Different aggregation approaches in the chironomid community and the threshold of acceptable information loss
2014
Аутори:
Milosevic, DjuradjStojkovic, Milica
Cerba, Dubravka
Petrovic, Ana
Paunović, Momir
Simic, Vladica
Тип документа:
Чланак у часопису (Објављена верзија)
Метаподаци
Приказ свих података о документуАпстракт:
Due to the problem of identification, Chironomidae larvae, although very
abundant, are often avoided or not properly used in bioassessment
programs. The aim of this work was to test how different aggregation
processes-taxonomic resolution and the random aggregation approach (best
practicable aggregation of species-BestAgg) affect the analysis of
chironomid communities regarding any information loss. The
self-organizing map method, together with classification strength
analysis and Spearman's rank correlation, revealed that the genus-level
and BestAgg-abundance matrix most accurately approximated the
species-level community pattern. The subfamily-level dataset was
ineffective at presenting the chironomid community structure, with a
substantially lower concordance with the species-level dataset. The
biologic environmental gradients analyses presented the same set of
important environmental variables for the species-level, genus-level,
and BestAgg-abundance matrix. The indicator values analysis showed that
indicator genera provide information very close to that gained from
species indicators. According to our results, the numeric relationship
between species and higher taxa influences taxonomic scaling, limiting
Chironomidae family aggregation, with acceptable information loss only
up to genus level. In addition, the BestAgg approach, with the maximum
level of aggregation, properly assesses the community structure and
consequently describes environmental conditions.
Кључне речи:
Aggregation; Taxonomic scaling; Chironomidae larvae; Bioassessment; Self-organizing map (SOM) methodИзвор:
Hydrobiologia, 2014, 727, 1, 35-50
DOI: 10.1007/s10750-013-1781-5
ISSN: 1573-5117