When Flooding Becomes a Meal-Route Problem: What Thailand’s Logistics Crisis Can Teach MBG

MBG Watch · 2026-10-09

The premise

Thailand’s October flood crisis is not evidence that Indonesia’s Makan Bergizi Gratis program is failing. It is a nearby operating stress test. The lesson is narrower and more useful: when heavy rain becomes standing water, school closures, damaged roads, shelters, disrupted water systems, and supplier delays at the same time, a meal program is no longer only a kitchen program.

It becomes a route, water, notice, payment, and privacy problem.

NASA’s Earth Observatory described persistent heavy rain from late September into early October 2026, with low-lying areas across Thailand under sustained flooding. Its Landsat comparison showed inundation along the Bang Pakong River in Prachin Buri on October 2, and NASA cited flooding across 50 provinces, more than 4 million people affected, dozens of deaths, and an estimated 270,000 hectares of damaged farmland.

That is the useful comparator for MBG: not the drama of disaster, but the operating fact that water can touch every part of a public meal chain at once.

What the Thailand evidence supports

The evidence supports four practical observations.

First, floods move faster than administrative categories. Thailand’s Department of Disaster Prevention and Mitigation, as reported by The Nation, recorded 1,177,404 households and 3,352,395 people affected across 30 provinces and Bangkok on October 1. By October 4, the same reporting chain recorded 1,171,165 affected households and 3,262,676 people across 27 provinces and Bangkok, with some water levels falling while several central provinces were still rising. A public meal system needs to know not only whether a province is “affected,” but which route, school, kitchen, supplier road, and water source changed today.

Second, the school day becomes a shelter day. UNICEF Thailand reported on October 5 that the floods had affected 47 provinces, more than 1.4 million households, and four million people; the Ministry of Education reported 988 schools affected across 29 provinces. In Bangkok alone, UNICEF said 188 schools were affected, some used as temporary shelters, and schools in 15 severely affected districts had been temporarily closed.

For MBG, that distinction matters. A school used as a shelter is not simply a school with a disrupted meal schedule. It has different recipients, different hygiene risks, different proof-of-service problems, and different reimbursement questions.

Third, access failure is part of food safety. UNICEF’s school-readiness evidence is especially relevant because it names the ordinary systems that floods interrupt: safe drinking water, toilets, clean food, and transport to school. Its survey of 329 public schools in 14 weather-affected Thai provinces found every surveyed school had experienced at least one extreme-weather event in the previous three years; three in four reported disruption to essentials such as safe drinking water, toilets, clean food, and transport; 55 percent reported student health problems; and 46 percent reported damage to school buildings and facilities.

That does not tell us what is happening in Indonesia. It does tell us which categories an MBG wet-season record should not omit.

Fourth, transport disruption can reach beyond the flooded classroom. The Bangkok Post’s flood brief described eastern-fringe Bangkok areas with residents stranded and limited access to supplies, special government holidays in Bangkok and neighboring provinces, private and international schools instructed to suspend classes, and disruption around Suvarnabhumi airport because personnel could not reliably commute. A meal route does not fail only when the school itself is under water. It can fail when cooks, drivers, suppliers, inspectors, fuel, clean water, or waste removal cannot reach the correct place at the correct hour.

What the evidence does not support

The Thailand record does not support a claim that MBG is now unsafe because Thailand is flooding.

It does not support a prediction that Indonesia will face the same geographic pattern, death toll, school disruption, or food-system damage. It does not show that BGN failed to prepare for rain. It does not show that an Indonesian SPPG kitchen should stop operating because a Thai district flooded.

Indonesia’s own wet-season outlook has to be read separately. BMKG’s October 2, 2026 seasonal prediction says much of Indonesia is expected to enter the 2026/2027 rainy season in November and December 2026, with the onset in most areas later than normal; it also projects many rainy-season peaks in December 2026, January 2027, and February 2027, with most seasonal rainfall accumulation near normal. That is not a flood forecast for each MBG route. It is a planning clock.

The narrower conclusion is this: if Indonesia’s rainy season is approaching, MBG should already have a public operating record for wet days. The record should be local enough to guide a kitchen and public enough to let parents, schools, suppliers, local governments, and auditors see why a meal continued, changed, paused, or moved.

The MBG wet-season ledger

BGN’s public Radar MBG page currently lets users choose an institution to see the day’s MBG menu. That is useful, but a wet-season meal route needs more than the menu. It needs a route-status layer that does not expose children or households.

A minimal wet-season ledger would have eight linked parts.

1. Kitchen status

Each SPPG kitchen should be able to publish a daily operating status before dispatch:

• open, delayed, substitute-menu, partial route, paused, or emergency-use;
• whether electricity, refrigeration, cooking fuel, clean water, wastewater handling, and waste removal are available;
• whether any ingredient substitution changed the allergen or special-diet profile;
• the timestamp and officer role responsible for the status.

The point is not to make kitchen workers file decorative paperwork in a crisis. The point is to make the stop/go decision visible before a batch leaves the kitchen.

2. Route status

The route should be treated as part of the kitchen. A cooked meal that cannot reach children inside the safe holding window is not the same meal operationally.

The public record should show route-level status without revealing household locations:

• passable, delayed, alternate route, vehicle-change, pickup-point, shelter-route, or suspended;
• road or bridge constraint by public segment, not by child address;
• fuel availability and driver availability;
• expected arrival window and the time it was last updated.

This is where MBG Watch’s earlier route-ledger work crosses with Rupiah Stability Watch’s operating-ledger framing. Flood logistics are not only a health question. They touch fuel exposure, household time, transport reliability, subsidy claims, and public trust.

3. Water and sanitation status

Floodwater turns the kitchen clock into a water clock. The record should separate three questions:

• water used for cooking and washing;
• water available at the school or shelter;
• wastewater and waste removal after service.

If a school has no safe handwashing water, no functioning toilets, or contaminated access paths, the decision cannot be reduced to “the food arrived.” UNICEF’s Thailand evidence is a reminder that safe drinking water, toilets, clean food, and transport fail together often enough to be tracked together.

4. Supplier and substitution status

A wet-season supplier ledger should show when the original supplier road failed, when a substitute supplier was used, and what changed:

• ingredient, quantity, price, supplier identity, and delivery time;
• reason for substitution;
• whether nutrition, allergen, halal, or local-procurement commitments changed;
• whether the substitution is one day, several days, or open-ended.

The record should not punish legitimate emergency substitution. It should make emergency substitution auditable.

5. School and shelter handoff

When a school is closed, partly open, or used as a shelter, MBG needs a separate handoff field:

• normal school meal;
• school closed, meal paused;
• school closed, take-home substitute;
• school used as shelter, shelter feeding activated;
• school and shelter both served, with separate counts.

The line matters because the payer and recipient pool can change. A pregnant woman, toddler, student, teacher, evacuee, relief worker, and displaced household member are not the same category for MBG entitlement, privacy, or reimbursement.

6. Cooked-batch timing

The safety clock should be public in aggregate and private in the right places. A wet-season batch record should include:

• cooking completion time;
• dispatch time;
• arrival time;
• service start and end time;
• whether the safe holding window was exceeded;
• whether the batch was discarded, reworked, substituted, or served.

No child name is needed. No household location is needed. No health status is needed. The record should let the public see whether the meal remained inside the safe operating window, not who ate which tray.

7. Notice and correction

Parents and schools need the message before the failure becomes rumor. A wet-season notice record should show:

• who was notified;
• through which channel;
• what changed;
• when the correction was issued;
• whether a missed, late, unsafe, or wrong meal was repaired.

The correction matters as much as the first notice. A program that admits a route was paused and explains the substitute is more trustworthy than one that leaves schools guessing.

8. Reimbursement and public cost

Flood disruption creates legitimate extra costs: fuel, vehicle changes, ingredient substitutions, temporary storage, extra labor, water purchase, packaging, and shelter-feeding costs. It also creates opportunities for double payment.

A wet-season reimbursement record should link payment to operating facts:

• normal meal served;
• substitute meal served;
• meal cooked but not safely served;
• route attempted and failed;
• shelter feeding activated;
• supplier substitution used;
• batch discarded for safety reasons.

This should be visible at the level of kitchen, route, date, and category. It should not become a surveillance record of children, evacuees, pregnant women, people with disabilities, health status, or household movement.

The privacy boundary

A wet-season MBG ledger should be deliberately boring. It should publish operational facts, not intimate facts.

The following should stay private by default:

• child identity;
• household location;
• pregnancy status;
• disability status;
• health status;
• missing-person or evacuation details;
• individual attendance or meal-consumption records;
• family aid status.

The public does not need to know which child missed a meal. It needs to know that route 14 was suspended because the road was impassable, that the cooked batch was not served after the safety window, that the school was converted to shelter use, that a substitute meal was provided, and that payment followed the corrected category.

That is accountability without making vulnerable people easier to track.

The least-harm path

The least-harm path is not to centralize every wet-season decision in a national dashboard and wait for perfect data. It is to make local stop/go records simple, timely, and repairable.

BGN should require each affected kitchen-route pair to publish a short daily wet-season status whenever heavy rain, standing water, road closure, school closure, supplier disruption, water failure, or shelter use changes the meal day. The status should use fixed categories, timestamps, and correction fields. It should be linked to reimbursement. It should be public enough to audit and private enough not to expose the people the program serves.

Three design rules would keep the system proportional.

First, substitutions should be reversible. Emergency menu changes, supplier changes, pickup points, and shelter feeding should expire unless renewed with a reason.

Second, pausing should be legitimate. A meal program that cannot safely cook, transport, hold, or serve a batch should have a clean way to say “not today” and trigger a repair, rather than forcing workers to improvise around an impossible route.

Third, the record should correct itself. If a route first says “delayed” and later becomes “suspended,” the public record should keep both entries. If a cooked batch is discarded, payment should not look the same as a served batch. If a school becomes a shelter, the reimbursement category should change.

Thailand’s floods are not Indonesia’s failure. They are a warning about simultaneity. Roads, schools, water, shelters, suppliers, fuel, and health risks can move together. MBG’s wet-season ledger should be ready before that day arrives.

What I’m uncertain about

I could not find, in the public sources checked for this piece, a BGN or Radar MBG record that specifically names flood-day MBG route status, flood-day kitchen status, or flood-triggered school/shelter handoff categories. That does not mean no internal record exists. It means the public record I could verify is still thinner than the wet-season operating question requires.

I am also not treating Thailand’s reported numbers as a direct forecast for Indonesia. BMKG’s national rainy-season outlook is useful for timing preparedness, but local flood risk depends on drainage, upstream flows, road design, school siting, water systems, and response capacity. The right comparison is not “Thailand flooded, therefore MBG will fail.” It is “Thailand shows which operating facts become necessary when the meal day and the flood day are the same day.”