Lecture 2

The RF-based PlanetScope-Sentinel-2 Data Fusion for Phenometrics Retrieval

Description: Monitoring crop phenology is crucial for optimizing agricultural management, resource allocation, and climate adaptation. Phenological shifts, influenced by climate change, extreme weather, and management practices, affect ecosystem services, agricultural productivity, and food security. Near real-time monitoring of crop cover and growth during early growth stages, enabling farmers to evaluate germination success and adapt practices to mitigate potential yield losses. RS time-series enable the extraction of Land Surface Phenology (LSP) metrics, such as Start-, Peak-, and End-of-Season (SoS, PoS, and EoS respectively), by capturing variations in crop characteristics like biomass and chlorophyll content. These metrics correspond to phenological transition dates and provide insights into crop growth dynamics at the field and sub-field levels. Both high temporal frequency and spatial detail in LSP retrieval remains a challenge due to trade-offs in satellite datasets.

Theoretical part (Lecture):

• Understanding the difference between phenology and LSP (including BBCH)
• Background on maize phenology and retrieval methods /software
• Vegetation indices
• Background on data fusion methods inkl. RF
• Validation and reference data (HR-VPP, DWD)

Practical part:

Develop skills to:

• Apply RF model on PlanetScope data to create synthetic S2 data and fill time-series gaps [Notebook 1]
• Build time-series of Sentinel-2 data and retrieve phenometrics (SoS, EoS, PoS) [Notebook 2]
• Compare retrieved LSP metrics to references [Notebook 3].

Duration: 3,5 hr (2,5 theoretical part, 1h practical exercises)

Lecture Flow (content):

2.1. Sentinel-2 and PlanetScope data, Vegetation indices
2.2. Wrap-up on optical data fusion methods
2.3. LSP definition and retrieval methods
2.4. Maize phenological stages and the BBCH Plant Classification System
2.5. References and validation
2.6. Discussion on methods advantages and limitations
2.7. The use of remote sensing based phenometrics
2.8. Summary and final thoughts
- Quiz

Outcomes:

• The ability to use optical RS time-series to understand crop phenology
• Acquire skills in RF-fusion of multimodal optical RS data with respect to spatial, temporal and spectral resolution
• Evaluate the impact of data fusion on phenometrics accuracy