In September 2024, the FRM4Fire team kicked off the project by collecting airborne data in northern Alberta and Saskatchewan, led by the NCEO (National Centre for Earth Observation) Airborne Earth Observatory (NAEO) at King’s College London.

FRM4Fire background and previous Sentinel-3 FRP validations

The FRM4Fire project is working to advance validation methods for fire radiative power (FRP) measurements, including by establishing reliable validation standards for active fire and FRP data obtained from satellite Earth observations.

The Earth Observation and Wildfire research group at King's have had a pivotal role in developing the Sentinel-3 FRP algorithm. In addition, King's and NAEO have conducted multiple airborne campaigns as part of the European Space Agency (ESA) Evaluation and Validation of Sentinel-3 Active Fire Detection and FRP Products: Fire Detection Experiment (FIDEX) project in 2018, 2023, and 2024. During the FIDEX campaigns, wildfires were surveyed with research aircraft equipped with a thermal camera mirroring the spectral response of Sentinel-3’s Sea and Land Surface Temperature Radiometer (SLSTR). These airborne surveys were aligned with satellite overpasses to assess their accuracy.

Figure 1 shows a still image of the thermal imagery collected over a wildfire. Click on the image to see a movie.

Figure 1. InfraTec ImageIR 8300 thermal images collected from an airborne survey of a wildfire in northwestern Ontario (52.3064°N, 94.8182°W) during the FIDEX-II campaign on 14 August 2023. This fire (RED070) was also measured by Sentinel-3A simultaneously with the images shown.

The FRM4Fire project is focused on uncertainties of satellite observations of FRP, but also involves identifying uncertainties associated with the airborne measurements used to validate the satellite observations, such as those taken during the FIDEX campaigns.

The CarbonARA Campaign in Fort McMurray

As part of ESA’s Carbon Amazon Rainforest Airborne (CarbonARA) project, an airborne campaign was conducted in Fort McMurray, Alberta, to measure greenhouse gas emissions from oil sands refineries. Recognising an opportunity to contribute to FRM4Fire, the team integrated wildfire data collection into this mission. The campaign utilised a de Havilland Twin Otter research aircraft owned by the British Antarctic Survey (BAS), outfitted with an InfraTec ImageIR 8300 mid-infrared (MIR) thermal camera. This specialised camera was equipped with a spectral filter replicating SLSTR’s S7 and F1 bands, which ensures compatibility between airborne and satellite FRP observations.

Figure 2. British Antarctic Survey de Havilland Twin Otter at Fort McMurray airport (56.6531°N, 111.2224°W) during the 2024 CarbonARA campaign. Credit: Farrer Owsley-Brown, King’s College London

Field Operations and Science Flights

The King's NAEO team travelled out to Calgary in early September 2024, where they installed the instruments on the aircraft for the campaign (see Fig. 2). From there, the team moved to Fort McMurray, strategically chosen for its proximity to active wildfires and the oil sands industry, where they were joined by further staff from King's and the FRM4Fire lead from the National Physical Laboratory (NPL).

Figure 3. Earth observation sensors installed in the camera viewing bay board of the British Antarctic Survey de Havilland Twin Otter during the 2024 CarbonARA campaign. The primary MIR thermal camera, InfraTec ImageIR 8300, can be seen in the bottom left of the viewing bay (red side). Credit: Callum Middleton, King’s College London.

Wetikoo Fire

On 11 September 2024, the team conducted a 2.5-hour survey of the Wetikoo wildfire (see Fig. 4; 56.8040°N, 107.3364°W), a 2,800-hectare wildfire in northern Saskatchewan. Coinciding with both Sentinel-3A and Sentinel-3B overpasses over the fire at 17:29 and 18:30 UTC, the airborne measurements provided direct comparisons with the satellite data.

Figure 4. Screenshot of FIRMS showing satellite hotspot detections in northern Saskatchewan on 11 September 2024. The location of the Wetikoo fire (centred at 56.8040°N, 107.3364°W) is highlighted.

During the survey, a massive pyrocumulonimbus cloud was observed (see Fig. 5) forming from a nearby wildfire 200 kilometers south of Wetikoo. Figure 6 shows RBG and shortwave infrared (SWIR) false colour composite images of the wildfire activity, with the Wetikoo fire identified by the yellow circle and the larger fire with the cloud by the purple circle. As the team was leaving the Wetikoo fire, a pyrocumulonimbus cloud began to form from its smoke plume as well (Fig. 7).

Figure 5. Images of a pyrocumulonimbus cloud over a large wildfire 200km to the south (located at 56.0755°N, 107.3364°W) of the Wetikoo fire being surveyed (56.8040°N, 107.3364°W) during a daytime flight on 11 September 2024. Credit: Farrer Owsley-Brown, King’s College London.

Figure 6. Sentinel-2 imagery of northern Saskatchewan on 11 September 2024: RGB (left), shortwave-infrared false colour composite (right). Both images show the same extent. The circles indicate the wildfire surveyed (56.8040°N, 107.3364°W; top/yellow) and a fire 200km to the south (56.0755°N, 107.3364°W; bottom/purple).

Figure 7. Photographs during the Wetikoo wildfire survey on 11 September 2024 showing a pyrocumulonimbus cloud beginning to form.

Flights over gas flares and another wildfire

While returning from the survey of Wetikoo, the team flew over the oil sands refineries (see Fig. 8) only 40km to the northwest of Fort McMurray, and determined suitable targets for a later survey, with a focus on observing the gas flares there and determining how stable the thermal signal they produced were.

Figure 8. Fort McMurray oil sands refineries during daytime (above; 11 September 2024) and nighttime (below; 15 September 2024) flights. Credit: Farrer Owsley-Brown, King’s College London.

A key aim of the FRM4Fire project is to characterise the uncertainties of the airborne measurements, distinguishing those arising from the environment and those from the sensor itself. Environmental effects include factors such as the terrain aspect or the amount of vegetation obscuring the view of the sensor at different angles. For instance, at nadir, more of the forest floor may be visible between the three trunks of a forest than when viewing at an angle. Factors associated with the sensor itself include the distortion from the lens.

A flight plan (see Fig. 9) was created so that a hotspot from a gas flare could be seen across the entire field of view of the thermal camera (see Fig. 10). As the gas flare measured was determined to be relatively stable, particularly when compared to the thermal emission from a wildfire, this data will enable the team to discover what effects stem from the sensor itself on the FRP retrieval. This flight over the gas flare was successfully conducted on the night of 15 September 2014, and the lucky crew were given an incredible show of the Northern Lights towards the end of the science flight (see Fig. 11).

Figure 9. Flight plan over a gas flare 40km to the northwest (57.0063°N, 111.4789°W) of Fort McMurray airport. Credit: Callum Middleton, King’s College London.

Figure 10. Thermal images from the night flight over an oil sands refinery 40km northwest (57.0063°N, 111.4789°W) of Fort McMurray airport on 15 September 2024. A thermally stable gas flare (circled) was measured across the entire field of view of the thermal camera.

Figure 11. Flight crew in front of the BAS Twin Otter, with the aurora borealis visible in the sky behind, after flying over the gas flares in the oil sands refineries on the night of 15 September 2024. Credit: Toby Wainwright, King’s College London.

This strategy (Fig. 9) was also repeated over a small wildfire (see Fig. 12) near Fort McMurray so that any environmental effects can be quantified, such as obscuration from vegetation. This flight was conducted on 11 September, the second science flight of the day, after the crew had landed from the Wetikoo survey and gas flare reconnaissance.

Figure 12. Wildfire 90km to the northeast of Fort McMurray airport (57.2560°N, 110.1639°W), surveyed on 11 September 2024. Credit: Farrer Owsley-Brown, King’s College London.

The FRM4Fire team have started processing the airborne data recorded from the CarbonARA campaign. Further analysis will provide information about uncertainties associated with airborne validations of satellite FRP retrievals. Watch this space!