Davis Strait Freshwater Flux Project
2006-2007 Upward Looking Sonar Sea-Ice Draft Data
NSF Grant OPP-02300381
APL/UW Mark 2 Upward Looking Sonar (ULS) Sea-Ice Draft Data
on the Davis Strait C2 Mooring, 2006-2007. Version 1.0
This archive contains the sea-ice draft data from one APL/UW ULS moored
at a fixed level for one year at site "C2" in Davis Strait. Other
instruments were installed on the mooring, including Aanderaa
Recording Current Meters, Sea-Bird Temperature-Conductivity
Recorders, and an RDI Acoustic Doppler Current Profiler (ADCP). Data from
these other instruments are collected in separate archive files.
-----------------------
Davis Strait 2006-2007 Bottom-Anchored Mooring
Mooring site "C2"
Position = 66° 45.66' N 60° 04.46' W
Corrected ocean depth = 700m (approx)
Mooring Deployed 2006, 16 October, 06:32:30 GMT
Mooring Recovered 2007, 4 October, 21:00:00 GMT
-----------------------
Nominal depth Instrument (serial number) Filename
------------- ------------------------- ---------------------------------------
83 m APL/UW ULS Mark 2 (s/n 2303) ULS_2303_DAVIS_ST_2006_2007_5min_V1.txt
83 m APL/UW ULS Mark 2 (s/n 2303) ULS_2303_DAVIS_ST_2006_2007_10s_V1.txt
Each instrument is represented by 2 ASCII files, with observations
sampled at 5-minute intervals and at 10-second intervals, respectively.
The 5-minute sampling is continuous, except for the observation at the
end of each hour, i.e. there are 11 5-minute sample per hour;
the 10-second sampling occurs during each of two 25-minute intervals
each day. Each file contains header information followed by the data
in space-delimited columns.
-------------------------------------------------------------------------------
Synopsis:
As part of the Davis Strait Freshwater Flux Array (NSF Grants OPP-9910305 and
OPP-0352754), oceanographic moorings are anchored to the ocean floor beneath
the drifting ice near in Davis Strait. Vertically distributed instruments
measure ocean properties at fixed depths, recording internally. The data are
retrieved annually when the mooring is recovered, and new moorings deployed.
The first Davis FW moorings were deployed in 2004. This archive contains the
sea-ice draft and depth data from the APL/UW Mark 2 ULS on the 2006-2007
mooring at site "C2".
For further information contact:
Dr. Richard E. Moritz dickm@apl.washington.edu (206) 543-8023
Polar Science Center, Applied Physics Lab, University of Washington
1013 NE 40th, Seattle, WA 98105-6698 USA FAX (206) 616-3142
-------------------------------------------------------------------------------
Meta data
=========
1) Platform: Bottom-anchored oceanographic mooring in Davis Strait.
2) Project title: AN OBSERVATIONAL ARRAY FOR HIGH RESOLUTION,
YEAR-ROUND MEASUREMENTS OF VOLUME, FRESHWATER, AND ICE FLUX
VARIABILITY IN DAVIS STRAIT
3) Data collection dates: October 2006 to September, 2007
4) PI: Richard E. Moritz (dickm@apl.washington.edu)
5) Data collection method:
Moored APL/UW Mark 2 Upward Looking Sonar sampling continuously at
5-minute intervals, and sampling at 10-second intervals for two
25-minute periods each day (00:00:00-00:25:00 GMT and 12:00:00-
12:25:00 GMT). The ULS measures the in-situ water pressure and
the range to the sea-ice bottom or open ocean surface directly
above the ULS. Depth is computed hydrostatically from the
pressure measurement. Sea-ice draft is computed as the difference
between depth and range. The draft of sea-ice directly
above the ULS changes over time as the sea-ice moves horizontally
on the surface of the ocean.
6) Data calibration method:
Calibration of the pressure gauge (for estimating depth) was performed by
ParosScientific Inc., Redmond WA. In general, pre-deployment calibrations
were accepted. Calibration of the beam pattern and sensitivity of the
acoustic transducer and lens system were performed at APL/UW using a
standard hydrophone/transmitter (Model E37, serial number 003).
ULS depth is calculated from the hydrostatic equation:
depth = (Puls - Ps)/(rhow * g) where
Puls = Pressure measured by the Paroscientific gauge
Ps = Sea Level Air Pressure interpolated to the mooring
position and the ULS observation time, from the NCEP
analyses available at 0000 and 1200 GMT
rhow = Average density of the water column above the ULS.
g = Gravitational acceleration 9.8 m/s
The density rhow is calculated as a function of temperature, salinity
and pressure at fixed depths above the ULS, using the equation of
state of seawater given in Appendix A3.1 of Gill, A.E., 1982, Atmosphere-Ocean
Dynamics (Academic Press). The temperature and salinity are interpolated
to the mooring position and observation time from the latest version of the
monthly Polar Hydrographic Climatology (PHC) of Ermold and Steele, available
at the website: http://psc.apl.washington.edu/Climatology.html. The
densities are then averaged vertically to produce the average density rhow.
The ULS measures the range to the bottom of the sea-ice (or to the
water-air interface) by measuring the time tau for a pulse of sound
to propagate up to the target and back to the ULS. The range is
calculated from the equation:
range = 0.5 * cbar * tau - 0.156 (meters) where
cbar = Average sound speed of the water column above the ULS
0.156 = 0.5 * (cbar - clens) * tau_lens
clens = Sound speed of the fluid in the lens of the ULS
tau_lens = Round-trip travel time for sound to propagate
between the ULS tranducer and the surface of the
acoustic lens.
The sound speed cbar is calculated as a function of temperature, salinity
and pressure at fixed depths above the ULS, using the sound speed
formula presented by Urick, 1983, Principles of Underwater Sound
(McGraw Hill, New York). The temperature
and salinity inputs are the same values from PHC that are used to compute rhow.
Sea-ice Draft is calculated from the formula:
draft = depth - range - 0.42 m where
0.42 m = Vertical distance from the pressure gauge port to
the ULS tranducer, when the ULS is oriented vertically.
Potentially important sources of error in the ice draft
computed as described above include the following:
* Systematic and random errors in density rhow
* Systematic and random errors in sound speed cbar
* Random errors in sea-level air pressure Ps
* Errors in the local value of earth's gravitational acceleration g
* Timing variations in the sonar hardware/software system
* Target identification errors for first-return echos from
targets not directly above the ULS, or from targets intervening
between the ULS and the intended target (ice bottom or sea surface).
* Non-hydrostatic (short) waves during open water intervals
* Non-vertical orientation of the sonar transducer caused by mooring motion.
Analysis [as described e.g. in Drucker, R., S. Martin and R. Moritz.
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), 2003,
18-1 - 18-18.] yields, for a ULS deployed at approximately
40 meters depth, a net systematic error of approximately
-0.2 meters, a standard deviation of random errors of approximately
0.05 meters during episodes of open water among ice floes,
and an annual variation of systematic error on the order of
0.10 meters for ULS 17 at NPEO 2001-2002. Both random and
systematic errors are expected to increase with the depth of
the ULS.
An offset correction was applied to both the 5-minute and
10-second time series, by identifying occurrence of open water
based on a combined analysis of the ULS and ADCP data.
The offset correction eliminates most of the net systematic
error, and reduces the random errors in summer.
The offset has been applied to the datasets provided here.
Once the offsets are applied, summer (June-August) ice drafts
between -0.20 m and +0.10 m are reset to 0 m, consistent
with the range of error variability for open water episodes and
with the model that ice thinner than 10 cm is rare during
these months. For all other months, ice drafts between
-.20m and 0 m are reset to zero. Ice drafts outside the
range -.20m < draft < 21 m are replaced by the bad
data flag -999. Intervals of 100% open water determined
independently from Canadian sea-ice charts exhibit variations
of apparent "ice draft" (these intervals
usually occur in recognizable envelope
patterns related to wind speed) on the order of +/- 1 meter
due to surface water waves. Drafts in these intervals,
at the beginning and end of the time series, are set equal to zero.
When no echo is detected by the sonar, the draft is assigned
the value -999. This results in datasets with all
ice drafts in the interval 0 m to 21 m, and all data deemed
"bad" or "missing" assigned the value -999. The 5-minute
observation at the end of each hour is also assigned the
value -999, because the ULS processor program writes
overhead information instead of the sonar echo at that time.
The ULS is programmed to look for ice drafts in a window
that opens approximately 20 meters below sea level, and
closes approximately 2 meters above sea level.
7) Instrumentation used
APL/UW Mark 2 Upward Looking Sonar, consisting of:
* Paroscientific Digiquartz pressure gauge (serial number 45454)
* APL/UW sonar transceiver: frequency 300 kHz, pulse length 1 millisecond.
* APL/UW Acoustic lens: -3dB beamwidth 2 degrees.
* As an integrated assembly, this ULS has serial number 2303.
8) Quality control procedures:
Occurrences of open water in the ULS time series were selected for
consistency with the ADCP ice velocity data. The ADCP data show
very distinct differences when retieving velocity from an ice
surface versus an open water surface.
9) Data format
Space delimited ASCII, formats from header (example below):
YEAR MO DAY HR MIN SEC TIME(days) DEPTH(m) DRAFT(m)
---- -- --- -- --- --- ---------- -------- --------
2005 12 29 14 20 0 363.59722 90.422 1.278
NOTES
-----
A. Columns 1 through 6 are in GMT, as measured by the
ULS clock. A ULS clock drift of +/- 10 minutes per
year is typical. No correction has been made in this dataset
for ULS clock drift.
B. TIME is given in decimal days. TIME = 1.0 days
at 00:00:00 GMT, 1 January, 2006. TIME = 366.0 days
at 24:00:00 GMT, 31 December, 2006.
C. DEPTH is the depth below local sea level of the
pressure gauge port on the ULS. When the ULS is
oriented vertically, this pressure port is
42 centimeters below the sonar transducer on the ULS.
D. DRAFT is the estimated vertical distance between the detected
sonar target (bottom of sea-ice) and local sea level directly above
the ULS.
E. The value -999.000 for DRAFT indicates missing or
bad data.
10) Data collection problems:
On downloading the data from ULS 2303 in 2007, it was found that
no echo was detected by the sonar in approximately 25% of the observations.
The frequency of occurrence of missing echos displays a definite
annual cycle as follows:
Percentage of undetected sonar echos
------------------------------------
Nov 10%
Dec 50%
Jan 35%
Feb 25%
Mar 25%
Apr 25%
May 17%
Jun 10%
Jul 8%
Aug 2%
Sep 10%
Oct 10%
Although the APL ULS-Mark2 is designed to detect all surface
targets (sea-ice, open water) at a range of 100 meters, there
is some variation from instrument to instrument, and it appears
that ULS2303 at the nominal depth of 83 meters was operating
near the maximum limit of its detection range during 2006-2007.
The sonar target strength at 300 kHz frequency varies for
different types of ice and open water, depending in a complicated
manner on the distribution and refractive indices of scatterers.
Target strength is relatively larger for open water, compared
to ice, and relatively smaller for frazil ice than for congelation
ice. There are likely to be differences between the target
strengths of growing and melting congelation ice, but these
differences are not well documented as a function of frequency.
All other factors being equal, horizontal surfaces of
congelation ice that are normal to the sonar beam have larger
target strength than steeply sloping ice surfaces. This is consistent
with analysis of the 25-minute segments of 10-second data,
which show a higher frequency of occurrence of missing data
in the midst of larger undulations in under-ice topography
among the remaining ("good") data points,
and fewer missing data in the midst of flatter portions of
the profiles. The larger undulations tend to occur preferentially
when the mean ice draft is larger. For these reasons, it is
likely that larger ice drafts are more likely to produce
undetected sonar echos than smaller ice drafts. This would
introduce a bias in statistics estimated from the data.
For example, by linearly interpolating the ice drafts to the
times of missing/bad data in the 25-minute
segments during days 400-500 (4 February, 2007 to 15 May, 2007),
we estimate mean drafts as follows:
Mean draft for all "good" data 1.22m
Mean draft for all interpolated data 1.28m
Mean draft of data interpolated to the times of
"bad" (-999) points 1.46m
When estimating statistics from the datasets, users are
cautioned to analyze biases that may be caused by
draft-dependence of the probability of occurrence of
the missing/bad data flag.
The author of these datasets is conducting ongoing investigation
of the bias, and will update this documentation when more
complete information is available.
11) Other related data sets:
Data from the Recording Current Meters and Sea-Bird Temperaure-Conductivity
Recorders on this same mooring are provided together with this archive.
Data from this mooring's ADCP and from Davis Strait Freshwater Flux
moorings covering subsequent years are being processed and will be
submitted when available. Information on the project is available at:
http://iop.apl.washington.edu
12) Conditions for use or citation:
In reports and publications using these data, please cite their source:
R. E. Moritz, Davis Strait Freshwater Flux Array,
NSF Grant OPP-02300381
13) Data qualifications or warnings:
Estimated accuracy: Depth .01% of the depth value.
Sea-ice draft, error standard deviation of 10 cm for
level and gently undulating ice. Ice drafts for
rough ice are biased high because the range is
determined by the first return, which comes from
the ice nearest the sonar, and within the beam
footprint. This ice is sometimes not directly
above the ULS as required for unbiased sampling.
This dataset has not been adjusted to account for
this rough ice bias. The bias associated with
first-return ranging is common to many ULS sea-ice
draft datasets acquired from moorings and
from submarines. Additional bias on the order of
10cm (see item (10) above).
14) Grant numbers:
NSF Grant OPP-02300381
15) Sample repository location:
AON CADIS http://aoncadis.ucar.edu/home.htm
-------------------------------------------------------------------------------