Phosphorus Reduction in Runoff Using a Steel Slag Trench Filter System

Abstract

Excessive concentrations of phosphorus (P) can lead to the deterioration of surface waters by eutrophication. The objective of this study was to evaluate the performance of a steel slag trench filter system developed for P reduction in runoff under field conditions. The experimental design for the study was a 2 x 2 factorial with factors of filter material (6 mm steel slag or 13 mm washed river gravel) and triple superphosphate fertilizer (P applied or no P applied) replicated within irrigation zones covered with `Astro' bermudagrass [Cynodon dactylon L. (Pers.)] mowed at 38 mm. Each irrigation block consisted of four plots that were 6.1 m wide with a uniform 5% slope that measured 24.4 m long. Four plastic containers (volume = 178 L each) were placed in a 5.2 m 1.2 m trench dug in the middle of each plot perpendicular to the slope to accommodate filter materials. Runoff caused by either natural rainfall or irrigation was collected for comparison of P concentrations before and after filtration. Flow rates were determined using ultrasonic depth detection devices as runoff flowed through calibrated Parshall flumes. From March through September 2012 a total of 14 runoff events were studied. Runoff filtered by steel slag contained from 17% to 43% less dissolved P than runoff filtered through an inert gravel control. The P concentrations in runoff from the fertilized treatment were always greater than the unfertilized treatment. The difference between P concentrations from the fertilized and unfertilized treatments declined with each runoff event from 9.48 mg L-1 in the initial event after fertilization to 0.20 mg L-1, 120 d and nine runoff events later. A previous P removal model developed by Penn and McGrath (2011) overestimated the maximum P loading potential of the slag filter structures (460 mg kg-1 vs. 269 mg kg-1 calculated from actual data), as well as the maximum P removal before the filters were spent (35.7 mg kg-1 vs 8.3 mg kg-1). The overestimation of the performance of the P removal structure probably occurred because the slag used in this study had "inferior" P sorption qualities than that used for model development.