Nonstationarity in seasonality of extreme precipitation: A nonparametric circular statistical approach and its application. Issue 6 (21st June 2015)

• Record Type: Journal Article
• Authors: Dhakal, Nirajan; Jain, Shaleen; Gray, Alexander; Dandy, Michael; Stancioff, Esperanza
• Is Part Of: Water resources research. Volume 51:Issue 6(2015:Jun.)
• Page Start: 4499
• Page End: 4515
• Publication Date: 2015-06-21
• Subjects: Hydrology -- Periodicals; 333.91
• DOI: 10.1002/2014WR016399
• Languages: English
• ISSNs: 0043-1397

Abstract

<abstract abstract-type="main"> <title>Abstract</title> <p>Changes in seasonality of extreme storms have important implications for public safety, storm water infrastructure, and, in general, adaptation strategies in a changing climate. While past research on this topic offers some approaches to characterize seasonality, the methods are somewhat limited in their ability to discern the diversity of distributional types for extreme precipitation dates. Herein, we present a comprehensive approach for assessment of temporal changes in the calendar dates for extreme precipitation within a circular statistics framework which entails: (a) three measures to summarize circular random variables (traditional approach), (b) four nonparametric statistical tests, and (c) a new nonparametric circular density method to provide a robust assessment of the nature of probability distribution and changes. Two 30 year blocks (1951–1980 and 1981–2010) of annual maximum daily precipitation from 10 stations across the state of Maine were used for our analysis. Assessment of seasonality based on nonparametric approach indicated nonstationarity; some stations exhibited shifts in significant mode toward Spring season for the recent time period while some other stations exhibited multimodal seasonal pattern for both the time periods. Nonparametric circular density method, used in this study, allows for an adaptive estimation of seasonal density. Despite the limitation of being sensitive to the smoothing parameter, this method can accurately characterize one or more modes of seasonal peaks, as well as pave the way toward assessment of changes in seasonality over time.</p> </abstract>