Are general circulation models representing processes controlling tropical and subtropical free tropospheric relative humidity properly? The added value of water vapor isotope measurements

Evaluating the representation of processes controlling tropical and subtropical tropospheric relative humidity (RH) in atmospheric general circulation models (GCMs) is crucial to assess the credibility of predicted climate changes. The goal of this study is to design diagnostics to detect and understand biases in the representation of these processes. We use water vapor isotope measurements from space and ground (4 satellite and 4 ground-based remote-sensing datasets, in-situ surface vapor sampling) to evaluate the 3D distribution of the tropospheric water vapor isotope composition simulated by several versions of the isotope-enabled GCM LMDZ. Excessive diffusion during water vapor transport leads to a moist bias and a reversed seasonality of the isotope composition throughout the free troposphere. On the other hand, subgrid scale variability of water vapor and precipitation efficiency in clouds significantly affects the RH without affecting the isotope seasonality much. Inter-comparing 7 isotope-enabled GCMs suggests that excessive diffusion during water vapor transport is a common cause of moist bias in GCMs. This study shows that water vapor transport and parameterization of microphysical and macrophysical cloud processes are major sources of uncertainty in simulating processes controlling tropospheric RH, and demonstrates the added value of water vapor isotope measurements to detect and understand misrepresentations of these processes.