Validating the validation: the influence of liquid water distribution in clouds on the intercomparison of satellite and surface observations

The intercomparison of LWP retrievals from observations by a geostationary satellite imager (SEVIRI onboard MSG) and ground-based MW radiometer is examined in the context of cloud inhomogeneity. Although the influence of cloud inhomogeneity on satellite observations has received much attention, relatively little is known about its impact on validation studies.

This paper quantifies the various validation uncertainties due to cloud inhomogeneities and proposes an approach to minimize these uncertainties. The study is performed by simulating both satellite and ground-based observations for a set of high-resolution (100 m) cloud fields that are derived from 1 by 1 km2 MODIS observations. Our technique for generating realistic high-resolution LWP fields preserves the information present in the original observations while creating extra LWP variation at smaller length-scales by considering clouds as simple fractals. To our knowledge, this is a new technique for creating high-resolution LWP fields.

Validation errors due to cloud inhomogeneity can be classified in two groups. The first group relates entirely to the retrieval process for satellite observations and includes the well-known plane parallel bias as well as mismatches between different channels. The second group relates to differences in the observed scene of both satellite and ground-based sensor. This includes systematic shifts in observed scene due to viewing conditions (parallax effect), offsets between satellite image and ground site as well as different field-of-views.

Of all the error contributions to the validation, the parallax effect easily tends to dominate for sites that are observed under large viewing angles (e.g. Northern Europe). We show that this error may be partly compensated by using information about cloud top heights and by spatial interpolation among an array of SEVIRI pixels to obtain the best estimate of the satellite retrieved LWP value over the ground site. Optimal intercomparison of satellite and ground-based observations is furthermore possible by matching the tracklength of the ground observations to the imager’s pixel size in the wind direction.
One surprising conclusion is that the LWP errors due to the second group (scene differences) are significantly larger than those due to the first group (satellite retrieval), even after we have applied corrections. Equally surprising, smaller satellite pixels do not alleviate the problem but rather aggravate it, unless the parallax error is corrected. Calibration errors are not considered in the present study.