Overview and Brief History - Sampling Spatial Units for Agricultural Surveys - page 47

trails, perennial crops, vineyards, orchards, and grassland annual crops. Finally,
artificial regions involve green artificial, paved, constructed or altered artificial.
For the predominant cover indicator, a grid is considered to have predominant
cover if more than 50 % of its points belong to a specific type of cover. For the
spatial organization indicator, each grid section is considered separately and a
statistical method is used to determine the number of proximities between the
three classes. Depending on the predominant proximities, each grid s land cover

is classified as homogeneous or heterogeneous. For the trend indicator, standard
trends are defined according to modifications in the frequency and homogeneity of
the three components.

'

Conclusions
Changes to land cover and land use are important for capturing the evolution
or dynamics of the countryside. The most feasible approach to surveying land
resources on a national scale is to use statistical sampling. A complete survey
is simply too expensive. A sampling based survey can also be repeated at
fixed intervals to provide information about changes in land resources.
During recent years, the use of spatial reference frames has become more
attractive for defining the sampling design. This approach appears to be suited
to the analysis of agricultural data, which are often geographically distrib-
uted. In this chapter, we have outlined some survey examples that make
effective use of spatial reference frames. This list is obviously not exhaustive.
For other examples, the reader can refer to FAO ( 1998 ). It contains a very
comprehensive description of spatial frame techniques (as well list survey
methods), and includes chapters that consider separate countries and their
individual implementations of geo-coded frames and crop surveys.
However, the first constraint when appropriately organizing a spatial
frame is the availability of up-to-date cartographic material (i.e., maps,
satellite images, aerial photos) that covers all the required land. The resolu-
tion of this material must be sufficient for stratification according to, for
instance, the proportion of land cultivated or predominance of certain crops.
These topics, which are of crucial importance to the definition of effective
spatial sampling for agricultural data, will be described in Chap. 3 (the
Geographic Information System) and Chap. 4 (remote sensing tools).

References

Arbia G (1989) Spatial data configuration in statistical analysis of regional economic and related
problems. Kluwer, North Holland
Arroway P, Abreu DA, Lamas AC, Lopiano KK, Young LY (2010) An alternate approach to
assessing misclassification in JAS. In: Proceedings of the section on survey research methods
JSM 2010, American Statistical Association, Alexandria, VA