North America Archives

Deeper Problems with Maps of “Excessive Drinking” in the United States

Article-Grid, Cultural Geography, North America, psychoactive substances / Martin W. Lewis

Oddly, maps of “excessive drinking” in the United States do not correspond particularly well with maps of per capita alcohol consumption. New Hampshire, Delaware, and Nevada, for example, are typically depicted as having the country’s highest levels of per capita alcohol consumption but only moderate levels of excessive drinking. In contrast, Iowa is usually depicted as having one of the highest levels of excessive drinking but only a moderate level of total alcohol consumption. (Unfortunately, maps of per capita alcohol consumption by county are not available.)

Maps of Excessive Drinking and Total Alcohol Consumption 1

Maps of Excessive Drinking and Total Alcohol Consumption 2

Such a seeming discrepancy can easily be explained by differences in drinking behavior. It is possible, for example, that on average Nevadans drink more than Iowans but generally do so in a more responsible manner, engaging less in “binge drinking.” Possible yes, but probable? I doubt it. I think it is more likely that the data used to make these maps are flawed. In the remainer of this post, I will critically examine the sources used for maps of “excessive drinking.” Admittedly, much of the evidence that I present was derived from “conversations” with AI; as such, my arguments should also be regarded with some skepticism.

The foundation of excessive-drinking maps is the data tabulated by the CDC’s Behavioral Risk Factor Surveillance System (BRFSS), derived from state-based telephone surveys. Sample sizes and exact questions, however, evidently vary by state and year, complicating comparisons over time and space. More significant, honest answers about such problematic behavior cannot always be expected. Supposedly, surveys routinely recover only a fraction of the amount of alcohol that sales data imply, often on the order of 30–60%. Such underreporting supposedly stems from forgotten drinks, underestimated pour sizes, and nonresponses by heavier drinkers. Many respondents, deemed “high impression managers,” want to appear respectable by reporting that they drink less than they actually do. According to one academic study, “high impression managers report 20 to 33% less consumption and are about 50% less likely to report risky drinking.” Such considerations worrisomely suggest that actual rates of excessive drinking are considerably higher than they appear to be on these maps.

But it is also known that the bias can run in the opposite direction, particularly among certain groups of college students for whom heavy drinking can be seen as a badge of belonging. Is it therefore coincidental that counties that are often mapped as having the highest rates of excessive drinking also have large college-student populations, such as Missoula and Gallatin in Montana and Dane in Wisconsin? More to the point, more under-reporting is to be expected in areas in which alcohol is deeply stigmatized (Utah and eastern Idaho; interior regions of the U.S. South) than in areas with alcohol-tolerant cultures (the upper Midwest and much of the interior West).

The definition of “excessive drinking” used in these surveys are also problematic. The CDC definition seems straightforward: either “binge drinking” (four or more drinks on one occasion for women, and five or more for men), or “heavy drinking” (eight or more drinks per week for women, and fifteen or more for men). Although the first threshold seems reasonable, the second is questionable. I doubt that most people would regard a woman who drinks a glass of wine with dinner every day, and two on Saturdays, as an “excessive drinker.” By the same token, is a man who drinks two beers a day, and three on Fridays, really a “heavy drinker?” Perceptions obviously differ widely from person to person. More important, actual alcohol consumption depends on what kind of beer the drinker in question consumes. A “light” beer, for example, typically has less than half the alcohol content of imperial IPA of the same size.

A similar problem is encountered with the concept of “binge drinking.” “Five or more drinks on one occasion” might seem like a high figure, but it depends on how long an “occasion” lasts. The standard figure for research purposes is two hours. This duration is used because a typical man who consumes five standard drinks (or 0.6 fluid ounces of ethanol) in two hours will reach a blood-alcohol concentration over 0.08%, which counts as legally “drunk” in most U.S. states. But the surveys on which excessive-drinking data are based do not seem to be rigorous in this regard. According to AI, “respondents are not walked through a two-hour clock. So five beers over a six-hour dinner can be coded the same as five shots in an hour. That is a real measurement gap, not a quibble.”

For both “heavy drinking” and “binge drinking,” individual weight and muscle-mass should also be considered. All other things being equal, a heavier person must drink more alcohol than a lighter person to reach the same blood-alcohol concentration, which is the real measurement of intoxication. A muscular person, moreover, will have a lower blood-alcohol concentration than a less-muscular person of the same weight who drinks the same amount of alcohol over the same period of time.

Ideally, alcohol-consumption surveys would account for such factors, but that is not necessarily the case. Considerations of weight and muscle mass are ignored, and a twelve-ounce beer is usually treated as standard unit regardless of alcohol content. After much back-and-forth, my AI “informant” concluded by telling me that the definition of excessive drinking employed by the CDC’s Behavioral Risk Factor Surveillance System (BRFSS) “is operational and crude on purpose. It is useful for comparing places and tracking a risk category. It is a poor description of how drunk someone got on a given night or how much alcohol they actually swallowed.”

Finally, alcohol-consumption surveys face the same problems of sample representativeness that bedevil all pollsters. Many if not most persons refuse to respond to such surveys, and it cannot simply be assumed that non-responders have the same experiences as responders. Researchers are well aware of this problem and try to compensate, but it is not clear whether they do so adequately when it comes to “excessive drinking.” According to AI:

BRFSS tries to account for refusal and uneven participation—but only in a limited way. It does not fully correct for the possibility that drinkers and non-drinkers refuse at different rates. Most people who are called never complete an interview. Recent median state response rates are around 44% (roughly 31–65% by state). That is normal for large telephone surveys now, not a secret. It assumes that, within those groups, people who answer the phone drink like people who hang up. Such an assumption is questionable.

All of this is not to argue that maps of excessive drinking in the United States are worthless. I find them interesting and regard them as partially valid. But they are more suggestive than definitive. I would also prefer that they be give more modest titles, such as “Rates of Excessive Drinking as Indicated by Public-Health Surveys.” A brief note on how “excessive drinking” is defined would ideally be included on all such maps.

Deeper Problems with Maps of “Excessive Drinking” in the United States Read More »

Could U.S. Counties with Low Levels “Excessive Drinking” Really Have High Levels of Alcohol-Induced Mortality and Low Levels of Health?

Article-Grid, Cultural Geography, Montana Atlas, North America, psychoactive substances / Martin W. Lewis

The map of excessive drinking in the United States that was recently discussed in GeoCurrents was originally posted on IntoxiStates.com and is based on information from the CDC’s Behavioral Risk Factor Surveillance System, as packaged by the University of Wisconsin’s Population Health Institute County Health Rankings & Roadmaps (CHR&R). It clearly indicates that Wisconsin has the country’s highest rate of excessive drinking and overtly states that among the many counties of the U.S., Gallatin in Montana occupies the top place. The IntoxiStates map is not dated, but a 2025 map based on 2022 data from the same source paints a different picture. Published in the The Hill, this map puts Montana in the top position and depicts Iowa as also having a higher rate of excessive drinking than Wisconsin. The county-based map from the same article indicates that Missoula in Montana is the ‘drunkest” county, with a 29% excessive-drinking rate. The most recent data from CHR&R’s own website puts both Missoula and Madison counties in Montana ahead of Gallatin, but just barely (26% vs. 25%). The “excessive drinking” data used to make these maps, in other words, change from year to year. More to the point, excessive drinking in Wisconsin does not seem to be nearly as pervasive or unusual as it appears to be on the IntoxiStates map.

The Hill Excessive Drinking Counties map

Montana Excessive Drinking by County map

The CHR&R’s Montana map is particularly interesting, as it reveals a clear county-level correlation between relatively low rates of excessive drinking and high percentages of indigenous residents. This association is significant, as it contradicts a widely held stereotype. It might not, however, be valid. It would be expected that low levels of excessive drinking would be associated with low levels of alcohol-induced mortality, but that is not what the Montana data indicate. A map of alcohol-induced mortality in Montana in the Covid-year 2021, derived from a map published in a PLOS Global Public Health article, shows a surprising negative relationship between excessive drinking and death by drinking. Another pair of maps indicates that Montana counties with high rates of excessive drinking are significantly healthier than counties with low rates. (Correlation, of course, does not imply causation.)

Montana Excessive Drinking Indigenous Population maps

Montana Excessive Drinking and Alcohol-Induced Deaths map

Montana Excessive Drinking and Health map

Montana does not seem to be alone in this regard. A pairing of the IntoxiStates map of excessive drinking with the PLOS Global Public Health map of alcohol-induced mortality in 2024 shows that several counties in other states have the same relationship. Curry County Oregon, Apache and La Paz counties in Arizona, and Torrence County New Mexico are all mapped as having low rates of excessive drinking and high rates of alcohol-induced death. Ethnic or racial differences do not seem to be a factor. Curry County is strongly white (83%), Apache County is strongly indigenous (75%), La Paz is highly mixed (61% white alone, 23% Hispanic/Latino, 14.5% American Indian) and Torrence is fairly evenly divided between whites and Hispanics (52% white alone, 42% Hispanic). As the final map in this post shows, many of the “driest” counties located in the interior U.S. south are also among the least healthy.

USA Excessive Drinking and Alcohol-Related Mortality map

Such surprising findings indicate that the “excessive drinking” data need to be examined more closely, as will be done in the next GeoCurrents post.

Could U.S. Counties with Low Levels “Excessive Drinking” Really Have High Levels of Alcohol-Induced Mortality and Low Levels of Health? Read More »

Does Gallatin County Montana Really Have the Nation’s Highest Percentage of Excessive Drinkers? If So, Does It Matter?

Article-Grid, Cultural Geography, North America / Martin W. Lewis

A map of U.S. counties that supposedly shows the percentage of residents who drink alcohol excessively has recently received a lot of attention. What immediately struck me when I first saw it was the prominence of Wisconsin, mapped as having a far higher proportion of problem drinkers than any other state. As I spent a year in Wisconsin and witnessed its drinking culture, I was initially inclined to regard this portrayal as reasonably accurate. But the more I examined the map, the more my doubts grew. The discrepancy between Wisconsin and its neighboring states is simply too stark to seem realistic. Several other states also seem to be overly differentiated from their neighbors. The low level of excessive drinking in Utah, for example, is understandable due to the prevalence of the LDS (Mormon) church, but several highly Mormon counties in southeastern Idaho are not mapped in the same way. Due to such seeming discrepancies, I plan to examine the data used to create the map and post my findings on GeoCurrents. For the time being, however, I will take this map as given and ask what it might reveal about the significance of regional drinking patterns.

It might be assumed that extremely high levels of excessive drinking would have some correlation with social disfunction. At the individual level, after all, immoderate consumption is associated with a host of health and behavioral problems. According to this map, more than a quarter of the (adult?) population drinks excessively in quite a few U.S. counties. With such high rates of drunkenness, wouldn’t these same counties be expected to exhibit some aspects of social disorder? By the same token, shouldn’t the soberest counties do relatively well according to social indicators?

What the map shows, however, does not always match this image. Many of the counties supposedly burdened by excessive drinking are prosperous, healthy, well educated, and safe. Prominent examples include the affluent Rocky Mountain counties of Colorado, Teton County, Wyoming, and Dane County, Wisconsin. By the same token, many counties in the South that are depicted as relatively free of excessive drinking are characterized by widespread poverty, low levels of health and education, and high crime rates.

Let us consider two counties on the opposite ends of the spectrum: Gallatin, in Montana, and McDowell, in West Virginia. Gallatin is mapped as having the highest rate of excessive drinking in the United States, at 26.8 percent. McDowell, in contrast, falls in the second lowest excess-drinking category. Although I have no experience with McDowell County, I know Gallatin County quite well. Its main city, Bozeman, is noted for its bars, brew pubs, and lively drinking culture. Yet it is also a safe and easy-going town embedded in a high-trust society. Gallatin County is the kind of place where people leave their trucks unlocked and running in parking lots on frigid winter mornings, unconcerned that they might be stolen. From what the data indicate, McDowell County in West Virginia is very different kind of place.

These differences can be easily elucidated through a series of maps. To begin with, Gallatin has a far higher educational level than McDowell. As can be seen, it falls into the highest tier of residents with college degrees, while McDowell falls into the lowest. Bozeman is home to Montana State University, noted for its engineering programs. It has a small but thriving tech sector, and many graduates stay in town, not wanting to leave such a pleasant and amenity-rich environment. As the second map shows, Gallatin County also has a high percentage of residents with graduate and professional degrees. Bozeman is, in other words, a “knowledge town.”

As the series of maps at the bottom of this post shows, Gallatin is far ahead of McDowell in other commonly used metrics. It is much wealthier, has a much lower homicide rate, far higher life expectancy, a strikingly lower rate of obesity, a much lower instance of tobacco-smoking, a far lower drug-overdose rate, and a significantly lower divorce rate. And as paired tables at the end of the series of illustrations show, Gallatin County is booming, attracting people from other areas at a rapid clip, whereas McDowell is losing residents at an alarming rate. In short, the “drunkest” county in the United States is remarkably successful by almost any standard, while one of the country’s “soberest” counties is withering away, plagued by numerous economic, social, and behavioral problems.

We must be wary, however, of reading too much into these comparisons, and I must again caution that the “excessive drinking” data used in this analysis may not be valid. Correlation, moreover, does not imply causation, and many other factors must also be considered. McDowell County, for example, has suffered grievously from the closure of its coal mines, which has no doubt intensified its many problems – although few of its remaining residents have evidently been driven to drink.

Yet for all of this, it is still safe to say that excessive drinking is a major cause for concern, at both the individual and societal levels. The difficulty comes in defining what “excessive” means. For a strident teetotaler, any alcohol consumption may seem excessive; for an unrepentant lush, excessive consumption is difficult to achieve. This is a complicated issue, as worrisome or dangerous levels of alcohol consumption vary from individual to individual and according to social context. At any rate, it remains to be seen how the author of this map defined the term. I suspect, however, that the definition used has a lower threshold than one that I would favor.

Despite these qualifications, I have no doubt that Gallatin County has a higher per capita level of alcohol consumption than the average American county. But is that necessarily a bad thing? The issue is debatable. Some scholars claim that moderate alcohol consumption has profound social benefits. Edward Slingerland goes so far as to argue that we would not have developed civilization without intoxication. I am not willing to go that far, but I can say that a geographical study of alcohol consumption in the United States offers no support for the position that drinking is intrinsically harmful.

Edward Slingerland’s Book Entitled “Drunk”

As a final note, Gallatin County also seems to have a significantly higher-than-average rate of cannabis consumption.* Dispensaries are ubiquitous; the Montana Department of Revenue lists 72 in and around Bozeman, a city with only around 60,000 residents. As is the case with alcohol, I do not see any evidence that the prevalence of cannabis has harmed Gallatin County, although I am sure than many residents would disagree vociferously.

*Reliable data on cannabis consumption is impossible to find. But an AI (Grok) assessment is worth considering: “Legal sales (good proxy for market intensity, not user prevalence) data from the Montana Department of Revenue consistently put Gallatin near the top of the state in monthly legal cannabis sales—often second only to much larger Yellowstone County. Combined with Gallatin’s population (~129,000) and very high dispensary density around Bozeman, per-capita legal sales are high. That is consistent with high use by residents plus tourist/visitor purchases. Montana as a whole also ranks among the highest states for legal cannabis sales and excise-tax collections per capita.

McDowell and Gallatin Demographically Compared

Does Gallatin County Montana Really Have the Nation’s Highest Percentage of Excessive Drinkers? If So, Does It Matter? Read More »

Personalized Climate Favorability Maps and the Problem of Scale

Article-Grid, Environmental Geography, North America / Martin W. Lewis

As noted in the two previous GeoCurrents posts, it is challenging to make a map of climate favorability partly because preferences vary significantly from person to person. This problem is largely avoided on the ingenious website myPerfectWeather, which allows visitors to construct personal maps based on their own preferences. Unfortunately, the variables that can be selected are limited to average daily high temperature, daily average dewpoint temperature, and daily average cloud cover. Despite this limitation, it is interesting to see how maps generated on the site differ significantly due to relatively small changes in the desired conditions. The first of these maps posted below surprised me by giving San Diego County a middling ranking, contrasting sharply with nearby top-ranking Los Angeles County. In both the Camelot Climate Index and the Yahoo! “perfect weather” map considered in the two previous posts, San Diego is ranked as having the most ideal climate in the United States.

As intriguing as it is, myPerfectWeather is useless for my own climate preference, as I care a lot about seasonality and precipitation, neither of which are considered. But I did enjoy changing the parameters and generating different maps. I wanted to make one that puts mild and cloudy San Francisco County in the top position, which was easy to accomplish (see the map posted below).

I focused on San Francisco in part because it is one of the smallest counties in the United States, coterminous with the city of the same name and covering only 47 square miles. Along the Pacific Coast of the United States, mapping climate at the level of the county is highly misleading,* as conditions vary considerably with proximity to the ocean. Even in miniscule San Francisco County, the west side is distinctly cooler and cloudier in the summer than the east side. In large Mendocino County (3,878 square miles) on California’s northern coast, such differences are pronounced, with a mean July high temperate of 65°F (18.3°C) in coastal Fort Bragg and 91.1°F (32.8°C) in interior Ukiah, the county seat.

To illustrate such county-level climatic variability, I extracted Los Angeles County from two different Köppen climate maps of California, both found in Wikipedia. The first of these maps shows five distinct climate zones in L.A. country and the second shows seven. Two main factors account for these significant differences: distance from the ocean and elevation.

Although I admire the Köppen climate-classification scheme and use it extensively, it is not without problems. On the final map posted below, note how most of the city of San Diego has a BSh climate (arid steppe hot), but its northwestern neighborhood of La Jolla is placed in the BSk category (arid steppe cold). Although La Jolla is slightly cooler than downtown San Diego, especially in the summer, the difference is too small to merit classifying the former as “cold” and the latter as “hot.” In the Köppen scheme, an annual mean temperature of 18°C (64.4°F) is the cutoff between “hot” and “cold” arid climates, which is essentially that of La Jolla. I think that the system could be improved by adding a “mild” category between these two extremes. Doing so would slightly simplify the climate map of southwestern California.

Climate Zones of Southwestern California Map

* myPerfectWeather acknowledges this problem by noting that “Data is shown at the county level, which may not reflect city-level variation.”

Personalized Climate Favorability Maps and the Problem of Scale Read More »

The Impossible Task of Mapping “Ideal” Weather, Which Is Actually “Ideal” Climate – Or Is It?

Article-Grid, Environmental Geography, North America / Martin W. Lewis

One reason why I like the “Camelot Climate Index Map” (discussed in the previous GeoCurrents post) is the fact that that it is explicitly framed as a climate map rather than a weather map. Most of its competitors claim to show permanent weather favorability, but weather varies from day to day and cannot be predicted beyond a week or two. Such maps actually depict climate patterns, which are relatively consistent from year to year. The Yahoo!News illustration posted below, for example, asks readers to examine the map if they want to find a place with “perfect weather year-round.” Even if one agrees with Yahoo’s evaluation of “perfect weather,” such a place cannot be found anywhere, as the map itself demonstrates.

Yahoo! Perfect Weather (Climate) Map of the United States

Like the Camelot Climate Index Map, the Yahoo! map locates the most favorable climate zone of the United States in coastal California. Unlike it, Yahoo depicts the southeast as having the country’s second-best climate, which may surprise those who dislike warm and humid days. The methodology employed, however, is unduly simplistic, with a “perfect weather day” defined as one in which the daily maximum temperature is between 60°F (15.5°C) and 85°F (29.4°C) and rainfall is below 1 mm. Most people, I think it is safe to say, would not regard a windy, cloudy day with temperatures in the 40s and 50s, peaking briefly at 60°F, as having “perfect weather.” Similarly, an extremely humid rainless day with a high temperature of 85°F and a low of 75°F (23.8°C) would probably be regarded by most people as unpleasant. To assess weather comfortability, wind, humidity, and other factors must be assessed along with maximum temperature and precipitation.

As Jan Null, author of the Camelot Climate Index, emphasizes, weather and climate preferences vary considerably from person to person. Average preferences also differ from culture to culture and region to region. I was recently surprised to learn that in the East Riding of Yorkshire in the United Kingdom, a “heat wave” is officially defined a period of at least three days in which the maximum temperature exceeds 79°F (26°C). In interior California, where I was reared, a three-day period in July or August with high temperatures reaching only 79°F (26°C) would be regarded as an unseasonal cold snap. Although many local people would be delighted by such mild conditions, quite a few would complain, asking “what happened to summer?”

Although maps of “ideal weather” actually depict climate zones, the methods used to construct them do tend to hinge on weather preferences. They are based, in other words, on the perception of what constitutes a single “nice,” “ideal,” or “perfect” day, rather than on preferred average conditions across the entire year. Many if not most people also like variation and would find a climate characterized by an unceasing sequence of “perfect days” somewhat boring. Distinct seasonal change is also highly appreciated by many. A single day of ideal weather, moreover, can be especially sweet if it comes after a period of inclement conditions. Similarly, delight is often found when the clouds clear and the wind drops on an otherwise cool and showery day. Such a change can be inspirational, as perfectly realized in “Here Comes the Sun,” the Beatles’ most streamed song. Such considerations help make sense of the seemingly contradictory statement traditionally attributed to King Charles II: “England has the best climate in the world and the worst weather.” But try mapping that!

The Impossible Task of Mapping “Ideal” Weather, Which Is Actually “Ideal” Climate – Or Is It? Read More »

The Fascinating Camelot Climate Index Map

Article-Grid, Environmental Geography, North America / Martin W. Lewis

My X (Twitter) feed has recently featured a number of maps that delineate and rank climate zones in the United States based on how comfortable or idyllic they supposedly are. I find these maps interesting but inevitably problematic. One of the most intriguing is the “Camelot Climate Index,” devised and mapped by meteorologist Jan Null. The person who posted it on X (VB Knives) claimed that it is a “legit climate quality map, which closely tracks to average preferences,” further arguing that it “rewards warm and sunny [climates].” I objected in my reply, noting that the map puts cool and foggy San Francisco in second place, after San Diego. I also foolishly misquoted Mark Twain, who is often said to have quipped that “The coldest winter I even spent was a summer in San Francisco.” Twain never made this statement, but it does seem fitting.

Embarrassed by my misquotation, I decided to examine the map in more detail. What I found is a well-considered and modest effort by a knowledgeable meteorologist to map his own climate preferences based on rigorous quantitative analysis. As Null frames his “Camelot Climate Index:

In this 1960’s musical King Arthur professes that Camelot has a perfect climate all the year; and by royal decree at that! But actually an “ideal” climate is extremely subjective, with one person’s idea of perfection being met with disdain by others. Some individuals may want warm beach weather all year round, while four distinct seasons are most desirable for others. What follows is just one person’s (the author’s) idea that an ideal climate is sunny and relatively mild with few extremes in temperature, humidity or precipitation. …

Null also notes that his map “is rudimentary at best given the limited dataset,” and concludes by admitting that other climate observers would come up with very different maps: “If a similar project was done by a storm chaser it might be called the Oz Climate Index with a bias towards the number of thunderstorm days, hail and tornadoes.”

At any rate, I have posted Null’s original map below. To make it more legible to those unfamiliar with isoline maps, I have colored in the areas between the lines, as can be seen in the second map.

Although my own climate preferences are quite different from those of Null, I admire his index and resulting map. He has taken on a complex task and has produced an interesting and informative map. But if I were to make a climate map based on the song “Camelot,” it would have a very different appearance – if it could be produced at all – because I would interpret the lyrics in a more literal manner. This is again merely a subjective preference, but exploring it might still prove worthwhile.

First, I would give more weight to the first climate-related line: in Camelot, “June, July, and August cannot be too hot.” As a result, the interior southwest would have a lower ranking on my hypothetical map than it does on Null’s. He gives Las Vegas, for example, a measurement of 78, the sixth most favorable in the country. But with an average July high temperature of 104ºF and an average low of 82ºF, Las Vegas seems distinctly “uncamelotic” to me. Its forecast high temperature for this coming Saturday is a blistering 115ºF, with a low of 89ºF.

The song also stipulates a relatively mild winter, one that does not begin until December and ends “March the second on the dot,” and in which there is a “legal limit to the snow.” But although a long and snowy cold season is clearly precluded, the lyrics imply that Camelot does have a distinct winter and does receive snow on a regular basis. As Null acknowledges, many people like a climate with four distinct seasons, and the song’s King Arthur was apparently one of them. By my reading, the coastal strip of southern and central California – which is placed in the top position on Null’s map – would have to be downgraded due to its lack of a real winter.

Precipitation is a particularly tricky issue in determining the climate of Camelot. On Null’s map, a large number of days with measurable rainfall counts against an idyllic climate. As a result, the driest area of the country, the southern interior southwest, has a relatively high ranking, while Washington’s Olympic Peninsula and Hilo, Hawaii are given very low rankings. But the song says nothing about total precipitation, specifying instead that in Camelot’s unique micro-climate rainfall occurs only at night. The common image of Camelot, it seems to me, includes a verdant environment, once that necessarily receives regular rainfall throughout the growing season. By this criterion, which is admittedly not specified in the song, the arid, semi-arid, and Mediterranean climate zones of the American west would have to be downgraded. It is also noteworthy that the areas of the United States that get more rain in the night than during the day include the north and west facing slopes of the Hawaiian Islands and the Great Plains. If this metric were included in the index, the “big” island of Hawaii would have to be upgraded, as would the northern Great Plains.

Regardless of what the song stipulates, my biggest problem with the innovative “Camelot Climate Index” is the way that it groups together areas with dissimilar climates. Eureka, California and Las Vegas, Nevada, for example, have almost identical scores, ranking in the fifth and sixth highest positions nationwide. Yet their climates are very dissimilar. The highest July temperature ever recorded in Eureka is a mere 77ºF, which is 5ºF lower than Las Vegas’s average July low temperature. Cool, cloudy, and windy Eureka, moreover, receives more than an order of magnitude more precipitation than arid, sun-blasted Las Vegas (40.4 inches as opposed to 4.18 inches). I would not want to live in either place, as the former is too cool and cloudy in the summer and latter too hot and dry. Other people, of course, have different preferences. But I doubt that many would view the two cities as having equally appealing climates.

The Fascinating Camelot Climate Index Map Read More »

The Changing Geography of Top Corporate Headquarters in the U.S., 1890-2025

Article-Grid, Economic Geography, North America / Martin W. Lewis

(Note: Last December I began work on a series of posts focusing on the geography of wealth and corporate power. After a single post on the location of corporate headquarters, I had to suspend my work on this topic. This post includes maps that I made at that time, based on data from late 2025.)

In the GeoCurrents post of December 11, 2025, I mapped the locations of the headquarters of world’s largest publicly traded corporations (by market capitalization). As these maps show, most of these firms are headquartered in the United States. Most of the top U.S.-based firms, in turn, are located in the West Coast and Texas. I was surprised at how few have their head offices in New York, the historical core of American capitalism. I then decided to make a historical series of maps showing the changing locations of the corporate headquarters of the top 25* publicly traded companies in the United States (again, by market capitalization). Finding the necessary data, however, proved impossible in the time period that I allotted for the task. As an expedient, I turned to AI, getting the information that I needed from ChatGPT. As such information is not necessarily reliable, these maps should be considered preliminary.

The first map shows the situation in 1890, when railroad companies dominated the list. As can be seen, New York City was then the core of American capitalism, with nine of 24 top corporate headquarters. Chicago was in clear second place, with two railroad-oriented firms and two meatpackers. (as Carl Sandburg’s famous poem opens, “Chicago: Hog Butcher for the World, … Player with Railroads and the Nation’s Freight Handler”). The only other city with multiple top-25 corporate headquarters was San Francisco, with only two. The sole city in the south was Durham, home of the monopolistic American Tobacco. The smallest city on the map was Butte, Montana, which at the time had 10,723 residents (although that figure would rise to almost 40,000 in twenty years). Butte reached this position because of the extraordinarily rich ores extracted by the Anaconda Copper Company. The key role of railroads in 1890 can be seen by the location of leading firms in secondary cities in the western Midwest: Omaha (population 140,452 in 1890), Topeka (31,007), and Saint Paul (133,156). New England, the original heartland of the American industrial revolution, is notable for its lack of major corporate headquarters.

1890: Location of the Top 25 Publicly Traded Companies in the United States

The map of major corporate headquarters in 1920 shows some major changes from that of 1890. Over the intervening thirty years, railroad companies declined in importance, largely replaced by heavy-industrial firms. New York and Chicago retained their positions, and a clear belt of corporate power had emerged across the mid-Atlantic and southern Great Lakes regions. The rise of the automobile industry is reflected by the addition of Detroit, South Bend, and Akron to the map (note that only General Motors is mapped in Detroit; Ford is missing because it was a privately held company at the time).

1920: Location of the Top 25 Publicly Traded Companies in the United States

The map showing the situation in 1960 is surprisingly little changed from that of 1920. A small city far from the corporate heartland hasd, however, joined the list: Bartlesville, Oklahoma, with 27,893 residents in 1960. Bartlesville owed this position to Phillips Petroleum, which was founded on the productive oilfields of northeastern Oklahoma. Today, its two descendant companies, Phillips 66 and ConocoPhillips, are based in Houston, which has emerged as the core city of the U.S. oil industry.

1960: Location of the Top 25 Publicly Traded Companies in the United States

By 1990, the New York City metropolitan area had further cemented itself as the center of American corporate capitalism. Slightly more than half of the country’s largest publicly traded companies were at the time headquartered there, as were seven of the top ten. In contrast, the upper Midwest saw a substantial drop from 1920, reflecting the relative decline of heavy industries. In contrast, pharmaceutical and consumer-oriented firms had come to occupy a more prominent position. Because of this shift, the South appears on the map for the first time since 1890, represented by Coca-Cola in Atlanta and Walmart in the small city of Bentonville, Arkansas (population 11,257 in 1990). The only corporation in the western half of the country that made the 1990 map was Chevron, an oil company based at the time in San Francisco. In 2024, Chevron announced that it would relocate to Houston.

1990: Location of the Top 25 Publicly Traded Companies in the United States

The 2025 map reveals a completely transformed geography of major corporate headquarters, deviating sharply from all other maps in this series. The New York City area dropped precipitously, to be replaced by the San Francisco Bay Area, home to most leading high-tech companies. Austin and Seattle gain a prominent position for the same reason. Together, the Bay Area and the Seattle metro area were home to the top six firms. Overall, this map is more geographically balanced than any of the preceding ones. Intriguingly, the only company and city found on both the 1890 and 2025 maps is the consumer-products giant Procter & Gamble of Cincinnati. Procter & Gamble’s roots extend to the time when fat from hogs was used to make soap. As early as 1840, Cincinnati had become the world’s largest pork-processing hub, earning it the nickname “Porkopolis.”

2025: Location of the Top 25 Publicly Traded Companies in the United States

The geography of leading corporate headquarters is continuing to experience rapid shifts. The prominence of California’s Bay Area might soon be heightened, owing to the widely anticipated initial public offerings (IPOs) of the AI firms OpenAI and Anthropic. Both are based in San Francisco, and both are expected to be valued at up to one trillion dollars, easily putting them in the top twenty-five. OpenAI, however, is currently embroiled in a major lawsuit with Elon Musk that might threaten its future.

But as the map also shows, several leading companies have recently left the San Francisco Bay Area. Their departure was largely due to concerns about taxation and regulation. Tesla relocated to Austin in 2021, as did Oracle. Texas will likely gain another top-25 firm when SpaceX undergoes its expected IPO. SpaceX began its relocation from Hawthorn, California to Starbase, Texas in 2024. Starbase, located near the southern tip of the state, is, however, more of an industrial facility than a city or even town, with an estimated population of 500. If SpaceX subsequently merges with Tesla, as many expect, its headquarters will almost certainly remain in Texas.

The pace of corporate relocation may be increasing. Oracle, for example, recently announced that it will be moving to Nashville by 2030, even though it had moved to Austin only in 2021. The military technology firm Palantir, with a current market capitalization of around $350 billion, transferred its headquarters from Palo Alto, California to Denver in 2020, largely because of the protests that it had experienced in California. Only four years later – after facing additional rounds of protest in Denver – it announced its relocation to Miami.

* Or 24 in the case of 1890, due to data limitations.

The Changing Geography of Top Corporate Headquarters in the U.S., 1890-2025 Read More »

Mapping the Return of the Gray Wolf to California

Article-Grid, Environmental Geography, North America / Martin W. Lewis

One of the most surprising aspects of the recovery of the gray wolf in the United States is the return of the species to California. Wolves had been eliminated from the state in the 1920s and were not expected to return any soon time, if at all. But in 2011, a single wolf made the arduous journey from northeastern Oregon to northeastern California, remarkably tracked by wildlife officials (see the map posted below). Other wolves followed, and before long several breeding packs had been established. California’s Department of Fish and Wildlife recently announced that in 2024 “a minimum of 30 pups born across five or six packs and that’s the largest annual litter count we’ve had thus far in the state.” Although most California wolves are found in the state’s northeastern reaches, in 2023 a new pack was confirmed in the southern Sierra Nevada Mountain Range some 200 miles to the south.

Gray Wolf Distribution in California 2024 map

The first map in this post also shows potential wolf habitat the United States (excluding Alaska), as well as the approximate range of wolves at roughly the time when California was reinhabited. Although it is not surprising that many areas marked as habitat are still without wolves, it is perplexing that wolves are depicted as living in areas outside their supposed habitat zone. Determining what constitutes “wolf habitat” is a difficult task, but I still do not think that the cartographer has done a good job on this map. The small dots of “habitat” scattered across Nevada, Utah, and Idaho are especially suspicious, as these areas are too small to support viable packs. I suspect that a much larger portion of the U.S. is potential wolf habitat, at least to the extent that people would allow wolves to repopulate it.

Even mapping the historic range of the grey wolf in the United States is a difficult and contentious issue. As can be seen in the figure posted below, different maps show very different distribution patterns. Most uncertain is the historical range of the gray wolves in the eastern United States. Much of the uncertainty here stems from the presence of a closely related but smaller species, the red wolf. But it is not even clear whether this critically endangered animal is a separate species or merely a gray-wolf subspecies. Red wolves have, moreover, hybridized extensively with coyotes in recent decades, adding to the general taxonomic and distributional confusion.

Historic Range of the Gray Wolf in North America maps

Another striking aspect of the historic distribution of the gray wolf depicted on these maps is the absence of the animal from central and western California (except the German-language map produced by the World Wildlife Fund). The reported lack of wolves here is curious, as California’s vast Central Valley and surrounding foothills had roughly 500,000 tule elk circa 1800, which could have supported a large wolf population. It seems most likely that the conventional mapping of the historic distribution of the gray wolf in California is simply incorrect. At any rate, the chatbots that I consulted all agreed that grey wolves were historically found across most the state. Grok, for example, informed me that:

GROK: Yes, wolves historically lived in the western areas of California. Gray wolves (Canis lupus) were once widespread across the state, including the western regions, such as the Sierra Nevada, coastal ranges, and Central Valley. Historical records, including accounts from early European settlers and Native American tribes, indicate their presence throughout California before the 20th century. However, due to habitat loss, hunting, and extermination campaigns, wolves were largely eradicated from California by the 1920s.

The best potential wolf habitat in California is probably located in the larger and more remote parts of the state that currently support elk (wapiti, or Cervus canadensis; see the map posted below, which probably exaggerates the current range.) But although elk are now widely distributed in California, their numbers remain small. According to current estimates, California has some 12,700 elk, whereas Idaho has around 120,000 and Montana up to 175,000. Due in part to the paucity of elk and other potential prey species larger than the black-tailed deer, California’s wolves have been extensively preying on domestic livestock. As can be seen in the “depredation report” posted below, such kills are occurring roughly every other day. It is thus hardly surprising that California ranchers are worried about the state’s expanding wolf packs.

Elk (Wapiti; Cervus canadiensis) Range in California map

Wolf Depredation Report California May 2024

The map entitled “U.S. Gray Wolf Distribution and Habitat” that is posted above depicts the central Sierra Nevada as one of the state’s largest areas of wolf habitat. This extensive and heavily forested area does not, however, have any elk. Its thick seasonal snowpack would be a challenge for both elk and wolves, requiring extensive areas of winter habitat in the more densely populated western foothills, presenting another challenge. I doubt that it is coincidental that California’s existing wolf packs are concentrated in the drier and more sparsely populated lands found to the east of the northern Sierra and southern Cascades, which I have illustrated on the final map posted below.

Gray Wolf Range in California and Precipitation map

Mapping the Return of the Gray Wolf to California Read More »

The Controversial Expansion of Gray Wolves in the United States

Article-Grid, Environmental Geography, North America / Martin W. Lewis

As discussed in the previous post, the gray wolf has made a remarkable but controversial recovery in Europe. The same is true in the United States. In the mid twentieth century, wolves were found only in Alaska, far northeastern Minnesota, and the western half of Michigan’s Upper Peninsula. A few may have lived in northern Wisconsin, but the state’s Department of Natural Resources claimed that the species had been locally extirpated by 1960. The western U.S. was essentially without wolves. The map posted below shows the range of the gray wolf terminating precisely along the U.S.-Canada border in western North America in the 1960s. Using a political border to delimit the range of a species is usually a lazy and misleading expedient, but it is probably appropriate here. At the time, wolves living in mountainous areas of southern British Columbia and Alberta occasionally crossed the border, but breading packs had been systematically eliminated from the western United States.

Range of the Gray Wolf in the U.S. Circa 1965

Wolves began to return to the western U.S. in the 1980s. In 1979, a lone female was encountered near Glacier National Park, and two years later she evidently mated with a migrant male. The so-called Magic Pack that they formed quickly expanded, forming the nucleus for wolf repopulation in northwestern Montana. In the 1990s, wolves were intentionally reintroduced into Yellowstone National Park, located mostly in northwestern Wyoming. Enjoying federal protection as an endangered species in all U.S. states but Alaska and Minnesota, wolves continued to expand their range and numbers over the next several decades, mostly through natural population increase and dispersion. Captive breeding programs for the smaller and almost extinct Mexican subspecies allowed its reintroduction into the mountainous area along the Arizona-New Mexico border.

Wolf numbers and ranges in the United States are carefully tracked by wildlife managers, allowing the creation of reasonably accurate distribution maps. Synthesizing spatial information from all the maps I could find and using numerical data from a Wildlife Informer article, I have mapped the distribution of wolves in the U.S. as of 2024 (see below). As can be seen, Alaska and the Lake Superior region in the upper Midwest still support the largest populations. In the western “Lower 48,” Montana and Idaho stand out.

Gray Wolf Distribution and Population in the U.S. 2024

The wolf resurgence is as controversial in the United States as it is in Europe. In 2020, Colorado voters opted to reintroduce the species by a razor-thin margin; since then, 25 wolves have been released in the state. But in the following year, federal protection was removed, letting some states manage their own wolf populations and allow hunting and trapping.

Currently, the governments of Montana and Idaho are seeking to substantially reduce their number of wolves. Over the past few years, some 200 to 300 wolves have been annually hunted or trapped in Montana, with the total state population, around 1,100, remaining roughly stable. Montana’s Department of Fish, Wildlife, and Parks is now proposing an increased cull, hoping to reduce the population to around 450. Similarly, Idaho’s Gray Wolf Management Plan calls for a reduction from around 1,200 to roughly 500, with a minimum of 350. Environmental activists opposing the Idaho plan argue that the actual numbers may be significantly less than current estimates. Ranchers point to the more than 200 annual livestock predation deaths in contending that the wolf population is too large, regardless of precise numbers. In response to such concerns, the Idaho Department of Fish and Game recently adopted a new statistically and genetically sophisticated method of population estimation called “Approximate Bayesian Computation,” which came up with a figure of 1,150.

Such wolf-reduction plans appall most environmentalists, who believe that wolves are important for ecosystem integrity. As argued in a recent article from the Center for Biological Diversity:

Most insidiously, some Republicans are trying to roll back the protection wolves do have, which would turn western states into bloodbaths like those in the northern Rockies, where wolves aren’t federally protected. But the current threats to legal protections for wolves are more than just an attack on a beloved wild animal. They have broader, more devastating implications for western lands. Wolves are an essential part of healthy, functioning wild landscapes. Like other keystone species, such as bears, beavers, bison and birds, wolves contribute to ecosystem restorations and help build a wilder landscape.

The argument that wolves help maintain ecosystem health by limiting the grazing and browsing activities of large herbivores, especially along watercourse, is well substantiated. The idea that “birds” are a “keystone species,” on the other hand, is both ecologically and taxonomically illiterate (although the alliteration is nice).

The wolf-reduction plans of Montana and Idaho are characteristic of the more conservative attitudes found in these Republican-voting, or “red,” states. As a left-leaning (if purple-trending) state, Minnesota is taking a different path. “Objective 1B” of its current management plan is to “maintain a population comparable to recent estimates (2,300-3,000, well above the federal recovery goals) and distributed across the majority of current wolf range.” Intriguingly, although wolf numbers in Minnesota have declined slightly since 2004, the species’ range in the state has continued to expand, as demonstrated by the figure posted below.

Gray Wolf Range and Population in Minnesota Map

Although the current wolf policies of Montana and Idaho seem harsh when compared to that of Minnesota, they seem positively pro-wolf when compared to that of Sweden, a country with a strong “green” reputation. As noted in the previous post, Sweden intends to reduce its wolf population from roughly 400 to only 170, whereas Idaho has set a far more robust target population of 500. Sweden, moreover, is almost twice as large an Idaho (173,860 sq mi vs. 83,571 sq mi), and a larger percentage of its land is potential wolf habitat.

The Controversial Expansion of Gray Wolves in the United States Read More »

Canadian Provinces and U.S. States Economically Compared

Article-Grid, Economic Geography, North America / Martin W. Lewis

As noted in the previous GeoCurrents post, Alberta is Canada’s most prosperous and economically productive province. But it does have, as might be expected, by the country’s highest cost of living. Although living costs in Alberta are above the national average, they are below those of Ontario, Canada’s most populous province. The major factor is the cost of housing. As can be seen on the map posted below, the average price of a house in both Ontario and British Columbia far exceeds Alberta’s figure, owing largely to expensive real estate of Toronto and Vancouver.

But despite its prominent economic position in Canada, Alberta ranks in a middling level when compared with the states of the U.S. Although the United States and Canada have historically maintained rough economic parity, their relative positions began to diverge in the second decade of the twenty-first century, as can be seen in the graphs posted below. As the following maps show, per capita GDP and median income levels are now substantially lower across the provinces of Canada than across the states of the United States. Canada’s Atlantic provinces occupy a particularly low position, below those of the poorest U.S. states. But it is also true that the cost of living is, on average, higher in the United States than in Canada, owing in part to elevated medical costs in the U.S.

(Note: In making the maps posted below, I used ChatGPT to convert Canadian dollars to U.S. dollars, which might have resulted in erroneous figures.)

U.S. and Canada Economically Compared graphs

U.S. States Canadian Provinces Per Capita GDP map

Canadian Provinces and U.S. States Economically Compared Read More »

Economic Factors Behind the Alberta Secession Movement in Canada

Article-Grid, Elections, North America, Regionalism / Martin W. Lewis

Several recent GeoCurrents posts have emphasized similarities in the electoral geographies of the United States and Canada. But there are also some major differences, particularly regarding state/province voting patterns and levels of economic development.

In the U.S., the wealthiest states tend to favor the left-leaning Democratic Party while the poorest states tend to favor the right-leaning Republican Party. As can be seen in the map first posted below, in 2024 presidential election all but one of the states that gave Kamala Harris the highest percentage of their votes have relatively high median household income figures. Intriguingly, however, that state that most strongly supported Harris, Vermont, has only a mid-level income figure. By the same token, states with the lowest median household income figures supported Donald Trump in 2024. Yet the state that most strongly supported Trump, Wyoming, has a relatively high median income figure (see the second map below).

2024 U.S. Election Income Level and Harris Vote map

2024 U.S. Election Income Level & Trump Vote map

In Canada this relationship is largely inverted. In the 2025 Federal Election, provinces with the lowest median income levels strongly supported parties on the left, whereas the province with the highest income level, Alberta, strongly supported the Conservative Party (see the first two maps posted below). The correlation between economic development and conservative political orientation is more clearly apparent when comparing the 2025 electoral map with the map of per capita Gross Domestic Product (GDP) by province. As can be seen, the two provinces that strongly supported the Conservative Party, Alberta and Saskatchewan, have much higher GDP figures than Canada’s other provinces. (Canada’s three northern territories admittedly have high GDP figures and low levels of support for the Conservative Party, but they also have very small populations.)

Conservative Vote in the 2025 Canadian Election map

This seeming economic-electoral discrepancy between the United States and Canada is easily explained, at least at the upper end of the GDP spectrum. The productive economies of the two conservative provinces, Saskatchewan and Alberta, are heavily based on natural resources and agriculture, and in Alberta fossil fuels play a particularly prominent role. In recent years, Canada’s Liberal government has pursued a carbon-reduction agenda that has restricted the oil and natural gas industries, and extractive processes more generally, angering many voters. Such policies have helped spur an independence movement in Alberta and, to a lesser extent, Saskatchewan

The independence movement in the western Prairie Provinces also gains strength from Canada’s system of “equalization payments,” which effectively transfers funds from wealthier to poorer provinces so that “reasonably comparable levels of public services can be provided at similar levels of taxation.” In 2025, the three westernmost provinces will not receive any equalization payments, whereas all the eastern provinces, including Ontario, Canada’s economic core, will receive such payments. The equalization-payments map, posted below, not surprisingly figures prominently in Alberta independence movement.

Economic Factors Behind the Alberta Secession Movement in Canada Read More »

Who Are “The Laurentian Elite?” Where Is “Laurentian Canada?” & Why Do Americans Ignore the Saint Lawrence River?

Article-Grid, Borders, Geopolitics, Historical Geography, North America, Physical Geography / Martin W. Lewis

Canada’s western separatists often refer disparagingly to “the Laurentian elite” and “Laurentian Canada,” terms that are not familiar to most readers in the United States. These terms derive from the region around the Saint Lawrence River, Canada’s political, economic, and demographic core, but they refer more specifically to the Canadian establishment, or governing class. As the Wikipedia article on the “the Laurentian elite” notes, this group has a “bias toward federal power,” which is one of the main reasons why is deeply distrusted by western regionalists. It is closely associated with the center-left Liberal Party.

I could not find a map of “Laurentian Canada,” as its geographical bounds are vague and unspecified. It could conceivably cover the entire “Quebec City-Windsor Corridor,” a well-mapped region that contains roughly half the country’s population (see the map below). ChatGPT, however, tells me that “Laurentian Canada” refers “particularly to southern Quebec and eastern Ontario,” a region focused on Ottawa, Toronto, and Montreal. The chatbot Grok also specifies the inclusion of Quebec City. I am dubious of this claim, however, as Quebec City strikes me as too provincial and too francophone to merit inclusion. It is also uncertain whether this informal region encompasses southwestern Ontario; cities such as Windsor strike me as too peripheral to be included. I have therefore impressionistically mapped a limited “Laurentian Canada” within the larger “Quebec City-Windsor Corridor” (the second map posted below).

The Saint Lawrence River, and larger Saint Lawrence corridor, is tremendously important in the history of Canada and for the country’s self-image. In the U.S., however, the river and its basin are generally overlooked, even though a large portion of the watershed is in the United States. For more than a hundred miles in northern New York, moreover, the river forms the border between the two countries. But consider its portrayal in a visually appealing but conceptually muddled National Geographic map of the “Watersheds of the United States,” the subtitle of which asks, “Where do the U.S. watersheds drain?” A casual interpretation of this map might lead to the conclusion that the rivers of northern New York drain into the “Great Lakes,” as this watershed (depicted in orange), is not mapped as reaching the sea. Bizarrely, a few limited areas in Canadian that drain either into the Great Lakes or directly into the Saint Lawrence are mapped as part of the same drainage basin, yet almost all of the northern portion of the watershed goes unmapped.

The general exclusion of the Saint Lawrence River from the U.S. geographical imagination stems in part from a heritage of nationalist thinking. The Great Lakes figure prominently, as four of them are partly in the U.S. and the fifth, Michigan, is entirely American. But the Saint Lawrence? It seems to be regarded as a Canadian river, and therefore of relatively little significance. Contributing to this partial erasure of the Saint Lawrence is the fact that its upper reaches, bordering New York, is something of a demographic and economic backwater. The two largest cities along this portion of the river, Cornwall and Brockville in Ontario, have modest populations (47,800 and 22,100 respectively), and those on the U.S. side are smaller still.

The reduced significance of the upper Saint Lawrence is partly rooted in physical geography and history. Before the development of canals, traveling upstream from Montreal would take one into Lake Ontario but not much farther, given the profound barrier of Niagara Falls. To reach the upper Great Lakes, and hence the interior of the continent, travelers would instead voyage up the Ottawa River and then make a short portage to gain access to Georgian Bay in Lake Huron (see the map below).

Ottawa-River Route to Interior North America

But regardless of such occlusion in the U.S. geographical imagination, the Saint Lawrence is one of the great rivers of the world, the tenth largest by average discharge. In terms of flow, the Mississippi and Saint Lawrence are in a league of their own as far as North America rivers are concerned, with average discharges into the sea of 21,300 and 17,600 cubic meters per second respectively. The next largest, the Mackenzie, flowing into the Arctic Ocean, has an average discharge of 9,800 m³/s, whereas that of the mighty Columbia is “only” 7,400 m³/s (see the map posted below).

Largest Drainage Basins of North America map

The relatively short length of the Saint Lawrence has probably contributed to its relative obscurity in the United States. But its actual length is the subject of debate. Does it include the massive estuary, the largest in the world according to some sources? Although most watershed maps of the Saint Lawrence include the estuary, sometimes even up to Anticosti Island (see the Wikipedia map posted below), I have covered only a small portion of it on my own map (the second figure posted above). Equally vexing is whether the Great Lakes, and the short rivers that flow between them, can be counted as part of the same river. Vanishingly few sources do so, but the Wikipedia article on the Saint Lawrence includes all options in reckoning its length:

Length: 500 km (310 mi) excluding the estuary. C. 928 km if included. (St. Lawrence River–Lake Ontario–Niagara–Lake Erie–Detroit–Lake St. Clair–St. Clair–Lake Huron–St. Marys River–Lake Superior–St. Louis–North River: 3,058 km).

Determining the length of a river, it turns out, can be a very tricky exercise.

Saint Lawrence Among Other Rivers of Central North America

Who Are “The Laurentian Elite?” Where Is “Laurentian Canada?” & Why Do Americans Ignore the Saint Lawrence River? Read More »

Canada’s Population Is Clustered Near the Southern Border, Yet Is Nonetheless Distanced from the United States

Article-Grid, North America, Population Geography / Martin W. Lewis

Although it is often claimed that 90 percent of the people of Canada live within 100 miles of the U.S. border, this assertion is of dubious validity. Reputable sources put the figure closer to 80 percent. Statistics Canada more convincingly claims that 66 percent of Canadians live within 100 kilometers (62 miles) of the border, based on the 2016 census. An excellent population-density map, derived from 2006 census data, depicts this “100 kilometer” limit, which passes through Winnipeg and Quebec City.

But if Canada’s population is not as clustered along the U.S. border as many people think, the lack of habitation across the northern three-quarters of the county is still noteworthy. Significantly, the zone of extremely low population density – defined on the map posted above as fewer than 0.4 persons per square kilometer – extends to the country’s southern limit in central and western Ontario. Demographically speaking, Canada is thus divided into western and eastern segments. As a result, population-based cartograms of Canada have a strange if not disconcerting shape, depicting the country as two imbalanced lobes. This demographic division is of some relevance to the separatist movements of Alberta and Saskatchewan.

As a cartographic experiment, I took the Wikipedia map of the results of Canada’s 2025 election and eliminated most areas of extremely low population density. This electoral map, like most others, exaggerates support for the Conservative Party and Bloc Quebecois, as these two parties find much of their support in agricultural areas of relatively low-density population density.

2025 Electoral Map of the More Densely Settled Areas of Canada

Even though Canada’s population is concentrated in the far south, density of settlement increases along most of the border as one heads north out of the United States. Consider, for example, the map detail posted below (taken from the second map posted below). From “Point A” in sparsely settled northern New York, for example, a trip into Canada would quickly take one into the densely inhabited Montréal metro area, with more than four million inhabitants. A trip south, in contrast, would take one into the sparsely settled Adirondak Mountains. From “Point B” in the wilds of northern Maine, a trip to the north, west, east, southwest, or southeast would take one into more densely inhabited areas in Canada. One would have to travel considerably farther to the south to reach areas of comparable population density in the United States.

Population Density of Northern New England and Maritime Canada map

Canada’s trans-border demographic advantage is also pronounced in the northern Great Plains, as is depicted in the “five-state, three-province” map posted below. Alberta alone has a larger population than Montana, Wyoming, North Dakota, South Dakota, and Nebraska combined. Canada’s Prairie provinces are also more urbanized than the northern Plains states of the United States. Calgary, Edmonton, and even Winnipeg far exceed any cities of the Dakotas, Montana, and Wyoming. It is also notable that the largest cities on the U.S. side of this trans-national region are on or near its eastern flank, unlike those of Canada.

Demographic Patterns of the Prairie Provinces of Canada and the Northern Great Plains in the U.S. map

In southwestern Saskatchewan and southeastern Alberta, rural population density increases markedly a few hundred kilometers north of the U.S. border. It does so largely because soil fertility and moisture also increase in this area, allowing more productive farming. Here one finds – or found, before agricultural settlement – the Aspen Parkland, a transitional biome between the grasslands of the Great Plains and the northern boreal forest (taiga). This agriculturally favorable zone terminates abruptly in north-central Alberta and Saskatchewan, as is reflected on the population density map. Farming is not feasible across most of the northern halves of the Prairie Provinces not because of climate but rather because of soil. But to the northwest of the belt of continuous agricultural settlement lies a secondary zone of fertile Aspen Parkland in the so-called Peace River Country, which is home to roughly a quarter-million persons.

Aspen Parklands and Population Density in Alberta and Saskatchewan map

The main body of the Peace River Country forms a relatively compact zone of settlement, but far to the north lies a smaller farming region, located to the east and southeast of the town of High Level. Approximately 20,000 persons live here, cultivating some 350,000 acres of rich farmland 455 miles (733 kilometers) north of Edmonton. This northernmost extension of the Peace River Country was first settled by farmers after World War II, perhaps making it the final agricultural frontier of North America. It also has two large oil and gas fields and two lumber mills.

Satellite Map of the High Level Region of Northern Alberta

Canada’s Population Is Clustered Near the Southern Border, Yet Is Nonetheless Distanced from the United States Read More »

Exploring the Urban/Rural Electoral Divide in Canada, and Its Exceptions

Article-Grid, Elections, North America / Martin W. Lewis

Although the Canadian electorate is more leftwing than that of the United States, the two counties have similar electoral geographies (see the previous GeoCurrents post). In both the U.S. and Canada, population density figures prominently in the political divide: the more urban the location, the higher the vote tends to be for parties and candidates on the left.

There are, however, some important exceptions to this generalization. Canada’s three very sparsely inhabited northern territories, along with the thinly settled northern reaches of Saskatchewan, Manitoba, and Quebec, opted for left-leaning parties in the 2025 federal election, as they generally do. Such voting behavior stems largely from the fact that many or most of their inhabitants are indigenous (First Nations), a group that generally supports the Liberal and NDP parties. A similar pattern is found in the United States, where sparsely inhabited but heavily indigenous areas in central and western Alaska, western South Dakota, and northern Arizona usually vote for candidates in the Democratic Party. (Urban Native Americans and those without reserved lands, however, have recently been trending in a decidedly Republican direction; according to exit polling, 68 percent of “Native American/American Indian” voters supported Donald Trump in 2024.)

The tendency for rural areas and small cities to support conservative parties is exemplified by southern Ontario, which has 94 percent of the province’s population on 14 percent of its land. As the map posted below shows, all but one of this region’s sizable cities supported the Liberal Party in the 2015 federal election. Southern Ontario’s rural areas turned, as usual, to the Conservative Party, as did Toronto’s northwestern suburbs and exurbs. In the Ottawa area, city and suburbs alike supported the Liberals. The one metropolitan area that went “blue” (“red” by U.S. conventions) in 2025 was Windsor, immediately south of Detroit. Intriguingly, the city of Windsor has long been a stronghold of the leftwing New Democratic Party. Note, however, the city’s light shade of blue on the map, showing that the Conservative victory came through minority support, owing to the multiple parties on the left. The Conservative Party’s success in the Windsor area has generally been attributed to the rising cost of living and concerns about public safety, but such issues are equally prominent in other Ontario cities, where the Conservatives gained votes but did not achieve victories. Windsor’s troubled manufacturing sector and its close economic ties with the United States may have contributed to its 2025 swing to the right.

Southern Canada’s major exception to the urban/rural political divide is found in Quebec. In the western half of southern Quebec, most rural districts usually support candidates in the separatist-oriented Bloc Quebecois, as they did in 2025. Although this party has some conservative aspects, it is usually classified as center-left (see the discussion forum for the previous GeoCurrents post). The Bloc also did well in sparsely settled areas of far eastern Quebec. But it was a different story in the lower Saint Laurence Valley in the southeast, where the Conservatives prevailed in rural areas and small cities, as they usually do. Although the Liberal Party triumphed in Quebec City in 2025, its suburbs voted blue (red, by U.S. conventions). Outside this southeastern region, support for the Conservative Party in Quebec is very limited.

The compelling question here is why most rural and small-city voters in areas surrounding Montréal support the socially liberal Bloc Quebecois whereas those near Quebec City turn instead to Conservative Party. Quebec City is much smaller and less cosmopolitan than Montréal, which is probably of some significance (Quebec City 2021 population, 549,459 city and 839,311 metro; Montréal 2021 population, 1,762,949 city and 4,291,732 metro). It is also much more francophone than largely bilingual Montréal. The French language and Quebecois culture have historically been more threatened in the west than in the east. Such precarity may have fostered protective ethno-nationalist politics in the west. Such reasoning, however, cannot account for the success of the Bloc Quebecois in far-eastern Quebec.

In discussion forums on Reddit, a different factor behind southeastern Quebec’s conservatism tends to be highlighted: rightwing talk radio. In the 1980s and 1990s, the radio-host turned politician Gilles Bernier gained influence across the Quebec City region. From 1984 to 1997, he represented the riding of Beauce, south of Quebec City, in the Canadian parliament. As Wikipedia notes, Beauce “has the highest percentage of people who answered “Canadian” as their ethnic origin in the 2006 Census” and has the “highest percentage of White people of European descent (99.3%)” in the country. The current influential radio figure in southeastern Quebec is Éric Duhaime, described by Wikipedia as both “openly gay” and as “a Canadian conservative columnist, radio host and politician serving as leader of the conservative party of since April 17, 2021.” The same article further claims that Duhaime “has been associated with radio poubelle(literally ‘garbage radio’), a style of provocative right-wing radio popular in Quebec.” But while it is tempting to attribute a conservative regional orientation to right-wing radio, one must ask why radio poubelle proved much more popular in the east than in the west. This importamt question deserves more investigation than I can provide.

British Columbia is another interesting place for examining the urban/rural electoral divide. Outside the three largest urban areas, Vancouver, Victoria, and Kelowna, only one electoral district did not vote Conservative in 2025* (Courtenay-Alberni, which supported instead the leftwing New Democratic Party.) But rural and small-city voters in this “left-voting” province proved to be much less conservative than rural and small-city voters in neighboring “right-voting” Alberta. In several districts of British Columbia, the Conservative candidate in 2025 won with less than 40 percent of the total vote (see the map posted below). The only dark-blue riding in British Columbia, with a more than 70 percent Conservative vote in 2025, is found in the northeast, an area demographically anchored on the agricultural “Peace River Country,” the core of which is in conservative Alberta.

British Columbia and Alberta 2025 Election map

Although the Conservative Party narrowly won most rural ridings in British Columbia in 2025, it had been a different story in previous recent federal elections. In 2021, the leftwing New Democratic was victorious across the province’s non-metropolitan western districts and triumphed as well in central Vancouver. Even in 2025, if one were to combine the vote tallies of the Liberal, NDP, and Green parties, western British Columbia would not be mapped in “conservative blue.” Some of this region’s electoral districts have large indigenous communities, which constitute more than a third of the population of the massive Skeena-Bulkley Valley riding. Overall, however, rural and small-city voters of all ethnic groups in western British Columbia tend in a leftwing direction The same pattern is found in several non-metropolitan counties in northwestern California and far-western Washington and Oregon.

Finally, most rural and small-city voters in Nova Scotia and Prince Edward Island in far eastern Canada also favor the political left. Contrastingly, most of New Brunswick follows the more typical pattern, with the larger cities voting “left” and the more rural areas voting “right.” The exception to this rule in New Brunswick are several mostly French-speaking rural ridings in the north and east, which opted for the Liberal Party in 2025, as they usually do.

I thought it would be interesting to compare recent voting patterns in Canada’s Maritime provinces and in neighboring northern New England in the U.S. The map that I made is crude and impressionistic, but it is still suggestive. Extensive areas on both sides of this trans-border region have left-voting but mostly rural elecorates. The adjacent low-density areas of interior Maine and central New Brunswick, however, have a decidedly conservative orientation.

Northern New England & Canadian Maritimes Political Orientation map

*The “West Vancouver-Sunshine Coast-Sea to Sky Country” riding might seem mostly rural on the map, but it encompasses West Vancouver as well as the ski-resort town of Whistler.

Exploring the Urban/Rural Electoral Divide in Canada, and Its Exceptions Read More »

A Major Win for the Left in Canada’s 2025 Election – But Not in Alberta or Saskatchewan

Article-Grid, Elections, North America / Martin W. Lewis

The Conservative Party was widely expected to win Canada’s April 28 federal election. As recently as January 2025, the party was polling well ahead of the governing Liberals (see the graph posted below). But after U.S. President Donald Trump began calling Canada the “fifty-first state,” support for the Conservative Party plunged while support for the Liberal Party surged. In the end, the Liberals scored a relatively narrow victory, taking 170 parliamentary seats with 43.8 per cent of the vote to the Conservatives 143 seats with 41.3 percent of the vote.

But if the election was fairly close in terms of party politics, it was a different matter regarding political ideology. In Canada, the center-right Conservative Party is the only major party with a right-leaning political stance. But three additional parties are generally located to the left of the center-left Liberal Party. All three suffered major losses in 2025, as left-leaning voters rallied with the Liberals to prevent a Conservative victory. The New Democratic Party lost 17 seats, the Bloc Québécois lost 11, and the Green Party lost one, retaining a single seat. All told, the election might be considered a landslide for the left, with its four parties taking 57.7 percent of the vote against the Conservative’s 41.3 percent.

Although the Liberal Party scored decisive victories in eight of Canada’s 10 provinces, it was a different matter altogether in the remaining two: Alberta and Saskatchewan. In these resource-rich provinces, it was the Conservative Party that scored landslide victories, taking 63.7 percent of the vote in Alberta and 64.7 percent in neighboring Saskatchewan. As can be seen in the second map posted below, the discrepancy in support for the Conservative Party between these two provinces and the rest of Canada is profound. Outside of Alberta and Saskatchewan, the Conservatives took more than 45 percent of the vote only in Manitoba, while in Quebec they received a meager 23.4 percent. As will be explored in later GeoCurrents posts, the nationwide defeat of the Conservative Party has reinvigorated a separatist movement in Alberta, and, to a lesser extent, in Saskatchewan. This development potentially threatens Canadian nationhood.

Canada 2025 Election Results by Province map

As is the case in the United States and Europe, rightwing political sentiments in Canada are concentrated in rural areas and in smaller cities and towns, whereas leftwing political sentiments are concentrated in urban areas. But again, Alberta and Saskatchewan are partial exceptions (see the first figure below). To be sure, the left did relatively well in 2025 in central Edmonton and Calgary, taking three parliamentary seats. In Strathcona, located in Edmonton’s university district, the social-democratic New Democratic Party came in first, as it has in every federal election since 2008. But as the map posted below shows, all other ridings (federal electoral districts) in the Edmonton and Calgary metro areas supported the Conservative Party. In Saskatchewan, the Conservative Party triumphed in all six metropolitan ridings, located in and around the smaller cities of Saskatoon and Regina, albeit with lower vote percentages than in adjacent rural areas. Such urban support for the Conservative Party sharply differentiates Alberta and Saskatchewan from Canada’s other western provinces, British Columbia and Manitoba (see the second figure below).

2025 Canada Election Metro Areas in Alberta and Saskatchewan map

2025 Canada Election Metro Areas British Columbia and Manitoba map

But as the map posted below shows, one highly rural Saskatchewan riding, which essentially covers the entire northern half of the province, gave landslide support to the Liberal Party (with 65.1 percent of the vote, as opposed to 25.5 percent for the Conservative Party). Not surprisingly, this district – Desnethé-Missinippi-Churchill River – is demographically dominated by indigenous people, who constitute more than 70 percent of its population. Other parts of Canada with indigenous (First Nations) majorities also voted for parties on the left, as will be examined in a later post.

Alberta and Saskatchewan in the 2025 Canadian Election map

A Major Win for the Left in Canada’s 2025 Election – But Not in Alberta or Saskatchewan Read More »

The Indo-European Controversy: Facts and Fallacies in Historical Linguistics	Diversity Amid Globalization: World Regions, Environment, Development (6th Edition)
The Myth of Continents: A Critique of Metageography	Green Delusions