The potential effects of climate change on vegetation are of increasing concern. In the Mediterranean region, the dominant impact of climate change is expected to be through the modification of water balance. In this paper we present a model developed to predict drought-induced mortality of woody plants under different climatic scenarios. The model is physiologically-based and simulates water transport within individual woody plants, which can be isolated or competing for a common water resource. The model assumes that plant mortality is controlled by the carbon balance: when the plant is unable to transport water to the leaves it ceases to acquire carbon and, if this situation lasts long enough, it can no longer survive. In the particular application that we report in this study, two evergreen species are compared, Quercus ilex and Phillyrea latifolia, which were very differently affected by the acute drought that occurred in E Spain in summer 1994. While in some Q. ilex populations the amount of individuals that dried completely was up to 80%, P. latifolia showed almost no damage. During the years 1999 and 2000, canopy transpiration was monitored using sap-flow sensors in individuals of these two species in a Holm-oak forest from NE Spain. A Generalised Likelihood Uncertainty Estimation (GLUE) approach was used to calibrate the model against sap-flow measurements. The only difference between species that was introduced 'a priori' was that Q. ilex was more vulnerable to xylem embolism than P. latifolia (based on our own measurements in the study area). During the calibration process the information provided by the measured sap flows was used to retain the more likely parameter sets for each species. These parameter sets were used in all the following simulations. The model was able to accurately simulate transpiration dynamics of the two species in the study area. When the meteorological conditions of summer 1994 were introduced, the model outputs also reproduced the differential impact that drought had on the two species studied. In the simulations under climate change two factors were explored: the increase in mean temperature (+ 1.5, + 3 and + 4.5 degreesC) through its effect on ET, and the duration of summer drought. Under any of the scenarios, mortalities were much higher for Q. ilex: while this species was predicted to survive with less than 5% mortality droughts of up to 84-94 days, the mortality of P. latifolia reached 5% between days 133 and 150. For droughts longer than 3 months, which is approximately the current drought duration in the study area for dry years, the mortality of Q. ilex increased sharply. These results are discussed in relation to the possible long-term impacts of climate change on Q. ilex-dominated forests. (C) 2002 Elsevier Science B.V. All rights reserved.
