Matches in UGent Biblio for { ?s ?p Stoneflies are macro-invertebrates that are sensitive water quality indicators. Here, their occurrence was modelled based on physical–chemical water characteristics, river morphology and land use with five different modelling techniques. In a case-study in Flanders, stoneflies were found in 219 samples and two sets of absence data were gathered: 219 random samples from sites without stoneflies and 219 samples from sites downstream of each sampling site where stoneflies were observed. With both random and downstream absences, logistic regressions, artificial neural networks, support vector machines, random forests and classification trees could all successfully predict stonefly occurrence. For most environmental parameters, significant differences were found between sites with and without stoneflies. As stoneflies were only detected in a few percent of the samples, the ecological water quality is obviously still too low in most watercourses. Based on planned water quality improvement measures, an ensemble forecast using the five mentioned modelling techniques predicted that stonefly prevalence will only increase marginally by 2015 and 2027. To meet the European Union Water Framework Directive requirements, which states that all surface waters should obtain a good ecological quality, a more ambitious management plan is needed to decrease nutrient concentrations and improve habitat quality.. }
Showing items 1 to 1 of
1
with 100 items per page.
- aggregation abstract "Stoneflies are macro-invertebrates that are sensitive water quality indicators. Here, their occurrence was modelled based on physical–chemical water characteristics, river morphology and land use with five different modelling techniques. In a case-study in Flanders, stoneflies were found in 219 samples and two sets of absence data were gathered: 219 random samples from sites without stoneflies and 219 samples from sites downstream of each sampling site where stoneflies were observed. With both random and downstream absences, logistic regressions, artificial neural networks, support vector machines, random forests and classification trees could all successfully predict stonefly occurrence. For most environmental parameters, significant differences were found between sites with and without stoneflies. As stoneflies were only detected in a few percent of the samples, the ecological water quality is obviously still too low in most watercourses. Based on planned water quality improvement measures, an ensemble forecast using the five mentioned modelling techniques predicted that stonefly prevalence will only increase marginally by 2015 and 2027. To meet the European Union Water Framework Directive requirements, which states that all surface waters should obtain a good ecological quality, a more ambitious management plan is needed to decrease nutrient concentrations and improve habitat quality.".