TY - JOUR
T1 - Measuring species richness on sandy beach transects: extrapolative estimators and their implications for sampling effort
AU - SCHOEMAN, DS
AU - Nel, Ronel
AU - Soares, Alexandre Goulart
PY - 2008/6/3
Y1 - 2008/6/3
N2 - Species richness is a measure that is fundamental to many studies in ecology, and it is particularly important on sandy beaches, where it underlies patterns described by the broadly accepted swash exclusion hypothesis. However, its estimation in practice is problematic. This has led ecologists in other fields to adopt extrapolative estimators of species richness, which project the total number of species present in a habitat by adjusting upward the number of species observed by an amount related to the number of rare species encountered in the samples. In so doing, the species richness can be estimated, with confidence intervals, at any level of sampling effort. Despite the availability and advantages of these methods, beach ecologists have continued to use the observed species richness as a point estimate of biodiversity for beaches. Here, we employ a Monte Carlo resampling approach over a range of routine transect designs used to sample sandy beaches, and evaluate the performance of seven non-parametric extrapolative estimators for species richness relative to that of the more conventionally used observed species richness. We find that the first-order Jackknife estimator (Jack 1) is the least biased, most accurate and most consistent across sites. Employing this estimator would allow accurate estimation of species richness on short (tens of metres) stretches of beach without exceeding the acceptable levels of sampling effort (4-5 m(2)). Spreading this effort evenly over three across-shore transects, each with a minimum of 13 equally spaced levels seems appropriately efficient. Although a greater number of research studies is required to ascertain the generality of these results beyond the beaches we sampled, we tentatively recommend the application of our results in biodiversity surveys on sandy beaches.
AB - Species richness is a measure that is fundamental to many studies in ecology, and it is particularly important on sandy beaches, where it underlies patterns described by the broadly accepted swash exclusion hypothesis. However, its estimation in practice is problematic. This has led ecologists in other fields to adopt extrapolative estimators of species richness, which project the total number of species present in a habitat by adjusting upward the number of species observed by an amount related to the number of rare species encountered in the samples. In so doing, the species richness can be estimated, with confidence intervals, at any level of sampling effort. Despite the availability and advantages of these methods, beach ecologists have continued to use the observed species richness as a point estimate of biodiversity for beaches. Here, we employ a Monte Carlo resampling approach over a range of routine transect designs used to sample sandy beaches, and evaluate the performance of seven non-parametric extrapolative estimators for species richness relative to that of the more conventionally used observed species richness. We find that the first-order Jackknife estimator (Jack 1) is the least biased, most accurate and most consistent across sites. Employing this estimator would allow accurate estimation of species richness on short (tens of metres) stretches of beach without exceeding the acceptable levels of sampling effort (4-5 m(2)). Spreading this effort evenly over three across-shore transects, each with a minimum of 13 equally spaced levels seems appropriately efficient. Although a greater number of research studies is required to ascertain the generality of these results beyond the beaches we sampled, we tentatively recommend the application of our results in biodiversity surveys on sandy beaches.
U2 - 10.1111/j.1439-0485.2008.00223.x
DO - 10.1111/j.1439-0485.2008.00223.x
M3 - Article
SN - 1439-0485
VL - 29
SP - 134
EP - 149
JO - Marine Ecology
JF - Marine Ecology
IS - Suppl.
ER -