Hydrologic Sustainability of a Dryland Tree-Grass Ecosystem in the Mediterranean Region Under Climate Change

Abstract

Dryland ecosystems are widely spread all around the world, and are characterized by their sensitivity to meteorological seasonal and decadal changes, which impacts water availability and ecosystem sustainability. For instance, in Mediterranean dryland ecosystems climate change occurred with an increase of air temperature and a decrease (mainly in wet seasons) of precipitation, which are key atmospheric forcing for grass and tree growths. Climate predictions of future scenarios of the Intergovernmental Panel on Climate Change (IPCC) are even worse, affecting, for instance, the central Mediterranean basin with a further decrease of rainfall in wet months and an increase of air temperature. The case study is a typical Mediterranean ecosystem in Sardinia, where wild olives and seasonal grass species grow on thin surface soil layer overlaying a fractured rock sublayer, and for which a long-term dataset of micrometeorological, tree transpiration, remote sensing data and soil water content measurements is available. Our objectives are: 1) detect trends and changes on the evolution of tree cover spatial distribution related to changes on climate conditions, and investigate the impact on soil water and evapotranspiration using a long ecohydrological database of a typical water-limited ecosystem; 2) develop an ecohydrological model for long-term predictions able to capture the evolution of tree cover spatial distribution, vegetation dynamics, and soil water balance interactions; 3) investigate the impact of future climate scenarios with increase of CO2 on soil water balance and tree hydrological sustainability of a Mediterranean dryland ecosystem. The Sardinian field site is characterized by a very attractive long database of almost 60 years of data, with micrometeorological and meteorological measurements, remote sensing data and aerial photography images, providing a unique opportunity to analyze the response of the tree-grass ecosystem to the historical climate and land cover changes. The proposed model was able to reproduce well the soil, vegetation and atmosphere interactions and dynamics, and their long-term evolution. The proposed update of the model was accurate for predicting the long-term dynamics of the tree cover fraction evolution, which have been reduced drastically (0.10) by a human induced fire almost 60 years ago, and restored naturally in almost 20 years, reaching the equilibrium value (0.33). The Sardinian tree-grass ecosystem suffered an historically significant reduction of the rain and a significant increase of air temperature in the last century, which produced dryer conditions but with a recent mean annual precipitation (MAP) still above 600 mm, apparently enough for sustain the tree growth. The GCM future scenarios are even worse, predicting a further decrease of MAP up to 400 mm, and an increase of air temperature up to +10 °C, which will cause a reduction of the tree cover fraction up to 0.10, and a strong decrease of the tree LAI. The soil water balance is predicted being drier, with also less grass and vegetation in general, with consequences on the landscape aspect, becoming more and more a savanna-like ecosystem. Water resources and environmental planning strategies need to be consider for increasing the resilience of the tree-grass ecosystems to the climate changes.