Hydrothermal systems can generate phreatic and/or phreatomagmatic explosions with little warning. Understanding the temporal and spatial evolution of geophysical and geochemical signals at hydrothermal systems is crucial for detecting precursory signs to unrest and to inform on hazard. Thermal signatures of such systems are poorly defined because data records are often too short or discrete compared to activity timescales, which can be decadal. La Fossa system of Vulcano has been monitored since the 1980s and entered a period of unrest in 2021. We assessed the thermal signature of La Fossa using ground-and satellite-based data with various temporal and spatial scales. While continuously-recording stations provided continuous but point-based measurements, fumarole field vent surveys and infrared images obtained from satellite-flown sensors (ASTER and VIIRS) allowed lower temporal resolution but synoptic records to be built. By integrating this multi-resolution data set, precursory signs of unrest could retrospectively be detected from February to June 2021. Intensity of all unrest metrics increased during summer 2021, with an onset over a few days in September 2021. By September, seismic, CO 2 , SO 2 and other geochemical metrics also indicated unrest, leading Civil Protection to raise the alert level to yellow on October 1. Heat flux, having been 4 MW in May 2019, increasing to 90 MW by September, and peaking at 120 MW in March 2022. We convolved our thermal data sets with all other monitoring data to validate a Vulcano Fossa Unrest Index (VFUI), framework of which can be potentially applied to any hydrothermal system. The VFUI highlighted four stages of unrest, none of which were clear in any single data set: background, precursory, onset and unrest. Onset was characterized by sudden release of fluids, likely caused by failure of sealed zones that had become pressurized during the precursory phase that began possibly as early as February 2021. Unrest has been ongoing for more than 18 months, and may continue for several more years. Our understanding of this system behavior has been due to hindsight, but demonstrates how multiparametric surveys can track and forecast unrest.