Reliability of climate model multi-member ensembles in estimating internal precipitation and temperature variability at the multi-decadal scale

Multi-member ensembles of climate models are used to study the role of internal climate variability and its potential impact on climate change impact studies. The reliability of such ensembles with respect to representing the internal climate variability is generally implicitly accepted. Using the latest version of the Climate Research Unit (CRU) data set as a baseline, this study verifies the reliability of multi-member ensembles in estimating the internal precipitation and temperature variability at the multi-decadal scale. To achieve this, multi-decadal variability calculated using climate model multi-member ensembles is first compared to that calculated using CRU data. The inter-member variability of a single climate model is then compared to the multi-decadal variability of CRU precipitation and temperature. Three climate models, with the number of members ranging between 5 and 40, are used to investigate whether the reliability is dependent upon the number of members of a climate model. The results show that multi-member ensembles are capable of capturing the observed spatial pattern of multi-decadal variability for both precipitation and temperature at the global scale. However, multi-member ensembles perform better over regions with smaller variability. For the variables and timescale investigated, all three multi-member ensembles show similar performances, suggesting that a five-member ensemble may be sufficient to estimate the internal multi-decadal climate variability for the chosen variables. Overall, this study indicates that the multi-member ensemble of a climate model can be used to estimate the internal climate variability of annual and seasonal precipitation and temperature at the multi-decadal and regional scales, if long historical records are not available.