Water levels fall where no one can cap pumping — and rose 200 feet where someone did

What is happening?

In Texas the rule of capture gives the owner of the land the water beneath it, and the owner may use as much of it as desired so long as it is not wasted or taken maliciously. Groundwater provides 54 percent of the total water supply of Texas, and availability is expected to fall 25 percent by 2070, with the Ogallala portion projected to fall 50 percent over the same horizon.

The contrast is measurable inside a single aquifer. In Harris and Galveston counties a subsidence district created in 1975 forced permit holders off groundwater, and from 1977 to 2026 water levels in Regulatory Areas One and Two rose by more than 200 feet, while in Regulatory Area Three, where the conversion deadline runs to 2035, they fell by more than 300 feet and by more than 400 feet around The Woodlands.

Elsewhere the regulator arrives only after the decline is large. Arizona designated the Ranegras Plain as its eighth Active Management Area in January 2026, after water levels had dropped more than 200 feet since the 1980s and annual withdrawals had come to exceed inflows by more than 900 percent.

Whose problem is this?

The body that decides whether a cap is legally possible is a state legislature, and the cost of having no cap lands on individual well owners decades later and one basin at a time.

Where does this problem end?

The boundary matters because the resource is falling in places that already have a regulator and in places that do not. What separates the two is not rainfall but whether anybody holds the power to cap the volume withdrawn.

What is the state now, and what should it be?

Now

Should be

The only numeric targets located are local. The Harris-Galveston Regulatory Plan caps groundwater at no more than 10 percent of total water demand in Regulatory Area One, no more than 20 percent in Regulatory Area Two, and no more than 20 percent in Regulatory Area Three by 2035.

Needs a new measurementno national target was found for how much of the country should sit under an enforceable pumping limit, nor any federal figure for an acceptable rate of aquifer decline. The district plans set a share of demand, not a level the aquifer must hold.

How big is it?

Fills with researchthe affected population could not be derived from the sources opened. No document located gives a count of people living over basins that lack an enforceable pumping limit, and no fifty-state tally of which jurisdictions hold that power was found. The one large confirmed headcount, about 34 million people on subsiding land, is scoped to the 28 largest urban areas and includes Houston and Phoenix, both of which do have a regulator, so it measures a different population than this problem defines.

Under what conditions does it arise?

1. Ownership is assigned before the limit is. Where the water belongs to the owner of the land above it, a cap reads as a taking. Texas groundwater districts monitor and permit but stop short of telling landowners they may not pump, and litigation risk is named as a reason. 2. The regulator is created basin by basin, after the fact. Designation follows the decline rather than preventing it. The Ranegras Plain waited until 2026, by which time levels had fallen more than 200 feet since the 1980s. 3. Depletion arrives as depth rather than as an outage, so the meter that would trip never trips. A falling water table does not stop service. It raises the cost of lifting water and strands the shallowest wells first, and those are usually domestic ones. 4. Weather noise hides the trend. The Kansas High Plains aquifer rose 0.2 feet in 2025 on above-average rainfall, the first overall increase since 2019, against an average decline of 0.57 feet per year since 1996. Nothing institutional changed in that year.

What has been tried?

The attempts divide into two kinds, and only one of them touches the volume withdrawn. Measuring, assessing and reporting are now widespread, while the power to cap remains rare and is granted one basin at a time.

What was found?

Why is it still unsolved?

Institutional gap — the decline is measured well and governed badly, because in most basins no body holds the authority to cap the volume withdrawn.

A limit on how much anyone may pump is the one instrument that most of these places have never had. The monitoring layer is in good order. Kansas measures about 1,400 wells every year, Arizona assesses each of its 51 basins on a five-year cycle, a subsidence district in Texas operates a network of about 190 satellite positioning stations, and Maine now runs a public survey to collect reports of dry wells. Every one of those systems can describe the fall precisely. None of them, on its own, can stop it.

Where the authority does exist, the direction of the curve changes. The Harris-Galveston comparison holds the aquifer, the climate and the decade constant and varies only the rule, and the two sides diverge by more than 500 feet. That is why this is filed as an institutional gap rather than a physical limit or a cost problem. The water came back where somebody had the power to require that it be left alone.

The gap closes one basin at a time and only after the damage is visible. Arizona designated its eighth management area in January 2026, its third since 2022, after more than 400 comments and after one well in the basin had dropped 242 feet in 40 years. Texas leaves the question to a property rule written long before the pumps could reach this deep, and its local districts, which do have permits and monitoring, decline to tell a landowner that the pumping must stop.

What observation would mean it is solved?

Candidates — (a) water levels stop falling in basins that have no pumping limit (b) every basin in measured decline sits under an authority that can cap withdrawals (c) dry domestic well reports and subsidence rates fall together.

Each fails alone. (a) is satisfied by weather: Kansas recorded plus 0.2 feet in 2025 on above-average rainfall while nothing about the rules changed, and the southwest district still lost 0.62 feet in the same year. (b) confuses an authority with a limit — Texas groundwater conservation districts already exist, hold permitting power, and still do not cap the volume. (c) lags: clay compaction continues after pumping slows, and a dry well count only exists where someone collects it, so a state that starts counting will appear to get worse.

What is it connected to?

Fills with researchirrigated agriculture economics, drinking water access for households on private wells, land subsidence damage to infrastructure, and interstate surface water allocation all appear adjacent. Relation type and evidence grade were not confirmed in this round.

What these sources do not say

• How many states hold an enforceable limit on the volume pumped. This was the central question and no opened source answers it. The fifty-state picture had to be assembled here from three states, which is not the same thing.
• How many people live over a basin with no pumping limit. Without that, the affected population cannot be derived at all, which is why block 5 carries a marker instead of a number.
• How many Irrigation Non-Expansion Areas Arizona has. The count of management areas and the count of basins came from two different documents published two years apart, so the coverage share is assembled rather than reported.
• Whether Maine collected dry well reports before 2025. The figure of 541 has no earlier-year baseline in the source that carries it, so it cannot be read as a rise or a fall against other drought seasons.
• What either single-site result means anywhere else. The Kansas measurements cover the Kansas portion of the High Plains aquifer only and are never extended to the whole Ogallala, and the Harris-Galveston recovery rests on converting users onto surface water drawn from three river basins, with nothing said about whether an inland agricultural basin has any comparable alternative to convert to.
• What a dry well costs a household. Deepening, replacing or hauling water are all named as consequences and none of them is priced.

See the evidence

The Harris-Galveston executive summary was read directly as a published document, and it is the only primary agency report opened in this round. Everything else is secondary reporting or an institutional news page, which is why the evidence tier stays at secondary. The Ranegras Plain figures were confirmed twice, once through the university water center and once through the state water news service, and the two accounts agree on the eighth designation and on the scale of decline while only the second carries the 242 feet and the 900 percent. The Maine figure of 541 comes from the state agency page dated 2025-12-19; a higher year-end figure appeared in secondary search results that were not opened, and it is not used here. The count of Arizona basins is two years older than the count of management areas, so the ratio of 8 to 51 is assembled from two documents rather than taken from one. Nothing in the Kansas measurement is extended beyond Kansas.

People affected

Estimated range Not derivable

The reason and what is missing are listed under “What is missing” below