Pitfalls in diagnosing temperature extremes

Lukas Brunner,Aiko Voigt

DOI: https://doi.org/10.1038/s41467-024-46349-x

IF: 16.6

2024-03-18

Nature Communications

Abstract:Abstract Worsening temperature extremes are among the most severe impacts of human-induced climate change. These extremes are often defined as rare events that exceed a specific percentile threshold within the distribution of daily maximum temperature. The percentile-based approach is chosen to follow regional and seasonal temperature variations so that extremes can occur globally and in all seasons, and frequently uses a running seasonal window to increase the sample size for the threshold calculation. Here, we show that running seasonal windows as used in many studies in recent years introduce a time-, region-, and dataset-depended bias that can lead to a striking underestimation of the expected extreme frequency. We reveal that this bias arises from artificially mixing the mean seasonal cycle into the extreme threshold and propose a simple solution that essentially eliminates it. We then use the corrected extreme frequency as reference to show that the bias also leads to an overestimation of future heatwave changes by as much as 30% in some regions. Based on these results we stress that running seasonal windows should not be used without correction for estimating extremes and their impacts.

multidisciplinary sciences

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the time -, region - and data - set - dependent biases introduced when using running seasonal windows to define temperature extreme events. Such biases may lead to a significant underestimation of the frequency of extreme high temperatures, and in some regions, the estimates of future heat wave changes may be overestimated by as much as 30%. By revealing the sources of these biases, the authors proposed a simple solution, namely removing the average seasonal cycle before calculating the extreme thresholds, thus basically eliminating these biases. The paper emphasizes that long running seasonal windows should not be used to estimate extreme events and their impacts without correction.

### Main problems:
1. **Biases of running seasonal windows**: The paper points out that long running seasonal windows (such as 15 - day or 31 - day) used in many studies can lead to systematic biases in the frequency of temperature extreme events. Such biases are particularly evident in regions with strong seasonal changes but weak diurnal variability.
2. **Underestimation of the frequency of extreme events**: These biases lead to a significant underestimation of the frequency of extreme high - temperature events, especially in the transition seasons.
3. **Overestimation of future heat wave changes**: Due to the biases in the historical period, the estimates of future heat wave changes may be overestimated, especially in some regions.

### Solutions:
1. **Removing the average seasonal cycle**: The authors suggest removing the average seasonal cycle before calculating the extreme thresholds, which can effectively eliminate the biases introduced by running seasonal windows.
2. **Using short windows**: As an alternative, shorter running windows (such as 5 - day) can be used, although this may lead to the problem of insufficient sample size.

### Main contributions of the paper:
- **Revealing the mechanism of biases**: It analyzes in detail how running seasonal windows introduce biases and shows how these biases change in different regions and times.
- **Providing solutions**: It proposes the method of removing the average seasonal cycle and proves its effectiveness through experiments.
- **Emphasizing the importance**: It emphasizes the potential risks of using long running seasonal windows without correction and calls on researchers to adopt more rigorous methods when conducting extreme event analysis.

Through these methods, the paper aims to improve the accurate estimation of temperature extreme events, so as to better understand the impacts of climate change.