Urban Heat Island Effect assessment by predicting land surface temperatures: A Case study of Karachi
Due to rapid urbanization, the concrete cover in urban areas has scaled up significantly in recent years. This increased concrete cover is becoming a source of the augmentation in temperature owing to the phenomenon known as Urban Heat Island.
Considering that Karachi is a major metropolitan city and the industrial hub of Pakistan, the city is readily moving towards urbanization which is leading towards an increase in the concrete cover. The resulting UHI effect due to this concrete infrastructure has led to an upsurge in the temperature levels in recent years, especially in the summer season. The aforementioned facts have given rise to some serious issues like heatstroke, respiratory difficulties, heat exhaustion, and heat cramps. This has also contributed to higher air pollution and energy costs (for cooling etc.). Thus, it is important to predict the potential UHIs of the region to help plan the city growth accordingly to avoid further escalation of this real-world issue.
In this regard, the UHI prediction has previously been done using various models, but we are using a new approach here by developing a new model named the long short-term memory (LSTM) model intending to eventually increase the prediction accuracy. To predict the UHI, LST data has been found which is one of the prime factors affecting the UHI. For this purpose, remote sensing data from Advanced Spaceborne Thermal Emission and Reflection Radiometer Global Digital Elevation Model (ASTER GDEM) and Moderate-Resolution Imaging Spectroradiometer (MODIS) sensors have been used. Moreover, to train the LSTM model, the LST data of 2000-2020 is used coupled with the road density, elevation, and enhanced vegetation index as input parameters.
The results suggest that the MAPE and MAE values for January are 0.15 and 0.27 and for May are 0.13 and 0.29, respectively. The prediction error is 1k (within allowable limits) and supreme quality as the MODIS LST quality criteria allow pixels having LST error below or equal to ±2 K for studies related to SUHI. In this paper, about 80 % of the pixels lie below the allowed limit of ±2 K. Thus, it is concluded that the proposed LSTM can be utilized for predicting LST of various areas for UHI studies by using good quality LST images of the previous year. Therefore, it can also be of significant use in the planning phase of city development projects for areas with greater UHI effect issues.
(00:00) – How to import data into GIS
(01:01) – How to draw maps
(03:34) – How to add details to the map
(04:53) – How to add Legend to the map
#UrbanHeatIsland #UHI #LST #Karachi #LSTM #ArcGIS