Late 20th-century trends in Iowa watersheds: an investigation of observed and modelled hydrologic storages and fluxes in heavily managed landscapes

• Published in: International journal of climatology 2012-07, Vol.32 (9), p.1373-1391
• Main Authors: Steffens, Kayla J., Franz, K. J.
• Format: Article
• Language: English
• Subjects: climate change; Earth, ocean, space; Exact sciences and technology; External geophysics; hydrologic modelling; hydrologic trends; land use; Meteorology; Midwest US
• ISSN: 0899-8418; 1097-0088
• DOI: 10.1002/joc.2361

Many climate change studies have shown that streamflow has been changing across the US over the last century. However, there are fewer studies examining trends in hydrologic variables such as soil moisture and snow, which tend to be poorly observed in some regions. Hydrologic models can be applied to understand changes in processes with limited spatial or temporal observational records. However, the application of hydrologic models to long time periods can be complicated by human influences on the hydrologic cycle. In this study, we explore hydrologic changes and the applicability of a commonly used hydrologic modelling system for climate change analysis in a heavily managed landscape. Historical streamflow observations and model‐derived time series of subsurface storage, frozen ground, and snow cover were analysed for 10 Iowa watersheds. The modelled data were generated using precipitation and temperature data spanning 1948–2003 as inputs to common hydrologic prediction models. Time series were tested for trends and significance using the Mann‐Kendall (MK) test with the Trend‐Free Pre‐Whitening procedure at a p = 0.1 significance level. Plots of 10‐year running averages were also examined. Significantly increasing trends in mean daily discharge and the number of high‐flow days were found for several watersheds. Increasing monthly subsurface water content was observed throughout the year in most basins, and was greatest during the winter. The number of modelled frozen‐ground days declined at all sites. Three different model calibration periods were tested and found to reproduce the general upward trend observed in the daily mean discharge. However, the model results varied from the observed in both magnitude and the number of sites identified to have MK trends. Additional functionality that considers land management and land cover change may make the models applied here more accurate for studying climate change impacts in regions such as Iowa. Copyright © 2011 Royal Meteorological Society