Forest degradation through logging is pervasive throughout the world's tropical forests, leading to changes in the three-dimensional canopy structure that have profound consequences for wildlife, microclimate and ecosystem functioning. Quantifying these structural changes is fundamental to understanding the impact of degradation, but is challenging in dense, structurally complex forest canopies.
We exploited discrete-return airborne LiDAR surveys across a gradient of logging intensity in Sabah, Malaysian Borneo, and assessed how selective logging had affected canopy structure (Plant Area Index, PAI, and its vertical distribution within the canopy).
LiDAR products compared well to independent, analogue models of canopy structure produced from detailed ground-based inventories undertaken in forest plots, demonstrating the potential for airborne LiDAR to quantify the structural impacts of forest degradation at landscape scale, even in some of the world's tallest and most structurally complex tropical forests.
Plant Area Index estimates across the plot network exhibited a strong linear relationship with stem basal area (R² = 0.95). After at least 11–14 years of recovery, PAI was ~28% lower in moderately logged plots and ~52% lower in heavily logged plots than that in old-growth forest plots. These reductions in PAI were associated with near-complete lack of trees >30-m tall, which had not been fully compensated for by increasing plant area lower in the canopy. This structural change drives a marked reduction in the diversity of canopy environments, with the deep, dark understorey conditions characteristic of old-growth forests far less prevalent in logged sites. Full canopy recovery is likely to take decades.
Synthesis and applications. Effective management and restoration of tropical forests requires detailed monitoring of the forest and its environment. We demonstrate that airborne LiDAR can effectively map the canopy architecture of the complex tropical forests of Borneo, capturing the three-dimensional impact of degradation on canopy structure at landscape scales, therefore facilitating efforts to restore and conserve these ecosystems.

1 INTRODUCTION

Degradation through logging is pervasive across the tropics, representing an important source of anthropogenic carbon emissions (Houghton, 2013) and land use change towards simplified production landscapes (e.g. oil palm, rubber, pulpwood and coffee) (Gaveau et al., 2016; Ordway & Asner, 2020). The island of Borneo hosts some of the largest tracts of intact forest within SE Asia, but the extent of forests here has declined by >30% from an estimated ~558,000 km³ in 1973 (Gaveau et al., 2014), with the deforestation front sweeping inland from the low-lying coastal regions (Gaveau et al., 2014). By 2010, >45% of the remaining forest had been subject to some degree of selective logging, including ~60% of the forested area in Sabah (Gaveau et al., 2014).

The direct impact of logging-driven degradation is to change the structure of the forest canopy. Trees that previously dominated the main canopy are removed, while crowns of the residual trees are damaged by felling of neighbouring trees (Pfeifer et al., 2015). Canopy structure contributes towards the regulation of microclimate (Hardwick et al., 2015; Jucker, Hardwick, et al., 2018), light availability (Kumagai et al., 2001; Montgomery & Chazdon, 2001) and canopy biogeochemical fluxes (Ellsworth & Reich, 1993; Flack-Prain et al., 2019). To a large extent, it also determines the environmental diversity within landscapes and therefore biodiversity (e.g. Coomes et al., 2009; Deere et al., 2020; Struebig et al., 2013). Degradation-driven shifts in canopy architecture therefore have the potential to propagate, affecting many different facets of ecosystem function.

To understand how forest structure responds to anthropogenic degradation—and therefore the wider impacts of degradation on tropical forests—it is critical to quantify the vertical distribution of foliage in the canopy. Foliar density is commonly quantified using Leaf Area Index (LAI, m²/m²), defined as the total (one-sided) leaf area per unit ground surface area (Watson, 1947). The vertical distribution of LAI is characterised by the distribution Leaf Area Density (LAD; units: m²/m³). The closely related Plant Area Index (PAI) and Plant Area Density (PAD) are estimated where methods do not distinguish between the leaves and branches or trunks (Gower et al., 1999). Harvesting vertical columns of foliage through the canopy is difficult and labour-intensive. Consequently, few direct estimates of vertical canopy structure in tropical forests exist (i.e. La Selva Biological Station: Costa Rica: Clark et al., 2008; Olivas et al., 2013; Tang et al., 2012; Adolfo Ducke Reserve, Brazil: Stark et al., 2012). Measurements from ground level are largely indirect, using estimates of gap fraction that do not resolve the vertical distribution of vegetation (Bréda, 2003). It is also difficult to map the degradation of canopies in dense tropical forests using optical or radar remote sensing techniques. Disturbances may be too small and regeneration of canopy cover too rapid to be captured by optical remote sensing (Milodowski et al., 2017), which also do not resolve vertical variations within canopy. The biomass and leaf area density supported by these forests exceed the signal saturation points of widely available radar products (Joshi et al., 2017).

Alternatively, canopy structure may be quantified at high spatial resolution using airborne remote sensing with LiDAR, which directly samples the three-dimensional structure of forest canopies at high spatial resolution (e.g. Stark et al., 2012; Tang et al., 2012; Vincent et al., 2017). To date, airborne LiDAR has not been applied to assess the impact of canopy degradation on the density and vertical structure of the hyper-diverse forests of Borneo. Airborne LiDAR-derived vertical canopy profiles have previously been used to investigate shifts in canopy structure driven by logging and regeneration in Costa Rica (Tang et al., 2012), finding that the PAI of secondary forests recovered to old-growth levels within 20 years. Other studies have examined the degradation driven by fires in Amazonian forests on the canopy profile (Almeida et al., 2016; Brando et al., 2019). However, the structural impact of degradation—and hence its wider environmental impact—is likely to be strongly dependent on the local context, including original old-growth canopy structure, and logging practices, which vary according to regulations and management decisions (Hosonuma et al., 2012). The dearth of studies documenting the basic structural impacts of degradation therefore represents a critical knowledge gap that undermines our ability to assess the resilience of these forests to future change (Mitchard, 2018).

We investigate the impact of degradation, through selective logging, on the canopy PAD profiles of eight 1-ha plots located in Sabah, Malaysian Borneo. These plots span a gradient of logging intensity, from undisturbed, old-growth forest to a series of degraded forest plots subject to different levels of logging (Both et al., 2019; Riutta et al., 2018). Logging predated the LiDAR survey by at least 11–14 years (Pfeifer et al., 2015). The LiDAR survey therefore provides a temporal snapshot of recovery following differing logging intensities. In addition, detailed field inventories of canopy architecture at the sites enable the development of independent models of canopy structure for cross-comparison, in the absence of harvested profiles for a true validation. This evaluation is important. To date, validation of LiDAR-derived tropical canopy profiles against vertically harvested profiles has been limited to two sites, namely one site in the Amazon, using discrete-return LiDAR (Stark et al., 2012), and the other site in Costa Rica, using full-waveform LiDAR (Tang et al., 2012). Together, the gradient in logging intensity and detailed field surveys present a unique opportunity to assess the ability of airborne LiDAR to detect canopy structural changes associated with degradation in structurally complex tropical forests. Specifically, we address the following questions:

How does the canopy structure of degraded forests, logged at various intensities, differ from old-growth forests, characterised by the PAI, and its vertical distribution (PAD) within the canopy?
How do the observed differences in PAI compare against classic structural attributes such as basal area?
What are the implications of these structural changes for the diversity of canopy environments within logged forests, compared to old-growth systems?

2 MATERIALS AND METHODS

For a more detailed description of the methods, please refer to the Supporting Information.

2.1 Field sites

Our study sites are located in the state of Sabah, Malaysian Borneo (Figure 1), comprising eight 1-ha plots (part of the Global Ecosystems Monitoring [GEM] network Marthews et al., 2014; Table 1). Each plot spans 1 ha, comprising a regular lattice of twenty-five 0.04-ha (20-m x 20-m) subplots. Four plots are located in undisturbed old-growth forest: two within the Maliau Basin Conservation Area (MLA-01 and MLA-02) two within the Danum Valley Conservation area inside the Danum 50-ha CTFS-ForestGEO plot (DAN-04 and DAN-05). The remaining four plots are located within logged forest fragments in the Kalabakan Forest Reserve. All three regions were originally within a connected tract of lowland dipterocarp rainforest.

Details are in the caption following the image — **FIGURE 1**
Open in figure viewer PowerPoint

Map of northern Borneo, indicating the locations of the field sites. The plotted above-ground carbon density (ACD) is from the Sabah-wide map published by Asner et al. (2018). Mangroves and plantations have been masked

TABLE 1. Summary characteristics for the 1-ha plots on which this study is based

GEM plot code (SAFE site names)	Location	Latitude (N)/Longitude (E)	Forest type	Basal area^a/(m²/ha)	Max. canopy height^b/m	LiDAR pulse density/pulses m⁻² ^{mean (min, max)}
MLA-01 Maliau Belian	Maliau Basin Conservation Area	4.747/116.951	Old-growth	41.6 ± 3.6	70.0	24.5 (9.0/34.6)
MLA-02 Maliau Seraya	Maliau Basin Conservation Area	4.737/116.951	Old-growth	34.7 ± 2.7	68.7	22.9 (14.8/34.2)
DAN-04 Danum Carbon 1	Danum Valley Conservation Area	4.953/117.795	Old-growth	32.0 ± 3.2	58.4	3.3 (2.7/4.1)
DAN-05 Danum Carbon 2	Danum Valley Conservation Area	4.958/117.795	Old-growth	30.6 ± 3.4	62.5	9.5 (5.2/18.6)
SAF-03 Fragment E	SAFE landscape, Kalabakan Forest Reserve	4.690/117.586	Moderately logged	19.6 ± 1.9	48.6	32.9 (26.7/51.0)
SAF-04 Fragment LF	SAFE landscape, Kalabakan Forest Reserve	4.765/117.702	Moderately logged	19.3 ± 1.7	33.0	19.6 (16.4/26.0)
SAF-02 Fragment B North	SAFE landscape, Kalabakan Forest Reserve	4.744/117.618	Heavily logged	11.1 ± 1.8	29.5	34.8 (22.3/49.7)
SAF-01 Fragment B South	SAFE landscape, Kalabakan Forest Reserve	4.729/117.618	Heavily logged	6.81 ± 1.0	28.5	39.5 (26.6/55.6)

^a Basal area for all trees with DBH ≥10 cm (Riutta et al., 2018).
^b 99th percentile of LiDAR first return heights.

The logged forest fragments have been subject to varying intensities of selective logging (Ewers et al., 2011). Two plots (SAF-03 and SAF-04) have been selectively logged twice (‘moderately logged’). Two plots (SAF-01 and SAF-02) have been subject to four rounds of selective logging (‘heavily logged’). The first round of logging occurred in the mid-1970s, with an estimated 114 m³/ha was removed under a modified uniform system (Struebig et al., 2013). Subsequent rounds took place during the 1990s and early 2000s, during which an additional 37 m³/ha of timber was removed at moderately logged sites. Lifting of restrictions at heavily logged sites resulted in further rounds of logging, removing a cumulative total of ~66 m³/ha (Struebig et al., 2013). Thus the last date of logging predated the LiDAR survey by 11–14 years (Pfeifer et al., 2015). The normal protocol of 60-year rotations was not followed, as the area had been set apart for conversion to oil palm plantation between 2015 and 2017 (Ewers et al., 2011), although the oil palm was never planted.

2.2 Field estimation of vertical canopy structure

In each plot, the positions and heights of all trees with a stem DBH ≥10 cm were mapped using ground-based Field-Map technology (IFER, Ltd.). We mapped individual tree crowns by measuring five to 30 spatial positions, representing the boundary of a crown projected onto the horizontal plane. Crown projections were smoothed using the ‘smooth contour line’ feature of Field-Map software v.11.

We use the canopy inventory survey to derive the estimates of vertical canopy structure independent of the LiDAR-based methods. Simple canopy volume models are clearly a simplification of true canopy structure. For example, tropical trees in the understorey have been found to have deeper crowns than their counterparts in the upper canopy (Kohyama et al., 2003; Montgomery & Chazdon, 2001). Nevertheless, field-based canopy crown models provide a useful and independent estimate for validation purposes where direct observations are not available, and have previously been used to help validate LiDAR-based structural metrics (Coops et al., 2007; Knapp et al., 2018). For each plot, we simulated a forest of ellipsoid model crowns, based on field-measured heights and crown areas, and crown depths determined using a regional allometric scaling relationship derived from the BAAD database (Falster et al., 2015). For comparison with the LiDAR canopy profiles, leaf area was assumed to be uniformly distributed within the crowns (e.g. Knapp et al., 2018); contributions from the trunks were ignored. To account for the predictive uncertainty associated with the allometric relationships, we used a Monte Carlo approach, producing 100 crown models for each plot.

2.3 LiDAR estimation of canopy structure

NERC’s Airborne Research Facility (ARF) undertook an airborne LiDAR survey in November 2014, using a Leica ALS50-II LiDAR sensor on-board a Dornier 228-201 (flight elevation: 1,400–2,400 m a.s.l., depending on the site; flight speed: 120–140 knots). The average density of the resultant point clouds varied between sites due to differing levels of flight line overlap (Table 1). We classified the points into ground and non-ground returns using LAStools (rapidlasso GmbH, Gilching, Germany) and normalised return heights to height above ground.

To quantify PAD distributions from airborne discrete LiDAR data, we use a variant of the 1D Beer–Lambert approximation for light propagation through a turbid medium (MacArthur & Horn, 1969; Stark et al., 2012). Beer–Lambert models have been widely applied to estimate the canopy PAD profiles from using both full-waveform (e.g. Tang et al., 2012) and discrete-return LiDAR (e.g. Stark et al., 2012). The resultant profiles have been validated against directly harvested foliage profiles in tropical forests in both the Brazilian Amazon (Stark et al., 2012) and Costa Rica (Tang et al., 2012).

The basic premise of the Beer–Lambert approximation is that for a laterally homogeneous canopy, with vertical distribution of plant density PAD(z), where z is the depth into the canopy from its top, the PAD for a given layer of thickness ∆z = |z_i–z_i₋₁| can be estimated based on the vertical column of LiDAR returns:

$urn:x-wiley:00218901:media:jpe13895:jpe13895-math-0001$ (1)

where w_i represents the points, weighted by the number of returns associated with their respective LiDAR pulse (e.g. Armston et al., 2013); κ is a correction factor accounting for canopy characteristics, such as clumping of vegetation within the canopy (Ni-Meister et al., 2001) and the leaf angle distribution (Detto et al., 2015). The number of returns entering the top of a canopy layer determines the numerator in the log term; the number of returns penetrating into underlying layers defines the denominator. We use a layer thickness, ∆z, of 1 m. We do not account for the azimuth of the returns.

We lack direct estimates of PAI for calibration of $urn:x-wiley:00218901:media:jpe13895:jpe13895-math-0002$ . However, Schneider et al. (2019) have published vertical profiles of PAI for an old-growth dipterocarp-dominated stand at Lambir Hills, also in Borneo based on a combination of ground-based and tower-mounted terrestrial LiDAR, with a cumulative PAI above 2 m of ~8.4 m²/m². As the forest at Lambir Hills is similar in character to the old-growth forests in Maliau Basin and Danum Valley (Riutta et al., 2018), we assume a value of $urn:x-wiley:00218901:media:jpe13895:jpe13895-math-0003$ (0.50) that results in a mean estimated PAI for our old-growth plots that matches this value. This assumption means the reported PAI estimates carry additional uncertainty, and complicate the interpretation of absolute PAI values against other studies. However, for a given model, we anticipate that the relative changes observed across the degradation gradient are more robustly comparable.

We provide a sensitivity analysis of the LiDAR metrics to pulse density and spatial resolution in the supplement.

2.4 Comparison against inventory-based crown volume distributions

To compare the similarity of the LiDAR-derived PAD profiles against the crown volume distributions derived from the field inventory, we employ a simple profile overlap test. The individual 1-ha PAD and crown volume profiles are normalised by dividing through by the total plot PAI and crown volume respectively. We calculate profile overlap based on the percentage overlap between the two profiles. In the absence of harvested foliage profiles for validation, this approach provides a simple test of agreement between independent approaches to estimate the vertical distribution of vegetation in the canopy, noting that there is a significant uncertainty attached to the field-based distributions.

2.5 Assessing the diversity of canopy environments

To assess the diversity of canopy environments across the degradation gradient, we use the canopy Shannon Index, which has previously been used to relate the canopy structural diversity to tropical forest dynamics (Stark et al., 2012). The Shannon index increases with the number of canopy layers, and as PAD is distributed more evenly between layers:

$urn:x-wiley:00218901:media:jpe13895:jpe13895-math-0004$

3 RESULTS

3.1 LiDAR-derived PAI and vertical PAD distributions

LiDAR-estimated PAI is substantially lower in forest plots degraded by logging compared to reference old-growth plots (Table 2). The gradient in degradation intensity is marked by a trend in decreasing PAI as logging intensity increases, following a linear relationship with basal area (R² = 0.95; Figure 2). In Maliau Basin, PAI reached 8.7 m²/m², with similar PAI measured in the other old-growth plots. PAI declined by 28% and 52% in moderately and heavily logged forests, respectively, compared to the mean old-growth forest PAI.

TABLE 2. Summary of PAI estimates across the degradation gradient. Profile overlap represents the percentage overlap between normalised crown volume profiles and LiDAR PAD profiles

Plot	Forest type	Crown volume/(m3/ha)	LiDAR-based PAI/(m²/m²)	Profile overlap/%
Plot	Forest type	Mean ± SD (100 iterations)	Mean + SE (assuming κ = 0.44)	Profile overlap/%
MLA-01	Old-growth	14.9 ± 1.0	8.7 ± 0.3	84.3
MLA-02	Old-growth	12.7 ± 0.5	8.3 ± 0.3	81.0
DAN-04	Old-growth	10.0 ± 0.6	8.3 ± 0.2	80.7
DAN-05	Old-growth	11.6 ± 1.0	8.3 ± 0.4	76.1
SAF-03	Moderately logged	4.8 ± 0.2	5.7 ± 0.4	86.9
SAF-04	Moderately logged	8.9 ± 0.3	6.4 ± 0.4	87.2
SAF-02	Heavily logged	4.0 ± 0.1	4.3 ± 0.4	83.5
SAF-01	Heavily logged	3.5 ± 0.2	3.6 ± 0.4	79.5

Vertical PAD profiles also revealed striking structural changes in the canopy across the degradation gradient. Old-growth forest plots were characterised by structurally complex canopies, stretching to 70-m height. In contrast, there was an almost complete loss of canopy material >30–40 m in moderately logged plots (SAF-03 and SAF-04) and >20–30 m in heavily logged plots (SAF-01 and SAF-02; Figure 3; equivalent plots for DAN-04 and DAN-05 presented in Figure S1). Across all forest plots, PAD is distributed throughout the canopy, but highest in the mid-lower canopy (<30-m height). PAD contrasted strongly with the distribution of the original point clouds (Figure 3), reflecting the increased probability of interception of LiDAR pulses higher in the canopy.

3.2 Cross-comparison of LiDAR-estimated PAD profiles with field-based canopy models

Aggregated crown volume estimates across the degradation gradient ranged from ~3.5 m³/m² in the most heavily logged plots to >10-m³/m² in the old-growth plots. Canopy volumes corresponded closely with LiDAR-based PAI estimates (R² = 0.89; Table 2). Vertical crown volume distributions mirrored the first-order patterns observed in the LiDAR-derived PAD distributions, with the loss of crown volume >30–40 m in the moderately logged plots, and further declines in crown volume >20–30 m for heavily logged plots (Figure 3). The morphology of the 1-ha crown volume distributions was similar to the LiDAR-derived PAD profiles at heavily logged and moderately logged plots (profile overlap >76%; Table 2). While differences were greater at old-growth plots, crown volume was distributed throughout the vertical profile, and highest in the mid- and lower canopy, consistent with the LiDAR estimates.

3.3 Shifts in the diversity of canopy environments

The Shannon index distributions show that the diversity of canopy environments is diminished in logged forest, relative to old-growth forest (Figure 4). To a large part, this is driven by lower overall canopy height, and loss of associated structure above ~30 m, limiting the number of subcanopy environments. The difference in the availability of subcanopy environments across the degradation gradient is illustrated by considering the variation in subcanopy volume as a function of cumulative overlying PAD (Figure 5). Recasting the canopy profiles like this reveals both logged forest and old-growth forest have similar canopy volumes at their surfaces (i.e. little overlying vegetation), where light is abundant. In contrast, there is divergence in the availability of understorey environments. Old-growth forest is characterised by deep, shaded understories, with two to three times greater subcanopy volumes than logged forests for a given level of overlying canopy PAD; understorey volumes are most greatly reduced in heavily logged forests.

4 DISCUSSION

4.1 Canopy structure in Borneo's old-growth forest

Old-growth forest plots within the Maliau Basin and Danum Valley Conservation Areas were characterised by high vegetation densities (PAI > 8, assuming κ = 0.50), with the tallest trees reaching >70 m, overtopping a dense understorey <~30 m. Profiles retrieved for the old-growth plots in Maliau Basin and Danum Valley show significant variation both within and between plots (Figure 4; Figure S6); however, they all share the common feature of gradually increasing PAD with canopy profile depth, and notable increases in understorey PAD (below 20–30 m). This general pattern is consistent with canopy profiles published at another old-growth forest site in Western Borneo based on a combination of terrestrial LiDAR surveys undertaken at ground level and above the canopy (from a tower; Lambir Hills; Schneider et al., 2019). Our PAD profiles do not discriminate the leaves from woody vegetation (branches, twigs, trunks), which may contribute to ~20% of the total PAI in tropical forests (Olivas et al., 2013), and will contribute particularly to increased PAD estimates in the understorey (Schneider et al., 2019).

Our analysis in Borneo contrasts with canopy profiles in old-growth forest elsewhere in Equatorial regions. Canopy PAD profiles reported for old-growth forests in Central Amazonia (direct harvest and discrete-return LiDAR; Stark et al., 2012), Costa Rica (direct harvest: Clark et al., 2008; full-waveform LiDAR: Tang et al., 2012) and French Guiana (3D inversion of small-footprint waveform LiDAR; Vincent et al., 2017) often exhibit closed canopies with peaks in PAD at ~25–30 m, and differing levels of understorey density. Across these Neotropical sites, canopy heights are limited to between 30 and 50 m, thus foliage density is distributed along a shorter vertical dimension compared to Bornean old-growth forests. Variability in old-growth canopy structure may limit the extent to which we can translate the ecological impacts of degradation from one region to another.

4.2 Logging intensity drives first-order changes in canopy structure

Logging practices in Borneo typically involve removal of the largest trees (Slik et al., 2013). This logging strategy results in a steep decline in the abundance of large-basal area trees relative to smaller sized stems (Riutta et al., 2018). The impact of this logging strategy on canopy structure is striking. PAI drops as a function of logging intensity, and is >50% lower at the most heavily logged sites relative to the average for pristine old-growth forest (Figure 2). The tallest cohort of trees, contributing PAD above ~30-m height, is virtually absent from logged plots, an effect that has persisted more than a decade after the final round of logging. We know the impact of logging on canopy structure extends beyond the trees removed; felling frequently causes substantial collateral damage to surrounding trees (Pfeifer et al., 2015). There is also a significant shift in the allocation of productivity to stem-wood production at the expense of canopy allocation in the logged forest plots (Riutta et al., 2018). However, our results indicate that the removal of large trees appears to be the principal mechanism driving the first-order changes in PAD distributions. Full recovery to pre-disturbance canopy structure will therefore likely take decades (Cannon et al., 1994), requiring reestablishment of the largest stems. Importantly, this loss of PAD from the mid-upper canopy has not yet been compensated by a similar increase in understorey PAD (at least above the 2-m threshold employed in this study). We caveat this conclusion with the uncertainty attached to lower canopy PAD estimates, which are particularly pronounced for the old-growth sites (see Supporting Information). The consequences of this loss of foliage from the canopy include reduced shade and buffering of subcanopy microclimate (Hardwick et al., 2015; Jucker, Hardwick, et al., 2018) and lower productivity (Riutta et al., 2018).

The picture emerging from our logged plots in Borneo is one of the slow canopy recovery. These plots provide a snapshot of forest recovery following two to four rounds of selective logging, 11–14 years after logging finished (Pfeifer et al., 2015; Riutta et al., 2018). Neither PAI, vertical profiles, nor structural diversity have recovered in this period. This contrasts against relatively rapid rates of recovery of PAI observed at La Selva, Costa Rica (Tang et al., 2012). The forest at La Selva comprises a mixture of old-growth, selectively logged forests (>30 years post-logging, Clark et al., 2004) and secondary forest recovering from clear-felling. Using full-waveform airborne LiDAR to map PAD profiles and PAI, Tang et al. (2012) found the PAI of selectively logged forests was close to old-growth values (within 10%), although direct comparisons are complicated by differences in the time since logging ceased. However, the secondary forest chronosequence suggests swift recovery rates: median PAI in young secondary forests at La Selva (age 6–17 years) was ~40% lower than old-growth forest, but returned close to old-growth values within ~20 years of clear-felling (Tang et al., 2012). Differences in recovery rates may reflect the differences in old-growth forest structure (e.g. height of dominant species), but also regional differences in logging practices and intensity (Hosonuma et al., 2012). Further exploration of factors driving differential recovery rates is critical to understanding the long-term impacts of logging and the resilience of degraded tropical forests.

4.3 Logging generates long-lasting shifts in available subcanopy environments

Logging results in a significant contraction in the length of the light gradient and diversity of canopy environments (Figures 4 and 5). In particular, the absence of a deep, shaded understorey represents a fundamental difference between logged and old-growth forest. The shifts in canopy structure observed here are intuitive and consistent with a model of canopy dynamics whereby: following disturbance, a thicket of light-demanding vegetation becomes established, from which pioneer trees, such as Macaranga spp., emerge through rapid vertical growth, combining with remnant trees to form a new canopy (Slik et al., 2002). As light availability is reduced, the density of light-demanding vegetation in the understorey declines, and although shade-tolerant trees may be present, as recruits or pre-disturbance remnants, growth is slower (Nicotra et al., 1999), occurring sporadically in response to new canopy openings, while the canopy continues to stretch upwards (Farrior et al., 2016). The absence of extensive deeply shaded understorey environments in logged forests (Figure 5) may limit habitat availability for specialists of these low-light conditions (Deere et al., 2020). Our findings suggest that restoration of old-growth structure is only complete once PAD has recovered. Recovery involves vertical packing of the understorey with species that can survive and grow even in the deepest shade (Kabakoff & Chazdon, 1996), and the establishment of tall emergent trees. Based on the time-scale of recovery at this snapshot (at least 11–14 years), full recovery of canopy diversity at logged sites is likely to take several decades.

4.4 Implications for remote sensing of degradation in tropical forests

The relative changes in PAI observed (Figure 2) are over three times greater than those suggested by previous estimates of PAI across the degradation gradient at SAFE, estimated using 5-m resolution RapidEye imagery calibrated against PAI estimates from hemispherical photographs (Pfeifer et al., 2016). Pfeifer et al. (2016) suggest a more moderate decline of ~15% across the degradation gradient. This result may reflect the saturation of PAI estimates in hemispherical photographs in tropical forests (Vincent et al., 2017), reducing the apparent impact of degradation on PAI. The close, linear correspondence between LiDAR-estimated PAI and basal area across the gradient highlights LiDAR’s value for studies assessing the environmental impacts of degradation in dense tropical forests. Given the close relationship between basal area and standing carbon stocks, PAI-based basal area models may facilitate efforts to map above-ground carbon density (Asner & Mascaro, 2014; Jucker, Asner, et al., 2018).

4.5 Management implications

Selective logging is pervasive through many of the world's tropical forests (Asner et al., 2005; Gaveau et al., 2014). The impact of the attendant degradation is spatially heterogeneous and of varying intensity (Berry et al., 2008). Given that canopy structure underpins many aspects of ecosystem function (e.g. energy balance, photosynthesis, transpiration) and is a key determinant of habitat diversity, effective management of these forests requires detailed knowledge of how canopy structure varies in space and time (Struebig et al., 2013). We demonstrate that canopy profiles derived from airborne LiDAR capture the structural impacts of degradation at high resolution and accuracy. Importantly, LiDAR products enable an assessment of three-dimensional and subcanopy variation in foliage density that will improve the understanding of local variations in microclimate (Hardwick et al., 2015), light environment (Kumagai et al., 2001; Montgomery & Chazdon, 2001), productivity (Riutta et al., 2018) and the way that different plant and animal species make use of these forest environments (Deere et al., 2020). This study highlights the role of LiDAR, through mapping of the full canopy profile of plant area density, to delineate the areas of forest that promote positive ecosystem functions, such as biodiversity retention, at landscape scale. LiDAR mapping therefore has clear potential to help prioritise regions for conservation and restoration, and to maximise the benefits of such interventions (Deere et al., 2020).

5 CONCLUSIONS

We used airborne LiDAR to quantify canopy architecture adjustments associated with logging and at least 11–14 years of recovery in Borneo's ultra-complex tropical forest. We found a decline in PAI of ~28% in sites logged twice, and ~52% at sites logged four times, relative to old-growth forest. This sharp decline is associated with the near-complete loss of PAD above ~30-m, with further reductions in PAD above 10–15 m at high logging intensities. One impact of these structural changes is a drop in the diversity of canopy environments, in particular, the loss of a deep, shaded understorey. These results are consistent with shifts in allocation away from foliage and into stems in logged forests (Riutta et al., 2018), and suggest that full recovery of foliage density, and its vertical distribution, is likely to take decades, leaving a long-lived legacy of logging in recovering forests in Borneo.

PAI estimates across the eight plots exhibited a strong linear relationship with independent measurements of basal area (R² = 0.95), highlighting the value of LiDAR to quantify the degradation impacts in dense, complex tropical forests and improve the estimates of above-ground carbon stocks (Asner & Mascaro, 2014; Jucker, Asner, et al., 2018). The sensitivity of LiDAR-based PAD distributions to logging-driven changes in canopy structure will facilitate landscape-level descriptions of forest condition in high-biomass tropical forests. LiDAR mapping can therefore facilitate management of these forests by helping prioritise conservation and restoration efforts in a manner that maximises the benefits to ecosystem services (Deere et al., 2020). The dominant drivers of degradation (timber logging, fuelwood extraction, fire) vary from region to region (Hosonuma et al., 2012), with potentially distinct impacts on canopy structure (e.g. Almeida et al., 2016; Brando et al., 2019; Tang et al., 2012). Future studies should expand the use of airborne LiDAR across a wider range of environmental settings to understand the detectability and impact of natural and human disturbance on canopy structure, and the consequent effects on wider ecosystem functions.

ACKNOWLEDGEMENTS

D.T.M., M.W.: UK NERC (NE/K016458/1, NE/M017389/1) and NCEO; T.J.: UK NERC (NE/S01537X/1); T.S.: Frank Jackson Foundation; Y.M.: ERC (GEM-TRAIT; Grant 321131); M.S.: INGO II (LG15051); J.K.: IGA (LDF_VP_2016040). Thanks to: Sime Darby Foundation to the SAFE Project; Yayasan Sabah; Maliau Basin and Danum Valley Management Committees; NERC Airborne Remote Sensing Facility; NERC Data Analysis Node; Rostin Jantan; SAFE Carbon Team; Danum 50-ha plot team; Laura Kruitbos, Unding Jami, Alexander Karolus, Ryan Gray and Reuben Nilus. Finally, we are very grateful for the comments from the Associate Editor and two anonymous reviewers.

AUTHORS' CONTRIBUTIONS

D.T.M. and M.W. designed the research, with input from D.A.C., T.J. and T.S.; D.T.M. conducted the analysis; D.A.C. and T.J. provided the LiDAR point cloud data; T.R., Y.M. and D.F.R.P.B. provided the stem inventories; M.S. and J.K. undertook the ground-based canopy surveys; Y.A.T., M.W., D.A.C., R.M.E., Y.M. and D.F.R.P.B. organised the wider research initiative; D.T.M. wrote the paper, with input from M.W. and T.S. and contributions from all the co-authors. All the authors approved the final manuscript.

Open Research

DATA AVAILABILITY STATEMENT

Source code (python) is available to via GitHub: https://github.com/DTMilodowski/LiDAR_canopy_structure_JAppEcol. The data are available via the SAFE archive: https://doi.org/10.5281/zenodo.4572775 (Milodowski et al., 2021).

Supporting Information

REFERENCES

Armston, J., Disney, M., Lewis, P., Scarth, P., Phinn, S., Lucas, R., Bunting, P., & Goodwin, N. (2013). Direct retrieval of canopy gap probability using airborne waveform lidar. Remote Sensing of Environment, 134, 24–38. https://doi.org/10.1016/j.rse.2013.02.021
10.1016/j.rse.2013.02.021
Web of Science®Google Scholar
Asner, G. P., Brodrick, P. G., Philipson, C., Vaughn, N. R., Martin, R. E., Knapp, D. E., Heckler, J., Evans, L. J., Jucker, T., Goossens, B., Stark, D. J., Reynolds, G., Ong, R., Renneboog, N., Kugan, F., & Coomes, D. A. (2018). Mapped aboveground carbon stocks to advance forest conservation and recovery in Malaysian Borneo. Biological Conservation, 217, 289–310. https://doi.org/10.1016/j.biocon.2017.10.020
10.1016/j.biocon.2017.10.020
Web of Science®Google Scholar
Asner, G. P., Knapp, D. E., Broadbent, E. N., Oliveira, P. J. C., Keller, M., & Silva, J. N. (2005). Selective logging in the Brazilian Amazon. Science, 310(5747), 480–482. https://doi.org/10.1126/science.1118051
10.1126/science.1118051
CASPubMedWeb of Science®Google Scholar
Asner, G. P., & Mascaro, J. (2014). Mapping tropical forest carbon: Calibrating plot estimates to a simple LiDAR metric. Remote Sensing of Environment, 140, 614–624. https://doi.org/10.1016/j.rse.2013.09.023
10.1016/j.rse.2013.09.023
Web of Science®Google Scholar
Berry, N. J., Phillips, O. L., Ong, R. C., & Hamer, K. C. (2008). Impacts of selective logging on tree diversity across a rainforest landscape: The importance of spatial scale. Landscape Ecology, 23(8), 915–929. https://doi.org/10.1007/s10980-008-9248-1
10.1007/s10980?008?9248?1
Web of Science®Google Scholar
Both, S., Riutta, T., Paine, C. E. T., Elias, D. M. O., Cruz, R. S., Jain, A., Johnson, D., Kritzler, U. H., Kuntz, M., Majalap-Lee, N., Mielke, N., Montoya Pillco, M. X., Ostle, N. J., Arn Teh, Y., Malhi, Y., & Burslem, D. F. R. P. (2019). Logging and soil nutrients independently explain plant trait expression in tropical forests. New Phytologist, 221(4), 1853–1865. https://doi.org/10.1111/nph.15444
10.1111/nph.15444
CASPubMedWeb of Science®Google Scholar
Brando, P. M., Silvério, D., Maracahipes-Santos, L., Oliveira-Santos, C., Levick, S. R., Coe, M. T., Migliavacca, M., Balch, J. K., Macedo, M. N., Nepstad, D. C., Maracahipes, L., Davidson, E., Asner, G., Kolle, O., Trumbore, S. (2019). Prolonged tropical forest degradation due to compounding disturbances: Implications for CO₂ and H₂O fluxes. Global Change Biology, 25(9), 2855–2868. https://doi.org/10.1111/gcb.14659
10.1111/gcb.14659
PubMedWeb of Science®Google Scholar
Bréda, N. J. J. (2003). Ground-based measurements of leaf area index: A review of methods, instruments and current controversies. Journal of Experimental Botany, 54(392), 2403–2417. https://doi.org/10.1093/jxb/erg263
10.1093/jxb/erg263
CASPubMedWeb of Science®Google Scholar
Cannon, C. H., Peart, D. R., Leighton, M., & Kartawinata, K. (1994). The structure of lowland rainforest after selective logging in West Kalimantan. Indonesia. Forest Ecology and Management, 67(1), 49–68. https://doi.org/10.1016/0378-1127(94)90007-8
10.1016/0378?1127(94)90007?8
Web of Science®Google Scholar
Clark, D. B., Olivas, P. C., Oberbauer, S. F., Clark, D. A., & Ryan, M. G. (2008). First direct landscape-scale measurement of tropical rain forest Leaf Area Index, a key driver of global primary productivity. Ecology Letters, 11(2), 163–172. https://doi.org/10.1111/j.1461-0248.2007.01134.x
10.1111/j.1461-0248.2007.01134.x
PubMedWeb of Science®Google Scholar
Clark, D. B., Read, J. M., Clark, M. L., Cruz, A. M., Dotti, M. F., & Clark, D. A. (2004). Application of 1-M and 4-M resolution satellite data to ecological studies of tropical rain forests. Ecological Applications, 14(1), 61–74. https://doi.org/10.1890/02-5120
10.1890/02-5120
Web of Science®Google Scholar
Coomes, D. A., Kunstler, G., Canham, C. D., & Wright, E. (2009). A greater range of shade-tolerance niches in nutrient-rich forests: An explanation for positive richness–productivity relationships? Journal of Ecology, 97(4), 705–717. https://doi.org/10.1111/j.1365-2745.2009.01507.x
10.1111/j.1365?2745.2009.01507.x
Web of Science®Google Scholar
Coops, N. C., Hilker, T., Wulder, M. A., St-Onge, B., Newnham, G., Siggins, A., Trofymow, & J. A. (.. (2007). Estimating canopy structure of Douglas-fir forest stands from discrete-return LiDAR. Trees, 21(3), 295. https://doi.org/10.1007/s00468-006-0119-6
10.1007/s00468?006?0119?6
Google Scholar
de Almeida, D. R. A., Nelson, B. W., Schietti, J., Gorgens, E. B., Resende, A. F., Stark, S. C., & Valbuena, R. (2016). Contrasting fire damage and fire susceptibility between seasonally flooded forest and upland forest in the Central Amazon using portable profiling LiDAR. Remote Sensing of Environment, 184(Suppl C), 153–160. https://doi.org/10.1016/j.rse.2016.06.017
10.1016/j.rse.2016.06.017
Web of Science®Google Scholar
Deere, N. J., Guillera-Arroita, G., Swinfield, T., Milodowski, D. T., Coomes, D. A., Bernard, H., Reynolds, G., Davies, Z. G., & Struebig, M. J. (2020). Maximizing the value of forest restoration for tropical mammals by detecting three-dimensional habitat associations. Proceedings of the National Academy of Sciences of the United States of America, 117(42), 26254–26262. https://doi.org/10.1073/pnas.2001823117
10.1073/pnas.2001823117
CASPubMedWeb of Science®Google Scholar
Detto, M., Asner, G. P., Muller-Landau, H. C., & Sonnentag, O. (2015). Spatial variability in tropical forest leaf area density from multireturn lidar and modeling. Journal of Geophysical Research Biogeosciences, 120(2), 2014JG002774. https://doi.org/10.1002/2014JG002774
10.1002/2014JG002774
Google Scholar
Ellsworth, D. S., & Reich, P. B. (1993). Canopy structure and vertical patterns of photosynthesis and related leaf traits in a deciduous forest. Oecologia, 96(2), 169–178. https://doi.org/10.1007/BF00317729
10.1007/BF00317729
CASPubMedWeb of Science®Google Scholar
Ewers, R. M., Didham, R. K., Fahrig, L., Ferraz, G., Hector, A., Holt, R. D., Kapos, V., Reynolds, G., Sinun, W., Snaddon, J. L., & Turner, E. C. (2011). A large-scale forest fragmentation experiment: The Stability of Altered Forest Ecosystems Project. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 366(1582), 3292–3302. https://doi.org/10.1098/rstb.2011.0049
10.1098/rstb.2011.0049
PubMedWeb of Science®Google Scholar
Falster, D. S., Duursma, R. A., Ishihara, M. I., Barneche, D. R., FitzJohn, R. G., Vårhammar, A., Aiba, M., Ando, M., Anten, N., Aspinwall, M. J., Baltzer, J. L., Baraloto, C., Battaglia, M., Battles, J. J., Bond-Lamberty, B., van Breugel, M., Camac, J., Claveau, Y., Coll, L., … York, R. A. (2015). BAAD: A biomass and allometry database for woody plants. Ecology, 96(5), 1445. https://doi.org/10.1890/14-1889.1
10.1890/14-1889.1
Web of Science®Google Scholar
Farrior, C. E., Bohlman, S. A., Hubbell, S., & Pacala, S. W. (2016). Dominance of the suppressed: Power-law size structure in tropical forests. Science, 351(6269), 155–157. https://doi.org/10.1126/science.aad0592
10.1126/science.aad0592
CASPubMedWeb of Science®Google Scholar
Flack-Prain, S., Meir, P., Malhi, Y., Smallman, T. L., & Williams, M. (2019). The importance of physiological, structural and trait responses to drought stress in driving spatial and temporal variation in GPP across Amazon forests. Biogeosciences, 16(22), 4463–4484. https://doi.org/10.5194/bg-16-4463-2019
10.5194/bg-16-4463-2019
CASWeb of Science®Google Scholar
Gaveau, D. L. A., Sheil, D., Husnayaen, Salim, M. A., Arjasakusuma, S., Ancrenaz, M., Pacheco, P., & Meijaard, E. (2016). Rapid conversions and avoided deforestation: Examining four decades of industrial plantation expansion in Borneo. Scientific Reports, 6(1), 1–13. https://doi.org/10.1038/srep32017
10.1038/srep32017
PubMedWeb of Science®Google Scholar
Gaveau, D. L. A., Sloan, S., Molidena, E., Yaen, H., Sheil, D., Abram, N. K., Ancrenaz, M., Nasi, R., Quinones, M., Wielaard, N., & Meijaard, E. (2014). Four decades of forest persistence, clearance and logging on borneo. PLoS ONE, 9(7), e101654. https://doi.org/10.1371/journal.pone.0101654
10.1371/journal.pone.0101654
PubMedWeb of Science®Google Scholar
Gower, S. T., Kucharik, C. J., & Norman, J. M. (1999). Direct and indirect estimation of leaf area index, fAPAR, and net primary production of terrestrial ecosystems. Remote Sensing of Environment, 70(1), 29–51. https://doi.org/10.1016/S0034-4257(99)00056-5
10.1016/S0034-4257(99)00056-5
Web of Science®Google Scholar
Hardwick, S. R., Toumi, R., Pfeifer, M., Turner, E. C., Nilus, R., & Ewers, R. M. (2015). The relationship between leaf area index and microclimate in tropical forest and oil palm plantation: Forest disturbance drives changes in microclimate. Agricultural and Forest Meteorology, 201, 187–195. https://doi.org/10.1016/j.agrformet.2014.11.010
10.1016/j.agrformet.2014.11.010
PubMedWeb of Science®Google Scholar
Hosonuma, N., Herold, M., De Sy, V., De Fries, R. S., Brockhaus, M., Verchot, L., Angelsen, A., & Romijn, E. (2012). An assessment of deforestation and forest degradation drivers in developing countries. Environmental Research Letters, 7(4), 044009. https://doi.org/10.1088/1748-9326/7/4/044009
10.1088/1748-9326/7/4/044009
Web of Science®Google Scholar
Houghton, R. A. (2013). The emissions of carbon from deforestation and degradation in the tropics: Past trends and future potential. Carbon Management, 4(5), 539–546. https://doi.org/10.4155/cmt.13.41
10.4155/cmt.13.41
CASWeb of Science®Google Scholar
Joshi, N., Mitchard, E. T. A., Brolly, M., Schumacher, J., Fernández-Landa, A., Johannsen, V. K., Marchamalo, M., & Fensholt, R. (2017). Understanding ‘saturation’ of radar signals over forests. Scientific Reports, 7(1), 1–11. https://doi.org/10.1038/s41598-017-03469-3
10.1038/s41598?017?03469?3
PubMedWeb of Science®Google Scholar
Jucker, T., Asner, G. P., Dalponte, M., Brodrick, P. G., Philipson, C. D., Vaughn, N. R., Teh, Y. A., Brelsford, C., Burslem, D. F. R. P., Deere, N. J., Ewers, R. M., Kvasnica, J., Lewis, S. L., Malhi, Y., Milne, S., Nilus, R., Pfeifer, M., Phillips, O. L., Qie, L., … Coomes, D. A. (2018). Estimating aboveground carbon density and its uncertainty in Borneo’s structurally complex tropical forests using airborne laser scanning. Biogeosciences, 15(12), 3811–3830. https://doi.org/10.5194/bg-15-3811-2018
10.5194/bg?15?3811?2018
Web of Science®Google Scholar
Jucker, T., Hardwick, S. R., Both, S., Elias, D. M. O., Ewers, R. M., Milodowski, D. T., Swinfield, T., & Coomes, D. A. (2018). Canopy structure and topography jointly constrain the microclimate of human-modified tropical landscapes. Global Change Biology, 24(11), 5243–5258. https://doi.org/10.1111/gcb.14415
10.1111/gcb.14415
PubMedWeb of Science®Google Scholar
Kabakoff, R. P., & Chazdon, R. L. (1996). Effects of canopy species dominance on understorey light availability in low-elevation secondary forest stands in Costa Rica. Journal of Tropical Ecology, 12(6), 779–788. https://doi.org/10.1017/S0266467400010038
10.1017/S0266467400010038
Web of Science®Google Scholar
Knapp, N., Fischer, R., & Huth, A. (2018). Linking lidar and forest modeling to assess biomass estimation across scales and disturbance states. Remote Sensing of Environment, 205, 199–209. https://doi.org/10.1016/j.rse.2017.11.018
10.1016/j.rse.2017.11.018
Web of Science®Google Scholar
Kohyama, T., Suzuki, E., Partomihardjo, T., Yamada, T., & Kubo, T. (2003). Tree species differentiation in growth, recruitment and allometry in relation to maximum height in a Bornean mixed dipterocarp forest. Journal of Ecology, 91(5), 797–806. https://doi.org/10.1046/j.1365-2745.2003.00810.x
10.1046/j.1365-2745.2003.00810.x
Web of Science®Google Scholar
Kumagai, T., Kuraji, K., Noguchi, H., Tanaka, Y., Tanaka, K., & Suzuki, M. (2001). Vertical profiles of environmental factors within tropical rainforest, Lambir Hills National Park, Sarawak, Malaysia. Journal of Forest Research, 6(4), 257–264. https://doi.org/10.1007/BF02762466
10.1007/BF02762466
Google Scholar
MacArthur, R. H., & Horn, H. S. (1969). Foliage profile by vertical measurements. Ecology, 50(5), 802–804. https://doi.org/10.2307/1933693
10.2307/1933693
Google Scholar
Marthews, T. R., Riutta, T., Oliveras Menor, I., Urrutia, R., Moore, S., Metcalfe, D., Malhi, Y., Phillips, O., Huarasca Huasco, W., Ruiz Jaén, M., Girardin, C., Butt, N., Cain, R., & colleagues from the RAINFOR and GEM networks. (2014). Measuring Tropical Forest Carbon Allocation and Cycling: A RAINFOR-GEM Field Manual for Intensive Census Plots (v3.0). Manual, Global Ecosystems Monitoring network. http://gem.tropicalforests.ox.ac.uk/
Google Scholar
Milodowski, D. T., Coomes, D. A., Swinfield, T., Jucker, T., Riutta, T., Malhi, Y., Svátek, M., Kvasnica, J., Reuben, N., Burslem, D. F. R. P., Ewers, R. M., Teh, Y. A., & Williams, M. (2021). Data from: The impact of logging on vertical canopy structure across a gradient of tropical forest degradation intensity in Borneo. Zenodo, https://doi.org/10.5281/zenodo.4572775
10.5281/zenodo.4572775
Google Scholar
Milodowski, D. T., Mitchard, E. T. A., & Williams, M. (2017). Forest loss maps from regional satellite monitoring systematically underestimate deforestation in two rapidly changing parts of the Amazon. Environmental Research Letters, 12(9), 094003. https://doi.org/10.1088/1748-9326/aa7e1e
10.1088/1748-9326/aa7e1e
Web of Science®Google Scholar
Mitchard, E. T. A. (2018). The tropical forest carbon cycle and climate change. Nature, 559(7715), 527. https://doi.org/10.1038/s41586-018-0300-2
10.1038/s41586-018-0300-2
CASPubMedWeb of Science®Google Scholar
Montgomery, R. A., & Chazdon, R. L. (2001). Forest structure, canopy architecture, and light transmittance in tropical wet forests. Ecology, 82(10), 2707–2718. https://doi.org/10.1890/0012-9658(2001)082[2707:FSCAAL]2.0.CO;2
10.1890/0012?9658(2001)082[2707:FSCAAL]2.0.CO;2
Web of Science®Google Scholar
Nicotra, A. B., Chazdon, R. L., & Iriarte, S. V. B. (1999). Spatial heterogeneity of light and woody seedling regeneration in tropical wet forests. Ecology, 80(6), 1908–1926. https://doi.org/10.1890/0012-9658(1999)080[1908:SHOLAW]2.0.CO;2
10.1890/0012-9658(1999)080[1908:SHOLAW]2.0.CO;2
Web of Science®Google Scholar
Ni-Meister, W., Jupp, D. L. B., & Dubayah, R. (2001). Modeling lidar waveforms in heterogeneous and discrete canopies. IEEE Transactions on Geoscience and Remote Sensing, 39(9), 1943–1958. https://doi.org/10.1109/36.951085
10.1109/36.951085
Web of Science®Google Scholar
Olivas, P. C., Oberbauer, S. F., Clark, D. B., Clark, D. A., Ryan, M. G., O’Brien, J. J., & Ordoñez, H. (2013). Comparison of direct and indirect methods for assessing leaf area index across a tropical rain forest landscape. Agricultural and Forest Meteorology, 177, 110–116. https://doi.org/10.1016/j.agrformet.2013.04.010
10.1016/j.agrformet.2013.04.010
Web of Science®Google Scholar
Ordway, E. M., & Asner, G. P. (2020). Carbon declines along tropical forest edges correspond to heterogeneous effects on canopy structure and function. Proceedings of the National Academy of Sciences of the United States of America, 117(14), 7863–7870. https://doi.org/10.1073/pnas.1914420117
10.1073/pnas.1914420117
CASPubMedWeb of Science®Google Scholar
Pfeifer, M., Kor, L., Nilus, R., Turner, E., Cusack, J., Lysenko, I., Khoo, M., Chey, V. K., Chung, A. C., & Ewers, R. M. (2016). Mapping the structure of Borneo’s tropical forests across a degradation gradient. Remote Sensing of Environment, 176, 84–97. https://doi.org/10.1016/j.rse.2016.01.014
10.1016/j.rse.2016.01.014
Web of Science®Google Scholar
Pfeifer, M., Lefebvre, V., Turner, E., Cusack, J., Khoo, M. S., Chey, V. K., Peni, M., & Ewers, R. M. (2015). Deadwood biomass: An underestimated carbon stock in degraded tropical forests? Environmental Research Letters, 10(4), 044019. https://doi.org/10.1088/1748-9326/10/4/044019
10.1088/1748-9326/10/4/044019
Web of Science®Google Scholar
Riutta, T., Malhi, Y., Kho, L. K., Marthews, T. R., Huaraca Huasco, W., Khoo, M. S., Tan, S., Turner, E., Reynolds, G., Both, S., Burslem, D. F. R. P., Teh, Y. A., Vairappan, C. S., Majalap, N., & Ewers, R. M. (2018). Logging disturbance shifts net primary productivity and its allocation in Bornean tropical forests. Global Change Biology, 24(7), 2913–2928. https://doi.org/10.1111/gcb.14068
10.1111/gcb.14068
ADSPubMedWeb of Science®Google Scholar
Schneider, F. D., Kükenbrink, D., Schaepman, M. E., Schimel, D. S., & Morsdorf, F. (2019). Quantifying 3D structure and occlusion in dense tropical and temperate forests using close-range LiDAR. Agricultural and Forest Meteorology, 268, 249–257. https://doi.org/10.1016/j.agrformet.2019.01.033
10.1016/j.agrformet.2019.01.033
Web of Science®Google Scholar
Slik, J. W. F., Paoli, G., McGuire, K., Amaral, I., Barroso, J., Bastian, M., Blanc, L., Bongers, F., Boundja, P., Clark, C., Collins, M., Dauby, G., Ding, Y. I., Doucet, J.-L., Eler, E., Ferreira, L., Forshed, O., Fredriksson, G., Gillet, J.-F., … Zweifel, N. (2013). Large trees drive forest aboveground biomass variation in moist lowland forests across the tropics. Global Ecology and Biogeography, 22(12), 1261–1271. https://doi.org/10.1111/geb.12092
10.1111/geb.12092
Web of Science®Google Scholar
Slik, J. W. F., Verburg, R. W., & Kessler, P. J. A. (2002). Effects of fire and selective logging on the tree species composition of lowland dipterocarp forest in East Kalimantan. Indonesia. Biodiversity and Conservation, 11(1), 85–98.
10.1023/A:1014036129075
Web of Science®Google Scholar
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. (2012). Amazon forest carbon dynamics predicted by profiles of canopy leaf area and light environment. Ecology Letters, 15(12), 1406–1414. https://doi.org/10.1111/j.1461-0248.2012.01864.x
10.1111/j.1461-0248.2012.01864.x
PubMedWeb of Science®Google Scholar
Struebig, M. J., Turner, A., Giles, E., Lasmana, F., Tollington, S., Bernard, H., & Bell, D. (2013). Quantifying the biodiversity value of repeatedly logged rainforests: Gradient and comparative approaches from Borneo. Advances in Ecological Research, 48, 183–224.
10.1016/B978-0-12-417199-2.00003-3
Web of Science®Google Scholar
Tang, H., Dubayah, R., Swatantran, A., Hofton, M., Sheldon, S., Clark, D. B., & Blair, B. (2012). Retrieval of vertical LAI profiles over tropical rain forests using waveform lidar at La Selva, Costa Rica. Remote Sensing of Environment, 124, 242–250. https://doi.org/10.1016/j.rse.2012.05.005
10.1016/j.rse.2012.05.005
Web of Science®Google Scholar
Vincent, G., Antin, C., Laurans, M., Heurtebize, J., Durrieu, S., Lavalley, C., & Dauzat, J. (2017). Mapping plant area index of tropical evergreen forest by airborne laser scanning. A cross-validation study using LAI2200 optical sensor. Remote Sensing of Environment, 198, 254–266. https://doi.org/10.1016/j.rse.2017.05.034
10.1016/j.rse.2017.05.034
Web of Science®Google Scholar
Watson, D. J. (1947). Comparative physiological studies on the growth of field crops: I. Variation in net assimilation rate and leaf area between species and varieties, and within and between years. Annals of Botany, 11(41), 41–76.
10.1093/oxfordjournals.aob.a083148
CASWeb of Science®Google Scholar

Citing Literature

Volume58, Issue8

August 2021

Pages 1764-1775

Journal list menu

Tools

Follow journal

The impact of logging on vertical canopy structure across a gradient of tropical forest degradation intensity in Borneo

Abstract

1 INTRODUCTION