Introduction: In recent years, lifestyle changes and urbanization of societies on the one hand and macro environmental changes such as climate change on the other hand have caused changes in the spatial structure and transfer of resources between different economic sectors and different regions as well. Due to the long-term and comprehensive nature of spatial development plans, they should be able to adapt to macro environmental changes and have the necessary flexibility. Spatial balance in the concept of adapting the supply and demand of the environment, as the ultimate goal and achievement of land development, is important that will be affected by these macro changes in the long term. The degree of compliance of an experimental program with these changes is an indicator of its compatibility with the changes, and it is necessary to measure this compatibility by using models and simulation methods of the effects of these changes on the environment.

Materials and methods: Machine Learning models including Cellular Automata, Random Forest and Regression Models in the form of PLUS model were used to simulate the land use map in 2000, 2010 and 2020 and finally predict the land use changes based on climate change scenarios (ICCP) for the future 25-year horizon for two planning areas of Kerman province, including Kerman-Bardsir and Baft-Rabar-Borzowieh planning areas. The 2010-2020 model with a kappa coefficient of 0.94 and overall accuracy of 0.98 was used as the selected model for this simulation. The overlap of changes in land use classes in four scenarios was done with the development plan of the agricultural sector in the study regions, and the share of driving forces of land use changes was also calculated. Finally, the ecosystem services supply and demand adaptation to climate change was modeled in the region.

Results and discussion: The evaluation results showed that the lands under human construction had the highest percentage of changes during the studied period. Due to the aridity of the region, there has not been an increase in cultivated land in line with the increase in population in this region. Also, the development potential map of each of the land cover classes in the study area shows significant changes in agricultural lands and man-made lands. Regarding the driving forces, depending on the type of land use, each of the driving factors showed a different contribution to the amount of changes. Altitude, rainfall, and temperature factors were the most important factors and forces affecting land use changes in the study area with coefficients of 0.45, 0.35, and 0.15, respectively. The modeling of land use changes in line with climate change scenarios showed that the designed plan did not comply with any of the possible scenarios, and in this respect, the designed plan has a spatial imbalance and as a result, future Spatial planning is not compatible with possible climate changes in this region. This resulted in the comparison of supply through production in agricultural lands and demand through the increase of human population in the region.

Conclusion: The integration of economic, social and environmental factors is necessary for predicting land uses in order to implement sustainable development, and machine learning models for simulating land cover changes and estimating environmental supply make it possible for predicting the land preparation program evaluated and created the necessary flexibility for different scenarios, especially in super dry and hot lands, and led to the goals of sustainable development in these areas through adaptation.