Héctor Hernández‐Alonso;Jaime Madrigal‐González;Fernando Sill

Area of Ecology, Faculty of Biology, Universidad de Salamanca, E-37007, Salamanca, Spai;Area of Ecology, Faculty of Biology, Universidad de Salamanca, E-37007, Salamanca, Spain; iuFOR-EiFAB, Campus de Soria, Universidad de Valladolid, E-42004, Soria, Spain

Legacy effects;Drought disturbances;Tree growth;Tree-ring width;Dendrochronology;Individual scal

The impact of lag effects produced by disturbances on primary production has been a major concern among ecologists during the last decade. Sudden and extreme climatic events are imposing drastic reductions in radial growth of trees as evidenced in tree-rings series Dendrochronological samples are obtained at tree level but analyzed at an aggregated scale (i.e., mean chronologies), although aggregating tree-ring chronology on a regional scale may reduce the possibility of studying the variability of individual tree response to drought, by amplifying the average population response. Here, we conducted experimental research in which 370 trees of 5 species were analyzed to assess the potential statistical and scaling issues that may occur when using regression-based methods to analyze ecosystem responses to disturbances. Drought legacy effects were quantified using individual and aggregated scales. Then, lag effects were validated using confidence and prediction intervals to identify values falling outside the certainty of the climate-growth model Individual scale legacy effects contrasted with confidence intervals were commonly distributed across species but were scarce when compared with prediction intervals. The analysis of aggregated scale legacies detected significant growth reductions when validated using prediction intervals; however, individual scale legacy lag effects were not detected. This finding directly contrasts the results obtained when using an aggregated scale. Our results provide empirical evidence on how aggregating ecological data to infer processes that emerge from an individual scale can lead to distorted conclusions. We therefore encourage the use of individual based statistical and ecological procedures to analyze tree rings as a means of further understanding the ecosystem responses to disturbances.