Improving Model Representation of Interactions Between Moisture and Tropical Convection

Realistically representing the multi-scale interactions between moisture and convection remains an ongoing challenge for weather prediction and climate models. In this study, we examine how such interactions evolve throughout the convective lifecycle, and as a function of spatiotemporal scale, by examining the relationship between precipitation and column saturation fraction (CSF). The convective to stratiform precipitation ratio is predominantly a function of CSF, not precipitation rate, suggesting that parameterizations of convective organization should maintain a strong relationship with CSF. Several models, all implementing versions of the Zhang-McFarlane deep convective parameterization, exhibit excessive drying of the large-scale environment in their representation of the transition from shallow to deep convection. The convective adjustment timescale (i.e. the timescale by which precipitation relaxes moisture towards its ``background'' state) lengthens with increasing spatiotemporal scale, allowing moisture anomalies of larger scale to persist longer than those of smaller scale. Models with conventional convective parameterizations struggle to reproduce this spatiotemporal scale dependence, likely as a result of insufficient CSF variance. Potential implications for equatorial wave dynamics are discussed.

A roughly exponential increase in precipitation occurs as CSF increases above a ``critical point'', which acts as an attractor in CSF-precipitation phase space (Peters and Neelin, 2006). For a wide variety of starting conditions, the system tends to evolve towards this moist, moderately precipitating state. Each perturbation of the system away from the attractor state, presumably driven by processes such as surface forcing or equatorial wave dynamics, causes it to evolve in a cyclical fashion around the attractor. This cyclic evolution is characterized by large-scale moistening as the convective ensemble transitions from shallow, to deep, to stratiform convection, followed by large-scale drying in a non-precipitating or lightly precipitating regime. This behavior exhibits properties consistent with self-organized criticality and characteristics of self-similarity, suggesting that shortcomings in model representation of the joint evolution of convection and large-scale moisture will negatively impact a broad range of spatiotemporal scales.