Ensemble based Forecast Sensitivity Observation Impact in the Hybrid 4DEnsemble Variational Data Assimilation System of the KIM-Global model

The initial condition is crucial for the accuracy of the Numerical Weather Prediction (NWP) model. In the data assimilation, the quality and characteristics of the observations play an important role of the NWP predictability. As the number of observations is gradually increased, many studies investigating the influence of observations on numerical weather prediction models are actively being  conducted. In addition, the research on tools for evaluating the impact of observations on numerical weather forecasting is also emerging.
Currently, several methods to carry out the examination of the impact of observational data used in the data assimilation: OSEs (Observing System Experiences), FSO (Forecast Sensitivity to Observations), and EFSO (Ensemble Forecast Sensitivity to Observations) using an ensemble predictions.
The Korean Meteorological Administration (KMA) has been operating the KIM (Korean Integrated Model)-Global model since April 2020, which adopts a Hybrid 4DEnVar. In the year of 2023, we developed an EFSO system using an ensemble forecast field and tested on the analysis the observations sensitivity of the KIM-Global model. When the EFSO applied for one month in July 2022 based on the low-resolution global analysis system of the KIM-Global model, the influence of radiosonde data and GNSS-RO (Global Navigation Satellite Systems Radio Occultation) data were 24.3% and 38.5%, respectively, which were the most influential observation data. 
Meanwhile KMA has been conducting an IOP (Intensive Observation Program) since 2020, to monitor and improve predictability of severe weather phenomena that occur in the West Sea and Gyeonggi Bay in summer and affect the Seoul metropolitan area. 
In this study we have investigated impact of radiosonde and dropsonde data observed in the IOP on the KIM-Global model’s forecast by using the EFSO. This results will be discussed in the presentation.