Evaluation of Simulated Tropical Convective Updraft Properties using HAIC-HIWC Aircraft Observations

Relationships between tropical deep convective dynamics and microphysics in aircraft observations during the High Altitude Ice Crystals-High Ice Water Content (HAIC-HIWC) campaign in Darwin, Australia in 2014 are compared with those in high-resolution WRF simulations. These simulations included 5 separate events, each of which provided unique observations, including long transects of high ice water content (IWC), multiple temperature levels, and strong updrafts. Most of the flights were performed at temperatures of -30 °C and -40 °C with the usage of a high-powered isokinetic evaporator probe (IKP) to measure total water content, providing an advantage over many past experiments in which IWC retrieval uncertainties result from ice size and density assumptions used with optical probes. The majority of updrafts are small and weak, but our analysis focuses on the more significant ones that contribute most to upward mass flux. Relationships between convective updraft core diameter and vertical velocity, as well as updraft diameter and ice microphysics, are weak. At temperatures around -30°C and -40°C, IWC quickly increases as a function of vertical velocity, but this rate slows as vertical velocity increases, and for updraft average vertical velocities exceeding 3 m s-1, observations and simulations agree that updraft average IWCs are typically between 1.5 and 2 g m-3. Greater disagreement between simulation output and limited observations occurs at temperatures around -10°C, where simulated graupel and liquid water are overproduced, which decreases the snow water content and contributes to a high reflectivity bias. For a given updraft vertical velocity and condensate content, not a single simulated updraft can reproduce observed low reflectivities in the four updraft cores sampled at -12°C. Increasing model horizontal grid spacing from 1 km to 333 m and doubling vertical levels does not change this result. These differences in observations and simulations are related to assumed hydrometeor properties in microphysics schemes and differences in observed and simulated ice number concentrations and sizes that impact microphysical process rates and phase partitioning.