The United States, like most of Europe and several other places around the world, is in the early stages of a third wave of rising Covid-19 cases. As of Oct. 29, nearly 9 million people across the country have contracted Covid-19, and almost 230,000 have died from it.

The first wave of infections began in early March and peaked towards the end of April. The second wave began around the second week of June and peaked at the end of July at a much higher daily average case count than the first.

The third wave appears to have begun during the second week of September. Its peak is nowhere in sight. But already average daily case totals are higher, at over 80,000, than the peaks of either of the first two.

In some respects, despite the enormous costs already incurred from Covid-19 in terms of lives, jobs, health, and well-being, the current wave is of greater concern than either of the first two. That is in part because the geographic area affected is substantially larger and more diverse.

The March-April period saw case spikes principally in major urban centers like New York, Boston, Detroit, Chicago, Atlanta, New Orleans, Miami, Seattle and Los Angeles. The June-July period saw a combination of regional hotspots driven by industries like meat packing and agriculture, plus enormous numbers of new infections across the Sunbelt, from Florida and the Carolinas to Texas, Arizona, Nevada, and much of Central and Southern California.

The current wave, however, appears to be spreading broadly throughout most areas of the country, including several that previously had brought the pandemic largely under control.

According to the latest model published by Facebook’s AI Research team, or FAIR, the country as a whole is likely to see as many as 1 million new infections between Oct. 26 and Nov. 8. (Direct Relief has published a dashboard of the results here.)

Comparison with actual changes in case counts in specific areas since Oct. 26 indicates that this forecast is likely to be largely accurate overall, even potentially conservative, although with variations in accuracy in different local areas.

Given the case fatality rate of 2.5% that has marked the Covid-19 pandemic in the United States thus far, an additional 1 million cases would equate to roughly 25,000 additional deaths by Nov. 8.

Viewed in terms of absolute growth in total case numbers, the areas most at risk during this larger, rising third wave include some of the key hotspots from the first two, particularly Chicago, Los Angeles, Houston, Dallas, Phoenix and Miami. In combination, those six cities alone represent a forecast increase of almost 100,000 cases over roughly the next week.

However, viewed in terms of percentage rate of change the map of expected Covid-19 impacts grows much larger, and includes many counties with relatively poor access to hospitals and other health services.

From Wisconsin and the upper half of Michigan, across Minnesota, the Dakotas, Montana, and down through Wyoming, Nebraska, Kansas and Oklahoma, communities representing a diverse mix of urban and rural areas are projected to see dramatic increases in Covid-19. These areas tend to be disproportionately older, with lower access to hospital beds and health resources than in major urban centers.

It is notable as well that Direct Relief has received a significant number of new requests for medical support throughout this area over the past two weeks, backing up anecdotally the underlying dynamics of the forecast model.

In some respects, the areas of greatest concern for the current wave, where high total projected case growth also represents high percentage change, are a pair of urban areas which thus far have not been at the forefront of the pandemic.

One of those is the city of El Paso, Texas. As of Oct. 26, more than 43,000 individuals in El Paso had been infected with Covid-19. The FAIR forecast calls for an increase of over 13,000 new cases by Nov. 8, which represents an increase of nearly one-third. News reports from El Paso, as well as feedback from Direct Relief partners, indicate that the influx of Covid-19 cases is causing enormous strains in the health system there, with hospital rationing and other measures already in place.

Like El Paso, Salt Lake City in Utah is expected to see an increase of over 9,000 new cases by Nov. 8, representing an increase of roughly 20% in only a few days. Hospitals and clinics in the Salt Lake City area are seeing similar strains and burdens.

Direct Relief continues to monitor a range of epidemiological models and datasets to make sure that the U.S. health care safety net is resilient and effective, particularly for the most vulnerable. The Facebook AI Research dashboard is a valuable resource for this effort. The forecast will be updated weekly for the foreseeable future.