In a context of global changes, modeling and predicting the dynamics of soil carbon stocks (CSs) in forest ecosystems are vital but challenging. Yasso07 is considered to be one of the most promising models for such a purpose. We examine the accuracy of its prediction of soil carbon dynamics over the whole French metropolitan territory at a decennial timescale.
We used data from 101 sites in the RENECOFOR network, which encompasses most of the French temperate forests. These data include (i) the quantity of above-ground litterfall from 1994 to 2008, measured yearly, and (ii) the soil CSs measured twice at an interval of approximately 15 years (once in the early 1990s and around 2010). We used Yasso07 to simulate the annual changes in carbon stocks (ACCs; in tC ha(-1) yr(-1)) for each site and then compared the estimates with actual recorded data. We carried out meta-analyses to reveal the variability in litter biochemistry in different tree organs for conifers and broadleaves. We also performed sensitivity analyses to explore Yasso07's sensitivity to annual litter inputs and model initialization settings.
At the national level, the simulated ACCs (+0.00 +/- 0.07 tC ha(-1) yr(-1), mean +/- SE) were of the same order of magnitude as the observed ones (+0.34 +/- 0.06 tC ha(-1) yr(-1)). However, the correlation between predicted and measured ACCs remained weak (R-2 < 0.1). There was significant overestimation for broadleaved stands and underestimation for coniferous sites. Sensitivity analyses showed that the final estimated CS was strongly affected by settings in the model initialization, including litter and soil carbon quantity and quality and also by simulation length. Carbon quality set with the partial steady-state assumption gave a better fit than the model with the complete steady-state assumption.
With Yasso07 as the support model, we showed that there is currently a bottleneck in soil carbon modeling and prediction due to a lack of knowledge or data on soil carbon quality and fine-root quantity in the litter.