DOI: 10.1007/s00382-016-2980-3
Scopus记录号: 2-s2.0-84954562838
论文题名: Atmospheric controls on Puerto Rico precipitation using artificial neural networks
作者: Ramseyer C.A. ; Mote T.L.
刊名: Climate Dynamics
ISSN: 9307575
出版年: 2016
卷: 47, 期: 2017-07-08 起始页码: 2515
结束页码: 2526
语种: 英语
英文关键词: Artificial neural networks
; Climate downscaling
; Daily Caribbean precipitation
; Predictor variables
英文摘要: The growing need for local climate change scenarios has given rise to a wide range of empirical climate downscaling techniques. One of the most critical decisions in these methodologies is the selection of appropriate predictor variables for the downscaled surface predictand. A systematic approach to selecting predictor variables should be employed to ensure that the most important variables are utilized for the study site where the climate change scenarios are being developed. Tropical study areas have been far less examined than mid- and high-latitudes in the climate downscaling literature. As a result, studies analyzing optimal predictor variables for tropics are limited. The objectives of this study include developing artificial neural networks for six sites around Puerto Rico to develop nonlinear functions between 37 atmospheric predictor variables and local rainfall. The relative importance of each predictor is analyzed to determine the most important inputs in the network. Randomized ANNs are produced to determine the statistical significance of the relative importance of each predictor variable. Lower tropospheric moisture and winds are shown to be the most important variables at all sites. Results show inter-site variability in u- and v-wind importance depending on the unique geographic situation of the site. Lower tropospheric moisture and winds are physically linked to variability in sea surface temperatures (SSTs) and the strength and position of the North Atlantic High Pressure cell (NAHP). The changes forced by anthropogenic climate change in regional SSTs and the NAHP will impact rainfall variability in Puerto Rico. © 2016, Springer-Verlag Berlin Heidelberg. Citation statistics:
资源类型: 期刊论文
标识符: http://119.78.100.158/handle/2HF3EXSE/53509
Appears in Collections: 过去全球变化的重建
There are no files associated with this item.
作者单位: Department of Geography, University of Georgia, Athens, GA, United States
Recommended Citation:
Ramseyer C.A.,Mote T.L.. Atmospheric controls on Puerto Rico precipitation using artificial neural networks[J]. Climate Dynamics,2016-01-01,47(2017-07-08)