Impact of mountain topography on potential evapotranspiration and water drainage

Hydrological models are currently used to simulate the water cycle at the catchment scale using climatic forcing. For water management purposes, one of the most important components to be determined is drainage. The estimation of drainage into an aquifer is directly related to the water input (precipitation), the outputs through evaporation and transpiration, and water flow dynamics in the unsaturated zone. The output fluxes are difficult to estimate through direct measurements and are often estimated using mathematical models build on climatic processes and data. Amongst the climatic data that are used, solar radiation is a key parameter since it estimates the energy available for open surface or soil evaporation and plant transpiration. Solar radiation can be computed directly knowing the sun’s position, provided by satellite surveys (with a spatial resolution down to 6x6km² and a time resolution of one hour) or interpolated from values measured at meteorological stations. Values based on observations should be preferred because direct computation is strongly biased due to the effects of weather conditions (cloud for example). In mountainous regions, the orientation of the hillslope regarding the sun's position can strongly impact the amount of solar energy arriving on the canopy or the soil.

The questions we address in this communication are the following: when applying physically based hydrological models to mountainous regions, is it really necessary to consider the potential sky obstruction due to the mountainous terrain of each grid cell to assess solar radiation? By rebound, does this strongly impact the estimation of evapotranspiration and water drainage to the aquifer?

To answer these questions, a mixed methodology that relies on two steps is proposed and tested. The first step consists of a theoretical computation of solar radiation for each grid cell of a given Digital Elevation Model using GIS tools. The second steps aim at correcting the first step computations to be consistent with measured or satellite data. For the first step, the (Scharmer, Greif 2000) model - which computes the 3 components of global radiation (direct, diffuse, and reflected by surrounding surfaces for clear sky conditions) – is used. The model also considers the local terrain to estimate the sky obstruction. In the second step, the data are averaged at the scale of the prescribed data (satellite or interpolated) and linearly corrected with a proportionality coefficient so that the average computed value fits the prescribed average value.

This methodology is then applied to a water catchment in the Vosges Mountains located close to Strasbourg (France). The proportionality coefficient varies locally between 0.2 and 2.5 showing that the local impact of topography on radiation is very significant. Using this correction coefficient in the Penman-Monteith formula, the relative difference in evapotranspiration is respectively -80% and +180% from the mean value for shaded areas and sunniest areas. For water drainage estimated through a conceptual model, the relative differences vary from -20% for the most exposed areas to +20% for the less exposed areas, demonstrating that orientation should be accounted for when simulating the response of mountainous watersheds.