wi                package:adehabitat                R Documentation

_C_o_m_p_u_t_a_t_i_o_n _o_f _S_e_l_e_c_t_i_o_n _R_a_t_i_o_s _f_o_r _H_a_b_i_t_a_t _S_e_l_e_c_t_i_o_n _S_t_u_d_i_e_s.

_D_e_s_c_r_i_p_t_i_o_n:

     These functions compute the resource selection ratios (wi) for
     design I, II and III data types, with resources defined by several
     categories. Basic tests are also provided.

_U_s_a_g_e:

     widesI(u, a, avknown = TRUE, alpha = 0.05)
     widesII(u, a, avknown = TRUE, alpha = 0.05)
     widesIII(u, a, avknown = TRUE, alpha = 0.05)
     print.wiI(x, ...)
     print.wiII(x, ...)
     print.wiIII(x, ...)
     plot.wi(x, caxis = 0.7, clab = 1, ylog = FALSE, errbar = c("CI", "SE"),
             main = "Manly selectivity measure", noorder = TRUE, ...)

_A_r_g_u_m_e_n_t_s:

       u: for 'widesI', a vector with named elements describing the
          sample of used resource units.  For 'widesII' and 'widesIII'
          a matrix or a data frame giving the number of used resource
          units for each animal (in rows) in each resource category (in
          columns)

       a: for 'widesI' and 'widesII', a vector with named elements
          describing the sample or the proportion of available resource
          units.  For 'widesIII' a matrix or a data frame giving the
          number or the proportion of available resource units for each
          animal (in rows) in each resource category (in columns)

 avknown: logical. 'TRUE' if the available proportions are known, and
          'FALSE' if they are estimated

   alpha: the threshold value for the tests and confidence intervals

       x: an object of class 'wi'

   caxis: character size on axes to be passed to 'par("cex.axis")'

    clab: character size of axes labels to be passed to
          'par("cex.lab")'

    ylog: logical.  If 'TRUE', the selection ratios are plotted on a
          log scale

  errbar: a character string.  Type of error bars: either '"CI"' for
          confidence intervals or '"SE"' for standard errors

    main: a character string. The title of the graph

 noorder: logical. If 'TRUE', the habitat categories are ordered on the
          graph in decreasing order of their preference. If FALSE, they
          are not ordered (i.e. they are in the same order as the
          columns in 'used' and 'available'

     ...: additionnal arguments to be passed to the function 'plot'

_D_e_t_a_i_l_s:

     'widesI' may be used to explore resource selection by animals,
     when designs I are involved (habitat use and availability are
     measured at the population level - animals are not identified). 
     The function tests habitat selection with the Khi2 of Pearson and
     log-likelihood Khi2 (recommended, see Manly et al. 2003).  The
     Manly selectivity measure (selection ratio = used/available) is
     computed, the preference / avoidance is tested for each habitat,
     and the differences between selection ratios are computed and
     tested. 

     'widesII' computes the selection ratios with design II data (same
     availability for all animals, but use measured for each one).
     Tests of identical habitat use for all animals, and of habitat
     selection are also provided.

     'widesIII' computes the selection ratios for design III data (when
     the use and the availability are measured for each animal -  see
     examples on the wild boar below).  Habitat selection is tested
     using a Chi-square for each animal, and the overall habitat
     selection is also tested.

_V_a_l_u_e:

     'widesI' returns a list of the class 'wiI'.  'widesII' returns a
     list of class 'wiII'.  'widesIII' returns a list of class 'wiIII'.
      These objects are all inheriting from the class 'wi'.  They have
     the following components: 

used.prop: the proportion of use for each resource type.

avail.prop: the proportion of available resources.

      wi: the Manly selectivity measure (selection ratio:
          used/available).

   se.wi: the standard error of the selection ratios.

comparisons: a list with the following components: .in +5

     '_d_i_f_f_w_i' a matrix with the differences of the selection ratios for
          each pair of resource type.

     '_I_C_d_i_f_f_u_p_p_e_r' a matrix containing the upper limit of confidence
          interval on the differences of the selection ratios between
          each pair of resource type.

     '_I_C_d_i_f_f_l_o_w_e_r' a matrix containing the lower limit of confidence
          interval on the differences of the selection ratios between
          each pair of resource type.

     '_s_i_g_n_i_f' the ranking matrix, with the sign of the differences
          between the resource type in row and the resource type in
          column.  When the difference is significant, the sign is
          tripled.

 profile: the profile of preferences: resource types are sorted so that
          the left type is the most preferred and the right type is the
          most avoided.  Habitats for which the selection ratios are
          not significant are connected by a line.

   alpha: the parameter 'alpha' of this function.

 avknown: the parameter 'avknown' of this function.

 se.used: only for designs I, the standard error of the proportion of
          use.

se.avail: only for designs I, the standard error of the available
          proportion.

chisquwi: only for designs I, the results of Chi-Square tests of the
          hypothesis that the selection ratios are in average equals to
          zero.

      Bi: only for designs I, equals to 'wi/sum(wi)'.

   Khi2P: only for designs I, test of random resource use (Pearson
          statistic).

   Khi2L: only for designs I, test of random resource use
          (Log-likelihood statistic).

  Khi2L1: only for designs II, test of identical use of habitat by all
          animals.

  Khi2L2: only for designs II, test of overall habitat selection.

Khi2L2MinusL1: only for designs II, test of hypothesis that animals are
          on average using resources in proportion to availability,
          irrespective of whether they are the same or not (= Khi2L2 -
          Khi2L1).

     wij: only for designs II and III, a matrix with the selection
          ratios for all animals and all resource categories.

ICwiupper: only for designs II and III, the upper limit of the
          confidence intervals on the selection ratios.

ICwilower: only for designs II and III, the lower limit of the
          confidence intervals on the selection ratios.

  Khi2Lj: only for designs III, the test of habitat selection for each
          animal.

   Khi2L: only for designs III, the global test of overall habitat
          selection.

_A_u_t_h_o_r(_s):

     Clment Calenge calenge@biomserv.univ-lyon1.fr

_R_e_f_e_r_e_n_c_e_s:

     Manly B.F.J., McDonald L.L., Thomas, D.L., McDonald, T.L. &
     Erickson, W.P. (2003) _Resource selection by animals - Statistical
     design and Analysis for field studies. Second edition_ London:
     Kluwer academic publishers.

     Thomas D. L. and Taylor E. J. (1990) Study designs and tests for
     comparing resource use and availability. _Journal of Wildlife
     Management_, *54*, 322-330.

_S_e_e _A_l_s_o:

     'compana' for another analysis with this kind of data.

_E_x_a_m_p_l_e_s:

     ############################
     ## Example of moose (Manly et al., 2003, p.52)
     ## Known available proportions on design I data
     moose.avail <- c(0.34, 0.101, 0.104, 0.455)
     moose.used <- c(25, 22, 30, 40)
     names(moose.used) <- c("InBurnInterior",
                          "InBurnEdge",
                          "OutOfBurnEdge",
                          "OutOfBurnFurther")
     names(moose.avail) <- names(moose.used)
     ## Computation of wi
     (wiRatio <- widesI(moose.used, moose.avail))

     ## plot the values of the selection ratios
     opar <- par(mfrow=c(2,2))
     plot(wiRatio)

     par(opar)



     ############################
     ## Example of Elk (Manly et al., 2003, p.62)
     ## Estimated available proportions on design I data
     elk.avail <- c(15, 61, 84, 40)
     elk.used <- c(3, 90, 181, 51)
     names(elk.used) <- c("0%", "1-25%", "26-75%", ">75%")
     names(elk.avail) <- names(elk.used)
     ## Computation of wi
     (wiRatio <- widesI(elk.used, elk.avail, avknown=FALSE))

     ## plot the values of the selection ratios
     opar <- par(mfrow=c(2,2))
     plot(wiRatio)

     par(opar)



     ############################
     ## Example of Bighorn (Manly et al., 2003, p.67)
     ## Known available proportions on design II data
     data(bighorn)
     ## Computation of wi
     (wi <- widesII(bighorn$used, bighorn$availT, alpha = 0.1))

     ## plot the values of the selection ratios
     opar <- par(mfrow=c(2,2))
     plot(wi)


     ############################
     ## Example of Bighorn (Manly et al., 2003, p.74)
     ## Estimated available proportions on design II data
     ## Computation of wi
     (wi <- widesII(bighorn$used, bighorn$availE, avknown = FALSE, alpha = 0.1))

     ## plot the values of the selection ratios
     plot(wi)

     par(opar)

     ############################
     ## Example of Wild boar
     ## Estimated available proportions on design III data
     data(puechdesIII)
     used <- puechdesIII$used
     available <- puechdesIII$available

     ## calculation of the selectio ratios
     ## with sampled availability
     (i <- widesIII(used,available, avknown = FALSE, alpha = 0.1))

     opar <- par(mfrow = c(2,2))
     plot(i)

     par(opar)

