Environmental Engineering Reference
In-Depth Information
Figure 1. Schematic of the MSEPS ensemble weather & wind power prediction system.
5.3.
The MSEPS Forecasting System
The Multi-Scheme Ensemble Prediction System (MSEPS) that has been used for our
studies and which will be described later in this chapter, is a limited area ensemble pre-
diction system using 75 different NWP formulations of various physical processes. These
individual “schemes” each mainly differ in their formulation of the fast meteorological pro-
cesses: dynamical advection, vertical mixing and condensation. The focus is on varying the
formulations of those processes in the NWP model that are most relevant for the simulation
of fronts and the friction between the atmosphere and earth's surface, and hence critical to
short-range numerical weather prediction. Meng and Zhang [31] found that a combination
of different parameterisation schemes has the potential to provide better background error
covariance estimation and smaller ensemble bias. Using an EPS for wind power prediction
is fundamentally different from using one consisting of a few deterministic weather pre-
diction systems, because severe weather and critical wind power events are two different
patterns. The severity level increases with the wind speed in weather, while wind power
has two different ranges of winds that cause strong ramping, one in the middle range and a
narrow one just around the storm level (the cutoff level). Wind power forecasting models
therefore have to be adopted to the use of the ensemble data. In general, a wind power
prediction model or module, that is directly implemented into the MSEPS is different from
traditional power prediction tools, because the ensemble approach is designed to provide
an objective uncertainty of the power forecasts due to the weather uncertainty and requires
adaptation to make use of the additional information provided by the ensemble. Figure 1
shows the principle of the MSEPS wind power forecasting system. Apart from the direct
Search WWH ::
Custom Search