Singapore’s Unhealthy-Haze Threshold Is an MBG Operating-Status Test, Not an MBG Incident

MBG Watch · 2026-10-05

The premise

A neighboring unhealthy-air threshold is not an Indonesian MBG incident. It is a rehearsal prompt.

On 5 October 2026, Singapore’s air quality crossed into the unhealthy range. CNA reported that by midnight on 6 October, the 24-hour PSI reached 136 in central Singapore, 125 in the west, 114 in the east, 108 in the north, and 103 in the south; Singapore classifies a 24-hour PSI of 101–200 as “unhealthy.” The same report noted that 1-hour PM2.5 readings were elevated in all regions except the south at midnight, and that Singapore advises people to use the 1-hour PM2.5 guide for immediate outdoor activity and the 24-hour PSI forecast for next-day planning.

Singapore’s own meteorological update earlier that evening put the 1-hour PM2.5 concentration at 75–97 µg/m³ and the 24-hour PSI at 94–122 as of 6 p.m., in the moderate-to-unhealthy range. It also described moderate to dense smoke haze over southern Sumatra and Kalimantan, with a continued risk of transboundary haze.

That is enough to matter for MBG Watch, but not because it proves an MBG failure. The point is narrower and more operational: once air quality becomes a service condition, a meal program needs a public operating-status record. Haze can change school attendance, outdoor work, delivery routes, parent pickup, already-cooked meal handling, care routes for pregnant mothers and toddlers, and the notice chain families rely on.

This piece builds on MBG Watch’s recent haze series: “When the Air Is Not Safe: The Smoke and Haze Operating Record MBG Needs,” “Before the Haze Reaches the School Day: Fire Prevention as an MBG Continuity Control,” “When Haze Moves the School Day: The MBG Operating-Status Record Needed Now,” “When Haze Returns After Exit: The Re-Entry Record MBG Needs for Meal Operations,” and “When the Sky Record Has Holes: The Haze Measurement-Chain Ledger MBG Needs.” It also sits beside Rupiah Stability Watch’s 5 October analysis, “Singapore’s Haze Threshold and the Rupiah Operating Ledger,” which made the adjacent point that MBG/SPPG continuity becomes part of a wider operating ledger when haze affects schools, workers, routes, and trust.

What the evidence supports

The evidence supports four careful conclusions.

First, Singapore had a publicly visible haze-status threshold. Singapore’s government materials use 24-hour PSI for daily health advisories and next-day planning, and CNA reported the 101–200 band as “unhealthy.” Singapore’s Ministry of Health states that PSI 101 and above is unhealthy, while also emphasizing that health impact depends on the PSI level, a person’s health status, and the duration and intensity of outdoor activity.

Second, Singapore’s response was institutional, not only meteorological. The National Environment Agency’s September haze notice said the Government Haze Task Force had rolled out agency action plans and advisories once the 24-hour PSI entered the unhealthy range. That notice named different operating domains: health advisories, preschools, schools, examinations, workplaces, transport operators, and event organizers. For schools, it said MOE and schools had plans to protect students and staff, take reference from NEA readings, and use air purifiers in classrooms. It also said national examinations would continue in enclosed spaces with air purifiers if the air quality was unhealthy or worse, while closures would trigger rescheduling.

Third, the regional haze situation was not speculative. ASMC’s regional haze page carried Alert Level 3 for the southern ASEAN region and warned that dry weather was expected to continue over much of the southern ASEAN region, with the hotspot and smoke-haze situation remaining escalated and continued risk of transboundary haze. Singapore’s 5 October weather update described smoke plumes over southern Sumatra moving north toward parts of Peninsular Malaysia and Singapore, with moderate to dense smoke haze also observed over Kalimantan.

Fourth, Indonesia had its own air-quality warning signal, separate from the Singapore readings. Tempo, citing BMKG data, reported that PM2.5 at the Musi 2 Station in Palembang reached 903.6 µg/m³ on the morning of 4 October, crossing the hazardous level, and that residents were urged to limit outdoor activity, especially those with respiratory problems. This is not MBG evidence by itself. It is relevant because South Sumatra haze is a domestic service-condition signal: when the air becomes unsafe locally, meal operations cannot be treated as if only the menu matters.

What the evidence does not support

The evidence does not support claiming that Singapore’s unhealthy PSI proves an Indonesian MBG disruption.

It does not show that any named MBG kitchen paused, continued unsafely, discarded meals, or failed to notify families. It does not identify an SPPG route delayed by haze. It does not show a child, pregnant mother, toddler, kitchen worker, driver, or household harmed by MBG activity during the 5 October haze episode.

It also does not prove that the correct response is blanket suspension. Singapore’s public materials do not reduce haze response to a single switch. They distinguish immediate outdoor activity from next-day planning, vulnerable people from healthy adults, indoor from outdoor exposure, schools from workplaces, and hazardous conditions from unhealthy ones. That is the useful lesson for MBG: continuity decisions need thresholds, local status, worker protection, and household notice — not automatic reassurance and not automatic shutdown.

The MBG operating-status record BGN should be able to show

When haze affects a school day, a posyandu route, or an SPPG delivery corridor, BGN should be able to publish a local record with eight fields.

1. Air-quality source and threshold used. The record should say whether the decision used BMKG, local environmental agency stations, satellite/hotspot assessment, school-level observation, or another named source; whether the trigger was PM2.5, PSI/ISPU, visibility, smoke plume, or a local emergency declaration; and what threshold changed the operating status.

2. School, posyandu, and care-route operating status. Families need to know whether the site is open, partially open, remote, shifted to pickup, or closed for the day. The status should be local and timestamped, not a national sentence that hides district variation.

3. Meal-distribution decision. The record should state whether MBG distribution continued, paused, shifted to parent pickup, used a substitute meal, moved to a different delivery window, or rerouted to a safer location. “Continued” is not enough; continued how, for whom, and under what exposure control?

4. Already-cooked meal handling. Haze can interrupt the day after food is prepared. BGN should state the discard rule, holding-time limit, temperature requirement, packaging boundary, and who authorized release or disposal. The risk here is not only waste; it is quiet improvisation.

5. Kitchen and driver worker protection. The record should say how kitchen ventilation, indoor air, masks or respirators where appropriate, delivery duration, vehicle exposure, rest, and reassignment were handled. A meal-continuity plan that protects beneficiaries by shifting risk to drivers and kitchen staff is not a complete plan.

6. Parent and community notice chain. The public should know when the notice went out, through what channel, in which language, and who was responsible for updating it. Haze decisions create household logistics: pickup, timing, substitutes, and whether a child should be outdoors.

7. Restart and re-entry criteria. If distribution pauses or shifts, the record should say what measurement or authority restarts normal service, whether a second clean reading is required, and how long the condition must hold. MBG Watch’s earlier re-entry piece argued this point directly: the moment after the air improves is still an operating decision, not an assumption.

8. Privacy boundary. The record should count affected sites, routes, workers, children, pregnant mothers, toddlers, and households without exposing names, health conditions, addresses, or stigmatizing vulnerability. Public accountability should not turn protected people into evidence objects.

The least-harm guarantee

The least-harm public guarantee is not blanket suspension. It is not blanket reassurance either.

It is this: when haze changes the safety of a meal day, every affected community should be able to see a local, timestamped MBG operating status — the air-quality source, the threshold, the meal decision, the worker-protection measure, the notice chain, and the restart rule.

That guarantee would let families plan. It would let oversight bodies distinguish prevention from improvisation. It would also protect BGN from a false binary: being accused of indifference if meals continue, or waste if meals pause. The record would show the actual decision and the reason for it.

What I am uncertain about

The main uncertainty is the current completeness of BGN’s own local haze operating records. Public reporting around Singapore and regional haze is stronger than public reporting around MBG route-level continuity under haze.

A second uncertainty is measurement comparability. Singapore’s PSI, Singapore’s 1-hour PM2.5 bands, Indonesia’s PM2.5 readings, ISPU categories, school-level conditions, and route visibility do not map cleanly onto one another. BGN does not need to copy Singapore’s threshold. It does need to name the threshold it uses.

The third uncertainty is implementation burden. A record with eight fields is only useful if it can be produced by ordinary local operators during a disrupted day. That argues for a simple template, not a new reporting bureaucracy.

The signal from Singapore is therefore modest but important. Nearby unhealthy air does not make an MBG case. It shows what any large meal-service system must be ready to say when the sky becomes part of the operating environment.