TerraTrace – Spatio-Temporal Vegetation Signatures for Land Use Analytics

Every year, humanity clears 10 million hectares of forests, which releases more than 5.6 billion tonnes of greenhouse gases. This significant contribution to climate change has led to the passage of global regulations, such as the EUDR, which aims to ensure that products linked to deforestation are excluded from the European market. Satellite-based remote sensing tools are popularly used for global monitoring to enable such compliance. However, they struggle to differentiate vegetation types in farms and orchards from forests (Fig.1.A). To solve this, we develop TerraTrace, a temporal signature mapping tool that combines Spectral Vegetation Indices, Satellite Imagery, and open data like Cropland Data Layer (CDL) to estimate historical land use. The key insight is that satellite-based spectral index data shows temporal variables like agricultural practices and plant growth cycles. Specifically, we demonstrate that yearly patterns of the Normalized Difference Vegetation Index (NDVI), based on plant photosynthesis, have temporal signatures unique to different crops, can distinguish forests from crops, and follow consistent patterns across different locations. Leveraging this we make the following contributions (Fig.2):