Referee Comment

"Reviewer B" (2010-06-04)


The paper deals with an important and largely unaddressed issue in cumulus parameterization, the effects of sub-grid organization in convective systems. The paper nicely outlines key observed characteristics of this organization and their possible implications for parameterized convection. A general variable related to organization, intended to capture the idealized behavior of organization as linearly forced with temporal damping, is introduced. The paper also applies the "org" variable to two aspects of organization, variations in boundary-layer temperature (and thus launch temperatures for cumulus parcels) and the effects of air detrained from earlier parcels on subsequent development. The approach is novel and merits discussion, but the attempt at generality in the "org" development may interfere with optimal design of parameterizations for specific forms of organization.

Major Issues

1. The greatest strength of this paper is its identification of a number of aspects of convective systems which are not treated by cumulus parameterization. The paper makes a strong case that parameterizing these aspects, related in various ways to organization of convective systems, will change the behavior of the models using the parameterizations in important ways. A reasonable path forward, based on that case, would be to attempt to parametrize these aspects of convective systems in a systematic way. Instead, a general variable "org" is introduced, applied to two specific aspects of organization, and advocated as useful for several others not yet studied. Lacking much theoretical or process-based information about organization in convective systems, proposing that organization is linearly forced and temporally damped is very reasonable. However, using the precipitation evaporation rate (EP) for all aspects of organization is probably not. It makes sense for the variability of boundary-layer temperature, although, even in that case, variability of humidity, which will also alter parcel lift properties, could also be considered. It's much less clear that using EP as a measure of the extent of parcel overlap for purposes of determining the properties of entrained air in later-developing cumulus is the best choice. For example, early-developing, non-precipitating plumes may be important in this respect, and the EP -dependent "org" might be a poorer choice than convective mass flux.

The main point here is to suggest that a better path forward might be to systematically proceed through forms of organization and treat each individually, as simply as possible. Using the same "org" variable in all cases seems to be a needless constraint, since not all organization will likely link to a single forcing. In fact, the ways in which "org" is used in the two applications discussed in the paper differ greatly, i.e., the parameterizations for the two different aspects are quite different, except for both using "org". Why not consider the best variable to characterize each type of organization, depending on its particular characteristics?

2. In addition to the general question as to whether a single "org" variable is the best way to drive parameterizations for multiple aspects of the organization of convective systems, the paper presents two candidate parameterizations. The evaluations are not complete. At the very least, the results of the parameterization for variations in parcel temperature, presented in Figures 3-6, should include observations. The parameterization for effects of prior parcels on entrainment uses an idealized SCM. Although useful for illustrating the behavior of the parameterization, results using an SCM with observed forcing and observations are required to assess the parameterization. This sort of test is certainly possible and strongly recommended.

Minor Issues

1. In the discussion following Eq. (1), α should be defined. If I understand the description correctly, it is defined as org/τ‾EP, where τ‾EP is the global-mean evaporating precipitation. Sensitivity to choice of this parameter is an important issue for the parameterization.

2. The paper frequently uses colloquial language. Although I generally found the meanings clear (and fun to read), it may be better for purposes of a wide readership to use more precise terminology, e.g., instead of "juicy" parcels, "more humid" or "warmer".

Summary

The paper aims at making three points: (1) importance of treating convective organization in parameterization, (2) utility of a single-parameter "org" as a control for a range of parameterizations dealing with various forms of organization, and (3) reasonableness of specific parameterizations using "org" for two forms of organization. The paper achieves (1) very well, but (2) remains speculative, and to demonstrate (3) requires comparing model results with observations and going beyond an idealized SCM for one of the parameterizations. Success with (3) would help to make the case for (2) also.



adv-model-earth-syst.org | james-p.org | james.iges.org