Fig. 1. Climate risk is defined by three elements: vulnerability, exposure, and hazard. While the level of certainty is high and the significance of evidence is well established regarding disparities between the global north and the global south in vulnerability and exposure, certainty is low and evidence is not well established when it comes to climate hazard. Examples of disparities in climate hazard are few, and the contrast between the global north and south shown by these examples is largely not sharp.
The geographically different impacts of climate change are not only limited to a specific sector but emerge in various fields. Some examples of disparities caused by climate change include differences in its impact on economic production (Burke et al. 2015; Callahan and Mankin 2022; Diffenbaugh and Burke 2019; King and Harrington 2018), disparities in how people allocate their time between work and leisure (Zivin and Neidell 2014), effects on poverty, urbanization, and migration (Burzyński et al. 2022), disparities in gender-based health (Sorensen et al. 2018), and contrasting consequences on income in urban and rural regions within countries (Paglialunga et al. 2022).
Compensating for disproportionate loss and damage imposed by climate change has been an integral, however, one of the most debatable and complex subjects in the international forums (e.g., the 27th Conference of Parties (COP27), held in Egypt in November 2022) that discuss climate change policies (Dorkenoo et al. 2022) (see text in the online supplementary material regarding the concept of loss and damage imposed by climate change). This topic is especially relevant considering the debate about the distribution of responsibility for causing climate change and who should bear the costs (Farber 2007), which are difficult to quantify (Rising et al. 2022), but certainly high (Dietz et al. 2018). Industrial and relatively rich countries (thereafter, the global north) have historically been the main emitters of greenhouse gases (GHGs) (Althor et al. 2016; Wei et al. 2016). Being legally liable for GHG emissions has been the main concern of the global north around the issue of responsibilities and compensation over climate change. However, the relatively poor developing countries (thereafter, the global south), which have contributed less to the problem, are disproportionately impacted by climate change (IPCC 2022).
In the current research, we emphasize that discussions over differential climate risk and compensation for the associated loss and damage have widely been tackled mainly from the lens of disparities between the global north and global south in terms of vulnerability and exposure. This is mainly because vulnerability and exposure display clear differences between these two groups of countries and the evidence for these disparities are highly significant, certain, and well-established (Fig. 1). But when it comes to climate hazard - the third component that defines climate risk - there are only a few significant evidence of sharp disparities between the global north and the global south. Additionally, evidence for such a differential pattern are either based on variables with relatively less significance to society or the contrast between the global north and south is not sharp. This area of research regarding north-south disparities in climate hazard has received relatively little attention (IPCC 2014) albeit its centrality in shaping climate change impact and defining losses, damages, and responsibilities.
2. Data and Methods
a. Observations and CMIP data
The observed 3-hourly temperature, dew point temperature and surface pressure data with a spatial resolution of 0.25° covering the 1959-2021 period were from the ERA5 reanalysis (Hersbach et al. 2020) (available at http://apps.ecmwf.int/datasets/). We used the daily output of temperature over the 1976–2100 period from 31 CMIP5 GCMs forced by historical forcing until 2005 and Representative Concentration Pathway 8.5 (RCP8.5) scenario thereafter (Taylor et al. 2012) and 29 CMIP6 GCMs forced by historical forcing up to 2014 and Shared Socioeconomic Pathway 5-8.5 (SSP5–8.5) scenario thereafter (Eyring et al. 2016). RCP8.5 represents a business-as-usual scenario with radiative forcing, reaching about 8.5 Wm-2 in 2100. SSP5-8.5 is approximately equivalent to RCP8.5 (Russo et al. 2019). For each model, one ensemble member was used. The list of global climate models and their details were provided in Table S1. All model outputs from CMIP6 (respectively from CMIP5) were re-gridded to a common 1° × 1° grid (respectively 1.5° × 1.5° grid). The anomalies in the model outputs are obtained for the period 1976-2005. The gridded population density was from the Gridded Population of the World (CIESIN 2018). The gridded global GDP was provided by Kummu et al. (2018).
b. Derivation of wet-bulb temperature
Wet-bulb temperature (TW) was computed adopting a method developed by Davies-Jones (Davies-Jones 2008) using surface temperature, humidity, and pressure, derived from the 3-hourly CMIP outputs and the ERA5 reanalysis.
c. Outdoor days
Our earlier study introduced the novel concept of “outdoor days” defined as the number of days with moderate temperature, neither too cold nor too hot, that allows most people to enjoy outdoor activities (Choi et al. 2023). Here, we adopt the same definition, following Choi et al. (2023), to estimate “outdoor days” based on dry-bulb and wet-bulb temperature. That is, the range of dry-bulb temperature from 10 ℃ to 25 ℃ is assumed to be suitable for outdoor activity (This range of temperature is comparable to a wet‐bulb temperature ranging from 8 ℃ to 15 ℃). We assume this definition is valid for all locations on Earth. However, the exact range of temperature defining an outdoor day may, in general, vary slightly depending on geographical location, tolerance levels of the local population, and what exact fraction of the population is meant when we say “most people”. Our definition of an outdoor day is discussed further below, where it is demonstrated that our results are not sensitive to the exact range of temperature we assume in this analysis. To explore this sensitivity, please visit https://eltahir.mit.edu/globaloutdoordays/.
Most existing studies that investigated concepts similar to outdoor days presented herein are limited to local and regional scales (Choi et al. 2023; Gao et al. 2018; Hanlon et al. 2021; Heng and Chow 2019; Lin et al. 2019; Spagnolo and de Dear 2003; Wu et al. 2017; Zhang et al. 2022; Zhang 2016). Previous studies referred to pleasant weather conditions as mild weather (Lin et al. 2019; van der Wiel et al. 2017; Zhang et al. 2022), mild days (Day et al. 2021), good weather (Zhang 2016), outdoor thermal comfort (Heng and Chow 2019), thermal comfort condition (Gao et al. 2018), comfortable days (Wu et al. 2017), and thermal comfort (Spagnolo and de Dear 2003).
Climate variables used in previous studies to analyze changes in the characteristics of mild weather include temperature (Choi et al. 2023; Hanlon et al. 2021; Zhang et al. 2022), dewpoint temperature (van der Wiel et al. 2017), wet-bulb temperature (Choi et al. 2023), precipitation (van der Wiel et al. 2017; Zhang 2016), relative humidity, wind speed, sunshine duration (Lin et al. 2019; Zhang et al. 2022), shortwave radiation, diffuse shortwave radiation, longwave radiation, and velocity (Spagnolo and de Dear 2003). Temperature is the primary and common variable used in most, if not all, of the previous studies that investigated the impact of climate change on concepts similar to outdoor days. However, the optimum ranges used for the daily maximum temperature to define mild weather vary considerably. The optimum daily maximum temperature considered in these studies ranges between 18 ℃ and 31.6 ℃. A summary of these literatures can be found in the online supplementary material (Table S2).
The studies of van der Wiel et al. (2017) and Zhang et al. (2023) are the only ones that investigated mild weather conditions on a global scale. These studies show a clear contrast in the change of mild weather between the global north and the global south. However, both studies considered applying only one or a few GCMs, limiting the reliability of future projections. Additionally, disparities between the global north and south were not the central focus of van der Wiel et al. (2017). To the best of our knowledge, our study is the first to provide more evidence of its scale in terms of data and models on the north-south disparities in outdoor days, improving our understanding of how disparities in climate hazard shape the contrasting risk of climate change on the global scale.
3. Results
In the current climate, outdoor days occur frequently in most regions of the world (Fig. 2). On average, across the land areas of the world, approximately 91 outdoor days are experienced every year (i.e., about 25% of the days of a year), although seasonal and/or regional differences are evident. If we restrict ourselves to residential areas (i.e., areas with a population density above 1 person per square kilometer), more outdoor days are found (165 days per year; i.e., about 45% of the days of a year). Particularly, the global south stands out due to more frequent outdoor days, compared to the global north. On a regional scale, high-latitude countries, such as Canada and Russia, are too cold (Fig. S2a), resulting in fewer outdoor days, especially during cold months, whereas countries such as Angola and regions like the southern parts of Brazil show frequent outdoor days regardless of season (Fig. 2).