Local‐scale temperature gradients driven by human disturbance shape the physiological and morphological traits of dung beetle communities in a Bornean oil palm–forest mosaic

Read the free Plain Language Summary for this article on Journal blog. Perubahan suhu sering diandaikan namun jarang diuji, sebagai mekanisma di mana spesies-spesies sensitif akan berkurang lanskap hutan berikutan kemerosotan habitat, fragmentasi dan kemusnahan hutan. Trait mengetengahkan bagaimana sesuatu spesies bertindakbalas kepada perubahan persekitaran, dengan trait fisiologi, morfologi tingkah laku merupakan utama yang menentukan tindakbalas ektoterma terhadap suhu. Kami mengumpul data mengenai berkait termal-sensitif (kritikal termal-maksima, saiz badan, kecerahan kutikel pilositi) bagi 46 kumbang tahi (Scarabaeinae) mozek kelapa sawit Borneo Malaysia. Dengan menggabungkan data-data ini persampelan komuniti berskala besar (>59, 0000 individu disampel daripada >600 perangkap) peta termal-udara LiDAR, kami menyelidik trait-pengantara peringkat menggunakan model hierarki, mendapati bahawa kritikal termal-maksimal meramalkan maksimum. Keputusan dicerminkan melalui analisis disamping corak sama pada trait-trait termal lain. Pertambahan badan pengurangan pilositi tinggi manakala menunjukkan perkaitan kompleks merentang gradien gangguan Penemuan menegaskan berpotensi dalam menyebabkan khusus telah diubahsuai mengakibatkan proses komuniti. Temperatures are rising in tropics, both from rapid land-use change (Senior et al., 2017) and anthropogenic climate (Malhi & Wright, 2004). Understanding scale magnitude of temperature effects biodiversity is more urgent than ever, as regional local may synergistically drive decline sensitive species (Frishkoff 2016). In particular, understanding temperature-related threats to insects importance given key role they play driving ecosystem functioning, their reported global declines (Wagner, 2020). Many tropical forests highly degraded fragmented (Haddad 2015), with remnant varying ability act microclimatic refugia hostile agricultural landscapes (Williamson 2021). As such, microclimate varies considerably across oil palm–forest mosaics due differences vegetation (Jucker 2018) and/or proximity habitat edges 2021), resulting a gradient increasing disturbance such Forests also exhibit floor canopy (Hardwick making lowland three-dimensional thermal complexity. Insect body temperature, development physiology intrinsically tightly linked at fine (Pincebourde Woods, Coupled fact that ectotherms already be operating close maxima (Deutsch 2008), tend isolated cool adapted limited variation (Wright 2009), outlook invertebrates bleak. Although many studies have investigated how respond environmental gradients, few identified underlying physiological or morphological traits, critical limits, size cuticle lightness, predict response communities local-scale increases (Bartomeus 2018; Boyle Dung beetles recognised useful ecological indicators forests. Not only community responses correlated those other taxa, but can sampled relative ease low cost (Nichols Gardner, 2011), carry out functions 2008). Tropical dung beetle been gradients associated land-use, topography (Gebert 2019; Larsen, 2012; Williamson Thermal niche differentiation an important factor influencing diversity (Verdú 2007), recently temporal spatial abundance activity (Gotcha 2020; Verdú 2019). Critical (CTmax) being used increasingly demonstrate invertebrate (Boyle Kaspari investigate ecology different biomes (e.g. Gotcha Machekano 2021; Sheldon Tewksbury, 2014; That said, to-date CTmax restricted investigating one handful (with notable exception early work by Gaston Chown, 1999), rather community-level patterns. exoskeleton thickness (Amore coloration (Stanbrook 2021) 2006), thought influence sensitivity. Cuticular lightness largely dictated melanin levels (True, 2003). There hypotheses melanism mediate temperature. First, darker absorb solar radiation, will prevalent cooler environments (Clusella-Trullas 2007). contrast, another hypothesis states increased radiation hotter environments, prevalent, mitigate deleterious UV-B (Bishop Similarly, finally, individuals less permeable cuticles, predicted dominant drier (and thus hotter) areas (Law Indeed, recent study ants showed Different systems, however 2016; Law Stanbrook relationship between complex, associations depending region Fritz 2009; Nichols 2013; Pocock, 2011). beetles, some found negative (Braga Gardner meta-analysis Fuzessy al. (2021) disturbance, although pattern was detected landscapes. Finally, hair retards loss heat body, meaning density should greater (Peters pilosity predicting bee Given stressors complex (Parrett Raine 2018), taxa similar sensitivities through combinations adaptations, resolving broad patterns trait-mediated challenging. Identifying vital, however, if we understand future affect processes. using traits paramount entomologists, natural histories most unknown, while measured (Brousseau 2018). To efficacy means collected mosaic Borneo. addition measurements, size, data. By combining our measurements Light Detection Ranging (LiDAR)-derived projections landscape, mediated traits. Based previous studies, hypothesised positive would higher 2020) lower We expected were unsure directionality competing relevant literature (body size—Gardner 2008; lightness—Bishop All carried across, to, Stability Altered Forest Ecosystems (SAFE) project (116.57′E–117.42′E, 4.38′N–4.46′N), large-scale fragmentation experiment Sabah, Malaysian standardised sampling points landscape palm plantation forest (Ewers The area dipterocarp underwent selective logging 1970s, areas, second round 2000s (Struebig 2013) followed salvage 2013 2015. Much now plantation, large tract continuous logged north. relatively aseasonal, there wetter season October March. mean annual rainfall ~2700 mm 26.7°C (Walsh Newbery, has become since these taken (Chapman Standardised 30 g human dung-baited pitfall traps (following Slade 2011) deployed 48 hr time (31 sites levels), series three trapping campaigns 2011, 2015 2017/2018 (hereafter referred campaigns, Table S1; Figure 1). conducted September April coincide marginally part year when active. Traps set riparian transects 10 spaced approximately 200 m apart alternating sides river (see Non-riparian SAFE second-order minimum 175 Ewers Beetles stored 70% ethanol freezer reference collections Universiti Malaysia Sabah Oxford University Museum Natural History. Samples under permits EPU Ruj. UPE: 40/200/19/2712; MBMC Project No. 129; SaBC Access Licences JKM/MBS.1000–2/2(381), JKM/MBS.1000–2/2JLD.6(63) JKM/MBS.1000–2/2JLD.7(83). measure physiology, additional targeting terrestrial arboreal beetles. Terrestrial non-lethal, Parrett 2019) along (Figure 1), two nearby plantations, December 2017, 2018, March 2019. Additional aerial within target Supporting Information, Methods). baited homogenised 24 hr. A total 623 (n = 1–56 per species, 13.5, median 10) transported laboratory cool, dark location maximum 4 before measuring tolerances. Individuals placed 50 ml polyethylene tubes submerged water bath (Grant Instruments TXF200) 34°C 5 min. then 0.1°C/min up 43.5°C, minimise delayed increases, limit death desiccation recorded which muscles started spasming uncontrollably could no longer right themselves walk down tube 2014). thermocouple ensured lag never exceeded 0.2°C. Ethical approval assays granted Queen Mary London's Animal Welfare Review Board. did not warrant ethical approval. Three related changes taken. Body length (clypeus pygidium), 364 37 (all had data, n 1–40 9.8, 3) 2018 campaigns. Measurements Microsoft Paint images generated Leica M205C stereo microscope MC190 HD camera. Lightness 8-bit luminance values (0 black, 255 white) front pronotum back elytra S1). Following Roquer-Beni (2020), pronotal elytral (hairs/mm2) estimated 0.1 × side Pronotal averaged give each individual. Functional guilds (daily nesting strategy) assigned following Gray (2014) Davis Sutton (1998) qualitative comparisons, statistically, strong non-independence arises account shared evolutionary history. assess whether tolerances varied constructed one-way ANOVAs species. Estimated marginal (EMMs) post hoc compare (Tukey-adjusted). Significance ascertained likelihood ratio tests (LRTs) comparison respective null lmtest package (Zeileis Hothorn, 2002). Mean daily (Tmax) extracted 629 samples 331 unique existing Jucker (2018). (2018) structural equation models ground-based, sensors vegetative (plant index height) topographic (topographic position index, aspect, elevation slope) airborne LiDAR metrics 2014. Generated resolution Tmax here over variables extremes (Vasseur defined separate mixed-effects CWM (CTmax, length, pilosity) campaign fixed effect, type effect point random (to control repeated structure) lme4 (Bates 2015). without communities, ascertain landscape-level results consistent sites. Additive containing run covariance. LRTs all (containing terms). test robustness models, fit independently. repeatedly campaign, linear regressions point. For >100 individuals, ran generalised (Poisson error-family) effects, glmmTMB (Brooks 2017). Estimates coefficient proxy species' subsequent analyses responses). responses, accounting phylogenetic relatedness, Markov chain Monte Carlo (MCMC) techniques MCMCglmm (Hadfield, 2010) maximal CTmax, covariance matrix distance (Supporting Methods; S3). Models iterated 5,000,000 times, thinning interval 500 iterations burn-in 50,000. An inverse-Gamma distribution covariate residual variance structure uninformative priors (V 1, ν 0.02). Each times seeds. Chain mixing assessed potential reduction factors checked equivalence. deviance information criterion selected representative. pMCMCs 2010). Non-significant sequentially removed re-fit until adequate remained. Posterior distributions, traces effective sample sizes examined fit. results, slopes any same methods described above models. R v4.0.5 (R Development Core Team, representing 11 genera 2), species-rich date. significantly among ( X 1 , 2 1,144.9, p < 0.001, LRT compared model) 598.3, 0.001). Of where ≥ 3, Oniticellus tesselatus Onthophagus agg. pacificus highest lowest (42.70°C ± 0.31 SE 38.19°C respectively), exhibiting genus (42.70 39.09°C respectively). Catharsius dayacus C. renaudpauliani—sister majority biomass Bornean (Gray 2011)—had (Tukey-adjusted EMMs, 3), undisturbed-forest specialist having generalist renaudpauliani (CTmax 39.94°C 0.09 41.56°C 0.07 This difference held true excluding (logged 41.23°C 0.09, Tukey's HSD test, trituber noteworthy second-highest (42.57°C 0.13) despite belonging 2; see S5). 59,597 traps, 59,004 belonged 58,658 (length, pilosity). increase 1°C 0.088°C (±0.018 SE) 23.1, model, S5), 0.52°C (from 40.27 40.79°C, 4) range (26.29–32.28°C). Oil 62.7, 40.97°C 40.41°C 0.02, respectively, 4; Community Tmax, 47.2, 1.08 0.15 S5, 88.4, 12.31 0.16 18.04 0.52 S5; 4). polynomial 52.4, polynomial, peaking intermediate temperatures extreme vary types (Table Pilosity decreased 18.1, −5.98 hairs/mm2 1.37 decrease 4), 54.4, 99.65 1.53 59.33 4.97, comparable considering S6), indicating just driven communities. robust single habitats, lighter S7). Species phylogeny (posterior 0.43, UCI 0.60, LCI 0.24, pMCMC <0.001, S8; 5). controlling S9). signal; S4). show statistical power, posterior qualitatively full mean, non-significant S8). LiDAR-derived maps, thermoregulatory change. Community-level (Tmax). addition, high tended locations temperatures. Together, species- linking upper limits sensitivity represent evidence disturbance. demonstrated thermoregulation gradient. uncover lighter-coloured temperatures, ones extremes. whole habitats. Importantly, abundances parts relatedness. Taken conjunction, cement idea determinant proxies (despite prominent criticisms—see Rezende (2020) 0.2°C every (compared here). Larger biology question. Ant workers flightless, exploring environment nest. volant able travel long distances Taxon-specific vagility alter drives distributions. Furthermore, bury soil strategy (tunnellers rollers), allowing them thermoregulate behaviourally. find anecdotal supporting idea, non-fossorial represented here. Ascertaining larval survival rates controlled pinpoint bottlenecks life cycles (Dahlke identify primary driver land-use. exhibited signal, several closely divergent physiologies. Sister largest Borneo, together comprising habitats (Slade From old-growth turnover dayacus, renaudpauliani. shown partitioning, it does sympatric intraspecific suggests plasticity (i.e. product acclimation), subpopulations (Rohr Additionally, tunneller, CTmax. best knowledge, first species-level certainly context. congenerics noteworthy, taxonomic identification possible genus, fail resolve drivers accurately safety margins (here CTmax) experienced organisms (Sinclair projected (32.3°C point) below (38.2°C), might naively assume unthreatened ongoing invertebrates, subject substrate exist upon, experiencing world millimetres soil–air boundary (Blonder widespread history remaining passed extinction filter, removing thermally values, absolute maximum, risk scenarios assumed margins. absent well It consider detrimental sub-lethal (van Heerwaarden Sgrò, individual fitness detrimentally affected completely inhibit behaviour ultimately cause death. Chown (1999) South African elevations site ~40.5 51.5°C—a good deal (36.9–43.2°C). indicate particularly preferences, contributing (Fuzessy distinctly 1999; perhaps result buffering canopies cuticles quadratic function These suggest proposed Clusella-Trullas 2007; study, darkest reflect beneficial 2020), becomes disturbed frequent gaps), protection against harmful exposure tolerance form outweigh benefits reflectivity predictive level level, findings (2021). surprising smaller animals equilibration shorter larger 2–2.5 heterothermic, abdominal pumping elevated thorax flying potentially mitigating size. (Larsen ours patchy mammal populations (Deere microrefugia larger-bodied persist spatially structured metapopulations. gradient, plantations keeping bees 2016), invertebrates. frequently insect-pollinator because its pollen acquisition (Roquer-Beni unaware hairiness thermoregulation. Here, quantify often-assumed, rarely empirically tested, here, significant, weak sizes. mismatch (taken 2017/2018) (generated 2014 scan). attempted methodological short-coming generally analysis Alternatively, small fertility limits; van alternative measures coloration; behavioural strategy, analyse statistically) determining presented Responses cumulative, trait. and, ultimately, processes, change, must gradients. traits-of-interest advocate extension contexts better ubiquity findings. studied measurable Peters universally mediating allow ecologists hard-to-survey case, threats, vitally ecosystems impacted climatic warming. funded Environment Research Council (NERC) Human Modified programme (NE/K016261/1; NE/K016377/1) Newton-Ungku Omar Fund via British Industry Government Group High Technology (216433953). NERC PhD studentship JW (NE/L002485/1) research fellowship TJ (NE/S01537X/1). EMS acknowledges funding BES Small Ecological Grant, No.: 3256/4035, Varley-Gradwell Travelling Fellowship Ecology. David Coomes instrumental generation projections. Thanks go Sol Milne Herry Heroin help field, Darren Mann assistance identification, Paul Eggleton his support advice, Rob Knell Kimberly feedback earlier version manuscript. doubly thank guidance. Sui Peng Heon translating abstract. grateful SEARRP staff who made possible: Unding Jami, Johnny Larenus, Amir, Anis, David, Didy, Dino, Joanni, Kiki, Loly, Mudin, Noy Zul Ong Xin Rui providing images. authors conflict interest declare. J.W., E.M.S., S.J.R. O.T.L. conceived study. E.M.S. led sampling, A.Y.C.C. E.B. J.W. J.P. specimen identification. E.T. T.J. provided M.B. variable evolution wrote manuscript input authors. available Zenodo repository. Specifically, https://doi.org/10.5281/zenodo.3906118 (Slade, Milne, https://doi.org/10.5281/zenodo.3906441 Williamson, respectively. non-riparian https://doi.org/10.5281/zenodo.3247492 Bush, 2019), https://doi.org/10.5281/zenodo.3247494 https://doi.org/10.5281/zenodo.3832076 Chung, https://doi.org/10.5281/zenodo.6477719 2022), https://doi.org/10.5281/zenodo.6477751 2022). Appendix S1. Please note: publisher responsible content functionality supplied Any queries (other missing content) directed corresponding author article.