Approach to reconstructing a target large-scale climate field from predictors employing multivariate regression methods. CFR methods have been applied both to filling spatial gaps in early instrumental climate data sets, and to the problem of reconstructing past climate patterns from ‘climate proxy’ data.