Preparing land surface models for the next generation of hyperspectral data

Land surface models that describe the global carbon cycle are essential for predicting future climate change. But in these models 300,000 global vascular plant species are represented by fewer than 20 plant functional types, contributing to massive prediction uncertainty. Hyperspectral remote sensing can identify and classify vegetation into species but this is less important to models than providing key information about canopy properties such as leaf nitrogen, specific leaf area, and lignin and chlorophyll concentrations. But are the models ready for these new data? And what is the best method to integrate this information into models?