The winter lingered, and then we went from snow on the ground one morning to eating lunch outside the next day. Things change fast: The upstairs is getting hot, and soon our Zoom calls are going to be made even more exciting by shouting over the air conditioning. The kids are chafing to get out; they’re not alone. But the virus is still out there, along with all the people who are not yet infected, and not yet immune.

Across the nation, “lockdowns” are being relaxed. There is a palpable hunger for the pandemic to be done with. Clever people who announced back in March that it was all overblown and would be gone in a few weeks are amazingly still claiming that it is all overblown and will be gone in a few weeks, having somehow not updated their assumptions or learned from the people who argued otherwise.

A lot of people want to know how to reopen safely. That depends on local conditions. What’s going on in San Francisco is not the same as what is happening in New York. But there are considerations that are in common wherever you are and regardless of what the pandemic has done to your community so far.

The first thing is to treat the virus with respect: We’ve seen what it can do. The second is to realize that we are still quite early on in the pandemic. Just because your community has been spared to date doesn’t mean that will continue, especially if the virus has more opportunities to spread. This doesn’t mean we need to maintain the status quo indefinitely; it does mean we have to be smart.

I agree that “lockdowns” cannot last forever — and I don’t think they should. But the beginning of this “reopening” phase doesn’t mean that the doors are about to be flung open and we will all gather in the sunlit uplands beyond the coronavirus. That is nonsense. The pandemic is still among us. After more than 20,000 deaths in New York City — about 1 in every 400 inhabitants — the virus is finally being wrestled to the mat there. But it is gathering pace in smaller communities. The virus will spread from them to others and relaxing our guard will help it.

In Senate testimony this week, National Institute of Allergies and Infectious Disease Director Anthony S. Fauci said that if areas “prematurely open up without having the capability of being able to respond effectively and efficiently, my concern is that we will start to see little spikes that might turn into outbreaks. … The consequences could be really serious.” Sen. Rand Paul (R-Ky.) responded, “We can listen to your advice, but there are people on the other side who say there’s not going to be a surge.”

But the response to both perspectives is the same: It would be disastrously foolish to ignore the potential for a surge. So the answer is to prepare for one and be able to detect it early. This, after all, is why we watch for hurricanes, tsunamis and other natural disasters. A pandemic that we know is already circulating is no different.

Surveillance is essential. Testing for the virus matters for a lot of reasons, including the important ability to do contact tracing. But we also need enough testing to detect what health care has coming its way in the next few weeks. German Chancellor Angela Merkel put it well when she pointed out that the higher the rates of transmission are, the shorter the period of time before cases build up into a surge that overwhelms hospitals and the entire health system, as it did in northern Italy and in New York.

There are a lot of ways this can be done. There are the numbers of new confirmed cases, which can help, but the cases need to be detected early, meaning we need community samples. In general, positive cases that are asymptomatic at the time of testing will be a useful measure: If this number is a large proportion of positive tests, that is a good sign. As it shrinks, that suggests more cases are being missed, because community transmission is increasing and more people are getting tested because they are sick, rather than for surveillance purposes.

After the debacle at the start of the pandemic, much more testing is available, but it’s not equally available everywhere. Environmental sampling of viral genetic material in sewage may be another way to estimate how much transmission is going on in the community, but this needs work to figure out exactly how the signal in the environment relates to transmission.

While surveillance matters, what arguably matters more is what happens once the data are in. This is where our leaders need to determine the triggers for decisive action to reduce transmission when it starts to build in future, as we can be sure it will. These must be decided in advance to avoid damaging delays while arguing over what to do. If the numbers of cases are doubling every few days, as has been the case early on in places hard hit, time to figure out how to respond is a luxury you do not have.

The response could be a staged set of different interventions. This has been the approach taken by New Zealand — which has so far had exceptional success in dealing with the pandemic. They have four levels of increasingly intense restrictions, leading to full lockdown. There are plans to relax from stage three to stage two Thursday.

Triggers and the rationale for reintroducing restrictions must be communicated effectively to the public and not subject to political manipulation in any direction. It’s too important. The “traffic light” approach taken by the city of Berlin is a good illustration. Three different pieces of data go into determining whether the situation is red (bad), yellow (less bad) or green (much less bad). The exact policy consequences of each are less well defined than I would like, but a red alert means that the loosening of restrictions ceases, and, indeed, elements of lockdown may be put back in place.

Red alert in Berlin is defined on the basis of the number of new cases (more than 30 for each 100,000 people), the extent to which the health-care system is already occupied with the pandemic (more than 25 percent of beds are occupied by covid-19 patients) and the effective reproductive number of the virus. This last metric tells us how much transmission is going on, by estimating how many new infections are produced on average for each case of disease. If the effective reproductive number is consistently more than 1, the outbreak will grow. In Berlin, the cutoff for code red is 1.2. That is a relatively low rate of spread. Right now, in most parts of the United States, this number is below 1, but it won’t stay there if we don’t handle reopening right.

Berlin’s approach might not be an ideal guide for us in the United States. For a start, it builds upon a testing rate and initial pandemic response that puts us to shame — Germany has a much better handle on the virus than we do (that’s why German soccer will probably be able to resume play this weekend, while almost every other major professional sports league in the world remains shut down). Wriggle room for politicians remains in the Berlin system, not least in exactly how the effective reproductive number is estimated. But their approach includes clear metrics and consequences. In contrast, the British government announced a nonsensical equation to measure the pandemic consequences of its own lockdown loosening, which attracted widespread mockery and sowed confusion among a public that has endured more pandemic deaths to date than any other European nation.

As U.S. states emerge into the summer and loosen their grip on the pandemic, they should be prepared for it to come roaring back. Nobody knows exactly when or where it will. We have to accept that there are a lot of things we don’t yet know about the virus, or how people will react to reopening. We will probably make different sorts of contacts with each other than we did in the winter before the pandemic: Some places will mandate the use of masks; others won’t. Some places will permit larger gatherings than others, and with that will come the risk of super-spreading events. The amount of time people spend together in the relatively close quarters that appear necessary for most transmission will be crucial. As will how many people are actually immune after infection — it’s not a given that they all will be.

I and a group of colleagues have put together a website that identifies 13 key areas that can form the basis for how we handle the next few months. They start with some of what I’ve talked about here and go on to discuss the other components of a long-term plan to deal with the pandemic. It’s our take on a possible way ahead that responsibly manages the crisis.

I started with talk of the winter easing. But the weekend after that happened here in Boston, it snowed again, and as I type, there is a frost advisory. Things that get better can get worse again. But unlike the temporary reversals of nature as the seasons change, we know that the virus will be back once it gets the opportunity. If we do the right thing now, we can be ready for it. Don’t panic. Do prepare.

Read more: