Identifying critical source areas for best management practice targeting in impaired Zumbro River watersheds using digital terrain analysis

The Zumbro River Watershed drains 1,421 sq. mi. (3,680 km2, 910,337 ac.) of land in southeast Minnesota. Sedimentation within the watershed, particularly in Lake Zumbro and Lake Shady, has raised concerns over sediment-laden runoff entering waterways. Floods in 2010 filled the 190-acre Lake Shady with sediment, necessitating costly removal of its dam, and the 600-acre Lake Zumbro reservoir's rising sediment levels has resulted in a planned $7 million dredging project. These issues have triggered public awareness campaigns in the watershed, including the "slow the flow" educational initiative designed to engage residents within the watershed to slow and reduce the amount of water running into the Zumbro River. Focus has also shifted upstream in order to reduce much of the sedimentation at the source – namely agricultural runoff – by pushing for conservation practice implementation. Currently, conservation practices in the watershed are implemented opportunistically, because a coordinated, watershed-wide approach for identifying critical sources of nonpoint source pollution, prioritizing sites, and planning implementation projects is absent. Critical source areas (CSAs) are small locations on a landscape that contribute a disproportionate amount of runoff to surface waters. Targeting CSAs can therefore give the best "bang for the buck" when optimizing best management practice cost/benefit ratios. The Zumbro watershed was therefore a prime candidate for CSA identification using a simple toolset that could be adopted by various agencies and conservationists throughout the state. Digital terrain analysis (DTA) – specifically the stream power index (SPI) – was chosen as the method to help locate CSAs based on its ease of application, simplicity, and documented success in similar studies. Three areas – each representing the main agroecoregions of the Zumbro River Watershed – were used to field validate the terrain analysis. Field accuracies associated with positively identifying surface erosional features using DTA methods ranged from 77-88%. DTA was 100% accurate when identifying features with the highest sediment delivery potentials (SDPs) for all three study areas.