Resettlement, like migration, is a product of fragility. Unlike migration, it's typically coordinated and forced, rather than sporadic and voluntary. The fragility that spurs it can be sociopolitical (as with harried minority groups), economic (as with smallholder farms during the Great Depression), climatic (as along the Louisiana coast), or any combination of these. Where climatic fragility leads to economic fragility, making climatic and economic incentives consonant, self-initiated movement is most likely, even in the absence of a disaster per se. But typically, it takes a disaster, and even then few people want to leave their homes and neighborhoods if at all possible. This post examines resettlement in the 20th century United States and its implications for the future we all face, which includes the challenges of extreme heat, sea-level rise, and stronger storms (among others).
To start, it's important to put resettlement in context. In a previous post, I explored the issue of climate-induced migration and resettlement in comparison with the adaptation measures that are more conventionally favored by planners. 'Adaptation' encompasses everything from human behavior to upgraded electrical grids and taller seawalls, and this will be sufficient for most of the population. The risks and costs of climate change are of course unevenly distributed geographically, but such is the raison d'etre of government -- to smooth out expenses and revenues across the full population, so that projects can be undertaken that are too complex, expensive, time-consuming, or unmarketable to fall within the province of private-sector activity. But of course government redistribution has its discontents, and particularly pertinent in the climate context is the disconnect between behaviors generated by current incentives and optimal behaviors from a utilitarian point of view. The basis of the objection is a simple economic axiom: the lower the cost of a good, the more of it is demanded, a prototypical example being taxpayer subsidy of the National Flood Insurance Program incentivizing the development of flood-prone areas. For the foreseeable future, much of this shared risk will probably remain politically and financially tolerable. But consider the most extreme cases: the sinking island, the isolated canyon community... these are the margins where serious public conversations need to be had about whether and how governments should provide an avenue (and perhaps additional incentives) for residents to relocate to less-fragile locations. The government thumb is already on the scales affecting where people live, in ways as varied as the infamous mortgage-interest deduction, the aforementioned NFIP, and the commitment to footing the entire bill for wildfire suppression without passing on any of the costs to homeowners. In the best case, climate-related challenges force a clear-eyed look at the purpose and consequences of these kinds of subsidies. If the major externalities are not internalized, market distortions and the resulting inefficiencies will continue indefinitely.
Although inexact, the closest parallel to the present-day issues posed by ongoing climate trends is probably the 1930s Dust Bowl and Great Depression. There was in fact a short-lived "Resettlement Administration", under the umbrella of the New Deal, whose aim was to resettle farmers as a response to rural poverty brought on by a combination of antiquated agricultural practices and adverse environmental conditions (heat and drought). As the former program administrator described it, the driving force of the program was explicitly economic, with the aim of addressing the suboptimal productivity of labor and land that the small struggling Great Plains farms represented. Buyouts were extensive, but the new suburban communities for the displaced that were envisioned were never realized on any consequential scale. Despite support from many small farmers, these new developments faced ultimately crippling opposition on both financial and sociocultural fronts. This provides a valuable lesson for the present, in that it highlights the importance of considering where people will go and how they will be assimilated into a new community, beyond simply getting them out of harm's way. An alternative is to move a group of people wholesale from one place to another, as happened with both Native Americans and African Americans. That neither example was terribly successful points to the critical role of economics, in addition to community cohesion.
An idealized 1920s vision of the progressive optimization of land use in New York State, with substantial implied movement of people from the less-productive to the more-productive areas. The report argued for government-managed resettlement to speed along this process and improve the economic fortunes of the poor in the less-productive areas, even without the motive force of climate-related disasters. Source: Jacobs 1989, quoting Mumford 1925.
At the same time, state initiatives were begun to alleviate the primarily economic hardship being suffered by smallholder farmers in the Northeast. With advances in technology and transportation stemming from the Industrial Revolution, their farms were no longer economically viable due to inherent climate and topographic factors, but many were trapped by debt. A decades-long political debate ensued between the 'social Darwinists' (advocating resettlement of the affected) and the 'repopulationists' (advocating investments allowing them to adapt to the changing times). The former group prevailed and buyouts began in earnest in the early 1930s in New York State under the governorship of FDR, who saw them as efficient ways to manage growing rural malaise. This is one of the few examples where a potential humanitarian disaster was foreseen and successfully avoided. However, they lasted just a few years before funding was cut back to skeleton levels.
A humbling lesson from these past efforts is that even in an era of substantial climatic stress and economic upheaval, these programs' achievements were ultimately narrower than their original visions. Key elements that seem necessary for long-term success are the preservation of some kind of social fabric, a plan for economic integration into the new community, and having community buy-in from the outset. A reliable and continuing revenue stream is also necessary from a programmatic point of view.
Looking to the future, the social-fabric issue seems particularly pertinent given rising inequality, especially in urban areas. Exacerbating it is the fact that, even without climate stressors, the wealthy and highly educated are much more mobile and have more transferable skills, so that the economic and social shock of resettlement is smaller for them. This means that the poor are often forced to make larger changes in response to climatic threats, despite not having the resources to comfortably do so. Subsistence farmers are arguably the worst off, being highly vulnerable to climatic whims, having the fewest resources, and having a livelihood which depends on the scarce resource of land (as opposed to practicing a trade or service-based skill). Fortunately, some efforts are underway in the international-policy sphere to smooth out the expected necessary transitions -- in fact, there's even a non-profit organization expressly dedicated to facilitating climate-related displacements. Climate and social-science researchers are brainstorming solutions, and localized efforts have been implemented and largely proven successful, though in all likelihood large-scale initiatives will only begin when the true scale of the issue comes into focus. While not something that's likely to affect most of us, the exigencies of resettlement will have a large-enough impact on those who do experience them that we all should be interested in finding efficacious and cost-effective solutions. While climate change exists on a bigger scale than anything seen before in human history, no disaster has ever been purely natural (including the Dust Bowl), which should give us heart that the answers are not out of the realm of experience either.
I recently attended the Rutgers Climate Symposium, where this year's theme was 'cities and climate' — especially appropriate for an event situated in the USA's finest megalopolis. A key takeaway from the day was the degree to which -- due to their high concentration of people, money, and physical assets -- cities are at the front lines of climate- and environment-related issues, more so than rural areas, provinces, or countries. One speaker, Bill Solecki, listed a number of unique challenges cities have faced in the last few centuries, which have been encountered and dealt with largely in serial fashion: fire, sanitation, air pollution, open space, and urban decay. From this perspective, sea-level rise or extreme heat is just another challenge that will mainly be felt first in cities, and that will strain the resources of some, but likely fail to change the larger arc of history that bends toward urbanization. The speaker Julie Pullen noted that this trend represents an opportunity across much of the globe to effect major change in urban form and lifestyles as the century progresses.
In fact, this positioning of cities on the front lines of global issues has been the case for about as long as cities themselves have existed, if for no reason other than that local governments are typically and of necessity concerned with more-prosaic matters than national leaders, who have more of a tendency to helicopter in, make a pronouncement, and helicopter out (hence Fiorello LaGuardia's quip that 'There's no Democratic or Republican way to pick up the garbage'). The greater complexity of cities than suburban or rural areas inspires and motivates a greater complexity of governance, and this fact is largely implicated in the strikingly complete sorting of the US (and many other places) into liberal urban and conservative rural halves.
Whatever their political leanings or economic status, another emergent theme was the need for municipalities to work together to tackle problems too big for any one of them to face alone. Not every city can afford an office of environment or climate, nor is there always the political will for it. Leveraging existing knowledge, financial, and political networks brings down the cost associated with addressing climate-related problems through economies of scale, and consequently reduces the potency of the argument that a given solution is too expensive or cumbersome to implement. For example, the various cities in a metropolitan area could all use information from the same study to assess their vulnerability to sea-level rise; could all finance seawall improvements through the same pooling they use to fund transportation infrastructure; could all coordinate these actions through the same conference of mayors they all belong to; etc. Organizations like the Global Covenant of Mayors and the Urban Climate Change Research Network also serve as conduits for resources and information while still allowing cities to take their own individualistic problem-solving approaches.
As a result, one of the main lessons of the symposium and of the most-recent Urban Climate Change Research Network report was that the most effective climate-related consortia all consider their programs as elements of larger programs or strategies, rather than being viewed ipso facto. Another aspect concerns disadvantaged populations: in environment-related issues as in political and financial ones, and in the developed world as in the developing, these communities are where there is the most room for improvement in terms of communication and implementation of cost-effective measures to address problems (whether done proactively or reactionarily). Thus, these populations must be explicitly focused on for substantial equity strides to be made.
Lastly, integrated and participatory decision-making around urban climatology is crucial for making decisions that are good (as these require the input of various leaders and organizations) and also decisions that are actually implemented (as this only happens when the aforementioned people consider the decisions legitimate and accurate). As speaker Cynthia Rosenzweig emphasized, this is exemplified by, essentially, keeping an open mind rather than letting pure science dictate the decisions. Risk reduction, reaction, and long-term planning should be considered hand-in-hand. It's doubly important to keep this integrated framework in mind in the frenzied weeks and months after an extreme event, which is when much of the political energy and financial muscle are available to enact changes. More broadly, Bill Solecki stressed that the intersection of extreme events, climate change, and urban areas is where there's the motivation and wherewithal for society-wide adaptation measures, and where these measures would have the biggest impact. And yet, they also illustrate one of the most fundamental challenges in applying the results of urban-climatological studies: that while we have lots of data on roughly what's expected in terms of climate in the next 100 years, we don't know what to do about it -- that is to say, we don't know how to react viscerally, and certainly not how to plan in an organized and effective manner. Which is exactly why these kinds of mutual brainstorming sessions between urban climatologists and everyone else with a stake in the future of cities in the 21st century are so important.
Note: this text was written for the website of the Consortium for Climate Risk in the Urban Northeast, where it will also be published by and by. The subject matter complements nicely many of the previous posts on this site, in the general sense of linking climate and social topics and particularly in exposing their deep interconnectedness.
Global sea level is trending inexorably upward, and at an increasing rate (Figure 1). This acceleration is expected to continue over the course of the 21st century, though with wide uncertainty (Figure 2) — an uncertainty proportionally much wider than that for global-average temperature, on which it’s loosely dependent. In this post we explore some of the reasons for the large uncertainty in projected sea-level rise [SLR], and CCRUN’s contribution to the scientific effort aimed at understanding and whittling away at it.
Figure 1: Increase in global-average sea level since 1700, estimated from salt marshes across the world (light purple), various tide gauges (dark purple, green, and orange), and satellite altimetry (light blue). These sea-level changes are calculated with respect to the geoid, i.e. local effects of uplift and glacial isostatic adjustment have been corrected for so the time series can be meaningfully compared across regions. Source: Church et al. 2013.
Figure 2: A timeseries of estimated global-average sea level, encompassing projections with a central range (dark gray shading) and wider uncertainty bounds (light gray shading) as well as a simplified time series starting in the year 1800. Source: Melillo et al. 2014 and Parris et al. 2012, via the National Centers for Environmental Information web portal (https://statesummaries.ncics.org/ny).
There are three primary contributors to SLR: ice melting, thermal expansion, and changes in land height. The first two are mostly global in nature, whereas the last depends on local geological forces. Global-average sea level is rising about 3-4 mm/yr, and the majority of this is due to thermal expansion — as water gets warmer, it gets less dense, i.e. the same mass takes up more space. Increases in ocean volume due to melting ice account for another 1 mm/yr or so. In CCRUN’s focus area of the Northeast US, additional local SLR of about 1 mm/yr is attributable to the continuing readjustment of the land to the freeing of the burden that the North American ice sheet imposed on it until about 15,000 years ago (Figure 3). This “glacial isostatic adjustment” means that Canada and the northernmost tier of the United States are rebounding, while areas on the periphery of the glacier, such as most of the US East Coast, are compensating by sinking. And the ice-weighting effect is not trivial: bedrock under the Antarctic Ice Sheet is currently depressed 500-1000 m by the weight of the ice (Anderson 1999). Like a bathtub being warped, this results in apparent sea-level changes independent of increases in the total volume of the oceans.
So what does the future hold? We know for certain that thermal expansion will continue apace, considering that it takes at least 1000 years for the ocean to come into thermal equilibrium with the atmosphere (Goodwin et al. 2015) – in other words, if we stopped emitting greenhouse gases today, the ocean would keep heating up for centuries. ocean volume will undoubtedly continue increasing in a warmer world -- but how much? The question is a thorny one because it’s not only the amount of melted ice that will determine future SLR, but the source: ice sheets are massive enough to have an appreciable gravitational effect on the oceans, so taking them away means a decrease in sea level nearby. Currently, Greenland is contributing somewhat more to SLR than Antarctica (Velicogna et al. 2014). As a consequence, North Atlantic SLR is slower than the global average (Figure 4). Projections that Greenland will continue to rapidly melt mean this regional dynamic could persist, so long as there are no big surprises from West Antarctica. Another Greenland-related factor also contributes to the Northeast US being a global hotspot of SLR uncertainty: the effect of its melting on the Atlantic Meridional Overturning Circulation (Little et al. 2015b). A slowdown in AMOC, resulting for example from a meltwater pulse, would tend to increase sea level in the North Atlantic (Goddard et al. 2015). Another significant dynamical factor is the North Atlantic Oscillation, which in its positive phase has southwesterly winds along the Northeast coast and consequent marginally lower sea level – any long-term changes in the NAO could thus have an impact on the local average rate of SLR as well.
However, these regional variations pale in comparison to the potential global increases stemming from Antarctic melting. The total amount of ice on Earth is enough to raise sea level about 80 m, with East Antarctica accounting for about 65 m of this. Most research suggests East Antarctica will experience little melting as far as we can predict, and that West Antarctica and Greenland will retain ice for at least several hundred years. But these are suggestions rather than guarantees. While our species and many others have proven on the whole able to adapt to the continental-scale changes in coast locations, vegetation, and temperatures associated with glacial-interglacial cycles (Grant et al. 2014), these last occurred when humans were few in number, nomadic, and nearly possessionless, rather than having $1.5 trillion in immobile housing investments on the glacial moraine of Long Island alone (U.S. Census Bureau 2016).
With all these social, ecological, and economic motivations in mind, CCRUN-affiliated scientists are working to estimate SLR on varying timescales. Radley Horton helped develop a model to predict the likelihood that a given coastal area will likely be able to adapt (by ecologic, depositional, or anthropogenic means) to seas as they rise, or whether the coastal ecosystem will simply become submerged (Lentz et al. 2016). For example, for a variety of reasons a marsh accrete sediment and remain marshy, or may “fall behind” SLR and disappear under the waves. The authors found that some locations are much more dynamic than others, allowing them more leeway to effectively maintain their ecosystem status quo (e.g. actively depositional barrier islands and beaches), whereas those that are more passive (e.g. marshes and impermeable surfaces) are most likely to be inundated (Figure 5). With relatively modest amounts of SLR, it’s presumed there will be successful anthropogenic efforts to keep most developed areas dry.
Figure 5: Predicted effect of sea-level rise on the coastline of Plum Island, MA, including the coast’s feedback response, in the 2080s relative to the present-day. Red areas are most likely to be functionally the same, while dark blue areas are currently above mean sea level but are likely to be flooded in the future. Source: Lentz et al. 2016.
The Lentz et al. study, however, did not take into account the possible effect of stronger storms lashing the beaches and eroding away any progress they might have made in keeping pace with SLR. However, two earlier CCRUN papers did just that (Horton et al. 2015; Little et al. 2015b). As it turns out, SLR and tropical-cyclone intensity are highly correlated in climate models, and including this correlation results in a longer tail of future flooding extremes than would otherwise be predicted (Figure 6). A large fraction of this correlation can be attributed to the underlying factor of upper-ocean warming. While tropical cyclones are more powerful, “Nor’easters” occur more frequently and often move more slowly, increasing their overall flooding potential (Horton et al. 2015). Many storm-specific factors affect the location and extent of flooding from a given storm, including wind speed, wind direction, storm speed, storm direction (i.e. a storm coming from the east will pile up more water on an eastward-facing coast), precipitation intensity, total precipitation, and storm timing relative to the daily and bimonthly tidal cycles. Because coastal inundation is a function of storm flooding overlaid on mean SLR, long-term changes in any combination of these factors can ‘stack the deck’ in terms of exacerbating or counterbalancing the effects of mean SLR. Recent trends have been upward for both frequency and intensity in recent decades, with the majority of evidence suggesting the strongest storms will get even stronger (Horton and Liu 2014).
CCRUN researchers are looking not just at the regional scale — using regional and global climate models, and leveraging correlations and other statistical tools — but at the level of the individual street. In this vein, a new nested model was designed specifically for New York Harbor (Blumberg et al. 2015). This enables questions to be answered about the very local impact of a given storm, sediment-dredging project, etc., and thus to make recommendations as to the most cost-effective measures to dampen waves and storm surges along complex coastlines. Such models are critical tools for assessing the local impact of a regional or global trigger such as accelerated ice-sheet melting. As one might expect from the highly varied times of tides even within a single short stretch of coastline, near-shore dynamics are inordinately complex and can result in feedbacks both positive and negative stemming from an initial perturbation. This is doubly true where humans have so highly modified the coastline, as in the Northeast US. Lack of awareness of this kind of nonlinear response was part of the reason for the unexpected severity of the flooding during Hurricane Katrina in New Orleans after some of the levees broke.
In any system that evolves dynamically, projections inherently get more uncertain further and further out into the future. In the case of SLR, there’s not only the atmosphere, ocean, and cryosphere changes to consider, but also the iterative human response to these changes. And in addition to the mean, storm surges, storm tides, and large waves are additional complicating factors. While the 90th-percentile estimate of SLR by 2100 in New York City is about 75” (Horton et al. 2015), the analogy with insurance purchased in case of a devastating but rare event should be kept in sight. And this is particularly so for a region where finance and insurance have long been a pillar of economic success, alongside maritime industries like shipping and whaling, and where property values are among the highest in the country (Figure 7).
To extend the finance analogy, in stocks the volatility index reflects investor sentiment about upcoming uncertainty in the markets. If such an index existed for the climate, it would surely be increasing, due largely to anthropogenic influences which are making it harder to predict the future even as understanding of the climate per se continues to get better and better. The most common investor response to expectations of high volatility is to pull out and wait for conditions to improve, but the Earth is a market that can’t be pulled out of, and that won’t recover naturally from an unnatural forcing. Thus CCRUN’s philosophy is to face such problems and their uncertainty head-on, with all the tools and information that can be mustered. Understanding issues of this magnitude, whether climatic or otherwise, is always a strong value proposition.
Anderson, J., (1999), Antarctic marine geology. Cambridge University Press, ISBN 0-521-59317-4.
Blumberg, A. F., N. Georgas, L. Yin, T. O. Herrington, and P. M. Orton (2015), Street-scale modeling of storm surge inundation along the New Jersey Hudson River waterfront. J. Atmos. Ocean. Tech., 32, 1486-1497, doi:10.1175/jtech-d-14-00213.1.
Church, J. A., et al. (2013), Sea-level change. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press.
Goddard, P. B., J. Yin, S. M. Griffies, and S. Zhang (2015), An extreme event of sea-level rise along the Northeast coast of North America in 2009-10. N. Commun., 6:6346, doi:10.1038/ncomms7346.
Goodwin, P., R. G. Williams, and A. Ridgwell (2015), Sensitivity of climate to cumulative carbon emissions due to compensation of ocean heat and carbon uptake. Nat. Geosci., 8, 29-34, doi:10.1038/ngeo2304.
Grant, K. M., et al. (2014), Sea-level variability over five glacial cycles. Nat. Commun., 5 (5076), doi:10.1038/ncomms6076.
Hay, C. C., E. Morrow, R. E. Kopp, and J. X. Mitrovica (2015), Probabilistic reanalysis of twentieth-century sea-level rise. Nature, 517, 481-492, doi:10.1038/nature14093.
Horton, R. M., and J. Liu (2014), Beyond Hurricane Sandy: What might the future hold for tropical cyclones in the North Atlantic? J. Extr. Even. 01, 1450007, 30 pp., doi:10.1142/s2345737614500079.
Horton, R. M., C. Little, V. Gornitz, D. Bader, and M. Oppenheimer (2015), New York City Panel on Climate Change 2015 Report Chapter 2: Sea-Level Rise and Coastal Storms. Ann. N.Y. Acad. Sci., 1336, 36-44, doi:10.1111/nyas.12593.
Lentz, E. E., E. R. Thieler, N. G. Plant, S. R. Stippa, R. M. Horton, and D. B. Gesch (2016), Evaluation of dynamic coastal response to sea-level rise modifies inundation likelihood. Nat. Clim. Change, 6, 696-701, doi:10.1038/nclimate2957.
Little, C. M., R. M. Horton, R. E. Kopp, M. Oppenheimer, G. A. Vecchi, and G. Villarini (2015), Joint projections of US East Coast sea level and storm surge. Nat. Clim. Change, 5, 1114-1121, doi:10.1038/nclimate2801.
Little, C. M., R. M. Horton, R. E. Kopp, M. Oppenheimer, and S. Yip (2015), Uncertainty in twenty-first-century CMIP5 sea-level projections. J. Clim., 28, 838-852, doi:10.1175/jcli-d-14-00453.1.
Melillo, J. M., T. C. Richmond, and G. W. Yohe, Eds. (2014), Climate change impacts in the United States: The Third National Climate Assessment. U.S. Global Change Research Program, 841 pp., doi: 10.7930/J0Z31WJ2.
Parris, A., P. Bromirski, V. Burkett, D. Cayan, M. Culver, J. Hall, R. Horton, K. Knutti, R. Moss, J. Obeysekera, A. Sallenger, and J. Weiss (2012), Global Sea-Level-Rise Scenarios for the US National Climate Assessment. NOAA Tech Memo OAR CPO-1. 37 pp.
Spiser, M. (2015), Resizing America based on property values. Business Insider. http://www.businessinsider.com/resizing-america-based-on-property-values-2015-7.
U.S. Census Bureau (2016), 2011-2015 American Community Survey 5-Year Estimates. https://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml?src=CF.
Velicogna, I., T. C. Sutterley, and M. R. van den Broeke (2014), Regional acceleration in ice mass loss from Greenland and Antarctica using GRACE time-variable gravity data. J. Geophys. Res. Space Physics, 41, 8130-8137, doi:10.1002/2014gl061052.
Back in July 2016 I first laid out the logic and methodology for computing "comfort scores" based on the deviation of temperatures on each day from the range considered ideal for human comfort, and I expanded upon this more practically in June 2017. (Charts for the most-recent periods are found in the Recent Weather section.) The true discomfort of low temperatures is estimated from the wind chill, and of high temperatures from the heat index. I've since made several refinements to the calculation. These include introducing a diurnal cycle to the range of ideal temperatures, so that now 64-70 F is considered ideal at night, and 70-80 F is considered ideal during the day. I also have run the calculation for North American Regional Reanalysis data to get higher-resolution results for the United States. The most-substantial change, however, is the inclusion of precipitation rate as an additional variable, along with the original ones of temperature, humidity, and wind speed. This means the discomfort is calculated as a sum of contributions from heat, cold, and wetness. From experience we all know that the effect of precipitation on comfort varies with temperature — cold rain is much more unpleasant than warm rain, and at very hot temperatures rain is welcome, as long as it's not very heavy. These relationships are captured in the below diagram illustrating the bivariate nature of the precipitation component of the discomfort score.
With this new definition, I recalculated the scores for the entire world using NCEP reanalysis, and for the continental US using NARR reanalysis. In the galleries below are the new climatologies for discomfort scores (red being most uncomfortable, blue most comfortable), encompassing the annual average as well as seasonal averages for winter, spring, summer, and fall, for the world (top) and the US (bottom). Key features include the complete unsuitability of Antarctica for human comfort, the appeal (climatologically speaking) of the subtropical ocean gyres, and the relative comfort of much of the tropics, especially those areas at several thousand feet elevation such as southern Africa, Central America, and Southeast Asia. Nonetheless, the hottest and most-humid areas (the northern Indian Ocean, western Pacific, and coastal Arabia) see summertime discomfort levels much greater than those elsewhere in the tropics, and in fact they approach those of Antarctica in austral summer (but are nothing like those in austral winter). Due to the inclusion of precipitation as discomfort-score factor, there is a greater value placed on dryness, resulting in the improved performance of the western US relative to the eastern. This makes the Southwest and Central Valley the most comfortable places on an annual-average basis, and even in summer they are much more comfortable than the humid Gulf of Mexico and Gulf of California.
I also used the revised discomfort-score climatology to compute time series over the period of available reanalysis data, 1979-2016. The next figure shows the trend in contributions to globally averaged discomfort for heat and cold. As the world has gotten warmer, discomfort has come more and more from heat rather than cold. Individual hot years also stand out, such as the El Ninos of 1983, 1998, and 2016. Note that discomfort is a nonlinear function of temperature due to the compounding effects of humidity and wind speed, as well as the threshold effects introduced by the calculation methodology. Regardless, a clear change is evident since about 1995, and these trends will likely continue as more locations experience uncomfortable heat while simultaneously experiencing less uncomfortable cold.
The last figure shows the same information, but broken down into latitudinal averages. Shown are percentiles relative to the 1979-2016 average for that latitudinal band; for example, a year falling in the 90th percentile means it was more uncomfortable overall than 90% of years in that location. Discomfort in the tropics and subtropics is primarily a function of heat, whereas in the mid-latitudes and polar regions it is a function of cold. Therefore, a trend toward greater comfort is expected poleward of 30 deg and toward less comfort equatorward of that — and that's exactly what is shown here. The last two years have exhibited record discomfort in the tropics and record comfort at the poles, which is again a trend very likely to continue.