Hourly Heat, Not Daily Heat: The Food-Safety Clock MBG Kitchens Need to Publish

MBG Watch · 2026-08-13

The premise

MBG Watch has already argued, in “Heat at the Kitchen Door,” that ambient heat is an operating condition for food safety, not an excuse to be discovered after children become ill. “The Power Behind the Plate” extended that argument to the infrastructure around the meal: refrigeration, water, backup power or fuel, route timing, and incident response. “Food Safety Crisis in MBG Rural Rollout” set the baseline: as of May 10, 2026, MBG Watch’s record cited 37,673 suspected victims across 445 incidents, linked to Ministry of Health reporting. “The First Incident After Suspension” and the Semarang reference in “The Power Behind the Plate” showed why time-and-temperature failures matter even when a kitchen is formally inside the program’s control system.

The new signal is narrower and useful: heat risk is clearer when measured by hours, not only by daily highs. A 2026 Nature Climate Change study on hourly heat extremes says daily temperature metrics “fail to capture the full dynamics of heat events.” Under a high-emissions scenario, the study projects hourly heat extremes to increase fourfold, population exposure to grow sixfold, and each hot day to gain about four hot hours, with another three hot hours occurring on days that daily metrics would not classify as hot. Ars Technica’s account of the same study makes the operational point plainly: if the unit of risk is the hour, the harmful part of the day becomes visible in a way daily maximum or mean temperature can hide.

For MBG kitchens, that is not an abstract climate metric. A meal can move through several heat-sensitive stages before it reaches a child: pre-dawn cooking, post-cook holding, dispatch, vehicle transport, school arrival, queueing, classroom distribution, delayed consumption, and sometimes being carried home. The kitchen does not fail because the official daily high is high. It fails when cooked food spends enough operational time in unsafe conditions without verified hot-holding, cold-holding, clean water, power, fuel, or a route short enough to keep the risk bounded.

What the evidence supports

The general food-safety rule is already time-based. WHO’s Five Keys to Safer Food include “keep food at safe temperatures,” alongside clean handling, separation of raw and cooked food, thorough cooking, and safe water and raw materials. The WHO poster gives the practical form: do not leave cooked and perishable food at room temperature for more than two hours; keep cooked food piping hot, above 60°C, before serving; refrigerate promptly, preferably below 5°C. Those thresholds are not an MBG-specific standard, and they do not replace Indonesian regulation. They do show the discipline MBG needs: temperature is not a certificate on the wall. It is a measured condition over time.

The current Indonesian response is moving in the right direction, but it remains incomplete if it stops at certification. On August 8, Antara reported BGN head Sudaryono saying the Hygiene and Sanitation Feasibility Certificate, or SLHS, is “not an administrative requirement” but an “absolute requirement” for MBG kitchens. The same report says BGN had received reports on around 950 kitchens suspected of failing hygiene and sanitation standards, and that kitchens failing the requirements would not be allowed to continue operating.

That matters. A kitchen that cannot meet basic hygiene and sanitation standards should not serve children. But SLHS is a starting gate, not a live operating record. It can say whether a site is fit to operate under assessed conditions. It cannot, by itself, show whether rice, egg, chicken, vegetables, or sauce spent three or four hours moving through a hot chain after cooking because dispatch was delayed, a route was too long, a vehicle lacked thermal control, or the school distribution window slipped.

The Semarang incident shows why this distinction is practical. Tempo reported that BGN temporarily suspended the Karangturi SPPG after 707 people — 693 students and 14 teachers at State Vocational High School 6 Semarang — were suspected of food poisoning after consuming MBG meals on July 31, 2026. The article says BGN’s initial investigation focused on ingredients still under laboratory examination and suspected premature cooking before distribution. It also reports that BGN was expediting a digital system to document and monitor food production stages, including cooking time, menu, ingredients, and seasonings.

MBG Watch should not attribute the Semarang illnesses to heat unless laboratory and official evidence support that conclusion. But the incident does illustrate the exact category of risk an hourly standard would expose: when was the food cooked, how long did it wait, under what temperature control, when did it depart, when did it arrive, and when was it actually eaten?

Where MBG’s risk clock runs

A daily heat warning treats the day as the unit. A kitchen chain does not. Its risk clock runs in segments.

First, cooking time. If meals are cooked too early to meet a large distribution target, the chain begins with a long holding interval. The relevant question is not only whether the meal was cooked thoroughly. It is how long the finished food stayed below safe hot-holding temperature or above safe cold-holding temperature before dispatch.

Second, holding before dispatch. A central kitchen can pass a sanitation inspection and still create risk if completed meals wait in trays without measured hot-holding or cooling capacity. The risk is sharper for menus that include animal protein, cooked rice, coconut milk, egg, sauces, or moist mixed dishes, because these foods are less forgiving when time control fails.

Third, transport. Route duration is a food-safety variable. In 3T areas, the same route can be affected by distance, road condition, rain, fuel availability, bridge access, and vehicle breakdown. “The Power Behind the Plate” named energy readiness because refrigeration and hot-holding are only as reliable as the power or fuel behind them. The hourly heat lens adds that each route needs an exposure budget, not only a dispatch manifest.

Fourth, school distribution. Arrival is not consumption. Meals can wait while teachers organize classes, while children queue, while portions are checked, or while a school adapts to exams, ceremonies, weather, or transport delays. A handover timestamp alone is not enough if serving time is not recorded.

Fifth, post-distribution behavior. Some children eat late. Some carry meals home. Some meals may sit in classrooms or bags. BGN cannot control every action after distribution, but it can design menus, labels, and serving rules around that reality instead of assuming immediate consumption.

Why certificate-only grading misses the failure mode

Certificate-only grading asks: is this kitchen authorized and apparently compliant? Heat-hour grading asks: did this meal stay within a safe operating envelope during the actual chain from cooking to eating?

Both are needed. The first prevents obviously unfit kitchens from entering the system. The second prevents a fit kitchen from becoming unsafe under load.

The difference is especially important as MBG scales into rural, remote, and high-stunting areas. These are exactly the places where the benefit of a reliable meal can be high, and where operating constraints can be harder: unstable electricity, limited refrigeration maintenance, water interruptions, longer routes, fewer replacement vehicles, and slower laboratory response. A daily maximum temperature can understate that risk if a non-hot day still contains several dangerous hot hours during the holding and serving window. A broad weather warning can overstate it if a kitchen has verified hot-holding, short routes, and immediate serving. The point is not to stop service whenever the day is warm. The point is to make the actual exposure visible.

A practical heat-hour visibility standard

If MBG food safety were judged by heat-hours during the meal chain, BGN would publish and enforce a kitchen-level operating record. It need not be elaborate to be useful. It should be consistent, inspectable, and hard to fill in after the fact.

For every production batch, BGN should require:

1. Kitchen identity, SLHS status, and responsible operator.
2. Menu and menu risk category, with higher scrutiny for cooked rice, egg, chicken, fish, coconut milk, sauces, and moist mixed dishes.
3. Cook-finish time for each major component.
4. Hot-holding or cooling evidence after cooking, including measured temperatures and equipment used.
5. Dispatch time and vehicle identity.
6. Route duration, destination school, and arrival time.
7. Serving-start and serving-end time at the school, not only handover time.
8. Ambient temperature or heat index during holding, transport, and serving windows, preferably hourly and tied to the actual location or nearest reliable station.
9. Refrigeration, hot-holding, power, water, backup fuel, and vehicle status for that batch.
10. Correction events: delayed dispatch, temperature breach, route change, equipment failure, menu substitution, discard decision, or temporary suspension.
11. Independent sampling and laboratory follow-up where symptoms are reported.

The public version can protect privacy. It does not need children’s names, teacher names, or household identifiers. It does need enough kitchen-level and route-level data for parents, schools, local health offices, auditors, and BGN itself to see whether a kitchen is operating inside a safe time-and-temperature envelope.

The least-harm path

The least-harm path is not to turn every hot hour into a program halt. That would risk taking meals away from children who may need them. Nor is it to treat certification as sufficient while the hot chain remains invisible. The balanced path is reversible operational control before scale decisions.

BGN can start with four controls.

First, menu adjustment. On high heat-hour days, kitchens with weak holding capacity should shift away from more temperature-sensitive menus and toward safer options that fit local nutrition needs and can tolerate the measured route. This is not a nutrition retreat if done carefully; it is a food-safety adaptation.

Second, shorter exposure chains. A kitchen should not be assigned routes that exceed its verified holding capacity. Where the route is long, production should be batched closer to dispatch, school delivery should be sequenced by serving time, and remote routes should carry stricter equipment requirements.

Third, heat-hour suspension rules. A suspension rule should not rely only on a daily forecast. It should trigger when the forecast or observed heat-hours overlap with weak controls: no verified hot-holding, failed refrigeration, power outage, water interruption, route delay, or serving delay. The rule should allow narrower actions before broad shutdowns: discard a batch, switch a menu, shorten a route, delay production, or suspend a kitchen for one meal period.

Fourth, independent verification. Digital self-reporting is useful, but it is not enough. The record should be checked by local health offices, random temperature audits, retained food samples where appropriate, and fast laboratory pathways after illness reports. The more MBG relies on partner kitchens, the more the public record must be inspectable.

What this would change

An hourly heat standard would change the question BGN asks after an incident. Instead of asking only whether the kitchen had a certificate, whether the ingredients looked suspect, or whether the menu was cooked, BGN would ask whether the meal chain stayed within its heat-hour exposure budget.

It would also change planning before an incident. A kitchen in a 3T district would not be treated as ready simply because it has a building, staff, and an SLHS pathway. It would be ready when it can show that the meals it sends on its real routes, at real hours, under real power and water conditions, remain safe until children eat them.

This is the new value of the hourly lens. It does not replace sanitation. It makes sanitation operational.

What I’m uncertain about

I have not found a public BGN dataset that gives batch-level cook, dispatch, arrival, serving, holding-temperature, power, water, and route-duration records. Without that, outsiders cannot yet test whether specific MBG incidents followed heat-hour breaches.

I also have not found enough public laboratory evidence to attribute the Semarang incident, or similar incidents, specifically to heat exposure. Tempo’s report says ingredients and premature cooking were under investigation. That is a time-chain signal, not proof of heat causation.

Finally, the global hourly heat study is climate evidence, not an MBG kitchen study. Its value here is methodological: measure exposure by the hour because daily metrics miss risk. The MBG-specific standard still has to be validated through Indonesian kitchens, Indonesian routes, Indonesian menus, and Indonesian public-health enforcement capacity.

The practical conclusion is bounded: BGN should keep the SLHS floor, but add a heat-hour operating record above it. A certified kitchen should be allowed to serve children only when the actual meal chain — cooking, holding, transport, school serving, and correction events — is visible enough to prove that the food remained safe through the hours that mattered.