Influence of habitat and climate on the spatial distribution of outbreaks of the Hylesia metabus moth, responsible for Lepidopterism, in coastal French Guiana

Influence of habitat and climate on the spatial distribution of outbreaks of the Hylesia metabus moth, responsible for Lepidopterism, in coastal French Guiana | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Influence of habitat and climate on the spatial distribution of outbreaks of the Hylesia metabus moth, responsible for Lepidopterism, in coastal French Guiana Raphaël Fougeray, Isaline Orhon, Manon Denux, Romane Ibanez, Romane Leseur, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7547284/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract Understanding the environmental context of insect outbreaks is crucial, particularly for pest species with significant impact on human health. The ashen moth Hylesia metabus is a generalist Lepidoptera with urticating scales ( setae ). This species causes severe dermatological reactions and its outbreaks pose serious public health challenges along the coastal regions of Venezuela and French Guiana. Despite the species' broad distribution throughout northern South America, outbreaks remain unpredictable and localized. Here, we explored factors that correlate with the recent spatial distribution of outbreaks by investigating 13 sites in French Guiana. We assessed forest structure, tree species composition, canopy cover and avian predation rates in the field. Additionally, we performed species distribution modeling to explore the effect of climate. Outbreak-prone sites were associated with overall low tree densities, high predation pressure, limited daily temperature variation, and pronounced seasonal changes between the dry and rain seasons. These conditions are more prevalent along the coast of French Guiana, contrasting sharply with the stable and diverse inland rainforest ecosystems where outbreaks are rarely reported. These findings highlight habitat features consistently associated with recent outbreak locations and provide a first step toward identifying ecological conditions that may influence outbreak propensity and can inform future monitoring strategies under changing environmental conditions. Ashen moth forest pest papillon cendre papillonite spatial heterogeneity species distribution model Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction A primary goal in insect ecology is to understand the drivers underlying spatial variability in the occurrence of insect outbreaks, particularly given their possible economic, ecological, and public health implications (Berryman 1987 ; Barbosa et al. 2012 ; Bradshaw et al. 2016 ). Spatial heterogeneity is known to profoundly influence the dynamics, severity, and frequency of insect population fluctuations (Turner 1989 ; Liebhold et al. 2004 ; Klemola et al. 2006 ), and this can be the result of abiotic factors, such as climate and habitat characteristics, and biotic factors, including host plant availability and predation pressure. Hence, understanding how ecological factors shape the spatial occurrence of insect outbreaks is crucial, particularly for pest species with significant impact on human health (Liebhold et al. 2004 ; Bradshaw et al. 2016 ). However, for widely distributed generalist species, predicting outbreaks remains difficult due to context-dependent responses to environmental variability and the limited availability of long-term monitoring data. But in the face of ongoing climate change, identifying the factors that determine why outbreaks occur only in certain regions or habitats, despite broad geographic distributions, becomes increasingly important as warming climates are expected to alter the frequency, intensity, and spatial distribution of insect outbreaks, exacerbating their impacts on ecosystems and societies (Raza et al. 2015; Lehmann et al. 2020 ). Furthermore, clarifying these ecological drivers can enable us to develop more effective management and mitigation strategies, ultimately reducing risks to public health and socioeconomic stability (Dent and Binks 2020 ). A notable example of a widely distributed, highly generalist insect pest is the ashen moth, Hylesia metabus (Cramer, 1775), which occurs across diverse habitats within the northern Amazon basin, including both forested and coastal regions. Within this broad geographic range, non-cyclical and unpredictable outbreaks occur specifically in coastal areas of the Guiana Shield, particularly in French Guiana and Venezuela (Lemaire 2002 ; Jourdain et al. 2012 ). Because H. metabus can cause severe cutaneous reactions, known as Lepidopterism (Leger and Mouzels 1918 ; Lamy and Lemaire 1983 ; Luce et al. 2023 ), these outbreaks necessitate frequent safety measures including curfews to reduce human interactions with the moths, and the use of large burners which attract and burn the moths, a method developed and used in the townships of Sinnamary and Iracoubo where outbreaks are especially problematic (CROPP 2015 ). Despite the health, social and economic importance of these outbreaks, the ecology of H. metabus remains poorly known and has rarely been the subject of in-depth studies. To understand why H. metabus outbreaks occur in some areas and not in others, and why some localities appear to be more prone to them, it is essential to identify the environmental factors that promote or constrain its population dynamics. Abiotic conditions, such as temperature, can directly affect larval growth and development (Davidson 1944 ; Ratte 1985 ), and influence the geographic range and phenology of the species (Kocsis and Hufnagel 2011 ; Checa et al. 2014 ). In turn, landscape structure can mediate the expression of these abiotic factors at local scales, for example through edge effects that alter microclimatic conditions (Didham and Lawton 1999 ). Landscape structure can also shape the movement and distribution of organisms through connectivity, thereby influencing population persistence (Turner 1989 ). Moreover, structural landscape features affect biotic interactions by determining community composition and spatial heterogeneity, which in turn can modulate the availability and quality of resources such as host plants (Pyke et al. 2001; LaPlante and Souza 2018), as well as predator community dynamics (Kareiva 1987 ). These interactions can regulate H. metabus populations through density-dependent effects, ultimately shaping the spatial distribution and potential for outbreaks of this species (Brown 1997; Dover and Settele 2008; Willems and Hill 2009). This study examines the interplay between abiotic (climatic conditions, landscape heterogeneity) and biotic (host plant composition, resource availability, predation pressure) factors in shaping H. metabus distribution and outbreak intensity in French Guiana. We hypothesize that spatial variation of recent outbreaks is driven by differences in the habitat. Specifically, we test how (1) habitat composition—including host plant availability, plant diversity, canopy cover, and predation pressure—as well as (2) climatic conditions and landscape structure influence H. metabus distribution and abundance, using recent outbreak localities as a proxy. Materials and methods Study area The study was carried out over 2 years (2020–2021) in French Guiana, an overseas department of France located on the northern coast of South America with a predominantly equatorial climate. Around 95% of French Guiana's area is occupied by dense rainforest with very low human population density. The coast of French Guiana, where the majority of the population is found, is instead composed of a mosaic of savanna, mangrove and coastal forest. It is on this coastal strip that moths of H. metabus can constitute a problem for public health. The most affected municipalities extend along the coastal road connecting Cayenne to Saint-Laurent-du-Maroni, via Kourou (Jourdain et al. 2012; Ciminera 2017). Coastal populations of H. metabus appear to differ from the forest populations (Ciminera et al. 2019), with both periodic and non-cyclical outbreaks rather than constant population cycles (Vassal 1989; Lemaire 2002), with some areas suffering from more frequent outbreaks of higher density (CNEV 2011; Jourdain et al. 2012; Ciminera 2017). Although no standardized data currently exists on moth population densities or outbreak severity, we used a qualitative outbreak classification based on three types of indirect sources: (1) local and regional news reports between 2006 and 2021 (Table S1), (2) municipal actions to mitigate risks to public health (e.g., use of moth burners, implementation of curfews), and (3) structured and informal discussions with local residents. While we acknowledge that such sources may be subject to biases, such as uneven media coverage, variation in public awareness, or differences in human population density, they nonetheless provide a first-order approximation of the perceived impact of H. metabus outbreaks at different sites. For this study, 13 sites (Fig. 1) were selected in order to cover a gradient of perceived outbreaks, and categorized into three propensity categories: (1) presence of the species but low number of outbreak reports (N = 4 sites); (2) moderate number of outbreak reports, with occasional outbreaks of mild effects on the human population (N = 6 sites); (3) high and recurrent reports of severe outbreaks, including widespread Lepidopterism and need for costly municipal actions (N = 3 sites). Habitat composition For each of the 13 sites, three line-intercept transects of 16 m were sampled. Line-intercept sampling is a type of transect whereby sampling quantifies the vegetation ( i.e. , in our case trees taller than 1 m) that intercepts a line segment (see Fig. S1). Each tree was identified to the highest taxonomic level, either directly in the field, or by collecting voucher specimens for comparison with reference material in the EcoFoG herbarium of Kourou. The circumference of each tree was also measured at a height of 1.3 m from the ground, and tree height and the area of intercept were estimated. This method is widely used in forest inventories and has been shown to produce an accurate general overview of the habitat studied (Caratti 2006). Transect accuracy improves with increasing length of transects used, but 16 m transects have been found to be suitable in dense vegetation (Fraver et al. 2018). Here the use of three transects of 16 m each, sufficiently spaced among them to avoid counting the same individual more than once and arranged randomly, allows to sample a total length of 48 m. Taxonomic data of tree species found was used to calculate three diversity indices per transect: species richness, Shannon diversity and Piélou evenness (using vegan R package, see electronic supplementary material, Methods for details). Moreover, the basal area (Ge), a relatively easily measured surrogate of total biomass (McElhinny et al. 2005), was calculated using the circumference (C) of each tree with the formula \(\:Ge=\frac{{C}^{2}}{4\pi\:}\). The total basal area for each transect was the sum of all the individual tree values, and the basal area specifically occupied by host plants within each transect was determined by summing the Ge values of all potential host plants (listed in Ciminera 2017 and references therein; see also Table S2). Total intercept length within a transect, and intercept length occupied by host plants, were the sum of intercept lengths of all trees and all host species intersecting with the transect, respectively (Fig. S1). In addition, canopy cover was measured by taking a photograph at each transect with a fish-eye lens (Samyang 7.5mm lens / Olympus OM-D E-M5). Images were converted into binary color schemes (sky in white and vegetation in black) using ImageJ (Schindelin et al. 2012) and the proportion of black pixels was calculated (see electronic supplementary material, Methods for details). Finally, for each site, habitat type was determined based on a census and forest habitat cartography by the French National Forest Office (Guitet et al. 2015). Predation pressure To assess predation rates, artificial clay caterpillars (3 cm long and 0.5 cm in diameter) resembling common cryptic palatable caterpillars (see e.g. , Lövei and Ferrante 2017 – Fig. S2) were made using light green plasticine (obtained by mixing modeling clay Van Aken yellow and plastilina Jovi green). They were attached with a thin metal wire (diameter 0.5 mm) on branches 1.5-2 m from the ground and at least 2 m apart. One hundred artificial clay caterpillars were placed at each site, evenly distributed over each transect (33 or 34 caterpillars per transect), for an overall total of 1300 caterpillars. The caterpillars were left at each site for approximately 72h. Predation marks on the malleable clay were categorized as either “avian” or “other/unknown” based on previous studies (see Eötvös and Lövei 2013; Low et al. 2014 – Fig. S3) and only attacks by avian predators were used for further analyses. Although invertebrates and other vertebrates can be potential predators of caterpillars, it is difficult to distinguish real predation attacks from territorial aggression (for example by ants) and clay consumption (Roels et al. 2018). Artificial caterpillars not recovered were scored as missing and excluded from further analyses (n = 26). As the data did not follow a normal distribution, a Kruskal-Wallis test was used to evaluate whether predation rates differed significantly among forest habitats using the rstatix R package (Kassambara 2020). Post-hoc pairwise comparisons were conducted using Dunn’s test (rstatix R package). Habitat comparison Given the categorical nature of the dependent variable (low, medium and high outbreak propensity), a spatial generalized additive model (sGAM) with ordinal response structure was done using the mgcv R package. To avoid multicollinearity, the variance inflation factor (VIF) was calculated for all predictor variables, and those with a VIF greater than 5 were excluded from further analysis (James et al. 2013), as were those with sampling inconsistencies (see Table S3). The following factors were kept as explanatory variables: tree species richness (S), Pielou's evenness index, predation rate, canopy cover (%), number of trees (as count data), amount of host plants (as count data) and total tree basal area (m²). To account for potential spatial autocorrelation in the data, a bivariate smooth term on geographic coordinates was included using a Gaussian process spline with a Matérn covariance structure, which allows for flexible modeling of spatial dependence between observations as a function of geographic distance (Apparicio et al. 2025). Model selection was performed by fitting all possible additive combinations of predictor variables and compared using the Akaike Information Criterion (AIC; Burnham and Anderson 2004). The best model (ΔAIC = 0) included predation rate, canopy cover, number of trees, and number of host plants as explanatory variables. Model residuals were evaluated using DHARMa and assumption were met with no evidence of overdispersion and a uniform distribution of residuals. Moreover, no significant spatial autocorrelation was detected (Moran’s I = -0.14, p = 0.9). Finally, a principal component analysis (PCA - FactoMineR R Package; Lê et al. 2008) was performed to visually assess the distribution of sites based on the variables identified as significant in the sGAM. The PCA biplot was colored by outbreak categories, and 95% confidence ellipses were added to illustrate the clustering of sites within each outbreak propensity category. Species distribution model To better understand how environmental conditions potentially influence the occurrence and abundance of H. metabus in French Guiana, species distribution modelling (SDM) was used with climate data collected between 1970 and 2000 (WorldClim 2 - Fick and Hijmans 2017). For comparison, models were also generated for three known host plants (Table S2) commonly found along the coastal strip and frequently reported as H. metabus caterpillar hosts. Two mangrove species, Avicennia germinans (black mangrove) and Laguncularia racemosa (white mangrove), were included due to their reported association with H. metabus in mangrove ecosystems (Lemaire 2002; Jourdain et al. 2012). Although Rhizophora mangle (red mangrove) is frequently reported as a primary host plant in Venezuela, in French Guiana, H. metabus seems to preferentially feed on A. germinans and L. racemosa (Jourdain et al. 2012). The third species, Tapirira guianensis , is a widespread tree found in both secondary and old-growth forests, selected because it was often reported as an important host by locals ( pers. obs. ). Furthermore, T. guianensis and A. germinans were commonly occurring host species in our transects (Table S2). Occurrence data of H. metabus (n = 65) were obtained from multiple sources, including the Barcode of Life Data System (BOLD: www.boldsystems.org; Ratnasingham and Hebert 2007), the French National Museum of Natural History (MNHN - Paris, France) and field sampling data compiled from Ciminera (2017). Occurrence data for host plants ( A. germinans : n = 63; L. racemosa : n = 41; T. guianensis : n = 63) were obtained from the Global Biodiversity Information Facility (GBIF). All spatial data were aligned using the WGS 84 (World Geodetic System 1984) - geographic coordinate system (EPSG:4326) at the 30 arc-seconds resolution (~ 1 km2) with the sf (Pebesma 2018) and raster (Hijmans 2023) R packages. Climate variables were selected using a Pearson correlation matrix, with a threshold of 0.7 to exclude collinear variables, followed by a variance inflation factor (VIF) analysis with a cutoff value of < 10 (Guisan et al. 2017). The climatic variables selected were: (1) Bio1: annual mean temperature; (2) Bio2: mean diurnal range; (3) Bio4: temperature seasonality, calculated as the standard deviation of monthly temperatures multiplied by 100, indicating the extent of temperature fluctuation across the year; (4) Bio18: precipitation during the warmest quarter; and (5) Bio19: precipitation during the coldest quarter. To minimize modeling bias due to uneven sampling, we refined the selection of pseudo-absence points ( i.e. , background points in MaxEnt) using the target-group background method (Phillips et al. 2009; Barber et al. 2022). This method involves generating a background dataset that accurately reflects sampling effort by creating a density map derived from the recorded occurrence of groups collected using similar sampling methods within the same geographic region. Specifically, a two-dimensional Kernel density estimation was applied, following Barber et al. (2022), with occurrence data from Saturniidae, Sphingidae, and Noctuidae moth families in French Guiana (obtained from the MNHN and GBIF; n = 5821; Fig. S4a). These families were chosen because their sampling methods—primarily light traps—match those used for collecting H. metabus . The resulting raster file, with the same spatial extent and grid resolution (30 arc-seconds) as the climatic variables, was rescaled to range from 1 to 20 following Elith et al. (2010) and assigned higher probabilities to background points drawn from areas with comparable sampling intensity to the H. metabus records. This approach effectively reduces the introduction of artifacts resulting from uneven survey efforts. The same approach was applied using Magnoliopsida occurrences (n = 75,262; Fig. S4b) to correct sampling bias in the host plants models. GBIF extractions (GBIF.org 2024) initially resulted in a total of 188,523 occurrences (moths and plants combined), which were then cleaned up by removing duplicates and excluding records with missing spatial coordinates or taxonomic identification using standard filtering procedures in R. Initially, optimal modeling settings (feature classes and regularization multiplier) for each species were determined using the ENMeval R package (Muscarella et al. 2014), selecting the model configuration with the lowest corrected Akaike Information Criterion (AICc). To ensure methodological consistency across species, we retained the same model settings (features = Linear + Quadratic, regularization multiplier = 1) for all final models (Table S4). This configuration showed a ΔAICc < 2 for all species except L. racemosa (ΔAICc = 2.38), which was nevertheless considered acceptable given the high predictive performance of the model (AUC = 0.94). These settings were subsequently used to build final species distribution models using MaxEnt software (Phillips et al. 2006) with a 10-fold cross-validation procedure. Finally, the predicted distribution of H. metabus was compared to those of its potential host plants through niche overlap analysis. Using the dismo R package (Hijmans et al. 2010), we calculated Schoener’s D and Hellinger-based I indices to quantify spatial similarity between species distribution models (Warren et al. 2008). Both indices range from 0 (no overlap) to 1 (complete overlap), but differ in their underlying assumptions and sensitivity. Schoener’s D measures absolute differences in predicted suitability across space, assuming that these values reflect relative habitat use. In contrast, the Hellinger-based I index treats model outputs strictly as probability distributions, without assuming a biological meaning, and is therefore more robust to extreme values and skewed predictions. By combining both metrics, we capture complementary aspects of niche similarity and minimize potential biases linked to any single interpretation of model outputs. We further assessed niche divergence between H. metabus and each host plant by performing both equivalency and similarity tests using the ENMTools R package (Warren et al. 2021). Both tests use Schoener’s D and Hellinger-based I indices to quantify niche overlap, but differ in their null hypotheses. The niche equivalency test evaluates whether H. metabus and its host plants occupy ecologically equivalent niches, by comparing the observed overlap to that expected under random allocation of occurrences. The niche similarity test assesses whether H. metabus ' niche is better predicted by the environmental distribution of its host plants than expected by chance, thus testing whether they share similar environments beyond spatial proximity. Together, these tests provide a statistical framework to evaluate whether niche similarity reflects true ecological overlap or arises from environmental availability alone. All models used in these tests were generated with Maxent (feature class = LQ, regularization multiplier = 1). Due to package limitations, cross-validation was not applied; instead, each test was based on 99 permutations per species. Results Habitat composition The 13 sites sampled differed in forest structure and composition (Table S4), but could be separated into four different categories (see Table 1 , Table S5) based on the forest habitat cartography of the French National Forest Office (ONF; Guitet et al. 2015 ). The first category is coastal forests, which include five of our sites, spanning all three risk levels of outbreak. Although consisting of many different forest habitats, all are strips of lowland forest (> 20 m in altitude) that extend inland > 40 km along the coastline. This habitat category is characterized by low species diversity, high density of small stems, and low forest cover (Guitet et al. 2015 ). Notably, amongst the tree species encountered, about half (49%) were potential host plants to H. metabus . Within the coastal forest category, the Mana and Larivot bridge sites were especially distinct. Mana is characterized by a white sand forest, with many endemic and rare species, generally uncommon elsewhere (Guitet et al. 2015 ). The Larivot bridge site consists of mangroves, which are periodically submerged by salty or brackish water, and have especially low diversity (it was the site with the lowest species richness) and tree species adapted to this unusual environment. Only two plant species were found in our Larivot transects, A. germinans and L. racemosa , and both are host plants for H. metabus . The second category is a mix of savanna and coastal forest. This included four of our sites and they had a moderate to severe risk for H. metabus outbreaks. These were a mix of stands of coastal forests and grassy woodland characterized by trees sufficiently widely spaced so that the canopy did not close. Canopy cover of our transects were in fact the lowest for this habitat (76% cover vs 87% on average for all sites) and the transects had the least number of trees counted (5.83 per transect vs the 7.92 overall mean), a large percentage (37%) of which were potential host plants. Table 1 Mean values (for the three transects) per site for selected ecological variables and diversity indices measured at 13 sites in French Guiana: values in bold indicate values greater than or equal to the median. The category outbreak propensity for each site is indicated in parentheses (low (1), medium (2) and high (3) outbreak propensity). ID Site Number of trees Number of host plants Proportion of host plants Canopy cover (%) Tree species richness (S) Forest habitat as per Guitet et al., 2015 1 Cacao (1) 11.33 2 0.16 95 9.33 Mid-altitude mountain forest 2 Mountain des singes (1) 12 0.33 0.05 89 8.33 Hill and valley forest 3 Mana (1) 6 2.67 0.48 94 4.67 Coastal forest (white sand forest) 4 Saint-Laurent-du-Maroni (1) 8.67 3 0.35 96 5.33 Coastal forest 5 Lake Petit-Saut (2) 13 0.33 0.04 95 5.33 Hill and valley forest 6 Kaw mountain (2) 8.67 3 0.33 94 6 Mid-altitude mountain forest 7 Roura (2) 7 4.67 0.61 79 3.67 Coastal forest 8 Bridge Larivot, Macouria (2) 5 5 1 88 2 Coastal forest (Mangroves) 9 Botanical garden, Macouria (2) 5.33 3.33 0.56 68 2.67 Savanna & coastal forest 10 Savanna of Matiti (2) 6 2.33 0.39 61 4.67 Savanna & coastal forest 11 Dégrad des Cannes (boat harbour, Rémire-Montjoly) (3) 8 2.33 0.31 97 4.33 Coastal forest 12 Iracoubo (3) 5.67 2.33 0.38 80 3 Savanna & coastal forest 13 Sinnamary (3) 6.33 1 0.15 95 5.33 Savanna & coastal forest Median 7 2.33 0.35 94 4.67 The third habitat category was the hill and valley forest, and it included two of our sites. These sites had fairly high tree density, and relatively abundant small stems. These forests are described as having dense undergrowth and mid to high canopy (30–35 m; Guitet et al. 2015 ). We found these sites to have a large number of trees (12.50 per transect vs the 7.92 overall mean) and a high species richness, but a very low percentage of these were potential hosts (4%). The fourth habitat category was the mid-altitudinal mountain forests and included two of our sites. These forests are described as having a high canopy (37 m) of irregular appearance, with high biomass, and where large trees are common and the undergrowth very diverse (Guitet et al. 2015 ). Similar to forests of hill and valleys, the mid-altitudinal forest had a great stem density from different species, but a large percentage of them (23%) were found to be potential host plants, although they were classified as of similarly low to moderate risk for outbreak. When comparing the different classes for outbreak propensity (Table 2 ), the sites most likely to suffer from frequent outbreaks (group 3, n = 3) were characterized by low tree count, whereas groups least likely to have outbreaks (group 1, n = 4) were characterized by high tree count and high diversity. Surprisingly, the highest proportion of host plants (43%) was found at sites of moderate risk (group 2, n = 6), although this difference appears to be solely due to the mangrove site where only two species, both potential hosts, were found. Table 2 Mean values for some of the ecological variables and diversity indices measured at 13 sites in French Guiana per category of outbreak propensity (low (1), medium (2) and high (3) outbreak propensity). Outbreak groups Number of sites Number of trees Number of host plants Proportion of host plants Canopy cover (%) Tree species richness (S) Proportion of savannas & coastal forests 1 4 9.50 2 0.26 93.44 6.92 0.5 2 6 7.50 3.11 0.49 80.92 4.06 0.67 3 3 6.67 1.89 0.28 90.76 4.22 1 Predation pressure Of the 1,284 artificial caterpillars retrieved, 291 exhibited signs of predation, of which only 85 could be attributed to avian attacks, resulting in an overall average avian predation rate of 0.07. Due to the generally low predation rates, differences between sites were moderate, ranging from 0 to 0.17 (Table S5). Moreover, the Kruskal-Wallis test did not detect a significant overall effect of habitat type on predation rates (Statistic = 6.25, p = 0.1). However, post-hoc Dunn’s test comparisons revealed significant differences between mid-altitude mountain forests and both coastal forests and savannas (statistic = -2.37, p = 0.02, statistic = -2.04, p = 0.04, respectively; Fig. 2 ), with predation rates being highest in mid-altitude forests (mean = 0.11). In contrast, predation rates were lowest in savannas (mean = 0.06) and coastal forests (mean = 0.04), although two sampled coastal forest sites exhibited marked variation (0.12 in white sand forests vs. 0 in mangroves). Predation rates in hill and valley forests were intermediate between those observed in coastal (savanna and coastal forest) and mid-altitude mountain forest, with an average predation rate of 0.09. Habitat comparison Spatial generalized additive model (sGAM) was fitted to account for the ordinal nature of the site categories for outbreak propensity, while controlling for spatial autocorrelation. The analysis revealed that only the predation rate, the canopy cover (%) and the number of trees were significantly associated with the potential for outbreak. Outbreak propensity increased with a higher predation rate (estimate = 177 ± 52, z = 3.38, p < 0.01), whereas it decreased with greater canopy cover (estimate = -0.30 ± 0.14, z = -2.17, p = 0.03) and with the number of trees (estimate = -3.94 ± 1.35, z = -2.91, p < 0.01). The number of host plants did not have a significant effect ( p = 0.63). The spatial smoothing term was highly significant ( edf = 5.77, χ² = 331.4, p < 0.001), suggesting a strong spatial structure in outbreak patterns. The final model explained 99.4% of the deviance. As such, sites associated with frequent outbreaks are characterized by habitats with low tree density (i.e. both reduced tree count and reduced canopy cover) and high predation rate (Fig. 3 , see supplementary material for details). Species distribution model All species distribution models (SDMs) for H. metabus and its host plants show reasonable or high performance with AUC values greater than 0.7 ( T. guianensis ) or 0.9 ( H. metabus , A. germinans , L. racemosa ) as shown in Table 3 (Swets 1988 ; Peterson et al. 2011 ). The models showed a higher probability for the presence of the moth H. metabus along the coastline (Fig. 4 a). Although H. metabus has been found to occur inland, these habitats appear less suitable for the species based on both collecting data and the distribution model. The areas with the highest probability of occurrence are situated along the coast between Iracoubo and a little past Cayenne and Rémire-Montjoly. These areas were also classified as being at moderate to high risk of outbreaks based on our classification. Although some occurrence is predicted on the eastern coast, around Saint-Laurent-du-Maroni and Mana, we have found this region to actually be at low risk. The climatic variable that contributed the most to the predicted distribution of H. metabus was the mean diurnal temperature range (PI = 88.9). The host plants A. germinans and L. racemosa were also predicted to occur along the coast (Fig. 4 b,c), consistent with their ecological restriction to mangrove habitats. Their distributions were also primarily influenced by the mean diurnal temperature range (PI = 79.3 and 69.8, respectively), with additional contributions from the annual mean temperature for A. germinans (PI = 11.7) and the precipitation of the warmest quarter for L. racemosa (PI = 26). The host plant T. guianensis had a broader predicted range (Fig. 4 d), with its distribution primarily associated with the mean diurnal temperature range (PI = 64.8) and precipitation of the warmest quarter (PI = 31.3). The highest similarity for spatial overlap was found with L. racemosa ( D = 0.84; I = 0.98), followed by T. guianensis ( D = 0.60; I = 0.89) and A. germinans ( D = 0.61; I = 0.87). These values indicate moderate to high spatial overlap between the moth’s predicted distribution and those of its potential host plants, especially L. racemosa , which shows near-complete similarity based on the I index (Table 3 ). Table 3 Results of SDMs with permutation importance (PI) of each variable, for the moth Hylesia metabus and three of its common host plants; Avicennia germinans, Laguncularia racemosa, Tapirira guianensis. H. metabus A. germinans L. racemosa T. guianensis AUC 0.92 0.97 0.93 0.76 Bio1 - Annual Mean Temperature 0.5 11.7 3.3 1 Bio2 - Mean Diurnal Range 88.9 79.3 69.8 64.8 Bio4 - Temperature Seasonality 0 1.2 0.5 0.3 Bio18 - Precipitation of Warmest Quarter 10 4.9 26 31.3 Bio19 - Precipitation of Coldest Quarter 0.5 3 0.5 2.7 Spatial overlap with the H. metabus model Hellinger's I 1 0.87 0.98 0.89 Schoener’s D 1 0.61 0.84 0.60 However, niche equivalency and niche similarity tests (Fig. 5 ) revealed important distinctions. The niche equivalency test did not detect significant differences between the niches of H. metabus and L. racemosa (p-values > 0.05), suggesting that their ecological niches may be equivalent. In contrast, the niches of H. metabus and A. germinans , as well as T. guianensis , were significantly different (p = 0.01 for both D and I), indicating that their niches are not statistically equivalent. Niche similarity tests indicated significant niche similarity between H. metabus and all three host plants (p = 0.01 for D and I in each case), suggesting that despite non-equivalent niches, the moth consistently occupies environments similar to those of its hosts. Discussion Habitat heterogeneity can affect where insect pest outbreaks occur (Turner 1989 ; Liebhold et al. 2004 ; Klemola et al. 2006 ), and understanding how these ecological factors shape these occurrences is crucial for both predicting outbreak prone sites, lessening their impacts, and guiding future monitoring efforts. Here we attempted to determine the biotic and abiotic factors that may explain the spatial differences in the non-cyclical outbreaks of H. metabus in French Guiana, a moth responsible for Lepidopterism. Sites categorized as experiencing more frequent or severe outbreaks were found to be characterized by lower tree densities (i.e. reduced number of trees and overall canopy cover) and high avian predation rate. These sites were more likely to occur in habitats consisting of coastal forests (including, but not limited to, mangroves) and savannas, which were generally correlated with low diversity and low overall tree count, of which a fairly high proportion were host plants. These habitats were also more likely to have pioneer plant species, many of which are potential hosts ( e.g. , Cecropia obtusa , T. guianensis ), although host plant density was not correlated with outbreak propensity, suggesting that this may not be a limiting factor. This is also consistent with other studies that have found that polyphagous insects like H. metabus prefer and/or do better in fragmented habitats of low spatial heterogeneity and complexity (Tscharntke et al. 2002 ; Benedick et al. 2006 ). Although surprising that outbreak-prone sites were correlated with high avian predation rate, it is important to note that overall predation rates were low across sites, and may have correlated more with habitat types than outbreak risk; for example, outbreak-prone sites of coastal forests had relatively high predation rates, but others like the mangroves and savannas did not. It is also worth noting that larvae of H. metabus have urticating spines, which may protect them from most avian predators. As such, avian predation rates may instead reflect other characteristics such as habitat fragmentation and edge effects that could be correlated with outbreak propensity (Kareiva 1987 ; Baggio et al. 2010 ; Batáry et al. 2014 ). For instance, in the forest tent caterpillar Malacosoma disstria , the sensitivity of outbreak duration to changes in forest structure has been hypothesized to reflect reduced effectiveness of natural enemies such as parasitoids and pathogens in fragmented forests (Roland 1993 ; Roland and Kauppp 1995 ; Roland and Taylor 1997 ). Although the parasitoids and diseases of H. metabus caterpillars are mostly unknown, future investigations should compare mortality rate and causes between coastal and inland habitats, as these may also explain differences in their population dynamics (Stamp and Bowers 1988 ; Zhou 2009 ). As for climatic conditions, species distribution modeling (SDM) identified mean diurnal temperature range and precipitation during the warmest quarter as key predictors of H. metabus occurrence. In particular, the coast of French Guiana, characterized by more of a tropical monsoon climate with a distinct dry season and minimal diurnal temperature fluctuations, was predicted to be the most favorable. In contrast, the inland forests, where outbreaks have rarely been reported, are characterized by a tropical rainforest climate, with greater climatic stability and less pronounced seasonality (Beck et al. 2023 ). This pattern likely reflects the role of climatic seasonality in H. metabus phenology, as shown in both temperate and tropical systems (Wolda 1988 ; Neuvonen et al. 1999 ; Nelson et al. 2013 ; Ward et al. 2019 ; Büntgen et al. 2020 ). Insects such as H. metabus may exploit predictable seasonal cues, such as rainfall onset, to synchronize development and reproduction. Although the three host plants used for the distribution models are known to be common and frequently exploited by H. metabus (Jourdain et al. 2012 ), only the spatial distribution of L. racemosa closely matched that of the moth. Nevertheless, our results revealed significant ecological similarities between the niches occupied by H. metabus and its three host plants, suggesting that the moth exploits a broader ecological niche than would be inferred solely from spatial overlap. This finding indicates that H. metabus likely occupies habitats ecologically analogous to multiple host species, even when their geographic co-occurrence is limited, suggesting ecological flexibility in the species allowing it to efficiently use resources distributed heterogeneously in space. For specialist insects (Opedal et al. 2020 ; Wilson et al. 2021 ), host plant availability can be an important driver of herbivore dynamics. However, for the polyphagous H. metabus , overall abundance of host plants was not found to be correlated with outbreak-prone sites. It is possible that the moth prefers habitats where a combination of host plants is present, or even sites with preferred or more suitable host plants. Nevertheless, H. metabus appears to consistently occupy environments that are similar to some of its host plants, implying a shared preference for certain climatic conditions. While the availability and distribution of host plants likely play a key role in shaping the moth’s range, climatic variables appear to be even more influential in determining its overall distribution and outbreak propensity. As global warming and increased land-use change continue to modify habitats, the potential future expansion or contraction of H. metabus’ range warrants further investigation (Lehmann et al. 2020 ). In conclusion, we show that sites with reported H. metabus outbreaks are mostly associated with habitats marked by overall low tree densities, minimal diurnal temperature fluctuations and distinct seasonality ( i.e. , marked dry and wet season). These characteristics are more likely to occur on the coast, and contrast markedly with the seasonally stable ( i.e. , with minimal differences between seasons) and diverse inland rainforest ecosystems, where outbreaks have rarely been reported. Avian predation, although correlated with outbreak occurrences, may instead reflect broader landscape structure such as fragmentation and edge effects. Future studies should also investigate the potential effect of other natural enemies, such as parasitoids and disease, which may be more negatively affected by habitat types and fragmentation, and which is more common along the coast, as well as host plant preferences and suitability. Targeted monitoring of both H. metabus populations and their natural enemies, combined with landscape management strategies that preserve forest complexity and reduce fragmentation, may help mitigate potential future risks. Long-term studies integrating ecological, climatic, and epidemiological data may be necessary to fully understand and anticipate the drivers of pest outbreak emergence in tropical systems. Declarations Conflict of Interest The authors have no relevant financial or non-financial interests to disclose. Ethical Approval This article does not contain any studies with human participants or animals performed by the authors. Funding This study was funded by a grant from the “mission pour les initiatives transverses et interdisciplinaires” (MITI) from the CNRS to MM & MA. This work also benefited from an “Investissement d’avenir” grant managed by the Center for the study of biodiversity in Amazonia (CEBA) (ANR-10-LABX-25-01) to MM. Acknowledgments We thank the Mission pour les Initiatives Transverses et Interdisciplinaires (MITI) of the CNRS for supporting this study through a grant awarded to MM and MA . We also acknowledge the support of the Investissement d’Avenir grant managed by the Center for the Study of Biodiversity in Amazonia (CEBA, ANR-10-LABX-25-01) awarded to MM . We thank Sandra Ianez, Rémi Mauxion, Pierre Lacoste, Léo-Paul Charlet and Guillaume Correa Pimpao for their precious help in the field. We would also like to thank Marina Ciminera for providing us with the Hylesia metabus occurrence data she collected during her PhD and Frédéric Bénéluz for all the information he provided. References Apparicio Philippe, Jérémy Gelb, Jean Dubé et Joan Carles Martori (2025). Méthodes de régression spatiale : un grand bol d’R . Université Laval et Université de Sherbrooke. fabriqueREL. Licence CC BY-SA. Baggio JA, Salau K, Janssen MA, Schoon ML, Bodin Ö (2010) Landscape connectivity and predator–prey population dynamics. Landsc Ecol 26(1):33–45. https://doi.org/10.1007/s10980-010-9493-y Barber RA, Ball SG, Morris RKA, Gilbert F (2022) Target-group backgrounds prove effective at correcting sampling bias in Maxent models. Divers Distrib 28(1):128–141. https://doi.org/10.1111/ddi.13442 Barbosa P, Letourneau DK, Agrawal AA (2012) Insect Outbreaks Revisited. John Wiley & Sons Bartoń K (2024) MuMIn: Multi-Model Inference Batáry P, Fronczek S, Normann C, Scherber C, Tscharntke T (2014) How do edge effect and tree species diversity change bird diversity and avian nest survival in Germany’s largest deciduous forest? For. Ecol. Manag. 319:44–50 Beck HE, McVicar TR, Vergopolan N, Berg A, Lutsko NJ, Dufour A, Zeng Z, Jiang X, van Dijk AIJM, Miralles DG (2023) High-resolution (1 km) Köppen-Geiger maps for 1901–2099 based on constrained CMIP6 projections. Sci Data 10(1):724. https://doi.org/10.1038/s41597-023-02549-6 Benedick S, Hill J, Mustaffa N, Chey V, Maryati M, Searle J, Schilthuizen M, Hamer K (2006) Impacts of rain forest fragmentation on butterflies in northern Borneo: species richness, turnover and the value of small fragments. J. Appl. Ecol. 43:967–977 Berryman AA (1987) Chapter 1 - The Theory and Classification of Outbreaks. In: Barbosa P, Schultz JC (eds) Insect Outbreaks. Academic Press, San Diego, pp 3–30 Bradshaw CJA, Leroy B, Bellard C, Roiz D, Albert C, Fournier A, Barbet-Massin M, Salles J-M, Simard F, Courchamp F (2016) Massive yet grossly underestimated global costs of invasive insects. Nat Commun 7(1):12986. https://doi.org/10.1038/ncomms12986 Büntgen U, Liebhold A, Nievergelt D, Wermelinger B, Roques A, Reinig F, Krusic PJ, Piermattei A, Egli S, Cherubini P, Esper J (2020) Return of the moth: rethinking the effect of climate on insect outbreaks. Oecologia 192(2):543–552. https://doi.org/10.1007/s00442-019-04585-9 Burnham KP, Anderson DR (eds) (2004) Model Selection and Multimodel Inference. Springer New York, New York, NY Caratti JF (2006) Line Intercept (LI). Lutes Duncan C Keane Robert E Caratti John F Key Carl H Benson Nathan C Sutherl. Steve Gangi Larry J 2006 FIREMON Fire Eff. Monit. Inventory Syst. Gen Tech Rep RMRS-GTR-164-CD Fort Collins CO US Dep. Agric. For. Serv. Rocky Mt. Res. Stn. P LI-1-13 164 Checa M, Rodriguez J, Willmott K, Liger B (2014) Microclimate Variability Significantly Affects the Composition, Abundance and Phenology of Butterfly Communities in a Highly Threatened Neotropical Dry Forest. Fla Entomol 97:1–13. https://doi.org/10.1653/024.097.0101 Ciminera M (2017) Identification spécifique et structure génétique des populations du papillon-cendre responsable des épisodes de papillonite en Guyane et au Vénézuela. Thesis Ciminera M, Auger-Rozenberg M-A, Caron H, Herrera M, Scotti-Saintagne C, Scotti I, Tysklind N, Roques A (2019) Genetic variation and differentiation of Hylesia metabus (Lepidoptera: Saturniidae): moths of public health importance in French Guiana and in Venezuela. J. Med. Entomol. 56:137–148 CNEV (2011) Réponse à la saisine “Stratégies et méthodes de lutte optimales contre Hylesia metabus”, agent de la papillonite en Guyane française. Avis à l’attention de la Direction Générale de la Santé. Centre National D’Expertise sur les Vecteurs, Montpellier CROPP (2015) La papillonite - Bulletin d’information n°10 Août 2015, Cellule régionale permanente d’observation et de prévention de la papillonite en Guyane. Davidson J (1944) On the Relationship between Temperature and Rate of Development of Insects at Constant Temperatures. J Anim Ecol 13(1):26–38. https://doi.org/10.2307/1326 Dent D, Binks RH (2020) Insect Pest Management, 3rd Edition. CABI Didham RK, Lawton JH (1999) Edge Structure Determines the Magnitude of Changes in Microclimate and Vegetation Structure in Tropical Forest Fragments. Biotropica 31(1):17–30. https://doi.org/10.1111/j.1744-7429.1999.tb00113.x Elith J, Kearney M, Phillips S (2010) The art of modelling range-shifting species. Methods Ecol Evol 1(4):330–342. https://doi.org/10.1111/j.2041-210X.2010.00036.x Eötvös C, Lövei GL (2013) Documenting predator marks on dummy caterpillars. In: Saska P, Knapp M, Honek A, Martinkova Z (eds) XVIth European Carabidologists Meeting - Book of Abstracts with Conference Programme: Carabids and man - can we live with(out) each other? European Carabidologists Meeting, p 31 Fick SE, Hijmans RJ (2017) WorldClim 2: new 1-km spatial resolution climate surfaces for global land areas. Int J Climatol 37(12):4302–4315. https://doi.org/10.1002/joc.5086 Fraver S, Ducey MJ, Woodall CW, D’Amato AW, Milo AM, Palik BJ (2018) Influence of transect length and downed woody debris abundance on precision of the line-intersect sampling method. For. Ecosyst. 5:1–10 GBIF.org (2024) GBIF Occurrence Download https://doi.org/10.15468/DL.28K84S Guisan A, Thuiller W, Zimmermann NE (2017) Habitat Suitability and Distribution Models: with Applications in R. Cambridge University Press Guitet S, Euriot S, Brunaux O, Baraloto C, Denis T, Dewynter M, Freycon V, Gonzales S, Jaouen G, Hansen CR (2015) Catalogue des habitats forestiers de Guyane Hijmans RJ (2023) raster: Geographic Data Analysis and Modeling Hijmans RJ, Phillips S, Leathwick J, Elith J (2010) dismo: Species Distribution Modeling. 1.3-16 Ianez S, Palisse M, Clerc-Renaud A (2021) Le quotidien avec Hylesia metabus (Cramer, 1789). Cohabiter avec le papillon cendre dans les communes du littoral guyanais. [Unpublished manuscript]. James G, Witten D, Hastie T, Tibshirani R (2013) An introduction to statistical learning. Springer Jourdain F, Girod R, Vassal JM, Chandre F, Lagneau C, Fouque F, Guiral D, Raude J, Robert V (2012) The moth Hylesia metabus and French Guiana lepidopterism: centenary of a public health concern. Parasite J Société Fr Parasitol 19(2):117–128. https://doi.org/10.1051/parasite/2012192117 Kareiva P (1987) Habitat fragmentation and the stability of predator–prey interactions. Nature 326(6111):388–390. https://doi.org/10.1038/326388a0 Kassambara A (2020) Pipe-Friendly Framework for Basic Statistical Tests [R package rstatix version 0.7.2] Klemola T, Huitu O, Ruohomäki K (2006) Geographically partitioned spatial synchrony among cyclic moth populations. Oikos 114(2):349–359. https://doi.org/10.1111/j.2006.0030-1299.14850.x Kocsis M, Hufnagel L (2011) Impacts of climate change on Lepidoptera species and communities. Appl Ecol Environ Res 9:43–72. https://doi.org/10.15666/aeer/0901_043072 Lamy M, Lemaire C (1983) Contribution à la systématique des Hylesia : étude au microscope électronique à balayage des «fléchettes » urticantes [Lep. Saturniidae]. Bull Société Entomol Fr :176–192 Lê S, Josse J, Husson F (2008) FactoMineR: A Package for Multivariate Analysis. J Stat Softw 25(1):1–18. https://doi.org/10.18637/jss.v025.i01 Leger M, Mouzels P (1918) Dermatose prurigineuse déterminée par des papillons Saturnides du genre Hylesia. Bull Soc Path Exot 11:104–107 Lehmann P, Ammunét T, Barton M, Battisti A, Eigenbrode SD, Jepsen JU, Kalinkat G, Neuvonen S, Niemelä P, Terblanche JS, Økland B, Björkman C (2020) Complex responses of global insect pests to climate warming. Front Ecol Environ 18(3):141–150. https://doi.org/10.1002/fee.2160 Lemaire C (2002) Saturniidae of America: Hemileucinae. Geocke & Evers Liebhold A, Koenig WD, Bjørnstad ON (2004) Spatial Synchrony in Population Dynamics. Annu Rev Ecol Evol Syst 35:467–490 Lövei GL, Ferrante M (2017) A review of the sentinel prey method as a way of quantifying invertebrate predation under field conditions. Insect Sci. 24:528–542 Low PA, Sam K, McArthur C, Posa MRC, Hochuli DF (2014) Determining predator identity from attack marks left in model caterpillars: guidelines for best practice. Entomol. Exp. Appl. 152:120–126 Luce AE, Couppie P, Michaud C, Clauteaux P, Blaizot R (2023) Épidémiologie, clinique et thérapeutique de la papillonite en Guyane, 2017–2023. Ann Dermatol Vénéréologie - FMC 3(8):A257. https://doi.org/10.1016/j.fander.2023.09.426 McElhinny C, Gibbons P, Brack C, Bauhus J (2005) Forest and woodland stand structural complexity: Its definition and measurement. For Ecol Manag 218(1):1–24. https://doi.org/10.1016/j.foreco.2005.08.034 Muscarella R, Galante PJ, Soley-Guardia M, Boria RA, Kass JM, Uriarte M, Anderson RP (2014) ENMeval: An R package for conducting spatially independent evaluations and estimating optimal model complexity for Maxent ecological niche models. Methods Ecol Evol 5(11):1198–1205. https://doi.org/10.1111/2041-210X.12261 Nelson WA, Bjørnstad ON, Yamanaka T (2013) Recurrent Insect Outbreaks Caused by Temperature-Driven Changes in System Stability. Science 341(6147):796–799. https://doi.org/10.1126/science.1238477 Neuvonen S, Niemelä P, Virtanen T (1999) Climatic Change and Insect Outbreaks in Boreal Forests: The Role of Winter Temperatures. Ecol Bull (47):63–67 Opedal ØH, Ovaskainen O, Saastamoinen M, Laine A-L, van Nouhuys S (2020) Host-plant availability drives the spatiotemporal dynamics of interacting metapopulations across a fragmented landscape. Ecology 101(12):e03186. https://doi.org/10.1002/ecy.3186 Pebesma E (2018) Simple Features for R: Standardized Support for Spatial Vector Data. R J 10(1):439–446. https://doi.org/10.32614/RJ-2018-009 Peterson AT, Soberón J, Pearson RG, Anderson RP, Martínez-Meyer E, Nakamura M, Araújo MB (2011) Ecological Niches and Geographic Distributions. Princeton University Press Phillips SJ, Anderson RP, Schapire RE (2006) Maximum entropy modeling of species geographic distributions. Ecol Model 190(3):231–259. https://doi.org/10.1016/j.ecolmodel.2005.03.026 Phillips SJ, Dudík M, Elith J, Graham CH, Lehmann A, Leathwick J, Ferrier S (2009) Sample selection bias and presence-only distribution models: implications for background and pseudo-absence data. Ecol Appl 19(1):181–197. https://doi.org/10.1890/07-2153.1 Ratnasingham S, Hebert PDN (2007) bold: The Barcode of Life Data System (http://www.barcodinglife.org). Mol Ecol Notes 7(3):355–364. https://doi.org/10.1111/j.1471-8286.2007.01678.x Ratte HT (1985) Temperature and Insect Development. In: Hoffmann KH (ed) Environmental Physiology and Biochemistry of Insects. Springer Berlin Heidelberg, Berlin, Heidelberg, pp 33–66 Raza MM, Khan ,Muhammad Aslam, Arshad ,Muhammad, Sagheer ,Muhammad, Sattar ,Zeeshan, Shafi ,Jamil, Haq ,Ehtisham ul, Ali ,Asim, Aslam ,Usman, Mushtaq ,Aleena, Ishfaq ,Iqra, Sabir ,Zarnab, and Sattar A (2015) Impact of global warming on insects. Arch Phytopathol Plant Prot 48(1):84–94. https://doi.org/10.1080/03235408.2014.882132 Roels SM, Porter JL, Lindell CA (2018) Predation pressure by birds and arthropods on herbivorous insects affected by tropical forest restoration strategy. Restor. Ecol. 26:1203–1211 Roland J (1993) Large-scale forest fragmentation increases the duration of tent caterpillar outbreak. Oecologia 93(1):25–30. https://doi.org/10.1007/BF00321186 Roland J, Kauppp WJ (1995) Reduced Transmission of Forest Tent Caterpillar (Lepidoptera: Lasiocampidae) Nuclear Polyhedrosis Virus at the Forest Edge. Environ Entomol 24(5):1175–1178. https://doi.org/10.1093/ee/24.5.1175 Roland J, Taylor PD (1997) Insect parasitoid species respond to forest structure at different spatial scales. Nature 386(6626):710–713. https://doi.org/10.1038/386710a0 Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, Preibisch S, Rueden C, Saalfeld S, Schmid B (2012) Fiji: an open-source platform for biological-image analysis. Nat. Methods 9:676–682 Stamp N, Bowers M (1988) Direct and indirect effects of predatory wasps (Polistes sp.: Vespidae) on gregarious caterpillars (Hemileuca lucina: Saturniidae). Oecologia 75:619–624 Swets JA (1988) Measuring the Accuracy of Diagnostic Systems. Science 240(4857):1285–1293. https://doi.org/10.1126/science.3287615 Tscharntke T, Steffan-Dewenter I, Kruess A, Thies C (2002) Contribution of Small Habitat Fragments to Conservation of Insect Communities of Grassland–Cropland Landscapes. Ecol Appl 12(2):354–363. https://doi.org/10.1890/1051-0761(2002)012[0354:COSHFT]2.0.CO;2 Turner MG (1989) Landscape ecology: the effect of pattern on process. Annu Rev Ecol Syst 20(1):171–197 Vassal J-M (1989) Biologie, écologie et pathologie d’Hylesia metabus (Cramer 1775) (Lépidoptères : Saturniidae), agent de la papillonite en Guyane Française : mise en place d’une structure de lutte intégrée. Thesis, USTL Venables WN, Ripley BD (2002) Modern Applied Statistics with S, Fourth. Springer, New York Ward SF, Moon RD, Aukema BH (2019) Implications of seasonal and annual heat accumulation for population dynamics of an invasive defoliator. Oecologia 190(3):703–714. https://doi.org/10.1007/s00442-019-04431-y Warren DL, Glor RE, Turelli M (2008) ENVIRONMENTAL NICHE EQUIVALENCY VERSUS CONSERVATISM: QUANTITATIVE APPROACHES TO NICHE EVOLUTION. Evolution 62(11):2868–2883. https://doi.org/10.1111/j.1558-5646.2008.00482.x Warren DL, Matzke NJ, Cardillo M, Baumgartner JB, Beaumont LJ, Turelli M, Glor RE, Huron NA, Simões M, Iglesias TL, Piquet JC, Dinnage R (2021) ENMTools 1.0: an R package for comparative ecological biogeography. Ecography 44(4):504–511. https://doi.org/10.1111/ecog.05485 Wilson JK, Casajus N, Hutchinson RA, McFarland KP, Kerr JT, Berteaux D, Larrivée M, Prudic KL (2021) Climate Change and Local Host Availability Drive the Northern Range Boundary in the Rapid Expansion of a Specialist Insect Herbivore, Papilio cresphontes. Front Ecol Evol Volume 9-2021 Wolda H (1988) Insect Seasonality: Why? Annu Rev Ecol Syst 19:1–18 Zhou Z (2009) A review on control of tobacco caterpillar, Spodoptera litura. Chin. Bull. Entomol. 46:354–361 Supplementary Files Supplementary.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 07 Oct, 2025 Reviewers invited by journal 28 Sep, 2025 Editor invited by journal 25 Sep, 2025 Editor assigned by journal 23 Sep, 2025 First submitted to journal 18 Sep, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7547284","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":521849426,"identity":"e01433f8-f3f3-438b-bf31-6207ff9ea772","order_by":0,"name":"Raphaël Fougeray","email":"data:image/png;base64,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","orcid":"https://orcid.org/0009-0002-8657-2386","institution":"Laboratoire Écologie, Évolution, Interactions des Systèmes Amazoniens (LEEISA), Université de Guyane, CNRS, IFREMER, 97300 Cayenne, France","correspondingAuthor":true,"prefix":"","firstName":"Raphaël","middleName":"","lastName":"Fougeray","suffix":""},{"id":521849427,"identity":"7d7bc2e9-1730-48ec-bc80-70e42ffbbfc1","order_by":1,"name":"Isaline Orhon","email":"","orcid":"","institution":"Laboratoire Écologie, Évolution, Interactions des Systèmes Amazoniens (LEEISA), Université de Guyane, CNRS, IFREMER, 97300 Cayenne, France","correspondingAuthor":false,"prefix":"","firstName":"Isaline","middleName":"","lastName":"Orhon","suffix":""},{"id":521849428,"identity":"06dd10a6-953d-44d4-b091-9389bdbfe6a1","order_by":2,"name":"Manon Denux","email":"","orcid":"","institution":"Laboratoire Écologie, Évolution, Interactions des Systèmes Amazoniens (LEEISA), Université de Guyane, CNRS, IFREMER, 97300 Cayenne, France","correspondingAuthor":false,"prefix":"","firstName":"Manon","middleName":"","lastName":"Denux","suffix":""},{"id":521849429,"identity":"02ec3ddb-a8c5-4070-a8d5-d26fedc933cb","order_by":3,"name":"Romane Ibanez","email":"","orcid":"","institution":"Laboratoire Écologie, Évolution, Interactions des Systèmes Amazoniens (LEEISA), Université de Guyane, CNRS, IFREMER, 97300 Cayenne, France","correspondingAuthor":false,"prefix":"","firstName":"Romane","middleName":"","lastName":"Ibanez","suffix":""},{"id":521849430,"identity":"d6636e3a-4640-4289-9462-8bfd52e6e293","order_by":4,"name":"Romane Leseur","email":"","orcid":"","institution":"Laboratoire Écologie, Évolution, Interactions des Systèmes Amazoniens (LEEISA), Université de Guyane, CNRS, IFREMER, 97300 Cayenne, France","correspondingAuthor":false,"prefix":"","firstName":"Romane","middleName":"","lastName":"Leseur","suffix":""},{"id":521849431,"identity":"39b805af-14b4-4523-bf5a-ad216aab4377","order_by":5,"name":"Liliana Ballesteros-Meija","email":"","orcid":"","institution":"DGD-REVE Direction Générale Déléguée à la Recherche, à l'Expertise, à la Valorisation et à l'Enseignement-Formation; CESAB, Centre de Synthèse et d'Analyse sur la Biodiversité, Montpellier, France","correspondingAuthor":false,"prefix":"","firstName":"Liliana","middleName":"","lastName":"Ballesteros-Meija","suffix":""},{"id":521849432,"identity":"e40af36c-c9d4-4590-a057-6df168f0fdbf","order_by":6,"name":"Rodolphe Rougerie","email":"","orcid":"","institution":"Institut de Systématique, Évolution, Biodiversité (ISYEB), CNRS, MNHN, EPHE, Sorbonne Université, Université des Antilles, Paris, France","correspondingAuthor":false,"prefix":"","firstName":"Rodolphe","middleName":"","lastName":"Rougerie","suffix":""},{"id":521849433,"identity":"40865436-c427-4c31-ac1a-58c02446d4c2","order_by":7,"name":"Giacomo Sellan","email":"","orcid":"","institution":"CIRAD, UMR EcoFoG (AgroPariTech, CNRS, INRAE, Université des Antilles, Université de Guyane), Campus Agronomique, Kourou, France","correspondingAuthor":false,"prefix":"","firstName":"Giacomo","middleName":"","lastName":"Sellan","suffix":""},{"id":521849434,"identity":"3f18b640-f6ce-4c66-b9ef-04a2c7c59c2b","order_by":8,"name":"Yi Moua","email":"","orcid":"","institution":"Espace-Dev, IRD, Univ Montpellier, Univ Guyane, Univ La Réunion, Univ Antilles, Univ Nouvelle Calédonie, Montpellier, France; Espace-Dev, Univ Guyane, Cayenne / Kourou, Guyane française, France","correspondingAuthor":false,"prefix":"","firstName":"Yi","middleName":"","lastName":"Moua","suffix":""},{"id":521849435,"identity":"9acb663d-f499-45f8-89ba-e7fac3c1361e","order_by":9,"name":"Mónica Arias","email":"","orcid":"","institution":"PHIM, Univ Montpellier, CIRAD, INRAE, Institut Agro, IRD, CEDEX 5, 34398 Montpellier, France","correspondingAuthor":false,"prefix":"","firstName":"Mónica","middleName":"","lastName":"Arias","suffix":""},{"id":521849436,"identity":"ead926cf-88f0-498f-a720-121651d5da3e","order_by":10,"name":"Melanie McClure","email":"","orcid":"https://orcid.org/0000-0003-3590-4002","institution":"Laboratoire Écologie, Évolution, Interactions des Systèmes Amazoniens (LEEISA), Université de Guyane, CNRS, IFREMER, 97300 Cayenne, France","correspondingAuthor":false,"prefix":"","firstName":"Melanie","middleName":"","lastName":"McClure","suffix":""}],"badges":[],"createdAt":"2025-09-05 22:08:25","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7547284/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7547284/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":93165383,"identity":"31f64366-21a3-431c-b759-52d9ea195d55","added_by":"auto","created_at":"2025-10-09 17:49:13","extension":"xml","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13421,"visible":true,"origin":"","legend":"","description":"","filename":"nentNENTD2500417.xml","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/ceb75ab82a9c4d985b122cb8.xml"},{"id":93165385,"identity":"2a905a69-0940-4bd3-bc25-83a3694ed45e","added_by":"auto","created_at":"2025-10-09 17:49:13","extension":"xml","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1116,"visible":true,"origin":"","legend":"","description":"","filename":"NENTD250041711650.go.xml","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/6ed1e16fdbece9fae1aac5b4.xml"},{"id":93165390,"identity":"57321a5d-c535-4fde-a8fa-607138b81aff","added_by":"auto","created_at":"2025-10-09 17:49:13","extension":"xml","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":927,"visible":true,"origin":"","legend":"","description":"","filename":"NENTD2500417Import.xml","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/b8263c2a58afbcb31d6c01be.xml"},{"id":93165399,"identity":"ebd9a57f-76b9-4b06-bfd3-2ea31e5c6634","added_by":"auto","created_at":"2025-10-09 17:49:13","extension":"xml","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":174573,"visible":true,"origin":"","legend":"","description":"","filename":"NENTD25004170enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/0afdba854138fc3cbf5cd57d.xml"},{"id":93165395,"identity":"08e63e94-e818-4ca7-a60d-b7029522ebea","added_by":"auto","created_at":"2025-10-09 17:49:13","extension":"pdf","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":201335,"visible":true,"origin":"","legend":"","description":"","filename":"Fig1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/d8974021de3bef06ad1c1c4f.pdf"},{"id":93166013,"identity":"2347afcb-6714-4330-b577-05daf155d988","added_by":"auto","created_at":"2025-10-09 17:57:13","extension":"pdf","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":47390,"visible":true,"origin":"","legend":"","description":"","filename":"Fig2.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/dbc479fe3f70852fe09a7832.pdf"},{"id":93166018,"identity":"f6c3542b-c4cc-4b80-abba-ae097bcb82ae","added_by":"auto","created_at":"2025-10-09 17:57:13","extension":"pdf","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":546123,"visible":true,"origin":"","legend":"","description":"","filename":"Fig3.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/82ab323512268505e72b0269.pdf"},{"id":93165398,"identity":"41a3b042-92fc-49f6-95a4-ec3a9df62ce6","added_by":"auto","created_at":"2025-10-09 17:49:13","extension":"pdf","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1903399,"visible":true,"origin":"","legend":"","description":"","filename":"Fig4.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/2fd00b576899608b58a6c018.pdf"},{"id":93166024,"identity":"f1501603-ca5b-4a88-9318-11c267c77ebc","added_by":"auto","created_at":"2025-10-09 17:57:13","extension":"pdf","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":364376,"visible":true,"origin":"","legend":"","description":"","filename":"Fig5.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/5709243e7bb3085c7937c492.pdf"},{"id":93167503,"identity":"c1b12c15-dc76-4b94-8b35-eb00afa29df9","added_by":"auto","created_at":"2025-10-09 18:21:13","extension":"jpeg","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":298098,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/82d006a714e79ff8606ed8e1.jpeg"},{"id":93166015,"identity":"d02a4c75-35fe-43dd-8197-88f545c92e1f","added_by":"auto","created_at":"2025-10-09 17:57:13","extension":"jpeg","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":256164,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/074e9e361ece998a7e67ef76.jpeg"},{"id":93166021,"identity":"395e51e2-7a22-46cf-879a-70ab3c0aa0de","added_by":"auto","created_at":"2025-10-09 17:57:13","extension":"jpeg","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":370372,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/9a94aa60087c5004a50b9049.jpeg"},{"id":93166294,"identity":"116ec0dc-368b-45de-9af1-02e13a6b6f35","added_by":"auto","created_at":"2025-10-09 18:05:13","extension":"jpeg","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1063816,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/422ab7f0e531ff350d77d07e.jpeg"},{"id":93166014,"identity":"543fee6c-7a04-4683-b351-f1d75c7c0895","added_by":"auto","created_at":"2025-10-09 17:57:13","extension":"jpeg","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":539570,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/8f634183d8724738fc0b9def.jpeg"},{"id":93165387,"identity":"33f5a8a3-d483-4474-9a93-bd063425e774","added_by":"auto","created_at":"2025-10-09 17:49:13","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":93333,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/2a25dd70059e13b30217a4a0.png"},{"id":93165401,"identity":"51371773-8be8-4bfd-bdbc-c61b61f37e33","added_by":"auto","created_at":"2025-10-09 17:49:13","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":62208,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/d6904a39aa0ff788d142de2f.png"},{"id":93166932,"identity":"6990049c-8faf-4cd4-9ecb-240cd352f819","added_by":"auto","created_at":"2025-10-09 18:13:13","extension":"png","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":90849,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/7d73d7ea7717561afc014989.png"},{"id":93166935,"identity":"f2ac6dd7-8da1-43dd-9a05-6ee0957c5eb2","added_by":"auto","created_at":"2025-10-09 18:13:13","extension":"png","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":306844,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/a5115c0e9bf07883461684f6.png"},{"id":93166297,"identity":"668502de-2389-48b1-a09e-695f617d569a","added_by":"auto","created_at":"2025-10-09 18:05:13","extension":"png","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":124358,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/2b0e383d06d49e93c440d02b.png"},{"id":93166025,"identity":"9363b85d-3d08-4189-bee4-4fdf00a69547","added_by":"auto","created_at":"2025-10-09 17:57:14","extension":"xml","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":175290,"visible":true,"origin":"","legend":"","description":"","filename":"NENTD25004170structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/5240e74331047ee27e3246c5.xml"},{"id":93165404,"identity":"15b7c58a-7d20-437e-af12-79a466ef62db","added_by":"auto","created_at":"2025-10-09 17:49:14","extension":"html","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":189226,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/cbd33623617c85b51c396fa8.html"},{"id":93166012,"identity":"3f2a7ead-70d7-4874-89f7-e6990fe086d9","added_by":"auto","created_at":"2025-10-09 17:57:13","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":137902,"visible":true,"origin":"","legend":"\u003cp\u003eThe spatial distribution of the 13 study sites; colors indicate categories of likelihood for outbreak of \u003cem\u003eHylesia metabus \u003c/em\u003eand negative human impact, based on news coverage, municipal records and information from the public. Green indicates species presence but low outbreak reports number (N=4; Saint-Laurent-du-Maroni, Mana, Mountain des Singes (Kourou), Cacao), yellow indicates moderate number of outbreak reports, with occasional outbreaks of mild effects on the human population (N=6; Dam \u0026amp; Lake Petit-Saut (Sinnamary), Savanna of Matiti (Macouria), Botanical garden of Macouria, Larivot bridge (Macouria), Roura, Kaw Mountain), and red indicates high and recurrent reports of severe outbreaks, including widespread Lepidopterism and need for costly municipal actions (N=3; Iracoubo, Sinnamary, Dégrad des Cannes (boat harbour, Rémire-Montjoly))\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/1f0718ed14c1328c2a06a6cd.png"},{"id":93166011,"identity":"f164e2e2-b7d3-46ee-9e86-7e78ba2f9567","added_by":"auto","created_at":"2025-10-09 17:57:13","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":63255,"visible":true,"origin":"","legend":"\u003cp\u003eDifferences in predation rates between savanna (N=4 sites), coastal forest (N=5 sites), hill and valley forest (N=2 sites), and mid-altitude mountain forest (N=2 sites). Coastal (light grey, including savannas and coastal forests) and inland (dark grey, including hill and valley forests and mid-altitude mountain forests) habitats are statistically different, as per the post-hoc Dunn’s test (α, αβ and β)\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/69ded8412ef20d8663261fd5.png"},{"id":93165378,"identity":"a855a9c4-02ee-43f1-a493-56d73e371736","added_by":"auto","created_at":"2025-10-09 17:49:13","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":148967,"visible":true,"origin":"","legend":"\u003cp\u003ePrincipal component analysis (PCA) grouping sites as a function of outbreak propensity, and showing the effect of predation rate, number of trees, and canopy cover.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/d25cfc3979953a405689dbd1.png"},{"id":93165380,"identity":"1ea53f9c-81d7-4c8b-a80b-fff67a9a698d","added_by":"auto","created_at":"2025-10-09 17:49:13","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":335492,"visible":true,"origin":"","legend":"\u003cp\u003eSpecies distribution model for a) \u003cem\u003eHylesia metabus\u003c/em\u003e, b) \u003cem\u003eAvicennia germinans\u003c/em\u003e, c) \u003cem\u003eLaguncularia racemosa\u003c/em\u003e, d) \u003cem\u003eTapirira guianensis, \u003c/em\u003ewith likelihood ranging from 0 (blue) to 1 (yellow). The mean diurnal temperature range from WorldClim (Fick and Hijmans 2017) is shown in e), ranging from 7 (orange) to 10 (purple)\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/3b617ddf12470e89536531bf.png"},{"id":93165382,"identity":"db63904d-a637-4ee9-b310-132957a7f2a6","added_by":"auto","created_at":"2025-10-09 17:49:13","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":141929,"visible":true,"origin":"","legend":"\u003cp\u003eNiche equivalency and similarity tests between \u003cem\u003eHylesia metabus\u003c/em\u003e and its host plants. Each panel shows the observed niche overlap between \u003cem\u003eH. metabus\u003c/em\u003e and a host plant (dashed line) compared to the distribution of overlap values from 99 random permutations (pink histograms). Tests were conducted using Schoener’s D (upper plots per species, D) and Hellinger-based I indices (lower plots per species, I). Asterisks (*) indicate significant differences for the niche equivalency test; and significantly higher similarity than expected under the null model for the niche similarity test (p \u0026lt; 0.05). a) \u003cem\u003eAvicennia germinans\u003c/em\u003e; b) \u003cem\u003eLaguncularia racemosa\u003c/em\u003e; c) \u003cem\u003eTapirira guianensis\u003c/em\u003e\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/b7503a6667271387aa77ab1a.png"},{"id":93225462,"identity":"89c0f91a-a1b3-4133-ada9-191781342db1","added_by":"auto","created_at":"2025-10-10 12:21:12","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1706676,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/622127cb-3e50-4e41-88d8-37108fd3b87f.pdf"},{"id":93166933,"identity":"b795d8ad-8c82-4799-9b93-993743373bc7","added_by":"auto","created_at":"2025-10-09 18:13:13","extension":"docx","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":4456562,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementary.docx","url":"https://assets-eu.researchsquare.com/files/rs-7547284/v1/bf4e3c0dd36c5d1d48556738.docx"}],"financialInterests":"","formattedTitle":"Influence of habitat and climate on the spatial distribution of outbreaks of the Hylesia metabus moth, responsible for Lepidopterism, in coastal French Guiana","fulltext":[{"header":"Introduction","content":"\u003cp\u003eA primary goal in insect ecology is to understand the drivers underlying spatial variability in the occurrence of insect outbreaks, particularly given their possible economic, ecological, and public health implications (Berryman \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e1987\u003c/span\u003e; Barbosa et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Bradshaw et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Spatial heterogeneity is known to profoundly influence the dynamics, severity, and frequency of insect population fluctuations (Turner \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e1989\u003c/span\u003e; Liebhold et al. \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2004\u003c/span\u003e; Klemola et al. \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2006\u003c/span\u003e), and this can be the result of abiotic factors, such as climate and habitat characteristics, and biotic factors, including host plant availability and predation pressure. Hence, understanding how ecological factors shape the spatial occurrence of insect outbreaks is crucial, particularly for pest species with significant impact on human health (Liebhold et al. \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2004\u003c/span\u003e; Bradshaw et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eHowever, for widely distributed generalist species, predicting outbreaks remains difficult due to context-dependent responses to environmental variability and the limited availability of long-term monitoring data. But in the face of ongoing climate change, identifying the factors that determine why outbreaks occur only in certain regions or habitats, despite broad geographic distributions, becomes increasingly important as warming climates are expected to alter the frequency, intensity, and spatial distribution of insect outbreaks, exacerbating their impacts on ecosystems and societies (Raza et al. 2015; Lehmann et al. \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Furthermore, clarifying these ecological drivers can enable us to develop more effective management and mitigation strategies, ultimately reducing risks to public health and socioeconomic stability (Dent and Binks \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eA notable example of a widely distributed, highly generalist insect pest is the ashen moth, \u003cem\u003eHylesia metabus\u003c/em\u003e (Cramer, 1775), which occurs across diverse habitats within the northern Amazon basin, including both forested and coastal regions. Within this broad geographic range, non-cyclical and unpredictable outbreaks occur specifically in coastal areas of the Guiana Shield, particularly in French Guiana and Venezuela (Lemaire \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2002\u003c/span\u003e; Jourdain et al. \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Because \u003cem\u003eH. metabus\u003c/em\u003e can cause severe cutaneous reactions, known as Lepidopterism (Leger and Mouzels \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e1918\u003c/span\u003e; Lamy and Lemaire \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e1983\u003c/span\u003e; Luce et al. \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), these outbreaks necessitate frequent safety measures including curfews to reduce human interactions with the moths, and the use of large burners which attract and burn the moths, a method developed and used in the townships of Sinnamary and Iracoubo where outbreaks are especially problematic (CROPP \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Despite the health, social and economic importance of these outbreaks, the ecology of \u003cem\u003eH. metabus\u003c/em\u003e remains poorly known and has rarely been the subject of in-depth studies.\u003c/p\u003e\u003cp\u003eTo understand why \u003cem\u003eH. metabus\u003c/em\u003e outbreaks occur in some areas and not in others, and why some localities appear to be more prone to them, it is essential to identify the environmental factors that promote or constrain its population dynamics. Abiotic conditions, such as temperature, can directly affect larval growth and development (Davidson \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e1944\u003c/span\u003e; Ratte \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e1985\u003c/span\u003e), and influence the geographic range and phenology of the species (Kocsis and Hufnagel \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; Checa et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). In turn, landscape structure can mediate the expression of these abiotic factors at local scales, for example through edge effects that alter microclimatic conditions (Didham and Lawton \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e1999\u003c/span\u003e). Landscape structure can also shape the movement and distribution of organisms through connectivity, thereby influencing population persistence (Turner \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e1989\u003c/span\u003e). Moreover, structural landscape features affect biotic interactions by determining community composition and spatial heterogeneity, which in turn can modulate the availability and quality of resources such as host plants (Pyke et al. 2001; LaPlante and Souza 2018), as well as predator community dynamics (Kareiva \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e1987\u003c/span\u003e). These interactions can regulate \u003cem\u003eH. metabus\u003c/em\u003e populations through density-dependent effects, ultimately shaping the spatial distribution and potential for outbreaks of this species (Brown 1997; Dover and Settele 2008; Willems and Hill 2009).\u003c/p\u003e\u003cp\u003eThis study examines the interplay between abiotic (climatic conditions, landscape heterogeneity) and biotic (host plant composition, resource availability, predation pressure) factors in shaping \u003cem\u003eH. metabus\u003c/em\u003e distribution and outbreak intensity in French Guiana. We hypothesize that spatial variation of recent outbreaks is driven by differences in the habitat. Specifically, we test how (1) habitat composition\u0026mdash;including host plant availability, plant diversity, canopy cover, and predation pressure\u0026mdash;as well as (2) climatic conditions and landscape structure influence \u003cem\u003eH. metabus\u003c/em\u003e distribution and abundance, using recent outbreak localities as a proxy.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cp\u003eStudy area\u003c/p\u003e\n\u003cp\u003eThe study was carried out over 2 years (2020\u0026ndash;2021) in French Guiana, an overseas department of France located on the northern coast of South America with a predominantly equatorial climate. Around 95% of French Guiana\u0026apos;s area is occupied by dense rainforest with very low human population density. The coast of French Guiana, where the majority of the population is found, is instead composed of a mosaic of savanna, mangrove and coastal forest. It is on this coastal strip that moths of \u003cem\u003eH. metabus\u003c/em\u003e can constitute a problem for public health. The most affected municipalities extend along the coastal road connecting Cayenne to Saint-Laurent-du-Maroni, via Kourou (Jourdain et al. 2012; Ciminera 2017). Coastal populations of \u003cem\u003eH. metabus\u003c/em\u003e appear to differ from the forest populations (Ciminera et al. 2019), with both periodic and non-cyclical outbreaks rather than constant population cycles (Vassal 1989; Lemaire 2002), with some areas suffering from more frequent outbreaks of higher density (CNEV 2011; Jourdain et al. 2012; Ciminera 2017). Although no standardized data currently exists on moth population densities or outbreak severity, we used a qualitative outbreak classification based on three types of indirect sources: (1) local and regional news reports between 2006 and 2021 (Table S1), (2) municipal actions to mitigate risks to public health (e.g., use of moth burners, implementation of curfews), and (3) structured and informal discussions with local residents. While we acknowledge that such sources may be subject to biases, such as uneven media coverage, variation in public awareness, or differences in human population density, they nonetheless provide a first-order approximation of the perceived impact of \u003cem\u003eH. metabus\u003c/em\u003e outbreaks at different sites. For this study, 13 sites (Fig.\u0026nbsp;1) were selected in order to cover a gradient of perceived outbreaks, and categorized into three propensity categories: (1) presence of the species but low number of outbreak reports (N\u0026thinsp;=\u0026thinsp;4 sites); (2) moderate number of outbreak reports, with occasional outbreaks of mild effects on the human population (N\u0026thinsp;=\u0026thinsp;6 sites); (3) high and recurrent reports of severe outbreaks, including widespread Lepidopterism and need for costly municipal actions (N\u0026thinsp;=\u0026thinsp;3 sites).\u003c/p\u003e\n\u003cp\u003eHabitat composition\u003c/p\u003e\n\u003cp\u003eFor each of the 13 sites, three line-intercept transects of 16 m were sampled. Line-intercept sampling is a type of transect whereby sampling quantifies the vegetation (\u003cem\u003ei.e.\u003c/em\u003e, in our case trees taller than 1 m) that intercepts a line segment (see Fig. S1). Each tree was identified to the highest taxonomic level, either directly in the field, or by collecting voucher specimens for comparison with reference material in the EcoFoG herbarium of Kourou. The circumference of each tree was also measured at a height of 1.3 m from the ground, and tree height and the area of intercept were estimated. This method is widely used in forest inventories and has been shown to produce an accurate general overview of the habitat studied (Caratti 2006). Transect accuracy improves with increasing length of transects used, but 16 m transects have been found to be suitable in dense vegetation (Fraver et al. 2018). Here the use of three transects of 16 m each, sufficiently spaced among them to avoid counting the same individual more than once and arranged randomly, allows to sample a total length of 48 m.\u003c/p\u003e\n\u003cp\u003eTaxonomic data of tree species found was used to calculate three diversity indices per transect: species richness, Shannon diversity and Pi\u0026eacute;lou evenness (using vegan R package, see electronic supplementary material, Methods for details). Moreover, the basal area (Ge), a relatively easily measured surrogate of total biomass (McElhinny et al. 2005), was calculated using the circumference (C) of each tree with the formula \\(\\:Ge=\\frac{{C}^{2}}{4\\pi\\:}\\). The total basal area for each transect was the sum of all the individual tree values, and the basal area specifically occupied by host plants within each transect was determined by summing the Ge values of all potential host plants (listed in Ciminera 2017 and references therein; see also Table S2). Total intercept length within a transect, and intercept length occupied by host plants, were the sum of intercept lengths of all trees and all host species intersecting with the transect, respectively (Fig. S1). In addition, canopy cover was measured by taking a photograph at each transect with a fish-eye lens (Samyang 7.5mm lens / Olympus OM-D E-M5). Images were converted into binary color schemes (sky in white and vegetation in black) using ImageJ (Schindelin et al. 2012) and the proportion of black pixels was calculated (see electronic supplementary material, Methods for details). Finally, for each site, habitat type was determined based on a census and forest habitat cartography by the French National Forest Office (Guitet et al. 2015).\u003c/p\u003e\n\u003cp\u003ePredation pressure\u003c/p\u003e\n\u003cp\u003eTo assess predation rates, artificial clay caterpillars (3 cm long and 0.5 cm in diameter) resembling common cryptic palatable caterpillars (see \u003cem\u003ee.g.\u003c/em\u003e, L\u0026ouml;vei and Ferrante 2017 \u0026ndash; Fig. S2) were made using light green plasticine (obtained by mixing modeling clay Van Aken yellow and plastilina Jovi green). They were attached with a thin metal wire (diameter 0.5 mm) on branches 1.5-2 m from the ground and at least 2 m apart. One hundred artificial clay caterpillars were placed at each site, evenly distributed over each transect (33 or 34 caterpillars per transect), for an overall total of 1300 caterpillars. The caterpillars were left at each site for approximately 72h. Predation marks on the malleable clay were categorized as either \u0026ldquo;avian\u0026rdquo; or \u0026ldquo;other/unknown\u0026rdquo; based on previous studies (see E\u0026ouml;tv\u0026ouml;s and L\u0026ouml;vei 2013; Low et al. 2014 \u0026ndash; Fig. S3) and only attacks by avian predators were used for further analyses. Although invertebrates and other vertebrates can be potential predators of caterpillars, it is difficult to distinguish real predation attacks from territorial aggression (for example by ants) and clay consumption (Roels et al. 2018). Artificial caterpillars not recovered were scored as missing and excluded from further analyses (n\u0026thinsp;=\u0026thinsp;26). As the data did not follow a normal distribution, a Kruskal-Wallis test was used to evaluate whether predation rates differed significantly among forest habitats using the rstatix R package (Kassambara 2020). Post-hoc pairwise comparisons were conducted using Dunn\u0026rsquo;s test (rstatix R package).\u003c/p\u003e\n\u003cp\u003eHabitat comparison\u003c/p\u003e\n\u003cp\u003eGiven the categorical nature of the dependent variable (low, medium and high outbreak propensity), a spatial generalized additive model (sGAM) with ordinal response structure was done using the mgcv R package. To avoid multicollinearity, the variance inflation factor (VIF) was calculated for all predictor variables, and those with a VIF greater than 5 were excluded from further analysis (James et al. 2013), as were those with sampling inconsistencies (see Table S3). The following factors were kept as explanatory variables: tree species richness (S), Pielou\u0026apos;s evenness index, predation rate, canopy cover (%), number of trees (as count data), amount of host plants (as count data) and total tree basal area (m\u0026sup2;). To account for potential spatial autocorrelation in the data, a bivariate smooth term on geographic coordinates was included using a Gaussian process spline with a Mat\u0026eacute;rn covariance structure, which allows for flexible modeling of spatial dependence between observations as a function of geographic distance (Apparicio et al. 2025). Model selection was performed by fitting all possible additive combinations of predictor variables and compared using the Akaike Information Criterion (AIC; Burnham and Anderson 2004). The best model (\u0026Delta;AIC\u0026thinsp;=\u0026thinsp;0) included predation rate, canopy cover, number of trees, and number of host plants as explanatory variables. Model residuals were evaluated using DHARMa and assumption were met with no evidence of overdispersion and a uniform distribution of residuals. Moreover, no significant spatial autocorrelation was detected (Moran\u0026rsquo;s I = -0.14, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.9). Finally, a principal component analysis (PCA - FactoMineR R Package; L\u0026ecirc; et al. 2008) was performed to visually assess the distribution of sites based on the variables identified as significant in the sGAM. The PCA biplot was colored by outbreak categories, and 95% confidence ellipses were added to illustrate the clustering of sites within each outbreak propensity category.\u003c/p\u003e\n\u003cp\u003eSpecies distribution model\u003c/p\u003e\n\u003cp\u003eTo better understand how environmental conditions potentially influence the occurrence and abundance of \u003cem\u003eH. metabus\u003c/em\u003e in French Guiana, species distribution modelling (SDM) was used with climate data collected between 1970 and 2000 (WorldClim 2 - Fick and Hijmans 2017). For comparison, models were also generated for three known host plants (Table S2) commonly found along the coastal strip and frequently reported as \u003cem\u003eH. metabus\u003c/em\u003e caterpillar hosts. Two mangrove species, \u003cem\u003eAvicennia germinans\u003c/em\u003e (black mangrove) and \u003cem\u003eLaguncularia racemosa\u003c/em\u003e (white mangrove), were included due to their reported association with \u003cem\u003eH. metabus\u003c/em\u003e in mangrove ecosystems (Lemaire 2002; Jourdain et al. 2012). Although \u003cem\u003eRhizophora mangle\u003c/em\u003e (red mangrove) is frequently reported as a primary host plant in Venezuela, in French Guiana, \u003cem\u003eH. metabus\u003c/em\u003e seems to preferentially feed on \u003cem\u003eA. germinans\u003c/em\u003e and \u003cem\u003eL. racemosa\u003c/em\u003e (Jourdain et al. 2012). The third species, \u003cem\u003eTapirira guianensis\u003c/em\u003e, is a widespread tree found in both secondary and old-growth forests, selected because it was often reported as an important host by locals (\u003cem\u003epers. obs.\u003c/em\u003e). Furthermore, \u003cem\u003eT. guianensis\u003c/em\u003e and \u003cem\u003eA. germinans\u003c/em\u003e were commonly occurring host species in our transects (Table S2). Occurrence data of \u003cem\u003eH. metabus\u003c/em\u003e (n\u0026thinsp;=\u0026thinsp;65) were obtained from multiple sources, including the Barcode of Life Data System (BOLD: www.boldsystems.org; Ratnasingham and Hebert 2007), the French National Museum of Natural History (MNHN - Paris, France) and field sampling data compiled from Ciminera (2017). Occurrence data for host plants (\u003cem\u003eA. germinans\u003c/em\u003e: n\u0026thinsp;=\u0026thinsp;63; \u003cem\u003eL. racemosa\u003c/em\u003e: n\u0026thinsp;=\u0026thinsp;41; \u003cem\u003eT. guianensis\u003c/em\u003e: n\u0026thinsp;=\u0026thinsp;63) were obtained from the Global Biodiversity Information Facility (GBIF). All spatial data were aligned using the WGS 84 (World Geodetic System 1984) - geographic coordinate system (EPSG:4326) at the 30 arc-seconds resolution (~\u0026thinsp;1 km2) with the sf (Pebesma 2018) and raster (Hijmans 2023) R packages.\u003c/p\u003e\n\u003cp\u003eClimate variables were selected using a Pearson correlation matrix, with a threshold of 0.7 to exclude collinear variables, followed by a variance inflation factor (VIF) analysis with a cutoff value of \u0026lt;\u0026thinsp;10 (Guisan et al. 2017). The climatic variables selected were: (1) Bio1: annual mean temperature; (2) Bio2: mean diurnal range; (3) Bio4: temperature seasonality, calculated as the standard deviation of monthly temperatures multiplied by 100, indicating the extent of temperature fluctuation across the year; (4) Bio18: precipitation during the warmest quarter; and (5) Bio19: precipitation during the coldest quarter. To minimize modeling bias due to uneven sampling, we refined the selection of pseudo-absence points (\u003cem\u003ei.e.\u003c/em\u003e, background points in MaxEnt) using the target-group background method (Phillips et al. 2009; Barber et al. 2022). This method involves generating a background dataset that accurately reflects sampling effort by creating a density map derived from the recorded occurrence of groups collected using similar sampling methods within the same geographic region. Specifically, a two-dimensional Kernel density estimation was applied, following Barber et al. (2022), with occurrence data from Saturniidae, Sphingidae, and Noctuidae moth families in French Guiana (obtained from the MNHN and GBIF; n\u0026thinsp;=\u0026thinsp;5821; Fig. S4a). These families were chosen because their sampling methods\u0026mdash;primarily light traps\u0026mdash;match those used for collecting \u003cem\u003eH. metabus\u003c/em\u003e. The resulting raster file, with the same spatial extent and grid resolution (30 arc-seconds) as the climatic variables, was rescaled to range from 1 to 20 following Elith et al. (2010) and assigned higher probabilities to background points drawn from areas with comparable sampling intensity to the \u003cem\u003eH. metabus\u003c/em\u003e records. This approach effectively reduces the introduction of artifacts resulting from uneven survey efforts. The same approach was applied using Magnoliopsida occurrences (n\u0026thinsp;=\u0026thinsp;75,262; Fig. S4b) to correct sampling bias in the host plants models. GBIF extractions (GBIF.org 2024) initially resulted in a total of 188,523 occurrences (moths and plants combined), which were then cleaned up by removing duplicates and excluding records with missing spatial coordinates or taxonomic identification using standard filtering procedures in R.\u003c/p\u003e\n\u003cp\u003eInitially, optimal modeling settings (feature classes and regularization multiplier) for each species were determined using the ENMeval R package (Muscarella et al. 2014), selecting the model configuration with the lowest corrected Akaike Information Criterion (AICc). To ensure methodological consistency across species, we retained the same model settings (features\u0026thinsp;=\u0026thinsp;Linear\u0026thinsp;+\u0026thinsp;Quadratic, regularization multiplier\u0026thinsp;=\u0026thinsp;1) for all final models (Table S4). This configuration showed a \u0026Delta;AICc\u0026thinsp;\u0026lt;\u0026thinsp;2 for all species except \u003cem\u003eL. racemosa\u003c/em\u003e (\u0026Delta;AICc\u0026thinsp;=\u0026thinsp;2.38), which was nevertheless considered acceptable given the high predictive performance of the model (AUC\u0026thinsp;=\u0026thinsp;0.94). These settings were subsequently used to build final species distribution models using MaxEnt software (Phillips et al. 2006) with a 10-fold cross-validation procedure. Finally, the predicted distribution of \u003cem\u003eH. metabus\u003c/em\u003e was compared to those of its potential host plants through niche overlap analysis. Using the dismo R package (Hijmans et al. 2010), we calculated Schoener\u0026rsquo;s \u003cem\u003eD\u003c/em\u003e and Hellinger-based \u003cem\u003eI\u003c/em\u003e indices to quantify spatial similarity between species distribution models (Warren et al. 2008). Both indices range from 0 (no overlap) to 1 (complete overlap), but differ in their underlying assumptions and sensitivity. Schoener\u0026rsquo;s \u003cem\u003eD\u003c/em\u003e measures absolute differences in predicted suitability across space, assuming that these values reflect relative habitat use. In contrast, the Hellinger-based \u003cem\u003eI\u003c/em\u003e index treats model outputs strictly as probability distributions, without assuming a biological meaning, and is therefore more robust to extreme values and skewed predictions. By combining both metrics, we capture complementary aspects of niche similarity and minimize potential biases linked to any single interpretation of model outputs. We further assessed niche divergence between \u003cem\u003eH. metabus\u003c/em\u003e and each host plant by performing both equivalency and similarity tests using the ENMTools R package (Warren et al. 2021). Both tests use Schoener\u0026rsquo;s \u003cem\u003eD\u003c/em\u003e and Hellinger-based \u003cem\u003eI\u003c/em\u003e indices to quantify niche overlap, but differ in their null hypotheses. The niche equivalency test evaluates whether \u003cem\u003eH. metabus\u003c/em\u003e and its host plants occupy ecologically equivalent niches, by comparing the observed overlap to that expected under random allocation of occurrences. The niche similarity test assesses whether \u003cem\u003eH. metabus\u003c/em\u003e\u0026apos; niche is better predicted by the environmental distribution of its host plants than expected by chance, thus testing whether they share similar environments beyond spatial proximity. Together, these tests provide a statistical framework to evaluate whether niche similarity reflects true ecological overlap or arises from environmental availability alone. All models used in these tests were generated with Maxent (feature class\u0026thinsp;=\u0026thinsp;LQ, regularization multiplier\u0026thinsp;=\u0026thinsp;1). Due to package limitations, cross-validation was not applied; instead, each test was based on 99 permutations per species.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eHabitat composition\u003c/p\u003e\u003cp\u003eThe 13 sites sampled differed in forest structure and composition (Table S4), but could be separated into four different categories (see Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, Table S5) based on the forest habitat cartography of the French National Forest Office (ONF; Guitet et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). The first category is coastal forests, which include five of our sites, spanning all three risk levels of outbreak. Although consisting of many different forest habitats, all are strips of lowland forest (\u0026gt;\u0026thinsp;20 m in altitude) that extend inland\u0026thinsp;\u0026gt;\u0026thinsp;40 km along the coastline. This habitat category is characterized by low species diversity, high density of small stems, and low forest cover (Guitet et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Notably, amongst the tree species encountered, about half (49%) were potential host plants to \u003cem\u003eH. metabus\u003c/em\u003e. Within the coastal forest category, the Mana and Larivot bridge sites were especially distinct. Mana is characterized by a white sand forest, with many endemic and rare species, generally uncommon elsewhere (Guitet et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). The Larivot bridge site consists of mangroves, which are periodically submerged by salty or brackish water, and have especially low diversity (it was the site with the lowest species richness) and tree species adapted to this unusual environment. Only two plant species were found in our Larivot transects, \u003cem\u003eA. germinans\u003c/em\u003e and \u003cem\u003eL. racemosa\u003c/em\u003e, and both are host plants for \u003cem\u003eH. metabus\u003c/em\u003e. The second category is a mix of savanna and coastal forest. This included four of our sites and they had a moderate to severe risk for \u003cem\u003eH. metabus\u003c/em\u003e outbreaks. These were a mix of stands of coastal forests and grassy woodland characterized by trees sufficiently widely spaced so that the canopy did not close. Canopy cover of our transects were in fact the lowest for this habitat (76% cover vs 87% on average for all sites) and the transects had the least number of trees counted (5.83 per transect vs the 7.92 overall mean), a large percentage (37%) of which were potential host plants.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eMean values (for the three transects) per site for selected ecological variables and diversity indices measured at 13 sites in French Guiana: values in bold indicate values greater than or equal to the median. The category outbreak propensity for each site is indicated in parentheses (low (1), medium (2) and high (3) outbreak propensity).\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"8\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eID\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSite\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNumber of trees\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNumber of host plants\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eProportion of host plants\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eCanopy cover (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eTree species richness (S)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e\u003cp\u003eForest habitat as per Guitet et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2015\u003c/span\u003e\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCacao (1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cb\u003e11.33\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.16\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e95\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e9.33\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eMid-altitude mountain forest\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMountain des singes (1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cb\u003e12\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.33\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.05\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e89\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e8.33\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eHill and valley forest\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMana (1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003e2.67\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e0.48\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e94\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e4.67\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eCoastal forest (white sand forest)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSaint-Laurent-du-Maroni (1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cb\u003e8.67\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003e3\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e0.35\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e96\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e5.33\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eCoastal forest\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eLake Petit-Saut (2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cb\u003e13\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.33\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.04\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e95\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e5.33\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eHill and valley forest\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eKaw mountain (2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cb\u003e8.67\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003e3\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e0.33\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e94\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e6\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eMid-altitude mountain forest\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eRoura (2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cb\u003e7\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003e4.67\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e0.61\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e79\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e3.67\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eCoastal forest\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eBridge Larivot, Macouria (2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003e5\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e88\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eCoastal forest (Mangroves)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eBotanical garden, Macouria (2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e5.33\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003e3.33\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e0.56\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e68\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e2.67\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eSavanna \u0026amp; coastal forest\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSavanna of Matiti (2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003e2.33\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e0.39\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e61\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e4.67\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eSavanna \u0026amp; coastal forest\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eD\u0026eacute;grad des Cannes (boat harbour, R\u0026eacute;mire-Montjoly) (3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cb\u003e8\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003e2.33\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.31\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e97\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e4.33\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eCoastal forest\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e12\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eIracoubo (3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e5.67\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003e2.33\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e0.38\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e80\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eSavanna \u0026amp; coastal forest\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSinnamary (3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e6.33\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.15\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e95\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e5.33\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eSavanna \u0026amp; coastal forest\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMedian\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2.33\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.35\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e94\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e4.67\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eThe third habitat category was the hill and valley forest, and it included two of our sites. These sites had fairly high tree density, and relatively abundant small stems. These forests are described as having dense undergrowth and mid to high canopy (30\u0026ndash;35 m; Guitet et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). We found these sites to have a large number of trees (12.50 per transect vs the 7.92 overall mean) and a high species richness, but a very low percentage of these were potential hosts (4%). The fourth habitat category was the mid-altitudinal mountain forests and included two of our sites. These forests are described as having a high canopy (37 m) of irregular appearance, with high biomass, and where large trees are common and the undergrowth very diverse (Guitet et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Similar to forests of hill and valleys, the mid-altitudinal forest had a great stem density from different species, but a large percentage of them (23%) were found to be potential host plants, although they were classified as of similarly low to moderate risk for outbreak.\u003c/p\u003e\u003cp\u003eWhen comparing the different classes for outbreak propensity (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e), the sites most likely to suffer from frequent outbreaks (group 3, n\u0026thinsp;=\u0026thinsp;3) were characterized by low tree count, whereas groups least likely to have outbreaks (group 1, n\u0026thinsp;=\u0026thinsp;4) were characterized by high tree count and high diversity. Surprisingly, the highest proportion of host plants (43%) was found at sites of moderate risk (group 2, n\u0026thinsp;=\u0026thinsp;6), although this difference appears to be solely due to the mangrove site where only two species, both potential hosts, were found.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eMean values for some of the ecological variables and diversity indices measured at 13 sites in French Guiana per category of outbreak propensity (low (1), medium (2) and high (3) outbreak propensity).\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"8\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eOutbreak groups\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eNumber of sites\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNumber of trees\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNumber of host plants\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eProportion of host plants\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eCanopy cover (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eTree species richness (S)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e\u003cp\u003eProportion of savannas \u0026amp; coastal forests\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e9.50\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.26\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e93.44\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e6.92\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0.5\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e7.50\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3.11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.49\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e80.92\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e4.06\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0.67\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e6.67\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1.89\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.28\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e90.76\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e4.22\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003ePredation pressure\u003c/p\u003e\u003cp\u003eOf the 1,284 artificial caterpillars retrieved, 291 exhibited signs of predation, of which only 85 could be attributed to avian attacks, resulting in an overall average avian predation rate of 0.07. Due to the generally low predation rates, differences between sites were moderate, ranging from 0 to 0.17 (Table S5). Moreover, the Kruskal-Wallis test did not detect a significant overall effect of habitat type on predation rates (Statistic\u0026thinsp;=\u0026thinsp;6.25, p\u0026thinsp;=\u0026thinsp;0.1). However, post-hoc Dunn\u0026rsquo;s test comparisons revealed significant differences between mid-altitude mountain forests and both coastal forests and savannas (statistic = -2.37, p\u0026thinsp;=\u0026thinsp;0.02, statistic = -2.04, p\u0026thinsp;=\u0026thinsp;0.04, respectively; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e), with predation rates being highest in mid-altitude forests (mean\u0026thinsp;=\u0026thinsp;0.11). In contrast, predation rates were lowest in savannas (mean\u0026thinsp;=\u0026thinsp;0.06) and coastal forests (mean\u0026thinsp;=\u0026thinsp;0.04), although two sampled coastal forest sites exhibited marked variation (0.12 in white sand forests vs. 0 in mangroves). Predation rates in hill and valley forests were intermediate between those observed in coastal (savanna and coastal forest) and mid-altitude mountain forest, with an average predation rate of 0.09.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eHabitat comparison\u003c/p\u003e\u003cp\u003eSpatial generalized additive model (sGAM) was fitted to account for the ordinal nature of the site categories for outbreak propensity, while controlling for spatial autocorrelation. The analysis revealed that only the predation rate, the canopy cover (%) and the number of trees were significantly associated with the potential for outbreak. Outbreak propensity increased with a higher predation rate (estimate\u0026thinsp;=\u0026thinsp;177\u0026thinsp;\u0026plusmn;\u0026thinsp;52, z\u0026thinsp;=\u0026thinsp;3.38, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01), whereas it decreased with greater canopy cover (estimate = -0.30\u0026thinsp;\u003cb\u003e\u0026plusmn;\u003c/b\u003e\u0026thinsp;0.14, z = -2.17, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.03) and with the number of trees (estimate = -3.94\u0026thinsp;\u003cb\u003e\u0026plusmn;\u003c/b\u003e\u0026thinsp;1.35, z = -2.91, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01). The number of host plants did not have a significant effect (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.63). The spatial smoothing term was highly significant (\u003cem\u003eedf\u003c/em\u003e\u0026thinsp;=\u0026thinsp;5.77, \u003cem\u003eχ\u0026sup2;\u003c/em\u003e = 331.4, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), suggesting a strong spatial structure in outbreak patterns. The final model explained 99.4% of the deviance. As such, sites associated with frequent outbreaks are characterized by habitats with low tree density (i.e. both reduced tree count and reduced canopy cover) and high predation rate (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, see supplementary material for details).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eSpecies distribution model\u003c/p\u003e\u003cp\u003eAll species distribution models (SDMs) for \u003cem\u003eH. metabus\u003c/em\u003e and its host plants show reasonable or high performance with AUC values greater than 0.7 (\u003cem\u003eT. guianensis\u003c/em\u003e) or 0.9 (\u003cem\u003eH. metabus\u003c/em\u003e, \u003cem\u003eA. germinans\u003c/em\u003e, \u003cem\u003eL. racemosa\u003c/em\u003e) as shown in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e (Swets \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e1988\u003c/span\u003e; Peterson et al. \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). The models showed a higher probability for the presence of the moth \u003cem\u003eH. metabus\u003c/em\u003e along the coastline (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). Although \u003cem\u003eH. metabus\u003c/em\u003e has been found to occur inland, these habitats appear less suitable for the species based on both collecting data and the distribution model.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe areas with the highest probability of occurrence are situated along the coast between Iracoubo and a little past Cayenne and R\u0026eacute;mire-Montjoly. These areas were also classified as being at moderate to high risk of outbreaks based on our classification. Although some occurrence is predicted on the eastern coast, around Saint-Laurent-du-Maroni and Mana, we have found this region to actually be at low risk. The climatic variable that contributed the most to the predicted distribution of \u003cem\u003eH. metabus\u003c/em\u003e was the mean diurnal temperature range (PI\u0026thinsp;=\u0026thinsp;88.9). The host plants \u003cem\u003eA. germinans\u003c/em\u003e and \u003cem\u003eL. racemosa\u003c/em\u003e were also predicted to occur along the coast (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb,c), consistent with their ecological restriction to mangrove habitats. Their distributions were also primarily influenced by the mean diurnal temperature range (PI\u0026thinsp;=\u0026thinsp;79.3 and 69.8, respectively), with additional contributions from the annual mean temperature for \u003cem\u003eA. germinans\u003c/em\u003e (PI\u0026thinsp;=\u0026thinsp;11.7) and the precipitation of the warmest quarter for \u003cem\u003eL. racemosa\u003c/em\u003e (PI\u0026thinsp;=\u0026thinsp;26). The host plant \u003cem\u003eT. guianensis\u003c/em\u003e had a broader predicted range (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ed), with its distribution primarily associated with the mean diurnal temperature range (PI\u0026thinsp;=\u0026thinsp;64.8) and precipitation of the warmest quarter (PI\u0026thinsp;=\u0026thinsp;31.3).\u003c/p\u003e\u003cp\u003eThe highest similarity for spatial overlap was found with \u003cem\u003eL. racemosa\u003c/em\u003e (\u003cem\u003eD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.84; \u003cem\u003eI\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.98), followed by \u003cem\u003eT. guianensis\u003c/em\u003e (\u003cem\u003eD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.60; \u003cem\u003eI\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.89) and \u003cem\u003eA. germinans\u003c/em\u003e (\u003cem\u003eD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.61; \u003cem\u003eI\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.87). These values indicate moderate to high spatial overlap between the moth\u0026rsquo;s predicted distribution and those of its potential host plants, especially \u003cem\u003eL. racemosa\u003c/em\u003e, which shows near-complete similarity based on the \u003cem\u003eI\u003c/em\u003e index (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eResults of SDMs with permutation importance (PI) of each variable, for the moth Hylesia metabus and three of its common host plants; Avicennia germinans, Laguncularia racemosa, Tapirira guianensis.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u003cem\u003eH. metabus\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eA. germinans\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cem\u003eL. racemosa\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cem\u003eT. guianensis\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAUC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.92\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.97\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.93\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.76\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBio1 - Annual Mean Temperature\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e11.7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBio2 - Mean Diurnal Range\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e88.9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e79.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e69.8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e64.8\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBio4 - Temperature Seasonality\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e1.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.3\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBio18 - Precipitation of Warmest Quarter\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e4.9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e26\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e31.3\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBio19 - Precipitation of Coldest Quarter\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e2.7\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSpatial overlap with the \u003cem\u003eH. metabus\u003c/em\u003e model\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHellinger's I\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.87\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.98\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.89\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSchoener\u0026rsquo;s D\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.61\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.84\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.60\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eHowever, niche equivalency and niche similarity tests (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e) revealed important distinctions. The niche equivalency test did not detect significant differences between the niches of \u003cem\u003eH. metabus\u003c/em\u003e and \u003cem\u003eL. racemosa\u003c/em\u003e (p-values\u0026thinsp;\u0026gt;\u0026thinsp;0.05), suggesting that their ecological niches may be equivalent. In contrast, the niches of \u003cem\u003eH. metabus\u003c/em\u003e and \u003cem\u003eA. germinans\u003c/em\u003e, as well as \u003cem\u003eT. guianensis\u003c/em\u003e, were significantly different (p\u0026thinsp;=\u0026thinsp;0.01 for both D and I), indicating that their niches are not statistically equivalent. Niche similarity tests indicated significant niche similarity between \u003cem\u003eH. metabus\u003c/em\u003e and all three host plants (p\u0026thinsp;=\u0026thinsp;0.01 for D and I in each case), suggesting that despite non-equivalent niches, the moth consistently occupies environments similar to those of its hosts.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eHabitat heterogeneity can affect where insect pest outbreaks occur (Turner \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e1989\u003c/span\u003e; Liebhold et al. \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2004\u003c/span\u003e; Klemola et al. \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2006\u003c/span\u003e), and understanding how these ecological factors shape these occurrences is crucial for both predicting outbreak prone sites, lessening their impacts, and guiding future monitoring efforts. Here we attempted to determine the biotic and abiotic factors that may explain the spatial differences in the non-cyclical outbreaks of \u003cem\u003eH. metabus\u003c/em\u003e in French Guiana, a moth responsible for Lepidopterism.\u003c/p\u003e\u003cp\u003eSites categorized as experiencing more frequent or severe outbreaks were found to be characterized by lower tree densities (i.e. reduced number of trees and overall canopy cover) and high avian predation rate. These sites were more likely to occur in habitats consisting of coastal forests (including, but not limited to, mangroves) and savannas, which were generally correlated with low diversity and low overall tree count, of which a fairly high proportion were host plants. These habitats were also more likely to have pioneer plant species, many of which are potential hosts (\u003cem\u003ee.g.\u003c/em\u003e, \u003cem\u003eCecropia obtusa\u003c/em\u003e, \u003cem\u003eT. guianensis\u003c/em\u003e), although host plant density was not correlated with outbreak propensity, suggesting that this may not be a limiting factor. This is also consistent with other studies that have found that polyphagous insects like \u003cem\u003eH. metabus\u003c/em\u003e prefer and/or do better in fragmented habitats of low spatial heterogeneity and complexity (Tscharntke et al. \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e2002\u003c/span\u003e; Benedick et al. \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2006\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eAlthough surprising that outbreak-prone sites were correlated with high avian predation rate, it is important to note that overall predation rates were low across sites, and may have correlated more with habitat types than outbreak risk; for example, outbreak-prone sites of coastal forests had relatively high predation rates, but others like the mangroves and savannas did not. It is also worth noting that larvae of \u003cem\u003eH. metabus\u003c/em\u003e have urticating spines, which may protect them from most avian predators. As such, avian predation rates may instead reflect other characteristics such as habitat fragmentation and edge effects that could be correlated with outbreak propensity (Kareiva \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e1987\u003c/span\u003e; Baggio et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Bat\u0026aacute;ry et al. \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). For instance, in the forest tent caterpillar \u003cem\u003eMalacosoma disstria\u003c/em\u003e, the sensitivity of outbreak duration to changes in forest structure has been hypothesized to reflect reduced effectiveness of natural enemies such as parasitoids and pathogens in fragmented forests (Roland \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e1993\u003c/span\u003e; Roland and Kauppp \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e1995\u003c/span\u003e; Roland and Taylor \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e1997\u003c/span\u003e). Although the parasitoids and diseases of \u003cem\u003eH. metabus\u003c/em\u003e caterpillars are mostly unknown, future investigations should compare mortality rate and causes between coastal and inland habitats, as these may also explain differences in their population dynamics (Stamp and Bowers \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e1988\u003c/span\u003e; Zhou \u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e2009\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eAs for climatic conditions, species distribution modeling (SDM) identified mean diurnal temperature range and precipitation during the warmest quarter as key predictors of \u003cem\u003eH. metabus\u003c/em\u003e occurrence. In particular, the coast of French Guiana, characterized by more of a tropical monsoon climate with a distinct dry season and minimal diurnal temperature fluctuations, was predicted to be the most favorable. In contrast, the inland forests, where outbreaks have rarely been reported, are characterized by a tropical rainforest climate, with greater climatic stability and less pronounced seasonality (Beck et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). This pattern likely reflects the role of climatic seasonality in \u003cem\u003eH. metabus\u003c/em\u003e phenology, as shown in both temperate and tropical systems (Wolda \u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e1988\u003c/span\u003e; Neuvonen et al. \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e1999\u003c/span\u003e; Nelson et al. \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Ward et al. \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; B\u0026uuml;ntgen et al. \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Insects such as \u003cem\u003eH. metabus\u003c/em\u003e may exploit predictable seasonal cues, such as rainfall onset, to synchronize development and reproduction. Although the three host plants used for the distribution models are known to be common and frequently exploited by \u003cem\u003eH. metabus\u003c/em\u003e (Jourdain et al. \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), only the spatial distribution of \u003cem\u003eL. racemosa\u003c/em\u003e closely matched that of the moth. Nevertheless, our results revealed significant ecological similarities between the niches occupied by \u003cem\u003eH. metabus\u003c/em\u003e and its three host plants, suggesting that the moth exploits a broader ecological niche than would be inferred solely from spatial overlap. This finding indicates that \u003cem\u003eH. metabus\u003c/em\u003e likely occupies habitats ecologically analogous to multiple host species, even when their geographic co-occurrence is limited, suggesting ecological flexibility in the species allowing it to efficiently use resources distributed heterogeneously in space.\u003c/p\u003e\u003cp\u003eFor specialist insects (Opedal et al. \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Wilson et al. \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), host plant availability can be an important driver of herbivore dynamics. However, for the polyphagous \u003cem\u003eH. metabus\u003c/em\u003e, overall abundance of host plants was not found to be correlated with outbreak-prone sites. It is possible that the moth prefers habitats where a combination of host plants is present, or even sites with preferred or more suitable host plants. Nevertheless, \u003cem\u003eH. metabus\u003c/em\u003e appears to consistently occupy environments that are similar to some of its host plants, implying a shared preference for certain climatic conditions. While the availability and distribution of host plants likely play a key role in shaping the moth\u0026rsquo;s range, climatic variables appear to be even more influential in determining its overall distribution and outbreak propensity. As global warming and increased land-use change continue to modify habitats, the potential future expansion or contraction of \u003cem\u003eH. metabus\u0026rsquo;\u003c/em\u003e range warrants further investigation (Lehmann et al. \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eIn conclusion, we show that sites with reported \u003cem\u003eH. metabus\u003c/em\u003e outbreaks are mostly associated with habitats marked by overall low tree densities, minimal diurnal temperature fluctuations and distinct seasonality (\u003cem\u003ei.e.\u003c/em\u003e, marked dry and wet season). These characteristics are more likely to occur on the coast, and contrast markedly with the seasonally stable (\u003cem\u003ei.e.\u003c/em\u003e, with minimal differences between seasons) and diverse inland rainforest ecosystems, where outbreaks have rarely been reported. Avian predation, although correlated with outbreak occurrences, may instead reflect broader landscape structure such as fragmentation and edge effects. Future studies should also investigate the potential effect of other natural enemies, such as parasitoids and disease, which may be more negatively affected by habitat types and fragmentation, and which is more common along the coast, as well as host plant preferences and suitability. Targeted monitoring of both \u003cem\u003eH. metabus\u003c/em\u003e populations and their natural enemies, combined with landscape management strategies that preserve forest complexity and reduce fragmentation, may help mitigate potential future risks. Long-term studies integrating ecological, climatic, and epidemiological data may be necessary to fully understand and anticipate the drivers of pest outbreak emergence in tropical systems.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eConflict of Interest\u003c/h2\u003e\n\u003cp\u003eThe authors have no relevant financial or non-financial interests to disclose.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical Approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis article does not contain any studies with human participants or animals performed by the authors.\u003c/p\u003e\n\u003ch2\u003eFunding\u003c/h2\u003e\n\u003cp\u003eThis study was funded by a grant from the “mission pour les initiatives transverses et interdisciplinaires” (MITI) from the CNRS to MM \u0026amp; MA. This work also benefited from an “Investissement d’avenir” grant managed by the Center for the study of biodiversity in Amazonia (CEBA) (ANR-10-LABX-25-01) to MM.\u003c/p\u003e\n\u003ch2\u003eAcknowledgments\u003c/h2\u003e\n\u003cp\u003eWe thank the \u003cem\u003eMission pour les Initiatives Transverses et Interdisciplinaires\u003c/em\u003e (MITI) of the CNRS for supporting this study through a grant awarded to \u003cstrong\u003eMM\u003c/strong\u003e and \u003cstrong\u003eMA\u003c/strong\u003e. We also acknowledge the support of the \u003cem\u003eInvestissement d’Avenir\u003c/em\u003e grant managed by the Center for the Study of Biodiversity in Amazonia (CEBA, ANR-10-LABX-25-01) awarded to \u003cstrong\u003eMM\u003c/strong\u003e. We thank Sandra Ianez, Rémi Mauxion, Pierre Lacoste, Léo-Paul Charlet and Guillaume Correa Pimpao for their precious help in the field. We would also like to thank Marina Ciminera for providing us with the \u003cem\u003eHylesia metabus\u003c/em\u003e occurrence data she collected during her PhD and Frédéric Bénéluz for all the information he provided.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eApparicio Philippe, Jérémy Gelb, Jean Dubé et Joan Carles Martori (2025). \u003cem\u003eMéthodes de régression spatiale : un grand bol d’R\u003c/em\u003e. Université Laval et Université de Sherbrooke. fabriqueREL. Licence CC BY-SA.\u003c/li\u003e\n \u003cli\u003eBaggio JA, Salau K, Janssen MA, Schoon ML, Bodin Ö (2010) Landscape connectivity and predator–prey population dynamics. Landsc Ecol 26(1):33–45. https://doi.org/10.1007/s10980-010-9493-y\u003c/li\u003e\n \u003cli\u003eBarber RA, Ball SG, Morris RKA, Gilbert F (2022) Target-group backgrounds prove effective at correcting sampling bias in Maxent models. Divers Distrib 28(1):128–141. https://doi.org/10.1111/ddi.13442\u003c/li\u003e\n \u003cli\u003eBarbosa P, Letourneau DK, Agrawal AA (2012) Insect Outbreaks Revisited. John Wiley \u0026amp; Sons\u003c/li\u003e\n \u003cli\u003eBartoń K (2024) MuMIn: Multi-Model Inference\u003c/li\u003e\n \u003cli\u003eBatáry P, Fronczek S, Normann C, Scherber C, Tscharntke T (2014) How do edge effect and tree species diversity change bird diversity and avian nest survival in Germany’s largest deciduous forest? For. Ecol. Manag. 319:44–50\u003c/li\u003e\n \u003cli\u003eBeck HE, McVicar TR, Vergopolan N, Berg A, Lutsko NJ, Dufour A, Zeng Z, Jiang X, van Dijk AIJM, Miralles DG (2023) High-resolution (1 km) Köppen-Geiger maps for 1901–2099 based on constrained CMIP6 projections. Sci Data 10(1):724. https://doi.org/10.1038/s41597-023-02549-6\u003c/li\u003e\n \u003cli\u003eBenedick S, Hill J, Mustaffa N, Chey V, Maryati M, Searle J, Schilthuizen M, Hamer K (2006) Impacts of rain forest fragmentation on butterflies in northern Borneo: species richness, turnover and the value of small fragments. J. Appl. Ecol. 43:967–977\u003c/li\u003e\n \u003cli\u003eBerryman AA (1987) Chapter 1 - The Theory and Classification of Outbreaks. In: Barbosa P, Schultz JC (eds) Insect Outbreaks. Academic Press, San Diego, pp 3–30\u003c/li\u003e\n \u003cli\u003eBradshaw CJA, Leroy B, Bellard C, Roiz D, Albert C, Fournier A, Barbet-Massin M, Salles J-M, Simard F, Courchamp F (2016) Massive yet grossly underestimated global costs of invasive insects. Nat Commun 7(1):12986. https://doi.org/10.1038/ncomms12986\u003c/li\u003e\n \u003cli\u003eBüntgen U, Liebhold A, Nievergelt D, Wermelinger B, Roques A, Reinig F, Krusic PJ, Piermattei A, Egli S, Cherubini P, Esper J (2020) Return of the moth: rethinking the effect of climate on insect outbreaks. Oecologia 192(2):543–552. https://doi.org/10.1007/s00442-019-04585-9\u003c/li\u003e\n \u003cli\u003eBurnham KP, Anderson DR (eds) (2004) Model Selection and Multimodel Inference. Springer New York, New York, NY\u003c/li\u003e\n \u003cli\u003eCaratti JF (2006) Line Intercept (LI). Lutes Duncan C Keane Robert E Caratti John F Key Carl H Benson Nathan C Sutherl. Steve Gangi Larry J 2006 FIREMON Fire Eff. Monit. Inventory Syst. Gen Tech Rep RMRS-GTR-164-CD Fort Collins CO US Dep. Agric. For. Serv. Rocky Mt. Res. Stn. P LI-1-13 164\u003c/li\u003e\n \u003cli\u003eCheca M, Rodriguez J, Willmott K, Liger B (2014) Microclimate Variability Significantly Affects the Composition, Abundance and Phenology of Butterfly Communities in a Highly Threatened Neotropical Dry Forest. Fla Entomol 97:1–13. https://doi.org/10.1653/024.097.0101\u003c/li\u003e\n \u003cli\u003eCiminera M (2017) Identification spécifique et structure génétique des populations du papillon-cendre responsable des épisodes de papillonite en Guyane et au Vénézuela. Thesis\u003c/li\u003e\n \u003cli\u003eCiminera M, Auger-Rozenberg M-A, Caron H, Herrera M, Scotti-Saintagne C, Scotti I, Tysklind N, Roques A (2019) Genetic variation and differentiation of Hylesia metabus (Lepidoptera: Saturniidae): moths of public health importance in French Guiana and in Venezuela. J. Med. Entomol. 56:137–148\u003c/li\u003e\n \u003cli\u003eCNEV (2011) Réponse à la saisine “Stratégies et méthodes de lutte optimales contre Hylesia metabus”, agent de la papillonite en Guyane française. Avis à l’attention de la Direction Générale de la Santé. Centre National D’Expertise sur les Vecteurs, Montpellier\u003c/li\u003e\n \u003cli\u003eCROPP (2015) La papillonite - Bulletin d’information n°10 Août 2015, Cellule régionale permanente d’observation et de prévention de la papillonite en Guyane.\u003c/li\u003e\n \u003cli\u003eDavidson J (1944) On the Relationship between Temperature and Rate of Development of Insects at Constant Temperatures. J Anim Ecol 13(1):26–38. https://doi.org/10.2307/1326\u003c/li\u003e\n \u003cli\u003eDent D, Binks RH (2020) Insect Pest Management, 3rd Edition. CABI\u003c/li\u003e\n \u003cli\u003eDidham RK, Lawton JH (1999) Edge Structure Determines the Magnitude of Changes in Microclimate and Vegetation Structure in Tropical Forest Fragments. Biotropica 31(1):17–30. https://doi.org/10.1111/j.1744-7429.1999.tb00113.x\u003c/li\u003e\n \u003cli\u003eElith J, Kearney M, Phillips S (2010) The art of modelling range-shifting species. Methods Ecol Evol 1(4):330–342. https://doi.org/10.1111/j.2041-210X.2010.00036.x\u003c/li\u003e\n \u003cli\u003eEötvös C, Lövei GL (2013) Documenting predator marks on dummy caterpillars. In: Saska P, Knapp M, Honek A, Martinkova Z (eds) XVIth European Carabidologists Meeting - Book of Abstracts with Conference Programme: Carabids and man - can we live with(out) each other? European Carabidologists Meeting, p 31\u003c/li\u003e\n \u003cli\u003eFick SE, Hijmans RJ (2017) WorldClim 2: new 1-km spatial resolution climate surfaces for global land areas. Int J Climatol 37(12):4302–4315. https://doi.org/10.1002/joc.5086\u003c/li\u003e\n \u003cli\u003eFraver S, Ducey MJ, Woodall CW, D’Amato AW, Milo AM, Palik BJ (2018) Influence of transect length and downed woody debris abundance on precision of the line-intersect sampling method. For. Ecosyst. 5:1–10\u003c/li\u003e\n \u003cli\u003eGBIF.org (2024) GBIF Occurrence Download https://doi.org/10.15468/DL.28K84S\u003c/li\u003e\n \u003cli\u003eGuisan A, Thuiller W, Zimmermann NE (2017) Habitat Suitability and Distribution Models: with Applications in R. Cambridge University Press\u003c/li\u003e\n \u003cli\u003eGuitet S, Euriot S, Brunaux O, Baraloto C, Denis T, Dewynter M, Freycon V, Gonzales S, Jaouen G, Hansen CR (2015) Catalogue des habitats forestiers de Guyane\u003c/li\u003e\n \u003cli\u003eHijmans RJ (2023) raster: Geographic Data Analysis and Modeling\u003c/li\u003e\n \u003cli\u003eHijmans RJ, Phillips S, Leathwick J, Elith J (2010) dismo: Species Distribution Modeling. 1.3-16\u003c/li\u003e\n \u003cli\u003eIanez S, Palisse M, Clerc-Renaud A (2021) Le quotidien avec Hylesia metabus (Cramer, 1789). Cohabiter avec le papillon cendre dans les communes du littoral guyanais. [Unpublished manuscript].\u003c/li\u003e\n \u003cli\u003eJames G, Witten D, Hastie T, Tibshirani R (2013) An introduction to statistical learning. Springer\u003c/li\u003e\n \u003cli\u003eJourdain F, Girod R, Vassal JM, Chandre F, Lagneau C, Fouque F, Guiral D, Raude J, Robert V (2012) The moth Hylesia metabus and French Guiana lepidopterism: centenary of a public health concern. Parasite J Société Fr Parasitol 19(2):117–128. https://doi.org/10.1051/parasite/2012192117\u003c/li\u003e\n \u003cli\u003eKareiva P (1987) Habitat fragmentation and the stability of predator–prey interactions. Nature 326(6111):388–390. https://doi.org/10.1038/326388a0\u003c/li\u003e\n \u003cli\u003eKassambara A (2020) Pipe-Friendly Framework for Basic Statistical Tests [R package rstatix version 0.7.2]\u003c/li\u003e\n \u003cli\u003eKlemola T, Huitu O, Ruohomäki K (2006) Geographically partitioned spatial synchrony among cyclic moth populations. Oikos 114(2):349–359. https://doi.org/10.1111/j.2006.0030-1299.14850.x\u003c/li\u003e\n \u003cli\u003eKocsis M, Hufnagel L (2011) Impacts of climate change on Lepidoptera species and communities. Appl Ecol Environ Res 9:43–72. https://doi.org/10.15666/aeer/0901_043072\u003c/li\u003e\n \u003cli\u003eLamy M, Lemaire C (1983) Contribution à la systématique des Hylesia : étude au microscope électronique à balayage des «fléchettes » urticantes [Lep. Saturniidae]. Bull Société Entomol Fr :176–192\u003c/li\u003e\n \u003cli\u003eLê S, Josse J, Husson F (2008) FactoMineR: A Package for Multivariate Analysis. J Stat Softw 25(1):1–18. https://doi.org/10.18637/jss.v025.i01\u003c/li\u003e\n \u003cli\u003eLeger M, Mouzels P (1918) Dermatose prurigineuse déterminée par des papillons Saturnides du genre Hylesia. Bull Soc Path Exot 11:104–107\u003c/li\u003e\n \u003cli\u003eLehmann P, Ammunét T, Barton M, Battisti A, Eigenbrode SD, Jepsen JU, Kalinkat G, Neuvonen S, Niemelä P, Terblanche JS, Økland B, Björkman C (2020) Complex responses of global insect pests to climate warming. Front Ecol Environ 18(3):141–150. https://doi.org/10.1002/fee.2160\u003c/li\u003e\n \u003cli\u003eLemaire C (2002) Saturniidae of America: Hemileucinae. Geocke \u0026amp; Evers\u003c/li\u003e\n \u003cli\u003eLiebhold A, Koenig WD, Bjørnstad ON (2004) Spatial Synchrony in Population Dynamics. Annu Rev Ecol Evol Syst 35:467–490\u003c/li\u003e\n \u003cli\u003eLövei GL, Ferrante M (2017) A review of the sentinel prey method as a way of quantifying invertebrate predation under field conditions. Insect Sci. 24:528–542\u003c/li\u003e\n \u003cli\u003eLow PA, Sam K, McArthur C, Posa MRC, Hochuli DF (2014) Determining predator identity from attack marks left in model caterpillars: guidelines for best practice. Entomol. Exp. Appl. 152:120–126\u003c/li\u003e\n \u003cli\u003eLuce AE, Couppie P, Michaud C, Clauteaux P, Blaizot R (2023) Épidémiologie, clinique et thérapeutique de la papillonite en Guyane, 2017–2023. Ann Dermatol Vénéréologie - FMC 3(8):A257. https://doi.org/10.1016/j.fander.2023.09.426\u003c/li\u003e\n \u003cli\u003eMcElhinny C, Gibbons P, Brack C, Bauhus J (2005) Forest and woodland stand structural complexity: Its definition and measurement. For Ecol Manag 218(1):1–24. https://doi.org/10.1016/j.foreco.2005.08.034\u003c/li\u003e\n \u003cli\u003eMuscarella R, Galante PJ, Soley-Guardia M, Boria RA, Kass JM, Uriarte M, Anderson RP (2014) ENMeval: An R package for conducting spatially independent evaluations and estimating optimal model complexity for Maxent ecological niche models. Methods Ecol Evol 5(11):1198–1205. https://doi.org/10.1111/2041-210X.12261\u003c/li\u003e\n \u003cli\u003eNelson WA, Bjørnstad ON, Yamanaka T (2013) Recurrent Insect Outbreaks Caused by Temperature-Driven Changes in System Stability. Science 341(6147):796–799. https://doi.org/10.1126/science.1238477\u003c/li\u003e\n \u003cli\u003eNeuvonen S, Niemelä P, Virtanen T (1999) Climatic Change and Insect Outbreaks in Boreal Forests: The Role of Winter Temperatures. Ecol Bull (47):63–67\u003c/li\u003e\n \u003cli\u003eOpedal ØH, Ovaskainen O, Saastamoinen M, Laine A-L, van Nouhuys S (2020) Host-plant availability drives the spatiotemporal dynamics of interacting metapopulations across a fragmented landscape. Ecology 101(12):e03186. https://doi.org/10.1002/ecy.3186\u003c/li\u003e\n \u003cli\u003ePebesma E (2018) Simple Features for R: Standardized Support for Spatial Vector Data. R J 10(1):439–446. https://doi.org/10.32614/RJ-2018-009\u003c/li\u003e\n \u003cli\u003ePeterson AT, Soberón J, Pearson RG, Anderson RP, Martínez-Meyer E, Nakamura M, Araújo MB (2011) Ecological Niches and Geographic Distributions. Princeton University Press\u003c/li\u003e\n \u003cli\u003ePhillips SJ, Anderson RP, Schapire RE (2006) Maximum entropy modeling of species geographic distributions. Ecol Model 190(3):231–259. https://doi.org/10.1016/j.ecolmodel.2005.03.026\u003c/li\u003e\n \u003cli\u003ePhillips SJ, Dudík M, Elith J, Graham CH, Lehmann A, Leathwick J, Ferrier S (2009) Sample selection bias and presence-only distribution models: implications for background and pseudo-absence data. Ecol Appl 19(1):181–197. https://doi.org/10.1890/07-2153.1\u003c/li\u003e\n \u003cli\u003eRatnasingham S, Hebert PDN (2007) bold: The Barcode of Life Data System (http://www.barcodinglife.org). Mol Ecol Notes 7(3):355–364. https://doi.org/10.1111/j.1471-8286.2007.01678.x\u003c/li\u003e\n \u003cli\u003eRatte HT (1985) Temperature and Insect Development. In: Hoffmann KH (ed) Environmental Physiology and Biochemistry of Insects. Springer Berlin Heidelberg, Berlin, Heidelberg, pp 33–66\u003c/li\u003e\n \u003cli\u003eRaza MM, Khan ,Muhammad Aslam, Arshad ,Muhammad, Sagheer ,Muhammad, Sattar ,Zeeshan, Shafi ,Jamil, Haq ,Ehtisham ul, Ali ,Asim, Aslam ,Usman, Mushtaq ,Aleena, Ishfaq ,Iqra, Sabir ,Zarnab, and Sattar A (2015) Impact of global warming on insects. Arch Phytopathol Plant Prot 48(1):84–94. https://doi.org/10.1080/03235408.2014.882132\u003c/li\u003e\n \u003cli\u003eRoels SM, Porter JL, Lindell CA (2018) Predation pressure by birds and arthropods on herbivorous insects affected by tropical forest restoration strategy. Restor. Ecol. 26:1203–1211\u003c/li\u003e\n \u003cli\u003eRoland J (1993) Large-scale forest fragmentation increases the duration of tent caterpillar outbreak. Oecologia 93(1):25–30. https://doi.org/10.1007/BF00321186\u003c/li\u003e\n \u003cli\u003eRoland J, Kauppp WJ (1995) Reduced Transmission of Forest Tent Caterpillar (Lepidoptera: Lasiocampidae) Nuclear Polyhedrosis Virus at the Forest Edge. Environ Entomol 24(5):1175–1178. https://doi.org/10.1093/ee/24.5.1175\u003c/li\u003e\n \u003cli\u003eRoland J, Taylor PD (1997) Insect parasitoid species respond to forest structure at different spatial scales. Nature 386(6626):710–713. https://doi.org/10.1038/386710a0\u003c/li\u003e\n \u003cli\u003eSchindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, Preibisch S, Rueden C, Saalfeld S, Schmid B (2012) Fiji: an open-source platform for biological-image analysis. Nat. Methods 9:676–682\u003c/li\u003e\n \u003cli\u003eStamp N, Bowers M (1988) Direct and indirect effects of predatory wasps (Polistes sp.: Vespidae) on gregarious caterpillars (Hemileuca lucina: Saturniidae). Oecologia 75:619–624\u003c/li\u003e\n \u003cli\u003eSwets JA (1988) Measuring the Accuracy of Diagnostic Systems. Science 240(4857):1285–1293. https://doi.org/10.1126/science.3287615\u003c/li\u003e\n \u003cli\u003eTscharntke T, Steffan-Dewenter I, Kruess A, Thies C (2002) Contribution of Small Habitat Fragments to Conservation of Insect Communities of Grassland–Cropland Landscapes. Ecol Appl 12(2):354–363. https://doi.org/10.1890/1051-0761(2002)012[0354:COSHFT]2.0.CO;2\u003c/li\u003e\n \u003cli\u003eTurner MG (1989) Landscape ecology: the effect of pattern on process. Annu Rev Ecol Syst 20(1):171–197\u003c/li\u003e\n \u003cli\u003eVassal J-M (1989) Biologie, écologie et pathologie d’Hylesia metabus (Cramer 1775) (Lépidoptères : Saturniidae), agent de la papillonite en Guyane Française : mise en place d’une structure de lutte intégrée. Thesis, USTL\u003c/li\u003e\n \u003cli\u003eVenables WN, Ripley BD (2002) Modern Applied Statistics with S, Fourth. Springer, New York\u003c/li\u003e\n \u003cli\u003eWard SF, Moon RD, Aukema BH (2019) Implications of seasonal and annual heat accumulation for population dynamics of an invasive defoliator. Oecologia 190(3):703–714. https://doi.org/10.1007/s00442-019-04431-y\u003c/li\u003e\n \u003cli\u003eWarren DL, Glor RE, Turelli M (2008) ENVIRONMENTAL NICHE EQUIVALENCY VERSUS CONSERVATISM: QUANTITATIVE APPROACHES TO NICHE EVOLUTION. Evolution 62(11):2868–2883. https://doi.org/10.1111/j.1558-5646.2008.00482.x\u003c/li\u003e\n \u003cli\u003eWarren DL, Matzke NJ, Cardillo M, Baumgartner JB, Beaumont LJ, Turelli M, Glor RE, Huron NA, Simões M, Iglesias TL, Piquet JC, Dinnage R (2021) ENMTools 1.0: an R package for comparative ecological biogeography. Ecography 44(4):504–511. https://doi.org/10.1111/ecog.05485\u003c/li\u003e\n \u003cli\u003eWilson JK, Casajus N, Hutchinson RA, McFarland KP, Kerr JT, Berteaux D, Larrivée M, Prudic KL (2021) Climate Change and Local Host Availability Drive the Northern Range Boundary in the Rapid Expansion of a Specialist Insect Herbivore, Papilio cresphontes. Front Ecol Evol Volume 9-2021\u003c/li\u003e\n \u003cli\u003eWolda H (1988) Insect Seasonality: Why? Annu Rev Ecol Syst 19:1–18\u003c/li\u003e\n \u003cli\u003eZhou Z (2009) A review on control of tobacco caterpillar, Spodoptera litura. Chin. Bull. Entomol. 46:354–361\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"neotropical-entomology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"nent","sideBox":"Learn more about [Neotropical Entomology](https://www.springer.com/journal/13744)","snPcode":"13744","submissionUrl":"https://www.editorialmanager.com/nent/default2.aspx","title":"Neotropical Entomology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Ashen moth, forest pest, papillon cendre, papillonite, spatial heterogeneity, species distribution model","lastPublishedDoi":"10.21203/rs.3.rs-7547284/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7547284/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eUnderstanding the environmental context of insect outbreaks is crucial, particularly for pest species with significant impact on human health. The ashen moth \u003cem\u003eHylesia metabus\u003c/em\u003e is a generalist Lepidoptera with urticating scales (\u003cem\u003esetae\u003c/em\u003e). This species causes severe dermatological reactions and its outbreaks pose serious public health challenges along the coastal regions of Venezuela and French Guiana. Despite the species' broad distribution throughout northern South America, outbreaks remain unpredictable and localized. Here, we explored factors that correlate with the recent spatial distribution of outbreaks by investigating 13 sites in French Guiana. We assessed forest structure, tree species composition, canopy cover and avian predation rates in the field. Additionally, we performed species distribution modeling to explore the effect of climate. Outbreak-prone sites were associated with overall low tree densities, high predation pressure, limited daily temperature variation, and pronounced seasonal changes between the dry and rain seasons. These conditions are more prevalent along the coast of French Guiana, contrasting sharply with the stable and diverse inland rainforest ecosystems where outbreaks are rarely reported. These findings highlight habitat features consistently associated with recent outbreak locations and provide a first step toward identifying ecological conditions that may influence outbreak propensity and can inform future monitoring strategies under changing environmental conditions.\u003c/p\u003e","manuscriptTitle":"Influence of habitat and climate on the spatial distribution of outbreaks of the Hylesia metabus moth, responsible for Lepidopterism, in coastal French Guiana","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-09 17:49:08","doi":"10.21203/rs.3.rs-7547284/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"","date":"2025-10-07T09:33:23+00:00","index":0,"fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-28T20:57:53+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"Neotropical Entomology","date":"2025-09-25T22:08:20+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-23T07:01:15+00:00","index":"","fulltext":""},{"type":"submitted","content":"Neotropical Entomology","date":"2025-09-18T11:14:21+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"neotropical-entomology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"nent","sideBox":"Learn more about [Neotropical Entomology](https://www.springer.com/journal/13744)","snPcode":"13744","submissionUrl":"https://www.editorialmanager.com/nent/default2.aspx","title":"Neotropical Entomology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"c349db92-7d4d-4329-8cb1-eee667f0cc17","owner":[],"postedDate":"October 9th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-18T19:42:32+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-09 17:49:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7547284","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7547284","identity":"rs-7547284","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}