Estimation of unbiased insect densities and density profiles with vertically pointing entomological radars

V. Alastair Drake

    Research output: Contribution to journalArticle

    8 Citations (Scopus)

    Abstract

    The variation with height of the number of insects in the air, and its relation to wind, weather, and other factors, is a topic of longstanding interest among entomologists concerned with migration and dispersal; it is now also receiving attention within the novel discipline of aeroecology. Radar, which provides one of the few practicable means of detecting insects hundreds of metres above the surface, is a proved source of information on insect density profiles. However, the ability of a radar to detect a target is strongly dependent on the target’s distance, and with the vertical-beam designs that are currently favoured for insect observation this range-dependence, if uncorrected, will significantly bias the estimated density profile. Detection can also depend on the target’s speed, and further biases can arise from systematic variations in echo-analysis success rates. In the procedure proposed here, targets are partitioned by size and densities are obtained independently for each class, with profiles being terminated at the maximum height at which all members of the class are detectable. Radar theory is drawn upon to calculate an ‘effective beam-width’ for detection of targets of a particular size at a particular height, and densities are ‘accumulated’ by forming a weighted sum incorporating individual widths and speeds. Accumulations are based on detections, using initial estimates of target size and, when available, speed. The subsample of fully analysed echoes can then be employed to further characterize the targets at different heights and to determine their direction of movement and orientation. Example applications illustrate how even prominent features, such as a pronounced layer echo, can be obscured, and artefacts generated, if one or more of the procedure’s components are omitted.
    Original languageEnglish
    Pages (from-to)4630-4654
    Number of pages25
    JournalInternational Journal of Remote Sensing
    Volume35
    Issue number13
    DOIs
    Publication statusPublished - 2014

      Fingerprint

    Cite this