The -cress_lola option is similar to the
-lola option in that it creates a regular LOngitude-LAtitude grid.
The former uses a Cressman analysis and the latter option uses a nearest-neighbor interpolation.
You need to specify the lower
left corner of the grid, the number of points in the zonal and meridional directions
and the latitude/longitude increments. Finally you need to specify the output file
and the format. WARNING: winds and other vector fields will not be
rotated. If the vector fields use a grid relative orientation,
then your interpolated winds will be using the original grid.
The interpolation to the lola grid is by a Cressman analysis. The Cressman analysis
is a multipass system which uses a user-specified "radius" for each pass.
A Cressman analysis can be computationally expensive so you may want to explore
-cress_lola LonSW:#lon:dlon LatSW:#lat:dlat file radius1:radius2:..:radiusN
LonSW Longitude of the South-West point, values from 0 .. 360
#lon number of longitude points
dlon spacing of the points in the zonal direction in degrees
LatSW Latitude of the South-West point, values from -90 .. 90
#lat number of latitude points
dlat spacing of the points in the meridional direction in degrees
file name of the output grib file
radiusM The radius in km for M-th pass.
Prelim - Cressman Analysis
defintion: input grid = observations = grid from the input grib file
output grid = grid as defined by LonSW:#lon:dlon LatSW:#lat:dlat
(1) compute mean of observations,save mean on background grid (output)
background(i,j) = average(observation)
(2) Repeat N times: PASS-M
(a) find background value for observations by bilinear interpolation of background grid
(b) remove the background from the observations: obs' = obs - interpolated_background
(c) inc'(i,j) = weighted average of obs'
inc'(i,j) is increment on output grid
weighted average of obs' depends on the distance between grid point and obsservation
(d) new background = background + inc'
Warning, this scheme doesn't handle handle rotated winds in a useful manner.
There will also be problem with analyzing winds near the poles.