DIKUL - logo
(UL)
PDF
  • Integrated network models for predicting ecological thresholds: Microbial e carbon interactions in coastal marine systems
    McDonald, Karlie S. ...
    This proof of concept study presents a Bayesian Network (BN) approach that integrates relevant bio-logical and physical-chemical variables across spatial (two water layers) and temporal scales to ... identify the main contributing microbial mechanisms regulating POC accumulation in the northern Adriatic Sea. Three scenario tests (diatom, nanoflagellate and dinoflagellate blooms) using the BN predicted diatom blooms to produce high chlorophyll a at the water surface while nanoflagellate blooms were predicted to occur also at lower depths (>5 m) in the water column and to produce lower chlorophyll a concentra-tions. A sensitivity analysis using all available data identified the variables with the greatest influence on POC accumulation being the enzymes, which highlights the importance of microbial community in-teractions. However, the incorporation of experimental and field data changed the sensitivity of the model nodes >=%25% in the BN and therefore, is an important consideration when combining manipulated data sets in data limited conditions.
    Source: Environmental Modelling & Software. - ISSN 1364-8152 (vol. 91, 2017, str. 156-167)
    Type of material - article, component part ; adult, serious
    Publish date - 2017
    Language - english
    COBISS.SI-ID - 4241999

    Link(s):

    https://authors.elsevier.com/a/1UZYr4sKhE9jiq

    aktivni loink do 07. aprila 2017


    DOI