I n s t r u m e n t s

a n d

M e t h o d s

Figure 2. Cleaned daily snow depth data from Agassiz Ice Cap, Ellesmere Island,N.W.T.

manufacturing.
Althoughtheinitialprototypeworkedwell,therewere
some basic design changeswhich had to be incorporated
in order to both improve the quality of the measured data
and make the sensor operational in remote areas.Some of
these changes were basic,involving powersupply (from
AC to DC operation),packaging and removal ofa built it
temperature sensor.The temperaturemeasurementwas
required to adjust the velocity ofsound forambient tem-
perature.It was difficult tobuild a reasonable radiation
screenforthetemperaturesensoraspartoftheoverall
sensor.Furthermore,atmostautomaticstationswhere
snowdepth measurementsarerequired,an airtempera-
ture measurement is usually available.
Thefirstsensordesigned and manufactured byCSCC
hadthemodeldesignationCSMAL01.Onlyalimited
numberof these sensors were built.They had the disad-
vantage of having the builtin temperature sensorwith a
radiationshield,the powerconsumptionwashigh and
the sensor enclosure was inadequatefor high humidityen-
vironments.The sensor subsequentlyhadtwo major design
changes,resultinginmodel designationUDG01 (Ultrasonic
Depth Gauge) and most recently the SR50 (Sonic Ranger).
Inthispresentation,onlythesetwomodelswillbe
discussed.

130

UDG01SENSOR

This was the sensor resulting from the first major redesign
of the ultrasonic sensor.The built in temperature sensor
was removedas wellas its associatedradiationshield.This
allowed a streamlining of the sensorpackaging.Figure 1
is a graph showing the raw snow depth data fora sensor
located on the Agassiz Ice Cap, on the northern region of
Ellesmere Island.The data are daily readings fromMay
1988toApril1989.There are periodswhen thesensor
wasnotcapableofmakingaproperreadingandthe
resultingdata valuewaszero.Thiswouldoccurifthe
sensorwas making a reading during a snow fall event or
duringperiodsofblowingsnow.Underthosecircum-
stances, the sensor would receivemultiple echoes and not
beabletomakea singledetermination.Thereisalsoa
fourmonth period when the sensorwas totally encrusted
with hoarfrostand could not make a reading.This isa
common problem with the site and it equally affected the
performance ofthe anemometerand the temperature ra-
diation shield.Figure 2,is forthe same data set but with
all the incorrect readings removed.
Figure 3issnowdepth data forthesamesitebutfor
severalyears.The1993-1994seasonisincompletebe-
cause ofmissingdata due toasevered communications