COS 105-3 - Effects of elevated atmospheric carbon dioxide on soil organic carbon in the Mojave Desert: A critical comparison of multiple analyses

Background/Question/Methods

Soils are a major component of the global carbon cycle, yet our understanding of how soil carbon stocks respond to global change remains limited. Determining the effects of elevated atmospheric CO₂ (eCO₂) on soil organic carbon (SOC) is critical to predicting changes in the global carbon cycle. Arid ecosystems have large spatial coverage globally and are known to respond to eCO₂, implying that arid soils may be important to future carbon cycling. However, SOC measurements in arid soils are challenging because of low overall organic carbon content, high spatial heterogeneity of vegetation, and the presence of inorganic carbon. As an example, the intensively studied Mojave Free-Air CO₂ Enrichment (FACE) site has reported conflicting results regarding the effects of eCO₂ on SOC. Here, we present measurements of SOC stocks from this site and compare it to previously published data from Mojave FACE in order to assess conflicting results. We attempt to interpret these opposing treatment effects by comparing statistical models and inorganic carbon removal methods used in each study.

Results/Conclusions

We found that the elevated CO₂ treatment resulted in a loss of -541.45 g C•m^-2 in SOC stocks under Larrea tridentata, the major vegetation cover type in this ecosystem. There was no CO₂ treatment effect any other vegetation type, though, resulting in no change to SOC stocks at the ecosystem level. This contrasts with previously published analysis showing a 20% increase in ecosystem SOC stocks. Our results show that statistical model design impacts estimated SOC stocks but does not change the CO₂ treatment effect. We further found that removing inorganic C using direct application of liquid H₃PO₄ can underestimate SOC in these soils by 20-30%. Both methodological decisions may have consequential effects for Earth System Models and meta-analyses that use data from the Mojave Desert FACE experiment. This study comparison highlights the importance of understanding the ecology of an ecosystem when designing statistical models and the necessity of using consistent inorganic carbon removal methods across studies. Although these factors can impact ecosystem SOC estimates, neither explain the conflicting ecosystem level treatment effects from multiple studies at the Mojave Desert FACE. Resolving this discrepancy is critical to our understanding of C cycling in global drylands under future climate change scenarios.