Testing convective transport on short timescales: comparisons with mass divergence and ozone anomaly patterns about high rain events

High rain events in the western tropical Pacific were identified using the TRMM 3B42 gridded rainfall dataset. Horizontal wind measurements from radiosonde arrays, cloud top frequency measurements from CALIPSO, and ozonesonde measurements from SHADOZ, were used to construct anomaly patterns of divergence, cloud top frequency, and ozone mixing ratio about the high rain events. The observed divergence anomaly pattern was compared with patterns produced by the GEOS-4 and GEOS-5 assimilated meteorological datasets. The mid-level (~ 6 km) divergence dipole seen in observations – consisting of a local maximum in divergence 8 hours before peak rainfall and a local maximum in convergence 8 hours after peak rainfall – is not represented in the GEOS-4 or GEOS-5 meteorology. The ozone anomaly pattern shows ozone decreases at mid-levels (3 – 8 km) up to 16 hours prior to peak rainfall. These decreases occur in association with increases in the frequency of mid-level cloud tops and are due to increased detrainment from cumulus congestus clouds. The observed ozone anomaly pattern was compared with anomaly patterns produced by the GEOS-Chem chemical transport model, driven by GEOS-4 and GEOS-5 assimilated winds and temperatures. The GEOS-Chem simulations also exhibit negative ozone anomalies at mid-levels in association with high rain events. However, these anomalies are nearly symmetric about peak rainfall, rather than strongest prior to peak rainfall. The upper level negative ozone anomalies of the models are more persistent than the observations. These results help characterize some of the difficulties of meteorological datasets in capturing the layered character of tropical convective outflow, and its timing with respect to high rain events.