Recent research have suggested that river and flooded land in Amazon basin are large sources of co2 to the atmosphere (Wofsy et al., 1988; Richey et al, 1990). The overall goal of this project is to explore the potential influence of co2 fluxes from river and inundated land on regional carbon balance in the Tapajos region. In addition to the prevailing trade wind regime, topographic and surface variations in the Tapajos region produce mesoscale circulations that have been documented from field data. These include a shallow diurnal river-breeze circulation forced by contrasts in sensible heat flux between the forest and the Tapajos and Amazon Rivers. We investigated this circulation using the CSU Regional Atmospheric Modeling System (RAMS). Several modifications have been made to RAMS Version 4.3: respiration rate and maximum photosynthesis are prescribed according to vegetation type; photosynthesis varies sinusoidally during daylight hours; Freitas et al (2000) convective transport scheme has been implemented to account for the deep tropical convections. A three month period (June-August, 2000) was simulated with two nested grids: The outer grid covers the northeastern part of Amazon and adjacent oceans at 100 km grid-spacing, and the inner grid covers an area of 1750 km x 1450 km, focusing on Santarem and adjacent area at 25 km grid-spacing. To investigate the effects of surface water co2 emission on the regional carbon balance, we performed these simulations with and without specifying the co2 fluxes from the river and inundated land. Smaller scale simulations that resolve the river breeze were performed for cases chosen from the regional simulations. These simulation reveal that heterogeneous vegetation produces easily detectable spatial structure in simulated co2 concentration. These gradients may be quantitatively interpreted to estimate area-averaged co2 fluxes under some circumstances, and form the basis of a field campaign using light aircraft which will be carried out in August of 2001. We found that surface water co2 flux modifies simulated co2 concentrations, especially at night. Experimental design for field campaign must treat trade wind and riverbreeze regime differently.