Early detection of human impacts using acoustic monitoring: An example with forest elephants

Abstract

The impacts of human activities and climate change on animal populations often take considerable time before they are reflected in typical measures of population health such as population size, demography, and landscape use. Earlier detection of such impacts could enhance the effectiveness of conservation strategies, particularly for species with slow population growth. Passive acoustic monitoring is increasingly used to estimate occupancy and population size, but this tool can also monitor subtle shifts in behavior that might be early indicators of changing impacts. Here we use data from an acoustic grid, monitoring 1250 km2 of forest in northern Republic of Congo, to study how forest elephants (Loxodonta cyclotis) assess risk associated with human impacts across a landscape that includes a national park as well as active and inactive logging concessions. By quantifying emerging patterns of behavior at the population level, arising from individual-based decisions, we gain an understanding of how elephants perceive their landscape along an axis of human disturbance. Forest elephants in relatively undisturbed forests are active nearly equally day and night. However, they become more nocturnal when exposed to a perceived risk such as poaching. We assessed elephant perception of risk by monitoring changes in the likelihood of nocturnal vocal activity relative to differing levels of human activity. We show that logging is perceived to be a risk on moderate time and small spatial scales, but with little effect on elephant density. However, risk avoidance persisted in areas with relatively easy access to poachers and in more open habitats where poaching has historically been concentrated. Increased nocturnal activity is a common response in many animals to human intrusion on the landscape. Provided a species is acoustically active, passive acoustic monitoring can measure changes in human impact at early stages of such change, informing management priorities.

Fig 1. Study area in northern Republic of Congo.

50 acoustic recorders (forest type: black = mixed, yellow = monodominant, orange = open) in southern national park and adjacent logging concession (darker purple active logging, lighter purple inactive logging). Basemap files courtesy of ESRI. ESRI reserves the right to grant permission for any other use of the image.

Fig 2. Zones of logging activity in the active logging stratum.

Exploitation areas for each year were drawn as minimum convex polygons around the actual sites of felled trees. Logging roads (red), rivers (blue), acoustic recording sites dots (black = mixed forest site, magenta = monodominant forest site). Active logging was situated southeast of Nouabalé-Ndoki National Park (green area).

Fig 3. Acoustically detected gunshot events by stratum.

Bar heights are standardized for total hours of recording in each stratum/year. Actual number of shot events over each bar. inactive logging—blue, active logging—red, national park—green.

Fig 4. Change over time in proportion nocturnal activity across the study system.

All comparisons between years significant at p<0.001 except year one and two (ns). Error bars are SE of the LSM.

Fig 5. Plots of marginal means from binomial models applied to data from each stratum.

Error bars are standard errors. Note that ordinant axes differ in scale, but all have a reference line at 50% probability of a call being at night. For all strata the plotted relationships were significant predictors of nocturnal calling. Interaction plots of forest type vs season: blue = dry season, red = wet season.

Fig 6. Behavioral response to exposure to logging activities in the active logging stratum.

Error bars are SE. Most contrasts significant at p<0.05 (ns contrasts: pre-exposure vs done 1^styear, done 2^ndyear; active vs done 2^ndyear, done 4^thyear; done 3^rdyear–done 5^thyear).

Fig 7. Call density effects on nocturnal behavior.

Shaded areas are 95% confidence intervals. Blue = mixed forest, green = monodominant forest, red = open forest. Computed at study year = 2.

Fig 8. Predicted probability of night calls as call density changes in active logging, sliced by levels of logging exposure.

Computed at study year = 2, forest type = mixed, with 95% confidence intervals indicated by shading.

Fig 9. Change in numbers of calls near recording sites versus exposure to logging activity.

Error bars are SE. Pre-exposure vs done 4^th yr, p = 0.04; active vs done 4^th yr, p<0.001; all other contrasts ns.

Fig 10. Relative call frequency in the three strata over 3.5 years of study.

Calls per week are standardized (see methods). blue = inactive logging, green = national park, red = active logging.