Large MAMMALS

Photo © Conservation International/Photo by John Martin

biodiversity sampling protocols 37
Medium and large-sized mammals
Maíra Benchimol
Universidade Estadual de Santa Cruz, PPG Ecologia e Conservação da Biodiversidade, Applied Ecology & Conservation Lab. Ilhéus, Bahia, Brazil.
Email: [email protected]

Introduction
Taxon Definition – Medium to large-bodied mammals, encompassing all terrestrial and arboreal non-
volant species but excluding small terrestrial mammals.

Suitability of the Taxa for Rapid Surveys – Large mammals are considered good bioindicators of intact
tropical landscapes and have therefore been increasingly used in large-scale monitoring programs
worldwide (Ahumada et al. 2011; Luzar et al. 2011; Nobre et al. 2013). They play vital roles in ecosystem
structure and functioning, participating in different trophic levels in food webs, contributing to herbivore
regulation, and acting as important seed dispersers for many tree species (Terborgh 1992). They are
also a vital economic resource for local human populations through their use as food, pets, artefacts
and tourism (Peres 2000; Costa et al. 2005). Indeed, mammals comprise an important source of protein
and income to local communities, especially the large-bodied species given their great amount of meat
(Redford 1992; Peres 2000). Moreover, they are widely hailed as regional conservation icons (e.g.,
pandas), as many species are charismatic and benefit from popular sympathy, which contributes to
promote conservation to the wider public (Cuaron 2000; Dirzo et al. 2014). Regardless of their appeal,
many mammal populations have gone extinct and many others are declining, requiring surveys and
monitoring that can inform conservation action to hinder continued population declines.

Methodology – The diurnal line-transect census is a well-recognized and cost-effective methodology for
surveying medium and large vertebrates in tropical forests and savannas (White 1994; Peres 1999; Carrillo
et al. 2000; Haugaasen and Peres 2005; Ogutu et al. 2006; Effion et al. 2013). It has the great advantage
of low financial costs, as only few items are required for surveys (see in ‘Equipment for surveys’ below).
Additionally, species identification is obtained directly during the surveys, without requiring collection of
any type of material for processing. The sampling protocol is simple, easily replicated, results comparable
among different sites, and provides reliable data for density estimates for several species (Peres 1999;
Haugaasen and Peres 2005). Line-transect census methodology is considered the only appropriate
technique to survey arboreal species (such as primates), yet provide low number of records for elusive and
rare species (Munari et al. 2011). Hence, combining line-transect censuses with indirect surveys (including
fresh tracks, feces, hair, burrows, digging and partly consumed fruits/seeds) can enhance detectability for
many mammal species, contributing to maximize the species lists.

38 biodiversity sampling protocols Medium and large-sized mammals

This chapter provides an easy, replicable and standardized sampling protocol to survey medium to
large-bodied mammals for rapid biological assessments in tropical forests and savannas, based on
diurnal line-transect censuses and indirect surveys. Additionally, it is recommended to use this protocol
concomitantly with the camera trapping protocol, to enable estimates of population-level abundances
and richness for a greater number of medium and large-bodied mammal species.

Core Methods
Sampling Protocols

Line-transect preparation – within each major habitat type (site), three 4-km long and 1-m wide linear
transects should be cut before the start date of surveys. The number of sampling sites and distances to
each other may have to vary according to the total area of the study landscape. Ideally, transects should
be established at least a week before the start of rapid assessments so that human disturbances do
not affect mammal behavior and results. At this step, transects should be measured (using a Hip-Chain®
or a 50-m tape) and marked (using a biodegradable flagging tape) every 50 m. Within each sampling
site, transects should be separated from each other by at least 1 km, and their location should take into
account accessibility, including the existence of rivers, streams and topography, that might hamper the
surveys. It is best to open transects more than 300 m from the base camp to avoid biased data due to
any species behavioral responses to camp activity. Shorter transects may be necessary in fragmented
forest sites where space constraints prevent long trails. Within fragmented forest landscapes, the length
and arrangement of transects should consider both area and shape of each forest patch, aiming to cover a
representative area (50% of a patch would be adequate).

Prior to the start of the surveys, a field sheet should be prepared to enable data records during the data
collection (Appendix I).

Diurnal line-transect census – two observers, preferably one trained researcher accompanied by a
local inhabitant with knowledge of species present, should walk at a constant speed (~1 km/h), with brief
stops (10 s) every 500 m, along each of the three transects established at each site (Peres 1999; Peres
and Cunha 2011). Transects should be walked in both directions, for a total of 24 km of sampling effort
per day (3 transects walked simultaneously x 8 km). In savannas, surveys can also be conducted using
a vehicle, at approximately 10 km/h. Surveys should be conducted in the morning (~6:15 – 10:30) and
afternoon (~14:00 – 17:30), and discontinued during rainy periods since these can affect results.

At the start of each survey, the lead observer should record the date, transect identity, name of
observers, general weather condition (sunny, overcast or cloudy) and start time. Observers will then start
walking along the transect, keeping a distance of ~15 m from each other, looking for target species in all
strata (in case of forest habitats) and on both sides of the transect (Fig. 1).

Medium and large-sized mammals biodiversity sampling protocols 39

Upon a visual detection event, observers should record: the time, species name, number of individuals,
sighting location along the transect, and the perpendicular distance from the animal (or first detected
individual, in case of groups) to the trail, which needs to be accurately measured (Hip-Chain® or a 50-m
tape; Fig. 1 and Appendix I). It is important that the observers see or hear the animal(s) before they
detect the observers – otherwise, the perpendicular distance may be inflated, directly affecting density
estimates. For each detection event, observers should remain on the transect line and spend no more
than 10 minutes to count individuals and record the data. The end time of each walk should be recorded
at the end of each morning and afternoon survey. In order to minimize biases related to the probability
of detection, each pair of observers should be rotated on a daily basis between different transects. The
number of sightings per km walked (sighting rate) should be used as a measure of abundance (both
for groups and individuals), and density estimates can be calculated using specific programs such as
Distance (Buckland et al. 1993). Probability of occupancy can be assessed by using a matrix of presence-
absence data per survey for each species, using programs such as Presence (Hines 2006; see Box I for
parameter terminology).

Figure 1
Demonstration of a line-transect census methodology conducted by two observers upon an animal detection,
showing in red some of the parameters to be recorded on the field sheet.

40 biodiversity sampling protocols Medium and large-sized mammals

Indirect surveys – concomitantly to
the diurnal line-transect surveys, the BOX I: Terminology
same two observers should search for
Abundance – number of individuals (or groups) per distance walked
any indirect evidence of target species
Density – number of individuals (or groups) per area
(Fig. 2), along and up to 5 m from the Richness – total number of species
transect. Local field guides should be Occupancy – presence of a focal species (also known as occurrence)
used to identify mammal tracks. Acoustic
records of identifiable species could
also be recorded. The field sheet can be
the same as that used to conduct line- Equipment for surveys
transect surveys, but the perpendicular
distance is not recorded as this Item Quantity
methodology cannot discern accurate Field sheets Several
density estimates (see Appendix I). Watch 01 per pair
However, indirect surveys may enhance Optical range finder or a 50-m tape 01 per pair
the number of mammal species recorded Binoculars 02 per pair
within a site, and enable occupancy Digital camera 01 per pair*
estimates that can be used to detect Field guides (color plates), if available 01 per pair*
changes over time. * Suggested but not essential.

Figure 2
Examples of indirect signs: an armadillo burrow, feces, hair and a footprint of different mammal species in an
Amazonian forest site.

Medium and large-sized mammals biodiversity sampling protocols 41

Selecting sampling sites – The location of sampling sites should cover a representative area of the study
landscape. A first required step is to obtain satellite images of the study landscape to acquire knowledge of
habitat distribution, existence of rivers, local villages and other site characteristics. Next, a visit to the area
should be performed prior to the survey in order to select the sampling sites. This visit should be used to
assess logistical challenges (feasibility for surveys), habitat heterogeneity (focus on the most representative
habitat types), anthropogenic disturbance (depending on the goals of the survey, which may be to focus
on intact sites, or to assess disturbances such as logging, fire and hunting), and accessibility (if the access
occurs by boat, transects should start close to rivers and streams in order to reduce time walking prior to
each survey). All transects should not traverse aquatic realms inaccessible by foot.

Effort required – A minimum of seven days of two-way surveys (i.e., morning and afternoon) along
each transect at each sampling site is required. This will provide a total of 168 km of cumulative effort.
This effort is expected to provide robust species richness for each sampling site – previous studies in
Neotropical forests recorded up to 93% of all species (extrapolated richness) considering a total survey
effort of 80-90 km (de Thoisy et al. 2008). However, some cryptic species are difficult to detect even
with higher sampling efforts, although abundances can be obtained with such effort, and provide a good
proxy of communities status. For occupancy models, one week of surveys will provide 14 ‘visits’ on the
presence-absence matrix, which is potentially adequate for analyses of site occupancy for most species.

For density estimates, however, a minimum of 40 detection events are recommended for robust
estimates, although 20 sightings may provide sufficient estimates (Peres 1999). If small sample sizes were
obtained at the end of rapid surveys, data from different sites can be pooled together to enable density
estimates using the Distance software. Sighting rates can also be calculated and compared among
different landscapes independent of sample size.

Context-Dependent Sampling Considerations

Environmental disturbances – Anthropogenic forms of disturbances such as hunting, logging and fire are
likely to affect the distribution, occurrence, density and detectability of several species within a site (Peres
2001, Benchimol and Peres 2015). Hence, it is recommended to consider these factors during the selection
of sampling sites and establishment of transects, which can be used to focus on undisturbed areas or to
assess environmental impacts. The level of hunting pressure on mammals in the study landscape should be
assessed through interviews with local people. This can be done by showing color plates or photographs
of all target species to local residents that might have access to the study sites, and assign a level of
hunting pressure to each site into three classes: [1] non-hunted; [2] lightly or occasionally hunted; and [3]
heavily hunted. Additionally, information on local hunting patterns, including name of hunted species,
number of hunters per family, and frequency of hunting events should be also assessed. These categories
of hunting pressure may be used for comparisons with other sites.

42 biodiversity sampling protocols Medium and large-sized mammals

Habitat considerations – For indirect surveys, the detectability of signs varies on different substrates and
biophysical conditions. Sandy soils can easily retain footprints, whereas seasonally-flooded and rocky habitats
are unfavorable for conducting sign surveys. Additionally, lowland habitats are prone to contain higher number
of signs for several species in tropical forests than plateau habitats. Weather condition is also an important
factor likely to affect this sampling performance, given that heavy rains can rapidly remove footprints.

Seasonal considerations – If only one rapid assessment can be conducted within the study landscape,
surveys should be conducted in the dry season (ideally less than 100 mm of rain for sites exhibiting
high seasonality). If feasible, two rapid assessments should be conducted during a year, one in the
peak of the dry season and another one in the peak of the wet season. Seasonality can determine
spatiotemporal shifts in mammal faunas, directly affecting detection probabilities for different species
(Bodmer 1990; Peres 1997; Lehman 2006). This occurs because a greater number of tree species is
producing fruits during the rainy season (Haugaasen and Peres 2005), which results in higher fruit
exploitations and therefore higher daily activity of frugivores at that period of the year, likely enhancing
detectability of several species. For instance, a lemur primate species was not recorded during the
dry season in Madagascar and most species showed significantly higher density estimates in the rainy
season compared to the same line-transect census protocol in the dry season (Lehman 2006).

In several Amazonian forests subjected to intense flooding regime, the prolonged seasonal inundation
strongly affects patterns of habitat use for different mammal species, directly affecting patterns of
species abundance (Branch 1983; Peres 1997; Haugaasen and Peres 2005). It is therefore expected that
surveys conducted in the dry season provide lower abundance estimates for several species and may
also reduce the detection of others. However, during the rainy season it is more difficult to perform rapid
surveys by conducting line-transect surveys, given the great probability of losing days of collection due
to the weather conditions.

Supplementary methods – Nocturnal line-transect census methodology is frequently used as a

complementary technique to survey nocturnal species. It has the disadvantage of visual limitation for the
observers during data collection, usually providing poor information for most night-time species (Munari
et al. 2011). However, this sampling technique can provide reliable density estimates for some arboreal
nocturnal species (Nekaris et al. 2008; Thornton et al. 2011) and should therefore be incorporated into
the sampling protocol for sites harboring a high diversity of arboreal nocturnal species. The sampling
protocol is similar to diurnal line-transect surveys, except that walking pace should be reduced to 400
m/h, and flashlights/headlamps are required. One-directional surveys are sufficient at night (4 km per
transect per day), starting at 18:00 hs, which will provide 84 km of sampling effort per site.

Footprint trap stations are another non-invasive method commonly used to survey medium and large
mammals (Wemmer et al. 1996; Lyra-Jorge et al. 2008). This permits both presence-absence and
abundance data for several terrestrial, cryptic species. Sandy soils should be disposed at each 2m2 (1 x 2 m)
track plot, set out at 500 m intervals along each transect. Once a day, each track plot should be checked,
footprints should be photographed to scale and measured, and then should be carefully cleared.

Medium and large-sized mammals biodiversity sampling protocols 43

Additionally, rapid interview surveys can provide information on species richness at a study landscape. It
is a cost-effective methodology, which enable acquisition of species occupancy and enhance community
participation (Parry and Peres 2015). Standardized questions on species presence/absence likely to
occur in the study area are asked to those local inhabitants with many years of local knowledge of
wildlife. This can be done concomitantly with the interviews on hunting pressure.

The protocol presented here should be used in combination with the camera trapping protocol for
obtaining a reliable biological assessment of an area for the target taxa. A more complete list of species
composition within the study landscape is likely to be obtained if diurnal line-transect surveys, indirect
surveys, footprint trap stations, interviews, and camera trapping are simultaneously adopted.

Data Management
The species identifications are obtained directly from the surveys. Data should be carefully recorded and
stored in an electronic data file, which needs to contain the following information for each survey:
• date
• observers name
• site name
• unique transect identifier
• general weather condition
• name of each recorded species and associated data
• time of record
• type of record (direct or indirect)
• number of individuals
• location along the transect
• perpendicular distance of the animal to the transect

Conservation Implications
Medium and large mammals are key components of forest and savanna communities and are therefore
considered good indicators of ecosystem health. Hence, data obtained from rapid surveys can provide
information on the current quality of the study landscape for current mammal populations. Because
anthropogenic disturbances are likely to affect occurrence and abundance of mammal species, these
surveys will also contribute to understanding the human impacts on mammal assemblages, identifying
local and global patterns of change. Additionally, data will contribute to assessments of the conservation
status of individual species, providing data on distribution, mapping the occurrence of endemic species,
and helping in the assessments of local and global threat of many species. Results are intended to support
programs related to conservation strategies – contributing, for instance, to the creation of protected areas
and development of new projects focused on protection/management of threatened species.

44 biodiversity sampling protocols Medium and large-sized mammals

Limitations
Rapid assessment techniques provide means of accelerating the acquisition of scientific data in
understudied areas. For medium and large mammals, however, this is not an easy task. The group
requires a high survey effort to achieve comprehensive sampling, due to the distribution, occupancy and
detectability of different species. A combination of different techniques is therefore required to enhance
detectability of different species and obtain an adequate estimation of the magnitude of species richness
and abundance. Nevertheless, all sampling techniques for the group present limitations for rapid surveys.
Diurnal line transect censuses are likely to provide good data for diurnal primates, but might provide few
records for many terrestrial species in rapid surveys. The inclusion of indirect surveys in the protocol
contributes towards gathering occupancy data for these terrestrial species, although the method fails in
providing density estimates.

Several assumptions of the line-transect census methodology are challenging (Buckland et al. 1993,
2001; Marshall et al. 2008). Firstly, sufficient and independent sightings are required for the estimation
of a detection function necessary for generating density analyses. Sampling effort needed can vary with
species and location, depending on their detectability, variation in group size and abundance. Secondly,
individuals should be detected with certainty, and measured accurately from their initial location to enable
density estimates. Lastly, good weather conditions are essential for census walks, thus rainy days preclude
data acquisition. For indirect surveys, detectability is related to the local substrate and the ability of
surveyors to detect and identify signs. Again, weather conditions might compromise data acquisition.

Acknowledgments
I sincerely thank three reviewers who provided constructive comments on the chapter.

Photo © Trond H. Larsen

Medium and large-sized mammals biodiversity sampling protocols 45

Literature Cited
Ahumada, J.A., Silva, C.E.F., Gajapersad, K., Hallam, C., Hurtado, J., Martin, E., McWilliam, A., Mugerwa, B., O’Brien, T., Rovero,
F., Sheil, D., Spironello, W.R., Winarni, N. and Andelman, S.J. 2011. Community structure and diversity of tropical forest
mammals: data from a global camera trap network. Phil. Trans. R. Soc. B. 366: 2703-2711.

Benchimol, M. and Peres, C.A. 2015. Predicting local extinctions of Amazonian vertebrates in forest islands created by a mega
dam. Biological Conservation 187: 61-72.

Bodmer, R.E. 1990. Responses of ungulates to seasonal inundation in the Amazon floodplain. Journal of Tropical Ecology 6:
191-201.

Branch, L. C. 1983. Seasonal and habitat differences in the abundance of primates in the Amazon (Tapajos) National Park,
Brazil. Primates 24: 424-431.

Buckland, S.T., Anderson, D.R., Burnham, K.P. and Laake, J.L. 1993. Distance sampling: estimating abundance of biological
populations. Chapman & Hall, London.

Buckland, S.T., Anderson, D.R., Burnham, K.P., Laake, J.L., Borchers, D.L. and Thomas, L. 2001. Introduction to distance
sampling: estimating abundance of biological populations. Oxford: Oxford University Press.

Carrillo, E., Wong, G. and Cuaron, A.D. 2000. Monitoring mammal populations in Costa Rican protected areas under different
hunting restrictions. Conservation Biology 14: 1580-1591.

Costa, L.P., Leite, Y.L.R., Mendes, S.L. and Ditchfield, A.D. 2005. Mammal conservation in Brazil. Conservation Biology 19: 672-
679.

Cuaron, A.D. 2000. A Global perspective on habitat disturbance and tropical rainforest mammals. Conservation Biology 14:
1574-1579.

De Thoisy, B., Brosse, S. and Dubois, M.A. 2008. Assessment of large-vertebrate species richness and relative abundance in
Neotropical forest using line-transect censuses: what is the minimal effort required? Biodiversity and Conservation 17: 2627–
2644.

Dirzo, R., Young, H.S., Galetti, M., Ceballos, G., Isaac, N.J.B. and Collen, B. 2014. Defaunation in the Anthropocene. Science
345:401-406.

Effiom, E.O., Nuñez-Iturri, G., Smith, H.G., Ottosson, U. and Olsson, O. 2013. Bushmeat hunting changes regeneration of African
rainforests. Proc R Soc B 280:20130246.

Haugaasen, T. and Peres, C.A. 2005. Mammal assemblage structure in Amazonian flooded and unflooded forests. Journal of
Tropical Ecology 21:133-145.

Hines, J.E. 2006. PRESENCE. Page Software to estimate patch occupancy and related parameters. U.S. Geological Survey,
Patuxent Wildlife Research Center, Laurel, Maryland.

Lehman, S.M. 2006. Effects of transect selection and seasonality on lemur density estimates in southeastern Madagascar.
International Journal of Primatology 27: 1041-1057.

Luzar, J.B., Silvus, K.M., Overman, H., Giery, S.T., Read, J.M. and Fragoso, J.M.V. 2011. Large-scale environmental monitoring by
indigenous people. BioScience 61:771-781.

Lyra-Jorge, M.C., Ciocheti, G., Pivello, V.R., Meirelles, S.T. 2008. Comparing methods for sampling large- and medium-sized
mammals: camera traps and track plots. Eur J Wildl Res 54:739–744.

Marshall, A.R., Lovett, J.C. and White, P.C.L. 2008. Selection of line-transect methods for estimating the density of group-living
animals: lessons from the primates. American Journal of Primatology 70:452–462.

Munari, D.P, Keller, C. and Venticinque, E.M. 2011. An evaluation of field techniques for monitoring terrestrial mammal
populations in Amazonia. Mammalian Biology 76:401-408.

46 biodiversity sampling protocols Medium and large-sized mammals

Nekaris, K.A.I., Blackham, G.V. and Nijman, V. 2008. Conservation implications of low encounter rates of five nocturnal primate
species (Nycticebus spp.) in Asia. Biodiversity and Conservation 17:733-747.

Nobre, R.A., Pereira, R.C., Roque, F.O., Constantino, P.A.L, Sabino, J. and Uehara-Prado, M. 2013. Monitoramento in situ da
biodiversidade: uma proposta para a composição de um sistema brasileiro de monitoramento da biodiversidade. Brasília/DF:
ICMBio.

Ogutu, J.O., Bhola, N., Piepho, H.P. and Reid, R. 2006. Efficiency of strip- and line-transect surveys of African savanna
mammals. Journal of Zoology 269: 149-160.

Parry, L., and Peres, C.A. 2015. Evaluating the use of local ecological knowledge to monitor hunted tropical-forest wildlife over
large spatial scales. Ecology and Society 20: 15. http://dx.doi.org/10.5751/ES-07601-200315.

Peres, C.A. 1997. Primate community structure at twenty western Amazonian flooded and unflooded forests. Journal of
Tropical Ecology 13: 381-405.

Peres, C.A. 1999. General guidelines for standardizing line-transect surveys of tropical forest primates. Neotropical Primates
7:11-16.

Peres, C.A. 2000. Effects of subsistence hunting on vertebrate community structure in Amazonian forests. Conservation
Biology 14:240-253.

Peres, C.A. 2001. Synergistic effects of subsistence hunting and habitat fragmentation on Amazonian forest vertebrates.
Conservation Biology 15:1490–1505.

Peres, C. A. and Cunha, A.A. 2011. Manual para censo e monitoramento de vertebrados de médio e grande porte por
transecção linear em florestas tropicais. Wildife Conservation Society, Ministério do Meio Ambiente e ICMBio, Brasil.

Redford, K.H. 1992. The empty forest. BioScience 42: 412-422.

Terborgh, J. 1992. Maintenance of diversity in tropical forests. Biotropica 24:283-292.

Thornton, D., Branch, L. and Sunquist, M. 2011. Passive sampling effects and landscape location alter associations between
species traits and response to fragmentation. Ecological Applications 21:817-829.

Wemmer, C., Kunz, T., Lundie-Jekins, G., McShea, W. 1996. Mammalian Sign. In: Wilson DE, Cole FR, Nichols JD, Rudran R,
Foster MS (eds) Measuring and monitoring biological diversity. Standard methods for mammals. Smithsonian Institution, USA,
pp 157–176.

White, L.J.T. 1994. Biomass of rain forest mammals in the Lope Reserve, Gabon. Journal of Animal Ecology 63:499-512.

Medium and large-sized mammals biodiversity sampling protocols 47