Relationships between the physicochemistry and macroinvertebrates of British upland streams: the development of modelling and indicator systems for predicting fauna and detecting acidity

1  Data on the macroinvertebrate riffle fauna and physicochemistry of 368 upland streams were collated for areas in Wales. Scotland and N.W. England. 2  TWINSPAN classification of the national macroinvertebrate data at the‘family' level produced either two or four groups of streams which differ...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Freshwater biology 1990-12, Vol.24 (3), p.463-480
Hauptverfasser: RUTT, G. P., WEATHERLEY, N. S., ORMEROD, S. J.
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:1  Data on the macroinvertebrate riffle fauna and physicochemistry of 368 upland streams were collated for areas in Wales. Scotland and N.W. England. 2  TWINSPAN classification of the national macroinvertebrate data at the‘family' level produced either two or four groups of streams which differed markedly in invertebrate fauna and acidity. The analysis also produced an indicator key based on the presence/absence of macroinvertebrate taxa which could be used to allocate new sites to the stream groups. 3  Multiple Discriminant Analysis (MDA) indicated that sets of environmental variables, which principally included pH or aluminium and calcium concentrations, discriminated effectively between the stream groups defined by TWINSPAN. These variables were incorporated into models which predicted faunal TWINSPAN groups from physicochemical data alone. 4  TWINSPAN classification and the MDA models were tested by application to new sites either reserved from the initial data set (RESERVED) or sampled during fieldwork in spring 1989 (FIELD). 5  For the RESERVED data there was 75–80% agreement between the two systems when two TWINSPAN groups were defined and 53–55% with four groups. Agreement was lower for the FIELD data, probably because chemical description was based on only one spot‐sample. 6  Distributions of pH and aluminium concentration for the groups of test sites classified by macroinvertebrate indicator key corresponded reasonably well to distributions of these determinands at sites in the four original TWINSPAN groups. Correspondence was better for the RESERVED than for the FIELD data set.
ISSN:0046-5070
1365-2427
DOI:10.1111/j.1365-2427.1990.tb00725.x