{
  "abstract": "Extreme heat is where urban adaptation needs kilometer-scale data the most, but the simulations training a downscaler can cost more than they save, and how much is needed has not been identified. We measured it with CASPER, a U-Net with a structure-preserving loss downscaling 32 km reanalysis to 1 km temperature, humidity and wind, across 24 configurations of one to eight months. Held-out error grows linearly with climatological distance to the training data, RMSE = 0.83 + 2.95 d, explaining 90% of its variance against 7% for volume and predicting unseen months in advance. On held-out extreme summer weeks CASPER preserves the fine-scale structure and cross-variable physics that matched-budget baselines degrade, and matches station observations during documented heat waves to within 1.8 K. Transfer to a new region degrades geographically; 11 days of local simulation cuts Vancouver's held-out error from 3.8 to 1.3 K. Training periods should span the target climate: the same accuracy for four times less simulation, putting kilometer-scale downscaling of extreme heat within reach of groups without large computing facilities.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2609.40140",
  "author_total": 9,
  "authors": [
    {
      "name": "Ahmed Marey"
    },
    {
      "name": "Henry Lu"
    },
    {
      "name": "Abhishek Gaur"
    },
    {
      "name": "Sherif Goubran"
    },
    {
      "name": "Malek Aloui"
    },
    {
      "name": "Theodore Potsis"
    },
    {
      "name": "David Rolnick"
    },
    {
      "name": "Alex Hernandez-Garcia"
    },
    {
      "name": "Liangzhu Leon Wang"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/cdjgd0b1jt8m2.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2609.40140",
  "fulltext_url": null,
  "id": "cdjgd0b1jt8m2",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2609.40140",
  "language": null,
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "article",
  "licence_url": "http://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cs.LG"
    ],
    [
      "arxiv",
      "physics.ao-ph"
    ],
    [
      "openalex",
      "1902|Meteorological Phenomena and Simulations"
    ]
  ],
  "openalex": "W7216313013",
  "pdf_url": "https://arxiv.org/pdf/2609.40140",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-01",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/cdjgd0b1jt8m2.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:26:03Z",
      "native_id": "2609.40140",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2609.40140"
    }
  ],
  "subjects": [
    {
      "field": "earth",
      "origin": "arxiv:physics.ao-ph",
      "subfield": "atmospheric"
    }
  ],
  "title": "Less is more: error-distance scaling relation for data-efficient kilometer-scale downscaling of extreme heat",
  "title_text": "Less is more: error-distance scaling relation for data-efficient kilometer-scale downscaling of extreme heat",
  "type": "preprint",
  "updated_at": "2026-10-30T12:26:03Z",
  "url": "https://research.draconitic.com/paper/cdjgd0b1jt8m2",
  "venue": "arXiv",
  "versions": []
}