G uided transect sampling is primarily intended for the sampling of sparse, geographically
scattered, populations for which there exist no list of the units. Basically, it consists of a twostage design, using wide strips in the first stage and a subsampling procedure in each strip in
the second stage. The subsampling is guided by prior information, e.g. in the form of remote
sensing image data. Different strategies can be used for the guidance, resulting in different
probabilities of inclusion of population units, and consequently in slightly different
estimators
The general principle for second stage subsampling guidance can be coupled with a
number of methods for how the samples should be selected along the "guided route". Strip
sampling, line transect sampling, adaptive cluster sampling, and plot sampling are examples
of methods that can be used. However, in the theoretical set-up of the method, it is assumed
that all objects in grid-cells passed by the survey transect are found. The grid-cells, covering
the entire area under study, also contain the covariate data that are used for directing the
sampling effort.