Land Surface water controls on Atmospheric CO2 growth

By: Prof. Keith Haines and Samantha Petch

Atmospheric CO2 levels are rising every year, primarily due to human activities, Fig 1. However, the rate of this increase does vary significantly from year to year, not because of human emissions, but mostly due to changes in how much CO2 is absorbed by land ecosystems. 

The land’s ability to absorb CO2 is determined by the balance between photosynthesis and respiration and is the least understood part of the global carbon cycle. This land carbon sink plays a crucial role in offsetting anthropogenic CO2 emissions, accounting for approximately 30 ±10% of these emissions each year (Friedlingstein et al., 2020). The controls on this land surface carbon sink are still poorly understood and would have a strong impact on future CO2 growth rates (CGR) if they were to change significantly.  There are notable increases in CGR during El Niño events and decreases during La Niña events (Keeling & Revelle, 1985) and there is also a widespread consensus that variations in tropical ecosystems exert the most significant influence on global CGR.  

Tropical temperatures have a strong positive correlation with CGR (Wang et al., 2013) and given that tropical temperatures are usually considered optimal for photosynthesis, any elevation in these temperatures under global warming could amplify the CGR. However it has recently been realised that water availability is also playing a really important role, see Fig 2.   

Figure 2 shows detrended interannual variations in the atmospheric CGR in black, and the total terrestrial water storage (TWS on a reversed axis) in blue, as measured from the GRACE satellite which detects water through mass redistributions over the Earth’s surface, (from Petch et al. 2024 following Humphreys et al. 2018).

Less water over land e.g. during the El Niño in 2015/6, goes with more rapid increases in atmospheric CO2 as the land surface absorption is reduced.  This range of uptake (up to 4GtC/yr) is really large and to understand this relationship better we really need to regionalise the surface response, see Fig 3.   

Figure 3 shows the regional interannual correlations in GRACE surface water storage and the global CGR (Petch et al. 2024).

Petch et al. (2024) found that the tropics alone can explain the entire global TWS-CGR relationship seen in Fig 2, with only minor offsetting contributions from the northern and southern hemisphere extratropics. Additionally, tropical forests, e.g. over the Amazon, were found to be the dominant land cover type governing this relationship, despite occupying only a relatively small portion of the land surface. While the negative correlations in Fig 3 (and other evidence) point to the importance of the tropics in the global relation seen in Fig 2, it would be important to have direct regional surface CO2 flux data to back this up.   

Having a verifiable way of comparing surface CO2 fluxes is of course of the highest scientific and political importance. Currently national anthropogenic CO2 emissions all rely on national inventories which are hard to verify or compare. CO2 in the atmosphere gets mixed relatively quickly, at least at the hemispheric level, and so only frequent reliable satellite based atmospheric measurements stand a chance of achieving this, e.g. ESA’s CO2M-(A,B,C) multi-satellite mission, launching 2025/6. In the meantime, less frequent satellite measurements e.g. GOSAT (2009 onwards) can be assimilated into atmospheric transport models which may then be used to infer surface sources and sinks which best match observed atmospheric CO2 concentrations. Given the importance of these estimates many teams produce such surface CO2 flux products, although many still rely on in situ measurements alone and show a lot of disagreement between regional results (Fig 4).

Figure 4 shows % contributions to global land carbon flux interannual variability over 2001-2023 from four atmospheric CO2 inverse model products. Note, the NISMON-CO2 map is depicted over a larger range due to higher contributions from tropical forests. While these products agree well with the global CGR from Fig 2, the regional agreements can be seen to be fairly poor, Petch et al. (2024).

We believe there is potential to bring in additional satellite-based information from the land surface such as the GRACE water data shown earlier, and perhaps other datasets such as ESA’s Climate Change Initiative Land surface temperature measurements (https://climate.esa.int/en/projects/land-surface-temperature/), to provide additional constraints to these atmospheric CO2 transport models and bring greater confidence to space-based estimates of regional CO2 fluxes.  This would help both improve understanding of ecosystem responses to climatic changes as well as provide better verification tools for monitoring anthropogenic emissions in a more verifiable way.  

References 

Friedlingstein, P., O’Sullivan, M., Jones, M. W., Andrew, R. M., Hauck, J., Olsen,  A., . . . Zaehle, S. (2020). Global carbon budget 2020. Earth Syst. Sci. Data,  12 , 3269–3340. https://doi.org/10.5194/essd-12-3269-2020 

Humphrey, V., Zscheischler, J., Ciais, P., Gudmundsson, L., Sitch, S., & Seneviratne, S. I. (2018). Sensitivity of atmospheric co2 growth rate to observed changes in terrestrial water storage. Nature, 560 , 628–631. https://doi.org/10.1038/s41586-018-0424-4 

Keeling, C. D., & Revelle, R. (1985). Effects of el nino/southern oscillation on the atmospheric content of carbon dioxide. Meteoritics, 20 , 437-450. https://doi.org/10.1093/nsr/nwab150 

Petch, S., Feng, L., Palmer, P. King, R. P., Quaife, T. & Haines, K.  (2024) Strong relation between atmospheric CO2 growth rate and terrestrial water storage in tropical forests on interannual timescales. Submitted to Glob. Biogeochem. Cyc. Also ESS Open Archive . https://doi.org/10.22541/essoar.172132358.83519012/v1  

Wang, W., Ciais, P., Nemani, R. R., Canadell, J. G., Piao, S., Sitch, S., . . ., Myneni, R. B. (2013). Variations in atmospheric CO2 growth rates coupled  with tropical temperature. Proc. Natl Acad. Sci., 110 , 13061–13066. https://doi.org/10.1073/pnas.1219683110