The effect of natural revegetation on streamflow in the Spanish Pyrenees

Abstract

Many of the Mediterranean lowlands depend for their fresh water supply on mountain areas. However, streamflow in mountain areas has been decreasing in the past century. Apart from changes in climate, land use changes have been identified as having as having an impact on the hydrology in mountain areas. One major land use change has been a strong increase in vegetation, which was the result of land abandonment associated with depopulation. The topic of this research is the change of vegetation and its effect on stream flow in the Upper Aragón basin between 1950 and 2050. First, the effect of vegetation on hydrology was modelled using the PyCatch hydrological model in a small catchment within the Upper Aragón basin. Secondly the sensitivity of the modeling outcomes to certain vegetation parameters was assessed. Thirdly, a non-linear vegetation growth equation was fitted through the pixels of twelve Landsat satellite images to determine vegetation growth in the Upper Aragon Basin for the 1950-2050 period. And finally, vegetation growth was coupled to streamflow, in order to determine changes in streamflow in the Upper Aragon Basin for the 1950-2050 period. Hydrological modeling for the Arnás catchment shows that vegetation growth has a large impact on streamflow. Especially initial vegetation growth (from LAI=0 to LAI=1) leads to almost complete drying up of summer streamflow. A complete vegetation recovery (from LAI=0 to LAI=4.5) will lead to a 100 percent decrease yearly streamflow in the Arnás catchment. The non-linear regression suggests a fast increase in vegetation in the UAB between 1960 and 2000. This translates to a stark decrease in streamflow for this period. This might be an overestimation of what is really happening in the UAB. There are several ways in which our research could be improved in the future. First, the NDVI-LAI saturation problem we encountered in our study, could be a solved by using a different method to find LAI values. For example, the Enhanced Vegetation Index (EVI) could be used. Modeling the vegetation growth could be improved by finding a regression equation that can more accurately describe vegetation growth in this region, and that does not result in missing values in case of constant vegetation or vegetation decrease. In future research also more attention should be given to changes in soil as a result of vegetation change, and its effect on hydrology. Another issue with the hydrological modeling is that, according to our results, almost all streamflow for all scenarios seemed to be the result of sub surface flow. Better ways to model Horton overland flow need to be sought to more realistically describe the hydrological fluxes in the Upper Aragon basin.