Temperature‐extreme precipitation scaling: a two‐way causality?. (5th December 2017)

• Record Type: Journal Article
• Authors: Barbero, R.; Westra, S.; Lenderink, G.; Fowler, H. J.
• Is Part Of: International journal of climatology. Volume 38(2018)Supplement 1
• Page Start: e1274
• Page End: e1279
• Publication Date: 2017-12-05
• Subjects: precipitation extreme -- temperature‐extreme precipitation scaling -- dew point temperature -- Australia -- climate change; Climatology -- Periodicals; Climat -- Périodiques; Climatologie -- Périodiques; 551.605
• DOI: 10.1002/joc.5370
• Languages: English
• ISSNs: 0899-8418

Abstract

ABSTRACT: Extreme precipitation events are widely thought to intensify in a warmer atmosphere through the Clausius‐Clapeyron equation. The temperature‐extreme precipitation scaling was proposed to analyse the temperature dependency of short‐duration extreme precipitation and since then, the concept has been widely used in climatology. Bao et al. (2017) suggest that the apparent scaling reflects not only how surface air properties affect extreme precipitation, but also reflects how synoptic conditions and localized cooling due to the storm itself affect the scaling – implying two‐way causality. We address here critical issues of this paper and provide evidence that dew point temperature drives extreme precipitation, with the direction of causality reversed only for the storm's peak intensity. This physical inference may serve as a basis to better quantify scaling rates and to help establish the relationship between extreme precipitation and environmental conditions in the current climate, and thereby provide insights into future changes to precipitation extremes due to climate change. Abstract : (a). Daily mean temperature observed in Darwin spanning 7 days before to 7 days after extreme precipitation events. Different colours from darkest blue (coolest) to darkest red (warmest) represent the ten different temperature bins. (b) Same as (a) but using a time series ( N = 10 000) with μ = 0, σ = 1 and the same skewness and first‐order autocorrelation observed in Darwin. The synthetic time series is then stratified into ten bins by quantile (from 0.1 to 1) and its temporal evolution before/after the values observed in each quantile is shown. (c) Distribution of observed daily temperature illustrating the negative skewness of the data. The horizontal line (boxplot) indicates the overall mean, as well as the standard deviation of the distribution.