Attribute-based classification of European hydrophytes and its relationship to habitat utilization

Citation

Willby N, Abernethy VJ & Demars BOL (2000) Attribute-based classification of European hydrophytes and its relationship to habitat utilization. Freshwater Biology, 43 (1), pp. 43-74. https://doi.org/10.1046/j.1365-2427.2000.00523.x

Abstract
1) Here we classify selected European hydrophytes into ‘attribute groups' based on the possession of homogenous sets of characteristics, and explore the correspondence between these attribute groups, or individual attributes, and habitat use. 2) Non-hierarchical clustering was used to assign 120 species to twenty groups based on a matrix of categorical scores for literature- and field-derived information covering seventeen intrinsic morphological and life-history traits. Subdivision of some of these traits produced a total of 58 attributes (i.e. modalities). The robustness of this classification was confirmed by a high rate of reclassification (92%) under multiple discriminant analysis (MDA). The phylogenetic contribution was explored using ordination methods with taxonomy at family level acting as a covariable. 3) Our approach differed from earlier classifications based on growth or life form because we regarded growth form plasticity as a property of the species and its range of growing conditions, rather than of each individual population, and we considered additional (e.g. regenerative) traits. However, some conventional life form groups were preserved (i.e. utricularids, isoetids, hydrocharids and lemnids). 4) Some parallels existed with established theory on terrestrial plant growth strategies, but we used strictly intrinsic attributes relevant specifically to hydrophytes and our groups could not be decomposed into three or four primary strategies. Only finer levels of partitioning appear to be of fundamental and applied ecological relevance in hydrophytes. 5) A principal components analysis ordination based on 26 attributes related to physical habitat utilization separated species and their attribute groups along axes relating to: (a) flow, substratum grade and organic matter content, scour frequency, and sedimentation; and (b) depth, water level stability and biotic disturbance. A MDA applied to species ordination scores indicated only a modest overall correspondence between attribute groups and habitat use (54% correct reclassification). Poor reclassification was the result of intergroup overlap (indicating alternative sets of attributes for a given habitat) or high intragroup variance in habitat utilization (indicating commonality of attributes between different habitats). These results are interpreted in terms of trade-offs between resistance and resilience traits, ‘functional plasticity' in traits, phylogenetic dependence in some groups and methodological constraints. The predictive potential of hydrophyte groups and their limitations are discussed. 6) Redundancy analysis revealed a highly significant correlation between traits and habitat use (P < 0.01). Our attribute matrix explained 72% of variation in physical habitat use with eight attributes (i.e. turions, anchored emergent leaves, high or low body flexibility, high root:shoot biomass ratio, free-floating surface or free-floating submerged growth form, and annual life history) explaining half of this variation. 7) Most attributes were mapped in accordance with habitat template predictions, although tests were confounded by the underlying correlation between spatial and temporal heterogeneity. The main features were: (a) a trade-off between resistance-type traits (related to stream lining, flexibility and anchorage) in more spatially heterogenous riverine and littoral zone habitats, and resilience type traits (i.e. turions, very small body size and free-floating growth forms) in spatially simple, rarely disturbed habitats, such as backwaters and canals; and (b) a shift from high investment competitive traits with a low reproductive output in deep stable habitats to classically ruderal and desiccation resistance traits in shallow fluctuating habitats.

Keywords
1 Here we classify selected European hydrophytes into ‘attribute groups’ based on the possession of homogenous sets of characteristics, and explore the correspondence between these attribute groups, or individual attributes, and habitat use. 2 Non-hierarchical clustering was used to assign 120 species to twenty groups based on a matrix of categorical scores for literature- and field-derived information covering seventeen intrinsic morphological and life-history traits. Subdivision of some of these traits produced a total of 58 attributes (i.e. modalities). The robustness of this classification was confirmed by a high rate of reclassification (92%) under multiple discriminant analysis (MDA). The phylogenetic contribution was explored using ordination methods with taxonomy at family level acting as a covariable. 3 Our approach differed from earlier classifications based on growth or life form because we regarded growth form plasticity as a property of the species and its range of growing conditions, rather than of each individual population, and we considered additional (e.g. regenerative) traits. However, some conventional life form groups were preserved (i.e. utricularids, isoetids, hydrocharids and lemnids). 4 Some parallels existed with established theory on terrestrial plant growth strategies, but we used strictly intrinsic attributes relevant specifically to hydrophytes and our groups could not be decomposed into three or four primary strategies. Only finer levels of partitioning appear to be of fundamental and applied ecological relevance in hydrophytes. 5 A principal components analysis ordination based on 26 attributes related to physical habitat utilization separated species and their attribute groups along axes relating to: (a) flow, substratum grade and organic matter content, scour frequency, and sedimentation; and (b) depth, water level stability and biotic disturbance. A MDA applied to species ordination scores indicated only a modest overall correspondence between attribute groups and habitat use (54% correct reclassification). Poor reclassification was the result of intergroup overlap (indicating alternative sets of attributes for a given habitat) or high intragroup variance in habitat utilization (indicating commonality of attributes between different habitats). These results are interpreted in terms of trade-offs between resistance and resilience traits, ‘functional plasticity’ in traits, phylogenetic dependence in some groups and methodological constraints. The predictive potential of hydrophyte groups and their limitations are discussed. 6 Redundancy analysis revealed a highly significant correlation between traits and habitat use (P < 0.01). Our attribute matrix explained 72% of variation in physical habitat use with eight attributes (i.e. turions, anchored emergent leaves, high or low body flexibility, high root:shoot biomass ratio, free-floating surface or free-floating submerged growth form, and annual life history) explaining half of this variation. aquatic plant; functional group; strategy; trait

Journal
Freshwater Biology: Volume 43, Issue 1