Estimating marine species richness: an evaluation of six extrapolative techniques

摘要: The number of species in an assemblage at a given point time is fundamental prop- erty ecological systems, yet it hard to quantify for many marine systems. We studied the perfor- mance 6 techniques ('estimators') extrapolating richness from limited numbers sam- ples, using 3 datasets which absolute value total could be determined. propose that ideal estimator should always slightly overestimate compared any observed maximum derived sampling, as sampling error will lead underestimation true richness. quantified performance estimators relative pled across range efforts up 80% required achieve asymptote accumulation. used measures: bias (mean deviation estimate known richness), precision (variance repeated estimates based upon subset avail- able pool samples), and overall accuracy (a combination precision). No single estima- tor performed best all cases, was affected by effort. esti- mator Chao1 intermediate efforts, with LAG S∞ also performing well high consistently underestimated, whilst Chao2 ICE both overestimated displayed poor accuracy, especially uneven patterns incidence. Species abundance incidence amongst samples dataset were shown affect most estimators, exception recently proposed family techniques. conclude represents compromise choice estimator, such nonparametric may represent useful tools rapid estimation some assemblages, on