Skip to main content
SLU publication database (SLUpub)
Doctoral thesis, 2017

Inventory strategies for monitoring and evaluation of forest damage

Roberge, Cornelia

Abstract

Under global change, increasing stresses on forests require strategies for monitoring and mitigation of damages caused by pests and diseases. While the threats to forests increase, so do the possibilities to set up efficient monitoring programmes and detect forest damage by utilising new technologies. This thesis focuses on strategies for forest damage inventories where different auxiliary data are combined to improve information for pest mitigation programmes. First, the efficiency of National Forest Inventories (NFIs; or similar inventories) for detecting and estimating state and change of forest damage across large regions was evaluated. NFIs were found efficient for assessing widely distributed damage, but unable to detect clustered and local outbreaks with adequate precision. Second, targeted forest damage inventories directed to areas with potential or suspected damage were investigated. It was found that two-phase sampling for stratification taking the first phase information from existing NFIs was an efficient strategy. Remotely-sensed auxiliary information and post-stratification was shown to further improve the precision. Third, the use of a new sampling design was evaluated: the local pivotal method (LPM), which spreads the sample in the multi-dimensional space of available auxiliary data. The LPM was found to be more efficient than simple random sampling in all scenarios and, depending on the allocation of the sample and the properties of the auxiliary data, it sometimes outperformed two-phase sampling for stratification. Thus, the LPM may be a valuable tool for practical forest damage inventories. Fourth, the cost-plus-loss method was applied to evaluate inventory strategies in a pest mitigation context. If inventory costs are large, it is especially important to quantify the inventory efforts necessary to evaluate the need for mitigation. The optimal sampling effort necessary for deciding whether or not a defoliator outbreak should be treated was quantified. Double sampling was found to be a cost-effective sampling strategy, i.e. the size of the second phase sample was determined based on the estimates from a small first phase sample. As an overall conclusion, the thesis points out the importance of making use of existing information in setting up effective inventories of forest damage and of using appropriate sampling strategies for making use of the information in the best possible way.

Keywords

cost-plus-loss; forest damage; forest inventory; forest pests; Monte-Carlo simulation; survey sampling; target tailored inventories

Published in

Acta Universitatis Agriculturae Sueciae
2017, number: 2017:2
ISBN: 978-91-576-8777-7, eISBN: 978-91-576-8778-4
Publisher: Department of Forest Resource Management, Swedish University of Agricultural Sciences