Data Fusion and Consistency of Fusing Data

Many remote sensing observations of vertical profiles of atmospheric variables are obtained with instruments operating on space-borne and airborne platforms, as well as from ground-based stations. When the same portion (or nearby portions) of atmosphere is observed more times by the same instrument or by different instruments the measurements can be combined in order to obtain a single vertical profile of improved quality with respect to that of the profiles retrieved from the single measurements. Recently, a new method of data fusion, referred to as Complete Data Fusion (CDF), was proposed for use in the combination of independent measurements of the same profile. This is an a posteriori method that uses standard retrieval products and with simple implementation requirements provides products equivalent to those of the simultaneous retrieval, which is considered to be the most comprehensive way of exploiting different observations of the same quantity. As part of the AURORA project, we have applied the CDF method to ozone profiles obtained from simulated measurements in the ultraviolet and in the thermal infrared in the framework of the Sentinel 4 mission of the Copernicus programme. We observe that the quality of the fused products is very good when we fuse consistent profiles, instead the quality is degraded when we fuse profiles that are either retrieved on different vertical grids or referred to different true profiles. In order to address this shortcoming, a generalization of the CDF method, which takes into account interpolation and coincidence errors, was developed. We determine the expressions of these errors and show how they enter in the CDF formula. This upgrade overcomes the encountered problems and provides products of good quality also when the fusing profiles are both retrieved on different vertical grids and referred to different true profiles. The approach developed to account for the interpolation and coincidence errors can also be followed to include other error components, such as forward model errors.