19Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33.DOI: 10.15201/hungeobull.70.1.2 Hungarian Geographical Bulletin 70 2021 (1) 19–33. Introduction In our days, the most important environ- mental phenomenon is climate change. At a global scale, the temperature change is al- ready observable, and by the end of the 21st century it is projected to likely exceed 1.5 °C (Stocker, T.F. et al. 2013). The temperature increase has complex environmental effects in global, regional and local (urban) scales, too. The local consequences of these are at least as fundamental as those of a global scale, as the majority of the population is al- ready concentrated in cities. The heat load in cities is supposed to get intensified as global temperature increase will be superimposed on urban temperature modification. Namely, owing to urban heat island (UHI) develop- ment the urban nocturnal temperature is usually higher than the rural one (Oke, T.R. et al. 2017). Overall, this can have far-reach- ing health effects (Baccini, M. et al. 2008; Bartholy, J. and Pongrácz, R. 2018). There- fore, the studies concerning the impact of global changes on local climate of cities are of a high significance for the urban inhab- itants’ health and well-being. Therefore, in order to plan and undertake the mitigation actions in particular cities, it is necessary to recognize the possible range of heat load in- crease there, in terms of both its magnitude and spatial extent. 1 Department of Climatology and Landscape Ecology, University of Szeged. H-6722 Szeged, Egyetem u. 2. E-mails: tgal@geo.u-szeged.hu, skarbitn@geo.u-szeged.hu, molnarge@geo.u-szeged, unger@geo.u-szeged.hu Projections of the urban and intra-urban scale thermal effects of climate change in the 21st century for cities in the Carpathian Basin Tamás GÁL , Nóra SKARBIT, Gergely MOLNÁR and János UNGER 1 Abstract This study evaluates the pattern of a night-time climate index, namely the tropical nights (Tmin ≥ 20 °C) during the 21st century in several different sized cities in the Carpathian Basin. For the modelling, MUKLIMO_3 mi- croclimatic model and the cuboid statistical method were applied. In order to ensure the proper representation of the thermal characteristics of an urban landscape, the Local Climate Zone (LCZ) system was used as land- use information. For this work, LCZ maps were produced using WUDAPT methodology. The climatic input of the model was the Carpatclim dataset for the reference period (1981–2010) and EURO-CORDEX regional model outputs for the future time periods (2021–2050, 2071–2100) and emission scenarios (RCP4.5, RCP8.5). As results show, there would be a remarkable increase in the number of tropical nights along the century, and there is a clearly recognizable increase owing to urban landform. In the near past, the number of the index was 6–10 nights higher in the city core than the rural area where the number of this index was negligible. In the near future this urban-rural trend is the same, however, there is a slight increase (2–5 nights) in the index in city cores. At the end of the century, the results of the two emission scenarios become distinct. In the case of RCP4.5 the urban values are about 15–25 nights, what is less stressful compared to the 30–50 nights accord- ing to RCP8.5. The results clearly highlight that the effect of urban climate and climate change would cause serious risk for urban dwellers, therefore it is crucial to perform climate mitigation and adaptation actions on both global and urban scales. Keywords: climate change, urban climate, Local Climate Zones, urban climate modelling Received November 2020; Accepted January 2021. Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33.20 In the last few years, local authorities and urban planners in Hungary have begun to pay growing attention to the latest climate change projections. Their interest, of course, focuses on urban-scale features and phenom- ena, but there are few basic research results for urban areas. Based on these trends, new basic research results are needed to produce climate projections at the local level in order to provide basic information to urban plan- ners on urban climate mitigation strategies or applied research in this field. The effect of climate change on tempera- ture is presented in IPCC reports (e.g. IPCC 2018). Recent climate model projections apply the Representative Concentration Pathways (Van Vuuren, D.P. et al. 2011), and the most commonly used scenarios are RCP4.5 and RCP8.5. These scenarios represent a global temperature increase of 2 °C and 4 °C by the end of the century, but at the regional level, the temperature change is highly variable. Considering the Carpathian Basin, it is essen- tial to evaluate climate projections for tempera- ture and temperature-related climate indices, as no further climate change mitigation and adaptation plans can be implemented with- out this information. In case of temperature change there are numerous regional model re- sults (e.g. Jacob, D. et al. 2014; Pieczka, I. et al. 2018). Based on these results, the temperature changes in this region are 1.5–2 °C (RCP4.5) and 3–4 °C (RCP8.5) by the end of this century. In order to help the evaluation of the future climate trends it is very suitable to utilize a cli- mate index projection, such as the number of tropical nights (TN, when the daily Tmin ≥ 20 °C). This particular index is a good indicator of the annual duration of adverse hot weather con- ditions, as a high minimum temperature also means a high daily temperature, taking into account the daily temperature cycle (Pieczka, I. et al. 2018). For TNs, the projected trends in the Carpathian Basin are as follows: In the period 2021–2040, their numbers are 10–15 (RCP4.5) and 10–20 (RCP8.5), and in the period 2081–2100 they are 20–30 (RCP4.5) and 40–60 (RCP8.5) (Pieczka, I. et al. 2018). It is important to highlight that these results are derived from regional climate models and the urban impact does not appear within these model outputs, so an evaluation of a detailed model experiment for urban areas in the Carpathian Basin would provide vital information for climate change related decisions. It is essential to use an urban climate model to predict the climate in urban areas. Recently, these models have evolved rapidly (Kusaka, H. et al. 2001; Martilli, A. et al. 2002; Lemonsu, A. et al. 2012; Lee, S.-H. et al. 2016; Ryu, Y.H. et al. 2016). Most of these model development initiatives are related to the Weather Research and Forecasting Model (WRF), Molnár, G. et al. (2020) briefly discusses these models. There are only a few other smaller-scale modelling options, e.g. ENVI-met (Bruse, M. and Fleer, H. 1998), Town Energy Balance (Masson, V. 2000) and MUKLIMO_3 models (Sievers, U. 1995). MUKLIMO_3 offers a great possibility for climate projection for urban ar- eas, since combined with the statistical cuboid- method it is capable for time effective simula- tion. The other advantage of this model is the representation of building arrays. In the urban parametrizations related to WRF the built-up is modelled with the urban canyon concept, however, in MUKLIMO_3 the built-up is re- garded as a porous volume. This concept is more close to reality in open urban built-up zones, where urban canyons cannot be defined properly. The computational time efficiency and the replacement of urban canyon concept were the main reasons for using this model. In the case of local-scale climate modelling, the selection of land cover data is crucial. There are a number of possible data sources for this, such as the CORINE land cover, the USGS land use dataset, and the Open Street Map. These databases have their advantages, however, none of them have been designed to represent urban thermal reactions. In the field of urban climatology, the Local Climate Zone (LCZ) classification (Stewart, I.D. and Oke, T.R. 2012) is widely accepted as a representation of urban land use (Table 1) and is used to characterize the environment of the measurement sites (e.g. Siu, L.W. and Hart, M.A. 2013; Stewart, I.D. et al. 2014; Lehnert, M. et al. 2015) or to map 21Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33. different urban neighbourhoods (e.g. Lelovics, E. et al. 2014; Zheng, Y. et al. 2018). Therefore, this scheme can also be used as surface input for numerical modelling (Žuvela-Aloise, M. 2017; Kwok, Y.T. et al. 2019). The application of this scheme is advantageous because it is based on the thermal characteristics of the urban are- as, i.e. it can be linked to the UHI phenomenon, which is the most important climate modifica- tion in these areas. Appropriate application of LCZ in local-scale climate modelling provides a good basis for global comparisons or valida- tion, and the trends obtained from the results can be generalized. In this study the c l i m a t e p r o j e c - tion outputs of the EURO-CORDEX re- gional models, the MUKLIMO_3 micro- climatic numerical model, and the cuboid method (Früh, B. et al. 2011) were used to explore the combined effects of regional climate change and urban climate. The modelling process is based on LCZs as ap- propriate urban pa- rameterization. Our previous studies of the thermal indices draw attention to the impor- tance of this topic: in case of Szeged there is a remarkable increase in different thermal indices by the end of the century, namely, a strong warming trend can be expected (Skarbit, N. and Gál, T. 2016; Bokwa, A. et al. 2018). The purpose of this study is twofold: (i) Analysis and comparison of the pat- terns of the annual values of tropical nights (TNs) quantifying the thermal load of the cities in the Carpathian Basin in the current (1981–2010) and future climate change peri- ods based on two different future emission scenarios (RCP4.5 and RCP8.5). (ii) Determining the overall additional thermal load in urban areas relative to their natural surroundings for each scenario dur- ing these periods. Study areas The investigation focuses on different sized cities with different geographical background in the Carpathian Basin, mainly on low-ly- ing areas with moderate relief (Figure 1). The population of the selected 26 cities is between 20,000 and 1,675,000 (Table 2). Table 1. Built and land cover LCZ classes* Built types Land cover types LCZ 1 – Compact high-rise LCZ 2 – Compact midrise LCZ 3 – Compact low-rise LCZ 4 – Open high-rise LCZ 5 – Open midrise LCZ 6 – Open low-rise LCZ 7 – Lightweight low-rise LCZ 8 – Large low-rise LCZ 9 – Sparsely built LCZ 10 – Heavy industry LCZ A – Dense trees LCZ B – Scattered trees LCZ C – Bush, scrub LCZ D – Low plants LCZ E – Bare rock / paved LCZ F – Bare soil / sand LCZ G – Water *After Stewart, I.D. and Oke, T.R. 2012. Fig. 1. Locations of the studied cities in the Carpathian Basin (modelling domains marked by black frames) Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33.22 An analysis of the current and future ther- mal situation of all 26 cities would go beyond the scope of this paper because of the length limitation. Therefore, we illustrate our results by selecting five city size categories, and we analyse the thermal situation of one city per category in detail (marked by italics in Table 2). Methods In order to get detailed information about the local scale future changes of thermal effects the MUKLIMO_3 model (Sievers, U. 1995) and cuboid method (Früh, B. et al. 2011) were applied to achieve high spatial resolution results inside the cities. The model is non- hydrostatic and calculates atmospheric tem- perature, relative humidity and wind field in a 3D grid by solving the Reynolds-aver- aged Navier-Stokes equations. Parametri- zations are used for unresolved buildings, short-wave and long-wave radiation, bal- anced heat and moisture budgets in the soil (Sievers, U. and Zdunkowski, W. 1985). The initial meteorology conditions were ensured by a 1D profile from a reference station with- in the study area. To run the model, high-res- olution orography and land use distribution data were needed. The horizontal resolution of 100 m was adjusted, while the vertical resolution changes in height and more ac- curate near to the surface, where the essen- tial processes occur. The vertical resolution near to the surface is 10 m and increases by several steps to the top of the model domain where it is 100 m. For most cities 25 vertical layers were enough, however, in some cases, where the topography was more variable we applied more layers (in Budapest 35, Eger 36, Miskolc 38, Novi Sad 33, Pécs 36, Salgótarján 40, Tatabánya 35 and Veszprém 30 layers). For our analysis, the orography data was provided by EU-DEM. The MUKLIMO_3 ap- plies custom land use categories, thus, any land use classification system is usable if the necessary surface, vegetation and built-up parameters are provided. This property of the model allows applying any urban land use classification as an input for urban land- forms. Using this advantage, we could apply the LCZ system to describe the land use. We used Bechtel-method for LCZ mapping of the selected cities (see Figure 1), which ap- plies free-access data and software (Bechtel, B. et al. 2015, 2019). This method is the basis of World Urban Database and Access Portal Tools (WUDAPT) which is a scientific initia- tive aiming to develop a global database for urban climate modelling. For this study, we used several Landsat images from different dates, in order to achieve more reliable LCZ Table 2. Population of the studied cities* City size category by range of population in 1,000 inhabitants City Population, 1,000 inhabitants 1 (over 1,000) Budapest 1,675 2 (between 200 and 400) Timisoara (RO) Novi Sad (SRB) Oradea (RO) Debrecen 315 215 207 201 3 (between 100 and 199) Arad (RO) Szeged Miskolc Pécs Nyíregyháza Kecskemét Subotica (SRB) 169 162 158 147 120 110 100 4 (between 50 and 99) Székesfehérvár Zrenjanin (SRB) Szolnok Tatabánya Kaposvár Békéscsaba Veszprém Eger 96 80 70 68 63 59 55 52 5 (between 20 and 49) Hódmezővásárhely Baja Salgótarján Szekszárd Siófok Makó 44 36 35 32 25 23 *Results of the cities written in italics are presented in details in this paper. Source: For Hungarian cities: https://nyilvantarto.hu, for other cities: https://worldpopulationreview.com/ 23Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33. classification. This approach ensures that the yearly changes of agricultural processes or vegetation cycle do not affect the final LCZ maps. The process was verified with field surveys in order to avoid misclassifications. To represent the thermal effects of cli- mate change several climate indices were calculated and to obtain these climate indi- ces the so-called cuboid method was used, which is a statistical dynamical downscaling method (Früh, B. et al. 2011; Žuvela-Aloise, M. et al. 2014). This process is basically a tri-linear interpolation of air temperature, relative humidity and wind fields derived by MUKLIMO_3 simulations and produces 30 year mean of annual number of 6 differ- ent climate indices. The method assumes that heat load situations can occur in case of specified weather situations, which can be described by the mentioned variables. The necessary inputs for the calculation of these climate indices are a 30-year daily climate dataset from a reference station and 8 single- day MUKLIMO_3 simulations for two pre- vailing wind direction (16 simulations). In this study the model outputs related to the tropical nights (TNs) are presented. The process of climate change was exam- ined through two future periods, 2021‒2050 and 2071‒2100 as well as period 1981‒2010 was considered as reference. To obtain in- put climate data for the cuboid method, the Carpatclim database (Szalai, S. et al. 2013) was applied for the reference period. It provides meteorological daily data for the Carpathian Region in spatial resolution of 0.1°. The data- base does not cover the entire area of Hungary, thus, cities in the western part of the country were excluded from the research. For the 21st century, data of EURO- CORDEX model simulations with resolu- tion 0.11° were used (Jacob, D. et al. 2014). The selection of the simulations was based on whether they include the necessary bias- corrected variables for the cuboid method e.g. air temperature, relative humidity wind speed and direction. Accordingly, 12 simula- tions were selected, which apply scenarios RCP4.5 and RCP8.5 also. The cuboid method was executed for all model simulations and the results were averaged by the scenarios. Results and discussion Examples of urban TN patterns by city categories Due to its dense built-up and large spatial ex- tent, the number of TNs is relatively high in Budapest in the period of 1981–2010, and the maximum value in the city centre exceeds 20 (Figure 2, b). The number of TNs exceeds 10 in the remarkable part of the urban area and 15 in the interior. The pattern of higher val- ues extends to the northeast due to the com- pact structure of this area (LCZ 3). Relatively high values also appear in the south-eastern part of the city, which may be caused by the prevailing wind direction. The high values of the southern city centre are the results of the LCZ 8, as this type of zone contains large impervious surfaces. This LCZ also appears scattered outside the urban area, with values above 5, especially in the south. In both scenarios, there are clear changes in the period 2021–2050 compared to the refer- ence period, but there is no considerable dif- ference between them (Figure 2, c-d). TNmax values are 31 and 35, respectively, according to RCP4.5 and RCP8.5. In most parts of the city, the value of TN in both cases exceeds 15. There are differences between the sce- narios in the patterns of the values of 20 and 25, which are more extended into the North, South, and downtown areas for RCP8.5. At the end of the century, there will be strong changes compared to the period 2021–2050, and the differences between the scenarios will become more remarkable (Figure 2, e-f). According to RCP4.5 the number of TNs is over 20 in almost the whole city and the area of TNs over 25 is also expanded (TNmax is 42). At the centre, the typical value exceeds 30, but even values above 40 appear in a smaller area. For RCP8.5, TNmax is 71 and the number of TNs is over 40 almost throughout the city. The ex- tension of values above 50 is also noteworthy, while in the city centre the number exceeds 60. Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33.24 Fig. 2. LCZ map (a) and patterns of the tropical nights in Budapest (Hungary) in 1981‒2010 (b), in 2021‒2050 by RCP4.5 (c), in 2021‒2050 by RCP8.5 (d), in 2071‒2100 by RCP4.5 (e) and in 2071‒2100 by RCP8.5 (f). The prevailing wind direction are NW and E. 25Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33. In the case of Debrecen, the smaller popu- lation and areal extent compared to the capi- tal is reflected in the number of TNs. Its value exceeds 5 in the more densely built-up west- ern part of the city and in the LCZ 8 between 1981 and 2010 (Figure 3, b). In the city centre, values above 10 occur mostly in the area of compact LCZs, but values above 15 can also be observed in a relatively small area (with a maximum of 16). The effect of the second dominant wind direction (north-east) also appears in the pattern. There will be minimal changes in the period 2021–2050, the scenarios show similar results, and the change between them is negligible (Figure 3, c-d). The only considerable change is the increased area of TNs over 10 and 15, which appear with slightly different magni- tudes by scenarios. The maximum values are 17 and 19 according to the different RCPs. Remarkable changes is observed in the period of 2071–2100 and the differences be- tween the scenario values are relatively large (Figure 3, e-f). According to RCP4.5, in most parts of the city the values are over 10 and the number of TNs in the interior exceeds 15, while in LCZ 2 it exceeds 20 (TNmax is 25). For RCP8.5, the values are more than twice as high: in most parts of the city, the number of TNs exceeds 30, while in the centre it is greater even 40 (TNmax is 51). Considering Arad, which represents the next category of cities, the magnitude of TNs is similar, but slightly lower than in Debrecen. During the reference period, the number of TNs exceeds 5 in densely built- up areas and 10 in the small area of LCZ 8 (TNmax is 11) (Figure 4, b). The effect of the prevailing wind directions is reflected in the north-west extent of the pattern. In the case of this city, too, the near future will not bring major changes and the differ- ence between the scenarios is minimal, the maximum values are 13 and 14 (Figure 4, c-d). The change compared to the reference period is the increase in areas above 5 and 10. This increase occurs especially around the green area in the south-east and in the north-west- ern parts of the city. In 2071–2100, the changes according to RCP4.5 are noticeable, but not salient (Figure 4, e). The pattern is similar to the previous period, but the values are higher of about 5 (TNmax is 21). The other change is the appearance of values above 5 in the western part of the study area, which may be the result of the prevailing wind directions and the dense tree zone (LCZ A). For RCP8.5, the changes are twice as high (Figure 4, f) in almost the entire city and in densely built-up areas the number of TNs exceeds 30 and 40, respectively (TNmax is 50). Although the population of Zrenjanin is lower than that of the previous two cities, the number of TNs in the reference period is higher due to its location at a lower latitude (Figure 5, b). In almost the entire area of the city the value exceeds 5, except for the south- ern part. In the LCZ 8 and LCZ 5 it is over 10, while LCZ3 it has more than 15 (TNmax is 23). In the period 2021–2050 the pattern val- ues exceed 5 and 10, especially according to RCP8.5 (Figure 5, c-d). In addition, values above 15 appear scattered across the western and southeastern areas. The TNmax is 21 and 24 for the different RCPs. The lack of increase of the maximum values in case of RCP4.5 is the result of the application of different cli- mate input for the reference and future peri- ods (Carpatclim and EURO-CORDEX). Considering the results between 2071 and 2100, the changes are of a similar magnitude as in previous cities (Figure 5, e-f). At RCP4.5, the number of TNs is over 15 throughout the city. Besides, values above 20 appear scattered throughout the pattern, especially in LCZ 8 and LCZ 5. In the core (dominated by LCZ 3), the values exceed 25 and even 30 in a small area (TNmax is 31). The outcomes of RCP8.5 during this period show far high values, which are slightly more than twice as high as those of RCP4.5: almost the entire area of the city has more than 40 TNs and values above 50 appear in the previously mentioned zones and are above 60 in the city core (TNmax is 62). Hódmezővásárhely is the smallest among the example cities, however, this is not re- flected in the number of TNs (Figure 6, b). Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33.26 Fig. 3. LCZ map (a) and patterns of the tropical nights in Debrecen (Hungary) in 1981‒2010 (b), in 2021‒2050 by RCP4.5 (c), in 2021‒2050 by RCP8.5 (d), in 2071‒2100 by RCP4.5 (e) and in 2071‒2100 by RCP8.5 (f). The prevailing wind directions are S and NE. 27Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33. Fig. 4. LCZ map (a) and patterns of the tropical nights in Arad (Romania) in 1981‒2010 (b), in 2021‒2050 by RCP4.5 (c), in 2021‒2050 by RCP8.5 (d), in 2071‒2100 by RCP4.5 (e) and in 2071‒2100 by RCP8.5 (f). The prevailing wind directions are S and SE. Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33.28 Fig. 5. LCZ map (a) and patterns of the tropical nights in Zrenjanin (Serbia) in 1981‒2010 (b), in 2021‒2050 by RCP4.5 (c), in 2021‒2050 by RCP8.5 (d), in 2071‒2100 by RCP4.5 (e) and in 2071‒2100 by RCP8.5 (f). The prevailing wind directions are SE and NW. 29Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33. In the reference period, the city boundary almost coincides with line 5, while in LCZ 8 the numbers exceed 10 (TNmax is 11). In the near future, minor changes will take place, which is reflected in the expansion of value areas above 5 and 10 (Figure 6, c-d). The main difference between the scenarios is the increased area of values above 10, which is even greater for RCP8.5: not only LCZ 5 and LCZ 8 are affected, but also the N-NE part of the city. The maximum values according to RCP4.5 and RCP8.5 are 12 and 13, respectively. For 2071–2100, the change in RCP4.5 is not outstanding: the number of TNs exceeds 10 in the entire urban area and is greater than 15 in most of the city, which means the densely built-up south-west, the south-east LCZ 8 and the aforementioned north-northeast (Figure 6, e). The TNmax does not exceed 20, which is exceptional among the presented cities. For RCP8.5, the values in the whole urban area exceed 30 and are higher than 40 in the pre- viously mentioned areas, and exceed 45 in the south-west and south-east (TNmax is 48) (Figure 6, f). Urban-rural heat load differences According to Table 3, the number of TNs in the reference period does not exceed 5 in the ru- ral areas. In the urban areas the dispersion is high: the values are between 5 and 10 in most cities, while they exceed 10 in larger cities and 15 only in the southernmost ones (Novi Sad and Zrenjanin). In the period 2021–2050, there will be minimal changes compared to 1981–2010, and the difference between the scenarios is also minimal. For RCP4.5, rural values are still below 5 with the exception of 3 cities, while urban values are between 10 and 15. The deviation of the TNs of RCP8.5 from RCP4.5 during this period is only 1–2 nights. Remarkable changes appear in 2071–2100 es- pecially in case of RCP8.5 and the difference between the scenarios is enormous. While the rural TNs of RCP4.5 are below 5 in most cases and do not exceed 15, RCP8.5 values are basically between 15 and 25 (except extreme cases). The urban TNs of RCP4.5 are usually between 15 and 25, however, for RCP8.5 there are very few cities where this number does not exceed 30. Typical TN values are between 40 and 50, but in four cases the number ex- ceeds even 50 (e.g. Budapest). The differences among the cities are mostly determined by the location, size, topography and built-up (LCZ) types. The highest val- ues appear for larger and/or southern cities such as Budapest, Novi Sad and Zrenjanin. For smaller cities and/or cities with higher altitudes and latitudes (e.g. Salgótarján), the TN values are generally lower. The results – particularly the rural values in Table 3 – can be compared to Pieczka, I. et al. (2018), which is the only example of tropical night extrapolation in the Carpathian Basin. According to Pieczka, I. et al. (2018), in case of period 2021–2040 and 2081–2100 the values are 10 and 15–30 days higher, respec- tively. There are some possible explanations of these differences. Firstly, the time periods are different, and in case of 2081–2100 it could cause major differences since in theory the first 10 years should be less warm than the last 20 within the period of 2071–2100. Secondly, in our study the outputs of 12 different regional models were applied meanwhile and Pieczka, I. et al. (2018) presented the results only of a single model. As the temperature extrapola- tion of the models are also different, it could also explain partly of the above mentioned dif- ferences. Finally, the values presented in Table 3 are the spatial mean of LCZ D areas within the domains, therefore several local and mi- cro-scale climate effects (nearby water surfaces or forest areas, small scale terrain forms) oc- cur, which are not implemented in regional scale models. Consequently, the comparison of these values is not entirely correct. The built-up environment causes remarka- ble differences in the number of TNs between the rural and urban area. Table 3 clearly shows that the maximum difference in each city depends on the size and location of the city and the time period and scenario. These results clearly support the motivation for lo- cal-scale climate modelling, as regional-scale Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33.30 Fig. 6. LCZ map (a) and patterns of the tropical nights in Hódmezővásárhely (Hungary) in 1981‒2010 (b), in 2021‒2050 by RCP4.5 (c), in 2021‒2050 by RCP8.5 (d), in 2071‒2100 by RCP4.5 (e) and in 2071‒2100 by RCP8.5 (f). The prevailing wind directions are S and NW. modelling can only determine rural condi- tions, whereas at the local scale, urban and intra-urban conditions can also be explored. The knowledge gained in this way is very valuable, as this type of projection of the in- creasing heat load in cities varying from dis- trict to district during the century, allows the authorities and partly the individuals to take appropriate preventive measures to mitigate the expected negative effects. 31Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33. Conclusions In this study the future changes in the number of TNs were examined through three time pe- riods in several cities in the Carpathian Basin. The results reveal that both the size and lati- tude of cities affect the values that are higher in southern and larger cities. Inside the cities the built-up types, the location and prevailing wind directions are determinative factors. In general, the change in the number of TNs and the difference between the scenarios are not re- markable in 2021–2050, the substantial change will occur in 2071–2100, especially for RCP8.5. Our results show the extent to which different built-up types modify (actually amplify) differently the effects of climate change. In this way, they provide detailed information on future processes not only for the regions (rural areas), but also for the cities, which are already, but will continue to be, the primary sites of human activity. Therefore, these results can serve as a guide for urban planners and local authorities to create neighbourhoods that are more liveable and better adapted to future changes. Table 3. General information on the urban-rural difference in the mean number of tropical nights during the 21st century by cities* City category Period 1981–2010 2021–2050 2071–2100 Scenario RCP4.5 RCP8.5 RCP4.5 RCP8.5 R U R U R U R U R U 1 Budapest 3 13 8 20 8 22 12 29 31 54 2 Timisoara (RO) Novi Sad (SRB) Oradea (RO) Debrecen 3 5 2 2 13 18 6 12 2 6 2 1 12 17 6 13 2 7 2 2 13 19 7 15 4 11 4 3 20 26 11 21 20 32 18 15 50 53 34 46 3 Arad (RO) Szeged Miskolc Pécs Nyíregyháza Kecskemét Subotica (SRB) 1 1 1 2 1 1 1 5 11 3 6 9 10 10 2 2 1 5 1 1 1 6 15 6 13 10 14 10 2 2 1 5 1 2 1 7 16 7 14 12 15 11 3 3 3 9 2 3 2 11 22 14 21 17 22 16 18 15 14 28 11 16 12 35 47 35 48 40 47 38 4 Székesfehérvár Zrenjanin (SRB) Szolnok Tatabánya Kaposvár Békéscsaba Veszprém Eger 2 1 2 0 1 2 0 0 7 17 9 0 2 10 1 0 3 2 2 2 2 1 2 1 10 16 14 3 4 11 4 3 4 3 2 2 2 1 2 1 11 18 15 4 5 12 5 3 6 5 3 4 4 2 4 2 17 26 21 7 8 18 8 5 24 23 17 17 19 13 17 10 41 56 46 24 29 43 28 19 5 Hódmezővásárhely Baja Salgótarján Szekszárd Siófok Makó 3 2 0 2 3 2 8 6 0 6 9 5 3 3 1 3 6 2 11 11 2 10 13 7 3 3 1 3 6 2 12 12 3 10 14 8 5 5 2 6 11 4 18 18 4 16 22 13 22 24 11 20 36 19 44 46 17 36 54 37 *Values are spatial means of different LCZs: R = LCZ D, U = the warmest LCZ in the given city. City cat- egories and cities written in italics are explained in Table 2. Acknowledgements: This research was supported by the National Research, Development and Innovation Office, Hungary (NKFI K-120346) and Ministry of Human Capacities, Hungary grant 20391-3/2018/FEKUSTRAT and TUDFO/47138-1/2019-ITM. We acknowledge the CORDEX project for producing and making available their model output and the Deutscher Wetterdienst (DWD) for making available the MUKLIMO_3 model.’ Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33.32 REFERENCES Baccini, M., Biggeri, A., Accetta, G., Kosatsky, T., Katsouyanni, K., Analitis, A., Anderson, H.R., Bisanti, L., D’Ippoliti, D., Danova, J., Forsberg, B., Medina, S., Paldy, A., Rabczenko, D., Schindler, C. and Michelozzi, P. 2008. Heat effects on mortality in 15 European cities. Epidemiology 19. 711–719. Doi: 10.1097/EDE.0b013e318176bfcd Bartholy, J. and Pongrácz, R. 2018. A brief review of health-related issues occurring in urban areas related to global warming of 1.5 °C. Current Opinion in Environmental Sustainability 30. 123–132. Bechtel, B., Alexander, P.J., Böhner, J., Ching, J., Conrad, O., Feddema, J., Mills, G., See, L. and Stewart, I. 2015. Mapping local climate zones for a worldwide database of the form and function of cities. ISPRS International Journal of Geo-Information 4. 199–219. Bechtel, B., Alexander, P.J., Beck, C., Böhner, J., Broussed, O., Ching, J., Demuzere, M., Fonteg, C., Gál, T., Hidalgo, J., Hoffmann, P., Middel, A., Mills, G., Ren, C., See, L., Sismanidis, P., See, L. Verdonck, M-L., Xu, G. and Xu, Y. 2019. Generating WUDAPT Level 0 data – Current status of produc- tion and evaluation. Urban Climate 27. 24–45. Doi: 10.1016/j.uclim.2018.10.001 Bokwa, A., Dobrovolný, P., Gál, T., Geletič, J., Gulyás, Á., Hajto, M.J., Holec, J., Hollósi, B., Kielar, R., Lehnert, M., Skarbit, N., Šťastný, P., Švec, M., Unger, J., Walawender, J.P. and Žuvela- Aloise, M. 2018. Urban climate in Central European cities and global climate change. Acta Climatologica 51–52. 7–35. Doi: 10.14232/acta.clim.2018.52.1 Bruse, M. and Fleer, H. 1998. Simulating surface– plant–air interactions inside urban environments with a three dimensional numerical model. Environmental Modelling & Software 13. 373−384. Doi: 10.1016/S1364-8152(98)00042-5 Früh, B., Becker, P., Deutschländer, T., Hessel, J.D., Kossmann, M., Mieskes, I., Namyslo, J., Roos, M., Sievers, U., Steigerwald, T., Turau, H. and Wienert, U. 2011. Estimation of climate-change impacts on the urban heat load using an urban climate model and regional climate projections. Journal of Applied Meteorology and Climatology 50. 167–184. Doi: 10.1175/2010JAMC2377.1 IPCC 2018. Global Warming of 1.5 °C. An IPCC Special Report on the impacts of global warming of 1.5 °C above pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of climate change, sustainable development, and ef- forts to eradicate poverty. Eds.: Masson-Delmotte, V., Zhai, P., Pörtner, H.-O., Roberts, D., Skea, J., Shukla, P.R., Pirani, A., Moufouma-Okia, W., Péan, C., Pidcock, R., Connors, S., Matthews, J.B.R., Chen, Y., Zhou, X., Gomis, M.I., Lonnoy, E., Maycock, T., Tignor, M. and Waterfield, T. Available at https://www.ipcc.ch/sr15/ Jacob, D., Petersen, J., Eggert, B., Alias, A., Christensen, O. B., Bouwer, L., Braun, A., Colette, A., Déqué, M., Georgievski, G., Georgopoulou, E., Gobiet, A., Menut, L., Nikulin, G., Haensler, A., Hempelmann, N., Jones, C., Keuler, K., Kovats, S., Kröner, N., Kotlarski, S., Kriegsmann, A., Martin, E., Meijgaard, E., Moseley, C., Pfeifer, S., Preuschmann, S., Radermacher, C., Radtke, K., Rechid, D., Rounsevell, 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. Doi: 10.1007/s10113-013-0499-2 Kusaka, H., Kondo, H., Kikegawa, Y. and Kimura, F. 2001. A simple single-layer urban canopy model for atmospheric models: comparison with multi-layer and slab models. Boundary-Layer Meteorology 101. 329–358. Doi: 10.1023/A:1019207923078 Kwok, Y.T., Schoetter, R., Lau, K.K-L., Hidalgo, J., Ren, C., Pigeon, G. and Masson, V. 2019. How well does the local climate zone scheme discern the thermal environment of Toulouse (France)? An analysis using numerical simulation data. International Journal of Climatology 39. 5292–5315. Doi: 10.1002/joc.6140 Lee, S.-H., Lee, H., Park, S.B., Woo, J.W., Lee, D.I. and Baik, J.J. 2016. Impacts of in-canyon vegetation and canyon aspect ratio on the thermal environment of street canyons: numerical investigation using a coupled WRF-VUCM model. Quarterly Journal of the Royal Meteorological Society 142. 2562–2578. Doi: 10.1002/qj.2847 Lehnert, M., Geletič, J., Husák, J. and Vysoudil, M. 2015. Urban field classification by “local climate zones” in a medium-sized Central European city: the case of Olomouc (Czech Republic). Theoretical and Applied Climatology 122. 531–541. Doi: 10.1007/ s00704-014-1309-6 Lelovics, E., Unger, J., Gál, T. and Gál, C.V. 2014. Design of an urban monitoring network based on Local Climate Zone mapping and temperature pattern modelling. Climate Research 61. 51–62. Doi: 10.3354/cr01220 Lemonsu, A., Masson, V., Shashua-Bar, L., Erell, E. and Pearlmutter, D. 2012. Inclusion of vegetation in the Town Energy Balance model for modelling urban green areas. Geoscientific Model Development 5. 1377–1393. Doi: 10.5194/gmd-5-1377-2012 Martilli, A., Clappier, A. and Rotach, M.W. 2002. An urban surface exchange parameterisation for mesoscale models. Boundary-Layer Meteorology 104. 261–304. Doi: 10.1023/A:1016099921195 33Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33. Masson, V. 2000. A physically-based scheme for the urban energy budget in atmospheric models. Boundary-Layer Meteorology 94. 357−397. Ministry of the Interior of Hungary. Available at https://nyilvantarto.hu (last accessed: 16.11.2020) Molnár, G., Kovács, A. and Gál, T. 2020. How does anthropogenic heating affect the thermal environ- ment in a medium-sized Central European city? A case study in Szeged, Hungary. Urban Climate 34.100673. Doi: 10.1016/j.uclim.2020.100673 Oke, T.R., Mills, G., Christen, A. and Voogt, J.A. 2017. Urban Climates. Cambridge, Cambridge University Press. Pieczka, I., Pongrácz, R. and Bartholy, J. 2018. Future temperature projections for Hungary based on RegCM4.3 simulations using new Representative Concentration Pathways scenarios. International Journal of Global Warming 15. 277–292. Ryu, Y.H., Bou-Zeid, E., Wang, Z.-H. and Smith, J.A. 2016. Realistic representation of urban trees in an urban canopy model. Boundary-Layer Meteorology 159. 193–220. Doi: 10.1007/s10546-015-0120-y Sievers, U. and Zdunkowski, W. 1985. A numerical simulation scheme for the albedo of city street canyons. Boundary-Layer Meteorology 33. 245–257. Sievers , U. 1995 . Veral lgemeinerung der Stromfunktionsmethode auf drei Dimensionen. Meteorologische Zeitschrift 4. 3–15. Siu, L.W. and Hart, M.A. 2013. Quantifying urban heat island intensity in Hong Kong SAR, China. Environmental Monitoring and Assessment 185. 4383–4398. Doi: 10.1007/s10661-012-2876-6 Skarbit, N. and Gál, T. 2016. Projection of intra-urban modification of nighttime climate indices during the 21st century. Hungarian Geographical Bulletin 65. (2): 181–193. Doi: 10.15201/hungeobull.65.2.8 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. 2013. Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate. Cambridge and New York, Cambridge University Press. Stewart, I.D. and Oke, T.R. 2012. Local Climate Zones for urban temperature studies. Bulletin of American Meteorological Society 93. 1879–1900. Doi: 10.1175/ BAMS-D-11-00019.1 Stewart, I.D., Oke, T.R. and Krayenhoff, E.S. 2014. Evaluation of the ‘local climate zone’ scheme using temperature observations and model simulations. International Journal of Climatology 34. 1062–1080. Doi: 10.1002/joc.3746 Szalai, S., Auer, I., Hiebl, J., Milkovich, J., Radim, T., Stepanek, P., Zahradnicek, P., Bihari, Z., Lakatos, M., Szentimrey, T., Limanowka, D., Kilar, P., Cheval, S., Deak, Gy., Mihic, D., Antolovic, I., Mihajlovic, V., Nejedlik, P., Stastny, P., Mikulova, K., Nabyvanets, I., Skyryk, O., Krakovskaya, S., Vogt, J., Antofie, T. and Spinoni, J. 2013. Climate of the Greater Carpathian Region. Final Technical Report. Available at www. carpatclim-eu.org/pages/download/ Van Vuuren, D.P., Edmonds, J., Kainuma, M., Riahi, K., Thomson, A., Hibbard, K., Hurtt, G.C., Kram, T., Krey, V., Lamarque, J-F., Masui, T., Meinshausen, M., Nakicenovic, N., Smith, S.J. and Rose, S.K. 2011. The representative concentration pathways: an overview. Climatic Change 109. 5–31. Doi: 10.1007/s10584-011-0148-z World Population Review. Available at https://world- populationreview.com/ (last accessed: 06.11.2020) Zheng, Y., Ren, C., Xu, Y., Wang, R., Ho, J., Lau, K. and Ng, E. 2018. GIS-based mapping of Local Climate Zone in the high-density city of Hong Kong. Urban Climate 24. 419–448. Doi: 10.1007/ s10661-020-08608-4 Žuvela-Aloise, M., Koch, R., Neureiter, A., Böhm, R. and Buchholz, S. 2014. Reconstructing urban climate of Vienna based on historical maps dating to the early instrumental period. Urban Climate 10. 490–508. Doi: 10.1016/j.uclim.2014.04.002 Žuvela-Aloise, M. 2017. Enhancement of urban heat load through social inequalities on an example of a fictional city King’s Landing. International Journal of Biometeorology 61. 527–539. Doi: 10.1007/s00484- 016-1230-z Gál, T. et al. Hungarian Geographical Bulletin 70 (2021) (1) 19–33.34