Fig. 9. Projected change in outdoor days for four seasons. Normalized
change in outdoor days in 2071–2100 with respect to 1976-2005 for four
seasons. The changes are normalized by the 1976-2005 mean. Superimposed
hatching indicates that more than 80% of models agree on the sign of
the change. These global maps are derived from 29 CMIP6 GCMs.
4. Discussion
Climate change studies usually focus on climate extremes and/or changes
in the mean climate conditions, which are certainly important aspects of
climate change (Choi et al. 2021, 2022; Choi and Eltahir 2022, 2023;
Fischer and Knutti 2015; Im et al. 2017; Kang and Eltahir 2018; Pal and
Eltahir 2016; Pfahl et al. 2017; Tuel and Eltahir 2020; Zhao et al.
2021). However, investigating mild weather, as done here for outdoor
days, is of significant importance to society (Lin et al. 2019; van der
Wiel et al. 2017; Zhang et al. 2022) since it could shape the
contrasting risk of climate change on the global scale as revealed in
this study. Meanwhile, future changes in rainfall and temperature – the
two main climate variables – due to climate change show no clear
north-south disparity in climate risk. According to IPCC (2022), the
overall trend for temperatures is to increase almost uniformly in the
future across the globe. An overall wetting trend in precipitation is
expected, except for a few regions, like the Mediterranean (Tuel and
Eltahir, 2020). Therefore, future changes in rainfall and temperature do
not provide significant evidence for a north-south disparity in climate
risk.
Although substantial scientific and public attention has focused on the
asymmetric distribution of vulnerability and exposure to climate change
(Diffenbaugh and Burke 2019), some climate extremes could possibly
result in somewhat north-south disparity (Carleton et al. 2020; Crowther
et al. 2016; Hirabayashi et al. 2013; Kam et al. 2021; Park et al. 2018;
Shiogama et al. 2019; Vicedo-Cabrera et al. 2021). For instance,
Shiogama et al. (2019) displayed that the regions with relatively large
increases in several hazard indicators, i.e., extremely hot days, heavy
rainfalls, and high stream flow coincide with countries characterized by
small CO2 emissions, low income, and high vulnerability.
The frequency and pattern of river flood occurrence due to climate
change exhibit a north-south gradient (Kam et al., 2021), with flood
frequency projected to increase in regions such as Southeastern Asia and
Eastern Africa and decrease in regions such as North America and Europe
(Hirabayashi et al., 2013). The effect of climate hazard might be
extended to soil systems with considerable losses in soil carbon stocks
in high-latitude areas and an expected increase in tropical regions
(Crowther et al., 2016). This suggests a reverse disparity where
developing countries are benefiting from climate change while developed
countries are losing. Moreover, the timing of climate hazards also
exposes a north-south contrast as reported by Park et al. (2018).
According to Park et al. (2018), while the time for the emergence of
aridification due to climate change in the global south is expected to
mostly occur before 2050, it might occur later than 2050 in the global
north. Until recently, climate hazard did not seem to affect the large
discrepancies in climate risk between countries − the global north and
the global south and/or rich and poor countries and/or high and low
emitters of greenhouse gases (Fig. S3).
Our results present some important caveats. First, the simulated
temperature from CMIP5/CMIP6 models has systematic biases of temperature
(Fig. S6). In turn, the biases may be transmitted to the historical runs
and future projections of outdoor days to some degree (Fig. S7).
Nevertheless, the overall pattern of outdoor days is well captured by
the climate models. Note that, compared to the CMIP5 models, the CMIP6
models tend to better represent the observed spatial patterns of
temperature and outdoor days over the world for the historical period,
due to their remarkable improvements, in terms of spatial resolution,
physical processes, and biogeochemical cycles (Eyring et al. 2016).
Second, the human feeling of weather is complex and a widely subjective
matter, and therefore, the definition of mild weather is non-trivial
(van der Wiel et al. 2017). However, it can broadly be defined as
pleasant weather conditions allowing most people to enjoy outdoor
activities such as walking, jogging, cycling, or those related to
construction and tourism industries (Lin et al. 2019; van der Wiel et
al. 2017; Zhang 2016). Although we, here, defined outdoor days assuming
a range of temperature from 10 ℃ to 25 ℃, the exact range of temperature
or variable used does not significantly affect the global distribution
of the climate risk induced by changes in outdoor days (Fig. 10; Table
S3; see interactive visualization at
https://eltahir.mit.edu/globaloutdoordays/). In particular, considering
other variables, such as wet-bulb temperature (Fig. S8) and different
ranges of temperature resulted in a broadly similar pattern, supporting
the north-south disparity. Third, analysis for some specific countries
might show contrasting responses of outdoor days to climate change
within the different climate regions of a country as shown by Choi et
al. (2023) for the United States.