Akademik Veri Yönetim Sistemi
Forestist, cilt.72, sa.2, ss.183-191, 2022 (ESCI, Scopus, TRDizin)
Unlike the traditional diversity indices, taxonomic diversity indices take into account the distances based on morphological or functional properties of species or the distances among the species based on the topology of Linnaean classifications. Various taxonomic diversity measures have been developed, among which the most frequently used are Warwick and Clarke’s taxonomic diversity (∆), taxonomic distinctness (∆*), and average taxonomic distinctness (∆+) indices. These well-known indices are a natural extension of Simpson diversity and use abundance or presence data of the species. More recently, using Deng entropy from the Dempster–Shafer evidence theory, a new taxonomic measure was proposed by Özkan (2018). By employing this new measure, weighted taxonomic diversity (TO), unweighted taxonomic diversity (uTO), weighted taxonomic distinctness (TO+ ), and unweighted taxonomic distinctness (uTO+ ) can be estimated. This study was addressed to compare the estimators (TO, uTO, TO+ , and uTO+ ) with the most popular traditional diversity measures (Species richness (S), Shannon entropy (H), and Simpson’s diversity index (1 − λ)) and taxonomic diversity measures (taxonomic diversity (∆), taxonomic distinctness (∆+), and average taxonomic distinctness (∆+). For comparisons, correlation analysis and principal component analysis (PCA) were applied using the plant data taken from 107 sample plots from Yazili Canyon Nature Park in Turkey. The results indicated that TO, uTO, TO+ , and uTO+ have significantly positive correlation with ∆, ∆*, ∆+, S, H, and 1 – λ (p < .01), and all these estimators that are located between taxonomic diversity/distinctness measures (∆, ∆*, and ∆+) and traditional diversity measures (S, H, and 1-λ) along the ordination diagram of the applied PCA. Thus, Deng entropy-based taxonomic diversity measures seem to represent not only traditional diversity indices but also taxonomic diversity measures. Reinforced estimators (maximum values of TO, uTO, TO+ , and uTO+ ) were also improved and used in the present study. Acceptable results were provided from computer-generated simulations by using a hypothetical community dataset. All the computations and simulations were performed using a spreadsheet program created by the authors of this paper. The program, TD_ÖMD, can be downloaded from http://www.kantitatifekoloji.net/takdivozkan.