Previously: Introduction
Slime Mold Time Mold proposes that obesity is caused by environmental contaminants that build up in the water supply and flow downstream. They have two main pieces of evidence for their thesis:
People who live at the mouths of major river systems and in endorheic basins are more obese than people who live elsewhere.
People who live at high altitudes have dramatically lower levels of obesity than people who live at lower elevations.
In this section we’ll just look at the river systems and basins. Altitude will be covered next time.
I: Sold Down the River
This is a map of the Mississippi River watershed in the United States:
This is a map of obesity rates by US county (data is slightly old, from 2008):
The lower portion of the Mississippi River watershed shows clearly in bright red. But is that because of environmental contamination?
The lower Mississippi River valley had excellent cotton-growing soil, and before the American Civil War more than half of its residents were slaves. Slavery was so profitable in the area that slaves from further north were literally “sold down the river” - a phrase that still signifies deep betrayal.
To this day, many of the descendants of those slaves live in the lower Mississippi River valley, along the Mississippi-Arkansas and Mississippi-Louisiana borders. They’re black, rural, and some of the poorest people in the US - all factors correlated with obesity.
In fact, the county-level obesity map I showed you earlier is mostly just a map of where disadvantaged ethnic minorities live. The blue circles are where black people live now, and the green circles are areas with large Native American populations.
Okay, forget about race for a second. What if we just look at obesity among white people?
I don’t have county-level data for this. On the state level, whites in Arkansas and Mississippi, near the lower Mississippi river, are in the most obese category. This seems unrelated to watersheds though:
Once you take race out of the picture, the obesity map you’re left with is mostly a map of socioeconomic status.
II: Beijing Belly
To their credit, Slime Mold Time Mold also looks for evidence outside of the United States. They offer this map of China:
This is mostly a map of affluence. Richer parts of the country, which have adopted more of the western diet, look more obese than poorer areas that are still eating a more traditional diet. There are two big exceptions:
Xinjiang is far more obese than you’d expect based on income alone. SMTM claims that this is because it’s in an endorheic basin at the bottom of a watershed. However, in 2010, obesity was far more common among members of the Uighur ethnic group than among Han Chinese in Xinjiang. Given what we saw about how obesity varies among ethnic groups in the US, and given what we know about Xinjiang, there could be a more parsimonious explanation here.
Shandong, the most obese province in China, has very average income levels and no obvious ethnic minorities (99% Han). However, as SMTM points out, it is on the mouth of the Yellow River. Suspicious indeed.
But although the Yellow River does meet the ocean in Shandong, only a small portion of the province is actually inside the Yellow River watershed. It’s a surprisingly mountainous area, with many small rivers originating within the province and leading directly to the ocean.
Is obesity in Shandong concentrated near the Yellow River watershed? At least for childhood obesity, it doesn’t look like it, according to this recent paper.
The paper starts with a sample of over 6 million schoolchildren whose height and weight were measured by local clinics. It does a bunch of statistics that I don’t know how to evaluate and concludes that there are two major obesity clusters in Shandong, in the cities of Jinan and Jining. Jinan is situated on the bank on the Yellow River, but Jining is well outside of its watershed. Other areas along the Yellow River were not especially affected, so my guess is that this is related to urbanization and western diet adoption.
I don’t know why Shandong is so obese, and I would be excited to hear from anyone who does. But it doesn’t look like it’s because of pollution in the Yellow River.
More importantly, though, is Shandong actually anomalously obese? I can’t find a source for Slime Mold Time Mold’s map of obesity in China. I asked a Chinese-speaking economist I know if he could think of any reason why obesity would be unexpectedly prevalent there. After looking around at Mandarin-language sources for a bit, he told me that most of the obesity maps he saw look more like this:
Shandong does not look any more obese than its neighbors on this map. The most obese areas of the country, immediately north of Shandong Province, are Beijing and its surroundings. The map shows a clear divide between northern and southern China, which might be due to well-known differences in diet between the regions.
III: I can’t think of a clever title related to Iran
SMTM also offers a map of Iran, where the provinces near the mouth of the Arvand Rud (the confluence of the Tigris and Euphrates rivers) and the Caspian Sea have more obese populations than elsewhere.
I know very little about Iran, which provides me with an opportunity to test my hypothesis. Are outlier provinces populated by ethnic minorities?
In the north, the Azeri, Gilaki and Manzandarani provinces definitely look more obese than the others. (The Gilaki and Mazandarani peoples are closely related.) These are not just the provinces nearest to the bottom of the Caspian Sea basin: Golestan, directly on the Caspian Sea, is not especially obese, and the obese Azeri-majority province of Azerbayjan Sharghi is at high elevation in different watershed.
It’s less clear what’s going on in the southwest of the country. Khuzestan, by the mouth of the Arvand Rud, seems to be populated by an Arab minority group, but the similarly-obese neighboring Bushehr province seems to be mostly Persian ethnically. But Bushehr is not in the Tigris/Euphrates watershed, so this is not evidence for watershed theory.
Southeastern Iran, home of the Baloch ethnic minority, has the lowest obesity rates in the country. A paper I found from 2004 claims that the Balochs were (at the time) the least obese ethnicity in Pakistan too. On the other hand, southeastern Iran is also extremely poor, so they might just not have enough food to be obese.
In general, obesity rates in Iran differ along both ethnic and geographic boundaries. In the northwest of the country ethnic background predicts fairly well and watershed position does not. In the southwest, neither ethnic background nor watershed position correctly predicts obesity.
Conclusions:
Obesity rates are sometimes different at different levels in watersheds, but not as often as Slime Mold Time Mold claims. When there is a relationship, it tends to be easily explained by other factors.
In the United States, people living in the lower Mississippi River watershed are extremely obese, but no more so than you would expect from a poor majority-black population in the US.
In China, the history of oppression of the Uighur residents of Xinjiang should probably be taken into account when trying to understand the province’s high obesity rates. Shandong province may or may not have anomalously-high levels of obesity, but most of Shandong is not in the Yellow River watershed, and child obesity near the Yellow River is not clearly higher than child obesity elsewhere.
In Iran, differences in provincial obesity rates fall partially along ethnic lines, and are not well-predicted by provincial watershed level.
Very interesting stuff! Loved SM TM's series too. Do you plan on writing any more of these? Pretty please with cherry on top?
I think your weird outlier in Illinois is created by a poor-quality dataset (with state-level numbers being extrapolated to individual counties): there are no demographic cut-offs associated with any of Illinois's borders with surrounding states (unlike the IA-MO boundary, which shows up in a whole bunch of data, like election results and religious identification). It would make sense for IL as a state to be less obese than its neighbors, because it has a high number of college-educated adults, but much of the south of the state is very poor and often has high black populations (particularly in Alexander county and St. Clair county, which don't cluster with other very black areas on your map). On specifically *childhood* obesity, Illinois probably clusters with its southern neighbors, according to this map from 2012: https://ci.uky.edu/kentuckyhealthnews/2012/08/31/kentucky-ranks-third-among-states-in/
If there's any state uniquely successful at fighting obesity, it's probably CO, since it seems to me like it does better than we might suspect just from altitude or demographics? Or perhaps not, but anyone who's ever measured this has found CO doing better than any other state, consistently. The IL anomaly is probably not real, on the other hand.