Print PageHeadwater streams provide important ecosystem services such as aquatic habitat, organic matter input, and nutrient processing. Despite their value, most headwater streams are not included in the national hydrography dataset (NHD), either because they were buried during early development or because they were not detected at the minimum mapping size. This omission leaves us with a considerable underestimation of total stream length and landscape connectivity. Here we present a systematic technique of mapping the natural potential location of channel heads (i.e., where streams begin) across broad and diverse areas. We first mapped over 250 channel heads in the field across five Omernik Level III ecoregions. We then compared these locations to a suite of topographic and soil characteristics provided by high-resolution geospatial data. Using logistic stepwise regression, we created a natural potential headwater map of the entire Potomac River Basin that takes into account landscape variability among different ecoregions.
Results/Conclusions
We found that explanatory variables for modeling the location of channel heads vary among ecoregions, with particular topographic controls (e.g., local slope) being more important for some ecoregions and particular soil characteristics (e.g., silt-clay percentage) being more important for others. The accuracy of our model also varied among ecoregions. The model correctly predicted the presence of 86% of all 10m stream segments in the Piedmont ecoregion, but only 64% of stream segments in the Coastal Plain ecoregion. By knowing the natural potential location of headwater channels, we have a better approximation of the ecosystem services that have been lost due to stream burial and development.
