A recent study found that FEMA failed to capture 75 percent of flood damages within their flood maps between 1999 and 2009. The FEMA flood maps are the primary source for how the government and insurance companies evaluate flood risk and insurance premiums. There are collectively trillions of dollars worth of property that rely on these maps being accurate.
The study, conducted by Rice University and Texas A&M University in Galveston, analyzed flood claims between the years of 1999 and 2009 during which there were five major floods. Each of these five floods did not meet the 100-year flood criteria and thus the FEMA flood maps which are based on 100-year floods should have been sufficient to cover Houston residents.
Unfortunately, 75 percent of flood damages caused by these five floods were not within FEMA's 100-year flood maps. The study was published just days before Harvey struck Houston causing devastating flooding throughout the city and many other cities across coastal Texas and Louisiana.
What went wrong? Why were the maps significantly different from reality in these five storms?
The study analyzed Hurricane Ike in 2008, Tropical Storms Erin and Allison in 2007 and 2001, respectively, and two unnamed rainstorms that resulted in flooding in 2006 and 2009. They compared these rain events to what hydrologists call a 100-year flood, meaning that a flood of that magnitude has a 1 percent or 1 in 100 chance of occurring. All five of the aforementioned flooding events did not meet the 100-year flood threshold.
Generally, FEMA models flood events by determining the amount of water required to flood rivers and streams and once flooded where that water will eventually go. However, this type of modeling can be limiting in low-lying areas such as Houston. First, imagine a town at the base of a broad mountain. If there was heavy rain on the mountain, FEMA's models would do a good job at predicting the amount of water that is likely to flow down the mountain and flood rivers and streams within the mountain town. That is, of course, if there is no significant avulsion of the rivers (where rivers abandon their current path for a new path during flood events).
In the second scenario picture heavy rainfall in a low-lying city with waterways throughout the city. In this scenario, the FEMA model doesn't do a great job at predicting where the water will go as the models are primarily used for overflow of rivers and streams. Flooding in low-lying areas can cause water to flow in practically any direction and is not dependent on the overflow of waterways.
A second limitation of FEMA's flood models is the granularity within the model compared to a complex landscape such as Houston. The models will typically classify whole neighborhoods or groups of neighborhoods with the same land use (green space, forest, concrete parking lot, etc.) and soil type (sand, silt, clay, etc.). This means the model is limited in predicting localized flood events within a specific neighborhood.
These two important limitations of FEMA's models make creating a 100-year flood map a difficult task. This is not to say the men and women of FEMA aren't experts at what they do and invaluable in their work. However, the task of modeling flood zones within a city like Houston requires a more complex, detailed, and innovative computational model. These models exist and it's time to start employing them in difficult to model areas like Houston.
As with many catastrophic disasters, we come out of them realizing the limitations of our tools and systems of protection. It's a similar lesson to the levee breaches from Hurricane Katrina in 2005, and a call for better and more robust protection against the inevitable next natural disaster.
