Tobias Ross
Investigating the Effects of CCN on the Timing of Convective Cold Pool Initiation During CACTI
Latent cooling from hydrometeor phase changes in precipitating downdrafts can spawn pools of cold air that spread outward at the surface. These cold pools exert controlling influences on the initiation, intensity, and organization of subsequent convection. Models requiring convective parameterizations lack adequate representation of cold pools and their effects, due in part to lack of consensus as to the most important physical drivers. Considerably little attention has been paid to microphysical influences on cold pool characteristics. We investigate the hypothesis that cold pools are generated sooner, relative to the initiation of their parent convection, in environments with fewer cloud condensation nuclei (CCN), that encourage a faster warm rain process and therefore expedite the earliest instances of rainfall. Multiple cases of well-observed cold pools forming in a range of ambient CCN concentrations are required to test the hypothesis. The Cloud, Aerosol, and Complex Terrain Interactions (CACTI) field campaign affords us this opportunity due to concentrated in-situ observations of CCN, microphysical measurements within developing cumuli, and radar observations of subsequent deep convection and cold pools.
CCN number concentrations, measured during nine CACTI research flights that targeted developing cumuli, are compared with hydrometeor concentrations within the clouds. Cloud fields ingesting fewer CCN consistently feature drizzle concentrations up to an order of magnitude greater than for those ingesting more CCN, denoting expedited collision-coalescence. This justifies our use of CCN concentrations as a robust predictor of the speed of the warm rain process in subsequent deeper convection. The timing of initiation of cold pools is quantified on three of these days using the CACTI radar network combined with RELAMPAGO and Argentina Mesonet observations. Cold pool onset is defined as the time elapsed between the first 30 dBZ echo and the appearance of a significant deviation in near-surface radial velocity beneath the precipitating cloud. Each cold pool is confirmed by matching the gust front to surface temperature decreases and wind shifts. These cases suggest that CCN number concentrations on a given day are not a good predictor of cold pool onset time, opposing the hypothesis. Idealized CM1 simulations of one case do show a sensitivity in cold pool onset time to ambient CCN concentrations: reducing CCN from the observed 900 cm-3 to 350 cm-3 reduces cold pool onset time by 30 minutes, supporting the primary hypothesis. Modelling additional cases across different convective environments will be performed in future work to understand these disparate initial findings.
Jiacheng Ye
Process-oriented Diagnostics on GEFSv12 Reforecasts
The Global Ensemble Forecast System v12 (GEFSv12) became operational in September 2020, which was accompanied by the release of a new 20-year reforecast dataset (2000-2019). A comprehensive evaluation of the GEFS v12 is carried out in order to provide guidance on future model improvement. Three levels of diagnostics are implemented: The Level-1 diagnostics focus on the evaluation of model systematic errors (moisture and precipitation, clouds, etc.); Level-2 consists of the evaluation of subseasonal predictability sources (the Madden-Julian Oscillation, weather regimes); Level-3 is about the evaluation of high-impact weather systems like tropical cyclones and blocking highs.
Our evaluation reveals that convection onset occurs too early in terms of column water vapor (CWV) accumulation, which leads to prevailing negative biases in CWV and positive biases in precipitation in the tropics. It is also shown that the MJO is skillfully predicted up to 16 days ahead, and the GEFSv12 also better represents the eastward propagation of the MJO over the Maritime Continent than an earlier model version. Weather regimes are another source of subseasonal predictability, and the GEFSv12 skillfully predicts the weather regime frequency beyond one week. Further analyses reveal that the teleconnections between MJO and weather regimes break down after Day 5 in the reforecasts, suggesting the model’s deficiencies on representing the tropical-extratropical interactions on longer time scales. Additionally, we show that the GEFS reproduces the spatial pattern of tropical cyclogenesis over the western Pacific but large biases in tropical cyclogenesis distribution are present over the eastern Pacific and Atlantic. The climatology of blocking highs is well captured by the model except over the Atlantic basin, where the occurrence is substantially underestimated. We hope our work would provide valuable information for model improvement.
