Standards for Hydrographic Surveys

8 Data Attribution

8.1 General

To allow a comprehensive assessment of the quality of survey data it is necessary to record or document certain information together with the survey data. Such information is important to allow exploitation of survey data by a variety of users with different requirements, especially as requirements may not be known when survey data is collected.

The process of documenting the data quality is called data attribution; the information on the data quality is called metadata. Metadata shall comprise at least information on:

  • The survey in general as e.g. date, area, equipment used, name of survey platform
  • The geodetic reference system used, i.e. horizontal and vertical datum; including ties to WGS 84 if a local datum is used
  • Calibration procedures and results
  • Sound velocity
  • Tidal datum and reduction
  • Accuracy achieved and the respective confidence levels.

Metadata shall preferably be in digital form and an integral part of the survey record. If this is not feasible, similar information shall be included in the documentation of a survey such as the final field report.

8.2 Point Data Attribution

All soundings should be attributed with a 95% statistical error estimate for both position and depth. Although this should preferably be done for each individual sounding, a global estimate will be provided for an entire dataset and the worst-case survey error must be shown.

In the case of positions, they shall be qualified by analyzing redundant lines of position (terrestrial systems) or independent positioning check (satellite systems); in the case of depth observations, they could be qualified by analyzing redundant depths observed at, for example, check line crossings.

It is understood that each sensor (i.e. positioning, depth, heave, pitch, roll, heading, seabed characteristic sensors, water column sensor parameters, tidal reduction sensor, data reduction models etc.) possesses unique error characteristics. Each survey system shall be uniquely analyzed to determine appropriate procedure(s) to obtain the required spatial statistics. For more information, see the Survey Management Guidelines.

8.3 Depth Accuracy

Depth accuracy is to be understood as the accuracy of the reduced depths. In determining the depth accuracy, the sources of individual errors need to be quantified. All error sources shall be combined to obtain a Total Propagated Uncertainty (TPU). TPU results from the combination of all contributing errors, which include among other things:

  1. measurement system and sound speed errors
  2. tidal measurement and modeling errors, and
  3. data processing errors.

A statistical method for determining depth accuracy by combining all known errors shall be adopted and checked. For example, CUBE processing can provide this information.

The TPU, determined statistically at the 95% confidence level, is the value used to describe the depth accuracy achieved. The TPU shall be recorded together with the sounding value.

Recognizing that there are both constant and depth dependent errors that affect the accuracy of depths, the formula under Table 1 is to be used to compute the allowable depth errors using a and b from row 3.

8.4 Geostatistics

When the seabed has not been totally searched during a survey, the soundings only provide samples of the seabed at discrete points. In such a case, it is necessary to interpolate depths derived from soundings to obtain a bathymetric model, which provides an estimate of depth information over the entire seabed surface. Geostatistical interpolation techniques may be used to estimate the error introduced by interpolation between soundings, taking into consideration the accuracy of reduced depths and positions as well as the spatial distribution of depth measurements. Using the values for a and b from Table 3 below, the formula under Table 1 is to be used to compute, at 95% confidence level, the allowable errors for the bathymetric model. If these errors are exceeded, the density of soundings shall be increased.

Table 3 Bathymetric Model Accuracy

ORDER Exclusive Special 1 2 3
Bathymetric a = 0.02 a = 0.50 a = 1.0 m a = 2.0 m a = 5.0 m
Model Accuracy (95% Confidence Level) b = 0.01 b = 0.01 b = 0.026 b = 0.05 b = 0.05

These interpolation techniques, based on an appropriate statistical error analysis that quantifies the roughness of the seabed, shall not be used as the only means to assess the quality of a survey, as they may not provide reliable estimates of the accuracy of the bathymetric model in all cases; particularly, if surveys were conducted with excessive line spacing or if there is a high likelihood that man-made features exist.

8.5 Error Sources and Budget

Although the following text focuses on errors of data acquired with multibeam systems, it should be noted that it is in principle applicable to data acquired with any echo sounding system.

With multibeam and multi-transducer echo sounding systems, the distance between the sounding on the seafloor and the positioning system antenna can be very large, especially in deep water with a wide swath system. Because of this, sounding position accuracy becomes also a function of the gyrocompass heading accuracy, beam angle (or transducer location for sweep systems) and the water depth (swath systems only).

Roll and pitch errors will also contribute to the relative error of the sounding from the transducer. Overall, it may be very difficult to generalize what is achievable as typical position accuracy for each sounding as a function of depth in some of these modern systems. The errors are not only a function of the echo sounder but also the vessel and the location and accuracy of the auxiliary sensors.

The use of non-vertical beams introduces additional errors caused by incorrect knowledge of the ship's orientation at the time of transmission and reception of sonar echoes. Errors associated with the development of the position of an individual beam must include the following:

  1. positioning system error,
  2. depth measurement error,
  3. the uncertainty associated with the ray path model (including the sound speed profile),
  4. the accuracy of the vessel heading,
  5. the accurate identification of system pointing errors resulting from transducer misalignment as determined by patch test,
  6. vessel motion sensor, i.e. roll, heave and pitch accuracy, and
  7. time latency.

Project managers responsible for the survey quality shall develop and document error budgets for their particular systems.

Date modified: