NRT Product Availabe since Mar 2008
|This operational product is fully documented in , in the Algorithm Theoretical Basis Document (ATBD), Product User Manual document (PUM) and the Product Output Format document (POF). The validation results for this product are available in the (VR) document and in . Users are encouraged to read references  and  linked to in the "References" list below that, respectively, provide a peer-reviewed description of the algorithm, product contents, performance evaluation/accuracy and demonstration of use of the FRPPIXEL product. It is requested that users cite these papers to describe the FRPPIXEL product if it is used within their studies.|
|The use of LSA SAF products in publications is kindly requested to be duly acknowledged:
FRPPIXEL was provided by the EUMETSAT Satellite Application Facility on Land Surface Analysis (LSA SAF; Trigo et al., 2011)
Trigo, I. F., C. C. DaCamara, P. Viterbo, J.-L. Roujean, F. Olesen, C. Barroso, F. Camacho-de Coca, D. Carrer, S. C. Freitas, J. García-Haro, B. Geiger, F. Gellens-Meulenberghs, N. Ghilain, J. Meliá, L. Pessanha, N. Siljamo, and A. Arboleda, 2011: The Satellite Application Facility on Land Surface Analysis. Int. J. Remote Sens., 32, 2725-2744, doi: 10.1080/01431161003743199
|The FRPPIXEL product records information on the location, timing and fire radiative power (FRP, in MWatts) output of landscape fires ("wildfires") detected every 15 minutes across the full Meteosat disk at the native spatial resolution of the SEVIRI sensor. It has been demonstrated in small-scale experimental fires that the amount of radiant heat liberated in a fire per unit time (i.e. the Fire Radiative Power) is well related to the rate at which fuel is being consumed . This is a direct result of the combustion process, whereby carbon-based fuel is oxidised to CO2 (and other gaseous and particulate products) with the accompanying release of a certain "heat yield". Measuring this FRP and integrating it over the lifetime of a fire provides an estimate of the total Fire Radiative Energy (FRE) released, which for landscape fires should be proportional to the total amount of biomass burned. The FRP approach therefore provides an alternative approach to calculating wildfire fuel consumption through methods that rely on the mapping of burned area and assuming, measuring or modelling the supposed fuel consumption per unit area. It is the basis of the global fire emissions service (GFAS) used within the Copernicus Atmosphere Monitoring Service.|
|Current methods to obtain fuel combustion estimates are based largely on burned area mapping approaches, with necessary assumptions regarding the fuel density and combustion completeness variables (which may vary with land-cover/climate/timing-of-fire) that control fuel consumption per unit area. The FRE approach in theory circumvents the requirement for these assumptions, providing a geophysical variable potentially more directly related to the amount of combusted biomass and thus also to the amount of a particular gaseous and particulate species released into the atmosphere. Geostationary observations allow very high temporal frequency FRP measurements, and thus a much-improved ability to estimate FRE via temporal integration of the FRP measures when compared to the far less-frequent FRP observations made from systems in low-Earth orbit (e.g. from the MODIS, VIIRS, and SLSTR sensors). Calculations indicate that the fire detection algorithm used to generate the FRPPIXEL product is extremely sensitive to fire, and can detect actively burning landscape fires whose flaming/smouldering areas cover down to 10-3 to 10-4 of a pixel. However, a limitation of geostationary observations results from the fact that the subtended pixel areas on the Earth are rather larger than those typical of low-Earth orbit systems, and even with the very strong sensitivity to fire displayed by the FRPPIXEL product, a greater fraction of the smaller and/or less intensely burning (i.e. lower FRP) fires typically remain undetected than when using Low Earth Orbit (LEO) sensors. However, this sensitivity advantage offered by LEO observations is counteracted by the far greater data availability provided from geostationary sensors (96 observations per day with Meteosat SEVIRI), which makes the chance of obtaining a cloud free view of a fire far more likely [2, 6], provides information across the widely varying fire diurnal cycle [4, 8], and which greatly simplifies the estimation of FRE [5, 8]. The development of the FRPPIXEL product is led by King's College London and uses the Wooster et al. (2003) "MIR radiance method" of FRP derivation, first proposed by  and adapted for use with geostationary systems by [2,3]. The same method is now used within the Collection 6 EOS-MODIS Fire Products (MOD14 and MYD14) generated by NASA.|
|Pixel containing sub-pixel active fire pixels burning at the time each SEVIRI image was taken are detected using the so-called geostationary Fire Thermal Anomaly (FTA) algorithm fully described in . At the heart of this process lie tests that attempt to detect the type of middle infrared (MWIR) and longwave infrared (LWIR) pixel signals indicative of a sub-pixel fire event. This basic criteria is used alongside a series of other tests that attempt to discriminate "true" fire detections from other phenomena that may induce similar MWIR and LWIR channel signals. A cloud mask and a land-water mask is used to identify and remove cloud and water body pixels from the analysis (that otherwise may induce false fire detections). The FTA algorithm works primarily on pixel-level brightness temperatures and areal statistics derived from the SEVIRI 3.9 µm and 11.0 µm channels. During the first stage a series of absolute thresholds are used with these data to detect "potential fire pixels" (PFPs), which in a second stage are then each assessed as "true" or "false" fire detections based on a series of further "contextual" tests whose thresholds are adjusted according to statistics derived from the immediately neighbouring non-fire "background window" signal surrounding each potential fire pixel. Each PFP must pass all tests to be confirmed as a "true" fire pixel, and a confidence measure is also assigned to each confirmed detection. The final stage focuses on the derivation of an atmospherically corrected FRP estimate for each confirmed fire pixel, together with the corresponding FRP uncertainty. FRP derivation is conducted using the Wooster et al. (2003) "MIR radiance method" of FRP derivation , and the LSA SAF implementation is fully detailed in .|
|The FRPPIXEL product is derived every 15 min at the native SEVIRI pixel spatial resolution and across the full SEVIRI imaging disk . The disseminated product includes two HDF5 format files. The first is the Quality Product file that for each SEVIRI pixel provides details of why it was or was not detected as an active fire pixel. The second is the List Product file that for all confirmed active fire pixels stores the location and time of the detection together with the fire pixels atmospherically corrected FRP and uncertainty (MW), the fire detection confidence measure (representing the level of confidence that the observation is indeed a "true" fire), and various other metrics including the individual spectral channel input signals. The Supplement to Reference  describes both files in detail. Because it stores data only at the location of confirmed fire pixel detections (in an ASCII-like table), the HDF5 List Product file is typically very much smaller (e.g. 100 to 1000x) than the corresponding HDF5 Quality Product file (which stores data across the full SEVIRI image disk in a 2D matrix). However, when compressed the two files are much more similar in size as the Quality Product file in particular compresses very strongly. Nevertheless, use of the List Product may be sufficient for many users, but some will want to access the Quality Product that can be very useful when interpreting the List Product data and for making comparisons to other datasets (e.g. polar orbiting active fire datasets) that benefit from an understanding of e.g. cloud cover etc at the time of the imaging slot. The Quality Product is a mandatory input into the LSA SAF FRPGRID product processing chain, since that product includes adjustment of the recorded FRP for the influence of cloud cover (and also an empirical correction for the impact of undetected "low FRP" fires that remain below SEVIRI's fire detection threshold).|
|The performance of the operational FRPPIXEL product has been assessed via comparisons to both other geostationary fire products derived from Meteosat data, and to near-simultaneous detections made from MODIS. The product was found to both meet its performance requirements with respect to MODIS, and to be the best performing geostationary fire product currently available at testing. The product [VR] along with  provide details of these comparisons.|
| Wooster, M. J., Roberts, G., Freeborn, P. H., Xu, W., Govaerts, Y., Beeby, R., He, J., Lattanzio, A., Fisher, D., and Mullen, R. (2015) LSA SAF Meteosat FRP products - Part 1: Algorithms, product contents, and analysis, Atmos. Chem. Phys., 15, 13217-13239, doi:10.5194/acp-15-13217-2015.
 Roberts, G., Wooster, M. J., Xu, W., Freeborn, P. H., Morcrette, J.-J., Jones, L., Benedetti, A., Jiangping, H., Fisher, D., and Kaiser, J. W. (2015) LSA SAF Meteosat FRP products - Part 2: Evaluation and demonstration for use in the Copernicus Atmosphere Monitoring Service (CAMS), Atmos. Chem. Phys., 15, 13241-13267, doi:10.5194/acp-15-13241-2015.
 Wooster, M. J., G. Roberts, G. L. W. Perry, and Y. J. Kaufman (2005), Retrieval of biomass combustion rates and totals from fire radiative power observations: FRP derivation and calibration relationships between biomass consumption and fire radiative energy release, J. Geophys. Res., 110, D24311, doi:10.1029/2005JD006318.
 Wooster, M. J., Zhukov, B., and Oertel, D., (2003) Fire radiative energy for quantitative study of biomass burning: derivation from the BIRD experimental satellite and comparison to MODIS fire products, Remote Sens. Environ. 86. 83-107.
 Freeborn, P.H., Wooster, M.J., Roberts, G. and Xu, W., 2014. Evaluating the SEVIRI fire thermal anomaly detection algorithm across the Central African Republic using the MODIS active fire product. Remote Sens., 6(3), pp.1890-1917.
 Roberts, G., Wooster, M.J. and Lagoudakis, E., 2009. Annual and diurnal African biomass burning temporal dynamics. Biogeosciences, 6, pp.849-866.
 Freeborn, P.H., Wooster, M.J., Roberts, G., Malamud, B.D. and Xu, W., 2009. Development of a virtual active fire product for Africa through a synthesis of geostationary and polar orbiting satellite data. Remote Sens. Environ., 113(8), pp.1700-1711.
 Andela, N., Kaiser, J. W., van der Werf, G. R., and Wooster, M. J. (2015) New fire diurnal cycle characterizations to improve fire radiative energy assessments made from MODIS observations, Atmos. Chem. Phys., 15, 8831-8846, doi:10.5194/acp-15-8831-2015.
Example of Product
Example Python code to read the product: