Despite deep learning being state of the art for data-driven model predictions, its application in ecology is currently subject to two important constraints: (i) deep-learning methods are powerful in data-rich regimes, but in ecology data are typically sparse; and (ii) deep-learning models are black-box methods and inferring the processes they represent are non-trivial to elicit. Process-based (= mechanistic) models are not constrained by data sparsity or unclear processes and are thus important for building up our ecological knowledge and transfer to applications. In this work, we combine process-based models and neural networks into process-informed neural networks (PINNs), which incorporate the process knowledge directly into the neural network structure. In a systematic evaluation of spatial and temporal prediction tasks for C-fluxes in temperate forests, we show the ability of five different types of PINNs (i) to outperform process-based models and neural networks, especially in data-sparse regimes with high-transfer task and (ii) to inform on mis- or undetected processes.
Understanding the response of marine organisms to temperature is crucial for predicting climate change impacts. Fundamental physiological thermal performance curves (TPCs), determined under controlled conditions, are commonly used to project future species spatial distributions or physiological performances. Yet, real-world performances may deviate due to extrinsic factors covarying with temperature (food, oxygen, etc.). Using a bioenergetic marine ecosystem model, we evaluate the differences between fundamental and realised TPCs for fish species with contrasted ecology and thermal preferences. Food limitation is the primary cause of differences, decreasing throughout ontogeny and across trophic levels due to spatio-temporal variability of low-trophic level prey availability with temperature. Deoxygenation has moderate impact, despite increasing during ontogeny. This highlights the lower sensitivity of early life stages to hypoxia, which is mechanistically explained by lower mass-specific ingestion at older stages. Understanding the emergence of realised thermal niches offers crucial insights to better determine population's persistence under climate warming.
Pollen limitation has a considerable influence on forest masting, the highly variable and synchronised seed production, on which forest regeneration and ecosystem dynamics largely rely. Depending on the various mechanisms possibly involved in pollen limitation, the consequences of climate change on masting could be very different. These mechanisms were investigated in 10 oak populations along a climatic gradient using surveys of airborne pollen and fruiting rate as a proxy of pollen limitation. We found no support for the widely accepted hypothesis of the intra-annual synchrony of flower phenology when considered in isolation. Instead, the fruiting rate was largely explained by a combination of intra-annual flower phenology synchrony, annual investment in flowering and the effects of weather on pollen maturation and diffusion. These findings highlight the need for a cohesive theoretical framework for pollen limitation to accurately predict the impact of climate change on oak-dominated ecosystems.
Understanding the role of climate in the assembly of rainforest tree communities is informative for predicting how future climates will impact species and communities. We surveyed rainforest tree communities across the Australian subtropics (spanning 600 to 2500 mm rainfall year−1) and measured functional traits on 285 (91%) of all recorded species. We used principal component analysis to create axes approximating species' hydraulic strategies, leaf economics and stature and included these as predictors in joint species distribution models, along with traits describing dispersal ability and leaf phenology. Hydraulic strategy and leaf phenology strongly modulated species' occurrence trends along the moisture availability gradient, while stature and leaf economics modulated species' responses to minimum temperature and soil variables, respectively. Overall, we quantify the occurrence trends of almost half of Australia's subtropical rainforest tree species based on their functional traits, providing a general foundation for prediction under ongoing climate change.
Elements are the basic substances that make up living organisms, and the element composition in plants quantitatively reflect the adaptation of plants to environment. However, the drivers that constitute the species-specific plant elementome, as well as the bivariate bioelemental correlations in determining the stability of different bioelements are yet unclear. Based on 1058 leaf observations of 84 plant species from 232 wetlands across large environmental gradients, we found that bioelements with higher concentration were more stable and evolutionary constrained. We proposed a stability of well-coordinated elements hypothesis, suggesting that bioelements that coordinate well in driving certain physiological functions constrain each other, thus maintaining relatively stable ratios in plants. In contrast, those functionally independent bioelements fluctuate greatly with environmental nutrient availability. Cold and saline stresses decreased plant stoichiometric network connectivity, complexity, and stability. Our research filled the gap in study of wetland plant elementome, and provided new evidences of plant–environment interactions in regions sensitive to climate change.
Commonly used two-sex discrete-time population projection models rely on mating functions developed for continuous-time frameworks that overestimate the number of unions between reproductive individuals. This has important consequences for our understanding of the evolution and demography of two-sex populations and consequently for management and conservation. Here, we propose a novel mating function that is robust by obeying all properties necessary to be ecologically valid and flexible by accommodating all mating systems and efficiency in mating encounters. We illustrate the usefulness of this novel function with an application to the sexually size-dimorphic and polygynous wild boar (Sus scrofa). We show that the population growth rate depends on the harem size, the operational sex ratio, and the mating efficiency. This novel function can be applied to all mating systems and tactics and is highly relevant in the context of global changes under which mating systems and mating efficiency are expected to change.
Soil microorganisms are crucial in terrestrial ecosystems, influencing carbon (C) sequestration, yet their metabolic activities are often constrained by nitrogen (N) and phosphorus (P) availability. Despite this, a global understanding of microbial nutrient limitation remains elusive. We synthesised 1245 observations from 225 articles to elucidate patterns and factors of microbial nutrient limitation. Contrary to convention, soil microbial P limitation is widespread (83.78% of observations), with N limitation mainly in temperate zones and pronounced P limitation in tropical and cold zones. Soil microbial P limitation correlates positively with mean annual precipitation and clay content, while N limitation correlates negatively with soil pH. Importantly, microbial nutrient limitation directly affects C cycling, as microbial C limitation increases with decreasing N or P limitation. This underscores the significance of microbial nutrient limitation in terrestrial C cycling and the need to incorporate it into Earth system models for accurate predictions under changing conditions.
Island biogeography theory provides key insights into biodiversity patterns across islands species–area relationships and conservation. However, classical island biogeography theory assumes that species are ecologically equivalent in terms of their dispersal ability. We evaluated the role of a key trait (hand-wing index, a proxy for dispersal ability in birds) in shaping species-area relationships of avifauna spanning 6706 species on 3894 islands. High community-weighted mean (CWM) dispersal ability in regional species pools had widespread but context-dependent effects on island species-area relationships. Among island archipelagos at smaller spatial extents, high CWM dispersal ability was associated with steeper species-area relationships. Among zoogeographical realms at larger spatial extents high CWM dispersal ability was associated with shallower species-area relationships and higher local species richness on small islands. Our study reveals that geographic variation in species' dispersal traits has strong effects on island species-area relationships and likely plays an important role in non-neutral community assembly.
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：