This paper describes five sets of regions intended for use in summarizing extreme weather over Earth’s land areas from a climate perspective. The sets differ in terms of their target size: ∼10 Mm2, ∼5 Mm2, ∼2 Mm2, ∼0.5 Mm2, and ∼0.1 Mm2 (where 1 Mm2= 1 million km2). The regions are based on political/economic divisions, and hence are intended to be primarily aligned with geographical domains of decision-making and disaster response rather than other factors such as climatological homogeneity. This paper describes the method for defining these sets of regions; provides the final definitions of the regions; and performs some comparisons across the five sets and other available regional definitions with global land coverage, according to climatological and non-climatological properties.

Despite being a well-established research field, the detection and attribution of observed climate change to anthropogenic forcing is not yet provided as a climate service. One reason for this is the lack of a methodology for performing tailored detection and attribution assessments on a rapid time scale. Here we develop such an approach, based on the translation of quantitative analysis into the “confidence” language employed in recent Assessment Reports of the Intergovernmental Panel on Climate Change. While its systematic nature necessarily ignores some nuances examined in detailed expert assessments, the approach nevertheless goes beyond most detection and attribution studies in considering contributors to building confidence such as errors in observational data products arising from sparse monitoring networks. When compared against recent expert assessments, the results of this approach closely match those of the existing assessments. Where there are small discrepancies, these variously reflect ambiguities in the details of what is being assessed, reveal nuances or limitations of the expert assessments, or indicate limitations of the accuracy of the sort of systematic approach employed here. Deployment of the method on 116 regional assessments of recent temperature and precipitation changes indicates that existing rules of thumb concerning the detectability of climate change ignore the full range of sources of uncertainty, most particularly the importance of adequate observational monitoring.

Impacts of recent regional changes in climate on natural and human systems are documented across the globe, yet studies explicitly linking these observations to anthropogenic forcing of the climate are scarce. Here we provide a systematic assessment of the role of anthropogenic climate change for the range of impacts of regional climate trends reported in the IPCC's Fifth Assessment Report. We find that almost two-thirds of the impacts related to atmospheric and ocean temperature can be confidently attributed to anthropogenic forcing. In contrast, evidence connecting changes in precipitation and their respective impacts to human influence is still weak. Moreover, anthropogenic climate change has been a major influence for approximately three-quarters of the impacts observed on continental scales. Hence the effects of anthropogenic emissions can now be discerned not only globally, but also at more regional and local scales for a variety of natural and human systems.

When working with large, multidimensional and multivariate data, science users are frequently interested in understanding variation in data, as opposed to the actual data values. Our work focuses on exploring how a simple statistical metric, the Coefficient of Variation (or Cv), can be used in several different ways to facilitate understanding variation in large data. As a statistical measure, it offers a key advantage over more widely accepted measures like standard deviation, namely to its ability to capture local variation properties. As a multidimensional projection operator, Cv is an effective way of reducing data size while preserving the key variational signal. Visualizations produced from Cv that target conveying variation in data are highly informative, especially compared to those produced with more widely known methods. We demonstrate these ideas within the context of a two-part application case study focusing on understanding long-Term trends in the the changes in precipitation and winds in large-scale climate model ensemble output.

An ongoing challenge in visual exploration and analysis of large, multi-dimensional datasets is how to present useful, concise information to a user for some specific visualization tasks. Typical approaches to this problem have proposed either reduced-resolution versions of data, or projections of data, or both. These approaches still have some limitations such as consuming high computation or suffering from errors. In this work, we explore the use of a statistical metric as the basis for both projections and reduced-resolution versions of data, with a particular focus on preserving one key trait in data, namely variation. We use two different case studies to explore this idea, one that uses a synthetic dataset, and another that uses a large ensemble collection produced by an atmospheric modeling code to study long-term changes in global precipitation. The primary findings of our work are that in terms of preserving the variation signal inherent in data, that using a statistical measure more faithfully preserves this key characteristic across both multi-dimensional projections and multi-resolution representations than a methodology based upon averaging.

The Weather Risk Attribution Forecast (WRAF) is a forecasting tool that uses output from global climate models to make simultaneous attribution statements about whether and how greenhouse gas emissions have contributed to extreme weather across the globe. However, in conducting a large number of simultaneous hypothesis tests, the WRAF is prone to identifying false “discoveries.” A common technique for addressing this multiple testing problem is to adjust the procedure in a way that controls the proportion of true null hypotheses that are incorrectly rejected, or the false discovery rate (FDR). Unfortunately, generic FDR procedures suffer from low power when the hypotheses are dependent, and techniques designed to account for dependence are sensitive to misspecification of the underlying statistical model. In this article, we develop a Bayesian decision-theoretical approach for dependent multiple testing and a nonparametric hierarchical statistical model that flexibly controls false discovery and is robust to model misspecification. We illustrate the robustness of our procedure to model error with a simulation study, using a framework that accounts for generic spatial dependence and allows the practitioner to flexibly specify the decision criteria. Finally, we apply our procedure to several seasonal forecasts and discuss implementation for the WRAF workflow. Supplementary materials for this article, including a standardized description of the materials available for reproducing the work, are available as an online supplement.

Assessing past impacts of observed climate change on natural, human and managed systems requires detailed knowledge about the effects of both climatic and other drivers of change, and their respective interaction. Resulting requirements with regard to system understanding and long-term observational data can be prohibitive for quantitative detection and attribution methods, especially in the case of human systems and in regions with poor monitoring records. To enable a structured examination of past impacts in such cases, we follow the logic of quantitative attribution assessments, however, allowing for qualitative methods and different types of evidence. We demonstrate how multiple lines of evidence can be integrated in support of attribution exercises for human and managed systems. Results show that careful analysis can allow for attribution statements without explicit end-to-end modeling of the whole climate-impact system. However, care must be taken not to overstate or generalize the results and to avoid bias when the analysis is motivated by and limited to observations considered consistent with climate change impacts.

Heatwaves with large impacts have increased in the recent past and will continue to increase under future warming. However, the implication for population exposure to severe heatwaves remains unexplored. Here, we characterize maximum potential human exposure (without passive/active reduction measures) to severe heatwaves in India. We show that if the global mean temperature is limited to 2.0 °C above pre-industrial conditions, the frequency of severe heatwaves will rise by 30 times the current climate by the end-21st century. In contrast, the frequency is projected to be about 2.5 times more (than the low-warming scenario of 2 °C) under conditions expected if the RCP8.5 'business-as-usual' emissions scenario is followed. Under the 2.0 °C low-warming target, population exposure to severe heatwaves is projected to increase by about 15 and 92 times the current level by the mid and end-21st century respectively. Strategies to reduce population growth in India during the 21st century may provide only limited mitigation of heatwave exposure mostly late in the century. Limiting global temperatures to 1.5 °C above preindustrial would reduce the exposure by half relative to RCP8.5 by the mid-21st century. If global temperatures are to exceed 1.5 °C then substantial measures will be required to offset the large increase in exposure to severe heatwaves in India.

In the midsummer of 2013, Central and Eastern China (CEC) was hit by an extraordinary heat event, with the region experiencing the warmest July-August on record. To explore how human-induced greenhouse gas emissions and natural internal variability contributed to this heat event, we compare observed July-August mean surface air temperature with that simulated by climate models. We find that both atmospheric natural variability and anthropogenic factors contributed to this heat event. This extreme warm midsummer was associated with a positive high-pressure anomaly that was closely related to the stochastic behavior of atmospheric circulation. Diagnosis of CMIP5 models and large ensembles of two atmospheric models indicates that human influence has substantially increased the chance of warm mid-summers such as 2013 in CEC, although the exact estimated increase depends on the selection of climate models.