Studies of impacts of urban land-use land-cover changes on the local hydrometeorology have been largely focused on the surface-layer dynamics. In this study, we aim to address the fundamental question: Can the influence of modified land surface processes in urban areas effectively “penetrate” into the overlying mixed layer via land-atmosphere interactions? Towards this objective, we couple a single column model (SCM) to a cutting-edge single-layer urban canopy model (UCM) with enhanced representation of urban hydrological processes. The land-surface transport of energy and moisture parametrized by the UCM provides realistic lower boundary conditions to the overlying atmosphere. The coupled UCM-SCM model is tested against field measurements of sensible and latent heat fluxes in the surface layer, as well as vertical profiles of temperature and humidity in the mixed layer under convective conditions. The model is then used to simulate urban land-atmosphere interactions by changing urban morphology, surface albedo, vegetation fraction and aerodynamic roughness. Results show that changes of landscape characteristics have a significant impact on the growth of the boundary layer as well as on the distributions of temperature and humidity in the mixed layer. In addition, we adopted an advanced stochastic procedure for analysing the model sensitivity, with uncertainties in input parameters characterized using prescribed probability distribution functions. In particular, results of sensitivity analysis reveal that the urban morphology, represented as the aspect-ratio between building height and street width in the UCM, exhibits a complex and non-linear effect on the evolution of the boundary-layer height. Overall, the proposed numerical framework provides a useful stand-alone modelling tool, with which the impact of urban land-surface conditions on the local hydrometeorology can be assessed via land-atmosphere interactions.

Urban population grew fast these last decades and is expected to still increase during the future decades. Urban areas, due to artificial materials, impacts both urban climate and hydrology: urban heat island, more frequent floods, longer droughts… Furthemore, these modifications related to urbanization can go together with those due to climate change. City planning evaluation in both hydrological and climate terms becomes crucial. Numerical models are useful tools for such evaluations. Nevertheless, even if numerous models dedicated to urban areas are used, very few models are able to simulate both detailed energy and water budgets. Most of them are specialized in one topic and simulate roughly the other one.

Introduction of vegetation in cities is supposed to be one of the solutions to limit urbanization impacts as urban heat island and/or floods, by increasing water infiltration and favoring evapotranspiration. Thus, such green infrastructures impact both energy and water budgets. Our objective is then to develop a model dedicated to urban areas and able to simulate both energy and water realistic budgets.

This model is based on the Town Energy Budget scheme that has known different evolutions these last years: introduction of vegetation inside the canyon, simulation of the greenroofs. Its water budget has been improved by introducing soil and groundwater under buildings and roads (thanks to the SVAT ISBA) and the sewer network (combined or stormwater) that transfers the surface runoff and the drained groundwater to the outlet. The groundwater balance under the different surface types (building, road and garden) is performed and may impact the evapotranspiration flux from the garden surfaces.

This paper will present these last model evolutions related to the water budget as well as their evaluation. The hydro-climate evaluation of greening scenarios of a large part of the City of Nantes (France) using different green infrastructures (as greenroofs, trees, varying vegetation rates) will be discussed.

When undertaking city-scale studies of weather and climate, the variability and complexity of urban surfaces presents a modelling challenge. In heavily urbanised sites, heat storage can consume around half of the net all-wave radiation. As heat takes time to diffuse through materials, the variation of heat storage is out of phase with net radiation. The delayed response, or thermal inertia, is often regarded as the key process in the genesis of the Urban Heat Island (UHI), and is clearly an important factor in accurately modelling urban climates through the diurnal cycle. Urban schemes often represent conduction using a discretised one-dimensional parameterisation where temperature nodes are located in the centre of layers of composite surfaces (the number of layers in urban models can range from one to over ten).

Here we will explore methods to adequately and efficiently represent heat conduction and storage in multi-layer urban surfaces. The number of temperature nodes required to adequately represent heat storage and conduction are examined. In addition, an alternative conduction paramaterisation that calculates temperature at layer boundaries (rather than layer centres) is introduced. Both methods were tested against half-hourly observations collected over 18 months for a medium density site in Melbourne, Australia using an urban canyon energy balance model modified from Masson (2000) Town Energy Budget approach (aTEB; Thatcher and Hurley, 2012).

For this site, the skin temperature conduction model was more successful in reproducing the observed cycle of heat storage and release, compared with the half-layer temperature conduction approach used in many urban schemes.

Realistically simulating the urban climate is challenging. Increasing scale, generating realistic urban structure, or manipulating green infrastructure in urban domains all require efficient computational methods to process the complex interactions between the urban structures and the environment. Integrating different computational models together into a single simulation can often impart sophisticated data communication and software development interfaces. Moreover, the results of the simulations must be communicated to a diverse set of users and stakeholders using a variety of visualization techniques that afford exploration of and interaction with the data. Our efforts to create this software infrastructure have resulted in QUIC EnvSim (QES), a modular framework for simulating and visualizing radiative turbulent transport and radiative land-surface interactions in urban domains.

QES is implemented in a software development framework for creating coupled, high performance, urban microclimate simulations. It leverages efficiency by managing simulation data and computations on heterogeneous computing hardware to take advantage of both multi-core, graphics hardware (GPUs) and traditional CPUs. QES supports several models important for urban microclimate simulations including sky-view factor, radiative transport computations for direct diffuse and specular shortwave energy, and longwave emission and transport. Turbulent transport and land-surface model computations are also coupled into the QES framework to model the transfer of energy and moisture scalars between surfaces and through the air. Computational efficiencies are gained in QES by utilizing SPMD (Single Program Multiple Data) GPU programming with NVIDIA's OptiX and CUDA GPU programming interfaces.

The presentation will focus on the use and integration challenges addressed by QES in a multi-disciplinary software development setting in which computer scientists, engineers, and climate scientists contribute to coding the model. One of the key contributions provided by the QES infrastructure is a focus on managing the computational resources on the GPU that allow model data to be shared across models. The presentation will provide a technical description and demonstrations for how the QES software development interfaces afford coupling between models in a variety of contexts to simulate and visualize different urban domains.