|
TECHNICAL ARTICLES
|
Published in issue No 104, April 2002 of The Hydrographic Journal
|
|
IHO S44 Standards for
Hydrographic Surveys and the Variety of Requirements for Bathymetric Data
David E Wells
Ocean Mapping Group
Department of Geodesy and
Geomatics Engineering
University of New Brunswick,
Canada
David
Monahan
Canadian Hydrographic Service
Ottawa,
Ontario, Canada
Abstract
In addition to its age-old use of abetting safe and efficient navigation,
hydrographic data is increasingly used in an ever-expanding array of marine
activities. Many non-hydrographic office agencies and companies are collecting
hydrographic data, for intended purposes other than the production of nautical
charts. All Hydrographic Offices wish to continue their prime role as providers
of navigation products and many aspire to have their data applied to a broader
list of intended uses. One of the strong points of hydrographic data is that it
generally conforms to a known standard.
For the past 40 years, this has usually been the International
Hydrographic Organisation’s S44. However, S44 is strongly biased towards
surface navigation and is inapplicable to many other uses of hydrographic data.
This paper will explain the uncertainty elements in bathymetry data and compare
S44 standards with alternatives that have recently emerged.
Introduction
The word 'standard' implies something that can be used as a basis for
comparison, such as a model or a set of rules, or an authorised measure of some kind. Along these lines, the
International Organisation for Standards (ISO) defines the term 'standards' as:
Rules, guidelines and definitions of characteristics, which
ensure that materials, products, processes and services are fit for their
intended purposes.
This paper will dissect the statement in three steps, as it applies to
bathymetric information. First it will
consider by what characteristics it is appropriate to judge whether bathymetric
data is 'fit'. Secondly it will
consider the variety of 'intended purposes' for which marine bathymetric data
may be fit. Finally, it will consider
the status of standards for bathymetric data, in particular those contained in
the 4th Edition of the International Hydrographic Organisation Special Publication
No. 44 (IHO, 1998), and some recent alternatives.
Intended uses for bathymetric data
The traditional mandate of hydrography has been to survey, chart and
supply all spatial information required to assist in safe navigation, and
safety of life at sea, primarily for those commercial shipping vessels which
fall under the conditions of the Safety of Life at Sea (SOLAS) convention
administered by the International Maritime Organisation (IMO).
However, driven by technological change, hydrographic needs and
capabilities are becoming more broadly concerned with the management of spatial
information concerning all marine features, processes and properties in four
dimensions (space and time) including the acquisition, analysis and
visualisation of this spatial information
(Kenny, 2000; Hecht, 2001;Monahan
et al, 2001). Bathymetry is that
aspect of hydrography that is concerned with delineating the marine floor,
including features of both natural origin and those due to human activity.
Bathymetric mapping has four broadly defined intended uses: to improve
knowledge and understanding; to establish sovereignty and security; economic
purposes (including offshore resource management and shipping) and
environmental management.
Hydrographic information, in particular bathymetric information, is used
to make informed decisions of several types: for example vessel navigation
decisions; resource management decisions; resource development decisions;
marine infrastructure decisions; marine construction decisions; coastal
development decisions; tactical and strategic military decisions and
environmental management decisions. The confidence with which such decisions
can be made depends on the confidence that can be placed on the hydrographic
(and other) information available to assist in making informed decisions. It is
consequently critical that users be informed of the uncertainty associated with
the data and with products constructed from it. For suppliers of bathymetry to
provide information about uncertainty, they must first assess it. They are aided in this assessment by
mathematical tools and an international standard, S44 of the International
Hydrographic Organisation (IHO, 1998).
Assessment of uncertainty in bathymetric data
The uncertainty associated with bathymetric measurements includes (a)
uncertainty in the location of a measured bathymetric data point; (b)
uncertainty in the depth associated with a bathymetric data point and (c)
uncertainty in the backscatter strength associated with a bathymetric
measurement.
Bathymetric uncertainty management involves both the design of a
bathymetric system and the evaluation of results and products derived from
bathymetric data. Measurements are always uncertain, to a greater or lesser
degree. Uncertainties are of three fundamentally different types: accidental, systematic and random. Each type must be dealt with
differently. A common characteristic
shared by all three, however, is that the reliability with which we can
determine uncertainty is completely dependant upon the degree to which the
bathymetric data is redundant (repeated measurements of the same seabed
feature, or even footprint, which can be directly compared to ascertain
consistency).
'Data cleaning' describes methods used to deal with 'accidental'
uncertainties, (also called mistakes, blunders, or outliers). Comparison of a suspected outlier with its
geographical nearest neighbouring data points (taking hydrographic judgment
into account) is the most powerful data-cleaning tool. A rule of thumb which has emerged for
cleaning high-density bathymetric data is that real features are distinguished
from points created accidentally according to whether multiple consistent data
points (multiple 'hits') in close proximity are observed or not.
'Artifact' describes the effect of a systematic uncertainty. 'Artifact detection' and, where possible,
'artifact removal' describe further steps in the data-cleaning process. Artifacts are most often manifested as
identifiable artificial features in a data series, with a strong correlation in
time or space with some other data series.
Effective artifact detection requires dense data, and powerful
visualisation tools.
Whatever remains after (perhaps incomplete) data cleaning and artifact
removal, are considered as random uncertainties, or noise, in the data. Sometimes it is appropriate and possible to
reduce the noise level by use of suitable filtering and smoothing of the data,
but this can be dangerous, re-introducing systematic uncertainties, due to the
filtering process itself.
In any case, in the best case some remaining 'random' uncertainties will
be left. Otherwise there will still be residual systematic uncertainties that
cannot be removed. In extremely unfortunate cases, there may still be blunders
or outliers which cannot be removed with certainty, because it is impossible to
decide whether these data points represent real features, or are accidents of
measurement.
To meet the requirements for informed decision-making, it must be
possible to describe these remaining uncertainties in some standard way. One
uncertainty descriptor is 'precision' which describes data consistency. Good precision indicates that outliers have
been successfully removed, and random uncertainties are small - but large
systematic effects may still exist.
Another uncertainty descriptor is 'accuracy', which in a perfect world
indicates the agreement of data with the 'truth' (whatever that may be). Good accuracy indicates that the systematic
effects have been reduced or eliminated, although occasional outliers may still
exist, and the random uncertainties may be large or small.
Both these uncertainty descriptors are based on statistical principles
and standards. The 'mean' and the
'standard deviation' are the two most common statistical descriptors of
measurement uncertainties. The mean describes the central tendency of a series
of measurements. The standard deviation describes the dispersion of a series of
measurements. If the mean value (or
perhaps a 'true value' if such is known) is subtracted from every measurement,
a series of 'residuals' or deviations from the mean will result. If the square
root of the sum of the squares of these residuals is calculated, the standard
deviation for that measurement series is obtained.
When discussing measurements that have a number of 'dimensions' or time-correlated
quantities (as is most certainly the case for a modern bathymetric survey),
then these concepts are extended into several dimensions by considering a 'mean
vector' and a 'covariance matrix'.
Data-sets containing many measurements tend to have a special statistical
character, known as a Gaussian distribution (the familiar 'bell-shaped curve'),
provided all accidental and systematic uncertainties have been removed, so that
the uncertainties are purely random. This Gaussian character is an approximate
model of reality, and becomes a better model the larger the number of values
which are being considered (something called the Central Limit Theorem), and
the more rigorous or successful the data cleaning process. An important descriptor of uncertainty, when
the data density permits, is the probability that the data residuals (the
random component of uncertainty) obey the Gaussian distribution.
But what does all this have to do with the confidence which can be placed
in the information or measurements? Another statistical principle that can be
predicted, under specific statistical conditions, is how often the measurement
uncertainties (or more specifically the measurement residuals) are likely to
exceed a certain value. The value (or values) in question is referred to as the
'confidence region', and the likelihood that the measurements lie inside this
confidence region is referred to as the 'confidence level'.
The international standard for confidence level is 95% – in other words
19 times out of 20. 95% is the
confidence level associated with weather predictions. 95% is the confidence level associated with election outcome
predictions or public polling results.
And 95% has become the standard for expressing the confidence level for
results derived from hydrographic measurements. If data has a Gaussian distribution, the 95% confidence region is
related to the standard deviation (in one dimension) or the covariance matrix
(in several dimensions) by a simple scale factor.
In summary, key quality factors in bathymetric survey design are
'coverage', 'resolution' and 'redundancy'. The key quality factor in
bathymetric data assessment is 'uncertainty' - what are the uncertainties in
the resulting bathymetric, positioning and sonar backscatter information, and
how do these uncertainties compare with the informed decision-making
requirements for the intended uses?
Bathymetric uncertainty management requires redundancy and consists of
two or three steps - data cleaning for both outliers and artifact removal, perhaps
followed by a noise reduction process, and finally an assessment of the 95%
confidence region associated with the remaining residual discrepancies.
Having applied the tools discussed in the previous section, it is
possible to arrive at numerical values for uncertainty of the bathymetry data,
either grouped by adjacent areas, or individually. One way to assess these
numbers (decide if they are fit for their intended purpose) is to compare them
against a standard.
A standard can be used as a planning document before data are collected
and as an evaluation document after the data are in. The a priori approach
tries to assess the uncertainty with which each piece of data could or should
be collected, before a survey is conducted.
This is implemented through an uncertainty prediction estimation process
or model. These predicted uncertainties
are compared with those required to meet the appropriate standard, and the
survey redesigned if they fall short.
The a posteriori approach
attempts to determine what uncertainties actually exist in the collected data,
using the data cleaning and assessment tools referred to earlier in this
paper. The results of these post-survey
checks are then compared with the appropriate standard, to determine whether
the survey results are actually 'fit for their intended purpose'. Sometimes it is claimed that a survey 'met'
the standard, but if no post survey check was carried out to verify this claim,
it is not supported by evidence.
Claiming that surveys were planned to meet the standard is not
enough. Planning and realisation are
not always the same thing.
In the following sections some of the standards that are available for
this assessment are considered. For
simplicity, just one of the many standard parameters required for assessing
hydrographic data will be addressed:
the uncertainty in determination of depth.
S44 - IHO standards for hydrographic surveys
The International Hydrographic Organisation (IHO) has issued standards
for hydrographic surveys (S44) since 1957, and most recently in 1998 (IHO,
1998). These are the standards used by
most producers of hydrographic data. Their stated purpose is:
To specify minimum standards for hydrographic surveys in order
that hydrographic data collected according to these standards is sufficiently
accurate and that the spatial uncertainty of data is adequately quantified to
be safely used by mariners (commercial, military or recreational) as primary
users of this information.
S44 identifies itself as a 'performance standard' and thus contains no
instructions on how to evaluate whether a survey meets the standard. Nor does it specifically require the
inclusion of redundancy, the most powerful tool for evaluating the uncertainty
of any set of measurements. These are
left to each agency to implement:
Equipment and procedures used to achieve the standards laid
down in this publication are left to the discretion of the agency responsible
for the survey quality.
Producing a standard like S44 is no easy task. Usual practice is for several
member states of the IHO to nominate specialists who not only have a profound
knowledge of the theory underlying the subject but are also aware of upcoming
improvements in the technology that may impact on the standard during its
lifetime. The group must also have a strong sense of the pragmatic: there is no
value in producing a standard that cannot be achieved or can only be achieved
at costs not sustainable by some member states. Finally, the members must
possess a thick skin, since their work can never please everyone.
The work of producing the standard is ongoing, in a periodic manner, with
the published intention of issuing a new edition every five years. An examination of the changes between
succeeding editions gives a strong indication the perceived progress in
hydrographic technology and evolution in users’ needs. For instance, the
current (4th) edition:
…departs from previous editions by specifying different
accuracy requirements for different areas according to their importance for the
safety of navigation. The most stringent requirements entail higher accuracies
than previously specified, but for areas of less critical nature for navigation
the requirements have been relaxed.
Improvements in positioning technology that allow vessels to determine
their locations at a level of uncertainty smaller than that required by the
previous standard, together with the development of high density bathymetric
mapping tools (such as multibeam sonar echo sounders and LIDAR), are reasons
behind this demand for higher accuracies in certain areas. Future editions will
likewise adapt the standard to evolving technology and users requirements.
S44 4th Edition classifies surveys into four different types (four
'intended uses'):
Special Order - for
specific critical areas with minimum under keel clearance and where bottom
characteristics are potentially hazardous to vessels (generally less than 40m),
such as harbours, berthing areas, and associated critical channels with minimum
under keel clearances.
Order 1 – for harbours, harbour approach channels, recommended
tracks, inland navigation channels, and coastal areas of high commercial
traffic density (less than 100m), such as harbours, harbour approach channels,
recommended tracks and some coastal areas with depths up to 100 m.
Order 2 – for areas with depths less than 200m not covered by
Special Order and Order 1.
Order 3 – for areas not covered by Special Order, and Orders 1
and 2 and in water depths in excess of 200m
For each of these it specifies Horizontal Accuracy, Depth Accuracy, 100%
Bottom Search, System Detection Capability and Maximum Line Spacing.
S44 4th Edition divides depth uncertainties into two contributing types:
fixed and variable. It makes no mention
of the primary classification of random, systematic and accidental, within
these fixed and variable types. Fixed
errors dominate the uncertainty budget in shallow water. Variable (depth-dependent) errors are
characterised as a fixed percentage of
water depth and thus grow larger with deepening water. The two types are combined in the
Root-Sum-of-Squares (RSS) sense to give the 95% uncertainty s. That
s = [a2+(bd)2]1/2
where
a = sum of all depth-independent errors, b = sum of all depth-dependent errors,
expressed as a fraction of water depth, and d = depth of water column in
metres.
S44 4th Edition draws a distinction between the sampling of the seabed
bathymetry represented by the measured depths, and the complete bathymetric
model which is presented (in some form) to the end user for informed decision
making. Unless the sampling density is
dense enough to delineate all seabed features, this model will be based, either
implicitly or explicitly, on some form of interpolation between the sampled
depths. Consequently the uncertainty
associated with a bathymetric model will include uncertainties introduced by
the interpolation process, and will be larger than the depth measurement
(sampling) uncertainty.
Table 1: Summary of Minimum
Standards for Depth Uncertainties
from S44 4th Edition (IHO, 1998)
In the case of the Special Order, this algorithm is somewhat more
demanding than the single depth uncertainty specification from S44 3rd Edition
(IHO, 1987), which was
30cm to the depth of 30m, and 1% of depth thereafter.
The S44 3rd Edition specification was at the 90% confidence level, and
did NOT include uncertainties in water level reduction, which are included in
the 4th edition specifications.
For Orders 1 to 3, this algorithm results in higher permitted
uncertainties than did the single 3rd Edition specification.
There are two ways in which the S44 4th Edition depth uncertainty
standards can be interpreted. In the
first interpretation, the word 'minimum' standards is taken as the operative
word, and the unlimited extension of each of the four S44 orders to deeper
depths is permitted, even though not mandatory. In the second, more limited, interpretation, each Order is
assigned a maximum depth to which it should be applied (Special Order to 40m,
Order 1 to 100m, Order 2 to 200m, and Order 3 in deeper water).
Fig. 1: Log-log plot of S44
3rd and 4th Editions.
In subsequent Figures, the
S44 Special Order plot is used as a reference.
Beyond S44 - other intended uses, other standards
S44 4th Edition broke a lot of new ground. It addresses the use of high density bathymetric methods, such as
multibeam, sweep and LIDAR. It
emphasises the need to determine and record ('attribute') depth and position
uncertainties. It distinguishes between
depth measurement uncertainty and bathymetric model uncertainty.
Previous S44 editions were based on the scale of a specified chart, and
the draughting skill of experienced marine cartographers. S44 4th Edition is based on uncertainty
budgets and (at least nominally) on intended uses. However, despite this nominal objective, the intended use for
which S44 4th Edition was created is still almost exclusively nautical
charting.
Some of those seeking depth uncertainty standards for other intended uses
of bathymetric information have referred to S44 4th Edition, as is (e.g. United
Nations, 1999). Others have extended,
modified and replaced the standards embodied in S44 4th Edition.
This paper will consider four examples of standards that go beyond S44
4th Edition:
• The Exclusive Order introduced
by the Swedish Maritime Administration (SMA).
• The US Army Corps of Engineers
shallow water standards.
• Standards proposed by Land
Information New Zealand for deep water multibeam echosounder surveys.
• Standards proposed by the
International Marine Contractors Association for offshore construction.
Swedish implementation of S44
IHO S44 are minimum standards. At least one hydrographic office, the
Swedish Maritime Administration, has defined standards which are based on S44
4th Edition, but which are more demanding than those minimum standards (SMA,
2000).
On 1 May 2000, these new standards came into effect for Swedish surveys,
and are being considered for adoption by other Baltic hydrographic offices.
• SMA extended S44 4th Edition in
four ways:
• A new Exclusive Order specification
was added, intended for the most demanding applications.
• 100% seafloor coverage is
required in all cases by SMA, whereas for S44 4th Edition 100% coverage is
specified only for Special Order and, if there is a grounding hazard, for other
Orders as well.
• Depth uncertainty in the
standards refer to both acoustic sounding measurements (topographical
reproduction) as well as determinations of the minimum depth by means of
mechanical sensors (sweeping bars).
• The SMA depth uncertainty
standards include the entire error budget from the surveying uncertainties up
to the final result - storage in the digital depth database. In this way, the SMA depth uncertainty
standards are conceptually closer to the S44 4th Edition bathymetric model
uncertainties, than to the depth measurement uncertainties. However the SMA standards are much tighter
than the S44 4th Edition standards, since the numerical values are derived from
the S44 4th Edition depth measurement uncertainties.
The SMA established two 'intended uses' - 'fairway areas' and
'other'. Fairway areas are defined as:
existing, proposed or planned fairways, traffic separations,
deepwater routes, ports and areas of anchorage or waiting.
Fairway area surveys require an initial acoustic sounding survey. This is followed by a mechanical bar sweep,
when the acoustic soundings indicate that the fairway depths are either:
• less than 150% of the minimum
existing, proposed or planned underkeel clearance safety margin, including
squat, or
• the underkeel clearance safety
margin is less than 1m.
Table 2. The Swedish implementation of S-44
USACE Hydrographic Manual 2001
The United States Army Corps of Engineers has published a Hydrographic
Manual, containing background information, field procedures, and survey
standards for Corps hydrographic projects since 1991. This document defines two categories of hydrographic surveys
(intended uses):
• Navigation and dredging support
surveys, including project condition surveys of navigation channels, dredging
contract plans and specifications surveys, dredging measurement, payment,
clearance, and acceptance surveys, and river charting surveys.
Fig. 2:
Log-log plot of SMA implementation of S-44
• General surveys and studies,
including general reconnaissance or planning surveys/studies, flood control
project surveys, reservoir sedimentation surveys, flood plain boundary surveys,
hydrological and hydraulic surveys, coastal engineering surveys, beach surveys,
environmental investigations, geotechnical investigations, and disposal area
surveys.
Based on the following principle:
• survey instrumentation
requirements, accuracy standards, and quality control procedures vary as a
function of bottom type in a navigation channel; as does the required accuracy
of dredge measurement and payment.
USACE navigation and dredging support surveys are further divided into
three categories:
• Hard bottom material and/or new work. Navigation projects where low
under-keel clearances are anticipated over potentially hazardous bottom
conditions, hazardous cargo is transported, or where bottom sediment could
adversely impact naval vessels transiting a project only a small number of
Corps projects fall under this category.
• Soft bottom material and/or
maintenance dredging. Navigation projects containing soft sand/silt bottoms not
judged to be hazardous to vessel hulls; or projects with soft, featureless, and
relatively continuous channel bottoms where gaps in coverage between survey
lines are unlikely to yield potential hazards/strikes. The vast majority of the
Corps deep- and shallow-draft navigation projects . . . fall within this
category.
• Underwater investigation
surveys. Precise investigation surveys of/around locks, dams, power plants,
abutments, piers, jetties, bulkheads, and other structures.
The USACE depth uncertainty standards include all uncertainty components
that make up a reduced elevation: uncertainties in datum, in tide/stage
modelling-extrapolation-interpretation, in dynamic-latency/roll/pitch/heave, in
acoustic measurement, sound speed, refraction, and beam forming, and
bathymetric mis-modelling through uncertainty in horizontal positioning (depth
georeferencing uncertainty). The Manual
notes that mechanical and acoustic depth measurement uncertainty increases with
increasing depth, that multibeam system uncertainties increase with increasing
beam angle, and that tide/stage and water level surface model uncertainties
will generally be smaller for shallow (<5m) projects than for deeper
(>12m) projects. The USACE depth
uncertainty standards are depth –dependent, but do not follow the S44 4th
Edition a/b coefficient model for depth
independent and depth dependent uncertainty components.
Table 3. USACE depth uncertainty standards (2001
draft version)
Fig. 3: Log-log plot of USACE
standards, and S44 4th Edition Special Order
The LINZ standard, specifically addressing MBES
performance
In response to a request from Land Information New Zealand (LINZ), John
Hughes Clarke, of the University of New Brunswick, prepared a set of 'Provisional
Swath Sonar Survey Specifications' (Hughes Clarke, 1999) for surveys involving
the use of multibeam sonar echosounders (MBES). The rationale for this project was as follows:
The [IHO S44 4th Edition] standards unfortunately contain significant
ambiguity and are drafted for the sole purpose of data collection for nautical
charting (a mandate much narrow than that of LINZ). One example of this broader mandate is that, as of July 1997,
LINZ has taken the responsibility for New Zealand’s Continental Shelf
Delimitation Project. This involves the
'measurement and analysis of seabed information according to internationally
agreed criteria developed by the United Nations Commission on the Law of the
Sea (UNCLOS)'. Unfortunately these
criteria do not include any specifications for the acquisition or delivery of
data that might be acquired by MBES.
The LINZ report explains that uncertainties associated with MBES depth
measurements, expressed as a percentage of water depth (coefficient 'b' in S44
4th Edition) are actually smaller in
deep water than in shallow water.
Depth-independent factors such as tide and heave, and one of the major
depth-dependent factors, unstable sound velocity profiles, all have larger
magnitudes in shallow (inshore) water than in deep (offshore) water. Consequently the depth uncertainties
resulting from imperfect measurement/recovery of these factors, are also far
more significant in shallow than in deep water. The report points out that uncertainties as small as 0.2% of
water depth have been reported for deep water MBES depths. To demand only 2.3%, as specified in S44 4th
Edition Order 3, ignores the capability of MBES, and is less appropriate than
S44 3rd Edition, which required 1% for both shallow and deep water depth
measurements.
The LINZ report also explains that MBES bottom detection, roll, and
refraction uncertainties are all larger for outer beams than for inner (near
nadir) beams. Bottom detection
uncertainties for the inner beams of a typical MBES are in the range of 40% to
60% of the S44 4th Edition Special Order depth measurement specifications. On the other hand, bottom detection
uncertainties alone will exceed the entire Special Order uncertainty limit
(from all sources) for outer beams (say those with a grazing angle of less than
30°). Therefore, a MBES survey designed
to meet a particular depth uncertainty standard for all beams (out to a certain
outer-beam cutoff), will likely outperform that uncertainty standard
significantly for the inner beam (near nadir) data.
This MBES beam-angle dependence is not addressed in S44 4th Edition. The LINZ report addresses this dependence
head-on by proposing MBES depth uncertainty specifications based on the
differences between inner-beam and outer-beam uncertainty performance. Rather than requiring that all depths from a
MBES survey meet the same uncertainty standard, inner-beam standards are
required to meet something closely related to S44 4th edition Special Order,
while the outer-beam standards are more relaxed. In addition, the permitted balance between inner-beam and
outer-beam coverage is allowed to relax as the survey specifications move from
LINZ Special Order to LINZ Order 3.
The expected performance of a MBES is divided into several sectors, from
the inner-beam sector to the outermost-beam sector. The number of sectors is allowed to increase from one to four,
and the specified coverage within each sector is partitioned more generously in
favour of the outer-beam sectors, as the survey order descends from Special to
Order 3. Since this approach could be
quite complex to design, realise and assess in practice, a simpler approach is
also proposed, which is based on the performance of the worst (outer beam)
sector. In each case, everything is
tied to the S44 4th Edition Special Order specification, and the lower order
S44 specifications are ignored. Four
uncertainty levels are specified: 1.0,
1.5, 2.0 and 2.5 times the S44 4th Edition Special Order depth uncertainty
specification, that is
For
1.0 x SO, a = 0.25 m, b = 0.75% of depth
For
1.5 x SO, a = 0.375 m, b = 1.125% of depth
For
2.0 x SO, a = 0.5 m, b = 1.5% of depth
For 2.5 x SO, a = 0.625 m, b = 1.875% of depth.
Table 4. Proposed LINZ Depth uncertainty
specifications
IMCA offshore construction standards
The International Marine Contractors Association (IMCA) have adapted the
S44 and LINZ standards to standards for informed decision making in offshore
construction activities (IMCA, 2000).
The intended uses associated with each of the four IMCA depth measurement
uncertainty orders are:
• IMCA First Order - site surveys for offshore engineering, requiring high
quality seafloor definition: Template
or jacket installations; Detailed route
engineering surveys; Route surveys in
confined areas; Surveys in ports and
harbours; Dredging and inshore
engineering surveys
• IMCA Second Order – site
surveys for offshore engineering, less stringent than First Order: Route reconnaissance surveys; Geo-Hazard and clearance surveys; Coastal engineering surveys; Deepwater geophysical and engineering
surveys (conducted by remote vehicle)
• IMCA Third Order – general
bathymetric surveys: Continental shelf
cable route surveys; Continental shelf
charting surveys; Export pipeline route
surveys
• IMCA Fourth Order – Reconnaissance
surveys: Deepwater cable route
surveys; Deepwater charting
surveys; Surveys for Exclusive Economic
Zone assessments and delineation
Table 5. Proposed IMCA Depth Measurement Uncertainty
Standards
Fig. 4: Log-log plot of Proposed LINZ worst-case
sector / IMCA depth uncertainty, and S44 4th Edition Special Order
What’s next for S44?
The IHO formally intends to reconsider S44 on a five year schedule, to
account for technological and procedural improvements as they occur. Hence work on S44 5th Edition is expected to
start soon. This review concludes with
speculation on the issues to be dealt with by the S44 working group tasked with
preparing S44 5th Edition.
Perhaps the most important issue is whether S44 5th Edition should aspire
to address all intended uses for hydrographic data, as was hinted at in S44 4th
Edition. As this paper has tried to
demonstrate, there are many non-nautical-charting uses for hydrographic data,
for which the depth uncertainty standards are quite different (often more
demanding) than the standards provided by S44 4th Edition. This brief review is by no means an
exhaustive survey of these other intended uses for bathymetric data.
An argument in favour of S44 5th Edition addressing all intended uses for
hydrographic data, is that many Hydrographic Offices aspire to be suppliers of
data/information/products to a broader clientele. It has even been argued that the survival of some HOs may depend
upon cultivating a broader user base (Monahan et al, 2001). It would be appropriate for the IHO to
establish data standards within S44 5th Edition which would facilitate these
aspirations.
On the other hand, this approach to a new edition of S44 would require
broader representation on the working group.
The working group would benefit from inclusion of members involved in
specifying the uncertainty requirements for several of the diverse intended
uses for hydrographic data, as listed in S44 4th Edition:
Coastal zone management, environmental monitoring, resource
development (hydrocarbon and mineral exploitation), legal and jurisdictional
issues, ocean and meteorological modelling, engineering and construction
planning.
Here are a few ideas for consideration, when work on S44 5th Edition
begins:
• Consider moving S44 from a
performance standard, to a document that provides guidance on how to apply the
performance standard, both a priori
for planning purposes, and a posteriori to
determine end use (informed decision making) uncertainty.
• Recognise, as the SMA seems to
have done, that the 'bathymetric model' introduced in S44 4th Edition is what
both navigational and non-navigational clients want and use for informed
decision making. Place more emphasis on
specifying, on methods for assessing and on methods for informing end users, of
the uncertainty associated with this model, and products based upon it (in
contrast to depth measurement uncertainty).
• Consider separating
navigational intended uses into use for (a) certified commercial navigation,
(b) uncertified commercial navigation, (c) recreational boating and (d)
military operations, with uncertainty management standards specific to each
category. Specify the quantity and
spatial distribution of redundant measurements, as well as methods of analysing
them.
• Clarify the issue of the
maximum depth to which the depth uncertainty associated with a particular order
of survey should be applied. Consider
removing all limits (essentially stressing that S44 represents minimum standards).
• Consider simplifying the
relationship between the various orders of survey, by tying the depth
uncertainty definitions for Orders 1, 2 and 3 to multiples of the Special Order
uncertainty, as has been done in the proposed LINZ and IMCA standards.
• Reconsider depth of water
column as the sole independent quality variable. For work from submerged submarines, ROVs and AUVs, depth under
the sensor would be a more appropriate quality variable than depth of the water
column. Accurate high resolution
bathymetry is often required in deep water for marine construction
surveys. Bottom slope and roughness,
area ensonified and multibeam beam angle should be considered as additional
quality variables.
• Consider providing guidelines
for managing all three types of uncertainties (accidental, systematic and
random) rather than providing a performance standard based on random
uncertainties alone.
Fig; 5: Compilation of all
depth uncertainty standards from Figures 1 to 4.
Conclusion
For simplicity, this paper has focused on one quality factor – depth
measurement uncertainty – in discussing the status of IHO S44, and other
hydrographic survey standards that have recently been proposed and
discussed.
However, these standards are more complex than has been represented here,
containing many other quality factors such as target detection uncertainty,
georeferencing uncertainty for depths, georeferencing uncertainty for navaids
and other features and tide and tidal stream uncertainty. It would be possible to prepare parallel
discussions to this paper, concentrating on each of these other factors in turn
(although there are no plans to do so!).
S44 4th Edition broke new ground, and has stimulated the development of
several alternative standards. Its
impact is still a long way from being fully felt or implemented.
However, there are sufficient challenges that have emerged for useful
work to begin on S44 5th Edition. This
paper has tried to identify and describe some of these challenges. The most important of these by far, is the
need to clearly identify in S44 5th Edition whether or not the standards should
apply to the broader intended uses of hydrographic data (only partially listed
in this paper).
References
Hare, R. 1997, Procedures for
Evaluating and Reporting Hydrographic Data Quality.Canadian Hydrographic Service, Internal Report, 176pp.
Hare, R. and Monahan, D. 1993. A modern quantification of historic
hydrographic data accuracy. Lighthouse,
No 48, pp. 1-14. and FIG XX International Congress, Melbourne, Australia
Hecht, H.2001.The Digital Hydrographic Office –
Challenges and Prospects of Hydrography in the Evolving Geographic Information
Infrastructure. US Hydrographic Conference, CD-ROM.
Hughes Clarke, J. 1999. Provisional Swath Sonar Survey
Specifications. TH Technical Report
No.2, August. Available online (2
November 2001) at www.omg.unb.ca/~jhc/LINZ_spec/
International Hydrographic Organisation, 1987. IHO Standards for Hydrographic Surveys, Classification Criteria for
Deep Sea Soundings and Procedures for Elimination of Doubtful Data. International
Hydrographic Organisation, Special Publication
No 44, 3rd Edition, 32pp.
International Hydrographic Organisation, 1998. IHO Standards for Hydrographic Surveys, International Hydrographic
Organisation, Special Publication No
44, 4th Edition, 23pp.
International Marine Contractors Association (IMCA), 2001. Guidelines
on the use of multibeam echosounders for offshore surveys. Presentation by
Edwin Danson, Chairman of the Institution of Civil Engineering Surveyors,
Hydrographic Survey Panel, to The Hydrographic Society in Scotland. See also www.imca-int.com/survey/
Kenny, M. 2000. For Hydrographic
Data, There’s a World Beyond Charting. Canadian Hydrographic Conference, 2000,
15-19 May 2000, CD-ROM.
Kielland, P. and Dagbert, P. 1992.
The Use of Spatial Statistics in Hydrography. International Hydrographic
Review, March, Vol. LXIX, No. 1,
pp. 71-92.
Monahan, D., Hecht, H., Wells, D., Kinney, M. and Campos, A. 2001
Challenges and Opportunities for Hydrography in the New Century. International Hydrographic Review,
Autumn 2001.
Monahan, D. 2001. Variable errors and fixed boundaries: the role of deep
echo-sounding in the United Nations Convention on Law of the Sea (UNCLOS) Lighthouse,
No. 59, pp17-23.
Monahan, D. and Wells, D. 2001. Deep water challenges to hydrography
stimulated by the United Nations Convention on Law of the Sea (UNCLOS). US Hydrographic Conference, 21-24 May,
Norfolk, Va, un-paginated CD-ROM.
Swedish Maritime Administration (SMA), 2001. The Swedish implementation of S-44. 9th Baltic Sea Hydrographic
Commission Conference, Estonia, 5-7 June 2001
United Nations, 1999. Scientific
and Technical Guidelines of the Commission on the Limits of the Continental
Shelf. UNCLOS document CLCS11, 13
May 1999, 91 pp.
United States Army Corps of Engineers (USACE), 2001. Hydrographic Surveying Manual (EM 1110-2-1003) – new draft version,
1 January 2001, Chapter 3: Corps
Accuracy Standards, Quality Control, and Quality Assurance Requirements. Available online (2 November 2001) at
www.thsoa.org/newcorpshsm.htm
Editor’s AcKNowledgement
This paper, previously published in Lighthouse
Edition 60 (Fall/Winter 2001), is reproduced by kind permission of the authors
and the Canadian Hydrographic Association.
Dave Wells
retired from the
University of New Brunswick in 1998, after two decades of teaching hydrographic
surveying there. He now follows the
seasons, and through 2001 taught at four universities: University of Southern
Mississippi from January to May; University of New Hampshire in May and
June; Universiti Teknologi Malaysia
during the summer; and from September to December back to the University of New
Brunswick, where is now Professor Emeritus, and where last autumn he introduced
three new hydrography courses (on tides, kinematic positioning, and
hydrographic data management). Dave and
three colleagues teach a short course on multibeam sonar surveying four times
each year at various locations around the world. Since 1990 Dave has been a member of the FIG/IHO International
Advisory Board on Standards of Competence in Hydrographic Surveying. He is also exploring the possibility of
doing some of his teaching via online (or CD) course delivery, thereby reducing
some of his hectic travel.
Dave Monahan
is Director of Ocean
Mapping in the Canadian Hydrographic Service in Ottawa and Hydrographer in
Residence with the Department of Geodesy and Geomatics Engineering at the University of New Brunswick. He is
Vice-Chairman of the International Hydrographic Organisation’s General
Bathymetric Chart of the Oceans project and has, over the past thirty years,
has worked in most elements of hydrography.
|
