Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Israel → Integrating GIS and numerical solution of flow equations to evaluate the impact of soil sealing on spatial and temporal soil moisture dynamics in a semi-arid hillslope

Ben Gurion University of the Negev (2014)

Integrating GIS and numerical solution of flow equations to evaluate the impact of soil sealing on spatial and temporal soil moisture dynamics in a semi-arid hillslope

Shai Sela

Titre : Integrating GIS and numerical solution of flow equations to evaluate the impact of soil sealing on spatial and temporal soil moisture dynamics in a semi-arid hillslope

Auteur : Shai Sela

Université de soutenance : Ben Gurion University of the Negev

Grade : Doctor of Philosophy (PhD) 2014

Résumé partiel
[1] Rainfall-induced soil surface sealing is a wide spread phenomenon in dry climates, both in agriculture fields and in natural environments. The seal layer has lower hydraulic conductivity than the underlying soil, and thus affects substantially hydrological fluxes. The majority of previous studies have focused mainly on the effect of surface sealing on runoff generation. However, previous studies have also found that the seal layer limited hydraulic conductivity can reduce evaporation rates from the soil, an aspect mostly disregarded in most studies. Thus the seal layer can shape the spatial and temporal water content distribution by being both a non-local control, affecting water content downslope by runoff generation ; and a local-control, affecting water content in-situ by reducing evaporation rates. The non-local control of the seal layer on spatial water distribution was studied extensively previously. This thesis is focused on the understudied effect of the local control of surface sealing on the soil water content. The rainfall regime in dry climates leads these two controls of the seal layer on soil water distribution to operate in different time scales : while the non-local control is active during the rainfall events (with a duration of minutes or hours), the local control operates for a much longer duration (days to weeks) during the drying periods between storms. Therefore, reduction of evaporation rates during the drying periods has the potential to substantially effect soil water content distribution, an affect which has not been studied yet. Furthermore, up to date explicit representation of surface sealing in hydrological models used to study soil water content dynamics has been rarely applied, and as such no data currently exists regarding the interactions between surface sealing and the soil water content at different soil profile depths, its variability, and the consequences for vegetation water availability.
[2] This thesis is the first attempt to address these research gaps, focusing on a semiarid hillslope scale. A hillslope was chosen (0.075 km2 ) at the Lehavim Long Term Ecological Research (LTER) site in the Negev desert at the middle of Israel (31020’ N, 34045’ E). Annual rainfall is 290 mm and the soils are brown lithosols and arid brown loess, prone to surface sealing. The hillslope was characterized by means of extensive field surveys sampling topographic and soil hydraulic parameters. This database was interpolated into continuous layers using GIS tools. Our modeling XVI approach included both 1D and 2D numerical solutions of the flow equations applying physically-based approaches. For the hillslope scales simulations a modeling framework was developed (aggregated Hydrus 1D runs). This framework (3X3 m, 8240 cells) accounts explicitly for topographic and soil hydraulic parameters sampled in the field. Hydraulic properties of the seal layer at the soil surface were modeled following Mualem and Assouline [1989]. A MATLAB code was written to automate and force the independent Hydrus1D runs with a suit of short and long term climatic data (33y). All independent water balance solutions were than aggregated to represent the hillslope scale water content dynamics. Two cases were considered : an undisturbed soil profile (unsealed case) and a profile fully topped with a seal layer. The modelling framework was validated extensively at the field during the 2010-2011 rainfall seasons with a satisfactory result (Nash-Sutcliff coefficient of 0.7). More importantly, accounting for surface sealing was found to improve the model predictions compared with an unsealed case.

Mots clés : Soil moisture — Arid regions — Simulation methods — Hillside planting – Research — Sarcopoterium — Effect of water levels on — Observations

Annonce (BGU)

Version intégrale

Page publiée le 6 novembre 2016, mise à jour le 21 novembre 2018