Dataset guides

Texas well, pipeline and survey shapefiles — downloading RRC GIS layers and converting to GeoPackage or KML

The Commission publishes its map data as shapefiles, one zip per county, in a 1927 datum. Here is what is in the well, pipeline, survey and base layers, how to get a statewide layer out of 255 files, how to reproject and convert them — and when a GeoPackage of the parsed permit or wellbore data is the better answer.

If you are looking for a Texas well shapefile, there are two quite different things you might mean, and this article covers both. The first is the Railroad Commission's own GIS export — the well, pipeline, survey and base map layers it draws its public map viewer from, published as shapefiles county by county twice a week. The second is a spatial file built from the Commission's tabular data: the permits and the wellbore database carry coordinates, and EZRRC will build either as a GeoPackage, GeoJSON, KML or shapefile with your filters applied.

Which one you want depends on whether you need the Commission's cartography — the surface-to-bottomhole arcs, the survey grid, the pipelines — or a point per well with the full record attached. Often it is both.

The five county layer sets

RRC's digital map data is published under five headings, each as one zip per county:

The five county layer sets
Dataset Contents Geometry
Well layers Surface locations, bottomhole locations, and the arcs joining them for directional and horizontal wells Points and lines
Pipeline layers Pipeline centrelines with commodity, status, system type, operator and T-4 permit Lines
Survey layers Original patent survey lines and abstract tracts Polygons
Base layers Roads, hydrography, municipalities and political boundaries Mixed
All layers The four above in one zip Mixed

The well layer is the one people search for. It is the Commission's own rendering of where wells are: a surface point, a bottomhole point, and for a deviated well the arc between them, which is the quickest way to see lateral direction and length across a county. Well type and status are encoded in the symbology, and the symbol scheme is documented in the Digital Map Information User's Guide (OGA094) linked from each dataset page.

The pipeline layer is a single layer per county — pipe001l for Anderson, and so on — with liquid, gas and abandoned lines told apart by attributes rather than split into files: COMMODITY1 to COMMODITY3 and CMDTY_DESC for what the line moves, STATUS_CD for its status, SYSTYPE for the system type, plus the operator name and P-5 number, diameter and T-4 permit. Abandoned lines are in there; filter on STATUS_CD if you do not want them. These are the Commission's mapped routes, not an as-built survey.

The survey layer matters more in Texas than it would almost anywhere else. Texas is not on the Public Land Survey System, so a legal description here is an abstract number and a survey name, not a township and range, and this grid is the only way to put one on a map. Permits and leases describe position this way, which is why the layer is worth having even if you only care about wells.

The all layers bundle is exactly the other four in one zip and nothing more. We checked: for Zavala County it holds 160 shapefile parts, and the well (24), survey (32), pipeline (8) and base map (96) bundles hold those same 160 between them, matching on size and checksum. Take it if you want everything for a county; take the individual layer if you do not — the pipeline layer for a county is around 0.6 MB against 5 MB for the bundle.

All five are served as the Commission's original zips. Upstream has no reader for them yet, so there is no CSV of a shapefile here and no filtering; you choose a county and you get its file. There is also a companion statewide API listing as dBase, one .dbf per county, for loading straight into GIS software — with the ten-character column names DBF imposes.

Two things that go wrong before you draw anything

The datum is NAD27. Every RRC map layer is in geographic coordinates on the North American Datum of 1927 — EPSG:4267 — which is not what web mapping expects. Overlay it on anything modern without reprojecting to WGS 84 and Texas wells land roughly a hundred metres from where they belong, consistently, in a way that looks like a data problem and is not. Set the source CRS explicitly when you convert; do not trust a missing or generic .prj.

"Statewide" is 255 files. The layers are packaged county by county under RRC's 8.3 naming convention, one zip per county on the odd county codes 001 to 507, plus a FED bundle covering federal offshore waters. A statewide layer is a merge of all 255, and merging the 254 county files alone silently drops everything offshore. County codes are explained in RRC districts and counties; the short version is that the number in the file name is the three-digit Texas county code, which is also the FIPS code, so 329 is Midland.

Converting to GeoPackage, GeoJSON or KML

The tool for this is GDAL's ogr2ogr, which every GIS install and most package managers carry. Unzip a county, then reproject and convert in one step:

# One county's well layer (329 is Midland), NAD27 -> WGS 84, into a GeoPackage.
# Point ogr2ogr at the unzipped folder and every .shp in it becomes a layer.
unzip -q well329.zip -d well329
ogr2ogr -f GPKG midland_wells.gpkg -s_srs EPSG:4267 -t_srs EPSG:4326 well329

# The same thing as KML, for Google Earth
ogr2ogr -f KML midland_wells.kml -s_srs EPSG:4267 -t_srs EPSG:4326 well329

Point ogr2ogr at the folder and each shapefile inside becomes a layer of the output; point it at one .shp for one layer. -s_srs is the part people forget: if the .prj is missing or ambiguous, stating the source datum is what makes the reprojection correct rather than a no-op.

A statewide layer is the same command in a loop, appending. The shapefile names inside each zip carry the county code — pipe001l is Anderson's pipeline layer — so stripping the digits from the name gives one layer name per kind of shapefile, and every county's like-named shapefiles land in the same layer:

for f in well*.zip; do
  d="${f%.zip}"; unzip -q -o "$f" -d "$d"
  for shp in "$d"/*.shp; do
    layer=$(basename "$shp" .shp)
    ogr2ogr -f GPKG texas_wells.gpkg -s_srs EPSG:4267 -t_srs EPSG:4326 \
            -append -update -nln "${layer//[0-9]/}" "$shp"
  done
done

-append -update keeps adding to the same file rather than replacing it. Include the FED zip in the glob or you have lost the offshore wells. If you would rather not script it, QGIS opens a zipped shapefile directly, and its Merge Vector Layers and Export tools do the same job with clicks.

Two shapefile limits worth knowing before you convert the other way. A shapefile truncates column names to ten characters, so a round trip through one renames your fields; and a single shapefile component cannot exceed 2 GB, which a statewide well layer with attributes can approach. GeoPackage has neither problem, which is why EZRRC leads with it.

The other Texas well shapefile: build one from the parsed data

The Commission's layers are its cartography. If what you want is a point per well with the full record on it — every permit issued in a county last quarter, or every wellbore in a district with its depth and plugging status — you do not need a shapefile from RRC at all. Three parsed datasets carry coordinates and can be downloaded in any spatial format:

  • The daily permit file — surface and bottomhole latitude and longitude for every permit, in decimal degrees on WGS 84, parsed out of the 14 and 15 trailer records. Filter by county, operator, application type or issue date, then choose GeoPackage, GeoJSON, KML, KMZ, Shapefile or FlatGeobuf as the format. Downloading drilling permits walks through it.
  • The Full Wellbore Database — about 1.2 million wells with WGS 84 coordinates, plus every casing, perforation, formation and plugging record hanging off them. This is the closest thing to a statewide well point layer with real attributes.
  • Completions — the W-2 and G-1 filings, with coordinates where the filing carries them.

At the options step you can also reproject: WGS 84, NAD 83, Web Mercator, the Texas Statewide Mapping System, and the five NAD 83 Texas State Plane zones in US survey feet. If the file is going into a county appraisal system or a survey package, ask for State Plane there and skip a step later.

One honest limitation. Roughly three percent of permits have no surface coordinate at all, and a point without coordinates cannot go into a spatial file. EZRRC drops those rows from spatial exports and reports how many were dropped — Midland County is 9,548 permits but 9,187 points — so a GeoPackage and a CSV of the same slice will disagree by a few percent, and the CSV is the complete one.

Which to use, in one paragraph

Use the Commission's well layer when you want its arcs, its symbology, or a picture that matches the RRC map viewer; use its survey layer whenever a legal description has to become a location; use its pipeline layer for infrastructure. For a well dataset on a map — points with depths, dates, operators and status you can filter and join — build it from the permits or the wellbore database in GeoPackage and reproject at the options step. And whichever you pick, remember that the RRC layers are NAD27 and the parsed tables are WGS 84: put both on one map without reprojecting and the same well will appear twice, a hundred metres apart.

Next: the recipes — starting with downloading Texas drilling permits as CSV, Excel or a GeoPackage.