arrow
返回

An Eulerian Framework for Event-Based Pattern Verification

delete2017-10-31
delete2
delete
OA
AI
J
Jason E. Nachamkin *
Y
Yi Jin
DOI:10.1175/WAF-D-17-0065.1delete
delete原文链接
delete分享
delete收藏
查看原文
摘要

摘要

En 中文
When consulting a forecast, users often ask some variant of the following questions: Will an event of interest occur? If so, when will it occur? How long will it last? How intense will it be? Standard verification measures often do not directly communicate the ability of a forecast to answer these questions. Instead, quantitative scores typically address them indirectly or in some combined form. A more direct performance measure grew from what started as a project for a high-school intern. The challenge was to evaluate aspects of forecast quality from a set of convection-allowing (1.67 km) precipitation forecasts over Florida. Although the output was highly detailed, evaluation became manageable by simply adding a series of static landmarks with range rings and radials. Using the targets'' as a guide, the student and the two authors successfully obtained quantitative estimates of model tendencies that had heretofore only been reported anecdotally. What follows is a description of the method as well as the results from the analysis. It is hoped that this work will stimulate a broader discussion about how to extract performance information from very complex forecasts and present that information in terms that humans can readily perceive.
Keyword:
OBJECT-BASED VERIFICATION
LIVED SQUALL LINES
PRECIPITATION FORECASTS
SPRING EXPERIMENT
MODEL
MESOSCALE
PREDICTION
ENSEMBLE
SYSTEM
ORGANIZATION
AI总结

AI总结

对已上传原文的论文进行重点信息的提取,主要内容包括:简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。

期刊

Weather and Forecasting 封面图
Weather and Forecasting
IF:
3.1
论文数:
2.9K
被引数:
7.9K

机构

United States Department of Defense 封面图
United States Department of Defense
学者数:
2.8W
论文数: 2.3W
被引数: 172
引用论文

引用论文

err分享
err收藏
err分享
err收藏
err分享
err收藏
学者 查看更多内容