Experts don’t know exactly how many people died when Hurricane Katrina plowed through New Orleans in 2005, but they do know it was a staggering number. Katrina, the costliest natural disaster in U.S. history and the deadliest hurricane since 1928, killed at least 971 people, according to a 2008 study. Higher estimates put the death toll at around 1,440.

Now, in Texas, Harvey is causing widespread damage and devastation, and the rain continues to dump on Houston. Its death toll is comparably low, at just six people, but storms kill people in surprising ways. According to the 2008 Katrina study, which was led by CDC researcher Joan Brunkard, drowning accounted for the plurality of deaths. But it was followed by injury and trauma at 25 percent, and finally heart conditions, accounting for 11 percent. Seniors were disproportionately affected, with half the victims being over 75, and blacks were 1.7 to 4 times more likely to die than whites.

To find out whether Harvey risks becoming more lethal, I spoke with John Mutter, a geophysicist at Columbia University who led an effort to count the dead in the wake of Hurricane Katrina. A lightly edited transcript of our conversation follows.

Olga Khazan: Describe how you studied the deaths after Katrina.

John Mutter: We’ll never know exactly what the [fatality number] is, but let’s say it’s close to 2,000. You have to go back to 1900 to have a similar event in the United States. It’s more typical of poor countries today, or our country a long time ago. So something weird was going on with Katrina. I was curious about exactly how the counting was being done, who was included, and why the fatality rate was so high.

I went to New Orleans several times, and I asked people how the determination of the cause of death was being done, to determine whether they were hurricane victims or not, how long after the event did they stop counting ... and tried to get a sense of what they were doing.

I also had a website where people could write in and describe the circumstances of the death of a relative. Sometimes it happened quite a few days or weeks after the event. Particularly among elderly people who were displaced, if they weren’t in good condition, which most of them were not, they would often die ... of exacerbation of existing conditions, like heart disease or respiratory failure. Particularly, say, if they lose track of their medications, which happens a fair bit.

Khazan: What were the main differences between Katrina in New Orleans and what we can tell of Harvey so far, in terms of lethality?

Mutter: What looks to be different is not the strength. It’s not unusual in wind speed. What seems unusual about Harvey is the extent of rainfall. When people talk about it being unprecedented, that’s what they’re talking about.

The fatalities that we saw in Sandy and in Katrina were mostly, not exclusively but mostly, seawater from storm surge. In Katrina, storm surge [a huge wave caused by high winds] came up the industrial canal, over top of the levees, and then breached a number of other levees—what came in was seawater from the ocean.

It seems like what’s going on [in Houston is] that the rainfall is biblical. If you’re inland, and it’s raining like hell, and streams that don’t normally flood start flooding, and you can’t figure out where to go, it can be deadly.

Khazan: How did people die during Katrina? Is that a risk here?

Mutter: A number of them were straightforward drowning. People were trapped in their houses, couldn’t get to upper levels, and drowned. But then there were a number who died from the exacerbation of existing injuries. If you have a weak heart, and you’re struggling to get out somewhere, you’re putting yourself at substantial risk.

Khazan: How does the geography of Houston, compared to New Orleans, affect the risks from tropical storms?

Mutter: Houston is flatter. If you’ve ever been to Houston, it’s really a flat place. New Orleans had what they call the bowl. A lot of it is below sea level. When the levees failed, and the seawater came in, it just stayed there. It took days to get out. [Houston] shouldn’t stay flooded as long or as extensively as New Orleans did.

Khazan: From the 2008 study, it looks like most of the Katrina deaths occurred in the first few days, and then they petered out. So far there are just six deaths from Harvey. Are we in the clear?

Mutter: Frankly that’s the number of bodies they’ve found so far. That doesn’t mean that there aren’t others that are as-yet-undiscovered. So people will be saying, “I haven’t heard from Grandma since the storm,” and they’ll be concerned, but they won’t know one way or another. So the numbers will go up. But don’t expect them to go up to 600. Because right here and now, you’d have something like 300, not 6. There aren’t many that are as-yet-undiscovered, but there’ll be some.

Khazan: If you’re right, why would Katrina have been so much more deadly by comparison?

Mutter: [Harvey] doesn’t come in like a storm surge—it’s not this enormous, long, 100-mile wall of water that just inundates. It’s raining, you can get inside, you can try to get yourself away from fast-flowing rivers.

Khazan: What do you recommend for people stranded in their homes in Houston right now?

Mutter: I hope what they’ve done is got themselves a lot of fresh water, canned food, and the like, because no matter how dedicated first responders are, it can take a long time to get to people. Make themselves visible to people in helicopters, with waving bedsheets and the like. Expect and be patient, wait, because no matter what these people do, it takes a while.