Re: axis points

[Chris Weaver <weaver@xxxxxxxxxxxxxxxxxxxxxxxxxxx>]

Hi,

Steve Hankin kindly provided me with some tools to determine what exactly 
happens when you perform certain types of regridding operations.  Specifically, 
I wanted to know, when transforming an hourly data set to a daily-mean data set 
using regridding, what the "central time" of the average was and which 
particular hourly points went into each daily average.

For example, if I define the following daily-mean axis:

define axis/t="1-jul-1989:15:00:00":"31-jul-1989:15:00:00":1/units=days t1day
let omega_daily = omega_hourly[gt=t1day@ave]

I can find out more about what happened using the following (Steve's tools):

LET ONE = omega_hourly*0 + 1           ! always one
LET LINEAR = omega_hourly*0 + L      ! linear in L
let delta_pt = some-L-value
LET DELTA = IF L EQ delta_point THEN 1    ! = 1 only at delta_point

So,
    one[gt=t1day@sum]   -- tells you how many points went into the average
    linear[gt=t1day@ave] -- tells you the center point of the averaging
    delta[gt=t1day@sum] -- tells you if the value at "delta_pt" is used in a
calculation

(You have to be careful about mixing world coordinates and simply numbered 
indices in the use of these tools to avoid incompatibilities using certain 
transformations.)

I hope this might be helpful to some out there.

Chris Weaver


-------------------
C.P. Weaver                                Internet:
Department of Environmental Sciences        weaver@cep.rutgers.edu
Rutgers University/Cook College             www.envsci.rutgers.edu/~weaver
14 College Farm Road                       Phone: 1 732 932 7902
New Brunswick, NJ 08901-8551               Fax:   1 732 932 8644 
USA