The curves on some Covid graphs are looking quite steep, again.

Reported levels of the virus in U.S. wastewater are higher than they have been since the first Omicron wave, according to data from the Centers for Disease Control and Prevention, though severe outcomes still remain rarer than in earlier pandemic winters.

“We are seeing rates are going up across the country,” said Amy Kirby, program lead for the C.D.C.’s National Wastewater Surveillance System. The program now categorizes every state with available data at “high” or “very high” viral activity.

The surge might reach its peak this week or soon after, modelers predict, with high levels of transmission expected for at least another month beyond that.

Hospitalizations and deaths have remained far lower than in previous years. There were around 35,000 hospitalizations reported in the last week of December — down from 44,000 a year earlier — and 1,600 weekly deaths as of early December, down from 3,000. (At the same time in 2020, there were around 100,000 hospitalizations and 20,000 deaths each week.)

Many of the metrics used early in the pandemic have become much less useful indicators of how widely the virus is spreading, especially since federal officials stopped more comprehensive data tracking efforts when they declared an end to the public health emergency last spring. Higher population-wide immunity has meant fewer hospitalizations even with high virus spread, and the sharp decline of Covid test results reported to authorities has made case counts far less relevant.

Wastewater testing remains one of the few reliable instruments still available to monitor the virus. It can signal the start of a surge before hospitalizations begin to rise, and it includes even people who don’t know they have Covid. For many who remain at higher risk from the virus — like those who are older, immunocompromised or already have a serious illness — it’s become a crucial tool helping them understand when to be particularly careful.

But it’s an imperfect metric, useful primarily for identifying if there’s an acceleration of virus spread, not for telling you exactly how much virus is circulating.

The data is often reported as normalized viral copies per milliliter or per gram, a number that is nearly impossible to translate into precise case counts, experts say. It’s also hard to know how comparable two different surges are: A peak in the data may not mean exactly the same thing this year as it did last year.

That’s why many scientists who study the data will say only that it shows the nation is in the middle of a large wave, not whether the surge this winter is bigger than previous ones.

(The C.D.C. doesn’t show the actual concentration levels — its dashboard instead shows how much they have increased relative to when spread was low. Above eight standard deviations is considered “very high.”)

Wastewater testing works at all because “everybody poops,” said David O’Connor, a virologist at the University of Wisconsin-Madison.

Samples of wastewater are captured at or on the way to treatment plants and tested for viral RNA in a lab. But no two samples are perfectly comparable. The amount of RNA in the sample will fluctuate depending on many factors, including the local population at any given time — think of a holiday influx into Miami or a college town emptying out for summer — and how much other material, such as industrial waste, is in the system.

What experts really want to know, said Marisa Eisenberg, a professor at the University of Michigan who runs a wastewater monitoring lab for five sites, is how much virus there is relative to the number of people around — the wastewater equivalent of the per-capita case count.

Some labs “normalize” the data — that is, they adjust the denominator — by looking at the number of gallons flowing through the plant, Professor Eisenberg said. But many sites use something called “pepper mild mottle virus,” a virus that infects pepper plants.

“People have studied this in human sewage and found we shed pretty consistent levels of this pepper virus,” she said. “So that’s a measurement of how many people went to the bathroom in the sewer shed today.”

Once Professor Eisenberg’s team normalizes the results, it sends data to the state and to the C.D.C., which collects information from sites across the country that together account for about 40 percent of the U.S. population.

The C.D.C. then aggregates its data and publishes state, regional and national trends. (Two companies that analyze wastewater, Verily Life Sciences and Biobot Analytics, also aggregate data from hundreds of sites and offer national and local pictures of virus spread.)

But those nationwide estimates can be tricky.

The sample population the C.D.C. looks at largely excludes people with septic tanks and cities with no wastewater testing. There can be data lapses, as when the C.D.C. switched contractors last year. Existing sites can stop testing, and new sites start up, as the network changes and expands.

And while Biobot and Verily can use the same methodology and normalization across all their sites, the C.D.C. has to determine trends from data that arrives from different sites with a variety of methodologies.

Finally, there are changes to the virus itself that could make comparisons over time more difficult. Scientists who track those changes say there are hints that this latest variant, JN.1, may be able to better replicate in the gut.

It’s still just a hypothesis, said Dr. O’Connor, the virologist. But it’s possible the virus is “a little more cozy in the gut” than it used to be, he said. If the hypothesis proves correct, it could mean that infected people shed more viral copies than they used to. In the wastewater data, the same number of infections could look like a lot more Covid.

All of that together creates significant uncertainty about how comparable the data is from year to year.

Michael Mina, an epidemiologist and chief science officer for eMed, estimates the real amount of Covid spreading could be quite a bit higher or lower than this time last year. But there’s no doubt that there’s a lot of virus, he said. And far more now than just a few months ago.

Many experts who study this data recommend dropping any notion of precision and just squinting a little at the line’s recent trajectory. And look at your city’s wastewater if possible, since data for a single site tends to be more reliable over time than a national estimate.

“If you have vulnerable people in your community or family, you want to be particularly aware when cases are going up, and take more precautions,” Dr. Mina said. “And when cases are going down or at a lull, relax those precautions.”

Those precautions include wearing a high-quality mask, getting vaccinated, testing and staying home if sick — and if someone at high risk is infected, taking Paxlovid.

Even in this new pandemic phase, people are still dying and can still get long Covid, said Maria Van Kerkhove, the technical lead on Covid for the World Health Organization. “While the crisis of Covid is over, the threat is not,” she said.