The carbon balance of forests worldwide is influenced strongly by human management through afforestation, deforestation, and harvesting. Shifts in their magnitude partly drive the global forest carbon sink on decadal time scales, and regional differences are large: China’s afforestation program increases carbon in forest biomass massively while renewed deforestation in Brazil led to carbon losses to the atmosphere. On top of these changes, shifts in climate means and extremes affect the forest carbon balance through enhanced growth (“fertilization”) by temperature and CO₂, as well as through reduced growth during heatwaves, sometimes associated with large losses through fires. Understanding the current global forest carbon sink, and quantifying and attributing its changes to these drivers is an important first step towards (a) predicting this sink for the coming decades, (b) maximizing the sink to limit the growth of CO₂ in the atmosphere, and (c) assessing the realism of large-scale carbon dioxide removal plans to help achieve climate neutrality in the EU.

The CarbonTracker Europe (CTE) data assimilation system has delivered annual estimates of the net carbon exchange (NEE) between forests and atmosphere worldwide for the past two decades. Its many studies have focused on atmospheric CO₂ variations from hundreds of surface measurement sites, and using these to quantify decadal carbon balances worldwide, and impact of droughts in Europe, North America, and the Amazon. Current developments in CTE focus on a new design of the data assimilation framework to explicitly separate the longer term (annual to decadal) time scale from the shorter (synoptic to seasonal) time scale. This allows high-precision but sparse climate time series to be used alongside high volume remote sensing products on e.g. fires and vegetation stress. A secondary line of development focuses on the integration of newly available above-ground biomass (AGB) data products into CTE, augmenting the atmospheric CO₂ view on net ecosystem exchange. This new data partly comes from remote sensing (combined space-based and aircraft) of AGB, but we also use the >100,000 national forest survey data points on tree diameters, interpreted by the European EFISCEN forestry model, in our data assimilation. The overall goal is to present decadal, annual, and seasonal changes in forest carbon balances that are maximally constrained by available datasets on the ground, from space, and from the atmosphere. The seminar will present our current efforts and first results, and their interpretation in light of the recent European drought(s) and the 2023/2024 El Nino Southern Oscillation.