Figure 1. Cumulus in the vicinity of Ascension Island, in a 100 x 100km image (which is close to global climate model spatial resolution) from MODIS onboard AQUA (22 July 2016)
Shallow cumulus cover large areas in the trade-wind marine regions and are considered to play an important role in the ocean–atmosphere exchanges and in the Earth radiative energy budget. Their small size (typically few hundreds of meters, Figure 1) and the heterogeneity of cumulus fields make it very challenging for global climate models which have relatively coarse spatial resolution (from tens to hundred of kilometres). Therefore, sub-grid physical processes (such as mass flux transport and entrainment) need to be parametrised into the models. Recent intercomparison studies between several climate models suggest that cumulus parametrisation is a source of a large discrepancy amongst models, causing uncertainty on climate sensitivity. To improve the representation of cumulus in models, more observations, especially at high-resolution, are needed.
In this work, we are focusing on cumulus occurring at Ascension Island (8°S, 14°W, see Figure 1) in the Southeast Atlantic Ocean. This region is doubly interesting as regards cumulus study. First, the occurrence of cumulus is relatively high (between 40 and 60%) from January to August, as the island is located in the southeasterly trade-wind region. Secondly, a large amount of biomass burning aerosols is transported from Southern Africa over the island, which makes it also interesting to study the interaction between clouds, aerosols and radiation.
Figure 2. 3-D cumulus reconstructed field from Ka-band scanning radar on 17 July 2017 and its 2-D projection. The size of the domain is 7 x 4km.
Capitalizing on a method developed at University of Reading (Fielding et al., 2014), we used reflectivity scans measured by a Ka-band (35 GHz) scanning radar to reconstruct 300 non-precipitating cumulus cloud fields (see an example in Figure 2). Those fields allow us to address macrophysical (cloud size distribution, thickness and cloud cover) and microphysical properties at a 50-m resolution. One of the key findings of this work is that, for the first time, the cloud size distribution of observed 3-D cumulus fields is described and shows that cumulus population follows a simple relationship (power-law), in agreement with satellite previous observations in 2-D. We also found that liquid droplet size is slightly smaller than what was observed in other Atlantic and Pacific regions. This is consistent with the presence of biomass burning aerosols that interact with clouds. We expect this unique dataset to help to improve our understanding of cumulus physical processes and to aid the low-cloud parametrisation efforts.