NHRP / Hazard themes / Weather, Flood and Coastal Hazards / Research Highlights 2013-14

Research Highlights 2013-14

Damaged irrigator in the aftermath of the Canterbury wind storm 2013. Photo: Dave Allen, NIWA.

Increasing Resilience to Weather Hazards

The costs of weather related hazards is increasing. During 2013 insured losses for weather related events were more than $174 million, the second most expensive year since records began in 1968. Of this amount, $74.5 million relate to the event of 10 – 11 September, when Canterbury was subjected to the strongest wind storm since 1 August 1975. This storm caused downed power lines and trees, damage to houses, businesses and farms, and overturned trucks, boats, and caravans. About 28,000 houses and businesses were without power overnight. Over 800 irrigators were damaged and there were significant business interruption costs. Recorded peak wind gusts ranged from 100 to 140 km/h. This event exposed the susceptibility of much of our built rural infrastructure to weather related hazards – which may be expected to occur more frequently in the future.

The most cost effective approach to improving resilience to weather hazards is through accurate forecasting that informs decisions which minimise the impact of such events. In the case of centre pivot irrigators, the loss of which led to significant insurance claims and impacts on farming operations in Canterbury, simple mitigation measures could have significantly limited the damage. This presupposes that accurate and specific forecasts of such events can be made, and that users have sufficient confidence in them that they will take action in advance of the predicted events. Here, we outline some of the research that will make our infrastructure more resilient to such weather related hazards in the future.

To solve the scientific problems associated with forecasting high impact weather events in the presence of New Zealand’s complex terrain, the following are required:

  • A computer model of the weather that can simulate the atmospheric processes acting at the scale of New Zealand’s landscape and location in the mid-latitudes;
  • Real time streams of satellite, aircraft, radar, balloon, ship and land surface observations that can be incorporated in the weather model to ensure that forecasts are accurate with respect to event intensity and timing;
  • A supercomputer that can quickly carry out the vast number of computations implied by the points above, ensuring end users have the longest possible time to prepare for predicted hazardous events.

During 2013 NIWA scientists developed and began testing a new ultra-high resolution numerical weather prediction model, the New Zealand Convective Scale Model (NZCSM). Every 6 hours, this model forecasts the weather on a 1.5 km horizontal grid covering the entire NZ land mass and adjacent ocean areas, from the surface to 40 km altitude, out to 36 hours ahead.

The NZCSM, a local configuration of the UK Met Office Unified Model, was developed at NIWA with international collaborators. It is the largest kilometre-scale weather forecast model internationally and can reveal atmospheric flow features caused by the interaction of weather systems with NZ’s complex terrain – features not previously seen.

NIWA fig 1 NZCSM data v2

Spatial composite of the maximum NZCSM 10 m (above ground level) 30 min wind gust simulations for 10 Sept 2013. Dashed line indicates location of eastern half of the vertical cross section plot showin in Fig 2 below. Data: NIWA.

To understand how outputs from the model could better inform decision making in advance of predicted high impact events, NZCSM simulations are being used to investigate the Canterbury windstorm event. Figure 1 (left) shows a spatial composite of the maximum 10 m wind gust simulations for the period 6 am to midnight on 10 September, together with observed maximum gust speeds for the same period. The NZCSM simulations indicate that some regions on the plains are much more susceptible to higher gust speeds (e.g. downstream from major river gorges) than others, perhaps explaining why some areas suffered major damage while others escaped. While the observation network resolution is too sparse to resolve the high wind regions, insurance claim data and the locations of downed power lines can be used to improve our understanding of where damaging winds occurred, and hence verify the accuracy of NZCSM forecasts.

The simulations indicate this event was a downslope wind storm, where energy and momentum in the flow above the mountains is forced down toward the surface. Figure 2 (below) shows the vertical winds (and potential temperature) of the lowest 5 km of the atmosphere along a cross section over the Alps at 8 pm on 10 September (Fig. 1). It indicates very strong mountain waves over the Southern Alps, with alternating zones of strong upward winds (red-orange) and downward winds (blue). Note the very strong downslope and surface winds (up to approximately 40 m/s, or 144 km/h) east of the mountains. The steep black potential temperature lines highlighted in the black ellipse indicate a large amplitude wave response and the strong possibility of wave-breaking and rotors with severe turbulence in the regions within the red ellipse.

NIWA fig 2 NZCSM data

NZCSM simulation of winds (arrows) in the plane of the cross section indicated in Fig 1. Data: NIWA

In the winter of 2014, the Deepwave experiment (Link ♦) will provide an opportunity to observe mountain waves and determine how well NZCSM is able to model them, and in turn, how well it can forecast the evolution of severe weather over the New Zealand landscape.

Contact: Michael Uddstrom, NIWA

Back to

Last updated 18 Nov 2014