Understanding the mechanisms that give rise to droughts and floods is essential to understanding their causes, character, probability of occurrence and potential predictability. Understanding the causes of projected future changes in hydroclimate is also essential in assessing the reliability of these projections and in efforts to assess whether human-induced hydroclimate change is already occurring. To achieve such understanding requires detailed analysis of the atmospheric branch of the hydrological cycle relating anomalies in moisture convergence (floods and wetting) or divergence (droughts and drying) to anomalies in moisture transports and atmospheric circulation. This requires analysis of the full three dimensional atmospheric state with high frequency data capable of capturing the transport of moisture by transient eddies (storm systems). Whether such analyses are performed on atmospheric reanalyses or climate models, it always presents some problems. Choices must be made in terms of numerical methods to be used (such as how to do vertical integrals), time resolution to use, numbers of vertical levels, even which version of the moisture conservation equation to use. In a soon to be published paper (Seager and Henderson 2013) we offer a detailed analysis of errors introduced by various choices and/or limits on data availability and suggest best practices.

• Seager, R., and N. Henderson, 2013: Diagnostic computation of moisture budgets in the ERA-Interim Reanalysis with reference to analysis of CMIP-archived atmospheric model data. J. Climate, in press.