Characteristics of road sediment fractionated particles captured from paved surfaces, surface run-off and detention basins.

This study presents the results of evaluating changes in properties of road sediment fractionated particles at the source (dry pavement), during transport (highway runoff) and after deposition (dry detention basin). A total of 38 solid particle samples were collected from paved surface, highway runoff, and from three locations within detention basins. Each sample was size-fractionated ranging from 38 µm to 1000 µm. Key results were that: (i) less than 2% of total particle mass in four vacuumed solid samples was associated with the <38 µm fraction compared to 47 to 82% in centrifuged highway runoff samples and 25% for detention basin sediments, (ii) based on number concentrations more than 90% of particles from all sources were smaller than 38 µm, (iii) the densities of fractionated particles were generally within the range from 1.5 to 2.2g/cm(3); (iv) all collected particles were predicted to resist aggregation in solution with a typical measure of surface potential, the zeta potential, typically ranging from -15 to -30 mV, (v) metal concentrations increased with decreasing particle size for vacuumed samples from the highway shoulder and parking lot, however, size dependent trends in metal concentrations were less apparent in centrifuged highway runoff samples and detention basin sediment samples; (vi) the distributions of metal mass across sieved size fractions generally followed patterns of particle mass distribution in which less than 10% of the total Cu, Pb, and Zn mass was associated with particles <38 µm; (vii) metal mass distributions tended toward increasing metal fractions in finer particle fractions along a gradient from the inlet to the outlet of the detention basins; and (viii) individual particle morphology (1 < d(p) < 10 µm) showed that particles were not smooth or spherical, clearly deviating from the assumptions typically used in estimating particle settling velocities using Stokes' Law.

Toxicity of urban highway runoff with respect to storm duration.

The toxicity of stormwater runoff during various time-based stages was measured in both grab and composite samples collected from three highly urbanized highway sites in Los Angeles, California between 2002 and 2005. Stormwater runoff samples were tested for toxicity using three freshwater species (the water flea Ceriodaphnia dubia, the fathead minnow Pimephales promelas, and the green algae Pseudokirchneriella subcapitatum) and two marine species (the purple sea urchin Strongylocentrotus purpuratus, and the luminescent bacteria Photobacterium phosphoreum using Microtox. Toxicity results varied substantially throughout the storm events for both freshwater and marine species toxicity tests. In general, however, the first few samples were found to be more toxic compared with those collected during later stages of each storm event. In most cases, more than 40% of the toxicity was associated with the first 20% of discharged runoff volume. Furthermore, on average, 90% of the toxicity was observed during the first 30% of storm duration. Toxicity identification evaluation results found copper and zinc to be the primary cause of toxicity in about 90% of the samples evaluated with these procedures. Surfactants were also found to be the cause of toxicity in less than 10% of the samples.

Characterization and prediction of highway runoff constituent event mean concentration.

Highway stormwater runoff quality data were collected from throughout California during 2000-2003. Samples were analyzed for conventional pollutants (pH, conductivity, hardness, and temperature); aggregates (TSS, TDS, TOC, DOC); total and dissolved metals (As, Cd, Cr, Cu, Ni, Pb, and Zn); and nutrients (NO(3)-N, TKN, total P, and ortho-P). Storm event and site characteristics for each sampling site were recorded. A statistical summary for chemical characteristics of highway runoff is provided based on statewide urban and non-urban highways. Constituent event mean concentrations (EMCs) were generally higher in urban highways than in non-urban highways. The chemical characteristics of highway runoff in California were compared with national highway runoff chemical characterization data. The results obtained in California were generally similar to those found in other states. The median EMC for Pb measured in studies conducted in previous decades was much higher than the current median Pb EMC in California. The lower Pb EMC in California compared to previous highway runoff monitoring is believed to be due to the elimination of leaded gasoline. An attempt was also made to identify surrogate constituents within a general family of water quality categories using Spearman correlations and selected pairs with Spearman coefficients greater than 0.8. The strongest correlations were observed among parameters associated with dissolved minerals (EC, TDS, and chloride); organic carbon (TOC and DOC); petroleum hydrocarbons (TPH and O & G); and particulate matter (TSS and turbidity). Within the metals category, total iron concentration was highly correlated with most total metal concentrations. The correlations between total and dissolved concentrations were all less than 0.8, even between total and dissolved concentrations of the same metals. Multiple linear regression (MLR) analyses were performed to evaluate the impact of various site and storm event variables on highway runoff constituent EMCs. Parameters found to have significant impacts on highway runoff constituent EMCs include: total event rainfall (TER); cumulative seasonal rainfall (CSR); antecedent dry period (ADP); contributing drainage area (DA); and annual average daily traffic (AADT). Surrounding land use and geographic regions were also determined to have a significant impact on runoff quality. The MLR model was also used to predict constituent EMCs. Model performance determined by comparing predicted and measured values showed good agreement for most constituents.

Characteristics of litter waste in highway storm runoff.

Litter characterization is an integrated part of the Caltrans First Flush Characterization Study. These data will provide a basis to develop potential treatment technologies and best management practices to control pollutants in runoff from freeways. During monitoring periods in Southern California areas, the first flush phenomenon was evaluated and the impacts of various parameters such as rain intensity, drainage area, peak flow rate, and antecedent dry period on litter volume and loading rates were evaluated. First flush phenomenon was generally observed for litter concentrations, but was not apparent with litter mass loading rates. Total captured gross pollutants, defined as larger than 0.5 cm, was 90% vegetation with only 10% being litter. The normalized cumulative litter loadings were determined from 1.25 to 13.39 kg/ha for dry litter weight and 0.40 to 8.99 kg/ha for dry biodegradable litter weight. The portions of biodegradable litter to non-biodegradable litter were roughly the same across the entire event. Event mean concentrations were ranged 0.0021 to 0.259 g/L for wet gross pollutants, 0.0001 to 0.027 g/L for wet litters and 0.00007 to 0.018 g/L for dry litters. The mass emission rates should be useful to estimate total litter production for developing total maximum daily loads.

Correlation analysis among highway stormwater pollutants and characteristics.

Stormwater runoff from highway land use is a common non-point source of pollutants. A large quantity of highway stormwater runoff characteristics were collected in California during the past three years. Correlations among various water quality parameters and constituents were performed using data sets collected over the 2000-2001, 2001-2002, and 2002-2003 wet seasons for 18, 21 and 23 storm events at three highway sites in west Los Angeles, California. In addition, statistical and graphical correlation analysis of the mass first flush ratio (MFF) with storm characteristics was made to determine if the first flush is related to site or storm characteristics. The results and analyses performed indicate that (1) TSS correlates well with most particulate-bound metals. However, TSS was poorly correlated with most other pollutants. (2) Strong correlations were also observed among dissolved and total metals; DOC, COD, TKN and oil and grease; conductivity and Cl. (3) Total metals, COD and DOC were generally well correlated with mass first flush, suggesting that BMPs that treat the early portion of runoff have an opportunity to remove high concentrations of these pollutants.

First flush and natural aggregation of particles in highway runoff.

Particle Size Distribution (PSD) in highway runoff was monitored in the 2004-2005 rainy season at three highway sites in west Los Angeles, California. PSD was measured for 200 grab samples for 18 storm events. Particles and especially larger particles showed a strong first flush. On average, the initial 20% runoff volume transported approximately 28% total number of particles between 0.5 and 2 microm in diameter, more than 30% of particles between 2 and 30 microm and more than 40% of particles larger than 30 microm. A naturally occurring aggregation was observed with smaller particles and mixing experiments were performed to determine the possible benefits for sedimentation and filtration. Samples composited from grab samples manually collected over the first hour of runoff were gently mixed (G = 38) and small particle concentration decreased by more than 50%. After 24 hours the number of particles with diameter between 0.5 and 7 microm decreased by 51% with gentle mixing and the same size particles decreased by only 14% without mixing. Number of particles with diameter larger than 20 microm increased by 6 and 4.5 times with and without mixing, respectively. Slow mixing can improve sedimentation efficiency by more than 40% for particles less than 20 microm in diameter.

A new modeling approach for estimating first flush metal mass loading.

The purpose of this study was to investigate pollutant mass loading from major highways in Southern California, with emphasis on interpretation of event mean concentrations and first flush effects. The results of monitoring eight sites during the 1999-2002 storm seasons found that metal contaminants had higher concentrations at the early stages of storm events compared with other stages of rain storms. A new washoff model was developed to predict the event mean concentrations of metal contaminants taking first flush effect into account. Model variable parameters included average daily traffic, antecedent dry period, rain intensity, total runoff volume, and runoff coefficient. The results obtained using the washoff model were compared with measured values and found to fit well for heavy metals with R2 ranging from 0.8 to 0.95.

Impact of non-detects in water quality data on estimation of constituent mass loading.

Often, fractions of stormwater constituents are not detected above laboratory reporting limits and are reported as non-detect (ND), or censored data. Analysts and stormwater modelers represent these NDs in stormwater data sets using a variety of methods. Application of these different methods results in different estimates of constituent mean concentrations that will, in turn, affect mass loading computations. In this paper, different methods of data analysis were introduced to determine constituent mean concentrations from water quality datasets that include ND values. Depending on the number of NDs and the method of data analysis, differences ranging from 1 to 70 percent have been observed in mean values. Differences in mean values were, as shown by simulation, found to have significant impacts on estimations of constituent mass loading.