Sampling grid
Sampling in the course of our national statistical forest inventory (NFI) occurs at the intersection points of a national grid with a 4 × 4 km mesh. The establishment of the sampling grid relied on our previous Forest Protection Network (FPN), which was set up in 1988. Sampling intensity was doubled by shifting grid nodes in two directions (SE-NE), and yielded a basic mesh (marked 'A') and an offset mesh (marked 'B'). Both comprise 4 × 4 squares and grid lines intersect each other at midpoint. Meshes 'A' and 'B' are rotated at 45 degrees to yield a 2.828 × 2.828 grid. Several rounds of sampling have been performed at the nodes that happen to be located on forest land of this systematic set of points. In the 2010-2014 forest inventory cycle, sampling was performed at the nodes of mesh 'A'.
Nodes
The full cycle was completed by sampling at 1/5 of the nodes of mesh 'A' each year in a cycle of 5 years. This covered in-terrain evaluation of fifteen hundred sampling points and the recording of the sub-set of points that satisfied the conditions of sampling.
| Year | Number of sampling points | |
|---|---|---|
| Total amount of potential sampling points | Sampling points visited on the field | |
| 1. cycle | ||
| 2010 | 4176 | 1396 |
| 2011 | 4659 | 1 460 |
| 2012 | 4694 | 1622 |
| 2013 | 4643 | 1307 |
| 2014 | 4176 | 1396 |
| 2. cycle | ||
| 2015 | 4637 | 1490 |
| 2016 | 4653 | 1581 |
| 2017 | 4650 | 1375 |
| 2018 | 4654 | 1465 |
| 2019 | 4638 | 1514 |
Taking into account expectations concerning spatial distribution and representativeness, the nodes of a sampling year have been selected as follows: The colour coding indicates the areas represented by tracts to be sampled in a certain year of sampling with coordinates indicated in kilometres.
A random section of Hungary in the NFI sampling grid
Theoretical grid nodes are at a distance of:
| Period | Mesh size | Represented area |
|---|---|---|
| 1 year | 8,9 × 8,9 | 2 000 ha |
| 5 year | 4 × 4 | 400 ha |
| 10 year | 2,8 × 2,8 | 200 ha |
Once sampling is completed across all of the nodes of mesh 'B', the outcomes of both grids will become simultaneously available.
SEL has tracts superimposed on each grid node. That way 4 sampling plots can be developed at each node, which are fundamentally independent of each other in terms of sampling. Tracts are not to be treated as single sampling units and their only purpose is to improve sampling efficiency. In each case, it is the SW corner point of a tract that corresponds to a node on mesh 'A' or 'B'. Tract identifiers can be derived this way (XXX-YYY). Individual sampling plots can be identified with tract and compass point information (XXX-YYY-N). As individual sampling plots in a tract are independent of each other, the number of sampled plots can vary between 0-4 depending on the initial and in-terrain rating of plots.
A schematic drawing of a tract with sampling circles at the corner points