Response of freshwater invertebrates to natural environmental and agricultural stressor gradients across larger scales
Final Report Abstract
Agrochemicals, especially pesticides, have been shown to be important contributors to the local loss of many sensitive species, for instance, up to 42% of invertebrates in freshwater ecosystems of different agricultural regions of the world. Most research has focused on pesticide effects on the individual and population level, which constrains the ability of scientists to appropriately assess the ecological risks and predict community-level effects. Trait-based approaches have been advocated to establish mechanistic trait-stressor relationships that allow for community prediction and for discrimination between stressors, which is of particular relevance for freshwater managers. Various studies scrutinised whether similar environmental conditions in a habitat lead to a predictable trait composition of freshwater invertebrate communities. The results show that the predictability is often context-dependent. One fact that has insufficiently been considered in many studies on traitenvironment relationships and that may explain part of the context-dependency, is that traits are often intercorrelated. In Work Package (WP) 1 we extracted trait profiles and groups for the taxa pools from different continents and regions, namely Europe, North America, Australia, New Zealand, and Bolivia using trait databases. The taxon pool and coverage of trait information was too small for Bolivia compared to the other regions and we excluded this region from the analysis. During data processing we identified a number of data-related issues that complicate the use of invertebrate traits for ecological studies such as our large-scale comparisons of trait profiles: (i) inconsistency in terminology (here we use the terms “trait” and “grouping feature” following the terminologies suggested by Schmera et al. (2015)), (ii) invertebrate trait databases include different grouping features and related traits (e.g., European database 10 traits for feeding mode; New Zealand database 6), (iii) coding of trait data is inconsistent across trait databases (e.g., fuzzy, categorical, binary), (iv) differences in the taxonomic resolutions (e.g., species, genus, or family) used when comparing among trait databases, or when linking observed taxonomic data to trait databases, and (v) availability of trait information varied strongly across grouping features and databases. Hence, we tackled these data-related issues first with the goal of establishing comparable harmonised trait datasets for each region. Subsequently, we developed harmonised trait datasets for the grouping features voltinism, body form, size, respiration, locomotion, feeding mode, and oviposition. When compared to non-harmonised data in a case study, we found that the harmonised trait datasets were generally effective in identifying trait - environmental relationships. After establishing these harmonised trait datasets, we extracted and compared trait profile groups (TPGs) and associated trait states across and within the continents and regions. We also added a recently developed trait dataset from South Africa to the analysis. Our hypothesis that invertebrate trait profiles have converged across continents and regions was partly confirmed. We found that trait spaces (based on principal coordinate analysis) of the considered continents and regions largely overlapped, except for Australia. However, with cluster analysis we could only identify two sets of traits that, in various combinations, defined TPGs on all continents and regions. TPGs that were similar across all continents and regions were characterised by the traits crawling, cylindrical body form, and univoltinism or by the traits respiration via plastron and spiracle and size small. Though the most important traits for grouping freshwater invertebrates into TPGs were rather specific for each continent and region we found that feeding mode traits were most important for the partitioning of families into different TPGs on all continents and regions. Furthermore, we found no strong taxonomic signal in the delineated TPGs. We also tested for differences in the trait structure between temperate and cold regions in Europe and found no difference. In WP2 we currently identify the most important response traits and trait profile groups (established in WP1) for predicting pesticide stress, also accounting for other agricultural stressors and different stream types. We will examine to which extent accounting for these intercorrelations in terms of trait profiles improves the consistency of trait-environment relationship across larger scales, defined here as continental and intercontinental scale, and consequently improves prediction of community structure. Afterwards, we will evaluate to which extent criteria for stressor-specificity are met by the response traits with the highest predictability for pesticide stress. Finally, we will scrutinise their specificity for pesticide stress in an agricultural multiple stressor context including additional stressors such as excessive nutrient loading. Overall, our project will inform on the utility and limitations of trait-based approaches, for instance, regarding the transferability of trait-environment relationships across regions and for use in agricultural multiple stress contexts.
Publications
-
Tackling inconsistencies among freshwater invertebrate trait databases: harmonising across continents and aggregating taxonomic resolution. Freshwater Biology, 67(2), 275-291.
Kunz, Stefan; Kefford, Ben J.; Schmidt‐Kloiber, Astrid; Matthaei, Christoph D.; Usseglio‐Polatera, Philippe; Graf, Wolfram; Poff, N. LeRoy; Metzeling, Leon; Twardochleb, Laura; Hawkins, Charles P. & Schäfer, Ralf B.
