From MSG-SEVIRI to MTG-FCI: Advancing Volcanic Thermal Monitoring from Geostationary Satellites

Highlights

What are the main findings?

• The MTG-FCI sensor provides enhanced spatial, spectral, and temporal resolution compared to MSG-SEVIRI, enabling more continuous and detailed observation of volcanic activity.
• The RSDF algorithm was adapted to FCI data and successfully applied to Mount Etna’s 2025 eruptions, allowing accurate retrieval of volcanic radiative parameters.
• The combination of FCI and polar-orbiting satellite data ensures consistent quantitative estimates of volcanic radiance and thermal anomalies.

What are the implications of the main findings?

• The improved performance of MTG-FCI enhances the capability for near-real-time monitoring of active volcanoes from geostationary orbit.
• These results contribute to the development of advanced operational systems for volcanic hazard detection and early-warning applications based on next-generation satellite missions.

Abstract

Continuous global monitoring of volcanic activity from space requires balancing spatial and temporal resolution, a long-standing trade-off between polar-orbiting and geostationary satellites. Polar sensors such as MODIS, VIIRS, and SLSTR provide high spatial resolution (375 m–1 km) but with limited temporal coverage. In contrast, geostationary sensors like SEVIRI offer high temporal resolution (5–15 min) but with coarser spatial detail (~3 km), often missing lower-intensity thermal events. The recently launched Flexible Combined Imager (FCI) aboard the geostationary Meteosat Third Generation (MTG-I) satellite represents a major improvement, providing images every 10 min with a spatial resolution of 1–2 km, comparable to that of polar orbiters. Here, we adapted the established Remote Sensing Data Fusion (RSDF) algorithm to exploit the enhanced capabilities of FCI for detecting volcanic thermal anomalies and estimating Volcanic Radiative Power (VRP). The algorithm was applied to Mount Etna during three different eruptive phases that occurred in 2025. The VRP derived from FCI data was compared with that obtained from the geostationary SEVIRI and the polar-orbiting MODIS, SLSTR, and VIIRS sensors. The results show that FCI provides a more detailed and continuous characterization of volcanic thermal output than SEVIRI, while maintaining close agreement with polar sensors. These findings confirm the capability of FCI to deliver high-frequency, high-resolution thermal monitoring, representing a major step toward operational, near-real-time volcanic surveillance from space.

1. Introduction

The wide variety of satellites orbiting Earth provides a continuous view of active volcanoes worldwide [1,2,3]. Thermal infrared sensors on board satellite platforms provide accurate estimates of high-temperature volcanic features, fundamental to improving our understanding of volcanic processes [4,5]. Satellite-based volcanic hotspot detection relies on algorithms that utilize various techniques, including spatial-based (or contextual) algorithms [6,7,8,9], temporal-based algorithms [10,11,12,13], spectral-based (or fixed-threshold) algorithms [14,15], and machine learning techniques [16,17,18]. Then, the volcanic radiative power (VRP) of the thermal anomalies can be quantified using the Middle Infrared (MIR) radiance method developed by Wooster et al. 2003 [19]. The VRP measures the heat radiated by volcanic activity, analogous to the Fire Radiative Power (FRP) [20,21] measure used for fire detection.

Reliably detecting thermal changes associated with volcanic eruptions requires high spatial resolution thermal imaging, which is typically provided by polar-orbiting satellites [22,23,24,25,26]. These sun-synchronous satellites orbit relatively close to Earth, offering moderate to high spatial resolution but limited temporal coverage. However, high temporal resolution is also essential for promptly detecting thermal anomalies and following the course of an eruptive crisis [27,28,29]. This is where geostationary satellites play a key role. Being geosynchronous, they continuously observe the same area, providing frequent updates every 10–30 min, though at the cost of lower spatial resolution [30].

The Spinning Enhanced Visible and Infrared Imager (SEVIRI), aboard the Meteosat Second Generation (MSG) satellites operated by EUMETSAT, has been extensively used for near-real-time volcanic monitoring and for tracking short-lived events such as lava fountains [31,32,33,34], particularly over Europe (e.g., Etna and Stromboli) [35,36] and Africa (e.g., Nyiragongo, Nyamuragira) [37,38]. Its high temporal resolution (one image every 15 min, or 5 min in rapid scan mode) makes it valuable for operational surveillance. Nevertheless, its relatively coarse spatial resolution (3 km at nadir) limits the detection of low-intensity thermal activity [39].

The launch of the Flexible Combined Imager (FCI) aboard the Meteosat Third Generation—Imaging (MTG-I) satellites in December 2022 marked a major technological advance. FCI offers improved spatial and temporal resolution compared with SEVIRI, as well as new spectral bands relevant for volcanic thermal monitoring [40]. Whereas SEVIRI operates with 12 spectral channels, FCI expands this to 16 channels, covering wavelengths from 0.4 to 13.3 µm and achieving spatial resolutions of 1–2 km. This is comparable to the nominal 1 km resolution of thermal bands on polar-orbiting sensors such as the Moderate Resolution Imaging Spectroradiometer (MODIS) aboard NASA’s Terra and Aqua satellites, and the Sea and Land Surface Temperature Radiometer (SLSTR) onboard the Copernicus Sentinel-3 missions. MODIS and SLSTR provide excellent spatial detail but limited temporal sampling, typically one or two passes per day [41]. In contrast, FCI delivers comparable spatial detail continuously, with full-disk imagery every 10 min (and even faster in rapid-scan mode over Europe). These features make FCI an innovative instrument for monitoring highly dynamic volcanic phenomena [42,43].

This study leverages the enhanced capabilities of the new FCI sensor to improve volcanic thermal monitoring. Specifically, we adapted the Remote Sensing Data Fusion (RSDF) algorithm [44], originally developed to detect thermal anomalies and estimate the VRP from multiple satellite sources. The RSDF algorithm has previously been applied to data from polar-orbiting sensors (MODIS, SLSTR, VIIRS) and the geostationary SEVIRI. Here, it is applied to FCI data to exploit its improved spatial resolution and extended spectral range, providing a more accurate and continuous characterization of volcanic activity. The algorithm was applied to three eruptive phases of Mount Etna in 2025: (i) February–March effusive activity, (ii) March–June explosive activity, and (iii) August–September effusive activity. The application enabled the comparison between FCI-derived VRP values and those obtained from SEVIRI (with similar temporal resolution) and MODIS, SLSTR, and VIIRS (with comparable spatial resolution). This study represents the first quantitative assessment to validate the capabilities of FCI, establishing its reliability for the detection of high-temperature volcanic features. By demonstrating the advantages of FCI for detecting and quantifying volcanic thermal anomalies, this work paves the way for its integration into future operational monitoring frameworks.

2. Etna 2025 Activity

Etna is a 3404 m high basaltic stratovolcano located in the east coast of Sicily (Italy) [45]. After a period of volcanic activity characterized by variable passive degassing from the summit craters (NEC = North-East Crater, VOR = Voragine crater, BN = Bocca Nuova crater, and SEC = South-East Crater), and approximately three months after the last lava fountain activity on 10 November 2024, Mt. Etna entered a new explosive and effusive eruptive phase on 6 February 2025. The eruption began with modest and episodic Strombolian activity produced by the SEC. Two days later, on 8 February, an effusive phase began, fed by an eruptive fissure that opened at the base of the BN crater at an altitude of ~3050 m a.s.l., with a lava flow advancing towards the southwest (Figure 1) [46]. The lava flow advanced rapidly, reaching an altitude of approximately 2150 m a.s.l on 11 February and 1960 m a.s.l on 13 February [47]. Explosive activity produced by at least two active vents was also observed at SEC [47]. On 19–20 February, lava flows reached an altitude of approximately 1840 m a.s.l. After a reduction in effusive activity and the cessation of explosive activity at the SEC, lava emission resumed on 22 February, accompanied by the return of Strombolian activity at the SEC on 23 February [48]. This activity continued until the morning of the 25th, when Strombolian activity at the SEC ceased again, and the effusion rate decreased significantly. On the evening of 27 February, the SEC reactivated with Strombolian explosions, followed by the resumption of effusive activity from the vent at the base of the BN. On the evening of 28 February, the explosive activity of the SEC gradually decreased and ended during the night. Glows at the effusive fissure below the BN were still visible at dawn on March 1, and a weak active lava flow was observed on 2 March [49].

In March, after two weeks of passive degassing, the SEC produced several Strombolian episodes: the first on 15–16 March [50], followed by further activity on 19–20 March and again on 24 March, when a small lava overflow descended the southern flank of the cone [51].

In early April, intermittent Strombolian activity resumed on 2 April, accompanied in the following days by ash emissions from VOR and BN [52]. A new sequence of short eruptive episodes occurred between 7 and 23 April, marked by increases in explosivity and repeated small lava overflows on the southern, eastern, and southeastern flanks of the SEC [53,54,55]. At the end of the month, on 29–30 April, activity intensified significantly, producing multiple lava overflows, pulsating lava fountains, and a well-fed lava flow on the southern flank, before ceasing around 07:00 on 30 April [56].

In May, two additional minor episodes occurred on 5 May, which generated small lava flows despite poor visibility [57], and on 12 May, when Strombolian activity produced two modest flows toward the south and east [58].

After the eruptive episode on 12 May, a new event occurred on 2 June, twenty days later. The SEC produced Strombolian activity that gradually intensified into a lava fountain, while lava flows were emitted from its southern and eastern flanks. Subsequently, an eruptive fissure opened on the northeast flank of the cone, feeding a third lava flow; in the same sector, a new fracture developed, emitting dense white steam that quickly turned into a reddish pyroclastic flow. The lava fountain then diminished, ending the eruptive episode at around 3:00 a.m. on 3 June [59].

During the night of 18 June, the SEC showed weak and intermittent Strombolian activity. A lava overflow was also observed, which fed a lava flow [60].

Between late June and early August, volcanic activity was characterized by variable degassing in the summit craters. On 9 August, after low explosive activity at the SEC, a new effusive phase began. In particular, the subterminal effusive activity was from an eruptive fracture located between BN and SEC at an altitude of ~3100 m a.s.l with a direction of approximately N-S [61,62,63,64].

The lava flows shown in Figure 1 were detected using the V-STAR (Volcanic Satellite Thermal Anomalies Recognition) application, a cloud-based tool designed for automated recognition of volcanic thermal features (https://www.ct.ingv.it/technolab/v-star, accessed on 16 December 2025) [65]. V-STAR operates on the Google Earth Engine (GEE) platform, leveraging its cloud computing capabilities to process large archives of multispectral Sentinel-2 imagery in near-real time [66]. Importantly, the tool automatically accesses the full Sentinel-2 historical catalogue, from the beginning of the mission, thereby providing both a long-term record of past activity and immediate updates whenever a new image becomes available. This feature ensures continuous situational awareness during ongoing eruptive phases. V-STAR employs a supervised machine-learning approach, specifically a Random Forest classifier, trained to discriminate thermally anomalous surfaces from the surrounding background. In this study, the eruptive events of February–March, March–June, and August–September were fully mapped using V-STAR. The resulting anomaly and lava-flow products were exported as GeoTIFFs and integrated into a high-resolution DEM of Mt. Etna [67]. This allowed enhanced visualization and interpretation of the lava flows, particularly regarding their elevation, the advancement of lava fronts, and spatial expansion. By combining satellite data, machine learning, and topographic context, V-STAR provides a robust and accessible framework for monitoring eruptive behavior and for supporting disaster-risk-reduction strategies. Relevant methodological details and comparable applications of machine-learning-based satellite thermal detection are discussed in [17,65].

Figure 1 displays the lava flows produced during three distinct eruptive phases: the effusive activity of February–March (red), the explosive activity of March–June (yellow), and the effusive activity of August–September (green). These maps were derived from Sentinel-2 imagery acquired on the following dates:

• February–March (Red): 10, 12, 17, 20, and 22 February 2025.
• March–June (Yellow): 8, 15, and 26 April; 13 May; and 2 and 20 June 2025
• August–September (Green): 16, 23, 24, 29, and 31 August 2025.

3. Satellite Data Sources

3.1. Flexible Combined Imager (FCI)

FCI is the core instrument of the geostationary imaging mission aboard the MTG-I satellites, continuing the legacy of the SEVIRI instrument. Launched in December 2022 and operational since September 2024, it provides global imagery every 10 min through the Full Disk Scanning Service (FDSS) and a Rapid Scanning Service (RSS) covering the upper quarter of the disk (mainly Europe) every 2.5 min. It has 16 spectral channels ranging from visible (VIS) to infrared (IR) wavelengths [68].

FCI provides visible (VIS) and near-infrared (NIR) channels at a nominal resolution (NR) of 1000 m. However, the VIS band centered at 0.6 µm (VIS 0.6) and the shortwave infrared band at 2.2 µm (SWIR 2.2) are also available in high-resolution (HR) mode at 500 m. Infrared (IR) channels are typically delivered at a coarser resolution of 2000 m, though exceptions include the 3.8 µm and 10.5 µm channels, which are additionally offered at 1000 m in high-resolution mode.

3.2. Comparison Between FCI and SEVIRI

FCI is the natural successor to SEVIRI, which operates aboard the current MSG satellites managed by the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT). Operating from a geostationary orbit, SEVIRI is a multispectral imaging radiometer providing continuous, high-resolution observations of the Earth’s atmosphere and surface, particularly over Europe and Africa. It scans the full Earth disk every 15 min and covers one-third of the disk every five minutes as part of the Rapid Scan Service (RSS) [39]. The main technical characteristics of SEVIRI and FCI are summarized in

Table 1. Technical characteristics of MSG-SEVIRI (Meteosat Second Generation—Spinning Enhanced Visible and Infrared Imager) and MTG-FCI (Meteosat Third Generation—Flexible Combined Imager) sensors.

Features	Spinning Enhanced Visible InfraRed Imager (SEVIRI)	Flexible Combined Imager (FCI)
Satellite	Meteosat Second Generation (MSG)	Meteosat Third Generation (MTG)
Number of Channels	12 channels	16 channels (Full Disc), 2 additional (Rapid Scan)
Spectral Range	0.4–14.4 µm (VIS, NIR, IR)	0.4–13.3 µm (VIS, NIR, IR)
Spatial Resolution	3 km (IR channels), 1 km (HRV)	2 km (IR), 1 km (VIS/NIR), 0.5 km (HRV)
Temporal Resolution	Full Disc every 15 min, Rapid Scan every 5 min	Full Disc every 10 min, Rapid Scan every 2.5 min
High-Resolution Channel	1 broadband HRV	2 HR channels (0.5 km resolution)

.

SEVIRI provides a total of 12 spectral channels: three dedicated to VIS and NIR observations, and eight focused on thermal infrared measurements. In addition, SEVIRI includes a High-Resolution Visible (HRV) channel, which combines a broad spectral range (approximately 0.4–1.1 µm) with improved spatial detail. Spatially, SEVIRI delivers imagery at a resolution of 3 km at the sub-satellite point, whereas the HRV channel achieves a nominal resolution of about 1 km, making it particularly useful for observing fine-scale cloud structures and surface features during daylight hours.

In contrast, FCI represents a significant step forward in both spectral and spatial capability. It operates with 16 spectral channels, offering seven dedicated NIR channels compared to SEVIRI’s three. This broader spectral diversity allows more detailed observations of clouds, aerosols, vegetation, and land–water surfaces. Additionally, FCI improves on SEVIRI’s infrared sensing with nine IR channels, including key atmospheric windows and absorption bands useful for temperature profiling, water vapor detection, and cloud analysis. FCI also introduces a second shortwave infrared (SWIR) band at 2.2 µm, enhancing capabilities for monitoring surface properties, fires, and cloud microphysics.

Spatial resolution is another area where FCI marks a substantial upgrade. Selected VIS and SWIR channels (notably 0.6 µm and 2.2 µm) are available at 500 m resolution, other VIS/NIR channels at 1 km, and most IR channels at 2 km. However, two IR bands (3.8 µm and 10.5 µm) are also provided at 1 km, enabling higher-resolution thermal observations than were possible with SEVIRI. SEVIRI collects images at 00, 15, 30, and 45 min past the hour, while FCI operates at 00, 10, 20, 30, 40, and 50 min. Therefore, to ensure a direct comparison, only simultaneous acquisitions at 00 and 30 min were used.

Table 2. Comparison of MSG-SEVIRI and MTG-FCI spectral channels and spatial resolutions. In yellow, the Middle Infrared (MIR) and Thermal Infrared (TIR) bands used to monitor volcanic thermal anomalies.

	MSG-SEVIRI			MTG-FCI
Spectral Channels	Band No.	Wavelength [µm]	Spatial Resolution [km]	Band No.	Wavelength [µm]	Spatial Resolution [km]
VIS 0.4	-	-	-	1	0.444	1
VIS 0.5	-	-	-	2	0.510	1–0.5 (HR)
VIS 0.6	1	0.635	3	3	0.645	1
VIS 0.8	2	0.81	3	4	0.865	1
VIS 0.9	-	-	-	5	0.914	1
NIR 1.3	-	-	-	6	1.380	1
NIR 1.6	3	1.64	3	7	1.61	1
NIR 2.2		-		8	2.25	1–0.5 (HR)
IR 3.8	4	3.90	3	9	3.8	2–1 (HR)
WV 6.3	5	6.25	3	10	6.3	2
WV 7.3	6	7.35	3	11	7.350	2
IR 8.7	7	8.7	3	12	8.7	2
IR 9.7	8	9.66	3	13	9.660	2
IR 10.5	9	10.8	3	14	10.5	2–1 (HR)
IR 12.0	10	12.0	3	15	12.3	2
IR 13.3	11	13.40	3	16	13.3	2
HRV	12	-	1	-	-	-

compares the spectral channels and spatial resolutions of SEVIRI and FCI. In yellow are the Middle Infrared (MIR) and Thermal Infrared (TIR) bands used by both sensors to monitor volcanic thermal anomalies.

3.3. Comparison Between FCI and Polar Satellite Sensors

FCI offers spatial resolutions comparable to those of polar-orbiting satellites. The Moderate Resolution Imaging Spectroradiometer (MODIS) instruments aboard NASA’s Terra and Aqua satellites provide global coverage every 1–2 days, capturing data across 36 spectral bands ranging from 0.4 to 14.4 µm, with spatial resolutions of 250 m, 500 m, and 1 km. MODIS scans a 2330 km-wide swath using a 55-degree scanning angle from an orbital altitude of 705 km. Its optical system includes a rotating double-sided scan mirror paired with an off-axis telescope cooled to 83 K for infrared observations. Since February 2000, Level 1 MODIS data have been available in a geographic projection at 1 km resolution. Radiance measurements are obtained from the mid-infrared (MIR, channel 22) and thermal infrared (TIR, channel 32) bands, with channel 21 used as a substitute when the MIR band exceeds its saturation limit of 328 K. Data are encoded as 16-bit unsigned integers and must be converted for analysis. These datasets are accessible through the LAADS web platform [69].

The Visible Infrared Imaging Radiometer Suite (VIIRS), onboard the Suomi NPP and JPSS satellites, collects detailed information on Earth’s land cover, surface temperatures, and atmospheric conditions across multiple spectral bands. It provides global coverage by collecting data over different regions during each orbit. A key feature of VIIRS is its ability to observe nighttime scenes using low-light bands, which is valuable for monitoring urban lighting and nocturnal fires. VIIRS Level 1B granules, accessible via the LAADS-DAAC system [69], contain mid-infrared (MIR, channel I4) and thermal infrared (TIR, channel I5) brightness temperature data at 375 m spatial resolution. These bands have saturation temperatures of 367 K and 380 K, respectively. When saturation occurs, data switch to the M13 (MIR) and M15 (TIR) bands at a coarser resolution of 750 m. Once resampled to a UTM grid, the data form a 67 × 67 pixel matrix, reflecting the sensor’s 750 m resolution.

The Sea and Land Surface Temperature Radiometer (SLSTR), onboard the Sentinel-3A and 3B satellites, is designed to measure land and sea surface temperature with a precision better than 0.2 K. SLSTR includes nine spectral channels spanning 0.55–12.0 µm. These comprise three channels in the visible and near-infrared (VIS/NIR), three in the shortwave infrared (SWIR), one in the mid-infrared (MIR), and two in the thermal infrared (TIR) range. The instrument acquires data in both nadir and oblique viewing geometries, providing spatial resolutions of 1 km and 0.5 km for VIS/SWIR channels over swath widths of 1400 km (nadir) and 750 km (oblique), respectively. Thermal activity is detected using the S7 (MIR) and S8 (TIR) bands, with automatic switching to fire channels F1 and F2 when saturation thresholds of 312 K and 500 K are exceeded, respectively. Data are provided in netCDF 4 format, covering both nadir and oblique views, with comprehensive metadata for each pixel [70].

4. Methods

The Remote Sensing Data Fusion (RSDF) algorithm [44] is a data-fusion-based approach designed to detect volcanic hotspots and estimate the emitted Volcanic Radiative Power (VRP) for each acquired image. Information is extracted from multispectral satellite data through combined spectral and spatial analyses, exploiting the potential of a bank of Gabor filters [71]. These filters are chosen because they simultaneously capture spatial structures and their frequency content, enhancing textures and directional patterns that may indicate subtle thermal anomalies. A Gabor filter is constructed by multiplying a sinusoidal plane wave by a Gaussian envelope, making it intrinsically sensitive to specific orientations and spatial frequencies. This structure allows the filter to highlight coherent, directionally organized features while suppressing irregular noise, such as that generated by clouds, thereby increasing the sensitivity and robustness of the detection process.

The RSDF algorithm is highly versatile and can integrate information from both geostationary sensors, such as SEVIRI, and polar-orbiting sensors, such as MODIS, VIIRS, and SLSTR. Combining data from multiple sensors with an advanced detection algorithm provides a powerful framework for monitoring volcanic activity [72]. This data fusion method accurately tracks thermal changes in a volcano, which is crucial for hazard forecasting and the issuance of timely warnings.

The initial step involves spectral analysis, deriving the Normalized Thermal Index (NTI) from the difference between the MIR (3.973–4.128 μm) and TIR (10.263–11.263 μm) bands [73]. A subsequent spatial analysis is then performed using a Spatial Standard Deviation (SSD) filter applied to each image pixel. Two masks are generated to detect volcanic hotspots:

• Mask1 flags pixels in the volcanic area (VA) with SSD values exceeding the maximum SSD of the non-volcanic area (NVA):

$$\mathrm{M}\mathrm{a}\mathrm{s}\mathrm{k}1:{\mathrm{S}\mathrm{S}\mathrm{D}}_{\mathrm{V}\mathrm{A}}>\left({\mathrm{S}\mathrm{S}\mathrm{D}}_{\mathrm{N}\mathrm{V}\mathrm{A}}\right)$$

(1)

• Mask2 is applied to differentiate and identify the “true hotspots” among the potential ones:

$${\mathrm{M}\mathrm{a}\mathrm{s}\mathrm{k}2:\mathrm{N}\mathrm{T}\mathrm{I}}_{\mathrm{V}\mathrm{A}}>\left({\mathrm{N}\mathrm{T}\mathrm{I}}_{\mathrm{V}\mathrm{A}}\right)+\mathrm{n}·\left({\mathrm{N}\mathrm{T}\mathrm{I}}_{\mathrm{V}\mathrm{A}}\right)$$

(2)

A bank of Gabor filters is then applied to the MIR images to extract spatial and thermal features, which are normalized and reduced using Principal Component Analysis (PCA) to highlight significant thermal anomalies [71]. The resulting Gabor image is multiplied by the NTI to produce the Gabor Weighted NTI (G-NTI), which is further refined through spatial weighting to obtain the Spatial Gabor Weighted NTI (SG-NTI), emphasizing true volcanic hotspots while suppressing false detections. A final mask (Mask3) is generated from the SG-NTI matrix:

$$\mathrm{M}\mathrm{a}\mathrm{s}\mathrm{k}3:\mathrm{S}\mathrm{G}-\mathrm{N}\mathrm{T}\mathrm{I}>\left(\mathrm{S}\mathrm{G}-\mathrm{N}\mathrm{T}\mathrm{I}\right)+\mathrm{n}·\left(\mathrm{S}\mathrm{G}-\mathrm{N}\mathrm{T}\mathrm{I}\right)$$

(3)

The final binary mask identifying true volcanic hotspots is obtained by multiplying the three previously generated masks. Finally, the VRP of each hotspot pixel is calculated using the MIR radiance approach proposed by Wooster [19]:

$$\mathrm{V}\mathrm{R}\mathrm{P}={\displaystyle \frac{{\mathrm{A}}_{\mathrm{p}\mathrm{i}\mathrm{x}\mathrm{e}\mathrm{l}}·\mathsf{\sigma }·\mathsf{\epsilon }}{\mathsf{\alpha }·{\mathsf{\epsilon }}_{\mathrm{M}\mathrm{I}\mathrm{R}}·{\mathsf{\tau }}_{\mathrm{M}\mathrm{I}\mathrm{R}}}}\left({\mathrm{L}}_{\mathrm{M}\mathrm{I}\mathrm{R}}-{\mathrm{L}}_{\mathrm{M}\mathrm{I}\mathrm{R},\mathrm{b}\mathrm{g}}\right)$$

(4)

where L_MIR and L_MIR,bg are, respectively, the spectral radiance (W·m⁻²·sr⁻¹·μm⁻¹) of the hot pixels and the non-fire background pixels recorded in the MIR band, A_pixel is the pixel area (m²), σ is the Stefan-Boltzmann constant (5.67∙10⁻⁸ W∙m⁻²∙K⁻⁴), τ_MIR is the atmospheric transmittance in the MIR band, ε is the emissivity, ε_MIR is the surface spectral emissivity in the MIR band, and α (W∙m⁻²∙sr⁻¹∙μm⁻¹∙K⁻⁴) is a constant. Following the approach of Wooster et al. 2003 [19], the MIR radiance emitted by sub-pixel hot sources is approximated by fitting a power-law relationship of the form L(λ) = aT^b to the Planck function over a representative fire-temperature range (600–1200 K), taking into account the sensor’s MIR spectral response. The empirical constant α used in Equation (4) is obtained directly from this calibration by regressing the MIR radiance increments against the corresponding temperature-dependent radiance values produced by the power-law approximation.

The total VRP, in Watt (W), of the analyzed image is then obtained by summing the VRP values of all detected hotspots.

4.1. Adaptation of RSDF to FCI

The RSDF algorithm was developed to process multispectral data from both polar and geostationary satellite sensors, making it readily adaptable to the new FCI dataset. The only parameters in Equation (4) that must be modified are A_pixel, corresponding to the spatial resolution of the data (in this case, 1000 m²), and α, which must be empirically determined. We followed the approach of Wooster [19], according to which, at very high temperatures (600–1500 K), the energy emitted by very hot objects such as volcanic hotspots at a specific infrared wavelength (4 µm) is directly proportional to their temperature raised to the fourth power (T⁴). The value of α found for FCI is 3.22 · 10⁻⁹ W∙m⁻²∙sr⁻¹∙μm⁻¹∙K⁻⁴.

We can define k as the constant that multiplies the excess MIR radiance $\left({\mathrm{L}}_{\mathrm{M}\mathrm{I}\mathrm{R}}-{\mathrm{L}}_{\mathrm{M}\mathrm{I}\mathrm{R},\mathrm{b}\mathrm{g}}\right)$, composed of fixed values (σ, τ_MIR, ε, and ε_MIR) and parameters dependent on the sensor (A_pixel, α). Therefore, the constant k is defined as

$$\mathrm{k}={\displaystyle \frac{{\mathrm{A}}_{\mathrm{p}\mathrm{i}\mathrm{x}\mathrm{e}\mathrm{l}}·\mathsf{\sigma }·\mathsf{\epsilon }}{\mathsf{\alpha }·{\mathsf{\epsilon }}_{\mathrm{M}\mathrm{I}\mathrm{R}}·{\mathsf{\tau }}_{\mathrm{M}\mathrm{I}\mathrm{R}}}}$$

(5)

is specific to each sensor and, for FCI, is equal to 1.76 · 10⁷ m²·sr·μm. The different values of α and k for each satellite sensor processed by the RSDF algorithm are available in

Table 3. Comparison of the temporal, spatial, and radiometric characteristics of selected satellite sensors, including calibration coefficients α [W·m⁻²·sr⁻¹·μm⁻¹·K⁻⁴] and k [m²·sr·μm].

Satellite Sensor	Temporal Resolution	Apixel [m2]	α	k	Saturation Temp	NetD
VIIRS-I	Twice per day	375 × 375	3.21 · 10−9 [74]	2.48 · 107	MIR (3.753 μm): 367 K TIR (11.469 μm): 380 K	MIR: <2.5 K at 270 K TIR: <1.5 K at 210 K
VIIRS-M	Twice per day	750 × 750	2.87 · 10−9 [74]	1.11 · 107	MIR (4.067 μm): 634 K TIR (10.729 μm): 363 K	MIR: 0.107 K at 300 K TIR: 0.070 K at 300 K
MODIS	Twice per day	1000 × 1000	3.0 · 10−9 [19]	1.89 · 107	MIR (3.959 μm): ~500 K TIR (11.030 μm): ~340 K	MIR: 0.07 K at 300 K TIR: 0.05 K at 300 K
SLSTR	Twice per day	1000 × 1000	3.30 · 10−9 [75]	1.70 · 107	MIR (3.74 μm): 311 K MIR (3.74 μm): 500 K TIR (10.8 μm): 321 K TIR (10.8 μm): 400 K	MIR (S7): <0.08 K at 270 K MIR (F1): Not specified TIR (S8): <0.05 K at 270 K TIR (F2): Not specified
FCI	Every 10 min	1000 × 1000	3.22 · 10−9	1.76 · 107	MIR (3.8 μm): 450 K TIR (10.5 μm): ~340 K	MIR: 0.2 K @ 300 K TIR: 0.1 K @ 300 K
SEVIRI	Every 15 min	3000 × 3000	1.38 · 10−8	3.70 · 107	MIR (3.9 μm): ~335 K TIR (10.8 μm): ~335 K	MIR: 0.35 K at 300 K TIR: 0.25 K at 300 K

.

4.2. Time Average Discharge Rate (TADR) and Volume Calculation

The time series of VRP is necessary for characterizing two eruptive parameters: the Time-Averaged Discharge Rate (TADR) and the volume of lava flows. The TADR represents the volume of lava emplaced during a given time interval divided by its duration. Like VRP, it depends on the active lava area within the satellite’s field of view. Therefore, a new empirical parameter, the radiant density (c_rad), is used to link the two quantities [76,77]:

$${\mathrm{c}}_{\mathrm{r}\mathrm{a}\mathrm{d}}={\displaystyle \frac{\mathrm{V}\mathrm{R}\mathrm{P}}{\mathrm{T}\mathrm{A}\mathrm{D}\mathrm{R}}}$$

(6)

The radiant density [J·m⁻³] summarizes the prevailing insulation, rheological and topographic conditions during lava emplacement, maintaining the proportionality between lava discharge rate and active area. The c_rad value commonly adopted for Etna can be found in the literature [78].

To account for the uncertainty in the derivation of the TADR parameter, we calculated its maximum, minimum, and mean values using three different c_rad values [J·m⁻³]. Assuming a ±50% uncertainty, minimum and maximum TADR values are computed to ensure accuracy [79]. The erupted lava volume Volume_i was retrieved by calculating the integral of the TADR of two sequential satellite acquisitions (at time i and i − 1) and the entire erupted lava volume is derived by cumulatively summing up Volume_i [80]:

$${\mathrm{T}\mathrm{A}\mathrm{D}\mathrm{R}}_{\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{n}}={\displaystyle \frac{{\mathrm{T}\mathrm{A}\mathrm{D}\mathrm{R}}_{\mathrm{i}}+{\mathrm{T}\mathrm{A}\mathrm{D}\mathrm{R}}_{\mathrm{i}-1}}{2}}$$

(7)

$${\mathrm{V}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{m}\mathrm{e}}_{\mathrm{i}}={\mathrm{T}\mathrm{A}\mathrm{D}\mathrm{R}}_{\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{n}}·\mathsf{\Delta }\mathrm{t}$$

(8)

All derived parameters are available through the Volc@Hazard web platform, freely accessible via the TechnoLab webpage of the INGV-Etna Volcano Observatory [81].

5. Results

To test FCI’s capability for monitoring volcanic activity, we analyzed data from Mount Etna between January and September 2025. February 2025 marked the onset of a varied eruptive period, characterized by a complex interplay between explosive and effusive activity across the summit craters, which lasted until early September (Figure 1) [46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64]. During this period of heightened activity, we identified three distinct phases: effusive activity (February–March), explosive activity (March–June), and a second effusive period (August–September).

We describe these volcanic phases using the VRP time series, which was also used to derive two key eruptive parameters: the TADR and the lava flow volume.

5.1. Effusive Activity (February–March 2025)

The first signs of unrest appeared on 6 February 2025, with modest Strombolian explosions at the South-East Crater (SEC). Two days later, on 8 February, a new effusive phase began from an eruptive fissure at the base of the Bocca Nuova crater, at approximately 3050 m above sea level (a.s.l.). This effusive activity generated multiple lava branches and concluded on 2 March, after 23 days of activity (Figure 1) [46,47,48,49].

The eruption was monitored in near real time using both polar-orbiting and geostationary satellite data. The SEVIRI sensor provided high-frequency (15 min interval) thermal measurements throughout the entire eruption. Figure 2a shows the temporal trend of VRP derived from SEVIRI (black crosses) and FCI (blue circles) from 8 February to 2 March. Both sensors captured the same overall trend, revealing several peaks of increased thermal activity around 11, 16, and 23 February. However, VRP values recorded by SEVIRI are consistently higher than those from FCI throughout the period.

Figure 2b presents the same data on a logarithmic scale, highlighting lower-power fluctuations while confirming that both instruments follow the same temporal pattern. A distinct and persistent offset is observed between the two datasets, with SEVIRI (black line) consistently above FCI (blue line).

When compared with polar satellite sensors (MODIS, VIIRS, SLSTR), the VRP retrieved from FCI shows excellent agreement (Figure 3). The dense cluster of blue circles from FCI provides a near-continuous depiction of the eruption’s intensity, while the sparser observations from the polar sensors—MODIS (red squares), VIIRS (purple diamonds), and SLSTR (green triangles)—consistently align with the FCI trend. This strong correlation demonstrates the high compatibility of FCI with other high-spatial-resolution sensors, confirming its reliability for operational volcanic monitoring.

The TADR and cumulative erupted volume were calculated using two datasets: one relying exclusively on FCI data and another combining measurements from MODIS, SLSTR, and VIIRS. The aggregated dataset compensates for the lower temporal resolution of the polar sensors relative to FCI’s 10 min acquisition cycle. Figure 4 shows the TADR and cumulative lava volume from 8 February 11:10 UTC to 1 March 21:00 UTC, 2025. The results indicate remarkable consistency between the two approaches. The total erupted lava volume derived from FCI data is approximately 3.74 Mm³, nearly identical to the 3.61 Mm³ obtained from the aggregated polar dataset. This strong agreement validates the combined use of geostationary and polar observations for quantifying erupted volumes.

5.2. Explosive Activity (March–June 2025)

Between March and June 2025, Mount Etna exhibited persistent explosive activity primarily concentrated at the summit craters and characterized by events of short duration, lasting from several hours to a few days. This included episodes of varying intensity, featuring Strombolian activity, lava fountains, and ash emissions that often produced short-lived lava flows (Figure 1) [50,51,52,53,54,55,56,57,58,59,60].

Figure 5a,b show thermal activity monitored using geostationary sensors (FCI and SEVIRI) and polar sensors (MODIS, SLSTR, VIIRS), respectively. The period from March to June was marked by a gradual intensification of explosive activity. In early March, INGV reports described mostly degassing from the summit craters. From mid-March, Strombolian activity resumed at the Southeast Crater (SEC), with moderate explosions ejecting coarse pyroclastic material within the crater area. This activity peaked on 20 March 2025, with a VRP maximum of 519.6 MW as measured by FCI. April and May featured several episodes of intense Strombolian activity, sometimes accompanied by minor ash emissions. In June, two short-lived, high-energy events occurred on 2 June and 13 June, corresponding to the highest VRP values recorded during this eruptive phase. Both episodes produced intense lava fountains, highlighting the strong thermal output associated with these events. Figure 6 focuses on these two events, showing zoomed panels for 2 June (Figure 6a,b) and 19 June (Figure 6c,d).

The TADR and cumulative erupted volume of the short-lived lava flow produced during the event of 2 June 2025 were calculated exclusively using FCI data and combining measurements from MODIS, SLSTR, and VIIRS. Figure 7 shows the TADR and cumulative lava volume from 2 June 02:00 UTC to 3 June 02:45 UTC 2025. The results indicate remarkable consistency between the two approaches. The total erupted lava volume derived from FCI data is approximately 0.30 Mm³, identical to that obtained from the aggregated polar dataset. This strong agreement validates the combined use of geostationary and polar observations for quantifying erupted volumes.

5.3. Effusive Activity (August–September 2025)

A new effusive phase occurred from mid-August to early September 2025 [61,62,63,64]. The eruption began on 14 August, characterized by Strombolian activity and lava effusion from a fissure between the Bocca Nuova and Southeast Craters (Figure 1). The intensity peaked between 25 and 26 August, then gradually declined in early September. By that time, the main effusive phase had ended. Subsequent activity was limited to variable degassing, weak ash emissions from Bocca Nuova, and low-level thermal anomalies.

Satellite data confirmed this timeline (Figure 8). VRP values derived from both geostationary (FCI, SEVIRI) and polar (MODIS, VIIRS, SLSTR) sensors show consistent trends, marking the onset, peak, and waning of the eruption.

The TADR and cumulative lava volume from 13 August 23:50 UTC to 1 September 20:20 UTC were calculated from two datasets, as illustrated in Figure 9. The first, based solely on FCI’s high-frequency data, yielded a total lava volume of approximately 2.69 Mm³, closely matching the 3.03 Mm³ derived from combined polar data (MODIS, SLSTR, VIIRS). Although the agreement between the two approaches remains high, the observed discrepancy in the total lava volume is attributable to a specific high-intensity VRP detection (1373.79 MW) by MODIS on 30 August 2025 at 19:45 UTC. This isolated peak resulted in a significantly elevated TADR, which consequently increased the calculated total erupted volume.

6. Discussions

Both FCI (aboard MTG) and SEVIRI (aboard MSG) are geostationary instruments that provide high temporal resolution for volcanic monitoring. While their main difference lies in spatial resolution (SEVIRI operating at 3 km and FCI at 1 km for the relevant thermal band), both sensors are capable of continuous, long-term observation of volcanic activity. As shown in Figure 2 and Figure 5a, and 8a, the temporal trends of Mount Etna’s VRP derived from FCI and SEVIRI are highly correlated, confirming their ability to consistently track eruptive events.

To ensure a fair comparison, only simultaneous acquisitions were used. SEVIRI collects images at 00, 15, 30, and 45 min past the hour, while FCI operates at 00, 10, 20, 30, 40, and 50 min. Consequently, we limited our comparative analysis to data acquired at 00 and 30 min to minimize temporal mismatches and ensure a direct one-to-one comparison. Moreover, no filtering was performed on cloudy images.

VRP values were analyzed separately for the two effusive phases (February–March and August–September 2025) and the explosive phase (March–June 2025), since these two eruptive styles exhibit distinct satellite responses. Figure 10a,b present scatter plots comparing the VRP estimated from SEVIRI and FCI imagery during effusive (Figure 10a) and explosive (Figure 10b) activity. In the plots, the x-axis corresponds to VRP derived from SEVIRI data (VRP_SEVIRI) and the y-axis to VRP derived from the FCI data (VRP_FCI).

During effusive activity (Figure 10a), VRP values from SEVIRI and FCI show a strong positive linear correlation. As VRP detected by SEVIRI increases, that detected by FCI increases proportionally. The data points are tightly distributed around the regression line y = 0.30x + 94.73, indicating that, on average, during effusive eruptions accounting for thermal anomalies broadly distributed in the spatial domain, FCI measures about 30% of SEVIRI’s VRP values, with a small constant offset. SEVIRI’s coarser resolution can cause adjacent pixels to hotspot regions to be flagged as “thermal anomalies” even if they only contain diffracted light, leading to an overestimation of the total VRP. In contrast, FCI’s finer spatial resolution better confines the thermal signal to the actual active area, resulting in a more constrained and accurate VRP calculation. The R-squared or coefficient of determination (R²), which measures the proportion of variance in the dependent variable explained by the independent variable, was calculated for the effusive activity phase (Figure 10a) and found to be 0.4. This indicates that only 40% of the variance in the SEVIRI data can be explained by the FCI data using a linear model. Consequently, while the correlation is strong at low VRP values, significant scattering is observed during high-intensity activity.

In contrast, Figure 10b, corresponding to explosive activity, displays a large scatter with no clear linear trend. Low-power values cluster densely, while higher values are more irregular. This behavior suggests that the two sensors respond differently to the transient and heterogeneous thermal conditions typical of explosive eruptions. The presence of ash, rapid temperature fluctuations, and differences in sensor resolution all contribute to these discrepancies. This different response aligns with the scattering behavior observed during the high-intensity effusive episodes shown in Figure 10a.

These systematic differences arise primarily from the different characteristics of the two instruments. Different sensors detect different numbers of pixels depending on the following:

• The thermal and spatial features of the monitored event (e.g., portions of lava at different temperatures, from cooling to incandescent lava flow portions);
• The technical characteristics of the sensors, including saturation temperatures for each channel, spatial resolution, available spectral bands, and noise equivalent temperature difference (NetD) [82].

As shown by Corradino et al. [82], even the same effusive event can appear differently depending on the instrument’s technical specifications. These discrepancies become even more pronounced when observing complex lava fields, where subpixel thermal heterogeneity and varying pixel coverage amplify the differences between sensors.

Another factor influencing the differences between VRP measurements from SEVIRI and FCI lies in the algorithm used to detect high-temperature volcanic features. Both datasets were processed with the same RSDF algorithm [44], which analyzes the intensity and spatial structure of the image using a bank of Gabor filters. These filters emphasize local spatial frequency patterns around each pixel, making them inherently sensitive to image resolution. Consequently, the sensitivity and precision of detection of volcanic phenomena, i.e., discriminating spatial features from the background, may be higher for the higher spatial resolution of FCI (1 km), allowing a more effective characterization of fine-scale volcanic features than SEVIRI (3 km).

Effusive and explosive activities exhibit markedly different behaviors. To better illustrate these differences, we compared simultaneous FCI and SEVIRI imagery for selected events. Two effusive cases were analyzed: one under cloudy conditions (19 February 2025) and one under clear-sky conditions (30 August 2025), shown in Figure 11a,b.

In Figure 11, the first two columns display the TIR (10.5 µm for FCI and 10.8 µm for SEVIRI) and MIR (3.8 µm for FCI and 3.9 µm for SEVIRI) bands, where the MIR image highlights the thermal anomaly. The third column shows the NTI, which enhances hotspot contrast. Visually, FCI’s higher spatial resolution provides sharper and more defined hotspots than SEVIRI. Under cloudy conditions (19 February 2025, 08:12 UTC), FCI measured a VRP of 1100 MW, while SEVIRI recorded about 900 MW, showing good agreement considering cloud obstructions. Under clear-sky conditions (30 August 2025, 21:42 UTC), SEVIRI measured 1813.46 MW and FCI 762.95 MW.

These discrepancies, often encountered for the effusive activity, are due to two main factors, (a) the broader pixel footprint of SEVIRI, which can integrate a larger portion of the heated area in clear-sky conditions, may produce higher VRP values; (b) the higher sensitivity of FCI can lead to partial pixel saturation during intense thermal events, causing underestimation of total VRP.

As regards explosive activity, we show two acquisitions (12 May 2025, 21:12 UTC; and 2 June 2025, 08:30 UTC), in Figure 12a,b. As before, the columns display TIR, MIR, NTI, and the resulting VRP masks. In the first case, FCI detected a VRP of 3977.01 MW, nearly double SEVIRI’s 2159.21 MW. This confirms the trend observed in Figure 5a, where FCI consistently measures higher VRP during explosive events. The 2 June 2025 event exemplifies this: FCI captured a distinct peak (4489.82 MW), while SEVIRI failed to detect any anomaly, likely due to ash obscuration combined with its lower spatial resolution. Following the outcomes from [82], the small dimension of the anomalous area and the very high temperature characterizing it make FCI respond more accurately than SEVIRI. Specifically, in the latter case, obscuration leads SEVIRI response to fall outside its working region.

Although FCI is geostationary, its 1 km spatial resolution matches that of polar sensors such as MODIS and SLSTR. Even though VIIRS is a polar-orbiting satellite sensor, it offers higher spatial resolution, with I-bands at 375 m and M-bands at 750 m. For a direct comparison, FCI data were cross-analyzed with simultaneous MODIS and SLSTR acquisitions.

Figure 13a,b show scatter plots comparing VRP_FCI with VRP_POLARS (VRP measurements from polar-orbiting sensors: MODIS in red and SLSTR in green) for effusive and explosive phases, respectively. In Figure 13a, the regression line y = 0.69x + 197.15 indicates a strong correlation between polar sensors and FCI. The positive y-intercept of +197.15 MW suggests that FCI is slightly more sensitive to lower VRP values (low-level thermal anomalies). In Figure 13b, the regression line y = 1.15x + 2.54 demonstrates near-perfect linearity, confirming the excellent consistency between FCI and polar sensors. The calculated R² values are 0.48 and 0.75 for the effusive (Figure 13a) and explosive (Figure 13b) phases, respectively.

Overall, differences between FCI and SEVIRI are much greater than between FCI and the polar sensors. This is primarily because FCI shares comparable spatial resolution (1 km) and similar saturation temperature and noise characteristics with MODIS and SLSTR, reducing systematic discrepancies.

This comparative analysis between FCI, SEVIRI, and polar sensors (MODIS, SLSTR, and VIIRS) confirms the strong potential of FCI for continuous volcanic monitoring. SEVIRI has played a central role in the past two decades, providing valuable information on the temporal evolution of volcanic radiative power thanks to its high acquisition frequency (one image every 15 min). Nevertheless, its coarse spatial resolution of about 3 km often limited the accuracy in locating active pixels, especially during small or spatially confined events. Polar sensors such as MODIS, SLSTR, and VIIRS offer higher spatial resolution and improved sensitivity to fine-scale thermal anomalies, but their limited temporal coverage (typically two acquisitions per day) prevents detailed tracking of rapidly evolving eruptions. Compared with these instruments, FCI reduces both spatial and temporal constraints, providing a more balanced dataset for quantitative volcanic monitoring. The resulting measurements demonstrate improved consistency across platforms and confirm that FCI can serve as a reliable geostationary reference for integrating multi-sensor observations.

Based on this study, we conclude that FCI represents a significant step forward in geostationary monitoring. Its improved spatial resolution, comparable to that of polar-orbiting sensors, allows for the detailed detection of thermal anomalies. FCI’s comparable spatial resolution combined with its high temporal frequency complements polar data, offering a synergistic approach to monitor volcanic phenomena. FCI’s superior spatial, spectral, and temporal capabilities make it the natural successor of SEVIRI, and it is expected to eventually replace SEVIRI as the primary geostationary instrument for volcanic monitoring.

7. Conclusions

The RSDF approach was adapted to the new FCI sensor data and successfully applied throughout 2025 to monitor volcanic activity at Mount Etna. The VRP values calculated from FCI were validated against those derived from SEVIRI, owing to their similar temporal resolutions (10 min for FCI vs. 15 min for SEVIRI), and from polar-orbiting sensors (MODIS, VIIRS, and SLSTR), which offer comparable spatial resolutions.

The use of the new geostationary FCI sensor represents a major technological advancement in volcanic activity monitoring. The combination of medium-to-high spatial resolution (1–2 km) and high temporal frequency (one acquisition every 10 min) provides an unprecedented capability for continuous observation of volcanic processes. This dual advantage makes FCI particularly effective in detecting and tracking thermal anomalies, quantifying VRP, and following rapid changes in eruptive behavior with both spatial accuracy and temporal consistency. Unlike its predecessor SEVIRI, limited by its coarser spatial resolution (3 km at nadir), or polar-orbiting sensors such as MODIS, SLSTR, and VIIRS, which provide only a few observations per day, FCI enables near-real-time surveillance and serves as a valuable bridge between these two observation strategies.

The improved spatial and temporal resolutions make FCI a powerful tool for supporting early warning systems, real-time hazard assessment, and long-term studies of volcanic activity trends. Future developments should focus on expanding FCI’s applications to include volcanic cloud detection and tracking. This will require the development of specific retrieval algorithms to identify and quantify ash and sulfur dioxide (SO₂) emissions, estimate cloud-top temperature and height, and assess plume dispersion in the atmosphere. Accurate retrieval of volcanic gas emissions is essential, as stratospheric SO₂ injections can induce global cooling through sulfate aerosol formation, while CO₂ emissions contribute to greenhouse warming. Understanding the impact of major eruptions on climate processes is therefore crucial, and satellite observations provide a unique means to obtain rapid and global information on these phenomena.

This study provides the first quantitative assessment of MTG-FCI performance for volcanic thermal monitoring. FCI demonstrates significant improvements in both temporal and spatial resolution compared with SEVIRI, enabling more detailed tracking of effusive and explosive activity. Although systematic differences in retrieved VRP remain, they can be corrected through sensor-specific calibration. In addition, the quantitative agreement achieved between FCI and polar sensors demonstrates the potential for integrated multi-sensor analyses, allowing consistent and cross-validated volcanic monitoring across orbital platforms. The forthcoming full operational deployment of FCI marks a key step toward an advanced, near-real-time volcanic surveillance system for Europe and beyond. We have integrated FCI data into the existing Volc@Hazard web platform, which was developed to monitor the activity of Etna and Stromboli in near-real time using satellite data. This integration will enhance real-time surveillance capabilities, allowing for more timely and precise alerts during the early stages of eruptive activity. Volc@Hazard is freely accessible via the TechnoLab webpage of the INGV-Etna Volcano Observatory [81].