Data sets of historically measured climate parameters, e.g. sea surface temperature (SST), sea ice (SI) concentration, marine and land air temperature, sea level pressure, surface winds, precipitation, sea surface heights, etc, extend for the last one or two centuries and constitute the observational basis for modern climate research. Yet due to irregular and time-dependent sampling patterns these observational data sets are hard to use in climate studies directly, thus they are often interpolated into regular grids with infilling of remaining gaps. Production of such interpolated data sets, a.k.a.reconstructions, also involves quality control of input data and sometimes correction of their systematic biases.
Wide use of historical (century or longer) reconstructed data sets in climate research and applications attracts a great deal of attention to methods used to produce them, to their actual uncertainties, and especially to the differences between existing multiple interpolated products for the same climate variable (e.g. SST). In contrast with paleoreconstructions, for instrumental data sets the theoretical connection of measured values with the target field is that of a direct sampling and generally does not have to be fit statistically. This feature makes the reconstruction problem simpler, on one hand. On the other hand, expectations of the product quality for instrumental reconstructions are generally higher than those for paleoreconstructions: we are trying to push limits in terms of higher spatial and temporal resolution, precision in uncertainty representation, and also in the data set homogeneity, despite drastic changes in data coverage and measurement methods. These goals motivate and renew research emphasis on better reconstruction and uncertainty representation methodology, blending in situ and satellite data, merging marine and land data, using statistical parameters, like observational error and target field covariances, and constraints estimated from abundantly observed modern data to improve reconstructions for earlier time periods. There are also specific technical problems associated with the large volume of observed data and reconstructed data sets as well.
This session calls for papers on all methodological aspects and existing problems of reconstructed climate data sets based on instrumental data. This focus also includes application and intercomparison papers that expose current problems with existing data sets or methodology. Papers comparing and contrasting gridded reconstructions based on paleoclimate proxies with those using instrumental data as well as papers comparing methods used for these two types of reconstruction are of particular interest.