Current Earth Observation systems provide operational tools to derive areal evapotranspiration (ET) for drought monitoring and sustainable management of agricultural water. But partitioning ET into transpiration T and evaporation E is also key for targeting plant water use efficiency and plant water stress conditions at landscape scale. T and E are estimated through land surface models (LSM) forced by visible/near infrared and thermal infrared (TIR) remote sensing (RS) data. However, water budget-based LSM face parameterization issues to constrain water limited T and E rates, while dual-source energy budget-based LSM forced by TIR observations provide separate estimates of T and E, but rely on specific assumptions to retrieve both components from a single composite surface temperature. Additional information is thus required, either specifically related to E (surface water content, cf. Sentinel-1 S1) or T (Shortwave infrared SWIR, cf. Sentinel-2). HiDRATE aims at determining how the existing and future (TRISHNA, LSTM) TIR observations can map T and E by synergistic use of RS observations of multiple wavelengths in conjunctions with LSMs of various complexities. This includes the directional RS signature in terms of soil and canopy cover fraction in the sensor field-of-view. The relevance of increasing modeling complexity or the number of RS constraints in inferring T and E will be assessed using in-situ experiments at local scale including independent ET, E/T, T and E estimates based on eddy covariance, lysimeters, sap flow and stable isotope measurements for several biomes and climates. Drone campaigns will be organized to mimic the revisit cycle of the future satellites. HiDRATE builds on the complementarity of two groups who share expertise in TIR ET retrieval: on the French side, experts in E+T measurements/modeling, as well as soil moisture mapping from S1; on the Luxembourg side, experts in airborne mapping as well as plant water stress assessment using SWIR.

Droughts cause more dieback in forests when they are associated with strong heat waves, as during the summer of 2003 in France. However, the physiological mechanisms of the exacerbating effect of water stress by temperature on tree mortality remain to be elucidated. The main objective of this project is to explore a new mortality hypothesis, whereby xylem hydraulic failure is caused by an abrupt and uncontrolled increase in residual cuticular water loss (g_min) beyond a critical temperature Tp. Tp values from literature and preliminary results from a new mechanistic model (SurEau) justify this hypothesis. Here, we will measure g_min and its temperature dependence for selected forest trees species spanning a large range of drought tolerance (including deciduous, evergreen, temperate and Mediterranean species). For two key forest species threatened by climate changes (Fagus sylvatica and Abies alba), we will explore the variability of g_min and Tp across provenances selected from contrasted origins in their natural distribution range by working both on provenance trails as well as on natural populations. g_min will also be compared for selected poplar genotypes differing for their drought tolerance. The phenotypic plasticity of this trait will be evaluated in a poplar genotype grown under contrasted controlled environments for watering, shading or temperature. To phenotype g_min under controlled temperature, air humidity and light conditions, we will develop and distribute across all the partners of the project a new tool (Drought-Box). We hypothesize (H1) that species, genotypes and phenotypes more resistant to hot droughts display lower and more thermo-stable g_min and higher Tp values. The underlying physical mechanisms will be investigated on real and biomimetic cuticles. The chemical composition of leaf cuticles will be measured on relevant plant material from the above experiments to identify the key structural components associated with the variation of g_min across and within species. The contribution of these components in g_min and its thermo-stability will be evaluated in Arabidopsis mutants modified for their biosynthesis. We hypothesize (H2) that the thermostability of g_min is caused by the presence or proportion of specific molecules in the cuticle. Test experiments will be conducted ex-situ under controlled conditions to assess the impact of a hot drought on hydraulic failure and mortality to validate the predictions of the SurEau model. This will require the parameterization of an explicit 3D model of leaf temperature combining both leaf cooling by latent heat and leaf warming due to radiative heating. Finally, we will implement the SurEau model into ecosystem models of forest fluxes to predict the risk of tree mortality on French ICOS-Ecosystem sites and project the future risk of desiccation caused by hot droughts at stand scale under several climatic scenarios. A consortium of researchers associating ecophysiologists, biomolecularists, chemists, physicists and modelers will ensure the success of the project. The main contribution of our project will consist in advancing our understanding of plant resistance to hot-drought episodes. We will propose new key physiological traits, new phenotyping tools and putative genes that will help breeder in their screening of genotypes better adapted to future climatic conditions. Other final products of the project will consist of a list of species and beech and fir provenances potentially better performing under hot-drought conditions. Finally, we will also improve ecosystem models used to assess climate change impacts on forest by including desiccation processes related to hot drought.

The EU Biodiversity Strategy 2020 aims to establish green infrastructures and to restore at least 15% of degraded ecosystems until 2020. In this strategy, forests play a key role since they provide multiple ecosystem services. European policy has invested great efforts in afforestation of former farmlands but has largely neglected opportunities for passive landscape restoration and defragmentation by spontaneous forest establishment (SFE). Yet, SFE is common in many parts of Europe due to the widespread abandonment of agricultural land use in past decades. SFE typically leads to many small forest patches that are not or little managed. Together with existing semi-natural forests, these new forest patches form a network of habitats that can help maintain biodiversity and ecosystem services. Although SFE may contribute to the creation of multifunctional, diverse landscapes, it has so far received little attention from ecological and social science research. In fact, SFE is often regarded as a challenge rather than an opportunity for landscape management and conservation. SPONFOREST will examine the potential of SFE as a cost-effective and politically feasible tool for reinforcing perennial green infrastructures of self-sustaining forests in fragmented landscapes. In-depth ecological and sociological studies will advance the understanding of forest regeneration in the landscape context, analyse ecosystem services and disservices of new forest patches and assess their perception by stakeholders and the greater public. Five case studies in Mediterranean and temperate landscapes will use this approach to investigate SFE under various environmental and socio-political conditions. The ecological research in SPONFOREST will analyse SFE with a broad spectrum of complementary approaches including dendroecology, population genetics, functional ecology, remote sensing, and landscape analysis. State-of-the-art field and laboratory methods will be used to gather high-quality data that inform a mechanistic framework aimed at forecasting SFE as a function of tree biology and the landscape context. The social science research of SPONFOREST will combine standardized surveys and in-depth expert interviews with stakeholders and policy makers to elucidate the societal perception of these new forests, their current use and the ecosystem services they supply from a demand perspective, including governance options to regulate this supply. SPONFOREST will place great emphasis on a detailed synthesis of the insights gained from ecological and sociological research, and will actively involve policy makers and experts in the transdisciplinary evaluation of key findings in view of policy recommendations. The comprehensive but distinct key deliverables address the scientific community, policy makers, forest and landscape managers. SPONFOREST should thus contribute to strategies that optimize future forest governance and management at local to European scales.

Environmental gradients and patchiness shape the level and the distribution of genetic diversity of adaptive significance. A rich literature describes how gene frequencies vary along gradients and deals with the interaction of local selection and global gene flow. The implicit assumption is that populations evolving according to divergent selective pressures have to be sufficiently isolated for long term selective divergence to take place. Genetic pools differentiating along gradients and between patches store large amounts of genetic variability and thus favour the maintenance of a species’ adaptive potential. Measuring and modelling the amount of divergence among populations under divergent ecological conditions is therefore central to the prediction of how populations may respond to future local and global environmental changes. The turnover of genetic variants over gradients and habitat patches has traditionally been studied on geographical scales between regional and global. However, it can be argued that, for long-lived organisms, the large amounts of observed within-population genetic diversity may be maintained at least partly by local selection. In some instances, environmental conditions vary on a spatial scale that is comparable to, or even shorter than, average gene flow distances, thus setting the conditions for the interplay between selection and dispersal in the generation of observed patterns of diversity. Under these conditions, local differentiation among sub-populations for adaptive traits and genes may be maintained in spite of, and in some cases thanks to, ongoing gene flow. A combination of landscape and ecological genetic approaches makes it possible to investigate these phenomena, which have been shown to occur only for a limited set of plant species. Strategies to infer and model the strength of selection at target genes require prior knowledge of the loci under selection; however, current sequencing technology, combined with intensive field sampling of natural populations, can lead to the identification of selectively significant loci by population genomic approaches, even for non-model species. Providing genotypes for thousands of Single Nucleotide Polymorphisms (SNP) for hundreds of individuals has currently become feasible and relatively cheap. Quantitative genetics tools can now be successfully used to validate the association between SNP frequencies, phenotypic values, and environmental gradients. The time is therefore ripe for starting to address these major ecological-genetic questions directly in ecologically relevant species. The present project aims at investigating the complex genome-wide effects of local adaptation in nine keystone tree species from four major terrestrial ecosystems. Here we address the question of whether and how genetic diversity in forest tree stands, occurring across environmental gradients, is spatially structured by selective forces, and we propose to estimate the proportion of the genome undergoing such processes as well as to identify genes under selection. We propose an original use of classical quantitative genetic tools applied to in-situ progeny tests (reciprocal transplant experiment and replicated provenance tests) to validate SNP-trait-environment associations. The intensity of migratory and adaptive processes will be modelled thanks to advanced modelling strategies and will allow us to provide predictions of the response of forests to climate changes. We focus here on water availability gradients in particularly sensitive forest areas, such as the Guiana shield, the Mediterranean basin, Sub-Saharan Africa and the Brazilian Cerrado; all these regions suffer broad seasonal changes in soil water availability and are expected to undergo abrupt rainfall changes in the near future. Therefore, studying how forest tree populations cope with ecological gradients is at the fore among tools to predict the impact of expected future environmental changes.

Forests are a major reservoir of biodiversity and trees, as keystone organisms, directly impact the diversity and functioning of forest communities. Predicting the response of trees to ongoing global change (GC) is thus a critical scientific and societal issue. Along with phenotypic plasticity and migration, genetic adaptation is a central component of this response, particularly in trees whose high levels of diversity and long distance gene flow facilitates the spread of favorable genes. However, the existence of abundant genetic variation does not guarantee adaptation: if the climate and environmental changes are too quick, or genetic modifications are too slow, the population would go extinct before it can adapt to the new environmental challenges. Our hypothesis is that there is a critical level of genetic diversity for stress responses, which, together with the demographic impact of stress, predicts the likelihood of adaptation or extinction. The main goal of TipTree is to identify tipping points in the demographic and micro-evolutionary dynamics of tree populations, and to assess how human actions interfere in the adjustment between the rate of evolution and the velocity of GC. TipTree benefits from the BiodivERsA project LinkTree (2009-2012) which investigates the evolutionary response of key forest tree species to GC by analyzing the spatial variation of stress tolerance candidate genes along environmental gradients. But TipTree brings a new and critical dimension, that of time, by focusing on regeneration. In trees, regeneration (from fertilization to early plant recruitment) is a key period of the life cycle, when selection is expected to be very strong and has the potential to catalyze the rapid spread of evolutionary novelties in the next generation. The amount of genetic variation available in adults and how it is transmitted, selected and expressed in juveniles will condition the ecological properties of the whole ecosystem in the next decades to centuries, which remains a challenging short and non-equilibrium term of evolution for long-lived organisms. Specifically, our consortium will: 1) Screen the ecological and geographical margins of widespread keystone forest trees from different ecoregions (Temperate, Boreal, Mediterranean and Tropical) to identify where recent environmental changes have provoked shifts in allele frequencies at adaptive genes and to quantify these shifts by contrasting parent and offspring genetic and phenotypic compositions. We will address key environmental drivers: water stress, temperature regime, storm/fire frequency, pest outbreaks. Using natural and controlled (reciprocal transplants, common gardens) populations from existing Pan-European networks, we will generate large arrays of genomic polymorphisms using innovative genomic approaches, 2) Test the existence and evaluate the magnitude of tipping points for tree population dynamics at micro-evolutionary scales, by using a new generation of models coupling biophysics, population dynamics and quantitative genetics. We will feed these models with (i) climate change scenarios provided by IPCC, (ii) forest management scenarios established by our stakeholder group and (iii) our experimental results on adaptive genetic diversity. Micro-evolution of tree populations will be simulated at local and regional scales, and will provide forecasts of ecosystem services (carbon budget and water balance) and decision support for management.