LiDAR-derived forest structure metrics show ecologically defined scales outperform grids when predicting mammal diversity

LiDAR-derived forest structure metrics show ecologically defined scales outperform grids when predicting mammal diversity | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article LiDAR-derived forest structure metrics show ecologically defined scales outperform grids when predicting mammal diversity Gene Estrada, Heiko Wittmer, Endro Setiawan, Andrew Marshall This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5479224/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Vegetation structure has emerged as a key determinant of terrestrial biodiversity based on studies using randomly placed sampling grids, often at broad spatial scales. The resultant grid cells often contain substantial heterogeneity in ecological conditions that are highly relevant for the taxa of interest, potentially undermining our ability to detect relevant drivers of diversity. Here we use 15 structural metrics measured using ground-based LiDAR to model mammalian diversity at 58 sampling locations across seven distinct tropical forest types in Indonesian Borneo. We conducted analyses at four spatial scales using over five years of camera trap data. Models predicting mammal diversity based on ecologically defined scales (i.e., forest type boundaries) outperformed models using a grid scale of comparable resolution. Our results underscore the value of LiDAR in capturing forest structural metrics relevant to mammals and highlight the importance of incorporating ecologically meaningful spatial scales in biodiversity studies. Biological sciences/Ecology/Forest ecology Earth and environmental sciences/Ecology/Tropical ecology LiDAR forest structure remote sensing mammal diversity spatial scale Figures Figure 1 Figure 2 Figure 3 Figure 4 INTRODUCTION Understanding the ecological factors that influence biodiversity is of considerable theoretical interest and has substantial value for conservation. Research on the predictors of animal diversity has spanned abiotic factors such as climate and temperature (Hawkins et al., 2003 ; Wiens, 2011 ), precipitation (Currie, 1991 ), sampling area (MacArthur & Wilson 1967), latitude (McCoy & Connor, 1980), elevation and topography (McCain, 2005 ), and biotic factors such as food availability (Kissling et al., 2010 ), species interactions (Chase & Leibold, 2003 ), and vegetation structure (Tews et al., 2004 ). The last has emerged as a key determinant of land-dwelling species richness (e.g., mammals, Stein et al., 2014 ), providing support for the ‘habitat heterogeneity hypothesis’, which predicts that habitats with higher structural complexity provide more niche space and thus support higher species richness than habitats with lower structural complexity (MacArthur & MacArthur, 1961 ). Most of the research investigating the influence of vegetation structure and habitat heterogeneity on species diversity has been conducted at broad spatial scales (> 1 km 2 ; Stein et al., 2014 ). Scale, however, has been identified as the single most important aspect for understanding both habitat associations and use for terrestrial species (Johnson, 1980 ). Research into diversity patterns at broad scales thus risks overlooking important ecological variation occurring at smaller spatial scales. For example, variation in structural characteristics or vegetation such as canopy height, canopy cover, and vegetation density significantly impacts mammal presence at micro-scales (< 80 m 2 ) in addition to larger scales (Püttker et al., 2008 ). Similar research found canopy cover predicts bird species diversity best at small scales (0.5 km 2 ; Hazell et al., 2021 ). To better understand how vegetation structure influences biodiversity, it is crucial to analyze these effects at smaller, potentially more ecologically relevant scales, which may reveal patterns missed by broader analyses. Tropical forests are particularly relevant for such an analysis. They exhibit the highest structural complexity of all forests (Ehbrecht et al., 2021 ) and house an estimated 62% of the world’s vertebrate species (Pillay et al., 2022 ) while covering only 7% of Earth’s land surface (Bradshaw et al., 2009 ). Despite their global importance, these ecosystems remain among the least understood and most highly threatened. Investigating their structural complexity is crucial, as it plays a pivotal role in shaping both ecosystem functioning and biodiversity (Penone et al., 2019 ; Stark et al., 2012 ). LiDAR technology is a powerful tool for describing forest structure at fine spatial scales. LiDAR surveys using aerial laser scanning (ALS) methods have been used for the past few decades to estimate structural characteristics such as forest height, aboveground biomass, and leaf area index (e.g., Drake et al., 2002 ; Ioki et al., 2014 ; Tang et al., 2012 ). While ALS efficiently surveys forest canopies, it is not an effective tool to characterize lower canopy structure or near-ground vegetation (White et al., 2016 ), especially in dense tropical rainforests. Terrestrial laser scanning (TLS), which uses a tripod-mounted LiDAR instrument to scan forests from the ground, is better suited to capture below-canopy structure characteristics. Indeed, TLS has been successfully used to describe forest structure across various tropical sites (e.g., Aiba et al., 2013 ; Ehbrecht et al., 2021 ; Palace et al., 2016 ). Here, we present an analysis of the terrestrial mammal community in Gunung Palung National Park (GPNP), West Kalimantan, Indonesia. Our study area spans over 1,000 m of elevation and contains 12 distinct forest type partitions (Fig. 1 ). We used forest structure metrics derived from TLS surveys at 58 locations to predict six mammal diversity metrics calculated from camera trap data across four spatial scales: individual camera trap/TLS survey locations (0.002 km 2 ), forest type partitions (~ 1.5 km 2 ), a grid with 0.75 km 2 cells, and a grid with 1.7 km 2 cells. We expect, based on predictions of the habitat heterogeneity hypothesis, that structure metrics will be positively associated with diversity metrics. We also expect structure metrics that are more ecologically relevant to terrestrial mammals, such as terrestrial vegetation density, to be more influential than metrics describing higher forest strata, such as canopy roughness (i.e., rumple index). Furthermore, we expect to find that different structure metrics best predict diversity metrics at different spatial scales. Our overall objective was thus to explore the scale-dependency between vegetation structure metrics and biodiversity measures. RESULTS Model output Eight of the 15 forest structure metrics we considered were reliable predictors of diversity metrics across the four spatial scales we modeled. Predictors differed between scales with little overlap. Only two structural metrics appear in models at multiple scales: rumple index which is a positive predictor at the camera trap and forest type partition scales, and mean tree height which is a negative predictor in both the 0.75 km 2 and 1.7 km 2 grid scale models. The other six forest structure metrics only appear in models at one spatial scale, including stand density and basal area in the camera trap scale models, mean dbh in the 0.75 km 2 scale models, vertical complexity in the 1.7 km 2 scale models, and vegetation volume and vertical point distribution in the partition scale models (Table 1 ). The patterns of reliable predictors were highly consistent across models for different diversity metrics, with certain structure metrics repeatedly emerging as reliable or near-reliable predictors. For brevity, we focus mainly on the results of the Shannon diversity models here. Detailed model results for all other metrics (species richness, species evenness, functional evenness, and functional divergence) are available in the supplementary materials. Due to model convergence issues and unstable coefficient estimates we do not report results from the functional richness models. To evaluate model fit, we focused on the forest type partition and 1.7 km 2 grid scales, as they are the most comparable in spatial resolution. Models at the partition scale consistently outperformed those at the 1.7 km 2 grid scale across all diversity metrics considered. For instance, in the Shannon diversity model, the partition scale yielded a deviance of 3.2, compared to 7.5 at the 1.7 km 2 grid scale. When comparing predictor effects across scales, it is useful to group the scales into two categories. The camera trap and partition scales can be grouped together since they are ecologically defined – both are only associated with a single forest type. The two grid scales form the second group based on the use of a randomly placed grid system. This approach allows us to compare the ecologically defined scales to those based on random grids. We can also create three groups of forest structure metrics based on the pattern of effect direction between the ecological and grid scales: if the direction of the effect is in the opposite direction, same direction, or mixed between the two scale groups. For instance, in the Shannon diversity models (Fig. 2 ), seven predictors show opposite direction of effects between the ecological and grid scales, with one group showing a directional effect while the other shows an effect in the opposite direction (e.g., rumple index), or one group showing no effect while the other shows a directional effect (e.g., mean tree height). Four predictors in the Shannon diversity models are grouped based on most models at different scales showing predictor effects in the same direction (e.g., standard deviation of tree heights shows a negative effect at all scales), including showing no effect. The final four metrics are grouped based on no consistent pattern apparent in predictor effect directions between the two scale groups. LiDAR-derived metrics Metrics such as maximum, average, and standard deviation of tree heights, average dbh, tree volume, canopy relief ratio, leaf area density, and vertical rumple follow a reverse C-shaped pattern across forest types (Fig. 3 ). These values are lower at low elevation locations in the peat swamp and freshwater swamp forests, increasing through the alluvial bench and lowland sandstone forests, and incrementally decrease as elevation increases, in many cases reaching their lowest values in the montane forests. Additionally, higher elevation forest types, especially montane forests, show the least amount variation in these metrics compared to other forests. Other metrics show a slightly different relationship with forest type and elevation. Basal area, tree density, and terrestrial vegetation metrics follow a pattern similar to the C-shape of the tree height metrics, for example, with low values in the peat swamp forests and higher values in the freshwater swamp and alluvial bench forests. However, these values increase in higher elevation forests, creating an S-shaped pattern in the elevation plots (Fig. S1). This indicates that tree densities, along with terrestrial vegetation density, increase in higher elevation forests, where trees are shorter and smaller in diameter, and account for an average basal area comparable to lower elevation forests (i.e., lowland granite and lowland sandstone) where taller and larger diameter trees are abundant. The vertical point distribution and vegetation volume metrics show a pattern of generally increasing values from lower to higher elevation forest types, with the lowest average and minimum values in peat swamp and freshwater swamp forests and the highest average values in the upland granite and montane forests. Variation in the metrics is higher in the low elevation forests and decreases with elevation. The patterns in these metrics describe vegetation that is more vertically clumped and with more empty space in the lower elevation forests, particularly in the alluvial bench and lowland sandstone forests, where large trees with extensive branching crowns are more common, compared to the higher elevation forests where smaller, lower volume tree crowns are common, resulting in a more even distribution of vegetation in 3D space. Mammal diversity metrics We used observations of 36 mammal species (n = 13,722) that we identified from camera trap videos to calculate species richness, Shannon diversity, and evenness metrics, and 35 mammal species (excluding unidentified rats, see Methods; n = 10,934) to calculate functional richness, functional divergence, and functional evenness metrics. Species richness, Shannon diversity, and functional richness values exhibit a reverse C-shaped pattern similar to many of the forest structure metrics, with average values by forest type lowest at low and high elevation forests, and the highest average values in lowland sandstone and lowland granite forests (Fig. S4). Species evenness and functional evenness show little variation in average values by forest type, with slightly lower average values in the lower elevation forests. This pattern is reflected in the nearly flat trend lines observed when these values are plotted as a function of elevation (Fig. S2). Average functional diversity values by forest type shows a unique pattern with the lowest values observed in the peat swamp forests, gradually increasing to the highest values in the lowland sandstone forests, and minimally decreasing in the upper elevation forests. DISCUSSION Our results show that the scale at which forest structure metrics are considered matters when predicting species diversity. A clear pattern emerged as we considered spatial scales from small areas (0.002 km 2 ) confined around camera traps to larger grid cells (1.7 km 2 ): larger spatial scales capture more ecological heterogeneity. This is evident in our models by the widening confidence intervals for forest structure predictors of species diversity as spatial scale increases, particularly at the 1.7 km 2 scale (Fig. 2 ). While this trend aligns with previous findings (Yao et al., 2023 ), the confidence intervals at the 1.7 km 2 scale are considerably wider than those at the forest type partition scale, despite their comparable spatial resolutions. This is further supported by adjusted coefficient of variation (CV) values, which are higher at the 1.7 km 2 scale than at the partition scale for 12 of 15 structure metrics (Fig. 4 ). Additionally, cells in the grid systems contained up to seven different forest type partitions (Fig. S3), indicating that the 1.7 km 2 grid cells contained substantially more heterogeneity than the forest type partitions in our study area. These findings suggest that partitions offered a more ecologically coherent scale for analysis. Grid systems that fail to account for important ecological or environmental boundaries may thus group survey locations inappropriately and impede our ability to understand species diversity patterns based on structural metrics. The use of grid systems yielded model results that differed markedly from those derived from more ecologically informed scales. Grid system model results show three metrics are reliable predictors of diversity that are not reliable predictors in the camera trap or partition models. Partition scale models also consistently outperformed grid scale models. If differences among forest types were not ecologically salient to mammals, we would expect to find no difference in model outcomes between the partition and comparable grid scales. A possible explanation for this observation is the strong scale dependency of animal-habitat associations and use (Johnson, 1980 ). Specifically, the diversity metrics we assess at community levels may reflect emergent properties derived from the aggregation of individual behavioral decisions made across different scales. While our study focuses on diversity metrics at broader scales (i.e., forest type partitions), these patterns may ultimately be shaped by fine-scale habitat use decisions made by individuals. This connection would further underscore the importance of considering ecologically defined scales in analyses such as ours, as these scales are likely more aligned with the decisions animals make than are arbitrary grid systems. We predicted that forest structure characteristics would be positively associated with diversity metrics, following the habitat heterogeneity hypothesis (MacArthur & MacArthur, 1961 ). Our findings support this prediction at the forest type partition scale, where all reliable predictors (rumple index, vegetation volume, and vertical point distribution) are positively associated with diversity metrics. However, at the camera trap scale, two of the three reliable predictors (tree density and basal area) show a negative association with diversity metrics, while rumple index is positively correlated with species richness. These results align with previous research indicating that negative effects of environmental heterogeneity are more likely to be observed at smaller spatial scales compared to larger ones (Stein et al., 2014 ). However, in the two grid scale models, mean tree height is negatively associated with all diversity metrics. Due to the issues with grid scales discussed earlier, we interpret the results of these models with caution. At the camera trap scale, only one structure metric - rumple index of roughness - was a positive predictor, with rumple index also emerging as a positive predictor of diversity metrics at the partition scale. Rumple index describes forest canopy surface heterogeneity (roughness) by calculating the ratio between canopy surface area and its projected area on the ground, specifically capturing forest canopy layer heterogeneity. This contrasts with our prediction that structural metrics describing canopy complexity would not be reliable predictors of terrestrial mammal diversity. However, if canopy heterogeneity is associated with overall forest complexity, as previously suggested (Kane et al., 2010 ), then rumple index may serve as a proxy for structural complexity across forest strata, explaining its positive association with diversity. Alternatively, the terrestrial mammals we consider here may have ecological requirements associated with a heterogenous forest canopy, such as access to a higher diversity of fruiting trees. Complex canopy structures in tropical forests support a wide range of epiphytes and invertebrates (Nadkarni, 1994 ) which may directly or indirectly enhance the diets of terrestrial mammals, thereby contributing to their species richness. Additionally, a structurally diverse canopy may foster a more complex understory in some dimension that we did not capture with our structure metrics, such as understory light availability (Stark et al., 2015 ), creating richer habitats that support a greater variety of terrestrial species. Two more positive predictors of diversity metrics appear in the partition scale models - vegetation volume and vertical point distribution. Vegetation volume, measured here as the ratio of vegetated voxels (3D pixels) to all voxels in a point cloud, and vertical point distribution, with higher values indicating an even vertical distribution of vegetation, both suggest increased forest complexity through enhanced microhabitat availability and surface area. This added niche space likely promotes species richness and diversity (Holt, 2009 ). These findings mirror previous research from Tanzania linking habitat surface area ratio with mammal occupancy (Gorczynski et al., 2023 ). Contrary to our expectations based on the habitat heterogeneity hypothesis, tree density and basal area at the camera trap scale were negative predictors of two diversity metrics. While the negative effect of these metrics may be in line with negative complexity-heterogeneity relationships found at small spatial scales (Stein et al., 2014 ), it is also possible that these relatively simple metrics are not effectively capturing forest structural complexity as well as other metrics considered here. For instance, areas with higher tree densities may still lack critical microhabitats and open niches that certain species require for foraging or nesting (Wilson & Cresswell, 2006 ). Further, this result may reflect habitat preferences of terrestrial mammals that use more open forest areas with lower tree densities for travel, reflecting a higher likelihood of being captured by cameras placed in more open areas and not an ecological requirement, per se (Slater et al., 2023 ). This result may also reflect camera trap sampling bias, with cameras in more open areas capturing more mammals than cameras in areas with higher tree density, due to an increase in camera viewable area (Hofmeester et al., 2017 ), leading to a negative association with density metrics. Our study period spans more than five years, encompassing both non-mast periods and one full masting cycle, a key feature of Southeast Asian tropical rainforests that occurs at irregular intervals of 2–10 years (Appanah, 1993 ). During masting events, numerous plant species from different families simultaneously flower and fruit in a 3–6 month span, resulting in a temporary period of resource abundance, followed by resource scarcity during the intervening non-mast periods (Ashton et al., 1988 ). These fluctuations significantly impact mammal populations, particularly seed predators, which experience population booms during masting and subsequent declines afterward (Curran & Leighton, 2000 ). Although our models do not explicitly account for mast events or forest productivity, the inclusion of both mast and non-mast periods in our study ensures that estimates of mammal diversity metrics are not biased, as they might be if only one type of period was considered. Using TLS-derived point clouds we extracted several useful structural metrics for the present analysis. Traditional metrics such as mean tree height, mean dbh, and tree density, serve as fundamental descriptors of forest structure (Wilkes et al., 2016 ) and have been collected by foresters and ecologists for decades without the need for advanced methods like TLS surveys (e.g., Grubb et al., 1963 ). However, many of the structure metrics that we consider here describe forest structure in more detail, capturing complexity beyond the simpler metrics of tree height and canopy cover (Calders et al., 2020 ). Specifically, our analysis includes point cloud- and voxel-level metrics that capture intricate forest characteristics such as vertical point complexity, the ratio of cubic vegetation volume to empty space, and the volume of understory gaps. Importantly, we find that these advanced metrics are crucial predictors of mammal diversity and other biodiversity indices. Their collection would not have been feasible without the capabilities offered by TLS technology. Our findings emphasize the value of using ecologically informed scales when modeling mammal diversity. These scales capture meaningful ecological variation, leading to more reliable predictions compared to randomly placed grids. Grid scales, by contrast, introduce artificial boundaries that often combine ecologically distinct areas, reducing the precision of biodiversity analyses. As our results show, scales aligned with ecological boundaries not only outperform grid scales but also provide more reliable insights into the relationships between biodiversity and forest structure. Research across a diversity of ecosystems would likely benefit from prioritizing the use of ecologically relevant spatial scales. METHODS Study site We conducted our research at the Cabang Panti Research Site (CPRS; 1°13′ S, 110°7′ E), a tropical rainforest site located in Gunung Palung National Park, West Kalimantan, Indonesia. CPRS covers an area of approximately 26 km 2 and contains seven distinct and contiguous forest types that vary in a number of physical and ecological characteristics including soil type, elevation, plant and animal species composition, and forest structure (Marshall, 2009 ; Marshall et al., 2014 , 2021 ). The distinct forest types are peat swamp forest (5–10 m asl), freshwater swamp forest (5–10 m asl), alluvial bench forest (5-100 m asl), lowland sandstone forest (20–200 m asl), lowland granite forest (200–400 m asl), upland granite forest (350–800 m asl), and montane forest (750-1,100 m asl). Forest types at CPRS can be further subdivided into spatially distinct forest type partitions (e.g., when they are on different mountain ridges), allowing for analyses at finer spatial resolution. Partitions differ in environmental conditions including rainfall, temperature, and elevation, and thus are likely to vary in structure as well. Forest types have been split into two partitions based on separation of forest types by mountain ridges and a river (lowland sandstone, lowland granite, upland granite, and montane forests) or by discontinuity (alluvial bench forests). Two forest types, peat swamp and freshwater swamp are not subdivided, yielding a total of 12 partitions (Fig. 1 ). Forest structure measurement We conducted LiDAR scan surveys across CPRS between 1 February and 24 February 2023 using a terrestrial laser scanner (TLS). We chose LiDAR survey locations based on previously established camera trap locations (see below). We used a FARO Focus Laser Scanner s350 (Faro Technologies Inc., Lake Mary, USA) to survey camera trap locations and create 3D point clouds of the forest structure, using scan settings of 1/8 resolution and 3x quality, following FARO guidelines for outdoor scanning. We placed the scanner on a tripod approximately 1.4 m above ground. We scanned 58 locations across all seven forest types and 11 of the 12 partitions at CPRS – six in peat swamp forests, ten in freshwater swamp and alluvial bench forests, nine in lowland sandstone and upland granite forests, eleven in lowland granite forests, and three in montane forests. We collected multiple scans at each location to ensure adequate 3D coverage, with an average of 22 scans per location (range 13–33 scans). We created scan areas by conducting scans 1 m − 15 m from trees on which camera traps were placed, taking multiple scans approximately 1 m − 2 m apart within this scan radius. We made abundant use of targets (checkerboards and mounted spheres; Fig. S6) in the scan areas to assist with scan registration, a process that allows multiple scans taken at one location to be merged into a single 3D point cloud (Fig. S7). Up to 25 0.25 m X 0.2 m checkerboards were affixed to trees throughout the scan area, and up to five white spheres with a diameter of 0.2 m were mounted on tripods at heights adjusted to be visible to the scanner (e.g., above terrestrial vegetation) at approximately 1 m − 2 m. We conducted scan registration using FARO Scene software (v2023.1; Faro Technologies Inc., Florida, USA) and exported the resulting point clouds as LAS files for further processing in R Statistical Software (v4.2.1; R Core Team 2024). We extracted multiple forest structure metrics for each scanned location using the FORTLS (Molina-Valero et al., 2022 ), lidR (Roussel et al., 2020 ), and lidRmetrics (Tompalski, 2024 ) packages in R, utilizing functions that apply tree detection algorithms and calculate tree-, stand-, and cloud-level variables. To ensure consistency across sites, we standardized all point clouds to a uniform 25 m radius (0.002 km 2 ), the default in FORTLS , before extracting structure metrics. The metrics we consider here include: maximum tree height (m); mean tree height (m); canopy relief ratio, the ratio of mean tree height to maximum tree height; standard deviation of tree heights (m), used here to describe the level of forest stratification; mean diameter at breast height (dbh; cm), measured at height of 1.3 m; tree basal area (m 2 /ha); tree density (trees/ha), tree volume (m 3 /ha), number of points below 2 m, a proxy for the proportion of terrestrial vegetation; vertical point distribution; vertical complexity; leaf area density; rumple index of roughness, which describes the horizontal heterogeneity of the forest canopy surface; vegetation density, a 3D (voxelized) approach to quantifying the amount of cubic vegetated space in each point cloud; and closed gap space, describing the volume (m 3 ) of cubic space beneath the canopy classified as vegetation gaps. Mammal surveys and diversity metrics We monitored mammal communities at CPRS between June 2015 and September 2020 using motion-triggered video cameras (Bushnell TrophyCam HD; Overland Park, KS). This ensured that we monitored species presence throughout a dipterocarp mast cycle that drives the distribution and abundances of mammals in Bornean rainforest systems (e.g., Curran & Leighton, 2000 ; Marshall et al., 2014 ; Marshall et al., 2021 ). We placed cameras at 177 sites (M = 260.1 days per camera, SD = 162.8, range 23–906) across all seven forest types at CPRS, along both established trails and off-trail placement sites (n = 46,044 trap nights). We mounted cameras on trees, placed at approximately 0.4 m from the ground, and programmed them to capture 20 s videos when triggered. We identified 55 non-human mammal species and assigned a terrestriality designation (terrestrial or arboreal) to each based on available data from the PanTHERIA database (Jones et al., 2009 ) and the Phillips’ Field Guide to the Mammals of Borneo text (Phillipps, 2016 ). Here, we only include terrestrial mammal observations, yielding 36 terrestrial mammals for subsequent analyses. This includes observations of rats (Family Muridae) and shrews (Family Soricidae) that we did not identify to species and so count as one species for diversity metrics. Rats, but not shrews, were excluded in the calculation of functional diversity metrics. We included shrews in functional diversity calculations since it is likely only two possible species range through GPNP and are closely related in ecology. We therefore averaged quantitative ecological traits between the two species to characterize Family Soricidae. We calculated six diversity metrics for each camera trap location and at larger spatial scales (see Spatial scales ). Using the vegan package in R (Oksanen et al., 2022 ) we calculated species richness, Shannon diversity index, and Pielou's species evenness values. Additionally, we used the FD package (Laliberté et al., 2014 ) to calculate functional richness, functional evenness, and functional divergence for 35 species (excluding rats) using six ecological traits (body mass, activity cycle, trophic level, diet breadth, group size, and litter size; see Table S1). Modeling approach To estimate the influence of forest structure on the mammal diversity metrics described above we fit generalized linear models (GLM) in R. All models used structural metrics as predictors and were specified using a gamma distribution. We include the number of days each camera trap location was operational as an offset in all models due to uneven survey effort across locations. We applied weights in models at all scales except the camera trap scale to account for variation in the number of scanned locations within each spatial subunit (i.e., grid cells and forest type partitions). We assessed all quantitative predictors for correlation before fitting models, excluding highly correlated pairs (r > 0.8) from the same model. We found significant spatial autocorrelation at the camera trap scale for all diversity metrics except functional richness and functional evenness, using Moran’s I. To account for this, we applied eigenvector-based spatial filtering by including Moran’s eigenvectors as predictors in the models with significant spatial autocorrelation. We used the ‘dredge’ function in the MuMin package (Bartoń, 2023) to fit candidate models, specifying models at all scales to include a maximum of two predictors, not including eigenvectors, to prevent overfitting and to ensure model complexity was the same across models at different spatial scales. We then used model averaging across all candidate models to assess the influence of all forest structure predictors. We determined predictors to be reliable indicators of outcome variables if the 95% confidence intervals (CI’s) of the beta coefficient estimate did not overlap zero. Spatial scales To assess the influence of structure metrics at different spatial scales, we fit all models at four spatial scales. In addition to the camera trap level scale, we created two grid systems that overlay the study area to aggregate locations for which we collected LiDAR (i.e., scanned locations) and camera trap data. We ensured each non-empty grid cell contained at least two scanned locations. One scale is a 6 X 4 grid containing 0.75 km 2 cells, creating 17 groups of scanned locations, with seven cells not containing any scanned locations. The other is 4 X 3 grid containing 1.7 km 2 cells, creating 10 groups of scanned locations. We also fit models using the forest type partition scale, which is comparable to the 1.7 km 2 grid scale in both area of sub-units (x̄ > 1.50 km 2 ) and groups of scanned locations (n = 11). Additionally, we grouped scanned locations in two larger spatial scales to assess variation in structure metrics across multiple scales. These additional grid scales include a 2 X 2 grid containing 6.5 km 2 cells and a scale comprising the entire study area, approximately 26 km 2 and were not used for modeling purposes. To assess the variation captured by spatial scales that aggregate TLS survey locations (i.e., forest type partition and grid scales), we calculated the adjusted coefficient of variation (CV) for all structure metrics and diversity metrics. The adjusted CV accounts for unequal sample sizes across subunits. For each metric at each scale, we first calculated the CV for each spatial subunit (partitions or individual grid cells), then averaged these values to obtain the mean CV for each scale. Limitations Since we were not able to identify observations of rats (Family Muridae) to the species level, we considered all observations as one species when calculating richness, diversity, and evenness metrics. We were not able to consider rats at all for functional diversity metrics. Consequently, we surely underestimated diversity metrics, given rats were the most frequently observed mammal group (n = 2,788 camera trap observations) and, based on species range maps and descriptions (Phillips, 2016), there are potentially up to 11 rat species at GPNP. These underestimates are likely most pronounced in the Lowland Granite and Upland Granite forest types, where most rat observations were made. Indeed, previous research involving small mammal trapping at GPNP identified five Muridae species, with the majority of observations and species occurring in low- and mid-elevation forest types (Gorog, 2003 ). Due to the limitations of TLS surveying, we may have underestimated structure metrics related to higher forest strata, such as maximum tree height, at several locations, as tree crowns are often out of the instrument’s direct line of sight and thus not captured in the resulting point clouds. This likely affected our ability to capture the crowns of the tallest trees, including emergents, which play an important structural and ecological role for many tropical animals (Cannon & Leighton, 1994 ). We expect this limitation to be most pronounced in the alluvial bench and lowland sandstone forests, where emergents of the Dipterocarpaceae family are most numerous and account for a significant portion of aboveground biomass in Borneo (Aiba et al., 2013 ). For the present study we were restricted in what forest structure metrics were available to be used for analyses by the relatively limited number of R packages and standalone software that currently exists to automate the point cloud analysis processes, and by the selection of structural metrics that are estimated by these packages and software. For example, while dozens of R packages exist that analyze LiDAR data for forest research purposes, according to a recent review of LiDAR-focused R packages only two are currently available that were designed to analyze TLS data specifically (Atkins et al., 2022 ). Standalone software packages are also limited in their functionality in analyzing TLS data, especially point cloud data from tropical rainforests which are highly structurally complex, as many software options compute metrics for single-tree point cloud files only. Furthermore, to our knowledge, no R package or software is currently capable of detecting and providing descriptive metrics of lianas or vines from TLS point clouds from tropical forests. Lianas and vines are important structural components of tropical forests for arboreal mammals (Arroyo-Rodríguez et al., 2015 ), and can impact forest biodiversity and dynamics in important ways (Schnitzer, 2015 ). Declarations Acknowledgements We would like to thank the Indonesian Ministry of Higher Education and Research and Technology and the Gunung Palung National Park Bureau for supporting research at the Cabang Panti Research Site. Our work was supported by funding from the University of Michigan, Victoria University of Wellington, the National Science Foundation (award no. 2216525), the Leakey Foundation, the Orangutan Conservancy, the Mohamed bin Zayed Species Conservation Fund, the AZA Ape TAG Initiative, the American Society of Primatologists, and the Lewis and Clark Fund Grant from the American Philosophical Society. The authors have no conflicts of interest to report. Author contributions GRE and AJM designed the study; GRE, HUW, and AJM raised funds for data collection; GRE, ES, HUW, and AJM collected the data; GRE wrote all analysis R code and produced the figures; GRE and AJM wrote the manuscript. All authors contributed to the drafts and gave final approval for publication. Data and code are available at: https://github.com/Estrada-Gene/LiDAR_Forest_structure References Aiba, S., Akutsu, K., Onoda, Y.: Canopy structure of tropical and sub-tropical rain forests in relation to conifer dominance analysed with a portable LIDAR system. Ann. Botany. 112 (9), 1899–1909 (2013). https://doi.org/10.1093/aob/mct242 Appanah, S.: Mass flowering of dipterocarp forests in the aseasonal tropics. J. Biosci. 18 , 457–474 (1993) Arroyo-Rodríguez, V., Asensio, N., Dunn, J.C., Cristóbal-Azkarate, J., Gonzalez-Zamora, A.: Use of lianas by primates: More than a food source. In: Ecology of Lianas, pp. 407–426. Wiley, Ltd (2015). https://doi.org/10.1002/9781118392409.ch27 Ashton, P.S., Givnish, T.J., Appanah, S.: Staggered Flowering in the Dipterocarpaceae: New Insights Into Floral Induction and the Evolution of Mast Fruiting in the Aseasonal Tropics. Am. Nat. 132 (1), 44–66 (1988). https://doi.org/10.1086/284837 Atkins, J.W., Stovall, A.E.L., Alberto Silva, C.: Open-Source tools in R for forestry and forest ecology. For. Ecol. Manag. 503 , 119813 (2022). https://doi.org/10.1016/j.foreco.2021.119813 Bartoń, K.: _MuMIn: Multi-Model Inference_. R package version 1.48.4, (2024). https://CRAN.R-project.org/package=MuMIn Bradshaw, C.J., Sodhi, N.S., Brook, B.W.: Tropical turmoil: A biodiversity tragedy in progress. Front. Ecol. Environ. 7 (2), 79–87 (2009). https://doi.org/10.1890/070193 Calders, K., Adams, J., Armston, J., Bartholomeus, H., Bauwens, S., Bentley, L.P., Chave, J., Danson, F.M., Demol, M., Disney, M., Gaulton, R., Moorthy, K., Levick, S.M., Saarinen, S.R., Schaaf, N., Stovall, C., Terryn, A., Wilkes, L., P., Verbeeck, H.: Terrestrial laser scanning in forest ecology: Expanding the horizon. Remote Sens. Environ. 251 , 112102 (2020). https://doi.org/10.1016/j.rse.2020.112102 Cannon, C.H., Leighton, M.: Comparative locomotor ecology of gibbons and macaques: Selection of canopy elements for crossing gaps. Am. J. Phys. Anthropol. 93 (4), 505–524 (1994). https://doi.org/10.1002/ajpa.1330930409 Chase, J.M., Leibold, M.A.: Ecological Niches: Linking Classical and Contemporary Approaches. University of Chicago Press (2003) Curran, L.M., Leighton, M.: Vertebrate Responses to Spatiotemporal Variation in Seed Production of Mast-Fruiting Dipterocarpaceae. Ecol. Monogr. 70 (1), 101–128 (2000). https://doi.org/10.1890/0012-9615(2000 )070[0101:VRTSVI]2.0.CO;2 Currie, D.J.: Energy and Large-Scale Patterns of Animal- and Plant-Species Richness. Am. Nat. 137 (1), 27–49 (1991). https://doi.org/10.1086/285144 Drake, J.B., Dubayah, R.O., Clark, D.B., Knox, R.G., Blair, J.B., Hofton, M.A., Chazdon, R.L., Weishampel, J.F., Prince, S.: Estimation of tropical forest structural characteristics using large-footprint lidar. Remote Sens. Environ. 79 (2), 305–319 (2002). https://doi.org/10.1016/S0034-4257(01)00281-4 Ehbrecht, M., Seidel, D., Annighöfer, P., Kreft, H., Köhler, M., Zemp, D.C., Puettmann, K., Nilus, R., Babweteera, F., Willim, K., Stiers, M., Soto, D., Boehmer, H.J., Fisichelli, N., Burnett, M., Juday, G., Stephens, S.L., Ammer, C.: Global patterns and climatic controls of forest structural complexity. Nat. Commun. 12 (1) (2021). Article 1. https://doi.org/10.1038/s41467-020-20767-z Gorczynski, D., Rovero, F., Mtui, A., Shinyambala, S., Martine, J., Hsieh, C., Frishkoff, L., Beaudrot, L.: Tropical forest mammal occupancy and functional diversity increase with microhabitat surface area. Ecology. 104 (12), e4181 (2023). https://doi.org/10.1002/ecy.4181 Gorog, A.J.: Historical biogeography of small mammals in the Sunda region of Southeast Asia [Ph.D., University of Michigan]. (2003). https://www.proquest.com/docview/287902251/abstract/20FCD055C4DE48D0PQ/1 Grubb, P.J., Lloyd, J.R., Pennington, T.D., Whitmore, T.C.: A Comparison of Montane and Lowland Rain Forest in Ecuador I. The Forest Structure, Physiognomy, and Floristics. J. Ecol. 51 (3), 567–601 (1963). https://doi.org/10.2307/2257748 Hawkins, B.A., Field, R., Cornell, H.V., Currie, D.J., Guégan, J.-F., Kaufman, D.M., Kerr, J.T., Mittelbach, G.G., Oberdorff, T., O’Brien, E.M., Porter, E.E., Turner, J.R.G.: Energy, Water, and Broad-Scale Geographic Patterns of Species Richness. Ecology. 84 (12), 3105–3117 (2003). https://doi.org/10.1890/03-8006 Hazell, R.J., Chmel, K., Riegert, J., Paul, L., Isua, B., Kaina, G.S., Fibich, P., Molem, K., Stewart, A.J.A., Peck, M.R., Weiblen, G.D., Novotny, V.: Spatial scaling of plant and bird diversity from 50 to 10,000 ha in a lowland tropical rainforest. Oecologia. 196 (1), 101–113 (2021). https://doi.org/10.1007/s00442-021-04925-8 Hofmeester, T.R., Rowcliffe, J.M., Jansen, P.A.: A simple method for estimating the effective detection distance of camera traps. Remote Sens. Ecol. Conserv. 3 (2), 81–89 (2017). https://doi.org/10.1002/rse2.25 Holt, R.D.: Bringing the Hutchinsonian niche into the 21st century: Ecological and evolutionary perspectives. Proceedings of the National Academy of Sciences, 106(supplement_2), 19659–19665. (2009). https://doi.org/10.1073/pnas.0905137106 Ioki, K., Tsuyuki, S., Hirata, Y., Phua, M.-H., Wong, W.V.C., Ling, Z.-Y., Saito, H., Takao, G.: Estimating above-ground biomass of tropical rainforest of different degradation levels in Northern Borneo using airborne LiDAR. For. Ecol. Manag. 328 , 335–341 (2014). https://doi.org/10.1016/j.foreco.2014.06.003 Johnson, D.H.: The comparison of usage and availability measurements for evaluating resource preference. Ecology. 61 (1), 65–71 (1980) Jones, K.E., Bielby, J., Cardillo, M., Fritz, S.A., O’Dell, J., Orme, C.D.L., Safi, K., Sechrest, W., Boakes, E.H., Carbone, C., Connolly, C., Cutts, M.J., Foster, J.K., Grenyer, R., Habib, M., Plaster, C.A., Price, S.A., Rigby, E.A., Rist, J., Purvis, A.: PanTHERIA: A species-level database of life history, ecology, and geography of extant and recently extinct mammals. Ecology. 90 (9), 2648–2648 (2009). https://doi.org/10.1890/08-1494.1 Kane, V.R., McGaughey, R.J., Bakker, J.D., Gersonde, R.F., Lutz, J.A., Franklin, J.F.: Comparisons between field- and LiDAR-based measures of stand structural complexity. Can. J. For. Res. 40 (4), 761–773 (2010). https://doi.org/10.1139/X10-024 Kissling, W.D., Field, R., Korntheuer, H., Heyder, U., Böhning-Gaese, K.: Woody plants and the prediction of climate-change impacts on bird diversity. Philosophical Trans. Royal Soc. B: Biol. Sci. 365 (1549), 2035–2045 (2010). https://doi.org/10.1098/rstb.2010.0008 Laliberté, E., Legendre, P., Shipley, B.: FD: measuring functional diversity from multiple traits, and other tools for functional ecology. R package version 1.0-12.3. (2014) MacArthur, R.H., MacArthur, J.W.: Bird. Species Divers. Ecol. 42 (3), 594–598 (1961). https://doi.org/10.2307/1932254 Marshall, A.J.: Gibbons Hylobates albibarbis? Biotropica. 41 (2), 257–267 (2009). https://doi.org/10.1111/j.1744-7429.2008.00461.x Are Montane Forests Demographic Sinks for Bornean White-bearded Marshall, A.J., Beaudrot, L., Wittmer, H.U.: Responses of Primates and Other Frugivorous Vertebrates to Plant Resource Variability over Space and Time at Gunung Palung National Park. Int. J. Primatol. 35 (6), 1178–1201 (2014). https://doi.org/10.1007/s10764-014-9774-4 Marshall, A.J., Farr, M.T., Beaudrot, L., Zipkin, E.F., Feilen, K.L., Bell, L.G., Setiawan, E., Susanto, T.W., Setia, M., Leighton, T., M., Wittmer, H.U.: Biotic and abiotic drivers of dispersion dynamics in a large-bodied tropical vertebrate, the Western Bornean orangutan. Oecologia. 196 (3), 707–721 (2021). https://doi.org/10.1007/s00442-021-04964-1 McCain, C.M.: Elevational Gradients in Diversity of Small Mammals. Ecology. 86 (2), 366–372 (2005). https://doi.org/10.1890/03-3147 Molina-Valero, J.A., Martínez-Calvo, A., Ginzo Villamayor, M.J., Novo Pérez, M.A., Álvarez-González, J.G., Montes, F., Pérez-Cruzado, C.: Operationalizing the use of TLS in forest inventories: The R package FORTLS. Environ. Model. Softw. 150 , 105337 (2022). https://doi.org/10.1016/j.envsoft.2022.105337 Nadkarni, N.M.: Diversity of Species and Interactions in the Upper Tree Canopy of Forest Ecosystems1. Am. Zool. 34 (1), 70–78 (1994). https://doi.org/10.1093/icb/34.1.70 Oksanen, J., Simpson, G., Blanchet, F., Kindt, R., Legendre, P., Minchin, P., O'Hara, R., Solymos, P., Stevens, M., Szoecs, E., Wagner, H., Barbour, M., Bedward, M., Bolker, B., Borcard, D., Carvalho, G., Chirico, M., De Caceres, M., Durand, S., Evangelista, H., FitzJohn, R., Friendly, M., Furneaux, B., Hannigan, G., Hill, M., Lahti, L., McGlinn, D., Ouellette, M., Ribeiro Cunha, E., Smith, T., Stier, A., Ter Braak, C., Weedon, J.: _vegan: Community Ecology Package_. R package version 2.6-4,. (2022) Palace, M., Sullivan, F.B., Ducey, M., Herrick, C.: Estimating Tropical Forest Structure Using a Terrestrial Lidar. PLOS ONE. 11 (4), e0154115 (2016). https://doi.org/10.1371/journal.pone.0154115 Penone, C., Allan, E., Soliveres, S., Felipe-Lucia, M.R., Gossner, M.M., Seibold, S., Simons, N.K., Schall, P., van der Plas, F., Manning, P., Manzanedo, R.D., Boch, S., Prati, D., Ammer, C., Bauhus, J., Buscot, F., Ehbrecht, M., Goldmann, K., Jung, K., Fischer, M.: Specialisation and diversity of multiple trophic groups are promoted by different forest features. Ecol. Lett. 22 (1), 170–180 (2019). https://doi.org/10.1111/ele.13182 Phillipps, Q.: Phillipps’ Field Guide to the Mammals of Borneo and Their Ecology: Sabah, Sarawak, Brunei, and Kalimantan. Princeton University Press (2016) Pillay, R., Venter, M., Aragon-Osejo, J., González-del-Pliego, P., Hansen, A.J., Watson, J.E., Venter, O.: Tropical forests are home to over half of the world’s vertebrate species. Front. Ecol. Environ. 20 (1), 10–15 (2022). https://doi.org/10.1002/fee.2420 Püttker, T., Pardini, R., Meyer-Lucht, Y., Sommer, S.: Responses of five small mammal species to micro-scale variations in vegetation structure in secondary Atlantic Forest remnants, Brazil. BMC Ecol. 8 (1), 9 (2008). https://doi.org/10.1186/1472-6785-8-9 R Core Team: _R: A Language and Environment for Statistical Computing_. R Foundation for Statistical Computing, Vienna, Austria. (2023). https://www.R-project.org/ Roussel, J.R., Auty, D., Coops, N.C., Tompalski, P., Goodbody, T.R.H., Sánchez Meador, A., Bourdon, J.F., De Boissieu, F., Achim, A.: lidR : An R package for analysis of Airborne Laser Scanning (ALS) data. Remote Sensing of Environment, 251 (August), 112061. <>10.1016/j.rse.2020.112061>. (2020) Schnitzer, S.A.: The Contribution of Lianas to Forest Ecology, Diversity, and Dynamics. In N. Parthasarathy (Ed.), Biodiversity of Lianas (pp. 149–160). Springer International Publishing. (2015). https://doi.org/10.1007/978-3-319-14592-1_9 Slater, H.D., Gillingham, P.K., Pratt, V., Eaton, B., Fletcher, S., Abdullah, A., Supriadi, Korstjens, A.H.: Living on the edge: Forest edge effects on microclimate and terrestrial mammal activity in disturbed lowland forest in Sumatra, Indonesia. Oryx: The International Journal of Conservation, 58(2), Article 2. (2023) Stark, S.C., Enquist, B.J., Saleska, S.R., Leitold, V., Schietti, J., Longo, M., Alves, L.F., Camargo, P.B., Oliveira, R.C.: Linking canopy leaf area and light environments with tree size distributions to explain Amazon forest demography. Ecol. Lett. 18 (7), 636–645 (2015). https://doi.org/10.1111/ele.12440 Stark, S.C., Leitold, V., Wu, J.L., Hunter, M.O., de Castilho, C.V., Costa, F.R.C., McMahon, S.M., Parker, G.G., Shimabukuro, M.T., Lefsky, M.A., Keller, M., Alves, L.F., Schietti, J., Shimabukuro, Y.E., Brandão, D.O., Woodcock, T.K., Higuchi, N., de Camargo, P.B., de Oliveira, R.C., Saleska, S.R.: Amazon forest carbon dynamics predicted by profiles of canopy leaf area and light environment. Ecol. Lett. 15 (12), 1406–1414 (2012). https://doi.org/10.1111/j.1461-0248.2012.01864.x Stein, A., Gerstner, K., Kreft, H.: Environmental heterogeneity as a universal driver of species richness across taxa, biomes and spatial scales. Ecol. Lett. 17 (7), 866–880 (2014). https://doi.org/10.1111/ele.12277 Tang, H., Dubayah, R., Swatantran, A., Hofton, M., Sheldon, S., Clark, D.B., Blair, B.: Retrieval of vertical LAI profiles over tropical rain forests using waveform lidar at La Selva, Costa Rica. Remote Sens. Environ. 124 , 242–250 (2012). https://doi.org/10.1016/j.rse.2012.05.005 Tews, J., Brose, U., Grimm, V., Tielbörger, K., Wichmann, M.C., Schwager, M., Jeltsch, F.: Animal species diversity driven by habitat heterogeneity/diversity: The importance of keystone structures. J. Biogeogr. 31 (1), 79–92 (2004). https://doi.org/10.1046/j.0305-0270.2003.00994.x Tompalski, P.: _lidRmetrics: Point cloud metrics for the lidR package._. R package version 0.1.3, (2024). https://ptompalski.github.io/lidRmetrics/ White, J.C., Coops, N.C., Wulder, M.A., Vastaranta, M., Hilker, T., Tompalski, P.: Remote Sensing Technologies for Enhancing Forest Inventories: A Review. Can. J. Remote. Sens. 42 (5), 619–641 (2016). https://doi.org/10.1080/07038992.2016.1207484 Wiens, J.J.: The niche, biogeography and species interactions. Philosophical Trans. Royal Soc. B: Biol. Sci. (2011). https://doi.org/10.1098/rstb.2011.0059 Wilkes, P., Jones, S.D., Suarez, L., Haywood, A., Mellor, A., Woodgate, W., Soto-Berelov, M., Skidmore, A.K.: Using discrete-return airborne laser scanning to quantify number of canopy strata across diverse forest types. Methods Ecol. Evol. 7 (6), 700–712 (2016). https://doi.org/10.1111/2041-210X.12510 Wilson, J.M., Cresswell, W.: How robust are Palearctic migrants to habitat loss and degradation in the Sahel? Ibis. 148 (4), 789–800 (2006). https://doi.org/10.1111/j.1474-919X.2006.00581.x Yao, Z., Yang, X., Wang, B., Shao, X., Wen, H., Deng, Y., Zhang, Z., Cao, M., Lin, L.: Multidimensional beta-diversity across local and regional scales in a Chinese subtropical forest: The role of forest structure. Ecol. Evol. 13 (10), e10607 (2023). https://doi.org/10.1002/ece3.10607 Table 1 Table 1 Reliable predictors of diversity metrics. Eight forest structure metrics are reliable predictors of five different diversity metrics: species richness, Shannon diversity, species evenness, functional evenness, and functional divergence. Functional richness models are excluded due to model convergence issues. The diversity metrics were modeled across four spatial scales. The table indicates the models where each structure metric was a reliable predictor and the direction of its effect. SCALE Structure metric Camera trap (0.002 km 2 ) Forest type partition (~ 1.5 km 2 ) 0.75 km 2 grid 1.7 km 2 grid Tree density Evenness (-) Func. Even. (-) Func. Diver. (-) Basal area Func. Even. (-) Func. Diver. (-) Rumple index Richness (+) All models (+) Mean tree height All models (-) All models (-) Mean tree dbh All models (+) Vertical point complexity All models (+) Vegetation volume Richness (+) Func. Even. (+) Vertical point dist. All models (+) Additional Declarations There is NO Competing Interest. Supplementary Files SUPPLEMENTARY.docx Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5479224","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":386270165,"identity":"6078ca7e-b6f2-4529-a916-47149f8a8ee9","order_by":0,"name":"Gene Estrada","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA0klEQVRIiWNgGAWjYNACNgYGfiB1gLEBRBJUzgzRItlAshYDkEqitPDznz/24UeZnd3mG9mJB37uYJDju5GAX4vkjGTmmT3nkpO33cjdcLD3DIOxJCEtBjeYmRl425iTzYBaDjO2MSRuIKTF/vxhZsa/bfXJxjMgWuoJajFgSGZm5m07bGcgAdGSYEBIi8SNZGNmmXPHEyTOvAX6pU3CcOaZB/i18PcffMz4pqzanr89d/OHn2028nzHCdgCA4kNUFuJUw4C9sQrHQWjYBSMghEHAHZrScsYA/WuAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-0212-7430","institution":"University of Michigan","correspondingAuthor":true,"prefix":"","firstName":"Gene","middleName":"","lastName":"Estrada","suffix":""},{"id":386270166,"identity":"6dc82b81-800b-488c-818d-958832367943","order_by":1,"name":"Heiko Wittmer","email":"","orcid":"https://orcid.org/0000-0002-8861-188X","institution":"Victoria University of Wellington","correspondingAuthor":false,"prefix":"","firstName":"Heiko","middleName":"","lastName":"Wittmer","suffix":""},{"id":386270167,"identity":"d035eedc-cc08-4d54-98bc-e5525dcf4ba2","order_by":2,"name":"Endro Setiawan","email":"","orcid":"","institution":"Gunung Palung National Park Office","correspondingAuthor":false,"prefix":"","firstName":"Endro","middleName":"","lastName":"Setiawan","suffix":""},{"id":386270168,"identity":"35b0bb02-eae5-45bb-b8d5-0ebc87d5c751","order_by":3,"name":"Andrew Marshall","email":"","orcid":"https://orcid.org/0000-0002-7703-8811","institution":"University of Michigan-Ann Arbor","correspondingAuthor":false,"prefix":"","firstName":"Andrew","middleName":"","lastName":"Marshall","suffix":""}],"badges":[],"createdAt":"2024-11-19 01:40:13","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5479224/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5479224/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":78826813,"identity":"e39ebe8e-fe13-478d-96df-ee5fd7158ba2","added_by":"auto","created_at":"2025-03-19 12:43:17","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":326741,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCabang Panti Research Site map\u003c/strong\u003e. CPRS is located on the island of Borneo in West Kalimantan, Indonesia and has an extensive trail system that spans 12 unique forest type partitions. Points show camera trap locations. White points show camera trap locations at which LiDAR surveys were conducted. The small gray block adjacent to Alluvial bench II is a small patch of heath (\u003cem\u003ekerangas) \u003c/em\u003eforest, that was not sampled.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-5479224/v1/c5d767832eb94613b7a222a6.png"},{"id":78827175,"identity":"bf7055e9-b984-443d-b8e1-6dd84b261cb2","added_by":"auto","created_at":"2025-03-19 12:51:18","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":427292,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eShannon diversity model results. \u003c/strong\u003eCoefficient plot showing estimates for all forest structure metrics at all spatial scales for Shannon diversity models. Models are grouped by ecologically defined scales (camera trap and forest type partition scales) and the two random grid scales. Predictors are grouped by the pattern of the direction of effects between the two scale groups – opposite, same, or mixed direction of effects between the ecologically defined and grid scale groups.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-5479224/v1/6faa27e361688ad6cc4ceb5d.png"},{"id":78826815,"identity":"c60cc46d-d1a1-45c9-8071-6f309119b6f8","added_by":"auto","created_at":"2025-03-19 12:43:18","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1561433,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eVariation of forest structure metrics across forest types. \u003c/strong\u003eStructure metrics of 58 LiDAR survey locations (points) grouped by forest type (box plots).\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-5479224/v1/9fc11b6a548db794e5917e72.png"},{"id":78826816,"identity":"8ef3c577-80bb-45b4-80fe-37260709adb9","added_by":"auto","created_at":"2025-03-19 12:43:18","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":608471,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eStructural metric variation increases with scale but not for forest type partitions. \u003c/strong\u003eMean adjusted coefficient of variation (CV), which controls for sample size, is compared across grid scales (cell size), shown with the black line. The green horizontal line shows the mean adjusted CV value for the forest type partition scale, comparable in resolution and number of sub-units to the 1.7 km\u003csup\u003e2\u003c/sup\u003e scale, which is lower for 12 of 15 structural metrics.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-5479224/v1/6200b65bbf7c547af929264e.png"},{"id":78827970,"identity":"7d48a9b7-3cc5-4047-8eff-6c9f1b3c1304","added_by":"auto","created_at":"2025-03-19 12:59:19","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3466934,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5479224/v1/7e9854dc-c239-47ba-ad1e-0b9e57a8e363.pdf"},{"id":78826818,"identity":"6c381388-8d63-42f1-9148-5fe90bf31294","added_by":"auto","created_at":"2025-03-19 12:43:18","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":4392513,"visible":true,"origin":"","legend":"","description":"","filename":"SUPPLEMENTARY.docx","url":"https://assets-eu.researchsquare.com/files/rs-5479224/v1/faa23a5c46782faebdef4c0d.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"LiDAR-derived forest structure metrics show ecologically defined scales outperform grids when predicting mammal diversity","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eUnderstanding the ecological factors that influence biodiversity is of considerable theoretical interest and has substantial value for conservation. Research on the predictors of animal diversity has spanned abiotic factors such as climate and temperature (Hawkins et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2003\u003c/span\u003e; Wiens, \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2011\u003c/span\u003e), precipitation (Currie, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e1991\u003c/span\u003e), sampling area (MacArthur \u0026amp; Wilson 1967), latitude (McCoy \u0026amp; Connor, 1980), elevation and topography (McCain, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2005\u003c/span\u003e), and biotic factors such as food availability (Kissling et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2010\u003c/span\u003e), species interactions (Chase \u0026amp; Leibold, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2003\u003c/span\u003e), and vegetation structure (Tews et al., \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2004\u003c/span\u003e). The last has emerged as a key determinant of land-dwelling species richness (e.g., mammals, Stein et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2014\u003c/span\u003e), providing support for the \u0026lsquo;habitat heterogeneity hypothesis\u0026rsquo;, which predicts that habitats with higher structural complexity provide more niche space and thus support higher species richness than habitats with lower structural complexity (MacArthur \u0026amp; MacArthur, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e1961\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eMost of the research investigating the influence of vegetation structure and habitat heterogeneity on species diversity has been conducted at broad spatial scales (\u0026gt;\u0026thinsp;1 km\u003csup\u003e2\u003c/sup\u003e; Stein et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Scale, however, has been identified as the single most important aspect for understanding both habitat associations and use for terrestrial species (Johnson, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e1980\u003c/span\u003e). Research into diversity patterns at broad scales thus risks overlooking important ecological variation occurring at smaller spatial scales. For example, variation in structural characteristics or vegetation such as canopy height, canopy cover, and vegetation density significantly impacts mammal presence at micro-scales (\u0026lt;\u0026thinsp;80 m\u003csup\u003e2\u003c/sup\u003e) in addition to larger scales (P\u0026uuml;ttker et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). Similar research found canopy cover predicts bird species diversity best at small scales (0.5 km\u003csup\u003e2\u003c/sup\u003e; Hazell et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). To better understand how vegetation structure influences biodiversity, it is crucial to analyze these effects at smaller, potentially more ecologically relevant scales, which may reveal patterns missed by broader analyses.\u003c/p\u003e \u003cp\u003eTropical forests are particularly relevant for such an analysis. They exhibit the highest structural complexity of all forests (Ehbrecht et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) and house an estimated 62% of the world\u0026rsquo;s vertebrate species (Pillay et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) while covering only 7% of Earth\u0026rsquo;s land surface (Bradshaw et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). Despite their global importance, these ecosystems remain among the least understood and most highly threatened. Investigating their structural complexity is crucial, as it plays a pivotal role in shaping both ecosystem functioning and biodiversity (Penone et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Stark et al., \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2012\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eLiDAR technology is a powerful tool for describing forest structure at fine spatial scales. LiDAR surveys using aerial laser scanning (ALS) methods have been used for the past few decades to estimate structural characteristics such as forest height, aboveground biomass, and leaf area index (e.g., Drake et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2002\u003c/span\u003e; Ioki et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Tang et al., \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). While ALS efficiently surveys forest canopies, it is not an effective tool to characterize lower canopy structure or near-ground vegetation (White et al., \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2016\u003c/span\u003e), especially in dense tropical rainforests. Terrestrial laser scanning (TLS), which uses a tripod-mounted LiDAR instrument to scan forests from the ground, is better suited to capture below-canopy structure characteristics. Indeed, TLS has been successfully used to describe forest structure across various tropical sites (e.g., Aiba et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Ehbrecht et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Palace et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eHere, we present an analysis of the terrestrial mammal community in Gunung Palung National Park (GPNP), West Kalimantan, Indonesia. Our study area spans over 1,000 m of elevation and contains 12 distinct forest type partitions (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). We used forest structure metrics derived from TLS surveys at 58 locations to predict six mammal diversity metrics calculated from camera trap data across four spatial scales: individual camera trap/TLS survey locations (0.002 km\u003csup\u003e2\u003c/sup\u003e), forest type partitions (~\u0026thinsp;1.5 km\u003csup\u003e2\u003c/sup\u003e), a grid with 0.75 km\u003csup\u003e2\u003c/sup\u003e cells, and a grid with 1.7 km\u003csup\u003e2\u003c/sup\u003e cells. We expect, based on predictions of the habitat heterogeneity hypothesis, that structure metrics will be positively associated with diversity metrics. We also expect structure metrics that are more ecologically relevant to terrestrial mammals, such as terrestrial vegetation density, to be more influential than metrics describing higher forest strata, such as canopy roughness (i.e., rumple index). Furthermore, we expect to find that different structure metrics best predict diversity metrics at different spatial scales. Our overall objective was thus to explore the scale-dependency between vegetation structure metrics and biodiversity measures.\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eModel output\u003c/h2\u003e \u003cp\u003eEight of the 15 forest structure metrics we considered were reliable predictors of diversity metrics across the four spatial scales we modeled. Predictors differed between scales with little overlap. Only two structural metrics appear in models at multiple scales: rumple index which is a positive predictor at the camera trap and forest type partition scales, and mean tree height which is a negative predictor in both the 0.75 km\u003csup\u003e2\u003c/sup\u003e and 1.7 km\u003csup\u003e2\u003c/sup\u003e grid scale models. The other six forest structure metrics only appear in models at one spatial scale, including stand density and basal area in the camera trap scale models, mean dbh in the 0.75 km\u003csup\u003e2\u003c/sup\u003e scale models, vertical complexity in the 1.7 km\u003csup\u003e2\u003c/sup\u003e scale models, and vegetation volume and vertical point distribution in the partition scale models (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe patterns of reliable predictors were highly consistent across models for different diversity metrics, with certain structure metrics repeatedly emerging as reliable or near-reliable predictors. For brevity, we focus mainly on the results of the Shannon diversity models here. Detailed model results for all other metrics (species richness, species evenness, functional evenness, and functional divergence) are available in the supplementary materials. Due to model convergence issues and unstable coefficient estimates we do not report results from the functional richness models.\u003c/p\u003e \u003cp\u003eTo evaluate model fit, we focused on the forest type partition and 1.7 km\u003csup\u003e2\u003c/sup\u003e grid scales, as they are the most comparable in spatial resolution. Models at the partition scale consistently outperformed those at the 1.7 km\u003csup\u003e2\u003c/sup\u003e grid scale across all diversity metrics considered. For instance, in the Shannon diversity model, the partition scale yielded a deviance of 3.2, compared to 7.5 at the 1.7 km\u003csup\u003e2\u003c/sup\u003e grid scale.\u003c/p\u003e \u003cp\u003eWhen comparing predictor effects across scales, it is useful to group the scales into two categories. The camera trap and partition scales can be grouped together since they are ecologically defined \u0026ndash; both are only associated with a single forest type. The two grid scales form the second group based on the use of a randomly placed grid system. This approach allows us to compare the ecologically defined scales to those based on random grids. We can also create three groups of forest structure metrics based on the pattern of effect direction between the ecological and grid scales: if the direction of the effect is in the opposite direction, same direction, or mixed between the two scale groups. For instance, in the Shannon diversity models (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e), seven predictors show opposite direction of effects between the ecological and grid scales, with one group showing a directional effect while the other shows an effect in the opposite direction (e.g., rumple index), or one group showing no effect while the other shows a directional effect (e.g., mean tree height). Four predictors in the Shannon diversity models are grouped based on most models at different scales showing predictor effects in the same direction (e.g., standard deviation of tree heights shows a negative effect at all scales), including showing no effect. The final four metrics are grouped based on no consistent pattern apparent in predictor effect directions between the two scale groups.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eLiDAR-derived metrics\u003c/h3\u003e\n\u003cp\u003eMetrics such as maximum, average, and standard deviation of tree heights, average dbh, tree volume, canopy relief ratio, leaf area density, and vertical rumple follow a reverse C-shaped pattern across forest types (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). These values are lower at low elevation locations in the peat swamp and freshwater swamp forests, increasing through the alluvial bench and lowland sandstone forests, and incrementally decrease as elevation increases, in many cases reaching their lowest values in the montane forests. Additionally, higher elevation forest types, especially montane forests, show the least amount variation in these metrics compared to other forests.\u003c/p\u003e \u003cp\u003eOther metrics show a slightly different relationship with forest type and elevation. Basal area, tree density, and terrestrial vegetation metrics follow a pattern similar to the C-shape of the tree height metrics, for example, with low values in the peat swamp forests and higher values in the freshwater swamp and alluvial bench forests. However, these values increase in higher elevation forests, creating an S-shaped pattern in the elevation plots (Fig. S1). This indicates that tree densities, along with terrestrial vegetation density, increase in higher elevation forests, where trees are shorter and smaller in diameter, and account for an average basal area comparable to lower elevation forests (i.e., lowland granite and lowland sandstone) where taller and larger diameter trees are abundant.\u003c/p\u003e \u003cp\u003eThe vertical point distribution and vegetation volume metrics show a pattern of generally increasing values from lower to higher elevation forest types, with the lowest average and minimum values in peat swamp and freshwater swamp forests and the highest average values in the upland granite and montane forests. Variation in the metrics is higher in the low elevation forests and decreases with elevation. The patterns in these metrics describe vegetation that is more vertically clumped and with more empty space in the lower elevation forests, particularly in the alluvial bench and lowland sandstone forests, where large trees with extensive branching crowns are more common, compared to the higher elevation forests where smaller, lower volume tree crowns are common, resulting in a more even distribution of vegetation in 3D space.\u003c/p\u003e\n\u003ch3\u003eMammal diversity metrics\u003c/h3\u003e\n\u003cp\u003eWe used observations of 36 mammal species (n\u0026thinsp;=\u0026thinsp;13,722) that we identified from camera trap videos to calculate species richness, Shannon diversity, and evenness metrics, and 35 mammal species (excluding unidentified rats, see Methods; n\u0026thinsp;=\u0026thinsp;10,934) to calculate functional richness, functional divergence, and functional evenness metrics. Species richness, Shannon diversity, and functional richness values exhibit a reverse C-shaped pattern similar to many of the forest structure metrics, with average values by forest type lowest at low and high elevation forests, and the highest average values in lowland sandstone and lowland granite forests (Fig. S4). Species evenness and functional evenness show little variation in average values by forest type, with slightly lower average values in the lower elevation forests. This pattern is reflected in the nearly flat trend lines observed when these values are plotted as a function of elevation (Fig. S2). Average functional diversity values by forest type shows a unique pattern with the lowest values observed in the peat swamp forests, gradually increasing to the highest values in the lowland sandstone forests, and minimally decreasing in the upper elevation forests.\u003c/p\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eOur results show that the scale at which forest structure metrics are considered matters when predicting species diversity. A clear pattern emerged as we considered spatial scales from small areas (0.002 km\u003csup\u003e2\u003c/sup\u003e) confined around camera traps to larger grid cells (1.7 km\u003csup\u003e2\u003c/sup\u003e): larger spatial scales capture more ecological heterogeneity. This is evident in our models by the widening confidence intervals for forest structure predictors of species diversity as spatial scale increases, particularly at the 1.7 km\u003csup\u003e2\u003c/sup\u003e scale (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). While this trend aligns with previous findings (Yao et al., \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), the confidence intervals at the 1.7 km\u003csup\u003e2\u003c/sup\u003e scale are considerably wider than those at the forest type partition scale, despite their comparable spatial resolutions. This is further supported by adjusted coefficient of variation (CV) values, which are higher at the 1.7 km\u003csup\u003e2\u003c/sup\u003e scale than at the partition scale for 12 of 15 structure metrics (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Additionally, cells in the grid systems contained up to seven different forest type partitions (Fig. S3), indicating that the 1.7 km\u003csup\u003e2\u003c/sup\u003e grid cells contained substantially more heterogeneity than the forest type partitions in our study area. These findings suggest that partitions offered a more ecologically coherent scale for analysis. Grid systems that fail to account for important ecological or environmental boundaries may thus group survey locations inappropriately and impede our ability to understand species diversity patterns based on structural metrics.\u003c/p\u003e \u003cp\u003eThe use of grid systems yielded model results that differed markedly from those derived from more ecologically informed scales. Grid system model results show three metrics are reliable predictors of diversity that are not reliable predictors in the camera trap or partition models. Partition scale models also consistently outperformed grid scale models. If differences among forest types were not ecologically salient to mammals, we would expect to find no difference in model outcomes between the partition and comparable grid scales. A possible explanation for this observation is the strong scale dependency of animal-habitat associations and use (Johnson, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e1980\u003c/span\u003e). Specifically, the diversity metrics we assess at community levels may reflect emergent properties derived from the aggregation of individual behavioral decisions made across different scales. While our study focuses on diversity metrics at broader scales (i.e., forest type partitions), these patterns may ultimately be shaped by fine-scale habitat use decisions made by individuals. This connection would further underscore the importance of considering ecologically defined scales in analyses such as ours, as these scales are likely more aligned with the decisions animals make than are arbitrary grid systems.\u003c/p\u003e \u003cp\u003eWe predicted that forest structure characteristics would be positively associated with diversity metrics, following the habitat heterogeneity hypothesis (MacArthur \u0026amp; MacArthur, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e1961\u003c/span\u003e). Our findings support this prediction at the forest type partition scale, where all reliable predictors (rumple index, vegetation volume, and vertical point distribution) are positively associated with diversity metrics. However, at the camera trap scale, two of the three reliable predictors (tree density and basal area) show a negative association with diversity metrics, while rumple index is positively correlated with species richness. These results align with previous research indicating that negative effects of environmental heterogeneity are more likely to be observed at smaller spatial scales compared to larger ones (Stein et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). However, in the two grid scale models, mean tree height is negatively associated with all diversity metrics. Due to the issues with grid scales discussed earlier, we interpret the results of these models with caution.\u003c/p\u003e \u003cp\u003eAt the camera trap scale, only one structure metric - rumple index of roughness - was a positive predictor, with rumple index also emerging as a positive predictor of diversity metrics at the partition scale. Rumple index describes forest canopy surface heterogeneity (roughness) by calculating the ratio between canopy surface area and its projected area on the ground, specifically capturing forest canopy layer heterogeneity. This contrasts with our prediction that structural metrics describing canopy complexity would not be reliable predictors of terrestrial mammal diversity. However, if canopy heterogeneity is associated with overall forest complexity, as previously suggested (Kane et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2010\u003c/span\u003e), then rumple index may serve as a proxy for structural complexity across forest strata, explaining its positive association with diversity. Alternatively, the terrestrial mammals we consider here may have ecological requirements associated with a heterogenous forest canopy, such as access to a higher diversity of fruiting trees. Complex canopy structures in tropical forests support a wide range of epiphytes and invertebrates (Nadkarni, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e1994\u003c/span\u003e) which may directly or indirectly enhance the diets of terrestrial mammals, thereby contributing to their species richness. Additionally, a structurally diverse canopy may foster a more complex understory in some dimension that we did not capture with our structure metrics, such as understory light availability (Stark et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2015\u003c/span\u003e), creating richer habitats that support a greater variety of terrestrial species.\u003c/p\u003e \u003cp\u003eTwo more positive predictors of diversity metrics appear in the partition scale models - vegetation volume and vertical point distribution. Vegetation volume, measured here as the ratio of vegetated voxels (3D pixels) to all voxels in a point cloud, and vertical point distribution, with higher values indicating an even vertical distribution of vegetation, both suggest increased forest complexity through enhanced microhabitat availability and surface area. This added niche space likely promotes species richness and diversity (Holt, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). These findings mirror previous research from Tanzania linking habitat surface area ratio with mammal occupancy (Gorczynski et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eContrary to our expectations based on the habitat heterogeneity hypothesis, tree density and basal area at the camera trap scale were negative predictors of two diversity metrics. While the negative effect of these metrics may be in line with negative complexity-heterogeneity relationships found at small spatial scales (Stein et al., \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2014\u003c/span\u003e), it is also possible that these relatively simple metrics are not effectively capturing forest structural complexity as well as other metrics considered here. For instance, areas with higher tree densities may still lack critical microhabitats and open niches that certain species require for foraging or nesting (Wilson \u0026amp; Cresswell, \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). Further, this result may reflect habitat preferences of terrestrial mammals that use more open forest areas with lower tree densities for travel, reflecting a higher likelihood of being captured by cameras placed in more open areas and not an ecological requirement, per se (Slater et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). This result may also reflect camera trap sampling bias, with cameras in more open areas capturing more mammals than cameras in areas with higher tree density, due to an increase in camera viewable area (Hofmeester et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), leading to a negative association with density metrics.\u003c/p\u003e \u003cp\u003eOur study period spans more than five years, encompassing both non-mast periods and one full masting cycle, a key feature of Southeast Asian tropical rainforests that occurs at irregular intervals of 2\u0026ndash;10 years (Appanah, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e1993\u003c/span\u003e). During masting events, numerous plant species from different families simultaneously flower and fruit in a 3\u0026ndash;6 month span, resulting in a temporary period of resource abundance, followed by resource scarcity during the intervening non-mast periods (Ashton et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e1988\u003c/span\u003e). These fluctuations significantly impact mammal populations, particularly seed predators, which experience population booms during masting and subsequent declines afterward (Curran \u0026amp; Leighton, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2000\u003c/span\u003e). Although our models do not explicitly account for mast events or forest productivity, the inclusion of both mast and non-mast periods in our study ensures that estimates of mammal diversity metrics are not biased, as they might be if only one type of period was considered.\u003c/p\u003e \u003cp\u003eUsing TLS-derived point clouds we extracted several useful structural metrics for the present analysis. Traditional metrics such as mean tree height, mean dbh, and tree density, serve as fundamental descriptors of forest structure (Wilkes et al., \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2016\u003c/span\u003e) and have been collected by foresters and ecologists for decades without the need for advanced methods like TLS surveys (e.g., Grubb et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e1963\u003c/span\u003e). However, many of the structure metrics that we consider here describe forest structure in more detail, capturing complexity beyond the simpler metrics of tree height and canopy cover (Calders et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Specifically, our analysis includes point cloud- and voxel-level metrics that capture intricate forest characteristics such as vertical point complexity, the ratio of cubic vegetation volume to empty space, and the volume of understory gaps. Importantly, we find that these advanced metrics are crucial predictors of mammal diversity and other biodiversity indices. Their collection would not have been feasible without the capabilities offered by TLS technology.\u003c/p\u003e \u003cp\u003eOur findings emphasize the value of using ecologically informed scales when modeling mammal diversity. These scales capture meaningful ecological variation, leading to more reliable predictions compared to randomly placed grids. Grid scales, by contrast, introduce artificial boundaries that often combine ecologically distinct areas, reducing the precision of biodiversity analyses. As our results show, scales aligned with ecological boundaries not only outperform grid scales but also provide more reliable insights into the relationships between biodiversity and forest structure. Research across a diversity of ecosystems would likely benefit from prioritizing the use of ecologically relevant spatial scales.\u003c/p\u003e"},{"header":"METHODS","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eStudy site\u003c/h2\u003e \u003cp\u003eWe conducted our research at the Cabang Panti Research Site (CPRS; 1\u0026deg;13\u0026prime; S, 110\u0026deg;7\u0026prime; E), a tropical rainforest site located in Gunung Palung National Park, West Kalimantan, Indonesia. CPRS covers an area of approximately 26 km\u003csup\u003e2\u003c/sup\u003e and contains seven distinct and contiguous forest types that vary in a number of physical and ecological characteristics including soil type, elevation, plant and animal species composition, and forest structure (Marshall, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Marshall et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2014\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The distinct forest types are peat swamp forest (5\u0026ndash;10 m asl), freshwater swamp forest (5\u0026ndash;10 m asl), alluvial bench forest (5-100 m asl), lowland sandstone forest (20\u0026ndash;200 m asl), lowland granite forest (200\u0026ndash;400 m asl), upland granite forest (350\u0026ndash;800 m asl), and montane forest (750-1,100 m asl). Forest types at CPRS can be further subdivided into spatially distinct forest type partitions (e.g., when they are on different mountain ridges), allowing for analyses at finer spatial resolution. Partitions differ in environmental conditions including rainfall, temperature, and elevation, and thus are likely to vary in structure as well. Forest types have been split into two partitions based on separation of forest types by mountain ridges and a river (lowland sandstone, lowland granite, upland granite, and montane forests) or by discontinuity (alluvial bench forests). Two forest types, peat swamp and freshwater swamp are not subdivided, yielding a total of 12 partitions (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eForest structure measurement\u003c/h3\u003e\n\u003cp\u003eWe conducted LiDAR scan surveys across CPRS between 1 February and 24 February 2023 using a terrestrial laser scanner (TLS). We chose LiDAR survey locations based on previously established camera trap locations (see below). We used a FARO Focus Laser Scanner s350 (Faro Technologies Inc., Lake Mary, USA) to survey camera trap locations and create 3D point clouds of the forest structure, using scan settings of 1/8 resolution and 3x quality, following FARO guidelines for outdoor scanning. We placed the scanner on a tripod approximately 1.4 m above ground. We scanned 58 locations across all seven forest types and 11 of the 12 partitions at CPRS \u0026ndash; six in peat swamp forests, ten in freshwater swamp and alluvial bench forests, nine in lowland sandstone and upland granite forests, eleven in lowland granite forests, and three in montane forests. We collected multiple scans at each location to ensure adequate 3D coverage, with an average of 22 scans per location (range 13\u0026ndash;33 scans). We created scan areas by conducting scans 1 m \u0026minus;\u0026thinsp;15 m from trees on which camera traps were placed, taking multiple scans approximately 1 m \u0026minus;\u0026thinsp;2 m apart within this scan radius. We made abundant use of targets (checkerboards and mounted spheres; Fig. S6) in the scan areas to assist with scan registration, a process that allows multiple scans taken at one location to be merged into a single 3D point cloud (Fig. S7). Up to 25 0.25 m X 0.2 m checkerboards were affixed to trees throughout the scan area, and up to five white spheres with a diameter of 0.2 m were mounted on tripods at heights adjusted to be visible to the scanner (e.g., above terrestrial vegetation) at approximately 1 m \u0026minus;\u0026thinsp;2 m. We conducted scan registration using FARO Scene software (v2023.1; Faro Technologies Inc., Florida, USA) and exported the resulting point clouds as LAS files for further processing in R Statistical Software (v4.2.1; R Core Team 2024).\u003c/p\u003e \u003cp\u003eWe extracted multiple forest structure metrics for each scanned location using the \u003cem\u003eFORTLS\u003c/em\u003e (Molina-Valero et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), \u003cem\u003elidR\u003c/em\u003e (Roussel et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), and \u003cem\u003elidRmetrics\u003c/em\u003e (Tompalski, \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) packages in R, utilizing functions that apply tree detection algorithms and calculate tree-, stand-, and cloud-level variables. To ensure consistency across sites, we standardized all point clouds to a uniform 25 m radius (0.002 km\u003csup\u003e2\u003c/sup\u003e), the default in \u003cem\u003eFORTLS\u003c/em\u003e, before extracting structure metrics. The metrics we consider here include: maximum tree height (m); mean tree height (m); canopy relief ratio, the ratio of mean tree height to maximum tree height; standard deviation of tree heights (m), used here to describe the level of forest stratification; mean diameter at breast height (dbh; cm), measured at height of 1.3 m; tree basal area (m\u003csup\u003e2\u003c/sup\u003e/ha); tree density (trees/ha), tree volume (m\u003csup\u003e3\u003c/sup\u003e/ha), number of points below 2 m, a proxy for the proportion of terrestrial vegetation; vertical point distribution; vertical complexity; leaf area density; rumple index of roughness, which describes the horizontal heterogeneity of the forest canopy surface; vegetation density, a 3D (voxelized) approach to quantifying the amount of cubic vegetated space in each point cloud; and closed gap space, describing the volume (m\u003csup\u003e3\u003c/sup\u003e) of cubic space beneath the canopy classified as vegetation gaps.\u003c/p\u003e\n\u003ch3\u003eMammal surveys and diversity metrics\u003c/h3\u003e\n\u003cp\u003eWe monitored mammal communities at CPRS between June 2015 and September 2020 using motion-triggered video cameras (Bushnell TrophyCam HD; Overland Park, KS). This ensured that we monitored species presence throughout a dipterocarp mast cycle that drives the distribution and abundances of mammals in Bornean rainforest systems (e.g., Curran \u0026amp; Leighton, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2000\u003c/span\u003e; Marshall et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Marshall et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). We placed cameras at 177 sites (M\u0026thinsp;=\u0026thinsp;260.1 days per camera, SD\u0026thinsp;=\u0026thinsp;162.8, range 23\u0026ndash;906) across all seven forest types at CPRS, along both established trails and off-trail placement sites (n\u0026thinsp;=\u0026thinsp;46,044 trap nights). We mounted cameras on trees, placed at approximately 0.4 m from the ground, and programmed them to capture 20 s videos when triggered. We identified 55 non-human mammal species and assigned a terrestriality designation (terrestrial or arboreal) to each based on available data from the PanTHERIA database (Jones et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2009\u003c/span\u003e) and the Phillips\u0026rsquo; Field Guide to the Mammals of Borneo text (Phillipps, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Here, we only include terrestrial mammal observations, yielding 36 terrestrial mammals for subsequent analyses. This includes observations of rats (Family Muridae) and shrews (Family Soricidae) that we did not identify to species and so count as one species for diversity metrics. Rats, but not shrews, were excluded in the calculation of functional diversity metrics. We included shrews in functional diversity calculations since it is likely only two possible species range through GPNP and are closely related in ecology. We therefore averaged quantitative ecological traits between the two species to characterize Family Soricidae. We calculated six diversity metrics for each camera trap location and at larger spatial scales (see \u003cem\u003eSpatial scales\u003c/em\u003e). Using the \u003cem\u003evegan\u003c/em\u003e package in R (Oksanen et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) we calculated species richness, Shannon diversity index, and Pielou's species evenness values. Additionally, we used the \u003cem\u003eFD\u003c/em\u003e package (Lalibert\u0026eacute; et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2014\u003c/span\u003e) to calculate functional richness, functional evenness, and functional divergence for 35 species (excluding rats) using six ecological traits (body mass, activity cycle, trophic level, diet breadth, group size, and litter size; see Table S1).\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eModeling approach\u003c/h2\u003e \u003cp\u003eTo estimate the influence of forest structure on the mammal diversity metrics described above we fit generalized linear models (GLM) in R. All models used structural metrics as predictors and were specified using a gamma distribution. We include the number of days each camera trap location was operational as an offset in all models due to uneven survey effort across locations. We applied weights in models at all scales except the camera trap scale to account for variation in the number of scanned locations within each spatial subunit (i.e., grid cells and forest type partitions). We assessed all quantitative predictors for correlation before fitting models, excluding highly correlated pairs (r\u0026thinsp;\u0026gt;\u0026thinsp;0.8) from the same model. We found significant spatial autocorrelation at the camera trap scale for all diversity metrics except functional richness and functional evenness, using Moran\u0026rsquo;s I. To account for this, we applied eigenvector-based spatial filtering by including Moran\u0026rsquo;s eigenvectors as predictors in the models with significant spatial autocorrelation. We used the \u0026lsquo;dredge\u0026rsquo; function in the \u003cem\u003eMuMin\u003c/em\u003e package (Bartoń, 2023) to fit candidate models, specifying models at all scales to include a maximum of two predictors, not including eigenvectors, to prevent overfitting and to ensure model complexity was the same across models at different spatial scales. We then used model averaging across all candidate models to assess the influence of all forest structure predictors. We determined predictors to be reliable indicators of outcome variables if the 95% confidence intervals (CI\u0026rsquo;s) of the beta coefficient estimate did not overlap zero.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eSpatial scales\u003c/h2\u003e \u003cp\u003eTo assess the influence of structure metrics at different spatial scales, we fit all models at four spatial scales. In addition to the camera trap level scale, we created two grid systems that overlay the study area to aggregate locations for which we collected LiDAR (i.e., scanned locations) and camera trap data. We ensured each non-empty grid cell contained at least two scanned locations. One scale is a 6 X 4 grid containing 0.75 km\u003csup\u003e2\u003c/sup\u003e cells, creating 17 groups of scanned locations, with seven cells not containing any scanned locations. The other is 4 X 3 grid containing 1.7 km\u003csup\u003e2\u003c/sup\u003e cells, creating 10 groups of scanned locations. We also fit models using the forest type partition scale, which is comparable to the 1.7 km\u003csup\u003e2\u003c/sup\u003e grid scale in both area of sub-units (x̄ \u0026gt; 1.50 km\u003csup\u003e2\u003c/sup\u003e) and groups of scanned locations (n\u0026thinsp;=\u0026thinsp;11). Additionally, we grouped scanned locations in two larger spatial scales to assess variation in structure metrics across multiple scales. These additional grid scales include a 2 X 2 grid containing 6.5 km\u003csup\u003e2\u003c/sup\u003e cells and a scale comprising the entire study area, approximately 26 km\u003csup\u003e2\u003c/sup\u003e and were not used for modeling purposes. To assess the variation captured by spatial scales that aggregate TLS survey locations (i.e., forest type partition and grid scales), we calculated the adjusted coefficient of variation (CV) for all structure metrics and diversity metrics. The adjusted CV accounts for unequal sample sizes across subunits. For each metric at each scale, we first calculated the CV for each spatial subunit (partitions or individual grid cells), then averaged these values to obtain the mean CV for each scale.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eLimitations\u003c/h2\u003e \u003cp\u003eSince we were not able to identify observations of rats (Family Muridae) to the species level, we considered all observations as one species when calculating richness, diversity, and evenness metrics. We were not able to consider rats at all for functional diversity metrics. Consequently, we surely underestimated diversity metrics, given rats were the most frequently observed mammal group (n\u0026thinsp;=\u0026thinsp;2,788 camera trap observations) and, based on species range maps and descriptions (Phillips, 2016), there are potentially up to 11 rat species at GPNP. These underestimates are likely most pronounced in the Lowland Granite and Upland Granite forest types, where most rat observations were made. Indeed, previous research involving small mammal trapping at GPNP identified five Muridae species, with the majority of observations and species occurring in low- and mid-elevation forest types (Gorog, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2003\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eDue to the limitations of TLS surveying, we may have underestimated structure metrics related to higher forest strata, such as maximum tree height, at several locations, as tree crowns are often out of the instrument\u0026rsquo;s direct line of sight and thus not captured in the resulting point clouds. This likely affected our ability to capture the crowns of the tallest trees, including emergents, which play an important structural and ecological role for many tropical animals (Cannon \u0026amp; Leighton, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e1994\u003c/span\u003e). We expect this limitation to be most pronounced in the alluvial bench and lowland sandstone forests, where emergents of the Dipterocarpaceae family are most numerous and account for a significant portion of aboveground biomass in Borneo (Aiba et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2013\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFor the present study we were restricted in what forest structure metrics were available to be used for analyses by the relatively limited number of R packages and standalone software that currently exists to automate the point cloud analysis processes, and by the selection of structural metrics that are estimated by these packages and software. For example, while dozens of R packages exist that analyze LiDAR data for forest research purposes, according to a recent review of LiDAR-focused R packages only two are currently available that were designed to analyze TLS data specifically (Atkins et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Standalone software packages are also limited in their functionality in analyzing TLS data, especially point cloud data from tropical rainforests which are highly structurally complex, as many software options compute metrics for single-tree point cloud files only. Furthermore, to our knowledge, no R package or software is currently capable of detecting and providing descriptive metrics of lianas or vines from TLS point clouds from tropical forests. Lianas and vines are important structural components of tropical forests for arboreal mammals (Arroyo-Rodr\u0026iacute;guez et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2015\u003c/span\u003e), and can impact forest biodiversity and dynamics in important ways (Schnitzer, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2015\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e "},{"header":"Declarations","content":"\u003cp\u003eAcknowledgements\u003c/p\u003e\n\u003cp\u003eWe would like to thank the Indonesian Ministry of Higher Education and Research and Technology and the Gunung Palung National Park Bureau for supporting research at the Cabang Panti Research Site. Our work was supported by funding from the University of Michigan, Victoria University of Wellington, the National Science Foundation (award no. 2216525), the Leakey Foundation, the Orangutan Conservancy, the Mohamed bin Zayed Species Conservation Fund, the AZA Ape TAG Initiative, the American Society of Primatologists, and the Lewis and Clark Fund Grant from the American Philosophical Society.\u003c/p\u003e\n\u003cp\u003eThe authors have no conflicts of interest to report.\u003c/p\u003e\n\u003cp\u003eAuthor contributions\u003c/p\u003e\n\u003cp\u003eGRE and AJM designed the study; GRE, HUW, and AJM raised funds for data collection; GRE, ES, HUW, and AJM collected the data; GRE wrote all analysis R code and produced the figures; GRE and AJM wrote the manuscript. All authors contributed to the drafts and gave final approval for publication.\u003c/p\u003e\n\u003cp\u003eData and code are available at:\u003c/p\u003e\n\u003cp\u003ehttps://github.com/Estrada-Gene/LiDAR_Forest_structure\u003c/p\u003e\n"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAiba, S., Akutsu, K., Onoda, Y.: Canopy structure of tropical and sub-tropical rain forests in relation to conifer dominance analysed with a portable LIDAR system. Ann. Botany. \u003cb\u003e112\u003c/b\u003e(9), 1899\u0026ndash;1909 (2013). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/aob/mct242\u003c/span\u003e\u003cspan address=\"10.1093/aob/mct242\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAppanah, S.: Mass flowering of dipterocarp forests in the aseasonal tropics. J. Biosci. \u003cb\u003e18\u003c/b\u003e, 457\u0026ndash;474 (1993)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArroyo-Rodr\u0026iacute;guez, V., Asensio, N., Dunn, J.C., Crist\u0026oacute;bal-Azkarate, J., Gonzalez-Zamora, A.: Use of lianas by primates: More than a food source. In: Ecology of Lianas, pp. 407\u0026ndash;426. Wiley, Ltd (2015). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/9781118392409.ch27\u003c/span\u003e\u003cspan address=\"10.1002/9781118392409.ch27\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAshton, P.S., Givnish, T.J., Appanah, S.: Staggered Flowering in the Dipterocarpaceae: New Insights Into Floral Induction and the Evolution of Mast Fruiting in the Aseasonal Tropics. Am. Nat. \u003cb\u003e132\u003c/b\u003e(1), 44\u0026ndash;66 (1988). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1086/284837\u003c/span\u003e\u003cspan address=\"10.1086/284837\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAtkins, J.W., Stovall, A.E.L., Alberto Silva, C.: Open-Source tools in R for forestry and forest ecology. For. Ecol. Manag. \u003cb\u003e503\u003c/b\u003e, 119813 (2022). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2021.119813\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2021.119813\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBartoń, K.: _MuMIn: Multi-Model Inference_. R package version 1.48.4, (2024). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://CRAN.R-project.org/package=MuMIn\u003c/span\u003e\u003cspan address=\"https://CRAN.R-project.org/package=MuMIn\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBradshaw, C.J., Sodhi, N.S., Brook, B.W.: Tropical turmoil: A biodiversity tragedy in progress. Front. Ecol. Environ. \u003cb\u003e7\u003c/b\u003e(2), 79\u0026ndash;87 (2009). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1890/070193\u003c/span\u003e\u003cspan address=\"10.1890/070193\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCalders, K., Adams, J., Armston, J., Bartholomeus, H., Bauwens, S., Bentley, L.P., Chave, J., Danson, F.M., Demol, M., Disney, M., Gaulton, R., Moorthy, K., Levick, S.M., Saarinen, S.R., Schaaf, N., Stovall, C., Terryn, A., Wilkes, L., P., Verbeeck, H.: Terrestrial laser scanning in forest ecology: Expanding the horizon. Remote Sens. Environ. \u003cb\u003e251\u003c/b\u003e, 112102 (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.rse.2020.112102\u003c/span\u003e\u003cspan address=\"10.1016/j.rse.2020.112102\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCannon, C.H., Leighton, M.: Comparative locomotor ecology of gibbons and macaques: Selection of canopy elements for crossing gaps. Am. J. Phys. Anthropol. \u003cb\u003e93\u003c/b\u003e(4), 505\u0026ndash;524 (1994). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/ajpa.1330930409\u003c/span\u003e\u003cspan address=\"10.1002/ajpa.1330930409\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChase, J.M., Leibold, M.A.: Ecological Niches: Linking Classical and Contemporary Approaches. University of Chicago Press (2003)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCurran, L.M., Leighton, M.: Vertebrate Responses to Spatiotemporal Variation in Seed Production of Mast-Fruiting Dipterocarpaceae. Ecol. Monogr. \u003cb\u003e70\u003c/b\u003e(1), 101\u0026ndash;128 (2000). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1890/0012-9615(2000\u003c/span\u003e\u003cspan address=\"10.1890/0012-9615(2000\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e )070[0101:VRTSVI]2.0.CO;2\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCurrie, D.J.: Energy and Large-Scale Patterns of Animal- and Plant-Species Richness. Am. Nat. \u003cb\u003e137\u003c/b\u003e(1), 27\u0026ndash;49 (1991). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1086/285144\u003c/span\u003e\u003cspan address=\"10.1086/285144\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDrake, J.B., Dubayah, R.O., Clark, D.B., Knox, R.G., Blair, J.B., Hofton, M.A., Chazdon, R.L., Weishampel, J.F., Prince, S.: Estimation of tropical forest structural characteristics using large-footprint lidar. Remote Sens. Environ. \u003cb\u003e79\u003c/b\u003e(2), 305\u0026ndash;319 (2002). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0034-4257(01)00281-4\u003c/span\u003e\u003cspan address=\"10.1016/S0034-4257(01)00281-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEhbrecht, M., Seidel, D., Annigh\u0026ouml;fer, P., Kreft, H., K\u0026ouml;hler, M., Zemp, D.C., Puettmann, K., Nilus, R., Babweteera, F., Willim, K., Stiers, M., Soto, D., Boehmer, H.J., Fisichelli, N., Burnett, M., Juday, G., Stephens, S.L., Ammer, C.: Global patterns and climatic controls of forest structural complexity. Nat. Commun. \u003cb\u003e12\u003c/b\u003e(1) (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003eArticle 1. https://doi.org/10.1038/s41467-020-20767-z\u003c/span\u003e\u003cspan address=\"Article 1. 10.1038/s41467-020-20767-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGorczynski, D., Rovero, F., Mtui, A., Shinyambala, S., Martine, J., Hsieh, C., Frishkoff, L., Beaudrot, L.: Tropical forest mammal occupancy and functional diversity increase with microhabitat surface area. Ecology. \u003cb\u003e104\u003c/b\u003e(12), e4181 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/ecy.4181\u003c/span\u003e\u003cspan address=\"10.1002/ecy.4181\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGorog, A.J.: Historical biogeography of small mammals in the Sunda region of Southeast Asia [Ph.D., University of Michigan]. (2003). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.proquest.com/docview/287902251/abstract/20FCD055C4DE48D0PQ/1\u003c/span\u003e\u003cspan address=\"https://www.proquest.com/docview/287902251/abstract/20FCD055C4DE48D0PQ/1\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGrubb, P.J., Lloyd, J.R., Pennington, T.D., Whitmore, T.C.: A Comparison of Montane and Lowland Rain Forest in Ecuador I. The Forest Structure, Physiognomy, and Floristics. J. Ecol. \u003cb\u003e51\u003c/b\u003e(3), 567\u0026ndash;601 (1963). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/2257748\u003c/span\u003e\u003cspan address=\"10.2307/2257748\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHawkins, B.A., Field, R., Cornell, H.V., Currie, D.J., Gu\u0026eacute;gan, J.-F., Kaufman, D.M., Kerr, J.T., Mittelbach, G.G., Oberdorff, T., O\u0026rsquo;Brien, E.M., Porter, E.E., Turner, J.R.G.: Energy, Water, and Broad-Scale Geographic Patterns of Species Richness. Ecology. \u003cb\u003e84\u003c/b\u003e(12), 3105\u0026ndash;3117 (2003). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1890/03-8006\u003c/span\u003e\u003cspan address=\"10.1890/03-8006\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHazell, R.J., Chmel, K., Riegert, J., Paul, L., Isua, B., Kaina, G.S., Fibich, P., Molem, K., Stewart, A.J.A., Peck, M.R., Weiblen, G.D., Novotny, V.: Spatial scaling of plant and bird diversity from 50 to 10,000 ha in a lowland tropical rainforest. Oecologia. \u003cb\u003e196\u003c/b\u003e(1), 101\u0026ndash;113 (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00442-021-04925-8\u003c/span\u003e\u003cspan address=\"10.1007/s00442-021-04925-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHofmeester, T.R., Rowcliffe, J.M., Jansen, P.A.: A simple method for estimating the effective detection distance of camera traps. Remote Sens. Ecol. Conserv. \u003cb\u003e3\u003c/b\u003e(2), 81\u0026ndash;89 (2017). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/rse2.25\u003c/span\u003e\u003cspan address=\"10.1002/rse2.25\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHolt, R.D.: Bringing the Hutchinsonian niche into the 21st century: Ecological and evolutionary perspectives. Proceedings of the National Academy of Sciences, 106(supplement_2), 19659\u0026ndash;19665. (2009). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1073/pnas.0905137106\u003c/span\u003e\u003cspan address=\"10.1073/pnas.0905137106\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIoki, K., Tsuyuki, S., Hirata, Y., Phua, M.-H., Wong, W.V.C., Ling, Z.-Y., Saito, H., Takao, G.: Estimating above-ground biomass of tropical rainforest of different degradation levels in Northern Borneo using airborne LiDAR. For. Ecol. Manag. \u003cb\u003e328\u003c/b\u003e, 335\u0026ndash;341 (2014). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.foreco.2014.06.003\u003c/span\u003e\u003cspan address=\"10.1016/j.foreco.2014.06.003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJohnson, D.H.: The comparison of usage and availability measurements for evaluating resource preference. Ecology. \u003cb\u003e61\u003c/b\u003e(1), 65\u0026ndash;71 (1980)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJones, K.E., Bielby, J., Cardillo, M., Fritz, S.A., O\u0026rsquo;Dell, J., Orme, C.D.L., Safi, K., Sechrest, W., Boakes, E.H., Carbone, C., Connolly, C., Cutts, M.J., Foster, J.K., Grenyer, R., Habib, M., Plaster, C.A., Price, S.A., Rigby, E.A., Rist, J., Purvis, A.: PanTHERIA: A species-level database of life history, ecology, and geography of extant and recently extinct mammals. Ecology. \u003cb\u003e90\u003c/b\u003e(9), 2648\u0026ndash;2648 (2009). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1890/08-1494.1\u003c/span\u003e\u003cspan address=\"10.1890/08-1494.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKane, V.R., McGaughey, R.J., Bakker, J.D., Gersonde, R.F., Lutz, J.A., Franklin, J.F.: Comparisons between field- and LiDAR-based measures of stand structural complexity. Can. J. For. Res. \u003cb\u003e40\u003c/b\u003e(4), 761\u0026ndash;773 (2010). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1139/X10-024\u003c/span\u003e\u003cspan address=\"10.1139/X10-024\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKissling, W.D., Field, R., Korntheuer, H., Heyder, U., B\u0026ouml;hning-Gaese, K.: Woody plants and the prediction of climate-change impacts on bird diversity. Philosophical Trans. Royal Soc. B: Biol. Sci. \u003cb\u003e365\u003c/b\u003e(1549), 2035\u0026ndash;2045 (2010). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1098/rstb.2010.0008\u003c/span\u003e\u003cspan address=\"10.1098/rstb.2010.0008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLalibert\u0026eacute;, E., Legendre, P., Shipley, B.: FD: measuring functional diversity from multiple traits, and other tools for functional ecology. R package version 1.0-12.3. (2014)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMacArthur, R.H., MacArthur, J.W.: Bird. Species Divers. Ecol. \u003cb\u003e42\u003c/b\u003e(3), 594\u0026ndash;598 (1961). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/1932254\u003c/span\u003e\u003cspan address=\"10.2307/1932254\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarshall, A.J.: Gibbons Hylobates albibarbis? Biotropica. \u003cb\u003e41\u003c/b\u003e(2), 257\u0026ndash;267 (2009). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1744-7429.2008.00461.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1744-7429.2008.00461.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e Are Montane Forests Demographic Sinks for Bornean White-bearded\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarshall, A.J., Beaudrot, L., Wittmer, H.U.: Responses of Primates and Other Frugivorous Vertebrates to Plant Resource Variability over Space and Time at Gunung Palung National Park. Int. J. Primatol. \u003cb\u003e35\u003c/b\u003e(6), 1178\u0026ndash;1201 (2014). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10764-014-9774-4\u003c/span\u003e\u003cspan address=\"10.1007/s10764-014-9774-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarshall, A.J., Farr, M.T., Beaudrot, L., Zipkin, E.F., Feilen, K.L., Bell, L.G., Setiawan, E., Susanto, T.W., Setia, M., Leighton, T., M., Wittmer, H.U.: Biotic and abiotic drivers of dispersion dynamics in a large-bodied tropical vertebrate, the Western Bornean orangutan. Oecologia. \u003cb\u003e196\u003c/b\u003e(3), 707\u0026ndash;721 (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00442-021-04964-1\u003c/span\u003e\u003cspan address=\"10.1007/s00442-021-04964-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcCain, C.M.: Elevational Gradients in Diversity of Small Mammals. Ecology. \u003cb\u003e86\u003c/b\u003e(2), 366\u0026ndash;372 (2005). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1890/03-3147\u003c/span\u003e\u003cspan address=\"10.1890/03-3147\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMolina-Valero, J.A., Mart\u0026iacute;nez-Calvo, A., Ginzo Villamayor, M.J., Novo P\u0026eacute;rez, M.A., \u0026Aacute;lvarez-Gonz\u0026aacute;lez, J.G., Montes, F., P\u0026eacute;rez-Cruzado, C.: Operationalizing the use of TLS in forest inventories: The R package FORTLS. Environ. Model. Softw. \u003cb\u003e150\u003c/b\u003e, 105337 (2022). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.envsoft.2022.105337\u003c/span\u003e\u003cspan address=\"10.1016/j.envsoft.2022.105337\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNadkarni, N.M.: Diversity of Species and Interactions in the Upper Tree Canopy of Forest Ecosystems1. Am. Zool. \u003cb\u003e34\u003c/b\u003e(1), 70\u0026ndash;78 (1994). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/icb/34.1.70\u003c/span\u003e\u003cspan address=\"10.1093/icb/34.1.70\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOksanen, J., Simpson, G., Blanchet, F., Kindt, R., Legendre, P., Minchin, P., O'Hara, R., Solymos, P., Stevens, M., Szoecs, E., Wagner, H., Barbour, M., Bedward, M., Bolker, B., Borcard, D., Carvalho, G., Chirico, M., De Caceres, M., Durand, S., Evangelista, H., FitzJohn, R., Friendly, M., Furneaux, B., Hannigan, G., Hill, M., Lahti, L., McGlinn, D., Ouellette, M., Ribeiro Cunha, E., Smith, T., Stier, A., Ter Braak, C., Weedon, J.: _vegan: Community Ecology Package_. R package version 2.6-4,. (2022)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePalace, M., Sullivan, F.B., Ducey, M., Herrick, C.: Estimating Tropical Forest Structure Using a Terrestrial Lidar. PLOS ONE. \u003cb\u003e11\u003c/b\u003e(4), e0154115 (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1371/journal.pone.0154115\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0154115\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePenone, C., Allan, E., Soliveres, S., Felipe-Lucia, M.R., Gossner, M.M., Seibold, S., Simons, N.K., Schall, P., van der Plas, F., Manning, P., Manzanedo, R.D., Boch, S., Prati, D., Ammer, C., Bauhus, J., Buscot, F., Ehbrecht, M., Goldmann, K., Jung, K., Fischer, M.: Specialisation and diversity of multiple trophic groups are promoted by different forest features. Ecol. Lett. \u003cb\u003e22\u003c/b\u003e(1), 170\u0026ndash;180 (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/ele.13182\u003c/span\u003e\u003cspan address=\"10.1111/ele.13182\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePhillipps, Q.: Phillipps\u0026rsquo; Field Guide to the Mammals of Borneo and Their Ecology: Sabah, Sarawak, Brunei, and Kalimantan. Princeton University Press (2016)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePillay, R., Venter, M., Aragon-Osejo, J., Gonz\u0026aacute;lez-del-Pliego, P., Hansen, A.J., Watson, J.E., Venter, O.: Tropical forests are home to over half of the world\u0026rsquo;s vertebrate species. Front. Ecol. Environ. \u003cb\u003e20\u003c/b\u003e(1), 10\u0026ndash;15 (2022). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/fee.2420\u003c/span\u003e\u003cspan address=\"10.1002/fee.2420\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eP\u0026uuml;ttker, T., Pardini, R., Meyer-Lucht, Y., Sommer, S.: Responses of five small mammal species to micro-scale variations in vegetation structure in secondary Atlantic Forest remnants, Brazil. BMC Ecol. \u003cb\u003e8\u003c/b\u003e(1), 9 (2008). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/1472-6785-8-9\u003c/span\u003e\u003cspan address=\"10.1186/1472-6785-8-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eR Core Team: _R: A Language and Environment for Statistical Computing_. R Foundation for Statistical Computing, Vienna, Austria. (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.R-project.org/\u003c/span\u003e\u003cspan address=\"https://www.R-project.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoussel, J.R., Auty, D., Coops, N.C., Tompalski, P., Goodbody, T.R.H., S\u0026aacute;nchez Meador, A., Bourdon, J.F., De Boissieu, F., Achim, A.: lidR : An R package for analysis of Airborne Laser Scanning (ALS) data. Remote Sensing of Environment, 251 (August), 112061. \u0026lt;\u0026gt;10.1016/j.rse.2020.112061\u0026gt;. (2020)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchnitzer, S.A.: The Contribution of Lianas to Forest Ecology, Diversity, and Dynamics. In N. Parthasarathy (Ed.), Biodiversity of Lianas (pp. 149\u0026ndash;160). Springer International Publishing. (2015). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-3-319-14592-1_9\u003c/span\u003e\u003cspan address=\"10.1007/978-3-319-14592-1_9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSlater, H.D., Gillingham, P.K., Pratt, V., Eaton, B., Fletcher, S., Abdullah, A., Supriadi, Korstjens, A.H.: Living on the edge: Forest edge effects on microclimate and terrestrial mammal activity in disturbed lowland forest in Sumatra, Indonesia. Oryx: The International Journal of Conservation, 58(2), Article 2. (2023)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStark, S.C., Enquist, B.J., Saleska, S.R., Leitold, V., Schietti, J., Longo, M., Alves, L.F., Camargo, P.B., Oliveira, R.C.: Linking canopy leaf area and light environments with tree size distributions to explain Amazon forest demography. Ecol. Lett. \u003cb\u003e18\u003c/b\u003e(7), 636\u0026ndash;645 (2015). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/ele.12440\u003c/span\u003e\u003cspan address=\"10.1111/ele.12440\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStark, S.C., Leitold, V., Wu, J.L., Hunter, M.O., de Castilho, C.V., Costa, F.R.C., McMahon, S.M., Parker, G.G., Shimabukuro, M.T., Lefsky, M.A., Keller, M., Alves, L.F., Schietti, J., Shimabukuro, Y.E., Brand\u0026atilde;o, D.O., Woodcock, T.K., Higuchi, N., de Camargo, P.B., de Oliveira, R.C., Saleska, S.R.: Amazon forest carbon dynamics predicted by profiles of canopy leaf area and light environment. Ecol. Lett. \u003cb\u003e15\u003c/b\u003e(12), 1406\u0026ndash;1414 (2012). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1461-0248.2012.01864.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1461-0248.2012.01864.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStein, A., Gerstner, K., Kreft, H.: Environmental heterogeneity as a universal driver of species richness across taxa, biomes and spatial scales. Ecol. Lett. \u003cb\u003e17\u003c/b\u003e(7), 866\u0026ndash;880 (2014). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/ele.12277\u003c/span\u003e\u003cspan address=\"10.1111/ele.12277\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTang, H., Dubayah, R., Swatantran, A., Hofton, M., Sheldon, S., Clark, D.B., Blair, B.: Retrieval of vertical LAI profiles over tropical rain forests using waveform lidar at La Selva, Costa Rica. Remote Sens. Environ. \u003cb\u003e124\u003c/b\u003e, 242\u0026ndash;250 (2012). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.rse.2012.05.005\u003c/span\u003e\u003cspan address=\"10.1016/j.rse.2012.05.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTews, J., Brose, U., Grimm, V., Tielb\u0026ouml;rger, K., Wichmann, M.C., Schwager, M., Jeltsch, F.: Animal species diversity driven by habitat heterogeneity/diversity: The importance of keystone structures. J. Biogeogr. \u003cb\u003e31\u003c/b\u003e(1), 79\u0026ndash;92 (2004). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1046/j.0305-0270.2003.00994.x\u003c/span\u003e\u003cspan address=\"10.1046/j.0305-0270.2003.00994.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTompalski, P.: _lidRmetrics: Point cloud metrics for the lidR package._. R package version 0.1.3, (2024). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://ptompalski.github.io/lidRmetrics/\u003c/span\u003e\u003cspan address=\"https://ptompalski.github.io/lidRmetrics/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWhite, J.C., Coops, N.C., Wulder, M.A., Vastaranta, M., Hilker, T., Tompalski, P.: Remote Sensing Technologies for Enhancing Forest Inventories: A Review. Can. J. Remote. Sens. \u003cb\u003e42\u003c/b\u003e(5), 619\u0026ndash;641 (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/07038992.2016.1207484\u003c/span\u003e\u003cspan address=\"10.1080/07038992.2016.1207484\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWiens, J.J.: The niche, biogeography and species interactions. Philosophical Trans. Royal Soc. B: Biol. Sci. (2011). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1098/rstb.2011.0059\u003c/span\u003e\u003cspan address=\"10.1098/rstb.2011.0059\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWilkes, P., Jones, S.D., Suarez, L., Haywood, A., Mellor, A., Woodgate, W., Soto-Berelov, M., Skidmore, A.K.: Using discrete-return airborne laser scanning to quantify number of canopy strata across diverse forest types. Methods Ecol. Evol. \u003cb\u003e7\u003c/b\u003e(6), 700\u0026ndash;712 (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/2041-210X.12510\u003c/span\u003e\u003cspan address=\"10.1111/2041-210X.12510\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWilson, J.M., Cresswell, W.: How robust are Palearctic migrants to habitat loss and degradation in the Sahel? Ibis. \u003cb\u003e148\u003c/b\u003e(4), 789\u0026ndash;800 (2006). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1474-919X.2006.00581.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1474-919X.2006.00581.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYao, Z., Yang, X., Wang, B., Shao, X., Wen, H., Deng, Y., Zhang, Z., Cao, M., Lin, L.: Multidimensional beta-diversity across local and regional scales in a Chinese subtropical forest: The role of forest structure. Ecol. Evol. \u003cb\u003e13\u003c/b\u003e(10), e10607 (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/ece3.10607\u003c/span\u003e\u003cspan address=\"10.1002/ece3.10607\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Table 1","content":"\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003e\u003cb\u003eReliable predictors of diversity metrics.\u003c/b\u003e Eight forest structure metrics are reliable predictors of five different diversity metrics: species richness, Shannon diversity, species evenness, functional evenness, and functional divergence. Functional richness models are excluded due to model convergence issues. The diversity metrics were modeled across four spatial scales. The table indicates the models where each structure metric was a reliable predictor and the direction of its effect.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c5\" namest=\"c2\"\u003e \u003cp\u003eSCALE\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStructure metric\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCamera trap\u003c/p\u003e \u003cp\u003e(0.002 km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eForest type partition\u003c/p\u003e \u003cp\u003e(~\u0026thinsp;1.5 km\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.75 km\u003csup\u003e2\u003c/sup\u003e grid\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.7 km\u003csup\u003e2\u003c/sup\u003e grid\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTree density\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEvenness (-)\u003c/p\u003e \u003cp\u003eFunc. Even. (-)\u003c/p\u003e \u003cp\u003eFunc. Diver. (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBasal area\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFunc. Even. (-)\u003c/p\u003e \u003cp\u003eFunc. Diver. (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRumple index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRichness (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAll models (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean tree height\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAll models (-)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAll models (-)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean tree dbh\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAll models (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVertical point complexity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAll models (+)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVegetation volume\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRichness (+)\u003c/p\u003e \u003cp\u003eFunc. Even. (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVertical point dist.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAll models (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"LiDAR, forest structure, remote sensing, mammal diversity, spatial scale","lastPublishedDoi":"10.21203/rs.3.rs-5479224/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5479224/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eVegetation structure has emerged as a key determinant of terrestrial biodiversity based on studies using randomly placed sampling grids, often at broad spatial scales. The resultant grid cells often contain substantial heterogeneity in ecological conditions that are highly relevant for the taxa of interest, potentially undermining our ability to detect relevant drivers of diversity. Here we use 15 structural metrics measured using ground-based LiDAR to model mammalian diversity at 58 sampling locations across seven distinct tropical forest types in Indonesian Borneo. We conducted analyses at four spatial scales using over five years of camera trap data. Models predicting mammal diversity based on ecologically defined scales (i.e., forest type boundaries) outperformed models using a grid scale of comparable resolution. Our results underscore the value of LiDAR in capturing forest structural metrics relevant to mammals and highlight the importance of incorporating ecologically meaningful spatial scales in biodiversity studies.\u003c/p\u003e","manuscriptTitle":"LiDAR-derived forest structure metrics show ecologically defined scales outperform grids when predicting mammal diversity","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-19 12:43:13","doi":"10.21203/rs.3.rs-5479224/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"communications-biology","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"commsbio","sideBox":"Learn more about [Communications Biology](http://www.nature.com/commsbio/)","snPcode":"","submissionUrl":"","title":"Communications Biology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Communications Series","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"c22231b0-0395-493d-8401-6d465cac398d","owner":[],"postedDate":"March 19th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":41157846,"name":"Biological sciences/Ecology/Forest ecology"},{"id":41157847,"name":"Earth and environmental sciences/Ecology/Tropical ecology"}],"tags":[],"updatedAt":"2025-11-14T23:25:31+00:00","versionOfRecord":[],"versionCreatedAt":"2025-03-19 12:43:13","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5479224","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5479224","identity":"rs-5479224","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}