What can the GRACE seasonal cycle tell us about lake-aquifer interactions?

Abstract

Lake-aquifer interactions have been the subject of investigation and debate for decades. Traditional investigation methods include measurement of water flux across the groundwater–surface water interface, application of heat and environmental tracer methods, conducting numerical simulations of the water flow, and mass balance–based approaches. We first review, evaluate and describe the different approaches that have been applied for examining lake-aquifer interactions and propose an additional complementary approach. While the leakage from lakes and its slow movement through porous media (laminar flow) is well characterized by existing methods, modeling the rapid and turbulent flow through preferred pathways (faults/karst) remains a challenging task. To better understand the nature, and the full scale, of the lake-aquifer interactions, and using Lake Nasser and the underlying Dakhla subbasin of the Nubian Sandstone Aquifer System (NSAS) in the Western Desert of Egypt (area: 0.66 × 106 km2) as a test site, we examine, model, and correlate temporally and spatially, the variations in Gravity Recovery and Climate Experiment terrestrial water storage (GRACETWS) solutions, with precipitation, lake levels, area, and water volume. We review current GRACE applications in hydrology and present our novel approach. Findings include: firstly, large seasonal fluctuations (peak: Nov./Dec.; trough: July/Aug.) in Lake Nasser's surface water levels (average annual fluctuations: from 3.7 m to 7.5 m), area (average area: 3622 km2 but up to 4530 km2), and volume (annual average: 13.4 km3, and up to 34 km3) are observed throughout years 2006 to 2015. These fluctuations are accompanied by an increase in GRACETWS (average: 50 ± 13 mm/yr, up to 77 ± 18 mm/yr) over Lake Nasser and by a progression of a front of increasing GRACETWS values (>50 ± 13 mm) away from the lake reaching distances of up to 700 km some 3 to 5 months following peak lake level periods. The areas witnessing the seasonal increase in GRACETWS display a progressive increase in phase and decrease in amplitude with distance from the lake. Secondly, the negligible precipitation over the Dakhla subbasin cannot account for the observed seasonal GRACETWS patterns and neither can the leakage signal from Lake Nasser. Thirdly, overall similarities in the distribution of GRACETWS seasonal spatial patterns are observed. During high lake level periods (e.g., Dec. 2007; Feb. 2015) additional, and more distant, areas from Lake Nasser saw increases in GRACETWS and vice versa during low lake level periods (e.g., July 2006; Aug. 2010). These observations are consistent with Lake Nasser being the main source of modern recharge for the Dakhla subbasin and suggest a new conceptual model for the subbasin: a slow groundwater flow through a porous matrix and a rapid groundwater flow along a network of faults, fractures, and karst topography across the matrix. We suggest that the proposed conceptual model for the interaction between Lake Nasser and the Dakhla subbasin could be applicable to aquifers of similar geologic, climatic, and hydrologic settings worldwide and that approaches similar to those advocated here could be used to investigate the validity of this suggestion.