Estimating Rooting Depth From Herbarium Specimens Might Be More Accurate Than Using Large Trait Databases

Global databases of plant functional traits are facing issues in data heterogeneity and taxonomical or geographical representativeness. To fill data gaps, natural history collections, such as herbaria, have become widely accepted as a potential source of data on functional traits. Surprisingly, root characteristics of plants still have not been studied on herbarium materials. We investigated whether rooting depth data from herbarium samples are realistic enough to be used in ecological studies. We measured original maximum rooting depth records on herbarium specimens and individuals from the field. Global data from the TRY database were also obtained. We tested the pairwise correlations between data from the three datasets. The effect of life form, taxonomic position, and average species height on rooting depth was also evaluated. Herbarium data showed strong correlation to field records, while records from the TRY database showed a weaker correlation with data measured on herbarium materials. Life form, taxonomic position, and height proved to be good predictors of rooting depth collected from the field or the herbarium; however, the model including data obtained from the TRY as the response variable performed weaker. We constructed an equation for predicting realistic average maximum rooting depth values of a given species based on herbarium data. Strong correlation among the field and herbarium datasets suggests that museal collections can be considered as resources of root trait data. Although herbarium-based rooting depth measurements usually represent lower values than field records, the correction of the herbarium-derived dataset is solvable. These corrected data might be more accurate than using large, global trait databases. Herbarium work might be a more sustainable, time- and cost-effective practice than field sampling. The inclusion of herbarium-derived information in trait-based studies, as well as in global databases, can improve these sources spatially, temporally, and taxonomically.