blockPredict {intamap}

Spatial block prediction

Description

blockPredict is a generic method for prediction of spatially aggregated variables within the intamap-package package.

Usage

 blockPredict(object, ...) 

Arguments

object

a list object of the type described in intamap-package

...

other arguments that will be passed to the requested interpolation method. See the individual interpolation methods for more information. The following arguments from object can be overrun through ...:

block

Block size; a vector with 1, 2 or 3 values containing the size of a rectangular in x-, y- and z-dimension respectively (0 if not set), or a data frame with 1, 2 or 3 columns, containing the points that discretize the block in the x-, y- and z-dimension to define irregular blocks relative to (0,0) or (0,0,0) - see also the details section of predict.gstat. By default, predictions or simulations refer to the support of the data values.

cellsize

size of cells for discretization of blocks for numerical simulation

Details

The function blockPredict is a wrapper around the spatialPredict.block function within the intamap-package package, to simplify the calls for block predictions.

Block predictions are spatial predictions assumed to be valid for a certain area. The blocks can either be given by passing SpatialPolygons as the predicitonLocations or by passing the block-argument through the parameters of the object or through the ...-argument.

There are esentially two ways to solve the problems of block predictions.

analytical

block predictions can be found directly by block kriging

numerical

block predictions can be found through numerical simulations over a set of points within the block, the requested output is found by averaging over these simulations

The analytical solutions are used when applicable. This is typically for ordinary kriging based methods and prediction types that can be found by linear aggregation (e.g. block mean).

If the prediction type necessitates simulations, this is done by subsampling the blocks. This can either be done block-wise, with a certain number of points within each block, with a certain cellsize, or with a certain number of points

automap Uses function autoKrige in the automap package. If object already includes a variogram model, krige in the gstat-package will be called directly.

Value

a list object similar to object, but extended with predictions at a the set of locations defined object.

Author(s)

Jon Olav Skoien

References

Pebesma, E., Cornford, D., Dubois, G., Heuvelink, G.B.M., Hristopulos, D., Pilz, J., Stohlker, U., Morin, G., Skoien, J.O. INTAMAP: The design and implementation f an interoperable automated interpolation Web Service. Computers and Geosciences 37 (3), 2011.

See Also

gstat,autoKrige

Examples

# This example skips some steps that might be necessary for more complicated
# tasks, such as estimateParameters and pre- and postProcessing of the data
data(meuse)
coordinates(meuse) = ~x+y
meuse$value = log(meuse$zinc)
data(meuse.grid)
gridded(meuse.grid) = ~x+y
proj4string(meuse) = CRS("+init=epsg:28992")
proj4string(meuse.grid) = CRS("+init=epsg:28992")

# set up intamap object:
obj = createIntamapObject(
	observations = meuse,
	predictionLocations = meuse.grid[sample(1:length(meuse.grid),10),],
	targetCRS = "+init=epsg:3035",
	class = "automap"
)

# do interpolation step:
obj = conformProjections(obj)
obj = estimateParameters(obj) 
obj = blockPredict(obj,block=c(100,100))  # blockPredict


# intamap object for which simulation is needed:
meuse$value = meuse$zinc
obj = createIntamapObject(
	observations = meuse,
	predictionLocations = meuse.grid[sample(1:length(meuse.grid),5),],
	params = list(ngrid = 16),
  class = "transGaussian"  # trans-Gaussian kriging method
)
obj = estimateParameters(obj, lambda = 0)  # lambda is optional, lambda = 0 gives lognormal kriging
obj = blockPredict(obj,block=c(100,100))  # blockPredict