Geospatial methods for aquatic conservation: Topeka shiner restoration site selection and the development of an Iowa watershed health assessment
Kevin J. Roe
The conversion of native ecosystems to agricultural landscapes often has profound impacts not only on the terrestrial environment, but also on aquatic ecosystems. Loss of perennial riparian vegetation, intensive installation of tile drainage, expansion of drainage ditches, and channelization of streams lead to the loss of lateral stream connectivity to the floodplain. In Iowa and Minnesota, restoration of off-channel habitats, also called oxbows, by the U.S. Fish and Wildlife Service (USFWS) are underway to both restore this connectivity and provide habitat for the conservation of the endangered Topeka Shiner (Notropis topeka). As oxbow restorations continue and expand to other areas of the Topeka Shiner’s range, there is a need for additional information to support site selection and prioritization.
I developed three studies to meet this need by 1) creating a repeatable, systematic methodology to identify former stream meanders and oxbow scars on the landscape in order to identify candidate restoration sites, 2) developing and applying species distribution models to rank potential restoration sites by their landscape level suitability for Topeka Shiners, and 3) synthesizing watershed health data to evaluate threats that may affect future restored sites. The models created to identify former stream meanders and oxbow scars as potential restoration sties were highly successful at discriminating target from non-target features on the landscape using Light Detection and Ranging (LiDAR) satellite imagery. The shape and relative height above the stream channel of riparian depressions were critical characteristics in identifying target features. I compensated for noise (misclassification of non-target features as target features) within models by developing a ranking system to help extract sites that were most likely to be stream meanders and oxbow scars. Next, species distribution models demonstrated strong predictive power to identify both stream and oxbow sites that would be suitable for Topeka Shiners. Last, the generation of the watershed health assessment ranked ten-digit hydrological unit code (HUC 10) watersheds within Iowa using five overarching components of geomorphology, hydrology, biology, connectivity, and water quality to visualize the relative threats to each watershed. Though the last study did not specifically address Topeka Shiner restoration site selection, it is applicable to many future conservation efforts as it allows managers to identify factors that may impact watershed restoration projects. Combined, these studies demonstrated diverse uses of geospatial data to support aquatic conservation and restoration efforts in highly modified landscapes.