Can local landscape attributes explain species richness patterns at macroecological scales?

Aim Although the influence on species richness of landscape attributes representing
landscape composition and spatial configuration has been well documented at
landscape scales, its effects remain little understood at macroecological scales. We
aim to assess the role of landscape attributes, and their relative importance compared
with climate, habitat heterogeneity and human influence (CHH) in particular,
in shaping broad-scale richness patterns.
Location Mainland China.
Methods Species richness data for mammals, birds, reptiles and amphibians were
derived from the China Species Information Service. Together with the richness
data, CHH variables and class- and landscape-level landscape metrics were calculated
using grain sizes of 50 km × 50 km, 100 km × 100 km and 200 km × 200 km.
At these multiple scales, the species richness of each taxonomic group was correlated
with CHH and landscape variables using both ordinary least square (OLS)
and simultaneous autoregressive (SAR) models; variation partitioning was used to
assess the relative strength of landscape attributes versus CHH variables.
Results In general, climate is the most influential factor shaping richness patterns.
Landscape attributes, especially class-level attributes, can also explain considerable
variation in richness.Variation partitioning showed largely overlapped fractions of
explained variation between landscape attributes and CHH variables. The pure
explanatory power of landscape attributes was small for mammals, reptiles and
amphibians, showing R2 of 1–3%, while it was considerably larger for birds,
showing R2 of 5–10%. The environment–richness correlations showed scale
dependency, but the pure explanatory power of landscape attributes appeared to
show small changes across the scale range used in this study.
Main conclusions In addition to CHH variables, landscape attributes can
explain some broad-scale richness patterns, especially for birds. The incorporation
of landscape attributes will be conducive to better understanding the drivers of
richness patterns and modelling species richness at macroecological scales.