The Texas Underground

How it is worked out

The depth figures are plain statistics of the driller reports filed with the state: the median and the middle half of the home wells around the point. The search widens from 2 to 15 miles until it finds at least 10 home wells, and always says how many it used.

For the formation, we take the ground elevation at each well from public terrain data, subtract the depth of the part of the well where water enters (the middle of the screened or open interval when the report has it, otherwise the bottom of the hole) and see which layer of the state's model that elevation falls in. Wells are reported by group of formations, because inside a group the model cannot tell the layers apart reliably.

In 7,977 wells in the area of the state's Northern Trinity and Woodbine model (North and Central Texas) where the State of Texas recorded the aquifer, our estimate of the formation the well draws from names the same layer of the state's model in about 74% of them, and that layer or the one next to it in at least 95%; among the 3,130 of those wells that the state places in a Trinity layer (Paluxy to Hosston) it is about 56%.

In 2,665 wells in the Texas Hill Country where the State of Texas recorded the aquifer, our estimate of the formation the well draws from names the same layer of the state's model in about 76% of them, and that layer or the one next to it in at least 98%.

In 14,333 wells in the area of the state's northern Gulf Coast Aquifer model (Houston and Southeast Texas) where the State of Texas recorded the aquifer, the formation our method says the well draws from (as grouped on these pages) is the one the state recorded in about 83% of them, and that formation or the one next to it in at least 99%; of the 3,578 of those wells that the state records in the Evangeline, it names the Evangeline for about 51% and the Chicot for most of the rest.

In 22,814 wells in the Texas part of the state's High Plains Aquifer System model (the Panhandle, the South Plains and the Permian Basin) where the State of Texas recorded the aquifer, the formation our method says the well draws from (as grouped on these pages) is the one the state recorded in about 75% of them; when it names the Dockum (4,542 wells), the state's record agrees only about 33% of the time, and most of the others are recorded as Ogallala wells.

The cost range is TerraFunded's estimate from published Texas driller prices ($25 to $60 per foot plus pump and equipment), applied to the median depth. Full methodology

What to keep in mind

It is the state's model, not a survey. Northern Trinity and Woodbine aquifers: quarter mile cells, summarized here to about 1.3 miles. Trinity Aquifer, Hill Country: 1 mile cells. Gulf Coast Aquifer System, northern part: cells of about 0.62 mile (1 km). High Plains Aquifer System (Ogallala, Edwards-Trinity, Dockum): half mile cells, read here on a grid of about 1.2 by 1.4 miles. Inside one cell the real rock can rise or fall. The layer under a lot is the likely one, not a certain one.

Thin layers and edges are approximate. In the northern model every layer is at least 30 feet thick, even where the real rock is thinner or absent.

The Antlers. North and west of Fort Worth (Parker, Wise, Montague, Cooke, Grayson and neighboring counties) the state maps the Trinity as one undivided unit, the Antlers. The model still draws its layers there, but the split is approximate: go by the depth ranges.

"Near or below the base". Some wells reach the bottom of the model or pass it, often by tens of feet: the base of the Trinity, of the Gulf Coast Aquifer or of the Dockum, depending on the model. Older rock that the model does not describe lies there. That is not an error.

Houston and Southeast Texas. The Burkeville, the Jasper and the Catahoula are reported together, because the state's records cannot tell them apart at a well. The line between the Chicot and the Evangeline is drawn differently by different studies, so the split between them is approximate: of the wells the state records in the Evangeline, our method finds the Evangeline for about 51% and the Chicot for most of the rest.

The High Plains. The Dockum is never a certainty here: where our method names it, the state's record agrees only about 33% of the time (61% for wells with a screened interval), so we say "at or just below the base of the Ogallala, in the top of the Dockum". The Edwards-Trinity group is the Cretaceous rock under the Ogallala: the Edwards-Trinity (High Plains) around Lubbock, the Edwards-Trinity (Plateau) around Midland and Odessa and the Rita Blanca in Dallam and Hartley. For old wells without a screened interval the method is only a little better than always answering the Ogallala (65% against 58%).

Younger formations. In the northern model the top layer lumps the Eagle Ford, Austin, Taylor and Navarro. The Washita and Fredericksburg layer includes part of the Edwards limestone in Williamson, Bell and Travis counties.

Water levels are measured, never modeled. The water level shown is what each driller measured when the well was finished. The models' own water levels are not used.

Depth is stretched in the figures. The cross-section and the 3D block stretch depth so the layers can be seen, and each one states by how much. The numbers are always written in feet.