Fore[st]cast - Nowcast and predict forest condition

Forests host about 90% of the world's terrestrial biomass in the form of carbon and are an important pool for global biodiversity. We need long-term monitoring but a near real-time data processing to fulfill the needs from science and the public. We propose here a further step toward near real-time assessment of the condition of our forests, as has  traditionally been done in meteorology or hydrology. Our goal is to further develop the TreeNet platform that reports on the current and predicted condition of trees (e.g., growth and drought stress) in relation to environmental variables.  

The milestones of this proposal aim to

1. implement a recently developed approach to gap-fill data with machine learning,
2. develop a method for anomaly detection,
3. interpolate point data to larger areas, and
4. develop tree growth and drought stress predictions.