Climate Researchers Crunch Data on Weather Extremes
Few data science problems can top weather forecasting and climate modeling for sheer complexity. Hence, big data is playing a key role in tracking changes in global climate and, with it, weather and climate extremes.
The application of data analysis to decades of weather and climate data can provide insights that inform billion-dollar investments in agriculture, commodities, insurance and the financial sector, experts stress.
More insights could be drawn from recent research at Northeastern University in Boston that used climate models and three “reanalysis” datasets to predict “a wider range of temperature extremes in the future.” The datasets used by the researchers were generated by combining the best available weather observations with numerical weather models.
The researchers, Evan Kodra and Auroop Ganguly of Northeastern University’s Sustainability and Data Sciences Lab, noted that recent data analysis has found further evidence of weather and climate extremes. The meteorological processes thought to be generating weather extremes, especially northern weather extremes, have been captured in global climate models.
The researchers applied big data techniques to unlock new insights about weather variability within these climate models. Their primary tool was a climate model suite dubbed CMIP5 that was used to extract more details about the mechanisms behind climate extremes. The 14-member ensemble consists of simulations from the most recent climate models generated for the Intergovernmental Panel on Climate Change.
The journal Phys.org called the extreme weather research the first of its kind to use “computational tools from big data science” to study extreme weather.
Global climate models “consistently show asymmetry in projected changes in seasonal” extreme temperatures, the researchers concluded based on the results of their data analysis of climate simulations. The findings suggest a wider range of “extreme temperature events across the globe in the future,” they warned.
Reliable climate modeling based on big data tools is increasingly seen as critical tools across a range of economic sectors, especially agriculture. Accurate climate models are increasingly used by financial institutions, insurers and other industries to gauge the impact of emerging weather and climate patterns on future investments and supply chains.
These industry sectors are increasingly seeking more granular climate data to make financial bets.
Big data analytics also is growing in importance as climate and weather datasets explode. According to one estimate by Computer Science Corp., about 15 petabytes of weather and climate data gathered so far is expected to grow exponentially to about 350 petabytes by 2030, with satellite data and modeling as the biggest contributors. In situ weather and climate observations also are growing, especially with new networks coming on line, according to Sharon Hays, CSC’s vice president of science and engineering.
Increasing the understanding of climate on weather is required, or at least better understanding of the inter-annual variability of weather driven by climate phenomena, Hays told a recent climate conference. “The value in these data that has to be unlocked through analysis.”
The Northeastern University extreme weather research applying big data tools was partially funded by a National Science Foundation “Expeditions in Computing” grant, the authors said.