Mean Sea Level Explained

This article opens the Elevation & Vertical Datums sub-hub. The /learn/the-geoid-explained and /learn/ellipsoid-vs-geoid articles in the Datums sub-hub cover the abstract reference surfaces; this sub-hub goes deeper on what people actually use for elevation in practice — chart datums, vertical reference frames, sea-level rise, the Mount Everest height controversy, and the satellite-altimetry record.

What MSL is

Mean Sea Level is conceptually simple: take a long series of water-level measurements at a coastal location, average them, get MSL. In practice the details matter:

• Measurement device: tide gauge — a fixed instrument recording water level continuously or hourly.
• Sampling rate: hourly or more frequently for modern gauges; historical records may be twice-daily or hourly.
• Averaging period: 19 years (one full lunar nodal cycle plus margin) is the canonical period for “mean” sea level.
• Location specificity: MSL is determined locally; different sites have different MSLs.

The result: a single number — the MSL at that gauge location, valid for the averaging epoch, with units of meters above some local reference (typically a permanent benchmark on land).

The lunar nodal cycle

Tides have several periodic components:

• Semi-diurnal (~12.42 hours): the dominant twice-daily tide.
• Diurnal (~24.84 hours): once-daily component.
• Spring/neap (~14.77 days): the fortnightly cycle of tidal range.
• Annual (1 year): seasonal sea-level changes.
• Lunar nodal cycle (18.61 years): slow variation in tidal range as the Moon's orbital plane precesses.

Averaging over a full lunar nodal cycle is needed to remove the longest periodic variation. The standard averaging period is 19 years (slightly longer than 18.61, for integer-year alignment).

Shorter averaging windows (one year, five years) leave residual variation from the nodal cycle. Modern operational MSL values from NOAA, UK Hydrographic Office, and similar agencies use 19-year averages.

Tide gauges

The primary measurement tool. Types:

Stilling-well gauges (traditional)

A vertical pipe (stilling well) is positioned to allow water in but damp the wave action. A float inside the well rises and falls with the water level; the float height is recorded mechanically or electronically.

Stilling wells are reliable, simple, and have a 100+ year track record at many sites. They're still in use at many global tide gauge stations.

Acoustic gauges

A downward-pointing acoustic transducer measures the distance from a fixed sensor head to the water surface via reflected sound waves. No moving parts; no stilling well needed. Higher precision and less maintenance than stilling-well gauges.

Pressure-based gauges

A submerged pressure sensor measures the water column above it; pressure is converted to water level using density and atmospheric pressure corrections. Common for offshore deployments (gauges at depth) and for arrays of gauges.

GPS-augmented gauges

Modern installations combine traditional gauges with co-located GPS to track the vertical motion of the land at the gauge site. This separates the rise (or fall) of the sea from the rise (or fall) of the land.

Major tide gauge networks

• NOAA NOS (US National Ocean Service): ~210 tide stations in the US and US territories.
• UK Tide Gauge Network: ~45 stations operated by the UK National Tidal and Sea Level Facility.
• GLOSS (Global Sea Level Observing System): international coordination of ~290 high-priority stations worldwide.
• PSMSL (Permanent Service for Mean Sea Level): the global tide gauge data archive, housed at the UK National Oceanography Centre. PSMSL holds records from ~2,000 stations, some dating to the early 19th century.

Some stations have very long records:

• Brest, France: 1807–present (the world's longest continuous tide gauge record).
• Stockholm, Sweden: 1774–present.
• San Francisco, USA: 1854–present.
• Sydney, Australia: 1886–present.

These long records are essential for sea-level-rise research.

The National Tidal Datum Epoch

The National Tidal Datum Epoch (NTDE) is the US official 19-year averaging window used for MSL and related datums. The current NTDE is 1983–2001.

NOAA revises the NTDE roughly every 20–25 years to keep the datum aligned with current MSL. A new NTDE covering ~2002–2026 is being computed and is expected to be adopted in the late 2020s.

Other countries have similar conventions:

• UK: 1996–2014 epoch.
• Australia: 1992–2011 standard epoch.
• Canada: shared with the US for some shared waters; separate elsewhere.

When the NTDE changes, MSL values at all stations update simultaneously. Charts, surveys, and elevation databases referenced to the old epoch must be revised or include an adjustment.

Why MSL varies by location

The ocean is not a smooth gravitational equipotential surface. Sources of variation:

Sea surface topography (SST)

Long-term variation in MSL across the Earth — typically ±1 m globally. Causes:

• Ocean currents: the Gulf Stream raises sea level on the US east coast by ~30 cm vs the west coast at the same latitude. The Antarctic Circumpolar Current produces similar effects in the Southern Ocean.
• Temperature: warmer water expands (thermosteric effect). Tropical Pacific is ~50 cm higher than cool subarctic seas.
• Salinity: lower-salinity waters are less dense; river-mouth areas have slightly elevated MSL.
• Atmospheric pressure: persistent low-pressure systems raise MSL beneath them (~1 cm per millibar of low pressure — the inverse barometer effect).

The Pacific-vs-Atlantic difference at Panama is a classic example: the Pacific Ocean is ~80 cm higher than the Atlantic Ocean across the isthmus, due to a combination of current patterns, density differences, and atmospheric pressure. The Panama Canal locks are designed to handle this.

Earth rotation and tides

Tides averaged out by definition, but the Earth's rotation imparts a centrifugal effect that's absorbed into the geoid surface.

Vertical land motion

The land itself rises or falls — earthquakes, glacial isostatic adjustment, tectonic motion. Tide gauges measure the relative sea level (sea relative to land); GPS co-located gauges separate the two contributions.

MSL vs the geoid

The geoid (see /learn/the-geoid-explained) is the equipotential surface of Earth's gravity that the ocean would form in static equilibrium without tides, currents, atmospheric pressure variations, etc.

MSL approximates the geoid but deviates by sea surface topography (SST) of up to ±1 m. The deviation is what creates the Pacific-Atlantic Panama difference, the equatorial bulge in the Pacific, etc.

For high-precision geodetic applications, the distinction matters: orthometric heights are geoid-relative, while MSL-based heights are tide-gauge-derived (and include SST). The two systems can differ by ~1 m globally.

For typical engineering purposes the difference is small enough to ignore.

Chart datums

For marine navigation, MSL is augmented by chart datums — reference levels designed to understate water depth for safety. A ship's draft (depth of keel below water) should not exceed the charted depth at the worst tide condition.

Mean Lower Low Water (MLLW)

The average of the lower of the two daily low tides over the 19-year NTDE. Used as the chart datum for US Pacific Coast waters. Conservative — most low tides are higher than MLLW.

Mean Low Water (MLW)

The average of all daily low tides. Used historically for US Atlantic Coast charts; some replaced by MLLW.

Lowest Astronomical Tide (LAT)

The lowest tide predicted from astronomical effects only (excluding weather). Used by the International Hydrographic Organization as the recommended chart datum for new charts since 2007. LAT is even more conservative than MLLW. UK Admiralty charts use LAT; many other countries are migrating.

Highest Astronomical Tide (HAT)

The opposite: the highest tide from astronomical effects. Used for bridge clearance charts and similar maximum-water-level applications.

The cascade of vertical references at a typical coastal site (US East Coast example):

HAT
- Mean High Water (MHW)
- Mean Higher High Water (MHHW)
- MSL (1983-2001 NTDE)
- Mean Low Water (MLW)
- Mean Lower Low Water (MLLW) ← chart datum
- LAT

These can differ by 2+ meters at sites with strong tides (Bay of Fundy, Cook Inlet).

Satellite altimetry

Since 1992, the TOPEX/Poseidon → Jason → Sentinel-6 mission series has measured global ocean surface height from space. The technique:

1. A radar altimeter on the satellite sends a pulse downward.
2. The pulse reflects from the ocean surface.
3. The round-trip time gives the satellite-to-ocean distance.
4. Combined with the satellite's known orbit, this yields the sea surface height relative to a reference ellipsoid.

The technique works over the open ocean only (coastal areas have complications from land radar returns and shallow-water effects). The accuracy is ~3 cm per measurement; averaging over many measurements yields millimeter-level precision for the global mean.

Mission lineage:

• TOPEX/Poseidon (NASA/CNES, 1992–2005)
• Jason-1 (NASA/CNES, 2001–2013)
• Jason-2 / OSTM (NASA/CNES/NOAA/EUMETSAT, 2008–2019)
• Jason-3 (NOAA/CNES/EUMETSAT, 2016–present)
• Sentinel-6 Michael Freilich (NASA/ESA/NOAA/EUMETSAT, 2020–present)
• Sentinel-6B (planned launch ~2026)

The 30+ year continuous record from this mission series is the foundation of modern sea-level-rise science.

Sea level rise

Global mean sea level has risen approximately 21–24 cm since 1880, with the rate accelerating in recent decades:

• 1880–1990: average ~1.4 mm/year.
• 1990–2010: ~3.0 mm/year.
• 2010–2024: ~3.4 mm/year (NASA Sentinel-6 data).
• Acceleration: ~0.1 mm/year² (the rate itself is increasing).

Two main contributions:

• Thermal expansion (~40% of recent rise): warmer ocean water occupies more volume.
• Land-ice melt (~60% of recent rise): glaciers, Greenland Ice Sheet, Antarctic Ice Sheet adding water to the oceans.

Local rates vary substantially:

• Tuvalu, Maldives: ~3–5 mm/year (rapidly rising ocean).
• Stockholm, Sweden: -3 mm/year (glacial isostatic rebound; land rises faster than sea).
• Jakarta, Indonesia: +50+ mm/year (massive land subsidence from groundwater extraction).
• Tokyo, Japan: ~+5 mm/year (subsidence + ocean rise).

Projections to 2100 from IPCC AR6 (2021):

• Low emissions scenario (SSP1-2.6): +0.43 m by 2100.
• High emissions scenario (SSP5-8.5): +0.84 m by 2100.
• High-end / less-likely scenarios: 1+ m possible.

The Mount Everest controversy

A vivid example of how reference choices affect heights. Different surveys have produced different “Mount Everest height” values:

• 1856 (Indian Trigonometric Survey): 29,002 ft = 8,839 m (later corrected to 29,029 ft = 8,848 m).
• 1955 (Indian survey): 8,848 m (matched mid-19th c. estimate; this became the “classic” value).
• 1999 (US National Geographic / Bradford Washburn): 8,850 m using GPS.
• 2005 (Chinese survey): 8,844.43 m rock height (snow varies).
• 2020 (Joint Nepal-China survey): 8,848.86 m snow height, the current official value used by both Nepal and China.

The variations come from:

• Reference datum: which MSL? Indian, Chinese, or Nepalese tide gauges produce slightly different values.
• Rock vs snow height: the snow on Everest's summit varies; rock height is more stable but harder to measure.
• Measurement technique: triangulation vs GPS vs satellite-only.
• Geoid model: the geoid undulation above the WGS 84 ellipsoid at Everest's latitude/longitude.

The 2020 Nepal-China joint measurement was designed to end the disagreement; both countries now use 8,848.86 m.

Common misconceptions

“Sea level is the same everywhere.” It's not. The ocean surface has ±1 m variation globally due to sea surface topography (currents, density, pressure differences). The Pacific is ~80 cm higher than the Atlantic across Panama; the Indian Ocean has its own pattern.

“MSL and the geoid are the same.” Close but distinct. The geoid is an equipotential surface; MSL is a tide-gauge-derived average that includes sea surface topography. They differ by up to ±1 m.

“Tide gauges measure absolute sea level.” They measure relative sea level — sea relative to the land at the gauge site. Vertical land motion is indistinguishable from sea-level change without co-located GPS. Modern installations include both.

“Satellite altimetry replaces tide gauges.” They're complementary. Satellites measure global/open-ocean sea level with full coverage; tide gauges measure local/coastal sea level with long historical records. Together they give a complete picture.

“Sea level rise is uniform globally.” It's not. Some regions experience much faster rise (tropical Pacific, eastern North America), others slower (eastern South America, parts of Northern Europe). Local vertical land motion adds further variability.

“The NTDE is fixed.” NOAA revises it roughly every 20–25 years. The current 1983–2001 NTDE will be replaced by a new epoch in the late 2020s.

“Chart datums are arbitrary.” They're deliberately conservative (designed to understate water depth for safety). The choice of MLLW vs MLW vs LAT reflects national/historical conventions. IHO recommends LAT for new charts globally.

“Mount Everest's height is precisely known.” Different measurement campaigns produce values varying by ~6 meters. The 2020 joint Nepal-China measurement (8,848.86 m) is the current official value but small revisions are possible.

“Sea level rise will displace coastal cities.” Most major coastal cities have adaptation plans (sea walls, managed retreat, etc.). The rate of rise is enough to require investment but not enough to instantly displace populations. Some small island nations (Tuvalu, Maldives) face existential risk that's much harder to mitigate.

“Global MSL is computed from one place.” It's computed from satellite altimetry over the open ocean plus tide gauge data, averaged globally. The result is a single global number (~3.4 mm/year rise rate) but the underlying data are spatially distributed.