Forest and tree survey sampling design

Forest and tree survey sampling design:

Forest inventory refers commonly to measurements of several important parameters of forests and trees and to the analysis of abundance and distribution of individuals. The inventory in forestry is, in most cases, based on sampling. Different options of sampling design exist to survey a given forest area in a country. Each option is chosen to fit given characteristics of the surveyed population and to satisfy specific needs of information on some selected parameters for a given budget. Random sampling is commonly used. It may be simple random sampling with or without stratification. It may also be double or multistage sampling, double or multiphase sampling, etc. Systematic sampling may also be unrestricted or stratified.

An inventory may be based on temporary or permanent sample plots depending on the objectives of the survey. Monitoring of resources relies often on permanent sample plots. In order to improve precision of results temporary plots may be added to existing permanent plots in successive surveys.

For a number of motives including practicality, data quality, cost implications, the systematic sampling was chosen for forest and tree data collection. The systematic sampling in forest inventory is: (i) easily planned, (ii) faster in execution and mostly cheaper; (iii) it gives better estimates of the mean than unrestricted random sampling and even stratified random sampling in large areas, because the variation which may be considerable in such areas is better represented in the sample for which the distribution of the sample plots within the surveyed population is homogeneous; (iv) it gives thus better precision compared to random sampling. Ir P.J.D Versteegh, (1976); Lee White and Ann Edwards, (2000);

The selection of the sample site (or tract ) population has been done on the basis of the latitude/longitude grid. Tract location is chosen in the intersections of degrees. If a higher number of tracts is required in a country, additional tracts may be selected every half degree or even every quarter degree. The density of tracts may be higher on one axis than on the other. The density of tracts may vary also according to the defined strata. For the purpose of standardising the approach, the south-west corner of the tract is taken to correspond to the intersection of the lat/long lines.

Figure 1: Illustration of systematic distribution of sample sites in the country

The sample site of 1 km2 area where a cluster of 4 sample plots are located is called 'Tract'.

The sampling design is built up around the following principles:

i) Stratification based on forest types and land use classes is discarded as strata may change over time and render stratified sampling irrelevant with permanent samples. But stratification on basis of relatively stable strata such as ecological zoning may be applied to countries with pronounced differences in ecosystems such as humid and dry forests (Cameroon), desertic and non-desertic climate (Algeria), etc.

ii) The sampling is designed to include tree resources outside forest and an array of biophysical and management/uses parameters of forest and TOF.

iii) The approach foresees establishing sample plots to monitor changes over time between forests and other land use classes and the processes within the forests;

iv) Biophysical data and the management and use information will be geo-referenced and integrated to enable assessing the size and state of the resources in relation to their geographic location in the country and their social context;

v) Data quality is guaranteed and measurement methods are harmonised;