Grateful to work with colleagues across @nasajpl.bsky.social, @caltech.edu , @stanforduniversity.bsky.social and partners worldwide.

Why it matters: CO₂ fertilization still props up land uptake, but climate-driven soil losses and accelerated turnover challenge the robustness of the sink. We need sustained monitoring + process-aware, observation-constrained models.

How we did it: Bayesian fusion of multi-mission Earth observations (OCO-2, GRACE/GRACE-FO, LAI, biomass, fires, runoff, etc.) to build an observation-informed terrestrial reanalysis.

Turnover matters: both CO₂ and climate shorten residence times (−2.6% and −1.3%, respectively) → a larger but faster-cycling sink with implications for resilience.

Where it went: gains accrue mostly to live biomass (+31.2 PgC). Climate pressures drive dead-organic (soil) losses (−8.8 PgC) and shift dead C from higher latitudes toward the tropics.

CO₂ rise added +39.4 PgC to land; climate trends removed −10.5 PgC → net +29.7 PgC over 2001–2021.

We used the CARDAMOM model–data fusion framework to separate impacts of rising CO₂ from contemporary climate trends (2001–2021), tracking not just fluxes but where carbon sits (live biomass vs dead organic C) and how long it stays.

CARDAMOM fuses satellites, field data, and models to better understand & predict ecosystem change. 🚀

#CarbonCycle #EcosystemModeling #DataAssimilation

Most models still use broad “tropical tree” or “grass” categories with fixed optics.

But real leaves vary in chlorophyll, thickness, & water content — and it matters. A lot.

Trait-based optics = better realism, better predictions.
#EarthSystem #CliMA #LeafTraits

Full paper: