lmd_Grandpeix2004_abstracts.html

2004 .

R. Tailleux and J. Y. Grandpeix. On the seemingly incompatible parcel and globally integrated views of the energetics of triggered atmospheric deep convection over land. Quarterly Journal of the Royal Meteorological Society, 130:3223-3243, October 2004. [ bib | DOI | ADS link ]

The energetics of the diurnal cycle of atmospheric deep convection over land remain difficult to understand and simulate accurately with current cumulus parametrizations. Furthermore, a proper formulation has remained elusive owing to seeming incompatibilities between, on the one hand, the parcel view of energetics which relies on such concepts as Convective Available Potential Energy (CAPE) and Convective Inhibition (CIN), and, on the other hand, the globally integrated view, which relies on such concepts as Moist Available Energy (MAE), reference states, and energy conversion terms. While the MAE is intuitively the global counterpart of the parcel-defined CAPE, there seems to be no global analogue to the parcel-defined concept of energy barrier attached to CIN. To gain insights into this issue, a new cost function PE is introduced to quantify the amount of positive or negative energy required for a given sounding to undergo an arbitrary adiabatic rearrangement of mass, and which encompasses both the parcel-defined and global energy concepts as particular cases. The function PE offers a complementary view of the stability and energy characteristics of atmospheric soundings, whose local minima are naturally associated with the reference states of the system. It is established that: (a) MAE is essentially equivalent to CAPE multiplied by a mass conversion factor Mb which scales as the amount of convectively unstable boundary-layer air. Using the available summer 1997 IOP data from the ARM- SGP site, Mb is found to correlate with CAPE, which suggests the existence of a functional relationship between CAPE and MAE; if further confirmed, this result would considerably simplify the computation of MAE. (b) A global counterpart to the parcel-defined concept of energy barrier can only be defined if the system considered admits several reference states, and not one as is classically assumed. In that case, energy barriers naturally arise as the amount of energy required to switch from one reference state to another. In the context of triggered deep convection, there must be at least two reference states: a shallow one, which is the actual state (or a slightly modified one if there is boundary layer CAPE), and a deep one associated with the release of MAE/CAPE; the energy barrier separating these two reference states naturally defines a generalized CIN. In the limited context of the above-mentioned IOP ARM data, it is further shown that: (c) Spatially averaged conditions exhibiting potential instability to deep convection may be associated with individual soundings having widely different stability characteristics, including absolute stability, potential instability, and absolute instability; this suggests that triggered deep convection may not necessarily be the result of a parcel's vertical kinetic energy exceeding its negative buoyancy, but rather from the destruction of convective inhibition (i.e. production of absolute instability) in a local region. (d) A few local soundings exhibit multiple reference states, corresponding roughly to multiple levels of neutral buoyancy. These may allow for convective clouds with cloud-top heights significantly lower than those classically predicted by the undiluted ascent of a boundary-layer air parcel up to its highest level of neutral buoyancy, even in the absence of complex entrainment scenarios.

J. Y. Grandpeix, V. Phillips, and R. Tailleux. Improved mixing representation in Emanuel's convection scheme. Quarterly Journal of the Royal Meteorological Society, 130:3207-3222, October 2004. [ bib | DOI | ADS link ]

Recent empirical and modelling studies suggest that mid-tropospheric relative humidity (RH) is an important controlling factor of deep atmospheric convection, which appears to be underestimated in present cumulus parametrizations. This indicates the possible presence of shortcomings in the way that entrainment is represented in such parametrizations. This matter was explored in the European Cloud Systems project (EUROCS) by means of an idealized humidity experiment in which the main controlling parameter is RH. In the latter study, cloud-resolving model (CRM) experiments suggested that a shallow/deep convection transition occurs when RH crosses a threshold value that ranges from about RH = 50% to RH = 60%. In this paper, we seek to increase the responsiveness of Emanuel's convection scheme to RH, and to reproduce the threshold behaviour of the idealized humidity case, by replacing the original uniform probability density function (PDF) for mixing fractions by a more flexible two-parameter bell-shaped function that allows a wider range of behaviour. The main result is that the parameters of this PDF can be tuned to allow a regime transition to occur near a threshold value of RH 55%. In contrast to CRM results, however, this transition is between two different regimes of deep convection rather than between a shallow and deep regime. Possible ways to obtain a shallow-to-deep transition with Emanuel's scheme are discussed.

S. H. Derbyshire, I. Beau, P. Bechtold, J.-Y. Grandpeix, J.-M. Piriou, J. L. Redelsperger, and P. M. M. Soares. Sensitivity of moist convection to environmental humidity. Quarterly Journal of the Royal Meteorological Society, 130:3055-3079, October 2004. [ bib | DOI | ADS link ]

As part of the EUROCS (EUROpean Cloud Systems study) project, cloud-resolving model (CRM) simulations and parallel single-column model (SCM) tests of the sensitivity of moist atmospheric convection to mid-tropospheric humidity are presented. This sensitivity is broadly supported by observations and some previous model studies, but is still poorly quantified. Mixing between clouds and environment is a key mechanism, central to many of the fundamental differences between convection schemes. Here, we define an idealized quasi-steady 'testbed', in which the large-scale environment is assumed to adjust the local mean profiles on a timescale of one hour. We then test sensitivity to the target profiles at heights above 2 km. Two independent CRMs agree reasonably well in their response to the different background profiles and both show strong deep precipitating convection in the more moist cases, but only shallow convection in the driest case. The CRM results also appear to be numerically robust. All the SCMs, most of which are one-dimensional versions of global climate models (GCMs), show sensitivity to humidity but differ in various ways from the CRMs. Some of the SCMs are improved in the light of these comparisons, with GCM improvements documented elsewhere.