A total of 25 flexural and 55 uniaxial compressive
strength tests were conducted in laboratory using landfast sea ice samples
collected in the Prydz Bay. Three-point bending tests were performed at ice
temperatures of
Sea ice flexural strength, effective elastic modulus, and uniaxial compressive strength are important ice engineering properties. They are always used to assess the ice load exerted on marine structures in ice-infested waters (Sinsabvarodom et al., 2020; Su et al., 2010) and the load that can be supported by the floating ice (Masterson, 2009). Scientific and commercial activities have been expanding in the polar regions in recent years (Arctic Council, 2009; Mayewski et al., 2005). Therefore, studies on sea ice mechanical properties are still required.
Sea ice flexural strength and effective elastic modulus are obtained simultaneously using two different bending tests: cantilever and simple beam tests (Ji et al., 2011; Karulina et al., 2019). The cantilever beam test is a full-scale measurement performed on a large ice beam with whole thickness through the ice cover. The simple beam test (three- or four-point supported) is carried out using cuboid samples cut free from the ice cover. Because of maintaining the intact ice state, the cantilever beam test gives in situ ice flexural strength. However, it is highly time- and labor-consuming to prepare sufficient full-thickness cantilever beams, especially in the polar regions. Compared with the bending test, the uniaxial compressive strength test is relatively easy to carry out because of the smaller size of the samples required. Moreover, the behavior of ice under compressive strength tests is affected by machine stiffness (Sinha and Frederking, 1979). So, the test is performed using a machine with a high-stiffness loading frame (Bonath et al., 2019; Moslet, 2007).
The mechanical strength of sea ice has a strong dependence on its physical
properties, to be more exact, on the multiphase structure. Timco and O'Brien (1994) compiled a database of 939 reported measurements on the flexural
strength of sea ice from polar and temperate regions, and proposed a widely
used empirical equation of sea ice flexural strength relying on brine volume
fraction. Recently, a new formula of flexural strength dependent on brine
volume was reported in Karulina et al. (2019), by performing a series of
full-thickness cantilever beam tests. The parameterization of sea ice
effective elastic modulus based on brine volume fraction was also given in
Karulina et al. (2019). While investigations showed that gas within sea ice
may occupy more space than brine, especially for warm ice (Frantz et al.,
2019; Wang et al., 2020), an overestimation will be produced by calculating
strength only with brine volume. Sea ice strength should depend more
accurately on the total porosity (Timco and Weeks, 2010). The studies of sea
ice flexural strength and effective elastic modulus related to porosity are
rare. On the contrary, previous researches have related sea ice uniaxial
compressive strength to porosity (Kovacs, 1997; Moslet, 2007; Timco and
Frederking, 1990). The commonly adopted formulae to estimate sea ice
uniaxial compressive strength were proposed by Timco and Frederking (1990),
where 283 small-scale strength tests performed mostly in the Arctic waters
were collected. A limitation in their equations is that the applicable
condition is only for a strain rate less than 10
The mechanical properties of Arctic sea ice have been widely investigated in the past century because of booming oil and gas exploration in the Northern Hemisphere polar regions, while understanding of mechanical properties of Antarctic sea ice is limited due to less human and industry activities than those developed in the Arctic. With current expansion in science and tourism in the Antarctic regions, sound knowledge of Antarctic sea ice engineering properties is urgent. It is thought that the ice strength depends on the conditions of ice cover formation and its development, which in turn determines the ice structure. Unlike Arctic sea ice, there is a large fraction of granular ice layer in the Antarctic sea ice (Carnat et al., 2013; Jeffries et al., 2001). Therefore, the empirical equations established based on Arctic sea ice strength tests may not be appropriate for Antarctic sea ice. Furthermore, because of the heterogeneous variability shown by sea ice in different Antarctic seas responding to climate change (Hobbs et al., 2016; Matear et al., 2015), sea ice in different sea areas may also behave differently in mechanical properties.
The landfast sea ice area is the only corridor for scientific expeditions to transport logistics cargos to the research stations in the shore of Prydz Bay. However, the available observations on the landfast sea ice conditions in the Prydz Bay is quite limited (Hui et al., 2017; Zhao et al., 2020), let alone the sea ice mechanical properties that are of importance to ensure safe activities on ice. The full-scale ice trial of an icebreaker was conducted in the landfast sea ice in the Prydz Bay in the late austral spring in 2019. A full-thickness ice block was extracted through the ice cover, part of which was used for on-site measurements, and the rest was stored for subsequent detailed investigations on sea ice mechanical properties in domestic laboratory. In this paper, we present the results of various mechanical experiments performed in laboratory, including crystal structure, flexural strength, effective modulus of elasticity, and uniaxial compressive strength of the Antarctic landfast first-year sea ice. The results will help to deepen the understanding of mechanical properties of Antarctic sea ice, especially the landfast sea ice in the Prydz Bay.
As part of the 36th Chinese National Antarctic Research Expedition, a site
(69.2
Vertical and horizontal sections were made for ice crystal structure measurements. One-centimeter-thick sections at approximately 10 cm intervals along the ice thickness were prepared first using a band saw. These thick sections were attached to glass plates and thinned down further to an approximate 2 mm thickness. A planer was then used to reduce the thickness to 0.5 mm. The thin slices were placed on a universal stage to observe the crystal structures under crossed polarized light, recorded by photography. Image sizes were calibrated using a ruler with a resolution of 1 mm. Moreover, the images of horizontal slices were further analyzed for grain sizes using an image processing software. The ice grains were distinguished separately and then were recognized as circles to determine the diameters from areas, all of which in a horizontal section were finally averaged.
The flexural strength and effective elastic modulus of sea ice were measured
using three-point supported beam tests. The rough-cut ice beams were
prepared using chain saw and were then machined carefully to section
dimensions of 7
The three-point bending test was conducted in the cold laboratory using
equipment as shown in Fig. 2. The device was powered by a hydraulic
actuator. A stainless-steel column was fixed on the bottom of a pressing plate
to give a line force on the midspan of the ice beam underlain by a simply
supported frame with a span of 60 cm. A force sensor with a capacity of 500 N and an accuracy of
Based on the linear elasticity theory, the flexural strength of ice
(
In the three-point supported beam tests, the effective elastic modulus of
ice (
Experiment equipment for the three-point bending test: (A1) hydraulic system; (A2) force sensor; (A3) aluminum plate; (A4) line-load pressing plate; (A5) laser displacement sensor (behind the ice sample); and (A6) simply supported frame.
Before loading, the mass of each sample was weighed using a balance (
The test conditions of bending and compression experiments.
Similar to sample preparation of bending tests, cuboid-shaped compression
samples with dimensions of 7
Because of the higher-rigidity setup required by the compressive strength test as
compared with the bending test, a universal testing machine with a portal
frame was used (Fig. 3). The machine was equipped with four columns
supporting the upper beams, all of which are made of welded steel plates.
So, it was expected to have a higher stiffness than the ice sample. The machine
also used a servo motor which can maintain a constant loading rate with an
accuracy of
Experiment equipment for the uniaxial compressive strength test: (B1) control system; (B2) loading frame; (B3) refrigerating system; (B4) cryostat container; and (B5) pressing plate.
Results of ice crystal measurements are shown in Fig. 4. The sea level was
approximately 19 cm below the ice surface. There was a snow-ice layer in the
top 28 cm of the ice sheet where a large number of fine-grained crystals
were in the shape of small polygons, shown in the vertical and horizontal
sections. The snow ice could also be judged by its white appearance. Their
diameters were too small (approximately
A total of three types of ice samples were measured in the bending tests
which consisted of snow ice, columnar ice, and a mixture of transition and
columnar ice. The snow-ice specimens could be distinguished easily by their
white appearance and light weight with mean density of 0.55
A total of 16 columnar-ice beams and 5 mixed-ice beams were conducted. In terms of strength, the flexural strength of mixed ice
ranged from 511 to 846 kPa with an average of 688
Typical curves of force versus deflection at the middle of a beam with corresponding broken sections depicted by black lines shown in subplot.
It was found that the flexural strength of congelation ice was dependent on
sea ice porosity rather than brine volume fraction (see Sect. 4.1.1 for
further discussion). Taking the square root of porosity of 0.02 as a bin,
the mean value and standard deviation were determined, and the relationship
between flexural strength and porosity is shown in Fig. 6. Sea ice flexural
strength decreased with increasing porosity. To further quantify the
dependence of flexural strength on porosity of our ice samples, regression
analysis was conducted using the form as the one for brine volume fraction
used in previously established equations (Table 2), and also other commonly
used functions including linear, logarithmic, and power functions. Results
showed a best-fit relationship with the form of Eq. (5) and with a
determination coefficient (
The relationship between average sea ice flexural strength and porosity for congelation ice. The confidence interval is selected based on significance level. The subplot shows the comparisons between estimated flexural strength using the empirical equations in previous studies and this work.
Empirical relationships of flexural strength and effective modulus
of elasticity based on brine volume fraction (
To compare our best-fit equation (Eq. 5) with existing equations in Table 2, the results of flexural strength calculated using these equations were plotted against the square root of the brine volume fraction (see subplot in Fig. 6). Results showed that the strength estimated using Karulina et al. (2019) was much lower than that estimated using ours. The strength estimated using Timco and O'Brien (1994) agreed better with ours, and only overestimated by 1.1 times.
The bending failure of ice is considered as a tensile failure in brittle
manner, and grain size affects ice failure under tension. Consequently, to
investigate the relationship between flexural strength and grain size, the
grain size of columnar ice was divided at an interval of 0.1 cm, and the
mean strength was determined. A clear trend is shown in Fig. 7a in which
the flexural strength of columnar ice decreased with increasing grain size. For
a homogeneous crack-free polycrystalline fresh-water ice with the same scale of grain
size as our samples, the fracture process at a high strain rate is
nucleation controlled. The crack propagates as soon as it forms, and the
stress at failure is inversely proportional to grain size (Sanderson, 1988).
It is noteworthy that probably because of the minor effect of grain size
compared with porosity, the best-fit trend was not significant at
The variations of average flexural strength with respect to
A total of four snow-ice beams were tested, and results showed that the
flexural strength of snow ice was much weaker than that of congelation ice,
with an average of 123
The effective elastic modulus of columnar-ice samples ranged from 0.4 to 2.3 GPa, with an average of 1.5
The effective elastic modulus of our ice samples had no statistically
significant dependence on brine volume fraction. Considering the effect of
sea ice porosity on the flexural strength, the dependence of effective
elastic modulus on porosity was also investigated. However, regression
analysis indicated a weak relationship between them with low
The effects of sea ice substructure on the effective elastic modulus of
columnar ice were investigated as shown in Fig. 8, where the relationships
between effective elastic modulus and grain size, as well as platelet spacing,
were examined using the same processing approaches as flexural strength.
Results showed that there was a negligible effect of grain size on effective
elastic modulus with
The variations of average effective elastic modulus with respect
to
The mean effective elastic modulus of snow ice was 0.4
The uniaxial compressive strength samples were divided into three types, i.e.,
congelation-ice samples consisting of columnar and/or transition ice,
snow-ice samples, and mixed-ice samples consisting of both snow and
congelation ice. The snow-ice samples had the least density with an average of
0.61
Figure 9 shows the typical stress–strain curves of sea ice during compressive strength tests. Different behaviors were shown by sea ice compressed under different strain rates and loading directions. At a low strain rate, stress increased linearly with strain until peak, and then decreased gently without abrupt change, indicating ductile behavior (curves B and D). Large deformation with local cracks can be seen in the samples at failure (samples B and D). With an increase in the strain rate, sea ice exhibited brittle behavior, where the stress dropped abruptly once reaching peak (curves A and C). No obvious deformation occurred at the time of failure (samples A and C). Similar stress–strain curves of brittle and ductile failures under compression were also seen in ridged ice (Bonath et al., 2019) and polycrystalline ice (Arakawa and Maeno, 1997).
In spite of stress–strain curves, there were several differences between vertically loaded and horizontally loaded samples at failure. For the vertically loaded samples compressed under a low strain rate (sample B), cracks developed along the long axes of the columnar grains, whereas for horizontally loaded samples (sample D), inclined cracks developed at both ends of the ice. For horizontally loaded samples, force was applied perpendicular to the long axes of the ice columns, and the sliding along the grain boundary was easily triggered, combined with the dislocation pile-up at grain boundaries, leading to crack development along the grain boundaries (Gold, 1997; Sinha, 1988), whereas for vertically loaded samples, force acted on the cross of the columns, and grain boundary sliding was suppressed. Therefore, local cracks developed parallel to the column axes due to the grain decohesion (Kuehn and Schulson, 1994). Alternatively, at a high strain rate, local cracks and deformation were not sufficient to relax the stress concentration. Therefore, for the vertically loaded samples, once the sea ice was split by a vertical crack, the slender columns suddenly became unstable and failed through buckling (sample A). For horizontally loaded samples, inclined cracks accumulated and formed a primary crack penetrating through the sample, and the sample failed through shear faulting (sample C).
Typical stress–strain curves and failure modes of
Since the test strain rate was wide ranging from 10
Variations of uniaxial compressive strength with strain rate
The mean strength and standard deviation were determined taking the porosity of 10 ‰ as a bin, and the variations of uniaxial compressive strength with respect to porosity are plotted in Fig. 10b. As the sea ice porosity increases, strength decreases since brine and gas inclusions cannot support the load. A declining trend in sea ice compressive strength with increasing porosity was also reported in Moslet (2007). Furthermore, regression analysis showed a statistically significant dependence of uniaxial compressive strength on porosity following a power law.
Since the uniaxial compressive strength of sea ice is affected by strain
rate and porosity, it is rational to parameterize the compressive strength
using these two parameters. Based on the aforementioned relationships
between strain rate, porosity, and uniaxial compressive strength, the form of the model
proposed in Kovacs (1997) to estimate the sea ice strength in the
rate-strengthening regime was used here (Eq. 6):
The fitting coefficients of Eq. (6).
The original model proposed in Kovacs (1997) was extended here to cover both
ductile and brittle regimes simply by changing the coefficients. All the
regressions showed good
The three-dimensional surfaces of sea ice uniaxial compressive strength
varying with strain rate and porosity are plotted in Fig. 11. It was found
that sea ice transitioned from ductile to brittle behavior over a range of
strain rates. For vertically loaded samples, the ductile-to-brittle range
was approximately 8.0
Indeed, the transition is a result of competition between stress relaxation and stress build-up (Schulson, 2001). Under vertical loading, the stress concentration is relaxed by grain decohesion, and for horizontal loading, it is relaxed by grain boundary sliding (Gold, 1997; Kuehn and Schulson, 1994; Sinha, 1988). Because the instantaneous elastic response of the decohesion is more sensitive to the strain, ductile-to-brittle strain rate under horizontal loading is higher (Ji et al., 2020). Moreover, the transition strain rate increased with increasing porosity. Brine and gas inclusions in sea ice can act as pre-cracks promoting stress concentration. On the other hand, with sea ice being more porous, both grain decohesion and grain boundary sliding is much more easily triggered (Schulson, 2001). It seemed that the effect of porosity on the grain boundary sliding was relatively slight, so that the transition range was narrow for horizontal loading.
Three-dimensional surfaces of uniaxial compressive strength of
A total of four mixed-ice samples were compressed under vertical loading
direction, of which two were tested at 2.0
Typical stress–strain curves and failure modes of mixed ice
samples
Similarly, snow ice samples compressed at different strain rates also showed
distinct stress–strain curves. Taking the vertically loaded snow ice samples
as examples, at low strain rates (e.g., 2.0
Sea ice flexural strength has often been related to brine volume fraction in previous studies (Karulina et al., 2019; Timco and O'Brien, 1994). To compare with previously reported empirical equations for sea ice flexural strength, the brine volume fractions of our congelation ice samples were determined and taken to the previously established equations listed in Table 2. Results shown in Fig. 13 indicate that the equation of Karulina et al. (2019) underestimated our flexural strength by approximately 40 % on average. The difference may be attributed to several factors. The first factor is that our tests were conducted in the laboratory, where samples were prepared with caution, while the tests in Karulina et al. (2019) were performed in field. Furthermore, the ice samples in Karulina et al. (2019) were much larger and contained more potential weaknesses than ours. Besides, the flexural strength in Karulina et al. (2019) was derived using cantilever beams, and stress concentrations at the root of the beam resulted in low strength. The equation provided by Timco and O'Brien (1994) was derived from various regions and test approaches, and thus its estimation agreed better with our data than the estimation using Karulina et al. (2019). Even so, it still overestimated our flexural strength by approximately 20 % because Timco and O'Brien's equation was also a conservative estimate.
Comparisons between estimated flexural strength, using previously established equations, and measured strength. The dashed lines are varying trends.
Furthermore, regression analysis was conducted not only using the
mathematical expressions as given in Table 2, but also other commonly used
functions (e.g., linear, logarithmic, and power functions) to validate the
relationship between flexural strength and brine volume fraction of our
congelation ice samples. However, all the fitting equations were accompanied
with low
The equations to estimate sea ice effective elastic modulus are rare, except where Karulina et al. (2019) proposed a mathematical relationship between sea ice effective elastic modulus and brine volume fraction, in which the effective elastic modulus decreased with the square root of the brine volume fraction in an exponential manner (Table 2). Taking our data into the equation, the comparison showed that the calculation using their formula gave 1.5 times overestimation than our measurements. As stated before, the size effect, sample preparation, and test techniques probably caused the differences.
Unlike effective elastic modulus, there are several empirical formulae to
estimate sea ice uniaxial compressive strength. The frequently used parameterization
was proposed by Timco and Frederking (1990), which has been validated by the
Arctic sea ice strength data, although it is uncertain whether their model is
appropriate for the Antarctic sea ice. Overestimates were obtained by taking
our test data into the previously established model, and the ratio was 1.4
for vertically loaded samples and 1.8 for horizontally loaded samples.
Furthermore, the applicable strain-rate range was below 10
From the viewpoint of field operations, it is easier to obtain the vertically
loaded uniaxial compressive strength than the horizontally loaded
compressive strength and flexural strength, because ice cores can be removed
from ice cover directly using an ice driller. However, the latter two
properties of sea ice are more useful than the former in engineering
applications. Since both compressive strength and flexural strength are
affected by porosity, taking the porosity of 10 ‰ as a
bin, ratios were obtained by comparing the mean strength located in the same
porosity interval. It was found that the ratio of vertically to horizontally
loaded uniaxial compressive strength was independent of porosity, and the
average was 3.1
In Sect. 3.2.1, a mathematical equation (Eq. 5) was proposed to determine
the flexural strength of landfast sea ice in the Prydz Bay based on sea ice
porosity. During the field measurements, the mean ice temperature and
salinity of the ice block extracted was
The field measured strength (719
Karulina et al. (2019) reported a range of sea ice flexural strength of 109–415 kPa by performing full-scale tests in the Arctic regions, which is lower than our field measured strengths. Our field measurements were performed using small-scale three-point supported beam tests, and as stated before, the differences in size effects and test techniques could result in different strength values.
Due to the severe ice conditions and unknown offshore water depth,
icebreakers usually have to stop in the landfast sea ice area in the Prydz
Bay (Fig. 1), and the logistics cargos for research stations have to be
transported by helicopters and trucks shortly after they are unloaded on
ice. Therefore, it is of paramount importance to estimate the ice bearing
capacity to ensure safety. A typical scenario of unloading cargos from an
icebreaker on landfast sea ice in the Prydz Bay is shown in Fig. 14. The
channel broken by the icebreaker acts as a wet crack penetrating the ice,
i.e., the entire ice side surface is exposed to water (Masterson, 2009), and
the cargos are loaded not far from the ship. The bending of the ice under
the load causes flexural stress to be imposed on the ice cross section. If
the maximum flexural stress does not exceed the ice strength, the load will
be supported. Following the industry standard required by ISO19906 (2019),
the extreme fiber stress in a cracked ice sheet due to a uniformly
distributed load is predicted by Eqs. (7) and (8):
Scenario of unloading cargos from an icebreaker on landfast sea ice in the Prydz Bay (Photo credit: Shiping Liu).
Variations of load with respect to sea ice porosity for a
landfast sea ice with congelation ice thickness of 1.3 m.
Noticeably, the flexural strength calculated using Eqs. (7) and (8) is
based on the elastics beam theory assumption. Kerr and Palmer (1972) deduced
the distributions of bending stresses considering varied elastic modulus
along ice thickness in a floating ice plate. It is difficult to obtain the
real distribution of elastic modulus along ice thickness, and Eqs. (7) and
(8) have been proved reliable by experience (ISO19906, 2019), and hence,
are adopted here. For the purpose of safe designing, a more conservative
prediction is preferred in practice. According to Eq. (7), the bearing
capacity increases with increasing flexural strength and decreasing
effective elastic modulus. Therefore, minimum flexural strength and maximum
effective elastic modulus are required. Section 3 talks about the variations
of flexural strength and effective elastic modulus with porosity in terms of
averages. So, all the measured data were plotted in Fig. 15a and b to
determine the lower and upper envelopes of flexural strength and effective
elastic modulus, respectively, providing the best-fit equations given by
Eqs. (9) and (10):
The above estimation is close to the actual scenario to some degree. As the load is applied on the ice sheet, the sheet is compressed at the top and tensioned at the bottom. Ice is a material which is strong in compression and weak in tension. So, the ice sheet deflects until the first crack or yielding develops in the underside of the sheet beneath the center of the load (Masterson, 2009). The low flexural strength often occurs at the bottom of the ice sheet because of high ice temperature near the freezing point; therefore, it is reasonable to use the lower envelope of flexural strength.
As stated in Sect. 3.1, the flexural strength of snow ice is much weaker
than that of congelation ice, making the snow cover negligible in terms of bearing
capacity. Therefore, the ice thickness was taken as 1.3 m, which is the
thickness of the congelation-ice layer of the ice block based on the crystal
texture (Fig. 4). Additionally, Eqs. (7) and (8) worked reliably when the
loaded radius was not large enough compared with the characteristic length
(
It is emphasized that there are uncertainties in the estimated bearing capacity. First, in this paper, sea ice mechanical properties were obtained using small-scale samples, which overestimates the full-scale cantilever beam tests preferred by the industry standard. Additionally, the load calculated by Eqs. (7) and (8) is for the short term, under which the creep of ice is ignorable. Although the goods are transported away shortly after being unloaded on the ice, the duration is at least hours because of manual operation, possibly resulting in ice creep. In the industry standard, in addition to ensuring that the ice sheet extreme fiber stress be less than the allowable flexural stress for ice, it is necessary to avoid submergence caused by sea ice creep deformation (ISO19906, 2019). Therefore, the mass of goods should be less than the estimations given in Fig. 15c.
A series of laboratory mechanical experiments, including 25 bending and 55 uniaxial compressive strength tests, was carried out on landfast sea ice collected in the Prydz Bay, East Antarctica. The crystal measurements showed that a snow-ice layer exists in the top of the ice followed by transition and columnar ice formed by congelation of seawater. The flexural strength, effective elastic modulus, vertically and horizontally loaded uniaxial compressive strength of congelation ice, snow ice, and a mixture of the two were measured. The strength of mixed ice is weaker than that of congelation ice because of the existence of snow ice, and that of ice consisting of pure snow ice is the lowest. The effects of sea ice substructure on columnar ice strength were investigated. Both flexural strength and effective elastic modulus increased with increasing platelet spacing, while the influence of grain size was not significant.
The commonly used estimating equations of sea ice strength derived from Northern Hemisphere polar regions is not fit for landfast sea ice in the Prydz Bay, and alternative models are established for flexural and compressive strength (Eqs. 5 and 6). Sea ice uniaxial compressive strength was parameterized with strain rate and porosity using the form proposed in Kovacs (1997), but it was extended to cover both ductile and brittle strain-rate regimes. Sea ice flexural strength is related to porosity in this work, rather than brine volume which was often adopted in previous models. The newly proposed parameterization for flexural strength based on sea ice porosity compensates for the lack of applicability to warm ice using previous models based on brine volume. Furthermore, the dependence of sea ice strength on total porosity, rather than brine content, can reflect the potential effects of climate change. Sea ice in polar regions may become warmer under the effects of global warming (Clem et al., 2020; Screen and Simmonds, 2010), and the gas within sea ice may occupy more space than brine (Wang et al., 2020).
As scientific investigations have flourished in the Antarctic regions, the mechanical properties of Antarctic sea ice need to be urgently elucidated for safe activities on ice. With minimum flexural strength and maximum effective elastic modulus of congelation ice, a method has been established to estimate the bearing capacity of landfast sea ice cover in the Prydz Bay following the industry standard (ISO19906, 2019). In this way, it is possible to estimate the magnitude of load that can be safely put on ice, based on the sea ice physical properties.
Admittedly, the mechanical tests performed as described in this paper were derived from the extracted ice block only, so the sample amount is limited. Nevertheless, the dataset provided here is of great value as it contains most industry-concerned sea ice mechanical properties, which is helpful to improve the understanding of Antarctic sea ice and support of safe marine activities.
A total of 25 bending and 55 uniaxial compressive strength tests were performed on
landfast sea ice samples. The detailed information of the ice samples is
listed in Tables A1–A3, where
Detailed information of bending test samples.
Detailed information of vertically loaded uniaxial compressive
strength test samples at the temperature of
Detailed information of horizontally loaded uniaxial compressive
strength test samples at the temperature of
All data are available at the website
QW, YX, and ZhiL designed the experiments, and QW and ZhaL carried them out. QW analyzed the test data and prepared the manuscript with contributions from all co-authors. PL reviewed and edited the manuscript.
The contact author has declared that neither they nor their co-authors have any competing interests.
Publisher’s note: Copernicus Publications remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We thank the crews of PRIC for their assistance with the fieldwork during ice sampling, Shiping Liu from the Xinhua News Agency for providing the photo of unloading cargos on ice, and Minghu Ding from CAMS for providing the historical meteorological data at the Zhongshan station. We are also grateful to editor Bin Cheng and three anonymous reviewers for their valuable suggestions which contribute considerably to the publication of this paper.
This research has been supported by the National Natural Science Foundation of China (grant nos. 52192692, 41906198, and 41922045), the Fundamental Research Funds for the Central Universities (grant no. DUT21RC3086), and the Liao Ning Revitalization Talents Program (grant nos. XLYC2007033 and XLYC1908027).
This paper was edited by Bin Cheng and reviewed by three anonymous referees.