In secr versions before 2.5.0, usage was defined as a binary value
(1 if trap $k$ used on occasion $s$, zero otherwise). In later versions, usage may take nonnegative real values and will be
interpreted as effort. This corresponds to the constant T_s used for
the duration of sampling by Borchers and Efford (2008). Effort is
modelled as a known linear coefficient of detection probability on the
hazard scale
(secr-varyingeffort.pdf;
Efford et al. 2013).
For replacement of usage, various forms are possible for value:
- a matrix in which the number of rows of value exactly
matches the number of traps K in object
- a vector of two values, the usage (typically 1) and the number of
occasions S (a K x S matrix will be filled with
the first value)
- a vector of R+1 values where R is the number of
sessions in a multi-session object and elements 2..R+1 correspond to
the numbers of occasions S1, S2,... in each session
- the usage only (typically 1) (only works when replacing an existing
usage matrix with known number of occasions).