Spatial modeling of an urban Brazilian heat island in a tropical continental climate

Abstract

Because climate aspects should be an important consideration when structuring urban spaces, this paper reports the results of combined field measurements and surface sensing data analyzed using statistical calculations. This was undertaken with the goal of detecting intra-urban air temperature patterns in a Brazilian city with a tropical continental climate (Paranavaí, Paraná State) through spatial modeling of an urban heat island. The air temperature data used were collected through mobile transects and used with Landsat 8 satellite as well as SRTM radar images for direct correlations of surface features (land use, NDVI, hypsometry and surface temperature). The model was executed in the IDRISI program based on linear regression and a final multi-criteria regression. In analyzing the heat island effect in Paranavaí, significant differences were found. The municipal public policy implemented in 2005 that chiefly addresses urban tree planting is assumed to have positive effects and can be considered as an example for other tropical cities. Based on our collective analyses and assessment, the modeling results lead to the conclusion that satisfactory results can be obtained by associating measured field data with surface characteristics, which represents an advance in the spatialization of local climatic data.