Discusión de cuatro métodos estadísticos para definir el patrón espacial, en poblaciones ecológicas

Abstract

Due to the importance of knowing precisely the space pattern of a popula- tion, this work was carried out in the ranch Los Angeles, in Buenavista, Saltillo, in the state of Coahuila, Mexico. It consisted in applying the statistical methods of greater importance. Among such methods were the variance/mean rate, plot, ne- gative binominal distribution and block quadrant variance. Species Yucca (Yuc- ca carnerosana) and lecheguilla (Agave lecheguilla) were selected for this work due to its economical importance, the unit sizes were 0.25 m2, 1.00 m2 y 4.00 m2 for lecheguilla and 100 m2 for yucca. This was done according to the proposal of Oosting (1956). As far as sample size is concerned the Pieper technique (1978) was used and in lecheguilla were found 76 units with 0.25 m2, 29 of 1 m2, 29 of 4 m2; and 75 units of 100.00 m2 in yucca. 

In the negative binomial distribution method it was found that the aggrega- tion estimator k3 is far better than k1 since it is a minimal variance estimator, and the sample size of 4.00 m2 resulted the best because of its smaller value of the variation coefficient. On the other hand, the plot method, in this study, is consi- dered as the non-efficient one due to its differences in defining the type of space pattern in relation to the other one. Besides the above mentioned, it was observed that indexes behaved in a homogeneous way in all but one of this methods; the block quadrant variance one. All other methods are considered reliable for describing the space pattern.