Comparing methods, species distribution data and spatial resolutions:
Peninsula effect study on amphibian of Malay Peninsula
Abstract
In macroecological studies, accurate species distribution information
and reliable methodologies are essential to illustrate the accurate
geographical distribution of organisms. The available distribution data
sources are generally limited to a particular geographical region and
studied taxa and often contain some degree of sampling bias that caused
a misrepresentation of species distribution. This study investigates how
the different distribution data, methods, and spatial resolution affect
the amphibian distribution pattern in the Malay Peninsula. Five types of
distribution data, namely point-based (simple point occurrences and
rarefied point occurrences) and range-based (expert-drawn range map,
modelled range map and hybrid range map), were assembled from the
published checklist, GBIF, natural history museum collections
information, and digital range maps. Species richness was point-to-grid
mapped by overlaying each distribution data with spatial resolutions of
100 km, 75 km, 50 km, 25 km and 15 km square grid and were evaluated
with descriptive and quantitative-based approaches. Overall results show
that richness maps created from range-based maps well-performed both
descriptive and quantitative. Spatial autocorrelation was present in all
distribution data; however, it is significantly reduced in 50 km
resolution in range-based maps. When mapped with different spatial
resolutions, the hybrid range map showed less variability in spatial
relatedness and structure and executed a better predictive performance.
The findings suggested that integrating the information from other
distribution sources reduces the erroneous information in particular
distribution data and can also evaluate the effects of spatial
resolutions before developing species distribution maps.