Estimating the volume of forest growing stock using auxiliary information derived from relascope or ocular assessments

The aim of this study is to demonstrate the potential of integrating probabilistic sampling and estimation with the conventional technique referred to as forest inventory by compartments. The objective of this paper is to propose two strategies for the assessment of growing stock volume using two-phase sampling, namely: (i) relascope basal area estimation performed on first-phase sampling points followed by volume estimation performed on a sub-sample of points selected in the second phase; (ii) ocular evaluation of growing stock volume performed on first-phase sampling plots of fixed size followed by volume estimation performed on a sub-sample of plots selected in the second phase. The effectiveness of using the auxiliary information gathered in the first phase is assessed by comparing the double-expansion estimator of total volume which depends solely on the second-phase sample with the two-phase ratio estimator which instead calibrates the double-expansion estimator on the basis of first-phase information. Conservative estimators of sampling variances and confidence intervals are derived for both the estimators. As is usual in forest inventories, first-phase sampling is assumed to be performed on a systematic random grid while three different schemes are considered for drawing the second-phase sample: simple random sampling without replacement, stratified sampling and 3-P sampling. The performance of double-expansion and ratio estimators under the three schemes adopted in the second phase is empirically checked by means of a simulation study performed on a real compartment in a beech forest of Central Italy. Simulation results show that the use of auxiliary information generated in the first phase constitutes a very effective way of increasing the accuracy of volume estimation at the compartment level, with a moderate increase of fieldwork.