The application of two-phased sampling, in conjunction with digital aerial-photographic interpretation and mensuration, for the purpose of inventorying forest stand conditions is an important focal point of research at the Institute of Forest Management and Yield Sciences of the University of Göttingen. The excellent results of these studies, which have been tested as to their economy, are very promising and the methodology can meanwhile be viewed as "ready to put to practice". It has been shown that the method of employing aerial photographs in large-scale inventory (e.g. forest enterprise inventories) of volume and increment estimation is, by far, superior to the time consuming and costly procedure of a purely terrestrial sampling.
In addition to the studies done up to now, which were essentially executed by means of aerial photographs to the scale of 1:50001:12,000, it was tested if, and to what measure, small-scale aerial photographs from high-altitude flights to the scale of 1:35,000 could be used in determining the dendrometric aerial-photographic data and to what extent the information gained could be employed in volume and increment estimation in two-phased sampling.
Two-phased sampling is executed in two parts: the first inventory phase, in which numerous, easily recorded auxiliary variables are collected; the second inventory phase, in which a subsample of relatively few sampling units is taken from the first phase and from which the variables of interest are determined. The dependence of variables of interest and auxiliary variables in the second phase is crucial in making assertions as to the variables of interest for the entirety. The variables of interest can be combined by means of the ratio estimator or regression estimator. In this study a regression model is to be developed which will combine both aerial-photographic evaluation and data obtained through terrestrial inventory.
The sampling method was executed with the aid of colored IR-aerial photographs shot from a high-altitude flight flown over the Lower Saxony State Forest Enterprises of Grünenplan to the scale of 1:35,000, using the fir tree species in example. The results of this inventory of the researched area were compared to those of a two-phased sampling, in which aerial photographs to the scale of 1:6000 had been employed.
Correspondence: A. Akca, Th. Beisch Institut für Forsteinrichtung und Ertragskunde, Göttingen, Germany