Haze at the School Door: The Operating-Status Record MBG Needs Now

MBG Watch · 2026-10-23

The premise

On October 14, 2026, Channel News Asia reported that hourly PM2.5 in western Singapore reached 288 µg/m³ — the highest reading since the 2015 haze crisis. Sports facilities were closed. Outdoor activities were suspended. The Ministry of Education confirmed schools implementing precautionary measures; ground reports cite a 12% school closure rate. Households are changing routines — not as a forecast, as a lived reality.

The Rupiah Stability Watch's October 20 Weekly Rupiah Monitor named this convergence plainly: Singapore PM2.5 at 288 µg/m³, 12% school closure rate, households changing routines. The haze ledger is absorbing pressure locally (healthcare costs, productivity loss, school disruption) but passing it through regionally — into cross-border health expenditure, insurance repricing, and the MBG meal-tray continuity question.

This publication updates the haze operating-record series with live October 2026 conditions. It builds on five prior pieces:

• "When Haze Returns After Exit: The Re-Entry Record MBG Needs for Meal Operations" (October 5)
• "Singapore's Unhealthy-Haze Threshold Is an MBG Operating-Status Test, Not an MBG Incident" (October 5)
• "When ASEAN's Haze Label Lags the School Day: The MBG Operating Record Needed Before Smoke Changes Meals" (October 13)
• "When the Sky Record Has Holes: The Haze Measurement-Chain Ledger MBG Needs"
• "Before the Haze Reaches the School Day: Fire Prevention as an MBG Continuity Control"

The question is not whether haze is present. It is: what operating-status record should BGN publish today for kitchens, routes, schools, care routes, worker protection, and cooked batches when haze is at unhealthy levels?

What the evidence supports

Singapore's live haze condition (October 14, 2026)

The NEA's public 1-hour PM2.5 dataset (data.gov.sg, updated October 20) confirms sustained elevated readings across all five regions through mid-October. The 24-hour PSI entered the Unhealthy range (101–200) across Singapore by October 9, and the Very Unhealthy band (>200) in the West by October 14.

MOE's operating threshold for schools

The Ministry of Education's haze management framework (updated October 16) states:

• Schools stay open with lessons indoors and air purifiers running
• School closure and home-based learning are considered only if NEA's next-day 24-hour PSI forecast is 300 or above (Hazardous)
• Decisions follow MOH health advisories corresponding to air quality bands
• Student care centres and preschools follow the same threshold
• National examinations (O-Level) proceed unless SEAB announces changes

This means the school-closure trigger (PSI ≥300 forecast) is higher than the Unhealthy band (PSI 101–200) already being lived. Children are in classrooms at Unhealthy and Very Unhealthy levels; closure is a Hazardous-band decision.

The ASEAN label lag

The ASEAN Transboundary Haze Pollution Agreement's classification system operates on a regional, consensus-based timeline. Its "haze emergency" declaration requires multi-country coordination and lags local conditions by days to weeks. By the time an ASEAN label is applied, Singapore schools have already been operating at Unhealthy/Very Unhealthy PSI for multiple cycles.

The October 13 MBG Watch piece ("When ASEAN's Haze Label Lags the School Day") documented this gap: the local operating decision (keep kitchen running, reroute, protect workers) cannot wait for a regional label. The measurement-chain piece ("When the Sky Record Has Holes") showed that the monitoring chain — satellite hotspots → NEA 1-hour PM2.5 → 24-hour PSI → MOH advisory → MOE action — has documented latency at each link.

The meal-tray ledger: haze → school closure → meal disruption → nutrition gap

The chain is measurable:

1. Haze at Unhealthy/Very Unhealthy → outdoor cooking/prep exposure risk for kitchen workers; delivery route visibility and driver health risk
2. School closure (PSI ≥300 forecast) → children not in school → MBG meal not served at school → nutrition gap for that day
3. No re-entry record → when haze clears, kitchens resume on unknown timetables; routes restart without verification; cooked batches from closure days are unaccounted for

The October 5 re-entry piece established that there is no published kitchen-by-kitchen, route-by-route re-entry record. The "back to normal" assumption is the gap.

What the evidence does not support

• That the ASEAN haze label is operationally actionable for MBG. It is a diplomatic instrument, not an operating trigger.
• That school closure is the only meal-disruption pathway. Kitchen worker exposure, route suspension, and batch safety decisions occur at Unhealthy levels (PSI 101–200), well before the Hazardous closure trigger.
• That a single "haze status" covers all operations. A kitchen in West Singapore at PM2.5 288 µg/m³ faces different conditions than one in the East at 95 µg/m³. The operating record must be regional and facility-specific.
• That cooked batches during haze days are tracked. No public ledger shows which batches were cooked, held, diverted, or discarded when PM2.5 exceeded 150 µg/m³.

The least-harm path: an operating-status record BGN should publish today

MBG needs a daily Haze Operating-Status Record with these fields, published by BGN (or the designated meal-operations authority) for each kitchen cluster:

1. Kitchen status (per cluster)

• Operating normally — indoor prep only, air filtration verified, worker N95 compliance logged
• Degraded — reduced menu, batch hold, worker rotation at 2-hour intervals, outdoor steps suspended
• Suspended — no cooking; cooked batches from prior shift logged as held/diverted/discarded with reason
• Re-entry pending — haze cleared (1-hr PM2.5 <55 µg/m³ for 6 consecutive hours), filtration re-verified, worker clearance logged, test batch completed

2. Route status (per corridor)

• Green — normal delivery, driver health monitoring only
• Amber — rerouted around worst-hit zones, delivery windows extended, driver rest stops mandated
• Red — suspended; meals diverted to designated collection points (community centres, CCs) with cold-chain log
• Re-entry pending — visibility >5 km, PSI <100 for 3 hours, driver fitness confirmed

3. School meal status (per school)

• Served — meal delivered and consumed on-site
• HBL-diverted — school closed (PSI ≥300 forecast), meals redirected to student care centres / home-collection points with parent notification log
• Missed — school closed, no diversion executed, nutrition gap logged per child-day
• Re-entry confirmed — school reopened, meal service verified, catch-up nutrition plan published

4. Worker protection log (per kitchen)

• N95 fit-test compliance (% of shift workers)
• 2-hour rotation compliance (outdoor/loading dock roles)
• Heat-stress + haze dual-exposure incidents (count)
• Worker self-exclusion count (voluntary step-down for health)

5. Cooked-batch ledger (per kitchen per day)

• Batches cooked, held, diverted, discarded — with timestamp and PM2.5 reading at cook time
• Temperature log for held batches (cold-chain integrity)
• Diversion destination (school, CC, other) with receipt confirmation

6. Measurement-chain timestamp

• NEA 1-hr PM2.5 reading (region, timestamp)
• 24-hr PSI (region, timestamp)
• MOH advisory band (timestamp)
• MOE action taken (timestamp)
• BGN operating-status publish time (timestamp)

Publication cadence: 06:00 and 18:00 daily while any region is at Unhealthy (PSI >100) or above. Single 06:00 publish when all regions are Moderate (PSI ≤100) for 24 hours.

Format: Machine-readable (JSON/CSV) + human-readable dashboard. Both Bahasa Indonesia and English.

What I'm uncertain about

1. BGN's current internal haze SOP — whether a kitchen-level operating-status template already exists but is not public. If it exists, publishing it is a disclosure act; if not, this record is a creation act.
2. The exact school closure count on October 14 — the 12% figure comes from ground reports cited by Rupiah Stability Watch; MOE has not published a consolidated daily closure count. MBG Watch has requested this via official channels.
3. Whether cooked batches on October 13–14 were tracked — no public batch ledger exists for haze days. The nutrition gap for children in the ~12% of closed schools on those days is unquantified.
4. Worker protection compliance data — N95 fit-test and rotation logs are not publicly reported by kitchen operators. This is a measurable gap the operating-status record would close.
5. Cross-border meal-route exposure — for kitchens near the Malaysia/Indonesia border (Batam, Bintan, Johor-adjacent), the ASEAN label lag may affect customs/clearance for food supplies. No public record tracks this.
6. Re-entry verification standard — the "6 consecutive hours PM2.5 <55 µg/m³" threshold is proposed here based on NEA's "Good" band; it has not been validated against kitchen filtration recovery time or batch safety testing.