[GRASS-user] SamplingStrata

Matteo Mura mur.teo at gmail.com
Thu Apr 3 08:11:38 PDT 2014


Dear all,

I am going to quickly explain you what I am trying to achieve.

I have a grid of 4693 hexagons (see
https://www.dropbox.com/s/rkim7v21i1p0bet/grid.jpeg). Each hexagon has
its univocal id (ID_ES).

We are planning a 2-phase sampling procedure. In the first phase we
identified some feature inside each hexagon (namely trees out of
forest features). Now we should carry out the second phase with ground
survey. Obviously we cannot visit each hexagon but just a sample of
them.

In doing so we need to divide the hexagon grid in strata composed by
compact (i.e. not too much elongated in vertical or horizontal
direction) groups of adjacent hexagons, of course without replacement,
inside of each strata an hexagon will be randomly choosen and visited.

This should be applied for different sampling rates (SR):

- SR 50% => divide the map in (4692/2=) 2345 strata composed by 2
adiacent hexagons + 1 strata of 3 hexagons, and get as a result a
vector table where each row contains the ID_ES of the 2 hexagons,
e.g.:

ID_ES1 | ID_ES2
2345 | 1654

(please note that the ID_ES is not spatially progressive, in the sense
that the ID_ES=1 is not adiacent to ID_ES=2).

Same thing for the others different SR:

- SR 25% => (4692/4=) 1172 strata composed by 4 adiacent hexagons + 1
starta of 5 hexagons, and get as a result a vector table where each
row contains the ID_ES of the 4 hexagons

- SR 10% => 466 strata of 10 adiacent hexagons + 3 strata of 11
hexagons and get as a result a vector table where each row contains
the ID_ES of the 10 hexagons + 2 rows with the ID_ES of the 11
hexagons.

Same procedure for SR of 5% (221 strata of 20 hex + 13strata of 21
hex) and 1% (40 strata of 100 hex + 7 strata of 99 hex).

My question is if there is a way to create these strata with its
vector tables in GRASS?

Hope I have been clear.

Thanks in advance!

Matteo


More information about the grass-user mailing list