Examining the 100-Year Floodplain as a Metric of Risk, Loss, and Household Adjustment

Authors


Wesley E. Highfield, Center for Texas Beaches and Shores, Texas A&M University at Galveston, 200 Seawolf Parkway, Galveston, TX 77553-6175, USA; highfiew@tamug.edu.

Abstract

An understudied, but central aspect in understanding flood impacts is the way we conceptualize, identify, and delineate risk. The 100-year floodplain is the longstanding metric in the United States for determining and acting upon the possibility of an area being inundated. This spatial delineation guides local planning and development decisions, triggers insurance purchases and other household adjustments, and serves as the fundamental indicator for whether it is safe to build a structure on a particular site. However, increasing evidence suggests that the 100-year floodplain is neither accurate nor sufficient in guiding communities and household decisions to mitigate the adverse economic impacts of floods. In this perspective, we examine the effectiveness of the 100-year floodplain as an appropriate marker of risk. First, we review existing studies on location and flood damage. Next, we apply these concepts to repetitive flood losses data in Harris County, Texas. We conclude that the apparent inability of the floodplain designation to effectively capture the likelihood of property damage and potential loss of human life in coastal areas has left potentially millions of property owners unaware of the flood risk and unprepared to mitigate their adverse impacts. Relying on traditional 100-year floodplain boundaries, local decision makers are hampered in their ability to ensure community development occurs in a resilient manner. Finally, we set forth an agenda for future research to better capture the conditions associated with flood risk and account for the large percentage of damage outside the designated floodplain.

1. INTRODUCTION

Flooding has been and continues to be one of the costliest and most chronic natural hazards in the United States. For example, from 1996 to 2007 insured residential flood losses alone totaled over $26 billion (approximately $2 billion per year). Furthermore, between 1960 and 2005, over 2,353 persons were killed and 17,129 people injured by flood events. And, all statistical indicators suggest that the flood problem in the United States has been steadily growing worse over time. For example, average annual property damage caused by floods has increased 54 times over the last four decades, from $51 million in 1960s, to $2.77 billion per year in the 2000s (2000–2008).(1) More recent storms, such as Hurricane Irene and Tropical Storm Lee in 2011 further illustrate this trend as each caused over $1 billion in damage. These negative economic impacts are particularly evident in low-lying coastal areas experiencing rapid population growth and development.

A central aspect in understanding flood impacts is the way we conceptualize, identify, and delineate risk. Traditionally, the process of risk assessment provides a framework for measuring hazard risk by focusing on the probability of events occurring and their expected impacts.(2,3) The first step in the process of flood risk assessment is the translation of occurrence probabilities into the development of hazard maps delineating areas most at risk from inundation.(4) The most ubiquitous outcome of this procedure is the 100-year floodplain, where there is a 1% chance of flooding every year. The floodplain has been the longstanding metric for determining and acting upon the possibility of an area being inundated in the United States. This spatial delineation guides local planning and development decisions, triggers insurance purchases and other household adjustments, and serves as the fundamental indicator for whether it is safe to build a structure on a particular site. In sum, the 100-year floodplain is the key spatial indicator for determining flood risk in the United States.

However, increasing evidence suggests that the 100-year floodplain is neither accurate nor sufficient in guiding communities and household decisions to mitigate the adverse economic impacts of floods. Data obtained from the Federal Emergency Management Agency (FEMA) and other data sources show an increasing amount of flood damages occurring outside of floodplain boundaries.(5) Evidence also suggests that the apparent inaccuracy of the floodplain in predicting actual losses is not solely a function of outdated or poorly constructed floodplain maps. Large percentages of flood loss occur outside floodplain boundaries even in cases where these delineations have been recently updated.

In this perspective, we build on previous national assessments by examining the effectiveness of the 100-year floodplain as an appropriate marker of risk at the local level in Texas. First, we review existing studies on location and flood damage. Next, we apply these concepts to a sample of repetitive flood losses data in Harris County, Texas. Finally, we set forth an agenda for future research to better capture the conditions associated with flood risk and account for the large percentage of damage outside the designated floodplain. On the basis of our analysis, we conclude that the apparent inability of the floodplain designation to effectively capture the likelihood of property damage and possibly human lives in coastal areas has left potentially millions of property owners unaware of the flood risk and unprepared to mitigate their adverse impacts. This problem is exacerbated in low-lying coastal areas as illustrated by our local-level analysis of repetitive flooding in southeast Texas. Furthermore, by relying on traditional 100-year floodplain boundaries, local decision makers are hampered in their ability to ensure community development occurs in a resilient manner.

2. PRINCIPLES AND PRACTICES OF FLOODPLAIN MANAGEMENT

By far the most comprehensive and widely implemented form of flood mitigation in the United States at the household level takes place through the National Flood Insurance Program (NFIP). The NFIP was established in 1968 to provide flood insurance to floodplain residents and businesses. As of 2010, the NFIP had over 23,000 participating communities and over 5.6 million flood insurance policies in force.(6) The NFIP now consists of three components: risk identification, floodplain management, and flood insurance. Risk identification identifies areas that are vulnerable to floods to define levels of risk and determine actuarial rates. The result of this analysis is the Flood Insurance Rate Map (FIRM), which contains, among other delineations, the boundaries of the regulatory 1% flood (often referred to as the “100-year flood” or base flood). Riverine flood risk is estimated in terms of the magnitude of a “design” flood or rainfall event. Areas within the 100-year flood boundary have a 1% chance of being reached or exceeded in any given year. Where sufficient records of streamflow are available, design floods are usually estimated by statistically analyzing past streamflow data. In areas without stream gauges or insufficient data, design floods are usually estimated by either interpolation based on regional streams or generating a design rainfall event through hydrologic modeling to estimate water flow. The resulting hydrologic models are used to estimate the depth and extent of the resulting flood through hydraulic models.(7) In comparison to riverine floodplains, floodplain delineation in coastal areas subject to storm surge is a more complex and uncertain process relying on simulation models.(8) The dynamic nature of lands influenced by coastal factors requires frequency based on historic storm surges, still water elevation levels, and models of wave generation, setup, overland propagation, run-up, and overtopping.(9) In addition, factors such as tides, erosion, and existing structures (e.g., seawalls and other barriers) must also be considered.

A problem with, and frequent criticism of, the NFIP are outdated FIRMS (and Digital FIRMS). In many cases, these maps are so old they do not reflect the true flood risk to households and communities, due primarily to increases in impervious surface coverage through urban and suburban development, which alters the boundaries of the 100-year floodplain. Recent figures indicate that 33% of the maps are more than 15 years old, and another 30% are 10–15 years old.1 The second component, floodplain management, is considered a requirement by FEMA in order for a community to participate in the NFIP. FEMA uses the 1% or 100-year floodplain maps as a basis for mitigation action and as a minimum requirement for participation of local governments.(10) Communities must enact a floodplain management ordinance that meets or exceeds NFIP minimum standards.(10,11)

The third component of the NFIP is its role in providing flood insurance. FEMA, through the NFIP, writes or underwrites flood insurance for participating NFIP communities in the United States. Individuals can purchase flood insurance directly through sanctioned FEMA representatives or through a traditional private insurer in what is known as the “Write Your Own” (WYO) program. There is a mandated requirement to purchase flood insurance for structures located within the SFHA that are being purchased by a federally regulated lending institution. As a result of habitual noncompliance, this requirement has been more forcefully implemented through lenders and loan servicers, requiring them to determine and document whether a structure is in the Special Flood Hazard Area (SFHA) and ensure that the mortgager maintains flood insurance throughout the life of the loan.(12)

3. THE 100-YEAR FLOODPLAIN AS A DRIVER OF FLOOD RISK AND MITIGATION ACTIVITIES

While the 100-year floodplain is a cornerstone of the NFIP and the marker of risk upon which local flood mitigation decisions are made, this delineation has not been a very effective predictor of vulnerability and subsequent adverse impacts from floods. First, the measure itself is susceptible to potential error upon implementation.(13) Sources of uncertainty include: limited hydrometeorological observations(14); spatial–temporal variations in precipitation(7); changes in climate(15); and assumptions made in statistical modeling techniques. Second, these models do not account for development and changing land use as a major contributor to reducing or increase flood effects.(16) However, development within a watershed has the potential to significantly increase the height of a 100-year flood.(17) Third, local conditions (both natural and human induced) tend to change floodplain boundaries more quickly than they can be revised, rendering many flood insurance rate maps outdated.(18) The FEMA Map Modernization Management Support program is the first major initiative to update the nation's flood boundaries and has been ongoing for over six years. Fourth, the 1% flood often occurs more frequently than expected and, in any case, most flood losses in the United States come from less frequent events.(16,18,19) The weaknesses of using the 100-year floodplain boundary for policy and mitigation decisions are only exacerbated in low-lying, rapidly developing coastal areas vulnerable to major storm events.

Despite the uncertainties inherent in delineating the 100-year floodplain, the boundaries are translated into “dichotomous decisions” about where development should be restricted, plans should be made, and whether insurance must be purchased.(7) That is, households and developments are treated equally whether they are 1 foot or 1 mile outside of the designated floodplain. Patterson and Doyle (2009) note that an unintended consequence of making policy decisions based solely on the 100-year floodplain delineation is the potential for development directly outside this boundary. Properties 1-foot outside the floodplain boundary are not required to purchase federal insurance, even though the probability of being flooded is only slightly lower than for those structures located 1-foot inside that boundary. There is no research on how close these properties are to the floodplain boundary, but evidence supports the notion that residential development is shifting to vulnerable edge locations where they are less affected by NFIP policy.(13) The present use of the 100-year floodplain as a strict boundary for triggering household adjustment conveys a false impression of safety to those outside the delineation. Although political motivations may act as a barrier, understanding and better communicating the variability of flood losses to property owners is a starting point for reducing the rising costs of floods in the United States.(20)

The uncertainty in delineating the 100-year floodplain in combination with treating this boundary as a dichotomous decision in terms of insurance requirements and other policies has led to an unexpectedly large amount of losses outside areas considered the most vulnerable. For example, one analysis found that between 1978 and 1996, 23% of insured losses occurred outside the 100-year floodplain.(21) More recent studies report that 30%(12) to 40%(22) of flood loss claims occur outside of the 1% boundary. Our own pilot study analysis examining a representative national sample of 450 jurisdictions shows that approximately 25% of flood insurance claims from 1999 to 2009 were located outside of the 100-year floodplain. This percentage greatly increases for coastal jurisdictions along the coast, where floodplain variability is especially responsive to human and natural changes on the landscape. Again, there is no research examining the proximity of these losses in relation to the floodplain boundary and other natural features. Empirically understanding these issues would lead to a more refined conception of vulnerability and losses associated with flooding.

4. REPETITIVE LOSS IN HARRIS COUNTY, TEXAS

As one of the most flood-damaged jurisdictions in the nation, Harris County, which encompasses the city of Houston, provides an ideal area in which to study chronic flooding. A large sprawling population combined with little slope, poorly drained soils, and episodes of heavy precipitation create prime conditions for repetitive flooding. From 1996 to 2007 alone, we catalogued over $1.1 billion of insured flood damage claims in Harris County, Texas. Among all coastal zip codes in Texas, six of the top ten most flood-damaged were located in Houston. In addition to property losses from floods, Harris County has also incurred the highest number of flood-related fatalities in the United States over the last 50 years.(1)

Out of all of the available flood data sets for the region, we selected for analysis repetitive losses as reported by the NFIP because they represent the properties most at risk to inundation. A repetitive loss (RL) property is any insurable building for which two or more claims of more than $1,000 were paid by the NFIP within any rolling 10-year period, since 1978. Currently there are over 122,000 RL properties nationwide.(23) Because RL damages are chronic, we presume claims for these properties would be located in the most vulnerable areas (such as the floodplain) and less likely the result of an unusual storm event or special circumstance. Focusing on these data thus represents an ideal opportunity to examine the effectiveness of the floodplain as an appropriate marker of risk.

In Harris County from 1978 to 2008, we catalogued 9,521 claims over 2,896 properties for a total of approximately $351 million in paid repetitive losses. On average, each property in the study area claimed 3.23 losses and received $36,881 per claim. Almost 35% of losses during this period were caused by a hurricane or tropical storm, most recently Hurricane Ike in 2008, which generated $20.4 million in RL damage. Of all losses paid, over 70% went to repair building-related damage (vs. contents) of which 94% was from single-family structures. Houston incurred by far the largest amount of total flood loss. But, on a per claim basis, the cities of La Porte and Webster ranked the highest within Harris County.

Unexpectedly, we found that over 47% of all claims were located outside of the designated 100-year floodplain boundary. This percentage holds even when examining claims subsequent to the most recent floodplain boundary update in 2001. Viewing these repetitive losses from an economic standpoint changes these percentages very little. As shown in Fig. 1, losses within the floodplain were concentrated in delineated floodplains (A zones) and accounted for over $201 million or 49% of losses. However, 41% of losses, or $147 million, were outside the 100-year floodplain designation (C and X zones) or in the 500-year floodplain (B zones).

Figure 1.

Repetitive losses by flood zone.

5. CONCLUSIONS AND A CALL FOR FUTURE RESEARCH

Based on our analysis of RL data for Harris County, Texas over the last 30 years, the 100-year floodplain has not had the expected predictive power of property damage across the region. Over 47% of claims and over $147 million dollars in losses paid outside the 100-year floodplain designation indicates that additional factors must be incorporated into risk assessments for flooding, particularly in low-lying coastal areas where the majority of insured damage occurs. This issue is of critical importance considering the floodplain boundary forms the basis of local hazard mitigation plans and triggers household preparedness in the form of flood insurance purchase. Improved measures of flood inundation risk could significantly reduce the amount of losses incurred over time by enabling communities and households to put mitigation measures in place.

As a first step, we propose that the following areas be studied to further explore and improve upon existing risk determinations and decision-making mechanisms. Results from these studies could then be incorporated into existing mitigation frameworks at the local level.

5.1. Spatial Proximity

As already mentioned, policy decisions associated with the floodplain designation are treated dichotomously (in/out) even though there is likely a spatial gradient predicting the frequency and severity of losses. Future research should use spatial analytical techniques to measure the distance of properties and property losses from each flood zone boundary. The findings from these analyses may help reduce what seems to be a false dichotomy of flood risk and provide critical information for mitigating negative impacts at the household level. This would effectively move floodplain management from a dichotomous policy decision to a gradient of risk.

5.2. Alternatives to Traditional Floodplain Delineation

The use of hydrologic and hydraulic models to delineate floodplain boundaries, and thus decision-making levers, is state of the art. However, these maps must be repeatedly developed to retain any accuracy. Changes in landcover, locally and upstream, have outpaced the development of floodplain boundaries. Even when boundaries are generated based on a full “build out” scenario more extreme levels of flooding alter the recurrence intervals of “100-year” floods, making them more probable. Incorporating concepts such as the spatial gradients noted above, local knowledge of flood prone areas, and geographic areas that traditionally flooded before altering riverine and coastal environments could effectively alter the way floodplain delineation is approached.

5.3. Better Understanding of Voluntary Flood Insurance Purchasing

Policy changes move slowly, and flood insurance is the frontline of flood mitigation for the vast majority of people. Although flood insurance is mandatory for those with a federally backed mortgage, purchasing is voluntary for those in floodplains without mortgage requirements and persons outside of floodplains. There is very little empirical research that examines the factors that drive households to purchase flood insurance in low-risk areas, despite high levels of flood damages outside of delineated floodplains. An important step in improved flood mitigation is a better understanding of why voluntary purchasing occurs, especially outside floodplains.

Increasing the number of low-risk flood insurance policies may provide a level of flood mitigation for many households at a low cost.

5.4. Conclusions

Flood losses are a frequent, widespread, and economically disruptive natural hazard—but only because households make decisions to settle in hazardous areas. Despite the best efforts of numerous fields of study, individuals, families, and businesses continue to absorb flood losses in areas where they are led to believe no risk exists. Future research on the preceding topics would be a critical step in addressing the risk gap and reducing the reliance on the dichotomous nature of floodplain delineation and management.

ACKNOWLEDGMENTS

This article is based on research supported in part by the U.S. National Science Foundation (1129998) and Houston Advanced Research Center (10–001). The findings and opinions reported are those of the author and are not endorsed by the funding organizations.

Footnotes

  • 1

    This problem is currently being addressed through FEMA's Map Modernization Management Support program, which aims to update maps to digital (GIS) formats for 92% of the U.S. population (FEMA, 2006).

Ancillary