| Title |
Analysis on the Species-Area Relationship of Benthic Macroinvertebrates Using Probability Distribution Models |
| Authors |
공동수(Dongsoo Kong) ; 권용주(Yongju Kwon) ; 김예지(Ye Ji Kim) ; 김명철(Myoung Chul Kim) ; 전영철(Yung-Chul Jun) ; 권순직(Soon Jik Kwon) ; 전용락(Yong Lak Jeon) |
| DOI |
https://doi.org/10.15681/KSWE.2025.41.3.199 |
| Keywords |
Benthic macroinvertebrates; Differential entropy; Microhabitat heterogeneity; Probability distribution models; Sample size; Species-area relationship |
| Abstract |
The species-area relationship in benthic macroinvertebrate assemblages from a clean stream in Korea was examined using 20 mathematical models, including 4 types of power and logarithmic functions and 16 probability distribution models. Probability models included two-parameter models (exponential, inverse exponential, monad, sine), three-parameter left-truncated models (normal, logistic, Gumbel), three-parameter models (generalized exponential, Weibull, inverse Weibull, lognormal, logistic power, gamma, sine power), and four-parameter models (generalized logistic power, beta). Among non-probability model groups, the logarithmic function was more suitable than the power function. It was also more appropriate to apply a shifted model with a threshold value than a basic function. While the power function has often been used in species?area analyses of open ecosystems such as terrestrial or marine environments, the logarithmic function appeared to be more appropriate for analyzing less mobile or sedentary benthic macroinvertebrates in a relatively closed habitat of a stream. The best-fitting models were inverse Weibull and generalized logistic distribution models, both of which showed the highest predictive performances. The power function model significantly overestimated species richness, while exponential, inverse exponential, normal, logistic, Gumbel, and sine distribution models underestimated it. Our results suggest that the maximum number of species estimated by inverse Weibull and generalized logistic distribution models might serve as an index of habitat species capacity. Additionally, the half-saturation area (the median of the probability distribution) and differential entropy might be useful indices for assessing microhabitat heterogeneity. |