Most large scale computer models are limited in their ability to predict the dynamics of forests under novel environmental conditions because the critical process of inter-plant competition and the physiological responses to water stress are not adequately represented. This project addresses these two fundamental problems by incorporating a more realistic hydrodynamic model of water movement in trees into a large scale model designed to predict changes in the composition and structure of forests. Direct field measurements of the conductivity of water in stems of trees and the point at which low water availability in the soil leads to a loss in plant vigor are required to constrain the model. This project will enable these critical field data to be collected.
This research addresses the fundamental mechanisms that cause a tree to die under extreme water stress, particularly the physiological traits that control plant water use and adjust under water stress. That information will be formalized within the modeling framework that can then be used to identify the fundamental processes and associated tradeoffs that lead to survival or mortality within a dynamic tropical forest under extreme drought, and predict how shifts in species composition will play out as precipitation patterns change.
This project will fill a large gap in our knowledge about water use and stress in tropical trees. With the models updated and coupled, advances can be made in terms of predicting how forests of the Amazon will evolve as climate change evolves over the next hundred years.
1. Introduction and background Recent climate model predictions of how rainfall may shift across the Amazon basin are converging toward considerable drying in the east and longer dry seasons across most of the basin by the end of this century (1, 2). However, regional vegetation models are poorly parameterized to capture the effects of severe and chronic drought on forest biomass dynamics (3, 4, 5). The goal of this doctoral dissertation project is to address key uncertainties regarding the response of Amazonian tree species to changes in rainfall in order to accurately predict the fate of the Amazon forest over the coming century. This goal will be accomplished by incorporating a physiologically based plant hydraulic model (6) into the Ecosystem Demography model (ED), which is a dynamic vegetation model that is structured to account for competitive interactions between organisms (7). Direct field measurements of xylem hydraulic conductivity and xylem and leaf vulnerability curves are required to constrain the parameterization of the new plant hydraulic model. The objective of this NSF Doctoral Dissertation Improvement Grant was to support the fieldwork component of this dissertation project. The project leveraged benefit from the NSF-funded Partnership for International Research and Education in the Amazon of Brazil (NSF award #OISE-0730305) and took substantial advantage of the extensive infrastructure already established at the two field stations located in the Tapajós (8) and Caxiuanã (9) National Forests in the eastern Amazon. 2. Key Results Species representative of four plant functional types (PFT) were measured at each site: drought tolerant versus intolerant combined with early versus late successional (Table 1). Samples were collected from the top of canopy trees. At Caxiaunã, for each species there was no significant difference between samples taken from the control and drought plots for either the pressure-volume curves (Figure 1) or xylem vulnerability curves (Figure 3); therefore, the curves for each species were constructed with the combined data of both plots. The change point in a pressure-volume curve indicates the turgor loss point for the leaf as it desiccates (Figure 1). The turgor loss point was used as a proxy for the point at which the stomata are 50% closed due to a reduction in leaf water potential (Ylf) (10). At each site, turgor loss points of the drought tolerant PFTs occurred at lower Ylf compared to the intolerant PFTs (Figure 2). However, within either the drought tolerant or intolerant PFT, turgor loss points were not different between the successional PFTs, except at the Tapajós where within the drought tolerant PFT, the turgor loss point of the late successional PFT (Licania) occurred at a lower Ylf compared to the early successional (Protium) PFT (Figure 2). At Caxiuanã, the xylem P50 values—i.e. the point when 50% of xylem conductance is lost—for the two species in the drought tolerant PFT were -2.3 MPa, which was 1.0 MPa lower than the P50 values for the two drought intolerant species (Figure 4). There were no significant differences between P50 values of early versus late successional PFTs that were contained within either the drought tolerant or intolerant PFTs—i.e. the P50 values of Inga vs. Eschweilera or Protium vs. Licania were equivalent (Figure 4). In contrast, the xylem P50 values for species measured at the Tapajós had a narrower range than the species measured at Caxiuanã (Figure 5). Also Inga, the drought intolerant early successional species, had the highest P50 value (-1.1 MPa), while Eschweilera, the drought intolerant late successional species, had the lowest P50 value (-1.9 MPa). The two drought tolerant species, Protium (early successional) and Licania (late successional), had intermediate P50 values of -1.7 and -1.6 MPa, respectively. These results indicate that traits controlling drought tolerance are orthogonal to those controlling succession, which is important for establishing axes of competition in dynamic vegetation models like ED. 3. Key Outcomes Improving our basic understanding of the physiological mechanisms determining tolerance and fitness during extreme drought events advances our ability to make informed decisions regarding i) watershed management, ii) forest conservation and iii) commercial and private forest management. The results from this study can also provide a field-based parameterization for new hydraulic formulations to be tested in terrestrial biosphere models, such as ED, which is essential for improving our ability to assess how changes in precipitation will affect composition and long-term carbon stocks of different forest types around the world. Moreover, the resulting improved terrestrial biosphere models will enhance the ability of Earth System Models to accurately predict the impacts and magnitude of climate change over the coming century.
