Ecology of benthic macro-invertebrates in soft-sediment environments: A review of progress towards quantitative models and predictions


  • Andrew J. Constable

    1. School of Aquatic Science and Natural Resources Management, Deakin University, Warrnambool, Vic. 3280, and Australian Antarctic Division, Channel Highway, Kingston, Tas. 7050, Australia
    Search for more papers by this author
    • *Present address: Australian Antarctic Division, Channel Highway, Kingston, Tas. 7050, Australia.


This paper reviews the ecology of soft-sediment macro-invertebrates and, in particular, evaluates how much progress has been made in the recent literature towards elaborating statistical models of the ecology of these biota and in using quantitative predictions derived from these models. Steps to formulating statistical models on the dynamics of populations and assemblages are discussed. Current models are mostly conceptual (~70% of studies surveyed), falling into 2 main classes: population dynamics (including recruitment, mortality, dispersal and availability of patches) and process-orientated studies (including the response of individuals to the physical environment, biogenic habitat modification, biological interactions and physical perturbations). Most recent studies were descriptive or on the population dynamics of species and were undertaken mostly at spatial scales of up to 1 km and temporal scales of months. The development of statistical models appears to be impeded by the limited scope of studies, an over-emphasis on conceptual models (but recognizing an important role for a rigorous experiment framework) and a difficulty in using outcomes from small-scale processes at the level of individual organisms to predict larger scale outcomes where many interactions contribute to variation in abundances. Currently, few studies undertake field assessments of the defining characteristics of habitats, the dynamics of those habitats and the relative importance of different habitats to individual populations. Much urgent work is required to develop large scale (space: > 1 km; time: > 1 years) statistical models. This is because attention needs to be given to those interactions and processes in the ecological systems that will provide for the greatest reduction in uncertainty in the quantitative predictions derived from these models.