Mean Sea Level, GPS, and the Geoid

By Witold Fraczek, Esri Applications Prototype Lab

The eoid appro!i"ates "ean sea level# The shape o$ the ellipsoid %as calc&lated based on the hypothetical e'&ipotential ravitational s&r$ace# A si ni$icant di$$erence e!ists bet%een this "athe"atical "odel and the real ob(ect# )o%ever, even the "ost "athe"atically sophisticated eoid can only appro!i"ate the real shape o$ the earth#

Frequently research and technology endeavors have unforeseen but positive outcomes. When European explorers set out to find a shortcut to India, they discovered the New World. When a staphylococci bacteria culture was mista enly contaminated with a common mold, the clear area between the mold and the bacterial colony led to the conclusion that the mold, Penicillin notatum, produced a compound that inhibited the growth of bacteria. !his chance discovery led to the development of the antibiotic penicillin. !hat the earth does not have a geometrically perfect shape is well established, and the geoid is used to describe the unique and irregular shape of the earth. "owever, only recently have the more substantial irregularities in the surface created by the global mean sea level #$%&' been observed. !hese irregularities are an order of magnitude greater than experts had predicted. (ontrolled by the gravitational potential of the earth, these irregularities form very gentle but massive )hills) and )valleys.) !his astonishing finding was made possible through the use of *+%, a technology designed by the ,nited %tates -epartment of -efense to revolutioni.e navigation for the ,.%. Navy and /ir Force. *+% has done that0and a lot more.

What *s Mean Sea Level+

The acc&racy o$ GPS hei ht "eas&re"ents depends on several $actors b&t the "ost cr&cial one is the ,i"per$ection, o$ the earth-s shape# )ei ht can be "eas&red in t%o %ays# The GPS &ses hei ht .h/ above the re$erence ellipsoid that appro!i"ates the earth-s s&r$ace# The traditional, ortho"etric hei ht .)/ is the hei ht above an i"a inary s&r$ace called the eoid, %hich is deter"ined by the earth-s ravity and appro!i"ated by MSL# The si ned di$$erence bet%een the t%o hei hts0the di$$erence bet%een the ellipsoid and eoid0is the eoid hei ht .1/# The $i &re above sho%s the relationships bet%een the di$$erent "odels and e!plains the reasons %hy the t%o hardly ever "atch spatially#

For generations, the only way to express topographic or bathymetric elevation was to relate it to sea level. *eodesists once believed that the sea was in balance with the earth1s gravity and formed a perfectly regular figure. $%& is usually described as a tidal datum that is the arithmetic mean of hourly water elevations observed over a specific 234year cycle. !his definition averages out tidal highs and lows caused by the changing effects of the gravitational forces from the moon and sun. $%& is defined as the .ero elevation for a local area. !he .ero surface referenced by elevation is called a vertical datum. ,nfortunately for mapma ers, sea level is not a simple surface. %ince the sea surface conforms to the earth1s gravitational field, $%& also has slight hills and valleys that are similar to the land surface but much smoother. "owever, .ero elevation as defined by %pain is not the same .ero elevation defined by (anada, which is why locally defined vertical datums differ from each other. !he $%& surface is in a state of gravitational equilibrium. It can be regarded as extending under the continents and is a close approximation of the geoid. 5y definition, the geoid describes the irregular shape of the earth and is the true .ero surface for measuring

elevations. 5ecause the geoid surface cannot be directly observed, heights above or below the geoid surface can1t be directly measured and are inferred by ma ing gravity measurements and modeling the surface mathematically. +reviously, there was no way to accurately measure the geoid so it was roughly approximated by $%&. /lthough for practical purposes, at the coastline the geoid and $%& surfaces are assumed to be essentially the same, at some spots the geoid can actually differ from $%& by several meters.

2i$$erin

Meas&re"ents

*+% has transformed how altitude at any spot is measured. *+% uses an ellipsoid coordinate system for both its hori.ontal and vertical datums. /n ellipsoid0or flattened sphere0is used to represent the geometric model of the earth.

The s&r$ace o$ lobal &nd&lations %as calc&lated based on alti"etric observations and very precise .&p to t%o centi"eters/ "eas&re"ents taken $ro" the T3PE45P3SE*231 satellite# This data %as represented in the Earth Geodetic Model .EGM67/, %hich is also re$erred to as the spherical har"onic "odel o$ the earth-s ravitational potential#

(onceptually, this precisely calculated ellipsoid, called an oblate ellipsoid of revolution, was intended to replicate the $%& as the main geodetic reference or vertical datum. If this ellipsoid vertical datum is used, height above the ellipsoid will not be the same as $%& and direct elevation readings for most locations will be embarrassingly off. !his is caused, in part, because the *+% definition of altitude does not refer to $%&, but rather to a gravitational surface called the reference ellipsoid. 5ecause the reference ellipsoid was intended to closely approximate the $%&, it was surprising when the two figures differed greatly. !he !6+E78+6%EI-6N satellite, launched in 2339, was specifically designed to perform very precise altimetric observations. !hese measurements have demonstrated that neither human error nor *+% inaccuracies are responsible for the sometimes substantial discrepancies between ellipsoid and $%& measurements. In fact, the three4dimensional surface created by the earth1s sea level is not geometrically correct, and its significant irregularities could not be mathematically calculated: this explains the difference between the ellipsoid4based *+% elevation readings and elevations shown on accurate topographic maps.

/ brief examination of elevation readings for Esri headquarters in ;edlands, (alifornia, demonstrates these differences. !he campus elevation is shown on topographic quadrangle maps and high4resolution digital elevation models #-E$s' for the area as approximately <== meters above $%&. "owever, a precise, nonad>usted *+% reading for the same location typically shows the elevation as ?@A meters.

The "ap sho%s the areas o$ the lobe that %o&ld have a sea level belo% the theoretical s&r$ace o$ the WGS89 ellipsoid, or the theoretical and eo"etrically correct sea level .sho%n in bl&e/# The sharp contrast bet%een the bl&e and reen indicates %here the ellipsoid and eoid intersect# With the continents displayed as opa'&e, the re"ainin area covered by %ater reveals %here sea level is act&ally at zero elevation relative to the WGS89 ellipsoid#

Why is there a ?94meter differenceB !he *+% receiver uses a theoretical sea level estimated by a World *eodetic %ystem #W*%A<' ellipsoid, which does not perfectly follow the theoretical $%&. !he $%&, approximated by an ellipsoid, is related to gravity or the center of mass of the earth. -iscrepancies between a W*%A< ellipsoid, and the geoid vary with location. !o continue with this example, elevation readings for Cucaipa, a city located less than 2= miles east of ;edlands, differ by ?2.D meters. !he geoid and ellipsoid intersect at the geoid undulations. ,ndulations result from several phenomena, the most significant of which is the existence of gravitational anomalies caused by the nonhomogeneous nature of the earth. !he density of magma in the earth1s crust is unevenly distributed. In areas of greater density, it can be significantly higher and, consequently, cooler. &ess dense areas are correspondingly lower and hotter. -ense magma exerts a stronger pull, which causes the accumulation of water masses. Not much is nown about whether these volumes relocate or how fast they move. If these locations do move, their movement would be at the same pace as other geologic events #i.e., extremely slow'. +recise measurements ta en from space are applied to *+% readings. !hese measurements are based on an ellipsoid surface that is a mathematically generated model of the earth obtained from a three4dimensional (artesian coordinate system. !he *+% receivers can provide only the ellipsoid #geometric' height.

"owever, most users expect accurate elevation readings that are related to $%&. (onsequently, newer *+% devices output orthometric #geoid' height measurements as a product of )behind the scenes) calculations based on a combination of formulas, tables, and matrices that use geographic coordinates as inputs. !he appropriate height for the geographic location ta en from a coarse or fine -E$ matrix is provided instead of a real measurement of the . value #or height'. %ome receivers use approximations of the geoid height to estimate the orthometric height from the ellipsoid height. %till other units, based on an older technology, provide direct readings of the . value based on the ellipsoid.

A Globally 2e$ined Geoid0Geoid66

This "ap sho%s a %orld %ith no ravitational ano"alies# A hypothetically ho"o eneo&s distrib&tion o$ "ass %ithin the earth %o&ld chan e the delineation o$ oceans and land"asses# This vie% o$ So&theast Asia depicts s&ch a spec&lative shoreline# 1ote the di$$erences bet%een the c&rrent and the theoretical o&tlines o$ the %orld ocean#

/ globally defined geoid was required so that *+% receivers could calculate the correct . value needed as a reference surface for a global vertical datum. *eoid33, a model with a submeter level of accuracy, was developed by the National *eodetic %urvey. It is used as a .ero surface to establish consistent and accurate elevations worldwide. "owever, despite this impressive level of accuracy, portions of the *eoid33 deviate from $%& due to gravity effects. !he Earth *eodetic $odel #E*$3@', developed in a collaborative effort by N/%/ *oddard %pace Flight (enter, the National Imagery and $apping /gency #NI$/', and 6hio %tate ,niversity, has been used to compute geoid undulations accurate to better than one meter #except in areas lac ing accurate surface gravity data'. !he values of this surface show at every location how much $%& varies from the ellipsoid used as the reference for *+% elevation readings. In other words, E*$3@ shows how uneven the gravitational potential of the ocean1s surface really is. !he maximum range of the *eoid33 undulations with respect to the W*%A< ellipsoid is 239 meters. !he biggest anomaly was discovered southeast of India where the geoid is 2=D meters below the ellipsoid, and its highest swelling was observed in eastern Indonesia.

,What *$, Si"&lations

This "ap sho%s the hypothetical appro!i"ate delineation o$ land"ass and ocean based on a spherically:shaped earth, %hich ass&"es that the c&rrent relie$ %o&ld be preserved#

"ow would continental coastlines change if there were no gravitational anomaliesB !he short answer would be that they would appear the same as if the earth were an ellipsoid. !he biggest undulations are concentrated in the northern Indian 6cean and within the Indonesian /rchipelago. (onsequently, areas currently covered by the /rafura %ea and (arpenteria 5ay, both north of /ustralia, would emerge. (onversely other areas, such as the lowlands formed by the deltas of some of the main %outh /sian rivers such as the *anges and 5rahmaputra as well as those of Indus, Irrawaddy, and $e ong, would sin . /s an intellectual exercise, imagine for a moment that the earth has the geometrically perfect shape of a spheroid and then an ellipsoid. !o visuali.e these )what if) scenarios, a computer simulation was generated using rasters that represented distances to the earth1s center and were generated with the ;aster (alculator in the /rc*I% %patial /nalyst extension. / standard digital elevation model for the globe was modified by removing the gravitational anomalies identified by E*$3@. !hese geometrically correct representations of a spheroid and ellipsoid were used to produce *I% simulations called the %+"E;6I- -E$ and E&&I+%6I- -E$. !he accompanying illustrations show the effects these shapes, as well as other scenarios, have on the distribution of landmass and ocean.

SP)E;3*2 2EM

What i$ the earth act&ally stood still+ The data appro!i"ates a theoretical elevation "odel %ith the elevations relative to the distance $ro" the earth-s center#

The GPS re$ers its direct elevation readin s to this shape# The cell val&es o$ the above ellipsoid<de" rid are the direct readin s o$ GPS# The displayed GPS readin s are based on the added ravitational correction as the shape o$ the eoid is based on ravitational potential .both positive and ne ative/#

!he %+"E;6I- -E$ was created based on the assumption that the earth1s shape, or more precisely, the shape of the earth1s surface as represented by $%&, is spherical. In other words, the globe1s radius connecting the center of mass with the hypothetical equipotential gravitational surface is identical everywhere on the planet. !he radius was set to @,?@E,<E? meters, which is the distance to the earth1s center from the ellipsoidal reference surface at <D degrees latitude. !his change in geometry to a spheroid geometry with its altered gravity would cause the global ocean to change. !he polar .ones would be relatively farther from the center of the earth, and these new higher altitudes would force seawater toward the equator. 5y the same to en, the equatorial region would be relatively closer to the earth1s center and would be more strongly affected by gravity. Increased gravitational force in the equatorial .one would pull oceanic water toward the equator and form a global equatorial ocean.

ELL*PS3*2 2EM
!he E&&I+%6I- -E$ depicts the earth as an ellipsoid. /lthough distances from the surface of an ellipsoid to the earth1s center along each latitude are identical, each latitude has its own unique value increasing from each pole toward the equator. !o loo at this ellipsoid earth from yet another perspective0its physical shape characteri.ed by the distance of any point on the surface to the center of the earth1s mass0a grid was generated using the W*%A< datum definition. Each grid cell value represents the distance in meters from the surface to the earth1s center of mass. / complex combination of trigonometric functions was applied to create the representation of the ellipsoid. !hen the -E$ representing the earth1s current relief was added to the ellipsoid raster.

The "ap sho%s distances to the earth-s center o$ "ass# The areas hi hli hted in red represent

"o&ntains that are the $arthest $ro" the center o$ "ass on the earth and are the areas o$ the lo%est ravitational p&ll on the earth-s s&r$ace#

!he difference between the ma>or and minor axes of the W*%A< ellipsoid is <9,EE= meters. !he difference between the length of the radius at the equator and one of the poles is 92,?AD meters0only =.?? percent of the radius. !he )flattening) of the earth is, geometrically spea ing, relatively insignificant but in terms of geography, it has a tremendous impact.

*$ the Earth Stood Still

What would happen if the earth stopped spinning and the centrifugal effect ceased to force oceans to accumulate around the equatorB It appears that the world1s ocean would split into two polar oceans and leave the equatorial area totally dry. !o model this hypothesis, a value of @,?E2,2<@ meters0the distance from the earth1s center indicates the approximate elevation of the sea level on the reference ellipsoid0was specified to separate water from land. For this )what if) simulation, the elevation of the sea level was based on the assumption that the volume of ocean water would be about the same as it is today.

Meltin

Glaciers

This "ap ill&strates %hat %o&ld happen i$ all the laciers "elted# The cell val&es o$ this ortho"etric s&r$ace rid are related to MSL#

;eturning to a geoid representation of the earth, one more simulation models an earth where all the glaciers have melted. !his simulation might predict a future that is only a few hundred years distant if the large glaciers of /ntarctica and *reenland, which currently cover approximately 2= percent of all land, melted as a result of global warming. If all the water in these glaciers were released, the $%& would rise about A= meters above its current level.

=oncl&sion
*I% ma es it possible to explore the effects of different conceptuali.ations of the earth1s shape and model a variety of global conditions.