Evaluation of precipitation forecasting with the limited area model ALADIN in an alpine watershed
Bodo Ahrens
Abstract
Bodo Ahrens
Abstract
The applicability of high-resolution precipitation forecasts in hydrological prediction systems for small- to medium-sized Alpine watersheds has to be evaluated. Here, this is performed for precipitation forecasts simulated by the operational limited area model (LAM) ALADIN (grid spacing Δx = 10 km) in one Alpine watershed (total area: 2627 km2). The impact of simple post-processing of the model forecast fields by up-, downscaling, and by grid box shifting is quantified. The forecasts are directly evaluated against daily and hourly precipitation analyses and indirectly by application in hydrological modeling with WaSiM-ETH (Δx = 500 m, Δt = 1 h). The application of a hydrological model makes precipitation comparable with well observed watershed discharge. Main results are: (a) LAM precipitation forecasts are useful in Alpine watershed hydrology, but (b) there is intermittent failure of precipitation forecasts within the scales of the meteorological forecast system, and (c) limited spatial grid resolution of the precipitation fields generates substantial uncertainties in peak precipitation and runoff (10% and more). Die Anwendbarkeit hochaufgelöster Niederschlagsvorhersagen in hydrologischen Prognosesystemen für Einzugsgebiete kleiner und mittlerer Größe muß evaluiert werden. In dieser Arbeit wird dies für Vorhersagen des operationellen Ausschnittsmodells ALADIN (Gitterpunktsabstand 10 km) in einem alpinen Einzugsgebiet (Fläche: 2627 km2) durchgeführt. Die Auswirkung einfacher Nachbearbeitung der Modellvorhersagen mittels up-, downscaling und Gitterboxverschiebung wird quantifiziert. Die Vorhersagen werden direkt evaluiert im Vergleich mit täglichen und stündlichen Niederschlagsanalysen und indirekt durch Anwendung des hydrologischen Modells WaSiM-ETH (Δx = 500 m, Δt = 1 h). Die Anwendung eines hydrologischen Modells macht Niederschläge mit beobachteten Gebietsabflüssen vergleichbar. Wichtigste Ergebnisse sind: (a) die Niederschlagsprognose des Ausschnittsmodells ist nützlich in der Hydrologie alpiner Einzugsgebiete, aber (b) Niederschlag ist zeitweise auf den Skalen des meteorologischen Systems schlecht vorhergesagt, und (c) die limitierte räumliche Auflösung der Niederschlagsprognosen generiert substantielle Fehler in Niederschlagsund Abflußspitzen (10% und mehr).
OpenAlex reports 15 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
The applicability of high-resolution precipitation forecasts in hydrological prediction systems for small- to medium-sized Alpine watersheds has to be evaluated. Here, this is performed for precipitation forecasts simulated by the operational limited area model (LAM) ALADIN (grid spacing Δx = 10 km) in one Alpine watershed (total area: 2627 km2). The impact of simple post-processing of the model forecast fields by up-, downscaling, and by grid box shifting is quantified. The forecasts are directly evaluated against daily and hourly precipitation analyses and indirectly by application in hydrological modeling with WaSiM-ETH (Δx = 500 m, Δt = 1 h). The application of a hydrological model makes precipitation comparable with well observed watershed discharge. Main results are: (a) LAM precipitation forecasts are useful in Alpine watershed hydrology, but (b) there is intermittent failure of precipitation forecasts within the scales of the meteorological forecast system, and (c) limited spatial grid resolution of the precipitation fields generates substantial uncertainties in peak precipitation and runoff (10% and more). Die Anwendbarkeit hochaufgelöster Niederschlagsvorhersagen in hydrologischen Prognosesystemen für Einzugsgebiete kleiner und mittlerer Größe muß evaluiert werden. In dieser Arbeit wird dies für Vorhersagen des operationellen Ausschnittsmodells ALADIN (Gitterpunktsabstand 10 km) in einem alpinen Einzugsgebiet (Fläche: 2627 km2) durchgeführt. Die Auswirkung einfacher Nachbearbeitung der Modellvorhersagen mittels up-, downscaling und Gitterboxverschiebung wird quantifiziert. Die Vorhersagen werden direkt evaluiert im Vergleich mit täglichen und stündlichen Niederschlagsanalysen und indirekt durch Anwendung des hydrologischen Modells WaSiM-ETH (Δx = 500 m, Δt = 1 h). Die Anwendung eines hydrologischen Modells macht Niederschläge mit beobachteten Gebietsabflüssen vergleichbar. Wichtigste Ergebnisse sind: (a) die Niederschlagsprognose des Ausschnittsmodells ist nützlich in der Hydrologie alpiner Einzugsgebiete, aber (b) Niederschlag ist zeitweise auf den Skalen des meteorologischen Systems schlecht vorhergesagt, und (c) die limitierte räumliche Auflösung der Niederschlagsprognosen generiert substantielle Fehler in Niederschlagsund Abflußspitzen (10% und mehr).
Key concepts: Precipitation, Watershed, Environmental science, Surface runoff, Snow, Quantitative precipitation forecast, Meteorology, Hydrological modelling