Evaluation of a stereo-derived cloud cover record for the greenland ice sheet

Abstract

The difficulties of using satellite imagery to determine the presence and type of cloud cover over snow and ice surfaces are well documented, introducing significant uncertainty in environmental and climate modelling for which cloud cover is an essential input. For example, an incomplete knowledge of the spatial and temporal distribution of cloud cover over Greenland precludes accurate predictions of the response of the ice sheet to climate change from energy balance models. This paper demonstrates the potential value of the technique of stereo- matching nadir and forward A TSR-2 thermalmages for the retrieval of the cloud fraction with the presentation of a one year record of cloud cover above Greenland. A number of empirical, observational, satellite-based and numerical reanalysis cloud cover datasets have been generated in recent years but this paper highlights their gross dissimilarity at high latitudes and their failure to identify regional variations in cloud cover, such as those associated with the North Atlantic Oscillation. The twice daily overpass of ATSR-2 and its 1 km resolution however permits the production of a detailed record of cloud cover that ground observations and radiosonde profiles indicate to be accurate in terms of resolving the presence and height of cloud to within 10 % and 500 m respectively. The value of the ATSR-2 derived cloud cover over other records is illustrated when the radiation balance is calculated, revealing the sensitivity of the zone around the equilibrium line to changes in cloud amount and type due to its position in the transition between the low albedo coastal margin and areas of lower cloud cover across the ice sheet interior. 'radiation paradox' have shown that although increased cloud cover reduces the shortwave radiation flux, the increase in longwave flux predominates (2). With a lack of detailed cloud data, many of the early energy balance models of the Greenland ice sheet over-parameterised the effect of clouds by using parameters based on Alpine conditions. However, at higher latitudes transmission due to clouds is larger due to their smaller optical depth. Moreover, it is known from regional investigations that incorporating independent variables of low, middle and high cloud gives better predications of both shortwave and, particularly, longwave radiation, and hence their different impacts (3). The complexity of the relationship between cloud cover and radiation budget was demonstrated by the measurements of Bintanja and van den Broeke (4) in both Greenland and Antarctica, illustrating the importance of surface albedo, and, in particular the opposing effects in the ablation and accumulation zones. In recognition of the importance of characterising cloud amount and type, a number of regional and global projects have been established to better understand the role of cloud cover in existing climatic and environmental processes, and also to estimate more reliably the contribution of changing cloud cover to a changing climate. The output of five of these records over the Greenland ice sheet is compared with a satellite based stereo-matching technique in order to ascertain the benefits of the accuracy and relatively high spatial and temporal resolution offered by this method.