11-20, 15:45–15:50 (Pacific/Auckland), WG403
ORA Reefs aims to restore degraded rocky reef ecosystems across Tīkapa Moana by removing large-scale kina barrens. To support marine habitat identification and monitoring, we are developing a classification pipeline in Python and QGIS. Using the Depth-Invariant Index to isolate seabed reflectance from water column reflectance.
ORA Reefs is a new initiative developed by Ocean Regeneration Aotearoa (ORA) Trust to actively restore degraded rocky reef and benthic ecosystems across Tīkapa Moana/the Hauraki Gulf, New Zealand. Sea urchin barrens are a dominant stressor on rocky reef ecosystems worldwide due to the overfishing of key predators. Evidence shows that releasing barren rocky reefs from kina (Evechinus chloroticus) grazing pressure in the Hauraki Gulf through large-scale removal enables biodiversity regeneration within 2 years. ORA Reefs’ initial focus is to pilot large-scale kina barren removal alongside the development of artificial reefs on degraded benthic ecosystems. If successful, these interventions, alongside the development of Blue Nature Credits, may offer a way to unlock capital for ecosystems that have historically lacked funding.
While dive surveys can provide high-resolution data including species composition and physical characteristics of the subtidal environment, they are expensive, require specialised scientific divers, and only cover small areas. Geospatial classification techniques could better support broad-scale habitat identification and monitoring over a significantly larger area at relatively little cost. However, at present it is difficult to accurately depict biodiversity and physical traits of subtidal marine habitats. Here, we aim to combine aerial imagery with diver surveys using machine learning and deep learning algorithms to classify near-shore broad-scale marine habitat.
We have begun developing a pipeline using Python and QGIS to classify these marine habitats, mitigating the issue of water column reflectance contribution to seabed reflectance by applying the Depth-Invariant Index (DII). The initial use-case trialled feeding DII layers through a Random Forest model with drone imagery that captured Blue, Green, Red, Red Edge, and Near Infrared (NIR) bands at ~5 cm resolution. Validation accuracy scores are promising and justify continual pipeline development to enhance marine habitat identification at different locations and times.
Laura is a geospatial analyst with expertise in GIS, statistical analysis, and user experience design, applying these skills to transdisciplinary problem-led research in marine and terrestrial ecosystems. She holds a BSc in Environmental Science and Geographic Information Science and a BBus in Sustainable Enterprise and Design. At EnviroStrat, she leverages her geospatial expertise to support impact investment projects, ecosystem restoration, and nature-based solutions. Day-to-day, Laura utilises Python, R, PostgreSQL, and QGIS. She brings a technical and analytical perspective to sustainable resource management and blue economy development, with a particular interest in utilising deep learning, remote sensing, and ocean condition models.
I am a remote sensing specialist who focuses on the environmental challenges we face today, with a particular interest in the maritime. I started in Physical Geography at Durham (UK), before moving on to a Masters in Environmental Science at Cambridge (UK). Since finishing a PhD in glaciology (Lund, Sweden) I worked in commercial geospatial intelligence, before moving back into academia with a focus on the validation and application of remotely sensed data (land surface temperature) at King’s College London (KCL). These projects involved collaboration with ILRI, NASA-JPL and ESA, working in diverse environments from East Africa to China and across Europe. After KCL I continued work on land surface temperature at the UK's National Centre for Earth Observation (NCEO). Most recently I spent time working in the maritime domain before taking up the post of lecturer at the University of Auckland.