American Society of Civil Engineers.

Transactions of the American Society of Civil Engineers (Volume 81) online

. (page 109 of 167)
Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 109 of 167)
Font size
QR-code for this ebook

including as it does not only storage in the channel proper but the
filling of side ravines and tributary valleys and the overflowing of
bottom lands back to high ground or levees. The natural impulse is
to give such storage great weight in holding down flood heights; but
this is clearly inadmissible. Channel storage is mainly absorbed in the
preliminary stages of a flood. The case is similar to that of a reservoir
which has been filled from the earlier rains and has no space left for
the later rains; and though it may still exercise a slight moderating
effect on run-off passing through it, such effect is not at all in propor-
tion to the full capacity of the reservoir; and the more prolonged the
crest of the flood is, the less will be even this attenuated effect.

The same conclusion may be reached by a hypothetical illustration:
Suppose the Ohio River from Pittsburgh to Cairo to be a closed condmt,
of cross-section equal to, say, the ordinary low-water section of the
open channel; and suppose that, by some mechanical process, sufficient
velocity could be imparted to the current in the conduit to carry away
all run-off as fast as it should present itself. It is manifest that the
discharge at Cairo during rising stages would be greater than with
an open channel, due to the fact that no storage had been allowed to
accumulate; but this difference would evidently disappear when the
maximum stage had been reached and there was no further accuraula-



tion of storage. At the critical point of the flood, therefore, the reser- Mr.
voir function of the channel ceases to have effect.

The writer feels the more freedom in- criticizing the Committee's
treatment of this matter because he is at the same time criticizing his
ovp-n previous utterances on the same subject.* The conditions are
abstruse and difficult to lay hold of definitely, but the writer believes
that they are fairly set forth in the following statement:

Storage in and adjacent to the channel of a stream, which results
automatically from an increase of stage, has little if any effect in
reducing the maximum stage. It operates to reduce somewhat the
rates of discharge during rising stages while storage is actively accumu-
lating, but this effect gradually disappears on approaching the maxi-
mum stage where storage accumulation ceases.

Myron L. Fuller,! Esq. (by letter). — The problem which most _^ Mr.
nearly attains successful solution, whether it be one of business,
science, or engineering, is that which takes into account, duly analyzes,
and gives proper weight to the greatest number of contributory factors.

Of the factors of flood control or prevention, geology is assuredly
one that must be reckoned with. The far-sightedness which is essential
to the satisfactory conclusion of any great undertaking demands that
every known factor of the problem be considered and the weight given
to it that mature investigation shows it has the right to claim.

Nothing is more true than the following statement by Morris
Knowles, M. Am. Soc. C. E., in the Minority Report of the Special
Committee. "It is unwise to assume that we have at any time obtained
all the knowledge possible upon a given subject." Every year sees
new advances in our knowledge of engineering problems; and our ideas
as to the precautions necessary in engineering undertakings are con-
tinually changing. Factors hitherto unknown or neglected are almost
daily found to be of importance.

Most of the great failures in engineering works, as well as in
other lines of business, have been due directly or indirectly to over-
looking some inconspicuous factor or neglecting some element of
apparent insigniflcance. For example, the lack of knowledge as to
the weakness and slipping planes in shales, or the failure to realize their
significance, has resulted in the failure of several large dams and reser-
voirs within the past few years, and the inadvertent omission of some
engineering factor has played an important part in the collapse of
bridges and other steel structures.

The writer does not maintain for a moment that the success or
failure of flood prevention or control measures is dependent on geo-
logical investigations, but he does contend that geology is one of

* "Flood Control, with Particular Reference to Conditions in the United States",
presented to the International Engineering Congress, San Francisco, September, 1915,
pp. 16 and 22.

t Boston, Mass.


Mr. the factors — sometimes one of marked importance — which should be
^""®'"- considered in any comprehensive investigation looking to the pre-
vention or control of floods. ■

Insufficiency of Precipitation Data.

Notwithstanding the great efficiency of the "Weather Bureau, rain-
fall data, chiefly because of the comparatively short periods during
which records have been kept, are admittedly insufficient for the
determination of the maximum possible or even probable precipitation
in a given area.

It is true that a certain approximation to the probable maximum
precipitation may be made from an inspection of the rainfall records
throughout the general province or climatological zone of which a
given district is a part, by applying the maximum precipitation re-
corded elsewhere to the area under consideration. As is well said in
the Report of the Special Committee (page 1220), however, "it is
improbable that the maximum precipitation that may occur at a
given locality has yet been recorded."

Taking the maximum rainfall recorded for the province or zone,
and the maximum known accumulation of snow for the same latitude
and altitude within the zone, it is possible to estimate the probable
height of a flood under existing conditions of channel. The height
calculated, however, is by no means free from likelihood of error,
and it is entirely possible for the water to rise many feet higher than
the estimated limits.

Geological Clues to Flood Heights.

Flood-Plains and Terraces. — In the absence of sufficient rainfall
data, it is fortunate that geology furnishes, in many instances, a
clue to the probable heights of flood-waters. The clue referred to is
that afforded by the flood-plains and terraces, which often give evi-
dences of heights of water greater than those which would be indicated
by the computations from greatest recorded rainfall.

Generally speaking, there is little evidence that floods are either
more or less pronounced and frequent at the present time than in
the immediate past. This, of course, does not apply to the more
remote geological past, when climatological conditions were radically

The evidence afforded by flood-plains is not altogether simple
or universally present. Mountain streams, having steep gradients,
are usually cutting rather than depositing, and flood-plains and
terraces are commonly absent, although even in mountainous regions
elevated pockets of sand often afford clues to the high-water limits,
In flatter regions, such as the greater portions of the Mississippi and
Ohio Basins, the streams are generally aggrading or filling their


valleys, at least locally, and are commonly bordered by flood-plains or Mr.



The flood-plains are fundamentally constructional in origin, being
built up of the relatively fine deposits laid down when the waters of
the streams overflow the normal channels and spread out with checked
velocity. The upbuilding continues, unless new geological or artificial
factors intervene, until the surface reaches nearly the level of the
highest floods. During the later stages, the upbuilding is very slow,
and the floods overtop the flood-plains only at long intervals, perhaps
of centuries. In the majority of localities the flood-plains have not
reached their final stages, especially where, because of constrictions
of the channels or other causes, the currents are unusually swift, but
are still overflowed to a depth of several feet by the greater floods.

The uncompleted stages of the flood-plains introduce large elements
of uncertainty, and the problem is still further complicated by the
fact that portions of the once completed flood-plains may have been
subsequently eroded away, leaving surfaces much lower than the
levels reached by the floods. In other words, a series of terraces is
left, only the uppermost members of which are of significance in
determining flood heights.

If this was all, the problem would still be comparatively simple,
but further difficulties are introduced by the fact that, in some in-
stances, the whole river valley has been cut in glacial or other sands
and gravels of a past geological period. In this down-cutting, terraces
similar in appearance to those of the flood-plains are commonly left
at various levels.

The determination of which of the general series of terraces
in a given valley are due to upbuilding, and are, therefore, of signifi-
cance in the determination of probable flood levels, and which are due
to down-cutting, and are without present significance, is often diffi-
cult, and the discrimination can seldom be left to the engineer with
safety, or even to the academic geologist of the average college faculty.
The discrimination can be made, but it requires a practical and expe-
rienced field geologist to make it with certainty.

The flood-plains and terraces, notwithstanding their complexities,
afford a line of evidence that should not be neglected in any com-
prehensive investigation of flood control. When the writer was engaged
in Ohio on Government geological surveys, ten years ago, the question
was several times asked in regard to towns or their water supplies:
"Is there any danger from floods?" The answer was that the towns
were situated on flood-plains built up by repeated overflows of the
streams, and that similar overflows were bound tc occur in the future
unless artificially prevented. The writer did not aecessarily expect


Mr. a flood of the proportions of that of 1913 so soon, but it was bound
^^' to come eventually.

Geological Obstructions to Streams.

Geological obstructions to streams, because of their obscurity, are
often overlooked — or their importance under-estimated — in flood-
control studies. Two distinct types may be mentioned: the boulder
delta and the rock barrier.

Boulder Delta. — Existing deltas or fans of boulders or coarse gravel
are manifest obstructions, and are recognized by engineers as such.
They occur chiefly where side streams having steep gradients issue
from hills or mountains into broader valleys where the streams are
more sluggish and are usually engaged in building up their flood-plains.
Often, the main streams are pushed aside and forced close against
the opposite banks, where they flow in comparatively narrow and
deep channels. They are obstructions which have an unmistakable
and calculable effect on flood height.

The deltas which are not pre-existing, on the other hand, introduce
important but unlaiown factors. A severe rain over the water-shed of a
small tributary is felt almost at once, and the stream commonly responds
long before the water rises very materially in the main valley. The
result is that enormous volumes of material are often swept into a
river, causing great accumulations of gravel or boulders where no
deposits may have previously existed. It is true that, as the main
river rises, many of these accumulations may be swept away, but, in
the meanwhile, they are a very real obstruction, with an important
influence on flood heights. Their final destruction may inaugurate
flood crests endangering the whole country below. Under other con-
ditions, the deposits may become permanent features, at least for a
considerable period of years.

The boulder or gravel deltas of this type are certainly worthy of
attention in flood-control studies, and geological investigations of their
possible effects are highly desirable.

Rock Barriers. — In many of the valleys in the Ohio and Mississippi
Basins there are buried rock channels much deeper than the general
rock floors over which the streams now flow. The latter commonly
mark the depth to which the scour of the river extends during flood
periods at the present time; the deeper channels mark the former
positions of the river under different conditions of level or discharge.
Sometimes, the present channels coincide with the older rock gorges,
in which case the scour may extend to somewhat greater depths.
Sooner or later, however, the streams have to return to the general
rock floors, usually accompanied by sudden rises in the bottoms.
This has an important effect in checking flow and in promoting


Although the importance of the rock barrier is less than that of Mr.
the delta, its influence is sufficient to warrant taking it into accoimt ^^'
in the flood-control problem.

Geological Factors of Run-Off.

The principal geological factors of run-off are: the grain of the
soil, the thickness of the soil, and the nature of the xm.derlying rock,
including composition and Assuring.

Grain of Soils. — The grain of soils, using the term "grain" in its
broader sense to include the arrangement and packing of the particles
as well as their size, is of the greatest importance in run-off and flood
studies of a given water-shed. A proper study of the grain includes
both laboratory tests and field examinations. The former consist of
sizing tests or analyses by passing the material through a series of
standard screens of various meshes, and from them may be computed
the rate of percolation under different conditions. Direct tests on
blocks of soil cut from the ground, where the material is sufficiently
firm to hold together, have been found, in the writer's experience,
to afford still more dependable results. All measurements or estimates
of rates of percolation apply, naturally, only to unfrozen soils, hence
are of value only in seasons where the ground is free from frost.

Classification of Soils. — Accompanying the percolation estimates
and tests mentioned — which are as much of an engineering as of a
geological nature — field examinations must be made, and the soils
classified, mapped, and their areas determined. These can be made
best by a geologist or soil expert. Samples, at the best, can be taken
at only a limited number of points, and must be supplemented by a
soil classification of the entire area, if the work is to have any real
value. The geologist's knowledge of the general character of residual
soils, or those resulting from the weathering of different types of
rock, of the nature of the various classes of glacial deposits, and of
the composition of the alluvial accixmulations, is of the greatest
assistance in this branch of subsidiary investigations in flood-control

Depth of Soil. — The depth of soil is of much greater significance
than is commonly realized. The writer's ground-water studies in some
thirty States have shown that rocks free from soil coverings seldom
contain much water, and that the same rocks, when covered by a
thick soil mantle, afford abundant supplies. Conversely, when the
soil is thin, the immediate run-off is much greater than where it
is thick. Therefore, geological studies to determine the depth of soil
are of considerable importance in rim-off studies.

Composition of Roch. — The absorptive capacity of the underlying
rock is a factor of significance in flood control. The porosity or
absorptive capacity varies greatly in rocks of different classes, ranging


Mr. from i% or less in quartzites to 5% in shales and limestones, and 25%
^'^' or more in sandstones. Many engineers can recognize the various
major varieties of rock, but the numerous errors in determination
which are indicated by the reading of engineering literature show
that it would be far safer to rely on a geologist than to leave rock
determinations to guesswork.

Fissuring of Bocks. — Fissures in rocks have been mentioned by
several engineers as a factor in natural flood regulation. Broadly
speaking, rock fissures are of four general types: cleavage planes,
bedding planes, joints, faults, or other fracture planes, and solution

The cleavage planes, limited chiefly to slates, and the bedding
planes of sedimentary rocks, are commonly potential rather than
actual fissures, and, even when open sufficiently to admit water, take
it up and deliver it very slowly. They affect the flood heights but
little, their office being that of feeders and regulators of low-water

The openings along joint or fault planes are rarely more than a
fraction of an inch in diameter, and although of more importance than
cleavage and bedding planes, they are likewise regulators of ground-
water flows in low-water stages, rather than factors in floods.

The solution passages in limestones, on the other hand, are often
many feet in diameter. The Mammoth Cave of Kentucky, with its
large passages several miles in length, is an example of this type.
It is now abandoned by the stream that formed it, but, elsewhere,
similar channels are still occupied by underground rivers. The writer
has seen such a stream on the outskirts of the Ozarks, in Missouri,
where a torrent several feet in width and a foot or more in depth
rushed along underground with rapids and falls like those of a
surface stream.

Such solution passages are natural tunnels capable of carrying
immense volumes of water. Nevertheless, notwithstanding their fre-
quency and large capacity, they are rarely elements of importance
in floods, as they replace rather than supplement the normal surface
drainage. Practically all of them are above the normal drainage
levels of the main streams, or those in which floods are likely to occur,
and lead from limestone sink-holes or other openings in the hills down
to the main valleys which determine the levels of their outlets. Al-
though they can lead water down to a flooded stream, they can seldom,
in the absence of lower outlets, take it away.

If clogged, a certain quantity of water is sometimes held back,
but this is often compensated by the breaking loose of obstructions
previously formed in other passages, and the setting free of waters
to be added to the flood.


It is to be noted, however, that in a few localities streams of con- Mr.
siderable size are kno-mi to flow at heights of several hundred feet
above adjacent streams. If in a limestone region, a solution passage
might take away a large proportion, if not the entire volume, of the
higher stream, but the quantity taken from the upper stream would
be added to that of the lower, where the flood conditions might be
even more serious.

It is apparent, therefore, that solution passages may be a real
factor in floods, and, though not a common element, are sufficiently
frequent to demand recognition. In storage reservoirs, where the
water is held above both the normal drainage and ground-water levels,
they are often of paramount importance, and great precautions are
necessary to prevent damage or loss through their agency. The reser-
voir failure at Johnson City, Tenn., on January 10th, 1913, is an
example of failure from neglect of solution passages.

Geological Factors of Ground-Storage.

The storage capacity of the sands and gravels of alluvial deposits
adjoining a stream is a factor of considerable importance in flood
regulation. The pore spaces of such materials commonly range from
25 to 35% of their volume, and although the residual moisture already
present may reduce this very materially, there is usually an available
porosity of from 10 to 25 per cent. In other words, 10 ft. of gravel
will take up the equivalent of from 1 to 2^ ft. of water. Of course,
if the rise of the flood is slow, the gravels will often fill with ground-
waters from the hills, but, in sudden floods, the ground-waters have
insufficient time to respond to change of level, and adjacent sands
and gravels will become filled from the river. The inflow into the
gravels is often quite rapid. In one case observed by the writer a
cellar 100 ft. from a stream subject to sudden fluctuations was filled
to stream level with a lag of less than 2 hours.

It is apparent, therefore, that, under favorable conditions, the
storage capacity of adjacent deposits may be a prominent factor in
floods, and geological studies, to determine their location, extent, and
availability, are highly desirable. In such studies, the aim should be
to determine, in addition, the presence or absence of buried rock
masses which might limit the storage capacity of a particular forma-
tion. Buried barriers of rock opposing the free movement of the
underflow and bringing it to the surface to be added to the floods
should likewise be located whenever possible. The position of buried
channels, if present, should be determined, for they may provide
outlets for the imderflow at high-water stages when not available
at low stages, or may at least afford additional storage.


Mr Limited Value of Sedimentation Basins.


Moving water will always carry its full load of sediment if the
material is available. It is a geological principle that, if it does not
already have such a load, a stream will pick it up from its bed or
banks until it is fully loaded. The greater portion of the coarse
material under transportation by a stream is carried only a short
distance. It is picked up where the current is accelerated, the scour
often extending to depths of 30 ft. or more, but it is deposited in
large measure as soon as the current is checked. Only the finer and
relatively harmless silts become a permanent addition to the water.
In the course of time, by innumerable repetitions of the process
described, the coarser material is carried to points far distant from
its source, but very little except the finer silts is transported to
any considerable distance during a single flood.

For these reasons, it appears that, as a general rule, sedimentation
basins may be expected to be of limited value, although, as local pro-
tection, they may occasionally be entirely successful.


A levee system, by confining the current of a stream, tends to
intensify the scour, and, for the same reason, subsequent deposition
may be hindered. In the long run, river bank erosion is reduced,
and such deposition as occurs is probably less harmful than along
an uncontrolled channel.


The value of natural or artificial outlets depends to a considerable
extent on geological conditions. An outlet may present a shorter
line of discharge and a straighter course than the existing channel;
but, unless the water can get into it freely, it is not likely to be
utilized by a stream. The inlet to it must be free from obstructions
and in line with the natural swing of the current, and the banks must
not cave (in proportion to the volume of water) more than those
along the existing outlet.

If any considerable volume of water follows a new outlet, tempo-
rary relief is, indeed, afforded, but the improvement is not permanent.
Either the new or the old channel will soon silt up, and the final
condition will be no better than before. The Hwang-Ho, in East
China, shifted its mouth a distance of 275 miles in 1852, but flood
conditions are said to have been accentuated rather than relieved.

Necessity for More Comprehensive Studies of Streams.

The Government, in connection with its comprehensive and detailed
studies of the Mississippi and its larger tributaries, has secured very


complete data on erosion, transportation, and deposition by low- Mr.
gradient streams, especially in their lower reaches. There is, however, ^ ^^'
a deficiency of data pertaining to the upper reaches and at the head-

It is evident that the conditions in the upper portions of most
large streams are entirely different from those in the lower portions.
In the latter, notwithstanding the wholesale caving of the banks,
the excavation of cut-offs, etc., deposition predominates, and stream
beds, flood-plains, and estuarine deposits are being gradually built up.
On the middle reaches erosion and deposition are more nearly balanced,
and the conditions as regards flood-control are essentially different.

Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 109 of 167)