Comments on the Regional Climate Variability Driven by Foehn Winds in the McMurdo Dry Valleys, Antarctica

TL;DR

This paper critiques a previous study on regional climate variability in Antarctica, arguing that its statistical analyses and conclusions about ENSO and Southern Annular Mode influences are flawed and unsupported.

Contribution

It provides a critical analysis of prior work, highlighting errors in statistical methodology and challenging the validity of its conclusions.

Findings

Previous study's statistical significance claims are invalid

Correlation coefficients were improperly calculated and interpreted

Main conclusions about climate influences are unfounded

Abstract

The main objection to Speirs, McGowan, Steinhoff and Bromwich [Int. J. Climatol. 33: 945-958] work arises from the lack of analyses of the probability distribution functions of underlying processes leading to wind formation of which velocities are measured by automated weather stations and reported in the paper. Mathematically a rigorous definition of calculating the correlation coefficient (Pearson product-moment correlation coefficient) of averages does not exist. Therefore the authors numbers as given in Table II represent a set of randomly calculated figures. The authors suggestion in relation to a few of these random numbers that some of them have statistical significance at the 95% level is erroneous since no relationship exists between correlation coefficient of averages and statistical significance. Therefore Speirs et al. main conclusion that the El Ni\~no Southern Oscillation (ENSO) and the Southern Annular Mode significantly influence foehn wind frequency at McMurdo Dry Valleys is entirely unfounded.