3.2 Patterns of non-endemic species richness
Non-endemic species (n = 170) accounted for 71.7% of all species and
were found in 1,659 grids. Their distribution in each grid ranges from
0–58 (mean: 25.65 ± 12.06 SD) species (Figure 1b; Table S5). The
species richness distribution pattern of non-endemic species was similar
to that of all rodent species. The Southern Hengduan Mountains, Qilian
Mountains, Tianshan Mountains, and Altai Mountains had high non-endemic
species richness. Similar to all species, species richness was low in
the Tibetan Plateau and Tarim Basin regions.
The analysis of non-endemic species richness and individual
environmental factors showed that VEG and AET had the greatest impact on
the non-endemic species richness pattern. This result was the same for
all the species (Figure 2c, d; Table S6).
The best model for predicting non-endemic species richness comprised
three predictor sets of five variables (EW1+ HH1+ HH2+ HE1+ HE2). OLS
and SAR explained 70% and 61% of the non-endemic species richness
(Table 1), respectively. As with the results for all species, both
regression models showed that HH1 and HE1 were the most important
predictors. However, OLS showed that, in addition to the two factors
mentioned above, EW1 is equally important. In variance partitioning, all
four predictor sets explained 70.10% of the variance in defining the
richness patterns of non-endemic species. Non-endemic species richness
was significantly correlated with the habitat heterogeneity predictor
set, explaining 51.21% of the variation. This was followed by human
factors (50.42%), the energy-water predictor set (27.67%), and the
climate seasonality predictor set (13.14%) (Figure 3a, c; Table S7).