人力飛行機長距離飛行におけるダウンスケール気象予測の適用とその検証

摘要

現今の航空気象予測分野においては，一般に気象予測はrn各観測地点で観測される気象情報をもとに，5kmメッシュrnを用いて3時間おきに実施され，1時間間隔の予測データrnが利用可能である．この予測データの問題点として，航空rn気象において重要なシビアな現象（地形性の乱気流や下層rnウインドシアなど）を捉えるには空間解像度が不十分であrnること，時間スケールが実際の飛行時間に対して大き過ぎrnることなどがあげられる．%Research of Long-Distance Human-Powered Flight has been performed through the following four items: Aircraft design, Risk management, Pilot performance, Weather prediction. Actual flight took place on August 12, 2009. The flight distance was 20.72 km. In this study, results of weather prediction using 1-way downscaling technique are validated by surface observation data and feasibility of weather prediction for Long-Distance Human-Powered Flight is discussed. The weather prediction with 1-km mesh decreases RMSE of wind speed by 0.1-0.2 m/s compared with that with 5-km mesh. The weather prediction with 1-km mesh also has a potential to reproduce nonstationary wind. The RMSE gradually increase with time, which is mainly caused by initial and boundary data given from coarse mesh model. To reduce the RMSE, it is desirable to use newest analysis data as possible for initial and boundary data.