Seasonal hydrologic trends in Iowa watersheds over the last half of the 20th century

Climate change is expected to cause significant changes in the global hydrologic cycle, yet the impact to regional hydrology is not well known for the Midwest U.S. As a first step in understanding the impact of climate change in Iowa watersheds, historical streamflow observations and model-derived time series of soil moisture, frozen ground and snow cover for ten Iowa watersheds were analyzed. The modeled data were generated using precipitation and temperature data spanning 1948 to 2003 as inputs to the conceptually-based hydrologic models of the National Weather Service river forecasting system. The models were calibrated for each watershed using observed discharge data. The time series were tested for trend significance using the Mann-Kendall test with the Trend-Free Pre-Whitening procedure at a p = 0.1 significance level. Results show an increasing trend in mean daily discharge, peak flow from rain, and low flows over the last 50 years. Monthly soil moisture content is also increasing, and is strongest during the warm seasons. Maximum daily flow from snow displays a decreasing trend and tendency to occur earlier in the year suggesting an earlier melt of the snowpack in the region. Results for frozen ground and snow cover show that the onset and conclusion has shifted to earlier in the year with more melt days occurring over the snow season. Although several of the hydrologic variables examined did not show statistical significance, trends in most hydrologic processes were observed. Different model calibration periods were tested and found to have minor influence on the average simulation accuracy, but did impact the simulated trends in streamflow. Model results from three calibrations indicate that the modeling system responds to changes in climate, but other factors that the model cannot account for (e.g. land-use change) may be reflected in the observed discharge from several basins