Anderson et al. (2025) What is a drought-to-flood transition? Pitfalls and recommendations for defining consecutive hydrological extreme events

Identification

• Journal: Hydrology and earth system sciences
• Year: 2025
• Date: 2025-11-07
• Authors: Bailey Anderson, Eduardo Muñoz‐Castro, Lena M. Tallaksen, Alessia Matanó, Jonas Götte, Rachael Armitage, Eugene Magee, Manuela I. Brunner
• DOI: 10.5194/hess-29-6069-2025

Research Groups

• Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland
• WSL Institute for Snow and Avalanche Research SLF, Davos Dorf, Switzerland
• Climate Change, Extremes and Natural Hazards in Alpine Regions Research Center CERC, Davos Dorf, Switzerland
• Department of Geosciences, University of Oslo, Oslo, Norway
• Institute for Environmental Studies (IVM), Vrije Universiteit Amsterdam, Amsterdam, the Netherlands
• UK Centre for Ecology and Hydrology, Wallingford, United Kingdom

Short Summary

This study assesses the suitability and differences of various threshold-level methods for defining drought-to-flood transitions using eight case study catchments, revealing that methodological choices significantly alter detected event characteristics and often fail to capture historically impactful transitions.

Objective

• To assess the suitability and differences of various threshold-level methods for defining drought-to-flood transitions, exploring how methodological choices influence detected event characteristics and their ability to capture historically impactful events.
• To determine how effectively different methods detect documented drought-to-flood transition events in each case study.
• To analyze how seasonality impacts the detection of these events.
• To understand how and why the number and type of transition events detected differ among methods.

Study Configuration

• Spatial Scale: Eight case study catchments across Australia, USA (California, Texas), Chile, England, Italy, Switzerland, and Norway, representing a range of hydrological regimes (monsoonal, snow-dominated, mixed, pluvial, semi-arid, flashy sub-alpine).
• Temporal Scale: Daily mean streamflow data, with lengths varying from 17 to over 40 years (six of eight catchments > 40 years with at least 95% coverage). Sub-daily streamflow data was used for qualitative analysis in the Swiss catchment. Drought events were required to have a minimum duration of 30 days, and transition events were defined by a maximum interval of 90 days (subdivided into rapid at 14 days) between drought end and flood start.

Methodology and Data

• Models used:
  • Threshold Level Method (TLM) for drought and flood detection.
  • Consecutive Dry Period Method (CDPM) combined with TLM for drought detection in intermittent rivers.
  • Generalized Extreme Value (GEV) model fitted using L-moments for estimating the probability distribution of time intervals between events.
• Data sources:
  • Publicly available large sample and national hydrological datasets (e.g., Arpae Emilia-Romagna, USGS, CAMELS-AUS, EStreams, NRFA, Norwegian Water Resources and Energy Directorate, CAMELS-CH, HydroCH, CAMELS-CL).
  • Media, governmental, and scientific reports for identifying and validating historical drought-to-flood case study events and their impacts.
  • Daily mean streamflow time series.
  • Sub-daily streamflow time series (for the Swiss case study).

Main Results

• Only three of the eight media-reported drought-to-flood transition events were consistently detected across all tested methodologies, indicating that current methods often fail to capture impactful real-world events.
• Daily mean streamflow data proved insufficient for detecting rapid, flash flood events in responsive regimes (e.g., the Emme River, Switzerland), where sub-daily data revealed extreme hourly peaks (255 m³ s⁻¹ vs. 17 m³ s⁻¹ daily mean).
• The choice of threshold level method significantly impacts the number and seasonality of detected transition events, particularly in highly seasonal catchments. Seasonal flood thresholds generally increased the number of detected transitions in such regimes.
• Fixed drought thresholds tend to identify events aligned with seasonal norms (e.g., droughts in low-flow seasons, floods in high-flow seasons), while variable thresholds capture anomalies within a season, which may not represent absolute extremes.
• In catchments with weak seasonality, the choice of threshold method has a marginal effect on the number of detected transitions, which are typically high and dispersed throughout the year.
• The probability of a transition occurring within a 14-day window varied substantially across methods and catchments, ranging from 4.1% to 12.7% in Norway and 14.7% to 19.8% in Switzerland, highlighting that a fixed time window may not reflect event extremity consistently.

Contributions

• This study provides the first comprehensive assessment of the suitability and differences of various threshold-level methods for defining consecutive drought-to-flood transition events.
• It critically demonstrates that commonly used methodologies often fail to detect historically impactful drought-to-flood transitions, highlighting a significant gap in current research approaches.
• The research elucidates the critical influence of methodological choices (e.g., threshold type, time interval, data resolution) on the characteristics, number, and seasonality of detected transition events across diverse hydrological regimes.
• It offers practical recommendations for improving transition detection strategies, including applying secondary filtering steps, adopting probabilistic definitions, and explicitly considering the specific impacts and drivers of interest.
• The paper emphasizes the necessity for location-specific approaches and calls for further research into the complex interactions between droughts and floods that lead to amplified impacts.

Funding

• Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung (grant no. 200021_214907)

Citation

@article{Anderson2025What,
  author = {Anderson, Bailey and Muñoz‐Castro, Eduardo and Tallaksen, Lena M. and Matanó, Alessia and Götte, Jonas and Armitage, Rachael and Magee, Eugene and Brunner, Manuela I.},
  title = {What is a drought-to-flood transition? Pitfalls and recommendations for defining consecutive hydrological extreme events},
  journal = {Hydrology and earth system sciences},
  year = {2025},
  doi = {10.5194/hess-29-6069-2025},
  url = {https://doi.org/10.5194/hess-29-6069-2025}
}