Kana Kamimura

A review of strategies for wind damage assessment in Japanese forests

Forest management for minimizing wind damage risk requires predicting future wind damage as accurately as possible. However, previous studies carried out in Japan mostly focused on field observation and not on an extensive estimate of damage in various regions. This paper, therefore, aims to understand better approaches to the assessment of wind damage in Japan. First, basic descriptions of wind damage were reviewed including the process and types of tree and stand failures. Several factors relating to wind damage were also described including biological factors and stand characteristics. Second, the current methods of wind damage risk assessment were classified such as (1) observational/empirical, (2) statistical, and (3) mechanistic methods. Of the current methods, the mechanistic methods were acceptable in terms of their prediction of wind damage using the mechanistic behaviour of tree and stand as a result of wind pressure. Third, this paper reviewed previous case studies in Japan and showed that most studies of wind damage focused on particular typhoon events. Their conclusions might be difficult to apply to other settings for the estimation of future wind damage risk with the changes of stand condition (thinning, gap creation, etc.). Therefore, the mechanistic methods would be one of the most powerful approaches to estimate the possibility of future wind damage risk with changes of stand conditions. Further studies are required to develop the methods of wind damage risk assessment in Japan including the mechanical behaviour of tree and stand as a result of wind (typhoon) phenomena.

A hybrid method for feature construction and selection to improve wind-damage prediction in the forestry sector

Catastrophic damage to forests resulting from major storms has resulted in serious timber and financial losses within the sector across Europe in the recent past. Developing risk assessment methods is thus one of the keys to finding forest management strategies to reduce future damage. Previous approaches to predicting damage to individual trees have used mechanistic models of wind-flow or logistical regression with mixed results. We propose a novel filter-based Genetic Programming method for constructing a large set of new features which are ranked using the Hellinger distance metric which is insensitive to skew in the data. A wrapper-based feature-selection method that uses a random forest classifier is then applied predict damage to individual trees. Using data collected from two forests within South-West France, we demonstrate significantly improved classification results using the new features, and in comparison to previously published results. The feature-selection method retains a small set of relevant variables consisting only of newly constructed features whose components provide insights that can inform forest management policies.