Center Logo

Natural Hazards Research and Applications Information Center
University of Colorado
482 UCB
Boulder, CO 80309-0482

Quick Response Report #164

Flood Damage Assessment and Survey of
Mitigation Efforts at Stump Lake, North Dakota:
A Study of a Closed-basin Lake Flood

Paul E. Todhunter
Bradley C. Rundquist

Department of Geography
University of North Dakota
Box 9020
Grand Forks, ND 58202-9020



Return to the Hazards Center Home Page

Return to the Quick Response Report Index

This QR report is available in its entirety only in PDF format. Below you
will find the abstract, introduction, and conclusion.

This material is based upon work supported by the National Science Foundation under Grant No. CMS-0080977. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation or the Natural Hazards Research and Applications Information Center.

Citation: Paul E. Todhunter and Bradley C. Rundquist. 2003. Flood Damage Assessment and Survey of Mitigation Efforts at Stump Lake, North Dakota: A Study of a Closed-basin Lake Flood. Quick Response Research Report #164. Boulder, Colorado: Natural Hazards Research and Applications Information Center, University of Colorado. URL:


Since 1993, Devils Lake, a terminal lake in northeastern North Dakota, has risen 7.3 m in response to an unprecedented ten-year regional wet cycle. Lake area has nearly tripled, resulting in more than $400 million dollars in direct flood damages. Beginning in 1999 Devils Lake began to overflow into Stump Lake, a smaller closed lake sub-basin located in Nelson County. Outflow from Devils Lake has combined with local runoff to produce a 1.5-m rise in Stump Lake over the past three years. Extensive rural flooding marked by an increase in the number, size, and permanence of wetlands has also occurred within both lake basins over the same period.

The authors document the flood history of Stump Lake and rural Nelson County, assess the flood damages resulting from the rise of Stump Lake and the growth of rural wetlands in Nelson County, and survey the flood mitigation efforts associated with this closed-basin flood hazard. Remote sensing image interpretation, field work, personal interviews and compilation of data from private, county, state, and federal agencies are used to quantify the direct, indirect, and secondary damages associated with terminal lake and rural wetland flooding in Nelson County. Results document the comparative magnitudes of direct, indirect, and secondary flood damages, as well as the relative contribution of rural wetland and terminal lake flood damages. The study provides a case history of a pervasive, chronic flood hazard not routinely addressed by federal flood mitigation programs.


Closed-basin lakes are one of the most dynamic hydrological systems in the world, and have been shown to be exhibit rapid and significant water surface elevation changes in response to both to climatic fluctuations and human impacts (Williams 1996). Closed basins lack a natural surface outlet to the sea and drain to a closed-basin lake (also called a terminal lake or saline lake). Although researchers have proposed that closed-basin lake levels provide an excellent natural monitor of climate change, their use for such purposes is hindered by the fact that they are also very sensitive to human modification of basin hydrologic conditions (Williams 1996).

Indigenous cultures have adapted to the non-equilibrium environment around closed-basins by developing human use systems and flexible livelihood strategies that are adjusted to the fluctuating lake levels (Evans and Mohieldeen 2002). In the United States, by contrast, resource managers normally view closed-basins through a stability bias lens that fails to recognize the natural variability of these systems (Morrisette 1988). Rigid and inflexible human use systems are forced onto a natural system that is characterized by non-equilibrium conditions. This assumption of climatic and hydrologic stability has led to the development of rigid human use systems for these variable natural systems, and has contributed to a growing national closed-basin lake flood hazard (Association of State Floodplain Managers 1986).

The most recent closed-basin lake natural disaster in the United States is the Devils Lake of North Dakota, where the direct damages for infrastructure repairs, residential and commercial relocations, utility lines, sewage facilities, and recreational facilities has approached $400 million. The water surface elevation time series for Devils Lake since European contact is given in Figure 1. Lake levels in Devils Lake dropped steadily until reaching a minimum of 427.0 m in 1940, when the lake was less than 1 m deep. A long-term wetting trend followed that produced a significant though erratic rise in the lake's water surface elevation through the 1980s. The torrential rains that occurred in the summer of 1993 initiated a dramatic shift to a rapid and steep rise in water surface elevation that has continued through 2001. Since 1993 the water surface elevation at Devils Lake has risen more than 7.0 m.

In July of 1999 Devils Lake reached an elevation of 440.89 m (Figure 1), and water began to spill from East Devils Lake into Stump Lake through the Jerusalem Outlet for the first time in approximately 700 years (Murphy et al. 1997). Stump Lake is a closed-basin lake located within the Stump Lake sub-basin to the east of the Devils Lake basin. Because of its lower WSE (approximately 429.2 m at the time of spillover from Devils Lake) and smaller potential lake volume Stump Lake provides an excellent opportunity to examine the full range of flood damages and flood mitigation efforts associated with a closed-basin lake natural hazard. Because the lake-rise flood hazard is fundamentally different from that experienced with riverine or coastal environments, their flood mitigation approaches will be unique and may be transferable to other closed-basin lake environments (Association of State Floodplain Managers 1986). Furthermore, the entire drought to deluge cycle that has taken more than 60 years to complete at Devils Lake (Figure 1) may be compressed into a much smaller number of years at Stump Lake, making the research problem more amenable to study.

We posed several specific research questions: (1) What is the nature and extent of flood damages due to the rising levels of Stump Lake?; (2) What is the magnitude of flood damages arising from the closed-lake flooding in comparison to flood damages resulting from rural wetland flooding?; (3) What is the relative scale of direct, indirect, and secondary flood damages?; (4) How can GIS and remote sensing be used to assess closed-basin and rural wetland flooding?; (5) What mitigation efforts and programs have proven successful for closed-basin and rural wetland flooding?; (6) Are there any unmet needs and unmitigated flood damages arising from either closed-basin lake or rural wetland flooding? Because the flooding problems around Stump Lake and within Nelson County are so interwoven with similar problems at Devils Lake and within the Devils Lake Basin, it is anticipated that lessons learned from the Stump Lake case study will be useful in providing insights to the larger Devils Lake Basin flood disaster. The study also provides a case study of a unique form of climate hazard associated with slow and unidirectional environmental change that produces pervasive but spatially ill-defined adverse effects (Riebsame 1985).


We return to the research questions posed in the Introduction. First, the primary flood damages resulting from the rising levels of Stump Lake are related to infrastructure, specifically maintaining the transportation network around the lake to provide for the movement of goods, peoples and services from areas south of the lake to the regional trade center northwest of the lake. Funds available through the CMC and PA programs have been adequate to this task, although they have consumed all available local cost share funds, resulting in the more rapid deterioration of the remaining county infrastructure. The lack of prime farmland immediately surrounding the lake has resulted in little direct agricultural damages due to the rising lake levels.

Second, the total flood-related damages for the county has been dominated by those resulting from the pervasive rural wetland flooding, as compared to the more intensive closed-basin flood hazard. Although closed-basin flooding is more likely to be communicated outside the region, rural wetland flooding dominates the Nelson County flood damage profile.

Third, direct flood damages are dominated by government payments for crop insurance claims, land enrolled in the CRP program, and the crop disaster program available through the Presidential Disaster Declaration process. Infrastucture costs are also a significant percentage of the total damages. Indirect costs associated with reduced purchased inputs arising from the large amount of land taken out of production are nearly as large as the direct federal payments. Secondary costs associated with economic multiplier effects resulting from the indirect damages are even larger. Although the majority (approximately 80%) of the direct damages are born by the federal government, all of the indirect and secondary damages are borne by the county or region. The reduced county tax base and income, and reduced flow of dollars through the regional economy, combined with the declining rural population and increased local cost share on infrastructure projects has placed an enormous strain upon the county and regional economy. How can the county and region maintain a viable rural economy when the primary agricultural economic sector is so devastated by pervasive flooding?

Fourth, GIS and remote sensing were found to be useful techniques in the analysis of both closed-basin and rural wetland flooding. In particular, they were helpful in determining the relative magnitude of closed-basin vs. wetland flooding, in establishing the land cover types that were inundated by each flood type, and in determining the future damages due to continued lake rise.

Fifth, the existing CRP, FCIC and PDD programs have proved successful in providing funds to the producers who are the base of the regional economy. Although not primarily intended as a primary flood-mitigation program, the CRP program has helped producers maintain their farm operations. At some point, however, this 'cash-flowing' approach will not be sustainable, particularly if the competition for federal dollars were to significantly change the structure of the Farm Bill, FCIC or CRP programs.

Riebsame (1985) hypothesized that slow onset hazards, such as closed-basin floods, would provide a wider range of mitigation alternatives than other hazard events with more rapid speed of onset, and would allow for more efficient loss mitigation. With the exception of the proactive culvert modifications, the bridge construction, and the reallocation of roads within the CMC system, however, this was generally not found to be the case. Because all of the infrastructure work tied to the CMC program requires a local cost share, work can only be done as the federal funding accrues in an account, and as Nelson County comes up with adequate local cost share matching funds. Similarly, infrastructure work tied to the PA program can only be expended once a PDD has been approved by FEMA, and as local cost share funds become available. The PA program, in particular, lacks a proactive component.

Sixth, the unusual nature of closed-basin floods and wetland expansion flooding, combined with the fact that FEMA flood policy is normally written with riverine or coastal flooding in mind, resulted in a wide number of unmet needs that fall outside of existing federal flood mitigation programs.

We anticipate that many of our findings for Stump Lake-Nelson County are also true for the larger flood event occurring at Devil Lake and the surrounding multi-county sub-basin. Many of the mitigation techniques may also be transferable to other closed-basin flood hazard environments.

Read Complete Report

Return to Top

Return to the Hazards Center Home Page

Return to the Quick Response Report Index

September 2003