Risk assessment of pesticide runoff from turf

The TurfPQ model was used to simulate the runoff of 15 pesticides commonly applied to creeping bentgrass (Agrostis stolonifera L.) fairways and greens on golf courses in the northeastern USA. Simulations produced 100-yr daily records of water runoff, pesticide runoff, and pesticide concentration in runoff for three locations: Boston, MA, Philadelphia, PA, and Rochester, NY. Results were summarized as annual and monthly means and annual maximum daily loads (AMDLs) corresponding to 10- and 20-yr return periods. Mean annual pesticide runoff loads did not exceed 3% of annual applications for any pesticide or site, and most losses were substantially less than 1% of application. However, annual or monthly mean concentrations of chlorothalonil, iprodione, and PCNB in fairway runoff often exceeded concentrations that result in 50% mortality of the affected species (LC50) for aquatic organisms. Concentrations of azoxystrobin, bensulide, cyfluthrin, and trichlorfon in extreme (1 in 10 yr or 1 in 20 yr) events often approached or exceeded LC50 levels. Concentrations of halofenozide, mancozeb, MCPP, oxadiazon, propiconazole, thiophanate-methyl, triadimefon, and trinexapac-ethyl were well below LC50 levels, and turf runoff of these chemicals does not appear to be hazardous to aquatic life in surface waters.     

Comparison of models of simazine transport and fate in the subsurface environment in a citrus farm

Contamination of groundwater by agrochemicals is now widely recognized as an extremely important environmental problem. Modern agricultural practices involve the combined use of irrigation with the application of large amounts of agrochemicals to maximize crop yield. Due to flood irrigation and natural runoff, agricultural activities might generate soil, surface water and groundwater contamination problems and leaching of pesticides. Modeling of the transport and fate of pesticides, such as simazine, may help understand the long-term potential risk to the subsurface environment. This paper illustrates a comparative study via the use of three different pesticide transport simulation models and the applicability of those models in determining the groundwater vulnerability to pesticides contamination in a citrus orchard located at the Lower Rio Grande Valley (LRGV). The three models used in the study are the pesticide root zone model-3 (PRZM-3), the pesticide analytical model (PESTAN) and integrated pesticide transport modeling (IPTM). The concentration values obtained from all three models are in agreement, and they show a decreasing trend from the surface through the vadose zone. The problem is how to use this information and, specifically, how to combine the testimony of a number of experts into a single useful judgment. With the aid of the fuzzy multiattribute decision making method, PRZM-3 is deemed as the most promising one for such precision farming applications.     

Estimating turf pesticide volatilization from simple evapotranspiration models

A previously developed model by Haith et al. (2002) related pesticide volatilization from turf to evapotranspiration (ET) by scaling factors determined from vapor pressures and heats of vaporization. Although the model provided volatilization estimates that compared well with field measurements, it relied on the Penman ET equation, requiring hourly temperature, wind speed, and solar radiation data, none of which are routinely available at field sites. The current study determined that the volatilization model works equally well with a simpler ET equation requiring only daily temperatures. Three daily temperature-based ET models were evaluated as vehicles for estimating pesticide volatilization from turf: Hamon, Hargreaves-Samani, and a modified Priestley-Taylor. When compared with field volatilization measurements for eight pesticides, volatilization estimates produced from the Hargreaves-Samani model most closely approximated both the field observations and the previous estimates based on the more data-intensive Penman model. Mean estimated volatilization exceeded mean observations by 15% and the coefficient of variation (R2) between estimates and observations was 0.65. The comparable values based on Penman ET were 17% and 0.63, respectively.     

Modeling pesticide volatilization from turf

Pesticide volatilization models are typically based on equilibrium partitioning of the chemical into solid, liquid, and gaseous phases in the soil environment. In turf systems direct vaporization from vegetation surfaces is a more likely source, and it is difficult to apply equilibrium methods to plant material due to the uncertainties of solid-liquid-gas partitioning. An alternative approach is to assume that pesticide volatilization is governed by the same processes that affect water evaporation. A model was developed in which evapotranspiration values, as determined by the Penman equation, were adjusted to chemical vaporization using ratios of water and chemical saturated vapor pressures and latent heats of vaporization. The model also assumes first-order degradation of pesticide on turf vegetation over time. The model was tested by comparisons of predictions with measurements of volatilization for eight pesticides measured during 3 to 7 d in 11 field experiments. Measured volatilization fluxes ranged from 0.1 to 22% of applied chemical. Pesticides were divided into two groups based on saturated vapor pressures and organic C partition coefficients. One pesticide was selected from each group to calibrate the model's volatilization constant for the group, and the remaining pesticides were used for model testing. Testing results indicated that the model provides relatively conservative estimates of pesticide volatilization. Predicted mean losses exceeded observations by 20%, and the model explained 67% of the observed variation in volatilization fluxes. The model was most accurate for those chemicals that exhibited the largest volatilization losses.     

Runoff Simulation of the Headwaters of the Yellow River Using The SWAT Model With Three Snowmelt Algorithms1

Abstract: The Soil and Water Assessment Tool (SWAT) model combined with different snowmelt algorithms was evaluated for runoff simulation of an 114,345 km² mountainous river basin (the headwaters of the Yellow River), where snowmelt is a significant process. The three snowmelt algorithms incorporated into SWAT were as follows: (1) the temperature‐index, (2) the temperature‐index plus elevation band, and (3) the energy budget based SNOW17. The SNOW17 is more complex than the temperature‐based snowmelt algorithms, and requires more detailed meteorological and topographical inputs. In order to apply the SNOW17 in the SWAT framework, SWAT was modified to operate at the pixel scale rather than the normal Hydrologic Response Unit scale. The three snowmelt algorithms were evaluated under two parameter scenarios, the default and the calibrated parameters scenarios. Under the default parameters scenario, the parameter values were determined based on a review of the current literature. The purpose of this type of evaluation was to assess the applicability of SWAT in ungauged basins, where there is little observed data available for calibration. Under the calibrated parameters scenario, the parameters were calibrated using an automatic calibration program, the Shuffled Complex Evolution (SCE‐UA). The purpose of this type of evaluation was to assess the applicability of SWAT in gauged basins. Two time periods (1975‐1985 and 1986‐1990) of monthly runoff data were used in this study to evaluate the performance of SWAT with different snowmelt algorithms. Under the default parameters scenario, the SWAT model with complex energy budget based SNOW17 performed the best for both time periods. Under the calibrated parameters scenario, the parameters were calibrated using monthly runoff from 1975‐1985 and validated using monthly runoff from 1986‐1990. After parameter calibration, the performance of SWAT with the three snowmelt algorithms was improved from the default parameters scenario. Further, the SWAT model with temperature‐index plus elevation band performed as well as the SWAT model with SNOW17. The SWAT model with temperature‐index algorithm performed the poorest for both time periods under both scenarios. Therefore, it is suggested that the SNOW17 model be used for modeling ungauged basins; however, for gauged basins, the SNOW17 and simple temperature‐index plus elevation band models could provide almost equally good runoff simulation results.     

SWMHMS - SMALL WATERSHED MONTHLY HYDROLOGIC MODELING SYSTEM1

ABSTRACT: SWMHMS is a conceptual computer modeling program developed to simulate monthly runoff from a small nonurban watershed. The input needed to run model simulations include daily precipitation, monthly data for evapotranspiration determination (average temperature, crop consumptive coefficients, and percent daylight hours), and six watershed parameter values. Evapotranspiration was calculated with the Blaney-Criddle equation while surface runoff was determined using the Soil Conservation Service curve number procedure. For watershed parameter evaluation, SWMHMS provides options for both optimization and sensitivity analysis. Observed runoff data are required along with the model input previously mentioned in order to conduct parameter optimization. SWMEIMS was tested with data from six watersheds located in different regions of the United States. Model accuracy was generally found to be very good except on watersheds having substantial snowfall accumulation. In having only six watershed parameters, SWMHMS is less complex to use than many other computer programs that calculate monthly runoff. Consequently, SWMHMS may find its greatest application as an educational tool for students learning principles of hydrologic modeling, such as parameter evaluation procedures and the impacts of input data uncertainty on model results.     

USING CURVE NUMBERS TO DETERMINE BASELINE VALUES OF GREEN-AMPT EFFECTIVE HYDRAULIC CONDUCTIVITIES1

ABSTRACT: Since the trend in infiltration modeling is currently toward process-based approaches such as the Green-Ampt equation, more emphasis is being placed on methods of determining appropriate parameters for this approach. The SCS curve number method is an accepted and commonly used empirical approach for estimating surface runoff, and is based on numerous data from a variety of sources. The time and expense of calibrating process-based infiltration parameters to measured data are often prohibitive. This study uses curve number predictions of runoff to develop equations to estimate the “baseline” hydraulic conductivities (K_b) for use in the Green-Ampt equation. Curve number predictions of runoff were made for 43 soils. K_b values in the Water Erosion Prediction Project (WEPP) model were then calibrated so that the annual runoff predicted by WEPP was equal to the curve number predictions. These calibrated values were used to derive an equation that estimated K_b based on the percent sand, percent clay, and cation exchange capacity of the soil. Estimated values of K_b from this equation compared favorably with measured values and values calibrated to measured natural runoff plot data. WEPP predictions of runoff using both optimized and estimated values of K_b were compared to curve number predictions of runoff and the measured values. The WEPP predictions using the optimized values of K_b were the best in terms of both average error and model efficiency. WEPP predictions using estimated values of K_b were shown to be superior to predictions obtained from the curve number method. The runoff predictions all tended to be biased high for small events and low for larger events when compared to the measured data. Confidence intervals for runoff predictions on both an annual and event basis were also developed for the WEPP model.     

Application of land-use data and screening tests for evaluating pesticide runoff toxicity in surface waters

Survey information on pesticide usage in New Zealand during 1985–1989 is summarized by regions and principal applications. Two screening tests, one based on a simple water-balance method and the other based on a semiempirical runoff formula, have been used to identify 18 pesticides with application rates that may yield runoff concentrations that are harmful to aquatic fauna. These are predominantly associated either with intensive applications in horticulture or extensive applications to cereal crops and pasture. The purpose of the screening tests was to calculate typical edge-of-field concentrations in runoff and, by comparing them with known aquatic toxicity values, determine which compounds are applied at rates that may yield toxic runoff. While it may be possible to extend these methods to calculate typical surface water concentrations, further studies will be needed to evaluate pesticide persistence and assimilation in stream channels.     

HYDROLOGIC PARAMETERIZATION OF WATERSHEDS FOR RUNOFF PREDICTION USING SWAT1

ABSTRACT: The use of continuous time, distributed parameter hydrologic models like SWAT (Soil and Water Assessment Tool) has opened several opportunities to improve watershed modeling accuracy. However, it has also placed a heavy burden on users with respect to the amount of work involved in parameterizing the watershed in general and in adequately representing the spatial variability of the watershed in particular. Recent developments in Geographical Information Systems (GIS) have alleviated some of the difficulties associated with managing spatial data. However, the user must still choose among various parameterization approaches that are available within the model. This paper describes the important parameterization issues involved when modeling watershed hydrology for runoff prediction using SWAT with emphasis on how to improve model performance without resorting to tedious and arbitrary parameter by parameter calibration. Synthetic and actual watersheds in Indiana and Mississippi were used to illustrate the sensitivity of runoff prediction to spatial variability, watershed decomposition, and spatial and temporal adjustment of curve numbers and return flow contribution. SWAT was also used to predict stream runoff from actual watersheds in Indiana that have extensive subsurface drainage. The results of this study provide useful information for improving SWAT performance in terms of stream runoff prediction in a manner that is particularly useful for modeling ungaged watersheds wherein observed data for calibration is not available.     

RUNOFF MODELING OF A MOUNTAINOUS CATCHMENT USING TOPMODEL: A CASE STUDY1

ABSTRACT: The rainfall‐runoff response of the Tygarts Creek Catchment in eastern Kentucky is studied using TOPMODEL, a hydrologic model that simulates runoff at the catchment outlet based on the concepts of saturation excess overland flow and subsurface flow. Unlike the traditional application of this model to continuous rainfall‐runoff data, the use of TOPMOEL in single event runoff modeling, specifically floods, is explored here. TOPMODEL utilizes a topographic index as an indicator of the likely spatial distribution of rainfall excess generation in the catchment. The topographic index values within the catchment are determined using the digital terrain analysis procedures in conjunction with digital elevation model (DEM) data. Select parameters in TOPMODEL are calibrated using an iterative procedure to obtain the best‐fit runoff hydrograph. The calibrated parameters are the surface transmissivity, T_O, the transmissivity decay parameter, m, and the initial moisture deficit in the root zone, S_r0. These parameters are calibrated using three storm events and verified using three additional storm events. Overall, the calibration results obtained in this study are in general agreement with the results documented from previous studies using TOPMODEL. However, the parameter values did not perform well during the verification phase of this study.     

Impact of data quality and model complexity on prediction of pesticide leaching

Accurate input data for leaching models are expensive and difficult to obtain which may lead to the use of "general" non-site-specific input data. This study investigated the effect of using different quality data on model outputs. Three models of varying complexity, GLEAMS, LEACHM, and HYDRUS-2D, were used to simulate pesticide leaching at a field trial near Hamilton, New Zealand, on an allophanic silt loam using input data of varying quality. Each model was run for four different pesticides (hexazinone, procymidone, picloram and triclopyr); three different sets of pesticide sorption and degradation parameters (i.e., site optimized, laboratory derived, and sourced from the USDA Pesticide Properties Database); and three different sets of soil physical data of varying quality (i.e., site specific, regional database, and particle size distribution data). We found that the selection of site-optimized pesticide sorption (Koc) and degradation parameters (half-life), compared to the use of more general database derived values, had significantly more impact than the quality of the soil input data used, but interestingly also more impact than the choice of the models. Models run with pesticide sorption and degradation parameters derived from observed solute concentrations data provided simulation outputs with goodness-of-fit values closest to optimum, followed by laboratory-derived parameters, with the USDA parameters providing the least accurate simulations. In general, when using pesticide sorption and degradation parameters optimized from site solute concentrations, the more complex models (LEACHM and HYDRUS-2D) were more accurate. However, when using USDA database derived parameters, all models performed about equally.     

THE APPLICATION OF HYDROLOGIC MODELS TO SMALL WATERSHEDS HAVING MILD TOPOGRAPHY1

ABSTRACT: The application of hydrologic models to small watersheds of mild topography is not well documented. This study evaluates the applicability of hydrologic models described by Huggins and the Soil Conservation Service to small watersheds by comparing the simulated and actual hydrograph for both gaged and ungaged situations. The annual maximum rainfall events plus storms exceeding 2.5 inches from 25 years of rainfall and runoff data for two small watersheds were selected for the model evaluations. These storms had a variety of patterns and occurred on many different watershed conditions. Simulated and actual hydrographs were compared using a parameter which contained volume, peak, and shape factors. One-half of the selected storms were used to calibrate the models. For both models, there were no significant differences between the simulated and actual runoff volumes and peak runoff rates. Parameters obtained during the calibration process and relationships developed to estimate antecedent moisture and to modify tabulated runoff curve numbers were used to simulate the runoff hydrograph from the remaining storms. These remaining storms or test storms were simulated only once in order to imitate an ungaged situation. In general, both the Huggins and SCS model performed similarly on the test storms, but the level of model performance was lower than that for the calibration storms. For both models, the two-day antecedent rainfall was more important than the five-day in determining antecedent moisture and modifying tabulated curve numbers. The time of concentration which resulted in good hydrograph simulations was about three times larger than that estimated using published empirical relationships.     

Runoff transport of pyrethroids from a residential lawn in central California

An irrigation runoff study on a residential lawn was conducted in California, northeast of Sacramento, during the summer and fall of 2008 to investigate the contribution of turf uses of pyrethroids to residues in Californian urban creek sediments. This study examined how over irrigation (i.e., irrigation that produces runoff) in the summer season may transport recently applied pyrethroids. The study included liquid and granular applications of both bifenthrin [(2-methyl-3-phenyl-phenyl) methyl 3-(2-chloro-3,3,3-trifluoro-prop-1-enyl)-2,2-dimethyl-cyclopropane-1-carboxylate] and beta-cyfluthrin [Cyano(4-fluoro-3-phenoxyphenyl)methyl 3-(2,2-dichloroethenyl)-2,2-dimethyl-cyclopropanecarboxylate]. Generally, runoff did not occur at irrigation rates of 2.03 cm/h (0.8 in/h) but did occur when the irrigation rates were increased to about 3.81 cm/h (1.5 in/h), generating chemical losses in the first runoff event of up to 0.58 and 0.08% of applied for beta-cyfluthrin and bifenthrin, respectively. Chemical runoff losses dropped significantly between over-irrigation events with the third over-irrigation event chemical runoff losses representing 0.026 and 0.015% of applied for beta-cyfluthrin and bifenthrin, respectively. Runoff losses were generally less for liquid formulations than granular formulations but within a factor of three. Additionally, the study included a simulated winter rainstorm 8 wk after application. The low runoff losses from turf seen in this study suggest that other sources could be contributing to observed residues in urban streams. Other sources could include pyrethroids ending up on impervious surfaces, such as concrete driveways from off-target applications to turf, spills, and other poor handling practices, or pyrechroids applied directly to impervious surfaces for insect control.     

Estimating Watershed Level Nonagricultural Pesticide Use From Golf Courses Using Geospatial Methods1

Abstract: Limited information exists on pesticide use for nonagricultural purposes, making it difficult to estimate pesticide loadings from nonagricultural sources to surface water and to conduct environmental risk assessments. A method was developed to estimate the amount of pesticide use on recreational turf grasses, specifically golf course turf grasses, for watersheds located throughout the conterminous United States (U.S.). The approach estimates pesticide use: (1) based on the area of recreational turf grasses (used as a surrogate for turf associated with golf courses) within the watershed, which was derived from maps of land cover, and (2) from data on the location and average treatable area of golf courses. The area of golf course turf grasses determined from these two methods was used to calculate the percentage of each watershed planted in golf course turf grass (percent crop area, or PCA). Turf‐grass PCAs derived from the two methods were used with recommended application rates provided on pesticide labels to estimate total pesticide use on recreational turf within 1,606 watersheds associated with surface‐water sources of drinking water. These pesticide use estimates made from label rates and PCAs were compared to use estimates from industry sales data on the amount of each pesticide sold for use within the watershed. The PCAs derived from the land‐cover data had an average value of 0.4% of a watershed with minimum of 0.01% and a maximum of 9.8%, whereas the PCA values that are based on the number of golf courses in a watershed had an average of 0.3% of a watershed with a minimum of <0.01% and a maximum of 14.2%. Both the land‐cover method and the number of golf courses method produced similar PCA distributions, suggesting that either technique may be used to provide a PCA estimate for recreational turf. The average and maximum PCAs generally correlated to watershed size, with the highest PCAs estimated for small watersheds. Using watershed specific PCAs, combined with label rates, resulted in greater than two orders of magnitude over‐estimation of the pesticide use compared to estimates from sales data.     

Runoff loss of pesticides and soil: a comparison between vegetative mulch and plastic mulch in vegetable production systems

Current vegetable production systems use polyethylene (plastic) mulch and require multiple applications of agrochemicals. During rain events, runoff from vegetable production is enhanced because 50 to 75% of the field is covered with an impervious surface. This study was conducted to quantify off-site movement of soil and pesticides with runoff from tomato (Lycopersicon esculentum Mill.) plots containing polyethylene mulch and a vegetative mulch, hairy vetch (Vicia villosa Roth). Side-by-side field plots were instrumented with automated flow meters and samplers to measure and collect runoff, which was filtered, extracted, and analyzed to determine soil and pesticide loss. Seasonal losses of two to four times more water and at least three times as much sediment were observed from plots with polyethvlene mulch (55.4 to 146 L m(-2) and 247 to 535 g m(-2), respectively) versus plots with hairy vetch residue (13.7 to 75.7 L m(-2) and 32.8 to 118 g m(-2), respectively). Geometric means (+/-standard deviation) of total pesticide loads for chlorothalonil (tetrachloroisophthalonitrile) and alpha-and beta-endosulfan (6,7,8,9,10,10-hexachloro-1,5,5a,6,9,9a-hexahydro6,9-methano-2,4,3-benzodioxathiepin 3-oxide) for a runoff event were 19, 6, and 9 times greater from polyethylene (800+/-4.6, 17.6+/-3.9, and 39.1+/-4.9 microg m(-2), respectively) than from hairy vetch mulch plots (42+/-6.0, 2.8+/-5.0, and 4.3+/-4.6 microg m(-2), respectively) due to greater concentrations and larger runoff volumes. The increased runoff volume, soil loss, and off-site loading of pesticides measured in runoff from the polyethylene mulch suggests that this management practice is less sustainable and may have a harmful effect on the environment.     

Response of turf and quality of water runoff to manure and fertilizer

Manure applications can benefit turfgrass production and unused nutrients in manure residues can be exported through sod harvests. Yet, nutrients near the soil surface could be transported in surface runoff. Our research objective was to evaluate responses of bermudagrass [Cynodon dactylon (L.) Pers. var. Guymon] turf and volumes and P and N concentrations of surface runoff after fertilizer or composted manure applications. Three replications of five treatments were established on a Boonville fine sandy loam (fine, smectitic, thermic Vertic Albaqualf) that was excavated to create an 8.5% slope. Manure rates of 50 and 100 kg P ha(-1) at the start of two monitoring periods were compared with P fertilizer rates of 25 and 50 kg ha(-1) and an unfertilized control. Compared with initial soil tests, nitrate concentrations decreased and P concentrations increased after two manure or fertilizer applications and eight rain events over the two monitoring periods. The fertilizer sources of P and N produced 19% more dry weight and 21% larger N concentrations in grass clippings than manure sources. Yet, runoff volumes were similar between manure and fertilizer sources of P. Dissolved P concentration (30 mg L(-1)) in runoff during a rain event 3 d after application of 50 kg P ha(-1) was five times greater for fertilizer than for manure P. Observations during both monitoring periods indicated that total P and N losses in runoff were no greater for composted manure than for fertilizer sources of P at relatively large P rates on a steep slope of turfgrass.     

HYDROLOGIC MODELING OF A BIOINFILTRATION BEST MANAGEMENT PRACTICE1

ABSTRACT: The goal of this research was to develop a methodology for modeling a bioinfiltration best management practice (BMP) built in a dormitory area on the campus of Villanova University in Pennsylvania. The objectives were to quantify the behavior of the BMP through the different seasons and rainfall events; better understand the physical processes governing the system's behavior; and develop design criteria. The BMP was constructed in 2001 by excavating within an existing traffic island, backfilling with a sand/soil mixture, and planting with salt tolerant grasses and shrubs native to the Atlantic shore. It receives runoff from the asphalt (0.26 hectare) and turf (0.27 hectare) surfaces of the watershed. Monitoring supported by the hydrologic model shows that the facility infiltrates a significant fraction of the annual precipitation, substantially reducing the delivery of nonpoint source pollution and erosive surges downstream. A hydrologic model was developed using HEC‐HMS to represent the site and the BMP using Green‐Ampt and kinematic wave methods. Instruments allow comparison of the modeled and measured water budget parameters. The model, incorporating seasonally variable parameters, predicts the volumes infiltrated and bypassed by the BMP, confirming the applicability of the selected methods for the analysis of bioinfiltration BMPs.     

COMPARISON OF 1-D AND 2-D MODELING OF OVERLAND RUNOFF AND SEDIMENT TRANSPORT1

ABSTRACT: One-dimensional and two-dimensional modeling approaches were compared for their abilities in predicting overland runoff and sediment transport. Both 1-D and 2-D models were developed to test the hypothesis that the 2-D modeling approach could improve the model predictions over the 1-P approach, based on the same mathematical representations of physical processes for runoff and sediment transport. The models developed in this study were applied to overland areas with cross slopes. A hypothetical case and an experimental study reported by Storm (1991) were used. Based on the simulation results from the selected hypothetical case and experimental study, the 2-D model provided better representation of spatial distribution of flow depths and sediment concentrations than the 1-D model. However, no significant differences in predictions of total runoff volume and sediment yield at the outlet area were found between the 1-D and 2-D models.