Global Ecology and Conservation 10 (2017) 139–146

Contents lists available at ScienceDirect

Global Ecology and Conservation

journal homepage: www.elsevier.com/locate/gecco

Original Research Article

Maxent modeling for predicting impacts of climate change on

the potential distribution of Thuja sutchuenensis Franch., an
extremely endangered conifer from southwestern China
Aili Qin a,c,1 , Bo Liu b,c,1 , Quanshui Guo a, *, Rainer W. Bussmann d ,
Fanqiang Ma a , Zunji Jian e , Gexi Xu e , Shunxiang Pei e
a
Research Institute of Forestry Ecology, Environment and Protection, Chinese Academy of Forestry, Beijing, China
b
College of Life and Environmental Science, Minzu University of China, Beijing, China
c
Institute of Botany, Chinese Academy of Science, Beijing, China
d
William L. Brown Center, Missouri Botanical Garden, St. Louis, USA
e
Experimental center of Forestry in North China, Chinese Academy of Forestry, Beijing, China

article info a b s t r a c t
Article history:
Objectives
Received 17 November 2016
Detailed and reliable information about the spatial distribution of species provides im-
Available online 16 March 2017
portant information for species conservation management, especially in the case of rare
Keywords: species of conservation interest. We aimed to study the consequences of climate change
Thuja sutchuenensis on geographical distributions of the tertiary rare tree species Thuja sutchuenensis Franch.
Maxent (Cupressaceae) to provide reference for conservation management of this species, includ-
Jackknife ing priority area selection for introduction and cultivation of the species. We expect that
Climate change this approach could be promising in predicting the potential distribution of other rare tree
Species distribution modeling
species, and as such can be an effective tool in rare tree species restoration and conservation
planning, especially species with narrow distribution or raw presence-only occurrence
data.
Methods
107 records covering the whole distribution range of T. sutchuenensis in the Daba
Mountains were obtained during a 3-year field survey. The principle of maximum entropy
(Maxent) was used to model the species’ potential distribution area under paleoclimate,
current and future climate background.
Results
The Maxent model was highly accurate with a statistically significant AUC value of
0.998, which is higher than 0.5 of a null model; The location of the potential distribution for
the last interglacial period is in southeastern China, with the largest optimal habitat area
being only 1666 km2 . In other periods, the central location of the potential distribution is
accordant with the real present distribution, but the model’s predicted optimal habitat area
is outside the current distribution.
Conclusions
Our findings can be applied in various ways such as the identification of additional
localities where T. sutchuenensis may already exist, but has not yet been detected; the

* Corresponding author.
E-mail address: guoqs@caf.ac.cn (Q. Guo).
1 These authors contributed equally to this work.

http://dx.doi.org/10.1016/j.gecco.2017.02.004
2351-9894/© 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/
licenses/by-nc-nd/4.0/).
140 A. Qin et al. / Global Ecology and Conservation 10 (2017) 139–146

recognition of localities where it is likely to spread to; the priority selection area for
introduction and cultivation and the conservation management of such rare tree species.
© 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC
BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

Climate change is projected to alter species’ natural distribution and drive biodiversity loss in forest ecosystems
(Millennium Ecosystem Assessment, 2005). To mitigate the effects of climate change on forest ecosystems, we can effectively
target conservation strategies by modeling species distributions to identify areas where sensitive species exist or likely
exist. To effectively model species distributions, detailed and reliable information about the spatial distribution of species
is needed. However, occurrence data tend to be very scarce for the vast majority of species, especially in the cases of rare
species where such information is either non-existent or very poor (Newbold, 2010; Marcer et al., 2013).
Species Distribution Models (SDMs) are commonly used to predict the geographic range of a species given presence-only
occurrence data and environmental variables assumed to influence its distribution (Raxworthy et al., 2003; Anderson and
Martinez-Meyer, 2004; Franklin and Miller, 2009; Thorn et al., 2009; Elith and Leathwick, 2009; Peterson et al., 2011; Wilson
et al., 2011). Of many species distribution model algorithm methods, Maxent (Maximum Entropy, Phillips, 2004) has proved
powerful when modeling rare species with narrow ranges and available scarce presence-only occurrence data (Phillips et al.,
2006; Elith et al., 2006; Pearson et al., 2007; Wisz et al., 2008; Rebelo and Jones, 2010; Elith et al., 2011; Sardà-Palomera et al.,
2012; Garcia et al., 2013; Marcer et al., 2013). A number of studies have been conducted on the distribution of tree species,
but most of them predict potential distribution areas under current climate conditions instead of taking the paleoclimatic
background into account (Ma et al., 2014). This makes it difficult to clearly assess the changes of a species’ distribution area in
the past, present and future climate fluctuations. In addition, few studies focus on tree species with very narrow distribution
areas and less presence-only occurrence data.
Thuja sutchuenensis Franch. (Cupressaceae) is a rare evergreen forest tree, naturally distributed in the Daba Mountains
range in northwestern Chongqing Municipality and eastern Sichuan Province, China. This species was once listed as extinct in
the wild by IUCN-SSC, until it was rediscovered in 1999 (Farjon and Page, 1999; Xiang et al., 2002). Previous studies suggested
that T . sutchuenensis originated in the middle to late Tertiary (Peng and Wang, 2008; Cui et al., 2015). It experienced huge
geological and climatic fluctuations in the late Tertiary and Quaternary. Currently the living individuals of this species are
distributed in subtropical evergreen broadleaf forest, with mountain cinnamon or brown soil on limestone soil type. This
area has an annual average temperature 6◦ -10◦ , an exposure of 1000–1200 h per year, frost-free days 150–200 d, average
Annual Precipitation 1200–1400 mm.
For understanding its possible reaction to climate change it is crucial to model the area it previously occupied; i.e. range
map and size under different climate background, including the middle to late Tertiary, Quaternary, now and future.
In the present study, we used Maxent to model the distribution of this ancient species in China. We aimed to use
species distribution models (SDMs) to identify additional localities where T. sutchuenensis may already exist but remains
undocumented or where it could potentially exist, ultimately providing targeted conservation strategies for this critically
endangered species.

2. Materials and methods

2.1. Study area

The rare tertiary-relic tree, T. sutchuenensis Franch, is distributed between 108.5◦ E–109.25◦ E and 31.5◦ N–31.83◦ N in the
Daba Mountains of southwestern China.
Our study area include major mountain chains of China, east to Changbai Mountain, west to the Tianshan, south to the
Hengduan Mountains, and north to Greater Khingan range.

2.2. Species occurrence data collection

The occurrence locations of T. sutchuenensis in the Daba Mountains were collected during a 3-year field survey across
southwestern China. We recorded and geo-referenced all natural populations and isolated individuals of T. sutchuenensis,
totaling 107 records in the Daba Mountains, resulting in a detailed distribution map (Fig. 1).

2.3. Environmental variables

Temperature, rainfall, geographical barriers and other ecological factors, such as underlying geological formations,
influence species distributions (Kaeslin et al., 2012). To determine which environmental variables most influence the
distribution of T. sutchuenensis, we included in our model 19 bioclimatic variables (Hijmans et al., 2005) and one biophysical
 A. Qin et al. / Global Ecology and Conservation 10 (2017) 139–146 141

Fig. 1. Present distribution of Thuja sutchuenensis in southwestern China based on 3-year field survey data.

Table 1
Environmental variables used in the study and their percentage contribution.
Code Environmental variables Unit % contribution
◦
Bio1 Annual Mean Temperature C
◦
Bio2 Mean Diurnal Range (Mean of monthly (max temp–min temp)) C 21.4
Bio3 Isothermality (Bio2/Bio7) (×100) –
Bio4 Temperature Seasonality (standard deviation ×100) C of V
◦
Bio5 Max Temperature of Warmest Month C
◦
Bio6 Min Temperature of Coldest Month C 17.3
◦
Bio7 Temperature Annual Range (Bio5–Bio6) C
◦
Bio8 Mean Temperature of Wettest Quarter C
◦
Bio9 Mean Temperature of Driest Quarter C
◦
Bio10 Mean Temperature of Warmest Quarter C
◦
Bio11 Mean Temperature of Coldest Quarter C
Bio12 Annual Precipitation mm 16.3
Bio13 Precipitation of Wettest Month mm
Bio14 Precipitation of Driest Month mm 19.1
Bio15 Precipitation Seasonality (Coefficient of Variation) 2.7
Bio16 Precipitation of Wettest Quarter mm
Bio17 Precipitation of Driest Quarter mm
Bio18 Precipitation of Warmest Quarter mm 23.3
Bio19 Precipitation of Coldest Quarter mm
The highlighted variables, selected through multi-collinearity test, were used in modeling.

variable (elevation) with a 30 s (ca. 1 km) spatial resolution, downloaded from WorldClim dataset (www.worldclim.org). The
current climate data over the period 1950–2000 and Paleoclimate data for the last interglacial (LIG; ∼120,000–140,000 years
BP) and the Last Glacial Maximum (LGM; ∼22,000 years BP) were included data on the Last Glacial Maximum (LGM;
∼22,000 years BP) were derived from CCSM4 (Collins et al., 2006) and MIROC (Hasumi and Emori, 2004), two different
global climate models. The future climate data for two representative concentration pathways (RCPs) for carbon dioxide for
2050 (average of predictions for 2041–2060) and 2070 (average of predictions for 2061–2080) were also included. These
were the most recent GCM climate projections that are used in the Fifth Assessment IPCC report.
To reduce multi-collinearity among the 19 bioclimatic variables and one elevation variable, highly correlated variables
(r ≥ 0.85 Pearson correlation coefficient) were eliminated from further models (Graham, 2003). This reduction of predictor
variables resulted in the inclusion of six variables for models (Table 1). These variables included mean diurnal temperature
range (Bio2), minimum temperature of coldest month (Bio6), annual precipitation (Bio12), precipitation of the driest month
(Bio14), precipitation seasonality (Bio15) and precipitation of warmest quarter (Bio18).
142 A. Qin et al. / Global Ecology and Conservation 10 (2017) 139–146

Fig. 2. The Jackknife test for evaluating the relative importance of environmental variables for Thuja sutchuenensis in southwestern China. (Note: ‘‘Bio12’’
is annual precipitation; ‘‘Bio14’’ is precipitation of driest period; ‘‘Bio15’’ is precipitation seasonality; ‘‘Bio18’’ is precipitation of warmest quarter; ‘‘Bio2’’ is
mean diurnal range; ‘‘Bio6’’ is min temperature of coldest month.)

2.4. Species distribution modeling

We used the maximum entropy model (Maxent version 3.3.3; (Phillips et al., 2006); http://www.cs.princeton.edu/
wschapire/maxent/) in this study because it has been shown to perform better with small sample sizes relative to other
modeling methods (Elith et al., 2006; Pearson et al., 2007; Kumar and Stohlgren, 2009). Maxent (Phillips et al., 2006) uses
presence-only data to predict the distribution of a species based on the theory of maximum entropy. The program attempts
to estimate a probability distribution of species occurrence that is closest to uniform while still subject to environmental
constraints (Elith et al., 2011). In our models, we selected 75% data for model training and 25% for model testing (Phillips,
2008), keeping other values as default. Jackknife analyses were performed to determine variables that reduce the model
reliability when omitted. We used the area under the Receiving Operator Curve (AUC) to evaluate model performance. The
value of AUC ranges from 0 to 1 (Fielding and Bell, 1997). An AUC value of 0.50 indicates that model did not perform better
than random, whereas a value of 1.0 indicates perfect discrimination (Swets, 1988). The model with the highest AUC value
was considered the best performer.
For display and further analysis, we imported the results of the Maxent models predicting the presence ofT. sutchuenensis
(0–1 range) into ArcGIS 9.3. With a reference to the classification proposed by Yang et al. (2013), five classes of potential
habitats were regrouped: unsuitable habitat (0–0.2); barely suitable habitat (0.2–0.4); suitable habitat (0.4–0.6); highly
suitable habitat (0.6–0.7); very highly suitable habitat (0.7–1.0). For each model, we calculated the area of the optimal
distribution, classified as highly or very highly suitable habitat (0.6–1).

3. Results

3.1. Model results

The Maxent model for T. sutchuenensis provided satisfactory results, with an AUC value of 0.998 (±0.001) which is higher
than 0.5 of a random model. Precipitation of warmest quarter (Bio18) contributed most to the model, followed by Mean
diurnal range (Bio2), precipitation of driest month (Bio14), minimum temperature of coldest month (Bio6) and mean annual
precipitation (Bio12) (Fig. 2, Table 1). The cumulative contribution of these five factors is 97.4%.

3.2. Predicted current potential distribution

Current suitable habitats forT. sutchuenensis were predicted in northeastern Chongqing Municipality, eastern Sichuan
province where populations ofT. sutchuenensis, here they are already known to exist, but suitable habitat was also predicted
in southern Shanxi province, where the species is known to be absent (Fig. 3(a)). Notably, the size of the current potential
distribution is significantly larger than the present occurrence ofT. sutchuenensis in the Daba Mountains of southwestern
China. The model predicted 2100 km2 of optimal habitat area, bounded on the north 34.47◦ , the south 29.37◦ , the west
102.92◦ and the east 110.41◦ (Table 2).

3.3. Predicted paleoclimate potential distribution

Paleoclimate (LIG) predictions show a large, continuous distribution area for T. sutchuenensis in southeastern China,
including large parts of Jiangxi and Fujian province, southern Zhejiang and Anhui province and eastern Hubei and Hunan
province. The distribution of T. sutchuenensis during LIG occurs to the southeast of the current distribution (Table 2, Fig. 3(b)).
Moreover, the model showed 16.66 km2 of optimal habitat, 7 times larger than the area of optimal habitat in the current
climate scenario (Table 2).
 A. Qin et al. / Global Ecology and Conservation 10 (2017) 139–146 143

Table 2
Range and area of distribution area of T. sutchuenensis under different climate scenarios.
Period Climate scenario East West South North Area/Km2
◦ ◦ ◦ ◦
Current 110.41 102.92 29.37 34.47 2100
LIG 122.84◦ 108.46◦ 22.28◦ 31.74◦ 1666
LGM MIROC 110.75◦ 106.44◦ 29.29◦ 32.43◦ 3450
CCSM4 105.12◦ 109.83◦ 29.65◦ 33.09◦ 6350
2050 Rcp45 112.76◦ 111.11◦ 28.20◦ 34.47◦ 3620
Rcp85 102.72◦ 110.83◦ 28.50◦ 34.48◦ 3040
2070 Rcp45 102.77◦ 110.57◦ 29.34◦ 34.46◦ 2030
Rcp85 102.63◦ 111.65◦ 28.46◦ 34.46◦ 3590

Fig. 3. Predicted current and historical period habitat for Thuja sutchuenensis. (3a, Current suitable habitats; 3b, Paleoclimate (LIG) predictions; 3c,
Paleoclimate predictions from CCSM4 climatic model for the last glacial maximum (LGM); 3d, Paleoclimate predictions from MIROC climatic model for
the last glacial maximum (LGM)).

Paleoclimate predictions from CCSM4 climatic model for the last glacial maximum (LGM) showed a significant change
in location and a great reduction in size of the predicted distribution compared with the last interglacial period (LIG),
resulting in a predicted location closer to that predicted by the current climate scenario, including northeastern Chongqing
Municipality, eastern Sichuan and southern Shanxi province (Fig. 3(c) and Fig. 3(d)). And also scattered distribution in
western Hubei and Anhui province (Fig. 3(d)). In Fig. 3(d), major optimal habitat (0.6–1) are in Central China, few yellow
ranges (0.4–0.6) fall into the area of Central East China near the Yellow sea, and boarder of Tibet and Sadia, India, probably
because the warm subtropical climate and complex geographic landform could be a shelter for this species. The CCSM4
model resulted in 6350 km2 of optimal habitat, whereas the MIROC model resulted in 3450 km2 of optimal habitat (Table 2).
144 A. Qin et al. / Global Ecology and Conservation 10 (2017) 139–146

Fig. 4. Predicted future habitat for Thuja sutchuenensis. Future predictions are based on two representative concentration pathways for 2050 (a, b) and 2070
(c, d). (Rcp45 and Rcp85).

3.4. Predicted future potential distribution for 2050 and 2070

Although the location of the future potential distribution of T. sutchuenensis is very similar to the current potential
distribution, our model results suggest that the geographic distribution would expand under predicted levels of climate
change. Compared with the area of the most optimal habitat under current climate prediction, the predictions for 2050
(Fig. 4(a), (b)) and 2070 (Fig. 4(c), (d)) using the Rcp45 and Rcp85 climatic models showed almost or slightly more than
the present predictions. For the Rcp45 scenario, the model predicts 3620 km2 of optimal habitat in 2050 and 2030 km2 of
optimal habitat in 2070. For the Rcp85 scenario, the model predicts 3040 km2 of optimal habitat in 2050 and 3590 km2 of
optimal habitat in 2070. Its change in the latitude and longitude see Table 2.

4. Conclusions and discussion

Our models’ output based on Maxent model reached an ‘‘excellent’’ and ‘‘accurate’’ level with higher AUC values (Shao
et al., 2009). Under different climate scenario, the central location of the potential distribution by the model predicted
are accordant with the real distribution except for LIG period which were warmer than other times. On the other hand,
T. sutchuenensis appears to prefer warm and humid environment as shown by the Jackknife results.
The potential distribution size change tendency of T. sutchuenensis is LIG >LGM>2070>2050>current, which is consis-
tent with climate change tendency. At the regional scale, the climate is the main factor to determine the species distribution.
The climate of LIG period is warmer than other period. Many Tertiary relict tree species such as Ginkgo biloba L., Metasequoia
glyptostroboides Hu and W.C. Cheng, Davidia involucrata Baill. and Cathaya argyrophylla Chun et Kuang developed well in this
period, and had a broader distribution area than current (Tang and Ohsawa, 2002; Tang et al., 2011, 2012; Guan and Chen,
1986).
 A. Qin et al. / Global Ecology and Conservation 10 (2017) 139–146 145

Our results suggest that a larger area is climatically suitable for T. sutchuenensis to naturalize and potentially spread in
south of China than its current distribution. However, this species has not aggressively colonized all suitable areas, in other
parts of China, despite a suitable climate? This could be due to other parameters such as soil, interspecific competition,
geographic barriers, human disturbance and other factors (Pearson and Dawson, 2005; Kunstler and Lepart, 2007; Peterson
and Vieglais, 2001; Phillips et al., 2006) but lack of sufficient data was the possible reason for incorporation in this study.
Tang et al. (2015) suggest that natural and human disturbances, as well as little competition from other species, influence
the distribution of T. sutchuenensis, and our models seem to reinforce this notion.
The relic species’ dispersal distances are very limited. After our observation in the actual habitat for many years, this
species only grow on very steep cliffs where there is cooler and dryer, and has more sunshine than in the valley. So the
Chinese name ‘‘Ya bai’’ literally means ‘‘cliff cypress’’. And the populations are very scattered at different mountain cliffs,
different populations are relatively isolated, so each October when the pollens come out, it will be very difficult for the pollen
grains to reach another individual. In the past ten years, we measured the fertilization rate, only less than 25% seeds fertilized,
and even the fertilized seeds are hard to germinate. It has many limitations for the habitat, soil, sunshine. Meanwhile, the
species’ timber has been used by local people for a long history, for Buddha beads, carving, furniture etc. Human distribution
of the population and the environments also severely affected the dispersal of this species.
Species distribution modeling generates valuable information for conservation management of this rare and endangered
Tertiary relict tree species. Our findings can be applied in various ways such as the identification of additional localities
where T. sutchuenensis may already exist, but has not yet been detected; the recognition of localities where it is likely to
spread to; the priority selection area for introduction and cultivation and the conservation management of such rare tree
species.

Acknowledgments

The authors are grateful to Prof. Xiaoquan Wang and Prof. Guangsheng Zhou for their valuable suggestions and comments
on the manuscript. We also thank members of Dabashan and Xuebaoshan National Nature Reserves sincerely for their help
in the field investigation. We are also grateful of colleagues at the Minzu University of China and Missouri Botanical Garden
for their helpful comments and discussions. This work was financially supported by the Special fund for basic scientific
research business of central public research institutes, Grant No. CAFYBB2014QB031 & CAFRIFEEP2015B07, National Natural
Science Foundation of China (31400474, 31400182), and Youth Academic Leadership Program in Minzu University of China
(2017MDYL32).

References

Anderson, R.P., Martinez-Meyer, E., 2004. Modeling species’ distributions for preliminary conservation assessments: an implementation with the spiny
pocket mice (Heteromys) of Ecuador. Biol. Cons. 116, 167–179.
Collins, W.D., Bitz, C.M., Blackmon, M.L., Bonan, G.B., Bretherton, C.S., Carton, J.A., Chang, P., Doney, S.C., Hack, J.J., Henderson, T.B., Kiehl, J.T., Large, W.G.,
McKenna, D.S., Santer, B.D., Smith, R.D., 2006. The community climate system model version 3 (CCSM3). J. Clim. 19, 2122–2143.
Cui, Y.M., Sun, B., Wang, H.F., Ferguson, D.K., Wang, Y.F., Li, C.S., Yang, J., Ma, Q.W., 2015. Exploring the formation of a disjunctive pattern between eastern
Asia and north America based on fossil evidence from Thuja (Cupressaceae). PLoS One 10, e0138544. http://dx.doi.org/10.1371/journal.pone.0138544.
Elith, J., Graham, C.H., Anderson, R.P., Dudik, M., Ferrier, S., Guisan, A., Hijmans, R.J., Huettmann, F., Leathwick, J.R., Lehmann, A., 2006. Novel methods
improve prediction of species’ distributions from occurrence data. Ecography 29, 129–151.
Elith, J., Leathwick, J., 2009. Species distribution models: ecological explanation and prediction across space and time. Ann. Rev. Ecol. Evol. System. 40,
677–697.
Elith, J., Phillips, S.J., Hastie, T., Dudík, M., Chee, Y.E., Yates, C.J., 2011. A statistical explanation of maxent for ecologists. Diversity and Distributions 17, 43–57.
Farjon, A. and Page, C.N. (1999) Conifers: Status Survey and Conservation Action Plan. IUCN-SSC Conifer Specialist Group, Gland, Switzerland and Cambridge,
UK: IUCN.
Fielding, A.H., Bell, J.F., 1997. A review of methods for the measurement of prediction errors in conservation presence/absence models. Environ. Conserv.
24, 38–49.
Franklin, J., Miller, J.A., 2009. Mapping Species Distributions-Inference and Predictions. Cambridge University Press, New York.
Garcia, K., Lasco, R., Ines, A., Lyon, B., Pulhin, F., 2013. Predicting geographic distribution and habitat suitability due to climate change of selected threatened
forest tree species in the Philippines. Appl. Geogr. 44, 12–22.
Graham, M.H., 2003. Confronting multicollinearity in ecological multiple regression. Ecology 84, 2809–2815.
Guan, Z.T., Chen, Y., 1986. A preliminary study on the Cathaya mixed forest in Jinfoshan, Sichuan. Acta Bot. Sin. 28, 646–656 (in Chinese).
Hasumi, H., Emori, S., 2004. K-1 Coupled GCM (MIROC) Description Center for Climate System Research. University of Tokyo, Tokyo.
Hijmans, R.J., Cameron, S.E., Parra, J.L., et al., 2005. Very high resolution interpolated climate surfaces for global land areas. Int. J. Climatol. 25, 1965–1978.
Kaeslin, E., Redmon, I. and Dudley, N., 2012. Wildlife in a changing climate. Rome, Italy: Food and Agriculture Organization Forestry Paper.
Kumar, S., Stohlgren, T.J., 2009. Maxent modeling for predicting suitable habitat for threatened and endangered tree Canacomyrica monticola in New
Caledonia. J. Ecol. Nat. Environ. 1, 94–98.
Kunstler, G., Lepart, J., 2007. Fagus sylvatica L. recruitment across a fragmented Mediterranean landscape, importance of long distance effective dispersal,
abiotic conditions and biotic interactions. Diversity Distrib. 13, 799–807.
Ma, S.M., Nie, Y.B., Geng, Q.L., Wang, R.X., 2014. Impact of climate change on suitable distribution range and spatial pattern in Amygdalus mongolica. Chin.
J. Plant Ecol. 38, 262–269.
Marcer, A., Sáez, L., Molowny-Horas, R., Pons, X., Pino, J., 2013. Using species distribution modelling to disentangle realised versus potential distributions
for rare species conservation. Biol. Cons. 166, 221–230.
Millennium Ecosystem Assessment, 2005. Ecosystems and Human Well-Being: Wetlands and Water Synthesis. World Resources Institute, Washington, DC.
Newbold, T., 2010. Applications and limitations of museum data for conservation and ecology with particular attention to species distribution models.
Progr. Phys. Geogr. 34, 3–22.
146 A. Qin et al. / Global Ecology and Conservation 10 (2017) 139–146

Pearson, R.G., Dawson, T.P., 2005. Long-distance plant dispersal and habitat fragmentation: identifying conservation targets for spatial landscape planning
under climate change. Biol. Cons. 123, 389–401.
Pearson, R.G., Raxworthy, C.J., Nakamura, M., Peterson, A. Townsend, 2007. Predicting species distributions from small numbers of occurrence records: a
test case using cryptic geckos in Madagascar. J. Biogeogr. 34, 102–117.
Peng, D., Wang, X.Q., 2008. Reticulate evolution in Thuja inferred from multiple gene sequences: Implications for the study of biogeographical disjunction
between eastern Asia and North America. Mol. Phylogenet. Evol. 47, 1190–1202.
Peterson, A.T., Soberón, J., Pearson, R.G., Anderson, R.P., Martínezmeyer, E., Nakamura, M., 2011. Ecological Niches and Geographic Distributions. Princeton
University Press, Princeton, NJ.
Peterson, A.T., Vieglais, D.A., 2001. Predicting species invasions using ecological niche modeling: new approaches from bioinformatics attack a pressing
problem. Bioscience 51, 363–371.
Phillips, S.J. (2004) A maximum entropy approach to species distribution modeling. In: Proceedings of the 21st International Conference on Machine
Learning, Banff, Canada.
Phillips, S.J., 2008. Transferability, sample selection bias and background data in presence-only modeling: a response to Peterson et al. and (2007). Ecography
31, 272–278.
Phillips, S.J., Anderson, R.P., Schapire, R.E., 2006. Maximum entropy modeling of species geographic distributions. Ecol. Model. 190, 231–259.
Raxworthy, C.J., Martinez-Meyer, E., Horming, N., et al., 2003. Predicting distributions of known and unknown reptile species in Madagascar. Nature 426,
837–841.
Rebelo, H., Jones, G., 2010. Ground validation of presence-only modelling with rare species: a case study on barbastelles Barbastella barbastellus (Chiroptera:
Vespertilionidae). J. Appl. Ecol. 47, 410–420.
Sardà-Palomera, F., Brotons, L., Villero, D., Sierdsema, H., Newson, S.E., Jiguet, F., 2012. Mapping from heterogeneous biodiversity monitoring data sources.
Biodivers. Conserv. 21, 2927–2948.
Shao, H., Tian, J.Q., Guo, K., Sun, J.X., 2009. Effects of sample size and species traits on performance of BIOCLIM in predicting geographical distribution of
tree species-a case study with 12 deciduous Quercus species indigenous to China. Chin. J. Plant Ecol. 33, 870–877.
Swets, J.A., 1988. Measuring the accuracy of diagnostic systems. Science 240, 1285–1293.
Tang, C.Q., Ohsawa, M., 2002. Tertiary relic deciduous forests on a humid subtropical mountain, Mt. Emei, Sichuan, China. Folia Geobot. 37, 93–106.
Tang, C.Q., Yang, Y., Ohsawa, M., Momohara, A., Hara, M., Cheng, S.L., Fan, S.H., 2011. Population structure of relict Metasequoia glyptostroboides and its
habitat fragmentation and degradation in south-central China. Biol. Cons. 144, 279–289.
Tang, C.Q., Yang, Y., Ohsawa, M., Momohara, A., Yi, S.R., Robertson, K., Song, K., Zhang, S.Q., He, L.Y., 2015. Community structure and survival of Tertiary
relict Thuja sutchuenensis (Cupressaceae) in the subtropical Daba Mountains, southwestern China. PLoS One 10, e0125307. http://dx.doi.org/10.1371/
journal.pone.0125307.
Tang, C.Q., Yang, Y., Ohsawa, M., Yi, S.R., Momohara, A., Su, W.H., Wang, H.C., Zhang, Z.Y., Peng, M.C., Wu, Z.L., 2012. Evidence for the persistence of wild
Ginkgo biloba (Ginkgoaceae) populations in the Dalou Mountains, southwestern China. Am. J. Bot. 99, 1408–1414.
Thorn, J.S., Nijman, V., Smith, D., Nekaris, K.A.I., 2009. Ecological niche modelling as a technique for assessing threats and setting conservation priorities for
Asian slow Lorises (Primates: Nycticebus). Diversity and Distributions 15, 289–298.
Wilson, C.D., Roberts, D., Reid, N., 2011. Applying species distribution modelling to identify areas of high conservation value for endangered species: a case
study using Margaritifera margaritifera (L.). Biol. Cons. 144, 821–829.
Wisz, M.S., Hijmans, R.J., Li, J., Peterson, A.T., Graham, C.H., Guisan, A., 2008. Effects of sample size on the performance of species distribution models.
Diversity-Distribut. 14, 763–773.
Xiang, Q.P., Farjon, A., Li, Z.Y., Fu, L.K., Liu, Z.Y., 2002. Thuja sutchuenensis: a rediscovered species of the Cupressaceae. Bot. J. Linnean. Soc. 140, 93.
Yang, X.Q., Kushwaha, S.P.S., Saran, S., Xu, J., Roy, P.S., 2013. Maxent modeling for predicting the potential distribution of medicinal plant, Justicia adhatoda
L. in Lesser Himalayan foothills. Ecol. Eng. 51, 83–87.