The spectrum of Asian Monsoon variability: A proxy system model approach to the hydroclimate scaling mismatch

Recent studies have found a significant mismatch between the amount of low-frequency variability in paleo-hydroclimate records and that simulated by climate models. However, uncertainties in how proxy systems record climate are significant, and the influence of low-frequency non-climate noise in paleoclimate archives is difficult to constrain. We use proxy system models to compare a network of mostly δ¹⁸O tree cellulose records from Monsoon Asia with isotope-enabled climate model output from (1) a coupled fully-forced version of iCESM (850–2005 CE), (2) isoGSM nudged toward the 20th Century Reanalysis (1871–2010 CE), and (3) an atmosphere-only run of iCAM5 with prescribed sea-surface temperatures from observations (1850–2014 CE). We identify a large discrepancy in scaling behavior between hydroclimate proxies and analogous iCESM pseudoproxies. In contrast, pseudoproxy networks based on climate models constrained by observations (isoGSM and iCAM5) have variance that continues to increase from decadal to multidecadal scales. This indicates that the mismatch arises from a failure of iCESM to simulate a sufficient amount of low-frequency hydroclimate variability. We test the influence of non-climate noise and photosynthate storage and find that they play a small role in elevating the low-frequency signal of proxies, but cannot explain the mismatch. Although our results are constrained to Monsoon Asia, the scaling of hydroclimate proxies across other regions and models have shown similar behavior, suggesting a proxy-model mismatch on a global scale that likely stems from an underestimation of low-frequency variability in fully coupled climate models.