How to Correct Elevation Easily: Five Steps to Avoid Mistakes in the Field

In the field, consultations about “elevations not matching,” “RTK heights shifting depending on the point,” or “not matching past deliverables” tend to occur more often for vertical measurements than for horizontal positions. In many cases the cause is not instability of the positioning itself, but that different elevation reference frames are being used, or that even if the reference is the same, the correction procedure is not being adopted in field operations. Elevation is directly linked to the direction water flows and is important for infrastructure and disaster prevention; however, the height obtained from satellite positioning (ellipsoidal height) uses a different reference surface and cannot be used as the practical “elevation” without conversion. 

Furthermore, from April 1, 2025, entity["country","日本","country in east asia"] will revise the elevation results of control points, and it has been clarified that the elevation system will transition to a framework based on satellite positioning. With this change, the definition of control point results moves from “Geodetic Datum 2011” to “Geodetic Datum 2024”; while horizontal positions (latitude and longitude) remain unchanged, the handling of elevation has been reorganized under the new framework. Mixing old and new results or confusing geoid models can cause seemingly non-reproducible field problems such as “a consistent offset for some reason” or “only remote islands don’t match.” 

This article summarizes, for practitioners who search “elevation correction method,” the shortest path to understanding the concept of elevation correction and five steps to avoid field mistakes. It explains the process in a way that does not depend on specific devices or software and can be applied when you want to display height in an RTK real-time view. At the end, it also touches on LRTK, a smartphone-mounted GNSS high-precision positioning device that makes it easier to incorporate elevation correction into field routines.

Table of contents

• Why elevation correction is necessary

• The elevation references to grasp first for correction

• Five steps to avoid mistakes in the field

• Common failure patterns and how to avoid them

• Tips to maintain elevation correction in RTK operations

• Making elevation correction a field routine with LRTK

Why elevation correction is necessary

The primary reason elevation correction is necessary is that the “height” people want to use in the field is the height referenced to mean sea level, which is influenced by gravity, whereas the height directly obtained by GNSS is a geometric height referenced to the Earth’s ellipsoid. entity["organization","国土地理院","japan mapping agency"] explains that Japan’s elevation reference is legally defined by the Survey Act as mean sea level, and the surface obtained by virtually extending that mean sea level onto land is called the geoid. It is also made clear that elevation can be obtained by subtracting geoid height from ellipsoidal height. 

A typical field mistake is treating ellipsoidal height as elevation. Ellipsoidal height is convenient for position calculations, but because the real Earth has gravitational anomalies, using ellipsoidal height as the everyday height reference can lead to contradictions that conflict with intuition—such as water flowing in unexpected directions—something introductory materials on satellite positioning warn about. Elevation correction resolves this contradiction and converts heights into values usable for design, construction, and inspection in the field. 

Elevation is also used in combination with multiple pieces of information such as “past results,” “known points,” “design drawings,” and “as-built checks.” If you omit elevation correction and use RTK heights in the field, differences in timing, personnel, or equipment can make heights appear to shift even at the same site, breaking quality control. The revision of elevation results effective April 1, 2025, states that the elevation system will be based on satellite positioning, which is expected to enable faster use of elevation results and improve efficiency in surveying and public works. Conversely, being able to perform elevation correction correctly will become a future field standard. 

The elevation references to grasp first for correction

The most important thing in understanding elevation correction is not memorizing terminology but having a single formula that shows which reference surface is used and what is subtracted in what order. The basic form is very simple: as repeatedly shown in materials and explanations, elevation is obtained by subtracting geoid height from ellipsoidal height. This is the core of elevation correction. 

However, in Japanese practice there is a tendency to assume “the same mean sea level is used nationwide as elevation 0 m.” Be careful: the reference surface is defined using the mean sea level of Tokyo Bay (elevation 0 m (0 ft)), and the Japan Vertical Datum (日本水準原点) is installed to fix that surface to the land. At the same time, Article 11 of the Survey Act provides a framework allowing, with approval of the Director-General of the Geospatial Information Authority of Japan, the use of a different origin (elevation) in special circumstances such as surveying remote islands. In other words, some remote islands may use a “local mean sea level” as the elevation reference, which may not coincide with elevations based on the Tokyo Bay mean sea level. 

To bridge this gap, a reference surface correction amount (reference surface correction parameter) is introduced. The data distribution pages define the reference surface correction parameter as the difference between the Tokyo Bay mean sea level and the remote island’s local mean sea level and describe it as data representing the reference surface correction amount at arbitrary positions. It is explicitly stated that in some remote islands, when deriving elevation by satellite positioning, both geoid height and the reference surface correction amount must be used. 

In practice, the procedure to obtain elevation from ellipsoidal height is concretely organized to align with the revision of elevation results. The “Nationwide Revision of Elevation Results” page shows that you should subtract the geoid height calculated from Geoid 2024 and the reference surface correction amount calculated from the reference surface correction parameter from ellipsoidal height, and also shows a method to subtract values calculated from an integrated file from ellipsoidal height. The integrated file is stated to be usable for all of Japan. With this understood, the formula to remember in the field collapses to a single line: Elevation = Ellipsoidal height − Geoid height (and if necessary, minus the reference surface correction amount). 

Five steps to avoid mistakes in the field

From here, I explain five steps to turn “knowing about elevation correction” into “being able to perform it with consistent quality every time,” covering everything from pre-deployment preparation to field checks. The key is to create operations that prevent mixing items that must not be mixed, even before worrying about calculation formulas. Because the definition of control point results changes with the revision of elevation results, field procedures must be updated accordingly. 

Step 1 is to fix the deliverable goal. Clarify whether your field needs ellipsoidal height or elevation, and align this with your deliverables, design, and as-built standards. In Japan, displaying height as height above mean sea level is legally positioned as the surveying standard, and the relationship between elevation and the geoid is organized. In other words, if your site handles “elevation,” operating on ellipsoidal height will inevitably fail to match later. Conversely, ellipsoidal height may be necessary for internal calculations or device interoperability, so separate uses by purpose and decide in advance which one to record. 

Step 2 is to align the generation (epoch) of control point and known point results. From April 1, 2025, control point survey results obtained will be “Geodetic Datum 2024,” and they will be distinguished from the pre-revision “Geodetic Datum 2011.” Even though horizontal positions (latitude and longitude) do not change, mixing generations is fatal for tasks that use height. Many field “consistent offset” issues are caused by known points still being based on old results while field positioning or the geoid model is assumed to be new. Confirm which system the known point appears in (Geodetic Datum 2011 or Geodetic Datum 2024) in the result tables and unify this within the project; this is the most important preliminary step in elevation correction. 

Step 3 is to determine which geoid model to use and whether reference surface correction is required. “Geoid 2024 Japan and Surroundings” is described as data representing geoid height at arbitrary positions given in latitude and longitude in the World Geodetic System (Japan Geodetic System 2024). Moreover, the reference surface correction parameter represents the difference between the Tokyo Bay mean sea level and a remote island’s local mean sea level, and it is stated that in some remote islands both geoid height and the reference surface correction amount are necessary. Be sure to confirm at the time of deployment whether your field falls within the area requiring a reference surface correction amount. Even if your area is not targeted, there is a Q&A noting that an integrated file is provided for nationwide use, so consider adopting the integrated file to simplify operations. 

Step 4 is to formalize field cross-checks. Even if the formula is correct on paper, discrepancies can arise from software settings in the field, file loading, units and rounding, antenna height input, and so on. The “Nationwide Revision of Elevation Results” page cautions that Geoid 2024, reference surface correction parameters, and the integrated file should be confirmed to be loadable by the software you will use before use. Therefore, from the first day of deployment, make it a rule to verify against known points. Specifically, confirm that repeated measurements of the same point do not scatter widely, elevations fall within expected ranges, and the display switches between ellipsoidal height and elevation as intended. Not checking is the largest cause of failure, so always perform this before starting field work. 

Step 5 is to record the adopted correction conditions and prepare for updates. The revision of elevation results indicates that the elevation system will shift to one based on satellite positioning and that related data and documentation will be provided. This means field elevation correction is not “set once and done”; you must follow updates to the result generation and provided data. Record which geoid model you used, whether you applied the reference surface correction amount, whether you used the integrated file or separate files, and which generation of known point results you used. With these records, even if values diverge months later, you can trace the cause. For accountability in operations, elevation correction is practical only when it includes both calculation and record-keeping. 

Common failure patterns and how to avoid them

Failures in elevation correction often present similar symptoms in the field, so misidentifying the cause can take a long time to resolve. Here, common failures are organized from the perspective of “what was mixed.” The conclusion is that many troubles arise from mixing references and settings before measurement precision itself. 

The most frequent failure is confusing ellipsoidal height with elevation. GNSS-derived height is the distance from the ellipsoidal surface to the ground, whereas elevation is the height from the geoid surface referenced to mean sea level—this distinction is consistently explained in guidance. In the field, operators sometimes assume that a screen label “height” means elevation. The avoidance measure is simple: decide in advance which type of height you will record, and make the display clearly indicate “ellipsoidal height” or “elevation” so it’s immediately identifiable. 

The next common issue is mixing result systems. Mixing “Geodetic Datum 2011” and “Geodetic Datum 2024” at the same site will cause height discrepancies. It is made clear that control point results obtained from April 1, 2025, onward are Geodetic Datum 2024, and you must distinguish these from pre-revision results. The mistaken belief that “horizontal positions don’t change so it’s fine” often leads to mixing, so ensure both the known point result tables and field settings use the same generation. For sites using pre-existing public control points, determine as a project rule how to handle elevations recorded under the old datum versus the new datum. 

A third failure is not applying the reference surface correction amount in remote islands, or conversely applying it where it is not needed. The reference surface correction amount is the difference between Tokyo Bay mean sea level and a remote island’s local mean sea level, and it is stated that in some remote islands both geoid height and the reference surface correction amount are required for satellite-derived elevation. The revised operational rules indicate that for Okinawa Island and certain remote islands, the geoid height used should be the geoid height from Geoid 2024 plus the value calculated from the reference surface correction parameter, and that the integrated file enables calculation nationwide. To avoid problems, confirm target areas before deployment and consider adopting the integrated file to keep operations from branching in the field. 

A fourth failure is proceeding without noticing software file-loading issues or format mismatches. It is advised to verify that Geoid 2024, the reference surface correction parameters, and the integrated file can be correctly loaded into the software you use before applying them. In the field, cases can occur where the software displays “applied” even though it failed to load, so make verification against known points mandatory. Skipping this check right after deployment can force you to redo field results and cause elevation correction to be disliked and become perfunctory. 

A fifth failure is insufficient record-keeping, preventing later cause tracing. Since related data and documentation are being provided with the revision of elevation results, the field side must also record “which data were used for correction.” At minimum, record the geoid model name, whether the reference surface correction amount was applied, the generation of the result system, and the basis for adopting known points; then, if values change months later, you can separate whether it was due to a datum update or a configuration error. 

Tips to maintain elevation correction in RTK operations

In RTK operations, whether elevation appears in the real-time display is an operational dividing line. If elevation appears, quick checks of as-built conformity and layout decisions are easier, but if elevation correction breaks here, people may mistakenly assume “realtime caused the error.” First, remember that elevation correction is independent of the RTK method (network RTK or standalone); it is simply the process of converting ellipsoidal height to elevation. As explained, the fundamental relation is that elevation is ellipsoidal height minus geoid height, and if needed, minus the reference surface correction amount. 

Next, if you will use “height” in RTK real-time displays, always synchronize the result system and the geoid model. Geoid 2024 is described as data representing geoid height at arbitrary points given in latitude and longitude in Japan Geodetic System 2024. If the result system differs, the geoid model and correction assumptions applied may be inconsistent. In the field, confirm that known points are Geodetic Datum 2024 and unify the geoid model used for conversion under the same assumption. Neglecting this can lead to a situation where the RTK fixed solution is stable but elevation alone has a consistent offset. 

Moreover, in remote islands and target regions, operation including the reference surface correction amount is essential. For some remote islands, both geoid height and the reference surface correction amount are required for satellite-derived elevation, and the rule revisions specify handling for Okinawa Island and some remote islands. A common RTK mistake is applying mainland operational rules unchanged to remote islands. As a countermeasure, use the guidance and Q&A that the integrated file can be used nationwide to prevent configuration branching in the field. 

Finally, if you intend to use real-time displayed elevation as a record, always combine checks against known points with reproducibility checks at the same point. The transition of elevation systems, the provision of data, and the pre-use loading check are all prerequisites for ensuring operational quality. Real-time display is convenient, but in sites without verification procedures “just record it for now” proliferates, and later consistency cannot be achieved. Elevation correction becomes reliably usable in the field only after performing the cross-checks described in Step 4 at deployment. 

Making elevation correction a field routine with LRTK

To embed the five-step elevation correction workflow in the field, the last important point is to make a system that yields the same results regardless of who performs it. Even if people understand the theory, checks can be missed in busy field conditions and everything quickly falls apart. Failures due to mixing result systems, confusing geoid models, forgetting the reference surface correction amount, and file loading problems all occur in gaps in human procedures. The revision shows that elevation results are transitioning to a satellite-positioning-based framework and that data such as Geoid 2024, the reference surface correction parameters, and the integrated file are being provided. On that premise, field operations must routinize configuration, inspection, and record-keeping. 

Smartphone-mounted GNSS high-precision positioning devices like LRTK, designed to attach to a smartphone for positioning, make it easier to create operations that “complete” inspections and checks because they are convenient tools. The LRTK lineup includes devices that attach to a smartphone for easy positioning and are designed for one-handed operation. Creating a mechanism that does not omit the known-point checks of Step 4, confirming positioning results on the spot, and maintaining the conditions for elevation correction will ultimately reduce mistakes most efficiently. Elevation correction’s value lies more in field reproducibility than in understanding an equation. Use LRTK to incorporate elevation correction into standard field procedures so that RTK real-time displays can be used to handle height without confusion. 