<![CDATA[BMKG’s Research and Development Center has developed a time series model prediction for climate forecasts as an anticipatory step in minimizing the impact of hydrometeorological disasters, both drought and wet conditions, namely Hybrid BMG (HyBMG). The performance of the HyBMG prediction model for the East Java region is tested using a verification method based on the percentage accuracy of the monthly rainfall categories based on SNI 8196:2015. The data used is the HyBMG monthly rainfall forecast data (ARIMA, ANFIS, WAVELET–ARIMA, and WAVELET–ANFIS) lag 0 to 11 from 197 observation points in East Java from August 2010 to April 2020. New data are added before HyBMG is re – run. The forecasting process is based on a simulation mechanism to copy what really happened at the operational level. Blank data were filled using IDW interpolation. Furthermore, the temporal analysis process is conducted for Neutral, El Niño, and La Niña. ENSO condition determined using ONI. The results of histogram analysis show that all four of the HyBMG methods do not indicate any improvement although updating data is applied. Meanwhile, ARIMA verification is at an average of 60 % with the other methods being lower than it. Generally, the verification of HyBMG’s outputs for La Niña episode is below 60% meanwhile for El Niño episode, only reaches 60 –70 %.]]>
Copyrights © 2022
