Using the algorithm I developed that's described on the Recent Weather page (and which I applied to the United States in an earlier post), here I extend the analysis globally using a gridded reanalysis product (NCEP). This data is a combination of all available stations, standardized to remove known effects like urban heat islands, and is produced at 6-hourly intervals. The algorithm I'm using computes comfort as a function of the heat index (if hot) or the wind chill (if cold), with the ideal conditions defined as being between 70 and 80 F, 24 hours a day. Notably, precipitation is not included -- though it may be in future iterations.
Doing this computation for world metropolitan areas with more than 2.5 million people (first figure below; click at top left to select a month to view) reveals some clear patterns. First and foremost, cities in the tropics (i.e. with nearly constant temperatures) at moderate elevations (i.e. are cooler than sea level) are comfortable all year long. This makes many of the South American cities rank very favorably. Spring in East Asia is pleasant (it's still relatively cool in North America and Europe), with nice summers particularly in coastal West Africa, then East Asia and Italy make appearances in autumn. Boston (July and August) is the only US city on the list, keeping San Diego from the stage in those months. November and December are omitted due to map limits, but are similar to March and February respectively. I also calculated the least-comfortable cities, though I don't show the map here; given the lack of large cities in the cold parts of the Southern Hemisphere, this list is studded with cold cities in boreal winter (Russia, northern China, and Canada) and then from spring through fall flips to tropical cities experiencing their dry season (e.g. India, West Africa) alongside perennially hot and moist areas (Indonesia, Singapore).
Then, I was interested in the question: given a particular location, when's the best time to visit it, according to this algorithm? Below is a first stab at an answer. In the map, reds and purples represent areas that are most comfortable in boreal winter (December-January-February) and greens represent areas that are most comfortable in boreal summer (June-July-August). Broadly speaking, the patterns track the warmest month in cool climates and the coolest month in warm climates. Note that this means cool ocean areas track SSTs and show up as September or October in the Northern Hemisphere, but this is interestingly not the case in the Southern Hemisphere. I speculate this characteristic may be related to the significantly stronger wind speed in the SH resulting in lower wind chills over the ocean during boreal-autumn storms, even if the SSTs are warmest then.
Because the map is based on a 2.5x2.5-degree grid, fine spatial details are not resolved. Nonetheless, one can still pick out regional features like the warmth of the Northeast-US corridor relative to inland (October vs August/September); the cool post-monsoon climate of southwest India; the island of cooler air over the Alps; the month-by-month progress of clear latitudinal bands of comfort stretching across Australia; and the dryness of western China relative to eastern, which results in the former being most comfortable in mid-summer while the latter is so in mid-spring and mid-fall. The data being of better quality than that of Mieczkowski 1985, and the algorithm more precise, in my view these results are a refinement of his work, and yet agree well with his "tourism climatic index" for January. While humans are capable of adapting to any climate, this comfort analysis is designed to reflect the ease of living in a particular location, climatologically speaking. It thus also is an index of the amount of energy needed to maintain equable indoor temperatures -- a place like Saudi Arabia, where 60% of summertime energy consumption goes toward air conditioning, clearly does not rank high on this list. With more and more of us on the planet, and continual discussions of how to achieve energy savings, perhaps some movement to the tropical highlands is in order. And, if further impetus is needed, consider that this is also where most of the world's coffee is grown.
Current American political dissatisfaction with the 2015 Paris climate accord hinges on its perceived economic risks to the industries that powered the country's 20th-century explosion in productivity and living standards for much of the population. While futurists promise everyone will be better off when the transition to cleaner and more-efficient energy is complete, many still fear those promises are hollow -- leading to a society dominated by overlords of the finance-technology-energy complex -- and are based on a false premise of imminent global catastrophe. Harborers of these reservations naturally advocate for applying the brakes to investments in measuring, preparing for, and adapting to climate changes of any stripe, and consider such efforts a gigantic boondoggle. Independent of climate projections, geopolitics, and the subtleties of national moods, I thought it worthwhile to take a moment to assess the indisputable evidence at hand for the observed connection between temperature and various economic measures. In other words, to point out there while there is a real cost to doing something about a warming world, there is also a cost to doing nothing.
The most-salient point is that there exists abundant evidence that high temperatures, like poverty, exact a kind of "cognitive tax" that seems to be an evolutionary adaptation to scarcity of a resource critical for survival. This temperature-productivity relationship has long been observed anecdotally, starting with Montesquieu in the 18th century (an "excess of heat" makes people "slothful and dispirited"). Quantitative studies have estimated that combating this heat with air conditioning results in increases in productivity from 5% (sedentary call-center work) to 25% (active kitchen work), and that the amount of outdoor labor -- from construction to gardening -- able to be performed during the hottest months could fall by half by the year 2200. Even in current conditions, outdoor laborers in Nicaragua opt to work about 20 fewer minutes per day (about 4%) when daily-maximum wet-bulb temperatures exceed the threshold of 26 C, according to public-policy graduate student Tim Foreman whom I spoke with at a recent conference. A listing of the results of some previous studies are shown in the figure below. Physiologically, these relationships are logical, given the body's reluctance to risk overheating itself and the decreasing effectiveness of our main cooling mechanism (sweat) at higher and higher temperatures.
On an international scale, for poorer (primary-resource-dependent) countries, every 1 C temperature increase is associated with a decrease in economic growth of 1.3 percentage points. A different study found productivity peaks at an annual-average temperature of 13 C, irrespective of a country's wealth, though it is only at daily maximum temperatures above 30 C that the relationship becomes meaningful. Thus warming is expected to help contribute to raising economic growth in cool countries, while subtracting from growth in warm ones. This is even the case within the United States, but as with the country comparisons, the definitive reason for this relationship's persistence despite electricity, air conditioning, and the increasing prevalence of indoor work is still at large.
Less expected, perhaps, is the link between temperature and mood, which is a nontrivial consideration on a personal level (e.g. in deciding where to accept a job or attend a school), but is rarely mentioned as it relates to society-wide impacts (e.g. with respect to possible changes in temperature or storminess in a particular region). I argue that this cost/benefit of climate change should be included in estimations of the overall effects. Leaving aside for a moment all other changes, the historical distribution of weather conditions sets a baseline for the culture of a region, month by month throughout the annual cycle, and this expectation may be considerably disrupted if monsoons, beach weather, snowstorms, etc. arrive at different times or not at all. It may be that this mood-weather relationship helps explain the observation that countries enduring clusters of repeated meteorological shocks are less able to recover on a unit basis than those for which the shocks are more evenly spaced out. And it's also likely a contributing factor to the expected erosion of economic gains in already-hot countries, which will face not only more-extreme heat but also more-severe storms.
All in all, it's abundantly clear that we not only shape the climate to some extent, but that the climate also shapes us -- our productivity, our mood -- in subtle yet unrelenting ways. It's even the case that temperature affects things as intimate, individual, and seemingly intrinsic as condom usage (and resultant HIV transmission rates). And this is not to mention the myriad other effects on our health. All of which goes to show that, especially going forward, the very term 'natural disaster' must be re-evaluated as regards anthropogenically-influenced changes in temperature, pollutants, and storms. While there may be reasonable debate about the location and magnitude of these changes, and the cost of the interventions intended to address them, there should be no doubt that they affect us all, no matter who we are or what we do.
This rather complex table summarizes the results of many earlier studies investigating the effects of temperature and pollutants on the quality of work performed in sedentary indoor settings, with strong negative effects reported especially for volatile organic compounds [VOCs] and high temperatures. Source: Mendell and Heath, 2005.
In billions of years,
Us. And this is what we do?
The world keeps turning.
In this post I'll recap the full year 2016 in the continental US — as well as winter 2017 — using comfort scores, as a supplement to the many other high-quality summaries out there that use more-traditional metrics (i.e. average temperature). Essentially, I intend it to capture the perceived pleasantness of a given 6-hour period, with temperatures between 70 F and 80 F considered ideal, and temperatures below 70 F or above 80 F accumulating 'discomfort' points in accordance with their wind chill or heat index respectively. Points are apportioned so that, for example, a wind chill of 30 F is considered roughly equivalent in comfort to a heat index of 92 F. This definition is applied to NCEP reanalysis data to produce the below figures and calculations. Additional comfort-score maps are found over on the Recent Weather page.
Averaged over the entire year (first figure below), it's clear that heat and humidity over the Gulf of Mexico combine to make it, and areas immediately adjacent to it, quite uncomfortable. In fact, due to high heat indices, these areas lead the country in discomfort three seasons of the year. With its lower moisture, despite high summer temperatures, most of the West is considerably more comfortable than the Southeast, with the sole exception of North-American-Monsoon-influenced southern Arizona. The entire Pacific coast, along with eastern New England, stand out as the most comfortable spots. In the northern tier of the country and into southern Canada, the annual-average discomfort is dominated by the winter months, and thus it increases going further north. In terms of individual cities, the most comfortable year-round are San Francisco (mild winters and cool summers) and Honolulu (cool winters and warm summers). The most uncomfortable are Brownsville, TX and Miami.
In the next figure, we can see the same data computed relative to the normal picture. This shows that 2016 was a warm year in nearly all of the country: the southern parts were less comfortable than normal (percentiles > 50), and the northern parts were more comfortable (percentiles < 50). Where the color scale maxes out in dark red — in the northern Gulf of Mexico and the western Atlantic — discomfort was record-high due to record high SSTs, whereas in the Pacific Northwest it was record-low, due mostly to a lack of cold wintertime temperatures in concurrence with the cloudiness and precipitation of the strong El Niño.
Plots for winter 2017 can be found on the Recent Weather page. In winter, moisture is low enough that it doesn't really play a role in affecting comfort, and so comfort is a temperature-only story. Thus, as is typical, the most comfortable areas were Hawaii and Florida, and the least comfortable the Upper Midwest across to the inland Northwest (with anomalous warmth in the inland Northeast taking it out of this competition in which it is normally a contender). The Southeast and lower Midwest were much more comfortable than normal due to persistent anomalous warmth that has also resulted in those areas having phenological spring arrive several weeks early. In contrast to last year, cool conditions have dominated the Pacific Northwest and made it less comfortable than normal (the conclusion also reached by the Weather Channel after a parade of storms there).
As in 2016, a recap and discussion of the past winter in terms of snowfall (also tracked regularly on the Recent Weather page) will happen once the snow stops falling at the highest ski resorts, probably sometime in May. Regular blog posts will of course occur in the interim, so keep an eye out for those!