A textural approach to snow depth distribution on Antarctic sea ice

2020 ◽  
Author(s):  
M. Jeffrey Mei ◽  
Ted Maksym
Keyword(s):  
Sea Ice ◽  
Snow Depth ◽  
Ice Thickness ◽  
Two Dimensional ◽  
Snow Surface ◽  
Altimetry Data ◽  
Laser Altimetry ◽  
Sea Ice Thickness ◽  
Antarctic Sea Ice ◽  
Surface Morphologies

<p>Understanding the distribution of snow depth on Antarctic sea ice is critical to estimating the sea ice thickness distribution from laser altimetry data, such as from Operation IceBridge or ICESat-2.  One major uncertainty in converting laser altimetry data to ice thickness is knowing the proportion of snow within the surface measurement. Snow redistributed by wind collects around areas of deformed ice, but it is not known how different surface morphologies affect this distribution. Here, we apply a textural segmentation algorithm to classify and group similar textures to infer the distribution of snow-ice ratios using snow surface freeboard measurements from Operation IceBridge (OIB) campaigns over the Weddell Sea. We find that texturally-similar regions have similar snow/ice ratios, but not similar snow depth measurements. This allows for the extrapolation of nadir-looking snow radar data using two-dimensional surface altimetry scans, providing a two-dimensional estimate of the snow depth. Using a convolutional neural network on an in-situ dataset, we find that local (~20 m) snow depth and sea ice thickness can be estimated with errors of < 20%, and that the learned convolutional filters imply that different surface morphologies have different proportions of snow/ice within the measured surface elevation. For the OIB data,  we show that at slightly larger scales (~180 m), snow depths can be estimated using the snow surface texture, and that the learned filters are comparable to standard textural segmentation filters. We also examine the statistical variability in the distribution of snow/ice ratios across different years to determine if snow distribution patterns on sea ice exhibit universal behaviour, or have significant interannual variations. These results suggest that surface morphological information can improve remotely-sensed estimates of snow depth, and hence sea ice thickness, compared to current methods. Such methods may be useful for reducing errors in Antarctic sea ice thickness estimates from ICESat-2.</p>

scholarly journals A Textural Approach to Improving Snow Depth Estimates in the Weddell Sea

2020 ◽  
Vol 12 (9) ◽  
pp. 1494
Author(s):  
M. Jeffrey Mei ◽  
Ted Maksym
Keyword(s):  
Sea Ice ◽  
Snow Depth ◽  
Thickness Distribution ◽  
Radar Data ◽  
Weddell Sea ◽  
Ice Thickness ◽  
Two Dimensional ◽  
Snow Surface ◽  
Sea Ice Thickness ◽  
Antarctic Sea Ice

The snow depth on Antarctic sea ice is critical to estimating the sea ice thickness distribution from laser altimetry data, such as from Operation IceBridge or ICESat-2. Snow redistributed by wind collects around areas of deformed ice and forms a wide variety of features on sea ice; the morphology of these features may provide some indication of the mean snow depth. Here, we apply a textural segmentation algorithm to classify and group similar textures to infer the distribution of snow using snow surface freeboard measurements from Operation IceBridge campaigns over the Weddell Sea. We find that texturally-similar regions have similar snow/ice ratios, even when they have different absolute snow depth measurements. This allows for the extrapolation of nadir-looking snow radar data using two-dimensional surface altimetry scans, providing a two-dimensional estimate of the snow depth with ∼22% error. We show that at the floe scale (∼180 m), snow depth can be directly estimated from the snow surface with ∼20% error using deep learning techniques, and that the learned filters are comparable to standard textural analysis techniques. This error drops to ∼14% when averaged over 1.5 km scales. These results suggest that surface morphological information can improve remotely-sensed estimates of snow depth, and hence sea ice thickness, as compared to current methods. Such methods may be useful for reducing uncertainty in Antarctic sea ice thickness estimates from ICESat-2.

scholarly journals Estimation of sea ice thickness distributions through the combination of snow depth and satellite laser altimetry data

2009 ◽  
Vol 114 (C10) ◽  
Author(s):  
Nathan T. Kurtz ◽  
Thorsten Markus ◽  
Donald J. Cavalieri ◽  
Lynn C. Sparling ◽  
William B. Krabill ◽  
...  
Keyword(s):  
Sea Ice ◽  
Snow Depth ◽  
Ice Thickness ◽  
Altimetry Data ◽  
Laser Altimetry ◽  
Sea Ice Thickness

scholarly journals On the retrieval of sea ice thickness and snow depth using concurrent laser altimetry and L-band remote sensing data

2018 ◽  
Vol 12 (3) ◽  
pp. 993-1012 ◽  
Author(s):  
Lu Zhou ◽  
Shiming Xu ◽  
Jiping Liu ◽  
Bin Wang
Keyword(s):  
Remote Sensing ◽  
Sea Ice ◽  
Snow Depth ◽  
Microwave Remote Sensing ◽  
Ice Cover ◽  
Ice Thickness ◽  
Laser Altimetry ◽  
Sea Ice Thickness ◽  
Passive Remote Sensing ◽  
L Band

Abstract. The accurate knowledge of sea ice parameters, including sea ice thickness and snow depth over the sea ice cover, is key to both climate studies and data assimilation in operational forecasts. Large-scale active and passive remote sensing is the basis for the estimation of these parameters. In traditional altimetry or the retrieval of snow depth with passive microwave remote sensing, although the sea ice thickness and the snow depth are closely related, the retrieval of one parameter is usually carried out under assumptions over the other. For example, climatological snow depth data or as derived from reanalyses contain large or unconstrained uncertainty, which result in large uncertainty in the derived sea ice thickness and volume. In this study, we explore the potential of combined retrieval of both sea ice thickness and snow depth using the concurrent active altimetry and passive microwave remote sensing of the sea ice cover. Specifically, laser altimetry and L-band passive remote sensing data are combined using two forward models: the L-band radiation model and the isostatic relationship based on buoyancy model. Since the laser altimetry usually features much higher spatial resolution than L-band data from the Soil Moisture Ocean Salinity (SMOS) satellite, there is potentially covariability between the observed snow freeboard by altimetry and the retrieval target of snow depth on the spatial scale of altimetry samples. Statistically significant correlation is discovered based on high-resolution observations from Operation IceBridge (OIB), and with a nonlinear fitting the covariability is incorporated in the retrieval algorithm. By using fitting parameters derived from large-scale surveys, the retrievability is greatly improved compared with the retrieval that assumes flat snow cover (i.e., no covariability). Verifications with OIB data show good match between the observed and the retrieved parameters, including both sea ice thickness and snow depth. With detailed analysis, we show that the error of the retrieval mainly arises from the difference between the modeled and the observed (SMOS) L-band brightness temperature (TB). The narrow swath and the limited coverage of the sea ice cover by altimetry is the potential source of error associated with the modeling of L-band TB and retrieval. The proposed retrieval methodology can be applied to the basin-scale retrieval of sea ice thickness and snow depth, using concurrent passive remote sensing and active laser altimetry based on satellites such as ICESat-2 and WCOM.

scholarly journals Snow depth uncertainty and its implications on satellite derived Antarctic sea ice thickness

2018 ◽  
Author(s):  
Daniel Price ◽  
Iman Soltanzadeh ◽  
Wolfgang Rack
Keyword(s):  
Sea Ice ◽  
Snow Depth ◽  
Snow Accumulation ◽  
In Situ Measurements ◽  
Ice Thickness ◽  
Growth Season ◽  
Sea Ice Thickness ◽  
Antarctic Sea Ice ◽  
Fast Ice

Abstract. Knowledge of the snow depth distribution on Antarctic sea ice is poor but is critical to obtaining sea ice thickness from satellite altimetry measurements of freeboard. We examine the usefulness of various snow products to provide snow depth information over Antarctic fast ice with a focus on a novel approach using a high-resolution numerical snow accumulation model (SnowModel). We compare this model to results from ECMWF ERA-Interim precipitation, EOS Aqua AMSR-E passive microwave snow depths and in situ measurements at the end of the sea ice growth season. The fast ice was segmented into three areas by fastening date and the onset of snow accumulation was calibrated to these dates. SnowModel falls within 0.02 m snow water equivalent (swe) of in situ measurements across the entire study area, but exhibits deviations of 0.05 m swe from these measurements in the east where large topographic features appear to have caused a positive bias in snow depth. AMSR-E provides swe values half that of SnowModel for the majority of the sea ice growth season. The coarser resolution ERA-Interim, not segmented for sea ice freeze up area reveals a mean swe value 0.01 m higher than in situ measurements. These various snow datasets and in situ information are used to infer sea ice thickness in combination with CryoSat-2 (CS-2) freeboard data. CS-2 is capable of capturing the seasonal trend of sea ice freeboard growth but thickness results are highly dependent on the assumptions involved in separating snow and ice freeboard. With various assumptions about the radar penetration into the snow cover, the sea ice thickness estimates vary by up to 2 m. However, we find the best agreement between CS-2 derived and in situ thickness when a radar penetration of 0.05-0.10 m into the snow cover is assumed.

scholarly journals On the Retrieval of Sea Ice Thickness and Snow Depth using Concurrent Laser Altimetry and L-Band Remote Sensing Data

2017 ◽  
Author(s):  
Lu Zhou ◽  
Shiming Xu ◽  
Jiping Liu ◽  
Bin Wang
Keyword(s):  
Remote Sensing ◽  
Sea Ice ◽  
Snow Depth ◽  
Ice Cover ◽  
Retrieval Algorithm ◽  
Ice Thickness ◽  
Laser Altimetry ◽  
Sea Ice Thickness ◽  
Passive Remote Sensing ◽  
L Band

Abstract. The accurate knowledge of sea ice parameters, including sea ice thickness and snow depth over the sea ice cover, are key to both climate studies and data assimilation in operational forecasts. Large-scale active and passive remote sensing is the basis for the estimation of these parameters. In traditional altimetry or the retrieval of snow depth with passive microwave sensing, although the sea ice thickness and the snow depth are closely related, the retrieval of one parameter is usually carried out under assumptions over the other. For example, climatological snow depth data or as derived from reanalyses contain large or unconstrained uncertainty, which result in large uncertainty in the derived sea ice thickness and volume. In this study, we explore the potential of combined retrieval of both sea ice thickness and snow depth using the concurrent active altimetry and passive microwave remote sensing of the sea ice cover. Specifically, laser altimetry and L-band passive remote sensing data are combined using two forward models: the L-band radiation model and the isostatic relationship based on buoyancy model. Since the laser altimetry usually features much higher spatial resolution than L-band data from Soil Moisture Ocean Salinity (SMOS) satellite, there is potentially covariability between the observed snow freeboard by altimetry and the retrieval target of snow depth on the spatial scale of altimetry samples. Statistically significant correlation is discovered based on high-resolution observations from Operation IceBridge (OIB), and with a nonlinear fitting the covariability is incorporated in the retrieval algorithm. By using fitting parameters derived from large-scale surveys, the retrievability is greatly improved, as compared with the retrieval that assumes flat snow cover (i.e., no covariability). Verifications with OIB data show good match between the observed and the retrieved parameters, including both sea ice thickness and snow depth. With detailed analysis, we show that the error of the retrieval mainly arises from the difference between the modeled and the observed (SMOS) L-band brightness temperature (TB). The narrow swath and the limited coverage of the sea ice cover by altimetry, as well the uncertainty associated with the radiation model are potential sources of error. The proposed retrieval algorithm (or methodology) can be applied to the basin-scale retrieval of sea ice thickness and snow depth, using concurrent passive remote sensing and active laser altimetry based on satellites such as ICESat and ICESat-2.

scholarly journals Snow-driven uncertainty in CryoSat-2-derived Antarctic sea ice thickness – insights from McMurdo Sound

2019 ◽  
Vol 13 (4) ◽  
pp. 1409-1422
Author(s):  
Daniel Price ◽  
Iman Soltanzadeh ◽  
Wolfgang Rack ◽  
Ethan Dale
Keyword(s):  
Sea Ice ◽  
Snow Depth ◽  
Snow Accumulation ◽  
In Situ Measurements ◽  
Ice Thickness ◽  
Growth Season ◽  
Mcmurdo Sound ◽  
Sea Ice Thickness ◽  
Antarctic Sea Ice

Abstract. Knowledge of the snow depth distribution on Antarctic sea ice is poor but is critical to obtaining sea ice thickness from satellite altimetry measurements of the freeboard. We examine the usefulness of various snow products to provide snow depth information over Antarctic fast ice in McMurdo Sound with a focus on a novel approach using a high-resolution numerical snow accumulation model (SnowModel). We compare this model to results from ECMWF ERA-Interim precipitation, EOS Aqua AMSR-E passive microwave snow depths and in situ measurements at the end of the sea ice growth season in 2011. The fast ice was segmented into three areas by fastening date and the onset of snow accumulation was calibrated to these dates. SnowModel captures the spatial snow distribution gradient in McMurdo Sound and falls within 2 cm snow water equivalent (s.w.e) of in situ measurements across the entire study area. However, it exhibits deviations of 5 cm s.w.e. from these measurements in the east where the effect of local topographic features has caused an overestimate of snow depth in the model. AMSR-E provides s.w.e. values half that of SnowModel for the majority of the sea ice growth season. The coarser-resolution ERA-Interim produces a very high mean s.w.e. value 20 cm higher than the in situ measurements. These various snow datasets and in situ information are used to infer sea ice thickness in combination with CryoSat-2 (CS-2) freeboard data. CS-2 is capable of capturing the seasonal trend of sea ice freeboard growth but thickness results are highly dependent on what interface the retracked CS-2 height is assumed to represent. Because of this ambiguity we vary the proportion of ice and snow that represents the freeboard – a mathematical alteration of the radar penetration into the snow cover – and assess this uncertainty in McMurdo Sound. The ranges in sea ice thickness uncertainty within these bounds, as means of the entire growth season, are 1.08, 4.94 and 1.03 m for SnowModel, ERA-Interim and AMSR-E respectively. Using an interpolated in situ snow dataset we find the best agreement between CS-2-derived and in situ thickness when this interface is assumed to be 0.07 m below the snow surface.

scholarly journals Sea-ice thickness and roughness in the Ross Sea, Antarctica

2001 ◽  
Vol 33 ◽  
pp. 187-193 ◽  
Author(s):  
Tina Tin ◽  
Martin O. Jeffries
Keyword(s):  
Surface Roughness ◽  
Sea Ice ◽  
Ross Sea ◽  
Geographical Location ◽  
Ice Thickness ◽  
Snow Surface ◽  
Sea Ice Thickness ◽  
Antarctic Sea Ice ◽  
Ice Roughness ◽  
Ice Floes

AbstractSea-ice thickness and roughness data collected on three cruises in the Ross Sea, Antarctica, showed interseasonal, regional and interannual variability. Variability was reduced to season, or age of ice floe, when sea-ice roughness values from around Antarctica were compared. There were statistically significant correlations between mean snow elevation and mean ice thickness; snow surface roughness and mean ice thickness; and snow surface roughness and ice bottom roughness, which appeared to be independent of season, geographical location and deformation history of ice floes. Our field data indicate that ice thickness can be predicted from snow elevation measurements with higher accuracy in summer. The feasibility of using snow surface roughness to infer ice thickness and ice bottom roughness is promising, and can provide us with a means to study the thickness and underside of Antarctic sea ice at good spatial and temporal resolution.

scholarly journals Uncertainties in Antarctic sea-ice thickness retrieval from ICESat

2015 ◽  
Vol 56 (69) ◽  
pp. 107-119 ◽  
Author(s):  
Stefan Kern ◽  
Gunnar Spreen
Keyword(s):  
Sea Ice ◽  
Snow Depth ◽  
Sensitivity Study ◽  
Weddell Sea ◽  
Ice Thickness ◽  
Sea Ice Thickness ◽  
Earth Observing System ◽  
Antarctic Sea Ice ◽  
Snow Properties ◽  
The Mean

AbstractA sensitivity study was carried out for the lowest-level elevation method to retrieve total (sea ice + snow) freeboard from Ice, Cloud and land Elevation Satellite (ICESat) elevation measurements in the Weddell Sea, Antarctica. Varying the percentage (P) of elevations used to approximate the instantaneous sea-surface height can cause widespread changes of a few to ˃10cm in the total freeboard obtained. Other input parameters have a smaller influence on the overall mean total freeboard but can cause large regional differences. These results, together with published ICESat elevation precision and accuracy, suggest that three times the mean per gridcell single-laser-shot error budget can be used as an estimate for freeboard uncertainty. Theoretical relative ice thickness uncertainty ranges between 20% and 80% for typical freeboard and snow properties. Ice thickness is computed from total freeboard using Advanced Microwave Scanning Radiometer for Earth Observing System (AMSR-E) snow depth data. Average ice thickness for the Weddell Sea is 1.73 ± 0.38 m for ICESat measurements from 2004 to 2006, in agreement with previous work. The mean uncertainty is 0.72 ± 0.09 m. Our comparison with data of an alternative approach, which assumes that sea-ice freeboard is zero and that total freeboard equals snow depth, reveals an average sea-ice thickness difference of ∼0.77m.

scholarly journals Estimating Early-Winter Antarctic Sea Ice Thickness From Deformed Ice Morphology

2019 ◽  
Author(s):  
M. Jeffrey Mei ◽  
Ted Maksym ◽  
Hanumant Singh
Keyword(s):  
Neural Network ◽  
Sea Ice ◽  
Relative Error ◽  
Snow Depth ◽  
Large Scale ◽  
Small Scale ◽  
Ice Thickness ◽  
Sea Ice Thickness ◽  
Mean Relative Error ◽  
Antarctic Sea Ice

Abstract. Satellites have documented variability in sea ice areal extent for decades, but there are significant challenges in obtaining analogous measurements for sea ice thickness data in the Antarctic, primarily due to difficulties in estimating snow cover on sea ice. Sea ice thickness can be estimated from surface elevation measurements, such as those from airborne/satellite LiDAR, by assuming some snow depth distribution or empirically fitting with limited data from drilled transects from various field studies. Current estimates for large-scale Antarctic sea ice thickness have errors as high as ~ 50 %, and simple statistical models of small-scale mean thickness have similarly high errors. Averaging measurements over hundreds of meters can improve the model fits to existing data, though these results do not necessarily generalize to other floes. At present, we do not have algorithms that accurately estimate sea ice thickness at high resolutions. We use a convolutional neural network with laser altimetry profiles of sea ice surfaces at 0.2 m resolution to show that it is possible to estimate sea ice thickness at 20 m resolution with better accuracy and generalization than current methods (mean relative errors ~ 15 %). Moreover, the neural network does not require specifying snow depth/density, which increases its potential applications to other LiDAR datasets. The learned features appear to correspond to basic morphological features, and these features appear to be common to other floes with the same climatology. This suggests that there is a relationship between the surface morphology and the ice thickness. The model has a mean relative error of 20 % when applied to a new floe from the region and season, which is much lower than the mean relative error for a linear fit (errors up to 47 %). This method may be extended to lower-resolution, larger-footprint data such as such as IceBridge, and suggests a possible avenue to reduce errors in satellite estimates of Antarctic sea ice thickness from ICESat-2 over current methods, especially at smaller scale.

scholarly journals Estimating early-winter Antarctic sea ice thickness from deformed ice morphology

2019 ◽  
Vol 13 (11) ◽  
pp. 2915-2934 ◽  
Author(s):  
M. Jeffrey Mei ◽  
Ted Maksym ◽  
Blake Weissling ◽  
Hanumant Singh
Keyword(s):  
Neural Network ◽  
Sea Ice ◽  
Snow Depth ◽  
Large Scale ◽  
Field Studies ◽  
Small Scale ◽  
Ice Thickness ◽  
Sea Ice Thickness ◽  
Antarctic Sea Ice ◽  
Potential Applications

Abstract. Satellites have documented variability in sea ice areal extent for decades, but there are significant challenges in obtaining analogous measurements for sea ice thickness data in the Antarctic, primarily due to difficulties in estimating snow cover on sea ice. Sea ice thickness (SIT) can be estimated from snow freeboard measurements, such as those from airborne/satellite lidar, by assuming some snow depth distribution or empirically fitting with limited data from drilled transects from various field studies. Current estimates for large-scale Antarctic SIT have errors as high as ∼50 %, and simple statistical models of small-scale mean thickness have similarly high errors. Averaging measurements over hundreds of meters can improve the model fits to existing data, though these results do not necessarily generalize to other floes. At present, we do not have algorithms that accurately estimate SIT at high resolutions. We use a convolutional neural network with laser altimetry profiles of sea ice surfaces at 0.2 m resolution to show that it is possible to estimate SIT at 20 m resolution with better accuracy and generalization than current methods (mean relative errors ∼15 %). Moreover, the neural network does not require specification of snow depth or density, which increases its potential applications to other lidar datasets. The learned features appear to correspond to basic morphological features, and these features appear to be common to other floes with the same climatology. This suggests that there is a relationship between the surface morphology and the ice thickness. The model has a mean relative error of 20 % when applied to a new floe from the region and season. This method may be extended to lower-resolution, larger-footprint data such as such as Operation IceBridge, and it suggests a possible avenue to reduce errors in satellite estimates of Antarctic SIT from ICESat-2 over current methods, especially at smaller scales.

Sign in / Sign up

Export Citation Format

Share Document