Application of the USGS soil-water-balance (SWB) model to estimate spatial and temporal aspects of groundwater recharge in north-central Iowa

Groundwater recharge (hereafter referred to as recharge ) is the process whereby water infiltrates at the land surface, flows through the unsaturated zone, and crosses the water table to enter the groundwater system. Because previous estimates of recharge have come from small scale studies, there is a need for a method or model that estimates recharge at larger scales. This study used the U.S. Geological Survey (USGS), Soil Water Balance (SWB) model to estimate the mean annual potential recharge (MAPR) in an area encompassing Boone, Story, Hamilton and part of Hardin and Webster counties in Iowa during the 20-year period from 1996 to 2015. The SWB model domain includes an area of 2,301 mi2 (5,960 km2 or 596,000 ha), a 328 ft (100m) grid spacing of 771 rows (47.9 mi or 77.1 km wide) by 773 columns (48.0 mi or 77.3 km wide) and contains 595,983 cells. It successfully estimated MAPR for the 20-yr period with a value of 6.3 in/yr (159 mm/yr), which is consistent with values from groundwater modeling studies in the region. On an annual basis, the lowest annual mean potential recharge (AMPR) occurred in 2000 with 0.3 inches (6.9 mm) and the highest was in 2008 with 14.3 inches (370 mm). There is a strong relationship between the amount of precipitation and AMPR with an R2 = 0.86 (p= 0.01). MAPR varies among land use/land cover types. For cultivated crops, the MAPR was 6.3 in/yr (160 mm/yr), ranging zero to 20.3 in/yr (515 mm/yr) during the period. In contrast, deciduous forest areas averaged 7.4 in/yr (188 mm/yr) and ranged from zero to 21.5 in/yr (546 mm/yr) – the maximum MAPR for all LULC types. Areas with Hydrologic Soil Group A showed the highest MAPR at 7.5 in/yr (191 mm/yr). Surficial geologic units were found to have differences in MAPR that were significantly different. Notable statistically significant differences at the p=0.05 level are those between units classified as till plain (Qtp) with MAPR of 6.3 in/yr (160 mm/yr), and till ridge (Qtr) with a MAPR of 6.5 in/yr (165 mm/yr). Differences were also found between modern floodplain (Qal) with MAPR of 6.3 in/yr (160 mm/yr) and alluvium found on terraces (Qalt) with MAPR of 6.1 in/yr (155 mm/yr). Glacial lake plain sediment (Qglp) showed the lowest MAPR of 5.6 in/yr (142 mm/yr). These differences suggest that MAPR values could be assigned to these units in a groundwater model. AMPR results from the SWB model were compared to baseflow values as a method to “calibrate” AMPR estimates, using the assumption that baseflow represents all of the recharge falling within a groundwater basin. Three baseflow-separation methods, PART, HYSEP-sliding-interval method, and HYSEP-fixed-interval methods, were used. Manual adjust of runoff curve numbers by ± 5 and 10% for the row crop LULU class, which covered 83.4 % of the model area, was performed, and the curve runoff numbers with the best overall fit to baseflow as evaluated by R2 were used. For the calibrated model, the overall comparison using annual baseflow estimates from all four watersheds, showed strong correlation with R2 = 0.78 for PART, 0.82 for HYSEP-sliding interval method, and 0.82 for HYSEP-fixed interval method, respectively. The best correlation was the South Fork Iowa River at New Providence (USGS gage 05451210) with an overall mean R2 =0.83 (significant at p=0.01). Values of R2 over all the watersheds ranged from 0.71 to 0.85 (significant at p=0.01). These results suggest that the AMPR values are reasonable and essentially “calibrated” to an independent estimate of baseflow. Sensitivity analysis was also performed by manually varying runoff curve numbers by ± 5 and 10% for the row crop LULU class. The results show that row crops – the largest land use type in the study area – are very sensitive to the value of the runoff curve number and may have a large impact on model results. Recharge estimates from SWB in this study are not without limitations. Future work should incorporate flow direction into the SWB model to provide a more accurate water budget for this study area. The model could also be used to project MAPR values under a warmer and wetter future climate in Iowa and the Midwest