Urban Evolution Predictions: Cellular Automata-Based Land Cover Change Analysis in Surabaya (2015-2021)

This study investigates the dynamic transformation of Surabaya's urban environment from 2015 to 2021 by employing satellite imagery and Cellular Automata (CA) simulations to predict alterations in land cover. This study provides essential perspectives for the fields of urban planning and environmental management, emphasizing significant urban expansion and its consequences, thus providing direction for sustainable development approaches in Surabaya. The study centers on critical components such as road networks, rivers, and the Central Business District through the analysis of land cover changes and the prediction of forthcoming shifts utilizing the CA model. The Random Forest algorithm was employed to analyze data obtained from Sentinel-2 satellite imagery on the Google Earth Engine cloud platform. The results indicate that Surabaya witnessed a decline of 40.60% in rice fields (6680.806 hectares) during the period from 2015 to 2021. In contrast, the city witnessed an expansion of 31.98% in buildings (5262.554 hectares) and 21.83% in roads (3592.240 hectares). The utilization of cellular automata to implement Artificial Neural Networks (ANN) to simulate land use change yielded a model that achieved an accuracy of 63.036%, as indicated by a kappa value of 0.49031. Water bodies are the land cover class most prone to change, as indicated by a probability value of 0.369. It is anticipated that structured land cover will predominate by 2030, comprising 57.774% of the total land area or 19373.40 hectares. Furthermore, under Surabaya's RDTR (Detailed Spatial Plan), the projected aggregate land utilization for 2030 is 46.10% (0.461), whereas non-compliant utilization accounts for 53.89% (0.5389).