APRIL 3-4, 2005 - DELAWARE RIVER FLOOD
Yardley, Bucks County, Pennsylvania
By Larry Hale - May 14, 2005

On Mon, Mar 28, 2005 the Delaware River started to rise. It crested at 17.4 feet in Trenton late Wed night but more rain was on the way, and there was deep snowpack in the upper part of the watershed. By early Saturday the river had dropped to only 14.8 feet when it started to rise a second time. The river crested Sunday morning, Apr 4, 2005 at 25.3 feet, the highest flood in 50-yrs. This article summarizes my recorded observations and supplements information presented in my previous article dated March 10, 2005.

CANAL OBSERVATIONS
From Mt. Eyre Road to Morrisville [Distance scaled 5.2-miles on Lower Makefield Map ]

1. Mt. Eyre Road to Lock 6:
Escaped damage from the river. I don't see any evidence of overflows over the towpath in either direction. There are well built high capacity waste gates up here. At other locations spillways did their job with no reverse flow. Furthermore, I couldn't see any evidence of significant overflow of canal waters over or around the top of the Lock 6.

2. Lock 6 to Mary Yardley Bridge:
Floodway overflowed into canal starting immediately below Lock 6. River water entered canal spread-out over 150-ft front, lifting all red stone surface material from the towpath surface and dropping them onto either the bank or side of the canal, but not concentrated enough to produce major damage ---by the time the floodwater entered the caanal at this location, the canal was probably already backed up and not moving very fast, so for the most part it was a gentle topping. There were similar toppings all the way to the Camelback Bridge most of them less than 6-inches. The rest of the way to Mary Yardley Bridge included more severe inflow from the river well over 12-inches at the worst location by the lake. Using a rating system 1 to 4, where 1 is "almost no damage" and 4 is "severe damage" the canal bank (next to the towpath) between Lock 6 and Mary Yardley Bridge varies between 2 and a 3 depending on depth of overflow. The entire opposite side of the canal has no damage. Canal damage is limited to some undermining of canal bank, and washing away of the red stone material, although much of the red stone and some of the clay bank are still in place.

3. Mary Yardley Bridge to Maple Street:
The towpath, and/or the berm, and the paved street are all high. There was no damage to the towpath and the canal bank was not undermined or significantly damaged, and looks the same as the canal bank on the opposite side of the canal.

4. Maple to Brown Street:
Severe damage (4 on our scale of 1-4) all along canal bank caused by riverwater overflowing the towpath into the canal... undermined the canal bank. The river was flowing fast into canal on average 12-in deep, probably the amount of the flow limited only by the total canal discharge capability, as follows:

As the river crested... river floodwater entering above Mary Yardley Bridge (25%?) + River Floodwater entering between Maple and the Aqueduct (75%?) = Discharge at Aqueduct Sluice Gate + Discharge at Lock 5 Sluice Gate + Discharge Over Towpath.

Once the level between the canal and the river flood stabilized, the open sluice gates probably achieved no effect, and the total amount of water escaping over the towpath would have remained the same.

5. Brock Creek Vicinity:
Floodway and canal were joined. The maximum water level reached at least the top of the concrete Aqueduct. Brock Creek is the approximate location where on April 3-4, 2005 the river flood level (which is on a slope) crosses the water surface of the canal (which is flat).

6. Afton Avenue to College:
Flood covered towpath right up against the high point on the berm with no overflowing past the berm for the first two-thirds of the distance to College, but serious damage nonetheless with collapse of the wall of the canal rather than undermining. Damage is rated 3 on scale of 1-4 where 4 is the worst damage. The canal bank was very seriously damaged the last one-third the distance to College where the water ran up against the stone bridge and was forced back into the canal at College Avenue.

7. College Avenue to Letchworth:
Direction of flow across towpath reverses ---the first reversal over a short distance was above College. Past College it's now the canal that overflows at locations where the towpath is low. The towpath is high behind the church on Pennsylvania Avenue and this part of the towpath remained above the flood. The worse flooding occurred at the usual location just up from Letchworth where a more permanent fix is needed. Here the towpath is repeatedly destroyed simply because it is at the lowest elevation and needs to be raised.   Letchworth "Failsafe" Pilot Project    However, a permanent repair, although preventing the damage to the canal at this location, would not have significantly reduced the floodlevel between the canal and the river during the crest of the April 2005 flood because most floodwater was coming directly from the river.

8. Letchworth to Lock 5:
Canal overflows at a number of low points and in the past has broken through at others. But a major overflow always naturally occurs over the last 150-ft of the towpath just up from Lock 5 (possibly as much as 1-ft in Apr 2005). There appeared to be no significant flow straight ahead over the top of the stone lock.

9. Lock 5 (lower side):
High velocity river water from the floodway tore through the canal towpath. The size of the flow and extent of the damage was determined by the level of water in the floodway, fed by the river at Brock Creek. [Front Page Photo Close-Up View, April 12, 2005 Courier Times]

10. Lock 5 Partway to Blackrock Road (36-inch Towpath):
The discharge from the Railroad Impoundment was successfully contained by the canal for the distance of the 36-inch towpath with little or no damage, while flowing near its brim. Alongside the canal, water level in the floodway was about 3-ft depth, representing mostly waters backed-up due to lack of grade along the bottom of the waterway.

11. Lock 5 Rest of the Way to Blackrock Road (24-Inch or Less Towpath):
There were overflows from the canal where the towpath was 24-inches or less and severe washouts where the towpath was 12-inches or less. Waters in floodway remain 3-ft deep. In some places a berm between the canal and the floodway is what prevented overflow. Toward end of Macclesfield there was deep overflow and almost a major breach. Halfway to Blackrock Road canal is high over the floodway and there is no berm. Water surface in floodway may be about 8-ft below the towpath. Half mile before Blackrock Road the Potato Creek is observed. The floodway may be 500-ft across. Where the picnic area begins I measured 4-5 inch deep overflow at the deepest location on the 50-ft front. It is worth noting that the canal is on high ground all the way from the railroad to Morrisville.

The canal overflows to the natural floodway at various low spots along the towpath between the railroad and Black Rock Road. Instead of allowing the canal to overflow randomly, this overflow needs to be controlled by constructing a spillway; including raising the towpath in whatever way is required to accommodate the spillway. This would prevent damage to the canal from the April 2005 flood, and allow the floodway to maintain its present function. Flow through the floodway from April 2005 event may possibly be increased to somewhat restore what nature intended.

There is no way of knowing the amount of flow passing through the canal at the canal culvert on Blackrock Road on April 3-4, 2005. This would depend on how much was being released or overflowing further down the canal. There was no towpath overflow between Blackrock and Morrisville, and it's likely the waste gate located directly above the dike was kept closed. In the worse case scenario, if there are no releases or overflows further down, this would mean the total discharge from the Railroad Impoundment and the wide open Lock 5 sluice gate is overflowing the towpath above Blackrock Road... this probably is the best location for the overflow but requires installation of a spillway.

Canal Observations Summary - April 2005:
Floodwaters from the river flowed over the towpath into the canal from Lock 6 to Mary Yardley Bridge at Fuld, damaging the canal but not severely. There was no flow in either direction where the towpath is high between Mary Yardley Bridge and Maple. The worst damage to the canal bank occurred between Maple and the Aqueduct where the canal was joined to the river 12-in deep over Edgewater Avenue between Brock Creek and Maple. Further down the canal the situation reversed with minor overflowing [just] below College Ave, no overflowing at all in the vicinity of the church on Pennsylvania Avenue, a major overflow with extensive damage just above Letchworth at the usual location. Below Letchworth the towpath is 12-inches for half the distance to Lock 5 and 18-inches for most of the rest, but drops down over 150-ft front just above Lock 5, however, the towpath is not destroyed because of a wide berm. The Brock Creek sluice gate was open and held canal flow in check up until the time the river started entering the canal. All the canal overflows consisted almost entirely of floodwater from the river, which had entered the canal between Lock 6 and the Aqueduct. Major damage and complete destruction of the canal towpath occurred just below Lock 5, where floodwaters mainly from the river flowed over the towpath into the canal.

Canal Geometry - Lock 6 to Lock 5 (As-Built)

Height of Canal Towpath...
It will be shown in a later section of this report, that the floodway at Yardley is naturally constricted...   The valley narrows, the land rises, and almost unbelievably the height of the canal towpath drops. It is unknown to me, whether or not the towpath problem is unique to Yardley.    Why is the towpath low?

The Canal:
The operating level of the canal is automatically maintained by the fixed elevation of the overflow weir at the Brock Creek Aqueduct. Standard design practice in the 1800s required a minimum towpath height 18-inches above the weir, and both the Aqueduct and Lock 5 were set at this height. But the towpath is much lower. The design follows the classical pattern… the fixed level of the water surface is more or less even with the terrain starting at Lock 6 but well above the adjacent terrain when it arrives downstream at Lock 5. Typically, floodwaters from the river might be expected to more readily enter just below each lock where the canal is low.

The River:
Unlike the canal surface which is flat, the bottom gradient of the river is sloped and the slope of the water surface (toward Trenton) greatly increases when the river floods. The relation between the canal and the river is complicated by the fact that the towpath varies from more than 2-ft above the required 18-inches height (at Lock 6) to less than 6-inches height at various low points.

Towpaths at Lock 6 & Lock 5 (Compared):
Inflow from the river isn't as bad as expected below Lock 6 because of the extra height of the towpath. Although the canal is about even with river road, the towpath is 2-ft higher than the required 18 inches for a considerable distance. At Lock 5 the situation is reversed ---although the towpath below the canal is 1-ft above the required 18-inches for a short distance, the water level in the canal is 2-ft lower than River Road. As a result, the river at the location where it is highest was only 6-in over the towpath, whereas at the location where the river was lowest, floodwaters destroyed the entire towpath. Nature definitely exploits the advantage of the extra height and "pressure head" available at Lock 5.

April 2005 Flood Scenario (Based on findings to date):
[The peak] overflow from the canal was probably less [than] one might surmize looking at the damage, and was certainly very much less than June 1996. A large part of the potential river inflow above Brock Creek was actually being short circuited by the direct connection to the river [in the vicinity of] Brock Creek once the river reached its maximum level. Breaches as they occurred drew water mostly from the river in the vicinity of Brock Creek, and much of that water rejoined the river at the various overflow points. The major overflow just above Lock 5 and possibly part of the overflow just above Letchworth are the exceptions, but even the sum of these two was only part of the waters escaping through the breach below Lock 5. Damage at the breach was due to high velocity through a limited cross section.

At Blackrock Road the amount passing through the canal culvert was probably nil, and the amount passing through Potato Creek culvert perhaps matched what passed through the culvert under the railroad. The flow over the top of Blackrock road although on a wide front was not deep. The main potential for flooding in the vicinity of Blackrock Road is not these waters coming down the floodway, but rather backup up of the river into this former channel now constituting the floodway.

The flood level along the relatively narrow stretch of flooded land above Brock Creek appears to have maintained a flood elevation closely allied with the river. The flood level drops 5.6-ft from Lock 6 to the railroad. The slope of the river is slightly less than average above the pumping station and slightly more between the pumping station and the railroad. Meanwhile, the towpath starts out below Lock 6 about 2-ft higher than the normal 18-inch height, but drops toward Mary Yardley Bridge on a slope greater than the river, so the river steadily gains on it looking for the low spots along the varying height of the towpath, with the main inflow occurring between the Aqueduct and Maple.

The relationship between the canal/towpath/berm and the floodlevel over the land and river becomes even more complicated over the last 1/2-mile or so before the railroad bridge. Unlike the upper half of the Borough, the flood over the land below Brock Creek is higher than the river due to the effect of the ridgeline (of the former island). Measurements along College Avenue at S. Bell indicate that the flood on the canal side of the ridgeline was 2.25-ft higher than the river level at College with a limited amount of overflow to the river across the intersection of College and Longshore, and northward along Longshore to College as well. A large flow escaped to the river through the fence between the old Eastburn Apartment Complex and Belmondo where a long line of leaves was found heavily indented on the fence. The floodlevel in Belmondo was 36.2-ft LMT datum, and escaped from Belmondo through the entranceway dropping another foot or more on its way to the river. Most of the ridgeline along the former island remained dry (except for a modest dip northward toward College where the road cuts through), and is believed to be the major factor causing waters on the canal side of the ridgeline to be higher than the river (except for the floodway between Belmondo and the canal, where the effect was less due to drawdown by the breakthrough at the canal). The potential total capacity of the discharge points exceeds the amount of water being delivered, accounting for the low flood elevation adjacent to Lock 5.

The key to understanding what's happening along the canal is the fact that whereas the canal is level, the river at flood stage is sloped, dropping from 40.6-ft LMT datum at Lock 6 to 35.0-ft LMT datum at the Railroad Bridge during the April '05 flood. Due to the open terrain, we can probably assume that the flood level of the main channel of the river nearest Lock 6 is probably the same as the measured flood level immediately below Lock 6. Since the flood elevation was 35-ft LMT datum at the Yardley railroad bridge, the river must level out as it passes between the various island and reaches Trenton Gage just above the Falls of Trenton. Finally, based on all the recent floods it's possible to reconcile recent data with the various eye witness accounts from Aug 1955... 1) In 1955 the slope of the river was definitely greater than April 2005, and 2) In 1955 a similar type of differential definitely existed across the ridgeline in the lower half of the Borough. However, the additional rise would probably not be linear, i.e. not as high as expected, due to additional discharge over the top of Lock 5, and to a much lesser extent by flow over the above mentioned ridgeline back to the river ---the 1955 flood level was 4-inches deep over the ridgeline at the high point on Letchworth according to one eyewitness and longtime resident on Letchworth.

THE DELAWARE - A BRAIDED CHANNEL RIVER

The Delaware River has many islands along its length where the river divides into two channels, and may be classified as a "braided channel" river   [Ref: “A View of the River” by Luna B. Leopold, Published by Harvard University Press, 1994, Page 56]. At one time the Delaware River was undoubtedly one single much wider and deeper channel, with a terrace along the lower Prophet 21 property line. As it subsided there was uneven deposition from the glacial melt, leaving the terrain in much the same configuration we see today.

Floodway Scenario:
Above Yardley a second channel was formed (where today there is a wide deep floodway), and this channel crossed back to the river at Yardley, at the same location where the extra channel below Yardley had its beginnings. This is the braiding effect... the crossover is caused by alternating huge flows and lesser flows manipulating the sand and gravel deposits left by the glacial melt. The channel below Yardley continued all the way to Morrisville, with a huge sand and gravel shoal (the island) emerging on the left side and separating the two channels. Due to the way the valley widens out and bends away from Yardley, the land between the river and the canal rises in Yardley and terraces were formed at various levels. But the bottom of the channels just above and just below Yardley remained wide and deep. As the river further subsided, these channels tended to dry up, thereafter serving mainly those creeks that still entered at what was formerly the river channel. Naturally these channels drain in the same direction as the river. Eventually they evolved into the wide deep natural floodways above and below Yardley, with river terraces at various levels in Yardley still subject to occasional flooding. There are almost no houses built today anywhere in this floodway, except in Yardley where the land is higher and classified as a terrace.

Natural floodway along the river...
[ A major player in Yardley's flooding problems, perhaps more important than the river itself ]

Hundreds of years ago this natural floodway was wide open, located parallel to the river, open to the river at the various creeks, mainly receiving discharge from some creeks, and serving as a second channel when the river flooded. Obstructions built in the floodway over several hundred years have essentially defeated nature by converting much of the natural floodway into a series of man-made impoundments.

My interpretation of the terrain in the immediate vicinity of Yardley is as follows ---identified in steps (Step 2 being higher elevation than Step 1, etc):

Step 1 - The part of Route 32 which frequently floods between the berm (on which the original bungalows are located) and the river---is definitely a flood plain.

Step 2 - Morgan Avenue located behind the berm is a low-lying terrace, which occasionally floods. [Morgan is separated from the river by a natural berm along North Delaware on which the original bungalows were built.]

Step 3 - Pennsylvania Ave. is a higher terrace that only rarely floods.

Step 4 - Canal Street is a much higher terrace near the edge of 100-yr flood zone.

Step 5 - S. Main St. appears to be on the natural berm of an ancient river and there are no known reports of river flooding since the settlers arrived.

Step 6. The most obvious (easy to identify) terrace in Yardley is located along the lower edge of Prophet 21 and the river probably hasn't flooded to this level in thousands of years.

Various flood frequencies up to and including Sep 2004 reactivate the tendency for Brock Creek (and flow from the floodway above Yardley) to permanently reconnect to the river at a location somewhere below the mouth of Brock Creek. And the higher Apr 2005 flood flowing over the land would tend to reestablish the old channel between Yardley Borough and Morrisville, again if the floodway was not blocked (mainly by the railroad). The Aug 1955 flood almost certainly would have succeeded in reopening the channel from Letchworth to Morrisville, if the floodway was not blocked.

FLOODWAY OBSERVATIONS

1) Floodway Above Yardley: Floodwaters tend to backup as the floodway approaches Yardley. Deposition during the glacial melt caused the land to rise toward College Ave, even more so on Pennsylvania ---the Sep 2004 flood reached College Ave. but not as high as Pennsylvania Ave. The next higher terrace formed by the ancient river is along Canal Street, which did not flood in April '05.

2) Floodway Letchworth to Railroad: The uppermost tip of the floodway probably begins near Letchworth. The floodway deepens toward the railroad, where floodwater is impounded behind the railroad. The outlet is the low-lying land along the canal immediately below the lock, where the river broke through the towpath.

3) Railroad Impoundment: Floodwaters trapped behind the railroad flowed over the towpath below Lock 5 at high velocity and destroyed the towpath. This overflow should be separately conducted directly to the floodway on the other side of the railroad. This could be achieved by the design and constructed of a 3-way junction including 1) inflow from canal, 2) a wide circular overflow structure from floodway, and 3) a reinforced concrete passage conducting the flow to the other side of the railroad embankment. It would be desirable to increase somewhat the April '05 flow through the floodway below the railroad to at least partially restore the original natural condition, but this needs to be approached cautiously because the floodway impoundment created by the railroad is fed by the Delaware River!

4) Floodway Railroad to Blackrock Road: Immediately past the railroad the floodway was filled-in and drainage is through a small ditch receiving water from the pipe under the railroad. The ditch connects to what may be the former channel of Silver Creek, which flows into the floodway. The floodway becomes extremely large and deep further down toward Blackrock Road. The ridgeline between the floodway and the river is higher than the April 2005 river level, and appears to have kept the floodway separate from the river ---the April '05 river level did not spill into Macclesfield. The original natural floodway fed by the river at the mouth of Brock Creek is cutoff (mostly by the railroad) from the flow it otherwise should naturally have received from the April 2005 river flooding.

5) Mini-Impoundment at Blackrock Road:

The Apr '05 flood depth over Blackrock Road appears to be caused mainly by the damming effect of the road itself. Although not deep, the floodwaters did on the margin extend over to part of Ivy Lane. If the road was not there and if the natural floodway was still its original size and shape, the April '05 floodlevel would likely have been 2-foot lower. That's the effect that needs to be achieved by an opening under the road no matter if Blackrock Road remains at its present elevation or if it is raised. My guess is that the capacity of the Potato Creek culvert matches the flow through the pipe under the railroad. If so, the flow over the top of Blackrock must include 1) the Lock 5 bypass through the sluice gate, plus 2) plus the discharge into the canal from the impoundment behind the railroad, all of which overflows the canal between the railroad and Blackrock Road.

The presence of the Trenton Gage river level at or near Blackrock Road in April '05 in combination with the potential size of the floodway, and relatively low flow in the floodway means that the April '05 should be accommodated, but the key word is "potential" because the original floodway in the vicinity of Blackrock Road has been constricted in many ways on both sides of Blackrock Road. Local storm runoff should not be a problem due to the elongated geometry of the watershed and the flatness of the floodway.

Even if nothing further is done at Blackrock Road, the spillway should still be constructed, and will not increase the flow in the floodway beyond the present amount overflowing from the canal into the floodway. The spillway needed to protect the canal between the railroad and Blackrock Road is best located not too far below the railroad where there is a solid berm between the canal and the floodway. In the event that it takes a long time for various other work envisioned in this article to be carried out, then install the required spillway at the end of the 36-inch towpath, and this will provide some immediate protection to the canal.

6) Floodway Below Blackrock Rd:
This wide deep floodway is subject to natural flooding from the Delaware River backing up through the mouth of Potato Creek and overland at Ferry Road. The Apr 2005 flood-level in the floodway remained parallel to the water surface of the canal almost to Blackrock Road and perhaps beyond.

7) Morrisville Dike: It's interesting to read that deposition had closed the floodway channel by the time of the early settlers, but that a massive flood in the late 1600s had reopened it, and that man had recently closed it again. The key to understanding what's happening at the lower end of the floodway is the fact that Potato Creek, which drains the floodway, connects to the river just above the Morrisville Dike. It is environmentally sound to allow floodwaters to backup in the floodway.

Floodway Restoration Plan:
Any floodway restoration either now underway or proposed, should place special emphasis on Potato Creek in the vicinity of Ferry Road, where "active restoration" is definitely needed, including but not limited to removing those piles of materials deposited in the floodway.

RIVER LEVEL

Aerial Photograph of Flood...
[Courier Times - April 5, 2005 - Large Photo on Front Page]

Photo shows peak flood condition (brown water) and the dry land (green grass), including the dry ridgeline well above the floodlevel at the high point on Letchworth Avenue. Photo also shows dry land separating the rear of Belmondo from the basin along the canal.

The "Trenton Gage" located just up from the Calhoun Street Bridge, reads the level of the river above mean sea level. When it reads 7.77-ft the flow in the river is zero. My observations are anchored to elevations at the Yardley Pumping Station. The sewer datum is not necessarily the same datum as any other, but the station is close to the river, and has the obvious advantage of being tied to manhole elevations throughout Yardley. The "common denominator" is the river's cubic feet per second (cfs) flow, which at high stage does not vary significantly between the Yardley Pumping Station and Calhoun Street. However, the level or stage of the river is different at various locations along the river, and this is one of the confusion factors leading to false rumors and erroneous news reports.

Just because the flow is assumed to be constant, this does not mean that the number of feet the water rises at Yardley is the same as it rises at Trenton. Obviously the flow is the product of the average velocity times the cross section area, and the product of the two numbers remains the same. The "catch-22" is that when the velocity slows the stage rises, and vice versa, but you can't see the average velocity! This helps explain why there is not always a 1:1 ratio between the rise in Trenton and the rise at various points in Yardley.

Trenton Gage Level versus Yardley Level Below College Ave

                 Trenton         Yardley 

Aug 55         28.6           0.00    Benchmark Elevation 

Difference      3.3           2.25   

Apr 05         25.3           2.25    Below Benchmark

Difference      1.9           2.00

Sep 04         23.4           4.25    Below Benchmark

Difference      1.2           2.00

Jan 96         22.2           6.25    Below Benchmark

As shown in the table, my hand measurements over the years at one specific location in Yardley (near River Road at a location 600-ft downriver from centerline of College Ave) indicate that the amount of the rise in Trenton corresponds to a different rise in Yardley. The situation between Trenton and Yardley appears to actually reverse with the rise at Yardley being greater than at Trenton for certain lower level floods, and apparently less than Trenton for certain maximum floods. In addition to different levels along the river, the flood level at the river is not necessarily the same as the flood level inland along the line perpendicular to the river. Floodwaters west of the ridgeline along the top of the former island were higher than the river level.

Various height differentials exist between the surface-level of the flood as it crossed over the land, versus the level of the flood at various locations along the river. The level tends to be higher on a line from the mouth of Brock Creek toward College, and is drawndown as itemized below.

1) On Longshore floodwaters were observed flowing several fps toward College Avenue ---this was the clue that led to understanding what was happening.

2) On College Avenue a 2.25-ft differential was measured between the flood at South Bell and the river level at the end of College.

3) The new storm sewer on Letchworth may also contribute to the drawdown. I noted during the Sep 2004 flood that there was zero backup into the storm sewer when the river rose.

4) At Belmondo there was strong flow through the fence between Riverquick and Belmondo... The floodwaters in Belmondo were witnessed escaping to the river through the entranceway. There was a 200-ft strip of dry land between Belmondo and the floodway basin along the canal where the waters were being drawn down through the opening torn through the canal towpath. An army NCO who lives at the rear of the complex said that the water entered the parking lot from the direction of the river [actually from RiverQuick] and not from the direction of the canal [true].

My measurements indicate the level dropped to 34.5-ft LMT datum just up from the overflow in the basin located next to Lock 5 and appears to have been 1.7-ft below the flood over the Belmondo parking areas.

SLOPE OF FLOOD SURFACE

It was only when I found out there was a 5.6-ft drop in the river surface between Lock 6 and the railroad on April 3-4 ---I previously thought it was a couple of feet--- that the importance of the slope of the river versus the varying height of the towpath came into focus. The transition point occurs at about 38-ft LMT datum near the Aqueduct, north of which the river level dominates.

The base variable is the natural gradient of the bottom of the river. Estimated based on USGS sheets showing a 20-ft contour crossing the river 6-miles above Trenton Gage... The average must be about 20-ft minus (7.77-ft + dry weather depth) equals roughly 10-ft. To make this rough estimate easy to follow, I'll assume the normal flow at Trenton is 10-ft (Gage Height) and allow a 4-ft drop at Scudders Falls ---this drop makes sense in terms of the probable water level that was maintained in the industrial water canal compared to the river at a point below the falls ---the wing dam for the canal is just above the falls. This leaves 6-ft so that the average slope (excluding the drop at Scudders Falls) is about 1-ft per mile (Slope = 0.0001) for the natural gradient. At high stage the slope increases.

Slope of River Varies with Discharge...
"Water surface slope does not change appreciably with discharge when measured over a distance equivalent to several channel widths."
[Ref: "A View of the River" by Luna B. Leopold, Harvard University Press, 1994]

The distance between Yardley and Trenton is more than a few channel widths. As a general rule, the greater the flood, the greater the slope, but at any specific location the situation may be more complicated that this implies. There are a number of variables that affect the elevation of the water. In addition to straightforward matters related to slope discussed so far, there is the question of backwater, in our case caused by high tide in the Delaware Bay. Intuitively one understands how an obstruction may cause the water to rise behind the obstruction, if the obstruction is large and/or continuous. What is less understood is how a "backwater curve" works.

Backwater Curve Defined...

"The longitudinal profile of the water surface in an open channel where the depth of flow has been increased by an obstruction, an increase in channel roughness, a decrease in channel width, or a flattening of the bed slope."  [FEMA 148, Glossary of Terms, Revised 2003]

HIGH TIDE IN THE DELAWARE BAY

As the previous Sep '04 flood neared its crest it was falsely rumored that the flood level was predicted to increase another 5-ft because of the high tide in the Delaware Bay. It is well known that high tide may increase the river level due to a backwater curve that develops on top of what the level would otherwise have been. Let's look at this in more detail, first the basics…

Tidal Variation is caused by the rotation of the earth every 24 hours and 50 minutes with respect to the moon. The moon causes a bulge in the oceans, thereby creating high tides on the two opposite sides of the earth that are aligned with the moon. The tide nearest the moon is usually somewhat higher than the tide on the side away from the moon. This translates into two high tides per day in the Delaware Bay as the earth makes one rotation on its axis. There is also a tidal effect caused by the sun and there are various monthly, yearly, and 19-year cycles as well. In addition there may be storm surges and various other anomalies. All of this is well known by ship captains and drawbridge operators.

Tidal Effect on River Level:
In terms of Mean Sea Level (MSL), the typical tidal variation occurs during 12 hours 25 minutes over a varying range. On April 1, 2005 the range at Burlington was +7.2 feet to -0.7 feet MSL for the tidal variation [this just happens to be the one I printed]. This particular high tide at 7.2 feet MSL just about intersects the 7.77 feet level of the Trenton Gage. The rising tide at Trenton can cause the river to backup all the way past Yardley. The height of the backup at any distance above Trenton can be estimated by working with the "backwater curve."

Exactly where the backup starts is one of the factors, which determines the effect of the backwater curve on the flood level at Yardley....

"Theoretically speaking, the backwater curve [caused by the tide in our example] extends indefinitely in the upstream direction; hence, it has no upstream end point… the backwater envelope curve starts at a point where the static pool level in the reservoir [in this case the high tide] at zero inflow intersects the channel floor." [Page 319, "Open-Channel Hydraulics" by Ven Te Chow, Ph.D., McGraw-Hill] The upper end of normal high tide is probably just below the Trenton Falls but the backwater envelope may sometimes start somewhere near the Trenton gage.

"As the inflow to the reservoir increases, the end point of the backwater curve may move either upstream or downstream, depending upon many factors…When the reservoir level is kept constant and when the channel is prismatic and has a simple cross section, it is most likely that the end point will move in a downstream direction as the discharge is increased. Increase in channel roughness usually results in downstream movement of the end point, since its effect is to reduce the length of the flow profile. The presence of flood plains has a similar effect." [Page 319, "Open-Channel Hydraulics" by Ven Te Chow, Ph.D., McGraw-Hill]

This appears to me to be favorable to Yardley because as the end point moves downstream, the backwater at Yardley drops. As a result, the rise at Yardley due to high tide is expected to be less than the rise at Trenton. However, this difference might not always show-up if other variables change, for example, possible increased height of river above the Yardley Bridge in 1955, depending on the buildup of debris in front of the bridge just before it collapsed.

Measuring the Backwater:
The actual crest of the river during April 3-4, 2005 was predicted to occur around 9:00 o'clock in the morning, presumably based on analysis of peak flows passing the upriver gages. However, the river did not start to fall as anticipated and news reports cited the high tide as holding it back. Actually I measured an additional rise by around noon of 4-inches just downriver from the Yardley Pumping Station. This is not definitive because there are just too many variables, but it is indicative of the problems involved in trying to match the rise at Trenton with the rise at Yardley. Factors such as precisely when the tide peaks, or what unusual anomalies such as a storm surge exist, or some unusual combination might surprise even the experts.

Marking Flood Levels on Poles:
Someone asked about marking the April 2005 flood level on poles throughout Yardley. This is presumably to help the homeowners and others predict the level of future floods near their homes… a dangerous practice for all but the most highly experienced and very well informed due to the many variables involved. The worst scenario would be a new homeowner making a judgement based on pole marks and unaware of unusual conditions such as heavy snowpack, or potential for a sudden large rise due to an ice jam. It's a case where incomplete advance knowledge might in some cases be a poor substitute for proper interpretation and rapid dissemination of reliable information. An additional problem is that flood levels vary in a number of ways inside the Borough and there is no fixed relationship between the stage at Yardley and the stage at the Trenton Gage. Homeowners, who mark their own poles, should do so only if they are replacing actual witnessed flood marks at specific locations.

National Weather Service:
The National Weather Service did fine job and used a cautious measured approach, gradually increasing their prediction on each of the days prior to and on the morning of the flood. The prediction kept rising in small increments but was never reduced. The predicted flow was plotted as "green" points with the actual flow following along as a continuous "blue" line. Each green dot disappeared when the blue line reached that same point in time. Just before my power failed the solid line was approaching one of the dots, and there was an offset indicating that the actual flow was trending lower than the predicted flow. Later that night by candlelight, I phoned my daughter who gave me an update as the blue line approached another green dot.