Knowledge extraction from large climatological data sets using a genome-wide analysis approach: application to the 2005 and 2010 Amazon droughts

• Main Authors: Heloisa Ruivo, Gilvan Sampaio, Fernando M. Ramos
• Format: Article in Journal/Newspaper
• Language: unknown
• Subjects: North Atlantic
• Online Access: http://hdl.handle.net/10.1007/s10584-014-1066-7

Today, the volume of data generated in almost all disciplines, particularly in meteorology and climate science, is dramatically increasing. Among the challenges generated by this “data deluge” is the development of efficient knowledge discovery strategies. Here, we show that statistical and computational tools used to analyze large data sets of genome-wide studies can be fruitfully applied to a climatic context. Although not as powerful as some techniques already in use by climatologists, these tools are simple and robust, and can easily be adapted to detect early warning signals for extreme events like droughts or be used to filter large data sets before applying other more advanced and computationally expensive methods. We test this approach in our investigation of the causes of the Amazon droughts of 2005 and 2010. Our results highlight the major role played in these extreme events by the warming of the sea’s surface temperature, mainly in the tropical North Atlantic. Our findings are in agreement with several analyses published in the literature. The main message we convey is that free and open-source data mining and visualization techniques routinely used in genetic studies can be useful in helping scientists to extract knowledge from large climatic data sets, particularly in regions of the world that are vulnerable to climate change but where the availability of technical expertise is critically scarce. Copyright Springer Science+Business Media Dordrecht 2014