Fig 1.
Contemporary vegetation distribution in China.
(A) Observed (white area are croplands, urban areas and planted forests); (B) modeled; (C) the proportions of each vegetation type to the terrestrial area of China (observed: light gray line; modeled: dark gray line) and their range fillings calculated as the ratios between the observed and modeled distribution ranges (red line). Due to strong human disturbance, veg 5, 6, and 8 have experienced significant deforestation at the present, and therefore the proportions of observed vegetation types do not sum up to 100%.
Table 1.
The AUC values of the Generalized Linear Models (GLMs) and MaxEnt for each vegetation type.
Fig 2.
Hindcasted past distribution of natural vegetation during (A) the Last Glacial Maximum (LGM) and (B) the Mid-Holocene (MH) and (C) the proportion (%) of each vegetation type (LGM: dark blue line; MH: orange line; modeled: dark gray line).
In (C), the proportion (%) of a vegetation type was calculated as the ratio of its distribution range relative to the terrestrial area of China.
Fig 3.
The regions with vegetation changes (dark green, A, B, C) and the dispersal directions of each vegetation type (D, E, F) from (A, D)the Last Glacial Maximum (LGM) to the Mid-Holocene (MH), (B, E) from the MH to the present and (C, F) from the LGM to the present.
Fig 4.
Percentage of persistence and changing vegetation distribution (A) from the Last Glacial Maximum (LGM) to the Mid-Holocene (MH), (B) from the MH to the present, and (C) from the LGM to the present.
Dark blue: percentage of a vegetation type that remained unchanged between two time periods. Orange: percentage of a vegetation type that was converted to the most abundant new vegetation type (indicated with the type number in the orange bar). Grey: percentage of a vegetation type that was converted to all other vegetation types. For example, 40% of the LGM distribution range of veg 1 (taiga forest) remained unchanged at the present, 49% was converted to veg 17 (grass and forb meadow) in the MH and 11% were converted to other vegetation types (C).
Table 2.
Kappa values between model-based and pollen-based vegetation reconstructions in the LGM and the MH for seven major vegetation types in both LGM and MH.
Fig 5.
Comparison between model-based and pollen-based vegetation reconstructions during the Last Glacial Maximum (LGM).
Panels (A) to (E) map the hindcasted climate suitability of the shown vegetation types, while panels (F) to (J) map the hindcasted binary distribution of each vegetation type. The blue dots in the maps represent the pollen sites for which the same vegetation type was identified during the LGM.
Fig 6.
Comparison between model-based and pollen-based vegetation reconstructions during the Mid-Holocene (MH).
Panels (A) to (G) map the hindcasted climatic suitability of the shown vegetation types, while panels (H) to (N) map the hindcasted binary distribution of each vegetation type. The blue dots in the maps represent the pollen sites for which the same vegetation type was identified for the MH.