Regional detection of multiple change points and workable application for precipitation by maximum likelihood approach
- PDF / 6,676,083 Bytes
- 16 Pages / 595.276 x 790.866 pts Page_size
- 8 Downloads / 216 Views
ORIGINAL PAPER
Regional detection of multiple change points and workable application for precipitation by maximum likelihood approach Qianfeng Wang 1,2
&
Jia Tang 1,2 & Jingyu Zeng 1,2 & Song Leng 3 & Wei Shui 1,2
Received: 26 December 2017 / Accepted: 4 September 2019 # Saudi Society for Geosciences 2019
Abstract Precipitation is an essential part of the hydrological cycle. Objective change-point detection plays an important role in the research on extreme climate events and risk assessment in the context of global climate change. The study can automatically identify and extract multiple change points over a lengthy time series and calculate trends over several segmentations by uniting the maximum likelihood approach and the nonparametric Mann-Kendall trend test which is also compared with ordinary least squares (OLS). The observed station precipitation data were compiled over Hebei Province in China from 1961 to 2014. Temporal-spatial characteristics were also investigated by using several indices: the number of change points, standard deviation, timing of change points, minimum and maximum trends for segmentations, and the generalization trend. Obvious change points were generally around 1974, 1981, and 1998, and occurred at few stations before 1974 and after 2000, indicating that precipitation was relatively stable in the study area during the periods 1961–1974 and 2000–2014; the minimum trend for segmentations decreased over all stations; the maximum trend for segmentations increased over all stations except Leting; and generalization trend weakened abrupt changes in specific time sections among the multiple segmentations. Change-point detection followed by trend analysis can detect an obvious increasing or decreasing trend over certain parts of the time series and the proposed method can serve as a management tool with proper measures to deal with climate change. The results for both segmentations and generalization can provide a workable reference for managing regional water resources and implementing strategies to mitigate meteorological risks. Keywords Precipitation . Change point . Maximum likelihood approach . Mann-Kendall . Trend; Hebei
Introduction Global climate change continues to influence our world’s ecosystem, water cycle and soil condition, crop yield, and other aspects of the natural environment in addition to social and economic activities (Fischer et al. 2012; Panda et al. 2013; Zhang 2013; Wang et al. 2018). Research on global climate Responsible Editor: Abdullah M. Al-Amri * Wei Shui [email protected] 1
College of Environment and Resource, Fujian Provincial Key Laboratory of Remote Sensing of Soil Erosion and Disaster Prevention, Fuzhou University, Fuzhou 350116, China
2
Key Lab of Spatial Data Mining & Information Sharing, Ministry of Education of China, Fuzhou 350116, China
3
Climate Change Cluster, University of Technology Sydney, Broadway, New South Wales 2007, Australia
change has grown particularly intense in recent years as the scientific community—and indeed, the public
Data Loading...
