As catastrophe modeling firms continue to update their earthquake research, experts say that man-made activity is being factored into some of their models.

Andrew O’Donnell, senior engineer at AIR Worldwide, a unit of Verisk Analytics Inc. said the Boston-based firm updated its U.S. earthquake model to include man-made or induced earthquakes that are caused by the disposal of waste water from fracking operations.

“People who live in areas like Oklahoma and Kansas, which previously had been thought of as dormant areas not necessarily impacted by seismicity, are now living in seismic areas that are on par with the Pacific Northwest in terms of seismicity and average annual loss,” Mr. O’Donnell said.

The Oklahoma Office Of the Secretary of Energy & Environment reported an increase in earthquakes over an eight-year period that ranges from 41 in 2010 to a high 903 in 2015. The number dropped to 304 in 2017.

“While we know that Oklahoma has historically experienced some level of seismicity,” the office said on its website, “we know that the recent rise in earthquakes cannot be entirely attributed to natural causes. Seismologists have documented the relationship between wastewater disposal and triggered seismic activity.”

The secretary’s office also said the Oklahoma Geological Survey has determined that most recent earthquakes in central and north-central Oklahoma “are very likely triggered by the injection of produced water in disposal wells.”

Mr. O’Donnell said the increase in seismic activity in fracking areas is “something that risk managers and insurance companies need to understand if they’re writing business in those areas — that this is a real hazard and it has its own intricacies.”

“It’s an evolving hazard,” he said. “There is a human element, so if it’s a result of our behavior it can also be mitigated by changes in our behavior.”

Maiclaire Bolton-Smith, Oakland, California-based senior leader of research and content strategy for CoreLogic Inc., said that the number of earthquakes has decreased in Oklahoma over the past few years because of a reduction in the amount of waste water being pumped into the ground.

“We haven’t incorporated this into our models yet, but it’s something we’re looking into,” she said. “The challenge is that when you usually model earthquakes, you look at a long-term forecast of what is expected to happen, because earthquakes don’t happen very frequently and the rate of earthquakes happening doesn’t change very frequently. With induced seismicity, it’s changing very rapidly.”

In order to produce a realistic model, Ms. Bolton-Smith said, it would have to be updated very frequently because the rate of earthquakes is changing quiet rapidly.

The Reston, Virginia-based U.S. Geological Survey said the number of earthquakes in the central and eastern United States “has increased dramatically over the past few years.”

Between 1973 and 2008, the USGS said, there was an average of 21 earthquakes of magnitude 3 and larger in the central and eastern United States. The rate jumped to an average of 99 M3+ earthquakes per year in 2009–2013.

Earthquakes in the magnitude 3-4 range are large enough to be felt but small enough to rarely cause damage, the USGS said on its website.

The USGS estimates that there are 500,000 detectable earthquakes in the world each year, of which 100,000 can be felt and 100 cause damage, the survey said.

“People often think we have more earthquakes now than we ever had before, and that’s not true at all,” Ms. Bolton-Smith said. “What’s changed is that there’s a lot more monitoring stations, so we’re able to detect more earthquakes that we potentially couldn’t have detected previously.”

In 2016, CoreLogic released its U.S. earthquake model, which was based on the Uniform California Earthquake Rupture Forecast, Version 3, or UCERF3.

“UCERF3 pushed boundaries of what was happening in the earthquake world and said there was a potential that earthquakes could essentially jump from one fault to another fault when they are rupturing if the faults were close one another,” she said.

Population shifts are also changing earthquake risk, said Carl Tape, associate professor of geophysics at the University of Alaska Fairbanks.

In the Pacific Northwest, for example, there has not been a magnitude 9 earthquake in the area since the Cascadia earthquake in 1700, he said.

“That’s eventually going to happen, and it’s going to happen at a time there are millions and millions more people living there than the last time,” Mr. Tape said.

The discussion of earthquake modeling often leads to the concept of earthquake prediction, something that experts said is unlikely.

“It’s an area of ongoing research, but it seems to be one of those Holy Grail things that we haven’t gotten a good answer for yet,” Mr. O’Donnell said.