# Authors report near-real-time results for regional weather model HRRRCast [Canonical reading](https://escapemeridian.com/readings/authors-report-near-real-time-results-for-regional-weather-1a2f17ab) The authors describe HRRRCast, a data-driven regional weather model trained on High-Resolution Rapid Refresh (HRRR) analysis data, and report tests of its performance on extreme weather and in near-real-time operation. They say training at native 3-km resolution improved fractions skill score at convective scales compared with 6-km subsampling, while sigmoid-weighted loss improved skill for high-threshold reflectivity and precipitation without harming common-event behavior. In seven months of independent observation-based verification, the authors report that HRRRCast ensemble means improved several near-surface metrics relative to operational HRRR, while maintaining strong precipitation and reflectivity performance. They also report competitive precipitation skill through 1 in/h and skill extending to about 3 in/h in the best recent configuration. A focused comparison with HREF found useful low-cost probabilistic guidance for 3-h precipitation. These results suggest potential value for regional forecasting, but the abstract does not establish how the system would perform across longer periods, other regions, or future extreme events. How does HRRRCast perform across longer periods and a wider range of regions and extreme events? Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims. ## Original source arXiv physics.ao-ph · Article Original source date: Oct 26, 2026 (2026-10-26). Date basis: Date basis not provided. Added to Escape Meridian: 2026-11-06T10:49:55.934Z. First observed: 2026-10-28T04:10:38.108Z. Retrieved: 2026-10-29T15:34:07.456Z. - [Towards operational data-driven regional weather forecasting: extreme-event skill and near-real-time verification with HRRRCast](https://arxiv.org/abs/2610.36186) ## Limits The evidence provided is the paper's abstract; it does not include the full evaluation methods or results. The reported independent observation-based verification covers seven months, so performance across longer periods is not established by this evidence. The abstract does not specify the range of regions or future extreme events over which performance was evaluated. ## Open question How does HRRRCast perform across longer periods and a wider range of regions and extreme events? ## Source-stated The authors evaluate HRRRCast, a data-driven convection-allowing emulator trained on HRRR analysis data. > HRRRCast, a data-driven convection-allowing emulator trained on the High-Resolution Rapid Refresh (HRRR) analysis data - [Supporting source](https://arxiv.org/abs/2610.36186) ## Source-stated The authors report that native-resolution training improved fractions skill score at convective scales compared with 6-km subsampling. > Native-resolution training avoids aliasing from 6-km subsampling and substantially improves fractions skill score at convective scales - [Supporting source](https://arxiv.org/abs/2610.36186) ## Source-stated The authors say sigmoid-weighted loss improved high-threshold skill while preserving common-event behavior. > The weighted loss improves high-threshold skill while preserving common-event behavior. - [Supporting source](https://arxiv.org/abs/2610.36186) ## Source-stated In seven months of independent observation-based verification, the authors report improvements in several near-surface metrics relative to operational HRRR. > In seven months of independent observation-based verification, HRRRCast ensemble means improve several near-surface metrics relative to operational HRRR - [Supporting source](https://arxiv.org/abs/2610.36186) ## Source-stated The authors report competitive precipitation skill through 1 in/h, extending to about 3 in/h in the best recent configuration. > Precipitation skill remains competitive through 1 in/h and extends to about 3 in/h in the best recent configuration. - [Supporting source](https://arxiv.org/abs/2610.36186) ## Escape Meridian interpretation The abstract does not establish performance across longer periods, other regions, or future extreme events. > In seven months of independent observation-based verification - [Supporting source](https://arxiv.org/abs/2610.36186) ## Source and contributor context Abdi, Daniel, Jankov, Isidora, Vargas, Vanderlei, Madden, Paul, Carpenter, Emily, Wilson, Travis, Hamilton, Jeff, Smith, Molly, Blank, Lindsay, Cui, Linlin, Wang, Jun, Mahalik, Matthew · Article author These are credits stated by the source; they do not establish a profile match or authorship of every passage. ## Source media No suitable media was found by the supported page extraction. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## TIME TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services. - [Explore this context](/entities/time) - [Reference](https://time.com/) ## arXiv arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed. - [Explore this context](/entities/arxiv) - [Reference](https://arxiv.org/) - [Reference](https://info.arxiv.org/help/api/index.html) ## Editorial provenance Version 1. Publication ID: 0e99b546-d6ea-4d48-844a-1b9f2f0d637b. Prepared by gpt-6-luna; grounding review by gpt-6-luna. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-06T10:49:55.934Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: material\_development. Observation: fd757a9be415378518e8783e90bab7eab66c2eb35118be8ff5cb4906fc3ea6e4. Source text hash: 1b0e4201b5b08b10d0daf7b3e446c4606ff90abfe759c2d02d3b825e038f60a6. ## Revision and source-check history ### Version 1 — 2026-11-06T10:49:55.934Z The authors describe HRRRCast, a data-driven regional weather model trained on High-Resolution Rapid Refresh (HRRR) analysis data, and report tests of its performance on extreme weather and in near-real-time operation. They say training at native 3-km resolution improved fractions skill score at convective scales compared with 6-km subsampling, while sigmoid-weighted loss improved skill for high-threshold reflectivity and precipitation without harming common-event behavior. In seven months of independent observation-based verification, the authors report that HRRRCast ensemble means improved several near-surface metrics relative to operational HRRR, while maintaining strong precipitation and reflectivity performance. They also report competitive precipitation skill through 1 in/h and skill extending to about 3 in/h in the best recent configuration. A focused comparison with HREF found useful low-cost probabilistic guidance for 3-h precipitation. These results suggest potential value for regional forecasting, but the abstract does not establish how the system would perform across longer periods, other regions, or future extreme events. How does HRRRCast perform across longer periods and a wider range of regions and extreme events? ## Original source arXiv physics.ao-ph · Article Original source date: Oct 26, 2026 (2026-10-26). Date basis: Date basis not provided. Added to Escape Meridian: 2026-11-06T10:49:55.934Z. First observed: 2026-10-28T04:10:38.108Z. Retrieved: 2026-10-29T15:34:07.456Z. - [Towards operational data-driven regional weather forecasting: extreme-event skill and near-real-time verification with HRRRCast](https://arxiv.org/abs/2610.36186) ## Limits The evidence provided is the paper's abstract; it does not include the full evaluation methods or results. The reported independent observation-based verification covers seven months, so performance across longer periods is not established by this evidence. The abstract does not specify the range of regions or future extreme events over which performance was evaluated. ## Open question How does HRRRCast perform across longer periods and a wider range of regions and extreme events? ## Source-stated The authors evaluate HRRRCast, a data-driven convection-allowing emulator trained on HRRR analysis data. > HRRRCast, a data-driven convection-allowing emulator trained on the High-Resolution Rapid Refresh (HRRR) analysis data - [Supporting source](https://arxiv.org/abs/2610.36186) ## Source-stated The authors report that native-resolution training improved fractions skill score at convective scales compared with 6-km subsampling. > Native-resolution training avoids aliasing from 6-km subsampling and substantially improves fractions skill score at convective scales - [Supporting source](https://arxiv.org/abs/2610.36186) ## Source-stated The authors say sigmoid-weighted loss improved high-threshold skill while preserving common-event behavior. > The weighted loss improves high-threshold skill while preserving common-event behavior. - [Supporting source](https://arxiv.org/abs/2610.36186) ## Source-stated In seven months of independent observation-based verification, the authors report improvements in several near-surface metrics relative to operational HRRR. > In seven months of independent observation-based verification, HRRRCast ensemble means improve several near-surface metrics relative to operational HRRR - [Supporting source](https://arxiv.org/abs/2610.36186) ## Source-stated The authors report competitive precipitation skill through 1 in/h, extending to about 3 in/h in the best recent configuration. > Precipitation skill remains competitive through 1 in/h and extends to about 3 in/h in the best recent configuration. - [Supporting source](https://arxiv.org/abs/2610.36186) ## Escape Meridian interpretation The abstract does not establish performance across longer periods, other regions, or future extreme events. > In seven months of independent observation-based verification - [Supporting source](https://arxiv.org/abs/2610.36186) ## Source and contributor context Abdi, Daniel, Jankov, Isidora, Vargas, Vanderlei, Madden, Paul, Carpenter, Emily, Wilson, Travis, Hamilton, Jeff, Smith, Molly, Blank, Lindsay, Cui, Linlin, Wang, Jun, Mahalik, Matthew · Article author These are credits stated by the source; they do not establish a profile match or authorship of every passage. ## Source media No suitable media was found by the supported page extraction. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## TIME TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services. - [Explore this context](/entities/time) - [Reference](https://time.com/) ## arXiv arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed. - [Explore this context](/entities/arxiv) - [Reference](https://arxiv.org/) - [Reference](https://info.arxiv.org/help/api/index.html) ## Editorial provenance Version 1. Publication ID: 0e99b546-d6ea-4d48-844a-1b9f2f0d637b. Prepared by gpt-6-luna; grounding review by gpt-6-luna. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-06T10:49:55.934Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: material\_development. Observation: fd757a9be415378518e8783e90bab7eab66c2eb35118be8ff5cb4906fc3ea6e4. Source text hash: 1b0e4201b5b08b10d0daf7b3e446c4606ff90abfe759c2d02d3b825e038f60a6. ### uncertain — 2026-11-06T10:49:55.934Z — anthropic/claude-opus-5.5 Publication: 0e99b546-d6ea-4d48-844a-1b9f2f0d637b attribution — pass: The reading attributes the findings to the paper's authors with phrasing like "the authors report" and "they say". It does not present the results as established fact. The arXiv page lists Daniel Abdi and colleagues as authors. > Authors:Daniel Abdi, Isidora Jankov, Vanderlei Vargas, Paul Madden, Emily Carpenter summary — uncertain: Several central terms go unexplained for a non-specialist: fractions skill score, convective scales, sigmoid-weighted loss, high-threshold reflectivity, ensemble means, and HREF (an acronym never expanded). The reading also never says that HRRR is the operational physics-based model being emulated, so the reader cannot see why a cheaper data-driven version matters. It omits the source's own storm-scale caveat: placement is better than morphology. Smallest repair: add the displacement-structure trade-off sentence, briefly gloss HRRR, HREF and the key metrics in plain words or term tooltips, and note that the system has run in near-real-time mode for a year. > At storm scale, HRRRCast exhibits a displacement--structure trade-off, with generally better storm placement than storm morphology. date — pass: The displayed originalPublished date of 2026-10-26 matches the arXiv v1 submission date. The later feed date of Oct 28 is correctly not used. > \[Submitted on 26 Oct 2026\] support — pass: Each factual claim closely paraphrases the abstract. The seven-month limitation is labeled as interpretation and keeps the gap between the evidence and any broader conclusion. The question about longer periods and other regions follows reasonably from the stated verification window. It is not presented as a finding. > In seven months of independent observation-based verification, HRRRCast ensemble means improve several near-surface metrics relative to operational HRRR media — pass: The selected media is the verified direct arXiv abstract page for this paper. It is the most direct way for a reader to access the source. ## Recorded references Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter. No public references recorded here yet. [Reference data (JSON)](/readings/1a2f17abc7b4f96cb7b6eb03b2d874cbe0371cb19eab2b9e55431c9e940b003b/references.json) ## Corrections and challenges [Send a correction or an opposing source](https://escapemeridian.com/corrections/) [JSON record](https://escapemeridian.com/readings/authors-report-near-real-time-results-for-regional-weather-1a2f17ab.json)