The main objectives of this study were to provide a proxy “validation” of the Soil Moisture and Ocean Salinity (SMOS) mission's vegetation optical depth product (τSMOS) on a global scale, to give a first indication of the potential of τSMOS to capture large-scale vegetation dynamics, and to contribute towards investigations into the possible use of optical vegetation indices (VI's) for the estimation of τ. The analyses were performed by comparing the spatial and temporal behaviour of τSMOS relative to four MODIS-based VI's, with that of the vegetation optical depth from a similar sensor, AMSR-E (τAMSR-E). 16-day and annual average values of the passive microwave optical depth (τ) for the year 2010 were obtained from SMOS (1.4 GHz) and AMSR-E (6.9 GHz) observations. The VI's chosen for this study were the Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Leaf Area Index (LAI) and Normalized Difference Water Index (NDWI). The highest global-scale, annual correlation was found between τSMOS and τAMSR-E from ascending orbits (Spearman's R = 0.80). On global, annual scales, τSMOS showed higher correlations with τAMSR-E than with the VI's, while τAMSR-E was more highly correlated with VI's than with τSMOS. Timeseries of both τ and the VI's were made per landcover class, for the northern hemisphere, tropics and southern hemisphere. Although the large-scale spatial and spatio-temporal behaviour of τSMOS and τAMSR-E is generally similar, the results highlight some notable differences in observing vegetation with optical vs. passive microwave sensors, and certain crucial differences between the two passive microwave sensors themselves. Overall, the results found in this study give a good first confidence in the SMOS L3 τ product and its potential use in vegetation studies. These results provide an essential general reference for future (global-scale) vegetation monitoring with passive microwaves, for the future inclusion of τSMOS in long-term, multi-sensor datasets, and for passive microwave algorithm development.< Réduire