Table 2 Challenges, advantages and uncertainties for each process of sampling strategy

Step 1: Delimitation of the study area

Mastery of the of Google Earth software

Knowing the boundaries of the study area and the remarkable morphological features (hydrography, orography, main arteries, stadiums, etc.)

Choice of the entire study area itself upstream

Is done randomly

Overview of the city and all potential sampling areas regardless of their surface area

The possibility of encountering areas without habitats

Step 2: Neighbourhoods selection

Mastery of Google Earth software

Delimited areas do not follow a regular grid

The distances between areas are randomly selected and obey only random sampling criteria

The choice is random and easy to make with the R software

Easy data cleaning process involving the previous steps

Possibility to select at the finest spatial resolution

Independence in the choice of sites

No redundancy in the choice of neighbourhoods

Do not know in advance the number of lots/neighbourhoods that will emerge

Step 3: Household selection

Have a good internet connection speed

Precise

The possibility of avoiding bias by clearly identifying habitats

Advanced knowledge of the number of houses to select

Random sampling and easy to make with R software

Step 4: Integration of points in the GPS

Mastering the conversion of file formats (Kml to GPX)

Automatically done from Qgis to GPS

Step 5: Field campaign

Training the investigator to master the use of GPS and map reading

Easy recognition on fields of previously identified sites and locations

Quite inexpensive because the investigator does not waste time in identifying the field

Possibility to find new habitats set up between the date of the images and the date of the fieldwork