Estimation of free-roaming dog populations using Google Street View: A methodological study

Area of Science:

• Veterinary Public Health
• Geospatial Analysis
• Citizen Science

Background:

• Accurate quantitative knowledge of dog populations is essential for controlling zoonotic diseases like rabies.
• Traditional dog population estimation methods are resource-intensive, time-consuming, and potentially hazardous.
• Developing efficient and reliable remote estimation techniques is crucial for public health initiatives.

Purpose of the Study:

• To evaluate the efficacy of Google Street View (GSV) as a remote sensing tool for estimating free-roaming dog populations.
• To compare GSV-based dog counts with traditional door-to-door (D2D) survey data.
• To assess the feasibility of a citizen science approach for dog population estimation.

Main Methods:

• Recruited and trained citizen scientists via social media to identify and count free-roaming dogs using GSV imagery.
• Collected GSV data across 20 urban and 6 peri-urban communities in Peru.
• Employed correlation metrics and negative binomial models to compare GSV counts with D2D survey data.

Main Results:

• Citizen scientists identified 862 dogs via GSV, with an adjusted estimate of 1,022 free-roaming dogs.
• A strong positive correlation (r=0.85, p<0.001) was found between GSV and D2D counts in urban areas.
• A weak correlation (r=0.36, p=0.478) was observed in peri-urban areas, with challenges due to image availability.

Conclusions:

• Google Street View, combined with citizen science, presents a viable and promising method for estimating dog populations, especially in urban environments.
• This remote sensing approach can supplement traditional methods, offering valuable data for disease control programs in resource-limited settings.
• The effectiveness of GSV may vary depending on community type and the availability of high-resolution imagery.