Rivers in watersheds dominated by agriculture throughout the US are impaired by excess sediment, a significant portion of which comes from non‐field, near‐channel sources. Both land‐use and climate have been implicated in altering river flows and thereby increasing stream‐channel erosion and sediment loading. In the wetland‐rich landscapes of the upper Mississippi basin, 20th century crop conversions have led to an intensification of artificial drainage, which is now a critical component of modern agriculture. At the same time, much of the region has experienced increased annual rainfall. Uncertainty in separating these drivers of streamflow fuels debate between agricultural and environmental interests on responsibility and solutions for excess riverine sediment. To disentangle the effects of climate and land‐use, we compared changes in precipitation, crop conversions, and extent of drained depressional area in 21 Minnesota watersheds over the past 70 years. Watersheds with large land‐use changes had increases in seasonal and annual water yields of >50% since 1940. On average, changes in precipitation and crop evapotranspiration explained less than one‐half of the increase, with the remainder highly correlated with artificial drainage and loss of depressional areas. Rivers with increased flow have experienced channel widening of 10–40% highlighting a source of sediment seldom addressed by agricultural best management practices. Copyright © 2013 John Wiley & Sons, Ltd.

For two watersheds in the northern Midwest United States, we show that landscape depressions have a significant impact on watershed hydrology and sediment yields and that the Soil and Water Assessment Tool (SWAT) has appropriate features to simulate these depressions. In our SWAT models of the Willow River in Wisconsin and the Sunrise River in Minnesota, we used Pond and Wetland features to capture runoff from about 40% of the area in each watershed. These depressions trapped considerable sediment, yet further reductions in sediment yield were required for calibration and achieved by reducing the Universal Soil Loss Equation (USLE) cropping-practice (P) factor to 0.40 to 0.45. We suggest terminology to describe annual sediment yields at different conceptual spatial scales and show how SWAT output can be partitioned to extract data at each of these scales. These scales range from plot-scale yields calculated with the USLE to watershed-scale yields measured at the outlet. Intermediate scales include field, upland, pre-riverine, and riverine scales, in descending order along the conceptual flow path from plot to outlet. Sediment delivery ratios, when defined as watershed-scale yields as a percentage of plot-scale yields, ranged from 1% for the Willow watershed (717 km) to 7% for the Sunrise watershed (991 km). Sediment delivery ratios calculated from published relations based on watershed area alone were about 5 to 6%, closer to pre-riverine-scale yields in our watersheds.

No abstract

No abstract