Land-Sea Warming Contrast Underestimated by Models Shapes Pacific IPO-Like Trends Since 1980s

The paper by Benjamin Johnson analyzes reanalysis data and CMIP-style simulations to demonstrate that the observed land-sea surface pressure gradient matches both the seasonal cycle and idealized land-heating experiments. Historical-forcing runs underestimate the land-ocean warming ratio by roughly 30-40 percent, while only the initial decades of abrupt 4xCO2 experiments recover the observed ratio and the associated oceanic high strengthening. This mismatch directly affects simulated Pacific trade-wind and SST trends. The discrepancy is not random but systematic across ensembles, pointing to a missing process in land-surface or boundary-layer schemes rather than forcing errors. Because the land-sea contrast sets the amplitude of the IPO-like response, resolving it would tighten constraints on near-term Pacific variability and global sensitivity. The authors note that persistence of the bias into future projections depends on whether the root cause is transient or structural. Targeted diagnostics separating land evapotranspiration, aerosol effects, and stomatal conductance changes could isolate the responsible parameterization within the next CMIP cycle.