Extratropical cyclones are the primary source of precipitation in the mid-latitudes. The physical mechanisms that drive cyclones are well understood, and a variety of studies have demonstrated strong relationships between cloud structures and cyclone dynamics. However, past research has focused on simplistic cloud categorizations, which lack spatial and textural information that is included in more sophisticated classification schemes. Unsupervised deep learning approaches may have significant advantages over these past methods, allowing them to discover previously unidentified cloud information in large datasets. One such approach is the rotation-invariant cloud clustering (RICC), which combines a dimensionality reduction deep learning technique with rotation-invariant clustering of input cloud images. We employ the RICC, along with two established cloud clusters, to investigate the relationship between extratropical cyclones and horizontal cloud distributions. We focus on comparing these different sets of clusters to each other. First the spatial distributions and the physical properties of the identified cloud types are examined around cyclones and the results corresponding to each classification are compared in detail. Then the similarities of these distributions are quantified using structural similarity. Additionally, the evolution of spatial distributions of cloud over the lifetime of cyclones is compared between the different classifications. Interestingly, we identify the same broad physical developments in all sets of clusters. Notably, identified differences are likely due to differences in measurement processes and resolutions of the corresponding datasets.

A document by Yusuf A. Bhatti. Click on the document to view its contents.

ERA5 reanalysis output is compared to WindSat measurements over cyclones at Southern Hemisphere mid- to high-latitudes. WindSat provides an independent measure of how well ERA5 represents cyclones, as WindSat is not assimilated into ERA5. We implement a tracking scheme to identify cyclone centres and tracks, before using cyclone composites to match concurrent data in ERA5 and WindSat. We find that both ERA5 and WindSat show comparable spatial structures for low level wind speed, total column water vapour, cloud liquid water and precipitation. Compared to WindSat, ERA5 underestimates total column water vapour by up to 5\% and cloud liquid water by up to 40\%. ERA5 underestimates precipitation in the warm sector by up to 15\%, but overestimates in the cold sector by up to 60\%. Similar biases in ERA5 are seen when comparing to AMSR-E data, even though AMSR-E radiances are assimilated into ERA5. Comparing ERA5 and WindSat across the cyclone lifecycle, a strong correlation is seen across the cyclone as it deepens and reaches peak intensity, before slightly declining as the cyclone decays. In the cold sector ERA5 shows underestimation of cloud liquid water, yet overestimates precipitation at all lifecycle stages. However, in the warm sector precipitation is underestimated. This potentially suggests the presence of biases within the ERA5 parameterisations of cloud and precipitation causing a disconnect between the two. Despite this, ERA5 shows strong correlation with WindSat and determines cyclone structure well across the cyclone lifecycle, showing its value for use in cyclone compositing analysis.

This study compares CL51 ceilometer observations made at Scott Base, Antarctica, with statistics from the ERA5, JRA55, and MERRA2 reanalyses. To enhance the comparison we use a lidar instrument simulator to derive cloud statistics from the reanalyses which account for instrumental factors. The cloud occurrence in the three reanalyses is slightly overestimated above 3km, but displays a larger underestimation below 3 km relative to observations. Unlike previous studies, we see no relationship between relative humidity and cloud occurrence biases, suggesting that the cloud biases do not result from the representation of moisture. We also show that the seasonal variation of cloud occurrence and cloud fraction, defined as the vertically integrated cloud occurrence, are small in both the observations and the reanalyses. We also examine the quality of the cloud representation for a set of synoptic states derived from ERA5 surface winds. The variability associated with grouping cloud occurrence based on synoptic state is much larger than the seasonal variation, highlighting synoptic state is a strong control of cloud occurrence. All the reanalyses continue to display underestimates below 3km and overestimates above 3km for each synoptic state. But, the variability in ERA5 statistics matches the changes in the observations better than the other reanalyses. We also use a machine learning scheme to estimate the quantity of super-cooled liquid water cloud from the ceilometer observations. Ceilometer low-level super-cooled liquid water cloud occurrences are considerably larger than values derived from the reanalyses, further highlighting the poor representation of low-level clouds in the reanalyses.

ERA5 reanalysis is used to examine extreme precipitation using a spatially dependent precipitation threshold applied within a cyclone compositing framework. This is used to account for regional variation in precipitation generating processes within Southern Hemisphere mid-latitude cyclones across the cyclone lifecycle. The spatial extent of extreme precipitation is limited to a smaller region around the cyclone centre compared to non-extreme precipitation, though extreme precipitation displays a good spatial correlation with non-extreme precipitation. Extreme precipitation occurs more often during the deepening phase of the cyclone before it reaches a maximum depth. Precipitation occurrence at the 90th and 98th percentiles reduces to 46% and 30% of the deepening value across the cyclone lifecycle, averaged over the composite. Precipitation fraction at the 90th and 98th percentile reduces to 80% and 60% of the deepening value. Our methodology provides a quantitative assessment of precipitation extremes both spatially and temporally, within a cyclone compositing framework.

We demonstrate that the relationship between the abundance of particulate surface area observed at sea-level and measurements of backscattered light by a ceilometer can be used to classify the mixing state of the atmospheric layer beneath the lowest observed cloud, where the relationship is defined by the Spearman Rank correlation. The accuracy of this correlation-based method was compared to two methods of detecting boundary layer decoupling based on radiosonde measurements. An optimized version of the new methodology correctly determined the mixing state of the below-cloud layer for 76 ± 4% of the radiosondes available for comparison. Further, it was more accurate than an alternative ground-based metric used to determine the below-cloud mixing state. For the majority of the time series in which the correlation analysis could be applied, the below-cloud boundary layer was well-mixed (54%), or else fog was present (27%), which indicated that aerosol particles observed at sea-level often have a direct pathway into low-cloud (81%). In the remaining analysis period, the near-surface atmospheric layer was stable and the atmospheric layer near the ocean surface was decoupled from the overlying cloud (19%). Forecasts from the Antarctic Mesoscale Prediction System also support our findings, showing that conditions that mix aerosol particles from the ocean surface to the lowest observed cloud occur 84% of the time over the open Southern Ocean. As a result, aerosol particles measured near sea-level are often tightly coupled to low-cloud formation over the Southern Ocean, highlighting the utility of shipborne aerosol observations in the region.

Analysis of Coupled Model Intercomparison Project Phase 6 (CMIP6) models has focused on their higher equilibrium climate sensitivity (ECS) relative to CMIP5 simulations, with higher ECS values attributed to changing cloud feedbacks. We examine the simulation of the shortwave cloud radiative effect (SW CRE) in CMIP6 and explore a potential link between SW CRE errors and ECS. We derive mean and compensating errors in model data relative to satellite observations using a cloud clustering methodology. A statistically significant negative relationship between the mean and compensating errors in SW CRE over the Southern Ocean is identified. This relationship is observed elsewhere, but is only significant over the Southern Ocean. This implies model tuning efforts potentially hide biases in the representation of clouds in this region. High ECS models tend to have lower mean and compensating errors over the Southern Ocean, suggesting model biases are unlikely to explain the high ECS values.

Modeling the shortwave radiation balance over the Southern Ocean region remains a challenge for Earth system models. To investigate whether this is related to the representation of aerosol-cloud interactions, we compared measurements of the total number concentration of sea spray generated particles within the Southern Ocean region to model predictions thereof. Measurements were conducted from a container laboratory aboard the R/V Tangaroa throughout an austral summer voyage to the Ross Sea. We used source-receptor modeling to calculate the sensitivity of our measurements to upwind surface fluxes. From this approach, we could constrain empirical parameterizations of sea spray surface flux based on surface wind speed and sea surface temperature. A newly tuned parameterization for the flux of sea spray particles based on the near-surface wind speed is presented. Comparisons to existing model parameterizations revealed that present model parameterizations led to over-estimations of sea spray concentrations. In contrast to previous studies, we found that including sea surface temperature as an explanatory variable did not substantially improve model-measurement agreement. To test whether or not the parameterization may be applicable globally, we conducted a similar regression analysis using a database of in situ whitecap measurements. We found that the key fitting parameter within this regression agreed well the parameterization of sea spray flux. Finally, we compared calculations from the best model of surface flux to boundary layer measurements collected onboard an aircraft throughout the Southern Ocean Clouds, Radiation, Aerosol Transport Experimental Study (SOCRATES), finding good agreement overall.

The 2B-CLDCLASS-LIDAR R05 (2BCL5) and the raDAR/liDAR (DARDAR) satellite retrievals of cloud occurrence are compared as a function of altitude and latitude. The largest disparities are observed at low altitudes over high southern latitudes. These datasets are cross referenced to ground–based measurements from the Atmospheric Radiation Measurement (ARM) West Antarctic Radiation Experiment (AWARE) campaign at McMurdo Station, Antarctica. Compared to AWARE observations, both 2BCL5 and DARDAR underestimate cloud occurrence below 1.5 km, with 2BCL5 and DARDAR distinguishing roughly one third of cloud occurrences observed by AWARE at 0.5 km. While DARDAR identifies greater cloud occurrences than 2BCL5 below 1.5 km, cloud occurrence values for the two datasets have similar differences relative to ground-based measurements. Therefore, the DARDAR retrievals of greater cloud occurrence at low altitudes are likely due to a larger quantity of false positives associated with radar ground clutter or attenuated lidar retrievals. DARDAR cloud occurrences match better with AWARE than 2BCL5 above 5 km. However, the likely underestimation of ground-based measurements at higher altitudes suggests DARDAR may underestimate high level cloud occurrence. Finally, both datasets indicate the presence of liquid containing clouds at temperatures within the homogeneous freezing regime, despite the fact that the ECMWF-AUX dataset implemented in their processing clearly indicates temperatures below -38 °C. Using AWARE radiosonde (ECMWF-AUX) temperature data, we find that 2BCL5 detects 13.3% (13.8%) of mixed phase clouds below -38 °C, while DARDAR detects 5.7% (6.6%) of mixed phase and 1.1% (1.3%) of liquid phase clouds below -38 °C.