|
Abstract:
|
Quantitative precipitation estimation(QPE) is a challenging and important task in several applications such as atmospheric science, hydrology, and environmental science. In this paper, we focus on estimating precipitation area rather than the amount, which makes the statistical challenge to model a binary process. Properly fusing datasets of different resolution has become increasingly important as it is easy to obtain data from different sources. For instance, data from both ground monitoring station and remote sensing can be jointly presented. These cases often pose a spatial misalignment problem and there are various ways to deal with such challenge. However, the majority of models in the literature focus on the implication of continuous measurement. In this paper, we propose a statistical downscaler for spatially misaligned binary data. Spatial binary responses are regressed on high resolution data using spatially varying coefficients. A simulation study is performed to assess the model performance based on various metrics. The proposed model is also illustrated in the analysis of rainfall data in South Korea.
|
