Yao, B; Teng, SW; Lai, RZ; Xu, XF; Yin, Y; Shi, CX; Liu, C (2020). Can atmospheric reanalyses (CRA and ERA5) represent cloud spatiotemporal characteristics?. ATMOSPHERIC RESEARCH, 244, 105091.

Atmospheric reanalysis normally provides a comprehensive record of Earth's weather and climate changes over time, and is extensively used in climate and energy researches. Among various atmospheric variables, clouds play a unique and important role in the atmosphere. However, there are still significant uncertainties on our understanding of cloud properties, and it is still unclear how those newly developed atmospheric reanalyses represent cloud spatiotemporal characteristics. This study evaluates the cloud properties in two atmospheric reanalyses, i.e., the China Meteorological Administration Reanalysis data (CRA) and the ECMWF's Fifth-generation Reanalysis (ERAS), and monthly mean cloud cover from satellite radiometers are considered as the reference. Both CRA and ERAS generally represent cloud temporal and spatial distributions. To be more specific, ERAS gives monthly mean cloud cover closely to the MODIS observation over land, but its value over ocean is similar to 10% smaller than MODIS. CRA underestimates global cloud cover for similar to 20%, especially over subtropical zones. The differences between the MODIS and reanalysis cloud cover are mostly significant over the polar regions, whereas they can be contributed by the uncertainties of both. Furthermore, the interannual climate variability of cloud cover influenced by ENSO can also be reasonably captured by the two reanalyses, especially ERAS, whereas the CRA and ERAS cloud cover anomalies are both spreading over a slightly larger region than those from MODIS. Such evaluations are suggested to be extended to even more detailed cloud properties and variations in the reanalysis for further improvements.