Search

Machine learning (ML) and in particular artificial neural networks (ANNs) push state-of-the-art solutions for many hard problems e.g., image classification, speech recognition or time series forecasting. In the domain of climate science, ANNs have good prospects to identify causally linked modes of climate variability as key to understand the climate system and to improve the predictive skills of forecast systems. To attribute climate events in a data-driven way with ANNs, we need sufficient training data, which is often limited for real world measurements. The data science community provides standard data sets for many applications. As a new data set, we introduce a collection of climate indices typically used to describe Earth System dynamics. This collection is consistent and comprehensive as we use control simulations from Earth System Models (ESMs) over 1,000 years to derive climate indices. The data set is provided as an open-source framework that can be extended and customized to individual needs. It allows to develop new ML methodologies and to compare results to existing methods and models as benchmark. Exemplary, we use the data set to predict rainfall in the African Sahel region and El Niño Southern Oscillation with various ML models. We argue that this new data set allows to thoroughly explore techniques from the domain of explainable artificial intelligence to have trustworthy models, that are accepted by domain scientists. Our aim is to build a bridge between the data science community and researchers and practitioners from the domain of climate science to jointly improve our understanding of the climate system.

We investigate how European heat waves and their associated heat stress on humans have changed over the 20th century. We find that the heat stress has increased, even in regions where heat waves have not become warmer. As heat stress increases over wide areas of Europe there is also an increase in the total population affected by the heat stress. Heat waves pose a serious health risk to humans by reducing our ability to shed heat. We have studied the occurrence and intensity of heat waves as well as a heat stress index based on simplified wet-bulb globe temperature using data from ERA-20C reanalysis 1900-2010. Over the 110 years of data we find an overall warming of the air temperatures and dew point. The 98th percentile of both air temperature has increased by more than 1.5°C over large areas of Europe. We find an overall increase in heat wave days per year as well as an increase of air temperature during heat waves over most of Europe. As such, many densely populated areas exhibit increased heat stress during heat waves. For example, the mean heat stress during heat wave days over Paris has increased by one level, from “alert” in 1900-1930 to “caution” in 1980-2010. The fraction of the population exposed to heat waves has increased by 10%/century in central Europe and 25%/century over the Mediterranean. We find more heat waves during 1920 - 1950, which may be related to the positive phase of the Atlantic Multidecadal Variation (AMV). This suggests that the heat stress during European heat waves may also be influenced by internal climate variability, and large-ensemble model simulations may be used to disentangle the effects of natural variability and anthropogenic forcing.