How does forest cover impact water flows and ecosystem services? Insights from “real-life” catchments in Wallonia (Belgium)

While planet boundaries are being crossed and ecosystems degraded, the Ecosystem Service (ES) concept represents a potential decision-making tool for improved natural resources management. The main aim of this paper is to assess the impact of forest cover on water related ES in Wallonia (Belgium) in terms of quantity and timing. We developed an approach based on easily accessible data, monitored in several countries and using straightforward statistical methods. This led us to study ES at “real-life” catchments scale: 22 catchments – from 30 to 250 km²– with mixed land covers were studied. We approached the water supply and flood protection services through 5 indicators extracted from 10 hydrological years (2005–2014) discharge data series. These were computed annually and seasonally (vegetation period from March to September and “non-vegetation” period the rest of the year). The water supply was assessed through the specific volume Vs, the baseflow index BFI and the specific discharge exceeded 95% of the time Q₉₅s whereas the flood protection service was approached through the specific discharge exceeded 5% of the time Q₀₅s and the flashiness index FI. Our study gives two main insights. First, statistical analyses show that forest cover negatively impact water supply when studying annual and “non-vegetation” period flows in general (Vs) but positively when studying low flows (Q₉₅s). Regarding flood protection a slightly negative impact of forest cover on high flows (Q₀₅s) was highlighted in the “non-vegetation” period. Results also show a negative impact of forests annually and in the vegetation period on the flashy behaviour of the catchment thus a positive impact on the flood protection ES. The “year” effect is overall highly significant testifying the importance of climatic factors. Rainfall is often significant and can be considered as a main driver of these ES. Secondly, analyzing the quality of the models produced and the results overall we assume that other variables characterizing the catchments such as topography or soil types do play a significant role in the delivery of these ES. This questions the use of land cover proxies in assessing and mapping of hydrological ES at a complex landscape scale. We thus recommend further research to keep improving land cover proxies if they are used.