Full implementation of non-orthogonal surfaces within the PALM model

The orthogonal 3D grid used in PALM and other atmospheric models brings challenges in representing resolved obstacles (such as terrain or buildings) with non-orthogonal surfaces. The simple binary form of representation, in which each grid cell is either entirely atmospheric or obstacle, causes aliasing effects (artificial stairs). It has been demonstrated that this leads to significant biases in radiation, surface energy balance and fluid dynamics.

The newly introduced PALM slanted surface option involves new methods in the Radiative Transfer Model (RTM) and the surface-related modules in PALM, and it enables a more faithful representation of non-orthogonal surfaces while preserving the PALM’s Arakawa-C model grid without inducing significant performance penalty. For surface-atmospheric interactions, the cut-cell approach is used. In the RTM, a new representation of surface elements is introduced for both raytracing in the preprocessing phase and radiosity-based interactions in the time-stepping phase of the simulation.

The format of the input data had to be extended to accommodate the representation of the slanted surfaces. The open-source input processor PALM-GeM is able to prepare slanted surface PALM inputs using various geographic data.

Validation and benchmarking results from synthetic and real-case scenarios are presented. The slanted surface model is also being prepared for wind tunnel validation.