Receptor-scale imprint of barometric pressure variability on community air monitoring near an elevated-temperature landfill: a multi-method analysis across two mitigation interventions.

Barometric pressure is an established control on landfill gas emission, with the rate of pressure change rather than absolute pressure governing advective transport through cover soils. It has been characterised by onsite flux chambers and eddy covariance, and by the tracer correlation method, which observes the whole-site plume from downwind but is campaign-based. Whether the dependence is recoverable from continuous downwind monitoring has not been tested. Two years of continuous hourly speciated monitoring at five community stations downwind of an elevated-temperature landfill in southern California were analysed, spanning gas-collection expansion and closure. Four methods with non-overlapping limitations were applied to a common analytical panel: conditional Granger causality, effective transfer entropy, wavelet coherence, and change-point detection on raw and meteorologically residualised series. A validity screen excluding physically impossible and censored observations retained 48 of 60 station-species-regime combinations. Sector-resolved correlation against an onsite station confirmed that community receptors detect the source, with correlations rising from 0.08 to 0.40 and from 0.17 to 0.51 at two stations when wind direction was resolved. Pressure-associated predictive structure conditional on wind speed was identified in 58 per cent of cells, retaining 54 per cent after false-discovery correction, with detection rising monotonically across the intervention sequence. Effective transfer entropy recovered five well-conditioned cells that linear analysis classified as null, and an inter-method rank correlation of 0.40 confirmed non-redundancy. Wavelet coherence located coupling in the diurnal and multi-day bands, consistent with the proposed transport mechanism. Change-point structure survived removal of meteorology, but only 12 per cent of breaks aligned with documented interventions, indicating that interventions altered pressure coupling without correspondingly shifting concentration level. Species resolved to distinct source sectors, and propene relocated north-west after closure at three stations while isopropyl alcohol did not, indicating spatially separable and differentially responsive source areas within a single landfill footprint.