Modeling species-area relationship with measurement uncertainty
In the present study, the role of measurement uncertainty of species richness has been taken into account when estimating the parameters (c and z) for the power-law species-area relationship (SAR). The nonlinear weighted estimator is used to quantify the influence of measurement uncertainty in the values of species richness, which can be derived from the observed and estimated species richness across different areas. As a comparison, the parameters are also estimated using the conventional nonlinear least-of-square (NLOS) estimator without considering data uncertainty and only the average species richness from estimated and observed values is used. Species richness for epigean arthropods (EAR), canopy arthropods (CAJ) and ground bryophytes (BD) over different areas at the Azores, Portugal are used as empirical data sets for comparing the proposed and conventional NLOS estimators. The results show that, both parameters c and z estimated by estimator are significantly different from those from NLOS respectively through the paired t-test in all the three empirical data sets except that c values are not significantly different for the BD data set when comparing both estimators. Given that fact that there are significant differences on the estimated parameters for the power-law SAR model when comparing both estimators, estimator is recommended for fitting SAR models so as to better capture the stochasticity of species richnes.