Plant radiocarbon across an urban-rural CO2 gradient matches surface and column CO2 observations

Atmospheric measurements are needed to verify progress in reducing fossil fuel carbon dioxide (ffCO2) emissions, especially in cities where most ffCO2 is emitted. However, because of complexities in atmospheric dynamics and large natural CO2 fluxes, ambient CO2 measurements alone cannot distinguish ffCO2 signals. Analysis of the radiocarbon (14C) content of urban plants can reveal ffCO2 patterns and is more cost-effective than air 14CO2 sampling, but its use has been limited because of uncertainty in the temporal integration period and because it has not been quantitatively evaluated against other atmospheric monitoring approaches. We analyzed the 14C content of turfgrasses collected along an urban to rural gradient in Southern California. The primary study sites are Downtown Los Angeles, Pasadena, Irvine, Riverside, and Beaumont, CA. We compare these plant 14C to multiscale, atmospheric CO2 measurements including surface CO2 and total column CO2 (XCO2). Surface CO2 was measured using cavity ringdown spectrometers, and XCO2 was measured using a portable Fourier Transform solar spectrometer. Our data shows that plant 14C is highly sensitive to local ffCO2 emissions at the intra-city scale and captures pronounced differences between urban to rural sites. We observe significant correlations between fossil fuel enhancements (Cff) derived from plant 14C and from atmospheric CO2. Our findings indicate that plant 14C analysis can be a very useful tool for quantifying ffCO2 trends in cities that lack surface CO2 and XCO2 measurement infrastructure.

Methods
Measurements of plant 14C, XCO2, and surface CO~2 ~were collected approximately every three months between May 2022 to March 2023. The study focused on five sites along an urban to rural gradient in the Greater Los Angeles area: Downtown Los Angeles, Pasadena, Irvine, Riverside, and Beaumont, CA. At each site, a series of turfgrass samples were collected on irrigated lawns, mostly on university campuses or public parks. A total of 147 plant samples were analyzed for 14C content at the University of California's Keck Carbon Cycle AMS facility.

We also collected ground-based XCO2 measurements using a portable Fourier Transform Spectrometer ("EM27/SUN", Bruker). XCO2 was measured at each site for two days per seasonal campaign during sunny conditions. The XCO2 data was calibrated based on side-by-side measurements with the Total Carbon Column Observing Network (TCCON) site at Caltech, Pasadena. TCCON data is publicly available at https://tccondata.org/.

Surface CO2 data for Downtown Los Angeles, Pasadena, and Irvine was provided by the Los Angeles Megacities (LAM) Carbon Project, which is a network of rooftop or tower sites managed by the National Institute of Standards and Technology (NIST). The surface CO2 data from the LAM network is publicly available at https://data.nist.gov/od/id/mds2-2388 (Verhulst et al., 2017; Kim et al., 2024). Surface CO2 data collected in Riverside and Beaumont is not managed by NIST and is publicated here. At Beaumont, we measured surface CO2 at a public park using a Picarro G2401 cavity ringdown spectrometer. Riverside surface CO2 measurements were collected on a rooftop at the University of California, Riverside. 

In addition to the five study sites along the urban-rural gradient, our analysis also relied on measurements at "background" sites. To estimate contributions from fossil fuel-derived CO2 emissions, we used 14C, surface CO2, and XCO2 data at several sites to quantify the atmospheric background. For 14C, the background estimates were based on monthly-integrated air samples collected at Utqiagvik, AK, a remote location far from fossil fuel emission sources (Xu, pers. comm 2023). For surface CO2, measurements at San Clemente Island and Victorville, CA were used to characterize the background (Verhulst et al., 2017; Kim et al., 2024). For XCO2, measurements at the NASA Armstrong Flight Research Center, a remote desert site near Edwards, CA, were used to estimate background. This site is part of the TCCON network and the data is available at https://tccondata.org/.