Modelling surface runoff and water fluxes over contrasted soils in pastoral Sahel: evaluation of the ALMIP2 land surface models over the Gourma region in Mali

Authors: GRIPPA, M. - Collaborator; KERGOAT, L. - Collaborator; BOONE, A. - Collaborator; PEUGOT, C. - Collaborator; DEMARTY, J. - Collaborator; CAPPELAERE, B. - Collaborator; GAL, L. - Collaborator; HIERNAUX, P. - Collaborator; MOUGIN, E - Collaborator; Anderson, Martha; HAIN, C. - University Of Maryland

Submitted to: Journal of Hydrometeorology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/3/2017
Publication Date: 6/22/2017
Citation: Grippa, M., Kergoat, L., Boone, A., Peugot, C., Demarty, J., Cappelaere, B., Gal, L., Hiernaux, P., Mougin, E., Anderson, M.C., Hain, C. 2017. Modelling surface runoff and water fluxes over contrasted soils in pastoral Sahel: evaluation of the ALMIP2 land surface models over the Gourma region in Mali. Journal of Hydrometeorology. 18:1847-1866. https://doi.org/10.1175/JHM-D-16-0170.1.
DOI: https://doi.org/10.1175/JHM-D-16-0170.1

Interpretive Summary: The African Sahel is a region where rainfall patterns and land-surface moisture conditions are strongly coupled, meaning soil moisture distributions on the land can significantly influence the weather. While rainfall is critical to maintaining agricultural production and moisture sensitive ecosystems in the Sahel, this coupling is not well understood and interesting contradictory patterns in the regional water balance have been observed over the past decades. While the region has been under periodic drought for the past 30 years, increases in streamflow, runoff and surface water have been measured. Some evidence points to land cover change and soil erosion as the key to recent modifications to the regional surface hydrology. To better understand the role of surface processes in water cycle and atmospheric coupling, a targeted effort to intercompare and improve land-surface models over this region has been undertaken as par of the AMMA (African Monsoon Multidisciplinary Analysis) Land surface Model Intercomparison Project (ALMIP). The second phase of this project, ALMIP2, compares model performance at key observation sites covering an eco-climatic gradient from Benin to Niger to Mali. This paper intercompares land-surface model performance at the Mali site, representing dry sparsely vegetated conditions at the northernmost site inthis gradient. Surface water evaporative fluxes (evapotranspiration, or ET) predicted by the ALMIP model suite ensemble are compared to measurements made at flux towers at the Mali site, and to three independent estimates from ET remote sensing approaches. The results show that while individual models within the suite may over or underestimate ET at the observation site, the ensemble average appears reasonable and in general agreement with the remote sensing estimates. The models, however, do not do a good job at distinguishing between shallow and deeper soils characteristic of the region, and these have a strong impact on runoff generation. The remote sensing models diagnose these soil conditions, and better represent spatial gradients in moisture fluxes across the landscape. The results of the study indicate that landsurface models running in this region either need very good soils maps, or they need to be coupled with remote sensing algorithms to better reproduce runoff and water fluxes in this part of the Sahel.

Technical Abstract: Land surface processes play an important role in West African monsoon variability and land –atmosphere coupling has been shown to be particularly important in the Sahel. In addition, the evolution of hydrological systems in this region, and particularly the increase of surface water and runoff coefficients observed since the 1950s, has had a strong impact on water resources and on the occurrence of floods events. Better understanding and modeling the spatial and temporal variability of the different components of the continental water cycle at the mesoscale is necessary to improve the representation of the coupled land-atmosphere system as well as to predict the evolution of Sahelian hydrological systems. This study addresses results from the second phase of the AMMA (African Monsoon Multidisciplinary Analysis) Land surface Model Intercomparison Project (ALMIP2), carried out to evaluate the capability of different state-of-the-art land surface models to reproduce surface processes at the meso scale. Runoff and water fluxes simulated by 20 land surface models over the northernmost site of the ALMIP2 project, the Mali site, are investigated. Model evaluation is carried out by comparing to runoff estimations over endorheic watersheds as well evapotranspiration derived by eddy covariance measurements and by three remote sensing based products (ALEXI, MODIS and GLEAM). It is found that, over deep sandy soils, surface runoff is generally overestimated but the ALMIP2 multi-model mean reproduces in situ measurements of ET and water stress events rather well, generally better than the remote sensing products analyzed here. However, ALMIP2 models are generally unable to distinguish among the two contrasted hydrological system typical of the study area, i.e. shallow soils generating runoff ending up into ponds and deep sandy soils. Employing as input a soil map which explicitly represents shallow soils, improves the representation of water fluxes, namely runoff and evaporation, for the models that can account for their representation. The effects of varying soil depth on evaporative loss were better captured by the diagnostic thermal-based ALEXI model, reflected in the higher local land surface temperatures. Therefore, a better representation of these soils, in soil databases, models’ parametrizations and remote sensing algorithms, is fundamental to improving the estimation of runoff and water fluxes in this part of the Sahel.