Regional and local components in land-use sampling: A case-study from the Brazilian Triangulo

Abstract

Using maps of the forest distribution, a number of selected standard recording methods were tested for their degree of reliability for an area of 10,000 square kilometers in central England. This area was chosen because its forest density is similar to that found in parts of the Brazilian Triângulo, and also to provide a verified estimate of the deviation from the actual value to be expected from the sample survey. The survey was performed at two scales: regional, on 1000 sq km blocks, and local, on 25 sq km blocks. For the regional survey, the survey covering 4% of the surface based on stratified but otherwise randomly drawn lines gave an average deviation value of ±1%; the local recording, which was carried out using transverse, randomly drawn lines, resulted in a deviation value of a similar magnitude. It was concluded that these drafts from selected recordings are sufficiently accurate to be used in the Brazilian Triângulo.