PREDICTING RUNOFF FROM A YARD WASTE WINDROW COMPOSTING PAD

Abstract

Composting is utilized increasingly as an aerobic treatment alternative to landfilling for bioconverting organicwastes such as municipal biosolids and yard wastes into soil amendments or landscaping materials. However, little informationis available to estimate composting operation runoff to corresponding design rainfall events. The objective of this workwas to evaluate dynamic versions of the simple . index and NRCS curve number runoff models currently used in agriculturalor urban stormwater management for predicting runoff from windrow composting operations. A 1.4 ha (approximately 48 300 m) windrow pad was constructed by placing 25 mm crusher-run aggregate on a tightly compacted Cecil sandy clay loam.The pad was sloped (<1% on the 300 m axis and 0% on the 48 m axis) and bermed to direct all runoff to a 0.2 ha 2 m deepcatchment pond with less than 3.175 mm/day seepage constructed with a regular trapezoidal shape, thus enabling astraightforward relationship between pond depth and volume. Precipitation, pond stage, and compost amount were monitoreddaily in accordance with environmental permit requirements. A dynamic NRCS curve number daily runoff model, in whichthe precipitation was adjusted with pad fraction covered and 7-day antecedent moisture, resulted in a validation R2 of 0.56and had a modeled vs. observed slope of 0.93 when applied to the validation data set. Applied to monthly data, the dynamiccurve number approach, which adjusted precipitation with average fraction covered with compost for the month, gave avalidation R2 of 0.75 and a validation slope of 0.76. Based on daily and monthly results, the dynamic NRCS-CN models withprecipitation adjustments yielded the best validation results. The more conventional dynamic NRCS-CN modeling approachwhere the curve number was modified performed nearly as well with daily and monthly runoff prediction. Theprecipitation-modified NRCS-CN equation tended to overpredict daily runoff at low precipitation rates, whereas theCN-modified version of the daily NRCS-CN equation was more consistent across the observed precipitation range. Bothdynamic NRCS-CN equations tended to overpredict monthly runoff.