On the choice of reference database and calibration period of bias-corrected simulations: A case study for Hungary 3Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21.DOI: 10.15201/hungeobull.74.1.1 Hungarian Geographical Bulletin 74 2025 (1) 3–21. Introduction Climate models have become key tools for climate research, providing not only infor- mation on past and present climate, but also numerical estimates of climate change (IPCC, 2013). General Circulation Models (GCMs) operate at a coarser horizontal resolution (100–500 km), therefore, they are unable to resolve complex topographical features that vary at finer scales. Regional climate models (RCMs), in contrast, are applied only to a lim- ited area with a higher (10–50 km) horizontal resolution, thus, representing extreme events with higher accuracy and providing added value, especially in regions with complex to- pography (Torma, Cs.Zs. et al. 2015, 2020; Di Luca, A. et al. 2016; Rummukainen, M. 2016; Fantini, A. et al. 2018; Ciarlo, J.M. et al. 2021). However, it is important to keep in mind that GCM and RCM simulations are encum- bered with uncertainties from a variety of sources (Giorgi, F. 2005), thus, using raw RCM simulations can lead to unrealistic results. These uncertainties can be quantified and re- duced by using bias-adjusted datasets and by evaluating several RCMs together, as mem- bers of an ensemble (Beniston, M. et al. 2007). 1 Department of Meteorology, Institute of Geography and Earth Sciences, ELTE Eötvös Loránd University, Pázmány Péter sétány 1/A, H-1117 Budapest, Hungary. E-mails: csilluss58@student.elte.hu, csaba.zsolt.torma@ttk.elte.hu, kiaqagt@staff.elte.hu On the choice of reference database and calibration period of bias-corrected simulations: A case study for Hungary Csilla SIMON1, Csaba Zsolt TORMA 1 and Anna KIS1 Abstract The aim of the present study is to investigate the accuracy of bias-adjusted regional climate model (RCM) simulations using various calibration periods, demonstrated for the region of Hungary. High-resolution (0.11°) RCM simulations of daily near-surface mean air temperature, daily minimum and maximum air temperature, and daily precipitation provided by the EURO-CORDEX community are analysed. The model ensemble consists of 5 RCM simulations driven by 4 different general circulation models for the historical time period 1976–2005. The publicly available, most accurate, measurement-based and quality-controlled HuClim is used as the reference dataset. The internationally widely used percentile-based quantile mapping method is applied for the bias-correction and it is performed on a monthly level. The novelty of the present study is that we used two different calibration periods to create bias-corrected datasets: an earlier and a more recent 30-year long period, and made these new datasets available in Zenodo. In addition to these HuClim-based bias-corrected databases, another database, containing bias-corrected RCM simulations and produced by the EURO-CORDEX community is also investigated. The assessment is carried out for the period 1993–2005, which is the overlapping time interval of the different calibration periods. According to our results, the accuracy of the bias-correction depends on the chosen calibration period and on the analysed climate index, and the choice of the validation period also affects the results. As next step, we plan to extend our research on projections under RCP4.5 and RCP8.5 scenarios. Keywords: EURO-CORDEX, HuClim, bias-correction, calibration period, validation, Hungary Received October 2024, accepted February 2025. mailto:csilluss58@student.elte.hu mailto:csaba.zsolt.torma@ttk.elte.hu Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21.4 The systematic bias of a climate model can be eliminated by post-processing the raw RCM data by applying a bias-correction method, which involves ensuring of equal mean val- ues between the observation-based reference dataset and the bias-corrected climate model simulations (Déqué, M. et al. 2007). Previous studies have confirmed that bias-correction is required to improve the quality of RCM simu- lations (e.g. Ngai, S.T. et al. 2016; Jaiswal, R. et al. 2022) – it is particularly important, when RCM simulations are used for impact studies (e.g. wind energy generation: Costoya, X. et al. 2020; hydrology: Faghih, M. et al. 2022). Several bias-correction methods have been developed to calibrate the raw RCM output against observations, and many studies have dealt with their comparison (Räty, O. et al. 2014; Casanueva, A. et al. 2020; Ji, X. et al. 2020; Mendez, M. et al. 2020). In addition to simpler approaches, including the delta method or linear scaling, there are also more complex methods that take into account the whole distribution of the meteorological var- iables (Themessi, M.J. et al. 2010). However, it is important to keep in mind that every method – even the best-estimated ones – has limitations since assumptions are made in all cases, such as the behaviour of the bias remains the same for the future with differ- ent climate conditions as it was in the past (Teutschbein, C. and Seibert, J. 2012; Van de Velde, J. et al. 2022). Moreover, a reliable, observation-based reference dataset of a good quality is required for a prosperous bi- as-correction (Casanueva, A. et al. 2020). The performance of the bias-correction method is sensitive to the choice of the length of the calibration period and at least a 30-year time period is recommended (Berg, P. et al. 2012; Reiter, P. et al. 2015; Ahn, K.H. et al. 2023). This study focuses on the effect of the choice of calibration period on the bias-cor- rected RCM data. Our aim was to compare bias-adjusted databases produced by using the same method with different calibration periods, as well as using another bias-correc- tion method with another calibration period and reference dataset (see Appendix, Table A1), and to investigate how the choice of differ- ent calibration periods affects the accuracy of the bias-correction. This is demonstrated by the validation of the different bias-corrected databases for the period 1993–2005. As far as we know, this latter aspect has never been analysed before with a special focus on the region of interest. Data and method Study area Hungary, the region of interest, is located in East-Central Europe, between latitudes 45.7°–48.7°N and longitudes 15.9°–22.9°E (Figure 1, A), surrounded by the Carpathi- ans to the north and east, and by the Alps to the west. The Carpathians and the territory surrounded by the mountain range together form the Carpathian Basin, one of the larg- est basins in the world, covering an area of about 500,000 km2, of which Hungary cov- ers roughly 93,000 km2. Although the Car- pathian Basin has a complex topography (the elevation varies between 75 m and 2655 m), the orography of Hungary is less complex: the highest peak of the country, called Kékes, is located in the North Hungarian Mountains with an altitude of 1014 m, and the lowest point is situated in the Great Hungarian Plain (75 m a.s.l.). It is also important to note that two-thirds of the Hungarian territory lies below 200 m a.s.l. (Figure 1, B). The climate of the country is characterised by oceanic, continental and mediterranean effects – the features of the humid oceanic climate cause slightly varying temperatures; more extreme temperatures are the result of dry, continental air masses. The precipi- tation maximum occurs in May-June, and the driest season is winter. The influence of Mediterranean air masses is mainly mani- fested in the second precipitation maximum in autumn, which is mostly observed in the south-western part of Transdanubia (Mezősi, G. 2017). Although the Carpathians are out- side of the borders of Hungary, its effect on 5Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21. the climate of the country is not negligible – an important example is the blocking of cold air masses of Siberian origin (Spinoni, J. et al. 2014). Due to the various climatic ef- fects, temperature and precipitation charac- teristics are investigated for Hungary on a yearly, seasonal and monthly scale. Reference dataset In this study, HuClim is used as a reference dataset for bias correction purposes and evaluation studies, which is produced by the HungaroMet Hungarian Meteorological Service and available on a daily basis and freely accessible via https://odp.met.hu/cli- mate/. The data is available from 1971 and it is updated to the latest year (it is 2022 in the version used in the present study) for mean air temperature, maximum air temperature, minimum air temperature and precipitation. HuClim is a measurement-based dataset, which covers Hungary on a 0.1° × 0.1° hori- zontal grid and builds upon 500 precipitation and 112 temperature stations’ data. Quality control is provided by the Multiple Analysis of Series for Homogenized Database (MASH) (Szentimrey, T. 2007) software, and the meth- od of Meteorological Interpolation based on Surface Homogenized Database (MISH) (Szentimrey, T. and Bihari, Z. 2008) is used for gridding and interpolating the meteoro- logical data. The importance of using HuClim data lies in the fact that this is the most accu- rate gridded, high-resolution, homogenized observational data currently available for the country: as it is well known that the quality of the reference database for bias adjustment is crucial (Casanueva, A. et al. 2020). Model simulations and databases Simulations of five RCMs driven by four different GCMs at a horizontal resolution of 0.11° are investigated in this study de- rived from the EURO-CORDEX framework (Jacob, D. et al. 2014). All historical simula- tions cover the period 1976–2005 and the pro- jections were accomplished under the 4.5 and 8.5 Representative Concentration Pathways scenarios (RCP4.5 and RCP8.5, respectively) (Moss, R.H. et al. 2010). Details of the selected GCM-RCM combinations are listed in Table 1. During the selection procedure we focused on those RCM-GCM combinations, which were available for both RCP scenarios as well as for raw and bias-corrected versions. In addition, model performance was taken into account based on previous studies for East-Central Europe (Mezghani, A. et al. 2017; Torma, Cs.Zs. 2019; Lazic, I. et al. 2021; Simon, Cs. et al. 2023). Fig. 1. The region of interest. A = Location of Hungary in Europe (filled with blue colour); B = The topography of Hungary on a 0.11° horizontal resolution. Source: Authors’ own editing. Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21.6 Four variables were used for this work: daily mean near-surface air temperature (tas), daily minimum near-surface air temperature (tasmin), daily maximum near-surface air tem- perature (tasmax), and daily precipitation (pr). Bias-adjusted model output from the EURO- CORDEX program was produced by using the MESAN reanalysis data (Häggmark, L. et al. 2000) for the time period 1989–2010, and a distribution scaling method (Yang, W. et al. 2010) was implemented for bias-correcting the RCM simulations. MESAN is an operational mesoscale analysis system developed by the Swedish Meteorological and Hydrological Institute (SMHI). The system is designed to provide high-resolution (about 11 km) anal- yses of meteorological variables, including precipitation and temperature. MESAN inte- grates various data sources such as weather radar observations, satellite data and ground- based measurements. Since climate model data and HuClim data are available on differ- ent horizontal resolutions, interpolation to a common 0.11° × 0.11° grid was performed us- ing the CDO (Climate Data Operators; https:// code.mpimet.mpg.de/projects/cdo/) software (Schulzweida, U. 2021) with a bilinear remap- ping method. For the purpose of creating a new bias-cor- rected RCM dataset for Hungary based on the HuClim database – the use of which is not wide- spread, only a few studies (e.g. Kern, A. et al. 2024) applied it for this territory –, we have also corrected the raw EURO-CORDEX simulations (see details in section Bias correction method). This bias-adjusted RCM data produced by the use of HuClim are publicly available in the Zenodo repository (Simon, Cs. et al. 2024). Note that the bias-correction was implemented for the RCM simulations of the historical (1976–2005) and the scenario (2006–2099) periods, but in this study only the analysis of the historical simula- tions is considered. Bias correction method In order to correct the systematic bias pre- sent in raw RCM outputs, the internation- ally accepted, non-parametric, percentile- based quantile mapping method was ap- plied, following the work of Mezghani, A. et al. (2017). This method is one of the most commonly used higher-skill bias-correction techniques in the climate research commu- nity (Teutschbein, C. and Seibert, J. 2013) which has been successfully applied in the East-Central European region (e.g. Torma, Cs.Zs. and Kis, A. 2022; Kern, A. et al. 2024). In general, the quantile mapping procedure matches the quantile-based distribution of the raw RCM simulations to that of the ob- served data. In the present study the bias- adjustment of the simulated time series was performed for each grid cell on the common 0.11° grid and the number of quantiles was set to 1000. In addition, the quantile mapping was performed for each month separately Table 1. Overview of the applied RCMs and their driving GCMs used in the present study RCM Driving GCM Modelling group CCLM4-8-17 (Rockel, B. et al. 2008) MPI-ESM-LR (Jungclaus, J.H. et al. 2010) Climate Limited-area Modelling Community, Germany HIRHAM5 (Christensen, O.B. et al. 1998) EC-EARTH (Hazeleger, W. et al. 2010) Danish Meteorological Institute, Denmark RACMO22E (Van Meijgaard, E. et al. 2012) HadGEM2-ES (Collins, W.J. et al. 2011) Royal Netherlands Meteorological Institute, The Netherlands RCA4 (Kupiainen, M. et al. 2014) CNRM-CM5 (Voldoire, A. et al. 2012) Swedish Meteorological and Hydrological Institute, Rossby Centre, Sweden REMO2009 (Jacob, D. et al. 2012) MPI-ESM-LR (Jungclaus, J.H. et al. 2010) Helmholtz-Zentrum Geesthacht, Climate Service Centre, Max Planck Institute for Meteorology, Germany 7Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21. with the aim of investigating the behaviour of the bias and the accuracy of the bias- correction on a finer timescale. The length and the quality of the reference dataset is also a key tool, because quantile mapping is considered to be sensitive to that (Fowler, H.J. and Kilsby, C.G. 2007). To perform the quantile mapping method, two different 30- year calibration periods were selected from the observation-based HuClim database: an earlier (1976–2005, BC-HUCLIM-1) and a more recent (1993–2022, BC-HUCLIM-2) 30-year long period with different climatic characteristics, thus, creating two different bias-adjusted databases. Noting that using different calibration peri- ods of the same length and the same bias ad- justment procedure can highlight the effect of the choice of the calibration period. However, the most recent period has characteristics of a warmer climate relative to the earlier pe- riod, which can lead to differences in rela- tive biases, when different datasets based on different calibration periods are investi- gated. HuClim was also used by Kern, A. et al. (2024) to construct the FORESEE-HUN v1.0 database, which contains bias-adjusted RCM projections for the period 2022–2100 for Hungary, and for which a longer calibration period (1971–2020) was chosen. Selected climate indices Beside the investigation of average tempera- ture and precipitation values, a total of eight climate indices were also chosen and analysed over the region of interest. Table A2 in Appen- dix contains the details about the set of these indices, which can be separated into two cat- egories: (1) threshold-related indices: count the number of days when a given (precipita- tion or temperature) threshold is exceeded; namely, summer days (SU), frost days (FD), tropical nights (TR) and wet days (RR1); (2) extreme-related indices: i.e. the warmest day (TXx) and the coldest night (TNn) of a period, the maximum of daily precipitation amount (RX1day), and extremely wet days (R99p). Results In this section, the performance of the differ- ent bias-adjusted databases is investigated for the evaluation period 1993–2005, which is the overlapping time interval of the three differ- ent calibration periods (1976–2005; 1989–2010; 1993–2022) used for the bias-corrections, fur- thermore, it contains only historical model simulations. Different metrics were selected for the evaluation: firstly, the mean precipitation and temperature characteristics are analysed on different timescales, and then the chosen climate indices are investigated over Hungary. Mean precipitation and temperature characteristics First of all, relative bias was calculated as the difference relative to the climatological aver- age (as defined e.g. in the work of Vogel, E. et al. 2023) of the precipitation in the reference period shown in the first column of Figure 2. Relative bias was obtained from average annual values over the evaluation period. In the case of precipitation, relative bias shows positive val- ues in most of the area for the raw simulations, especially in the North Hungarian Mountains with a positive bias of 35–55 percent, whereas in the south-western part of the country a neg- ative bias of 5–15 percent occurs. BC-MESAN shows lower relative bias in the northern area, but the negative values are more pronounced. In terms of the two HuClim-based bias-correct- ed datasets the relative bias is closer to zero in comparison to the above-mentioned cases, but for the BC-HUCLIM-1 a negative bias of 5–10 percent is dominant over the country, while BC-HUCLIM-2 shows the same amount of positive bias in most of the area. In summary, the warming of recent decades has also affect- ed annual precipitation totals. For temperature (tas, tasmin, tasmax) absolute biases are shown (columns 2–4 of Figure 2), which were calculat- ed as the difference between the simulated and the observation-based values. Absolute biases are small (around 0.5 °C) for BC-HUCLIM-1 and BC-HUCLIM-2, but with an opposite sign, which can be related to the different climatic Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21.8 conditions of the two calibration periods, i.e. while the database calibrated on the basis of a warmer climate shows an overestimation, the database bias corrected on the basis of a colder (earlier) period shows an underestima- tion. BC-MESAN is the most accurate for tas (±0.6 °C), but a relatively large bias appears in the case of tasmin (1–2 °C). The performance of each RCM simulation was analysed by the difference of spatially averaged seasonal precipitation sum between the simulated values and HuClim, calculat- ed and displayed for each database and ex- pressed as a percentage (Figure 3). The differ- ence between the climate models is higher in all seasons for the raw simulations, while for the bias-corrected results, these differences are reduced. Most of the raw RCM simula- tions underestimate summer precipitation by 15–30 percent, whereas in the other seasons an overestimation by 5–30 percent is found. For the two HuClim-based bias-corrected datasets, the difference between the individu- al RCMs is proved to be the smallest in spring and autumn. RACMO22E was found to be the most accurate among the RCMs and the worst performing models are HIRHAM5 and CCLM4-8-17. Based on the multi-model aver- age of the differences, the variation is negli- gible in autumn for BC-HUCLIM-2 (-0.3%), and BC-HUCLIM-1 shows the best perfor- mance (-4%) in the case of winter. However, for spring and summer the results most con- sistent with observations were found in the case of the BC-MESAN multi-model average (+4.5% and -3.9%, respectively). The performance of the individual RCM simulations was also investigated for the tem- perature-related variables. The average sea- sonal temperature characteristics were calcu- lated based on the RCM simulations and com- pared to HuClim, which served as reference Fig. 2. Biases of the raw and bias-adjusted RCM simulations based on the multi-model ensemble mean for each variable and database for the period 1993–2005. Source: Authors’ own editing. 9Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21. (results can be seen in Figure A1 in Appendix). The multi-model average and standard devi- ation of the variations have also been calcu- lated and analysed. Similar to precipitation, raw RCM outputs show the largest standard deviation (between 0.6–2 °C), except for aver- age summer tasmin (0.17 °C), which is com- parable with BC-MESAN (0.15 °C). For BC- MESAN, the standard deviation is the smallest in autumn for all variables, and the most negli- gible for tasmin (0.08 °C), however, in the case of average seasonal tas and tasmax, the highest standard deviation values occur for all seasons in comparison with the other bias-corrected databases. The standard deviation is compara- ble for BC-HUCLIM-1 and BC-HUCLIM-2 and it ranges from 0.15 °C to 0.25 °C. Based on the multi-model average of the differences of the individual RCM simulations, BC-HUCLIM-2 shows the poorest performance characterised by a general overestimation. The best perfor- mance was found for BC-HUCLIM-1 in terms of average seasonal tas and tasmin, with an av- erage difference of ±0.3 °C. For BC-MESAN a slight overestimation is more common for tas and tasmax. In the case of BC-HUCLIM-1 and BC-HUCLIM-2, CCLM4-8-17 was obtained to be the most accurate RCM simulation, and the performance of RCA4 was found to be the poorest in winter. For the other seasons, we cannot highlight any climate model as being the best one or an absolute outlier. Finally, we evaluated the raw and bias-ad- justed RCM data on a monthly basis. The annual cycle of the average monthly mean, minimum and maximum temperature and the average monthly precipitation sum over Hungary was investigated for the validation Fig. 3. The spatially averaged seasonal precipitation totals compared to HuClim for the period 1993–2005 dis- played for the individual RCM simulations (indicated by different colours) and for the databases considered in this study. The differences are expressed as a percentage. Source: Authors’ own editing. Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21.10 period according to the multi-model ensemble of the raw and the bias-cor- rected RCM simulations (Figure 4). In the one hand, for the temperature-relat- ed variables and for the raw RCM simulations, the spread of the models was found to be the larg- est (1.4–1.9 °C) in summer in the case of tasmax, but on the other hand, it was minimal (0.3–0.7 °C) for tasmin. The variance be- tween the RCM simu- lations ranges between 0.6 °C and 1.7 °C for tas, with the greatest extent in winter and spring, and the smallest in autumn based on the raw data. The uncertainty was re- duced by bias-adjustment regardless of the choice of the calibration period. The performance of BC- HUCLIM-1 was the best for temperature values in autumn and in the first part of spring (March and April), however, a gener- al underestimation (with a median of 0.1–0.6 °C) can be observed in May and in the summer months (JJA). In the case of BC- HUCLIM-2 an overes- timation by 0.4–1.7 °C is dominant except for May and for October. BC-MESAN has the best performance in autumn and the poorest from February to April. For precipitation, a substantial overestimation (10–44 mm) was shown by the raw RCM simulations, espe- Fig. 4. The annual cycle of the average monthly temperatures (tas, tasmin, tasmax) and precipitation in Hungary during the period of 1993–2005 accord- ing to the raw and the different bias-adjusted RCM simulations (marked with different colours) in comparison with the measurement-based HuClim data (black horizontal lines). Source: Authors’ own editing. 11Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21. cially in winter months, moreover, in May and October, when the uncertainty is the highest. A general underestimation of 9–35 mm was found for July, August and September based on the raw RCM simulations. After the bias-correc- tion procedure, the variance between the RCM simulations decreased, and it was found to be the smallest in January and March in the two, HuClim-based bias-adjusted databases, but in some cases (in August and December) it re- mained comparable with the uncertainties of the raw simulations. The performance of the different bias-adjusted databases varies over the months: BC-HUCLIM-1 and BC-HUCLIM-2 show similar results in January, March, June, August and December, however, the boxes rep- resent higher (lower) values for BC-HUCLIM-2 in comparison with BC-HUCLIM-1 in February, May, July, September and October (April and November). A clear overestimation (underesti- mation) appears in the case of May and October (July) regardless of the applied bias-correction and calibration periods. Climate indices This section presents the validation of the selected climate indices for Hungary. First, the spatial distribution of the annual number (amount) of threshold-based and extreme, temperature-related (precipitation-related) cli- mate indices was investigated for the different datasets. Figure 5 shows the results for summer days, tropical nights, frost days and wet days averaged over the period 1993–2005. The an- nual number of SU varies between 10–100 days over Hungary, with the minimum (10–25 days) in the mountainous areas. The highest occur- rence (4–7 days per year) of the annual number of TR was observed at higher altitudes and on the southern slopes of the mountain ranges. This result can be explained by the presence of inversion stratification and as an effect of foehn wind, which occurs on the lee side of a moun- tain range (Brinkmann, W.A.R. 1971). The an- nual number of FD and its spatial distribution is also consistent with orography: over the highest peaks it reached 140–150 days, while in the southern part of Hungary it remained below 100 days per year. The annual frequency of RR1 is found to be relatively homogeneous across the country with 80–100 days. Figure A2 in Appendix shows the spatial distribution of the bias fields with respect to the HuClim dataset. On the one hand, the ensemble mean of BC-HUCLIM-1 is in good agreement with the reference values for every threshold-based climate index apart from the underestimation of SU with 5–15 days in the Great Hungarian Plain and the slight underes- timation of TR, especially in areas with higher altitudes. On the other hand, the average an- nual number of TR is overestimated by all da- tabases except for BC-HUCLIM-1. In the case of SU, a general underestimation was found for BC-MESAN and a general overestimation appears based on BC-HUCLIM-2, especially in the south-eastern region of the target area. Raw simulations show 20–30 days overestima- tion for RR1 (mostly in the mountains), and the same extent of underestimation appears for FD compared to the reference values. These results are consistent with a warming trend in the re- gion, i.e. the database calibrated to the most recent period gives an overestimation of the relevant indices compared to the earlier period. Figure 6. compares the values of ex- treme-related climate indices and their spa- tial distribution over the period 1993–2005 based on the different databases investigated in this study. According to the reference data, the absolute minimum temperatures (around -28 °C) were detected in areas prone to frost, such as the north-eastern region and the northern valleys. Among the bias-corrected databases BC-HUCLIM-2 and BC-MESAN show relatively better agreement in terms of both spatial distribution and values. BC- HUCLIM-1 assumes lower temperatures over an extensive area. The highest temperatures (39–40 °C) occurred in the south-eastern part of the Great Hungarian Plain, while in the mountains TXx values of 30–32 °C were found. This index is best represented by BC- HUCLIM-1, however, BC-MESAN, as well as the raw simulations, overestimates TXx by 1–2 °C, mainly in the Great Hungarian Plain. Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21.12 Turning our attention to the extreme, precipitation-related climate indices, the highest daily precipitation sum (110–130 mm) was clearly related to Mátra mountain range, where Kékes is located. However, for R99p – which varies between 18–30 mm over Hungary –, the higher values were more prevalent in the south-western Transdanubian region and in the west- ern border. These spatial patterns are well represented by BC-HUCLIM-1 and BC- HUCLIM-2, but according to BC-MESAN, a much more homogeneous spatial distribu- tion appears for RX1day with a strong un- derestimation, especially in the mountains, where the values for this index are almost half as much as the reference. The spatial distribution of the bias fields with respect to the HuClim dataset is also shown in Figure A3 in Appendix. Normalized Taylor diagrams (Taylor, K.E. 2001) were also created in order to determine the degree of statistical similarity between the HuClim reference dataset and the various cli- mate model simulations for each climate in- dex. The closer a symbol is to this reference point (indicated by a black square), the better the performance of the related RCM simula- tion ensemble. Figure 7 presents these statisti- cal metrics for the average annual number of threshold-based climate indices for the target domain for the period 1993–2005. It can be Fig. 5. Threshold-based climate indices (SU, TR, FD, RR1) over Hungary based on the multi-model averages of the different bias-corrected simulations (rows 2–4) and raw outputs (last row) in comparison with the HuClim reference data (first row) for the validation period of 1993–2005. Source: Authors’ own editing. 13Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21. seen that bias-correction based on HuClim data (regardless the calibration period) has obviously a positive effect, in addition, the two HuClim-based datasets provide similar statis- tical metrics, except for TR, where standard deviation values were found to be different – which means that BC-HUCLIM-2 exhibits larger spatial variability for tropical nights than BC-HUCLIM-1. These databases show the highest degree of similarity for SU and FD compared to the HuClim reference, for which the correlation coefficients are found to be above 0.99 and the RMSE values are minimal (< 1.2). It is interesting to see that the symbols of the multi-model ensemble of BC-MESAN and raw simulations are located on similar lines of correlation for each climate index. Taylor diagrams for extreme-related climate indices for the period 1993–2005 can be seen in Figure A4 in Appendix. In this case the effect of bias-adjustment using HuClim was also found to be favourable but less successful than for threshold-related indices. Similar statistical metrics were obtained for the HuClim-based databases in terms of extreme, precipitation- related climate indices, but more pronounced differences appeared for TNn and TXx. The degree of similarity regarding the spatial dis- tribution of the lowest temperature was higher for BC-MESAN compared to BC-HUCLIM-1, however, BC-HUCLIM-1 showed the best performance in the case of TXx, for which the correlation coefficient is around 0.99 and the RMSE was found to be the smallest (0.15). Fig. 6. The same as in Figure 5, but for extreme-related climate indices (TNn, TXx, RX1day, and R99p). Source: Authors’ own editing. Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21.14 Summary and final conclusions The effect of the choice of the calibration pe- riod on the accuracy of the bias-correction was analysed in this study for Hungary, through the validation of three different bias- adjusted databases. Five RCMs were investi- gated from the framework of EURO-CORD- EX at a horizontal resolution of 0.11° for the historical time period of 1976–2005 for four variables: daily mean temperature, minimum and maximum temperature, and precipita- tion. The percentile-based quantile mapping method was applied for the bias-correction, and was performed on a monthly scale. The observation-based HuClim dataset was used as a reference for the bias correction and the validation. Two, 30-year long time periods were selected from the HuClim database: 1976–2005 and 1993–2022, and as a result of the bias-correction, two different bias-adjust- ed databases were created based on these calibration periods. A third bias-adjusted database produced by the EURO-CORDEX community was also examined in this study. Two groups of climate indices were also as- sessed: (1) threshold-related climate indices: SU, TR, FD and RR1; (2) extreme-related cli- mate indices: TXx, TNn, RX1day and R99p. The period 1993–2005 was selected as the Fig. 7. Statistical characteristics summarized by Taylor diagrams for raw and bias-corrected multi-model data (coloured symbols) with respect to HuClim (black square) for the period 1993–2005. The four panels refer to the threshold based climate indices (SU, TR, FD, and RR1). Source: Authors’ own editing. 15Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21. validation period, since it is the overlapping time interval of the three calibration periods and contains only historical simulations. In the validation period, the relative bias of the mean annual precipitation was the closest to zero (5–10%) in the cases of the two, HuClim- based bias-corrected databases, but with the opposite sign. This sign of absolute bias also appears for mean annual temperatures, since BC-HUCLIM-1 (BC-HUCLIM-2) has a bias of around -0.5 °C (+0.5 °C) over Hungary. The av- erage seasonal temperature characteristics are similarly well approximated by BC-HUCLIM-1 and BC-MESAN, but a general overestimation appears for BC-HUCLIM-2. For the annual cycle of the average monthly mean, minimum and maximum temperature, BC-HUCLIM-1 is the most accurate bias-adjusted database, especially in autumn and in the first part of spring (March and April), however, a slight underestimation (with a median of 0.1–0.6 °C) appears during the summer months (JJA). For precipitation, the performance of each data- base shows a large variability between seasons and months. Note that the variation between the individual RCM simulations is reduced for each bias-corrected database in comparison with the raw model simulations. The annual number of threshold-based climate indices was in good agreement with the reference values in the case of BC-HUCLIM-1. The spatial dis- tribution of the precipitation-related climate indices (RR1, RX1day, R99p) are well repre- sented by the HuClim-based bias-corrected datasets, however, an excessively homogene- ous spatial distribution appears for RX1day with a strong underestimation according to BC-MESAN. In general, the choice of calibra- tion period is clearly influenced by the ongoing climate change. That is, the database corrected for the warmer period overestimates the av- erage temperature and precipitation patterns compared to an earlier (and cooler) period, while the thresholds for the cold period are underestimated. As a final conclusion, it can be said that the performance of the bias-corrected RCM simulations clearly depends on the analysed variable and chosen calibration period, as the results of the validation reflect the different climatic conditions of the calibration peri- ods. (For example, the overestimation of the temperature-related variables or the tropical nights when using a more recent time period with more extreme events for bias-correcting the raw RCM data.) On the other hand, the re- sults for precipitation are less affected by the choice of the calibration period, but they are more sensitive to the reference database. This can be explained by the fact that precipitation is one of the most variable meteorological el- ements not only in time but also in space. It means that using a database produced by a higher number of stations’ measurement data provides more accurate results for precipita- tion. Overall, using the earlier calibration pe- riod (1976–2005) from the HuClim database proved to be the most accurate in the most cases during the validation. The next step in our research is to analyse the different bias- adjusted RCM simulations for the future. Acknowledgements: The research has been supported by the Hungarian Scientific Research Fund (OTKA FK-142349). All data from EURO-CORDEX modelling group used in this study © European Commission– JRC 2013, along with GTOPO30 data provided by the U.S. Geological Survey are acknowledged. The raw and bias-corrected RCM data provided by EURO- CORDEX community and used in this work was downloaded from the following web site: https:// esgf-data.dkrz.de/. The HuClim database is freely available at: https://odp.met.hu © HungaroMet REFERENCES Ahn, K.H., de Padua, V.M.N., Kim, J. and Yi, J. 2023. Impact of diverse configuration in multivariate bias correction methods on large-scale hydro- logical modelling under climate change. Journal of Hydrology 627. Part A, 130406. https://doi. org/10.1016/j.jhydrol.2023.130406 Beniston, M., Stephenson, D.B., Christensen, O.B., Ferro, C.A.T., Frei, C., Goyette, S., Halsnaes, K., Holt, T., Jylhä, K., Koffi, B., Palutikof, J., Schöll, R., Semmler, T. and Woth, K. 2007. Future extreme events in European climate: An exploration of re- gional climate model projections. Climatic Change 81. 71–95 https://doi.org/10.1007/s10584-006-9226-z Berg, P., Feldmann, H. and Panitz, H.J. 2012. Bias correction of high resolution regional climate model https://doi.org/10.1016/j.jhydrol.2023.130406 https://doi.org/10.1016/j.jhydrol.2023.130406 https://doi.org/10.1007/s10584-006-9226-z Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21.16 data. Journal of Hydrology 448–449. 80–92. https:// doi.org/10.1016/j.jhydrol.2012.04.026 Brinkmann, W.A.R. 1971. What is a foehn? We a t h e r 2 6 . ( 6 ) : 2 3 0 – 2 4 0 . h t t p s : / / d o i . org/10.1002/j.1477-8696.1971.tb04200.x Casanueva, A., Herrera, S., Iturbide, M., Lange, S., Jury, M., Dosio, A., Maraun, D. and Gutiérrez, J.M. 2020. Testing bias adjustment methods for regional climate change applications under ob- servational uncertainty and resolution mismatch. Atmospheric Science Letters 21. e978. https://doi. org/10.1002/asl.978 Christensen, O.B., Christensen, J.H., Machenhauer, B. and Botzet, M. 1998. Very high resolution region- al climate simulations over Scandinavia – Present climate. Journal of Climate 11. 3204–3229. https:// doi.org/10.1175/1520-0442(1998)011<3204:VHRR CS>2.0.CO;2 Ciarlo, J.M., Coppola, E., Fantini, A., Giorgi, F., Gao, X.J., Tong, Y., Glazer, R.H., Torres Alvarez, J.A., Sines, T., Pichelli, E., Raffaele, F., Das, S., Bukovsky, M., Ashfaq, M., Im, E.S., Nguyen-Xuan, T., Teichmann, C., Remedio, A., Remke, T., Bülow, K., Weber, T., Buntemeyer, L., Sieck, K., Rechid, D. and Jacob, D. 2021. A new spatially distributed added value index for regional climate models: The EURO-CORDEX and the CORDEX-CORE highest resolution ensembles. Climate Dynamics 57. 1403– 1424. https://doi.org/10.1007/s00382-020-05400-5 Collins, W.J., Bellouin, N., Doutriaux-Boucher, M., Gedney, N., Halloran, P., Hinton, T., Hughes, J., Jones, C.D., Joshi, M., Liddicoat, S., Martin, G., O’Connor, F., Rae, J., Senior, C., Sitch, S., Totterdell, I., Wiltshire, A. and Woodward, S. 2011. Development and evaluation of an Earth System model, HADGEM2. Geoscientific Model Development 4. (4): 1051‒1075. https://doi. org/10.5194/gmd-4-1051-2011 Costoya, X., Rocha, A. and Carvalho, D. 2020. Using bias-correction to improve future projections of offshore wind energy resource: A case study on the Iberian Peninsula. Applied Energy 262. 114562. https://doi.org/10.1016/j.apenergy.2020.114562 Déqué, M., Rowell, D.P., Lüthi, D., Giorgi, F., Christensen, J.H., Rockel, B., Jacob, D., Kjellström, E., Castro, M. and van den Hurk, B. 2007. an in- tercomparison of regional climate simulations for Europe: Assessing uncertainties in model projections. Climatic Change 81. 53–70. https://doi. org/10.1007/s10584-006-9228-x Di Luca, A., Argüeso, D., Evans, J.P., de Elía, R. and Laprise, R. 2016. Quantifying the overall added val- ue of dynamical downscaling and the contribution from different spatial scales. Journal of Geophysical Research: Atmospheres 121. (4): 1575–1590. https:// doi.org/10.1002/2015JD024009 Faghih, M., Brissette, F. and Sabeti, P. 2022. Impact of correcting sub-daily climate model biases for hydrological studies. Hydrology and Earth System Sciences 26. (6): 1545–1563. https://doi.org/10.5194/ hess-26-1545-2022 Fantini, A., Raffaele, F., Torma, C.Z., Bacer, S., Coppola, E., Giorgi, F., Ahrens, B., Dubois, C., Sanchez, E. and Verdecchia, M. 2018. Assessment of multiple daily precipitation statistics in ERA- Interim driven Med-CORDEX and EURO-CORDEX experiments against high resolution observa- tions. Climate Dynamics 51. 877–900. https://doi. org/10.1007/s00382-016-3453-4 Fowler, H.J. and Kilsby, C.G. 2007. Using regional climate model data to simulate historical and future river flows in northwest England. Climatic Change 80. 337–367. https://doi.org/10.1007/s10584- 006-9117-3 Giorgi, F. 2005. Climate change prediction. Climatic Change 73. 239–265. https://doi.org/10.1007/s10584- 005-6857-4 Häggmark, L., Ivarsson, K.I., Gollvik, S. and Olofsson, P.O. 2000. Mesan, an operational mesoscale analysis system. Tellus A: Dynamic Meteorology and Oceanography 52. (1): 2–20. https:// doi.org/10.3402/tellusa.v52i1.12250 Hazeleger, W., Severijns, C., Semmler, T., Stefanescu, S., Yang, S., Wang, X., Wyser, K., Dutra, E., Baldasano, J.M., Bintanja, R., Bougeault, P., Caballero, R., Ekman, A.M.L., Christensen, J.H., van den Hurk, B., Jimenez, P., Jones, C., Kallberg, P., Koenigk, T., McGrath, R., Miranda, P., van Noije, T., Palmer, T., Parodi, J.A., Schmith, T., Selten, F., Storelvmo, T., Sterl, A., Tapamo, H., Vancoppenolle, M., Viterbo, P. and Willén, U. 2010. EC-Earth: A seamless earth-system predic- tion approach in action. Bulletin of the. American Meteorological Society 91. (10): 1357–1364. https:// doi.org/10.1175/2010BAMS2877.1 IPCC 2013. Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Eds.: Stocker, T.F., Qin, D., Plattner, G.K., Tignor, M., Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex, V and Midgley, P.M. Cambridge, United Kingdom and New York, NY, USA, Cambridge University Press. https://doi. org/10.1017/CBO9781107415324 Jacob, D., Elizalde, A., Haensler, A., Hagemann, S., Kumar, P., Podzun, R., Rechid, D., Remedio, A.R., Saeed, F., Sieck, K., Teichmann, C. and Wilhelm, C. 2012. Assessing the transferability of the regional climate model REMO to different coordinated re- gional climate downscaling experiment (CORDEX) regions. Atmosphere 3. (1): 181–199. https://doi. org/10.3390/atmos3010181 Jacob, D., Petersen, J., Eggert, B., Alias, A., Christensen, O.B., Bouwer, L.M., Braun, A., Colette, A., Déqué, M., Georgievski, G., Georgopoulou, E., Gobiet, A., Menut, L., Nikulin, https://doi.org/10.1016/j.jhydrol.2012.04.026 https://doi.org/10.1016/j.jhydrol.2012.04.026 https://doi.org/10.1002/j.1477-8696.1971.tb04200.x https://doi.org/10.1002/j.1477-8696.1971.tb04200.x https://doi.org/10.1002/asl.978 https://doi.org/10.1002/asl.978 https://doi.org/10.1175/1520-0442(1998)011%3C3204:VHRRCS%3E2.0.CO;2 https://doi.org/10.1175/1520-0442(1998)011%3C3204:VHRRCS%3E2.0.CO;2 https://doi.org/10.1175/1520-0442(1998)011%3C3204:VHRRCS%3E2.0.CO;2 https://doi.org/10.1007/s00382-020-05400-5 https://doi.org/10.5194/gmd-4-1051-2011 https://doi.org/10.5194/gmd-4-1051-2011 https://doi.org/10.1016/j.apenergy.2020.114562 https://doi.org/10.1007/s10584-006-9228-x https://doi.org/10.1007/s10584-006-9228-x https://doi.org/10.1002/2015JD024009 https://doi.org/10.1002/2015JD024009 https://doi.org/10.5194/hess-26-1545-2022 https://doi.org/10.5194/hess-26-1545-2022 https://doi.org/10.1007/s00382-016-3453-4 https://doi.org/10.1007/s00382-016-3453-4 https://doi.org/10.1007/s10584-006-9117-3 https://doi.org/10.1007/s10584-006-9117-3 https://doi.org/10.1007/s10584-005-6857-4 https://doi.org/10.1007/s10584-005-6857-4 https://doi.org/10.3402/tellusa.v52i1.12250 https://doi.org/10.3402/tellusa.v52i1.12250 https://doi.org/10.1175/2010BAMS2877.1 https://doi.org/10.1175/2010BAMS2877.1 https://doi.org/10.1017/CBO9781107415324 https://doi.org/10.1017/CBO9781107415324 https://doi.org/10.3390/atmos3010181 https://doi.org/10.3390/atmos3010181 17Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21. G., Haensler, A., Hempelmann, N., Jones, C., Keuler, K., Kovats, S., Kröner, N., Kotlarski, S., Kriegsmann, A., Martin, E., van Meijgaard, E., Moseley, C., Pfeifer, S., Preuschmann, S., Radermacher, C., Radtke, K., Rechid, D., Rounsevel, M., Samuelsson, P., Somot, S., Soussana, J.F., Teichmann, C., Valentini, R., Vautard, R., Weber, B. and Yiou, P. 2014. EURO- CORDEX New high resolution climate change projections for European impact research. Regional Environmental Change 14. 563–578. https://doi. org/10.1007/s10113-013-0499-2 Jaiswal, R., Mall, R.K., Singh, N., Kumar, T.V.L. and Niyogi, D. 2022. Evaluation of bias correction methods for regional climate models: Downscaled rainfall analysis over diverse agroclimatic zones of India. Earth and Space Science 9. e2021EA001981. https://doi.org/10.1029/2021EA001981 Ji, X., Li, Y., Luo, X., He, D., Gou, R., Wang, J., Bai, Y., Yue, C. and Liu, C. 2020. Evaluation of bias correction methods for APHRODITE data to im- prove hydrologic simulation in a large Himalayan basin. Atmospheric Research 242. 104964. https://doi. org/10.1016/j.atmosres.2020.104964 Jungclaus, J.H., Lorenz, S.J., Timmreck, C., Reick, C.H., Brovkin, V., Six, K., Segschneider, J., Giorgetta, M.A., Crowley, T.J., Pongratz, J., Krivova, N.A., Vieira, L.E., Solanki, S.K., Klocke, D., Botzet, M., Esch, M., Gayler, V., Haak, H., Raddatz, T.J., Roeckner, E., Schnur, R., Widmann, H., Claussen, M., Stevens, B. and Marotzke, J. 2010. Climate and carbon cycle variability over the last millennium. Climate of the Past 6. 723‒737. https://doi.org/10.5194/cp-6-723-2010 Kern, A., Dobor, L., Hollós, R., Marjanovic, H., Torma, Cs.Zs., Kis, A., Fodor, N. and Barcza, Z. 2024. Seamlessly combined historical and projected daily meteorological datasets for impact studies in Central Europe: The RORESEE v4.0 and the FORESEE-HUN v1.0. Climate Services 33. 100443. https://doi.org/10.1016/j.cliser.2023.100443 Kupiainen, M., Jansson, C., Samuelsson, P., Jones, C., Willén, U., Hansson, U., Ullerstig, A., Wang, S. and Döscher, R. 2014. Rossby Centre regional atmospheric model, RCA4. Rossby Centre Newsletter, June 2014. Norrköpping, Sweden, Rossby Centre. Lazic, I., Tošic, M. and Djurdjevic, V. 2021. Verification of the EURO-CORDEX RCM his- torical run results over the Pannonian Basin for the summer season. Atmosphere 12. 714. https://doi. org/10.3390/atmos12060714 Mendez, M., Maathuis, B., Hein-Griggs, D. and Alvarado-Gamboa, L.F. 2020. Performance evalu- ation of bias correction methods for climate change monthly precipitation projections over Costa Rica. Water 12. (2): 482. https://doi.org/10.3390/w12020482 Mezghani, A., Dobler, A., Haugen, J.E., Benestad, R.E., Parding, K.M., Piniewski, M., Kardel, I. and Kundzewicz, Z.W. 2017. CHASE-PL Climate Projection dataset over Poland – Bias adjustment of EURO-CORDEX simulations. Earth System Science Data 9. (2): 905‒925. https://doi.org/10.5194/ essd-9-905-2017 Mezősi, G. 2017. Climate of Hungary. In The Physical Geography of Hungary. Geography of the Physical Environment. Ed.: Bleier, D., Cham, Springer, 101– 119. https://doi.org/10.1007/978-3-319-45183-1_2 Moss, R.H., Edmonds, J.A., Hibbard, K.A., Manning, M.R., Rose, S.K., van Vuuren, D.P., Carter, T.E., Emori, S., Kainuma, M., Kram, T., Meehl, G.A., Mitchell, J.F.B., Nakicenovic, N., Riahi, K., Smith, S.J., Stouffer, R.J., Thomson, A.M., Weyant, J.P. and Wilbanks, T.J. 2010. The next generation of sce- narios for climate change research and assessment. Nature 463. (7282): 747–756. https://doi.org/10.1038/ nature08823 Ngai, S.T., Tangang, F. and Juneng, L. 2016. Bias correction of global and regional simulated daily precipitation and surface mean temperature over Southeast Asia using quantile mapping method. Global and Planetary Change 149. 79–90. https://doi. org/10.1016/j.gloplacha.2016.12.009 Räty, O., Räisänen, J. and Ylhäisi, J.S. 2014. Evaluation of delta change and bias correction methods for fu- ture daily precipitation: intermodel cross-validation using ENSEMBLES simulations. Climate Dynamics 42. 2287–2303. https://doi.org/10.1007/s00382-014- 2130-8 Reiter, P., Gutjahr, O., Schefczyk, L., Heinemann, G. and Casper, M. 2015. Bias correction of ENSEMBLES precipitation data with focus of the effect of the length of the calibration period. Meteorologische Zeitschrift 25. (1): 85–96. https://doi.org/10.1127/ metz/2015/0714 Rockel, B., Will, A. and Hense, A. 2008. Special issue: Regional climate modeling with COSMO-CLM (CCLM). Meteorologische Zeitschrift 17. (4): 347‒348. https://doi.org/10.1127/0941-2948/2008/0309 Rummukainen, M. 2016. Added value in regional climate modeling. WIREs Climate Change 7. (1): 145–159. https://doi.org/10.1002/wcc.378 Schulzweida, U. 2021. CDO user guide. CDO Version 2.0.0. Geneva, Zenodo, CERN. Available at https:// code.mpimet.mpg.de/projects/cdo/embedded/ cdo.pdf Simon, Cs., Kis, A. and Torma, Cs.Zs. 2023. Temperature characteristics over the Carpathian Basin – Projected changes of climate indices at regional and local scale based on bias-adjusted CORDEX simulations. International Journal of Climatology 43. (8): 3552–3569. https://doi.org/10.1002/joc.8045 Simon, Cs., Torma, Cs.Zs. and Kis, A. 2024. Bias- corrected EURO-CORDEX daily temperature and precipitation dataset for Hungary (Data set). Zenodo, Geneva, CERN. https://doi.org/10.5281/zeno- do.10925529 https://doi.org/10.1007/s10113-013-0499-2 https://doi.org/10.1007/s10113-013-0499-2 https://doi.org/10.1029/2021EA001981 https://doi.org/10.1016/j.atmosres.2020.104964 https://doi.org/10.1016/j.atmosres.2020.104964 https://doi.org/10.5194/cp-6-723-2010 https://doi.org/10.1016/j.cliser.2023.100443 https://doi.org/10.3390/atmos12060714 https://doi.org/10.3390/atmos12060714 https://doi.org/10.3390/w12020482 https://doi.org/10.5194/essd-9-905-2017 https://doi.org/10.5194/essd-9-905-2017 https://doi.org/10.1007/978-3-319-45183-1_2 https://doi.org/10.1038/nature08823 https://doi.org/10.1038/nature08823 https://doi.org/10.1016/j.gloplacha.2016.12.009 https://doi.org/10.1016/j.gloplacha.2016.12.009 https://doi.org/10.1007/s00382-014-2130-8 https://doi.org/10.1007/s00382-014-2130-8 https://doi.org/10.1127/metz/2015/0714 https://doi.org/10.1127/metz/2015/0714 https://doi.org/10.1127/0941-2948/2008/0309 https://doi.org/10.1002/wcc.378 https://code.mpimet.mpg.de/projects/cdo/embedded/cdo.pdf https://code.mpimet.mpg.de/projects/cdo/embedded/cdo.pdf https://code.mpimet.mpg.de/projects/cdo/embedded/cdo.pdf https://doi.org/10.1002/joc.8045 https://doi.org/10.5281/zenodo.10925529 https://doi.org/10.5281/zenodo.10925529 Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21.18 Spinoni, J., Szalai, S., Szentimrey, T., Lakatos, M., Bihari, Z., Nagy, A., Németh, Á., Kovács, T., Mihic, D., Dacic, M., Petrovic, P., Krzic, A., Hiebl, J., Auer, I., Milkovic, J., Stepánek, P., Zahradnicek, P., Kilar, P., Limanowka, D., Pyrc, R., Cheval, S., Birsan, M.V., Dumitrescu, A., Deak, G., Matei, M., Antolovic, I., Nejedlík, P., Stastny, P., Kajaba, P., Bochnícek, O., Galo, D., Mikulová, K., Nabyvanets, Y., Skrynyk, O., Krakovska, S., Gnatiuk, N., Tolasz, R.,Anto1fie, T. and Vogt, J. 2014. Climate of the Carpathian Region in the period 1961–2010: climatologies and trends of 10 variables. International Journal of Climatology 35. (7): 1322–1341. https://doi. org/10.1002/joc.4059 Szentimrey, T. 2007. Manual of Homogenization Software MASHv3.02. Budapest, Hungarian Meteorological Service. Szentimrey, T. and Bihari, Z. 2008. MISH (Meteorological Interpolation based on Surface Homogenized Data Basis). In The Use of Geographical Information Systems in Climatology and Meteorology. COST Action 719, Final Report. Eds.: Tveito, O.E., Wegehenkel, M., Van der Wel, F. and Dobesch, H., Brussels, Office for Official Publication of the European Communities, 54–56. Taylor, K.E. 2001. Summarizing multiple aspects of model performance in a single diagram. Journal of Geophysical Research 106. (D7): 7183–7192. https:// doi.org/10.1029/2000JD900719 Teutschbein, C. and Seibert, J. 2012. Bias correction of regional climate model simulations for hydrological climate-change impact studies: Review and evaluation of different methods. Journal of Hydrology 456–457. 12–29. https://doi. org/10.1016/j.jhydrol.2012.05.052 Teutschbein, C. and Seibert, J. 2013. Is bias correction of regional climate model (RCM) simulations possible for non-stationary conditions? Hydrology and Earth System Sciences 17. (12): 5061–5077. https:// doi.org/10.5194/hess-17-5061-2013 Themessl, M.J., Gobiet, A. and Leuprecht, A. 2010. Empirical-statistical downscaling and error correction of daily precipitation from regional climate models. International Journal of Climatology 31. (10): 1530–1544. https://doi.org/10.1002/joc.2168 Torma, Cs.Zs., Giorgi, F. and Coppola, E. 2015. Added value of regional climate modeling over areas characterized by complex terrain – Precipitation over the Alps. Journal of Geophysical Research: Atmospheres 120. (9): 3957–3972. https:// doi.org/10.1002/2014JD022781 Torma, Cs.Zs. 2019. Detailed validation of EURO- CORDEX and Med-CORDEX regional climate model ensembles over the Carpathian Region. Időjárás 123. (2): 217–240. https://doi.org/10.28974/ idojaras.2019.2.6 Torma, Cs.Zs., Kis, A. and Pongrácz, R. 2020. Evaluation of EURO-CORDEX and Med CORDEX precipitation simulations for the Carpathian Region: Bias corrected data and projected changes. Időjárás 124. (1): 25–46. https://doi.org/10.28974/ idojaras.2020.1.2 Torma, Cs.Zs. and Kis, A. 2022. Bias-adjustment of high-resolution temperature CORDEX data over the Carpathian region: Expected changes including the number of summer and frost days. International Journal of Climatology 42. (12): 6631–6646. https://doi. org/10.1002/joc.7654 Van de Velde, J., Demuzere, M., De Baets, B. and Verhoest, N.E.C. 2022. Impact of bias nonstationarity on the performance of uni- and multivariate bias-adjusting methods: a case study on data from Uccle, Belgium. Hydrology and Earth System Sciences 26. (9): 2319–2344. https://doi. org/10.5194/hess-26-2319-2022 Van Meijgaard, E., van Ulft, L.H., Lenderink, G., de Roode, S.R., Wipfler, E.L., Boers, R. and van Timmermans, R.M.A. 2012. Refinement and Application of a Regional Atmospheric Model for Climate Scenario Calculations of Western Europe. Research report KVR 054/12. WIMEK, Wageningen Environmental Research. Wageningen, KVR. Vogel, E., Johnson, F., Marshall, L., Bende-Milch, U., Wilson, L., Peter, J.R., Wasko, C., Srikanthan, S., Sharples, W., Dowdy, A., Hope, P., Khan, Z., Mehrotra, R., Sharma, A., Matic, V., Oke, A., Turner, M., Thomas, S., Donnelly, C. and Duong, V.C. 2023. An evaluation framework for downscaling and bias correction in climate change impact studies. Journal of Hydrology 622. Part A, 129693. https://doi.org/10.1016/j. jhydrol.2023.129693 Voldoire, A., Sanchez-Gomez, E., Salas y Mélia, D., Decharme, B., Cassou, C., Sénési, S., Valcke, S., Beau, I., Alias, A., Chevallier, M., Déqué, M., Deshayes, J., Douville, H., Fernandez, E., Madec, G., Maisonnave, E., Moine, M.P., Planton, S., Saint-Martin, D., Szopa, S., Tyteca, S., Alkama, R., Belamari, S., Braun, A., Coquart, L. and Chauvin, F. 2012. The CNRM-CM5.1 global climate model: Description and basic evaluation. Climate Dynamics 40. 2091‒2121. https://doi.org/10.1007/ s00382-011-1259-y Yang, W., Andréasson, J., Graham, L.P., Olsson, J., Rosberg, J. and Wetterhall, F. 2010. Distribution based scaling to improve usability of RCM regional climate model projections for hydrological climate change impacts studies. Hydrology Research 41. (3–4): 211–229. https://doi.org/10.2166/nh.2010.004 https://doi.org/10.1002/joc.4059 https://doi.org/10.1002/joc.4059 https://doi.org/10.1029/2000JD900719 https://doi.org/10.1029/2000JD900719 https://doi.org/10.1016/j.jhydrol.2012.05.052 https://doi.org/10.1016/j.jhydrol.2012.05.052 https://doi.org/10.5194/hess-17-5061-2013 https://doi.org/10.5194/hess-17-5061-2013 https://doi.org/10.1002/joc.2168 https://doi.org/10.1002/2014JD022781 https://doi.org/10.1002/2014JD022781 https://doi.org/10.28974/idojaras.2019.2.6 https://doi.org/10.28974/idojaras.2019.2.6 https://doi.org/10.28974/idojaras.2020.1.2 https://doi.org/10.28974/idojaras.2020.1.2 https://doi.org/10.1002/joc.7654 https://doi.org/10.1002/joc.7654 https://doi.org/10.5194/hess-26-2319-2022 https://doi.org/10.5194/hess-26-2319-2022 https://doi.org/10.1016/j.jhydrol.2023.129693 https://doi.org/10.1016/j.jhydrol.2023.129693 https://doi.org/10.1007/s00382-011-1259-y https://doi.org/10.1007/s00382-011-1259-y https://doi.org/10.2166/nh.2010.004 19Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21. Appendix Table A1. Overview of the databases analysed in our work Name Type Bias-correction method Reference dataset Calibration period RAW raw – – – BC-MESAN BC-HUCLIM-1 BC-HUCLIM-2 bias-adjusted distribution scaling quantile mapping quantile mapping MESAN HuClim HuClim 1989–2010 1976–2005 1993–2022 Table A2. Description of the temperature and precipitation based climate indices used in this study Label Name Category Description Unit SU Summer days Threshold Let TX be the daily maximum temperature on day i in period j. Count the number of days when TXij > 25 °C. Days FD Frost days Let TN be the daily minimum temperature on day i in period j. Count the number of days when TNij < 0 °C. TR Tropical nights Let TN be the daily minimum temperature on day i in period j. Count the number of days when TNij > 20 °C. RR1 Wet days Let R be the daily precipitation amount on day i in period j. Count the number of days when Rij ≥ 1 mm. TXx The warmest day Extreme Let TXx be the daily maximum temperature in month k, period j. The maximum daily maximum temperature each month is then: TXxkj = max(TXxkj). °C TNn The coldest night Let TNn be the daily minimum temperature in month k, period j. The minimum daily minimum temperature each month is then: TNnkj = min(TNnkj). °C RX1day The highest daily precipitation sum Let R be the daily precipitation amount on day i in period j. The highest daily precipitation sum over a time series is then: RX1day = max(Rij). mm R99p Extremely wet days Let R be a time series of the daily precipitation amount. Then R99p is the 99th percentile of the daily precipitation amount on wet days for a refer- ence period. mm Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21.20 Fig. A1. The spatially averaged seasonal temperature characteristics compared to HuClim for the period 1993–2005 displayed for the individual RCM simulations indicated by different colours and for the four databases considered in this study. The differences are expressed in °C. Source: Authors’ own editing. Fig. A2. Differences of threshold-based climate indices (SU, TR, FD, RR1) over Hungary based on the multi- model averages of the different bias-corrected simulations (rows 1–3) and raw outputs (last row) with respect to the HuClim reference data for the validation period of 1993–2005. Source: Authors’ own editing. 21Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21. Fig. A3. The same as in Figure A2, but for extreme-related climate indices (TNn, TXx, RX1day, and R99p). Source: Authors’ own editing. Fig. A4. Statistical characteristics summarized by Taylor diagrams for raw and bias-corrected multi-model data (coloured symbols) with respect to HuClim (black square) for the period 1993–2005. The four panels refer to the extreme-related climate indices (TNn, TXx, RX1, and R99p). Source: Authors’ own editing. Simon, Cs. et al. Hungarian Geographical Bulletin 74 (2025) (1) 3–21.22