Data collection

Sampling design

NFI Cape Verde project uses stratified sampling design. First-phase sample is based on regular systematic grid of equidistant points of 150×150 m. This sampling grid is used to estimate land-use classes’ area using manual aerial photo interpretation. Second-phase sample is based on the network of equidistant points of 150×150 m for Santiago Island, 450×450 m for Santo Antão, São Nicolau, São Vicente, Boavista, Maio, Fogo islands and 300x300m for the islands of Brava and Sal. This sampling grid is used as a ground survey to provide additional information for main land-use classes (i.e. Forest, Agroforestry, Open forest and Shrubland).

Table 1 shows number of inventory plots per islands.

Table 1: Number of inventory plots per island

* All generated plots were visited during ground survey, the difference is due to inaccessible plots or plots indicated in the field as “no-forest plots” (i.e. plots, which do not belong to one of four main forest land-use classes).

In case any of the original inventory plots were inaccessible, replacement plots were created located at a distance of 150m to the north, east, south or west from the original co-ordinates.

Land use classification

The system of classification, used to define classes of land-use in the field was based on the classification system developed by the Global Forest Resources Assessment of FAO in order to ensure harmonization between countries, in the regional or global assessments adapted to the specifications of Cape Verde.

The classification of use and soil occupation was carried out in the field on rectangular plots of 2500 m2, following the scheme on the figure 1.

Figure 1 Schema of classification of land use and soil occupation

Scheme of inventory plot

On Santiago Island variable-size circular plots were used based on land-use class identified in the field. The field team established circular plots of 1000 m2 for agroforestry and open forest land-use class, and circular plots of 500 m2 for forestry and shrubland land-use class.

On other islands all plots were circular with area of 1000 m2 regardless land-use class.

Field data survey

The location and navigation to the sample plots centre were accomplished by the aid of GNSS equipment (i.e. GPS II SXBlue), connected to a field computer that contained field application (i.e. software Field-Map Data Collector), the digital background map and sampling grid. Once in the centre, the calibration of the electronic equipments was performed (i.e. calibration of the electronic compass).

Then, the classification of land use was made in an area of 2500 m2 based on the land-use classification scheme (see Figure 1). The information about date of measurement, fieldteam and land-use classification was stored in the database.

Four pictures were taken, from the center to four major directions (North, East, South and West), in each inventory plot.

The additional information was collected for following land-use classes:

• Forest
• Open forest
• Agroforestry
• Shrubland

Table 2: Data collected on selected land-use classes

The list of attributes (variables) predefined in the database is found in the Annex 1. When a plot was not accessible a replacement plot was selected following the same procedures for data collection.

All trees in the plot with a height more than a meter were mapped and described.

The main dendrometric parameters observed were:

• Tree height (m):

The measurement was done on all trees in the plot with a height exceeding 1 meter, using a laser rangefinder/hypsometer ForestPro or laser rangefinder/hypsometer Impulse.

• Projection of the crown (horizontal):

The measurement of the tree was made ​​using at least 5 points projected from the same place where the tree was mapped using the laser rangefinder Hypsometer ForestPro and/or Hypsometer Laser rangefinder Impulse electronic compass and reflector.

As an alternative for the dense forests, when the visibility was very poor for the reflector, measurement of the projection of the crown was made ​​from the center of the tree, to the four directions (North, East, South and West) indicating their respective distance.

• Diameter at Breast Height (DBH), cm (only trees with DBH> = 5 cm):

For measurement of DBH, it was used electronic caliper DigiTech (Haglöf). The diameter measurement was made ​​during the mapping of trees; the diameter value was submitted by online communication via radio receiver, by the operator of the caliper to the computer.

– Number of trunks and mean diameter of small trunks (trunks only ≥ 2 cm and <5 cm):

All the stems whose diameter were between 2 to 5 cm, were counted and their respective mean Diameter at Breast Height was calculated based on measurement with DigiTech (Haglöf) electronic caliper;

The collected descriptive parameters of trees were:

• Species
• Tree damage (i.e. presence of stem break, animal damage, insect damage, fungi damage, other type of damage)
• Wood quality

Before leaving the plot, the team checked if all the data collected were correct and the attributes filled in the sampling design. In the same place, backups of all collected data were created.