Tier 2 Clausius–Clapeyron Test CMIP6

1 figure · 1990–2014 → 2040–2049

CMIP6 Multi-Model Mean Context

Comparison with CMIP6 conventional-resolution ensemble mean (up to 8 models under SSP3-7.0, regridded to 0.25°).

Synthesis

IFS-NEMO and IFS-FESOM exhibit hydrological sensitivities (4.1–4.5 %/K) that exceed the upper bound of the CMIP6 envelope, indicating a stronger water vapor feedback closer to theoretical Clausius-Clapeyron scaling.
The DestinE coupled models (IFS-FESOM and IFS-NEMO) exhibit a notably stronger hydrological sensitivity than the standard-resolution CMIP6 ensemble under SSP3-7.0 forcing. Both high-resolution models produce local fractional Total Column Water Vapor (TCWV) scaling rates—4.1 %/K and 4.5 %/K, respectively—that exceed the upper bound of the CMIP6 P5–P95 envelope (2.3–3.9 %/K) and are significantly steeper than the multi-model median (3.3 %/K). This places the high-resolution response entirely outside the range of conventional model behavior for this metric, suggesting a distinct thermodynamic adjustment in the eddy-rich simulations. Physically, while both models remain below the theoretical Clausius-Clapeyron limit of ~7 %/K due to continental moisture constraints, their positioning indicates a closer adherence to theoretical expectations than coarser models. The analysis suggests that the high-resolution configurations may maintain relative humidity more effectively or simulate stronger air-sea moisture fluxes, particularly in the dense cluster of ocean-dominated points at lower warming levels (<2 K). While the response flattens at higher warming levels consistent with land-based limitations, the aggregate behavior represents a 'moister' warming trajectory. This deviation from the CMIP6 envelope likely reflects added value from improved process resolution rather than error. The elevated sensitivity implies that for every degree of warming, DestinE models predict a larger increase in precipitable water than their lower-resolution counterparts. This has direct implications for precipitation intensity and the global energy budget, potentially driven by better-resolved mesoscale convective processes or ocean-atmosphere coupling that sustains higher evaporation rates.

Related diagnostics

pr_sensitivity global_energy_budget relative_humidity_feedback

Clausius–Clapeyron Test

Clausius–Clapeyron Test
Variables avg_tcwv, avg_2t
Models ifs-fesom, ifs-nemo
Units kg/m2
Baseline 1990-2014
Future 2040-2049
Method δ_TCWV = ln(TCWV_future / TCWV_hist); ΔT = T_future − T_hist. Sub-sampled to 50k cells.

Summary high

This diagnostic evaluates the hydrological sensitivity of IFS-FESOM and IFS-NEMO by correlating fractional Total Column Water Vapor (TCWV) change with surface warming (ΔT), comparing the results against theoretical Clausius-Clapeyron (C-C) scaling and the CMIP6 ensemble. Both DestinE models demonstrate a moisture scaling response (4.1%/K and 4.5%/K) that is significantly stronger than the CMIP6 median (3.3%/K) and exceeds the upper bound of the CMIP6 envelope.

Key Findings

  • IFS-NEMO (4.5%/K) and IFS-FESOM (4.1%/K) exhibit moisture scaling rates notably higher than the CMIP6 multi-model median (3.3%/K) and outside the CMIP6 P5–P95 range (2.3–3.9%/K).
  • Both models remain below the theoretical pure C-C rate of 7%/K, consistent with global constraints (e.g., land limited evaporation), but are closer to the theoretical limit than the standard resolution CMIP6 ensemble.
  • IFS-NEMO shows a slightly steeper slope and wider scatter than IFS-FESOM, indicating a stronger hydrological sensitivity in this configuration.

Spatial Patterns

The scatter plots reveal two distinct regimes: a dense cluster at lower warming levels (0–2 K, likely dominated by oceans) which follows a steeper slope close to the C-C line, and a more diffuse distribution at higher warming levels (>3 K, likely land/high-latitudes) where the moisture response flattens, depressing the global regression slope.

Model Agreement

The two DestinE models qualitatively agree on a stronger-than-CMIP6 water vapor feedback. The divergence from CMIP6 suggests that the high-resolution coupled systems may maintain relative humidity more effectively or simulate stronger air-sea moisture fluxes, resulting in a 'moister' warming trajectory than coarser models.

Physical Interpretation

While the Clausius-Clapeyron relation predicts a ~7%/K increase in water vapor holding capacity, realized scaling is often lower due to relative humidity reductions (drying) over land and dynamic constraints. The fact that DestinE models exhibit slopes (4.1–4.5%/K) closer to the 7%/K limit than CMIP6 (3.3%/K) implies they simulate less reduction in relative humidity or stronger evaporation feedbacks, potentially due to better resolved mesoscale convective processes or ocean-atmosphere coupling.

Caveats

  • The linear regression aggregates all grid points globally, conflating different physical regimes (e.g., moisture-limited land vs. energy-limited ocean).
  • The CMIP6 reference slope (3.3%/K) is lower than typical global-mean scaling estimates (~6–7%/K), indicating that the local cell-by-cell regression method used here yields different sensitivities than analyzing global means.
  • Internal variability over the short 10-year analysis period may introduce noise into the local ΔT vs ΔTCWV relationship.