Connecticut River Meanders Study-1985Connecticut River Meanders at Northampton and
Northfield: A Comparision of Geomorphology and R cent
Changes in Bank Locations
A Report to the City Planner
City of Northampton, Ivlassachusetts
By
_John B. Reid jr.
Associate Professor of Geology
and
Students of Evolution of the Earth (NS 1 07)
Hampshire College
Amherst, Massachusetts
It4
°
CONNECTICUT RIVER MEANDERS AT NORTHAMPTON AND NORTHFIELD, MASSACHUSETTS:
A comparison of their geomorphology and changes in bank location for the
period 1939-
John B. Reid. Jr., Associate Professor of Geology, and students of the
class, Evolution of the Earth (NS107), Hampshire College, Amherst, MA,
01002
September 23, 1985
Purpose of the Report
About one mile below the Coolidge Bri.dge in Northampton, the right bank
of the Connecticut River is retreating southward at about 10 feet per
year. The city of Northampton has accepted proposals for bank
stabilization that involve the use of automobile tires strapped together
and chinked with stones as bankreinforement. The method was tested by
the V.S. Army Corps of Engineers in 1982 on a bend at Northfield,
Massachusetts; today, vegetation and sediment coveys the exposed portion
of the tire blanket at Northfield and as a result the teSt has been
deem - d successful. In this report we document the changes in bank
�� location (determined from historic aerial photographs and field surveys
made in September, 1985) along with bottom topography for the two
meanders. The intent of this document is to assist the city planner in
deciding the best course of action concerning the erosion problem.
Specificall we conclude the following:
(1) The Sandy Beach bend has retreated southward about 490 feet since
'1939, while photographs and recent surveying show that the bend at
Northfield has undergone no measurable outerbank retreat in the same
interval. As a result, the apparent success of the tire-blanket may not
be r!sal, and the test results inapplicable to an eroding bend like the
Sandy Beach bend, In recent geologic time (the last several centuries)
the Northfield meander has been the siteof deppsitibn on the outer
concave bank. This anomalous behavior is exhibited by a number of other
xea//del.s on the CunueLti River (at Hatfield, Hadley, Easthampton and
Longmeadow), and results from the river's interaction with isolated
resistant obstructions in the flpodplain (discussed below). The stability
of the Northfield bend probably signifies that the outerbank deposition
phase has come to a close, and the • river has achieved equilibrium with
respect to either deposition pr erosion. Since the tire-blanket is only
three years old, the stability (dating back at least to 1939) cannot be
considered the result of the the test installation.
(2) The cros= sectional profile of the river at Northampton is typical
of actively eroding rivers; the thalweg (deepest part of the channel) is
very close to the eroding outer bank as expected from classic meander
theory. At Northfield in the vicinity of the test reinf/orcements, the
thalweg follows the inner bank which shows exposures of underlying
glacio ..... fluvial deposits, rather than the new sediment expected in a
typical meander. As will be seen in the photo reconstruction, the
outerbank site is not an eroding.bank and thus is inappropriate for a
test of bank stabili techniques. Sediments and vegetation which now
cover the tires at Northfield could not be expected to develop at a =i+�
of active erosion like Northampton. If the proposed tire ..... blanket were to
remain intact, jt would probably remain fully exposed to view. Erosion at
Northampton is most active at the toe of the outer bank, with consequent
collapse of the overlying bank material. Tires attached to'this bank
could not avoid slumping with it. If the tire-blanket were to
disaggregate, the cost and difficulty of retrieving the individual tires
downriver would be considerable.
(3) The land lost at Northampto. since 1939 is.a crescent shaped parcel
totalling 28 acres in extent. The maximum retreat rate, measured at the
widest portion of the crescent in the vicinity of the large isolated
silver maple has been steady at about ten feet per year since 1939.
During this 46 year period, the loss rate has been . slightly more than a
half an acre each year. • At present land values there, this loss amounts
to about $900/year. Proposed reinforcement with the tire method has been
projected to cost on the order of $300,000. Annual interest, lone, on
.
this amount Hs about 35 times the value of the land being lost each year.
(4) Concern has been expressed about the possibility of an oxbow cutoff
that would isolate the eastern portion of the meadows. Our observations
suggest that such an event is very unlikely for the follpwing reasons:
(a) Reconstruction of ancient channel positions for the Connecticut
River in the Northampton meadows shows that oxbow cutoffs have never
•occurred except at the southern extreme of the valley where the
downvalley progress of meanders is impeded by the Holyoke Range.
• (b) Downvalley migration rates for the Rainbow Beach bend are about
1.5 times those for the Sandy Beach bend, meaning that the Rainbow Beach
loop is widening, not narrowing witb time. Oxbow cutoffs generally take
place when the downvalle; progress of the leading limh (the Rainbow Beach
. bend in this case) is somehow impeded, and the trailing bend (Sandy
Beach) catches up from behind. Since the loop is widening, it is unlikely
that a cutoff will occur in the meadows until the Rainbow Beach loop has
migrated to the south end of the valley •stalled against the mountain,
and has been caught from behind by the Sandy Beach meander. At their
present migration rates, this evolution is likely to require on the order •
of 1000-2000 years.
(c) For a cutoff to develop, it is necessary that headward eroSion
takes place along the future shortcut •channel during flood. The headward
erosion generally begins where the shortcut channel-to be and the
existing channel converge. If such a cutoff event werelikely, evidence
for the initiation of this process should be found as gullies or
• channelways cutting their ways northward into the river's present bank in
the vicinity of Mitch' s Is1and where floodwaters crossing the meadows re-
. enter the existing channel. Even following the flood of May, 1984, no
such evidence can be-found. 8n the contrary, several inches to a foot •of
new sediment was deposited by this flood in the eastern part of the
meadow in the oldest (westernmost) section of the Rainbow Beach woods.
Because of these considerations, we strongly recommend against the use
Of the tire blanket method at Northampton. Since channel depths do not
exceed 11 feet along this bank, the use of stone riprap beneath the
waterline is a preferable solutipn. At the site of most active erosion,
depths do not exceed nine feet. •
However, given the relatively low erosion rates, the unlikelyhood of an
oxbow cutoff at this site, • and the present attractiveness of the river
bank, the best alternative may be to do nothing.
'
Introduction •
When a river flows across a floodplain free of isolated resistant
obstructions (as the Connecticut River does at Northampton), its channel
evolves in the classic manner. Erosion occurs .on the outer (concaye) bank
while new floodplain land is created by deposition across the river on
the convex pointbar (as at Sandy Beach in Hadley). Since the floodplain
is tilted seaward, erosion is most active along the downvalley portion of
the outer bank of each bend, and meanderstend to migrate downvalley as
sets of waves, much as a se,ies of pulses moves along a rope shaken back
and forth on a smooth floor. Reconstructions based on U,S. •Soil
Conservation Service photographs (1939-1980) show that the Sandy Beach
bend is migrating southward at about 10 feet per year while its
downstream neighbor at Rainbow Beach is advancing downvalley at about'15
feet per year (see below) Twenty-eight �
. acres have been lost in
Northampton' along the Sandy Beach section since 1939; seventy acreshave
3:::' een lostby Hadley opposite Rainbow Beach in the same interval (Fig. 1).
Many parts of the Connecticut River's floodplain in Masschuset.s,
however, are not uniformly erodable. By and large:, the valley consists of
the unconsolidated clay and silt deposts of glacial Lake Hitchcock (that
drained about 11,000 years ago). These easily eroded materials are
punctuated here and there by resistant knobs of bedrock (the high-
standing portions of the underlying glaciated bedrock surface), and by
boulder-gravels deposited in the lake by sub glacial meltwater streams.
These resistant materials act as obstructions to the normal downvalley
migration of meanders much as a post in. the floor would impede the waes
travelling along a rope. As a pulse on the rope, facing for . example to
, the right, encounters the post, it will stall and then reflect back
leftward as the next pulse behind it (facing to the left) arrives and
passes the post. Connecticut River mean'erS in several locations have
evolved in much this way. Our researc river over the last eight
6
Figure 1. Channel locations for the Northampton reach
of the Connecticut River in 1980 superimposed on the 1939
USSCS photograph. Twenty -eight acres of land have been
lost by Northampton, while 70 acres have disappeared in
Hadley opposite Rainbow beach on the east side of the river
since 1939. '
7
Figure 2. The Northfield bend as it appeared in 1952.
The eastern bank is the site of a series of bank reinforcement
tests performed by the U.S. Army Corps of Engineers in 1982.
As discussed in the text, this bank has not changed location
since 1939 (the date of the earliest available aerial photograph),
making it inappropriate for tests of bank stabilization procedures.
0
Cd
4
U .0 .0
ro +-■ a
ra o G
m •�
•d W o
G t7
of 1u a)
1a
sa a
aJ o
ri
+.1
o rt +.i ,x
4
•r1 co
L1-1 ,.s~ of
a.) 0
r1 a bD
w G aJ
o W r1 .0
.0 +.1
a .0 cn 0
a
•d G ao a�
aJ •r1 bA .d
at a
v u-1
a) .0 .'o
G up
N U
bD (1) O U)
1 ,C r cd
1i H a p
cj
G cll
m .W
o r1
r1 u (I) +.J
cd a cd
•� c u)
H �+ +.) u
•
ui
+E
a) cd ao
a •r1
0 "d o1
bO f. 0
W 4 •r1 aJ
The geomorphologic data discussed in this report were obtained in the
following way. Bottom topography was measured using a Lowrance model X15A
recording fathometer mounted on a 15 ft. Boston Whaler. The end points of
each traverse •on the river were located by triangulation using two shore
based transit operators in radio communication with each other and with
the boat operator. Transit locations (on opposite banks of the river)
were determined using • standard surveying practice. Recent USSCS vertical
aerial photos, showing fixed landmarks and large individual trees were
used as the b�se. for mapping. Reconstructions of erosion rates for both •
bends were obtained from USSCS photographsfor the period (1939-198O),
coupled with • surveying data obtained this month (September, 1985).
Channel reconstructions based on nineteenth century town maps Were found
' years has centered on the '
e � ver s response to such obstructions, and in
particular on the deposits that are created along the outer concave bank
'just upstream of the obstructions. The results are summarized in a report
published by the American Society of Civil Engineers (Reid, 1983,
attached). Our findings for the bends at Northfield, Hatfield, Hadley,
Easthampton and Longmeadow are similar; in each case a bend of the river
has stalled against a relatively small, non -erodable obstruction, and has
then witnessed deposition on the outer concave bank immediately upstream
of the obstruction as the next meander arrives from upvalley. As shown
below, the specific bend at Northfield (Fig. ' on which the . test
reinforcements were made is the site of recent outer bank deposition of
this sort. As a result, the test results ,are probably not applicabl ` to
//u,o/ally behaving meander like the Sandy Beach bend that is currently
undergoing outerbank erosion.
~
m
Results
Bottom Topography
to be unreliablev, As early as 1831, careful surveying was done to locate
roads, buildings and town boundaries; the river's location was in general
located with considerable artist's license.
Bottom topography for the Sandy Beach bend and the Northfield meander,
• determined from a series of sonar generated bottom cross sectiont is
shown •in Figs. 3 & 4; typical cross-sections are' shown in Fig. 5. The
Sandy Beach bend' displays the characteristic assymetric profile of
normally behaving meanders with the thalweg (deepest part of the channel)
close to the outer eroding bank and a broad apron of sediments extending
toward the thalweg from the inner point bar. The maximum depth measured
for the Sandy Beach bend •was about 11 feet. On several traverses the
thalweg was encountered within 30 feet of the eroding bank (Fig. 3). (Fu,
comparion, the more rapidly migrating Rainbow Beach bend shows maXimum
depths of about 39 feet, also in general close to the outer [Hadley]
bank). Observations of the exposed portion of the Sandy Beach bend's
outerbank indicate that erosion near the base of the chann : el has
undermihed the bank; it has slumped as large units downward into the
river. Slump blocks vary in size, and have moved varying distances
downward. At medium water, the top surfaces of many blocks are at the
•water line, though topographic profiles reveal features well
beneath the surface (Fig. 3 & 5), implying as one would expect, that the
toe of the outer bank is the site of most active erosion.
The Northfield reach of the river is best thought of as thrpp
cd v i
CU • CU
4J n
•1-I • to
O • O
G ai
O
U 4-1
0
O • cd
+-1 U
• •r•+
O a
cd
4 4)
4i
0
z
4J
cd -1
4-1
4
a a
cd
• U
b0 r4
O $-1
a .0
O aJ
4-1
O En
O m
4-1 cd
1J
O N
PI 4
H
M • •
N +.1
• O
O a1
b0 4a
F=+ N
12
Figure 4. Bottom topography at Northfield (contour interval 5
feet). The bottom is unusual in that the thalweg (the deepest
channel) follows the inside of the curve opposite the bank
that was reinforced in 1982 by the U.S. Army Corps of Engineers.
consecutive bends, with the central bend being the one On which tho toot
reinforcements were made (Fig. 4). The profile for the middle bend is
anomalous and strikingly diffm,ent from profiles for the Northampton
meander. The dominant thalweg of the middle Northfield bend follows the
inside of the curve. (The downstream portion of the outer concave bank
[the site of the tire reinforcement] exhibits a smaller channel. A
shallow bar occ urs between these channels, that can be seen in many of
the USSCS photos, and appears as an island in 1909 town map.) At the
water line, the inner bank consists of the coarse glacial gravels that
were deposited into Lake Hitchcock by meltwater streams. Normally, the
inner bank of river bends consists of the newly deposited materials of
the gently shelving point bar, as is the case at Sandy.Beach in Hadley.
The glacial gravels exposed at the base of the inner bank at Northfield
represent the "bedrock" into which the river is carvihg its channel, and
demonstrate that the inner bank is the site of recent erosion -- an
anomalous circumstance.
Immediately downstream from the series of test reinforcemeH s on the
outer (east) bank is another outcrop of coarse glacial gravel which
projects as a gently curved point into. river. The boulders in this
deposit --up' to about' the size of basketballs -- exceed the maximum size
that the 'river is capable of transporting. The gravel deposit therefore
constitutes an immovable obstruction, and for. perhaps the last several
centuries, has stalled the downvalley progress of the meander on which
the test reinforcements were made. Using the meander scars visible in
photographs of the plowed fields east of the river (Fig. 2), we have
reconstructed the previous positions of the channel as the river evolved
into its present configuration. This reconstruction is shown in•Fig. 6.
Recent evolution has consisted of deposition on the outer bank and
erosion on the inner point the opposite of what is usually the case.
r,?
14
, 1
■
a
* bedrock obstructions
\-t
4 0
oar v. coarse gravel
rec ent
utbank deposit
i. s
Figure 6. Hypothetical evolution of the Northfield reach
of the Connecticut River. Downvalley meander migration has
been arrested by a series of non- eroable obstructions, re-
sulting in the development of anomalous outerbank deposits
upstream of the lower two obstructions. One of the two sites
of recent outerbank deposition is the site of U.S. Army Corps
of Engineers bank stabilization tests done in 1982.
goologic-pact^- Two - additional-obsLrucLi ohs - cumpuseJ - of bedrock occur on
15
the west bank of the river, one at the site of the old Route 10 bridge,
and the other downstream at the Northfield Mount Hermon School boat
l°nding. Together, these three obstructions have apparently stalled the
downvalley progress of all three bends at Northfield. The entire reach is
now locked in position as a result, as will be seen in the 1939-1985
channel reconstruction below.
Reconstructions of channel positions for the period 1939 1985
Using historic USSCS aerial photographs for the period 1939 1980 along
with surveving data obtained in September, 1985, we have determined the
rates of migrations for the river's channel at Northampton and
Northfield. These results are shown in Fig. 7. The maximum rate of bank
retreat at Northampton (near the large silver maple) averages about 10
• feet per year for the 46 year interval, while the position of the bank at
Northfield that recently received the test reinforcement has shown no
change in the same inte (With the advent of the pump-storage
facility at Northfield Mountain, some bank •degrad. has occurred on
both banks of this reach at Northfield. However, no significant retreat
is discernable from a comparison of our 1985 survey data with the most
• recent USSCS photo (1971) available to us, meaning that the retreat is
probably less than about 0 meters.)
Given its recent anomalous history, the Northfield bank is therefore
not appropriate for a test of bank stabilization. In particular, the
results of test reinforcements made there are probably not meaningful
when applied to the Northampton erosion problem.
NORTHFIELD BEND
2.2
2.1 --
1.8 -
1935 1945
1935 1945
NORTHAMPTON - BEND
retreat rate opposite Sandy Beach
vertical scale as for Sandy Beach
1955
16
year
year
1965
1975
1985
1955 • 1965 1975 1985
Figure 7. Changes in bank location at Northampton and Northfield
since 1939. The Northampton bend has migrated southward about
490 feet since 1939, while there has been no measurable change
in bank location at Northfield. The Northfield bend is therefore
inappropriate for a test of bank stabilization procedures.
Rec
We recommend against the of the tire blanket at Northampton for the
following reasons:
(1) In terms of installation costs, environmental degradation and
possible clean up expenses, the costs of the proposed tire-blanket method
are very high and make it a poor choice for bank stabilization at
Northampton. In addition it is not certain that the method would succeed
since it was tested in an environment quite different from the
Northampton bend one undergoing no measureable erosion since 1939. •
(2) The relatively shallow thalweg at the Northampton bend (about 11
feet) sugge�ts that bank stabilization could be achieved by reinforcement
• of the toe of the eroding bank with stone riprap. If this material were
to be applied only tb the surface be.ew the waterline, the expense would
be minimized, and the upper bank would remain in its present attractive
state. Since erosion occurs largely by undermining the toe of the
outerbank, and the subsequent sluMp of the overlying material, it is
likely that below - waterlineriprap would suffice to stall the erosion.
(3) In terms of acreage lost per year, the Northampton bank retreat is
not severe (about 1/2 acre per year). The value of the land lost each
' year (ab $900) is about 1/35 the annual interest (at 10%) that could
be earned on the estimated $300,000 cost 'of the tire-blanket
reinforcement. Whether reimbursement for land loss could b� made from the
money saved is a matter for consideration but is beyond the scope of this
_
(4) The observable history of the Connecticut River in the meadows is
such that oxbow cutoffs have never occurred except at the downvalley
extreme 'of the floodplain against the Holyoke Range (Fig. 8).. Given
therefore that a cutoff is not a likely event in the meadows before the
Rainbow Beach meander migrates to the south end of the valley (perhaps
1000 2000 years in the future), the best solution for the Sandy Beach
meander may be to perform no corrective measures at all�
19
Figure 8. Meander scars on the Northampton meadows as they
appeared in 1940. There are no oxbow lakes in this record.
Oxbows only occur at the southern extreme of the floodplain
where downvalley meander migration has been arrested by the
presence of the Holyoke Range. As a result, it is very un-
likely that an oxbow cutoff will occur near the Rainbow Beach
meander (see text).