Extrapolating carbon dynamics of seasonally dry tropical forests across geographic scales and into future climates: improving simulation models with empirical observations

Seasonally dry tropical forests (SDTFs) occur in areas with warm temperatures and a pronounced dry season with little to no rainfall that lasts 3 to 7 months. The potential area covered by this biome is vast: globally, 47% of all forest occurs in tropical and subtropical latitudes, and of all tropical forests approximately 42% are classified as dry forests. Throughout the last several centuries, the area covered by SDTFs has been dramatically reduced through conversion to grazing and croplands, and they are now considered the most threatened tropical biome. However, in many regions STDFs are now growing back. These forests are valuable because they are reservoirs of unique biodiversity. Additionally, because growing tropical forests store so much carbon in vegetation and soils, these ecosystems may play an important role in climate mitigation. The extent to which SDTFs are vulnerable to ongoing and future global changes, such as climate change and nitrogen deposition, is not yet known.

Despite their global importance, assessing the response of SDTF carbon dynamics to altered climate or nutrient deposition is extremely challenging. Many ecosystem models have not resolved SDTFs from wetter tropical rainforests or drier savannahs. Model improvements are strongly limited by availability of empirical data, especially for belowground processes such as root production and nutrient uptake. Therefore, the central objectives of this research are to qualitatively and quantitatively describe how belowground processes mediate the response of SDTF carbon dynamics to environmental change. To do so, we propose an interdisciplinary approach that links field observations across a range of dry forest sites, manipulative experiments, and model simulations that quantify sensitivity of ecosystem carbon cycling to external forcings. Our research will integrate specific, targeted ecosystem measurements with a generalizable, trait-based framework for predicting ecosystem responses to climate change. These data will then be incorporated into predictive models used by the global change community.

SDTFs span an incredible variety of natural habitats, with large variation in precipitation regime, soil nutrient availability, and plant community composition across the biome. To capture this natural variation, we will partner with collaborators who manage existing networks of data-rich forests in Costa Rica, Mexico, Puerto Rico and Colombia. Our sampling strategy will quantify the relationships among above- and belowground plant traits, and the ways in which plant properties impact soil C, nitrogen (N), and phosphorus (P) cycling. At all sites, we will measure a suite of leaf, wood, and root traits, tree growth, litterfall, fine root production, nutrient fluxes and soil C stocks. Simultaneously, we will also establish two experiments at a well-characterized site in Costa Rica: a nutrient addition and drought experiment in plantations, and a full-factorial fertilization experiment with N and P in diverse mature forests. Together, these manipulative experiments will provide generalizable data on responses to altered nutrient and water availability, and allow us to extrapolate these responses to the ecosystem scale.

Next, our empirical data will be used to parameterize and validate two ecosystem models, Ecosystem Demography 2 (ED2) and the Accelerated Climate Modeling for Energy (ACME). Both models use information on plant functional types (PFTs) to simulate land-atmosphere exchange of carbon, water, radiation, and energy. However, ED2 and ACME have different ways of representing variations in resource availability, species succession, and plant-water interactions. Therefore, comparison of the two models will help us to quantify structural uncertainty in model predictions of ecosystem responses to altered climates. We will then use these models to explore the potential responses of SDTF ecosystems to various global change scenarios. Evaluation of model simulations will also be used to identify outstanding knowledge gaps and to suggest future field experiments.

As a whole, this body of research complements DOE's ongoing research in tropical wet forests. This work has the potential to generate significant conceptual advances in the fields of functional ecology, global change biology, and ecosystem modeling. Ultimately, our combined measurement and modeling approach will elucidate controls on C cycling in SDTFs, leading to better understanding of atmospheric carbon dioxide dynamics, nutrient deposition, and climate change feedbacks.