雪プロダクト作成のための離散的時系列モデルの開発

抄録

There are many areas where snow is main water resource and snowfall affects plant growth depending on moisture and water conditions. Therefore, it is important to understand snowfall fluctuations in snowfall areas for environmental monitoring. This paper describes a novel method for generating snow products by using SPOT/vegetation data. We engineered a discrete time-series model using a self-organizing map (SOM) and a hidden Markov Model (HMM) to reduce the influence of clouds in order to improve the accuracy of the snow products. The method developed in this research improved the processing accuracy and determined snowfall with an accuracy as high as 95% as well as reduced the calculation amount to less than one tenth of that required by continuous time-series model processing.