Modeling rates of ecosystem recovery after fires by using Landsat TM data

[摘要] The aim of this work consists of monitoring the recovery process after five by means of satellite imagery. The objectives are to assess the regrowth pathways followed by different species populations after a disturbance, to analyze the speed of recovery in the years following fire, and, finally, to estimate rates of regrowth. The test area is located in the north of the province of Alicante, on the Mediterranean coast of Spain. This area, especially prone to forest fires, shows a remarkable land-use history and human pressure. The test areas belong to different microclimatic zones, show diverse vegetation communities, and have different degrees of stoniness; so we attempted to discover their postfire behaviors according to their biogeographical conditions. To accomplish these objectives, we used nine Landsat 5 thematic mapper images from 1984 to 1994 to which geometric and radiometric corrections were applied. Once the comparability between images was guaranteed, we generated a normalized difference vegetation index (NDVI) for each date. First, we demonstrated that the differences between NDVI images were suitable for mapping burned areas. Second, we undertook a nonlinear regression analysis between NDVI values and the time elapsed since the fire to assess the recovery processes. The exponential adjustment between NDVI and time was in accord with the asymptotic behavior observed when the recovery process is complete. The parameters supplied by the proposed method are helpful in quantifying the effects of fire on different ecosystem processes. (C) Elsevier Science Inc., 1997.

[发布日期] 1997-09-01