Urban green space spatio-temporal change influences on land surface temperature in Kuala Lumpur, Malaysia

Urban green space (UGS) is a nature-like environment established in the urban structure of a city. It plays a vital role in providing vegetation cover to provide shade and act as a natural cooling eco-system to reduce the city’s heat by releasing oxygen for sustaining a healthy ecological environment. However, given the developments brought about by urbanisation, UGS has been sacrificed to allow for the urban growth activity. The continual development of new construction, road networks and buildings has eradicated UGS areas thus contributing to the rising of land surface temperature (LST). Accordingly, this study aims to monitor the UGS changes and LST pattern in Kuala Lumpur (KL) for the past six years and to develop an automated prediction model of these scenario for the year 2025 via temporal and spatial variation, using high-resolution aerial imagery data supported by the use of advanced technology mapping. The research utilised high-resolution aerial imagery for 2014, 2016, and 2019 that firstly used to map the spatial-temporal evolution of UGS over the past six years and to examine the UGS loss within the boundary of KL city. Secondly, to assess the pattern of LST change for the past six years and investigating the correlation between UGS changes and the effect on the LST. Thirdly, to develop an automated spatial prediction model that could potentially predict the UGS changes and their effect on the LST pattern. This research also tested the suitability of object-based classification methods of high-resolution aerial imagery using the support vector machine (SVM) classifier regarding its capability to correctly classify and recognise UGS patterns. The study also applied land surface emissivity (LSE) algorithm to determine the LST value extracted from the Band 10 parameter of Landsat 8 OLI/TIRS. A linear regression technique was employed to investigate the correlation between both scenarios using spatial statistical analysis and further predicting the UGS pattern and LST gradient for 2025 using the Artificial Neural Network - Cellular Automaton (ANN-CA) model. This model confidently predicted these scenarios logically, in which the expansion of built-up areas (BUA) in KL for following six years increased by body areas (WBA) slightly decreased by 4.57%. This led to an increase in the mean LST gradient for 2025 (32.15°C, which was about 3.22°C higher than the value recorded in 2019 (28.93°C). The prediction model employed in this study provides a significant benefit in monitoring the UGS changes and impact on the LST pattern for the past, present and future scenarios. The new automated model utilising highresolution aerial imagery has great potential to assist city planners and professionals in extracting, updating and detecting land use changes, particularly for UGS by applying a comprehensive procedure through a geographical information system (GIS) platform. The broad range of output generated from the multiple temporal of high-resolution aerial imagery could henceforth improve the reliability of collected data and develop a high-performance outcome in interpreting real visualised scenarios.11.62%, the UGS decreased by 28.88%, and water body areas (WBA) slightly decreased by 4.57%. This led to an increase in the mean LST gradient for 2025 (32.15°C, which was about 3.22°C higher than the value recorded in 2019 (28.93°C). The prediction model employed in this study provides a significant benefit in monitoring the UGS changes and impact on the LST pattern for the past, present and future scenarios. The new automated model utilising highresolution aerial imagery has great potential to assist city planners and professionals in extracting, updating and detecting land use changes, particularly for UGS by applying a comprehensive procedure through a geographical information system (GIS) platform. The broad range of output generated from the multiple temporal of high-resolution aerial imagery could henceforth improve the reliability of collected data and develop a high-performance outcome in interpreting real visualised scenarios.

Text FRSB 2021 11 IR.pdf Download (1MB)

Actions (login required)

View Item