Below are short summaries of selected key publications from my past research.


Whether it rains more over wet or over dry soils is set by the representation of convection

Whether an increase in soil moisture leads to an increase (positive feedback) or a decrease (negative feedback) of precipitation has long been debated. Positive feedbacks are of particular relevance for the climate system as they can amplify an existing perturbation, thus affecting the climate system on longer time scales. One difficulty in diagnosing the strength and sign of the feedback between soil moisture and convection is that it involves convection, which, in traditional climate model, is not explicitly resolved but parameterized. There is no guarantee that a convective parameterization can faithfully capture this feedback. In this study, we for the first time investigated the soil moisture-precipitation feedback in regional climate simulations using an explicit representation of convection and compared the results to similar integrations conducted at coarser resolution with parameterized convection. Whereas the simulations with convective parameterizations revealed a positive feedback, in agreement with past studies, the simulations with explicit convection revealed a negative feedback. The negative feedback was the result of a strong capping inversion at the top of the planetary boundary layer that inhibits the development of convection over wet soils doe to missing boundary layer growth. Moreover, using different versions of the convective parameterization, either positive, neutral or negative feedbacks could be obtained, depending whether the convective parameterization was able to see the capping inversion. Even though biases in the representation of the soil moisture-precipitation feedback in the explicit simulation cannot be excluded, in any case, the feedback sign should not depend on the design of the convective parameterization and on the chosen representation of convection.

Hohenegger C., P. Brockhaus, C. S. Bretherton, and C. Schär, 2009: The soil-moisture precipitation feedback in simulations with explicit and parameterized convection. J. Climate22, 5003-5020, doi:10.1175/2009JCLI2604.1


Congestus moistening is too inefficient to explain the daily transition to deep convection

As the buoyant convective plume rises in the atmosphere, it mixes with its environment, leading to the detrainment of updraft air in its environment.  This process moistens the environment, making it less hostile to the ascent of the next convective plume, a process called preconditioning. But other processes, in particular moisture-convergence and forced vertical ascent resulting from larger scale forcing, can also promote the transition to deep convection. In this study, we investigated the importance of  moistening by previous congestus clouds, a thermodynamic control on the transition to deep convection, versus moisture convergence, a dynamical control on this transition. We used the idea that only fast enough processes can be of relevance for the transition as it has to happen within one day. Comparing theoretical estimates of transition time due to congestus moistening or moisture convergence to observed transition times revealed that congestus moistening, with a theoretical transition time scale of  35-38 h, is too slow to explain observed transition times (2-4 h). The results are confirmed by estimates derived from idealized large-eddy simulations.  Further evidence speaking against the efficiency of preconditioning by congestus clouds were the fact that (i) most congestus clouds do not transition to deep convection actually; that (ii) the probability of transition does not increase with their lifetime and that (iii) the presence of cumuli congestus over a region does not enhance the likelihood for deep convection development. The results indicate that previous work might have overemphasized the idea of convection randomly developing in a moist and unstable atmosphere, on its own, through moistening by previous shallower clouds.

Hohenegger C. and B. Stevens, 2013: Preconditioning deep convection with cumulus congestus. J. Atmos. Sci., 70, 448-464 doi:10.1175/JAS-D-12-089.1.


The ability of the soil to dry out ensures a homogeneous precipitation distribution over land

Imagine the simplest planet with only three fundamental processes at play: radiation, that acts to destabilize the atmosphere, convection, that acts against it and stabilizes the atmosphere, and soil moisture. How would precipitation be spatially distributed on such a planet? Past studies that have used this conceptualization of radiative convective equilibrium (RCE) over a surface of fixed SST have found that precipitation ends up strongly organized in a blob. Does this remain true over land? A priori, two distinct final precipitation distributions are possible.  Soil moisture may act to keep precipitation to the precipitating region through a positive thermodynamical soil moisture-precipitation feedback, or to homogenize the precipitation distribution through a negative dynamical feedback requiring the generation of strong enough soil moisture-induced circulations. We found that, at least when the soil dries out, the disaggregating effect of soil moisture wins. This result could be explained theoretically combining a bulk model of the planetary boundary layer, the linearized version of the surface energy budget equation and gravity current theory. The results are surprising as they indicate that, on its own, the drying out of the soil terminates heat waves and acts against the formation of deserts. We nevertheless showed that the results are consistent with observations of the propagation of the monsoon.

Hohenegger C. and B. Stevens, 2018: The role of the permanent wilting point in controlling the spatial distribution of precipitation. Proc. Natl. Acad. Sci., 115, 5692-5697, doi:10.1073/pnas.1718842115.


Many basic climate statistics already converge at relatively coarse resolution

Convective processes happen on scales  that are too fine to be explicitly resolved by state-of-the-art General Circulation Models (GCMs) with their typical grid spacing of O(100 km). But how fine is fine enough to capture basic climate statistics? In this study, we made use of a unique dataset of eight global storm-resolving models integrated for 40 days, called DYAMOND, and of simulations conducted with the model ICON using an explicit representation of convection across resolution, to answer this question. As the determination of the convergence is actually ill-posed for the climate system, we proposed a novel measure of convergence. We quantified the convergence by comparing resolution-induced differences to the spread obtained in the DYAMOND ensemble. Looking at large-scale features of the atmosphere, such as location of the ITCZ, precipitation amount, energy budget, we found that the resolution dependency is much smaller than initially thought. A grid spacing of 5 km already allows capturing 26 out of the 27 investigated statistics. Resolution down to 5 km especially matters for radiation, which systematically increases with resolution because of reductions in the low cloud amount over the subtropical oceans. Even for nine of the investigated statistics, including the global mean precipitation amount, the latitudinal position and  the width of the Atlantic ITCZ or the longitudinal position of the Pacific ITCZ, a grid spacing of 80 km does not lead to significant differences. On the one hand and given the cost of high-resolution simulations, this is good news for the prospect of conducting storm-resolving simulations on climatic time scales. On the other hand, the results do question the importance of the small scales for determining large-scale features of the atmosphere.

Hohenegger, C., L. Kornblueh, D. Klocke, T. Becker, G. Cioni, J. F. Engels, U. Schulzweida and B. Stevens, 2020: Climate statistics in global simulations of the atmosphere, from 80 to 2.5 km grid spacing. J. Meteorol. Society Japan, 98, 73-91, doi:10.2151/jmsj.2020-005.