the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Barriers of urban hydro-meteorological simulation: a review
Abstract. Urban areas, characterized by dense populations and many socio-economic activities, increasingly suffer from floods, droughts, and heat stress due to land use and climate change. Traditionally, the urban thermal environment and water resources management have been studied separately, using urban land surface models (ULSMs) and urban hydrological models (UHMs). However, as our understanding deepens and the urgency to address future climate disasters grows, it becomes clear that hydrological disasters—such as floods, droughts, severe urban thermal environments, and more frequent heat waves—are actually not isolated events but compound events. This underscores the close interaction between the water cycle and the energy balance. Consequently, the existing separation between ULSMs and UHMs creates significant obstacles to better understanding urban hydrological and meteorological processes, which is crucial for addressing the high risks posed by climate change. Defining the future direction of process-based models for hydro-meteorological predictions and assessments is essential for better managing climate disasters and evaluating response measures in densely populated urban areas. Our review focuses on three critical aspects of urban hydro-meteorological simulation: similarities, differences, and gaps among different models; existing gaps in physical process implementations; and efforts, challenges, and potential for model coupling and integration. We find that ULSMs inadequately represent water surfaces and hydraulic systems, while UHMs lack explicit surface energy balance solutions and detailed building representations. Coupled models show potential for simulating urban hydro-meteorological environments, but face challenges at regional and neighborhood scales. Our review highlights the need for interdisciplinary communication between the urban climatology and urban water management communities to enhance urban hydro-meteorological simulation models.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Hydrology and Earth System Sciences.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.- Preprint
(1241 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 07 Apr 2025)
-
RC1: 'Comment on egusphere-2024-3988', Zhi-Hua Wang, 25 Jan 2025
reply
This paper presents a comprehensive review of the state-of-the-art urban climate and urban hydrological modeling. The review is timely and of great interest to the urban study community, especially because urban hydrological modeling has been lagging behind practices for sustainable urban development. While some technical details need to be further clarified (see my specific comments below), overall, the paper is technically sound and well written, and I thoroughly enjoyed reading the manuscript. I therefore recommend the paper to be accepted for publication after the following comments to be adequately addressed.
Specific comments
1. Lines 112: on the compound urban climate mitigation mechanisms, the following study provides a mathematical formalism that may worth considering:
Wang, Z.H. (2021). Compound environmental impact of urban mitigation strategies: Co-benefits, trade-offs, and unintended consequence. Sustainable Cities and Society, 75, 103284. https://doi.org/10.1016/j.scs.2021.103284
2. Line 160-161: “However, the multilayer models solve the vertical profiles of the atmospheric conditions”, the phrase “atmospheric conditions” is too general to describe the multilayer UCM, it is more precise to use “canopy-layer flows and momentum transport”.
3. Table 1: under SLUCM, the citation Wang et al., 2021 should be referring to
Wang, C., Wang, Z.H., & Ryu, Y.H. (2021). A single-layer urban canopy model with transmissive radiation exchange between trees and street canyons. Building and Environment, 191, 107593. https://doi.org/10.1016/j.buildenv.2021.107593
which is not included in the reference list.
4. Figure 1: I wonder if it is necessary to separately indicate the surface temperature and net radiation for heterogeneous landuse. For by the same token, sensible (and latent) heat fluxes for these facets are also different and should also be separately indicated. In addition, there is a downwelling Rnet and 7 upwelling Rnets, while the net radiation from the urban canopy should be the combination of them, and none of the individual components can be called Rnet I’d suggestion to keep the separate representation of surface temperatures as it is, but combine the radiation into a single Rnet (or with a downwelling radiation as Rdown and a upwelling component as Rup).
5. Lines 226-229, “The single-layer urban canopy model developed by Kusaka et al. (2001) (SLUCM) is very similar to TEB. The only differences are that the SLUCM in this version (Kusaka et al., 2001) includes the canyon orientation and diurnal change of solar azimuth angle, and the surface consists of several canyons with different orientations.” This statement is incomplete. In fact, Kusaka’s SLUCM, as implemented into WRF, contains a different parameterization scheme of radiation by discretizing the canyon facet and computing radiation on individual gridcells, whereas TEB uses the analytical formulae for in-canyon view factors. For simple rectangular canyons with only walls and roads (and short vegetation), Kusaka’s radiation scheme is a setback to the analytical formulation, but it opens the possibility to include radiative exchange by roughness elements presented in street canyons such as trees/blocks/vehicles.
6. Table 3, I don’t really understand how the temporal resolution of different ULSMs is determined. If the temporal resolution refers to the time intervals/steps used to solve the parameterization scheme, it varies widely depending on the discretizing (forward- or central-in-time finite difference) schemes, running platforms (offline or imbedded in regional climate models such as WRF), and applications. The time steps used to solve parameterization schemes can be as small as 1s (e.g. WRF-UCM for it used both spatial-temporal discretization for land-atmosphere interactions), or as large as 30 min. If the temporal resolution refers to the time scale for sampling the output, it is a rather arbitrary choice of the users. For instance, output of WRF-UCM is often sampled in hourly scale, like what is indicated in the table, but it can also be sampled at 3-hourly or 6-hrouly intervals for longterm (monthly to annual) simulations, but it can also be, in theory, sampled at 1s interval. My understanding is that the temporal resolution of all UCMs have no essential difference as their parameterization schemes represented by partial differential equations of land surface processes are all similar. The spatial scale for them does vary for SLUCM resolves the physical structure of the canyon, and building-resolving models has to resolve individual buildings, while slab models represent the aggregated urban landscapes.
7. Table 5: SLUMC should be SLUCM (same typo in Tables 6 and 7). Also, I am concerned about the naming of the urban canopy models. The discussion of the single-layer urban canopy models in this paper is largely based TEB, Kusaka’s UCM implemented in WRF, and the Arizona Single Layer Urban canopy Model (ASLUM, the name is used in Wang et al., 2021, Lipson et al., 2024, and Jongen et al., 2024). Yet TEB is separately discussed in this table, and SLUCM seemingly groups ASLUM and Kusaka’s model. Given the fact that Kusaka’s single-layer model is not further developed in a separate line, the representative SLUCM should be more properly named after ASLUM, as the latter is a coherent family of models developed by the same group of model developers in a continuous manner (Wang et al., 2013; Wang, 2014; Yang et al., 2015; Ryu et al., 2016; Wang et al., 2024).
8. Section 6: this part presents some thought-provoking questions that need to be pursued in future development of urban climate and urban environment modeling in depth. I would suggest the authors also include a brief discussion of the potential of AI and machine learning application in the field, given that these tools are increasingly adopted and some promising results generated from pioneering work in this field.
Citation: https://doi.org/10.5194/egusphere-2024-3988-RC1 -
RC2: 'Comment on egusphere-2024-3988', Anonymous Referee #2, 15 Mar 2025
reply
The manuscript reviews the main features of (1) urban land surface models (U-LSMs) mostly used for studying urban canyon microclimate and in urban climate studies and (2) urban hydrological models (U-HMs), mostly used in urban flood analysis. By comparison these two families of models, the article shows how distant the two modeling communities are, while theoretically approaching a problem – the solution of urban energy-water-carbon fluxes – that has a lot of common elements. The authors advocate for a better integration of the two modeling approaches with better communication across communities and better integration of hydrological processes and land-atmosphere interactions.
While it is fine to develop a model for a specific purpose, I overall agree with the call made in this manuscript and the importance to review for different communities (I suppose many scientists are not aware of the full landscape of models) the differences and similarities in the various modeling approaches – with the hope that soon rather than later models across the two fields could be developed. Based on my knowledge, the article is also mostly accurate in presenting the different modeling components and I have mostly minor comments listed below. The only broader comment is that the article is currently very long, and I suppose it will benefit by some shortening, even though admittedly I could not easily identify parts that can be trimmed, even though I give a few suggestions below.
Specific Comments.LL 6 and also LL 103-104. Please note that not all hydroclimatological events listed here will lead to “compound events”. “Compound events” is used a bit loosely in the manuscript.
LL 63. “…small scale heterogeneity” is also characteristic of many natural systems not only urban.
LL 72. At this stage is not clear what a distinction between a “urban meteorology tool” and “urban land surface model” is. A U-LSMs to be an urban meteorology tool needs to be coupled with a mesoscale weather model, otherwise at the best it can explore “micrometeorology” in the canyon, not the overall city meteorology.
LL 82-82. It might be worth referring to Jongen et al 2024, which is looking to “hydrological aspects” of these U-LSMs, already at this stage and move the later text here.
LL 89-90. I think this sentence is not clear, I would suggest rephrasing it.
LL 163-164. While it is true that they represent all urban facets, many CFD approaches prescribe surface temperature of urban facets or of a subset of urban facets, which means they are not solving for energy budget (or at least not fully solving for it).
LL 172-173. I would suggest rewriting this sentence, it reads awkwardly.
Table 1. For UT&C please see also Meili et al 2025, with further model developments.
LL 221. How are lateral water flows solved in SUEWS?
LL 255. Please note that a main advancement of UT&C is the capability to consider physiological and biophysical properties of vegetation, and thus being able to consider different vegetation types – at least in principle - which was not the case in any of the other models except partially VTUF-3D
LL 261. They take “wind speed” too as input. I suggest having the complete list of six environmental variables. Pr, Ta, RH, Ws, Rsw, Ldw, some models need atmospheric pressure too.
LL 276. The second acronym should be TUF-3D, without “V”.
LL 313 and below. Maybe using “grid cell” rather than “square” would be more aligned with the previous literature. Overall, I think this part can be shortened.
LL320-333. I find all this part a bit convoluted and not very clear, please consider rewriting and probably shortening, as it might not be so essential.
LL346-347. Please note that “urban block scale” and “neighborhood area” are quite vague and actually they might largely overlap, a clarification about scale might be helpful here.
LL 350. “present” rather than “utilise”
LL 402-403. Please note that data availability might be an issue to define branching and drainage structure in many urban environments, as sewage systems locations are often poorly known in old cities or not available in modern cities. This can be also discussed later on.
LL 435. Are you referring to water surface as “urban lakes” or “ponding water in various surfaces” the latter is likely solved in most U-LSMs.
LL 458. Please note that the shortwave and longwave radiation budget is different and not only sky view factors but view factors among different urban facets are required, at least for longwave exchange, it is also not true that all U-LSMs do not compute multiple reflections.
LL 459. “radiosity method” is unclear.
Table 5. Maybe it could be interesting subdividing between models using a multi-layer vs two- or three-layers approach for solving the urban canyon, as more than one layer is not lumped anymore.
LL 495. “area” rather than “ares”.
LL 537. Urban Hydrological Models not “Urban Heat Models”, which does not mean anything.
LL 598. Applying 1D Richards equation allows to solve for infiltration, but it is much more than an infiltration method, as it allows the solution of variably saturated flow in multiple soil layers.
LL 604-605. The description of the “Horton method” is wrong. It is the other way around, infiltration decreases exponentially up to an asymptotic constant value at large times, which is typically the saturated hydraulic conductivity.
LL 615. Please note that on most impervious surfaces (e.g., roofs, paved streets), the saturated hydraulic conductivity will be fundamentally zero, so why one should compute infiltration? This is not clear to me.
LL623. Do you mean Table 7?
LL 638. In VTUF-3D and UT&C transpiration is indeed simulated using an explicit solution of stomatal conductance, which is a function of photosynthesis, so plant photosynthesis need to be solved.
LL646-656. I think here, there is a bit of confusion around Penman-Monteith equation method. If a model solves the energy budget (i.e., latent heat) by solving for surface temperature/s, there is no need to use Penman-Monteith equation, which is actually a simplification of the overall energy budget and it does not guarantee energy budget closure. I think which models use Penman-Monteith and which aren’t should be better described and specified.
LL 692-694. Does this refer to canopy or ground vegetation? This is not clear.
LL723. The end of the sentence is not clear, which empirical models? Which calibration?
LL770-772. I am not sure I understand the question here and especially why it is formulated as a question. It is mostly a parameterization issue how to deal with these processes.
LL 778. I am not sure I understand why ones would like to have two-ways feedback from the hydraulic part to the atmosphere, being the flood response mostly a very fast process and occurring largely in impermeable channels and drains.
LL 875. As the authors know, total evaporation and latent heat flux is the same thing, just different units, the sentence is awkward.
ReferencesMeili, N., Zheng, X., Takane, Y., Nakajima, K., Yamaguchi, K., Chi, D., Zhu, Y., Wang, J., Qiu, Y., Paschalis, A. and Manoli, G., 2025. Modeling the effect of trees on energy demand for indoor cooling and dehumidification across cities and climates. Journal of Advances in Modeling Earth Systems, 17(3), p.e2024MS004590.
Citation: https://doi.org/10.5194/egusphere-2024-3988-RC2 -
RC3: 'Comment on egusphere-2024-3988', Anonymous Referee #3, 29 Mar 2025
reply
This manuscript is a timely and valuable review of urban land surface and hydrology models, bringing together diverse modelling approaches. The authors successfully highlight both achievements and persistent gaps in the field. The paper’s organisation is logical and the narrative engaging, providing a thoughtful synthesis across historically fragmented modelling communities - I thoroughly enjoyed reading this review paper.
I believe, however, the paper would benefit from addressing two key aspects that would further enhance its clarity, structure, and practical utility.
Major Comment 1: The paper would benefit from a conceptual schematic and stronger directory-style guidance, ideally integrated into Section 2.
While the manuscript does mention its rationale for selecting representative and newly developed models and those at the climatology-hydrology interface, it currently lacks an overarching conceptual framework that visually summarises the overall structure of the review. Including a schematic diagram early in Section 2, clearly illustrating how different urban processes (such as radiation, evapotranspiration, runoff, and soil moisture) relate to the various model classes reviewed (bulk, SL-UCM, ML-UCM, hydrology-focused models) and coupling strategies, would provide valuable directory-like guidance. Such a visual roadmap would help readers, especially those less familiar with the field, to better navigate and contextualise the rich and diverse content of subsequent sections.
Major Comment 2: The future directions section should be more closely tied to specific scientific tasks and supported by clear technical mechanisms.
Currently, the call for integration and collaboration in Section 6 is valuable but somewhat general. To strengthen its practical impact, the authors could explicitly frame future development around specific urban hydroclimate challenges (e.g. urban flood forecasting, heat mitigation, compound event analysis), clarifying the necessary coupling strategies and modelling approaches.Two concrete, actionable suggestions could help operationalise this vision:
- Develop a common modelling protocol—similar to the NetCDF CF conventions widely used in climate sciences—which is currently missing for urban hydroclimate modelling. Such a protocol could standardise inputs, outputs, metadata, and resolution criteria to greatly facilitate interoperability and cross-model comparisons. Initial steps toward this goal have already been taken within the Urban-PLUMBER initiative, but these efforts should be expanded and formalised into more broadly accepted community practices.
- Establish a practical technical framework to overcome barriers in collaborative model development. Leveraging open-source collaboration platforms like GitHub could foster transparency, modular development, and community engagement. Additionally, considering WRF’s wide adoption and flexibility, it might serve effectively as a common base framework to host and test integrated urban hydroclimate components.
These steps would provide concrete pathways to improve the current fragmented landscape, enabling more coherent and coordinated model advancements.
Citation: https://doi.org/10.5194/egusphere-2024-3988-RC3 -
RC4: 'Comment on egusphere-2024-3988', Anonymous Referee #3, 30 Mar 2025
reply
L213: The authors state that “Wang et al. (2016) added an irrigation scheme to the model” (referring to SUEWS). This attribution is incorrect. The irrigation functionality was first introduced in Järvi et al. (2011), which documented the initial release of the Surface Urban Energy and Water Balance Scheme (SUEWS). The cited Wang et al. (2016) paper does not pertain to SUEWS development and focuses instead on vegetation cooling in desert cities. I suggest correcting this reference to reflect the accurate model development history.
Citation: https://doi.org/10.5194/egusphere-2024-3988-RC4
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
283 | 60 | 7 | 350 | 5 | 6 |
- HTML: 283
- PDF: 60
- XML: 7
- Total: 350
- BibTeX: 5
- EndNote: 6
Viewed (geographical distribution)
Country | # | Views | % |
---|---|---|---|
United States of America | 1 | 107 | 33 |
Netherlands | 2 | 55 | 16 |
Germany | 3 | 41 | 12 |
China | 4 | 18 | 5 |
India | 5 | 12 | 3 |
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1
- 107