{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-11-04-026",
    "source_story_id": "tmp-story-flood-depth-highwater-marks",
    "edition_id": "mp-2026-11-04-morning-0117",
    "edition_url": "https://themachinepress.com/edition/2026-11-04",
    "position": 15,
    "story_type": "dispatch",
    "section": "research",
    "editorial_classification": "editorial",
    "headline": "High-Water Marks Teach the Model Before the Peak",
    "slug": "high-water-marks-teach-the-model-before-the-peak",
    "dek": "A fusion network trained on 5,689 field surveys predicts flood depth six hours ahead with 0.39-meter mean absolute error on sampled events.",
    "summary": "A fusion network trained on 5,689 field surveys predicts flood depth six hours ahead with 0.39-meter mean absolute error on sampled events.",
    "body_text": "The dataset spans 39 events from 2011 through 2025 across 24 U.S. states and Puerto Rico, pairing USGS high-water marks with terrain, geospatial embeddings and 144 hours of meteorological forcing. At surveyed marks, the model reaches 0.39-meter mean absolute error six hours before peak, while a measured-water-level-minus-terrain baseline has 0.93-meter error. Attribution assigns 77 percent of the reduction to forcing. But on 11 held-out events, the model still separates flooded from rain-free cases while depth resolution collapses to an R-squared of 0.03. The authors therefore limit the reported skill to sampled regions and storms.",
    "why_it_matters": "A fusion network trained on 5,689 field surveys predicts flood depth six hours ahead with 0.39-meter mean absolute error on sampled events.",
    "limitations": [],
    "importance": 9,
    "canonical_url": "https://themachinepress.com/story/mp-2026-11-04-026/high-water-marks-teach-the-model-before-the-peak",
    "json_url": "https://themachinepress.com/story/mp-2026-11-04-026.json",
    "first_published_at": "2026-11-04T09:00:00.000-05:00",
    "modified_at": "2026-11-04T09:00:00.000-05:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-11-04-026-001",
        "text": "A fusion network trained on 5,689 field surveys predicts flood depth six hours ahead with 0.39-meter mean absolute error on sampled events.",
        "source_ids": [
          "source-2026-11-04-015"
        ],
        "qualification": null
      }
    ],
    "source_ids": [
      "source-2026-11-04-015"
    ],
    "tags": [
      "flooding",
      "Earth observation",
      "forecasting"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-11-04-015",
      "title": "arXiv preprint 2611.05220, original v1",
      "publisher": "arXiv",
      "url": "https://arxiv.org/abs/2611.05220v1",
      "canonical_url": "https://arxiv.org/abs/2611.05220v1",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-11-03T12:05:45.000-05:00",
      "accessed_at": "2026-11-04T08:28:27.560-05:00",
      "supports_claim_ids": [
        "claim-mp-2026-11-04-026-001"
      ]
    }
  ],
  "corrections": [],
  "publisher": {
    "name": "The Machine Press",
    "url": "https://themachinepress.com",
    "description": "A daily newspaper for the age of artificial intelligence."
  },
  "cite_this_report": {
    "title": "High-Water Marks Teach the Model Before the Peak",
    "publisher": "The Machine Press",
    "published_at": "2026-11-04T09:00:00.000-05:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-11-04-026/high-water-marks-teach-the-model-before-the-peak"
  }
}