effects of dam impoundment on the flood regime of natural floodplain communities in the upper...

16
ABSTRACT: Understanding the effects of dams on the inundation regime of natural floodplain communities is critical for effective decision making on dam management or dam removal. To test the implications of hydrologic alteration by dams for floodplain natural communities, we conducted a combined field and modeling study along two reaches in the Connecticut River Rapids Macrosite (CRRM), one of the last remaining flowing water sections of the Upper Connecticut River. We surveyed multiple channel cross sec- tions at both locations and concurrently identified and surveyed the elevations of important natural communities, native species of con- cern, and nonnative invasive species. Using a hydrologic model, HEC-RAS, we routed estimated pre- and post-impoundment dis- charges of different design recurrence intervals (two year through 100 year floods) through each reach to establish corresponding reductions in elevation and effective wetted perimeter following post-dam discharge reductions. By comparing (1) the frequency and duration of flooding of these surfaces before and after impound- ment and (2) the total area flooded at different recurrence inter- vals, our goal was to derive a spatially explicit assessment of hydrologic alteration, directly relevant to natural floodplain com- munities. Post-impoundment hydrologic alteration profoundly affected the subsequent inundation regime, and this impact was particularly true of higher floodplain terraces. These riparian communities, which were flooded, on average, every 20 to 100 years pre-impound- ment, were predicted to flood at 100 100 year intervals, essential- ly isolating them completely from riverine influence. At the pre-dam five to ten year floodplain elevations, we observed smaller differences in predicted flood frequency but substantial differences in the total area flooded and in the average flood duration. For floodplain forests in the Upper Connecticut River, this alteration by impoundment suggests that even if other stresses facing these com- munities (human development, invasive exotics) were alleviated, this may not be sufficient to restore intact natural communities. More generally, our approach provides a way to combine site specif- ic variables with long term gage records in assessing the restora- tive potential of dam removal. (KEY TERMS: riparian; dams; floods; watershed management; aquatic ecosystems.) INTRODUCTION Riverine plant and animal communities are strong- ly influenced by hydrologic regime. Numerous studies have demonstrated that the abundance and distribu- tion of floodplain species can be predicted on the basis of the magnitude, frequency, and duration of flood flows (Hupp and Osterkamp, 1985). In the United States, however, most large rivers have been dammed for decades (Graf, 1999), ultimately changing histori- cal flow regimes and potentially altering natural com- munities (Molles et al., 1998; Graf, 2001). Because riparian habitats are associated with unique distur- bance regimes (i.e., flooding) and distinct physical conditions, they contain a disproportionately high number of rare species and rare natural community types. Restoring and conserving these species and natural communities is a goal of many river restora- tion efforts, including dam removal. These conservation and river regulation issues are highly relevant to the New England region of the northeastern United States. New England has a long and unique history of river regulation and land use change, (Foster, 1990; Graf, 1999), and possesses dis- tinctive natural floodplain communities (Metzler and Damman, 1985; Bechtel and Sperduto, 1998; Kears- ley, 1999). While recent studies have measured the effects of hydrologic alteration on riparian natural communities in other regions (Barnes 1997; Howe and Knopf, 1991), information on the effect of hydrologic alteration in New England is lacking. Region specific information is critical to conserve and restore these communities and habitats, and to understand the 1 Paper No. 01189 of the Journal of the American Water Resources Association. Discussions are open until June 1, 2003. 2 Respectively, Research Scientist, Northeastern Research Station, USDA-U.S. Forest Service, Amherst, MA 01003; Professor and Research Assistant, Dartmouth College, 6017 Fairchild, Hanover, New Hampshire 03755; Research Scientist, The Nature Conservancy, 21/2 Beacon Street, Concord, New Hampshire 03301; and The Nature Conservancy, 27 State Street, Montpelier, Vermont 05602-2934 (E-Mail/Nislow: [email protected]). JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 1533 JAWRA JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION VOL. 38, NO. 6 AMERICAN WATER RESOURCES ASSOCIATION DECEMBER 2002 EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER 1 Keith H. Nislow, Francis J. Magilligan, Heidi Fassnacht, Doug Bechtel, and Ana Ruesink 2

Upload: keith-h-nislow

Post on 23-Jul-2016

214 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

ABSTRACT: Understanding the effects of dams on the inundationregime of natural floodplain communities is critical for effectivedecision making on dam management or dam removal. To test theimplications of hydrologic alteration by dams for floodplain naturalcommunities, we conducted a combined field and modeling studyalong two reaches in the Connecticut River Rapids Macrosite(CRRM), one of the last remaining flowing water sections of theUpper Connecticut River. We surveyed multiple channel cross sec-tions at both locations and concurrently identified and surveyed theelevations of important natural communities, native species of con-cern, and nonnative invasive species. Using a hydrologic model,HEC-RAS, we routed estimated pre- and post-impoundment dis-charges of different design recurrence intervals (two year through100 year floods) through each reach to establish correspondingreductions in elevation and effective wetted perimeter followingpost-dam discharge reductions. By comparing (1) the frequency andduration of flooding of these surfaces before and after impound-ment and (2) the total area flooded at different recurrence inter-vals, our goal was to derive a spatially explicit assessment ofhydrologic alteration, directly relevant to natural floodplain com-munities.

Post-impoundment hydrologic alteration profoundly affected thesubsequent inundation regime, and this impact was particularlytrue of higher floodplain terraces. These riparian communities,which were flooded, on average, every 20 to 100 years pre-impound-ment, were predicted to flood at 100 ≥ 100 year intervals, essential-ly isolating them completely from riverine influence. At thepre-dam five to ten year floodplain elevations, we observed smallerdifferences in predicted flood frequency but substantial differencesin the total area flooded and in the average flood duration. Forfloodplain forests in the Upper Connecticut River, this alteration byimpoundment suggests that even if other stresses facing these com-munities (human development, invasive exotics) were alleviated,this may not be sufficient to restore intact natural communities.More generally, our approach provides a way to combine site specif-ic variables with long term gage records in assessing the restora-tive potential of dam removal.(KEY TERMS: riparian; dams; floods; watershed management;aquatic ecosystems.)

INTRODUCTION

Riverine plant and animal communities are strong-ly influenced by hydrologic regime. Numerous studieshave demonstrated that the abundance and distribu-tion of floodplain species can be predicted on the basisof the magnitude, frequency, and duration of floodflows (Hupp and Osterkamp, 1985). In the UnitedStates, however, most large rivers have been dammedfor decades (Graf, 1999), ultimately changing histori-cal flow regimes and potentially altering natural com-munities (Molles et al., 1998; Graf, 2001). Becauseriparian habitats are associated with unique distur-bance regimes (i.e., flooding) and distinct physicalconditions, they contain a disproportionately highnumber of rare species and rare natural communitytypes. Restoring and conserving these species andnatural communities is a goal of many river restora-tion efforts, including dam removal.

These conservation and river regulation issues arehighly relevant to the New England region of thenortheastern United States. New England has a longand unique history of river regulation and land usechange, (Foster, 1990; Graf, 1999), and possesses dis-tinctive natural floodplain communities (Metzler andDamman, 1985; Bechtel and Sperduto, 1998; Kears-ley, 1999). While recent studies have measured theeffects of hydrologic alteration on riparian naturalcommunities in other regions (Barnes 1997; Howe andKnopf, 1991), information on the effect of hydrologicalteration in New England is lacking. Region specificinformation is critical to conserve and restore thesecommunities and habitats, and to understand the

1Paper No. 01189 of the Journal of the American Water Resources Association. Discussions are open until June 1, 2003.2Respectively, Research Scientist, Northeastern Research Station, USDA-U.S. Forest Service, Amherst, MA 01003; Professor and Research

Assistant, Dartmouth College, 6017 Fairchild, Hanover, New Hampshire 03755; Research Scientist, The Nature Conservancy, 21/2 BeaconStreet, Concord, New Hampshire 03301; and The Nature Conservancy, 27 State Street, Montpelier, Vermont 05602-2934 (E-Mail/Nislow:[email protected]).

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 1533 JAWRA

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATIONVOL. 38, NO. 6 AMERICAN WATER RESOURCES ASSOCIATION DECEMBER 2002

EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURALFLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER1

Keith H. Nislow, Francis J. Magilligan, Heidi Fassnacht, Doug Bechtel, and Ana Ruesink2

Used Mac Distiller 5.0.x Job Options
This report was created automatically with help of the Adobe Acrobat Distiller addition "Distiller Secrets v1.0.5" from IMPRESSED GmbH. You can download this startup file for Distiller versions 4.0.5 and 5.0.x for free from http://www.impressed.de. GENERAL ---------------------------------------- File Options: Compatibility: PDF 1.3 Optimize For Fast Web View: Yes Embed Thumbnails: Yes Auto-Rotate Pages: No Distill From Page: 1 Distill To Page: All Pages Binding: Left Resolution: [ 600 600 ] dpi Paper Size: [ 612 792 ] Point COMPRESSION ---------------------------------------- Color Images: Downsampling: Yes Downsample Type: Bicubic Downsampling Downsample Resolution: 144 dpi Downsampling For Images Above: 216 dpi Compression: Yes Automatic Selection of Compression Type: Yes JPEG Quality: << /VSamples [ 2 1 1 2 ] /HSamples [ 2 1 1 2 ] /Blend 1 /QFactor 0.9 /ColorTransform 1 >> Bits Per Pixel: As Original Bit Grayscale Images: Downsampling: Yes Downsample Type: Bicubic Downsampling Downsample Resolution: 144 dpi Downsampling For Images Above: 216 dpi Compression: Yes Automatic Selection of Compression Type: Yes JPEG Quality: << /VSamples [ 2 1 1 2 ] /HSamples [ 2 1 1 2 ] /Blend 1 /QFactor 0.9 /ColorTransform 1 >> Bits Per Pixel: As Original Bit Monochrome Images: Downsampling: Yes Downsample Type: Bicubic Downsampling Downsample Resolution: 300 dpi Downsampling For Images Above: 450 dpi Compression: Yes Compression Type: CCITT CCITT Group: 4 Anti-Alias To Gray: No Compress Text and Line Art: Yes FONTS ---------------------------------------- Embed All Fonts: Yes Subset Embedded Fonts: Yes Subset When Percent Of Characters Used is Less: 35 % When Embedding Fails: Warn and Continue Embedding: Always Embed: [ ] Never Embed: [ ] COLOR ---------------------------------------- Color Management Policies: Color Conversion Strategy: Convert All Colors to sRGB Intent: Default Working Spaces: Grayscale ICC Profile: Adobe Gray - 20% Dot Gain RGB ICC Profile: sRGB IEC61966-2.1 CMYK ICC Profile: None Device-Dependent Data: Preserve Overprint Settings: No Preserve Under Color Removal and Black Generation: No Transfer Functions: Preserve Preserve Halftone Information: No ADVANCED ---------------------------------------- Options: Use Prologue.ps and Epilogue.ps: No Allow PostScript File To Override Job Options: No Preserve Level 2 copypage Semantics: Yes Save Portable Job Ticket Inside PDF File: No Illustrator Overprint Mode: Yes Convert Gradients To Smooth Shades: Yes ASCII Format: No Document Structuring Conventions (DSC): Process DSC Comments: Yes Log DSC Warnings: No Resize Page and Center Artwork for EPS Files: No Preserve EPS Information From DSC: No Preserve OPI Comments: No Preserve Document Information From DSC: Yes OTHERS ---------------------------------------- Distiller Core Version: 5000 Use ZIP Compression: Yes Deactivate Optimization: No Image Memory: 524288 Byte Anti-Alias Color Images: No Anti-Alias Grayscale Images: No Convert Images (< 257 Colors) To Indexed Color Space: Yes sRGB ICC Profile: sRGB IEC61966-2.1 END OF REPORT ---------------------------------------- IMPRESSED GmbH Bahrenfelder Chaussee 49 22761 Hamburg, Germany Tel. +49 40 897189-0 Fax +49 40 897189-71 Email: [email protected] Web: www.impressed.de
Adobe Acrobat Distiller 5.0.x Job Option File
<< /ColorSettingsFile () /LockDistillerParams true /DetectBlends true /DoThumbnails true /AntiAliasMonoImages false /MonoImageDownsampleType /Bicubic /GrayImageDownsampleType /Bicubic /MaxSubsetPct 35 /MonoImageFilter /CCITTFaxEncode /ColorImageDownsampleThreshold 1.5 /GrayImageFilter /DCTEncode /ColorConversionStrategy /sRGB /CalGrayProfile (Adobe Gray - 20% Dot Gain) /ColorImageResolution 144 /UsePrologue false /MonoImageResolution 300 /ColorImageDepth -1 /sRGBProfile (sRGB IEC61966-2.1) /PreserveOverprintSettings false /CompatibilityLevel 1.3 /UCRandBGInfo /Remove /EmitDSCWarnings false /CreateJobTicket false /DownsampleMonoImages true /DownsampleColorImages true /MonoImageDict << /K -1 >> /ColorImageDownsampleType /Bicubic /GrayImageDict << /HSamples [ 2 1 1 2 ] /VSamples [ 2 1 1 2 ] /Blend 1 /QFactor 0.9 >> /CalCMYKProfile (None) /ParseDSCComments true /PreserveEPSInfo false /MonoImageDepth -1 /AutoFilterGrayImages true /SubsetFonts true /GrayACSImageDict << /VSamples [ 2 1 1 2 ] /HSamples [ 2 1 1 2 ] /Blend 1 /QFactor 0.9 /ColorTransform 1 >> /ColorImageFilter /DCTEncode /AutoRotatePages /None /PreserveCopyPage true /EncodeMonoImages true /ASCII85EncodePages false /PreserveOPIComments false /NeverEmbed [ ] /ColorImageDict << /HSamples [ 2 1 1 2 ] /VSamples [ 2 1 1 2 ] /Blend 1 /QFactor 0.9 >> /AntiAliasGrayImages false /GrayImageDepth -1 /CannotEmbedFontPolicy /Warning /EndPage -1 /TransferFunctionInfo /Preserve /CalRGBProfile (sRGB IEC61966-2.1) /EncodeColorImages true /EncodeGrayImages true /ColorACSImageDict << /VSamples [ 2 1 1 2 ] /HSamples [ 2 1 1 2 ] /Blend 1 /QFactor 0.9 /ColorTransform 1 >> /Optimize true /ParseDSCCommentsForDocInfo true /GrayImageDownsampleThreshold 1.5 /MonoImageDownsampleThreshold 1.5 /AutoPositionEPSFiles false /GrayImageResolution 144 /AutoFilterColorImages true /AlwaysEmbed [ ] /ImageMemory 524288 /OPM 1 /DefaultRenderingIntent /Default /EmbedAllFonts true /StartPage 1 /DownsampleGrayImages true /AntiAliasColorImages false /ConvertImagesToIndexed true /PreserveHalftoneInfo false /CompressPages true /Binding /Left >> setdistillerparams << /PageSize [ 576.0 792.0 ] /HWResolution [ 600 600 ] >> setpagedevice
Page 2: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

effects of dam management. The Connecticut River,the largest river in New England, provides a usefulillustration of these research challenges. The Con-necticut has been dammed for power and flood controlfor the last half of the 20th century (Fallon-Lambert,1998). Like many rivers in the region, the ConnecticutRiver has a long stream gage record, encompassingboth pre- and post-regulation time periods, which canbe used to measure hydrologic alteration at the wholeriver scale (km to100 km) (Richter et al., 1996; Mag-illigan and Nislow, 2001). However, stream gage anal-yses alone are incapable of assessing changes in theflood frequency and duration at specific locations onthe floodplain, or changes in overall area flooded atdifferent recurrence intervals. To accomplish this sitespecific analysis requires detailed knowledge ofspecies distributions, channel dimensions, and therelationship between river discharge and water sur-face elevation at local (m to 100 m) scales.

Establishing a site specific approach is particular-ly relevant to dam removal, where maintenance andrestoration of natural riparian communities is oftenan explicit goal. Determining the effects of removingdams has generally been based on fluvial theory orconceptual models (Bednarek, 2001; Doyle et al.,2002). These models, however, still require fundamen-tal knowledge about initial boundary conditionsand/or the geomorphic and ecological impacts of damsbefore potential effects can be estimated. In thispaper, we present a case study of hydroecologicalimpacts resulting from long term impoundment in theConnecticut River, and we further provide techniquesfor estimating the impacts of dams, which can then beused as a basis for assessing riparian adjustments fol-lowing dam removal. We used a three part researchapproach. First, in order to determine the overalleffects of flow regulation on flood magnitude and fre-quency, we applied two techniques, the Indicators ofHydrologic Alteration (IHA) (Richter et al., 1996) andlog-Pearson Type III flood frequency analysis, to thepre- and post-regulation stream gage records for theUpper Connecticut River. Second, we conducted floris-tic surveys to note the locations and elevations of nat-ural floodplain communities and target species at twoimportant floodplain sites and surveyed channel crosssections to establish channel dimensions at multiplelocations within these reaches. Lastly, we used ahydrologic model, HEC-RAS, to route estimated pre-and post-impoundment discharges of different designrecurrence intervals (2, 5, 10, 20, 50, and 100 yearfloods), and determined the discharge needed to inun-date the specific floodplain surfaces on which differ-ent species and communities occurred. By comparingthe frequency and duration that these surfaces wereflooded, and the area flooded at different recurrenceintervals, our goal is to obtain a “spatially explicit”

assessment of hydrologic alteration, directly relevantto natural floodplain communities, that can helpguide future dam management decisions.

STUDY AREA

Our study was conducted in the Connecticut RiverRapids Macro Site (CRRM) (Figure 1). The CRRM, a24 km stretch of the Upper Connecticut beginning atthe confluence of the Ottaqueechee and ConnecticutRivers, currently supports several regionally andglobally threatened floodplain species and communi-ties. This CRRM is the largest remaining section ofthe Upper River that has not been essentially con-verted from fluvial, running water, into lacustrine,stillwater, habitat by impoundments (Fallon-Lambert,1998). The geologic setting of the Upper Connecticutin this section of river imparts major constraints onfloodplain communities. The river flows down strikeof a synclinal trough of highly resistant rocks, produc-ing a constrained valley throughout most of the Ver-mont and New Hampshire border and inhibitingsignificant lateral channel migration. This confine-ment generates a channel of low sinuosity with a verynarrow meander plain. The geologically narrow mainvalley is further constrained by a series of late Pleis-tocene and Holocene terraces. As a result, appropriatesites for floodplain communities are limited, and com-munity occurrences are highly patchy. Geomorphiclimitations are further exacerbated by extensive agri-cultural and residential development, particularly onhigher floodplain terraces (Bechtel and Sperduto,1998).

The Connecticut River has a long history of flowregulation with several hydroelectric dams in theupper part of the basin dating back to the late 1920s(Fallon-Lambert, 1998). Wilder Dam, within theupper reaches of the CRRM, is a medium head hydro-electric facility built in 1951. While this single facilityhas minimal flood storage capability, its constructionallowed the entire series of dams on the Upper Con-necticut to be operated as an integrated systemdesigned for both power generation and flood control(Fallon-Lambert, 1998). In addition, its close proximi-ty to the CRRM increases the importance of this facil-ity in controlling overbank floods. As a result, we usethe Wilder Dam completion data (1951) date todemarcate the pre-dam period (prior to 1951) from thepost-dam period (1951 to present). A U.S. GeologicalSurvey (USGS) gage was installed in 1911 approxi-mately 3 km downstream of the present day locationof the dam, and daily discharges have been collectedregularly since then providing a robust pre- and post-dam hydrologic data set. The Connecticut River

JAWRA 1534 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

Nislow, Magilligan, Fassnacht, Bechtel, and Ruesink

Page 3: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

drains 10,600 km2 at the gage and includes runofffrom the southern and eastern Green Mountains of Vermont and the western flanks of the White

Mountains in New Hampshire. In addition to main-stem dams, most major tributaries have flood con-trol structures that operate either as constant pool

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 1535 JAWRA

Effects of Dam Impoundment on the Flood Regime of Natural Floodplain Communities in the Upper Connecticut River

Figure 1. Map of Study Sites. Map of northeastern United States adapted from the Connecticut River Watershed Council Inc.Watershed map displayed at http: //www.ctriver.org/infofaq.html. Close up map of the Connecticut River adapted

from the U.S. Department of Agriculture, Soil Conservation Commission Cartographic Division map.

Page 4: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

operation or as run of the river during the summerthat have significant impacts on mainstem hydrology(Magilligan and Nislow, 2001).

Two specific reaches along the CRRM, the HartIsland Rivershore Complex and Burnaps Island, wereselected for study based on their support of importantnatural communities and species of concern, and theirproximity to Wilder Dam (Figure 1). The Hart IslandRivershore Complex is comprised of Hart Island, a975 m long, 244 m wide island situated along the mid-dle reach of the CRRM which harbors a range of flood-plain species and community types, an adjacentfloodplain rivershore forest, and a rocky rivershoreoutcrop which contains one of only four known occur-rences of the globally endangered Astragalus robbin-sii var. jessupi (Jessup’s milk vetch). Burnaps Island,at the upper end of the CRRM, is a smaller, lower ele-vation island that also contains a range of floodplaincommunities. In addition, the state listed Cincindelamarginipennis (cobblestone tiger beetle), which is con-fined to river islands (Nothnagle, personal communi-cation), occurs at both sites.

Conservation Targets: Natural Communities, RareSpecies, and Nonnative Invasive Species

Two levels of biological organization have been rec-ognized by conservation scientists as appropriate tar-gets for conservation: species (fine filter approach)and natural communities (coarse filter approach)(Noss and Cooperrider, 1994). Natural communitiesare defined by three distinguishing characteristics: (1)a definite plant species composition, (2) a consistentphysical structure, and (3) a specific set of naturalconditions (such as different combinations of nutri-ents, drainage and climate) (Sperduto, 1997). Whilethe majority of species (85 to 90 percent by some esti-mates) and natural processes should be protected byfocusing on patterns of natural community distribu-tion, rare species falling outside such a network needto be recognized and protected as well (Noss, 1987).This study used a combination of fine filter and coarsefilter approaches. Based on recent data from bothNew Hampshire and Vermont Natural Heritage Pro-gram studies on floodplain forest communities (Bech-tel and Sperduto, 1998; Kearsley, 1999), we identifiedand delineated major natural community types. Wesubsequently identified all occurrences of species thatare presently listed as threatened or endangered ateither the state or the federal level.

In addition to rare native species and unique natu-ral communities, riparian areas also commonlyprovide habitat for nonnative invasive plants species.Invasive plants take advantage of high sunlight along

edges, often spread through high volumes of winddispersed and waterborn seeds, and tend to thrive inbare mineral and disturbed soils, all commonattributes in riparian habitats. Conservation man-agers have recognized that the control of non-nativeinvasive species is necessary to maintain natural pat-terns of native vegetation. In fact, exotic species areone of the greatest threats to native species and tohuman disturbed ecosystems (e.g., impounded rivers)in the world (Reid and Miller, 1989). For the purposesof this study, rare plant and animal species, and non-native invasive species are henceforth referred to as“target species.” While this is not meant to implygoals for conservation, we do imply that these speciesare necessary to understand for management of biodi-versity.

METHODS

Hydrologic Analysis

We used two techniques to characterize the pre-versus post-impoundment flow regime of the UpperConnecticut River. The Indicators of Hydrologic Alter-ation (IHA) (Richter et al., 1996) uses mean daily dis-charges and calculates 32 indices that describe thehydrologic regime for that station. In essence, themodel evaluates changes in maxima, minima, timing,and rate of change, at several temporal scales (oneday to 90 day). Differences in these indicators repre-sent the magnitude of hydrologic alteration associatedwith pre- versus post-impoundment periods of record.The IHA, however, does not effectively documentchanges in the frequency and magnitude of extreme,rare flood events that may be particularly importantto floodplain forest communities. To address this, weused a log-Pearson Type III analysis (U.S. WaterResources Council, 1976) for the annual maximumflood series and calculated the estimated 2, 5, 10, 20,50, and 100 year discharges for the pre- and post-impoundment streamflow record.

Channel Surveys

Field data were gathered between July and Decem-ber 1999, at the Hart Island Rivershore Complex(hereafter referred to collectively as Hart Island) andBurnaps Island sites. Cross section elevation datawere collected along transects oriented approximatelyperpendicular to the main river valley walls. Six crosssections were established at Hart Island: one

JAWRA 1536 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

Nislow, Magilligan, Fassnacht, Bechtel, and Ruesink

Page 5: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

upstream of the island before the channel splits, fourevenly spaced cross sections characterizing the island,and one downstream of the island after the singlechannel reforms. The same format was used at Bur-naps Island. Since the island is somewhat smaller,only two cross sections were needed to characterizethe island, for a total of four cross sections at the site.Cross sections were surveyed using an automaticlevel, survey tape, and stadia rod. Surveyed eleva-tions at both sites were tied into USGS benchmarksto establish absolute elevations for all surveyedpoints. In addition to channel bed and floodplain ele-vations, water surface elevations were surveyed ateach transect to calculate low flow water surfaceslope. Following the mid-September 1999 regionalflood associated with Hurricane Floyd, the high waterline at each transect was marked and surveyed forsimilar calculations for a higher discharge. Distances

between transects were determined using GPS datacollected with a hand held GPS unit. These data weredifferentially corrected using base station data fromMontpelier, Vermont.

Community Surveys

We surveyed plant communities in floodplain areasin the Hart and Burnaps Island study sites during athree day period in June 1999. Surveys consisted oftwo basic activities. First, lateral transects, perpen-dicular to the river channel were established. Alongthese transects, upper and lower elevational limits ofeight community types (Table 1) were then identifiedand flagged. Second, the locations of three species ofconcern (Table 1) were marked wherever theyoccurred. Locations of Jessup’s milk vetch were taken

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 1537 JAWRA

Effects of Dam Impoundment on the Flood Regime of Natural Floodplain Communities in the Upper Connecticut River

TABLE 1. Natural Communities Characterized in June 1999 Surveys of Burnaps Island and the Hart Island Complex WithCharacteristic Dominant Plant Species, and Constituent Rare Species and Nonnative Exotic Plants. Natural community

classifications based on recent data from both New Hampshire and Vermont Natural Heritage Program studies on floodplainplant communities (Bechtel and Sperduto, 1998; Kearsley, 1999). Indicator species reflect dominance (percent cover) within eachnatural community type. Species of concern are federally threatened or endangered (T or E) or state listed (SL) species. Invasive

species are those that are known to be present and problematic within the natural community type at the study sites.

Community Type Indicator Species Species of Concern Invasive Species

Bare Cobble, Riverside Prunus pumila (Sand cherry) Cincindela marginipennis Euphorbia cyparissusCobble, Barren Populus sp. (Cottonwoods) (Cobblestone tiger beetle) (Cypress spurge)

seedlings and saplings Cyanchum nigrum(Black swallowwort)

Riverside Meadow Calamagrostis canadensis Cincindela marginipennis Cyanchum nigrum(Blue joint grass) (Cobblestone tiger beetle) (Black swallowwort)Phalaris arundinacea Physostegia virginiana Fallopia japonica(Reed canary grass) (Obedience plant) (Japanese knotweed)

Riverside Thicket Cornus amomum (Silky dogwood) Cyanchum nigrumVitis aestivalus (Summer grape) (Black swallowwort)

Fallopia japonica(Japanese knotweed)

Silver Maple Floodplain Acer saccharinum (Silver maple) Cyanchum nigrumForest Matteucia struthiopteris (Ostrich fern) (Black swallowwort)

Fallopia japonica(Japanese knotweed)Lonicera morrowii(Morrow’s honeysuckle)

Sugar Maple Floodplain Acer saccharum (Sugar maple) Staphylea tridentata (Bladdernut) Lonicera morrowiiForest Tilia americana (Basswood) (Morrow’s honeysuckle)

Disturbed Floodplain Forest Acer negundo (Box elder) Lonicera morrowii(Morrow’s honeysuckle)

River Terrace Floodplain Mix of upland and floodplain tree Staphylea tridentata (Bladdernut) Lonicera morrowiiForest species (Morrow’s honeysuckle)

Page 6: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

from permanent markers established for a long termpopulation study (Nothnagle, unpublished data). Ele-vations of all flagged locations were measured bytying in survey measurements at the locations toabsolute elevations on USGS benchmarks.

Hydrologic Modeling

Water surface elevations were estimated at bothstudy sites for a series of pre- and post-dam designdischarges having recurrence intervals (RI) rangingfrom 2 to 100 years. Using the one dimensional stepbackwater model HEC-RAS, we estimated the dis-charges necessary to inundate surfaces where naturalcommunities and target species were observed, andwe determined the total wetted perimeter at each ofthe transects for the 2 to 100 year RI floods, pre- andpost-dam. HEC-RAS, an updated version of HEC-2, isa hydraulic model developed by the U.S. Army Corpsof Engineers that calculates water surface elevationand other hydraulic variables from discharge andfield derived channel data using a standard step iter-ative process (Hoggan, 1989), based on Manning’sequation:

V = [(R.67) * (Se.5)]/n

where V = velocity, R = hydraulic radius (equal todepth in wide shallow channels like the ConnecticutRiver), Se = energy slope, and n = Manning’s rough-ness coefficient.

To calibrate the model, we ran HEC-RAS for flowsof known discharges and corresponding water surfaceelevations. These discharges included the relativelylow flow conditions that occurred in mid-October1999, and the higher flow conditions associated withHurricane Floyd in mid September 1999. Water sur-face elevations predicted by the model were comparedto measured values. Manning’s roughness coefficientswere then adjusted within the model to improve thecorrelation between modeled and measured watersurface elevations. Once the model was calibrated inthis manner, it was run for pre- and post-dam dis-charges for the estimated 2, 5, 10, 20, 50, and 100year floods.

Analysis of Spatially Explicit Hydrologic Alteration

Natural community and target species elevationand location data were used in combination with thehydrologic model to answer three major questionsconcerning the influence of hydrologic alteration.

These comparisons formed the basis of our assess-ment of spatially explicit hydrologic alteration.

1. Do discharges necessary to inundate floodplainspecies and communities have different probabilitiesof occurrence pre- and post-impoundment?

We used HEC-RAS to determine the dischargesnecessary to inundate the range of eachspecies/community elevations (low, mid, andhigh). We then compared the probability ofachieving these discharges based on pre- versuspost-impoundment gage records.

2. Do these discharges have different average dura-tions pre- and post-impoundment?

Daily discharge data from the USGS gage inWest Lebanon, New Hampshire, were used toassess pre- and post-impoundment flood dura-tion differences. For each year of record, we cal-culated the number of days when dischargeequaled or exceeded that necessary to inundatea specific species or community occurrence. Wethen calculated the difference in the averageduration between pre- and post-impoundmentrecords for each community/species at low, mid,and high elevation levels. These calculationswere made in two ways: (1) for all pre-post years,and (2) for only those years where inundationdischarges were achieved, to remove the effectsof zero years (which essentially repeats the floodfrequency information) from the analysis of floodduration.

3. Does total flooplain area inundated at specificintervals differ for pre- versus post-dam conditions?

Because HEC-RAS is based on discrete cross sec-tions, our estimates of changes in total areainundated are based on changes in total wettedperimeter (WPT) calculated at each cross sec-tion, for each recurrence interval flood. In addi-tion, since our interest was in assessing therelative changes in inundated area above bank-full, we subtracted out the bankfull wettedperimeter (WPbf) from both pre- and post-damoutput. The effect of impoundment on floodedperimeter for a specific recurrence interval floodis therefore expressed as:

JAWRA 1538 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

Nislow, Magilligan, Fassnacht, Bechtel, and Ruesink

Page 7: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

RESULTS

Overall Shifts in Hydrologic Regime Based on GageRecords

Impoundment in the Upper Connecticut River hassignificantly reduced the frequency and magnitude oflarge floods. Log-Pearson Type III analysis indicatesthat while pre- and post-dam flood discharges weresimilar for bankfull discharges (approximately twoyear flood) (Figure 2a and Table 2), differencesbecame progressively magnified at higher magnitude,less frequent events. For example, the pre-dam fiveyear flood now has a recurrence interval of approxi-mately 20 years, while the pre-dam ten year floodwould have a less than 1 percent chance of occurringin any given year (i.e., the 100 year flood). Theseresults were reflected in the IHA analysis which indi-cated an approximate 15 percent decrease in the mag-nitude of the average annual one day maximumdischarge (Figure 2b).

These effects manifest not just in modeled esti-mates, but also in the actual flow record. In the past49 years of impoundment, the old, pre-dam two yearbankfull discharge has been equaled or exceeded atleast 18 times using the annual flood record and 26times using the partial duration record (Table 3).Although the partial duration record increases thenumber of discharges greater than bankfull, almostall of these additional events are still between the twoyear and five year flood. The largest post-dam flowalong the CRRM occurred during the 1973 water yearwhen a discharge of 2230 m3/s (between a five yearand 10 year event) occurred during a late June storm.By comparison, during the same interval (1951 to pre-sent), the adjacent unregulated White River in Ver-mont (drainage area = 1800 km2) has experiencedthree 10 year floods and a 20 year flood.

Overall Shifts in River Stage

Differences between pre- and post-impoundmentdischarges of greater than five year RI floods resultedin substantial differences in river stage at both studysites. While pre- and post-dam water surface eleva-tions were nearly identical for the bankfull flood, theremainder of the post-dam water surface elevations

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 1539 JAWRA

Effects of Dam Impoundment on the Flood Regime of Natural Floodplain Communities in the Upper Connecticut River

Percent Change in Overbank WPWP WP WP WP

WP WP

T pre dam bf pre dam T post dam bf post dam

T pre dam bf pre dam=

−( ) − −( )[ ]−( )

− − − −

− −

( ) ( ) ( ) ( )

( ) ( )* 100

TABLE 2. Pre- and Post-Dam Discharges for Design RecurrenceIntervals. Discharges based on daily peak flow data obtained

for USGS stream gage 01144500 (Connecticut River atWest Lebanon, New Hampshire). 1950 demarcates

pre- from post-impoundment gage record.

Q Pre-Dam Recurrence Interval Q Post-Dam(cms) (years) (cms)

4,389 100 2,1783,710 050 2,0982,917 020 1,9652,401 010 1,8381,940 05 1,6761,376 02 1,351

Figure 2. (a) Results of Log Pearson Type III Analysis of Pre- andPost-Impoundment Floods. Upper line (triangles) gives themagnitude/frequency relationship pre-impoundment; lowerline (boxes) gives magnitude/frequency post-impoundment.

(b) IHA Analysis of the Magnitude of the One-DayMaximum Discharge Pre- and Post-Impoundment.

Page 8: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

were greater than 0.5 m below corresponding pre-damlevels. The difference increased progressively withdischarge with the estimated post-dam 100 year floodapproximately 3 m below pre-dam levels (Table 4 andFigure 3). To put in perspective the magnitude ofthese differences, consider the following statistics: theHartland Rivershore Floodplain Forest lies approxi-mately 5 to 7 m above the channel bottom, and thetop of Hart Island is a maximum of 12 m above thechannel bottom along our cross sections. BurnapsIsland is somewhat lower, with its maximum eleva-tion being only around 7 m above the channel bottom.Therefore, a decrease in stage of 2 to 3 m is substan-tial. When actual releases are evaluated, we find thatin the past 50 years of impoundment, the largest dis-charge observed (2230 m3/s), would barely inundatethe pre-disturbance 10 year flood surface. At Hartand Burnaps Islands, riparian surfaces at elevationsgreater than about 98 m and 101 m, respectively,have not been inundated since Wilder Dam beganoperating.

Elevations of Natural Communities and Speciesof Concern

Natural community and species of interest occur-rences at both study sites were generally well distin-guished on the basis of elevation (Figures 4 and 5).Herbaceous communities (Cobble Barren, RiversideMeadow, Riverside Thicket) dominated the less thantwo year floodplain, giving way to floodplain forests athigher elevations. Among the herbaceous communi-ties, Riverside Thicket occurred at higher elevationsthan meadows or barrens; but these community typeswere poorly distinguished on the basis of elevation.Among floodplain forests, silver maple dominatedfloodplain forests occurred at the lowest, and riverterrace forest occurred at the highest elevations, withsugar maple floodplain forests intermediate betweenthese two. Elevations of silver maple floodplainforests were substantially higher on the Hart Island

JAWRA 1540 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

Nislow, Magilligan, Fassnacht, Bechtel, and Ruesink

TABLE 3. Partial Duration Peak Flows Following Impoundment,Compared to the Predicted Recurrence Interval of the Flood

Pre-Impoundment. USGS gage 01144500, ConnecticutRiver at West Lebanon, New Hampshire.

Water Discharge Pre-Dam RecurrenceRank Year (cms) Interval (years)

1 1973 2,231 Less than 10year flood2 1953 2,0673 1976 1,9854 1987 1,929 Five year flood5 1984 1,8646 1972 1,7487 1969 1,7488 1998 1,6899 1994 1,653

10 1958 1,61911 1982 1,60712 1960 1,60713 1952 1,59314 1973 1,48315 1983 1,48116 1959 1,45217 1973 1,44718 1954 1,43019 1986 1,42420 1982 1,42121 1955 1,42122 1996 1,41823 1996 1,39624 1996 1,39325 1976 1,39026 1979 1,382 Two year flood

Figure 3. Schematic Depiction of Water Surface ElevationAchieved at Different Recurrence Intervals Pre-

and Post-Impoundment at Burnaps Island.

Page 9: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

(20 to 50 year floodplain) than on the rivershore (2 to10 year floodplain). Disturbed floodplain forests werenot well distinguished by elevation, and may repre-sent a primary successional variant of any of thethree other forest types. Where communities orspecies occurred at both study sites, elevations werehigher for Burnaps occurrences due to this islandbeing upstream of the Hart Island Complex. Withrespect to species of concern, obedience plant occur-rences and cobblestone tiger beetle habitat corre-sponded to the herbaceous riverside communities,while bladdernut occurred in association with highterrace floodplain forest communities. Elevations ofJessup’s milk vetch corresponded roughly to silvermaple floodplain forests, even though this speciesoccurred in a different community type (Rocky Out-crop Community; not considered in this study). Withrespect to invasive exotics, honeysuckle occurredmostly in conjunction with sugar maple and river ter-race floodplain forests, and knotweed occurred mostlyin association with riverside meadow and thicketcommunities.

Flood Regimes of Species and Communities

Hydrologic alteration profoundly affected thehydrologic regime of floodplain communities (Figures6, 7, and 8). This was particularly true of sugar maple

and river terrace floodplain forests, which includedbladdernut and honeysuckle. These communitieswhich were flooded on average every 10 to 100 yearspre-impoundment, were predicted to lack inundationfollowing impoundment. At intermediate elevations (5to 10 year pre-dam floodplain), where silver maplefloodplain forests, riverside thickets, and Jessup’smilk vetch were found, we observed smaller differ-ences in predicted flood frequency. Differences in floodfrequency were found only at the high end of the ele-vational range of these species and communities, andgenerally involved shifts from the five year to the 10to 20 year RI flood (Figure 8).

In addition to changes in flood frequency, we alsoobserved changes in flood duration (the average num-ber of days per year a surface was predicted to beinundated). Differences in flood duration are mostapparent for two to five year floods. While differencesin flood frequency were relatively minor for theserecurrence interval floods, the average duration ofthese floods were 20 to 40 percent longer beforeimpoundment (Figure 9a), essentially shifting from afour to seven, to a one to three day average duration(Figure 9b). Floods of greater than five year RI occurso infrequently that flood duration impacts were notdiscernable; floods of less than two year RI exhibitedno discernable changes in either frequency or dura-tion of flooding.

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 1541 JAWRA

Effects of Dam Impoundment on the Flood Regime of Natural Floodplain Communities in the Upper Connecticut River

TABLE 4. Modeled Water Surface Elevations (m) for Design Discharges.

Two Five 10 20 50 100Year RI Year RI Year RI Year RI Year RI Year RI

Transect Pre Post Pre Post Pre Post Pre Post Pre Post Pre Post

HART ISLAND

1 95.50 95.45 96.40 95.95 96.92 96.22 97.64 96.44 98.46 96.47 99.00 96.60

2 95.39 95.34 96.29 95.83 96.81 96.11 97.54 96.33 98.39 96.36 99.24 96.49

3 95.12 95.07 95.96 95.53 96.45 95.79 97.15 96.01 97.98 96.01 98.89 96.13

4 94.98 94.93 95.79 95.36 96.26 95.63 96.96 95.83 97.71 95.83 98.43 95.95

5 94.80 94.75 95.57 95.15 96.02 95.41 96.72 95.61 97.54 95.59 98.25 95.71

6 94.66 94.62 95.40 94.98 95.81 95.24 96.47 95.43 97.18 95.39 97.86 95.51

Average -0.05 -0.43 -0.64 -1.14 -1.94 -2.60Difference

BURNAPS ISLAND

1 100.28 100.22 100.93 100.61 101.37 100.75 102.13 100.97 103.00 101.02 103.88 101.25

2 99.83 99.76 100.35 99.95 100.82 100.15 101.70 100.41 102.73 100.49 103.73 100.64

3 99.63 99.55 100.05 99.67 100.50 99.82 101.45 100.10 102.53 100.20 103.57 100.36

4 99.33 99.27 99.49 99.28 100.03 99.28 100.89 99.54 101.81 100.36 102.74 99.95

Average -0.07 -0.33 -0.68 -1.29 -2.14 -2.94Difference

Page 10: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

JAWRA 1542 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

Nislow, Magilligan, Fassnacht, Bechtel, and Ruesink

Figure 4. Absolute Elevations of Natural Communities on(a) Hart and (b) Burnaps Island Study Sites. “Low” and

“High” are the lowest and highest elevations wherewe observed the community; “Mid” refers to the

midpoint between lowest and highest occurrences.

Figure 5. Absolute Elevations of Species-of-Concern on(a) Hart and (b) Burnaps Island Study Sites at Low, Mid,

and High Elevation Occurrences (see previous figure).“N” indicates native species; “E” indicates exotic species.

Page 11: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

Hydrologic alteration also had major influences onthe total area flooded at any given recurrence inter-val. Although we found substantial differences in thechanges in wetted perimeter among transects withina site, dependent on local topography, overbankwetted perimeter was substantially decreased at all

transects for both study sites. At the intermediate(pre-dam 5 to 10 year) RI floodplain, overbank wettedperimeters were decreased on average approximately60 percent following impoundment at Burnaps Island,and decreased approximately 50 percent at the HartIsland Complex (Figure 10).

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 1543 JAWRA

Effects of Dam Impoundment on the Flood Regime of Natural Floodplain Communities in the Upper Connecticut River

Figure 6. Flood Recurrence Probabilities (1 * RI-1; RI = Recurrence Interval) for High-Terrace Natural CommunitiesPre- and Post-Impoundment on Hart and Burnaps Islands at Low, Mid, and High Elevation Occurrences.

Page 12: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

DISCUSSION

We found that long term impoundment of theUpper Connecticut River has altered the historicalflow regime experienced by important riparian natu-ral communities. This work is the first direct assess-ment of the implications of hydrologic alterationimpacts on these communities for the New Englandregion. Due to the greater effect of impoundment oninfrequent, high intensity floods, both the magnitudeand type of hydrologic impact varied systematicallywith floodplain elevation. Moving down from highestto lowest floodplain terraces, observed impacts shiftedfrom major changes in flood frequency, to changes inflood duration and total area flooded, and, at the low-est elevations, to no discernable changes. Obviously,this pattern of hydrologic alteration (major effects onlarge floods, minor effects on smaller floods) and theparticular channel topographies of our sites, are spe-cific to this study. However, our approach provides away to incorporate these variables into analyses ofdam effects at other locations, yielding an assessmentof the restorative potential of dam removal or changes

in dam management. This site specific methodology,as opposed to models derived from general popula-tions of river systems, has also been advocated bylarge scale efforts, such as the multidisciplinaryKissimmee River restoration project (Warne et al.,2000)

For the specific case of the Upper ConnecticutRiver, floodplain forests and species of concern occur-ring on higher floodplain terraces, have essentiallybeen completely isolated from riverine influence inthe post-dam environment. At the regional scale,these high terrace floodplain forests represent some ofthe most floristically diverse habitats in the UpperConnecticut River basin (Bechtel and Sperduto,1998). They are also among the most threatened byagricultural and residential development, and cur-rently occupy only a small portion of their formerrange. Our results indicate that in the Upper Con-necticut River, hydrologic alteration may threaten thelong term stability of the few remaining higher ter-race floodplain forests. In addition, our analysis indi-cates that even if other stresses were alleviated,hydrologic alteration may prevent recovery and

JAWRA 1544 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

Nislow, Magilligan, Fassnacht, Bechtel, and Ruesink

Figure 7. Flood Recurrence Probabilities (1 * RI-1; RI = Recurrence Interval) for High-Terrace Species of concernPre- and Post-Impoundment on Hart and Burnaps Islands at Low, Mid, and High Elevation Occurrences.

Page 13: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

restoration of intact natural communities on higherfloodplain terraces. For example, at Transect 5 onHart Island Rivershore we observed a broad terrace(now in agricultural use) at approximately the sameelevation as the sugar maple floodplain forest on HartIsland. Our evidence currently suggests that simplyallowing the forest to reestablish would not be suc-cessful, given the present day isolation of this surfacefrom current riverine influence.

Similarly, effects on total area flooded may haveimportant implications for the viability of floodplainspecies and communities, and is an important comple-ment to understanding changes at specific floodplainsurfaces. While species and communities may moveup and down elevational gradients in response tochanges in flood regime, reduction in the total areasubject to a specific recurrence interval flood will sig-nificantly reduce species ability to find appropriategermination and recruitment sites, and decrease theaverage size of floodplain vegetation occurrences,making them more vulnerable to invasion by exotics.This is a particularly important issue for the Con-necticut River, where opportunities for floodplain

species recruitment are already severely limited byhuman development pressure and underlying geologi-cal constraints. Expansion of this analysis beyondthese two study sites, along with more detailed spa-tial analysis to look at changes in the size and spatialarray of floodplain vegetation patches at a broaderscale, will significantly improve our understanding ofthe effects of hydrologic alteration in the Connecticutbasin.

We have demonstrated that combining stream gagerecord analysis, ecological and channel surveys, andhydrologic modeling, revealed important aspects ofhydrologic alteration impacts to floodplain communi-ties. This approach, however, has several importantlimitations.

First, if channel dimensions change significantly asa function of impoundment, hydrologic models maynot accurately predict the water surface elevations ofpre-dam flood discharges. For example, if significantbed aggradation occurs following impoundment, the model would overpredict river stage for pre-dam floods based on current channel dimensions. In the Upper Connecticut River, due to geological

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 1545 JAWRA

Effects of Dam Impoundment on the Flood Regime of Natural Floodplain Communities in the Upper Connecticut River

Figure 8. Flood Recurrence Probabilities (1 * RI-1; RI = Recurrence Interval) for Intermediate Terrace Natural CommunitiesPre- and Post-Impoundment on Hart and Burnaps Islands at Low, Mid, and High Elevation Occurrences.

Page 14: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

constraints on channel form, along with the sedimentcontributions of the White River, a major, unimpound-ed tributary entering the Connecticut just belowWilder Dam, we feel that this is not a major issue inour study. Furthermore, the Connecticut River flowson resistant bedrock throughout the CRRM, therebyinhibiting channel incision. However, for rivers withmore mobile beds, or for larger impoundments, post-dam channel change may significantly complicateinterpretation of model results.

Second, it is likely that only considering the fre-quency and duration of inundation is insufficient tofully characterize hydrologic alteration. In our study,

this insufficiency may be particularly significant forfloodplain forests at relatively low floodplain eleva-tions. For example, Hughes and Cass (1997) foundthat the silver maple floodplain forest at Little OtterCreek, an unimpounded river in western Vermont,was flooded approximately every two years. However,almost all the individuals were recruited to the popu-lation in a single year, when a major overbank(greater than two year RI) flood occurred. This sug-gests that while two year floods are sufficient tomaintain existing stands, successful recruitmentrequires a much larger event, possibly to scour ordeposit floodplain sediments, and/or remove potentialcompetitors (Bendix, 1999). Hydrologic alteration ofthe type we observed, while having minor effects onfrequency and duration of flooding where some exist-ing stands occur, might have major effects on theprobability of successful recruitment and long-termviability of the population. Determining whether sil-ver maple recruitment is consistently associated withflood events of specific recurrence intervals is logisti-cally possible, via analysis of tree cores from popula-tions along rivers with long term stream gages.Obtaining these data is essential for understandingthe impact of hydrologic alteration on this importantfloodplain community type.

Finally, a spatially explicit analysis may be compli-cated if the distribution of floodplain species and com-munities tracks changes in flood frequency caused byhydrologic alteration. This is of particular concern formobile animals and for annual plant species withrapid generation times, and less of a concern for slow-er growing woody vegetation. In our study sites, silverand sugar maple floodplain forests appear to be sig-nificantly older than our pre-impoundment date of1950, suggesting that they were established under thepre-impoundment hydrologic regime. Additionally,because of the lack of suitable floodplain surfaces, theability of species and communities to alter their dis-tribution is limited in this system.

In terms of broader management issues within theConnecticut River system, this study highlights thetype of concerns identified by Kondolf and Wilcock(1996). Prescriptive approaches in habitat manage-ment have focused on controlled releases to maintainand restore specific riparian habitat. The timing andmagnitude of these controlled releases are critical forspecific habitat (Wu 2000; Schmidt et al., 2001), yetno single flow or discharge will universally ameliorateor mitigate existing disturbed environmental condi-tions. For the post-impoundment hydrologic regimewithin the CRRM, the broadly different style andmagnitude of hydrologic disturbance has dispropor-tionately affected specific species and communities,thereby precluding a simple and distinct managementsolution, short of completely re-establishing the

JAWRA 1546 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

Nislow, Magilligan, Fassnacht, Bechtel, and Ruesink

Figure 9. (a) Percentage Change in Flood Frequency andFlood Duration Post-Impoundment Versus Discharge.Stippled lines indicate pre-impoundment discharges

of different recurrence intervals as landmarks.(b) Frequency distribution of the number of days

flooded for years in which the two to five yearsurface was flooded at least once duringthe year. N = 11 years pre-impoundment;

N = nine years post-impoundment.

Page 15: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

natural flow regime identified by Poff et al. (1997).Analyses, as shown herein, identify the type and mag-nitude of those disturbances and may help to priori-tize the prescriptive remedies for restoring criticalaspects of the riparian zone.

ACKNOWLEDGMENTS

We would like to thank Jen Stamp, Matt Kelley, Adam Sepulve-da, Jolyon Rivoir-Pruszinski, Alice Hartley, Phil Nothnagle, ChrisFichtel, Elizabeth Thompson, Pat McCarthy, Jeanne Anderson, and

Mike Stevens for help with field surveys, and two anonymousreviewers for comments on the manuscript. This work was support-ed by a grant from the U.S. Environmental Protection AgencyRegional Geographic Initiative Grant Program (July 1996) to TheNature Conservancy and by a Reiss Grant from the Dartmouth Col-lege Rockefeller Center for Social Sciences to F. J. Magilligan.

LITERATURE CITED

Barnes, W. J., 1997. Vegetation Dynamics on the Floodplain of theLower Chippewa River in Wisconsin. Journal of the TorreyBotanical Society 124:189-197.

Bechtel, D. A. and D. D. Sperduto, 1998. Floodplain Forest NaturalCommunities Along Major Rivers in New Hampshire. NewHampshire Natural Heritage Inventory. Concord, New Hamp-shire, 56 pp.

Bednarek, A. T., 2001. Undamming Rivers: A Review of the Ecologi-cal Impacts of Dam Removal. Environmental Management27(6):803-814.

Bendix, J., 1999. Stream Power Influence on Southern CalifornianRiparian Vegetation. Journal of Vegetation Science 10(2):243-252.

Doyle, M. W, E. H. Stanley, and J. M. Harbor, 2002. GeomorphicAnalogies for Assessing Probable Channel Response to DamRemoval. JAWRA 38(6):1567-1579.

Fallon-Lambert, K., 1998. Instream Flow Uses, Values and Policiesin the Upper Connecticut River Watershed. Connecticut RiverJoint Commissions Report, 178 pp.

Foster, D., 1992. Land Use History (1730-1990) and VegetationDynamics in Central New England, USA. Journal of Ecology80(4):253-772.

Graf, W. L., 1999. Dam Nation: A Geographic Census of AmericanDams and Their Large-Scale Hydrologic Impacts. WaterResources Research 35(4):1305-1311.

Graf, W. L., 2001, Damage Control: Restoring the Physical Integrityof America's Rivers. Annals Of the Association of AmericanGeographers 91(1):1-27.

Hoggan, D. H. 1989. Computer-Assisted Floodplain Hydrology andHydraulics. McGraw Hill Book Company, New York, New York.

Howe, W. H. and F. L. Knopf, 1991. On the Imminent Decline of RioGrande Cottonwoods in Central New Mexico. SouthwesternNaturalist 36:218-224.

Hughes, J. W. and W. B. Cass, 1997. Pattern and Process of a Flood-plain Forest, Vermont, USA: Predicted Responses of Vegetationto Perturbation. Journal of Applied Ecology 34(3):594-612.

Hupp, C. R. and W. R. Osterkamp, 1985. Bottomland VegetationDistribution Along Passage Creek, Virginia, in Relation to Flu-vial Landforms. Ecology 66(3):670-81.

Kearsley, J. B., 1999. Inventory and Vegetation Classification ofFloodplain Forest Communities in Massachusetts. Rhodora101(906):105-135.

Kondolf, G. M. and P. R. Wilcock, 1996. The Flushing Flow Prob-lem: Defining and Evaluating Objectives. Water ResourcesResearch 32(8):2589-2599.

Magilligan, F. J. and K. H. Nislow, 2001. Long-Term Changes inRegional Hydrologic Regime Following Impoundment in aHumid-Climate Watershed. JAWRA 37(6):1551-1569.

Metzler, K. J. and A. W. H. Damman, 1985. Vegetation Patterns inthe Connecticut River in Relation to Frequency and Duration ofFlooding. Naturaliste Canadien 12:535-549.

Molles, M. C., C. S. Crawford, L. M. Ellis, H. M. Valett, and C. N.Dahm, 1998. Managed Flooding for Riparian EcosystemRestoration – Managed Flooding Reorganizes Riparian ForestEcosystems Along the Middle Rio Grande in New Mexico. Bio-Science 48(9):749-756.

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 1547 JAWRA

Effects of Dam Impoundment on the Flood Regime of Natural Floodplain Communities in the Upper Connecticut River

Figure 10. Percentage Change in Overbank Wetted Perimeterat (a) Burnaps Island and (b) Hart Island Complex

Sites at River Cross Sections (Burnaps Island1 to 4; Hart Island Complex (1 to 6).

Page 16: EFFECTS OF DAM IMPOUNDMENT ON THE FLOOD REGIME OF NATURAL FLOODPLAIN COMMUNITIES IN THE UPPER CONNECTICUT RIVER

Noss, R. F., 1987. From Plant Communities to Landscapes in Con-servation Inventories: A Look at the Nature Conservancy (USA).Biological Conservation 41(1):11-37.

Noss, R. F. and A. Y. Cooperrider, 1994. Saving Nature’s Legacy,Protecting and Restoring Biodiversity. Island Press, WashingtonD.C.

Poff, N. L., J. D. Allan, M. B. Bain, J. R. Karr, K. L. Prestegaard, B. D. Richter, R. E. Sparks, and J. C. Stromberg, 1997. The Nat-ural Flow Regime. BioScience 47(11):769-784.

Reid, W. V. and K. R. Miller. 1989. Keeping Options Alive – The Sci-entific Basis for Conserving Biodiversity. World Resources Insti-tute, Washington, D.C.

Richter, B. D., J. V. Baumgartner, J. Powell, and D. P. Braun, 1996.A Method for Assessing Hydrologic Alteration Within Ecosys-tems. Conservation Biology 10(4):1163-1174.

Schmidt, J. C., R. A. Parnell, P. E. Grams, J. E. Hazel, M. A.Kaplinski, L. E. Stevens, and T. L. Hoffnagle, 2001. The 1996Controlled Flood in Grand Canyon: Flow, Sediment Transport,and Geomorphic Change. Ecological Applications 11(3): 657-671.

U.S. Water Resources Council, 1976. Guidelines for DeterminingFlood Flow Frequency. Bulletin 17, U.S. Government PrintingOffice, Washington, D.C.

Warne, A. G., L. A. Toth, and W. A. White, 2000. Drainage-Basin-Scale Geomorphic Analysis to Determine Reference Conditionsfor Ecologic Restoration – Kissimmee River, Florida. EcologicalApplications 112:884-899.

Wu, F. C., 2000. Modeling Embryo Survival Affected by SedimentDeposition Into Salmonid Spawning Gravels: Application toFlushing Flow Prescriptions. Water Resources Research36(6):1595-1603.

JAWRA 1548 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

Nislow, Magilligan, Fassnacht, Bechtel, and Ruesink