HighResMIP models across Europe Gerard van der Schrier Antonello Squintu and Primavera team Record high temperature in Europe in 2019 28 April 2020 Royal Netherlands Meteorological Institute ID: 1024355
Download Presentation The PPT/PDF document "28 April 2020 Evaluation of trends in ex..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
1. 28 April 2020Evaluation of trends in extreme temperatures simulated by HighResMIP models across EuropeGerard van der Schrier, Antonello Squintu and Primavera team
2. Record high temperature in Europe in 201928 April 2020Royal Netherlands Meteorological Institutesource: Getty Imagessource: annual State of the Climate, C3SRanking of 2019 highest daily max temperature The Europe-averaged daily max temperature breaks more heat records than theoretically expected in a stationary cimate
3. Can climate models reproduce the trend toward more hot extremes? 28 April 2020Royal Netherlands Meteorological Institutesource: Min et al. (2013)ExampleComparison of model trends against observationsMetric: warmest day in the year (TXx)Dataset with gridded station observationsConclusion: models strongly under estimate trends in hot extremes
4. But is this a fair comparison?28 April 2020Royal Netherlands Meteorological InstituteMinimum of minimum temperatureVery localE-OBSv20.0e, 0.1°Yearday of maximum of maximum temperatureTiming is not homogeneous at allInstead of TXx/TNn, use TX90p/TN10p
5. Observed Data may have problems tooRelocation of stations, often from the city center to the airport (urban heat island effect removed) or with change of altitude.Change in the instrumental features (new screen, manual to automatic, analog to digital, etc.)Gradual changes of the surrounding (growing vegetation, expansion of urban area)Step-like signals in series introduced by: 420 m405 m~6kmuntil 1961Geneva, Switzerlandsince 1962
6. HOMOGENIZATIONTN 10p of Geneva (blended), original and homogenized with running mean. The lower band indicates the donating series.Homogenized versionOriginalPortion coming from Geneva Cointrin (Airport)Portion coming from Geneva Observatoire (City)Adjustments calculated via Quantile Matching betweenAirport and City pdf, with the help of hom. ref. series GENEVA OBS.+AIRPORT
7. Effects of homogenization on TX trends in Europe28 April 2020Royal Netherlands Meteorological Institutebeforeafter
8. 28 April 2020Royal Netherlands Meteorological Institutebeforeafter
9. How well do the Primavera models?28 April 2020Royal Netherlands Meteorological InstituteCMCCECMWFCNRMHadGEM3EC-EarthMPISimple biassummer average of TX
10. Summer TX trend (1961-2010)28 April 2020Royal Netherlands Meteorological InstituteHow well do the Primavera models?CMCCECMWFCNRMHadGEM3EC-EarthMPI
11. Trends in summer TX90p (1961-2010)28 April 2020Royal Netherlands Meteorological InstituteHow well do the Primavera models?CMCCECMWFCNRMHadGEM3EC-EarthMPI
12. Comparison between HR and LR28 April 2020Royal Netherlands Meteorological InstituteRegrid the LR to the HR grid (nearest neighbour)Calculate: | - | - | - |If metric < 0: HR trend bias is smaller than LRIf metric > 0: HR trend bias is larger than LR
13. 28 April 2020Royal Netherlands Meteorological InstituteImprovement between HR and LR?CMCCECMWFCNRMHadGEM3EC-EarthMPI
14. Some observations28 April 2020Royal Netherlands Meteorological InstituteBias in TX: CMCC & ECMWF cold bias, others have a warm bias in NW and a cold bias in SEBias in TX summer trends: quite ok, ECMWF & EC-Earth underestimate trends in Italy, Adriatic coast and over the Ukraine, HadGem underestimates trends in ScandinaviaBias in hot extremes trends: all models show NW overestimation/SW underestimation. For some models: entire Mediterrean too weak trends in hot extremesSome improvement in HR wrt LRIs it possible that the warm bias in the Mediterranean of most models gives a negative bias in soil moisture, which then produces too weak trends in hot extremes? Cold extremes: generally models have too strong warming trends. Problems with snow dynamics?