5:15 PM - 6:45 PM
[AAS08-P09] Probability Prediction by the JMA's Atmospheric Transport Model for Volcanic Ash with Ensemble Forecast Data
Keywords:ATM, EPS, QVA
Effects of tephra (volcanic ash and lapilli) blown by the wind from volcanic plumes are more widespread as the scale of the eruption becomes larger. For the prevention of disasters caused by tephra, volcanic ash advisories (VAAs) and volcanic ash fall forecasts (VAFFs), which predict volcanic ash in the atmosphere and ash fall and lapilli on the land and sea surface, respectively, are issued by the Japan Meteorological Agency (JMA). In the production of these information, the offline atmospheric transport model (JMA-ATM; Tech. Rep. MRI 84 (2021)) is executed with input deterministic forecast values derived from the global spectral model (GSM), meso-scale model (MSM) or local forecast model (LFM), corresponding to the forecast domain and the forecast time. However, while VAFF has issued a quantitative forecast showing the amount of expected ash fall since March 2015, VAA has issued a qualitative advisory showing only the expected volcanic ash dispersion area since April 1999, and no probability information are added to indicate the reliability of the predictions.
Uncertainty in the tephra transport prediction by the JMA-ATM comes from the initial values of emission source parameters (ESPs), the input grid-point values (GPVs) of the atmospheric field and the imperfection of the ATM itself. For numerical weather prediction GPVs, ensemble prediction systems (EPSs; Kishou Kenkyu Note 201 (2002)) have already been applied to quantify their uncertainties, the GEPS and MEPS are currently in operation (Suchi Yoho Kaisetsu Shiryoshu 56 (2024)). The members of the EPSs consist of a non-perturbed control run and perturbed runs in which the initial perturbations are derived from singular vectors (and a local ensemble transform Kalman filter in the GEPS) and mixed by a variance minimum rotation (Sep. Vol. Annu. Rep. NPD 52 (2006), 62 (2016)). Therefore, for all the perturbed runs, the growth rates are equivalent and the forecast results are equally comparable. For the uncertainty of the atmospheric field in the ATM prediction, probability information can be added by the input of an arbitrary extract from the ensemble GPVs of the GEPS or MEPS, depending on the number of ensemble members available in the operational VAA and VAFF.
The Meteorological Research Institute (MRI) of the JMA is planning to develop a technology for the probability prediction of tephra transport under one of the MRI research plans which start in April 2024. The establishment of prediction technologies for probability combined with atmospheric concentration could contribute to the information of quantitative volcanic ash (QVA; Kawaguchi et al., JpGU2022; Furukawa et al., JpGU2024) that will be recommended in the future. In this presentation, we will report on differences in the ATM predictions with high-resolution deterministic GPV or ensemble GPVs, perturbation growth rates for longer forecast time, sensitivity experiments for ensemble mean, spread, expected maximum concentration, etc. using different numbers of ensemble members concerning the probability prediction of tephra transport due to the uncertainty of the atmospheric field. Probability prediction in combination with ensemble ESPs is a future subject.
Uncertainty in the tephra transport prediction by the JMA-ATM comes from the initial values of emission source parameters (ESPs), the input grid-point values (GPVs) of the atmospheric field and the imperfection of the ATM itself. For numerical weather prediction GPVs, ensemble prediction systems (EPSs; Kishou Kenkyu Note 201 (2002)) have already been applied to quantify their uncertainties, the GEPS and MEPS are currently in operation (Suchi Yoho Kaisetsu Shiryoshu 56 (2024)). The members of the EPSs consist of a non-perturbed control run and perturbed runs in which the initial perturbations are derived from singular vectors (and a local ensemble transform Kalman filter in the GEPS) and mixed by a variance minimum rotation (Sep. Vol. Annu. Rep. NPD 52 (2006), 62 (2016)). Therefore, for all the perturbed runs, the growth rates are equivalent and the forecast results are equally comparable. For the uncertainty of the atmospheric field in the ATM prediction, probability information can be added by the input of an arbitrary extract from the ensemble GPVs of the GEPS or MEPS, depending on the number of ensemble members available in the operational VAA and VAFF.
The Meteorological Research Institute (MRI) of the JMA is planning to develop a technology for the probability prediction of tephra transport under one of the MRI research plans which start in April 2024. The establishment of prediction technologies for probability combined with atmospheric concentration could contribute to the information of quantitative volcanic ash (QVA; Kawaguchi et al., JpGU2022; Furukawa et al., JpGU2024) that will be recommended in the future. In this presentation, we will report on differences in the ATM predictions with high-resolution deterministic GPV or ensemble GPVs, perturbation growth rates for longer forecast time, sensitivity experiments for ensemble mean, spread, expected maximum concentration, etc. using different numbers of ensemble members concerning the probability prediction of tephra transport due to the uncertainty of the atmospheric field. Probability prediction in combination with ensemble ESPs is a future subject.