资源环境科技发展态势分析平台

Observed hourly data from New York City and San Francisco are examined, and the role of hourly changes in the occurrence of daily extreme temperatures is assessed. The tails of the conditional distribution of daily extreme temperatures are modeled with a class of extreme value models that incorporate information on changes in hourly temperature, and location-specific behavior is found. The proposed statistical analyses, which are easily carried out using open-source software, could be used to assess whether the hourly down-scaled data necessary for many impact and adaptation studies accurately reproduce the relationship between observed hourly temperatures and daily temperature extremes at a given site.