Abstract: To address the challenge of estimating the return level of hydrological drought events due to the limited sample size of drought events that can be obtained from measured streamflow data, we applied three commonly utilized annual runoff probability distribution functions,namely, Log-normal, Gamma, and Normal,to measured runoff data obtained from the Waizhou station on the Ganjiang River. Theoretical probability distribution functions (PDFs) for drought characteristics, including duration and severity, were derived using statistical properties of annual runoff. The return period, defined as the mean interarrival time of drought events surpassing a certain severity threshold, was computed and validated through Monte Carlo simulation. Results demonstrate that deriving return periods of hydrological drought events using PDFs of drought duration and severity establishes a robust statistical basis with credible accuracy. The proposed method partially mitigates sample bias in estimating drought return periods based on limited observed hydrological series, offering a novel approach to assessing future drought risk.

Key words: hydrological drought, return period, calculation method, annual runoff, Ganjiang River basin