HETEROGENEITY, MODELS, AND SITING OF MULTI-SPECIES RIPARIAN BUFFERS

Multi-species riparian buffers have been proposed to minimize the impacts of agricultural production and improve water quality in Iowa and the Midwest. Although broadly viewed as improving groundwater quality, the effect of buffers on nutrients in groundwater is discussed mainly for idealized hydrogeologic settings. Since, 1990, the hypothesis of improved groundwater quality via nitrate-N removal has been tested in the Bear Creek watershed, a 7,656 ha agricultural watershed in central Iowa. Multi-species riparian buffers, consisting of 20- to 40-m-wide strips of switchgrass, shrubs, and trees up to 15 years old, have been installed in sections of the watershed. Three sites – the Risdal North and South sites and the Strum site representing a chronosequence of buffers – were chosen because of their apparent geological similarities. All were constructed on top of about 4 m of loamy and coarse-grained (sand and gravel) alluvium of the Holocene DeForest Formation. An aquitard composed of late Wisconsinan till and shale/limestone underlies the alluvium and overlies the Mississippian aquifer, such that groundwater is effectively channeled through a 2-m zone beneath the buffer on the way to Bear Creek. These would appear to be ideal hydrogeologic settings to evaluate buffers; however, buffer effectiveness differs among the sites due differences in the age and composition of sediments. This, in turn, causes differences in groundwater flow paths, hydraulic gradients, velocities, and residence times at these and nearly every subsequent site investigated in the watershed. In the context of planned watershed-wide buffer implementation, heterogeneity may be less important to quantify than regional groundwater flow. For example, an analytic element, groundwater flow model suggests that many channelized stream segments of the creek are not connected to the water table. At other sites, the regional model can be used to define boundary conditions and parameters for smaller scale “generic” models. In place of knowing the exact distribution of sediments everywhere, the “generic” models can be used to simulate “idealized” heterogeneous buffers and predict their effect on groundwater quality.