PIONEER: A Prototype Service for Forest Inventory and Health Monitoring using Endurance Drones and Citizen Science

by N. Teferle and A. König

Forest ecosystems can play important roles as habitats for diverse species, as drivers of bio-geo-chemical cycles including the water and nutrient cycles, they can promote water infiltration in soil and help the build-up of ground water bodies, and as such they can contribute also to the resilience of regional ecosystems against impacts of extreme weather events, such as droughts.

At the same time, individual trees can be vulnerable to predicted impacts of anthropogenic climate change, including old and new plant pests and diseases. The timely identification, mapping and characterization of trees has thus gained in importance for evidence-based and future-oriented management and policies. Traditional field-based inventory surveys are limited in scope and highly time-consuming and expensive to monitor the forests in the Grand Duchy of Luxembourg (GDL). A paradigm shift in forestry monitoring is called for, with forests needing to be understood as complex socio-ecological systems that require adaptive management and governance to make forests and forest management more resilient and sustainable. New technologies for remote sensing and citizen science promise innovative approaches that may be able to help respond to these new needs.  The interdisciplinary project PIONEER explores the development of an innovative approach to produce timely, accurate and accessible evidence on forest ecosystems and their services by combining remote sensing and citizen science.

Firstly, PIONEER will investigate the use of endurance unmanned aerial vehicle (UAV) platforms, i.e., new systems that can cover up to 100 km2 per day, with light detection and ranging (LiDAR) and hyperspectral image (HSI) sensors to collect data on the forest canopy. The approach will rely on developing deep learning (DL) approaches for improving the extraction of known and new forest metrics. LiDAR data will be drawn upon to describe canopy architecture, encompassing the horizontal and vertical structure at scales of individual trees, while HSI data does so in describing physiological traits of individual trees from their spectral signature. The developed DL algorithms will extract 3D ITCs and integrate them with their spectral signatures for an efficient way to classify species, while providing a digital model for precision forestry to identify tree growth rates and health at a scale of individual trees, investigating spectral characteristics associated with early infestations, and providing the forest metrics that enhance monitoring and sustainability of forestry. A critical aspect for DL is the gathering of enough labeled bespoke training and validation data for successful and efficient classification, segmentation and identification models.

Secondly, a citizen science approach will be co-designed with experts, professionals, forest owners and citizen volunteers to allow the collection of complementary data from below the canopy that presents data on both the state of health of individual trees and estimated ecosystem services such as carbon sequestration and water infiltration in the ground. Furthermore, PIONEER will explore whether data for deep learning can be collected with citizen science. A citizen science tool set for participatory monitoring will help to enrich official data pools but also provides educational means that can be used by schools and in life-long learning offered as part of leisure activities to develop an appreciation of complexity and renegotiation of what is valued in the community.”

Both remote sensing and citizen science will be leveraged in tandem to develop a tree-centric approach to carbon density estimates, biodiversity, and forest diseases based on identifying individual tree crowns (ITCs) and species by utilizing the complementarity in these data sources. This combined approach promises to respond to new data needs for new official environmental accounting approaches considering ecosystems and ecosystem services as well as economic aspects that are being proposed at the level of the UN and discussed within the EU.

Prof. Norman TEFERLE

Prof. Ariane KÖNIG