$175,900top of the range in California · middle $99,070 / yr
AI is transforming this role
Climate Scientists in the United States earn a median of $99,070 a year. Pay starts near $53,060. Pay reaches $175,900 at the top of the range in California, the best-paying state for this work among those with at least 500 people in the job.
Source: U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics, May 2025 (Atmospheric and Space Scientists, SOC 19-2021). Last checked 9 September 2026.
Entry level
$53,060
Top of the range · California
$175,900
Education
Master's or Doctoral degree
Wages — U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics, May 2025 (Atmospheric and Space Scientists). Top of the range is the highest state-level figure among states with at least 500 people in the job. AI-impact rating is PayCrunch's editorial assessment. Updated September 2026.
🆕 New & Trending AI Tools for Climate ScientistReviewed September 2026
We track new AI-tool launches every week and refresh this list — here’s what’s gaining traction for Climate Scientist work right now.
Julius AINEWFree / $20 mo
AI data analyst that runs statistics and charts from plain-language prompts.
How a Climate Scientist uses it: analyze datasets and generate figures without writing code
NotebookLMNEWFree / $7.99 mo
Google tool that answers questions grounded only in the documents you give it — with citations.
How a Climate Scientist uses it: load your own manuals, policies, or PDFs and ask questions that stay accurate to the source
ElicitFree / $12 mo
AI research assistant that finds and summarizes papers.
How a Climate Scientist uses it: run a literature review and extract findings across dozens of papers fast
ConsensusFree / $9 mo
AI search that answers questions from peer-reviewed research.
How a Climate Scientist uses it: get evidence-backed answers with the studies behind them
SciSpaceFree / paid
AI that explains papers and helps with literature review.
How a Climate Scientist uses it: decode dense papers and trace citations quickly
SciteFree / $20 mo
Shows whether other studies support or contradict a paper's claims (Smart Citations).
How a Climate Scientist uses it: check if a finding is actually backed by the wider literature before you cite it
ChatGPTFree / $20 mo
The most-used AI assistant — writing, analysis, research, and images from a plain-language chat.
How a Climate Scientist uses it: draft emails and documents, summarize long files, and get instant answers to on-the-job questions
ClaudeFree / $20 mo
AI assistant known for careful writing, long-document analysis, and coding.
How a Climate Scientist uses it: analyze big reports or spreadsheets and turn messy notes into clean, finished writing
Google GeminiFree / $20 mo
Google's AI assistant, built into Gmail, Docs, and Search.
How a Climate Scientist uses it: draft and reply inside Google Workspace and research without leaving the page
You want the long record and the model run, the kind of science that argues about decades, and you are ready to leave a job that only needed tomorrow's forecast. Climate science is that longer argument. You study how the atmosphere, the ocean, and the land surface behave, you test ideas in models and in observations, and you write the result so another scientist can challenge it. A career change from physics, ecology, oceanography, statistics, software, or a forecast office can work. The door into a research seat is still a graduate degree, and there is no licence to hide behind.
Research weeks, models, and the record
A research week is unequal on purpose. One day is code and a model experiment. The next is a dataset that will not line up with the run. The next is a seminar where someone asks why your region looks different from the global mean. You might work with satellite records, weather-station histories, ocean measurements, or paleoclimate evidence such as ice and sediment, depending on the group. You might run a global model, a regional model, or a simpler calculation that exists to test one mechanism. The product is a figure, a paper, a technical report, or a dataset with a methods note honest enough that a stranger can rerun the idea.
The tools are a programming language the group already trusts, files of gridded data, a cluster or a cloud allocation when the run is large, versioned code, and statistics you can explain without hiding the assumption. The people are other scientists, a principal investigator who owns the grant, graduate students and postdoctoral researchers if you are in a university, program managers if you are in an agency, and sometimes a stakeholder who needs a risk number for a coastline or a crop. The decisions are which experiment is worth the compute, which comparison would falsify your claim, and when a result is solid enough to show outside the group. Career changers who can already code still have to learn which comparison the science considers fair. Career changers who know the science and freeze at a script will stall in the same way.
Places shape the week. A university group lives on papers, teaching, and the next proposal. A national laboratory or an agency lab such as one inside NOAA or NASA lives on missions, model development, and products other offices use. A private firm that prices climate risk for insurers, banks, or infrastructure lives on a deadline and a client. The underlying science can rhyme. The definition of finished work does not. Ask, before you apply, whether success this year is a paper, an operational dataset, or a client deliverable.
Separate from the daily forecast desk
Research seat, forecast office
A forecast office produces the daily weather the public uses. A climate research seat studies longer behavior of the atmosphere, ocean, and land surface, and tests that behavior in models. The skills overlap. The job you are switching into is the research seat.
If you already forecast, you know observations, model guidance, and how to speak carefully when the public is listening. That discipline transfers. What changes is the question. You stop owning tomorrow morning's high and start owning a claim about variability, trends, extremes, or a process inside a model. If you have never forecast, you do not need to become a television meteorologist to do climate research. You do need enough atmospheric or earth-system literacy to know when a beautiful plot is physically silly. Take that literacy from graduate coursework and from the group's own seminars, and be humble in the first year when an old forecast hand corrects your reading of a map.
The degree a research seat expects
A graduate degree is the usual requirement for a research seat. A master's can open applied and support roles, and some agency jobs, especially if your thesis already dealt with climate data. A doctorate is the common door to leading a research project, advising students, or holding a scientist title in a lab. There is no state licence for this work. Employers treat the degree, the analysis you can defend, and the papers or technical reports with your name on them as the proof. A portfolio of private Jupyter notebooks that never met a review will not substitute for that record, though the notebooks are how you get the record.
Preparation is the degree itself: coursework in the earth system or a neighboring science, a thesis with a real dataset or a real model, and the habit of writing methods while you still remember the choices. If you are changing careers and you already hold a doctorate in a neighbor field, a postdoctoral project inside a climate group can be the bridge. If you hold a bachelor's in software or statistics, the honest path into a research seat runs through graduate school, sometimes after a year as a research assistant or data specialist who learns the science from the inside. Say which of those you are. Hiring managers can smell a candidate who wants the title and hopes the degree is optional.
Choose the group for the question, not for the logo. Read two recent papers from the adviser or the lab lead. If you cannot explain what they trusted the model to do, you are not ready to write the application essay. If you can, your essay should say so in concrete nouns: the dataset, the experiment, the uncertainty you care about. Leave slogans about saving the planet for a sentence at the end, if you need them at all. The reader is trying to picture you in the lab meeting.
How a group decides to take a career changer
University hiring for students is an application, a statement, and letters. Hiring for a research assistant or a staff scientist is closer to an industry interview plus a science talk. Agency jobs add a federal-style application in which your resume must echo the duties in plain language. Private risk firms interview like technology companies and then ask you to critique a figure. In every case, bring one analysis you can discuss from raw idea to caveat. Strip anything your current employer owns. A small public dataset and a clear plot will beat a mysterious claim about a proprietary model you cannot show.
Expect to be asked what would change your mind. Climate research is full of results that look dramatic and fail a second test. Candidates who defend a plot as if it were an identity lose to candidates who name the next check. Also expect a question about collaboration. Models and observations are usually group sports. If your old job rewarded lone heroics, practice saying how you share code, credit, and bad news early.
Funding is part of the adult version of this job. Principal investigators write proposals. Staff scientists support them and sometimes lead work packages. If you want that life, learn to read a call for proposals and to sketch a work plan with tasks, data, and a deliverable. You can start that practice during graduate school by helping a mentor for one cycle. The skill is scarce, and it is how research seats stay open.
From the first analysis to leading the project
The path often runs from research assistant or graduate student, to postdoctoral researcher, to scientist or assistant professor, to the person who leads a group. Some people leave the academic ladder for an agency career with a clearer grade structure. Some move into climate services, where the job is to translate model output for a city, a utility, or a company. Some join a private risk team and never write another journal article. All of those can be honorable. The spine is the same: you remain someone who can say what the evidence supports and where it stops.
What moves you up is a result other people use. A dataset with documentation, a model improvement that stays in the code, a paper that changed how the group frames a problem, a briefing that kept a non-scientist from misreading the figure. Keep a short list of those. Promotion and the next job will ask for it. Teaching, outreach, and review work matter in universities. Operational reliability matters in agencies. Client clarity matters in firms. Pick the scoreboard of the place you want, and stop apologizing for it.
While you are still in the old job, build one public analysis you can talk about without betraying an employer. Pick a published dataset, document every choice, and write a page on what would make the result collapse. That page is your interview. If you need coursework, take the atmospheric or statistics classes your target program actually lists, rather than a random sampler of climate headlines. Learn to read a methods section until the experiment is visible. The career change fails when the enthusiasm is wide and the one figure you can defend is missing.
Communication is part of the job earlier than people expect. A city staffer, a journalist, or a program manager will ask what a result means for a decision. Your answer has to include the uncertainty in words a non-specialist can repeat accurately. Practice that with a friend who does not work in science. If they leave believing the model "proved" something you never claimed, rewrite the sentence. Research seats go to people who can hold that line in a meeting, not only in a comment thread.
An offer on the atmospheric and space science chart
If a research group offers you a climate-scientist salary, read it against the May 2025 Occupational Employment and Wage Statistics chart for atmospheric and space scientists, SOC 19-2021, a series wider than climate research alone. Entry on that chart is $53,060. The national median is $99,070. The gap from entry to median is $46,010. That gap is large, and for a career changer it often tracks the distance between a first professional or support role and the middle of a field that expects graduate training. Use $53,060 when the seat is an early analyst role and your climate research record is just beginning. Use $99,070 when you already hold the graduate degree and you can defend a body of work, and name $46,010 as the span between those two descriptions.
California's high end on this chart is $175,900, the top of the published range where the Bureau releases a figure for the job. That figure is the top of the range in California. It is separate from the national median. The gap from the median to that high end is $76,830. Use $175,900 when the role is a senior scientist, a lab lead, or a scarce specialist in that state, and you can point to scope that matches the top of the range. A first postdoctoral or analyst offer that opens with $175,900 is using the wrong end of the chart. These facts include no separate list of state medians and no employment count, so do not invent a typical California paycheck or a headcount to decorate the talk. The California number you may cite is the high end, $175,900, and only for a job that belongs there.
In the negotiation, say what the offer is buying: a support analysis, an independent research program, or a client-facing risk product. Match $53,060, $99,070, or $175,900 to that scope. Mention that the chart covers atmospheric and space scientists only in the one comparison above, then stay with the dollars and the duties. If the employer is outside California, keep $175,900 out of the ask unless you are truly discussing that state's high end for a role you would do there.
Pick one paper in the corner of climate science you hope to join, read the methods until you can say what the model was trusted to do, and aim your next application at that kind of group. If you cannot yet say it, the next step is the reading, or the graduate program, before you borrow the title.
The top of Climate Scientist pay — and how to get there with AI
$175,900what Climate Scientist pay reaches in California
Highest state-level top-of-range annual wage for Atmospheric and Space Scientists, among states with at least 500 people in the job. U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics, May 2025.
And the role it leads to — Physical Scientists, All Other — reaches $225,810 in Colorado.
$53,060entry$99,070middle$175,900top end
Toward the top of this occupation you stop being handed questions and start choosing them — which records get built, which timescale the organisation is willing to fund, and who else works on it.
Most of this job is upstream of the science: pulling observations from surface stations, upper air stations, satellites and radar, cleaning them, and getting them into a form a model can eat. Scripting that in Perl or C++ on Linux, and letting Claude review the script and write the edge cases you would not have thought of, can hand back a large share of a week. The trap is spending the returned hours on more of the same requests. People who reach the top of the range spend them on something the employer cannot buy elsewhere — a paleoclimate reconstruction, a long homogenised record, an air quality analysis with a defensible uncertainty statement — and then use the managerial half of the role, schedules and staff training and matching expertise to situations, to put other people on it too.
Your playbook, by where you are now
Just startingAutomate the intake, and count what you saved
Write down how long each recurring data gather takes you this month, station by station and source by source, before you change anything.
Script the ingest and the quality checks on Linux so a rerun is free, and keep the scripts under version control from day one.
Have Claude read your script back to you and list what it would do wrong on a missing day, a duplicated sounding or a units change, then fix those cases yourself.
Move your routine statistics out of hand-run steps into a saved IBM SPSS Statistics job so results reproduce exactly.
Re-time the same tasks after three months and write the difference in hours in a note you keep.
What proves it: A reproducible ingest and quality-control pipeline for one data source, with a before-and-after time record.
Realistic span: months one through eighteen
A few years inSpend the reclaimed hours on a longer record
Pick a question your employer cannot answer today because nobody has assembled the record for it, and assemble it.
Work paleoclimate or long instrumental series into your analysis so your conclusions carry a timescale the daily product cannot.
Learn one air quality modelling package well enough to run it unsupervised, so you can take work outside the forecast desk.
Build your standard map products once in ESRI ArcInfo and stop redrawing them, so a new question costs analysis time rather than graphics time.
Write one scientific climate report a year that stands on its own, and give the talk that goes with it.
What proves it: A published climate report built on a record you assembled, cited internally when the question comes up again.
Realistic span: roughly the third through the seventh year
ExperiencedConvert the record into a brief you staff
Put the hours you saved and the work they produced side by side, in writing, before you ask for anything.
Ask for a named line of work rather than a title, and say what it would produce in its first year.
Take the scheduling, the staff training and the expertise-matching that go with running an office, so you decide who works on which record.
Turn your ingest scripts into the team's standard, and train each new analyst on them in their first month.
Do the public work as well — the interviews, the questions from non-scientists — because the brief only survives if people outside the group can describe it.
What proves it: A funded programme with your name on it and at least one other climate scientist working inside it.
Realistic span: from about year eight
The next 90 days
Take the single most repetitive thing you do — for most climate scientists it is gathering data from surface and upper air stations, satellites and radar and forcing it into a common format — and time it honestly for two weeks. Then script it end to end, including the checks you currently do by eye, and have a model review the script for the failure modes you skipped. When it runs clean, do not absorb the free time. Write down what it bought you, in hours per month, and start one long-record analysis you have wanted to do and never had a week for. In three months you will hold two things almost nobody in this field holds: a measured throughput gain and a piece of science that only exists because of it. That pair is the argument for scope.
Wage figures: BLS OEWS, May 2025. The playbook is PayCrunch editorial guidance, not a guarantee of pay or placement.
Every figure is the national median from the U.S. Bureau of Labor Statistics (OEWS) shown on that role’s own page.
Never used AI before? Start here (2 minutes).
Start with Google Earth Engine and a general AI as your coding copilot.Earth Engine (free for research) puts petabytes of satellite, reanalysis, and climate data one script away, and ChatGPT or Claude writes and explains the Earth Engine and Python (xarray) so you spend time on the science, not the syntax. Together they take you from question to map or time series in an afternoon.
For learning, lean on the free stack: the Pangeo ecosystem and Jupyter for big geoscience data, Elicit or Consensus for literature, and open ML weather models (GraphCast, FourCastNet) you can study and run. Always validate AI and ML output against physical models and observations.
The one rule, forever: AI weather and climate emulators are fast approximations, not physics — validate every output against established models, reanalysis, and observations, and never present a projection without its uncertainty and assumptions. Don't let a model extrapolate beyond its training regime unchecked, guard against fabricated citations in AI-written text, and be precise about confidence when your analysis informs public-safety or investment decisions.
The plays — exact steps, exact prompts
Do these in order. Each one is copy-paste ready. You do not need to know anything about AI going in.
1
Run and validate ML weather and climate emulators
Why this pays: ML emulators like GraphCast are reshaping forecasting and climate services — the scientists who can run and validate them are in demand across national labs and the private sector.
Run an open ML weather model on reanalysis data and benchmark its skill against a physical model for your variable and region.
2
Write the evaluation pipeline with your copilot.
Copy-paste this prompt
I want to evaluate an ML weather emulator ([GraphCast / FourCastNet]) against [ERA5 or a physical model] for [2-meter temperature over a region]. Write the Python/xarray workflow to load both, regrid to a common grid, compute skill metrics (RMSE, ACC, bias) by lead time, and plot the comparison. Note the caveats in interpreting emulator skill.
Emulators approximate physics — always validate against reanalysis and observations and report where skill breaks down. Never present output without uncertainty.
What you'll haveCredible command of the tools redefining the field — the skill set behind top climate-services and private-sector pay.
2
Do geospatial and remote-sensing analysis at scale
Why this pays: Turning satellite archives into impact metrics — heat, drought, flood, emissions — is the core deliverable for climate-risk clients: high-value, in-demand work.
Google Earth EnginePython (xarray/rioxarray)QGIS
1
Use Google Earth Engine to compute a climate or impact indicator over a region and time period, then export it for analysis.
2
Have AI write the Earth Engine code for your specific indicator.
Copy-paste this prompt
Write a Google Earth Engine script to quantify [urban heat / drought / flood exposure] over [area] from [Landsat / Sentinel / ERA5]: the indicator definition, preprocessing and cloud masking, the time-series computation, trend analysis, and how to export results and visualize them for a stakeholder report.
Validate indicators against ground data where possible, and document data sources and limitations for any client-facing result.
What you'll haveSatellite-scale impact analysis on demand — the deliverables that justify private climate-risk salaries.
3
Build climate-risk analytics for finance and insurance
Why this pays: The private climate-risk market — insurers, banks, real assets — pays well above academia. It needs scientists who can turn projections into asset-level, decision-ready risk.
Python (scikit-learn)CMIP6 dataChatGPT
1
Downscale and bias-correct CMIP6 projections to asset locations, then translate hazard into an exposure or loss metric.
2
Design the analysis and the uncertainty communication up front.
Copy-paste this prompt
Help me design a physical climate-risk analysis for [a portfolio of real-asset locations] under [SSP2-4.5 and SSP5-8.5]: which CMIP6 variables and hazards to use ([heat, flood, wildfire, wind]), how to downscale and bias-correct, how to translate hazard into an exposure or damage indicator, and how to communicate uncertainty to non-scientists.
State scenario assumptions and uncertainty explicitly — risk figures inform real financial decisions and must not overstate confidence.
What you'll haveDecision-ready climate risk for paying clients — the private-sector pivot that reaches the $175,900 tier.
4
Accelerate literature, proposals, and analysis code
Why this pays: In research and consulting, funding and publications gate advancement. AI compresses the literature review, the proposal draft, and the analysis pipeline.
ElicitNotebookLMClaude
1
Use Elicit or Consensus to build an evidence table, then load the key papers into NotebookLM for cited synthesis.
2
Draft the framing from your own references.
Copy-paste this prompt
Draft the background and significance for a proposal on [the climate impact of X on Y region]. Here are my key references and findings: [paste]. Establish the knowledge gap, state the significance for [decision-makers], and frame 3 specific aims with methods. Flag any claim needing a stronger citation.
Verify every AI-provided citation against the source — models fabricate references. The scientific claims and framing are yours.
What you'll haveFaster, better-supported proposals and papers — the funding and publication record behind advancement.
5
Communicate and visualize for decision-makers
Why this pays: Impact and salary in climate work increasingly come from making complex science actionable. Scientists who communicate risk clearly get the client-facing, higher-paid roles.
ChatGPTPython (Matplotlib/Plotly)Quarto
1
Use AI to draft plain-language briefs and to build clear, honest figures from your data.
2
Translate a technical result for a specific audience.
Copy-paste this prompt
Turn this technical result into a one-page brief for [a city planning board / an investment committee]: [paste the result and key numbers]. Explain what it means, the confidence and caveats, the decision-relevant thresholds, and the recommended actions. Avoid jargon and keep the uncertainty honest.
You own the accuracy — check that plain-language simplification never becomes an overstatement of confidence.
What you'll haveScience that drives decisions — the translator role that commands the top of the pay band.
Your 12-month sequence to the top of the range
How the plays above stack into a path from median pay toward the $175,900 tier.
Month 1
Run one Google Earth Engine analysis with a copilot writing the code — go from question to map yourself.
Months 2-3
Run and validate an open ML weather emulator against reanalysis for a variable you know well.
Months 3-6
Build a climate-risk mini-project — downscale CMIP6 to locations and produce an exposure metric.
Months 6-12
Use AI to accelerate a paper or proposal, and build a public portfolio of climate-risk analyses.
Year 2
Position for a private-sector climate-risk or climate-services role — the pivot that reaches the $175,900 tier.
Gear for this job
As an Amazon Associate, PayCrunch earns from qualifying purchases. Links to books and tools are for the job on this page; we only recommend what we’d use in the work.
Same live O’Reilly 3rd already on data-scientist / python-developer / market-research-analyst / oceanographer. This leftover page names Python (xarray) as a play tool and the emulator prompt is Write the Python/xarray workflow to load both, regrid to a common grid. Not CompTIA Data+ and not leftover Ross Exam P (that is actuary). Confirm 109810403X. Live page HTTP 200, no PC_GEAR / amazon.com/dp / tag=paycrunch-20 at 2026-09-17 3:31 PM PT.
Next steps for a Climate Scientist
Some links below are affiliate or partner links. PayCrunch may earn a commission if you enroll or subscribe through them, at no extra cost to you. Wage figures on this page still come from the Bureau of Labor Statistics, not from these programs.
Climate Scientist work is specific enough that a stamped 'check out these courses' block would be noise. BLS files this work as Atmospheric and Space Scientists (SOC 19-2021). O*NET Job Zone 4 is typical: a bachelor's degree, so the honest next credential is a professional certificate or bachelor's-level coursework — not a random catalog dump.
The occupation's listed knowledge areas include Physics and Geography; the links search those subjects, not a generic 'career courses' list.
Climate Scientists in this dataset list C++ among the tools in use, so a program that names that stack is a better fit than a survey course.
Coursera search for physics — a professional certificate or bachelor's-level coursework that lines up with science, not a generic professional-development aisle.
FlexJobs screens remote, hybrid, freelance, and flexible listings so you are not wading through unverified ads. This is a job-board search for Climate Scientist work, not a claim that they list a counted SOC 19-2021 inventory.
Write a Climate Scientist resume, or one aimed at Physical Scientists, All Other, instead of a blank template. Resume Now is a resume builder; we are not claiming a counted template set for this SOC.
A Climate Scientist resume that names the actual tasks on this page, or the step-up title Physical Scientists, All Other, beats a blank template when you apply.
What Climate Scientists earn by state
This page does not show a state table, and the reason is worth stating: the Bureau publishes this occupation nationally, but fewer than five states employ enough people in it to report a median we would stand behind. Scaling the national median by a cost-of-living index would produce a number for every state, but it would be an estimate of living costs wearing a wage’s clothes, and PayCrunch would rather show you nothing than that.
What the national figures say: pay starts near $53,060, the median is $99,070, and the top of the range is $175,900. Those national figures come from U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics, May 2025.
No. AI emulators are fast approximations trained on the output of physical models and observations — someone has to build, validate, and interpret them, judge physical plausibility, and own the uncertainty. AI raises the value of that judgment and opens new roles in ML emulation and climate-risk analytics. Scientists who add these tools move up; asking the right question stays human.
Can I trust an ML weather or climate emulator's output?
Only after validation. Emulators can be strikingly skillful and also fail outside their training regime or on extremes. Benchmark against physical models, reanalysis, and observations for your variable and region, and never present a projection without its uncertainty and assumptions.
I'm a domain scientist, not a software engineer — can I use these tools?
Yes — that's the shift. A general AI copilot writes and explains the Earth Engine, xarray, and ML code while you supply the climate science. Start with one Earth Engine analysis; you learn the stack by using it, not by becoming a programmer first.
How does AI move a climate scientist's pay up?
The top of the band is increasingly the private sector — insurance, finance, energy, tech — paying for decision-ready climate risk. AI emulators, geospatial ML, and risk analytics are exactly those skills, and AI also speeds the papers and proposals that drive academic advancement. It's the bridge from research to the higher-paid applied roles.
Is it okay to use AI to write my papers and proposals?
To draft and organize, yes; to source facts, carefully. Models fabricate citations and can overstate confidence — verify every reference and number against the source, and keep the uncertainty honest. The tool accelerates writing; the scientific integrity is yours.
Methodology & sources
Salary (median, 10th, top of the range) — U.S. Bureau of Labor Statistics, OEWS.
By state — the Bureau of Labor Statistics’ own state medians, limited to states employing at least 500 people in the occupation. No cost-of-living arithmetic is applied to a wage anywhere on this page.
The plays — PayCrunch's own step-by-step guidance using publicly available AI tools. Tool names/URLs are real and current as of August 2026; prompts written to work as-is. Verify any professional output before relying on it.