TL;DR
This study leverages a novel laser imaging system (GreenLITE™) and advanced atmospheric modeling to analyze CO2 concentrations in Paris. It reveals that while models can capture suburban trends, they struggle with the complex urban environment, particularly in winter, highlighting the need for improved urban canopy parameterizations and a better understanding of vertical mixing.
What Problem Does This Paper Address?
Accurately quantifying CO2 emissions from urban areas is crucial for climate change mitigation. However, the complex interplay of emissions, urban morphology, and atmospheric transport makes it challenging to model and measure CO2 concentrations within cities. Traditional ground-based monitoring networks are often too sparse to capture the fine-scale spatial and temporal variability. This paper addresses this by using the GreenLITE™ system, which provides integrated measurements along multiple horizontal paths across Paris, to evaluate the performance of a high-resolution atmospheric model (WRF-Chem) with different urban canopy schemes. The goal is to improve our understanding of urban CO2 dynamics and support the development of city-scale CO2 emission inversion systems.
Methods
The study employs a multi-faceted approach:
- GreenLITE™ Laser Imaging: This system uses continuously operating laser transceivers and retroreflectors to measure CO2 concentrations along 30 horizontal "chords" (2.3-5.2 km long) covering a 25 km² area of central Paris. This provides a unique, spatially integrated view of CO2 distribution.
- In Situ Measurements: Six conventional ground-based CO2 monitoring stations within Paris and its surroundings are used for comparison and validation.
- WRF-Chem Model: A high-resolution (1 km) atmospheric model coupled with a chemistry transport model is used to simulate CO2 concentrations.
- Urban Canopy Schemes: Two different parameterizations for representing urban surface processes are tested within WRF-Chem: the Urban Canopy Model (UCM) and the Building Effect Parameterization (BEP). This allows for an assessment of how different representations of urban physics impact CO2 modeling.
- Bias Correction: A specific method was developed to correct for potential drifts in the GreenLITE™ measurements, ensuring better comparability with in situ data and model outputs.
Datasets
- GreenLITE™ Data: Near-surface atmospheric CO2 concentrations measured along 30 horizontal chords in Paris from December 2015 to November 2016.
- In Situ CO2 Data: Hourly CO2 measurements from six monitoring stations in the Île-de-France region.
- WRF-Chem Model Outputs: Simulated CO2 concentrations at 1 km horizontal resolution.
- Anthropogenic Emissions Inventories: AIRPARIF inventory for the Île-de-France region and the IER inventory for outside the region, used as input for the WRF-Chem model.
- Meteorological Data: ERA-Interim reanalysis data for boundary conditions and surface weather station data for assimilation.
- Biogenic Flux Data: Vegetation Photosynthesis and Respiration Model (VPRM) outputs for biogenic CO2 fluxes.
- Global CO2 Inversion Product: CAMS global CO2 atmospheric inversion product for initial and lateral boundary conditions.
Key Insights and Ideas
- Urban Canopy Matters: The choice of urban canopy scheme significantly impacts the model's ability to reproduce observed CO2 concentrations, especially within the city. The BEP scheme generally performed better than UCM.
- Winter vs. Summer Dynamics: CO2 concentrations exhibit distinct behaviors in winter and summer. Winter's low vertical mixing leads to higher CO2 concentrations and larger spatial gradients, making it more sensitive to emission and transport modeling. Summer's increased mixing and convection tend to smooth out these gradients.
- GreenLITE™'s Value: The GreenLITE™ system, by providing integrated measurements over larger areas, is less susceptible to local emission uncertainties than point measurements and is thus better suited for evaluating kilometer-scale models.
- Model Limitations: While WRF-Chem can reproduce seasonal cycles and suburban trends, it struggles with accurately capturing the fine-scale spatial and temporal variations of CO2 within the complex urban fabric of Paris, particularly concerning vertical mixing.
- Summer Discrepancies: During summer, GreenLITE™ observed larger horizontal CO2 differences than both in situ measurements and model results, suggesting potential issues with either the measurements themselves (noise/bias) or the model's representation of summer atmospheric processes.
Key Results
- Both UCM and BEP urban canopy schemes in WRF-Chem were capable of reproducing seasonal cycles and most synoptic variations in CO2 over suburban areas.
- During winter, the BEP scheme provided a significantly better description of temporal variations and horizontal CO2 differences within Paris compared to the UCM scheme.
- In summer, GreenLITE™ measurements indicated larger CO2 horizontal differences than both in situ measurements and model results, with systematic east-west variations.
- The models generally reproduced cross-city upwind-downwind differences in CO2 concentrations well, with BEP showing better agreement than UCM.
- However, both models struggled to reproduce the observed CDS-JUS CO2 difference, suggesting limitations in representing urban emissions and transport at this scale.
Limitations and Future Work
- Summer GreenLITE™ Issues: The paper notes potential increases in measurement noise and bias for some GreenLITE™ chords during the summer, which needs to be resolved for future assimilation into inversion systems.
- Model Resolution: The 1 km resolution of WRF-Chem may be insufficient to capture the fine-scale (sub-kilometer) CO2 concentration features within Paris, especially those influenced by complex urban morphology and street canyons.
- Emission Inventory Accuracy: The accuracy of the emission inventories, particularly their temporal profiles and spatial distribution within the city, remains a potential source of model-measurement discrepancies.
- Vertical Mixing Uncertainty: The study highlights that uncertainty in modeling vertical mixing during summer is likely larger than uncertainty in emissions, posing a significant challenge for constraining emission inventories using urban CO2 measurements.
Future work should focus on improving urban canopy parameterizations, better representing vertical mixing processes in atmospheric models, and addressing potential biases in the GreenLITE™ measurements, especially during summer. Further analysis of the spatial distribution of emissions and their impact on observed CO2 gradients is also warranted.
Summary updated