Liquid Water and Vapor Flow in Arid Soil: Comparison of Weighing Lysimeter Data with Simulations from a Process-Based Model

Abstract:

Desert soils account for about a third of the Earth’s land surface and are believed to be important players in terrestrial energy balance. However, the mechanisms that govern energy and mass fluxes across the land-atmosphere interface of hot deserts remain poorly understood. This knowledge gap also spills over to our insufficient understanding of the ecology and hydrology of deserts. A recently constructed weighing lysimeter (3 m deep and 2.26 m in diameter) located in Boulder City, NV, provides data of water and energy fluxes across the soil-atmosphere boundary of the Mojave Desert. The lysimeter has been filled with homogenized desert soil from nearby Eldorado Valley, instrumented with a suite of more than 150 sensors at multiple depth between 2.5 and 250 cm and under continuous operation since July 2008. In this study, we report on water content, water potential, and temperature data from one hydrologic year at high spatial and temporal resolutions. The data was used to develop, calibrate and validate a coupled, process-based water flow and storage model using Hydrus-1D. The model simulates liquid water flow, heat flow, and non-isothermal vapor flow along the soil profile. Detailed soil bulk density and porosity profiles are known based on soil mass and volume determined during lysimeter soil installation. Water retention property was determined from concurrent volumetric water content and matric potential measurements. A density-dependent scaling relation was developed to adjust water retention properties to the different soil bulk densities in the profile. The water flux across the soil-atmosphere boundary was determined from high-resolution lysimeter scale data. The saturated hydraulic conductivity was estimated via inverse modeling, using a subset of the soil moisture data. The calibrated model was validated using the remainder of the data set. The model accurately captures the soil temperature dynamics through the year and across the profile. The water content dynamics was also captured very well, except when the water potential of the near surface soil dropped below -100 MPa. This suggests that the bundle-of-capillaries based coupling of water retention and hydraulic conductivity breaks down at such low matric potentials, where vapor flow is likely the dominant water flow mechanism.