Imagens EOS-MODIS e LANDSAT 5 TM no estudo da dinâmica das comunidades de macrófitas na várzea amazônica / EOS-MODIS and LANDSAT 5 images for the study of macrophyte community dinamics in the Amazon varzea

AUTOR(ES)
DATA DE PUBLICAÇÃO

2004

RESUMO

Wetlands encompass about 4 to 6% of the terrestrial surface, occurring in all climates, from the tropics to the tundra, in all continents, and exhibit considerable function in the global carbon cycle, being responsible for up to 72% of the global total methane emissions by non-anthropogenic sources. The Amazon river floodplain is one of the most extensive areas of wetlands in the planet, covering more than 300,000 km2. Due to its great extension, this region is responsible for a significant contribution to the global methane flux to the atmosphere. Currently, more accurate estimates of emission are limited by the uncertainties resulting of little knowledge concerning the seasonal and interanual dynamics of the flooded areas. It is known, however, that areas of aquatic plant (macrophytes) growth present some of the highest values of methane emission. Remote sensing data represents one of the most adequate tools for the study of macrophyte community dynamics in the Amazon floodplain. In this study the use of MODIS and TM sensor data to monitor spatial and temporal changes in the macrophyte cover of the Amazon Lakes is examined. Moreover, a methodology for image processing and integration of remote sensing data into ecological models for methane emission is developed and tested.. For TM images, the application of a pre-classification step, incorporating altimetry and vegetation indexes information, to delineate areas of interest, coupled with a minimum distance classifier algorithm have yielded the best classification results. Regarding MODIS images, the use of image restoration techniques together with the previous procedure allowed classification results very close to those provided by TM images. The estimates of the area occupied by macrophyte stands at minimum and maximum water level periods have shown significant differences between dates. This variation, when related to in situ biomass measurements and published methane emission data, reveal the importance of further studies on the spatio-temporal dynamics of Amazonian aquatic plant communities. The results presented here demonstrate the capability of MODIS data for estimating interseasonal changes in the area occupied by macrophyte stands, at regional scale, in the Amazon floodplains, and opens up new possibilities on the study of such plant communities through orbital data.

ASSUNTO(S)

image classification metano accuracy carbon cycle aquatic plants classificação de imagens Áreas alagáveis ciclo do carbono variação anual modis methane plantas aquáticas landsat 5 remote sensing acurácia amazon region sensoriamento remoto landsat 5 annual variation wetlands região amazônica modis

ACESSO AO ARTIGO

Documentos Relacionados