Fatores que influem na emissão de CO2 e CH4 em áreas alagáveis interfluviais do médio Rio Negro

Abstract

Wetlands in the Amazon basin have been shown to be a globally significant source of methane and carbon dioxide, important greenhouse gases. However interfluvial savanna wetlands of Negro River basin have not been characterized. Here we describe the hydrometeorology and their influence on emissions of CO2 and CH4 from these wetlands to improve regional emission estimations. Water level and rain fall were daily measured in three wetlands and also in Aracá and Negro rivers. In two of these wetlands, one permanently flooded (RNL) and one seasonally flooded (Itu), oxygen profiles of water column, surficial temperature, depth and CO2 and CH4 fluxes were monthly measured from February 2005 to January 2006. Fluxes were measured by terrestrial chambers when soil was dry and by funnels and flooding chambers when it was flooded and also estimate by Fick´s low. To estimate emissions of both wetlands, inundation models were developed for these two wetlands using temporal sequences of C-band radar images taken during 2004 and 2005. The thirst wetland (Aracá), a savanna with limited access, was sampled only at July, August and November of 2005 and their CO2 and CH4 emissions were estimate by Fick´s low. Local, habitat, depth, variations in water table level, oxygen concentration and surface water temperature were related to diffusive emissions from tree sites and ebullitive emissions from RNL and Itu site. The existence of methane transport through grass parenchyma was investigated in these two wetlands using paired emission measurements with and without grasses and a paired T-test of Student. All wetlands were very shallow, with maximum depth of 1.8 m. Rain fall was higher in May and lesser in October in all wetlands. Wetland water table was influenced daily by rain fall and seasonally by water level of nearest river in all wetlands. Total area calculated for RNL field was 1685 km2 and for Itu field was 1295 km2. Yearly, about 52% of the area of both wetlands was inundated. Oxygen concentrations fell with depth but the hypolimnion was never anoxic. The average surface temperature was 28°C. The average ebullitive CO2 emission rate for RNL and Itu sites was 11 (±11) mg CO2 m-2d-1. Diffusive CO2 average for all tree sites was 7729 (±5198) mg CO2 m-2d-1. CO2 diffusive and ebullitive emissions were higher in the RNL site, and into witch site in higher density vegetation habitats. Bubble CH4 emissions were higher in RNL site and when water table was falling. There was diffusive emission of 26 mg CH4 m-2d-1 in wet environments and absorption of 5 mg CH4 m-2d-1 in dry environments. In Itu site, where emergent plants were totally rooted in soil, emission was higher in presence of emergent plants (pared T-test: n = 38; p = 0.0134), with an emission of CH4 thought parenchyma of 7 mg CH4 m-2d-1. Diffusive and ebullitive fluxes of CO2 and of CH4 were higher when dissolved oxygen in water was depleted (p < 0.05). Space temporal integration shows emissions were higher in May, when most of these wetlands areas were flooded. It was estimated Amazon interfluvial wetlands emit 56 Tg C y-1 de CO2 e de 1.5 Tg C y-1 de CH4. Total emission of methane per area was lower in interfluvial wetlands than in river floodplains. The permanent flooded field emitted 15 Mg C km-2y-1 and the seasonally flooded Field emit 4.1 Mg C km-2y-1de CH4. These results show it is necessary to recalculate methane emission estimated before for Amazon wetlands to correct overestimation.