||Tropical ecosystems (rainforests and grasslands) play a dominant role in regulating the global gross primary productivity, contributing 85 PgC/year that is 70% of the global value of 122 PgC/year. They account for about 60% of the total terrestrial land surface metabolism that is much greater than the 8 PgC/year and 10 PgC/year values for boreal and temperate forests respectively. However, predicting their response to future climate change (warming and increased precipitation) is highly uncertain due to limited observations, complex ecosystem-hydrologic-atmospheric couplings and the primitive state of the parameterizations of tropical ecosystem interactions in climate models. Our goal is to gather the first ever “regional scale” (10-50km) observations of atmospheric column CO2, CH4, H2O, HOD, N2O and CO concentrations over Amazonian tropical forests (downwind of the Manaus urban plume) to measure the atmospheric variability arising from the coupled carbon-water-nitrogen terrestrial cycle on time scales the span minutes, hours, days, seasons and years. Our first of the kind data in the Amazon will provide “new constraints” to develop process level mechanisms and evaluate terrestrial ecosystem parameterizations in climate models. Variability in temperatures (from seasons) sunlight (from cloud cover), soil water (from rainfall) and nitrogen (from fires and urban plumes) will be used to diagnose specific responses of the model to these parameters. Furthermore, we will evaluate how the heterogenic response of the carbon fluxes to variations in forest and subsurface properties observed by long term in situ tower measurements near Manaus averages (smears out) in our larger scale observations. The observed scale dependence of atmospheric composition will provide a new metric to link ecosystem processes to larger scale parameterizations in coupled land-atmosphere models (e.g. CLM/ESM). Our core activity entails the deployment of a solar tracking Fourier Transform Spectrometer (Bruker 125HR) that records high-resolution solar spectra in the near infrared in Manaus, Brazil during the collaborative multi-program OBER funded GOAMAZON campaign. Our system locks in and tracks the sun collecting spectra in the near infrared every 1-2 minutes continuously from sunrise to sunset and is automated. The atmospheric solar absorption spectra are fitted using laboratory spectra of species to retrieve atmospheric column abundances of constituents that include CO2, CH4, N2O, H2O, HOD and CO. The LANL solar FTS is part of the global Total Column Carbon Observing Network (TCCON) network and uses stringent operational and retrieval protocols that allows us to achieve very accurate and precise observations (better than 0.25% or 1 ppm for CO2) that are necessary to evaluate regional carbon cycle mechanisms and models. TCCON is also crucial for the calibration of satellite observations of CO2 that are being made by Japan’s GOSAT and will be made by NASA’s OCO-2 starting in middle of 2014. Unfortunately, while TCCON has good coverage at high and mid latitudes, it has a critical gap in the tropics. Therefore our Manaus FTS deployment will have a very high impact on NASA and Japanese satellite products for global scale model evaluation and will fertilize inter-agency and international collaborations. Our observatory also has in situ sensors to measure local meteorology, and capability to monitor CO2, CH4, CO and H2O in situ at the site that would be valuable for separating boundary layer signals from the total column observations.