A multi-scale assessment of the impacts of vegetated buffer strips on surface water quality in southern Illinois

Author

William Joseph Beck, Southern Illinois University CarbondaleFollow

Date of Award

5-1-2010

Degree Name

Master of Science

Department

Forestry

First Advisor

Schoonover, Jon

Abstract

Intensive agricultural practices have the potential to introduce significant loads of sediment, nutrients, and fecal contamination into surface waters. A two phase study was conducted to assess the impacts of vegetated buffer strips on surface water quality in southern Illinois, USA. Phase one entailed a plot scale study that examined the ability of various vegetated buffer strip species to reduce fecal coliform bacteria and E. coli in overland flow, while also assessing fecal bacteria persistence in associated buffer soils. In July 2008, giant cane (Arundinaria gigantea (Walt.) Muhl.), Kentucky Bluegrass (Poa pratensis), Orchardgrass (Dactylis glomerata L.), and control (i.e., non-vegetated) buffers were established in sloped, 1x10 m plots. During June 2009, the upslope end of each plot was flooded with cattle (Bos taurus) manure laden water with a mean E. coli concentration of 2,111 MPN 100 mL ^-1, and a mean fecal coliform concentration of 2,634 CFU 100 mL ^-1. Mean E. coli concentrations in overland flow exiting the giant cane, Kentucky Bluegrass, Orchardgrass, and non-vegetated control plots were 14,387, 16,689, 16,045, and 10,256 MPN 100 mL ^-1 respectively. Overland flow samples were significantly higher than the initial tank for both E. coli (p < 0.01) and fecal coliform levels (p < 0.01). Vegetative treatment had no overall significant effect on E. coli levels, however, vegetative treatment did have an overall significant effect on fecal coliform levels (p = 0.04). Mean overland flow fecal coliform levels were significantly lower in non-vegetated control plots. To assess E. coli persistence in plots, overland flow was collected from a 50.8 mm hr ^-1 simulated rainfall event 4 days following the initial manure slurry application, and from natural rainfall events 22, 46, 93, 97, and 111 days after the application. Over the course of the 6 precipitation events, mean E. coli levels in overland flow exiting the giant cane, Kentucky Bluegrass, Orchardgrass, and non-vegetated control plots were 14,310, 10,366, 6,144, and 5,538 MPN 100 mL ^-1 respectively. Vegetative treatment had no overall significant effect on E. coli persistence during this time period. The seemingly poor microbial attenuation performance of study VBS may have resulted from the presence of concentrated flow paths and high ambient levels of fecal indicator bacteria within plots. Phase two entailed a paired-watershed approach that compared the chemical and physical attenuation abilities of giant cane and forested vegetated buffer strips at the watershed scale. This phase of the study is currently in year two of a three year calibration period. Water samples within three small (< 100 ha) agricultural watersheds were collected during both stormflow and baseflow and analyzed for total suspended solids (TSS), turbidity, nitrate-N, dissolved reactive phosphorus (DRP), ammonium-N, pH, conductivity, and fecal indicator bacteria. Stormflow and baseflow event mean concentrations were used to perform inter and intra-watershed comparisons. Stormflow event mean concentrations for TSS, nitrate-N, and ammonium-N were significantly higher than baseflow concentrations for two of the three watersheds. Inter-watershed significant differences were exhibited exclusively during stormflow, and occurred only for TSS and DRP. Watershed scale studies that examine the water quality impacts of vegetated buffer strips establishment are rare, and the ability of giant cane buffers to attenuate fecal indicator bacteria in overland flow is an area that has received no research attention to date. Results from this research will provide managers with more information on a broader range of species to incorporate into mixed species vegetated buffer strip designs.