A two-stage optical fusion framework for wildfire severity mapping across the conterminous United States

Abstract

Accurate wildfire severity mapping (WSM) is essential for post-fire recovery planning, erosion risk assessment, ecosystem monitoring, and disaster risk reduction. Although Landsat and Sentinel optical imagery have been widely used for burn severity assessment, the added value of fusing multiple optical sensors has not been sufficiently quantified across diverse fire events, particularly since the launch of Landsat-9. This study evaluates whether multisensor optical fusion improves wildfire severity mapping relative to single-sensor baselines using Sentinel-2, Landsat-8, and Landsat-9 imagery across 40 wildfire events in the conterminous United States. We tested a two-stage fusion framework that combines feature-level fusion with pixel-level dimensionality reduction. First, feature-level fused datasets were created through early fusion by combining standardized post-fire bands from each sensor into a single predictor stack. Both raw reflectance bands and pairwise spectral transforms were retained to capture within- and cross-sensor spectral interactions. Second, Linear Discriminant Analysis was applied to both single-sensor and fused datasets to produce comparable low-dimensional feature spaces. Six machine-learning classifiers were then used to benchmark model performance with repeated spatially buffered train–test splits. Results show that Landsat-9 was the strongest single-sensor baseline. Among the fusion strategies, Sentinel-2 + Landsat-9 produced the most consistent improvement and reduced performance variability. Landscape-condition analysis further showed that this fusion was most beneficial in shrubland-dominated and high-terrain fires, where it achieved the highest overall mean accuracy and the fewest failures. In contrast, its benefits were less reliable in evergreen forests, mixed vegetation, and low- to moderate-elevation terrain. In operational settings, the Sentinel-2 + Landsat-9 configuration offers a practical solution for post-fire recovery planning, erosion-risk assessment, watershed management, and ecological monitoring when field observations are available and timely satellite-based information is needed.

Keywords

Wildfire severity mapping / Optical data fusion / Machine learning / Landsat / Sentinel

References

[1] Addison, P., Oommen, T., 2018. Utilizing satellite radar remote sensing for burn severity estimation. Int. J. Appl. Earth. Obs. Geoinf. 73, 292-299. doi: 10.1016/j.jag.2018.07.002.

[2] Arango, E., Jiménez, P., Nogal, M., Sousa, H.S., Stewart, M.G., Matos, J.C., 2024. Enhancing infrastructure resilience in wildfire management to face extreme events: insights from the Iberian Peninsula. Clim. Risk Manage. 44, 100595. doi: 10.1016/j.crm.2024.100595.

[3] Belenguer-Plomer, M.A., Tanase, M.A., Chuvieco, E., Bovolo, F., 2021. CNN-based burned area mapping using radar and optical data. Remote Sens. Environ. 260, 112468. doi: 10.1016/j.rse.2021.112468.

[4] Brown, A.R., Petropoulos, G.P., Ferentinos, K.P., 2018. Appraisal of the Sentinel-1 & 2 use in a large-scale wildfire assessment: a case study from Portugal’s fires of 2017. Appl. Geogr. 100, 78-89. doi: 10.1016/j.apgeog.2018.10.004.

[5] Burnett, J.T., Edgeley, C.M., 2023. Factors influencing flood risk mitigation after wildfire: insights for individual and collective action after the 2010 Schultz Fire. Int. J. Disaster. Risk. Reduct. 94, 103791. doi: 10.1016/j.ijdrr.2023.103791.

[6] De Luca, G., Silva, J.M.N., Modica, G., 2021. A workflow based on Sentinel-1 SAR data and open-source algorithms for unsupervised burned area detection in Mediterranean ecosystems. GIsci. Remote Sens. 58 (4), 516-541. doi: 10.1080/15481603.2021.1907896.

[7] Demir, S., Başayiğit, L., 2024. Digital mapping burn severity in agricultural and forestry land over a half-decade using sentinel satellite images on the Google Earth Engine platform: a case study in Isparta province. Trees. For. People 16, 100520. doi: 10.1016/j.tfp.2024.100520.

[8] Fisher, R.A., 1936. The use of multiple measurements in taxonomic problems. Ann. Eugen. 7 (2), 179-188. doi: 10.1111/j.1469-1809.1936.tb02137.x.

[9] Geng, J., Tan, Q.Y., Lv, J.W., Fang, H.J., 2024. Assessing spatial variations in soil organic carbon and C:N ratio in Northeast China’s black soil region: insights from Landsat-9 satellite and crop growth information. Soil. Tillage Res. 235, 105897. doi: 10.1016/j.still.2023.105897.

[10] Ghassemian, H., 2016. A review of remote sensing image fusion methods. Inf. Fusion. 32, 75-89. doi: 10.1016/j.infus.2016.03.003.

[11] Hosseini, M., Lim, S ., 2023. Burned area detection using Sentinel-1 SAR data: a case study of Kangaroo Island, South Australia. Appl. Geogr. 151, 102854. doi: 10.1016/j.apgeog.2022.102854.

[12] Howe, A.A., Parks, S.A., Harvey, B.J., Saberi, S.J., Lutz, J.A., Yocom, L.L., 2022. Comparing Sentinel-2 and Landsat 8 for burn severity mapping in western North America. Remote Sens. 14 (20), 5249. doi: 10.3390/rs14205249.

[13] Hu, X.K., Zhang, P.Z., Ban, Y.F., Rahnemoonfar, M., 2023. GAN-based SAR and optical image translation for wildfire impact assessment using multi-source remote sensing data. Remote Sens. Environ. 289, 113522. doi: 10.1016/j.rse.2023.113522.

[14] Jiao, L.P., Bo, Y.C., 2022. Near real-time mapping of burned area by synergizing multiple satellites remote-sensing data. GISci. Remote Sens. 59 (1), 1956-1977. doi: 10.1080/15481603.2022.2143690.

[15] Jodhani, K.H., Patel, H., Soni, U., Patel, R., Valodara, B., Gupta, N., Patel, A., Omar, P.J., 2024. Assessment of forest fire severity and land surface temperature using Google Earth Engine: a case study of Gujarat State, India. Fire Ecol. 20 (1), 23. doi: 10.1186/s42408-024-00254-2.

[16] Kabir, S., Pahlevan, N., O’Shea, R.E., Barnes, B.B., 2023. Leveraging Landsat-8/-9 underfly observations to evaluate consistency in reflectance products over aquatic environments. Remote Sens. Environ. 296, 113755. doi: 10.1016/j.rse.2023.113755.

[17] Kadakci Koca, T., Küçükuysal, C., Gül, M., Esetlili, T., 2024. A comprehensive approach to soil burn severity mapping for erosion susceptibility assessment. Catena 245, 108302. doi: 10.1016/j.catena.2024.108302.

[18] Kim, B., Lee, K., Park, S., 2024. Burned-area mapping using post-fire PlanetScope images and a convolutional neural network. Remote Sens. 16 (14), 2629. doi: 10.3390/rs16142629.

[19] Liu, W.J., Guan, H.D., Hesp, P.A., Batelaan, O., 2023. Remote sensing delineation of wildfire spatial extents and post-fire recovery along a semi-arid climate gradient. Ecol. Inform. 78, 102304. doi: 10.1016/j.ecoinf.2023.102304.

[20] Lucas Borja, M.E., Zema, D.A., 2024. Delayed application of straw mulching increases soil erosion in Mediterranean pine forests burned by wildfires. Catena 236, 107714. doi: 10.1016/j.catena.2023.107714.

[21] Martins, V.S., Roy, D.P., Huang, H., Boschetti, L., Zhang, H.K., Yan, L., 2022. Deep learning high resolution burned area mapping by transfer learning from Landsat-8 to PlanetScope. Remote Sens. Environ. 280, 113203. doi: 10.1016/j.rse.2022.113203.

[22] Masek, J.G., Wulder, M.A., Markham, B., McCorkel, J., Crawford, C.J., Storey, J., Jenstrom, D.T., 2020. Landsat 9: empowering open science and applications through continuity. Remote Sens. Environ. 248, 111968. doi: 10.1016/j.rse.2020.111968.

[23] Miller, J.D., Thode, A.E., 2007. Quantifying burn severity in a heterogeneous landscape with a relative version of the delta normalized burn ratio (dNBR). Remote Sens. Environ. 109 (1), 66-80. doi: 10.1016/j.rse.2006.12.006.

[24] Montero, D., Aybar, C., Mahecha, M.D., Martinuzzi, F., Söchting, M., Wieneke, S., 2023. A standardized catalogue of spectral indices to advance the use of remote sensing in Earth system research. Sci. Data 10, 197. doi: 10.1038/s41597-023-02096-0.

[25] Montorio, R., Pérez-Cabello, F., Borini Alves, D., García-Martín, A., 2020. Unitemporal approach to fire severity mapping using multispectral synthetic databases and Random Forests. Remote Sens. Environ. 249, 112025. doi: 10.1016/j.rse.2020.112025.

[26] Moody, J.A., Ebel, B.A., 2012. Hyper-dry conditions provide new insights into the cause of extreme floods after wildfire. Catena 93, 58-63. doi: 10.1016/j.catena.2012.01.006.

[27] Nguyen Van, L., Lee, G., 2025. Optimizing stacked ensemble machine learning models for accurate wildfire severity mapping. Remote Sens. 17 (5), 854. doi: 10.3390/rs17050854.

[28] Nguyen Van, L., Lee, G., 2024. Underutilized feature extraction methods for burn severity mapping: a comprehensive evaluation. Remote Sens. 16 (22), 4339. doi: 10.3390/rs16224339.

[29] Ousmanou, S., Martial, F.E., Jules, T.K., Ludovic, A.M., Agnès Blandine, K.T., Sufinatu, A., Mohamed, R., Maurice, K., 2024. Mapping and discrimination of the mineralization potential in granitoids from Banyo area (Adamawa, Cameroon), using Landsat 9 OLI, ASTER images and field observations. Geosyst. Geoenviron. 3 (1), 100239. doi: 10.1016/j.geogeo.2023.100239.

[30] Palmer, J ., 2022. The devastating mudslides that follow forest fires. Nature 601 (7892), 184-186. doi: 10.1038/d41586-022-00028-3.

[31] Pfoch, K.A., Pflugmacher, D., Okujeni, A., Hostert, P., 2023. Mapping forest fire severity using bi-temporal unmixing of Sentinel-2 data -towards a quantitative understanding of fire impacts. Sci. Remote Sens. 8, 100097. doi: 10.1016/j.srs.2023.100097.

[32] Porto, P., Callegari, G., 2021. Using¹³⁷Cs measurements to estimate soil erosion rates in forest stands affected by wildfires. Results from plot experiments . Appl. Radiat. Isot. 172, 109668. doi: 10.1016/j.apradiso.2021.109668.

[33] Quintano, C., Fernández-Manso, A., Fernández-Manso, O ., 2018. Combination of Landsat and Sentinel-2 MSI data for initial assessing of burn severity. Int. J. Appl. Earth. Obs. Geoinf. 64, 221-225. doi: 10.1016/j.jag.2017.09.014.

[34] Roy, D.P., Huang, H.Y., Boschetti, L., Giglio, L., Yan, L., Zhang, H.H., Li, Z.B., 2019. Landsat-8 and Sentinel-2 burned area mapping -a combined sensor multitemporal change detection approach. Remote Sens. Environ. 231, 111254. doi: 10.1016/j.rse.2019.111254.

[35] Running, S.W., 2006. Is global warming causing more, larger wildfires? Science 313 (5789), 927-928. doi: 10.1126/science.1130370.

[36] Schmitt, M., Zhu, X.X., 2016. Data fusion and remote sensing: an ever-growing relationship. IEEE Geosci. Remote Sens. Mag. 4 (4), 6-23. doi: 10.1109/MGRS.2016.2561021.

[37] Seydi, S.T., Sadegh, M., 2023. Improved burned area mapping using monotemporal Landsat-9 imagery and convolutional shift-transformer. Measurement 216, 112961. doi: 10.1016/j.measurement.2023.112961.

[38] Shakesby, R.A., 2011. Post-wildfire soil erosion in the Mediterranean: review and future research directions. Earth-Sci. Rev. 105 (3-4), 71-100. doi: 10.1016/j.earscirev.2011.01.001.

[39] Singh, R., Saritha, V., Pande, C.B., 2024. Monitoring of wetland turbidity using multitemporal Landsat-8 and Landsat-9 satellite imagery in the Bisalpur wetland, Rajasthan, India. Environ. Res. 241, 117638. doi: 10.1016/j.envres.2023.117638.

[40] Tanase, M.A., Belenguer-Plomer, M.A., Roteta, E., Bastarrika, A., Wheeler, J., Fernández-Carrillo, Á., Tansey, K., Wiedemann, W., Navratil, P., Lohberger, S., Siegert, F., Chuvieco, E., 2020. Burned area detection and mapping: intercomparison of Sentinel-1 and Sentinel-2 based algorithms over tropical Africa. Remote Sens. 12 (2), 334. doi: 10.3390/rs12020334.

[41] Tucker, C.J., 1979. Red and photographic infrared linear combinations for monitoring vegetation. Remote Sens. Environ. 8 (2), 127-150. doi: 10.1016/0034-4257(79)90013-0.

[42] Van, L.N., Lee, G., 2025. Tree-based regressor comparison for burn severity mapping: spatially blocked validation within and across fires. Remote Sens. 17 (22), 3756. doi: 10.3390/rs17223756.

[43] Van, L.N., Tran, V.N., Nguyen, G.V., Yeon, M., Do, M.T., Lee, G., 2024. Enhancing wildfire mapping accuracy using mono-temporal Sentinel-2 data: a novel approach through qualitative and quantitative feature selection with explainable AI. Ecol. Inform. 81, 102601. doi: 10.1016/j.ecoinf.2024.102601.

[44] Vetrita, Y., Diwyacitta, K., Sukarno, K.M., Albar, I., Usman, A.B., Ritonga, R.P., Santoso, I., Lestari, A.I., Kartika, T., Novresiandi, D.A., Prasasti, I., Ulfa, K., Novita, N., Siwi, S.E., Augusto, S., Prakoso, E.T., Rahmi, K.I.N., Nugroho, U.C., Widodo, J., Cochrane, M.A., 2025. Evaluating the capabilities of high-resolution PlanetScope and Sentinel-2 images for mapping burned area and vegetation regrowth in Indonesia’s savannas. Int. J. Remote Sens. 46 (18), 6803-6825. doi: 10.1080/01431161.2025.2546154.

[45] Wang, X.G., He, H.S., Li, X.Z., 2007. The long-term effects of fire suppression and reforestation on a forest landscape in northeastern China after a catastrophic wildfire. Landsc. Urban. Plan. 79 (1), 84-95. doi: 10.1016/j.landurbplan.2006.03.010.

[46] White, D.C., Morrison, R.R., Wohl, E., 2022. Fire and ice: winter flooding in a Southern Rocky Mountain stream after a wildfire. Geomorphology 413, 108370. doi: 10.1016/j.geomorph.2022.108370.

[47] Xue, J.R., Su, B.F., 2017. Significant remote sensing vegetation indices: a review of developments and applications. J. Sens. 2017, 1353691. doi: 10.1155/2017/1353691.

[48] Yilmaz, O.S., Acar, U., Sanli, F.B., Gulgen, F., Ates, A.M., 2023. Mapping burn severity and monitoring CO content in Türkiye’s 2021 Wildfires, using Sentinel-2 and Sentinel-5P satellite data on the GEE platform. Earth Sci. Inform. 16 (1), 221-240. doi: 10.1007/s12145-023-00933-9.

[49] Zhang, Q., Ge, L.L., Zhang, R.H., Metternicht, G.I., Du, Z.Y., Kuang, J.M., Xu, M., 2021. Deep-learning-based burned area mapping using the synergy of Sentinel-1&2 data. Remote Sens. Environ. 264, 112575. doi: 10.1016/j.rse.2021.112575.