Land use and land cover (LULC) change analysis is a systematic technique that aids in the comprehension of physical and non-physical interaction with the natural habitat and the pursuit of environmental sustainability. Research regarding LULC’s spatiotemporal changing patterns and the simulation of future scenarios offers a complete view of present and future development possibilities. To simulate the spatiotemporal change transition potential and future LULC simulation, we utilized multi-temporal remotely sensed big data from 1990 to 2020 with a 10-year interval. Independent variables (DEM, slope, and distance from roads) and an integrated CA-ANN methodology within the MOLUSCE plugin of QGIS were utilized. The findings reveal that physical and socioeconomic driving variables have a substantial effect on the patterns of the terrain. In the last three decades, the study area had a significant rise in impervious surface from 10.48% to 26.91%, as well as a minor increase in water from 1.30% to 1.67%. As a result, forest cover decreased from 12.60% to 8.74%, green space decreased from 26.34% to 16.57%, and barren land decreased from 49.28% to 46.11%. Additionally, the predictions (2030–2050) support the increasing trend towards impervious surface at the expense of significant quantities of forest and green space.

Frequent gradient exchange and heterogeneous data distribution in federated learning can lead to serious privacy leakage risks. Traditional privacy-preserving strategies fail to meet the personalized privacy needs from different users and may cause a decrease in model accuracy and convergence difficulties. The symmetry of federated learning may lead to the insufficiency of contribution evaluation mechanisms in protecting the privacy of sensitive data holders. However, federated learning avoids the risk of privacy leakage caused by data centralization because the raw data is always stored on the local device during the training process, and only encrypted model parameters or gradient updates are exchanged. To address these issues, this paper proposes an adaptive personalized differential privacy federated learning scheme APDP-FL. First, we propose an adaptive noise addition method that scores each round of training based on the parameters generated during training and dynamically adjusts the noise level for the next round. This method adds larger noise scales in the early stages of training, consuming less privacy budget, and gradually reduces noise addition during training to accelerate model convergence. Second, we design a personalized privacy protection strategy that adds noise tailored to individual needs for participating clients based on their privacy preferences. This solves the problem of insufficient or excessive privacy protection for some participants due to identical privacy budget sets for all clients, achieving personalized privacy protection for clients. Finally, we conduct extensive experimental simulations, comparisons, and analyses on three real federated datasets, MNIST, FMNIST, and CIFAR-10, verifying the advantages of APDP-FL in terms of privacy protection, model accuracy, and convergence speed.