The project has two major objectives. The first is geared towards improving climate datasets (both instrumental and proxy) and gridded products that can be derived from them. The second is using these data to assist in the validation of output from GCMs and RCMs. The first of these objectives includes the continued updating of the monthly global temperature record, incorporating as much additional station (preferably homogenized) data from around the world. Considerable amounts of daily station temperature and precipitation data are now becoming available. In this project, we will analyse these data and compare them with similar data developed from the latest Reanalyses (ERA-INTERIM). The aim of this aspect of the work is to see if these new Reanalyses are of climate quality, and whether they can be used in the real-time climate monitoring of extremes. Gridded versions of daily station data are being developed for some regions of the world for validation of RCMs. This proposal will assess how well these gridded products reproduce the whole distribution, as there is much evidence that they perform less well for extremes. Transformations of the daily temperature and precipitation distributions will be investigated which both perform better for extremes, but still do equally well with the majority of the distribution. Finally, with respect to longer-term climatic data, more and more proxy data are becoming available. In this project, many of the long European documentary and natural proxy records will be investigated in the light of slight reductions to the summer temperature values (developed for the period before 1850 in previous projects) due to poorer exposure of the instruments before the advent of screens. The second major objective of the proposal is to use these improved climate datasets for climate model validation, both of RCMs and GCMs on continental and global space scales, respectively. The emphasis with RCMs at the continental scale will be on the daily timescale and will assess a set of RCM simulations for Europe that have been forced with ERA-40 boundary conditions. These runs can be used both for assessing which models perform better than others, but they can also be used to determine upper limits of the probable maximum quality in the current generation of RCM simulations. Longer-timescale variability with GCMs will be assessed using simulations of the last 500 to 1000 years, and also longer duration Reanalyses that will soon be completed from 1892 (using only surface observational data, principally of pressure and wind data). With the longer simulations, we will take part in the Paleoclimatic Reconstruction Challenge, and we will compare the longer Reanalyses with daily temperature and precipitation records available for Europe and North America.