KARA DERE HAVZASININ (ERZİNCAN) ANALİTİK HİYERARŞİ SÜRECİ METODUYLA HEYELAN DUYARLILIK ANALİZİ


Creative Commons License

YANIK M. E., Kopar İ.

Black Sea Journal of Engineering and Science, cilt.9, sa.4, ss.1954-1969, 2026 (TRDizin)

Özet

Doğa kaynaklı afetlerin başında gelen heyelanlar, can ve mal kayıplarına neden olan önemli bir kütle hareketidir. Bu çalışmanın amacı, Erzincan Ovası'nın kuzeybatısında yer alan Kara Dere Havzası'nın heyelan duyarlılık analizini gerçekleştirerek olası riskli alanları mekânsal olarak belirlemektir. Araştırmada Çok Kriterli Karar Verme (ÇKKV) yöntemlerinden biri olan Analitik Hiyerarşi Süreci (AHS) ve Coğrafi Bilgi Sistemleri (CBS) teknolojisi kullanılmıştır. Model kapsamında heyelanı tetikleyen eğim, yağış, litoloji, NDVI, eğim şekli, akarsuya mesafe, arazi kullanımı, yükselti, büyük toprak grupları, bakı ve yollara mesafe olmak üzere 11 farklı parametre değerlendirilmiştir. AHS yöntemiyle yapılan ağırlıklandırma sonucunda heyelan oluşumunda en etkili faktörlerin eğim (%25,10) ve yağış (%19,20) olduğu tespit edilmiştir. Üretilen duyarlılık haritasının doğruluk değerlendirmesi için çalışma alanındaki geçmiş heyelan envanteri rastgele ikiye bölünerek %70 eğitim ve %30 test verisi olarak kullanılmıştır. Alıcı İşletim Karakteristiği (ROC) eğrisi ve Eğri Altında Kalan Alan (AUC) tekniği ile yapılan doğrulama sonucunda 0.799 AUC değeri elde edilmiş olup, modelin bilimsel olarak kabul edilebilir ve güvenilir bir tahmin kapasitesine sahip olduğu kanıtlanmıştır. Analiz sonuçlarına göre havzanın %43'ünün çok fazla ve riskli, %34'ünün orta derecede riskli, %23'ünün ise az ve çok az riskli alanlardan oluştuğu saptanmıştır. En riskli bölgelerin havzanın kuzeyi ve kuzeydoğusundaki yüksek eğimli araziler olduğu belirlenmiştir. Ayrıca, Erzincan-Kelkit karayolunun belirli kesimlerinin yüksek riskli bölgede yer aldığı görülürken, geçmişteki heyelan sonrası kurulan Ahmetli Köyü afet evlerinin düşük riskli alanda bulunduğu doğrulanmıştır. Elde edilen bu duyarlılık haritasının, gelecekteki bölgesel ve mekânsal planlamalar ile afet risk yönetimi çalışmalarına önemli bir bilimsel altlık oluşturması öngörülmektedir.
Landslides, as one of the primary natural hazards, are significant mass movements that cause considerable casualties and property damage. The objective of this study is to perform a landslide susceptibility analysis of the Kara Dere Basin, located in the northwest of the Erzincan Plain, and to spatially delineate potential risk areas. In this research, the Analytic Hierarchy Process (AHP), a Multi-Criteria Decision-Making (MCDM) method, and Geographic Information Systems (GIS) technology were utilized. Within the scope of the model, eleven different landslide-conditioning parameters were evaluated: slope, precipitation, lithology, NDVI, slope curvature, distance to streams, land use, elevation, great soil groups, aspect, and distance to roads. As a result of the weighting process applied via the AHP method, slope (25.10%) and precipitation (19.20%) were identified as the most influential factors in landslide occurrence. For the validation of the susceptibility map produced, the historical landslide inventory in the study area was randomly partitioned, utilizing 70% as training data and 30% as independent testing data. The validation procedure, performed using the Receiver Operating Characteristic (ROC) curve and the Area Under the Curve (AUC) technique, yielded an AUC value of 0.799, demonstrating that the model possesses a scientifically acceptable and reliable predictive capacity. According to the spatial analysis results, it was determined that 43% of the basin consists of high and very high-risk areas, 34% of moderate-risk areas, and 23% of low and very low-risk areas. The most susceptible regions were identified as the high-gradient terrains in the north and northeast of the basin. Furthermore, while certain sections of the Erzincan-Kelkit highway were observed to be located in high-risk zones, it was verified that the Ahmetli Village disaster housing, constructed following a historical landslide event, is appropriately situated in a low-risk area. It is anticipated that the resulting susceptibility map will serve as a crucial scientific baseline for future regional and spatial planning, as well as disaster risk management strategies.