Sample-based estimation of tree species richness in a wet tropical forest compartment

Abstract

Petersen's capture-recapture ratio estimator and the well-known bootstrap estimator are compared across a range of simulated low-intensity simple random sampling with fixed-area plots of 100 m? in a rich wet tropical forest compartment with 93 tree species in the Western Ghats of India. Petersen's ratio estimator was uniformly superior to the bootstrap estimator in terms of average error (bias) and mean absolute error. The observed richness always had the largest negative bias. A large negative bias of 25 percent persisted even when approximately 10 percent of the area was sampled. Estimated confidence intervals had poor coverage rates. A proposed variance estimator for the observed richness performed well.

Parent Publication

• Proceedings of the seventh annual forest inventory and analysis symposium

Citation

Magnussen, Steen; Pelissier, Raphael. 2007. Sample-based estimation of tree species richness in a wet tropical forest compartment. In: McRoberts, Ronald E.; Reams, Gregory A.; Van Deusen, Paul C.; McWilliams, William H., eds. Proceedings of the seventh annual forest inventory and analysis symposium; October 3-6, 2005; Portland, ME. Gen. Tech. Rep. WO-77. Washington, DC: U.S. Department of Agriculture, Forest Service: 269-272.