Quantifying and modeling water availability in temperate forests: a review of drought and aridity indices

Climatic water availability is a major determinant of forest structure and composition, while drought events may severely impact forest dynamics. In recent decades, an increasing number of severe drought events has been reported in forests around the world. In the future, climate models project increasingly dry conditions in many temperate regions. Various tools have been applied to better understand the effects of drought on forests, such as dendrochronological analyses, climatic trend analyses and dynamic models. With these approaches, water availability is often summarized as a single scalar, termed a drought or aridity index. As droughts are complex phenomena, such indices are always associated with a loss of information. Many different such indices exist, and have been developed for various purposes. This review asks whether some of these indices are more suitable than others to quantify water availability in temperate forests. In a first step, the rationale and theoretical background of different drought indices are spelled out and compared among each other. Then, evaluations and intercomparisons of drought indices from the literature are reviewed. The implementation of drought indices in dynamic forest models is also discussed. Finally, two current research questions are identified: the role of dry air for physiological drought, and the suitability of various drought indices under climate change. It appears from this review that indices accounting for evaporative demand generally perform better than indices based on precipitation alone. When comparing sites with different edaphic conditions, indices accounting for soil moisture storage are more suitable. Nevertheless, results from intercomparisons show considerable divergence, and it is not possible to clearly favor one index. Furthermore, a differential response of tree species to different drought indices is often observed, although no clear pattern emerges from this comparison. More intercomparisons of indices, across climates and species, might provide valuable knowledge. Another key finding is that the properties of indices also depend on choices regarding, e.g., the calculation of evaporative demand, or the underlying water balance model. Reporting such methodological details could greatly increase the value of future evaluations of drought indices.