Effects of drought and passive nighttime warming on soil water balance of shrubland ecosystems

Understanding and predicting the climate driven changes in ecosystem functioning requires realistic studies at ecosystem level, involving experimental manipulation of temperature, water and CO2. A new non-intrusive experimental method has been developed and employed within 2 European projects (VULCAN and INCREASE). Climate manipulations, which involve a field scale drought and nighttim warming technique, were carried out at 6 shrubland sites located in Denmark (DK), Hungary (HU), Italy (IT) Spain (ES), United Kingdom (UK) and The Netherlands (NL) during the period 2001-2012. In this study we used data from the climate manipulation experiments in combination with a simplified soil water balance model (SWBM) in order to figure out the effects of drought and warming treatments, compared to control treatment, on the water balance of shrubland ecosystems. The SWBM operates with a daily iteration time step, and is driven by time series of corrected rainfall and actual evapotranspiration (ETa). Daily ETa values are obtained from the calculation of reference evapotranspiration (ETo) in combination with empirically determined vegetation coefficients (Kc), which are calculated splitting the computation into 2 components, transpiration and evaporation (dual coefficient methodology). The determination of Kc values and the calibration of SWBM were made by trial and error comparing model output to soil water content measurements in order to achieve the best model performance. Analysis of data shows important differences between treatments and sites, with the largest increase in ET rate occurring in the warming treatment (+4%) and at the northern sites (DK, UK, and NL). The effect of the drought treatment on ETa rates was much larger, with a decrease ranging from -10% (IT) to -35% (DK and UK). Warmer conditions determine a general reduction of the ETa/ETo ratio (from -2%, northern sites, to -8% in Spain). The effect of the drought treatment on ETa/ETo is much more relevant, with a reduction ranging from -10% (HU and IT) to -35% (DK and UK). The drought treatment determines also a general reduction of the available water content (from -5%, HU, to -30%, ES), which, conversely, is not significantly affected by the warming treatment. In addition, SWBM shows a very good agreement between model soil water content and actual observations with a modeling efficiency index larger than 0.80. These results suggest that using a simplified soil water model combined with climate manipulations allows the assessmen of the potential effects of climate change and variability on the water balance of shrubland ecosystems.