Respiratory Illness, Weather and Air Pollution in Greater Jakarta During Indonesia's Fire and Haze Season: An Ecological Assessment With National Influenza Forecasting

Background Indonesia’s 2026 forest-fire and haze season began early and has coincided with concerns about respiratory illness and rising influenza activity. While we know that exposure to PM2.5 in fire-affected areas may increase respiratory health risks and compromise airway defenses, influenza A detections have also risen in areas not yet directly affected by the consequences of fires, including Greater Jakarta and its surrounding areas. These concurrent trends suggest that respiratory illness in Greater Jakarta should be interpreted in light of infection, particulate exposure, and weather-related changes. Objective We aim to provide an exploratory national influenza A positivity forecast through publicly accessible data on respiratory and environmental indicators within the range of 16 July 2026 to 16 October 2026, and offering a repeatable methodology for ongoing local monitoring. Methods We compared 13 national surveillance reports and local public communications within an equal 28-day windows and examined 92 complete days of ECMWF IFS weather and Jakarta CAMS atmospheric fields. In contrast to persistence and random–walk baselines, a beta-binomial state model with damped logit drift was supported by eleven overlapping specimen weeks. We also provided separate weather and particulate model forecasts for environmental context. Results Relative temperature in Jakarta went from 29.24°C to 29.36°C, relative humidity went from 61.0% to 68.4%, and wind speed went from 6.59 km/h to 5.49 km/h. Modeled PM2.5 rose from 42.7 µg/m³ to 50.3 µg/m³; though not a station-confirmed increase. The national influenza A positive cases increased from 12/115 (10.4%) to 2025/337 (60.8%) in the August 15, 2026, and October 3, 2026, reports, respectively. An unresolved difference in the percentage denominator was found in a later study. The approximate projection of the exploratory model estimates 72% positivity for the week ending October 24, 2026 (95% prediction interval, range 49% − 89%). However, this refers to specimens collected and tested for national surveillance, not Jakarta population risk. Errors rose with lead time and there was a limitation for back-testing. Conclusions Available evidence does not prove a local causal relationship or a smoke-attributable illness burden, but it does show contemporaneous viral activity and shifting environmental conditions. Geographically-matched disease counts, pathogen-specific testing denominators, calibrated pollution measures, and report-vintage tracking are all necessary for ongoing monitoring needs. The forecast is a research example that has to be evaluated in the future.