GAS RELEASE BELOW BALBINA DAM. Alexandre Kemenes*, Bruce Rider Forsberg* & John Michael Melack # * Instituto Nacional de Pesquisas da Amazônia, INPA - CPEC - BADPI, C.P. 478, Manaus, AM, Brazil, cep: 69011-970. Correspondence and requests for material should be addressed to A.K. email: alekemenes@yahoo.com.br # Donald Bren Hall 4424, UCSB, Santa Barbara, California 93106-5131 USA Introduction As scientists and governments become increasingly concerned about the growing evidence of global warming, reducing the emissions of greenhouse gases is gradually emerging as a common goal. Hydroelectric power has been promoted as an environmentally clean energy source. However, this view has begun to change. Some hydroelectric reservoirs have been found to release more greenhouse gases (GHGs) per megawatt than analogous fossil fuel plants (1, 2). While most research has focused on gas emissions from reservoirs, it has been argued that emissions downstream from hydroelectric dams, resulting from the rapid depressurization of reservoir bottom waters, could be considerably higher (1, 2, 3). The water intake for most hydroelectric dams is located near the bottom of the reservoir to guarantee sufficient head pressure to operate the turbines. As gas-rich water passes through the turbines and is exposed to the atmosphere, the hydrostatic pressure drops immediately, and a large portion of the gas is released as bubbles. This occurs quickly, offering little opportunity for oxidizing bacteria to consume the methane. Here we present the results of detailed measurements of GHG emissions downstream of Balbina Dam, one of the largest hydroelectric power plants in the Brazilian Amazon. Materials and Methods To provide an accurate estimate of the total flux downstream from the dam and improve estimates of the total release of GHGs from the entire system, we made regular estimates of the two principle downstream emission components: 1) gas ebullition as the reservoir waters passed through the turbines and 2) diffusive gas losses in the river channel below the dam. The total initial downstream gas flux for both methane and CO 2 was estimated from the product of turbine discharge and the gas concentration at the turbine intake (TIF). Immediate gas ebullition upon passage through the turbines was estimated from the product of water discharge through the turbines and the difference in dissolved gas concentrations at the intake and outlet of the turbines. Diffusive gas emission in the downstream channel was measured at approximately 4000 m intervals along a 70 km reach immediately below the dam, using a drifting static chamber method. Methane and CO 2 concentrations were determined sequentially using a dual column gas chromatograph (4). Results Positive linear relationships were encountered between turbine discharge and both total initial downstream gas flux (CH 4 TIF = -275.2 + 0.97 [discharge], n= 8, R 2 = 0.76, p< 0.05; and CO 2 TIF = - 147.1 + 0.85 [discharge], n= 8, R 2 = 0.92, p< 0.05) (Fig. 1A) and immediate ebullitive gas emissions (CH 4 emission = -197.4 + 0.63 [discharge], n= 8, R 2 = 0.83, p< 0.05; and CO 2 emission = -145.2 + 0.56 [discharge], n= 8, R 2 = 0.87, p< 0.05) (Fig. 1B). These relationships were used together with daily discharge values to estimate the total annual TIF and 663 Proceedings of 8 ICSHMO, Foz do Iguaçu, Brazil, April 24-28, 2006, INPE, p. 663-667.