EXPOSED
A hydrologist's week splits between screen work — running MODFLOW/HEC-RAS models, processing gauge and well data, writing permit reports and impact assessments — and field work installing instruments, sampling wells, and surveying streams. The modeling and report tiers are exactly where AI is fastest: calibration scripting, literature synthesis, and regulatory boilerplate are already largely automatable. What persists is field judgment about messy sites, defending conclusions to regulators and courts, and owning the call when a model says a floodplain or aquifer is safe.
Part 2020 shock, part continued decline in the years since.
Median pay $81,270 → $96,600 -4.9% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
-0.1% 6,300 → 6,300 on the projections basis
Shrinking, but not obviously because of AI
The BLS projects -0.1% by 2034, but at 48/100 this work is only moderately exposed — not the profile of a job current AI can simply do. Occupations shrink for many reasons, and the score does not point at automation as this one's cause.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~500 openings a year on average, including replacing people who leave.
ScientistHydrologistSeismologistVolcanologistHydrogeologistHydraulic EngineerPhysical ScientistHydrologic EngineerIsotope HydrologistSurface HydrologistResearch HydrologistGroundwater ConsultantProject HydrogeologistWater Resources ScientistSource Water Protection Specialist
Holding it up: judgment & accountability . Weakest point: trust premium .
Mixed — a routine tier and a judgment tier Model calibration runs, rating-curve fitting, USGS gauge data QA, and the standard-format sections of a 404 permit or groundwater impact report can be scripted or generated today, but siting a monitoring well network on a specific site, deciding which aquifer test failed and why, and constructing a conceptual model of a fractured-rock system from sparse borehole logs still needs a hydrologist thinking about that basin — hence 11, not the 5 a pure data-processing analyst would get.
Some physical or field component Field weeks are real — wading streams for discharge measurements, bailing and purging wells, deploying pressure transducers and dataloggers, storm-event sampling in weather that does not wait — but they are a minority of billable hours against modeling and reporting, and much of the continuous record now arrives by telemetry, which is why this sits at 11 rather than up with a drilling crew.
Certification preferred, not legally required Most states have no hydrologist licence; you sign as a Professional Geologist, PE, or Certified Professional Hydrologist where the state requires a stamp on a well permit or dam safety report, and plenty of work is issued under a firm's or an agency's name instead — a real but partial gate, well short of the personal, non-delegable exposure of a licensed structural engineer.
Meaningful discretion You choose the design storm, the boundary conditions, the recharge assumption, and the safe yield number — calls where the data underdetermines the answer and the consequence is a floodplain designation, a contaminant plume capture zone, or a withdrawal permit — but they land inside FEMA, EPA, and state regulatory frameworks that constrain method and get peer-reviewed before they bind anyone, which keeps this at 12.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (11/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (7/20) is whether the law requires a licensed human to sign. Trust premium (7/20) is whether buyers specifically pay for a person. Judgment and accountability (12/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 26 of this occupation's 48 points (54%).
Embodiment (11/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
No occupation passed every test: close enough to hydrologists on skills and subject matter, at least 10 points more resistant, no big jump in training, no new licence, no pay cut, and not shrinking on its own. That happens for 223 of the 654 occupations here that aren't SAFE, and it is worth stating plainly rather than leaving the section off.
The usual reason is that exposure travels with the skill profile. The jobs most similar to yours tend to be exposed for the same reasons yours is, so the near neighbours don't clear the gap — and the ones that do are a different kind of work, not a transfer of what you already know. Read that as a limit of this method, not a verdict that you're stuck: it only compares whole occupations, and it cannot see specialisation, industry, or anything you'd bring that isn't in a federal skill survey.
Here is that claim on your own job rather than in the abstract. These are the three occupations closest to this one by skill and subject matter — the places the work would most naturally transfer — with what the register scores them:
That is the whole problem in three lines. The nearest work is not meaningfully safer, so there is no move here that trades a similar skill set for a better verdict. This is not us running out of ideas — it is what the neighbourhood looks like.
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 66/100, still EXPOSED.
Task-mix shift is genuine here: the occupation has a clear routine tier (calibration scripting, gauge QC, permit boilerplate, literature synthesis) and a judgment tier (deciding whether a site conceptual model is even right, defending a floodplain delineation under Daubert challenge in water-rights adjudication or flood-litigation, calling an aquifer safe for a municipal supply). If the routine tier is largely automated, the residual role is dominated by consequential calls under ambiguity, and this dimension rises without any new law.
Same two-tier logic: once scripting and report drafting are absorbed, remaining hours concentrate in conceptual-model formulation from ambiguous field evidence, expert testimony preparation, and negotiating assumptions with opposing consultants — work current systems cannot do at usable quality. Note the countervailing risk: this shrinks headcount even as per-worker resistance rises.
FEMA's Letter of Map Revision (MT-2) and Conditional LOMR packages already require a registered PE or licensed land surveyor's seal on hydraulic analyses; extending equivalent seal requirements to groundwater work — e.g. more states adopting Professional Geologist/Geologist-in-Charge licensure (currently ~30 states, ASBOG-based) with an explicit requirement that aquifer test interpretations, wellhead protection delineations and CERCLA/RCRA groundwater fate-and-transport models carry a named licensee's stamp — would put personal liability on a human for exactly the model outputs AI generates. Watch state geology board rulemaking and any board statement that AI-generated modeling output constitutes unlicensed practice unless sealed after independent review.
State dam-safety and water-rights agencies requiring a licensed professional's certification on model files themselves, not just the report — e.g. Texas TCEQ or California DWR conditioning permit acceptance on a signed statement that the modeler personally verified calibration, boundary conditions and parameter choices. Parallel to engineering boards' emerging positions on AI-assisted design documents.
Narrow route only: litigation and adjudication markets (interstate compact disputes, Colorado River and Rio Grande accounting, CWA/CERCLA cost-recovery cases) where courts and opposing counsel require a testifying human whose credentials and cross-examinability are the product. Federal Rule of Evidence 702 amendments (2023) tightening expert-basis scrutiny reinforce this. No plausible route to a general public-facing human premium for permit or consulting work.
The limit. Realistic ceiling is mid-60s. Even with full PG-seal coverage, hydrology's liability shield is weaker than engineering's because much work is federal-agency in-house or consultant deliverables reviewed by agency staff rather than seal-gated. The bigger threat is not per-worker capability but volume: a 5,850-person occupation where the routine tier absorbs most billable hours can see the score rise while positions fall.
| Phoenix-Mesa-Chandler, AZ | 260 | $90,000 -7% |
| Minneapolis-St. Paul-Bloomington, MN-WI | 210 | $94,580 -2% |
| Columbus, OH | 170 | $98,200 +2% |
| Denver-Aurora-Centennial, CO | 140 | $108,680 +13% |
| Seattle-Tacoma-Bellevue, WA | 140 | — |
| New York-Newark-Jersey City, NY-NJ | 130 | $88,100 -9% |
| Sacramento-Roseville-Folsom, CA | 110 | $106,860 +11% |
| Austin-Round Rock-San Marcos, TX | 100 | $73,810 -24% |
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 50 | $152,730 +58% |
| San Francisco-Oakland-Fremont, CA | 50 | $147,080 +52% |
| San Jose-Sunnyvale-Santa Clara, CA | 30 | $142,640 +48% |
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 48. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.