Simulating Plant Water Stress and Phenology in Seasonally Dry Tropical Forests: Plant Hydraulics and Trait-Driven Trade-Offs

Friday, 19 December 2014: 2:40 PM
Xiangtao Xu1, David Medvigy1, Jennifer S Powers2 and Justin M Becknell3, (1)Princeton University, Geosciences, Princeton, NJ, United States, (2)University of Minnesota Twin Cities, Department of Ecology, Evolution and Behavior, Minneapolis, MN, United States, (3)University of Alabama, Department of Biology, Tuscaloosa, AL, United States
Seasonally dry tropical forests account for over 40% of the forested area in tropical and subtropical regions. Previous studies suggest that seasonal water stress is one main driver of phenology and related vegetation dynamics in seasonally dry tropical forests. Species that coexist in seasonally dry tropical forests have different plant traits, experience different degrees of plant water stress and show distinctive phenological patterns. However, the observed diversity in plant phenology and related vegetation dynamics is poorly represented in current dynamic vegetation models. In this study, we employ a new modeling approach to enhance our model skills in seasonally dry tropical forests. First, we implement a new plant hydraulic module under the framework of a state-of-the-art dynamic vegetation model, Ecosystem Demography 2 (ED2). Second, we link plant water stress with several key coordinated plant traits. Unlike previous models, the updated ED2 does not prescribe leaf phenology (deciduous or evergreen) and plant water stress is not determined by empirical water stress factors or by soil moisture alone. Instead, the model tracks more mechanistic indicators of plant water stress like leaf water potential, accounts for different abilities to tolerate water stress among plant functional types and predicts dry season leaf deciduousness and related vegetation dynamics. The updated model is then tested with in-situ meteorological data and long-term ecological observations. We also perform numerical experiments to explore the possible biases of ignoring the observed diversity in seasonally dry tropical forests. We find that (i) variations of several key plant traits (specific leaf area, wood density, turgor loss point and rooting depth) can account for the observed distinctive phenological patterns as well as inter-annual variations in vegetation growth among species. (ii) Ignoring the trait-driven trade-offs and diversity in seasonality would introduce significant amount of biases in model predictions of ecosystem energy and water fluxes.