The story reads as a political one: the U.S. Global Change Research Program, and the National Climate Assessment it produces, is now directed by Matthew Wielicki, a geochemist who publicly disputes mainstream climate science. For water-resources practitioners, the more useful question is narrower. What does an assessment do inside our design work, and what happens to that work if the assessment stops being reliable?
What the assessment actually feeds
The National Climate Assessment (NCA) is a congressionally mandated report, produced roughly every four years under the Global Change Research Act, that synthesizes how climate change is altering conditions across the United States. It is not, by itself, a design manual. But it is the authoritative national statement that regional guidance, agency policy, and downscaled projections lean on. When a stormwater team justifies a larger culvert, or a levee district revisits its freeboard, the chain of reasoning frequently traces back to the direction of travel the NCA documents: heavier short-duration rainfall, shifting snowpack, longer dry spells punctuated by more intense events.
That direction of travel matters because the standard tools assume it does not exist. Flood-frequency analysis, the return-period estimates behind most infrastructure sizing, rests on stationarity: the premise that the statistical distribution of floods is fixed, so the past is a fair sample of the future. A warming atmosphere breaks that premise. The assessment is one of the few national vehicles that says so plainly.
Where nonstationarity bites
Consider the numbers a design engineer pulls without much thought. Precipitation-frequency estimates, in the United States drawn largely from NOAA Atlas 14, convert a location and a return period into a design storm depth. Those estimates are fitted to historical gauge records. If the underlying climate is drifting, an Atlas value labelled as a 1-percent annual chance event may already understate present risk, and understate future risk further over an asset's service life.
The assessment is where that drift is supposed to be named, quantified at regional scale, and translated into guidance a practitioner can defend. Strip out or soften that signal and the design record does not become wrong overnight. It becomes quietly outdated, in a way that is hard to detect from inside a single project. The failure mode is not a dramatic error. It is a slow accumulation of infrastructure sized to a climate that no longer exists.
Planning without a trusted federal signal
Practitioners are not left with nothing if the national synthesis loses credibility. The primary science it summarizes still exists. Global climate projections from the CMIP model ensembles are open, and the downscaling methods that turn them into catchment-scale precipitation and temperature series are documented and reproducible. State climate offices, regional consortia, and academic centres already publish projection products that can substitute for, or cross-check, federal guidance. The cost is real: this work moves from a single authoritative reference to a patchwork that each organization has to assemble and defend on its own.
The practical response is to make climate uncertainty explicit rather than inherited. That means stating which precipitation-frequency source a design uses and how old it is, testing sensitivity to a plausible range of intensification rather than a single Atlas value, and adding freeboard where consequences are high and projections diverge. These are ordinary engineering habits. They simply have to be applied to the climate input, not only to the hydraulic one.
The wetland footnote
There is a related reason not to treat this as an abstract policy dispute. A recent study quantifying the cost of wetland loss put a figure on something engineers already model: natural floodplain and wetland storage attenuates peaks and buys downstream capacity. When flood risk is understated at the source, the value of the natural buffers that offset it is understated too, and those buffers are easier to permit away. Good hydrologic accounting and good climate accounting are the same accounting.
None of this is a verdict on any individual appointment. It is a reminder that the assessment is not a press release. It is a design input, and design inputs that drift silently are the expensive kind. For related coverage, see our note on nature-based flood-risk reduction and our explainer on what a 100-year flood really means. More about this publication is on the about page.