Ecology

Sampling sites and sample preparationCulms of three- to four-year-old of Bambusa vulgaris Schrad. ex J.C., B. maculata Widjaja, and Gigantochloa atter (Hassk) Kurz ex Munro were collected from naturally bamboo forests at four sites in Lombok Island, Indonesia23. The culm age was estimated based on some morphological features (the presence of culm sheath, color, and sound created by tapping with fingers) checked by an experienced bamboo farmer. Figure 1 shows the map of sampling sites and climatic conditions of the sites. Ten individual culms in each species at each site were collected from different clumps and cut 20 cm above the ground (Fig. 2). A total of 120 culms (three species × four sites × 10 individual culms from 10 individual clumps) were collected in the present study (Fig. 2). To determine the longitudinal variations of the anatomical characteristics and mechanical properties, the internode section was collected at 2-m intervals from 2 to 8 m above the ground; a total of 480 internode sections. (120 culms × four heights) were obtained from three species (Fig. 2). The collection of bamboo culms was permitted by Indonesian Institute of Science (Reference no. B-206/SKIKH/KS.02.04/X/2020) and complied with relevant guidelines and regulations of Indonesian CITES Management Authority, Ministry of Environment and Forestry, Indonesia. In addition, the voucher specimen was deposited at the Herbarium Lesser Sunda, University of Mataram, Indonesia under the voucher number of DSR01, 02, and 03 (specimens were identified by Mr. Niechi Valentino). Table 1 shows the culm diameter at 1.3 m above the ground, total culm height, and mean value of culm thickness at four positions23.Figure 1Locations and climate conditions of sampling sites in the present study23. Note: Site I, Tempos (8°41′59″ S, 116°8′40″ E); Site II, Kabul (8°47′21″ S, 116°10′21″ E); Site III, Keruak (8°45′45″ S, 116°28′54″ E); Site IV, Genggelang (8°23′16″ S, 116°15′35″ E). *, mean annual precipitation. The value in the bracket is the mean annual temperature. Climate data were provided from Nusa Tenggara River Basin Management I, Indonesia. Mean monthly temperature and precipitation were calculated by averaging monthly values from 2016 to 2018. Bars indicate the mean values of precipitation. Circles indicate the mean values of temperature. The graph was originally created by R27 (version 4.0.3, https://www.R-project.org/).Full size imageFigure 2Photographs of the clumps in three bamboo species (a–c) and schematic diagrams of experimental procedures (d). Note: a, B. vulgaris; b, B. maculata; c, G. atter. The specimens of fiber area measurement and mechanical properties have the whole culm thickness (including the cortex and inner part of the culm) in the radial direction.Full size imageTable 1 Mean values and standard deviations of growth characteristics in three bamboo species at each site23.Full size tableAnatomical characteristicsThe internode sections were split into two parts: the strips (10 mm in the longitudinal direction) and the small blocks (10 [T] mm by 10 [L] mm by culm thickness in the radial direction) (Fig. 2). The strips and small blocks were the samples for measuring fiber length and fiber area, respectively. In the present study, the fiber area was defined as the sheaths area around the vascular bundles24.To determine the fiber length, small sticks (not including the cortex and the most inner part of the culm) were obtained from the strips with a razor blade (Fig. 2). Randomly selected sticks from each height position (without separation of collected positions of the samples within the radial direction of the culm in a height) were macerated with Schultze’s solution (100 mL of 35% nitric acid containing 6 g potassium chloride) at 70 °C for two hours. The length of 50 fibers was measured in each sample with a digital caliper (CD-15CX, Mitutoyo, Kawasaki, Japan) on a microprojector (V-12B, Nikon, Tokyo, Japan).To measure the fiber area, one block was taken at each height position on each individual culm (Fig. 2). The transverse sections of the blocks were polished with sandpaper sheet (#180, 3 M Japan, Tokyo, Japan), and then their images were captured using a microscope digital camera (DS-2210, Sato Shouji Inc., Kawasaki, Japan) attached to a stereo microscope (SZX12, Olympus, Tokyo, Japan). The fiber area was determined by ImageJ25 (version 1.53e). Binarized images were prepared by ImageJ to distinguish as clearly as possible between the vascular bundle and the background (Fig. 3). The darker area of binarized images in Fig. 3 was identified as fiber sheaths. The fiber area was calculated as follows:$$FAleft( % right) , = A_{fs} /A_{c} times {1}00$$
(1)
where FA = fiber area (%), Afs = the transverse-sectional area of fiber sheath in bamboo culm (mm2), and Ac = the transverse-sectional area of bamboo culm (mm2).Figure 3The photomicrographs of transverse section in B. vulgaris (a and d), B. maculata (b and e), and G. atter (c and f). Note: a, b and c, original image; d, e and f, binarized image processed by ImageJ25 (version 1.53e, https://imagej.nih.gov/ij/). The darker area in photomicrographs (d, e and f) is fiber sheath area.Full size imageMechanical propertiesThe following mechanical properties of culm were measured: bending properties (MOE and MOR), CS, and tensile properties (TM and TS). A total of 480 specimens (one specimen × four heights in an individual × ten individuals × three species × four sites) without node were obtained in each property (Fig. 2).The strips (10 [T] mm × 200 [L] mm × varied culm thickness in the radial direction) were prepared as the specimens for the static bending test (Fig. 2). The static bending test was conducted using a universal testing machine (MSC 5/500–2, Tokyo Testing Machine, Tokyo, Japan). A load was applied to the center of the specimen on the outer cortex surface with 180 mm span and 3 mm min−1 load speed. Due to larger thickness (exceeded 12.9 mm = 180 mm of span / 14) in the radial direction, the span / depth ratio in some specimens was less than 14, indicating that MOR in some specimens might be underestimated due to the occurrence of the shearing strength26. Of 480 specimens, the large culm thickness exceeded 12.9 mm was total 19 specimens from B. vulgaris species collected at 2 m height position from different sites (Site I = four specimens, Site II = six specimens, Site III = four specimens, and Site IV = five specimens). However, all these 19 specimens were broken at the tension side of the specimens during static bending test, which was the normal breaking forms of bending specimens with span / depth ratio less than 14.The load and deflection were recorded with a personal computer, and then MOE and MOR were calculated by the following formulae:$$MOE , left( {GPa} right) , = Delta Pl^{3} / , 4Delta Ybh^{3} , times 10^{ – 3}$$
(2)
$$MOR , left( {MPa} right) , = , 3Pl/ , 2bh^{2}$$
(3)
where ΔP = difference between upper and lower proportional limit within the range of elasticity (N), l = length of the span (mm), ∆Y = deflection due to ∆P (mm), b = width of the specimen (mm), h = height of the specimen (mm), and P = maximum load (N).The compressive test specimen (10 [T] mm × 20 [L] mm × culm thickness in the radial direction) was also prepared (Fig. 2). The test was conducted using a universal testing machine (RTF-2350, A&D, Tokyo, Japan) with a load speed of 0.3 mm min−1. The compressive strength parallel to grain (CS) was calculated by the following formula:$${text{CS }}left( {{text{MPa}}} right) , = P/A_{0}$$
(4)
where P = maximum load (N), and A0 = the cross-sectional area of the specimen (mm2).The tensile tests were conducted using bone-shaped specimens (Fig. 2). The specimen length was 230 (L) mm with a 20 (T) mm width of the specimen grip. The cross-sectional area of the specimen was 2 mm in the tangential direction by culm thickness in the radial direction. A strain gage type extensometer (SG25-10A, A&D, Tokyo, Japan) was used to detect the elongation in the test specimen. The specimen grip sections were attached to small boards (75 mm in length × 40 mm in width × 5 mm in thickness) and then were clamped between the metal grip of a universal testing machine (RTC-2410, A&D, Tokyo, Japan). The tensile load was applied at 1 mm min−1. The tensile strength (TS) and Young’s modulus (TM) were calculated by the following formulae:$${text{TS }}left( {{text{MPa}}} right) , = P/A_{0}$$
(5)
$${text{TM }}left( {{text{GPa}}} right) , = Delta Pl/A_{0} Delta l times {1}0^{{ – 3}}$$
(6)
where P = maximum load (N), A0 = the cross-sectional area of the specimen (mm2), ∆P = difference between upper and lower proportional limit within the range of elasticity (N), l = gauge length (mm), and ∆l = elongation of the original gauge length (mm).The moisture content and air-dry density of each specimen were measured after each mechanical testing by the oven-dry method. The moisture content and air-dry density of the specimen at testing were listed in Table S1.Statistical analysisThe statistical analyses were conducted using R software (version 4.0.3)27. To evaluate the longitudinal variations of the measured properties in each species, the y-intercept, linear, and nonlinear mixed-effects models with each measured property value as a responsible variable, the height position as a fixed effect, and site and individual culm as random effects were developed by the “lmer” function in “lme4” packages28 and the “nlme” function in the “nlme” package29. The following four full models were developed and compared:Model I (y-intercept model):$$Y_{ijk} = alpha_{{1}} + Site_{{{1}k}} + Culm_{{{1}jk}} + e_{ijk}$$
(7)
Model II (linear model):$$Y_{ijk} = , (beta_{0} + Site_{0k} + Culm_{0jk} )X_{ijk} + beta_{{1}} + Site_{{{1}k}} + Culm_{{{1}jk}} + e_{ijk}$$
(8)
Model III (logarithmic model):$$Y_{ijk} = , (gamma_{0} + Site_{0k} + Culm_{0jk} ){text{ ln }}left( {X_{ijk} } right) + gamma_{{1}} + Site_{{{1}k}} + Culm_{{{1}jk}} + e_{ijk}$$
(9)
Model IV (quadratic model):$$begin{gathered} Y_{ijk} = , (zeta_{0} + Site_{0k} + Culm_{0jk} )X_{ijk}^{{2}} + , (zeta_{{1}} + Site_{{{1}k}} + Culm_{{{1}jk}} )X_{ijk} hfill \ + zeta_{{2}} + Site_{{{2}k}} + Culm_{{{2}jk}} + e_{ijk} hfill \ end{gathered}$$
(10)

where Yijk is measured property at the ith height position from the jth individual culm within the kth site, Xijk is the ith height position from the jth individual culm within the kth site, α1, β0, β1, γ0, γ1, ζ0, ζ1, and ζ2 are the fixed effects, Site0k, Site1k, and Site2k are the random effect at the site level, Culm0jk, Culm1jk, and Culm2jk are the random effects at the individual culm level, and eijk is residual. Total 36 derived models (three y-intercept models, 15 linear models, nine logarithmic models, and nine quadratic models) were developed. The model selection was conducted using the Akaike information criterion30. The model with the minimum AIC value was regarded as the most parsimonious model among developed models. In addition, the differences in AIC (ΔAIC) ≤ 2 indicate no significant differences between models, and a simpler model with fewer parameters is preferred31. To evaluate the longitudinal variation, estimated values of each property was calculated at 0.1 m interval from 2.0 to 8.0 m above the ground using fixed-effect parameters of the selected models. Mean value and standard deviation were obtained from the estimated values from 2.0 to 8.0 m in each property. In addition, the coefficient of variation was also calculated from the mean value and standard deviation. The longitudinal variation patterns were classified into four types (Types A to D) based on the model selection (Fig. 4). Although model II to IV was selected, longitudinal variation with the coefficient of variation less than 3.0% was regarded as stable (Type A in Fig. 4).Figure 4Classification of longitudinal variation of bamboo culm property. Note: Lines or curves indicate formulae with fixed-effect parameters in the selected mixed-effect model for explaining longitudinal variation (Tables 3, 4, 5). Coefficient of variation calculated from mean values and standard deviation from 2 to 8 m above the ground estimated by fixed-effect parameters values less than 3.0% is regard as stable variation (Type A), even in selected model is Model II to IV.Full size imageGeographic variations in each bamboo property were estimated by evaluating the variance component of sites and culms as random effects by using the intercept-only linear mixed-effects model. The full model is described as follows:$$Y_{ijk} = mu + Site_{k} + Culm_{jk} + e_{ijk}$$
(11)
where Yijk is the bamboo property at the ith height position of the jth individual culm within kth site, μ is the model intercept or grand mean, Sitek is the random effect of the kth site, Culmjk is random effect of jth individual culm within kth site, and eijk is the residual. The contribution of each level of variation was calculated as a percentage of the total random variation in the best model32,33. More

Biodiversity on Earth is threatened by land-use changes, overexploitation of resources, pollution, biological invasions, and current and projected climate change. Understanding how species will respond to these stressors is difficult, in part because stressors don’t occur in isolation, and because responses can trickle through ecological networks due to interactions among species. More

Gartland, L. A., Firth, J. A., Laskowski, K. L., Jeanson, R. & Ioannou, C. C. Sociability as a personality trait in animals: Methods, causes and consequences. Biol. Rev. https://doi.org/10.1111/brv.12823 (2021).Article 

Google Scholar 
Bergmüller, R. & Taborsky, M. Adaptive behavioural syndromes due to strategic niche specialization. BMC Ecol. 7, 12. https://doi.org/10.1186/1472-6785-7-12 (2007).Article 

Google Scholar 
Massen, J. J. M., Sterck, E. H. M. & de Vos, H. Close social associations in animals and humans: Functions and mechanisms of friendship. Behaviour 147, 1379–1412. https://doi.org/10.1163/000579510X528224 (2010).Article 

Google Scholar 
Haller, J., Harold, G., Sandi, C. & Neumann, I. D. Effects of adverse early-life events on aggression and anti-social behaviours in animals and humans. J. Neuroendocrinol. 26, 724–738. https://doi.org/10.1111/jne.12182 (2014).Article 
CAS 

Google Scholar 
Carlson Bruce, A. Early life experiences have complex and long-lasting effects on behavior. Proc. Natl. Acad. Sci. 114, 11571–11573. https://doi.org/10.1073/pnas.1716037114 (2017).Article 
ADS 
CAS 

Google Scholar 
Zablocki-Thomas, P. B. et al. Personality and performance are affected by age and early life parameters in a small primate. Ecol. Evol. 8, 4598–4605. https://doi.org/10.1002/ece3.3833 (2018).Article 

Google Scholar 
Langenhof, M. R. & Komdeur, J. Why and how the early-life environment affects development of coping behaviours. Behav. Ecol. Sociobiol. 72, 34–34. https://doi.org/10.1007/s00265-018-2452-3 (2018).Article 

Google Scholar 
Daros, R. R., Costa, J. H. C., von Keyserlingk, M. A. G., Hötzel, M. J. & Weary, D. M. Separation from the dam causes negative judgement bias in dairy calves. PLoS One 9, e98429. https://doi.org/10.1371/journal.pone.0098429 (2014).Article 
ADS 
CAS 

Google Scholar 
Grześkowiak, ŁM. et al. Impact of early-life events on the susceptibility to Clostridium difficile colonisation and infection in the offspring of the pig. Gut Microbes 10, 251–259. https://doi.org/10.1080/19490976.2018.1518554 (2019).Article 

Google Scholar 
Schmauss, C., Lee-McDermott, Z. & Medina, L. R. Trans-generational effects of early life stress: The role of maternal behavior. Sci. Rep. 4, 4873. https://doi.org/10.1038/srep04873 (2014).Article 
ADS 
CAS 

Google Scholar 
Brask, J. B., Ellis, S. & Croft, D. P. Animal social networks: An introduction for complex systems scientists. J. Complex Netw. 9, cnab001. https://doi.org/10.1093/comnet/cnab001 (2021).Article 

Google Scholar 
Almeling, L., Hammerschmidt, K., Sennhenn-Reulen, H., Freund, A. M. & Fischer, J. Motivational shifts in aging monkeys and the origins of social selectivity. Curr. Biol. 26, 1744–1749. https://doi.org/10.1016/j.cub.2016.04.066 (2016).Article 
CAS 

Google Scholar 
Borgeaud, C., Sosa, S., Sueur, C. & Bshary, R. The influence of demographic variation on social network stability in wild vervet monkeys. Anim. Behav. 134, 155–165. https://doi.org/10.1016/j.anbehav.2017.09.028 (2017).Article 

Google Scholar 
Cantor, M. et al. The importance of individual-to-society feedbacks in animal ecology and evolution. J. Anim. Ecol. 90, 27–44. https://doi.org/10.1111/1365-2656.13336 (2021).Article 

Google Scholar 
Sosa, S., Sueur, C. & Puga-Gonzalez, I. Network measures in animal social network analysis: Their strengths, limits, interpretations and uses. Methods Ecol. Evol. 12, 10–21. https://doi.org/10.1111/2041-210X.13366 (2021).Article 

Google Scholar 
Farine, D. R. & Whitehead, H. Constructing, conducting and interpreting animal social network analysis. J. Anim. Ecol. 84, 1144–1163. https://doi.org/10.1111/1365-2656.12418 (2015).Article 

Google Scholar 
Neethirajan, S. & Kemp, B. Social network analysis in farm animals: Sensor-based approaches. Animals 11, 434. https://doi.org/10.3390/ani11020434 (2021).Article 

Google Scholar 
Whitehead, H. Analyzing Animal Societies. University of Chicago Press, Chicago, IL, USA (2008).
Smith, J. E. & Pinter-Wollman, N. Observing the unwatchable: Integrating automated sensing, naturalistic observations and animal social network analysis in the age of big data. J. Anim. Ecol. 90, 62–75. https://doi.org/10.1111/1365-2656.13362 (2021).Article 

Google Scholar 
Chen, S., Ilany, A., White, B. J., Sanderson, M. W. & Lanzas, C. Spatial-temporal dynamics of high-resolution animal networks: What can we learn from domestic animals?. PLoS One 10, e0129253. https://doi.org/10.1371/journal.pone.0129253 (2015).Article 
CAS 

Google Scholar 
Atton, N., Galef, B. J., Hoppitt, W., Webster, M. M. & Laland, K. N. Familiarity affects social network structure and discovery of prey patch locations in foraging stickleback shoals. Proc. R. Soc. B Biol. Sci. 281, 20140579. https://doi.org/10.1098/rspb.2014.0579 (2014).Article 
CAS 

Google Scholar 
Ilany, A. & Akçay, E. Personality and social networks: A generative model approach. Integr. Comp. Biol. 56, 1197–1205. https://doi.org/10.1093/icb/icw068 (2016).Article 

Google Scholar 
Romano, V. et al. Modeling infection transmission in primate networks to predict centrality-based risk. Am. J. Primatol. 78, 767–779. https://doi.org/10.1002/ajp.22542 (2016).Article 

Google Scholar 
Ren, K., Bernes, G., Hetta, M. & Karlsson, J. Tracking and analysing social interactions in dairy cattle with real-time locating system and machine learning. J. Syst. Archit. 116, 102139. https://doi.org/10.1016/j.sysarc.2021.102139 (2021).Article 

Google Scholar 
Boyland, N. K., Mlynski, D. T., James, R., Brent, L. J. N. & Croft, D. P. The social network structure of a dynamic group of dairy cows: From individual to group level patterns. Appl. Anim. Behav. Sci. 174, 1–10. https://doi.org/10.1016/j.applanim.2015.11.016 (2016).Article 

Google Scholar 
Chopra, K. et al. Proximity interactions in a permanently housed dairy herd: Network structure, consistency, and individual differences. Front. Vet. Sci. 7, 583715 (2020).Article 

Google Scholar 
Šárová, R. et al. Pay respect to the elders: Age, more than body mass, determines dominance in female beef cattle. Anim. Behav. 86, 1315–1323. https://doi.org/10.1016/j.anbehav.2013.10.002 (2013).Article 

Google Scholar 
Foris, B., Haas, H. G., Langbein, J. & Melzer, N. Familiarity influences social networks in dairy cows after regrouping. J. Dairy Sci. 104, 3485–3494. https://doi.org/10.3168/jds.2020-18896 (2021).Article 
CAS 

Google Scholar 
Foris, B., Zebunke, M., Langbein, J. & Melzer, N. Comprehensive analysis of affiliative and agonistic social networks in lactating dairy cattle groups. Appl. Anim. Behav. Sci. 210, 60–67. https://doi.org/10.1016/j.applanim.2018.10.016 (2019).Article 

Google Scholar 
de Freslon, I., Martínez-López, B., Belkhiria, J., Strappini, A. & Monti, G. Use of social network analysis to improve the understanding of social behaviour in dairy cattle and its impact on disease transmission. Appl. Anim. Behav. Sci. 213, 47–54. https://doi.org/10.1016/j.applanim.2019.01.006 (2019).Article 

Google Scholar 
Bolt, S. L., Boyland, N. K., Mlynski, D. T., James, R. & Croft, D. P. Pair housing of dairy calves and age at pairing: Effects on weaning stress, health production and social networks. PLoS One 12, e0166926. https://doi.org/10.1371/journal.pone.0166926 (2017).Article 
CAS 

Google Scholar 
Koene, P. & Ipema, B. Social networks and welfare in future animal management. Animals (Basel) 4, 93–118. https://doi.org/10.3390/ani4010093 (2014).Article 

Google Scholar 
Raussi, S. et al. The formation of preferential relationships at early age in cattle. Behav. Proc. 84, 726–731. https://doi.org/10.1016/j.beproc.2010.05.005 (2010).Article 

Google Scholar 
Weary, D. M., Jasper, J. & Hötzel, M. J. Understanding weaning distress. Appl. Anim. Behav. Sci. 110, 24–41. https://doi.org/10.1016/j.applanim.2007.03.025 (2008).Article 

Google Scholar 
Lopes, P. C., Block, P. & König, B. Infection-induced behavioural changes reduce connectivity and the potential for disease spread in wild mice contact networks. Sci. Rep. 6, 31790. https://doi.org/10.1038/srep31790 (2016).Article 
ADS 
CAS 

Google Scholar 
Ripperger, S. P., Stockmaier, S. & Carter, G. G. Tracking sickness effects on social encounters via continuous proximity sensing in wild vampire bats. Behav. Ecol. 31, 1296–1302. https://doi.org/10.1093/beheco/araa111 (2020).Article 

Google Scholar 
McGuirk, S. M. & Peek, S. F. Timely diagnosis of dairy calf respiratory disease using a standardized scoring system. Anim. Health Res. Rev. 15, 145–147. https://doi.org/10.1017/s1466252314000267 (2014).Article 

Google Scholar 
Callan, R. J. & Garry, F. B. Biosecurity and bovine respiratory disease. Vet. Clin. N. Am. Food Anim. Pract. 18, 57–77. https://doi.org/10.1016/S0749-0720(02)00004-X (2002).Article 

Google Scholar 
Sewio. Tag Leonardo iMU/Personal. https://docs.sewio.net/docs/tag-leonardo-imu-personal-30146967.html (2022).Barker, Z. E. et al. Use of novel sensors combining local positioning and acceleration to measure feeding behavior differences associated with lameness in dairy cattle. J. Dairy Sci. 101, 6310–6321. https://doi.org/10.3168/jds.2016-12172 (2018).Article 
CAS 

Google Scholar 
Team, R. C. R: A Language and Environment for Statistical Computing.Hadfield, J. D. MCMC methods for multi-response generalized linear mixed models: The MCMCglmm R package. J. Stat. Softw. 33, 1–22. https://doi.org/10.18637/jss.v033.i02 (2010).Article 

Google Scholar 
Franks, D. W., Weiss, M. N., Silk, M. J., Perryman, R. J. Y. & Croft, D. P. Calculating effect sizes in animal social network analysis. Methods Ecol. Evol. 12, 33–41. https://doi.org/10.1111/2041-210X.13429 (2021).Article 

Google Scholar 
Bates, D., Mächler, M., Bolker, B. & Walker, S. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67, 1–48. https://doi.org/10.18637/jss.v067.i01 (2015).Article 

Google Scholar 
Bell, D. C., Atkinson, J. S. & Carlson, J. W. Centrality measures for disease transmission networks. Soc. Netw. 21, 1–21. https://doi.org/10.1016/S0378-8733(98)00010-0 (1999).Article 

Google Scholar 
Benjamini, Y. & Hochberg, Y. Controlling the false discovery rate: A practical and powerful approach to multiple testing. J. R. Stat. Soc. Ser. B (Methodol.) 57, 289–300. https://doi.org/10.1111/j.2517-6161.1995.tb02031.x (1995).Article 
MathSciNet 
MATH 

Google Scholar 
PercieduSert, N. et al. Reporting animal research: Explanation and elaboration for the ARRIVE guidelines 2.0. PLoS Biol. 18, e3000411. https://doi.org/10.1371/journal.pbio.3000411 (2020).Article 
CAS 

Google Scholar 
Hulbert, L. E. & Moisá, S. J. Stress, immunity, and the management of calves. J. Dairy Sci. 99, 3199–3216. https://doi.org/10.3168/jds.2015-10198 (2016).Article 
CAS 

Google Scholar 
Sweeney, B. C., Rushen, J., Weary, D. M. & de Passillé, A. M. Duration of weaning, starter intake, and weight gain of dairy calves fed large amounts of milk. J. Dairy Sci. 93, 148–152. https://doi.org/10.3168/jds.2009-2427 (2010).Article 
CAS 

Google Scholar 
Rault, J.-L. Friends with benefits: Social support and its relevance for farm animal welfare. Appl. Anim. Behav. Sci. 136, 1–14. https://doi.org/10.1016/j.applanim.2011.10.002 (2012).Article 

Google Scholar 
Ishiwata, T., Kilgour, R. J., Uetake, K., Eguchi, Y. & Tanaka, T. Choice of attractive conditions by beef cattle in a Y-maze just after release from restraint. J. Anim. Sci. 85, 1080–1085. https://doi.org/10.2527/jas.2006-405 (2007).Article 
CAS 

Google Scholar 
Ede, T., von Keyserlingk, M. A. G. & Weary, D. M. Social approach and place aversion in relation to conspecific pain in dairy calves. PLoS One 15, e0232897. https://doi.org/10.1371/journal.pone.0232897 (2020).Article 
CAS 

Google Scholar 
Cantor, M. C., Renaud, D. L., Neave, H. W. & Costa, J. H. C. Feeding behavior and activity levels are associated with recovery status in dairy calves treated with antimicrobials for Bovine Respiratory Disease. Sci. Rep. 12, 4854. https://doi.org/10.1038/s41598-022-08131-1 (2022).Article 
ADS 
CAS 

Google Scholar 
Kappeler, P. M., Cremer, S. & Nunn, C. L. Sociality and health: Impacts of sociality on disease susceptibility and transmission in animal and human societies. Philos. Trans. R. Soc. Lond. B Biol. Sci. 370, 20140116. https://doi.org/10.1098/rstb.2014.0116 (2015).Article 

Google Scholar 
Ezenwa, V. O. et al. Host behaviour–parasite feedback: An essential link between animal behaviour and disease ecology. Proc. R. Soc. B Biol. Sci. 283, 20153078. https://doi.org/10.1098/rspb.2015.3078 (2016).Article 
CAS 

Google Scholar 
Klein, S. L. Parasite manipulation of the proximate mechanisms that mediate social behavior in vertebrates. Physiol. Behav. 79, 441–449. https://doi.org/10.1016/S0031-9384(03)00163-X (2003).Article 
ADS 
CAS 

Google Scholar 
LavistaFerres, J. M. et al. Social connectedness and movements among communities of giraffes vary by sex and age class. Anim. Behav. 180, 315–328. https://doi.org/10.1016/j.anbehav.2021.08.008 (2021).Article 

Google Scholar 
VanderWaal, K. L., Wang, H., McCowan, B., Fushing, H. & Isbell, L. A. Multilevel social organization and space use in reticulated giraffe (Giraffa camelopardalis). Behav. Ecol. 25, 17–26. https://doi.org/10.1093/beheco/art061 (2014).Article 

Google Scholar 
Sato, S. & Wood-Gush, D. G. Observations on creche behaviour in suckler calves. Behav. Process. 15, 333–343. https://doi.org/10.1016/0376-6357(87)90017-9 (1987).Article 
CAS 

Google Scholar 
Lecorps, B., Kappel, S., Weary, D. M. & von Keyserlingk, M. A. G. Social proximity in dairy calves is affected by differences in pessimism. PLoS One 14, e0223746. https://doi.org/10.1371/journal.pone.0223746 (2019).Article 
CAS 

Google Scholar 
Carslake, C., Occhiuto, F., Vázquez-Diosdado, J. A. & Kaler, J. Repeatability and predictability of calf feeding behaviors—Quantifying between- and within-individual variation for precision livestock farming. Front. Vet. Sci. 9, 827124 (2022).Article 

Google Scholar 
Occhiuto, F., Vázquez-Diosdado, J. A., Carslake, C. & Kaler, J. Personality and predictability in farmed calves using movement and space-use behaviours quantified by ultra-wideband sensors. R. Soc. Open Sci. 9, 212019. https://doi.org/10.1098/rsos.212019 (2022).Article 
ADS 

Google Scholar 
Carslake, C., Occhiuto, F., Vázquez-Diosdado, J. A. & Kaler, J. Indication of a personality trait in dairy calves and its link to weight gain through automatically collected feeding behaviours. Sci. Rep. 12, 19425. https://doi.org/10.1038/s41598-022-24076-x (2022).Article 
ADS 
CAS 

Google Scholar 
Planas-Sitjà, I., Deneubourg, J.-L. & Cronin, A. L. Variation in personality can substitute for social feedback in coordinated animal movements. Commun. Biol. 4, 469. https://doi.org/10.1038/s42003-021-01991-9 (2021).Article 

Google Scholar 
Stockmaier, S., Bolnick, D. I., Page, R. A. & Carter, G. G. Sickness effects on social interactions depend on the type of behaviour and relationship. J. Anim. Ecol. 89, 1387–1394. https://doi.org/10.1111/1365-2656.13193 (2020).Article 

Google Scholar 
Smith, L. A., Swain, D. L., Innocent, G. T., Nevison, I. & Hutchings, M. R. Considering appropriate replication in the design of animal social network studies. Sci. Rep. 9, 7208. https://doi.org/10.1038/s41598-019-43764-9 (2019).Article 
ADS 
CAS 

Google Scholar 
Shin, D. H., Kang, H. M. & Seo, S. Social relationships enhance the time spent eating and intake of a novel diet in pregnant Hanwoo (Bos taurus coreanae) heifers. PeerJ 5, e3329. https://doi.org/10.7717/peerj.3329 (2017).Article 
CAS 

Google Scholar  More

Study areaIn Colombia, the Amazon region represents 42.3% of the territory with an estimated area of 483,164 km2. In this area, 14% is dominated by agricultural lands, secondary vegetation and fragmented forests. Currently, 86% of the area corresponds to natural areas in a good state of conservation, where forests are the dominant coverage6. In the northwest area, the region borders the Andean Cordillera and Orinoquía to the north. The political-administrative division includes the departments Amazonas, Caquetá, Guainía, Guaviare, Putumayo and Vaupés, and part of the departments Cauca, Meta, Nariño and Vichada. The human population is estimated at ~ 1.4 million, with a density of 2.5 inhab/km2. Internal conflict and poverty make this region one of the most important population dynamics in the country in terms of displacement36. The geographical location of the study area and the spatial pattern of the loss of forests that occurred between 2002 and 2016 are shown in Fig. 1.Figure 1Study area. Colombian Amazon and location of Amazonian tropical forests that were lost between 2002 and 2016. (Maps were generated using software ArcGis 10.7.1 https://www.esri.com).Full size imageLand cover maps and variables for change analysisThematic land cover maps used in this research were produced by the Colombian Amazon Land Cover Monitoring System (SIMCOBA) of the Amazon Institute for Scientific Research SINCHI (https://siatac.co/simcoba/). SIMCOBA has prepared land cover maps for the periods 2002, 2007, 2012, 2014, 2016 and 2018. Three of the land cover maps prepared were used in this study: 2002, 2016 and 2018 a scale of 1:100,00033. The maps were generated from the visual interpretation of a mosaic of Landsat 5 Thematic Mapper (TM) and Landsat 8 Operational Land Imager (OLI) images, using the PIAO technique (Photo Interprétation Assistée par Ordinateur). The classification categories of the land cover maps were based on the Corine land cover methodology adapted for Colombia37.The SIMCOBA system calculates the annual rates of Amazon forest loss (forest loss/ha/annual) by comparing the cover maps of the last two periods and subtracting from the previous map those forests that are no longer present in the most current map (Fig. 3). This process only considers the forests loss and the permanent forests. New forests due to natural regeneration or restoration are omitted in the calculations6.To facilitate the interpretation of changes and cover transitions, the classification categories of the maps were re-categorized into 7 types: “Amazon forests”, “floodplain forests”, “fragmented forests and secondary vegetation”, “grasslands and shrublands”, “water bodies and wetlands”, “pastures and crops” and “urban and artificialized cover”. The land cover maps were resampled at a resolution of 60 m × 60 m to facilitate the computational analysis of the explanatory model, the simulations of the scenarios, and to keep the detailed spatial resolution of the coverage and explanatory variables16.A geospatial database was created with a set of variables for the cover changes to create an explanatory model for each transition. Driving factors of change are grouped into the following variables: (1) accessibility, (2) climate, (3) landscape features, (4) production practices and environmental degradation, (5) landscape management, (6) socioeconomy, and (7) soil characteristics. We considered 41 explanatory variables (see supplementary information Table S1).Accessibility variables such as roads and navigable rivers were obtained from the geodatabase at a scale of 1:100,000 of the Agustín Codazzi Geographical Institute of Colombia (IGAC). Bioclimatic temperature data were obtained from Worldclim v1.438. Cover variables (e.g., patch sizes Amazon forests and distance to pastures and crops) were created using the software ArcGis (v.10.7.1)39 from the 2002 land cover map to understand which drivers were more influential in the dynamics of land-use changes since 2002 that resulted in the distribution of land cover in 2016.Degradation variables, such as advances of the agricultural frontier, were obtained from the Territorial Environmental Information System of the Colombian Amazon (SIAT-AC)40; livestock density data came from the Colombian Agricultural Institute (ICA); the fire density were processed from MODIS and VIIRS images (https://siatac.co/puntos-de-calor/); and the location of mining titles was obtained from the National Mining Agency.The information on the landscape features and socioeconomic variables was obtained from different sources: (1) the limit of the protected natural areas was provided by the National System of Protected Areas (SINAP)41, (2) the Amazon Forest Reserve areas (Second Law of 1959) were obtained from the Ministry of Environment and Sustainable Development (MADS), (3) the location of the indigenous reservations was provided by the Ministry of the Interior, and (4) the limits of the areas of Indigenous Reservations and the legal status of the Amazonian region were obtained from the SINCHI cartographic database40.Socioeconomic information was spatialized from data from the National Administrative Department of Statistics (DANE). Soil-type data were obtained from IGAG, and topographic and altitudinal variables were derived from a DEM at 100 m resolution from the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER V003) sensor42. All explanatory variables were resampled at a resolution of 60 m.Patterns of land cover changes and transitionsThe transformation patterns of territory are mainly defined by human intentions and the activities that these groups plan to develop after making the land cover changes, as well as the dynamics of vegetation regeneration43. In this study, these changes in the study area were obtained and analyzed employing the Land Change Modeller (LCM) module of TerrSet34 and using the land cover maps for 2002 and 2016 as input information (Fig. 2).Figure 2(Source: Open Data—SINCHI Institute https://datos.siatac.co/pages/coberturas) (Maps were generated using software ArcGis 10.7.1 ).Land cover maps 2002, 2016 and 2018, produced by the Colombian Amazon Land Cover Monitoring System (SIMCOBA) of the Amazonian Research Institute SINCHIFull size imageTo represent dynamics and changes in the vegetation during the study period, a total of 14 transitions of greater importance in terms of area were considered (transitions with an area  More

Our objective in developing the Eco-ISEA3H database37 was to compile a coordinated, global set of tabular data, characterizing environmental conditions and the geographic distributions of large mammalian species. The database was built on the ISEA3H DGGS, a multi-resolution system of global grids, each grid dividing the Earth’s surface into discrete, equal-area hexagonal cells. These cells constitute areal units of observation, uniformly resampling data provided in different coordinate reference systems, spatial resolutions, geographic data models, and file formats. We included data at six consecutive ISEA3H resolutions, in which cell centroid spacing ranges from 29 kilometers to approximately 450 kilometers.Eco-ISEA3H themes and variables were derived from 17 geospatial data sources, and represent 3,033 features to be used for ML-based predictive modeling. Source datasets were published in raster or vector format, data models built on fundamentally different representations of spatial phenomena. Raster datasets comprise regular arrays of pixels, each pixel holding a value, while vector datasets comprise point, line, and polygon features, each feature defined by one or more (x, y) coordinate pairs and attributed with one or more values. Our task was to integrate these disparate source datasets, resampling and summarizing the values of raster pixels and vector features via the discrete, equal-area cells of the ISEA3H global grid system. The hexagonal cells on which the Eco-ISEA3H database37 is built thus serve as unifying observational units for SDM and ecometric analysis and modeling.From the statistical and ML perspective, each areal observational unit is characterized by (1) a set of environmental variables, representing climatic conditions, soil and near-surface lithology, land cover, and physical geography; and (2) a set of occurrence variables, representing the present and estimated natural distributions of large mammalian species. Predictive modeling tasks for statistical and ML modeling can be formulated in two directions: predicting species’ occurrences as a function of climatic and other environmental conditions (as in SDM studies), or predicting climatic and other environmental conditions as a function of species’ occurrences and functional traits (as in ecometric studies).Spatial units of observationTo study continuous spatial phenomena over a region of interest, it is often necessary to divide the region into a number of discrete, areal observational units, which may be used in statistical summaries and/or modeling. Machine learning methods for ecometric and species distribution modeling require discrete observational units, each characterized by two sets of variables, one describing environmental conditions, the other species’ geographic distributions. A major question in data representation concerns the form of these units; defining discrete spatial units of observation constitutes a well-known problem in geography, termed the modifiable areal unit problem (MAUP)38. As we change the size of proposed observational units, or change the boundaries between units while holding unit areas constant, measures of interest within these units – and derived summary statistics and model parameters – may differ; these are termed the “scale” and “zone” effects, respectively38.Our objective in utilizing the ISEA3H DGGS34 was to implement a robust spatial division of the Earth’s surface. The grid cells of the DGGS discretize the Earth’s sphere, forming, at each DGGS resolution, a global set of areal observational units with which to sample and summarize source datasets. To be optimally effective in the observation, simulation, and visualization of spatial phenomena, such a grid must meet certain structural criteria. We propose, modifying the Goodchild Criteria39, the DGGS grid must contain (1) contiguous, (2) equivalent observational units, (3) minimizing intra-unit variability, (4) having uniform topology with neighboring units, and (5) being visually effective, facilitating interpretation and communication. Each criterion will be discussed in detail; further, we will argue the ISEA3H DGGS selected for this study satisfies these criteria.Contiguity & congruencyWe suggest that a regular tiling maximally satisfies the criteria of (1) contiguity and (2) equivalence. A tiling is simply a set of shapes which cover a plane without gaps or overlaps40. A regular tiling is one of a class of tilings in which the tiles – our observational units – are highly equal; such tilings are monohedral, and composed of congruent, regular (equiangular and equilateral) polygons. Thus, regular tilings are also highly symmetrical, being vertex-, edge-, tile-, and flag-transitive. Three regular polygons may be used to create a regular tiling: the equilateral triangle, the square, and the regular hexagon40.With this suggestion, we follow common convention; in ecology, grids of square (or rectangular) cells are most often utilized, motivated in part by the use of raster datasets41, made of rectilinear rows and columns of pixels. However, it should be noted that while the square cells of these grids are equal in the coordinate reference system in which they are defined, such cells are rarely congruent, or indeed even square, on the Earth’s surface. The properties of the ISEA projection selected for this DGGS – area preservation, and relatively low angular distortion – serve to retain considerable congruency when inversely projecting grid cells to the spherical surface of the Earth.CompactnessTo accurately represent the spatially continuous phenomena of the Earth system, the grid cells of a DGGS – the areal observational units used in summarizing, modeling, and visualizing – must effectively discretize these phenomena. Thus, the DGGS must be structured such that (3) intra-unit variability is minimized, and inter-unit variability is maximized. In this way, patterns of variation among units more accurately represent patterns of variation inherent in the phenomena.Intra-unit variability may be minimized, in expectation, by compact observational units. Tobler’s oft-cited first law of geography serves as explanation: “everything is related to everything else, but near things are more related than distant things”42. Thus, compact units, in which all portions of the interior are nearer each other, are expected to contain less interior variability than elongated units, in which portions of the interior may be more distant. Given these properties, compact units are optimal in the context of DGGS development, elongated units in the context of efficient ecological sampling.Regular hexagons are the most compact of the three polygons – the equilateral triangle, square, and regular hexagon – admitting regular tilings. This compactness may be expressed in several related and complementary ways. First, of any equal-area tiling, regular hexagons have the minimum possible ratio of perimeter to area43. In minimizing perimeter length per unit area, regular hexagons are thus the most circle-like of the polygons admitting equal-area tilings. Relatedly, regular hexagonal packing is the highest-density arrangement of equal-area circles on a plane44.Finally, a regular hexagonal lattice optimally quantizes a plane; of the polygons admitting regular tilings, regular hexagons minimize the mean squared distance of any point to the nearest polygon centroid45. This distance, or “dimensionless second moment,” quantifies the more qualitative notion of interior nearness discussed in relation to Tobler’s Law.TopologyIn addition to maximally satisfying the (3) compactness criterion, regular hexagons have a topological advantage over equilateral triangles and squares. Of these three regular polygons, hexagons have the simplest relationship with neighbors in a tiling or grid, each (4) uniformly sharing an edge with the six adjacent hexagons forming its first-order neighborhood. Triangles and squares, in contrast, share only a single vertex with three or four neighbors, respectively, and an edge with three or four neighbors, complicating the definition of neighborhood in these grids.It follows that hexagonal topology has greater angular resolution than edge-based triangular or square topologies; movement may be simulated between cells in six directions, rather than in three or four, respectively. These properties – neighborhood simplicity and angular resolution – were confirmed by Golay46, in the context of pattern transformation operations on two-dimensional arrays. Further, these properties likely account for the widespread use of hexagonal grids in strategy board games, since these grids were introduced in the early 1960s47.Differing grid topologies affect the results of ecological models simulating dispersal. White and Kiester48, for example, found the topology of the network of communities in a neutral community ecology model – in which simulated communities had hexagonal neighborhoods, or von Neumann, Moore, or Margolus neighborhoods – affected modeled species abundances and diversities, but in complex ways, which differed given different model parameter values. (Note that the four neighbors with which a square cell shares an edge are termed its rook, or von Neumann neighborhood, and these plus the four neighbors with which it shares a single vertex its queen, or Moore neighborhood.)VisualizationFinally, in addition to these gains in representational accuracy, (5) hexagonal tilings are more visually effective than square tilings. Whether used in cartography or other two-dimensional data visualization, tilings inevitably create visual lines, artifacts of the lattice of shared edges between tiles49. Given our “sense of gravitational balance,” Carr et al.49 argue the horizontal and vertical lines of square tilings strongly distract the human eye, obscuring data-driven patterns in a dataset so visualized. The non-orthogonal lines of hexagonal tilings, however, feature less prominently, and thus distract less from patterns of interest49.Note that this is not an issue of aesthetics only: maps are often essential tools in scientific reasoning and communication, and effective visualization is important. Indeed, Carr et al.49 suggest this visual advantage makes a stronger case for hexagonal tilings than the representational advantages discussed previously.DGGS sampling workflowsThe set of scripted workflows developed to incorporate spatial datasets into the Eco-ISEA3H database37 utilize published spatial libraries and packages for Python and R, and include several validation steps, intended to verify the integrity of source datasets and the fidelity of the transfer to the DGGS. Workflows developed for raster datasets are presented in Fig. 1, and workflows for vector datasets in Fig. 2.Fig. 1Workflow developed to incorporate raster datasets into the ISEA3H DGGS.Full size imageFig. 2Workflow developed to incorporate vector datasets into the ISEA3H DGGS.Full size imageTo begin, one general principle guides each workflow: each source dataset is processed in its native coordinate reference system. In all cases, a representation of the DGGS is developed in the coordinate reference system of the source dataset, and used in summarizing that dataset. The guiding premise here is that the spatial dataset is as the authors intended it in the coordinate reference system in which it is published and distributed.This is especially relevant for vector polygon datasets. Consider, for example, certain species’ range polygons published by the IUCN Red List50; these polygons are defined only roughly, having relatively few, widely spaced vertices, connected by arcs many hundreds of kilometers in length. These arcs are “straight” in the plate carrée projection with which the dataset’s WGS84 latitude/longitude coordinates are visualized by default. If vertex coordinates were projected into another coordinate reference system, the arcs would be similarly “straight” in this new system, and thus potentially trace very different paths across the Earth’s surface. Absent information to the contrary, we assume the arcs are as intended in the reference system in which the data are distributed.The spatial structure of raster datasets depends similarly on each dataset’s coordinate reference system; rasters are made of rows and columns of pixels, rectilinear and orthogonal only in the raster’s native coordinate reference system. We assume raster pixels are “atomic” units, each indivisible and representative of the area it natively covers. Thus, we query the DGGS at each pixel’s centroid, and assign the pixel wholly to the coincident DGGS cell.Raster dataset processingIf necessary, source raster datasets were first converted to the GeoTIFF file format, so that the files were readable in the open-source GIS software used later in the processing workflow. GeoTIFF files are simply Tag Image File Format (TIFF) image files with embedded georeferencing information, describing the dataset’s spatial extent and coordinate reference system. Hierarchical Data Format Release 4 (HDF4) files were converted to GeoTIFF format using the Geospatial Data Abstraction Library (GDAL) translate utility51.Next, raster tiles containing ISEA3H hexagon identification (HID) indexing numbers were generated; these integer HIDs uniquely identify each cell at a given ISEA3H resolution. A set of HID raster tiles was required for each source raster dataset, for each ISEA3H resolution, because (1) GeoTIFF rasters are able to hold only a single value at each pixel; and (2) HIDs sequentially number cells at a given ISEA3H resolution, from 1 to the number of cells present at that resolution. Thus, HIDs are not unique between resolutions; HID 84, for example, identifies a cell at each ISEA3H resolution 2 and higher.The HID raster tiles generated for a source raster dataset matched that dataset’s grid resolution, extent, and coordinate reference system precisely; thus, there was a one-to-one correlation between the pixels of the HID raster tiles and the source raster dataset tiles. For each tile, pixel centroid coordinates were passed to the dggridR package52 for R, which returned the ISEA3H cell identification number for that location. In this way, the pixels of the source raster were treated as indivisible units, assigned wholly to a particular HID on the basis of each pixel’s centroid. HID rasters were written in GeoTIFF format using the raster package53 for R.In equal-area projected coordinate reference systems, simple counts of the number of raster pixels assigned to each HID were sufficient to determine each ISEA3H cell’s total area. In all other cases – for example, for raster datasets using the World Geodetic System 1984 (WGS84) coordinate reference system – raster tiles containing pixel areas were generated. These areas were calculated by passing each pixel’s corner coordinates to the GeographicLib library54 for Python.Finally, source raster dataset tiles, HID raster tiles, and area raster tiles (for source rasters using non-authalic coordinate reference systems) were superimposed to generate summary tabular files, describing the features of the source raster dataset by ISEA3H cell. The specifics of this process, which utilized functions of the raster package53 for R, depended on whether the source raster contained discrete, categorical values, or continuous, real-numbered values.Discrete themesFor each source raster dataset containing discrete pixel values, one or more of the following summary statistics were calculated. While the centroid attribute requires a simple point sample, the fraction and mode attributes are area-integrated, and involve a multiple-step sampling process. For rasters using an authalic coordinate reference system, the raster package’s crosstab function53 was used to generate a contingency table for each tile; applied to source raster and HID raster tiles, the function tallied the number of pixels of each class coincident with each HID, for each tile. These tile-specific tables were then summed, to obtain total counts of pixels of each class within each HID.For rasters using a non-authalic coordinate reference system, area raster tiles were required as well. For each tile, a vector of classes present in the source raster was assembled. For each of these classes in turn, a mask raster tile was generated, retaining pixels belonging to the class, and screening pixels belonging to all other classes. This mask was applied to the area raster tile, and retained pixels were summed within each HID using the raster package’s zonal function53. Thus, a contingency table was compiled for each raster tile, containing the area of each class within each HID. Finally, these tile-specific tables were summed, to obtain the total area of each class within each HID.

Centroid. The centroid attribute records the categorical value occurring at each ISEA3H cell’s centroid. Where the source raster dataset contains a null value at a centroid, the cell is assigned a flag signifying no value is available.

Fraction. The fraction attributes record the proportion of each ISEA3H cell’s area covered by each categorical value. For example, the Köppen-Geiger climate classification system, as implemented by Beck et al.55, includes 30 classes, listed in Table 4. Thus, each ISEA3H cell has an associated set of 30 fraction attributes for this dataset, recording the proportions of the cell’s area covered by the 30 categorical values, from tropical rainforest (Af) to polar tundra (ET).

Mode. The mode attribute records the categorical value covering the greatest proportion of each ISEA3H cell’s area. For example, if an ISEA3H cell had a fraction value of 0.4 for some hypothetical categorical value A, 0.3 for B, and 0.3 for C, it would be assigned a mode value of A. A mode attribute is specified for cells in which the sum of the fraction attributes is greater than or equal to 0.2; where fraction attributes total less than 0.2, a flag signifying no value is assigned.

Continuous variablesFor each source raster dataset containing continuous pixel values, one or more of the following summary statistics were calculated. Again, the centroid attribute requires only a simple point sample, while the mean attribute is area-integrated, requiring area raster tiles for source rasters using a non-authalic coordinate reference system.

Centroid. The centroid attribute records the continuous value occurring at each ISEA3H cell’s centroid. Where the source raster dataset contains a null value at a centroid, the cell is assigned a flag signifying no value is available.

Mean. The mean attribute records the area-weighted arithmetic mean of the continuous values of raster pixels within each ISEA3H cell. For raster datasets in authalic coordinate reference systems, the area-weighted mean is equivalent to the simple mean of the values of raster pixels within each cell; however, in all other cases, pixel values are weighted by pixel areas per the equation below, in which wi and xi indicate the area and value, respectively, of each pixel i within an ISEA3H cell containing n pixels.

$$overline{x}=frac{{sum }_{i=1}^{n}{w}_{i}{x}_{i}}{{sum }_{i=1}^{n}{w}_{i}}$$For each tile, source raster values and area values were multiplied, pixel by pixel, using the raster package’s * arithmetic operator53. The resulting product raster tile, as well as the area raster tile, were then summed within each HID using the raster package’s zonal function53. Finally, these tile-specific tables were summed, to obtain both the numerator (summed product values) and denominator (summed area values) for the above equation, for each HID.Vector dataset processingSource vector datasets incorporated into the Eco-ISEA3H database37 contain polygon features, discrete areas assigned a categorical value. A dataset may (1) contain polygons of several different classes; for example, the vector shapefile published by Olson et al.56 contains ecoregion polygons, each assigned to one of several biogeographic realms. Alternatively, a dataset may (2) represent a single class, with polygons indicating class presence; for example, the shapefiles published by the IUCN Red List50 each represent a species’ geographic range, with polygons indicating regions the species is present. In both cases, the summary statistics discussed in reference to raster datasets containing discrete values may be calculated.Prior to inclusion in the Eco-ISEA3H database37, source vector datasets were preprocessed. To simplify the geographic representation of the class(es) of interest – that is, to remove unnecessary polygon boundaries – dataset polygons were dissolved, either on the class attribute in case (1), or globally in case (2), using the QGIS open-source desktop GIS application. The geodesic areas of dissolved polygons were then calculated using the GeographicLib library54. Finally, the geometries of dissolved polygons were checked for conformance with the OGC Simple Feature Access standard57 using the Shapely library58 for Python, ensuring these features served as valid input in the processing workflow to follow.The intersection of source dataset polygons and ISEA3H cell polygons is central to the vector processing workflow. Source polygons result from the preliminary simplification and verification steps just discussed; cell polygons result from polygonizing a set of HID raster tiles for the ISEA3H resolution of interest. The polygonizing procedure utilized the open-source GDAL command-line tools polygonize and ogrmerge51, as well as the GeographicLib54 and Shapely58 libraries. Polygonizing HID raster tiles of the appropriate coordinate reference system (specifically, the system matching that of the source polygon dataset) ensured HID polygon boundaries displayed both proper geodesic curvature and the shape distortion induced by the ISEA map projection.Intersection is a set-theoretic operation, returning polygons representing each coincident class/HID combination. The operation was implemented via the Shapely library58, and the geodesic areas of intersected polygons were calculated via the GeographicLib library54. Note that the scripted intersection tools developed for the Eco-ISEA3H database37 allow limiting the ISEA3H cells included in a single tool run, to break the processing of large datasets into manageable pieces. Runs may be limited to a user-specified range of HIDs. Additionally, if cells at the next coarser or finer ISEA3H resolution have been intersected with the source dataset, cells retained by the operation may be used as a spatial index; a list of coincident HIDs at the ISEA3H resolution of interest may be generated, and used to limit tool runs.An output shapefile is written, containing intersected polygons attributed with the geodesic area, the HID, and in case (1), the source class. Next, an additional verification of the geometries of these intersected polygons is performed. Each intersected polygon is superimposed over the original ISEA3H cell polygon having the same HID. If intersected polygons have too few vertices to be valid, or are not contained by the original cell polygon from which each was derived, these polygons are flagged for review and revision. This step was implemented to catch geometry errors observed early in the development of the Eco-ISEA3H intersection tools.Finally, the geodesic areas of intersected polygons are totaled, and the total area of each class within each HID is calculated. Dividing by the geodesic areas of the original ISEA3H cell polygons, these class totals are expressed as fractions of each cell’s total area. In two final verification steps, (1) the total intersected area of each class, across all HIDs, is compared to the area of the same class in the source dataset; and (2) class fraction values are confirmed to be less than or equal to unity within each HID. Deviations are flagged for review and revision.Data sources & themesThe Eco-ISEA3H database37 incorporates 17 source datasets, characterizing the Earth’s climate, geology, land cover, and physical geography, as well as human population density and the geographic ranges of nearly 900 large mammalian species. Data sources are listed in Table 1. We first present a brief overview of these sources, and describe sources and themes in greater detail in the following sections.Table 1 Source datasets and themes included in the Eco-ISEA3H database37. Each dataset is described by full and abbreviated name, source, spatial resolution (for datasets published/distributed at more than one resolution), version, and scenario. Each theme is described by full and abbreviated name and type (whether it contains discrete, categorical values or continuous, real-valued variables).Full size tableClimate is characterized primarily by temperature- and precipitation-based averages and extremes, summarized over the past 50 to 70 years, and forecasted for 40 to 60 years in the future under the RCP 8.5 climate change scenario59; data sources include WorldClim30,31, ENVIREM60, and the ETCCDI extremes indices derived by Sillmann et al.61,62 from ERA-4063 and CCSM464. Additionally, present climate is classified via the Köppen-Geiger climate classification system, from GLOH2O55. Geological data include soil types, from the Digital Soil Map of the World (DSMW)65; near-surface rock types, from the Global Lithological Map (GLiM)66; and sedimentary basin types67. Human geography is quantified by human population density, from the Gridded Population of the World (GPW)68. Land cover is described by the International Geosphere-Biosphere Programme (IGBP) cover classification scheme, from MCD12Q169; and by percent tree, non-tree, and non-vegetated cover, from MOD44B70. The Earth’s physical geography is characterized by continental and island landmasses, from Natural Earth; lakes and wetlands, from the Global Lakes and Wetlands Database (GLWD)71; biogeographic realms56; and terrestrial topography and ocean bathymetry, from ENVIREM60 and SRTM30_PLUS72. Finally, distributional data include the present and estimated natural ranges of large mammalian species, from the IUCN Red List50 and the Phylogenetic Atlas of Mammal Macroecology (PHYLACINE)73,74.Climate

ENVIREM. The ENVIREM (ENVIronmental Rasters for Ecological Modeling) dataset60 contains 16 climatic variables derived from WorldClim v1.4 monthly temperature and precipitation30, and extraterrestrial radiation. These are intended to compliment the WorldClim v1.4 bioclimatic variables30, capturing additional environmental features directly relevant to floral and faunal physiology and ecology60. Source rasters at 30 arc-second resolution were summarized by area-weighted mean at ISEA3H resolutions 8 and 9. Variable codes, descriptions, and units are listed in Table 2. Title and Bemmels60, and references therein, provide full definitions and calculation methods for these variables.Table 2 Codes, descriptions, and units for the 16 ENVIREM climatic variables, from Title and Bemmels60.Full size table

ETCCDI. A comprehensive set of 27 climate extremes indices was defined by the Expert Team on Climate Change Detection and Indices (ETCCDI); these generally capture “moderate” extremes, having recurrence intervals of a year or shorter, and are based on observed/simulated daily temperature and precipitation61,62. Sillmann et al.61,62 derive these indices from results of a number of global climate models and atmospheric reanalyses, several of which were incorporated in the Eco-ISEA3H database37. Given the relatively low-resolution grids used in modeling and reanalysis, these source rasters were interpolated to ISEA3H cell centroids by inverse (geodesic) distance weighting (IDW). Variable codes, descriptions, and units are listed in Table 3. Sillmann et al.61 provide full definitions and calculation methods for these indices.Table 3 Codes, descriptions, and units for the 27 ETCCDI climate extremes indices, from Sillmann et al.61,62.Full size table

The Eco-ISEA3H database37 includes ETCCDI variables based on results of the ERA-40 reanalysis63, produced by the European Centre for Medium-Range Weather Forecasts (ECMWF). The reanalysis combines past meteorological observations with a weather forecasting model, producing a global representation of the state of the atmosphere for each reanalysis time step, usually a six-hour interval63. These were averaged for the period 1958 to 2001, the 44 full years for which the ERA-40 reanalysis was conducted, and were interpolated to ISEA3H resolutions 5 to 9.Additionally, the database includes ETCCDI variables based on results of the Community Climate System Model v4 (CCSM4), a global climate model developed for CMIP564. These were averaged for the period 1950 to 2000, to match the approximate period covered by WorldClim v1.4, and for the period 2061 to 2080, to match the final interval for which CCSM4 model results were downscaled/debiased using WorldClim v1.430. Variables were interpolated to ISEA3H resolution 9.ETCCDI variables for this latter period represent conditions under Representative Concentration Pathway (RCP) 8.5, the RCP resulting in the highest radiative forcing (8.5 W/m2) by 210059. This scenario was selected such that future conditions maximally different from the present might be considered; in RCP 8.5, rapid population growth, and relatively slow growth in per capita income and technological development, lead to high energy demand without associated climate mitigation policies, resulting in greenhouse gas emissions and atmospheric concentrations increasing significantly in the coming decades59.

Köppen-Geiger Climate Classification. As implemented by Beck et al.55, the Köppen-Geiger system classifies the Earth’s terrestrial climates into five primary classes, and further into 30 subclasses, based on a set of threshold criteria referencing monthly mean temperature and precipitation. These climate classes are ecologically significant, as regions within each class support floral communities sharing common characteristics. Beck et al.55 utilize four climatic datasets, including WorldClim v1.x and v2.x, adjusted to the period 1980 to 2016, to define the present-day classes incorporated in the Eco-ISEA3H database37. The source raster at 30 arc-second resolution was summarized by fraction and mode at ISEA3H resolution 9. Variable codes and descriptions are listed in Table 4.Table 4 Codes and descriptions for the 30 Köppen-Geiger climate classes, from Beck et al.55.Full size table

WorldClim v1.4. The first-generation WorldClim dataset30 contains four monthly themes, each with 12 variables, characterizing monthly temperature and precipitation; additionally, it contains 19 bioclimatic variables, derived from the monthly variables, capturing biologically relevant seasonal and annual features of the climate system. These bioclimatic variables, first developed for the BIOCLIM species distribution modeling (SDM) package75, are used extensively in SDM studies; a recent synthesis found most were included in more than 1,000 published MaxEnt SDMs (of 2,040 reviewed)76.

WorldClim monthly temperature and precipitation rasters are interpolated from weather station observations averaged for the approximate period 1950 to 2000. The interpolation was done using thin plate smoothing splines, with latitude, longitude, and elevation as predictor variables30. These rasters characterize present-day climate, and further served as an observational baseline with which the predictions of CMIP5 global climate models were downscaled and bias-corrected.The 19 bioclimatic variables, for both present-day and future conditions (the latter averaged for the period 2061 to 2080, from the CCSM4 RCP 8.5 simulation), were incorporated into the Eco-ISEA3H database37; source rasters at 30 arc-second resolution were summarized by area-weighted mean at ISEA3H resolution 9. Variable codes, descriptions, and units are listed in Table 5. O’Donnell and Ignizio77 provide full definitions and calculation methods for these variables.Table 5 Codes, descriptions, and units for the 19 WorldClim bioclimatic variables, from v1.430 and v2.031.Full size table

WorldClim v2.0. The second-generation WorldClim dataset31 contains seven monthly themes, each with 12 variables, characterizing monthly temperature, precipitation, solar radiation, wind speed, and vapor pressure; additionally, it contains the standard set of 19 bioclimatic variables, derived from monthly temperature and precipitation.

As in the first-generation dataset, monthly rasters were interpolated from weather station observations, averaged here for the approximate period 1970 to 200031. Again, thin plate smoothing splines were used in the interpolation, but with additional covariates included for one or more interpolated features: distance to coast, computed extraterrestrial radiation, and three satellite-derived observations – cloud cover, and maximum and minimum land surface temperature, from the Moderate Resolution Imaging Spectroradiometer (MODIS) instrument.The 12 source rasters for each of the seven monthly themes, at 30 arc-second resolution, were summarized by centroid at ISEA3H resolutions 5 to 10. Additionally, the 19 source bioclimatic rasters, at 30 arc-second resolution, were summarized by centroid at ISEA3H resolutions 5 to 10, and by area-weighted mean at ISEA3H resolutions 6 to 9. Codes, descriptions, and units for the bioclimatic variables are listed in Table 5.Geol10ogy

DSMW. The Digital Soil Map of the World (DSMW)65 describes the geographic distribution and physical and chemical properties of the world’s soils. The DSMW was digitized from the FAO-UNESCO Soil Map of the World, printed at 1:5,000,000 scale. Each digitized mapping unit is assigned a number of soil attributes; here we classify units via the DOMSOI attribute, the dominant soil or land unit code. The DSMW includes 117 soils in 26 major soil groupings, as well as six other land units, for a total of 123 DOMSOI classes. The source vector dataset was summarized by fraction and mode at ISEA3H resolutions 5, 6, and 9. Variable codes and descriptions are listed in Table 6.Table 6 Codes and descriptions for the 123 DSMW soil and land units, from the FAO65.Full size table

GLiM. The Global Lithological Map (GLiM)66 represents the rock and unconsolidated sediments at or near the Earth’s terrestrial surface; this geological material is a source of geochemical flux to the Earth’s soils, biosphere, and hydrosphere. Hartmann and Moosdorf66 compiled the map and accompanying database from 92 regional geological maps and 318 literature sources. Rock was classified into 16 first-level lithological classes; 12 second-level and 14 third-level subclasses further describe specific mineralogical and physical properties.

The source vector dataset was summarized by centroid at ISEA3H resolution 9. Variable codes and descriptions are listed in Table 7. The attribute assigned each ISEA3H cell takes the form xxyyzz; underscore characters (_) in the yy and/or zz slots indicate the second- and/or third-level subclasses were undefined.Table 7 Codes and descriptions for the 16 first-level, 12 second-level, and 14 third-level GLiM lithological classes, from Hartmann and Moosdorf66.Full size table

Sedimentary Basins. Sedimentary basins are areas of subsidence in the Earth’s crust, in which sediments eroded from uplands are deposited and potentially preserved for a million or more years67, thus entering the planet’s long-term geological record. Nyberg and Howell67 delineate active sedimentary basins, covering both the Earth’s terrestrial surface and marine areas over continental crust. The authors operationally defined basins as low-relief areas containing Quaternary Period sediments, and further classified the basins by tectonic setting, identifying backarc, forearc, foreland, extensional, intracratonic, passive margin, and strike-slip basins on the basis of published literature and geological maps67.

Terrestrial basins were incorporated in the Eco-ISEA3H database37. Note that no terrestrial backarc basins were delineated. The source vector dataset was summarized by fraction and mode at ISEA3H resolution 9.Human geography

GPW. Human population density is one of several measures of human presence and activity which together define the human “footprint,” associated with profound, adverse effects on natural systems78. Given this pervasive impact, data characterizing degree of human influence are used as predictors in some ecological models, including SDMs28. The Gridded Population of the World (GPW)68 density dataset represents the global distribution of human population density, developed using census records, population registers, and the administrative boundaries of approximately 13.5 million national and subnational units. Density, measured by population count per square kilometer, was estimated every five years, from 2000 to 2020, inclusive. The source raster dataset for each year, at 30 arc-second resolution, was summarized by area-weighted mean at ISEA3H resolutions 6 to 9.

Land cover

MCD12Q1. The Moderate Resolution Imaging Spectroradiometer (MODIS) land cover type (MCD12Q1) dataset69 describes land cover globally, via six different classification schemes. The Eco-ISEA3H database37 includes land cover classified via the International Geosphere-Biosphere Programme (IGBP) scheme, initially developed for the DISCover dataset79; the IGBP scheme includes 16 land cover classes, 13 natural and three anthropogenically modified. The MCD12Q1 dataset is derived from reflectance data collected by the MODIS instruments aboard the Terra and Aqua satellites; the two instruments observe the entirety of the Earth’s surface every one to two days, recording reflectance in 36 spectral bands.

MCD12Q1 land cover is estimated annually. For each year, reflectance time-series data are smoothed and gap-filled via smoothing splines; derived spectro-temporal features are used as input to a random forest classifier; and output land cover classifications are post-processed, to incorporate prior knowledge and reduce inter-annual variability69. The source raster dataset for 2001 and 2014 to 2018, inclusive, at approximately 500 meter resolution, was summarized by centroid, fraction, and mode at ISEA3H resolutions 5 to 10. Variable codes and descriptions are listed in Table 8.Table 8 Codes and descriptions for the 16 IGBP land cover classes, from Friedl and Sulla-Menashe69.Full size tableMOD44B. The MODIS vegetation continuous fields (VCF) dataset (MOD44B)70 describes global land cover quantitatively, as fractions of three cover components: tree canopy, non-tree canopy, and non-vegetated, barren cover. Note that canopy cover, as defined here, indicates the area over which light is intercepted; this differs from crown cover, which indicates the area covered by a plant’s crown regardless of light interception/penetration. The MOD44B dataset is derived from reflectance data collected by the MODIS instrument aboard the Terra satellite; for each annual VCF estimate, reflectance time-series data are used as input to a bagged ensemble of linear regression trees70. The source raster dataset for 2018, at approximately 250 meter resolution, was summarized by area-weighted mean at ISEA3H resolution 9.Physical geography

Biogeographic Realms. As defined by Olson et al.56, the eight terrestrial biogeographic realms are the broadest divisions of the Earth’s terrestrial flora and fauna; these may be further subdivided into biomes and ecoregions, the latter containing distinct natural communities. Olson et al.56 developed this hierarchical system primarily for global and regional conservation planning. Realm, biome, and ecoregion delineations are based on expert knowledge, contributed by more than 1,000 scientists working in relevant fields; these divisions thus incorporate knowledge of endemic taxa, unique species assemblages, and local geological and biogeographical history56. Realms were included in the Eco-ISEA3H database37 to provide a high-level classification of the Earth’s biogeography, from a source frequently cited in the scientific literature. The source vector dataset was summarized by fraction and mode at ISEA3H resolutions 5 to 9. Variable codes and descriptions are listed in Table 9.Table 9 Codes and descriptions for the eight biogeographic realms, from Olson et al.56.Full size table

ENVIREM. In addition to the climatic variables discussed previously, the ENVIREM dataset60 contains two topographic variables, derived from SRTM30_PLUS. These two indices characterize terrain roughness, a measure of variability in local elevation; and topographic wetness, a function of slope and upgradient contributing area. Source rasters at 30 arc-second resolution were summarized by area-weighted mean at ISEA3H resolutions 8 and 9. Variable codes, descriptions, and units are listed in Table 10.Table 10 Codes, descriptions, and units for the two ENVIREM topographic variables, from Title and Bemmels60.Full size table

GLWD. The Global Lakes and Wetlands Database (GLWD)71, Level 3, represents the maximum extent of lakes, reservoirs, rivers, and a number of wetland types, comprising 12 waterbody classes in total. Lehner and Döll71 compiled the three levels of the GLWD by combining seven source map and attribute datasets, and suggest Level 3 may be useful as input in global hydrologic and climatic modeling. The source raster dataset at 30 arc-second resolution was summarized by fraction and mode at ISEA3H resolution 9. Variable codes and descriptions are listed in Table 11.Table 11 Codes and descriptions for the 12 GLWD waterbody classes, from Lehner and Döll71.Full size table

Natural Earth. Natural Earth is a public-domain collection of raster and vector datasets developed for production cartography. Three vector themes describing physical geography were incorporated: Land, which includes continents and major islands; Islands, which includes additional minor islands; and Lakes, which includes lakes and reservoirs. Source vector datasets at 1:10,000,000 scale were summarized by fraction at ISEA3H resolutions 5 to 9. Further, fractions for a Terra theme were calculated, by adding per-cell Land and Islands, and subtracting Lakes. The Terra theme may be thresholded (for example, at a fraction value ≥0.5) to identify terrestrial ISEA3H cells, excluding cells covered primarily by ocean or freshwater habitat.

SRTM30_PLUS. The SRTM30_PLUS dataset72 is a global digital elevation model (DEM), representing the Earth’s terrestrial topography and ocean bathymetry. A number of elevation sources were incorporated in developing the DEM; terrestrial topography was derived from the Shuttle Radar Topography Mission (SRTM) at latitudes between ±60°, from GTOPO30 in the Arctic, and from GLAS/ICESat in the Antarctic. Ocean bathymetry was derived from satellite radar altimetry, calibrated on 298 million corrected ship-based depth soundings, gathered from several sounding sources72. The source raster dataset at 30 arc-second resolution was summarized by area-weighted mean at ISEA3H resolutions 6 to 10.

Species rangesFrom the Red List and the Phylogenetic Atlas, the geographic ranges of species belonging to four mammalian orders were sampled: Artiodactyla (even-toed ungulates), Perissodactyla (odd-toed ungulates), Primates, and Proboscidea (elephants). These species are primarily large-bodied herbivores, and as such are frequently the subject of dental ecometrics research; for example, averaged dental traits of communities of these mammals have been used to predict measures of local precipitation, at both global3 and regional11 scales.

IUCN Red List. The International Union for Conservation of Nature’s (IUCN) Red List of Threatened Species50 comprises global assessments of the conservation status of nearly 150,000 floral, faunal, and fungal species. The Red List includes expert-delineated geographic ranges for most of these species, including most extant mammalian species. For each species, portions of the range for which the species’ presence was coded extant, and for which its origin was coded native or reintroduced, were sampled. Source vector datasets were summarized by fraction at ISEA3H resolutions 8 to 9 (Artiodactyla and Perissodactyla), 9 (Primates), and 7 to 9 (Proboscidea).

PHYLACINE. The Phylogenetic Atlas of Mammal Macroecology (PHYLACINE)73,74 includes trait, phylogeny, and geographic range data for all mammalian species known from the last interglacial period (approximately 130,000 years ago) to the present, both extant and recently extinct. PHYLACINE includes species’ ranges under two scenarios, both of which were incorporated: present-day ranges, from the IUCN v2016.3; and “present-natural” ranges, for which each species’ present-day range was modified to estimate its distribution under current climatic conditions, but absent anthropogenic pressure. This included, among eight modification categories, reconnecting fragmented ranges, by filling suitable intervening habitat; and expanding ranges reduced by human activity, by filling suitable adjacent habitat. Present-natural range modifications are documented for each species in PHYLACINE’s metadata, and intended to mitigate human impact on the results of macroecological analysis and modeling. Source rasters at approximately 100 kilometer resolution were summarized by centroid at ISEA3H resolution 9. More

Kennedy, E. V. et al. Avoiding coral reef functional collapse requires local and global action. Curr. Biol. 23, 912–918. https://doi.org/10.1016/j.cub.2013.04.020 (2013).Article 
CAS 

Google Scholar 
Beck, M. W. et al. The global flood protection savings provided by coral reefs. Nat. Commun. 9, 2186. https://doi.org/10.1038/s41467-018-04568-z (2018).Article 
ADS 
CAS 

Google Scholar 
Kuffner, I. B. & Toth, L. T. A geological perspective on the degradation and conservation of western Atlantic coral reefs. Conserv. Biol. 30, 706–715. https://doi.org/10.1111/cobi.12725 (2016).Article 

Google Scholar 
Allemand, D. et al. Biomineralisation in reef-building corals: From molecular mechanisms to environmental control. C. R. Palevol. 3, 453–467. https://doi.org/10.1016/j.crpv.2004.07.011 (2004).Article 

Google Scholar 
Glynn, P. W. Bioerosion and coral-reef growth: A dynamic balance. In Life and Death of Coral Reefs (ed Birkeland, C.) 68–95 (Chapman & Hall, 1997).Eyre, B. D., Andersson, A. J. & Cyronak, T. Benthic coral reef calcium carbonate dissolution in an acidifying ocean. Nat. Clim. Change 4, 969–976. https://doi.org/10.1038/nclimate2380 (2014).Article 
ADS 
CAS 

Google Scholar 
Enochs, I. C. et al. Upwelling and the persistence of coral-reef frameworks in the eastern tropical Pacific. Ecol. Monogr. 91, e01482. https://doi.org/10.1002/ecm.1482 (2021).Article 
CAS 

Google Scholar 
Alvarado, J. J., Grassian, B., Cantera-Kintz, J. R., Carballo, J. L. & Londoño-Cruz, E. Coral reef bioerosion in the eastern tropical Pacific. In Coral Reefs of the Eastern Tropical Pacific (eds Glynn, P. W., Manzello, D. P., Enochs, I. C.) 369–403 (Springer, 2017).Perry, C. T. et al. Caribbean-wide decline in carbonate production threatens coral reef growth. Nat. Commun. 4, 1402. https://doi.org/10.1038/ncomms2409 (2013).Article 
ADS 
CAS 

Google Scholar 
Oppenheimer, M. et al. Sea level rise and implications for low-lying islands, coasts and communities. In IPCC Special Report on the Ocean and Cryosphere in a Changing Climate (2019).Alvarez-Filip, L., González-Barrios, F. J., Pérez-Cervantes, E., Molina-Hernández, A. & Estrada-Saldívar, N. Stony coral tissue loss disease decimated Caribbean coral populations and reshaped reef functionality. Commun. Biol. 5, 440. https://doi.org/10.1038/s42003-022-03398-6 (2022).Article 

Google Scholar 
Perry, C. T. et al. Loss of coral reef growth capacity to track future increases in sea level. Nature 558, 396–400. https://doi.org/10.1038/s41586-018-0194-z (2018).Article 
ADS 
CAS 

Google Scholar 
van Woesik, R. & Cacciapaglia, C. W. Carbonate production of Micronesian reefs suppressed by thermal anomalies and Acanthaster as sea-level rises. PLoS ONE 14, e0224887. https://doi.org/10.1371/journal.pone.0224887 (2019).Article 
CAS 

Google Scholar 
van Woesik, R. & Cacciapaglia, C. W. Thermal stress jeopardizes carbonate production of coral reefs across the western and central Pacific Ocean. PLoS ONE 16, e0249008. https://doi.org/10.1371/journal.pone.0249008 (2021).Article 
CAS 

Google Scholar 
van Woesik, R. & Cacciapaglia, C. W. Keeping up with sea-level rise: Carbonate production rates in Palau and Yap, western Pacific Ocean. PLoS ONE 13, e0197077. https://doi.org/10.1371/journal.pone.0197077 (2018).Article 
CAS 

Google Scholar 
Eakin, C. M. Where have all the carbonates gone? A model comparison of calcium carbonate budgets before and after the 1982–1983 El Niño at Uva Island in the eastern Pacific. Coral Reefs 15, 109–119. https://doi.org/10.1007/BF01771900 (1996).Article 
ADS 

Google Scholar 
Perry, C. T. & Morgan, K. M. Bleaching drives collapse in reef carbonate budgets and reef growth potential on southern Maldives reefs. Sci. Rep. 7, 40581. https://doi.org/10.1038/srep40581 (2017).Article 
ADS 
CAS 

Google Scholar 
Connell, J. H. Disturbance and recovery of coral assemblages. Coral Reefs 16, S101–S113. https://doi.org/10.1007/s003380050246 (1997).Article 

Google Scholar 
Hughes, T. P. et al. Spatial and temporal patterns of mass bleaching of corals in the Anthropocene. Science 359, 80–83. https://doi.org/10.1126/science.aan8048 (2018).Article 
ADS 
CAS 

Google Scholar 
Dana, T. F. Development of contemporary eastern Pacific coral reefs. Mar. Biol. 33, 355–374. https://doi.org/10.1007/BF00390574 (1975).Article 

Google Scholar 
Cortés, J. Eastern tropical Pacific coral reefs. In The Encyclopedia of Modern Coral Reefs: Structure, Form and Process. 351–358 (2011).O’Dea, A., Hoyos, N., Rodríguez, F., Degracia, B. & de Gracia, C. History of upwelling in the tropical eastern Pacific and the paleogeography of the Isthmus of Panama. Palaeogeogr. Palaeoclimatol. Palaeoecol. 348–349, 59–66. https://doi.org/10.1016/j.palaeo.2012.06.007 (2012).Article 

Google Scholar 
Glynn, P. W. & Colgan, M. W. Sporadic disturbances in fluctuating coral reef environments: El Niño and coral reef development in the Eastern Pacific. Am. Zool. 32, 707–718. https://doi.org/10.1093/icb/32.6.707 (1992).Article 

Google Scholar 
Manzello, D. P. et al. Poorly cemented coral reefs of the eastern tropical Pacific: Possible insights into reef development in a high-CO2 world. Proc. Natl. Acad. Sci. USA 105, 10450–10455. https://doi.org/10.1073/pnas.0712167105 (2008).Article 
ADS 

Google Scholar 
Eakin, C. M. & Glynn, P. W. Low tidal exposures and reef mortalities in the eastern Pacific. Coral Reefs 15, 120 (1996).Article 

Google Scholar 
Glynn, P. W. Some physical and biological determinants of coral community structure in the eastern pacific. Ecol. Monogr. 46, 431–456. https://doi.org/10.2307/1942565 (1976).Article 

Google Scholar 
Toth, L. T., Macintyre, I. G. & Aronson, R. B. Holocene reef development in the eastern tropical Pacific. In Coral Reefs of the Eastern Tropical Pacific (eds Glynn, P. W., Manzello, D. P., Enochs, I. C.) 177–201 (Springer, 2017).Cortés, J., Macintyre, I. G. & Glynn, P. W. Holocene growth history of an eastern Pacific fringing reef, Punta Islotes, Costa Rica. Coral Reefs 13, 65–73. https://doi.org/10.1007/BF00300763 (1994).Article 
ADS 

Google Scholar 
Glynn, P. W. et al. Eastern pacific coral reef provinces, coral community structure and composition: An overview. In Coral Reefs of the Eastern Tropical Pacific (eds Glynn, P. W., Manzello, D. P., Enochs, I. C.) 107–176 (Springer, 2017).Glynn, P. W. & Macintyre, I. G. Growth rate and age of coral reefs on the Pacific coast of Panama. In Proceedings of the 3rd International Coral Reef Symposium, Miami, vol. 2, 251–259 (1977).Glynn, P. W. & Stewart, R. H. Distribution of coral reefs in the Pearl Islands (Gulf of Panama) in relation to thermal conditions. Limnol. Oceanogr. 18, 367–379. https://doi.org/10.4319/lo.1973.18.3.0367 (1973).Article 
ADS 

Google Scholar 
Glynn, P. W., Druffel, E. M. & Dunbar, R. B. A dead Central American coral reef tract: Possible link with the Little Ice Age (Costa Rica, Gulf of Papagayo, Gulf of Panama). J. Mar. Res. 41, 605–637. https://doi.org/10.1357/002224083788519740 (1983).Article 

Google Scholar 
Glynn, P. W. & Leyte Morales, G. E. Coral reefs of Huatulco, west México: Reef development in upwelling Gulf of Tehuantepec. Rev. Biol. Trop. 45, 1033–1047 (1997).
Google Scholar 
Tribollet, A. & Golubic, S. Cross-shelf differences in the pattern and pace of bioerosion of experimental carbonate substrates exposed for 3 years on the northern Great Barrier Reef, Australia. Coral Reefs 24, 422–434. https://doi.org/10.1007/s00338-005-0003-7 (2005).Article 
ADS 

Google Scholar 
D’Croz, L. & O’Dea, A. Variability in upwelling along the Pacific shelf of Panama and implications for the distribution of nutrients and chlorophyll. Estuar. Coast. Shelf S. 73, 325–340. https://doi.org/10.1016/j.ecss.2007.01.013 (2007).Article 
ADS 

Google Scholar 
Randall, C. J., Toth, L. T., Leichter, J. J., Maté, J. L. & Aronson, R. B. Upwelling buffers climate change impacts on coral reefs of the eastern tropical Pacific. Ecology 101, e02918. https://doi.org/10.1002/ecy.2918 (2020).Article 

Google Scholar 
Tyberghein, L. et al. Bio-ORACLE: A global environmental dataset for marine species distribution modelling. Glob. Ecol. Biogeogr. 21, 272–281. https://doi.org/10.1111/j.1466-8238.2011.00656.x (2012).Article 

Google Scholar 
Assis, J. et al. Bio-ORACLE v2.0: Extending marine data layers for bioclimatic modelling. Glob. Ecol. Biogeogr. 27, 277–284. https://doi.org/10.1111/geb.12693 (2018).Article 

Google Scholar 
R Development Core Team. R Core Team (2022). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. R Foundation for Statistical Computing vol. 2.Wickham, H. ggplot2: Elegant Graphics for Data Analysis. (Springer, 2016). ISBN 978-3-319-24277-4, https://ggplot2.tidyverse.org.Dunnington, D. ggspatial: Spatial Data Framework for ggplot2. (2022). https://paleolimbot.github.io/ggspatial/, https://github.com/paleolimbot/ggspatial.Toth, L. T. et al. ENSO drove 2500-year collapse of eastern Pacific coral reefs. Science 336, 81–84. https://doi.org/10.1126/science.1221168 (2012).Article 
ADS 
CAS 

Google Scholar 
Toth, L. T. et al. Climatic and biotic thresholds of coral-reef shutdown. Nat. Clim. Change 5, 369–374. https://doi.org/10.1038/nclimate2541 (2015).Article 
ADS 

Google Scholar 
Guzman, H. & Cortés, J. Arrecifes coralinos del Pacífico oriental tropical: revisión y perspectivas. Rev. Biol. Trop. 41, 535–557 (1993).
Google Scholar 
Chollett, I., Mumby, P. J. & Cortés, J. Upwelling areas do not guarantee refuge for coral reefs in a warming Ocean. Mar. Ecol. Prog. Ser. 416, 47–56. https://doi.org/10.3354/meps08775 (2010).Article 
ADS 

Google Scholar 
Glynn, P. W., Maté, J. L., Baker, A. C. & Calderón, M. O. Coral bleaching and mortality in Panama and Ecuador during the 1997–1998 El Niño-Southern Oscillation event: Spatial/temporal patterns and comparisons with the 1982–1983 event. Bull. Mar. Sci. 69, 79–109 (2001).
Google Scholar 
Paz-García, D. A., Hellberg, M. E., García-de-León, F. J. & Balart, E. F. Switch between morphospecies of Pocillopora corals. Am. Nat. 186, 434–440. https://doi.org/10.1086/682363 (2015).Article 

Google Scholar 
Tortolero-Langarica, J. J. A., Rodríguez-Troncoso, A. P., Cupul-Magaña, A. L. & Carricart-Ganivet, J. P. Calcification and growth rate recovery of the reef-building Pocillopora species in the northeast tropical Pacific following an ENSO disturbance. PeerJ 2017, e3191. https://doi.org/10.7717/peerj.3191 (2017).Article 

Google Scholar 
Medellín-Maldonado, F. et al. Calcification of the main reef-building coral species on the Pacific coast of southern Mexico. Cienc. Mar. 42, 209–225. https://doi.org/10.7773/cm.v42i3.2650 (2016).Article 
CAS 

Google Scholar 
Cabral-Tena, R. A. et al. Calcification of coral assemblages in the eastern Pacific: Reshuffling calcification scenarios under climate change. Ecol. Indic. 95, 726–734. https://doi.org/10.1016/j.ecolind.2018.08.021 (2018).Article 
CAS 

Google Scholar 
Glynn, P. Coral growth in upwelling and nonupwelling areas off the Pacific coast of Panama. J. Mar. Res. 35, 567–585 (1977).
Google Scholar 
Guzman, H. M. & Cortes, J. Growth rates of eight species of scleractinian corals in the eastern Pacific (Costa Rica). Bull. Mar. Sci. 44, 1186–1194 (1989).
Google Scholar 
Cabral-Tena, R. A. et al. Functional potential of coral assemblages along a typical eastern tropical Pacific reef tract. Ecol. Indic. 119, 106795. https://doi.org/10.1016/j.ecolind.2020.106795 (2020).Article 

Google Scholar 
Manzello, D. P. Coral growth with thermal stress and ocean acidification: Lessons from the eastern tropical Pacific. Coral Reefs 29, 749–758. https://doi.org/10.1007/s00338-010-0623-4 (2010).Article 
ADS 

Google Scholar 
González-Barrios, F. J. & Álvarez-Filip, L. A framework for measuring coral species-specific contribution to reef functioning in the Caribbean. Ecol. Indic. 95, 877–886. https://doi.org/10.1016/j.ecolind.2018.08.038 (2018).Article 

Google Scholar 
Jokiel, P. L., Maragos, J. & Franzisket, L. Coral growth: Buoyant weight technique. In Coral Reefs: Research Methods (eds Stoddart, D. R., & Johannes, R. E.) 529–541 (UNESCO, 1978).Kuffner, I. B., Hickey, T. D. & Morrison, J. M. Calcification rates of the massive coral Siderastrea siderea and crustose coralline algae along the Florida Keys (USA) outer-reef tract. Coral Reefs 32, 987–997. https://doi.org/10.1007/s00338-013-1047-8 (2013).Article 
ADS 

Google Scholar 
Cornwall, C. E. et al. Global declines in coral reef calcium carbonate production under ocean acidification and warming. Proc. Natl. Acad. Sci. USA 118, e2015265118. https://doi.org/10.1073/pnas.201526511 (2021).Article 
CAS 

Google Scholar 
Rose, C. S. & Risk, M. J. Increase in Cliona delitrix infestation of Montastrea cavernosa heads on an organically polluted portion of the Grand Cayman fringing reef. Mar. Ecol. 6, 345–363. https://doi.org/10.1111/j.1439-0485.1985.tb00142.x (1985).Article 
ADS 

Google Scholar 
Prouty, N. G. et al. Vulnerability of coral reefs to bioerosion from land-based sources of pollution. J. Geophys. Res-Oceans 122, 9319–9331. https://doi.org/10.1002/2017JC013264 (2017).Article 
ADS 
CAS 

Google Scholar 
Eakin, C. M. The damselfish-algal lawn symbiosis and its influence on the bioerosion of an El Niño impacted coral reef, Uva Island, Pacific Panama. ProQuest Dissertations and Theses (1991).Alvarado, J. J., Reyes-Bonilla, H. & Benítez-Villalobos, F. Diadema mexicanum, erizo de mar clave en los arrecifes coralinos del Pacífico Tropical Oriental: Lo que sabemos y perspectivas futuras (Diadematoida: Diadematidae). Rev. Biol. Trop. 63(Suppl 2), 135–157. https://doi.org/10.15517/rbt.v63i2.23140 (2015).Article 

Google Scholar 
Glynn, P. W. Widespread coral mortality and the 1982–83 El Niño warming event. Environ. Conserv. 11, 133–146. https://doi.org/10.1017/S0376892900013825 (1984).Article 

Google Scholar 
Glynn, P. W. El Niño warming, coral mortality and reef framework destruction by echinoid bioerosion in the eastern Pacific. Galaxea 7, 129–160 (1988).
Google Scholar 
Eakin, C. M. A tale of two ENSO events: Carbonate budgets and the influence of two warming disturbances and intervening variability, Uva Island, Panama. Bull. Mar. Sci. 69, 171–186 (2001).ADS 

Google Scholar 
Russ, G. R., Questel, S. L. A., Rizzari, J. R. & Alcala, A. C. The parrotfish–coral relationship: Refuting the ubiquity of a prevailing paradigm. Mar. Biol. 162, 2029–2045. https://doi.org/10.1007/s00227-015-2728-3 (2015).Article 

Google Scholar 
Wellington, G. M. & Glynn, P. W. Responses of Coral Reefs to El Niño-Southern Oscillation Sea-Warming Events. In Geological Approaches to Coral Reef Ecology (ed Aronson, R. B.) 342–385 (Springer, 2007).Guzmán, H. M. & Cortés, J. Changes in reef community structure after fifteen years of natural disturbances in the eastern Pacific (Costa Rica). Bull. Mar. Sci. 69, 133–149 (2001).
Google Scholar 
Guzman, H. M. & Cortés, J. Reef recovery 20 years after the 1982–1983 El Niño massive mortality. Mar. Biol. 151, 401–411. https://doi.org/10.1007/s00227-006-0495-x (2007).Article 

Google Scholar 
Edmunds, P. J. et al. Why more comparative approaches are required in time-series analyses of coral reef ecosystems. Mar. Ecol. Prog. Ser. 608, 297–306. https://doi.org/10.3354/meps12805 (2019).Article 
ADS 

Google Scholar 
Enochs, I. C. et al. Enhanced macroboring and depressed calcification drive net dissolution at high-CO2 coral reefs. Proc. R. Soc. B 283, 20161742. https://doi.org/10.1098/rspb.2016.1742 (2016).Article 
CAS 

Google Scholar 
Roff, G. Reef accretion and coral growth rates are decoupled in Holocene reef frameworks. Mar. Geol. 419, 106065. https://doi.org/10.1016/j.margeo.2019.106065 (2020).Article 
ADS 
CAS 

Google Scholar 
Perry, C. T. et al. Regional-scale dominance of non-framework building corals on Caribbean reefs affects carbonate production and future reef growth. Glob. Change Biol. 21, 1153–1164. https://doi.org/10.1111/gcb.12792 (2015).Article 
ADS 

Google Scholar 
IPCC. Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. IPCC (2014).Neumann, A. C. & Macintyre, I. Reef response to sea level rise: Keep-up, catch-up or give up. In Proceedings 5th International Coral Reef Congress, Tahiti 3, 105–110 (1985).Macintyre, I. G. Modern coral reefs of western Atlantic: New geological perspective. AAPG Bull. 72, 1360–1369. https://doi.org/10.1306/703C99A1-1707-11D7-8645000102C1865D (1988).Article 

Google Scholar 
Hallock, P. & Schlager, W. Nutrient excess and the demise of coral reefs and carbonate platforms. Palaios 1, 389–398. https://doi.org/10.2307/3514476 (1986).Article 
ADS 

Google Scholar 
Kleypas, J. A. Coral reef development under naturally turbid conditions: Fringing reefs near Broad Sound, Australia. Coral Reefs 15, 153–167. https://doi.org/10.1007/BF01145886 (1996).Article 
ADS 

Google Scholar 
van Woesik, R. & Done, T. J. Coral communities and reef growth in the southern Great Barrier Reef. Coral Reefs 16, 103–115. https://doi.org/10.1007/s003380050064 (1997).Article 

Google Scholar 
Sully, S. & van Woesik, R. Turbid reefs moderate coral bleaching under climate-related temperature stress. Glob. Change Biol. 26, 1367–1373. https://doi.org/10.1111/gcb.14948 (2020).Article 
ADS 

Google Scholar 
Grottoli, A. G., Rodrigues, L. J. & Palardy, J. E. Heterotrophic plasticity and resilience in bleached corals. Nature 440, 1186–1189. https://doi.org/10.1038/nature04565 (2006).Article 
ADS 
CAS 

Google Scholar 
Romero-Torres, M. et al. Coral reef resilience to thermal stress in the eastern tropical Pacific. Glob. Change Biol. 26, 3880–3890. https://doi.org/10.1111/gcb.15126 (2020).Article 
ADS 

Google Scholar 
Martínez-Castillo, V., Rodríguez-Troncoso, A. P., Mayfield, A. B., Rodríguez-Zaragoza, F. A. & Cupul-Magaña, A. L. Coral recovery in the central Mexican Pacific 20 years after the 1997–1998 El Niño Event. Oceans 3, 48–59. https://doi.org/10.3390/oceans3010005 (2022).Article 

Google Scholar 
Anton, A. et al. Differential thermal tolerance between algae and corals may trigger the proliferation of algae in coral reefs. Glob. Change Biol. 26, 4316–4327. https://doi.org/10.1111/gcb.15141 (2020).Article 
ADS 

Google Scholar 
Roth, F. et al. High summer temperatures amplify functional differences between coral- and algae-dominated reef communities. Ecology 102, e03226. https://doi.org/10.1002/ecy.3226 (2021).Article 

Google Scholar 
Roik, A., Röthig, T., Pogoreutz, C., Saderne, V. & Voolstra, C. R. Coral reef carbonate budgets and ecological drivers in the central Red Sea—A naturally high temperature and high total alkalinity environment. Biogeosciences 15, 6277–6296. https://doi.org/10.5194/bg-15-6277-2018 (2018).Article 
ADS 
CAS 

Google Scholar 
Wyatt, A. S. J. et al. Heat accumulation on coral reefs mitigated by internal waves. Nat. Geosci. 13, 28–34. https://doi.org/10.1038/s41561-019-0486-4 (2020).Article 
ADS 
CAS 

Google Scholar 
Smith, T. B., Glynn, P. W., Maté, J. L., Toth, L. T. & Gyory, J. A depth refugium from catastrophic coral bleaching prevents regional extinction. Ecology 95, 1663–1673. https://doi.org/10.1890/13-0468.1 (2014).Article 

Google Scholar 
Guest, J. R. et al. A framework for identifying and characterising coral reef “oases” against a backdrop of degradation. J. Appl. Ecol. 55, 2865–2875. https://doi.org/10.1111/1365-2664.13179 (2018).Article 

Google Scholar 
Courtney, T. A. et al. Disturbances drive changes in coral community assemblages and coral calcification capacity. Ecosphere 11, e03066. https://doi.org/10.1002/ecs2.3066 (2020).Article 

Google Scholar 
Bachman, S. D., Kleypas, J. A., Erdmann, M. & Setyawan, E. A global atlas of potential thermal refugia for coral reefs generated by internal gravity waves. Front. Mar. Sci. 9, 1346. https://doi.org/10.3389/fmars.2022.921879 (2022).Article 

Google Scholar 
Dixon, A. M., Forster, P. M., Heron, S. F., Stoner, A. M. & Beger, M. Future loss of local-scale thermal refugia in coral reef ecosystems. PLoS Clim. 1, e0000004. https://doi.org/10.1371/journal.pclm.0000004 (2022).Article 

Google Scholar 
Kuffner, I. B., Stathakopoulos, A., Toth, L. T. & Bartlett, L. A. Reestablishing a stepping-stone population of the threatened elkhorn coral Acropora palmata to aid regional recovery. Endanger. Species Res. 43, 461–473. https://doi.org/10.3354/esr01083 (2020).Article 

Google Scholar 
Perry, C. T., Lange, I. D. & Januchowski-Hartley, F. A. ReefBudget Indo Pacific: Online resource and methodology. http://geography.exeter.ac.uk/reefbudget/ (2018).Nava, H. & Carballo, J. L. Chemical and mechanical bioerosion of boring sponges from Mexican Pacific coral reefs. J. Exp. Biol. 211, 2827–2831. https://doi.org/10.1242/jeb.019216 (2008).Article 

Google Scholar 
Carballo, J. L., Bautista, E., Nava, H., Cruz-Barraza, J. A. & Chávez, J. A. Boring sponges, an increasing threat for coral reefs affected by bleaching events. Ecol. Evol. 3, 872–886. https://doi.org/10.1002/ece3.452 (2013).Article 

Google Scholar 
Smith, T. B. Temperature effects on herbivory for an Indo-Pacific parrotfish in Panamá: Implications for coral-algal competition. Coral Reefs 27, 397–405. https://doi.org/10.1007/s00338-007-0343-6 (2008).Article 
ADS 

Google Scholar 
Glynn, P. W., Enochs, I. C., Afflerbach, J. A., Brandtneris, V. W. & Serafy, J. E. Eastern Pacific reef fish responses to coral recovery following El Niño disturbances. Mar. Ecol. Prog. Ser. 495, 233–247. https://doi.org/10.3354/meps10594 (2014).Article 
ADS 

Google Scholar 
Palacios, M. M., Muñoz, C. G. & Zapata, F. A. Fish corallivory on a pocilloporid reef and experimental coral responses to predation. Coral Reefs 33, 625–636. https://doi.org/10.1007/s00338-014-1173-y (2014).Article 
ADS 

Google Scholar 
Toth, L. T. Holocene Coral-Reef Development in the Tropical Eastern Pacific. (Florida Institute of Technology, 2013).Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D. & R Core Team. nlme: Linear and nonlinear mixed effects models. https://cran.r-project.org/package=nlme.R-project (2021).Holgate, S. J. et al. New data systems and products at the permanent service for mean sea level. J. Coast. Res. 29, 493–504. https://doi.org/10.2112/JCOASTRES-D-12-00175.1 (2013).Article 

Google Scholar 
Permanent Service for Mean Sea Level. Balboa Tide Gauge Data. http://www.psmsl.org/data/obtaining/ (2022). More

Loreau, M. et al. Biodiversity and ecosystem functioning: current knowledge and future challenges. Science 294, 804–808 (2001).Article 
ADS 
CAS 

Google Scholar 
Pimm, S. L., Russell, G. J., Gittleman, J. L. & Brooks, T. M. The future of biodiversity. Science 269, 347–350 (1995).Article 
ADS 
CAS 

Google Scholar 
Newbold, T. et al. Global effects of land use on local terrestrial biodiversity. Nature 520, 45–50. https://doi.org/10.1038/nature14324 (2015).Article 
ADS 
CAS 

Google Scholar 
Cardoso, P. et al. Scientists’ warning to humanity on insect extinctions. Biol. Conserv. 242, 108426. https://doi.org/10.1016/j.biocon.2020.108426 (2020).Article 

Google Scholar 
Hallmann, C. A. et al. More than 75 percent decline over 27 years in total flying insect biomass in protected areas. PLoS ONE 12, e0185809. https://doi.org/10.1371/journal.pone.0185809 (2017).Article 
CAS 

Google Scholar 
Seibold, S. et al. Arthropod decline in grasslands and forests is associated with landscape-level drivers. Nature 574, 671–674. https://doi.org/10.1038/s41586-019-1684-3 (2019).Article 
ADS 
CAS 

Google Scholar 
Sánchez-Bayo, F. & Wyckhuys, K. A. G. Worldwide decline of the entomofauna: A review of its drivers. Biol. Cons. 232, 8–27. https://doi.org/10.1016/j.biocon.2019.01.020 (2019).Article 

Google Scholar 
Yang, L. H. & Gratton, C. Insects as drivers of ecosystem processes. Curr. Opin. Insect Sci. 2, 26–32. https://doi.org/10.1016/j.cois.2014.06.004 (2014).Article 

Google Scholar 
Bowler, D. E., Heldbjerg, H., Fox, A. D., de Jong, M. & Böhning-Gaese, K. Long-term declines of European insectivorous bird populations and potential causes. Conserv. Biol. 33, 1120–1130. https://doi.org/10.1111/cobi.13307 (2019).Article 

Google Scholar 
Biesmeijer, J. C. et al. Parallel declines in pollinators and insect-pollinated plants in Britain and the Netherlands. Science 313, 351–354. https://doi.org/10.1126/science.1127863 (2006).Article 
ADS 
CAS 

Google Scholar 
Tscharntke, T., Klein, A. M., Kruess, A., Steffan-Dewenter, I. & Thies, C. Landscape perspectives on agricultural intensification and biodiversity – ecosystem service management. Ecol. Lett. 8, 857–874. https://doi.org/10.1111/j.1461-0248.2005.00782.x (2005).Article 

Google Scholar 
Scherber, C. et al. Bottom-up effects of plant diversity on multitrophic interactions in a biodiversity experiment. Nature 468, 553–556. https://doi.org/10.1038/nature09492 (2010).Article 
ADS 
CAS 

Google Scholar 
Siemann, E., Tilman, D. & Haarstad, J. Insect species diversity, abundance and body size relationships. Nature 380, 704–706. https://doi.org/10.1038/380704a0 (1996).Article 
ADS 
CAS 

Google Scholar 
Borer, E. T., Seabloom, E. W. & Tilman, D. Plant diversity controls arthropod biomass and temporal stability. Ecol. Lett. 15, 1457–1464. https://doi.org/10.1111/ele.12006 (2012).Article 

Google Scholar 
Ebeling, A. et al. Plant diversity effects on arthropods and arthropod-dependent ecosystem functions in a biodiversity experiment. Basic Appl. Ecol. 26, 50–63. https://doi.org/10.1016/j.baae.2017.09.014 (2018).Article 

Google Scholar 
Ebeling, A. et al. Plant diversity induces shifts in the functional structure and diversity across trophic levels. Oikos 127, 208–219. https://doi.org/10.1111/oik.04210 (2018).Article 

Google Scholar 
Ebeling, A. et al. Plant diversity impacts decomposition and herbivory via changes in aboveground arthropods. PLoS ONE 9, e106529. https://doi.org/10.1371/journal.pone.0106529 (2014).Article 
ADS 
CAS 

Google Scholar 
Marquard, E. et al. Plant species richness and functional composition drive overyielding in a six-year grassland experiment. Ecology 90, 3290–3302 (2009).Article 

Google Scholar 
Tilman, D. et al. Diversity and productivity in a long-term grassland experiment. Science 294, 843–845. https://doi.org/10.1126/science.1060391 (2001).Article 
ADS 
CAS 

Google Scholar 
Simons, N. K. et al. Resource-mediated indirect effects of grassland management on arthropod diversity. PLoS ONE 9, e107033. https://doi.org/10.1371/journal.pone.0107033 (2014).Article 
ADS 
CAS 

Google Scholar 
Wardle, D. A., Nicholson, K. S., Bonner, K. I. & Yeates, G. W. Effects of agricultural intensification on soil-associated arthropod population dynamics, community structure, diversity and temporal variability over a seven-year period. Soil Biol. Biochem. 31, 1691–1706 (1999).Article 
CAS 

Google Scholar 
Luff, M. L. & Rushton, S. P. The ground beetle and spider fauna of managed and unimproved upland pasture. Agr. Ecosyst. Environ. 25, 195–206 (1989).Article 

Google Scholar 
Dennis, P., Young, M. R., Howard, C. L. & Gordon, I. J. The response of epigeal beetles (Col, Carabidae, Staphylinidae) to varied grazing regimes on upland Nardus stricta grasslands. J. Appl. Ecol. 34, 433–443 (1997).Article 

Google Scholar 
Murdoch, W. W., Evans, F. C. & Peterson, C. H. Diversity and pattern in plants and insects. Ecology 53, 819–829 (1972).Article 

Google Scholar 
Siemann, E., Tilman, D., Haarstad, J. & Ritchie, M. Experimental tests of the dependence of arthropod diversity on plant diversity. Am. Nat. 152, 738–750 (1998).Article 
CAS 

Google Scholar 
Joern, A. & Laws, A. N. Ecological mechanisms underlying arthropod species diversity in grasslands. Annu. Rev. Entomol. 58, 19–36. https://doi.org/10.1146/annurev-ento-120811-153540 (2013).Article 
CAS 

Google Scholar 
Hunter, M. D. & Price, P. W. Playing chutes and ladders: Heterogeneity and relative roles of bottom-up and top-down forces in natural communities. Ecology 73, 724–732 (1992).Article 

Google Scholar 
Knops, J. M. H. et al. Effects of plant species richness on invasion dynamics, disease outbreaks, insect abundances and diversity. Ecol. Lett. 2, 286–293 (1999).Article 
CAS 

Google Scholar 
Thiele, H. U. Carabid beetles in their environment. A study on habitat selection by adaptions in physiology and behaviour. (Springer- Verlag, 1977).Harvey, J. A., van der Putten, W. H., Turin, H., Wagenaar, R. & Bezemer, T. M. Effects of changes in plant species richness and community traits on carabid assemblages and feeding guilds. Agr. Ecosyst. Environ. 127, 100–106 (2008).Article 

Google Scholar 
Luff, M. L. Use of Carabids as environmental indicators in grasslands and cereals. Ann. Zool. Fenn. 33, 185–195 (1996).
Google Scholar 
Kotze, D. J. et al. Forty years of carabid beetle research in Europe—from taxonomy, biology, ecology and population studies to bioindication, habitat assessment and conservation. ZooKeys https://doi.org/10.3897/zookeys.100.1523 (2011).Article 

Google Scholar 
Barnes, A. D. et al. Biodiversity enhances the multitrophic control of arthropod herbivory. Sci. Adv. 6, eabb6603. https://doi.org/10.1126/sciadv.abb6603 (2020).Article 
ADS 

Google Scholar 
Bianchi, F. J. J. A., Booij, C. J. H. & Tscharntke, T. Sustainable pest regulation in agricultural landscapes: A review on landscape composition, biodiversity and natural pest control. Proc. R. Soc. B: Biol. Sci. 273, 1715–1727. https://doi.org/10.1098/rspb.2006.3530 (2006).Article 
CAS 

Google Scholar 
Lövei, G. L. & Magura, T. Ground beetle (Coleoptera: Carabidae) diversity is higher in narrow hedges composed of a native compared to non-native trees in a Danish agricultural landscape. Insect Conserv. Divers. 10, 141–150. https://doi.org/10.1111/icad.12210 (2017).Article 

Google Scholar 
Loreau, M. Consumers as maximizers of matter and energy flow in ecosystems. Am. Nat. 145, 22–42. https://doi.org/10.1086/285726 (1995).Article 

Google Scholar 
Mielke, L. et al. Nematode grazing increases the allocation of plant-derived carbon to soil bacteria and saprophytic fungi, and activates bacterial species of the rhizosphere. Pedobiologia 90, 150787. https://doi.org/10.1016/j.pedobi.2021.150787 (2022).Article 

Google Scholar 
Holland, J. M. & Luff, M. L. The effects of agricultural practices on Carabidae in temperate agroecosystems. Integr. Pest Manag. Rev. 5, 109–129. https://doi.org/10.1023/A:1009619309424 (2000).Article 

Google Scholar 
Roscher, C. et al. The role of biodiversity for element cycling and trophic interactions: An experimental approach in a grassland community. Basic Appl. Ecol. 5, 107–121 (2004).Article 

Google Scholar 
Weisser, W. W. et al. Biodiversity effects on ecosystem functioning in a 15-year grassland experiment: Patterns, mechanisms, and open questions. Basic Appl. Ecol. https://doi.org/10.1016/j.baae.2017.06.002 (2017).Article 

Google Scholar 
Freude, H., Harde, K. W. & Lohse, G. A. Die Käfer Mitteleuropas Bd.1–11. (Goecke & Evers, 1965–83).Koch, K. Die Käfer Mitteleuropas. Ökologie Bd.1–6. (Goecke & Evers, 1989–95).R: A language and environment for statistical computing (R Foundation for Statistical Computing, Vienna, Austria, 2021).Bates, D., Mächler, M., Bolker, B. & Walker, S. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67, 1–48. https://doi.org/10.18637/jss.v067.i01 (2015).Article 

Google Scholar 
Schmid, B., Baruffol, M., Wang, Z. & Niklaus, P. A. A guide to analyzing biodiversity experiments. J. Plant Ecol. 10, 91–110. https://doi.org/10.1093/jpe/rtw107 (2017).Article 

Google Scholar 
Zuur, A., Ieno, E. N., Walker, N., Saveliev, A. A. & Smith, G. M. Mixed effects models and extensions in ecology with R. (Springer, 2009).Oksanen, J. et al. vegan: Community Ecology Package v. 2.6–2 (2022).Lenth, R. et al., emmeans: Estimated Marginal Means, aka Least-Squares Means v. 1.8.1-1 (2022).Lovei, G. L. & Sunderland, K. D. Ecology and behavior of ground beetles (Coleoptera: Carabidae). Annu. Rev. Entomol. 41, 231–256 (1996).Article 
CAS 

Google Scholar 
Ravenek, J. M. et al. Long-term study of root biomass in a biodiversity experiment reveals shifts in diversity effects over time. Oikos 123, 1528–1536. https://doi.org/10.1111/oik.01502 (2014).Article 

Google Scholar 
Root, R. Organization of a plant -arthropod association in simple and diverse habitats: the fauna of collards (Brassica oleracea). Ecol. Monogr. 43, 95–124 (1973).Article 

Google Scholar 
Duelli, P. & Obrist, M. K. Regional biodiversity in an agricultural landscape: The contribution of seminatural habitat islands. Basic Appl. Ecol. 4, 129–138 (2003).Article 

Google Scholar 
Perner, J. & Malt, S. Assessment of changing agricultural land use: Response of vegetation, ground-dwelling spiders and beetles to the conversion of arable land into grassland. Agr. Ecosyst. Environ. 98, 169–181 (2003).Article 

Google Scholar 
Purtauf, T., Dauber, J. & Wolters, V. Carabid communities in the spatio-temporal mosaic of a rural landscape. Landsc. Urban Plan. 67, 185–193 (2004).Article 

Google Scholar 
Eisenhauer, N. et al. Biotic interactions, community assembly, and eco-evolutionary dynamics as drivers of long-term biodiversity–ecosystem functioning relationships. Res. Ideas Outcomes https://doi.org/10.3897/rio.5.e47042 (2019).Article 

Google Scholar 
Guerrero-Ramirez, N. R. et al. Diversity-dependent temporal divergence of ecosystem functioning in experimental ecosystems. Nat. Ecol. Evol. 1, 1639–1642. https://doi.org/10.1038/s41559-017-0325-1 (2017).Article 

Google Scholar 
Reich, P. B. et al. Impacts of biodiversity loss escalate through time as redundancy fades. Science 336, 589–592. https://doi.org/10.1126/science.1217909 (2012).Article 
ADS 
CAS 

Google Scholar 
Isbell, F. I., Polley, H. W. & Wilsey, B. J. Biodiversity, productivity and the temporal stability of productivity: Patterns and processes. Ecol. Lett. 12, 443–451. https://doi.org/10.1111/j.1461-0248.2009.01299.x (2009).Article 

Google Scholar 
Blake, S., Foster, G. N., Fisher, G. E. J. & Ligertwood, G. L. Effects of management practices on the carabid faunas of newly established wildflower meadows in southern Scotland. Ann. Zool. Fenn. 33, 139–147 (1996).
Google Scholar 
Boetzl, F. A., Krimmer, E., Krauss, J. & Steffan-Dewenter, I. Agri-environmental schemes promote ground-dwelling predators in adjacent oilseed rape fields: Diversity, species traits and distance-decay functions. J. Appl. Ecol. 56, 10–20. https://doi.org/10.1111/1365-2664.13162 (2019).Article 

Google Scholar 
Knapp, M., Seidl, M., Knappová, J., Macek, M. & Saska, P. Temporal changes in the spatial distribution of carabid beetles around arable field-woodlot boundaries. Sci. Rep. 9, 8967. https://doi.org/10.1038/s41598-019-45378-7 (2019).Article 
ADS 
CAS 

Google Scholar  More

Global Forest Resources Assessment 2020 (FAO, 2020).Taubert, F. et al. Global patterns of tropical forest fragmentation. Nature 554, 519–522 (2018).ADS 
CAS 

Google Scholar 
Vancutsem, C. et al. Long-term (1990–2019) monitoring of forest cover changes in the humid tropics. Sci. Adv. 7, eabe1603 (2021).ADS 

Google Scholar 
Foley, J. A. et al. Amazonia revealed: Forest degradation and loss of ecosystem goods and services in the Amazon Basin. Front. Ecol. Environ. 5, 25–32 (2007).
Google Scholar 
Barlow, J. et al. The future of hyperdiverse tropical ecosystems. Nature 559, 517–526 (2018).ADS 
CAS 

Google Scholar 
Brandon, K. Ecosystem services from tropical forests: Review of current science. SSRN J. https://doi.org/10.2139/ssrn.2622749 (2014). Article 

Google Scholar 
Indarto, J. & Mutaqin, D. J. An overview of theoretical and empirical studies on deforestation. MPRA. Paper No. 70178 (2016).Geist, H. J. & Lambin, E. F. Proximate causes and underlying driving forces of tropical deforestation: Tropical forests are disappearing as the result of many pressures, both local and regional, acting in various combinations in different geographical locations. Bioscience 52, 143–150 (2002).
Google Scholar 
Angelsen, A. & Kaimowitz, D. Rethinking the causes of deforestation: Lessons from economic models. World Bank Res. Obs. 14, 73–98 (1999).CAS 

Google Scholar 
Contreras-Hermosilla, A. The Underlying Causes of Forest Decline (Center for International Forestry Research, 2000).
Google Scholar 
Turner, B. L. et al. Two types of global environmental change: Definitional and spatial-scale issues in their human dimensions. Glob. Environ. Change 1, 14–22 (1990).
Google Scholar 
Meyer, W. B. & Turner, B. L. Human population growth and global land-use/cover change. Ann. Rev. Ecol. Syst. 2, 39–61 (1992).
Google Scholar 
Miyamoto, M., Mohd Parid, M., Noor Aini, Z. & Michinaka, T. Proximate and underlying causes of forest cover change in Peninsular Malaysia. For. Policy Econ. 44, 18–25 (2014).
Google Scholar 
Lim, C. L., Prescott, G. W., De Alban, J. D. T., Ziegler, A. D. & Webb, E. L. Untangling the proximate causes and underlying drivers of deforestation and forest degradation in Myanmar. Conserv. Biol. 31, 1362–1372 (2017).
Google Scholar 
Carodenuto, S. et al. A methodological framework for assessing agents, proximate drivers and underlying causes of deforestation: Field test results from southern cameroon. Forests 6, 203–224 (2015).
Google Scholar 
Curtis, P. G., Slay, C. M., Harris, N. L., Tyukavina, A. & Hansen, M. C. Classifying drivers of global forest loss. Science 361, 1108–1111 (2018).ADS 
CAS 

Google Scholar 
Hosonuma, N. et al. An assessment of deforestation and forest degradation drivers in developing countries. Environ. Res. Lett. 7, 044009 (2012).ADS 

Google Scholar 
Köthke, M., Leischner, B. & Elsasser, P. Uniform global deforestation patterns—An empirical analysis. For. Policy Econ. 28, 23–37 (2013).
Google Scholar 
Busch, J. & Ferretti-Gallon, K. What drives deforestation and what stops it? A meta-analysis. Rev. Environ. Econ. Policy 11, 3–23 (2017).
Google Scholar 
Ferrer Velasco, R. F., Köthke, M., Lippe, M. & Günter, S. Scale and context dependency of deforestation drivers: Insights from spatial econometrics in the tropics. PLoS One 15, e0226830 (2020).CAS 

Google Scholar 
Lambin, E. F. et al. Effectiveness and synergies of policy instruments for land use governance in tropical regions. Glob. Environ. Change 28, 129–140 (2014).
Google Scholar 
Börner, J., Schulz, D., Wunder, S. & Pfaff, A. The effectiveness of forest conservation policies and programs. Ann. Rev. Resour. Econ. 12, 45–64 (2020).
Google Scholar 
Bemelmans-Videc, M.-L., Rist, R. C. & Vedung, E. Carrots, Sticks & Sermons: Policy Instruments and their Evaluation (Transaction Publishers, 1998).
Google Scholar 
Seymour, F. & Harris, N. L. Reducing tropical deforestation. Science 365, 756–757 (2019).ADS 
CAS 

Google Scholar 
Lambin, E. F. et al. The role of supply-chain initiatives in reducing deforestation. Nat. Clim. Change 8, 109–116 (2018).ADS 

Google Scholar 
Wolff, S. & Schweinle, J. Effectiveness and economic viability of forest certification: A systematic review. Forests 13, 798 (2022).
Google Scholar 
Müller, R., Pistorius, T., Rohde, S., Gerold, G. & Pacheco, P. Policy options to reduce deforestation based on a systematic analysis of drivers and agents in lowland Bolivia. Land Use Policy 30, 895–907 (2013).
Google Scholar 
Tegegne, Y. T., Lindner, M., Fobissie, K. & Kanninen, M. Evolution of drivers of deforestation and forest degradation in the Congo Basin forests: Exploring possible policy options to address forest loss. Land Use Policy 51, 312–324 (2016).
Google Scholar 
Hoffmann, C., García Márquez, J. R. & Krueger, T. A local perspective on drivers and measures to slow deforestation in the Andean-Amazonian foothills of Colombia. Land Use Policy 77, 379–391 (2018).
Google Scholar 
Henders, S., Ostwald, M., Verendel, V. & Ibisch, P. Do national strategies under the UN biodiversity and climate conventions address agricultural commodity consumption as deforestation driver?. Land Use Policy 70, 580–590 (2018).
Google Scholar 
Salvini, G. et al. How countries link REDD+ interventions to drivers in their readiness plans: implications for monitoring systems. Environ. Res. Lett. 9, 074004 (2014).ADS 

Google Scholar 
Bos, A. B. et al. Integrated assessment of deforestation drivers and their alignment with subnational climate change mitigation efforts. Environ. Sci. Policy 114, 352–365 (2020).CAS 

Google Scholar 
Fritz, S. et al. A continental assessment of the drivers of tropical deforestation with a focus on protected areas. Front. Conserv. Sci. https://doi.org/10.3389/fcosc.2022.830248 (2022).Article 

Google Scholar 
Lawrence, D. & Vandecar, K. Effects of tropical deforestation on climate and agriculture. Nat. Clim. Change 5, 27–36 (2015).ADS 

Google Scholar 
Fedele, G., Locatelli, B., Djoudi, H. & Colloff, M. J. Reducing risks by transforming landscapes: Cross-scale effects of land-use changes on ecosystem services. PLoS One 13, e0195895 (2018).
Google Scholar 
Yackulic, C. B. et al. Biophysical and socioeconomic factors associated with forest transitions at multiple spatial and temporal scales. Ecol. Soc. https://doi.org/10.5751/ES-04275-160315 (2011).Article 

Google Scholar 
Loran, C., Ginzler, C. & Bürgi, M. Evaluating forest transition based on a multi-scale approach: Forest area dynamics in Switzerland 1850–2000. Reg. Environ. Change 16, 1807–1818 (2016).
Google Scholar 
Moonen, P. C. et al. Actor-based identification of deforestation drivers paves the road to effective REDD+in DR Congo. Land Use Policy 58, 123–132 (2016).
Google Scholar 
Strassburg, B. The tragedy of the tropics: A dynamic, cross-scale analysis of deforestation incentives. Working Paper—Centre for Social and Economic Research on the Global Environment No. 07-02 (2007).López-Carr, D. et al. Space versus place in complex human–natural systems: Spatial and multi-level models of tropical land use and cover change (LUCC) in Guatemala. Ecol. Model. 229, 64–75 (2012).
Google Scholar 
Hoang, N. T. & Kanemoto, K. Mapping the deforestation footprint of nations reveals growing threat to tropical forests. Nat. Ecol. Evol. 5, 845–853 (2021).
Google Scholar 
Pendrill, F. et al. Agricultural and forestry trade drives large share of tropical deforestation emissions. Glob. Environ. Change 56, 1–10 (2019).
Google Scholar 
Ferrer Velasco, R. et al. Towards accurate mapping of forest in tropical landscapes: A comparison of datasets on how forest transition matters. Remote Sens. Environ. 274, 112997 (2022).ADS 

Google Scholar 
Jayathilake, H. M., Prescott, G. W., Carrasco, L. R., Rao, M. & Symes, W. S. Drivers of deforestation and degradation for 28 tropical conservation landscapes. Ambio 50, 215–228 (2021).
Google Scholar 
Minang, P. A. et al. REDD+Readiness progress across countries: Time for reconsideration. Clim. Policy 14, 685–708 (2014).
Google Scholar 
Current pledges | Bonn challenge. https://www.bonnchallenge.org/pledges. Accessed: 15th August 2022.Nansikombi, H. et al. Can de facto governance influence deforestation drivers in the Zambian Miombo?. For. Policy Econ. 120, 102309 (2020).
Google Scholar 
Sullivan, A., York, A., White, D., Hall, S. & Yabiku, S. D. Jure versus de facto institutions: Trust, information, and collective efforts to manage the invasive mile-a-minute weed (Mikania micrantha). Int. J. Commons 11, 171–199 (2017).
Google Scholar 
Busch, J. & Amarjargal, O. Authority of second-tier governments to reduce deforestation in 30 tropical countries. Front. For. Glob. Change https://doi.org/10.3389/ffgc.2020.00001 (2020).Article 

Google Scholar 
Sandström, C., Eckerberg, K. & Raitio, K. Studying conflicts, proposing solutions—Towards multi-level approaches to the analyses of forest conflicts. For. Policy Econ. 33, 123–127 (2013).
Google Scholar 
Hoogstra-Klein, M. A., Permadi, D. B. & Yasmi, Y. The value of cultural theory for participatory processes in natural resource management. For. Policy Econ. 20, 99–106 (2012).
Google Scholar 
de Jong, W., Ruiz, S. & Becker, M. Conflicts and communal forest management in northern Bolivia. For. Policy Econ. 8, 447–457 (2006).
Google Scholar 
Eckerberg, K. & Sandström, C. Forest conflicts: A growing research field. For. Policy Econ. 33, 3–7 (2013).
Google Scholar 
Sierra, R., Calva, O. & Guevara, A. La Deforestación en el Ecuador, 1990–2018. Factores promotores y tendencias recientes, 216 (2021).Wasserstrom, R. & Southgate, D. Deforestation, agrarian reform and oil development in Ecuador, 1964–1994. Nat. Resour. 04, 31 (2013).
Google Scholar 
Wiebe, P. C., Zhunusova, E., Lippe, M., Ferrer Velasco, R. & Günter, S. What is the contribution of forest-related income to rural livelihood strategies in the Philippines’ remaining forested landscapes?. For. Policy Econ. 135, 102658 (2022).
Google Scholar 
Le, H. D., Smith, C. & Herbohn, J. What drives the success of reforestation projects in tropical developing countries? The case of the Philippines. Glob. Environ. Change 24, 334–348 (2014).
Google Scholar 
Carandang, A. P. et al. Analysis of key drivers of deforestation and forest degradation in the Philippines. Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) (2013).Phiri, D., Morgenroth, J. & Xu, C. Four decades of land cover and forest connectivity study in Zambia—An object-based image analysis approach. Int. J. Appl. Earth Obs. Geoinf. 79, 97–109 (2019).ADS 

Google Scholar 
Nansikombi, H., Fischer, R., Kabwe, G. & Günter, S. Exploring patterns of forest governance quality: Insights from forest frontier communities in Zambia’s Miombo ecoregion. Land Use Policy 99, 104866 (2020).
Google Scholar 
Zhang, H., Wang, P. & Wood, J. Does institutional quality matter for the nexus between environmental quality and economic growth?: A tropics perspective. In Business, Industry, and Trade in the Tropics (eds Wood, J. et al.) (Routledge, 2022).
Google Scholar 
Reed, J., Van Vianen, J., Deakin, E. L., Barlow, J. & Sunderland, T. Integrated landscape approaches to managing social and environmental issues in the tropics: Learning from the past to guide the future. Glob. Change Biol. 22, 2540–2554 (2016).ADS 

Google Scholar 
Fischer, R. et al. Interplay of governance elements and their effects on deforestation in tropical landscapes: Quantitative insights from Ecuador. World Dev. 148, 105665 (2021).
Google Scholar 
Torres, B., Vasco, C., Günter, S. & Knoke, T. Determinants of agricultural diversification in a hotspot area: Evidence from colonist and indigenous communities in the Sumaco biosphere reserve Ecuadorian Amazon. Sustainability 10, 1432 (2018).
Google Scholar 
Ojeda Luna, T., Zhunusova, E., Günter, S. & Dieter, M. Measuring forest and agricultural income in the Ecuadorian lowland rainforest frontiers: Do deforestation and conservation strategies matter?. For. Policy Econ. 111, 102034 (2020).
Google Scholar 
Kazungu, M. et al. Effects of household-level attributes and agricultural land-use on deforestation patterns along a forest transition gradient in the Miombo landscapes Zambia. Ecol. Econ. 186, 107070 (2021).
Google Scholar 
Kleemann, J. et al. Deforestation in continental ecuador with a focus on protected areas. Land 11, 268 (2022).
Google Scholar 
Mulenga, M. M. & Roos, A. Assessing the awareness and adoptability of pellet cookstoves for low-income households in Lusaka, Zambia. J. Energy South. Afr. 32, 52–61 (2021).
Google Scholar 
Eguiguren, P., Ojeda Luna, T., Torres, B., Lippe, M. & Günter, S. Ecosystem service multifunctionality: Decline and recovery pathways in the amazon and chocó lowland rainforests. Sustainability 12, 7786 (2020).CAS 

Google Scholar 
Vasco, C., Torres, B., Pacheco, P. & Griess, V. The socioeconomic determinants of legal and illegal smallholder logging: Evidence from the Ecuadorian Amazon. For. Policy Econ. 78, 133–140 (2017).
Google Scholar 
van der Ploeg, J., van Weerd, M., Masipiqueña, A. B. & Persoon, G. A. Illegal logging in the Northern Sierra Madre Natural Park, the Philippines. Conserv. Soc. 9, 202–215 (2011).
Google Scholar 
Liu, D. S., Iverson, L. R. & Brown, S. Rates and patterns of deforestation in the Philippines: Application of geographic information system analysis. For. Ecol. Manag. 57, 1–16 (1993).
Google Scholar 
Boquet, Y. Environmental challenges in the Philippines. In The Philippine Archipelago (ed. Boquet, Y.) 779–829 (Springer International Publishing, 2017). https://doi.org/10.1007/978-3-319-51926-5_22.Chapter 

Google Scholar 
MAGAP. ATPA: Reconversión Agro productiva Sostenible en la Amazonía Ecuatoriana (2014).Jones, K. W. et al. Forest conservation incentives and deforestation in the Ecuadorian Amazon. Environ. Conserv. 44, 56–65 (2017).
Google Scholar 
Lindsey, P. A. et al. Underperformance of African protected area networks and the case for new conservation models: Insights from Zambia. PLoS One 9, e94109 (2014).ADS 

Google Scholar 
Fischer, R. et al. Effectiveness of policy instrument mixes for forest conservation in the tropics – a stakeholder perspective from Ecuador, the Philippines and Zambia. Land Use Policy https://doi.org/10.1016/j.landusepol.2023.106546 (2022).Article 

Google Scholar 
Gurney, G. G. et al. Biodiversity needs every tool in the box: Use OECMs. Nature 595, 646–649 (2021).ADS 
CAS 

Google Scholar 
Maxwell, S. L. et al. Area-based conservation in the twenty-first century. Nature 586, 217–227 (2020).ADS 
CAS 

Google Scholar 
Priebe, J. et al. Transformative change in context—Stakeholders’ understandings of leverage at the forest–climate nexus. Sustain. Sci. 17, 1921–1938 (2022).
Google Scholar 
Höhl, M. et al. Forest landscape restoration—What generates failure and success?. Forests 11, 938 (2020).
Google Scholar 
Köthke, M., Ahimbisibwe, V. & Lippe, M. The evidence base on the environmental, economic and social outcomes of agroforestry is patchy—An evidence review map. Front. Environ. Sci. https://doi.org/10.3389/fenvs.2022.925477 (2022).Article 

Google Scholar 
Fischer, R., Giessen, L. & Günter, S. Governance effects on deforestation in the tropics: A review of the evidence. Environ. Sci. Policy 105, 84–101 (2020).
Google Scholar 
Bare, M., Kauffman, C. & Miller, D. C. Assessing the impact of international conservation aid on deforestation in sub-Saharan Africa. Environ. Res. Lett. 10, 125010 (2015).ADS 

Google Scholar 
Vuohelainen, A. J., Coad, L., Marthews, T. R., Malhi, Y. & Killeen, T. J. The effectiveness of contrasting protected areas in preventing deforestation in Madre de Dios. Peru. Environ. Manag. 50, 645–663 (2012).ADS 

Google Scholar 
Hull, V. & Liu, J. Telecoupling: A new frontier for global sustainability. Ecol. Soc. 23, 41 (2018).
Google Scholar 
Aitchison, J. The statistical analysis of compositional data. J. Roy. Stat. Soc. 44, 139–160 (1982).MathSciNet 
MATH 

Google Scholar 
Norman, G. Likert scales, levels of measurement and the “laws” of statistics. Adv. Health Sci. Educ. 15, 625–632 (2010).
Google Scholar 
Day, M., Gumbo, D., Moombe, K. B., Wijaya, A. & Sunderland, T. Zambia Country Profile: Monitoring, Reporting and Verification for REDD+ Vol. 113 (CIFOR, 2014).
Google Scholar 
Piotrowski, M. Nearing the tipping point. Drivers of Deforestation in the Amazon Region (2019).Sarker, P. K., Fischer, R., Tamayo, F., Navarrete, B. T. & Günter, S. Analyzing forest policy mixes based on the coherence of policies and the consistency of legislative policy instruments: A case study from Ecuador. For. Policy Econ. 144, 102838 (2022).
Google Scholar 
Likert, R. A technique for the measurement of attitudes. Arch. Psychol. 22(140), 55–55 (1932).
Google Scholar 
Altinsoy, M. et al. Ambulatory ECG monitoring for syncope and collapse in United States, Europe, and Japan: The patients’ viewpoint. J. Arrhythm. 37, 1023–1030 (2021).
Google Scholar 
R Core Team. R: A language and environment for statistical computing. https://www.R-project.org/. (R Foundation for Statistical Computing, Vienna, Austria, 2022).Kassambara, A. rstatix: Pipe-friendly framework for basic statistical tests. R package version 0.7.0 (2021).Kassambara, A. & Mundt, F. factoextra: Extract and visualize the results of multivariate data analyses. R package version 1.0.7 (2020).Komsta, L. & Novometsky, F. moments: Moments, cumulants, skewness, kurtosis and related tests. R package version 0.14.1 (2022).Zhang, Y., Zhou, M. & Shao, Y. mvnormalTest: Powerful tests for multivariate normality. R package version 1.0.0 (2020).Kassambara, A. ggpubr: ‘ggplot2’ Based Publication Ready Plots. R package version 0.4.0 (2020).Wickham, H. et al. Welcome to the Tidyverse. JOSS 4, 1686 (2019).ADS 

Google Scholar 
Bache, S. M. & Wickham, H. magrittr: A Forward-Pipe Operator for R. R package version 2.0.3 (2022).Ushey, K., Allaire, J., Wickham, H. & Ritchie, G. rstudioapi: Safely Access the RStudio API. R package version 0.13 (2020).Wickham, H. ggplot2: Elegant Graphics for Data Analysis (Springer International Publishing, 2016). https://doi.org/10.1007/978-3-319-24277-4.Book 
MATH 

Google Scholar 
Wilkins, D. treemapify: Draw Treemaps in ‘ggplot2’. R package version 2.5.5 (2021).Shapiro, S. S. & Wilk, M. B. An analysis of variance test for normality (complete samples). Biometrika 52, 591–611 (1965).MathSciNet 
MATH 

Google Scholar 
Mardia, K. V. Measures of multivariate skewness and kurtosis with applications. Biometrika 57, 519–530 (1970).MathSciNet 
MATH 

Google Scholar 
Kruskal, W. H. & Wallis, W. A. Use of ranks in one-criterion variance analysis. J. Am. Stat. Assoc. 47, 583–621 (1952).MATH 

Google Scholar 
Dunn, O. J. Multiple comparisons using rank sums. Technometrics 6, 241–252 (1964).
Google Scholar 
Conover, W. J. & Iman, R. L. Rank transformations as a bridge between parametric and nonparametric statistics. Am. Stat. 35, 124–129 (1981).MATH 

Google Scholar 
Student,. The probable error of a mean. Biometrika 6, 1–25 (1908).MATH 

Google Scholar 
Tukey, J. W. Comparing individual means in the analysis of variance. Biometrics 5, 99–114 (1949).MathSciNet 
CAS 

Google Scholar 
Jolliffe, I. T. Principal Component Analysis (Springer, 2002).MATH 

Google Scholar  More