fields homepage


UCAR \| NCAR Statistics Software Data Publications Projects
CISL \| IMAGe \| Statistics \| Contact Us \| Visit Us \| People Search	advanced

About Fields

Home

Tutorial / Demo

Contents

Related Software

fields

A collection of programs written in the R language for curve and function fitting with an emphasis on spatial data.

Authors/Contributors and Citation

Fields has drawn heavily from its predecessor, FUNFITS, to the extent that it is hard to separate who is primarily responsible for what. With that in mind, giving credit to everyone involved in FUNFITS and Fields seems appropriate. In no particular order, the Fields Development Team is: Doug Nychka, Barbara Bailey, Stephen Ellner, Perry Haaland, Michael O'Connell, Sarah Hardy, Jungmin Baik, Wendy Meiring, J. Andrew Royle, Montserrat Fuentes, Tim Hoar, Claudia Tebaldi, and Eric Gilleland.

Citing fields
Fields Development Team (2004). fields: Tools for Spatial Data. National Center for Atmospheric Research, Boulder, CO. URL http://www.cgd.ucar.edu/Software/Fields.

Availability

The R version of Fields is currently available for UNIX, Linux and Windows through the Comprehensive R Archive Network (CRAN). [See the contributed packages section]. The Windows version is available as a binary, the others can be readily built in UNIX/LINUX. Currently, due to time constraints, we no longer offer fields for Splus.

A development version

fields_2.1.1.tar.gz (posted SEP-06-2005) This may be slightly different than the version available from CRAN and not checked completely.

The major methods include:

Tps	Thin Plate spline regression
Krig	Spatial process estimation (Kriging)
krig.image	Spatial process estimate for large problems

The Kriging functions allow you to supply a covariance function that is written in native code.

cover.design	Space-filling designs
as.image image.add image.plot smooth.image	functions for working with image data
sreg, qsreg	1-D smoothing splines and 1-D quantile splines

There are also generic functions that support these methods such as:

plot	diagnostic plots of fit
summary	statistical summary of fit
print	shorter version of summary
surface	graphical display of fitted surface
predict	evaluation fit at arbitrary points
predict.se	prediction standard errors at arbitrary points.

FORTRAN verses native code

The main functions have been written to use only native code. In particular, Krig, Tps, and krig.image. In just a few places are calls made to suporting FORTRAN routines, most notably for the cubic smoothing function, sreg.