Search

Georgia farmers are increasing preemergence applications of soil residual herbicides to control glyphosate resistant weeds. To improve efficacy these herbicides are often activated by post-application irrigation. Broiler litter is commonly applied to fields before these herbicides. The herbicide wetting-in practice increases surface soil water content and may increase runoff and transport of broiler litter borne fecal bacteria into surface waters during subsequent storm events. Our objective was to determine differences in loads of fecal bacteria, Escherichia coli and Salmonella spp., in runoff from conventional tillage (CT) and no-till (NT) systems after herbicides were watered into an Ultisol fertilized with broiler litter. On replicated 6 m 2 -plots ( n = 3) simulated rainfall was applied for 70 min with composite runoff samples collected every 5 min and analyzed for E. coli and Salmonella spp. Although total runoff volume from the CT plots was significantly greater than from NT plots, no significant differences in total load of E. coli and Salmonella or the percent of total loads of E. coli and Salmonella recovered in runoff between tillage systems were observed. Total percentage of Salmonella recovered in runoff from both tillage systems was, however, four log 10 orders of magnitude greater than the percentage of E. coli that was recovered. Difference in percentage recovered between the fecal indicator bacterium, E. coli , and the pathogen, Salmonella , underscores an apparent difference in hydrologic transport characteristics of these two fecal bacteria and casts doubts on the efficacy of E. coli as an indicator of risk to public health.

Assessment of erosion impact on soil carbon and nitrogen loss and redistribution within landscapes is needed to develop estimates of soil carbon sequestration potential, soil quality management plans, and to evaluate potential for transport of sediment bound agrochemicals. We used variable intensity rainfall simulations to quantify the effects of tillage, conventional and strip, and antecedent soil water content on sediment-bound carbon and nitrogen loss from a Tifton loamy sand located in the southeastern Atlantic Coastal Plain (USA). Carbon and nitrogen loss via erosion of silt + clay sized versus sand sized sediment particles were quantified. Antecedent water content had no effect on mean sediment loss within tillage treatments, but losses from conventional till treatments were significantly greater than from strip till. Sediment lost as silt + clay was from 58 to 78% of the total under conventional and from 30 to 39% under strip tillage. The fraction of sediment lost as silt + clay versus sand was greater under conventional tillage than under strip tillage. Within-event sediment carbon enrichment compared to the top 2 cm of soil was 0.9–7.2 for conventional and 0.6–3.7 for strip tillage. The strip till silt + clay fraction had significantly higher carbon content than the strip till sand fraction and the conventional till silt + clay fraction. Carbon loss from treatments was directly proportional to sediment loss. However, the conventional till treatments lost 4.6–6 times more carbon from the silt + clay sized fraction and 1.9–4.8 times more carbon from the sand sized fraction than strip till treatments. Results suggest that the higher proportion of silt + clay fraction sediment loss from conventional till may deplete nitrogen enriched organic matter while decreased erosion from strip till may serve to increase retention of organic nitrogen. Findings also indicate that an approach that adjusts loss estimates of organic carbon and nitrogen by using wet-sieved subsamples for analysis followed by standardization against total bulk sediment loss accounts for introduced errors from both sub-sampling efficiency and disturbance.

A rainfall simulator was used to apply 5 cm of rainfall in 2 hours to two replicate 624 m2 plots at six times during each of the growing seasons of 1992 and 1993. Because the simulator generated reproducible and time-invariant rainfall intensities, the resulting 24 hydrographs reproducibly reveal the effects of tractor wheel compaction, tillage, soil reconsolidation, surface sealing, and corn canopy development. A time series data set including weather, crop development, soils properties, evapotranspiration, and antecedent soil water is available. These data should provide hydrologie modelers, particularly those interested in modeling runoff with time resolutions of

Laboratory studies provide the best opportunity for studying water and pesticide partitioning for a range of soil conditions while maintaining reproducible conditions. We developed and evaluated a laboratory technique for determining pesticide partitioning among infiltration, runoff, and sediment. Simulated rainfall (44 mm h -1 ) was applied to a sandy clay loam 14 days after pesticide (atrazine, chlorpyrifos, and 2,4-D) application. The laboratory technique provided controlled test conditions and reproducible results on the partitioning of rainfall, runoff, and pesticides at the soil surface and on sediment and pesticide transport. Average atrazine and chlorpyrifos half-life (t 1/2 ) values determined from corresponding concentrations in the 0- to 1-cm soil layer were approximately 30 and 10 d. Atrazine, chlorpyrifos, and 2,4-D concentrations remaining in the 0- to 1-cm surface layer decreased logarithmically with rainfall duration. Atrazine and 2,4-D concentrations in the 0- to 1-cm layer were correlated positively with those associated with splash (S), runoff (R), and sediment yield (E) (R 2 = 0.69 to 0.99). Relationships for all pesticide concentrations in the 0- to 1-cm layer and infiltration (INF) were correlated negatively. Relatively poor correlations (R 2 = 0.10-0.63) were found between chlorpyrifos losses and corresponding concentrations in the surface layer even though positive correlations were found between R and E (R 2 = 0.91) and E and S, (R 2 = 0.77). Measured effective partitioning coefficients (K deff ) increased gradually for each event, and were at least 31 times greater than reported K d values for atrazine, chlorpyrifos, and 2,4-D. The methodologies and experimental design described provide a means of evaluating how changes in pesticide concentrations in the surface layer are influenced by soil, rainfall, and pesticide characteristics, which will enhance the development of process-based models to predict pesticide fate and transport for event-based storms.

Many pesticide fate and transport models, including the GLEAMS model, overestimate pesticide degradation during its later stages of residence in soil. Except for runoff events shortly after pesticide application, models using equilibrium sorption kinetics often underestimate sediment-transported pesticides. To address concerns about transport of low levels of pesticides by runoff to sensitive ecosystems over annual or longer time cycles, GLEAMS was modified using a two-compartment pesticide-state model and algorithms for adjustment of degradation rates for temperature and soil water contents. Two pesticide pools, labile and nonlabile, were linked using first-order kinetics, with the forward and reverse rate constants between these pools as the only two additional inputs required for the model, GLEAMS-TC. GLEAMS-TC was calibrated successfully using 3 years of field data. Long-term pesticide persistence in soil was represented. Comparing model predictions with observed data, GLEAMS-TC simulated observed pesticide sediment transport, whereas GLEAMS underestimated observed data by a factor of 4. Sensitivity to rates of pesticide exchange between the two pools and impact on runofflosses were demonstrated with 50-year simulations. GLEAMS-TC is a research tool for investigating changes in pesticide state in soils as a function of exposure time, environmental variables, and interrelationships between pesticide degradation, mobility, and fate and transport in the environment.

Fenamiphos, a nematicide used on corn and sorghum, quickly oxidizes into two metabolites which have similar activities and toxicities, yet are more mobile and persistent than the parent compound. Given the soil and climatic conditions of the southeastern U.S., fenamiphos and its metabolites could be transported from the application site and contaminate off-site water bodies. A three-year study was conducted to evaluate (1) degradation and transport of the fenamiphos parent (Fp) and its metabolites (sulfoxide, Fx, and sulfone, Fo) from a 0.34 ha field site, and (2) the utility of the GLEAMS (Groundwater Loading Effects of Agricultural Management Systems) model in describing system response and simulating pesticide transport. Each year, fenamiphos was applied at 6.7 kg ha–1 a.i., broadcast and incorporated into the upper 100 mm soil layer before planting each crop. Concentrations of fenamiphos and its metabolites were determined from soil samples taken within the root zone at 50 mm intervals to a depth of 300 mm and from subsurface tile outflow at selected times throughout each sweet corn (Zea mays L.) and hybrid pearl millet (Pennisetum glaucum (L.) R. Br.) growing season. The GLEAMS model was used to simulate runoff, lateral subsurface flow (LSF), and Fp, Fx, and Fo losses from the Cowarts loamy sand. An average of 6 and 21% of the total rainfall + irrigation was measured as runoff and LSF, respectively. GLEAMS model simulations were correlated with measured runoff (R2 = 0.81) and LSF (R2 = 0.89). Field half-lives (t1/2) were determined by comparing observed concentrations in soil by depth and time to those simulated with the GLEAMS model. Average t1/2 values from measured field data were 5, 28, and 14 days for Fp, Fx, and Fo, respectively. For the three-year study, about 6.2% of the total amount of applied fenamiphos (Ftot = Fp + Fx + Fo) was measured in LSF, while less than 0.1% of the applied fenamiphos was measured in surface runoff. Fx was the dominant compound measured and simulated in the root zone and LSF, with 70 to 99% of measured Ftot being Fx. Calibration of the GLEAMS model provided fit of the field data that indicated (1) Fp dissipated rapidly while the two metabolites (Fx and Fo) formed (average Fp t1/2 = 5.5 d); (2) t1/2 values for all compounds remained relatively constant during 1987 and 1988, then numerically decreased in 1989; (3) coefficient of transformation (CT) values for Fx and Fo decreased from 1987 to 1989; and (4) CT values describing transformational changes from Fp to Fx were greater than those describing transformational changes from Fx to Fo. Decreases in t1/2 and CT values for Fp, Fx, and Fo with continued use over the three-year study is characteristic of enhanced microbial degradation.

There is need to evaluate the effects of agricultural production activities on sedimentation, pesticide, and nutrient losses under controlled, simulated rainfall on plots large enough (generally exceeding 50 m2) to incorporate realistic slope lengths and dominant processes that control runoff and sediment yield from “field-size” areas. Therefore, a rainfall simulator system was developed for 600-m2 plots (mesoplots) to evaluate runoff (water and sediment) and agrichemical movement from fields for different tillage practices and chemical applications. The rainfall simulator applies water with irrigation sprinklers spaced 3 m apart on two irrigation laterals arranged 14.6 m apart along the plot length. Runoff and sediment were collected in a V-shaped trough and directed to a flume for measuring and sampling. Simulated rainfall of 25 mm/h with a median drop diameter of 1.52 mm had a coefficient of uniformity of 91 over the plot area. Water was applied and sediment and runoff collected from two mesoplots for six events over a corn-growing season.

Laboratory soil pans and rainfall simulators have been used to study fundamental erosion processes affecting interrill soil loss. Problems exist when laboratory results are extended to those determined under field conditions. Experimental methodology influences a soil's interrill erodibility and interrill erodibility ranking among soils. We compared interrill soil loss and erodibility data from two erosion pans with field data from identical soils. Three plot size-rainfall simulator methodologies were used: (1) 0.14-m 2 lab pan under a constant drop size (4.6 mm) rainfall simulator, (2)0.32-m 2 lab with soil border areas under an oscillating nozzle (2.3 mm median drop diameter) rainfall simulator, and (3) 1-m 2 field plots under the same rainfall simulator as used in Method 2. Soil loss was measured at 5-min intervals. Interrill erodibility (K i ) was calculated from two equations (E=K ii and E=K iq Iq) using measured soil loss (E), rainfall intensity (I), and flow discharge (q) values. The interrill earosion equation, methodology, time, and initial water content influenced calculated K i values. Soil loss and K i values from Method 1 did not correlate with and were greater tan corresponding values from Methods 2 and 3. Soil loss and K i values from Method 2 were correlated (r=0.56 to 0.79) with corresponding values from Method3. The K iq values decreased with time and were a function of soil properties related to soil detachment and sediment transport. The K ii values (i) increased with time, (ii) were primarily a function of soil properties related to soil detachment only, and (iii) did not account for infiltration and runoff differences among soils

Cotton (Gossypium hirsutum L.) production in the Southeast United States can be limited by periodic drought. Irrigation and furrow diking tillage may improve economic yield and water use efficiency of cotton. Timing of rainfall may interfere with the efficiency of irrigation. Field studies were conducted during 2001 to 2010 near Shellman, Georgia to examine four irrigation rates based on Irrigator Pro for Cotton consisting of 100, 66, 33, and 0%. The objectives were to determine the value of irrigation and the economic water use efficiency of irrigation among irrigation rates over years. In-season rainfall ranged from 270 to 760 mm and irrigation volume ranged from 110 to 455 mm. Total water applied (rainfall + irrigation) ranged between 560 and 870 mm. In all but 2003, irrigation improved yield by 247-645 kg lint ha -1 . Years with average or below average rainfall had incrementally higher yields as irrigation rate increased except for 2007. Water use efficiency for irrigation was highest for the 33% rate in 2001 and 2002 and higher in 2004 and 2007 for the 66% rate. In severe drought years, the 33% rate did not sufficiently relieve drought stress. Irrigation provided profit in all but one year of the study. Although 100% irrigation is not the most efficient irrigation level, it often provides the most economic return. Furrow diking improved yield and water use efficiency in 3 of 6 years tested.

Coastal Plain soils in Georgia are susceptible to runoff, sediment, and chemical losses from short duration-high intensity, runoff producing storms at critical times during the growing season. We quantified runoff and sediment losses from a Tifton loamy sand managed under conventional- (CT) and strip- (ST) tillage and planted to peanuts. Simulated rainfall was applied at planting, 30 days after planting, and after harvest during the peanut growing season with rainfall events comprised of variable intensity (Iv) patterns representative of each time or season (spring=IvSPR, summer=IvSUM, fall=IvFALL). Simulated rainfall was applied to 2x3m plots (n=3) for each treatment. Runoff and sediment were measured from each 6-m 2 plot. Runoff ranged from 9-22% of the rainfall applied for the three events. The most runoff occurred from CT-IvFALL plots; the least occurred from ST-IvSUM plots. Maximum runoff rates were 7-20% of the maximum intensity and occurred 3-8 min after maximum intensity peaks. Sediment yields ranged from 105-1420 kg ha -1 . The most sediment occurred from CT-IvSPR plots; the least occurred from ST-IvSUM plots. Runoff and sediment curves had similar shapes as their corresponding rainfall intensity pattern. As for tillage, CT plots had 38% more runoff and 2.7-fold more sediment than ST plots over the three events. The largest difference in runoff (2.4-fold) and sediment (3.8-fold) among CT and ST plots occurred in the fall (IvFALL). Results improve our understanding of when runoff, sediment, and chemical losses are highest at critical times during a peanut growing season, and show how ST is effective in limiting those losses.

For the Front Range region of Colorado, quantifying rainfall partitioning under current and/or proposed farming practices and changing precipitation patterns is the first step to understanding how to efficiently conserve water and soil resources to meet crop water demands. We hypothesize that processes controlling runoff, sediment, and agrichemical losses would be better understood and predicted if variable rainfall intensity patterns derived from natural rainfall were used in rainfall simulation studies to evaluate runoff and sediment losses from conventionally tilled, intensively cropped Alfisols in the north-central region of Colorado. Our objective was to quantify the effects of constant (Ic) and more realistic, observed variable (Iv) rainfall intensity patterns (most frequent, Ivf; extreme, Ive) on runoff (R) and sediment (E) losses from a Fort Collins sandy clay loam (Aridic Haplustalf) cropped to conventional-till corn (Zea mays L.). Five treatments were evaluated: three Ic treatments (20, 47, and 65 mm h-1) and two Iv treatments representing the most common and most extreme events occurring in the region. Field plots, each 2 m wide by 3 m long, were established on each treatment. Each 6 m2 plot received simulated rainfall (Ic or Iv) for 60 min. For Ic events, increasing Ic 2.3-fold (20.3 vs. 47.3 mm h-1) and 1.4-fold (47.3 vs. 65.0 mm h-1) increased runoff 5.9- and 2.4-fold; sediment 4.9- and 4.1-fold; Rmax 5.8- and 1.9-fold; and increased Emax 3.1- and 3.9-fold, respectively. For Ivf vs. Ic (20 mm h-1) events, runoff, Rmax, and sediment losses were similar. However, Emax values for Ivf events were 33% less than those for Ic events. Also, tRmax and tEmax for Ivf events were 40-min (3-fold less) and 33-min (2.2-fold less) earlier than those for Ic (20 mm h-1) events. For Ive vs. Ic (65 mm h-1) events, runoff losses were similar. However, Ive events had 1.7-fold more sediment, 28% higher Rmax, and 2.9-fold higher Emax losses than those for Ic (65 mm h-1) events. Also, tRmax and tEmax for Ive events were 32-min (2.1-fold less) and 28-min (2.1-fold less) earlier than those for Ic (65 mm h-1) events. For Ivf vs. Ive events, runoff, Rmax, sediment, and Emax values for Ive events were 11.7-, 16.1-, 36- and 48-fold greater than corresponding values for Ivf events, respectively. Surface sealing, expressed as dINF (maximum infiltration-minimum infiltration), increased 19-fold as Ic increased 3.3-fold. The Iv patterns had at least 52% greater dINF values than corresponding values for Ic patterns; dINF values for Ive events were 3.1-fold greater than those for Ivf events. Our results show the pronounced effect of rainfall intensity patterns, particularly variable intensity patterns, on detachment and transport processes controlling runoff and sediment losses. A more accurate measure of rainfall partitioning, soil detachment, and sediment transport was obtained when Iv patterns derived from natural rainfall were utilized and evaluated.

Rainfall intensity (I), antecedent water content, and slope steepness affect soil loss from interrill areas. Objectives of this field study were to (i) test the validity of the assumption that interrill soil loss is a function of I 2 , (ii) determine effect of antecedent water content on interrill erodibilities (K i ) for two slope steepnesses, (iii) evaluate the slope factor used in the water erosion prediction project (WEPP), and (iv) evaluate the product of I, flow discharge, and slope steepness as a possible interrill erosion model. Four soils, ranging in texture from sandy loam to clay, were exposed to sequences of simulated rainfall at three intensities (50, 70, and 100 mm h −1 ). Soil loss and runoff were measured from airdried and prewetted flat plots and from air-dried and prewetted ridged plots [...]

Nutrients from agricultural fields can contaminate runoff and groundwater and enhance eutrophication. Runoff, soil loss, soluble P (SP), and bioavailable P (BP) associated with runoff and soil loss were measured from six 2 h simulated rainfall events (1 in./h) to determine amount of P transported during a corn (Zea mays L.) growing season and to develop relationships that predict P amounts and forms leaving different size field plots. Two 60-sq-ft microplots and two 6500-sq-ft mesoplots (3% slope) were located on a Tifton loamy sand (fine-loamy, siliceous, thermic Plinthic Kandiudult). Phosphorus was broadcast at 31 Ib P/acre, then incorporated to a depth of 6 in. Corn (conventional tillage) was planted after fertilizer application. Six simulated rainfall events occurred 5 d before and 1, 14, 29, 49, and 108 d after fertilizer application [...]

The relationship between rainsplash and three measures of soil strength was investigated for surface soil seals using a laboratory rainfall simulator and erosion pans. A significantly high correlation was found between pocket penetrometer and torvane strengths ( R 2 = 0.49) , and a lower correlation between both pocket penetrometer ( R 2 = 0.24) and torvane ( R 2 = 0.24) strengths and the fall-cone strength was found. Rainsplash was highly correlated ( R 2 = 0.49) with rainfall kinetic energy (KE) divided by the fall-cone shear strength (FC) but was not related to penetrometer and torvane strengths. Addition of a silt content term to the KE/FC linear equation slightly decreased R2. Improvements in the prediction of rainsplash can be achieved by including terms defining the macro- and microtopography, such as slope steepness, surface roughness, and depressional storage.

Potential pesticide loss in soil percolate is influenced by pesticide persistence and sorption by soil constituents (organic matter). Pesticide persistence, expressed as half-life (ti/2), changes with soil depth as microbial activity and soil properties change. Little is known, however, how these changes influence potential pesticide transport out of the root zone. Objectives of this study were to investigate relative differences in potential pesticide losses from the root zone by percolation due to 1) different soil surface and subsurface textures and pesticide ti/2» and 2) interactions between pesticide ti/2 and timing of rainfall after pesticide application. The GLEAMS (Groundwater Loading Effects of Agricultural Management Systems) model and a 50-year historical rainfall record at Tifton, Georgia, were used to simulate pesticide losses by percolation from three soils ranging in surface texture from sand to sandy clay loam. Hypothetical pesticides had surface ti/2 of 5, 15, 30, and 60 d and a range of subsurface ti/2 (2.5-360 d), and were applied to continuous com {Zea maize, L.) at 2 kg ha~ ^ as surface spray at planting each year on 1 April. Simulated pesticide losses by percolation increased with increased surface and subsuiface t|/2» and decreased with increased KQC (adsorption constant based on soil organic matter) values. Potential pesticide leaching was greatest for Lakeland sand and least for Greenville sandy clay loam. Rainfall timing affected simulated pesticide loss by percolation, especially for nonpersistent pesticides. For short pesticide ti/2 (0-5 d), excessive rainfall events within 1 ti/2 were largely responsible for simulated pesticide loss by percolation. Results indicate that changes in pesticide t|/2 in surface and subsurface horizons of different soils influence potential pesticide leaching from the root zone, and models (i.e., GLEAMS) can be used to provide comparative analysis of soil-pesticide-climate interactions. For example, depending on soil type and pesticide K^^ and surface Xyp^ values, potential leaching losses increased two to seven times as subsurface tiy2 increased six times.

The effects of soil and water conservation practices on water quality are being simulated with hydrologic models. Input parameters for physically-based models are commonly taken from generalized property data. However, field-measured input parameters provide a higher level of assurance that model output uncertainty will be affected by model structure rather than by input parameter uncertainty. The objectives of this study were to calculate curve numbers for conventional and conservation strip tillage and to compare curve numbers for growing and dormant seasons. Precipitation - runoff data collected from 1999 - 2005 at a field study site in south Georgia were used to calculate curve numbers by an averaging method and by a log-normal method. For average antecedent moisture conditions, the curve numbers calculated by the average curve number method for conventional and strip tillages were 89 and 84, respectively, and by the log-normal method were 89 and 83, respectively. The curve numbers derived from the short-term field data are higher than standard table values, 81 and 75, respectively, which were developed from a long-term series of annual maximum runoff. Curve numbers for strip tillage were 81 and 86 for growing season and dormant season, respectively; however, there was no difference between growing and dormant season curve number, 89, for conventional tillage. Under wetter than normal soil conditions during the dormant season, the curve number for both conventional and strip tillage was 96. The results indicate that hydrologic modeling of Coastal Plain watersheds would be improved if curve number were varied seasonally.