Premium
Predicting vascular plant richness patterns in C atalonia ( NE S pain) using species distribution models
Author(s) -
Pérez Nora,
Font Xavier
Publication year - 2012
Publication title -
applied vegetation science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.096
H-Index - 64
eISSN - 1654-109X
pISSN - 1402-2001
DOI - 10.1111/j.1654-109x.2011.01177.x
Subject(s) - species richness , plant species , quadrat , statistics , range (aeronautics) , species distribution , generalized linear model , multivariate statistics , vascular plant , ecology , mathematics , geography , biology , habitat , materials science , shrub , composite material
Question Given the current state of knowledge on plant species richness in C atalonia, how can we improve definition of sampled species richness patterns? We propose a concrete methodology that will highlight the most speciose areas and those areas that are insufficiently sampled. Location C atalonia, covering 31 980 km 2 in northeast S pain. Methods This study provides a quantitative assessment of plant richness using sampling units of 10 km × 10 km ( n = 319). Generalized linear models ( GLM ), multivariate adaptive regression splines ( MARS ) and a maximum entropy model ( M axent) were used for all plant species contained in the C atalonia D atabase of B iodiversity ( B anc de D ades de B iodiversitat de C atalunya, BDBC ). The projected presence/absence maps were combined to create species richness maps based on distribution models for 2738 species. Results The modelling techniques were highly correlated, and we did not distinguish any differences in projections of geographic patterns of species richness among the modelling algorithms. However, MARS and M axent achieved the best prediction success, whereas GLM tended to over‐predict species number per quadrat. The MARS map gave the highest predictive performance based on both K appa and the true skill statistic, and for the two components of disagreement (quantity and allocation disagreement). We were able to identify one previously known region of high diversity, the P yrenees, and two additional areas, the coastal range and the pre‐coastal mountain range. Conclusions We obtain improved distribution patterns for plant species in C atalonia over the previously sampled patterns, and, most importantly, we provide an estimate of the number of species present in those areas where sampling data are incomplete, with an expected 300 species. The model‐predicted richness maps presented here can be used to detect zones with low and high species richness and to develop strategies for either restoring or protecting landscape biodiversity as part of national conservation plans.