Albert Oliver

Improving shadows detection for solar radiation numerical models

Solar radiation numerical models need the implementation of an accurate method for determining cast shadows on the terrain or on solar collectors. The aim of this work is the development of a new methodology to detect the shadows on a particular terrain. The paper addresses the detection of self and cast shadows produced by the orography as well as those caused by clouds. The paper presents important enhancements on the methodology proposed by the authors in previous works, to detect the shadows caused by the orography. The domain is the terrain surface discretised using an adaptive mesh of triangles. A triangle of terrain will be under cast shadows when, looking at the mesh from the Sun, you can find another triangle that covers all or partially the first one. For each time step, all the triangles should be checked to see if there are cast or self shadows on it. The computational cost of this procedure eventually resulted unaffordable when dealing with complex topography such as that in Canary Islands thus, a new methodology was developed. This one includes a filtering system to identify which triangles are those likely to be shadowed. If there are no self shadowed triangles, the entire mesh will be illuminated and there will not be any shadows. Only triangles that have their backs towards the Sun will be able to cast shadows on other triangles. Detection of shadows generated by clouds is achieved by a shadow algorithm using satellite images. In this paper, Landsat 8 images have been used. The code was done in python programming language. Finally, the outputs of both approaches, shadows generated by the topography and generated by clouds, can be combined in one map. The whole problem has been tested in Gran Canaria and Tenerife Island (Canary Islands – Spain), and in the Tatra Mountains (Poland and Slovakia).

Characterization of Geographical and Meteorological Parameters

This chapter is devoted to the introduction of some geographical and meteorological information involved in the numerical modeling of wind fields and solar radiation. First, a brief description of the topographical data given by a Digital Elevation Model and Land Cover databases is provided. In particular, the Information System of Land Cover of Spain (SIOSE) is considered. The study is focused on the roughness length and the displacement height parameters that appear in the logarithmic wind profile, as well as in the albedo related to solar radiation computation. An extended literature review and characterization of both parameters are reported. Next, the concept of atmospheric stability is introduced from the Monin–Obukhov similarity theory to the recent revision of Zilitinkevich of the Neutral and Stable Boundary Layers (SBL). The latter considers the effect of the free-flow static stability and baroclinicity on the turbulent transport of momentum and of the Convective Boundary Layers (CBL), more precisely, the scalars in the boundary layer, as well as the model of turbulent entrainment.

Wind Field Diagnostic Model

This chapter describes Wind3D, a mass-consistent diagnostic model with an updated vertical wind profile and atmospheric parameterization. First, a description of Wind3D is provided, along with their governing equations. Next, the finite element formulation of the model and the description of the solver of the corresponding linear system are presented. The model requires an initial wind field, interpolated from data obtained in a few points of the domain. It is constructed using a logarithmic wind profile that considers the effect of both stable boundary layer (SBL) and the convective boundary layer (CBL). One important aspect of mass-consistent models is that they are quite sensitive to the values of some of their parameters. To deal with this problem, a strategy for parameter estimation based on a memetic algorithm is presented. Finally, a numerical experiment over complex terrain is presented along with some concluding remarks.

Wind Field Probabilistic Forecasting

Probabilistic wind forecasting is a methodology to deal with uncertainties in numerical weather prediction models (NWP). In this chapter, we describe the need for ensemble forecasting, the different techniques used to generate the different initial conditions, and the operational ensemble models that are used nowadays in meteorological agencies. Then, we develop an ensemble method designed for the downscaling wind model described in Chap. 4 coupled with the AROME–HARMONIE mesoscale model, a non-hydrostatic dynamic forecast model described in Chap. 5. As we have explained in Chap. 4, some parameters need to be estimated since we do not know its exact value. These parameters are, basically, the roughness length and the zero plane displacement (explained in Chap. 2), as well as the Gauss moduli parameter (\(\alpha \)) used in the diagnostic wind model. This estimation is the main source of uncertainties in the model; therefore we will estimate some of these parameters using different forecast values of the AROME–HARMONIE. Finally, an example of the approach is applied in Gran Canaria island with a comparison of the ensemble results with experimental data from AEMET meteorological stations.