MSc thesis topic: How Does Density Matter? The Impact of Sentinel-1 Time Series Density on Detecting Various Types of Tropical Forest Disturbances.

With Sentinel-1A and 1B in orbit, revisit times reached up to 6 days in the tropics. However, the failure of Sentinel-1B in late 2021 reduced data coverage and observation frequency in many regions. This thesis investigates the impact of this decreased temporal density on the detection of different forest disturbance events, including fires, selective logging, windthrow, and large-scale clearings. The findings will help improve the understanding and interpretation of radar-based forest disturbance detections and products for the period when only a single Sentinel-1 satellite was operational (2022–2024).

Radar systems can acquire data day and night and penetrate cloud cover, making them ideal for monitoring tropical forests and detecting disturbances (De Sy et al., 2012). In 2014, Sentinel-1 was launched, providing the first large-scale, freely accessible radar data, enabling extensive forest monitoring. Since 2016, two identical Sentinel-1 satellites have been in orbit, offering revisit times of up to 6 days in the tropics (ESA, 2012). This has enabled the emergence of near-real-time products leveraging Sentinel-1 data (e.g. RADD alerts) (Reiche et al., 2021, 2024), providing large-scale, free information on forest disturbances. However, since late 2021, one of the identical Sentinel-1 satellites failed, effectively reducing revisit times and, in some areas, preventing observations altogether (ESA, 2022).

Studies have shown that for disturbance types characterized by small canopy gaps and rapidly regenerating forest cover, fast revisit times or dense Sentinel-1 time series are essential for precise detection (Hethcoat et al., 2021). However, such studies have focused only on selective logging and have not investigated other disturbance types, such as forest fires, wind throw, or other events that also exhibit rapid recovery of the C-band SAR signal. This fast recovery can complicate detection or even result in missing the disturbance entirely. Additionally, no studies have explored the impact of reduced time series on the timeliness of disturbance detection.

This study will apply existing methods (Reiche et al., 2015, 2018; Verbesselt et al., 2012) for disturbance detection on simulated, reduced time series of Sentinel-1 data from 2016-2020, focusing on known disturbance events in South America (Balling et al., 2023). Using data from 2016 to 2020 prevents any impact from Sentinel-1B's failure. Various scenarios of sparse time series will be tested to measure the effects on detection accuracy and timeliness. The disturbance types include common events in the tropics, such as forest fires, selective logging, wind throw, and large-scale logging. Additionally, the student may explore the potential benefits of incorporating additional information, such as texture features (e.g., GLCM) (Hall-Beyer, 2017; Haralick et al., 1973), to mitigate the limitations caused by reduced time series density.

The insights gained will help understand the limitations of Sentinel-1 in detecting disturbance types with reduced time series and contribute to refining current radar-based disturbance products (e.g., RADD Alerts).

Note: This study may also be conducted using optical remote sensing data, such as Landsat or Sentinel-2.

Relevance to research/projects at GRS or other groups

• Understanding the impact of reduced Sentinel-1 time series due to the failure of Sentinel-1B on the detection capability of radar for tropical forest disturbances.
• Gaining insights into the results of RADD Alerts from 2022 to the present.

Objectives and Research questions

• Assessing the impact of Sentinel-1 time series density on the detection capability for various forest disturbance events.
• Investigating the effect of varying time series density levels on the timeliness of detection.
• Optional: Exploring the added benefit of incorporating additional information, such as texture, to mitigate the limitations of reduced time series.

Requirements
Courses:

• Geo-scripting (GRS33806)
• Advanced Earth Observation (GRS32306)

Software:

• Google Earth Engine, R, Python

Literature and information

• De Sy, V., Herold, M., Achard, F., Asner, G. P., Held, A., Kellndorfer, J., & Verbesselt, J. (2012). Synergies of multiple remote sensing data sources for REDD+ monitoring. Current Opinion in Environmental Sustainability, 4(6), 696–706.
• ESA. (2012). Sentinel-1: ESA’s Radar Observatory Mission for GMES Operational Services. In ESA Special Publication. ESA Communications.
• ESA. (2022, August 3). Mission ends for Copernicus Sentinel-1B satellite.
• Hall-Beyer, M. (2017). Practical guidelines for choosing GLCM textures to use in landscape classification tasks over a range of moderate spatial scales. International Journal of Remote Sensing, 38(5), 1312–1338.
• Haralick, R. M., Shanmugam, K., & Dinstein, I. (1973). Textural Features for Image Classification. IEEE Transactions on Systems Science and Cybernetics, 4, 610–621.
• Reiche, J., de Bruin, S., Hoekman, D., Verbesselt, J., & Herold, M. (2015). A Bayesian approach to combine landsat and ALOS PALSAR time series for near real-time deforestation detection. Remote Sensing, 7(5), 4973–4996.
• Reiche, J., Hamunyela, E., Verbesselt, J., Hoekman, D., & Herold, M. (2018). Improving near-real time deforestation monitoring in tropical dry forests by combining dense Sentinel-1 time series with Landsat and ALOS-2 PALSAR-2. Remote Sensing of Environment, 204(November 2017), 147–161.
• Reiche, J., Mullissa, A., Slagter, B., Gou, Y., Tsendbazar, N.-E., Odongo-Braun, C., Vollrath, A., Weisse, M. J., Stolle, F., Pickens, A., Donchyts, G., Clinton, N., Gorelick, N., & Herold, M. (2021). Forest disturbance alerts for the Congo Basin using Sentinel-1. Environmental Research Letters, 16(2), 024005.

Expected reading list before starting the thesis research

• Balling, J., Herold, M., & Reiche, J. (2023). How textural features can improve SAR-based tropical forest disturbance mapping. International Journal of Applied Earth Observation and Geoinformation, 124, 103492.
• Hethcoat, M. G., Carreiras, J. M. B., Edwards, D. P., Bryant, R. G., & Quegan, S. (2021). Detecting tropical selective logging with C-band SAR data may require a time series approach. Remote Sensing of Environment, 259, 112411.
• Reiche, J., Balling, J., Herold, M., Slagter, B., & Tsendbazar, N. (2024). RADD Forest Disturbance Alert. Retrieved May 14, 2024, from
• Reiche, J., Mullissa, A., Slagter, B., Gou, Y., Tsendbazar, N.-E., Odongo-Braun, C., Vollrath, A., Weisse, M. J., Stolle, F., Pickens, A., Donchyts, G., Clinton, N., Gorelick, N., & Herold, M. (2021). Forest disturbance alerts for the Congo Basin using Sentinel-1. Environmental Research Letters, 16(2), 024005.
• Verbesselt, J., Zeileis, A., & Herold, M. (2012). Near real-time disturbance detection using satellite image time series. Remote Sensing of Environment, 123, 98–108.

Theme(s): Sensing & measuring; Modelling & visualisation