How long should a remote sensing-based time series of above ground biomass be to confidently detect gains ?

Monitoring the above-ground biomass on a global scale is essential to improve our understanding of long-term carbon cycle dynamics. Variations of carbon stock occur over different time scales, and in the absence of disturbance vegetation can provide long term carbon storage. While remote sensing can help measuring these annual changes, detecting significant trends is challenging with slow gains or high AGB estimation uncertainty. We show that using current retrieval methods, at least ~30 year long AGB time series data are required to detect density gains of 1 - 2 Mg/ha/year. We further explore the relationship between detectable gains and different retrieval methods and compare these requirements to the time series length currently available from remote sensing-based products.