Wildings research applied to Molesworth

With financial support from Forest Grower Research, Scion has developed machine learning algorithms to detect wilding conifers infestations by using remote sensing platforms like drones (for smaller scale assessments of infestations) and manned airplanes (for large regional surveys). The captured high-resolution imagery from these platforms are processed with Scion’s detection algorithms to map and quantify wilding infestations. The research has shown that wildings can be successfully detected before they start coning, which in turn will support better planning and management of wilding conifer infestations.

Using the outputs of this research, the National Wilding Conifer Programme has now embarked on a trial to test aerial remote sensing detection on a large scale, by surveying the vast area of Molesworth Station located in the upper part of the South Island. This area has been a hotspot of wilding incursion, with growing infestations threatening the unique ecosystems found in the area. So far, aerial imagery has been captured for more than 30,000 hectares to detect wilding infestations and ground truthing is underway to test the quality of the maps generated by the various algorithms developed over the last years.

Once the mapping is complete and the quality determined, the maps and various metrics (e.g. estimated area of occupancy and infestation densities) will be used to plan control operations. It will also provide possible critical insights how infestations develop and spread across such large areas by analysing the distribution of wildings at such large scales. The work will highlight the ability to operationalise promising technologies and research to a large spatial scale and the results will interest forest owners that are involved in managing wilding conifers.