Biological Conservation 220 (2018) 245–253 Contents lists available at ScienceDirect Biological Conservation journal homepage: www.elsevier.com/locate/biocon Do responsibly managed logging concessions adequately protect jaguars and T other large and medium-sized mammals? Two case studies from Guatemala and Peru ⁎ Mathias W. Toblera, , Rony Garcia Anleub, Samia E. Carrillo-Percasteguia, Gabriela Ponce Santizob, John Polisarc, Alfonso Zuñiga Hartleya,d, Isaac Goldsteinc a San Diego Zoo Global, Institute for Conservation Research, 15600 San Pasqual Valley Rd., Escondido, CA 92027, USA b Wildlife Conservation Society – Guatemala Program, Avenida 15 de Marzo, Casa #3, Ciudad de Flores, Petén 17001, Guatemala c Wildlife Conservation Society, Southern Boulevard, Bronx, NY 10460, USA d Servicio Nacional Forestal y de Fauna Silvestre, Avenida Siete #229, Urb. Rinconada Baja, La Molina, Lima, Peru A R T I C L E I N F O A B S T R A C T Keywords: Large areas of tropical forest have been designated for timber production but logging practices vary widely. Reduced-impact logging Reduced-impact logging is considered best practice and third-party certification aims to ensure that strict Multi-species occupancy model standards are met. This includes minimizing the number of roads constructed, avoiding sensitive areas and Spatial capture-recapture strictly regulating hunting. Large scale camera trap grids were utilized in Guatemala and Peru to evaluate the Madre de Dios impact of reduced-impact logging in certified concessions upon the large and medium-sized mammal fauna with Petén Maya Biosphere Reserve special emphasis on jaguars (Panthera onca). Spatial capture-recapture models showed that jaguar density in Camera traps Peru (4.54 ± 0.83 ind. 100 km−2) was significantly higher than in Guatemala (1.52 ± 0.34 ind. 100 km−2) but Forest management in both regions, densities were comparable to protected areas. Camera traps detected 22 species of large and medium sized mammals in Guatemala and 27 in Peru and a multi-species occupancy model revealed that logging had no negative impact on any of the species studied and actually had an initial positive impact on several herbivore species. We found no avoidance of logging roads; in fact, many species, especially carnivores, fre- quently used logging roads as movement corridors. Our results indicate that well-managed logging concessions can maintain important populations of large and medium-sized mammals including large herbivores and large carnivores as long as hunting is controlled and timber volumes extracted are low. Responsible forest manage- ment would therefore be an ideal activity in the buffer zones and multiple use zones of protected areas creating much less impact and conflict than alternatives such as agriculture or cattle ranching while still providing economic opportunities. Logging concessions can also play an important role in maintaining landscape con- nectivity between protected areas. 1. Introduction Management practices of logging operations vary greatly, ranging from clear-cutting to selective reduced-impact logging. Many countries Over the last few decades deforestation of humid tropical forests have established forest reserves, logging concessions or other systems around the world has continuously increased (Achard et al., 2014; for leasing state owned forests to private companies for the extraction Asner et al., 2009; Hansen et al., 2013). In Southeast Asia and Central of timber with the goal of managing these forests sustainably for long- America over 70% of the original humid tropical forest has been lost or term production (Blaser et al., 2011). Forest certification was created as greatly degraded (< 50% tree cover) and in South America this figure is an independent third-party verification of responsible forest manage- 36% (Asner et al., 2009). About 40% of the remaining forests are af- ment with strict standards. The Forest Stewardship Council (FSC), fected by commercial logging that often leads to forest degradation, loss which was established in 1993, has a global forest certification system of carbon stock, increased vulnerability to fire and increased access to that accredits companies that use sound social and environmental such areas by hunters and small farmers (Asner et al., 2009; Blaser practices for forest management (FSC, 2016). FSC-certified logging et al., 2011; Laurance et al., 2014; Nepstad et al., 1999). operations are required to practice reduced impact logging, control or ⁎ Corresponding author. E-mail address: mtobler@sandiegozoo.org (M.W. Tobler). https://doi.org/10.1016/j.biocon.2018.02.015 Received 8 September 2017; Received in revised form 12 January 2018; Accepted 9 February 2018 0006-3207/ © 2018 Elsevier Ltd. All rights reserved. M.W. Tobler et al. Biological Conservation 220 (2018) 245–253 prohibit hunting within the concession, set aside high conservation in two FSC certified logging concessions (Forestal Otorongo and value forest, and avoid, repair, or mitigate environmental impacts (FSC, Aserradero Espinoza) south of the Tahuamanu river (Fig. 2). These 2016). While this comes at a significant cost for the logging company concessions are part of a large block of logging concessions towards the (Gullison, 2003), certified wood sells for a higher price in international north, south and west and are bordered by agriculture land and Brazil markets and companies get increased market access, resulting in a net nut concessions (for the extraction of Brazil nut from mature forests) to financial benefit. Although the global area of certified forests has con- the east. Logging in these concessions started in 2003 but was preceded tinuously increased over the last two decades, the largest increases by unregulated selective extraction of mahogany and a few other high- happened in boreal forests in Europe and North America (FSC, 2016; value hard-wood species for almost a decade. The average volume of PEFC, 2017). In 2011, only 13% of tropical forests were considered timber extracted from the concessions is between 2 and 3 m3/ha. sustainably managed and only 4–5% were certified (Blaser et al., 2011). Hunting is strictly prohibited within the concessions. The conservation of biodiversity is an explicit goal of the FSC cer- The topography is flat with elevation ranging from 150 m to 300 m tification scheme (FSC, 2016), but the number of studies that have and the vegetation is lowland Amazonian moist forest with several evaluated how well certified forest management under the FSC label areas dominated by large patches of bamboo. The mean annual tem- protects biodiversity are few in number. Several studies have looked at perature is 24 °C and mean annual rainfall is between 2500 and responses by either single species or small numbers of species of large 3500 mm. and medium-sized mammals to certified forest management ap- proaches, leading to the generation of management recommendations 2.1.2. Guatemala (Clark et al., 2009; Davies et al., 2001; Polisar et al., 2017; Rayan and Over a thousand years ago Guatemala's lowland Department of Mohamad, 2009), but far fewer have examined the effects on the entire Petén was the epicenter of the Maya culture. In the 20th century, the mammal community (Roopsind et al., 2017; Sollmann et al., 2017). economy of the northern Petén was dominated by extraction of gum Moreover, several studies have found negative impacts of logging on from chicle trees (Manilkara zapota), a market that has since dwindled. species richness with effects varying greatly by taxonomic group, geo- Until recently, this, the largest of Guatemala's 22 departments was graphic region, and logging intensity (Burivalova et al., 2014; isolated from the rest of the country due to the lack of well-maintained Chaudhary et al., 2016; Gibson et al., 2011). For tropical forests, re- access routes and long distances from principal cities (Hodgdon et al., duced impact logging has been found to have the least negative effect, 2015). with some forests under reduced-impact logging retaining between 80% In 1990, the Guatemalan government via the Consejo Nacional de and 100% of their species richness (Bicknell et al., 2014; Chaudhary Áreas Protegidas (CONAP, Guatemala's National Council of Protected et al., 2016; Gibson et al., 2011; Putz et al., 2012). Areas) created the Maya Biosphere Reserve (MBR) in the northern In this study, we use large-scale camera trap surveys to evaluate portion of the Petén with the goal of “combining the conservation and terrestrial mammal communities in FSC certified logging concessions in sustainable use of natural and cultural resources in order to maximize Guatemala and Peru. Camera traps are ideally suited to assess mammal the ecological, economic and social benefits for Guatemala” (Secaira communities in tropical forests and, unlike other methods such as line et al., 2015). The reserve was divided into three zones: (a) the core zone transects, they are also able to collect data on cryptic and nocturnal (36% of the MBR) is formed by national parks where only scientific species (Ahumada et al., 2013; Tobler et al., 2008; Tobler et al., 2015). investigation and low impact tourism are allowed, (b) a 15 km-wide We used multi-species occupancy models (Dorazio and Royle, 2005; buffer zone (24% of the MBR) along the southern border of the MBR Dorazio et al., 2006; Yamaura et al., 2011) to examine community where agriculture, farming, and other productive activities are per- structure and distribution of mammals in the logging concessions, and mitted with the aim of reducing the pressure on the other two zones, assessed the density of the top predator, the jaguar, using spatial cap- and (c) a 848,440 ha multiple use zone (40% of the MBR) where sus- ture-recapture models (Borchers and Efford, 2008; Efford et al., 2009; tainable and low-impact land uses are allowed including controlled Royle and Young, 2008). logging of hardwood tree species in forest concessions (Hodgdon et al., 2015; Radachowsky et al., 2012; Secaira et al., 2015). 2. Methods Between 1994 and 2002, CONAP granted 533,132 ha of the multiple use zone (MUZ) of the MBR to 14 forest concessions for a period of 2.1. Study areas 25 years. They included two industrial concessions (private companies), six non-resident community concessions (communities in the buffer 2.1.1. Peru zone), two resident community concessions with forest-based history Peru has 62.5 million ha of lowland tropical rainforest with his- (communities established as chicle harvesting centers more than a torically low annual deforestation rates (around 0.2% per year between century ago) and four resident community concessions for recent im- 1990 and 2015 (FAO, 2015)). In 2000, the Peruvian government passed migrants (Hodgdon et al., 2015; Radachowsky et al., 2012; Secaira a new law of Forestry and Wildlife (Ley Forestal y de Fauna Silvestre, et al., 2015). Three of the four resident community concessions for Ley N° 27308) that designated about 8 million ha of permanent pro- recent immigrants were cancelled or suspended due to a lack of com- duction forest. Within these areas the government can grant conces- pliance with the contract agreements while the other 11 concessions sions of between 5000 and 50,000 ha for durations of up to 40 years. were granted FSC certification between 1998 and 2004 (Carrera et al., The concession holders are required to develop a five-year management 2006; Hodgdon et al., 2015; Radachowsky et al., 2012). The harvest plan and an annual operating plan in which they agree to restrictions intensities in these concessions (1.2–3.0 m3/ha) are among the lowest including limits on timber extraction to 5% of the available basal area in the world. and limits on subsistence hunting (commercial hunting is strictly pro- Our study was carried out in the territory of five non-resident hibited). Each concession is divided into 20 blocks representing a 20- community concessions: La Unión, Las Ventanas, Chosquitan, Rio year harvest cycle with timber being extracted from one block annually. Chanchic and Yaloch managed by Sociedad Civil Custodios de la Selva, In the department of Madre de Dios there are 1.3 million ha of Árbol Verde, Sociedad Civil Laborantes del Bosque, Sociedad Civil logging concessions of which 422,959 ha are FSC certified (DGFFS, Impulsores Suchitecos, and Sociedad Civil El Esfuerzo respectively 2013). These concessions go through an annual review process under- (Fig. 1). These concessions are exclusively used for logging; no people taken by an outside certification organization that evaluates com- besides the workers are living inside the concessions and there is no pliance with all the FSC standards to ensure sustainable management hunting. Forty-three percent of the study area was harvested before the practices. sampling period and the entire area reported 0% of deforestation Our study was carried out in the north-eastern part of Madre de Dios during 2000–2013 (Hodgdon et al., 2015). The MBR is classified as 246 M.W. Tobler et al. Biological Conservation 220 (2018) 245–253 89°30’0"W 89°15’0"W 17°30’0"N 17°30’0"N Camera Traps Roads Year Logged Unlogged 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 17°15’0"N 17°15’0"N 2010 2011 2012 2013 0 5 10 km 89°30’0"W 89°15’0"W Fig. 1. Guatemala camera trap locations and logging concessions. Colors of logging blocks indicate what year trees were harvested. Unlogged areas are as of 2013. Petén-Veracruz Moist Forest with trees reaching heights of 25–35 m. 2.3. Jaguar densities Elevations range between 100 and 420 m, average temperatures be- tween 20°-30° and annual precipitation between 1324 and 1350 mm Individual jaguars were identified based on their coat pattern and with a dry season from December to April. Forest fires are a major the sex of each individual was determined whenever possible. If a photo threat to the MBR, registering an average of 2010 MODIS hot points could not be clearly assigned to an individual, it was excluded from the (i.e. fire detected by NASA satellites across the entire MBR) from 2010 density analysis. to 2015; however only five of those occurred inside the sampling area We estimated jaguar densities using a spatial capture-recapture (CONAP-WCS, 2015). (SCR) model (Efford et al., 2009; Royle et al., 2013b; Royle and Young, 2008). SCR models use the spatial information of jaguar detections to estimate the parameters of the half-normal detection function (detec- 2.2. Camera trap surveys tion probability at the center of the home range g0 and movement parameter σ) in order to estimate the density of jaguars. We modeled Our camera trap survey design followed recommendations by both detection probability and home-range size independently for Tobler and Powell (2013) for jaguar surveys. In both study areas we males and females. We also included a detection covariate to account used a paired camera trap setup in a regular grid with cameras spaced for the difference in detection probability for cameras that were placed 2–3 km apart. In Guatemala we had 50 camera stations with Reconyx on active logging roads, on old roads and off roads and modeled an Hyperfire RM 45 and HC 500 camera traps whilst in Peru we had 89 interaction with sex in order to evaluate whether road preference dif- camera stations using Bushnell TrophyCam HD (2012 and 2013 fered for males and females. Old roads are roads that are not used models) cameras (Table 1). Camera trap polygons covered an area of anymore and have vegetation growing back but are still passable (we 520 km2 in Guatemala and 645 km2 in Peru. Jaguars normally have also included trails within this category). Given that jaguars extensively higher detection probabilities on roads and trails (Sollmann et al., 2011; use the road network for travel we hypothesized that a non-Euclidian Tobler et al., 2013) so we set up cameras along logging roads whenever distance model would better be able to explain the movement of jaguars possible. Camera traps were in the field for between 90 and 120 days in the landscape (Royle et al., 2013a; Sutherland et al., 2015). This during the dry season and were active for 24 h a day. All photos were model can account for jaguars traveling longer distances along roads entered into Camera Base 1.7, a database specifically developed for than off road. We created a binary cost surface where roads had a value managing and analyzing camera trap data (Tobler, 2015). 247 M.W. Tobler et al. Biological Conservation 220 (2018) 245–253 70°W 69°45’W 69°30’W 11°15’S 11°15’S 11°30’S 11°30’S Camera Traps Roads Year Logged Unlogged 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 - 2014 11°45’S 11°45’S 0 5 10 km 70°W 69°45’W 69°30’W Fig. 2. Peru camera trap locations and logging concessions. Colors of logging blocks indicate what year trees were harvested. Unlogged areas are as of 2014. Table 1 been logged) and distance from active logging road. We also evaluated Data for the camera trap surveys carried out in Guatemala and Peru. distance to the nearest river and the normalized difference vegetation index (NDVI) based on Landsat images but neither was significant for Survey Start End date Stations Camera days Camera date polygona (km2) any of the species studied so we dropped them from the model. We included a detection covariate to account for the difference in detection Guatemala 18 April 16 July 50 4406 520 probability for cameras placed on active roads, old roads and off-road. 2013 2013 Continuous covariates were standardized to a range between 0 and 1. Peru 25 June 23 October 89 8688 645 2014 2014 We calculated the percentage of area occupied (PAO) across the full extent of the concessions by creating raster grids with a 100 × 100 m a Minimum convex polygon without buffer. resolution for all the covariates, calculating the occupancy probability for each species for each grid cell and then averaging across all cells. We of zero and forested areas a value of one. Models were run in a max- did this for each sample of our MCMC results to estimate credible in- imum-likelihood framework using the package secr (Efford, 2016) in R tervals for the resulting occupancy values. For wide-ranging species (R Development Core Team, 2015) and we subsequently compared such as large carnivores this value is interpreted as the percentage of models using the Akaike Information Criterion (AIC). the area used by the species. We ran the model in JAGS (Plummer, 2003) through R (R 2.4. Species richness and occupancy Development Core Team, 2015). We ran three chains with 150,000 iterations, a burn-in of 50,000 and a thinning rate of 100. We visually To evaluate the effect of logging on the whole large and medium- inspected the chains for convergence. Covariates were considered sig- sized mammal community we took advantage of the fact that during nificant when the 95% Bayesian credible interval did not include zero. each year one or two blocks in each study area were being harvested. This creates a mosaic of unlogged blocks and blocks harvested at dif- ferent points in time. We used this spatial replication to evaluate the 3. Results effect of logging over time. While temporal replication would be pre- ferable (evaluating blocks pre- and post-logging), such a design would 3.1. Jaguar densities take 10 to 15 years to carry out. We believe that our study areas are homogenous enough and the logged blocks are distributed enough to We obtained 203 records of 23 jaguars (14 males and 9 females) not affect our inferences. from Guatemala and 215 records of 43 jaguars (19 males, 22 females For the community analysis we used a Bayesian multi-species oc- and 2 of unknown sex) from Peru. The estimated density in Guatemala cupancy model to estimate community structure and occupancy for all was 1.52 ± 0.34 ind. 100 km−2 and in Peru 4.54 ± 0.83 ind. species (Dorazio and Royle, 2005; Dorazio et al., 2006). These models 100 km−2 with the highest ranking non-Euclidian distance model and account for imperfect detection and, by combining data from all species 3.00 ± 0.54 with the Euclidian distance model. For both surveys, de- of interest, can provide improved parameter estimates for rarer species tection probability g0 was higher on active roads than on old roads, (Zipkin et al., 2009). To cope with the large level of heterogeneity much lower off road, and higher for male than for female jaguars generally present in camera trap data due to differences in the local (Table 2). The movement parameter σ and therefore the home range abundance or non-random movement of animals we used the Royle- size was larger for males than for females, and the respective estimates Nichols version of the multi-species occupancy model (Tobler et al., of σ were similar for the two surveys (Table 2). For the Peru data, the 2015; Yamaura et al., 2011). We used the following occupancy cov- model with the lowest AICc value also included both the non-Euclidian ariates: logged (yes/no), years since logged (zero if the site had not distance as well as an interaction term for road type and sex, indicating 248 M.W. Tobler et al. Biological Conservation 220 (2018) 245–253 Table 2 Results from spatial capture-recapture models for jaguar surveys in Guatemala and Peru. Values show as mean, standard error and 95% confidence intervals. Parameter Guatemala Peru no-Euclidian Peru Euclidian D 1.52 ± 0.34 (0.96–2.25) 4.54 ± 0.83 (2.78–7.44) 3.00 ± 0.54 (2.11–4.26) σM 5306 ± 483 (4465–6334) 6922 ± 467 (6066–7899) 5605 ± 382 (4906–6404) σF 4073 ± 667 (3043–5673) 4491 ± 574 (3500–5763) 3276 ± 349 (2659–4034) g0 M road 0.0421 ± 0.0058 (0.0318–0.0540) 0.0522 ± 0.0083 (0.0380–0.0712) 0.0490 ± 0.0072 (0.0366–0.0653) g0 M old road 0.0230 ± 0.0065 (0.0132–0.0400) 0.0161 ± 0.0040 (0.0099–0.0262) 0.0070 ± 0.0015 (0.0046–0.0107) g0 M off road 0.0080 ± 0.0042 (0.0029–0.0192) 0.0054 ± 0.0031 (0.0018–0.0164) 0.0010 ± 0.0005 (0.0004–0.0027) g0 F road 0.0092 ± 0.0029 (0.0047–0.0163) 0.0087 ± 0.0023 (0.0051–0.0147) 0.0091 ± 0.0023 (0.0055–0.0151) g0 F old road 0.0049 ± 0.0019 (0.0021–0.0098) 0.0037 ± 0.0012 (0.0019–0.007) 0.0029 ± 0.0009 (0.0016–0.0054) g0 F off road 0.0017 ± 0.0011 (0.0005–0.0045) 0.0019 ± 0.0011 (0.0006–0.0058) 0.0012 ± 0.0006 (0.0005–0.003) Cost off road M 8.72 ± 1.86 (5.77–13.2) Cost off road F 2.49 ± 0.91 (1.25–4.98) σ: movement parameter in meters, g0: detection probability, D: density in ind. 100 km−2, Cost: cost-estimate for off-road travel by the non-Euclidian distance model, M: male, F: female. 1.0 that road use is different for male and female jaguars. Neither of these were included in the highest-ranking model for Guatemala but this 0.8 could be because of the lower number of cameras and because no fe- Occupancy male jaguars were detected off-road, making parameter estimates un- 0.6 reliable for the interaction model. 0.4 3.2. Species richness and occupancy 0.2 In Guatemala we recorded 24 species including 22 species of large and medium-sized terrestrial mammals and two species of terrestrial 0.0 birds. In Peru, we recorded 27 species comprising of 25 species of terrestrial large and medium-sized mammals and two species of ter- 0 1 2 3 4 5 6 7 8 9 10 11 12 restrial birds. In Guatemala, three species showed a significant increase Years since logged in occupancy in logged areas compared to unlogged areas: red brocket deer (Mazama temama), white-tailed deer (Odocoileus virginianus) and Fig. 4. Marginal occupancy probabilities of 27 species in relation to time after logging in the tapir (Tapirus bairdii). Three species showed an initial increase and certified logging concessions in the Peruvian Amazon. Zero indicates unlogged blocks. then a slow decline over time: paca (Cuniculus paca), brown agouti Think black lines indicate species that have a significant response (95% credible intervals do not include zeroes) and the thick black lines shows the mean response across all (Dasyprocta punctata) and common opossum (Didelphis marsupialis), species. Fig. 3). In Peru, there were seven species that showed a significant increase: paca, brown agouti, ocelot (Leopardus pardalis), razor-billed curassow (Mitu tuberosum), puma (Puma concolor), Brazilian rabbit cinereoargenteus) (Table S6). In Peru this was the case for six species: (Sylvilagus brasiliensis) and the lowland tapir (Tapirus terrestris); Fig. 4. ocelot, razor billed curassow, jaguar, puma, Brazilian rabbit and the The initial increase was generally followed by a decrease in occupancy lowland tapir (Table S4). Of these, three also showed an increased over time after logging. No species showed a negative initial response to detection probability on old roads and trails. In addition, three species logging. We found no species avoiding areas close to logging roads and had a lower detection probability on roads in Peru (paca, nine-banded the jaguar was found more often closer to roads. In Guatemala, nine armadillo (Dasypus spp.), giant anteater (Myrmecophaga tridactyla), species had a higher detection probability on active logging roads than pale-winged trumpeter (Psophia leucoptera) but none in Guatemala. on old roads and away from roads: great curassow (Crax rubra), brown In Peru, the total percentage of area occupied across the extent of agouti, common opossum, ocelot, red brocket deer, ocellated turkey the concessions ranged from 15% for the pacarana to 95% for the (Meleagris ocellata), jaguar, puma and the grey fox (Urocyon lowland tapir and the red brocket deer (Fig. 5 and Table S1). Jaguars and pumas both used about 75% of the area. Results were similar in Guatemala with values ranging from 19% for the striped hog-nosed 1.0 skunk to 96% for the ocelot (Fig. 6 and Table S2). Jaguars and pumas used between 70% and 75% of the area. 0.8 Occupancy 0.6 4. Discussion 0.4 4.1. Jaguar densities 0.2 Tobler and Powell (2013) found that many jaguar camera trap surveys covered too small an area to collect reliable data on jaguar 0.0 densities and made a number of design recommendations. We subse- 0 1 2 3 4 5 6 7 8 9 10 12 quently implemented these recommendations resulting in two of the largest camera trap survey areas for jaguars with some of the highest Years since logged numbers of individuals recorded to date (review of previous studies in Fig. 3. Marginal occupancy probabilities of 24 species in relation to time after logging in Tobler and Powell, 2013). These robust datasets lead to improved certified logging concessions in the Maya Biosphere Reserve in Guatemala. Zero indicates density estimates with smaller confidence intervals for all parameters. unlogged blocks. Thin black lines indicate species that have a significant response (95% While densities were lower in Guatemala (1.52 ± 0.34 ind. 100 km−2) credible intervals do not include zeroes) and the thick black lines shows the mean re- than in Peru (4.56 ± 0.83 ind. 100 km−2), this is likely due to habitat sponse across all species. conditions. The Guatemala study site receives less precipitation than 249 M.W. Tobler et al. Biological Conservation 220 (2018) 245–253 Tapirus terrestris home ranges and higher detection probabilities than females, indicating Mazama americana that they travel further on a daily basis. This is consistent with findings Dasyprocta punctata Leopardus pardalis from jaguars tracked with GPS collars (Morato et al., 2016). We also Mitu tuberosum Cuniculus paca found that detection probabilities on active, open logging roads were Didelphis marsupialis Psophia leucoptera higher than on old roads and trails, and much higher than off road. A Puma concolor Pecari tajacu preference of roads can be seen for both sexes, although it is stronger Panthera onca Priodontes maximus for males than for females. In Peru, male jaguars had a 10-fold higher Dasypus spp. detection probability on roads compared to off roads and females had a Sylvilagus brasiliensis Myrmecophaga tridactyla two-fold higher detection probability. In Guatemala, the detection Eira barbara Leopardus wiedii probability on roads was around five times higher for both sexes but Myoprocta pratti Puma yagouaroundi interestingly no females were detected off road. This shows the im- Nasua nasua Procyon cancrivorus portance of placing cameras on trails or roads to maximize detection Atelocynus microtis Mazama nemorivaga probabilities. At least in the dense forests of the Amazon, roads seem to Tayassu pecari be important movement corridors for jaguars as reflected by the better Dinomys branickii Speothos venaticus Galicitis vittata fit of the model using non-Euclidian distance and again male jaguars showed a stronger preference for traveling along roads than females. 0.0 0.2 0.4 0.6 0.8 1.0 Ignoring these effects of the landscape structure on movement can lead Percentage of Area Occupied (PAO) to an underestimation of densities (Royle et al., 2013a; Sutherland et al., 2015). Fig. 5. Percentage of area occupied (PAO) for 27 species in reduced-impact logging concessions in Peru (mean and the 95% credible intervals estimated with a multi-species 4.2. Species richness and occupancy occupancy model). We confirmed previous findings that for large and medium-sized Leopardus pardalis terrestrial mammals the impact of reduced-impact logging is generally Crax rubra small and, in some cases, logging can have positive effects on their Agouti paca Urocyon cineroargenteus diversity and abundance (Azevedo-Ramos et al., 2006; Burivalova et al., Mazama temama Tapirus bairdii 2014; Clark et al., 2009; Gibson et al., 2011; Meijaard and Sheil, 2007; Panthera onca Roopsind et al., 2017). At both study sites we detected all the species Meleagris ocellata Puma concolor known to occur in the region, and in Peru we detected an additional Dasyprocta punctata Nasua narica species (Pacarana (Dinomys branickii)) that was missed by previous Didelphis surveys (Tobler et al., 2015). Occupancy rates across the two study sites Leopardus wiedii Odocoileus virginianus were very similar with logging resulting in an increase in herbivores Pecari tajacu including tapir, deer, rabbit, paca and agouti. Logging had no negative Mazama pandora Yaguarundi effect on any of the other species. The pattern showed an initial increase Dasypus novemcinctus Eira barbara in occupancy after logging and a slow decrease over the following years Tayassu pecari back to pre-logging levels. Logging opens up the forest canopy in areas Procyon lotor Canis latrans where timber is harvested and along logging roads, which in turn leads Conepatus semistriatus to an increase in understory vegetation that is consumed by herbivores. Philander opossum Over time the canopy closes again and understory vegetation di- 0.0 0.2 0.4 0.6 0.8 1.0 minishes. Similar results have been found in studies in Africa and Percentage of Area Occupied (PAO) Southeast Asia where generalist herbivores benefited from low-impact logging but some smaller-bodied frugivores were negatively affected Fig. 6. Percentage of area occupied (PAO) for 24 species in reduced-impact logging (Clark et al., 2009; Davies et al., 2001). concessions in Guatemala (mean and the 95% credible intervals estimated with a multi- species occupancy). The cause for the increase in the two large rodent species is less clear since they are mainly frugivores (Dubost et al., 2006). However, an increase of rodents in logged forests can be explained by an increase the Peru site (1324-1350 mm compared to 2500–3500 mm) and due to in microhabitat diversity, increased cover and an increase in resource its porous karstic substrate, that simply filters rain down into aquifers, abundance in the form of fruit and insects (Fredericksen and possesses a scarcity of surface water when compared to the Peru site. Fredericksen, 2002; Lambert et al., 2006). Many of the herbivores and However, density estimates from Guatemala were on the high end of large rodents are important prey species for jaguars and pumas estimates from even drier sites such as the Bolivian Chaco (SCR esiti- (Emmons, 1987; Foster et al., 2010; Weckel et al., 2006) which is re- mate range: 0.46–1.46 ind. 100 km−2, Noss et al., 2012) and higher flected by the high density and high occupancy of these large cats in the than an estimate from the Emas National Park in the Cerrado of Brazil studied logging concessions. An increase in rodent abundance could (SCR estimate: 0.29 ± 0.10 ind 100 km−2, Sollmann et al., 2011). also explain the increase in occupancy of ocelots in logged areas. Density estimates for Peru were similar to those from the llanos of Most carnivore species showed an increased detection probability Venezuela (4.44 ± 1.16 ind. 100 km−2, Jędrzejewski et al., 2017) but on logging roads at both study sites confirming previous results that were higher than jaguar densities along the coast of French Guiana they actively use trails and roads for travel (Di Bitetti et al., 2014; (3.22 ± 0.30 ind 100 km−2, Petit et al., 2017) and significantly higher Harmsen et al., 2010; Sollmann et al., 2012; Tobler et al., 2015). At than in a logging concession in Guyana (1.72 ind 100 km−2; Roopsind both sites, the curassow had a higher detection probability on roads et al., 2017). Higher jaguar densities were only found in the Pantanal of than away from them as did the red brocket deer in Guatemala and the Brazil (6.6 ind. 100 km−2, non SCR estimate, Soisalo and Cavalcanti, tapir in Peru. These findings indicate that even active logging roads 2006). However, estimates were lower than previous estimates from the with trucks passing several times a day have a relatively small impact same area in Peru (4.9 ± 1.0 ind. 100 km−2; Tobler et al., 2013) likely upon our study species and do not pose much of a disturbance or barrier due to the increase in the extent of the camera trap polygon. for movement. On the contrary, they might actually increase long-dis- In agreement with previous camera trap studies (Sollmann et al., tance movement and possibly facilitate dispersal of animals into logged 2011; Tobler et al., 2013), male jaguars were found to have much larger areas. In Peru, we observed jaguars, tapirs and many other species 250 M.W. Tobler et al. Biological Conservation 220 (2018) 245–253 starting to use a new road only days after it was created by bulldozers. activities while still protecting intact ecosystems. The key to success is While we observed a wide variation in PAO values across species, strict control and enforcement of management practices by govern- several patterns can be seen at both sites. Large ungulates such as deer ments along with third-party organizations such as the FSC or the and tapirs are common and widespread with the exception of the white- Program for the Endorsement of Forest Certification. Programs such as lipped peccary (Tayassu pecari) in Peru and Guatemala and the brown the REDD+ (Reducing Emissions from Deforestation and Forest De- brocket deer in Peru (Mazama nemorivaga). This results in a healthy gradation), that pay concession owners for maintaining high carbon prey base for jaguars and puma as indicated by our density estimates. stocks by reducing logging impact, can complement forest certification Among the medium-sized carnivores we can see that the ocelot is the to ensure the economic viability of low impact forest management and most common species followed by the margay (Leopardus wiedii) and help maintain high biodiversity forests in logging concessions. In con- the yaguarundi (Puma yagouaroundi). Occupancy values in our studies clusion, increasing the area of tropical forests in Latin America that were generally higher than in a comparable study from logging con- strictly adhere to certification standards has the potential for large cessions in Guyana (Roopsind et al., 2017) which could be due to the conservation benefits. complete absence of hunting in our study areas. In Peru occupancy values for many species, especially large mammals, were comparable to values from other sites, including protected areas (Tobler et al., 2015), Acknowledgements but comparison is limited by differences in forest composition that can affect species distribution and occupancy independent of logging. For the work in Peru we would like to thank the Disney Worldwide Conservation Fund and WWF Switzerland for generous funding. We are 4.3. The importance of sustainably managed logging concessions grateful to Victor Espinoza, Elisabeth Espinoza and Victor Leoni Espinoza from Aserradero Espinoza and Vittorio De Dea Peña and The impact of logging on species diversity and abundance has been Nelson Melendez Ascaño from Maderera Bozovich for allowing us to the subject of much debate (Burivalova et al., 2014; Chaudhary et al., work in their concessions and for helping us with logistics. The fol- 2016; Edwards et al., 2014; Gibson et al., 2011). However, the value of lowing people helped with the camera trap work in Peru: Ruben well-managed logged forest and logging concessions for biodiversity Aviana, Marlon Guerrra, Javier Huinga, Juan, Racua Remigio Yumbato, conservation as a complement to protected areas is becoming more Denisse Mateo, Diego Acosta, Victoria Palomino, and Angela Arapa. We recognized (Dent and Wright, 2009; Dickinson et al., 1996; Edwards also thank Kate Lambert, Kathleen Esra, Carina Graham, and Gloria et al., 2014; Putz et al., 2012; Struebig et al., 2013; Wilson et al., 2010). Marselas for their hard work processing all the camera trap images. We For larger and more mobile species, existing networks of protected would like to thank the DGSPFFS-SERFOR (formerly DGFFS) for areas alone may not be not enough to ensure long-term conservation. granting us permission through the R.D. No 0206-2014-MINAGRI- Well-managed forests can provide both important habitat and con- DGFFS/DGEFF to carry out this research in Peru. We greatly appreciate nectivity among protected areas (Chazdon et al., 2009; Clark et al., the logistic and administrative support provided by WWF-Peru and 2009; Edwards et al., 2014; Wilson et al., 2010). SDZG-Peru. Many thanks also to George Powell who initiated our long- An exceptionally important aspect for effective wildlife conserva- term jaguar research program in Peru. tion in logging concessions is a strict regulation of hunting inside the For the work in Guatemala we would like to thank the Liz Claiborne concessions and no access to the logging roads for outsiders (Bennett, and Art Ortenberg Fundation, United States Agency for International 2004; Davies et al., 2001; Polisar et al., 2017). Several studies have Development (USAID), Rainforest Alliance and Programa Clima, shown that the largest impact of logging is not the change in forest Naturaleza y Comunidades en Guatemala (CNCG). We are grateful to structure but the increase in hunting due to the easy access by people CONAP (Consejo Nacional de Áreas Protegidas de Guatemala), along logging roads (Meijaard et al., 2005; Poulsen et al., 2011; Sociedad Civil Árbol Verde, Sociedad Civil Custodios de la Selva, Robinson et al., 1999; Roopsind et al., 2017). In the concessions studied Sociedad Civil Laborantes del Bosque, Sociedad Civil El Esfuerzo and in Peru and Guatemala all the main access roads to the concessions are Sociedad Civil Impulsores Suchitecos for allowing us to work in the gated with a 24-h watchman on guard. This results in virtually no Maya Biosphere Reserve and their community concessions (respec- hunting inside the concessions which is better than in most protected tively). We also thank Christian Rossell from Pecorino Ristorante in areas. Guatemala City, Fabio Diaz from WCS Nicaragua, and Dr. Francisco We found a complete terrestrial mammal community and good ja- Estrada-Belli from Proyecto Arqueológico Holmul for their kindly sup- guar populations in all the logging concessions sampled and could not port in the field and Guatemala City. The following people helped with detect negative impacts of logging on any of the study species. the camera trap work in Guatemala: Marcial Córdova, Juventino López, However, we would like to make it clear that our results cannot be Fabio Diaz, Efraín Alberto Soza, José Rafael Garrido and Luis Miguel generalized to all logging concessions across tropical forests. The vo- Recinos. lume of wood harvested in the concessions studied (1.2–3 m3/ha) is We would like to thank Andy Royle and Chris Sutherland for input much lower than in many other logging operations around the world that led us to try out the non-Euclidian distance models and Matt where volumes may be as high as 150 m3/ha, and well below the Anderson, Francis Putz and two anonymous reviewers for valuable threshold of 10 m3/ha identified by Burivalova et al. (2014) where comments on the manuscript. negative effects on mammals start to be noticeable. Furthermore, the FSC certification ensures that the logging cycles of 20 years are re- spected and strict reduced impact logging practices are employed. At Appendix A. Supplementary data the time of this study only half of the blocks had been harvested al- lowing for fast recolonization of logged areas by animals from sur- Additional detailed results from the multi-species occupancy models rounding blocks. (Appendix S1), and model selection results for the spatial capture-re- We believe that well-managed, certified logging concessions have capture models (Appendix S2) are available online. The authors are far less negative effects on forest ecosystems than alternative land uses solely responsible for the content and functionality of these materials. such as cattle ranching, palm plantations and mono-cultures that result Queries (other than absence of the material) should be directed to the in deforestation and drastic reductions in biological diversity (Gaveau corresponding author. Supplementary data associated with this article et al., 2013; Oliveira et al., 2007; Polisar et al., 2017; Radachowsky can be found in the online version, at https://doi.org/10.1016/j. et al., 2012). Logging concessions would therefore be ideal in buffer biocon.2018.02.015. zones and multiple use zones of protected areas, allowing for economic 251 M.W. Tobler et al. Biological Conservation 220 (2018) 245–253 References are irreplaceable for sustaining tropical biodiversity. Nature 478, 378–381. Gullison, R.E., 2003. Does forest certification conserve biodiversity? Oryx 37, 153–165. Hansen, M.C., Potapov, P.V., Moore, R., Hancher, M., Turubanova, S.A., Tyukavina, A., Achard, F., Beuchle, R., Mayaux, P., Stibig, H.-J., Bodart, C., Brink, A., Carboni, S., Thau, D., Stehman, S.V., Goetz, S.J., Loveland, T.R., Kommareddy, A., Egorov, A., Desclée, B., Donnay, F., Eva, H.D., Lupi, A., Raši, R., Seliger, R., Simonetti, D., 2014. Chini, L., Justice, C.O., Townshend, J.R.G., 2013. High-resolution global maps of Determination of tropical deforestation rates and related carbon losses from 1990 to 21st-century forest cover change. Science 342, 850–853. 2010. Glob. Chang. Biol. 20, 2540–2554. Harmsen, B.J., Foster, R.J., Silver, S., Ostro, L., Doncaster, C.P., 2010. Differential use of Ahumada, J.A., Hurtado, J., Lizcano, D., 2013. Monitoring the status and trends of tro- trails by forest mammals and the implications for camera-trap studies: a case study pical forest terrestrial vertebrate communities from camera trap data: a tool for from Belize. Biotropica 42, 126–133. conservation. PLoS One 8, e73707. Hodgdon, B., Hughell, D., Ramos, V., McNab, R., 2015. Deforestation Trends in the Maya Asner, G.P., Rudel, T.K., Aide, T.M., Defries, R., Emerson, R., 2009. A contemporary as- Biosphere Reserve, Guatemala. Rainforest Alliance, pp. 14. sessment of change in humid tropical forests. Una Evaluación Contemporánea del Jędrzejewski, W., Puerto, M.F., Goldberg, J.F., Hebblewhite, M., Abarca, M., Gamarra, G., Cambio en Bosques Tropicales Húmedos. Conserv. Biol. 23, 1386–1395. Calderón, L.E., Romero, J.F., Viloria, Á.L., Carreño, R., Robinson, H.S., Lampo, M., Azevedo-Ramos, C., de Carvalho Jr, O., do Amaral, B.D., 2006. Short-term effects of re- Boede, E.O., Biganzoli, A., Stachowicz, I., Velásquez, G., Schmidt, K., 2017. Density duced-impact logging on eastern Amazon fauna. For. Ecol. Manag. 232, 26–35. and population structure of the jaguar (Panthera onca) in a protected area of Los Bennett, E., 2004. Seeing the Wildlife and the Trees: Improving Timber Certification to Llanos, Venezuela, from 1 year of camera trap monitoring. Mamm. Res. 62, 9–19. Conserve Tropical Forest Wildlife. Wildlife Conservation Society and the World Bank, Lambert, T.D., Malcolm, J.R., Zimmerman, B.L., 2006. Amazonian small mammal New York, New York, USA. abundances in relation to habitat structure and resource abundance. J. Mammal. 87, Bicknell, J.E., Struebig, M.J., Edwards, D.P., Davies, Z.G., 2014. Improved timber harvest 766–776. techniques maintain biodiversity in tropical forests. Curr. Biol. 24, R1119–R1120. Laurance, W.F., Sayer, J., Cassman, K.G., 2014. Agricultural expansion and its impacts on Blaser, J., Sarre, A., Poore, D., Johnson, S., 2011. Status of tropical forest management tropical nature. Trends Ecol. Evol. 29, 107–116. 2011. In: ITTO Technical Series 38. pp. 418. Meijaard, E., Sheil, D., 2007. The persistence and conservation of Borneo's mammals in Borchers, D.L., Efford, M.G., 2008. Spatially explicit maximum likelihood methods for lowland rain forests managed for timber: observations, overviews and opportunities. capture-recapture studies. Biometrics 64, 377–385. Ecol. Res. 23, 21–34. Burivalova, Z., Şekercioğlu, Çağan H., Koh, Lian P., 2014. Thresholds of logging intensity Meijaard, E., Scheil, D., Nasi, R., Augeri, D., Rosenbaum, B., Iskandar, D., Setyawati, T., to maintain tropical forest biodiversity. Curr. Biol. 24, 1893–1898. Lammertink, M., Rachmatika, I., Wong, A., Soehartono, T., Stanley, S., O'Brien, T., Carrera, F., Stoian, D., Campos, J.J., Morales, J., Pinelo, G., 2006. Forest certification in 2005. Life After Logging: Reconciling Wildlife Conservation and Production Forestry Guatemala. In: Confronting Sustainability: Forest Certification in Developing and in Indonesian Borneo. Cifor. Transitioning Countries, pp. 363–406. Morato, R.G., Stabach, J.A., Fleming, C.H., Calabrese, J.M., De Paula, R.C., Ferraz, Chaudhary, A., Burivalova, Z., Koh, L.P., Hellweg, S., 2016. Impact of forest management K.M.P.M., Kantek, D.L.Z., Miyazaki, S.S., Pereira, T.D.C., Araujo, G.R., Paviolo, A., De on species richness: global meta-analysis and economic trade-offs. Sci. Rep. 6, 23954. Angelo, C., Di Bitetti, M.S., Cruz, P., Lima, F., Cullen, L., Sana, D.A., Ramalho, E.E., Chazdon, R.L., Harvey, C.A., Komar, O., Griffith, D.M., Ferguson, B.G., Martínez-Ramos, Carvalho, M.M., Soares, F.H.S., Zimbres, B., Silva, M.X., Moraes, M.D.F., Vogliotti, A., M., Morales, H., Nigh, R., Soto-Pinto, L., Van Breugel, M., Philpott, S.M., 2009. May Jr., J.A., Haberfeld, M., Rampim, L., Sartorello, L., Ribeiro, M.C., Leimgruber, P., Beyond reserves: a research agenda for conserving biodiversity in human-modified 2016. Space use and movement of a neotropical top predator: the endangered jaguar. tropical landscapes. Biotropica 41, 142–153. PLoS One 11, e0168176. Clark, C.J., Poulsen, J.R., Malonga, R., Elkan, J.P.W., 2009. Logging concessions can Nepstad, D.C., Verssimo, A., Alencar, A., Nobre, C., Lima, E., Lefebvre, P., Schlesinger, P., extend the conservation estate for Central African tropical forests. Conserv. Biol. 23, Potter, C., Moutinho, P., Mendoza, E., Cochrane, M., Brooks, V., 1999. Large-scale 1281–1293. impoverishment of Amazonian forests by logging and fire. Nature 398, 505–508. CONAP-WCS, 2015. Monitoreo de incendios en la Reserva de Biosfera Maya para la Noss, A.J., Gardner, B., Maffei, L., Cuéllar, E., Montaño, R., Romero-Muñoz, A., Sollman, temporada 2015. In: Programa Clima, Naturaleza y Comunidades en Guatemala. R., O'Connell, A.F., Altwegg, R., 2012. Comparison of density estimation methods for Guatemala, Wildlife Conservation Society, pp. 12. mammal populations with camera traps in the Kaa-Iya del Gran Chaco landscape. Davies, G., Heydon, M., Leader-Williams, N., MacKinnon, J., Newing, H., 2001. The ef- Anim. Conserv. 15, 527–535. fects of logging on tropical forest ungulates. In: Fimbel, R.A., Grajal, A., Robinson, Oliveira, P.J.C., Asner, G.P., Knapp, D.E., Almeyda, A., Galvan-Gildemeister, R., Keene, S., J.G. (Eds.), The Cutting Edge: Conserving Wildlife in Logged Tropical Forest. Raybin, R.F., Smith, R.C., 2007. Land-use allocation protects the Peruvian Amazon. Columbia University Press, New York, USA, pp. 93–124. Science 317, 1233–1236. Dent, D.H., Wright, J.S., 2009. The future of tropical species in secondary forests: a PEFC, 2017. https://www.pefc.org/. quantitative review. Biol. Conserv. 142, 2833–2843. Petit, M., Denis, T., Rux, O., Richard-Hansen, C., Berzins, R., 2017. Estimating jaguar Development Core Team, R., 2015. R: A Language and Environment for Statistical (Panthera onca) density in a preserved coastal area of French Guiana. In: In Computing. R Foundation for Statistical Computing, Vienna, Austria. Mammalia. Di Bitetti, M.S., Paviolo, A., De Angelo, C., 2014. Camera trap photographic rates on roads Plummer, M., 2003. JAGS: a program for analysis of Bayesian graphical models using vs. off roads: location does matter. Mastozool. Neotropical 21, 37–46. Gibbs sampling. In: Proceedings of the 3rd International Workshop on Distributed Dickinson, M.B., Dickinson, J.C., Putz, F.E., 1996. Natural forest management as a con- Statistical Computing. Vienna, Austria. servation tool in the tropics: divergent views on possibilities and alternatives. Polisar, J., de Thoisy, B., Rumiz, D.I., Santos, F.D., McNab, R.B., Garcia-Anleu, R., Ponce- Commonw. For. Rev. 75, 309–315. Santizo, G., Arispe, R., Venegas, C., 2017. Using certified timber extraction to benefit Dirección General Forestal y de Fauna Silvestre del Ministerio de Agricultura y Riego, jaguar and ecosystem conservation. Ambio 46, 588–603. 2013. Operaciones forestales con certificación forestal voluntaria en el Perú. Poulsen, J.R., Clark, C.J., Bolker, B.M., 2011. Decoupling the effects of logging and Dorazio, R.M., Royle, J.A., 2005. Estimating size and composition of biological commu- hunting on an Afrotropical animal community. Ecol. Appl. 21, 1819–1836. nities by modeling the occurrence of species. J. Am. Stat. Assoc. 100, 389–398. Putz, F.E., Zuidema, P.A., Synnott, T., Peña-Claros, M., Pinard, M.A., Sheil, D., Vanclay, Dorazio, R.M., Royle, J.A., Soderstrom, B., Glimskar, A., 2006. Estimating species richness J.K., Sist, P., Gourlet-Fleury, S., Griscom, B., Palmer, J., Zagt, R., 2012. Sustaining and accumulation by modeling species occurrence and detectability. Ecology 87, conservation values in selectively logged tropical forests: the attained and the at- 842–854. tainable. Conserv. Lett. 5, 296–303. Dubost, G., xe, rard, Henry, O., 2006. Comparison of diets of the acouchy, agouti and Radachowsky, J., Ramos, V.H., McNab, R., Baur, E.H., Kazakov, N., 2012. Forest con- paca, the three largest terrestrial rodents of French Guianan forests. J. Trop. Ecol. 22, cessions in the Maya Biosphere Reserve, Guatemala: a decade later. For. Ecol. Manag. 641–651. 268, 18–28. Edwards, D.P., Tobias, J.A., Sheil, D., Meijaard, E., Laurance, W.F., 2014. Maintaining Rayan, D.M., Mohamad, S.W., 2009. The importance of selectively logged forests for tiger ecosystem function and services in logged tropical forests. Trends Ecol. Evol. 29, Panthera tigris conservation: a population density estimate in peninsular Malaysia. 511–520. Oryx 43, 48–51. Efford, M.G., 2016. secr: spatially explicit capture-recapture models. In: R Package Robinson, J.G., Redford, K.H., Bennett, E.L., 1999. Conservation - wildlife harvest in Version 2.10.4. logged tropical forests. Science 284, 595–596. Efford, M.G., Dawson, D.K., Borchers, D.L., 2009. Population density estimated from lo- Roopsind, A., Caughlin, T.T., Sambhu, H., Fragoso, J.M.V., Putz, F.E., 2017. Logging and cations of individuals on a passive detector array. Ecology 90, 2676–2682. indigenous hunting impacts on persistence of large Neotropical animals. Biotropica Emmons, L.H., 1987. Comparative feeding ecology of felids in a neotropical rain-forest. 49, 565–575. Behav. Ecol. Sociobiol. 20, 271–283. Royle, J.A., Young, K.V., 2008. A hierarchical model for spatial capture-recapture data. FAO, 2015. Global Forest Resources Assessment 2015. FAO, Rome, Italy. Ecology 89, 2281–2289. Foster, R.J., Harmsen, B.J., Valdes, B., Pomilla, C., Doncaster, C.P., 2010. Food habits of Royle, J.A., Chandler, R.B., Gazenski, K.D., Graves, T.A., 2013a. Spatial capture–- sympatric jaguars and pumas across a gradient of human disturbance. J. Zool. 280, recapture models for jointly estimating population density and landscape con- 309–318. nectivity. Ecology 94, 287–294. Fredericksen, N.J., Fredericksen, T.S., 2002. Terrestrial wildlife responses to logging and Royle, J.A., Chandler, R.B., Sollmann, R., Gardner, B., 2013b. Spatial capture-recapture. fire in a Bolivian tropical humid forest. Biodivers. Conserv. 11, 27–38. Academic Press, Waltham, MA, USA. FSC, 2016. https://us.fsc.org/. Secaira, E., Ariano, D., Barrientos, T., Radachowsky, J., 2015. Plan Maestro Reserva de la Gaveau, D.L.A., Kshatriya, M., Sheil, D., Sloan, S., Molidena, E., Wijaya, A., Wich, S., Biósfera Maya. Segunda Actualización, p. 370. Guatemala: Wildlife Conservation Ancrenaz, M., Hansen, M., Broich, M., Guariguata, M.R., Pacheco, P., Potapov, P., Society (WCS) and Programa para el Desarrollo de Petén y la Conservación de la Turubanova, S., Meijaard, E., 2013. Reconciling forest conservation and logging in Reserva de la Biósfera (PDP-CRBM). CONAP-DGPCN. Indonesian Borneo. PLoS One 8, e69887. Soisalo, M.K., Cavalcanti, S.M.C., 2006. Estimating the density of a jaguar population in Gibson, L., Lee, T.M., Koh, L.P., Brook, B.W., Gardner, T.A., Barlow, J., Peres, C.A., the Brazilian Pantanal using camera-traps and capture-recapture sampling in com- Bradshaw, C.J.A., Laurance, W.F., Lovejoy, T.E., Sodhi, N.S., 2011. Primary forests bination with GPS radio-telemetry. Biol. Conserv. 129, 487–496. 252 M.W. Tobler et al. Biological Conservation 220 (2018) 245–253 Sollmann, R., Furtado, M.M., Gardner, B., Hofer, H., Jácomo, A.T.A., Tôrres, N.M., Tobler, M.W., Carrillo-Percastegui, S.E., Pitman, R.L., Mares, R., Powell, G., 2008. An Silveira, L., 2011. Improving density estimates for elusive carnivores: accounting for evaluation of camera traps for inventorying large- and medium-sized terrestrial sex-specific detection and movements using spatial capture-recapture models for ja- rainforest mammals. Anim. Conserv. 11, 169–178. guars in central Brazil. Biol. Conserv. 144, 1017–1024. Tobler, M.W., Carrillo-Percastegui, S.E., Zúñiga Hartley, A., Powell, G.V.N., 2013. High Sollmann, R., Furtado, M.M., Hofer, H., Jácomo, A.T.A., Tôrres, N.M., Silveira, L., 2012. jaguar densities and large population sizes in the core habitat of the southwestern Using occupancy models to investigate space partitioning between two sympatric Amazon. Biol. Conserv. 159, 375–381. large predators, the jaguar and puma in central Brazil. Mammal. Biol. Z. Saugetierkd. Tobler, M.W., Zúñiga Hartley, A., Carrillo-Percastegui, S.E., Powell, G.V.N., 2015. 77, 41–46. Spatiotemporal hierarchical modelling of species richness and occupancy using Sollmann, R., Mohamed, A., Niedballa, J., Bender, J., Ambu, L., Lagan, P., Mannan, S., camera trap data. J. Appl. Ecol. 52, 413–421. Ong, R.C., Langner, A., Gardner, B., Wilting, A., 2017. Quantifying mammal biodi- Weckel, M., Giuliano, W., Silver, S., 2006. Cockscomb revisited: jaguar diet in the versity co-benefits in certified tropical forests. Divers. Distrib. 23, 317–328. Cockscomb Basin wildlife sanctuary, Belize1. Biotropica 38, 687–690. Struebig, M.J., Turner, A., Giles, E., Lasmana, F., Tollington, S., Bernard, H., Bell, D., Wilson, K.A., Meijaard, E., Drummond, S., Grantham, H.S., Boitani, L., Catullo, G., 2013. Quantifying the biodiversity value of repeatedly logged rainforests: gradient Christie, L., Dennis, R., Dutton, I., Falcucci, A., Maiorano, L., Possingham, H.P., and comparative approaches from Borneo. Adv. Ecol. Res. 48, 183–224. Rondinini, C., Turner, W.R., Venter, O., Watts, M., 2010. Conserving biodiversity in Sutherland, C., Fuller, A.K., Royle, J.A., 2015. Modelling non-Euclidean movement and production landscapes. Ecol. Appl. 20, 1721–1732. landscape connectivity in highly structured ecological networks. Methods Ecol. Evol. Yamaura, Y., Andrew Royle, J., Kuboi, K., Tada, T., Ikeno, S., Makino, S.i., 2011. 6, 169–177. Modelling community dynamics based on species-level abundance models from de- Tobler, M.W., 2015. Camera Base Version 1.7. San Diego Zoo Global. tection/nondetection data. J. Appl. Ecol. 48, 67–75. Tobler, M.W., Powell, G.V.N., 2013. Estimating jaguar densities with camera traps: pro- Zipkin, E.F., DeWan, A., Royle, A.J., 2009. Impacts of forest fragmentation on species blems with current designs and recommendations for future studies. Biol. Conserv. richness: a hierarchical approach to community modelling. J. Appl. Ecol. 46, 159, 109–118. 815–822. 253