WP4 – Climate Module

The climate module (WP4) work package consists of three main objectives: (a) research and development, (b) integration and (c) implementation of the climate model in the two case studies.

 

Research and development

The climate module can be understood as the anchor point for calibrating and re-evaluating system outputs from all simulation modules to detect climate mitigation strategies or to perform urban system adaptations within the interactive evaluation processes. Therefore a main objective is to obtain a physically correct view on urban climate energy flows within an urban fabric. Multi-scale climate modelling approaches enable the systemic evaluation of urban microclimate in respect to the urban heat island effect. The computed information improves the low granularity of existing climate data sensing usually found in cites for an enhanced understanding of the current climatic behaviour of a city. Further the outputs from this module shall improve all connected system modules with an initial description of the urban environment through the shared database while the model receives feedback from the modules transportation/land use, sensing, building stock and urban development strategies.

 

Integration

Computation costs are serious concerns for climatic evaluations on urban scale and usually require large-scale high performance computing infrastructures. Additionally, for this project the provision of interactive feedback is crucial for the exchange between climate module, other modules and for evaluation purposes. Therefor the climate module will be tightly coupled with the GIS/data-warehouse to provide a simplified data exchange between modules. Data from the landuse/transport module, the building stock module, the sensing module and strategies from the urban development module will have direct impact on the climate module and vice versa. Further the Climate Module will mainly work with the multiscale resolutions of the procedural 3D city model as inputs for the computation. A special challenge will be to publish the climate module to cloud computing providers. Further an interface to web-based services needs to be developed to make the computed results interactively explorable. Therefor a direct connection to the visualization frontend will be developed.

 

Implementation

Computation costs are serious concerns for climatic evaluations on urban scale and usually require large-scale high performance computing infrastructures. Additionally, for this project the provision of interactive feedback is crucial for the exchange between climate module, other modules and for evaluation purposes.

Therefor key investigations will concentrate on minimizing computation costs for urban heat island assessment while the simulation process has to be accelerated.

  1. Recently multi-scale urban climate evaluations have been performed with computational-fluid-dynamics (CFD) simulations with promising results. We will investigate to tailor existing non-proprietary CFD environments, such as OpenFOAM, to optimize computation costs with graphics processing unit (GPU) based parallel computing infrastructures for the calculation of the urban climate. These infrastructures seem to offer accelerated calculation time while increasing packet size and decreasing calculation costs.
  2. Simplified CFD algorithms, which deal more efficiently with calculation time and provide better interactivity, are recently described in literature. Hence we will investigate the implementation within our HPC infrastructure.
  3. Another trend in high performance computing is to compress calculation time with collections of pre-processed model samples. We will therefore investigate the frequency-based pre-calculation of urban typology samples (building mass, street canyon, and topological configurations) grouped into the scales building, street canyon, quarter, and city. This multi scale information shall provide the tile boundaries and physical boundary variables for the physically correct CFD calculations and structure the overall calculation process into optimized simulation tasks.  The procedural model provided by the GIS/data-warehouse seems to be an ideal candidate for the data compression and spatial organization of the tiling and sampling processes.

 

Finally, a very important objective is to investigate if the samples could directly provide statistical climate data to the connected system modules and the decision support tools. This process seems to be very promising in terms of diminishing calculation efforts. It will be further investigated if the calculation samples can be designed as a generic urban climate grammar, which can be then possibly transferred and applied to a multitude of urban regions for statistical evaluation and as initialization data for otherwise time-consuming CFD model set ups.