Produção Nacional
Revisado por pares
CO 2 emissions from a tropical hydroelectric reservoir (Balbina, Brazil)
2011; American Geophysical Union; Volume: 116; Issue: G3 Linguagem: Inglês
10.1029/2010jg001465
ISSN2156-2202
AutoresAlexandre Kemenes, Bruce R. Forsberg, John M. Mélack,
Tópico(s)Marine and coastal ecosystems
ResumoJournal of Geophysical Research: BiogeosciencesVolume 116, Issue G3 Free Access CO2 emissions from a tropical hydroelectric reservoir (Balbina, Brazil) Alexandre Kemenes, Alexandre Kemenes [email protected] Instituto Nacional de Pesquisas da Amazonia, Manaus, Brazil Now at Embrapa Meio-Norte, Teresina, Brazil.Search for more papers by this authorBruce R. Forsberg, Bruce R. Forsberg Instituto Nacional de Pesquisas da Amazonia, Manaus, BrazilSearch for more papers by this authorJohn M. Melack, John M. Melack Bren School of Environmental Science and Management, University of California, Santa Barbara, California, USASearch for more papers by this author Alexandre Kemenes, Alexandre Kemenes [email protected] Instituto Nacional de Pesquisas da Amazonia, Manaus, Brazil Now at Embrapa Meio-Norte, Teresina, Brazil.Search for more papers by this authorBruce R. Forsberg, Bruce R. Forsberg Instituto Nacional de Pesquisas da Amazonia, Manaus, BrazilSearch for more papers by this authorJohn M. Melack, John M. Melack Bren School of Environmental Science and Management, University of California, Santa Barbara, California, USASearch for more papers by this author First published: 21 July 2011 https://doi.org/10.1029/2010JG001465Citations: 104AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract [1] Hydroelectric reservoirs can release significant quantities of CO2, but very few results are available from the tropics. The objective of the present study was to estimate the emission of CO2 from the Balbina hydroelectric reservoir in the central Brazilian Amazon. Diffusive and ebullitive emissions were estimated at regular intervals, both above and below the dam, using a combination of static chambers and submerged funnels. Gas releases immediately below the dam were calculated as the difference between gas flux at the entrance and the outflow of the hydroelectric turbines. An inundation model derived from a bathymetric map and daily stage readings was used for spatial and temporal interpolation of reservoir emissions. Annual emissions of CO2, upstream and downstream of Balbina dam for 2005, were estimated as 2450 and 81 Gg C, respectively, for a total annual flux of 2531 Gg C. Upstream emissions were predominantly diffusive with only 0.02 Gg C yr−1 resulting from ebullition. On average, 51% of the downstream emission was released by degassing at the turbine outflow, and the remainder was lost by diffusion from the downstream river. The total annual greenhouse gas emission from Balbina dam, including the CO2 equivalent of previously estimated CH4 emissions, was 3 Tg C yr−1, equivalent to approximately 50% of the CO2 emissions derived from the burning of fossil fuels in the Brazilian metropolis of São Paulo. Key Points Annual CO2 emissions from Balbina Dam in the Brazilian Amazon were quantified Emissions below the dam were significant Reservoirs could contibute significantly to anthropogenic CO2 emissions 1. Introduction [2] Increases in the concentrations of CO2 and other greenhouse gases (GHGs) in the Earth's atmosphere during the last century have resulted in a rise in average air temperatures and could lead to major environmental and socio-economic impacts if these trends continue [Intergovernmental Panel on Climate Change, 2007; Hansen et al., 2006]. The impoundment of rivers for hydroelectric power generation has contributed to these changes by converting terrestrial ecosystems into wetlands and transforming impounded and inflowing terrestrial and aquatic organic matter into biogenic gas emissions. It has been estimated that emissions from reservoir surfaces represent approximately 4% of CO2 released from all anthropogenic sources, but very few tropical reservoirs were included in this assessment [Saint Louis et al., 2002]. Ebullitive gas release immediately below hydroelectric turbines, due to the rapid depressurization of reservoir waters, and diffusive fluxes further downstream have been shown recently to contribute significantly to GHG emissions from tropical reservoirs [Abril et al., 2005; Guérin et al., 2006; Kemenes et al., 2007]. Few studies have considered all potential components of gas flux from hydroelectric systems, and information on these fluxes is especially important in tropical regions where many new dams are expected to be built. [3] Recent syntheses of the role of inland waters in the global carbon cycle have identified significant processing of organic carbon in lakes, wetlands and rivers [Cole et al., 2007; Tranvik et al., 2009; Aufdenkampe et al., 2011]. A portion of the organic carbon fixed by photosynthesis in these systems and imported from uplands in their catchments, is sequested in aquatic sediments, a portion is exported via rivers and a large fraction is often released to the atmosphere. As these functions of inland waters have become increasing apparent, the paucity of relevant information from tropical systems is a serious hiatus in light of the high rates of autotrophic and heterotrophic processes expected in these warm regions. [4] Balbina is one of the largest reservoirs in tropical South America. Due to its low energy yield per unit of flooded area (∼0.14 W m−2), Balbina generates a large amount of greenhouse gases with a relatively small energy return [Fearnside, 1989]. Regionally significant CH4 emissions have been measured both upstream and downstream from Balbina dam [Kemenes et al., 2007]. Diffusive losses of CO2 were also reported below the dam by Guérin et al. [2006], based on results from one sampling. Here, we present results of a detailed investigation of CO2 emissions from the Balbina hydroelectric system, which include year-round measurements of CO2 emission from multiple sampling stations, both upstream and downstream from the dam. The total annual emission of CO2 from the Balbina system is estimated, together with the CO2 equivalent of emissions of both CO2 and CH4. Potential errors in these emission estimates are discussed, as is with the influence of preimpoundment emissions or sequestration on the emission estimate. Our results contribute to understanding of impacts of reservoirs on GHG emissions and, more generally, to the role of tropical waters in carbon processing. 2. Methods Study Area [5] Balbina reservoir was formed in 1987 by impounding the Uatumã River in the central Amazon basin. The flooded area was dominated by tropical forest growing on the dissected Guianense Complex Formation [RADAM BRASIL, 1978]. This reach of the Uatumã River had a narrow floodplain and no significant alluvial wetlands, and no significant interfluvial wetlands within the drainage [RADAM BRASIL, 1978]. Balbina is the second largest Amazonian reservoir with an average flooded area of 1770 km2, an average depth of 10 m and an average water residence time of ∼12 months [Kemenes, 2006]. During the study period, water level varied approximately 2 m, which resulted in a surface area change of about 220 km2 between high and low water periods. Balbina dam has an average discharge of 570 m3 s−1 and an installed generating capacity of 250 MW [Fearnside, 1989]. The average depth of the turbine intake is about 30 m. Meteorological conditions varied seasonally at the reservoir with the highest rainfall and wind speeds occurring between December and May and the lowest occurring between June and November (Manaus Energia, unpublished data, 2010, from a meteorological station installed on the dam, equipped with a three-cup anemometer (Campbell Scientific, Inc.) and Ville de Paris rain gauge). Average rainfall and wind speed were 13.8 mm d−1 and 12.3 km h−1 in the rainy period (December–May) and 7.6 mm d−1 and 6.4 km h−1 during the drier season (June–November), respectively. Limnological Methods [6] Limnological characteristics were monitored at 10–14 points on the reservoir (Figure 1) and in the Uatumã River, immediately below the dam, at approximately monthly intervals between September 2004 and February 2006. Measurements on the reservoir were made in conjunction with gas flux estimates at points in open water and inundated forests. Surface and near bottom temperatures, conductivity and dissolved oxygen concentrations were measured with a YSI Model 85 m at all sites. Secchi depth transparency was determined with a 20 cm diameter white disk. Profiles of dissolved oxygen and temperature were obtained just upstream from the turbine intake. Measurements in the Uatumã River were made at the surface near the turbine outflow. pH was measured in situ with an Orion model 250a field pH meter. Water samples for dissolved organic carbon (DOC) were filtered through precombusted (400°C for 12 h) Whatman GFF glass fiber filters and stored in precombusted glass bottles with Teflon cap liners and 10 mg L−1 HgCl2 as a preservative. DOC concentrations were determined with a Shimadzu 500 TOC analyzer. Daily values of turbine discharge, reservoir stage height, rainfall and wind speed were provided by Manaus Energia. Figure 1Open in figure viewerPowerPoint Balbina Reservoir showing location of the dam and reservoir sampling points (solid circles). Extent of flooded area derived from high water L band synthetic aperture radar image acquired by the JERS-1 satellite in 1995. Gas Fluxes [7] To determine CO2 emissions from the Balbina hydroelectric system, we made regular measurements of surface emissions from the reservoir, gas discharge through the turbines, degassing at the turbine outflow and diffusive emissions in the river channel below the dam. 2.3.1. Reservoir Emissions [8] Surface emissions from Balbina reservoir were estimated at approximately monthly intervals at 10–14 points between September 2004 and February 2006 (Figure 1). Emissions were measured with static chambers and inverted funnels. The static chambers measured both diffusive and ebullitive fluxes, and the results derived from these measurements are referred to as total emissions. The funnels measured only ebullitive fluxes. The static chamber consisted of a floating cylinder covered with reflective thermal insulation and equipped with an internal fan to improve circulation and maintain ambient temperature. It was allowed to drift freely during deployment. The chamber had 0.23 m2 of open area in contact with the water surface and a small diameter connecting tube for withdrawing gas samples. The lower 3 cm of the vertical chamber wall was continually submersed during deployment. During each flux measurement, four gas samples were collected from the chamber at 5 min intervals with 60 ml polyethylene syringes and stored in 20 ml glass serum vials with high density butyl rubber stoppers until analysis [Devol et al., 1990]. All measurements were made in duplicate. The inverted funnels, used to estimate ebullition, had 30 cm diameter mouths and a fitting at the narrow end for collecting gas samples. Funnels were incubated underwater at all sample sites for 24 h. Gas samples collected from the funnels were stored in 20 ml glass serum vials until analysis. Near-surface and near-bottom water samples were collected at each sampling site for determining dissolved gas concentrations. Samples were collected near the bottom with a Ruttner sampler and subsampled at the surface with a 60 ml syringe. Near-surface samples were collected directly with a syringe. Dissolved gas concentrations were determined with the headspace method by equilibrating equal volumes of water and air in the sampling syringe [Johnson et al., 1990]. The equilibrated air samples were transferred to serum vials and stored until analysis. CO2 concentrations were determined using a dual column gas chromatograph (Shimadzu, model GC-14A) and methodology described by Hamilton et al. [1995]. Standards of 335 mg L−1 (SD = 105) and 995 mg L−1 (SD = 104) were used to calibrate the analyses. The detection limit was 100 mg L−1. Chamber emissions for each deployment were estimated from the regression of CO2 concentration against time. All regressions had r2 > 0.90, indicating that emissions were predominantly diffusive. [9] Aquatic habitats in the central Amazon basin, including Balbina reservoir, were classified by Hess et al. [2003] using 100 m resolution JERS-1 L band synthetic aperture radar mosaics acquired during high and low water periods. A modification of the procedure developed by Hess et al. [2003] was used by Kemenes [2006] to distinguish open water and dead forests with varying extent of inundation in Balbina reservoir. Four inundated habitat classes were defined: open water (OW), nearly submerged forest (NSF), moderately inundated forest (MIF), and slightly inundated forest (SIF). A bathymetric map and inundation model were derived from a geo-referenced bathymetric survey conducted with a Lowrance Inc. LMS - 320 recording sonar. Approximately 3000 km of sonar transects were surveyed covering all parts of the lake basin which were accessible by small boat at high water. Depth measurements were collected at 30 m intervals along these transects, yielding about 100,000 georeferenced data points. All depth values were normalized to peak water level (50 m stage height). The perimeter of the lake at peak stage height was delineated from a wetland mask developed by Hess et al. [2003] for the central Amazon basin. The data were interpolated spatially (Spatial Analyst, ESRI, Inc.) to obtain a digital bathymetric map for the lake. The digital bathymetric map was used to develop a nonlinear regression model between reservoir stage height and surface area which was utilized to calculate seasonal variations in reservoir surface area. The influences of depth, temperature, dissolved oxygen, transparency, wind velocity, rainfall, water level change and habitat type on gas fluxes were investigated using an ANCOVA [Sneath and Sokal, 1973]. 2.3.2. Degassing at the Turbine Outflow [10] Degassing immediately below the turbine outflow was estimated as the difference between the gas fluxes at the inflow and immediately below the outflow of the turbines. The gas fluxes at the turbine intake and outflow were calculated as the product of water discharge and gas concentrations measured at the turbine intake and just below the outflow, respectively. Water samples near the turbine inflow were collected at a depth of ∼30 m immediately upstream of the dam using a Ruttner sampler that was subsampled at the surface with a 60 ml syringe. Samples were also collected at 1 m intervals above this point to describe the vertical profile of CO2 concentrations. Water samples at the turbine outflow were collected at a depth of ∼0.5 m within 50 m of the dam with a 60 ml syringe. Gas concentrations in these samples were determined by the headspace method described above. The influences of turbine discharge, rainfall and wind velocity on gas flux upstream and downstream of the dam and on turbine emissions were evaluated in separate stepwise multiple linear regression analyses. Only significant variables (p < 0.05) were included in the final regression equations used to estimate daily flux rates. Due to the rapid drop in hydrostastatic pressure at the turbine outflow and the large rapid drop in CO2 concentrations (∼52%) observed at this point, turbine emissions were assumed to be predominantly ebullitive. 2.3.3. Sampler Comparison [11] We developed a new sampler, called the Kemenes sampler, designed to avoid degassing losses during the collection of deepwater samples. The sampler consists of a weighted housing which secures a 60 ml polyethylene syringe in a vertical position. Before lowering, a solenoid valve at the mouth of the syringe is closed and a vacuum is applied to the syringe by pulling the syringe piston and securing it in an extended position. At the sample depth the solenoid valve is opened to collect the sample. The sampler and syringe are maintained in a vertical position during retrieval to ensure that any gas bubbles released during the process are retained in the syringe. At the surface, the syringe is treated as indicated in the headspace method described above, ensuring that both the equilibrated gas and any released bubbles are included in the final gas sample. [12] To quantify the degassing error, we made 11 parallel collections of hypolimnetic waters near the turbine inflows of Balbina, Tucuruí and Samuel reservoirs, using both the Ruttner and Kemenes samplers. The samples were collected between September 2004 and February 2006, following the methodology described above. For depths greater than 20 m, which included most of the depth range expected to contribute to turbine flow in these three systems, the CO2 and CH4 concentrations in water samples collected with the Kemenes sampler averaged 34% and 116% higher than those collected with the Ruttner sampler, respectively. The greater difference encountered for CH4 reflects the lower solubility of this gas and greater tendency to form bubbles when hydrostatic pressure is reduced. 2.3.4. Downstream Diffusive Emissions [13] After the large ebullitive loss at the turbine outflow, no further evidence of ebullitive emissions was encountered downstream from the dam; hence, surface fluxes estimated with static chambers in the river channel below this point were assumed to represent diffusive emissions. Annual average gas concentrations and diffusive emissions fell gradually from 2450 to 1610 mg C m−3 and 417 mg C m−2 h−1 to 215 mg C m−2 h−1, respectively, until approximately 30 km below the dam after which they dropped only about 50 mg C m−3 and 30 mg C m−2 h−1, respectively, for the next 40 km. The drop in concentrations during the first 30 km was assumed to represent the degassing of excess CO2 derived from the dam while the stable value encounter below this point represented the balance between in stream respiration and diffusive emissions. Eight downstream measurement campaigns were done between November 2004 and November 2005. Measurements were made at logarithmically increasing intervals downstream from the dam to represent the large initial decline in values observed immediately below the dam. Diffusive emissions were measured with a drifting static chamber. Surface water samples for dissolved gases were collected with a 60 ml syringe and processed using the headspace method described above. The daily emission for each subreach (area between sampling points) was estimated from the product of the subreach area and the average areal emission rate along the subreach. The values for all subreaches were then summed to estimate daily diffusive emission for the entire reach between the dam and 30 km downstream (see discussion below). The average areal emission was estimated by dividing the total diffusive emission by the total surface area of the study reach. No significant relationships were found between daily emission and environmental variables which could be used for temporal interpolation. Hence, the total annual diffusive emission for the 30 km reach was calculated from the average of the 8 daily estimates times 365. The residual gas flux at 30 km was estimated as the product of water discharge and the dissolved CO2 concentration measured at that distance. 3. Results Limnological Characteristics [14] Average near-surface water temperature in the reservoir varied from 28°C during the rainy season to 30°C during the drier season. The reservoir was thermally stratified most of the year with an oxycline located around 8 m and a hypoxic hypolimnion (Figure 2). Near-surface dissolved oxygen concentrations ranged from 4 to 7 mg l−1 while near-bottom concentrations varied between 0.5 and 4 mg l−1. Secchi disk transparency in the reservoir varied from 1 to 3 m. Average turbine discharge at the dam ranged from 694 m3 s−1 during the rainy season to 433 m3 s−1 during the drier season. Conductivity in the Uatumã River downstream from the dam varied from 6 to 10 μS cm−1, pH ranged from 6.0 to 7.4, dissolved organic carbon concentrations varied from 2.7 to 7.1 mg L−1 and dissolved oxygen ranged from 4.5 to 7 mg L−1. Figure 2Open in figure viewerPowerPoint Average vertical profiles of CH4 and CO2 concentrations, temperature and dissolved oxygen for rainy and drier seasons at Balbina Reservoir, immediately upstream of the turbines. CO2 Distributions in the Reservoir [15] The observed ranges of gas concentrations and emission in Balbina reservoir are summarized in Table 1. All variables had normal distributions (Kolmogorov-Smirnov test, p < 0.05) and means are arithmetic averages. Concentrations of dissolved CO2 ranged from 42 to 180 μM at the surface and from 52 to 375 μM near the bottom with average values of 99 and 161 μM, respectively. During the rainy season, the reservoir was weakly stratified with multiple thermoclines above 20 m, and CO2 concentrations were relatively low and constant in this region (Figure 2). Below 20 m, thermal stratification was stronger and CO2 concentrations increased with depth. During the drier season, the water column was strongly stratified, and CO2 concentrations in hypolimnetic waters were higher than those encountered in the rainy season. Table 1. Range of CO2 Concentrations and Emissions Measured at Balbina Reservoir Between September 2004 and February 2006a Maximum Minimum Superficial concentration 180 42 Near-bottom concentration 608 128 Total emission (chamber) 8500 343 Diffusive emission 8500 343 Ebullitive emission 1.3 0 a Here n = 72. Emissions are expressed as mg C m−2 d−1, and concentrations are expressed as μM. CO2 Fluxes From Reservoir [16] Total emission rates of CO2 (chamber measurements) from the reservoir surface varied from 343 to 8529 mg C m−2 d−1 with an average value of 3776 mg C m−2 d−1 and were due almost entirely to diffusive fluxes (Table 1). Ebullitive emissions accounted, on average, for less than 0.02% of the total flux and were not considered separately in the annual flux estimates. Inundated habitats varied seasonally with MIF (37%) predominating at low water, followed by SIF (36%), NSF (24%) and OW (3%), and NSF (50%) predominating at high water, followed by MIF (32%), OW (9%) and SIF (9%). However, variations in habitat type and other environmental parameters had no significant effects on ebullitive emission rates (ANCOVA; F14,56 = 0.5, p = 0.50). An ANCOVA of these same variables against total emission rate was significant (F14,56 = 3.2, p < 0.05), but only due to surface and bottom dissolved oxygen which both had a strong negative effect on flux. Integrated Reservoir Emissions [17] Since oxygen values were measured only occasionally, the statistical relationship between dissolved oxygen and emission could not be used to interpolate total emissions in time. An inundation model and average monthly emission rates were therefore used for this purpose. A quadratic regression model (F = −0.088 S4 + 17.8 S3 − 1321 S2 + 4296 S − 517590), relating total flooded area (F) to daily reservoir stage height (S), was derived from the bathymetric data normalized to stage height (Figure 3). Daily emission from the entire reservoir surface (Mg C d−1) was estimated by multiplying the total flooded area, estimated daily with this model, by the average monthly areal emission rate (Table 2). During most of year, daily reservoir emissions were positively correlated with flooded area. The few exceptions included unexpected peaks in CO2 emissions in July, November and December. The total annual emission from the reservoir surface was estimated by integrating daily emissions throughout the year, and during 2005 was estimated to be 2450 Gg C yr−1. Figure 3Open in figure viewerPowerPoint Relationship between lake water level and total flooded area in Balbina Reservoir (p < 0.05, r2 = 0.88, n = 44). Table 2. Average CO2 Emissions and Flooded Areas at Balbina Reservoir in 2005 Month Average Emission (mg C m−2 d−1) Average Flooded Area (km2) Jan 3570a 1350 Feb 2160 1400 Mar 2710a 1600 Apr 3260 1870 May 3440 2230 Jun 3770a 2320 Jul 4100 2320 Aug 3870 2220 Sep 2810 2070 Oct 3890a 1980 Nov 4980 1720 Dec 3570a 1620 a Emissions estimated from mean of adjacent monthly values. Degassing of CO2 at the Turbine Outflow [18] The average concentrations of CO2, measured above and below the turbines, were 378 (SD 153) and 215 (SD 66) μM, respectively. The daily discharge of CO2 at the turbine inflow (vCO2inflow) and outflow (vCO2outflow) were estimated by multiplying the upstream and downstream concentrations by the daily discharge of water through the turbines. Degassing at the turbine outflow was estimated by the difference between vCO2inflow and vCO2outflow. The average rate of degassing determined from the measured fluxes was 100 Mg C d−1. The degassing rate varied during the year with the lowest values occurring during the rainy period (December–May) and the highest levels occurring at the end of the drier season (June–November) (Figure 4a). Figure 4Open in figure viewerPowerPoint Seasonal variation of (a) initial degassing and (b) diffusive emissions of CO2 downstream of Balbina dam. Data collected between September 2004 and February 2006. Months without data are indicated with an asterisk. [19] To improve the estimate of the annual degassing, we explored correlations with environmental parameters for which we had daily values and found that the best relationships for estimating daily CO2 discharge at the turbine inflow (r2 = 0.48, n = 13, p < 0.05) and outflow (r2 = 0.42, n = 13, p < 0.05) were multiple regressions against daily wind speed and rainfall: These equations were used together with daily wind and rainfall data (Manaus Energia) to estimate the daily gas discharges at the turbine inflow and outflow and the resulting daily ebullitive emissions during 2005. Total annual CO2 discharge at the turbine inflow was estimated as 78 Gg C yr−1 while degassing at the turbine outflow was estimated to be 41 Gg C yr−1, 53% of the total flux through the turbines. CO2 Fluxes in the Uatumã River [20] The average surface concentration and real emission rate of CO2, measured along the 30 km study reach below the dam, were 161 μM and 4790 mg C m−2 d−1, respectively. Both of these values were higher than the average values encountered in the reservoir upstream of the dam. Gas concentrations and fluxes declined slowly below the dam reaching relatively constant values at the end of the 30 km reach (Figure 5). The daily diffusive emission, integrated along the 30 km study reach (Figure 4b), had little seasonal variation. No significant relationships were found between these daily fluxes and measured environmental parameters (ANCOVA, p < 0.05). The total annual diffusive emission for the 30 km reach was, therefore, calculated from the average of the 8 daily estimates multiplied by 365 days. The resulting annual flux was 40 Gg C yr−1. The residual discharge of dissolved CO2 at 30 km was calculated from the mean of 8 measured values, resulting in an annual flux of 41 Gg C yr−1 for 2005. The difference between the residual discharge of dissolved CO2 at the turbine outflow and the sum of the residual discharge at 30 km and the diffusive loss above this location provided an estimate of CO2 inputs to the river along the 30 km reach. The total residual discharge of dissolved CO2 at the turbine outflow in 2005 was estimated to be 37 Gg C yr−1, and the input of CO2 along the study reach was estimated at 44 Gg C yr−1. Figure 5Open in figure viewerPowerPoint Mean annual variation in surface concentration and diffusive emission of CO2 downstream of Balbina dam. Data collected between November 2004 and May 2005. Total CO2 Fluxes for the Balbina Hydroelectric System [21] The total annual CO2 emission for the Balbina hydroelectric system, including emissions from the reservoir, the turbine outflow and the river channel to 30 km below the dam for 2005 was 2531 Gg C yr−1 (Table 3). Only 3% of the total flux occurred downstream of the dam. The remaining 97% was released from the reservoir surface. Degassing at the turbine outflow accounted for 51% of the downstream release and the remainder was due to diffusion from the river surface. Table 3. CO2 Fluxes for the Balbina Dam System in 2005a Value Fluxes From Reservoir Total emission from reservoir surface 2450 Fluxes Below the Dam Degassing at the turbine outflow 41 Residual discharge at the turbine outflow 37 Diffusive emission from river channel for first 30 km reach below dam 40 Residual discharge at 30 km 41 Net gain over 30 km 44 Total emission below dam for 30 km 81 Total emission from Balbina system 2531 a All fluxes are in Gg C yr−1. 4. Discussion CO2 Fluxes From the Reservoir [22] The average daily CO2 emission rate from Balbina reservoir was close to the values reported for Tucuruí and Samuel reservoirs [Lima et al., 2002] but considerably higher than those cited for Curuá-Una [Duchemin et al., 2000] and Petit Saut [Abril et al., 2005] (Table 4). Balbina's value was also significantly higher than the average value of 960 mg C m−2 d−1 cited by Saint Louis et al. [2002] for CO2 in tropical
Referência(s)