Premium
Apparently we do need phytosociological classes to calibrate Ellenberg indicator values!
Author(s) -
Wamelink G.W.W.,
Dobben H.F.,
Berendse F.
Publication year - 2003
Publication title -
journal of vegetation science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.1
H-Index - 115
eISSN - 1654-1103
pISSN - 1100-9233
DOI - 10.1111/j.1654-1103.2003.tb02190.x
Subject(s) - indicator value , sigmoid function , vegetation (pathology) , mathematics , statistics , vegetation types , linear relationship , ecology , environmental science , biology , habitat , medicine , pathology , machine learning , computer science , artificial neural network
. We stated (Wamelink et al. 2002) that mean Ellenberg indicator values are biased towards expectations of phytosociologists. Witte & von Asmuth (2003; this volume) have two major points of criticism: (1) the data we used would be systematically biased; (2) in calibrating Ellenberg indicator values for moisture availability against mean spring groundwater level we should have assumed a sigmoid response instead of a linear one. As to (1), a bias in the data would require that wet vegetation types were visited in wet years and dry vegetation types in dry years. We do not see any evidence for this. As to (2), our data do not provide strong evidence for a sigmoid relation instead of a linear one. Neither is there any indication that the bias in the Ellenberg indicator values would disappear when a sigmoid function would be fitted. We do agree with Witte & von Asmuth that it is preferable to characterize the species’ response by those variables to which they most directly respond.