<%BANNER%>

Constitutive Properties of Particulates in Urban Dry Deposition and Source Area Rainfall-Runoff Loadings

Permanent Link: http://ufdc.ufl.edu/UFE0021356/00001

Material Information

Title: Constitutive Properties of Particulates in Urban Dry Deposition and Source Area Rainfall-Runoff Loadings
Physical Description: 1 online resource (202 p.)
Language: english
Creator: Ying, Gaoxiang
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: deposition, distribution, dry, matter, particulate, partitioning, rainfall, runoff
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: Urban highway surfaces are sources of contaminants, such as metals, phosphorus, trash and debris, and oxygen demanding substances, in part, related to atmospheric deposition. Rainfall-runoff from urban highway surfaces often transports significant loads of pollutants in a complex heterogeneous mixture that includes particulate and dissolved solids, metals, phosphorus and organic compounds. Control of runoff particulate matter is challenging due to the variability in flow, variable particulate matter transport and, varying mass and concentration during a wet weather event as compared to conventional wastewater treatment loadings. This study investigated physical indices of particulate matter delivered in source area rainfall-runoff as a function of hydrologic transport and quiescent settling on an event basis. Four mass and flow-limited events, whose volume and particulate load were fully-captured and recovered from settling tanks receiving runoff from a 1088 m2 paved urban watershed in Baton Rouge, LA, were analyzed. Results indicate that mass and flow-limited behavior results in separate particulate matter delivery and relationships which are transformed to a single particulate matter relationship after unit operations such as settling. While flashed from urban surface into storm water runoff and transferred to natural water body during a storm event, equilibrium partition of phosphorus, COD and metals in dissolved and particulate phase in urban rainfall-runoff was examined before and after 60 minutes of quiescent settling for both two classes of events. Phosphorus has a high partition in particulate phase and can be removed with solids by quiescent settling, metals can be partially removed by settling process but needs a further treatment in order to reach national recommended water quality criteria, and COD cannot be effectively removed by quiescent settling. Granulometry, transport and solubility of dry deposition particles was investigated with a purpose of discovering the source and pathway of contaminants in rainfall-runoff from an urban highway land use. In addition, metal species bonded with dry deposition particles were examined and arithmetic mean concentration of metals across the entire size gradation was ranked as Cd < As < Cu < Pb < Zn < Mg < Fe < Ca constantly. Cumulative mass distributions (CMDs) of particulate matter, all modeled with a cumulative gamma distribution function, were translated a finer particle size gradation during the transport from dry deposition into rainfall-runoff, with D50m decreasing from 330.7 ?m to 13.7 ?m along the runoff stream. The settling velocity is a key variable while studying the transport mechanism of particles throughout their entire pathways from dry deposition, carried into urban rainfall-runoff, and finally discharged to different water bodies. In this study, settling velocity of particles were measured as a function of salinity, particle size and suspended sediment concentration (SSC) to illustrate various scenarios of in situ particle delivery. Salinity was found to have no significant effect on settling velocity, while particle size and SSC were dominant effects in the particle transport.
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 Gaoxiang Ying.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Sansalone, John.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021356:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021356/00001

Material Information

Title: Constitutive Properties of Particulates in Urban Dry Deposition and Source Area Rainfall-Runoff Loadings
Physical Description: 1 online resource (202 p.)
Language: english
Creator: Ying, Gaoxiang
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: deposition, distribution, dry, matter, particulate, partitioning, rainfall, runoff
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: Urban highway surfaces are sources of contaminants, such as metals, phosphorus, trash and debris, and oxygen demanding substances, in part, related to atmospheric deposition. Rainfall-runoff from urban highway surfaces often transports significant loads of pollutants in a complex heterogeneous mixture that includes particulate and dissolved solids, metals, phosphorus and organic compounds. Control of runoff particulate matter is challenging due to the variability in flow, variable particulate matter transport and, varying mass and concentration during a wet weather event as compared to conventional wastewater treatment loadings. This study investigated physical indices of particulate matter delivered in source area rainfall-runoff as a function of hydrologic transport and quiescent settling on an event basis. Four mass and flow-limited events, whose volume and particulate load were fully-captured and recovered from settling tanks receiving runoff from a 1088 m2 paved urban watershed in Baton Rouge, LA, were analyzed. Results indicate that mass and flow-limited behavior results in separate particulate matter delivery and relationships which are transformed to a single particulate matter relationship after unit operations such as settling. While flashed from urban surface into storm water runoff and transferred to natural water body during a storm event, equilibrium partition of phosphorus, COD and metals in dissolved and particulate phase in urban rainfall-runoff was examined before and after 60 minutes of quiescent settling for both two classes of events. Phosphorus has a high partition in particulate phase and can be removed with solids by quiescent settling, metals can be partially removed by settling process but needs a further treatment in order to reach national recommended water quality criteria, and COD cannot be effectively removed by quiescent settling. Granulometry, transport and solubility of dry deposition particles was investigated with a purpose of discovering the source and pathway of contaminants in rainfall-runoff from an urban highway land use. In addition, metal species bonded with dry deposition particles were examined and arithmetic mean concentration of metals across the entire size gradation was ranked as Cd < As < Cu < Pb < Zn < Mg < Fe < Ca constantly. Cumulative mass distributions (CMDs) of particulate matter, all modeled with a cumulative gamma distribution function, were translated a finer particle size gradation during the transport from dry deposition into rainfall-runoff, with D50m decreasing from 330.7 ?m to 13.7 ?m along the runoff stream. The settling velocity is a key variable while studying the transport mechanism of particles throughout their entire pathways from dry deposition, carried into urban rainfall-runoff, and finally discharged to different water bodies. In this study, settling velocity of particles were measured as a function of salinity, particle size and suspended sediment concentration (SSC) to illustrate various scenarios of in situ particle delivery. Salinity was found to have no significant effect on settling velocity, while particle size and SSC were dominant effects in the particle transport.
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 Gaoxiang Ying.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Sansalone, John.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-08-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021356:00001


This item has the following downloads:


Full Text

PAGE 1

1 CONSTITUTIVE PROPERTIES OF PARTICULATES IN URBAN DRY DEPOSITION AND SOURCE AREA RAINFALL-RUNOFF LOADINGS By GAOXIANG YING A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007

PAGE 2

2 2007 Gaoxiang Ying

PAGE 3

3 To my father, Changlun Ying, and my mother, Wenfang Xiao who always stand behind me, supporting me, encouraging me, and believing th ere is nothing that I cannot achieve

PAGE 4

4 ACKNOWLEDGMENTS First and foremost, I thank my supervisor, co mmittee chair, Dr. John J. Sansalone, for his consistent guidance, encouragement and support. His mentoring helped me to shape my career and strengthened my capability to pursue my future goals.. I thank my committee members, Dr. James P. Heaney, Dr. Chang-Yu Wu, and Dr. Willie G. Harris, for their reviews and valuable advice. I thank Dr. Hong Lin, Dr. Jia Ma, Dr. Ti anpeng Guo, Dr. Xuheng Kuang and Dr. JongYeop Kim for their valuable discussions and help ful ideas. I thank my student worker, Mr. Aaron S. Lancaster for his great help in my experi ments. I thank my colleagues including Mr. SubbuSrikanth Pathapati, Ms. Natalie Magill, Mr. Robe rt W. Rooney, Mr. Saurabh N. Raje, Mr. Ruben A. Keztesz and Ms. Tingting Wu who made this research possible. I express my special and warmest thanks to my best friend, Bo Liu, for his enormous help and his sincere friendship.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES................................................................................................................ .........9 LIST OF ABBREVIATIONS........................................................................................................13 ABSTRACT....................................................................................................................... ............19 CHAPTER 1 GLOBAL INTRODUCTION.................................................................................................21 2 DIFFERENTIATION OF PARTICULATE MATTER RELATIONSHIPS IN URAN RAINFALL-RUNOFF BETWEEN FLOW AND MASS LIMITED EVENTS....................27 Introduction................................................................................................................... ..........27 Objectives..................................................................................................................... ..........28 Background Study............................................................................................................... ...28 Mass-Limited Vs. Flow-Limited Behavior.....................................................................28 Particle Size Distributions (Psds)....................................................................................29 Turbidity And Particul ate Matter Indices........................................................................30 Methodology.................................................................................................................... .......31 Experimental Site............................................................................................................31 Sampling....................................................................................................................... ...32 Differentiation of Particulate Fractions...........................................................................33 Particle Size Distribution (PSD) Methods.......................................................................34 Quiescent Settling Protocol (Dissolve d, Suspended And Settleable Fractions)..............34 Mechanical Sieve Analysis (Settl able And Sediment Fractions)....................................35 Suspended Sediment Concentration (SSC) And Volatile SSC.......................................35 Particle Size Indices........................................................................................................36 Results........................................................................................................................ .............37 Mass-Limited Vs. Flow-Limited Hydrology...................................................................37 Differentiation Of Suspended, Settleable And Sediment Fractions................................37 Turbidity And SSC Relationship.....................................................................................38 Influence Of Quiescent Settling On Settleable And Suspended Particles.......................40 Turbidity...................................................................................................................... ....41 Suspended Sediment Concentration (SSC).....................................................................42 Conclusions.................................................................................................................... .........43

PAGE 6

6 3 EQUILIBRIUM PARTITIONING OF M ETALS, PHOSPHORUS AND COD IN URBAN RAINFALL-RUNOFF INFLUE NCED BY SEDIMENTATION..........................59 Introduction................................................................................................................... ..........59 Objective...................................................................................................................... ...........62 Background Study............................................................................................................... ...62 Methodology.................................................................................................................... .......70 Experimental System.......................................................................................................70 Phase Fractionation.........................................................................................................72 Particulate Fractionation..................................................................................................72 Acid Digestion.................................................................................................................73 Metal Inductively Coupled Plasma-Ma ss Spectrometry (ICP-MS) Analysis.................74 Digestion And Analysis For Phosphorus........................................................................74 Chemical Oxygen Demand (COD) Analysis..................................................................74 Results........................................................................................................................ .............75 Phosphorus Partition In Dissolv ed And Particulate Phase..............................................75 Chemical Oxygen Demand Partition In Dissolved And Particulate Phase.....................76 Phosphorus And COD Partitioning In Suspe nded, Settleable And Sediment Solids......76 Partitioning Of Metals In Disso lved And Particulate Phase...........................................77 Metal Partitioning In Suspended, Se ttleable And Sediment Solids................................79 Conclusions.................................................................................................................... .........79 4 GRANULOMETRY TRANSPORT AND SO LUBILITY OF DRY DEPOSITION PARTICLES FOR AN URB AN HIGHWAY LAND USE.................................................102 Introduction................................................................................................................... ........102 Objective...................................................................................................................... .........103 Background Study............................................................................................................... .104 Dry Deposition Sampling..............................................................................................104 Transfer Mechanism And Solubility Test Of Dry Deposition Particles........................105 Methodology.................................................................................................................... .....106 Dry Deposition Sampling..............................................................................................106 Previous Dry Hours, Dry Deposition Solid Mass And D50m.........................................107 Particle Size Distribution...............................................................................................108 Modeling Of Mass-Based Particle Size Distributions (PSDs)......................................108 Surface Area And Specific Surface Area (SA/SSA).....................................................109 Solubility Test...............................................................................................................109 Results........................................................................................................................ ...........110 Relationship Between Previous Dry Hour s And Mass Loading Of Dry Deposition Material......................................................................................................................110 Granulometric Distribution Of Dry Deposition Particles..............................................111 Specific Surface Area (SSA) And Surface Area (SA) For Dry Deposition Solids.......111 Comparison Of Cumulative Mass Distributions and D50m For Urban RainfallRunoff And Dry-Deposition Solids...........................................................................112 Influence Of Initial Ph And Particle Size On Equilibrium Ph.......................................112 Influence Of Initial Ph And Particle Size On Equilibrium Alkalinity...........................113 Influence Of Initial Ph And Particle Size On Equilibrium TDS And Conductivity.....114

PAGE 7

7 Influence Of Contact Time On Ph, TDS, Conductivity................................................114 Conclusions.................................................................................................................... .......116 5 CHARACTERISTICS, TRANSPORT AND SOLUBILITY OF DRY PARTICULATE METALS DEPOSITED ON URBAN HIGHWAY AREA.................................................129 Introduction................................................................................................................... ........129 Objective...................................................................................................................... .........130 Background Study............................................................................................................... .130 Methodology.................................................................................................................... .....133 Dry Deposition Sampling..............................................................................................133 Solubility Test For Metal Species In Dry Deposition Solids........................................135 Total/Dissolved Mass And Concentrati on Of Metal Species In Dry Deposition Solids......................................................................................................................... .135 Modeling Of Cumulative Mass Di stribution For Metal Species...................................137 Results........................................................................................................................ ...........138 Mass And Concentration Distribution Of Metal Species In Dry Deposition Particles.138 Transport And Solubility Of Metal Species In Dry Deposition Particles.....................140 Influence Of Ph On Transport Of Meta l Species In Dry Deposition Particles.............143 Conclusions.................................................................................................................... .......144 6 EFFECTS OF SALINITY AND SSC ON S ETTLING VELOCITY OF PARTICLES IN URBAN RAINFALL-RUNOFF..........................................................................................155 Introduction................................................................................................................... ........155 Objective...................................................................................................................... .........157 Background Study............................................................................................................... .157 Settling Columns...........................................................................................................157 Discrete Settling (Type I Settling).................................................................................159 Flocculant Settling (Type II Settling)............................................................................162 Single floc settling..................................................................................................162 Multi-flocs settling.................................................................................................164 Hindered Settling (Type III Settling)............................................................................166 Methodology.................................................................................................................... .....167 Results........................................................................................................................ ...........170 Conclusions.................................................................................................................... .......173 7 GLOBAL CONCLUSIONS.................................................................................................185 LIST OF REFERENCES.............................................................................................................190 BIOGRAPHICAL SKETCH.......................................................................................................202

PAGE 8

8 LIST OF TABLES Table page 2-1 Summary of particulate fr actions for 8 rainfall-runoff events with complete runoff volume recovery at Baton Rouge site................................................................................45 2-2 Summary of solid fractions and hydrographs for 8 rainfall-runoff events at I-10 site, Baton Rouge.................................................................................................................... ...46 2-3 Median diameter and size indices for suspended solids in rainfall-runoff as untreated influent (designated as time, t = 0 minute) and after 60 minutes of quiescent settling (designa ted as time, t = 60 minutes) at Baton Rouge............................47 2-4 Summary of statistical characteristics of mass-based particle size distribution for suspended solids in rainfall-runoff as untreat ed influent (Time = 0 minute) and after 60 minutes of quiescent settling (Tim e = 60 minutes) at Baton Rouge.............................48 2-5 Mean ( ) Standard deviation ( ) and R-square (r2) of turbidity and SSC in Gaussian distribution (Sample number = 60 for each ev ent) for rainfall-runoff as untreated influent (Time = 0 min) and after 60 minutes of qui escent settling...................................49 3-1 Water chemistry data for 8 rainfall-runoff events with complete runoff volume recovery at the Baton Rouge site.......................................................................................81 3-2 Concentrations of 10 metal species in in fluent runoff for 8 rainfall-runoff events at the Baton Rouge site..........................................................................................................82 3-3 Partitioning and distribution of metals in dissolved and particulate phase (sediment + settleable + suspended) for unsettled influe nt rainfall-runoff and after 60 minutes of quiescent settling............................................................................................................. ...83 3-4 Rainfall chemistry data for 4 storm events collected at Baton Rouge site........................84 3-5 Metals in particulate fractions of infl uence runoff for storm events at the Baton Rouge site. (Note units for Ca, Fe, Mg).............................................................................85 4-1 Summary of particulate mass of dry depos ition solids collected at 5 sampling sites during the period from 17 January to 16 July 2004.........................................................117 5-1 Concentration of metal species in dry deposition Particles and transport of metal species from dry deposition particles into rainfall-runoff, compared with metal concentrations in dissolved and particulat e phase of rainfall-runoff at Baton Rouge.....146

PAGE 9

9 LIST OF FIGURES Figure page 2-1 Plan view of I-10 site, Baton Rouge, LA...........................................................................50 2-2 Hydrographs of 4 flow-limited events and 4 mass-limited events with cumulative volume captured at I-10 site...............................................................................................51 2-3 Solid fraction in rainfall-runoff for 8 ra infall-runoff events at Baton Rouge site.............52 2-4 Particle size distribution for particulate matter captured in the settling basin for rainfall-runoff events at I-10 site, Baton Rouge. ( e: mass balance error).........................53 2-5 Differentiation of the constitutive relationship between turbidity and suspended sediment concentration (SSC) for 8 ra infall-events in Baton Rouge, LA.........................54 2-6 Total volume concentration (TVC) for 4 flow-limited and 4 mass-limited events at I10 site, Baton Rouge, LA...................................................................................................55 2-7 Cumulative volume concentration of unsettl ed (time = 0 min) and after 60 minutes settling (time = 60 min) rainfall-runoff and settling efficiency based on mass for 8 events at the Baton Rouge site...........................................................................................56 2-8 Probability density function (pdf ) of turbidity for rainfall-runoff.....................................57 2-9 Probability density function ( pdf) of SSC for rainfall-runoff............................................58 3-1 Equilibrium of phosphorus partitioning in dissolved and particulate phase before and after settling for 8 rainfall-runoff events, n = 60 for each event........................................86 3-2 Equilibrium of COD partitioning in dissolved and particulate phase before and after settling for 8 rainfall-runoff events, n = 60 for each event................................................87 3-3 The fd values and equilibrium coefficient, Kd values of phosphorus in unsettled influent rainfall-runoff (t0) and rainfall-runoff after 60 minutes settling (t60)...................88 3-4 The fd values and equilibrium coefficient, Kd values of COD in unsettled influent rainfall-runoff (t0) and rainfall-runoff afte r 60 minutes settling (t60).................................89 3-5 Phosphorus fraction (%) of suspended, sett leable and sediment solids in rainfallrunoff, n = 60 for each event..............................................................................................90 3-6 Chemical oxygen demand fraction (%) of suspended, settleable and sediment solids in rainfall-runoff, n = 60 for each event.............................................................................91 3-7 Equilibrium partitioning between dissolved As, (As)d and particulate As, (As)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60)......................................................................................................92

PAGE 10

10 3-8 Equilibrium partitioning between dissolved Cd, (Cd)d and particulate Cd, (Cd)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60)......................................................................................................93 3-9 Equilibrium partitioning between dissolved Cr, (Cr)d and particulate Cr, (Cr)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60)........................................................................................................................94 3-10 Equilibrium partitioning between dissolved Cu, (Cu)d and particulate Cu, (Cu)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60)......................................................................................................95 3-11 Equilibrium partitioning between dissolved Fe, (Fe)d and particulate Fe, (Fe)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60)........................................................................................................................96 3-12 Equilibrium partitioning between dissolved Mn, (Mn)d and particulate Mn, (Mn)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60)......................................................................................................97 3-13 Equilibrium partitioning between dissolved Pb, (Pb)d and particulate Pb, (Pb)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60)........................................................................................................................98 3-14 Equilibrium partitioning between dissolved Zn, (Zn)d and particulate Zn, (Zn)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60)......................................................................................................99 3-15 Equilibrium partitioning between dissolved Mg, (Mg)d and particulate Mg, (Mg)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60)....................................................................................................100 3-16 Equilibrium partitioning between dissolved Ca, (Ca)d and particulate Ca, (Ca)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60)....................................................................................................101 4-1 Dry deposition sampling duration time a nd rainfall records during the period from January to July 2004........................................................................................................118 4-2 Relationship of previous dry hours (pdh) with dry particle mass deposited for urban transportation land use site in Baton Rouge, Louisiana...................................................119 4-3 Solid gradations for 17 dry traffic de position events at the Baton Rouge site, simulated by a cumulative gamma distribution function.................................................120 4-4 Specific surface area (SSA) and Surface ar ea (SA) for dry deposition solids with different particle size gradation.......................................................................................121

PAGE 11

11 4-5 Measured granulometry of runoff (q) a nd dry deposition (DD) at the Baton Rouge site........................................................................................................................... .........122 4-6 Equilibrium pH of solubility test for 17 gradation dry deposition particles in 5 different initial pH solutions............................................................................................123 4-8 Equilibrium TDS of solubility test fo r 17 gradation dry deposition particles in 5 different initial pH solutions............................................................................................125 4-9 Equilibrium conductivity of solubility test for 17 gradation dry deposition particles in 5 different initial pH solutions.........................................................................................126 4-10 pH value as a function of time for 3 diffe rent sizes of dry de position particles in 5 different pH solutions......................................................................................................127 4-11 TDS value as a function of time for 3 diffe rent sizes of dry deposition particles in 5 different pH solutions......................................................................................................128 5-1 Mass and chemical concentration distri bution of Cu, Cd, Pb and Zn across size gradation of dry deposition particles................................................................................147 5-2 Mass and chemical concentration distri bution of As,Cr, Fe and Mn across size gradation of dry deposition particles................................................................................148 5-3 Specific surface area (SSA) and Surface area (SA) for dry deposition particles across the entire particle size gradation......................................................................................149 5-4 Leaching capability of metals (Cd, Cu, Pb and Zn) for dry deposition (DD) particles across entire size gradation at an equilibrium partitioning time of 60 minutes in rainfall....................................................................................................................... .......150 5-5 Leaching capability of metals (As, Fe, Ca and Mg) for dry deposition (DD) particles across entire size gradation at an equilibrium partitioning time of 60 minutes in rainfall....................................................................................................................... .......151 5-6 Site mean metal distribution (cumul ative mass for Cr, Cd, As, Pb, Cu, Zn) by different particle size gradation after tran sported from dry deposition into rainfallrunoff at Baton Rouge, LA. DD: dry deposition; q = runoff...........................................152 5-7 Relative incremental dissolved mass range s of metals leached from dry deposition (DD) particles at an equilibrium part itioning time of 60 minutes in rainfall...................153 5-8 Dissolved concentrations of metal sp ecies (Cd, As, Mn, Fe, Pb, Mg, Cu, Zn) in rainfall solutions after equal amount of particles in different size gradation transported from dry depositi on into urban rainfall-runoff..............................................154

PAGE 12

12 6-1 Measured median terminal settling veloci ties of particles across the size range from 1 m to 4750 m, considering the effect of vari ous salinity under discrete settling conditions..................................................................................................................... ....175 6-2 Settling velocity distribution of part icles in suspended fr action, shown as a probability density function (pdf) for each specific size gradation (1, 2, 5, and 10 m). Results were compared against various salinities...................................................176 6-3 Settling velocity distribution of part icles in suspended fr action, shown as a probability density function (pdf) for each specific size gradation (25, 50, and 75 m). Results were compared against various salinities...................................................177 6-4 Settling velocity distribution of part icles in suspended fr action, shown as a probability density function (pdf) for each specific size gradation (100, 150, and 200 m). Results were compared against various salinities...................................................178 6-5 Settling velocity of particles in Type II (flocculant) settli ng region under a quiescent settling condition, illustrate d as 3 exponential trends for original particle concentrations of 5, 10, 20 g/L respectively....................................................................179 6-6 Interface height of Type III (hindered) settling region for particles with initial concentrations of 5, 10, 20 g/L, respectively..................................................................180 6-7 Interface height of Type III (hindered) settling region for particles with initial concentrations of 50, 100, 200 g/L, resp ectively, under a quies cent condition with salinity of 1 ppt.............................................................................................................. ..181 6-8 Interface height of Type III (hindered) settling region for particles with initial concentrations of 50, 100, 200 g/L, respect ively, under a quiescent condition with salinity of 10 ppt............................................................................................................. .182 6-9 Interface height of Type III (hindered) settling region for particles with initial concentrations of 50, 100, 200 g/L respec tively, under a quiescent condition with salinity of 30 ppt............................................................................................................. .183 6-10 Relationship of settling velocity and in itial particle concentration in Type III (hindered) settling zone under a quiescent settling condition..........................................184

PAGE 13

13 LIST OF ABBREVIATIONS I : inclusive graphic standard deviation GK : kurtosis kS : skewness nd50 : median particle size ba sed on particle distribution vd50 : median particle size based on particle volume : Phi size, index of grain size distribution in sediment x : sample mean : volume concentration a, b : regression parameters AADT : annual average daily traffic cd : dissolved -bound concentrations Cd : Newtons drag coefficient CDS : continuous deflective separation Ch,t : particle concentration at height of h and time of t CICP-MS,D,i,j : measured concentration of dissolved metal i for size j from ICP-MS, ppb; CICP-MS,i,j : measured concentration of me tal i for size j from ICP-MS, ppb; CII : particle concentration in the flocculant settling zone CIII : particle concentration in the hindered settling zone CMCs : Criteria Maxium Concentrations CMDs : cumulative mass distributions; Cmetal,D,i,j : concentration of dissolved metal i in DD particles with size j, mol/m2; Cmetal,i,j : concentration of metal i in DD particles with size j, mol/L;

PAGE 14

14 COD : Chemical oxygen demand cp : particulate-bound concentrations CSF : Corey shape factor d* : dimensionless particle parameter d16, d84 : the 16th and 84th percentiles of particle size distribution d50 : median particle size based on particle mass DD : dry deposition; Df : floc diameter dg : geometric mean diameter Di : intermediate diameter Di,j : dilution number for metal i in DD particles with size j; Dl : long diameter DMM : dissolved metal mass; Dn : Nominal diameter, 3 s i l nD D D D dp : particle diameter, mm Ds : short diameter E : porosity E : shape factor f(x) : gamma distribution function F(x) : cumulative gamma distribution function FD : hydrodynamic drag force fd : dissolved fraction Fg : gravity force

PAGE 15

15 fp : particulate fraction G : gravitational force, 9.81m/s2 GTM : grey tank management H : traveling distance of particles h0 : height of water surface hi : height of interface hi,max : maximal height of interface i : metal species, such as Cu, Cd, Pb, Zn, Fe, Mn, As, Cr, Mg, Ca; ICP : inductively coupled plasma ICP-MS : Inductively coupled plasma-mass spectroscopy; j : particle size range, from >9500 m to <25 m; Kd : equilibrium coefficient Kso : heterogeneous equilibrium constant Kturb : specific turbidity Md : dry deposition mass Mdigested,j : Mass of DD particles with size j, g; mi : mass of solids having particle diameter I g Mmetal,i,j : concentration of metal i in DD particles with size j, g; Molar Massi : Molar mass for metal i, g/mol; Msolild,j : Mass of DD particles with size j as a dosage for solubility test, g; nf : floc fractal dimension NTUd : NTU found in diluted sample PCC : Portland cement concrete

PAGE 16

16 PDF : Probability density function; PSD : particle size distribution QA : Quality Assurance; QC : Quality Control; Qd : dry deposition flux Red : Reynolds number Redif : Reynolds number at incipient fluidization ref : reference volume concentration s : standard deviation SA : surface area, m2 SAi : surface area of solid s having particle diameter I m2 SSA : specific surface area, m2/g SSAi : specific surface area of solids having particle diameter i m2/g SSAj : specific surface area for DD particles with size j, m2/g; SSC : suspended sediment concentration, mg/L SSE : sum of squared errors; T : duration time T0 : time = 0 minute T60 : time = 60 minutes TSS : total suspended solid V0 : settling velocity of the equivalent sphere Vd : volume of dilution water, mL Vdigested,i,j : digested volume of metal i in for DD particles with size j, mL;

PAGE 17

17 Vh,t : settling velocity at height of h and time of t Vif : settling velocity at incipient fluidization VII : flocculant settling velocity right VIII : hindered settling velocity Vp : measured terminal settling velocity Vs : sample volume taken for dilution, mL Vsample : sample volume, mL Vsample : sample volume, mL Vsolution,i,j : rainfall solution volume for metal i for DD particles with size j, L; VSSC : volatile suspended sediment concentration, mg/L Wa : measured weight of EGME retained by sample, g Wafter : weight of filter with aluminum pan after ignition, g Wbefore : weight of filter with aluminum pan before ignition, g Wfinal : final weight of filter + residue with aluminum pan, g Ws : measured weight of dried sample, g Wtare : tare weight of filter with aluminum pan, g X : particle diameter, m Z : gamma function parameter : scale factor : scaling parameter : shape parameter if : Voidage at incipient fluidization : dynamic viscosity, 2/ m s N

PAGE 18

18 : kinematics viscosity, m2/s f : floc density, g/cm3 p : particle density, g/cm3 w : water density, 1.0 g/cm3 50 : central tendency x incomplete gamma function gamma function

PAGE 19

19 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CONSTITUTIVE PROPERTIES OF PARTICULATES IN URBAN DRY DEPOSITION AND SOURCE AREA RAINFALL-RUNOFF LOADINGS By Gaoxiang Ying August 2007 Chair: John J. Sansalone Major: Environmental Engineering Sciences Urban highway surfaces are sources of contam inants, such as metals, phosphorus, trash and debris, and oxygen demanding substances, in part, related to at mospheric deposition (Harned, 1988). Rainfall-runoff from urban highwa y surfaces often transports significant loads of pollutants in a complex heterogeneous mixture that includes particulate and dissolved solids, metals, phosphorus and organic compounds. Control of runoff particulate matter is challenging due to the variability in flow, variable pa rticulate matter transpor t and, varying mass and concentration during a wet weather event as co mpared to conventional wastewater treatment loadings. This study investigated physical indices of particulate matter delivered in source area rainfall-runoff as a function of hydrologic transport and quiescent settling on an event basis. Four mass and flow-limited events, whose volume and particulate load were fully-captured and recovered from settling tanks receiving runoff from a 1088 m2 paved urban watershed in Baton Rouge, LA, were analyzed. Results indicate that mass and flow-limited behavior results in separate particulate matter de livery and relationships whic h are transformed to a single particulate matter relationship af ter unit operations such as settling. While flashed from urban surface into storm water runoff and transferred to natural water body dur ing a storm event,

PAGE 20

20 equilibrium partition of phosphorus, COD and metals in dissolved a nd particulate phase in urban rainfall-runoff was examined before and after 60 minutes of quiescen t settling for both two classes of events. Phosphorus ha s a high partition in particulate phase and can be removed with solids by quiescent settling, metals can be par tially removed by settling process but needs a further treatment in order to reach national r ecommended water quality criteria, and COD cannot be effectively removed by quiescent settling. Granulometry, transport and sol ubility of dry deposition particles was investigated with a purpose of discovering the source and pathway of c ontaminants in rainfall-runoff from an urban highway land use. In addition, metal species bonde d with dry deposition particles were examined and arithmetic mean concentration of metals ac ross the entire size gradat ion was ranked as Cd < As < Cu < Pb < Zn < Mg < Fe < Ca consta ntly. Cumulative mass dist ributions (CMDs) of particulate matter, all modeled with a cumulativ e gamma distribution function, were translated a finer particle size gradation dur ing the transport from dry depos ition into rainfall-runoff, with D50m decreasing from 330.7 m to 13.7 m along the runoff stream. The settling velocity is a key variable while studying the transport mechanism of particles throughout their entire pathways from dry depos ition, carried into urban rainfall-runoff, and finally discharged to different water bodies. In this study, settling veloci ty of particles were measured as a function of salinity, particle size and suspended sediment concentration (SSC) to illustrate various scenarios of in situ particle delivery. Salinity was found to have no significant effect on settling velocity, while particle size and SSC were domi nant effects in the particle transport.

PAGE 21

21 CHAPTER 1 GLOBAL INTRODUCTION Urban land uses, specifically tr ansportation land use combined with vehicular activities are a dominant source of particulate matter (PM). PM accumulates within and is transported from transportation land uses. Th e coupling of hydrologic phenomen a with the higher surface conveyance capacity of anthropoge nic interfaces (such as pavement and compacted soils), and vehicular activities results in th e delivery of significantly higher loads of PM. For example, in Cincinnati Ohio, annual loads of PM transported in rainfall-runo ff from interstate and arterial roadway paved surfaces (represen ting approximately 10 % of urban area) are greater than those delivered in untreated domestic wastewater from the population of 800,000 (Sansalone et al. 1998). Although this paved area represents a relatively small proportion of the geographic landscape, the high degree of imperviousness comb ined with anthropogenic load generation has a significant impact on the quantity and qua lity of rainfall-runoff (Drapper 2000). The pavement and vehicular activities are majo r sources of metals, initially as abraded material. Vehicles are the major source of abraded metal deposition on transportation land use surfaces (Bourcier and Hinden 1979). Metals most frequently reported include Al, Cd, Cr, Cu, Fe, Ni, Pb, and Zn (Sansalone and Buchberger 1997). These metals can adversely impact receiving waters by increasing toxi city in the water column and sediments, with bioaccumulation in the food chain (Yousef et al. 1987). For toxico logical impacts from urban rainfall-runoff Zn, Cd, Cu and Pb are commonly examined with resp ect to discharge regulations. Transportation land use runoff concentrations of Zn, Cu, Cd, Pb, Cr and Ni are sign ificantly above ambient background levels, and for some heavily traveled roadways, Zn, Cu, Pb and Cd often exceed USEPA and State EPA surface water discharge criteria on an event basis (Sansalone and Buchberger 1997). Depending on fact ors such as pH, alkalinity, re sidence time, and complexing

PAGE 22

22 agents, these metals can partition into the aqueou s phase as ionic species with the potential to exert an acute toxicity impact (Sansalone and Cristina 2004). While PM and metals are most commonly sourced to specific transportation land uses and activities, elevated nutrient lo ads with commensurate levels of eutrophication are also identified (USEPA 1993). Specific sources are biogenic, for example right-of-way (ROW) vegetation, pavement materials and admixtures such as phosphogypsum, and right-of-way management practices such as fertilizer a ddition, irrigation wastewater re use and mowing. Transportation land use discharges combined with receiving wa ters that support recreational and contact activities has resulted in cases of neurological damage for indivi duals exposed to the highly toxic volatile chemical produced by dinoflagellates; in creasing public awarenes s of eutrophication and the need for solutions. Waschbusch et al. (1999) found that lawns and st reets are the largest sources of total and dissolved phosphorus and th eir combined contribution was approximately 80 % with respect to mass. Other sources of the total phosphorus mass in the runoff include fertilizers and biogenic material including leaves, twig s, and other organic debris (Ray 1997). Constituents such as metals, phosphorus, pesticides, organics, and oxygen demanding substances, can be associated with PM, organi c material or mineral phases transported in rainfall-runoff (USEPA 1999). The granulometry, sp ecifically density, part icle size distribution (PSD) and surface characteristics of PM s in the rainfall-runoff plays an important role in determining transport and fate of PM, as well as the associated contaminants including metals and phosphorus (Sansalone et al. 2005). Rainfall-r unoff can delivery PM from their source areas of generation and disperse them through the watershed, potentially im pairing larger areas (Magnuson et al. 2001). The transp ort of metallic pollutant in ra infall-runoff is a function of partitioning and the distribution across the PSD.

PAGE 23

23 While granulometric measurements such as PSD s and unit operations such as separation of PSDs from runoff have a fundamental physical basis, the adoption of automated sampling methods, initially developed to monitor primary tr eatment of wastewater, has resulted in almost four decades of results based on runoff PM meas ured as total suspended solids (TSS). While automated sampling, and PM measured as TSS, can potentially characterize the lighter and more organic PM in a steady-flow wastewater primar y clarifier, such methods generate a very truncated PSD mis-representing the denser, more discrete and highly hete ro-disperse gradations generated by urban source areas such as transpor tation land uses. In addition to truncating the PSD, the common combination of automatic sampling and PM measured as TSS can also misrepresent the distribution of metals with PSD fractions. For example, results from automated sampling and TSS methods will demonstrate that me tals are predominately associated with the TSS fraction from transportation land use runo ff (Sansalone et al. 1995). However, when the entire PSD is sampled and analyzed, results indica te that the predominance of metal mass can be associated with the sediment-size fraction (> 75 m) for the same urban transportation land use (Turer et al. 2001, Sansalone and Cristina 2005). For the coarser gr avel-size fracti on (larger than 2000 m) the relative importance of these particles in the transport of adsorbed metals from urban areas is diminished (Sartor and Boyd 1972). This dissertation will quantify the highly hetero-disperse nature of source area PM, distribu tion of metals across the entire PSD, and how an urban catchment such as a highway ROW or un it operation such as set tling, act as a modifier of PSDs. Control of particulate matter mass and con centration is challenging due to the wide variability in flow, hetero-disperse nature of PM, variable mass and concentration transport during a wet weather event as compared to the far more steady-state hydr aulic and organic PM

PAGE 24

24 loadings to a conventional wastewater treatment plant. The historical reliance on automatic sampling and TSS results have the unintended consequence of mis-representing PSDs, metal partitioning and distribution, PM and metal load s. As a result such non-representative monitoring and analysis is likely to lead to improper sel ection, design, operation, maintenance, with a commensurate impacts on effluent concentr ation and loads.. Therefore, a primary goal of this research was to provide an examination of PM across the entire gradation with respect to PSDs, partitioning, distribution and the PM size-fractions of suspe nded, settleable and sediments. Atmospheric deposition, such as dust fall and rain fall, as well as more localized traffic resuspension (dry transportation deposition) plays a major role in pollutant circulation and transport in urban areas (Col man and Breault 2000; Yun et al. 2002). Transportation land use systems represent significant line loads of pollu tants that dominate adjacent land uses. On-road vehicles are the major emission sources for PM 10 according to the 1997 Alachua County PM10 emission inventory (Chuaybamroong et al. 2007). Dry deposition (dust fall) is considered one major pathway for removal of contaminants fr om the urban atmosphere or airshed (Kobriger 1984); however, instead of eliminating these constitu ents from the environment, these pathways only transport constituents to terrestrial surfaces of a watershed and eventually a significant fraction are delivered to receiving waters. While contaminants are removed from the airshed the overall pollutant loadings are not reduced to th e environment (Barrett et al. 1995). Contaminants accumulated on impervious urban surfaces from dr y traffic deposition can be easily washed off during a storm and create a signifi cant loading to urban rainfallrunoff (Sansalone et al. 1998). With respect to loads of metals, phosphorus and PM, most untreated rainfall-runoff is incorporated in the urban ecosystem, at least te mporarily, and therefore constituents transported

PAGE 25

25 in runoff can eventually impact the terrestri al and aquatic system once deposited (Glenn and Sansalone 2002). A number of studies have examined the pave ment surface as a temporal reservoir of dry traffic deposition and traffic re-suspended PM. St udies have reported that this build-up function between wash-off events can be represented as a first-order exponentia l accumulation function (Vaze and Chiew 2002). Recovery and characteri zation of this dry traffic deposition PM indicates that most of the PM is inorganic a nd with a median size in the sand-size range (Characklis 1997; Sator and Boyd 1972; Sator et al. 1974). However, the contribution of dry deposition PM generated by traffic has been genera lly difficult to measure because of the large variation of particle size, pa vement surface characteristics, and the turbulent atmospheric interface generated by traffic (Yun et al. 2002; Caffr ey et al. 1998), and in particular, traffic activities. Other than the studies just mentione d previous studies have generally not addressed the granulometric distribution of dry deposition PM and a very limited number of studies have characterized the entire PSD of transportation land use runoff at the edge of the pavement or even within the ROW. In addition, the fate of th e metals in dry traffic deposition and pavement runoff is rarely investigated. Few studies have focused on the contribution of metals from dry traffic deposition to urban rainfall -runoff. Therefore a goal of this research was to examine the change in partitioning from dry traffic depos ition to source area runoff as a function of granulometry, specifically PSDs a nd particle surface area from an inter state catchment in urban Baton Rouge, Louisiana. With respect to fate of transported PSDs a nd associated distribution of contaminants, settling velocity is a fundamental variable in the transport of particles throughout their entire pathway from source to accretion areas (Jimen ez and Madsden 2003). Many empirical equations

PAGE 26

26 have been utilized to describe settling velocities, an accurate description being critical in the design and performance of unit operations and pr ocesses (UOPs) for st ormwater runoff (Cheng 1997). In most treatment UOPs multiple settling m echanisms can be identified in UOP settling zones at a given time by considering PSDs, pa rticle concentration, hydrodynamics and tendency of particles to interact with each other (Metcalf and Eddy 1991). Discrete particle settling velocity is a constant with specific particle characteristics such as particle size, shape and density (Hawley 1982). Salinity, as a physic al and chemical parameter, alters both water density and i onic strength. Most anthropogeni c and biogenic particles are irregular with variable shape a nd roundness, and these significan tly influence settling velocity (Baba 1981b). In order to simulate the settling velocity of natural pa rticles, Newtons Law requires modification to incorporate the aspect of variable particle shapes. Particle interactions occur more frequen tly at higher concentrations, generating aggregation or flocs with various structures and dimensions. In th ese areas, the effect of salinity on the flocculant settling velocity appeared to be significant. Particle collisions are usually generated by three different mechanisms: Br ownian motion, fluid shear, and differential sedimentation (McCave 1975; Hunt 1980). The sett ling velocity of flocs may vary with the particle concentration, depth, velocity gradients, and range of particle sizes until those flocs fall into the hindered settling zone (Lau and Krishnappan 1992). The eff ect of irregular particles or flocs on hindered settling was greater than th at for spherical ones, leading to a larger sedimentation exponent in the Richardson and Zaki expression and also leading to lower than expected concentration gradients (Tomkins 2005).

PAGE 27

27 CHAPTER 2 DIFFERENTIATION OF PARTICULATE MATTER RELATIONSHIPS IN URAN RAINFALL-RUNOFF BETWEEN FLOW AND MASS LIMITED EVENTS Introduction Rainfall-runoff is capable of transporting significant loads of particulate matter and chemical constituents from urban surfaces, in particular, pavement surfaces. For example, in Cincinnati Ohio, particulate matte r transported in rainfall-runoff from interstate and arterial pavement surfaces (approximately 10 % of urban ar ea) transports an annual load of particulate matter greater than particulate matter (measured as total suspended solids, TSS) delivered in untreated domestic wastewater from the populat ion of 800,000 (Sansalone et al. 1998). Although this paved land use represents a relatively sma ll proportion of the geograp hic landscape, the high degree of imperviousness and load has a sign ificant impact on the quantity and quality of rainfall-runoff (Drapper 2000). A la ter study indicated that the di rectly connected impervious area (DCIA), which covers 44 % of the catchment contributes 72 % of the total runoff volume during 52 years (Lee and Heaney 2003). In addi tion, constituents such as metals, phosphorus, pesticides, organics, and oxygen demanding substances, can be a ssociated with particles or mineral phases transported in rainfall-runoff (USEPA 1999). Depending on the partitioning, particulate matter can be an important vehicle for transport of these constituents (Muschack 1990, Sansalone and Buchberger 1997). Particulate matter in rainfall-runoff is heterogeneous in size, minera l and organic makeup as well as density. Most floatable materials are largely organic (anthropogenic or biogen ic) material, including oil and grease, vegetative matter, and products such as cigarettes, plastics and styrofoam. In urban runoff from pavement areas suspended, settleab le and sediment fractions are predominately inorganic (Sansalone et al 1998). The partiti oning and distribution of constituents on this particulate matter influences the selection, desi gn, operation, maintenanc e and performance of

PAGE 28

28 unit operations and processes (UOPs). Separation of particulate matter mass and concentration is challenging due to the wide variability in flow particulate matter granulometry, transport and resulting variability for mass and concentration delivery during a wet weather event as compared to conventional wastewat er treatment loadings. Objectives This study had three objectives. The first objective wa s to capture and collect the entire volume of runoff from mass-limited and flow-limite d events in order to examine the physical characteristics (particle fractions, concentra tion and particle volume and particle number distribution) of the source area particulate matter. The second objective was to examine, if the relationship between SSC and turbidity for ma ss-limited and flow-limited events could be differentiated, how this relationship was influenc ed by quiescent settling. The third objective was to examine the role of 1 hour of quiescent settli ng on particle volume and particle number distribution. Background Study Mass-Limited Vs. Flow-Limited Behavior Previous research for small and paved urban watersheds has demonstr ated that rainfallrunoff events can be differentiated into two lim iting types of behavior: mass-limited and flowlimited based events. Criteria for differentiation, generally based on hydrologic characteristics or examined with respect to pollutant loading ch aracteristics and phases, are briefly explained herein, but described in detail elsewhere (S ansalone et al. 1998, Cr istina and Sansalone 2003, Sansalone and Cristina 2004). Mass-limited ev ents generate a first-order exponential relationship between cumulative runoff volume and transported constituent mass (for example, SSC). Such events usually generate high runo ff volume, have low load generation during the event, for example as characterized by a relativ ely low vehicle to runoff volume ratio (VPV) for

PAGE 29

29 watersheds dominated by tra ffic loadings, and have a low correlation between constituent pollutographs and the corresponding hydrographs. In source area watersheds, this limit of event classification produces the concept of a mass or c oncentration first-flush. In contrast, flowlimited events generate a zero-order exponentia l relationship between cumulative runoff volume and transported constituent mass (for example, SSC), are generally low volume and low runoff intensity events, and do not produce a mass or concentration first-f lush. Such events usually generate low runoff volume, have high vehicle to runoff volume ratio (VPV) and have a high correlation between a constituent pollutographs and the corresponding hydrographs. In flowlimited events, constituent mass is still available for delivery sin ce flow is limiting transport of mass, while in mass-limited events, initial depl etion of constituent mass by flow limits the subsequent transport of particul ate matter throughout the event (Sansalone et al. 1998). In this study, hydrology was utilized to identif y mass and flow-limited behavior. Particle Size Distributions (Psds) Many studies have summarized granulometri c PSDs transported either in runoff or deposited on pavement surfaces (Sartor and Boyd, 1972; Sartor et al., 1974; Shaheen, 1975; CH2M-Hill, 1982; NCDNRCD, 1983; WCC, 1993; WCC, 1994; Sansalone et al., 1998, USEPA 1999). All PSDs are reasonably similar with the d50m (mass-based median particle size) ranging from 300 to 700 m in runoff, and 250 to 400 m for deposition on pavement surfaces. Shaheen (1975) found that for the entire gradation of particles deposited on urban pavement surfaces, 10 % were less than 75 m, 32 % between 75 to 250 m, 24 % between 250 and 420 m, 19 % between 420 and 850 m, and 15 % between 850 and 3350 m based on mass. These results were similar to results by Sansalone et al. ( 1998) for source area urban pavement rainfall-runoff which indicated that 10 % of the mass was less than 100 m, 25 % was from 100 to 400 m, 15

PAGE 30

30 % was from 400 to 600 m, 20 % was from 600 to 1000 m and 30 % was from 1000 to 9500 m. Turbidity And Particulate Matter Indices Turbidity is an important indictor of suspende d particulate levels in rainfall-runoff. Lewis (1996) indicates that runoff loadi ng estimated with continuous remote turbidity data were more accurate than that estimated by discharge. More than particle concentration, turbidity is also influenced by particulate charac teristics, dissolved organic matter and acids, temperature, and other coloring not associated with particulates (Gippel, 1989; USEPA 1999). The relationship between turbidity and suspended solids (typically measured as total suspended solids, TSS) has been examined in several studies (Anderson and Potts 1987; Gippel 1989; Gippel 1995). In these studies, the relationship between turbid ity and TSS depends on 4 main factors including particle size, shape, composition (organic and inorganic) and degree on water color. Gippel (1995) found in his study that particle size variati ons could cause turbidity to vary by a factor of four for the same gravimetric value of TSS. Resu lts also indicated that color influenced turbidity to a small degree; colo r altered turb idity by less then 10 % (Gippe l 1995). Particle shape and refractive index cause differences in scatteri ng efficiency between mineral particles which directly affects the turb idity value (USEPA 1999). There are two equations summarized by Gippel (1995) to express the relationship between turbidity and TSS. If the particle size distribut ion and particle compositi on do not vary with TSS, the equation is expressed as below. TSS K Turbidityturb (2-1) If the particle size varies w ith concentration, the relations hip is given in equation 2-2. cTSS Turbidity ) ( (2-2)

PAGE 31

31 In these empirical expressions, turbidity uni ts are NTU and TSS units are [mg/L]. The coefficient is a function of the con centration of dissolved orga nic matter, 1 for attenuance turbidity, and 1 for nephelometric turbidity. is equal to 0 for either nephelometric or attenuance instruments, or where the color increas es from zero with par ticle concentration. The exponent c is less than 1 when th e particle size increases with particle con centration, and greater than 1 when the particle size de creases with particle concentr ation. Packman (1999) developed a log-linear model which illustrated a strong correl ation between TSS and turbidity with R square value around 0.96, expressed as equa tion 2-3 where C is a constant. C Turbidity TSS ) ln( 32 1 ) ln( (2-3) Methodology Experimental Site The study watershed is a completely-paved area located in urban Baton Rouge, Louisiana. The watershed area of this site is 1088 m2 and the annual average daily traffic (AADT) load is 141,000 vehicles (east and westbound). The watershe d pavement consisted of Portland cement concrete (PCC). Mean annual precipitation at the site is 1460 mm/year. East Baton Rouge Parish is designed as a National Pollutant Di scharge Elimination System (NPDES) Phase II entity. Rainfall-runoff from the watershed was co llected by catch basins at the edge of the pavement and a PVC trough located under each of th e bridge expansion joints to pick up any flow through the expansion joint. Catch basin fl ow drained from the bottom of the catch basin by existing vertical piping whose flows were co llected by PVC piping with a 10 % slope, and the PVC troughs were connected to this same pipi ng that was approximately 15 m in length for eastbound and westbound catchments. All pipi ng and watershed runoff was combined approximately 3 m upstream of a calibrated 50.8 mm Parshall Flume. Flow heights were

PAGE 32

32 measured at 1-second intervals and reported at 1-minute intervals using a 70-kHz ultrasonic sensor and data logger. For this study, all flow discharged from the Parshall flume into a drop box used for manual sampling was diverted into a conical bottom stor age basin supported by a steel frame on a concrete slab at the site, inst ead of through experimental treatment systems at the site. The storage basin was equipped with sewage pumps and an internal and external recirculation system to maintain a well-mixed c ondition during the experiment s. The external recirculation system discharged into the piping upstr eam of the Parshall Flume so that external recirculation flow rates could be monitored. A da ta-logging rain gage was located within 100 m of the site in an open grassed ar ea adjacent to the lake, and lo cated away from any vertical obstructions. The tipping bucket rain gage measured and collected rainfall data in increments of 0.254 mm (0.01 inches). Sampling Once the entire event volume was captured in the storage basin and hydrograph flow ended, the basin was well-mixed utilizing both in ternal and external re -circulation. Complete basin mixing and re-circulation was ensu red through the use of three 400 L-min-1 sewage pumps. A minimum of 30 sample sets of replicated sa mples per event were taken manually in the drop box from flow externally recircul ated from the internally well-mixed basin. Each sample set contained three independent samples that were re plicated. Replicate 1.1-L samples were taken for Imhoff Cone quiescent settling tests. Dupli cate 500-ml samples were taken for dissolved or particulate analysis (Temperatu re, pH, SSC, VSSC, turbidity, TD S, Conductivity, PSD). A 12-L sample was also taken to obtain the sediment-siz e fraction and mass. After 30 sample sets were taken, particles in the basins we re either allowed to settle for 24 hours or until the particle concentration was less than 5 L/L in the basin supernatant as determined by supernatant laser

PAGE 33

33 diffraction analysis. The basin supernatant was then siphoned over an additional 24 hour period. Once the supernatant was siphoned, the basin part iculate slurry was collected. Slurries were dried at 40 C in shallow Plexiglas tanks (2 m 1.5 m 0.5 m). The system was cleaned and prepared to capture the next runoff event. Differentiation of Particulate Fractions Particulate matter in rainfall-runoff can be categorized into four fractions: dissolved, suspended, settleable and sediment (Lin 2003). The nominal 1 m size is considered to be the separation size criterion that di fferentiates total dissolved so lids (TDS) and the suspended particulate fraction (glass fiber filters separa ting suspended from dissolved solids are commonly specified as a nominal 1 m glass fiber filter), following the Standard Method 2540 (APHA 1995). Nominal differentiation size between suspe nded and settleable fractions was determined by analysis of particulate matter remaining in an Imhoff Cone supernatant suspension after 60 minutes of quiescent settling. The particle size gradation of the suspended fraction was measured by laser diffraction (Imhoff Cone supernat ant). The settleable fraction was determined by Standard Method 2540F. The settleable mass was recovered, dried and measured. Settleable fraction measurement was (mL/L), in an aqueous volume of 1.0 L, and specific gravity (g/cm3) of the settleable solids determined. The sedi ment fraction was nominally determined to be 75 m and was separated before Im hoff cone testing. Therefor e, each 12 L sample was prescreened using a #200 sieve and retained sedime nts recovered, dried and analyzed for each sample set. Differentiation between settleable and sediment fractions is in keeping with conventional designations separating fine and coarse fractions at 75 m (ASTM, 1993). All particulate matter was dried at temperatures of less than 40 C.

PAGE 34

34 Particle Size Distribution (PSD) Methods There are many methods to determine particle si ze distributions (PSD) in rainfall-runoff. Two methods were utilized in this stu dy. For the suspended fraction, where number concentration dominated mass c oncentration and particles co uld remain entrained during analyses, laser diffraction was u tilized. The laser diffraction an alysis system was used to determine particle number concentr ation for the settleable and very fine sediment fraction. This system provided total volume concentrations (TVC) in 32 logarithmica lly-spaced increments with analysis limits that ranged from 1 to 250 m. In contrast, for the sediment fraction where mass concentration dominated number concentrati on, mechanical sieve analysis was utilized. The mechanical sieve an alysis ranged from 9500 m to the settleable size fractions of 25 m, and utilized 17 discrete size increments. Conve rsion between the two sy stems of measurement was made through knowledge of particle dens ity and particle volume concentration. Quiescent Settling Protocol (Dissolved, Suspended And Settleable Fractions) Quiescent settling was facilitated utilizing an Imhoff Cone procedure for each of the 60 (30 samples with duplicates) well-mixed 1.1-L sample s per rainfall-runoff event. These samples were pre-screened with a #200 sieve to separate out the sediment-size fr action. Laser diffraction analysis quantified TVC di stributions from 1 to 250 m at time 0 and after 60 minutes of quiescent settling. Approximately 100 mL was re quired for each TVC analysis. In addition, each Imhoff Cone sample was analyzed for turbid ity, suspended sediment concentration (SSC), volatile suspended sediment c oncentration (VSSC), and total dissolved solid (TDS). The methodology created a step-wise process so suspe nded, settleable, and sediment fractions could be identified through separation of all particles fr om the entire 1-L volume. All particles from

PAGE 35

35 the Imhoff Cone supernatant were filtered, dried, gravimetrica lly-measured, and retained for analysis. Mechanical Sieve Analysis (Settlable And Sediment Fractions) The entire particulate matter gradation from the entire runoff event volume from each Baton Rouge event went through granulometric an alyses. After air-drying at 40C, particulate matter was disaggregated and sieved through a set of graded sieves ranging from 9500 m (#3/8) through 25 m (#500). Analysis followed ASTM D422-63 with the addition of many more sieve sizes (ASTM, 1993). Based on lase r diffraction size and number analyses of the suspended and settleable fraction, the settl eable fraction was determined to have a lower size limit of approximately 25 m (based on the cumulative d90m of the suspended fraction) and an upper limit of approximately 75 m. This upper limit corresponds to a US Standard #200 sieve differentiating silt and sand, and also the fine and coarse designation for soils and sediments (ASTM, 1993). The lower limit of 25 m corresponds to the #500 sieve. Mass balances errors remained below 2 % with respect to the original mass of the gradation. Suspended Sediment Concentrat ion (SSC) And Volatile SSC Because of the wide gradation of particulate ma tter in rainfall-runoff, questions have been raised regarding the commonly-u tilized total suspended solids (TSS) method of analysis for representation of particulate ma tter (Gray et al. 2000). Studies of fluvial and highway-runoff sediment data indicate that T SS analysis under-represented the tr ue sediment concentration by 25 to 34 % and the total load by seve ral orders of magnitude (Bent et al. 2001). The United States Geological Survey (USGS) has recommended th e suspended sediment concentration (SSC) method of analysis be utilized for rainfall-r unoff instead of TSS analysis (ASTM D3977-97, Bent et al. 2001, Gray et al 2000). The SSC me thod of analysis differs from the TSS method

PAGE 36

36 (Standard Method 2540 C) by performing the analyses of particulate mass from the entire sample volume while the TSS method measures particulate ma tter from an aliquot of the original sample that results in a high probability of unrepresent ative sampling of the particle gradation in particular the settleable and sediment fractions. The volat ile fraction of SSC (VSSC) was determined by ignition at 500 50C in a muffle furnace. Particle Size Indices Several indices can be used to describe the shape and dimension characters of particles. Innman (1952) suggested using the geometri c mean of the particle diameters for nonsymmetrical or lognormal distributions The geometric mean diameter dg is calculated in equation 2-4. 84 16d d dg (2-4) The standard deviation g is also obtained from d16 and d84 in equation 2-5: 16 84d dg (2-5) The -size is often used in desc ription of particle size. -size is the negative natural logarithm value of corresponding diameter. Central tendency 50 is given in equation 2-6. ) ( log50 2 50d (2-6) The other factors, including uniformity factorI symmetry factor Sk and normality factor KG, are determined respectively in equation 2-7 to 2-9. ) 6 6 4 (5 95 16 84 I (2-7) ) ( 2 2 ) ( 2 25 95 50 95 5 16 84 50 84 16 kS (2-8)

PAGE 37

37 ) ( 44 225 75 5 90 GK (2-9) Results Eight rainfall-runoff events were fully-captu red at the urban Bat on Rouge watershed and examined for physical particulate matter charac teristics and indices. These events were differentiated into two classes: mass-limited or flow-limited. Mass-Limited Vs. Flow-Limited Hydrology The hydrographs for each event are summari zed in Figure 1. The hydrographs for the flow-limited events (left-hand side of figure) a nd mass-limited events (right-hand side of figure) are significantly different with respect to peak flow rate. For mass-limited events, the peak flow rate values were, with one exception, an order of magnitude larger than the peak flow rate for flow-limited events. For both classes of events, the duration of runoff was similar and relatively short, with flow durations lasting less than 2 hours. Event based hydrologic statistics are summarized at the top of Table 1. Differentiation Of Suspended, Se ttleable And Sediment Fractions For each sample set, three samples (and duplic ates) of differing volumes were collected; 0.5 L, 1.1 L and 12.0 L. These sample sets were collected for the purpose of quantifying dissolved, suspended, settleable and sediment frac tions. Results for each event, on an event basis, are summarized in Table 1. Results in Tabl e 1 indicate that, in ge neral, settleable and sediment concentrations were greater for ma ss-limited events as compared to flow-limited events. Previous dry hours (PDH) appear to be independent of event di fferentiation and results regarding mass or concentration. A summary of the event-based results for each particulate fraction or index is shown in Table 2. Resu lts indicate that the se ttleable concentration dominated the suspended concentration as well as the sediment concentrati on for the site.

PAGE 38

38 On a mass basis, the relative predominance of the suspended fraction was examined with respect to the settleable and se diment fractions, and these results are summarized in Figure 2. Results indicate that the suspende d fraction represented a third of the total mass or greater for flow-limited events, while for the mass-limited events the suspended fraction represented a third or less of the total mass. As shown in Figure 2, with one exception, flow-limited events mobilize and deliver a smaller fraction of sediment as co mpared to mass-limited results. Mass results in Figure 2, as with the concentration results in Ta ble 1 and 2, indicate that the settleable fraction was the dominant fraction across both classes of events. The entire gradation of particles is illustrated in Figure 3. Figu re 3 also illustrates that the predominance of mass resided in the sett leable and sediment fractions. The d50m for each event was either in the settleable or sediment fracti on of the gradation. The sediment fraction is removed very rapidly under quiescent settling, by definition the settleable fraction is largely removed under quiescent settling, and by definition, the suspended fraction remains. Turbidity And SSC Relationship Relationships between turbidity and SSC were examined, and results illustrated in Figure 4 as either a function of event classification or under the influence of qui escent settling. Results illustrate three distinct trends, and results were consistent whether examined on an event or sample basis for well-mixed conditions. For ma ss-limited events, the re lationship between SSC and turbidity illustrated the influence and role of larger particles, and this relationship was distinctly different (higher) as compared to the flow-limited relationship between SSC and turbidity; a relationship that was predominately influenced by suspended particles. These SSCturbidity relationships were also compared to an SSC-turbidity relations hip after 60 minutes of quiescent settling where sediment a nd settleable particles did not influence the relationship. The influence of settling of the sediment and settleab le particles resulted in a separate distinctly

PAGE 39

39 different (lower) relationship than the flow-lim ited relationship. The uppe r two plots are samplebased and the lower plots are event-based with ut ilized means and standard deviations from the sampling data. Time 0 represents the influent SSC before any settling occurs and time 60 minutes represents the settled SSC after 60 minutes of quiescent settling. All relationships between SSC and turbidity could be explained to a significant degree by a lin ear relationship. In the plots on the left side of Figure 4 there are distinct linear relationships illustrated in each plot with statistically significantly different slopes. The equations of these straight lines are expressed in equation 2-10 and 2-11 respectively. SSC [mg/L] = 2.83 (NTU) (2-10) SSC [mg/L] = 1.23 (NTU) (2-11) The coefficients of determination are high a nd a visual examination of the data also confirms two distinctly differe nt relationships for mass-limited and flow-limited events. The four events located on the upper trend line with slopes equal to 2.83 are all mass-limited events and the other four events on the lower trend line with slope of 1.23 are all flow-limited events. The plots on the right side of Figure 4 provide the relationship of turbidity and SSC after 60 minutes of quiescent settling for both masslimited and flow-limited events. Due to the modification of the each PSD by unit operations su ch as quiescent settling (resulting in removal of a majority of the sediment fraction, and set tleable fraction) the tw o differing relationships between turbidity and SSC rotate towards a single new trajectory of smaller slope. This relationship is expre ssed in equation 2-12. SSC [mg/L] = 0.98 (NTU) (2-12)

PAGE 40

40 The slope of this relationship is numerically smaller than for either mass or flow-limited events because larger size par ticles are removed by the settling operation. The separation of sediment and settleable particle s yields a nearly 1:1 relations hip between SSC and turbidity. Relationships between SSC and turbidity are di stinctly different depending on particle size distributions (PSDs), as selected in this study by either the clas sification of the event or the selection of PSDs based on a unit operation such as settling. Influence Of Quiescent Settling On Settleable And Suspended Particles PSDs (measured as total volume concentrati on, TVC, and particle number density, PND) across the suspended and settleable fractions were quantified for ea ch sample at initial conditions (time 0) and after 60 minutes of quiescent settling in an Imhoff Cone. Results are summarized in Figure 5 for flow-limited events and mass-limited ev ents. Each bar represents the mean of the measured TVC for 60 separate samples and the ra nge bars represent the standard deviation. Gradation indices for initial conditions and afte r 60 minutes of settling are summarized in Table 3 and 4. For all events, results i ndicate that particles towards the larger settleable fraction sizes were settled to a greater extent than the smalle r particles in the suspe nded size range. Across the gradation, mass-limited events had a higher reduction in TVC ( x = 14.32, = 12.16), while flow-limited events had a lower reduction in TVC ( x = 12.33, = 6.73). While the relative reduction in larger particles in comparison to sm aller particles is expected due to quiescent settling, the 05 March 2004 event app ears to be an exception to results from other events. This event had the highest initial con centration of suspende d particles of all events, and the relative increase in larger particles reflects the likely generation of flocs through differential sedimentation and commensurate floc generation w ith sizes in the coarse suspended fraction and settleable fractions. Results in Table 3 and Tabl e 4 indicate that quiesce nt settling resulted in a reduction in d50v and d50m indices for all events except for 05 March and 10 July. Across all

PAGE 41

41 events, the mean d50n, based on number of particles, did not change more than 1 m (slightly over 1 m for 1 June) after quiescent settling, indicating the influen ce of the large number of particles in the finer suspended fraction that did not settle or coagulate. In contrast, indices that are based on volume or mass indicated that the d50v, d50m decreased, illustrating the settling of larger particles for which the predominance of vol ume and mass of the gradation are associated. Results in Figure 6 illustrate the quiescent sett ling behavior as a func tion of size across the suspended and settleable particle size ranges. Separation effici ency increased as size increased. Turbidity The probability density functions (pdf) for turbidity were determined from 60 well-mixed samples for each event. It was hypothesized that representative samples taken from the entire well-mixed volume of the event should follow a Gaussian distribution. Using computed mean and standard deviation data a pdf was develope d for the untreated influent population and the treated population after 60 minutes of quiescent settling. A summary of the mean and standard deviation are shown in Table 5. Both populations f it Gaussian distributions with coefficients of determination summarized in Table 5. The turbidity pdf results are plotted in Figur e 7 for each event for the untreated influent and settled populations. All pdfs fit a Gaussian distribution (n = 60). The left column illustrates results from flow-limited events and the right column illustrates mass-limited events. The ratios of mean settled and untreated turbidity ( T60/ T0) are different for both classes of events. In masslimited events, ( T60/ T0) ranged from 0.33 to 0.58 (67 to 42 % reduction of m ean turbidity), while in flow-limited events, ( T60/ T0) ranged from 0.58 to 0.78 (42 to 21 % reduction of mean turbidity). These results indicate that for e qual settling times, mass-limited events turbidity is more effectively treated through quiescent settling than for flowlimited events. The ratio also

PAGE 42

42 infers that mass-limited events generally transport a larger mass fraction of large particles, and the relative predominance of thes e particles in mass-limited events suggest that unit operations such as quiescent settling will produce different re sults for different classes of storms for the same surface loading rates. Suspended Sediment Concentration (SSC) The probability density functions (pdf) for SSC were determined from 60 well-mixed samples for each event. As with turbidity, it wa s hypothesized that repres entative samples taken from the entire well-mixed volume of the even t should follow a Gaussian distribution. Using computed mean and standard de viation data a pdf was developed for the untreated influent population and the treated population after 60 minut es of quiescent settling. A summary of the mean and standard deviation are shown in the lowe r half of Table 5. Both untreated and settled populations of SSC sample results fit Gaussian distributions with coefficients of determination summarized in Table 5. Standard deviations for SSC were larger than turbidity, illustrating the potential difficulty with sampling the entire grad ation of particles, in particular the coarser particles that have a mass that is relatively la rge compared to the rest of the gradation. The SSC pdf results are plotted in Figure 8 fo r each event for the untreated influent and settled populations. All pdfs fit a Gaussian distribution (n = 60) The left column illustrates results from flow-limited events and the right column illustrates mass-limited events. The ratios of mean settled and untreated SSC ( T60/ T0) are different for both classes of events. In masslimited events, ( T60/ T0) ranged from 0.14 to 0.27 (86 to 73 % reduction of m ean turbidity), while in flow-limited events, ( T60/ T0) ranged from 0.40 to 0.65 (60 to 35 % reduction of mean turbidity). These results indicate that for e qual settling times, mass-limited events SSC is much more effectively treated through quiescent settling than for flowlimited events. The ratio also infers that mass-limited events generally transport a larger mass fraction of large particles, and

PAGE 43

43 the relative predominance of thes e particles in mass-limited events suggest that unit operations such as quiescent settling will produce different re sults for different classes of storms for the same surface loading rates. In addition, for a gi ven settling time, since mu ch of the SSC mass is associated with the coarser fraction of particle s it would be expected that mass-limited events, which transport larger particles would illustra te greater mass-based performance than flowlimited events, which transport a predominance of finer particles. Conclusions The differentiation of stormwater events into mass-limited and flow-limited will benefit the selection, design, operation, maintenance and performance of unit operations and processes (UOPs) for rainfall-runoff, where the hydrographs for the flow-limited and mass-limited events are significantly different with respect to peak flow rate. The relationships between TSS and turbidity for raw water in flow-limited and masslimited events could be expressed as two linear equations with stat istically significantly different slopes of 1.23 and 2.83. The slopes dropped down to a numerically lower value (0.98) after 60 minutes of quiescent sett ling for either mass-limited or flow-limited events. The particulate matter in mass-limited events had a higher proportion of settleable and sediment mass (around 70 % to 85 % total by mass), as compared to flow-limited events (generally less than 70 %). The particles towards th e larger settleable fractions were settled to a greater extent than the finer particles in su spended fractions as expected. The mass removal efficiency for mass-limited events was from 66 % to 76 % after 60 minutes of quiescent settling, much higher than flow-limited events where it was only around 39 % to 57 %. Event-based ratios of settled and unsettled turbidity or SSC mass ( T60/ T0) were significantly different for mass-limited and flow -limited events, which indicated that mass-

PAGE 44

44 limited events demonstrate higher treatment efficien cies as compared to flow-limited events for the same settling time.

PAGE 45

45Table 2-1. Summary of particulate fractions for 8 rainfall-runoff events with complete runoff volume recovery at Baton Rouge si te Storm Date 05 Mar 04 08 Jul 04 10 Jul 04 06 Apr 04 29 Mar 04 30 Apr 04 1 Jun 04 21 Jun 04 PDH (hr) 5810452162124 10328884 VPV (#/L) 14102022 213Volume (L) 37143877417504716 389310086697Duration T (min) 4038907313 497223Dissolved fraction < 1 m (Total Dissolved solids, TDS) x 215941015569 4012861Dissolved [mg/L] s 11411 111Suspended fraction (1 25 m) x 29621912577159 119113142SS [mg/L] s 4325151017 131943 x 8767443153 143239VSS [mg/L] s 820888 5114 x 34830416376118 109249326Turbidity (NTU) s 228975 4138 x 4863662415 223851TVC [ L/L] s 77892 255Settleable fraction (25 75 m) and Sediment fraction (75 4750 m) x 217207134122259 218399724Settleable solids [mg/L] s 3514122021 122142 x 2971252131 49242118Sediment solids [mg/L] s 16631524 89332Particle size indices for entire gradation d50m (m) [SS] 141318 3027 312019d50m (m) [SSC] N/A 828520691 2117764 PDH: previous dry hours x s: mean and standard deviation of 30 influent replicate runoff samples per event T: duration of runoff SSC: suspended sediment concentration SS: suspended solids TVC: total volume concentration (volume of particles pe r volume of liquid) VSS: volatile suspended solids d50m: mass-based median particle diameter VPV: vehicles per runoff volume

PAGE 46

46 Table 2-2. Summary of solid fractions and hydrogra phs for 8 rainfall-runoff events at I-10 site, Baton Rouge Event based (n = 8) Volume based Solid fraction Index x s x s TDS [mg/L] 95.51.067.0 0.8 Dissolved TDS (g) 142.81.8142.8 1.8TSS [mg/L] 156.223.0129.2 20.6VSS [mg/L] 45.88.836.3 6.9VSS/TSS 0.30.10.3 0.1Turbidity (NTU) 211.79.7169.1 7.1TVC [L/L] 41.05.631.5 4.8Suspended Total TSS (g) 275.443.8275.4 43.8Settleable Settleable [mg/L] 285.022.1315.8 22.9 Settleable (g) 672.948.7672.9 48.7Sediment [mg/L] 80.024.589.3 22.7Sediment Sediment (g) 190.248.45190.2 48.45TS [mg/L] 616.7260.1601.3 67.0Total solids TS (g) 1281.21184.11281.2 1184.1 Event based (n = 8) Hydrologic statistics x s PDH (hr) 121.8 75.9 VPV (#/L) 6.8 7.2Volume (L) 2130.9 1598.8Peak flow (L/min) 323.6 345.1Mean flow (L/min) 37.4 34.1Duration T (min) 49.8 26.6 TDS: total dissolved solids TSS: total suspended solids VSS: volatile suspended solids TVC: total volume concentration PDH: previous dry hours VPV: vehicles per runoff volume TS: dissolved + suspended + settleable + sediment

PAGE 47

47 Table 2-3. Median diameter and size indices for suspended solids in rainfall-runoff as untreated influent (designated as time, t = 0 minute) and after 60 minutes of quiescent settling (designated as time, t = 60 minut es) at Baton Rouge. (Sample number = 60 for each event) d50n d50v d50m 50n 50v 50m 5-Mar-04 1.2613.5614.206.684.30 4.25 6-Apr-04 1.4130.0331.726.563.51 3.458-Jul-04 1.3713.2213.846.594.33 4.2810-Jul-04 1.3218.5019.466.633.99 3.9429-Mar-04 1.2426.7628.246.693.62 3.5730-Apr-04 1.5130.8832.626.503.48 3.421-Jun-04 6.1520.5421.635.093.89 3.8321-Jun-04 5.7418.7319.705.163.98 3.93mean 2.5021.5322.686.243.89 3.83Time = 0 min std. dev 2.136.947.390.690.33 0.335-Mar-04 3.0319.0320.025.803.96 3.916-Apr-04 1.7614.4615.166.344.24 4.198-Jul-04 1.128.268.576.794.80 4.7610-Jul-04 1.4121.0222.136.563.86 3.8129-Mar-04 1.149.069.426.784.70 4.6730-Apr-04 1.4612.9413.546.534.35 4.301-Jun-04 4.9312.6913.285.314.37 4.3221-Jun-04 5.0310.6811.145.294.54 4.50mean 2.4913.5214.166.184.35 4.31Time = 60 min std. dev 1.654.544.830.620.33 0.34 d50n: number-based median particle diameter, m 50n: number-based size d50v: volume-based median pa rticle diameter, m 50v: volume-based size d50m: mass-based median particle diameter, m 50m: mass-based size 50: ) ln(50d d in millimeter

PAGE 48

48 Table 2-4. Summary of statistical characterist ics of mass-based partic le size distribution for suspended solids in rainfall-runoff as untreat ed influent (Time = 0 minute) and after 60 minutes of quiescent settling (Time = 60 minutes) at Baton Rouge. (Sample number = 60 for each event) d50m 50 dg ( m) g I Sk Kg 5-Mar-04 13.56 4.3216.644.061.410.11 1.06 6-Apr-04 30.03 3.5128.323.091.14-0.08 0.858-Jul-04 13.22 4.3312.482.060.78-0.03 1.2410-Jul-04 18.5 4.0118.662.581.020.08 1.2129-Mar-04 26.76 3.6220.853.591.32-0.22 1.1630-Apr-04 30.88 3.4829.572.781.03-0.07 1.051-Jun-04 20.54 3.8920.962.150.750.01 1.0721-Jun-04 18.73 3.9818.462.300.79-0.01 0.96mean 21.53 3.8920.742.831.03-0.03 1.08Time = 0 min std. dev 6.94 0.345.730.710.250.10 0.135-Mar-04 19.03 3.9718.552.180.82-0.02 1.136-Apr-04 14.46 4.2313.292.671.03-0.12 1.128-Jul-04 10.68 4.5510.111.870.630.04 0.8810-Jul-04 8.26 4.768.422.190.78-0.02 1.0129-Mar-04 9.06 4.717.274.171.36-0.15 0.9530-Apr-04 12.94 4.3611.692.520.94-0.11 1.061-Jun-04 11.75 4.4311.212.470.98-0.11 1.2121-Jun-04 12.69 4.3711.011.970.64-0.14 0.81mean 12.36 3.9216.131.990.6-0.58 0.77Time = 60 min std. dev 3.38 4.4211.142.480.87-0.15 0.98 d50m: mass-based median particle diameter Sk: skewness dg: geometric mean diameter KG: kurtosis g: standard deviation of geometric mean diameter I: inclusive graphic standard deviation 50: ) ln(50d d in millimeter

PAGE 49

49 Table 2-5. Mean ( ) Standard deviation ( ) and R-square (r2) of turbidity and SSC in Gaussian distribution (Sample number = 60 for each ev ent) for rainfall-runoff as untreated influent (Time = 0 min) and after 60 minutes of quiescent settling (Time = 60 minutes) at I-10 site, Baton Rouge Turbidity SSC r2 r2 5-Mar-04 478.3043.950.86589.1198.98 0.97 6-Apr-04 131.828.560.74193.389.60 0.908-Jul-04 302.587.470.86339.8015.39 0.9310-Jul-04 163.259.440.95235.1713.02 0.9629-Mar-04 203.6215.370.94590.40105.59 0.8330-Apr-04 109.254.880.97279.6047.20 0.501-Jun-04 250.0516.320.95595.7664.65 0.94Time = 0 min 21-Jun-04 325.889.550.971015.9040.84 0.975-Mar-04 347.2821.540.96295.3743.19 0.706-Apr-04 76.276.700.8576.549.67 0.878-Jul-04 194.908.850.84219.2724.79 0.8710-Jul-04 127.477.660.71125.2014.98 0.9029-Mar-04 118.755.600.95159.2516.97 0.8930-Apr-04 42.692.830.9447.635.07 0.801-Jun-04 103.289.990.96112.6019.01 0.64Time = 60 min 21-Jun-04 108.13 7.83 0.92 141.72 12.91 0.92 SSC: Suspended sediment concentration : Mean r2: Coefficient of determination : Standard deviation

PAGE 50

50 10 cmPVC delivery pipe @ 10% Data logger Drop box Ultrasonic sensor Recirculation system 5 cm ParshallFlume Tee Drainage flow (15 cm PVC pipe) Primary tank V = 4000 L Secondary tank V = 4000 LpumpsTank drain Diameter= 244 cm Watershed at I-10,BTR m2PCC pavement 2% surface slope AADT = 142,00010cmweir 31 cm Trough @ 6% Catchment= 544 m2Catchment= 544 m2 Figure 2-1. Schematic plan view of I-10 site, Baton Rouge, LA.

PAGE 51

51 Flow-limited events Mass-limited events 05 March 04 Flow rate (L/min) 0 20 40 60 80 100 Cumulative Volume ( L ) 0 100 200 300 400 500 29 March 04 Flow rate (L/min) 0 200 400 600 800 1000 Cumulative Volume ( L ) 0 1000 2000 3000 4000 5000 06 April 04 Flow rate (L/min) 0 20 40 60 80 100 Cumulative Volume ( L ) 0 500 1000 1500 2000 30 April 04 Flow rate (L/min) 0 200 400 600 800 1000 Cumulative Volume ( L ) 0 1000 2000 3000 4000 5000 08 July 04 Flow rate (L/min) 0 20 40 60 80 100 020406080 Cumulative Volume ( L ) 0 100 200 300 400 500 01 June 04 Flow rate (L/min) 0 50 100 150 200 250 020406080 Cumulative Volume ( L ) 0 400 800 1200 1600 2000 10 July 04 Time (minutes) 020406080 Flow rate (L/min) 0 20 40 60 80 100 020406080 Cumulative Volume ( L ) 0 200 400 600 800 1000 21 June 04 Time (minutes) 020406080 Flow rate (L/min) 0 200 400 600 800 1000 Cumulative Volume ( L ) 0 2000 4000 6000 8000 10000 Figure 2-2. Hydrographs of 4 flow-limited even ts and 4 mass-limited events with cumulative volume captured at I-10 site.

PAGE 52

52 Flow-limited events Mass-limited events 05-Mar-04 06-Apr-04 08-Jul-04 10-Jul-04 29-Mar-04 30-Apr-04 01-Jun-04 21-Jun-04 Particulate Fraction (%) 0 20 40 60 80 100 55 40 29 47 31 49 31 56 15 53 14 74 50 48 46 49 Suspended (1-25 m) Settleable (25-75 m) Sediment (>75 m) Flow-limited events Mass-limited events 05-Mar-04 06-Apr-04 08-Jul-04 10-Jul-04 29-Mar-04 30-Apr-04 01-Jun-04 21-Jun-04 Particulate Fraction (%) 0 20 40 60 80 100 55 40 29 47 31 49 31 56 15 53 14 74 50 48 46 49 Suspended (1-25 m) Settleable (25-75 m) Sediment (>75 m) Figure 2-3. Solid fraction in rainfall-runoff fo r 8 rainfall-runoff events at Baton Rouge site.

PAGE 53

53 Particle diameter (mm) 10100100010000 % finer by mass 0 20 40 60 80 100 29 March 04 (e = 3.6%) 30 April 04 (e = 1.5%) 01 June 04 (e = 4.4%) 21 June 04 (e = 3.9%) 06 April 04 (e = 2.3%) 08 July 04 (e = 2.6%) 10Jul04 (e = 1.7%) Median Figure 2-4. Particle size distri bution for particulate matter capt ured in the settling basin for rainfall-runoff events at I-10 site, Baton Rouge. ( e: mass balance error)

PAGE 54

54 Sample Basis Time = 0 min n = 480 SSC [mg/L] 0 200 400 600 800 1000 1200 Flow limited events Mass limited events SSC [mg/L] = 2.83 (NTU) R2 = 0.85 SSC [mg/L] = 1.23 (NTU) R2 = 0.87 Sample basis Time = 60 min n = 480 SSC [mg/L] = 0.98 (NTU) R2 = 0.83 Event Basis Time = 0 min n = 480 Turbidity (NTU) 0200400600800 SSC [mg/L] 0 200 400 600 800 1000 1200 Flow limited ( A ) Mass limited ( A ) Flow limited ( B) Mass limited ( B ) SSC [mg/L] = 2.84 (NTU) R2 = 0.92 SSC [mg/L] = 1.23 (NTU) R2=0.96 Replicate A Event Basis Time = 60 min n = 480 Turbidity (NTU) 0200400600800 SSC [mg/L] = 0.98 (NTU) R2= 0.86 Figure 2-5. Differentiation of the constitutive relationship between turbidity and suspended sediment concentration (SSC) for 8 ra infall-events in Baton Rouge, LA.

PAGE 55

55 Flow-limited events Mass-limited events 05 March 2004 Rm(%) = 45.9 Rn(%) = 90.6 Nt(0) = 5.4 TVC [ L/L 0 30 60 90 120 150 29 March 2004 Rm(%) = 65.5 Rn(%) = 0.7 Nt(0) = 1.8 Cumulative TVC [ L/L] 0 300 600 900 1200 06 April 2004 Rm(%) = 39.2 Rn(%) = 0.1 Nt(0) = 0.4 TVC [ L/L] 0 30 60 90 120 150 30 April 2004 Rm(%) = 75.6 Rn(%) = 20.5 Nt(0) = 0.4 Cumulative TVC [ L/L] 0 300 600 900 1200 08 July 2004 Rm(%) = 39.5 Rn(%) = 0.4 Nt(0) = 2.5 TVC [ L/L] 0 30 60 90 120 150 01 June 2004 Rm(%) = 72.1 Rn(%) = 5.0 Nt(0) = 0.1 Cumulative TVC [ L/L] 0 300 600 900 1200 10 July 2004 Rm(%) = 56.7 Rn(%) = 69.1 Nt(0) = 1.5 110100 TVC [ L/L] 0 30 60 90 120 150 Paticle diameter ( m) 21 June 2004 Rm(%) = 66.0 Rn(%) = 11.2 Nt(0) = 0.1 110100 Cumulative TVC [ L/L] 1200 900 600 300 0 Particle diameter ( m)TVC ( t = 0 min) Cumulative TVC ( t = 0 min) TVC ( t = 60 min) Cumulative TVC ( t = 60 min) Rm: mass removalRn: number removal Nt(0): particle number density (1010/mL) (t = 0 min)Figure 2-6. Total Volume Con centration (TVC) for 4 flow-limited and 4 mass-limited events at I-10 site, Baton Rouge, LA.

PAGE 56

56 Flow-limited events Mass-limited events 05 March 04 % finer by mass 0 20 40 60 80 100 29 March 04 Settling Efficiency (%) 0 20 40 60 80 100 06 April 04 % finer by mass 0 20 40 60 80 100 30 April 04 Settling Efficiency (%) 0 20 40 60 80 100 08 July 04 % finer by mass 0 20 40 60 80 100 01 June 04 Settling Efficiency (%) 0 20 40 60 80 100 10 July 04 Particle Diameter ( mm ) 110100 % finer by mass 0 20 40 60 80 100 21 June 04 Particle Diameter ( mm ) 110100 Settling Efficiency (%) 0 20 40 60 80 100 Time = 0 minute Time = 60 minutes Settling Efficiency Figure 2-7. Cumulative Volume Concentrati on of unsettled (time = 0 min) and after 60 minutes settling (time = 60 min) rainfall -runoff and settling efficiency based on mass for 8 events at the Baton Rouge site.

PAGE 57

57 Flow-limited events Mass-limited events 05 March 2004 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 478.3 s = 43.95 R2 = 0.86 Time = 60 min = 347.3 s = 21.54 R2 = 0.96 06 April 2004 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 131.8 s = 8.56 R2 = 0.74 Time = 60 min = 76.3 s = 6.70 R2 = 0.85 08 July 2004 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 302.6 s = 7.47 R2 = 0.86 Time = 60 min = 194.9 s = 8.85 R2 = 0.84 10 July 2004 Turbidity, NTU 0100200300400500600 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 163.2 s = 9.44 R2 = 0.95 Time = 60 min = 127.5 s = 7.66 R2 = 0.71 29 March 2004 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 203.6 s = 15.37 R2 = 0.94 Time = 60 min = 118.8 s = 5.60 R2 = 0.95 30 April 2004 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 109.2 s = 4.88 R2 = 0.97 Time = 60 min = 42.7 s = 2.83 R2 = 0.94 01 June 2004 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 250.1 s = 16.32 R2 = 0.95 Time = 60 min = 103.3 s = 9.99 R2 = 0.96 21 June 2004 Turbidity, NTU 0100200300400500600 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 325.9 s = 9.55 R2 = 0.97 Time = 60 min = 108.1 s = 7.82 R2 = 0.92 Figure 2-8. Probability density function (pdf) of turbidity for rainfall-runoff as untreated influent (time = 0 minute) and after 60 minutes of quiescen t settling (time = 60 minutes). All pdfs fit a Gaussi an distribution (n = 60). 55 160 0T T 42 260 0T T 01 360 0T T 28 160 0T T 73 160 0T T 56 260 0T T 38 160 0T T 71 160 0T T

PAGE 58

58 Flow-limited events Mass-limited events 05 March 2004 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 589.1 s = 99.0 R2 = 0.97 Time = 60 min = 295.4 s = 43.19 R2 = 0.70 06 April 2004 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 193.4 s = 9.6 R2 = 0.90 Time = 60 min = 76.5 s = 9.67 R2 = 0.87 08 July 2004 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 339.8 s = 15.39 R2 = 0.93 Time = 60 min = 219.3 s = 24.79 R2 = 0.87 10 July 2004 SSC [mg/L] 020040060080010001200 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 235.2 s = 13.02 R2 = 0.96 Time = 60 min = 125.2 s = 14.98 R2 = 0.90 29 March 2004 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 590.4 s = 105.6 R2 = 0.83 Time = 60 min = 159.2 s = 16.97 R2 = 0.89 30 April 2004 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 279.6 s = 47.20 R2 = 0.50 Time = 60 min = 47.6 s = 5.07 R2 = 0.80 01 June 2004 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 595.8 s = 64.65 R2 = 0.94 Time = 60 min = 112.6 s = 19.01 R2 = 0.64 21 June 2004 SSC [mg/L] 020040060080010001200 pdf 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 min = 1015.9 s = 40.83 R2 = 0.97 Time = 60 min = 141.7 s = 12.90 R2 = 0.92 Figure 2-9. Probability density function (pdf) of SSC for rainfa ll-runoff as untreated influent (time = 0 minute) and after 60 minutes of quiescent settling (time = 60 minutes). All pdfs fit a Gaussian distribution (n = 60). 17 760 0T T 88 160 0T T 29 560 0T T 55 160 0T T 87 560 0T T 53 260 0T T 71 360 0T T 99 160 0T T

PAGE 59

59 CHAPTER 3 EQUILIBRIUM PARTITIONING OF METALS PHOSPHORUS AND COD IN URBAN RAINFALL-RUNOFF INFLUENCED BY SEDIMENTATION Introduction Rainfall-runoff from urban highway surfaces co ntains and transports significant loads of pollutants in a complex heterogeneous mixture which includes particulate materials, metals, phosphorus and organic compounds. From urban in terstate highway pavement alone, annual heavy metal, total suspended solids (TSS), chemical oxygen demand (COD) loadings have been shown to equal or exceed annual loadings from un treated domestic wastewater for a given urban area (Sansalone et al. 1998). Constituents on urban highway surfaces are ge nerated mainly from traffic activities, vehicular component wear, fluid l eakage, pavement paint and eros ion, roadway maintenance, as well as atmospheric deposition (pr ecipitation and dustfall) (Shinya et al. 2000). Vehicles are the major source of heavy metal deposition on hi ghway surfaces (Bourcier and Hinden 1979). The most-often cited heavy metal contaminants are Cd Cr, Cu, Fe, Ni, Pb, and Zn. These metals can adversely impact receiving waters by increasing toxicity in the water column and/or sediments, and bioaccumulation in the food chain (Yousef et al. 1987). For urban storm water, Zn, Cd, Cu and Pb are commonly examined for comparison to discharge regulations. Roadway runoff levels of Zn, Cu, Cd, Pb, Cr and Ni are significantl y above ambient background levels, and for some heavily traveled roadways, Zn, Cu, Pb and Cd often exceed USEPA and State EPA surface water discharge criteria on an event basis (Sansalone and Buchberg er 1997). Depending on factors such as pH, redox conditions, alkalinity, residenc e time, complex agents and suspended solids, these heavy metals can preferenti ally dissolve in ionic form an d exert an immediate toxicity impact (Sansalone and Cristina 2004).

PAGE 60

60 Urban runoff contains high phosphorus con centrations (USEPA 1993) that may be accelerating the eutrophication in most cases. Ne urological damage in people exposed to the highly toxic volatile chemical produced by dinoflagellate has dr amatically increased public awareness of eutrophication and the need for so lutions. Waschbusch et al. (1999) found that Lawns and streets are the largest sources of to tal and dissolved phosphorus and their combined contribution was approximately 80 %. Other sources including fertilizers, leaves, twigs, and other organic debris are also accounted for the total phosphorus mass in the runoff. Approximately 25 % of the total phosphorus mass in each size fraction attri butes to leaves (Ray 1997). To reduce the impact of contam inants in rainfall-runoff to ambient environs, different type of treatment units are selected, designed, set up and operated in situ area. The performance of unit operations and processes (UOPs) may be infl uenced by the stochastic runoff loadings, which could generate various partitioning and distributi on of constituents, such as metals, phosphorus, and COD between dissolved and particulate phase as well as among the particle fractions. The partitioning of constituents is difficult to simulate due to the large variation in different events, causing the control of these cons tituents in rainfall-runoff sti ll challenging (Sansalone and Buchberger 1997). However, rainfall-runoff events can be differentiated into two limiting types of behavior: mass-limited and flow-limited based events, based on hydrologic characteristics or examined with respect to pollutant loading ch aracteristics and phases (S ansalone et al. 1998, Cristina and Sansalone 2003, Sansalone and Cristina 2004). The nature and size distribution of particles in the rainfall-runoff plays an important role in determining their transport and fate, as well as that of any associated contaminants including metals and phosphorus (Sansalone et al. 2005). The larger solids in the runoff stream may settle

PAGE 61

61 into storm sewers, urban waterways, and so on, where the particles, metals and phosphorus can be mobilized by subsequent storms. Smaller, coll oidal particles may remain in suspension and be transported much greater dist ances (Characklis and Wiesner 1997). Previous work focused directly on the urban runoff and reported that ap proximately 50 % of total metals by mass in storm water were associated with particles (Hunter et al. 1981). The transport of metallic pollutant through water system is due to adsorptio n of pollutant onto suspended particles in the water. A strong correlation exists between metals and suspended solids in runoff released from highway surfaces during storms (Sansalone et al. 1995). Storm events can flush the particles from their site of generation a nd disperse them through the wate rshed, potentially contaminating vast areas (Magnuson et al. 2001 ). Results of cumulative hea vy metal mass trends across the particle size gradations for each site indicate that the predom inance of heavy metal mass is associated with the coarse size particle fraction (>250 m) (Turer et al. 2001). However, from both a hydrological and geochemical viewpoi nt, coarse sediment larger than 2000 m is generally considered to be less important in the transport of adsorbed metals from urban areas (Sartor and Boyd 1972). Sartor and Gaboury (1984), Hvitved-Jacobsen and Yousef (1991), and Hvitved-Jacobsen et al. (1994) investigated road runoff polluta nt characteristics and found 60 % to 80 % of phosphorus in road runoff to be associated with particulates. The bulk of the phosphorus load results from the greater partic le-size fractions (Ray 1997), an d the largest amount of total phosphorus was found in the > 250 m size fraction (nearly 50 %). Ho wever, other studies have shown that large phosphorus concentrations corr espond with small particle sizes because of the high surface area to mass ratio for sm all particles (Sartor and Boyd 1972).

PAGE 62

62 The wide variation of parti tioning and distribution data in previous studies may lead improper design and selection of UOPs for the in-s itu runoff, causing low removal efficiencies of constituents and remaining the potential impacts to ambient environs. Therefore, it is critical to investigate and simulate the partitioning and dist ribution of these contamin ants in rainfall-runoff properly accurately, whic h contributes the main purpose of this study. Objective There are three objectives in this study. The fi rst objective is to dete rmine the fractions of phosphorus and COD associated with the suspended, settleable, sediment solids in rainfall-runoff for 8 events at I-10 site, Bat on Rouge, LA from March to July 2004 where the entire volume of flow was captured. The second objective is to examine equilibrium partitioning of phosphorus, COD and metals between dissolved and particulate phases in rainfall-runoff before and after 60 minutes of quiescent settling. The last objective is to investigate the role of quiescent settling on separation of metals, COD and phosphorus. Background Study Phosphorus reaching the receiving waters from ur ban rainfall runoff can be in a variety of organic and inorganic forms, de pending on the pH condition and redox potential encountered in the runoff and the characteristics of particles cont ained in the runoff. The most frequent soluble forms of phosphorus in the runoff are ort hophosphate, represented by phosphoric acid (H3PO4) and its dissociation products, 4 2PO H, 2 4HPO, 3 4PO, and ion complexes of these ions with some of the other constituents of runoff. Se veral of other classes of phosphorus-containing compounds in the runoff are condensed phospha tes and organic phosphates. The condensed phosphates are formed by the condensation of tw o or more orthophosphates groups and have the characteristic P-O-P linkage, while polyphospha tes are linear molecules and metaphosphates are

PAGE 63

63 cyclic. The organic phosphates vary in many types, including sugar phosphates, nucleotise, phospholipids etc. which might origin from the leaves, lawn and other sources. Presumably, any compound when exposed to th e combined action of air and moisture over a sufficiently long length of time will be transf ormed into orthophosphates. The distribution of the phosphate ions in terms of concentrations could be calculated from the mass-balance equation below, [P]T = [PO4 3-]T + [HPO4 2-]T + [H2PO4 -]T + [H3PO4]T Since the orthoand polyphosphate anion are multiply charged negative species they react readily with multiply charged positive ions to give precipitate. The sorption is often related to the precipitation processes, in that electrostatic forc es between the negatively charged phosphate and a positively charged surface are often involved. As indicated by the Ostwald-Gay-Lussac step rule (according to which an unstable form is us ually produced before the stable form appears), phosphates tend to precipitate first as an amor phous mass of variable composition a mass of material which slowly converts to one or more stable crystall ine species. Both orthoand polyphosphates also form soluble complex with metal ions. The orthophosphates complexes are very weak for the alkali and alkaline earth me tals and only become im portant for transition metals such as iron. But the polyphosphates form relative strong complexes with all of them. In summary, it is probably safe to say that the em phasis in the chemistry of phosphorus in urban runoff will be on the orthophosphate ion and its so rption on the surface of particles, colloidal, suspended or making up settable and sediment. The behavior of metals in urban rainfall r unoff is also a topic of ongoing interest and concern. One key characteristic of metals in aqua tic systems is their ability to exist as either dissolved or solid species and to combine with numerous other species The ionic charge on a

PAGE 64

64 metal is called its oxidation number or oxidation state. Many metals ions can exist in stable forms in two or more oxidation states (e.g., Fe2+ or Fe3+, Cr3+ or Cr6+). The significance of different charges deals with oxidation-reduc tion reactions. Many meta l ions can act as multiprotic acids, releasing protons from water mol ecules in the inner hydration sphere until they are surrounded by OHinstead of H2O molecules. As the concentration of metal and hydroxyl ions in a solution in creases, some of the metals start forming larger complexes. In addition to that, metals can form solid with a variety of ligands. For instance, many metals can form so luble complexes and/or solids with hydroxide, carbonate, phosphate, sulfide ions. The metal par titioning between the dissolved and particulate bound fractions in storm water is a dynamic process. Whether in pavement runoff, urban storm water or any aqueous system, there is a tempor al partitioning between heavy metals in solution and solids whether these solids are in suspension (TSS, VSS) or as settleable solids that may be part of a fixed or mobile bed load. This part itioning includes specific ma ss transfer mechanisms of sorption, ion exchange and surface complexa tion with both organic an d inorganic sites on the solid matter. Partitioning reactions are generally non-linearly reversible between the solid-phase and soluble-phase concentrations (Turer et al. 2001). For the stor m water runoff with pH average at 7.5, the soluble inorganic phosphate could be regarded as being present exclusively as H2PO4 and HPO4 2, with negligible amounts of PO4 3 and H3PO4. Under the certain range of chemical propertie s of runoff, phosphorus is typically combined with compounds (such as Ca, Fe, Al and Mg) asso ciated with the partic le, making it unavailable to many plants. Phosphorus adsorbed onto the surf ace of solid particles or associated with any organic matter will act as reserves of phosphorus and will become available over time. For urban rainfall runoff, these particles will be carried down to the receiving water, slowly releasing

PAGE 65

65 phosphorus until they eventually settle. The representative heterogeneous equilibria of phosphates are shown as below: CaHPO4(S) = Ca2+ + HPO4 2pKSO = +6.66 Ca(H2PO4)2(S) = Ca2+ + 2H2PO4 2pKSO = +1.14 -Ca3(PO4)2(S) = 3Ca2+ + 2PO4 3pKSO = +24.0 FePO4(S) = Fe3+ + PO4 3pKSO = +21.9 AlPO4(S) = Al3+ + PO4 3pKSO = +21.0 The dynamic equilibrium between particulate phosphates and dissolved phosphates can be applied to predict solution ions in contact with a solid under th e specific temperature, pH and redox condition. Consider the equilibrium expressi on for dissolution of a metal (Me) ligand (L) solid above and the corresponding equilibrium constant expression, as follows: Mem(L)n(s) = mMen+ + nLmKs0 = { Men+}m{ L-}n / { Me(L)n(s)} This partitioning, which vari es throughout a rainfall runoff event is a function of pH, alkalinity, residence time and so lids characteristics, each of wh ich vary significantly between hydrologic events and traffic patterns (Sansalone and Buchberger 1997). For any component in the runoff (P, Metal, COD), the total concentr ation here is the su m of the dissolved (cd) and the particulate-bound c oncentrations (cp): cT = cd + cp Operationally, the soluble or dissolved fr action is that fracti on that passes the 0.45 m membrane filter and therefore contains both th e dissolved and part of the colloidal-bound compounds. The solid phas e concentration, cp is defined as the product of the chemical

PAGE 66

66 concentration on the solid phase, cs in terms of mass/mass of solid s and the concentration of the adsorbing solid material in the aqueous system m typically measured as TSS in terms of mass/volume of aqueous solution: m c cs p Under equilibrium conditions, when the rate of sorption and de-s orption are equal, concentration equilibrium exists between the di ssolved and solid-phase concentrations of a chemical (Sansalone et al. 1998). The ratio of these phases at equilibrium is referred to as the partitioning coefficient, Kd for a particular compound at a particular pH and redox level: Kd = cp/cd The particulate-bound fraction (fp) is defined as: ) 1 ( 1 ) /(m K c c P D D fd T d d where, D is the dissolved mass of a compound ( g/L) and P is the particulate-bound mass of a compound ( g/Kg). For fd > 0.5, the compound mass is mainly in dissolved form. The product of m Kd is dimensionless and Kd is usually expressed as lite rs per kilogram (L/kg). The larger the Kd value, the greater the partitioning to th e solid phase. Heavy metals in pavement runoff have Kd values that range from 102 to over 106 (Sansalone and Buchberger 1997). It has been shown empirically that Fe3+ is reduced to the Fe2+ form in the 100 to 200 mV range of measured redox potential in a variet y of soils (Turner and Patrick 1968). The reduction of Fe3+ to Fe2+ has been demonstrated as the mechanism for the release of phosphorus and reduction in the phosphorus sorpti on capacity of the soil (Sherwood and Qalls 2001). It has also been reported for heavy metals (Yeoman et al. 1989; Yeoman et al. 1992) that ratios of contaminants to particle mass were not increased in the small size fractio ns. Therefore, from a practical point of view, it is important to re move particle mass efficiently throughout the whole

PAGE 67

67 particle size range. Particle-size analysis has s hown that, by number, the particles have a median size of < 1 m but are settling within 24 h under laboratory conditions. By volume, the particle sizes have a median around 100 m. This relatively short settling time is contrary to the expected behavior of particles this size, but may be due to the presence of metals in solution. Under controlled conditions the particles > 0.5 m have settled out of the runoff sample within 24 h, without any chemical addition. It is hypothesized that interact ions with the dissolved heavy metals may assist in the floccu lation and sedimentation of roadwa y particulates (Drapper et al. 2000). The study found the dry street surface material as containing significant concentrations of metals, and reported that the smallest size class analyzed (< 43 m), while accounting for only a small percentage of the total solid mass, contained over 50 % of the total metals (Characklis and Wiesner 1997). Studies of sediments on an urban ri ver in New Jersey yielded the similar results, finding a disproportionate fraction of total metals associated with smallest sediment particles (Wilber and Hunter 1979). Tanizaki et al. (1992) analyzed the meta ls concentration in individual size fractions of samples taken from an urban st ream under dry weather condition. Zn existing in the dissolved fraction ranged from 44 % to 83 % of total Zn. Considerable variability was observed in the partitioning of individual metals Concentration of iron and the macro-colloidal particles (0.45-20 m) were also highly correlated th roughout the duration of both storms (Characklis and Wiesner 1997). Cd is present in both the dissolved and colloid al fractions; Cr is soluble, mobile, and can be stable for long periods in wate rs with low levels of organic ma tter; it can also form stable complexes, especially chromium hydroxide; Cu is typically associated with dissolved solids and colloidal material; Fe is primarily associated with suspended solids. Pb is mostly associated with suspended solids or as a carbonate precipitate; Unlike other metals, Zn is mostly associated with

PAGE 68

68 the dissolved fractions, although it can also ad sorb to suspended sediment and colloidal particles(Makepease et al. 1995). For Zn, Pb and Cu the concentrations generally increase with increasing specific surfaces area (SSA) or decr easing particle size. Metals are primarily associated with the finer particle sizes, less than 150 m. This observation suggests that sedimentation basins, and perhaps even filtration systems, may be ineffective at removing metals from highway runoff. Silts and colloids, whic h have particle sizes ranging from 1 to 50 m and 0.001 to 1 m, respectively, have high sorption cap acities for heavy metals and non-polar organics (Sansalone and Buchbe rger 1997; Krein and Schorer 2000). Native phosphorus has been shown to be releas ed as soils are reduced under anaerobic conditions after flooding (Sanyal and Dedatta 199 1; Krairapanond et al. 1993). The smaller the particle size in soil b ecomes, the more total phosphorus (TP) is found (Sa et al. 2003). Kann (1998) postulated that under hi gh pH, hydroxide ions compete with phosphate ions for sorption sites on iron hydroxides, thereby releasing phosphorus into the wa ter column. Thus, a wind event strong enough to resuspend sediments has the potential to enhance mechanisms for mobilizing phosphorus in the bottom sediments by virtue of both the increased sediment surface for phosphorus desorption processes and the large mass of phosphorus that bottom sediments can carry into the water co lumn. The results of the Resuspende d Sediment Study showed that even when diffusion limitations on the pr ocesses responsible for the tr ansfer of phosphorus across the sediment-water interface were eliminated, the phosphorus in the sediments did not rapidly repartition into the aqueous phase. This was tr ue under all the conditions tested: ambient and photo-synthetically elevated pH, very low sol uble reactive phosphorus (SRP) concentration in the aqueous phase, and elevated pH accompanied by a large increase in alkalinity (Fisher and Wood 2001).

PAGE 69

69 Polyphosphate-metabolizing bacteria store polyphosphates under aer obic conditions when phosphate is available in excess of immediate requirements for growth. These bacteria use the stored polyphosphate as an en ergy source when conditions turn anaerobic (Gachter et al. 1988; Davelaar 1993; Khoshmanesh et al. 2002), which releases phospha te to the surrounding medium. Fe3+ phosphates or Fe3+ complexes, which absorb phos phorus, most obviously play an important role in the cycling of phosphorus in the environment (Spi vakov et al. 1999). Their data indicate that almost half of the heavy metals found on street sediments are associated with particles of 60 to 200 m in size, and 75 % are associated with particles finer than 500 m in size. Dempsey et al. (1993) showed that the hi ghest recorded concentrations of Cu, Zn, and phosphorus were associated with sand particles between 74 and 250 m in size. Results from the dry sieving analysis indicate that almost all the TP in the dry surface pollutant is attached to particles between 53 and 300 m. Results from the wet sieving analysis show that about 25 % of the total TP is dissolved less than 0.45 m, little TP between 0.45 and 11 m, about 10 % between 11 and 53 m, about 30 % between 53 and 150 m, and more than 50 % between 53 and 300 m. As expected, more TP is associated with the finer particle sizes in the wet sieving compared to the dry sieving because of the disso lution of the larger par ticles. The dry sieving analysis provides an indication of the availabil ity of TP in the surface pollutant while the wet sieving analysis gives an indica tion of the potential TP that can be washed off if there is sufficient water to dissolve the pollutants. This has important implications because an effective reduction in sediment or total suspended solid load does not necessarily equate to a similar reduction in nutrient loads. 60 % of TP in the storm water samples is attached to sediments between 11 and 150 m, and 40 to 50 % between 11 and 53 m. TP in solution contributs about 25 % of the total TP. Practically all the particul ate TP in storm water samples are attached to

PAGE 70

70 sediments between 11 and 150 m. This suggests that to eff ectively remove TP, pollutant treatment facilities must be able to remove pollutants down to 11 m (Vaze and Chiew 2004) due to the strong association of suspended solids with waste water COD or phosphorus (Levine et al. 1985). Furthermore, th e biological degradation rate in terms of COD reduction is influenced by particle size di stribution (Levine et al 1991). Fractionated filtration of wastewater samples revealed that consider able amounts of COD and phosphorus were particulate. 91 % of the raw sewage COD, 46 to 71 % of the primar y effluents COD, and 14 to 46% of the final effluents COD were attributed to suspended solids. In case of phosphorus, the particle-associated fraction was 79 % in raw sewage, 35 to 67 % in the primary effluents, and 54 to 81 % in the final effluents. The distribution of phosphates and metals in a water column of a sedimentation basin is generally not homogeneous. The dissolved ph osphates, of course, tend to reach uniform concentration throughout the column but this tendency is counte racted by various particulate phosphate compounds. Aluminum and iron phosphates we re associated with the finer particles carried by runoff movement before settling; calcium phosphate was associated with the coarse particles and so was deposited in the sediment. Th e formation of these lo w-soluble phosphates is used for chemical removal of phosphates from wate r. In practice, this causes problems in water treatment by coagulation when already rather low concentrations of polyphosphates can cause improper agglomeration of colloidal pa rticles into settleable flocs. Methodology Experimental System The experimental site located at Interstate 10 over City Park Lake in urban Baton Rouge, Louisiana. The watershed is 1088 m2 area of Portland cement concrete (PCC) pavement with 2 % surface slope to each catch basin and expans ion joint located on th e east end of each catchment. The watershed is composed of two identical ca tchments, each 544 m2. Mean annual

PAGE 71

71 precipitation at the site is 1460 mm/year and the annual average daily traffic (AADT) load is 142,000 considering both eastbound and westbound vehicles. The site drainage is designed as a gravitati onal collection system. Rainfall-runoff generated from pavement surface during a storm event is collected by two PVC troughs with 6 % slope. These troughs drain into a 10 mm PVC delivery pi pe, leading to a 5 mm Parshall flume. The ultrasonic sensor located at 2/3A measured real-time depths that recorded by a Sigma 950 data logger. Rainfall-runoff discharge from Parshall flume can either be directly into hydrodynamic separator for pretreatment or directly bypasse d into a series of 4000 L settling tanks, where internal pumps were utilized for internal and external mixing (re-circulation). The two internal mixing pumps kept the tanks well mixed during sampling time and the external mixing pump pumped runoff back to the entrance of the Parshall flume for sampling. During the sampling period from March 5th 2004 to July 10th 2004, rainfall-runoff was captured and saved by the in-situ collection system with entire volume from 4 flow-limited and 4 mass-limited events respectively. All the hydrologic and loading parameters for each event were measured and shown in Table 1, including previo us dry hours, traffic loading, rainfall volume, duration time. 30 homogeneous runoff samples for each event were taken from drop box and each sample set was collected in 3 separate cont ainers for different use, 500 mL, 1 L and 12 L. During each storm event, average 2-minute traffi c load on either bound of bridge was recorded every 15 minutes and temperature of runoff at different time was another real-time measurement was also recorded. The runoff start time, end time and primary tank storage depth are 3 critical parameters to calculate total rainfall-runoff vol ume and examine mass balance in the tank. In addition, there is a data-logging ra in gage located 100 m from the watershed and used to measure and collect rainfall data in increments of 0.25 mm.

PAGE 72

72 Phase Fractionation By convention, filtration using a 0.45 m membrane is typically used to define the operationally dissolved and particulate fractions of materials in storm water or wastewater. However, it is recognized that this operati onally defined dissolved fraction can include colloidal particles (Bull and Williamson, 2001). A pr essure filtration method was applied in this study to fractionate rainfall-runoff samples into di ssolved and particulate phase for further metal and phosphorus analysis. The samples were filtered at time of collection using a stainless steel pressure filter, containing a filter support of stainless steel, and a supported cellulose acetate membrane filter with 0.45 m pore diameter. For later analysis of the particulate fraction, the filtration volume was measured. Samples were filt ered at a pressure of 150 to 350 KPa to avoid the filter membrane broken. For each filtered sa mple (dissolved part), 25 mL was used for dissolved phosphorus measurement, 2 mL for di ssolved COD measurement and rest sample was acidified by adding 5-mL 5 % trace metal nitric acid for ICP-MS analysis of metals. The filter with residual particles was folded carefully in an aluminum pan by using a tweezers to keep the particles from unnecessary contamina tion before chemical analysis. Particulate Fractionation Samples in this study are well mixed before th ey were poured into th e Imhoff cones for the 60-minutes quiescent settling. 1 L sample is placed in an lmhoff cone and allowed to settle for 45 minutes. Then while gently agitating the sample near the sides of the cone, it is allowed to settle 15 minutes longer. The volum e of solid portion settled on the bottom of cone is recorded to estimate the amount of settable and sediment solid. Under specific condition, part of the rest solids suspended in the water column will finally settle after 60 minutes. According to the different gravitational settling velo city of non-colloidal particles, total suspended solid (TSS) and total settable and sediment solid are separated based on the 60 minutes settling time. In addition

PAGE 73

73 to that, the large floating part icles or submerged agglomerates of non-homogeneous materials are allowed to include in the desi red suspended fraction result. The sediment fraction of SSC is referred as the solid portion retained on the mechanic sieve with size 75 m (#200). SSC solid obtained from quiescent settling a nd mechanic sieving were air-dried in the hot room with a relatively constant temperature around 40oC. Acid Digestion After mechanical sieve analysis and particle size separation, 0.5 g particles of a given size range are transferred into the labeled clean gla ss beakers. For each par ticle size digested and analyzed, control (blanks) s hould be carried out. For all sa mples, blank and standards (PriorityPollutnTTM/CLP inorganic solids, Environmenta l Resource Associates) digestion involved adding 9 mL of concentrated HNO3 and 3 mL of concentrated HCl. The sample is covered with a ribbed watch glass or other suita ble covers and heated on a hot plate or other heating source at 150 oC. The samples are taken to a slow boil for 90 minutes and then an additional 30 minutes at in creased temperature at 175 oC so that a greater reflux action access. The digestion beakers and then allowed to cool do wn. Dilute the digested samples with 10 mL of DI water. After digestion, the silicates and othe r insoluble material that could clog the nebulizer in ICP-MS analysis were removed from the digested solution by filtration where the digested sample was transferred to a funne l with glass-fiber filter into a labeled volumetric flask. The filter was rinsed with 2 % nitric acid solution at least 3 times until all the insoluble material was completely transferred to the filter, and the overa ll volume of rinsed solu tion was finally adjusted to 100 mL. 10 mL well-mixed solution was taken out and added into an ICPMS vial for required metal concentration analysis.

PAGE 74

74 Metal Inductively Coupled Plasma-Ma ss Spectrometry (ICP-MS) Analysis Inductively Coupled Plasma-Mass Spectro metry (ICP-MS) system is capable to simultaneously determine up to 80 metal species in a single run, with extend dynamic detector range over a full 8 orders of magnitude and detec tion limits down to ppt level. 8 metals were analyzed in this study, which include Cr, Cu, Zn, Pb, Cd, As, Mn, and Fe. Raw rainfall-runoff samples cant go into ICP-MS system directly and must to be digested in a mixture of concentrated nitric and hydrochloride acid to release metal species into dissolved phase. 100 L internal standard was added into the digested so lution for calibration of metal concentrations in each single sample. A four-point st andard calibration (10, 20, 100, 500 g/L) and a blank (5 % nitric acid) were applie d to calibrate each measured metal sp ecie before analysis of samples. Analytical controls, including an alysis of control samples (QC 50), blank samples, mass balance checks, and standard solution quantity checks fo r every 9 samples, were taken throughout the process. Digestion And Analysis For Phosphorus Phosphorus acid digestion followed the Persul fate Digestion Method (Standard Methods, 1995). 1 mL concentrated sulfate acid and 10 mL 1: 10 persulfate were added into 50 ml influent rainfall-runoff sample and refluxed ge ntly on a preheated hot plate at 300 0C for about 90 minutes until the sample volume remaining in the beaker was around 10 mL. Add 1 mL sodium hydroxide in the beaker and dilute sample back to 50 mL with D.I. water. Phosphorus was measured by HACH DR-2000 Spectrophotometer us ing PhosVer 3 Ascorbic Acid Method for both dissolved and total phosphorus. Chemical Oxygen Demand (COD) Analysis For each rainfall-runoff sample replicate of total COD, 1 mL of influent sample was transferred into a COD digestion reagent vial and diluted with 1 mL D.I. water. COD vials with

PAGE 75

75 sample were then placed in COD reactors and heated at 150 oC. For each runoff sample replicate of the dissolved fraction of COD, 2 mL homogeni zed sample was taken directly from the sample filtrate and put in the vials. After mixing the samp le solution with reagents in the vial, the vials were placed in the preheated reactors and heat ed for two hours. All analyses (samples and blanks) were run from the same lot of vials sinc e each lot has a separate calibration curve. After two hours of reaction, vials were cooled to room temperature. For selected samples, such as samples taken during mass-limited events, dilutions were required, a colorimetric procedure for the 0-1500 mg/L COD vial range with a calibration was used to determine the chemical oxygen demand concentration expressed as mg/L. Results Phosphorus Partition In Dissolved And Particulate Phase Equilibrium of phosphorus partit ioning in dissolved and particulate phase of rainfall-runoff before (t0) and after 60 minutes of quiescent settling (t60) was investigated and results were illustrated in Figure 1. Phosphorus concentration in particulate phase of influent runoff was in the range of 1 to 3 mg/L, while in dissolved phase which was only 0.1 to 0.3 mg/L, indicating that phosphorus most likely part itioned in particulate phase rather than dissolved phase for both type events. Considering the settling effects, it was no signi ficant difference of phosphorus concentration in dissolved phase before and after 60 minutes settling, but for particulate phosphorus, the concentration decrea sed significantly after quies cent settling, especially for mass-limited events. The reason for the decrease is that phosphorus adsorb ed on the settleable solids can be easily separated from the aqueous pha se through the settling process, compared to the dissolved phosphorus. Dissolved fraction, fd values shown in Figure 3 were less then 0.06, and equilibrium coefficient, Kd values of phosphorus were mostly larger than 50 L/Kg for both t0 and t60 samples

PAGE 76

76 in each specific event, which also indicated that only little amount of phosphorus existed in the dissolved phase and most phosphorus partitioned in the particulate phase. After 60 minutes of quiescent settling, settleable and sediment so lids were removed from aqueous phase and phosphorus re-distributed between dissolved and pa rticulate phase in the runoff, causing lower fd values and larger Kd values for t60 runoff samples, compared with t0 samples. Chemical Oxygen Demand Partition In Dissolved And Particulate Phase COD concentration of influent runoff sample s was in the range of 50 to 200 mg/L for 8 storm events, which was lower than that in municipal waste water. Equilibrium partitioning of COD between dissolved and particulate phase in Figure 2 showed that dissolved COD was mostly equal or larger than particulate COD, wh ich indicated that settli ng process might not be efficient to remove COD from runoff due to its high partitioning in the dissolved phase. After 60 minutes settling, COD concentration in dissolved phase varied by a tiny amount, but a significant decreasing was found in particul ate phase for most events. fd values and equilibrium coefficient, Kd values of COD in rainfall-runoff before and after 60 minutes settling was shown in Figur e 4. It illustrated that most fd values were larger than 0.6, which means that COD partitioned in dissolved pha se was greater than in particulate phase. Kd values of COD were less than 10 L/Kg in all 8 ev ents, which also indicated a high partitioning of COD in dissolved phase rather than particulate phase, comparing with phosphorus in the runoff. fd and Kd values after 60 minutes settling were mostly larger than that of influent samples, because of the removal of particulate COD with settleable and sediment solids during the settling process, making an increase of COD partiti oning in dissolved phase of rainfall-runoff. Phosphorus And COD Partitioning In Suspe nded, Settleable And Sediment Solids Figure 5 gave out the phosphorus factions in suspended, settle able and sediment solids of influent rainfall-runoff. Results illustrated that high proportion of phosphor us distributes to the

PAGE 77

77 suspended fraction, especially fo r flow-limited events, which contributed more than 50 % of total phosphorus (TP) in rainfall-runoff, and to a lesser extent for the sediment fraction which has the lowest proportion of phosphorus for both classes of events, less than 20 % of TP as shown in Figure 5. Illustrated from Figure 6, COD also had a larg e proportion in suspended part, in particular, for flow-limited events where COD fractions were all more than 60 %, indicating that COD were more likely adsorbed on suspended solids than any other fractions. Resu lts also indicated a lowest proportion of COD in sediment part, which was less than 10 % for all 8 events. Phosphorus and COD fractions varied significantly between mass-limited and flow-limited events, due to the large differen ce of solid fractions generated by stochastic runoff loadings from these events. Partitioning Of Metals In Dissolved And Particulate Phase Ten metals were investigated for each event with results summarized in Table 2, which included total concentrations of As, Ca, Cd, Cr Cu, Fe, Mg, Mn, Pb, and Zn. According to the Criteria Maxium Concentrations (CMCs) of those metals shown in Table 3, total concentrations of Cu, Zn, and Cr exceeded CMCs respectively for al l the events. Moreover, it also indicated that Pbs total concentration exceeded CMC for 4 ev ents (3 of them belonged to the mass-limited events), and Cd was over CMC value for only one mass-limited event. There are no CMCs of Ca, Mg, Fe and Mn in national recommended water qu ality criteria for priority toxic pollutants in EPA 822-Z-99-001 (USEPA 1999). Metal concentrations in dissolved and partic ulate phase were measur ed and summarized in Table 3, based on event mean values for rainfall-runoff before (t0) and after 60 minutes settling (t60). Results indicated that most metal concentrations in disso lved phase were under CMCs for both unsettled and settled runoff samp les, except dissolved Zn of t60 samples. While in

PAGE 78

78 particulate phase, Cu, Zn concentr ations were over CMCs for both t0 and t60 samples and Pb exceeded CMC value for only t0 samples. The separation efficiencies of metals by quiescent settling were summarized in Table 3, suggesting that settling process was an effective method to remove particulate metals from rainfall-runoff with removal efficiencies over 50 % for most metals. However, further treatment of dissolved metals was necessary in order to reduce dissolved metal concentrations below CMCs for a ll the metals in rainfall-runoff, especially for Cu, Zn. Moreover, results in Table 3 also indicated that fd values of metals in influent runoff samples were mostly less than 0.3, excluding Ca and Mg with fd values of 0.78 and 0.5 respectively. All the fd values of metals increased signi ficantly after 60 minutes of quiescent settling, mainly caused by the removal of particul ate metals with settleable and sediment solids during the settling process. For anothe r partitioning parameter, most of Kd values shown here were more than 10 L/Kg with maximal value of 118.4 L/Kg, and the lowest Kd value appeared for Ca only, which was equal to 0.5 L/Kg for unse ttled runoff and 0.6 L/Kg for settled runoff. Metal partitioning between di ssolved and particulate bound fr actions in urban rainfallrunoff is a dynamic process influenced by specifi c mass transfer mechanisms of sorption, ion exchange and surface complexation with both or ganic and inorganic sites on the particulate matter. Event-based equilibrium partitioning of metal species between di ssolved and particulate phase were presented in Figure 7 to 16, expressed as pdf values of 60 replicate samples before and after 60 minutes of quiescen t settling. Results indicated that there was no significant difference of metal concentrations in either diss olved or particulate phas e of influent rainfallrunoff between different type events for all the metal species. While taken 60 minutes of quiescent settling, metal concentrations in the di ssolved phase of runoff varied within a small

PAGE 79

79 range for both type events, while particulate metal concentrations de creased significantly, especially for mass-limited events. Metal Partitioning In Suspended, Settleable And Sediment Solids Metal factions of suspended, se ttleable and sediment solids in influent rain fall-runoff could be derived by metal concentrations ( g/g) of each fraction multiplying the solid concentration (mg/L) in runoff for that specific fraction, wh ich were determined a nd summarized in Table 1 and Table 5 respectively for both type events. Results indicated that metals had a high distribution in the suspended fraction for only flow-limited ev ents, which contributed around 43 to 64 % of concentrations for total particulate metals in influent rainfall-runoff, while on the contrary, which were only 19 to 40 % for mass-lim ited events. In addition, it also indicated that metal species dominated in the settleable fract ion, especially for mass-limited events with a proportion from 53 to 66 %, and the lowest propor tion of metal species appeared in the sediment fraction for both classes of ev ents, less than 30 % of particulate metal concentrations. Conclusions This study investigated the equilibrium partitioning of phosphorus, COD and metals in urban rainfall-runoff at I-10 site, Baron Rouge and got following conclusions. For both flowlimited and mass-limited events, phosphorus partitione d in particular phase rather than dissolved phase of influent runoff, where concentration of particulate phosphorus was in the range of 1 to 3 mg/L and there was only 0.1 to 0.3 mg/L for dissolved phosphorus. Part iculate phosphorus was dominant in suspended fraction of influence runoff for most even ts, especially the flow-limited events with a proportion of more than 50 %, and the lowest proportion was found in sediment fraction for both classes of events which was less than 20 %. COD concentration in rainfall-runoff was in the range of 50 to 200 mg/L for studied events, which was lower than that in genera l municipal wastewater. COD has a high proportion

PAGE 80

80 in the dissolved phase of runoff with fd values larger than 0.6 and Kd values less than 10 L/Kg. In the particulate phase, COD has a higher correlati on with finer particles rather than coarse particles, causing a larger partitioning of COD in suspended solids comparing with settleable and sediment solids. Concentrations of Cu, Zn and Cr exceeded CM Cs for all the research events, even after 60 minutes of quiescent settling, while Pb and Cd exceeded CMCs only for some of the events. Metal concentrations in either dissolved or particulate phase of influent rainfall-runoff varied between different type events for all the me tal species. While taken 60 minutes of quiescent settling, metal concentrations in the dissolved phase of runoff va ried within a small range for both type events, but concentratio ns of particulate metal decreased significantly, especially for mass-limited events. fd values of metal species were in the range of 0.1 to 0.3 mg/L and Kd values from 10 to 100 L/Kg. The high partition of metals in suspended fraction appeared for only flow-limited events with a cont ribution of 43 to 64 %, however, which was only 19 to 40 % for mass-limited events. Metals in settleable fraction dominated in most events, especially for masslimited events with a proportion of 53 to 66 %.

PAGE 81

81Table 3-1. Water chemistry data for 8 ra infall-runoff events with complete runoff volume recovery at the Baton Rouge site Event type Flow-limited Mass-limited Event Date 05 Mar 04 08 Jul 04 10 Jul 04 06 Apr 04 29 Mar 04 30 Apr 04 1 Jun 04 21 Jun 04 Hydrologic and loading parameters PDH (hr) 581621045212410328884VPV (#/L) 14.69.75.88.81.92.00.61.7Runoff volume (L) 37143877417504716389310086697Duration t (min) 4073389013497223Water chemistry parameters x 23.223. 620.821.921.222.423.125.8Temperature (oC) 0.20.40.80.20.20.10.10.6 x 7.57.47.07.07.67.27.58.1pH 0.00.10.00.10.10.00.00.1 x 450.4214.4118.9199.4145.8230.6270.0128.8Conductivity ( s/cm) 1.96.40.11.62.10.40.41.6 x 214.755.294.3101.268.940.42128.160.9TDS [mg/L] 0.80.40.83.51.00.70.30.8 x 296.0219.3125.276.7159.2119.1112.6141.7Suspended solids [mg/L] 43.324.815.09.617.012.719.042.9 x 217.2206.7134.2121.9259.1217.7398.8724.3Settleable solids [mg/L] 35.014.211.920.220.712.320.741.8 x 28.87.312.151.6131.049.2242.4117.6Sediment solids [mg/L] 15.75.62.815.223.87.793.232.3 x 152.9157.8178.6190.7170.652.6230.2183.2Redox (+mV) 18.22.86.36.710.80.30.620.4 x 40.017.517.529.944.023.937.289.5Alkalinity [mg/L] 2.10.12.13.65.30.10.416.6 x 1050.2537.5607.8481.8953.5223.4584.6396.8Hardness [mg/L] 30.711.521.226.862.914.117.911.0 x N/A 172.2234.2168.8132.6290.162.0173.4CODt [mg/L] N/A 10.618.811.927.840.725.427.9 PDH: previous dry hours TDS: total dissolved solids VPV: vehicles per runoff volume t: duration time of runoff CODt: total chemical oxygen demand x : mean and standard deviation

PAGE 82

82Table 3-2. Concentrations of 10 metal species in influent runoff for 8 rainfallrunoff events at the Baton Rouge site Event type Flow-limited Mass-limited Event Date 05 Mar 04 06 Apr 04 08 Jul 04 10 Jul 04 29 Mar 04 30 Apr 04 1 Jun 04 21 Jun 04 Metal, [ g/L] (CMC) Water chemistry parameters (metal concentrations) x 8.64.88.05.43.6 2.14.83.0Total As (360) 1.14.82.42.91.7 0.42.50.4 x 105.870.097.463.086.9 65.0128.4110.8Total Cu (17) 3.423.71.25.111.3 9.721.911.6 x 2.32.92.01.24.3 1.33.62.2Total Cd (3.7) 0.60.60.50.40.5 0.50.50.5 x 79.030.059.937.055.8 86.099.985.3Total Pb (65) 3.22.01.91.36.6 9.47.12.0 x 762.9784.2730.2441.7719.8 441.61018.11120.4Total Zn (110) 129.173.927.552.785.3 64.722.958.0 x 54.144.951.041.846.3 30.980.055.9Total Cr (550/15) 4.07.911.48.77.3 4.27.75.6 x 299.4127.0234.6139.7362.4 160.1469.2486.8Total Mn 9.28.59.213.639.8 21.823.484.8 x 11039.316673.19788.23971.614095.6 6855.114145.918690.5Total Fe 412.2203.7438.4238.31463.2 1089.3892.3261.4 x 49697.521008.423040.724958.544676.4 11983.331457.728353.3Total Ca 1538.110362.2585.91488.327476.1 959.62374.41399.7 x 2813.3941.21266.41114.12703.7 857.31964.42012.3Total Mg 111.1331.640.4117.01238.6 89.5263.2119.2 x : mean and standard deviation

PAGE 83

83Table 3-3. Partitioning and distribution of metals in dissolved a nd particulate phase (sediment + settleable + suspended) for u nsettled influent rainfall-runoff and after 60 minutes of qu iescent settling Metal species Cr Cu Zn As Cd Pb Fe Mn Ca Mg CMC [ g/L] 550(III)/ 15(VI) 17 110 360 4 65 N/A N/A N/A N/A Influent rainfall-runoff x 2.1 11.188.61.30.71.7 3820.191.222808.1860.3Dissolved [ g/L] 2.4 6.748.70.41.01.8 795.353.112447.5426.4 x 53.8 88.7737.58.21.972.1 10739.7215.26588.4848.4Particulate [ g/L] 19.1 16.0192.98.21.06.1 1376.554.04428.8777.5fd 0.04 0.110.110.140.270.02 0.260.300.780.50 Kd (L/Kg) 102.5 28.031.016.420.4118.4 25.718.40.52.9After 60 minutes of quiescent settling x 3.2 14.9115.32.12.81.6 2501.059.417831.5602.6Dissolved [ g/L] 3.7 6.768.21.23.91.7 490.739.66578.2281.9 x 37.7 45.2323.53.30.929.2 3231.085.21667.5302.4Particulate [ g/L] 13.0 9.8170.02.01.34.1 764.524.21980.6202.2fd 0.08 0.250.260.390.750.05 0.440.410.91 0.66Kd (L/Kg) 382.3 29.915.613.16.4111.5 27.622.20.62.6Separation efficiency by settling (%) Particulate 30.0 49.056.159.652.859.4 69.960.4 74.764.4 Total 27.1 39.846.943.439.558.2 60.652.833.747.0 CMC: Criteria Maximum Concentration x : mean and standard deviation fd: dissolved fraction Kd: equilibrium coefficient Total: dissolved + particulate

PAGE 84

84Table 3-4. Rainfall chemistry data for 4 st orm events collected at Baton Rouge site 29 Jan 04 12 Feb 04 26 Feb 04 14 Mar 04 Events x x x x Water chemistry Temperature (oC) 21.50.221.70.322.4 0.222.40.1pH 4.70.24.80.24.8 0.14.30.2Conductivity ( s/cm) 6.30.17.40.011.6 0.018.30.1TDS [mg/L] 2.00.13.00.15.0 0.19.00.1DO [mg/L] 8.70. 39.30. 49.3 0.28.60.3Alkalinity [mg/L] 0.60.20.70.30.7 0.10.30.1Hardness [mg/L] 0.10.10.10.10.1 0.10.10.1Turbidity (NTU) 4.11.21.80.52.5 0.81.60.4Metal species [ g/L] Total Cu 2.50.31.90.41.7 0.41.90.7Total Cd 0.10.10.20.10.1 0.10.10.1Total Cr 0.40.10.50.10.6 0.10. 60.1Total Fe 26.91.118.03.418.0 0.817.90.8Total Mn 2.30.42.00.32.1 0.42.10.3Total Pb 0.20.20.10.10.2 0.20.20.1Total Zn 13.20.27.70.37.2 0.18.60.4Total As 0.50.10.30.10.2 0.10.20.2Total Ca 32.52.025.13.218.7 2.738.03.1Total Mg 3.50.11.20.13.1 0.21.10.5 TDS: total dissolved soli ds DO: dissolved oxygen x : mean and standard deviation n: 6 replicate samples for each event

PAGE 85

85Table 3-5. Metals in particul ate fractions of influence runoff for storm events at the Baton Rouge site (Note units for Ca, F e, Mg) Event type Flow-limited Mass-limited Event Date 05 Mar 04 06 Apr 04 08 Jul 04 10 Jul 04 29 Mar 04 30 Apr 04 1 Jun 04 21 Jun 04 Metal Fraction x x x x x x x x Suspended 4.9 0.6 4.80.64.90.65.20.54.6 0.75.00.65.20.55.20.5Settleable 4.3 0.6 4.30.74.30.74.30.74.3 0.74.30.74.40.74.30.7As [ g/g] Sediment 2.6 0.2 2.60.72.80.73.00.73.0 0.72.50.73.10.73.10.7Suspended 1.1 0.1 1.00.21.00.11.10.11.0 0.21.10.11.10.11.10.2Settleable 1.3 0.6 1.30.71.40.71.30.61.3 0.71.30.71.40.61.30.7Cd [ g/g] Sediment 0.9 0.1 0.80.71.40.71.50.70.9 0.70.70.71.70.71.50.7Suspended 32.4 4.1 33.34.233.34.226.63.539.5 4.830.13.825.63.424.73.3Settleable 16.5 2.6 15.92.522.33.613.72.321.6 3.322.43.522.23.526.64.1Cr [ g/g] Sediment 24.9 1.4 24.51.429.21.329.51.029.0 1.020.51.831.01.231.30.9Suspended 265.4 20.4 258.920.8259.320.8305.517.5215.8 23.9281.619.2312.617.0318.816.6Settleable 72.3 5.3 69.54.8101.09.458.03.7101.0 10.1103.59.795.16.0127.814.1Cu [ g/g] Sediment 90.5 4.8 87.54.4122.312.5116.75.1108.8 5.373.08.9119.95.5121.94.0Suspended 218.4 28.8 219.929.7219.829.6209.023.6230.1 35.3214.626.7207.322.7205.821.9Settleable 121.9 21.1 117.920.0164.429.5100.118.2161.0 27.2164.727.6162.528.0188.930.6Mn [ g/g] Sediment 192.5 14.8 190.815.3207.210.8210.310.0213.3 14.2174.018.1221.413.4219.111.4Suspended 233.6 28.9 229.428.6229.728.6259.531.1201.7 26.3244.129.8264.131.4268.031.8Settleable 276.5 34.0 278.334.7266.330.0278.134.9266.1 31.2270.231.1279.231.9267.430.4Pb [ g/g] Sediment 298.0 21.8 281.421.4426.921.5483.028.6438.7 44.3193.916.2537.131.0552.938.6Suspended 340.1 5.9 337.56.5337.76.5355.72.0320.8 10.7346.44.3358.51.3360.90.7Settleable 239.1 28.9 231.626.9320.144.4196.823.7312.8 43.1322.442.5320.536.4359.647.8Zn [ g/g] Sediment 251.3 8.9 243.99.2303.66.7336.86.4332.1 7.3198.314.6354.96.7367.84.8 Suspended 20.5 2.0 20.42.020.42.020.91.620.0 2.420.61.820.91.621.01.5Settleable 22.6 2.9 22.63.022.02.422.93.121.9 2.621.92.522.42.521.62.5Ca [mg/g] Sediment 22.6 2.2 22.62.224.93.220.31.320.9 1.733.45.221.51.520.21.3Suspended 14.3 1.8 14.01.814.01.816.01.912.2 1.715.01.816.31.916.61.9Settleable 6.8 1.0 6.61.09.21.55.60.98.9 1.49.21.49.11.410.41.6Fe [mg/g] Sediment 13.8 0.9 13.70.914.30.814.50.814.9 0.712.41.315.21.015.00.8Suspended 2.4 0.2 2.30.22.40.22.50.12.3 0.32.40.12.50.22.40.2Settleable 2.3 0.3 2.20.32.30.32.40.32.2 0.32.30.32.30.32.20.3Mg [mg/g] Sediment 1.4 0.1 1.40.11.60.11.60.11.6 0.21.30.21.70.11.70.1

PAGE 86

86 Flow-limited events Mass-limited events 06 April 2004 Particulate TP [mg/L] 0 1 2 3 4 Dissolved TP [mg/L] 0.00 0.05 0.10 0.15 0.20 30 April 2004 Particulate TP [mg/L] 0 1 2 3 4 Dissolved TP [mg/L] 0.00 0.05 0.10 0.15 0.20 29 March 2004 Particulate TP [mg/L] 0 1 2 3 4 Dissolved TP [mg/L] 0.00 0.05 0.10 0.15 0.20 05 March 2004 Particulate TP [mg/L] 0 1 2 3 4 Dissolved TP [mg/L] 0.00 0.05 0.10 0.15 0.20 t0pt60pt0dt60d 10 July 2004 Particulate TP [mg/L] 0 1 2 3 4 Dissolved TP [mg/L] 0.00 0.05 0.10 0.15 0.20 21 June 2004 Particulate TP [mg/L] 0 1 2 3 4 Dissolved TP [mg/L] 0.00 0.05 0.10 0.15 0.20 01 June 2004 Particulate TP [mg/L] 0 1 2 3 4 Dissolved TP [mg/L] 0.00 0.05 0.10 0.15 0.20 08 July 2004 Particulate TP [mg/L] 0 1 2 3 4 Dissolved TP [mg/L] 0.00 0.05 0.10 0.15 0.20 t0pt60pt0dt60d Figure 3-1. Equilibrium of phosphorus partitioning in dissolved and particulate phase before and after settling for 8 rainfall-runoff events, n = 60 for each event. TP: total phosphorus t0p: particulate TP at time = 0 minute t60p: particulate TP at time = 60 minutes t0d: dissolved TP at time = 0 minute t60d: dissolved TP at time = 60 minutes

PAGE 87

87 Flow-limited events Mass-limited events 29 March 2004 COD [mg/L] 0 50 100 150 200 250 300 05 March 2004 COD [mg/L] 0 50 100 150 200 250 300 06 April 2004 COD [mg/L] 0 50 100 150 200 250 300 30 April 2004 COD [mg/L] 0 50 100 150 200 250 300 01 June 2004 COD [mg/L] 0 50 100 150 200 250 300 21 June 2004 COD [mg/L] 0 50 100 150 200 250 300 t0pt60pt0dt60d 08 July 2004 COD [mg/L] 0 50 100 150 200 250 300 10 July 2004 COD [mg/L] 0 50 100 150 200 250 300 t0pt60pt0dt60d Figure 3-2. Equilibrium of COD partitioning in di ssolved and particulate phase before and after settling for 8 rainfall-runoff events, n = 60 for each event. t0p: particulate COD at time = 0 minute t60p: particulate COD at time = 60 minutes t0d: dissolved COD at time = 0 minute t60d: dissolved COD at time = 60 minutes

PAGE 88

88 05 March 2004 06 April 2004 08 July 2004 10 July 2004 29 March 2004 30 April 2004 01 June 2004 21 June 2004 Flow-limited Mass-limited fd 0.00 0.02 0.04 0.06 0.08 0.10 Time = 0 minute (t0) Time = 60 minutes (t60) 05 March 2004 06 April 2004 08 July 2004 10 July 2004 29 March 2004 30 April 2004 01 June 2004 21 June 2004 Kd (L/Kg) 0 500 1000 1500 2000 Flow-limited Mass-limited ( a ) ( c ) ( d ) ( b ) Figure 3-3. fd values and equilibrium coefficient, Kd values of phosphorus in unsettled influent rainfall-runoff (t0) and rainfall-runoff afte r 60 minutes settling (t60), n = 60 for each event.

PAGE 89

89 05 March 2004 06 April 2004 08 July 2004 10 July 2004 29 March 2004 30 April 2004 01 June 2004 21 June 2004 05 March 2004 06 April 2004 08 July 2004 10 July 2004 29 March 2004 30 April 2004 01 June 2004 21 June 2004 Kd, (L/Kg) 0 5 10 15 20 Flow-limited Mass-limited fd 0.0 0.2 0.4 0.6 0.8 1.0 Time = 0 minute (t0) Time = 60 minutes (t60) Flow-limited Mass-limited ( a ) ( b ) ( d ) ( c ) Figure 3-4. fd values and equilibrium coefficient, Kd values of COD in unsettled influent rainfall-runoff (t0) and rainfall-runoff afte r 60 minutes settling (t60), n = 60 for each event.

PAGE 90

90 05 March 2004 06 April 2004 08 July 2004 10 July 2004 29 March 2004 30 April 2004 01 June 2004 21 June 2004 Phosphorus Fraction (%) 0 20 40 60 80 100 Flow-limited events Mass-limited events 72 26 50 40 51 40 50 45 30 54 26 68 67 32 64 34 Suspended Settleable Sediment Figure 3-5. Phosphorus fraction (%) of suspended, settleab le and sediment solids in rainfallrunoff, n = 60 for each event.

PAGE 91

91 05 March 2004 06 April 2004 08 July 2004 10 July 2004 29 March 2004 30 April 2004 01 June 2004 21 June 2004 COD Fraction (%) 0 20 40 60 80 100 Flow-limited events Mass-limited events 88 11 70 27 64 31 69 29 47 42 38 58 75 24 73 26 Suspended Settleable Sediment Figure 3-6. COD fraction (%) of suspended, settleable and sedime nt solids in rainfall-runoff, n = 60 for each event.

PAGE 92

92 Flow-limited events Mass-limited events 10 July 2004 As [ g/L] 024681012 pdf for (As)d 0.0 0.5 1.0 1.5 2.0 t0d : = 1.9, = 0.6 t60d : = 1.6, = 0.5 t0p : = 48.5, = 4.6 t60p : = 45.4, = 3.6 08 July 2004 pdf for (As)d 0.0 0.5 1.0 1.5 2.0 t0d : = 1.7, = 0.6 t60d : = 1.7, = 0.4 t0p : = 4.2, = 0.9 t60p : = 3.8, = 0.7 05 March 2004 pdf for (As)d 0.0 0.5 1.0 1.5 2.0 t0d = 2.5, = 0.7 t60d = 4.8, = 0.6 t0p = 6.4, = 0.6 t60p = 4.2, = 0.7 06 April 2004 pdf for (As)d 0.0 0.5 1.0 1.5 2.0 t0d : = 0.6, = 0.2 t60d : = 2.4, = 0.7 t0p : = 3.0, = 0.7 t60p : = 2.7, = 0.4 29 March 2004 pdf for (As)p 0.0 0.5 1.0 1.5 2.0 t0d = 1.8, = 0.3 t60d = 1.3, = 0.2 t0p = 4.7, = 0.7 t60p = 2.6, = 0.4 30 April 2004 pdf for (As)p 0.0 0.5 1.0 1.5 2.0 t0d : = 1.3, = 0.2 t60d : = 1.3, = 0.3 t0p : = 4.0, = 0.8 t60p : = 0.9, = 0.2 01 June 2004 pdf for (As)p 0.0 0.5 1.0 1.5 2.0 t0d : = 1.9, = 0.9 t60d : = 2.0, = 0.7 t0p : = 10.1, = 0.7 t60p : = 3.1, = 0.6 21 June 2004 As [ g/L] 024681012 pdf for (As)p 0.0 0.5 1.0 1.5 2.0 t0d : = 1.6, = 0.2 t60d : = 1.2, = 0.2 t0p : = 7.2, = 0.7 t60p : = 2.0, = 0.3 d p Figure 3-7. Equilibrium partitioning between dissolved As, (As)d and particulate As, (As)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60). Each pdf was developed from a sample population of n = 60 for each event. Each pdf for the dissolved pha se is associated with the left pdf axis and each pdf for the particulate phase is associated with the right pdf axis.

PAGE 93

93 Flow-limited events Mass-limited events 10 July 2004 Cd [ g/L] 024681012 pdf for (Cd)d 0.0 0.5 1.0 1.5 2.0 t0d : = 1.4, = 0.4 t60d : = 0.86, = 0.2 t0p : = 1.0, = 0.2 t60p : = 0.5, = 0.2 08 July 2004 pdf for (Cd)d 0.0 0.5 1.0 1.5 2.0 t0d : = 1.3, = 0.3 t60d : = 1.1, = 0.2 t0p : = 1.8, = 0.7 t60p : = 1.1, = 0.4 05 March 2004 pdf for (Cd)d 0.0 0.5 1.0 1.5 2.0 t0d = 1.1, = 0.4 t60d = 1.7, = 0.5 t0p = 3.3, = 0.9 t60p = 2.4, = 0.6 06 April 2004 pdf for (Cd)d 0.0 0.5 1.0 1.5 2.0 t0d : = 0.7, = 0.2 t60d : = 1.4, = 0.3 t0p : = 2.5, = 0.6 t60p : = 1.4, = 0.4 29 March 2004 pdf for (Cd)p 0.0 0.5 1.0 1.5 2.0 t0d = 2.7, = 0.7 t60d = 4.1, = 1.0 t0p = 1.8, = 0.7 t60p = 0.9, = 0.2 30 April 2004 pdf for (Cd)p 0.0 0.5 1.0 1.5 2.0 t0d : = 0.7, = 0.2 t60d : = 0.4, = 0.3 t0p : = 1.3, = 0.5 t60p : = 0.4, = 0.2 01 June 2004 pdf for (Cd)p 0.0 0.5 1.0 1.5 2.0 t0d : = 1.3, = 0.5 t60d : = 0.4, = 0.2 t0p : = 1.4, = 0.5 t60p : = 0.8, = 0.2 21 June 2004 Cd [ g/L] 024681012 pdf for (Cd)p 0.0 0.5 1.0 1.5 2.0 t0d : = 1.4, = 0.4 t60d : = 1.2, = 0.3 t0p : = 2.4, = 0.5 t60p : = 1.2, = 0.2 d p Figure 3-8. Equilibrium partitioning between dissolved Cd, (Cd)d and particulate Cd, (Cd)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60). Each pdf was developed from a sample population of n = 60 for event. Each pdf for the dissolved phase is associated with th e left pdf axis and each pdf for the particulate phase is associated with the right pdf axis.

PAGE 94

94 Flow-limited events Mass-limited events 10 July 2004 Cr [ g/L] 020406080100 pdf for (Cr)d 0.0 0.2 0.4 0.6 0.8 1.0 t0d : = 1.9, = 0.6 t60d : = 1.6, = 0.5 t0p : = 48.5, = 4.6 t60p : = 45.4, = 3.6 08 July 2004 pdf for (Cr)d 0.0 0.2 0.4 0.6 0.8 1.0 t0d : = 1.9, = 0.6 t60d : = 1.3, = 0.3 t0p : = 55.7, = 3.0 t60p : = 54.8, = 4.0 05 March 2004 pdf for (Cr)d 0.0 0.2 0.4 0.6 0.8 1.0 t0d = 1.1, = 0.4 t60d = 9.0, = 1.1 t0p = 60.1, = 5.5 t60p = 58.9, = 4.4 06 April 2004 pdf for (Cr)d 0.0 0.2 0.4 0.6 0.8 1.0 t0d : = 1.5, = 0.5 t60d : = 5.4, = 1.0 t0p : = 48.2, = 3.9 t60p : = 28.9, = 2.4 29 March 2004 pdf for (Cr)p 0.00 0.06 0.12 0.18 0.24 0.30 t0d = 8.5, = 1.3 t60d = 6.1, = 1.2 t0p = 42.0, = 5.2 t60p = 33.1, = 3.6 30 April 2004 pdf for (Cr)p 0.00 0.06 0.12 0.18 0.24 0.30 t0d : = 2.2, = 0.6 t60d : = 1.2, = 0.4 t0p : = 32.9, = 4.3 t60p : = 16.6, = 1.5 01 June 2004 pdf for (Cr)p 0.00 0.06 0.12 0.18 0.24 0.30 t0d : = 1.5, = 0.5 t60d : = 2.5, = 0.6 t0p : = 67.3, = 5.3 t60p : = 31.6, = 3.1 21 June 2004 Cr [ g/L] 020406080100 pdf for (Cr)p 0.00 0.06 0.12 0.18 0.24 0.30 t0d : = 2.6, = 0.8 t60d : = 2.1, = 0.7 t0p : = 60.3, = 8.8 t60p : = 27.8, = 3.4 d p Figure 3-9. Equilibrium partitioning between dissolved Cr, (Cr)d and particulate Cr, (Cr)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60). Each pdf was developed from a sample population of n = 60 for each event. Each pdf for the dissolved phase is associated with the le ft pdf axis and each pdf for the particulate phase is a ssociated with the right pdf axis.

PAGE 95

95 Flow-limited events Mass-limited events 10 July 2004 Cu [ g/L] 020406080100120140 pdf for (Cu)d 0.0 0.2 0.4 0.6 0.8 1.0 t0d : = 8.3, = 2.7 t60d : = 13.1, = 1.8 t0p : = 60.7, = 7.3 t60p : = 48.6, = 6.0 08 July 2004 pdf for (Cu)d 0.0 0.2 0.4 0.6 0.8 1.0 t0d : = 9.7, = 1.2 t60d : = 19.4, = 3.2 t0p : = 97.4, = 8.3 t60p : = 61.1, = 7.7 05 March 2004 pdf for (Cu)d 0.0 0.2 0.4 0.6 0.8 1.0 t0d = 3.5, = 0.5 t60d = 3.1, = 0.6 t0p = 113.7, = 10.7 t60p = 92.6, = 8.0 06 April 2004 pdf for (Cu)d 0.0 0.2 0.4 0.6 0.8 1.0 t0d : = 4.8, = 0.7 t60d : = 19.1, = 2.7 t0p : = 72.4, = 8.9 t60p : = 26.9, = 1.5 29 March 2004 pdf for (Cu)p 0.00 0.06 0.12 0.18 0.24 0.30 t0d = 26.5, = 3.1 t60d = 14.4, = 1.7 t0p = 67.1, = 5.4 t60p = 40.4, = 2.4 30 April 2004 pdf for (Cu)p 0.00 0.06 0.12 0.18 0.24 0.30 t0d : = 9.3, = 1.6 t60d : = 5.6, = 0.7 t0p : = 61.8, = 4.2 t60p : = 15.8, = 1.5 01 June 2004 pdf for (Cu)p 0.00 0.06 0.12 0.18 0.24 0.30 t0d : = 17.6, = 1.7 t60d : = 27.7, = 2.0 t0p : = 123.1, = 8.5 t60p : = 43.7, = 4.6 21 June 2004 Cu [ g/L] 020406080100120140 pdf for (Cu)p 0.00 0.06 0.12 0.18 0.24 0.30 t0d : = 8.7, = 0.9 t60d : = 16.6, = 1.9 t0p : = 133.4, = 12.8 t60p : = 32.6, = 2.6 d p Figure 3-10. Equilibrium partitioning between dissolved Cu, (Cu)d and particulate Cu, (Cu)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60). Each pdf was developed from a sample population of n = 60 for each event. Each pdf for the dissolved pha se is associated with the left pdf axis and each pdf for the particulate phase is associated with the right pdf axis.

PAGE 96

96 Flow-limited events Mass-limited events 10 July 2004 Fe [*1000 g/L] 05101520 pdf for (Fe)d 0.0 0.5 1.0 1.5 2.0 t0d : = 0.7, = 0.3 t60d : = 2.4, = 0.3 t0p : = 4.0, = 0.3 t60p : = 1.8, = 0.6 08 July 2004 pdf for (Fe)d 0.0 0.5 1.0 1.5 2.0 t0d : =1.9, =0.6 t60d : =1.3, =0.3 t0p : =55.7, =3.0 t60p : =54.8, =4.0 05 March 2004 pdf for (Fe)d 0.0 0.5 1.0 1.5 2.0 t0d = 0.7, = 0.2 t60d = 0.5, = 0.3 t0p = 12.0, = 0.5 t60p = 7.0, = 0.6 06 April 2004 pdf for (Fe)d 0.0 0.5 1.0 1.5 2.0 t0d : = 0.8, = 0.2 t60d : = 0.6, = 0.3 t0p : = 17.6, = 1.1 t60p : = 3.0, = 0.5 29 March 2004 pdf for (Fe)p 0.0 0.5 1.0 1.5 2.0 t0d =8.5, =1.3 t60d =6.1, =1.2 t0p =42.0, =5.2 t60p =33.1, =3.6 30 April 2004 pdf for (Fe)p 0.0 0.5 1.0 1.5 2.0 t0d : = 4.6, = 0.6 t60d : = 3.6, = 0.4 t0p : = 5.1, = 0.8 t60p : = 2.4, = 0.4 01 June 2004 pdf for (Fe)p 0.0 0.5 1.0 1.5 2.0 t0d : = 8.1, = 0.6 t60d : = 2.2, = 0.4 t0p : = 11.2, = 0.7 t60p : = 1.7, = 0.4 21 June 2004 Fe [*1000 g/L] 05101520 pdf for (Fe)p 0.0 0.5 1.0 1.5 2.0 t0d : = 8.6, = 0.8 t60d : = 3.4, = 0.5 t0p : = 16.0, = 1.3 t60p : = 7.0, = 0.4 d p Figure 3-11. Equilibrium partitioning between dissolved Fe, (Fe)d and particulate Fe, (Fe)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60). Each pdf was developed from a sample population of n = 60 for each event. Each pdf for the dissolved pha se is associated with the left pdf axis and each pdf for the particulate phase is associated with the right pdf axis.

PAGE 97

97 Flow-limited events Mass-limited events 10 July 2004 Mn [ g/L] 0100200300400 pdf for (Mn)d 0.00 0.05 0.10 0.15 0.20 t0d : = 44.6, = 3.8 t60d : = 28.9, = 2.8 t0p : = 105.6, = 8.1 t60p : = 61.6, = 5.9 08 July 2004 pdf for (Mn)d 0.00 0.05 0.10 0.15 0.20 t0d : =98.4, =10.1 t60d : =46.4, =3.8 t0p : =151.3, =14.2 t60p : =124.8, =14.0 05 March 2004 pdf for (Mn)d 0.00 0.05 0.10 0.15 0.20 t0d =17.1, =3.6 t60d =9.2, =2.2 t0p =313.7, =10.3 t60p =181.6, =7.8 06 April 2004 pdf for (Mn)d 0.00 0.05 0.10 0.15 0.20 t0d : =37.9, =5.0 t60d : =31.2, =3.8 t0p : =98.9, =7.7 t60p : =28.9, =2.4 29 March 2004 pdf for (Mn)p 0.00 0.05 0.10 0.15 0.20 t0d = 149.4, = 13.0 t60d = 111.9, = 15.9 t0p = 236.6, = 14.3 t60p = 60.8, = 6.2 30 April 2004 pdf for (Mn)p 0.00 0.05 0.10 0.15 0.20 t0d : = 15.4, = 2.9 t60d : = 11.0, = 2.2 t0p : = 171.9, = 9.2 t60p : = 35.2, = 3.6 01 June 2004 pdf for (Mn)p 0.00 0.05 0.10 0.15 0.20 t0d : = 201.2, = 14.8 t60d : = 153.6, = 14.6 t0p : = 297.8, = 25.6 t60p : = 67.1, = 5.6 21 June 2004 Mn [ g/L] 0100200300400 pdf for (Mn)p 0.00 0.05 0.10 0.15 0.20 t0d : = 175.4, = 13.8 t60d : = 123.2, = 9.6 t0p : = 345.9, = 22.6 t60p : = 120.1, = 8.0 d p Figure 3-12. Equilibrium partit ioning between dissolved Mn, (Mn)d and particulate Mn, (Mn)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60). Each pdf was developed from a sample population of n = 60 for each event. Each pdf for the dissolved pha se is associated with the left pdf axis and each pdf for the particulate phase is associated with the right pdf axis.

PAGE 98

98 Flow-limited events Mass-limited events 10 July 2004 Pb [ g/L] 020406080100120140 pdf for (Pb)d 0.0 0.2 0.4 0.6 0.8 1.0 t0d : = 0.2, = 0.4 t60d : = 0.3, = 0.5 t0p : = 40.9, = 4.5 t60p : = 25.4, = 3.5 08 July 2004 pdf for (Pb)d 0.0 0.2 0.4 0.6 0.8 1.0 t0d : = 0.3, = 0.4 t60d : = 0.3, = 0.5 t0p : = 66.2, = 4.2 t60p : = 42.3, = 3.6 05 March 2004 pdf for (Pb)d 0.0 0.2 0.4 0.6 0.8 1.0 t0d = 3.0, = 0.6 t60d = 2.8, = 0.4 t0p = 87.7, = 5.6 t60p = 59.0, = 4.6 06 April 2004 pdf for (Pb)d 0.0 0.2 0.4 0.6 0.8 1.0 t0d : = 3.0, = 0.4 t60d : = 4.1, = 0.7 t0p : = 32.9, = 2.9 t60p : = 13.8, = 4.7 29 March 2004 pdf for (Pb)p 0.00 0.06 0.12 0.18 0.24 0.30 t0d = 7.3, = 1.0 t60d = 2.5, = 0.4 t0p = 53.8, = 6.0 t60p = 21.9, = 3.6 30 April 2004 pdf for (Pb)p 0.00 0.06 0.12 0.18 0.24 0.30 t0d : = 0.2, = 0.4 t60d : = 0.2, = 0.4 t0p : = 95.3, = 10.6 t60p : = 19.4, = 2.0 01 June 2004 pdf for (Pb)p 0.00 0.06 0.12 0.18 0.24 0.30 t0d : = 3.3, = 0.3 t60d : = 1.3, = 0.7 t0p : = 107.4, = 9.9 t60p : = 30.2, = 2.2 21 June 2004 Pb [ g/L] 020406080100120140 pdf for (Pb)p 0.00 0.06 0.12 0.18 0.24 0.30 t0d : = 1.9, = 0.3 t60d : = 2.4, = 0.4 t0p : = 92.7, = 7.3 t60p : = 22.0, = 3.9 d p Figure 3-13. Equilibrium partitioning between dissolved Pb, (Pb)d and particulate Pb, (Pb)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60). Each pdf was developed from a sample population of n = 60 for each event. Each pdf for the dissolved pha se is associated with the left pdf axis and each pdf for the particulate phase is associated with the right pdf axis.

PAGE 99

99 Flow-limited events Mass-limited events 10 July 2004 Zn [ g/L] 020040060080010001200 pdf for (Zn)d 0.00 0.02 0.04 0.06 0.08 0.10 t0d : = 71.1, = 15.2 t60d : = 165.7, = 11.8 t0p : = 411.8, = 62.1 t60p : = 341.5, = 46.0 08 July 2004 pdf for (Zn)d 0.00 0.02 0.04 0.06 0.08 0.10 t0d : = 88.9, = 7.6 t60d : = 128.3, = 10.2 t0p : = 712.6, = 47.9 t60p : = 442.6, = 33.8 05 March 2004 pdf for (Zn)d 0.00 0.02 0.04 0.06 0.08 0.10 t0d = 22.6, = 4.1 t60d = 17.1, = 5.2 t0p = 822.6, = 50.9 t60p = 768.9, = 60.7 06 April 2004 pdf for (Zn)d 0.00 0.02 0.04 0.06 0.08 0.10 t0d : = 44.2, = 4.4 t60d : = 158.5, = 16.2 t0p : = 822.2, = 81.8 t60p : = 265.5, = 16.8 29 March 2004 pdf for (Zn)p 0.00 0.01 0.02 0.03 0.04 0.05 t0d = 141.6, = 12.4 t60d = 153.0, = 12.8 t0p = 642.5, = 23.8 t60p = 268.9, = 17.3 30 April 2004 pdf for (Zn)p 0.00 0.01 0.02 0.03 0.04 0.05 t0d : = 34.3, = 6.7 t60d : = 36.8, = 7.5 t0p : = 452.6, = 72.5 t60p : = 79.4, = 12.3 01 June 2004 pdf for (Zn)p 0.00 0.01 0.02 0.03 0.04 0.05 t0d : = 136.4, = 12.8 t60d : = 130.2, = 11.6 t0p : = 979.6, = 45.5 t60p : = 206.8, = 21.3 21 June 2004 Zn [ g/L] 020040060080010001200 pdf for (Zn)p 0.00 0.01 0.02 0.03 0.04 0.05 t0d : = 170.0, = 16.3 t60d : = 132.9, = 12.6 t0p : = 1055.9, = 73.4 t60p : = 214.5, = 13.3 d p Figure 3-14. Equilibrium partitioning between dissolved Zn, (Zn)d and particulate Zn, (Zn)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60). Each pdf was developed from a sample population of n = 60 for each event. Each pdf for the dissolved pha se is associated with the left pdf axis and each pdf for the particulate phase is associated with the right pdf axis.

PAGE 100

100 Flow-limited events Mass-limited events 10 July 2004 Mg [ g/L] 0500100015002000 pdf for (Mg)d 0.00 0.01 0.02 0.03 0.04 0.05 t0d : = 784.3, = 90.1 t60d : = 808.8, = 70.2 t0p : = 329.3, = 29.5 t60p : = 214.5, = 16.5 08 July 2004 pdf for (Mg)d 0.00 0.01 0.02 0.03 0.04 0.05 t0d : = 641.5, = 40.6 t60d : = 646.8, = 30.3 t0p : = 624.2, = 22.8 t60p : = 343.0, = 10.8 05 March 2004 pdf for (Mg)d 0.00 0.01 0.02 0.03 0.04 0.05 t0d = 1623.3, = 30.0 t60d = 1405.4, = 50.1 t0p = 1190.0, = 40.4 t60p = 684.5, = 20.0 06 April 2004 pdf for (Mg)d 0.00 0.01 0.02 0.03 0.04 0.05 t0d : = 1126.1, = 56.0 t60d : = 541.8, = 42.5 t0p : = 399.2, = 29.5 t60p : = 149.1, = 12.0 29 March 2004 pdf for (Mg)p 0.00 0.01 0.02 0.03 0.04 0.05 t0d = 1628.2, = 60.3 t60d = 1285.3, = 79.2 t0p = 1074.7, = 60.7 t60p = 316.5, = 14.7 30 April 2004 pdf for (Mg)p 0.00 0.01 0.02 0.03 0.04 0.05 t0d : = 211.0, = 11.9 t60d : = 207.2, = 10.1 t0p : = 645.5, = 35.9 t60p : = 103.4, = 10.4 01 June 2004 pdf for (Mg)p 0.00 0.01 0.02 0.03 0.04 0.05 t0d : = 897.0, = 40.7 t60d : = 906.0, = 53.3 t0p : = 1067.3, = 84.0 t60p : = 222.8, = 46.7 21 June 2004 Mg [ g/L] 0500100015002000 pdf for (Mg)p 0.00 0.01 0.02 0.03 0.04 0.05 t0d : = 459.8, = 22.8 t60d : = 444.0, = 10.7 t0p : = 1552.0, = 48.2 t60p : = 246.0, = 21.0 p d Figure 3-15. Equilibrium partit ioning between dissolved Mg, (Mg)d and particulate Mg, (Mg)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60). Each pdf was developed from a sample population of n = 60 for each event. Each pdf for the dissolved pha se is associated with the left pdf axis and each pdf for the particulate phase is associated with the right pdf axis.

PAGE 101

101 Flow-limited events Mass-limited events 10 July 2004 Ca [mg/L] 0102030405060 pdf for (Ca)d 0.0 0.1 0.2 0.3 0.4 0.5 t0d : = 23.0, = 4.3 t60d : = 18.8, = 2.2 t0p : = 9.2, = 0.7 t60p : = 2.2, = 0.8 08 July 2004 pdf for (Ca)d 0.0 0.1 0.2 0.3 0.4 0.5 t0d : = 20.5, = 3.6 t60d : = 20.0, = 2.3 t0p : = 2.6, = 0.5 t60p : = 1.1, = 0.7 05 March 2004 pdf for (Ca)d 0.0 0.1 0.2 0.3 0.4 0.5 t0d = 39.4, = 6.0 t60d = 37.3, = 4.1 t0p = 10.3, = 4.4 t60p = 5.7, = 1.0 06 April 2004 pdf for (Ca)d 0.0 0.1 0.2 0.3 0.4 0.5 t0d : = 18.4, = 2.5 t60d : = 18.6, = 6.0 t0p : = 2.6, = 0.8 t60p : = 0.8, = 0.6 29 March 2004 pdf for (Ca)p 0.0 0.2 0.4 0.6 0.8 1.0 t0d = 35.5, = 4.3 t60d = 18.8, = 2.2 t0p = 9.2, = 0.7 t60p = 2.2, = 0.8 30 April 2004 pdf for (Ca)p 0.0 0.2 0.4 0.6 0.8 1.0 t0d : = 8.6, = 1.9 t60d : = 8.1, = 1.1 t0p : = 3.4, = 1.0 t60p : = 0.6, = 0.4 01 June 2004 pdf for (Ca)p 0.0 0.2 0.4 0.6 0.8 1.0 t0d : = 21.9, = 2.7 t60d : = 22.7, = 3.3 t0p : = 9.5, = 0.8 t60p : = 1.0, = 0.7 21 June 2004 Ca [mg/L] 0102030405060 pdf for (Ca)p 0.0 0.2 0.4 0.6 0.8 1.0 t0d : = 15.1, = 2.8 t60d : = 14.3, = 1.7 t0p : = 13.2, = 2.4 t60p : = 0.8, = 0.8 p d Figure 3-16. Equilibrium partitioning between dissolved Ca, (Ca)d and particulate Ca, (Ca)p phases for influent (time = 0 minute, t0) and after 60 minutes of quiescent settling (time = 60 minutes, t60). Each pdf was developed from a sample population of n = 60 for each event. Each pdf for the dissolved pha se is associated with the left pdf axis and each pdf for the particulate phase is associated with the right pdf axis.

PAGE 102

102 CHAPTER 4 GRANULOMETRY TRANSPORT AND SOLUBILI TY OF DRY DEPOSITION PARTICLES FOR AN URBAN HIGHWAY LAND USE Introduction Atmospheric deposition, defined as pollution from the atmosphere associated with dry deposition in the form of dust, wet deposition in the form of rain and snow, or as a result of vapor exchanges, plays a major role in pollutant circulation and transport around urban areas (Colman and Breault 2000; Yun et al. 2002). Dry depos ition (dust fall) is considered a main path for removal of most contaminants from the at mosphere (Kobriger 1984); however, instead of eliminating these constituents from the environm ent, these pathways only transport constituents to terrestrial surfaces and rece iving waters, and the overall pollu tant loadings are not reduced (Barrett et al. 1995). Contamin ants accumulated on impervious urban surfaces from dry deposition can be easily washed off during a st orm and create a significa nt loading to urban rainfall-runoff (Sansalone et al. 1998). Most untr eated rainfall-runoff goes di rectly into the urban ecosystem, and so constituents carried with runoff can seriously impact the terrestrial and aquatic system once deposited from the atmosphere (Glenn and Sansalone 2002). Dry deposition rates of contaminants such as particulate matters and metals vary dramatically with land use. For example, partic ulate matter and metals from an urban area are deposited at a much higher rate than a rura l area (Paode et al. 1998). Deposition rates in highway areas have been shown to decrease with increasing dist ance away from traffic lanes, suggesting that anthropogenic ac tivities at urban areas, and esp ecially highways, are important sources of contaminants related to atmosphe ric deposition (Hedges and Wren 1987; Harned 1988). These anthropogenic constitu ents are generated mainly from traffic activities, vehicular component wear, fluid leaks, pavement erosi on and roadway maintenance (Shinya et al. 2000; Sansalone and Cristina 2004).

PAGE 103

103 Highway surface material was investigated b ecause highways are a la rge receiver of dry deposition solids and are an important source of rainfall-runoff solids. Results indicated that most of the highway surface material was inor ganic and similar to common sand (Characklis 1997; Sartor and Boyd 1972; Sartor et al. 1974). However, the contribution of dry deposition solids to surface material was difficult to measure b ecause of the large variation in particle size, meteorology (atmospheric stability, relative humid ity and wind speed), surface characteristics (Yun et al. 2002), and in particular traffic activities. Due to the di fficulty of separating out these influential factors, previous st udies have not addressed the gr anulometric dist ribution of dry deposition solids near highway areas, and the trans port of dry deposition particles into rainfall run-off. Therefore, in this st udy, we collected dry deposition so lids at highway areas in Baton Rouge during a half-year period, with the purpose of investigati ng the granulometric distribution of dry deposition particles; we also sought to de termine the influence of particle size gradation and initial pH of rainfall water on the potential transport and solubility capabilities of dry deposition particles that are moved into rainfall-runoff. Objective There are three objectives in this research. The first objective is to examine dry deposition flux at different sampling sites on the Interstate10 Bridge in Baton Rouge and the relationship of dry deposition solids mass and previous dry hours. The second objective is to investigate the granulometry of dry deposition solids (size grad ation, specific gravity, surface area) for 17 sampling events from January to July 2004, a nd to apply optimized Gamma distribution in modeling particle size distributions (PSDs) of dry deposition solids. The third objective is to determine the solubility (conductivity, TDS, pH a nd alkalinity) of dry deposition particles as a function of time and pH based on particle size gradation.

PAGE 104

104 Background Study Dry Deposition Sampling According to different purposes, sampler desi gn of dry deposition a nd material choice of sampler surface are also different. Three types of dry deposition sampler are widely used at present: impactor samplers, bucket collectors and standard surface collectors,. Impactor samplers are more suitable for catching fine particles. Rotary-impactor samplers are used to collect particles from 6.5 to 36.5 m and conventional ca scade-impactor samplers are used for particles from 0.43 to 10 m (Holsen and Noll 1992). Buck et collectors and standard surface collectors can work for whole size-range particles, especi ally for large particle flux. A limitation of the bucket samplers is that they only catch vertic ally falling particles. Any horizontally moving particle will be blocked by the bucket wall a nd prevented from being captured inside. Thus particles inside the bucket may not reflect the exact scenario of dry deposition in situ. The standard surface sampler is better than a bucket sa mpler in terms of reproducibility of results and correspondence to total depositi on where turbulence is at a mi nimum. While applying this collector in a highway setting, the effect of traffic-induced turbulence and highway surface roughness needs to be considered and dry deposition on standard surface also needs to be adjusted. The adjustment depends on the particle size of deposited materials. Dry deposition, especially for small part icle deposition, depends on various physical characteristics of the collection surface us ed. Vandenberg and Knoerr (1985), for instance, measured the following variation for sulfate: 1.1, 3.2 and 5.8 mg SO4 2-/m2/d in a Petri plate, from the interior surface of a bucket, and from a Te flon filter surface. In Lins study (1993), a PVC plate with a sharply beveled leading edge (10 de grees) was used as a standard surface, on which mylar strips were covered with a thin layer of Apezion L grease. A sim ilar kind of PVC plate, with a sharply beveled leading edge (less than 10 degree) was developed that can minimize air

PAGE 105

105 disruption and provide an estimate of the lower li mit of deposition (Davidson et al. 1985; Holsen et al. 1991). They also affixed Mylar strips on the plate in orde r to prevent the bouncing-off and resuspension of particles. Transfer Mechanism And Solubility Test Of Dry Deposition Particles Dry deposition of highway-generated particle s and associated constituents onto areas adjoining to roadway surfaces were found to be re lated to average daily traffic, wind speed and direction, and available surface load (Kobriger, 1984). In other research conducted by Kobriger and Geinopolos (1984) at a rural highway site in North Carolina, atmospheric deposition was 8 % of the total material loading on the highway surface, with vehicle deposition and highway maintenance (including sanding and salti ng) at 25 % and 67 %, respectively. There are two mechanisms for removal of materials from the highway surface. One mechanism is caused by natural and vehicle-indu ced winds which constan tly sweep the highway surface and remove dirt and debris from the traffic lane to the cu rb and shoulder. This explains why the majority of material on highway surfaces ar e found within 3 feet of the curb, rather than in the traffic lane (Sartor et al. 1972). In this study, a dry deposition sampler was erected on the shoulder of highway and was supposed to colle ct both atmospheric deposition and resuspended sweep-off materials, which are syncrously created from the highways surface. The second mechanism is intense urban rainfall which washes materials from a highways surface into rainfall runoff. Vaze and Chiew (2003a ) showed that common storms only remove a small proportion of the total surface pollutant load and that street sweeper s can release the finer material but only partially removes them, renderi ng the fine sediment available for washoff by the next storm. They also stated that the su rface pollutants are much coarser compared to the sediment in washoff samples (Vaze and Chiew 2003b). Of the dry samples collected on an urban road surface in Melbourne, Australia only up to 15 % measured finer than 100m, compared to

PAGE 106

106 washoff samples where almost all the sediment s measured finer than 100m (Drapper et al. 2000). Water-quality properties, such as pH and specific conductivity, ar e main factors for chemical reactivity of atmospheric deposition on highway surfaces (Bricker, 1999). Atmospheric deposition which is usually acid ic, particularly in the eastern United States (Rice and Bricker, 1992; Bricker, 1999)could enhance the geochemical mobility of many hi ghway contaminants (Colman et al., 2001). Particles play an important role in release and adsorption of contaminants, such as heavy metals, to and from highway runo ff (Sansalone et al. 1998). Ellis et al. (1987) found that leachate levels are in dependent of pH, and that pH remained relatively constant throughout the time period of the experiment. Other factors, such as ion exchange processes, aeration and agitation of the samples, and me tal binding would also affect contaminants solubility during the leaching experiments (Ellis et al. 1987). Methodology Dry Deposition Sampling Five sampling sites were chosen for this st udy with consideration given to the different urban highway pathways for dry deposition cont aminants: materials swept up and resuspended by natural wind and highway traffic, and vertic al gravitational sett ling from background and traffic sources. At one site, located adjacent to the traffic lane of Interstate-10 East at City Park Lake in Baton Rouge, a 1-m2 PVC sampler with 100 mm ver tical sides was placed at the highway shoulder and directly adjacent to the bri dge railing. The sampler was at the same level as the traffic lane and there was no embankment between the sampler and the traffic lane. This adjacency allowed entrained particles to be cap tured by the sampler. Due to the short distance between the traffic lane and the sampler, material collected from this site were considered to be representative for illustrating the effect of highway traffic and dry deposition contaminants on

PAGE 107

107 urban highway areas. The second sampling location was between the eastbound and westbound lanes of Interstate-10; two 1-m2 samplers were placed at this site. The samplers were located one meter below the bridge railing and vertically de posited dry deposition from vehicular turbulence. In this case, a comparatively smaller amount of dry deposition particle s were collected by the samplers. The third site was located on the roof of CEBA annex building, about 50 feet away from the roadside, where two samplers were placed to collect control or background measurements. Previous Dry Hours, Dry Deposition Solid Mass And D50m Samplers were set up only if there were more than four previous dry days for deposition sampling, and samplers were recovered and covere d before the beginning of the next rain event. The sample collection period in this study was from 17 January 2004 to 16 July 2004. During the six month period, dry deposition particles from different sampling sites were retrieved on 17 occasions. The particles in each sampler were rec overed with a very soft brush and a stainless steel recovery tool, and saved in a glass samp le jar with a known weight. Based on solid mass data and relative duration time, the dry deposit ed flux of each sampling site (sampler) can be calculated through equation 4-1: t M Qd d/ (4-1) Where Qd is the dry deposited flow rate, Md is mass of dry deposite d solid, and t is the duration time. Results indicated th at the dry deposited flow rate is not a constant value in different previous dry periods. Solid mass had a nonlinear relati onship to duration time, which can be modeled and described as an e xponential equation shown in equation 4-2. ) 1 () ( t b de a M (4-2)

PAGE 108

108 The D50m is a common index for particle grad ation, and the relationship between D50m and previous dry hours (PDH) will illustrate the e ffect of PDH on gradat ion of dry deposition particles. Particle Size Distribution The particles collected in each event were sift ed through a set of graded sieves, we used a No. 3/8 sieve (> 9500 m) at the top, and a finer sieve, No. 500 (25 m) at the bottom. Eighteen gradations of particles were separated for la ter analysis. Analysis followed ASTM D422-63. Mass balance provided a 98 % recovery of the original mass of the gradation. Solids collected from each sieve were transfe rred into a glass sample jar with known tare weight and kept them separately. Percentage of fi ner particles by mass, which is used to describe cumulative particle size distribution, is defined as the percentage of total mass for solids less than a specific particle size to overall mass for solids in entire size range for different sampling event. Modeling Of Mass-Based Particle Size Distributions (PSDs) Cumulative mass distribution of metal species for dry deposition particles were fit with a cumulative gamma distribution model. The prob ability density function (PDF) of a gamma distribution is expresse d in equation 4-3 where represents a shape parameter and is a scaling parameter. The cumulative gamma distri bution is given in equation 4-4 where is the gamma function and x is the incomplete gamma functi on, shown in equation 4-5 and equation 4-6. ) ( / ) (/ 1 xe x x f (4-3) ) ( / ) ( ) ( xx F (4-4) 0 ) ( 1) ( dz e zz (4-5)

PAGE 109

109 x z xdz e z0 ) ( 1) ( (4-6) The independent variable (x-value) in e quation 4-1 and 4-2 was normalized before regression modeling. The main objective of the modeling proce ss was to optimize parameter and b in the cumulative gamma distribution function. Initial and b values could be set as the values from a statistical estimat e as shown in equation 4-7 where x and s are the sample mean and standard deviation. The statistical method for optimization of parameter and b is to minimize the sum of squared errors (SSE), maximi ze the correlation coefficient, and ensure that the p-values were greater than 0.05 betw een the modeled and measured data. 2 s x, x s2 (4-7) Surface Area And Specific Surface Area (SA/SSA) The modified EGME method (Sansalone et al. 1998) was applied to measure SSA of particles with different size gradations. Granular activated carbon (GAC) se rved as the control. Based on monolayer surface coverage and molecu lar weight of the EG ME molecule, SSA was calculated by Equation 4-8, S aW W SSA 000286 0 (4-8) SSA results were summed over each size grad ation to yield a total surface area (SA) distribution, shown in equation 4-9, ) )( (i i iSSA m SA (4-9) Solubility Test In the experimental matrix for testing the solubility of particles, we examined the particle size, initial pH and residence ti me. The factors between events and between watersheds vary for

PAGE 110

110 different cover conditions. Partic le size was classified into various size ranges [> 9500, 9500 ~ 4750 38 ~ 25, < 25 m], representing the entire gradati on of dry deposition particles. The recorded pH factors at the site were [3, 4, 5, 6, 7] which encompassed the range of pH values for recorded rainfall at the site. Th e residence time factors were [0 5, 10, 20, 40, 60 minutes], in order to examine the solubility of particles and metal species. The initial pH value of rainfall solution was adjusted by 0.01N hydrochloric acid and 0.01N potassium hydroxide to 3, 4, 5, 6, and 7 respectively. Rainfall solutions with differe nt prepared initial pH values were stored separately in sealed refriger ated polypropylene containers. In each experiment, for separate particle sizes 5 labeled 1-L glass beakers were filled up with rainfall solutions of different pH values and the solutions were well-mixed throughout each experiment. Throughout each experiment, TDS, conductivity, pH and temperature were measured and recorded. Immediately before each e xperiment, initial values of these factors were recorded before particulates were added to th e solution. A 1.0 gram dry deposition particle mass of a precise mass was added into each beaker an d the starting time initialized. As solubility of dry deposition particles was examined as a f unction of residence time, conductivity, TDS, pH and alkalinity were measured at interv als of 0, 5, 10, 20, 40, 60 minutes. Results Relationship Between Previous Dry Hours A nd Mass Loading Of Dry Deposition Material Dry deposition flux of particulat es in urban highway areas vari es as a function of previous dry hours (PDH). The relationship between overall dr y deposition mass of different event and the relative previous dry hour could be expressed as an exponentia l function shown in figure 2. While considering particles in differe nt fractions, such as less than 25 m, 25 to 75 m and larger than 75 m, a simultaneous variation of dry de position flux were found with relative

PAGE 111

111 previous dry hours, illustrating a similar exponen tial relationship of ma ss flux and previous dry hour in each size fraction, as shown in figure 2. Granulometric Distribution Of Dry Deposition Particles Dry deposition particles were collected from sampling sites beside highway I-10 during different sampling periods and the granulometric distribution of dry deposition particles was measured throughout their whole size range. Re sults were examined as a cumulative mass distribution (CMD) of pa rticles in each dry de position event, which coul d be described as a cumulative gamma distribution (CGD) f unction. Depicted in figure 3, D50m values were found ranging from 257 to 369 m, and more than 95 % of particles were larger than 75 m for all the events. Since dry deposition solids contained a la rge proportion of coarse pa rticles, they can be easily separated from water by their fast settli ng capabilities when tran sported into rainfall runoff. The mean and standard deviation of sh ape and scale parameters for all the cumulative gamma distribution functions were calculated, shown as (2.187, 0.396) and (185.18, 48.98), indicating that sampling period a nd previous dry hour have little effect on CMDs which made the variation of CMDs insignifi cant between different events. Specific Surface Area (SSA) And Surface Area (SA) For Dry Deposition Solids Throughout the whole size range of dry deposition solids, sp ecific surface area (SSA) of particles increased gradually as particle size de creased, shown in figure 4. The increasing trend indicated that SSA values were well correlated with particle size, and the most likely reason for this is that discrete sand and grit were dominant in dry deposi tion particles, with rare porous structures on their surface contributing to the SSA values. Total surface area (SA) was calculated as a function of specific surface area and the ma ss distribution and result s were shown in the right side of figure 4 wi th PDF curve of SA values. Results i llustrate that most high SA values fell in to the coarse pa rticle gradation with size ranges from 75 to 2000 m, this indicates that the

PAGE 112

112 effect of coarse particles on SA were significant a lthough their SSA values were much lower, compared to the fine particles. Comparison Of Cumulative Mass Distributions and D50m For Urban Rainfall-Runoff And Dry-Deposition Solids Based on the results of CMDs for dry deposi tion and rainfall-runoff particles across the particle size range from larger than 9500 m to less than 25 m, all modeled with a cumulative gamma distribution function in figu re 5, results indicate that the CMD curves are translated to a finer particle size gradation after particles were transported from dry deposition into upstream runoff. While the shape and uniformity are changed to a degree, D50m decreased significantly from 330.7 to 98.6 m. Along the runoff stream, larger size pa rticles settle more easily and were removed from the aqueous phase automatically, which caused the CMD curve of runoff particles to be skewed in favor of the finer sizes, when their path of travel was plotted from upstream to downstream. The downstream CMD and CMD afte r 60 minutes of quiescent settling were investigated across the partic le size range from 250 to 1 m by using a laser particle size analyzer and results are depicted in the same figure. The downstream runoff still contained a large proportion of settleable pa rticles, with around 50 % partic les by mass in downstream runoff greater than 25 m as observed in figure 5. After 60 mi nutes of quiescent settling, suspended particles were dominant in the downstream runof f, leading the CMD curve fall into a finest particle size region with 80 % particles less than 25 m by mass, and D50m value decrease to 13.7 m. Influence Of Initial Ph And Part icle Size On Equilibrium Ph When transported into rainfall-runoff, dry de position solids could signi ficantly change the water chemistry parameters of rainfall solutions. Rainfall solutions with various initial pH value (3, 4, 5, 6, 7) and whole size range particles (18 gradations) were em ployed in this study to

PAGE 113

113 examine the effect of initial pH and particle size on equilibrium pH alkalinity, TDS and conductivity. Equilibrium pH decreased exponentially with increasing particle size, especially in the size range from less than 25 m to 300 m, which have a more significant decreasing trend for the equilibrium pH, compared with the size range above 300 m in figure 6. Equilibrium pH goes constant when the partic le size is larger than 300 m, indicating that particles larger than 300 m have no significant effect on pH value while they are transported into rainfall solutions. Equilibrium pHs of rainfall solutions were al l higher than the corresponding initial pH values throughout the whole rang e of particle sizes, due to the many alkalinity-contributing species dissolved into rainfall solution from dry deposition solids. The maximal increase of pH value was found in the finer size range (less than 25 m) jumping from 3 ~ 7 to 9 ~ 10 respectively, and the minimal increase was obser ved in the opposite size region which only had a difference of 2. Results also illustrated that th e effect of the initial rainfall solution pH on equilibrium pH was influenced by the particle si ze, while in the finer size ranges (such as less than 25 m) equilibrium pHs from five initial pH solutions were observed to increase from 9 to 10, which is a more narrow range than the initial pH range which was from 3 to 7. As particle sized increased, the effect of pa rticle size on equilibrium pH decr eased and made the effect of initial pH appear to be more significant and dominant in the la rger size ranges (larger than 300 m). Influence Of Initial Ph And Particle Size On Equilibrium Alkalinity While transported into rainfall solution, alka linity-contributing speci es in dry deposition solids, such as CaCO3, MgCO3, CaHCO3, MgHCO3, partially dissolve into an aqueous phase and caused the pH and alkalinity to increase signific antly. Figure 7 depicts equilibrium alkalinities influenced by the initial pH of various rainfall solutions and particle sizes for different dry deposition solids gradations. As a result of high sp ecific surface area, fine particles generally had

PAGE 114

114 more contact area in the liquid phase, leading to a higher solubility of alkalinity-contributing species inside these particles. Equilibrium al kalinity in the fine size range (less than 25 m) was found to have the highest value of all th e rainfall solutions, around 40 to 50 mg CaCO3/L respectively. As particle sizes increased, the equilibrium alka linity decreased significantly. Particles above 300 m behaved differently: the alkalinity became constant and fell into a range of 5 to 15 mg CaCO3/L. The decreasing trend of equilibrium alkalinity with particle size could be expressed as an exponential decay function fo r each rainfall solution with specific initial pH, simulated in figure 7. The initial pH of rainfall solution appeared to have no significant effect on equilibrium alkalinity. Influence Of Initial Ph And Particle Si ze On Equilibrium TDS And Conductivity Equilibrium TDS and conductivity was investigat ed under similar controlled conditions of initial pH and particle size gradation, and resu lts are illustrated in figure 8. TDS varies with particle size in an expone ntial trend for all the pH rainfall solutions, jumping from 30 ~ 40 mg/L in fine size range (less than 25 m) to 0 ~ 5 mg/L in coarse size range (larger than 300 m). Results illustrated that TDS values have a str ong correlation with conductiv ity (Sansalone et al. 2005), which shows that conductivity ha s a similar trend with particle size and initial pH, this is illustrated in figure 9. TDS was also found to be correlated with alkalinity since they have a similar trend while varying with initial pH and particle size. Influence Of Contact Time On Ph, TDS, Conductivity At the beginning of the test, the equilibrium st atus was still far away to set up, and there was a strong potential for the so luble material to dissolve into the rainfall solution, leading the TDS and conductivity value to incr ease significantly during the fi rst 4 or 5 minutes. This is shown in figures 11 and 12. The soluble material in dry deposition solids contains a large portion of alkalinity-contributin g species, which could directly a ffect the pH, TDS and conductivity

PAGE 115

115 values if leaching into rainfall solutions The effect of initial pH on the variation of pH, TDS and conductivity could be neg ligible, compared to th e effect of a large am ount alkalinity -contributing species which can easily dissolve into rainfall so lutions and cause a signif icant increase of pH values,. The relationship of pH and cont act time was investigated und er different initial pH and particle size conditions and result s are illustrated in figure 10. Th e plot at the top of the figure show that the pH value increased rapi dly for the particles in size range 25 m, and reached a peak value of around 9 to 10 within 5 minutes. Th is value then held constant as further time elapsed and could be considered as an equilibrium state for pH in rainfall solutions. Medium particles in size range 75 m yielded a clear declining trend of pH value over time after reaching a peak value of 8 to 10, this is shown in the plot in the middle of figure 10. After peaking, the pH value decreased gradually to 7 or 8 after the experimental tim e period of 60 minutes elapsed. A different scenario was found for large particles of size range 180 m, where pH increased comparatively slower in the initi al five minutes, shown in the lower plot of figure 10. After 10 minutes passed, the pH appeared to be constant within the range of 6 to 8. When particles dissolve into rainfall solutions, TDS varies in different ways with different amounts of contact time for pa rticles of size ranges 25 m, 75 m, and 180 m. This is illustrated as 3 different trendlines in figure 11. In the plot at the top of the figure for 25 m particles, TDS increased rapidly at the beginning, and within 5 minutes the value jumped from ~25 mg/L to ~55 mg/L. Then TDS values decrea sed gradually as time elapsed until they went down to the range of 40 to 50 mg/L at the end of the 60-minute period. In comparison, a slowly growing trend was observed fo r particles of size range 75 m, when after 5 minutes they reached the TDS level of 25 mg/L; this is illustrated in th e middle plot of figure 11. Since it is very hard

PAGE 116

116 to release the soluble material in side large-size particles to dissolve into rainfall solutions, TDS values gradually increased with contact time, show n as an exponential trend in the lower plot of figure 11, which kept the increas ing trend even after the60-minut e time period. A similar trend was found for the relationship of conductivity and contact time under different particle size gradation, as a result of the high correlation of conductivity and TDS values in rainfall solutions. This is shown in figure 12. Conclusions Granulometry, transport and sol ubility of dry deposition particle s was investigated in this study and conclusions are summarized as following. Dry deposition flux of particles in different size gradations were found to be correlated with previous dry hours in an exponential growth function during the whole sampling period. Dry deposition particles collected from sampling sites beside highway I-10 was measur ed for each sampling event, which D50m values were in a range of 257 to 369 m, and more than 95 % of particles were found to be larger than 75 m. The granulometric distribution of particles wa s examined for each dry deposition period, and results could be well described by a cu mulative gamma distribution function. Cumulative mass distributions (CMDs) of particle s were translated to a finer particle size gradation after being trans ported from dry deposition into rainfall runoff, while D50m decreased from 330.7 m to 13.7 m along the runoff stream. Since ther e is a high potenti al for soluble material in dry deposition solids to dissolve into rainfall solu tion, the pH, TDS and conductivity values in rainfall solution increased significantly at the beginning of mass transport, reached a peak value within 5 minutes, and then rested in equilibrium for the remainder of the period. Considering the effect of different particle gradation, equilibrium pH, alkalinity, TDS and conductivity of rainfall solutions decreased significan tly while particle size increased, especially

PAGE 117

117 in the size range from less than 25 m to 300 m. Conversely, the initial pH of rainfall had little effect on water chemistry parameters of rainfall solutions. Table 4-1. Summary of partic ulate mass of dry deposition solid s collected at 5 sampling sites during the period from 17 January to 16 July 2004 Sampler ID I-10 Center-a I-10 Center-b I-10 East bound Control a Control b Sampling date Duration time, hour Traffic Volume Particulate mass of dry deposition solids, g 17 January 2004 145.0 4.26E+0567.5155.361669.42 0.06 0.2825 January 2004 92.7 2.69E+0546.0442.16806.97 0.16 0.1909 February 2004 79.2 2.38E+0561.4349.10616.46 0.03 0.0520 February 2004 75.8 2.30E+0572.5060.961016.85 0.07 0.0729 February 2004 82.0 3.23E+0569.0056.69483.12 0.03 0.0404 March 2004 58.0 1.82E+0523.8720.60194.00 0.01 0.0113 March 2004 172.0 5.05E+0588.2066.731134.68 0.05 0.3023 March 2004 166.0 4.85E+0585.2770.54623.56 0.03 0.0128 March 2004 124.0 3.73E+0558.8372.35569.45 0.15 0.4106 April 2004 144.0 4.23E+05106.8988.30812.52 0.21 0.2817 April 2004 90.5 2.60E+0551.6043.17617.94 0.01 0.0223 April 2004 147.0 4.34E+0560.7257.80687.03 0.01 0.0129 April 2004 67.5 1.92E+0532.8332.56314.23 0.10 0.0110 May 2004 172.0 5.10E+0592.3372.38802.15 0.11 0.0430 May 2004 241.0 7.09E+05102.1291.78840.23 0.01 0.0124 June 2004 77.0 2.32E+0545.1346.50282.83 0.01 0.0116 July 2004 102.0 3.04E+0530.3031.17386.95 0.01 0.01

PAGE 118

118 Sampling Period Sampling Duration Time (day) 0 1 2 3 Precipitation (mm) 0 50 100 150 200 Sampling Time Precipitation Jan. Feb. Mar. Apr. May Jun.Jul. Figure 4-1. Dry deposition samp ling duration time and rainfall records during the period from January to July 2004.

PAGE 119

119 Particle loading (g/m2) 0 400 800 1200 1600 Particle loading (g/m2) 0 400 800 1200 1600 M = 1272*(1 e [-0.0085*pdh] ) Previous dry hour (pdh) 050100150200250300 Particle loading (g/m2) 0 5 10 15 20 Previous dry hour (pdh) 050100150200250300 Particle loading (g/m2) 0 5 10 15 20 Total mass < 25 m > 75 m 25 m 75 m M = 1.3*(1 e [-0.0076*pdh] ) M = 13.5*(1 e [-0.0246*pdh] ) M = 1076*(1 e [-0.0122*pdh] ) Figure 4-2. Relationship of pr evious dry hours (pdh) with dry particle mass deposited for urban transportation land use si te in Baton Rouge, Louisiana.

PAGE 120

120 Particle Diameter ( m) 10100100010000 {cdf}: finer by mass, F(D) 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 17 Apr. 2004 23 Apr. 2004 29 Apr. 2004 10 May 2004 30 May 2004 24 Jun. 2004 16 Jul. 2004 13 Mar. 2004 Model 17 Jan. 2004 25 Jan. 2004 09 Feb. 2004 20 Feb. 2004 29 Feb. 2004 04 Mar. 2004 28 Mar. 2004 Figure 4-3. Solid gradations for 17 dry traffi c deposition events at the Baton Rouge site, simulated by a cumulative gamma distribution function.

PAGE 121

121 SSA Particle Diameter, m SSA, m2/g 0 20 40 60 80 100 120 140 Particle Diameter, m SA, m2 0 5e+3 1e+4 2e+4 2e+4 PDF, % 2 3 4 5 6 SA pdf 9500 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 9500 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 Figure 4-4. Specific surface area (SSA) and Surf ace area (SA) for dry deposition solids with different particle size gradation.

PAGE 122

122 Particle Di ameter, D ( m) {cdf}: % finer by mass, F(D) 0 20 40 60 80 100 Dry Deposition q (up) q (down) downstream upstream q (settled) 1 10 100 1000 10000 Figure 4-5. Measured granulometry of runoff (q ) and dry deposition (DD) at the Baton Rouge site.

PAGE 123

123 Particle diameter (mm) 10100100010000 Equilibrim pH 0 2 4 6 8 10 12 14 pH = 6 pH = 7 pH = 3 pH = 4 pH = 5 < 25 25~75 75~300 300~9500 Figure 4-6. Equilibrium pH of solubility test for 17 gradation dry deposition particles in 5 different initial pH solutions

PAGE 124

124 Particle diameter (mm) 10100100010000 Alkalinity (mg CaCO3/L) 0 20 40 60 80 pH = 6 pH = 7 pH = 3 pH = 4 pH = 5 < 25 25~75 75~300 300~9500 Figure 4-7. Equilibrium alkalinity of solubility test for 17 gradation dry deposition particles in 5 different initial pH solutions

PAGE 125

125 Particle diameter ( m) 10100100010000 Equilibrim TDS (mg/L) 0 10 20 30 40 50 pH = 6 pH = 7 pH = 3 pH = 4 pH = 5 < 25 25~75 75~300 300~9500 Figure 4-8. Equilibrium TDS of solubility test for 17 gradation dry deposition particles in 5 different initial pH solutions

PAGE 126

126 Particle diameter ( m) 10100100010000 Equilibrim Conductivity ( s/cm) 0 20 40 60 80 100 pH = 6 pH = 7 pH = 3 pH = 4 pH = 5 < 25 25~75 75~300 300~9500 Figure 4-9. Equilibrium conductivity of solub ility test for 17 gradation dry deposition particles in 5 different initial pH solutions

PAGE 127

127 75 m pH value 0 2 4 6 8 10 12 180 m Time (minutes) 0102030405060 pH value 0 2 4 6 8 10 12 25 m pH value 0 2 4 6 8 10 12 pH = 3 pH = 4 pH = 5 pH = 6 pH = 7 Figure 4-10. pH value as a functi on of time for 3 different sizes of dry deposition particles in 5 different pH solutions

PAGE 128

128 25 m TDS (mg/L) 0 10 20 30 40 50 60 pH = 3 pH = 4 pH = 5 pH = 6 pH = 7 75 m TDS (mg/L) 0 10 20 30 40 50 60 180 m Time (minutes) 0102030405060 TDS (mg/L) 0 10 20 30 40 50 60 Figure 4-11. TDS value as a functi on of time for 3 different sizes of dry deposition particles in 5 different pH solutions

PAGE 129

129 CHAPTER 5 CHARACTERISTICS, TRANSPORT AND SOLUBILITY OF DRY PARTICULATE METALS DEPOSITED ON URBAN HIGHWAY AREA Introduction Heavy metals, transported from ambient and vehicle-related atmospheric deposition into highway runoff, are important contaminants becau se of their significant contributions to storm runoff as well as their persistent and potential toxicity in aquatic systems (Sansalone and Buchberger 1997). There are mainly two kinds of sources for heavy metal contaminants emitted to the atmosphere: natural processes and anthro pogenic activities. Natu ral processes, such as volcanic emissions and weatheri ng of earth-surface materials, contribute particulate-bound trace metals to the atmosphere (Motelay-Massei 2005 ). Anthropogenic activit ies include combustion of municipal solid waster and fossil fuels in power plants, re leases from metal smelters, motor vehicle emissions and industrial st ack emissions (Shaheen 1975). In urban highway area, concentrations of heavy metal contaminants from atmospheric deposition particles were found d ecreasing with increasing distan ce from highway surface, which indicates that highway are large sources of these contaminants re lative to atmospheric deposition (Harned 1988). In other research, it was found th at motor vehicles played a major role in pollutant emissions to the atmosphere and also ha d a significant effect on air and water quality (Bullin and Moe 1982; Ellis et al 1987). The most-often cited heavy metal contaminants are Cd, Cr, Cu, Fe, Ni, Pb, and Zn, which can adversely a ffect receiving waters by in creasing toxicity in the water column and/or sediments, and bioaccu mulation in the food chain (Yousef et al. 1987). Urban highways have a large, well drained and impervious ar ea where atmospheric deposited contaminants can be easily washed off and tran sported into urban rainfa ll runoff during a storm (Colman 2001). Influenced by factors such as pH, redox conditions, alkalinity, residence time,

PAGE 130

130 complex agents and suspended solids, these heavy me tals can be preferentially dissolved in ionic form where they can exert an immediate toxi city impact (Sansalone and Cristina 2004). As a source of pollutants to urban rainfa ll-runoff, dry deposition may be especially important because of significant amount of trace metals and other pollutants emitted into the atmosphere daily (Sabin et al. 2005). However, the ultimate fate of the trace metals in dry deposition particles was rarely investigated, especially in urba n highway area, and few studies focused on the contribution of metals from dr y deposition to urban rainfall-runoff. With a purpose to illustrate the impact of metals from dry deposition pa rticles on urban environs, a study was employed to examine characteristics, transp ort and solubility of dr y deposition particulate metals at Baton Rouge hi ghway transportation area. Objective In order to investigate the tr ansport mechanisms and contribu tion of metal species from dry deposition particles to urban rainfall runoff, th ree objectives were carried out in this study. The first objective was to examine characteristics of metal species in dry deposition solids which were collected at the I-10 site, Baton Rouge fr om January to July 2004. The second objective is to investigate the solubility of metal species from dry deposition particles in urban rainfallrunoff. Third objective is to examine the cumu lative mass distribution of total and dissolved metal species for dry deposition part icles and model these distributions. Background Study Dry deposition rates of heavy metals varied si gnificantly in different areas. For example, in recent research by Paode et al. (1998), the aver age measured Pb, Cu, and Zn fluxes were 0.07, 0.06, and 0.20 mg/m2/day in Chicago (urban areas) and 0.004, 0.007, and 0.004 mg/m2/day in South Haven (rural areas). The dry deposition rat es of heavy metals were extremely higher in urban areas than that in rural areas (Paode et al. 1998). In N.W. London, both dry deposition and

PAGE 131

131 gutter sediment samples were collected from roads and analyzed separately for concentrations of Cd, Cu, Fe, Mn, Pb and Zn. Results suggested th at concentrations of Cd and Pb reached the highest level for coarse particle s with size range from 100 to 500 m during the rush hour on highway (Ellis and Revitt 1982). Dry deposition particles, recove red from a pavement section at I-75 in Cincinnati, were analyzed for heavy me tals. Concentrations increased with decreasing particle size for Zn, Cd, Pb and Cu, compared with metal mass which are not dominant in fine particles, but coarse and mid-ra nge particles (Sansalone and Tri bouillard 1999). A study taken at a small city in Korea, investigated dry depositio n fluxes of ambient particulate heavy metals in 2000 (Yun et al. 2002). It indicates that the average fl ux of metals during the daytime was higher than the nighttime caused by higher wind speed a nd large ambient contributions. The average flux of Al and Ca, typical crusta l species, were 1 to 2 orders of magnitude higher than Mn and anthropogenic metals, such as Mn, As, Cd, Cu, Ni, Pb and Zn (Yun et al. 2002). Atmospheric bulk deposition of traces to the Seine river basi n in Paris was inves tigated during 1998 to 2001 (Motelay-Massei et al. 2005), which showed that concentrations of heavy metals in atmospheric deposition followed the order of Zn > Pb > Cu > Ni > Cd for both dissolved and particulate form. Concentrations of Cd, Cu, Pb and Zn in atmosphe ric wet deposition was investigated at a coastal station in western Europe, and a constant ra nking was found as Zn >> Pb > Cu >> Cd for 4 sampling sites (Deboudt et al. 2004). The amount of heavy metals c ontributed by vehicles or ambien t atmospheric deposition to urban highway runoff is still a hot topic in r ecent researches. In Irwin and Loseys (1978) research, ambient bulk deposition contributed 18 % for Pb to 38 % for Cu to the bridge runoff at a highway bridge in Florida. Other research shows that major ions (Na+, K+, Mg2+, Ca2+, Cl-, and SO4 2) from precipitation loads into highway runoff were only 2 % in average (Bellinger et al

PAGE 132

132 1982). Dry deposition was a dominant source of h eavy metals in urban areas compared with wet deposition. Concentration of Pb from dry deposition was 0.13 mg/m2/day measured from June to October 1991 in Chicago (Lin et al. 1993), and 0.0029 mg/m2/day from wet deposition in the same area in 1992 (Vermettea et al. 1995), which is 45 times different. In semi-rid regions like Los Angeles, dry deposition was recognized as a dominant mechanism for transfer of atmospheric pollutants to watershed surfaces becau se of the low rainfall quantity in this area (Sabin et al., 2005). Average daily flux of dry deposition ( g/m2/day) in Los Angeles being Cr 1.3, Cu 9.4, Pb 5.8, Ni 3.7 and Zn 39 with the rate of dry deposition to st orm runoff is 0.69, 0.70, 1.07, 1.06 and 0.52 respectively. Results suggest that dry depos ition may contribute 52-100 % of heavy metal loads to annual stormwat er discharge (Sabin et al. 2005). Bricker, in one of his res earches (1999), found that many heavy metals in atmospheric deposition were in bio-availabl e dissolved form. Chemical reac tivity of atmospheric deposition on the highway surface can be affected by waterquality properties, such as pH and specific conductivity (Bricker 1999) In the eastern United States, atmospheric deposition is usually acidic, particulate (Rice and Bricker 1992; Bricker 1999), which could enhance the geochemical mobility of heavy metal contaminants (Harned 1988). Dry deposition particles have reactive sites and relative large surface-to-volume ratio where heavy metals can be released or adsorbed from particles to or from highway runoff (Sansalone et al. 1998). Leaching ca pability of heavy metals was independent with particle size and pH value of highway runoff. Extraction effici encies for the five metals from dry deposition particles into urban rainfall runoff are found to be independent of ro ad type and in the order Cd > Zn, Cu > Mn > Pb. Pb may concentrate in sedimen t of urban runoff and cant be easily released from particulate-bound fraction into dissolved frac tion, compared with other metals (Ellis and

PAGE 133

133 Revitt, 1982). Similar results were provided by research for leaching behavior of heavy metals at the edge of A-71 motorway in Sologne. It is f ound that extraction efficiencies of Cd, Zn and Pb were ranked as an order of Cd>Zn>>Pb in car bonate-free systems during a moderately acid rainstorm. If considering Fe and Mn, the mobility rank becomes Mn ~= Cd > Zn >> Pb > Fe, for artificial soils as well as for suspensions, a lthough there was a big di fference in heavy metal levels from those two sources (Lee and Touray, 1998). There are several factors which may affect the mobility of metal extraction from dry deposition into storm water runoff, such as pH value, ion exchange, ae ration and agitation, and metal bindings. As found in several researches, so lubility of most heavy metals in given soils was inversely related to pH value (Sinha et al., 1978; Mc Bride and Blasiak, 1979; Reddy and Patrick, 1977). An experiment taken by Nai du et al. (1998) examin ed the adsorption and desorption of heavy metals (Pb, Cu, Ni and Zn) from pH-adjusted soils. Results indicated that extraction rates of heavy metals were depende nt upon the pH of soil samples with retention ability of heavy metals signifi cantly increasing above a pH value equal to 7.0 to 7.5. Soils usually have specific adsorption sites for Pb, Ni and Cu but little or none for Zn, based on subsequent extractabili ty (Naidu et al. 1998). Methodology Dry Deposition Sampling Three sampling sites were chosen in this re search by considering di fferent sources for dry deposition contaminants at urban highway area, such as materials swept and resuspended by natural wind and highway traffi c, and vertical gravitational se ttling from natural and ambient sources. The first site was located on the road shoulder next to the bridge railing, adjacent to the traffic lane of I-10 East at C ity Park Lake, Baton Rouge. A 1-m2 PVC sampler with 4-inch edges was set up at the first site as same level of th e traffic lane, without any embankment between

PAGE 134

134 them except the sampler edge, making it easier an d more efficient to collect particles. In addition, the shortest distance from sampler to traffic lane made material collected from this site more accurate and representative to illustrate the effect of highway traffic to dry deposition contaminants in urban highway area. The second s ite was beside a pier, with two samplers set up on frame work between I-10 East an d West bridges. Samplers were suspended in the air, parallel to the driving lane and 5 feet lower from the top of the bridge railing. Particles resuspended by highway traffic were embanked by the bridge rai ling if they didnt reach the top of railing and jump over it. In this case, only small amounts of resuspended particles could be collected by the sampler and gravitational settling particles ge nerated from anthropogenic and natural sources became a dominant part in the samples from this s ite. The third site was located on the roof of Annex Laboratory, 50 meters aw ay from Nicholson Dr. Extens ion, with two samplers, as a control for background investigation. There were a total of 17 collection events for dry depositions dur ing the sampling period from January 17th to July 16th, 2004. When four or more dry-w eather days appeared ahead of storms in the weather for ecast, samplers were set up at each si te and started collecting. Solids in each sampler were retrieved by a soft brush and stainless iron shovel right before the storm came and saved in a glass sample jar with known weight. All sampl es were transferred back to the lab and stored separately be fore employing any further measu rements on them. Samples from one event had to be combined af ter characteristics an alysis. Otherwise the potential mass-limited issue would show up, affecting the results accuracy in future experiments. Following standard method ASTM D422-63 (ASTM 1990), samples of dry deposition solids were sieved by 17 different size sieves sequentially into 18 size gradations (from >9500 to <25 m) and each size particles were saved in a la beled sample jar separately.

PAGE 135

135Solubility Test For Metal Species In Dry Deposition Solids A solubility test was designed to investigate the transportation mechanism of metal species from dry deposition materials into urban rainfa ll runoff, as well as th e contribution of dry deposition on metal species in urban rainfall runo ff. In this test, a total of 18 experiment sets were taken to examine the effect of particle size on the solubility of metal species where one set represents one particle size gradation. Five ra infall solutions with different initial pH value (from 3 to 7) were chosen which represent all pos sible pH rainfalls on record. Initial pH values of rainfall solutions were ad justed by 0.01N hydrochloride ac id and 0.01N potassium hydroxide to 3, 4, 5, 6, and 7 respectively. 1 liter ra infall solution of each pH value was taken out and dropped into a labeled 1L glass beaker. The so lution was well mixed continuously throughout the experiment. 1 gram particles from each size dry deposition solid were added into the solution and timed (start point is Time 0). At time 0, 5, 20, 40, 60 mi nutes, 20 mL of solution was sucked out from each beak er and fractionated by a 0.45 m membrane filter. The filtrate was acidified, diluted and combined with internal st andard before it went in to ICP-MS analysis. Total/Dissolved Mass And Concentration Of Metal Species In Dry Deposition Solids Inductively coupled plasma-m ass spectroscopy (ICP-MS) was u tilized to analyze metal species in dry deposition samples with different size gradation. 10 me tals were concerned in this study, Cu, Cd, Pb, Zn, As, Cr, Fe, Mn, Ca and Mg which are all measurable in ICP-MS. As the measurement range limit of ICP-MS equipment, 5 mL digested solution for each size particles was diluted to 10 mL by 5 % nitric acid and combined with 100 L internal standard in a 15 mL centrifuge tube. QA/QC was require d for every ten measurements of ICP-MS analysis in every experiment set. Since the ICP-MS equipment can only be utili zed on dissolved metal samples, particulate bound metal species in dry deposition solids had to be wholly transferred into dissolved phase

PAGE 136

136 for ICP-MS analysis. For each size dry deposition sample, 0.5 to 1 gram particles were prepared for acid digestion, while a blank and standard sample was added to each set of acid digestion as QA/QC requirements. Metal acid digestion pr ocedure followed standa rd method 3030F (1995) which was modified in this research to improve ex traction efficiencies of me tal species. Digested solutions were filtrated, diluted by 2 % nitric acid to 100 mL, and well kept for the future experiment. The results obtained from ICPMS analysis dont directly illustrate the real mass or concentration of metal species in dry deposition so lids. This is due to the necessity of several steps for the preparation of ICP-MS samples from dry deposition solids. Each procedure needed to be considered very carefully while doing calculations, fo r even one small error, such as a wrong unit conversion, in the sequence would ca use a big mistake in the final results. The calculation of concentrati on of metal species (unit: mol/L) is shown in equation 5-1. 1000, , i j digested j i j i digested j i ICP j i metalMass Molar M D V C C (5-1) Total mass of metal i in dry deposition solids with different size ranges were calculated by equation 5-2. j j i metal j i metalM C M , (5-2) The function of cumulative mass distribution for metal i was expressed by equation 5-3. % 100 ,%18 1 1 j j i metal j t t i metal j iM M mass cumulative (5-3) Concentration of dissolv ed metal species (unit: mol/m2) in dry deposition solids was calculated by equation 5-4.

PAGE 137

137 j i j solid j i j i solution j i D ICP j i D metalSSA Mass Molar M D V C C , , (5-4) Total dissolved mass and cumulative dissolv ed mass distribution followed the same calculation method shown in equation 5-2 and 5-3. Modeling Of Cumulative Mass Distribution For Metal Species Cumulative mass distribution of metal species in dry deposition solid s shows a very good trend to fit with a cumulative gamma distribu tion model. The general formula for probability density function (PDF) of gamma distribu tion is expressed in equation 5-5 where a represents a shape parameter and b is a scaling parameter. ) ( / ) (/ 1a b e b x x fb x a (5-5) The cumulative distribution function of gamma distribution is given in equation 5-6 where a is the gamma function and x a is the incomplete gamma func tion, expressed as equation 5-7 and 5-8 respectively. ) ( / ) ( ) ( a a x Fx (5-6) 0 ) ( 1) (dt e t at a (5-7) x t a xdt e t a0 ) ( 1) ( (5-8) The independent variable (x value) in equa tion 5-5 and 5-6 should be normalized before running the regression model. Th e main point of modeling process is to optimize parameter a and b in the cumulative gamma distribution function. Initial a and b values could be set as the values from statistical estimation sh own in equation 5-9 and 5-10, where x and s are sample mean and standard deviation.

PAGE 138

138 2 s x a (5-9) x s b2 (5-10) The statistical method for optimization of parameter a and b is to minimize the sum of squared errors (SSE), maximize the correla tion coefficient, and make sure that p -values is greater than 0.05 between the modeled and measured data. Results Mass And Concentration Distribution Of Met al Species In Dry Deposition Particles Metal concentration distributi ons across size gradation for different metal species were measured and results illustrated different scenar ios in Figure 1 and Figur e 2. The concentration of Cu has a steadily increas ing trend with particle si ze decreasing, less than 2.4 mol/g for particles larger than 75 m, 2.4 to 3.1 mol/g in the size range of 25 to 75 m, and average 5.2 mol/g with size less than 25 m. The concentration of Cd has no significant variation trend for particles larger than 106 m, which was mostly in the range of 0.002 to 0.007 mol/g, except that for 4750 m particles. The highest concentration of Cd, found at the size gradation of 75 m, was equal to 0.029 mol/g, which was almost 3 times hi gher than the average value of concentrations in the size range less than 75 m. The concentration of Cd appeared to be relatively constant in the particle size range less than 75 m, varying from 0.008 to 0.016 mol/g. The concentration of Pb was found at a relatively low level (less than 0.2 mol/g) for particles larger than 425 m, compared with that in the ot her size gradation. It increased significantly in the size range from 180 to 75 m, with concentration va lue jumping from 0.3 to 4.5 mol/g. A relatively constant value of Pb con centration was depicted as an average of 1.1 mol/g in particle gradations less than 75 m. Except particle size of 425 m and 600 m, Zn

PAGE 139

139 concentration for particles larger than 150 m has a steadily increasing trend with values from 0.2 to 3.4 mol/g. The maximal Zn concentration was in the range of 5.8 to 7.6 mol/g, found for particles with size between 45 and 106 m only. For particles less than 45 m, the concentration of Zn was ge nerally constant, around 5.0 mol/g. The concentration of As was lower than 0.012 mol/g for particles larger than 4750 m, between 0.025 and 0.043 mol/g for particles with size from 75 to 2000 m, and around 0.06 mol/g for particles less than 75 m. The concentration of Cr was relatively lower for particles larger than 2000 m, illustrated as less than 0.16 mol/g, and in the range of 0.28 to 0.80 mol/g for particles less than 2000 m, where it was over 0.6 mol/g for specific size gradations, including 800, 106, 75, and 25 m. The relatively lower concentration of Fe (less than 25 mol/g), was also found for particles larger than 4750 m. The highest concentration of Fe was equal to 463 mol/g, which appeared to correspond with particles in size range between 600 and 800 m. While for particles less than 600 m, the concentration of Fe varied in a narrow range from 175 to 296 mol/g. The concentration of Mn has a si milar pattern to Fe across the entire size gradation, illustrated as less than 0.4 mol/g for particles larger than 4750 m, maximum of 5.95 mol/g for particles of 600 m, and relatively an equal value of 4.0 mol/g for particles less than 600 m. Total metal mass in particles from each pa rticle size gradation was calculated by the corresponding metal concentration, multiplying th e particle mass of this gradation separated from the overall particle mass collected during the entire sampling period. Cumulative mass of metal species were also examined and results were described by an optimized cumulative gamma distribution function respectively, summarized in Figure 1 and Figur e 2. Results illustrated that mass of Cu was dominant in dry depos ition particles with size range from 75 m to 800 m, with

PAGE 140

140 maximum of 0.0285 g out of 0.8506 g of total mass for particles of 300 m, and approximate 95 % of total mass was contributed by particles larger than 75 m. The mass of Cd was shown as a relatively equal distribution in the size range between 75 m to 800 m, which contributed totally 0.0053 g from the overall mass of 0.0060 g across the entire size gr adation. The mass of Cd from particles less than 75 m was less than 5 % of total ma ss. The mass of Pb had two peak values, 0.31 and 0.35 g for particles of 300 and 106 m respectively, and less than 10 % of total mass was contributed by part icles larger than 300 m. While for particles less than 75 m, the corresponding mass of Pb was only 3 % of total mass, indicating that Pb is more likely to be correlated with coarse particles larger than 75 m. The mass of Zn illustrated a parabolic trend across the whole size range, and 97 % of total ma ss was from particles w ith size between 75 and 2000 m, with maximal value of 0.6 g for particles of 425 m. A similar trend of metal mass distributed across the size grada tion existed for metal species, such as As, Cr, Fe, and Mn, where metal mass was less from pa rticles larger than 2000 m, and the maximal mass was correlated with particles of 800 m. Metal mass of As, Cr, Fe, and Mn dropped significantly with particle size decrease for part icles less than 800 m, in particular, considered to be negligible from particles less than 75 m. Transport And Solubility Of Metal Species In Dry Deposition Particles While transported from dry deposition into ra infall solution, particles in different size gradation illustrated various capabilities of l eaching metal species into dissolved phase. Equal amount of particles from each size gradation was employed to test th e solubility of metal species into rainfall solutions, describe d as dissolved molar mass (DMM: mol/m2) in Figure 3 and Figure 4 for Cu, Cd, Pb, Zn, As, Mg, Fe, Ca resp ectively. Results illustrated that DMM of Cu distributed as a parabolic trend across the entire size gradation, dominant in the size range from 106 to 600 m with values of 0.0015 to 0.0023 mol/m2, around 0.0010 mol/m2 for particles

PAGE 141

141 less than 75 m, and decreasing significantly with particle size increase for particles larger than 600 m. For dissolved molar mass of Cd, two p eak values were found throughout the whole particle size range, one is equal to 9.2e-5 mol/m2 at the gradation of 150 m, and the other is 1.6e-4 mol/m2 for particles less than 25 m. DMM of Cd were all less than 5.0e-5 mol/m2 for particles in the rest of gradations. Solubility of Pb for dry deposition particles had an inverse proportional relationship with partic le size gradation, shown as an increase of DMM of Pb with particle size decreasing while particles are larger than 75 m. Two peak values were determined as 4.6e-4 and 4.7e-4 mol/m2 for particles in the size range of 75 m and 106 m, and the DMM of Pb was relatively constant for particles less than 75 m, around 2.0e-4 mol/m2. The DMM of Zn illustrated a different scenario of the relati onship with particle size gradations. A steadily increasing trend of DMM was found with particle size decrease for particles larger than 250 m, reaching the peak value of 0.01 mol/m2 at the gradation of 250 m. The DMM of Zn significantly decreased and kept a relatively low level through size ranges from 180 to 53 m, which was less than 0.004 mol/m2 respectively. The DMM value increased significantly in the finer size gradations (less than 45 m) with the maximal value of 0.018 mol/m2 for particles less than 25 m. Since most of the DMM of As varied in a narrow range of 3.7e-5 to 4.7e-5 mol/m2, no significant distribution trend was depicted across the entire size gradation. The two highest DMM of As value of 6.9e-5 and 6.8e-5 mol/m2 occurred for particles of 75 m and less than 25 m, while the lowest value was 2.45e-5 mol/m2 for particles larger than 4750 m. Compared with other metal species except Ca, the leaching capability of Mg for dry deposition particles was relatively larger, with most DMM values in the range of 0.05 to 0.1 mol/m2. The DMM of Fe was at a high value level of 0.03 to 0.04 mol/m2 for particles in the coarser size range from

PAGE 142

142 300 to 2000 m, but found to have no significant trend in the rest size grad ations, varying from 0.002 and 0.022 mol/m2 randomly. Found to be relative cons tant for particles less than 180 m, the DMM of Ca was mostly in the range of 1.0 to 2.0 mol/m2, and a steadily increasing trend was depicted for particles larger than 180 m, with values jumping from 0.5 to 2.3 mol/m2. By considering the mass distribution of dry de position particles, the DMM of metals across the entire size gradation could be described as a gamma distribu tion trend, where the cumulative DMM of metals was simulated by a cumulative gamma distribution function and results were depicted at the right side of Fi gure 3 and 4. Results indicated that more than 95 % of metal mass was released from particles larger than 75 m into rainfall solutions and the contribution from particles less than 75 m was negligible for all the metal sp ecies. The total dissolved metal mass was calculated based on an entire mass of 12. 586 kg for dry deposition particles collected throughout the full sampling period, shown on the plots for each corresponding metal specie. After being transported into rainfall-runoff, cumulative mass distri bution of particulate metal species translated to a finer size range of dry deposition particles, and both of them could be described by a cumulative gamma distributio n function. Results show that 0.1052 g Cr remained in the particulate phase during the transport while total 0. 2765 g Cr was bonded with dry deposition particles in the original situation. After the transport, the proportion of Cr increased from 5 % to 20 % fo r particles less than 75 m. For Cd, there was 4.218 mg out of 5.977 mg transported in to rainfall-runoff and remaining in the particulate phase, with d50m of Cd decreasing from 300 to 170 m and the proportion increasing from 10 to 25 %, in the fine size range which is less than 75 m. Arsenic was found to have the highest d50m of 450 m for particles before th e transport and 240 m after that, compared with the other 5 metals in Figure 5. 0.0134 g of As was transported into the part iculate phase of runoff from dry deposition phase

PAGE 143

143 which contained 0.0340 g As originally. Althoug h the cumulative mass distribution of Pb was similar with that for other metals during the tr ansport, there was only 0. 1501 g transported into the particulate phase of rainfall-runoff, while total 1.384 g Pb was bonded with dry deposition particles initially. d50m of Pb decreased from 290 to 170 m after the transport. Particulate Cu measured in runoff was 0.1760 g out of 0.8506 g of total Cu mass bonded with particles transported from dry deposition into rainfall-runoff, with d50m decreasing from 280 to 170 m. The proportion of Cu in particles less than 75 m increased from 5 % to 25 % after the transport. The cumulative mass distribution of Zn was also investigated and results illustrated that there was 1.597 g Zn transported into the particulate phase of runoff, corres ponding to the total 2.559 g Zn in dry deposition particles. An increasing proportion of Zn was found for particles less than 75 m which rised from 5 % to 20 % after the transport, and the d50m decreased from 350 to 180 m. Influence Of Ph On Transport Of Metal Species In Dry Deposition Particles A large variation of dissolved metal mass wa s found in Figure 6 for particles transported from specific gradation into rainfall with initial pH in a range of from 3 to 7. However, it is difficult to employ any equation to describe the relationship of initial pH and dissolved metal mass because large amount of alkalinity-contribu ting material found in DD particles generally has an opposite effect on the tr ansport of metal species. Rela tive incremental dissolved mass ranges of metals leached from dry deposition partic les in rainfall were depicted as box plots in Figure 6 at an equilibrium partitioning time of 60 minutes. Each box plot illustrates the effect of a range of initial pH of rainfall (3, 4, 5, 6 and 7) on the dissol ved mass loading for a specific size gradation. Results are based on 1000 g of DD particles and determined based on the mass proportion represented by each particle size with respect to the entire DD particle gradation. Cumulative mass distributions of dissolved metals were i nvestigated and modeled by a

PAGE 144

144 cumulative gamma distribution fit to the arithme tic mean of the incremental results. The influence of initial pH on dissolved metal concentration (expressed as g/L) in rainfall was also investigated at an equilibrium partitioning time of 60 minutes by transporting 2.0 gram DD particles from each size gradation into rainfa ll. The initial pH has no significant effect on dissolved metal concentrations for Cd, As, Mn Fe, Pb, Mg, Cu, Zn across the entire size gradation, although a random vari ation caused by the initi al pH of rainfall was found in some specific gradation. Conclusions Characteristics, transport and solubility of me tal species was investig ed for dry deposition particles collected from I-10 e xperimental site at Baton R ouge. Arithmetic mean of metal concentrations across the entire size gradation in dry deposition partic les was ranked as Cd < As< Cr < Cu < Pb < Zn < Mn < Mg < Fe < Ca, with value range from 0.5 g/g for Cd to 26092.0 g/g for Ca, all summarized in table 1 respectiv ely. Cumulative mass distributions (CMDs) of particulate metals were well described by a cumulative gamma distribution, and CMDs translated a finer size gradati on after transported from dry de position into rainfall-runoff. The proportion of metal mass correlated w ith fine particles (less than 75 m) increased from 5 % to 20 % during the transport, and d50m of metals decreased from 300-450 m to 150-220 m for most metal species. Results indicated that 0.105 g Cr went into pa rticulate phase from dry deposition with total Cr mass of 0.276 g after transported into rainfa ll-runoff, while 0.004 g Cr was found in dissolved phase at an equilibrium partitioning time of 60 minutes. 4.22 mg and 1.00 mg of Cd was transported into particulate and dissolved phase respectively from DDPS with total Cd mass of 5.98 mg. As has a similar transport capability as Cd, where 13.40 mg and 0.80 mg were found in particulate and dissolved phase re spectively during the transport of total As mass of 34.00 mg in

PAGE 145

145 dry deposition particles. There is only 0.150 g Pb transported into particulate phase of runoff from total mass of 1.378 g, while 0.012 g Pb was found in dissolved phase. 0.176 g and 0.044 g of Cu out of 0.851 g total mass was found in part iculate and dissolved phases respectively during the transport, while 1.597 g Zn was transported into particulate phase and 0.162 g Zn went into dissolved phase from dry deposition pa rticles (toal Zn mass of 2.599 g). Dissolved mass of metals was found to follow a similar ranking of As < Cd < Pb < Cu < Zn < Mg < Fe < Ca. The slight change of order indicated varying leaching capability of different metals. The initial pH of rainfall has no significa nt effect on dissolved metal concentrations for most metal species during the transport, even though a random variation was found to be caused by the initial pH in some specifi c gradation for dissolved metals.

PAGE 146

146Table 5-1. Concentration of metal species in dry deposition Particles and transport of metal species from dry deposition parti cles into rainfall-runoff, compared with metal con centrations in dissolved a nd particulate phase of rain fall-runoff at Baton Rouge, LA. Metal species Cr Cu Zn As Cd Pb Fe Mn Ca Mg CMC [ g/L] 550(III)/ 15(VI) 17 110 360 4 65 N/A N/A N/A N/A Dry deposition x 23.9 326.3363.75.31.1272.5 16916.5207.821066.22455.4<25 m [ g/g] 0.5 2.53.30.80.14.2 192.72.2144.753.8 x 33.9 160.0430.34.51.2253.8 12655.1240.221548.02298.025 -75 m [ g/g] 0.5 1.34.60.80.12.1 168.73.4262.031.7 x 21.2 65.5198.02.30.4147.7 15576.4212.826165.81362.0> 75 m [ g/g] 0.5 1.32.00.80.12.1 222.33.4355.117.3 x 21.7 66.8202.03.00.5149.8 15533.6212.826092.01377.1Total [ g/g] 0.5 1.32.60.80.12.1 223.43.4353.417.5Rainfall x 0.5 2.09.20.30.10.2 20.22.128.62.2Dissolved [ g/L] 0.1 0.50.30.10.10.1 1.50.42.80.2Runoff Volume based x 2.1 11.188.61.30.71.7 3820.191.222808.1860.3Dissolved [ g/L] 2.4 6.748.70.41.01.8 795.353.112447.5426.4 x 53.8 88.7737.58.21.972.1 10739.7215.26588.4848.4Particulate [ g/L] 19.1 16.0192.98.21.06.1 1376.554.04428.8777.5Runoff Mass based x 1.1 5.846.20.70.40.9 1991.047.511887.6448.4Dissolved [ g/g] 1.3 3.525.40.20.50.9 414.527.76487.6222.2 x 28.0 46.2384.44.31.037.6 5597.5112.23433.9442.2Particulate [ g/g] 10.0 8.3100.54.30.53.2 717.428.12308.3405.2 x 29.1 52.0430.65.01.438.5 7588.6159.715321.5890.6Toal [ g/g] 9.6 7.892.53.70.53.1 638.028.05551.0313.1 CMC: Criteria Maxium Concentration x : mean and standard deviation fd: dissolved fraction Total: dissolved + particulate

PAGE 147

147 Cu 0.0 2.0 4.0 6.0 8.0 10.0 Cu 0.000 0.100 0.200 0.300 0.400 0.500 0 20 40 60 80 100 Total Mass = 0.8506 g = 9.2377 = 0.0462 R2 = 0.99 Cd 0.00 0.01 0.02 0.03 0.04 0.05 Pb 0.00 1.00 2.00 3.00 4.00 5.00 Zn 0.0 2.0 4.0 6.0 8.0 10.0 9500 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 Cd 0.000 0.001 0.002 0.003 0.004 0.005 0 20 40 60 80 100 Total Mass = 0.0060 g = 4.5651 = 0.0906 R2 = 0.99 Pb 0.000 0.200 0.400 0.600 0.800 1.000 0 20 40 60 80 100 Total Mass = 1.3784 g = 11.691 = 0.0428 R2 = 0.98 Zn 0.000 0.200 0.400 0.600 0.800 1.000 0 20 40 60 80 100 9500 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 Total Mass = 2.5592 g = 6.4511 = 0.0602 R2 = 0.99 Particle Diameter, m Particle Diameter, m Figure 5-1. Mass and chemical concentration di stribution of Cu, Cd, Pb and Zn across size gradation of dry deposition particles. Mass of metal species, g Cumulative mass, % Chemical concentration, [ mol/g]

PAGE 148

148 As 0.00 0.02 0.04 0.06 0.08 0.10 As 0.000 0.002 0.004 0.006 0.008 0.010 0 20 40 60 80 100 Total Mass = 0.0340 g = 4.6080 = 0.0741 R2 = 0.99 Cr 0.00 0.20 0.40 0.60 0.80 1.00 Fe 0 100 200 300 400 500 Mn 0.0 2.0 4.0 6.0 8.0 10.0 9500 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 Particle Diameter, m Cr 0.000 0.020 0.040 0.060 0.080 0.100 0 20 40 60 80 100 Total Mass = 0.2765 g = 4.7845 = 0.0758 R2 = 0.99 Fe 0.0 20.0 40.0 60.0 80.0 100.0 0 20 40 60 80 100 Total Mass = 197.09 g = 5.1885 = 0.0631 R2 = 0.99 Mn 0.000 0.200 0.400 0.600 0.800 1.000 0 20 40 60 80 100 9500 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 Total Mass = 2.655 g = 5.2606 = 0.0640 R2 = 0.99 Particle Diameter, m Figure 5-2. Mass and chemical concentration distribution of As,Cr, Fe and Mn across size gradation of dry deposition particles. Mass of metal species, g Cumulative mass, % Chemical concentration, [ mol/g]

PAGE 149

149 SSA Particle Diameter, m SSA, m2/g 0 20 40 60 80 100 120 140 Particle Diameter, m SA, m2 0.0 5.0e+3 1.0e+4 1.5e+4 2.0e+4 2.5e+4 PDF, % 2 3 4 5 6 SA pdf 9500 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 9500 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 Figure 5-3. Specific surface area (SSA) and Surface area (SA) for dry deposition particles across the entire partic le size gradation.

PAGE 150

150 Cd dissolved mass, mg 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Cumulative mass, % 0 20 40 60 80 100 TDM = 1.0 mg = 13.65 = 0.034X Axis 3 Cu dissolved mass, mg 0 3 6 9 12 15 18 Cumulative mass, % 0 20 40 60 80 100 TDM = 44.3 mg = 7.041 = 0.055 Pb dissolved mass, mg 0 1 2 3 4 5 6 Cumulative mass, % 0 20 40 60 80 100 TDM = 12.4 mg = 8.654 = 0.055X Axis 3 Zn dissolved mass, mg 0 10 20 30 40 50 60 Particle diameter, m Cumulative mass, % 0 20 40 60 80 100 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 TDM = 162.5 mg = 8.685 = 0.042 Cu Cu [ mol/m2] 0.000 0.002 0.004 0.006 0.008 Cd Cd [ mol/m2] 0.00000 0.00005 0.00010 0.00015 0.00020 Pb Pb [ mol/m2] 0.0000 0.0002 0.0004 0.0006 0.0008 0.0010 Zn Zn [ mol/m2] 0.000 0.005 0.010 0.015 0.020 0.025 0.030 Particle diameter, m4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 Figure 5-4. Leaching capability of metals (C d, Cu, Pb and Zn) for dry deposition (DD) particles across entire size gradation at an equilibr ium partitioning time of 60 minutes in rainfall.

PAGE 151

151 Mg dissolved mass, mg 0 50 100 150 200 250 Cumulative mass, % 0 20 40 60 80 100 Fe dissolved mass, mg 0 50 100 150 200 250 X Axis 3 Cumulative mass, % 0 20 40 60 80 100 Ca dissolved mass, g 0 2 4 6 8 Cumulative mass, % 0 20 40 60 80 100 As dissolved mass, mg 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Cumulative mass, % 0 20 40 60 80 100 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 TDM = 0.8 mg = 8.046 = 0.035 TDM = 837.9 mg = 5.062 = 0.071 TDM = 642.4 mg = 6.985 = 0.046 TDM = 21.58 g = 6.575 = 0.061Particle diameter, m Fe Fe [ mol/m2] 0.00 0.02 0.04 0.06 0.08 0.10 As As [ mol/m2] 0.00000 0.00002 0.00004 0.00006 0.00008 0.00010 Mg Mg [ mol/m2] 0.00 0.05 0.10 0.15 0.20 Ca Ca [ mol/m2] 0.0 0.5 1.0 1.5 2.0 2.5 3.0 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25Particle diameter, m Figure 5-5. Leaching capability of metals (A s, Fe, Ca and Mg) for dry deposition (DD) particles across entire size gradation at an equilibr ium partitioning time of 60 minutes in rainfall.

PAGE 152

152 Cr (DD = 0.2765 g, q = 0.1052 g) Cumulative mass, % 0 20 40 60 80 100 q DD DD = 4.78 DD = 0.08 R2 = 0.99 q = 5.56 q = 0.09 R2 = 0.99 As (DD = 0.0340 g, q = 0.0134 g) Cumulative mass, % 0 20 40 60 80 100 DD = 4.61 DD = 0.07 R2 = 0.99 q = 5.23 q = 0.10 R2 = 0.98 Cd (DD = 5.977 mg, q = 4.218 mg) Cumulative mass, % 0 20 40 60 80 100 DD = 4.56 DD = 0.09 R2 = 0.99 q = 6.56 q = 0.08 R2 = 0.99 Zn (DD = 2.559 g, q = 1.597 g) Particle diameter, m Cumulative mass, % 0 20 40 60 80 100 9500 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 DD = 6.45 DD = 0.06 R2 = 0.99 q = 6.13 q = 0.08 R2 = 0.98 Pb (DD = 1.378 g, q = 0.1501 g) Cumulative mass, % 0 20 40 60 80 100 DD = 5.61 DD = 0.08 R2 = 0.97 q = 8.23 q = 0.07 R2 = 0.99 Cu (DD = 0.8506 g, q = 0.1760 g) Particle diameter, m Cumulative mass, % 0 20 40 60 80 100 DD = 9.24 DD = 0.05 R2 = 0.99 q = 6.84 q = 0.08 R2 = 0.989500 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 9500 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 9500 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 Figure 5-6. Site mean metal distribution (c umulative mass for Cr, Cd, As, Pb, Cu, Zn) by different particle size gradation after tran sported from dry deposition into rainfallrunoff at Baton Rouge, LA. DD: dry deposition; q = runoff.

PAGE 153

153 Mg dissolved mass, mg 0 3 6 9 12 15 0 20 40 60 80 100 Fe 0 4 8 12 16 20 % finer by mass 0 20 40 60 80 100 Pb dissolved mass, mg 0.0 0.1 0.2 0.3 0.4 0.5 0 20 40 60 80 100 Ca 0 100 200 300 400 500 % finer by mass 0 20 40 60 80 100 As 0.00 0.01 0.02 0.03 0.04 0.05 % finer by mass 0 20 40 60 80 100 Cd dissolved mass, mg 0.00 0.01 0.02 0.03 0.04 0.05 0 20 40 60 80 100 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 Zn Paticle diameter, m 0 1 2 3 4 5 % finer by mass 0 20 40 60 80 100 4750 2000 800 600 425 300 250 180 150 106 75 63 53 45 38 25 <25 Cu Particle diameter, m dissolved mass, mg 0.0 0.3 0.6 0.9 1.2 1.5 0 20 40 60 80 100 TDM = 0.08 mg = 13.65 = 0.034 TDM = 0.06 mg = 8.05 = 0.035 TDM = 66.58 mg = 5.06 = 0.071 TDM = 51.04 mg = 6.98 = 0.046 TDM = 0.98 mg = 8.65 = 0.055 TDM = 1714 mg = 6.57 = 0.061 TDM = 3.50 mg = 7.04 = 0.055 TDM = 12.90 mg = 8.68 = 0.042 Figure 5-7. Relative incremental dissolved mass ranges of metals leached from dry deposition (DD) particles at an equilibrium part itioning time of 60 mi nutes in rainfall.

PAGE 154

154 Pb Concentration ( g/L) 0 3 6 9 12 15 Cu Particle Diameter ( m) 10100100010000 Concentration ( g/L) 0 5 10 15 20 25 30 Zn Particle Diameter ( m) 10100100010000 Concentration ( g/L) 0 50 100 150 200 250 300 Cd Concentration ( g/L) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 pH = 6 pH = 7 pH = 3 pH = 4 pH = 5 Mn Concentration ( g/L) 0 3 6 9 12 15 As Concentration ( g/L) 0.0 0.2 0.4 0.6 0.8 1.0 Fe Concentration ( g/L) 0 100 200 300 400 Mg Concentration ( g/L) 0 50 100 150 200 250 300 Figure 5-8. Dissolved concentra tions of metal species (Cd, As, Mn, Fe, Pb, Mg, Cu, Zn) in rainfall solutions after equal amount of particles in different size gradation transported from dry depositi on into urban rainfall-runoff.

PAGE 155

155 CHAPTER 6 EFFECTS OF SALINITY AND SSC ON SETTLING VELOCITY OF PARTICLES IN URBAN RAINFALL-RUNOFF Introduction The settling velocity is a key variable of the transport mech anism of particles throughout their entire pathways (Jimenez and Madsden 2003) Many empirical equations were predicted to describe accurately settling veloci ties in different scenarios, which was critical in the design and performance of unit operations and processes fo r stormwater runoff (Cheng 1997). While in a stormwater pretreatment system, one or more t ypes of settling can be found in different settling zones at a given time by considering particle con centration and tendency of particles to interact (Metcalf and Eddy 1991). Discrete settling (type I settli ng) is dominant for particles with a concentration of about 100 mg/L or less, and wh ere the interaction betw een particles can be ignored for most size gradations. Di screte settling velocity is a constant with specific particle characteristics such as particle size, shap e and density (Hawley 1982) Salinity, as a water physic-chemical parameter, coul d change both water density a nd ion strength, but only water density has a significant effect on the particle se ttling velocity in discrete settling zone. Balance of gravitational and frictional drag forces ar e key considerations when creating modeling equations, which were fundamentally expressed by Newtons Law or Stoke s Law for spherical particles (Jimenez 2003). Most na tural particles are irregular with variable shape and roundness, which could significantly affect a particles falling rate and cause errors and problems in predicting its settling velocity (B aba 1981b). In order to simulate th e settling velocity of natural particles, Newtons Law or Stokes Law needs to be modified to inco rporate the aspect of variable particle shapes. Furt hermore, Schulz at al. (1954) in dicated that various Reynolds numbers could cause different effects of partic le shape upon the drag coefficient, which was small with low Red and large with high Red.

PAGE 156

156 Particle interactions occur more frequently in higher concentr ation areas, generating aggregation or flocs with various structures and dimensions. In th ese areas, the effect of salinity on the flocculant settling velocity appeared to be significant. Particle collisions are usually generated by three different mechanisms: Br ownian motion, fluid shear, and differential sedimentation (McCave 1975; Hunt 1980). Brownian motion may cause particles to collide and form aggregates, but this is only significant for small particles less than 1 m. Differential sedimentation occurs when rapidly falling part icles overtake comparatively slowly falling particles and through the process of colliding, aggregate together. However, the capture efficiency was very small, shown in one study to be less than 0.2 %, and it decreased with the magnitude of fractal dimensions (Li and L ogan 1997; Winterwerp 2002) In the freshwater system, collision efficiency of particles was dete rmined in the range of 0.001 to 0.1, which was lower than that in the marine system (Jiang 1991 ). Fluid shear and turbulence was suggested to be dominant in the process of coagulation of part icles into larger aggregates or flocs in both freshwater and marine aquatic systems (Hunt 1980; Jackson 1990; Hill 1992; Winterwerp 1998). The effects of Brownian motion a nd differential sedimentation could be negligible in estuarine environments where most particulates were or iginated from domestic drainage (Winterwerp 2002). Flocs created in the process of particle aggr egation have various fractal dimensions, which is the most important parameter for characterizing floc self-similarity and measuring floc rarefied structure (Noever and Nikora 1995). It was also shown in a recent finding that fractal dimension was incorporated into relationships for floc strength, maximal size, and settling velocity (Kranenburg 1994, 1999; Winterwerp 1998; Winterwerp 2002). One relationship predicted by Winterwerp (1998) suggested that the fractal dimension might influence the floc settling velocity

PAGE 157

157 only if its equal to or greater than 2; while a fractal dimension of less than 2 for the fractal structure of sediment flocs has no affect upon the settling rate. The range of measured floc fractal dimensions for natural flocs was found to be from 1.3 to 2.7 (Winterwerp 2002). Furthermore, the settling velocity of flocs may vary with the particle concentration, depth, velocity gradients, and range of particle sizes un til those flocs fall into the hindered settling zone (Lau amd Krishnappan 1992). The effect of irregula r particles or flocs on hindered settling was greater than that for spherical one s, leading to a larger sediment ation exponent in the Richardson & Zaki expression and also leading to lower th an expected concentration gradients (Tomkins 2005). The effect became more significant when th e irregularity of particles or flocs was increased. At volume concentrations around 0.4, the measured hindered settling velocity of natural particles of a medium to small size re duced to about 70 %, compared to the value predicted by existing empirical expressions (Tomkins 2005). Objective There are three objectives in this study. The first objective is to investigate settling velocities of particles transpor ted from dry deposition into ra infall-runoff as a function of salinity, particle size an d particle concentration (SSC). Th e second objective is to compare measured settling velocities in different settli ng zones with data modeled from theoretical and empirical equations. The last objective is to examin e the behavior of particles in different settling zones (such as discrete, flocculant and hi ndered) during the quie scent settling period. Background Study Settling Columns Settling columns are widely used in the field or lab as effective apparatuses to determine particle settling velocity, as well as the removal efficiency of particles. For each individual case, a settling column was designe d with specific dimensions for different purposes and

PAGE 158

158 requirements. According to the various dimens ions and different mixi ng and sampling methods, several types of settling columns were categorized, such as a traditional column, a CERGRENE column, a UFT column and an Aston column. The traditional column, also named a long column, has a height of 1.8 to 2.5 m in order to simulate the effective settling depth in a sedime ntation tank, which is typically over 2.5 m. The diameter of a long column is in the range of 20 to 30 cm with total volume around 40 to 80 L, making it difficult to reach and maintain the homoge neous status in the column at the initial sampling time. An approach used by Pisano et al. (1984) was to mount the long column on a device that allowed axial rotation in order to reduce the measurem ent error of initial particle concentrations. The CERGRENE column was designed with the purpose of filling quickly and mixing completely at the time of initial sampling. A seri es of small columns, 1 m high and 65 mm in diameter, were set up to analyze th e particle settling velocity. This column is able to measure the settling velocity of particles ac ross a wide size range, and provides a more accurate prediction of parameters for the system design. It is also cap able of both in-situ measurements and laboratory use. However, the settling velocities measured in the field were interpreted as more accurate than those taken back to the lab, because a smalle r time interval elapsed between sampling and measurement made it much easier to preserve the original matrix of samples without agglomeration occurring. The UFT-type column was developed by the German company Umwelt and Fluid Technik, it used a typical top-loading sett ling column for either unmodified or pre-settled samples (Pisano and Brombach1996). It has a great advantage for th e analysis of effluents with very different settling behaviors and solids con centrations due to its specific construction, which includes two

PAGE 159

159 conjoint sections, an upper c onventional settling column and a lower Imhoff cone, with a total height of 1.3 m (Michelbach and Whrle 1993; Wong 2000). According to the basic concept of utilizing a UFT-type column, a water sample is first pre-settled in multiple Imhoff cones to separate the settleable solids from the liquid phas e, and then only the settleable material is collected and induced into the column for testing. The experimental data from this column test could generate a mass-based settling curve, where the settling velocity is illustrated as a function of cumulative mass fraction (Wong and Piedrahita 2000). The Aston column has a specific use for meas uring settling velocity gradations of both urban rainfall-runoff and dry weather flow se wage often with rela tively high particle concentrations. Tyack et al. ( 1996) developed a method to inves tigate the settli ng velocity of storm sewage by using an Aston column, typically 2.2 m high and 25 to 50 mm in diameter. He suggested that a larger diameter column c ould create a more accura te settling velocity distribution from a sample, by taking into consid eration the boundary effect of column walls to the inside particle and flow motions. There are tw o valves in each side of the column with the purpose of convenient sampling and refilling, and an axle jointed with th e column in the middle was fixed on two bearings for column rotation during the test (Ibid). Discrete Settling (Type I Settling) In type I settling, the settling velocity of spherica l particles was only a function of particle size while other parameters ( p, w, and ) keep constant, which can be described as a power law function of particle diameter with varied power index. Settling velocity will not vary with depth and TSS in the settling column. Newtons law and Stokes law are widely used to determine the discrete settling velocity of s pherical particles, ex pressed by equation 6-1 and 6-2 respectively. Newtons law can be used in all the flow regions (laminar, transitional, turbulent), and Stocks law is only fit in the Laminar flow region.

PAGE 160

160 The equation of Newtons Law is p w w p d Pd C g V ) ( 3 4 (6-1) The equation of Stocks Law is 18 ) (2 p w p Pd g V (6-2) 34 0 3 24 ed ed dR R C where ed dR C24, Red < 1; 34 0 dC Red > 1000 (6-3) p P edd V R (6-4) Drag coefficient Cd, a critical parameter for calculating settling velocity, varies with Red in different fluid regions. Neve rtheless, in equation 6-4 Red is determined by the calculated settling velocity from Cd itself. Therefore, iterative computation is employed in the calculation of settling velocity with corresponding drag coefficient and Red until the iterated velocity value converges. According to the assumption that particles are spherical, Newton s Law and Stocks Law cannot be applied directly for natural particles because of their irregular shapes. Hence, many revised equations were develope d to predict the settling veloc ity of natural particles more accurately by also taking into considerat ion the effect of th eir irregularities. Two shape factors are often used to describe th e irregular shape of par ticles, E shape factor and Corey shape factor. E shape factor is expressed as 2 1 2 2 23 l i s sD D D D E which is equal to 1 for a sphere and between 0 and 1 for nonspherical grains. According to the definition equation, 2 1) (l i sD D D CSF Corey shape factor is suggested as 0 for a two-dimensional plane

PAGE 161

161 and 1 for a sphere (Le Roux 2004). The study by Ba ba (1981a) suggested that irregular particles would keep their maximum project ion area vertically towards the settling direction while falling down individually, which is probabl y the reason why most irregular particles settled faster than regular ones with same weight and nominal size. Irregularly shaped particles typically score a Corey value of less than 1, and have an average score of a bout 0.7 (Cheng 1997). Almost no correlation was found between sett ling velocity and roundness for irregular particles in the low Reynolds number (Red) region, and the controlli ng factors considered here were only the shape of the particle and Red (Baba and Komar 1981a). The settling velocity Vp can be calculated with the modified Stokes eq uation for particles havi ng Reynolds numbers less than 0.5, expressed as equation 6-5 (Ibid). This equation can be employed to predict the settling velocity of very fine partic les or in regions of very hi gh viscosity flow, which usually corresponds to a low Red. ) 318 0 672 0 ( ) 318 0 672 0 ( ) ( 18 10 2E V E gD Vn w p p (6-5) Equation 6-5 shows that the se tting velocity has a linear rela tionship with E shape factor and could be expressed as the same equation of St okes Law with E shape factor equal to 1. For sieve-size range particles from 350 m to 1680 m, the relationship between settling velocity and E shape factor dropped to another linear tr end, shown in equation 6-6, because the high Red those particles created may cause spinning, tumbling, and oscillations, making the settling velocity different from that at Low Red (Baba and Komar 1981b). 008 0 992 00 E V Vp, 192 0 808 00 CSF V Vp (6-6)

PAGE 162

162 Cheng (1997) found that th e relationship between Cd and Red was different for natural sediment particles in the intermediate Red region, which can be expressed as an empirical equation (equation 6-7), compar ed to regular particles. 5 1 5 1 / 11 32 ed dR C (6-7) Derived from Stokes Law, Cd also can be express as 2 3 23 4 3 4ed p n dR d V gD C (6-8) Where dimensionless diameter d* was defined by n sD g d 3 / 1 2 *) ) ( ( Substituting equation 6-8 into equation 6-7, a simplified formul a for predicting the settling velocity of natural sediment particles was develope d as equation 6-9, which was s uggested have the highest degree of prediction accuracy when compared with other published equations (Cheng 1997). 5 1 2 *) 5 2 1 25 ( d D Vn p (6-9) Flocculant Settling (Type II Settling) Particle dimension, surface area and density changed eventually due to aggregation, making the simulation of settling velocity for aggregated particles more comprehensive by considering these unpredictable va riables (Droppo et al. 2000). Single floc settling Created by the coagulation of small compact pa rticles into large porous aggregates, flocs usually have different fractal structures and heterogeneous ma ss distributions (Li and Logan 1997). Fractal dimension correlate with floc properties such as strength, maximal size and settling velocity (Logan a nd Kilps 1995). In the prediction of floc settling velocity in a real

PAGE 163

163 sedimentation system, the key point is to measur e the floc fractal dimension accurately, which is difficult to obtain because of the low-accuracy inst ruments for in situ measurement (Nikora et al. 2004). To resolve this problem, LISST-ST was recently utilized to measure the in-situ sediment settling characteristics with little disturbance of se diment flocs, but it is not widely used due to its high price (Thonon et al. 2005). While corresponding to the degree of irregular ity and complexity or the space-filling capacity of an aggregate (Li and Logan 1997) floc fractal dimension is defined as ) ln( ) ln( lim L N nL L f where NL is the number of primary objects as a function of its linear size L (Winterwerp 2002). With this definition, the num ber of particulate monomers with diameter Dn in a larger floc of diameter Df could be expressed as fn n fD D N ) ( The porosity ( e ) of flocs then can be calculated as a function of nf, shown in equation 6-10. As know n that the porosity of flocs is a key factor in the determination of its densit y, the differential density of flocs was derived in equation 6-11 and also expressed as a function of fr actal dimension nf (Kranenburg 1994). 3 3) ( 1 ) ( 1 fn n f f n f p fD D D D N V V N V e (6-10) e e V V N V V V N V m m V mw p f p f w f p p f w p f f f ) 1 ( 3 3) ( ) )( ( f fn n f p n n f w p w f fD D D D (6-11) From the balance between the gravity force Fg and the hydrodynamic drag force FD, floc settling velocity was derived a nd shown in equation 6-12, where is the 3-dimensional shape factor and is the 2-dimensional (plate) shape factor.

PAGE 164

164 g D Ff f g 36 2 24 2f f d w DV D C F 2 2 34 2 6f f d w f f D gV D C g D F F 2 / 13 4 1 f w f d fD g C V 2 / 1 2) ( 3 4 1 n n n f w p d fD D D g C Vf (6-12) Assuming that Cd can be described by the relation ) 15 0 1 ( 24687 0 ed ed dR R C for regular particles, equation 6-12 is revised and expressed as equation 6-13, 687 0 1 315 0 1 ) ( 18ed n f n n w p fR D D g Vf f (6-13) For spherical ( = = 1), massive ( nf = 3) particles in the Stokes regime (Red<<1), the equation 6-13 transferred into the well know n equation, Stokes Law (Winterwerp 1998). Multi-flocs settling In a flocculent sedimentation sy stem, floc settling velocity is influenced by the depth of the basin, the velocity gradie nts, the concentration of particles, and the range of particle sizes, which could only be determined by sedimentation te sts (Matcalf and Eddy 1991). As no theoretical equation is available to describe flocculation sett ling, empirical functions are often used in the design of sedimentation tanks (Ozer 1994). The relationship betw een concentration percentage, time and depth can be expressed as d b t ht ah C C 0 where a, b, and d are parameters that will be

PAGE 165

165 determined for each set of column-test (Ozer 1994). The differential concentration at depth h and time t can be calculated by the following steps: t h t h t hC C dC, d b d b d bt ah dt t dh h a dt dh t abdh ) ( ) (1 1 Assuming thatdt V dht h ,, replace dh in the above equation with dt Vt h and the settling velocity is derived and expressed as a si mplified function, shown in equation 6-14. 1 ) 1 )( 1 (, t dt d h dt V b ht dt dt V bdt h t h t h b d Vt h (6-14) Lau and Krishnappan (1992) found that the concentration gradient at h and t can be expressed as a function of concen tration, depth and settling veloci ty, shown in equation 6-15. By intergrading Equation 6-15, the m ean settling velocity over a give n time period can be calculated from the concentration data as: dt C dC h V h dt V C dCt t t t t t dt dt C dC h t t dt V t t Vt t t t t t t t 2 1 1 2 2 1 1 2) ( 1 1 (6-15) 2 1ln1 2 t t tC C t t h V (6-16) While t1 = 0, the average settling velocity at time t is given by t tC C t h V0ln

PAGE 166

166 Hindered Settling (Type III Settling) Hindered settling usually occurs as a continuing process afte r discrete and flocculation settling, but the mechanism of hi ndered settling for discrete part icles and flocculating particles are different. For discrete particles, where interac tion could rarely cause flocculation of particles, the hindered settling velocity reduc es with particle c oncentration in creasing, as known described by Richardson & Zaki (1954) by their famous function expressed as equation 6-17. n nV V ) 1 (0 (6-17) Under a volume concentration the exponent n was found to be significantly influenced by the particle shape, which can be determined by ) log( ) log( ) log( ) log(0 0 if ed if ed if ifR R V V n where V0 and if are both corresponding to the partic le irregularity (Tomkins 2005). For flocculating particles, three hinde ring effects were suggested for the settling velo city, which are return flow and wake formati on, viscosity and buoyancy or re duced gravity. Combining these three effects, the effective settling velocity V is given by equation 6-18, where the factor ) 1 ( accounts for the return-flow effect, ) 1 (p represents the effect of buoyancy or reduced gravity, and ) 5 2 1 ( is correlated with the viscos ity effect (Winterwerp 2002). 5 2 1 ) 1 )( 1 (0 pV V (6-18) This relationship can also be de scribed by equation 6-19, in which ref and n are two parameters affected by the process of hindered settling (Br ooks et al. 2000).

PAGE 167

167 n refV V ) 1 ( 10 (6-19) Methodology Salinity, particle size and SSC are chosen as three main factors in th is study to investigate the effects on settling velocity of particles. Particles in urban rainfall-runoff have a wide size range from larger than 9500 m to less than 25 m, pre-separated into 18 gradations by mechanical sieves for experimental usage (San salone and Buchberger 1997). Receiving water bodies have various salinities, which may cause different settling mechanisms of particles discharged from runoff sewage of salinity different with these wa ter bodies. Four salinity values were employed to represent different water bodies, such as stormwater sewage, river, estuary and seawater, with salinity of 0 1.4 10 and 20 respectively. Pa rticle concentration (SSC) is a critical factor that affects the tendenc y of particles to aggreg ate or coagulate in the sedimentation process. According to the different concentration, settling behaviors of particles can be differentiated into four types: discrete, floccu lant, hindered, and co mpression settling. To fully investigate the settling behaviors of partic les with different concentrations, a series of concentrations were selected in this study, including 0.1, 1, 5, 10, 20, 50, 100, 200 g/L. Use of a traditional long column was chosen for this study because of its unique multisampling ports, where samples could be taken simultaneously at diffe rent fractions in the column during the sedimentation process. In addition, its long travel dist ance for particles and large inner diameter may significantly reduce the measuremen t error of settling velocity of particles, especially for the large r-sized and heavier particles. Three long columns, all with the same dimensions, were set up on a movable base, boun d on an aluminum frame and connected with each other through the serial ports at the bottom. Particles were mixed t horoughly with water in a

PAGE 168

168 200 L tank right behind the frame and pumped into th e columns through the seri al ports, if all the ports were open, they filled up simultaneously. While measuring the discrete settling velocity of particles with different size ranges, two methods were employed for partic les larger or less than 180 m respectively. First, 100 individual particle s larger than 180 m, were randomly selected from each size gradation,and then immersed in the pre-settled runoff sample to remove any attached bubbles before they were placed in the columns. The partic les were released one by one from the top of the columns, just below the waters surface. The traveling time of each individual particle was measured by a stopwatch from the moment the particle was re leased to the time the particle reached the sampling port. For particles less than 180 m, a certain amount of particles (around 3.8 g) were added into 100 ml distilled water, mixed and then sonicated for at least 20 minutes to break up any particle clogs and aggregates. The mixture was dropped fr om the top of the column and samples were taken from sampling port at each designed samp ling time. Samples were analyzed by a laser diffraction particle analyzer (LISST) as partic le size distribution on volume base. The top liquid level of the column will gradually go down while sampling is being performed, so the decrease in height was also measured. In this experiment an approximately 110 ml sample was taken each time, and it decreased about 10 mm from the beginning top level of column. Three sampling frequencies were selected for the different sized particles. In column the sampling frequency was once per 20 seconds and the overall sampling time was 15 minutes; this experiment was designed to inves tigate coarse particle settling ve locity, where particles fell in the size range of 55 to 250 m, In order to catch middle size particles around 10 to 65 m, the sampling frequency in column was once per 10 minutes, and the sampling time was about 8

PAGE 169

169 hours. In column the sampling frequency was once per 4 hours, and the entire sampling time lasted at least 8 days, thus most fine particles ranging in size from 2 to 14 m could be caught from the bottom port. The discrete settling velocity could be calculated directly by equation 6-20, where H is the traveling distance for particles which was the height from the waters surface to the sampling port, and t is the traveling time of particles as measured from the moment of release to the moment it arrived at the sampling port. t H Vp (6-20) To investigate the settling velocity of part icles in flocculant, hindered and compressed zones, high initial particle c oncentrations from 5 to 200 g/L were selected to examine the relationship between settli ng velocity and particle concentrati on in different regions, and whether the relationship varies as a function of elapsed time and salinity. Dry particles were added into the 200 L tank and mixed with 50 L water of different salinities to get the exact concentrations of 5, 10, 20, 50, 100, 200 g/L respectively. The mi xture was pumped into the column from the bottom until reaching the height of 2466 mm. Particles start settling down under a quiescent condition right after the column was filled. The elapsed time and the drop down distance of interface were recorded simultane ously at each time interval. Flocculant settling was dominant in the se ttling zone above the interface, where two methods are generally employed on calculating the flocculant settling velocity. One is based on concentration gradient versus time in this regi on which could be expressed as equation 6-16, and the other is mass transfer balance method acros s the interface between fl occulant and hindered settling zones, derived in equation 6-21. III III II IIV C V C

PAGE 170

170 dt dh h h h h Vi i i i II max 0 (6-21) Where CII and VII are particle concentration and settli ng velocity in the flocculant settling zone right above the interface, CIII and VIII are particle concentration and settling velocity in the hindered settling zone right below the interface, h0 is the height of water surface, hi is the height of interface, and hi,max is the maximal height of interface. The diameter of a single floc could be cal culated by equation 6-22, which was derived from equation 6-13. 1 1 3 687 0) ( 18 ) 15 0 1 ( f fn n n w p ed f fD g R V D (6-22) Hindered settling occurred undern eath the interface which ve locity was defined as the tangent slope at any point on the settling curve of interface height versus elapsed time, expressed as equation 6-23. dt dH t H Vp (6-23) Results Settling velocities of particle s under discrete settling condition s were investigated across the size range from 1 to 4750 m with SSC of 130 mg/L by cons idering the effect of salinity. Measured terminal settling velo city was compared with the modeled values by a modified Newtons Law for subject to differ ent salinity, shown in figure 1. Results indicated that salinity has no significant effect on both modeled and meas ured values. Since gravitational and frictional drag forces are two dominant factors on behavior of particles larger than 5 m, particle settling velocity could be described by Newtons Law or Stokes Law, leading a well match between

PAGE 171

171 measured and modeled values. Newtons Law equa tion was modified for a pplication to natural particles with irregular sh apes (see equation 6-5). When particles were less than 5 m, their interaction couldn t be negated because they have a very high numerical concentration, desp ite a very low overall mass. The interaction breaks the discrete settl ing condition, causing problems when using Newtons Law or Stokes Law to predict accurately what the particle sett ling velocity might be. Ion strength generated by salinity has a great effect on the chemical and electrostatic particle-on-particle interactions. High ion strength compresses the thickness of double laye r for particles with ne gative surface charge, which makes it easier for particles to aggregate a nd settle down more rapi dly than did individual ones. Our results illustrated that measured terminal settling velocities for particles less than 5 m were all larger than modeled va lues calculated by modified St okes Law under different salinity conditions, shown in figure 1. The deviation betw een measured and modeled values for salinity of 20 ppt were significantly greater than that for 1 and 10 ppt in the fine particle size range of less than 5 m. However, no significant difference was found for the deviation between 1 and 10 ppt, suggesting that the effect of salinity on aggr egation may only be signif icant when salinity is greater than 10 ppt. Aggregates/flocs created by fine particle s have different fract al structures and heterogeneous mass distribution, contributing various flocculant dimensions. Diameter of flocs was calculated based on equation 622, and by assuming that floc fr actal dimension was equal to 2.0 (Winterwerp 2002) and 3-dimensional shape f actor was 0.7 (Cheng 1997) for each original particle size. Under quies cent conditions with salinity about 1 ppt, aggregation wa s most likely to occur for particles less than 3 m, and flocs sizes ranged from 3.2 to 6.4 m. A similar result was found for conditions where salinity was10 ppt, a nd where floc size varied from 2.9 to 7.0 m.

PAGE 172

172 For salinity of 20 ppt, aggregation appeared to be more significant and floc size jumped into the range of 4.4 to 13.8 m for particles of 5 m or less. The distribution of terminal settling velocity for part icles of a specific size was investigated under discrete settling conditions, de picted as a probability density function (pdf) with different salinities shown in figures 2, 3, and 4 respectivel y. All pdfs could be described by an optimized log normal distribution function. Results indicated that settling velocities for particles in different size range s generally distributed in a ra nge of one or two orders of magnitude across the median value. Salinity has no significant effect on these distributions, but particle shape and density were considered to be two main effects on the settling velocity distribution. Flocculant settling velocities of particle s were examined under different initial concentrations with values of 5, 10 and 20 g/L respectively. Resu lts in figure 5 illustrated three increasing trends for settling velocities as time elapsed, which can be described by a set of power law functions with parameters all correlated proportionally with particle concentrations. Coefficients were equal to 48.1, 82.0, and 115. 6, and exponents were 0.6, 0.7, and 0.8 for concentrations of 5, 10 and 20 g/L. Within the flocculant settling period, settling velocity for particles with initial concentration of 5 g/L increased from 15 mm/s to 85 mm/s. Settling rate jumped from 30 to 160 mm/s for concentration of 10 /L, and 25 to 215 mm/s for concentration of 20 g/L in the same time interval. Particles under the interface settled down with a constant settling velocity, noted as hindered settling which varies with initial part icle concentration only. To investigate the relationship of hindered settling ve locity and initial concentrati on, the height of interface for different concentrations was measured as a functi on of elapsed time, which is illustrated in figure

PAGE 173

173 6. Results indicated that the hi gher the initial concentration, the faster settling particles would settle, and settling velocities were recorded to be 0.01, 0.04 and 0.21 mm/min for concentrations of 5, 10, and 20 g/L respectively. The height of interface was also measured for particles with initial concentrations larger th an 50 g/L under different salinity conditions, shown in figure 7, 8 and 9 for salinities of 1, 10 and 30 ppt, respectively. The eff ect of salinity was found to be significant when particle concen trations reached 200 g/L, but sa linity was less significant for particle concentrations of 100 and 50 g/L. For th e first 7 hours, particle interface was relatively constant when the salinity was 1 ppt and the pa rticle concentration wa s 200 g/L; but then it dropped rapidly over the next 7 hours from 2466 mm down to 1720 mm. When salinity was 10 ppt, the constant interface level of particles was kept for the first few hours, but then it dropped down to 1720 mm line at the time of 14 hours. When the salinity was 30 ppt, the interface started dropping down from the beginning and reached 172 cm level at the time of 14 hours also. Hindered settling velocities were 4.23, 1.90 a nd 0.80 mm/min for part icles with initial concentration of 50, 100, 200 g/L under conditions of 1 ppt salinity, which were much higher than those in relatively low particle concentrations of less than 50 g/L. The relationship of hindered settling velocity and initial concentrat ion is depicted in fi gure 10, where results illustrated two reversed trends. When particle concentrations were greater than 50 g/L, the relationship of settling velocity and initial particle concentration followed an increasing trendline, whereas when particle concentrations were less than 50 g/L, the relationship followed a decreasing trendline. Conclusions Influence of SSC and salinity on settling velocity of particles was investigated in a series of long columns under quiescent condition. Results indicated that settling velocities modeled by a modified Newtons law function ma tched measured values very we ll for particles larger than 5

PAGE 174

174 m with initial concen tration of 130 mg/L, where salinity had no significant effect on both modeled and measured values under discrete sett ling conditions. For fine particles of a size less than 5 m, a significant deviation was found betw een measured and modeled values, indicating that aggregation was more likely to occur in this size region, and wh ere Newtons Law is not suitable to predict the settling velocity for these colloidal particles. In addition, the effect of salinity on settling velocity appeared to be sign ificant, because ion stre ngth generated by salinity could significantly influence the interaction be tween fine particles. Flocculant settling only occurred in the first few hours for high particle concentrations such as 5, 10 and 20 g/L, where three increasing trends were found for settling ve locities as time elap sed, which could be described by a set of power law functions w ith parameters all correlated with particle concentrations proportionally. Hi ndered settling velocity was re latively constant for specific particle concentrations, illustrating an increasi ng trend with values of 0.01, 0.04, and 0.21 mm/min for concentration of 5, 10, 20 g/L resp ectively, and followed by a decaying trend with values equal to 4.23, 1.90, and 0.80 mm/min for 50, 100, and 200 g/L. Resu lts indicated that the effect of salinity on hindered settling velocity was not significant for mo st concentrations less than 200 g/L.

PAGE 175

175 Particle Diameter, D ( m) 110100100010000 Settling Velocity, Vs (mm/s) 0.001 0.01 0.1 1 10 100 1000 1 ppt 10 ppt 20 ppt 1 ppt model 10 ppt model 20 ppt model Temperature : 24.5 oC + 0.2 oCLaminar Transitional Turbulent Particle Diameter, D ( m) 110100 110100 Cumulative PND [#/mL] 1e+2 1e+4 1e+6 1e+8 1e+10 1e+12 % finer by number 0 20 40 60 80 100 cumulative PND [#/mL] % finer by number SSC : 130 + 5 mg/L s : 2.50 + 0.06 g/cm3 44 0 Re 3 Re 24 D D DC Figure 6-1. Measured median terminal settli ng velocities of particles across the size range from 1 m to 4750 m, considering the effect of various salinity under discrete settling conditions.

PAGE 176

176 Salinity = 1 ppt pdf 0.00 0.05 0.10 0.15 0.20 Salinity = 10 ppt pdf 0.00 0.05 0.10 0.15 0.20 Salinity = 20 ppt Settling velocity, Vs (mm/s) 0.00010.0010.010.1110100 pdf 0.00 0.05 0.10 0.15 0.20 1 m 5 m 2 m 10 m 1 m 5 m 2 m 10 m 1 m 5 m 2 m 10 m Figure 6-2. Settling velocity di stribution of particles in su spended fraction, shown as a probability density function (pdf) for each specific size gradation (1, 2, 5, and 10 m). Results were compared against various salinities.

PAGE 177

177 Salinity = 1 ppt pdf 0.00 0.05 0.10 0.15 0.20 Salinity = 10 ppt pdf 0.00 0.05 0.10 0.15 0.20 Salinity = 20 ppt Settling velocity, Vs (mm/s) 0.0010.010.11101001000 pdf 0.00 0.05 0.10 0.15 0.20 25 m 50 m 75 m 25 m 50 m 75 m 25 m 50 m 75 m Figure 6-3. Settling velocity di stribution of particles in su spended fraction, shown as a probability density function (pdf) for each specific size gradation (25, 50, and 75 m). Results were compared against various salinities.

PAGE 178

178 Salinity = 1 ppt pdf 0.00 0.05 0.10 0.15 0.20 Salinity = 10 ppt pdf 0.00 0.05 0.10 0.15 0.20 Salinity = 20 ppt Settling velocity, Vs (mm/s) 0.010.111010010001000 0 pdf 0.00 0.05 0.10 0.15 0.20 100 m 150 m 200 m 100 m 150 m 200 m 100 m 150 m 200 m Figure 6-4. Settling velocity di stribution of particles in su spended fraction, shown as a probability density function (pdf) for each specific size gradation (100, 150, and 200 m). Results were compared against various salinities.

PAGE 179

179 Type II settling Time (hours) 0.00.51.01.52.02.5 Settling velocity, Vs (mm/s) 0 50 100 150 200 250 5 g/L (0.5 %) 10 g/L (1.0 %) 20 g/L (2.0 %) Vs = 115.6 t0.8Vs = 82.0 t0.7Vs = 48.1 t0.6 Figure 6-5. Settling velocity of particles in Type II (f locculant) settling region under a quiescent settling condition, illustrated as 3 exponential trends for original particle concentrations of 5, 10, 20 g/L respectively.

PAGE 180

180 Time (hours) 020406080100120 Height of interface (cm) 0 5 10 15 20 25 30 5 g/L ( 0.5% ) {Vs5 = 0.05 cm/hr} 10 g/L ( 1% ) {Vs10 = 0.22 cm/hr} 20 g/L ( 2% ) {Vs20 = 1.27 cm/hr} s : = 2.5 g/cm3, = 0.06 g/cm3 Salinity : 1.4 ppt Temperature : 23oC Vs20 Vs10Vs5 Figure 6-6. Interface height of Type III (hinde red) settling region for particles with initial concentrations of 5, 10, 20 g/L, respectiv ely. Settling velocity in hindered settling region is a constant for each specific concentration, determined by the ratio of the distance of interface drop and the settling elapsed time, and expressed as the slope of interface graph within the region.

PAGE 181

181 Time (hours) 050100150200 Height of interface (cm) 50 100 150 200 250 300 50 g/L ( 5%) 100 g/L (10%) 200 g/L (20%) s : = 2.5 g/cm3, = 0.0017 g/cm3 Salinity : 1 ppt Temperature : 23oC Time (hours) 05101520 Height of interface (cm) 50 100 150 200 250 Figure 6-7. Interface height of Type III (hinde red) settling region for particles with initial concentrations of 50, 100, 200 g/L, re spectively, under a qui escent condition with salinity of 1 ppt.

PAGE 182

182 Time (hours) 050100150200 Height of interface (cm) 50 100 150 200 250 300 200 g/L ( 20% ) 100 g/L ( 10% ) 50 g/L ( 5% ) s : = 2.5 g/cm3, = 0.0017 g/cm3 Salinity : 10 ppt Temperature : 24.5oC Time (hours) 05101520 Height of interface (cm) 50 100 150 200 250 Figure 6-8. Interface height of Type III (hinde red) settling region for particles with initial concentrations of 50, 100, 200 g/L, resp ectively, under a quies cent condition with salinity of 10 ppt.

PAGE 183

183 Time (hours) 050100150200 Height of interface (cm) 50 100 150 200 250 300 200 g/L ( 20% ) 100 g/L ( 10% ) 50 g/L ( 5% ) s : = 2.5 g/cm3, = 0.0017 g/cm3 Salinity : 30 ppt Temperature : 24.5oC Time (hours) 05101520 Height of interface (cm) 50 100 150 200 250 Figure 6-9. Interface height of Type III (hinde red) settling region for particles with initial concentrations of 50, 100, 200 g/L resp ectively, under a quies cent condition with salinity of 30 ppt.

PAGE 184

184 Hindered settling Initial particle concentration [g/L] 1101001000 Settling velocity, Vs (mm/min) 0.001 0.01 0.1 1 10 Figure 6-10. Relationship of settling velocity a nd initial particle concentration in Type III (hindered) settling zone under a quiescent settling condition.

PAGE 185

185 CHAPTER 7 GLOBAL CONCLUSIONS This dissertation investigated constitutive properties of particulate matter (PM) in urban dry deposition (largely dry traffic deposition) and source area rainfall-runoff loadings from inter state 10 over City Park Lake in Baton Rouge, Louisiana. PM fr om eight wet weather events and 17 dry weather events was recovered and analyzed Wet weather event ca pture was unique from all other studies in that all runoff was collect ed and all PM transported in the runoff was recovered for each event. This chapter summ arizes the conclusions reached from this dissertation. Results indicate that the delivery of PM loads or size fractions of these loads in source area runoff can be differentiated into mass-limited and flow-limited behavior. Wh ile such behavior is generally not known a-priori with current technology, understanding that load transport can be resolved into two limiting behaviors, largely a function of hydrology and granulometry, benefit the selection, design, operation, maintenance a nd performance of unit operations and processes (UOPs). The relationships between PM and turbid ity for untreated source area runoff for flowlimited and mass-limited events could be expressed as two linear equations with statistically significantly different slopes of 1.23 and 2.83. The two different sl opes were resolved to a single statistically different and lowe r value (0.98) for the linear relationship after 60 minutes of quiescent settling. In addition to PSDs recovered across the entire gradation, the gradation was resolved into a suspended fraction (1 to 25 m), a settleable fraction (25 to 75 m) and a sediment fraction ( > 75 m). Results indicate that the PM in mass-li mited events had a higher proportion of settleable and sediment mass (around 70 % to 85 % total by mass), as compared to flow-limited events (generally less than 70 %). Results indicated th at the settlable partic les had higher settling

PAGE 186

186 velocities than the suspended fraction of particles. The PM mass removal efficiencies for masslimited events was from 66 % to 76 % after 60 minut es of batch quiescent settling, much higher than flow-limited events with efficiencies in the range of 39 % to 57 %. Results illustrate that delivered PSDs are a function of hydrology and therefore unit operation e fficiency, even for a given hydraulic residence time or surface load ing rate, is a function of hydrology unless upstream equalization is provided. Event-based ratios of settled and unsettled turbidity or SSC mass ( T60/ T0) were significantly different for masslimited and flow-limited events, also indicating that mass-limited events demonstrate higher treatment efficien cies as compared to flow-limited events for the sa me hydrodynamic conditions For both flow-limited and masslimited events, phosphorus st rongly partitioned to the particular phase rather than di ssolved phase of source area influe nt runoff. For this concretepaved source area, concentrations of particulate phosphorus were in the range of 1 to 3 mg/L and 0.1 to 0.3 mg/L for dissolved phosphorus. Partic ulate phosphorus predominately partitioned to the suspended fraction of influent runoff for most events, in part icular for flow-limited events with a proportion of more than 50 %, and the lo west proportion was found in sediment fraction for both classes of events, whic h was less than 20 % by mass. Chemical oxygen demand (COD) concentration for the combined dissolved and suspended fractions of rainfall-r unoff was in the range of 50 to 200 mg/L for captured wet weather events as an event mean concentration (EMC), typicall y of primary clarifier in fluent of a municipal wastewater.treatment plant. C OD has a high proportion in the di ssolved phase of runoff with fd values larger than 0.6 and Kd values less than 10 L/Kg. Concentrations of Cu, Zn and Cr exceeded criteria maximum con centrations (CMCs) (USEPA 1999) for all the wet weather events captured, even after 60 minutes of quiescent

PAGE 187

187 settling, while Pb and Cd exceeded CMCs only for x of the eight events. fd values of metal species were in the range of 0.1 to 0.3 and Kd values from 10 to 100 L/Kg. The high degree of metals in the suspended fraction occurred for onl y flow-limited events with a contribution of 43 to 64 %. The fraction associated with the su spended fraction was 19 to 40 % for mass-limited events. Metals in settleable fraction dominated for most events, especially for mass-limited events with a proportion of 53 to 66 % by mass. Dry traffic deposition fluxes of particles in different PM frac tions (suspended, settleable and sediment) were found to be correlated w ith previous dry hours in an exponential growth function during the entire dry deposition samp ling period for each of the 17 events. The D50m of dry deposition particles were in a range of 257 to 369 m, with more than 95 % of particles larger than 75 m for all the dry deposition periods. The granulometric distribution of particles was described by a cumulative gamma distribut ion function. Cumulativ e mass distributions (CMDs) of particles were translated to a finer particle size gr adation after being transported from dry deposition into rainfall runoff. The tr ansportation land use ROW translated the D50m (median diameter based on mass) from 330.7 m for dry traffic deposition to 98 mm for upstream runoff to 23 mm for downstream runoff, and finally to 13.7 m after 1 hour of batch retention. Since dry transportation deposition PM will part ition, in part, to the aqueous phase once wet by rainfall and transported from the pavement surface by runoff, the solubility of this PM was investigated. It was hypothesi zed that the solubility of the PM would result in higher pH, higher total dissolved solids (TDS ), higher conductivity and higher metal concentrations in the solution phase. Under well-mixed conditions typical of the pavement surface impacted by rainfall drops, tire suction, and vehicular sp ray, results demonstrated that pH, TDS and

PAGE 188

188 conductivity values increased rapidly, and reached a peak value within 5 minutes, before finally dropping to an equilibrium level for the remai nder of the period. Considering the effect of particle gradations, equilibriu m pH, alkalinity, TDS and conductivity of rainfall solutions decreased significantly with partic le size increasing, and went cons tant when particle size is over 300 m. Conversely, the initial pH of rainfall ha d a minor influence t on these water chemistry indices of rainfall solutions. Arithmetic means of metal concentration across the entire size gradation in dry deposition particles was ranked as Cd < As< Cr < Cu < Pb < Zn < Mn < Mg < Fe < Ca. Cumulative mass distributions of dry traffic de position PM metals were well described by a cumulative gamma distribution. Such a cumulative ga mma distribution is also applicable for PM metals in rainfallrunoff, which is finer than dry traffic depos ition PM The proportion of metal mass correlated with fine particles (less than 75 m) increased from 5 % to 20 % during the transport, while d50m by metal mass generally decreased from 300 450 m down to 150 220 m. At an equilibrium partitioning time of 60 minutes after transport, dissolved mass of metals was found to follow similar ranking as dry traffic deposition As < Cd < Pb < Cu < Zn < Mg < Fe < Ca. The slight change of order indicated vary ing leaching capability of differ ent metals. The initial pH of rainfall has no significant effect on dissolved metal concentrations for Cd, As, Cu, Pb, Zn, Mg, Fe, Ca when they were transported from partic ulate phase of dry deposition particles into dissolved phase of rainfall-runoff. One of the important constitutive properties of runoff PM is settling velocity. Settling velocities were modeled by a modified Newtons law function. Results demonstrated that measured and modeled values matched very well for particles larger than 3 m. At a gravimetric concentration of 130 mg/L, salinity had only a sma ll effect on settling ve locities unde r discrete

PAGE 189

189 (Type I) settling conditions. For suspended particles of a size less than 3 m, a significant deviation was found between measured and mode led values, indicating that Brownian motion began to have a significant influence and wher e Newtons Law is not suitable to predict the settling velocity for these particles approaching collo idal size particles. In addition, the effect of salinity on settling velocity appeared to be sign ificant, because ion stre ngth generated by salinity could significantly influence the inte raction between fine particles. Flocculant settling was examined at higher co ncentrations, typical of stormwater sludge PM. Flocculant settling only occurr ed in the first few hours for high particle concentrations at 5, 10 and 20 g/L. Higher concentrations led to high er hindered settling veloci ties (Type III) within this range of concentrations. Results could be described with a set of power law functions with parameters all correlated with particle concentrations. Hindered settling velocity was relatively constant for a specific particle concentrations, illustrating an increasing trend with values of 0.01, 0.04, and 0.21 mm/min for concentration of 5, 10, 20 g/L, and followed by a decaying trend in settling velocities with values equal to 4.23, 1.90, and 0.80 mm/min for 50, 100, and 200 g/L respec tively. Results indicated that the effect of salinity on hindered settling veloc ity was not significant for most concentrations less than 200 g/L.

PAGE 190

190 LIST OF REFERENCES Alon, G. and Adin, A. (1994). Mathematical mode ling of particle size di stribution in secondary effluent filtration. Water Environmental Research, 66, 836-841. Anderson, B., and Potts, D. F. (1987). Suspe nded sediment and turb idity following road construction and logging in western Montana. Water Resources Bulletin, 23, 681-690. ASTM (1993). Standard Practice for Dry Preparation of So il Samples for Par ticle-Size Analysis and Determination of Soil Constants. West Conshohocken, Pa. Baba, J. and Komar, P. D. (1981a ). Settling velocities of irre gular grains at low Reynolds numbers. Journal of Sedimentary Petrology, 51(1), 121-128. Baba, J. and Komar, P. D. (1981b). Measurements and analysis of settli ng velocities of natural quartz sand grains. Journal of Sedimentary Petrology, 51(2), 631-640. Ball, J. E., Jenks, R., and Aubourg, D. (1996). D ry weather build-up of constituents on road surfaces. Proceedings of the 7th Internati onal Conference on Ur ban Storm Drainage Hannover, Germany, 785-790. Barrett, M. E., Zuber, R. D., Collins, E. R., Malina, J. F., Charbeneau, R. J., and Ward, G. H. (1995). A review and evaluation of literatur e pertaining to the qua ntity and control of pollution from highway r unoff and construction. 2nd edition, Center for Research in Water Resources, University of Texas at Austin, 162. Bellinger, E. G., Jones, A. D., and Tinker, J. (1 982). The character and dispersal of motorway run-off water. Water Pollution Control, 81(3), 372-390. Bent, G. C. (2001). Effects of forest-manag ement activities on r unoff components and groundwater recharge to Quabbin Rese rvoir, central Massachusetts. Forest Ecology and Management, 143(1-3), 115-129. Bourcier, D. R., and Hinden, E. (1979). Lead, iron, chromium and zinc in road runoff at Pullman, Washington. Science of the Total Environment, 12, 205-215. Bricker, O. P. (1999). An overview of the factor s involved in evaluating the geochemical effects of highway runoff on the environment. U.S. Geological Survey Open-File Report 98-630, 28. Brooks, K. P., Rector, D. R., and Smith, P. A. (2000). Gravity settling of hanford single-shell tank sludges. U.S. Department of Energy, PNNL-12010. Bullin, J. A., and Moe, R. D. (1982). Measurement and analysis of aerosols along Texas roadways. Environmental Science and Technology, 16(4), 197-202. Characklis, G. W., Wiesner, M. R. (1997). Particles, metals, a nd water quality in runoff from large urban watershed. Journal of Environm ental Engineering, 123(8), 753-759.

PAGE 191

191 Cheng, N. S. (1997). Simplified settling ve locity formula for sediment particle. Journal of Hydrology Engineer Division, ASCE, 123 (2), 149-152. Chuaybamroong, P., Cayse, K., Wu, C. Y., and Lundgren, D. A. (2007). Ambient aerosol and its carbon content in Gainesville, a mid-scale city in Florida, Environmental Monitoring and Assessment, 128, 421-430. Colman, J. A., and Breault, R. F. (2000). Sampling for mercury at subnanogram per litre concentrations for load estimation in rivers. Canadian Journal of Fisheries and Aquatic Sciences, 57, 1073-1079. Colman, J. A., and Friesz, P. J. (2001). G eohydrology and limnology of walden pond, Concord, Massachusetts. U.S. Geological Survey Water-Resour ces Investigations Report 01-4137, 61. Cristina, C. M. and Sansalone J. J. (2003). First flush, power law and particle separation diagrams for urban storm-wate r suspended particulates. Journal of Environmental Engineering, 129(4), 298-307. Davelaar, D. (1993). Ecological significan ce of bacterial polyphosphate metabolism in sediments. Hydrobiologia, 253, 179-192. Davidson, C. I., Santhanam, S ., Fortmann, R. C. (1985). Atmos pheric transport and deposition of trace elements onto the greenland ice-sheet. Atmospheric Environment, 19(12), 20652081. Deboudt, K., Flament, P., Bertho, M. L. (2004) Cd, Cu, Pb and Zn concentrations in atmospheric wet deposition at a coas tal station in western Europe. Water, Air, and Soil Pollution, 151, 335 -359. Deletic, A. B., and Maksimovic, C. T. (1998). Evaluation of water qua lity factors in storm runoff from paved areas. Journal of Environmental Engineering, 124(9), 869-879. Dempsey, B. A., Tai, Y. L., and Harrison, S. G. (1993). Mobilization and removal of contaminants associated with urban dust and dirt. Water Science Technology, 28(3-5), 225-230. Drapper, D., Tominson, R., and Williams, P. (2000) Pollutant concentrations in road runoff: southeast Queensland case study. Journal of Environm ental Engineering, 126(4), 313320. Droppo, I. G., Walling, D. E., Ongley, E. D. ( 2000). The influence of floc size, density and porosity on sediment and contaminant transport. The Role of Erosion and Sediment Transport in Nutrient and Contaminant Transfer, 141-147. Ellis, J. B., Revitt, D. M. (1982). Incidence of heavy-metals in street surface sediments solubility and gr ain-size studies. Water, Air, and Soil Pollution, 17(1), 87-100.

PAGE 192

192 Ellis, J. B., Revitt, D. M., Harrop, D. O., a nd Beckwith, P. R. (1987). The contribution of highway surfaces to urban stormwater sediments and metal loadings. The Science of the Total Environment, 59, 339-349. Fisher, L. H. and Wood, T. M. (2001). Effect of water-column pH on sediment-phosphorus release rates in upper Klamath Lake, Oregon. Water-Resources Investigations Report 034271, U.S. Department of the Interi or, and U.S. Geological Survey. Gachter, R., Meyer, J. S., and Mares, A. (1988). Contribution of bacteria to release and fixation of phosphorus in sediments. Limnology and Oceanography, 66(6-2), 1542-1558. Gippel, C. J. (1989). The use of Turbidim eters in suspended sediment research. Hydrobiologia, 77, 465-480. Gippel, C. J. (1995). Potential of turbidity m onitoring for measuring the transport of suspended solids in streams. Hydrological Processes, 9, 83-97. Glenn, D. W., and Sansalone J. J. (2002). Accret ion and partitioning of heavy metals associated with snow exposed to urban traffic and winter storm maintenance activities. II Journal of Environmental Engineering, 128(2), 167-185. Gray, J. R., Glysson, D., Turcios, L., and Sc hwarz, G. (2000). Comparability of suspendedsediment concentration and total suspended solids data. Water-Resources Investigation Report 00-4191, USGS. Greenberg, A. E., Clesceri, L. S., and Eaton, A. D. (1992). Standard Methods for the Examination of Water and Wastewater, 18th edition. American Pub lic Health Association, Washington D.C. Greenberg, A.E., (1995). Standard Methods for the Examin ation of Water and Wastewater, 19th edition. American Public Health Association, Washington. DC. Harned, D. A. (1988). Effects of highway r unoff on streamflow and water quality in the Sevenmile Creek Basin, a rural area in the Piedmont Province of No rth Carolina, July 1981 to July 1982. U.S. Geological Survey Water-Supply Paper 2329, 33. Hawley, N., (1982). Settling velocity di stribution of natural aggragates. Journal of Geophysical Research-Ocean and Atmospheres, 87(12), 9489-9498. Hedges, P. D., and Wren, J. H. (1987). The te mporal and spatial vari ations in the aerial deposition of metals for a resident ial area adjacent to a motorway. The Science of the Total Environment, 59, 351-354. Hill, P. S. (1992). Reconciling aggregation th eory with observed vertical fluxes following phytoplankton blooms. Journal of Geophysical Research, 97, 2295-2308.

PAGE 193

193 Holsen, T. M., and Noll, K. E. (1992). Dry deposition of atmospheric particles Application of current models to ambient data. Environmental Science and Technology, 26(9), 18071814. Holsen, T. M., Noll, K. E., Lui, S. P., and Lee, W. J. (1991). Dry deposition of polychlorinated biphenyls in urban areas. Environmental Science and Technology, 25, 1075-1081. Hunt, J. R. (1980). Prediction of oceanic part icle size distributions from coagulation and sedimentation mechanisms. Advances in Chemistry Series No. 189, 243-257. Hunter, J. V., Balmat, J., Wilber, W., and Saba tino, T. (1981). Hydrocar bons and heavy metals in urban runoff. Urbanization, stormwater runo ff and the aquatic environment, George Mason University. Hvitved-Jacobsen, T., and Yousef, Y. A. (1991) Road runoff quality, environmental impacts and control in road pollution. R. S. Hamilton and R. N. Harrison, eds., Elsevier, London, 165-209. Hvitved-Jacobsen, T., Johansen, N. B., and Yous ef, Y. A. (1994). Treatmentsystems for urban and highway runoff in Denmark. Science of the Total Environment, 146-147, 499-506. Innman, D. L. (1952). Measures for descri bing the size distribution of sediments. Journal of Sedimentary Petrology, 22(3), 125-145. Irwin, G. A., Losey, G. T. (1978). Water-qual ity assessment of runoff from a rural highway bridge near Tallahassee, Florida USGS Water-Resources Investigations Report 79-1. Isano, W.C., and Brombach, H. (1996). Solids se ttling curves: wastewater solids data can aid design of urban runoff controls. Water Environment Technology, 8(4), 27-33. Jackson, G. A. (1990). A model of the formatio n of marine algal flocs by physical coagulation processes. Deep-Sea Research, 37, 1197-1211. Jacopin, C., Bertrand-Krajewski, J., and Desbordes, M. (1999). C haracterization and settling of solids in an open, grassed, stormwat er sewer network detention basin. Water Science and Technology, 39(2), 135-144. Jiang, Q. and Logan, B. (1991). Fractal dimensi ons of aggregates determined from steady-state size distribution. Environmental Science and Technology, 25(12), 2031-2038. Jimenez, J. A. and Madsen, O. S. (2003). A si mple formula to estimate settling velocity of natural sediments. Journal of Waterway Port C oastal and Ocean Engineering-ASCE, 129(2), 70-78. Johnson, C., Li, X.Y., and Logan, B. (1996). Settling velocities of fractal aggregates. Environmental Science and Technology, 30(6), 1911-1918.

PAGE 194

194 Kann, J. (1998). Ecology and water quality dynamics of a shallow hypereutrophic lake dominated by cyanobacteria. Ph.D. dissertation, Chapel Hill, University of North Carolina, 110. Kavanaugh, M. C., and Leckie, J. O. (1980). Use of particle size distri bution measurements for selection and control of solid /liquid separation processes in particulates in water. American Chemical Society, Washington, D.C. Kavanaugh, M. C., Tate, C. H., Trussell, A. R., Tr ussell, R. R., and Treweek, G. (1980). Use of particle size distribution measur ements for selection and control of solid/liquid separation processes. Advances in chemistry series 189, ACS, 305-328. Khoshmanesh, A., Hart, B.T., Duncan, A., and Beckett, R. (2002). Luxury uptake of phosphorus by sediment bacteria. Water Resources Research, 36, 774-778. Kobriger, N. P. (1984). Volume I. Sources and migration of highway runoff pollutants. Federal Highway Administration, Rexnor d, EnvironEnergy Technology Center, FHWA/RD-84/057, Milwaukee, Wisconsin. Kobriger, N. P., and Geinopolos, A. (1984). Sour ces and migration of highway runoff pollutants Research report volume III. U.S. Federal Highway Ad ministration Final Report FHWA/RD-84/059, 358. Krairapanond, A., Jugsujinda, A., and Pa trick, W. H. (1993). Phosphorus sorption characteristics in acid sulfate soils of Tha iland effect of uncontr olled and controlled soil redox potential (EH) and pH. Journal of Plant Soil, 157(2), 227-237. Kranenburg, C. (1994). The fractal structure of cohesi ve sediment aggregates. Estuarine Coastal and Shelf Science, 39, 451-460. Kranenburg, C. (1999). Effects of floc streng th on viscosity and de position of cohesive sediment suspensions. Continental Shelf Research, 19, 1665-1680. Krein, A. and Schorer, M. (2000). Road Runoff Pollution by Polycyclic Aromatic Hydrocarbons and its Contri bution to River Sediments. Water Research, 34(16), 41104115. Lau, Y. L., and Krishnappan, B. G. (1992). Size distribution and settling velocity of cohesive sediments during settling. Journal of Hydraulic Research, 30(5), 673-684. Le Roux, J. P. (2004). A hydrodynamic classification of grain shapes. Journal of Sedimentary Research, 74(1), 135-143. Lee, H., Lau, S., Kayhanian, M., and Stenstrom, M. K. (2004). Seasonal first flush phenomenon of urban stormwater discharges. Water Research, 38, 4153-4163. Lee, J. G., Heaney, J. P. (2003). Estimation of urban imperviousness and its impacts on storm water systems. Journal of Water Resources Planning and Management, 129(5), 429-426.

PAGE 195

195 Lee, P. K., and Touray, J. C. (1998). Character istics of a polluted artif icial soil lo catd along a motorway and effects of acidification on the leaching behavior of heavy metals (Pb, Zn, Cd). Water Research, 32(11), 3425-3435. Levine, A.D., Tchobanoglous, G. and Asano, T. ( 1985). Characterization of the size distribution of contaminants in wastewater: tr eatment and reuse implications. Journal Water Pollution Control Federation, 57, 805-816. Levine, A.D., Tchobanoglous, G., and Asano, T. (1991). Size distribu tion of particulate contaminants in wastewater a nd their impact on treatability. Water Research, 25, 911922. Lewis, J. (1996). Turbidity-controlled suspe nded sediment sampling for runoff-event load estimation. Water Resources Research, 32(7), 2299-2310. Li, D. H., and Ganczarczyk, J. (1989). Fractal ge ometry of particle aggregates generated in water and wastewater treatment processes. Environmental Science and Technology, 23(11), 1385-1389. Li, X.Y., and Logan, B. (1997). Collision freque ncies between fractal aggregates with small particles by differen tial sedimentation. Environmental Science and Technology, 31(4), 1229-1236. Lin, H. (2003). Granulometry, chemistry and physical inte ractions of non-colloidal particulate matter transported by urban storm water. Ph.D. dissertation, Baton Rouge, Louisiana State University, 155. Lin, J. M., Fang, G. C., Holsen, T. M., and No ll, K. E. (1993). A comparison of dry deposition modeled from size distribution da ta and measured with a smooth surface for total particle mass, lead and calcium in Chicago. Atmospheric Environment, 27A, 1131-1138. Logan, B., and Kilps, J. (1995). Fractal dimens ions of aggregates formed in different fluid mechanical environments. Water Research, 29(2), 443-453. Magnuson, J. J., Safina, C., a nd Sissenwine, M. P. (2001). Ecology and conservation whose fish are they anyway? Science, 293(5533), 1267-1268. Makepease, D. K., Smith, D. W. and Stanley, S. J. (1995). Urban stormw ater quality: summary of contaminant data. Environmental Scie nce and Technology, 25(2), 93-139. Mason, Y., Ammann, A. A., Ulrich, A., and Sigg, L. (1999). Behavior of heavy metals, nutrients, and major components dur ing roof runoff infiltration. Environmental Science and Technology, 33, 1588-1597. McBride, M. B., and Blasiak, J. J. (1979). Zinc and copper solubility as a function of pH in an acid soil. Soil Science Society of America Journal, 43(5), 866-870. McCave, I. N. (1975). Vertical fl ux of particles in the ocean. Deep-Sea Research, 22, 491-502.

PAGE 196

196 Metcalf and Eddy INC. (1991). Wastewater Engineering Tr eatment, Disposal, and Reuse, 3rd Edition. McGraw-Hill, INC., New York. Michelbach, S. and Whrle, C. (1993). Settleab le solids in a combined sewer system, settling characteristics, heavy metals, e fficiency of storm water tanks. Water Science Technology, 27(5-6), 153-164. Motelay-Massei, A., Ollivon, D., Tiphagne, K., Garban, B. (2005) Atmospheric bulk deposition of trace metals to the seine river basin: concentrations, sources and evolution from 1998 to 2001 in Paris. Water, Air, and Soil Pollution, 164, 119-135. Muschake, W. (1990) Pollution of street run-off by traffic and local conditions. Science of the Total Environment, 93, 419-431. Naidu, R., Sumner, M. E., and Harter, R. D. (1998). Sorption of heavy metals in strongly weathered soils: an overview. Environmental Geochemistry and Health, 20(1), 5-9. Nikora, V., Aberle, J., and Green, M. (2004). S ediment flocs: settling velocity, flocculation factor, and optical backscatter. Journal of Hydraulic Engineering, 130(10), 1043-1047. Noever, D., and Nikora, V. (1995). Rarefied solids. Microgravity Science and Technology, 7(4), 343. North Carolina Department of Natural Res ources and Community Development (NCDNRCD) (1983). An Evaluation of Street Sweeping as a Runoff Pollution Control. PB85102507, USEPA, Washington, D.C. Ozer, A. (1994). Simple equations to express settling column data. Journal of Environmental Engineering, 120(3), 677-682. Packman, J. J., Comings, K. J., and Booth, D. B. (1999). Using turbid ity to determine total suspended solids in urbanizing streams in the Puget Lowlands. Canadian Water Resources Association Annual Meeting, 158-165. Paode, R. D., Sofuoglu, S. C., Si vadechathep, J., Noll, K. E., Holsen, T. M., and Keeler, G. J. (1998). Dry deposition fluxes and mass size dist ributions of Pb, Cu, and Zn measured in Southern Lake Michigan during AEOLUS. Environmental Science and Technology, 32, 1629-1635. Pierzynski,G. M. (2000). Methods of phosphorus analysis fo r soils, sediments, residuals, and waters. ISBN: 1-58161-396-2, Southern Cooperative Series Bulletin, No. 396. Pisano, W.C., Connick, D. J., and Aronson, G. L. (1984). Swirl and helical bend regulator/concentrator for storm and combined sewer overflow control. EPA-600/284/151 (NTIS PB85-102 523), U.S. EPA. Pitt, R., (1982). Street particulate data collection and analysis. Report prepared for WASHOE Council of Governments.

PAGE 197

197 Ray, H. (1997). Street dirt as a phos phorus source in urban stormwater. Ph.D. dissertation, University of Alabama at Birmingham, 125. Reddy, K. R., and Patrick, W. H. (1977). Eff ect of placement and concentration of applied 15NH4-N on nitrogen lo ss from flooded soil. Soil Science, 123,142-147. Rice, K. C., and Bricker, O. P. (1992). Aci d rain and its effect on streamwater quality on Catoctin Mountain, Maryland. U.S. Geological Survey Open-File Report 92-168, 2. Richardson, J. F., and Zaki, W. N. (1954). Sedimentation and fluidization: part 1. Transaction of the American Institute of Chemical Engineers, 32, 35-53. Sa, S., Masuda, T., and Hosoi, Y. (2003). A study of SS size distribution during runoff and fractionation of phosphates depending on so il size in agricultural watershed. Diffuse Pollution Conference, 13(8), 13-15. Sabina, L. D., Limb, J. H., Stolzenbachb, K. D., Schiffa, K. C. (2005) Contribution of trace metals from atmospheric deposition to stor mwater runoff in a small impervious urban catchment. Water Research, 39, 3929-3937. Sakadevan, K., and Bavor, H. J. (1998). Phosphat e adsorption characteristics of soils, slags and zeolite to be used as substrates in constructed wetland systems. Water Research, 32(2), 393-399. Sansalone, J. J., and Buchberger, S. G. (1997). Partitioning and Fi rst Flush of Metals in Urban Roadway Storm Water. Journal of Environmental Engineering, 123(2), 134-143. Sansalone, J. J., and Cristina, C. M. (2004). Fir st flush concepts for suspended and dissolved solids in small impervious watersheds. Journal of Environmental Engineering, 130(11), 1301-1314. Sansalone, J. J., and Tribouillard, T. (1999). V ariation in characteristics of abraded roadway particles as a function of part icle size implications for water quality and drainage. Transportation Research Board, Washington, D.C. Sansalone, J. J., Buchberger, S. G., and Koechlin g, M. T. (1995). Correlations between heavy metals and suspended solids in highway runoff: R-18 implications for control strategies. Transportation Research Record N1483, 112-119. Sansalone, J. J., Hird, J. P., Cartledge, F. K., and Tittlebaum, M. E. (2005). Event-based stormwater quality and quantity loadings from elevated urba n infrastructure affected by transportation. Water Environmental Research, 77(4), 348-365. Sansalone, J. J., Koran, J. M., Smithson, J. A. (1998). Physical characteristics of urban roadway solids transported during rain event. Journal of Environmental Engineering, 124(5), 427440.

PAGE 198

198 Sanyal S. K., and DeDatta S. K. (1991). Chemi stry of phosphorus transformations in soils. Advanced Soil Science, 16, 1-50. Sartor, J. D., and Boyd, G. B. (1972). Water pol lution aspects of street surface contaminants. EPA-R2-72-081, USEPA, Washington, D.C. Sartor, J. D., and Gaboury, D. R. (1984). Street sweeping as a water poll ution control measure: lessons learnt over the last ten years. Science of the Total Environment, 33, 171-183. Sartor, J. D., Boyd, G. B., and Agardy, F. J. (1972). Water pollution aspects of street surface contaminants. Proceedings of the 45th Annual Conf erence of the Water Pollution Control Federation, Atlanta, Georgia, 26. Sartor, J. D., Boyd, G. B., and Agardy, F. J. (1974). Water pollution aspects of street surface contaminants. Journal Water Polluti on Control Federation, 46(3), 458-467. Schulz, S. F., Wilde, R. H., and Albertson, M. L. (1954). Influence of shape on the fall velocity of sedimentary particles. M.R.D. Sediment Service, No. 5, Missouri River Div., U.S. Army Corps of Egnrs. Shaheen, D. (1975). Contribution of urban roadway usage to water pollution. EPA-600/2-75004, USEPA, Washington, D.C. Sherwood, L. J., and Qualls, R. G. (2001). Stability of phosphorus within a wetland soil following ferric chloride treatm ent to control eutrophication. Environmental Science and Technology, 35, 4126-4131. Shinya, M., Tsuchinaga, T., Kitano, M., Yamada, Y., a nd Ishikawa, M. (2000). Characterization of heavy metals and polycyclic aromatic hydrocarbons in urban highway runoff. Water Science and Technology, 42(7-8), 201-208. Sinha, M. K., Dhilllon, S. K., Dhillon, K. S., and Dyanand, S. (1978). Solubility relationships of Fe, Mg, Cu, and Zn in alkaline and calcareous soils. Australian Journal of Soil Research, 16(1), 19-26. Smith, S. V. (1984). Phosphorus versus nitr ogen limitation in the marine environment. Limnology and Oceanography, 29(6), 1149-1161. Spivakov, B. Y., Maryutina, T. A., and Muntau H. (1999). Phosphorus sp eciation in water and sediments. Pure and Applied Chemistry, 71(11), 2161-2176. Stevenson, D. G. (1997). Flow and filtration th rough granular media The effect of grain and particle size dispersion. Water Research, 31(2), 310-322. Tanizaki, Y., Shimokawa, T., and Yamazaki, M. (1992). Physico-chemical speciation urban runoff program. Water Planning Division NTIS: PB84-185545.

PAGE 199

199 Thonon, I., Roberti, J. R., Middelkoop, H., van der Perk, M., Burrough, P. A. (2005). In situ measurements of sediment settling charac teristics in floodplains using a LISST-ST. Earth Surface Processes, 30(10), 1327-1343. Toet, C., Hvitvedjacobsen, T., Yousef, Y. A. ( 1990). Pollutant removal and eutrophication in urban runoff detention ponds. Water Science and Technology, 22(10-11), 197-204. Tomkins, M. R. (2005). Hinde red settling of sand grains. Sedimentology, 52, 1425-1432. Turer, D., Maynard, B. J., and Sansalone, J. J. (2001). Heavy metal contamination in soils of urban highways: Comparison between runoff and soil concentrations. Journal of Water, Air and Soil Pollution, 132(3), 214-230. Turner, F. T., and Patrick, W. H. (1968). Chemi cal changes in waterlogged soils as a result of oxygen depletion. International of Society of Soil Science Transport 9th Congress, 4, 5365. Tyack, J. N., Hedges, P. D., and Smisson, R. P. (1996). The relati onship between settling velocity grading and the characterist ics of the contributing catchment. Water Science and Technology, 33(9), 135-142. U.S. Environmental Protection Agency (1986). Quality Criteria for Water 1986. EPA 440/586-001, Office of Water Regulation and Standards, Washington, D.C. U.S. Environmental Protection Agency (1993) Handbook: Urban Runoff Pollution Prevention and Control Planning. EPA l625/R-93/004, Office of Research and Development, Washington, DC. U.S. Environmental Protection Agency (1999) Development of Bench-Scale Settling Apparatus: Settling Velocity Data for Desi gn and Operation of Wet-Weather Flow SolidsLiquid Separation Processes. EPA/600/X-99/031, Washington, D.C. U.S. Environmental Protection Agency (1999). National Pollutant Discharge Elimination System Regulations for Revision of the Water Pollution Control Program Addressing Storm Water Discharges. [FRL], RIN 2040AC82, Federal Register, Vol. 64, No. 235. U.S. Environmental Protection Agency (1999). U SEPA Issue Paper Measurement of TSS in Runoff. I:189-FR977001, URS Greiner Woodward Clyde. Vandenberg, J. J., and Knoerr, K. R. (1985). Comparison of surrogate surface techniques for estimation of sulfate dry deposition. Atmospheric Environment, 19(4), 627-635. Vaze, J. and Chiew, F. H. (2004). Nutrient load s associated with different sediment sizes in urban stormwater and surface pollutants. Journal of Environmental Engineering, 130(4), 391-396.

PAGE 200

200 Vaze, J., and Chiew F. H. (2003a). Study of pollutant wash-off from small impervious experimental plots. Water Resources Research, 39(6), Article No. 1160. Vaze, J., and Chiew F. H. (2003b). Comparati ve evaluation of urban storm water quality models. Water Resources Research, 39(10), Article No. 1280. Vermettea, S. J., Pedenb, M. E., Willoughbyc T. C., Lindbergd, S. E., Aaron, D., and Weisse, A. D. (1995). Methodology for the sampling of me tals in precipitation: Results of the national atmospheric deposition pr ogram (NADP) pilot network. Atmospheric Environment, 29(11), 1221-1229. Waschbusch, R. J., Selbig, W. R., and Bannerm an, R. T. (1999). Sources of phosphorus in Stormwater and street dirt fr om two urban residential basi ns in Madison, Wisconsin, 199495. Water-Resources Investigations Report 99-4021, U.S. Geological Survey, Middleton, Wisconsin. Wiesner, M. R. (1999). Morphology of particle deposits. Journal of Environmental Engineering, 125(12), 1124-1132. Wilber, W. G., and Hunter, J. V. (1979). The impact of urbanization on the distribution of heavy metals in bottom sediments of the Saddle River. Water Research Bulletin, 15(3), 790-800. Winterwerp, J. C. (1998). A simple model for turbulence induced flocculation of cohesive sediment. Journal of Hydraulic Research, 36(3), 309-326. Winterwerp, J. C. (2002). On the flocculati on and settling velocity of estuarine mud. Continental Shelf Research, 22, 1339-1360. Winterwerp, J. C., and Kranenburg, C. (2002) Fine sediment dynamics in the marine environment. eds., Elsevier, Amsterdam. Winterwerp, J. C., Manning, A. J., Martens, C., Mulder, T., and Vanlede, J. (2006). A heuristic formula for turbulence-induced floc culation of cohesive sediment. Estuarine, Coastal and Shelf Science, 68, 195-207. Wong, K. B., and Piedrahita, R. H. (2000). Set tling velocity characte rization of aquacultural solids. Aquacultural Engineering, 21, 233-246. Woodfield D., and Bickert G. ( 2001) An improved permeability model for fractal aggregates settling in creeping flow. Water Research, 35(16), 3801-3806. Woodward-Clyde (1994). San Jose st reet sweeping equipment evaluation. Report prepared for City of San Jose, California. Wu, J. S., Allan C. J., Saunders, W. L., and Evett, J. B. (1998). Charact erization and pollutant loading estimation for highway runoff. Journal of Environmental Engineering, 124(7), 584-592.

PAGE 201

201 Yeoman, S., Lester, J. N., and Perry, R. ( 1992). Phosphorus removal and its influence on particle size and heavy metal distri bution during wastewater treatment. Environmental Technology, 13, 901-924. Yeoman, S., Sterritt, R. M., Rudd, T., and Lest er, J. N. (1989). Parti cle size fractionation and metal distribution in sewage sludges. Water, Air and Soil Pollution, 45, 27-42. Yousef, Y. A., Hvitved-Jacobsen, T., Wanielista, M. P., and Harper H. H. (1987). Removal of contaminants in highway runoff through swales. The Science of the Total Environment, 59, 391-397. Yun, H. J., Yi, S. M., and Kim, Y. P. (2002). Dry deposition fl uxes of ambient particulate heavy metals in a small city, Korea. Atmospheric Environment, 36, 5449-5458.

PAGE 202

202 BIOGRAPHICAL SKETCH Gaoxiang Ying received his bachelors degr ee in environmental engineering from Tsinghua University, Beijing, P.R.China in July 1999. From the same university, he got his masters degree in environmental engineering in July 2002, under the guidance of Dr. Yongqi Lu in the Department of Environmental Science and Engineering. Gaoxiang Ying will receive his Doctor of Philosophy in Environmental Engineering Sciences from the University of Florida in August 2007. His doctoral research was focused on c onstitutive properties of particulates in urban dry deposition and source ar ea rainfall-runoff loadings. He worked under the guidance of Dr. John J. Sansalone in the Department of Environmental Engineering and Sciences.