Wolter, A.; Rosser, B.J.; Lin, S.-L.; Boyes, A.F.; Townsend, D.B., Jones, K.E., Choi, E. 2022 Phase I: Reconnaissance report on landslides caused by the 16–18 July 2021 rainstorm in the Marlborough region. Lower Hutt, N.Z.: GNS Science. GNS Science report 2022/08. 68 p.; doi: 10.21420/C7DK-BQ35
Abstract
An extreme winter storm travelled across New Zealand on the 16th to 18th of July 2021, affecting the Marlborough, Tasman, Nelson, West Coast and Wellington regions. Flooding and landslides resulted in evacuation of numerous communities and damage to transport networks and property in these regions. Multiple roads were closed and houses impacted by landslides, particularly in the Buller River Valley and Marlborough. Following the storm, GNS Science and Marlborough District Council staff completed reconnaissance in the area to determine the extent of damage from landslides to infrastructure and communities, including a five-hour helicopter flight and three days of detailed site visits and surveys. The site visits involved collecting UAV (drone) imagery to create 3D photogrammetric models, characterising 11 landslides, and assessing the damage and impact to 18 residential buildings, including the state or condition of damage sustained from landslides. In addition, the safety colour tagging and habitability of buildings were recorded. Analysis of rainfall data indicates that 400–500 mm of rain fell over 54 hours (annual return intervals [ARIs] > 250 years) for most of the Wairau north bank hill country and Marlborough Sounds regions, with localised rainfall intensities exceeding 25 mm/hr in several large areas (ARIs 2–10 years). Antecedent soil moisture conditions were relatively high following frequent rainfall over the previous months. Landslides occurred in all vegetation cover types in areas of high rainfall, including indigenous forest, regenerating indigenous forest and scrub, exotic forestry and pasture. Recently harvested or replanted exotic forest plantations were most susceptible to landsliding, but slopes with exotic forest and pasture, as well as with regenerating indigenous forest and scrub, were also commonly affected. Debris avalanches and debris flows were the most common landslide types and frequently initiated at the contact between bedrock and overlying colluvium. Many of the observed landslides entrained significant large woody debris, which caused substantial damage to buildings. Of the 18 residential buildings surveyed, 10 were deemed uninhabitable. Fortunately, no deaths or injuries were reported, as most of the affected houses were vacant during the event. Data collected will be included in the development of a rainfall-induced landslide prediction tool. In addition, the information presented here will be used in combination with numerical landslide runout model hazard intensity outputs (e.g. debris flow height) to develop landslide vulnerability functions for New Zealand buildings. The information may also be used to better understand landslide risks in the study area and aid planning for and mitigating against natural hazards. (auth)
