On July 11, 2017 Leah Binkovitz, a staff writer for the Rice Kinder Institute for Urban Research, published a story on The Urban Edge blog entitled: “Why the 100-Year Floodplain Needs to be Rethought.” This post provides some comments.

The story states that “areas inside the 100-year floodplain have a 1 percent chance of flooding each year and are required to buy federally-regulated insurance. But that measure has been increasingly called into question, along with the role of the flood insurance program.”

The story correctly notes that floodplain managers in the United States have used the 1% annual chance threshold to define the arbitrary boundary between high risk and low risk areas along bayous and rivers since 1968. It is policy decision we’ve made, not a technical one.

It is not true that all structures with an annual chance of flooding of 1% or greater are required to buy flood insurance. There is no law or regulation that requires that. It is almost always required by the lender, in order to replace the structure, which serves as collateral for the loan, if it is damaged in a flood. Owners of structures with an annual chance of flooding lower than 1% are also not required to obtain insurance by regulation or by lenders; but many can and do.

We should encourage discussion about whether the 1% risk level is a good dividing line between “high risk” and “low risk” areas, when, obviously, the risk gradually declines as you move away from each creek or bayou. We should also encourage discussion about the exposure to risk from flooding which occurs in areas outside of bayou floodplains. Local streets and storm drains in neighborhoods can only handle a certain depth of rainfall over a certain period of time. If that, so called, “level of service,” is exceeded, or if drains are clogged or broken, local flooding can occur. Our policy discussion should also consider the public’s risk tolerance and their willingness to pay to reduce their risk exposure.

If we decided to make the policy choice to define the high risk areas as those with a 0.5% or a 0.1% annual chance of flooding, and mandated the purchase of insurance on homes or structures in those areas, what would happen to the size of the floodplain and the community cost burden? It would get much bigger. It would encompass many more homes and structures. Flooding events would trigger more claims and much larger insurance payouts. All of these facts would drastically increase the cost of flood insurance or the cost to federal tax payers, especially in flat, coastal areas like Houston. [This might even be considered analogous to the way the Affordable Care Act mandated the purchase of health insurance so that healthy people (those with a low flood risk) would help subsidize the less healthy people or those with pre-existing conditions (those with a high flood risk).]

The story then describes how many insured losses are located outside the 100-year floodplain, quoting from a study, published online earlier this year and in print in the August issue of Natural Hazards Review.

Why would so many homes and businesses located outside of the 100-year floodplain have flood insurance? Many people obtain insurance on a voluntary basis because it is very inexpensive when compared to risk of loss and the cost of such losses. It is inexpensive because it is subsidized by the federal tax payer (you and me).

The story suggests that floodplain models “don’t take into account changing land use, like loss of wetlands or additional impervious surfaces — both important factors in the fast-growing Houston metropolitan area.”

Floodplains are defined using the most up to date and current information about rainfall intensity and depths, topography, land use, and land cover available at the time. Floodplain models are created by licensed professional engineers using current data and then formally adopted by floodplain administrators. These modeling efforts are extensive and relatively costly, therefore, they are not broadly updated very often. Changes in the watershed outside the floodplain are not addressed until there is restudy of the floodplain associated with that watershed and bayou. So its true we don’t have “real-time” watershed models and floodplain maps. They are only accurate for a short time after they are published in a fast growing watershed. In watersheds without significant development activity, they are accurate for a much longer period of time.

Floodplain models and maps are, however, frequently updated as a result of new development or other changes inside the existing floodplain through a process called a “Conditional Letter of Map Revision (CLOMR)” and a “Letter of Map Revision (LOMR).” The CLOMR is a prediction of the increase or decrease in the base flood elevation and areal extent, before to the project being built. The LOMR revises the map after the project is built. This process allows floodplain managers to update the model as a result of floodplain changes and to inform the affected land owners and communities know about the pending and actual changes. Project Brays is a great local example of this process illustrating the reduction of the floodplain as a result of channel changes.

Wholesale updates of the model for a particular watershed occur less frequently mainly due to funding constraints.  Floodplain managers would conduct these updates more frequently is funding was made available.

So discussions about our risk tolerance, our willingness to pay for reduced risks, our willingness to invest in more frequently updated risk assessments, and who bears the costs of floodplain management and losses are all important.  What do you think?