When the Air Is Not Safe: The Smoke and Haze Operating Record MBG Needs

MBG Watch · 2026-08-19

The premise

MBG already has to answer questions at the kitchen door: whether food was kept cold, whether the route was safe, whether a school or home handoff could happen, and whether beneficiaries were told the truth in time. Smoke and haze add a different operating condition. The hazard is not only the fire. It is the air around the kitchen, the route, the school yard, the posyandu line, and the worker who must keep preparing and delivering meals while others are advised to stay indoors.

On 19 August 2026, BMKG’s public air-quality page showed a national PM2.5 monitoring panel with Sintang at 318.1, Kubu Raya at 239.7, Mempawah at 180.6, Banjarbaru at 178.1, and Palangkaraya at 158.5. BMKG’s own category table places PM2.5 above 250.5 in “berbahaya,” 150.5–250.4 in “sangat tidak sehat,” and 55.5–150.4 in “tidak sehat.” BMKG also publishes an “early warning for extreme air quality” and a three-day haze-particulate spread forecast for smoke from forest and land fires.

That does not prove an MBG failure. It does prove that MBG’s operating environment can change within the hours when food is prepared, dispatched, handed over, eaten, or sent by a care route to people outside school. The relevant question is therefore narrow: what must BGN make visible when unsafe air intersects with public nutrition delivery?

What the evidence supports

First, smoke is a health hazard, not only a visibility problem. WHO describes wildfire smoke as a mixture of hazardous pollutants including PM2.5, nitrogen dioxide, ozone, aromatic hydrocarbons, and lead; it also notes that wildfire events can disrupt transport, communications, water supply, power, and other services. WHO’s ambient-air-pollution guidance links outdoor air pollution to respiratory infections, heart disease, stroke, chronic obstructive pulmonary disease, acute lower respiratory infections, and lung cancer, with emerging evidence on pregnancy and birth outcomes.

Second, children and pregnant people are not incidental beneficiaries in this program. They are among the groups for whom smoke exposure matters most. EPA’s wildfire-smoke-and-pregnancy page states that pregnant women exposed to wildfire-smoke PM2.5 faced higher risk of preterm birth than women not exposed, and advises avoiding smoke exposure, checking local air-quality reports, and following instructions for sensitive individuals. This matters directly to MBG’s newer care route for pregnant mothers, breastfeeding mothers, toddlers, and 3T beneficiaries. It is the same logic MBG Watch named in “The Care Has to Travel”: when the beneficiary is not a schoolchild sitting in a classroom, the last mile becomes part of the care itself.

Third, workers need their own record. NIOSH’s wildland-fire-smoke guidance says employers and workers can reduce exposure by monitoring air quality, relocating or rescheduling work to less smoky areas or times, reducing strenuous work where possible, giving breaks in smoke-free places, and accommodating workers indoors or in lower-exposure locations. MBG kitchen workers, drivers, loaders, school handoff staff, and posyandu/home-delivery workers are not protected by a beneficiary-only dashboard. If the public record says only “meals delivered,” it hides whether delivery was achieved by shifting exposure onto workers.

Fourth, Indonesia has current smoke and haze signals that plausibly matter for operations. ASMC activated Alert Level 2 for the southern ASEAN region on 3 August 2026, citing escalating hotspot and smoke-haze activity, moderate smoke haze from clusters in western and southern Kalimantan, smoke plumes in southern Sumatra, and hundreds of hotspots detected in Kalimantan on 1–2 August. Its 18 August regional haze page reported widespread hotspots in West, Central, and South Kalimantan, moderate to dense smoke plumes in those areas, and forecast continued dry conditions in parts of Sumatra, Borneo, Java, and the Lesser Sunda Islands. ANTARA, citing BMKG Sumatera Selatan, reported on 14 August that Palembang’s air quality was at a red, very unhealthy level in the morning because of smoke from forest and land fires; BMKG’s Musi 2 Palembang PM2.5 station reached 144.4 µg/m³, in the unhealthy category.

Fifth, BGN is already publishing some operational transparency, but not this one. Radar MBG shows daily menus and SPPG providers by location. BGN’s own news page also lists recent transparency, food-safety, kitchen-closure, and beneficiary-validation stories, including Radar MBG, food that should be returned if unfit, and kitchen closures. In the material checked for this brief, I did not find a BGN/MBG public record that links air quality or haze to kitchen status, worker protection, route changes, school/posyandu handoff decisions, home-delivery substitution, or meal discard logic. That absence should be treated carefully: it is not proof that no internal SOP exists. It means the public cannot inspect the operating decision from the outside.

What the evidence does not support

The evidence does not support saying that smoke has caused an MBG food-safety incident. I found current Indonesian haze and PM2.5 signals, but not a sourced case showing that MBG meals were contaminated, delayed, discarded, or served unsafely because of smoke.

The evidence also does not support collapsing all smoke-day decisions into one rule such as “stop MBG.” Nutrition support may be more important, not less, during disrupted days, especially for pregnant mothers, toddlers, and children whose households lose income or mobility. The least-harm path is not a slogan. It is a set of proportional controls: shift time, shift location, protect workers, change handoff method, substitute shelf-stable food when needed, and pause only the part of the chain that cannot be made safe.

Nor does the evidence support importing foreign wildfire guidance as if it were an Indonesian regulation. WHO, EPA, and NIOSH are useful because they name the biological and occupational hazards. The operating threshold in Indonesia should come from Indonesian sources where possible: BMKG PM2.5 monitoring and warnings, KLHK or other Indonesian environmental-health standards where they apply, Kemenkes health advice, local disaster instructions, and SPPG-specific conditions.

The minimum smoke-and-haze record BGN could publish

BGN does not need to publish child-level exposure data. It should not. “Seen Without Being Watched” and “Before the Number Becomes a Fact” remain the boundary: public accountability should show the decision record, the source of the number, and the correction trail, not personally identifiable health or movement data.

A minimum public smoke-and-haze operating record could be one row per affected SPPG, service day, and delivery window:

1. Air-quality source and time stamp. The record should name the station or forecast used, the PM2.5 value or category, the source agency, the time observed, and whether it came from BMKG near-real-time monitoring, BMKG early warning, local government guidance, or another named source.

2. Kitchen indoor-air status. The row should say whether the kitchen was operating normally, operating with closed openings and filtration/ventilation controls, preparing only sealed or lower-exposure items, or pausing preparation. If the kitchen lacks filtration or safe enclosed work areas, that should be visible as a readiness gap, not hidden behind meal-count totals.

3. Worker-protection status. The record should state the mitigation used: adjusted shifts, reduced outdoor loading time, protected rest area, mask/respirator provision where appropriate, route rescheduling, or temporary reassignment. This is not a medical disclosure. It is an employer-duty record.

4. Route and handoff decision. The record should distinguish school service, posyandu service, and home/care-route delivery. A school can be open while outdoor queuing is unsafe. A home route can protect a toddler while exposing a driver. A posyandu line can be shortened, moved indoors, split by appointment, or replaced with a safer handoff. Each is a different decision.

5. Meal substitution or discard logic. The record should say whether meals were served as planned, replaced with a safer sealed option, held for a shorter window, returned, or discarded. It should distinguish respiratory risk from food contamination risk. Poor ambient air does not automatically mean a sealed meal is unsafe; visible smoke or ash exposure during preparation, staging, or open handoff may create a different decision.

6. Beneficiary communication. The record should show when schools, posyandu cadres, pregnant mothers, parents, and local health workers were told of the change. It should include the public message template, not names or phone numbers.

7. Correction trail. If a PM2.5 reading was wrong, a source changed, a route was misclassified, or a meal count was later corrected, the row should keep the earlier entry and add the correction. That is the provenance standard applied to smoke days.

This record would extend, not replace, MBG Watch’s earlier standards. “Hourly Heat, Not Daily Heat” argued that food-safety risk changes by the hour. “Weather at the Dispatch Door” argued that routes need a stop-go record. “The Power Behind the Plate” and “Not the Panel, the Cold Chain” argued that readiness is the operating capacity behind the public promise. Smoke and haze belong in the same family: not as a new political argument, but as a condition that changes whether the meal can be prepared, carried, received, and eaten with least harm.

The least-harm operating choices

A good smoke-day MBG decision does not start with a national declaration. It starts with the local operating unit and a time-stamped threshold.

If air quality is moderate but worsening, an SPPG may continue with monitoring, closed staging, worker breaks, and an advance message to schools and cadres. If it reaches unhealthy levels, outdoor waiting, open-air staging, and strenuous loading should be reduced; handoffs should move indoors where possible; delivery windows should shorten; workers should have protected rest and masks suited to the risk. If it reaches very unhealthy or hazardous levels, MBG should have authority to change the service mode: sealed substitution, delayed handoff, school/posyandu pickup by appointment, home delivery only where it reduces total exposure, or temporary pause for the unsafe route segment.

The important distinction is that continuity of nutrition support is not the same as continuity of the usual route. During smoke, the program may need to keep feeding while changing the pathway. That is exactly why the care route for pregnant mothers and toddlers needs an air-quality field: not to surveil the beneficiary, but to document why today’s care had to travel differently.

What remains uncertain

I am uncertain whether BGN has internal smoke or haze SOPs that are not public. The material checked here did not surface a public MBG air-quality operating record, but absence from public search is not proof of absence inside BGN or local government.

I am also uncertain how consistently BMKG station coverage maps onto every SPPG catchment area. A public record should therefore name the source used and allow a “nearest available station / forecast product” field, rather than pretending every kitchen has a monitor at the door.

Finally, I am uncertain how often smoke-day risk will change the meal itself rather than the route, staging, or worker-protection plan. That is why the record should not assume contamination. It should force the decision to be inspectable: what was the air condition, which part of the chain was affected, what changed, who was told, and what was corrected later.

The air is not always part of a nutrition program. On smoke days, it is. MBG’s accountability standard should be able to say, calmly and publicly, how the program knew that — and what it changed before the meal reached the child, the pregnant mother, the toddler, or the worker carrying it.