نوع مقاله : علمی - پژوهشی
نویسندگان
1 دانشجوی دکتری آبخیزداری، دانشگاه ارومیه، ارومیه، ایران.
2 دانشیار، گروه مرتع و آبخیزداری، دانشکده منابع طبیعی، دانشگاه ارومیه، ارومیه، ایران
3 محقق پسا دکتری، دانشگاه جزیره پرنس ادوارد، کانادا
چکیده
کلیدواژه‌ها
عنوان مقاله [English]
نویسندگان [English]
Background and objectives: Meteorological drought is one of the most disastrous types of drought, with extensive direct and indirect consequences for society. Its occurrence can also impact other natural phenomena. Iran, due to its geographical location within the arid and semi-arid belt, is highly prone to drought. Over the past three decades, the country has experienced frequent and severe droughts in its watersheds and forested areas. Predicting the exact temporal and spatial occurrence of drought-induced damages at both global and national scales remains highly challenging. However, research indicates that the impacts of meteorological drought are more pronounced in forested areas compared to watersheds. Identifying influential factors, measuring, forecasting, and defining a natural hazard such as drought is often complex and impractical due to its multifaceted nature. Therefore, implementing adaptive measures tailored to specific regions requires identifying key components, understanding complex interrelationships, and comprehensively assessing drought risks. This can be achieved by developing drought hazard and susceptibility maps.
Materials and Methods: Kurdistan province is geographically located between 34°44' to 36°30' N latitude and 45°31' to 48°16' E longitude from the Greenwich meridian. The drought occurrence distribution map was prepared using the Drought Severity Index (DSI) in the Google Earth Engine environment. Natural, anthropogenic, and morphometric factors contributing to drought occurrence were considered. A digital elevation model (DEM) was downloaded from the US Geological Survey (USGS) website to generate elevation, slope, and morphometric index layers. Precipitation, temperature, evapotranspiration, and climate data were collected from rain gauge, synoptic, and climatology stations across the province and interpolated using the Kriging method. The Normalized Difference Vegetation Index (NDVI) layer was extracted from Sentinel-2 images, while the soil map of Kurdistan province was obtained from the FAO soil database. The thematic layers were processed using ArcGIS and SAGA-GIS software. The drought susceptibility map was generated using the Bagging–Random Forest machine learning model, and the impact of each factor on drought occurrence was analyzed using the permutation index. The model’s performance was evaluated using the Receiver Operating Characteristic (ROC) curve in the R programming environment.
Findings: The model evaluation results, based on the ROC curve after generating the drought susceptibility map using the Bagging–Random Forest method, indicate that the model performs at an excellent level with an Area Under the Curve (AUC) value greater than 0.9. The analysis of the percentage of areas susceptible to drought within the study region reveals that approximately 60% of the watersheds and forested areas of Kurdistan province fall within moderate to very high vulnerability zones. The permutation analysis for the model indicates that climatic factors such as precipitation, temperature, climate, and evapotranspiration have the highest impact on drought occurrence, whereas the Normalized Difference Vegetation Index (NDVI) and mass balance index have the least influence.
Conclusion: The results of the Bagging–Random Forest model for Kurdistan province indicate that areas with high drought susceptibility are primarily located in the eastern and, to some extent, southern parts of the province. Since most of the high-quality forest and rangeland areas are concentrated in the western part of the province, these regions exhibit lower drought sensitivity compared to other areas. The evaluation of the model’s predicted map for the western part of the province, which contains valuable forest and rangeland ecosystems, indicates that forested areas are in a more critical condition compared to watersheds. Therefore, future planning and management strategies should prioritize these regions.
کلیدواژه‌ها [English]
)