At Interpine we offer a wide range of forest inventory products, all with same goal; to provide accurate data to decision makers. Here is a few tips to consider on this pathway to an accurate and precise final dataset.
Before starting the process of inventory planning a key component is to outline the inventory objectives since these assumptions define how the data is to be used, now and in the future. Conducting a forest inventory for a woodlot can have very different objectives compared due diligence on 20,000ha. Narrowing these down early helps avoid problems in the statistical design of the inventory later on.
Figure 1: Objectives
Obtain Prior Information
Reliable and accurate data requires good planning and implementation. A recent NZIF article names one of the major pitfalls of volume inaccuracy as error in the calculation of NSA (Woollons, 2014). Since NSA changes over time (Figure 2), one of the initial steps in forest inventory planning should be forest NSA mapping. NSA is a horizontal area based on an aerial assessment, which must match actual measurements on the ground. Another key component is the removal of gaps hereby reducing variability. Also ensure that field crews replicate these assumptions of NSA when installing plots on the ground and therefore are briefed on what boundary definitions are being used and have an upto date map.
Design a Sampling Strategy and Sampling Population (Stratification)
Another cost efficient/cost beneficial method to decrease variety is stratification. Stratification is the segregation of a sampling frame (forest) into area known homogenous sub-populations (strata). Foresters typically stratify based on species, age, productivity, silviculture etc.
Figure 3: Forest stratified into individual strata’s
Sampling design and plots to establish varies depending on inventory type and objective, typical a sampling error of 5-10 percent difference in predicting volumes is accepted in inventories close to harvest, while 20 percent or more might be a reasonable sampling error for mid-rotation inventories. Mapping of plot locations are usually established by the use of Simple Random Sampling (SRS), Plots is located in a systematic pattern (e.g. grid) rather than at random. Grid start point is often randomly assigned to avoid bias.
Woollons, R. 2014. Woollons, R. 2014. Don’t blame the tree volume equations. New Zealand Jurnal of Forestry, 50(3): 37-39
Pieres Maclaren, J. 2000. How much wood has your woodlot got? Forest Research Bulletine No. 217