With Sentinel-1A and -1B (launched in 2014 and 2016) for the first time, dense and regular SAR time series data are provided globally and openly. Such potential needs to be utilized.
In the past 10 years, satellite-based alert systems (Reiche et al., 2021, Hansen et al. 2016) have emerged as the primary tool to provide near real-time information on newly disturbed tropical forest areas. Limited data availability due to persistent cloud cover in the tropics, however, limits the capacity of optical-based systems (e.g. GLAD alerts; Hansen et al, 2016). With the new RADD (Radar for Detecting Deforestation) alerts (Reiche et al., 2021) for the first time high resolution (10 m) radar-based forest disturbance information (every 6 – 12 days) are provided for the humid tropics.
While optical and radar-based forest monitoring systems show good accuracies for the humid tropics where seasonality and dynamics of forest are moderate, accuracies in dry forest and boreal forest areas with more seasonality are much poorer.
To make full use of the available Sentinel-1 data stream for forest monitoring outside of the humid tropics, we need to study the seasonality of the Sentinel-1 radar signal and how it is affected by climate events (e.g. droughts, El Nino).
This thesis will assess the seasonality/dynamics of the Sentinel-1 C-band radar backscatter signal across a variety of forest biomes globally (e.g. tropical humid forest, tropical dry forest, temperate broadleaved forest, boreal forest ...). This thesis will look at dense forest only (e.g. tree cover above 60%) and not consider disturbances.
First, a sample-based dataset of Sentinel-1 time series for different forest ecosystems will be retrieved from Google Earth Engine. Second, different seasonality measures and models (e.g. harmonic function a used in the BFAST package) will be used to model and characterize the seasonality of the backscatter signal over the past 4-5 years. Third, the robustness of such seasonal model in the case of unexpected climate extreme events (e.g. droughts, El Nino) will be studied.
Software: Google Earth Engine and R (potentially also python)
- Generate a sample-based (pixel-based) dataset of Sentinel-1 backscatter time series across global forest ecosystems
- Model and characterize the seasonality at a pixel level and compare results across forest ecosystems
- Assess the robustness of the seasonal models in the case of climate extreme events
- Reiche et al. (2021). Forest Disturbance alerts for the Congo Basin using Sentinel-1. Environmental Research Letters https://doi.org/10.1088/1748-9326/abd0a8
- Vebesselt et al. (2011): Detecting trend and seasonal changes in satellite image time series, RSE, https://www.sciencedirect.com/science/article/abs/pii/S003442570900265X
- Reiche et al., (2018): Characterizing Tropical Forest Cover Loss Using Dense Sentinel-1 Data and Active Fire Alerts. Remote Sensing. https://www.mdpi.com/2072-4292/10/5/777
- Dostalova et al., (2018): Annual seasonality in Sentinel-1 signal for forest mapping and forest type classification. International Journal of Remote Sensing https://www.tandfonline.com/doi/full/10.1080/01431161.2018.1479788
- Global Ecological Zones map for FRA 2010. http://www.fao.org/3/ap861e/ap861e.pdf
- Advanced Earth Observation course
- Geo-scripting course (Good knowledge in scripting is an asset; e.g. Google Earth Engine, R, python, java script)
Theme(s): Sensing & measuring; Modelling & visualisation