See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/253991544

Measuring Ice Lens Growth and Development of Soil Strains during Frost
Penetration Using Particle Image Velocimetry (GeoPIV)

Article · January 2008

CITATIONS READS

17 913

4 authors:

Tezera Firew Azmatch Lukas U. Arenson

SNC Lavalin BGC Engineering Inc.
10 PUBLICATIONS 298 CITATIONS 104 PUBLICATIONS 3,488 CITATIONS

SEE PROFILE SEE PROFILE

David Sego Kevin W. Biggar

University of Alberta K.W. Biggar Engineering Consulting
201 PUBLICATIONS 4,407 CITATIONS 84 PUBLICATIONS 1,276 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Geotechnical Study of Oil Sands Dewatering using Thin Layer Freeze-Thaw View project

https://era.library.ualberta.ca/items/0a0ac1f8-5159-43cf-a0a9-9d9b0b2f8483 View project

All content following this page was uploaded by David Sego on 02 July 2014.

The user has requested enhancement of the downloaded file.

 Measuring Ice Lens Growth and Development of Soil Strains during Frost
Penetration Using Particle Image Velocimetry (GeoPIV)
Tezera Firew Azmatch
UofA Geotechnical Centre, Department of Civil and Environmental Engineering, University of Alberta, Edmonton, AB,
Canada
Lukas U. Arenson
BGC Engineering Inc., Vancouver, BC
David C. Sego
UofA Geotechnical Centre, Department of Civil and Environmental Engineering, University of Alberta, Edmonton, AB,
Canada
Kevin W. Biggar
BGC Engineering Inc., Edmonton, AB

Abstract
The formation of ice lenses and the water migration during freezing of frost susceptible, fine-grained soils is a dynamic
and complex process. Horizontal ice lenses and vertical ice veins form at different spacing, intervals, and growth rates
as the pore freezing front penetrates into the ground forming a three dimensional ice lens pattern. One-dimensional
step-freezing tests were carried out in a transparent freeze cell with frost susceptible Devon silt to study this effect
in detail. Time-lapse photography and particle image velocimetry are used to identify the dynamics of the freezing
processes. Three dimensional movements are used to measure the changes in ice lens growth and axial soil strain in-
between ice lenses with time. The formation and growth rate of vertical ice veins are studied and local vertical strains
are determined. Horizontal soil deformations are also investigated. The detailed strains and ice lens growth rates within
freezing soils help in improving formulations for frost heave, strength, and deformation behavior of freezing and frozen
soils.

Keywords: frost heave; frost susceptibility; ice lens formation.

Introduction ice veins form as the soil freezes. The vertical ice veins
develop a hexagonal crack pattern similar to that observed
The formation of ice lenses and the water migration during
in drying soils (Arenson et al. in press).
freezing of frost susceptible, fine-grained soils is a dynamic
The reticulate ice structure, also observed under natural
and complex process. Horizontal ice lenses and vertical ice
freezing conditions (Mackay 1974), during freezing
veins form at different spacing, intervals, and growth rates as
deforms the soil horizontally and vertically. When the final
the freezing front penetrates into the ground forming a three
ice lens starts to form, part of the soil below the final ice
dimensional ice lens pattern. This pattern and the growth
lens consolidates, and part of the soil above the final ice
rates of the ice lenses strongly depend on the freezing
lens heaves. Arenson et al. (2007) measured the vertical
boundary conditions, such as temperature gradient, cooling
deformations in freezing soil and were able to explain the
rate, or vertical pressures, and the characteristics of the soil,
freezing process using these deformation results.
i.e. the grain size distribution, hydraulic conductivity, water
This work was continued, and this paper focuses on the
availability, and pore water salinity (Konrad & Morgenstern
formation of the vertical ice veins and the horizontal soil
1982, Konrad 1994, Miller 1973, Penner 1972).
deformations. Results from the time-lapse photography
The formation of ice lenses is possible with the presence
presented by Xia et al. (2005) are re-analyzed using particle
and migration of unfrozen water at temperatures below its
image velocimetery (PIV) to develop a better understanding
freezing point (Konrad and Morgenstern 1980). It is believed
of the formation and growth of ice structures.
that migration of the free water is induced by a temperature
gradient within the soil that induces suction. The suction
then drives the movement of free water to the growing ice Laboratory Tests
lens. The flow of water towards the warmest ice lens is One-dimensional open system (access to water) step-
assisted by the formation of tension cracks, which increase freezing tests were carried out on Devon silt under different
the vertical hydraulic conductivity of the soil (Chamberlain boundary conditions. The Devon silt has a liquid limit of
& Gow 1979). 32%, plastic limit of 20%, and specific gravity of 2.65.
Freezing tests conducted at the University of Alberta The samples were frozen from the top downward. Two
(Xia 2006) have shown that ice lens growth is not a one- temperature baths control the temperature conditions at the
dimensional process. Horizontal ice lenses as well as vertical top and the bottom of the sample to establish one dimensional

89
90   Ninth International Conference on Permafrost

Table 1. Test conditions.

Test # T1grad S2 σ3
(°C/m) (g/L) (kPa)
1 58 0 0
4 58 0 100
8 58 10.2 0
1
Temperature gradient at thermal steady state
2
Salinity (NaCl)
3
Vertical pressures during freezing

Image Analysis using GeoPIV

The digital images were analyzed using GeoPIV software,
a MatLab module which implements particle image
velocimetery (PIV) in a manner suited to geotechnical testing.
Figure 1. Original picture (left), and visible ice lens structure This code has been programmed and successfully used for
(right). geotechnical laboratory testing to measure deformations in
soils (White et al. 2003). It has also been used to determine
the deformations in freezing soils (Arenson et al. 2007).
Figure 1 shows the original image and the corresponding
200 1 2 3 4 5 6 ice lens structure for one of the images.
GeoPIV was used to measure the soil deformation. Geo
PIV uses image texture to follow patches over a time series
400 of pictures. A number of patches were defined on the sample
7 8 9 10 11 12
and their movement was followed over time (Fig. 2).
600
13 14 15 16 17 18
Test Results and Image Analysis Results
800
Three tests from Xia et al. (2005) were analyzed in this
paper. These tests were chosen because of their well defined
1000 19 20 21 22 23 24
99 vertical ice veins. The freezing test conditions are shown in
Table 1. All the samples were initially consolidated at 100
1200 kPa vertical pressure.

1400 Ice Lens Growth and Deformation

The reticulate ice structures and cross sections of the
1600 25 26 27 28 29 30
frozen samples showing the hexagonal ice vein pattern are
presented in Arenson et al. (in press). Figures 3 through 10
1800 31 32 33 34 35 36
100
show various deformation results.

2000
200 400 600 800 1000 1200
Figure 2. Initial patch locations; the scale is in pixels (1 pixel ~10
µm).
vertical freezing from the top downward.
A fluorescent tracer (C20H12O5), which appears green in
unfrozen water but colorless in ice under ultra violet light
(Arenson & Sego 2006), was used to determine the frozen
and unfrozen zone of the sample during freezing.
Time-lapse photography provided digital photo records
to visually observe and document the freezing process.
Details of the time-lapse photography technique and of the
laboratory freezing tests used herein are presented in Xia et
al. (2005) and Xia (2006).
Vertical deformations
In the plots for vertical deformations (Figs. 3, 7)
positive values indicate heave and negative values indicate
consolidation.
It can be seen from Figure 3 (Test #1) that heaving starts at
about 42 hours after the start of the test. By the time heaving
has started the consolidation process is nearly complete.
This means that heaving and consolidation are not occurring
in parallel. Heaving starts almost after consolidation is
complete. The heave rate and the consolidation rate are
constant.
Figure 3 also indicates that P10 is undergoing heaving,
and P22, P24, and P25 are consolidating. Hence, row
2 (containing patches P7–P12) is heaving, and row 4
(containing patches P19–P24) is consolidating. The final ice
lens for this test is located between row 2 and row 4. This
 Azmatch et al.   91  

80 100

60
80

Horizontal deformation (pixel)

Vertical deformation (pixel)

40 P23
P 10
20 60 P21
P 22
0 P15
P25 40
-20 21 41 61 81 101
P24
-40 20
-60
0
-80
-100 -20
Tim e (h) 5.5 6 6.5 7 7.5
Tim e (h)

Figure 3. Vertical deformations for Test #1. P10 is located above the Figure 4. Horizontal deformations for Test #1.
final ice lens. P22, P24 and P25 are located below the final ice lens.

90 10
80
0
Horizontal deformation (pixel)

70 Vertical deformation (pixel)

60
P 22 -10 P3
P 24
50 -20 P4
40 P5
-30
30
P14
20 -40
P17
10 -50
0
-60
-10
21 41 61 81 101 -70
Tim e (h) 22.65 32.65 42.65 52.65 62.65 72.65
Tim e (h)
Figure 5. Horizontal deformations while a final vertical ice vein is Figure 6. Horizontal deformations while an intermediate vertical
being formed for Test #4. ice vein is being formed for Test #4.

140 80

120 60
Horizontal deformation (pixels)

Horizontal deformation (pixel)

100 40 P18

80 P12 20 P16

P18 P13
60 0
P17 22.65 24.65 26.65 28.65 30.65
40 -20
P10
20 -40

0 -60

-20 -80
9 9.5 10 10.5 11 11.5 12 -100
Tim e (h)
Tim e (h)

Figure 7. Vertical deformations while the final ice lenses are being Figure 8. Horizontal deformations while a final vertical ice vein is
formed for Test #8. P3, P4 and P5 are located above the final ice being formed for Test #8. P13 is located to the left of the vertical ice
lens. P14 and P17 are located below the final ice lens. vein. P16 and P18 are to the right of the vertical ice vein.
 92   Ninth International Conference on Permafrost
80
60
Horizontal deformation (pixel)
40
20
0
-20
-40
-60
-80
-100
22 32 42 52 62
Tim e (h)
Figure 9. Horizontal deformations while a final horizontal ice lens
is being formed for Test #8. P13 is located to the left of the vertical
ice vein. P16 and P18 are to the right of the vertical ice vein.
explains that water is being drawn from the region below the
final ice lens. This water is then used to form the final ice
lens. As the final ice lens forms, the regions above it start to
heave, as observed by the movement of P10. Similar trends
have also been observed by tracing other patches in the same
test.
In Figure 7, all the patches are consolidating during
the first 42 hours. After 42 hours, patches P3 and P4 start
heaving, and patches P14 and P17 do not deform anymore.
The final ice lens for this test started forming after about
42 hours. This indicates that the development of the final
ice lens, hence frost heave, starts after the consolidation is
complete. The final ice lens for this test is located between
row 2 (containing patches P7–P12) and row 3 (containing
patches P13–P18)
Figures 3 and 7 also indicate that the rate of heaving is
smaller than the rate of consolidation. It can be concluded
that the rate of ice formation is lower than the rate of water
extraction from the soils.
Horizontal deformations
In Figure 3, it can be seen that heaving starts around 42
hours after the start of the test. This implies that the final
horizontal ice lens starts to form after 42 hours. In Figure
4, it can be noted that the horizontal strains remain constant
while the final ice lens is being formed. The same observation
can be made in Figures 5, 6, and 8. This implies that the
horizontal strains are due to the formation of the tension
cracks and not due to the formation of the final horizontal
ice lenses. The horizontal strains could therefore be assumed
to originate from the formation of the vertical tension cracks,
which are a result of the suction created by the temperature
gradient. According to the authors, the mechanism of the
formation of the vertical ice veins can be described as
follows: First, suction is created in the frozen fringe, and
tension cracks form; these tension cracks are then filled
with water being removed from the soil, and finally, part of
the water in the tension cracks freezes to form the vertical
ice veins and part of it moves up to form the horizontal ice
10
9
Depth of vertical ice vein (mm)
8
7
P13 Test #4 IIV
6
P18
5 Test #4 FIV
P16
4 Test #1 FIV
3 Test #8 FIV
2
1
0
72 0 2 4 6 8 10
Tim e (h)
Figure 10. Rate of growth of the vertical ice veins (FIV=Final Ice
Vein, IIV=Intermediate Ice Vein).
lenses behind the vertical ice veins. The formation of the
tension cracks facilitates the flow of water by increasing the
hydraulic conductivity of the soil. It is observed that tension
cracks are formed prior to the formation of horizontal ice
lenses (Arenson et al. in press).
While the soils are straining horizontally, water is being
drawn to form the vertical ice veins and the horizontal ice
lenses above their crack tip. Mackay (1974) states, “The
vertical and horizontal ice veins are believed to have grown
in shrinkage cracks with much of the water being derived
from the adjoining clay.”
In most of the tests, the magnitude of the horizontal
displacement is between 60 and 80 pixels (approximately
0.6–0.8 mmm). These could indicate that the thickness of
the vertical tension cracks formed is almost constant. It can
be observed from the digital images that the thickness of the
vertical ice veins is the same for almost all the tests.
From Figures 8 and 9, it can be observed that the
movements of P13 are in the opposite direction to that of P16
and P18. This is because the patches are located on different
sides of the vertical ice vein. P13 is located to the left of the
vertical ice vein, and P16 and P18 are located to the right of
the vertical ice vein. Hence, their corresponding horizontal
displacement is in a different direction but it displaces the
same amount, indicating the opening of the vertical crack
and the formation of the ice vein.
Rate of growth of vertical ice veins
Tension cracks always formed before the horizontal ice
lenses. These tension cracks, which facilitate the flow of
unfrozen water, are then filled partly by vertical ice veins.
Figure 10 shows the rate of growth of the vertical ice
veins. The final ice veins (FIV) are defined as the vertical ice
veins that form immediately before the start of the formation
of the final horizontal ice lens. The intermediate ice veins
are the vertical ice veins that form before the FIV. The rate
of growth of the final vertical ice veins ranges from 1.22
mm/h to 1.57 mm/h. For Test #4, the rate of growth of the
intermediate ice vein is 5.78 mm/h. Hence, the rate of growth

Table 2. Depth and growth rate of vertical ice veins.
Test # Height of vertical Growth rate
ice vein (mm) (mm/h)
1 FIV 8.60 1.22
4 FIV 8.67 1.37
IIV 8.60 5.78
8 FIV 7.91 1.57
of the IIV is higher than the rate of growth of the FIV.
The height of the vertical ice veins for the tests under
consideration is about the same for all tests.
Conclusions
Particle image velocimetry (PIV) was used to measure
soil deformations. Patches with a distinct characteristic were
followed as the ice lenses formed and grew. The deformations
measured were used to explain the freezing process and the
resulting ice being formed.
The heaving process and the consolidation process were
observed to take place at different times. Heaving followed
immediately after consolidation was complete.
The horizontal strains were observed to be a result of the
formation of the vertical tension cracks as the freezing front
advanced.
The suction produced as a result of the temperature gradient
caused tension cracks in the soil. The tension cracks, which
increase the vertical permeability of the soil, were then filled
with water. A channel formed where water can migrate from
the unfrozen zone to the warmest ice lens. Part of the water
moved up enhancing the development of a horizontal ice
lens above the crack, and part of it froze inside the cracks,
leading to the formation of the vertical ice veins.
The vertical tension cracks always formed before the
horizontal ice lenses formed. Thus, the formation of the
tension cracks, and hence the tensile strength of the soils,
play a very important role in the freezing process.
Acknowledgments
The authors would like to thank Dr. Andy Take from
Queens University for his support with the GeoPIV analysis.
Tezera Firew Azmatch appreciated the funding through the
NSERC Discovery Grant held by Dr. Sego.
References
Arenson, L.U., Firew Azmatch T. & Sego, D.C. 2008. A new
hypothesis on ice lens formation in frost susceptible
soils. Proceedings of the Ninth International
Conference on Permafrost, Fairbanks, Alaska, June
29–July 3, 2008 (this proceedings).
Arenson, L.U., Sego, D.C. & Take, W.A. 2007. Measurement
of ice lens growth and soil consolidation during frost
penetration using particle image velocimetry (PIV).
60th Canadian Geotechnical Conference, Ottawa,
ON, October 21–24, 2007: in press.
View publication stats
Azmatch et al.   93  
Arenson, L.U., Xia, D., Sego, D.C. & Biggar, K.W. 2006.
Change in ice lens formation for saline and non-saline
Devon silt as a function of temperature and pressure.
13th International Conference on Cold Regions
Engineering. American Society of Civil Engineers
(ASCE), Orono, Maine, USA: CD-Rom.
Arenson, L.U. & Sego, D.C. 2006. The effect of salinity
on freezing of coarse grained sands. Canadian
Geotechnical Journal 43(3): 325-337.
Chamberlain, E.J. & Gow, A.J. 1979. Effect of freezing and
thawing on the permeability and structure of soils.
Engineering Geology 13(1-4): 73-92.
Konrad, J.M. 1989. Pore water-pressure at an ice lens:
Its measurement and interpretation. Cold Regions
Science and Technology 16(1): 63-74.
Konrad, J.M. 1994. 16th Canadian Geotechnical Colloquium
- Frost Heave in Soils - Concepts and Engineering.
Canadian Geotechnical Journal 31(2): 223-245.
Konrad, J.M. & Morgenstern, N.R. 1980. A mechanistic
theory of ice lens formation in fine-grained soils.
Canadian Geotechnical Journal 17(4): 473-486.
Konrad, J.M. & Morgenstern, N.R. 1982. Effects of applied
pressure on freezing soils. Canadian Geotechnical
Journal 19(4): 494-505.
Mackay, J.R. 1974. Reticulate ice veins in permafrost,
Northern Canada. Canadian Geotechnical Journal
11(2): 230-237.
Miller, R.D. 1973. Soil freezing in relation to pore water
pressure and temperature, Second International
Conference on Permafrost. National Academy of
Sciences, Washington, Yakutsk, Siberia: 344-352.
Nixon, J.F.D. 1991. Discrete ice lens theory for frost heave
in soils. Canadian Geotechnical Journal 28(6):
843-859.
Penner, E. 1959. The Mechanism of Frost Heaving in Soils.
Bulletin 225, U.S. Highway Research Board.
Penner, E. 1972. Influence of Freezing Rate on Frost
Heaving. Highway Research Record 393, U.S.
Highway Research Board.
Taber, S. 1930. The mechanics of frost heaving. Journal of
Geology 38: 303-317.
White, D.J., Take, W.A. & Bolton, M.D. 2003. Soil deformation
measurement using particle image velocimetry (PIV)
and photogrammetry. Geotechnique 53(7): 619-631
Xia, D. 2006. Frost Heave Studies Using Digital
Photographic Technique. M.S. Thesis. Edmonton,
AB, Canada: University of Alberta, 165 pp.
Xia, D., Arenson, L.U., Biggar, K.W. & Sego, D.C.
2005. Freezing process in Devon silt – using time-
lapse photography. Proceedings, 58th Canadian
Geotechnical Conference, Saskatoon, Saskatchewan.
September 18–21. CD-ROM