Fig 1.
Location and elevation of the study region showing the names of the main mountain ranges.
The red dots indicate the locations of climate observation stations used by WorldClim (WC).
Fig 2.
Difference between the remotely-sensed (RS) and WorldClim (WC) datasets for 19 bioclimatic variables generated from temperature (units of °C * 10) and precipitation (units of mm).
Positive differences indicate higher values for the RS dataset, and negative differences indicate higher values for the WC dataset. The outlines on each panel are the boundaries of the different mountain ranges and are provided for reference. The bioclimatic variables are long-term averages of annual mean temperature (bio1); mean diurnal range (bio2); isothermality (bio3); temperature seasonality (bio4); maximum temperature of the warmest month (bio5); minimum temperature of the coldest month (bio6); annual temperature range (bio7); mean temperature of the wettest (bio8), driest (bio9), warmest (bio10), and coldest (bio11) quarter; annual precipitation (bio12); precipitation of the wettest (bio13) and driest (bio14) month; precipitation seasonality (bio15); and precipitation of the wettest (bio16), driest (bio17), warmest (bio18) and coldest (bio19) quarter.
Fig 3.
Pearson pairwise correlation matrix between the bioclimatic variables for the WorldClim (WC) dataset (upper-right matrix) and for the remotely-sensed (RS) dataset (lower-left matrix) shown in colored circles, and the correlation between the RS and WC datasets for the same bioclimatic variables (diagonal values from upper-left to lower-right).
The numbers from 1 to 19 stand for the 19 bioclimatic variables, while “E” stands for elevation. Blue colors indicate positive correlations and red colors indicate negative correlations as indicated in the figure legend. Both the size and the color of the circles represent the magnitude of the correlation. The bioclimatic variables are long-term averages of annual mean temperature (bio1); mean diurnal range (bio2); isothermality (bio3); temperature seasonality (bio4); maximum temperature of the warmest month (bio5); minimum temperature of the coldest month (bio6); annual temperature range (bio7); mean temperature of the wettest (bio8), driest (bio9), warmest (bio10), and coldest (bio11) quarter; annual precipitation (bio12); precipitation of the wettest (bio13) and driest (bio14) month; precipitation seasonality (bio15); and precipitation of the wettest (bio16), driest (bio17), warmest (bio18) and coldest (bio19) quarter.
Table 1.
The varimax-rotated principal component loadings for the WorldClim (WC) and remotely-sensed (RS) datasets.
“RC” refers to rotated principal components with eigenvalues > 1. Four principal components were extracted for the WC dataset and five for the RS dataset. The variables shown in bold were used in the calibration of the species distribution models.
Fig 4.
Contribution of each of the top three bioclimatic variables for the three model calibrations.
“WC4” indicates the model calibration using the WorldClim baseline climate information and four bioclimatic variables; “RS4” refers to the model calibration using the remotely-sensed baseline climate information and four bioclimatic variables; and “RS5” refers to the model calibration using the remotely-sensed baseline climate information and five bioclimatic variables. The number on the top of each bar refers to the bioclimatic variables: bio1, annual mean temperature; bio4, temperature seasonality; bio7, temperature annual range; bio12, annual precipitation; and bio15, precipitation seasonality. The height of a bar is the percent contribution of a particular variable.
Fig 5.
Biplots for the 21 bamboo species of the change in the likelihood of occurrence obtained from the three model calibrations expressed as the difference between the future (2061–2080) and baseline climate conditions under RCP 8.5 for 17 downscaled GCMs.
“WC4” indicates the model calibration using the WorldClim baseline climate information and four bioclimatic variables; “RS4” refers to the model calibration using the remotely-sensed baseline climate information and four bioclimatic variables; and “RS5” refers to the model calibration using the remotely-sensed baseline climate information and five bioclimatic variables.
Fig 6.
Projected differences in the relative likelihood of occurrence between future (2061–2080) and baseline climate conditions for Fargesia denudata, as estimated by model simulations calibrated from the WorldClim (top) and remotely-sensed (bottom) datasets using four bioclimatic variables as predictors (abbreviated as WC4 and RS4).
The results shown here used the “clamping” option in MaxEnt where variables outside the training range are treated as though they are at the limit of the training range. For each calibration, the individual panels represent the outcomes obtained from downscaled climate projections for 17 global climate models (GCMs), and the abbreviated model names are provided to facilitate comparison between the WC4 and RS4 calibrations. The 17 GCMs are: ACCESS1-0 (ac), BCC-CSM1-1 (bc), CCSM4 (cc), CNRM-CM5 (cn), GFDL-CM3 (gf), GISS-E2-R (gs), HadGEM2-AO (hd), HadGEM2-CC (hg), HadGEM2-ES (he), INMCM4 (in), IPSL-CM5A-LR (ip), MIROC-ESM-CHEM (mi), MIROC-ESM (mr), MIROC5 (mc), MPI-ESM-LR (mp), MRI-CGCM3 (mg), and NorESM1-M (no).
Fig 7.
Projected change in the climatically-suitable area for the 21 bamboo species by the three model calibrations developed from the WorldClim (WC) and remotely-sensed (RS) datasets using either four for five bioclimatic variables as predictors (WC4, RS4, and RS5).
The projected change is expressed as the ratio of the difference in climatically-suitable area between the future (2061–2080) and baseline climate conditions to the climatically-suitable area for the baseline conditions (the values can be multiplied by 100 to obtain a percentage change). Each box and whisker plot includes projections obtained from 17 global climate models (GCMs) and 11 conversion thresholds.
Fig 8.
Proportion of sum-squared error to total squared error for a three-way analysis of variance (ANOVA) of the projected change in suitable area for each of the 21 bamboo species (left), and summarized over all 21 bamboo species (right).
The analysis includes two model calibrations developed from the WorldClim (WC) and remotely-sensed (RS) datasets and using four bioclimatic variables as predictors (referred to as WC4 and RS4), 17 future climate projections downscaled from global climate models (GCMs), and 11 conversion thresholds.