Frederick B. Lotspeich
Eldor W. Schallock
Working Paper No. 27
June 1974
Associate Laboratory
U.S. Environmental Protection Agency
National Environmental Research Center
Office of Research and Development
Corvallis, Oregon

A Working Paper presents results of investigations which are, to some
extent, limited or incomplete. Therefore, conclusions or recommendations
expressed or implied are tentative. Mention of commercial products or
services does not constitute endorsement.

4	The CO2 sampling technique in action under winter
5	Employment of equipment on hand to remove a frozen
6	A core of frozen gravel removed from under the ice
by the COg freezing technique.
1	Equipment necessary for winter use of CO2 sampling
2	All equipment needed for this procedure includes
sampling gear, ice auger, "Akhio" and snow machine.
3	A core removed from under about 5 feet of ice with
flowing water between the ice and top of sediment.	6
A core of sand and fine gravel obtained by CO?
freezing.	IJ
A core about 3 feet in length representing a
wide sedimentary range, from fine gravel to
pebble size particles.	14
Particle size versus accumulative percent for
several representative cores collected during late
winter of 1973.

Any study of streams, whether the emphasis is on biology or stream
dynamics, must consider the sediments because they are of vital importance
in understanding many stream phenomena. Phillips (1973) clearly points
out how stream substrates are important to salmonidSj a group requiring
clean gravel as spawning areas. In many of Alaska's streams, such gravel
substrates predominate.
Sand-bottomed streams have been studied in temperate climates and
much of their dynamic nature is understood; however, gravel-bottomed
streams have been somewhat neglected. Although it has long been observed
that salmon spawning requires a gravel bottom, little quantitative data on
why and how gravel substrates influence productivity are available. This
absence of data is at least partly caused by a lack of accurate substrate
sampling procedures. Development of such procedures must precede more
fundamental studies of sediment-biology interrelationships. The ideal
sediment samples would be truly representative of that portion of the stream
from which it is extracted. Several sampling techniques, from narrow spades
to large cylinders inserted in the gravel with hand-picking in place, have
been used, but all result in inaccurate samples as much of the fine material
is lost.
An apparently superior method for sampling sediments has been developed
and described by Walcotton (1973). Basically this method consists of ex-
tracting a frozen core of in situ material from the depth to which the
sampler can be inserted. A rigid copper pipe is forced into the stream
bottomt then a smaller open-ended flexible copper tube is placed inside
the longer pipe. Liquid C02 is injected through this smaller tube to the
bottom of the larger one. When the CO2 is injected; it cools the outer

tube rapidly, freezing the adjacent sediments in position. All material
within the frozen core is retained, including water, all sizes of sediment,
and occasionally living organisms.
A similar procedure, using liquid nitrogen as the cooling agent, has
been described by Stoker and Williams (1972). (This method avoids occa-
sional clogging by solid C02,) These workers also used a power winch to
remove the frozen core from the stream bed, a practice we have considered
but not tested.
Since most streams of Alaska are ice-covered for a large part of the
year, this procedure was tested under winter conditions to fulfill a need
in an existing project. The equipment was fabricated and the procedure
tested during the summer of 1972. It operated just as described, with the
limitations and capabilities noted by.Walcotton. This report describes
the equipment, procedures, modifications and results of winter use.

Our original equipment followed Walcotton's design closely because of
frequent communication with him during his early testing of the technique,
and was found satisfactory for summer use or under non-ice conditions. How-
ever, for winter use several modifications were made on his original design.
Because longer tubes were required, the male portion of a brass union was
/ brazed to the upper end of the pointed tube; the other part of the union was
brazed to an 8-10 inch tube, which ended in a "T" (Figure 1). Intermediate
sections of the same tubing (about 4 feet in length) were provided with
similar unions. This permitted the length of the sampler to be quickly
extended, in 4-foot sections, when a longer tube was needed for use under
thick ice and/or deep water. The inner tube was also lengthened but without
sectioning. All the equipment was easily carried on an "Akhio" sled towed
by a snow machine. Figure 1 shows a general view of the entire set of samp-
ling equipment situated in position at the hole,which was bored by the ice
auger shown in the background. The ice here was 3 to 4 feet thick. Figure
2 shows the snow machine with the "Akhio" in the foreground and the sampler
emplaced prior to freezing. Figure 3 shows the core extracted and the ice
coated union joining the two sections.
These modifications enabled sampling under, ice but did introduce some
minor problems which effected slight procedural changes from those suggested
by Walcotton. A new problem was the increased flexibility of the outer,
rigid copper tube when it was lengthened to 8 or 12 feet. This prevented
putting much force on the sampler during emplacement; hence, it could not
penetrate consolidated sediments. However, no problem was found in samp-
ling freshly deposited, unconsolidated sediments even when three sections

Figure 1. Equipment necessary for winter use of CO2 sampling procedure.

Figure 2. All equipment needed for this procedure includes sampling
gear, ice auger, "Akhio" and snow machine.

Figure 3. A core removed from under about 5 feet of ice with flowing water
between the ice and top of sediment. Total depth of sediment, water, and
ice is about 7 feet.

of rigid tubing were used. Walcotton suggests that 2 or 3 samples could
be frozen with 15 lbs of C02 however, to reliably get a sample under ice
and greater water depth with the extended tubes required more CO2. After
some experimentation, including several failures to get a sample, the
procedure finally developed was to expend one 15 lb. bottle of CO2 per
sample. Such a rate of expenditure of gas may appear somewhat extravagant
but is actually inexpensive when all other costs are considered. Figure 4
shows the sampler in operation using about 8 feet of rigid tubing. Notice
here that the inner flexible tube extends several feet beyond the vertical
tube. This additional portion permits the same inner tube to be used when
another section of rigid tube is needed; such use of one long tube eliminates
the need to carry inner tubes of several lengths. Note also that the opera-
tor is holding the inner tube with his hand instead of tying it as suggested
by Walcotton; either way is satisfactory, but securing of this tube is a
requirement or it will blow out of the sampling tube when CO2 is expended.
Once the core is frozen in place, it becomes something of a problem
to extract it from the stream bottom. Several approaches are available.
For summer use and with short cores, the core was simply pulled out by
lifting with bent arms under the "T" of the tube. This requires some
vigorous pulling and the risk of a wrenched back is always present, espe-
cially irr uncertain footing. Walcotton is presently experimenting with a
tripod with a light weight block-and-tackle. Figure 5 shows a method using
the equipment normally carried on such a sampling trip. The shank of a
heavy ice chisel was lashed to the llT" handle and used the C02 cylinder
as a fulcrum of a second class lever. Such an arrangement gave substan-
tial mechanical advantage and enabled safe removal of several long cores
when straight lifting was unsuccessful. When taking core samples during

Figure 4. The COg sampling technique in action under winter conditions.
Expenditure of CO2 in this scene is probably more than required but, to get
a reliable core, we found that a rapid rate of CO2 emission produced maximum
length and diameter of cores.

Figure 5 Employment of equipment on hand to remove a frozen core. Note
the water over the ice, a common occurrence when holes are bored in river
ice and wnich contributes to uncertain footing (later winter, 1973).

warmer weather, they must be removed within a few minutes or the water
may thaw the material immediately adjacent to the tube and it may slip
out without core.
In winter, cores are easily preserved in their frozen state, can be
handled while on the tube, and transported to the laboratory for further
processing. Once in the lab the tube can be filled with warm tap water
whtch quickly thaws the core around the copper tube and it is easily slipped
out. The intact core can then be allowed to thaw or be stored in a freezer
for future work. Longer cores can be cut into sections where definite changes
of material occur, placed in plastic bags, and stored frozen until particle
size analysis can be run on all cores collected during the winter season.
The analyses made of cores taken during our evaluation phase consisted of
sieving through a set of sieves using the international system with square
holes based on 1 mm as; a standard. Sizes ranged downward by halving each
sieve to .062 mm and upward by doubling to 16 mm. A final screen of 22 mm
was added to collect the larger gravel. Sieving was by a hand-operated
mechanical shaker secured to a lab bench top.

Three cores, illustrating a range of particle-size distribution, are
presented in Figures 6, 7, and 8. It is obvious that the material in
Figure 6 is principally gravel that appears to be uniformly distributed
with depth. The cores shown in Figures 7 and 8 are composed of much smal-
ler particles, with Figure 7 showing very uniform distribution with depth
until a gravel substrate is encountered. Figure 8 shows a mixture of
material with the bulk of material in the sand and fine gravel size range.
Such a view showing the range in size of material that is reliably sampled
by this technique illustrates the usefulness of this method and permits
field interpretation of changes in substrate with depth and position.
Figure 9 is a graphical portrayal of particle-size distribution from
a sieve analysis of the cores shown in Figures 6 (core #30), 7 (core #29),
and 8 (cores #27A and 27B). In this Figure, size is given on the "x" axis
on a logarithmic scale, and quantity as percent accumulation on the "y"
axis. Such a diagram clearly shows that the appearance of the cores, as
removed from the stream bed, reflects differences in size distribution.
The curve for core #30 is significantly different from the other three and
contains no material finer than 0.125 mm in diameter. Moreover, most of
the material (71 percent) is larger than 4 mm.
The curve for core #29 is quite different from that for core #30 and
verifies the uniform appearance of the core materialas removed from the
stream bed. Only about 22 percent of this core is greater than 8 mm in
diameter and remains low (a total of 28 percent) down to 0.5 mm. The bulk
of the material in this core, about 78 percent, is <0.5 mm in diameter.
Curves for cores #27A and 27B were taken from the core shown in
Figure 8. This core was divided into two sections after being trans-

Figure 6. A core of frozen gravel removed from under the ice by the C02
freezing technique.

Figure 7. A core of sand and fine gravel obtained by C02 freezing. Note
here the pebbles at the bottom; this represents dramatic changes in sedi-
mentation from a stable gravel substrate to a more ephemeral rapid
sedimentary regimen (length of core is about 15 inches).

Figure 8. A core about 3 feet in length representing a wide sedimentary
range, from fine gravel to pebble size particles.

ported and stored in the lab. Such sectioning illustrates another ad-
vantage of this technique; it permits quantitative analysis of portions
of the substrate without mixing of adjacent sections. Although these
curves are similar in shape, the material in #27A is clearly,much finer,
with 97 percent being finer than 1 mm and 57 percent finer than 0.25 mm
1n diameter. The material in core #27B is coarser with 28 percent less
than 0.5 mm in diameter.
This freezing procedure permits a nearly undisturbed sample to be
withdrawn from a stream bed and allows the investigator to make prelimi-
nary field Interpretation at that time. It also permits quantitative lab
analysis of an entire core or individual section where discontinuities
occur, with reliable assurance that contamination by adjacent sections or
loss of some component is negligible. Both the field interpretation and
interpretation of graphs such as Figure 9 enable an investigator to draw
important conclusions about the characteristics of the material whether it
is a biological or river mechanics study. A substrate such as that shown
by the curve for core #29, will be vastly different in physical properties
from that shown in the other three curves and may greatly affect the bio-
logy of a stream.
The freezing technique reported here allows a core to be taken at
least 84 cm long weighing at least 900 grams. These samples can be con-
sidered quite representative of the actual bed material because of minimal
disturbance during sampling activities, although there is no such thing as
absolute non-disturbance.
As the length of the sampling tube increases, the flexibility also
increases which may prevent penetration into compacted material. However,
for the unconsolidated material sampled here, a tube length of 10 to 12
feet can be used. More work is needed on the sampling of consolidated

10	/	0.1
size distribution (mm)
Figure 9. Particle size versus accumulative percent for several
representative cores collected during later winter of 1973.

sediments using this procedure. Removal of the frozen core also introduces
problems but this can be overcome by mechanical devices such as a tripod
and block-and-tackle.
A final point regarding any substrate sampling procedure or program;
1t becomes highly desirable, even mandatory, that the program be under the
field supervision of the most qualified person available. This will de-
crease the total number of samples, while improving their quality as they
better represent a given physical or biological parameter. No amount of
subsequent analyses, whether in the field or in the lab, will enhance the
final outcome of a program if the samples do not represent what is under
study. The highest qualified technician is seldom capable of directing
the location of sampling stations simply because he does not have the
training or experience to understand the systems under study. It may seem
uneconomical to have the highest paid or valuable person directing field
sampling, but it will result in higher reliability and credibility of re-
sults. Optimum sampling sites are best selected by a person who thoroughly
understands the interrelationships of the system under study. Once the sites
are selected, this freezing procedure should aid the investigator in getting
representative samples from which to get reliable analyses.

Phillips, Robert W., 1970. Effects of Sediment on the Gravel Environment
and Fish Production. Proceedings of Symposium on Forest Land Uses and
Stream Environment, Oregon State University Press, Corvallis, October
19-20, pp. 64-74.
Stoker, Z. S. J., and Williams, D. Dudley, 1972. A Freezing Core Method
for Describing the Vertical Distribution of Sediments in a Streambed.
Limnol. Oceanogr. 17:136-138.
Walcotton, William J., 1973. A Freezing Technique for Sampling Streambed
Gravel. U.S. Forest Service Research Note, PNW-205, Portland, Oregon,
7 pp.