『Summary
　This is the first of three related papers that summarizes streamflow characteristics of a set of 1221 global rivers. The rivers are well distributed world-wide, are unimpacted by upstream reservoirs or diversions for the period of data collection and have at least 10 years of continuous monthly and annual streamflow data. The following key features of annual flows are examined: mean, variability and skewness, distribution type (Gamma or Lognormal), flow percentiles and dependence. High and low frequency persistence is examined through the Empirical Mode Decomposition technique. Low flow run length, magnitude and severity are also explored, where severity is based on extended Deficit Analysis. It has been observed elsewhere that there are large differences in hydrologic characteristics between Australia and southern Africa in contrast to the rest of the world. This issue is tested further in this paper. The range of analyses and results presented herein also form a suite of empirical evidence that future unified theories of hydrology at the catchment scale must be able to adequately describe.

Keywords: Global hydrology; Global streamflow data; Global streamflow characteristics; Global rivers』

Introduction
Literature
Annual streamflow data
What probability distribution function?
Mean, variability and skewness
Flow percentiles
Dependence
Persistence characteristics
Low flow run length, magnitude and severity
Summary and conclusions
Acknowledgement
Appendix A. Frequency distribution of annual streamflows
　Lognormal case
　　L-Cv versus Cv
　　L-Skew versus Skewness
　Gamma case
　　L-Cv versus Cv
　　L-Skew versus Skewness
References