Long-Time Memory in Drought via Detrended Fluctuation Analysis


TATLI H. , Dalfes H. N.

WATER RESOURCES MANAGEMENT, cilt.34, ss.1199-1212, 2020 (SCI İndekslerine Giren Dergi) identifier identifier

  • Cilt numarası: 34 Konu: 3
  • Basım Tarihi: 2020
  • Doi Numarası: 10.1007/s11269-020-02493-9
  • Dergi Adı: WATER RESOURCES MANAGEMENT
  • Sayfa Sayıları: ss.1199-1212

Özet

The persistence of drought events largely determines the severity of socioeconomic and ecological impacts, unfortunately the performance of current weather forecasting models (WFM) to simulate such events is subject to great uncertainties. This study is investigating time-domain characteristics of drought persistence over Turkey by applying the detrended fluctuation analysis (DFA) method to the Palmer drought severity index (PDSI). The existence of long-range power-law correlation in PDSI fluctuations is demonstrated for time scales ranging from monthly to decadal. Understanding of such statistical patterns in PDSI values can definitely be a step forward in drought predictability. From a climatological point of view, it is found that the areas with high level DFA scaling exponent (generalized Hurst) indicate the areas of higher sensitivity to droughts and associated risks. Furthermore, the characteristics of the persistence of the PDSI in climate zones have also been examined by applying the Holdridge Life Zones (HLZ) classification. HLZ classification over Turkey leads to two climate-zones: cool-temperate and warm-temperate. In addition, when topography is taken in account, montane (cool-temperate) and lower-montane (warm-temperate) climate zones can be treated as two different zones. It has been observed that the predictable index (PI) of the PDSI derived from the DFA Hurst exponent is relatively high in the cool-temperate and montane climate zones compared to others. In fact, very different PI values were also obtained in a few HLZ climate classes within the same climate zone and with same vegetation index (i.e. steppe, dry-forest, warm-forest etc.).