krige                package:sgeostat                R Documentation

_K_r_i_g_i_n_g

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

     Carry out spatial prediction (or kriging).

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

     krige(prdpnt, point.obj, v1, var.mod.object,maxdist=NULLi,extrap=F) 

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

  prdpnt: a point object, generated by `point()', at which prediction
          is carried out

point.obj: a point object, generated by `point()', containing the
          sample points and data

      v1: the variable, contained in `point.obj', for which prediction
          will be carried out

 maxdist: an optional maximum distance. If entered, then only sample
          points (i.e, in point.obj) within maxdist of each prediction
          point will be used to do the prediction at that point. If not
          entered, then all n sample points will be used to make the
          prediction at each point.

  extrap: logical, indicates if prediction outside the convex hull of
          data points should be done, default `FALSE'

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

     A point object which is a copy of the prdpnt object with two new
     variables, `zhat' and `sigma2hat', which are, repspectively, the
     predicted value and the kriging variance.

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

     http://www.gis.iastate.edu/SGeoStat/homepage.html

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

     `est.variogram',`fit.variogram'

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

     # a single point:
     prdpnt <- point(data.frame(list(x=180000,y=331000)))
     prdpnt <- krige(prdpnt, maas.point, 'zinc', maas.vmod)
     prdpnt

     # kriging on a grid (slow!)
     grid <- list(x=seq(min(maas$x),max(maas$x),by=100),
                  y=seq(min(maas$y),max(maas$y),by=100))
     grid$xr <- range(grid$x)
     grid$xs <- grid$xr[2] - grid$xr[1]
     grid$yr <- range(grid$y)
     grid$ys <- grid$yr[2] - grid$yr[1]
     grid$max <- max(grid$xs, grid$ys)
     grid$xy <- data.frame(cbind(c(matrix(grid$x, length(grid$x), length(grid$y))),
                  c(matrix(grid$y, length(grid$x), length(grid$y), byrow=T))))
     colnames(grid$xy) <- c("x", "y")
     grid$point <- point(grid$xy)
     grid$krige <- krige(grid$point,maas.point,'zinc',maas.vmod,
                         maxdist=1000,extrap=F)
     op <- par(no.readonly = TRUE)
     par(pty="s")
     plot(grid$xy, type="n", xlim=c(grid$xr[1], grid$xr[1]+grid$max),
                         ylim=c(grid$yr[1], grid$yr[1]+grid$max))
     image(grid$x,grid$y,
           matrix(grid$krige$zhat,length(grid$x),length(grid$y)),
           add=T)
     contour(grid$x,grid$y,
             matrix(grid$krige$zhat,length(grid$x),length(grid$y)),
             add=T)
     data(maas.bank)
     lines(maas.bank$x,maas.bank$y,col="blue")
     par(op)

