Skip to content

Optimal interpolation client

Thomas Nipen edited this page Jun 18, 2020 · 7 revisions

Command-line example

When the input file is deterministic, or does not contain enough ensemble members (specified by minEns), the deterministic mode is automatically used. In this case, a static covariance structure is used.

Parameters

The parameter file contains the observation at points as well as information about its uncertainty and covariance structure (for precipitation from radar). The first parameter is always the observation. The second is optional and represents a scaling factor for sigma. A value of 0.5 indicates that a particular observation has half the standard error. For precipitation, a third parameter can be used, which is simply a flag that indicates that the observation error covaries with all other observations for which the flag is set. This is needed for radar values, since these errors are often correlated in space.

Settings for different variables

MET Norway uses the following settings for temperature:

d=10000 h=100 sigma=0.5 delta=2 wmin=0.5 elevGradient=-0.0065

MET Norway uses the following settings for precipitation:

d=10000 h=-999 sigma=0.15 sigmaC=0.15 wmin=1 dc=3000 maxElevDiff=-999 elevGradient=0 transform=boxcox

Clone this wiki locally