Trends in Floods

Mary Fran Myers
Natural Hazards Research and Applications Information Center
University of Colorado
Boulder, CO


It has been reported that floods account for more losses than any other natural disaster in the United States both in terms of loss of lives and damage to property and crops (FIFMTF, NHC). Trying to characterize what these losses mean in terms of impacts over time, in terms of the magnitude and frequency of events, and in terms of future socio-economic problems for the nation is difficult at best. This brief paper addresses the questions raised by the workshop organizers by presenting a review of some of the existing data that shed some light on the issues. I rely mainly on the sources available from the Natural Hazards Center's (NHC) careful review of the last 20 years of NOAA's Storm Data set (the analysis is one part of the Center's development of a loss data base for the "Second Assessment of Research and Applications for Natural Hazards" project), other analyses done on data available from the National Weather Service (Yen and Yen, Nishenko, Pielke), and data from the U.S. Army Corps of Engineers' annual flood damage reports to Congress.

Trends in Flood Impacts

Economic Impacts Using data compiled by the National Weather Service from 1903 to 1993, Yen and Yen conclude that average annual flood damages in the U.S. are $2.41 billion. They report that 10-year moving averages of the data show an increase in losses over time in constant dollars. Their linear regression of the damage data shows a 1.17% annual increase in damage value (pg. 559). Pielke's analysis of the same data over a 25-year moving average reveals a similar, steadily increasing trend in losses due to floods. (pg. 77).

The NHC review of Storm Data records on damage due to floods indicates between $19.5 and $195 billion (in 1994 dollars) of property was lost to floods during the 20 year review (Figure 1). Seven of those years had property damages exceeding $1 to $10 billion. The most costly year was 1993 when floods damaged between $3.3 and $33 billion of property. Not surprisingly, 1988, the most costly year for droughts, was the least costly year for floods. During 1988, property damages ranged from $30 to $300 million dollars.

In addition to these property losses, between $8.1 and $81 billion damage was done to crops. The trend line indicates this is an escalating figure (Figure 2). Again, 1988 had the fewest crop losses due to flooding. By adding property and crop losses together, a total of some $27.6 billion to $276 billion dollars in 1994 dollars were lost to floods (Figure 3).

Loss of life and injuries Figure 4 shows the number of deaths per year due to floods from 1975-1994 as reported in Storm Data. The trend indicates a slight reduction in the number of deaths. Pielke draws the same conclusion though also points out that a review of a longer record of data over a moving 25-year period beginning in 1927 suggests that the number of deaths due to floods have been increasing in the second half of the century (Pielke, pg. 74). The Corps of Engineers reported an average of 94 deaths per year due to floods from Fiscal Year 1986-95 (U.S. Army Corps of Engineers, pg 13). Nishenko's review of cumulative flood fatalities since the beginning of the century indicates a similar average (personal communication).

An important point to note is that many flood deaths and injuries occur in single catastrophic events. For example, the number of deaths in 1976 is largely due to the Big Thompson Canyon, CO flash flood that killed 156 people and the 1985 peak is chiefly due to a flood in Puerto Rico that killed 180 people. In contrast, the 1993 Midwest floods that affected millions of more people over a much longer time killed far fewer people than these flash floods.

Figure 5 shows that injuries due to floods have also declined slightly in the past two decades. Injuries were highest in 1982 and lowest in 1988 and 1989. The lowest number of injuries coincides with the lowest death totals.

Environmental Impacts Floods are naturally occurring phenomena that are part of the physical and biological processes which have shaped our nation's landscapes. Smith claims "More than any other environmental hazard, floods bring benefits as well as losses." He points especially to the importance of floods in maintaining ecosystem habitats and soil fertility. (Smith, pg 257). Consideration of the negative impacts of floods on the environment more often appraise the detrimental aspects of human attempts to manage flood prone areas which, in turn, disrupt the natural flood cycle. For example, the Federal Interagency Floodplain Management Task Force reports that human activities have deeply affected floodplains and the nature of flooding. Activities such as drainage of wetlands and land clearance for farming; upstream development that replaces natural vegetation with paved asphalt; and construction of channels, levees, reservoirs change the flood cycle and often result in increased runoff, destruction of riparian habitat, and increased water pollution--all of which can be considered destructive to the natural environment (FIFMTF, pp. 19-21). However, systematic assessment of the environmental benefits and costs of flooding and the environmental benefits and costs of floodplain management programs has not been done.

Indirect Impacts In addition to the impacts described above, floods also have indirect impacts that disrupt society. As is the case for all disasters, these impacts and their costs (or value) are not well documented. However, they include such things as business disruption and loss of income, loss in tax revenues, transportation delays, and spread of illness--both physical and mental. There is also the indirect impact on other government-funded programs. Each tax dollar spent on flood response, relief, and recovery is a dollar not spent on other publicly funded programs such as education.

Trends in Frequency and Magnitude

Are floods occurring more often and are they bigger? From a social perspective, it appears that because flood damages are increasing, one might conclude intuitively that the answer is yes. The NHC review of Storm Data over the last 20 years shows a steady increase in the number of events that have caused at least $50,000 in damages (Figure 6).

However, these numbers must be viewed with caution. They may be on the rise due to something as simple as better reporting. More important, though, damage resulting from floods depends on a variety of factors ranging from the specific characteristics of the flood itself (e.g., velocity and duration) to the specific characteristics of the area inundated (e.g., land use, mitigation measures in place, etc.) to the time at which a flood occurs (e.g., agricultural flood damages can vary greatly based on the stage of crop maturation). Without taking all these kinds of factors into consideration, it is not possible comment on the trends in flood frequency and magnitude.

Limitations of Data

Gilbert White, in an invited comment on the "Floodplain Management in the United States: An Assessment Report" summed the problem, "The report candidly recognizes the severe handicaps of incomplete and inconsistent collection of data on which policy judgments must be based." (FIFMTF, pg. 67). This view is shared by many; each of the sources referenced in this paper include a similar caution about the data being discussed. Staff at the NHC working on the Second Assessment of Hazards Research and Applications project found that the nation has no systematic method of accounting for direct losses much less for indirect losses, has no single repository for loss statistics, and data that is collected by various agencies use different criteria making it impossible to accurately compare loss figures between agencies. Further, the data that is gathered is open to criticism because of the scales involved and the subjectivity with which collection criteria are applied.

An example is the Storm Data which form the basis for some of the material presented above. The categories used to classify damage lump all events between $50,000 and $500,000; between $500,000 and $5 million; between $5 million and $500 million; between $500 million and $5 billion! By using powers of ten, it is obvious that actual damages in any category could vary considerably. Other limitations inherent in Storm Data involve the naming and subsequent coding of events. It appears that in some instances, there was no consistent naming of meteorological events. Further, much of the information in the Storm Data set is gathered by volunteers and is, therefore, subjective in nature.

Based on the analysis of Storm Data and several other loss data bases (e.g., insurance claims, FEMA data) the NHC believes that the reliability of the existing data is questionable at best and is filled with gaps and inconsistencies. Data sets from different agencies cannot be compared much less combined. As a result, losses due to natural hazards, including floods, may be greatly underestimated. Unfortunately, we cannot assess how serious the situation is because losses have never been consistently reported or compiled.

For Better, For Worse?

Are things getting better or worse in regard to the flood hazard? In terms of dollar losses, despite the steady increase in damages described above, there is evidence that in a broader context the picture is not so bleak. The Federal Interagency Floodplain Management Task Force reported over a 70 year period (1916-1985) "Property losses from floods appear to have been fairly constant in relation to the overall national economy." (FIFMTF, pg. 18). Yen and Yen's analysis is similar: "The result shows clearly that the damage as a percentage of the GNP decreases slightly despite the fact that the damage in terms of constant-value dollars keeps increasing.." This led them to conclude that flood mitigation measures put in place over the same time have been effective (Yen and Yen, pg. 560).

One must be careful about such conclusions. First, the reliability of the data on which they are made is very questionable. Second, it is important to note that the conclusions are backward looking. They measure what has happened in the past and do not take into consideration the myriad of factors which may contribute to increasing flood losses in both the short- and long-term future. Some of these factors include an increase in the population and built environment at risk, aging infrastructure which becomes more vulnerable to damage from flood, changing climate patterns which may increase the frequency of flood events, and changing development patterns that increase runoff.

Without careful scrutiny of these types of factors, along with several of the "fallacies of floods" noted by Pielke (pp 13-29) one could be lulled into complacently thinking that things are not getting worse, that we are "holding our own", and that things might be improving slightly. I would argue that is not the case. Rather, we do not know if things are getting better or worse.

Socio/Economic Problems

What, then, are the socio-economic issues for the nation in regard to floods? I believe that this complex question can only be answered if several other questions are answered first. Briefly, these are:

Until such questions can be satisfactorily answered, the jury will remain out on the issue of how serious a threat floods are to the nation.


Federal Interagency Floodplain Management Task Force (FIFMTF), 1992. "Floodplain Management in the United States: An Assessment Report," Volume 1, Summary Report. Washington, DC: Federal Emergency Management Agency, 69 pp.

Natural Hazards Research and Applications Information Center (NHC). 1997. "Assessing Losses and Costs Over the Last 20 Years." unpublished draft report prepared by Betsy Forrest and Alice Fothergill.

Nishenko, Stuart. 1997. personal communication.

Pielke, Jr., Roger A, 1996. "Midwest Floods of 1993: Weather, Climate and Societal Impacts." Boulder, CO: National Center for Atmospheric Research. 159 pp.

Smith, Keith. 1996. "Environmental Hazards." 2nd Edition. London and New York: Routledge. 389 pp.

U.S. Army Corps of Engineers. 1996. "Annual Flood Damage Report to Congress for Fiscal Year 1995." Washington, DC: U.S. Army Corps of Engineers. 17 pp.

Yen, Chin-lien and Ben-Chie Yen. 1996. "A Study on the Effectiveness of Flood Mitigation Measures" in Rivertech 96, Volume 2, proceedings of the 1st International Conference on New/Emerging Concepts for Rivers, pp. 555-562. Urbana, IL: International Water Resources Association. 931 pp.

Societal Aspects of Weather

Workshop's Main Page

Table of Contents