TL;DR
This paper explores the effectiveness of a mesoscale atmospheric model coupled with a Lagrangian particle dispersion model for reconstructing regional CO2 fluxes. The study demonstrates that tower measurements are crucial for significantly reducing uncertainties in CO2 flux estimates, with aircraft data providing a noticeable, though secondary, contribution.
What Problem Does This Paper Address?
Accurately quantifying carbon dioxide (CO2) fluxes at regional scales is a critical challenge for understanding the global carbon cycle. Traditional "bottom-up" approaches rely on aggregating local flux measurements, which can be difficult to scale up and validate. "Top-down" approaches infer fluxes from atmospheric CO2 concentration measurements, but these methods face challenges with atmospheric transport models, especially at mesoscale resolutions (tens to hundreds of kilometers). The ill-conditioning of inverse problems and the smoothing effect of transport at these scales make it difficult to pinpoint flux sources. This paper addresses this by investigating how to best utilize available mesoscale observational data (from towers and aircraft) to constrain CO2 flux inversions within a defined regional domain.
Methods
The core of the methodology involves a "top-down" inversion system. This system couples the non-hydrostatic atmospheric model Meso-NH with the Lagrangian Particle Dispersion Model (LPDM).
- Meso-NH: This model simulates atmospheric dynamics at high resolution (8 km and 2 km) to provide the necessary wind and turbulence fields.
- LPDM: This model simulates the backward transport of air parcels from measurement locations. By tracing these particles back in time, the model can determine the "footprint" or the region of the surface that influences a particular measurement. This is crucial for attributing observed CO2 concentrations to specific surface fluxes.
- Matrix Inversion Framework: The coupled model is integrated into a matrix inversion framework. This framework uses the LPDM's calculated "influence functions" (essentially a Jacobian matrix) to relate unknown surface fluxes and boundary conditions to the observed CO2 concentrations. The goal is to minimize a cost function that balances the agreement between model simulations and observations with prior knowledge of flux uncertainties.
The paper focuses on evaluating the reduction in error of CO2 flux estimates, rather than providing the final flux estimates themselves. This error reduction is a measure of how much the observations constrain the prior uncertainties.
Datasets
The study utilizes data from the CERES (CarboEurope Regional Experiment Strategy) campaign conducted in the South West of France during May-June 2005. Specifically, it uses:
- CO2 concentration measurements:
- From two instrumented towers (Marmande and Biscarosse), providing hourly data.
- From two aircraft (Piper-Aztec and ECO-Dimona), providing measurements at high frequency (1 Hz) along various flight paths.
- Meteorological data: Analyzed data from ECMWF (European Centre for Medium-Range Weather Forecasts) for initial and boundary conditions for the Meso-NH model.
The paper does not introduce new benchmark datasets but rather uses the CERES campaign data to evaluate the performance of their mesoscale inversion system.
Key Insights and Ideas
- Towers are King (for this scale): The study strongly suggests that for mesoscale inversions (around 8 km resolution), tower measurements are the primary drivers of error reduction for surface CO2 fluxes. Their continuous nature and proximity to the surface allow for a more direct constraint on local fluxes.
- Aircraft Data: A Valuable, but Different, Contribution: While aircraft measurements are valuable and contribute to error reduction, their impact is less dominant than towers. They are particularly useful for constraining boundary conditions and providing information over larger areas, especially when flying at higher altitudes.
- Nocturnal Data is Crucial (and Difficult): The paper highlights that nighttime CO2 concentrations at tower sites are critical for constraining fluxes. However, modeling these nocturnal accumulations and their associated errors is challenging, and improving this aspect could lead to significant gains in inversion performance.
- Boundary Conditions Matter Less Than Expected (for surface fluxes): For this specific mesoscale domain and the chosen observational strategy, imperfect knowledge of boundary conditions had a surprisingly limited impact on the error reduction of surface fluxes. This suggests that the surface fluxes themselves were the dominant source of uncertainty being addressed.
- Flight Strategy Optimization: The research explores how to optimize aircraft flight paths and altitudes to maximize their impact on flux inversions, suggesting that measurements within the boundary layer are most informative.
Key Results
- Significant Error Reduction: The combination of tower and aircraft measurements allowed for a significant reduction in uncertainties (error reduction) for 4-day averaged CO2 fluxes over about half of the 300x300 km² domain.
- Tower Dominance: Most of the error reduction came from the tower measurements, even though aircraft measurements had a noticeable impact.
- Aircraft Impact: Aircraft measurements improved surface flux error reduction by 15-20% compared to a tower-only scenario.
- Nocturnal vs. Diurnal: The study found that nighttime observations provided significant constraints on both diurnal and nocturnal fluxes, and that the error reduction for nocturnal fluxes was generally larger, despite the challenges in modeling them.
- Spatial Extent of Influence: The paper quantifies how the spatial footprint of influence varies with measurement type (tower vs. aircraft) and altitude, showing that tower influence is more localized, while aircraft can probe larger areas.
Limitations and Future Work
- Nocturnal Modeling Challenges: The paper acknowledges that modeling nocturnal CO2 concentrations at tower sites is a significant challenge, and improvements in this area are needed for better error reduction.
- Spatial and Temporal Correlations: The study notes that assumptions about spatial and temporal correlations of flux errors are important and require further investigation. The current assumption of uncorrelated prior flux errors is a simplification.
- Boundary Layer Turbulence: The authors suggest that a more sophisticated planetary boundary layer scheme is needed to accurately represent vertical mixing, which is crucial for mesoscale inversions.
- Tall Tower Impact: The experiment with a hypothetical tall tower indicated that while easier to model, they might be more influenced by large-scale transport, potentially making them less effective for constraining local fluxes in mesoscale inversions.
- Future Work: The authors plan to conduct actual flux inversions using validated data in a future paper. They also aim to investigate more complex atmospheric transport schemes and explore the impact of different prior error assumptions.
Summary updated