Sky Islands Climate is a one-person, hobbyist project. It started with a storm that nobody saw coming. On the evening of March 31, 1999, I was buying a Ford Bronco — planning to tow my boat up to Lake Powell that weekend. The forecast for Flagstaff called for sunshine, light winds, and highs in the mid-70s.
Instead, it started snowing that night and didn’t let up for days, burying the city under 41.3 inches by the time it was done — the 7th-biggest snowstorm on record there. Not a dusting nobody planned for; a complete miss, sunshine-and-70s to feet of snow. Watching that gap — between what the models said and what actually happened on the ground — is what got me hooked, and I’ve been tracking Arizona weather ever since: first on an AOL blog, then on my own site, stu-in-flag.net, and now here. I’ve also been a CoCoRaHS observer for years, contributing daily precipitation reports as ground-truth data for the National Weather Service and researchers.
That gap between what extended-range models promise and what actually happens over Arizona’s complicated terrain is the reason this site exists.
The problem
Extended-range (7–10 day) forecasts from GFS, ECMWF/IFS, ICON, and the NBM show significant divergence and unreliability over Arizona’s complex, orographically-influenced terrain — a gap compounded by public narratives around large-scale teleconnections (ENSO, PDO, AMO, NZI, SOI, MJO) that oversimplify or misstate their actual regional influence.
What this project does
Sky Islands Climate maintains an ongoing, model-verified record for Anthem, AZ and the greater Flagstaff area — comparing extended-range model output (QPF heatmaps across GFS/IFS/NBM/ICON) against ground-truth precipitation observations at selected sites, contextualized against teleconnection indices using the McCabe/Mamalakis/Redmond-Koch/Adams-Comrie research lineage. Analysis is published as recurring blog posts and standing tracker tools, and syndicated to social media.
This is an independent hobby project, not affiliated with NWS, NOAA, or any other agency.
On Wikipedia
If you are still using wikipedia, consider moving your searches to Artificial Intelligence (AI) platforms. Wikipedia’s content is becoming aged and stale. In many places, it is subject to exceptional bias. AI platforms can be effectively used, with appropriate guardrails.
Provide the AI with strict guidance on sourcing and disallow confabulation. For sourcing, limit their references to reliable sources. Reliability can vary by topic or problem. For instance, when writing an article on the history of the climate of the southwestern US, peer-reviewed articles are the sources required. When trying to open a stuck jar lid, many more sources, like blogs or reddit could be appropriate.
Confabulation, or hallucination is a significant issue with most AI platforms. I once observed it create a fictional “peer-reviewed” article. By giving AI a strict rule, or guidance of “no confabulation allowed,” it can self-limit these types of errors. If you are working around a fact-based topic, this is a critical, but not sufficient starting place. It still requires human review for accuracy.
Off Wikipedia
Check out these encyclopedic articles I have written. They are extensively referenced and written following the style guidelines of Wikipedia.