Search

The vast volume of currently available unstructured text data, such as research papers, news, and technical report data, shows great potential for ecological research. However, manual processing of such data is labour-intensive, posing a significant challenge. In this study, we aimed to assess the application of three state-of-the-art prompt-based large language models (LLMs), GPT-3.5, GPT-4, and LLaMA-2-70B, to automate the identification, interpretation, extraction, and structuring of relevant ecological information from unstructured textual sources. We focused on species distribution data from two sources: news outlets and research papers. We assessed the LLMs for four key tasks: classification of documents with species distribution data, identification of regions where species are recorded, generation of geographical coordinates for these regions, and supply of results in a structured format. GPT-4 consistently outperformed the other models, demonstrating a high capacity to interpret textual data and extract relevant information, with the percentage of correct outputs often exceeding 90% (average accuracy across tasks: 87–100%). Its performance also depended on the data source type and task, with better results achieved with news reports, in the identification of regions with species reports and presentation of structured output. Its predecessor, GPT-3.5, exhibited slightly lower accuracy across all tasks and data sources (average accuracy across tasks: 81–97%), whereas LLaMA-2-70B showed the worst performance (37–73%). These results demonstrate the potential benefit of integrating prompt-based LLMs into ecological data assimilation workflows as essential tools to efficiently process large volumes of textual data. • Large Language Models (LLMs) are tested for extracting biodiversity data from text. • Tasks include identification of relevant sources, geocoding, and data structuring. • LLMs, especially GPT-4, show over 90% accuracy on multiple tasks. • LLMs promise effective data assimilation from unstructured text in ecology. [ABSTRACT FROM AUTHOR]

The Common Agricultural Policy (CAP) has shown difficulties in meeting its environmental objectives, namely in supporting biodiversity-friendly farming systems that remain under pressure to intensify or abandon. Proposals to address this have ranged from increasing the focus on highly tailored and targeted agri-environment schemes, to promoting broad-brush policies such as those recently implemented in the Greening of the CAP. Both options have been criticised due to questionable cost-effectiveness. Alternatives based on agri-environment policies oriented to support conservation-relevant farming systems have been suggested, but they have faced operational difficulties related primarily to obtaining the necessary data to define farming system typologies. Here we investigated whether a simplified approach based on a coarse farming system typology built from incomplete data on land-use and livestock, such as that available in CAP paying agencies, could be used to infer on a wider range of conservation-relevant farm management practices and, ultimately, to select the farming systems qualifying for premium payments. Based on data collected by a farm-survey on a High Nature Value farmland area in southern Portugal, we show that some farming systems are consistently associated with conservation-relevant practices related to the use of herbicides, stubble grazing, creation of wildlife plots and early cereal harvest. The traditional system involving the rotational production of cereals and sheep grazing on fallows showed the most favourable balance of land uses and farm management practices with positive conservation effects. Results underlined the potential of farming systems as a framework for developing agri-environment policy. [ABSTRACT FROM AUTHOR]

Abstract: In the Euro-Mediterranean region, mechanical fuel reduction is increasingly used in response to the mounting occurrence of catastrophic wildfires, yet their long-term ecological effects are poorly understood. Although Mediterranean vegetation is resilient to a range of disturbances, it is possible that widespread fuel management at short intervals may threaten forest structural complexity and the persistence of some plant species and functional types, with overall negative consequences for biodiversity. We used a chronosequence approach to infer woody vegetation changes in the first 70 years after understory clearing in upland cork oak (Quercus suber) forests, and to assess how these are affected by treatment frequency. Across the chronosequence there was a shift between plant communities with contrasting composition, structure and functional organization. Understory cover increased quickly after disturbance and a community dominated by pioneer seeder and dry-fruited shrubs (Cistus ladanifer, C. populifolius, Genista triacanthos, and Lavandula stoechas) developed during about 15 years, but this was slowly replaced by a community dominated by resprouters and fleshy-fruited species (Arbutus unedo, Erica arborea) >40 years after disturbance. During the first 15 years there were rapid increases in woody species richness, vertical structural diversity, cover by Q. suber juveniles and saplings, and shrub cover at <1.5m strata, which levelled off or slightly declined thereafter. In contrast, tree species richness, tree density and density of arboreal A. unedo and E. arborea, vertical structural evenness, and cover at >1.5m strata increased slowly for >50 years. Treatment frequency showed strongly negative relationships with species richness, structural diversity and evenness, and horizontal and vertical understory cover, particularly that of slowly recovering species. These findings suggest that fuel reduction programs involving widespread and recurrent understory clearing may lead to the elimination at the landscape scale of stands with complex multi-layered understory occupied by large resprouters and fleshy-fruited species, which take a long time to recover after disturbance. Fuel management programs thus need to balance the dual goals of fire hazard reduction and biodiversity conservation, recognizing the value of stands untreated for >50 years to retain ecological heterogeneity in Mediterranean forest landscapes. [Copyright &y& Elsevier]

Abstract: Afforestation of agricultural land is increasingly used to deliver environmental benefits, but their effects on biodiversity remain poorly understood. This paper tests the hypothesis that afforestation changes predation processes in surrounding farmland, examining how the characteristics and landscape context of forest plantations affect predator (birds and mammalian carnivores) and key prey (rabbits and hares) abundances, and bird nest predation rates in Iberian cereal-steppes. Lagomorphs and predators were surveyed in fallow fields around 50 forest plantations, where predation rates were estimated using artificial nests set at 0, 100, 200 and 300m from the forest edge. Recent plantations structurally similar to sparse (oak) or dense (pine) shrublands were associated with the highest hare and rabbit abundances, respectively, whereas both species avoided landscapes with high eucalyptus cover. In contrast, mature eucalyptus plantations showed strong positive effects on typical nest predators such as corvids and carnivores. Open farmland fragmentation favoured the abundance of lagomorphs and carnivores. Despite these effects and the high predation rate on artificial nests (49%), there was neither evidence for increased predation near plantation edges nor higher predation in fields with more lagomorphs and predators. However, predation tended to increase with cover by young oak plantations and overall forest plantation cover, to decrease with eucalyptus cover at both the local and landscape scales, and to peak in landscapes with intermediate edge densities. These results suggest that afforestation may have strong effects on bird nest predation rates by changing landscape composition and configuration, rather than by inducing local increases in predator and prey populations. Nevertheless, increased abundances of generalist predators associated with forest plantations may still be considered of conservation concern, thus supporting the recommendation for strongly restricting afforestation in areas important for open grassland birds. Where this is unavoidable, monitoring should be undertaken to provide early signals for bird population declines associated with predator increases, eventually triggering conservation action such as predator exclusion or removal. [Copyright &y& Elsevier]

[Display omitted] • Afforestation of previously unforested land impacts grassland biodiversity. • We used camera traps to study the effects of different plantations on nest predation. • Common crow was the main predator predating ca. half of the artificial quail nests. • Nest predation risk and predator abundance was higher near eucalyptus plantations. • Slow-growing native trees (oak and pine) have lower impacts on grassland birds. Large scale afforestation (i.e., establishment of forests on farmland, grassland and other land not previously forested) is increasingly regarded as a cost-effective option to mitigate climate change by promoting carbon sequestration. However, this strategy can have negative biodiversity impacts, potentially causing the loss and fragmentation of open habitats of conservation value, as well as edge effects that can extend well beyond forest boundaries. Mitigating impacts of afforestation programs thus requires a detailed understanding of the mechanisms whereby they affect biodiversity, particularly where new plantations are established close to areas important for open habitat species. Here, we examined how afforestation with oak, pine or eucalyptus trees, commonly planted in Mediterranean Europe, may have distinct effects on nest predation risk of grassland birds in southern Portugal. We used artificial nests with quail eggs and camera traps to: (i) identify nest predators and estimate predation risk of ground-nesting birds in grasslands adjacent to oak, pine and eucalyptus plantations across a distance gradient (0 m to 300 m) from the plantation edge; and (ii) examine how predation risk was influenced by avian predator abundance, sward height and vegetation clearing for photo-monitoring (nest-site manipulation), and their interactions with plantation type. Carrion crow was by far the main egg predator identified, followed by Montagu's harrier, raven and magpie. Overall, nest predation risk and predator abundance were higher around eucalyptus plantations, compared to either pine or oak plantations. Predation decreased with increasing sward height, but this effect was less noticeable when vegetation around nests was cleared. In pine and oak plantations, predation risk was slightly higher close to edges, and decreased with distance from the edge. Conversely, in eucalyptus plantations predation was lower close to edges, and increased with distance to edges. Distance effects around eucalyptus plantation were, however, absent in manipulated nests. Overall, results show that afforestation with different tree species has distinct impacts on predation risk of artificial nests in Mediterranean farmland. This is probably a consequence of an increasing structural complexity of the plantations, from oak to eucalyptus, driven by differences in tree growth rates and undergrowth cover. These findings suggest that farmland afforestation should favour agro-forest plantations with slow growing species (oak and pine) to mitigate negative effects on breeding grassland birds in adjacent open habitats. [ABSTRACT FROM AUTHOR]

• The perception, attitudes and behaviours of stakeholders determine the effectiveness of IAS prevention and management. • The effective prevention and management of IAS is depends on long-term funding of scientific projects and management actions, which could guarantee a constant monitoring by trained staff. • We recommend involving managers and citizens; using updated information to inform decisions; funding long-term projects; invest in restoration, and improve legislation. Invasive alien species are considered a major threat to biodiversity and seriously impact the economy and public health of recipient regions. Identifying the determinants of protected areas' vulnerability to the establishment of invasive alien species (IAS) may improve management, prevention and early warning systems of biological invasions. We elicited the perceptions of 106 experts working in 56 protected areas located in 6 European countries, to gather insights on the determinants of ecosystem vulnerability to IAS. We implemented an online questionnaire to understand the following: i) the expectations and perceived drivers of the future entry of IAS to protected areas, ii) the most common policy frameworks and management options used, and iii) the main challenges encountered when managing or preventing biological invasions. Experts project an increased entrance of IAS in protected areas located in Corsica, Portugal, and Spain. Areas where the entry of IAS is expected to increase comprise i) low to average anthropogenic disturbance, ii) average biodiversity richness, with endemic/threatened/protected species, iii) current presence of IAS, and iv) anthropogenic activities in surrounding regions. Respondents expecting increased entrance of IAS in protected areas also consider corridors as the main pathway of their entry, whereas those expecting IAS entrance to decrease believe IAS enter protected areas mainly through intentional introductions. Our results also underline the complexity of IAS management, which is often hampered by the lack of clear policies, shortage of economic resources and specialised staff, and poor awareness of biological invasions. Future research should further strive to understand which areas are more vulnerable to IAS, especially in regions with free trade and movement agreements, such as the EU. Meanwhile, vulnerability to IAS can be minimized through a stricter management of both protected areas and surrounding anthropogenic activities, and by raising manager and citizen awareness of the ecological and economic problems caused by IAS. [ABSTRACT FROM AUTHOR]