American attitudes to changing weather, and therefore to climate change, have been analysed on the basis of US migration patterns since the 1970s. The findings have implications for the success of global climate policies. See Letter p.357
For decades, studies have described how climate change might have catastrophic global consequences. Yet motivating individuals to make climate-friendly decisions in daily life seems to be challenging, and insufficient action to mitigate climate change has been taken. On page 357 of this issue, Egan and Mullin1 report a study that might help to explain why. From an analysis of domestic human-migration patterns in the United States, the authors conclude that the net effects of climate change up to now are perceived by Americans to have been positive. Unless future climate threats are better understood and communicated, this perception may limit further action — until it is too late.
Estimates of the effects of future climate change cannot fully encompass the complexity of the changes. As the economist Nicholas Stern put it2, “It is these hard-to-predict impacts that are the most troubling potential consequences of inaction”. This might partly explain why the risks of climate change have been insufficient to provoke the general public to become a low-carbon-emitting society. It has been reported3 that technological innovation, 'green' jobs, and an understanding of the societal impact and health benefits of climate-change mitigation can stimulate public change through personal engagement — although other studies (see ref. 4, for example) report that such simple reframing is unlikely to gain much support from the public. But whether attitudes towards changing weather can inspire greater public action has yet to be proved, even if those attitudes become increasingly negative in the future.
In their study, Egan and Mullin estimated the US public's weather preferences by using an index score that measures the extent to which US migration patterns are associated with weather at different locations, adjusting for confounding factors using a previously reported method5. The index thus describes climate effects that are meaningful to people's daily lives, and which stimulate responses. It incorporates trade-offs between different weather conditions, taking into account seasonal variation that may influence migration.
More specifically, the index included information about January's maximum temperature, July's heat index (a measure that combines maximum temperature and relative humidity), July's mean relative humidity, annual precipitation and the number of days with precipitation per year5. The authors weighted their weather preference index (WPI) to better reflect the geographical density and distribution of populations across the United States. They did this using a method6 developed by geographers to determine the exposure of human populations to weather, in which the amount of exposure is scaled to the inverse of the distance of weather observations from population centres of US counties.
Egan and Mullin find that America's domestic-migration patterns reflect a general dislike for warm, humid summers but an appreciation of warm winters. This means that the effects of climate change since the 1970s are perceived overall to have been an improvement — potentially undermining public support for actions that limit climate change. Their findings were largely unaffected when data from different sets of studies were used to calculate the WPI. Notably, the authors also show that the warmer winters currently enjoyed by US populations (Fig. 1) probably foreshadow warmer (and therefore negatively perceived) summers in the future, and that future perceptions of the weather will probably become even more negative in scenarios that involve higher greenhouse-gas emissions.
A strength of the study is that it isolates the effect of long-term changes in daily weather on behaviour. But it does not incorporate the effects of weather extremes, which might be seen as a limitation. Public opinion, and potentially behaviour, is affected by extreme events and may catalyse support for emissions reductions7 — although such effects can be short-lived, as the authors have shown previously8.
The period studied in the retrospective part of the analysis is relatively short by the standards of climate studies, encompassing only 40 years. Over such short periods, naturally occurring climate variability can confuse the climate-change signal. Similarly, the reported signal cannot be differentiated into the effects of environmental changes that cause local warming, such as urbanization, and the outcomes of global phenomena related to the greenhouse effect. The prospective part of the study is robust — the climate-change projection takes into account the predictions of 35 climate models.
Another limitation of the study is that populations are assumed to remain in counties for which the WPI indicates public displeasure with climate. According to the definition of the index, such displeasure should contribute to migration out of those areas. The future, population-weighted change in the WPI is therefore likely to be less extreme than that reported, solely because people will move away from regions that develop unsatisfying climate conditions. On the other hand, the exaggerated population-weighted changes in the WPI might allow components of migration patterns to be estimated in future climate scenarios, assuming the weather preferences remain unchanged.
It matters globally if US populations feel that they have benefited from the effects of climate change up to now.
Egan and Mullin show that it may be hard for Americans to understand how changes in climate can be a catastrophe when such changes apparently make daily life more pleasant. The United States is one of the most influential countries in global policymaking, and one of the greatest emitters of greenhouse gases9. It therefore matters globally if US populations feel that they have benefited from the effects of climate change up to now. Moreover, many Americans may not recognize that these effects will change, as Egan and Mullin show, and that even if emissions are reduced now, it will be a long time — not until roughly the middle of the twenty-first century10 — before climate alters in response.
The study's findings might also be relevant to the attitudes of European populations. In colder regions, such as in Canada, Russia and China, future winter warming may also be perceived as a benefit of climate change, but the WPI would need to consider the effects of climate change on snowfall and ice formation, and how people respond to those changes. It would also be interesting to examine whether attitudes to climate change are different in tropical or subtropical regions of Asia and Africa — perhaps an initially positive experience has already reversed in some areas, causing displeasure or emigration from affected regions. More broadly, studies that provide insight into the public experience of weather will help inform global climate policy.Footnote 1
Egan, P. J. & Mullin, M. Nature 532, 357–360 (2016).
Stern, N. Nature 530, 407–409 (2016).
Watts, N. et al. Lancet 386, 1861–1914 (2015).
Bernauer, T. & McGrath, L. F. Nature Clim. Change https://doi.org/10.1038/nclimate2948 (2016).
Rappaport, J. Reg. Sci. Urban Econ. 37, 375–398 (2007).
Hanigan, I., Hall, G. & Dear, K. B. G. Int J. Health Geogr. 5, 38 (2006).
Spence, A., Poortinga, W., Butler, C. & Pidgeon, N. F. Nature Clim. Change 1, 46–49 (2011).
Egan, P. J. & Mullin, M. J. Polit. 74, 796–809 (2012).
Matthews, H. D. et al. Environ. Res. Lett. 9, 014010 (2014).
IPCC. Climate Change 2013: The Physical Science Basis. Working Group I Contribution to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (eds Stocker, T. F. et al.) (Cambridge Univ. Press, 2013).