Towards operational post-processing of probabilistic temperature forecasts at MeteoSwiss

2020 ◽  
Author(s):  
Jan Rajczak ◽  
Regula Keller ◽  
Jonas Bhend ◽  
Christoph Spirig ◽  
Stephan Hemri ◽  
...  
Keyword(s):  
Substantial Improvement ◽  
Model Output ◽  
Post Processing ◽  
Boundary Layer Height ◽  
Medium Range Forecast ◽  
Direct Model ◽  
Medium Range ◽  
Ensemble Model Output Statistics ◽  
Model Output Statistics ◽  
Gaussian Regression

<p>MeteoSwiss is currently developing a post-processing suite for the territory of Switzerland. The system aims to provide optimized multi-variable (i.e. temperature, precipitation, wind and cloud cover), spatial and probabilistic predictions. The system will combine information in a seamless manner from the in-house short range and regional (COSMO-E/1) of 1 resp. 2 km resolution and the medium range ECMWF IFS NWP systems. At the example of probabilistic temperature forecasts, this contribution discusses recent advances and experiences at developing, applying and operationalizing non-homogenous Gaussian regression, also known as ensemble model output statistics (EMOS).</p><p>Over the complex terrain of Switzerland, postprocessing leads to a substantial improvement of temperature forecasts by up to 30% in terms of CRPS with respect to elevation-corrected direct model output (DMO) even by a basic EMOS only relying on DMO of temperature. Incorporating suitable predictors, such as the atmospheric boundary layer height, leads to a further gain in forecast quality. Results also show that combining high- (COSMO-E) and coarse-resolution (IFS) NWP output can not only provide a seamless medium-range forecast, but also further increase prediction skill during the time horizon when both models are available. Finally, we discuss first attempts to produce high-resolution spatial PP fields for arbitrary locations by exploiting a global EMOS framework with multiple static (e.g. geographic characteristics) and dynamic predictors derived from NWP data.</p>

scholarly journals Calibrated ensemble forecasts of the height of new snow using quantile regression forests and ensemble model output statistics

2021 ◽  
Vol 28 (3) ◽  
pp. 467-480
Author(s):  
Guillaume Evin ◽  
Matthieu Lafaysse ◽  
Maxime Taillardat ◽  
Michaël Zamo
Keyword(s):  
Quantile Regression ◽  
Specific Humidity ◽  
Ensemble Model ◽  
Model Output ◽  
Post Processing ◽  
Ensemble Forecasts ◽  
Mountain Areas ◽  
Processing Methods ◽  
Ensemble Model Output Statistics ◽  
Model Output Statistics

Abstract. Height of new snow (HN) forecasts help to prevent critical failures of infrastructures in mountain areas, e.g. transport networks and ski resorts. The French national meteorological service, Météo-France, operates a probabilistic forecasting system based on ensemble meteorological forecasts and a detailed snowpack model to provide ensembles of HN forecasts. These forecasts are, however, biased and underdispersed. As for many weather variables, post-processing methods can be used to alleviate these drawbacks and obtain meaningful 1 to 4 d HN forecasts. In this paper, we compare the skill of two post-processing methods. The first approach is an ensemble model output statistics (EMOS) method, which can be described as a nonhomogeneous regression with a censored shifted Gamma distribution. The second approach is based on quantile regression forests, using different meteorological and snow predictors. Both approaches are evaluated using a 22 year reforecast. Thanks to a larger number of predictors, the quantile regression forest is shown to be a powerful alternative to EMOS for the post-processing of HN ensemble forecasts. The gain of performance is large in all situations but is particularly marked when raw forecasts completely miss the snow event. This type of situation happens when the rain–snow transition elevation is overestimated by the raw forecasts (rain instead of snow in the raw forecasts) or when there is no precipitation in the forecast. In that case, quantile regression forests improve the predictions using the other weather predictors (wind, temperature, and specific humidity).

scholarly journals Calibrated ensemble forecasts of the height of new snow using quantile regression forests and Ensemble Model Output Statistics

2021 ◽  
Author(s):  
Guillaume Evin ◽  
Matthieu Lafaysse ◽  
Maxime Taillardat ◽  
Michaël Zamo
Keyword(s):  
Quantile Regression ◽  
Specific Humidity ◽  
Ensemble Model ◽  
Model Output ◽  
Post Processing ◽  
Ensemble Forecasts ◽  
Mountain Areas ◽  
Processing Methods ◽  
Ensemble Model Output Statistics ◽  
Model Output Statistics

Abstract. Height of new snow (HN) forecasts help to prevent critical failures of infrastructures in mountain areas, e.g. transport networks, ski resorts. The French national meteorological service, Meteo-France, operates a probabilistic forecasting system based on ensemble meteorological forecasts and a detailed snowpack model to provide ensembles of HN forecasts. These forecasts are however significantly biased and underdispersed. As for many weather variables, post-processing methods can be used to alleviate these drawbacks and obtain meaningful 1-day to 4-day HN forecasts. In this paper, we compare the skill of two post-processing methods. The first approach is an ensemble model output statistics (EMOS) method, which can be described as a Nonhomogeneous Regression with a Censored Shifted Gamma distribution. The second approach is based on quantile regression forests, using different meteorological and snow predictors. Both approaches are evaluated using a 22-year reforecast. Thanks to a larger number of predictors, the quantile regression forest is shown to be a powerful alternative to EMOS for the post-processing of HN ensemble forecasts. The gain of performance is important in all situations but is particularly marked when raw forecasts completely miss the snow event. This type of situations happens when the rain-snow transition elevation is overestimated by the raw forecasts (rain instead of snow in the raw forecasts) or when there is no precipitation in the forecast. In that case, quantile regression forests improve the predictions using the other weather predictors (wind, temperature, specific humidity).

scholarly journals Multimodel probabilistic prediction of 2 m-temperature anomalies on the monthly timescale

2017 ◽  
Vol 14 ◽  
pp. 123-129 ◽  
Author(s):  
Alfonso Ferrone ◽  
Daniele Mastrangelo ◽  
Piero Malguzzi
Keyword(s):  
Logistic Regression ◽  
Model Output ◽  
Probabilistic Prediction ◽  
Temperature Anomalies ◽  
Skill Scores ◽  
Direct Model ◽  
Prediction Systems ◽  
Model Output Statistics ◽  
Gaussian Regression

Abstract. The 2 m-temperature anomalies from the reforecasts of the CNR-ISAC and ECMWF monthly prediction systems have been combined in a multimodel super-ensemble. Tercile probability predictions obtained from the multimodel have been constructed using direct model outputs (DMO) and model output statistics (MOS), like logistic and nonhomogeneous Gaussian regression, for the 1990–2010 winter seasons. Verification with ERA-Interim reanalyses indicates that logistic regression gives the best results in terms of ranked probability skill scores (RPSS) and reliability diagrams for low–medium forecast probabilities. Also, it is argued that the logistic regression would not yield further improvements if a larger dataset was used.

Will post-processing always improve my forecasts?

2020 ◽  
Author(s):  
Jon Olav Skøien ◽  
Peter Salamon ◽  
Fredrik Wetterhall
Keyword(s):  
River Basin ◽  
Ensemble Model ◽  
Model Output ◽  
Post Processing ◽  
Continental Scale ◽  
Post Process ◽  
Basin Scale ◽  
Verification Period ◽  
Ensemble Model Output Statistics ◽  
Model Output Statistics

<p>Different statistical techniques are frequently employed to post-process the outcome of ensemble forecasting models. The main reason is to compensate for biases due to errors in model structure or initial conditions, and as a correction for under- or overdispersed ensembles.</p><p>Here we present analyses of the results from one these methods. We use the Ensemble Model Output Statistics method (EMOS; Gneiting et al., 2005) to post-process the ensemble output from a continental scale hydrological model - LISFLOOD (Van Der Knijff et al., 2010; De Roo et al., 2000). The model was calibrated at approximately 700 stations based on long term observations of runoff and meteorological variables. We use the same locations for calibration and verification of the 1-10 days forecasts of the model, based on ensemble and deterministic meteorological forecasts from ECMWF (51 ensemble members + 1 high-resolution), DWD (1 member) and COMSO-LEPS (16 ensemble members).</p><p>We calibrated the EMOS-parameters using the Continuous ranked probability score (CRPS). Whereas the post-processing improved the results for the first 1-2 days lead time, the improvement was less for increasing lead times of the verification period. As the post-processing is based on assumptions about the forecast errors, we will here present analyses of the ensemble output that can give some indications of what to expect from the post-processing.</p><p> </p><p>Gneiting, T., Raftery, A. E., Westveld, A. H. and Goldman, T.: Calibrated Probabilistic Forecasting Using Ensemble Model Output Statistics and Minimum CRPS Estimation, Mon. Weather Rev., 133(5), 1098–1118, doi:10.1175/MWR2904.1, 2005.</p><p>Van Der Knijff, J. M., Younis, J. and De Roo, A. P. J.: LISFLOOD: a GIS‐based distributed model for river basin scale water balance and flood simulation, Int. J. Geogr. Inf. Sci., 24(2), 189–212, doi:10.1080/13658810802549154, 2010.</p><p>De Roo, A. P. J., Wesseling, C. G. and Van Deursen, W. P. A.: Physically based river basin modelling within a GIS: The LISFLOOD model, in Hydrological Processes, vol. 14, pp. 1981–1992, John Wiley & Sons Ltd. [online] Available from: http://www.scopus.com/inward/record.url?eid=2-s2.0-0034254644&partnerID=tZOtx3y1, 2000.</p><p> </p>

LQ-Warn – Entwicklung optimierter Vorhersagen verschiedener Luftqualitätsparameter

2021 ◽  
Author(s):  
Sabine Robrecht ◽  
Robert Osinski ◽  
Ute Dauert ◽  
Andreas Lambert ◽  
Stefan Gilge ◽  
...  
Keyword(s):  
Root Mean Square Error ◽  
Mean Square Error ◽  
Root Mean Square ◽  
Model Output ◽  
Mean Square ◽  
Post Processing ◽  
Eine Hohe ◽  
Erste Ergebnisse ◽  
Model Output Statistics ◽  
Monitoring Service

<p>Schlechte Luftqualität gefährdet die Gesundheit der Bevölkerung. Zur Information und zur Ergreifung kurzfristiger Maßnahmen zur Luftqualitätsverbesserung (z.B. Verkehrslenkung) ist eine möglichst genaue und – insbesondere in städtischen Gebieten – möglichst räumlich hochaufgelöste Luftqualitätsvorhersage notwendig. Numerische Luftqualitätsmodelle haben für diese Aufgabe in der Regel eine zu geringe räumliche Auflösung.</p> <p>Daher ist es Ziel des Projektes „LQ-Warn“ die Luftqualitätsvorhersage insbesondere im Hinblick auf die Überschreitung von Grenzwerten zu verbessern. Basierend auf den Modellergebnissen für Luftqualitätsparameter des Copernicus Atmospheric Monitoring Service (CAMS) werden zwei Ansätze verfolgt: Einerseits werden Vorhersagen mit dem regionalen chemischen Transportmodell „REM-CALGRID“ (RCG) unter Einbeziehung von CAMS-Ergebnissen und regionalen Emissionsdaten berechnet. Dabei kann eine hohe horizontale Auflösung von 2 km erzielt werden und Prognosen können für verschiedene Luftschadstoffe in stündlicher Auflösung mit bis zu 72 Stunden Vorlaufzeit erstellt werden, unter anderem für NO<sub>2</sub>, O<sub>3</sub>, PM<sub>10</sub> und PM<sub>2.5</sub>. Andererseits wird die statistische Post-Processing-Methode „Model Output Statistics“ (MOS) angewandt, um Punktvorhersagen für die Massenkonzentration der Spezies NO<sub>2</sub>, O<sub>3</sub>, PM<sub>10</sub> und PM<sub>2.5</sub> mit einer Vorlaufzeit von bis zu 96 Stunden zu berechnen. Dafür werden luftqualitätsbezogene Messungen, CAMS-Modellergebnisse und meteorologische Parameter aus dem numerischen Wettervorhersagemodell des ECMWF als Prädiktoren verwendet.</p> <p>Es werden erste Ergebnisse der mit den o.g. Ansätzen errechneten Vorhersagen präsentiert und die Vor- und Nachteile der jeweiligen Verfahren hervorgehoben. Durch die statistische Post-Processing-Methode MOS wird an den Vorhersagepunkten vor allem für die Massenkonzentration von O<sub>3 </sub>und NO<sub>2</sub> eine signifikante Verringerung des RMSE (Root Mean Square Error) im Vergleich zu den Vorhersagen des numerischen CAMS-Modells erreicht. Diese deutliche Verbesserung der Luftqualitätsvorhersage sinnvoll auf die Fläche auszudehnen ist jedoch noch eine Herausforderung. Im Gegensatz dazu zeigt die Vorhersage mit dem RCG-Modell eine geringere Verbesserung der Vorhersagegüte an einzelnen Vorhersagepunkten als der MOS-Ansatz. Stattdessen bietet das RCG-Modell zeitlich und räumlich konsistente Vorhersagen an allen Modellgitterpunkten. Kleinskalige Konzentrationsunterschiede können aufgrund der höheren Modellauflösung deutlich realistischer vorhergesagt werden als mit den CAMS-Vorhersagen. Ein weiterführendes Ziel des LQ-Warn-Projektes ist es die beiden Ansätze zu kombinieren, um die Vorteile beider nutzen zu können und eine präzise Luftqualitätsvorhersage flächendeckend für Deutschland bereitstellen zu können.</p>

Sign in / Sign up

Export Citation Format

Share Document