Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach
... To analyse the impact of human activities on the behaviour and breeding success of Bonelli's eagles, we ran a set of generalized linear mixed models (GLMM). We applied an information theoretical approach using Akaike's information criterion (AIC; Burnham & Anderson, 2002). Models were ranked using AICc differences (Δi) and Akaike weights (wi). ...
... Δi was calculated as the difference between the AICc of each model and the AICc of the best model. Models with Δi < 2 can be considered alternative models to the best one (Burnham & Anderson, 2002). We computed all of our models and fitted them using the maximum likelihood method with the Laplace approximation, using the 'glmer' function in the 'lme4' package (Bates et al., 2015) in R v.4.3.3. ...
... In the first model selection (AIC; Burnham & Anderson, 2002), we analysed whether eagle behaviour was affected by disturbing activities; the characteristics of the territory and the birds' reaction using logistic binary regression models, with a 'probability of reaction' as the response variable (one data point for each activity detected), using a binomial distribution and a logit link function. ...
There has been increasing recognition of the adverse effects of outdoor recreation on the behavioural ecology, breeding success and ultimate abandonment of territories in several raptor species, such as the Bonelli's Eagle (Aquila fasciata).
We assessed the effects of human activities on the behaviour and breeding success of Bonelli's eagles breeding in south‐eastern Spain, using a 14‐year long‐term data set.
Our results showed a weak behavioural response to human activities. The probability of eagle reaction varied among the human activities assessed and increased when they occurred close to a nest, being most marked for ecotourism activities. In general, the probability of a bird reacting increased sharply when an activity occurred closer than 480 m from the nest, but reactions were negligible when an activity occurred 1000 m away. Ecotourists only affected Bonelli's eagle behaviour at distances <300 m from the nest.
We found a negative relationship between human activity and Bonelli's eagle behaviour: the probability of a reaction decreased with the increasing frequency of human activities, suggesting a degree of eagle habituation. Breeding success did not vary with a pair's experience of disturbance, nor between the periods before, during and after the COVID‐19 lockdown but did increase with the frequency of human activities.
Our results indicated that Bonelli's eagles can develop a degree of tolerance to human activities during the nesting season. This behavioural plasticity offers the opportunity to harmonize anthropogenic activities with the application of specific conservation measures for this species in humanized landscapes. Our findings led to suggestions to minimize the detrimental effects of human activities and to optimize the conservation management of Bonelli's eagles in Southeastern Spain. More generally, our study emphasises the value of behavioural ecology studies for guiding conservation.
... We included an auto-regressive correlation structure of order three to account for auto-correlation in the residuals (lower or higher auto-regressive orders led to a higher Akaike information criterion or AIC; Burnham & Anderson, 2002;Zuur et al., 2009). We also adjusted the variance structure to account for residual F I G U R E 1 Legend on next page. ...
... This component assumes that the inter-island variation around the intercept α is normally distributed (with mean 0 and variance D). ε represents the (normally distributed) errors and subscript i the island. For both the nest number and clutch size analyses, we considered different models characterized by different fixed effect structures and adopted a model comparison approach based on the AIC (Burnham & Anderson, 2002) to determine whether the number of nests or average clutch size remained constant or varied over time, and whether they were affected by changes in spring sea ice. We considered both linear and quadratic shapes for the long-term trend in these variables. ...
... We also considered the interaction with island category to test whether the trends in the number of eider nests, or in the average clutch size, and the sea ice effect varied between early and late ice-free islands. If the difference in AIC values between two models was <2, the models had equal statistical support, and in the case of nested models, the simplest one (i.e., with lowest df) was preferred according to the principle of parsimony (Burnham & Anderson, 2002). For this model selection, following Zuur et al.'s (2009) approach for comparing models with the same random effect structure but different fixed effects, we fitted models using maximum likelihood (ML). ...
The Arctic is warming four times faster than any other region on Earth, leading to a dramatic reduction in sea ice. Even though sea ice plays a key role in the ecology of many Arctic species, few studies have assessed the consequences of its disappearance on the dynamics of Arctic wildlife populations. Moreover, the potential intra‐population variations in such effects remain largely overlooked. Here, using a 40‐year time series, we assessed how sea ice changes in a High Arctic fjord affected the population dynamics of common eiders Somateria mollissima via changes in their fine‐scale breeding distribution and how these effects varied among breeding sites. More specifically, some islands within the fjord used to be connected by landfast ice to the shore most of the spring and thus to be accessible to one of the main eider predators, the Arctic fox Vulpes lagopus. Following the disappearance of spring sea ice in the fjord, these islands recently became disconnected much earlier in the season and thus inaccessible to foxes. We tested the prediction that these islands now represent favorable nesting grounds for common eiders and that the breeding eider populations on these islands increased following the sea ice retreat. Our results support our prediction and the role played by fox predation in mediating the sea ice effects. Even though the overall eider population in the fjord has slightly declined in the last decades, the recent sea ice reduction has led to a rapid colonization of newly available breeding habitats and to an increasing number of breeding eiders there. Inter‐annual changes in sea ice did not significantly affect the number of eiders breeding on the islands that were historically isolated from fox predation. Ignoring such intra‐population variation between breeding sites in predation risk masks the effects of sea ice reduction on eider dynamics. Our study illustrates the complex and fine‐scale effects of sea ice disappearance on Arctic wildlife and the potential importance of predation in mediating these effects.
... Muff et al. (2022) explains that describing results using a more gradual notion of evidence is a more nuanced approach to reporting statistics than using strict p-values. Prior to model fitting we tested the effect the detection covariates had on total bat activity using Akaike's Information Criterion scores with a correction for small sample size (AICc) (Burnham and Anderson, 2002). Model testing showed that daily rainfall and nightly temperature both had strong impacts on total bat activity, and thus were included in all further models as covariates. ...
... We tested five tree size thresholds: ≥40 cm DBH, ≥50 cm DBH, ≥60 cm DBH, ≥70 cm DBH and ≥ 80 cm DBH. We selected the most parsimonious model for each community metric using AICc with a correction for small sample size (Burnham and Anderson, 2002) (Table S1). The model formula for each model set followed the following formula: Community_Metric ~ Tree_Size_Threshold + Region + Temperature + Daily_Rainfall. ...
... For each metric we employed multi-model inference to compare five models incorporating region, daily rainfall, temperature, the tree size threshold identified in the previous step, and woody vegetation cover. We selected the most parsimonious model for each community metric using AICc scores with a correction for small sample size (Burnham and Anderson, 2002) (Table S2). Where the interaction model (i.e., model e) was the topranked model, we plotted the interactive effect in Fig. 2. We plotted the effect sizes for top-ranked conditional models in Fig. S1 and for topranked hurdle models in Fig. S2. ...
Large trees are keystone structures for a range of taxa. Insectivorous bats depend on large trees for roosting habitat, such as hollows and peeling bark, as well as habitat for their prey. However, large trees are in decline in urban areas globally. We sought to determine if the richness and activity of insectivorous bats in urban environments is associated with the occurrence of large trees; and in which urban landscape context large trees have greatest benefit for insectivorous bats. Using ultrasonic bat detectors set at 83 sites spanning temperate and subtropical cities in eastern Australia, we identified 20,026 bat passes from 16 microbat species and four species complexes. We found strong positive associations between four bat metrics (mean nightly activity, richness and activity of edge-adapted bats and vespertilionid bat activity) and the number of large trees ≥50 cm DBH. We also found evidence that large trees supported a higher richness of edge-adapted bats in areas with lower woody vegetation cover. Our data indicate that the value of large trees for edge-adapted bats is enhanced when large trees are isolated-a relationship previously demonstrated for birds but not bats. Large trees in urban greenspace, especially trees in isolation, offer valuable habitat that supports a substantial community of insectivorous bats. Our results highlight the importance of retaining large, isolated trees (both native and non-native) in urban greenspace for bat conservation.
... At times, researchers may want to predict unobserved outcomes from new input data by using training data to optimize parameter estimation such that a set of input features predict output values that most closely match observed data output values in verification data ("Prediction-Based Model Selection and Forecasting" in Figure 1). Predictive studies rely on procedures that emphasize model evaluation and selection through predictive performance, including model averaging that derives inferences from several plausible models (i.e., multi-model inference; Burnham and Anderson, 2010). Results from models selected for high prediction accuracy are often believed to produce more meaningful parameter estimates for inference than models with low prediction accuracy (Harrison et al., 2018), which has spurred the popularity of machine learning approaches touted to provide "data-driven" understandings of complex natural processes (Christin et al., 2019;Olden et al., 2008). ...
... Proper model specification is crucial for valid causal conclusions (Burnham and Anderson, 2010), thus more attention must be invested in the process of designing studies and building models using pre-existing knowledge to make causal claims from experimental and observational studies. To formalizing pre-existing knowledge in causal analysis, researchers may use two widely used tools: directed acyclic graphs (DAGs) and thought experiments based on ideal randomized controlled trials (RCTs). ...
This work presents a conceptual synthesis of causal discovery and inference frameworks, with a focus on how foundational assumptions -- causal sufficiency, causal faithfulness, and the causal Markov condition -- are formalized and operationalized across methodological traditions. Through structured tables and comparative summaries, I map core assumptions, tasks, and analytical choices from multiple causal frameworks, highlighting their connections and differences. The synthesis provides practical guidance for researchers designing causal studies, especially in settings where observational or experimental constraints challenge standard approaches. This guide spans all phases of causal analysis, including question formulation, formalization of background knowledge, selection of appropriate frameworks, choice of study design or algorithm, and interpretation. It is intended as a tool to support rigorous causal reasoning across diverse empirical domains.
... Model selection was based on information criteria and backward elimination of variables (Burnham and Anderson 2002). ...
... Competing models were constructed beginning with the full set and removing variables one by one. Models were compared by AIC value, and a variable was removed when the difference in AIC was lower than 2 (Burnham and Anderson 2002). ...
In the last decades, the population of the squat lobster Munida gregaria inhabiting the water column in the Patagonian Sea, Argentina, has expanded. This crustacean shares its habitat with the Argentine anchovy Engraulis anchoita , a key species in the marine food web. Existing information suggests that M. gregaria may compete with anchovy causing spatial segregation or interfere with the formation of fish schools. This study examines anchovy schools and squat lobster swarms sharing the water column in San Matías and Nuevo gulfs, in northern Patagonia. A total of four acoustic zigzag surveys were performed in San Matías gulf and five in Nuevo gulf, covering all the seasons. The echoes of both species were characterized and discriminated, measuring a set of shape, location, and density properties. A linear model was built to determine which factors affect anchovy school morphology. Anchovy schools presented well‐defined edges, were amoeboid in shape, and occupied a wide range of depths in the water column. M. gregaria formed large and irregular aggregations, mainly in the upper half water column. Anchovy schools that interacted with M. gregaria tended to be larger and less dense, especially in spring and summer. This may result from either physical interference between individuals, leading to greater laxity in anchovy schools, or the partial dispersal of anchovies feeding on juvenile M. gregaria .
... We additionally fitted models to the base variables to compare those fits to models, including climate and weather variables. Thereafter, we performed a backward selection of the full models based on the Akaike Information Criteria (AIC; Burnham and Anderson, 2002) to remove uninformative covariates. We fitted the models with the "glmmTMB" package (Brooks et al., 2017) in R (Core Team, 2022). ...
... In contrast to the discrete probability distributions used for the quantity of tree recruits, variables linked to the sampling design were removed due to perfect separation between the species groups which would lead to a violation of MLRM model assumptions. Uninformative covariates were removed based on the AIC (Burnham and Anderson, 2002). The models were fitted using the function "multinom" from the "nnet" package (Venables and Ripley, 2002). ...
Tree recruitment forms an essential process in forest growth models as it determines the amount and composition of the next generation of trees and, hence, the provision of forest ecosystem services over long time spans. With global change and the hereby associated changes in environmental conditions and forest management adaptations, the common static tree recruitment modelling approaches have become largely obsolete and necessitated the development of more dynamic models. Limited by the availability of data for the parameterisation of tree recruitment processes, such models have only been developed for single species or national frameworks and largely failed to detect climatic influences. In this study, we developed a dynamic tree recruitment model for Europe, utilising National Forest Inventory data from 8 countries with more than 95,000 repeated plot observations and nearly 138,000 individual tree recruitment events. We investigated the effect of forest management, forest structure, soil characteristics, nutrient deposition and five groups of weather and climate variables on the quantity and the species composition of recruiting trees. The climatic groups spanned annual averages, intra annual averages, annual variability, intra annual extremes and a combination of the aforementioned groups. The model with the combination of climate and weather variables outperformed all other groups. We found distinct climatic effects on tree recruitment quantities linked to water limitations and temperature extremes. The results as such showed that tree recruitment quantities benefit from stable climatic conditions, high precipitation and suffer from high maximum temperatures. Increasing temperatures also facilitate the share of recruiting broadleaves. The recruitment species was largely determined by the lead species in a plot, indicating the importance of seed limitation. Furthermore, the results confirm the important role of forest structure in tree recruitment and enable forest managers to steer the next generation of trees. Especially multi-species stands show a clear advantage over single species stands regarding tree recruitment quantities and diverse species compositions. Our research enables dynamic and state-of-the-art recruitment simulations across forests in Europe. It presents a reproducible method that can be applied to forest simulation modelling frameworks.
... Spatial terms were included as a smoothed thin plate regression spline of latitude, smoothed thin plate regression spline of latitude by longitude, or tensor product smooth of latitude by longitude. We evaluated goodness of fit for both the random effect structure and fixed effects using second-order Akaike information criterion (Burnham & Anderson, 2002) suited for small sample size (AIC c ) and Bayesian information criterion (BIC). The best model was refit using restricted maximum likelihood to better estimate model parameters for models with random effects. ...
... For top models with ΔAIC c <2, we averaged model coefficients using only the top models and the explanatory parameters included therein with the "model.avg" function in the MuMln package (Bartoń, 2022;Burnham & Anderson, 2002). Additionally, we summed each parameter weight to interpret as a measure of the relative importance of the parameter. ...
Objective
Apex-predator-initiated trophic cascades occur in many nearshore marine habitats that simultaneously serve as critical habitat and food sources for commercially and ecologically important species, including juvenile Pacific salmon Oncorhynchus spp. Yet the potential relationships among apex predators (e.g., sea otters Enhydra lutris), submerged aquatic vegetation, and juvenile salmonids are not well understood. In Southeast Alaska, we investigated (1) juvenile salmonid abundance in eelgrass Zostera marina meadows and understory kelp beds and (2) potential drivers of juvenile Chum Salmon Oncorhynchus keta and Pink Salmon O. gorbuscha abundance in eelgrass meadows.
Methods
We analyzed historic (1998–2007) beach seine surveys to compare juvenile salmonid abundance in nearshore habitats. We then employed the same survey (2017, 2019) in eelgrass to quantify juvenile salmonid abundance alongside the influence of sea otter density (number/km2), distance from anadromous stream (km), seasonality, sediment categorization, and aboveground eelgrass biomass (g/m2).
Results
We found greater abundance of Chum Salmon in understory kelp compared with eelgrass, whereas Pink Salmon abundance was similar between habitats. In eelgrass, Chum Salmon abundance peaked in June and was positively associated with sea otter density. Pink Salmon abundance varied seasonally, peaking in May. We found increased Pink Salmon abundance with increasing sea otter density and distance from anadromous stream and decreased abundance with increased eelgrass biomass.
Conclusion
Growth and survival while juvenile salmonids are out-migrating from streams and relying on nearshore vegetated habitats can determine if they recruit to fisheries as adults. Here, we highlight the use of multiple habitats, eelgrass and understory kelp, indicating that both should be explored as critical nursery habitat. We present evidence of indirect effects of sea otters influencing the abundance of juvenile salmonids, with potential further implications as sea otter populations expand. Apex predators, quality of vegetated habitats, and their structuring roles in the nearshore are critical for informing adaptive coastal fisheries management.
... Models including one or various of the explanatory variables were compared by ranking them according to their AICc using the function dredge from the MuMIn package (version 1.46.00) (Burnham & Anderson, 2002). When using Akaike's Information Criterion for model selection, AICc will sometimes support the inclusion of uninformative variables in the "best" models if a variable with poor explanatory power is added to an otherwise good model and the result is a model with ΔAIC < 2 (Arnold, 2010). ...
... All statistical analyses were performed using the lme4, MuMIn, and ggeffects packages in R statistical software version 4.1.2 (Bates et al., 2015;Burnham & Anderson, 2002;Lüdecke, 2018). We checked for normality and homogeneity of variance by using residual diagnostics following the procedures described in the DHARMa package for the initial model (Hartig, 2022). ...
Vulture populations have dramatically declined worldwide over the past few decades. In Europe, various programs have been implemented to reintroduce or restock these birds. In the Alps, the Bearded Vulture (Gypaetus barbatus) was eradicated in the early 20th century, prompting the launch of an international reintroduction program in 1986. Through intensive monitoring of Bearded Vulture breeding territories over 26 years (1995–2021), we observed spatial and temporal changes in population size and assessed environmental and individual factors influencing breeding performance across four countries in the Alpine arc. The number of breeding pairs increased steadily from 1 in 1995 to 65 in 2021, and the number of fledglings from 0 in 1995 to 42 in 2021, although we observed significant differences across areas, with slower growth in the Eastern Alps. The breeding success (proportion of clutches leading to a fledgling) of the entire Alpine population was 68%, higher than that of other European populations. Our analysis, which included individual and environmental variables, revealed that nest elevation positively influenced breeding success, while the proportion of open habitat in the territory had a negative effect. Additionally, breeding success significantly increased with the coverage of not strictly protected areas around the nest. In terms of individual factors, longer pair bonds positively affected breeding success, and pairs composed of wild‐hatched individuals had higher breeding success compared to those including released individuals. Our results confirm the overall positive trend of the reintroduced Alpine population but also highlight the increased vulnerability of the population breeding in the Eastern Alps. Additionally, our findings suggest that an effective network of protected areas can support the establishment of reintroduced populations. We also identified directions for further research to better understand the drivers of Bearded Vulture breeding success in the Alps.
... We then selected and ranked models based on the cumulative difference values of the Akaike Information Criterion (ΔAICc < 2). The candidate models are considered to have substantial support (Burnham and Anderson 2002). Next, we calculated the Akaike weight (w i ) for each candidate model based on the ΔAICc values. ...
... Next, we calculated the Akaike weight (w i ) for each candidate model based on the ΔAICc values. The Akaike weight (w i ) refers to the probability that the model is the best model in candidate models (w i > 0.9) (Burnham and Anderson 2002;Guthery et al. 2003). However, the Akaike weights indicated that none of the candidate models were the best model (Table S5). ...
The noise pollution, habitat loss, and human disturbance caused by urbanization have damaged bird communities. Research on the relationship between urbanization and birds has predominantly focused on highly urbanized areas, with relatively few studies in underdeveloped urbanized areas. Here, we conducted bird surveys along the urban–rural continuum by utilizing 150 line transects within a 51,385 km² area from June to August in 2022 and 2023, aiming to explore the impact of urbanization on bird species diversity and functional traits during the breeding season in the Huanghuai Plain of China. We found significant differences in species diversity and functional traits among three habitats along the urban–rural continuum (i.e., urban, suburban, and rural). Additionally, a measure combining several aspects of urbanization (the urban synthetic index) had significant negative correlations with species richness and the Shannon‐Wiener index, while it had no significant correlation with functional traits. We then assessed that the environmental noise, the distance to the county center, and the proportion of building area within a 250‐m radius were critical factors affecting species diversity, as well as environmental noise and the distance to the county center were the best predictors for functional traits. The composition and proportions of diets and nest types of birds were similar across the urban, suburban, and rural habitats. Our study highlights the importance of environmental noise, the distance to the county center, and the building index in protecting urban birds in the Huanghuai Plain. The research findings filled a gap in the study area regarding the relationship between urbanization and avian communities based on the urban–rural continuum.
... Lower AIC indicates a better fit. If ΔAIC >10, there is strong evidence that the model with the lower AIC is better (Burnham and Anderson, 2002). Crude fiber estimation from ADF of grass forages Table 3 presents the difference between the actual and estimated CF values of 11 grass forages. ...
Because of the cost and inaccessibility of laboratory facilities, animal feed formulation at the farm level, in many parts of Ethiopia, is based on feed database information. However, nowadays many laboratories are phasing out the Weende crude fiber (CF) method of analysis. The fiber content of feeds available in most feed databases (including the sub-Saharan Africa feeds composition database) are a result of detergent method analysis (NDF, ADF and lignin). However, CF is still used in poultry feed formulation and forage analysis for horses, in addition to the neutral detergent fiber (NDF) fraction for determining fiber in different countries. Since there is a statistically (P<0.01) difference between the CF and acid detergent fiber (ADF) value of a feed, ADF can’t be used directly in place of CF. Therefore, this work aims to formulate a regression equation that could roughly estimate the CF level of a feed from its NDF and ADF values. Considering the strong multicollinearity between NDF and ADF, this study developed separate models for ADF and NDF and compared them based on R² and Akaike Information Criterion (AIC), and the ADF-based model provided a better fit. The equations 0.79×ADF-0.46, 0.01+0.79×ADF, and 1.37+0.62×ADF have effectively predicted CF for cereal grains and beans, pulses and byproducts, and also oilseed meals and cakes, respectively. For grass forages, the equation 3.38+0.76×ADF, tested on 10 forages, showed potential but remains unreliable due to its R² value below 0.8. Finally, it is concluded that this approach provides a practical alternative for estimating CF where laboratory services or database information are unavailable. Keywords: Crude Fiber, Estimation, Feed database information, Prediction, Regression.
... The best substitution model for each gene was identified using PartitionFinder2 (Lanfear et al. 2017), with linked branch lengths via PhyML 3.0 (Guindon et al. 2010). Model selection used the corrected (for finite sample size) Akaike Information Criterion (AICc) (Burnham and Anderson 2002). Oligodon fasciolatus was used as outgroup (Pyron, Burbrink and Wiens 2013). ...
The analysis of mtDNA data (16S ribosomal RNA) shows that the Fowlea (formerly Xenochrophis) of the Andaman islands, formerly referred to as Fowlea (or Xenochrophis) tytleri are conspecific with F. piscator (Schneider, 1799). Specimens from the Andaman islands form a mixed clade together with individuals from the mainland and differ from the latter by genetic distances of 0.2–0.5% (except for 1.7% from a single mainland specimen that also differs from other mainland samples of F. piscator by 1.4–1.5% genetic distance). The genetic distances between individuals of F. piscator and its supposed sister taxon F. flavipunctatus vary from 4.8 to 7.3%. In pholidosis, F. tytleri and F. piscator are essentially identical. Based on the evidence from mtDNA and morphological data (the latter taken from the literature), we formally place Fowlea tytleri (Blyth, 1863) in the synonymy of F. piscator (Schneider, 1799).
... We used Akaike's information criterion adjusted for sample size (AIC c ) to evaluate the relative support for competing hazard models using AIC c > 2.00 as a threshold for model support (Burnham and Anderson 2002). First, we considered the potential confounding influence of sampling area (1-4) as a categorical fixed effect. ...
Reservoir predation has been identified as a leading mortality source for smolts migrating through impounded river systems. We investigated smolt predation risk for an endangered Atlantic salmon (Salmo salar) population in the Weldon Dam reservoir in the Penobscot River, Maine, USA. In spring 2022, we characterized the fates of 390 tethered smolts. Smolts were exclusively predated by two predator species not native to the study area: chain pickerel (Esox niger, n = 43) and smallmouth bass (Micropterus dolomieu, n = 42). Using Cox-proportional hazard analysis, we estimated that 23% (95% CI = 15%–29%) of tethered smolts were expected to be predated within a one-hour deployment. Water temperature was the primary driver of predation risk as predation probability increased from 10% to 33% when temperature increased from 5 to 15 °C. Smolts also incurred above-average predation risk when they were within 40 m of shore. We demonstrate that non-native fish predation may drive patterns of high impoundment mortality and that risk is spatially and temporally heterogeneous within these systems. Collectively, this study offers direct evidence of species-specific predation on Atlantic salmon smolts and illuminates potential strategies to mitigate predation risk during reservoir migration.
... Second, we constructed multiple models by leaving or removing insignificant (p > 0.05) paths from the initial model (Table S1). Next, we computed the χ 2 measure of goodness-of-fit (except for the full model due to lack of degrees of freedom) and the corrected Akaike information criterion for small sample sizes (AIC C ; Burnham & Anderson, 2002) for the full and each improved model. ...
Most studies of trade‐offs between defence, growth and reproduction have examined pairwise correlations between these processes, used ratio‐based measures for defence allocation such as allelochemical concentration and trichome density and estimated resource allocation to growth and reproduction in terms of biomass. However, for statistical and biological reasons, it may be preferable to analyse these processes holistically, to use absolute amounts of resistant traits and leaf mass/area, and to measure growth and reproduction in terms of nodes with or without flowers.
We aimed to identify how leaf stinging hairs as resistant structures and leaf area as a functional trait affected growth and reproduction in the Japanese stinging nettle Urtica thunbergiana.
We conducted a greenhouse experiment with nettles derived from a population that has been historically exposed to heavy browsing by Sika deer (Cervus nippon) in Nara Park, Japan. We analysed causal relationships between stinging hair number, leaf area, growth rate, growth performance and reproductive output using structural equation modelling (SEM). In this analysis, we adopted newly developed indices for a plant's stinging hair number and leaf area, and measures of growth and reproductive traits in terms of nodes instead of biomass.
There was a significant covariation between stinging hair number and leaf area. Stinging hair number had direct negative effects on growth rate and performance, while leaf area had positive direct effects on growth rate, growth performance and reproductive output. The growth rate had a significant direct positive effect on reproductive output, resulting in a significant indirect negative effect of stinging hair number on reproductive output.
This indicates that there is not only a trade‐off between resistance and growth but also an indirect trade‐off between resistance and reproduction through reduced growth rate and suggests that U. thunbergiana sacrifices growth rather than reproduction to increase resistance. Our study provides future work on trade‐offs between defence, growth and reproduction with a new methodological framework that can assess indirect as well as direct trade‐offs, together with the effects of leaf area as a functional trait on these processes.
Read the free Plain Language Summary for this article on the Journal blog.
... This produces a proxy for metabolic heating specific to this nesting site (Monsinjon et al. (2017a) for details). The best model was selected based on the lowest AICc (Burnham and Anderson, 2002) from models using a daily lag with air or sea temperature varying from 0 (i.e., synchronous relationship with nest temperature) to 5 days (i.e., lagged relationship with nest temperature at day+5). We used the standard deviation of the coefficients of the random effect as a proxy of nest thermal heterogeneity (TH) at the nesting beach scale (Monsinjon et al., 2019). ...
Knowledge of sex ratios of species with temperature-dependent sex determination (TSD) is key to provide baseline information which can be used to inform management strategies and predictions of how climate change can affect populations. In marine turtles, increased incubation temperatures can lead to extreme biases in sex ratios and reduced hatching success. Here we present a long-term analysis (34 years) of incubation durations of leatherback turtles (Dermochelys coriacea) in Brazil, the only population of this species breeding in the Southwest Atlantic, and estimate offspring sex ratios. We deployed data loggers recording nest temperatures in 2015, 2016, 2017 and 2019 (n = 28 clutches), to predict offspring sex ratios based on incubation temperatures during the thermosensitive period when sex is determined. The overall mean incubation duration for leatherback turtle clutches in Brazil (1988-2021) was 66.3 days (range 52-91, SD = 6.4, n = 867), decreasing by 4.4 days between the first and last 10 years of monitoring and varying latitudinally across the nesting range of the population. When modelled to the overall nesting season and accounting for nesting seasonality, we estimated the current (2012− 2021) mean season-wide primary sex ratio to be 46.9 % female (range 32.7 % to 84.8 %). Hindcasting for the first ten years of monitoring (1988-1997) showed the average predicted offspring sex ratios would have been 34.6 % female (range 7.7 % to 68.1 %). This population has not shown a phenological shift in the timing of nesting over the period 1988-2021. These findings suggest that, although the primary sex ratio of this population has likely become more female-biased in recent decades, the spatial and temporal variation observed demonstrates some resilience to the effects of increasing temperatures under climate change.
... Using the following equation [31], AIC was calculated: ...
This study aimed to describe the growth pattern in partridges using nonlinear models. Eight nonlinear mathematical functions (Bridges, Janoschek, Richards, Schumacher, Morgan, Lomolino, Sinusoidal, and Weibull) were used. The parameters of nonlinear models were estimated by fitting the models to partridge body weight records using the NLIN and MODEL procedures in the SAS program. Model performance was assessed and model behavior was examined during the process of fitting nonlinear regression curves. The overall goodness of fit of each model to various data profiles was assessed using the adjusted coefficient of determination, root mean square error, Akaike’s information criterion, and Bayesian information criterion. The adjusted coefficient of determination values for each model are generally high, indicating that the models fit the data well overall. Based on goodness of fit criteria, the Morgan model was found to be the most appropriate function for fitting the growth curve of male and female partridges. Furthermore, the Lomolino model had the worst fit to the growth curves of male and female partridges. While the predictions of the final body weight from all the models were good, the Morgan function outperformed the others in this regard. Based on the first derivative of the Morgan model, the absolute growth rates for male and female partridges as a function of time revealed that these values gradually increased with increasing age until 42 and 35 days of age, respectively, and then declined. The Morgan function is a useful replacement for conventional growth functions when describing the growth curves of different partridge breeds.
... We ran models with all combinations of effects (including interactions between ban and diet, and between ban and family for models over the whole study period) that could be derived from the general model using the dredge function in R (Bart on, 2015). We ranked models using the Akaike information criterion (AICc; Burnham and Anderson 2002). We deemed the effects of variables to be significant when the 95% confidence intervals of their estimate did not include 0. Effects are given ± one Standard Error (SE). ...
Lead pollution is a worldwide environmental and health issue causing persistent detrimental effects on humans and wildlife. Despite having been identified as detrimental to waterfowl a century ago, in France lead shotgun pellets for hunting were banned only in 2006 and only in wetlands. We used long‐term monitoring data from the Camargue (southern France) to (1) assess the local effectiveness of the French regulation at reducing the ingestion of lead shotgun pellets by waterfowl and to (2) assess local hunter compliance with the regulation. We used data on 2187 gizzards from 13 waterfowl species collected from 38 hunters in the Camargue over 20 hunting seasons (1998 to 2017). In addition, from 2008 to 2019, we systematically collected shotgun shell cases at three communal wetland hunting sites and used the ratio of cases indicating lead versus nontoxic shotgun shells to estimate hunter compliance with the ban. Across the 20‐year study period that spanned pre‐ and post‐ban, gizzards of harvested waterfowl had a mean lead shotgun pellet prevalence of 12% across the 13 species with no significant reduction over time. There was a slow increase in the use of nontoxic shot shells after the ban; lead shells continued to be used widely, suggesting inadequate policy enforcement. Because the ban was only applied for hunting in wetlands, hunters have to be caught with lead shells in the shotgun while in wetland to be fined. It is still legal to carry lead shells, favoring the circumvention of the regulation. Our results support the need for a complete ban of lead shotgun shells for both wetland and terrestrial hunting. A complete ban would facilitate policy enforcement and would favor compliance, eventually leading to a phase‐out of lead shell production and use.
... To this aim, different covariance structures can be tested, reflecting available information or alternative biological hypotheses. For the peach progeny used in this study, a diagonal covariance matrix, corresponding to the parameter independence, proved to be the best choice according to the AIC criterion (Burnham and Anderson, 2002), compared to a full-rank matrix (full covariance matrix AIC = 2396.1 vs diagonal covariance AIC = 2299.7). This suggests that the genetic control of the biochemical reactions involved in peach sugar metabolism is split over independent regulatory modules (Desnoues et al., 2016, Carreno-Quintero et al., 2013. ...
Integrating genetic information into dynamical models is crucial for understanding genotype variation and
improving crop performance in the face of current agronomic and ecological constraints. Calibration over a large
number of genotypes is a key step in building gene-to-phenotype models but remains challenging. We compared
two strategies for calibrating an Ordinary Differential Equations (ODE) kinetic model simulating sugar accumulation during peach fruit development. In the first strategy, a Genotype-Based (GB) approach, the model was
calibrated independently for each genotype, using either a Single-Objective Optimization (GBS) or a MultiObjective Optimization (GBM) formulation. In the second strategy, a Population-Based (PB) approach, the
model was calibrated for all genotypes simultaneously. The two strategies were applied to simulated data and to
a real dataset from 106 peach genotypes. Results showed that the GB strategy allowed for a high goodness of fit
for most genotypes, especially in the GBS formulation. However, the estimated parameters suffered from a lack of
practical identifiability as independent repetitions of the estimation algorithm did not always converge to the
same value for most genotypes. The PB calibration strategy overcame this issue showing a good identifiability of
the population parameter values, a goodness of fit comparable to the one obtained with the GB strategy, and a
good characterization of parameter variations within the progeny, which is key to assess the inter-individual
genetic variability. This study demonstrates the value of the PB calibration strategy in capturing interindividual genetic variability, a critical step toward developing robust gene-to-phenotype models.
... Additionally, the null model was also included. Model selection was based on the corrected Akaike information criterion for small samples (AICc), focusing on models with a ΔAICc < 2 units (Burnham and Anderson 2002). All analyses were conducted using RStudio version 2024.04.2 Build 764 ("Chocolate Cosmos" Release, e4392fc9, 2024-06 -05) with Quarto 1.4.555. ...
Agricultural practices significantly influence riparian landscapes, increasing contaminant inputs into aquatic ecosystems, with profound effects on biodiversity and ecological processes. This study evaluated the structural and functional responses of autotrophic biofilms to nitrogen (10 µg/L as urea fertilizer), glyphosate (8.64 µg/L of active ingredient, Green II Herbicide Soluble Concentrate GREEN-GO®), and their combination in a 15-day laboratory microcosm experiment conducted during May and June 2021. The experiment used 1.5 cm × 1.5 cm frosted glass substrates, which were colonized in the laboratory with a natural algal community inoculated from the Chocancharava River, Córdoba, Argentina (33°03′14″ S, 64°22′23″ W). We assessed total biomass, chlorophyll a, algal community structure, and metabolic activity under four treatments: control, nitrogen, glyphosate, and nitrogen-glyphosate. Glyphosate significantly reduced taxa diversity and altered community structure, leading to a decline in diatoms and the dominance of Monoraphidium minutum, likely due to its tolerance mechanisms. Nutrient enrichment and glyphosate exposure simplified algal assemblages, favoring tolerant taxa. The density of Monoraphidium minutum significantly influenced net primary production, underscoring its functional role under stress conditions. These findings demonstrate how tolerant species drive biofilm responses to diffuse pollution, with implications for trophic connectivity and ecosystem functioning due to community simplification.
Graphical Abstract
... Models with catchments as a fixed effect were tested but did not perform as well as those using catchment or climate properties. Model performance was evaluated using correlation coefficients, the Akaike Information Criterion (AIC) (Akaike 1974;Burnham and Anderson 2002), and the Shapiro-Wilk test for residual normality (Shapiro and Wilk 1965). The best models had high correlation coefficients, low AIC values, and normally distributed residuals. ...
The importance of subsurface water dynamics, such as water storage and flow partitioning, is well recognised. Yet, our understanding of their drivers and links to streamflow generation has remained elusive, especially in small headwater streams that are often data-limited but crucial for downstream water quantity and quality. Large-scale analyses have focused on streamflow characteristics across rivers with varying drainage areas, often overlooking the subsurface water dynamics that shape stream-flow behaviour. Here we ask the question: What are the climate and landscape characteristics that regulate subsurface dynamic storage, flow path partitioning, and dynamics of streamflow generation in headwater streams? To answer this question, we used streamflow data and a widely-used hydrological model (HBV) for 15 headwater catchments across the contiguous United States. Results show that climate characteristics such as aridity and precipitation phase (snow or rain) and land attributes such as topography and soil texture are key drivers of streamflow generation dynamics. In particular, steeper slopes generally promoted more streamflow, regardless of aridity. Streams in flat, rainy sites (< 30% precipitation as snow) with finer soils exhibited flashier regimes than those in snowy sites (> 30% precipitation as snow) or sites with coarse soils and deeper flow paths. In snowy sites, less weathered, thinner soils promoted shallower flow paths such that discharge was more sensitive to changes in storage, but snow dampened streamflow flashiness overall. Results here indicate that land characteristics such as steepness and soil texture modify subsurface water storage and shallow and deep flow partitioning, ultimately regulating streamflow response to climate forcing. As climate change increases uncertainty in water availability, understanding the interacting climate and landscape features that regulate streamflow will be essential to predict hydrological shifts in headwater catchments and improve water resources management.
... Furthermore, the statistical method developed by Okamura and Semba (2009) was applied to validate the periodicity of the band pair formation using three models: (1) acyclic (N); (2) annual (A); (3) and biennial (B). The model that best fit the data was selected based on the lowest value of the Akaike Information Criterion (AIC) (Burnham and Anderson 2002). ...
Context. Carcharhinus falciformis (silky shark) ranks third in elasmobranch landings in the Ecuadorian Pacific. However, there is limited and outdated information about its life history, making it challenging to propose management strategies. Aims. To determine the age and growth parameters of C. falciformis in the southern Ecuadorian Pacific. Methods. At the artisanal port 'Playita Mía', vertebral samples were collected from 290 specimens (161 females and 129 males) and sectioned sagittally to estimate age. A Bayesian and frequentist multimodel approach (von Bertalanffy, Gompertz, and Logistic) was used for length-at-age analysis. Results. Sizes ranged from 79.2 cm to 238 cm total length (TL), with age groups estimated between 0 and 19 years. The Bayesian von Bertalanffy growth model provided the best fit according to the leave-one-out information criterion (LOOIC). The estimates were: (1) for combined sexes, L ∞ = 271.54 cm TL, k = 0.09 years −1 , and L 0 = 72.21 cm TL; (2) for females, L ∞ = 271.83 cm TL, k = 0.09 years −1 , and L 0 = 71.61 cm TL; and (3) for males, L ∞ = 273.60 cm TL, k = 0.08 years −1 , and L 0 = 72.99 cm TL. Conclusions. This study demonstrates that C. falciformis is a slow-growing species, making it susceptible to overfishing. Implications. The findings provide valuable information for future population and demographic analyses of this species.
... We ranked models using the second-order Akaike information criterion (AICc), which favours more parsimonious models through maximum likelihood penalized by the number of variables used in the model's composition (Burnham and Anderson, 2002). Models were also ranked by ΔAICc (the difference between AICc of the best model and other models) and wAICc, which represents the proportional weight of each model, that is, how representative the contribution of a model is. ...
Species abundances and richness are central parameters in ecology and crucial for describing diversity and composition across environments. Understanding how they vary in natural environments is critical for informed conservation decisions, especially in the face of anthropogenic pressures, such as deforestation and climate change. We evaluate the influence of landscape and local habitat variables on the richness and abundances of lizards in the Caatinga, the largest continuous block of seasonally dry tropical forests. We sampled seven lizard communities for three months using visual encounters along transects. We recorded landscape and microhabitat variables and evaluated their influence on lizard species richness, diversity, and occurrence using model selection. Ten lizard species were recorded, with Tropidurus semitaeniatus, Ameivula ocellifera, and Tropidurus hispidus being the most abundant. Topographic complexity and the number of rocky outcrops positively affect species richness and diversity by promoting environmental heterogeneity and hence increasing refuges, shelters, and thermoregulation sites. Different microhabitat and landscape variables were important predictors of the occurrences of individual lizard species. The quantity of rocks significantly increased the likelihood of Tropidurus semitaeniatus occurrence, while litter negatively affected Tropidurus hispidus, and fallen logs increased the probability of Ameiva ameiva occurrence. We argue that preserving topographically complex regions is essential for maintaining the diversity of lizards in the Caatinga biome.
... Models with catchments as a fixed effect were tested but did not perform as well as those using catchment or climate properties. Model performance was evaluated using correlation coefficients, the Akaike Information Criterion (AIC) (Akaike 1974;Burnham and Anderson 2002), and the Shapiro-Wilk test for residual normality (Shapiro and Wilk 1965). The best models had high correlation coefficients, low AIC values, and normally distributed residuals. ...
Keywords: flow path partitioning | headwater catchments | runoff generation | shallow and deep subsurface flow | snow | storage-discharge sensitivity | streamflow flashiness | subsurface dynamic storage ABSTRACT The importance of subsurface water dynamics, such as water storage and flow partitioning, is well recognised. Yet, our understanding of their drivers and links to streamflow generation has remained elusive, especially in small headwater streams that are often data-limited but crucial for downstream water quantity and quality. Large-scale analyses have focused on streamflow characteristics across rivers with varying drainage areas, often overlooking the subsurface water dynamics that shape stream-flow behaviour. Here we ask the question: What are the climate and landscape characteristics that regulate subsurface dynamic storage, flow path partitioning, and dynamics of streamflow generation in headwater streams? To answer this question, we used streamflow data and a widely-used hydrological model (HBV) for 15 headwater catchments across the contiguous United States. Results show that climate characteristics such as aridity and precipitation phase (snow or rain) and land attributes such as topography and soil texture are key drivers of streamflow generation dynamics. In particular, steeper slopes generally promoted more streamflow, regardless of aridity. Streams in flat, rainy sites (< 30% precipitation as snow) with finer soils exhibited flashier regimes than those in snowy sites (> 30% precipitation as snow) or sites with coarse soils and deeper flow paths. In snowy sites, less weathered, thinner soils promoted shallower flow paths such that discharge was more sensitive to changes in storage, but snow dampened streamflow flashiness overall. Results here indicate that land characteristics such as steepness and soil texture modify subsurface water storage and shallow and deep flow partitioning, ultimately regulating streamflow response to climate forcing. As climate change increases uncertainty in water availability, understanding the interacting climate and landscape features that regulate streamflow will be essential to predict hydrological shifts in headwater catchments and improve water resources management.
... The main objective is to compare these techniques as well as to assess their effectiveness in predicting survival probabilities and determining factors that significantly affect breast cancer mortality. Models chosen based on statistical criteria such as **Akaike Information Criterion (AIC)**, **Bayesian Information Criterion (BIC)**, etc. help to find the best-fitting model for survival analysis Burnham and Anderson (2002). By comparing different techniques of survival analysis, a research goal will definitely gain insight into breast cancer in the future, as far as scrutinizing the statistical methodologies used in medical research is concerned. ...
This study offers a comparative analysis of parametric, semi-parametric, and non-parametric techniques of survival data analysis. It focuses on assessing the methods in terms of their efficiency in estimating survival probabilities and hazard functions. Alternate real and simulated datasets are then utilized to assess the relative strengths and weaknesses of different approaches with respect to efficiency, flexibility, and interpretability. The results indicate that the choice of technique is a factor of the underlying data characteristics, with parametric models working quite well under specific assumptions, semi parametric methods balancing between the structure and flexibility, while the nonparametric ones have the best performance in scenarios driven by the data. The AIC and BIC values determined from the model selection process demonstrated that the model finding fits the best under the given conditions as it has the smallest AIC and BIC scores. This is because AIC and BIC balance between goodness of fit and complexity in the model. For the Cox proportional hazard model results, most variables tested in the analysis showed very few statistically significant p- values. Results from Kaplan-Meier survival analysis strengthened the model by providing critical survival probability key stages over time, indicating the most occurrences beyond 60 months marked significant declines in survival rates. Hence, since the determined model com- bines predictive accuracy with interpret-ability, it should have as few but robust elements as would be meaningful in survival prediction without over-fitting. Therefore, the final model selection is based on the best compromise between explanatory power and statistical validity.
... The effects of covariates were judged as significant or statistically significant if their confidence interval (CI) did not cross 0, and as non-significant effect otherwise. We used the Akaike information criterion corrected for small sample size (AICc) for model ranking and regarded those with delta ΔAICc <2 as the best candidate models (Burnham & Anderson, 2002). We applied 'MuMIn' R package to examine possible combinations of the covariate effects (Bartón, 2022). ...
The garden dormouse (Eliomys quercinus) is an understudied mammal species endemic to Europe. Once distributed widely across the continent, its eastern populations have significantly declined, with the species now found in only 50% of its former range. In Germany, it also occurs in spruce forests in low mountain ranges. However, this habitat has recently been lost due to bark beetle infestation following a prolonged period of drought. In some places, not only the dead trees are removed, but also all the branches and topsoil.
We assessed habitat requirements of a garden dormouse population in such a changing habitat in the Harz Mountains, Germany. To assess garden dormouse occupancy, we conducted transect surveys using footprint tunnels during July and October 2022 for the presence of garden dormice. Additionally, we measured covariates such as coverage of tree, shrub and herb layers, as well as deadwood, soil characteristics and reforestation. Using single‐season occupancy modelling, we evaluated the effects of these habitat covariates on the occupancy of the garden dormouse.
Our results indicated that garden dormouse occupancy was negatively affected by herb cover but positively associated with plant successional stages. Furthermore, occupancy of the garden dormouse was positively coupled with the occurrence of the hazel dormouse (Muscardinus avellanarius). Detection probability increased with higher weekly minimum temperatures.
Practical implications: For the conservation of garden dormice in low mountain ranges like the Harz, practical measures should include establishing interconnected forest edges and strips. Moreover, we suggest adopting a dynamic, mosaic approach to felling, avoiding topsoil removal and promoting regeneration that fosters semi‐open successional habitats as essential strategies.
... bifurcilinea had VIF values below 2, indicating no collinearity issues. Models were compared using the corrected Akaike information criterion (AICc) for small sample sizes (Burnham and Anderson 2004) using the model.sel function in the R package MuMIn (Barton 2009). ...
The adaptive significance of female ornamentation remains a central question in evolutionary biology, with ultraviolet (UV)-induced fluorescence emerging as a key area of interest. This study investigates the potential adaptive advantages of female-specific UV-induced fluorescence in male mate choice and predation risk, as fitness costs, using two species of ornate jumping spiders Phintella vittata and Ph. bifurcilinea. In these species, the palps of adult females exhibit UV-induced fluorescence, offering a compelling model to explore the interplay of sexual and natural selection acting on female ornamentation. In male mate-choice trials, males were presented with a choice between a fluorescent (F+, UV-visible) and a non-fluorescent (F-, UV-blocked) female. Males showed pronounced mate preference for F+ females over F-females, that is, spending significantly more time interacting with F+ females, suggesting that fluorescence serves as a sexually selected signal. To assess the potential costs of fluorescence, we tested its effect on predation risk using the spider-eating jumping spider Portia xishan as a predator under F+ and F-conditions. Predation rates were significantly higher for F+ females than for F-females, indicating that UV-induced fluorescence increases detectability by predators. These findings provide empirical evidence of a trade-off: While fluorescence enhances male mate preference, it also increases predation risk. This study is the first to demonstrate the dual roles of fluorescence in sexual signaling and predation in female jumping spiders, challenging traditional male-centric perspectives on mate choice. By integrating behavioral and ecological approaches, this work offers new insights into the evolutionary trade-offs associated with female sexually selected traits.
... Impacts on hatching success of loggerhead sea turtles in Italy and our selected models prioritize those with fewer parameters that avoid overfitting [53]. The best model (Model 549) included F_Dune, LST, and Nat/Ant as predictors. ...
Coastal biodiversity is globally threatened by climate change and human pressures, including habitat destruction, overfishing, eutrophication, and pollution, which alter natural coastal ecosystem processes. Caretta caretta, hereafter referred to as loggerhead sea turtle, is listed as “Vulnerable” at global level and “Least Concern” in the Mediterranean Sea by the International Union for Conservation of Nature (IUCN). This species is the only sea turtle nesting along Italian coasts, making it crucial to understand the factors affecting its reproductive success for effective conservation strategies. However, key aspects of the ecology and life cycle in Italian waters, such as spatial distribution, reproductive site selection, and factors influencing reproductive outcomes and migratory movements, remain unknown. This study aimed to identify factors influencing the reproductive success of the loggerhead sea turtle. Data from 237 nests between 2019 and 2023 across 14 Italian regions were recovered, quality-checked, and analyzed. A statistical model predicting reproductive success, represented by hatching success, was developed, incorporating various environmental variables from marine and terrestrial spheres, along with local pressures from urbanization and anthropization. These predictors were related to hatching success using a generalized linear model (GLM) accounting for zero-inflated data. The best models identified both environmental variables, such as temperature and extreme wave events, and human-controlled factors, including the presence of dunes and coastal urbanization, as key predictors of hatching success. Coastal anthropization and beachfront disturbances were particularly important. While human activities can pose notable challenges to the loggerhead sea turtle, their identification also offers opportunities for enhancing its reproductive success through targeted management actions focused on mitigating pressures. Our findings highlight the urgent need for targeted conservation efforts to address both local and global challenges to protect and enhance the reproductive success of the loggerhead sea turtle and possibly other coastal species. Effective management can and should focus on mitigating human-induced pressures. Policymakers and conservationists need to work together to implement strategies that consider both the immediate human impacts and the long-term effects of climate change, ensuring the sustainable management of coastal ecosystems and the protection of endangered species like the loggerhead sea turtle.
... We used the Akaike information criterion for small samples (AICc) to select the most parsimonious models (Martin and Fahrig 2012). Models with ∆ ≤ 2 were considered equally plausible (Burnham and Anderson 2002). We considered that an independent variable was causing an effect on the response if the 95% confidence interval of its regressor coefficient did not include zero. ...
The Cerrado is a highly threatened biome that harbors a great portion of Brazilian biodiversity, including many endemic species. Such richness may reflect its environmental heterogeneity, which includes different vegetation types (from grasslands to forests). Several endemic species in the Cerrado are strongly associated with open vegetation types. Recent studies have shown that woody encroachment can negatively impact several components of biodiversity. In the present study, we provide information on variation in lizard diversity across different vegetation types in a protected area located in the southern portion of the Cerrado and evaluated the effect of vegetation density on lizard taxonomic, functional, and phylogenetic diversity. We predicted that the three aspects of lizard diversity would be negatively affected by vegetation density and that Beta diversity would vary significantly across the vegetation gradient. We confirmed nearly all of our predictions, except for functional diversity, which, contrary to our expectations, was positively influenced by vegetation density. We also found that Beta diversity indices based on both species composition/abundance and phylogenetic features of communities experience a significant turnover across the vegetation density gradient. This study greatly strengthens previous studies that suggest a negative impact of vegetation density (and, consequently, woody encroachment) on lizard diversity. Therefore, managers of protected areas in the Cerrado should consider the use of strategies to avoid woody encroachment, like fire management with prescribed fires.
... We used the Akaike information criterion for small sample size (AIC c ) to determine the best-approximating model. In cases where a single best-fitting model could not be determined and no other models had ΔAIC c < 2, the effects of individual variables were estimated via multi-model inference (Burnham and Anderson 2002), and model predictions were calculated using conditional model averaging whereby parameters are only averaged over models in which they occur (Grueber et al. 2011). Explanatory variables with unconditional 95% CI overlapping zero were noted as uninformative (Arnold 2010). ...
The nutritional quality of nestling diets, influenced by parental prey selection, determines offspring health and development. In agricultural landscapes, nestling diets may be impacted by changes in prey availability from agricultural intensification. In 2020 and 2021, we compared landscape-level effects on nestling prey resources in two sympatric species of breeding aerial insectivores, Tree Swallow (Tachycineta bicolor) and Barn Swallow (Hirundo rustica). Our study was conducted across an agricultural crop and wetland gradient in Saskatchewan, Canada. Using hydrogen stable isotope analyses (δ²H) of nestling feathers, as an indicator of differential diet composition of terrestrial vs. aquatic-emergent insects, we hypothesized that Tree Swallows, as aquatic insect specialists, would have lower feather δ²H values indicative of more aquatic diets. We predicted that increasing annual-row crop and wetland cover around nests would differentially impact each species with Tree Swallows being more sensitive to landscape differences given known reliance on wetland-derived prey. Wetland waters and select aquatic and terrestrial insects showed high variation presumably due to seasonal stochastic evaporation requiring greater sampling effort. However, evidence of differential use of aquatic resources was consistently found between the two swallow species and between wet versus dry years. Lower average Tree Swallow nestling feather δ²H values suggested their diet was more reliant on aquatic-emergent prey, unrelated to the land use around the nest (at 500–2000 m). In contrast, Barn Swallows had higher average feather δ²H, which decreased with greater standing water cover in proximity of the nest (< 500 m), suggesting more terrestrial diets with opportunistic use of aquatic-emergent prey resources. For both species, we found no effect of crop cover extent on the isotopic indicator of prey source. These results contribute to the growing body of evidence that multiple species of aerial insectivores rely on aquatic insect resources regardless of local agricultural land use. Our results highlight the importance of conservation of small, interspersed wetlands, especially in intensive cropland-dominated landscapes, to benefit multiple species of aerial insectivores.
... Sometimes, it is suggested that if the difference between two AIC/BIC scores is greater than 2, then the model with the smaller AIC/BIC score is more supported. If the difference is less than 2, then both models are equally well supported (Burnham and Anderson 2002;Burnham et al. 2011;Berchtold 2010;Yin and Miller 2022). The quantilequantile (Q-Q) plots and observed versus fitted density are popular techniques for model diagnostics. ...
Bangladesh is highly susceptible to the impacts of climate change, particularly extreme rainfall during the monsoon season, leading to severe floods and landslides. This study introduces a nonstationary generalized extreme value (GEV) modeling framework, which integrates atmospheric dry bulb temperatures as a covariate to capture the seasonal and dynamic characteristics of extreme rainfall events. Using daily rainfall and temperature data from Dhaka (1990–2015) and Chattogram (1999–2015), the study identifies optimal models based on AIC, BIC, and goodness-of-fit criteria. Uncertainties in the predictions are quantified using the delta method and parametric bootstrap approaches. The results indicate a higher likelihood of extreme rainfall events in Chattogram compared to Dhaka, as reflected in the predictions and probabilities in return levels. Diagnostic evaluations confirm that the models effectively capture the variability in monthly maximum rainfall during the monsoon. These findings offer valuable information for flood risk management, urban planning, and disaster preparedness. By incorporating temperature effects and quantifying prediction uncertainties, the study addresses key limitations in existing methodologies. Future work will expand this framework to assess spatiotemporal rainfall variability in Bangladesh and explore advanced machine learning approaches to simultaneously model the bulk and tail of rainfall distributions.
... These included models with additional lags and the incorporation of the number of tourist arrivals into both the economic contribution model and the tourist behaviour model. The Akaike Information Criterion (AIC) values were used to compare the model fit, with lower values indicating a better fit [7]. ...
... For all GLMMs, the best model was selected based on the lowest Akaike information criterion (AIC) using backward elimination in the dredge function in the MuMIn package (Burnham and Anderson 2002). Furthermore, the best pair of climatic variables was selected based on the lowest AIC (Zuur et al. 2010), and we did not observe any patterns in the residuals. ...
Aim
Norway spruce ( Picea abies ) is a dominant tree species across Eurasia, and is known to regenerate on decaying logs. Understanding its regeneration dynamics is crucial for predicting forest sustainability under climate change. The decay type of the logs, indicative of fungal decay capabilities of lignin and holocellulose and traditionally categorised into white rot and brown rot, markedly influences spruce seedling density on the logs, because brown rot reduces their establishment. As fungal distribution and wood decay types are affected by macroclimate, the success of spruce seedling establishment is expected to vary continentally and may be further influenced by climate change. The goal of the present study is to examine the relationship between decay type frequencies and spruce seedling densities on decaying logs, assessing effects on regeneration success along a climatic gradient.
Location
Fifteen sites in six European countries.
Taxa
Norway spruce ( Picea abies [L.] H. Karst.).
Methods
We collected wood samples from spruce logs naturally decaying in the forest to evaluate their decay types. Spruce seedling numbers were recorded on the logs. The occurrence of certain decay types, seedling density, and their relationships with climate, stand, and log variables were analysed using generalised linear mixed models (GLMM) and structural equation modelling (SEM).
Results
White rot type of decay was dominating in central Europe compared with southern and northern Europe, where brown rot was prevalent. Spruce seedling densities were positively correlated with white rot in heartwood, which was explained by high precipitation seasonality in central Europe.
Main Conclusions
Spruce seedling regeneration on decaying logs may be indirectly affected by climate through its influence on fungal‐mediated wood decomposition. Despite predictions of northward shifts in Norway spruce distribution due to climate change, slower regeneration rates on brown rot logs may limit northern expansion, potentially leading to a contraction of the species' range.
... We reported parameter estimates and 95% confidence intervals for each model and considered parameters with 95% confidence intervals that included 0 not significant (Burnham and Anderson 2002). However, we also reported predicted values and 95% confidence intervals for all parameters included in our models. ...
Advancements in tracking technologies have improved our knowledge of migratory movements for many songbirds and simultaneously raised questions about how tracking devices may influence an individual’s survival and behavior. During a five-year study (2017–2021) at two study sites in central Texas, USA, we quantified the potential impacts of miniaturized light-level geolocators on a small (~10 g) Nearctic-Neotropical migratory songbird, the Golden-cheeked Warbler (Setophaga chrysoparia; hereafter warbler). We used generalized linear models to examine the effects of geolocator, age, and year on warbler pairing success, fledging success, and apparent annual return rates (hereafter return rates) at each study site. Most parameters we considered in our pairing success, fledging success, and return rate models were not statistically significant. However, at one study site, we found lower return rates for second year warblers with geolocators compared to second year and after second year warblers without geolocators. We also described warbler behavior immediately after we attached the geolocators, body condition upon their return, and body mass before and after carrying the devices. Most warblers resumed normal activities (e.g., singing, foraging) within 30 min after we attached the devices, showed no visible signs of injuries, and had similar body mass before and after carrying the geolocators. Overall, we did not find consistent evidence across sites or age classes that miniaturized light-level geolocators negatively impacted the Golden-cheeked Warbler responses we measured. However, we encourage continued research on this topic, especially as it relates to age-dependent effects of geolocator attachment on birds.
... To account for a small sample size, we used a modified Akaike's information criterion (AICc) and calculated the difference in AICc values ( AICc) and the AIC weight (w i ). Models with AICc values of less than two are considered to have less support (Burnham and Anderson 2002). To describe the general trend of the height growth-gap size relationship, we summed the weight of evidence (i.e., w i values) for linear and power models to indicate monotonically increasing height growth; we summed the quadratic and saturating models to indicate nonmonotonic height growth. ...
Where frequent fires are the primary driver of gap dynamics, gap-based silviculture can be deployed to restore fire-maintained structures. We studied the growth of regeneration for six species in a western US mixed conifer forest, following 12 years of growth across a range of gap sizes (0.1–1 ha). We found that severe edge effects limited growth within 3 to 4 m of surrounding canopy trees. Using an edge zone where competition is expected to include more moderate levels of growth suppression, we compared 12-year growth in edge zones versus interior zones. Ponderosa pine was the tallest among edge trees, followed by giant sequoia, Douglas-fir, sugar pine, incense-cedar, and white fir. Following experimental gap expansions (i.e., femelschlag harvest), we observed that increases in light availability were substantial along southern edges, whereas northern edges had no increase. Saplings in edge zones grew substantially in response to expansion. Our results demonstrate that using gap-based silviculture can create coarse-scale heterogeneity while regenerating all species, including shade-intolerants in smaller gaps. Gap expansions can be used to restore some of the structural complexities that fire historically maintained. Colonnades of tree-free space may enhance objectives of heterogeneity and fuel discontinuity that frequent fires used to maintain.
... Additionally, for abundance analyses of forest species, we first used offset(log(Area)) as an offset 27 to control for variable site area. We conducted 2 goodness-of-fit tests and constructed response curves for each covariate using model-averaged predictions with 95% confidence intervals 28 . To assess the impact of altered land cover on the avian community, we also recorded the occurrences of all avian taxa within woody sites (including study taxa) and used these to estimate avian functional diversity based on Bregman et al. 29 . ...
Globally, managing invasive plants and habitat transformation often constitutes a ′wicked problem′ due to highly variable impacts on threatened biodiversity. In naturally patchy habitats, invasion can blur edges, pushing species beyond natural habitats. In biphasic habitats like forest-grassland mosaics, complementary sets of specialist species offer a unique configuration to examine such impacts from both directions. In the Shola Sky Islands, extensive woody invasive stands and agriculture/production landscapes have created a set of transformed closed and open habitats structurally similar to their natural counterparts. We expect the species′ usage of such novel habitats to reflect their dietary and habitat specialisation.
We conducted 4,519 surveys across 1,204 randomly selected grid cells for the avian community, covering the global distribution of five species. For a select set of specialist and generalist species, we estimated patch occupancy and abundance using hierarchical models. We used acoustics to assess species persistence across invasion stages with automated recorders deployed across a year. We used community occurrence data to examine functional traits that correspond to habitat colonisation.
Forest species occur year-round in transformed woody habitats across all invasion stages, and habitat overstory determines avian functional diversity. Forest specialists decline in transformed habitats across high-contrast edges, while generalists increase across both high- and low-contrast edges, indicating a ′supertramp′ strategy. Grassland specialist, however, decline strongly beyond all edges, creating a duality of losses and gains across the biphasic matrix.
We highlight the diversity of threatened species′ responses to invasion and habitat transformation, underscoring the importance of nuanced approaches to habitat restoration, particularly in biphasic habitats.
... Four functions (linear, quadratic, cubic, and inverse) were used to relate the climate and response. The corrected Akaike information criterion (AICc) was used to select the most parsimonious models following an information-theoretic approach (Burnham and Anderson, 2004). Then, the difference between the AICc of the selected model and that of the null model was calculated (ΔAICc). ...
... Additionally, in the mixed-effect models, we incorporated the study site as a random factor and the statistical interaction between sea distance and the biodiversity variables to address potential impacts of stress and disturbance on the BEF relationships (La Bella et al., 2024). For each response variable, we applied a model averaging approach using the "dredge" function (MuMIn R package; Barton & Barton, 2023) based on Akaike Information Criterion (ΔAIC < 3; Burnham & Anderson, 2002;Richards, 2005) to identify the best predictors. Model residuals underwent visual inspection to ensure homoscedasticity and normality (Zuur et al., 2010). ...
Protected areas are generally designed to conserve biodiversity. However, how well they also contribute to maintaining ecosystem functions that plant diversity supports has rarely been explicitly tested, often due to the lack of historical ecosystem function data. Here, we used a trait-based approach to reconstruct past ecosystem functioning and examine changes in functions over the last 15 years in protected and unprotected coastal dune ecosystems, checking where functions remain stable over time. First, we resurveyed vegetation in quasi-permanent plots and measured in the present several ecosystem functions related to productivity, carbon, water, nutrient cycling, erosion control, and invasion resistance across six coastal dune sites in Central Italy. Second, using these data, we quantified Biodiversity-Ecosystem Function (BEF) relationships and employed them to hindcast past ecosystem functions based on historical vegetation surveys. Finally, as a case study, we applied this method to assess temporal changes in ecosystem functioning under three protection regimes: national protected areas (i.e. strict protection), Natura 2000 sites (loose protection), and non-protected areas. Productivity, carbon, and water regulation increased over time in non-protected areas, likely due to an expansion of ruderal and non-native species, that are usually more productive. Within Natura 2000 sites, communities showed a decrease in erosion control and invasion resistance potential, due to the loss of important dune-building species and the spread of non-natives. Only within national protected areas, ecosystem functions did not undergo significant temporal changes, and invasion resistance even increased. Our results suggest that in the study area ecosystem functioning remain stable over time only in areas under strict protection. More broadly, our study demonstrates the potential for using revisitation data in combination with locally estimated BEF relationships to hindcast past ecosystem functioning, providing a valuable tool for monitoring long-term functional changes in response to conservation measures.
Changes in landscapes due to the conversion of natural vegetation into agroecosystems are causing an alarming and rapid loss of biodiversity on a global scale. We assess how local environmental factors and landscape structure influence Squamata reptiles’ abundance, and alpha and beta diversities in agricultural landscapes in the Brazilian Cerrado. We sampled 20 sites using active visual search and pitfall traps to record Squamata reptiles’ abundance and richness. For each sampling site, we measured seven variables of vegetation structure at local scale and calculated four landscape metrics at five spatial scales. We recorded 145 individuals of 30 Squamata species comprising 15 lizards, 12 snakes, and three amphisbaenians. Agriculture cover had a negative effect on lizards’ abundance and richness. No effect was found on the abundance and richness of snakes. Species turnover was the major process responsible for Squamata reptile dissimilarity among sites. Reptile total beta diversity and species turnover were influenced by variations in the number of trees at the local scale. Our results highlight the negative effects of agriculture expansion on species abundance and richness in lizard communities. Moreover, vegetation structure may act as an environmental filter and predict the variation in Squamata reptiles’ species turnover between sites. Our findings show that maintaining well-preserved patches of natural vegetation in agricultural landscapes is of utmost importance for Squamata reptiles’ conservation and should be considered in landscape conservation and management design.
Preferences for animal species may influence public engagement in conservation efforts. However, urban residents often have limited opportunities to learn about and connect emotionally with a wide range of wildlife. Zoos can help bridge this gap, with enclosures and interpretive systems playing a key role in fostering understanding and positive attitudes towards animals. At present, little is known about how the characteristics of enclosures and interpretive systems work together to influence visitors' learning about wildlife. Over a 12‐month period, we assessed the features of enclosures and interpretive facilities at Guangzhou Zoo, China, and captured visitors' wildlife preferences before and after their visits using an open‐ended question. By examining how these features and the overall zoo experience influenced changes in preferences, we aimed to evaluate the zoo's potential for conservation education. Our results showed that zoo visits significantly impacted visitors' wildlife preferences, increasing their fondness for a broader range of species and fostering a greater affinity for relatively unpopular species or taxa. Importantly, the interaction between the number of interpretive signs and three other factors—enclosure size, enclosure type, and the interactivity of interpretive facilities—emerged as key determinants of these preference changes. This study provides valuable insights into improving the design of zoo enclosures and interpretive facilities to support conservation education.
Background
Among early-onset conduct problems (CP), associated with more disruptive behaviors of greater intensity and stability, several domains have been proposed from a variable-centered perspective to capture their heterogeneity: oppositional defiant disorder (ODD), attention-deficit/hyperactivity disorder (ADHD), and callous-unemotional traits (CU). Using a person-centered approach, the present study aims to identify distinct profiles of child CP, examining different predictors and developmental outcomes.
Methods
Data included parent- and teacher-reported questionnaires from an ongoing longitudinal study (ELISA). Latent profiles were identified first in a community sample (n = 2,103; age 4–8 years; 50.9% boys) and replicated in a high-CP subsample (n = 168; 70.24% boys).
Results
Four profiles emerged in the community sample (Normative Development, Daring/Impulsive, Low prosociality + Fear; Low prosociality + Psychopathic Traits [PP]), and three in the high-CP sample (same except the normative). The identified CP profiles aligned with the ADHD and CU domains, but not the ODD domain. Differences in activity, punitive and inconsistent parenting emerge as the most significant predictors. Regarding the outcomes, the Low prosociality + PP profile stands out as the group with the most severe emotional, social and behavioral maladjustment.
Conclusions
These findings highlight the heterogeneity within CP, and the importance of designing specific and tailored interventions for each identified profile.
Inorganic mercury (iHg) is an anthropogenic pollutant that forms monomethylmercury, a neurotoxicant affecting human health through seafood consumption. Despite iHg emissions reductions, the impact on oceanic concentrations remains unclear due to limited long-term data. Here, we present a four-year weekly time series of oceanic iHg concentrations at Scripps Pier in La Jolla, California, capturing interannual and seasonal variability. Interannual variability is driven by wet season precipitation, with wet years exhibiting sevenfold higher iHg concentration variance than dry years, potentially linking to El Niño-Southern Oscillation. Seasonally, precipitation and wave dynamics influence iHg inputs, with wet seasons driven by precipitation and runoff and dry seasons by upwelling. These parameters informed a model built to reconstruct a 20-year record of iHg concentrations, suggesting a long-term decline of 0.005 pM yr ⁻¹ due to climate-driven effects alone. This study highlights challenges in detecting long-term trends and emphasizes the need for sustained monitoring of oceanic iHg.
Background/Objectives: Clinical studies have shown a marked reduction in tumor control in prostate cancer treated with radically hypofractionated high-dose-rate brachytherapy (HDR-BT). The purpose of this study was to analyze the dose–response of prostate cancer treated with HDR-BT, specifically aiming at investigating the potential failure of the linear–quadratic (LQ) model to describe the response at large doses-per-fraction. Methods: We collated a dataset of dose–response to HDR-BT (3239 patients). The analysis was conducted separately for low and intermediate risk, resulting in 21 schedules (1633 patients) and 23 schedules (1606 patients), respectively. Data were fitted to tumor control probability models based on the LQ model, the linear–quadratic–linear (LQL), and a modification of the LQ model to include the effect of reoxygenation during treatment. Results: The LQ cannot fit the data unless the α/β is allowed to be high (∼[20, >100] Gy, 95% confidence interval). If the α/β is constrained to be low (≤8 Gy), the LQ model cannot reproduce the clinical results, and the LQL model, which includes a moderation of radiation damage with increasing dose, significantly improves the fitting. On the other hand, the reoxygenation model does not match the results obtained with the LQL. The clinically observed reduction in tumor control in prostate cancer treated with radical HDR-BT is better described by the LQL model. Using the best-fitting parameters, the BED for a 20 Gy × 1 treatment (128 Gyα/β) is far less than that of a conventional 2 Gy × 37 fractionation (196 Gyα/β). Conclusions: Our analysis showed that the substantial loss of tumor control observed in extremely hypofractionated HDR-BT trials can only be explained by the LQ model if the α/β is very large (≥100 Gy), in clear disagreement with the limits set in the analysis of external radiotherapy data. It seems more reasonable that there is a moderation of the LQ-predicted effect with increasing dose per fraction. These results may assist in the design of radical HDR-BT treatments.
Animals living in seasonal environments have adopted a wide array of tactics used to deal with seasonal resource scarcity. Many species migrate between habitats to reach areas where food resources are more plentiful as an attempt to address energetic demands through foraging. We assessed the winter behavioral adaptations of Caribou (Rangifer tarandus), a large ungulate inhabiting Arctic and sub-Arctic
regions known for seasonal resource scarcity. Movement rates of Caribou are the lowest of the year during winter, revealing 1 mechanism
individuals use to reduce energy expenditures. However, migratory individuals moved nearly twice as much as nonmigratory individuals
during winter, suggesting that migratory individuals rely more upon income (forage), whereas nonmigratory individuals rely more upon capital (bodily reserves). Lichens are the primary winter forage for large, migratory herds of Caribou, and migratory individuals experienced more than 2.5 times greater lichen cover than nonmigratory individuals. We documented that both groups slowed their movement in areas of greater lichen cover, suggesting increased foraging time in these areas. Movement rates were greater near villages, which may be suggestive of disturbance, but the effect was weak. Overall energy saved by reduced movement rates was modest. However, energy savings were 11% of daily body energy lost or 47% of the demands of early pregnancy, which potentially could affect individual condition and/or fetal growth if not offset by increases in forage intake.
Radioteleskope sind große Messinstrumente, die in der Geodäsie zur Interferometrie auf langen Basen eingesetzt werden. Ihre Messungen werden zur Bestimmung von Erdrotationsparametern, zum Detektieren von Veränderungen an der Erdkruste aber insbesondere auch zum Ableiten von sternenfesten und erdfesten Referenzrahmen eingesetzt. Radioteleskope empfangen emittierte Radiosignale von räumlich verteilten Quasaren. Ändert sich die Orientierung des Radioteleskops, kommt es aufgrund der Dimension und des Eigengewichts des Hauptreflektors zu schwerkraftbedingten Verformungen. Unberücksichtigt führen diese zu Signalwegvariationen und verfälschen abgeleitete Produkte systematisch. Die bisherige Modellierung von Hauptreflektorverformungen erfolgt häufig in Anlehnung an die geplante Sollgeometrie durch einfache rotationssymmetrische Paraboloide mit variierender Brenn-weite. Zur Verformungsanalyse des 26 m Radioteleskopes am Mount Pleasant Radio Observatory wurde die Oberfläche des Hauptreflektors in mehr als 200 Experimenten photogrammetrisch erfasst. Statt den Hauptreflektor lediglich durch ein Paraboloid zu approximieren, wird in diesem Beitrag ein innovativer Ansatz mit Zernike-Polynomen vorgeschlagen, der auch zonale Verformungen bei der Modellierung berücksichtigt. Bereits für Elevationswinkel von unter 60° liefern diese Zernike-Polynome gegenüber dem konventionellen Ansatz signifikant bessere Ergebnisse und beschreiben die auftretenden Verformungen zuverlässiger.
Urbanization in South Korea has significantly impacted the country’s sandy beach ecosystems. In our study, we investigated the population responses of sand bubbler crabs (Scopimera globosa) to beach urbanization and assessed the impact of adjacent terrestrial ecological quality. We employed the urbanization index to measure the urbanization levels of two sandy beaches and the remote-sensing ecological index (RSEI) to evaluate the ecological quality of adjacent terrestrial areas. The sampling of sand bubbler crab populations was conducted over five months. Our results show that urbanization significantly influences sand bubbler crab populations. While the ecological quality of adjacent terrestrial areas does not directly affect the crab populations, the land surface temperature (LST) of surrounding lands has a significant impact on sand bubbler crab biomass. These results suggest that sand bubbler crabs may serve as a useful indicator of anthropogenic disturbance on sandy beaches. This study provides critical ecological insights and offers a novel perspective for developing effective conservation strategies in South Korea’s sandy beach ecosystems.
Context Roads act as significant disruptors to wildlife movement through landscapes, even in relatively undisturbed areas. Large terrestrial vertebrates often choose to cross roads, which brings risk of injury or mortality from vehicle collision. In the Wet Tropics bioregion in Queensland, Australia, the range of the southern cassowary (Casuarius casuarius) spans a diversity of habitats and intensity of road network development. There has been a large increase in the human population within cassowary habitat in recent decades and at local scales there remains concern over increasing mortality caused by vehicle collision. Aims To inform management decisions on where is the best to direct collision mitigation resources based on identifying environmental factors that correlate with higher collision risk. Methods We collated 28 years of data on cassowary–vehicle collisions from across the bioregion and sought to identify ecological, physical and anthropogenic correlates of collision at both landscape- and location-specific scales. Key results We identified a major hotspot of reported cassowary–vehicle collisions in a large area of coastal lowland between Innisfail and Mission Beach. At a landscape scale, vehicle collisions increased significantly when the proportion of cassowary core habitat in the landscape was >50% and as the density of major roads increased. To a lesser extent, collisions increased when minor road density exceeded 400 m/km2. At a location-specific scale, vehicle collisions are more likely to occur on straight sections of major roads that have high canopy cover on both sides of the road and were less likely to occur on roads passing through land use types with low canopy cover. Conclusions Our findings highlight the significant challenges faced in effectively addressing the cassowary–vehicle collision problem in the Wet Tropics bioregion. Where cassowary habitat and roads intersect, cassowary mortality will occur. Implications Construction of new roads within rainforest should be avoided where possible and over- or under-passes should be constructed on new and existing major roads that intersect cassowary habitat and where location-specific correlates of cassowary collision exist. In addition, further collision mitigation measures such as traffic slowing should be targeted with these results in mind.
Wood-pastures are a type of agroforestry system consisting of assemblages of cattle-grazed pastures and scattered trees. They maintain biodiversity because of their environmental heterogeneity due to irregular tree cover that promotes diversity of resources and habitats. Agricultural intensification threatens the fragile balance of these traditional systems. Thus, we aimed to assess which factors associated with trees, shrubs and pasture heterogeneity within wood-pastures influence the community composition and diversity of different groups of animal species. We selected 45 wood-pasture sites of 20 ha along a gradient of tree density in the Swiss Jura mountains. In each site, we used remote sensing data and agricultural inventories to characterize the tree, shrub and herb layers’ structural and compositional heterogeneity. In study sites we also conducted survey of birds and hoverflies (Syrphidae) and evaluated the relationships between species composition and diversity with environmental variables. Tree density and tree spatial distribution were the main drivers of differences in community specific composition for the two groups of species. For hoverflies, botanically rich pastures and elevation were also driving differences in species composition between communities. Increased tree density showed a generally positive effect on birds and Syrphidae diversity indices. Moreover, for birds, scattered trees’ spatial distribution and a high percentage of deciduous trees also promoted species diversity. In this study, we highlight the importance of trees in pastures for bird and hoverfly diversity, as well as the necessity for trees to be heterogeneously distributed across the pastures and composed of a mix of deciduous and coniferous species.
Factorial invariance is critical for ensuring consistent relationships between measured variables and latent constructs across groups or time, enabling valid comparisons in social science research. Detecting factorial invariance becomes challenging when varying degrees of heterogeneity are present in the distribution of latent factors. This simulation study examined how changes in latent means and variances between groups influence the detection of noninvariance, comparing Bayesian and maximum likelihood fit measures. The design factors included sample size, noninvariance levels, and latent factor distributions. Results indicated that differences in factor variance have a stronger impact on measurement invariance than differences in factor means, with heterogeneity in latent variances more strongly affecting scalar invariance testing than metric invariance testing. Among model selection methods, goodness-of-fit indices generally exhibited lower power compared to likelihood ratio tests (LRTs), information criteria (ICs; except BIC), and leave-one-out cross-validation (LOO), which achieved a good balance between false and true positive rates.
ResearchGate has not been able to resolve any references for this publication.