Pollinator monitoring: comparing standardised and novel survey methods

This thesis aims to explore the methods we use to survey insect pollinator populations, together with their biases, and their potential for development as part of a future national monitoring scheme. I will begin by focusing on one of the most commonly-used survey methods for assessing insect pollinator diversity: pan trapping. Pan trapping is a passive survey technique and is thus considered to be standardised due to its lack of collector bias. However, different protocols are employed by users of pan trapping worldwide, making comparisons between studies problematic; in addition, there is little research available to suggest which set of protocols maximises sampling ability. Chapter 2 will compare different pan trapping protocols in terms of their ability to sample bee and hoverfly populations, and make recommendations for a standardised protocol that can be used by future studies and monitoring schemes. Chapter 3 will build upon the theme of Chapter 2, focusing upon the sampling biases inherent to different survey methods. All survey methods have their own biases that affect the composition of the samples that they collect. Studies of insect pollinator-related survey methods, so far, have tried to quantify these biases by directly comparing methods to one-another, with pan trapping and transect surveys being the most often compared. But these studies lack any independent knowledge concerning the relative abundance of the insect populations present, and are thus of limited use. I will use a mark-release-recapture experiment in a closed island ecosystem to provide an independent source of relative abundance data, against which the rank abundance of samples collected via pan trapping and transect surveys can be compared, in an attempt to accurately quantify their respective sampling biases. I then move on to explore the development and performance of a novel method for surveying and monitoring insect pollinator communities: acoustics. Taxonomic skills are in decline worldwide, which is restricting our ability to generate reliable species-level identifications for data records, especially those collected by citizen scientists, and thus our ability to accurately monitor insect pollinator population trends. Chapter 4 will focus upon the novel use of bioacoustics to identify flower-visiting insect taxa at varying levels of taxonomic resolution, using the sound generated by their wing beats during foraging flights. Chapter 5 will build upon Chapter 4, focusing on the application of passive acoustic monitoring to survey for flower-visiting insects at a landscape scale. I will use commercially available automated detection software to extract the number of instances of insect flight sound from soundscape recordings, which can then be compared to the number of insects sampled by traditional sampling methods, in order to test the performance of this novel survey method. I also explore the performance of the automated detection software in terms of both false positive and false negative error rates. Chapter 6 will then synthesise the findings from Chapters 2-5, and put them in the context of the need for future systematic, standardised monitoring in response to insect pollinator decline, while suggesting avenues for future research.