4.1 IMPROVEMENTS IN FOREST INVENTORIES
4.2 FIELD MEASUREMENTS FOR DEVELOPING BIOMASS REGRESSION EQUATIONS
The data requirements for estimating aboveground biomass density are basically no different from those for estimating volumes from forest inventories. Careful measurement and reporting of all tree diameters and volume in inventory plots as part of a statistically sound sampling design are the basic primary data needs, data that foresters have been collecting for decades. To improve the data for biomass estimation, all inventories should be conducted according to some agreed-upon standard. The effort and resources involved in conducting forest inventories is difficult to match by other field data collection programs, thus it is important that as new inventories are planned the need for new types of data be considered, such as change in biomass for use in making greenhouse gas emission inventories. By including a few extra measurements or analyses at marginal costs, so much more information useful for biomass estimation can be obtained. This section will focus on what additional measurements need to be taken to improve the utility of forest inventories for biomass estimation. As with all inventories, the forested area under study will need to be stratified into distinct strata, and the corresponding areas estimated. The product of biomass per ha for a given strata and the corresponding area will result in an estimate of the total biomass for the region.
· As a minimum, all trees of all species, whether presently commercial or not, to a minimum diameter of 10 cm or lower should be measured. Palm trees should also be inventoried where they occur. In drier forest formations or secondary forests, a larger proportion of the biomass will be in smaller diameter trees. Therefore, all trees to smaller diameters should be measured, e.g., to at least 5 cm minimum diameter.
· Stand tables should be reported with diameter classes of even width and no larger than 10 cm wide. All field measurements need to be archived so that follow-up studies or questions could be addressed.
· Stand tables must not group large diameter trees into one diameter class. Because of the importance of large trees to biomass density, serious errors could be introduced when assumptions about the number of large trees/ha are made.
· Estimation of volume should be standard for all inventories. However, standards are likely to vary among closed forests, open woodlands, and secondary forests, and the definition of measured volume must be clearly stated in inventory reports.
· For open forests/woodlands, young secondary forests, plantations, and individual trees plantings, it is preferable to use stand tables using small diameter classes or individual tree measurements.
· All inventory plots need to be located exactly on a map; this would allow for follow-up studies to measure biomass change for example.
· Information about inventoried areas and plots should be recorded, such as average height of stand, climatic zone, presence of natural disturbances (e.g., wildfires, tropical storms), average soil type (e.g., standard soil classification system and soil texture), elevation, and general topography (lowland, hilly, mountainous, steep slopes, etc.).
· Information on the degree of human disturbance, either past or present, in the inventory area is needed because it can affect aboveground biomass estimates and aid in explaining the estimates. Evidence of human disturbance includes presence of tree stumps, harvesting activity, logging roads or trails (old or new), other roads or tracks, cleared or young secondary forest patches (e.g., resulting from slash-and-burn agriculture), charcoal pits, and identification of tree species favored by humans.
· Other information such as distance to human settlements and fragmentation of forest area are also useful for assessing likely human impacts. A useful index for assessing fragmentation of forest is the ratio of the length of the perimeter of the forest to the area of the forest (perimeter-area ratio). This can be measured from aerial photos (often used in forest inventories) or remote sensing imagery if available.
The aboveground oven-dry-weight of trees can be measured directly by felling them, oven-drying all components and then weighing them. However, it is not realistic to do this for all inventories. Instead, a practical solution is to develop regression equations based on data from felled trees where this is possible. Such functions should use some easily measurable dimension such as diameter (and sometimes height) as presented in Section 3.2.1. As discussed above, the equations presented in this primer are based on a relatively limited data base, especially for dry forest and conifer forest formations, and improvements in biomass estimation can be made with additional tree data.
During many forest inventories, a number of trees are felled to generate local volume equations. These same felled trees can be used for developing local biomass equations using the methods given below. In other situations, trees need to be selected for felling. The selection of these trees in multi-species forests poses a challenging sampling design, and it is recommended that the assistance of a forest biometrician be sought. However, as a guide, the selected trees must come from the population of interest, represent the major species in the forest, and represent all size classes. It is particularly important that trees in the larger diameter classes be well represented even though estimating their biomass is very time consuming because of their large size. If resources are limited, it is recommended that a couple of trees representing small, medium, and large diameters be selected and their biomass measured as described below. These could then be compared to the estimates derived from the appropriate regression equation, and if they are within acceptable limits no further sampling is needed. If, however, they are not within acceptable limits further sampling and biomass measurements would be needed to develop local biomass equations.
1 As for volume equations, dbh, or diameter above the buttress, must be measured. In the case of multi-stemmed trees, common in dry or open forests and woodlands, the diameter at 0.3 m above the ground is often used instead of dbh, and height measurements are also recommended (for further details see Stewart et al. 1992).
2 After the trees are felled as close to the ground as possible, they should be divided into their components, including main stem, branches of different size classes, leaves and twigs, and fruits.
3 Small branches (<10 cm basal diameter), leaves and twigs, and fruits should be weighed fresh in the field. Several sub-samples (at least five) of each component must be collected and their fresh weight determined. Then they must be oven dried to constant weight at 105°C. Weighing of dried sub-samples should be done as soon as possible after removing them from the oven because they soon absorb moisture and gain weight. For each sub-sample, a ratio of oven-dry-to-fresh weight can be calculated and an average ratio calculated. Multiplication of the total fresh weight of each component by the corresponding oven-dry-to-fresh-weight ratio will result in an estimate of the dry weight of the component.
4 For the larger branches and main stem (>10 cm diameter), it is generally not practical to weigh these fresh in the field. Instead, they should be cut into sections and the volume of each section calculated. The oven-dry-weight of these sections is determined as the product of volume and density (oven-dry-weight per unit of green volume). To estimate density, a disk of wood from each section should be removed. The volume of the disk can be calculated as the cross-sectional area of the disk times the thickness (measured at four points, 90° to each other) or by the water-displacement method. The water-displacement method is based on the principle that an immersed object displaces its own volume of water. The disk is carefully immersed in a container of water (pushing it down with a sharp pointed object), and the increase in water level when the disk is fully immersed is used to measure the increase in volume. To improve accuracy the container of water can be made absolutely full and as the disk is immersed all the water displaced by the disk is collected in a previously weighed container. The weight of the water displaced is weighed in grams and equals the volume of the disk in cm3, because 1 g of water has a volume of 1 cm3. Thus if the weight of the water displaced is 20 g, then the volume of the disk is 20 cm3. After volume measurement the disk is oven dried to constant weight at 105°C; this weight divided by its volume gives density. The weight of the stem and branches is then calculated, making sure all the measurements are in the same units (volume in cm3 and density in g/cm3).
5 The sum of weights of all the components results in the total oven-dried weight of the tree, generally expressed in kg.
Once the biomass of all selected trees has been determined, a regression equation similar to those given in the table in Section 3.2.1 can be developed using an available statistical package. Once again, it is recommended that the assistance of a forest biometrician or statistician be sought at this stage. For trees from the dry zone it is recommended that the oven-dry weight of the tree be correlated to the height (H) as well as diameter (D). Such an equation could take the form:
mass of tree = a + b * HD2, where a and b are regression coefficients
To develop local regression equations for palm trees, height is a better measure of biomass than diameter (see Section 3.2.1). At least three individuals from several height classes should be felled. For each palm tree, the height of the main stem should first be measured. Then the stem should be cut into three to four approximately equal length sections, depending upon the height, and their volume estimated. A disk from each section should be removed and its density (oven dry weight per green volume) measured as described above in step 4. The weight of the stem is then the sum of the product of the volume of each section and its density. The weight of the leaves then has to be determined. This can be determined by first counting the total number of leaves. Then about three to four leaves should be selected and the average of their oven-dry weight (at 105°C) determined. The total oven-dry weight of all leaves is calculated as the number of leaves multiplied by the dry weight per leaf. The oven-dry weight of the inflorescence (flower or fruit) should also be measured and added to obtain the total weight. A simple linear regression equation of total oven-dry biomass of the palm versus its height should be developed as given above in Section 3.2.1.