Geographic data is comprised of geographic features – such as wells, fields or licence areas – located by coordinates, referenced to the earth by a coordinate reference system (CRS). The management of geographic data is reliant upon understanding CRSs and successful management of the data’s CRS properties.
The potential impact on oil and gas businesses caused by CRS mismanagement and inaccuracies can be high:
- Disputes over licence boundaries as definitions are specified as coordinates
- Inaccurate risk assessments in relation to drilling operations
- Potential damage to existing infrastructure
- Difficulties integrating geographic data with other E&P datasets.
The Esri ArcGIS product suite, in common with most GIS systems, can visualise and manage geographic data specified by coordinates referenced to the earth in different CRSs. The desktop mapping component of ArcGIS Desktop, ArcMap, can correctly align layers of geographic data referenced and stored in different CRSs within the same data frame. This functionality is often referred to as “projecting on-the-fly” and ArcMap undertakes coordinate operations to align layers of geographic data without changing the underlying data itself. To facilitate this, each geographic dataset in ArcGIS has a CRS property indicating the CRS in which the coordinates are referenced to the earth (or “unknown” if the CRS is either not known or has not been set).
The dialog above is from ArcCatalog and displays the CRS property of a feature class. In this case, the CRS property indicates the coordinates of the feature class are referenced in European Datum 1950.
Exporting data for use in other technical systems
Converting a dataset’s feature coordinates to a different CRS is a common requirement of a GIS. The data can then be exported and used in other technical systems that are only capable of displaying data referenced in the same CRS.
As can be seen above, the Feature Class Properties Dialog allows the CRS property of the data to be changed using the Select, Import, New, Modify and Clear buttons. So does this convert feature coordinates of the data to a different CRS?
Well, no! It only updates the CRS property of the dataset – it does not alter the actual values of feature coordinates. As identical coordinate values referenced in different CRS do not refer to identical locations on (or under) the earth, the outcome is ArcGIS will incorrectly interpret the CRS and the data will be misaligned. Effectively, the features will be incorrectly located.
The magnitude of the misalignment depends on the CRSs and the region involved. For example, with two geographic CRSs, WGS 1984 and European Datum 1950, coordinates in the region of the Brent oil field in the North Sea are misaligned by ~110m – a significant amount if the GIS is the definitive source of truth for subsea infrastructure.
With projected CRSs, resetting the CRS property from ED50 / UTM zone 31N to ED50 / UTM zone 32N can lead to gross and significant errors. The location of the Brent oil field is found around 320 km away on the Norwegian mainland.
To make things worse, changing the CRS property in this way does not leave an audit trail in the dataset’s metadata record, making it difficult for future users to diagnose and trace the positioning errors.
Changing the CRS of a dataset’s coordinates
There are two ways to convert feature coordinates to a different CRS using ArcGIS. Both methods create a new dataset and do not change the original dataset.
1. ArcMap’s Export Data functionality.
First, the CRS of the data frame should be set to the target CRS, then the data exported using the same CRS as the data frame. If a coordinate transformation is required, this must be set in the data frame prior to exporting the data, otherwise the exported data may be incorrectly located.
2. ArcToolbox’s Project (for vector data) or Project Raster (for raster data) tools.
Enter the input dataset, output dataset, target CRS and, if required, a coordinate transformation. The advantage of using this method is the tool automatically detects whether a coordinate transformation is required. It also supports concatenated coordinate transformations (transformations requiring more than one step) and the tool can also be embedded within automated workflow models or scripts. Raster datasets require additional parameters – such as the resampling method – but the overall approach is the same.
The Project tool and Project Raster tool can be found in the Projections and Transformations toolset in the Data Management toolbox.
Summary
It is vitally important to understand the difference between changing the CRS property of a geographic dataset and converting the feature coordinates of a geographic dataset to a different CRS. Confusing them can lead to incorrectly located data and potentially some gross errors in the position of the features in your geographic data.
Posted by Ian Milligan, Senior Consultant, Exprodat.