Flood Exposure Situation Report · Southern Mexico (Veracruz, Oaxaca, Chiapas)

Capybara Center for Risk Analysis · situation report · under JANUS

Flood Exposure Situation Report · Southern Mexico

Veracruz, Oaxaca and Chiapas, Bayesian flood model, concurrent with a national intense-rain alert

HazardFlood, short-horizon exposure

AreasVeracruz, Oaxaca, Chiapas (three runs)

Runs2026-07-08, 16:33 / 17:39 / 17:46 UTC

HorizonGFS short-term, 168 h (~7 days)

Provenance. Legacy Capybara Bayesian flood model (Google Earth Engine), model=bayesian, prior_balance=balanced, profile=standard. ENSO baseline: el_nino (2019, 2023, 2024), consistent with the current NOAA/IDEAM El Nino advisory. pop=WorldPop100m · Code as FDO: derivedFrom: capybara-legacy [PID pending].

Legend flood risk: low → high (grey, blue, deep blue, dark purple = highest)

Three state-level flood runs, issued as Mexico's Coordinacion Nacional de Proteccion Civil placed Veracruz, Oaxaca and Chiapas under an intense-rain alert. The model concentrates exposure where it should: the Veracruz Gulf coastal plain, western and coastal Oaxaca, and eastern and northern Chiapas. Across the three states, roughly 4.4 million people fall within the flagged flood surface. The figures are scenario estimates of exposure; the corroboration is concurrent observation, consistency, not predictive skill.

Three Bayesian flood runs, 8 Jul 2026, El Niño baseline. Purple marks the highest-risk class. Top: Veracruz. Middle: Oaxaca. Bottom: Chiapas.

Population is now reliable at state scale. Unlike recent national runs (where population returned N/A or absurd totals), these three states report correct in-domain populations (Oaxaca 3.89M, Veracruz 7.41M, Chiapas 5.22M), so the people-exposed figures below are usable. The remaining aggregation issue is specific to national and big-country runs and does not affect these.

§AVeracruz — the Gulf coastal plain

Where it concentrates. The highest-intensity surface covers the length of the Gulf coastal lowland: the Huasteca and Panuco basin in the north (Tampico border, Tuxpan), the Poza Rica and Papantla corridor, and, most heavily, the southern Sotavento and Papaloapan lowlands around Alvarado, Cosamaloapan, Tres Valles, Coatzacoalcos and Minatitlan. This is Mexico's most flood-prone coastal state, and it is the most exposed of the three by every measure.

Asset	Share exposed	Expected (in-domain)
Protected areas	48.6%	~249,600 ha
Cropland	47.2%	~461,900 ha
Urban area	45.2%	~23,000 ha
Oil palm	39.1%	~1,651 ha
Population	38.2%	~2,828,000 ppl
Electric grid	35.7%	~4,648 km
Built-up footprint	35.7%	~2,461,000 ha
Roads	34.6%	~9,146 km

Triangulation · concurrent (7-8 Jul 2026)

Mexico's CNPC and the Servicio Meteorologico Nacional issued an alert for 7 July for very heavy to intense rain of 75 to 150 mm over Veracruz (centre and south), Oaxaca, Chiapas and Tabasco, warning of flooding, river rises and landslides, driven by tropical wave number 16 and cyclonic circulations over the western Gulf of Mexico (CNPC/SSPC; SMN). The centre-and-south emphasis matches the model's heaviest flagging in the Veracruz lowlands. Coastal swell of 1 to 2 metres was also forecast for the Gulf.

{ "@id": "nanopub:tri_MX_veracruz_flood_2026-07",
  "assertion": { "subject": "model flood surface, Veracruz Gulf coastal plain (Sotavento/Papaloapan)",
    "predicate": "coincidesWith", "object": "CNPC/SMN 75-150mm intense-rain alert, Veracruz centre/south, 7 Jul 2026" },
  "provenance": { framing: "concurrent observation, consistency", driver: "tropical wave 16",
    ensoBaseline: "el_nino (matches state)", source: "gob.mx/sspc; smn 2026-07-07",
    derivedFrom: "capybara-legacy [PID pending]" },
  "pubInfo": { "identifier": PID pending, "license": "CC-BY-4.0" } }

§BOaxaca — the western Mixteca and coast

Where it concentrates. The flagged surface concentrates in western Oaxaca, the Mixteca and the Costa around Santiago Pinotepa Nacional, Santiago Juxtlahuaca and Putla, with additional exposure along the Pacific coast. The central valleys and the Isthmus are lower. About 1.09 million people fall in the flagged surface, roughly 28 percent of the in-domain population, with grid and road exposure near a third.

Asset	Share exposed	Expected (in-domain)
Electric grid	32.0%	~3,574 km
Roads	31.7%	~6,464 km
Population	28.2%	~1,094,000 ppl
Built-up footprint	27.7%	~2,537,000 ha
Oil palm	22.7%	~20 ha
Urban area	18.6%	~2,620 ha
Cropland	16.9%	~83,600 ha
Protected areas	3.3%	~32,300 ha

Triangulation · partial (divergence logged)

Oaxaca is named in the same CNPC/SMN alert, and earlier tropical waves 13 and 14 flagged Oaxaca's north-west and south-west. Note a spatial nuance: the 7 July SMN bulletin emphasizes northern and north-eastern Oaxaca (the Tuxtepec and Papaloapan side bordering Veracruz), whereas this run's heaviest flagging is in the western Mixteca and Costa. The overlap is the state-level alert and the Pacific coastal swell; the model's western concentration is the more structural, susceptibility-driven signal and is not the exact footprint of today's rain. This divergence is recorded, not smoothed.

{ "@id": "nanopub:tri_MX_oaxaca_flood_2026-07",
  "assertion": { "subject": "model flood surface, western Oaxaca (Mixteca/Costa)",
    "predicate": "partiallyCorroboratedBy", "object": "CNPC alert (state) + waves 13/14; SMN 7 Jul emphasizes N/NE Oaxaca" },
  "provenance": { framing: "state-level consistency; sub-state divergence logged",
    source: "gob.mx/sspc; smn; nmas 2026-07", derivedFrom: "capybara-legacy [PID pending]" },
  "pubInfo": { "identifier": PID pending, "license": "CC-BY-4.0" } }

§CChiapas — the eastern Lacandon and the north

Where it concentrates. The flagged surface sits in eastern Chiapas, the Lacandon lowlands and the Montes Azules biosphere reserve, Marques de Comillas and the Guatemala border around Benemerito de las Americas and Frontera Corozal, with additional exposure in northern Chiapas (Reforma, Pichucalco, toward Tabasco) and pockets of the Soconusco coast near Motozintla and Tapachula. Overall population exposure is lower (8.8 percent), but protected-area exposure is high (36 percent) because the flagged east is precisely the Lacandon rainforest reserve system.

Asset	Share exposed	Expected (in-domain)
Protected areas	36.2%	~474,300 ha
Built-up footprint	18.5%	~1,363,000 ha
Oil palm	15.4%	~6,564 ha
Electric grid	13.0%	~1,481 km
Roads	11.1%	~2,060 km
Population	8.8%	~457,000 ppl
Urban area	3.9%	~803 ha
Cropland	3.4%	~14,800 ha

Triangulation · concurrent

Chiapas is named in the CNPC/SMN alert, with the 7 July bulletin emphasizing the state's north-west (the Pichucalco and Reforma zone toward Tabasco), which matches the model's northern-Chiapas flagging. The heavy eastern Lacandon flagging is the structural, high-rainfall rainforest signal rather than the specific footprint of today's system. The high protected-area exposure is the operationally distinctive finding here, a flood story about ecosystems and reserves as much as people.

{ "@id": "nanopub:tri_MX_chiapas_flood_2026-07",
  "assertion": { "subject": "model flood surface, eastern Lacandon + northern Chiapas",
    "predicate": "coincidesWith", "object": "CNPC/SMN alert, Chiapas NW, 7 Jul 2026; 36% protected-area exposure" },
  "provenance": { framing: "consistency; eastern signal is structural",
    source: "gob.mx/sspc; smn 2026-07-07", derivedFrom: "capybara-legacy [PID pending]" },
  "pubInfo": { "identifier": PID pending, "license": "CC-BY-4.0" } }

§DReading the three together

The three runs share one weather system and one El Nino baseline, and together they map where a single intense-rain episode lands on people, land and lifelines across Mexico's south. Veracruz carries the human exposure (2.83 million people, near half its cropland and protected land), Oaxaca the lifeline exposure in the west (about a third of grid and roads), and Chiapas the ecosystem exposure (a third of its protected areas, the Lacandon reserves). A useful discipline the runs also expose: the model's structural, susceptibility-driven surface is broader than any single day's rainfall footprint, so it is best read as where flooding concentrates over the season, cross-checked against the daily SMN bulletins for the specific pulse.

§ERecommended use

1. Treat these as state readiness maps for the rainy season, prioritizing the Veracruz Sotavento, western Oaxaca and northern/eastern Chiapas.
2. Cross the flagged surface with the daily SMN and CNPC bulletins to align the structural picture with the specific system (currently tropical wave 16).
3. Route the Chiapas protected-area figure to environmental authorities; this is as much an ecosystem-flood story as a human one.
4. Now that population is reliable at state scale, the people-exposed figures can be used directly; withhold them only for national and big-country runs pending the remaining fix.
5. Freeze and date the three runs so they can be evaluated against the season.

§FAssumptions & how to read the figures

The model returns probability-weighted exposure fields, not point predictions. Percentages are the expected share of each in-domain asset within the flagged surface under the stated configuration. They are planning estimates, not counts of confirmed events, and are not validated against realized outcomes. The El Nino baseline matches the current advisory; note that current flooding is driven by tropical waves and monsoon moisture, which occur despite El Nino's suppression of named Atlantic hurricanes.

§GSources

• CNPC / SSPC, gob.mx. Alerta por lluvias muy fuertes en Veracruz, Oaxaca, Chiapas y Tabasco (7 Jul 2026). gob.mx/sspc
• Servicio Meteorologico Nacional (Conagua), pronostico general, onda tropical 16 (7 Jul 2026). smn.conagua.gob.mx
• UnoTV, alerta por lluvias intensas Veracruz/Oaxaca/Chiapas/Tabasco (7 Jul 2026). unotv.com
• Diario de Mexico, CNPC alerta 75-150 mm (7 Jul 2026). diariodemexico.com
• N+ / nmas, ondas tropicales 13-14, estados afectados (1 Jul 2026). nmas.com.mx

Capybara Center for Risk Analysis · colectivo de ciencia ciudadana · marco JANUS (Feged-Rivadeneira, 2025; archived release doi.org/10.5281/zenodo.21170613).
Legacy flood model runs. Probabilistic exposure estimates, not deterministic forecasts of inundation or damage. Triangulation is concurrent observation. Consistency, not independently established skill.