Abstract. Daily meteorological data such as temperature or precipitation from climate models is needed for many climate impact studies, e.g. in hydrology or agriculture but direct model output can contain large systematic errors. Thus, statistical bias adjustment is applied to correct climate model outputs. Here we review existing statistical bias adjustment methods and their shortcomings, and present a method which we call EQA (Empirical Quantile Adjustment), a development of the methods EDCDFm and PresRAT. We then test it in comparison to two existing methods using real and artificially created daily temperature and precipitation data for Austria. We compare the performance of the three methods in terms of the following demands: (1): The model data should match the climatological means of the observational data in the historical period. (2): The long-term climatological trends of means (climate change signal), either defined as difference or as ratio, should not be altered during bias adjustment, and (3): Even models with too few wet days (precipitation above 0.1 mm) should be corrected accurately, so that the wet day frequency is conserved. EQA fulfills (1) almost exactly and (2) at least for temperature. For precipitation, an additional correction included in EQA assures that the climate change signal is conserved, and for (3), we apply another additional algorithm to add precipitation days.

Abstract. Daily meteorological data from climate models is needed for many climate impact studies, e.g. in hydrology or agriculture but direct model output can contain large systematic errors. Thus, statistical bias correcting is applied to correct the raw model data. However, up to now no method has been introduced that fulfills the following demands simultaneously: (1) The long term climatological trends (climate change signal) should not be altered during bias correction, (2) the model data should match the observational data in the historical period as accurate as possible in a climatological sense and (3) models with too little wet days (precipitation above 0.1 mm) should be corrected accurately, which means that the wet day frequency is conserved. We improve the already existing quantile mapping approach so that it satisfies all three conditions. Our new method is called empirical percentile–percentile mapping (EPPM) which uses empirical distributions for meteorological variables and is therefore computationally inexpensive. The correction of precipitation is particularly challenging so our main focus is on precipitation. EPPM corrects the historical model data so that precipitation sums and wet days are equal to the observational data.

Abstract. Daily meteorological data such as temperature or precipitation from climate models are needed for many climate impact studies, e.g., in hydrology or agriculture, but direct model output can contain large systematic errors. A large variety of methods exist to adjust the bias of climate model outputs. Here we review existing statistical bias-adjustment methods and their shortcomings, and compare quantile mapping (QM), scaled distribution mapping (SDM), quantile delta mapping (QDM) and an empiric version of PresRAT (PresRATe). We then test these methods using real and artificially created daily temperature and precipitation data for Austria. We compare the performance in terms of the following demands: (1) the model data should match the climatological means of the observational data in the historical period; (2) the long-term climatological trends of means (climate change signal), either defined as difference or as ratio, should not be altered during bias adjustment; and (3) even models with too few wet days (precipitation above 0.1 mm) should be corrected accurately, so that the wet day frequency is conserved. QDM and PresRATe combined fulfill all three demands. For (2) for precipitation, PresRATe already includes an additional correction that assures that the climate change signal is conserved.

Interactive comment on "An improved statistical bias correction method that also corrects dry climate models" by Fabian Lehner et al.Fabian Lehner et al.

<p>Im Zuge des Klimawandels ändern sich die Baumarten, die für das jeweilige Klima als passend angesehen werden ("Klimafitter Wald"). Da für die Eignung der Baumarten die Klimatologie relevant ist, braucht es eine Vielzahl an meteorologischen Daten in einer hohen räumlichen Auflösung und teilweise auf Tagesbasis, nicht nur für die historische Periode, sondern auch für ausgewählte Szenarien aus dem Ensemble der RCP 4.5 und 8.5 EUROCORDEX-Regionalmodelle.</p> <p>Benötigt werden Tageszeitreihen aus den Parametern Höchsttemperatur, Tiefsttemperatur, Niederschlag, Globalstrahlung, Luftfeuchtigkeit und mittlere Windgeschwindigkeit. Als Datengrundlage für vergangenes Klima dienen einerseits, soweit vorhanden, Gitterdatensätze der ZAMG (SPARTACUS und APOLIS), teilweise zusätzlich mit Stationsdaten kombiniert und andererseits selbst erstellte Gitterdatensätze direkt aus Stationsdaten. Die tägliche, räumliche Interpolation auf ein 200x200 m Gitter erfolgt vor allem mit linearen Regressionen von Variablen, die von der Topographie abgeleitet werden.</p> <p>Die Klimaszenarien werden mit statistischen Methoden fehlerkorrigiert und lokalisiert, die auf den Methoden PresRAT (Pierce et al., 2015) und EDCDFm (Li et al., 2010) basieren. Mit diesen Methoden werden die statistischen Verteilungsfunktionen der Variablen so angepasst, dass sie im historischen Zeitraum für jeden Gitterpunkt den Beobachtungsdaten entsprechen.</p> <p>Aus den Tagesdaten werden im letzten Schritt auch Klimaindikatoren berechnet, die mit lokaler linearer Regression in Abhängigkeit von der Seehöhe auf das finale Gitter mit 10x10 m Auflösung interpoliert werden. Beispiele für die Klimaindikatoren sind die Länge der Vegetationsperiode, die Anzahl der Hitzetage oder Frosttage oder die klimatische Wasserbilanz.</p>