| Popis: |
During the week of 25–31 March 2019, the West Coast of the South Island experienced an intense rainstorm. The rain was widespread, extending from Haast to Hokitika, with the highest intensity of rainfall recorded in the range front south of Hokitika. The March 2019 storm resulted in a record amount of rainfall recorded over a 48-hour period in the Cropp River Catchment. The result of this extreme rainfall event was widespread flooding and landslides, including the washout and destruction of the Waiho Bridge on State Highway 6. An aerial reconnaissance was flown by GNS Science on 17 and 18 April 2019, as part of GeoNet landslide response, to identify landslides triggered by the storm. High-resolution SkySat satellite imagery was acquired for a c. 10,000 km2 area covering the area of most intense total rainfall, which extended from inland of Hokitika south to Harihari. This satellite imagery, combined with pre storm event aerial imagery obtained from Land Information New Zealand and Google Earth, was used to map the distribution of landslides that were triggered by the storm. The storm event triggered 1290 landslides within the study area, with the highest spatial densities observed in the hill country surrounding Mount O’Connor, southeast of Hokitika. The landslide distribution was compared against key physiographic attributes such as rainfall, geology, slope angle, slope aspect and vegetation type to assess controlling influence(s) on landslide failure. The results of our analysis show that the areas of more intense rainfall do not correlate with a greater density of landslides. This disparity could be the result of the interpolated rainfall data not being representative of the true rainfall during the event and may indicate that rainfall was higher than calculated by the linear interpolation of the rain gauge data on slopes nearer to the coast (where there is a higher density of landslides) than those further inland. Additionally, this disparity may be due to variations in rainfall intensity during this 48-hour period, which are not captured in the 48-hour rainfall totals from rain gauge data. Additionally, the limitations of the satellite imagery, which include snow cover and shadow, may have prevented the identification of landslides at higher altitudes. While the density distribution of landslides in the study area does not correlate with rainfall intensity, it does correlate with slope angle and aspect. Landslide density increases with steeper slopes, and north-, northwest- and northeast-facing slopes all display a higher landslide density, which may be related to the characteristics of the rainfall in the storm event. For the major land-cover types, sub-alpine shrub displays the highest landslide density, with 2.96 landslides per km2, and accounts for 16.5% of the study area. The dominant land-cover type of indigenous forest, accounting for 51.3% of the study area, only displays a landslide density of 1.64 km2(auth) |
