Tramblay et al. (2018) Future evolution of extreme precipitation in the Mediterranean

Identification

• Journal: Climatic Change
• Year: 2018
• Date: 2018-09-25
• Authors: Yves Tramblay, Samuel Somot
• DOI: 10.1007/s10584-018-2300-5

Research Groups

• Universidad de León (ULE), Atmospheric Physics Group (GFA), Environmental Institute, León, Spain
• University of Castilla-La Mancha (UCLM), Institute of Environmental Sciences (ICAM), Faculty of Environmental Sciences and Biochemistry, Earth and Space Sciences (ESS) Group, Toledo, Spain

Short Summary

This study evaluates the hourly precipitation performance of the Weather Research and Forecasting (WRF) model for 45 extreme precipitation events (EPEs) in northeastern Spain, classifying events by their hourly distribution and testing various microphysics and planetary boundary layer (PBL) parameterizations. It finds that microphysics schemes are more critical than PBL schemes, with Goddard and Thompson microphysics combined with MYNN PBL generally performing better, though no single optimal combination exists for all event types.

Objective

• To appraise the WRF model's hourly precipitation performance for 45 extreme precipitation events (EPEs) in northeastern Spain.
• To design a WRF ensemble using three microphysics and two planetary boundary layer (PBL) schemes for each EPE.
• To classify EPEs based on their hourly precipitation distribution to study model behavior for different precipitation types.
• To investigate WRF hourly precipitation performance by cluster using statistical goodness-of-fit measures and density plots.

Study Configuration

• Spatial Scale: Ebro Basin, northeastern Spain. Three nested domains with horizontal resolutions of 27 km, 9 km, and 3 km. Observations from 367 rain gauge stations distributed across the basin.
• Temporal Scale: 45 extreme precipitation events (EPEs) selected from 2010 to 2018. Each simulation covered 30 hours (24 hours for the event plus a 6-hour spin-up period). Hourly precipitation assessment.

Methodology and Data

• Models used:
  • Weather Research and Forecasting (WRF) model version 4.1 (non-hydrostatic Advanced Research WRF).
  • Microphysics schemes: Goddard, Thompson, Morrison 6-class double moment.
  • Planetary Boundary Layer (PBL) schemes: Mellor-Yamada-Janjic (MYJ), Mellor-Yamada Nakanishi and Nino Level 2.5 (MYNN).
  • Other physics schemes: Dudhia (shortwave radiation), Rapid Radiative Transfer Model (longwave radiation), Eta surface layer, Noah Land Surface Model, Kain-Fritsch cumulus scheme (for the inner domain).
• Data sources:
  • Initial and boundary conditions: National Oceanic and Atmospheric Administration / National Centers for Environmental Prediction Global Forecasting Model reanalysis (0.25° horizontal grid spacing, 6-hour intervals).
  • Observational data: 367 tipping-bucket rain gauge stations in the Ebro Basin, providing 10-minute data (hourly and daily precipitation retrieved) from 2010 to 2018. Data quality control performed using the R package reddprec.
  • EPE definition: Days where total precipitation exceeded the station-specific 99th percentile of daily precipitation on "wet" days (precipitation ≥ 1 mm), with at least 10 stations meeting this criterion.
  • Evaluation metrics: Modified Kling-Gupta efficiency (KGE) and its components (Pearson correlation 'r', bias ratio 'β', variability ratio 'γ'), Kolmogorov-Smirnov (K-S) test (D statistic and p-value), and density plots for precipitation intensity, temporal distribution, and spatial location of maximum precipitation.

Main Results

• The 45 EPEs (965 station-events) were classified into 5 distinct clusters based on their observed hourly precipitation distribution, representing different event types (e.g., short intense convective, prolonged stratiform).
• WRF model performance for hourly precipitation verification against point observations showed generally poor results (low KGE values), indicating significant challenges in accurately simulating EPEs at sub-daily scales.
• Microphysics parameterizations were found to be more important than PBL schemes in influencing model performance for hourly EPEs.
• The Goddard and Thompson microphysics schemes, particularly when combined with the MYNN PBL scheme, generally yielded better results across most analyzed characteristics (KGE, precipitation amount, temporal distribution, and spatial location of maximum precipitation).
• The Morrison microphysics scheme consistently showed the poorest performance among those tested.
• The model exhibited a tendency to underestimate hourly precipitation (bias ratio β < 1) and to smooth the temporal distribution of precipitation, spreading it over a longer duration than observed.
• Stratiform events (Cluster 5) and organized mesoscale convective events (Cluster 2) were better simulated by WRF compared to highly localized convective events (Cluster 4).
• No single optimal parameterization combination was identified for all EPE types, as model performance was highly dependent on the specific characteristics of the event.
• Spatial deviations between modeled and observed maximum precipitation were typically less than 10 km for stratiform rainfall but larger for convective events.

Contributions

• This study provides the first comprehensive hourly precipitation verification of the WRF model for a large dataset of 45 extreme precipitation events (EPEs) in northeastern Spain.
• It introduces a novel classification of EPEs based on their hourly precipitation distribution, allowing for a differentiated analysis of model performance across various rainfall types (convective vs. stratiform).
• The research employs a robust evaluation methodology combining traditional statistical goodness-of-fit measures (KGE, K-S test) with advanced density plots to assess precipitation intensity, temporal distribution, and spatial location of maximum precipitation.
• It quantitatively demonstrates the greater influence of microphysics parameterizations compared to planetary boundary layer schemes in simulating hourly EPEs in the study region.
• The findings offer critical insights into the strengths and limitations of specific WRF parameterization combinations for different EPE characteristics, highlighting ongoing challenges in sub-daily EPE forecasting and providing guidance for future model improvements.

Funding

• Project LE240P18 (Consejería de Educación, Junta de Castilla y León)
• Project CGL2016-78702-C2-1-R (Ministerio de Economía y Competitividad)
• Project PID2019-108470RB-C22 (Ministerio de Economía y Competitividad)
• Project PID2019-108470RB-C21 (Ministerio de Economía y Competitividad)

Citation

@article{Tramblay2018Future,
  author = {Tramblay, Yves and Somot, Samuel},
  title = {Future evolution of extreme precipitation in the Mediterranean},
  journal = {Climatic Change},
  year = {2018},
  doi = {10.1007/s10584-018-2300-5},
  url = {https://doi.org/10.1007/s10584-018-2300-5}
}