Close

Service Interruption

The NSIDC Web site and data services are currently having intermittent problems and may be unavailable. We are working to restore these services as soon as possible and apologize for any inconvenience this may cause. Please contact NSIDC User Services for assistance.

Submarine Upward Looking Sonar Ice Draft Profile Data and Statistics

Summary

This data set consists of upward looking sonar draft data collected by submarines in the Arctic Ocean. It includes data from both U.S. Navy and Royal Navy submarines. Maps showing submarine tracks are available. Data are provided as ice draft profiles and as statistics derived from the profile data. Statistics files include information concerning ice draft characteristics, keels, level ice, leads, undeformed and deformed ice. Data from the U.S. Navy's Digital Ice Profiling System (DIPS) have been interpolated and processed for release as unclassified data at the U.S. Army's Cold Regions Research and Engineering Laboratory (CRREL) in Hanover, New Hampshire. Data from the analog draft recording system were digitized and then processed by the Polar Science Center, Applied Physics Laboratory, University of Washington. Data from British submarines were provided by the Scott Polar Research Institute, University of Cambridge. All data sources used similar processing methods in order to ensure a consistent data set.

Access to the Submarine Upward Looking Sonar Ice Draft Profile Data and Statistics data set is unrestricted, but users are encouraged to register for the data. Registered users will receive e-mail notification about any product changes.

Citing These Data

National Snow and Ice Data Center. 1998, updated 2006. Submarine upward looking sonar ice draft profile data and statistics. Boulder, Colorado USA: National Snow and Ice Data Center. http://dx.doi.org/10.7265/N54Q7RWK

We kindly request that you cite the use of this data set in a publication using the following citation example. For more information, see our Use and Copyright Web page.

Overview Table

Category Description
Data format Data files are ASCII (text) format.
Spatial coverage and resolution Arctic Ocean (see Table 1)
Temporal coverage and resolution See Table 1
File size The entire data set is 150 MB.
Parameters Sea ice deformation
Sea ice draft/thickness
Sea ice roughness
Leads
Metadata access View metadata
Data access Data are available via FTP.

Table of Contents

  1. Contacts
  2. Overview
  3. Detailed Data Description
  4. Data Acquisition and Processing
  5. Data Access and Related Collections
  6. References and Related Publications
  7. Acknowledgements
  8. Document Information

1. Contact

NSIDC User Services
National Snow and Ice Data Center
CIRES, 449 UCB
University of Colorado
Boulder, CO 80309-0449  USA
phone: +1 303.492.6199
fax: +1 303.492.2468
form: Contact NSIDC User Services
e-mail: nsidc@nsidc.org

2. Overview

Background

This data set includes submarine data collected in the Arctic Ocean by U.S. Navy and Royal Navy submarines. U.S. Navy guidance has stated that previously classified, submarine-collected ice draft data may be declassified and released according to set guidelines. Those guidelines include restrictions stating that positions of the data must be rounded to the nearest 5 minutes of latitude and longitude, and date is to be rounded to the nearest third of a month. The guidelines also specify a region in which the data may be released. The Chief of Naval Operations has expanded the release area beyond the original "Gore Box" (so called because of Vice President Gore's advocacy for releasing the data). See the map below (click on the image to see the full size map).

Gore Box

The SCience ICe EXercise (SCICEX) is a program that uses U.S. Navy submarines for research. SCICEX data are not classified and do not have restrictions on reporting the precise location and date for the data; therefore the SCICEX ice draft data in this collection are reported with their date of acquisition, and position is reported to six decimal places. For more information about SCICEX, see the NOAA@NSDIC SCICEX Web site.

Since 1967 U.S. submarines have employed a narrow beam sonar transducer. Since 1976 data have usually been recorded digitally on U.S. Navy submarines with the Digital Ice Profiling System (DIPS). All U.S. Navy data in this data set come from the DIPS system, unless they are part of the analog portion. In processing, data are corrected for depth errors, erroneous drafts are removed, and data are spatially interpolated. The interpolation routine integrates submarine speed and position to obtain drafts at uniform spatial intervals. This is a labor-intensive interactive process, during which segments in which the submarine changed depth or course must be removed from the data. The majority of the cruise data were interpolated and processed for release as unclassified data at the U.S. Army's Cold Regions Research and Engineering Laboratory (CRREL) in Hanover, New Hampshire. SCICEX-97 and SCICEX-98 data were processed at the University of Washington, Polar Science Center, in cooperation with CRREL and using similar processing steps.

Data from British submarines were processed by the Scott Polar Research Institute (SPRI), University of Cambridge, in the same way as were the U.S. submarine data.

ULS draft data acquired on U.S. submarines prior to 1976 were recorded only as traces on paper rolls. In 1976 and thereafter, data were recorded both on analog paper roles and using DIPS. Polar Science Center investigators developed a method to scan and digitize the analog draft data so that they are as equivalent to the digitally recorded DIPS data as possible (Wensnahan and Rothrock, 2005). These data were added in 2006. This portion of the collection is referred to as the analog portion.

The map below shows submarine tracks from the non-analog portion of the data set (click on the image to see the full size map). To access the map legend, click here.

submarine tracks

3. Detailed Data Description

Background

Data are in two types of files, one for ice draft profiles, and the other for statistics derived from the profile data. Ice draft files include a header that gives date and location information followed by a sequential list of drafts spaced at 1.0 m intervals that comprise the bottom-side sea-ice roughness profile. Data in each file fall along a straight-line (great circle) track between the two end points given in the header. The length of the profile in any given file can be up to 50 km, but may be shorter if data dropouts create gaps greater than 0.25 km, or if changes in course cause deviations from a straight-line track. Statistics files include information on ice draft characteristics, keels, level ice, leads, un-deformed, and deformed ice. For background information on scientific uses of ice draft data such as these statistical measures of ice deformation, see Analysis of Arctic Ice Draft Profiles Obtained by Submarines, a note provided by W. Tucker and S. Ackley, CRREL, Hanover, NH, in July 1998.

Table 1 shows the cruise reference name (click to see cruise track), dates, number of segments, the size of the directory containing the data (after uncompressing and untarring), and examples of naming conventions for the data files. NSIDC is told how we may refer to each cruise by the data providers. We have agreed to adhere to this naming convention. Therefore NSIDC cannot provide the submarine names for all cruises to users of this data set. Note that permission was obtained to release some SCICEX-99 data acquired outside the previously mentioned release box. A legend for the submarine cruise tracks is also available.

Table 1. Cruise Information
Cruise Reference Name Start Date End Date Number of Draft Segments Size of Untarred/
Uncompressed File Directory
File Name Convention/
Examples
Provided to NSIDC for publication by* Date published by NSIDC Raw data source
1975 (analog) May 1975 May 1975       Wensnahan June 2006 September 2006 USchart
UK-76 (Gurnard) 07 April 1976 10 April 1976 27 22.0 MB 0476drft.002
0476stat.013
Davis February 1999 May 1999 UKDIPS
1976 (analog) April 1976 April 1976       Wensnahan June 2006 September 2006 USchart
1979 (analog) April 1979 April 1979       Wensnahan June 2006 September 2006 USchart
1981 (analog) October 1981 October 1981       Wensnahan June 2006 September 2006 USchart
1982a (analog) November 1982 November 1982       Wensnahan June 2006 September 2006 USchart
1983a (analog) August 1983 August 1983       Wensnahan June 2006 September 2006 USchart
1984b (analog) September 1984 September 1984       Wensnahan June 2006 September 2006 USchart
1984c (analog) November 1984 November 1984       Wensnahan June 2006 September 2006 USchart
1984d (analog) October 1984 November 1984       Wensnahan June 2006 September 2006 USchart
1986a May 1986 June 1986 111 19.1 MB 1986adrft.053
1986astat.090
February 2001 (?) December 2001 USDIPS
1986b 02 April 1986 03 April 1986 82 21.0 MB 1986bdrft.001
1986bstat.001
Tucker February 2001 (?) (original)

March 2001

USDIPS
Tucker May 2004 (corrected) July 2004
UK-87 (analog) 08 May 1987 26 May 1987 130 82.8 MB 0587drft.a41
0587stat.b13
Davis February 1999 May 1999 UK, A/H**
1987 02 April 1987 03 April 1987 64 17.7 MB 1987drft.035
1987stat.035
February 2001 March 2001 USDIPS
1987c (analog) May 1987 June 1987       Wensnahan June 2006 September 2006 USchart
1988a 03 May 1988 03 May 1988 32 10.5 MB 1988drft.050
1988stat.064
Tucker March 2000 (original)

May 2000

USDIPS
Tucker May 2004 (corrected) July 2004
1988b 01 August 1988 03 August 1988 47 12.9 MB 1988bdrft.018
1988bstat.018
February 2001 March 2001 USDIPS
1988c (analog) April 1988 May 1988       Wensnahan June 2006 September 2006 USchart
1989b September 1989 September 1989 47 13.6 MB 1989bdrft.018
1989bstat.099
Tucker April 2002 June 2002 USDIPS
1990 March 1990 April 1990 35 5.3 MB 1990drft.131
1990stat.143
Tucker November 2001 December 2001 USDIPS
1990c (analog) September 1990 September 1990       Wensnahan June 2006 September 2006 USchart
UK-91 (analog) 20 April 1991 22 April 1991 16 11.6 MB 0491drft.012
0491stat.021
Davis February 1999 May 1999 UK, A/H**
1991 03 March 1991 02 May 1991 142 20.2 MB 1991drft.047
1991stat.107
Tucker March 2000 May 2000 USDIPS
Grayling-1992 April 1992 April 1992 9 2.9 MB g92drft.032
g92stat.0431992a
Tucker February 1998 (original)

February 1998

USDIPS
Tucker May 2004 (corrected) July 2004
1992a May 1992 May 1992 17 2.9 MB 1992adrft.015
1992astat.021
Tucker September 2001 December 2001 USDIPS
1992b August 1992 September 1992 38 8.8 MB 1992badrft.038
1992bstat.027
Tucker September 2001 December 2001 USDIPS
L2-92 02 April 1992 02 April 1992 64 16.5 MB L292drft.010
L292stat.052
Eppler October 1998 November 1998 USDIPS
SCICEX-93 01 September 1993 12 September 1993 139 43.2 MB sc93drft.041
sc93stat.131
Eppler October 1998 November 1998 USDIPS
1993 02 April 1993 03 April 1993 86 24.1 MB 1993drft.034
1993stat.034
Tucker February 2001 March 2001 USDIPS
1993c (analog) April 1993 April 1993       Wensnahan June 2006 September 2006 USchart
1994 01 April 1994 01 April 1994 85 30.1 MB 1994drft.146
1994stat.161
Tucker September 1999 October 1999 USDIPS
1994b (analog) September 1994 September 1994       Wensnahan June 2006 September 2006 USchart
SCICEX-96 20 September 1996 22 October 1996 217 64.9 MB sc96drft.095
sc96stat.141
Eppler February 1999 (original)

March 1999

USDIPS
Tucker May 2004 (corrected) July 2004
SCICEX-97 03 September 1997 02 October 1997 217 64.9 MB sc97drft.111
sc97drft_sheba.15
sc97stat.054
sc97stat_sheba.167
Yu September 1999 (original)

October 1999

USDIPS
Yu June 2002 (corrected) June 2002
SCICEX-98

Note
02 August 1998 16 August 1998 129 44.2 MB sc98drft.020
sc98drft_sheba.036
sc98stat.116
sc98stat_sheba.028
Yu May 2002 June 2002 USDIPS
SCICEX-99
02 April 1999 13 May 1999 41 (plus subsegments) 119.0 MB sc99drft.404_002.002
sc99stat.404_002.002
Tucker November 2004 June 2005 USDIPS
2000a (analog)
SCICEX SAM
October 2000 October 2000   50 MB   Wensnahan June 2006 September 2006 USchart
2005a (analog) July 2005 July 2005   11 MB   Wensnahan April 2010 December 2010 USchart
2005e (analog)
SCICEX SAM
November 2005 November 2005   30 MB   Wensnahan April 2010 December 2010 USchart

*Also see acknowledgements

**Analog/Hand-digitized

The description of the data in this section is applicable to the non-analog portion of the data set. The analog portion of the data set is described in a document provided by M. Wensnahan: Documentation for G01360 Analog Portion (PDF, 115 KB). This documentation will be integrated with the rest of the product documentation when resources allow.).

File Naming Convention

For non-analog U.S. Navy cruise data, the file name begins with four characters denoting the cruise. The next four characters are either drft (for draft files) or stat (for statistics files). Each file name is followed by a three-digit extension that corresponds to an ice segment. The extensions were assigned in the order in which the segments were acquired by the submarine. Each draft file contains data for one ice segment. Each statistics file contains data (19 parameters) for one ice segment. For SCICEX-97 and SCICEX-98 data, data files for segments acquired in the vicinity of the Surface Heat Balance of the Arctic Ocean (SHEBA) experiment have _sheba added to their names. For SCICEX-99 data, the naming convention is as follows: sc99drft.404_002.002 indicates data collected on April 4 (404), and this is the second segment processed for this day (_002). The segment required processing in parts, and this is the second (.002) part of the segment.

See Notes on U.K. Data Files, and Documentation for G01360 Analog Portion (PDF, 115 KB) for information on the naming convention for Royal Navy and U.S. Navy analog portion files.

Format

Data files are ASCII text format.

File Header Format and Information

File headers at the beginning of the draft and statistics files give the following information concerning the data segment from which information in the archive file was generated (Fig. 1):

  1. The name of the source file from which the data were generated.
  2. Date information consisting of the exact date (year, month, and day) the data were acquired, for SCICEX data; in the case of previously classified data, the year, month, and the third of the month (1=Days 1 to 10, 2=Days 11 to 20, 3=Days 21 to 31) in which the data were acquired.
  3. Geographic coordinates of the first and last drafts in the file; in the case of previously classified data, the coordinates are rounded to the nearest 0.1° north latitude and east longitude.
  4. The number of drafts in the profile segment.
  5. The length of track included in the profile segment in kilometers.
  6. The length and location within the draft file of any gaps that exceed 10.0 m in length. Profile length is the great circle distance between unrounded latitude-longitude coordinates of the first and last drafts, for both SCICEX and previously classified data. Individual draft measurements are equally spaced at approximately 1.0 m intervals along the great circle arc.

    Note that the number of draft values at 1.0 m spacing can sometimes be greater or less than the profile length in meters. Two reasons for this are rounding error (draft values are nominally every 1 meter, but may be slightly more or less), and the fact that minor turns or turns of short duration in the submarine track may not have been edited from the data record.

 

SOURCE FILE: lnall.026

--------DATE--------

          Year: 1992

         Month:  APR

Third of Month:    2

--------------------

-----SEGMENT DESCRIPTION-----

  Beginning Latitude:  80.8

 Beginning Longitude: 210.3

     Ending Latitude:  81.2

    Ending Longitude: 210.2

    Number of Drafts:   49999

Length of Track (km):  49.999

-----------------------------

__________________________________

     DATA GAPS IN DRAFT FILE      

 (The following gaps greater than 

     0.010 km were detected.)       

----------------------------------

  0.012 km gap follows draft 35328

----------------------------------

  3.45

  2.99

  2.92

  3.27

  2.89

  2.62

  3.02

  2.99

  2.75

  2.91

    .

    .

    .

    .

    .	

        
Figure 1. Ice draft file showing header and first 10 drafts in meters

Notes on U. K. Data Files

The header file for the U.K. data (the equivalent of Fig. 1) has a slightly different format. The naming convention for the U.K. files is XXYYzzzz.xyy, where XX designates month, YY is the year, zzzz is drft or stat, for draft or statistics file, x is a placeholder that designates which survey of the cruise the data are from when a cruise has more than one survey, and yy is the segment number.

The 1976 SPRI data are from USS Gurnard in the Beaufort Sea with approximate latitude/longitude coordinates supplied. Centroids were not determined. An experimental narrow-beam sonar was used (Wadhams and Horne, 1980).

The 1987 SPRI data are from the HMS Superb in the Greenland Sea & Eurasian basin. Data are in two legs (a and b). Segments a03 and b46 have insufficient data for analysis. For segments b45, b55, b58, b60 and b61 the ice regime is not conducive to standard analyses; therefore, these segments were processed with level ice slope of 0.05 and minimum lead width set to zero.

File Size

The entire data set is 150 MB.

4. Data Acquisition and Processing

The description of the data in this section is applicable to the non-analog (digital) portion of the data set. The analog portion of the data set is described in a document provided by M. Wensnahan: Documentation for G01360 Analog Portion (PDF, 115 KB).

Also see Processing of the SCICEX '98 Submarine Data, by Y. Yu and S Dickinson, for information on corrections that needed to be applied to the SCICEX 98 data due to errors that were caused by an improperly working depth gauge.

Ice Draft Files

In order to statistically analyze these data, they were interpolated to even spatial intervals. The raw, digital data contain information only about ice draft and time, which is not useful for statistical, fractal, or spectral analysis. To obtain ice drafts at uniform spatial intervals, the speeds and positions of the submarine were integrated with the interpolation routine. Segments of the data during which the submarine changed course and/or depth were removed. For some cruises, only segments greater than 10 km in straight-line length were retained for this data set.

Raw top-sounder profiles, from which data presented here are derived, were created by sampling ice draft with top-sounder profilers at intervals spaced equally in time as the submarine moved beneath the ice cover. Adjacent drafts in the raw profile, though recorded at intervals that are constant in time, represent spot measurements separated by non-constant distances, the length of which vary with changes in vessel speed. In this raw format, profiles from different tracks (or even from different segments of the same track) are not directly comparable because the same feature (keel, lead, etc.) sampled twice will have a different shape depending on whether the sensor platform was moving rapidly or slowly. Keels and other roughness elements in raw top-sounder profiles thus appear compressed at high speeds, and stretched out at low speeds. Such apparent differences in sampling rate bias summary statistics (mean draft, variance, etc.) and spectral characteristics (Fourier transforms, auto- and cross-correlation, etc.) because the bottom-side ice profile represented in one section of data is over- or under-sampled with respect to that in another section.

To eliminate this problem, interpolated profiles composed of drafts spaced equally in distance (as opposed to time) are created. Navigation data combined with speed and bearing information give good estimates of the geographic location of each draft. Great circle distances between points, calculated from geographic coordinates using standard mapping equations, provide a basis for interpolating a derivative set of equidistant drafts using a cubic spline algorithm [spline( ) and splint( )] (Press et al. 1992). The interpolated profiles that result, consisting of drafts spaced equally with respect to distance (nominally 1.0 m apart), form the basis of this data archive (Fig. 1).

Individual ice draft files represent data acquired continuously over straight-line tracks that span distances up to 50 km in length. Data acquired while the vessel was turning have been removed. Gaps within archived profiles, resulting from dropouts and other sensor malfunctions, are shorter than 0.25 km; their length and location within the profile is noted in header information described above. When gaps greater than 0.25 km in length were encountered, one file was closed and the next opened. Draft measurements are given in meters, and the distance between consecutive drafts is 1.0 m.

Ice Statistics Files, Ice Draft PDF, and General Statistics

Basic statistical analysis was performed on the processed, interpolated data. Data of lengths 10 km to 50 km were retained. Although 50 km segments are preferable (Wadhams 1984), shorter segments were included because they add value to the data set, especially in regions where the ice morphology changes rapidly. Because these shorter segments were included, caution must be exercised when analyzing regional, seasonal, and interannual variations. The statistics data files are ASCII text files. Probability Density Functions (PDFs) (Fig. 2) are derived from the frequency distribution of all drafts in the track segment. Bin width is 0.1 m. Counts in each bin are normalized by the total number of drafts in the segment to give the probability of occurrence of drafts of any given depth. Bins for which no drafts occur have probability of 0.0 and are omitted from the listing to save storage space (see, for example, BIN 276 for drafts between 27.5 and 27.6 m, Fig. 2). This convention is used for all other pdfs in the statistics archive.

General statistics calculated for ice drafts in each segment include standard parametric descriptors of central tendency and dispersion (mean and median draft, variance, standard and average deviation, standard error, skewness, kurtosis, and root-mean-square draft, see Fig. 3). Note that the mean is that of all ULS measurements, including open water.

 

PROBABILITY DENSITY

-FUNCTION OF ICE DRAFTS-

  Bin Width (m):    0.1

 Number of Bins:    279

|----|-----|-----------|

      LOWER             

      BOUND             

  BIN  (m)  PROBABILITY 

|----|-----|-----------|

    1   0.0  0.00630013

    2   0.1  0.00110002

    3   0.2  0.00094002

    4   0.3  0.00062001

    5   0.4  0.00064001

    6   0.5  0.00086002

    7   0.6  0.00264005

    8   0.7  0.00220004

		.

		.

		.

  274  27.3  0.00004000

  275  27.4  0.00006000

  277  27.6  0.00004000

  278  27.7  0.00002000

  279  27.8  0.00002000

|----|-----|-----------|

        
 

----GENERAL DRAFT STATISTICS----

              Mean (m):    3.250

            Median (m):    2.200

 Average Deviation (m):    1.700

Standard Deviation (m):    2.627

    Standard Error (m):    0.012

              Variance:    6.904

              Skewness:    3.263

              Kurtosis:   15.624

         RMS Draft (m):    4.179

--------------------------------

        
Figure 2. Probability density function (pdf) of ice drafts Figure 3. Ice draft statistics

Specific formulae used to calculate these values are as follows (code used in these calculations borrows heavily from that given in the moment ( ), select ( ), and middle ( ) functions of Press et al.,1992):

formulae used to calculate values

Autocorrelation

Function Autocorrelation measures the correlation between pairs of consecutive drafts within a profile. Pairs may consist of adjacent drafts, or drafts separated by a particular distance (lag). This process compares the ice draft profile with itself. Successive comparisons with increasing values of lag in effect slide the profile past itself and allow one to determine whether periodicities exist that lead to higher correlations at some offsets than at others. Such periodicities, if they exist, may arise from periodic noise in the profile, or may reflect geophysical phenomena that produce recurring features.

First-order autocorrelation considers correlation between the set of all pairs of adjacent drafts:

(X1, X2), (X2, X3), (X3, X4), ....., (Xi-1, Xi), ....., (Xn-1, Xn).

This assumes that the distance between consecutive drafts is constant; drafts used here are interpolated to a nominal spacing of 1.0 m, so this requirement is met. Higher order autocorrelations are calculated in sequence by comparing pairs of drafts separated by successively greater distance or lag. In the case where lag=2, for example, the set of adjacent pairs is represented by:

(X1, X3), (X2, X4), (X3, X5), ....., (Xi-2, Xi), ....., (Xn-2, Xn),

and for lag=5:

(X1, X6), (X2, X7), (X3, X8), ....., (Xi-5, Xi), ....., (Xn-5, Xn).

Autocorrelation r as a function of lag is defined as:

.

The analog to this procedure in conventional correlation analysis is calculation of a correlation coefficient associated with the cluster of points produced by plotting, in a scatter diagram, all possible pairs of drafts that are separated by a given lag. The statistics archive lists autocorrelation as a function of lag from 0 to 150, inclusive (Fig. 4). Inasmuch as the spatial separation between individual draft measurements is 1.0 m, this corresponds to a range of lags from 0.0 m to 150.0 m. In addition, a variable called Correlation Length, defined as the lag at which rlag less than or equal to 1/e, is given as a basis for making general comparisons between autocorrelation functions calculated for different profile segments.


______________

AUTOCORRELATION

---------------

Evaluated Lags From 0 To 150

Criterion for Correlation Length = 0.367879

Correlation Length = 45

|----|--------|

  LAG  R[LAG]

|----|--------|

    0  1.00000

    1  0.98696

    2  0.96349

    3  0.94327

    4  0.92245

    5  0.90072

    6  0.87947

    7  0.85825

    8  0.83758

    9  0.81778

	.

	.

	.

    
Figure 4. Autocorrelation function
Keels

Keel detection is accomplished using an algorithm developed by A.W. Lohanick (unpublished). Lohanick's routine, which was originally written to detect ridges in laser profilometer data acquired during Project Birdseye, uses a Rayleigh criterion to identify local maxima (or, in the case of ice draft data, minima) that correspond to ridges (or keels). To qualify as a keel, an ice draft must be at least twice as deep as the local minimum draft measured from an undeformed ice datum (2.5 m), it must be the deepest draft among all local drafts, and it must be deeper than 5.0 m. Two or more keels that occur adjacent to each other are identified as independent features if they are separated by at least one draft that is less than half the depth of the first keel in the pair, as measured from the undeformed ice datum (2.5 m). Otherwise they are identified as a single feature with a draft equal to the local maximum.

Keels detected using this routine are listed in a table giving the record number at which the keel occurs in the draft file, the depth of the keel, and the distance to the previous keel (Fig. 5). Additional tables give pdfs of keel depths with a bin width of 1.0 m (Fig. 6) and of spacings between adjacent keels with a bin width of 50.0 m (Fig. 7). Summary statistics calculated for keel depths and keel spacings using equations given above for draft statistics give mean, median, maximum and minimum draft and spacing, average and standard deviation, and variance, skewness, and kurtosis (Fig. 8).

 

	      KEELS

-----------------------------------

Number of Drafts Examined:  49999

 Number of Keels Detected:    306

Minimum Keel Depth Cutoff:   5.00 m

     Undeformed Ice Datum:   2.50 m

-----------------------------------

 LIST OF DETECTED KEELS

|------|------|--------|

         KEEL    KEEL

 RECORD  DEPTH  SPACING

 NUMBER   (m)     (m)

|------|------|--------|

     15   8.04    44.12

     59   9.16    87.39

    147  17.09    25.45

    172   9.19   363.12

    535   9.01    74.66

    610   6.82    16.12

    626   6.06   121.32

    747   8.05    16.97

    764   8.97   263.86

   1028   5.24    55.99

   1084   5.92   878.11

   1962   8.68   244.34

   2206   8.07    43.27

   2249   8.58    53.45

   2303  10.36   215.49

   2518  11.25   145.07

   2663   6.31   617.62

   3281   5.45   266.39

   3547   7.34   466.60

   4014   5.29   112.83

   4127   6.21   719.42

   4846  10.37   100.11

   4946   6.36   399.58

   5344   5.01   246.03

   5590   7.48    78.90

   5669   5.03   525.14

		.

		.

		.

		.

		.

        
 

---PDF OF KEEL DEPTHS---

------------------------

  Bin Width (m):    1.0

 Number of Bins:     27

|----|-----|-----------|

      LOWER             

      BOUND             

  BIN  (m)  PROBABILITY 

|----|-----|-----------|

    6   5.0  0.21895425

    7   6.0  0.21895425

    8   7.0  0.14052288

    9   8.0  0.13398693

   10   9.0  0.06535948

   11  10.0  0.04248366

   12  11.0  0.05555556

   13  12.0  0.02941176

   14  13.0  0.02614379



   15  14.0  0.01307190

   16  15.0  0.00653595

   17  16.0  0.00980392

   18  17.0  0.00980392

   19  18.0  0.00653595

   20  19.0  0.00326797

   22  21.0  0.00653595

   23  22.0  0.00326797

   26  25.0  0.00653595

   28  27.0  0.00326797

|----|-----|-----------|

          
Figure 5. List of keels Figure 6. Probability density function of keel depths
 

---PDF OF KEEL SPACINGS---

--------------------------

  Bin Width (m):   50.0

 Number of Bins:     28

|----|-------|-----------|

       LOWER             

       BOUND             

  BIN   (m)   PROBABILITY 

|----|-------|-----------|

    1     0.0  0.26470588

    2    50.0  0.26470588

    3   100.0  0.15686275

    4   150.0  0.07516340

    5   200.0  0.04575163

    6   250.0  0.04575163

    7   300.0  0.02941176

    8   350.0  0.01960784

    9   400.0  0.01960784

   10   450.0  0.00326797

   11   500.0  0.01633987

   12   550.0  0.00326797

   13   600.0  0.01960784

   14   650.0  0.00653595

   15   700.0  0.00653595

   17   800.0  0.00326797

   18   850.0  0.00980392

   23  1100.0  0.00653595

   29  1400.0  0.00326797

|----|-------|-----------|

    
 

	     KEEL STATISTICS

------------------------------------------

                      KEEL        KEEL    

     STATISTIC      DEPTH (m)  SPACING (m)

------------------|-----------|-----------

              Mean       8.50      163.74

            Median       7.38       89.93

           Minimum       5.01        5.09

           Maximum      27.86     1411.66

 Average Deviation       2.56      132.80

Standard Deviation       3.63      198.89

          Variance      13.15    39556.33

          Skewness       2.23        2.71

          Kurtosis       6.39        9.18

------------------|-----------|-----------

    
Figure 7. Probability density function of keel spacings Figure 8. Keel depth and spacing statistics
Level Ice Segments

Level ice segments are defined as a series of consecutive drafts spanning a distance greater than 10 m in length over which the slope between any two adjacent drafts is less than or equal to 0.050 (Fig. 9). The magnitude of individual drafts is not a criterion. Level ice defined on this basis thus does not necessarily indicate thin ice or lead ice but can occur (and occasionally does occur) within thick first-year ice, multiyear ice, and regions of heavily deformed ice. Parameters given for each level ice segment include the record number within the draft file at which the segment begins, segment length, the mean of drafts within the segment, the mean of slopes between adjacent drafts within the segment, and the distance (spacing or separation) from the end of the previous level ice segment to the start of the current segment. Separate tables list pdfs of mean draft (Fig. 10), level ice spacing (Fig. 11), and level ice segment length (Fig. 12). The bin width used for mean draft pdfs is 0.5 m, for mean spacing pdfs is 50.0 m, and for mean level ice segment length is 10.0 m.


-------------LEVEL ICE SEGMENTS-------------

Criteria Used to Define Level Ice Segments:

	Maximum draft-to-draft slope: 0.050

	Maximum ice draft: NONE

	Minimum segment length:  10.0 m

|------|-------|-------|-------|-----------|

 FIRST  SEGMENT  MEAN           DISTANCE TO

 RECORD  LENGTH  DRAFT   MEAN    PREVIOUS

 NUMBER   (m)     (m)    SLOPE  SEGMENT (m)

|------|-------|-------|-------|-----------|

    879   14.01    2.28  0.0043       0.00 

   1667   11.01    1.97  0.0082     773.98 

   1747   11.01    2.09  0.0082      68.95 

   3199   10.01    2.07  0.0110    1441.47 

   3415   10.01    1.20  0.0090     205.94 

   3443   14.01    0.09  0.0071      18.01 

   3474   18.01    0.05  0.0089      17.01 

   3659   17.01    2.01  0.0106     167.02 

   3679   14.01    1.97  0.0093       3.00 

				.

				.

				.

    
 

PDF: LEVEL ICE MEAN DRAFT

-----------------------------

  Bin Width (m):    0.5

 Number of Bins:      6

|----|-----|----|-----------|

      LOWER                  

      BOUND                  

  BIN  (m)    N  PROBABILITY 

|----|-----|----|-----------|

    1   0.0    7  0.09333333

    2   0.5    2  0.02666667

    3   1.0    4  0.05333333

    4   1.5   35  0.46666667

    5   2.0   26  0.34666667

    7   3.0    1  0.01333333

|----|-----|----|-----------|

    
Figure 9. List of level ice segments Figure 10. Probability density function of mean draft in level ice segments
 

PDF: LEVEL ICE SEGMENT SPACINGS

-------------------------------

  Bin Width (m):   50.0

 Number of Bins:    108

|----|-------|----|-----------|

       LOWER             

       BOUND             

  BIN   (m)    N   PROBABILITY 

|----|-------|----|-----------|

    1     0.0   14  0.18666667

    2    50.0    9  0.12000000

    3   100.0    2  0.02666667

    4   150.0    3  0.04000000

			.

			.

			.

   46  2250.0    1  0.01333333

   53  2600.0    1  0.01333333

  109  5400.0    1  0.01333333

|----|-------|----|-----------|

    
 

-PDF: LEVEL ICE SEGMENT WIDTHS-

-------------------------------

  Bin Width (m):   10.0

 Number of Bins:      7

|----|-------|----|-----------|

       LOWER             

       BOUND             

  BIN   (m)    N   PROBABILITY 

|----|-------|----|-----------|

    2    10.0   70  0.93333333

    3    20.0    2  0.02666667

    4    30.0    2  0.02666667

    8    70.0    1  0.01333333

|----|-------|----|-----------|

    
Figure 11. Probability density function of separation between level ice segments Figure 12. Probability density function of the width of level ice segments
Leads

Leads are defined as a series of consecutive drafts, all of depth less than 0.3 m, that span a distance 10.0 m or greater in length. Parameters given for each lead segment include the record number within the draft file at which the segment begins, lead width, the mean of drafts within the segment, and the distance (spacing or separation) from the end of the previous lead to the start of the current lead (Fig. 13). Separate tables list pdfs of mean draft within leads (Fig. 14) and distance between adjacent leads (Fig. 15). The bin width used for pdfs of mean lead draft is 0.05 m, and for mean spacing pdfs is 50.0 m.

 

----------------LEADS---------------

Criteria Used to Define Leads:

	Maximum ice draft:   0.3 m

	Minimum ice draft:   0.0 m

	Minimum width:  10.0 m

|------|-------|-------|-----------|

 FIRST   LEAD    MEAN   DISTANCE TO

 RECORD  WIDTH   DRAFT   PREVIOUS

 NUMBER   (m)     (m)   SEGMENT (m)

|------|-------|-------|-----------|

   3438   25.02   0.094       0.00 

   3469   26.02   0.046       6.00 

   9209   93.96   0.020    5716.49 

  27943   18.02   0.166   18636.52 

  27964   12.01   0.177       3.00 

  28495   38.04   0.012     518.93 

  32759   22.02   0.042    4221.41 

  33409   18.02   0.048     627.98 

  36596   42.04   0.040    3179.51 

  42019   50.04   0.019    5377.14 


PDF: LEAD ICE MEAN DRAFT

-----------------------------

  Bin Width (m):    0.1

 Number of Bins:      3

|----|-----|----|-----------|

      LOWER                  

      BOUND                  

  BIN  (m)    N  PROBABILITY 

|----|-----|----|-----------|

    1 0.000    7  0.70000000

    2 0.050    1  0.10000000

    4 0.150    2  0.20000000

|----|-----|----|-----------|

Figure 13. List of leads Figure 14. Probability density function of mean draft in leads
 

------PDF: LEAD SPACINGS-------

-------------------------------

  Bin Width (m):   50.0

 Number of Bins:    372

|----|-------|----|-----------|

       LOWER             

       BOUND             

  BIN   (m)    N   PROBABILITY 

|----|-------|----|-----------|

    1     0.0    3  0.30000000

   11   500.0    1  0.10000000

   13   600.0    1  0.10000000

   64  3150.0    1  0.10000000

   85  4200.0    1  0.10000000

  108  5350.0    1  0.10000000

  115  5700.0    1  0.10000000

  373 18600.0    1  0.10000000

|----|-------|----|-----------|

Figure 15. Probability density function of distances between leads

The depth criterion used to define lead segments effectively excludes ice that has undergone significant deformation. Adjacent lead segments separated by short distances, although listed here as separate features, thus may be part of the same lead. In the absence of sound criteria with which to distinguish ridged ice within a lead from thick ice between two adjacent but separate leads unambiguously, we leave it to the user community to establish their own rules to be applied to the draft profiles and lead statistics for discriminating between these two cases.

Undeformed and Deformed Ice Undeformed ice is defined as a series of consecutive drafts, all of depth less than 5.0 m, that span a distance 10.0 m or greater in length over which the slope between adjacent drafts does not exceed 0.050; deformed ice is all ice that is not classified as undeformed on the basis of these criteria. Undeformed and deformed ice segments are listed in different tables of the same format. Parameters given include record numbers within the draft file at which segments begin and end, segment width, the mean of drafts within the segment, the mean of slopes between adjacent drafts within each segment, and the distance (spacing or separation) from the end of the previous segment to the start of the current segment (Fig. 16). Separate tables list pdfs of mean draft within undeformed and deformed ice segments (Fig. 17), distance between adjacent segments (Fig. 18), and segment lengths (Fig. 19). The bin width used for pdfs of mean draft is 0.5 m, for mean spacing pdfs is 50.0 m, and for segment length is 10.0 m.

 

--------------UNDEFORMED ICE SEGMENTS--------------

Criteria Used to Define Undeformed Ice Segments:

	Maximum draft-to-draft slope: 0.050

	Maximum ice draft:  5.0 m

	Minimum segment length:  10.0 m

|------|------|-------|-------|-------|-----------|

    RECORD     SEGMENT  MEAN           DISTANCE TO

    NUMBER      LENGTH  DRAFT   MEAN    PREVIOUS

 (STRT)  (END)   (m)     (m)    SLOPE  SEGMENT (m)

|------|------|-------|-------|-------|-----------|

    878    892   14.01    2.28  0.0043       0.00 

   1666   1677   11.01    1.97  0.0082     773.98 

   1746   1757   11.01    2.09  0.0082      68.95 

   3198   3208   10.01    2.07  0.0110    1441.47 

   3414   3424   10.01    1.20  0.0090     205.94 

   3442   3456   14.01    0.09  0.0071      18.01 

   3473   3491   18.01    0.05  0.0089      17.01 

   3658   3675   17.01    2.01  0.0106     167.02 

   3678   3692   14.01    1.97  0.0093       3.00 

   3712   3730   18.01    1.85  0.0067      19.90 

   4151   4164   13.01    1.51  0.0108     420.99 

   4782   4798   16.01    2.17  0.0113     618.04 

   5269   5281   12.01    1.89  0.0092     472.92 

   6873   6897   24.02    2.07  0.0067    1592.02 

   7657   7670   12.90    2.14  0.0094     760.58 

   8088   8102   13.90    1.95  0.0101     418.09 

   8658   8669   11.01    1.12  0.0091     556.09 

   9214   9292   77.95    0.00  0.0022     544.97 

					.

					.

					.

					.

					.

 

PDF: UNDEFORMED ICE MEAN DRAFT

-----------------------------

  Bin Width (m):    0.5

 Number of Bins:      4

|----|-----|----|-----------|

      LOWER                  

      BOUND                  

  BIN  (m)    N  PROBABILITY 

|----|-----|----|-----------|

    1   0.0    7  0.09459459

    2   0.5    2  0.02702703

    3   1.0    4  0.05405405

    4   1.5   35  0.47297297

    5   2.0   26  0.35135135

|----|-----|----|-----------|

Figure 16. List of undeformed ice segments (List of deformed ice segments given in identical format) Figure 17. Probability density function of mean draft within undeformed ice segments

PDF: UNDEFORMED ICE SEGMENT SPACINGS

-------------------------------

  Bin Width (m):   50.0

 Number of Bins:    108

|----|-------|----|-----------|

       LOWER             

       BOUND             

  BIN   (m)    N   PROBABILITY 

|----|-------|----|-----------|

    1     0.0   14  0.18918919

    2    50.0    9  0.12162162

    3   100.0    2  0.02702703

    4   150.0    3  0.04054054

    5   200.0    7  0.09459459

    6   250.0    2  0.02702703

    7   300.0    1  0.01351351

    9   400.0    5  0.06756757

   10   450.0    3  0.04054054

   11   500.0    1  0.01351351

   12   550.0    2  0.02702703

   13   600.0    1  0.01351351

   15   700.0    2  0.02702703

   16   750.0    2  0.02702703

   17   800.0    1  0.01351351

   18   850.0    1  0.01351351

   21  1000.0    2  0.02702703

   23  1100.0    1  0.01351351

   26  1250.0    2  0.02702703

   27  1300.0    1  0.01351351

   29  1400.0    3  0.04054054

   32  1550.0    1  0.01351351

   35  1700.0    1  0.01351351

   38  1850.0    1  0.01351351

   39  1900.0    1  0.01351351

   41  2000.0    1  0.01351351

   45  2200.0    1  0.01351351

   46  2250.0    1  0.01351351

   53  2600.0    1  0.01351351

  109  5400.0    1  0.01351351

|----|-------|----|-----------|

 

PDF: UNDEFORMED ICE SEGMENT WIDTHS

-------------------------------

  Bin Width (m):   10.0

 Number of Bins:      7

|----|-------|----|-----------|

       LOWER             

       BOUND             

  BIN   (m)    N   PROBABILITY 

|----|-------|----|-----------|

    2    10.0   69  0.93243243

    3    20.0    2  0.02702703

    4    30.0    2  0.02702703

    8    70.0    1  0.01351351

|----|-------|----|-----------|

Figure 18. Probability density function of distance between adjacent undeformed ice segments Figure 19. Probability density function of the width of undeformed ice segments
Note on Data Intervals and Segment Length

The following information was added to the documentation on 21 August 2003. It was provided by D. Eppler, Bronson Hills Associates, on 27 September 1999 in response to a user's question regarding why the track distance based on track endpoints is sometimes less or greater than would be expected based on number of meter-spaced data values in that segment. The text provided by D. Eppler was edited slightly by F. Fetterer:

There are three possible explanations for why the track distance based on track endpoints is sometimes less or greater than would be expected based on number of meter-spaced data values in that segment. Two of the explanations arise from certain aspects of these data that cannot be changed. The third explanation involves errors we may have introduced by failing to detect turns in what we otherwise thought were long straight-line course segments.

1. Rounding Error: We create the profiles using an algorithm that converts time and speed in the raw data set to distance, which in turn allows us to apply a cubic spline technique to interpolate a series of equally spaced points (drafts) that are located 1.0 m apart. This entails a series of non-trivial calculations involving trig functions, square roots, and other library functions that introduce rounding errors. The nominal spacing between adjacent drafts thus is 1.0 m, but the actual spacing may be slightly greater than or less than this. I would expect that the sum of all errors over a long profile would approach 0.0 m, but this might not be the case. If, for example, the error tends to be negative more often than it is positive, the outcome would be a profile with more drafts in it than you would otherwise expect if the spacing was exactly 1.0 m between consecutive drafts. I do not think that the this type of imprecision in the exact location of a draft will have significant impact on most end-users of the data set, especially where the user is interested in summary statistics calculated for all drafts in an entire segment.

2. Navigation Uncertainty: We determine the location of drafts in the profile using a set of tie points taken from navigation logs provided to us by the Arctic Submarine Laboratory. At best, these points are recorded 30 minutes apart, but in some cases the time gap between successive points is on the order of an hour or more. That is to say that in the ideal case, we know exactly where the boat was twice in an hour; but we really don't know with certainty where the boat was in between successive navigation tie points. In the absence of conflicting information (from navigation notes, bearing or information recorded in the raw profile data set) we assume the course taken is a straight line between the successive tie points. As a check on this we look at the ship's heading that is recorded in the raw profile data provided us. If we see a course change, we break off the current straight-line segment and begin a new one after the turn ends. Recognize, however, that even a straight-line course typically deviates a bit--plus or minus two or three degrees from a mean heading is typical. Barring deviations greater that this we assume that a straight course is followed.

The straight line course between navigation points is of course the shortest distance between them. Given that we know the actual course is not perfectly straight, it is likely that many profiles will have more points in them than would be expected if the nominal spacing is absolutely constant at 1.0 m. A 50 km segment thus may in fact end up with slightly more than 50,000 drafts because, in reality, the boat sailed a distance further than 50 km to get to the next tie point.

3. We erred in creating the segments: Occasionally we err when we put together a segment by including data taken while the boat was turning. Abrupt, tight 360 degree turns where the boat changes course, circles, and then comes back immediately to it's previous course heading are common in some of the cruises. If we miss such momentary excursions from a straight line course, this leads to segments in which there are many more points than there should be for the distance supposedly traveled. We believe we removed most if not all of these bad segments, but some may have been overlooked.

5. Data Access and Related Collections

Data Access

Data are available via FTP.

Related NSIDC Data Collections

Segment length: Differences were always less than 6 m and most were less than 3 m. APL and BHA routines were consistent with respect to distances calculated from the raw top sounder data records.

Ice draft statistics: The mean and standard deviation compared well, but values of RMS draft departed significantly because the two software packages used different formulae for the RMS calculation.

Keel location: APL software selects more keels than the BHA software. Most discrepancies appear to arise from keel picks associated with broad keels characterized by multiple closely spaced peaks. APL software identifies these as separate keels and the BHA software a single keel.

Keel statistics: The APL software consistently provided mean keel drafts that exceeded the BHA values by 2.0 to 2.5 m. Standard deviations were consistent. This difference is thought to have occurred from the slightly different application of the Rayleigh criterion used for keel detection and APL interpolation methods (Fred Tanis, ERIM International, Yanling Yu, University of Washington, and Dennis Farmer, Bronson Hills Associates, provided this information.).

Other Related Collections

Mooring data from the Beaufort Gyre Exploration Project

6. References and Related Publications

References

Press, W. H., S. A. Teukolsky, W. T. Vetterling, and B. P. Flannery. 1992. Numerical recipes in C: the art of scientific computing. Cambridge University Press, 994.

Rothrock, D. A. and M. Wensnahan. 2007. The Accuracy of Sea-Ice Drafts Measured from U.S. Navy Submarines. Journal of Atmospheric and Oceanic Technology 24 (11): 1936-49. doi: 10.1175/JTECH2097.1.

Wadhams, P., and R. J. Horne. 1980. An analysis of ice profiles obtained by submarine in the Beaufort Sea. Journal of Glaciology 25: 401-424.

Wadhams, P. 1984. Arctic sea ice morphology and its measurement. Arctic Technology and Policy. I. Dyer and C. Chryssostomidis, eds., Washington, D.C., Hemisphere Publishing Corp., 179-195.

Wensnahan, M., and D. A. Rothrock, 2005. Sea-ice draft from submarine-based sonar: Establishing a consistent record from analog and digitally recorded data. Geophysical Research Letters 32, L11502, doi:10:1029/2005GL022507.

Related Publications

(This short bibliography of upward looking sonar references was compiled in 2000 and has not been substantially added to since then.)

Ackley, S. F., W. D. Hibler III, F. K. Kugzruk, A. Kovacs, and W. F. Weeks. 1976. Thickness and roughness variations of Arctic multiyear sea ice. CRREL Report 76-18, U.S. Army Cold Regions Research and Engineering Laboratory, Hanover, NH, 25 pp.

Alam, A. 2000. Surface turbulent fluxes over arctic leads. Ph.D. Thesis, University of Colorado, 234 pp.

Alam A., J. A. Curry, and M. A. Tschudi. 2001. A parameterization of the lead-width distribution and turbulent surface heat flux for arctic sea ice. J. Geophys. Res., in press, (SHEBA special issue).

Bourke, R. and C. Garrett. 1987. Sea ice thickness distribution in the Arctic Ocean. Cold Reg. Sci. and Technol. 13: 259-280.

Bourke, R. H., and R. G. Paquette. 1989. Estimating the thickness of sea ice, J. Geophys. Res. 94, 919-923.

Bourke, R. H. and A. S. McLaren. 1992. Contour mapping of Arctic Basin ice draft and roughness parameters. J. Geophys. Res. 97, 17,715-17,728.

Comiso, J. C., P. Wadhams, W. B. Krabill, R. N. Swift, J .P. Crawford, and W. B. Tucker III. 1991. Top/bottom multisensor remote sensing of arctic sea ice. J. Geophys. Res. 96 (C2): 2693-2709.

Davis, N. R. and P. Wadhams. 1995. A statistical analysis of arctic pressure ridge morphology, J. Geophys. Res., 100, C6 10915-10925.

Drucker, R., S. Martin, and R. Moritz. 2003. Observations of ice thickness and frazil ice in the St. Lawrence Island polynya from satellite imagery, upward looking sonar, and salinity/temperature moorings. J. Geophys. Res., 108: C5, 3149, doi:10.1029/2001JC001213.

Kerman, B., P. Wadhams, D. Norman, J. Comiso. 1999. Informational equivalence between synthetic aperture radar imagery and the thickness of Arctic pack ice. J. Geophys. Res. 104 29,721-29,731.

Kvambekk, A. S., and T. Vinje. 1992. Upward looking sonar ice draft series from the Greenland Sea. In Report of the Sea Ice Thickness Workshop (A.S. Thorndike, C. Parkinson, and D.A. Rothrock, eds.), Polar Science Center, University of Washington, Seattle, WA, B25-B28.

LeShack, L. A., W. D. Hibler III, and F. H. Morse. 1971. Automatic processing of Arctic pack ice data obtained by means of submarine sonar and other remote sensing techniques, in Propagation Limitations in Remote Sensing, (J.B. Lomax, ed.), AGARD Conference Proceedings No. 90, NATO Advisory Group for Aerospace Research and Development.

Lyon, W. K. 1984. Submarine exploration of the North Pole region; history problems, positioning and piloting. J. of Navigation 37 (2): 155-179.

McLaren, A. S., P. Wadhams, and R. Weintraub. 1984. The sea ice topography of M'Clure Strait in winter and summer of 1960 from submarine profiles. Arctic 37: 110-120.

McLaren, A. S., 1988. Analysis of the under-ice topography in the arctic basin as recorded by the USS Nautilus during August 1958. Arctic 41 (2): 117-126.

McLaren, A. S., 1989. The under-ice thickness of the Arctic Basin as recorded in 1958 and 1970, J. Geophys. Res. 94: 4971-4983.

McLaren, A. S., R. G. Barry, and R.H. Bourke, 1990. Could Arctic ice be thinning? Nature 345: 762.

McLaren, A. S., J. E. Walsh, R. H. Bourke, R. L. Weaver, and W. Wittmann. 1992. Variability in sea-ice thickness over the North Pole from 1977 to 1990. Nature, 358: 224-226.

McLaren, A. S., R. H. Bourke, J. E. Walsh, and R. L. Weaver. 1994. Variability in Sea-Ice Thickness Over the North Pole From 1958 to 1992. The Polar Oceans and Their Role in Shaping the Global Environment, pp. 363-371, The American Geophysical Union.

Melling, H., and D. A. Riedel. 1996. Development of seasonal pack ice in the Beaufort Sea during the winter of 1991-1992: A view from below. J. Geophys. Res. 101 (C5): 11,975-11,991.

Moritz, R. E. 1992. Sampling the temporal variability of sea ice draft distribution. Report of the Sea Ice Thickness Workshop (A.S. Thorndike, C. Parkinson, and D.A. Rothrock, eds.), Polar Science Center, University of Washington, Seattle, WA, B29-B38.

Rothrock, D. A., Y. Yu and G. A. Maykut, 1999. Thinning of the Arctic Sea-Ice Cover. Geophys. Res. Let. 26 (23): 3469.

Rothrock, D. A., J. Zhang, and Y. Yu. 2003. The arctic ice thickness anomaly of the 1990s: A sonsistent view from observations and models. J. Geophys. Res. 108(C3): 3083. doi:10.1029/2001JC001208.

Thorndike, A. S., C. Parkinson, and D.A. Rothrock (eds.). 1992. Report of the Sea Ice Thickness Workshop. Polar Science Center, University of Washington, Seattle, WA, 41pp + App.

Tucker, W. B. III, and W. D. Hibler III. 1986. Variability of Arctic sea ice drafts. Proc. of Second Workshop on Ice Penetration Technology, CRREL Special Rpt. 86-30, U.S. Army Cold Regions Research and Engineering Laboratory, Hanover, NH, 237-256.

Tucker, W. B. III, R. Anderson, J. Newton, C. Wales, G. Newton, and T. Luallin. 1992. Accuracy of submarine ice draft measurements. Report of the Sea Ice Thickness Workshop (A.S. Thorndike, C. Parkinson, and D.A. Rothrock, eds.), Polar Science Center, University of Washington, Seattle, WA, B22-B24.

Tucker, W. B. III, J. W. Weatherly, D. T. Eppler, D. Farmer and D. L. Bentley. 2001. Evidence for the rapid thinning of sea ice in the western Arctic Ocean at the end of the 1980s. Geophys. Res. Let.28 (14): 2851-2854.

Wadhams, P., and R. J. Horne. 1980. An analysis of ice profiles obtained by submarine in the Beaufort Sea. J. Glaciol. 25: 401-424.

Wadhams, P. 1981. Sea-ice topography of the Arctic Ocean in the region 70°W to 25°E. Philos. Trans. R. Soc. London Ser. A (302): 45-85.

Wadhams, P. 1984. Arctic sea ice morphology and its measurement, In Arctic Technology and Policy (I.Dyer and C. Chryssostomidis, eds.) Washington, DC, Hemisphere Publishing Corp., 179-195.

Wadhams, P., A. S. McLaren, and R. Weintraub. 1985. Ice thickness distribution in Davis Strait from submarine profiles. J. Geophys. Res. 90: 1069-1077.

Wadhams, P., and T. Davy. 1986. On the spacing and draught distributions for pressure ridge keels. J. Geophys. Res. 91: 10,697-10,708.

Wadhams, P. 1990. Evidence for thinning of the Arctic ice cover north of Greenland. Nature 345: 795-797.

Wadhams, P., N. R. Davis, J. C. Comiso, R. Kutz, J. Crawford, G. Jackson, W. Krabill, C. B. Sear, R. Swift and W. B. Tucker. 1991. Concurrent remote sensing of arctic sea ice from submarine & aircraft. Intnl. J. Remote Sensing 12 (9): 1829-1840.

Wadhams, P. 1992. Sea ice thickness distribution in the Greenland Sea and Eurasian Basin, May 1987. J. Geophys. Res. 97: 5331-5348.

Wadhams, P., W. B. Tucker III, W. B. Krabill, R. N. Swift, J. C. Comiso, and N. R. Davis. 1992. Relationship between sea ice freeboard and draft in the Arctic Basin and implications for ice thickness monitoring. J. Geophys. Res. 97: 20,325-20,334.

Wadhams, P. and N. R. Davis. 1994. The fractal properties of the underside of sea ice. Marine, Offshore and Ice Technology (Murthy, Wilson and Wadhams, eds.),. Fifth International Conference on Computer Aided Design, Manufacture and Operation in the marine and Offshore Industries, Incorporating the Fourth Ice Technology Workshop. Computational Mechanics Publications, Southampton & Boston: 353-363.

Wadhams, P., 1995. Arctic sea ice extent and thickness. Phil. Trans. R. Soc. Lond. A 352: 301-319.

Wadhams, P., 1997. Ice thickness in the Arctic Ocean: the statistical reliability of experimental data. J. Geophys. Res. 102: 27,951-27,959.

Wadhams, P., and N. R. Davis. 2000. Further evidence of ice thinning in the Arctic Ocean. Geophysical Research Letters 27 (24): 3973-3975

Wadhams, P., and R. J. Horne. 1978. An analysis of ice profiles obtained by submarine sonar in the AIDJEX area of the Beafort Sea. Scott Polar Research Institute Technical Report 78-1. Cambridge: SPRI.

Walsh, J. E., R. S. Weaver, R. Colony, A. S. McLaren, and R. H. Bourke. 1995. Correlation of ice thickness variability to atmospheric circulation over the North Pole 1958-1992. Proc. of Fourth Conf. on Polar Meteorol

Wensnahan, W., D. Rothrock, and P. Hezel. 2007. New Arctic sea ice draft data from submarines. EOS Transactions American Geophysical Union 88 (5), doi:10.1029/2007EO050003.

Williams, E., C. Swithinbank, and G. de Q. Robin. 1975. A submarine study of Arctic pack ice. Glaciol. 73: 349-362.

Williams, M. D. 1998. Submarines under ice: the U.S. Navy's polar operations. Naval Institute Press, Annapolis, MD, xii, 223.

Winsor, P. 2001. Arctic sea ice thickness remained constant during the 1990s. Geophysical Research Letters 28 (6): 1039-1041.

7. Acknowledgements

The National Science Foundation Office of Polar Programs project "Analysis of Arctic Ice Draft Profiles Obtained by Submarines," W. B. Tucker III and S. F. Ackley, principal investigators, supported preparation of data for those cruises identified in Table 1 as provided to NSIDC by Tucker or Eppler. The upward looking sonar data were interpolated and processed for release as unclassified data at the U.S. Army's Cold Regions Research and Engineering Laboratory (CRREL) in Hanover, New Hampshire. The U.S. Navy's Arctic Submarine Laboratory (ASL) provided the original data to CRREL. ASL approved declassification on behalf of the Chief of Naval Operations. Software and processing algorithms for these data were developed by D. Eppler and D. Farmer of Bronson Hills Associates, Hanover, NH. Bronson Hills Associates also provided technical documentation.

Preparation of the U.K. data (identified in Table 1 as provided to NSIDC by Davis) was funded by a subcontract under the same National Science Foundation Office of Polar Programs project "Analysis of Arctic Ice Draft Profiles Obtained by Submarines." The data were processed by the Scott Polar Research Institute, University of Cambridge, with the cooperation of the Royal Navy and the U.K. Hydrographic Office. N.R Davis and P. Wadhams were involved in the production of the U.K. data. SCICEX-97 and SCICEX-98 data were provided by D.A. Rothrock and Y. Yu with the support of National Science Foundation (OPP-9617343). These are identified in Table 1 as provided by Yu. The original data were provided by the U.S. Navy's Arctic Submarine Laboratory and were subsequently processed at the Polar Science Center, Applied Physics Laboratory, University of Washington. The software and processing algorithms were provided by B. Markham of ASL and by Bronson Hills Associates, making the data compatible with other submarine data archived previously by NSIDC. SCICEX-99 data were delivered to NSIDC by Tucker, and were processed by Bronson Hills Associates (D. Farmer) through the support of the Applied Physics Laboratory, University of Washington, and NSF grant OPP-9910331.

The U.S. analog data were processed at the Polar Science Center at the University of Washington and provided with documentation by M. Wensnahan and D. A. Rothrock (identified in Table 1 as provided to NSIDC by Wensnahan). These data were prepared with funding from NSF Office of Polar Programs grant OPP-9910331.

Researchers making use of these invaluable data owe a debt of gratitude to the present and past staff of the Arctic Submarine Laboratory, San Diego, California, for their long-term stewardship of the data. Without guidance from ASL, and in particular without the collaboration of D. Bentley, J. Gossett, and T. Luallin release of these data to the scientific community would not be possible. The Arctic Submarine Laboratory holds raw data from all U.S. submarine cruises beginning with the first cruise under the ice in 1958.

This data set is maintained at NSIDC with support from the NOAA NESDIS National Geophysical Data Center.

8. Document Information

Acronyms

Table 2 lists acronyms used in this document.

Table 2. Acronyms
Acronym Description
APL Applied Physics Lab
ASL Arctic Submarine Lab
BHA Bronson Hills Associates
CRREL Cold Regions Research and Engineering Laboratory
DIPS Digital Ice Profiling System
ERIM Environmental Research Institute of Michigan
EWG Environmental Working Group
NESDIS National Environmental Satellite, Data, and Information Service
NOAA National Oceanic and Atmospheric Administration
NSF National Science Foundation
NSIDC National Snow and Ice Data Center
PDF Probability Density Function
PSC Polar Science Center
SAM Science Accommodation Mission
SCICEX Science Ice Exercise
SPRI Scott Polar Research Institute
SHEBA Surface Heat Balance of the Arctic Ocean

 

Document Authors

This documentation was originally drafted by NSIDC’s M. Marquis, based on information and incorporating written documentation provided by D. Eppler, Bronson Hills Associates. Supplementary documentation (information linked in the documentation) was provided by Y. Yu and S. Dickinson (Processing of the SCICEX '98 Submarine Data, on 14 May 2002), by W. Tucker and S. Ackley, (Analysis of Arctic Ice Draft Profiles Obtained by Submarines on 6 July 1998) and by M. Wensnahan (Documentation for G01360 Analog Portion, 17 July 2006).

Document Creation Date

24 July 1998

Document Revision Date

January 2011: A. Windnagel added a table of acronyms and links to the SCICEX Web site.
December 2010: A. Windnagel added two new cruises to Table 1: 2005a and 2005e.
April 2008: L. Ballagh added new submarine track images created by B. Raup.
February 2007: L. Ballagh fixed three broken links in Table 1.
2006: F. Fetterer extensively edited and reformatted the documentation. This revision to the documentation coincided with the addition of the analog portion of cruise data.
The documentation was minimally revised as data from several additional cruises were provided by CRREL and the Polar Science Center after the initial 1998 release.

Document URL

http://nsidc.org/data/docs/noaa/g01360_upward_looking_sonar/index.html