Fabien Laroche

Tree-related microhabitats (TreMs) are key features for forest biodiversity, and knowing their accumulation rate is essential to design integrative management strategies. Many types of TreMs are associated to large old trees and show slow ontogenical processes. The rarity of such TreMs (particularly in intensively managed forests) hinder the estimation of their occurrence rate along tree growth. Here, we used a continental meta-analysis on TreMs occurrence rate along tree growth to build informative priors for a model of trunk-base rot-hole occurrence on oaks within the Grésigne forest, France --- a context where stand management and tree DBH were confounded. We explored whether the use of informative priors could improve the identifiability, the precision of estimates and the predictive abilities of the model. Without prior information, the low variance of tree DBH within management modalities rendered the model poorly identifiable and prevented the detection of an effect of tree DBH per se across the range of explored tree DBH. By contrast, using informative priors contributed to improve the precision of estimates and lead to detecting a positive effect of tree DBH per se. Informative priors did not degrade the model fit and clearly improved predictive abilities on new stands. In particular, while the model without prior information did not predict the occurrence of trunk-base rot-holes significantly better than a purely random guess, the model with informative priors did. Irrespective of the prior used, models suggested that the high recruitment of trunk-base rot-holes in Grésigne may be a temporary management effect in stands undergoing conversion from coppice-with-standards to high forest through sprout thinning, which will lead to conservation issues for cavicolous saproxylic species when all conversions are complete. Because using informative priors was simple and beneficial in our study, it should be further explored in other local applied contexts to orientate forest management.

Etude de patrons de diversité, des gènes aux espèces, dans les métacommunautés : apports d’une théorie quasi-neutre.: La théorie neutre a été proposée en écologie pour expliquer la diversité biologique observée au sein des communautés. Postulant l’équivalence écologique des espèces, elle explique les patrons de diversité par le hasard, la migration et la spéciation. Elle propose un « modèle nul » réfutable pour tester des hypothèses sur les règles d’assemblage des communautés mais présente cependant deux limitations: (i) elle décrit souvent très bien la diversité des espèces dans les communautés naturelles, si bien que cette seule information est insuffisante pour tester des hypothèses plus complexes ; (ii) il existe peu d’hypothèses alternatives menant à des prédictions directement comparables à celles du modèle neutre, et ces hypothèses manquent souvent d’arguments biologiques, en particulier évolutifs, justifiant les niches postulées. Cette thèse propose de nouvelles approches liant écologie et évolution et visant à surmonter ces deux difficultés. Tout d’abord, la diversité génétique des espèces est intégrée à la théorie neutre afin d’étendre la base empirique sur laquelle tester l’existence de niches à de nouveaux patrons, tels que les corrélations entre diversité des gènes et des espèces. Le développement d’un modèle neutre unifiant les niveaux génétique et spécifique permet de préciser les conditions menant à de telles corrélations. Par ailleurs, on propose aussi une approche statistique plus générale de ces corrélations, en lien avec des variables environnementales, permettant d’affiner notre compréhension des conditions sous-jacentes, en s’appuyant sur l’exemple de communautés antillaises de mollusques d’eau douce. Dans un deuxième temps, on développe des hypothèses alternatives, des modèles « quasi-neutres », qui considèrent des organismes variant suivant un faible nombre d’axes de niche. Deux approches possibles sont explorées. La première porte sur les processus évolutifs sous-tendant des différences écologiques, en analysant l’évolution de la dispersion dans une métacommunauté et ses conséquences sur les patrons de diversité. Plus descriptive, la seconde approche vise à identifier des espèces à rôle écologique spécifique en s’intéressant à l’influence de la présence du thym dans des communautés de garrigue. L’ensemble du travail développé dans cette thèse plaide pour une théorie des communautés fondée sur la « présomption de neutralité » qui intègre espèces, gènes et traits, ainsi que les contraintes évolutives qui les unissent.

Inferring assembly processes from species co-occurrence data is a long-standing challenge in community ecology. Approaches that focus on detecting non-random spatial covariance between species occurrences are limited by the fact that spatial patterns can deviate from randomness for many reasons. Process-based null hypotheses are needed to overcome this limitation. Here, we explored the neutral theory of community ecology as a promising candidate. We built upon a robust property of neutral co-occurrences, the ‘rank consistency’: within a common regional pool, the presence probabilities of two species should be ordered similarly across local sites. We suggested performing pairwise tests of species rank consistency along ecological gradients of interest and creating a species network where rank-consistent species are connected. Network mo­dules then indicate species groups that do not co-occur neutrally with one another, hence making an important step towards the understanding of assembly processes. These modules can be further interpreted by relating their composition to species traits. We tested our framework on virtual data and successfully retrieved pre-defined functional groups without generating false positive detections. Then, we analysed two published examples on tropical trees and Mediterranean herbaceous communities. We found ecologically meaningful modules in both cases, hence illustrating the potential of our approach.

Biological pest control is a major ecosystem service and is known to depend on landscape heterogeneity. The composition and configuration of landscapes can affect natural enemy communities, trophic interactions, and pest density within agroecosystems. However, local agricultural management can interfere with natural enemy activity, so the positive effects of landscape heterogeneity may be disrupted by farming practices. Here, we studied the influence of landscape context and management options on the biological control of Lobesia botrana, one of the main insect pests of grapes. We focused on two complementary measures: predation rates, which reflect part of biological control potential, and plant damage, which reflects pest density and the associated infestation. We used a set of sentinel prey (eggs, caterpillars, pupae) to quantify predation rates across different developmental stages of the pest. The study was carried out in a landscape‐scale experimental set‐up consisting of 38 vineyards in Southwestern France. Using structural equation models, we show that predation rates on sentinel prey were affected by both landscape heterogeneity and local management practices. Higher pest predation rates were observed in landscapes with smaller vineyards and in vineyards with low applications of synthetic pesticides. We observed limited relationships between predation rates and grape infestation levels. However, our results suggest that predation rates at the pest pupae stage are significantly shaping infestation levels. Additionally, pest damage in spring and summer was primarily influenced by the intensity of local pesticide use and the grass cover in the field and exacerbated by the decreasing size of vineyards, while semi‐natural habitats had no effect on pest damage. We conclude that links between L. botrana infestation and biological control potential appear tenuous in our study region. This is likely due to the high local management intensity, as evidenced by the negative association observed between pesticide applications and predation rates. Nevertheless, both predation and infestation respond to landscape or field heterogeneity and pesticide use. Reducing the use of pesticides should be combined with multi‐scale diversification measures at field and landscape levels to amplify the predation potential.

Tree-related microhabitats (TreMs) are key features for forest biodi-versity, and knowing their accumulation rate is essential to design inte-grative management strategies. Many types of TreMs are associated to large old trees and show slow ontogenical processes. The rarity of such TreMs (particularly in intensively managed forests) hinder the estimation of their occurrence rate along tree growth. Here, we used a continental meta-analysis on TreMs occurrence rate along tree growth to build in-formative priors for a model of trunk-base rot-hole occurrence on oaks within the Grésigne forest, France — a context where stand management and tree DBH were confounded. We explored whether the use of infor-mative priors could improve the identifiability, the precision of estimates and the predictive abilities of the model. Without prior information, the low variance of tree DBH within management modalities rendered the model poorly identifiable and prevented the detection of an effect of tree DBH per se across the range of explored tree DBH. By contrast, using informative priors contributed to improve the precision of estimates and lead to detecting a positive effect of tree DBH per se . Informative priors did not degrade the model fit and clearly improved predictive abilities on new stands. In particular, while the model without prior information did not predict the occurrence of trunk-base rot-holes significantly better than a purely random guess, the model with informative priors did. Ir-respective of the prior used, models suggested that the high recruitment of trunk-base rot-holes in Grésigne may be a temporary management ef-fect in stands undergoing conversion from coppice-with-standards to high forest through sprout thinning, which will lead to conservation issues for cavicolous saproxylic species when all conversions are complete. Because using informative priors was simple and beneficial in our study, it should be further explored in other local applied contexts to orientate forest man-agement.

Quantifying the autocorrelation range of species distribution in space is necessary for applied ecological questions, like implementing protected area networks or monitoring programs. However, the power of spatial sampling designs to estimate this range is negatively related with other objectives such as estimating environmental effects acting upon species distribution. Mixing random sampling points and systematic grid (‘hybrid’ designs) is a classic solution to make a trade-off. However, fractal designs (i.e. self-similar designs with well-identified scales) could make an even better compromise, because they cover a wide array of possible autocorrelation range values across scales. Using maximum likelihood estimation in an optimal design of experiments approach, we compared errors of hybrid and fractal designs when simultaneously estimating an effect acting upon a response variable and the residual autocorrelation range. We found that Pareto-optimal sampling strategies depended on the feasible grid mesh size (FGMS) over the study area, given the sampling budget. When the FMGS was shorter than expected autocorrelation range values, grid design was the best option on all criteria. When the FMGS was around or larger than expected autocorrelation range values, the choice of designs depended on the effect under study. Fractal designs outperformed hybrid designs when studying the effect of a monotonic environmental gradient across space, while grid design was more efficient for other types of question. Beyond the niche identified in our analysis, fractal designs may also appear interesting when studying response variables with more heterogeneous spatial structure across scales, and when considering more practical criteria of performance such as the distance needed to cover the design.

Population monitoring programmes typically rely on sampling because it is impossible to survey all the sites within the study area. In such a situation, the general recommendation to obtain unbiased estimates of population trends is to select monitoring sites using probability sampling. However, site selection not based on probability sampling, such as selecting sites with the largest abundance of individuals at the beginning of the monitoring programme, is common in practice. Nevertheless, these methods carry the risk of obtaining biased trend estimates. Using simulations, McClure & Rolek (2023) investigated whether three non‐probability sampling site selection methods can yield unbiased trend estimates under some specific conditions. For two of these methods, that is selecting high quality sites and selecting sites known to be occupied, the authors conclude that there is a major risk of obtaining biased trend estimates. For the third method, that is selecting sites with the largest initial abundance, they found conditions in which unbiased estimates can be obtained. They conclude that the general recommendation to use probability sampling should be revised. Here, we show that the authors' results, although perfectly correct, do not invalidate this recommendation. First, we point out that the authors made strong assumptions about the populations' functioning in their simulations, especially that inter‐annual variance in abundance is similar for all sites, which is unlikely in most real populations. We show through simple simulations that even slightly relaxing this assumption invalidates the authors' results. We also point out that for most of the hypotheses made by the authors, it is generally not known at the beginning of a study whether they will be respected. Furthermore, the authors did not provide evidence that selecting sites based on high initial abundance leads to more precise trend estimates than probability sampling methods. Therefore, neither the benefits nor the risks of this method are known. We conclude that until evidence is provided that abundance‐based site selection improves estimate precision and the situations in which it provides unbiased estimates are clearly identified, using probability sampling should remain the rule.

We present a data set resulting from the first round of a national monitoring program of forest reserves. It contains 9538 permanent plots, distributed across 111 study sites in mainland France (including Corsica). Notably focusing on dead wood measurement, this protocol has primarily been applied in strict forest reserves and special nature reserves (sensu Bollmann & Braunisch 2013), with 68% (6494) of the plots being currently located in strict forest reserves (unmanaged) and 24.7% (2363 plots) in forests unmanaged for at least 50 years. Sites cover a large variety of ecological conditions, from lowland to subalpine forests, but with an underrepresentation of Mediterranean forests (Table 1). The protocol assesses all the stages of a tree's life cycle, from seedling to decomposed lying dead wood. On each plot, a combination of three sampling techniques was used: (1) fixed‐area inventory for regeneration, standing dead trees, living trees, and coarse woody debris (CWD) with diameter over 30 cm; (2) transect lines for CWD with diameter 30 cm (using a relascopic angle of 3%). Measurements include exact tree location (azimuth, distance), species, diameter(s), tree‐related microhabitats, decay stage and bark cover, and seedling cover. With ongoing climate change, the program network can also provide important information to monitor changes in forest ecosystems. It can also be used as forest management monitoring or conservation status assessment. These data are freely available for noncommercial scientific use (Creative Commons Attribution 4.0 CC BY SA 4.0) with attribution, and this paper must be cited if this material is reused.

Understanding the colonization process of species living in a dynamic fragmented habitat is essential to assess their persistence. In the metapopulation theory, the colonization of a species can be quantified using the turnover of occupancy in habitat patches. However, this approach is often limited by the feasible size of surveyed areas. Because many species are capable of long‐distance dispersal, such areas often constitute open systems undergoing colonization of propagules coming from outside, the ‘background deposition'. We focus on disentangling background deposition from local colonization among surveyed patches when analyzing turnover. We consider two spatial scales: 1) focal areas where all patches are monitored over time; 2) a larger extent, encompassing the focal areas, over which the distribution of the target species is quantified with a coarse spatial grain. Our key idea is to use the regional connectivity of focal areas within the larger scale as a covariate when analyzing colonization events within focal areas. A positive effect of regional connectivity on the colonization probability of patches may indicate background deposition. We applied this approach to the epiphytic bryophyte Dicranum viride in a managed temperate deciduous forest, considering phorophyte trees as patches, forest stands as focal areas and the whole forest as the larger scale. We combined a fine‐grained turnover survey of occupied trees within three forest stands (~ 3 ha) with a coarse‐grained snapshot of D. viride distribution over the forest (~ 15 000 ha). Regional connectivity came out as the most significant factor, with a strong positive effect on colonization probability within stands. However, it was attributed to sources in the immediate vicinity of focal stands, suggesting a short‐ranged colonization process occurring across stands' borders rather than long‐distance background deposition. Our results thus call for maintaining a stepping‐stone of habitat across the forest through time to improve D. viride persistence.

Quantifying the autocorrelation range of species distribution in space is necessary for applied ecological questions, like implementing protected area networks or monitoring programs. However, the power of spatial sampling designs to estimate this range is negatively related with other objectives such as estimating environmental effects acting upon species distribution. Mixing random sampling points and systematic grid (‘hy-brid’ designs) is a classic solution to make a trade-off. However, fractal designs (i.e. self-similar designs with well-identified scales) could make an even better compromise, because they cover a wide array of possible autocorrelation range values across scales. Using maximum likelihood estimation in an optimal design of experiments approach, we compared errors of hybrid and fractal designs when simultaneously estimating an effect acting upon a response variable and the residual autocorrelation range. We found that Pareto-optimal sampling strategies depended on the feasible grid mesh size (FGMS) over the study area, given the sampling budget. When the FMGS was shorter than expected autocorrelation range values, grid design was the best option on all criteria. When the FMGS was around or larger than expected autocorrelation range values, the choice of designs depended on the effect under study. Fractal designs outperformed hybrid designs when studying the effect of a monotonic environmental gradient across space, while grid design was more efficient for other types of question. Beyond the niche identified in our analysis, fractal designs may also appear interesting when studying response variables with more heterogeneous spatial structure across scales, and when considering more practical criteria of performance such as the distance needed to cover the design.