Geospatial modelling integration in landscape character analysis the case of büyük menderes basin

This study aimed to develop three new methodological approaches to landscape character analysis. i) the production of landscape units by including hydrological factors, ii) the determination of landscape character areas by different clustering methods, and iii) the transfer of landscape character analysis results into land use planning with an innovative approach, in the example of suitable area analysis for Fig tree (Ficus carica L.). The Büyük Menderes Basin, Turkey's seventh largest basin, was selected as the study area due to its rich biodiversity resulting from its diverse climate and morphological characteristics. In addition, the regions with the highest production of fig trees, which are of economic importance to Turkey, are located within the boundaries of this basin. During the landscape character analysis process, in addition to traditional input data, water flow representing the hydrological structure was included in the process to produce landscape units, which are the smallest units representing the landscape. The relationship and connectivity between landscape units was established through surface flow.In addition, both categorical and numerical data were used to create landscape character areas with the help of different clustering methods, and the most homogeneous and representative landscape character areas were obtained using the DBSCAN and SOM (Self-Organizing Maps) clustering methods. Finally, suitable areas for fig cultivation were identified based on landscape types, and spatial changes in fig cultivation areas were assessed in the context of climate change. In the study, suitable areas for fig trees were determined using the Convolution Neural Network (CNN) model, one of the artificial neural networks. In general, plains and the foothills of the Aydın Mountains exhibited high suitability. However, when evaluating the Representative Concentration Pathways (RCP) 4.5 and 8.5 scenarios used in the HadGEM2-AO Global Circulation Model (GCM) for the period 2050-2100 within the framework of climate change, it is projected that suitable areas will decrease across the basin. Soil and climate variables were found to be statistically significant in terms of growing conditions. Additionally, the Emberger bioclimatic classification, which has been proven to be related to Mediterranean vegetation, was included in the study. The findings can help predict the effects of climate change on fig trees and provide a basis for making predictions about fig production at local and regional scales, as well as developing climate-compatible agricultural application strategies.