Agricultural Economics Report No. 432 December, 1999 The Feasibility of Wetland Restoration to Reduce Flooding in the Red River Valley: A Case Study of the Maple (ND) and Wild Rice (MN) Watersheds Steven D. Shultz Department of Agricultural Economics Agricultural Experiment Station North Dakota State University Fargo, North Dakota Acknowledgements Funding was provided by the Red River Basin Task Force of the International Joint Commission (IJC), the North Dakota State Water Commission, the Minnesota Department of Natural Resources, and the Agricultural Experiment Station at North Dakota State University. Much of the research reported here is closely related to, and dependent upon, the work of others. Melanie Bengtson and Pad (G. Padmanabhan) from the Civil Engineering Department at NDSU estimated hydrologic models to quantify the relationships between wetland storage and flooding in the Maple and Wild Rice Watersheds. And, Mike Kjelland, for his MS Thesis in the Department of Agricultural Economics at NDSU, quantified flood-related damage over time in the two watersheds. Dr. Jay Leitch at NDSU was involved in all aspects of the IJC research and provided substantial assistance and input into this research project. I would also like to thank various faculty members, students, and staff in the Department of Agricultural Economics at NDSU for their help and assistance. In particular, Kevin Kermes and Pat Fridgen for compiling land rental data and making maps, and to Dave Lambert and Carol Jensen for helping to review and prepare this document. I would also like to thank Curtis Borchert of the Norman County (MN) Soil and Water Conservation Board for providing GIS data, the Wild Rice Watershed District for providing watershed information, Moore Engineering of Fargo for providing data on the Maple Watershed, and to Brett Hovde at the North Dakota State Water Commission for providing information about the Devils Lake impoundment-storage program. Any errors or omissions in the report are the sole responsibility of the author. We would be happy to provide a single copy of this publication free of charge. You can address your inquiry to: Carol Jensen, Department of Agricultural Economics, North Dakota State University, P.O. Box 5636, Fargo, ND, 58105-5636, Ph. 701-231-7441, Fax 701-231-7400, e-mail [email protected] . This publication is also available electronically at this web site: http://agecon.lib.umn.edu/ndsu.html NOTICE: The analyses and views reported in this paper are those of the author. They are not necessarily endorsed by the Department of Agricultural Economics or by North Dakota State University. North Dakota State University is committed to the policy that all persons shall have equal access to its programs, and employment without regard to race, color, creed, religion, national origin, sex, age, marital status, disability, public assistance status, veteran status, or sexual orientation. Information on other titles in this series may be obtained from: Department of Agricultural Economics, North Dakota State University, P.O. Box 5636, Fargo, ND 58105. Telephone: 701-231-7441, Fax: 701-231-7400, or e-mail: [email protected]. Copyright © 1999 by Steven D. Shultz. All rights reserved. Readers may make verbatim copies of this document for non-commercial purposes by any means, provided that this copyright notice appears on all such copies. Table of Contents Page List of Figures and Tables.............................................................................................................ii Highlights ..............................................................................................................................iii Abstract ..............................................................................................................................iv Chapter 1: Introduction.................................................................................................................1 Chapter 2: Literature Review and Background Information........................................................4 2.1 2.2 2.3 2.4 The Maple River Watershed ......................................................................................4 The Wild Rice River Watershed ................................................................................7 Historical Flood Damage in the Watersheds..............................................................8 The Hydrologic Modeling of Wetland Storage and Flooding..................................16 Chapter 3: The Feasibility of Wetland Restoration-Storage Programs......................................21 3.1 3.2 3.3 3.4 3.5 Quantities and Locations of Previously Drained Wetlands .....................................21 The Costs of Wetland Restoration-Based Storage Programs...................................22 The Potential Benefits of Wetland Restoration-Storage Options.............................29 Net Present Values and Benefit-Cost Ratios of Wetland Restoration Options........30 Wetland Restoration Feasibility Considering Additional Benefits of Wetlands......30 Chapter 4: The Feasibility of Impoundment-Storage Programs ............................................... 37 4.1 An Introduction to Impoundments for Flood Control..............................................37 4.2 The Feasibility of Impoundment Storage in the Maple and Wild Rice Watersheds.........................................................................................38 Chapter 5: Summary and Conclusions.......................................................................................43 References .............................................................................................................................45 Appendices A: Abbreviations Used in the Report .............................................................................48 B: Impoundment-Storage Programs in North Dakota and Minnesota ...........................49 List of Figures and Tables Figure 1 2 3 4 5 6 7 8 9 10 11 Page Study Site Location: Watersheds within the Red River Valley ........................................2 Land Use in the Maple Watershed ....................................................................................5 Wetland Types in the Maple Watershed ...........................................................................6 Land Use in the Wild Rice Watershed..............................................................................9 Wetland Types in the Wild Rice Watershed ...................................................................10 Historical Flood Damage in the Maple Watershed .........................................................13 Historical Flood Damage in the Wild Rice Watershed...................................................14 Residential Flood Damage by Township in the Wild Rice Watershed...........................15 Drained Wetlands and Hydrology, Maple Watershed ....................................................23 Drained Wetlands and Hydrology,Wild Rice Watershed ...............................................24 Dollars Per Acre for Water Storage on Croplands in the Wild Rice Watershed ............27 Table 1 2 3 4 5 6 7 8 9 10 11 12 13 Page Flood Damage and Characteristics, Maple Watershed (1989-1998) ..............................12 Flood Damage and Characteristics, Wild Rice Watershed (1989-1998) ........................12 Effects of Wetland Storage in Reducing Peak Flows During Flood Events in the Maple River at Enderlin, ND ................................................................................17 Wetland Acres by Size Classes in the Maple and Wild Rice Watersheds ......................22 Wetland Restoration Costs in the Maple and Wild Rice Watersheds Over a 20-year Period .....................................................................................................28 Historical and 20-Year Expected Reductions in Flood Damage Due to Wetland Storage in the Maple Watershed.......................................................................31 Historical and 20-Year Expected Reductions in Flood Damage Due to Wetland Storage in the Wild Rice Watershed.................................................................32 Benefit-Cost Ratios of Alternative Wetland Restoration-Storage Options in the Maple and Wild Rice Watersheds.............................................................................33 Required Reductions in Peak Flood Stage and Flood Damage for Alternative Wetland Restoration-Storage Options to Break Even..................................34 Required Values of Other (Nonflood-Related) Wetland Benefits in Order for Wetland Restoration with Full Environmental Services to be Feasible ....................36 Cost of Impoundment Storage in the Maple and Wild Rice Watersheds........................39 Historical and 20-Year Expected Reductions in Flood Damage Due to Impoundment-Storage in the Maple Watershed..............................................................41 Historical and 20-Year Expected Reductions in Flood Damage Due to Impoundment Storage in the Wild Rice Watershed........................................................42 ii Highlights The economic feasibility of wetland restoration and the use of impoundments to store water and reduce flood damage, were evaluated in two Red River Valley watersheds: the Maple River Watershed in North Dakota and the Wild Rice River Watershed in Minnesota. The economic evaluations utilized information from past and concurrent hydrologic modeling and wetland restoration studies, National Wetland Inventory data, local land rental values, and site-specific historical flood damage. Three wetland restoration options were evaluated in each watershed: simple restoration (one foot of available storage ‘bounce’ per surface acre), restoration with outlet control devices (two feet of bounce), and full restoration providing a full range of wetland-based environmental goods and services and with two feet of bounce. For each alternative, the use of average and lower-end (cheaper) land rental rates was evaluated. Finally, a highly optimistic scenario was evaluated where the least expensive alternative (simple restoration on cheap land) was hypothesized to provide two feet of storage bounce. Economic feasibility was evaluated through benefit-cost ratios calculated for a future 20year period. Costs included land rentals, and restoration and construction costs, which varied by type of restoration and as a function of wetland size. Benefits (reduced flood damage) for a hypothetical 20-year future period were based on historical flood damage (1989 to 1998) both within and downstream of the watersheds along with hydrologic modeling estimates of reductions in peak flood stage with alternative storage scenarios and flood events in the Maple River Watershed (Bengtson and Padamanabahn, 1999). None of the wetland restoration options were found to be economically feasible in reducing flood damage in either of the two watersheds. Benefit-cost ratios ranged from 1:1.4 for the most optimistic scenarios, to 1:4.1 for restoration with the provision of full environmental services. Reductions in peak flood stage and damage would need to be reduced from 3.4% to 10.2% for different wetland restoration options to break even. The inclusion of additional (nonflood-related) wetland values did not make wetland restoration economically feasible. However, these additional wetland benefits have the potential to improve the feasibility of wetland restoration-storage activities, especially at the local (sitespecific) level. More research in this area is warranted. Finally, the use of impoundments to store water and reduce flood damage in the two watersheds was evaluated. Impoundments can store more water at a cheaper cost than wetlands but they do not provide many additional goods and services. With benefit-cost ratios which ranged from 1:1.1 to 1:1.7, impoundments are also not considered economically feasible. In conclusion, neither the restoration of wetlands nor the construction of impoundments on a large scale, are likely to be an economically feasible way to reduce flood damage in the Red River Valley. However, future research is warranted with respect to the use of impoundment storage on a site-specific basis and the calculation of the nonflood-related benefits of restored wetlands. iii Abstract The economic feasibility of wetland restoration and the use of impoundments to store water and reduce flood damage were evaluated in two Red River Valley watersheds: the Maple River Watershed in North Dakota and the Wild Rice River Watershed in Minnesota. The economic evaluations utilized information from past and concurrent hydrologic modeling and wetland restoration studies, National Wetland Inventory data, local land rental values, and sitespecific historical flood damage. None of the wetland restoration options were found to be economically feasible in reducing flood damage in either of the two watersheds. Benefit-cost ratios ranged from 1:1.4 for the most optimistic scenarios, to 1:4.1 for restoration with the provision of full environmental services. Reductions in peak flood stage and damage would need to be reduced from 3.4% to 10.2% for different wetland restoration options to break even. The use of impoundments was also found not to be an economically feasible method in reducing flood damage. Future research is warranted with respect to the use of impoundment-storage for flood control on a site-specific basis and the calculation of the nonflood-related benefits of restored wetlands. Key Words: wetland restoration, flooding, Red River Valley iv Chapter 1. Introduction Steven D. Shultz∗ Flooding in the Red RiverValley (RRV) has historically caused large-scale physical and economic damage to public and private property. There is wide consensus that new strategies are needed to improve the prediction, control, and mitigation of future flood events (IJC, 1997). However, there is a great deal of uncertainty and disagreement regarding the most efficient ways to reduce flood damage. There has been considerable speculation and debate regarding the relationships between wetlands and flooding in the RRV. On one side of the issue are those who believe that wetlands, by storing water, can effectively reduce flood impacts and that the wetland drainage in the last century has exacerbated recent low frequency flood events (Sierra Club, 1997 and 1998). Others believe that wetlands have no significant impacts on floodwater levels in the RRV, especially during larger (low frequency) flood events during the early spring when wetlands are already full. This present study will evaluate the economic feasibility of restoring previously drained wetlands to reduce downstream flood damage both within and on the main-stem of the RRV. This will include an evaluation of ‘simple’ wetland restoration (primarily for flood control), and more ‘complex’ restoration efforts that include the construction of outlet control devices and the restoration of a full range of wetland-based environmental goods and services. A secondary objective is to evaluate the economic feasibility of establishing impoundments (earthen dikes, constructed in agricultural fields with outlet control devices) for storage in order to reduce downstream flood damage both within sub-watersheds and on the main-stem of the RRV. Many of the same doubts and uncertainties concerning the effectiveness of wetland storage to reduce downstream flooding exist for impoundment-storage projects. However, impoundment-storage programs have already been initiated in the nearby Devils Lake area of North Dakota and the Wild Rice Watershed of Minnesota. This study will focus on two RRV Watersheds: the Maple River Watershed in North Dakota and the Wild Rice River Watershed in Minnesota which are located west and northeast of Fargo/Moorhead and south of Grand Forks/East Grand Forks (Figure 1). They are typical RRV watersheds, dominated by agricultural land use and with large acres of both existing and drained wetlands, and with frequent spring flood events. ∗ Assistant Professor, Department of Agricultural Economics, North Dakota State University, Fargo. Figure 1. Study Site Location: Watersheds within the Red River Valley. Watersheds in the Red River Basin Alternative wetland restoration and impoundment options are evaluated only for the middle to upper reaches of the two watersheds (hereafter referred to as the upper watershed) for a variety of reasons. First, most of the recently drained wetlands identified by the National Wetland Inventory (NWI) are located in the upper sections of the watersheds (93% in the Maple and 74% in the Wild Rice). Second, land values in the upper watersheds are lower. Finally, a previous hydrological-based study of the same watersheds (based partly on simulated data), found that upper wetland storage in contrast to wetland storage in the middle and lower reaches of the watersheds provided the greatest reductions on streamflows (Meyer, 1998). Evaluating the feasibility of wetland restoration and impoundment options to reduce flood damage in these two watersheds involves three distinct components. First, the costs of alternative wetland restoration options and impoundment construction are determined. Second, benefits associated with reduced flood damage both within and outside the watersheds are estimated. Finally benefit-cost ratios under alternative scenarios are calculated. The costs of restoring drained wetlands and establishing storage impoundments will be estimated under various scenarios in the upper sections of the watersheds. Data for these estimates are based on the NWI, local agricultural land prices, previously conducted wetland 2 restoration efforts (Sip, 1998 and Eppich et al., 1998), and the experiences of two ongoing impoundment-storage projects in, and near, the RRV. Alternative types of restoration will be evaluated, specifically simple wetland restoration, more complex restoration with outlet control devices, and complete restoration that captures a full range of ecological benefits associated with wetlands. It will be assumed that wetland construction costs are subject to economies of scale or, more specifically, that per acre construction costs are lower for large wetlands. The quantity of wetlands potentially available for restoration is based on wetlands classified as ‘drained’ in the NWI database. It is likely that this database may under-represent the actual quantity of drained wetlands in the watersheds, but this should not effect our feasibility results, which are based on benefit-cost ratios. In other words, if the quantity of wetlands available for restoration increases, it is assumed that associated benefits and costs will increase proportionately. Regarding impoundments, it is assumed an equal quantity of land available for wetland restoration is available for the construction of impoundments. The potential benefits of alternative wetland restoration and impoundment-storage programs are reduced flood-related damage resulting from the creation of wetland restorationand impoundment-based storage. Avoided flood damage is based on annual flood damage from 1989 to 1998, best estimates of available storage capacity or ‘bounce’ of wetlands, and the hydrologic relationships between wetland storage and historical peak flood stage in the Maple Watershed calculated in a concurrent study by Bengtson and Padmanabahn (1999). These reduced flood damage benefits from 1989 to 1998 are then extrapolated to a hypothetical 20-year time period in the future. Benefit-cost ratios are used to measure the feasibility of alternative wetland restoration and impoundment-storage options over a discounted 20-year period in the future. Finally, required reductions in peak flood stage (and reduced flood damage), and minimum values of nonflood-related wetland benefits required for alternative wetland restoration-storage options to break even are estimated. These ‘what if’ estimates are intended to facilitate future evaluations of the economic feasibility of wetland restoration based on improved hydrologic modeling procedures and/or improved estimates of the values of wetland functions. 3 Chapter 2. Literature Review and Background Information The physical and socioeconomic characteristics of the two sub-watersheds, which are the focus of the study, are summarized below. This is followed by an evaluation of the historical flood-related damage inside and just downstream of the watersheds, and a review of hydrological-based models and related literature which forms the basis for many of the assumptions used in this study. 2.1. The Maple River Watershed The Maple River Watershed (hereafter referred to simply as the Maple Watershed) is a sub-watershed of the RRV, located in eastern North Dakota, approximately in the geographic middle of the RRV basin (Figure 1). The watershed transects parts of five counties (Steele, Barnes, Cass , Ransom, and Richland), and 63 townships. Its total area is approximately 1,024,000 acres (1,600 square miles) and it contains 48 sub-watersheds (as classified by Moore Engineering). The Maple Watershed originates in the upper reaches of a five-county drainage area (northern Richland, northern Ransom, southern Steele, northeastern Barnes, and northwestern Cass counties). The Maple River flows southward through a well-defined valley between the towns of Chafee and Enderlin. At Enderlin, the river changes direction and flows to the northeast, meandering across the flat, RRV Plain. Approximately five miles to the north of West Fargo, the Maple River joins the Sheyenne River. Approximately 25% of the eastern watershed is classified as Glacial Lake Plain. Moving west, the watershed rises out of the RRV. This transition area is classified as the Beach Ridge Area and accounts for roughly 15% of the watershed. Glacial Lake Aggassiz fluctuated throughout time, leaving distinct markings on the landscape in the form of beaches. The western and northern portions of the upper watershed are to the west of the Beach Ridge Area. This sub-basin area is characterized by rolling uplands and is developed on a recessional moraine and accounts for approximately the remaining 60% of the watershed. The landscape characteristics of this area were formed by the soil, rocks, and debris deposited by the glaciers (Houston Engineering, Inc., 1999). Land use in the Maple Watershed is influenced by its geophysical characteristics. Agriculture predominates in the basin and the principal agricultural crops include wheat, barley, oats, sugarbeets, potatoes, corn, beans, forage grasses, and sunflowers (see Figure 2). Wheat is rotated with beans and potatoes and is planted mostly in the northern half of the basin while corn and soybeans are planted mostly in the southern third of the basin. Based on the NWI compiled by the United States Fish and Wildlife Service (USFWS), within the Maple Watershed there are 48,342 acres of wetlands of which 2,900 are drained (Figure 3). 4 Figure 2. Land Use in the Maple Watershed. Source: National Agricultural Statistics Service. 5 Figure 3. Wetland Types in the Maple Watershed. 6 Prior to several U.S. Army Corps of Engineers (USACE) flood mitigation projects in the late 1980s, flooding in the Maple Watershed occurred regularly in the city of Enderlin where the Maple and South Branch Rivers converge (International Coalition, 1989). East of Enderlin, waterflows have a high velocity until slowing greatly in the plains area north and east of Leonard. In this flat, Glacial Lake Plain area, channel capacity decreases and the potential for flooding greatly increases. The Cass County Joint Water Resources District (CCJWRD) is currently evaluating alternatives that may alleviate flooding in the watershed, particularly along the lower Maple River. According to the 404 Permit for Maple River Dam Draft Environmental Impact Statement (DEIS), water storage through main-stem dams, tributary dams, or wetland restoration are the only alternatives that are both technically feasible and meet the purpose and need of the project (USACE, 1998). Eleven plans were evaluated, with the preferred plan involving one main-stem dam located downstream of Enderlin, approximately half way between Enderlin and Leonard. This plan would protect 7,750 acres at a cost of $1,806 per acre protected. 2.2. The Wild Rice River Watershed The Wild Rice River Watershed [hereafter referred to simply as the Wild Rice Watershed (WRW)] is a sub-watershed of the Red River Watershed. It is located in northwestern Minnesota, just south of the approximate mid-line of the RRV basin. The watershed transects 6 counties (Becker, Clay, Clearwater, Mahnomen, Norman, and Polk) and 86 townships with a total area of 1,293,467 acres (2,021 square miles). Approximately 25% of the western watershed is Glacial Lake Plain (lakebed region of the Glacial Lake Aggassiz) and is nearly flat. Moving east, the watershed rises out of the RRV in a transition area classified as the Beach Ridge Area, which accounts for roughly 15% of the watershed. Glacial Lake Aggassiz fluctuated throughout time, leaving distinct markings on the geography in the form of beaches. This Beach Ridge Area follows a north-south corridor approximately eight miles wide (Houston Engineering, Inc., 1999). The glacial moraine is to the east of the Beach Ridge Area and accounts for approximately the remaining 60% of the watershed. The landscape characteristics of this area were formed by the soil, rocks, and debris deposited by the glaciers (Houston Engineering, Inc., 1999). Lakes, ponds, bogs, and wetlands dot the landscape. The terrain is gently rolling with occasional steep grades. In close parallel to the physiographic regions, there are three distinctive soil groupings in the watershed (Houston Engineering, Inc., 1999). Clay soils primarily dominate the Glacial Lake Plain in the western watershed. These soils are highly productive for agriculture, have low permeability, and have poor internal drainage. The soils in the Beach Ridge Area are more diversified, ranging from clay loams and sandy loams to sands and gravels. Clays and silts essentially make up the glacial moraine. Drainage within the soils of this area varies from poor to good. 7 Land use in the watershed is largely influenced by its geophysical characteristics. Cropland predominates in the western basin, while the far eastern basin is largely forested. Approximately 67% of the watershed is cultivated, 18% is forest, and 7% is in pasture (Figure 4 and IJC-NDSU, 1999). The principal crops grown are wheat, barley, oats, sugarbeets, potatoes, corn, beans, forage grasses, and sunflowers. Agriculture largely influences the economy of the western and central parts of the watershed while tourism, lake cottages, and related recreational activities are a major part of the local economy in the eastern part of the watershed which contains many lakes and forests. Based on the NWI compiled by the USFWS, within the WRW there are 120,500 acres of wetlands (Figure 5). In addition, there are 17,000 acres of drained wetlands and about 800 impounded acres. Within the upper watershed, there are about 100,000 total wetland acres, 12,600 drained acres, and 350 impounded acres. As discussed later, most of the drained wetlands are associated with large-size wetlands (greater than 5 acres). Flooding in the WRW usually occurs during April to June and occasionally in March and July. Snowmelt is the primary contributor to flooding within the Wild Rice basin, with spring rains often adding to the runoff. Saturated soil, frozen ground, and ice-covered rivers also contribute to flooding. During the early stages of snowmelt, river channels are often clogged by hard packed snow and ice which can increase local flood stages by several feet (USACE, 1980 and Kranz and Leitch, 1993). Water flows are generally contained in definable areas of the upper watershed (USACE, 1980). Stream channels in the Beach Ridge Area generally are deeper and more defined with water contained within stream banks. Channel capacity decreases upon entering the flat valley plain resulting in overland flooding. The western portion of the lakebed region is subject to regular flooding by the Red River of the North (Houston Engineering, Inc., 1999). 2.3. Historical Flood Damage in the Watersheds Flood damage reports in the RRV have, for the most part, tended to be very general in nature or they focussed on single flood events at a particular site (most often Fargo or Grand Forks). Examples of such damage reports can be found in Krenz and Leitch, 1993; IJC, 1997; and Carlson, 1999. Of relevance to this study is flood damage that has occurred both within the Maple and Wild Rice Watersheds and damage outside the watersheds downstream of the mainstem of the Red River. 8 9 Figure 4. Land Use in the Wild Rice Watershed. 10 Figure 5. Wetland Types in the Wild Rice Watershed. Flood Damage Within the Maple and Wild Rice Watersheds The quantification of economic damage associated with flooding in the Maple and Wild Rice Watersheds over time (from 1989 to 1998) was undertaken specifically to assist with this present study. This required the collection, assembly, and modification of data under varying assumptions, from local, state, federal and nongovernmental agencies (over 40 different agencies and institutions in both the public and private sector were contacted). Of particular importance were data obtained from the Federal Emergency Management Agency (FEMA). Agricultural damage data from the USDA was not provided because it was not possible to separate floodrelated from other damage and assistance data. In many cases, data were reported to us only at the county or regional level of analysis. In order to obtain needed flood damage data at the watershed level of analysis, the following tasks were undertaken. First, township-specific damage report data obtained from FEMA were manually classified to identify in which counties and/or townships the damage was located. Second, agricultural damage for the watersheds was estimated by manipulating and extrapolating the ‘Computerized Agricultural Crop Flood Damage Assessment System’ (CACFDAS) model previously estimated by the USACE in the WRW and by a local engineering firm in the Maple Watershed. Third, county-level agricultural damage data occurring in a watershed were estimated by dividing the total agricultural damage reported by individual counties by the percentage of the county actually located in the watershed. These data were then compared and/or combined with the previously modeled agricultural damage data. Fourth, county-level flood damage expenditures by FEMA, insurance agents, agencies, grants, loans, charity, etc., were divided by the percentage of the population of the county in a watershed to determine watershed-specific damage. Finally, flood damage data were estimated for years when they were not available based on data from other years and annual flood volume data (peak flows and flood stages and duration and antecedent moisture conditions). Additional details concerning the estimation of historical flood damage in the watersheds can be found in a recent Master’s thesis in the Department of Agricultural Economics at North Dakota State University (Kjelland, 1999). Annual flood damage (both total and agricultural only damage) are summarized in Tables 1 and 2. Agricultural flood damage makes up 55% and 39% of total flood damage in the Maple and Wild Rice Watersheds, respectively (Figures 6 and 7). This damage is usually associated with summertime floods that last several days or more. In contrast, residential flood damage is most often associated with very large single event floods during the early spring. As expected, residential flood damage (FEMA public assistance payments) tend to be concentrated in the lower reaches of a watershed where higher concentrations of towns and people are located (Figure 8). 11 During the ten-year period from 1989 to 1998, flood damage in the Maple Watershed totaled approximately $29.3 million, while in the WRW total damage was $102 million (both in 1998 dollars). It should be noted that this 10-year period is part of a pronounced ‘wet cycle’ in the region. Table 1. Flood Damage and Characteristics, Maple Watershed (1989-1998)* 1989 1992 1993 1994 1995 1996 1997 1998 Total Damage $0.3 $0.1 $12.4 $0.5 $0.9 $1.8 $12.5 $0.8 ($ millions) Agricultural Damage 64% 100% 62% 87% 52% 68% 10% 2% (% of total) Peak CFS & 981 710 3,770 3,040 1,830 3,860 4,060 1,730 Stage (ft) 8.4 ft 7.3 ft 12.7 ft 11.5 ft 10.1 ft 12.3 ft 12.6 ft 9.9 If Wet Prior to Flood No No Yes No Yes Yes Yes Yes Notes: - At Enderlin, ND, the flood stage is at 9 feet. - In 1990 and 1991, no serious flooding or flood damage occurred in the watershed. Table 2. Flood Damage and Characteristics, Wild Rice Watershed (1989-1998)* 1989 1990 1993 1994 1995 1996 1997 1998 Total Damage $2.6 $0.6 $15 $2.1 $0.9 $3.3 $68.8 $8.7 ($ millions) Agricultural Damage 60% 0% 72% 21% 0% 61% 12% 82% (% of total) Peak CFS & 5,480 1,100 3,680 2,600 3,200 5,750 10,600 6,550 Stage (ft) 29.6 ft 14.3 ft 27.2 ft 21.2 ft 25.3 ft 28.7 ft 33.9 ft 26.9 ft If Wet Prior to Flood No No Yes No No Yes Yes No Notes: - At Hendrum, MN, the flood stage is 17 feet. - In 1991 and 1992, no serious flooding or flood damage occurred in the watershed. 12 PUBLIC UTILITY SYSTEMS 6% PUBLIC FACILITIES 0% MISCELLANEOUS 2% COMMERCIAL/IND. 7% AGRICULTURAL 39% WATER CONTROL FACILITIES 10% 13 TRANSPORTATION 19% PROTECTIVE MEASURES 5% RESIDENTIAL 13% DEBRIS CLEARANCE 0% Figure 6. Historical Flood Damage in the Maple Watershed. PUBLIC FACILITIES 1% MISCELLANEOUS 6% AGRICULTURAL 22% PUBLIC UTILITY SYSTEMS 4% 14 RESIDENTIAL 17% COMMERCIAL/IND. 41% DEBRIS CLEARANCE 1% WATER CONTROL FACILITIES 5% PROTECTIVE MEASURES 4% TRANSPORTATION 3% Figure 7. Historical Flood Damage in the Wild Rice Watershed. 15 Figure 8. Residential Flood Damage by Township in the Wild Rice Watershed. Flood Damage Downstream on the Main-Stem of the Red River Damage outside the sub-watersheds (downstream on the main-stem of the Red River) are assumed here to be associated only with Grand Forks, ND, and East Grand Forks, MN, (hereafter referred to simply as 'Grand Forks'). Grand Forks is about 70 river miles (or 40 road miles) downstream of the Maple and Wild Rice Watersheds. The USACE (1978) has estimated that approximately 15% of river flow volume in the Red River at Grand Forks is associated with drainage from the Maple Watershed (7%) and the WRW (8%). Communities further downstream such as Pembina, ND, and Winnipeg, Manitoba, are considered influenced by many other sub-watersheds and a variety of other hydrological systems within the RRV. During the flood of 1997 (a 100-plus-year flood), Grand Forks received about $98 million in Federal assistance for residential and infrastructure flood damage (FEMA estimates reported in Carlson, 1998). To approximate flood damage associated with each of the two watersheds, this figure is multiplied by the percentage of Red River flow volume at Grand Forks associated with each of the watersheds. Therefore, Grand Forks flood damage over a typical 10year period associated with water flows from the Maple Watershed are $6,860,000 and for the WRW the figure is $7,840,000. 2.4. The Hydrologic Modeling of Wetland Storage and Flooding Many of the assumptions in this present study to evaluate the feasibility of wetland restoration and impoundment storage are based on previous studies of the relationships between wetlands and flooding. Most of these studies were focussed in other areas of the country, in the upper Midwest, and other parts the RRV. However, one study (a companion study to this report) was conducted in the Maple and Wild Rice Watersheds. Previous Studies on Wetland Storage and Flooding Perhaps the most extensive written evaluation of flooding issues in the RRV is the report by Miller and Frink (1984) for the United States Geological Society (USGS). This report notes that due to its large size, geology, and climate, major flood events in the Red River are most often associated with climatic conditions and, more specifically, with early spring snowmelt especially after a wet fall. This is confirmed from the historical flood damage estimates within the Maple and Wild Rice Watersheds (Tables 2 and 3). Miller and Frink and others (McCombs-Knutson, 1984) note that because the Red River basin is as wide as it is long and because the river flows northward, the timing of the flows on the tributaries closely coincide with that on the main-stem adding to flood peaks. 16 Table 3. Effects of Wetland Storage in Reducing Peak Flows During Flood Events in the Maple River at Enderlin, ND High Frequency Flood Event (2-10 Year) Medium Frequency Flood Event (25 Year) Low Frequency Flood Event (50 Year) Very Low Frequency Flood Event (100+ Year) 1) Peak CFS & Flood Stage Baseline Situation 1,535 9.14 2,893 11.25 3,824 12.35 4,977 13.44 2) Peak CFS & Flood Stage 2,700 AF Wetland Storage (1 AF Bounce) Reduction in Flood Stage 1,357 8.79 2,706 11.00 3,630 12.14 4,737 13.23 3.8% 2.2% 1.7% 1.6% 3) Peak CFS & Flood Stage 5,400 AF Wetland Storage (2 AF Bounce) Reduction in Flood Stage 1,292 8.65 2,627 10.89 3,549 12.05 4,622 13.12 5.4% 3.2% 2.4% 2.4% 4) Peak CFS & Flood Stage 1,171 2,518 3,404 4,578 10,800 AF Wetland Storage 8.39 10.74 11.89 13.08 (4 AF Bounce) Reduction in Flood Stage 8.5% 4.5% 3.7% 2.7% * All wetland storage options assume that a maximum of 25% of the flow out of each sub-basin is diverted into wetland storage. Finally Miller and Frink summarize several previous studies that have attempted to explain relationships between changing land use, drainage, and flooding in the RRV. In summary, these studies have found that increased drainage does have an effect on small (highfrequency) flood events but a diminished or negligible effect on large-scale, low frequency (100 and 500 year) flood events. It is also noted that studies such as Moore and Larson (1980) have concluded that statistical (regression) analyses of historical data involving changes in land use and drainage and peak flows is problematic due to data that are limited both spatially and temporally. A study by the USACE in 1976 of the Charles River Watershed near Boston, Massachusetts, found that wetland storage had a strong effect on peak flood events. However, it should be noted that the climatic and hydrological conditions of this watershed differ greatly from those of the RRV. Other studies and commentary have noted that historical levels of drainage could have made significant reductions in total flood levels in the Mississippi (Hey and Philippi, 1995) and RRV Watersheds (Sierra Club, 1997). However, these analyses make the unrealistic assumption 17 that wetlands are empty at the time of flood events or that all potential wetland storage contributes to reducing peak flood flows. A more rigorous analysis of Mississippi River Watersheds found that restoring upland wetlands could help reduce flood peaks by 1% to 23% with deep wetlands and from 5% to 9% with shallow wetlands. They also noted that wetland restoration was most effective in reducing flood damage during typical (25 year or less) storm events (USACE, 1994). The hydrologic effects of wetland drainage were recently modeled with a hydrologic simulation program model and detailed watershed level data in the Little Cobb River subwatershed of the LeSuer River Watershed in Minnesota (Miller, 1999). The study concluded that wetlands drainage accounted for up to a 57% increase in annual peak discharge. However, the study also concluded that potential decreases in flooding due to the addition of wetlands is dependent on the quantity of available storage in wetlands prior to a flood event, and that wetland storage had very little effect on reducing the magnitude of large (low frequency) flood events. Finally, a nonauthored and difficult to find hydrological analysis using empirical data for three restored wetlands in the RRV concluded that: “The restoration of small drained wetlands (1.8, 2.5, and 10.8 surface acres), when done using a simple ditch plug and spillway, does not appear to provide any flood control benefits. In fact, based on our sample, it appears more likely that downstream flood control benefits will be reduced when compared to those, which were provided by the drained wetland basin. The exception may be during those flood events following a dry period when starting pond elevations are low” (Red River Management Board, 1993). Storage Volume of Wetlands A critical factor in the role of wetlands in reducing peak flood flows is their storage potential, or ‘bounce’ especially during major flood events. Bounce is a function of wetland volume and depth as well as antecedent soil moisture conditions. Hubbard (1982), in a survey of 213 small wetlands in northeastern South Dakota immediately after the vernal thaw, determined that the wetlands on average were .67 acres in size with an average maximum depth of 1.4 feet and an average volume of 0.73 acre feet (AF) of water. It was also noted that much of the storage probably came from snowmelt and that excess storage capacity existed meaning that the maximum volume of water that these wetlands can store is a little less than an AF of water per surface acre of wetland. A study in the Devils Lake area of North Dakota (Ludden, Frink, and Johnson, (1983), found that wetlands on average store 0.6 AF per surface acre (AFPSA) during a 2-year flood event, and 0.45 AFPSA during a 100-year flood event. From this data, one can deduce that 18 bounce, or available wetland storage during a major (100-year) flood event is about half an AFPSA. A study of three restored wetlands with simple plug ditches and earthen spill-ways in the RRV found that the difference in storage between a 2-year and 100-year flood event ranged from 0.5 to 1.5 AFPSA of wetland (Red River Management Board, 1993). As reported in Eppich, Apfelbaum, and Lewis (1998), the Minnesota DNR reports at least 1.5 AFPSA in Red River basin wetlands during wet periods (Terry and Adland, 1998), while the DNR Division of Fish and Wildlife estimates wetland storage volume at 1 AFPSA. However, the details regarding these studies is not provided, and it is not known whether their term of storage volume refers to actual storage or available storage (bounce). Finally, Epich, Apfelbaum, and Lewis (1998), a group of wetland restoration experts recommend that wetlands constructed with drainage outlets not withdrawn (discharged) more than 2 feet in order to maintain reasonably diverse and dense wetland vegetation. Modeling Wetlands and Flooding in the Maple Watershed Preliminary results of a hydrological modeling exercise quantifying relationships between drained wetland storage and mainstream flood levels in the Maple Watershed (Bengtson and Padmanabahn, 1999) is summarized below. The Maple Watershed was subdivided into 48 sub-basins, and modeled using HEC-1, a quasi-distributed lumped parameter hydrologic model that accounts for wetland storage, land use, hydrologic conditions, and rainfall events. Wetland storage was modeled as diversions, which permanently retained some of the outflow from each sub-basin. As water arrived at the outlet of a sub-basin, 75% of the flow was routed downstream, while 25% was diverted until the wetland storage was filled. After that, 100% of the flow would enter the stream. The 25% diversion assumes that 25% of the overland flow in the watershed will be intercepted by wetlands. Since the drained wetlands only comprise about 0.3% of the surface area of the watershed, this is a generous assumption. Using the 25% diversion assured that all wetland storage would be utilized. The effect of changing the timing of the diversions was modeled by restricting their outflow from each sub-basin after reaching 50% of the peak. This had the effect of reducing the volume of runoff at a time nearer the peak flow and, thus, more effectively reducing the peak flow. However, they considered it unlikely that all of the overland flow will somehow not be intercepted by wetlands until the outflow reaches 50% of peak. The assumption was that all NWI drained wetlands would be restored and their estimated volume (on average 1 AFPSA of wetlands) is entirely available for flood storage. From this, it was concluded that increased wetland storage gained by restoring drained wetlands in the Maple Watershed did not materially affect peak flows at the watershed outlet for low-frequency flood events, even when the storage available was increased four-fold. Changing the timing of the 19 diversions so that they would occur nearer the hydrograph peak only slightly decreased the peak flow. Table 3 summarizes reduction in peak flows and stages for 3 precipitation flood events under a baseline condition (no wetland storage) and wetland storage of 2,700 AF (1 foot bounce), 5,400 AF (2 foot bounce), and 10,800 AF (4 foot bounce). Reductions in peak flood stages increased (albeit at a nonproportional and diminishing rate) as wetland bounce increased and decreased as flood events became larger. The study concluded that it is unlikely that storage in randomly located, uncontrolled storage basins would “slice off the peak” of a flood hydrograph. For this to occur, each wetland basin would need to remain empty until just the right time so that retained runoff does not contribute to maximum peak flow. It is more likely that wetlands will begin to fill as snowmelt and/or rainfall runoff begins to move overland and is intercepted by these wetlands. Thus, the runoff potentially stored in available wetland basins would reduce the volume of runoff under the rising limb of a flood hydrograph, which results in a slightly lower peak flow for low-frequency flood events. 20 Chapter 3. The Feasibility of Wetland Restoration-Storage Programs The economic feasibility of alternative wetland restoration options to reduce flood damage in the Maple and Wild Rice Watersheds are evaluated in three ways. First, the costs of wetland restoration will be compared to benefits (reduced flood damage) based on historical flood damage in the watersheds and the results of hydrologic modeling exercises in the Maple Watershed. Second, estimates will be made regarding minimum levels of flood stage and damage reduction required for wetland restoration options to break even which do not rely on the aforementioned hydrologic modeling results in the Maple Watershed. Finally, the per acre annual values of additional (nonflood-related) wetland functions required for wetland restoration options to break even are calculated. In all three cases, feasibility estimates will be made over a hypothetical 20-year period in the future and focus on flooding both within and outside the watersheds. The downstream focus is, however, limited to the main-stem of the Red River at Grand Forks. Costs and benefits are discounted to1999 dollars using a 5% discount rate. The evaluation focuses on three wetland restoration options. First, simple flood storage using wetland plugs with earthen spillways. Second, more complicated and expensive wetland restoration which relies on drainage outlets that permit wetland volumes to be drawn down in advance of flood events and, hence, increase available storage. And third, wetland restoration intended to restore the full environmental services of wetlands using established restoration criteria which is even more complicated and expensive as it involves making sure that soil, slope, water, and vegetative characteristics of a restored wetland are very similar to natural wetlands in a particular area. 3.1. Quantities and Locations of Previously Drained Wetlands The location and size of previously drained wetlands in the upper sections of the watersheds were identified though a GIS-based analysis of the NWI. It is possible that additional wetlands, especially those drained many years ago, are not captured by the NWI. However, these missing wetlands should not affect the feasibility estimates in this study because wetland restoration costs and benefits are assumed to increase proportionately with wetland quantities. Also, the restoration of many drained wetlands missing from the NWI is probably not economically and/or politically feasible due to historical land use developments. It should also be noted that for the option where wetland restoration is hypothesized to take place on low cost lands in the upper sections of the watersheds, it is assumed that there are actually more drained wetlands than in the NWI. Previously drained wetlands and hydrology in the Maple and Wild Rice Watersheds are illustrated in Figures 9 and 10 and summarized in Table 4. In the upper Maple Watershed there are approximately 2,700 acres of drained wetlands. Their average size is 1.3 acres while the most frequent (common) size is 0.5 acres, and wetland acreage is evenly split among small- (< 1 acre), medium- (1-5 acres), and large- (> 5 acres) size classes. 21 In the upper WRW there are approximately 12,600 acres of drained wetlands whose mean size is 2.6 acres and the most frequent size is 2.5 acres. The total costs of restoring wetlands include construction costs in the first year and annual land rental payments throughout the life of the project, which in this present evaluation is 10 years. Operating and/or maintenance costs are uncertain and assumed to be either insignificant and/or the responsibility of landowners renting their land. Table 4. Wetland Acres by Size Classes in the Maple and Wild Rice Watersheds* Upper Maple Upper Wild Rice Total Acres 2,700 12,600 Small Wetland Acres (< 1 Acre) 700 acres 1,200acres Medium Wetland Acres (1-5 Acres) 1,100 3,000 Large Wetland Acres (> 5 Acres) 900 8,400 3.2. The Costs of Wetland Restoration-Based Storage Programs Construction Costs The construction costs of restoring wetlands vary considerably depending on what type of restoration is performed (plugging drains, establishing outlet flows, or restoring wetlands with their full ecological services). These costs are also subject to economies of scale or, more specifically, that per acre restoration costs decrease with the size of wetlands being restored. Baseline construction costs used here are based on actual wetland restoration costs in northwest Minnesota as well as the results of a large-scale wetland restoration/impoundment in the Alice Lake area of North Dakota. As reported in Sip (1998), the 1995 Minnesota Wetland Replacement/Mitigation Cost Summary published by the Minnesota Board of Water and Soil Resources, estimates wetland restoration costs to range from $95 to $ 30,000 per acre with an average of $3,000. In contrast, the cost of impounding new wetlands ranges from $200 to $20,000 per acre with an average of $1,500 per acre. Eppich, Apfelbaum, and Lewis (1998) estimate that establishing a depressional wetland control without land acquisition costs or an outlet control is approximately $1,500 per acre. The cost of including an outlet control is $2,750 again excluding land acquisition costs. They also estimate that the cost of constructing impounded wetlands (building dikes or berms) with outlet controls is approximately $10,000 per acre excluding land acquisition costs. The authors also note that many of these per acre wetland construction costs (especially permitting, design, and engineering costs) can be reduced when large and/or many wetlands are being restored/constructed. 22 Figure 9. Drained Wetlands and Hydrology, Maple Watershed. 23 24 Figure 10. Drained Wetlands and Hydology, Wild Rice Watershed. An example of economies of scale in restoring wetlands is the Alice Lake Wetland in North Dakota, where according to Renner (1999), dikes and outlet flow structures were constructed in order to restore 3,500 wetland acres at a cost of approximately $930,000, excluding land acquisition costs ($265 per acre). Finally, Hammer et al., (1993) estimate that farmers should be able to construct an effective wetland, including planting of limited vegetation, for a little less than $3,000 per acre. For this present study, the construction costs of three types of wetland restoration will be estimated: simple plugs and spillways, outlet control devices, and full ecological restoration. Construction cost will be adjusted to account for economies of scale (small, medium, and large wetlands) for each restoration option. Simple wetland restoration/construction costs for plugging a wetland drain with an earthen spillway are estimated to be $300 per acre for small-size wetlands (< 1 acre), $200 per acre for medium-size wetlands (1-5 acres), and $100 per acre for large-size wetlands (> 5 acres). Using wetland size classes for the two watersheds summarized in Table 3, total simple wetland restoration/construction costs for the upper Maple Watershed are $520,000 and for the WRW the cost is $1.8 million. Restoring wetlands with outlet control is estimated to be twice as expensive ($600 per acre for small-size wetlands, $400 per acre for medium-size wetlands, and $200 per acre for large-size wetlands) resulting in construction costs in the Maple Watershed being $1,040,000 and $3.6 million in the WRW. Wetland restoration/construction costs providing full ecological services are estimated at $3,000 per acre for small-size wetlands, $1,000 per acre for medium-size wetlands, and $500 per acre for large-size wetlands. This results in total construction costs of $3.6 million in the Maple Watershed and $10.8 million in the WRW. Land Rental Values Payments to farmers for the storage of water on their lands are considered to be the same regardless of the wetland restoration option being evaluated. Annual per acre rental payments are used because they require less up-front costs, it is not known how long the storage programs will be required, and because reasonably accurate annual rental data exist for the watersheds being studied. Finally, the land rental values reported here are considered low because due to topological and ownership conditions it is likely that wetland restoration would require more land than is actually restored. Alternatively, it is not likely that all procured land could be used for wetland restoration Land rental values in the upper Maple Watershed were obtained from county level cropland rental values from 1993-1997 as reported by the North Dakota Agricultural Statistics Service. A weighted average of cropland and pastureland rental rates was estimated. Also, to account for lower quality land in the upper section of the watershed, land rental values were adjusted on the basis of rental values of adjacent counties. Total rental value costs were 25 estimated by multiplying average county rental rates in the upper watershed ($40/acre) by the number of drained wetlands in each county. In the upper WRW, annual land values were estimated by making adjustments to assess tillable land values reported by the Minnesota Department of Revenue. It was noted that in the Maple Watershed annual rental values were 7% of total values, and this same percentage was, therefore, used to convert capitalized land values to annual rental values in the WRW. The map in Figure 11 shows the variation in township level rental values across the township in the WRW and illustrates that variations in rental values are primarily a result of geophysical characteristics. Specifically, rental values are highest in the lower watershed (the lake plain) followed by middle (Beach Ridge) and upper (Glacial Moraine) parts of the watershed. Total land rental costs in the upper WRW were estimated by multiplying township rental values by the number of drained wetland acres in each township and on average are $40 per acre per year. Rental costs on “cheap” lands in the upper sections of each of the two watersheds involved the assumption that wetland restoration would occur only in townships or areas with rental values below the average for the watershed. On average, these lands cost $35 per acre a year to rent in the upper Maple Watershed and $26.50 per acre in the upper WRW. This requires the assumption that there are more drained wetlands in the watershed than indicated by the NWI. The total costs of restoring drained wetlands under different scenarios in the upper sections of each of the two watersheds over 10 years are a summation of the present value of construction costs in year 1 and land rental payments in years 1 through 20 (Table 5). Wetland restoration costs over a 20-year period range from $1.8 million to $4.7 million in the upper Maple Watershed and between $5.9 million and $12.8 million in the upper WRW. As expected, the total costs of simple wetland restoration (plugging up a wetland) is cheaper than restoring wetlands with outlets which is cheaper than restoration with full ecological services. A more useful indicator of the relative cost of alternative restoration scenarios is the annual storage cost per AF of water stored. Based on the literature previously described, wetlands restored with a simple plug are expected to provide at most 1 AF of water storage per surface acre of wetland during a major springtime flood event. In contrast, a wetland restored with an outlet control device is expected to be able to hold up to 2 AFPSA (actually more storage is possible but this would require seasonal drainage of wetlands that could possibly damage wetland ecosystems). 26 27 Table 5. Wetland Restoration Costs in the Maple and Wild Rice Watersheds Over a 20-year Period Total Annual Rent 2 Construction Costs (Year 1) Total Costs (20 Years) 3 Annual Cost per Acre Foot Stored Simple Restoration Simple Restoration (Cheap Land) $108,000 $94,500 $520,000 “ $1,841,157 $1,673,393 $34 $31 Restoration with Outlets Restoration “ “ (Cheap Land) $108,000 $94,500 $1 Million “ $2,298,300 $2,130,536 $43 $39 Full Restoration Full Restoration (Cheap Land) $108,000 $94,500 $3.6 Million “ $4,774,490 $4,606,250 $88 $85 Simple Restoration Simple Restoration (Cheap Land) $430,000 $334,000 $1.8 Million “ $7,073,036 $5,876,664 $28 $23 Restoration with Outlets Restoration “ “ (Cheap Land) $430,000 $334,000 $3.6 Million “ $8,787,322 $7,590,950 $35 $30 Upper Maple Watershed (2,700Acres)1 28 Upper Wild Rice Watershed (12,600 Acres)1 Full Restoration $430,000 $9.1 Million $14,025,417 Full Restoration (Cheap Land) $334,000 “ $12,829,045 NOTES: 1. Wetland acreage based on the National Wetland Inventory. 2. Average annual rent in the upper Maple Watershed is $40 and $35 (cheap land) while in the Wild Rice Watershed, land rents are $32 and $26 (cheap land). 3. Present values calculated using a 5% discount rate. $56 $51 Finally, full wetland restoration efforts are also assumed to include outlet controls and are, therefore, also expected to store 2 AFPSA of wetland. It can, therefore, be deduced that wetland restoration in the upper Maple Watershed costs between $31 and $85 per AF of stored water. The cheapest restoration-storage option appears to be the use of outlets. The same is true for the upper WRW where restoration-storage costs range from $23 to $56 per AF of water stored. Wetland restoration-storage costs are cheaper in the upper WRW than in the upper Maple Watershed because land values are lower and there are larger wetlands with cheaper per acre restoration-construction costs. 3.3. The Potential Benefits of Wetland Restoration-Storage Options To estimate the economic feasibility of alternative wetland restoration options, it is necessary to calculate the potential benefits of storage as represented by reduced flood damage both within and downstream of the Maple and Wild Rice Watersheds. This requires determining how historical flood damage in the watersheds would have been reduced by alternative wetland restoration options on the basis of hydrologic modeling exercises in the Maple Watershed and then extrapolating this reduced damage into the future. From 1989 to 1998, flood damage within the Maple Watershed was $29.3 million and $102 million within the WRW. Flood damage outside the watersheds in Grand Forks that are associated with water flows from the Maple Watershed is $6.9 million and $7.8 million for the WRW (Chapter 2, Section 2.3). The estimation of stage-damage curves that quantify the relationship between flood levels and damage was not estimated for this study due to a lack of detailed elevation and sitespecific flood damage data. It is, therefore, assumed that annual reductions in peak flood stage estimated for the Maple Watershed by Bengtson and Padmanabahn (1999) correspond directly with reductions in flood damage. This assumption is probably more valid for agricultural-based flood damage than residential flood damage, but without detailed stage-damage curves, no other alternatives exist. This assumption will likely increase estimates of the benefits of wetlands for reducing flooding because flood damage usually increases at an increasing rate with higher flood stages. Finally, without any such hydrologic modeling completed for the WRW, it is necessary to assume that the percentage relationships between wetland storage and flood stage levels are the same for each watershed. Potential reductions in flood damage in the two watersheds are based on estimated reductions in peak flood stages of historical (1989-1998) flood events (high-, medium-, low-, and very low-frequency floods), and alternative wetland storage (different levels of wetland acres restored and alternative levels of storage bounce). Recall from Chapter 2 and Table 3 (hydrological modeling in the Maple Watershed), that wetland storage with 1 foot of bounce reduces peak flood stage by 3.8% (high-frequency floods), 2.2% (medium-frequency floods), 1.7% (low-frequency floods) and 1.6% (very low-frequency floods). The corresponding 29 reductions in peak flood stage for 2 feet of storage bounce are 5.4%, 3.2%, 2,4%, and 2.4% (Bengtson and Padmanabahn, 1999). This reduced flood damage (1989-1998) is extrapolated to a hypothetical 20-year period in the future by assuming that they will repeat themselves twice, and that the annual average of this reduced flood damage occurs each year in the future. This future reduced flood damage is then discounted using a 5% discount rate. Therefore, reduced 20-year flood damage in the Maple Watershed with 1 foot of bounce and 2,700 AF of storage is $800,800 and $1,170,000 with 2 feet of bounce (5,400 AF of storage). The corresponding values for the WRW (albeit based on hydrologic modeling results from the Maple Watershed) are $2.4 million with 1 foot of bounce and 12,600 AF storage, and $3.6 million for 2 feet of bounce and 25,200 AF storage (Tables 6 and 7). These values are considered project benefits over a 20-year period. 3.4. Net Present Values and Benefit-Cost Ratios of Wetland Restoration Options All of the wetland storage options evaluated have costs that greatly exceed their benefits. Net present values are all negative and benefit-cost ratios range from 1:1.4 to 1:4.1 in the Maple Watershed and 1:1.6 to 1:3.9 in the WRW (Table 8). Even with the optimistic scenario with restoration on the cheapest lands and 2 feet of bounce, economic feasibility defined as benefits greater than costs, is not achieved. Alternatively, wetland restoration storage options would have to reduce peak annual flood stage and total historical flood damage by between 3.4% and 9.6% in the Maple Watershed and by 4.3% to 10.2% in the WRW before breaking even (Table 9). These estimates of required reductions in peak flood stage and flood damage required for economic feasibility may be useful to future and/or alternative hydrologic modeling studies of the effects of wetland storage on flooding. 3.5. Wetland Restoration Feasibility Considering Additional Benefits of Wetlands Wetland restoration for the purpose of flood control in two typical sub-watersheds of the RRV has been demonstrated as nonfeasible from an economic standpoint. However, some people would argue that these restored wetlands would likely provide additional (nonfloodrelated) benefits such as services to animal and plant species, improving water quality, and recreational and aesthetic services. To evaluate if the inclusion of nonflood-related wetland benefits would make wetland restoration-storage activities economically feasible, the minimum required values of these ‘additional’ wetland benefits in order for restoration-storage alternatives to break even were estimated. These minimum required wetland values were then compared to available estimates of wetland benefits found in the literature. 30 Table 6. Historical and 20-Year Expected Reductions in Flood Damage Due to Wetland Storage in the Maple Watershed NPV (20 Years) 31 1989 1992 1993 1994 1995 1996 1997* 1998 Total Damage ($ millions) 0.3 0.1 12.4 0.5 0.9 1.8 19.4 0.8 Peak CFS 981 710 3,770 3,040 1,830 3,860 4,060 1,730 8.4 7.3 12.7 11.5 10.1 12.3 9.9 High High Low Medium High Low 12.6 Very Low High $11,400 $3,800 $210,800 $11,000 $34,200 $30,600 $310,400 $30,400 $800,800 $16,200 $5,400 $297,600 $16,000 $48,600 $43,200 $465,600 $43,200 $1,170,000 Peak Flood Stage (ft) Type of Flood Event (Frequency) Reduced Flood Damage with Simple Restoration (1 ft bounce, 2,700 AF storage) Reduced Flood Damage with Restoration using Outlets (2 ft bounce, 5,400 AF storage) Notes: - Flood stage at Enderlin, ND, is 9 feet. - No major flood damage recorded in 1990 or 1991. - Hydrological relationships from a study in the Maple Watershed (Bengtson and Padmanabahn, 1999). - From Table 3, storage with 1 foot of bounce reduces peak flood stage by 3.8% (high-frequency floods), 2.2% (medium-frequency floods), 1.7% (low- frequency floods), and 1.6 (very low-frequency floods). The corresponding reductions for storage with 2 feet of bounce are 5.4%, 3.2%, 2.4%, and 2.4%. - Historical flood damage based on estimates by Kjelland (1999) and expanded 20 years into the future using a 5% discount rate. - Assumption of direct (linear) relationship reductions in flood stage (ft) and flood damage ($). - *1997 Flood damage includes Grand Forks damage associated with flows from the watershed. Table 7. Historical and 20-Year Expected Reductions in Flood Damage Due to Wetland Storage in the Wild Rice Watershed (Based on hydrological modeling in the Maple Watershed) NPV (20 years) 32 1989 1990 1993 1994 1995 1996 1997* 1998 Total Damage ($ millions) $2.6 $0.6 $15 $2.1 $0.9 $3.3 $76.6 $8.7 Peak CFS & Stage (ft) 5,480 1,100 3,680 2,600 3,200 5,750 10,600 6,550 29.6 14.3 27.2 21.2 25.3 28.7 33.9 26.9 Low High Medium High High Low Very Low Low $44,200 $22,800 $330,000 $79,800 $34,200 $56,100 $1,225,000 $147,900 $2.4 (million) $62,400 $32,400 $480,000 $113,400 $48,600 79,200 $1,838.000 $208,800 $3.6 (million) Peak Flood Stage (ft) Type of Flood Event (Frequency) Reduced Flood Damage with Simple Restoration (1 ft bounce, 2,700 AF storage) Reduced Flood Damage with Restoration using Outlets (2 ft bounce, 5,400 AF storage) Notes: - Flood stage at Hendrum, MN, is 17 feet. - No flooding in 1991 and 1992. - Hydrological relationships from a study in the Maple Watershed (Bengtson and Padmanabahn, 1999). - From Table 3, storage with 1 foot of bounce reduces peak flood stage by 3.8% (high-frequency floods), 2.2% (medium-frequency floods), 1.7% (low-frequency floods), and 1.6 (very low-frequency floods). The corresponding reductions for storage with 2 feet of bounce are 5.4%, 3.2%, 2.4%, and 2.4%. - Historical flood damage based on estimates by Kjelland (1999) and expanded 20 years into the future using a 5% discount rate. - Assumption of direct (linear) relationship reductions in flood stage (ft) and flood damage ($). - *1997 Flood damage includes Grand Forks damage associated with flows from the watershed. Table 8. Benefit-Cost Ratios of Alternative Wetland Restoration-Storage Options in the Maple and Wild Rice Watersheds Total Costs (20 Years) Reduced Flood Damage (20 years) BenefitCost Ratio $1.8 million $800,800 1:2.3 $1.6 million $800,800 1:2.1 $2.3 million $1,170,000 1:2 $2.1 million $1,170,000 1:1.8 $4.7 million $1,170,000 1:4.1 $4.6 million $1,170,000 1:3.9 $1.7 million $1,170,000 1:1.4 $7 million $2.4 million 1:2.9 $5.9 million $2.4 million 1:2.4 $8.8 million $3.6 million 1:2.5 $7.6 million $3.6 million 1:2.1 $14 million $3.6 million 1:3.9 $12.9 million $3.6 million 1:3.6 $5.9 million $3.6 million 1:1.6 Upper Maple Watershed (2,700 Acres of Upper Wetlands Restored)1 Simple Restoration (1 ft bounce, 2,700 AF storage) Simple Restoration on Cheapest Land (1 ft bounce, 2,700 AF storage) Restoration with Outlets (2 ft bounce, 5,400 AF Storage) Restoration “ “ on Cheapest Land (2 ft bounce, 5,400 AF Storage) Full Restoration (2 ft bounce, 5,400 AF Storage) Full Restoration on Cheapest Land (2 ft bounce, 5,400 AF Storage) Most Optimistic Scenario: (Simple Restoration on Cheapest Land, 2 feet of bounce, and 5,400 AF Storage) Upper Wild Rice Watershed (12,600 Acres of Upper Wetlands Restored ) Simple Restoration (1 ft bounce, 12,600 AF storage) Simple Restoration on Cheapest Land (1 ft bounce, 12,600 AF storage) Restoration with Outlets (2 ft bounce, 12,600 AF Storage) Restoration “ “ on Cheapest Land (2 ft bounce, 12,600 AF Storage) Full Restoration (2 ft bounce, 12,600 AF Storage) Full Restoration on Cheapest Land (2 ft bounce, 12,600 AF Storage) Most Optimistic Scenario: (Simple Restoration on Cheapest Land, 2 feet of bounce, and 12,600 AF Storage) Notes: 1. Wild Rice Watershed benefit estimates are based in part on hydrologic modeling results conducted in the Maple Watershed. 2. Nonflood-related benefits of wetlands are not considered here. 33 Table 9. Required Reductions in Peak Flood Stage and Flood Damage for Alternative Wetland Restoration-Storage Options to Break Even Total Flood Damage (20 years) Required Reduction in Flood Damage (Peak Flood Stage)1 to ‘Break Even’ $1.8 million $1.6 million $2.3 million $2.1 million $4.7 million $4.6 million $1.6 million $36.5 Million “ 3.7% “ 4.6% “ 4.3% “ 9.6% “ 9.2% $7 million $5.9 million $8.8 million $7.6 million $14 million $12.9 million $5.9 million $121.6 Million “ 5.2% “ 6.4% “ 5.5% “ 10.2% “ 9.4% Total Costs (20 years) Upper Maple Watershed (2,700 Acres of Upper Wetlands Restored )1 Simple Restoration (1 ft bounce, 2,700 AF storage) Simple Restoration on Cheapest Land (1 ft bounce, 2,700 AF storage) Restoration with Outlets (2 ft bounce, 5,400 AF Storage) Restoration “ “ on Cheapest Land (2 ft bounce, 5,400 AF Storage) Full Restoration (2 ft bounce, 5,400 AF Storage) Full Restoration on Cheapest Land (2 ft bounce, 5,400 AF Storage) Most Optimistic Scenario: (Simple Restoration on Cheapest Land, 2 ft of bounce, and 5,400 AF Storage) Upper Wild Rice Watershed (12,600 Acres of Upper Wetlands Restored ) Simple Restoration (1 ft bounce, 2,700 AF storage) Simple Restoration on Cheapest Land (1 ft bounce, 2,700 AF storage) Restoration with Outlets (2 ft bounce, 5,400 AF Storage) Restoration “ “ on Cheapest Land (2 ft bounce, 5,400 AF Storage) Full Restoration (2 ft bounce, 5,400 AF Storage) Full Restoration on Cheapest Land (2 ft bounce, 5,400 AF Storage) Most Optimistic Scenario: (Simple Restoration on Cheapest Land, 2 ft of bounce, and 12,600 AF Storage) Notes: 3.4% 3.4% 4.3% 4.3% 1. Reductions in peak flood stage are assumed to be directly related to reduced flood damage. 2. Nonflood-related benefits of wetlands are not considered here. 34 The value of nonflood-related wetland benefits (on an annual per acre level) required for alternative wetland restoration-storage options to break even, range from $9 to $67 in the Maple Watershed and from $21 to $97 in the WRW (Table 10). The higher end values ($67 and $97) are associated with the ‘full’ wetland restoration options, which are the most likely restoration alternatives that will provide a full range of wetland functions. In the upper Great Plains a limited amount of wetland valuation research has been conducted (Hubbard, 1989; Leitch and Hovde, 1996; and Roberts and Leitch, 1997). The results of these studies indicate that combined user, owner, and societal values of selected wetlands in the region (including flood control values) range from $50 to $326 per acre per year with an average value of $140 (Leitch and Hovde, 1996). From this present study, gross annual floodrelated benefits of wetlands (gross flood reduction benefits divided by 20 years and the acres of restored wetlands) range from $15 to $22 per acre in the Maple Watershed and $11 to $24 per acre in the WRW. This would seem to imply that if the nonflood-related benefits of wetlands are considered, some of the cheaper wetland restoration options to reduce flood damage could be economically feasible. However, this is a premature conclusion for three reasons. First, existing estimates of the nonflood-related benefits of wetlands have focussed on only a few individual wetland locations, and there is no guarantee that all of the wetlands in the Maple or WRWs have similar characteristics. Second, if large numbers of wetlands were restored in these watersheds, then the per acre value of wetland benefits would likely decline significantly due to the law of diminishing marginal returns (there is a finite number of wetland acres that birds or recreationists can utilize). Finally, many of the nonflood-related wetland benefits are comprised of consumer surplus values (not necessarily associated with actual money) making the provision of this public good difficult to finance. It may be that only full wetland restoration will provide substantial wetland function values, which is still infeasible due to their high costs. In summary, it is unlikely that the inclusion of nonflood-related wetland benefits would make wetland restoration-storage projects in the RRV economically feasible. However, more research is needed regarding the valuation of wetland benefits. Of particular interest is how per acre wetland values change when large numbers of wetlands are restored, and whether it is valid to transfer wetland benefit estimates across time and space. Additional research is also needed to compare nonflood-related wetland function values among different types of restored wetlands (i.e., simple restoration, wetlands with outlets, and full restoration). 35 Table 10. Required Values of Other (Nonflood-Related) Wetland Benefits in Order for Wetland Restoration with Full Environmental Services to be Feasible Net Benefits of Wetland Required Annual Per Acre Restoration for Value of ‘Other’ Wetland Flood Control Benefits for Restoration to (20 Years) Break Even Upper Maple Watershed (2,700 Acres of Upper Wetlands Restored )1 Simple Restoration (1 ft bounce, 2,700 AF storage) Simple Restoration on Cheapest Land (1 ft bounce, 2,700 AF storage) Restoration with Outlets (2 ft bounce, 5,400 AF Storage) Restoration “ “ on Cheapest Land (2 ft bounce, 5,400 AF Storage) Full Restoration (2 ft bounce, 5,400 AF Storage) Full Restoration on Cheapest Land (2 ft bounce, 5,400 AF Storage) Most Optimistic Scenario: (Simple Restoration on Cheapest Land, 2 ft of bounce, and 5,400 AF Storage) - $1 million $19 - $ 873,000 $16 - $1.1 million $21 - $960,000 $18 -$3.6 million $67 - $3.4 million $64 - $1 million $9 - $5.9 million $43 - $4.6 million $32 - $6.2 million $48 - $5 million $37 - $11.4 million $97 - $ 10.2 million $86 - $3.3 million $21 Upper Wild Rice Watershed (12,600 Acres of Upper Wetlands Restored ) Simple Restoration (1 ft bounce, 12,600AF storage) Simple Restoration on Cheapest Land (1 ft bounce, 12,600 AF storage) Restoration with Outlets (2 ft bounce, 25,200 AF Storage) Restoration “ “ on Cheapest Land (2 ft bounce, 25,200 AF Storage) Full Restoration (2 ft bounce, 25,200 AF Storage) Full Restoration on Cheapest Land (2 ft bounce, 25,200 AF Storage) Most Optimistic Scenario: (Simple Restoration on Cheapest Land, 2 ft of bounce, and 25,200 AF Storage) 36 Chapter 4. The Feasibility of Impoundment-Storage Programs The use of impoundments to store water in low-lying agricultural fields with outlet control devices is closely related to and often confused with wetland restoration activities. In this chapter, impoundments for flood control will be explained and their economic feasibility in reducing flood damage will be evaluated. 4.1. An Introduction to Impoundments for Flood Control Impoundments utilizing outlet control devices can store more water than restored wetlands (usually about 4 AF of water per acre of impoundment versus the 1 to 2 AF assumed for restored wetlands) primarily because they are designed and managed solely for the purpose of flood control. Impoundments are most likely to be constructed in low-lying areas and their size is expected to be relatively large in order to take advantage of economies of scale. When used in an area of seasonal flooding such as the RRV, their water levels would likely be drawn down significantly well in advance of flood events. Impoundments for flood control purposes have been used for several years in the Devils Lake region of North Dakota and are currently being proposed for use in the WRW in Minnesota. In North Dakota, the state government through the Available Storage Acreage Program (ASAP) has, since 1996, paid upland farmers to store water on their fields in order to prevent additional flooding in the Devils Lake basin. ASAP stored approximately 8,000 AF of water in 1996, 22,000 in 1997, and 21,000 in 1998. This storage has had the effect of reducing inflows into Devils Lake by approximately 40,000-50,000 AF and kept the lake from rising about four to five inches (NDSWC, 1998, 1999). The average size of an impoundment was 87 acres, with a range in size from 2 to 858 acres. The ASAP program pays farmers up to $50 per acre annually to store water, which is higher than published rental rates in the region (and the rental rates used to evaluate the feasibility of wetland restoration in the Maple and Wild Rice Watersheds). This price premium is likely a result of the uncertainty associated with the longer term effects of water storage on farmland ecology and because farmers are responsible for much of the required impoundment construction and maintenance activities. In fact, as noted from ASAP accounting documents, most impoundment sites incurred zero or minimal ($4 to $15 per acre) structural construction costs. More information concerning the details of the ASAP program is contained in Appendix B. In Minnesota, the Wild Rice Watershed Management Board (WRWMB) is currently developing a pilot project utilizing natural water storage areas in the upper watershed to reduce local and watershed flood damage. They are currently identifying natural water detention sites that are linked to downstream flood damage and where only minor structural changes and improvements are required to impound water. 37 As the project has just recently been proposed, no specific impoundment cost data have been obtained. However, according to Houston Engineering (1999), a break-even cost estimate of $1,000/AF has been made. Also, the project intends to purchase land from farmers based on tax assessment values of agricultural land while criteria for the design and construction of the water storage sites will be based on the ‘Minnesota Wetland Restoration Guide’ (1992). More information about this proposed impoundment-storage project is also contained in the Appendix B 4.2. The Feasibility of Impoundment Storage in the Maple and Wild Rice Watersheds To evaluate the economic feasibility of using impoundments to reduce flooding in the Maple and Wild Rice sub-watersheds of the RRV, the following assumptions will need to be made. First, the land area available for impoundments is assumed to be equal to the amount of potential wetland restoration analyzed in the Maple and Wild Rice Watersheds (2,700 acres and 12,600 acres, respectively). This allows direct comparisons with the feasibility of impoundments and wetland restoration options. It also permits the utilization of existing hydrological modeling results in the Maple Watershed (Bengtson and Padmanabahn, 1999). In reality, a great deal of additional land is likely to be available for impoundments in each of the watersheds. However, our focus here is on the relative economic efficiency of impoundments in reducing flood damage (i.e., a benefit-cost ratio), and it is assumed that an increase in the number of impoundment acres would be associated with a proportional increase in costs and benefits. Second, based on the experiences of the Devils Lake and Wild Rice impoundment programs, impoundments in this analysis are assumed to have a storage capacity (bounce) of 4 feet per surface acre. And, this storage is assumed to reduce peak flood stage as in the same way as wetland storage (Table 3). Third, based on a summary of construction costs by Eppich, Apfelbaum, and Lewis (1998) and previous estimates of the costs of restoring large wetlands with outlet control devices, the cost of such impoundments are assumed to be $200 per acre plus land rental costs Finally, land rental costs are the same as used in renting cropland for wetland restoration and are based on reported rental rates by the National Agricultural Statistics Service (NASS) in the Maple Watershed and by county assessors in the WRW. Evaluations of impoundments are made using average rental values and lower cost (cheaper) land rentals. The total cost of impounding 2,700 acres in the Maple Watershed over 20 years is $1.8 million or $1.7 million using lower priced land (Table 11). With storage of 10,800 AF of water annually, the annual per acre foot cost of storing water is $9 on regular priced land and $8 on cheaper land. In the WRW, the total cost of impounding 12,600 acres over 20 years is $7.4 million and $6.5 million, if only lower priced land is utilized. Because such impoundments would store 50,400 AF of water annually, the annual cost per AF stored is $7 (regular land prices) and $6 (on cheaper land). 38 Table 11. Cost of Impoundment-Storage in the Maple and Wild Rice Watersheds Annual Rent 2 Construction Costs (Year 1) Total Costs (20 Years) 3 Annual Cost per Acre Foot Stored $108,000 $95,00 $1.3 million “ $1.8 million $1.7 million $9 $8 $403,000 $334,000 $6 million “ $7.4 million $6.5 million $7 $6 Upper Maple Watershed (2,700Acres)1 Impoundments using Average Land Prices Impoundments using Cheapest Land Prices 39 Upper Wild Rice Watershed (12,600 Acres)1 Impoundments using Average Land Prices Impoundments using Cheapest Land Prices NOTES: 1. Impoundment acreages are hypothetical and expected to be potentially much higher and each acre of impoundment is expected to store 4 AF of water. 2. Average annual rent in the upper Maple Watershed is $40 and $35 (cheap land) while in the Wild Rice Watershed, land rents are $32 and $26 (cheap land). 3. Present values calculated using a 5% discount rate. Again, this reduced flood damage is extrapolated to a hypothetical 20-year period in the future by assuming that they will repeat themselves twice and that annual average values occur each year in the future. Therefore, reduced 20-year flood damage associated with impoundment storage is $1.6 million in the Maple Watershed and $4.5 million in the WRW (Tables 12 and 13). Therefore, the benefit-cost ratio for using impoundment to reduce flood damage in the Maple Watershed is 1:1.2 for regular priced land and 1:1.1 for cheap land, while in the WRW the benefit-cost ratios are 1:1.7 and 1:1.5 (cheap land). These benefit-cost ratios of impoundments indicated that costs are approximately equal to benefits. It is likely, therefore, that impoundments for flood control have the potential to be feasible. Such feasibility is likely to occur when land can be rented or purchased at prices markedly lower than watershed averages, when construction costs are lower than $200 per acre, and when there are direct hydrological relationships between impoundments and downstream flooding. The inclusion of additional (nonflood-related) wetland benefits are not included here because it is unlikely that any will be provided by impoundments which are subject to seasonal drainage. Finally, it is interesting to note that two cases of actual impoundment-storage programs being undertaken in the RRV (the Devils Lake ASAP Program and in the proposed WRW impoundment demonstration program), closely mirror the above scenarios for impoundmentstorage feasibility (impoundments on low cost lands and which are limited in scale and are sitespecific). 40 Table 12. Historical and 20-Year Expected Reductions in Flood Damage Due to Impoundment-Storage in the Maple Watershed Total Damage ($ millions) Peak CFS & Stages (ft) 41 Type of Flood Event (Frequency) Reduced Flood Damage with 2,700 Acres of Impoundments (4 ft bounce, 10,800 AF storage) Notes: 1989 1992 1993 1994 1995 1996 1997* 1998 $0.3 981 8.4 ft $0.1 710 7.3 ft $12.4 3,770 12.7 ft $0.5 3,040 11.5 ft $0.9 1,830 10.1 ft $1.8 3,860 12.3 ft $0.8 1,730 9.9 ft High High Low Medium High Low $19.4 4,060 12.6 ft Very Low $25,500 $8,500 $458,800 $22,500 $76,500 $66,600 $523,800 $68,000 NPV (20Years) High $1.6 million - Flood stage at Enderlin, ND, is 9 feet. - No major flood damage recorded in 1990 or 1991. - Hydrological relationships from a study in the Maple Watershed (Bengtson and Padmanabahn, 1999). - From Table 3, Storage with 4 feet of bounce reduces peak flood stage by 8.5% (high-frequency floods), 4.5% (medium-frequency floods), 3.7% (low-frequency floods),and 2.7% (very low-frequency floods). - Historical flood damage based on estimates by Kjelland (1999). - Assumption of direct (linear) relationship reductions in flood stage (ft) and flood damage ($). - *1997 Flood damage includes Grand Forks damage associated with flows from the watershed. Table 13. Historical and 20-Year Expected Reductions in Flood Damage Due to Impoundment-Storage in the Wild Rice Watershed (Based on hydrological modeling in the Maple Watershed) 42 1989 1990 1993 1994 1995 1996 1997* 1998 Total Damage ($ millions) $2.6 $0.6 $15 $2.1 $0.9 $3.3 $76.6 $8.7 Peak CFS & Stages 5,480 1,100 3,680 2,600 3,200 5,750 10,600 6,550 Low High Medium High High Low Very Low Low $96,200 $51,000 $675,000 $178,500 $76,500 $122,100 $2,068,200 $321,900 Type of Flood Event (Frequency) Reduced Flood Damage with 12, 600 acres of Impoundments (1 ft bounce 50,400 AF storage) Notes: - Flood stage at Hendrum, MN, is 17 feet. - No flooding in 1991 and 1992. - Hydrological relationships from a study in the Maple Watershed (Bengtson and Padmanabahn, 1999). - From Table 3, Storage with 4 feet of bounce reduces peak flood stage by 8.5% (high-frequency floods), 4.5% (medium-frequency floods), 3.7% (low-frequency floods),and 2.7% (very low-frequency floods). - Historical flood damage based on estimates by Kjelland (1999). - Assumption of direct (linear) relationship reductions in flood stage (ft) and flood damage ($). - *1997 Flood damage includes Grand Forks damage associated with flows from the watershed. NPV (20 Years) $ 4.5 million Chapter 5. Summary and Conclusions This study evaluated the economic feasibility of restoring wetlands and constructing impoundments in order to reduce flood damage in two sub-watersheds of the RRV. The benefits and costs of the following storage options were evaluated: simple wetland restoration, wetland restoration with outlet control devices to increase storage capacity, wetland restoration with the provision of full environmental services, and finally, the construction of impoundments with outlet flow structures. Construction costs were estimated from wetland restoration guidelines and the experiences of previous water storage projects in the region. Land rental values were based on rental value surveys in North Dakota and county tax assessments in Minnesota. The expected benefits of storage programs were based on expected reductions in historical (1989-1998) flood damage both within and downstream of the watersheds extrapolated over a hypothetical 20-year period in the future. Specifically, reduced peak flood stage associated with alternative levels of wetland restoration- and impoundment-based storage options, were based on hydrological modeling estimates in the Maple Watershed by Bengtson and Padmanabahn (1999). With the absence of stage-damage curves, it was assumed that reduced peak flood stage reduced monetary flood damage proportionally. Reduced flood damage was extrapolated and discounted for a hypothetical 20-year period in the future meaning that the historical wet cycle of the last 10-year cycle (which included a ‘100-year’ flood) was assumed to continue in the future. This study utilized site-specific biophysical and socioeconomic data and a variety of assumptions that were justified by previous studies and the existing literature. Most of the assumptions used were either beneficial or neutral to the cause of using wetland restoration to reduce flood damage. These include the assumptions that the recent wet cycle (1989-1998) will continue for the next 20 years, and that reductions in peak flood stage was assumed to reduce historical flood damage proportionately. These assumptions, which increase the potential benefits of wetland and impoundment programs, strengthen or provide additional confidence in the major conclusion this study: that large-scale wetland restoration programs are not an economically feasible way to reduce flood damage in the RRV. None of the wetland restoration options evaluated were economically feasible. Benefitcost ratios ranged from 1:1.4 for the most optimistic scenario (simple, low cost, and high storage), to 1:4.1 for more complex restoration (intended to provide a full range of wetlandbased goods and services). Similarly, wetland restoration options would need to reduce peak flood stage and related damage by between 3.4% to 10% to break even. The inclusion of additional (nonflood-related) wetland benefits did not make wetland restoration economically feasible. More research on this topic is warranted, especially with regard to the creation of nonflood-related wetland benefits when large numbers of wetlands are restored across entire watersheds and/or regions. Finally, the use of impoundments to store water and reduce flood damage was evaluated. Impoundments can store more water than wetlands but they are not likely to provide many 43 additional goods and services. Impoundments were found to be more cost effective in reducing flood damage than wetlands, but with benefit-cost ratios ranging from 1:1.1 to 1:1.7, they also were not considered economically feasible. In conclusion, neither wetland restoration alternatives, nor the construction of impoundments are considered economically feasible methods to store water and reduce flood damage in the RRV. However, it is possible (although not evaluated fully in this present study) that such storage options could be feasible if implemented on a site-specific basis. This would likely involve cases when storage sites can be procured at below average rental rates in locations with direct hydrological impacts on downstream flooding, especially when they provide significant levels of additional (nonflood-related) goods and services. For this reason, more research is warranted regarding strategies to improve both the valuation of the goods and services provided by wetlands and the site-specific modeling of the interactions between wetland storage and flooding. 44 References Bengtson, M., and G. Padmanabahn. 1999. A Hydrological Model for Assessing the Influence of Wetlands on Flood Hydrographs in the Red River Basin. Report Submitted to the International Joint Commission. Civil Engineering Department at North Dakota State University, Fargo. Carlson, G. 1999. The Economic Effects of the 1997 Flood on the Red River Valley. Master’s Thesis, University of North Dakota, Department of Geography, Grand Forks. Eppich, D., S. Apfelbaum, and L. Lewis. 1998. “Small Wetlands Use for Stormwater Runoff Management in the Red River of the North Basin.” In: Compendium of Technical and Scientific Advisory Committee Working Papers. Report Prepared in Support of the Red River Basin Damage Reduction Workgroup. Hammer, D. A., B. P. Pullin, T. A. McCaskey, J. Eason, and V.W. Payne. 1993. “Treating livestock wastewater with constructed wetlands.” In Constructed wetlands for water quality improvement. Edited by Gerald Moshiri, Boca Raton, Fl. Lewis Publishers, Inc. Hey, D., and N. Philippi. 1995. “Commentary on Flood Reduction Through Wetland Restoration: The Upper Mississippi River Basin as a Case History.’” Restoration Ecology 3(1), 1995. Houston Engineering, Inc. 1999. Personal Communications. Fargo, North Dakota. Houston Engineering, Inc. 1999. Flood Storage Easements Pilot Program Proposal. Fargo, North Dakota. Hubbard, D. 1982. “Surface Area-Volume Relationships of Wetland Basins in the Interior of the Coteau Des Prairies.” Proceedings of the South Dakota Academy of Science Vol. 61. Hubbard, D. 1989. “Wetland values in the prairie pothole region of Minnesota and the Dakota’s. Biological Report 88(43). Brookings, SD: U.S. Fish and Wildlife Service, Cooperative Research Service. IJC. 1997. Interim Report of the International Joint Commission Red River Basin Task Force. Red River Flooding, Short-Term Measures. December 1997. IJC-NDSU. 1999. The Collection of GIS-Based Data in the Maple (ND) and Wild Rice (MN) Sub-Watersheds of the Red River Valley. Report compiled by the Agricultural Economics Department, North Dakota State University for the Red River Basin Taskforce of the International Joint Commission. Fargo, North Dakota. International Coalition for Land/Water Stewardship in the Red River Basin. 1989. Land and Water Guide. Moorhead, Minnesota. 45 Kjelland, M. 1999. Unpublished Master’s Thesis. Department of Agriculture Economics. North Dakota State University, Fargo. Krenz, G., and J. A. Leitch. 1993. A River Runs North: Managing an International River. Red River Water Resources Council. Fargo, North Dakota. Leitch, J. A., and B. Hovde. 1996. “Empirical valuation of prairie potholes: Five case studies.” Great Plains Research 6:25-39. Ludden, A, D., Frink, and D. Johnson. 1983. “Water Storage Capacity of Natural Wetland Depressions in the Devils Lake Basin of North Dakota” Journal of Soil and Water Conservation 38(1):45-48. McCombs-Knutson. 1984. Water Resources Engineering/Planning Program for the Red River of the North Basin in Minnesota. Prepared by McCombs-Knutson Associates, Inc. for the Lower Red River Watershed Management Board (LRRWMB), May 1984. Meyer, A. J. 1998. Estimating the Effect Wetlands Have on Flood Hydrographs, Through Computer Simulation. Master’s Thesis, Department of Agricultural Economics, North Dakota State University, Fargo. Miller, R. 1999. Hydrologic Effects of Wetland Drainage and Land Use Change in a Tributary Watershed of the Minnesota River Basin: A Modeling Approach. Master’s Thesis, University of Minnesota, St. Paul. Miller, J. E., and D. L. Frink. 1984. Changes in Flood response of the Red River of the North Basin, North Dakota-Minnesota. Water-Supply Paper 2243, U.S. Geological Survey, Alexandria, Virginia. Minnesota Board of Water and Soil Resources. 1992. Minnesota Wetland Restoration Guide. St. Paul, Minnesota. Moore, I, and C. Larson. 1980. “Hydrologic impact of draining small depressional watersheds.” Journal of Irrigation and Drainage. Proceedings of the ASCE, 106(IR4):345363. North Dakota State Water Commission. 1996. UPDATE: Governor’s Emergency Plan for Devils Lake. Bismarck. North Dakota State Water Commission. 1997. UPDATE: Governor’s Emergency Plan for Devils Lake. Bismarck. North Dakota State Water Commission. 1998. UPDATE: Governor’s Emergency Plan for Devils Lake. Bismarck. 46 North Dakota State Water Commission. 1999. UPDATE: Governor’s Emergency Plan for Devils Lake. Bismarck. Red River Management Board. 1993. Hydrologic Analysis of Restored Wetlands. Internal Report (author unknown), December, 17, 1993. Renner, R. 1999. “Wetland Projects Help Reduce Flooding in North Dakota.” Waterfowl 2000, 12(1) page 7. Roberts, L. A., and J. Leitch. 1997. Economic valuation of some wetland outputs of Mud Lake Minnesota-South Dakota. Ag. Econ. Report #381, Department of Agricultural Economics, North Dakota State University, Fargo. Sierra Club. 1997. Red River Rampage: How restoring wetlands and moving homes from floodplains can reduce future flood risk. A Report by the Sierra Club (Midwest Office) and the Clean Water Network, May, 1997. Sierra Club. 1998. Red River Valley: Future Flooding or Sensible Solutions? How Basin Wide Coordination and Wetland Protection Can Reduce the Risk of Flooding in the Red River Valley. Report by the Agassiz Basin Group of the Sierra Club, Sheridan WY, August 3, 1998. Sip, R. L. 1998. An Economic Assessment of Wetland Mitigation in Northwest Minnesota. Master’s Thesis, Department of Agricultural Economics, North Dakota State University, Fargo. Terry, K., and L. Aadland. 1997. Unpublihsed DNR Memorandum dated August, 1997. U.S. Army Corps of Engineers. 1976. Natural Valley Storage: A partnership with Nature. New England Division, Waltham Massachusetts. U.S. Army Corps of Engineers. 1978. Interim Feasibility Study, Red River of the North Main Stem, Hydrologic Data. St. Paul District, St. Paul, Minnesota, 49 p. U.S. Army Corps of Engineers. 1980. Red River of the North Reconnaissance Report-Wild Rice Marsh Rivers Subbasin. St. Paul District, St. Paul, Minnesota. U.S. Army Corps of Engineers. 1994. Sharing the Challenge: Floodplain Management into the 21st Century. Report of the Interagency Floodplain Management Review Committee to the Administration Floodplain Management Taskforce. Washington, DC. U.S. Army Corps of Engineers. 1998. 404 Permit for Maple River Dam Draft Environmental Impact Statement. Omaha District, Nebraska. 47 Appendix A: Abbreviations Used in the Report AF Acre Feet ASAP Available Storage Acreage Program BWSR Board of Water and Soil Resources CACFDAS Computerized Agricultural Crop Flood Damage Assessment System CCJWRD Cass County Joint Water Resources District DEIS Draft Environmental Impact Statement DNR Department of Natural Resources EBI Environmental Benefits Index FEMA Federal Emergency Management Agency GIS Geographical Information Systems NDSWC North Dakota State Water Commission NWI National Wetlands Inventory RIM Reinvest In Minnesota RRV Red River Valley USFWS United States Fish and Wildlife Service USGS United States Geological Society WRW Wild Rice Watershed WRWMB Wild Rice Watershed Management Board 48 Appendix B: Impoundment-Storage Programs in North Dakota and Minnesota 1) The Devils Lake ASAP Program in North Dakota In the Devils Lake area of North Dakota, the Governor’s Emergency Plan for Action, which includes the Available Storage Acreage Program (ASAP), was initiated in 1996. The plan is intended to help reduce flooding associated with high water levels in Devils Lake, which is located in northeastern North Dakota, approximately 90 miles west of Grand Forks, North Dakota (NDSWC, 1997). At the time of inception, the upper-basin storage component had a potential storage capacity of approximately 75,000 AF, which could hypothetically reduce the lake level by approximately one foot. The goal of the ‘smaller tract water storage’ portion of the ASAP program is to store approximately 35,000 AF by creating smaller tracts of water storage on available public and private lands. This part of the component is called the Available Storage Acreage Program (ASAP). ASAP’s inception was a joint effort between the North Dakota State Water Commission and pertinent federal agencies having jurisdiction in water-related issues. The federal agencies involved include: 1) Natural Resources Conservation Service (NRCS), 2) U.S. Department of Agriculture (USDA), 3) U.S. Army Corps of Engineers (USACE), 4) U.S. Fish and Wildlife Service (USFWS), and 5) Environmental Protection Agency (EPA). Under agreement by the aforementioned authorized agencies, the newly created water storage areas and the expanded or restored wetlands will not be regulated provided three criteria are adhered to. The following guidelines and rules have been applied to the ASAP program: 1) One-year leases will be negotiated to retain floodwater. Lease extensions may be available if funding is available and it is agreeable to the landowner. 2) Only land areas that contribute water to Devils Lake are eligible for ASAP. 3) The NDSWC evaluates sites on the basis of cost to implement, lease cost, and volume of water stored. Since program funding is limited, areas holding the largest volumes of water for the least amount of money will be given first priority. Payment will be made only for acres flooded. 4) Approved participants will be offered ASAP contracts. The participant will be paid the contract price after spring runoff was captured at the site verification is completed by NDSWC staff. 5) For sites involving multiple landowners, all affected landowners must agree to participate and are eligible for payment for any impacted land. 6) Water control structures will be earthen berms or drain plugs with or without gated control at the landowner’s preference. 7) Control structures will be left in place after the agreement expires for landowners to use as they wish. If removal of the structure is desired at the end of the lease, it will 49 be the responsibility of the landowner and should be reflected in the bid. 8) Landowners may use the land immediately adjacent to the impounded water as they wish; there is no upland requirement. 9) ASAP does not provide or require the replanting of the water retention area after the agreement expires. 10) Agreement has been secured to ensure nonwetland areas will not become classified as jurisdictional wetlands. 11) Landowners who are interested in participating in ASAP should submit an application to the NDSWC or the Devils Lake Joint Board. The application is nonbinding. ASAP stored approximately 8,000 AF of water in 1996. Approximately 22,000 and 21,000 AF of water were stored, respectively, in 1997 and 1998 (NDSWC, 1998). As of January 15, 1999, it was estimated that inflows into Devils Lake were reduced by approximately 40,00050,000 AF, which kept the lake from rising about four to five inches (NDSWC, 1999). The minimum lease payment for ASAP was set at $50 per acre (NDSWC, 1996). Governor Schafer determined this payment level in December 1995 after hearing comments from three public meetings. Actual payments are a function of the landowner’s bid and volume of water storage. Although ASAP compensated landowners based on competitive bids, the need for water storage was greater than the interest. Therefore, most landowners received the maximum payment for water withheld (NDSWC, 1999). The maximum payment was set according to the damage the volume of water would expect to have caused if added to Devils Lake. At that time, 75,000 AF of water added to Devils Lake would have caused a one-foot rise at a cost of $5 million, which equated to $60 per AF (NDSWC, 1999). The maximum payment was reduced to $50 per AF as the basin became wetter and interest among landowners grew. This value is approximately 170% higher than average rental rates in the area (as reported by the North Dakota Agricultural Statistics Service). The difference between average rental rates and the ASAP payments is considered a price ‘premium’ that farmers require due to uncertainty concerning the potential effects of water storage on the long- term productivity and value of agricultural land. However, this premium might be relatively high because apparently, many farmers in the region knew in advance the maximum rental payments the state would be willing to make. Presently, however, the damage caused by a one-foot increase in Devils Lake would equate to approximately 100,000 AF and cause an average of $25 million in damage (NDSWC, 1999). Current annual water storage in ASAP is approximately 20,000-22,000 AF. With the current lake levels, costs beyond the current storage levels would likely increase rapidly and may not be considered feasible (NDSWC, 1999). Through 1998, approximately $2,578,000 has been spent on ASAP. An additional $950,000 has been allocated for 1999 ASAP. Costs are primarily for payments to participants, though secondary costs involve preparation of sites. Approximately $430,000 was spent on water storage in 1996 with 1997 site renewal costing $403,000. Sites completed in the fall of 1996 cost $47,000, while sites originated in 1997 cost approximately $691,000 (NDSWC, 1997). ASAP was allocated $1,150,000 for 1998 of which roughly $1,000,000 was spent (NDSWC, 1997, 1998). 50 2) The Proposed Wild Rice Storage Program The Wild Rice Watershed Management Board (WRWMB) is currently developing a pilot project utilizing natural water storage areas in the upper watershed to reduce local and watershed flood damage. More specifically, the project identifies natural water detention sites in areas that are linked to downstream flood damage and where only minor structural changes and improvements are necessary to enhance the water storage capabilities at these sites. Flood storage easements are then intended to be purchased which would involve regular flood-storage activities. According to Houston Engineering (1999), the primary project benefits would include: 1) flood protection, 2) water quality improvement, 3) erosion control, 4) wildlife habitat enhancement, 5) recreational opportunities, and 6) aesthetic enhancement. Flood protection should be enhanced through the project's water storage sites, which are intended to reduce peak flows downstream through runoff retention near its origin. Retained runoff will then be slowly released, reducing peak flows, decreasing erosion, and reducing downstream property damage. It is also hypothesized that a variety of related water quality, wildlife, and recreational benefits will result from the project (Houston Engineering, 1999). The specific rules of the easement involve farmers providing the WRWMB the right of entry and exit to paid-for lands for the purposes of structural construction and maintenance and the right to retain runoff on the land until a suitable time for controlled discharge. Possible storage sites will be selected based on location in the upper watershed, geophysical land characteristics, and overall cost effectiveness of the flood storage site. The location should provide the maximum potential flood storage with the least amount of structural improvements. Utilizing existing topography, soil characteristics, and natural water retention areas will help to reach this criterion. The overall cost-effectiveness of each site will be measured as a function of project costs and volume of water storage obtained. According to Houston Engineering (1999), if overall costs are less than $1,000/AF of storage obtained, the project could be deemed feasible. As a guideline, this measure would take into account the entire project rather than be limited to individual detention sites within the project. Easement acquisition will be a multi-step procedure contingent on previous steps: 1. Problem identification is the first step in the procedure and targets existing drainage problems. Often landowners are the best source identifying drainage problems. Therefore public advertising of the program is vital. In addition to landowners, township, county, state, federal, and private organizations all can be beneficial in problem identification. 2. After drainage problems have been identified, the specific storage site(s) will be identified and proposed. USGS quadrangle maps, aerial photos, and other methods will be essential for this step in order to identify approximate flood pool sizes, affected areas, and affected landowners. This is required in order to meet with landowners. 3. Once a problem and corresponding storage area has been identified, representatives from the Watershed District will meet with the affected landowner(s) to determine interest in the 51 project. Though only a rough conceptual framework will be known at this point, enough information should be available for a landowner to determine interest or lack of in the program. Individuals and circumstances vary and, therefore, some may require more information in order to feel comfortable with an informed decision. The main objective at this stage is to gain an option for an easement or purchase agreement for the required land if the land is deemed feasible before proceeding to the next step. 4. The option for easement or purchase allows the process to continue with assurance that the landowner has agreed to the easement or purchase of the land if determined to be feasible. This guarantees that future funds will not be expended studying unavailable sites. Upon attainment of this option, a detailed survey of the site can be initiated, including preliminary design, a cost estimate, and a hydrologic and hydraulic analysis. 5. Overall cost effectiveness of the site(s) in storing water will determine the decision to continue or abandon the project. The cost estimate will need to include but not be limited to construction, engineering, administration, legal counsel, land acquisition, permitting, and meetings (Houston Engineering, 1999). A preliminary guideline of $1,000/AF of storage or less will determine the cost effectiveness of the project. 6. If a decision to implement the project is determined, the option(s) for easement/purchase will be exercised. Once legal obligations are finalized, final design and construction can be implemented. A variety of regulations, situations, and locations will likely be encountered with potential water storage sites, including lands currently participating in government programs such as the Conservation Reserve Program (CRP). Therefore, it is necessary to have a dynamic rather than a uniform payment system. Following is a list addressing potential situations and recommended actions: 1. Flood Storage Easement: The easement is the primary proposed method of flood storage acquisition. The landowner retains ownership of the property and may continue to use for agricultural and recreational purposes. The easement will be purchased at a price not to exceed the County Assessor’s Township Average Market Value available annually through the Minnesota Department of Revenue, or the individual parcel assessed value as recorded by the county auditor (Houston Engineering, 1999). The landowner will continue to pay property taxes on the land while the easement will be transferred to the Watershed District for the life of the agreement. 2. Fee Simple Land Purchase: In this situation, the land would become the property of the Watershed District with all landowner rights and responsibilities being transferred from the landowner to the Watershed District. Payment for the land would follow the same criterion previously outlined. 3. Flood Storage Easement on lands currently involved in CRP: The land acquisition method will need to meet the usage requirements of the CRP contract. CRP payments for individuals are based on the duration of the contract and the bid the 52 landowner had approved upon acceptance into the program. Due to this variation, flood storage easements would be based on the value for adjacent land not enrolled in CRP or a similar program. The price paid would follow the previously outlined criterion. In the event the landowner prefers to sell the land parcel rather than granting an easement, the purchase price would be the same as the easement price and the CRP contract would transfer to the Watershed District (Houston Engineering, 1999). It should be noted that there is some debate whether land enrolled in CRP can be used for water storage. The main question focuses on whether water storage would be incongruent with the accepted bid and resulting CRP contract. Currently, there is not a consensus on this issue. It can be reasonably argued that water storage on current and new CRP contracts likely would not be allowable within the framework of the contract as signed. However, expiring contracts that are up for continuous renewal or future sign-ups may allow water storage if written into the contract. The answer likely will come from the interpretation of how water storage is viewed within the Environmental Benefits Index (EBI) of the CRP bid. 4. Flood Storage Easements on existing Reinvest In Minnesota (RIM) program lands: Acquiring flood storage easements on existing RIM lands would require coordination and approval through the Board of Water and Soil Resources (BWSR). RIM program requirements would need to be met in addition to approval of the overall flood storage plan. Once again, the price paid would follow the previously outlined criterion. Criteria for design and construction of the water storage sites will be based on the “Minnesota Wetland Restoration Guide,” (1992), prepared by the Minnesota Board of Water and Soil Resources. Though the guide addresses both flood control and wildlife production, flood control is the primary objective of this project. Therefore, recommendations for flood control will be emphasized though designs may vary from site to site and from the guide. Deviations from the guide will be followed if deemed advisable by a licensed professional engineer. Each design will depend on location, topography, available storage, and costeffectiveness. Earthen embankments and principal and emergency spillway systems will be the primary methods for water control. The principal spillway controls inflow and outflow from the detention site while the emergency spillway safely discharges runoff that exceeds the maximum capacity of the site (Houston Engineering, 1999). Monitoring Procedures: Monitoring will be the responsibility of the Wild Rice Watershed District and consists of inspections and maintenance. Specifically, inspections will be based on criteria set forth in the “Minnesota Wetland Restoration Guide,” (1992). These criteria emphasize seasonal inspections for three years following initial construction. Inspections should also follow any flood events. Observations and corrective actions or plans should be documented. Following the first three years, inspections should be performed annually in May or June, after major flood events, or after any condition that may negatively affect the integrity of the site. 53
© Copyright 2026 Paperzz