- Mar 21, 2001
>===== Original Message From Barbara.Hock@... =====

[snip]

>The aim of the work is to compare three sites representing

in >the 2nd rotation site. Sample sites of approx 5ha have been located, and

>pre-plantation/undisturbed conditions, 1st rotation plantation trees (Pinus

>radiata), and 2nd rotation trees. An area has been found where examples of

>these sites occur close to one another on the same soil type, elevation and

>slope. The confounding factor is the site preparation that was used: for a

>number of reasons, the topsoil gets pushed together into long mounds about 2

>metres apart. These mounds are obvious in the 1st rotation site, but less so

an >unbalanced hierachical sampling schema was decided on (from Webster and>Olivier, if I recall correctly):

Since the origin of the grid is chosen at random, you may use standard

>24 primary centres are located at the nodes of a 40m grid of random origin

>all other plots are at random orientation from each other according to the

>following schema:

>- from the primary centre, locate a 2nd plot site at 13.5m (random

>orientation)

>- from each of these plot sites, locate a 3rd set of sites at 4.5m

>- and from each of the above, locate a 4th set at 1.5m

>This gives a total of 144 plots per study site.

estimators from sampling theory (no need for spatial modelling). The chosen

design is rather complex, with 4 stages as i can see from your description.

Perhaps it can be simplified to a 2-stage random systematic design and then

use well known estimators found in, for example Cochran 1977 Sampling

Techniques, page 279.

On the other hand, spatial modelling can be undertaken under the transitive

method, in which the structural tool is the transitive covariogram, where the

estimation variance of the zonal mean is computed by considering all estimates

which can be obtained by shifting the grid rigidly across the region (Petitgas

1993 Ices J Mar Sci 50:285-298, Petitgas and Lafont 1997 EVA2: Estimation

Variance version 2, this is the manual of the EVA software, widely used to

estimate the geoestatical zonal mean and its variance in the fisheries

community).

Since the purpose is hypothesis testing (the comparison of the 3 situations

with regard to pine forest disturbance) you will need to obtain interval

estimates for the zonal mean, and see if those intervals overlap, and

therefore assume a distribution for the estimator. I think the normal

distribution can be assumed in both the spatial transitive framework and the

standard sampling theory.

Ruben

--

* To post a message to the list, send it to ai-geostats@...

* As a general service to the users, please remember to post a summary of any useful responses to your questions.

* To unsubscribe, send an email to majordomo@... with no subject and "unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list

* Support to the list is provided at http://www.ai-geostats.org - << Previous post in topic