Bias in return period of droughts estimated from tree-ring reconstructions

Abstract

This article investigates the bias in return period of droughts of extreme duration evolved from trun-cating tree-ring reconstructed precipitation. 239 years of tree-ring reconstructed annual precipitation for Madaba region in Jordan, covering the period 1743–1981, was used to explore the bias. Furthermore, 100 traces of simulated annual precipitation, each of 239 years, were generated using stochastic simulation model for comparison purposes. The results show that the return period estimated from the analysis of the tree-ring reconstructed precipitation departs increasingly downward from the expected results provided by simulation or theoretical models as the drought duration increases (becomes more extreme). The bias in the estimated return period can be neglected when the drought duration is fairly short, i.e. 1–2 years for the case under study.