Selecting Markov chain orders for generating daily precipitation series across different Köppen climate regimes

Markov chain models are a commonly used statistical technique to generate realistic sequences of precipitation, but the choice model order can strongly affect their performance. Although it is widely accepted that first-order reproduces precipitation occurrence in temperate latitudes quite well, also well known have several shortcomings. These include limited memory rare events and inaccurately reproducing distribution dry-spell lengths, performance outside regions less understood. We present, therefore, first assessment model-order optimization which both global extent uses four evaluation methods: Bayesian information criterion (BIC) each model-order's ability reproduce wet- interannual variability occurrence. As as analysis, we assessed selection separately within five climate regimes based on Köppen classification system: tropical, dry, temperate, continental polar. metrics were determine best performing time series across different regimes. find most sensitive metric dependent regime. Across all regimes, show performs when evaluated with BIC for generating wet-spell distributions except where third best. third-order observed second once again Our findings highlight benefits flexible tailored approach constructing series.