the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 10 Feb 2025
Impact of urban canopy parameters on urbanization induced modifications of climate
Abstract. Urban areas are characterized by modifications of the local climate leading to a so called urban meteorology island (UMI). UMI is the results of different physical properties of surfaces in cities compared to their rural surroundings. In this study we performed a set of multi-year simulations with the Weather Research and Forecast (WRF) model and two urban schemes to investigate the sensitivity of the urban climate modifications (or UMI) on changes in characteristics of the urban environment, described in models by the so-called urban canopy parameters (UCP). Our results reveal a high sensitivity of urban-induced changes in all mentioned meteorological variables to the alterations of UCP. Temperature in urban areas is mainly influenced by changes in urban fraction, roof albedo, green roofs with irrigation and also by anthropogenic heat in winter, with a magnitude around 0.5 °C. On the contrary, urban wind speed is impacted rather by parameters that describe the urban morphology. Our study also shows substantial differences between both urban models used, mainly in urban-induced temperature in winter. The results of the study can also be used as a primary evaluation of different mitigation strategies represented by changes in UCP values. The decrease of urban fraction and the increase of roof albedo seem to be the most suitable possibilities to reduce the intensity of the urban heat island in summer, vegetation-covered roofs have a noticeable impact only if they are also irrigated.
- Preprint
(4846 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-388', Anonymous Referee #1, 03 Mar 2025
Review of the manuscript egusphere-2025-388
Impact of urban canopy parameters on urbanization induced modifications of climate
By Jan Karlicky, Jachym Bareš, and Peter Huszar
Summary: in this study the WRF mesoscale model is employed over a domain in mid to western Europe in order to study the model sensitivity of the modelled urban meteorology island to urban morphological parameter values.
Recommendation: Major revision
Major remarks:
- In the methodology and discussion sections I miss a reasoning how the range of which the UCPs have been varied? Are all parameter values reasonable?
- In the modelling strategy misses how the model simulations were set up in the time domain. I.e. is the experiment based on 1 long free running simulation? Or does it consist of a series of e.g. 48 or 72 h forecasts that have been glued together? Would a different strategy result in a different answer to the research question?
- The paper does not mention spinup of temperature of buildings (brick). In hot summer days, the spin up can take 3 days or more if you make a cold start. Could the authors analyse this in more detail, e.g. through time series analysis of the storage heat flux. If the daily max storage flux is more or less the same for a number of consecutive days, then the spin up is well established.
- On temperature evaluation: at what height was the observation taken, and is that the same height as representative in the model. In the model, did you evaluate the T2M or the air temperature within the canyon? In general the paper could discuss in more detail the discrepancy between local observations, gridded observations (E-OBS) and model results. Their “footprint” is not the same, which cannot be avoided, but should be discussed.
- Ln 115: The resulting UHI intensity… The UHI is not well defined yet in this study. Do you evaluate the statistics on all hourly UHI values? Or do you take the maximum daily UHI (lets say 2-3 h after sunset) and evaluate the statistics over multiple years? Also the study generates UHI statistics for multiple cities across the European domain, so then which grid cells are used to categorize “urban” and “rural”. I.e. how many km near a city is considered and the rural area belonging to a certain city/urban area.
- Concerning the visualizations of the model sensitivities: The model results are now presented as differences in meteorological values between simulations. Why not plot UHI as function of albedo, and other properties, so we learn about a slope between the UHI and the varied parameter, like in Steeneveld et al. (2019). I think it may attracts a wider audience from the field of the city planning and design.
- Apart from the caveats above I find the research quite solid, but I miss a bit where is the innovation in the manuscript. The study does not introduce new and novel observational datasets, neither contributes to new model code. Can this be put forward in more detail?
Minor remarks
Ln 2: UMI is the results -> UMI is the result
Ln 3: Forecast -> Forecasting
Ln 38: builtup -> built up
Ln 42: I think this paper (particularly Fig 5) is a reasonable first attempt (despite in a 1D WRF context): https://doi.org/10.1016/j.resconrec.2016.12.002
Ln 53: horizontal resolution -> I have a little preference here to replace resolution with “grid spacing”. The resolution is then about 5*the grid spacing
Ln 54: 40 model layers : is this not little small these days? Please also document what is the height of the first model level, and the typical number of layers in the daytime PBL.
Ln 61-66: The list of parameterizations is well documented for reproducibility reasons, but at the same time they have not been justified or defended. Why is exactly this combi of settings suitable for answering the research question?
Figure 1: Please add scale bar and North Arrow
Ln 81: “The sensitivity of model simulations on specific urban canopy parameters” -> “The sensitivity of model results to specific urban canopy parameter values”
Table 5: the header needs to say we look at statistics for air temperature. In addition, is it 2-m air temperature? And also: if one verifies against E-OBS, I would expect to only use rural grid cells, not urban grid cells, since E-OBS does not represent cities well.
Section 3.1: I agree these statistics should be presented, but it would be good too to say here something whether the biases are in line with similar WRF studies for European cities. To put the results in a wider context.
Ln 117: but in winter there is an overestimation of simulation with BEP+BEM model with a magnitude above 1.5 ◦C. Do you understand why? Is that due to a poor representation of the anthropogenic heat flux, or a difference in target temperature between model and the real world?
Table 6: I do not understand what the column “stations” represents here. Also I suggest to merge Tables 5 and 6, since the manuscript has many relatively small tables scattered over the manuscript. It is more comprehensive to bring then together.
Figure 2: please label the panels a,b,c, and d. In the caption add that we look at “2-m air temperature” and define winter and summer (DJF and JJA respectively). It is serving the reader that (s)he does not need to go back to the text how winter and summer have been defined.
Ln 123: Add “The” before Planetary Boundary Layer Height
Ln 151: “Significant differences … ”. Has this been tested using a statistical test? Rain is a challenging variable to get sufficient events to make differences significantly visible.
Figures 4-7: It would be helpful for the reader to include in the caption whether this effect is measured over all cities (so both spatially and temporal means) or for Prague only.
Ln 158: scientific notation. Here you use “ms −1” which is perfectly correct, but before I also have seen “m/s” and “g/kg”. Please make consistent throughout the whole manuscript. In this piece of the analysis maybe underline that the diagnostic 10-m wind speed is analsysed.
Ln 162: most important component of UMI in view of human living impact. Is that correct? It is the most used and most well-known UMI component, but the wind island is also important for ventilation against heat and air pollution… Maybe reword.
Figure 8: Caption: Maybe advance the caption towards “Modelled diurnal cycle of 2-m UHI intensity in dependence on urban fraction for city of Prague (or all cities if all cities were used)” . Similar upgrades to the caption can be applied to Figures 9 and 10. Also indicate whether the time axis is in UTC or local time. The model domain is wide enough to have an hour difference in start of the UHI between the west and the east of the domain so this may affect your statistics.
Figure 8: Could error bars or confidence intervals be added to these plots.4
Ln 169: it is important to remark that only the BEP+BEM scheme generates a negative UHI in the early morning till noon, while SLUCM does not. This is important information, since this urban cool island has been observed in multiple cities across Europe. Apparently BEP+BEM does a much better thing in representing this dynamics.
Ln 187: SLUCM assumes constant anthropogenic heat independent to outdoor conditions. Although this is true, SLUCM does have a diurnal and seasonal cycle on AH implemented, which may still favour the simulation.
Reference:
Steeneveld, G.J., J.O. Klompmaker, R.J. Groen,. A.A.M. Holtslag, 2018: An Urban Climate Assessment and Management tool for combined heat and air quality judgements at neighbourhood scales. Resources, Conservation, and Recycling, 132, 204-217.
Citation: https://doi.org/10.5194/egusphere-2025-388-RC1 -
AC1: 'Reply on RC1', Jan Karlický, 15 May 2025
The author’s response to Anonymous Referee #1
We would like to thank to Anonymous Referee #1 for all comments, suggestions and corrections in his review of our manuscript. We addressed all and below our point-by-point responses follow:
MAJOR REMARKS
Referee’s Comment #1: In the methodology and discussion sections I miss a reasoning how the range of which the UCPs have been varied? Are all parameter values reasonable?
Author’s response: Some explanation of UCPs settings is on lines 87–88: “Further, altered values of chosen UCP are set with regard to their potential realism, limits and possibilities constrained by urban models.” It means that values are set following WRF model limits (e.g. in case of building height), physical limits (e.g. albedo between 0 and 1) or realistic range of values (e.g. urban fraction values below 0.6 are physically possible, but not reasonable and very unusual for central-European cities). The sentence in lines 87–88 is improved in the revised version of the text.
Referee’s Comment #2: In the modelling strategy misses how the model simulations were set up in the time domain. I.e. is the experiment based on 1 long free running simulation? Or does it consist of a series of e.g. 48 or 72 h forecasts that have been glued together? Would a different strategy result in a different answer to the research question?
Author’s response: The experiment is not based on long free-running simulations through the whole 2015–2019 period (which would be the ideal option), but based on 10 three-month simulations covering winters and summers in the given period – this choice was made to decrease the computational demand. The information will be better explained in the revised manuscript. Such a strategy has only a minor effect on results, because the domain is not so large thus a general meteorology is strictly driven by ERA-5 data and spin-up interval of urban schemes is short (two or three days in maximum), and could be neglected in three months simulation. However, the mentioned strategy of gluing of 48 or 72 hour forecasts would have a significant impact on results, in case of starting of all simulations from one initial conditions.
Referee’s Comment #3: The paper does not mention spinup of temperature of buildings (brick). In hot summer days, the spin up can take 3 days or more if you make a cold start. Could the authors analyse this in more detail, e.g. through time series analysis of the storage heat flux. If the daily max storage flux is more or less the same for a number of consecutive days, then the spin up is well established.
Author’s response: As mentioned above, the experiment is based on free-running three-month simulations, so the effect of initial setup (conditions) of variables in urban scheme is significantly suppressed. Therefore, we believe that no special analysis is necessary to determine if the adjustment (spin-up period) within urban scheme happens after two or three days.
Referee’s Comment #4: On temperature evaluation: at what height was the observation taken, and is that the same height as representative in the model. In the model, did you evaluate the T2M or the air temperature within the canyon? In general the paper could discuss in more detail the discrepancy between local observations, gridded observations (E-OBS) and model results. Their “footprint” is not the same, which cannot be avoided, but should be discussed.
Author’s response: Station data provided by the Czech Hydro-Meteorological Institute (CHMI) include temperature in the height of 2 m, same as in the E-OBS data and model outputs used in the study. This information is added to the reviewed version of the text. Model 2-m temperature is standardly compared against E-OBS temperatures in studies including model validation. Also in case of urban areas, 2-m temperature is considered in model results and observations. We agree that local observations or gridded data should be used carefully for validation purposes, because they represent either specific locations or some interpolation method, leading to imperfect comparison of values in model grid-boxes for many reasons (although correction on the same height is included in stations). The note is added into the revised text, but it have to also be mentioned that the model validation is not the main aim of the study.
Referee’s Comment #5: Ln 115: The resulting UHI intensity… The UHI is not well defined yet in this study. Do you evaluate the statistics on all hourly UHI values? Or do you take the maximum daily UHI (lets say 2-3 h after sunset) and evaluate the statistics over multiple years? Also the study generates UHI statistics for multiple cities across the European domain, so then which grid cells are used to categorize “urban” and “rural”. I.e. how many km near a city is considered and the rural area belonging to a certain city/urban area.
Author’s response: In general, UHI intensity means difference between urban and rural temperature. In line 115 and Table 6, there is the overall daily mean UHI intensity, separately for summer and winter over the whole simulation period. Here, model urban and rural grid-boxes closest to stations are taken into consideration, as stated in line 115. On the other hand, the process of determining of urban and rural grid-boxes of/around specific cities for results in section 3.2 is described in paragraph including lines 92–99 in section 2.2. The location of the ring of considered rural grid-boxes around cities is in the distance of three grid-boxes (i.e. 27 km) from urban grid-boxes. A short note that reminds the readers the application of this procedure is added to the start of section 3.2, also the sentence in line 115 is modified to highlight the difference of UHI intensity determination in this validation part.
Referee’s Comment #6: Concerning the visualizations of the model sensitivities: The model results are now presented as differences in meteorological values between simulations. Why not plot UHI as function of albedo, and other properties, so we learn about a slope between the UHI and the varied parameter, like in Steeneveld et al. (2019). I think it may attracts a wider audience from the field of the city planning and design.
Author’s response: Even though we appreciate this proposed presentation as clear and transparent, we decided to keep the original form for two reasons: First, we want to avoid making a function from few points in graph and secondly we want to limit number of figures – we would have to present 7 dependencies or functions, which means 7 figures for every meteorological variable instead of the two in the current version manuscript. However, we thank for bring into our attention another study focusing on similar topic, we will use it for another comparison in the discussion.
Referee’s Comment #7: Apart from the caveats above I find the research quite solid, but I miss a bit where is the innovation in the manuscript. The study does not introduce new and novel observational datasets, neither contributes to new model code. Can this be put forward in more detail?
Author’s response: We agree that the study does not introduce a new observation neither new modeling approach. The novelty or added value of the study lies in its complexity, where impact of several UCPs on urban effects in several meteorological variables is investigated, while previous studies showed impact of changing of one or two UCPs or impact of more UCPs on only one variable as e.g. Steeneveld et al. (2019).
MINOR REMARKS
Referee’s Comment: Ln 2: UMI is the results -> UMI is the result
Author’s response: Thanks for finding the mistake, corrected.
Referee’s Comment: Ln 3: Forecast -> Forecasting
Author’s response: Repaired.
Referee’s Comment: Ln 38: builtup -> built up
Author’s response: Repaired to “built-up” as adjective.
Referee’s Comment: Ln 42: I think this paper (particularly Fig 5) is a reasonable first attempt (despite in a 1D WRF context): https://doi.org/10.1016/j.resconrec.2016.12.002
Author’s response: We agree, the study is added in the Introduction and also to the Discussion.
Referee’s Comment: Ln 53: horizontal resolution -> I have a little preference here to replace resolution with “grid spacing”. The resolution is then about 5 times the grid spacing
Author’s response: We accept the comment, words “with horizontal resolution 9 km” are changed to “with grid spacing of 9 km”.
Referee’s Comment: Ln 54: 40 model layers: is this not little small these days? Please also document what is the height of the first model level, and the typical number of layers in the daytime PBL.
Author’s response: Number of vertical layers is a compromise between increasing of accuracy and computational costs. Moreover, more vertical layers and thus lower first layer creates numerical problems within SLUCM scheme. In the model setup used, first model layer ends in 49 m above ground, and if we consider the height of daytime PBL as 1 km, 8 model layer are under this height. The information about first model layer height is added into the text.
Referee’s Comment: Ln 61-66: The list of parameterizations is well documented for reproducibility reasons, but at the same time they have not been justified or defended. Why is exactly this combi of settings suitable for answering the research question?
Author’s response: This combination of parameterization setting was chosen based on previous study of the main author (Karlický et al., 2020), which targeted on the same model domain. This information is added into the revised text.
Referee’s Comment: Figure 1: Please add scale bar and North Arrow
Author’s response: This is added in revision version.
Referee’s Comment: Ln 81: “The sensitivity of model simulations on specific urban canopy parameters” -> “The sensitivity of model results to specific urban canopy parameter values”
Author’s response: Corrected in the text.
Referee’s Comment: Table 5: the header needs to say we look at statistics for air temperature. In addition, is it 2-m air temperature? And also: if one verifies against E-OBS, I would expect to only use rural grid cells, not urban grid cells, since E-OBS does not represent cities well.
Author’s response: Values in Table means 2-m air temperature, this information is added into table caption. Although E-OBS does not represent cities on street canyon scale, it has a similar grid spacing as model simulations (0.1° vs. 9 km), thus effect of big cities should be also captured and there is no reason for excluding of urban grid-boxes from the comparison.
Referee’s Comment: Section 3.1: I agree these statistics should be presented, but it would be good too to say here something whether the biases are in line with similar WRF studies for European cities. To put the results in a wider context.
Author’s response: Some discussion of results presented in section 3.1 is in the beginning of discussion section, but we agree that general WRF biases are not commented in a wider context. This is added in the revised version of text.
Referee’s Comment: Ln 117: but in winter there is an overestimation of simulation with BEP+BEM model with a magnitude above 1.5 ◦C. Do you understand why? Is that due to a poor representation of the anthropogenic heat flux, or a difference in target temperature between model and the real world?
Author’s response: The explanation is mentioned in the beginning of discussion section – “The high winter overestimation of the UHI intensity by the BEP+BEM simulation is probably caused by an incorrect description of heat fluxes from buildings (or explicit anthropogenic heat calculation) and was also noticed in previous studies (Liao et al., 2014; Karlický et al., 2018), with a similar magnitude.”
Referee’s Comment: Table 6: I do not understand what the column “stations” represents here. Also I suggest to merge Tables 5 and 6, since the manuscript has many relatively small tables scattered over the manuscript. It is more comprehensive to bring then together.
Author’s response: We find merging of both tables problematic, because they describe different quantities (2-m air temperature vs. UHI intensities thus temperature differences) and validation against different reference data (E-OBS vs. station data). The column “stations” stands for values derived from observations. The name of the column is changed to observation, with an explanation in table caption.
Referee’s Comment: Figure 2: please label the panels a,b,c, and d. In the caption add that we look at “2-m air temperature” and define winter and summer (DJF and JJA respectively). It is serving the reader that (s)he does not need to go back to the text how winter and summer have been defined.
Author’s response: Labels added and caption modified by the proposed explanation.
Referee’s Comment: Ln 123: Add “The” before Planetary Boundary Layer Height
Author’s response: Added in the revised text.
Referee’s Comment: Ln 151: “Significant differences … ”. Has this been tested using a statistical test? Rain is a challenging variable to get sufficient events to make differences significantly visible.
Author’s response: Test of statistical significance was not performed and would be hardly applicable, when we investigate “changes in differences” in Fig. 6. The sentence is rephrased to avoid word “significant” to prevent understanding as ”statistical significant.”
Referee’s Comment: Figures 4-7: It would be helpful for the reader to include in the caption whether this effect is measured over all cities (so both spatially and temporal means) or for Prague only.
Author’s response: These figures include averaged effects within all selected cities, this is added in captions of the mentioned figures and also Figures 8–10.
Referee’s Comment: Ln 158: scientific notation. Here you use “ms −1” which is perfectly correct, but before I also have seen “m/s” and “g/kg”. Please make consistent throughout the whole manuscript. In this piece of the analysis maybe underline that the diagnostic 10-m wind speed is analyzed.
Author’s response: Following instructions of ACP, units should be formatted in form of negative exponent, thus all units are in form “m s-1” in the revised manuscript. Wind speed is considered in the first model layer (as stated in caption of Fig. 3), this information and the reason why (wrong values of 10-m wind speed in SLUCM simulations) is added into the text.
Referee’s Comment: Ln 162: most important component of UMI in view of human living impact. Is that correct? It is the most used and most well-known UMI component, but the wind island is also important for ventilation against heat and air pollution… Maybe reword.
Author’s response: This sentence is rephrased to avoid potentially unclear statements about importance of specific components of UMI for human life.
Referee’s Comment: Figure 8: Caption: Maybe advance the caption towards “Modelled diurnal cycle of 2-m UHI intensity in dependence on urban fraction for city of Prague (or all cities if all cities were used)” . Similar upgrades to the caption can be applied to Figures 9 and 10. Also indicate whether the time axis is in UTC or local time. The model domain is wide enough to have an hour difference in start of the UHI between the west and the east of the domain so this may affect your statistics.
Author’s response: The caption is modified, time axis added (UTC). The selected cities (Fig. 1) are not very distinct in terms of longitude, with a maximal deviation till 7° from Prague’s latitude, which corresponds to time shifts less than half hour, thus no transition to local time is adopted.
Referee’s Comment: Figure 8: Could error bars or confidence intervals be added to these plots.
Author’s response: To preserve good visibility of the figure, no error bars or confidence intervals are added.
Referee’s Comment: Ln 169: it is important to remark that only the BEP+BEM scheme generates a negative UHI in the early morning till noon, while SLUCM does not. This is important information, since this urban cool island has been observed in multiple cities across Europe. Apparently BEP+BEM does a much better thing in representing this dynamics.
Author’s response: We add a sentence including this finding.
Referee’s Comment: Ln 187: SLUCM assumes constant anthropogenic heat independent to outdoor conditions. Although this is true, SLUCM does have a diurnal and seasonal cycle on AH implemented, which may still favour the simulation.
Author’s response: Only diurnal profile of anthropogenic heat is included within SLUCM in mentioned WRF version (4.3.3), no seasonal profiles (on difference on anthropogenic latent heat), so the sentence is correct.
Karlický, J., Huszár, P., Halenka, T., Belda, M., Žák, M., Pišoft, P., and Mikšovský, J.: Multi-model comparison of urban heat island modelling approaches, Atmospheric Chemistry and Physics, 18, 10 655–10 674, 2018.
Karlický, J., Huszár, P., Nováková, T., Belda, M., Švábik, F., Ďoubalová, J., and Halenka, T.: The “urban meteorology island”: a multi-model ensemble analysis, Atmospheric Chemistry and Physics, 20, 15 061–15 077, 2020.
Liao, J., Wang, T., Wang, X., Xie, M., Jiang, Z., Huang, X., and Zhu, J.: Impacts of different urban canopy schemes in WRF/Chem on regional climate and air quality in Yangtze River Delta, China, Atmospheric Research, 145-146, 226–243, 2014.
Citation: https://doi.org/10.5194/egusphere-2025-388-AC1
-
RC2: 'Comment on egusphere-2025-388', Anonymous Referee #2, 31 Mar 2025
In this study, the impacts of urban canopy parameters on urban climate were analyzed with sensitivity numerical experiments using WRF model. The methodology and results are too vague in current status and lots improvements are needed.
- The main result that “the urban temperature is mainly influenced by roof albedo, green roof (which modifies roof albedo and available soil moist) and anthropogenic heat release, and urban wind speed is impacted by urban morphology” is not a new finding, which has been documents in lots of publications using both off-line urban canopy models and UCM-coupled simulations (e.g. Taha 1997). It is also not suitable to take urban fraction as an UCP parameters, because it describes the urban land cover. And the decrease of urban fraction definitely leads to a decreasing of UHI, because there is no cities when urban fraction decrease to zero.
- The manuscript did not give a clear definition on how UMI was calculated. Previous studies have shown the urbanization impacted regional climate (e.g. Trusilova et al. 2008, Zhang et al. 2010, Wang et al. 2013), the intensity of UMI will strongly depend on how ‘urban’ and ‘rural’ area are selected. The simulation was carried out on a relative coarse resolution (9-km), how urban areas are described when the urban fraction is fixed (base case in Table 2)
- Due to the different design of the parameterization schemes, the T2m and Wind speed at 10m have different physical meaning in SLUCM and BEP/BEM (Sun et al. 2021), it should be careful to select the observations to evaluate the model output. I am not sure whether E-OBS is suitable for the model performance validation, but the bias pattern between the base model and E-OBS also (underestimation/overestimation over different areas) will impact on the intensity of UMI, this should be discussed.
Citation: https://doi.org/10.5194/egusphere-2025-388-RC2 -
AC2: 'Reply on RC2', Jan Karlický, 15 May 2025
The author’s response to Anonymous Referee #2
We would like to thank to Anonymous Referee #2 for all comments, suggestions and corrections in his review of our manuscript. We addressed all and below our point-by-point responses follow:
Referee’s Comment #1: The main result that “the urban temperature is mainly influenced by roof albedo, green roof (which modifies roof albedo and available soil moist) and anthropogenic heat release, and urban wind speed is impacted by urban morphology” is not a new finding, which has been documents in lots of publications using both off-line urban canopy models and UCM-coupled simulations (e.g. Taha 1997). It is also not suitable to take urban fraction as an UCP parameters, because it describes the urban land cover. And the decrease of urban fraction definitely leads to a decreasing of UHI, because there is no cities when urban fraction decrease to zero.
Author’s response: We agree that the study describe the research that was partly investigated before, but the novelty or added value of the study lies in its complexity, where impact of several UCPs on urban effects in several meteorological variables is investigated, while previous studies showed impact of changing of one or two UCPs or impact of more UCPs on only one variable. Urban fraction is one of important characteristics of urban areas or cities, therefore it is also considered, although not directly associated with the street canopy geometry.
Referee’s Comment #2: The manuscript did not give a clear definition on how UMI was calculated. Previous studies have shown the urbanization impacted regional climate (e.g. Trusilova et al. 2008, Zhang et al. 2010, Wang et al. 2013), the intensity of UMI will strongly depend on how ‘urban’ and ‘rural’ area are selected. The simulation was carried out on a relative coarse resolution (9-km), how urban areas are described when the urban fraction is fixed (base case in Table 2)
Author’s response: The procedure how the magnitude of urban-induced effects on various meteorological variables (or UMI) is calculated is described in section 2.2, third paragraph. A short note reminding the use of this procedure is also added to the start of section 3.2. The cities selected for the study are listed in Table 3 and shown on Fig. 1, together with urban grid-boxes.
Referee’s Comment #3: Due to the different design of the parameterization schemes, the T2m and Wind speed at 10m have different physical meaning in SLUCM and BEP/BEM (Sun et al. 2021), it should be careful to select the observations to evaluate the model output. I am not sure whether E-OBS is suitable for the model performance validation, but the bias pattern between the base model and E-OBS also (underestimation/overestimation over different areas) will impact on the intensity of UMI, this should be discussed.
Author’s response: E-OBS data are used only for general validation of model simulations in case of 2-m air temperature, not specially for urban areas. The bias pattern is discussed more in section 4 of the revised manuscript, also in relation to urban areas. For evaluating of UHI intensities, station data for 2-m air temperature in and around Prague are used. In section 3.2, 2-m air temperature and also surface temperatures are considered, but the pattern is qualitatively similar, only with higher magnitudes in case of surface temperatures.
Citation: https://doi.org/10.5194/egusphere-2025-388-AC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 174 | 51 | 18 | 243 | 12 | 15 |
- HTML: 174
- PDF: 51
- XML: 18
- Total: 243
- BibTeX: 12
- EndNote: 15
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1