Ashton,
I have encountered the same issue dealing with random field simulations
of precipitation. My thought, though I have not yet tried it, would be
to look at the distribution of the extremes of the observed values and
to randomly select from them for the backtransform max/min values.
Values that were too low (in the case of the max values) would have to
be discarded; similarly for the mins if they were too high. It's very
clear that the selection of the value makes a significant difference. My
thought was 'back-of-the-envelope' as well…
Regards,
Tom
Ashton Shortridge wrote:
Hi all,
I am employing colocated cokriging to develop realizations of elevation for a
study area. I have scattered reference elevations and exhaustive, but
definitely inaccurate, gridded elevations serving as the secondary data. I'm
using GSLIB's sgsim module to do this. To conform with the Gaussian framework
the data (hard and soft) were transformed to normal deviates using GSLIB's
nscore prior to the simulation.
The question arises on back-transforming the realizations. GSLIB's backtr
program requires min and max values for the tail extrapolation. I adopted the
following approach to estimate these values:
1) Develop a linear model using OLS:
reference = B0 + B1 * secondary elevation
2) Identify the minimum and maximum elevations in the exhaustive secondary
dataset
3) Use the model coefficients to predict the reference values for those min
and max elevations; plug those predictions into the parameter file for
backtr.
This is admittedly a pretty back-of-the-envelope approach, and I'd welcome the
list's thoughts or suggestions.
Thanks,
Ashton
--
Thomas E Adams
National Weather Service
Ohio River Forecast Center
1901 South State Route 134
Wilmington, OH 45177
EMAIL: [EMAIL PROTECTED]
VOICE: 937-383-0528
FAX: 937-383-0033
+
+ To post a message to the list, send it to [email protected]
+ To unsubscribe, send email to majordomo@ jrc.it with no subject and "unsubscribe
ai-geostats" in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the
list
+ As a general service to list users, please remember to post a summary of any
useful responses to your questions.
+ Support to the forum can be found at http://www.ai-geostats.org/
