Studying wind gusts in Switzerland
We have recently published a paper on the ability of various wind gust parameterizations (WGP) to produce realistic wind gusts in historic and recent wind storms in Switzerland. The full title of the paper (which by the way is open access) is Evaluation of downscaled wind speeds and parameterised gusts for recent and historical windstorms in Switzerland, and is a collaboration between KUP and the Geography department, both at the University of Bern.
Wind gusts are a very important phenomenon with a strong potential to produce disastrous situations leading to large economical losses and personal damage. They are a great concern in central Europe (where storms as Lothar or Vivial are still remembered), and must be carefully taken into account in all risk assessments to protect critical infrastructures. However, gusts are a complex phenomenon hard to comprehend and forecast. Although they form under rather well known and even predictable meteorological circumstances, they are ultimately driven by turbulent fluxes within the planetary boundary layer.
Turbulence is an uncomfortable outcome of the equations of fluid dynamics. Albeit the physical laws that control this phenomenon were established long time ago, the resulting differential equations turn out to be a convoluted problem without analytical solution. Therefore, the study of turbulence has to be carried out with the help of semi empirical and numerical methods. But even this is problematic, since turbulent eddies that lead to wind gusts evolve at temporal and spatial scales that are beyond the scope of current meteorological models. For instance, a spatial resolution of 2 km, which is in the edge of what can be nowadays implemented with state-of-the-art RCMs, is not yet sufficient for this purpose. This is so because turbulent eddies leading to wind gusts range between few to hundred of meters. Therefore, current RCMs alone are not good enough to be used in the study of wind gusts.
Thus, the current approach to deal with wind gusts consists of parameterizing the turbulent flux. Thereby, wind gusts are diagnosed offline (this is, outside the RCM simulation) as:
\[ \textrm{gust}(\vec{r},t) = \textrm{wind}_{10}(\vec{r},t) + f(u, v, P, Z, T, \nabla T, e, \dots) \]
Where \(\textrm{wind}_{10}\) is the instantaneous 10-meter wind speed customary diagnosed by all RCMs, and the function \(f\) can be defined in a variety of ways, from a simple rescaling to very elaborated semi empirical models that have into account the Turbulent Kinetic Energy. But regardless of the details of the implementation, each method ultimately requires a calibration of the parameterization against observations.
And this is precisely the main objective of this paper: evaluate two RCM simulations (that produce the first term in the right hand side of the equation above) and four WGP (the second term) to evaluate to what extent this model chain is able to reproduce, compared to observations, the wind and wind gusts speeds in recent and historical wind storms in Switzerland. The paper is actually more ambitious, since another important aspect it tackles is to evaluate to what extent driving WRF with the so-called twenty century reanalysis (20CR) is feasible, in contrast with more orthodox reanalysis such as ERA Interim.
We conclude that yes, using the 20CR dataset is sensible, as well as necessary when you want to investigate historical events not covered by the relatively short periods of other reanalysis. Regarding the WGP, we demonstrate that, although the sustained wind (this is, the first term in the right hand side of the equation above) is systematically overestimated, the outcome of the WGP is an underestimation of the wind gust. This implies that all parameterizations underestimate the turbulent flux in the complex orography of the Swiss terrain, and therefore we suggest that they should be carefully recalibrated taking into account the characteristic of each area of application.
Tags: paper publications
Categories: Science
subscribe via atom