Testing the reliability of an AI-based large language model to extract ecological information from the scientific literature.

Artificial intelligence-based large language models (LLMs) have the potential to substantially improve the efficiency and scale of ecological research, but their propensity for delivering incorrect information raises significant concern about their usefulness in their current state. Here, we formally test how quickly and accurately an LLM performs in comparison to a human reviewer when tasked with extracting various types of ecological data from the scientific literature. We found the LLM was able to extract relevant data over 50 times faster than the reviewer and had very high accuracy (>90%) in extracting discrete and categorical data, but it performed poorly when extracting certain quantitative data. Our case study shows that LLMs offer great potential for generating large ecological databases at unprecedented speed and scale, but additional quality assurance steps are required to ensure data integrity.

Identifying the generalizable controls on insect associations of native and non-native trees.

Trees growing outside their native geographic ranges often exhibit exceptional growth and survival due in part to the lack of co-evolved natural enemies that may limit their spread and suppress population growth. While most non-native trees tend to accumulate natural enemies over time, it remains uncertain which host and insect characteristics affect these novel associations and whether novel associations follow patterns of assembly similar to those of native hosts. Here, we used a dataset of insect-host tree associations in Europe to model which native insect species are paired with which native tree species, and then tested the model on its ability to predict which native insects are paired with which non-native trees. We show that native and non-native tree species closely related to known hosts are more likely to be hosts themselves, but that native host geographic range size, insect feeding guild, and sampling effort similarly affect insect associations. Our model had a strong ability to predict which insect species utilize non-native trees as hosts, but evolutionarily isolated tree species posed the greatest challenge to the model. These results demonstrate that insect-host associations can be reliably predicted, regardless of whether insect and host trees have co-evolved, and provide a framework for predicting future pest threats using a select number of easily attainable tree and insect characteristics.

Evolutionary history of host trees amplifies the dilution effect of biodiversity on forest pests.

Biodiversity appears to strongly suppress pathogens and pests in many plant and animal systems. However, this "dilution effect" is not consistently detected, and when present can vary strikingly in magnitude. Here, we use forest inventory data from over 25,000 plots (>1.1 million sampled trees) to quantify the strength of the dilution effect on dozens of forest pests and clarify why some pests are particularly sensitive to biodiversity. Using Bayesian hierarchical models, we show that pest prevalence is frequently lower in highly diverse forests, but there is considerable variability in the magnitude of this dilution effect among pests. The strength of dilution was not closely associated with host specialization or pest nativity. Instead, pest prevalence was lower in forests where co-occurring tree species were more distantly related to a pest's preferred hosts. Our analyses indicate that host evolutionary history and forest composition are key to understanding how species diversity may dilute the impacts of tree pests, with important implications for predicting how future biodiversity change may affect the spread and distribution of damaging forest pests.

A global analysis of tree pests and emerging pest threats.

SignificanceThe introduction of trees outside their native ranges has greatly expanded the potential ranges of their pathogens and insect pests, which risk spilling over and impacting native flora. However, we often lack a strong understanding of the host, climatic, and geographic factors that allow pests to establish outside their hosts' native ranges. Using global datasets of pest occurrences and the native and nonnative ranges of tree hosts, we show there are strong generalizable trends controlling pest occurrences and can predict the occurrence of pests outside their hosts' native ranges with >75% accuracy. Our modeling framework offers a powerful tool to identify future invasive pest species and the ecological mechanisms controlling the accumulation of pests outside their hosts' native ranges.

Host phylogenetic diversity predicts the global extent and composition of tree pests.

Tree pests cause billions of dollars of damage annually; yet, we know little about what limits their regional composition and distribution. Here, we model the co-occurrence of 4510 pests and 981 tree host genera spread across 233 countries. We show the composition of tree pests is primarily driven by the phylogenetic composition of host trees, whereas effects of climate and geography tend to be more minor. Pests that utilise many hosts tend to be more widespread; however, most pests do not fill the geographic range of their hosts-indicating that many pests could expand their extents if able to overcome barriers limiting their current distribution. Our results suggest that the establishment of pests in new regions may be largely dictated by the presence of suitable host trees, but more work is needed to fully understand the influences climate has on the distributions of individual pest species.

Towards a phylogenetic ecology of plant pests and pathogens.

Plant-pathogens and insect pests, hereafter pests, play an important role in structuring ecological communities, yet both native and introduced pests impose significant pressure on wild and managed systems, and pose a threat to food security. Global changes in climate and land use, and transportation of plants and pests around the globe are likely to further increase the range, frequency and severity of pest outbreaks in the future. Thus, there is a critical need to expand on current ecological theory to address these challenges. Here, we outline a phylogenetic framework for the study of plant and pest interactions. In plants, a growing body of work has suggested that evolutionary relatedness, phylogeny, strongly structures plant-pest associations-from pest host breadths and impacts, to their establishment and spread in new regions. Understanding the phylogenetic dimensions of plant-pest associations will help to inform models of invasive species spread, disease and pest risk in crops, and emerging pest outbreaks in native plant communities-which will have important implications for protecting food security and biodiversity into the future. This article is part of the theme issue 'Infectious disease macroecology: parasite diversity and dynamics across the globe'.

Sequence of flower and leaf emergence in deciduous trees is linked to ecological traits, phylogenetics, and climate.

While much research has focused on the timing of individual plant phenological events, the sequence of phenological events has received considerably less attention. Here we identify drivers and patterns of flower and leaf emergence sequence (FLS) in deciduous tree species of the Great Lakes region of North America. Five hypotheses related to cold tolerance, water dynamics, seed mass, pollination syndrome, and xylem anatomy type were compared for their ability to explain FLS. Phylogenetic and geographic patterns of FLS were also assessed. We identified additional traits associated with FLS using Random Forest models. Of the hypotheses assessed, those related to species' water dynamics and seed mass had the greatest support. The spatial pattern of FLS was found to be strongly related to minimum monthly temperature and the phylogenetic pattern was clustered among species. Based on results from Random Forest models, species' fruiting characteristics were found to be the most important variables in explaining FLS. Our results show that FLS is related to a suite of plant traits and environmental tolerances. We emphasize the need to expand phenological research to include both the timing and sequence of plant's entire phenology, in particular in relation to plant physiology and global change.

Dissemination of Goss's Wilt of Corn and Epiphytic Clavibacter michiganensis subsp. nebraskensis from Inoculum Point Sources.

Goss's wilt of corn, caused by Clavibacter michiganensis subsp. nebraskensis, has reemerged since 2006 as an economically important disease of corn in in the Midwestern United States. In 2012 and 2013, field plot studies were conducted with a pathogenic, rifampicin-resistant C. michiganensis subsp. nebraskensis isolate and a Goss's wilt-susceptible corn hybrid to monitor epiphytic C. michiganensis subsp. nebraskensis population densities and the temporal and spatial spread of Goss's wilt incidence originating from inoculum point sources. The randomized complete block trial included three treatments: noninoculated control, inoculum point sources established by wound inoculation, and inoculum point sources consisting of C. michiganensis subsp. nebraskensis-infested corn residue. Epiphytic C. michiganensis subsp. nebraskensis was detected on asymptomatic corn leaves collected up to 2.5 m away from inoculum sources at 15 days after inoculation in both years. The percentage of asymptomatic leaf samples on which epiphytic C. michiganensis subsp. nebraskensis was detected increased until mid-August in both years, and reached 90, 55, and 35% in wound-, residue-, and noninoculated plots, respectively, in 2012; and 50, 11, and 2%, respectively, in 2013. Although both growing seasons were drier than normal, Goss's wilt incidence increased over time and space from all C. michiganensis subsp. nebraskensis point sources. Plots infested with C. michiganensis subsp. nebraskensis residue had final Goss's wilt incidence of 7.5 and 1.8% in 2012 and 2013, respectively; plots with a wound-inoculated source had final Goss's wilt incidence of 16.6 and 14.0% in 2012 and 2013, respectively. Our findings suggest that relatively recent outbreaks of Goss's wilt in new regions of the United States may be the result of a gradual, nondetected buildup of C. michiganensis subsp. nebraskensis inoculum in fields.