engleski

Optimal Experimental Design for Sampling of Plant Pests for Early Detection

Pests are distributed unevenly across a location, therefore choosing where to sample, and the number and frequency of samples, is key to increasing the chances of detection. Sentinel plantings are a promising tool for detecting potential pests prior to their introduction into a country. Identification of the majority of species associated with the plant is important in order to identify potential pests before they enter and establish in a country. Detection of a wide range of species, including rare species, requires intensive sampling which increases the costs of sampling. It is thus important to balance sampling effort and costs. Given that the sentinel planting concept is relatively new, there is a lack of information and guidance on sampling for sentinel studies. Here we explore statistical sampling designs for the assessment of species diversity associated with woody plants. We outline factors that may affect diversity on different spatial and temporal scales and illustrate their influence on the fraction of the expected diversity that may be captured using several different case studies. (i) Differences in fungal diversity among and within seed lots of woody plants from three continents were assessed to illustrate local and regional diversity. (ii) We illustrate how variation in biodiversity within and among years can be captured by looking at long-term data from 18 year- long sampling of early oak defoliators. (iii) We examine between year variations in the fungal diversity of herbaceous plants and (iv) within year diversity of insect in nurseries. We discuss, with reference to previous sentinel studies, how the factors that cause the spatial and temporal differences in the diversity may be included in the sampling design and suggest an optimal sampling design for sentinel plantings. We also assess the effect of increasing costs in order to increase species detection.

Biodiversity, Risk based sampling, Spatial and temporal distribution, Sentinel plantation

nije evidentirano