Percolation observations in an arid piedmont watershed and linkages to historical conditions in the Chihuahuan Desert

Abstract

Abstract. A critical hydrologic process in arid and semiarid regions is the interaction between ephemeral channels and groundwater aquifers. Generally, it has been found that ephemeral channels contribute to groundwater recharge when streamflow infiltrates into the sandy bottoms of channels. This process has traditionally been studied in channels that drain large areas (tens to hundreds of square kilometers). Since the water table in arid and semiarid regions is typically far from the surface, measured streamflow losses or percolation into the deep vadose zone is equated to groundwater recharge. In this study, we use a water balance approach to estimate deep percolation in a first-order, instrumented watershed (4.7 ha) on a piedmont slope of the Jornada Experimental Range (JER) in the Chihuahuan Desert. Results indicate that runoff generated within the piedmont slope contributes significantly to deep percolation. During the short-term 6-yr study period, we estimated 385 mm of total percolation, 62 mm/yr, or a ratio of percolation to rainfall of 0.26. Based on the instrument network, we identified that percolation occurs inside channel areas when these receive overland sheetflow from hillslopes. We observed less streamflow leaving the watershed as compared to percolation during the study period, leading to an outlet streamflow to rainfall ratio of 0.02. Using long-term data sets available from the JER, we estimate that over the last 100 yr, 48 mm/yr of percolation occurs at the study site, a ratio of percolation to rainfall of 0.19. We then scale this up to determine the contribution of similarly sized watersheds on the piedmont slopes of the JER. These observations highlight the importance of arid piedmont slopes for generating groundwater recharge, in particular during above-average rainfall periods.

On the Sensitivity of Hillslope Runoff and Channel Transmission Losses in Arid Piedmont Slopes

“…In contrast, channel sediments are sandy in texture with the presence of large cobbles at depth. The main channel is about 400 m in length, ranges from 0.5 to 1 m in width and has several small tributaries (see Schreiner‐McGraw & Vivoni, , for subwatershed properties). Channel width ( W c in meter) estimates at varying drainage areas ( A d in square kilometer) were used to obtain a relation W c = 3.59 A d 0.35 applicable to the channel network.…”

“…and an open‐path infrared gas analyzer (LI‐7500, LI‐COR) installed at 7‐m height. Flux corrections followed procedures described in Anderson and Vivoni (), and energy balance closure was forced using the Bowen ratio method of Twine et al () in order to match the approach of Schreiner‐McGraw and Vivoni () to estimate percolation and the assumption of energy balance closure in the model physics (Ivanov et al, ). As reported by Schreiner‐McGraw et al () and expanded to the full time period for this study, energy balance closure of 83% was obtained.…”

“…Independent, split‐sample model testing was performed over 1 January 2014 to 30 September 2016. By calibrating and validating the model with the full set of observations simultaneously, we ensure a consistent representation of the simulated watershed response with the objective of reproducing the main features of the rainfall and runoff response described by Schreiner‐McGraw and Vivoni (). Model performance was assessed using the mean absolute error (MAE), correlation coefficient (CC), and bias (B) as goodness of fit metrics for the aforementioned observations at daily and monthly resolutions, as described in Pierini et al, (, their Appendix A).…”

“…Tooth, ). Few studies have been aimed at understanding the relative roles of hillslope and channel processes on watershed‐scale runoff generation and their implication on transmission losses (or percolation, Schreiner‐McGraw & Vivoni, ) in dryland systems. For instance, Puigdefabregas et al () found that semiarid hillslopes in Spain mainly contributed to runoff through the infiltration‐excess mechanism occurring during intense but short storms and had a limited role in the recharge generated during longer‐lasting storms in downstream channel areas.…”

“…Nevertheless, recent observations in an arid piedmont watershed indicate that infiltration losses from hillslope surfaces and transmission losses in first‐order channels account for a large fraction of the water balance (Schreiner‐McGraw & Vivoni, ). Monthly losses from the small watershed (4.7 ha) were estimated using coordinated measurements of precipitation, streamflow, ET, and changes in soil water storage over a 6‐year period (2010–2016).…”