The study develops and validates a theoretical model to predict thermal mortality under natural conditions, based on measurements of mortality performed in the laboratory at multiple constant temperatures. The theoretical model first fits a thermal tolerance landscape, which describes how survival probability is affected by both temperature and exposure time, to the empirical measurements of mortality obtained in the laboratory under controlled conditions. Then, employing a numerical approximation to the analytical solution based on differential calculus, it combines this tolerance landscape with ambient temperature records in natural settings to predict the survival probability curve under these thermal conditions. These predictions were validated by contrasting predicted and observed mortality curves in 11 Drosophila species under three different warming rates, reported in the literature, which were virtually indistinguishable. Having validated the model, the study then examines how mortality should be affected by climate change in a natural population of Drosophila subobscura from Santiago, Chile, employing temperature records for this location during 1984 - 1991 and 2014 - 2018. The cumulative mortality predicted from temperature records closely resemble the periods of population collapse recorded for this population during the Austral summer and, according to the model, warming temperatures in the past 30 years may have advanced this period by almost a month. This methodology is highly general and can in principle be employed to predict temperature mortality in small ectotherms under any varying thermal regime.  

Datasets of Drosophila thermal mortality under constant and warming conditions, as well as data on seasonal abundances of Drosophila subobscura in the field,  were compiled from the literature. Ambient temperature records for 1984 - 1991 and 2014 - 2018 were obtained from climatological data available online from the Chilean government websites. The detailed sources are provided in the attached Readme.txt and in the original paper.     

Additional information is available in the attached Readme.txt