441
1 INTRODUCTION
Theneedtoevaluatethepossibilityofpassagealong
theNorthernSeaRoute(NSR)arosewhenvesselsof
relativelylow ice classappeared on theNSR. Rapid
changes in sea ice cover cause periodic blocking of
individualsections ofthe NSRby driftingice fields.
Thisappliesespeciallytothespaceconstraintsatthe
nodes of the route and especially to the narrow
passages. In these pla
ces there occur specific
phenomenon of ice cover (concentration processes,
pressure of ice, hummocking and drift of sea ice)
whichreducethesafetyofvessels,especiallythoseof
low ice‐class. For this reason, hydro‐meteorological
data,especiallyinformat
ionabouticeconditionsand
specific risksarising ineach section ofthe NSR,are
veryimportanttoshipping.Thereexistvarioustypes
oficedata requiredtoevaluatethepassageofvessels
throughtheiceandtheshipʹssafespeed.The above
datawereinit
iallydescribedandusedover20years
ago in theʺIce Passportʺ for Russian vessels. The
content of theʺIce Passportʺ is now defined by the
IMOandthesamedocumenttooktheofficialnameof
theʺIce Certificateʺ (IMO, 2011). The Ice Certificate
describesseaworthiness ofvessels in ice,ta
king into
account structure of the hull, dimensions,
displacement, propulsion system, characteristics of
thepropeller,theageandactualconditionofthehull.
The Ice Certificate contains a vessel’s speed in
conjunction with sea ice cover thickness,
concentrationoficefloe,icefloesize,icepressureand
hummocking. It facilitates the ma
king of well‐
founded and documented decisions (IMO 2011).
Calculations of the above mentioned vessel’s safe
speed on ice have been described by Ryvlin and
Chejsin(1980).
These sources of limited content or accessibility
wereusedforrouteplanninginiceformanyyears.In
recent yearsthere haveappeared a large number of
new data sources, mainly in the form of processed
sat
ellite images. However, the quality of the
Functionality of Sea Ice Data Sources on the NSR
T.Pastusiak
GdyniaMaritimeUniversity,Gdynia,Poland
ABSTRACT:ThefunctionalityofavailableofficialsourcesofseaicedatafortheNorthernSeaRoutetodateis
low.Inrecentyearsalargenumberofnewpubliclyavailablesourceshaveappeared.Theirfunctionalityfor
purposes of route planning has yet to be evaluated. This study presents results of qualit
ative and expert
analysesofvarioussources.Itisproposedtousenewindicatorstoenablecomparisonoffunctionalityofdata
sources.Newsourcesprovidethetechnicalprogressthatisinstrumentalinreducingtheamountofeffortand
influenceofthehumanfactorinthedecision‐makingsystem.Thestudyalsopresentssolutionstotheproblem
oflimit
edbandwidthavailableathighlatitudeswithIridiumsatellitesystem.Presentedsolutionscanbeused
onanyvesselbyanycompanyornavigatortoimplementordesignthedecisionsupportsystemrelatedtoroute
planninginiceinaccordancewiththerequirementsoftheISMCodeandconceptofe‐Navigation.
http://www.transnav.eu
the International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 10
Number 3
September 2016
DOI:10.12716/1001.10.03.09
442
information available for use in the procedure of
planning and monitoring a voyage from an e‐
Navigation point of view has not been assessed.
Assumed in this study is that the scheme of route
planningthroughicecoveredareasofthe NSRshall
take into account the safety of the
decision process,
facilitation of downloading more reliable and
complete data in the most useful format and
functionality,whichistominimiseinvolvementofthe
user (IMO, 2008; Patraiko, 2008; Jurdzinski and
Pastusiak,2009).Theuser(navigator)shouldbeable
to control a vessel traffic safety according to
applicable rules and
should not be engaged in
activitieswhicharenotdirectlyrelatedtothevessel’s
movement control. It is possible to minimise
workload during downloading data sources,
digitalisation, processing and evaluation of results.
The decision‐making process should affect not only
content of the data sources but also their accuracy,
resolution,availability
and informationon reliability
and overall quality. Determined in this study are
criteria for assessing the operational sources of
information related to navigation in sea ice covered
areasontheNSRandqualityassessmentprocedures
ofdesignateddatasources.Thesecriteriacontainthe
information content of data sources, accuracy,
resolution,
reliability,generalqualityofdatasources,
workloadandweightofthecontent.
2 RESEARCHMETHODS
In orderto assess theavailability and the quality of
information about current hydro‐meteorological
conditions, information from the services GMDSS
(Safety NETforecasting areas of METAREA XX and
XXI and also NAVTEX data) was systematically
collectedforaoneyearperiodoftime,followedbyan
evaluation of their content. There was found a high
levelofgeneralisationoficedataprovidedbyvarious
means from the official sources. This indicates that
they cannot be used as a sole basis for passage
planningonthe
NSR,especiallyinnarrowpassages.
Therefore, other available data sources on the Web
wereanalysed, whichcontainmorepreciselydefined
parametersoftheseaicecover recognisedintheIce
Certificate or classified as hazards that occur on the
NSR. In total, 80 groups of data sources were
analysed. Data sources
were evaluated using
qualitonomicanalysis(Holodnik‐Janczurak,2007)by
newly developed indicators. Values of adopted
quality scale are in the range from 0 (worst) to 1
(best).Theweightsassignedforeachdatasourcesets
are according to their ascending or descending
importance for their intended purpose, which is to
evaluatethepossibilityofpassingapartoftheNSR.
The cumulative level of quality reached by various
datasourcesisdefinedbytheformulafortheaverage
value of quality indicators (Holodnik‐Janczurak,
2007). Firstly is taken into account the safety and
effectiveness of the decision‐making process of the
route planning, helping to obtain information more
reliableandmorecomplete inamoreuseful format,
next functionality and automatic operation with
minimalinvolvementofuser(e‐Navigationcriterion).
Secondly it is assumed that the information and
commands for direct implementation on a vessel’s
bridgebynavigatorsʺshouldbegiven
inaclearand
simplemannerʺ(ISMcriterion).
The author created a mathematical model of
assessmentbasedonexpertresearch.Thisarosefrom
the experience of the author, interviews with
practitioners and experts as well as a qualitative
analysis of problem and conclusions. Appropriate
softwarewasusedfortypeoffile
ormethodofdata
storage during study. In first place was the use of
Windows software as the most widespread and the
software supported by it or freeware software
supportedbyWindowsora licensedsoftwareoflow
priceofpurchaseandthereforeeasilyaccessibletoa
navigatoronthe
shiportoship‐owner.Datasources
arerelatedtothetimeperiod2011‐2012.
3 CHARACTERISTICSOFSEAICEDATA
SOURCES
The evaluation of functionality of analysed sea ice
datasourcesforvoyageplanningpurposestakesinto
accountseveralcharacteristics.Themostimportantof
theseareoutlinedbelow.
3.1 Weight
ofthecontent
Weight of the content (Table 1) was determined on
the basis of its usefulness for voyage planning. The
highestweightwasgiventotheparametersdescribed
intheIceCertificateandparametersconsideredtobe
the most serious threat to navigation on the NSR.
Theydirectlyinclude
strengthofthevesselʹshulland
ability to overcome ice with a safe speed under
definediceconditions. Theweight ofthe icedrift in
Table1isequal0.5.Thisvaluewasadoptedforopen
sea conditions. In case of narrow straits the
importance of ice drift data increases
due to the
existenceoflocalphenomenaofclosinga leadandthe
icejeteffect.Increasedvaluesoftheweightcoefficient
upto1.0shouldbeconsideredintheseregions.
Table1.Weightcoefficientofcontent
_______________________________________________
WeightWeight
_______________________________________________
Iceunderpressure 1.0 Openingsinice1.0
Hummockedice 1.0 Formationofsnow‐and‐ice
“cushion”atvessel’shull 1.0
Iceforms1.0 Stageoficemelting1.0
Thicknessofice 1.0 Icedrift0.5
Thicknessofsnow 1.0 Fog0.25
Concentration1.0 Icingofvessel
0.25
Floesizes
1.0
_______________________________________________
3.2 Resolution‐characteristicsofpositionprecisionon
maps
Datasourceshavevariousformsandmethodsofdata
recordinganddatapresentation.Mostsourcesdonot
containinformationonthescaleofthemap.Therefore
anewconceptfortheresolutionhasbeenintroduced‐
the minimum identifiable distance. Resolution was
adopted
asacriterionforcomparingtheprecisionof
position of various kinds of data sources. Three
443
methodsto determine theresolution of data sources
weredevelopedfor:
vector files (SIGRID‐3, KML, KMZ), raster image
graphics and their vector transformations (BMP,
JPGPNG,GIF,TIF,EPS,PDF) –theshortestlength
of a straight line, which approximates to isoline
curveofthesmallestobservedradius,
griddedfiles(GRIB, NetCDF,HDF)–thelengthof
thesidesofasinglegrid,
raster files (BMP, JPG, PNG, GIF, TIF) with an
averagedgridofdatathatdoesnotcoincidewith
themeridiansandparallels‐thelengthoftheside
of a single grid that is
specified by the
manufacturer.
3.3 Volumeofdatastoragesourcesanddatatransfer
indicators
Inordertocompareabilitytodownloaddatasources
were developed indicators (measures) of volume of
the internet file transfer
V
f
and the volume of
internet file transfer with the necessary elements of
the website
V
w
. File data transfer rate indicator
Q
f
(Formula 1) is the quantity of files able to be
downloaded during visibility of one satellite of the
Iridiumsatellitesystemabovethehorizon.
1
3600
f
f
B
Q
V
(1)
where:
Q
f
– file data transfer rate indicator,
V
f
–
volume of the internet file transfer [kB],
B
– data
transfer rate of Iridium satellite system [kB/s].
Bandwidth adopted for calculations is equal to 7.2
kb/s, equal to 0.95 kB/s,
τ
I
– the average time of
availabilityofoneIridiumsatelliteabovethehorizon
[hours]. According to information from the Iridium
company
τ
I
is a pproximately equal to 10 minutes.
Forthe purpose of calculations the value
τ
I
= 0.1667
hourshasbeenadopted.
Replacing
V
f
intheformula(1)to
V
w
allowsthe
obtainmentofformulaforrateindicator
Q
w
ofthefile
datatransferwithnecessaryelementsofthewebsite.
Data transfer rate indicators
Q
f
and
Q
w
characterise
ability to download a file. In practice it has been
observed that visibility of the next Iridium satellite
does not guarantee the continuity of a download.
Usually a download is aborted. Data transfer rate
indicators
Q
f
and
Q
w
take into account factors
attributable to satellite communication system
(download time and time of visibility of single
satellitebythereceiverofsatellitesystem).
Bothindicators
Q
f
and
Q
w
expresshowmanytimes
aparticularfileorafilewiththeaccompanyingweb
site during visibility of a single satellite by the
receiver for satellite systems can be downloaded. A
comparisonof indicators
Q
f
and
Q
w
candetermine
the cause of reduced data transfer and at the same
time the opportunities to improve it. Under real
conditions the limiting criterion of ability to
downloadafileisvalueofindicator
Q
w
equalto1.
Figure1.Schemeofdownloadingprocedure
Figure2.Schemeofdigitisingprocedure
3.4 Workload
Various data sources were examined and finally
universal procedures for their acquisition and
digitising have been developed. They are shown in
Figures1and 2with anindication ofa samplepath
for a HDF file and NetCDF file. The sequence of
actions in the procedure of downloading and
digitising data sources represents the sequence of
444
necessarysteps(Figure1)requiredtodownloadand
save information XYZ (latitude, longitude and
parameteroficenavigation)onacomputerinaform
acceptable for further mathematical calculations
(Figure2).
It has been observed that data sources come
together in three groups with similar downloading
and digitising patterns. These
are: (1) – vector files
KMZ,SIGRID‐3andShapefile,(2)–rasterandvector
mapswithgridcoordinatesBMP,JPG,TIF,GIF,PDF,
EPSandrastergeoreferencedmapsGEOTIFFand(3)
–griddedfilesGRIB‐1,GRIB‐2,NetCDFandHDF.
Workloadisthenumberof operationsperformed
in
diagrams(Figure 1and Figure2), which involves
anoperatordownloadingafilefromtheInternetand
digitisingthedata.Workloadindicator
Q
e
istheratio
of workload of downloading and digitalisation of
examined file
e
i
to the highest workload occurring
amongtheexaminedfiles
e
max
.Ithasbeendescribed
byformula2.
max
i
e
e
Q
e
(2)
3.5 Accuracy
Thefilesareavailabletotheuserforacertainperiod
of time after the moment from which they were
issued. In the considerations was adopted the
indicator
Q
a
(Formula 3) referring to the number of
hours of delay
t
a
and the number of hours of
operational planning cycle
t
o
, which was assumed
equal to 168 hours. Time shift is always negative
when consideringʺanalysisʺ. Time shift, as a rule,
must be positive when consideringʺforecastsʺ. This
indicator prefers forecast over analysis (the future
statemorethanthehistoricalone).
2
oa
a
o
tt
Q
t
(3)
Based on different scales of the operational
planning for the navigation in ice (Khvochtchinski
andBatskikh, 1998; Timco et al., 2005), the assumed
averagedurationoftheoperationalplanningcycleis
equalto7days(168hours).
3.6 Qualitycharacteristicsofadatasource
Theabilitytoexportor
writethedataina formatthat
isfitforreadingandprocessingsoftware.
Table2.Exportofdataindicator
Q
1
_______________________________________________
Indicator
Q
1
_______________________________________________
Unabletoexportdata0.00
Cannotexportnorsavefilesinsimpletextformats 0.25
Textfiles(TXT,ASCII)possibleforusebyspecialised0.50
software(suchasMATLAB)
Textfiles(TXT,PRN,CSV,XLSX)withacomplex 0.75
structureofdatastorage(storedinseparatetabsof
MSExcel)
Text
files(TXT,CSV,XLSX)withasimpleCSV 1.00
structuresuitableforeasyuseinMSExcel
_______________________________________________
The simplicity of transferring data and their
application in reading and processing software was
adopted as a quality criterion of the data source
(Table2). Simplicity and clarity received the highest
position in evaluation scales due to requirements of
theISMCodeandthee‐Navigationconceptcriteria.
3.6.1 Explanation
ofthescaleoftherelevantinformation
(legend)
It is assumed that the navigator on board the
vessel uses what he found on board. A equipment
deficiency would cause difficulties for the navigator
tocarryouthisorhertasksandimpairthequalityof
work done. Availability of scale
of presented
information
Q
2
(Table 3) helps determine a proper
value in relation to limits and increases precision of
the assessment. In the same way it helps to avoid
improperreading.Scaleclarityindicator
Q
3
allowsto
assignthepropervalueofthisparameter(Table4)to
establishedstandardscalesofjudgment.
Table3.Legendavailabilityindicator
Q
2
_______________________________________________
Indicator
Q
2
_______________________________________________
Legendofinformationisnotavailableinsourcefile 0.0
Legendofinformationisnotavailableinsourcefile.0.5
Itishoweveravailableinthedigitalpartofinformation
onawebpageorinapublicationonsourcefile
Legendofinformationisavailableinsourcefile 1.0
_______________________________________________
Table4.Legendclarityindicator
Q
3
_______________________________________________
Indicator
Q
3
_______________________________________________
Thescaleisnotclearlyunderstoodorthereis 0.000
noreference
Simplifiedscaleinrelationtopreciselydefined 0.333
establishedstandard
Scaleofparameterisdescribedincommonly 0.667
knownunits
Scaleofinformationisinaccordancewith 1.000
establishedstandard
_______________________________________________
3.6.2 Typeoffile
File type indicator
Q
4
(Table 5) reflects usability
and simplicity to direct use in the evaluating
possibilities of the NSR passage and also to data
processingwhenusingcomputation.Thescaleoffile
type indicator
Q
4
measures workload for the
application of data sources in the automatic
evaluationofpossibilitytonavigatetheNSR.
Table5.Filetypeindicator
Q
4
_______________________________________________
Indicator
Q
4
_______________________________________________
Rastergraphicfileorotherwithoutgridof 0.000
coordinatesorwithoutgeoreferencedpositions
Textfileofgeographicalpositionsorpositions 0.143
relatedtoareadescribedinanotherdocument
(plaintext)
Rastergraphicfilewithgridofcoordinates0.286
Vectorgraphicfilewithgridofcoordinates 0.429
Rastergeoreferencedgraphic0.572
Vectororgriddedgeoreferencedfile0.715
Textalphanumericfilewithrecordedgeographical0.858
positionsasunordered(complex)dataseries
Textalphanumerictiedpositionsinthisorin 1.000
anotherfileasanordered(simple)dataseries
_______________________________________________
445
3.6.3 Reliabilityofdatadescribed
Toassessthereliabilityofinformationonnautical
charts the concept of Zones of Confidence has been
introduced.Onthis basis, onecandetermine quality
of data on nautical charts for safe navigation. The
concept of confidence level was also developed for
thecoverageof
unsurveyedregionsbynauticalcharts
and publications such as those for research vessel
equipmentsupportfornavigationinpoorlysurveyed
regions (Pastusiak, 2011). Sources of hydro‐
meteorologicalandicedatashouldalsobeevaluated
bytheindividualreliabilityindicator.Asameasureof
reliabilityofaparticularparameter ofice
navigation
wasadoptedtheconceptofthepossibilitytomakea
routingdecisiontakingintoaccountthereliabilityof
theparametervaluesreceivedfromprovidersandnot
only the value of a single parameter. Information
obtained from satellite imagery transformations
should be provided with information on quality for
each message (file)
taking into account spatial
distribution.Inthiscase,itispossibletoincludethis
information into data processing and evaluation of
route calculated. Two indicators were adopted in
ordertoassessthereliability(quality)ofthedescribed
parameter: indicator on availability of reliability of
data described as
Q
5
(Table 6) and indicator on
availabilityofqualityscaleofthisparameter
Q
6
(Table
7). Assuming that information about the parameter
qualityallowsitsevaluation,avaluecanbeplacedon
ascaleofquality“theworst‐thebest.ʺ
Table6. Indicator on availability of reliability of data
described
Q
5
_______________________________________________
Indicator
Q
5
_______________________________________________
Noinformationaboutqualityofparameter0.0
Nodataonqualityparameterinthefilebutitis 0.5
availableinaseparatesourceofinformation
Informationaboutthequalityofparameteris 1.0
availableinfile
_______________________________________________
Table7. Indicator on availability of reliability scale of
parameter
Q
6
_______________________________________________
Indicator
Q
6
_______________________________________________
Nodatainfileonscaleofquality0.0
Noqualityscaleinfilecontainingdatabutscaleis 0.5
availableinaseparatesourceofinformation
Informationonqualityscaleisavailableinfile 1.0
containingdataonquality
_______________________________________________
3.6.4 Digitalisation
Theconceptofdigitalisationsetsouthowtomove
data from a source file (usually visualised using
software) to data processing software. For each
separate file indicators of accessibility were
considered (as described in Table 8 for the position
coordinates
Q
7
and Table 9 for the parameter of
navigationinice
Q
8
).Thesamemadeforcomplexity
of data reading (described in Table 10 for position
coordinates
Q
9
andTable11foraparameter
Q
10
).
Table8. Indicator on availability of geographical position
data
Q
7
_______________________________________________
Indicator
Q
7
_______________________________________________
Nodataonpositioncoordinates0.00
Manualreadingsofpositioncoordinatesonscale0.25
onthemapbyinterpolationproportionwith
calculatorandmanuallytypingfromkeyboard
tothecomputer
Visualreadingofpositioncoordinatesfromopen0.50
text(handwriting)orfromsoftwareonscreenand
manualentrytocomputerby
keyboard
Copyandpastepositioncoordinatestocomputer0.75
software(copy&paste)
Independentdatatransferbycomputersoftware 1.00
_______________________________________________
Table9.Indicatoronavailabilitytoreadparameter
Q
8
_______________________________________________
Indicator
Q
8
_______________________________________________
Cannotreadtheparameter0.000
Visualreadingofparameteraccordingto0.333
appropriatescaleandtypingfromkeyboardto
computer
Copyandpasteparameterdatatocomputer 0.667
software(copy&paste)
Independentdatatransferbycomputersoftware 1.000
_______________________________________________
Table10. Indicator of complexity of reading geographical
coordinates
Q
9
_______________________________________________
Indicator
Q
9
_______________________________________________
Cannotreadgeographicalcoordinatesofposition 0.000
Readingofgeographiccoordinateswithout 0.333
interpolation(discrete)
Readingofgeographiccoordinatesmanuallyby 0.667
interpolation
Readingofgeographicalcoordinatesbythe 1.000
software
_______________________________________________
Table11.Indicatorofcomplexityofreadingparameter
Q
10
_______________________________________________
Indicator
Q
10
_______________________________________________
Cannotreadparameter0.000
Readingofparameterwithoutinterpolation 0.333
(discrete)
Readingofparametermanuallybyinterpolation0.667
Readingofparameterbysoftware1.000
_______________________________________________
3.7 Totalqualityofadatasource
Based on evaluation of the public data sources
indicator
Q
T
relating to total quality of individual
sourcesoficenavigationontheNSRwasdeveloped.
It is an average of the 10 indicators describing
characteristics of the data source (Formula 4). Thus
was obtained the qualitative evaluation indicator,
which allows to compare a wide variety of data
sources.
10
1
10
i
i
T
Q
Q
(4)
446
3.8 Qualityofinformationaccessofindividualseaonthe
NSR
Indicatoronavailableinformationaccessqualityofa
selected sea
Q
R
(Formula 5) represents the highest
valuesofavailablesourcesthere.Ittakesintoaccount
the weight of each content of 1.0, 0.5 and 0.25
respectively. Attached indexes are addressed to the
followingmeaning:P–pressure,H–hummocking,T
–thicknessofice,S–thicknessofsnow,O–
openings
inice,C–“cushion”,M–melting,U–concentration,
F–formoficefloe,D–drift,V–visibility,G–icing,j
–indexofindividualregionlikeKaraSea,LaptevSea,
EastSiberianSeaandChukchiSea.
124
10
P
HTSOCMUF VG
D
j
R
QQ QQQQQ QQ QQ
Q
Q
(5)
3.9 QualityofinformationaccessonthewholeNSR
Qualityofdatasourcesfortheappointedrouteacross
theentireNSRisgivenbyformula6.Averagevalue
ofqualityindicatorsofindividualregions reflectthe
indicator of data sources for the entire NSR. All
sources of information are placed on the graph in
ascendingordervaluesofquality
Q
E
.Theresultisa
steadilyincreasinglinewithoutspikes.Itisassumed,
therefore, that this indicator well reflects examined
relations.
4
1
4
j
R
j
E
Q
Q
(6)
4 ANALYSISOFRELATIONSHIPSBETWEEN
INDICATORSANDWEIGHTS
Theabovementionedqualityindicatorsandweights
ofdifferentsourcesofdatawerestatisticallyanalysed.
Results of this analysis are shown in the graphs of
dependenciesandcorrelations.Toobtainthesegraphs
MS Excel and CurveExpert software were used. For
each
ofthegraphsbasedonCurveExpertsoftwareare
given the values of correlation coefficient (r) and
standarddeviation(S)ofdependentvariable(y).
Thevarietyofpropertiesofthesourcefilesevenin
thesamegroupresultedinlowcorrelationlinesthat
reflecttrends.Forexample,GRIBfilesincludedinthe
studyarerelatedtooneorbothhemispheres,whichis
twice the difference in volume of file transfer. Some
filescontainonlyoneparameteranalysisassignedto
onemomentoftimeandotherfilesareacompilation
of a few or several slides of different parameters or
different moments of
time. The problem is further
complicatedbythefactthatsomefilesareinGRIB‐1
and other files are in a compressed format; GRIB‐2.
Another difficulty is the diversity of data grid
resolution.
Thetypeoffile(filetypeindicator)showsthelevel
of the feasibility of implementation of
geographical
position coordinates and the parameters included in
file when determining the route of the vessel. The
graphinFigure3ashowsthatwithanincreaseofthe
file type indicator values, the overall quality of the
datasourcealsoincreases.Thisrelationshipishighly
statisticallysignificant (p <0.000001). For
almost any
typeof the ice parameter itis possible to select files
with a low or high workload of downloading and
digitalisation(Figure3b).
Figure3. The relationship of quality indicator (a) and
workload(b)withfiletypeindicator
Themajorityofthefilesarecharacterisedbyhigh
workload. These include mainly raster maps with
grid coordinates and raster georeferenced maps.
Duringtheanalysistherewasnotedsomedecreasein
the workload with increased weight of content
(Figure4).
Figure4.Therelationshipbetweenworkloadindicatorand
weightofinformationcontainedinthedatasources
Figure5.Filedatatransferindicator
Q
f
andindicatoroffile
data transfer with necessary elements of a website
Q
w
in
ascendingorder
The only significant difference in the resulting
workload is the need to manually enter the
geographicalposition coordinates and parameters to
acomputer(Figure5).Thisdifferencedidnotdepend
ontypeoffile.Itshould,however,benotedthatthe
447
benefitsofbypassingthemanualentryprocedurecan
be utilisedonly if there is software that can directly
implement files containing these data sets. The
internet bandwidth of the IRIDIUM satellite system
greatlyreducesthepossibilityofusingdatasourcesof
higherqualityindicatorvalues(Figure5).
Bandwidthdistributionof
files without necessary
elements of a website adopts linear character for
values below 7 on the graph. However, bandwidth
distribution for files with necessary elements of the
websitesignificantlyreducesbandwidthforalmostall
data sources. Volume of additional transfer of
websites seems to be a significant difficulty for
acquiring
files. Extremely high values were omitted
when considering relationships for
Q
f
and
Q
w
indicators. In this way was obtained many more
details on the graph than on the graph covering all
datasources.Itwasnotedthatdistributionofdatahas
two boundary lines (Figure 6a). Many of the more
complexfunctionsdisplayedacurvedlinesuchason
Figure 6b. Such an
example of a polynomial curve
(Figure6b)showsthatforvalueof
Q
f
abovevalue1
(forfilesthatcan bedownloaded during one period
of visibility of the IRIDIUM satellite above the
horizon) indicator
Q
w
assumes a constant value
independentoffurtherincreasesof
Q
f
.
Figure6.Therelationshipbetweenindicators
Q
f
and
Q
w
:a
‐ approximation by rectilinear function, b‐approximation
bypolynomialfunctionofthethirddegree
Figure7. The relationship between the data source quality
indicatorandtheweightofcontent
There are a large number of data sources of the
highestimportancetotheroutinginalmostthewhole
range, from the lowest attainable quality to the
highest (Figure 7). There are a lot of data sources
availabletotheuser.
Theweightofcontentspecifiestheimportanceof
data type
for the route calculation process and for
assessment of the ice and hydro‐meteorological
conditions.Withtheincreaseofweightofcontent,the
delay of delivered “analysis” data also increases. A
comparison of maps used commercially by Transas
Marine and their free equivalents available on the
internetleadstothe
deductionthatsomeinformation
is provided for commercial purposes on a regular
basis,withoutdelay.However,thesameinformation
becomes publicwhen it no longer has a commercial
value(forvoyageplanningonNSR).
Withincreaseofdatacontentandresolutioncomes
an increase also in file size. There are two
principal
tendenciesingroupsoffiles.Theseareverydetailed
filesi.e. ofgoodresolution(seemtobeusedprimarily
ashore,whereaverylargevolumetransferispossible
forhigh‐speedconnections)andfileswitharelatively
lowtransfervolume(theyseemtobepredestinedto
be received by
vessels with very low bandwidth
connection in the Arctic, when using the Iridium
satellitesystem(Figure8).
Figure8.Relationshipbetweenthevolumeofafiletransfer
V
f
andresolution
Thewidely understood qualityof data sources is
directly proportional to the volume of internet
transfer.TransferofdatainArcticregionsthatisfully
covered only by the IRIDIUM satellite system is
significantly reduced. This reduces quality of the
information available to a vessel under way on the
NSR. The
analysis of about 200 maps indicate that
basic access to information about ice conditions
(concentration, ice forms, floe size and limit of ice
edge)providesjustafewofthem.Table12presentsa
comparison between the ma in characteristics of
different sea ice data services from various sources
(abbreviations: IP‐
Ice under pressure, HI‐
Hummockedice,IF‐Iceforms,C–Concentration,FS
‐Floesizes,D‐Icedrift).Itshouldhelpnavigatorsto
identifythemost usefuldatasources andgive more
information on agencies providing ice maps and
services. The SIGRID‐3 format files published by
AARI contain information
about concentration, ice
formsandfloesizeanalysisdataforspecificregions
andall of Arctic. They have a very good resolution,
arequiteeasytodownloadandhavearelativelyhigh
total quality indicator. A very low workload value
couldbeacceptablebuttheaccuracy(delay)rangeof
0.571‐1.714
fortheoperationalplanningcycleduring
448
the navigational season makes them useless. The
same situation concerns AARI files for prognosis of
ice under pressure (compression), hummocked ice
andiceformsfortheregionoftheBarentsandKara
Seas. A resolution of 25,000 km could be acceptable
forgeneralrouteplanningbutadelayrangeof
1.5‐2.0
for the operational planning cycle during the
navigational season make them useless. In this
situation a much better offer is provided by a
screenshot from the website ocean8x.aari.nw.ru
representing ice floe concentration. The very low
resolutioncanbepartlycounterbalancedbyusingthe
isolineformatofthemaps.
Regionalmapswithicefloeconcentrationanalysis
fortheBarentsandKaraSeasinPDFformatavailable
on the website dnmi.no provide good support for
voyageplanning.Theyareavailablewithonlyavery
short delay and are easily downloadable. The
resolution and total quality indicator are on an
acceptablelevel.
Analmostsimilarqualityoficefloe
concentration analysis maps of the Chukchi Sea is
published by NIC on the website bsis‐ice.de. The
limitation is the availability of these maps on a
weeklybasismakingthemusefulonlyforashorttime
aftertheyarepublished.
An interesting offer
of separate files of
concentrationandiceformatforthewholeArcticisin
GRIB format available on the website osisaf.met.no.
Theresolutionislowerat10,000km.Thetotalquality
andworkloadindicatorsaregood.Thefilesareeasily
downloadablewithashortdelay. Filesrelatedtothe
quality of
the data are also available on the site.
Everythingtogethermakethesesourcesa quitegood
alternative for files covering only specific regions
(DNMI and NIC in PDF format) and a good
supplementforallthosementionedearlierincluding
concentration maps from the ocean8x.aari.nw.ru
website.
MarginalIceZonerepresentsa
separategroupof
fileswithalimitedconcentrationscale.TheseareNIC
filesinGEOTIFF formatconsisting of regional maps
ofthewholeArcticwithverygoodresolution,easily
downloadable, with a low workload and accuracy
(delay)andofacceptabletotalqualityindicator.They
were available on polarview.aq website (until the
project was closed). Currently they are available on
the natice.noaa.gov website. In the same group are
maps issued by NIC in shapefile format on the
polarview.aq (actually on natice.noaa.gov) website
andinKMZformatonthenatice.noaa.govwebsite.
Also worth mentioning are ice floe concentration
maps available on the iup.physik.uni‐
bremen.de
website in PNG, GEOTIFF and HDF formats. Their
delayisveryshort.Thebestindicatorsofferregional
mapsinPNGandGEOTIFFformat.Theusercanuse
alsoregional maps in HDF format.Thefiles contain
information on quality of data but the workload is
veryhigh.
Therearealso
varioussourcesofdatarelatedtoice
drift. The most suitable files for users are in raster
imageformat.Allofthemhaveaverylowresolution
of 31.25‐62.50 kilometres. The only important
attribute that differentiates them is the accuracy. Ice
drift analysis maps can be found on the
ifremer.fr
websitewithverylongdelay.Itmakesthemuseless
forvoyage planning. The maps of analysis available
on the osisaf.met.no website have the worst
resolution.Inthiscasethemostsuitableareicedrift
prediction maps available on ocean8x.aari.nw.ru.
Thereispossibilitytoselectamapinadvanceupto
sevendaysatintervalsofthreehours.
This study did not include the availability of
recent raster maps of ice cover analysisdeveloped
bytheRussianprojectPlaneta.Theyareavailableona
dailybasiswithoutsignificantdelay.Theirresolution
isestimatedtobeveryhigh,atleastthesame
asthe
abovementionedsourcesofdata. However,reliability
andqualityindicatorsofthesemapsarelowandthe
workloadisveryhigh.Alsowerenotincludedrecent
KMZfilesoficedriftforecastmapsdevelopedbyU.S.
NIC.
Table12.Maincharacteristicsofselectedseaicedatasources
__________________________________________________________________________________________________
SourceData Nameofsamplefile/datasource Weight Qf Qw ResolutionQe Qa QT
__________________________________________________________________________________________________
aari.ruC,IF,FS aari_arc_20100706_pl_a.ZIP (general) 1.00 0.43 0.34 1,008 0.015‐0.571 0.657
aari.ruC,IF,FS aari_kar_20110927_pl_a.ZIP(regional) 1.00 2.08 0.74 849 0.021‐1.714 0.657
ocean8x.aari.nw.ruC screenshotfromocean8x.aari.nw.ru 1.00 1.62 0.93 55,560 0.768+0,429 0.434
ocean8x.aari.nw.ruD screenshotfromocean8x.aari.nw.ru 0.50 4.16 0.84 55,560 0.768+0,429 0.421
aari.ruIP 20090918.BMP(Barentsand
KaraSeas) 1.00 4.26 1.12 25,000 0.764‐2.000 0.447
aari.ruHI 20090918.BMP(BarentsandKaraSeas) 1.00 4.07 1.14 25,000 0.764‐2.000 0.447
aari.ruIF 20090117.PNG(BarentsandKaraSeas) 1.00 1.47 1.47 25,000 0.761‐1.500 0.392
osisaf.met.noC ice_conc_nh_201105211200.grb.gz 1.00 0.30 0.28 10,000 0.764‐0.054 0.733
osisaf.met.noIF ice_type_nh_201109301200.grb.gz 1.00
1.00 0.82 10,000 0.764‐0.054 0.683
osisaf.met.noD ice_drift_nh_polstere‐625_multi‐oi_ 0.50 3.12 3.12 62,500 0.761‐0.054 0.517
201110051200‐201110071200_arc.PNG
dnmi.noC c_map1.PDF(BarentsandKaraSeas) 1.00 0.68 0.64 6,945 0.764 ‐0.009 0.487
NIC(bsis‐ice.de)C,IF,FS chukcurrentcolor.PDF1.00 1.99 1.49 3,902 0.764 ‐1.000 0.445
NIC(polarview.aq)C icechart_nic_current_arctic.ZIP
1.00 0.19 0.19 1,070 0.013 ‐0.051 0.546
(general)
NIC(polarview.aq)C arctico.tif.tar.gz(regional)1.00 0.91 0.91 2,000 0.761 ‐0.051 0.409
natice.noaa.govC arctic_2011286.kmz1.00 0.30 0.26 2,360 0.015 ‐0.048 0.546
iup.physik.uni‐bremen.de C asi‐n3125‐20111003_nic.png(regional) 1.00 6.27 2.88 3,125 0.772 ‐0.012 0.490
iup.physik.uni‐bremen.de C asi‐n3125‐20111003.tif
(regional) 1.00 1.73 1.38 3,125 0.772 ‐0.012 0.575
iup.physik.uni‐bremen.de C asi‐n3125‐20111003.hdf(regional) 1.00 0.45 0.42 3,125 0.024 ‐0.012 0.826
polarview.aqFS WSM_SS_20110925_153307_4417_3.dim1.00 0.01 0.01 96 0.761 ‐0.116 0.423
polarview.aqFS WSM_SS_20110930_123111_6248_1. 1.00 0.01 0.01 104 0.761 ‐0.116 0.295
dim.final.jpg
ifremer.frD 20110101‐20110201.PNG0.50 3.19 1.29
31,250 0.777 ‐1.143 0.462
__________________________________________________________________________________________________
449
For most regions of the Arctic (Laptev Sea, East‐
SiberianandChukchi)therewasnotedalackofmaps
for the most difficult parameters of ice navigation,
thesebeingpressureoficeandhummockingofice.At
themomentthisproblemcanbesolvedonlybyusing
the maps of
ice drift prediction available on the
ocean8x.aari.nw.ru website. The shortest regular
changesoficeunderpressure(compression)depend
on semidiurnal tides. The compression reach
maximum twice daily in between Low Water and
HighWater(Mironovet al2010 followingBuinitskii
1951 and Legen’kov 1988). The amplitude of the
compression changes
due to the effect of the tides
does not excide number 0.9 (Mironov et al 2010
following Kagan et al 2007). This is the possible
reason why the problem of compression due to tide
effectdidnotattractseriousattention(Mironovetal
2010followingProshutinsky1993).Basedoncollected
information
regarding the speed, direction and
duration of ice drift the navigator should use the
“Guide on dangerous hydro‐meteorological and ice
phenomenaalongtheNorthernSeaRoute”available
ontheaari.ruwebsitetoassessthelocationsandlevel
of compressed ice as well as for leads opening or
closing.However,
itrequiresmuchexperiencebyan
IceNavigator.
5 DISCUSSION
Itispossibletodevelopapassageplanforatransport
vessel of specified ice class and technical properties
thoughtheNSRwithinaspecifiedtimeperiod,under
specified ice conditions using studied hydrological
and meteorological data sources. The quality of
the
proposedroutedependsonqualityandcompleteness
ofthedatasourcesconcerninganalysesandforecasts
ofnavigationinice.Itisalsopossibletominimisethe
workload in the procedures of downloading,
digitalising, processing and evaluating data sources
using chosen data sources. Reducing the amount of
the human factor
introduces technical progress into
the decision making process and increases safety of
maritimetransport.
Therestillexistdeficienciesinusefuldatafor the
planningofroutesontheNSR.Theyarerelatedwith
thehydro‐meteorologicalandiceconditions.Onlythe
mainandwideststraitsontheNSRaresatisfactorily
provided
withdatasources.Officialdatasourcesare
ofmuchlowerqualityinrelationtonewformsofdata
storage.Therearealargenumberofdiversesources
ofinformationorfilesontheWorldWideWeb.New
formsofdatastoragerequiretheuseofsoftwarethat
isnotcommonly
known.
Itisimportanttopossessdataaboutparametersof
navigation in ice and its quality for correct decision
making.Afewsourcesofraster‐basedmapsprovide
verypoorinformationonthequalityoficeconditions,
or remark on their quality only. They are issued by
the Norwegian Meteorological Institute,
GMDSS
METAREAandOSISAFEUMETSAT.Russiansources
ignore issues of parameter quality. U.S. facsimile
maps provide unspecified information regarding
whole maps. Such files as GRIB (developed by
OSISAF), NetCDF (developed by OSISAF), HDF
(developedbytheUniversityofBremen)andcontain
individual information about quality of parameter.
This information is
contained inside the file or in a
separate file. The highest quality of data sources is
relatedtonewformsofdatastorage.
Data quality indicators for particular Russian
Arctic seas and the whole NSR achieve very low
values. Among 10 parameters of the highest
importance, four of them are practically
not
achievable in current operational practices (snow
thickness, “cushion”, melting and openings in the
ice). For this reason, a quality indicator of data
sourcesforthe wholeNSR cannotexceed avalueof
5.5.with10beingthemaximum(foricedriftweight
equal 0.5). Parameters of the ice conditions
that
present the greatest difficulties and require the
avoidanceof zones of theiroccurrence – pressure of
iceandhummocking‐are availableonly inasingle
file,andonlyfortheKaraSea.Inthecaseofomitting
data unavailable in the assessment types of ice
parameters, the quality indicator of
data sources for
the whole NSR, elaborated on basis of quality,
achievesonly63%ofthemaximumvalue.Inthecase
of taking into account bandwidth limitation, the
qualityindicator achieves an outcome twice worse‐
just33%.
When selecting sources of information according
toquality criteriamostly files offormat
HDF,GRIB,
SIGRID‐3andtext‐typerequiringspecialisedsoftware
wereaccepted.Whenselectingsourcesofinformation
according to bandwidth criteria mostly raster maps
with grid coordinates and with attributes of paper
maps and maps with significantly worse resolution
wereaccepted.TheKaraSeaismuchbetterequipped
in sources on
hydrological, meteorological and ice
data than other NSR seas. For this reason, there is
considerabledisparitybetweenpossiblequalityofthe
route planning on the NSR and assessment of
situations by shore side (for example, the
AdministrationoftheNSR)andthevessel.Thisisdue
tothelowbandwidth
oftheIridiumsatellitesystem,
which affects quality difference. It limits availability
of files with higher quality or resolution for vessels.
Thus, the main limitation of quality of the available
datasourcesontheinternetisthelowbandwidthof
theIridiumsatellitesystem.
It is possible to solve the problem
of limited
bandwidth of the Iridium satellite system. Internet
data transfer of a single connection through the
Iridiumsystemis2.4kB/s.Itispossibletoincreasethe
capacity by using specialised software enabling
transfer through several Iridium telephone
connections running simultaneously. An
improvement can also be achieved through action
from the website owner, by reducing the volume of
the website. A few data providers allow for this
opportunitybymakingdoubleversionsofwebsites‐
for connections with high bandwidth and low
bandwidth.Theseincludeʺpolarview.aqʺ(closed) and
ʺnoaa.natice/omb.ʺ The second possibility is the
involvement of the vessel’s company by
collecting
files ashore using fast internet connections, dividing
files into smaller portions or selecting small areas
needed,compressingthemanddividingthesefilesby
usingforexampleRARsoftwareandsendingthemto
the vessel using the low bandwidth of the Iridium
satellite system. Recently, Microsoft Download
Managersoftwarehas been
made availableto users,
450
whichpermitstheautomaticdownloadoflargerfiles
even following the uncontrolled interruption of
connections. This method does not involve other
parties in the procedure for downloading files and
providethevesselfullindependenceinthisarea.
6 CONCLUSIONS
This paper is concerned with the safety of the
maritime transport
in high‐latitude regions of the
RussianArcticseas,wheretheseafreezescompletely
during the winter season. There are very difficult,
rapidlychangingiceconditionsandtherearerisksto
shipping that are not found in other parts of the
world.AvesselnavigatingthroughtheNSRhasvery
limited access to information on the hydro‐
meteorologicalconditions,includingiceconditions.
Innovativeapproachesto theabove problems are
the use of new sources of information that are
different from commonly known image maps on
paperorinelectronicform,thedevelopmentofnew
indicators and methods to compare usability of
the
data sources for route planning purposes in ice‐
covered regions of the NSR, the introduction of
technicalprogressmeaningthereductioninworkload
and the improvement of maritime safety by finding
newways to reduce theimpact of the human factor
ondecision‐makingandoutputs.
The presented algorithm
for downloading and
digitisingisusedforthepurposeofrouteplanningin
ice on the NSR using publicly available sources of
dataoncurrentandforecastconditionsofnavigation
in ice. The presented algorithm quantifies the
workload. In this way, information is available to
compare the workload between various sources
of
information and quality of information used to
prescribe the route. A vessel’s captain may take a
decisionindependently,selectingaroutebychoosing
appropriate data sources following his or her
knowledge,experienceandexistingcircumstancesof
navigation. This meets the requirements of the
conceptofe‐Navigation.
Themost
importantachievementsofthisstudyare
the analysis of functionality of various data sources
oncurrentandforecastconditionsofthenavigationin
iceontheNSR.Thisallowsdevelopmentstosupport
maritimesafetythrough implementationoftechnical
progress. It reduces workload, facilitates the
automatisation of data processing and analysis for
decision support systems associated with route
planning in ice on the NSR, in accordance with
requirements of the ISM Code and the e‐Navigation
concept. Another result is the development of
mathematical tools to assess data sources related to
navigationiniceconditionsandtoplantheroutesof
ships
on the NSR. The next achievement is the
determinationofmethodsonhowtofixtheproblem
oflimitedbandwidthavailableathighlatitudesusing
theIridiumsatellitesystem.Methods,proceduresand
algorithmscanbeusedwidelyon eachship, by any
companyornavigatortodesignthedecisionsupport
system
for navigation in ice, with the occurrence of
diverse,incompleteoruncertaininformationrelating
tonavigationiniceandtheabilitytoovercomeiceby
vesselsofanyiceclass.
Forthesereasons,itisexpectedthatthepaperwill
serve navigators on conventional and non‐
conventionalvesselsasa
guidelineonavailablenew,
free of charge, modern data sources with a larger
range of information than official sources. At the
same time information services and agencies
providing ice maps in new formats are a source of
information providing new opportunities for their
use.Theyincludesimilardataasinthe
conceptofthe
Zones of Confidence that have been introduced on
electronic sea charts. This paper should also be
helpfulforanyonewhoapprecia testhepossibilitiesof
using independent and free sources to assess ice
conditionsor the possibilityto verify or supplement
informationprovidedbyofficialsources.
REFERENCES
BuinitskiiV.Kh.(1951).Formationanddriftoficecoverin
theArcticbasin,Proceedingsofdriftingexpeditionl/
pʺG.Sedovʺ(inRussian),Glasevmorput,Vol.4: 74‐179.
IMO (2011). Development of a mandatory code for ships
operating in polar waters, The ice certificate concept.
IMO,DE55/12/11,14
January2011:6p.
IMO (2008). Draft strategy for the development and
implementation of e‐navigation. IMO Resolution A.
989(25),NAV54/25Annex12:14p.
Hołodnik‐Janczura G. (2007). Examining the quality of
informaticsproductusingvaluationmethod,Operations
ResearchandDecisions,No2/2007:55‐69.
JurdzinskiM.,
PastusiakT.(2009).E‐navigationasaprocess
of modelling of safety of navigation, Economic (in
polish),socialand legalchallengesof maritime statein
the European Union, Srodkowopomorska Rada NOT,
Koszalin:85‐93.
Kagan B.A., Romanenkov D.A., Sofina E.V. (2007).
Simulationoftidalicedriftandice‐generatedchangesin
thetidaldynamicsontheSiberiancontinental shelf (in
Russian). Izvestiya of Academy of Sciences, Russia.
Atmospheric and Oceanic Physics, Vol. 43, Nо 6:831‐
850.
KhvochtchinskiN.I.,BatskikhY.I.(1998).TheNorthernSea
Routeasanelementofthe ICZMsystemin the Arctic:
problems and perspectives. Ocean &
Coastal
Management,41:161‐173.
Legen’kov A.R. (1988). Shifts and tidal deformation of
driftingice,Gidrometeoizdat:104.
Mironov E.U. (editor), Kljackin S.V., Gudkovič Z.M., May
R.I., Frolov S.V. (2010). Dangerous ice phenomena for
shipping in the Arctic (in Russian). AANII, St.
Petersburg:320p.
PastusiakT.(2011).Ship’sNavigationalSafetyin
theArctic
UnsurveyedRegions,MaritimeNavigationandSafetyat
SeaTransportation.Miscellaneousproblemsinmaritime
navigation,transportandshipping,CRCPressTaylor&
FrancisGroup,London:59‐64.
Patraiko D.J. (2008). e‐Navigation, Digital Ship, Posidonia,
July2008:26p.
Proshutinsky A. Yu. (1993). The Arctic Ocean level
oscillations (in
Russian).Gidrometeoizdat, St.
Petersburg:216p.
RyvlinA.Ya.,ChejsinD.E.(1980).Testsofvesselsinice(in
Russian).Sudostroenie,Leningrad:208p.
TimcoG.W.,GormanB.,FalkinghamJ.,O’ConnellB.(2005).
ScopingStudy:IceInformationRequirementsforMarine
Transportation of Natural Gas from the High Arctic.
Technical Report CHC‐TR‐029,
Canadian Hydraulics
Centre,Ottawa:124p.
