Automated Forest Restoration (AFR) Workshop

Brainstorming a research agenda to automate
tasks of restoring tropical forest ecosystems.

28-31/10/15, Grand View Hotel, Chiang Mai, Thailand

THIS WORKSHOP IS FINISHED. BUT YOU ARE WELCOME TO VIEW THE OUTPUTS BY CLICKING ON THE INFORMATION BUTTON ABOVE.

Rationale


Restoring tropical forests is of huge economic and ecological importance; to mitigate climate change, avert biodiversity losses, stabilize watersheds and supply forest products and services. In 2014, the UN called for forest restoration over 350 million ha by 2030, but how could it be done on such a vast scale? Most accessible land is already cultivated, leaving only remote sites for forest restoration. People are reluctant to carry trees and materials long distances across rugged terrain. The labour needed for tree planting, weeding, fertilizer application and monitoring often exceeds the local supply. Workers are unwilling to do such laborious and low paid work.
 

Tropical forest restoration techniques have advanced considerably over the past 20 years. In many areas, we now know which tree species to plant, when to collect seeds, how to grow planting stock and how to plant and maintain trees, until canopy closure and biodiversity recovery occur. But implementation still relies largely on Stone-Age technologies – hauling tree saplings and materials on foot, over long distances, across rough terrain.

"If we are to achieve the UN goal, forest restoration practices must be dragged
out of the Stone Age and into the Drone Age."

 
Therefore, in this brainstorming workshop, we invited restoration ecologists and technologists to 
explore the potential for "auto-restoration". We developed a research agenda, for the next 5 years which we hope will turn AFR from a pipedream into reality, by improving drone and visual recognition technologies and combining them with proven forest restoration practices.
 
VDO IntroductionSome of the questions addressed at the workshop were: could a new generation of autonomous "dendrones" spot seed trees and transmit their GPS co-ordinates to seed collectors? Or could they collect seeds themselves with robotic arms? How about aerial seeding by drones or even auto-weeding? Could drones be programmed to spray a non-residual, systemic herbicide on weeds, without harming young trees?

 
<<<<<< VDO Concept - click on the image



Objectives:
 
1)   to assess the current states of technologies that could be used to automate forest restoration tasks and identify               limitations or gaps in those technologies;
2)   to design research programs to improve technologies or combine them in innovative ways, leading to prototypes             for testing;
3)   to evaluate the feasibility of automated forest restoration from technical, ecological and socio-economic                         perspectives and
4)   to facilitate collaboration among technologists and restoration ecologists and the formation of interdisciplinary                 research teams.
 
Outputs:
 
1)    A research agenda, for the advancement of automated forest restoration, 2015-2020, based on the expertise of a            quorum of technologists and restoration ecologists.
2)    Material for the first edition of a multi-authored textbook on “Automated Forest Restoration”

Format
 
The workshop comprisd introductory presentations, by leading technologists and ecologists on the current state of technologies that could be used to automate various forest restoration tasks. Subsequently participants contributed to breakout discussion sessions to decide on research topics of most importance in advancing AFR. A field day to view restoration plots and demonstrate drone technology was included. Participants voted to prioritize research proposals on the final day. 
 
Topics will included:
 
1)    auto-seed collection- combining UAV1technologies with plant-recognition 
software to locate seed trees;
2)    aerial seeding – precise delivery, of seeds within degradable containers, with gels and coatings to maximize                      seedling establishment;
3)    auto-weed control – combining UAV and plant-recognition technologies to spray weeds with non-residual,                          herbicide, without harming trees, and/or developing herbicides, which kill weeds but not trees, and/or  selecting            herbicide-resistant trees;
4)    auto fertilizer application– using UAV’s to deliver precise fertilizer doses to trees and

5)    auto-monitoring– using UAV’s, with plant-recognition software, to assess tree survival/ growth, canopy closure,                plant diversity recovery and using remote cameras and/or microphones to monitor recolonization of restored plots          by bird/mammal species
 
Drone Building Master Class

An optional extra Master Class on drone building was run after the main workshop closed.
Our tutor was Mr Lot Amoros, artist, engineer and master drone builder. You can view his previous work here: https://vimeo.com/77119470,   https://vimeo.com/54890280.

Lot Amorós made his debut in the art world by programming reactive visual algorithms for live performances. He investigates the intervention of technology in the physical world, focussing on use of public air space and practical applications of unmanned aerial vehicles. He has developed data visualization interfaces, mixed reality performances and interactive audiovisual instruments. As an activist for open-source software and public data, he creates public-access wireless networks and disruptive devices. His creations have been exhibited in Brazil, Canada, Germany, the Netherlands, Egypt, and Spain.
 
Outputs of the workshop can be viewed here and if you have any questions about the workshop, please email the workshop secretariat: 

                           afr_secretariat@outlook.com


Primary Facilitators

The workshop’s primary facilitators were specialists in particular technologies, which may be applied to automating some of the tasks of restoring tropical forest ecosystems. They presented PowerPoints to update all participants on the latest technological advances and identify gaps or deficiencies for the break-out discussion groups to consider in more detail. They also drafted review papers, to prepare participants for the workshop.


Dr Steve Elliott – Forest Restoration Ecology

Stephen Elliott gained a PhD in forest ecology from Edinburgh University in 1985 and became a lecturer in wildlife conservation and tropical plant ecology at The Biology Department, Chiang Mai University, Thailand, in 1986. In 1994, he co-founded the Forest Restoration Research Unit and has concentrated on research and outreach to develop appropriate methodologies for restoring tropical forest ecosystems, ever since. 



Dr Lian Pin Koh - UAV’s in Ecology and Conservation
 
Lian Pin received his doctorate in Ecology and Evolutionary Biology from Princeton University in 2008. In 2014, he was appointed as Associate Professor and Chair of Applied Ecology & Conservation at the University of Adelaide, where he works on emerging environmental and socioeconomic challenges facing tropical developing nations, including the threats of intensifying land-use conflicts to natural ecosystems and wildlife. He is also Regional Technical Advisor for Conservation International, a TED speaker and Founding Director of the non-profit, ConservationDrones.org.

 
 
Computer Aided Plant Recognition

Dr. Dawn Frame - Canopy Science and Auto-Plant-Recognition


Dawn Frame gained her Ph.D. in Biology at City University of New York in 1990. She is a classical botanist and morphologist and has conducted research on tropical ecology for over 30 years. She developed her view of tropical forest ecosystems during 5 "canopy-raft" expeditions. She specializes in tree pollination and, along with colleagues at AMAP (botany, bioinformatics & plant architecture), Montpellier, France, developed plant recognition software (PlantNet-Mobile & Pl@ntNet-Identify) for general use and for fine-scale remote sensing studies. She sees remote sensing with UAVs as important tools for tropical forest ecologists and canopy scientists.

Dr. Carol Garzon-Lopez - Remote Sensing for Tree Species Recognition

Dr Carol Garzon-lopez, from Columbia, is a postdoctoral researcher at Fondazione Edmund Mac