Seasonal trend analysis and ARIMA modeling of relative humidity and wind speed time series around Yamula Dam


Eymen A. , Köylü Ü.

METEOROLOGY AND ATMOSPHERIC PHYSICS, cilt.131, sa.3, ss.601-612, 2019 (SCI İndekslerine Giren Dergi) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 131 Konu: 3
  • Basım Tarihi: 2019
  • Doi Numarası: 10.1007/s00703-018-0591-8
  • Dergi Adı: METEOROLOGY AND ATMOSPHERIC PHYSICS
  • Sayfa Sayıları: ss.601-612

Özet

Local climate change is determined by analysis of long-term recorded meteorological data. In the statistical analysis of the meteorological data, the Mann-Kendall rank test, which is one of the non-parametrical tests, has been used; on the other hand, for determining the power of the trend, Theil-Sen method has been used on the data obtained from 16 meteorological stations. The stations cover the provinces of Kayseri, Sivas, Yozgat, and Nevehir in the Central Anatolia region of Turkey. Changes in land-use affect local climate. Dams are structures that cause major changes on the land. Yamula Dam is located 25 km northwest of Kayseri. The dam has huge water body which is approximately 85 km(2). The mentioned tests have been used for detecting the presence of any positive or negative trend in meteorological data. The meteorological data in relation to the seasonal average, maximum, and minimum values of the relative humidity and seasonal average wind speed have been organized as time series and the tests have been conducted accordingly. As a result of these tests, the following have been identified: increase was observed in minimum relative humidity values in the spring, summer, and autumn seasons. As for the seasonal average wind speed, decrease was detected for nine stations in all seasons, whereas increase was observed in four stations. After the trend analysis, pre-dam mean relative humidity time series were modeled with Autoregressive Integrated Moving Averages (ARIMA) model which is statistical modeling tool. Post-dam relative humidity values were predicted by ARIMA models.