TG-05 Abstract

Modeling the Effects of Land Use Change and Surface Hydrology on Carbon and Trace Gas Fluxes for the Amazon Region

Claudio Jose Reis de Carvalho, EMBRAPA/CPATU (SA-PI)
Raimundo Cosme de Oliveira, EMBRAPA (SA-PI)
Christopher S. Potter, NASA/ARC (US-PI)
Carlos Schenato, (SA-PI)

The Amazon region includes the largest remaining tropical forest ecosystem on Earth.  Despite the Amazon’s potential importance for climate regulation, the precise pattern of a terrestrial sources and sinks for CO2 and other “greenhouse gas” compounds like methane remain uncertain for the region.  These large gaps in our knowledge still exist to a great degree because many impacts of land cover change and geochemical controls related to surface hydrology have not been understood in adequate detail to determine precise regional fluxes for biogenic trace gases.  Therefore, the fundamental approach we propose for LBA-ECO II research involves integrated studies and predictive modeling for Amazon carbon, water, and nutrient cycles using a regional ecosystem model called NASA-CASA (Carnegie-Ames Stanford Approach), which has been uniquely developed and applied using satellite data drivers as a component of the NASA LBA-ECO I program.  Our proposed work thereby builds upon over four years of coordinated data assimilation and model development, as part of the LBA Science Team.  The research plan we now propose is designed to extend the value of new LBA site (tower and small plot flux) data and remote sensing products by up-scaling with NASA-CASA to the regional level using a variety of different land cover products available for the Amazon region.  In this manner, we propose to make new linkages in LBA-ECO project science between site-based investigation and regional scale modeling and remote sensing.

A unique aspect of our study approach is to treat land cover inputs to our ecosystem carbon modeling as variable  according to time, spatial scale, and spatial resolution.  In the proposed research plan, the NASA-CASA ecosystem model will make a series of simulation runs using multiple sequences of land cover change for the entire Amazon regional extent.  The simulations will include the following land cover/land use class settings  -- moist primary forest, dry primary forest, drought-deciduous forest, savanna, disturbed (burned), secondary forest, pasture, and annual crop.  Nominal spatial resolution for hydro-meteorologic inputs to our LBA regional simulations will be 8x8 km pixel size.  The percentage cover-weighted results from these individual model runs for land cover change sequences will be aggregated to reconstruct a history of regional heterogeneity in Amazon ecosystem dynamics.  Trace gas fluxes (e.g., CO2, N2O) will be weighted by land cover change frequency distributions derived from the highest spatial resolution data available (e.g., from MODIS and Landsat sensors) for the entire region and for selected smaller areas around intensive LBA studies.  High resolution (< 1-km) land cover images for 'footprint' areas of LBA tower sites will be used to define the proportions of each land cover class with which to weight ecosystem model estimates and validate against measured tower fluxes of carbon and water exchange.  Particular emphasis will be placed on Tapajós National Forest (TNF) tower sites, where our Brazilian co-investigators are actively making field measurements of energy, water, and carbon exchange, which are vital to ecosystem model evaluation.

A second integrative aspect of this study will be to extend NASA-CASA model predictions of carbon and trace gas fluxes to (seasonally) inundated areas of the region.  Through close collaboration with other LBA investigator teams whose studies have developed or used radar remote sensing products or surface water models, we will adapt our CASA model to include Amazon mainstem and floodplain dynamics, with focus on up-scaling wetland ecosystem methane and CO2  emission fluxes.  At the smaller scale of the TNF watershed, we will complete the development of a new parameterization scheme within the NASA-CASA model for coupling a surface water routing model called HYDRA (Univ. Wisconsin), and validating simulations of surface water flow against measured flow and dissolved carbon chemistry of the Rio Moju discharge network from the TNF.