Construction and application of integrated entropy model for measuring precipitation complexity

Measuring the complexity of precipitation can help understand the intrinsic nature of precipitation changes. Entropy as a tool of measuring complexity, has been widely applied to the precipitation complexity. However, previous study mostly used a single entropy model, resulting in differences in precipitation complexity measurement for the same region. In this study, the permutation entropy (PE), sample entropy (SE), wavelet entropy (WE) and fuzzy entropy (FE) were used to measure the precipitation complexity respectively. Then, this study presents an integrated entropy model on the set pair analysis (SPA) and fuzzy analytic hierarchy process (FAHP) to combine different entropy models for precipitation complexity measurement. Finally, the integrated entropy model was applied for analyzing the spatial differences among complexity for precipitation and its influencing factors in Fujian. Results indicated that: (1) the integrated entropy model is superior to single entropy model, and it can provide more accurate measurement of precipitation complexity, (2) the spatial pattern of integrated entropy shows a declining trend from the northwest (maximum entropy of Ningde is 0.953) toward southeast (minimum entropy of Zhangzhou is 0.861), corresponding to the northwest and middle area with less predictable for precipitation and the more predictable in southeast coast. Additionally, the terrain and distance from sea have become the factors influence the precipitation complexity in Fujian. The research results may provide scientific reference for measurement of regional precipitation complexity and its optimal.