Coordinated Biological, Chemical,and Atmospheric Investigations of the Amazon as a Carbon Sink
David Medvigy, assistant professor of geosciences, and Lars Hedin, professor of ecology and evolutionary biology, are coordinating field- and model-based assessments of the response and resilience of tropical ecosystems to global environmental change. Their study will seek to understand how nutrient feedbacks can affect the strength of the tropical forest carbon sink in the future, to better resolve the processes responsible for the conversion of soil carbon to atmospheric carbon dioxide (CO2), and to investigate how plant diversity impacts the response of tropical forests to climate change. This project is supported by the Carbon Mitigation Initiative (CMI), a partnership between Princeton University and BP.
A Princeton Institute for Rainforests and the Amazon including their Nutrients, Hydrology, and the Atmosphere (PIRANHA)
Medvigy and Hedin, in coordination with their CMI funded project, are creating a tropical rainforest research community by bringing together Princeton faculty, undergraduates, and GFDL researchers called the PIRANHA Consortium. This Consortium will serve as a platform for integrated biological, biogeochemical, and atmospheric investigations relevant to tropical forests and to the impacts of large-scale deforestation. The project will give Princeton students opportunities to actively engage in consortium meetings, summer internships, independent research projects, and new course offerings centered on tropical ecosystems such as the Amazon. Support for this project comes from Siebel Energy Challenge.
Seasonal variability in forest leaf area and its consequences for terrestrial carbon budgets and ecosystem structure
One example of Medvigy’s recent work on temperate forests considered the timing of leaf emergence in the spring (Jeong et al. 2011). In previous studies, poor understanding of the physical and biological controls on leaf emergence has led to large errors in the carbon fluxes simulated by numerical models. With funding from NOAA/CICS, Medvigy and his postdoc Su-Jong Jeong have been investigating species- and climate-dependence of leaf emergence. They used two complementary datasets: a 17-year record from a single site, and a 2-year record from ~50 sites. They also used a powerful statistical approach, reversible-jump Markov chain Monte Carlo, that had not previously been used in terrestrial biosphere modeling. Jeong et al. (2011) concluded that pioneer-type tree species had a markedly different leaf emergence date (by up to two weeks) than other deciduous trees. They also showed that this had important implications for ecosystem carbon budgets. Representing all deciduous trees with a single, aggregate leaf emergence scheme instead of a species-specific scheme was found to lead to errors in ecosystem productivity of up to 20%. In collaboration with Elena Shevaliakova, Sergey Malyshev, and Ron Stouffer, this new representation of leaf emergence has been implemented into the GFDL LM3 model.
Deforestation in the Amazon
The Medvigy Laboratory is also investigating how future deforestation in the Amazon can alter precipitation statistics in South America using a novel scaling approach. Previous global-model studies of deforestation have used ~250 km resolution, which is too coarse to resolve actual spatial patterns of deforestation or to simulate “deforestation breezes”. Previous limited-area models studies have used a ~25 km resolution, but these studies require lateral boundary conditions that are totally unconstrained and exert a significant influence on model simulations. In collaboration with researchers from the University of Miami, Medvigy attempted to combine the best features of both approaches by using a variable-resolution GCM, OLAM (Ocean-Land-Atmosphere Model). The model’s grid mesh was set up to cover South America and nearby oceans at ~25 km resolution, and then to gradually coarsen and cover the rest of the world at ~200 km resolution. Because of the computational efficiency of this approach, it was possible to carry out the first decadal-scale simulations of Amazon deforestation at a resolution that can resolve some regional-scale atmospheric circulations as well as landscape heterogeneity. The results indicated that deforestation reduces simulated precipitation in the Amazon, but this reduction (~5%) was much smaller than that seen in most previous global-model studies (~20%). Interestingly, a sub-continental re-distribution of precipitation was found whereby the northwest Amazon becomes drier and the southeast Amazon becomes wetter. This has potentially important implications for agriculture, because most agriculture is currently located in the southeast.
Medvigy also investigates how deforestation can affect extreme event frequency and intensity. In order to capture smaller scales within the context of a GCM simulation, Medvigy and colleagues used the OLAM model to simulate South America at 25 km resolution. They found that deforestation induced large changes in the frequency of June-July-August extreme cold events. Large increases in cold event frequency and intensity occurred in the western Amazon and, surprisingly, in parts of southern South America, far from the actual deforested area. One possible mechanism for these remote effects involved changes in the position of the subtropical jet stream, caused by temperature changes in the Amazon. Because much of southern South America is under agriculture and this region is extremely sensitive to frosts, deforestation in the Amazon has the potential to remotely effect agricultural productivity.
Medvigy is also carrying out research on the importance of day-to-day climate variability for terrestrial ecosystems. One aspect of particular interest is the significance of high-frequency variability of environmental parameters (sunlight, precipitation, temperature) for the structure and function of terrestrial ecosystems under current and future climate. In a recent study, Medvigy and colleagues used the ED2 model and the long-term record of carbon fluxes measured at Harvard Forest in Massachusetts. They found that fluctuations of sunlight and precipitation are strongly and nonlinearly coupled to ecosystem function, with effects that accumulated through annual and decadal timescales. Increasing variability in sunlight and precipitation led to lower rates of carbon sequestration and favored broad-leaved deciduous trees over conifers. Temperature variability has only minor impacts by comparison. They also found that projected changes in sunlight and precipitation variability have important implications for carbon storage and ecosystem structure and composition. Based on IPCC model estimates for changes in high-frequency meteorological variability over the next 100 years, we found that terrestrial ecosystems will be affected by changes in variability almost as much as by changes in mean climate.
Abrupt changes in terrestrial carbon uptake
A series of recent studies have suggested that the Earth underwent significant synchronous climatic shifts around 1988/89 towards enhanced warming in the northern latitudes and increased solar radiation associated with reduced cloudiness in the tropics. A new synthesis analysis based on the growth rate of atmospheric CO2, fossil fuel emission estimates, and modeled ocean CO2 uptake reveals that there may have been also a major shift towards greater land carbon uptake around this time frame. The detection and attribution of rapid large ‐ scale shifts in the global carbon cycle is of primary concern for society since they lead to feedbacks that either amplify or diminish physical climate change. In collaboration with colleagues from Princeton, UCLA, and NASA, Medvigy is developing a framework for the detection and attribution of rapid regime or large ‐ scale shifts in the terrestrial carbon cycle. The framework consists of statistical methods to detect changes in the fundamental behavior of the carbon cycle as well as terrestrial ecosystem modeling components to unravel the processes that are responsible for these shifts.