Classification of Reynolds phytoplankton functional groups using individual traits and machine learning techniques

KRUK C.; DEVERCELLI M; HUSZAR V; HERNANDEZ E.; BEAMUD SG; DIAZ M; SILVA L.; SEGURA A.

FRESHWATER BIOLOGY (PRINT)

WILEY-BLACKWELL PUBLISHING, INC

Lugar: Londres; Año: 2017

0046-5070

1. The Reynolds Functional Groups (RFG)classification scheme is an informative and widely used method in ecologicalstudies of freshwater phytoplankton. It clusters species with similar traits,as well as common environmental sensitivities and tolerances. However,researchers face the difficulty to classify species into RFG because it relies in expert opinion,taxonomical knowledge and environmental information, which are not alwaysaccessible. Thus, a step forward is to build general statistical models toclassify species into RFG.2. Under the hypothesis that an organism?s response toenvironmental conditions determines their functional traits, here representedby the RFG, we predict that morphology and classification into broad taxonomicgroups will explain RFG independently fromenvironmental information and expert knowledge.3. To evaluate the predictive ability ofmorphological traits (e.g. volume) and taxonomic affiliation (e.g. chroococcalCyanobacteria) as discriminant variables of RFG, we compiled 1,300 species (264waterbodies) and applied Random Forest (RF) and Classification and RegressionTrees (CART). We divided the data to train the models and test theirperformance.4. RF successfully classified species intothe 28 RFG (only c. 10% test error) with an average individual RFGsuccess rate of 84.6 (range = 33%?100%). This is a relatively high percentage of successfrom an ecological point of view. It suggests that the selected variables areable to reconstruct the RFG and represent well environmental preferences,without including information about local environmental conditions asclassifiers, information that was available for the classification by experts.5. Our results reinforce the functionalbasis of the RFG as defined by Reynolds (Reynolds et al., 2002, J. PlanktonRes. 24, 417?428) and supportboth morphological traits and taxonomic classification as good proxies ofphytoplankton responses to environmental conditions. A dichotomous key based inthe CART was constructed and an R code to classify species into the RFG is freelyavailable. This work may help users to classify species into the RFG, includingthose that were not previously listed in Reynolds classification system.