Why Bandung heard the September 2026 Anak Krakatau eruption and Jakarta did not: eruption timing, ray tracing and search-interest evidence, compared with April 2020

Ryan Andri Wijaya 

During the night of 4–5 September 2026, residents of the Bandung area, 240–300 km from Anak Krakatau, reported repeated booms and rattling windows, while Jakarta, 150 km from the volcano, reported almost none. The Indonesian meteorological agency (BMKG) first judged the volcano too distant and the winds unfavourable, then attributed the booms to the eruption. We test the volcanic explanation with three independent lines of open evidence. Timing: the volcanological agency (PVMBG) recorded continuous eruption tremor from 23:07 WIB on 4 September to 00:04 WIB on 6 September, and Darwin VAAC advisories show the ash column rising from about 2 km to 13.7–15.2 km between 23:20 and 01:10 WIB; the booms, reported from about 23:30 to 05:00 WIB, fall within this high-intensity phase and are consistent with atmospheric travel times. Propagation: ray tracing through ERA5 extended upward with the HWM14 wind and NRLMSIS atmosphere models, with 200 random gravity-wave realisations, returns sound to Bandung in 88 % of ensemble members at the eruption hour and in 84–97 % in every hour of the reported window, mostly through eastward winds near the stratopause (about 55–60 km), whereas Bogor (11 %), Jakarta (3 %) and Serang (0 %) lie in a shadow zone; with a windless upper atmosphere the shadow zone persists but the return to Bandung is thermospheric, a path along which absorption would leave only sub-audible infrasound. Search behaviour: all 26 cities with measurable Google search interest in “dentuman” (boom) lay 206–371 km from the volcano, with 96 % of the interest at 200–300 km, while every large city within 200 km showed none; searches for “krakatau” instead declined with distance from the volcano and peaked a day later. Applied to a 2025 meteor of known location, the same measure placed 96 % of “dentuman” interest within 100 km of it. For the April 2020 eruption, whose volcanic origin PVMBG rejected in favour of thunderstorms, eastward stratopause winds also favoured returns, at about 150–210 km; 70 % of “dentuman” interest lay at 100–200 km, and rain within thunder range of Jakarta was light during the reported booms, although a thunderstorm had passed close to Jakarta about two hours earlier. A matched-control analysis of eleven boom reports shows that low-level acoustic ducts occur on most nights and cannot by themselves explain when booms are heard. Long-range sound and infrasound from sustained explosive eruptions, returning beyond a shadow zone, explain both episodes, the 2020 case with less certainty.