Bark beetle detection method using electronic nose sensors. A possible improvement of early forest disturbance detection?

Abstract

Forest ecosystems are long-term exposed to dry periods in Europe, which leads to a significant loss of vitality and higher mortality, especially in coniferous forests. Identifying stress in the early stages when measures can be taken to protect the forest and living trees is crucial. Current detection methods are based on field surveys by forest workers or remote sensing methods to cover larger areas, which use changes in spectral reflectance of the forest canopy. In some cases, the attacked trees do not change their appearance, and based on calculations of vegetation indices from remote sensing data, the attack cannot be mapped. We present an innovative methodology based on non-optical analysis, namely identifying a group of volatile compounds and microclimate signs in forest stands that indicate stress factors in forest stands. An attacked tree by a bark beetle produces increased amounts of biogenic volatile organic compounds associated with defense, and the microclimate changes due to interrupted transpiration. In addition, the bark beetle uses the aggregation pheromone to attract more individuals and to attack the tree massively. In this study, we tested three electronic noses (Miniature Bosch sensor device with 25,419 samples, Sensory device for environmental applications with 193 samples, Handheld VOC Detector Tiger with 170 samples) in a freshly infested spruce stand. The measurement was conducted at ground level with the help of a human operator and was repeated six times to verify the detection capability of the electronic noses. To verify the capability of electronic noses to predict tree infestation, we used machine learning Random Forest. The results demonstrated that electronic noses can detect bark beetle infestation start (within 1 week of the first attack). The Miniature Bosch sensor device achieved the highest accuracy with a value of 95%, in distinguishing forest sections that are healthy and infested; the second most accurate electronic nose is the Sensory device for environmental applications, with an accuracy of 89%. Our proposed methodology could be used to detect bark beetle presence.