Tropical free-tropospheric humidity differences and their effect on the clear-sky radiation budget in global strom-resolving models

Abstract

Free-tropospheric water vapor strongly impacts the Earth's outgoing longwave radiation (OLR) and therefore plays a key role in controlling the clear-sky response of the climate system to an increase in greenhouse gases. It is now widely accepted that this response is described by a warming and moistening of the atmosphere that is implied if the relative humidity (RH) and lapse rate were to depend on temperature alone, which corresponds to a warming at approximately constant RH (e.g., Held & Soden, 2000;Po-Chedley et al., 2019;Romps, 2014). This reduces the radiative response compared to a warming at constant absolute humidity, and can be described as a positive water-vapor-lapse-rate feedback. While general circulation models (GCMs) agree on this basic response (e.g., Bony et al., 2006;Soden & Held, 2006), there is still an appreciable inter-model spread in the magnitude of the water-vapor-lapse-rate feedback. This spread, which primarily originates from the tropics, contributes a non-negligible (about 30%) uncertainty to the climate sensitivity (Vial et al., 2013).

How model uncertainties influence tropical humidity in global storm-resolving simulations

“…In the following we focus on R differences in the mid troposphere (4 km to 8 km, indicated by the gray lines in Figure 1). Although mid-tropospheric R differences are, similar as in the DYAMOND ensemble, not particularly large, Lang et al (2021) showed that R differences in this region are particularly important for differences in OLR.…”

“…Comparably large R changes occur in the TTE and 2-mom experiments, in which the parameterizations of turbulence and microphysics were changed. The largest changes occur in the lower and mid troposphere, where they have a larger impact on the clearsky OLR than those in the upper troposphere (Lang et al, 2021). Changing to the TTE turbulence scheme results in a strong increase in R of up to 8% over a broad altitude layer between 2 km to 6 km.…”

“…First it is worth noting that the R spread produced by our experiments is similar to the inter-model spread in the DYAMOND ensemble (Figure 1c). Based on the DYA-MOND ensemble, Lang et al (2021) showed that the R spread across GSRMs is reduced compared to classical GCMs. This is possibly related to the omission of convective parameterisations, which represent a major source of uncertainty in GCMs.…”

“…To obtain a better understanding of the physical mechanisms behind the humidity changes produced in our experiments we use a last-saturation framework based on back-trajectories. For this analysis we focus on the altitude region between 4 km and 8 km, where R differences in the DYAMOND ensemble were shown to have a comparably large effect on the clear-sky radiation budget (Lang et al, 2021). A main goal is to understand to what extent the changes in R are explained by changes in the properties of the source regions of air parcels, i.e.…”

“…At present, due to the high computational effort, storm-resolving simulations are limited to time scales of days to months. A first intercomparison of GSRMs, the DYnamics of the Atmospheric general circulation Modeled On Non-hydrostatic Domains (DYAMOND; project, indicates that the inter-model spread in tropical free-tropospheric humidity is indeed reduced compared to GCMs (Lang et al, 2021). While this is a promising result and highlights the benefit of even approximately resolving deep convection, the study also showed that the remaining differences in relative humidity are still an important source of uncertainty for the clear-sky outgoing longwave radiation (OLR).…”