PhD research project

The Gola forest is one of the largest remaining fragments of the Upper Guinean Rainforest of West Africa. The forest lies on the border between Sierra Leone and Liberia and, over the past century, has been subject to strong anthropogenic pressures which include logging and mining as well as subsistence and commercial hunting. These activities are threatening the incredibly rich biodiversity of the region.

 

The Gola Rainforest National Park (GRNP) in Sierra Leone has been protected since December 2011. The Gola Forest National Park (GFNP) in Liberia was inaugurated at the time of study, in April 2018.

 

North of the GFNP in Liberia, the community forest is part of the GolaMa project for alternative livelihoods. The project, implemented by the Society for Conservation of Nature Liberia (SCNL) and the Royal Society for the Protection of Birds (RSPB), aims at reducing bushmeat hunting and providing financial loans within the community to encourage alternative income activities to reduce pressure on the forest.

Study Site

Gola Forest of Liberia and Sierra Leone

My research spanned across the Gola region, across different levels of forest enforcement (GRNP - GFNP - Community forest).

A pilot study was conducted in Liberia in May and June 2017, and the core data collection period was between October 2017 and June 2018.

Study species

IMG_9094.jpg

Diana monkey Cercopithecus diana

  • Arboreal and diurnal

  • Restricted to high forest canopy in undisturbed forest, highly vocal species

  • Habitat sensitivity: high

  • Home range: 0.5 - 1 km2 

  • Body mass: 3.5 - 5 kg (medium size)

  • Group size: 15 - 25 individuals

  • Major threats: habitat loss and hunting

Q1 What are the main commonalities and what the key differences between maps of hunting pressure built using different methods?

Amongst anthropogenic pressures threatening biodiversity at a global scale, hunting for wild meat or trade is becoming a significant one, alongside habitat destruction and degradation (caused by agricultural intensification, logging and mining amongst other things). With an increasing human population, hunting activities pose a strong threat to prey species, as the off-take is often too high to allow for population persistence over time. The use of weapons such as shotguns and the large commercial scale of hunting, alongside poor law enforcement are all major reasons why hunting pressure increasingly threatens many species.

The majority of primate species in West Africa are showing steady declines in numbers, largely due to the high demand for protein in growing human populations. But how is hunting pressure estimated? How do we know how much hunting is going on in a particular region? What signs can tell us something about the level of hunting?

Different methods have been used to estimate hunting pressure. My first research chapter builds maps of hunting pressure distribution from data collected using different methods to see how they compare and to evaluate which methods are more accurate.

1) Distance from human infrastructure

 

This method estimates the level of hunting pressure based on the distance from the closest inhabited area. The idea is that hunters will exert strongest pressure in areas that are closer to home. Humans are central-place foragers, which means that they return to their homes (central-place) with hunted prey. The main issue with this method is that areas around villages/towns are often depleted of resources and hunters are forced to head further for successful hunting trips. 

3) Gunshot frequency

 

The main method used to hunt primates in West Africa is with shotguns. Using passive acoustic monitoring (PAM) as a method to record gunshot frequency across a landscape is a growing body of research and is a promising resource to aid ranging patrols in National Parks. Currently, the main limitations for this method are the expense and the battery life of the devices but cheaper alternatives are available, accessible to a broader research audience. I used Audiomoth recording devices to record gunshot frequencies across the Gola landscape.

2) Hunting signs

The encounter rate and density of hunting signs can be calculated for a specific area of interest and compared with that of other areas. The main issue with this method is that it is labour intensive as it requires the area to be surveyed appropriately. Knowledge of hunting signs is also important as some may be old and not be representative of current levels of hunting pressure.

Hunters will always leave traces of their presence across the landscape. Whether a hunting trail, a cartridge or a trap, all these signs can be used as indicators of hunting pressure. 

4) Participatory mapping

 

Working within communities and asking people in which areas they hunt more frequently, and using that information to construct a map, is another method used to estimate hunting pressure. This relies heavily on accurate reporting by the hunters and produces qualitative data that has to be groundtruthed to produce a spatial map of hunting pressure.

Q2 How do behavioural responses to human presence change along a gradient of hunting pressure?

Animals often adapt their behaviours to external cues. Human hunting represents a strong predatory pressure for primates across West Africa and different species are expected to adapt their behaviour accordingly to avoid predation. Anti-predatory behaviours, in fact, may change depending on predator and surrounding habitat (see Density Estimates page).

If primates indeed adapt their behaviour to human hunting pressure, one might expect behavioural variation across a landscape to be a good indicator of hunting pressure. Many other factors however might be influencing the observed behaviour. 

My second and third chapters explore the effects of hunting pressure on monkey behaviour.

Q3 How do changes in behaviour impact density estimates?

The most widespread method used to estimate primate densities is distance sampling (line transect sampling). This method consists in dividing the region of interest into lines that are walked at a slow speed (around 0.5km/h). For each detected animal the perpendicular distance from the line is measured (Figure 1). The distances can then be used to model a detection function, with detectability decreasing as a function of distance, i.e. the majority of animals that are detected are close to the line (Figure 2). From the detection function, number of animals missed can then be estimated and consequently the density of animals in the area of interest (Equation below). One of the main assumptions of this method is that detectability is only a function of distance.

 

 

Hunting pressure by humans has been shown to affect the behaviour of animals, including primates. In areas of high hunting pressure, primates have been shown to alter their anti-predatory behaviours. There is evidence, for example, of a decrease in vocalisation in Diana monkeys, of an increasing in freezing and fleeing behaviours and some evidence for reduced group size (Figure 3). All of these behavioural changes may have an impact on species detectability.

 = estimated density

n = number of animals seen

Pˆa = estimate of proportion of animals seen

w = width of the transect

= total length of all the transects

Using a model simulation, my final research chapter adds behavioural parameters (derived from the literature and from my fieldwork in West Africa) such as movement (fleeing) to a simulated population with known density, to make predictions on how this behavioural change may impact detectability and final density estimates.

Results from this research may result in more precise density estimates across a landscape with varying levels of hunting pressure and may help make more informed decisions regarding estimates of hunting sustainability and how species are categorised under international conventions (i.e. IUCN red list).

Figure 1. Line transect of length L and width W. Filled dots represent detected primates and x is the perpendicular distance between the line and that primate/s. Empty dots represent the undetected ones.

Figure 2. Example detection function - The probability of detecting a primate will decrease with distance from the line

Figure 3. Predicted effects of changes in anti-predator behaviours on the detectability of primate species

The different data chapters from my PhD research are currently in prep for submission to multiple scientific journals, watch this space for results and publications!!