THE DESIGN AND ANALYSIS OF METHODS FOR SAMPLING MICROCLIMATIC FACTORS
成果类型：
Article
署名作者：
WILM, HG
刊物名称：
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
ISSN/ISSBN：
0162-1459
DOI：
10.2307/2280489
发表日期：
1946
页码：
221-232
关键词：
摘要：
Microclimatic readings vary locally and in time so that excessively laborious and expensive sampling may be required to obtain adequate average values for any temporal or spatial units. The amt. of sampling for long time averages may be reduced by using few instruments which, after each sampling, are moved among k randomized positions. This technique, in conjunction with unbiased fixed stations which will completely measure temporal variations at individual points, may be used to obtain average values in [eta] short time intervals, providing data from the fixed and moving stations are closely correlated. By this concept, any factor has a true magnitude, [mu], for a given area over a long period of time, and variations within the area result from a component, [eta], due to the short time unit of sampling while variations, within the long time interval are attributed to a component, [epsilon], associated with the location of the observation. The estimated variance of the observations likewise consists of a positional and a temporal component which may be separated by analysis of variance. A numerical example dealing with rainfall to a forest illustrates the efficiency of the proposed sampling technique.