Background

Forest degradation involves the progressive deterioration of forest structure and function without complete land-use change [1], and contrary to deforestation does not necessarily entails the permanent conversion of forested land to non-forest uses. It reduces ecosystem quality and affects biodiversity, water regulation, and carbon sequestration. In tropical regions, the main drivers are selective logging, fire, and pests or diseases, which differently impact canopy structure, moisture, and biochemical properties. Monitoring forest degradation in Near Real-Time (NRT) is therefore fundamental, as it often represents an early stage or precursor of deforestation, providing critical information for preventive action and sustainable forest management.

Several methods for NRT deforestation monitoring exist today, relying either on multispectral [2], or SAR imagery [3][4][5]. A recent study proposed a multi-source fusion approach combining SAR and multispectral data to improve detection reliability across diverse environmental conditions [6]. However, monitoring forest degradation in NRT remains considerably more challenging. Degradation processes are often progressive rather than abrupt [7], and can occur at very fine spatial scales, i.e., below 0.1 ha [8]. In addition, degradation does not always result in a clear binary change between forest and non-forest states, but rather in subtle alterations of canopy structure, biomass, and spectral properties.

Today, the growing constellation of Earth observation satellite missions provides new opportunities to address these challenges. SAR sensors operating at different frequencies (C-band with Sentinel-1, L-band with ALOS-4, SAOCOM and NiSAR, and P-band with BIOMASS) capture complementary information on forest structure and moisture, while multispectral and thermal missions (Sentinel-2, Landsat-8, MODIS, and VIIRS) provide detailed indicators of vegetation condition. The integration of these multi-resolution, multi-sensor datasets opens promising avenues monitoring of tropical forest degradation.

Description

The proposed PhD aims to design and evaluate advanced methods for multi-frequency and multi-source monitoring of tropical forest degradation, with an emphasis on the integration of radar, multispectral, and thermal satellite data.

Main objectives:

1) A technique should be designed for multi-frequency and multi-source integration by combining C-, L-, and P-band SAR data with multispectral and thermal imagery to characterize structural and spectral indicators of degradation caused by fire, logging, and disease.

2) The detection of high-resolution degradation (selective logging) will be investigated by developing advanced and automatic image coregistration and fusion methods to detect subtle disturbances typically smaller than 0.1 ha, while preserving spatial resolution.

3) The monitoring of fire-induced degradation shall explore the integration of SAR with infrared and multispectral indicators to characterize canopy damage and post-fire recovery dynamics.

4) The monitoring of forest diseases and environmental stress shall be conducted by discriminating anomalous behaviors within long time-series [9] to detect gradual canopy stress or mortality from combined SAR and spectral indicators.

[1] Bourgoin, C., Ceccherini, G., Girardello, M. et al. Human degradation of tropical moist forests is greater than previously estimated. Nature 631, 570–576 (2024). 

[2] Hansen, M., Krylov, A., Tyukavina, A. et al. Humid tropical forest disturbance alerts using Landsat data. ERL 11, 034008 (2016).

[3] Bouvet, A., Mermoz, S., Ballère, M. et al. Use of the SAR Shadowing Effect for Deforestation Detection with Sentinel-1 Time Series. Remote Sensing 10, 1250 (2018).

[4] Reiche, J., Verhoeven, R., Verbesselt, J. et al. Characterizing Tropical Forest Cover Loss Using Dense Sentinel-1 Data and Active Fire Alerts. Remote Sensing, 10, 777 (2018).

[5] Bottani, M., Ferro-Famil, L., Doblas, J. et al. Novel unsupervised Bayesian method for Near Real-Time forest loss detection using Sentinel-1 SAR time series: Assessment over sampled deforestation events in Amazonia and the Cerrado. RSE, 331, 115037 (2025).

[6] Bottani, M., Ferro-Famil, L., Tourneret, J.Y. Multi-source fusion using Bayesian online change detection: application to deforestation monitoring using SAR-optical time series. In Proceedings of CAMSAP (2025).

[7] Bottani, M., Ferro-Famil, Poccard-Chapuis, R. et al. Continuous monitoring of fire-induced forest loss using Sentinel-1 SAR time series and a Bayesian method: A case study in Paragominas, Brazil. Remote Sensing 17, 2822 (2025).

[8] Dupuis, C., Fayolle, A., Bastin, J.F. et al. Monitoring selective logging intensities in central Africa with sentinel-1: A canopy disturbance experiment. RSE, 298, 113828 (2023).

[9] Leon-Lopez, Kareth M., et al. "Anomaly detection and classification in multispectral time series based on hidden Markov models." IEEE TGRS, 60, 1-11 (2021).

=================

For more Information about the topics and the co-financial partner (found by the lab!); contact Directeur de thèse - Laurent.Ferro-Famil@isae-supaero.fr

Then, prepare a resume, a recent transcript and a reference letter from your M2 supervisor/ engineering school director and you will be ready to apply online before March 13th, 2026 Midnight Paris time!