elephant poaching in space and time
Parinaz Rashidi is a PhD student in the department of Natural Resources (NRS). Her supervisor is prof.dr. A.K. Skidmore from the faculty of Geo-information Science and Earth Observation (ITC).
Poaching of elephants is an increasingly rampant problem both in Kenya and across Africa. It is a key driver of elephant population decline in certain regions of Africa. Quantifying spatial and temporal dynamics of elephant poaching risk at different scales provides the information needed for setting conservation priorities as well as for concentrating management resources where they are most needed. However, few studies exist that are aimed at quantifying the spatial and spatiotemporal patterns linked to the criminal act of wildlife poaching. Obtaining an absolute measure of poaching levels based on direct observation is practically impossible due to the covert nature of poaching. Detailed data are scarce, and many poaching reports are collected incidentally and may be only indirectly obtained. Such reports may be challenging to analyze because of the absence of sampling design and uncertainty concerning locations. It is therefore important to apply methods that can overcome the scarce data problem, which emerges when unstable estimations occur due to low counts of incidents and high sampling variation.
Using different spatial and spatio-temporal modelling methods, with the ability to estimate poaching risk in areas with zero counts by borrowing information from neighboring areas and experts, we investigated how elephant poaching risk may change at different locations, times or for combinations of space and time at a local and national level, and we tried to identify key factors influencing elephant poaching risk. Applying this kind of modelling, which provides a flexible framework for borrowing information over space and time from adjacent areas by using spatial and temporal random effects, provides an effective way to form estimates for poorly sampled areas.
Poaching data were obtained from the Kenya Wildlife Service’s database on elephant mortality. Based on the literature and prior scoping discussion, we selected the potential risk factors associated with poaching. Thirty experts from the Kenya Wildlife Service were interviewed regarding elephant and poaching These factors included: (1) distance to roads, (2) distance to settlements, (3) distance to rivers and streams, (4) density of waterholes, (5) elevation, (6) slope, (7) mean normalized difference vegetation index (NDVI), (8) standard deviation of NDVI, (9) elephant population density, (10) livestock density, (11) distance to international border, and (12) seasonal timing of elephant poaching (i.e., poaching probabilities in the dry and wet seasons).
Our results indicate that although hotspot predictions varied for the different methods, some areas were consistently identified as encompassing poaching hotspots. Moreover, our analysis at national level shows that high poaching risk areas in Kenya have varied annually, shifting from the south-east to the west between 2002 and 2012. They are shifting from areas with high recorded elephant poaching incidents to areas where minimal poaching incidents have been historically reported. The results also demonstrate that the Tsavo ecosystem is the main persistently high-risk area at national level in Kenya between 2002 and 2012. It was also found that the mean trend in elephant poaching is increasing in the Tsavo ecosystem over time. The results further showed that two blocks, located in the Tsavo west national park and the Taita ranches, have been consistently identified as poaching hotspots irrespective of the model used. Furthermore, our results indicate that areas with the highest poaching risk differ between dry and wet seasons.
In addition, the results revealed that adding risk factors enhances the model fit for assessing poaching risk. Furthermore, results obtained from the poaching risk factor analysis indicated that similarities occurred between risk factors detected at the local and national level, but also that some differences emerged. For example, the seasonal timing of elephant poaching, density of waterholes, distance to a road, and distance to an international border are all key factors persistently influencing elephant poaching risk at local and national level in Kenya. Distance to an international border seems to have the greatest influence on elephant poaching at the local level, while the distance to a road is most influential at the national level, with a variation greater than other risk factors. Moreover, it was found that distance to settlements was significantly identified as a key factor that influences the spatial patterns of elephant poaching risk at the national level while was not significant at the local level. On the other hand, livestock density was a significant predictor, which significantly influenced the spatial variation in trends regarding elephant poaching risk over time at a local level but was not significant at the national level.
These findings can be used to guide the deployment of policing resources in areas with relatively increasing poaching trends, or to improve or alter management actions. The Kenyan Wildlife Service (KWS) should continue to carry out frequent surveillance and intensify patrolling within and around the detected elephant poaching hotspots especially targeting the areas which exist along international borders and areas with human settlements in and around them. The findings could also be incorporated in future national and regional management programs to further -reduce the poaching of elephants.
