Mean diurnal cycle of low, mid-level, and high clouds (gray shading) from Meteosat, a CRM, and a CPM for a summer period. Also shown is the mean diurnal cycle of observed and modeled surface precipitation. In contrast to the CPM, the CRM improves the timing of rainfall and reduces the too large fraction of high clouds.
The larger computational costs of convection-permitting simulations (CRMs) can only be justified for regional climate modeling if a significant added value with respect to coarse-grid parameterized simulations (CPMs) is identified. A process-level evaluation is conducted in this paper to compare CRMs, CPMs, and observations. Here, we are particularly interested in the interaction between thermally-driven mountain flows (which set in in the morning) and orographic deep convection (which often follows in the afternoon).
Why do we care?
Thermally driven flows at a variety of scales are undoubtly of relevance to the initiation of deep convection over topography. The latter in turn is key to the hydrological cycle during persistent high-pressure episodes in summer when rainfall is limited to topographic regions. To inform about future water resources, regional climate models thus need to replicate the physics behind the interaction of thermally driven flows and convection.
How do we approach this?
A 18-day high-pressure period with daily re-occuring deep orographic convection is studied. CRMs with 1.1 km and 2.2 km horizontal grid spacing and a CPM with 7.7 km grid spacing are evaluated against observations provided by MeteoSat-8 (cloud cover and height), station measurements across the Alps (valley winds and pressure gradient), and gridded precipitation in Switzerland. Mean diurnal cycles are evaluated.
What do we find?
Our main findings from this comparison are that the diurnal evolution of convective rainfall and cloud cover are significantly improved in CRMs, since the parameterization scheme for deep convection is switched off and convective instabilities are solved explicitly based on first principles. More importantly, however, we show that an added value is reflected also in an improved physical consistency at a process-level. We demonstrate that the observed link between the mass convergence by the large-scale plain-mountain flow during the morning, ensuing valley flows, and deep orographic convection during the afternoon is only captured by the CRMs.