Manual And Guide 14 21 09 06

User Manual:

Open the PDF directly: View PDF PDF.
Page Count: 88

Manuals and Guides 14
Intergovernmental Oceanographic Commission
.......... . . . . . . . . . .
JCOMM Technical Report No. 31
WMO/TD. No. 1339
Manual on Sea Level
Measurement and Interpretation
Volume IV: An Update to 2006
Intergovernmental Oceanographic Commission (IOC)
United Nations Educational, Scientific and Cultural Organization
1, rue Miollis
75732 Paris Cedex 15, France
Tel: +33 1 45 68 10 10
Fax: +33 1 45 68 58 12
Website: http://ioc.unesco.org
.......... . . . . . . . . . .
Manual on Sea Level
Measurement and Interpretation
Volume IV: An Update to 2006
Manuals and Guides 14
Intergovernmental Oceanographic Commission
JCOMM Technical Report No. 31
WMO/TD. No. 1339
.......... . . . . . . . . . .
The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part
of the Secretariats of UNESCO and IOC concerning the legal status of any country or territory, or its authorities, or concerning the delimitation of the frontiers
of any country or territory.
Layout and design by Eric Loddé
For bibliographic purposes, this document should be cited as follows:
Manual on Sea-level Measurements and Interpretation, Volume IV : An update to 2006. Paris, Intergovernmental Oceanographic Commission of UNESCO. 78 pp.
(IOC Manuals and Guides No.14, vol. IV ; JCOMM Technical Report No.31; WMO/TD. No. 1339) (English)
Rev. 2006/09
Printed in 2006
by the United Nations Educational, Scientific and Cultural Organization
7, place de Fontenoy, 75352 Paris 07 SP
© UNESCO 2006
Printed in France
(SC-2006/WS/38)
Sea Level Measurement and Interpretation
1. Introduction .......................................................................................................... 1
2. The Nature of Sea Level Variations ..................................................................... 2
2.1 Introduction ...................................................................................................................................2
2.2 Surface Waves ................................................................................................................................3
2.3 Seiches ...........................................................................................................................................4
2.4 Tides ..............................................................................................................................................4
2.4.1 Tidal Analysis .................................................................................................................................5
2.5 Storm surges ..................................................................................................................................6
2.6 Tsunamis ........................................................................................................................................7
2.7 Mean Sea Level and Trends ............................................................................................................7
2.8 Estimation of Extreme Sea Levels ....................................................................................................7
2.8.1 Introduction ..................................................................................................................................7
2.8.2 The Annual Maximum Method (AMM) .........................................................................................8
2.8.3 The Joint Probabilities Method (JPM) .............................................................................................8
2.8.4 The Revised Joint Probabilities Method (RJPM) ..............................................................................9
2.8.5 The Exceedance Probabilities Method ...........................................................................................9
2.8.6 Spatial Estimation of Extremes ......................................................................................................9
3. Instruments for the Measurement of Sea Level .............................................. 10
3.1 Introduction .................................................................................................................................10
3.1.1 The Choice of a Tide Gauge Site ................................................................................................
10
3.2 The Stilling Well ...........................................................................................................................12
3.2.1 Datum Switches ..........................................................................................................................12
3.3. Pressure Gauges ..........................................................................................................................13
3.3.1 Pneumatic Bubbler Gauges .........................................................................................................13
3.3.2 Pressure Sensor Gauges ..............................................................................................................14
3.3.3 The Datum of a Pressure System .................................................................................................15
3.3.4 Multiple Pressure Transducer Systems (‘B’ gauges) ......................................................................16
3.3.5 Pressure Transducers in Stilling Wells ...........................................................................................16
3.3.6 Bottom-mounted Pressure Gauges .............................................................................................16
3.4 Acoustic Tide Gauges ...................................................................................................................17
3.4.1 Acoustic Gauges with Sounding Tubes .......................................................................................17
3.4.2 Acoustic Gauges without Sounding Tubes ..................................................................................17
3.5 Radar Gauges ...............................................................................................................................19
3.6 Summary of the Merits of Different Technologies .........................................................................21
iii
Table of contents
4. Datum Control and Levelling ............................................................................. 28
4.1 Datums and Benchmarks ..............................................................................................................28
4.1.1 Tide Gauge Benchmark (TGBM) ..................................................................................................29
4.1.2 GPS Benchmark (GPSBM) ............................................................................................................29
4.1.3 Gauge Contact Point (CP) ...........................................................................................................29
4.1.4 Tide Gauge Zero (TGZ) ................................................................................................................29
4.1.5 Revised Local Reference (RLR) Datum ..........................................................................................29
4.1.6 National Levelling Network .........................................................................................................29
4.1.7 Chart Datum ...............................................................................................................................30
4.1.8 Working Datums .........................................................................................................................30
4.2 Levelling Between Local Benchmarks ............................................................................................30
4.3 Levelling Between Wider Area Marks ...........................................................................................30
4.4 Geodetic Fixing of Tide Gauge Benchmarks ..................................................................................31
4.4.1 Introduction ...............................................................................................................................
31
4.4.2 GPS Measurements .....................................................................................................................32
4.4.3 DORIS Measurements ................................................................................................................
33
4.4.4 Absolute Gravity Measurements ................................................................................................
33
4.4.5 Geocentric Co-ordinates and Vertical Land Movements of Tide Gauge Benchmarks ..................
34
5. Real Time Data Transmission ............................................................................. 35
5.1 Introduction .................................................................................................................................35
5.2 Choice of a System ......................................................................................................................38
5.3 Data Transmission Systems ......................................................................................................................39
5.3.1 Systems already well established .................................................................................................39
5.3.2 Systems now being applied or considered for application
in the transmission of sea level data ............................................................................................39
5.3.3 The Global Telecommunications System (GTS) ............................................................................41
5.4 Data Transmission Formats ...........................................................................................................41
6. Quality Control of Data ...................................................................................... 42
7. Training Materials and Contacts ........................................................................ 43
8. New Techniques for Sea Level Measurements ................................................. 45
8.1 GPS on Buoys ...............................................................................................................................45
8.2 GNSS Reflectometry .....................................................................................................................46
References ............................................................................................................... 48
Appendices .............................................................................................................. 52
I. GLOSS Requirements for Tide Gauges .............................................................................................52
II. Previous Volumes of the IOC Manual ..............................................................................................53
III. List of Websites ..............................................................................................................................55
IV. List of Acronyms ............................................................................................................................56
iv
Sea Level Measurement and Interpretation
Sea Level Measurement and Interpretation
V. Contributed Practical Experiences with Various Tide Gauge Technologies ...................................................58
FLOAT GAUGES IN STILLING WELLS: EXPERIENCE IN NORWAY ...........................................................58
D. Hareide, H. Hodnesdal, T. Tørresen and T. Ellef Hansen Østebøvik
THE ESEAS-RI SEA LEVEL PILOT STATION IN VILAGARCÍA DE AROUSA ...............................................61
B. Martín, B. Pérez, E. Alvarez Fanjul
COMPARISON OF RADAR DEVICES IN GERMANY ...............................................................................65
C. J. Blasi and U. Barjenbruch
EXPERIENCE WITH SRD TIDE GAUGES AND REASONING
BEHIND CHANGE TO RADAR TIDE GAUGES .......................................................................................67
R. Farre
PRESSURE GAUGE BASED GLOSS SEA LEVEL STATION
AT TAKORADI HARBOUR (GHANA, WEST AFRICA): EXPERIENCES OVER A YEAR ................................69
A. Joseph, P. Mehra, J. Odammetey and N. E. Kofi
CHILEAN SEA LEVEL NETWORK ..........................................................................................................72
J. Fierro
GAUGES FOR TSUNAMI WARNING ....................................................................................................75
B. Kilonsky
ODINAFRICA TIDE GAUGE SPECIFICATION AUGUST 2005 ..................................................................76
P. Foden
TIDETOOL – A SOFTWARE PACKAGE TO DISPLAY AND DECODE SEA LEVEL DATA
TRANSMITTED OVER THE WMO GLOBAL TELECOMMUNICATIONS SYSTEM ......................................77
L. Kong
v
IOC Manuals and Guides No 14 vol IV
This is the fourth in the series of IOC Manuals on Sea
Level Measurement and Interpretation. It incorporates
the changes in tide gauge technology and measure-
ment techniques in the five years since the third
manual was written, and includes material from the
Workshop on New Technical Developments in Sea and
Land Level Observing Systems (UNESCO, Paris, 14–16
October 2003). In addition, it reflects to a great extent
the changes in priorities for tide gauges in a global
network which have taken place in recent years. For
example, it is inconceivable now that most gauges
installed in the GLOSS network will be without a real-
time reporting capability and a capacity to provide data
of use to a tsunami warning system.
The manual includes some sections of text from the
earlier editions, updated as appropriate. However,
for reasons of space it does not include some other
sections from the earlier versions, even though they
are still valid and useful (e.g. the discussion of data
quality control and filters in Volume III, see the present
Appendix II). The earlier editions continue to be readily
available on the web at http://www.pol.ac.uk/psmsl/
manuals/.
In order to provide a fresh perspective, this volume
has been largely written by new people. A consultant
(Dr. Ian Vassie) produced a first draft. Drs. Tilo Schöne
and Georg Beyerle of GFZ, Postdam, contributed the
text for section 8. The first drafts were commented
on and edited by the GLOSS Technical Subcommittee
(Chair Dr. Begoña Pérez) and the volume was subse-
quently reviewed by members of the GLOSS Group
of Experts and Mr David Meldrum provided additional
comments on Section 5.
The following section provides a brief overview of
sea level variations which may be of general interest,
including a discussion of estimation of extreme levels
that was missing from earlier editions. However, the
volume is largely concerned with tide gauge and data
communications technologies and aimed at people
who work in those fields. These are rapidly develop-
ing topics, and ones in which the sharing of expertise
among groups is essential. Some readers of this volume
may, therefore, have different perspectives on sea level
measurements. Some of these independent views are
expressed in the contributions given in Appendix V.
Each of these authors has expressed willingness to pro-
vide advice to others as required.
We thank everyone who contributed material for, and
advice on, this volume. In particular, we thank Robert
Smith of the Proudman Oceanographic Laboratory for
his technical assistance and Ray C. Griffiths for editorial
assistance.
Thorkild Aarup
(GLOSS Technical Secretary)
Mark Merrifield
(Chair GLOSS Group of Experts)
Begoña Pérez
(Chair GLOSS Technical Subcommittee)
Ian Vassie
(Consultant)
Philip Woodworth
(Director, Permanent Service for Mean Sea Level)
June 2006
Sea Level Measurement and Interpretation
1
1. Introduction
IOC Manuals and Guides No 14 vol IV
2.1 Introduction
The study of sea level has many different facets. It is not
simply the measurement of the sea level that requires
technical expertise. The data must be carefully calibrated,
checked and evaluated. The measurements should be tied
to local benchmarks that in turn are fixed into a country’s
national levelling network and further fixed into the global
network using modern geodetic techniques. The recorded
data need to be archived, documented and protected
for future studies. Only then is it of benefit as a valuable
resource and can be used for studies ranging from local
engineering projects to long-term global climate change.
Variations in sea level contain contributions from dif-
ferent physical sources that are usually distinguished by
their period. Components range from surface gravity
waves with periods of 1 to 20 seconds; seiches and
tsunamis with periods of minutes to over an hour; tides
centred around 1/2 and 1 day; meteorological effects
of several days to 1 year; interannual and decadal
variability; and long-term trends in the mean level
caused by geological and climatological effects. The
magnitudes of these components vary enormously.
Surface gravity waves can have amplitudes up to 30 m.
Tsumanis tend to be less than 1 m in the deep ocean
but may be several metres near the coast. Tides are
relatively small in the ocean but may be 10 metres near
the coast. Storm surges may be of the order of a few
metres in shallow seas. Within this mix one is trying
to estimate long-term trends in the mean level of the
order of 1 mm per year. The fact that this is possible,
and has been for over 100 years, is testimony to the
expertise and dedication of the engineers and scientists
who are involved in sea level research.
The majority of historical sea level data were collected from
float and stilling-well tide gauges with analogue charts,
many of which are still in existence, but superseded by the
modern trend to the digital systems described below. With
digital technology it is possible to improve the accuracy
and reliability of the data and make the data available to
the user in real time.
In analogue form the charts were always available for
re-analysis and errors could be rectified by reappraisal of
the chart and re-sampling of the pen-trace, if necessary.
In digital form a corresponding re-analysis is not always
possible. The decision has to be made in advance as to
what is a reasonable sampling (or averaging) interval. One
cannot return and re-sample the data at a more frequent
interval. In the past, the generally accepted sampling (or
averaging) rate was 1 hour, since this allowed the study
of all processes, from tides to mean sea level (IOC, 1990).
Waves were, by their nature, considered a different sci-
entific province and were filtered out of the data. More
recently, the sampling frequency has been increased to 15
minutes, 6 minutes and even higher rates.
The disastrous tsunami of 26 December 2004 in the
Indian Ocean made it clear that the normal tide gauge
sampling would be inadequate and that it would be nec-
essary to increase it to 1 minute or ideally to 15 seconds.
This places constraints on the tide gauge technology and
increases the demand on the storage and transmission
requirements of a tide gauge network. There is a balance
to be struck between the need to capture the essence of
the data and the need to store and perhaps transmit large
volumes of data.
2. The Nature of Sea Level
Variations
2
Sea Level Measurement and Interpretation
IOC Manuals and Guides No 14 vol IV
A second important issue is that, historically, a tide gauge
was attended continuously by a trained observer who
collected ancillary tide-pole information, and height and
datum corrections were appended to the chart weekly.
This produced a very stable reference and of course meant
that faults were quickly identified. In modern systems the
datum and calibrations tend to be checked less frequently.
Thus greater reliance is placed on the accuracy and stability
of the measuring equipment. Fortunately, modern techno-
logical improvements have allowed this, not only through
better equipment, but with two-way communication the
sea level station can be interrogated and its operational
characteristics adjusted as necessary.
The need for an operator to be permanently at the tide
gauge has been removed. Perhaps one can speculate that
it is time to withdraw all manual intervention. Certainly,
with the growing requirement for real-time data, manual
intervention will not always be possible. In the future, the
only viable approach might be to check and authenticate
the data automatically at source before transmission. It can
then be passed to the end user and be placed in a form
that can be entered directly into the global sea level data
banks without intervention.
2.2 Surface Waves
Surface waves are probably the most noticeable variation
of the sea surface to a casual observer. They have been
relatively little discussed in previous editions of this manual,
as most tide gauges are designed to filter out such waves.
However a brief description of their characteristics is worth
including, as the design of a tide gauge relies on an under-
standing of their general characteristics.
Waves are characterized as wind-waves or swell. Wind-
waves are generated by the effect of the wind on the local
sea surface and have a relatively broad spectrum. Swell is
produced when the waves propagate out of a storm area.
They occupy a narrower part of the spectrum. In general,
wind waves have periods from 1 to 15 seconds, and swell,
from 12 to 25 seconds, although this definition is not
exclusive. Outside this range of periods, wave amplitudes
are small. Wave period is usually calculated via the time
between successive zero up-crosses of the wave (Tz).
Wave heights are usually defined in terms of their
peak-to-trough range in height, although wave ampli-
tude is sometimes calculated as the height above a
mean level. Significant wave height (Hs) is the usu-
ally quoted parameter which closely approximates
the height of the highest one-third of the waves in a
given period of time. Traditionally, a wave record has
a duration of 20 minutes and is re-sampled every 3
hours, choices which were derived originally from the
stochastic properties of storm duration. It is difficult
to give an overall figure for maximum wave height,
as it depends critically on location. Waves are subject
to amplification, dispersion, refraction and focusing.
In general, significant wave heights of several metres
are common during a storm, but individual waves up
to 30 metres have been measured.
Figure 2.1 Spectrum of Sea Level Variations. The long-period variations and mean sea level changes are part of the
enhanced energy at low frequencies.
Sea Level Measurement and Interpretation
3
IOC Manuals and Guides No 14 vol IV
Wave activity with a period of a few minutes can be
caused by non-linear effects; e.g. when the waves
encounter a current or a change in bottom topography.
These longer-period waves occur because the height of
successive waves is not uniform; they occur in groups
of higher or lower waves. This leads to the popular
misconception that every seventh wave is the highest.
In fact, the wave groups are not of equal length but
they do produce non-linear effects that have periods
related to the period of the wave groups. The most
significant effect of this, as far as the study of sea level
is concerned, is that the wave groups produce ‘set-up’
of the sea level near the coast. The degree of ‘set-up’
depends on many factors, of which the shape of the
beach is the most critical. Set-up can be of the order of
a few tens of centimetres during a severe storm.
Waves have directional properties as well as a magni-
tude. Many early recordings were only concerned with
wave height, because instruments capable of measur-
ing direction were not available. Wave riders from this
era were moored to the sea bed on a flexible coupling
and contained accelerometers which were integrated
twice to obtain wave height. However, modern moor-
ings are now available which are capable of measuring
pitch and roll of the surface buoy, from which direc-
tional information can be derived.
Coastal tide gauges tend not to be located optimally
to measure wave conditions in the nearby deep ocean.
However, they can at times provide useful information
with the correct (pressure) gauge technology. Vassie
et al. (2004) provide a recent description of the use
of pressure tide gauges to measure swell at ocean
islands.
2.3 Seiches
Seiches are periodic variations in the surface level usually
set in motion by a disturbance such as a strong wind or cur-
rent, a sudden change in atmospheric pressure or even a
tsunami. In lakes and gulfs their period is controlled by the
dimensions of the basin and their lifetime is determined by
frictional effects. Typical periods are in the range of a few
minutes to a few hours (between wind waves and tides),
and typical amplitudes are centimetric to decimetric. They
can be seen on tide gauge records from almost all regions.
Seiches have largely been ignored in most sea level stud-
ies, owing to their primarily local origin, but knowledge
of them is important for coastal and harbour engineering
as well as for harbour operations, where small-amplitude
seiches may be associated with strong currents at the
entrance of the harbour. On the other hand, they can have
a major effect on other sea level studies. For example, if
their amplitude is large enough, and if the sampling rate of
the tide gauge is insufficiently high, then their energy can
be aliased into tidal and other sea level signals.
2.4 Tides
The oceans respond to the gravitational attraction of the
Moon and the Sun, and the solar radiation, to produce
the tides, which are normally the predominant signals
in sea level records. The tides are easy to distinguish
from other components of sea level variation (e.g. storm
surges) because they have well defined periods, whereas
other processes tend to occur at irregular intervals.
An examination of the forces causing the tides leads
some way towards an understanding of their nature. This
examination is usually via discussion of the Equilibrium
Tide (Doodson and Warburg, 1941; Forrester, 1983;
Pugh, 1987; Open University, 1989). The gravitational
attraction of the Moon and Sun on the Earth produces
a semi-diurnal (2 cycles per day) tidal bulge’, which
is usually oriented at an angle to the equator produc-
ing the diurnal (1 cycle per day) tidal components. The
diurnal and semi-diurnal waves both have a planetary
space scale. As the Earth rotates about its axis, signals
containing the above periods, but usually dominated by
the semi-diurnal component, should appear in the sea
level record. A lunar day is slightly longer than a solar day
by approximately 50 minutes, leading to lunar and solar
tides of differing periods which interact over 14 days to
produce the Spring-Neap cycle.
Study of the celestial motion of the Earth–Moon–Sun
system leads to a more complex form of the tidal poten-
tial (or Equilibrium Tide) in which the main constituents
are modulated at periods of 1 month, 1 year, 8.85 years,
18.61 years and 21,000 years. The effect of the modula-
tion is to split the tides into additional constituents but
with periods close to 1 and 2 cycles per day. This grouping
is termed ‘tidal species’.
The tidal potential so far discussed explains only the diurnal
and semi-diurnal species of the tide, but can be extended
to include ter-diurnal (third of a day period) tides and tides
of even shorter period. A power spectrum of a tidal record
clearly shows that higher-order species do exist, except
sometimes when measurements are made at an oceanic
location. These ‘compound tidesare primarily generated
by the main tidal components in shallow water as they
encounter frictional forces. They have periods of 2, 4 and 6
cycles per day (and even 12 cpd in very shallow areas), with
each species demonstrating separate tidal characteristics.
The tidal regime varies enormously in different parts
of the world. In most regions the tide is dominated by
semi-diurnal components, reflecting the importance of
the main semi-diurnal terms in the Equilibrium Tide.
However, there are many areas where the tides are pre-
dominantly diurnal (e.g. Persian Gulf), and some where
the regime is ‘mixed’ (i.e. the diurnal and semi-diurnal
components have a comparable magnitude). Examples
of these various regimes are shown in Figure 2.2.
4
Sea Level Measurement and Interpretation
IOC Manuals and Guides No 14 vol IV
Sea Level Measurement and Interpretation
5
While the temporal characteristics of the tide in the real
ocean are similar to those of tidal potential (Equilibrium
Tide), their spatial characteristics are very different. This
difference is caused by the dynamical response of the
ocean basins, causing the tides to propagate as progres-
sive waves and to generate standing waves in some areas.
Tides in the deep ocean have amplitudes of typically
1 m or less, considerably lower than the amplitudes on
continental shelves where local resonances can produce
large amplitudes. In all oceans (deep oceans as well as the
enclosed sea areas of continental shelves) there are regions
of no tide, called amphidromic points, which are a conse-
quence of the standing waves.
Tide gauges, such as those described in this manual,
remain the primary source of tidal knowledge in coastal
regions, although new techniques are under continuous
development (section 8). The tides of the deep ocean
can also now be well measured, with the use of bottom
pressure recorders (Cartwright et al., 1980; Filloux, 1980;
Spencer and Vassie, 1997), and more recently by means of
altimeter satellites (Shum et al., 1997).
2.4.1 Tidal Analysis
The model that has been derived for the Equilibrium Tide is
not completely without use, as it does provide the knowl-
edge that the tide is composed of a finite number of con-
stituents of calculable frequency. It also provides a measure
of their relative amplitudes so that we have an idea which
constituents are important in the real tide.
The analysis consists in reducing a set of measure-
ments, which amounts to 8,760 hourly values in a
normal year, to a manageable set of parameters which
completely specify the tidal component of the record.
The tides can then be removed to reveal the remaining
Figure 2.2 Tidal characteristics at five stations, showing different regimes: diurnal, mixed, semi-diurnal with strong
spring-neap modulation in the Indian Ocean, semi-diurnal with smaller amplitudes at a N.Atlantic site, and shallow
water distortions.
IOC Manuals and Guides No 14 vol IV
components of the sea level variations (e.g. storm surges,
tsunami) and the long-term trend.
Many organizations have developed their own method
of tidal analysis. Apart from the Response Method (Munk
and Cartwright, 1966), these methods generally fit, in
some optimal way, a set of harmonic constituents to
the data. This can be done in several different ways. The
Admiralty Semi-Graphic Method and those of Doodson
(1928) were designed for hand calculations. Most mod-
ern techniques (Murray, 1963; Foreman, 1997) rely on the
ability of the computer to solve large sets of simultaneous
equations. Many have been converted to ‘user friendly’
packages and are available from the following website:
http://www.pol.ac.uk/psmsl/training/analysis.html.
2.5 Storm Surges
The exchange of energy between the atmosphere and the
ocean is one of the most important topics in geophysics.
Storm surges are among the more spectacular examples
of energy transfer in which the energy contained in winds
and time-dependent changes in air pressure are absorbed
by the sea to produce strong currents and high sea levels.
In the open sea these currents decay by the action of dis-
sipative forces. Where the current is impeded by the pres-
ence of a continental shelf or other discontinuity in depth,
or by a coastline, more of the kinetic energy of the sea
tends to be converted into potential energy. Abnormal
elevations of sea level may then occur, with disastrous
results if the coast is low-lying.
Physically, the atmosphere acts on the sea in two distinct-
ly different ways. The first is the ‘Inverse Barometer (IB)
Effect’ wherein a 1-hPa (mbar) increase of atmospheric
pressure decreases sea level by 1 centimetre. (Dynamical
effects can complicate this simple IB description at short
time-scales.) The second is due to the drag (or ‘stress’)
of the wind on the sea surface, which is proportional (to
a first approximation) to the square of the wind speed.
This force sets up sea level gradients which are propor-
tional to wind stress divided by water depth, and which
result in the storm surges in shallow water regions. The
dynamics of surges in shallow water result in flow being
in the direction of the wind, differing from a deeper
water situation in which the transport is at right angles
to the wind (to the right in the northern hemisphere).
Recordings of sea level at any coastal station contain
some evidence of the influence of winds and pressure,
but some areas are particularly susceptible to large
surges. The Baltic, being virtually an enclosed sea and
subject on occasion to severe gales, experiences large
surges. In 1924 St. Petersburg (Leningrad) was flooded
by a surge 4 m high. The North Sea, with its southern
extremity almost closed, responds readily to northerly
winds; the vulnerable coastlines of the German Bight,
eastern England and more particularly the Low Countries
have repeatedly been inundated by great surges. The
storm surge of 1953 resulted in many deaths in The
Netherlands and England. The Hamburg disaster of
1962 was more localized, mainly affecting the German
Bight and the River Elbe, where the surge reached more
than 3 m in height.
Hurricanes travelling towards the Atlantic seaboard of
the United States are no less effective in generating
destructive surges. The Japanese islands are also subject
to typhoon surges. Events on this scale demand as com-
plete an understanding of the phenomena as possible
so that they may be forecast (using forecast meteoro-
logical information) and their consequences mitigated.
After the immediate danger of flooding, the subsequent
dislocation of normal services, such as water supplies
and sewerage, gives rise to serious dangers. Also, once
flooded by sea water, previously fertile lands are unsuit-
able for growing crops for several years because of
the saline deposit which remains after the floods have
receded.
For scientific analysis and for systems designed for surge
prediction, it is usual to distinguish between tropical and
extra-tropical surges.
Tropical surges are generated by tropical storms that are
small and very intense. These storms are generated at
sea, from where they move in an irregular way until they
meet the coast. Here they produce exceptionally large
flood levels over a region of perhaps 10–50 km of coast-
line. Tropical storms are difficult to monitor offshore and
their effects on a particular stretch of coastline cannot be
estimated from the statistics of observed floods because
such storms are relatively rare events in any particular
region. A combination of numerical and statistical mod-
els may be used to estimate the maximum flood levels,
but their exact location depends on the track of each
individual storm.
Extra-tropical surges are generated by storms which
extend over several hundred kilometres and which are
generally slow moving. They affect large areas of coast
over periods that may extend to several days. At their
centre is a region of low atmospheric pressure. They are
more amenable to study by hydrodynamic modelling
taking into account the distribution of atmospheric pres-
sure and wind fields, sea bed bathymetry, the coastal
topography and the effects of the Earth’s rotation.
A tide gauge network by which the storm surge can
be monitored is of key importance in providing data to
enhance the performance of operational hydrodynamic
tide–surge models used in flood warning. Data can
be used in the verification of the models and for data
assimilation into them (Flather, 2000; Alvarez Fanjul,
2001). Such a network clearly has to be capable of
remote telemetry on a near-real-time basis.
6
Sea Level Measurement and Interpretation
IOC Manuals and Guides No 14 vol IV
Sea Level Measurement and Interpretation
7
2.6 Tsunamis
A tsunami is a wave train generated by a vertical displace-
ment of the water column. Earthquakes, landslides,
volcanic eruptions, explosions, and even the impact
of cosmic bodies, such as meteorites, can gener-
ate tsunamis. Where they impact a coastline, they
can cause severe property damage and loss of life.
Tsunamis may have wavelengths in excess of 100 km
and periods of minutes to over an hour, depending
on the generation mechanism. As a result of its long
wavelength compared to the water depth, a tsunami
behaves like a shallow-water wave and propagates at
a speed that is equal to the square root of the product
of the acceleration of gravity (9.8 m.s-2) and the water
depth. In a typical ocean depth of 4,000 m, a tsunami
travels at about 200 m.s-1, or over 700 km.hr-1. Because
the rate at which a wave loses its energy is inversely
related to its wavelength, tsunamis not only propagate
at high speeds, they can also travel great distances
without loss of energy (Figure 2.3). Tsunamis are only
about a metre high, at the most, in the open ocean.
However, where they impact the coast, amplitudes are
significantly higher and can be as large as 10 m (30
m in extreme cases). Wave refraction, caused by seg-
ments of the wave moving at different speeds as the
water depth varies, can cause extreme amplification in
localized areas.
The ability to warn of the approach of a tsunami depends
on a variety of measurements (especially seismic data), but
also on a network of tide gauges to monitor the progress
of the wave and thereby forecast the time of arrival at a
distant coast and the likely affected areas. Because the
propagation speed of the waves is large, it is essential to
have real-time data transmission without any significant
time delay. Decision-making and mitigation procedures
have to be considered before warnings are issued to the
relevant authorities.
2.7 Mean Sea Level and Trends
The determination of mean sea level (MSL) and its long-
term trend is probably the most exacting component of a
tide gauge data set. Whereas the accuracy of an instru-
ment in determining the properties of the tides or a storm
surge need only be about 1 cm, the long-term trend in sea
level has a magnitude of around 1 mm per year. Hence
precise measurement not only relies on the accuracy of the
instrument but also on its long-term stability. This in turn
implies an ability to maintain the datum of a tide gauge
within a local levelling network. The levelling between,
and geocentric fixing of, tide gauge benchmarks, is dealt
with in section 4.
The data from the existing global network of tide gauges
clearly shows a rise in sea level over the last century. Their
data are fundamental in studies of climate change, and
especially as an aid in the development of atmosphere–
ocean general circulation models that have a capability to
predict future sea level change. The mean value is extract-
ed from the observed data by the application of numeri-
cal filters discussed in Volume 1 of the IOC Manual on
Sea Level Measurement and Interpretation. Monthly and
annual mean sea level series are collected and published
by the Permanent Service for Mean Sea Level (PSMSL),
together with details of gauge location, and definitions
of the datums to which the measurements are referred.
Data are held for over 2,000 stations, of which 112 have
data from before 1900. The longest record held is from
Brest, France, which begins in 1806. The physical location
of gauges on the network is not ideal: the vast majority of
gauges operate in the northern hemisphere and careful
analysis is necessary to avoid bias in the interpretation of
their data. There is a continuing need for more data from
the southern hemisphere, and from oceanic islands.
The change in mean sea level relative to a fixed point on
land is a measure of the difference between the vertical
movements of the sea’s surface and of the land itself. Long-
term changes of measured sea level are termed ‘secular
changes. Global changes in the mean sea level are called
eustatic’ changes. Vertical land movements of regional
extent are called eperiogenic movements. Examples of
such long-term changes can be obtained from the PSMSL
website. Study of the records will show that there are
many similarities between stations which can be considered
nearby’ relative to ocean and geological space-scales. The
close agreement between stations using different kinds
of instruments shows that the oceanographic variability is
much greater than the errors in the measurements.
2.8 Estimation of Extreme Sea Levels
2.8.1 Introduction
The aim of this section is to summarize the key meth-
ods which can be used for the estimation of extreme
Figure 2.3 The 26 December 2004 Sumatra tsunami
signal at a distant tide gauge (Port Louis, Mauritius) with
an amplitude over 1 m.
8
Sea Level Measurement and Interpretation
IOC Manuals and Guides No 14 vol IV
sea levels. It begins with the classical method of Annual
Extremes, which first appeared in the early 1960s and
continued to be developed for some time thereafter.
Following this, the Joint Probability Method, which was
developed in the late 1970s, is considered. This makes
more efficient use of data by incorporating our exten-
sive knowledge of the tides and storm surges, which
are the two main components of sea level, as a part of
the estimation procedure.
More recent work on the Annual Exceedance Method is
discussed, followed by a revision of the Joint Probability
Method to correct its deficiencies in areas where the sea
level is dominated by the meteorological surge compo-
nent. Finally, very recent work on the spatial estimation
of extremes is mentioned. References are given at each
stage so that the reader can examine any of the meth-
ods in greater depth. Although extreme high sea levels
are considered, results for extreme low sea levels can be
obtained in an analogous way.
2.8.2 The Annual Maximum Method (AMM)
This is the classical general method of analysis of
extremes having been applied to sea level estimation
since 1963 (Lennon, 1963; Suthons, 1963). It is based
on a result from probabilistic extreme value theory
which states: if X1,... Xn is a sequence of independent
and identically distributed random variables, then
max(X1,... Xn), suitably linearly normalized, converges
as n , to a random variable with a distribu-
tion function which is one of the so called extreme-
value distributions. The general case is known as the
Generalized Extreme Value (GEV) distribution. An
important special case is the Gumbel distribution.
The Annual Maximum Method takes the GEV to be
the distribution function of the maximum sea level in
a year. Therefore, for a place of interest, the annual
maximum for each year is extracted from hourly obser-
vations and is used as data to estimate the parameters
of the distribution that they follow. From the estimated
distribution one can obtain the sea level corresponding
to a chosen ‘Return Period’. In practice, return periods
of 50, 100 and 1,000 years are common. The basic
method assumes that there is no trend in the data, but
it can be extended to deal with those cases where a
trend is present.
A recent extension of the annual maximum method
involves using probabilistic extreme value theory to
obtain the asymptotic joint distribution of a fixed num-
ber (r) of the largest independent extreme values, for
example the five largest in each year. Essentially the
approach is the same as above except that more rele-
vant data are included in the analysis thereby improving
the estimation. Care must be taken to ensure that the
number of annual maxima ‘r’ is not excessive, such that
the lower extremes fall outside the tail of the extreme
value distribution.
This method of estimating sea level extremes is highly
inefficient in its use of data, since it extracts very few
values from each yearly record. This is particularly impor-
tant when the sea level record is short, since it yields
return level estimates with unacceptably large standard
errors. In addition, it makes no use of our knowledge of
the sea level and storm surge processes. However, the
advantage of annual maxima methods is that they do
not require knowledge of tide–surge interaction which
can sometimes be a significant feature of the data.
Consequently the methods are relatively straightforward
to apply.
2.8.3 The Joint Probabilities Method (JPM)
This method of analysis was introduced to exploit our
knowledge of the tide in short data sets to which the
annual maxima method could not be applied (Pugh
and Vassie, 1979). At any time, the observed sea level,
after averaging out surface waves, has three compo-
nents: mean sea level, tidal level and meteorologically
induced sea level. The latter is usually referred to as a
storm surge. Using standard methods, the first two of
these components can be removed from the sea level
sequence leaving the surge sequence, which is just the
time-series of non-tidal residuals. For simplicity these are
assumed to be stationary. Because the tidal sequence
is deterministic, the probability distribution for all tidal
levels can be generated from tidal predictions. This
distribution can be accurately approximated using 18.6
years of predictions.
The probability distribution of hourly sea levels can be
obtained either directly using an empirical estimate or
by combining the tidal and surge probability density
functions (
pdf
). The latter is preferable, as it smoothes
and extrapolates the former. However the nature of
the combination of the
pdf
s depends on whether
there is dependence between the tide and surge
sequences. Initially, consider the case in which they
are independent.
By combining the
pdf
s of tide and surge, the distribution
function of hourly (instantaneous) sea levels is obtained.
From this, the distribution function of the annual maxi-
ma is required. If hourly values were independent, which
is approximately the case where the tide dominates the
regime, then this is straightforward.
The method has been widely applied. It makes better
use of the data and of our extensive knowledge of the
tides, and accounts for surges that could have occurred
on high tide but by chance did not. Most successful
applications have been to sites which have several years
of hourly records (>10 years) and where the site is tidally
dominant, i.e. where the tidal range is large in compari-
son to the surge amplitude. Least successful applications
have been to sites with both short lengths of data and
where the site is surge dominant.
Sea Level Measurement and Interpretation
9
IOC Manuals and Guides No 14 vol IV
2.8.4 The Revised Joint Probabilities Method
(RJPM)
Particular emphasis was given to two principal
improvements that make the revised method more
widely applicable than the original joint probabilities
method (Tawn et al., 1989). It was principally direct-
ed at sites where the storm surge was responsible
for a respectable proportion of the sea level and to
improve the estimation procedure for sites where
less than 10 years of data were available.
The first issue was that of converting the hourly dis-
tribution into annual return periods. It is clear that
each hourly value of sea level is not independent of
its predecessor or successor. Of the 8,760 hourly val-
ues in a year, it is necessary to determine the effective
number of independent observations per year. This
was done through an Extremal Index which is derived
from the mean overtopping time of a level for each
independent storm which exceeds that level. In fact
the Extremal Index can be shown to be a constant
in the region of the extremes. Because large values
tend to cluster as storms, it should be expected that
the Extremal Index >1; for example, in the North
Sea, it is 1.4. This effectively reduces the number of
independent observations from 8,760 to 8,760/1.4.
If the site is tidally dominant then the Extremal Index
is considerably smaller than if the site is surge domi-
nant. The immediate advantages of this modifica-
tion are: firstly, that no assumption about the local
dependence of the process is required; secondly, that
the conversion from the hourly distribution to annual
maxima is invariant to sampling frequency.
The second modification enabled probabilities for
levels beyond the existing range of the surge data
to be obtained, in addition to providing smoothing
for the tail of the empirical distribution. The method
is based on the idea of using a fixed number of
independent extreme surge values from each year to
estimate probabilities of extreme surges. The proce-
dure involves two important steps. Firstly, the iden-
tification of independent extreme surges. Secondly,
the selection of a suitable number of independent
extreme surges from each year of data, perhaps five
per year. Using these surge data, estimates can be
made of the parameters of the distribution of the
annual maximum surge (Smith, 1986).
Using the ideas for extremes of dependent sequenc-
es, this can be related to the distribution function
of hourly surge levels, and then the empirical surge
density function can be replaced by the adjusted
density. Using the adjusted density function, the
convolution can be performed to combine the tidal
and surge distributions to obtain the hourly sea level
distribution and hence the return periods can be
calculated for different levels.
When interaction is present, the level of the tide
affects the distribution of the surge. In particular, the
tail of the surge
pdf
depends on the corresponding
tidal level. Thus the convolution of tide and surge
can be adapted so that the surge parameters are
functions of tidal level. This formulation also enables
statistical tests of independence to be performed.
2.8.5 The Exceedance Probability Method
(EPM)
An alternative method of obtaining extreme sea
level estimates from short data sets is called the
exceedance probability method (EPM) (Middleton
et al., 1986; Hamon et al., 1989). The EPM, like the
RJPM, involves combining the tide and surge distri-
butions and accounting for dependence in the sea
level sequence. The approach differs in the way that
it handles extreme surges. The EPM uses results for
continuous time processes and makes assumptions
about the joint distribution of the surge and its deriv-
ative. Improvement is achieved by allowing flexibility
in the surge tail through the use of a contaminated
normal distribution.
2.8.6 Spatial Estimation of Extremes
Extreme sea levels along a coastline are typically gener-
ated by the same physical mechanisms, so the param-
eters that describe the distribution are likely to be
spatially coherent. Models that describe the separate
constituents of the sea level are best suited to exploit-
ing this spatial coherence, as the individual parameters
should change smoothly along a coastline.
The joint distribution of annual maxima over sev-
eral data sites can be modelled using a multivariate
extreme-value distribution (Tawn, 1992). Changes in
each of the parameters of the distribution, over sites,
can be modelled to be consistent with the properties
of the underlying generating process identified from
the RJPM. The main advantage of the spatial method
is that it can utilize data sites with extensive sea level
records and augment these with data from sites with
shorter records of a few years.
IOC Manuals and Guides No 14 vol IV
3. Instruments for the
Measurement of Sea Level
10
Sea Level Measurement and Interpretation
3.1 Introduction
This section contains information on the types of instru-
ment that are presently available for the measurement of
sea level. The reason that so many different technologies
have evolved is connected with the difficulty of measur-
ing a fluid that is in constant motion due to the pro-
cesses discussed in section 2. In general, sea level mea-
surements are not concerned with the measurement of
surface gravity waves which must be filtered out of the
system. Waves can be appreciable in amplitude and can
cause problems for most forms of tide gauge technol-
ogy. Therefore, their potential effects on a ‘sea level’
measurement must always be kept in mind. Another
factor that needs to be considered is that the properties
of sea water (salinity, temperature and hence density)
may change on a regular or irregular basis. How this
affects an instrument depends much on the technology
used to acquire the observations. These are discussed
along with the merits of each tide gauge.
There are fundamentally four types of measuring tech-
nology in common use:
A stilling well and float: in which the filtering of the
waves is done through the mechanical design of
the well.
Pressure systems: in which sub-surface pressure
is monitored and converted to height based on
knowledge of the water density and local accelera-
tion due to gravity. Such systems have additional
specific application to ocean circulation studies in
which pressure differences are more relevant than
height differences.
Acoustic systems: in which the transit time of a
sonic pulse is used to compute distance to the sea
surface.
Radar systems: similar to acoustic transmission, but
using radar frequencies.
Within each of these four types, different technologies
have been employed, leading to different designs.
In addition, there are direct measuring devices based
on resistance or capacitance rods, but these have found
less widespread use because of their lack of robustness
in hostile regions. Recent advances in technologies,
such as Global Positioning System (GPS) reflection
methods, have lead to other elaborate ways of measur-
ing sea level which might be important in the future.
At the present time, many of the above systems are
undergoing tests and inter-comparisons by agencies
worldwide (IOC, 2004). It would appear that most
systems for measuring sea level have a precision
approaching 1 cm, given sufficient care and attention.
This value is adequate for the measurement of most
of the hydrodynamic processes discussed in section 2.
However, this precision does not necessarily imply an
accuracy for adequate measurement of the mean level.
The determination of the mean level depends as much
on the long-term stability of the measuring system.
There are practical constraints that govern the choice
of an instrument for a particular application. These
include cost, degree of difficulty of installation, ease
of maintenance and repair, support facilities etc. For
example, the installation of a highly complex electronic
instrument with sophisticated software control would
be unwise without technical support staff who possess
Sea Level Measurement and Interpretation
11
IOC Manuals and Guides No 14 vol IV
the ability to maintain its operation. Another important
consideration in the choice of an instrument is the site
at which it is to be located. This is discussed in the next
section.
Traditionally, permanent sea level stations around the
world have been mainly devoted to tide and mean sea
level applications, and this has been the main objec-
tive of GLOSS. This implies that not only wind waves
are filtered out from the records by mechanical or
mathematical procedures, but any oscillation between
wind waves and tides (e.g. seiches, tsunamis etc.) has
not been considered a priority, and in fact not properly
monitored, owing to the standard sampling time of
more than 5–6 minutes. If this range of the spectrum
should be covered from now on, it would be necessary
to consider this when choosing a new instrument and
designing the sea level stations.
3.1.1 The Choice of a Tide Gauge Site
In many cases, the site for a tide gauge may be speci-
fied (e.g. it has to be located in a port area). However,
in many instances, the choice of site will not be clear
and can only be made by judging which of the con-
straints listed below are more significant and which
should be given greater emphasis. This emphasis may
depend on, for example, whether the gauge is intended
for oceanographic research, in which case one clearly
requires it to be located with maximum exposure to the
open ocean, and not situated in a river. Most GLOSS
Core Network sites have been selected with this aspect
in mind. For local programmes, where the process
to be studied may be coastal erosion or storm surge
activity, then clearly the gauge will have to be situated
optimally for that purpose. In most cases, some of the
following constraints are still valid:
The installation must be capable of withstanding
the worst environmental conditions (winter ice,
storms etc.) likely to be encountered. This is clearly
an issue relevant to the type of instrument and to
its intended position. Positions exposed to envi-
ronmental extremes should clearly be avoided to
enable the eventual accumulation of a long time-
series of data.
The ground on which the installation is to be erect-
ed should be ‘stable’ as far as possible, not being
liable to subsidence because of underground work-
ings or land subsidence (e.g. due to the area being
reclaimed land). It must also not be liable to slip-
page in the event of heavy prolonged rain (i.e. the
area must be adequately drained) or being eroded
by river or sea action. An installation on solid rock
is the ideal.
River estuaries should, if possible, be avoided.
Estuarine river water can mix with sea water to
varying extents during a tidal cycle and at differ-
ent times of the year, resulting in fluctuations in
water density. This may have important impacts
on float gauge measurements in stilling wells
because of ‘layering’ of water drawn into the well
at different times causing a difference in density
inside and outside the well. It will also impact on
pressure measurements, as the density assumed
for the conversion of pressure to sea level will not
be constant. Currents associated with river flow
can also cause drawdown in stilling wells and in
the stilling tubes of acoustic gauges. Following
heavy rain-storms, debris floating down-river
could damage a gauge.
Areas where impounding (isolation from the open
sea) can occur at extreme low-tide levels should be
avoided. Similarly, sandbars slightly below the sur-
face between the site and the open sea can result in
uncharacteristic levels being measured. Monitoring
across long shallow sloping beaches should also be
avoided for the same reasons.
Sharp headlands and sounds should be avoided,
since these are places where high tidal currents
occur which tend to result in unrepresentative tidal
constants and in a drop of MSL (Pugh, 1987).
Proximity to outfalls can result in turbulence, cur-
rents, dilution and deposits, and should be avoided.
Places where shipping passes or moors close to the
proposed site, since there will be a risk of collision
and propeller turbulence causing silt movement; a
study should be made of this possible factor.
Places where construction work in the area at some
future time may affect the tidal regime at the site
(e.g. by construction of new quays or breakwa-
ters); investigations should be made to determine
whether there is a possibility of this occurring. This
might necessitate the relocation of the tide gauge,
thus interrupting the sea level time-series. This is
something very difficult to avoid in some harbours.
A site should have continuous mains electrical
power (or adequate storage batteries/solar panels
or generator supply) and telephone or satellite
access for transmission of data to an analysis
centre.
There must be adequate access to the site for instal-
lation and maintenance and the site must be secure
from vandalism or theft.
The area of the site must be capable of containing
the benchmarks required for geodetic control of
the sea level data. In particular, it must have good
TGBM and GPSBM marks, which must also be
secure from accidental damage.
If stilling well or acoustic gauges are to be installed,
then the stilling well or acoustic tube must be tall
enough to record the highest sea levels. This may
require permission from port authorities if, for
example, the installation is on a busy quayside.
The water depth must extend at least two metres
beneath Lowest Astronomical Tide (LAT) for the
successful operation of a stilling well. The outlet of
the stilling well should be clear of the sea bed and
IOC Manuals and Guides No 14 vol IV
be set deep enough to allow the float to operate
about one metre below LAT.
Finally, it is clear that tide gauge datum control is an
essential issue for any installation. Consequently, even
if the station is equipped with the most modern equip-
ment, it is common sense to provide confirmation of
the datum from time to time by means of an inexpen-
sive tide ‘pole’ or ‘staff’ to guard against gross errors
in the datum.
3.2 The Stilling Well
A stilling well gauge is probably the most common of
all sea level recording systems on a worldwide basis.
These gauges were at one time employed at every port
installation and were the primary technology by which
sea level records were compiled. Recent stilling well
installations are less common, since they require a con-
siderable amount of costly engineering work, so that
they have often been superseded by one of the other
technologies discussed below. In some circumstances it
may not be possible to install a well, e.g. on a shelving
beach, and other methods have to be adopted.
The function of a well is to filter out, ‘to still’, the wave
activity, so that the tides and longer-period processes
can be recorded accurately. It is most commonly asso-
ciated with having a float gauge in the well driving a
pen and chart recorder or, in more recent years, a shaft
encoder such that the readings of sea level height can
be digitized automatically. It is not uncommon for other
types of instrument, e.g. a pressure sensor, to also be
placed in the well.
The well itself is a vertical tube about 1 m in diameter
constructed of concrete, coated steel or plastic, with a
hole or, less frequently, a pipe connection to the sea.
The ratio of the hole diameter or pipe length and diam-
eter to that of the well gives it the characteristics of a
low pass filter (Noye,1974a, b, c). In other words, it acts
as a mechanical filter. Care has to be exercised in trying
to measure processes such as tsunami waves, as the
frequency response is not 100% for periods 4 hours.
The stilling well suffers from amplitude attenuation
and a phase lag at shorter periods which are critically
dependent on the design of the well and sometimes
difficult to change.
The characteristics, installation and use of a stilling well
were covered in substantial detail in Volume 1 of the
Manual of Sea-Level Measurement and Interpretation
(IOC, 1985). The reader is advised to refer to that publi-
cation, and for additional information on the character-
istics of the stilling well, to Noye (1974). Lennon (1971)
dealt in detail with errors that arise in the operation of
such a system.
A schematic diagram of a float gauge in a stilling well
is shown in Figure 3.1. The float wheel is shown driving
a pen recorder, but the same pulley could equally drive
a digital shaft encoder or a potentiometer, which can
then be recorded by a local data logger or interfaced
to a telemetry system. The well is shown with a conical
inlet at its base, since this is the most common con-
figuration and is to some extent self-cleaning. Many
other configurations of the inlet are acceptable, and
although the conical orifice does restrict the inflow
relative to the outflow, this does not appear to have a
significant effect on the records even in the presence
of waves.
3.2.1 Datum Switches
In common with all other types of sea level recording
systems, the setting and control of datums is of cru-
cial importance. This topic is dealt with in section 4.
Stilling well tide gauge installations were, at one time,
attended on a continuous basis. Under these circum-
stances visual comparisons were made with a fixed tide
gauge staff on a regular basis and appropriate time and
datum corrections were applied to the data. Without
this, alternative means of fixing the datum have to be
found. One alternative is to site a level switching device
as part of the installation at approximately mean sea
level. The switch indicates the instant at which the sea
Figure 3.1 Stilling well tide gauge.
12
Sea Level Measurement and Interpretation
IOC Manuals and Guides No 14 vol IV
crosses the level of the switch, a level that is known rel-
ative to all other datums of the tide gauge. Ideally the
switch, which can be mechanical, optical or acoustic,
should be sited outside the well in its own mini-stilling
well. The switch provides a correction for any datum
shift that previously would have been manually record-
ed by an operator. Although the switch will not work
correctly under all conditions, e.g. when high waves or
a seiche is present, there will usually be sufficient days
of calm to obtain an accurate datum check.
3.3 Pressure Gauges
Instruments that measure subsurface pressure instead
of sea level directly have found widespread use. A
knowledge of seawater density and gravitational accel-
eration is required to make the conversion from pres-
sure to sea level, but in spite of this, the instruments
have many practical advantages as sea level recorders.
The most commonly used types are the pneumatic
bubbler gauges and pressure sensor gauges in which
sensors are mounted directly in the sea. The two types
have much in common and a choice of which type is
suitable is usually based on practical considerations at
a proposed site.
3.3.1 Pneumatic Bubbler Gauges
The pneumatic bubbler tide gauge has been success-
fully used worldwide for several decades. It replaced
many of the float-operated harbour gauges as the pri-
mary standard for sea level measurement in countries
such as the United States and the United Kingdom,
although in the USA they have since been superseded
by acoustic gauges (section 3.4). The UK still operates
its National Tide Gauge Network (44 stations) based on
this technology. It has been shown to be robust, both
in terms of accuracy and datum stability. It has demon-
strated its value in situations where there are no vertical
structures on which to attach the equipment, e.g. on
coral atolls (Pugh, 1978), as the part of the equipment
installed in the sea and on land can be several hundred
metres apart, which is not the case with many other
types of instrument.
Figure 3.2 shows the basic essentials of a bubbler sys-
tem. Air is passed at a metered rate along a small-bore
tube to a pressure point fixed underwater well below
the lowest expected sea level. The pressure point
normally takes the form of a short vertical cylinder
with a closed top face and open at the bottom. A
small ‘bleed hole’ is drilled about half way down its
length and metered air is entered through a connec-
tion on the top surface. As air from the tube enters
the pressure point it becomes compressed and pushes
the water down inside the chamber until the level
of the bleed hole is reached at which time the air
bubbles out through the hole and back to the surface.
Provided that the rate of air flow is low and the air
supply tube is not unduly long, the pressure of air in
the system will equal that of the pressure due to the
depth of the sea water above the bleed hole cou-
pled with atmospheric pressure. A pressure-recording
Figure 3.2 Bubbler pressure gauge.
Sea Level Measurement and Interpretation
13
14
Sea Level Measurement and Interpretation
IOC Manuals and Guides No 14 vol IV
instrument connected into this supply tube at the
landward end records the changes in water level as
changing pressures, according to the law:
h=(p-pa)/(ρg)
where h = height of sea level above the bleed hole
p = measured pressure
pa = atmospheric pressure
ρ = seawater density
g = gravitational acceleration
Most pneumatic instruments use a pressure sensor
as part of the recording equipment to monitor the
changes in pressure and hence sea level. It is common
to use a sensor operating in the differential mode, sen-
sors being so constructed that the system pressure is
opposed by atmospheric pressure. Hence, the resultant
pressure experienced by the sensor becomes (p–pa),
making the measured pressure directly proportional to
the required sea level.
A knowledge of the seawater density (ρ) is impor-
tant. This is normally obtained from separate water
sampling, and, where the water is well mixed, can be
considered constant. In estuarine locations, the den-
sity may change during a tidal cycle or seasonally, and
density corrections will have to be included in the data
processing.
Several other effects on the measured pressure have
to be considered. These include a ‘static’ effect, which
is a function of the height of the gauge above sea
level, and a ‘dynamic’ effect, which results from the
dynamics of gas flow. The latter can be calculated in
terms of tube length and radius and the minimum
air-flow consistent with preventing water from enter-
ing the system (Pugh, 1972). The effect of waves on
the system is to introduce a positive bias during storm
conditions (i.e. sea level is measured too high). These
effects can perturb the sea level measurements at the
sub-centimetre level during average conditions, but
measurements may be incorrect by several centimetres
under extreme waves.
In common with all pressure measuring systems, there
is a need to establish a datum for the observed time
series. This can be achieved in several ways: (a) from
a knowledge of the exact depth of the pressure point
bleed hole during installation, combined with accurate
calibration of the pressure transducer; (b) using datum
level switches similar to those described above for
stilling wells which trigger at a known sea level; (c) by
having a parallel system (called a ‘B’ gauge; section
3.3.4) with a second and more accessible pressure
point fixed near mean sea level. Comparison of the dif-
ferences between the two bubbling systems when both
are submerged gives an accurate measure of the datum;
method ‘c’ is the most accurate.
Air is normally supplied to a bubbler from a compressor
to afford continuous operation of the installation. In the
event of electrical supply failure, a reserve air capacity
capable of sustaining the system for at least several days
is necessary. For sustained operation under fault condi-
tions, an alternative low power backup system in the
form of a pressure transducer mounted directly in the
sea is necessary. Transducers, compressors, data loggers
etc. can be purchased from the major gauge manufac-
turers with ready-to-go packages. An all-bubbler system
has an advantage that most components are under-
water, and that all components are both robust and, if
damaged, relatively inexpensive to replace.
3.3.2 Pressure Sensor Gauges
Pressure sensors can be fixed directly in the sea to
monitor sub-surface pressure in a similar fashion to the
bubbler gauge. The sensor is connected by a cable that
carries power and signal lines to an onshore control
and logging unit. In the sea, the active sensor is usually
contained within a copper or titanium housing with
the cable entering through a watertight gland. Material
used for the housing is chosen to limit marine growth.
The assembly is contained in an outer protective tube
to provide a stable fixation to a sea wall or rock out-
crop. Where this is not possible, the pressure sensor
may be placed securely on the sea bed, but this method
has disadvantages, as deployment and maintenance
usually require a diving team.
Pressure-based instruments can be operated from bat-
teries for periods of a year or more, as they consume
a very small amount of power. This can be advanta-
geous even where electrical supplies are available but
subject to long periods of failure. Therefore, they have
been used extensively in remote areas, such as oceanic
islands, where access is limited. In polar regions, they
offer the best alternative if the area is ice covered or if
the gauge is to be left unattended for long periods. The
main disadvantage is the lack of a fixed datum level,
which has to be found by alternative means.
Pressure sensors are available in two varieties that
provide either an absolute or differential signal. If an
absolute transducer is employed, the sensor provides a
measurement of the total pressure including sea level
and atmosphere. Therefore, a separate barometer is
required usually in the form of an identical transducer
open to the atmosphere. Both sensors are synchronized
to the same clock so they can readily be subtracted to
yield sea level (with subsequent correction for density
and acceleration due to gravity). Differential pressure
transducers have a vented cable in which the refer-
ence side of the transducer is open to the atmosphere.
Vented systems are known to suffer from occasional
blockage and are used less frequently in hazardous
environments. In addition, a record of barometric pres-
sure is valuable for oceanographic studies, so the two-
transducer option is most frequently employed.
Sea Level Measurement and Interpretation
15
IOC Manuals and Guides No 14 vol IV
Relatively inexpensive pressure sensors use strain gauge
or ceramic technology in which changes in water pres-
sure cause changes in resistance or capacitance in the
pressure element. The most accurate, but expensive,
sensors use a quartz element, the resonant frequency
of which varies with the strain applied to it. The result-
ing signal, which is normally a frequency proportional
to the applied pressure, is carried down the signal
cable to the control electronics where it is converted
into physical units and can be displayed and stored by
a data logger.
All pressure transducers are sensitive to temperature. Some
have an in-built temperature sensor to allow compensa-
tion of the pressure signal. If this is not the case, then it
is important that temperature is monitored independently
and used as a correction. In general, sea temperature
varies much less than atmospheric temperature and com-
pensation by either of the above methods is effective.
Users with access to a test facility can also subject the
instruments to a range of temperatures and pressures to
ensure that calibration values are correct. Experience has
shown that the calibration coefficients supplied by leading
manufacturers are accurate and constant over periods of
several years. Drift in the various properties of pressure
sensors is confined to changes in its datum value (i.e.
there is usually no change in scale). However, even for a
high-quality low-pressure sensor suitable for coastal work,
instrumental drift can be an important issue (of the order
of 1 mm per year) which has to be addressed through
regular checks of some kind.
Single transducer systems can be deployed in environ-
mentally hostile areas where other forms of gauge will
not work. For example, they can be safely positioned
on the sea bed under the winter ice at polar sites with
the signal cable to the tide gauge hut on the shore
protected by a steel pipe. They can be operated at
sites with harsh weather conditions where the exposed
structures of a stilling well or acoustic gauge may
be subject to extreme forces of winds and waves. In
tropical locations, where equipment may be prone to
mechanical damage by falling trees etc., single trans-
ducer systems can be deployed safely below the sea
surface. Even in locations with excessive marine growth
or silt deposits, pressure systems appear to work cor-
rectly for long periods of time.
Pressure sensors have a fast response time and have
been used to measure wave heights at periods of a few
seconds. In tide gauge applications, the signal is usually
averaged by the control electronics to a more relevant
period, such as 1, 6 or 15 minutes. This method of
averaging allows a great deal of flexibility, since the
sampling period can be easily altered to suit the appli-
cation. Changes can be made remotely if an installation
is connected by a telephone link or to a two-way com-
munication network.
As with the bubbler gauge, seawater density is needed
to convert measured pressures into heights. The com-
ments made in section 3.3.1 are equally valid.
3.3.3 The Datum of a Pressure System
The major problem with a single pressure transducer
is establishing a datum for its measurements. A good
approximation can be obtained with differential trans-
ducers by careful calibration within a test facility. It
is less accurate with absolute sensors because atmo-
spheric pressure introduces an offset that may prevent
a sufficiently low pressure being reached during the
calibration. In general, other means of fixing the datum
are preferred.
A method frequently adopted is to make visual mea-
surements against a tide staff over a period of one day
and repeat this at regular intervals. Individual measure-
ments should be accurate to 2–3 cm and on average
Figure 3.3 Pressure gauge.
(a) The pressure sensor is mounted directly in the sea.
(b) In this case, it is fastened to a pier in Port Stanley
harbour.
a
b
16
Sea Level Measurement and Interpretation
IOC Manuals and Guides No 14 vol IV
should fix the datum to approximately centimetre accu-
racy. However it is tedious and can only be carried out
infrequently in remote areas.
3.3.4 Multiple Pressure Transducer Systems
(‘B’ gauges)
A method was developed at POL in the early 1990s
for precise datum control of sea level records from
pressure tide gauges. An additional pressure point
was located at approximately mean sea level and
fixed relative to the contact point of the gauge. The
juncture at which the tide fell below this second
sensor could be used to fix the datum of the record
from the principal sensor. The technique was found
to be extremely reliable and accurate and now forms
the basis of gauges, called ‘B’ gauges, in POLs South
Atlantic and Antarctic networks (Spencer et al.,
1993). The principle of the technique was described
in detail in Volume 2 of the Manual (IOC, 1994) and
in the scientific literature (Woodworth et al., 1996).
At the time of writing, it is not possible to purchase
a ‘B’ gauge although expressions of interest in their
manufacture have been obtained from major sup-
pliers.
A schematic ‘B’ gauge setup is shown in Figure 3.4,
with an absolute pressure sensor in the water (‘C’)
and another in the atmosphere (‘A’). Paroscientific
digiquartz sensors are employed throughout, although
less expensive sensors should work reasonably well
and are being investigated. The difference C–A gives
sea level, after corrections for seawater density and
acceleration due to gravity are applied. A third sensor
is placed at ‘Datum B’ which is near mean sea level.
The height of ‘Datum B’ has to be known accurately
relative to the contact point of the installation and to
the local land levelling network. The difference B–A is
again a sea level height, but only when the sea level
is above ‘Datum B’. The top part of this record can be
fitted to the equivalent part of the record from the
principal sensor to transpose the known datum to the
full sea level record. It is important that all sensors are
driven from the same control and logging system to
maintain synchrony. Sampling the data at 15-minute
intervals or less is preferred for the identification of
the inflexion points, i.e. the time at which the sea level
falls (or rises) below (or above) ‘Datum-B’.
The essential feature is that, while any pressure mea-
sured by a sensor at B will contain an offset, and per-
haps a drift, the vertical height of its effective pressure
point can be positioned at ‘Datum B’ very accurately.
So, although it is not known what it is measuring to
within perhaps a few hectopascals (centimetres), it is
known where it is measuring with millimetric precision.
The flat part of B–A and its inflexion points provide an
extremely precisely defined shape which is immune to
any problems with datum offsets and low-frequency
instrumental drifts. Experience with several instruments
at different sites suggests that datums can be fixed to
within a few millimetres by this technique.
To work properly, the method needs a sizable tidal range,
so that B will spend half its time in water and half in air.
It will not work in lakes or microtidal areas, but most
coastal and many island sites have usable tidal ranges,
even if only at spring tides. In the presence of waves,
the flat portion of the ‘B’ gauge is reduced in length
and may be unuseable under large wave conditions.
However, there are is always a sufficient number of calm
days during which the technique can be applied.
In practice, the two pressure sensors in the sea are co-
located near the base of the installation with a rigid
tube connecting the ‘B’ gauge to its appropriate datum
point. This avoids the ‘B’ sensor being subject to atmo-
spheric temperature variations that are more severe
than those of the sea. The barometric sensor may also
be installed at the same position with a tube open to
atmosphere. Alternatively it may be installed as part
of the data logger in the tide gauge hut. The method
does not require the actual installed height of C or A to
be known. Where it is difficult to install a fixed gauge
C below the water, because of shallow gradients per-
haps, then a pop-up or bottom-mounted gauge could
equally well be used.
3.3.5 Pressure Transducers in Stilling Wells
A variant of the ‘B’ gauge method described above is
to install an absolute pressure sensor below low water
in a stilling well that has been used hitherto in a float
system. This transducer will be functionally the same as
sensor ‘C’ and will be complemented by a transducer
‘A that records atmospheric pressure, as described
above. Alternatively, a ‘differential’ sensor could be
used. Instead of a third sensor employed in the ‘B’
gauge, datum control for the C–A pressure-difference
time-series is provided by means of regular, prefer-
ably daily, electronic datum probe checks of the level
in the well relative to the tide gauge CP and TGBM.
Comparison of the values of C–A, corrected for density
and acceleration due to gravity, with the well sound-
ings, provides an ongoing datum for the time-series
which can accommodate transducer drift and varia-
tions in the properties of the sea water.
This method has many of the advantages of pressure
systems and of electronic datum probes, combined
with the recognized disadvantages inherent in the
use of stilling wells (Lennon, 1971). It may be a
preferred option if measurements are required from
a well that has produced long-term measurements
from a float gauge.
3.3.6 Bottom-mounted Pressure Gauges
Bottom pressure gauges rest on the sea bed and
record pressure at intervals over periods of a year or
Sea Level Measurement and Interpretation
17
IOC Manuals and Guides No 14 vol IV
more. They are self-contained instruments powered
by batteries. They have little application to the long-
term measurement of coastal sea level but have been
used extensively to obtain initial tidal knowledge of
an area where a coastal gauge is planned. Their main
problem in the GLOSS context is the lack of a datum.
They have principally proved their value offshore and
in the deep ocean (Spencer and Vassie, 1997).
3.4 Acoustic Tide Gauges
A number of acoustic tide gauges have been devel-
oped which depend on measuring the travel time
of acoustic pulses reflected vertically from the sea
surface. This type of measurement can theoretically
be made in the open with the acoustic transducer
mounted vertically above the sea surface, but in cer-
tain conditions the reflected signals may be lost. To
ensure continuous and reliable operation the sensor
is located inside a tube that provides some degree
of surface stilling and protects the equipment; some
sensors even constrain the acoustic pulses within a
narrow vertical tube, which is contained inside the
previous one. This outer tube does not completely
filter out wave action but, by averaging a number of
measurements, the desired filtering is achieved.
The velocity of sound in air varies significantly with
temperature and humidity (about 0.17%/°C) and
some form of compensation is necessary to obtain
sufficient accuracy. The simplest method is to mea-
sure the air temperature continuously at a point in
the air column and use this to calculate the sound
velocity. To account for temperature gradients in the
air column, temperature sensors may be required at
a number of different levels.
A more accurate method of compensation is by use
of an acoustic reflector at a fixed level in the air col-
umn. By relating the reflection from the sea surface
to that from the fixed reflector, direct compensation
for variation in sound velocity between the acoustic
transducer and the fixed reflector can be achieved.
However this still does not account for any variation
in sound velocity between the fixed reflector and
the sea surface. To achieve full compensation would
require, in principle, a number of fixed reflectors
covering the full tidal range, but none of the known
acoustic sensors has this possibility.
3.4.1 Acoustic Gauges with Sounding Tubes
The National Oceanic and Atmospheric Administration
(NOAA), National Ocean Service (NOS) in the USA,
initiated over a decade ago a multi-year implementa-
tion of a Next-Generation Water Level Measurement
System (NGWLMS), both within the US national
tide gauge network and at selected sites around
Figure 3.4 (a,b) Schematics of operation of a ‘B’ gauge.
b
a
IOC Manuals and Guides No 14 vol IV
the world (Gill et al., 1993). These systems were
operated alongside existing (float or bubbler) tide
gauges at many stations for a minimum period of
one year to provide datum ties and data continuity.
Dual systems were maintained at a few stations for
several years to provide a long-term comparison.
Tide gauges using the same technology have been
deployed in a number of other countries, such as
Australia, where they are known as SEAFRAME sys-
tems (Lennon et al., 1993).
The NGWLMS tide gauge uses an Aquatrak water
level sensor developed by Bartex Inc. and acquired by
Aquatrak Corporation, together with a Sutron data-
processing and transmission system. The Aquatrak
sensor sends a shock wave of acoustic energy down
a 1/2-inch-diameter PVC sounding tube and mea-
sures the travel time for the reflected signals from
a calibration reference point and from the water
surface. Two temperature sensors give an indica-
tion of temperature gradients down the tube. The
calibration reference allows the controller to adjust
the measurements for variations in sound velocity
due to changes in temperature and humidity. The
sensor controller performs the necessary calculations
to determine the distance to the water surface. The
sounding tube is mounted inside a 6-inch-diam-
eter PVC protective well which has a symmetrical
2-inch-diameter double cone orifice to provide some
degree of stilling. The protective well is more open
to the local dynamics than the traditional stilling
well and does not filter waves entirely. In areas of
high-velocity tidal currents and high-energy sea swell
and waves, parallel plates are mounted below the
orifice to reduce the pull-down effects (Shih and Baer,
1991). Figure 3.5 is a schematic of a typical NGWLMS
installation. To obtain the best accuracy, the acoustic
sensor is calibrated by reference to a stainless steel
tube of certified length, from which the zero offset is
determined.
The NGWLMS gauges have the capability of handling
up to 11 different ancillary oceanographic and mete-
orological sensors. The field units are programmed to
take measurements at 6-minute intervals with each
measurement consisting of 181 one-second-interval
water level samples centred on each tenth of an
hour. Software in the instrument rejects outliers etc.
which can occur as a result of spurious reflections.
Measurements have a typical resolution of 3 mm.
The instrument contains the necessary hardware for
telephone and satellite communications.
Papers by Gill et al. (1993) describe the operation-
al performance of the NGWLMS instrumentation.
Lennon et al. (1993) and Vassie et al. (1993) present
comparisons between NGWLMS and conventional
stilling well or bubbler systems in Australia and the
UK. Most comparisons show small differences, of the
order of a few millimetres, for the various tidal and
datum parameters, which are generally within the
uncertainty of the instrumentation. Such differences
are very small when compared to typical tidal ranges
and even seasonal and interannual sea level varia-
tions. NGWLMS systems are considered sufficiently
accurate for mean sea level studies.
Figure 3.5 NGWLMS tide gauge.
18
Sea Level Measurement and Interpretation
IOC Manuals and Guides No 14 vol IV
A modern version of the NGWLMS is called a Sea
Ranger which is claimed to have a number of advan-
tages over the earlier technology including self cali-
bration (IOC, 2004)
3.4.2 Acoustic Gauges without Sounding Tubes
Several acoustic instruments have been produced
that are operated without a sounding tube, normally
located inside an existing stilling well or inside a
plastic tube some 25 cm in diameter. Some of them
may operate in the open air, but are not normally
employed for high-quality sea level measurements
(see Table 3.1 in section 3.6). These acoustic instru-
ments operate at a frequency of 40–50 kHz and have
a relatively narrow beam width of 5°. Their measure-
ment range is approximately 15 m and an overall
accuracy of 0.05% is claimed by the manufacturers
(see websites below).
Contradictory experiences can be found with this type
of acoustic sensor, from some problems in achieving
the stated accuracy under all environmental conditions
(e.g. see presentation by Ruth Farre, in IOC, 2003), to
the high-quality and continuous operation of 15 tide
gauges in the REDMAR network (Spain), most of them
installed in 1992 and still in operation (e.g. see pre-
sentation by Begoña Pérez in IOC, 2003).
A crucial aspect of this type of sensor is the depen-
dence of the velocity of sound on the environmental
conditions, such as the air temperature. On the other
hand, tubes tend to increase the temperature-gradi-
ent between the instrument and the sea surface
unless special precautions are taken to ensure that
the air is well mixed in the tube. A complementary
and necessary method is to compensate for sound
velocity variations using a reflector mounted at a
suitable distance below the transmitter, as is the case
for the SRD gauges employed in the REDMAR net-
work. A careful design of the installation, avoiding
different ambient conditions along the tube and fol-
lowing the maker’s requirements about the minimum
distance to the water surface, become crucial for the
final accuracy of the data.
The performance of one of these sensors (SRD) over
an existing stilling well inside a hut or small building
in Santander (Spain), has been incredibly good (nearly
perfect and continuous during 15 years). The condi-
tions of this installation are probably perfect, perhaps
because the temperature inside the building is rather
homogeneous. Data from this acoustic sensor have in
fact helped to correct malfunctions of the float gauge
that operates inside the same stilling well.
Studies of mean sea levels from 12 years of data in
Spain, comparing this type of acoustic sensor (SRD)
with the traditional float gauges, has shown their
high quality and has even helped to identify refer-
ence jumps in the older float gauges. This is, again,
a contradictory experience to the one in South Africa
(see article by Farre in Appendix V of this volume).
Nevertheless, it seems that radar gauges will replace
this type of acoustic sensor everywhere, in the near
future.
3.5 Radar Gauges
Radar tide gauges have become available during the
last few years from several manufacturers. Although this
technology is relatively new, radar gauges are being pur-
chased and installed by a number of agencies as replace-
ments for older instruments or for completely new
networks. The reason is that they are as easy to operate
and maintain as acoustic sensors, without their main
disadvantage: their high dependence on the air temper-
ature. Radar gauges have a relatively low cost and the
engineering work necessary to install them is relatively
simple compared to other systems. The instruments are
supplied with the necessary hardware and software to
convert the radar measurements into a sea-level height.
In addition, the output signals are often compatible with
existing data loggers or can be interfaced to a communi-
cation network. Like many modern systems they can be
set up using a portable computer.
The active part of the gauge is located above the
water surface and measures the distance from this
point to the air–sea interface. A diagram is given in
Figure 3.6. The gauge has to be mounted in such a
way that there are no restrictions or reflectors in the
path of the radar beam, between the gauge mount-
ing and the sea surface. It has to be positioned above
the highest expected sea level and preferably above
the highest expected wave height, so as to prevent
physical damage.
It has many advantages over traditional systems in
that it makes a direct measurement of sea level.
The effects of density and temperature variations,
even in the atmosphere, are unimportant. The main
constraint is that the power consumption may be
relatively large in radar systems if used on a continu-
ous basis in a rapid sampling mode. Averages are
typically taken over periods of minutes. This may
limit its use in some applications (e.g. tsunami warn-
ing) where observations are required on a continuous
high-frequency (e.g. 15-second) basis. In such areas,
pressure gauges may be more appropriate, although
work and research is still being done concerning this
particular application.
Radar gauges fall into two categories. Those that
transmit a continuous frequency and use the phase
shift between transmitted and received signal to deter-
mine sea level height (frequency-modulated continuous
Sea Level Measurement and Interpretation
19
IOC Manuals and Guides No 14 vol IV
20
Sea Level Measurement and Interpretation
Figure 3.6 Radar tide gauges.
(a) diagram comparing a radar and a bubbler gauge (Woodworth and Smith, 2003);
(b) an OTT Kalesto test installation at Liverpool.
b
a
IOC Manuals and Guides No 14 vol IV
Sea Level Measurement and Interpretation
waves: FMCW). The OTT Kalesto, Miros and Radac instru-
ments use this method. The VEGA and SEBA systems use
pulsed transmissions and time-of-flight measurement.
All these gauges have undergone initial tests and inter-
comparisons by various agencies in different countries.
Details of these tests can be found in IOC Workshop
Report No 193. Details of the individual instruments can
be found on the websites shown below.
In principle, the instruments are self calibrating,
as far as a datum value is concerned. However, to
provide confidence that the datum remains constant
over long time periods, alternative means are being
investigated. These take the form of a reflector that
can be placed in the radar beam at appropriate
intervals. The reflector is placed at a known distance
below the contact point of the installation for a short
period. Over a period of a year or more the datum
value can be verified and used to adjust the measure-
ments, if necessary.
Initial indications that these instruments can pro-
vide acceptable measurements for the purposes of
GLOSS are promising. As with all tide gauges, practi-
cal considerations related to a particular application
often dominate other considerations. For example,
they have very limited application in polar regions.
They have not yet been used in extremely hostile
environments, for example on remote islands, where
extreme waves may overtop the gauge by several
metres. However, for a normal application in which a
stilling well or bubbler gauge is presently in use, they
appear to operate satisfactorily.
3.6 Summary of the Merits of Different
Technologies
In this section, we summarize the relative merits
of different tide gauge technologies for scientific
research, operational oceanography and for localized
practical purposes, such as harbour operations.
The GLOSS programme has scientific research as
its raison d’être, although it is intended that the
development of the GLOSS networks should serve to
improve standards overall (see IOC, 1997). We can
use the designation ‘GLOSS’ to indicate the most
demanding requirement of scientific-quality perfor-
mance of a gauge (Appendix I).
There are also sea level requirements from opera-
tional users of oceanographic data in such topics
as marine infrastructure (e.g. offshore industry,
transport, coastal recreation) and coastal defences
(e.g. flood protection from surges, and studies of
coastal erosion or sea level rise impacts). Many of
these applications overlap GLOSS interests, the
study of secular changes in sea level being an obvi-
ous example. However, the particular applications
will vary from country to country. Therefore, such
gauges will be capable of deployment for extended
periods, but perhaps not to the same high standards
as those intended for GLOSS, and will be affordable
for use in larger numbers than for GLOSS, especially
by developing countries.
Finally, there will be applications which require a
cheap instrument capable of showing the state of
the tide at any moment but certainly not accurate
enough for GLOSS.
Table 3.1 presents a summary of the main conclu-
sions on the relative merits of each gauge technol-
ogy based on the previous sections of this Manual.
The Table also includes an estimate of the likely cost
of a basic system with gauge, data transmission (e.g.
modem) and meteorological package, although this
is an extremely difficult item to quote given the large
number of manufacturers, monetary exchange rates
etc. For example, the cost of a pressure transducer
will vary by a factor of 3 depending on the qual-
ity. With these reservations in mind, Cost Band 3
has been set as the highest cost, which might be
12,000–20,000 US$ (at the time of writing and
within a large band, say 30%); Band 2 might be
8,000–12,000 US$; and Band 1, 5,000–8,000 US$.
However, in our experience, the real costs of any tide
gauge station are those of installation (e.g. some kind
of engineering support will be needed for installation
of a stilling well, acoustic sounding tube gauge, or
‘B’ gauge; diver support will be needed for pressure
gauge installations etc.), ongoing maintenance and
data analysis (with implications for staff resources).
Anyone planning a gauge installation has, therefore,
to take into account all the local costs as well as the
up-front costs of gauge hardware. Agencies partici-
pating in GLOSS which require the input of expertise
may wish to explore the possibilities of collaboration
with other GLOSS participants.
Our recommendations are:
If one is planning a new GLOSS tide gauge station
in a mid- or low-latitude location, one should prob-
ably opt for:
- an acoustic gauge with sounding tube or
- a radar tide gauge or
- a ‘B’ pressure gauge
If low tidal range or other factors preclude the use of
a ‘B’ gauge, then a single transducer pressure gauge,
a bubbler pressure gauge or a pressure transducer in
a stilling well would be options. In addition, in most
cases, the main tide gauge should be accompanied
by a pressure sensor installed in the sea and capable
21
of sampling (or averaging) at the high frequencies
(once every 15 seconds or 1 minute) required for
tsunami-warning purposes, although, in the case of
some of the FMCW radars, this high frequency is also
possible at the main tide gauge. The advantages of
running two or more sea level sensors in parallel are
also: (i) improved data recovery; (ii) improved data
quality assessments by comparing redundant records;
and (iii) ability to optimize sampling strategy for dif-
ferent processes. Real time data from GLOSS sites
need to be made available in real time to the GLOSS
Real Time Centre at the University of Hawaii.
If one is planning a new GLOSS station at a higher-
latitude site which has sea ice cover for part of the
year, one should probably opt for:
- an absolute transducer pressure gauge and
accompanying barometer
- a bubbler pressure gauge
Although it is true to say that float gauges have been
operated in Antarctica, and the longest tide gauge
record in Antarctica is from the Faraday/Vernadsky
float gauge in a heated stilling well, we do not rec-
ommend their future use in ice areas. Bubblers and
acoustic gauges have also been tried in Antarctica,
but our recommendation is to use the absolute
transducer system if possible, with summer-time
datum control using either tide poles or ‘temporary
B gauges’.
If one is planning to upgrade an existing float
gauge GLOSS installation at most places, then we
would recommend the following approach.
First, consider simply upgrading the existing system
to electronic data acquisition and transmission.
(Charts must be abandoned as the main recording
system, although they may be retained to provide
ancillary information.) This will provide instructive
experience with real time data.
Second, consider the use of a pressure gauge system
within the stilling well.
Third, consider installation of a new station alongside
the old one (either acoustic sounding tube, radar or
‘B’ gauge etc. as described above) but keep both
of them operational for inter-comparison of their
data for an extended period (possibly as much as a
decade). Probably, the installation of a radar gauge
over the well would be the easier option.
If one is planning to use relatively cheap gauges
(but perhaps many units) for ‘coastal’ purposes,
then we would recommend:
- single-transducer pressure gauges (either absolute
or differential)
- if existing (or easily installed) stilling wells are avail-
able, fairly inexpensive shaft encoder float systems
are now on the market
- if stilling wells are available, pressure transducers in
the well
- cheap radar gauges (normally pulse technology),
both in the case of existing stilling wells or in the
open air.
If one required a ‘cheap and cheerful’ gauge for
‘practical’ harbour operations or approximate flood
level estimates, then we would recommend:
- differential transducer pressure gauges
- cheaper radar gauges.
Such installations would not need the ancillary
parameters needed for GLOSS (Appendix I, point
vi), but they may require such components as ‘user
friendly’ real-time displays.
Whichever type of gauge is selected, advice will
be needed, and groups such as the PSMSL and the
GLOSS Technical Committee will be pleased to help.
Something important to take into account is the
correct installation and a good knowledge, by the
maintenance technicians, of the problems that any
particular sensor can present and how to avoid them
with adequate operation.
22
Sea Level Measurement and Interpretation
IOC Manuals and Guides No 14 vol IV
Acoustic Gauges with Sounding Tubes
Equipment Complete ready-to-go package (acoustic transducer, sounding tube, met. package, ancillary sub-sea pressure sensor, modem
and satellite communications) can be purchased from several manufacturers.
Operation The device measures the time of flight of an acoustic pulse along a vertical sounding tube. Transit time is compared to a reflec-
tion from a calibration hole at a known distance from the acoustic transducer to obtain sea level height. Sound velocity is tem-
perature-sensitive, therefore temperature is measured in the support tube by two thermistors mounted some distance apart.
Installation
Requirements
The length of the sounding tube is altered to suit the application. The sounding tube is fastened to an Aquatrak acoustic trans-
ducer and inserted in an outer protective stilling tube. The full assembly is then fixed to a vertical sea wall.
Mains power or batteries and solar panels.
An auxiliary pressure sensor is normally fitted as part of the installation. This is a vented cable-type transducer.
Location Requires a sea wall or vertical structure for installation.
Calibration Calibration is performed during manufacture prior to delivery.
Accuracy Better than 1 cm of sea level.
Cost Bands 2–3.
Record of Use Used extensively in the United States as a replacement for bubbler systems in the National Tide Gauge Network. Used in a large
part of the Australian network and at island sites in the Pacific Ocean.
Comments For best accuracy, a calibration facility is required. In areas of large tidal range, a long sounding tube is needed which may result
in magnified temperature and/or temperature-gradient effects.
Table 3.1 Merits and drawbacks of each tide gauge technology.
Acoustic Gauges in the Open Air
Equipment A ready-to-go package can be purchased from several manufacturers.
Operation The device measures the time of flight of an acoustic pulse from a transducer to the sea surface. The time is converted to
a sea level height using a known value of the velocity of sound in air. Sound velocity is temperature sensitive, which can
cause significant errors if it is not taken into account.
Installation
Requirements
The installation requirements are relatively simple. The device requires a rigid structure to position it above the sea with
sufficient clearance to avoid spurious reflections from any adjacent structures. As with many tide gauges, all ancillary
equipment (data logger, modem, satellite communications, battery backup), needs to be housed in an adjacent building.
Location Requires a site with vertical clearance sufficient to mount the device clear of the maximum sea surface, including wave action.
Calibration Accurate calibration is one difficulty because of the sensitivity of sound velocity to air temperature. A calibration can be
achieved by placing a reflective bar at a known position in the acoustic beam. The results are usually inferior to the sound-
ing tube method.
Accuracy Greater than 1 cm of sea level.
Cost Band 1.
Record of Use Without independent confirmation of the accuracy and datum, these gauges are less applicable for GLOSS purposes. They
have been used successfully on offshore rigs to record sea level over periods of several years.
Comments For best accuracy an independent calibration method is required. In areas of large tidal range, or where the transducer is
high above the sea surface, there are secondary effects caused by the acoustic beam sounding different surface areas of
the sea at the peak and trough of a wave.
Sea Level Measurement and Interpretation
23
IOC Manuals and Guides No 14 vol IV
IOC Manuals and Guides No 14 vol IV
24
Sea Level Measurement and Interpretation
Acoustic Gauges without Sounding Tube inside Protective Tube or Well
Equipment A ready-to-go package can be purchased from SRD that includes the sensor, the data logger and configuration unit, and
the communications system.
Operation The device measures the time of flight of an acoustic pulse from a transducer to the sea surface. A bar is fixed at a known
distance from the transducer, which is used for self-calibration and computation of the velocity of sound before each mea-
surement. The time is converted to a sea level height using the value of the velocity of sound in air previously computed
by means of the fixed bar.
Installation
Requirements
This type of acoustic sensor has proved to be accurate enough if placed over an existing well, or inside a protective PVC
tube of 300 mm diameter. The transducer must be located at a minimum distance of around 2–3 metres from the water
surface at any moment. As with many tide gauges, all ancillary equipment (data logger, modem, battery backup), needs
to be housed in an adjacent building.
Location Requires a sea wall or vertical structure for installation.
Calibration Calibration of the reference is performed during manufacture, prior to delivery.
The calibration of the velocity of sound is made by means of the reflective bar at a known position in the acoustic beam.
Accuracy 1 cm of sea level.
Cost Band 2.
Record of Use They have been used successfully in the REDMAR network, the Spanish Harbour Authority’s sea level network, for nearly
16 years. The long-term means seem to be as accurate or better than the standard float gauges operating in Spain.
Comments In areas of large tidal range a long protective tube is needed which may result in magnified temperature and/or tempera-
ture-gradient effects. Very sensitive to the careful design of the installation.
Single Transducer Pressure Gauges
Equipment Complete ready-to-go package (sub-sea pressure sensor, cabling and data logger) can be purchased from several manu-
facturers.
Operation Two different options are available: (a) an absolute pressure sensor measuring the total pressure due to sea level and atmo-
sphere; (b) a differential sensor which has a vented cable measuring pressure changes due to sea level alone. Conversion
of pressure to sea level height requires knowing seawater density. Generally, an average value can be used unless there
are significant seasonal or tidal variations. Pressure sensors are also temperature sensitive, but, since sea temperature varies
much less than atmospheric temperature, this normally has a small effect.
Sensors vary in cost by up to a factor of 20. Relatively inexpensive sensors use strain gauge technology. Top-of-the-range
sensors are constructed using quartz crystals. For the latter, the temperature sensitivity of low-pressure sensors is around
1 mm/°C. Instrumental drift of the same sensor is about 1 mm per year.
Many pressure sensors produce a frequency-modulated output. This can be counted (integrated) by relatively simple elec-
tronics to produce the required measurements. Resolution therefore depends on the integration period, which is typically
15 or 6 minutes, but can be as short as 1 minute and still provide sufficient accuracy. Some manufacturers provide equip-
ment that does not integrate over the full sampling interval, in order to conserve battery power.
Installation
Requirements
These devices can be used virtually anywhere, even on shelving beaches. They are normally mounted in an outer protec-
tive tube fastened to a sea wall but can be fixed directly on the sea bed and connected to the shore by armoured cable.
Pressure sensors require very little power and can be run for periods of 1–2 years on non-rechargeable batteries.
Location Pressure sensors can be used at virtually any site, even in hostile environments, such as the polar regions. Regions with large
variations in seawater density may cause significant errors.
Calibration Calibrations traceable to National Physical Laboratory (UK) standards can be obtained from pressure sensor manufacturers
and have been shown to remain stable over many years. However, drift in the datum value of a sensor may cause changes
to its ‘zero’ value. Re-calibration at intervals may be necessary. Alternatively, the difficulty of establishing a datum can be
rectified by using alternative means (e.g. from annual tide pole measurements). These have proved adequate, since the
drift is normally linear with time.
(Continued on next page)
IOC Manuals and Guides No 14 vol IV
Sea Level Measurement and Interpretation
Accuracy Resolution of a low pressure sensor is typically better than 1 mm of sea level. However, instrumental drift may degrade
this, so that the accuracy is approximately 1 cm of sea level.
Cost Varies by a large factor depending on type. Band 1–2.
Record of Use Used frequently as a temporary exploratory tide gauge. Extensively used at remote island sites and in hostile environments,
such as the Antarctic.
Comments Datum fixing is the major problem and other types of tide gauge are preferred for permanent installations.
25
Multiple Pressure Transducer Systems (B Gauges)
Equipment These instruments are used only by POL and were developed to produce a high precision tide gauge. They are constructed
in-house from commercially available components but cannot be obtained as a complete ready-to-go package. The
instrument requires three high quality pressure transducers which results in a relatively expensive system. A less expensive
construction is presently being considered.
Operation The instrument contains three pressure sensors which measure respectively a) atmospheric pressure b) Half-Tide pressure
and c) Full-Tide pressure. All three sensors are positioned in the sea with a rigid tube to the appropriate measuring point
above. Since the position of the top of the Half-Tide tube is known accurately this can be used to calibrate the datum
of the Full-Tide pressure. Data is fed by an armored cable to a data logger and control unity sited nearby. Most of the
comments relating to a single pressure sensor are applicable but drift in the pressure sensors is inconsequential to its
operational capability. Temperature compensation of the pressure sensors is obtained from components integrated into
the pressure sensors.
Installation
Requirements
The instrument is pre-assembled and requires fixing to a vertical sea wall or marine structure.
Mains power or batteries and solar panels.
Location Requires a sea wall or vertical structure for installation.
Calibration Manufacturers calibrations of the pressure sensors are sufficiently accurate. The Half-Tide point should be levelled to local
benchmarks.
Accuracy Precision and accuracy of a few millimetres has been achieved.
Cost Band 3.
Record of Use Used in the United Kingdom and extensively at remote island sites in the Atlantic as well as in the Antarctic.
Comments Extremely accurate system with automatic datum control and as a by-product air pressure, air temperature and sea
temperature are recorded.
For operational reasons the instrument will only work in a region where the tidal range is 1 metre or greater.
Pressure Transducers in Stilling Wells
Equipment Pressure transducers are often placed in stilling wells, where these are available. This provides a protected and secure
environment for the sensors and can augment measurements made by a float gauge. The comments above on pressure
sensors are equally valid for this type of installation.
Accuracy Approximately 1 cm of sea level. The absolute accuracy may be limited by the characteristics of the stilling well.
Cost Band 2.
Comments Problems associated with the use of stilling wells are well documented. (see Float Gauges).
Bubbler Pressure Gauges
Equipment Complete systems are available commercially, but considerable assembly work is required to construct an operational tide
gauge. The equipment comprises an air supply (normally from a compressor), a gas control system, a connecting pipe, a
sub-sea pressure outlet, a pressure transducer at the landward end and the various data logging and support electronics.
Operation The instrument supplies air from the high pressure supply at a reduced pressure and at a constant rate through the system.
The pressure required to bubble the air through the sub-sea outlet at this rate is a measure of the sea level above the
outlet. A differential pressure transducer vented to the atmosphere alleviates the need to measure atmospheric pressure
separately, thereby producing a pressure reading proportional to sea level height. The sub-sea outlet is open at the base,
has a large surface area relative to its volume and has a small exit port approximately half way from the base. This design
reduces the effect of wave action and provides a very stable datum.
Installation
Requirements
The outlet and part of the connecting tube are the only components in the sea. Such a configuration increases the reli-
ability of the system and makes replacement relatively simple. All other components of the system are housed nearby.
The system requires external power for continuous operation, and backup operation is relatively limited, owing to the
limited air supply.
Location Bubbler systems can be used at virtually any location, even on shelving beaches. Connecting tubes can be several hundred
metres in length. As with most pressure measuring systems, regions with large variations in seawater density may cause signifi-
cant errors.
Calibration Calibration is concerned with the pressure sensor accuracy and may need to be repeated at intervals. Calibrations sup-
plied by leading pressure transducer manufacturers are acceptable provided occasional means of fixing the datum value
are used.
Accuracy In general, an average accuracy of 1 cm of sea level is achievable, but this may degrade under large-wave conditions.
Cost Band 2.
Record of Use Used extensively in the United States and the United Kingdom for their national tide gauge networks.
Comments At a few locations, a secondary bubbler system has been installed at the mid-tide level as part of the United Kingdom
network. This can be used to fix the datum of measurements in the same fashion as the ‘B’ gauges discussed above.
26
Sea Level Measurement and Interpretation
IOC Manuals and Guides No 14 vol IV
Sea Level Measurement and Interpretation
27
Float Gauges
Equipment A float in a stilling well is the tried and tested method of measuring sea level directly, rather than through an indirect
parameter such as pressure or sound.
Operation A stilling well filters out wave activity at periods shorter than the maximum tidal period, which might be 2 hours in shal-
low water regions. In modern installations the float drives a shaft encoder or potentiometer the output of which is fed to
an electronic data logger. In the past, chart recorders were extensively used, but are no longer acceptable as the principal
data-recording method, as they contain many sources of inaccuracy and require labour-intensive digitization.
Installation
Requirements
Stilling well installations require heavy civil engineering work in areas of large tidal range. Many stilling wells exist through-
out the world, as they are of robust construction, but new installations are less common, owing to the engineering cost.
A suitable building is required above the well to protect the well and its associated measuring equipment.
Location Requires a sea wall or vertical structure for installation.
Calibration Stilling wells can suffer from several defects which have been well documented. For example, density variations between
the inside and outside of the well in regions of stratification cause errors. Siltation and marine growth can cause changes
to the dynamic response of the well. Absolute calibration usually involves dipping the well with a calibrated tape at peri-
odic intervals.
Accuracy Approximately 1 cm of sea level.
Cost Band 1–2.
Record of Use Used extensively in the United States and the United Kingdom for their national tide gauge networks.
Comments Stilling wells have been used worldwide for a considerable period and are still used, both as the primary system and as
backup system for a modern tide gauge.
Radar Gauges
Equipment Radar tide gauges have so far been little used for GLOSS purposes, because it is a very recent technology. However, they
offer a complete ready-to-go package which is relatively easy to install above the sea surface and seem to have advantages
with respect the acoustic sensors.
Operation Radar gauges measure the time of flight either from a pulsed radar or the phase change between a transmitted and
received carrier wave, to determine the distance to the sea surface. They are much less affected by air temperature than
acoustic gauges.
Installation
Requirements
The installation requirements are relatively simple. The device requires a rigid structure to position it above the sea with
sufficient clearance to avoid spurious reflections from any adjacent structures. As with many tide gauges, all ancillary
equipment (data logger, modem, satellite communications, battery backup), needs to be housed in an adjacent building.
No need of a protective tube.
Location Requires a site with vertical clearance sufficient to mount the device clear of the maximum sea surface, including wave action.
Calibration In essence the device is self calibrating. However, for GLOSS purposes, a reflective target is mounted at a known distance
below the radar transmitter.
Accuracy Accuracy is expected to be approximately 1 cm of sea level.
Cost Band 2–3.
Record of Use So far these gauges have been used for relatively short periods experimentally by Spain and the United Kingdom.
Comments Radar tide gauges may consume excessive power if used in a continuous mode. In burst mode, they provide sufficient
accuracy for measuring most tidal parameters, but their use in a rapid sampling mode may be limited by this, although
tests are being made in Spain for higher-frequency sampling.
IOC Manuals and Guides No 14 vol IV
4. Datum Control and Levelling
28
Sea Level Measurement and Interpretation
It should be clear that the measurements made by a
tide gauge provide the relative movement of the sea
level with respect to the land. Of course, neither land
nor sea levels are constant over long periods of time.
There are vertical movements of the land associated
with a range of natural processes, such as co-seismic
activity (earthquakes), in addition to glacial isostatic
adjustment (post-glacial rebound) and plate tecton-
ics and with a range of human activities (e.g. ground
water pumping). For a review of the geological signals
in tide gauge records, see Emery and Aubrey (1991).
Long-term changes in sea level relate to variations in
ocean currents, to changes in the volume of water in
the oceans and therefore to climate change. It is clear
that, to understand sea level changes properly, the dif-
ferent sea level and land signals have to be decoupled.
This is achieved by careful definition of the tide gauge
datums, by local levelling procedures, and by making
independent measurements of changes in the land
levels, using modern geodetic techniques. Such tech-
niques derive from the use of very high resolution GPS
receivers and absolute gravimeters.
4.1 Datums and Benchmarks
For sea level observations, a land benchmark is used
as the primary reference point. The benchmark is a
clearly marked point located on a stable surface, such
as exposed rock, a quay wall or a substantial building.
When a benchmark is on a horizontal surface, it nor-
mally takes the form of a round-headed brass bolt, the
highest point of the domed head being the reference
level (Figure 4.1). When on a vertical surface, it can
be in the form of a horizontal groove in the surface
or on a metal frame attached to the surface, having a
horizontal reference edge to which a measuring staff
support can be fixed.
It is poor practice to depend upon the stability of a
single benchmark. It is recommended that there be a
minimum of five within a few hundred metres, or at
most one kilometres, of the tide gauge. These should
be connected individually by high-precision levelling
and shown to maintain the same relative elevation
as time progresses. If no changes are observed over
long periods, it is safe to assume that the area of land
around the gauge is ‘stable’. The area could, of course,
exhibit vertical movement with respect to a much wider
area. This can be demonstrated by wide-area levelling
or from surveys using space geodetic techniques.
It is desirable, although not essential, that all benchmarks
be tied into a country’s national levelling network, and
periodically checked with respect to that network. The
benchmarks will then be given elevations referred to
the datum of the national network. However, national
levelling networks tend to be redefined at intervals. For
that reason, in sea level studies, it is best not to rely on
national levelling for any scientific purpose, although,
Figure 4.1 A brass bolt
benchmark at Newlyn, UK,
which functions as a reference
point for height measure-
ments in the UK and as the
TGBM of the Newlyn gauge.
IOC Manuals and Guides No 14 vol IV
Sea Level Measurement and Interpretation
of course, it may provide useful ancillary information. It
is important that the benchmarks be clearly identified,
by the inscription of a name or number. In addition,
they should be unambiguously documented in the tide
gauge metadata, with a description of the mark itself,
photographs, national grid reference and a local map.
4.1.1 Tide Gauge Benchmark (TGBM)
The tide gauge benchmark (TGBM) is chosen as
the main bench mark for the gauge from the set of
approximately five marks described above. The TGBM
is extremely important, since it serves as the datum to
which the values of sea level are referred. The choice
of TGBM is somewhat subjective; in principle, it should
be the ‘most stable’ or ‘most secure’ mark of the set,
although, if the area is largely stable, then the choice
should be fairly arbitrary. Often the nearest mark to the
gauge is chosen. Over a period of time it may be neces-
sary to redefine the TGBM, if the original is destroyed
as a result of local development. The benefit of having
a set of five local marks, regularly interconnected by
high-precision levelling, is that it allows a new TGBM
to be defined in terms of the old one, if circumstances
require it.
In some countries the historical practice has been
not to define one mark as the TGBM, but to use a
weighted average of several marks. For GLOSS, it is
recommended that the single, unique TGBM approach
be adopted as the standard.
4.1.2 GPS Benchmark (GPSBM)
The GPS benchmark (GPSBM) is another special mark
of the available set that is the reference mark for GPS
measurements near the gauge. In some busy ports, the
GPSBM may be several hundred metres from the TGBM
and the gauge. As with the other marks, it must be
connected by high-precision levelling to the TGBM at
regular intervals. (See section 4.4.1 for details on GPS
measurements at tide gauges).
4.1.3 Gauge Contact Point (CP)
The contact point (CP) of a tide gauge is a type of
‘benchmark’, or vertical reference mark, associated
with the gauge itself. After a geodetic connection has
been made between the TGBM and the CP, the gauge’s
sea level data can be expressed in terms of the TGBM
datum. The essential point to note is that the CP comes
with the gauge; if a different type of gauge is installed
at the site, it will have a different CP which will require
re-levelling to the TGBM.
For conventional float and stilling well gauges, the
CP is often located at the top of the well inside the
tide gauge hut. Sometimes, in older stations, the CP
is located in a most difficult and inaccessible location
for levelling purposes and new stations should take
care to provide ready access. For acoustic gauges with
sounding tubes, the CP is located at a point at the top
of the gauge on the container holding the acoustic
transducer. Similarly, for radar gauges, the CP will be
a mark on the transducer. For ‘B’ gauges, the CP will
be at the top of the vertical supporting tube which is
known relative to the ‘B’ datum level.
In the case of float gauges located in a tide gauge hut,
the CP should not be used as the TGBM itself, as it is
always possible for the building and the well to gradu-
ally settle over a long period. With a good set of local
benchmarks, this settling will be evident by check level-
ling between TGBM and CP.
4.1.4 Tide Gauge Zero (TGZ)
The tide gauge zero (TGZ) is the level for which the
gauge would record zero sea level. In practice, the sea
level may not fall to this level. In a conventional float
gauge arrangement, the TGZ can be related to the CP
after dipping checks in the well have been performed.
This is done using a calibrated tape set to zero at the
CP. Measurements are made by lowering the tape until
it reaches the water and an electrical circuit is com-
pleted. The level of sea water in the well can then be
related to the CP and to all other local datums.
4.1.5 Revised Local Reference (RLR) Datum
The revised local reference (RLR) datum at a gauge site
is a datum defined as a simple offset from the TGBM,
such that values of sea level expressed relative to the RLR
datum have numerical values around 7,000 mm. The
concept of the RLR datum was invented by the PSMSL
so that long time-series of sea level change at a site
could be constructed, even if parts of the time-series had
been collected using different gauges and different, but
geodetically connected, TGBMs. The approximate value
of 7,000 mm was chosen so that the computers of the
time (the late 1960s) would not have to store negative
numbers. The RLR datum is defined for each gauge site
separately and the RLR at one site cannot be related to
the RLR at any other site, without additional knowledge
of connections between TGBMs at the different sites.
When sea level data are contributed to the PSMSL, or
to a sea level centre, it is essential that full information
on the geodetic relationships between TGBM and TGZ
etc. accompany the data. Without this information, it
is impossible for the PSMSL to include such data in the
RLR data set.
4.1.6 National Levelling Network
Most countries have, during the last one hundred
years, implemented national levelling networks that are
defined usually in terms of mean sea level (MSL) at one
29
IOC Manuals and Guides No 14 vol IV
IOC Manuals and Guides No 14 vol IV
or more stations. Levelling connections within these
networks then allow the heights of objects (e.g. moun-
tains) to be related to MSL at the coast. For example,
the UK national levelling network expresses heights in
terms of ‘Ordnance Datum Newlyn’ (ODN), which was
the average level of the sea at Newlyn in southwest
England during 1915–21. ODN can be thought of as an
imaginary datum plane extending over a large area (i.e.
over the whole of Great Britain). The heights of bench
marks, for example, can be expressed in terms of ODN
as can, therefore, the Chart Datum at the port.
The concept of a national levelling network has under-
gone revolutionary change during the last decade, pri-
marily due to the advent of GPS. However, it was already
a defective concept from the point of view of sea level
studies, for several reasons. First, sea level has risen at
Newlyn since 1915, as it has done at many other places
around the world, so ODN no longer represents the
present average Newlyn levels. Second, the mean sea
surface around a coast is not ‘flat’, i.e. it does not follow
the geoid, but varies due to ocean currents, density dif-
ferences, meteorological effects etc. Consequently, MSL
was never a perfect choice for a national datum plane.
Third, rates of change of MSL are different at different
locations, thereby complicating the time-dependence
of the network. Fourth, all national levelling networks
(with the possible exception of that of The Netherlands,
Finland and Sweden) contain multi-decimetric errors due
to systematic, instrumental errors in the levelling. Fifth,
as levelling networks tended to be redefined at intervals,
their redefinition in itself was a potential source of error,
as ‘heights’ were redefined.
Consequently, while interaction between sea level spe-
cialists and national surveyors is inevitable, we advise
most sea level specialists to take great care with the
concept of a national levelling system.
4.1.7 Chart Datum
The chart datum (or Admiralty Chart Datum in the UK)
is the low-water plane below which the depths on a
nautical chart are measured and above which, tidal lev-
els are often presented for practical purposes, such as
tide tables for harbour operations. The chart datum is
a horizontal plane over a limited area and the elevation
of this plane will vary around the coastline, depending
on the tidal ranges at the places considered. In the
UK, the chart datum at a port is the same as ‘Lowest
Astronomical Tide’ (Pugh, 1987).
4.1.8 Working Datums
Practical working datums are often used in ports
where they describe sea level (or water depth) more
clearly than perhaps a scientifically rigorous reference
to a benchmark. Examples of such datums include
the levels of the sill of a lock or a shallow point in the
entrance to a harbour. The sea level from a tide gauge
then indicates the depth of water above these hazards.
Working datums often functioned as the first TGBMs
for Europe’s sea level records (e.g. the ‘Old Dock Sill’
datum at Liverpool).
4.2 Levelling Between Local
Benchmarks
High-precision levelling will need to be made between
all the marks of the local network at regular intervals.
For GLOSS purposes, the recommendation is that the
exercise be repeated at least annually, with results fully
documented by the responsible agency. The exact fre-
quency of required levelling will depend on the geology
of the area. On unstable ground, more frequent level-
ling may be necessary.
Personnel familiar with the best practices of the tech-
nique should perform levelling with a good-quality level
and staff. For example, if marks are far apart, it will be
necessary to establish ‘staging points’ clearly identified
and about 50 m apart on a hard surface. This can be
done by painting a small ring around the point and,
on softer surfaces, by driving in a round-headed pin.
The levelling instrument can then be set up between a
benchmark and the first staging point and readings of
the staff taken at the two positions. This is then repeat-
ed throughout the whole network. It is important that
the pairs of readings be taken in the correct sequence,
otherwise an erroneous height difference will result.
Modern levelling instruments with built-in data loggers
can remove most of the tedious arithmetic associated
with the use of a simple level.
As with many other aspects of tide gauge operations,
the main principle of levelling is that ‘practice makes
perfect’. For advice on good levelling methods, the
PSMSL website (www.pol.ac.uk/psmsl) contains a set of
notes used by Prof. Charles Merry at the University of
Cape Town GLOSS Training Course in 1998.
4.3 Levelling Between Wider Area
Marks
The height of the TGBM should also be related to a
wider area network extending typically 10 km. This
provides a verification of whether the sea level mea-
sured relative to the TGBM is consistent with that of
the surrounding area.
First-order geodetic levelling is accurate to 1 or 2 mm
over distances of a few kilometres and, therefore,
annual campaigns can detect any vertical movements
of the TGBM with respect to the local benchmarks.
30
Sea Level Measurement and Interpretation
IOC Manuals and Guides No 14 vol IV
Levelling over longer distances has been found to con-
tain significant systematic errors that can cause appar-
ent spurious changes in the height of the TGBM. For
this reason the PSMSL requires MSL data to be defined
with respect to the TGBM rather than with respect to
national datum levels.
Consequently, while it is desirable in principle to per-
form regular wide area levelling, their accuracy has
always to be considered, especially as the areas consid-
ered become wider. At a distance of some 10 km, the
errors involved in levelling become comparable to those
achievable by means of space geodetic techniques.
Therefore, while the choice of technology for the wider
area surveys is clearly evolving, the principle that the
relative sea level measurements provided by the gauge
data are applicable to studies for the surrounding
area is still valid. Table 4.1 summarizes the accuracy
obtained by the different techniques.
4.4 Geodetic Fixing of Tide Gauge
Benchmarks
4.4.1 Introduction
Over the past decade, advances in modern geodetic
techniques have provided new methods for fixing tide
gauge bench marks. These are the techniques of space
geodesy, using the satellites of the Global Positioning
System (GPS) and those of the Doppler Orbitography and
Radiopositioning Integrated by Satellite (DORIS) system.
Absolute gravity measurements provide collateral evi-
dence of vertical crustal movements. The space geodesy
measurements can be used to fix into a geocentric refer-
ence frame the GPSBM, which should be connected to
the TGBM by levelling. Therefore, the MSL at the tide
gauge can be defined in a global geocentric reference
frame. This furnishes an absolute measure of mean sea
level, rather than MSL relative to each local TGBM, or
even to the wider surrounding area. Measurements of sea
level are then defined in the same geocentric reference
frame as that used for satellite altimetry and can therefore
be directly compared with altimetric sea levels.
Repeated space geodesy measurements at a tide
gauge, annually over a decade for example, enables
the vertical crustal movement to be determined and
therefore removed from the mean sea level trend to
give the true sea level change due to climatic influenc-
es. Measuring changes of gravity near the tide gauge
using an absolute gravimeter allows a completely
independent determination of any vertical crustal
movements. Figure 4.2 shows a schematic diagram of
a local levelling network within a tide gauge system to
measure absolute sea levels.
Technique Accuracy
Primary Levelling of Local
Benchmarks
0–1 km: <1 mm
1–10 km: <1 cm
GPS from TGBM to SLR/VLBI
Reference Frame
<1 cm
Absolute Gravity near Tide
Gauges and at SLR/VLBI
Station
<2 µgal
(approx. 1 cm)
Table 4.1 Accuracy of geodetic fixing of TGBMs.
Figure 4.2 Schematic of levelling required between various benchmarks at a tide gauge station.
Sea Level Measurement and Interpretation
31
32
Sea Level Measurement and Interpretation
IOC Manuals and Guides No 14 vol IV
An international working group was set up in the late
1980s by the International Association for the Physical
Sciences of the Ocean, under its Commission on Mean
Sea Level and Tides, to recommend a strategy for the
geodetic fixing of tide gauge bench marks. These
resulted in the ‘Carter reports’ (Carter et al., 1989;
Carter, 1994). The following sections provide a sum-
mary and describe recent developments. The reader is
referred to Neilan et al. (1998) and Bevis et al. (2002)
for further details.
4.4.2 GPS Measurements
Over the past decade, the GPS technique has devel-
oped rapidly to the extent that it is of fundamental
importance to many areas of geophysical research (see
links documented on the PSMSL training web page).
The International GNSS Service (IGS) receives data
from a global network of GPS stations and produces
information on the orbits of the GPS satellites which
is significantly more precise than the ephemerides
routinely transmitted by the satellites themselves. This
information is employed by researchers to produce pre-
cise positioning computations. GPS data from the IGS
network are archived at the IGS Central Bureau.
Ideally, all tide gauge sites should be equipped with
a permanent continuous receiver (CGPS). However,
in practice, the financial resources required are often
large. Many countries adopt the procedure of install-
ing permanent GPS receivers at strategic tide gauges
and then densifying the network with regular GPS
campaign measurements (Neilan et al., 1998). There
is clearly an advantage in concentrating CGPS work
at sites with long duration PSMSL RLR mean sea level
records. The GLOSS Implementation Plan refers to this
set as the GLOSS Long-Term Trends (GLOSS-LTT) net-
work. The campaigns can then concentrate on other
tide gauges in the network for which the records are
shorter. The exact mix between permanent and cam-
paign GPS tide gauges will change as the cost of GPS
receivers continues to decrease.
For studies involving sea level, it is recommended that
a dual-frequency CGPS receiver should be installed
directly at the tide gauge so that it monitors any move-
ment of the TGBM. If the receiver is placed exactly at
the TGBM, then the GPSBM and the TGBM will coin-
cide, eliminating the need for levelling between the
two benchmarks. The TGBM is then the fundamental
point that is geocentrically located by the GPS mea-
surements and to which all the sea level measurements
are related. In practice, tide gauge sites are not always
ideal for making GPS measurements. This may be due
to obscured sky visibility, excessive multipath transmis-
sions or because of radio interference, in which case a
site should be chosen that is as close as possible to the
tide gauge. Ideal