INTEGRATED MULTI-CRITERIA AGRICULTURAL LAND SUITABILITY ASSESSMENT IN TASHKENT REGION USING SENTINEL-2, RUSLE, AND AHP-BASED MODELING

Land suitability assessment is a critical component of sustainable agricultural planning, particularly in semi-arid regions exposed to climatic variability and soil degradation processes. This study presents an integrated multi-criteria evaluation framework for agricultural land suitability in the Tashkent region (Uzbekistan) using Sentinel-2 imagery, topographic, climatic, and erosion-related factors within a Google Earth Engine (GEE) environment. A total of 257 Sentinel-2 images (2023–2024 vegetation period) were processed to derive vegetation indices (NDVI, NDWI, EVI). Terrain characteristics were extracted from SRTM DEM, climatic variables from CHIRPS precipitation and MODIS LST datasets, and soil erosion risk was estimated using the RUSLE model. All criteria were normalized and integrated through a Weighted Linear Combination (WLC) approach based on Analytical Hierarchy Process (AHP) weights. The results indicate that 8,065.20 ha are highly suitable (S1), 595,143.66 ha moderately suitable (S2), 485,628.54 ha marginally suitable (S3), 101,540.48 ha conditionally unsuitable (N1), and no areas were classified as permanently unsuitable (N2). The proposed framework provides a reproducible and scalable spatial decision-support tool for sustainable agricultural land management.土地适宜性评价是可持续农业规划的关键环节，对于面临气候变异性与土壤退化问题的半干旱地区而言尤为重要。本研究以乌兹别克斯坦塔什干地区为研究区域，构建了一套集成式多准则评价框架以开展农业土地适宜性评价：研究依托谷歌地球引擎（Google Earth Engine，GEE）平台，整合哨兵-2号（Sentinel-2）遥感影像、地形、气候及侵蚀相关因子。研究共处理了2023—2024年植被生长期内的257景哨兵-2号影像，以提取归一化植被指数（NDVI）、归一化水体指数（NDWI）与增强型植被指数（EVI）。地形特征数据从航天飞机雷达地形测绘任务数字高程模型（SRTM DEM）中获取，气候变量来自气候灾害集团红外降水监测数据集（CHIRPS）降水数据与中分辨率成像光谱仪地表温度产品（MODIS LST），并通过修正通用土壤流失方程（RUSLE）模型估算土壤侵蚀风险。所有评价准则均完成归一化处理，并基于层次分析法（Analytical Hierarchy Process，AHP）确定权重，通过加权线性组合（Weighted Linear Combination，WLC）方法完成集成。结果表明，研究区内高度适宜（S1）的土地面积为8065.20公顷，中度适宜（S2）为595143.66公顷，勉强适宜（S3）为485628.54公顷，条件不适宜（N1）为101540.48公顷，未出现永久不适宜（N2）的土地。本研究提出的评价框架可为可持续农业土地管理提供一套可复现、可扩展的空间决策支持工具。