Some Sampling Methods in Forest Surveys

In this paper, an attempt has been made to give a brief account of the various sampling techniques in vogue in forest surveys. Different aspects of systematic sampling have been discussed in detail and a systematic sampling with two random starts is recommended as it is expected to give more precise results than a stratified random sampling with two units per stratum. When sampling is done in two or more stages, a partially systematic line-plot survey is recommended because it provides a valid estimate of the sampling error. Kadambi has used circular plots within a strip in Mysore state; Seth has tried and initiated various types of plots within each strip in U.P. Ratio and regression methods of estimation can safely be used for sample sizes greater than thirty. When the forest has been completely enumerated in the past, one can use a double ratio method to estimate the ratio of class I to class II trees utilising the information of the past complete enumeration. If the units of a random sample are selected with probability proportional to the size of the unit and with replacement the mean of the volume per acre of the units included in the sample will provide an unbiased estimate of the population mean volume per acre.