Experiments in estimating density from joint point- and nearest-neighbour distance samples
- Forest and Range Experiment Station, Rangiora
A new model has been developed for estimating density of plant or animal populations. For uniform, randomly-distributed or contagious populations which are distributed with one or two orders of aggregation, it employs the distance from each sample point to the nearest population member, from that individual to the nearest neighbour and from that neighbour to its nearest neighbour. The principle of the model is that an estimate of density is first obtained from the sample point-nearest population member distances. This point-distance estimate is then corrected for bias arising from non-randomness by using the Gums and frequencies of first and second neighbour distances. Procedures are given for extending the model to populations distributed with three or more orders of contagion.
The model was tested on five populations of dots, of which four were contagious, two pine plantations of uniform distribution and a population of the contagiously-distributed Composite, Celmisia spectabilis. Estimates were all within 25% of the true values. The model is considered to be suitable for estimating a wide variety of populations.