An Study of Severe Storms and Wind Gusts on High-Power Transmission Towers during a Southern Brazil Campaign

The west of Parana State in Brazil have been affected by outages due the severe local storms. The outage index in this area is about 30% higher than adjacents areas because the high incidence of severe storms. Furthermore, high-power transmission towers have been systematically blown down by wind gust during strong storms. To understanding the mechanisms that development extreme and explosive convection that produces high values of rainfall and winds, the Meteorological System of Parana (Simepar) established research projects with the Power Company of Parana (Copel), the Center for Natural Disaster monitoring and Alert of Brazil (Cemaden) and the University of Pelotas (UFPel). The projects include the installation of sonic anemometers in four towers, the analysis of the storms and QPE using data from the Simepar Dual-Pol S-Band Weather Radar and high-resolution simulations with WRF and WRFDA radar assimilation. In three towers, the sonic anemometers were installed in three levels, 10m, 20m and 30m. In another tower it was installed an additional sensor at 44m. The towers were located within a 60km radius distance of the dual-polarization S-band weather radar operated by the Meteorological System of Parana State (Simepar). High-resolution simulations (3 km) using WRF/NCAR MARS (Model Rapid Assimilation of Simepar) and 3DVAR data assimilation of radar reflectivity and radial velocity were performed for strong and moderate convection events.  The assimilation module of MARS runs under a updated cycle with radar data input every 15 minutes. Lateral boundary conditions were updated every three hours  using the GFS model. Fifteen meteorological stations from Simepar basic network was used to obtain 10m wind gusts and precipitation as well as  other variables over about. On the area of interest there are 15 station in a 25 000 km2 area. The QPE was obtained by a Poisson's Equation methodology which remove the bias of the radar estimations and integrated it with rain gauges. The advantage of this technique is keeping the the rain gauge value in the pixels which are close to the meteorological stations.

The investigation shows, until the moment,  that gust winds above 20 ms-1 could be generated by Squall Lines and somes cells embedded in it. Supercells and tornadoes were found in radar data and in photoes, respectively, but its winds were not recorded in the anemometers installed in towers.  Since february 2016, the maximum wind registered was 27.6 ms-1 at 28m. It was generated by a convective-cell embedded in a Squall Line. By the radar data it could be found that the registered winds were located in the northeast flank of the storm indicating it could be generated more intense gusts below the high-reflectivity areas. In radial velocity it was found strong divergence signatures and some anticyclonic rotation at low levels. The simulations of this event is detailed in another article in this conference by Inoue et al.

A tentative of associating reflectivity intensity and radial velocity with wind gusts registered in towers has been made in the project. Despite the statistics was poor yet, it was possible to find some qualitative evidences between weak reflectivity signals and moderate gust measurements (10 to 15 ms-1). High wind peaks (20 – 27 ms-1) were associated with strong long-lived squall lines and with severe convection in cold fronts. During the development and propagation of an extratropical cyclone at South Atlantic Ocean in 25-26 April 2016, 15-20 ms-1 wind gusts were registered in all anemometers in all levels, opening an evidence that this phenomenon should be considered in the studies, specially if the propagation comes inland.

Another interesting result found in the campaign was that during events with strong gusts (greater than 15 ms-1) the vertical structure of the wind becomes linear, probably due to the propagation of the gust fronts. Therefore, while the average wind profile has an exponential shape, during the storms the profile shift to linear with similar values from the 10m up to 44m level.

The results of QPE during the severe storms propagation show a high-dependence of the rain gauge information because the reflectivity-only could overestimate to an unreal values. At this moment, a dual-pol R-Z relationship have been tested to improve the rainfall estimations in severe storms events.

So, more results of the measurements are expected, mainly in austral spring season when the most explosive and intense phenomenon occur, to improve the observed sample and data to be used in studies and simulations.