A 1-D canopy interface model (CIM) has been developed recently and integrated in the meso-scale meteorological WRF v3.5 model in order to improve the surface representation. One of the objectives of such a model is to prepare for the coupling of micro-scale models with meso-scale models so as to improve building energy comsumption estimates at the urban scale as well as improve meteorlogical variables calculation in urban canyons.

The objective of the present study is to evaluate the value of the use of a module able to produce highly resolved vertical profiles of these variables. The coupling methodology is detailed here and its evaluation is done using a reference run based on a fine resolution WRF simulation. In order to keep both the CIM and the meso-scale model in coherence, a new methodology is developed and an additional term is added to the calculation of the CIM.

Two major conclusions can be drawn from this work: the coupling improves the simulations of the meso-scale model and the WRF-CIM system provides highly resolved vertical profiles while at the same time improving significantly computational time. The data from these preliminary results are very promising as it provides the foundations for the CIM to act as an interface between meso-scale and micro-scale models.

Additionally, we will discuss the strategy that will be used to couple the WRF-CIM system with the CitySim software. It is expected that on the one hand the feedback from the CitySim software would improve the meteorological variables calculations while on the other hand the WRF-CIM system can provide enhanced meteorological profiles to CitySim. This coupled sytem, could be used by urban planners or architects, as it would provide a significant advantage in the evaluation of building energy consumption and urban planning scenarios.

Keywords: building energy-use, canopy model, meso-scale models, micro-climate, multi-scale modelling, urban climate

The ability to accurately forecast local weather conditions in cities is important for a wide variety of applications. For example, alerting local authorities to low temperatures so that roads can be gritted, or issuing health-related warnings to vulnerable residents. Until recently, a lack of urban meteorological observations above roof height has meant that verification of numerical weather prediction models in urban areas can be difficult. This presentation reports results from a comparison of the UK Met Office forecast model with observations for a) a one-year comparison of the operational model at a resolution of 1.5 km, and b) preliminary results for runs with higher resolution (100m) for strong urban heat island cases.

Observations of the boundary layer of central London were carried out at the BT Tower (height 191 m) and at a rooftop (height 19 m a.g.l) from summer 2010 to spring 2013. A Doppler lidar was used over the same period to obtain profiles of wind and turbulence, and to determine mixing height, although the most significant overlap with model data was only for November 2011 to February 2012. Observations of surface meteorology and mixing height using Doppler lidar were also available at a rural site (Chilbolton, 125 km south-west of London). These observations have been used to evaluate the behaviour of the Met Office 1.5 km forecast model (UKV) over one year within the observational period. A focus was put upon days when a strong urban heat island was occurring, to test the model performance when the urban surface energy balance was having the maximum effect. For these days, both sensible and latent heat fluxes were underestimated at the BT Tower, and the timing of the diurnal cycle was delayed by two hours. The underestimate was also seen at the Chilbolton rural site, but not the delay. Nocturnal urban temperature was well modelled, whereas daytime urban temperatures were underestimated on average by 1-2C. Daytime rural temperatures performed similarly, but night-time minima were too warm. For the shorter winter period available, the model did not capture the observed increase in urban boundary layer depth compared to the rural site. For the higher resolution simulations, the effect of resolution on resolved vs. sub-grid scale fluxes is demonstrated, and turbulent characteristics of the simulations are compared with Doppler lidar observations. It is tested whether increased resolution produces improved model performance in terms of fluxes and boundary layer depth.

Mesoscale weather and climate models are useful tools for providing important insights and improved understanding regarding climate change mitigation and adaptation in urban environments for institutional stakeholder and urban planner communities. For urban applications, the complexities of the urban fabric within each urban area require scale-dependent descriptions of the land use and activity pattern in order to account for the effects of subgrid scale urban surface cover and buildings. This enables improved predictions of the wind, turbulence, and concentration fields. A commonly used method is the inclusion of a set of urban canopy parameters (UCPs) into mesoscale meteorological models to parameterize building-induced drag and turbulence production and the building-modified surface energy balance. Current UCPs used in mesoscale models attempt to capture major struc¬tural and material features considered to control the momentum and ther¬modynamics of the flow. It is anticipated that through the WUDAPT (Ching et al., 2014), being initiated by IAUC, scale-dependent UCPs will soon become available for model applications.

In this presentation, we illustrate and explore the sensitivity of model outputs to scale-dependent UCP inputs to better understand and articulate their implication to model user communities. The MRA (Mouzourides et al, 2013; 2014) provides a powerful means to perform a weather and climate model scale-dependent sensitivity study for urban applications of models. The MRA is a method that can take into consideration the inherent information residing in urban landscapes, and convey this information to multi-scale modelling studies (without discarding redundant details) in a manageable coherent and structured way (Ching, 2012; Mouzourides et al., 2012). For example in a nested grid simulation over an urban area, the partitioning of the urban heterogeneity between the part resolved by the grid of the model and the subgrid part treated by the UCP must be derived in a consistent manner across all scales in order to obtain meaningful UCP values at all scales.

Given that the majority of the world’s populations now reside in urban areas, we focus on demonstrating scale sensitivity in the context of energy usage and urban activities on climate, heat island intensity and other like issues. For this, scale-dependent energy-related attributes will be derived for the city of London using building energy data. It can be further shown how using this gridded data and the BEP-BEM urban option in WRF we can obtain predictions of anthropogenic heating from buildings, spatially resolved surface energy budget distributions, heat Island intensity and flows for London.

A common simplification of the conservation equations of momentum and heat in the Planetary Boundary Layer (PBL) used in mesoscale models consists in neglecting the horizontal component in the divergence of the turbulent fluxes. This is usually justified in two ways: 1) because the sink of momentum, and source/sinks of heat, located at the surface, induce stronger vertical than horizontal gradients of the atmospheric variables; 2) because the strongest turbulent motions are buoyancy driven (and so directed in the vertical). In addition, the horizontal resolution used in mesoscale atmospheric modeling until the last decade of the 20th century (when most of the PBL schemes were developed) was of several kilometers, something that, by itself, prevented the resolution of sharp horizontal gradients. These are the main reasons why standard PBL schemes are 1D in the vertical. These 1D PBL schemes were frequently tested/tuned/validated against LES with horizontally homogeneous surface fluxes. The main scientific question motivating the current work is: to what extent neglecting the horizontal divergence of turbulent fluxes is justified for PBLs over regions with strongly horizontally-heterogeneous surface fluxes, like cities? The answer to this question is particularly relevant if we consider that, thanks to the increase of computational power, we are now able to perform simulations with horizontal spatial resolutions of the order of several hundreds of meters, which creates the possibility of resolving sharp horizontal gradients. We use the NCAR Large Eddy Simulation (LES) model (Sullivan and Patton, JAS, 2011) to simulate turbulent flow over a 2.5 km wide hot and rough strip (surface heat flux of 360 W/m2 and roughness length of 1m), representing a city, surrounded by a colder and smoother (120 W/m2 and roughness length of 0.1m) strip of the same width, representing a rural area. The horizontal resolution is 20m, vertical is 8m, and the PBL is capped at 1000m. Mean fields are obtained by averaging over lines parallel to the direction of the stripes, and the turbulent fluxes are deduced from the perturbations from these means. Results are then averaged over time. Analysis of these results reveals that the horizontal divergence of the horizontal fluxes is found to be of the same order as the vertical divergence of the vertical fluxes, particularly close to the boundary of the city; therefore in the vicinity of city-induced horizontal heterogeneity, the horizontal divergence of horizontal fluxes cannot be neglected. A budget analysis of the horizontal and vertical turbulent fluxes is also performed to guide parameterization of such fluxes.

The land surfaces take an important role to provide dynamical and thermal energy to the atmosphere above. In order to forecast the appropriate amount of momentum, heat, and vapor fluxes, the MRI’s NHRCM (Non-Hydrostatic Regional Climate Model) selected a sophisticated vegetation scheme of the SiB (Simple Biosphere) as its land surface scheme. However, non-vegetation but urbanized grids became obvious as the resolution of the model became higher up to several km along with the rapid progress of dynamical downscaling technique and computational technology. Reproducibility of the climatology on urban area seems to be insufficient even if the model devises treatment of the SiB as dried bare ground to express the so-called urban deserts.

In this study, an urban canopy parameterization scheme, called SPUC (Square Prism Urban Canopy), will be applied to the 4km-resolution NHRCM in order to improve the representation of radiation and heat budgets of urban surfaces. Using the SPUC- and SiB- coupled NHRCM, present climate (from 2001 to 2006) experiments were executed and the impact of the SPUC scheme to the climatic reproducibility was evaluated.

The targeted area was Kanto-Koshin region including Tokyo metropolitan area, which is one of the most urbanized cities in Japan. The JMA’s regional analysis (RANAL) dataset was used as initial and boundary conditions of the simulation. The RANAL was downscaled once by NHRCM10km with SiB scheme for all land grids. The 10km resolution dataset was also downscaled by NHRCM4km. The 4km experiments were executed using SiB scheme for all land grids (NHRCM-SiB), and using both SiB for natural surface grids and SPUC for urban surface grids (NHRCM-SPUC). Time integration was continuously executed for about 5 years from August 1, 2001 to September 1, 2006.

The preliminary evaluation of the reproducibility can be concluded as follows. The five year mean surface temperature reproduced by NHRCM-SiB showed a certain level of minus biases in the Tokyo metropolitan area. The minus biases around Tokyo were changed to be plus when the SPUC scheme was applied there. Although the area averaged bias changed to be worse from 1.3 °C by NHRCM-SiB to 1.55 °C by NHRCM-SPUC, the correlation factor between the simulation and observation was improved from 0.73 (NHRCM-SiB) to 0.86 (NHRCM-SPUC) implying the better reproducibility of NHRCM-SPUC on horizontal distribution of temperature. On the other hand, few differences were seen in total amount of precipitation between the two experiments.

This study explores the sensitivity of high-resolution mesoscale urban heat island (UHI) simulations and its relation with the lake breeze, focusing on the Chicago metropolitan area (CMA) and its environs with a series of climate downscaling experiments using the Weather Research and Forecasting (WRF) model at 1-km horizontal resolution. This study has the following research objectives: (i) perform a robust analysis of different urban physical parameterizations for the Chicago region and tune urban physics for the region; (ii) study the impact of urbanization and its relation to UHI and lake breeze; (iii) study the influence of land data assimilation for initialization of regional climate model; (iv) and study the impact of sub-grid scale land cover variability based on dominant and mosaic approach. Comparison of simulations with station observations and MODIS satellite data showed that WRF model was able to replicate the measured surface temperature and wind speeds with above numerical modeling improvements. It was found that numerical models need better representation of surface characteristics and correct initialization of land surface observational data to accurately capture near surface meteorology. In addition, changes in near surface temperatures were more significant during nighttime when urban heat island was high. Inclusion of the effects of sub-grid scale variability in sub-urban areas improved the near surface temperatures in those regions. Results have shown that the above objectives have helped to capture complex interactions between UHI and lake breeze and reduced uncertainties in numerical modeling techniques.