The Cucconi Test for Location-scale Alternatives in Application to Asymmetric Hydrological Variables

Abstract

To identify location-scale trends, which environmental data often exhibit, location-scale tests have to be addressed. The aim of this article was to estimate size and power of the Cucconi rank-based test when applied to various skewed distributions, typical in hydrology. Results of the Monte Carlo simulation revealed great power for series with low coefficient of variation, time of change close to the middle, not very heavy tail, and with length of at least 60. Comparison to the Lepage test discovered larger usefulness of the Cucconi test for short series and change close to the middle. Several practical applications were presented.

Keywords:

• Hydrological time series
• Monte Carlo simulation
• Nonparametric framework
• Power and size of the test
• Test for location-scale change.

Mathematics Subject Classification:

• 62G10
• 65C05
• 62P12.