Site icon Conservation news

Portable scanners that identify timber species could detect illegal logging

  • Scientists have harnessed near infrared spectroscopy (NIRS) in a portable device to efficiently distinguish among structurally similar wood species.
  • NIRS could help environmental and trade agents in forests, on roads and at control posts, ports and borders detect unlawful timber in real time, strengthening forest law enforcement around the globe.
  • With continued testing of the scanners and associated identification model over time, combined with developments such as procedures for offsetting wood humidity, NIRS could facilitate future wood identification.

Many timber woods of conservation concern are hard to distinguish anatomically from species resembling them, especially when the wood sample lacks leaves or other botanical features. For instance, big-leaf mahogany (Swietenia macrophylla), classified as vulnerable on the IUCN Red List, is confused with the structurally similar Crabwood (Carapa guianensis), Cedar (Cedrela odorata), Cedrinho (Erisma uncinatum) and Curupixá (Micropholis melinoniana). It’s critical that law enforcement accurately discriminate among woods to prevent the illegal logging and trading of threatened tree species without hindering legal trade.

Wood anatomist José Arlete Camargos identifying cedar wood log in Acre, Brazil. Photo credit: Project NIRS ID.

A team of scientists at the Brazilian Forest Service’s (SFB) Laboratory of Forest Products (LPF) has harnessed near-infrared spectroscopy (NIRS) in a portable device to quickly and correctly identify wood to the species level.  The technology is an indirect chemical analysis involving a complex interpretation of the near-infrared region of the material’s electromagnetic spectrum and the application of multivariate statistics, or chemometrics. Researchers hope it will help environmental and trade agents in forests, on roads, at control posts, ports, borders and other stages of the supply chain detect unlawful timber in real time and strengthen forest law enforcement around the globe.

Fieldwork team preparing wood surface and taking NIRS spectra of big-leaf mahogany in Acre, Brazil. Photo credit: Project NIRS ID.

“Considering law enforcement will have an agile tool for wood identification in the forest, we believe it can help fight illegal logging, contribute to the preservation of native forests and add value to companies working within the law,” stated Tereza Pastore, Jez Braga and Vera Coradin, collaborators on the project at the University of Brasilia. “Our ultimate goal is to see this technology used to combat illegal logging and help loggers correctly identify forest species.”

So how does this wood scanner work? According to Pastore, an expert in wood’s analytical chemical analysis and spectroscopy, just as every human has a distinctive fingerprint, every tree species has a distinctive wood surface spectrum; individual trees of a species have nearly identical structural arrangement and traits such as growth rings, grain, texture, figure, brightness and color.

“The spectrum provides information on both the chemical composition and physical distribution of cellulose, lignin and extractives on the surface being studied,” she said.

NIRS determines a wood’s species by comparing its surface spectrum to a database of surface spectra for different tree species, particular to its area of origin, one at time. For example, the Brazilian wood discrimination model includes seven species similar to mahogany, the high-value timber species; the technology checks the unknown sample against every species starting with mahogany until a match occurs. A robust, representative identification model incorporates thousands of samples; mahogany alone has samples from about 3000 spectra across its occurrence zone in northern Brazil.

Two different portable devices, MicroPhazir and MicroNIr, being used to acquire NIRS spectra of woods in Santa Cruz, Bolivia. Photo credit: Project NIRS ID.

NIRS is advantageous for chemically identifying the heterogeneous and complicated surfaces of timber. First off, it’s accurate to species level 95% of the time. It’s also fast; in terms of preparation, samples just need to be sanded and cleaned before testing. The equipment the scientists chose, either one of two portable devices called MicroPhazir and MicroNIr, is commercially available and user-friendly. The chemical analysis is also non-destructive and doesn’t generate waste. Last but not least, it costs only about half of conventional wet chemical analysis, chemistry performed on liquid samples.

“The time needed to prepare the wood sample, fire the NIRS device’s trigger and obtain the spectrum with the result on the screen takes only a few seconds,” explained Pastore. “There is much work built into the process: 1) obtaining samples of different trees, 2) obtaining samples of different origins for all species we want to identify, 3) assembly of the database of a large number of samples, and 4) construction of a statistical model that will compare [the] drawn spectrum of your sample with database spectra.”

Wood anatomist Vera Coradin identifying lumber before the NIRS measurement is taken at a sawmill in Brasilia, Brazil. Photo credit: Project NIRS ID.

The research team has completed three field trials with the timber scanners. The first was on a farm in Acre state authorized to log mahogany and cedar; the pilot incorporated variables such as wood moisture and surface cutting direction into the identification model, which includes the various parameters the scanners assess. The second test occurred in sawmills in Brasilia, the third in sawmills in Santa Cruz, Bolivia. The group’s latest work uses NIRS to determine the national origin of mahogany from Brazil, Bolivia, Guatemala, Mexico or Peru. NIRS spectra can detect small variations resulting from the trees’ different locales and classify them by country of origin. Further technological refinement could allow accurate identification at finer scales, which would help find out if high-value wood comes from legally harvestable areas.

The rise of computing in the 1980s, with the optimization of electronic components and advent of microprocessors, popularized statistical chemical interpretations of near-infrared spectra. In 2000, Pastore conceived the idea of applying NIRS to identifying wood in a congress about methods in analytical chemistry that introduced her to the technology. Here, she helped Coradin, an LPF research associate and wood anatomist experienced in identifying Brazilian wood, examine chemical characteristics of commercial Brazilian timber for identification. Aware of the challenge environmental inspectors faced in discriminating among the huge variety of Brazilian timber species, Pastore realized NIRS could help detect native woods. In 2007, Braga, a spectroscopist and chemometrist who conducts multivariate analysis on the results of the wood’s spectra, arrived at the University of Brasilia and joined Coradin and Pastore to apply NIRS to Brazilian timber identification. Organizations for tech development research such as the International Tropical Timber Organization (ITTO)-CITES program, National Council for Scientific and Technological Development (CNPq) and National Institute of Science and Technology of Bioanalytics (INCTBio) funded the project.

Work team with NIRS scanners. From left to right: Diego Carvalho (chemistry student), Tereza Pastore (wood chemist), Vera Coradin (wood anatomist), Alexandre Gontijo (wood anatomist), Liz Soares (chemistry student), Nayara Guimarães (chemistry student) and Jez Braga (chemiometrist). Photo: Project NIRS ID.

As with any breakthrough innovation, developing and using the NIRS wood identifier hasn’t been without setbacks. For one, initial costs were high, financial resources low. What’s more, the initiative lacked backing in its early phase. It also required acquiring high-precision equipment, collecting tedious and hard-to-obtain samples of each target forest species in saw mills and exploited forests and formulating the identification model in laboratory and field conditions. In addition, the technology requires regular model and equipment maintenance and monitoring of reliability of results, as any such method would. Another limitation is that NIRS’ timber identification accuracy decreases as the number of species in the global model increases. To ensure correctness, the technology can only compare a number of species that are actually confused with one another and relevant to logging at a regional rather than global scale.

Though NIRS has been applied to food and pharmaceutical quality control and soil chemistry assessment for over thirty years, its widespread application to timber identification has a ways to go. Pastore, Braga and Coradin have reached the end of the experimental stage, evaluating the portable device’s performance and global model’s stability over time. They still need to create procedures for offsetting the humidity in wood samples and for updating the model. Next year they’ll pilot the technology with the Brazilian Institute of Environmental and Renewable Resources (IBAMA)’s enforcement officials.

“In the future we will have various methods to be applied in different situations,” projected the researchers. “No method of identification is a panacea. We are trying to give our small contribution to solving this major issue.”