deplet(catch = NULL, effort = NULL, method = c("l", "d", "ml",
"hosc", "hesc", "hemqle", "wh"), kwh=NULL, nboot = 500)l= Leslie estimator, d= effort corrected
Delury estimator, ml= maximum likelihood estimator of Gould and Pollock (1997), hosc= sampling coverage
wh. NULL for all possible capture parameters.ml,
hosc,hesc, and hemqle methods.out are created for each method and contain tables of various statistics associated with the method.l) estimator (Leslie and Davis, 1939), the effort-corrected Delury (d) estimator (Delury,1947; Braaten, 1969),
the maximum likelihood (ml) method of Gould and Pollock (1997), sample coverage estimator for the homogeneous model (hosc) of Chao and Chang (1999),
sample coverage estimator for the heterogeneous model (hesc) of Chao and Chang (1999), and the maximum quasi-likelihood estimator for the heterogeneous model (hemqle) of Chao and Chang (1999).
The variable effort models can be applied to constant effort data by simply filling the effort vector with 1s. Three removals are required to use the Leslie, Delury, and Gould
and Pollock methods.
The constant effort model is the generalized removal method of Otis et al. 1978 reviewed in White et al. (1982: 109-114).
If only two removals, the two-pass estimator of N in White et al. (1982:105) and the variance estimator of Otis et al. (1978: 108) are used.
Note: Calculation of the standard error using the ml method may take considerable time.
For the Delury method, zero catch values are not allowed because the log-transform is used.
For the generalized removal models, if standard errors appear as NAs but parameter estimates are provided, the inversion of the Hessian failed.
If parameter estimates and standard errors appear as NAs, then model fitting failed.
For the Chao and Chang models, if the last catch value is zero, it is deleted from the data. Zero values between positive values are permitted.data(darter)
deplet(catch=darter$catch,effort=darter$effort,method="hosc")
hosc.outRun the code above in your browser using DataLab