Biology subjects

Changenet, A.

Publications and source records attributed to Changenet, A..

3 recordsLinked to original sources

bioRxiv2024

Evaluating genomic offset predictions in a forest tree with high population genetic structure

Predicting how tree populations will respond to climate change is an urgent societal concern. An increasingly popular way to make such predictions is the genomic offset (GO) approach, which aims to use genomic and climate data to identify populations that may experience climate maladaptation in the near future. More precisely, GO tries to represent the change in allele frequencies required to maintain the current gene-climate relationships under climate change. However, the GO approach has major limitations and, despite promising validation of its predictions using height data from common gardens, it still lacks broad empirical testing. In the present study, we evaluated the consistency and empirical validity of GO predictions in maritime pine (Pinus pinaster Ait.), a tree species from southwestern Europe and North Africa with a marked population genetic structure. First, gene-climate relationships were estimated using 9,817 SNPs genotyped in 454 trees from 34 populations; and candidate SNPs potentially involved in climate adaptation were identified. Second, GO was predicted using four methods, namely Gradient Forest (GF), Redundancy Analysis (RDA), latent factor mixed model (LFMM) and Generalised Dissimilarity Modeling (GDM), two sets of SNPs (candidate and control SNPs) and five climate general circulation models (GCMs) to account for uncertainty in future climate predictions. Last, the empirical validity of GO predictions was evaluated within a Bayesian framework by estimating the associations between GO predictions and two independent data sources: mortality data from National Forest Inventories (NFI), and mortality and height data from five common gardens in contrasting environments. We found high variability in GO predictions across methods, SNP sets and GCMs. Regarding validation, GO predictions with GDM and GF (and to a lesser extent RDA) based on the candidate SNPs showed the strongest and most consistent associations with mortality rates in common gardens and NFI plots. We found almost no association between GO predictions and tree height in common gardens, most likely due to the overwhelming effect of population genetic structure on tree height in this species. Our study demonstrates the imperative to validate GO predictions with a range of independent data sources before they can be used as informative and reliable metrics in conservation or management strategies.

evolutionary biology↗

bioRxiv2020

Occurrence but not intensity of mortality rises towards the climatic trailing edge of tree species ranges in European forests.

AimTree mortality is increasing worldwide, leading to changes in forest composition and altering global biodiversity. Yet, due to the multi-faceted stochastic nature of tree mortality, large-scale spatial patterns of mortality across species ranges and their underlying drivers remain difficult to understand. Our main goal is to describe the geographical patterns and drivers of the occurrence and intensity of tree mortality in Europe. We hypothesize that the occurrence of mortality represents background mortality and is higher in the margin than the core populations, whereas the intensity of mortality could have a more even distribution according to the spatial and temporal stochasticity of die-off events. LocationEurope (Spain, France, Germany, Belgium, Sweden and Finland) Time period1981 to 2014. Major taxa studiedMore than 1.5 million trees belonging to 20 major forest tree species MethodsWe develop hurdle models to tease apart the occurrence and intensity of tree mortality in National Forest Inventory plots at range-wide scale. The occurrence of mortality indicates that at least one tree has died in the plot and the intensity of mortality refers to the number of trees dead per plot. ResultsThe highest mortality occurrence was found in peripheral regions and the climatic trailing edge linked with drought, whereas the intensity of mortality was driven by competition, drought and high temperatures and was uniformly scattered across species ranges. Main conclusionsOur findings provide a new perspective in our understanding of tree mortality across species ranges. We show that tree background mortality but not die-off is generally higher in the trailing edge populations, but whether other demographic traits such as growth, reproduction and regeneration would also decrease at the trailing edge of European tree populations needs to be explored.

ecology↗

bioRxiv2019

Similar patterns of background mortality across Europe are mostly driven by drought in European beech and a combination of drought and competition in Scots pine.

AimBackground tree mortality is a complex demographic process that affects forest structure and long-term dynamics. We aimed to test how drought intensity interacts with interspecific and intraspecific competition (or facilitation) in shaping individual mortality patterns across tree species ranges. LocationEuropean latitudinal gradient (Spain to Finland). Time period1985 - 2014. Major taxa studiedScots pine (Pinus sylvestris L.) and European beech (Fagus sylvatica L.). MethodsWe performed logistic regression models based on individual tree mortality recorded in five European National Forest Inventories. We computed the relative importance of climatic drought intensity, basal area of conspecific and heterospecific trees (proxy of indirect intra- and interspecific competition or facilitation) and the effects of their interactions on mortality along the latitudinal gradient of both species range. ResultsIncrease in drought intensity over the study period was associated with higher mortality rates in both species. Drought was the most important driver of beech mortality at almost all latitudes while Scots pine mortality was mainly driven by basal area. High conspecific basal area was associated with high mortality rates in both species while high heterospecific basal area was correlated with mortality rates that were high in Scots pine but low in beech. Main conclusionsBeech mortality was directly affected by drought while Scots pine mortality was indirectly affected by drought through interactions with basal area. Despite their different sensitivity to drought and basal area, the highest predicted mortality rates for both species were at the ecotone between Mediterranean and cool temperate biomes, which can be explained by the combined effect of drought and competition. In the context of global warming, which is expected to be particularly strong in the Mediterranean biome, our results suggest that populations at the southern limit of species ranges may experience increased mortality rates in the near future. BIOSKETCHThe authors research is focused on functional trait ecology and global change, with special attention to mortality and demography processes. The authors use modelling multidisciplinary approaches to understand complex processes in ecology on a large geographical scale.

ecology↗