This blog post discusses the seminar given by Don Weller of the Smithsonian Environmental Research Center (SERC), at the IAN Seminar Series on October 28, 2010.

The variable proportion of nitrate removed by riparian buffers in the different physiographic regions was noted in the discussion. Also noted was the relative constancy of the absolute amount of nitrate removed by riparian buffers (~10 mg/L) in all physiographic regions. The question of what mechanism(s) led to the variable proportion and constant absolute amount of nitrate removal was asked. Don indicated that a hypothesis was that the hydrological flow paths were different in the different physiographic regions. He indicated that the values were averages of many different watershed sites, and nitrate removal efficiencies were variable within each physiographic region. The fractured rock and especially karst (porous calcium carbonate rocks) regions in the piedmont and Appalachian mountains were also factors in 'short-circuiting' flow paths that could account for less effective nitrate removal.

Another factor in nitrate removal efficiency discussed was the importance of saturated soils near the stream. Saturated soils often have anaerobic/aerobic interfaces and available carbon substrates to promote denitrification, the conversion of nitrate to nitrogen gas. Don pointed out that there were several potential mechanisms of nitrate removal in riparian buffers including denitrification, conversion into biomass and recalcitrant nitrogen forms as soil storage.

In terms of management applications, the 'eye-popping' efficiencies of nitrate removal by coastal plain buffers lead to a strong recommendation to insure complete riparian coverage in this region. In addition to high efficiencies, the proximity to Chesapeake Bay make the coastal plain streams that much more important for targeting for riparian vegetation. Don did say that there are more extensive gaps in riparian buffers in the piedmont region, thus there were more opportunities for implementation of riparian restoration projects in this region.

The issue of the using riparian buffers in concert with other best management practices, in particular, cover crops was discussed. Since riparian buffers primarily intercept nitrate in overland flows (and shallow groundwater) and cover crops primarily capture nitrate that would enter groundwater, the combination of these two best management practices would be particularly powerful.

The focus of Don Weller's research was nitrogen and he used nitrate as a proxy for total nitrogen. Don noted that in disturbed watersheds (e.g., watersheds with fertilized crops) that nitrate was good proxy for total nitrogen, but in forested watersheds, various other forms of nitrogen would need to be monitored. Phosphorus uptake by riparian buffers was also discussed and Don felt that any sediment interception by forested buffers would likely reduce phosphorus loading to streams, but this was not assessed in his research.

The issue of localized differences in nitrate removal efficiencies within each of the physiographic regions was discussed. Being able to develop a landscape scale assessment and compare, for instance, eastern shore coastal plain riparian efficiencies with western shore coastal plain efficiencies would be of interest. Don provided an important caveat with extrapolating data to the entire Chesapeake watershed in the sampling strategy did NOT obtain representative random sampling of the entire watershed. While there were a substantial number of replicate watersheds studied, there were several factors that could influence the basin wide extrapolations. The age or health of the buffer was not assessed, nor were buffers less than about 30 m wide (due to resolution of surface topography and forest cover data used to calculate flow paths) considered. Higher spatial resolution data are needed, especially to begin to address the question of optimal buffer width for nutrient removal.

The application of information theory to compare different models was a rather novel approach for watershed data. Don mentioned that the information theory allows for more effective comparisons between different models than simple regression comparisons. The fact that forested buffers and crop lands accounted for 75% of the variability of the data was elucidated by this approach.