A novel approach for predicting burned forest area

ÖNCEL ÇEKİM H., Guney C. O., ŞENTÜRK Ö., ÖZEL KADILAR G., Ozkan K.

NATURAL HAZARDS, cilt.105, sa.2, ss.2187-2201, 2021 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 105 Sayı: 2
  • Basım Tarihi: 2021
  • Doi Numarası: 10.1007/s11069-020-04395-w
  • Dergi Adı: NATURAL HAZARDS
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, IBZ Online, PASCAL, Aerospace Database, Agricultural & Environmental Science Database, Aquatic Science & Fisheries Abstracts (ASFA), CAB Abstracts, Communication Abstracts, Environment Index, Geobase, INSPEC, Metadex, PAIS International, Pollution Abstracts, Sociological abstracts, Veterinary Science Database, DIALNET, Civil Engineering Abstracts
  • Sayfa Sayıları: ss.2187-2201
  • Anahtar Kelimeler: Forest fire, Singular spectrum analysis, Vector SSA, Mediterranean, SINGULAR SPECTRUM ANALYSIS, CLIMATE-CHANGE, FIRE, IMPACT, PATTERNS, INDEXES, VECTOR, REGION, MODEL, RISK
  • Hacettepe Üniversitesi Adresli: Evet

Özet

Forest fire hazard is a major problem in the Mediterranean region of Turkey and has a significant effect on both the climate system and ecosystems. During the last century, many forest fires accounted for the majority of the Mediterranean region in Turkey. Vector singular spectrum analysis (V-SSA) and vector multivariate singular spectrum analysis (V-MSSA) are relatively novel but powerful time series analysis techniques. The present study addresses how to forecast burned forest area (BFA) by V-SSA. One of the most important factors affecting forest fires is weather conditions. The prediction of BFA is therefore also obtained by V-MSSA using meteorological covariates (i.e., relative humidity (RH), temperature (T) and wind speed (WS). In the study, forest fire data records covering the years 2005-2019 were collected and analyzed. To gain forecast accuracy, the years 2017-2019 were used as testing data, and forecast values for 1, 3, 6, 12, 24 and 36 months were obtained. Then, V-SSA and V-MSSA models were compared via the root mean square errors (RMSEs) to reach the best model explaining BFA. Our results indicated that the RMSEs of the eight models were low and close to each other. Further, forecasts for the months of the years 2020-2022 were obtained and compared with actual BFA values by means of the RMSEs. According to RMSEs, the best forecasts are obtained using the V-MSSA model with meteorological covariates BFA, WS and T.