Downloaded from https://royalsocietypublishing.org/ on 07 April 2022 royalsocietypublishing.org/journal/rsblResearch Cite this article: Lamarre GPA et al. 2022 More winners than losers over 12 years of monitoring tiger moths (Erebidae: Arctiinae) on Barro Colorado Island, Panama. Biol. Lett. 18: 20210519. https://doi.org/10.1098/rsbl.2021.0519Received: 26 October 2021 Accepted: 7 March 2022Subject Areas: ecology Keywords: climate change, functional traits, population trend, rainforest, PanamaAuthors for correspondence: Greg P. A. Lamarre e-mail: greglamarre973@gmail.com Nicholas A. Pardikes e-mail: nickpardikes@gmail.com© 2022 The Author(s) Published by the Royal Society. All rights reserved.†These authors contributed equally to the paper. A contribution to the special feature ‘Insect Decline’ organised by Martin Gossner, Florian Menzel and Nadja Simons. Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare. c.5896928.Global change biology More winners than losers over 12 years of monitoring tiger moths (Erebidae: Arctiinae) on Barro Colorado Island, Panama Greg P. A. Lamarre1,2,7,†, Nicholas A. Pardikes1,3,†, Simon Segar4, Charles N. Hackforth5, Michel Laguerre6, Benoît Vincent6, Yacksecari Lopez7, Filonila Perez7, Ricardo Bobadilla7, José Alejandro Ramírez Silva7 and Yves Basset1,2,7,8 1Department of Ecology, Institute of Entomology, Biology Centre, Czech Academy of Sciences, Ceske Budejovice 37005, Czech Republic 2Faculty of Sciences, University of South Bohemia, Ceske Budejovice, Czech Republic 3Department of Life and Earth Sciences, Perimeter College, Georgia State University, Atlanta, USA 4Agriculture and Environment Department, Harper Adams University, Newport, Shropshire TF10 8NB, UK 5Department of Geography, University College London, Gower Street, London WC1E 6BT, UK 6Muséum National d’Histoire Naturelle, Département Systématique et Évolution, Entomologie, 57 rue Cuvier, Paris, France 7ForestGEO, Smithsonian Tropical Research Institute, Apartado 0843-03092, Balboa, Ancon, Panamá City, Republic of Panamá 8Maestria de Entomologia, Universidad de Panamá, Apartado 3366, Panamá 4, Panamá GPAL, 0000-0002-7645-985X; NAP, 0000-0002-9175-4494; SS, 0000-0001-6621-9409; BV, 0000-0002-2515-2284; YB, 0000-0002-1942-5717 Understanding the causes and consequences of insect declines has become an important goal in ecology, particularly in the tropics, where most terrestrial diversity exists. Over the past 12 years, the ForestGEO Arthropod Initiative has systematically monitored multiple insect groups on Barro Colorado Island (BCI), Panama, providing baseline data for assessing long-term popu- lation trends. Here, we estimate the rates of change in abundance among 96 tiger moth species on BCI. Population trends of most species were stable (n = 20) or increasing (n = 62), with few (n = 14) declining species. Our analysis of morphological and climatic sensitivity traits associated with population trends shows that species-specific responses to climate were most strongly linked with trends. Specifically, tiger moth species that are more abundant in warmer and wetter years are more likely to show population increases. Our study contrasts with recent findings indicating insect decline in tropical and temperate regions. These results highlight the significant role of biotic responses to climate in determining long-term population trends and suggest that future climate changes are likely to impact tropical insect communities.1. Introduction Ongoing declines in insect biodiversity threaten to destabilize ecosystems world- wide [1]. Climate change and other threats affect insect population dynamics in temperate regions [2,3], but similar data are lacking in the species-rich tropics [4]. Tropical insects may be impacted by global mean temperatures and extreme climatic events. For example, many insect species shift their geographic range poleward or to higher elevations in response to increased mean temperatures [5–7]. Temperature changes may affect insect activity, development, phenology 2 Downloaded from https://royalsocietypublishing.org/ on 07 April 2022 royalsocietypublishing.org/journal/rsbl Biol. Lett. 18: 20210519and survival directly or indirectly through host phenological shifts or effects of temperature on plant chemistry [8]. Many tropical insects are extreme resource or microhabitat specialists and may be more susceptible to such changes [9,10]. Poiki- lothermic organisms cannot regulate their body temperature, and temperatures that exceed their thermal safety margin may thus result in significant fitness declines [11–14]. However, insect functional traits may be associated with potential declines in tropical communities, but the extent is unclear. Tiger moths include contrasting tribes with high morpho- logical and ecological variation [15]. They comprise generalists and specialist consumers, including the only known lineages capable of sequestrating secondary compounds from lichens, used to defend against predators and pathogens [16]. Arctiinae exhibit awide range of wing colouration, lightness and size [17]. Such high inter-species variation may lead to divergent responses to climate change and help predict insect population dynamics in the face of climate change [18–20]. Here, we exam- ine population trends among 96 tiger moth species over the past 12 years in Panama and test for their association with mor- phological and climatic sensitivity (e.g. sensitivity to mean monthly precipitation) traits and phylogenetic relatedness. Due to dispersal limitations, we predict that smaller wingspan moths may be more sensitive to climate changes [7]. However, larger species may be more prone to thermal exhaustion due to higher energy requirements [21,22]. We also predict that species with darker colouration may not favour increased solar radi- ation, particularly during the prolonged dry season [23,24]. Resource specialists such as lichen feeders are suspected to be particularly impacted by recent climate anomalies even if little data exist on lichen feeders in tropical regions. We predict that the effects of climatic sensitivity traits on temporal trends may depend on morphology.2. Material and methods (a) Study site and climate data We performed this study on Barro Colorado Island (BCI) in Panama (9.15° N, 79.85° W; approximately 140 m elevation), a tropical lowland rainforest. The island is mainly preserved and covered by lowland tropical forests with few anthropogenic disturbances. BCI receives an average of 2662 mm rainfall per year and an annual average daily maximum and minimum air temperatures of 31°C and 23.6°C, respectively (see [25]). (b) Arctiinae data and functional traits Since March 2009, the ForestGEO Arthropod Initiative has mon- itored several insect groups, including Arctiinae, using a standardized approach. The protocol consists of automatic black- light traps installed in the forest understory at 10 sites [26]. The traps operate for two non-consecutive nights at each site during four surveys in March, May, September and November (total 80 trap-nights/year). The two non-consecutive sampling nights within each month were combined for this analysis. We accumulated 12 years of continuous monitoring for a total of 47 data points for each species (due to the pandemic, we missed one sample date). We also collated morphometry, phylo- geny and functional traits for 188 Arctiinae species [27]. We argue that these functional traits are directly related to popu- lation density under an assumption of climate change. We also quantified species-specific sensitivity traits to several climatic variables, represented as beta coefficients extracted from a Baye- sian negative binomial regression model run separately for eachspecies. Description of traits and predictors are provided in elec- tronic supplementary material, S1 and S2. Each species was characterized with a unique DNA barcode, deposited in the public library BOLD [28]. (c) Data analysis We restricted the statistical analysis to common species observed during at least 6 of the 12 study years to get more robust estimates of population trends [29]. This reduced the number of species from 188 to 96. To examine rates of change through time, we modelled the sum abundance of all 96 species together andmoth species indi- vidually as a function of year. We accounted for seasonality by including month as a cofactor in all subsequent models. The extracted year coefficients served as the estimates of population trends through time. We also investigated how sensitive our results were to the period selected by removing the first and second years of sampling from our analyses. We used a Bayesian linear model and implemented continuous probabilities to provide a ‘degree of belief’ in population trends. We calculated the ‘degree of belief’ that the parameter for ‘year’ was greater than one or between specific values by counting the number of posterior draws that met our criteria and dividing that by the total number of posterior samples. We reported the number of species for whom the ‘degree of belief’ (probability) fell below 33.3% (i.e. ‘decrease’ category, twice as much confidence in a decrease than an increase), fell above 66.7% (i.e. ‘increasing’ category, twice as much confidence in an increase than a decrease) and the number of species whose degree of belief fell in between 33.3% and 66.7% per cent (i.e. ‘stable’ category, no strong evidence of increase or decrease). With quarterly abundance data for each species per year, the total abun- dance ofmoths and counts of eachmoth speciesweremodelled as a negative binomial distributionwith a logarithmic function. This dis- tribution is appropriate for overdispersed count data, which was observed in many of our species. All models met assumptions of uniformity of residuals, autocorrelation and zero-inflation using simulated residuals from the DHARMa package [30]. Year coefficients from these models are on a logarithmic scale and can be interpreted directly or as multiplicative rates of change after being exponen- tiated. We calculated the mean, 95% and 80% credible intervals from each species’ posterior probability distribution. (d) Analysis of associations between traits and trends After estimating population trends using Bayesian linear models, we examined associations between species traits (see [27]) and population trends using generalized least-squared (GLS) and phylogenetic generalized least-squared (PGLS) analyses using the nlme package in R [31]. We predicted population trends (extracted means of the posterior distributions as the estimates of change through time) as a function of a set of functional response traits, sensitivity to climate variables, or a combination of both using a GLS, with a Gaussian error distribution. All con- tinuous predictors were mean-centred to improve interpretation and model performance. We also used a PGLS model with Brow- nian motion correlation among species to account for any phylogenetic signal in population trends. We tested for phyloge- netic signal in the residuals of the GLS model using the R package picante [32]. All models included total abundance or the proportion of sample periods observed to account for differ- ences in commonness and density among species. We checked and met model assumptions (normality of residuals, heterosce- dasticity and autocorrelation) using simulated residuals from the DHARMa package [30]. We compared and evaluated GLS and PGLS model performances using AICc, root mean squared error (RMSE) and variance explained (R2). We acknowledge the switch from Bayesian to frequentists paradigms; we use trends estimated with a reasonable degree of certainty and consider PGLS the most appropriate approach here. (a) (b) 3 25 30 20 mean trend = 1.07 (d) no. species > 1.01 = 65 sp. 20 15 3 mean = 0.71 10 no. species > 66.7% = 62 spp. 4 10 2 5 Pionia affinis 3 Aclytia punctata 0 0 1 0.6 0.8 1.0 1.2 1.4 1.6 0 0.25 0.50 0.75 1.00 2 exp (estimate year) prob. slope is > 1 (c) Pionia affinis 0 1 Timalus leucomela Idalus lucaii 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 Poliopastea cyllarus year Aclytia punctata Hermierius testaceus Melese asana 0 Arctiinae sp. Balbura intervenata 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 Dysschema perplexa year Sarosa azia Talara sp. near esperanza Agyrta sp. Saurita anselma 6 Chrysostola moza Lophocampa maroniensis buchwaldi Flammeus walkeri Xanthyda saron Odozana sixola 150 Euthyone sp. Xanthyda chalcostica Hyalurga fenestra 4 Loxophlebia sp. Xanthyda chalcosticta Idalus critheis Episcepsis hypoleuca Hypocladia restricta Saurita mora Opharus gemma 100 Draudtius sp. Gaudeator paidicus 2 Loxophlebia imitata Tricypha imperialis Moeschlerius sp. Lophocampa amaxiaeformis Laguerreius pseudarchias Prepiella pexicera 50 Eucereon aeolum 0 Talara violescens Cosmosoma paAuccliyptuian cstpa. 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 Virbia mentiens year Leucopleura ciarana Correbidia elegans Melese sixola sixola Melese laodamia 0 Euthyone grisescens 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 Amaxia rocioecheverriae year Ecdemus Aobgsyclluar astpa. Agylla bioptera Talara minynthadia Lycomorphodes bicolor Lepiobneiva caecum Calonotos sp. 100 Ormetica sicilia Epeiromulona lephina Rhabdatomis cora coroides Hypocladia restricta Virbia mentiens Idalus aleteria Odozana sp. Amaxia carinosa Episcepsis venata Leucanopsis tabernilla 50 Nevrolongus thysbodes Psoloptera thoracica Agaraea sp. VirbTiaa lmarean tciaernas 75 Cosmosoma remotum Lepidoneiva teuthras OdozanNai tdoercise pstpa. Acridopsis sp. Scaptius chrysoperina 0 Scaptius chrysoperina Robinsonia sanea 50 Andrenimorpha salvini 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 Haemanota rosacea year Robinsonia flavomarginata Balbura dorsisigna Hypocharis sp. Meiese incertus Episcepsis thetis DolEiclhyessiuia 25 s sspp.. Laguerrellus banoca Episcepsis sp. Illice sp. Hyalurga urioides Ammalo helops Coccabdominis sp. 0 Idalus occidentalis Nodozana sp. 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 Callisthenia truncata year Talara sp. near caecina Correbidia costinotata Biturix sp. Pleurosoma sp. 0.5 1.0 1.5 2.0 population trend Figure 1. (a) Distribution of Arctiinae population trends over the 12 years from the negative binomial Bayesian regression. Shaded bars in the histogram represent estimates of population trends (mean of posterior distribution) that are less than 0.98, suggesting population declines. The dashed line denotes where trends are stable (e.g. exp(Neg. Binomial.Year.Coefficient) = 1). (b) The histogram displays the distribution of ‘degree of belief’ that exponentiated means of the posterior probability distribution were greater than one. This was calculated as the proportion of posterior draws for each species that were greater than one. (c) Estimates of population trends over the past 12 years among 96 Arctiinae species monitored on BCI. Each point represents the exponentiated mean of the posterior distribution from the negative binomial Bayesian regression for each species. The horizontal lines represent the 95% credible intervals, and the vertical lines display the 80% credible intervals. (d ) Examples of the temporal dynamic of six commonly collected Arctiinae on BCI using abundance-based time-series (see electronic supplementary material, figure S4). Downloaded from https://royalsocietypublishing.org/ on 07 April 2022 species no. species no. species abundance abundance abundance abundance abundance abundance royalsocietypublishing.org/journal/rsbl Biol. Lett. 18: 202105193. Results (a) Population trends in Neotropical Arctiinae Estimates of population trends in abundance over the past 12 years at BCI revealed that the entire tiger moth community had increased by 6% (95% CI: 1.01,1.11) per year (electronic supplementary material, figure S5). The probability that tiger moth abundance increased by at least 1% per year is 98%. Estimates of species-specific responses among tiger moth species revealed that most species (82 out of 96) were either stable or increased in abundance (figure 1). Sixty-two of the 96 species showed a strong degree of belief (greater than 66.7%) that their population trend was increasing (i.e. greater than one). Only 14 species showed strong evidence of declines (less than 33.3%) and the remaining 20 species did not have strong evidence of increased or decreased trends, suggesting stable dynamics. Removing 2009 or both 2009 and 2010 did not significantly alter the number of species present in each category mentioned above ofpopulation trends (electronic supplementary material, figures S6 and S7; table S1). Of the 20 species whose trends were stable, the mean probability that their trend lies within ±1% per year was 10% (electronic supplementary material, figure S8). The high degree of uncertainty in these 20 ‘stable’ species may be due to their low abundances across sample periods. They were commonly observed in our traps but were gener- ally not abundant when sampled. For 16/20 species, the average number of individuals collected in each sample period (n = 48) was less than one. This suggests that their esti- mates of population trends are uncertain, and more data may be necessary to predict their trends more accurately. (b) Association between species-specific traits and trends Models that accounted for correlations in population trends among species (PGLS) generally explained more variance but showed consistently higher AICc values. There was no Table 1. Results from the top GLS model (e.g. climate sensitivity) after 4 AICc model selection. We indicate significant associations ( p < 0.05) in italics. We modelled 93 species since three did not have genetic information and were not included in the PGLS. Standardized estimates and 95% confidence intervals are presented. CV abundance represents the coefficient of variation in abundance. predictors estimates 95% CI p-value intercept 1.06 1.04–1.08 <0.001 log(total abundance) 0.01 −0.02–0.04 0.635 CV abundance −0.01 −0.04–0.02 0.513 maximum temperature 0.04 0.02–0.07 0.002 minimum temperature −0.03 −0.05–0.00 0.019 average precipitation 0.03 0.00–0.05 0.023 geographic range −0.02 −0.04–0.01 0.153 observations 93 R2 Nagelkerke 0.359 Downloaded from https://royalsocietypublishing.org/ on 07 April 2022 royalsocietypublishing.org/journal/rsbl Biol. Lett. 18: 20210519phylogenetic signal in the residuals of any GLS models (elec- tronic supplementary material, table S2). Table 1 shows a detailed model output for the top-performing physiological (e.g. climatic sensitivity). The climatic sensitivity model rep- resented 91.2% of the AICc weight among all models. Morphological traits did not significantly predict changes in abundance over time and only explained 7% of the variation in population trends. The best combined morphological and climate sensitivity model explained 41% of the variation in population trends but was 4.7 AICc units below the best-per- forming physiological model. Our strongest predictors of population trends were variables measuring climatic sensi- tivity (table 1). A positive association existed between population trends and sensitivity to average monthly precipi- tation (figure 2a). Based on standardized beta coefficients, the strongest predictor of population trends was sensitivity to the average monthly maximum temperature (figure 2b), with a standardized coefficient of 0.04 (CI 95% 0.02–0.07), suggesting that moths that are more abundant during warmer years are more likely to show increasing population trends.4. Discussion Our results highlight the vital role of specific differences in climate sensitivity in explaining variation in population trends in this tropical moth community. We indicate that most species have either increased or remained temporally stable over the past 12 years but are likely to be further influenced by future climate changes in Panama. The increas- ing and stable population dynamics contrast with observed sharp declines in caterpillar density in Costa Rica [33,34]. Cli- mate change may have been a driving force in the decrease observed in the Costa Rican studies, but other factors, such as land-use changes and agricultural practices, likely induced decline. Our survey on BCI, an isolated protected forest island, indicates that common tiger moths showed wide- spread increases and temporal population stability. Although several species show strong evidence of decline (figure 1), the overall temporal stability in arctiinepopulations highlights that insect declines are not homo- geneous. Since more than 60% of tiger moth species have strong evidence of increasing in abundance since 2009, our results also contrast with other studies [33–35]. We hypoth- esize that this pattern may have important implications locally, with cascading impacts driven by herbivory and predation at higher trophic levels. Although it is well- established that climate change affects species distributions and abundances of insect herbivores [33,34,36], the impacts of climate change on trophic interactions have been less studied [10]. Outbreak species may benefit from climate changes, as reported for two Panamanian species [5]. In a previous study on BCI, we observed that populations of some large Saturniidae species are increasing [28]. We also showed that recent climate anomalies occurring in the tro- pics, such as increasing average precipitation on BCI [25], have significant and positive effects on the abundance of tiger moths (figure 2a). A similar trend has been observed in the United Kingdom [37] but is also likely driven by differ- ential responses to land-use change. We expected that morphological traits relating to climate, especially thermal tolerance, would predict temporal trends. Our results contra- dict this expectation, and while phylogenetic information does increase the proportion of variance explained, this comes at a high cost in terms of model parameters. Hence, it is unlikely that any of the morphological traits that we measured may be significant predictors of response to cli- mate, although we cannot rule out that such traits exist. Few studies have found that functional traits predict population trends [22,35]. Species-specific climate sensitivity traits were the best pre- dictors of temporal trends of tiger moths on BCI. Sensitivity to average precipitation showed a significant and positive relationship with population trends. Species that were more abundant in months with higher precipitation showed positive population trends (figure 2). Sensitivity to average maximum temperatures also predicted temporal trends, indicating that population abundances of species that were twice as abundant in months with a one-degree increase in temperature have increased by 5% each year. Increased temperatures facilitate more frequent, longer or more effective territorial and mate-locating behaviours [38]. Pro- longed exposure at extreme temperatures can also influence the pace of insect life cycles, thus affecting developmental time and population growth rates [39,40]. The inclusion of thermal tolerance measurements is primordial to correctly interpret moth population dynamics patterns [11–14]. Our analysis provides evidence of a stable and increasing tropical moth community. Still, it highlights the potential future impact of climate change, as climatic sensitivity traits were the best predictors of population trends. Since 1981, BCI has experienced a 17.9% increase in mean annual pre- cipitation [19], and we showed that moth populations that respond to increasing precipitation in Panama are also increasing. With increasing air temperature also predicted for tropical regions by recent models [41,42], this species group may indeed be favoured by future environmental conditions. However, future phenotypic responses and upper levels of thermal tolerance are hard to predict. Should the rate of warming exceed physiological response capacities, we can expect sharp declines in population density for many tropical insect species. moths decrease with moths increase with 5 (a) more precipitation more precipitation 1.4 b = 10.93 std. b = 0.03 1.2 moths increasing 1.0 moths decreasing 0.8 0.990 0.995 1.000 1.005 sensitivity to avg. monthly precip. (beta coef.) (b) moths decrease with moths increase with higher max. temperatures higher max. temperatures 1.4 1.2 moths increasing 1.0 moths b = 0.06 decreasing std. b = 0.042 0.8 1 2 3 4 sensitivity to avg. monthly max temp (beta coef.) Figure 2. Exponentiated rates of change in Arctiinae abundance regressed against the exponentiated coefficients of (a) sensitivity to average monthly precipitation and (b) sensitivity to average maximum monthly temperatures. The fitted line and 95% confidence intervals are from multiple linear regression, and the raw and standardized beta coefficients are shown within each figure. The dashed horizontal and vertical lines at 1 for each axis represent coefficient values when there is no multiplicative change in trend over the years (y-axis) or no multiplicative change in abundance in response to either average precipitation or maximum temperature. A value of two (x-axis) suggests that species are twice as abundant in months with a 1°C increase in average monthly maximum temperature or a 1 mm increase in average monthly precipitation. Downloaded from https://royalsocietypublishing.org/ on 07 April 2022 population trend population trend royalsocietypublishing.org/journal/rsbl Biol. Lett. 18: 20210519Data accessibility. The dataset is publicly available on FigShare: https:// smithsonian.figshare.com/articles/dataset/More_winners_than_losers_ over_12_years_of_monitoring_tiger_moths_Erebidae_Arctiinae_on_ Barro_Colorado_Island_Panama/16850218 [29]. Authors’ contributions. G.P.A.L.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualiza- tion, writing—original draft and writing—review and editing; N.A.P.: conceptualization, formal analysis, investigation, methodology, project administration, resources, software, supervision, validation, visualiza- tion, writing—original draft and writing—review and editing; S.T.S.: conceptualization, formal analysis, investigation, methodology, resources, software, supervision, validation, visualization, writing— original draft and writing—review and editing; C.N.H.: conceptualiz- ation, data curation, investigation, methodology, resources, software, supervision, validation, visualization, writing—original draft and writing—review and editing; M.L.: conceptualization, data curation, investigation, methodology, validation, visualization, writing— original draft andwriting—reviewandediting; B.V.: conceptualization, data curation, formal analysis, investigation, methodology, validation, visualization,writing—original draft andwriting—reviewand editing;Y.L.: conceptualization, data curation, investigation, methodology, project administration, resources, software, supervision, validation, writing—original draft and writing—review and editing; F.P.: concep- tualization, data curation, investigation, methodology, project administration, resources, software, supervision, validation, writ- ing—original draft and writing—review and editing; R.B.: conceptualization, data curation, investigation, methodology, project administration, resources, software, supervision, validation, visualiza- tion, writing—original draft andwriting—review and editing; J.A.R.S.: conceptualization, data curation, investigation, methodology, project administration, resources, software, supervision, validation, visualiza- tion, writing—original draft and writing—review and editing; Y.B.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, soft- ware, supervision, validation, visualization, writing—original draft and writing—review and editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests. We declare we have no competing interests. Funding. Grants of the Czech Science Foundation to G.P.A.L. (grant no. GAČR19-15645Y) and Y.B. (grant no. GAČR20-31295S). 6 Downloaded from https://royalsocietypublishing.org/ on 07 April 2022 Acknowledgements. We thank ForestGEO and the Smithsonian Tropical Research Institute for logistical support. Y.B. was supported by theSmithsonian Barcoding Opportunity and the Sistema Nacional de Investigación SENACYT (Panama).royalsocietypReferencesublishing.org/journal/rsbl Biol. Lett. 18: 202105191. Dirzo R, Young HS, Galetti M, Ceballos G, Isaac NJB, Collen B. 2014 Defaunation in the Anthropocene. Science 345, 401–406. (doi:10.1126/science. 1251817) 2. Sánchez-Bayo F, Wyckhuys KAG. 2019 Worldwide decline of the entomofauna: a review of its drivers. Biol. Conserv. 232, 8–27. (doi:10.1016/j.biocon. 2019.01.020) 3. Wagner DL, Fox R, Salcido DM, Dyer LA. 2021 A window to the world of global insect declines: Moth biodiversity trends are complex and heterogeneous. Proc. Natl Acad. Sci. USA 118, e2002549117. (doi:10.1073/pnas.2002549117) 4. Basset Y, Lamarre GPA. 2019 Toward a world that values insects. Science 364, 1230–1231. (doi:10. 1126/science.aaw7071) 5. Van Bael S, Aiello A, Valderrama A, Medianero E, Samaniego M, Wright JS. 2004 General herbivore outbreak following an El Niño-related drought in a lowland Panamanian forest. J. Trop. Ecol. 20, 625–633. (doi:10.1017/S0266467404001725) 6. Battisti A, Stastny M, Netherer S, Robinet C, Schopf A, Roques A, Larsson S. 2005 Expansion of geographic range in the pine processionary moth caused by increased winter temperatures. Ecol. Appl. 15, 2084–2096. (doi:10.1890/04-1903) 7. Sekar S. 2012 A meta-analysis of the traits affecting dispersal ability in butterflies: can wingspan be used as a proxy? J. Anim. Ecol. 81, 174–184. (doi:10.1111/j.1365-2656.2011. 01909.x) 8. Detto M, Wright JS, Calderon O, Muller-Landau H. 2018 Resource acquisition and reproductive strategies of tropical forest in response to the El Niño–Southern Oscillation. Nature Comm. 9, 913. (doi:10.1038/s41467-018-03306-9) 9. Eggleton P. 2020 The state of the World’s insects. Annu. Rev. Environ. Resour. 45, 61–82. (doi:10. 1146/annurev-environ-012420-050035) 10. Schleuning M et al. 2020 Trait-based assessments of climate-change impacts on interacting species. Trends Ecol. Evol. 35, 319–328. (doi:10.1016/j.tree. 2019.12.010) 11. Deutsch CA, Tewksbury JJ, Huey RB, Sheldon KS, Ghalambor CK, Haak DC, Martin PR. 2008 Impacts of climate warming on terrestrial ectotherms across latitude. Proc. Natl Acad. Sci. USA 105, 6668–6672. (doi:10.1073/pnas.0709472105) 12. Kaspari M, Clay NA, Lucas J, Yanoviak SP, Kay A. 2015 Thermal adaptation generates a diversity of thermal limits in a rainforest ant community. Glob. Chang. Biol. 21, 1092–1102. (doi:10.1111/gcb. 12750) 13. García-Robledo C, Kuprewicz EK, Staines CL, Erwin TL, Kress WJ. 2016 Limited tolerance by insects to high temperatures across tropical elevationalgradients and the implications of global warming for extinction. Proc. Natl Acad. Sci. USA 113, 680–685. (doi:10.1073/pnas.1507681113) 14. Pincebourde S, Casas J. 2019 Narrow safety margin in the phyllosphere during thermal extremes. Proc. Natl Acad. Sci. USA 116, 588–596. (doi:10.1073/ pnas.1815828116) 15. Zenker M, Wahlberg N, Brehm G, Teston JA, Przybylowicz L, Pie MR, Freitas AV. 2017 Systematics and origin of moths in the subfamily Arctiinae (Lepidoptera, Erebidae) in the Neotropical región. Zool. Scripta 46, 348–362. (doi:10.1111/zsc.12202) 16. Chialvo SCH, Holland JD, Anderson TJ, Breinholt JW, Kawahara ZX, Liu S, Zaspel JM. 2018 A phylogenomic analysis of lichen-feeding tiger moths uncovers evolutionary origins of host chemical sequestration. Mol. Phylogenetics Evol. 121, 23–34. (doi:10.1016/j.ympev.2017.12.015) 17. Gawne R, Nijhout HF. 2020 The Arctiid Archetype: a new Lepidopteran Groundplan. Front. Ecol. Evol. 8, 175. (doi:10.3389/fevo.2020.00175) 18. Parmesan C. 2006 Ecological and evolutionary responses to recent climate change. Annu. Rev. Ecol. Evol. Syst. 37, 637–669. (doi:10.1146/annurev. ecolsys.37.091305.110100) 19. Lavorel S et al. 2013 A novel framework for linking functional diversity of plants with other trophic levels for the quantification of ecosystem services. J. Veg. Sci. 24, 942–948. (doi:10.1111/ jvs.12083) 20. Kissling WD et al. 2018 Towards global data products of essential biodiversity variables on species traits. Nat. Ecol. Evol. 2, 1531–1540. (doi:10. 1038/s41559-018-0667-3) 21. Slade EM, Merckx T, Riutta T, Bebber DP, Redhead D, Riordan P, Macdonald DW. 2013 Life-history traits and landscape characteristics predict macro-moth responses to forest fragmentation. Ecology 94, 1519–1530. (doi:10.1890/12-1366.1) 22. Coulthard E, Norrey J, Shortall C, Harris EW. 2019 Ecological traits predict population changes in moths. Biol. Conserv. 233, 213–219. (doi:10.1016/j. biocon.2019.02.023) 23. Heidrich L, Friess N, Fiedler K, Brandle M, Hausmann A, Brandl R, Zeuss D. 2018 The dark side of Lepidoptera: colour lightness of geometrid moths decreases with increasing latitude. Glob. Ecol. Biogeogr. 8, 1–10. (doi:10.1111/geb.12703) 24. Clusella-Trullas S, Nielsen M. 2020 The evolution of insect body coloration under changing climates. Curr. Opin. Insect. Sci. 41, 25–32. (doi:10.1016/j. cois.2020.05.007) 25. Anderson-Teixeira KJ et al. 2015 CTFS ForestGEO: a worldwide network monitoring forests in an era of global change. Glob. Chang. Biol. 21, 528–549. (doi:10.1111/gcb.12712)26. Lucas M, Forero D, Basset Y. 2016 Diversity and recent population trends of assassin bugs (Hemiptera: Reduviidae) on Barro Colorado Island, Panama. Insect Conserv. Divers. 9, 546–558. (doi:10. 1111/icad.12191) 27. Hackforth CN et al. In preparation. Functional classification of Neotropical tiger moths (Erebidae- Arctiinae) provides insight on vulnerability to climate change. Manuscript in preparation for Funct. Ecol. 28. Basset Y et al. 2017 The Saturniidae of Barro Colorado Island, Panama: a model taxon for studying the long-term effects of climate change? Ecol. Evol. 7, 1–14. (doi:10.1002/ece3.3515) 29. Lamarre GPA et al. 2022 Data from: More winners than losers over 12 years of monitoring tiger moths (Erebidae: Arctiinae) on Barro Colorado Island, Panama. FigShare. See https://smithsonian.figshare.com/articles/ dataset/More_winners_than_losers_over_12_years_ of_monitoring_tiger_moths_Erebidae_Arctiinae_on_ Barro_Colorado_Island_Panama/16850218. 30. Hartig F. 2019 DHARMa: residual diagnostics for hierarchical (multi-level/mixed) regression models. R package version 0.2.4. 31. Pinheiro J, Bates D, DebRoy S, Sarkar D, Heisterkamp, S, Van Willigen B, Maintainer R. 2017 Package ‘nlme’. Linear and nonlinear mixed effects models, version, 3(1). 32. Kembel S, Cowan P, Helmus M, Cornwell W, Morlon H, Ackerly D, Blomberg S, Webb C. 2010 Picante: R tools for integrating phylogenies and ecology. Bioinformatics 26, 1463–1464. (doi:10.1093/ bioinformatics/btq166) 33. Salcido DM, Forister ML, Garcia Lopez H, Dyer LA. 2020 Loss of dominant caterpillar genera in a protected tropical forest. Sci. Rep. 10, 422. (doi:10. 1038/s41598-019-57226-9) 34. Janzen DH, Hallwachs W. 2021 To us insectometers, it is clear that insect decline in our Costa Rican tropics is real, so let’s be kind to the survivors. Proc. Natl Acad. Sci. USA 118, e2002546117. (doi:10. 1073/pnas.2002546117) 35. Forister ML et al. 2021 Fewer butterflies seen by community scientists across the warming and drying landscapes of the American West. Science 371, 1042–1045. (doi:10.1126/science.abe5585) 36. Fox J et al. 2012 Package ‘car’, p. 16. Vienna, Austria: R Foundation for Statistical Computing. 37. Fox R et al. 2021 The state of Britain’s larger moths 2021. Wareham, UK: Butterfly Conservation, Rothamsted Research and UK Centre for Ecology & Hydrology. See https://butterfly-conservation.org/sites/ default/files/2021-03/StateofMothsReport2021.pdf. 38. Hayes MP, Hitchcock GE, Knock RI, Lucas CBH, Tuerner EC. 2019 Temperature and territoriality in the Duke of Burgundy butterfly, Hamearis lucina. J. 7 Downloaded from https://royalsocietypublishing.org/ on 07 April 2022 royalsocietypublishing.orgIns. Cons. 23, 739–750. (doi:10.1007/s10841-019- 00166-6) 39. Rebaudo F, Rabhi VB. 2018 Modeling temperature-dependent development rate and phenology in insects: review of major developments, challenges, and future directions. Entom. Exp. Appl. 166, 607–617. (doi:10.1111/eea.12693) 40. Iltis C, Louapre P, Pecharova K, Thiery D, Zito S, Bois B, Moreau J. 2019 Are life-history traitsequally affected by global warming? A case study combining a multi-trait approach with fine-grain climate modeling. J. Ins. Physiol. 117, 103916. (doi:10.1016/j.jinsphys. 2019.103916) 41. IPCC. 2021 Climate change 2021: the physical science basis. Contribution of working group I to the sixth assessment report of the Intergovernmental Panel on Climate Change (eds V. Masson-Delmotteet al.). Cambridge, UK: Cambridge University Press. See https://www.ipcc.ch/report/ar6/wg1/ #FullReport. 42. Stocker T et al. 2014 Climate change 2013: the physical science basis. Contribution of working group I to the fifth assessment report of the Intergovernmental Panel on Climate Change. Cambridge, UK: Cambridge University Press. See https://www.ipcc.ch/report/ar5/wg1/./journal/rsbl Biol. Lett. 18: 20210519