CFE 2019: Start Registration
View Submission - CMStatistics
B1719
Title: Estimating multispecies biodiversity indicators using hidden Markov models Authors:  Takis Besbeas - Athens University of Economics and Business (Greece) [presenting]
Abstract: Biodiversity indicators play a vital role in the assessment of changes in biodiversity. Several indicators have been recently developed, based on derived indices of abundance for individual species. We describe a state-space model formulation for the calculation of multispecies biodiversity indicators. The new approach has several advantages, including the ability to deal with unavailable data when species enter late in the study, incorporate estimates of uncertainty of the species index values when these are available, and allow smoothing within the indicator, rather than posthoc, as is frequently the case. We show how hidden Markov modelling may be used to efficiently fit the model in general, and illustrate the approach using abundance data from a 'citizen science' scheme on butterflies in the UK.