1

Quantifying the Impact of Hydropower Operations on Shoreline Erosion throughout the

Turners Falls Impoundment (Connecticut River)

By Timothy Sullivan, Regulatory Specialist, Gomez and Sullivan Engineers, Henniker, NH;

Dr. Andrew Simon, P.E., Fluvial Geomorphologist, Cardno, Oxford, MS; and

Dr. Robert Simons, P.E., Fluvial Geomorphologist, Simons & Associates, Midway, UT

ABSTRACT

As part of the Federal Energy Regulatory Commission relicensing of the Turners Falls

Hydroelectric Project and the Northfield Mountain Pumped Storage Project, a multi-year study

was conducted to identify, evaluate, and quantify the causes of erosion throughout the Turners

Falls Impoundment (Connecticut River, which serves as the lower reservoir for the pumped

storage project) and determine the extent to which erosion was related to hydropower operations.

To achieve the goals of the study a wide variety of field data collection efforts, data analyses, and

computational modeling were conducted. HEC-RAS and River2D hydraulic models were

developed to analyze the complex hydrologic and hydraulic characteristics of the Turners Falls

Impoundment as well as the location, duration, and magnitude of hydraulic forces associated with

erosion. The results of the HEC-RAS model, combined with a wide array of field collected data,

were then used as input parameters for the Bank Stability and Toe Erosion Model (BSTEM).

Site-specific BSTEM results from twenty-five detailed study sites along the Turner Falls

Impoundment riverbank were then extrapolated throughout the entire Impoundment to identify the

dominant and contributing causes of erosion at each riverbank segment and to determine the

impact hydropower operations have on erosion, if any. The results of the study found that naturally

occurring high flows were the dominant cause of erosion at almost all riverbank segments with

boat waves also having an impact in the lower portion of the study area.

INTRODUCTION

The Northfield Mountain Pumped Storage Project (Northfield Mountain) (FERC No. 2485) and

the Turners Falls Hydroelectric Project (Turners Falls) (FERC No. 1889), collectively referred to

as the Project, are located on the Connecticut River in the towns of Montague and Erving, MA,

respectively. The Turners Falls Dam creates the approximately twenty-mile-long Turners Falls

Impoundment (Impoundment), which also serves as the lower reservoir and tailrace for Northfield

Mountain. The Turners Falls Dam includes two hydroelectric projects operating from a canal

including Station No. 1 and Cabot Station. The Vernon Dam, which is part of the Vernon

Hydroelectric Project (Vernon) (FERC No. 1904) and located in Vernon, VT, is located at the

upstream extent of the Impoundment. The Turners Falls and Northfield Mountain Projects are

owned and operated by FirstLight Hydro Generating Company (FirstLight) and are currently

licensed with the Federal Energy Regulatory Commission (FERC or the Commission). The

licenses for Turners Falls and Northfield Mountain were issued on May 14, 1968 and May 5, 1980,

respectively, with both set to expire on April 30, 2018. FirstLight has initiated the process of

relicensing the Project with the Commission using FERC’s Integrated Licensing Process. The

figure found on page 2 depicts the Impoundment from the Turners Falls Dam to Vernon Dam.

2

Streambank erosion throughout the Impoundment has been a contentious issue for decades. Since

operation of the Northfield Mountain Project commenced in 1972, local stakeholders have

contended that water level

fluctuations associated with

the Project’s pumping and

generation cycles have been

the cause of erosion

throughout the

Impoundment. Under the

current license, FirstLight is

responsible for remediating

all erosion in the

Impoundment, regardless of

cause. Due to the history of

this resource issue, Simons &

Associates, Cardno, and

Gomez and Sullivan

Engineers, DPC1

(collectively referred to as the

Study Team), conducted the

multi-year Northfield

Mountain/Turners Falls

Operations Impact on

Existing Erosion and

Potential Bank Instability

relicensing study (the study).

The results from this study

were used to identify,

evaluate, and quantify the

causes of erosion throughout

the Impoundment and to determine the extent to which hydropower operations impact erosion, if

at all.

METHODOLOGY

To achieve the study goals a wide variety of field data collection efforts, data analyses, and

computational modeling were conducted throughout the study area. The study area included the

entire Impoundment from Vernon Dam to Turners Falls Dam; the study period encompassed 2000-

2014. Five potential primary causes of erosion were examined in-depth including: (1) hydraulic

shear stress due to flowing water; (2) water level fluctuations due to hydropower operations; (3)

boat waves; (4) land management practices; and (5) ice. Secondary causes of erosion such as

1 Additional study partners included The National Center for Computational Hydroscience at the University of

Mississippi, New England Environmental, and Dr. Kit Choi, P.E.

3

animal activity; freeze-thaw; wind waves; and seepage and piping were thought to have minimal

to no influence on erosion in the Impoundment except in specific, localized areas where they may

occur. As such, these potential secondary causes were not examined in-depth.

An extensive combination of historic data from the 1990s and 2000s combined with newly

collected field data provided the foundation from which all analyses and modeling were conducted.

Datasets for this study included

hydraulic and hydrologic data;

geomorphic and riverbank

geotechnical data; Project

operations data; and cross-

sectional and bathymetric

survey data. Given the vast

study area (i.e., over forty miles

of riverbank), twenty-five

detailed riverbank study sites

were established throughout

the Impoundment with in-

depth field data collection,

investigation, and modeling

occurring at each site. The

detailed study sites (shown in

the adjacent figure) spanned

the geographic extent of the

Impoundment, included a

representative range of

riverbank features,

characteristics, hydraulic, and

erosion conditions, included

both non-restored and restored

sites, and were located at a

combination of existing

transects (surveyed annually

since the late 1990s) and newly

identified sites. Detailed study sites located at existing transects provided an opportunity to

calibrate the Bank Stability and Toe Erosion Model (BSTEM) with actual erosion amounts or

changes in bank geometry as it occurred over time.

Prior to evaluating the causes of erosion, a 1-dimensional unsteady HEC-RAS model and 2-

dimensional River2D hydraulic model were developed to analyze the complex hydraulic and

hydrologic characteristics of the Impoundment as well as their associated impact on erosion

processes. The hydraulic model results provided valuable insight into the hydraulic forces

associated with erosion and were used for several analyses including delineation of hydraulic

reaches as well as shear stress, water level duration, stage-discharge, and flow exceedance

4

analyses. The outputs from the HEC-RAS model (specifically hourly Energy Grade Line Slope

(EGL slope) and water surface elevation), combined with field collected data, were used as input

parameters for BSTEM. BSTEM was run on an hourly timestep at each of the twenty-five discrete

detailed riverbank study sites located throughout the Impoundment over the period 2000-2014. In

the event that a study site had been previously restored, BSTEM was run for both the pre-

restoration and post-restoration condition.

The site-specific BSTEM results, combined with the results of the hydraulic models and other

supplemental analyses, were used to quantify the dominant and contributing causes of erosion at

each detailed study site. Once the causes of erosion were determined for each site, an

extrapoloation approach was utilized to assign the cause(s) of erosion for each riverbank segment2

throughout the Impoundment.

RESULTS

Although a wide array of data analyses and computational modeling were conducted for this study,

discussion presented in this section focuses primarily on the results of the hydraulic and BSTEM

modeling. The results of the hydraulic modeling provided valuable insight into the hydraulic

forces associated with erosion and the influence that hydropower operations have on the complex

hydraulics in the study area. Site-specific BSTEM results quantified the amount of bank erosion

which occurred over the course of the study period and identified dominant and contributing causes

of erosion at each site. Results from the hydraulic model analyses and BSTEM are discussed in

greater detail below.

Hydraulic Modeling Results

The results of the hydraulic modeling were used for several analyses to better understand the

complex hydrologic and hydraulic characteristics of the study area and to determine the extent to

which hydropower operations impact the hydraulic forces associated with erosion. Analyses

pertaining to flow, water level, and shear stress as well as delineation of hydraulic reaches were

conducted. An overview of each of these analyses is presented below.

Delineation of hydraulic reaches: Review of the hydraulic model results, and more specifically

the EGL slope, revealed four distinct hydraulic reaches within the Impoundment including: the

Upper (Reach 4), Middle (Reach 3), Northfield Mountain (Reach 2), and Lower (Reach 1) reaches

(figures page 5 and 10). The delineation of hydraulic reaches was significant in that the results of

the hydraulic and BSTEM models indicated that hydropower operations can only potentially

impact erosion processes within the hydraulic reach where a given project is located due to the

varying hydraulic characteristics of the Impoundment. In other words, the models showed that

Vernon operations can only potentially impact erosion processes at riverbank segments in Reach

4, Northfield Mountain operations can only potentially impact erosion processes in Reach 2, and

Turners Falls operations can only potentially impact erosion processes in Reach 1. Hydropower

operations do not impact erosion processes at all in Reach 3. Although hydropower operations

can impact flows and water levels beyond their given hydraulic reach, the impacts at flows which

2 Riverbank segments were delineated during an earlier FERC relicensing study (i.e., the 2013 Full River

Reconnaissance survey).

5

cause erosion (as determined by BSTEM) are minor enough that they do not alter the EGL slope,

and therefore the velocity or shear stress, outside of their reach.

Flow Analysis: The models assessed the erosive impacts of flows within thresholds established

by the hydraulic characteristics of each reach. Through further analysis of the various modeling

results two flow thresholds were established in the Upper reach of the Impoundment (Reach 4):

(1) <17,130 cubic feet per second (cfs), and (2) >17,130 cfs. This threshold value was identified

as it corresponds with the total turbine hydraulic capacity of the Vernon Hydroelectric Project and

is consistent with the hydraulic characteristics of this more riverine reach of the Impoundment.

In the remaining three hydraulic reaches, three flow thresholds were established, including: (1)

Low Flow (<17,130 cfs), during these periods flows and water levels are controlled by

hydroelectric operations; (2) Moderate Flow (17,130 to 37,000 cfs), during these periods several

hydraulic influences are observed including hydropower operations, Turners Falls Dam water level

management, and a natural hydraulic control (i.e., the French King Gorge3 at flows greater than

30,000 cfs); and (3) High Flow (>37,000 cfs). 37,000 cfs was chosen as the high flow threshold

as it represents the combined hydraulic capacity of Vernon and Northfield Mountain, is at a flow

above which the French King Gorge becomes the hydraulic control for the mid and upper portions

of the Impoundment (see figure on page 3), and represents periods when Northfield Mountain

operates less frequently. Flows greater than 37,000 cfs are considered naturally occurring.

The establishment of flow thresholds was vital when determining the impact Project operations

may have on erosion processes. Given that BSTEM was run on an hourly timestep for the full

study period, the site-specific BSTEM results quantified the amount of erosion which occurred at

every flow and determined the flow at which the majority of erosion occurred at a given site. Flow

analyses were run using the BSTEM results to determine the erosion flow threshold at which 50%

and 95% of all erosion occurred at a given site. Based on the results of this analysis, and the

established flow thresholds, a determination was made as to the sites where natural moderate or

3 The French King Gorge is located in Reach 2; the river width narrows considerably here causing backwatering to

occur in Reaches 2, 3, and 4 at flows greater than approximately 30,000 cfs.

6

high flows were found to be a dominant or contributing cause of erosion. The results of the HEC-

RAS model were then used to determine the percent of time flows at which the majority of erosion

occurred. By delineating hydraulic reaches and establishing flow thresholds, the Study Team

could determine erosion caused by natural flows versus erosion caused by hydropower operations.

Water Level Analysis: Hydraulic erosion processes associated with hydraulic shear stress due to

flowing water; water level fluctuations due to hydropower operations; boat waves; and ice occur

at or below the water surface. As such, it is vital to understand where on the riverbank the water

surface rests and for what duration.

Impoundment riverbanks are typically

characterized by a clearly defined lower

and upper bank (see adjacent figures). The

lower bank is typically a relatively flat,

beach-like feature that is submerged or

experiences daily water level fluctuations

during low to moderate flows because of

hydropower peaking operations. As one

moves up-gradient from the normal edge-

of-water, the lower bank transitions to an

upper bank; the toe of which is clearly

identifiable. The upper bank is typically

steep, has some degree of vegetation, and is

usually above the water surface except

during high flows. The distinction between

the upper and lower bank is important as

the vast majority of erosion in the

Impoundment occurs when the water

surface reaches the upper bank.

To determine the amount of time the water

surface rests on the upper bank, and

therefore the amount of time the vast

majority of erosion occurs, a water level

duration analysis was conducted at a representative subset of the twenty-five detailed study sites.

The water level duration analysis entailed using the HEC-RAS results to develop stage-discharge

relationships at the representative sites to determine the flow at which the water level exceeds the

toe of the upper bank (toe-of-bank elevations were derived from cross-section survey data). Flow

duration curves were then developed and examined to determine the percent of time flows which

resulted in a water surface elevation which exceeded the toe of the upper bank could have occurred.

The results of the stage-discharge analysis were then compared against the results of the BSTEM

flow analysis which examined the flows at which 50% and 95% of all erosion occurred at a given

site.

7

The results of this analysis found that the water level rested on the lower bank, where minimal to

no erosion occurs, the vast majority of the time (i.e., 78-90% of the time during the study period).

The flow required for the water surface to exceed the toe of the upper bank was generally found

to be within the moderate flow range (i.e., 17,130 cfs to 37,000 cfs); however, the flows at which

95% of all erosion occurred were found to be close to, or greater than, the natural high flow

threshold (37,000 cfs) at almost all sites. The percent of time which the 95% erosion flow

threshold occurred at these sites ranged from 3%-7%, with the corresponding water surface

elevations well above the toe of the upper bank. In other words, as observed from the models,

minimal to no erosion occurred during low flow periods when river flows are less than the

hydraulic capacity of the hydroelectric projects. It was not until the water surface reached the

upper bank during naturally occurring moderate to high flows that the vast majority of erosion

occurred.

The results of this analysis demonstrated the minimal impact which hydropower operations have

on erosion in the Impoundment by determining the location, duration, and magnitude of the

hydraulic forces associated with erosion processes.

Shear Stress: The results of the River2D hydraulic model were used to evaluate velocity and shear

stress in the near-bank area at each of the detailed study sites as well as other areas of interest (i.e.,

areas with unique hydraulic conditions such as eddies). Given that BSTEM determines the

boundary shear stress along each node of the wetted perimeter, the River2D analysis was

conducted as a supplemental analysis to verify and confirm the BSTEM findings.

The River2D analysis utilized the results of six steady-state simulations which examined a range

of conditions, including normal operating conditions, commonly occurring flows that might occur

every few years, and more extreme events including the 100-year flood. Boundary shear stress

values derived from the River2D model were compared against the critical shear stress at each site

which was derived from field collected data. The results of this analysis found that the hydraulic

shear stress is only sufficient to cause erosion when flows at Turners Falls Dam exceed

approximately 30,000 to 65,000 cfs, and may be insufficient to cause erosion at approximately half

of the detailed study sites under a 100-year return period event (i.e., 157,700 cfs).

BSTEM Results

BSTEM is a state-of-the-science deterministic model that simulates hydraulic and geotechnical

erosion processes responsible for bank erosion, including the effects of vegetation, pore-water

pressure, and the confining forces due to flow in the channel. BSTEM was the principal tool used

to evaluate the potential primary causes of erosion including hydraulic shear stress due to flowing

water, water level fluctuations, and boat waves. The remaining potential primary causes of erosion

(i.e., ice and land management practices) were evaluated through separate, supplemental analyses.

Multiple hourly BSTEM runs for the full study period (2000-2014) were executed at each detailed

study site and included the Baseline Condition, which represented observed conditions during the

model period, and Scenario 1, which modeled Northfield Mountain as being idle (no pumping or

generation) for the full model period. In addition, model runs for both the Baseline Condition and

Scenario 1 were run with the boat wave module “turned on” and “turned off” to determine the

impact of boat waves. If a detailed study site had previously been restored, BSTEM was run for

8

both the pre-restoration and post-restoration condition. BSTEM outputs quantified the total

erosion under the Baseline Condition in terms of ft3/ft/yr and determined the erosion flow

thresholds at which 50% and 95% of all erosion occurred at each detailed study site. Site specific

BSTEM results are presented in the figures below. The top figure depicts the average annual rate

of erosion at each detailed study site, while the bottom figure depicts the discharge at which 95%

of erosion occurs.

To determine the impact hydropower operations had on erosion and bank instability, the potential

primary causes of erosion were further broken down into the following categories: (1) natural

moderate or high flows; (2) boat waves; (3) Vernon operations; (4) Northfield Mountain

operations; and (5) Turners Falls operations. Table 1 provides an overview of how each primary

cause of erosion was determined. Based on these classifications dominant and contributing causes

of erosion were identified at each detailed study site. For a cause to be considered Dominant, it

needed to be responsible for at least 50% of the erosion at a given site, as determined by the

modeling results. Conversely, for a cause to be considered Contributing, it must be responsible

for at least 5% but less than 50% of the bank erosion at a given site. Causes responsible for less

9

than 5% of the erosion at a given site were considered immeasurable and were within the accuracy

of the analysis. The site-specific BSTEM results identified the bank erosion rate, dominant cause,

contributing cause(s), contributing factors, and contributing processes for each detailed study.

Table 1 - Determination of the Primary Causes of Erosion in the Turners Falls Impoundment

Primary

Cause Description

Moderate or

High Flows

Hydraulic shear stress due to flowing water. A flow analysis was conducted

which resulted in identification of the erosion flow threshold at which 50% and

95% of all erosion occurred at a given site. Based on the results of this analysis,

and the flow thresholds discussed earlier, a determination was then made as to

the sites where natural moderate or high flows were found to be a dominant or

contributing cause of erosion.

Boat waves

BSTEM was enhanced with a built-in boat wave module for this study. Two

BSTEM runs were executed utilizing this module, one with boat waves “turned

on” and the other with boat waves “turned off”. The difference in observed

erosion between the two model runs determined the sites where boat waves were

a cause of erosion.

Vernon

operations

Hydraulic shear stress due to flowing water, water level fluctuations associated

with hydropower operations. The results of the flow analysis were used to

identify areas within the Upper hydraulic reach where erosion was observed at

flows below 17,130 cfs.

Northfield

Mountain

operations

Hydraulic shear stress due to flowing water, water level fluctuations associated

with hydropower operations. Two BSTEM runs were executed, one

representing Baseline Conditions and one representing Northfield Mountain as

idle. The difference in observed erosion between the two model runs determined

the sites where Northfield Mountain operations were a cause of erosion.

Turners

Falls

operations

Hydraulic shear stress due to flowing water, water level fluctuations associated

with hydropower operations. Due to the hydraulic characteristics of the Lower

hydraulic reach (i.e., lake-like downstream portion (Barton Cove) and a riverine

upstream portion), a combination of site-specific BSTEM results, geomorphic

assessment, and hydraulic model analysis were used to determine the causes of

erosion in the Lower hydraulic reach and the impact, if any, of Turners Falls

operations.

SUMMARY EVALUATION OF THE CAUSES OF EROSION

After determining the dominant and contributing causes of erosion at each detailed study site, the

BSTEM results, combined with the results of the supplemental analyses conducted for this study,

were extrapolated across the Impoundment to over 593 riverbank segments. The extrapolation

process was a multi-step process that included analysis of the riverbank features, characteristics,

10

and erosion conditions at each segment, the variability of hydraulic forces throughout the

Impoundment, and the adjacent land-use. The result of this task was the quantification, based on

relative percentages, of the dominant and contributing causes of erosion at each detailed study site

and the Impoundment overall.

The results of the extrapolation process found that naturally occurring high flows were the

dominant cause of erosion in the Impoundment at 78% of the riverbank segments (~33 miles)

followed by boat waves at 13%

(~6 miles). Northfield

Mountain or Turners Falls

operations were not found to

be a dominant cause of erosion

at any riverbank segment in the

Impoundment. Analysis of the

contributing causes of erosion,

found that the majority of

riverbank segments in the

Impoundment did not have a

contributing cause of erosion

(68% of the riverbank

segments or ~29 miles) given

that natural high flows were

such a dominant factor in

erosion processes. At

riverbank segments that did

have contributing causes of

erosion, boat waves were

found to be the most common

(16% or ~7 miles) followed by

naturally occurring moderate

flows (10% or ~4 miles),

natural high flows (9% or ~4

miles), and Northfield

Mountain operations (4% or

~1.5 miles)4. Turners Falls or

Vernon operations were not found to be a contributing cause of erosion at any riverbank segment

in the Impoundment. The spatial distribution of the causes of erosion are presented in the adjacent

figure.

4 Note that since moderate flows and boat waves are contributing causes of erosion at a number of the same riverbank

segments, the total percentage for contributing causes does not equal 100%. In other words, given that a riverbank

segment can have more than one contributing cause of erosion, the percentages do not add to 100%.

11

CONCLUSIONS

The Northfield Mountain/Turners Falls Operations Impact on Existing Erosion and Potential

Bank Instability Study successfully utilized state-of-the-science technology and a wide array of

existing and newly collected data, data analyses, and computational modeling to identify and

quantify the causes of erosion at 593 riverbank segments spanning over forty miles of shoreline.

In addition, the study successfully quantified and identified the locations where hydropower

operations are a dominant or contributing cause of erosion.

The two primary tools used to conduct this study (1-dimensional unsteady HEC-RAS model and

BSTEM) provided a comprehensive analysis of the causes of erosion, and forces associated with

them, in the study area. Results of the hydraulic model analyses were vital in identifying the

hydraulic reaches where a given hydroelectric project could or could not potentially impact erosion

processes; identifying flow thresholds that were vital in determining the causes of erosion and the

impact of hydropower operations; and determining the location, duration, and magnitude of

hydraulic forces associated with erosion. Site-specific BSTEM results determined bank erosion

rates and erosion flow thresholds as well as dominant and contributing causes of erosion,

contributing factors, and contributing processes at each detailed study site.

The successful execution of the study provides a template from which other erosion causation

studies can follow to evaluate and identify the causes of erosion and determine the extent to which

hydropower operations impact erosion processes.

AUTHORS

Timothy Sullivan, GISP

Mr. Sullivan is a regulatory specialist with Gomez and Sullivan Engineers, DPC with experience

related to both traditional and pumped storage hydroelectric projects. He has served as the Project

Manager or technical lead for several erosion and sediment transport studies in the Northeast and

Mid-Atlantic United States and has experience in geomorphology; sediment transport; shoreline

erosion; hydrology; and hydraulics, including HEC-RAS modeling.

Andrew Simon, PhD, PE

Dr. Andrew Simon is an internationally recognized geomorphologist at Cardno in Oxford, MS.

He has 35 years of research experience which includes 16 years with the U.S. Geological Survey

and 16 years at the USDA-Agricultural Research Service, National Sedimentation Laboratory. He

is the author of more than 100 technical publications, has edited several books and journals and is

the senior developer of BSTEM.

Robert Simons, PhD, PE

Dr. R. K. Simons of Simons & Associates has extensive experience on hundreds of projects

covering various aspects of civil engineering focusing on the interaction and effect of projects on

watersheds, rivers, and estuaries related to changing hydrology, hydraulics, fluvial

geomorphology, sediment transport, erosion and sedimentation, flooding, and channel

stabilization.