Résumé : Due to the large yet uncertain source of methanol, its atmospheric oxidation by hydroxyl radicals (OH) affects the oxidizing capacity of the atmosphere (and therefore the lifetime of the climate gas methane) and contributes to the budget of formaldehyde. Here we report a global budget of methanol constrained by multi-platform observations. Our focus is on continental sources, and particularly on biogenic emissions which represent by far the largest component of the global budget. The spaceborne methanol column data newly retrieved from the IASI satellite sensor are used as constraints on biogenic and pyrogenic emissions in the global chemistry-transport model MAGRITTEv1.1 (Müller et al. 2019). The IASI data are based on an improved version of the Artificial Neural Network for IASI (ANNI) retrieval framework, which relies on a hyperspectral range index (HRI) for the quantification of the gas spectral signature and on an artificial feedforward neural network to convert the HRI into a gas total column (Franco et al. 2018). The vertical profile shapes of methanol concentrations used in ANNI rely on a synthesis of aircraft measurements over land and ocean. Direct comparison of FTIR methanol columns with co-located IASI data (at St Petersburg, Jungfraujoch, Toronto, Porto Velho and Reunion Island) shows good agreement (r~0.8) despite a slight underestimation of large columns (>5x1016 molec.cm-2).The MAGRITTE model and its adjoint are used to derive top-down methanol emissions constrained by IASI data over land for several years between 2010 and 2019. The extratropical biogenic emissions are substantially enhanced (+80%) by the inversion, while tropical emissions decrease over rainforests (e.g. Indonesia) and increase in dry ecosystems. The model is also used as intercomparison platform to assess the consistency between the IASI, FTIR, aircraft and surface in situ data. Additional inversion experiments are conducted using adjusted IASI datasets (a) bias-corrected against FTIR columns, and (b) bias-corrected against aircraft data over Northern America from the campaigns SEAC4RS, SENEX, DISCOVER and NOMADSS. The comparison indicates the existence of a probable bias between optical and in situ methanol data, which remains so far unresolved.