Anuário do Instituto de Geociências (Jan 2019)
Spatial Transformations and Impacts on the Land-Surface Temperature in the Rio de Janeiro Petrochemical Complex (COMPERJ) Area
Abstract
The present work studies the land-surface temperature (LST) variation associated with the land cover change, supported by normalized difference vegetation (NDVI) and based built-up (IBI) indices for the COMPERJ installation region. This study is based on remote sensing data, namely image time series from the Landsat-5, 7 and 8 satellites, for the years 2002 to 2013, which as divided into two groups: pre-construction (2002 to 2007) and post-construction (2008 to 2013) period, totalizing 12 years with 58 images available. Two images were selected as the basis for making two land cover maps, one for each period, taking into account urban, soil, vegetation, and water land cover classes. Subsequently, LST was correlated with each index, exhibiting an inverse (direct) relation with NDVI (IBI). The highest correlation for the post-period (R = 0.85) was found with IBI, suggesting that this index could work as an additional metric to analyze the urban thermal patterns. An LST, NDVI, and IBI characterization (average and standard deviation) was performed in accordance with each class for both pre and post-period. The LST difference between COMPERJ (urban class) and vegetated areas was 1.9°C (3.2°C) for the pre-period (post). The comparison between these two values shows an intensification of 1.3°C for the postperiod. The IBI (NDVI) has increased (decreased) from -0.16 to 0.05 (0.49 to 0.15). In addition, the separability index M was used to analyze the parameters distinction between the classes. All parameters have shown M < 1 for the pre-construction period whereas all parameters have M values near to 1 for the post-construction period.
Keywords