Report to:

FLP Maine Hydro LLC

150 Main Street, Lewiston, ME 04220

and

E/PRO Engineering and Environmental Consulting

249 Western Avenue, Augusta, ME 04330

by

Andrew M. Tuthill, P. E.

and

Kathleen D. White, P.E., PhD.

Ice Engineering Group

Cold Regions Research and Engineering Laboratory

U. S. Army Research and Development Center

72 Lyme Rd. Hanover, NH 03755

22 April 2004

The 95-year-old Ft. Halifax Dam, currently owned by FPL Energy Maine Hydro LLC (FPLE), is

located on the Sebasticook River just upstream of its confluence with the Kennebec River in

Winslow, Maine (Figs 1 & 2). The next dam on the Sebasticook is located 6 miles upstream at

Benton, ME. Removal of an 87-ft-long central section of the Fort Halifax Dam is proposed by

FPLE as a less costly alternative to constructing a fish elevator. The change will decrease the

low-flow pool height by about 24 ft. This study examines the effects of the proposed dam

modification on the ice regime of the lower Sebasticook and Kennebec Rivers.

On rivers with dams, sheet ice typically first appears in the pools immediately upstream of the

structures. Ice cover formation upstream of the pools follows, either as border ice that grows in

from the channel sides, or as frazil ice, formed in upstream open water reaches, that drifts down

and accumulates against the upstream edge of the advancing ice cover. Depending on the

magnitude of the downstream forces, pan-shaped frazil ice floes may accumulate edge-to-edge in

a "juxtaposed" ice cover, or collapse to form a multi-layered "shoved" ice accumulation.

Arriving frazil ice may be entrained beneath the leading edge of the ice cover. By this process,

pools above dams can store large volumes of frazil ice and block the movement of frazil ice to

downstream locations. At the onset of breakup, these thickened ice covers upstream of dams

commonly remain in place longer than the thinner ice found in steeper reaches. The frazil

deposits can block the breakup ice run, often forming ice jams at the upper end of the dam

impoundment and thereby reducing the ice volume reaching downstream ice jam sites.

The lowering or removal of a dam typically results in a longer free-flowing upstream reach,

typically with increased frazil ice production. At the same time, the pool length is shortened,

resulting in decreased frazil ice storage capacity. Because of these two factors, frazil ice is more

likely to pass the dam site following its removal or lowering. This frazil may cause freezeup ice

jams at downstream locations where these problems previously did not exist when the dam was

in place. The frazil accumulations displaced downstream may also block the breakup ice run,

causing ice jam flooding in reaches where the ice previously passed through without incident.

For these reasons, it is important to examine the impact of dam removal on the ice regime of a

river, specifically how it will affect the ice jam flood potential downstream. More information on

the effects of dam removal on the ice regime can be found in White and Moore (2002).

Since the 1999 removal of the Edwards Dam on the Kennebec River in Augusta ME, freezeup

ice jams have formed annually at the head of tide just downstream from the former dam site. As

previously mentioned, freezeup jams can block breakup ice runs later on in the winter or early

spring. Despite relatively mild conditions during ice cover breakup, several breakup ice jams

have formed on the Kennebec River in Augusta since the removal of the Edwards Dam,

generally for short periods of time. The breakup jam of 2003 was longer in duration and

increased stages to near flood level (Fig. 3).

Since ice jam conditions in Augusta currently approach flood stage, the Army Corps of

Engineers required, as part of the permit process, that the dam owner study the additional

changes to the ice regime due to the breaching of the Fort Halifax Dam.

The work was carried out as a joint effort between CRREL and FPLE’s engineering consultant

E/PRO Engineering and Environmental Consulting. Objectives were to characterize the ice

regimes on the lower Sebasticook and Kennebec Rivers under the pre and post Fort Halifax Dam

breaching cases. The approach combined research of historical information, field observation

and numerical ice-hydraulic modeling. Fig. 4 shows the extent of the study area.

2.1.Existing Ice Regime

CRREL and E/PRO collected and analyzed available information describing the existing ice

regime on the Sebasticook and Kennebec Rivers, including ice cover formation and progression,

as well as the nature of the ice breakup. The goals were to determine where sheet ice covers

form, which reaches produce frazil ice, how the frazil ice is transported, and where it deposits.

Also important were the timing and sequence of ice breakup and identification of sites where ice

jams typically form. Information sources included USGS records, the CRREL Ice Jam Database,

Ice Engineering Archives, newspapers articles, and discussions with people familiar with the two

rivers. E/PRO and FPLE provided relevant information from their files and their experience

with the rivers. In addition, CRREL, E/PRO and FPLE observed ice conditions from the air and

on the ground during the 2003-2004 winter. Periodic river inspections identified areas of sheet

ice, open water reaches, and frazil ice accumulations. Ice thickness measurements at key

locations complemented the field observations.

2.2 Review of Historical Information

The CRREL Ice jam Database reports 16 breakup ice jams on the Kennebec River between

Skowhegan and Richmond since 1795 (Table 1). Because many ice jams go unreported or

records are lost, the Ice Jam Database tends to under represent the number of actual ice jams.

The historical data suggest that, the Kennebec breaks up in sections, with multiple jams forming

upstream of dams, islands, bends and channel constrictions. An event on 13-14 March 1936

provides a worst-case example with ice jams or ice jam flooding reported upstream of Waterville

at Hinkley, and below the Edwards Dam at Augusta, Randolph, Gardiner and Richmond. In

1936, the pre-breakup ice on the Kennebec was at least 2-ft-thick, and breakup discharge was on

the order of 75,000 cfs. By the time the peak flow of 154,000, had occurred a week later, many

bridges and structures along the Kennebec had been destroyed with total damages in the millions

of dollars. The 1936 ice jam caused the record high stage at Hallowell, which is 0.5 ft higher

than peak stage caused by the 2 April 1987 open water flood of record.

In late January 1996, massive jams formed upstream of the dams at Waterville and also upstream

of Swan Island at Richmond. The Waterville jam extended 3 miles upstream to Fairfield and was

reported to be nearly 30-ft thick in places. These jams froze in place, and caused much concern

before thawing out gradually in late February.

The CRREL Ice Jam Database contains only three entries for the Sebasticook River, two of

which are ice-affected annual peak stages reported at the Pittsfield USGS Gage in 1930 and

1973. The third entry describes the destruction of the railroad bridge at Burnham on 13 March

1936. All three reported incidents occurred well upstream of the Benton Dam and the study

reach.

Mrs. Jane Edwards, a long time resident of Winslow says that the worst open water floods on the

Sebasticook at Winslow occur below the Ft. Halifax Dam, usually as the result of backwater

from the Kennebec River. Edwards reported seeing ice jams on the Sebasticook behind the

railroad bridge below the Fort Halifax Dam, but believes this ice may have come from the

Kennebec River rather than the Sebasticook. Ice does pass over the Ft. Halifax Dam nearly

every year, often damaging the 4-ft-high flashboards. The Sebasticook watershed is relatively

flat, with 16 percent lake storage, and the river typically peaks 12 to 24 hours after the

Kennebec 1 . This suggests that, by the time the ice moves out of the Sebasticook, the Kennebec

is open in the confluence area, and able to convey the Sebasticook ice downstream.

2.3 Field Observations, winter of 2003-2004

Field observations were made from the air and on the ground as a joint effort between CRREL,

E/PRO and FPLE. The USGS also provided Kennebec River ice data collected as part of their

normal operations. Reconnaissance flights were made on 19 Dec., 8 Jan., 25 Feb and 1 March.

Ice conditions were inspected on the surface on 9, 16, 21 and 26 Jan., as well as 11 and 26 Feb.

The real time USGS web camera images of the Kennebec River at Augusta supplemented the

field observations.

The 2003-2004 winter was somewhat unusual in terms of ice formation. An early winter ice

cover formed and broke up in December, causing ice jams and flooding at many locations in

northern New England. The ice cover rapidly reformed during a period of above-average flows,

extreme cold air temperatures, and abundant of open water areas during the first half of January.

By the end of February, regional ice covers were above average in extent and thickness. The

month of March was unusually dry, with consistently moderate air temperatures resulting in a

gradual thermal melt out rather than a dynamic breakup. Fig. 5 shows air temperature,

precipitation and flow data for the 2003-2004 winter.

An early December cold spell produced in 6-8-inch thick ice on the lower Sebasticook and much

of the Kennebec by mid-December. A major thaw with rain on 18 December broke up many

rivers in Northern New England, including the lower Kennebec. This event caused serious ice

jam floods in many towns including Canton, Farmington and Bethel, ME and left ice rubble and

shear walls along the banks of the Kennebec from Waterville down through Augusta. Kennebec

River discharge at North Sydney reached 80,000 cfs. Although discharge on the Sebasticook

reached 8000 cfs at the mouth, the ice on the Fort Halifax Pool remained intact to a point within

1 mile of the Benton Dam.

In early January, a combination of extreme cold, high flows and the large open water areas

produced high frazil concentrations on the Kennebec. Fig. 6 shows frazil ice pans accumulating

at Augusta on 8 Jan. By the next day, the ice had jammed at Augusta, and also 5 miles upstream

1 Discussion with Andrew Straz, E/PRO Engineering

above the Sevenmile Island Rips (Fig. 7). The ice had also jammed against work barges moored

in the openings of the new Third Bridge, about 3/4 miles upstream of the site of the former

Edwards Dam. As frazil floes continued to accumulate, the ice covers rapidly progressed

upstream of these jams. The ice edge had reached a point just downstream of the mouth of

Messalonskee Stream (2.5 miles below Waterville) by 26 Jan. and remained there through the

month of February. Open sections also remained from the former Edwards Dam site up to the

new Third Bridge, and from a point 2 miles upstream of the new bridge to the jam above the

Rips. A small ice cover also formed in the Ticonic Bay area between the Lockwood Dam

tailwater and the mouth of the Sebasticook. At the same time, ice covers formed more or less

simultaneously behind the dams at Waterville, Shawmut and Skowhegan, cutting off the frazil

ice supply from upstream.

During the same period, the sheet ice cover on the Fort Halifax Dam pool thickened and

progressed to about 2000 ft below the Benton Dam (Fig. 8). By the end of February, ice had also

formed on the Sebasticook below the Ft. Halifax Dam. Upstream of Benton, the Sebasticook was

covered in what appeared to be sheet ice to beyond Pittsfield.

On 26 Feb. 2004, 25-inch-thick thermally-grown ice was measured on the Fort Halifax Pool

about 500 ft upstream of the dam. At the upper end of the pool, at the transmission line crossing,

26.5 inches of white ice was measured, the color indicating a frazil origin. On the Kennebec

River at the site of the ice motion detector 2 about 400 ft upstream of the Memorial Bridge in

Augusta, the freezeup ice accumulation was, on average 30-inches-thick, with significant frazil

slush deposited beneath. Accessing the river by hovercraft, the USGS drilled six holes about

200 ft upstream of the CRREL measurements, finding the average thickness of hard ice to be 29

inches with about 5 ft of frazil slush deposited beneath. At the Sidney boat launch, the site of a

second ice motion detector about 7.5 miles upstream of Augusta, the ice was 24 inches thick and

white in color, indicating a frazil ice origin. This supports the field measurements and aerial

observations of the ice forming on this reach on 8-9 January.

Moderate temperatures during the month of March with little or no precipitation resulted in a

long gradual deterioration of the ice cover, with most of the ice simply melting in place. By the

morning of 26 March, the Kennebec River was open at Augusta. The iceout was so gradual that

the ice motion detectors installed at Augusta and upstream at Sydney did not even trigger.

The HEC-RAS model was used to analyze the ice formation and breakup processes within the

study reach, which is defined as the Sebasticook River from Benton Dam downstream to the

mouth, and the Kennebec River from Waterville to the Richmond-Gardiner town line. E/PRO

provided the Sebasticook River HEC-RAS geometry used in the model and the USGS provided

bathymetry data for the Kennebec from Augusta to Waterville in HEC-RAS, WSPRO and E431

input formats.

2 Ice motion detectors are switch devices installed on shore and attached by nylon cord to the river ice cover. When

3.1 Freezeup Modeling

HEC-RAS was used to calculate average channel velocities for the freezeup discharge range.

Simple water velocity criteria can help identify likely areas of thermally-grown sheet ice, frazil

ice accumulations, and areas where open water is likely to persist throughout the winter. In slow-moving

sections of river where mean channel velocity is in the 0.5 - 1.0 ft/s range, sheet ice will

typically grow out from the banks as it does on a lake. Where mean velocity is between 1.25 and

2.25 ft/s the ice cover can be expected to form as arriving floes that come to rest edge to edge at

the upstream end of the cover (juxtaposition) or tilt and stack up against the leading edge of the

ice cover (shoving).

At water velocities greater than about 2.25 ft/s, or Froude numbers 3 greater than about 0.1,

arriving floes can be expected to be entrained by the current beneath the leading edge of the ice

cover and either be carried past or deposited beneath. This frazil thickening, or hanging dam

formation resists flow, progressively increasing upstream stage and slowing the ice approach

velocity. Velocity may decrease to the to the point where arriving floes are no longer entrained

beneath the ice and again begin to accumulate on the surface.

Fig. 9 shows a HEC-RAS-calculated water surface profile, with average water velocities

computed for the Kennebec River at the average December-January discharge of 6763 cfs. A

mid-tide water surface elevation of 2.8 ft was used as the downstream boundary. The calculated

water velocities are consistent with the field-observed ice formation of the 2003-2004 winter. In

only a few places upstream of Augusta is the water velocity below the 1.0 ft/s sec upper limit for

thermally grown sheet ice. One would therefore expect the ice cover to form predominantly

through accumulation of frazil floes, as was observed on 8-9 Jan. 2004. Note that calculated

water velocity is low at the second ice jamming area observed upstream of the Sevenmile Island

Rips. In sections of river where open leads were observed throughout the winter, such as the

Sevenmile Island Rips and below, the calculated velocity was at least 3 ft/s.

Fig. 10 shows water surface profiles and calculated average open water velocities for the Fort

Halifax Pool on the Sebasticook, before and after the proposed dam breaching. In the existing

conditions case, water velocity is well within the expected range for thermal ice growth in all but

uppermost 3000 ft below the Benton Dam. In the winter of 2003-2004 the observed frazil ice

portion of the cover was longer, due probably to the above-average early winter discharges (Fig.

5). In the post breach case, the calculated velocities indicate that there will be a short slow

section in the vicinity of station 3500 and a longer pool between stations 6500 and 13,500. One

would expect thermal ice growth on these pools to block downstream movement of frazil ice,

and promote the upstream progression of a frazil ice cover. Once an ice cover had formed on

this pool, the amount of frazil ice entering the Kennebec River would probably not be

significantly greater than it is with the dam in place.

Thermal ice growth and the distribution of maximum estimated ice thickness (ti) were calculated,

based on the long-term record of maximum accumulated freezing degree-days (max AFDD).

3 The Froude number F is defined as v F gh

=where v is average channel velocity, h is depth and g is

max i t C AFDD =

Where C is a coefficient usually in the 0.3 to 0.6 range. Figure 5 shows the maximum

accumulated freezing degree-days for the 2003-2004 winter leveling off after 28 Feb at about

1100 (Fig. 5). Compared to the 47-year period of record at the Augusta Airport, the 2003-2004

winter was about average in terms of coldness and ice thickness (Fig. 11). Based on the 25 Feb.

2004 measured ice thickness of 24 inches on the lower Fort Halifax Pool, a coefficient C of 0.7

was used to estimate thermal ice growth.

Severe winter freezeup ice profiles and ice volumes were calculated for the existing conditions

and post-breaching cases on the Sebasticook and also for the Kennebec River from Waterville

downstream to the Richmond-Gardiner town line. Based on the 2003-2004 field observations of

the Kennebec River and the HEC-RAS water velocity calculations, the ice cover forms primarily

due to frazil accumulation, and, to a lesser extent, thermal ice growth. Initially, the HEC-RAS

model was used to calculate freezeup ice accumulation profiles. In the slow moving sections of

river, we assumed thermal ice growth would predominate, with a maximum ice cover thickness

of about 2 ft.

Further ice thickening due to frazil deposition beneath the initial freezeup ice cover was

estimated, using the methods outlined by White and Acone (1998) and Shen and Wang (1995).

This theory defines the conditions for frazil stability beneath the ice cover in terms of a

dimensionless ice transport capacity vs. dimensionless flow strength curve. As ice discharge and

ice concentration increase, deposition is favored, while an increase in flow strength and under ice

water shear encourages under ice erosion and frazil transport. In this study, we adjusted the

thickness of the deposited frazil ice beneath the initial freezeup ice cover to achieve a theoretical

equilibrium condition between frazil deposition and erosion.

Figure 12 shows a hypothetical maximum freezeup ice cover for the Kennebec River from

Gardiner to Waterville. The dotted blue line traces the bottom of the initial freezeup ice

accumulation calculated by HEC-RAS. Important ice parameters used are an ice accumulation

internal strength angle Ö = 45 º, a Mannings roughness for the ice underside nice = 0.03, an ice

accumulation porosity e = 0.4, and a maximum non-eroding, under ice water velocity veros = 3

ft/s. USACE (1999) Fig. 4.2 provides typical roughness ranges for different types of ice

accumulations and Liu and Shen (2000) Table 1 gives typical Ö and e values 4 . The dashed red

line shows areas of possible under-ice frazil deposition. The theoretical frazil slush depth

compares fairly well to the USGS measurements of 26 Feb. 2004 at Augusta. The deposition

theory applies less well 7.5 miles upstream of Augusta where 8-ft-thick frazil slush was

predicted beneath the 2-ft-thick freezeup ice cover, but no frazil slush was actually observed. A

possible explanation is that the initial freezeup ice cover progressed rapidly upstream, quickly

reducing the open water frazil-producing area. The open water sections upstream of the former

Edwards Dam and below the Rips are included in the modeled profile. Downstream of Augusta,

the cover was assumed to be 2-ft thick sheet ice. Fig. 13 shows cumulative ice volumes for the

freezeup ice cover depicted in Fig. 12, with and without the frazil deposition beneath.

4 Note that porosity e = 1- c, where c is ice accumulation maximum concentration.

By the same process, maximum freezeup ice covers were developed for the Fort Halifax Dam

Pool on the Sebasticook for the pre and post breach cases (Fig. 14). The existing conditions case

is mostly thermal sheet ice with some frazil deposition predicted near the upstream end. For the

post breach case, the HEC-RAS results predict a shove-thickened frazil ice accumulation

upstream of a thermal ice covered pool section. The dashed red line shows the area of possible

frazil slush deposition. Figure 15 shows cumulative ice volumes for the Fig. 14 cases. An

interesting result is that although the post-breach ice cover is much thicker, the cumulative ice

volume of 1.51 x 10 7 ft 3 is about half the existing conditions case ice volume of 3.0 x 10 7 ft 3 as a

result of the reduced channel width.

3.2 Breakup Modeling

The range of breakup discharges was defined based on the history of severe ice breakups on the

Sebasticook and Kennebec Rivers. Daily average discharge data are available for USGS

Waterville Gage (01048500) for 1893-1935, and annual peak data are recorded for 1893-1954.

The record of daily flows resumes in 1978 at the North Sydney gage (01049265), with a gap

from 1993 to 2000 when the gage was moved upstream to a site near the Waterville Wastewater

Treatment Plant (01049205).

Unfortunately, daily discharge data for the lower Kennebec River (Waterville and North Sydney)

are unavailable from 1955 to 1978, a period during which a number of historic ice jams

occurred. Upstream at the Bingham gage record for the Kennebec goes back to 1907, but, low to

intermediate discharges do not transpose well downstream, due to regulation at intermediate

dams. The Pittsfield gage on the Sebasticook has a good record back to 1928, and these flows

multiplied by 1.654 were used to estimate discharge at the mouth 5 . Daily discharges

surrounding the March 1936 ice jam flood were estimated by comparison to flows upstream at

Bingham, and on the Androscoggin at Auburn. (Fig. 16).

In general, the thicker the ice, the greater the flow required to break it up and move it

downstream. A rapid time-to-peak also contributes to ice breakup severity. Because 1936 was a

severe winter, one can assume that the ice on the Kennebec was at least 2-ft-thick. Although the

flow eventually exceeded 150,000 cfs, estimated discharge on March 13th, the day of breakup,

was about 75,000 cfs. This breakup had sufficient energy to transport the Kennebec ice from

below Waterville past the Edwards Dam at Augusta and into a jam at Browns Island 2 miles

below Hallowell. The ice supply to this jam was limited by an upstream jam reported in the

vicinity of Hinkley. Ice from below Browns Island ran and jammed at the head of Swans Island

at Richmond. The rise on the Sebasticook River during the 1936 event was more gradual,

peaking at about 20,000 cfs a few days later than the Kennebec River. It is not know if or when

the ice ran out of the Sebasticook in 1936.

On 21 Jan. 1996, a one-day rise in discharge to 47,000 cfs broke up the 1-ft-thick ice cover,

forming jams at Fairfield and Farmingdale. Much of the reported ice jam flooding corresponds

with the following peak to 59,000 cfs that occurred on 27 Jan. Based on this limited

information, the discharge range for severe breakups on the Kennebec is estimated to be about

50,000 to 80,000 cfs. A representative value of 65,000 cfs was used in the HEC-RAS breakup

ice jam modeling.

5 Discussion with Andrew Straz of E/PRO

Using HEC-RAS, severe winter breakup ice jam profiles were calculated for the Kennebec River

from Gardiner upstream to Waterville. Based on review of historical breakups, it was assumed

that the upstream limits of the ice supply reaches were the Benton Dam on the Sebasticook

River, and the Lockwood Dam at Waterville for the Kennebec River. Pre-breakup ice volumes

were estimated from the maximum freezeup ice profiles (Figs. 12-14). Important ice parameters

used in the breakup modeling were an ice accumulation internal strength Ö = 45 º, a Mannings

roughness for the ice underside nice = 0.04, an ice accumulation porosity e = 0.4, and a

maximum non-eroding under ice water velocity veros = 5 ft/s. (USACE ,1999, and Liu and Shen,

2000). Transport losses due to shear wall formation, overbank deposition, and melting were

assumed to be 50 %. Because jam location is somewhat uncertain, three locations were modeled:

the first at Nehumkeag Island (4.5 miles below Gardiner) the second at Hallowell and the third at

the Memorial Bridge in Augusta. Fig 17 shows the resulting breakup ice jam profiles.

Under existing conditions, the ice cover on the Sebasticook appears quite resistant to breakup as

shown in the thaw of 18 December 2003, which broke up the Kennebec ice but left most of the

ice cover on the Fort Halifax Dam Pool intact. The ice on the post-breach Sebasticook would

probably be more prone to breakup and run into the Kennebec than under existing conditions

because of the lower water storage capacity. However, two factors would postpone the ice

release. The first is the sheet ice on the pool between stations 6500 and 13,500, and the second

is the sharp bend to the right and channel constriction near the confluence with Outlet Stream.

Either or both of these could slow or stop ice passage from the Fort Halifax Dam pool.

As a worst-case post Fort Halifax Dam breaching scenario, the entire ice volume of the

Sebasticook below the Benton was added to a simulated breakup ice jam on the Kennebec at

Hallowell (Fig. 18). The resulting water surface profile was not significantly higher than the

simulated jam without the Sebasticook ice. Proportionally, the post-breach Sebasticook ice

volume below the Benton Dam of 1.5 x 10 7 ft 3 (Fig. 15) represents about 6 percent of the

maximum estimated ice volume on the Kennebec River between Richmond and Waterville of 2.5

x 10 8 ft 3 .

Based on the ice-hydraulic modeling done in this study, the proposed changes to the Ft. Halifax

Dam will not adversely affect ice processes on the lower Sebasticook and Kennebec Rivers, and

are not likely to significantly increase the ice jam flood potential on the lower Kennebec River.

Following breaching, probably a small portion of the frazil ice that is now retained upstream of

the Ft. Halifax Dam will pass into the Kennebec River in early winter. A thermal ice cover is

expected to form on a mile-long pool section about 4000 ft upstream of the dam. Once formed,

this ice cover will intercept additional frazil floes, and the frazil ice discharge into the Kennebec

River will be small. Total late season ice volume on the post-breach Fort Halifax Dam Pool is

expected to be about half that with the dam in place, due to the reduced ice surface area.

Based on the winter 2003-2004 observations, freezeup ice covers initiate at Augusta and in the

slower moving water upstream of the Sevenmile Island Rips. Any post-breach additional frazil

ice escaping from the Sebasticook River will most likely contribute to the upper ice cover on the

Kennebec rather than the freezeup jam that forms in Augusta.

Following breaching, the ledge outcrops, frazil-thickened freezeup ice covers, and channel bends

and constrictions will help delay the breakup ice run on the lower Sebasticook River. In the

event that all the Sebasticook River ice below the Benton Dam runs into the Kennebec, the ice-hydraulic

models indicate that added ice volume in the Kennebec River jams will be relatively

small (about 6 percent) and the resulting stage increase should not be significant.

White, K. D. and S.E. Acone (1998) Ice Cover Thickening at River-Reservoir Confluences: A

case Study. Proceedings, ASCE Cold Regions Engineering Conference, Duluth, MN September,

1998

White, K. D. and J. N. Moore (2002) Impacts of Dam Removal on Riverine Ice Regime. ASCE

Journal of Cold Regions Engineering, Vol. 16, No. 1, March 2002.

Shen, H. T. and D. H. Wang (1995) Under cover transport and accumulation of frazil granules.

Journal of Hydraulic Engineering, 121(2): 148-195.

Liu, L. and H. T. Shen (2000) Numerical Simulation of River Ice Control with Booms.

ERDC/CRREL Technical Report TR-00-10.

http://www.crrel.usace.army.mil/techpub/CRREL_Reports/reports/TR00-10.pdf

USACE (1999) Ice Engineering Manual. EM 1110-2-1612.

http://www.usace.army.mil/inet/usace-docs/eng-manuals/em1110-2-1612/c-4.pdf

Get the facts on the dam.

Opportunities for Public Input on Fort Halifax Dam Fisheries Restoration Plans and/or Dam Licensing Decisions.

TU/Kennebec Coalition Filings at FERC  Regarding Fort Halifax Dam

Summary of FERC and ACOE related activity during 2004 Pertaining to the Fort Halifax Project No. 2552-065

ORDER DENYING REHEARING AND STAY