Examination of inconsistencies in the surface-observed cloud amount record over China: 1954–2008

Surface-observed cloudiness records are valuable tools in assessing regional and global climate changes. For most of the world, these records predate satellite cloud observations by at least several decades, and most countries still compile these data as part of routine synoptic weather observations, making for quite lengthy records. China is one such country, with records beginning generally in the 1950s and continuing to the present. A quality assured "national" dataset for China compiled by the China Meteorological Administration (CMA), comprised of 6-hourly estimates of cloud amount, has been used in several studies since the late 1990s. These studies have shown significant decreases in cloud amount over much of China from the 1950s through the 1990s. Several global cloud amount studies have incorporated more "raw" cloud observations (less quality control at the original data source) transmitted via the Global Telecommunications System (GTS) and reached the same general decreasing cloud amount finding for China, although these records do not commence until 1971 (but extend later, through 2009). In this study we compare and contrast 6-hourly observations from the CMA and GTS datasets, the latter being extracted from NOAA's Integrated Surface Database (ISD). Aside from the differences in cloud amount units (CMA uses tenths of sky cover; ISD uses octas (1/8s of sky cover)) - which ideally would not cause major disagreements when converted to percentages and averaged over sufficient time scales - we find a significant degree of large, simultaneous differences in cloud amount among many of the roughly 200 stations common to both datasets. In addition to attempting to reconcile such differences, we will present findings regarding which of the datasets gives a more accurate picture of cloud amount trends, including an updated assessment of our earlier findings which used changes in the frequency of occurrence of specific cloud amount classes to illustrate some significant effects of increasing anthropogenic aerosols on surface-observed cloud amount estimates.