Spatial prediction of dog population distribution in Kenya

Area of Science:

• Veterinary epidemiology
• Spatial analysis
• Public health

Background:

• Free-roaming dogs are a public health concern due to disease transmission, particularly rabies, which is endemic in Kenya.
• Existing methods for estimating dog populations in Kenya are not standardized or sustainable.
• Accurate population estimates are crucial for effective disease control strategies.

Purpose of the Study:

• To develop and apply spatial interpolation techniques for estimating the free-roaming dog population in Kenya.
• To identify key environmental and demographic factors influencing dog distribution.
• To create a spatial framework to support rabies control and public health interventions.

Main Methods:

• Kriging and co-kriging spatial interpolation techniques were employed to predict dog distribution.
• Environmental (temperature, NDVI) and demographic (human density) predictors were integrated into the models.
• Village-level dog population data were collected via an online survey of veterinary professionals across 34 counties.
• A spherical semivariogram model was used, incorporating covariates to refine spatial predictions.

Main Results:

• Kenya's free-roaming dog population was estimated at 7.46 million, with a median national density of 12.13 dogs/km².
• The co-kriging model incorporating human density demonstrated the best fit, minimizing prediction errors.
• Dog densities were lower in arid, sparsely populated pastoral areas and higher in peri-urban, agricultural, and densely populated regions.

Conclusions:

• Spatial interpolation, particularly co-kriging with human density, provides a robust method for estimating free-roaming dog populations.
• The study generated a valuable spatial distribution map of dogs in Kenya.
• This framework is essential for targeted rabies control programs and public health interventions in Kenya and similar endemic regions.