This paper focuses on how to extract physically meaningful information from climate data,with em-phases placed on adaptive and local analysis. It is argued that many traditional statistical analysis methods with rigorous mathematical footing may not be efficient in extracting essential physical information from climate da-ta;rather,adaptive and local analysis methods that agree well with fundamental physical principles are more ca-pable of capturing key information of climate data. To illustrate the improved power of adaptive and local analy-sis of climate data,we also introduce briefly the empirical mode decomposition and its later developments.