Attribution of Temperature Increases: A Probabilistic Analysis

Instructors who include a module on climate science might want to draw upon a new study in the journal Climate Risk Management. The study addresses one of the most frequent questprobabilityion of students, as well as the general public: could current warming trends (through July 2014, consecutive months in which global land and ocean temperatures have exceeded the 20th century monthly average) be primarily a function of natural system fluctuation? The study, conducted by scientists at the Commonwealth Scientific and Research Organization and the University of Wollongong, seeks to facilitate the development of probabilistic statements to assess the likelihood that anomalous warming (in the study, defined as the unbroken sequence of 346 months, from March 1985-December 2013, of average monthly temperatures exceeding the 20th Century average) is a function of anthropogenic greenhouse gas emissions. It extends the methodology of attribution studies by using a statistically robust approach that incorporates time series modeling, validation and 100,000 bootstrap simulations of temperature time series (which can facilitate sampling distribution of statistics). The model correlates global temperature to well-recognized drivers of temperature variation, including El Niño, solar radiation, volcanic aerosols and greenhouse gas concentrations. The model was run using the historic record and re-run without the anthropogenic influence of greenhouse gas emissions.
Among the study’s conclusions:
1. The chances of observing 304 months or more of temperatures exceeding the 20th century average when anthropogenic greenhouse gases are excluded are less than 0.001%;
2. As was the case of the conclusions in a previous study, solar radiation was found to be “an insignificant contributor to global warming over the last century;”
3. Even the eleven cases in which global 10-year temperatures declined during this period were consistent with predictions of models of climate change. The study concluded that “there is only a 0.01 per cent change of observing the record 11 events (or fewer) in the absence of recent global warming.” Moreover, the researcher’s concluded, “even when GHG emissions are included, the observed number of cooling periods is low compared with an average of 15 events simulated. Thus, rather than being an indicator that global warming is not occurring … the observed number of cooling periods reinforces the case in support of recent global warming due to human influence.”

Be Sociable, Share!