newtons law is valid up to a Reynolds number of 104. (Fair et al. 1968, Reynolds 1996). Newtons Law p p d p w w p d t pd 1 sgC 3 g 4 d C 3 g 4 V ( 21) Coefficient of Drag 0.34 Re 3 Re 24 Cd d d ( 22) Reynolds Number d V d V Rep p w p p d ( 23) Stokes Law 18 d 1 sg g 18 d g 2 p p 2 p w p p ( 24) Filtration is a UOP that is commonly required downstream of primary or secondary clarification to separate finer settleable size and suspended PM. Granulometric parameters including PSD particle density, surface charge and gravimetric or PND are input parameters for filtration UOPs.

PAGE 27

27 Coagulation/flocculation (C/F) is generally used prior to sedimentation or filtration. The PM surfaces have some combination of an amphoteric charge and for clay minerals such as smectites a permanent change due to isomorphic substitution. These charges can be negative for many PM minerals and help sustain suspended matter and colloidal PM in suspension through repulsive forces. The stability of the fine suspended and colloids is generally measured as the electrostatic potential of the shear plane of the particle or the zeta potential (Kim and Sansalone 2008). Adsorption UOPs entail the use of an adsorbent to remove a given substance from a solution by e ither physical or chemical means. Zeta potential D qdL 4 ( 25) Methodology and Laboratory Analysis S ampling Locations Frozen snow samples were collected from six source area sites over four winter seasons. Multiple years of data are available from three of the sites, while the other three sites were only sampled once due to inaccessibility. Table 2 1 lists the site numbers, sampling dates and locations. The National Weather Service Cooperative Observer Program (COOP) public acces s database at the Western Regional Climate Center was the source for climate data for the region (WRCC 2002). Data w ere compiled from nine stations including mean high and low monthly temperatures and monthly total precipitation and snowfall. Traffic data w ere compiled for highway US 50 in the Lake Tahoe region for both California and Nevada. Data were taken over the course of a year, with no bias to season. California data consisted of a 21.4 km section from Echo Lake Road to the state line. Peak t raffic values ranged from 14,700 to 43,500 vehicles a day and collected in 2004 (Transportation

PAGE 28

28 Management Association 2002). Nevada data consisted of a 8.0 km section from the state line to Zephyr Cove ranging from 14,800 to 32,100 vehicles per day and w as collected in 2000 (NDOT 2006). S ampling Method Snow samples were taken from plowed roadway snow banks or parking areas snow piles. Samples were taken from 6 source area sites listed in Table 2 1. The snow banks and piles were generated from pavement plowing. Snow was packed into 4 L wide mouth polypropylene (PP) bottles. Sampling consisted of taking a width of the cross sectional area of the snow banks until sufficient volume was obtained to fill the 4 L bottles. Samples remained frozen during tra nsport to the laboratory and stored in the freezer until analyzed. Samples were allowed to melt at room temperature for 12 hours before beginning analysis S uspended, Settleable, and Sediment Fractions The snowmelt samples were well mixed and poured int o 1L Imhoff cones to remove the settleable PM from the supernatant. After 1 hour of quiescent settling, the settleable PM concentration as mL/L was determined. The supernatant was decanted into 1 L wide mouth PP bottles and stored for suspended PM analy sis, leaving the settleable solids in the base of the Imhoff Cone. The settleable PM (operationally defined as 75 < d< 25 m) was removed from the Imhoff cones and placed into Petri dishes and dried at 40C. The sediment PM (operationally defined as d> 75m) was removed from the 4 L PP bottles and dried in the oven at 40C. For each site the PM from the duplicate samples were composited to ensure sufficient mass for PSD analyses, with the exception of the 5 January 2003 samples.

PAGE 29

29 Particle Size Distribu tion (PSD) Analysis A dry mechanical sieve analysis method was utilized to determine the particle size distribution of the settleable and sediment fractions of the samples. The method of analysis was ASTM D422 modified to include a greater number of siev es (Sansalone et al. 1998); all samples were prepared using ASTM D421 (ASTM 1990, ASTM 1993). The wet slurry of both the settleable and sediment fractions were dried at 40C and desiccated. The sample was disaggregated with a mortar and pestle before th ey were sieved through a series of 17 sieves ranging from the 9750m (#3/8) to 25m (#500) and a pan to collect particles smaller than 25m. The sieves were agitated for 7 minutes with a 2 mm amplitude. The dried and desiccated PM was weighed for an i nitial mass before sieving, the PM remaining on each sieve was collected and weighed, and a mass balance was calculated for quality control to ensure that the criterion of 2% was met. The percent of mass finer was plotted against the particle diameter after all mass values were normalized to 1000g. A cumulative gamma distribution was utilized to model the PSDs on a gravimetric basis. The gamma distribution was fit by optimizing the gamma and beta parameters to minimize the SSE (sum of square errors) between the measured and modeled curves. Reflected cumulative mass distribution function d 1 e d 1 1 d; f 1 ( 26) The granulometric indices including the geometric mean, central tendency, skewness, kurtosis, and uniformity were calculated for each site using E quations 27 through 2 12 (Ying and Sansalone 2008). Geometric Mean 84 16 gd d d ( 27) Standard De viation of the Geometric Mean 16 84 gd d ( 28)

PAGE 30

30 Central Tendency 50 2 50d log ( 29) Skewness 5 95 50 95 5 16 84 50 84 16 k2 2 2 2 S ( 210) Kurtosis 25 75 5 95 G2.44 K ( 211) Uniformity 6.6 4 5 95 16 84 I (2 12) Particle Density of PM Dried PM separated into size ranges was desiccated overnight to remove any remaining moisture and humidity P article density of the PM was analyzed with a helium gas pycnometer using ASTM D5550 (ASTM 1994). This procedure measures the true volume of the sample using a series of pressure measurement s a known dry mass of PM, a calibrated cell volume and the ideal gas law. Helium gas was used to infiltrate the smallest pores of the particle (Lowell and Shields 1991). The particle density is calculated by dividing the PM volume as measured by the pycnometer by the sample mass. Specific Surface Area and Total Surface Area An EGME (ethylene glycol monoethyl ether, HOCH2CH2OCH2CH3) method was used to determine the surface area (SA) (Carter 1986, Sansalone et al. 1998,). The ratio of the surface area of a particle to the mass or volume of the particle, specific surface area, was calculated from the SA data (Carter 1986, Sansalone et al. 1998, Arnepalli et al. 2008). This method measures both internal and e xternal surface area (Arnepalli et al. 2008). Granular activated carbon (GAC) was used as a control PM (1100m2/g, Calgon 1995). Each gradation was measured in triplicate with a minimum of 0.2 g of sample for each replicate. For grains of larger diamet er, a minimum of 7 separate grains were analyzed. The measured SSA was multiplied by the measured mass of

PAGE 31

31 that particle size to determine total surface area (SA), and indication of the capacity and potential for surface complexation within ionic solutions For plotting purposes, all SA values were normalized to 1000 g. g Sample of Mass m g 0.000286 g Sample on EGME of Mass SSA2 ( 213) im SSA SAi i ( 214) Surface Charge and Point of Zero Charge (PZC) Surface charge was determined using a modified potentiometric titration with PM of a given size suspended in solutions of differing pH values. A known mass of PM, nominally 0.5 g, were suspended in a 0.01M potassium chloride buffer solution and agitated for 24 hours at a known initial pH ranging from 5 to 10 and then allowed to settl e overnight. The pH values were chosen to both mimic natural systems and to surround the PZC of the samples. A portion of the solution was then titrated back to the original pH with either 0.001M HCl or KOH solution (Van Reij and Peech, 1972). The surfa ce charge was measured for multiple pH values, typically ranging between 5 10, and the PZC was calculated for each PM size, using a linear regression between the positive and negative data points. When adequate sample mass was unavailable to run each PM size individually, PM sizes were combined and the weighted mean of the individual SSAs were used to determine the surface charge of the composite PM size range. All samples were run in triplicate; blanks were run for each pH value and buffer solution prepared. The surface charge of a particle is needed to determine the potential for solutes to complex to the surface of the PM.

PAGE 32

32 6 Sample Total Titrated Blank Base Base Acid Acid Sample Baee Base Acid Acid i10 1000 W V V V M V M V M V M C (2 15) i i iSSA C SC (2 16) The pH value where the negative and positive surface charges of t he particle balance is the point of zero charge (PZC). This value was calculated with a linear interpolation between the two surface charge points closest to zero but on opposite sides of zero. Settling Velocity Settling velocity was analyzed with two measurement methods; both utilized settling columns that were 0.10 m (4 in) in diameter and 0.91 m (3 ft) tall with a sampling port located 0.15 m (6 in) from the bottom of the column. The column setup remained refr igerated at a constant air and water temperature. The first and simplest method was the visually observable method that utilized PM greater than 100 m; these particles were visually observable settling in the Plexiglas column. The settling velocity was measured by measuring the time the particle took to fall a given distance in the column. The velocity was determined by dividing the distance by the time. A second method was used for PM smaller than 100 m that were not visually discernable. This PM was pre soaked in DI water and cooled in the same refrigerator. The slurry was added to the top of the column and samples were taken from a sampling port 0.15 m (6 in) from the bottom of the column at known sampling intervals depending on the PM size range o f interest. The sampling intervals along with the change in water level as the samples were taken were used to calculate the mean settling velocity of a particle within the sample. The samples were analyzed with a laser diffraction system capable of meas uring PM with a range of 0.1 to 2000m. The frequencies were plotted as a function of settling velocity and a log normal

PAGE 33

33 distribution was fitted to the data to determine the peak for that particle size fraction The measured PSD and settling velocity data were modeled assuming spherical Euclidian particle geometry with a measured particle density, assumed to be constant for all particles of that size at temperatures of 1, 5, and 10 C using Newtons Law. Treatability Study To model the solids removal efficiency of the existing South Lake Tahoe basins, climate data for the region were obtained from the South Lake Tahoe airport, a distance of approximately 3 km from the basin. Daily total precipitation data were collected for time period for 20012008. Daily rainfall totals were plotted in a frequency graph; all daily totals over 2.54 cm (1 inch) of rain were analyzed for rainfall intensity, to determine which year showed the highest intensity rainfall events. The 2008 data were chosen because it showed the highest intensity storms, which would maximize the volume of flow into the basin. Climate data were obtained at five minute intervals for the region surrounding the existing treatment facility from January 1 through July, 31 2008. Basin characteris tics, temperature and rainfall data were input into the EPA Storm Water Management Model (EPA SWMM Version 5.0) to model rainfall and snowmelt runoff flow rates into the existing sedimentation basin (Rossman 2007, Huber and Dickinson 1988). The sedimentat ion basin has a maximum volume of 975 m3. Hazens theory (N=1, Continuous stirred tank reactor, CSTR) (Malcom 1989, Sansalone 2003a) was then used to determine the solids removal efficiency of each particle diameter and summed over the PSD weighted by the mass concentration of each diameter for the existing sedimentation basin using real time flow data and a surface area of 550 m2. In this expression Vp is the measured particle settling velocity, SA is the sedimentation basins surface area, Q is the efflu ent flow rate, and N

PAGE 34

34 is Hazens N used to quantify the extent of the nonideal conditions of the basin (Hazens N=1, Continuous stirred tank reactor, (CSTR)) (Malcolm 1989, Sansalone 2003a). Hazens Law N PN Q SA V 1 1 (%) Removal (2 17) Within the EPA SWMM model, the snow storage basin was given an area of 2.06x104 m2 (5.1 acres), with an external watershed of 1.46x105 m2 (36.1 acres) also draining into the basin. The storage basin was assumed to be completely impervious with a slope of 0.94%, and the exter nal basin was set at 5% imperviousness and a 5% slope. The slope and imperviousness data is utilized in the model the overland flow of rainfall and snowmelt. The external watershed was found to contain 4 soil types classified as loamy coarse sands with s lopes ranging from 0 to 9%: Christopher loamy coarse sand, Christopher Gefo complex, Jabu coarse sandy loam, and Marla coarse sandy loam (Soil Survey Staff, 2008). The soil types in the area are somewhat excessively drained soils (87.7% of the watershed) with the exception of the poorly drained Marla coarse sandy loam which occupies only 6.5% of the external watershed. All of the soil types have frost free periods ranging from 20 to 90 days per year. For the external watershed, infiltration was calculate d using GreenAmpt parameters and an aquifer set just below the ground to minimize infiltration and maximize runoff to the storage basin for a more conservative estimate. The efficiency of the basin was calculated on both a mass and particle number basis (Cristina et al. 2002). While the majority of the PSDs were analyzed with PM greater than 25 m, a median number based PSD with d > 1 m was also utilized. This gradation was produced by modeling the existing cumulative PSD (d > 25 m) with an exponential decay curve to determine the fractions in the range of 1 < d < 25 m.

PAGE 35

35 R esults R elative M ass of S uspended, S ettleable and S ediment P articles Shown in Table 22 is the summary of the relative fractions of the suspended, settleable and sediment PM for each site. S ediment (>75 m) PM gravimetrically dominate these results, ranging from 8 3 to 99% with a mean of 92% The settleable PM ranges from 0.78 to 14% with a mean of 6%, while suspended PM fraction ranged from 0.1 to 8.2% with a mean of 2.2%. The total aqueous volume and total mass associated with each sample are also summarized in Table 22. Particle Density Figure 21 illustrates the particle density as a function of particle size for all sites, along with the mean for both the sediment and settleable. There was not sufficient material available for measuring the density of suspended solids in the majority of the samples, with only 3 of the 11 samples retaining sufficient mass in the pan. In this study, the particle density ranged from 0.29 to 3.60 g/cm3 across the PSD. Density values below 1 were found in two samples; both were in the largest gradation (9500 m) and were composed of biogenic matter such as leaves, bark and twigs. Excluding the largest gradation (and clearly discernable biogenic material) the minimum particle density measured was 1.97 g/cm3. Sediment sizes had a mean density of 2.64 g/cm3, while the settleable PM had a mean density of 2.65 g/cm3. These values are lower than measured values for Cincin nati, OH snowmelt PM with sediment PM sizes having a particle density of 2.75 g/cm3 and 2.86 g/cm3 for settleable PM (Cristina et al. 2002 ), but higher than rainfall runoff measured values for Baton Rouge, LA with sediment, settleable, and suspended PM sho wing mean values of 2.14, 2.24, and 2.40 g/cm3, respectively (Lin et al. 2008).

PAGE 36

36 PSD Analysis Figure 22A s ummarizes gravimetric results of mechanical sieve analysis modeled with a gamma distribution. While the accreted PM is hetero disperse the total dry mass is dominated by PM greater than 250 m. For all sites the d50 (Table 2 3) value ranges from 27 7 to 1389 m, with a median of 604 m. P M larger than 2000 m (typically considered gravel) had a mean of 18.5% and ranged from 3.1% to 33.9%. The sand range of 75 to 2000 m dominated the samples with 65.788.9% of the total mass falling into this sandsize range. For suspended PM, mass fractions were typically below 10% of the total mass of PM, with only one site at 11.5%. This indicates that the suspended material is generally less than 10% of the tot al mass for each sample. The geometric mean, central tendency, skewness, kurtosis, and uniformity of the samples are summarized in Table 2 3 along with descriptions of the indices. Geometric means range from 285 to 1334 m with central tendencies ranging from gravel to medium sand. Nine of the eleven sites show symmetry when skewness was measured and seven of the eleven sites show mesokurtic values when kurtosis was measured. The uniformity index showed that all eleven sites are very well sorted (hetero disperse), see Table 2 3 for descriptions. Individual site PSDs were successfully modeled with a cumulative gamma distribution along with a median PSD for all sites, the model parameters are shown in Table 2 4. The mass balances for all sites were less than 0.60%, indicating that less than 0.60% of the initial mass was lost during the sieving process. S SA and Total SA Calculating SSA of PM typically requires three assumptions: constant specific gravity, spherical particle shape, and solid particles (no porosity). For biogenic or anthropogenic particles, these assumptions are rarely valid for snowmelt PM as demonstrated in Figure 2 3.

PAGE 37

37 Measured SSA is typically 1000 times higher in magnitude when compared to calculated SSA. These results are similar to that for rainfallrunoff (Sansalone et al. 1998). Figure 2 2 summarize the incremental SSA (Figure 2 2B) and total SA (Figure 22C) along with the cumulative total SA as a percent. Total SA is normalized to a 1000 g dry mass basis. The measured SSA vari es significantly with most values exceeding 10 m2/g, one hundred times larger than the largest calculated SSA for spherical particles, while some PM sizes demonstrate values approaching or exceeding 50 to 60 m2/g. SSA typically increases with decreasing p article size. However with respect to total SA this trend reverses (after the SSA is multiplied by the mass for each PM size) to obtain the total surface area of the sample. While SSA is a function of the individual PM size (smaller particles have higher SSA by definition) the total SA is a function of the mass distribution across a hetero disperse PSD. The majority of the total SA is associated with the sediment PM, only because the sediment PM dominates the source area loads. For nine of the eleven si tes, over 80% of the total SA was associated with sediment size PM. The 2004 sample from site 4 did not have sufficient PM to examine PSDs less than 75 m, and the 2005 sample from site 6 had 77.3% of the total SA associated with sediment size PM. On ave rage 75% of the total surface area is attributed to particle sizes with diameters greater than 151 m (data range of 80 to 398 m ) It is common to find total SA for a single PM size from 425 to 2000 m to exceed 1000 m2 per kg of total PM sample. The in dividual sites cumulative total surface area was successfully modeled with a cumulative gamma distribution, along with a median SA PSD for all sites. A summary of the model parameters are also shown in Table 2 4. Figure 2 4 illustrates the surface area found in sampled PM with four grains chosen to shown the range of surface area found in the 1103B sample.

PAGE 38

38 Surface Charge and PZC Figure 2 5 depicts the variations of surface charge across all sites for particle diameters of 425 m and 75 m. The mean and variability in the PZC is also shown in both graphs. Figure 26 summarizes the PZC for each sampling event. Horizontal range bars were used to designate when adjoining gradations were combined for surface charge analyses. The values range from 7.00 to 9.65. The mean value for se diment PM was 8.28 while the mean for s ettleable solids was 8.24. The 05 Jan uary 20 03 sampling eve nt demonstrated the highest values of all the events while the other 3 events showed similar ranges and mean values Settling Velocity Figure 2 7 compares measured and modeled settling velocities for PM at temperatures below 10 C. Calculated settling ve locities were determined at 1, 5, and 10 C using the mean measured particle densities from Figure 2 1. The model consistently overpredicts the settling velocities for coarse particles (d > 1000 m). Settling velocities for particles with diameters smalle r than 1000 m are in good agreement with calculated values. Modeled values show very little deviation across the PSD with respect to temperature differences. Treatability Study The typical temperature profile for t he region is shown in Figure 28. The increase in temperatures to average above 0C in late February coincides with the increase in fl ow rates also seen in Figure 2 8. The flow rates range from 0.0003 to 0.0645 m3/s, with median flow values of 0.0075 and 0.0096 m3/s for the influent and effluent respectively. This is not to imply that the basin increases the flow rate, only that the periods of no flow were removed when calculating the statistics. The graphs also show that the basin will reach capacity and steady state by March.

PAGE 39

39 The treatabi lity study shows that the primary sedimentation basin is very effective at removing PM from snowmelt, with the basin operating at over 97% removal efficiency when calculated on a mass basis as seen in Figure 2 9 (left). A similar trend is seen when the re moval efficiency is calculated on a particle number basis with over 96% removal (Figure 2 8, right). Both plots i n Figure 2 9 consider particle sizes greater than 25 m. When PM smaller than 25 m (and greater than 1 m), the removal efficiencies drop c onsiderably when calculated on a number basis, and the dependence of the removal efficiency on flow rate is shown in Figure 210 showing a median removal efficiency of 4.39% with a range of 3.4 to 27.7%. The range of influent and effluent PSDs for the bas in is shown in Figure 211. C onclusions This paper focused on analyzing the characteristics of a comprehensive gradation of particulate matter found in urban snowmelt. Representative sampling was undertaken across varied geographic locations, over a period of four winter seasons. PM was analyzed in order to characterize the particle density, surface area, surface charge, and settling velocity across the gradation. PM with an average diameter greater than 75 m, classifiable as sediment, was found to predominate on a dry mass basis ranging from 83 to 92% of total PM. While all sites exhibited particle size distributions that were hetero disperse, the majority of total PM (65.7 to 88.9 % of total PM) could be classified as coarse sand. The average par ticle density was in the range of 2.64 to 2.65 g/cm3 for sediment and settleable PM respectively. It was also found that particles with diameters in the range of 25 to 75 m had a slightly higher particle density, compared to particles coarser than 75 m. This result is consistent with previous studies of particle density of urban PM in Cincinnati and Baton Rouge (Cristina et al. 2002, Lin et al. 2008). It was found that the SSA of PM generally increases with decreasing particle size, but the

PAGE 40

40 majority of th e SA is associated with larger particles due to the dependency of the SA on the PSD. The surface charge was measured as a function of particle size, and was found to range from 1.60 to 6.64 mol/m2. No trend was found on the effect of particle size on PZ C. An important finding directly applicable to any coagulation based UOP was that the magnitudes of the PZC values vary by site and sampling event. Settling velocities were measured as function of particle diameter at room temperature (25C) and were foun d to be in very good agreement with the modeled values assuming Newtonian settling, with the most pronounced deviations in particles larger than 1000 m in diameter. The effect of temperature on settling velocities was also studied at 1, 5 and 10C. No si gnificant difference was found between modeled settling velocities at 1, 5 and 10C as function of particle diameters, although a nominal difference was observed for particles with diameters finer than 100m. Local precipitation conditions (rainfall and snowmelt) result in low flow rates and a very high removal of PM in the primary sedimentation basin, typically over 99% removal of particles diameters greater than 25m, based on both a mass and number calculation. The lowest removal efficiencies are seen at the end of April and coincide with the highest flow rates that were also seen at the end of April. This is entirely due to snowmelt from the storage basin because there was very little precipitation seen in April. High intensity storms seen in Janua ry and February had little influence on the effluent flow from the sedimentation basin, indicating that the majority of flow volume is due to snow melt alone, with stormwater having minor effects. This situation is confirmed by noticing that the high inte nsity storms in late June resulted in very low flow rates (Q < 0.01 m3/s). Much lower removal efficiencies were seen when the addition of particles 1 < d < 25 m was considered on a number basis, the minimum removal efficiency was 55.3%, with a median val ue of 81.5%.

PAGE 41

41 D iscussion Parallel to rainfallrunoff treatment, there exists the necessity of a comprehensive approach to separation of particulate matter in urban snowmelt. Such an approach should consider the coupled quantitative and qualitative aspects of the PM, which in turn are highly site specific. In addition, the highly variable influencing factors such as snow residence time, vehicular traffic loadings, and snow maintenance practices translate into highly variable data for particle characteristic variables that influence selection of an appropriate separation BMP/UOP. Determination of parameters such as surface area and surface charge has a direct implication not only on the selection of a UOP, but also on long term efficacy. The average particle diameter was found to be i n the sediment range, (range of 94.55 to 99.97, average of 98.48%) which suggests that the appropriate cost effective PM separation UOP would be a sedimentation operation or any UOP designed to remove sediment sized material. Settling velocity measuremen ts matched modeled values very well and temperature dependence was negligible for particles greater than 100 m, indicating that the temperature effects on settling velocity would be minimized for the majority of the PM examined in this study. The majorit y of the total surface area is in the larger particles, showing that if the larger gradations were removed, that the majority of components adsorbed to the particles would also be removed and will be settled out based on the treatability study, but might r equire regular maintenance to account for partitioning and speciation of sediment bound nutrients Overall, it is imperative that site specific particle characteristics be determined as part of a comprehensive PM treatment and management system for urban snowmelt.

PAGE 42

42 Nomenclature BMP best management practice dC drag coefficient iC Surface Charge (mol/ g ) C/F coagulation/flocculation COOP Cooperative Observer Program CSTR continous stirred tank reactor D dielectric constant of a liquid d particle diameter (L) gd geometric mean dp particle diameter (L) dL thickness of the layer surrounding the shear surface of a particle d# the particle diameter at which #% of the total mass is found below that particle size e.g. d84 is the particle diameter at which 84% of the mass is found in particles finer than that size EGME ethyleneglycol monoether g acceleration due to gravity (L T1) GAC granular activated carbon GK kurtosis L length M mass MBE Mass balance error

PAGE 43

43 mi mass of a specific particle size N Hazens N ( ) NDOT Nevada Department of Transportation PM particulate matter PND particle number density PP polypropylene PSD particle size distribution PZC point of zero charge q charge per unit area Q Flow Rate (V/T) dRe Reynolds number SA surface area (L2) Sk skewness psg specific gravity ( ) SSA specific surface area (L2 M1) SSC suspended solids concentration (M V1) TDS total dissolved solids TSS total suspended solids concentration (M V1) UOP Unit Operation and Process V volume pV Settling velocity of a particle (L/T) t pV terminal particle velocity (L/T) WRCC Western Regional Climate Center

PAGE 44

44 Gamma distribution alpha Gamma distribution beta Gamma function zeta potential dynamic viscosity (M T L2) kinematic viscosity (L2 T1) p particle density (M L3) w density of water (M L3) g the standard deviation of the geometric mean I uniformity index # the negative log base 2 of d#

PAGE 45

45 Table 21. Location of snow sampling sites. Site # Sample Location Sample Date Notes 1 103A Sierra Blvd@Barbara Ave, S. Lake Tahoe, CA 1/5/2003 Roadway Snow Dump 1 103B Sierra Blvd@Barbara Ave, S. Lake Tahoe, CA 1/5/2003 Roadway Snow Dump 1 1203 Sierra Blvd@Barbara Ave, S. Lake Tahoe, CA 12/20/ 20 03 Roadway Snow Dump 2 1203 HWY US50W@Tunnel, NV 12/20/ 20 03 Road Surface 2 1204 HWY US50W@Tunnel, NV 12/27/2004 Road Surface 2 1205 HWY US50W@Tunnel, NV 12/21/2005 Road Surface 3 1203 HWY US50W@Firestation #5, NV 12/20/2003 Road Surface 4 1204 HWY US50W@Zephyr Cove, NV 12/28/2004 Parking Lot 4 1205 HWY US50W@Zephyr Cove, NV 12/21/2005 Parking Lot 5 1204 HWY US50W@Stateline, CA 12/29/2004 Road Surface 6 1205 HWY US50W@Lake Pkwy, S. Lake Tahoe, NV 12/21/2005 Road Surface Table 22. Total PM dry mass and aqueous volume along with relative sediment, settleable, and suspended particle mass. Site # Total Mass Total Volume Suspended PM Settleable PM Sediment PM (kg) (L) (%) (%) (%) 10103 3.89 6.76 8.2 3.4 88.4 11203 0.51 6.79 1.8 6.9 91.3 21203 8.75 5.86 0.0 2.8 97.2 21204 4.41 7.18 0.9 1.1 97.9 21205 2.62 5.81 0.7 4.4 95.0 31203 3.06 6.37 1.5 4.2 94.3 41203 0.17 7.49 1.6 4.0 94.4 41204 1.75 8.06 0.2 0.4 99.4 51204 5.23 5.75 0.2 0.6 99.3 61205 0.87 7.25 0.5 3.0 96.6

PAGE 46

46 Table 2 3. Geometric mean (dg), central tendency ( 50), skewness (Sk), kurtosis (KG), and I). The explanations of the parameters are shown below. Site # dg (m) g (m) 50 Sk KG I 1103A 539 4.49 0.73 0.03 0.88 2.02 1103B 703 4.69 0.43 0.01 0.95 2.10 11203 285 3.79 1.85 0.03 0.76 1.76 2 1203 644 4.34 0.73 0.04 0.96 2.05 21204 345 3.35 1.74 0.17 1.02 1.73 21205 1334 3.12 0.47 0.02 0.95 1.67 31203 864 2.15 0.43 0.03 2.15 1.22 4 1204 494 4.30 0.99 0.01 0.83 2.01 41205 1216 2.76 0.37 0.17 1.07 1.49 51204 832 4.68 0.29 0.02 0.94 2.18 61205 489 3.80 1.01 0.07 0.99 1.95 50 50 < 1 1< 50 <0 0< 50 <1 1< 50 <2 2< 50 <3 3< 50 <4 4< 50 <8 8< 50 Sediment Type Gravel Very coarse sand Coarse sand Medium sand Fine sand Very fine s a nd Silt clay SK 0.30
PAGE 47

47 Table 2 4. Summary of gamma distribution parameters for individual sites and the median PSD. The first column indicates values for the mass PSD plot (Figure 2 4 top). The mass balance error (MBE) is also shown for the mass PSDs. The second column (PSD by SA) indicates values for the SA PSD plot (Figure 24 bottom). Site ID PSD by Mass PSD by SA r 2 MBE (%) r 2 1 103A 0.932 0.106 0.99 0.54 0.741 0.159 0.96 1 103 B 0.897 0.137 0.99 0.26 0.816 0.126 0.97 1 1203 0.974 0.047 0.98 0.18 1.086 0.030 0.95 21203 0.946 0.108 0.99 0.16 0.719 0.123 0.97 2 1204 1.278 0.036 0.99 0.18 1.070 0.051 0.99 2 1205 1.229 0.157 0.99 0.16 1.229 0.059 0.97 3 1203 1.704 0.057 0.99 0.08 1.297 0.093 0.99 41204 1.278 0.036 0.99 0.18 1.054 0.093 0.99 4 1205 1.229 0.157 0.99 0.16 1.765 0.026 0.97 5 1204 0.801 0.190 0.99 0.20 0.575 0.241 0.98 6 1205 1.085 0.069 0.99 0.38 1.052 0.059 0.94 Median PSD 1.001 0.201 0.99 N/A 0.895 0.202 0.99 Particle Diameter ( m) 10 100 1000 10000 Particle Density (g/cm3) 0 1 2 3 4 5 Coarse Fine Figure 21. Particle density as a function of particle diameter

PAGE 48

48 Mass Retained (g/kg) 0 100 200 300 400 500 Percent Retained (%) 0 20 40 60 80 100 Mass Measured Modeled n = 1.001 = 0.201 r2 = 0.99 SSA (m2/g) 0 10 20 30 40 50 60 70 Particle Diameter ( m) 4750 2000 850 600 425 300 250 180 150 106 75 63 53 45 38 25 SA (m2/kg) 0 1000 2000 3000 4000 5000 Cumulative Percent (%) 0 20 40 60 80 100 Surface Area Measured Modeled = 0.895 = 0.202 r2 = 0.99 n = 11 Figure 22. Particle size distributions for all eleven sites are shown in the top plot; values shown are normalized to 1000g. Parameters listed are for a transpose cumulativ e gamma distribution used to model the cumulative particle size distribution. The middle plot shows incremental specific surf ace area. The bottom plots show incremental total surface area and cumulative percent, values shown are normalized to 1000g. A B C

PAGE 49

49 Particle Diameter ( m) 4750 2000 850 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 SSA (m2/g) 0 20 40 60 80 100 Measured Calculated value x 1000 Figure 23. A comparison of calculated and measured SSA values for snow particulate matter. Measured values are the average and standard deviation of all SSA data presented; calculated values assume constant specific gravity, spherical particles, and solid particles. (Sansalone et al 1998).

PAGE 50

50 Figure 2 4. Image of 4 grains from the 4750 < d< 2000 mm 1103B sample. The paperclip in the lower part of the i mage has a diameter of 1 mm. The samples shown represent A) a smooth grain with minimal surface area, B) a grain with debris that would contribute to surface area, C) a porous grain with a large surface area, and D) a quartz grain with moderate surface ar ea. 425 m pH 6 7 8 9 Surface Charge ( mol/m2) -2 -1 0 1 2 3 4 75 m pH 6 7 8 9 -2 -1 0 1 2 3 4 Figure 25. Surface charge as a function of pH for a 425 m (left) and a 75 m (right) particle, the average and standard deviation of the point of zero charge is also shown. A B 1 mm A B C D

PAGE 51

51 S. Lake Tahoe Snow Dump, CA 10 100 1000 10000 pH @ PZC 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 05Jan03 A 05Jan03 B 20Dec03 US 50@Tunnel, NV 10 100 1000 10000 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 US 50@Zephyr Cove, NV Particle Diameter ( m) 10 100 1000 10000 pH @ PZC 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 28Dec04 21Dec05 Particle Diameter ( m) 10 100 1000 10000 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 20Dec03 US HWY 50 @ Firestation #5 29Dec04 US HWY 50 @ Stateline 21Dec05 US HWY 50 @ Lake Pkwy 20Dec03 27Dec04 21Dec05 Figure 26. Point of zero charge by site. Plot (a) shows the 3 samples from Site 1, plot (b) shows the 3 samples from Site 2, plot (c) shows the 2 samples from site 4, plot (d) shows the remaining Sites 3,5 and 6. Particle Diameter, D ( m) 10 100 1000 10000 Settling Velocity, VP (mm/s) 0.01 0.1 1 10 100 1000 Modeled @ 1C Modeled @ 5C Modeled @ 10C Measured @ 5C Turbulent Transitional Laminar 34 0 Re 3 Re 24 d d dC Figure 27. Measured and modeled settling velocities.

PAGE 52

52 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Temperature (C) -10 0 10 20 30 40 Average Maximum Temperature Mean Temperature Average Minimum Temperature Flow Rate (m3/s) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 Jan Feb Mar Apr May Jun 0.00 0.01 0.02 0.03 0.04 0.05 0.06 Influent Effluent Figure 28. (Top) Average temperature data. Average high and low data from Western Regional Climate Center (WRCC 2002), m ean temperature from (WCI 2008). (Bottom) Influent and effluent flow rates for the sedimentation basin. Mass-Based d>25 m Jan Feb Mar Apr May Jun Removal Efficiency (%) 96 97 98 99 100 Number-Based d>25 m Feb Mar Apr May Jun Figure 2 9. The mass and number based removal efficiency of the basin for PM greater than 25 m in diameter given the 2008 precipitation and temperature conditions.

PAGE 53

53 Flowrate (m3/s) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 Removal Efficiency (%) 0 20 40 60 80 100 Figure 2 10. The effect of flow rate on the number based removal efficiency of the basin for PM with diameters greater than 1 m. Particle Size ( m) 0.1 1 10 100 1000 10000 Removal Efficiency (%) 0 20 40 60 80 100 Influent Effluent Figure 2 11. Range of influent and effluent number based PSDs.

PAGE 54

54 Figure 2 12. EDS (Energy dispersive spectrometer) analysis on 3 particle diameters. The EDS shown for the 2000 mm particle diameter is the average of 3 separate grains. The EDS for the 425 and 63 um particle diameters are bulk EDS plots and were taken over multiple grains with areas of 27 mm2 and 4.3 mm2 respectively.

PAGE 55

55 CHAPTER 3 DISTRIBUTION OF PARTICULATE BOUND METALS IN SOURCE AREA SNOW IN THE LAKE TAHOE WATERSHED Introduction Snow snowmelt is a vital water source for the western United States. As western areas were settled water resources such as Lake Tahoe, in part supplied by snowmelt, were promoted as a water supply for agriculture, industry and domestic demands; and a water source for silviculture growth while providing waterway nourishment for aquatic and terrestrial life (Feth et al. 1964). However, as pavement and the motor vehicle began to dominate urban areas in the 20th century, snow and snowmelt on pavement became a hazard and bare pavement policies were instituted since the 1960s to provide safe and uninterrupted traffic in regions with snow (Novotny et al. 1998, Glenn and Sansalone 2002). While snowmelt is a water source, plowed snow banks alongside roadways form a temporary linear reservoir for traffic generated constituents such as metals and PM. In cold climates with urban pavement, a significant fraction of the annual metal and PM based loads occurs during snowmelt and spring runoff event sequences (Oberts 2000). Highway impacted snow generally melts in a snowmelt sequence: pavement melt, roadside (impervious) and finally pervious area melt. As part of this sequence, any stochastic rain onsnow melt events can tran sport the highest loads of PM based pollutants (Oberts 2000). Pavement surfaces result in hydrologic modifications including larger volumes, reduced lag time (Huber 1993). This in turn promotes pollutant load transport. These pollutant loads are generat ed by anthropogenic activities such as traffic, as well as pollutants generated by pavement and associated infrastructure ( Heaney and Huber 1984; Hamilton and Harrison 1991; Arnold and Gibbons 1996) These concentrations can be deleterious to receiving water bodies, aquatic life and potentially the natural food web. Sources of metals on roadways include wet and dry

PAGE 56

56 deposition, traffic activities, direct deposition of deicing salts and antiskid material, and highway construction and repair materials. A cidic snowmelt can reduce the pH and increase the concentration of dissolved metals in receiving streams, but such impacts have been difficult to detect due to the sporadic nature of snowmelt (Servos et al. 1987). An example of metal generation from traff ic activities include Cd from brakes and tires; Cu from brakes and tires; Pb from brakes, tires, fuels and oils, and de icing salts; and Zn from brakes, tires, frame and vehicular body components (Huber et al. 2001, Sansalone and Buchberger 1997). A cont ributor of Zn to highway runoff is from coatings on the roadway guardrail surface (Field and Sullivan 2003). Infrastructure such as brick buildings have also been shown to be a source of Pb and Zn in a study with synthetic rain water (Davis et al. 2001). A study by Heijerick et al. (2002) concluded that 94 to 99% of Zn in runoff from infrastructure materials was ionic Zn2+. Copper roofing materials have also been shown to leach during rain events (Field and Sullivan 2003). De icing compounds contain additives that may generate high concentrations of metals and cyanide on roadway runoff (Sansalone and Glenn 2002). Construction activities have been shown to contribute pollutant loads in stormwater, including solids, metals, and deicing chemicals (Field an d Sullivan 2003). The pavement itself can leach metals into stormwater, when recycled waste materials are used for both economic and waste reduction purposes (Nelson et al. 2003). A 1998 Cincinnati highway study analyzed 10 sites for metals in snow as a f unction of particle size indicating most metal mass was particulate bound (Glenn and Sansalone 2002). A highway study in Cincinnati, OH examined variations of PM based Na, Zn, Pb, Al in rainfall runoff and snowmelt with higher event mean concentrations (E MCs) for Cd, Cu and Pb in snowmelt as opposed to higher Zn concentrations in rainfall runoff (Sansalone and Buchberger

PAGE 57

57 1996). In another study with rainfall and snowmelt interactions, the highest concentrations were measured during rain onsnow events (We sterlund et al. 2003). Anthropogenically impacted snow is a temporal maintenance challenge. Common urban snow maintenance practices continue to include plowing snow into banks along the roadside and application of traction or anti skid material such as sand or grit and/or de icing salts (Glenn and Sansalone 2002, Oberts 2003). In addition, accumulated snow can be left to melt insitu, or excavated and transported to a central location for melting, or directly dumped into a nearby water body depending on local pollution control regulations. Reinosdotter and Viklander (2006) report that some cities in Sweden have tried snow separation strategies to minimize the negative effects of snow dumping in receiving water bodies, wherein heavily polluted snow such as from highways is transported to a locality that is designed, located, and operated to minimize negative environmental effects while cleaner snow was left in place to melt or moved to a local snow deposit. To differentiate levels of snow pollution, urban areas are divided according to snow quality (Reinosdotter and Viklander 2006). In order to propose an effective snowmelt management strategy, for example excavation, trucking and central treatment by melting and sedimentation, PM pollutants such as met als require quantification. Since the 1980s, there have been many structural best management practices (BMPs) ( based largely on volumetric control such as detention) that are also capable of providing a sedimentation mechanism (Nix et al. 1988; Clar et al 2004). Such quantification establishes the level of snow pollution, for example as a function of land use or exposure for such management decisions, and allows the generation of constitutive models between pollutant parameters that can be integrated wit h treatment mechanisms such as sedimentation.

PAGE 58

58 Objectives There are three objectives of this study focused on measuring and modeling the mass of PM based metal species in transportation source areas of the Lake Tahoe watershed. The first objective is to m easure metal mass as a function of granulometry (specifically, particle size distributions, PSD, and surface area) to illustrate mass distributions. The second objective is to represent the metal mass distribution with a gamma distribution as a constitutive model of the PM metal mass relationship Ultimately, snow treatment for PM based metals requires melting, and some form of separation, commonly sedimentation. In cold climates where treatment is supplied by a centralized mass accumulat ion of snow with onsite treatment, the snow melting process can persist into late spring or early summer. The final objective is to utilize the PM metal mass constitutive model combined with SWMM (Storm Water Management Model) and rainfall loadings to mo del the eff ectiveness of an existing sedimentation basin for separation of PM based metals during the melting process. Partitioning is not examined in this study. M ethodology and Laboratory Analysis Sampling Locations Data w as collected from six source area sites over all or part of four consecutive winter seasons. Three sites have data from multiple years and data from three other sites were not collected in multiple years due to inaccessibility. Table 31 lists the site designations, locations and sam pling dates. Local climate conditions for the area are gathered from the National Weather Service Cooperative Observer Program (COOP) (WRCC 2002). Data are compiled in a public access database at the Western Regional Climate Center. Historical climate data for this study are collected from 9 stations in the Lake Tahoe Basin region. Reported parameters include mean high and low monthly temperatures, monthly total precipitation, total monthly snowfall, and

PAGE 59

59 mean monthly snow depth. The period of record for each site varies with the earliest date in 1914 and the most recent in 2007 with most sites having at least 20 years of data. The mean daily minimum temperature in January is 8.8 C with mean daily low temperatures typically below 0 C from November to April and daily mean temperatures below 0 C from December to February. The annual rainfall for the area ranges from 212 to 1289 mm (8.36 to 50.73 inches) with a mean of 708 40.8 mm (27.8 16.1 in) and total snowfall ranging from 0.5 to 10.5 m (18.7 to 412.5 inches) annually, with a mean of 4.7 3.8 m (185 150 inches). Traffic data for the region consist of the average daily traffic, averaged over the course of a year. Traffic along the 21.4 km (15 mile) section of US Highway 50 through the La ke Tahoe Basin sampling section from Echo Lake Road (California mile marker 65.62) to the California/Nevada state line; ranges from 12,200 to 36,000 vehicles per day measured during 2004. Maximum values of average daily traffic (ADT) traffic can range fro m 14,700 to 43,500 vehicles per day, this data was taken from the month of the heaviest traffic (Transportation Management Association 2002). An 8.0 km (5 mile) section of US 50 in Nevada from the state line to Zephyr Cove, NV has monitored traffic rangi ng from 14,800 to 32,100 vehicles per day measured during 2000 (NDOT 2006). Snow Sampling Snow samples were taken from linear snow banks at the edge of a roadway or parking area. These snow banks were generated from plowing pavement snow off of roadways a nd parking pavement. Snow was packed into 4 L wide mouth polypropylene (PP) bottles. Sampling consisted of taking a thin width of the leading cross sectional area of the snow banks (Glenn and Sansalone 2002). All samples were taken in duplicate. All sa mples were taken while snow was still frozen and kept in ice chests in a solid phase during transport to the laboratory.

PAGE 60

60 The samples were then stored frozen until analyses. Samples were thawed for 12 hours at room temperature before analysis were initiated. All snow samples were examined as melted snow. Particle S ize D istributions (PSD) The mechanical dry sieve analysis test method used to determine the PSD was ASTM D422 modified to include a greater number of sieves (Sansalone et al. 1998); all samples were prepared using ASTM D421 (ASTM, 1990, ASTM, 1993). Upon melting, supernatant wa s separated and refrigerated and the wet slurry of the entire PM phase from each snow sample was dried at 40 C and desiccated. The PM trapped in the snow was heterodisperse and non cohesive and was disaggregated with a mortar and pestle. These samples w ere sieved through a series of 17 sieves ranging from the 9500 m (#3/8) to 25 m (#500) and a pan to collect PM smaller than 25 m. The sieves were agitated with the sieve shaker for 7 minutes with a 2 mm amplitude. The initial mass of PM was measured and the mass remaining on each sieve was collected and weighed; mass balances were calculated. A mass balance criterion of 2% was established for quality control purposes. PSDs were modeled on a gravimetric basis using a cumulative gamma distribution. Pe rcent finer by mass at each PM size was plotted as a function of particle size. All mass values were normalized to 1000 g. Specific Surface Area and Total Surface Area The specific surface area (SSA) is the ratio of the surface area of the particle to the mass or volume of the particle. SSA was determined using an EGME (ethylene glycol monoethyl ether, HOCH2CH2OCH2CH3) method (Carter et al. 1986, Sansalone et al. 1998, Arnepalli et al., 2008). EGME is a very polar liquid with a high vapor pressure. A control PM with a known SSA (1100 m2/g, Calgon, 1995) was utilized for all tests in triplicate, granular activated carbon (GAC). Total surface area (SA) distributions were determined by multiplying the measured SSA by the

PAGE 61

61 mass of that particle size. The tot al SA is an index of the capacity and potential for adsorption and ion exchange with PM. All total SA values were normalized to 1000 g. g m g g SSA Sample of Mass 000286 0 Sample on EGME of Mass2 ( 31) i i im SSA SA ( 32) Metals Analysis A nitric hydrochloric acid method was used to extract metals from dry PM using SW 846 Method 3015 (USEPA 1990). A dry spiked soil standard sample (Environmental Resources Associates, Trace Metals in Soil) and a blank (empty vial) were digested with every eighteen samples to ensure a 90 to 110% me thod recovery and check for method contamination, respectively. A known mass of sample was added with replicates in 125 mL Erlenmeyer flasks and then 9 mL of trace metal grade nitric acid and 3 mL of hydrochloric acid were added to each flask, including t he blank and standard flasks. Flasks were covered with a watch glass and heated at 150 C for 90 minutes followed by heating at 175 C for 30 minutes. After the two hour period, 10 mL of deionized water was added to the digested sample for dilution, then t he sample was filtered through a glass fiber filter into a 100 mL volumetric flask, a solution of 2% nitric acid was added to a flask until the total volume reached 100 mL. For the 20032004 samples the digested PM filtrate was analyzed on a Perkin Elmer Elan 9000 ICP MS. The protocol outlined by the EPA in SW 846 Method 6020 for ICP MS analysis was utilized (USEPA 1996). For the 2005 samples, the filtrate was analyzed with a PerkinElmer Elan 3200 ICP AES. Data Analysis Cumulative metal distributions w ere modeled with a gamma distribution. While the gamma distribution has not been applied to snowmelt metal mass PSDs, it has been successfully

PAGE 62

62 applied to stormwater PM PSDs and urban traffic dry deposition PM (Kim and Sansalone 2008, Sansalone and Ying 2008). The probability density function (pdf) is shown in Equation 33 and the gamma cumulative density function (cdf) is shown in Equation 34. The gamma function and incomplete gamma function are shown in Equations 35 and 36 respectively. In these equ ations is the distribution shape parameter, is the scaling parameter, d is the normalized particle diameter, and x is a substitution for d (PM diameter) in the integrals. The gamma distribution is normalized by the largest particle size (9500 m) and metal mass is modeled as a percentage. The parameter estimates were based on minimizing the sum of squared errors (SSE) and maximizing the coefficient of determination. 0 1 1dx e x e d d fx d (3 3) dd F 1 1 (3 4) 0 1dx e xx (3 5) d x ddx e x0 1 (3 6) The metal mass concentration in each size range (CMe, i) normalized to 1000 g of dry PM was calculated using Equation 3 7: where CICP is the concentration measured by the ICP, Vd is the total volume of sample after digestion, MN is the mass retained on the sieve normalized to a total mass of 1000 g (1 kg), and Md is the mass of sample digested. mg g g M kg g kg kg M L V L g C kg mg Cd N d 3 3 ICP i Me,10 10 ( 37)

PAGE 63

63 The number of moles of metal on a surface area basis is calculated with Equation 3 8, where SA is the total surface area and MW is the molecular weight of the metal species. mol g MW kg m SA mol mol mg g kg mg C m mol C2 6 3 i Me, 2 i Me,10 10 1 (3 8) Metal mass distributions were modeled using a cumulative gamma distribution and minimizing SSE by changing and The gamma function (Equation 39) was used with particle diameter used in place of the x value, normalized to the largest particle diameter (9500 m). Percent of mass finer for each size distribution was plotted against particle size. All metal mass values were normalized to 1000 g of total PM before plotting. C umulative gamma distribution xe x x f11 ; (3 9) Modeling Basin Treatability for PM based Metals Separation of PM based metal for a centralized snow storage site was facilitated with an onsite sedimentation system of two adjacent basins in series (identified as the basin) collecting snow deposit snowmelt from snow trucked and mounded at the site. The snow storage site (2.1 ha) is located in South Lake Tahoe, CA. From site measurements, mounded snow was approximately 300 m in length and varied from 3 to 5 m in height. Based on LIDAR data and analysis in GIS (Lake Tahoe Data Clearinghouse 2005) the site also received runon from an adjacent 14.6 ha watershed as shown i n Figure 31. The storage site was paved with asphalt and densified gravel generating a highly impervious surface. The primary soil types are Christopher loamy sand, Christopher Gefo complex, Jabu sandy loam, and Marla sandy loam. All four soil types ar e loamy sands for the first 50150 cm, with a shallow (27 cm) organic top layer and

PAGE 64

64 slopes ranging from 0 to 9 percent (Soil Survey Staff, 2008). The snowmelt and runoff to the basin was modeled using SWMM (Rossman 2007, Huber and Dickinson 1988), histor ical rainfall time series, and Hazens Law as a modified Type I overflow rate settling model (Malco l m 1989, Cristina et al. 2002). A Hazens N value of 1 (assumes basin operates as a single CSTR, continuously stirred tank reactor) and is described by the following equation. N PN Q s A V 1 1 (%) Removal ( 310) In this expression Vp is the measured settling velocity of a particle size, AS is the basin surface area (1909 m2 + 550 m2), Q is the effluent flow rate, and N provides an estimate of nonideal basin mixin g. The two basins in series are connected by a 460 mm pipe (18.1 in.); 13 m (42.6 ft) in length. Seven months of historical climate data from 01 January through 31 July 2008 were collected for the storage site from the South Lake Tahoe airport, a dist ance of approximately 3 km from the storage site watershed and at a similar elevation (1925 m and 1918 m, respectively, above sea level). Rainfall intensity and temperature served as inputs into SWMM to generate flow rates from rainfall and to generate s nowmelt into the basin and for evaporation. The modeled inflow data, sedimentation basin geometry, stage storage, overflow rates and PM metal mass constitutive model for snow were utilized to determine the separation of PM based metal as a function of par ticle diameter. Separation efficiency was weighted on a granulometric mass basis for each PM size to generate gradation based basin behavior for individual PM sizes. R esults For each metal, data from all sites and all sampling years are combined and plot ted in a nonparametric format for the entire gradation. A summary of these results are presented in the

PAGE 65

65 Figure 32 (left) illustrating that the highest PM based metal concentration was Fe and the lowest was Cd. The granulometric distribution (as a PSD) of PM across which these metals are distributed is illustrated in Figure 32 (right). The PSD plot illustrates the hetero dispersivity of the PM and that the predominance of PM mass and therefore total PM surface area (SA) is associated with the sediment size fraction (> 75 m) similar to results for Cincinnati source area snow PM (Sansa lone and Cristina 2004). Ranking the total metal mass (per kg of total PM mass) from highest to lowest demonstrated the following PM based order: Fe > Al > Zn > Cu > Cr > Pb > As > Cd as summarized in Figure 3 2 and Table 32. Fe and Al exceeded 1000 mg/ kg, while Zn, Cu, Cr, Pb, and As are in the 1 to 100 mg/kg range and Cd was below 1 mg/kg. These summary results are further examined across the granulometry. Metal concentrations are plotted in two formats; both illustrating metal as a function of partic le diameter. The left hand plots of Figure 33 (Cr, As, Fe, and Al) and Figure 34 (Cu, Zn, Cd, Pb) illustrate each metal on a molar basis, normalized to measured SA as a function of PM diameter. The right hand plots illustrate each metal on a gravimetric basis as a function of PM diameter normalized to 1000 g of total PM. On a molar basis, Fe and Al illustrate the highest PM based values as a function of SA of PM. When averaged across the entire PSD, the mean molar metal species values ranked from high est to lowest are: Fe > Al > Zn > Cu > Cr > Pb > As > Cd. There is a mild trend of increasing molar concentration with decreasing PM size. The mass concentration values as a function of particle size are shown in a box and whisker format (explanation found in Figure 39). A peak mode occurred in the coarse PM range between 300 and 2000 m (seen in Figures 33 and 34, Table 33). Another smaller increase in concentration was seen for all metal elements around the 75 to 106 m range, followed by decreasi ng median concentration values as particle diameter decreased. The

PAGE 66

66 measured d50m (mass based) values for all metals range from 179 (for Cu) to 542 (for As) m. The measured d10m, d25m, and d90m were typically higher than the corresponding modeled values, while the d50m and d75m were typically lower than modeled values as summarized in Table 33. ANOVA analyses indicate no significant difference in the means between the measured and modeled for the d10, d25, d50, or d75. The lowest values for each granulom etric index are typically associated with Cu and Pb while the largest values are typically associated with Al, As, and Cd. PM based metal mass data are modeled with cumulative gamma distribution. Model parameters ( ) are individually shown in the second column of the plots and compared graphically (Figures 33 and 34, right column) and in a tabular summary for the entire dataset in Figure 35. Values for the scaling parameter, ranged from 0.529 to 1.763 for the metals examined, while the shape factor, ranged from 0.012 to 0.220. ANOVA analyses (at = 0.05) was conducted to determine if and values were significantly different between metals. Results indicated that values were not significantly differen t (p = 0.56), while the values were statistically different (p < ). The mean total metal mass on a 1.0 kg of total PM basis for each metal is summarized for each site in Table 32. Using ANOVA analyses ( = 0.05) to determine if the means of each site were equivalent, the results failed to reject the null hypothesis for Cr (p = 0.39), Cu (p = 0.29), and Zn (p = 0.13). The remaining metals, As, Al, Cd, Fe, and Pb, had pvalues less than 0.05 and the hypothesis rejected; with statistically different mea n concentrations. Utilizing continuous simulation with SWMM to model flows into the basin, Hazens algorithm, and the assumption of Type I settling, results indicate the existing basin is very effective at removing PM bound metals. Precipitation and watershed yield data for the basin are shown in Figure 36. Precipitation is shown both incrementally and cumulatively with the flow rate shown only cumulatively. The graph illustrates several precipitation events before March.

PAGE 67

67 These events did not increase flow immediately b ecause they were snow and only after March did snowmelt increase. Temperature data for the model is shown in Figure 37 along with the daily mean temperature and the 20 year mean daily temperature. Results indicate that 2008 had periods from January to M arch where the temperature was below average, decreasing snowmelt into the basin. Temperatures remained above freezing after March. After March the basin is loaded by the majority of flow volume. Since the 14.6 ha watershed is wooded and sparsely popula ted the modeling assumes negligible PM bound metals in rainfall runoff. Influent and effluent flow rates, surface overflow rates, as well as the measured and modeled stagestorage curves for the basins are shown in Figure 3 8. Median influent and effluent are 0.00765 and 0.00765 m3/s respectively. Surface overflow rates varied based on effluent flow rates and stage basin surface area relationship across the variation in stage above the outflow invert. Surface overflow rates show a maximum value of 0.086 m/s with mean and median values of 0.018 and 0.013 m/s, respectively. At the outflow invert the maximum storage of the basins is approximately 1323 and 975 m3. Influent and effluent PM metal mass and PM loads from 01 January through 31 July 2008 are show n in Figure 39, with metals arranged in decreasing order by median value. The simulation only focused on influent and effluent loads generated by snowmelt since the adjacent watershed was largely undeveloped forested area and snow samples taken from this watershed had PM and metals that were at trace levels compared to the snow dump material. Simulation results indicated that Fe generated the highest influent and effluent loads, while Cd generated the lowest influent and effluent loads. The median and r anges of influent and effluent PSDs show significant reduction in coarse PM; with an effluent d50 of approximately 50 m. On a gravimetric basis the separation efficiency of the basin was 99% for removal of the median mass of each metal and 99% removal of PM over the period of the

PAGE 68

68 continuous simulation from January through July 2008. The 99% represented the entire gradation, but was dominated by the coarse PM fraction since the basis was gravimetric. Discussion While it can be shown that coarse PM and ass ociated metal loads are readily separating either by a basin design or in the urban drainage system, the potential for leaching of metals from this PM is significant (Ying and Sansalone 2008). Leaching can be promoted by changing water chemistry; for exam ple in below grade BMPs where reducing conditions can be generated between drainage events. Leaching can also be promoted by de icing salts or exchangeable ions in snowmelt. If an adsorbed species forms outer sphere complexes, then background ions would compete for the adsorption sites and cations in solution. Whether separated PM or engineered media systems for snowmelt; a decrease in adsorption and increase in aqueous complexes for metals is a potential result of higher de icing salt concentrations. C onversely, if a species forms inner sphere complexes directly coordinated to adsorbent surfaces background de icing salt ions would have less of an impact. Leaching or conversely adsorption would be less affected. However, for runoff or snowmelt PM, spe cifically coarser more labile PM captured by BMPs, adsorption is physical outer sphere complexes predominate and leaching is promoted. Therefore regular maintenance of management strategies such as a sedimentation basin at a centralized snow storage site or cleaning of the urban drainage system can help mitigate any eventual transport to receiving waters. Furthermore, maintenance practices such as pavement cleaning can reduce PM and metal loads to plowed snow, to receiving waters and to the local atmosphere. While the majority of PM and metal loads are associated with coarser PM fractions that are readily separated, the greater challenge is the finer suspended fraction that is not separated by most unit operations or in drainage systems. This suspended fraction has a more important turbidity impact and is more bio available than the settleable and sediment PM

PAGE 69

69 fractions in pristine receiving waters such as Lake Tahoe. With more centralized management strategies such as the storage site basin, secondary t reatment through adsorptive filtration represents a potentially viable consideration to further improve treatment for managed snowmelt. The effluent gradation with a d50 of approximately 50 m in the settleable PM range suggests the need for adsorptive fil tration after sedimentation; in particular for receiving waters such as Lake Tahoe or tributaries. Centralized treatment systems for snow allow maintenance and documentation of PM based metal management as opposed to site by site unit operations that are largely unmaintained. Further improvements to snowmelt discharges can be facilitated by a combination of regular pavement cleaning, seasonal cleaning of drainage appurtenances and de centralized unit operations, as well inclusion of secondary treatment suc h as adsorptive filtration. Winter and summer seasonal recovery and management of PM is critical to ensure that PM based metals do not partitioned back to unit operation or drainage appurtenance discharges. Summary and Conclusions Source area snow load ed by traffic can be a significant reservoir and source for PM and PM based metals. Over a period of four years across six traffic source area sites in the Lake Tahoe watershed, snow plowed to the edge of the pavement was sampled and analyzed for the gran ulometric distribution of metals associated with the hetero disperse PM The resulting distributions were examined nonparametrically on a molar basis per unit of SA and on a gravimetric basis per unit of PM mass. The metal mass was associated predominat ely with coarser PM with this metalbased d50m in the sediment size range (179 to 542 m). The median d50 value of 341 m indic ates that removal of particles greater than this size would remove at least 50% mass removal of all metals. Metal mass distributions with PM were modeled with a constitutive model as a cumulative gamma distribution with all r2 values exceeding 0.94. Values

PAGE 70

70 of the gamma parameter showed no statistical difference in the means for each metal species indicating that the scaling parameter does not differ significantly with the species of metal investigated given that the land use conditions of each of the sites are similar. Total metal mass distribution trends are similar across all sites and sampling periods with the means of three metals (Cr, Cu, Zn) illustrating no statistical difference between sites. The highest PM based concentrations of metals were Fe and Al while the lowest was Cd. Granulometric results were utilized with SWMM to model the PM based separation of metal s by a sedimentation basin at a snow storage site in S. Lake Tahoe. With granulometric data, local climate data during the snowmelt duration at the snow storage site as well as basin stagesurface area, stage storage data were input into SWMM to model bas in inflow and surface overflow rates. While the continuous simulation was conducted from January through July of 2008 to examine rainfall runoff impacts on surface overflow from the basin for that period, the majority of the snow melt occurs from March to June. During this time period the sedimentation basin reached maximum inflow rates of 0.064 m3/s in late May. These results were coupled with a surface overflow rate model, Hazens model, to quantify sedimentation and PM based metal removal, assuming no re partitioning in the basin. The treatability study demonstrated that the majority of particulate bound metals were separated by the existing basin through a combination of large basin surface area, low flow rates and a coarse gradation Effluent mass loads ranged from 0.03 g for Cd to 1.37 x 105 g for Fe as compared to their respective influent mass loads of 375 g and 1.39 x 109 g for the simulation period. The resulting removal efficiencies exceeded 99%. This gravimetricbased efficiency was based on the entire gradation from >4750 m to <25 m although the PM mass smaller than 25 m was less than 1% of the gradation on a gravimetric basis. Results indicate that the basin is effective

PAGE 71

71 for significant mass load reductions of settleable and sediment PM and PM based metals. While the >25 m fraction in the effluent is less than 1% on a gravimetric basis, this PM fraction is the most bio available fraction and the PM fraction remaining suspended in the water column contributing to turbidity. Nomenclature ANOVA Analysis of Variance CICP metal concentration from raw ICP data (M L3) CMe, i metal concentration on particle size i on a mass per mass or moles per area basis cdf Cumulative Density Function COV Coefficient of variation (100 x Standard Deviat ion/Mean) CSTR Continous stirred tank reactor d normalized particle diameter (L) EGME ethyleneglycol monomethyl ether EMC event mean concentration (M L3) d f probability density function of the gamma distribution d F cumulative gamma distribution ICP inductively coupled plasma L length (L) mi mass of a specific particle size (M) M mass (M) Mcf cumulative PM (M) Md mass of solids digested (M)

PAGE 72

72 MS total PM (M) Mecf cumulative metal mass (M) MeS total metal mass (M) MN normalized mass (M) MW molecular weight pdf Probability Density Function PM particulate matter PSD particle size distribution SA surface area (L2) SSA specific surface area (L2 M1) SSE Sum of squared errors psg specific gravity ( ) V volume (L3) Vd volume of filtrate from digestion of solid sample (L3) Vp Settling velocity of a particle x normalized particle diameter in gamma function equations (L) parameter in the cumulative gamma distribution gamma function d incomplete gamma function

PAGE 73

73 Table 3 1. Summary of sites, locations, and sampling dates. Site # Sample Location Sample Date Notes 1 103A Sierra Blvd@Barbara Ave, S. Lake Tahoe, CA 1/5/2003 Roadway Snow Dump 1 103B Sierra Blvd@Barbara Ave, S. Lake Tahoe, CA 1/5/2003 Roadway Snow Dump 1 1203 Sierra Blvd@Barbara Ave, S. Lake Tahoe, CA 12/20/03 Roadway Snow Dump 2 1203 HWY US50W@Tunnel, NV 12/20/03 Road Surface 2 1204 HWY US50W@Tunnel, NV 12/27/2004 Road Surface 2 1205 HWY US50W@Tunnel, NV 12/21/2005 Road Surface 3 1203 HWY US50W@Firestation #5, NV 12/20/2003 Road Surface 4 1204 HWY US50W@Zephyr Cove, NV 12/28/2004 Parking Lot 4 1205 HWY US50W@Zephyr Cove, NV 12/21/2005 Parking Lot 5 1204 HWY US50W@Stateline, CA 12/29/2004 Road Surface 6 1205 HWY US50W@Lake Pkwy, S. Lake Tahoe, NV 12/21/2005 Road Surface Table 3 2. Summary of total metal mass (mg per 1.0 kg of PM) and total PM (MT) (units of g) measured for each site and the mean of all sites. The total liquid volume of the melted sample is shown in the VT column (units of L). Site ID Al As Cd Cr Cu Fe Pb Zn MT Vs 1103A 2252 4.17 0.14 16.6 17.9 15233 6.6 61.4 1518 2.93 1103B 2501 4.20 0.19 15.6 18.4 15139 6.8 65.1 2054 3.70 11209 2552 2.72 0.08 20.1 35.6 18658 5.7 57.8 488 6.79 21203 3701 5.18 0.31 14.6 19.9 15054 8.7 148.1 8747 5.86 21204 4348 2.00 0.20 24.5 53.7 15005 12.6 53.5 4346 7.18 21205 1122 4.60 0.05 9.0 25.5 10883 2.8 87.9 2608 5.81 31203 3418 4.46 0.19 10.2 14.8 12128 6.2 45.7 3009 6.37 41204 9756 4.17 0.27 19.9 72.2 23964 9.7 111.0 169 7.49 41205 6034 5.93 0.25 10.0 20.5 11494 3.8 54.9 1748 8.06 51204 4803 3.01 0.13 9.6 25.9 17325 13.1 51.5 5061 5.75 61205 7511 11.79 0.37 15.4 35.6 19254 15.9 98.1 861 7.25 Median 3701 4.20 0.19 15.4 25.5 15139 6.8 61.4 2054 6.37

PAGE 74

74 Table 3 3. Measured vs. modeled parameters from the gamma distribution. Cumulative mass (g) distribution was summed from largest to smallest particle diameter. The dX m val ues represent the diameter at which X percent by mass is greater Metal Species d 10m d 25m d 50m d 75m d 90m Measured Modeled Measured Modeled Measured Modeled Measured Modeled Measured Modeled As 1684 1835 1336 1239 542 540 193 218 83 77 Al 1716 1590 833 934 430 451 143 176 75 60 Cd 1679 1962 882 1104 458 487 134 169 70 49 Cr 1409 1020 659 614 268 308 110 128 64 47 Cu 1239 957 597 505 179 229 64 76 43 20 Fe 1646 1247 750 751 414 375 139 156 72 57 Pb 1139 871 582 530 237 268 101 113 57 42 Zn 1298 1044 649 612 264 298 98 119 59 41 Maximum 1716 1962 1336 1239 542 540 193 218 83 77 Median 1527 1145 704 682 341 341 122 141 67 48 Mean 1476 1315 786 786 348 369 122 144 65 49 Minimum 1139 871 582 505 179 229 64 76 43 20

PAGE 75

75 Parameter Watershed Storage Basin Physical Characteristics Area (m 2 ) 146092 20630 Slope (%) 5 % 0.94 % Impervious Area (%) 5 % 100 % Green Ampt Parameters Suction Head (mm) 239 239 Conductivity (mm/hr) 0.6 0.6 Initial Deficit ( ) 0.321 0.321 Figure 3 1. The snowmelt basin in South Lake Tahoe, CA and shown in white. The surrounding watershed that contributes runoff to the basin is shown in grey. Coordinates at centroid of the watershed are: 38 54 55 N and 119 58 51 W.

PAGE 76

76 Metal Species Fe Al Zn Cu Cr Pb As Cd Total Metal Mass [mg/kg] 1e-2 1e-1 1e+0 1e+1 1e+2 1e+3 1e+4 1e+5 Particle Diameter ( m) 4750 2000 850 600 425 300 250 180 150 106 75 63 53 45 38 25 Incremental PM Mass (as g/kg of PM) 0 70 140 210 280 350 420 % Finer by Mass 0 20 40 60 80 100 Figure 32. Total metal Mass measured for selected metals normalized to 1000 g of total PM on the left. PSD of PM for all sites and percent finer by mass of the median values is shown on the right.

PAGE 77

77 As ( mol/m2) 0.00 0.02 0.04 0.06 0.08 As As [mg/kg] 0 1 2 3 4 5 Cumulative Mass (%) 0 20 40 60 80 100 As = 0.962 = 0.086 r2 = 0.99 TM = 4.75 mg/kg Al [mg/kg] 0 500 1000 1500 2000 2500 Cumulative Mass (%) 0 20 40 60 80 100 Al = 0.920 = 0.077 r2 = 0.99 TM = 4418 mg/kg Fe [mg/kg] 0 1000 2000 3000 4000 5000 Cumulative Mass (%) 0 20 40 60 80 100 Fe = 0.999 = 0.057 r2 = 0.99 TM = 15021 mg/kg Fe ( mol/m2) 0 30 60 90 120 Fe Al ( mol/m2) 0 15 30 45 60 Al Particle Diameter ( m) 9500 4750 2000 850 600 425 300 250 180 150 106 75 63 53 45 38 25 Cr [mg/kg] 0 2 4 6 8 10 Cumulative Mass (%) 0 20 40 60 80 100 Cr = 0.981 = 0.047 r2 = 0.98 TM = 15.72 mg/kg Particle Diameter ( m) 4750 2000 850 600 425 300 250 180 150 106 75 63 53 45 38 25 Cr ( mol/m2) 0.00 0.06 0.12 0.18 0.24 Cr Figure 3 3. Distribution of metal mass as a function of particle diameter for selected metals (Al, Fe, As, Cr). Results are illustrated on a surface area (column 1) and a mass basis (column 2) with measured cumulative distribution modeled with a cumulative gamma distribution. (TM = Total Mass of Metal Species)

PAGE 78

78 Pb ( mol/m2) 0.00 0.01 0.02 0.03 0.04 0.05 Pb Cd ( mol/m2) 0.0000 0.0003 0.0006 0.0009 0.0012 0.0015 Cd Cu ( mol/m2) 0.0 0.1 0.2 0.3 0.4 Cu Pb [mg/kg] 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Cumulative Mass (%) 0 20 40 60 80 100 Pb = 1.031 = 0.039 r2 = 0.98 TM = 8.36 mg/kg Cd [mg/kg] 0.00 0.03 0.06 0.09 0.12 0.15 Cumulative Mass (%) 0 20 40 60 80 100 Cd = 0.786 = 0.104 r2 = 0.98 TM = 0.20 mg/kg Cu [mg/kg] 0 3 6 9 12 15 Cumulative Mass (%) 0 20 40 60 80 100 Cu = 0.860 = 0.120 r2 = 0.99 TM = 33.10 mg/kg Particle Diameter ( m) 4750 2000 850 600 425 300 250 180 150 106 75 63 53 45 38 25 Zn ( mol/m2) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Zn Particle Diameter ( m) 9500 4750 2000 850 600 425 300 250 180 150 106 75 63 53 45 38 25 Zn [mg/kg] 0 5 10 15 20 25 Cumulative Mass (%) 0 20 40 60 80 100 Zn = 0.931 = 0.050 r2 = 0.98 TM = 78.80 mg/kg Figure 3 4. Distribution of metal mass as a function of particle diameter for selected metals (Cu, Cd, Pb, Zn). Results are illustrated on a surface area (column 1) and mass basis (column 2) with measured cumulative distribution modeled with a cumulative gamma distribution. (TM = Total Mass of Metal Species)

PAGE 79

79 As Al Cd Cr Cu Fe Pb Zn Values 0.0 0.5 1.0 1.5 2.0 2.5 As Al Cd Cr Cu Fe Pb Zn Values 0.00 0.05 0.10 0.15 0.20 0.25 Site As Al Cd Cr Cu Fe Pb Zn 1 103A 0.89 0.10 0.86 0.09 1.26 0.02 0.94 0.06 0.96 0.05 0.96 0.06 1.11 0.03 1.14 0.03 1 103B 0.88 0.12 0.87 0.11 1.18 0.03 0.84 0.09 0.91 0.06 0.98 0.06 1. 01 0.04 1.09 0.04 11209 1.03 0.04 1.04 0.03 1.50 0.01 1.30 0.02 1.16 0.01 1.22 0.02 1.50 0.01 1.31 0.02 2 1203 0.99 0.06 0.80 0.10 0.97 0.04 1.01 0.04 0.85 0.05 1.17 0.04 1.13 0.03 1.03 0.03 2 1204 1.46 0.02 1.21 0.03 1.76 0.02 1.61 0.02 1.38 0.02 1.64 0.02 1.33 0.03 1.39 0.02 2 1205 1.55 0.11 1.41 0.11 1.13 0.15 1.15 0.10 1.16 0.12 1.35 0.10 1.62 0.03 1.32 0.08 31203 1.47 0.06 1.56 0.06 1.35 0.07 1.22 0.05 0.78 0.11 1.43 0.05 1.37 0.03 1.32 0.05 4 1204 0.91 0.07 1.02 0.05 0.93 0.06 1.19 0.04 0.91 0.03 1.13 0.05 1.04 0.05 1.10 0.04 4 1205 1.45 0.09 0.86 0.18 1.40 0.13 0.84 0.15 0.88 0.15 0.90 0.15 0.86 0.12 0.88 0.14 5 1204 0.79 0.11 0.82 0.13 0.70 0.10 0.81 0.07 0.53 0.12 0.93 0.07 0.80 0. 05 0.73 0.09 61205 0.77 0.20 1.27 0.04 0.70 0.22 1.41 0.02 1.14 0.04 1.29 0.04 1.25 0.03 1.18 0.03 Median 0.96 0.09 0.92 0.08 0.79 0.10 1.00 0.05 0.74 0.05 1.00 0.06 1.03 0.04 0.93 0.05 Figure 3 5. Summary of gamma distribution parameters ( ), all coefficient o f determinations (r2) exceed 0.94. The median values shown (white circles) are the result of taking the median value of incremental metal mass data for all eleven sites, then modeling the incremental metal mass with a gamma distribution.

PAGE 80

80 Jan Feb Mar Apr May Jun Jul Intensity (mm/hr) 2 4 6 8 10 Cumulative Rainfall (mm) 0 100 200 300 400 500 Cumulative Watershed Yield (m3) 0 300 600 900 1200 1500 Incremental Precipitation Cumulative Precipitation Cumulative Yield Figure 3 6. The incremental precipitation for the watershed, cumulative precipitation for the watershed, and cumulative watershed yield into the basin. The period of simulation was from 1Jan08 through 31July08. Jan Feb Mar Apr May Jun Jul Temperature Degrees C -30 -20 -10 0 10 20 30 40 Hourly Temperature Data 2008 Mean Daily Temperature 20 Year Mean Daily Temperature Figure 3 7. Measured temperature values for the modeled year. Hourly 2008 data is shown along with daily mean temperature values for both 2008 and a twenty year mean for 19892008. The period of simulation was from 1Jan08 through 31July08.

PAGE 81

81 Jan Feb Mar Apr May Jun Jul Surface Overflow Rate (mm/s) 0.00 0.02 0.04 0.06 0.08 0.10 Surface Overflow Rate (mm/s) 0.00 0.02 0.04 0.06 0.08 Frequency (%) 0.01 0.1 1 10 100 Storage, S (m 3 ) 200 400 600 800 1000 1200 S = 379.8 x Z 1.1808R2 = 0.9981 SMAX = 975.02 m 3 Storage, S (m 3 ) 0 200 400 600 800 1000 1200 1400 Stage, Z (m) 0.0 0.5 1.0 1.5 2.0 2.5 Incremental Storage Cumulative Storage Stage/Storage Curve S = 139.39 x Z 2.4488R2 = 0.9997 SMAX = 1322.95 m3 Basin 1 Basin 2 Influent Jan Feb Mar Apr May Jun Incremental Flow (m 3 /s) 0.00 0.02 0.04 0.06 0.08 Effluent Jan Feb Mar Apr May Jun Jul Cumulative Flow (m3/s) 0 20000 40000 60000 80000 Figure 38. The top left graph shows influent flow rates over the modeled period, effluent flow rates are shown in the top right graph. Surface overflow rates are shown in the center figures, with the time series on the left and the frequency distribution on the right. The stage/storage curves for the sedimentation basins are shown in the bottom plots. The period of simulation was from 1Jan08 through 31July08.

PAGE 82

82 Figure 3 9. Total effluent metal and PM mass. Values assume no partitioning of metals into the aqueous phase. Note the difference in scale where total metal mass is illustrated in g and total mass in kg. Mass concentrations are the summation of PM (d > 25m) and across the entire simulation with a total effluent volume of 59825 m3. The period of simulation was from 1Jan08 through 31July08. Particle Diameter ( m) 4750 2000 850 600 425 300 250 180 150 106 75 63 53 45 38 25Percent Retained by Mass (%) 0 20 40 60 80 100 Effluent Influent Figure 3 10. The median and range of influent and effluent PSDs. Fe Al Zn Cu Cr Pb As Cd PM Total Metal Mass (g) 1e-2 1e-1 1e+0 1e+1 1e+2 1e+3 1e+4 1e+5 1e+6 1e+7 1e+8 1e+9 1e+10Total Mass (kg) 1e-2 1e-1 1e+0 1e+1 1e+2 1e+3 1e+4 1e+5 1e+6 1e+7 1e+8 1e+9 1e+10 Effluent Influent 5 th percentile 10 th percentile 25th P ercentile Mean Median 75 th P ercentile 90 th percentile 95 th percentile

PAGE 83

83 CHAPTER 4 PARTITIONING AND SPECIATION OF METALS IN URBAN SNOWMELT EXPOSED TO TRAFFIC ACTIVITIES IN THE LAKE TAHOE WATERSHED I ntroduction As a natural resource and as a source of water to sustain the built environment, s now snowmelt is vital for the western United States. As western areas were settled water resources such as Lake Tahoe, in part supplied by snowmelt, were promoted as a water supply for agriculture, industry and domestic demands; and a water source for silviculture growth while providing waterway nourishment for aquatic and terrestrial life (Feth et al. 1964). However, as pavement and the motor vehicle now dominate our built environs, including those in the Lake Tahoe watershed snow and snowmelt on pavement is a hazard and bare pavement policies are common to provide safe and uninterrupted traffic in regions with snow (Glenn and Sansalone 2002). While snowmelt is a water source, plowed snow banks alongside roadways form a temporary linear reservoir for traffic generated constituents such as metals and particu late matter ( PM ) In urban environs with cold climates, a significant fraction of the annual metal and PM based loads occurs during snowmelt and spring runoff event sequences (Oberts 2000). Whether generated by rainfall or snow, runoff impacted by tra nsportation infrastructure and activities such as traffic transports metals at levels that can impact receiving water bodies, aquatic life and potentially the food web (Glenn and Sansalone 2002). Similar to rain, s now is poorly buffered with an acidic pH. However, source area runoff and snowmelt from impervious surfaces can contain high levels of particulate matter (PM) and increase d metal concentrations C haracterization is difficult due to the variability of the snowmelt process (Servos 1987) and for pa vement source areas the intermittency of snowmelt sheet flow Higher volumetric runoff loadings and gravimetric pollutant concentrations are often associated with rainonsnow runoff or snowmelt (Oberts 2000). Sources of pollutants in transportation corr idors include traffic dry

PAGE 84

84 deposition, pavement tire and vehicular abrasion, deicing and anti skid agents, roadway surface deterioration, and urban litter (Oberts 2000). A study in Sweden shows that a significant fraction (32 to 57%) of metals (Al, Cd, Cu, Fe, Pb, and Zn) is within the one meter of the edge of the pavement (Reinosdotter et al. 2006). Ying and Sansalone (2008) illustrated an order of magnitude decrease in dry deposition flux for PM within five meters of the pavement edge. M etals in snowmelt will partition between aqueous and particulate phases. In the aqueous phase metals can be further differentiated as separate species The concentration of metal species along with water chemistry is a basic consideration when considering the bioavailabi lity of metal species and the effects on biota (Morrison 1989) along with fate of these metal species For example, divalent Cu is toxic while Cu complexed with natural organic matter (Cu NOM) is less toxic (Morrison 1989) and less reactive. H ydrated Zn ions are highly bioavailable species and there is a high correlation between the amount of total Zn in runoff and the environmental impacts of Zn (Wallinder 2001). When a cidic snowmelt is a major component in spring runoff aqueous aluminum species can b e mobilized by low pH impacting aquatic species, especially juvenile stages of salmon (Campbell 1992). M etal mobility can also be greatly influenced by winter maintenance practices, such as NaCl as a d e icing salt, increasing the mobility of Cd, Pb, and Z n and lowering pH (Nelson et al. 2006). Deicing salts, depending on their source and processing, can contain Zn, Cu, Ni and As (Goldman and Hoffman 1975). A study in Ontario reports the concentrations of selected metals in river water, rain and snow ; while also indicating low particulate fractions for Al (1.2%), Cu (19.5%), Fe (2.6%), and Ni (5.3%) in rooftop snowmelt (Lu et al. 1996). Wallinder et al. (2001) report that at pH values between 4.5 and 7, over 95% of total Zn identified in the runoff as ioni c for Zn roof panels exposed to runoff. A study of urban snowmelt exposed to average daily traffic (ADT) (ADT >

PAGE 85

85 20,000) in two cities in Sweden reports dissolved Cu, Pb, and Zn of 7.0, 0.08, 1.5 g/L and total concentrations at 1,022; 217; and 2,233 g/L, respectively (Reinosdotter and Viklander 2006). Aqueous metals are mobile, and treatment options such as adsorptive filtration are commonly required (Sansalone and Teng 2005, Sansalone 1999). Passive infiltration treatment methods for snowmelt can be mo re viable in less developed areas compared to developed areas where metals tend to adsorb to PM (Oberts 2003), unless adsorptive filtration systems are designed for snowmelt with de icing and grit loadings. In cold climates, infiltration systems are less e ffective due to frozen soils which reduce infiltration rates (Al Houri et al. 2009, Oberts 2003). O bjectives The goal of this study is to examine the speciation of metal elements measured in traffic impacted snowmelt. Towards this goal snow samples impacted by traffic in the Lake Tahoe watershed and control snow samples are collect ed and examined for water chemi stry parameters including pH, alkalinity, PM, total dissolved solids (TDS) anions and metal concentrations required for partitioning and speciation. The first objective i s to determine the partitioning of metal elements between the dissolved and particulat e phases. The second objective i s to determine, quantify and compare metal species for 6 locations and control locations in the Lake Tahoe watershed. The final objective is to examine the influence of temperature on speciation. Background The partitio ning of metals between dissolved and particulate phases is a function of pH, alkalinity, hardness, residence time, hydrodynamics, chemistry of the metal element, and entrained PM properties (Glenn and Sansalone 2002, Sansalone and Buchberger 1997). Metal partitioning is important for transport, bioavailability considerations, and in the design of treatment systems facilitating the mass transfer or separation of metals (Sansalone and

PAGE 86

86 Buchberger 1997). If equilibrium partitioning is considered as a two pha se model the dissolved and particulate fractions, fd and fp respectively, can be defined. Accounting for the gravimetric concentration of PM, the equilibrium partitioning coefficient, Kd (L/kg), as the ratio of these phases can also be defined (Thomann and Mueller 1987) at a particular pH and redox level. The mass component of the Kd equation differs for the unsettled and settled calculations where SSC is used in the unsettled calculations and TSS is used in the settled calculations. D m P C C Kd s d ( 41) In this expression, D is the dissolved mass of a metal ( g/L), and P is the particulate bound mass of a metal ( g/L). The dissolved fraction (fd) and the particulate bound fraction (fp) a re also defined m K 1 1 C C P D D fd T d d ( 42) m K 1 m K C C P D P fd d T p p (4 3) fd + fp = 1 ( 44) In these expressions m in this case is the total suspended solid (TSS) mass in terms of mass/volume of aqueous solution (OConnor 1988, Glenn and Sansalone 2002). Within the aqueous phase, metal speciation is a funct ion of the metal and aqueous matrix water chemistry including parameters of pH, chemical concentrations and aqueous complexes. Common complexing agents include: hydroxides, carbonates, organic compounds or indices such as dissolved organic matter (DOM), phosphates, chlorides, nitrates, and sulfates. Equilibrium speciation of aqueous or adsorbed species has been commonly examined with the

PAGE 87

87 model, MINTEQ (EPA 2005) for ambient waters and runoff. The model utilizes an iterative method to solve reaction equat ions and mass action expressions using log equilibrium constants to determine a numerical solution to equilibrium concentrations (Allison et al. 1991). Required inputs for this model include temperature, pH, aqueous concentrations of cations, anions, liga nds and PM along with a charge balance to determine the measurement of the major ionic species (Dean et al. 2005). An extensive thermodynamic database is available for equilibrium constants at 25C, these values are adjusted for temperature using the van t Hoff equation. 2 1 0 r T1 eq T2 eqT 1 T 1 R H K K ln (4 5) Keq is the equilibrium constant at T1 and T2, 0 rH is the standard state molar enthalpy, R is the universal gas constant, and T1 and T2 are 25C and the target temperature. At temperatures greatly exceeding 25C, this method is found to have significant errors due to the assumption that the enthalpy of reaction is independent of temperature (Allison et al. 1991). The system is modeled as an open system with species interacti ng with the atmosphere. Altitude corrections on the atmospheric CO2 pressure are an option within the model but were not utilized in this study. The elevation correction in the model corrects for the partial pressure of CO2 in the atmosphere. There are t hree DOC sub models: Gaussian, NICA Donnan, and SHM. The most commonly used submodel is the Gaussian model (Hesterburg et al. 2006, Dean et al. 2005, Birceanu et al. 2008, Gardner et al. 2007, Buerge Weirich and Sulzberger 2004, Gnecco et al. 2008). The Gaussian sub model was used in this study and assumes that NOM is a complex mixture of functional groups and that the frequency of occurrence of a binding site is normally distributed with respect to the log K value for metal or proton binding (Grimm et al. 1991). Therefore each reaction involving DOM has an associated mean and standard deviation for log

PAGE 88

88 K; as provided in the database of DOM reactions with Al, Cd, Cr(III), Cu, Fe(III), Mg, Pb, Ni, and Zn proposed by Susetyo et al (1991). The NICA Donnan submodel (NonIdeal Competitive Adsorption) accounts for nonideal binding to heterogene ous ligands and strongly influenced by intrinsic chemical heterogeneity of the humic material (Benedetti et al. 1995). Required inputs for this model inc lude proton dissociating groups for type 1 and 2 (mol/g), Donnon volume parameter, width of distribution for type 1 and 2 sites, and nonideality parameters for the hydrogen ion for both type 1 and 2 sites (Allison et al. 1991). This model tends to underes timate the H+/M+2 exchange ratio at high pH values and for Cu binding (Kinniburgh et al. 1996). The SHM submodel is a di s crete ligand model with 8 protonbinding sites for humic or fulvic acids; there are seven adjustable parameters required for the prot on dissociation reactions including: proton dissociating groups (mol/g), S tern layer capacitance, spherical radius, site density, gel fraction parameter, concentration of B sites (% of A type sites), log K of both type A and B groups, the the delta pKa for both type A and type B groups (Allison et al. 1991). The Type A groups include s the 4 strongest acid sites (typically carboxylic acid groups), while the Type B groups typically represent weaker acids (Gustafsson 2001). It is very difficult to optimize the protonbinding parameters simultaneously (Gustafsson 2001). M ethodology Sampling Locations The six sampling locations are along US Highway 50 within the Lake Tahoe watershed traversing the California and Nevada state line All samples represent plo wed pavement snow, whether taken directly from the edge of the pavement or from the local snow management site in South Lake Tahoe. Samples taken from the South Lake Tahoe snow management site represent a combination of US 50 and local roadway snow from S outh Lake Tahoe. Sampling location and dates are summarized in Table 1. Control snow, obtained from areas removed from

PAGE 89

89 direct anthropogenic influences such as vehicular or human traffic is also exampled with respect to these roadway snow locations. Tra ffic data along US 50 is similar for both states; the data is averaged over a yearly duration, without examining seasonal influences. A 21.4 km section of US 50 in California from Echo Lake Road to the state line has traffic averages of 12,200 to 36,000 ve hicles per day measured in 2004 (TMA 2004). The 8 km section of US 50 in Nevada from the California state line to Zephyr Cove has traffic of 14,800 to 32,000 vehicles per day measured in 2000 (NDOT 2006). Snow Sampling Snow samples are taken from each of the sampling locations as plowed roadside snow or from the snow management site where snow is windrowed by blowers; 300 m in length and varied from 3 to 5 m in height with a total volume of approximately 8500 m3 at the time of sampling. Snow that had no apparent anthropogenic terrestrial or vehicular influences, but is subject to atmospheric deposition or biogenic deposition from the surrounding coniferous canopy is sampled as a control. For all locations samples are packed in 4L wide mouth polypropy le ne bottles. The snow subsamples to fill each bottle are taken from approximately a 1 m2 area unless sampled from a vertical snow bank; in which case the entire depth was manually cored and recovered. All samples were taken in duplicate. Table 41 summa rizes all site locations All samples are taken while snow is frozen and snow is kept in a solid state during transport to the laboratory. The samples are stored in the freezer below 0C until analyzed. Frozen samples are thawed overnight to room tempe rature before analys is Water chemistry parameters are analyzed on the total sample. To separate settleable PM samples are poured into 1L Imhoff cones. After 1 hour, t he supernatant i s decanted into 1L polypropylene bottles and stored for analysis (T = 60 samples)

PAGE 90

90 Laboratory Analysis Water chemistry p arameters Water chemistry parameters such as pH, suspended sediment concentration (SSC), total dissolved solids (TDS), turbidity, dissolved chemical oxygen demand (CODd), and alkalinity are measured within 24 hours of thawing. Conductivity pH and TDS are measured with a conductivity meter in triplicate; standardized with a 3 point curve prior to analyses with an Orion 4 Star pH ISE meter. The entire volume of sample supernatant is filtered through a pre weighed nominal 1 m glass fiber filter and allowed to dry to determine SSC (ASTM 1999). Turbidity is measured with a Hach 2100AN turbidimeter and also measured in triplicate. The supernatant is further fractionated to obtain the dissolved phase by pressure filtering through a 0.45 m membrane and also analyzed for CODd and alkalinity. The membrane filters are biologically inert and made a mixture of cellulose acetate and cellulose nitrate that will not contaminate filtrate with fibers or partic les. Alkalinity is measured in triplicate using Method 2320B (APHA 1998). Control snow samples are analyzed without filtration. Sample digestion for metal partitioning The dissolved phase samples are acidified to a pH of 2 using trace metal grade nitric a cid. The particulate phase captured on the filter is digested using a nitric hydrochloric acid method using SW 846 Method 3015 (USEPA 1990). For method QC a soil standard (Environmental Resource Associates, QC D048 540) and a blank (empty vial) is digest ed with every eighteen samples. The soil standard consists of certified values of Al, As, Ba, Cd, Cr, Co, Cu, Fe, Pb, Mg, Mn, Hg, Mo, Ni, K, Se, Ag, Na, and Zn. The values for the measured standard are within the "performance acceptance limits" for each metal element within 20% of the "certified value". The filtrate is then analyzed on a Perkin Elmer Elan 9000 Inductively Coupled Plasma Mass

PAGE 91

91 Spectrophotometer (ICP MS). The protocol outlined by the EPA in SW 846 Method 6020 for ICP MS analysis is utiliz ed (EPA 1996). I on analysis Snowmelt supernatant is pressure filtered through a 0.45 m membrane filter to obtain the dissolved phase for ion analysis. The di ssolved fraction of the sample i s analyzed for PO4 1, NO3 1, NH3, SO4 2, Cl1 tannic acid and CODd using a DR/5000 spectrophotometer. A 5point standard calibration curve i s d etermined for each parameter. Dissolved organic carbon (DOC) i s analyzed with a Shimadzu TOC 5050A analyz er. The DOC aliquot i s fractionated as described above but separately acidified to a pH of 2 with trace metal sulfuric acid. Speciation analysis The water chemistry from all snowmelt and control samples a re input into MINTEQ version 2.53. The MINTEQ DOC Gaussian submodel is chosen because it is extensively used in t he literature and the parameter requirements for the NICA Donnan, and SHM submodels are significant (Christensen and Christensen 1999) and are unavailable in this study. Control snow samples are analyzed without filtration. Equilibrium speciation is mo deled at constant values of 1, 5, 10, and 25C. Temperature values of 1, 5, and 10C are common to the Lake Tahoe region. The 25C temperature i s used for comparison purposes and represents a reasonable upper limit of temperature that asphalt pavement so urce area rainfall runoff achieves during the summer in paved developed catchments of Lake Tahoe (Magill and Sansalone 2010) Model inputs included temperature, pH, Cl1, CO3 2, DOC, NH3N NO3 1, PO4 3, SO4 2, Na+1, Mg+2, Ca+2, Ag+1, Ba+2 (Table 42), and the metals reported herein (Figure 4 2). Water chemistry data for snowmelt resulted in charge balances from 1 to 30%. Speciation results a re compiled for all snowmelt and control locations and sampling events. The speciation results of nine metals a re reported herein.

PAGE 92

92 Results A summary of the water chemistry parameters for snowmelt and control samples is shown in Tables 4 2 and 43, respectively. The mean, median, minimum, and maximum values for phosphate, nitrate, ammonia nitrogen, sulfate, chlor ide, tannic acid, CODd, DOC, dissolved silica and turbidity are tabulated. All parameters are utilized in the speciation modeling with the exception of tannic acid, silica, CODd, turbidity, suspended PM, settleable PM, and sediment PM The inputs into MI NTEQ are shown in Tables 44 and 45 for the snowmelt and control snow samples, respectively. A table of stability constants (log Ki) for the metal ligand complexes seen in the MINTEQ output at 25 C is shown in Table 46. Results for the particulate bound and dissolved metal concentrations are shown in Figure 42. Fe has the highest while Cd has the lowest median values. Both particulate bound and dissolved median concentrations illustrate similar trends: Fe > Al > Zn > Cd. The mean concentrations of dissolved Cu, Pb, and Zn exceed reported values from Reinsodotter and Viklander from high traffic areas in Sweden (2006). Dissolved fractions (fd) for snowmelt are s ummarized in Figure 43 and the metal mean and median are below 0.25, and few samples have dissolved metal fractions higher than 0.5. Metals species are shown in order of decreasing median value. The equilibrium partitioning coefficient (Kd) values in L/kg, summarized in Figure 4, are also shown in order of decreasing median value. Both the unsettled (t = 0 min.) and settled (t = 60 min.) are summarized. The unsettled snowmelt exhibits much higher Kd values than the settled snowmelt as a result of PM bound metals settling in 60 minutes. Unsettled Kd value s range from 5.01x107 to 5.25 x104 for Pb and Zn, respectively, while the settled values range from 5.87x104 to 1.79x103 for Fe and Cd, respectively. With the loss of PM bound metals from solution, the predominance of the

PAGE 93

93 dissolved metals increases, decreasing the value of Kd. Additionally, much smaller variation in partitioning is observed in the settled samples. With respect to the role of temperature, metals species are shown in order of decreasing medi an value in Figure 4 5 through 4 7. The 1, 5, and 10C temperature are typical for the Lake Tahoe region during periods of snowmelt and storage, whereas 25C represents the reasonable upper limit of rainfallrunoff from paved areas and serves as a comparison level. A fifth plot on the right of each met al specific series of speciation plots illustrates species change with increasing temperature with respect to 1C for divalent metal species of each metal element except for Al and Fe. For Al and Fe, Al(OH)2 +1 and Fe(OH)3(aq) are shown, respectively. The se Al and Fe hydroxide species are identified as potentially toxic to freshwater salmon and trout (Poleo 1991, Poleo et al. 1995) or cause physiological changes in fish (Brenner and Cooper, 1978), respectively. In all plots, the snowmelt speciation result s are shown as nonparameter ( box and whisker) plots, while control snow are also shown as a bolded scatter plots in each plot. The term (aq) following a species indicates an uncharged aqueous species. All speciation data are modeled at sea level. Figure 45 summarizes the results for the primary species of Al, Cd, and Cr. The dominant species comprising 79 to 100% of dissolved snowmelt Al are Al(OH)4 1, Al(OH)3 (aq), Al2(OH)2CO3 +2, and Al(OH)2 +1. The dominant Al species for snowmelt is Al(OH)4 1; illustrating a 25% increase in median concentration from 1 to 25C. In contrast, a reverse trend (53% decrease) is observed for control snowmelt across this temperature range. Al is the only metal examined where the dominant species (79 to 100% of diss olved Al) and the control samples (0 to 99.9% of dissolved Al) are not identical. Although not graphically illustrated for control snowmelt, ionic Al+2 and Al DOM are the dominant species for control snow (0.03 to 33.22%

PAGE 94

94 and 0.04 to 94% of dissolved Al, r espectively). At 25C the median value of Al2(OH)2CO3 +2 decreases to 1% of median value at 1C value. The primary snowmelt Cd species (86.3 to 99.9% of dissolved Cd) are: Cd+2, CdCl +1, CdDOM, CdCO3 (aq), and CdCl2 (aq). Cd+2 is the dominant species w ith a 3.7% decrease from 1 to 25C. CdCl2 is not present in control snow at any temperature and CdCO3 is only present at 25C. The medium concentration of CdCO3 increases 78% and while CdCl2, increases by 10% from to 1 to 25C. Cr+2 and CrOH+1 comprise 100% of the snowmelt Cr species. In contrast to Al and Cd species, there are no significant differences with increasing temperature for snowmelt or control samples with a 2% increase in median value for Cr+2 from 1 to 25C. Figure 4 6 illustrates the spec iation results for Cu, Fe and Pb. The primary snowmelt Cu species comprise 43 to 99.9% of dissolved Cu as CuDOM, CuCO3 (aq), Cu+2, and CuOH+1. CuDOM is the dominant species with a median value that decreases 23% decrease from 1 to 25C. The median val ue of CuOH+1 and CuCO3(aq) increase with increasing temperature; 67 and 61% in the snowmelt samples, respectively. The control samples show an increase of 107% from 1 to 25C for CuCO3 and 104% for CuOH+1 from 1 to 25C. Fe results in Figure 4 6 illustrate FeOH+2, Fe(OH)2 +1, Fe(OH)3 (aq), and Fe(OH)4 1 species comprising 99.80 to 100 % of dissolved Fe The median value of the dominant species, Fe(OH)2 +1 decreases by 12% with increasing temperature while the control increase by 14.1% from 1 to 25C. Steady decreases in median values occur for snowmelt and control samples for FeOH+2 with reductions of 34.3% and 21.7%, respectively, from 1 to 25C Fe(OH)4 1 illustrates a large temperature dependence and the me an snowmelt value increases 180% over the 1 to 25C range for the snowmelt samples Significant increases occur for Fe(OH)3(aq); a species shown to

PAGE 95

95 cause physiological changes in fish (Brenner and Cooper, 1978); for snowmelt and control samples with increases of 78.4% and 95.8%, respectively Pb DOM, PbCO3 (aq), Pb(OH)+1, Pb+2, and PbCl+1 comprise 63.2 to 99.9% of dissolved Pb. A steady decrease in the snowmelt median of Pb+2 and PbDOM occur over the 1 to 25C range; decreasing by 24 and 15%, respectively. Alternatively, the control samples show increases of 0.2% and 9% for PbDOM and Pb+2, respectfully, from 1 to 25C. Increases in snowmelt median values occur for PbOH+1 and PbCO3 over the temperature range, with values increasing by 80 and 25%, respectively. The control snow also illustrates increases, the PbOH+ 1 median increases by 69% and PbCO3 by 115%. Figure 4 7 summarizes the speciation of Mn, Ni and Zn. The primary Mn species are Mn+2, MnCO3 (aq), MnCl+1, and MnSO4 (aq) representing 98.2 to 99.9% of dissolved Mn. The dominant snowmelt species is Mn+2 with a 4.8% decrease in the median concentration from 1 to 25C. With respect to control snow, the median value of MnCl(aq) remained relatively constant with a 12% increase for Mn+2 from 1 to 25C. As temperature increases fr om 1 to 25C the median concentration increases 84% for MnCO3(aq), 4% for MnCl and 21% for MnSO4(aq). For control snow median values for MnCO3 and MnSO4 increased 111% and 27% respectively from 1 to 25C. With respect to Ni species, Ni+2, NiDOM, NiCO3(aq ) and NiHCO3 +1 are the primary species with 80.4 to 99.9% of dissolved Ni. Snowmelt Ni+2 and Ni DOM illustrate 1.2 and 2.4% increases, respectively, from 1 to 25C. Median values of NiCO3(aq) increase 76% in snowmelt from 1 to 25C with an increase of 45% in control snow, primarily from 10 to 25C. Snowmelt NiHCO3 +1 increases 99% from 1 to 25C, and the control samples increase 31% from 1 to 25C. Zn results indicate that 87.6 to 99.9% of dissolved Zn is Zn+2, Zn DOM, Zn(OH)2(aq) ZnOH+1, and ZnCO3(aq). With increasing temperature, snowmelt Zn+2, Zn DOM, Zn(OH)2

PAGE 96

96 (aq), and control Zn+2 and Zn(OH)2 (aq) median values decrease 7.4, 4.9, 4.1, 1.0, and 1.6%, respectively. Snowmelt ZnOH+1, ZnCO3 (aq) and control Zn DOM median values increase 555, 70, and 3%, respectively. ZnCO3 in control snow is increases 82% from 5 to 25C. Conclusions S now exposed to urban activities, specifically vehicular traffic, is a r eservoir and source for dissolved and particulate bound metals. For a six year period, traffic impacted areas in the Lake Tahoe watershed are analyzed for metal s Results indicate that Al and Fe are of the highest concentration in both phases and Cd is the lowest. All results indicate that the majority of the metal mass is associ ated with the par ticulate phase; fd values typically below 0.25, in agreement with Swedish and American studies (Reinosdotter et al 2006, Glenn and Sansalone 2002). Speciation results illustrate several temperature based trends for metal elements complexes with carbonat es, DOM, or remain as ionic species. The median values of all CO3 2complexed metals increases with increasing temperature, and is due to the decrease in the CO3 2 pKa value with increasing temperature, shifting the carbonate system equilibrium. This results in the carbonate species increasing in dominance and complexation with temperature. The reverse trend is observed for DOM complex metals in snowmelt where th e median values decrease with increasing temperature with the exception of Cd with less than a 1%. All divalent snowmelt species median values decrease with increasing temperature which is the expected trend as more species tend to become less soluble wit h increasing temperature (Benjamin 2002). Based on the California Toxics Rule (CTR) Cu exceeds the both the Criterion Maximum Concentration (CMC) and Criterion Continuous Concentration (CCC), and Pb exceeds the CCC for snow samples, but not control snow Fe also exceeds the NPDES permit for discharge to a surface water body. The dissolved fraction of Cd, Ni and Zn in snowmelt or control snow do not exceed

PAGE 97

97 the CTR discharge limits (USEPA 2000). Total Fe exceeds the NPDES permit for discharge to surface w aters (0.50 mg/L) with particulate and dissolved values ranging from 14 to 4282 mg/L (mean of 381 mg/L) and 0.1 to 3.0 mg/L (mean of 0.67 mg/L) (USEPA 2003). Fe is a concern based Goldman and Hoffman (1975) who indicate that Lake Tahoe is Fe limited Fe influences algal productivity much like N and P, and is a requirement for algae to carry out nitrate reduction (Chang et al. 2002). Discussion This study illustrates that while the majority of the metal mass is associated with PM, there is sufficient di ssolved concentration, to exceed specific regulatory levels for Lake Tahoe However, the suspended PM and dissolved fractions of metals are typically either weakly treated or not treated by most UOPs or drainage systems and discharged into receiving water s Secondary treatment of snowmelt through adsorptive filtration is a viable UOP to follow primary clarification in a centralized management strategy of snowmelt; for example at the snow management site in S. Lake Tahoe Additionally, the particulate bound metals representing most of the metal mass can repartition to aqueous species with water chemistry conditions in Lake Tahoe; or re partition from the labile coarser fraction of PM during drainage transport in the drainage system to Lake Tahoe or within below grade unit operations that are not maintained. All metal elements exhibited median fd values less than 0.25 due to very high concentration of PM in source area snow, higher pH and higher alkalinity. While results are contrary to rainfall runoff studies that indicate that source area runoff metals (Cd, Cu, Pb and Zn) are primarily dissolve d (Sansalone and Buchberger 1997), snowmelt fd values ranging from 0.006 to 0.45 are reported in other studies (Sansalone et al. 1996, Lancaster et al. 2009). With respect to control, a centralized treatment system would allow for recovery of metals and PM as well as maintainability of UOPs as opposed to decentralized systems that typically

PAGE 98

98 lack oversight, monitoring and management. Based on results presented; a combination of primary clarification followed by secondary adsorptive filtration for metals u nder competitive conditions is required. Given the high levels of PM and need to physically remove the snow in many cases from roadway right of ways and parking areas; centralized treatment will also facilitate maintenance and residual recovery. During p eriods of dry pavement, pavement cleaning is a critical source control practice will reduce pollutant loads on a continual basis. N omenclature (aq) uncharged aqueous species cd dissolved metal concentration cp particulate metal concentration CCC Criterion Continuous Concentration ( g/L), the highest concentration of a pollutant which aquatic life can be exposed for 4 days CMC Criterion Maximum Concentration ( g/L), the highest concentration of a pollutant to which aquatic life can be exposed for a short pe riod of time CTR California Toxic Rule D dissolved mass of a metal element (mg) DOC dissolved organic carbon DOM dissolved organic matter fd dissolved fraction ( ) fp particulate fraction ( ) ICP inductively coupled plasma ICP AES inductively coupled plas ma atomic emission spectroscopy ICP MS inductively coupled plasma mass spectrometer

PAGE 99

99 Kd partition coefficient between particulatebound mass and dissolved mass (L/kg) Ki stability constant for metalligand complexes Keq equilibrium constant NICA nonideal competitive adsorption P particulate bound metal mass (mg) R universal gas constant QA/QC quality assurance/quality control SHM Stockholm humic model SSC suspended solids concentration (mg/L) T temperature TDS total dissolved solids (mg/L) TOC total organic carbon UOP unit operation and processes 0 rH standard state molar enthalpy

PAGE 100

100 Table 41. Summary of sites, locations, and sampling dates. Site # Sample Location Sample Date 1 103A Sierra Blvd@Barbara Ave, S. Lake T ahoe, CA 1/5/2003 1 103B Sierra Blvd@Barbara Ave, S. Lake Tahoe, CA 1/5/2003 1 1203 Sierra Blvd@Barbara Ave, S. Lake Tahoe, CA 12/20/03 2 1203 HWY US50W@Tunnel, NV 12/20/03 2 1204 HWY US50W@Tunnel, NV 12/27/2004 2 1205 HWY US50W@Tunnel, NV 12/21/2005 3 1203 HWY US50W@Firestation #5, NV 12/20/2003 4 1204 HWY US50W@Zephyr Cove, NV 12/28/2004 4 1205 HWY US50W@Zephyr Cove, NV 12/21/2005 5 1204 HWY US50W@Stateline, CA 12/29/2004 6 1205 HWY US50W@Lake Pkwy, S. Lake Tahoe, NV 12/21/2005

PAGE 101

101 Table 42. Summary of water chemistry parameters in snowmelt samples. Parameter Mean Median Min Max Count Method Tannic Acid (mg/L) 1.04 0.963 0.263 1.86 6 Tyrosine Method COD d (mg/L) 161.2 130.4 21.8 610.0 17 Reactor Digestion Method Silica (mg/L) 3.26 2.67 1.28 7.00 6 Heteropoly Blue Method Turbidity (NTU) 10743 1999 99050 470 17 Standard Method 2310B Alk (mg/L as CaCO3) 306 122 19.5 1308 17 Standard Method 2320B TSS (mg/L) 9556 277 4.62 51037 17 Standard Method 2540 D Settleable Solids (g/L) 182 74.4 2.91 1040 16 Standard Method 2540F Sediment (g) 2043 826 95 8761 16 ASTM D3977 97

PAGE 102

102 Table 4 3. Summary of water chemistry parameters in control snow samples. Parameter Mean Median Min Max Count Method Tannic Acid (mg/L) -0.015 --1 Tyrosine Method Silica (mg/L) -0.030 --1 Reactor Digestion Method COD d (mg/L) 54.1 50.0 5.26 130.0 5 Heteropoly Blue Method Turbidity (NTU) 11.9 4.50 0.452 45.8 5 Standard Method 2130B Alk (mg/L as CaCO3) 1.71 2.00 1.00 2.00 5 Standard Method 2320B TSS (mg/L) 6.33 5.00 1.00 14.5 5 Standard Method 2540 D

PAGE 103

103 Table 4 4. Summary of MINTEQ input data from snowmelt samples. The median charge balance for the snowmelt samples is 4.98% (pH= log{H+}) Parameter Mean Median Minimum Maximum Count Method Ag +2 (mmol/L) 1.13E 05 1.25E 05 1.00E 06 2.51E 05 14 Standard Method 3030B Al +3 (mmol/L) 2.41E 02 1.42E 02 2.59E 04 8.35E 02 17 Standard Method 3030B Ba +2 (mmol/L) 9.20E 04 4.50E 04 8.88E 05 5.59E 03 17 Standard Method 3030B Ca +2 (mmol/L) 9.20E 01 3.57E 01 5.03E 02 8.24E+00 17 Standard Method 3030B Cd +2 (mmol/L) 1.26E 05 1.51E 05 1.34E 08 3.24E 05 17 Standard Method 3030B Cr +2 (mmol/L) 7.93E 04 2.14E 04 2.51E 05 5.11E 03 17 Standard Method 3030B Cu +2 (mmol/L) 5.36E 04 4.21E 04 5.17E 05 2.49E 03 17 Standard Method 3030B Fe +3 (mmol/L) 1.19E 02 9.76E 03 2.14E 03 3.93E 02 17 Standard Method 3030B Mg +2 (mmol/L) 1.25E 01 5.75E 02 1.06E 02 9.87E 01 17 Standard Method 3030B Mn +2 (mmol/L) 3.16E 03 1.21E 03 7.50E 05 1.63E 02 17 Standard Method 3030B Na +1 (mmol/L) 1.76E+01 1.37E+01 1.18E 01 4.27E+01 8 Standard Method 3030B Ni +2 (mmol/L) 2.57E 04 1.24E 04 8.81E 06 8.58E 04 17 Standard Method 3030B Pb +2 (mmol/L) 6.27E 05 1.38E 05 4.84E 09 1.99E 04 17 Standard Method 3030B Zn +2 (mmol/L) 1.09E 03 8.84E 04 1.19E 04 3.67E 03 17 Standard Method 3030B Cl 1 (mmol/L) 2.45E+01 1.23E+01 5.45E 01 1.33E+02 17 Mercuric Thiocyanate Method CO 3 2 (mmol/L) 8.63E 01 7.33E 01 1.95E 01 2.08E+00 17 Standard Method 2320B DOC 2.8 (mmol/L) 3.16E+00 2.13E+00 5.22E 01 1.01E+01 16 Standard Method 5310B NH 4 +1 (mmol/L) 1.69E 02 7.06E 03 1.31E 03 6.06E 02 9 Salicylate Method NO 3 1 (mmol/L) 8.39E 03 6.86E 03 7.65E 04 1.87E 02 17 Cadmium Reduction Method PO 4 3 (mmol/L) 3.16E 03 1.99E 03 9.08E 04 9.92E 03 17 Ascorbic Acid Method SO 4 2 (mmol/L) 2.73E 01 1.14E 01 7.07E 03 2.18E+00 16 SulfaVer 4 Method pH ( ) 8.08 7.91 6.91 9.58 17 Standard Method 4500 H + B

PAGE 104

104 Table 4 6. Summary of MINTEQ inputs for control snow samples. The median charge balance for the control snow is 48.46%. (pH= log{H+}) Parameter Mean Median Minimum Maximum Count Method Ag +2 (mmol/L) 1.3E 05 1.3E 05 8.8E 07 2.7E 05 5 Standard Method 3030B Al +3 (mmol/L) 1.3E 05 1.3E 05 8.8E 07 2.7E 05 5 Standard Method 3030B Ba +2 (mmol/L) 3.8E 03 1.5E 03 1.5E 04 1.5E 02 5 Standard Method 3030B Ca +2 (mmol/L) 1.4E 04 7.3E 05 1.1E 05 4.4E 04 5 Standard Method 3030B Cd +2 (mmol/L) 8.0E 02 1.2E 02 6.7E 03 3.6E 01 5 Standard Method 3030B Cr +2 (mmol/L) 6.7E 06 2.9E 06 3.9E 07 2.2E 05 5 Standard Method 3030B Cu +2 (mmol/L) 3.9E 04 6.5E 05 2.3E 05 1.5E 03 5 Standard Method 3030B Fe +3 (mmol/L) 3.8E 03 1.4E 03 1.0E 03 1.4E 02 5 Standard Method 3030B Mg +2 (mmol/L) 6.9E 03 1.3E 03 1.2E 03 2.9E 02 5 Standard Method 3030B Mn +2 (mmol/L) 1.5E 04 5.2E 05 3.5E 05 5.6E 04 5 Standard Method 3030B Na +1 (mmol/L) 3.3E 02 3.3E 02 1.3E 02 5.4E 02 5 Standard Method 3030B Ni +2 (mmol/L) 1.4E 04 1.4E 04 4.0E 05 2.4E 04 5 Standard Method 3030B Pb +2 (mmol/L) 5.9E 05 1.9E 05 1.6E 06 2.1E 04 5 Standard Method 3030B Zn +2 (mmol/L) 7.7E 04 5.5E 04 2.0E 04 1.8E 03 5 Standard Method 3030B Cl 1 (mmol/L) 6.7E 06 2.9E 06 3.9E 07 2.2E 05 5 Mercuric Thiocyanate Method CO 3 2 (mmol/L) 7.4E 02 2.5E 02 2.5E 03 2.8E 01 5 Standard Method 2320B DOC 2.8 (mmol/L) 3.9E 04 6.5E 05 2.3E 05 1.5E 03 5 Standard Method 5310B NH 4 +1 (mmol/L) 9.7E 04 9.7E 04 9.7E 04 9.7E 04 5 Salicylate Method NO 3 1 (mmol/L) 9.1E 03 7.5E 03 8.3E 04 2.2E 02 5 Cadmium Reduction Method PO 4 3 (mmol/L) 2.9E 04 2.5E 04 1.7E 04 5.1E 04 5 Ascorbic Acid Method SO 4 2 (mmol/L) 1.0E 02 6.9E 03 1.1E 03 2.4E 02 5 SulfaVer 4 Method pH ( ) 5.45 5.17 4.32 7.02 5 Standard Method 4500 H + B

PAGE 105

105 Table 4 6. Stability constants for metal ligand complexes at 25 C (Benjamin 2002, Stumm and Morgan 1996) Reaction Stability Constant log K i Al +3 + OH 1 +2 9.01 AlOH +2 + OH 1 2 +1 8.89 Al(OH) 2 +1 + OH 1 3 (aq) 8.10 Al(OH) 3 (aq) + OH 1 4 +1 7.00 Cd +2 + Cl 1 +1 1.98 CdCl +1 + Cl 1 2(aq) 2.6 Cd +2 + CO 3 2 3(aq) 5.4 Cu +2 + CO 3 2 3(aq) 6.73 Cu +2 + OH 1 +1 6.00 Fe +3 + OH 1 +2 11.81 FeOH +2 + OH 1 2 +1 10.52 Fe(OH) 2 +1 + OH 1 3 (aq) 6.07 Fe(OH) 3 (aq) + OH 1 4 1 6.00 Pb +2 + Cl 1 +1 1.6 Pb +2 + CO 3 2 3(aq) 7.24 Pb +2 + OH 1 1 6.29 Mn +2 + Cl 1 +1 0.6 Mn +2 + CO 3 2 3(aq) 10.4 Mn +2 + SO 4 2 4(aq) 2.3 Ni +2 + CO 3 2 3(aq) 6.87 Ni +2 + HCO 3 1 3 +1 12.47 Zn +2 + CO 3 2 3(aq) 5.3 Zn +2 + OH 1 +1 5.04 Zn(OH) +1 + OH 1 2(aq) 6.06

PAGE 106

106 Figure 4 1. Location of sampling sites within the Lake Tahoe watershed.

PAGE 107

107 Al Cd Cr Cu Fe Mn Ni Pb Zn Metals [ g/L] 1e-3 1e-2 1e-1 1e+0 1e+1 1e+2 1e+3 1e+4 1e+5 1e+6 1e+7 Al Cd Cr Cu Fe Mn Ni Pb Zn Snow samples Control Snow Cd Cu Fe Ni Pb Zn 1e-3 1e-2 1e-1 1e+0 1e+1 1e+2 1e+3 1e+4 1e+5 1e+6 1e+7 CMC CCC NPDES PM-bound Dissolved Regulations Figure 4 2. Non parametric plot of dissolved and particulate bound metals. Total control snow values are shown in the dissolved metals plot as an open circle. Applicable regulations for total dissolved metals are shown in the right most plot. California Toxic Rule values, CMC (Criterion maximum concentration) and CCC (Criterion continuous concentration), are shown for Cd, Cu, Fe, Pb, and Zn. The NPDES permitted effluent limit for discharge to a collection system or surface water for Ni is also shown. Cd Pb Cr Cu Ni Mn Zn Al Fe fd 0.001 0.01 0.1 1 Figure 4 3. Dissolved fractions (fd) for all sites.

PAGE 108

108 Roadside Sample Pb Cd Fe Mn Al Ni Cr Zn Cu Kd (L/kg) 1e+4 1e+5 1e+6 1e+7 1e+8 1e+9 After Imhoff Cone Fe Al Zn Ni Mn Cr Pb Cu Cd 1e+1 1e+2 1e+3 1e+4 1e+5 1e+6 Figure 4 4. Kd (equilibrium partition coefficient) values for all sites.

PAGE 109

109 Temperature 5 10 25 Percent Difference from 1C (%) -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 Cd+2 Cd +2 CdCl +1 Cd DOM CdCO3 (aq) CdCl2 (aq) Concentration (mmol/L) 1e-11 1e-10 1e-9 1e-8 1e-7 1e-6 1e-5 1e-4 Cd +2 CdCl +1 Cd DOM CdCO3 (aq) CdCl2 (aq) Cd +2 CdCl +1 Cd DOM CdCO3 (aq) CdCl2 (aq) 1C 5C 10C Cadmium Cd +2 CdCl +1 Cd DOM CdCO3 (aq) CdCl2 (aq) 25C 1C CrOH +1 Cr +2 Concentration (mmol/L) 1e-8 1e-7 1e-6 1e-5 1e-4 1e-3 1e-2 5C CrOH +1 Cr +2 10C CrOH +1 Cr +2 Chromium 25C CrOH +1 Cr +2 Temperature 5 10 25 Percent Difference from 1C (%) 0.0 0.4 0.8 1.2 1.6 2.0 Cr+2 Temperature 5 10 25 Percent Difference from 1C (%) -100 -80 -60 -40 Al(OH)4 -1 Al2(OH)2CO3 +2 Al(OH)3 (aq) Al(OH)2 +1 Concentration (mmol/L) 1e-7 1e-6 1e-5 1e-4 1e-3 1e-2 1e-1 Al(OH)4 -1 Al2(OH)2CO3 +2 Al(OH)3 (aq) Al(OH)2 +1 Al(OH)4 -1 Al2(OH)2CO3 +2 Al(OH)3 (aq) Al(OH)2 +1 1C 5C 10C Aluminum Al(OH)4 -1 Al2(OH)2CO3 +2 Al(OH)3 (aq) Al(OH)2 +1 Snowmelt Control 25C Al(OH)2 +1 Figure 4 5. Speciation results for Aluminum (top), Cadmium (center), and Chromium (bottom) for 1, 5, 10, and 25C. Snowmelt samples are shown as box nwhisker plots while control samples are shown as a scatter plot.

PAGE 110

110 Temperature 5 10 25 Percent Difference from 1C (%) -28 -24 -20 -16 -12 -8 -4 0 PbDOM PbCO3 (aq) PbOH +1 Pb +2 PbCl +1 Concentration (mmol/L) 1e-10 1e-9 1e-8 1e-7 1e-6 1e-5 1e-4 1e-3 PbDOM PbCO3 (aq) PbOH +1 Pb +2 PbCl +1 PbDOM PbCO3 (aq) PbOH +1 Pb +2 PbCl +1 1C 5C 10C Lead PbDOM PbCO3 (aq) PbOH +1 Pb +2 PbCl +1 25C Pb+2 Temperature 5 10 25 Percent Difference from 1C (%) 0 20 40 60 80 100 120 Fe(OH)2 +1 Fe(OH)4 -1 Fe(OH)3 (aq) FeOH +2 Concentration (mmol/L) 1e-7 1e-6 1e-5 1e-4 1e-3 1e-2 1e-1 Fe(OH)2 +1 Fe(OH)4 -1 Fe(OH)3 (aq) FeOH +2 Fe(OH)2 +1 Fe(OH)4 -1 Fe(OH)3 (aq) FeOH +2 1C 5C 10C Iron Fe(OH)2 +1 Fe(OH)4 -1 Fe(OH)3 (aq) FeOH +2 25C Fe(OH)3 (aq) Temperature 5 10 25 Percent Difference from 1C (%) -18 -16 -14 -12 -10 -8 -6 -4 -2 0 CuDOM CuCO3 (aq) Cu +2 CuOH +1 Concentration (mmol/L) 1e-9 1e-8 1e-7 1e-6 1e-5 1e-4 1e-3 1e-2 1e-1 1e+0 CuDOM CuCO3 (aq) Cu +2 CuOH +1 CuDOM CuCO3 (aq) Cu +2 CuOH +1 1C 5C 10C Copper CuDOM CuCO3 (aq) Cu +2 CuOH +1 Snowmelt Control 25C Cu+2 Figure 4 6. Speciation results for Copper (top), Iron (center), and Lead (bottom) for 1, 5, 10 and 25C. Snowmelt samples are shown as box nwhisker plots while control samples are shown as a scatter plot.

PAGE 111

111 Temperature 5 10 25 Percent Difference from 1C (%) -5.0 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 Mn +2 MnCO3 (aq) MnCl +1 MnSO4 (aq) Concentration (mmol/L) 1e-10 1e-9 1e-8 1e-7 1e-6 1e-5 1e-4 1e-3 1e-2 1e-1 Mn +2 MnCO3 (aq) MnCl +1 MnSO4 (aq) Mn +2 MnCO3 (aq) MnCl +1 MnSO4 (aq) 1C 5C 10C Manganese Mn +2 MnCO3 (aq) MnCl +1 MnSO4 (aq) Snowmelt Control 25C Mn+2 Temperature 5 10 25 Percent Difference from 1C (%) -1.6 -1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 Ni+2 Ni +2 NiDOM NiCO3 (aq) NiHCO3 +1 Concentration (mmol/L) 1e-10 1e-9 1e-8 1e-7 1e-6 1e-5 1e-4 1e-3 1e-2 Ni +2 NiDOM NiCO3 (aq) NiHCO3 +1 Ni +2 NiDOM NiCO3 (aq) NiHCO3 +1 1C 5C 10C Nickel Ni +2 NiDOM NiCO3 (aq) NiHCO3 +1 25C Temperature 5 10 25 Percent Difference from 1C (%) -10 -8 -6 -4 -2 0 2 Zn+2 Zn +2 ZnDOM Zn(OH)2 (aq) ZnOH +1 ZnCO3 (aq) Concentration (mmol/L) 1e-8 1e-7 1e-6 1e-5 1e-4 1e-3 1e-2 Zn +2 ZnDOM Zn(OH)2 (aq) ZnOH +1 ZnCO3 (aq) Zn +2 ZnDOM Zn(OH)2 (aq) ZnOH +1 ZnCO3 (aq) 1C 5C 10C Zinc Zn +2 ZnDOM Zn(OH)2 (aq) ZnOH +1 ZnCO3 (aq) 25C Figure 4 7. Speciatio n results for Manganese (top), Nickel (center), and Zinc (bottom) for 1, 5, 10 and 25C. Snowmelt samples are shown as box nwhisker plots while control samples are shown as a scatter plot.

PAGE 112

112 CHAPTER 5 SPECIATION OF NUTRIENTS AND ANIONS IN URBAN SNOWMELT EXPOSED TO TRAFFIC ACTIVITIES IN THE LAKE TAHOE WATERSHED I ntroduction Urban snowbanks are a reservoir for particulate matter (PM), metals, nutrients, organic chemicals, and deicing salts species and residuals (Oberts 2003, Zinger and Delisle 1988). Sum mer rainfall runoff typically has significantly lower values of conductivity, sodium, potassium, calcium, inorganic carbon, chloride, sulfate, and alkalinity when compared to snowmelt runoff (Bckstrom et al. 2003). PM, commonly reported as total suspende d solids (TSS), concentrations vary by location and age of snowpack, but are consistently greater than rainfall runoff PM concentrations. Snowmelt concentrations of PM are found to be more variable and generally higher. Urban snowmelt in Sweden had PM le vels ranging from 11 to 7,889 mg/L, and is comparable to snow in Montreal ranging from 86 to 8,546 mg/L (Reinosdotter and Viklander 2006, Zinger and Delisle 1988). In contrast, the event mean concentration (EMC) of rainfall runoff TSS samples at a Cincinnati site is reported as 130.7 57.2 mg/L (Sansalone et al. 1998), and 423.7 273.3 at a Baton Rouge site (Sansalone and Kim 2008). The use of de icing salts can results in structural and ecological impacts. Sodium chloride salts enhance metal oxidatio n and pavement deterioration, as well as negatively impact roadside vegetation, soils, and receiving waters (Amrhein et al. 1992). Chlorides in Montreal snowmelt range from 56 to 10,000 mg/L (Zinger and Delisle 1988) ; a similar range was found in four ur ban roadway sites in Cincinnati (Sansalone and Glenn 2002). Application rates seen on California roads in the Lake Tahoe region range from 0.87 to 24.28 tons/lane per km in the 19741975 winter season (Goldman and Hoffman 1975). Sansalone and Glenn (2002) report lower application rates ranging from 0.088 to 0.143 tons/lane per km. Bckstrom et al. (2003) report

PAGE 113

113 mean snowmelt chloride values from two Swedish cities as 1,490 and 3,490 mg/L. While increasing lake salinity may not directly generate acute tox icity of freshwater aquatic species, there is evidence that such salinity generates chronic susceptibility to environmental stresses (Goldman and Hoffman 1975). Chemicals in road salt mixtures can be either stimulants or inhibitors for algal and bacteria l metabolisms; such dichotomous effects are dependent on the water chemistry of the receiving water body and organisms (Goldman and Hoffman 1975). Increased salinity inputs into a lake can cause serious ecological problems including the formation of a monimolimnon, or favorable conditions to the growth of blue green algae; a nuisance and an indicator of health hazards (Goldman and Hoffman 1975). Monimolimnon formation, deep denser layers of water often caused by high salinity waters, can lead to circulati on failure, lack of turnovers (meromixis) between layers. As a result, deep layers can become anoxic, acts as a sink for nutrients, and accumulate H2S (Goldman and Hoffman 1975). If a meromixis occurs and the layer is subsequently mixed with the lake dur ing spring or fall turnovers, acute lethality to fish and other aquatic organisms can occur (Goldman and Hoffman 1975). Campbell (1992) reports on the effects of acidic spring snowmelt on metals and nutrients on Quebec North Shore salmon rivers. Results demonstrate the inverse relationship between pH and alkalinity with increasing rate of flow, while DOC concentrations and color intensity mimicked the passage of the hydrograph (Campbell 1992). Zinger and Delisle (1988) report that Montreal snowmelt values for sulfates and ammonium nitrogen range from 25 to 295 mg/L and 0.1 to 0.8 mg/L, respectively. Bckstrom et al. report two Swedish urban snowmelt levels for sulfate, alkalinity and TOC of 8.26 and 45.2 mg/L, 0.55 and 0.71 meq/L, and 6.36 and 8.23 mg/ L, respectively. Jassby et al. (1995) report that measured (rainfall and snowmelt) runoff into

PAGE 114

114 Lake Tahoe during 19891991with total nitrogen (TN) and total phosphorus (TP) mean flux levels of 8.8 3.6 and 0.80 0.29 mmol/m2 per year. These measurements are based on a watershed area of 812 km2 and a lake surface area of 501 km2. Atmospheric deposition rates during the study indicate TN and TP rates of 23 3 and 0.73 0.00 mmol/m2 per year (Jassby et al. 1995). Nitrogen and phosphorus inputs into the lake are of importance as the eutrophication of the lake has become a national concern. A 19731993 study indicates that nitrate values in the lake are the highest from January to April (Jassby et al. 1995), but it is unclear whether this trend is due dec reased productivity or seasonal inputs. Several effluent limits are listed in the NPDES permit for the Lake Tahoe region. Effluent limits are found for both land treatment systems and collection systems/surface waters for total nitrogen, total phosphorus turbidity, and suspended solids (permit CAG616001, USEPA 2003). These values are shown in Table 51. B ackground Chemical speciation is influenced by water chemistry. The primary water chemistry parameters influencing speciation are pH, redox potential, ionic strength, and available ligands in solution. Typical ligands utilized in speciation models are dissolved organic matter (DOM) such as fulvic acid, humic acid, dissolved organic carbon (DOC); carbonate species (HCO3 1, CO3 2); h ydroxides; and anions (Cl1, SO4 2, NO3 1, PO4 3). The MINTEQ model is a widely used speciation model capable of determining the equilibrium concentrations of dissolved and adsorbed species and multiple solid phases and a gas phase ( US EPA 2005). Typical model inputs include pH, redox potential, and ionic strength and, to complete the charge balance, all major dissolved ionic measurements (Dean et al. 2005). Reaction equations, log equilibrium constants, mass action expressions and mole balances are utili zed to iteratively determine a mathematical solution to the equilibrium concentrations of all species (Allison et al. 1991).

PAGE 115

115 Equilibrium constants in the thermodynamic database are referenced to 25C and are adjusted for temperature using the vant Hoff e quation (equation 1), this assumes that enthalpy of reaction is independent of temperature which can result in significant errors at temperatures that far exceed 25C (Allison et al. 1991). 000588 0 25 log log0 25T H K Kr T (1) Of the 3 available sub models for DOC in MINTEQ, the Gaussian model is utilized. Necessary parameters to utilize the NICA Donnan and SHM submodels are unavailable: proton dissociating groups (mol/g), Donnan volume parameter, width of distribution of sites nonideality parameter for H+, Stern Layer capacitance, spherical radius, site density, etc. The Gaussian model is the predominately used submodel ((Hesterburg et al. 2006, Dean et al. 2005, Birceanu et al. 2008, Gardner et al. 2007, Buerge Weirich and S ulzberger 2004). Ideal gas law, shown in equation 52, is used to convert gaseous concentrations from atm ospheres to mmol/L. In this equation P is the pressure in atmospheres, V is the volume in liters, n is the number is moles, T is the temperature in K and R is the gas constant of 0.08206 L atm/mol K. PV=nRT ( 52) The three species in this study most affected by pH are the ammonium, carbonate and phosphate species. The chemical reactions and equilibrium constants (25 C) for these species are shown i n equations 53 through 5 8 (Stumm and Morgan 1996) NH4 +1 +1 + NH3 (aq) pKa1 = 9.25 ( 53) H2CO3 (aq) +1 + HCO3 1 pKa1 = 6.35 ( 54) HCO3 1 +1 + CO3 2 pKa2 =10.33 ( 55) H3PO4 (aq) +1 + H2PO4 1 pKa1 = 2.16 ( 56)

PAGE 116

116 H2PO4 1 +1 +HPO4 2 pKa2 = 7.20 ( 57) HPO4 2 +1 + PO4 3 pKa3 = 12.35 ( 58) Alpha notation is used later in the manuscript where 0 is defined as the fraction of the i is the fraction of the species that has lost i values must equal 1. The term (aq) indicates that the species is an uncharged aqueous species. The alpha notation equations are shown in Equations 59 through 5 12 for the phosphate system. The equations for the other two species are similarly derived (Benjamin 2002). 0 = [H3PO4]/TOT PO4 ( 59) 1 = [H2PO4 1]/TOT PO4 ( 510) 2 = [HPO4 2]/TOT PO4 ( 511) 3 = [PO4 3]/TOT PO4 ( 512) O bjectives The objectives of this study focus on modeling the speciation of anions associated with urban snowmelt to determine if snowmelt will need to be treated for removal of anions. The first objective i s to collect water quality data such as pH, alkalinity, SSC TDS, nutrients, and metal concentrations for snow melt samples. The second objective i s to utilize MINTEQ to determine the speciation of metals in the sample. Snowmelt in the Lake Tahoe area will ultimately enter the lake, either through direct runoff or following a separation (commonly sedimentation) system. The data will be analyzed to determine if they exceed permits or discharge limits for Lake Tahoe.

PAGE 117

117 M ethodology and Laboratory Analysis Sampling Locations The Lake Tahoe watershed, along highway US 50, covering both California and Nevada, is the region of focus. Data are collected from 6 source area sites and over 5 nonconsecutive winter seasons. Two sites are found in California, four in Nevada. Four locations are roadside samples, one is a par king lot sample, and one is the local snow storage area for removed urban snow. All locations are sampled at least twice. To simplify the data set, all samples are referred to with a Site ID number. A summary of site ID, locations and sampling dates can be found in Table 5 2. Traffic data for the region is listed as average daily traffic, averaged over the course of a year. Traffic data was obtained for a 21.4 km stretch of highway in California in 2004, from Echo Lake Road to the state line, and an 8 km stretch of highway in Nevada in 2000, from the state line to Zephyr Cove. Annual average daily traffic counts range from 12,200 to 36,000 vehicles per day for California and 14,800 to 32,000 for Nevada (TMA 2004, NDOT 2006). Seasonal traffic data cou ld not be identified. Snow Sampling and Preparation Samples are taken from roadside snow banks or parking areas with a total of 17 snowmelt samples. The snowbanks are generated by plowing snow from roadways and parking pavement. A thin width of the cross sectional areas of the roadside or parking lot snow bank is removed and packed into 4 L polypropylene bottles until full. Samples are taken in duplicate while still frozen and kept below 0C during transport to the laboratory and kept in a frozen st ate until analysis. Before analysis, frozen samples are thawed at room temperature overnight. Water quality parameters SSC and COD are analyzed on the total sample. Thawed samples are then agitated

PAGE 118

118 and decanted into 1 L Imhoff Cones for settleable solids removal. Supernatant in the Imhoff cones after 60 minutes of quiescent settling is removed and stored in 1 L bottles for analysis. Laboratory Analysis Water quality p arameters Several water quality parameters are measured within 24 hours of thawing the snow samples including: pH, suspended solids concentration (SSC ), total suspended solids (TSS), total dissolved solids (TDS), conductivity, alkalinity, and chemical oxygen demand (COD) A conductivity and pH meter are used to measure conductivity, TDS, sa linity, and pH in triplicate and both meters are calibrated with a 3 point calibration curve prior to analyis. A sodium hydroxide titration method is used to determine alkalinity (Standard Method 2320B) in triplicate (APHA 1998). A filtration method is used on samples before Imhoff Cone Analysis for SSC; samples are filtered through a nominal 1 m glass fiber filter and dried to determine SSC (ASTM 1999). After Imhoff Cone analysis, TSS is measured using Standard Method 2540D (APHA 1998). Nephelometri c turbidity measurements are also made in triplicate on a Hach 2100AN turbidimeter. A summary of water quality, water chemistry, solids analysis, and measurement methods are shown in Tables 53 through 56. Unsettled (T0, total sample) measurements inclu de CODT, redoxT, conductivityT, salinityT, and SSC. Settled (T60) measurements include pH, turbidity, CODS, redoxS, conductivityS, salinityS, and TSS. The settled sample is filtered through a 0.45 m membrane filter for the measurement of dissolved components. The membrane filters are a biologically inert mixture of cellulose acetate and cellulose nitrate and contain no fibers or particles t hat can contaminate filtrate. Dissolved measurements include CODd, alkalinity, and metal and ion analysis. The control samples are analyzed as total sample without any filtration.

PAGE 119

119 PSD a nalysis Particle size distributions are determined using ASTM D422, a mechanical dry sieve analysis, samples are prepared using ASTM D421 (ASTM 1990, ASTM 1993). After the supernatant is removed from the unsettled sample, the remaining wet slurry is dried at 40C and desiccated. The solids are then disaggregated with a mortar and pestle. The PM samples are sieved through a series of 17 sieves (9500 m to 25 m) and a pan to collect PM smaller than 25 m and allowed to shake in an oscillating shaker for 7 minutes with 2 mm amplitude. The mass remaining on each sieve is co llected and weighed; the sum of all retained mass is compared to the initial weight to determine if the mass balance criterion of 2% is maintained. Metals analysis The dissolved fraction of metallic elements is determined by filtering the supernatant sampl es through a 0.45 m membrane filter and then acidified with nitric acid. The control samples are analyzed as a total sample without any filtration and reported as dissolved concentrations. The acidified samples are then analyzed on an ICP MS or ICP AES for dissolved metals. The ICP is calibrated with a multielement standard solution with 7 calibration points. All analyses are performed in triplicate. I on a nalysis Supernatant samples a re filtered through a 0.45 m membrane filter to obtain the dissol ved fraction The control samples are analyzed as total sample without any filtration and reported as dissolved. The dissolved fraction is then analyzed on a DR/5000 spectrophotometer for phosphate, nitrate, ammonia, sulfate, chloride, tannins and lignins, and CODd. A summary of the methods used is shown in Tables 53 through 56. Each parameter is calibrated with a 5 point calibration curve. A separate dissolved fraction sample i s acidified with sulfuric acid to a pH of

PAGE 120

120 below 2 and analyzed for DOC on a Shimadzu TOC 5050A. All analyses are performed in triplicate. Speciation analysis Visual MINTEQ version 2.53 i s used to analyze all 17 snowmelt and 5 control samples for speciation. The program is run at 4 different temperatures, 3 that are common t o the Lake Tahoe region during snowmelt (1, 5, and 10C), and a higher temperature for comparison (25C). Program inputs include pH, 14 metals, DOC, NH4 +1 and 5 anions: Cl1, CO3 2, NO3 1, PO4 3, and SO4 2 as seen in T ables 53 and 55. The charge on each metallic element is assumed to be the most stable isotope. The speciation results of ammonium and the 5 anions are reported here for all 17 snowmelt and 5 control samples. The system was modeled as an open system with the pH remaining constant at the measured pH for each of the samples. While the effect of altitude can be accounted for in the model by correcting the atmospheric CO2 pressure, the results reported here are modeled at sea level. The effect of the elevation on the carbona te system was studied, as this system is the most affected by the atmospheric CO2 pressure. R esults A summary of water quality parameters including phosphate, nitrate, sulfate, chloride, DOC, alkalinity and pH is shown in Figure 52. Note that the values shown are in mg/L of that parameter with the exception of alkalinity (mg/L as CaCO3) and DOC (mg/L as C). Snowmelt phosphate, nitrate, sulfate, and chloride values range from 0.086 to 0.942 mg/L, 0.139 to 1.162 mg/L, 0.679 to 209 mg/L, 0.013 to 0.411 mg/ L, and 19.3 to 4732 mg/L, respectively. Dissolved organic carbon ranges from 4.5 to 121 mg/L as C with mean and median snowmelt values of 52.8 and 45.7 mg/L as C, respectively. A summary table of all snowmelt MINTEQ inputs is shown in Table 52. Other m easure snowmelt water quality values are listed in Table 5 3, including turbidity, solids parameters, redox and conductivity values, dissolved COD, tannic

PAGE 121

121 acid, and silica. A summary table of control snow MINTEQ inputs and other water quality parameters i s shown in Tables 55 and 5 6, respectively. Tannic acid and silica are only measured for the 2008 sampling event; the average control snow value for that sampling event is shown as the median value. A box and whisker summary of the 17 snowmelt samples a nd 5 control snow samples is shown in Figure 53 for SSC, TSS, TDS, and Turbidity. Snowmelt suspended solids values range from 4.6 to 51037 mg/L while the control snow samples ranges from 1 to 14.5 mg/L. Turbidity values range from 470 to 99050 NTU for snowmelt samples and 0.45 to 45.84 NTU for control snow samples. A summary of chemical oxygen demand (COD) is shown in Figure 54 with total suspended and dissolved COD values shown for the snowmelt samples, and total COD shown for the control snow sampl es. Dissolved COD values range from 21.8 to 610 mg/L for snowmelt samples and 5.26 to 130 mg/L for control snow total COD. Alkalinity and hardness components are shown in Figure 55 and demonstrate that calcium hardness is the most prevalent form of hardness in snowmelt and control snow samples. A summary of PSDs for 11 of the sites is shown in Figure 56. The majority of the mass is associated with sediment and settleable PM. Speciation results are shown in Figures 5 7 and 58. They are shown at four temperature values 1, 5, 10, and 25C, increasing from left to right. Temperatures common to the Lake Tahoe region during the peak snowmelt period are: 1, 5, and 10C, the 25C value is only for comparison purposes. Species are shown in order of decre asing median value determined at 1C. Snowmelt values are shown in box and whisker format while the control snow values are shown as a scatter plot. Figure 5 7 shows anions that contain the nutrients nitrogen and phosphorus: ammonia, nitrate, and phospha te Figure 58 contains other anions important to water quality: carbonates, chlorides, and sulfates. Figure 59 illustrates the compounds that evolve from

PAGE 122

122 solution: CO2 (g) and NH3 (g). A frequency distribution of the measured snowmelt pH values is sh own in Figure 510 along with the expected predominance in the snowmelt samples of 3 species given measured pH values. If the pH was allowed to change in the MINTEQ model, the final pH will increase. This increase is shown in Figure 511 for both the snowmelt and control snow samples. The first row of Figure 5 7 shows that the dominant species of ammonia are NH4 +1, NH3 (aq), and NH4SO4 1, which encompasses 41.39 to 100% of total ammonia nitrogen. Median values of NH4 +1 and NH4SO4 1 show a decreasing t rend with increasing temperature with decreases of 3.5 and 9.4%, respectively. The NH4SO4 1 control median shows no change with temperature, while the NH4 +1 control median show very little change with temperature with 0.4% decrease. Ammonia (NH3 (aq)) sh ows a dramatic increase in median values with a temperature increase from 1C to 25C, with an increase of 479% for snowmelt and 232% for control snow. The middle row of F igure 57 shows the three dominant species (100%) of nitrate: NO3 1, CaNO3 +1 and NaN O3 (aq). Both NO3 1 and NaNO3 (aq) demonstrate very little change in the snowmelt median value over the temperature range with differences of +0.02 and 1.10%, respectively. Snowmelt median values for CaNO3 +1 show a steady decrease (19.1%) from 1 to 25C The control values for NO3 1 indicate little change over the temperature range (+0.1%) and CaNO3 +1 show s a steady decrease of 18.3% in value from 1 to 25C. Control values for the NaNO3 (aq) species did not exist at any temperature. The bottom row o f F igure 57 shows the five dominant species of phosphate (56.94 to 99.97% of total phosphate): HPO4 2, H2PO4 1, CaHPO4 (aq), CaPO4 1 and NaHPO4 1. The control species are varied in response to temperature change: HPO4 2 increases by 8.7%, H2PO4 1 decreases by 0.53%, NaHPO4 1 is only seen at 10 and 25C and increases 36.4% in that temperature range, CaHPO4 (aq) initially increases by

PAGE 123

123 22.4% from 1 to 10C, then decreases from 10 to 25C for an overall increase of 13.0%. The remaining species, CaPO4 1, i s not observed at any temperature for the control samples. Decrease in snowmelt median values are seen for HPO4 2, H2PO4 1, and CaPO41 with decreases of 24.8, 9.7 and 28.1%, respectively. Both CaPO4 1 and NaHPO4 show increasing snowmelt median valu es with overall increases of 56.8 and 73.2%. The top row of F igure 5 8 shows the 5 dominant species of carbonate (97.10 to 100% of total carbonate): HCO3 1, H2CO3* (aq), CaCO3 (aq), CaHCO3 +1, and CO3 2. The snowmelt median values for HCO3 1 and H2CO3* both show a decreasing trend with increasing temperature with decreases of 5.9 and 38.2%, respectively. The control values show HCO3 1 values increasing by 18.3% and the H2CO3* values decreasing by 4.8%. A steady increase for the temperature range i s see n for both the snowmelt and control medians for CO3 2 with increases of 105 and 110%, respectively. CaCO3 (aq) oscillate s though the overall trend from 1 to 25C is an increase of 35.1%. Snowmelt CaHCO3 medians show a steady increase of 81.4%, and in the control snow is only seen at 10 and 25C with an increase of 90% in that temperature range. The middle row of Figure 58 shows the dominant species of chloride (99.98 to 100% of total chloride) which are Cl1, CaCl+1, NaCl (aq), and MgCl+1. No significa nt change i s seen for snowmelt or control Cl1. The median snowmelt value for CaCl+1 decreases from 1 to 5C, then increases from 5 to 25C with an overall increase of 2.1%, the control show s a steady increase with temperature with an overall increase of 14.7%. A steady decrease of 19.0% in snowmelt median val ue with increasing temperature i s seen for NaCl (aq) while NaCl (aq) is only seen in the control at 25C. A steady decrease of 9.9% i s seen for the snowmelt median value for MgCl+1. The bottom row of Figure 58 contains the sulfate speciation results with SO4 2, CaSO4 (aq), and NaSO4 1 containing over 98% of total sulfate. SO4 2 show s a minor decrease with an

PAGE 124

124 overall decrease of 0.30%, CaSO4(aq) show s a steady increase of 25.24%, NaSO4 1 show s constant values from 1 to 10C follow ed by at slight increase of 0.18% from 10 to 25C. Both CaSO4 (aq) and NaSO4 1 snowmelt values show an increasing trend of 6.9 and 1.7%, respectively. Snowmelt SO4 2 median values show a small but steady decrease of 0.4% from 1 to 25C. Figure 59 summarizes the percent change in values with respect to 1C for the dominant species. All snowmelt species demonstrate median decreases with increasing temperature, with the exception of NO3 1 which shows a slight increase of 0.02%. HPO4 2 shows the largest decrease with 18.3%, while Cl1 indicates the smallest decrease with 0.05%. Chloride shows no change in control median value with temperature changes. Both SO4 2 and NH4 +1 exhibit slightly decreasing control median values with increasing temperature with values of 0.30 and 0.11%. Control medians for HCO3 1, HPO4 2, and NO3 1 increase with increasing temperature with increases of 18.3, 8.7, and 0.01%, respectively. The evolution of CO2 (g) and NH3 (g) in Figure 510 from the solution illustrates that the change in solubility with temperature. An increase in evolution with increasing temperature is seen for both species in both the snowmelt and control snow samples. A more prominent increase of 18 07% in the snowmelt median value is seen in the gaseous ammonia, while the gaseous carbon dioxide shows a median increase of 25% for the snowmelt samples. The pH frequency distribution shown in Figure 511 represents a mean pH value of 8.2570.623, while the control snow has a mean pH value of 5.1380.543. Also in this figure the alpha values for 0 values indicate the fraction of total ammonia that is present as ammonium ion, the most protonated species. 0 values for the phosphate system were outside of the pH range of the figure and therefore, not shown in the

PAGE 125

125 figure. The results show that at the given pH values the dominant species for these systems will be NH4 +1, HCO3 1, and HPO4 2; this agrees w ith the speciation results from MINTEQ. Figure 5 12 illustrates the difference between the measured pH values and the final pH value if MINTEQ was set to calculate the pH of the sample based on the mass balance. The modeled pH values for the snowmelt samples ranges from 9.5 to 11, values outside of the typical pH of a natural water body. When the pH was allowed to change, the modeled pH of the sample decreased with increasing temperature. Tables 5 7 and 5 8 illustrate the difference in median snowmelt concentrations and average control snow concentrations on the carbonate system when MINTEQ was allowed to account for the elevation of the Lake Tahoe Basin, atmospheric pressure was adjusted to 80% of sea level atmospheric pressure. Control snow samples i llustrate increases in the average concentration of all carbonate species with increased elevation at all temperatures (lower atmospheric pressure), with the exception of HCO3 1 at 25 C. The effect of increased elevation on the median snowmelt sample conc entration varies by species. Both HCO3 1 and CaCO3(aq) show increases in concentration at elevation with an increasing percent difference with increasing temperatures. The remaining species (H2CO3*(aq), CaHCO3 +1, CO3 2) indicate decreases in median concentration with increased elevation, and increasingly decreasing values with increasing temperatures. The effect of temperature on the equilibrium constants on the ammonia, carbonate, and phosphate systems is shown in Table59. C onclusions Source area snow affected by traffic is both a sink and a source for solids and pollutants. Six urban sites along with a control site in the South Lake Tahoe area are analyzed for solids and anions over a 6 year period. Solids parameters such as SSC, TDS, and turbidity are measured

PAGE 126

126 along with water quality parameters such as pH, phosphates, nitrates, sulfates, chlorides and DOC. Turbidity values are typically between 1,000 to 10,000 NTU, well above the NPDES permit for effluent discharges to a land treatment system or su rface waters (200 and 20 NTU respectively); SSC values also exceed the effluent discharge limit to surface waters of 50mg/L (USEPA 2003). Dissolved COD values are at least one order of magnitude lower than the total COD values indicating that the suspended solids contain a very high oxygen demand. The snowmelt samples typically show higher hardness values when compared to alkalinity, and the majority of the hardness is from magnesium. Total nitrogen and total phosphorus are both listed in the NPDES per mit for the Lake Tahoe Hydrologic Unit (Table 51) with maximum values of 0.5 and 0.1 mg/L for discharge to surface waters (USEPA 2003). The total nitrogen averages are below the discharge limit, but the range of values did exceed the limit (0 to 0.67 mg/ L as N). Total phosphorus values range from 0.03 to 0.31 mg/L as P, with the mean value of 0.098 mg/L just below the effluent limit. Both snowmelt and control snow ammonia (NH3 (aq)) demonstrate the largest temperature dependence with increases exceeding 200% with an increase in temperature from 1 to 25C. All species of sulfate, chloride, and nitrate species show very little sensitivity to temperature changes with less than a 25% change in median value for both snowmelt and control snow samples. The sp eciation results show that the dominant form of each parameter for snowmelt samples will be ammonium ion, bicarbonate ion, ionic chloride, nitrate, hydrogen phosphate, and sulfate. Discrepancies between the dominant species did occur between the snowmelt and control samples for the carbonate and phosphate species and would are explained with the pH differences between the two sample types. Due to the low pH of the control snow samples,

PAGE 127

127 H2CO3* and H2PO4 1 species are dominant. The carbonate ion (CO3-2) in creases in concentration with increasing temperature due to the effect of temperature on the equilibrium coefficients seen in Table 9. The pKa value for the HCO3 1/CO3 2 reaction decreases with increasing temperature, resulting in a higher CO3 2 concentration. Species found in this snowmelt study that are studied for aquatic ecotoxicity include: ammonia, ammonium, calcium sulfate, sodium sulfate, ammonium sulfate, and sodium nitrate (USEPA 2007), though these species are found at concentrations much lower than found in the ecotoxicity studies. The effect of both pH and temperature on the evolution of CO2 and NH3 shows that the evolution increases with increasing temperature and the effect of pH is a larger factor in the evolution of NH3 when compared to CO2. This is seen in the difference in magnitude between the snowmelt and control snow samples. Two options were tested on how to input pH into MINTEQ. The condition of setting the pH at a constant measured value better represents what occurs in nature. While the evolution of CO2 will raise the pH of the sample in the natural environment, the increases predicted by MINTEQ are unrealistic with values ranging from 9.5 to 11.0 for the snowmelt samples and 5.9 to 9.9 for the control snow samples. M aintaining the pH at the measured values for the sample gave reasonable results for the speciation of anions with the expected dominant pH dependant species agreeing with the model results. D iscussion This study illustrates that the there is cause for treatment for removal of the suspended PM that is not typically removed with most drainage systems or unit operations. Median nutrient levels in this study do not exceed current effluent limits but are close enough to warrant concern fo r at risk lakes such as Lake Tahoe, since the maximum measured value exceed the current limits. There is also a significant turbidity impact that is more bio available in receiving water bodies like Lake Tahoe.

PAGE 128

128 A 1975 study reports no evidence of a moni molimnon in Lake Tahoe, and the volume dilutes influent chloride levels to manageable levels, the lake maintains a fairly constant concentration of 1mg/L, with occasional values of 2 mg/L measures below 100m deep (Goldman and Hoffman 1975). The same study reports that Lake Tahoe is typically nitrogen and iron limited (Goldman and Hoffman 1975). A later study shows that the lake had become phosphorus limited indicating a large influx of nitrogen into the lake (Goldman et al 1993). Inputs into the lake alt ered the water chemistry enough to switch the lake from nitrogen limited to phosphorus limited in less than 20 years. These studies reinforce the need for treatment for all nutrient inputs into Lake Tahoe. Jassby et al. (1995) reports atmospheric deposit ion rates of 23 and 0.73 mmol m2 yr1 (1989 to 1991) in the Lake Tahoe watershed for total nitrogen and total phosphorus, respectively. The high total nitrogen atmospheric deposition rates should explain why total nitrate values for snowmelt and control s now are extremely similar and may explain why the lake switched from nitrogen and phosphorus limited. There is evidence to show that increased treatment of snowmelt is necessary for the Lake Tahoe area to reduce inputs of nutrients, toxic species, turbidi ty, suspended solids, and chlorides. A centralized treatment system designed to manage both nutrients and fine solids should be considered and would allow more maintenance than often ignored site by site unit operations. A more centralized system would allow for secondary treatment such as absorptive filtration to aid in the removal of nutrients. No menclature Alk alkalinity (aq) uncharged aqueous species COD chemical oxygen demand

PAGE 129

129 DOC dissolved organic carbon DOM dissolved organic matter EMC event mean concentration fd dissolved fraction ICP inductively coupled plasma ICP MS inductively coupled plasma mass spectrometer ICP AES inductively coupled plasma atomic emission spectroscopy K25 equilibrium constant at 25 C Keq equilibrium constant KT equilibrium constant at temperature T MINTEQ Mineral thermodynamic equilibrium model n number of moles in the ideal gas law NICA nonideal competitive adsorption NPDES National Pollutant Discharge Elimination System NTU Nephelometric turbidity unit P pressu re in atmospheres in the Ideal Gas Law PM Particulate Matter PSD Particle Size Distribution QA/QC quality assurance/quality control R Gas Constant in the Ideal Gas Law, R = 0.08206 L atm/mol K SHM Stockholm Humic Model SSC suspended solids concentration T temperature

PAGE 130

130 T0 initial melted snow sample, without any settling T60 melted snow sample after 1 hour of quiescent TDS total dissolved solids TN total nitrogen TOC total organic carbon TP total phosphorus TSS total suspended solids V Volume in liters in the Ideal Gas Law 0 rH standard state enthalpy of reaction Table 5 1. NPDES effluent discharge limits. (USEPA 2003). NPDES permit numbers CAG616001, CAG6161002, CAG 6160032. Parameter Land Treatment System Collection Systems/Surface Waters Total Nitrogen 5 mg/L (as N) 0.5 mg/L (as N) Total Phosphorus 1 mg/L (as P) 0.1 mg/L (as P) Turbidity 200 NTU 20 NTU Suspended Solids -50 mg/L

PAGE 131

131 Table 5 2. Summary of sites, locations, and sampling dates. Site # Sample Location Sample Date Notes 1 103A Sierra Blvd@Barbara Ave, S. Lake Tahoe, CA 1/5/2003 Duplicate samples taken 1 103B Sierra Blvd@Barbara Ave, S. Lake Tahoe, CA 1/5/2003 Duplicate samples taken 1 1203 Sierra Blvd@Barbara Ave, S. Lake Tahoe, CA 12/20/2003 Snow dump site 1 1208 Sierra Blvd@Barbara Ave, S. Lake Tahoe, CA 12/28/2008 Snow dump site 2 1203 HWY US50W@Tunnel, NV 12/20/03 Road surface 2 1204 HWY US50W@Tunnel, NV 12/27/2004 Road surface 2 1205 HWY US50W@Tunnel, NV 12/21/2005 Road surface 2 1208 HWY US50W@Tunnel, NV 12/28/2008 Road surface 3 1203 HWY US50W@Firestation #5, NV 12/20/2003 Parking Lot 3 1208 HWY US50W@Firestation #5, NV 12/28/2008 Parking Lot 4 1204 HWY US50W@Zephyr Cove, NV 12/28/2004 Road surface 4 1205 HWY US50W@Zephyr Cove, NV 12/21/2005 Road surface 4 1208 HWY US50W@Zephyr Cove, NV 12/28/2008 Road surface 5 1204 HWY US50W@Stateline, CA 12/29/2004 Parking Lot 5 1208 HWY US50W@Stateline, CA 12/28/2008 Parking Lot 6 1205 HWY US50W@Lake Pkwy, S. Lake Tahoe, NV 12/21/2005 Road surface 6 1208 HWY US50W@Lake Pkwy, S. Lake Tahoe, NV 12/28/2008 Road surface

PAGE 132

132 Table 5 3. Summary of MINTEQ input data for snowmelt samples. The median charge balance for the snowmelt samples is 4.98%. (pH = log{H+}) Parameter Mean Median Minimum Maximum Count Method Ag +2 (mmol/L) 1.13E 05 1.25E 05 1.00E 06 2.51E 05 14 Standard Method 3030B Al +3 (mmol/L) 2.41E 02 1.42E 02 2.59E 04 8.35E 02 17 Standard Method 3030B Ba +2 (mmol/L) 9.20E 04 4.50E 04 8.88E 05 5.59E 03 17 Standard Method 3030B Ca +2 (mmol/L) 9.20E 01 3.57E 01 5.03E 02 8.24E+00 17 Standard Method 3030B Cd +2 (mmol/L) 1.26E 05 1.51E 05 1.34E 08 3.24E 05 17 Standard Method 3030B Cr +2 (mmol/L) 7.93E 04 2.14E 04 2.51E 05 5.11E 03 17 Standard Method 3030B Cu +2 (mmol/L) 5.36E 04 4.21E 04 5.17E 05 2.49E 03 17 Standard Method 3030B Fe +3 (mmol/L) 1.19E 02 9.76E 03 2.14E 03 3.93E 02 17 Standard Method 3030B Mg +2 (mmol/L) 1.25E 01 5.75E 02 1.06E 02 9.87E 01 17 Standard Method 3030B Mn +2 (mmol/L) 3.16E 03 1.21E 03 7.50E 05 1.63E 02 17 Standard Method 3030B Na +1 (mmol/L) 1.76E+01 1.37E+01 1.18E 01 4.27E+01 8 Standard Method 3030B Ni +2 (mmol/L) 2.57E 04 1.24E 04 8.81E 06 8.58E 04 17 Standard Method 3030B Pb +2 (mmol/L) 6.27E 05 1.38E 05 4.84E 09 1.99E 04 17 Standard Method 3030B Zn +2 (mmol/L) 1.09E 03 8.84E 04 1.19E 04 3.67E 03 17 Standard Method 3030B Cl 1 (mmol/L) 2.45E+01 1.23E+01 5.45E 01 1.33E+02 17 Mercuric Thiocyanate Method CO3 2 (mmol/L) 8.63E 01 7.33E 01 1.95E 01 2.08E+00 17 Standard Method 2320B DOC 2.8 (mmol/L) 3.16E+00 2.13E+00 5.22E 01 1.01E+01 16 Standard Method 5310B NH4 +1 (mmol/L) 1.69E 02 7.06E 03 1.31E 03 6.06E 02 9 Salicylate Method NO3 1 (mmol/L) 8.39E 03 6.86E 03 7.65E 04 1.87E 02 17 Cadmium Reduction Method PO4 3 (mmol/L) 3.16E 03 1.99E 03 9.08E 04 9.92E 03 17 Ascorbic Acid Method SO4 2 (mmol/L) 2.73E 01 1.14E 01 7.07E 03 2.18E+00 16 SulfaVer 4 Method pH ( ) 8.08 7.91 6.91 9.58 17 Standard Method 4500H + B

PAGE 133

133 Table 5 4. Water quality, water chemistry, and solids summary for the snowmelt samples. Parameter Mean Median Minimum Maximum Count Method Turbidity (NTU) 10743 1999 470 99050 17 Standard Method 2130B Alk (mg/L as CaCO3) 306 122 19.5 1308 17 Standard Method 2320B pH ( ) 8.08 7.91 6.91 9.58 17 Standard Method 4500 H + B SSC (mg/L) 14723 22861 1435.0 56820 9 ASTM D3977 97 TSS (mg/L) 2567 9938 317.00 59124 17 Standard Method 2540 D Settleable Solids (g/L) 182 74.4 2.91 1040 17 Standard Method 2540F Settleable Solids (mL/L) 76 42.0 2.50 531 17 Standard Method 2540F Sediment (g) 2043 826 95 8761 16 ASTM D3977 97 Redox T (+mV) 23.45 77.9 100.8 449.7 15 Combo Redox/ORP probe Redox S (+mV) 1.10 115.3 97.6 388 17 Combo Redox/ORP probe Conductivity T (mS/cm) 2151 7182 160 67700 15 4 Electrode Conductivity Cell Conductivity S (mS/cm) 1757 2386 175 14760 17 4 Electrode Conductivity Cell Salinity T (ppt) 1.0 0.7 0.1 3.7 12 4 Electrode Conductivity Cell Salinity S (ppt) 1.1 0.8 0.1 3.8 12 4 Electrode Conductivity Cell COD T (mg/L) 6431 6523 711 10355 9 Reactor Digestion Method COD S (mg/L) 2362 1728 259 9542 17 Reactor Digestion Method COD d (mg/L) 161.2 130.4 21.8 610.0 17 Reactor Digestion Method Tannic Acid (mg/L) 1.04 0.963 0.263 1.86 6 Tyrosine Method Silica (mg/L) 3.26 2.67 1.28 7.00 6 Heteropoly Blue Method

PAGE 134

134 Table 5 5. Summary of MINTEQ inputs for control snow samples. The median charge balance for the control snow samples is 48.46%. (pH = log{H+}) Parameter Mean Median Minimum Maximum Count Method Ag +2 (mmol/L) 1.3E 05 1.3E 05 8.8E 07 2.7E 05 5 Standard Method 3030B Al +3 (mmol/L) 1.3E 05 1.3E 05 8.8E 07 2.7E 05 5 Standard Method 3030B Ba +2 (mmol/L) 3.8E 03 1.5E 03 1.5E 04 1.5E 02 5 Standard Method 3030B Ca +2 (mmol/L) 1.4E 04 7.3E 05 1.1E 05 4.4E 04 5 Standard Method 3030B Cd +2 (mmol/L) 8.0E 02 1.2E 02 6.7E 03 3.6E 01 5 Standard Method 3030B Cr +2 (mmol/L) 6.7E 06 2.9E 06 3.9E 07 2.2E 05 5 Standard Method 3030B Cu +2 (mmol/L) 3.9E 04 6.5E 05 2.3E 05 1.5E 03 5 Standard Method 3030B Fe +3 (mmol/L) 3.8E 03 1.4E 03 1.0E 03 1.4E 02 5 Standard Method 3030B Mg +2 (mmol/L) 6.9E 03 1.3E 03 1.2E 03 2.9E 02 5 Standard Method 3030B Mn +2 (mmol/L) 1.5E 04 5.2E 05 3.5E 05 5.6E 04 5 Standard Method 3030B Na +1 (mmol/L) 3.3E 02 3.3E 02 1.3E 02 5.4E 02 5 Standard Method 3030B Ni +2 (mmol/L) 1.4E 04 1.4E 04 4.0E 05 2.4E 04 5 Standard Method 3030B Pb +2 (mmol/L) 5.9E 05 1.9E 05 1.6E 06 2.1E 04 5 Standard Method 3030B Zn +2 (mmol/L) 7.7E 04 5.5E 04 2.0E 04 1.8E 03 5 Standard Method 3030B Cl 1 (mmol/L) 6.7E 06 2.9E 06 3.9E 07 2.2E 05 5 Mercuric Thiocyanate Method CO3 2 (mmol/L) 7.4E 02 2.5E 02 2.5E 03 2.8E 01 5 Standard Method 2320B DOC 2.8 (mmol/L) 3.9E 04 6.5E 05 2.3E 05 1.5E 03 5 Standard Method 5310B NH4 +1 (mmol/L) 9.7E 04 9.7E 04 9.7E 04 9.7E 04 5 Salicylate Method NO3 1 (mmol/L) 9.1E 03 7.5E 03 8.3E 04 2.2E 02 5 Cadmium Reduction Method PO4 3 (mmol/L) 2.9E 04 2.5E 04 1.7E 04 5.1E 04 5 Ascorbic Acid Method SO4 2 (mmol/L) 1.0E 02 6.9E 03 1.1E 03 2.4E 02 5 SulfaVer 4 Method pH ( ) 5.45 5.17 4.32 7.02 5 Standard Method 4500H + B

PAGE 135

135 Table 5 6. Water quality, water chemistry, and solids summary for control snow samples. Parameter Mean Median Minimum Maximum Count Method Turbidity (NTU) 11.90 4.50 0.45 45.84 5 Standard Method 2130B Alk (mg/L as CaCO3) 1.71 2.00 1.00 2.00 5 Standard Method 2320B pH ( ) 5.45 5.17 4.32 7.02 5 Standard Method 4500H+ B SSC (mg/L) 6.33 5.00 1.00 14.50 5 Standard Method 2540 D Redox (+mV) 397.5 300.0 52.4 608.7 5 Combo Redox/ORP probe Conductivity (mS/cm) 9.57 12.4 2.71 73.2 5 4Electrode Conductivity Cell Salinity (ppt) 0.0 0.0 0.0 0.0 3 4Electrode Conductivity Cell CODd (mg/L) 54.1 50.0 5.26 130.0 5 Heteropoly Blue Method Tannic Acid (mg/L) -0.015 --1 Tyrosine Method Silica (mg/L) -0.030 --1 Reactor Digestion Method

PAGE 136

136 Table 57. The percent difference in value between the MINTEQ model run at sea level and S. Lake Tahoe elevation for the snowmelt samples. An increase in percentage indicates an increase in median value of snowmelt at elevation. Temper ature (C) H 2 CO 3 (aq) HCO 3 1 CaCO 3 (aq) CaHCO 3 +1 CO 3 2 1 23.7% 34.1% 69.1% 19.1% 2.8% 5 27.7% 31.2% 67.7% 30.6% 8.2% 10 33.5% 27.0% 65.7% 87.9% 15.4% 25 56.2% 10.5% 58.0% 124.6% 41.6% Table 5 8. The percent difference between the MINTEQ model run at sea level and S. Lake Tahoe elevation for the control snow samples for the carbonate system species. A positive percentage value indicates a higher average concentration at S. Lake Tahoe elevation as compared to sea level. Temperature (C) H 2 CO 3 (aq) HCO 3 1 CaCO 3 (aq) CaHCO 3 +1 CO 3 2 1 74.3% 49.9% 77.3% 79.4% 77.4% 5 72.1% 42.4% 75.7% 77.6% 75.8% 10 69.2% 32.2% 73.5% 75.1% 73.6% 25 58.8% 3.5% 66.0% 66.6% 66.1% Table 5 9. The equilbrium constants (pKa) values of the ammonia, carbonate, and phosphate species as a function of temperature. (Benjamin 2002, Goldberg et al. 2002, Plummer and Busenberg 1982, Stumm and Morgan 1996) Reaction 0C 5C 10C 25C NH 4 +1 +1 + NH 3 (aq) 10.0 9.9 9.73 9.26 H 2 CO 3 (aq) +1 + HCO 3 1 6.58 6.52 6.46 6.35 HCO 3 1 +1 + CO 3 2 10.63 10.55 10.49 10.33 H 3 PO 4 (aq) +1 + H 2 PO 4 1 2.056 2.073 2.088 2.16 H 2 PO 4 1 +1 +HPO 4 2 7.31 7.28 7.25 7.2 HPO 4 2 +1 + PO 4 3 12.69 12.63 12.56 12.35

PAGE 137

137 Figure 5 1. The locations of the six sampling sites. Phosphate Nitrate Sulfate Chloride DOC Alkalinity pH Concentration [mg/L] 0.01 0.1 1 10 100 1000 10000 pH 4 5 6 7 8 9 10 Snowmelt Control Snow Figure 5 2. Parametric plot of dissolved parameters: anion concentrations in mg/L of the respective anion, DOC in units of mg/L as C, alkalinity in units of mg/L as CaCO3 and pH on the right axis. Snowmelt samples are shown in box nwhisker format while control snow samples are shown as a scatter plot.

PAGE 138

138 SSC TSS TDS Concentration (mg/L) 1e-1 1e+0 1e+1 1e+2 1e+3 1e+4 1e+5 1e+6 Turbidity Turbidity (NTU) 1e-1 1e+0 1e+1 1e+2 1e+3 1e+4 1e+5 1e+6 Snowmelt Control Figure 5 3. SSC (total sample), TSS (supernatant sample), TDS, and turbidity for both snowmelt and control snow samples. Total Suspended Dissolved COD (mg/L) 1e+0 1e+1 1e+2 1e+3 1e+4 1e+5 Snowmelt Control Figure 5 4. Total, suspended, and dissolved COD the snowmelt samples, the total COD value for the control snow is also shown. The median COD dissolved fraction, fd, equals 0.05.

PAGE 139

139 Alkalinity Hardness Ca Mg Other Alkalinity/Hardness (mg/L as CaCO3) 0.1 1 10 100 1000 Snowmelt Control Snow Hardness Components Figur e 5 5. Dissolved alkalinity and hardness components for snowmelt and control snow samples. Particle Diameter ( m) 9500 4750 2000 850 600 425 300 250 180 150 106 75 63 53 45 38 25 Mass Retained (g per kg total solids) 0 100 200 300 400 Cumulative Percent Retained (%) 0 25 50 75 100 n = 11 Figure 5 6. Particle size distribution for 20032005 snowmelt samples.

PAGE 140

140 1C NO3 -1 CaNO3 +1 NaNO3 (aq) Concentration (mmol/L) 1e-7 1e-6 1e-5 1e-4 1e-3 1e-2 1e-1 5C NO3 -1 CaNO3 +1 NaNO3 (aq) 10C NO3 -1 CaNO3 +1 NaNO3 (aq) Nitrate 25C NO3 -1 CaNO3 +1 NaNO3 (aq) 1C HPO4 -2 H2PO4 -1 CaHPO4 (aq) CaPO4 -1 NaHPO4 -1 Concentration (mmol/L) 1e-8 1e-7 1e-6 1e-5 1e-4 1e-3 1e-2 5C HPO4 -2 H2PO4 -1 CaHPO4 (aq) CaPO4 -1 NaHPO4 -1 10C HPO4 -2 H2PO4 -1 CaHPO4 (aq) CaPO4 -1 NaHPO4 -1 Phosphate 25C HPO4 -2 H2PO4 -1 CaHPO4 (aq) CaPO4 -1 NaHPO4 -1 1C NH4 +1 NH3 (aq) NH4SO4 -1 Concentration (mmol/L) 1e-7 1e-6 1e-5 1e-4 1e-3 1e-2 1e-1 1e+0 5C NH4 +1 NH3 (aq) NH4SO4 -1 10C NH4 +1 NH3 (aq) NH4SO4 -1 Ammonia 25C NH4 +1 NH3 (aq) NH4SO4 -1 Snowmelt Control Figure 5 7. Speciation results for Ammonia (top), Nitrate (center), and Phosphate (bottom), for 1, 5, 10, and 25C. Snowmelt samples are shown as box nwhisker plots, while control samples are shown as a scatter plot.

PAGE 141

141 1C HCO3 -1 H2CO3 (aq) CaCO3 (aq) CaHCO3 +1 CO3 -2 Concentration (mmol/L) 1e-7 1e-6 1e-5 1e-4 1e-3 1e-2 1e-1 1e+0 1e+1 1e+2 5C HCO3 -1 H2CO3 (aq) CaCO3 (aq) CaHCO3 +1 CO3 -2 10C HCO3 -1 H2CO3 (aq) CaCO3 (aq) CaHCO3 +1 CO3 -2 Carbonate 25C HCO3 -1 H2CO3 (aq) CaCO3 (aq) CaHCO3 +1 CO3 -2 Snowmelt Control 1C Cl -1 CaCl +1 NaCl (aq) MgCl +1 Concentration (mmol/L) 1e-7 1e-6 1e-5 1e-4 1e-3 1e-2 1e-1 1e+0 1e+1 1e+2 1e+3 5C Cl -1 CaCl +1 NaCl (aq) MgCl +1 10C Cl -1 CaCl +1 NaCl (aq) MgCl +1 Chloride 25C Cl -1 CaCl +1 NaCl (aq) MgCl +1 1C SO4 -2 CaSO4 (aq) NaSO4 -1 Concentration (mmol/L) 1e-7 1e-6 1e-5 1e-4 1e-3 1e-2 1e-1 1e+0 1e+1 5C SO4 -2 CaSO4 (aq) NaSO4 -1 10C SO4 -2 CaSO4 (aq) NaSO4 -1 Sulfate 25C SO4 -2 CaSO4 (aq) NaSO4 -1 Figure 5 8. Speciation results for Carbonates (top), Chlorides (center), and Sulfate (bottom) for 1, 5, 10, and 25C. Snowmelt samples are shown as box nwhisker plots while control samples are shown as a scatter plot.

PAGE 142

142 Figure 5 9. The percent difference of temperature values with respect to the 1C value for the most dominant species. Both snowmelt and control snow samples are shown.

PAGE 143

143 CO2 Evolution (atm) 1e-6 1e-5 1e-4 1e-3 1e-2 Snowmelt Control Snow CO2 (g) Evolution (mmol/L) 0.0001 0.001 0.01 0.1 1 Temperature (C) 0 5 10 15 20 25 NH3 (g) Evolution (atm) 1e-13 1e-12 1e-11 1e-10 1e-9 1e-8 1e-7 1e-6 1e-5 Temperature (C) 0 5 10 15 20 25 NH3 (g) Evolution (mmol/L) 1e-13 1e-12 1e-11 1e-10 1e-9 1e-8 1e-7 1e-6 1e-5 1e-4 Figure 5 10. Evolution of gaseous CO2 and NH3 from the samples at sea level. Snowmelt samples (n=17) are shown as box and whisker plots while control snow (n=5) samples are shown as a scatter plot. Values are shown in atmospheres and mmol/L units (calculated using ideal gas law, R = 0.08206 L atm/mol K)

PAGE 144

144 Ammonia pH 6 7 8 9 10 Carbonate pH 6 7 8 9 10 Phosphate pH 6 7 8 9 10 11 Alpha Value 0.00 0.25 0.50 0.75 1.00 pH 6 7 8 9 10 Frequency (%) 0 10 20 30 40 = 8.257 = 0.623 n = 17 pH Figure 511. The frequency distribution of measured pH values in snowmelt samples. The mean, standard deviation, and count are shown in the plot. The statistics for the pH of the control snow samples are =5. most affected by pH are also shown here along with the snowmelt pH distribution using equilibrium constants at 25C. The pH frequency distribution is shown in the background of the following plots for comparison purposes. Snowmelt Temperature (C) 0 5 10 15 20 25 pH 4 6 8 10 12 Final pH Average pHi Std. Dev. pHi Control Snow Temperature (C) 0 5 10 15 20 25 pH 4 6 8 10 12 Final pH Average pHi Std. Dev. pHi Figure 512. The mean and one standard deviation of the snowmelt (n=17, 8.2570.623) and control snow (n=5, 5.1380.543) samples compared to the final pH values seen if the pH was allowed to c hange in MINTEQ.

PAGE 145

145 CHAPTER 6 SETTLING INDUCED CHANGES TO THE GRANU LOMETERY OF PARTICUL ATE MATTER IN SNOW IMPAC TED BY TRAFFIC IN TH E LAKE TAHOE WATERSH ED Introduction Snowmelt can have very high and variable values of sediment, settleable, suspended and dissolved PM, as well as very high turbidity values. There are currently very few snowmelt treatment systems in place, and snowmelt is generally treated via stormwater systems if they exist within the surrounding infrastructure. Urban snowmelt is generally treated by either disposing of the snow into a local water body, transporting the snow to a local snow deposit or grass swale to melt, or left to melt in place. The practice of snow dumping has been studied by various researchers and have concluded that the high variability of snow demands further investigation or snow separation strategies to protect the receiving water body (Bckstrm 2003, Reinodotter and Viklander 2006, Zinger and Delisle 1988). Few studies exist on treating snowmelt that has been transported to a local snow deposit, with the majority of them focusing on stormwater inputs. Urban drainage in cold climates is significantly impacted by snowfall; sn owmelt and rain onsnow events typically cause peak event runoff volumes and maximum pollutant loads (Oberts et al. 2000). Snow in urban areas tends to accumulate various pollutants including solids, metals, nutrients, and chlorides. The levels of these pollutants depend on the residence time of the snow and the traffic loadings to which the snow is exposed. Bckstrm (2003) reports that ditches and grass swales appear to be ideal locations for the storage of urban snow as the melt water will be contained within them and have showed 78 to 99% retention of suspended solids in a study in Sweden. This treatment option is beset with the problem that dissolved substances will receive no treatment and ice growth can prevent flow and cause blockages (Bckstrm 2003). Sedimentation based retention/detention basins are effective measures for treatment of

PAGE 146

146 stormwater runoff, provided there is available land (Field and Sullivan 2003). For instance, Aldehhimer and Bennerstedt (2003) report a study using sedimentation ta nks that store stormwater for 36 hours for significantly reduced effluent solids, with a median reduction of suspended solids of 84%. It would stand to reason that if a sedimentation system were utilized for snowmelt flows, which are generally lower than stormwater flows, would show equivalent removal values. A snowmelt study in Cincinnati showed that temporal variations in TDS and SSC are highly variable in snowmelt with values ranging from 1 to 10,000 mg/L for both TDS and SSC (Sansalone and Glenn 2002). Reinosdotter and Viklander have proposed a system to define different treatment methods for snowmelt based on average daily traffic (2006). Snow samples in Sweden that analyzed for suspended solids and are found to be highly dependent on the traffic loa d and area, housing areas ranging from 11mg/L with no traffic to 5667 mg/L in high traffic areas, and a central area with averages ranging from 31 to 7889 mg/L for no and high traffic areas, respectively (Reinosdotter and Viklander 2006). A study in Mont real investigates the quality of used snow that is being discharged into the St. Lawrence River and shows suspended solids levels ranging from 86 to 8546 mg/L while turbidity values range from 5 to 90 NTU (Zinger and Delisle 1988). There is a need for a q uantitative and qualitative analysis of the efficacy of sedimentation as a treatment option for snowmelt, as a function of particle size distributions of PM. O bjectives The focus of this study is the settling induced clarification of snowmelt PM. The fir st objective of this study was to analyze the snowmelt samples before and after one hour of settling to determine the removal efficiency of solids. The second objective was to determine if the particle size distribution (PSD) is affected by increased hold ing time (24, 48, 72, and 96 hours)

PAGE 147

147 before analysis on composite samples. The third objective focuses on the effect of detention or settling time (1, 24, 48, 72, and 168 hours) on the PSD and subsequent removal of solids. B ackground A snow storage depos it currently located in South Lake Tahoe, CA utilizes two sedimentation basins in series for the removal of PM in snowmelt. Sedimentation is a commonly used for removing solids from liquids. Concentration, settling velocity, and time are the primary vari ables that influence the overall sedimentation efficiency for separation of PM. Settling velocity is typically modeled with either Newtons Law (Equation 1) or Stokes Law (Equation 2) depending on the Reynolds number (Equation 3). Stokes law is utilize d in the laminar settling regime with Reynolds numbers less than 5, and is derived from Newtons Law with a constant drag coefficient (Equation 4). (Fair et al. 1968, Reynolds 1996). Newtons Law p p d p w w p d t pd 1 sgC 3 g 4 d C 3 g 4 V ( 61) Stokes Law 18 d 1 sg g 18 d g V2 p p 2 p w pt p ( 62) Coefficient of Drag 0.34 Re 3 Re 24 Cd d d ( 63) Reynolds Number d V d V Rep p w p p d ( 64) Turbidity considerations are important from an aesthetic point of view, and directly influence the applicability of filtration or disinfection as treatment options. Turbidity is

PAGE 148

148 generally difficult to treat with sedimentation and typically requires coagulation and/or filtration. Turbidity can reduce the effectiveness of disinfection as pathogens can become encased in the turbidity and inadvertentl y protected from the disinfectant. Turbidity requirements for the Lake Tahoe region in the state of California f or discharge to a collection system or surface water are 20NTU and 200 NTU for land treatment systems (USEPA 2003). Turbidity levels are incr easing in Lake Tahoe with mean annual Secchi disk depths decreasing at a rate of 0.27 meters per year, with a total decrease of 10 meters from 1967 to 2001 (Jassby et al. 2003). Suspended solids discharges to natural water are regulated for aesthetic reaso ns and to prevent sediment buildup. In addition, total suspended solids (TSS) is generally used as an index to quantify the strength of domestic wastewater (Sawyer et al. 2003) Suspended solids in snowmelt generally have more particulate bound metals in comparison with stormwater; these particulate bound metals can pose a toxicity threat to receiving water bodies (Viklander et al. 2003, Sansalone and Buchberger 1996). The NPDES permitted level for suspended solids effluent discharge to collection systems o r surface water in the state of California is listed as 50 mg/L (USEPA 2003). M ethodology and Laboratory Analysis Sampling Locations The Lake Tahoe watershed contains sections of both California and Nevada and the lake is encircled by interstate highway US 50. Sampling sites range from South Lake Tahoe, CA and Glenbrook, NV and are taken from snowbanks along this highway from the roadside snow banks and parking lot snow piles with one sampling location representing a snow storage site in California. Locat ions and sampling locations can be found in Table 61. Sampling locations are shown in Figure 1, with snow sampling locations shown in Figure 1A and snowmelt sampling locations shown in the inset Figure 1B.

PAGE 149

149 Regional traffic loadings ranged from 12,200 to 36,000 cars per day in California along a 21.4 km stretch on highway US 50 (TMA, 2008). An 8 km section of highway US 50 in Nevada has traffic loadings ranging from 14,800 to 32,000 cars per day (NDOT 2006). This data is averaged over the course of a yea r, with no bias to season. Snow Sampling and Preparation Snow samples were taken from roadside or parking lot snow banks. Cross sectional samples were packed into 4 L polypropylene bottles. All samples were taken in duplicate. Samples were kept frozen du ring transport to the laboratory and stored in the freezer until analysis. Prior to analysis, samples were removed from the freezer and allowed to melt in the container overnight. The samples were agitated with a stir plate and stir bar and analyzed; thi s sample is referred to as T0 or Time zero. Samples were then agitated manually and poured into 1 L Imhoff cones to remove the settleable solids from the supernatant. After one hour of quiescent settling, the supernatant was poured off into 1 L bottles a nd analyzed again. This second set of data is referred to as T60 or Time 60. Sample Analysis Water q uality analysis Water quality parameters were measured in the laboratory in triplicate. Parameters measured included suspended solids concentration (SSC) volatile suspended solids (VSSC), turbidity, and particle size distribution (PSD). These parameters were measured on the initial samples, and then the samples were manually agitated and poured into 1L Imhoff Cones (ASTM 1999). The samples were allowed to s ettle under quiescent conditions for 1 hour, and then the supernatant was decanted and analyzed for PSD, SSC, VSSC, and turbidity. Parameters measured on the initial sample are referred to as Unsettling Influent, while parameters

PAGE 150

150 measured on the superna tant sample are referred to as Settled Effluent. A filtration method was utilized to determine SSC (APHA 1998); the filters were ignited in a furnace at 550C for 30 minutes for the removal of organics to determine VSSC. Nephelometric turbidity was mea sured in triplicate on a Hach 2100AN turbidimeter using standard method 2130B (ASTM 1999). Composite s amples Composite samples were prepared using 6 (collected in 2008) samples for both T0 and T60 samples. These samples were analyzed for PSD, SSC, VSSC, a nd turbidity initially, then at 24 hour increments for 4 days, with five times in total. Samples were stored refrigerated at 4C and were allowed to reach room temperature before sub sampling. Composite samples were agitated with a stir bar and stir plate before analysis, and stored in the refrigerator again until the next sampling time. Downstream samples An additional set of samples was collected for a quantitative and qualitative analysis of the PSDs downstream of the snowmelt flow and how that PSD would change with respect to increased retention time. Snowmelt was sampled as flow directly off the snowbank. The storm sewer sample was taken approximately 692 m south and downstream of the snowbank where the flow entered the storm sewer. The Lake Tah oe Dock sample was taken at the outfall of the storm sewer 762 m west of the stormsewer sample. These locations are shown in Figure 1B. Samples were analyzed for PSD, SSC, VSSC (volatile suspended solids concentration), and turbidity before being decante d into the Imhoff Cones. Subsamples were obtained from the Imhoff cone supernatant after 1 hour, 24 hours, 48 hours, 72 hours, and 96 hours and analyzed for PSD, SSC, VSSC, and turbidity. Samples were left to remain in the Imhoff Cones at room temperatur e between sampling events.

PAGE 151

151 Particle size a nalysis All samples are analyzed on a Malvern Mastersizer 2000 Hydro G particle size analyzer to determine aqueous PSD. A light scattering technique is utilized to determine particle size of solids in the range of 0.02 to 2000m. The scattering of laser light in combination with Mie theory determines both the particle diameter and the percent of total volume occupied by particles of that diameter. Well mixed subsamples are diluted with deionized water until the l aser obscuration is within 1 to 10% for analysis. The diluted sample is agitated at 75% of the stirring capacity, and pumped through the laser light at 75% of the pumping capacity. These speeds were chosen to be high enough to keep PM suspended without entraining air bubbles. All samples are made in triplicate, and all measurements are made in triplicate, for a total of 9 PSDs per sample. The resulting PSD data is converted to mass concentrations using appropriate SSC data and analyzed as mass based PS Ds. Modeling of PSDs All samples were modeled with a cumulative gamma distribution. The equations below show the probability density function (Equation 65), the cumulative density function (Equation 66), the gamma function (Equation 67), the incomplet e gamma function (Equation 68) and the cumulative gamma distribution (Equation 69). 0 1 1dx e x e d d fx d ( 65) dd F 1 1 ( 66) 0 1dx e xx ( 67)

PAGE 152

152 dx e xx d 1 ( 68) xe x x f11 ; ( 69) In equations 65 through 69, is the distribution shape parameter, is the scaling parameter, d is the particle diameter and x is a substitution variable for d in the integrals. Parameters were selected by minimizing the sum of squared errors (SSE) bet ween the measured and modeled PSD. Treatability Study To model the treatability of the measured PSDs, three influent gradations were analyzed in the treatability study: unsettled influent from the water quality samples (n = 9), pavement snowmelt and storm sewer snowmelt samples from the watershed. These three were chosen to quantify the efficacy of the sedimentation basin across differing influent mass loads and influent PSDs. The removal efficiency and effluent concentration of the basin was modeled using surface overflow rate (Malcolm 1989) in a 1,000 m2 basin, and the settling velocity of the particles was calculated using and iterative method with Stokes Law at a temperature of 5C. Using surface overflow rate, all particles with Vt > V0 are consider ed completely removed from the basin, while the other particles are removed at a ratio shown in Equation 10. Climate data was obtained at 5 minute increments and utilized to determine overland flow rates into the basin. The solids removal efficiency and effluent SSC was calculated for a range of flow rates (0.00028 to 0.06456 m3/s) and equivalent surface overflow rates. Particle reentrainment by flow was not considered. Surface Overflow Rate 100 x V V (%) Removal0 t (6 10)

PAGE 153

153 R esults Water Quality Samples The variation of SSC, VSSC, and turbidity before and after one hour of quiescent settling is shown in Figure 62. A significant decrease is visible for both with median removals of 87% and 85% for SSC and VSSC, respectively. A lower yet still significant reduction in turbidity is also visible, with a decrease of 68% in the median value. A summary of the mean, median, minimum, and maximum values for SSC, VSSC, and turbidity, along with percent reduction can be found in Table 62. Aqueous PSDs as cumulati ve percent finer by mass for all nine snowmelt samples are shown in Figure 63. A noticeable change in the PSD with one hour of settling is seen for all samples with the d50 values decreasing by 72%. A summary of the d10, d50, and d90 value for the cumula tive PSDs is shown in Table 63 along with the percent reduction of each with one hour of settling allowed. The statistics on the gamma distribution parameters are also shown for both the unsettled influent and settled effluent. The incremental PSDs show n in Figure 6 4 also illustrate the removal of particles and a peak shift toward the finer particle diameters with one hour of settling. Composite Samples The particle size distributions of composited samples are shown in Figure 6 5 for both the unsettl ed influent and settled effluent, for both incremental and cumulative PSDs. There was no noticeable change in the PSD with increased holding time, for either the unsettled influent or settled effluent samples using either the incremental or cumulative gr aphs. The gamma distribution and d50 parameters were analyzed with increased holding time, and no statistically significant difference was found, indicating that no trend was found with increased snowmelt holding times.

PAGE 154

154 Watershed Samples Particle size dis tributions for downstream samples are shown in Figure 66. The first column illustrates the decrease in particles with 1 hour of settling. The second column illustrates the change in particle size distributions with 24, 48, 72, and 168 hours (1, 2, 3, and 7 days) of settling and shows a decrease in the magnitude of the peak with increased settling times. Please note that this data is not available for the Lake Tahoe Dock sample. The third column shows the change in the cumulative PSD with respect to time modeled with a gamma distribution. There is a noticeable difference between time 0, 1 and 24 hours, with the changes hardly noticeable after 24 hours. Cumulative PSDs of the storm sewer outfall sample were not measured after 24 hours. The gamma distribution and d50 parameters were analyzed as a function of increased settling time in Figure 6 7. Pavement snowmelt and storm sewer snowmelt were analyzed at all times, while the storm sewer outfall sample could only be analyzed at 0 and 1 hour. Gamma values (shape parameter) increase for the first 24 to 48 hours and then decrease with increased settling time. The beta values (scaling parameter) showed a steady decrease for the first 24 hours, and then leveled off. A similar trend was seen for the d50 va lues where there is a marked decrease in the first 24 hours and then continued to decrease at a much lower rate. A summary of the values of SSC, VSSC, turbidity, d10, d50, d90, and gamma distribution parameters for all samples for each settling time is se en in Table 6 4. Treatability Study The treatability study shows that sedimentation basins located in Lake Tahoe are very effective at removing course particles from snowmelt as seen in Figure 9, but are not effective enough for discharge standards. It also shows the extreme variability in basin performance with the three different influent PSDs. The nine unsettled influent samples indicate source area snow, with extreme variability in both the influent and effluent PSDs, with influent d50 values rangin g

PAGE 155

155 from 17.8 to 57.4 m. The pavement snowmelt and storm sewer snowmelt samples show little to no variability in the influent, while varying significantly in the effluent. These two samples represent the downstream PSD and the PSD entering the lake from the storm sewer system. The downstream sample is associated with the coarsest PSD, with a median d50 of 51.6 m, followed by the source area (median d50 of 45.1 m), and the outfall (median d50 of 7.7 m). The median effluent PSDs of the samples are clos er to the lower range of the effluent PSDs because of the prevalence of higher flow rates. The effluent concentrations for the source areas indicate extreme variability with increasing surface overflow rate. The range of influent SSC values is also shown in the plots, indicating the concentrations that could be reached with large surface overflow rates. Equivalent particle diameters for each surface overflow rate indicate the smallest particle that will be 100% by the system. C onclusions This paper focu ses on the PSD of snowmelt and how it is affected by increased settling time and the stormwater conduit system (hydrologic cycle). Snow samples were collected at six source area sites during two separate sampling events and analyzed for water quality parameters and PSD. Imhoff Cone analysis indicated that SSC, VSSC, and turbidity were significantly reduced with one hour of quiescent settling with median values of 87.8, 84.8, and 68.2% respectively. The d50 of the median PSDs also decreased from 17.94 m to 4.26 m (74.6%). Results also indicated that increasing the holding time did not affect the PSD, gamma distribution parameters, or d50 of the sample when agitated on a stir plate, for up to 96 hours when samples were stored at 4C. When settling time was increased, the cumulative PSD of the samples did not vary with time after 24 hours of sampling, while the incremental PSDs showed decreases in peak concentration. Granulometric indices (d10, d50, d90) and gamma distribution

PAGE 156

156 parameters ( ) showed ve ry little change from 24 to 168 hours with all coefficients of variance (COV) below 11%. From a design perspective, a minimum of 24 hours of settling time is required to meet the NPDES permit for suspended solids, with a subsequent treatment method necessa ry to reach the turbidity requirement. Snowmelt flow downstream of the snowpack showed SSC values lower than values from snowmelt that had been allowed to settle, indicating that much of the PM could be removed during flow conditions. Additional studies would be necessary to determine required treatment for such particulate bound contaminants. The treatability study shows that the effluent PM concentration and PSD of the basin is significantly influenced by the source PSD. The high variability in the unsettled effluent samples indicates a design challenge for secondary treatment systems such as filtration. Even at the high concentrations seen in Figure 9, the Type I settling assumption is still valid, as significant Type II settling behavior was not se en for 96 hours as seen in Figure 6. Storm sewer snowmelt samples exhibited a much finer PSD and lower effluent concentrations from the basin when compared to the pavement snowmelt samples. The storm sewer sample demonstrates the need for additional trea tment as sedimentation is not capable of removing solids to the levels acceptable for discharge to Lake Tahoe. Using the outfall sample as an example, to reach a target loading rate of 50 mg/L for filtration, design surface overflow rates must remain below 0.000632 mm/s (based on 1,000 m2). Discussion The proposed treatment recommendations by Reinosdotter and Viklander (2006) for urban snow exposed to over 10,000 cars per day should be transported to a snow deposit and the snowmelt should be considered a possible contaminant to groundwater and local waterbodies.

PAGE 157

157 This study illustrates that sedimentation is a very effective treatment method for snowmelt. Primary treated effluent levels for turbidity and PM are reached even with 24 hours of settling. Howev er, secondary treatment, possibly coagulation/flocculation and filtration would be required to reduce turbidity levels below discharge standards for the region. While a significant change in the PSD was observed after one day of settling, the additional change was minimal with additional settling time. Laboratory analyses suggest analysis of PSD is not time sensitive as long as the samples are thoroughly mixed. In essence, there was no evidence of coagulation within the samples that could not be overcom e with sufficient mixing. This result also indicates the lack of Type II settling behavior in these samples regardless of the high PM concentration. The extreme variability of effluent PSDs and effluent concentrations show that sedimentation basin solids removal is highly dependent on the influent PSD and flow rate. Passive treatment of snowmelt with sedimentation basins is very effective at the remov al of coarse particles, but additional treatment will be necessary to remove fine PM. Secondary or advanced treatment such as C/F and filtration would be required to remove the finer particles (d < 10 m). The treatability study indicates that a sedimentation basin of 1,000 m2 will not treat the stormsewer sample to acceptable discharge limits with surface overflow rates greater than 0.000632 mm/s. The treatability study indicates that even the low flow rates seen in this study are capable of keeping fine particles entrained in flow and prevent sedimentation seen in the quiescent settling study. Nomenclature COV Coefficient of variation (%) d particle diameter ( m)

PAGE 158

158 N Hazens N NPDES National Pollutant Discharge Elimination System NTU Nephelometric turbidity unit PM particulate matter PSD particle size distribution RPD relative percent difference (%) SSC suspended solids concentration (mg/L) SSE sum of squared errors T0 Time zero, parameters associated with the initial sample T60 Time sixty minutes, parameters associated with a sample that has undergone 1 hour of quiescent settling TDS total dissolved solids VSSC volatile suspended solids concentration (mg/L) x substitution for d in gamma distribution integrals alpha, substitution for in gamma di stribution function beta, parameter in the gamma distribution function gamma, parameter in the gamma distribution function gamma function d incomplete gamma function

PAGE 159

159 Table 61. Summary of sites, lo cations, and sampling dates. Sites represented with Site #s were sampled as snow, while the remaining 3 were sampled as snowmelt. Site # Sample location Sample Date Description 11208 Sierra Blvd @ Barbara Ave, S. Lake Tahoe, CA 12/28/ 08 Snow storage site 21205 HWY US50W@Tunnel, NV 12/27/ 04 Paved Shoulder 21208 HWY US50W@Tunnel, NV 12/28/ 08 Paved Shoulder 31208 HWY US50W@Firestation #5, NV 12/28/ 08 Paved Shoulder 41205 HWY US50W@ZephyrCove, NV 12/21/ 05 Parking Lot Pile 41208 HWY US50W@Zephyr Cove, NV 12/28/ 08 Parking Lot Pile 51208 HWY US50W@Stateline, CA 12/28/ 08 Paved Shoulder 61205 HWY US50W @ Lake Parkway, S. Lake Tahoe, CA 12/21/ 05 Paved Shoulder 61208 HWY US50W @ Lake Parkway, S. Lake Tahoe, CA 12/28/ 08 Paved Shoulder Pavement Snowmelt HWY US50W @ Lake Parkway, S. Lake Tahoe, CA 12/28/ 08 Direct snowmelt from pile Storm Sewer Snowmelt HWY US50W 691 m south of Pavement Snowmelt 12/28/ 08 Snowmelt downstream of pile Storm Sewer Outfall at Lake Outfall of storm sewer 762 m west of Storm Sewer 12/28/ 08 Storm sewer outfall into lake

PAGE 160

160 Table 62. Summary of the unsettled (0 minutes), batch settled (60 minutes), and percent reduction for SSC, VSSC, and turbidity. Results represent 12 observations for 9 sites. Category Parameter Mean Median Min Max Unsettled Influent SSC [mg/L] 20779 14075 1604 52167 VSSC [mg/L] 2792 2772 302 6417 Turbidity (NTU) 16542 11106 1225 39954 Settled Effluent SSC [mg/L] 2467 1527 325 7063 VSSC [mg/L] 429 405 79 940 Turbidity (NTU) 4814 2702 470 14060 Reduction (%) SSC 88.47 87.83 79.72 95.15 VSSC 84.73 84.81 73.76 93.63 Turbidity 68.55 68.25 42.64 93.12

PAGE 161

161 Table 63. Mean, m edian, m aximum, and m inimum values for particulate matter (PM) indices ( d10, d50, d90) with a ssociated reduction percentages. (Samples collected as snow.) PM Index Category Mean Median Max Min d10 (m) Unsettled Influent 2.34 2.19 4.56 1.25 Settled Effluent 1.06 1.10 1.28 0.80 Reduction (%) 50.39 51.21 72.50 31.02 d50 (m) Unsettled Influent 16.75 17.94 26.76 6.73 S ettled Effluent 4.22 4.26 5.52 2.47 Reduction (%) 72.14 74.66 83.59 56.97 d90 (m) Unsettled Influent 81.76 88.78 119.07 45.47 Settled Effluent 17.36 12.59 54.24 9.26 Reduction (%) 79.26 78.42 90.15 54.45 Category Index Mean Median Max Min Unsettled Gamma 0.829 0.777 0.699 1.006 Influent Beta 35.606 40.096 18.023 56.362 Settled Gamma 1.413 1.376 1.056 2.021 Effluent Beta 4.369 4.409 2.229 7.159

PAGE 162

162 Table 64. Values of SSC, VSSC, turbidity parametric indexes (d10, d50, d90), and modeling parameters ( ) with increased settling times. (Samples collected as snowmelt). Sample Type Settling Time (hours) SSC (mg/L) VSSC (mg/L) Turbidity (NTU) d10 (m) d50 (m) d90 (m) Pavement Snowmelt 0 9083 1242 11475 1.88 20.95 94.65 0.73 48.02 1 1212 258 3465 0.80 2.93 12.16 1.32 3.16 24 56.5 36.0 122 0.53 1.34 3.53 2.22 0.74 48 45.7 16.9 68.4 0.50 1.29 3.43 2.17 0.73 72 34.5 13.1 35.4 0.46 1.25 3.48 2.01 0.78 168 25.2 7.8 36.0 0.41 1.20 3.47 1.81 0.84 Storm Sewer Snowmelt 0 679 156 1358 0.80 3.48 20.67 0.98 5.94 1 446 124 901 0.83 2.90 7.89 1.64 2.25 24 48.9 26.3 112 0.49 1.17 2.88 2.50 0.56 48 39.9 13.0 92.2 0.51 1.20 2.91 2.61 0.54 72 26.9 6.8 68.2 0.45 1.07 2.69 2.49 0.52 168 18.0 7.9 43.7 0.42 1.01 2.57 2.39 0.51 Storm Sewer Outfall at Lake 0 41.3 24.9 11.0 5.75 37.13 292.5 0.77 91.61 1 10.4 5.0 11.9 2.54 17.17 65.32 0.97 26.54

PAGE 163

163 Figure 61. The snow sampling locations are shown in part A and the inset shown in part B illu strates the snowmelt sampling locations The distance of flow from the upstream US HWY 50 source area snowmelt location to the downstream storm sewer sampling location is 690 m. The distance of flow from the downstream storm sewer to the outfall sampling location at Lake Tahoe is 760 m. A B A

PAGE 164

164 SSC 0 60 SSC or VSSC [mg/L] 10 100 1000 10000 100000 VSSC Batch Settling Time (Min) 0 60 Turbidity 0 60 Turbidity (NTU) 10 100 1000 10000 100000 n = 12 n = 12 n = 12 Figure 62. SSC, VSSC and turbidity for source area snow (sampled as snow). Time 0 represents the unsettled snowmelt sample and settled results are determined after 60 minutes of quiescent batch settling. Particle Diameter ( m) 0.01 0.1 1 10 100 1000 10000 Percent finer by mass (%) 0 20 40 60 80 100 Unsettled Influent Settled Effluent Distribution Category R2Unsettled 0.770 36.300 0.998 Settled 1.318 4.313 0.999 Particle Diameter ( m) 1 10 100 1000 Settling Veloctity (mm/s) 0.01 0.1 1 10 100 1000 Modeled Measured 0.34 Re 3 Re 24 Cd d d Figure 63. Mean and standard deviation of all nine unsettled (t = 0, T0) and settled (t = 60 min., T60) snow sample PSDs (sampled as snow). PSDs are modeled with a cumulative gamma distribution; the parameters for the model are shown. Measured and modeled (Newt ons Law) settling velocity at 5C is shown in the right plot. Settling velocity for discrete PM sizes are compared to model results utilizing Newtons Law.

PAGE 165

165 2-1205 Concentration (mg/L) 0 100 200 300 400 500 600 700 Unsettled Influent Settled Effluent 4-1205 6-1205 1-1208 Concentration (mg/L) 0 100 200 300 400 500 600 4-1208 3-1208 0 10 20 30 40 50 60 2-1208 Particle Diameter ( m) 0.1 1 10 100 1000 Concentration (mg/L) 0 200 400 600 800 1000 1200 1400 1600 1800 5-1208 Particle Diameter ( m) 0.1 1 10 100 1000 6-1208 Particle Diameter ( m) 0.01 0.1 1 10 100 1000 Figure 64. Incremental PSDs and incremental concentrations of PM for all nine source area sites. (Sampled as snow.)

PAGE 166

166 Particle Diameter ( m) 0.01 0.1 1 10 100 1000 Concentration [mg/L] 0 200 400 600 800 1000 1200 Unsettled Influent Settled Effluent Particle Diameter ( m) 0.01 0.1 1 10 100 1000 Percent Finer by Mass (%) 0 20 40 60 80 100 Unsettled Influent Settled Effluent Figure 6 5. Incremental and c umulative PSDs for 2008 source area sample composites (n = 6) unsettled (t = 0 min) and settled (t = 60 min) PM samples. The samples are analyzed at 0, 24, 48, 72, and 96 hours after melting. (Sampled as snow.) Snowmelt Holding Times (hrs) 0 24 48 72 96 0 4 8 12 16 20 24 28 32 0 24 48 72 96 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Unsettled Influent Settled Effluent d50M 0 24 48 72 96 d50 0 2 4 6 8 10 12 14 16 Figure 66. The role of snowmelt sample holding time on the gamma distribution parameters ( ) and the d50m index. There is no stastically significant difference ( = 0.05) between the slopes of the trend lines and zero indicating no significant coagulation or flocculation during snowmelt holding time.

PAGE 167

167 Particle Diameter ( m) 0.01 0.1 1 10 100 1000 0 1 2 3 4 Concentration [mg/L] 0 100 200 300 400 0 hr 1 hr 0 1 2 3 4 24 hr 48 hr 72 hr 168 hr Percent Finer by Mass (%) 0 25 50 75 100 0 1 24 48 72 168 Pavement Snowmelt 0.01 0.1 1 10 100 1000 Concentration [mg/L] 0 10 20 30 40 0.01 0.1 1 10 100 1000 Percent Finer by Mass (%) 0 25 50 75 100 Storm Sewer Snowmelt Figure 67. Inc remental and cumulative PSDs for settling time series for the source area snowmelt and storm sewer (sampled as snowmelt). The first column shows the initial unsettled (t = 0) and settled (t = 60 min.) samples. The second column of results illustrates settling at 24, 48, 72, and 168 hour s for these samples. The third column of results represents the cumulative percent finer by mass for all samples and all results are modeled with cumulative gamma distributions. All hourly refere nces refer to hours of settling allowed before the sample is analyzed. (Samples collected as snowmelt.)

PAGE 168

168 Snowmelt Settling Time (hrs) 0 1 24 48 72 168 0 10 20 30 40 50 60 0 1 24 48 72 168 d50M 0 5 10 15 20 25 30 0 1 24 48 72 168 0.5 1.0 1.5 2.0 2.5 3.0 Pavement Snowmelt Storm Sewer Snowmelt Figure 68. The change in and d50M with increased settling time for the source area pavement snowmelt and s torm sewer snowmelt samples (Samples collected as snowmelt.)

PAGE 169

169 Surface Overflow Rate (mm/s) 0.0001 0.001 0.01 Effluent Concentration (mg/L) 0 100 200 300 400 500 Effluent Concentration (mg/L) 0 500 1000 1500 2000 2500 Equivalent Particle Diameter ( m ) Effluent Concentration (mg/L) 0 3000 6000 9000 12000 15000 Range Median Ci = 20,779 15,698 mg/L Source Areas (n = 9) Downstream Ci = 9083.3 328.4 mg/L Outfall Ci = 679.4 16.4 mg/L % Finer by Mass 0 20 40 60 80 100 0.42 1.31 4.16 % Finer by Mass 0 20 40 60 80 100 Effluent Median Influent Median Particle Diameter ( m) 0.1 1 10 100 1000 10000 % Finer by Mass 0 20 40 60 80 100 Figure 69. Influent and effluent PSDs for snow source areas (sampled as snow), the downstream (gutter snowmelt flow, sampled as snowmelt), and storm sewer outfall to Lake Tahoe (sampled as snowmelt) are shown in the left column. Effluent PSDs calculated by utilizing Newtons law (5C, specific gravity = 2.60) to determine settling velocity. The role of Type I settling utilizing surface overflow rate to de termine effluent concentration is shown in the right column (based on a surface area of 1,000 m2) for influent PSDs and concentrations at each location.

PAGE 170

170 CHAPTER 7 GLOBAL CONCLUSIONS Urban snowmelt impacted by traffic activities is a significant source of solids, metals, and other contaminants. A comprehensive study of urban snow along Highway 50 in the Lake Tahoe Basin has shown high variability in potential PM and metal loads, along with significant soluble components. Treatment schemes utilizing sedime ntation are effective at removing both solids and particulate bound metals with median removal efficiencies exceeding 99% for particles with diameters greater than 25 m. Sedimentation was also shown to be effective at reducing suspended PM and turbidity values, though to a much smaller extent. It was found that the majority of the solids in snowmelt are classifiable as sediment, with 83 to 92% to total PM associated with particles larger than 75 m. The average particle density of snowmelt particles wa s 2.64 and 2.65 g/cm3 for sediment and settleable PM, respectively. The specific surface area (SSA) of PM generally increases with decreasing particle size, with the majority of the total surface area (SA) associated with particles greater than 75 m, due to the dependence of SA on PSD. No significant effect of temperature was found on settling velocity at 1, 5, and 10C. Mass and number based removal efficiencies typically exceed 99% for particle diameters greater than 25 m when modeled in the existing basin. Lower number based removal efficiencies were seen for particles greater than 1 m, with median efficiency rates at 81.5%. Metal mass for Al, As, Cd, Cd, Cu, Fe, Pb, and Zn was found to be predominantly associated with coarser PM, with d50m rangin g from 179 to 542 m with a median of 341 m. The metal mass distributions were modeled with a cumulative gamma distribution. The mean gamma parameter, representative of the scaling function, showed no significant difference across metal elements. The tot al metal mass for Cr, Cu, and Zn indicated no statistical difference

PAGE 171

171 between sites. Iron and aluminum were shown to be the most prevalent metals in snowmelt, while cadmium was the least prevalent. Median removal efficiencies of all metal elements exceed 99% on a mass basis for particles greater than 25 m, indicating that the existing sedimentation basin would be effective at removing settleable and sediment based metals. Both dissolved and particulate bound metal elements show similar trends with decr easing median values showing the following order: Fe, Al, Zn/Mn, Cu, Cr, Ni, Pb, and Cd. Total iron levels exceeded NPDES discharge limits for discharge to a surface water body. Copper and lead exceeded Criterion Continuous Concentration (CCC) levels found in the California Toxic Rule. Partitioning data indicate that all elements are prominently associated with the particulate phase; fd values were typically below 0.25 for all metal elements. Metal speciation was analyzed at 1, 5, 10, and 25C. All ionic metal species showed decreases in concentration with increasing temperature. Ionic cadmium, manganese, nickel, and zinc were found to be the most prevalent species for each element. Snowmelt treatment would need to address both dissolved and particul atebound metal elements before it would be safe to discharge into Lake Tahoe. Control snow total nitrate levels are near the median snowmelt total nitrate levels, indicating that atmospheric deposition of nitrogen is a significant fraction of total nitra te in urban snowmelt. Turbidity and SSC values found in snowmelt exceed discharge limits for effluent to surface water (USEPA 2003). While neither snowmelt total nitrogen phosphorus median concentrations exceeded discharge limits, the upper range of snowm elt values did exceed the limits. The largest effect of temperature on speciation was seen with ammonia, with increases in concentration exceeding 200%. Very little sensitivity to temperature changes was seen with sulfate, chloride, and nitrate speciatio n. The pH difference between the snowmelt and control snow samples explains the discrepancy between the dominant species of carbonate and

PAGE 172

172 phosphate groups. Toxic compounds such as ammonia, ammonium, calcium sulfate, sodium sulfate, ammonium sulfate, and sodium nitrate, were identified as existing species (USEPA 2007). The removal of potentially toxic dissolved compounds as well as fine solids removal would have to be considered in treatment design options to treat this snowmelt to a point where it could be safely discharged into Lake Tahoe. A significant reduction in solids (both suspended PM and turbidity) can be seen with only one hour of quiescent settling. Water quality differences between the raw snowmelt and settled snowmelt indicate that sus pended PM and turbidity indicated median decreases of 87.8% and 68.2%, respectively. As expected, the settled PSD was significantly finer than the raw PSD, with the d50m of the resulting PSDs also show a significant decrease from 17.94 to 4.26 m. Both the unsettled influent and settled effluent were tested to determine if sample holding time affected the PSD. No significant difference in gamma distribution parameters or granulometric indices was found with holding times of 24, 48, 72, and 96 hours. Increased settling time had little effect on PSDs, gamma distribution parameters, or granulometric indices after 24 hours and up to 168 hours. Sedimentation was found to be an effective form of treatment for the removal of suspended PM and turbidity. A minimum of 24 hours of sedimentation was found to be effective at reducing suspended PM below discharge limits, but settling times of168 hours was ineffective at reducing turbidity levels below NPDES discharge limits. The inherent variability in snowmelt samples results in extreme variability when treatment was modeled in the existing sedimentation basin. The treatability study utilizing surface overflow rate illustrate the difficulties in the treatment of snowmelt, as the fine particle sizes are difficu lt to remove with sedimentation even at low flow rates and will require filtration for removal.

PAGE 173

173 The treatment of urban snowmelt is difficult due to the inherent variability of quality, high concentrations of both particulate bound and dissolved contaminants, and limited treatment options due to the near freezing temperatures. Sedimentation methods are effective at removing the majority of the PM and particulate bound contaminants. The remaining snowmelt still contains fine solids and dissolved pollutants as well as nutrients that must be treated before discharging to Lake Tahoe. Nomenclature CCC Criterion Continuous Concentration NPDES National Pollutant Discharge Elimination System PM Particulate Matter SA Total surface area SSA Specific sur face area

PAGE 174

174 LIST OF REFERENCES Al Houri, Z.M., Barber, M.E., Yonge, D.R., Ullman, J.L., and Beutel, M.W. (2009) Impacts of frozen soils on the performance of infiltration treatment facilities. Cold Regions Science and Technology (In Press) DOI:10.1016/j.coldregions.2009.06.002. Allison, J.D., Brown, D.S., and K.J. NovoGradac. (1991) MINTEQA2/PRODEFA2, A Geochemical Assessment Model for Environmental Systems: Version 3.0 Users Manual. EPA/600/391/021, 115 pp. http://epa.gov/cea mpubl/mmedia/minteq/index.html Accessed [19 Aug09] Aldeheimer, G., and Bennerstedt, K. (2003). Facilities for treatment of stormwater runoff from highways. Water Science and Technology 48(9), 113121. American Public Health Association (APHA). (1998). Standard Methods of the examination of water and wastewater 20th Ed., A. D. Eaton, L. S. Clesceri, and A. E. Greenberg, eds., American Public Health Association, American Water Works Association and Water Environment Federation, Washington, D.C. American Society for Testing and Materials (ASTM), Designation D 422 63 (Reapproved 1990), Standard Test Method for Particle Size Analysis of Soils, Annual Book of Standards, Vol. 04.08,1016. American Society for Testing and Materials, Designation D 421 85 (Reapproved 1993), Standard Practice for Dry Preparation of Soil Samples for ParticleSize Analysis and Determination of Soil Constants, Annual Book of Standards, Vol. 04.08, 89. American Society for Testing and Materials (1999). Standard Test Method for Determining Sediment Concentration in Water Samples, Annual Book of Standards, Water and Environmental Technology ASTM D397797, Vol. 11.02, p. 389394. American Society for Testing and Materials, Designations D555094 (1994). Standard test method for Specific Gravity of Soil Solids by Gas Pycnometer. D5550 94, West Conshohocken, PA., 04.08, 1016. Anderson, J., Estabrooks, T., and McDonnell, M. (2000). Duluth Metropolitan Area Streams Snowmelt Runoff Study. Minnesota Pollution Control Agency. Duluth, Minn., 130. Amrhein, C., Strong, J.E., and Mosher, P.A. (1992). Effect of de icing salts on metal and organic matter mobilization in roadside soil. Environmental Science and Technology 26, 703709. Arenpalli, D.N., Shanthakumar, S., Hanumantha Roa, B., and Singh, D.N. (2008). Comparison of Methods for Determination of Soil Constants. D42185. West Conshohocken, PA., 04.08, 89.

PAGE 175

175 Arnold, C., and Gibbons, J. (1996). Impervious surface coverage: the emergence of a key environmental indicator. J ournal of the American Planning Association 62(2):243258. Backstrom, M. (2003). Grassed swales for stormwater pollution control during rain and snowmelt. Water Science and Technology 48(9), 123134. Bckstrom, M., Nilsson, U., Hkannon, K., Allard, B., and S. Karlsson. (2003). Speciation of heavy metals in road runoff and roadside total deposition. Water, Air, and Soil Pollution, 147, 343366. Benedetti, M.F., Milne, C.J., Kinniburgh, D.G., Van Reinsdijk, W.H., and L.K. Koopal. (1995). Metal Ion Binding to Humic Substances: Application of the NonIdeal Competitive Adsorption Model. Environmental Science and Technology 29(2), 446457. Benjamin, M.M. (2002) Water Chemistry. McGraw Hill: New York. Birceanu, O., Chowdhury, M.J ., Gillis, P.L., McGeer, J.C., Wood, C.M., and Wilkie, M.P. (2008). Modes of metal toxicity and impaired branchial ionoregulation in rainbow trout exposed to mixtures of Pb and Cd in soft water. Aquatic Toxicology 89(4), 222231. Brenner, F.J. and W.L. Cooper. (1978). Effect of Suspended Iron Hydroxide on the Hatchability and Embryonic Development of the Coho Salmon. Ohio Journal of Science, 78(1), 3438. Buerge Weirich, D., and B. Sulzberger. (2004). Formation of Cu(I) in Estuarine and Marine Waters: Application of a New Solid Phase Extraction Method to Measure Cu(I). Environ. Sci. Technol., 38, 18431848. Calgon Carbon Corporation. (1995). F400 Specification Sheet. Pittsburgh, PA. Campbell, P.G.C., Hansen, H. J., Dubreuil, B., and Nelson, W. O. (1992) Geochemistry of Quebec North Shore Salmon Rivers during Snowmelt: Organic Acid Pulse and Aluminum Mobilization, Can. J. Fish. Aquat. Sci., 49, 19381952. Carter, D.L., Mortland, M.M., and Kemper, W.D. (1986). Specific Surface. In: Methods of Soil Analysis, Part 1: Physical and Mineralogical Methods. 2nd Ed. Arnold Klute, Eds., American Society of Agronomy, Inc., Soil Science Society of America, Inc., Madison, WI. Chang, C.C. Kuwabara, J.S. and S.P. Pasilis. (1992). Phosphate and Iron Limitation of Phytoplankton Biomass in Lake Tahoe. Canadian Journal of Fisheries and Aquatic Sciences 49(6), 12061215. Christensen, J.B. and T.H. Christensen (1999) Complexation of Cd, Ni, and Zn by DOC in Polluted Groundwater: A Comparison of Approaches Using Resin Exchange, Aquifer

PAGE 176

176 Material Sorption, and Computer Speciation Models (WHAM and MINTEQA2) Environmental Science and Technology. 33(21), 38573863. Clar, M. L., Barfield, B. J., OConnor, T. P. (2004). Stormwater Best Management Pr actice Design Guide. Vol. 1 (4), 41 425. Cristina, C., Tramonte, J. and Sansalone, J. J. (2002) A Granulometry based Selection Methodology for Separation of Traffic generated Particles in Urban Highway Snowmelt Runoff. Water, Air, and Soil Pollution. 136, 3353. Davis, A. P., Shokouhian, M., and Ni, S. (2001) Loading estimates of lead, copper, cadmium, and zinc in urban runoff from specific sources. Chemosphere, 44, 997 1009. Dean, C. M., Sansalone, J. J., Carteledge, F. K., and Pardue, J. H. (2005). Influence of hydrology on rainfall runoff metal element speciation. Journal of Environmental EngineeringASCE, 131, 632642. Environmental Protection Agency ( EPA ) (2006) Exposure Assessment Models: MINTEQA2. 31 Jan 2005. http://www.epa.gov/ceampubl/mmedia/minteq/index.htm Environmental Protection Agency (1996) Method 6020: ICP MS Metals Analysis. Test Methods for Evaluating Solid Wastes Physical/Chemical Methods ( SW 846). Fair, G.M., Geyer, J.C., and Okun, D.A. (1968). Water and wastewater engineering, water purification and wastewater treatment and disposal. Vol. 2, 251 to 2512. Wiley, New York. Feth, J.H., S.M. Rogers, and C.E. Roberson. (1964). Chemical Composition of snow in the northern Sierra Nevada and other areas. USGS Water Supply Paper. pp.1535J Field, R. and D. Sullivan. Eds. (2003) Wet Weather Flow in the Urban Watershed: Technology and Management. RC Press LLC, Boca Raton, FL. Gardner, K.H., Tsiatsios, C.J., Melton, J., and T.P. Seager. (2007) Leaching behavior of estuarine sediments and cement stabilized sediments in upland management environments. Waste Management 27(11), 1648 1654. Glenn, D., and J. Sansalone. (2002). Accretion and pa rtitioning of heavy metals associated with snow exposed to urban traffic and winter storm maintenance activities. II Journal of Environmental Engineering ASCE, 128, 167185. Gnecco I., Sansalone J., Lanza L. (2008) Speciation of Zn and Cu in Stormwater Pavement Runoff from Airside and Landside Aviation Land Uses, J ournal of Water, Air and Soil Pollution 192, 321336.

PAGE 177

177 Goldman C.R., and R.W. Hoffman. (1975). A study of the Influence of highway Deicing Agents on the Aquatic Environment in the Lake Tahoe Basin and Drainage Along Interstate 80. State of California, Department of Transportation, CA DOT TL 7153175 27194134. Goldman, C.R., Jassby, A.D., and S. Hackley. (1993). Decadal, injterannual, and seasonal variability in enrichment bioassays at Lake Tahoe, California Nevada, USA. Canadian Journal of Fisheries and Aquatic Sciences 50, 14891496. Grimm, D.M., Azarraga, L.V., Carreira, L.A., and W. Susetyo. (1991) Continous multiligand distribution model used to predict the stability constant of copper(II) metal complexation with humic material from fluorescence quenching data. Environmental Science and Technology 25(8), 1427 1431. Gustafsson, J.P. (2001). Modeling the AcidBase Properties and M etal Complexation of Humic S ubstances with the Stockholm Humic Model, Journal of Colloid and Interfac e Science, 244, 102112. Hatch, L.K., Reuter, J.E., and C.E. Goldman. (2001). Stream Phosphorus Transport in the Lake Tahoe Basin, 19891996. Environm ental Monitoring and Assesment 69, 6383. Hamilton, R. S., and Harrison, R. M. (1991). Highway pollution, Elsevier, New York. Heijerick, D.G., Janssen, C.R., Karln, C., Wallinder, I.O., and C. Leygraf. (2002). Bioavailability of zinc in runoff water from roofing materials. Chemosphere. 47, 10731080. Heaney, J. P., and Huber, W.C. (1984). Nationwide assessment of urban runoff impact on receiving water quality, Water Resources Bulletin, 20(1), 35 42. Hesterberg, D., de Vos, B., and P.A.C. Raats. (2006) Chemistry of subsurface drain discharge from an agricultural polder soil. Agricultural Water Management 86(12), 22228. Huber, W.C., and R.E. Dickinson (1988) Storm Water Management Model User's Manual, Version 4. EPA/600/ 388/001a (NTIS PB88236641/AS). U.S. Environmental Protection Agency, Athens, GA. Huber, W. (1993). Contaminant transport in surface water, D.R. Maidment, Ed., Handbook of Hydrology, McGraw Hill Inc., New York, N.Y., 14.1 14.50. Huber W., Nelson P., Eldin N., Williamson K., and Lundy J. (2001). Envi ronmental impact of runoff from highway construction and repair materials, J. of the Transportation Research Record, 1743, 19. Jassby, A.D., Reuter, J.E., and C.R. Goldman (2003). Determining long term water quality change in the presence of climate va riability: Lake Tahoe (U.S.A) Canadian Journal of Fisheries and Aquatic Sciences 60, 14521461.

PAGE 178

178 Jassby, A.D., Goldman, C.R., and J.E. Reuter. (1995) Long term change in Lake Tahoe (California Nevada, U.S.A.) and its relation to atmospheric deposition of algal nutrients. Archiv fur Hydrobiologie 135(1), 121. Jassby, A.D., Goldman, C.R., Reuter, J.E., and R.C. Roberts. (1999). Origins and scale dependence of temporal variability in the transparency of Lake Tahoe, California Nevada. Limnology and Oceanography 44(2), 282294. Jones, H.G. (1999). The ecology of snow covered systems: a brief overview of nutrient cycleing and life in the cold. Hydrological Processes 13, 21352147. Kim, J Y, and J.J. Sansalone. (2008). Event based size distributi ons of particulate matter transported during urban rainfall runoff events.Water Research, 42(10 11), 27562768. Kinniburgh, D.G., Milne, C.J., Benedetti, M.F., Pinheiro, J.P., Filius, J. Koopal, L.K., and W.H. Van Reimskijk. (1996). Metal Ion Binding by Humic Acid: Application of the Nica Donnan Model, Environmental Science and Tech nology 30(5), 16871698. Lake Tahoe Data Clearinghouse (2005), Western Geographic Science Center, United States Geological Survey. GIS Data. Available online at http://ta hoe.usgs.gov/GIS.html A ccessed [02/03/2009]. Lancaster, C.D., Beutel, M.W., and D. Yonge. (2009) Evaluation of Roadside Infiltration to Manage Stormwater Runoff in Semiarid Eastern Washington. Environmental Engineering Science 26(5), 935940. Lin, H., Ying, G., and Sansalone, J. (2008) Granulometry of Noncolloidal Particulate Matter Transported by Urban Runoff. Water Air and Soil Pollution. (In Press). Lowell, S., and Shields, J. (1991) Powder and Surface Area and Porosity 3rd Ed. Chapman and Hall, New York. 1 256. Lu, J. Y., Chakrabarti, C. L., Back, M. H., and Sekaly, L. R. (1996). Speciation of Some Metals in River Surface Water, Rain and Snow, and the Interactions of These Metals With Selected Soil Matrices. Journal of Analytical At omic Spectrometry, 11, 11891201. Magill, N. and J.J. Sansalone. (2010). Distribution of Particulate Bound Metals in Source Area Snow in the Lake Tahoe Watershed. Journal of Environmental Engineering, (In Press.) Malcolm, H. R. (1989). Elements of Urban Storm Water Design, North Carolina State University, Raleigh, North Carolina. Morrison, G.M.P., Batley, G.E., and Florence, T.M. (1989). Metal Speciation and toxicity. Chemistry in Britain 25, 791796.

PAGE 179

179 National Climatic Data Center (NCDC). NCDC W eather station Climate Data Inventories 14 Jul 08. http://www.ncdc.noaa.gov/oa/ncdc.html Nelson, P.O., Huber, W.C., Eldin, N.N., Williamson, K.J., Lundy, J.R., Azizan, M.F., Thayumanavan, P., Quigley, M.M., Hesse, E.T., Frey, K.M., and R.B. Leahy. (2003) NCHRP Report 448: Environmental Impact of Construction and Repair Materials on Surface and Ground Waters. Summary of Methodology, Laboratory Results, and Model Development. National Academy Press, Washington, D.C. Nelson, S. S., Yonge, D. R., Ba rber, M. E., Beutel, D.W. and Al Houri, Z. M. (2006). Performance Evaluation of Cold Weather Flow Control and Runoff Treatment BMPs. Cold regions engineering 2006, Current practices in cold regions engineering: Proceeding of the 13th international confer ence. Pg. 1 14. Nevada Department of Transportation (NDOT) (2006). Annual Traffic Report http://www.nevadadot.com/reports_pubs/traffic_report/2000/pdfs/Douglas.pdf Accessed [17 Mar 2008] Nix, S.J., Heaney, J.P., and Huber, W.C. (1988). Suspended soli ds removal in detention basins. J. Environ. Eng., 114(6), 13311343. Novotny, V., D. Muehring, D. Zitomer, D. Smith, and R. Facey. (1998). Cyanide and metal pollution by urban snowmelt: Impact of deicing compounds Water Science and Technology, 38(10), 223 230. Oberts, G. L. (2003) Cold climate BMPs: solving the management puzzle Water Science and Technology, 48, 2132. Oberts, G.L. (2000) Influence of Snowmelt Dynamics on Stormwater Runoff Quality Watershed Protection Techniques. 1(2) pp. 5561. Oberts, G.L., Marsalek, J., and M. Viklander. (2000). Review of Water Quality Impacts of Winter Operation of Urban Drainage, Water Quality Research Journal of Canada, 35(4), 781808. OConnor, D. J. (1988). Models of sorptive toxic substances in fres hwater systems. II: Lakes and Reservoirs. J. Environmental Eng ineeringASCE, 114 (3), 533551. Plummer, L.N. and E. Busenberg. (1982). The solubilities of calcite, aragonite and vaterite in CO2 H20 solutions between 0 and 90C, and an evaluation of the aqueous model for the system CaCO3 CO2 H20. Geochemica et Cosmochimica Acta 46, 10111040. Polo, A.B.S., Lydersen, E., and I.P. Muniz. (1991). "The influence of temperature on aqueous aluminum chemistry and survival of Atlantic salmon (Salmo salar L.) fingerlings." Aquatic Toxicology 21, 267278.

PAGE 180

180 Polo, A.B.S. (1995). "Aluminum polymerization a mechanism of acute toxicity of aqueous aluminum to fish." Aquatic Toxicology 31, 347356. Reino sdotter, K. and Viklander, M. (2005). A comparison of snow quality in two Swedish municipalities Lulea and Sundsvall. Water Air and Soil Pollution. 167, 316. Reinosdotter, K., and M. Viklander. (2006) Handling of urban snow with regard to snow quality Journal of Environmental Engineering ASCE, pp. 271278. Reinosdotter, K, Viklander, M., and P.A. Malmqvist. (2006). Polycyclic aromatic hydrocarbons and metals in snow along a highway. Water, Science, and Technology 54 (67), 195203. Reynolds, T. (1996). Unit Operations and Processes in Environmental Engineering, 2nd Ed., Boston, PWS Publishing Company. Rossman, L. A. (2007), Stormwater Management Model Users Manual, Version 5.0, USEPA/600/ R 05/040, Cincinnati, OH, USA. Sansalone J.J. (1999) Adsorptive Infiltration of M etals in Urban Drainage Media Characteristics, S cience of the Total Environment 235 (1 3) 179 188. Sansalone J.J. and D.W. Glenn (2002). Accretion and partitioning of pollutants associated with urban traffic activities in roadway snow Part I. Journal of Environmental Engineering, 128 (2), 151166. Sansalone, J.J. and J Y Kim. (2008). Transport of Particulate Matter Fractions in Urba n Source Area Pavement Surface Runoff. Journal of Environmental Quality 37, 18831893. Sansalone J.J. and S. G. Buchberger. (1996). Characterization of metals and solids in urban highway snow and rainfall runoff. Transportation Research Record, 1523, 147159. Sansalone, J.J. and C.M. Cristina. (2004). First Flush Concepts for Suspended and Dissolved Solids in Small Impervious Watersheds. Journal of Environmental Engineering, 130(11), 13011314. Sansalone, J.J., and S.G. Buchberger. (1997). Partiti oning and First Flush of Metals in Urban Roadway Storm Water. Journal of Environmental Engineering, A.S.C.E, Vol. 123(2), pp. 134143. (Feb.) Sansalone, J.J., and G. Ying. (2008). Partitioning and granulometric distribution of metal leachate from urban traffic dry deposition particulate matter subject to acidic rainfall and runoff retention. Water Research, 42(15), 41464162. Sansalone, J.J., J.M. Koran, J.A. Smithson, and S.G. Buchburger. (1998). Physical Characteristics of urban roadway solids trans ported during rain events. Journal of Environmental Engineering. 124(4), 427440.

PAGE 181

181 Sansalone, J. J., Glenn, D. W., and Tribouillard, T. (2003b). Physical and Chemical Characteristics of Urban Snow Residuals Generated from Traffic Activities. Water, Air and Soil Pollution, 148, 4560. Sansalone, J. J., Dean, C. M., and Howerter, K. M. (2003a). The Potential for Insite Treatment by Primary Clarification Through Detention of Transportation Land Use Snowmelt Runoff. TRBAnnual Meeting Session A2A03. Wa shington, D.C. Sansalone, J.J. and Z.Teng (2005), Transient Rainfall Runoff Loadings to a Partial Exfiltration System: Implications for Water Quantity and Quality, J. Environ. Eng. ASCE., 131(8), 11551167. Sawyer, C.N., McCarty, P.L., and G.F. Parkin. (2003). Chemistry for Environmental Engineering and Science. 5th edition, McGraw Hill, New York. Scott, W.S., and N.P. Wylie. (1980). The Environmental Effects of Snow Dumping: A Literature Review. Journal of Environmental Management 10, 219240. Servos, M. R., Malley, D. F., Mackie, G. L., and LaZerte, B. D. (1987). Lack of Bioaccumulation of Metals by Elliptio complanata (Bivalva) During Acidic Snowmelt in Three SouthCentral Ontario Streams. Bulletin of Environmental Contamination and Toxicology. 38, 762768. Sharma, M and McBean, E. (2002). Atmospheric PAH deposition: Deposition velocities and washout ratios. Journal of Environmental EngineeringASCE, 128(2), 186 195. Soil Survey Staff (2008) Natural Resources Conservation Service, Unite d States Department of Agriculture. Web Soil Survey. Available online at http://websoilsurvey.nrcs.usda.gov/ A ccessed [03/12/2009]. StummW. and J.J. Morgan. (2002). Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters 3rd ed. John Wiley & S ons: New York. Susetyo, W., Carreira, L.A., Azarraga, L.V., and D.M. Grimm. (1991). Fluorescence techniques for metal humic interactions Fresenius Journal of Analytical Chemistry 339(9), 624635. Teng, Z., Smithson, J., Ping, Z, and Sansalone, J.J. ( 2002). Geospatial Distribution of Heavy Metals in Transportation Land Use Soils. Journal of Transportation Research Board, 1797, 1122. Thomann, R.V. and Mueller, J.A. (1987). Principles of Surface Water Quality Modeling and Control. New York: Harper an d Row, Pub., Inc.

PAGE 182

182 Transportation Management Association. (2004). US 50 Traffic Counts http://www.50corridor.com/traffic_counts.asp. Accessed [17 Mar 2008] USEPA. (2007). ECOTOX User Guide: ECOTOXicology Database System. Version 4.0. http:/www.epa.gov/ecotox/ Accessed [31 Aug09] USEPA (1996) Method 6020: ICP MS Metals Analysis. Test Methods for Evaluating Solid Wastes Physical/Chemical Methods (SW 846). USEPA (1990) Test methods for evaluating solid waste, physical/chemical meth ods. SW 846, 3rd Ed., Final Update, Washington, DC. US EPA. (2000) Water Quality Standards; Establishment of Numeric Criteria for Priority Toxic Pollutants for the State of California; Rule Part III. Environmental Protection Agency, Office of Water, Washington, DC. 40 CFR Part 131. 3168231719. USEPA (2003). WatershedBased National Pollutant Discharge Eliminating System (NPDES) Permitting Implementation Guidance. EPA 833 B 03 004 http://www.water.rutgers.edu/Projects/trading/kick_off_Mtg/Passaic_kickoff_meeting/ref erences/EPA guidance Watershed based permitting.pdf Accessed [3 Aug 2008] Van Raij, B., and Peech, M. (1972). Electrochemical Properties of Some Oxisols and Alfisols of the Tropics. Soil Science Society of America Proceedings 36, 587593. Viklander, M, Marselek, J, Malmquist, P.A., and W.E. Watt. (2003). Urban Drainage and Highway Runoff in Cold Climates: Conference Overview Water, Air and Soil Pollution, 48(9), 110. Wallinder, I. O., Leygraf, C., Karlen, C., Heijerick, D., and Janssen, C.R. (2001). Atmospheric corrosion of zinc based materials: runoff rates, chemical speciation, and ecotoxicity effects, Corrosion Science 43, 809816. The Weather Channel Interactive, Inc. (WCI) (2008). Monthly Averages for South Lake Tahoe CA < http://www.weather.com/outlook/recreation/outdoors/wxc limatology /monthly/USCA1083> Accessed [28 Mar 2008] Western Regional Climate Center (WRCC). (2002). Western US Historical Summaries. http://www.wrcc.dri.edu/Climsum.html Accessed [ 28 March 2008] Westerlund, C., and Viklander, M. (2006). Particle and associated metals in road runoff during snowmelt and rainfall. Science of the Total Environment 362, 143156. Westerlund, C., Viklander, M., and Bckstrm, M. (2003). Seasonal variation in road runoff quality in Lule, Sweden. Water Science and Techonology. 48(9), 93 101.

PAGE 183

183 White, P. Rasmussen, J., and Blaise, C. (1995). Genotoxicity of Snow in the Montreal Metropolitan Area. Water, Air and Soil Pollution 83, 315334. Ying, G and Sansal one, J.J. (2008). Granulometric Relationships for Urban Source Area Runoff as a Function of Hydrologic Event Classification and Sedimentation. Water, Air and Soil Pollution, 193(1), 229246. Zinger, I., and Delisle, C. (1988). Quality of UsedSnow Disc harged in the St. Lawrence River, in the Region of the Montreal Harbor. Water, Air and Soil Pollution, 39, 4757.

PAGE 184

184 BIOGRAPHICAL SKETCH Natalie Magill was born i n 1982 in Baton Rouge, LA. At the age of 8, the Magill family moved to Poplar Bluff, MO. She graduated from Poplar Bluff High School in 2000. She attended Louisiana State University after high school. Her Engineering in Training certification was earned in April 2004. She received her b achelors degree in e nvironmental e ngineering in December 2004. She received her Doctoral of Philosophy in from University of Florida in December 2009. Her doctoral research focused on urban snowmelt contaminants and treatment. She worked under the guidance of Dr. John Sansalone the Department of Envi ronmental Engineering Sciences.