Public Urban Park Quality Assessment Using Fuzzy C Means Classification of Land Surface Temperature and Social Function

Abstract

AbstractThis study describes the process to assess public urban park (PUP) quality in Bogor Municipality, Indonesia, using social function key performance indicator (KPI) and land surface temperature (LST) from Landsat 8. Eighteen PUPs located in high LST were analyzed to observe their social function based on KPI values. Three fuzzy linguistic classes of LST have been generated from the fuzzy c means (FCM) classification technique. A method called fuzzy class separation analysis is proposed in this study to select the most acceptable fuzzy separation among three fuzzy membership values ranges. This fuzzy class separation analysis used Jaccard Similarity Index (JSI) to identify the least intersected raster cells from three fuzzy images, namely (a) > 0.7–1, (b) > 0.75–1, and (c) > 0.8–1. The result from JSI shows that fuzzy LST image with > 0.75–1 membership value has the smallest number of intersected cells, and therefore high LST raster image was extracted from > 0.75 to 1. Within images from > 0.75 to 1 fuzzy membership, 30.99% raster cells were classified as high LST, 37.14% as medium LST, 27.27% as low LST, and 4.60% unclassified. High LST image was integrated with PUPs vector datasets using the extraction menu in ArcMap 10.3 and yielded in eighteen PUP locations. Based on social function KPI values in each PUP, seven PUP locations within high LST can be classified as compatible with KPI standard value (0.68–1), while eleven locations can be associated with less compatible (0.34–0.67). Therefore, seven PUP locations with compatible standards need proper maintenance to keep their social function quality. In contrast, the other eleven locations with less compatible standards need improvement to increase their social function quality.KeywordsPublic urban parkFuzzy c meansLand surface temperatureSocial functionKey performance indicator