PARTNERSHIPS FOR ENHANCED ENGAGEMENT IN RESEARCH (PEER) Cycle 8 (2019 Deadline) Development of wood identification system and timber tracking database to support legal trade PI: Ratih Damayanti (ratih_turmuzi@yahoo.com), Forestry Research Development and Innovation Agency (FORDA) of the Ministry of Environment and Forestry, Indonesia, in partnership with Bogor Agricultural University USDA/FS Collaborator: Michael Wiemann, U.S. Forest Service, Forest Products Laboratory Project Dates: February 2020 - October 2021 Project Overview:   Wood sample collection (photo courtesy of Dr. Damayanti). This project is aimed at developing a wood identification system using a combination of methods to obtain quicker and more accurate results. Tools for improved wood identification accuracy at the species level could also help to differentiate wood origins and estimate when the tree was cut, something that is not currently possible. The technology will be integrated with the Indonesian Forest Product Administration Information System (SIPUHH) online, as part of the Indonesian Timber Legality Assurance System. The methods to be applied involve a combination of computer vision and spectrometry, integrated with an online digital database. The expertise of U.S. partners Dr. Wiemann and Dr. Hermanson from the United States Department of Agriculture will be critical in developing and testing the new system. Dr. Wiemann is a wood anatomist in the Center for Wood Anatomical Research at the Forest Products Laboratory, and Dr. Hermanson has developed a machine vision-based wood species classifier (XyloTron) and worked in other domains of data-driven machine-learning for classification and database creation, management, and mining. This PEER project also aims to establish a wood species database for timber tracking in supporting legal trade. Data from the existing wood species database from the Xylarium Bogoriense could also be integrated with new data from other regions in Indonesia. The digital collection should be useful as a guide in mapping the distribution of wood biodiversity in Indonesia, providing information on carbon stock based on wood density, and listing active compounds identified by spectrometry that could be useful for scientific and commercial purposes. The proposed project has strong relevance to Indonesian and USAID objectives for enhancing the country’s resilience and capacity for sustainable management of a rapidly depleting resource base of bio-diverse forests. In 2009, the Ministry of Forestry issued created the Timber Legality Assurance System with the goal of improving forestry management governance and eradicating illegal logging and trade. With regard to international trade with the United States, the Lacey Act requires verification of timber origins. Importers must declare the trade and botanical names of the wood species, the country of origin, the size and volume of the wood, and its value. Difficulties are regularly encountered in meeting these requirements, however, due to limited knowledge, technology, and staff capabilities. Because wood identification currently can only be carried out by trained or experienced researchers or officers, personal perception factors can also influence the accuracy of results and the length of time required for identifications. The development of an automated wood identification system and wood species databases would provide significant positive impact on the legal timber verification process. If the system is incorporated into the mandatory Indonesian system, it will support the trade in legal timber and help reduce illegal logging and preserve forests. Moreover, the establishment of an integrated wood database would create a comprehensive information source to improve the management of forest product utilization by various stakeholders, including the Ministry of Environment and Forestry, Customs, forest concessions, local governments, other related stakeholders, and communities. Final Summary of Project Activities: This project, which ended on October 31, 2021, had five main outputs. Following is a brief overview of progress made on each: Improving the existing automatic wood identification system (AIKO-KLHK Version 1): The PI Dr. Damayanti and her team collected an additional 350 wood specimens from the Xylarium Bogoriense and developed the system and supporting information in two languages. AIKO-KLHK Version 2 has been launched for public use. It accommodates 1,180 wood species and is available in two languages (Bahasa Indonesian and English). A national patent has been registered. Developing a more objective wood identification system based on wood capacitance, namely WIDER (Wood Identifier): The researchers completed specimen preparation, sensor design, material characterization, and data analysis for 15 wood species; a national patent has been registered. WIDER is a capacitance spectroscopy measurement system using a machine learning-based identification software. WIDER can measure wooden objects (identifying the wood species) in just milliseconds with an accuracy of greater than 95%. WIDER is portable and battery-powered, so it is suitable for field use. Further research is needed to improve the database and modify the sensor to increase its flexibility in identifying various shapes of timber. Assessing the current wood legality system in Indonesia and implementation/integration of the wood identification system into the Indonesian Timber Legality Assurance System (SVLK): The researchers have carried out trials in representative industries to whether the AIKO-KLHK can be integrated into the SVLK system. Focus Group Discussions were conducted twice, first with the Indonesian Timber Legality Assurance System (TLAS) Auditor and second with policymakers, including the Director General of Sustainable Forest Management of the Ministry of Environment and Forestry, Republic of Indonesia. Developing the Integrated Xylarium Bogoriense Database: Project researchers have recorded 232,020 wood specimens representing 6,679 species, 170 families, and 1,105 genera; a national copyright has been registered. Creating the ECVT 4D Dynamic Tree Monitoring System: A national patent has been registered. The physiological processes of trees are strongly influenced by environmental conditions. By understanding the physiological processes that occur, the appropriate silvicultural treatment can be determined, thus increasing forest productivity. Information on tree resilience to disturbances can also be useful for individual or population selection in tree breeding programs and climate change adaptation. The technology for monitoring tree physiological processes based on the nature of electrical permittivity allows the tree's response to the environment to be known in real-time. This was a fairly small project that was only originally expected to last for one year, and unfortunately it began only a month before the start of the COVID pandemic, so Dr. Damayanti and her team had to cope with many challenges, including lockdowns and serious illnesses on the part of team members. Thanks to their dedicated efforts and the additional time they received from PEER under a no-cost extension, they accomplished even more than they expected. Here are just a few data points: 40 research assistants and private company staff were trained for wood identification 27 research scientists and project team members were trained how to develop a wood identification system by using Deep/Machine Learning 43 research scientists and project team members were trained for patent drafting 4,054 users registered for AIKO-KLHK (four times more than the number of users at the start of the project) One product (the automatic wood identification system AIKO-KLHK Version 2, containing 1,180 wood species listed with wood properties and other supporting information in two languages) has been launched for the public One prototype of WIDER (Wood Species Identification System Based on Resistance, Capacitance, and Inductance Properties) has been produced and is ready for further development and implementation One Monitoring System for Tree Physiological Processes Based on Electrical Conductivity (ECVt 4D Dynamic) has been produced and is ready for further development and application One set of policy documents on the improvement of the Indonesian TLAS and the potential for integration of the wood identification system into the TLAS is ready for publication One integrated Database of the Xylarium Bogoriense Wood Collection for mapping Indonesian wood diversity is ready to be shared with the public Six wood industry organizations have been informed about and used AIKO-KLHK Four new research grants have been received by the PI (three from Indonesian funders and one from the JAPAN-ASEAN Science, Technology and Innovation Platform of Sustainable Development Research), totaling about $82,000 Although the PI Dr. Damayanti is not based at a university but instead is head of the Lignocellulose Anatomy Laboratory and curator of the Xylarium Bogoriense, a research and development institution under the Indonesian Ministry of Environment and Forestry, the results from her PEER projects have nevertheless had an impact on higher education. After learning of her project’s results, other research centers and universities began to pay more attention to the wood identification system. A new university course was created, including the development and integration of wood identification systems and wood forensics, and Dr. Damayanti was invited to serve as a lecturer. Overall, Dr. Damayanti characterizes the importance of PEER support as follows: “The PEER project made all ideas and concepts that had previously been impossible to implement using government funding finally possible to be implemented and paved the way for many methods to be developed for easier, cheaper, faster, and more accurate wood identification technology. Another important output from this project is the initiation of integration between the wood identification system and the timber legality assurance system. The latter was a very beneficial suggestion from the PEER Cycle 8 project reviewer, which changed our way of thinking about the importance of the wood identification system, not only at the national but also at the regional and international levels, for sustainable forest management and legal timber trade. Delivering this research output at many events has attracted international offers of cooperation, and it brought a significant positive impact on the development of this research area. The development of WIDER and 4D Dynamic opened the researchers’ and the government’s minds to how from this invention, many applications of wood capacitance, for instance, can be implemented in many areas in forestry, especially for non-destructive testing and development of a device for tree monitoring to study the adaption of trees to climate change. The PEER project has opened a new network that makes everything possible.” A 14-minute video on the work of Dr. Damayanti and her team, including the aspects supported by PEER, is available on YouTube. Publications, Patent Applications, and Database Rahmanto, R G H, Damayanti, R, Agustiningrum, D A, Oktapiani, C, Satiti, E R, Tutiana, Dewi, L M, Krisdianto, Andianto, Djarwanto, Pari, G, Karlinasari, L, Bramasto, Y, Aminah, A, Novriyanti, E, Siregar, I Z, Teruno, W P, Huda, M A, Rohmadi, Yusuf, A, and Nugraha, H. 2021. Anatomical comparison of branches and trunks of seven commercial wood species. IOP Conf. Series: Earth and Environmental Science 914 (2021) 012071. IOP Publishing. doi:10.1088/1755-1315/914/1/012071. Sistem Identifikasi Jenis Kayu Berbasis Sifat Resistansi, Kapasitansi, dan Induktansi (Wood Species Identification System Based on Resistance, Capacitance, and Inductance Properties), Indonesian Patent Number P00202106065 Sistem Pemantauan Proses Fisiologi Pohon Berbasis Sifat Dasar Hantar Listrik (Monitoring System for Tree Physiological Processes Based on Electrical Conductivity), Indonesian Patent Number P00202106083 Metode dan Sistem Identifikasi Jenis Kayu Menggunakan Convolutional Neural Network (CNN) dan Tensorflow (Methods and Systems for Wood Species Identification using CNN and Tensorflow ), Indonesian Patent Number P00202106087 Aplikasi Database Xylarium Bogoriense Terintegrasi (Integrated Xylarium Database), Copyright EC00202138462 AIKO-KLHK database: http://aiko.menlhk.go.id/dashboard/ Back to PEER Cycle 8 Grant Recipients