Host: Japan Society of Hydrology and Water Resources
Spatial pooling of data in various spatial scales using different similarity measures is studied here which to detect robust changes of 100 year rainfall in a future climate using a high resolution GCM. For spatial pooling procedure, the idea of Region-of-Influence (ROI) method is used which each grid having its own region. Spatial proximity, seasonality, and daily rainfall statistics are used as similarity measures to define each region. The approach is applied to a 20 km GCM in present and future climate condition. As a result, ROI approach in 1000 km scale revealed a good performance in estimating future changes of 100-year return values of precipitation.