Testing a reanalysis‐based infilling method for areas with sparse discontinuous air temperature data in northeastern Canada

Abstract

Abstract This study tests various applications of a new technique for infilling sparse monthly climate data which combines temperature anomalies from gridded observational and reanalysis data sets with baseline climatologies from short‐instrumental records. Out‐of‐sample comparisons between infilled and observed climate data for 53 stations in northeastern Canada suggests that mean absolute errors using the proposed method are ±1 and ±0.5 °C on monthly and annual timescales, respectively. Evaluation of several gridded data sets used to guide infilling suggests that ERA‐Interim ( ERAI ) and modern‐era retrospective reanalysis ( MERRA ) reanalyses are the most suitable for this purpose in the Labrador‐Ungava region.