The length of time that carbon remains in forest biomass is one of the largest uncertainties in the global carbon cycle. Estimates from six contemporary models found this time to range from 12.2 to 23.5 years for the global mean for 1985–2014. Future projections do not give consistent results, but 13 model-based hypotheses are identified, along with recommendations for pragmatic steps to test them using existing and novel observations, which would help to reduce large current uncertainty.

Ecosystem CO₂ uptake changes in time depending on climate conditions. In this study, we analyze how different climate variables affect the timing when CO₂ uptake is at a maximum (DOY_GPPmax). We found that the joint effects of radiation, temperature, and vapor pressure deficit are the most relevant controlling factors of DOY_GPPmax and that if they increase, DOY_GPPmax will happen earlier. These results help us to better understand how CO₂ uptake could be affected by climate change.

Carbon dioxide is produced by soil microbes through respiration, which is particularly fast when soils are moistened by rain. Will respiration increase with future more intense rains and longer dry spells? With a mathematical model, we show that wetter conditions increase respiration. In contrast, if rainfall totals stay the same, but rain comes all at once after long dry spells, the average respiration will not change, but the contribution of the respiration bursts after rain will increase.

The Arctic tundra represents a vast store of carbon that may be broken down by microbial activity into greenhouse gases such as CO₂ and CH₄. Though microbes are less active in winter, the long duration of the cold season makes this period very important for carbon cycling. We show that, under conditions of warmer winter air temperatures and greater snowfall, deeper soils can remain warm enough to sustain significantly enhanced CH₄ emission. This could have large implications for future climates.

Soil biogeochemical models (SBMs) are needed to predict future soil CO₂ emissions levels, but we presently lack statistically rigorous frameworks for assessing the predictive utility of SBMs. In this study, we demonstrate one possible approach to evaluating SBMs by comparing the fits of two models to soil CO₂ respiration data with recently developed Bayesian statistical goodness-of-fit metrics. Our results demonstrate that our approach is a viable one for continued development and refinement.

We assess trade-offs between different types of biological observations, specifically satellite ocean color and BGC-Argo profiles and the benefits of combining both for optimizing a biogeochemical model of the Gulf of Mexico. Using all available observations leads to significant improvements in observed and unobserved variables (including primary production and C export). Our results highlight the significant benefits of BGC-Argo measurements for biogeochemical model optimization and validation.

We combine several lines of evidence to provide a robust estimate of historical CO₂ emissions from land use change. Our novel approach leads to reduced uncertainty and identifies key remaining sources of uncertainty and discrepancy. We also quantify the carbon removal by natural ecosystems that would have occurred if these ecosystems had not been destroyed (mostly via deforestation). Over the last decade, this foregone carbon sink amounted to about 50 % of the actual emissions.

Temperate forest soils (sites dominated by European beech, Fagus sylvatica) from cold and snowy sites in northern Poland release more nitrogen and phosphorus after soil freeze–thaw cycles (FTCs) than soils from warmer, snow-poor conditions in northern Germany. Our data suggest that previously cold sites, which will lose their protective snow cover during climate change, are most vulnerable to increasing FTC frequency and magnitude, resulting in strong shifts in nitrogen leaching.

We describe a method for correcting errors in oxygen optode measurements on autonomous platforms in the ocean. The errors result from the relatively slow response time of the sensor. The correction method includes an in situ determination of the effective response time and requires the time stamps of the individual measurements. It is highly relevant for the BGC-Argo program and also applicable to gliders. We also explore if diurnal changes in oxygen can be obtained from profiling floats.