Acta Oecologica 49 (2013) 23e31

Contents lists available at SciVerse ScienceDirect

Acta Oecologica
journal homepage: www.elsevier.com/locate/actoec

Original article

Predicting impacts of climate change on medicinal asclepiads of

Pakistan using Maxent modeling
Rizwana Khanum a, *, A.S. Mumtaz a, Sunil Kumar b
a
Quiad-i-Azam University, Islamabad 44000, Pakistan
b
Natural Resource Ecology Laboratory, Colorado State University, Fort Collins, CO 80523, USA

a r t i c l e i n f o a b s t r a c t

Article history: Maximum entropy (Maxent) modeling was used to predict the potential climatic niches of three medici-
Received 25 November 2012 nally important Asclepiad species: Pentatropis spiralis, Tylophora hirsuta, and Vincetoxicum arnottianum. All
Accepted 14 February 2013 three species are members of the Asclepiad plant family, yet they differ in ecological requirements,
Available online
biogeographic importance, and conservation value. Occurrence data were collected from herbarium
specimens held in major herbaria of Pakistan and two years (2010 and 2011) of field surveys. The Maxent
Keywords:
model performed better than random for the three species with an average test AUC value of 0.74 for
Climate change
P. spiralis, 0.84 for V. arnottianum, and 0.59 for T. hirsuta. Under the future climate change scenario, the
Niche modeling
Maxent
Maxent model predicted habitat gains for P. spiralis in southern Punjab and Balochistan, and loss of habitat
Species distribution modeling in south-eastern Sindh. Vincetoxicum arnottianum as well as T. hirsuta would gain habitat in upper Peaks of
Species vulnerability northern parts of Pakistan. T. hirsuta is predicted to lose most of the habitats in northern Punjab and in
Risk analysis parches from lower peaks of Galliat, Zhob, Qalat etc. The predictive modeling approach presented here may
be applied to other rare Asclepiad species, especially those under constant extinction threat.
Ó 2013 Elsevier Masson SAS. All rights reserved.

1. Introduction group of perennial herbs, twining shrubs, lianas, stem succulents or

It is widely accepted that biodiversity is being lost at an unprec-
edented rate (Pimm et al.,1995). Climate change is a significant driver
for biodiversity loss as it may affect species’ natural distribution,
cause temporal reproductive isolation, and increase pest and disease
outbreak frequencies (Millennium Ecosystem Assessment, 2005).
The Intergovernmental Panel on Climate Change (IPCC) estimates a
0.2
C increase in temperature for each future decade (IPCC, 2007).
This increase in temperature would have harmful consequences for
ecosystems (Millennium Ecosystem Assessment, 2005). To mitigate
the impacts of climate change on ecosystems, biodiversity conser-
vation is a key objective that will require both quantifying biodiver-
sity and monitoring its losses (Balmford and Bond, 2005). Modeling
species distribution can provide a useful means of commissioning
future surveys in predicted species distribution area, consequently
prioritizing conservation needs (Guisan and Thuiller, 2005).
Subcontinents, especially India and Pakistan, are a secondary
center of diversity for Asclepiads. Briefly stating, asclepiads form a
Abbreviations: P. spiralis, Pentatropis spiralis; T. hirsuta, Tylophora hirsuta;
V. arnottianum, Vincetoxicum arnottianum; GB, Gilgit Baltistan; PMNH, Pakistan
Museum of Natural History.
* Corresponding author. Tel.: þ92 51 2520788; fax: þ92 51 929087.
E-mail addresses: rizvana.khan@gmail.com (R. Khanum), asmumtaz@
qau.edu.pk (A.S. Mumtaz), sunil.kumar@colostate.edu (S. Kumar).
rarely trees of dicotyledonous plants that contains over 3000
known species (Meve, 2002). The name originated from type genus
Asclepias (milkweeds). Previously they belonged to the family
Asclepiadaceae, but now classified as the subfamily Asclepiadoi-
deae of the dogbane family Apocynaceae (Endress and Bruyns,
2000). Asclepiads are named for their milky juice. Pollination in
this genus is accomplished in an unusual manner. Pollen is grouped
into complex structures called pollinia. These species produce their
seeds in follicles ((Nasir and Ali, 1982).
Pakistan has a unique assemblage of species that are under
serious threat of extinction from habitat loss and rapidly developing
climatic changes. Surprisingly, there have been no studies investi-
gating the impact of climatic change on the distribution of Ascle-
piads. Preliminary studies report explorations in the Lal Suhanra
Park in southern Punjab (Hameed et al., 2002), Harboi rangeland
near Quetta, and Baluchistan (Durrani and Hussain, 2005). These
include studies on phytosociology of grass-dominated communities
of Karachi (Khan and Shaukat, 2005), and phytosociology and
structure of Himalayan forests (Ahmed et al., 2006). Based on As-
clepiads collected from the study areas and herbaria, the three most
medicinally important species are Pentatropis spiralis, Tylophora
hirsuta, and Vincetoxicum arnottianum.
Among these species, P. spiralis has antimicrobial and antifungal
properties, and its tubers are edible. T. hirsuta has laxative, expec-
torant, diaphoretic and purgative properties. This plant is used for

1146-609X/$ e see front matter Ó 2013 Elsevier Masson SAS. All rights reserved.
http://dx.doi.org/10.1016/j.actao.2013.02.007
24 R. Khanum et al. / Acta Oecologica 49 (2013) 23e31

the treatment of various disorders such as asthma and retention of 2.2. Climatic data
urine (Srivpuri et al., 1972; Thiruvengadam et al., 1978; Udupa et al.,
1991; Reddy et al., 2009, 2010). Other species of Tylophora, Climatic data were downloaded from the WorldClim database,
including Tylophora indica also known as Tylophora asthmatica, and available at approximately 1 km2 spatial resolution (Hijmans et al.,
Tylophora vomitoria, are used for traditional medicines and are 2005; http://www.worldclim.org/). The WorldClim data are
mentioned in pharmacopoeia texts such as Bengal Pharmacopoeia derived from measurements of altitude, temperature and rainfall
and Chinese Materia Medica (Shiu-Ying, 1980). Similarly, from weather stations across the globe (Period 1950e2000). We
V. arnottianum along with other species of this genus, are used for used 19 bioclimatic variables (Table 1) from the WorldClim dataset
ailments like gastritis, malaria, cholera, asthma, skin diseases, ulcer to assess current climatic conditions. These variables are frequently
and constipation (Jan et al., 2008). Vincetoxicum arnottianum is also used in modeling species distributions (e.g., Kumar et al., 2009;
used to treat external wounds and injuries in humans and animals Evangelista et al., 2011; Sanchez et al., 2011), and capture annual
(Zaidi and Crow, 2005). Phytochemically it contains floral volatile ranges, seasonality, and limiting factors such as monthly and quar-
(Jurgens et al., 2009), alkaloids (Stockel et al., 1969) and other terly temperature and precipitation extremes (Hijmans et al., 2005).
medicinal properties (Ertug , 2000). Future climate scenario data for 2050 (A2a emission scenario) were
This study used ecological niche modeling (ENM) to predict the obtained from Consultative Group on International Agricultural
distribution of the three medicinally important Asclepiads Research (CGIAR)’s Research Program on Climate Change, Agricul-
(P. spiralis, T. hirsuta and V. arnottianum) in Pakistan using combined ture and Food Security (CCAFS) climate data archive (http://ccafs-
occurrence data from field and historical (herbaria) specimens. climate.org). These future climate projections are based on IPCC
Ecological niche modeling provides a strong predictive framework 4th assessment data and were calibrated and statistically down-
for locating additional field populations of a species (Menon et al., scaled using the data for ‘current’ conditions. To reduce the uncer-
2010), and for relocating threatened populations to suitable habi- tainty in single global circulation model (GCM) predictions, we used
tats. ENM merges known occurrence records for a species with data from three different global climate models. These models were:
environmental data to estimate species ecological requirements Canadian Center for Climate Modeling and Analysis (CCCMA),
and potential geographic distribution patterns. This estimation Hadley Coupled Model V3 (HadCM3), and Commonwealth Scientific
helps to narrow down sets of possible occurrence sites for more and Industrial Research Organization (CSIRO). We ran separate
targeted field surveys (Menon et al., 2010). Several authors have models for each species using future climate data from these three
described using ENM to locate poorly known species (e.g., de- GCMs. Future potential habitat predictions for each species were
Siqueira et al., 2009; Buchmann et al., 2010; Sanchez et al., 2011). obtained by averaging results (ensemble approach; Araujo and New,
These species were chosen as a case study to represent taxa with 2007) from the CCCMA, HadCM3 and CSIRO future climate models.
contrasting ecological requirements, biogeographic importance
and conservation value. This approach may also be applied to other
2.3. Predictive modeling
taxa in the geographic region to assess and manage their respective
conservation needs. We believe that ENM modeling is an important
We used the maximum entropy model (Maxent version 3.3.3;
conservational tool for present and future distribution of species.
Phillips et al., 2006; http://www.cs.princeton.edu/wschapire/
maxent/) because it performs better with small sample sizes
2. Materials and methods
Table 1
2.1. Occurrence data collection Percent contributions of the bioclimatic variables in the Maxent models for the three
target species; values shown are averages over 10 replicate runs. Variables without
A total of 553 records for all three Asclepiad species were any values (indicated by -) were removed because of high cross-correlations.
collected from major herbaria [Karachi University Herbarium (KAR), P. spiralis T. hirsuta V. arnottianum
National Herbarium Program, National Agricultural Research Council Precipitation seasonality (CV) (Bio15) 27.2 e e
(NARC), the Quaid-i-Azam University Herbarium (IBD) and Pakistan Precipitation of driest quarter (Bio17) 21.6 e e
Museum of Natural History (PMNH)] in Pakistan. Encarta Atlas (2011) Temperature seasonality (SD  100) 21.2 18.8 5.3
and Google Earth (2011) were used to geo-reference herbarium re- (Bio4)
Precipitation of wettest month (Bio13) 13.3 e e
cords. The locality of each specimen was transformed into Mean temperature of wettest quarter 12.7 e e
geographic coordinates (WGS84 datum) using National Geographic (Bio8)
Tylor software (2003). Based on herbarium records, fresh speci- Mean diurnal range in temperature 4.0 e e
mens were collected during field surveys in 2010 and 2011, from (Bio2)
Annual mean temperature (Bio1) e 70.0 e
March to September, in the Potohar plateau that lies between 33
280
Precipitation of warmest quarter e 9.0 e
and 33
480 N latitude, and 72
480 and 73
220 E longitude. (Bio18)
The Potohar plateau covers Rawalpindi and its adjoining areas: Precipitation of driest month (Bio14) e 2.2 e
Attock, Jhelum, Salt range, Murree, and Dungagali. The climate in this Mean annual precipitation (Bio12) e e 72.3
area varies from sub-temperate to temperate with arid conditions, Temperature annual range (Bio7) e e 13.9
Mean temperature of warmest quarter e e 8.5
and allows for a wide variety of trees and plants to flourish (Blood, (Bio10)
1996). Walking transects were used in field surveys in the moun- Isothermality (Bio3) e e e
tainous terrain of the Margalla Hills, regions of the lower and outer Maximum temperature of warmest e e e
Himalayas, and the Hazara and Kala Chitta Ranges. A Global Posi- month (Bio5)
Minimum temperature of coldest e e e
tioning System (GPS, Garmin 12) was used to determine the co-
month (Bio6)
ordinates for all collections in the field. Appendix A shows number of Mean temperature of driest quarter e e e
occurrence records collected for three plant species from different (Bio9)
regions of Pakistan. After maintaining one record per 1-km2 cell; as Mean temperature of coldest quarter e e e
Maxent automatically removed duplicates, we had 99 unique re- (Bio11)
Precipitation of wettest quarter (Bio16) e e e
cords for P. spiralis, 26 for T. hirsuta, and 36 for V. arnottianum (Detail Precipitation of coldest quarter (Bio19) e e e
of data are available from the first author on request).
 R. Khanum et al. / Acta Oecologica 49 (2013) 23e31
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 occur-
rence that is closest to uniform while still subject to environmental
constraints (Elith et al., 2011). Maxent automatically includes var-
iable interactions and can handle continuous and categorical pre-
dictor variables. It uses a set of features (e.g., linear, quadratic,
product, threshold and hinge) that are functions of environmental
variables that constrain the geographic distribution of a species. It
also uses a regularization parameter, which is determined empiri-
cally, to control model over fitting. Maxent generates an estimate of
habitat suitability for the species that varies from 0 (lowest suit-
ability) to 1 (highest suitability). Finally, Maxent generates
response curves for each predictor variable and has a jackknife
option that estimates the relative influence of individual predictors.
We used 10-fold cross-validation for testing the model per-
formance. We ran 10 replicates for each species and averaged the
results. We used the jackknife procedure and percent variable
contributions to estimate the relative influence of different pre-
dictor variables. The Area under the ROC (receiver operating
characteristic) curve (AUC) was used to evaluate model perfor-
mance. AUC is a measure of model performance and varies 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); Maxent calculates
AUC value slightly differently (see Phillips et al., 2006). Since our
species occurrence data were not randomly collected we gener-
ated a Kernel Density Estimator (KDE) surface following Elith et al.
(2010) to draw 10,000 random background points in Maxent. We
did this using Software for Automated Habitat Modeling (SAHM;
Morisette et al., 2013). Model thus trained were projected to the
entire area of Pakistan for future and current potential habitat
predictions.
Current 19 bioclimatic variables were reduced to fewer variables
for each species after examining cross-correlations among them to
account for multicollinearity (Graham, 2003). We used r  0.80
(Pearson correlation coefficient) as a cut-off threshold to determine
the exclusion of highly correlated variables. The reduced number of
predictor variables for P. spiralis was six, four for T. hirsuta, and four
for V. arnottianum (Table 1). Using Maxent-generated response
curves, we also examined relationships between the habitat suit-
ability for a species and bioclimatic variables. To convert from the
continuous suitability index maps to binary habitat and non-
habitat maps, a probability threshold is needed to determine po-
tential changes in future habitat for a species. The choice of a
threshold value is critical because model results and outputs vary
based on the applied threshold. We used “10th percentile training
presence threshold” to define habitat and non-habitat (or unsuit-
able areas) for all species (Pearson et al., 2004). To avoid erroneous
predictions of suitable habitat under future climate scenarios for
2050 we used the ‘fade-by-clamping’ option in Maxent, which
removes heavily clamped pixels from the final predictions (Phillips
et al., 2006).
3. Results
Models for the three species performed better than random,
with average test AUC values ranging from 0.59 to 0.83. The current
distributions of P. spiralis and V. arnottianum were largely affected
by the precipitation variables, whereas T. hirsuta distribution was
strongly predicted by temperature variables (Table 1). Appendices
B, C and D show the bioclimatic profiles (i.e., range of different
bioclimatic variables) for the three plant species. Among three
25
species P. spiralis had higher temperature tolerance varying from a
minimum of 5.1
C to a maximum of 27.9
C (Appendix B).
The current and future distributions of the species as modeled
are as follows. The Maxent model for P. spiralis performed very well
with an average AUC value of 0.74 (0.10). Current suitable habitats
for P. spiralis were predicted in northern Punjab and the southern
part of Sindh, with patches in the central parts of Khyber Pakhtun
Khuwa (KPK) and south-eastern Balochistan (Fig. 1a). Precipitation
seasonality (Bio15) and precipitation of driest quarter (Bio17)
provided the most useful and unique information for P. spiralis,
information that was not present in other variables (Fig. 2a). These
variables were the top two predictors in the Maxent model with
contributions of 27.2% and 21.6% (Table 1). Averaged future pre-
dictions from three climatic models for 2050 (A2a emission sce-
nario) showed an increase in suitable habitat for P. spiralis in
southern Punjab and southern Balochistan, and a loss of habitat in
south-eastern Sindh (Fig. 1b). The suitability of habitat for P. spiralis
initially increased sharply with increasing precipitation season-
ality; it increased slowly with increasing precipitation of driest
quarter (Fig. 3a, b).
The Maxent model for T. hirsuta had an AUC value of 0.59
(0.14). Current suitable habitats for T. hirsuta were predicted in
northern parts of Pakistan, including Gilgit Baltistan, tribal areas,
upper parts of Balochistan (Zhob), and in the upper Northern
Punjab (patches in Islamabad and Murree) (Fig. 1c). The mean
annual temperature (Bio1) and temperature seasonality (Bio 4)
showed higher effects on the distribution of T. hirsuta relative to
other bioclimatic variables (Fig. 2b). This was evident through the
higher training gain and training AUC for these variables. Mean
annual temperature (Bio1) had the top contribution of 70%
(Table 1), while temperature seasonality (Bio 4) had 18% contri-
bution (Table 1). Averaged future predictions from three climatic
models for 2050 (A2a emission scenario) showed predictions of
complete loss of habitat for T. hirsuta in northern Punjab (Islam-
abad, Murree hills), KPK (Abbotabad, Galliat) and some parts of
Baluchistan (Zhob, Dera Murad Jamali, Lower edges of Qalat)
(Fig. 1d). There was a gain in habitat in upper peaks of northern
parts of Pakistan (Gligit, Skardu). The habitat suitability of T. hirsuta
increased slowly with increases in the mean annual temperature
(Bio1) and temperature seasonality (Bio 4) (Fig. 3c, d).
The model performance for V. arnottianum was better than
random with an AUC value of 0.83 (0.10). Current suitable habitats
for V. arnottianum were predicted in hilly parts of Punjab (the
Murree hills) and upper parts of Balochistan (Quetta, Panjgur,
Turbat) (Fig. 1e). This model performed well in the Sindh province,
except in the southern portion where there was evidence of over
estimation. Mean annual precipitation (Bio12) and Temperature
annual range (Bio7) had higher training gain and AUC values than
other bioclimatic variables for V. arnottianum (Fig. 2c). Mean annual
precipitation (Bio12) contributed 72% and temperature annual
range (Bio7) contributed 13.9% to the Maxent model for
V. arnottianum (Table 1). Averaged future predictions from three
climatic models for 2050 (A2a emission scenario) showed loss of
habitat in patches from the northern parts of Punjab (Murree Hills),
edges of tribal areas and southern parts of Baluchistan. There was
gain in habitats in upper peak areas of the country (Fig. 1f). The
probability of presence of V. arnottianum increased with an increase
in mean annual precipitation (Bio12), but decreased sharply with
increasing temperature annual range (Bio7) (Fig. 3e, f).
In sum, changes in habitat distribution for the three species
examined under future climate were spatially heterogeneous. More
specifically, the model predicted that T. hirsuta would lose sub-
stantial habitat in the southern parts of Pakistan while
V. arnottianum as well as T. hirsuta would gain habitat in upper
northern parts of Pakistan.
26 R. Khanum et al. / Acta Oecologica 49 (2013) 23e31

Fig. 1. Predicted current (suitable and unsuitable) and future (suitable/stable, lost, gained and unsuitable) habitat for P. spiralis (a, b), T. hirsuta (c, d) and V. arnottianum (e, f). Future
predictions are based on an ensemble of predictions from three global circulation models for 2050 (A2a emission scenario).

4. Discussion
In regard to future distribution of the three medicinal Asclepiads
in this study, modeling suggests that their geographic distributions
would shrink under predicted levels of climate warming. The
Maxent model performed better than random for all three impor-
tant medicinal Asclepiads in Pakistan, with reasonably good AUC
values. Maxent predicted gains in suitable habitat for P. spiralis in
southern Punjab and southern Balochistan, and loss of habitat in
south-eastern Sindh (Fig. 1b). T. hirsuta showed loss of habitat in
northern parts of Punjab (Islamabad, Rawalpindi and Murree hills),
as well as patchy distribution in KPK and Baluchistan provinces
(Fig. 1d). Vincetoxicum arnottianum showed habitat expansions to-
wards northern areas, Galiats, and other mountainous parts of
Pakistan (Murree) (Fig. 1f). In this study, ENM models using future
climate scenarios predicted the greatest risk of habitat loss for
T. hirsuta, followed by P. spiralis then V. arnottianum (Fig. 1f). Since
Asclepiads are highly dependent on local climate, the plants sprout
only when there are moist to moderately moist conditions and
most of the species rapidly diminish as the precipitation decreases
(personal observation; R. Khanum). As earlier studies show, seed-
lings are not resistant to drought and might not be able to establish
(Wickens, 1982). As the model shows, suitable habitat under the
current climatic conditions would become unsuitable in the future,
resulting in local extinction, the ultimate consequence.
Some species may be capable of adapting to future climatic
conditions through phenological or physiological changes, or
through adaptations to microclimate conditions responsible for
population survival. In this study, V. arnottianum as well as T. hirsuta
showed range expansions towards higher peaks of the country,
which may be facilitated through adaptations. Examples of these
adaptations may be found in the literature like dispersal of seed is
also a significant factor for predicting climatic changes on plant
species (Neilson et al., 2005; Midgley et al., 2006). A study across 26
 R. Khanum et al. / Acta Oecologica 49 (2013) 23e31
Fig. 2. Relative predictive power of different bioclimatic variables based on the jackknife of regularized training gain in Maxent models for P. spiralis (a), T. hirsuta (b) and
V. arnottianum (c). Values shown are average over 10 replicate runs.
mountains in Switzerland showed that alpine flora have expanded
their ranges upward toward peaks (Grabherr et al., 1994; Pauli et al.,
1996), and a similar study indicated the upward movement of
treelines in Siberia Moiseev and Shiyatov (2003). Another consid-
eration is the future utilization pressure on these medicinal plants as
other plant species fail to cope with the climatic changes. This could
cause extinction of the more utilized plants despite the current and
future suitable habitats (Svenning et al., 2009). Our study shows that
medicinal Asclepiads are threatened by climatic changes in
Pakistan; similar findings have been reported from other parts of the
world (e.g., a study on Baobab trees; Sanchez et al., 2011; Wickens
and Lowe, 2008). Apart from climatic changes, other parameters
such as soil or land transformations would also contributing factors
(Pearson and Dawson, 2005; Kunstler et al., 2007) but lack of suf-
ficient data was the possible reason for incorporation in this study.
Our results may also be affected by the lack of data from species’
entire distributional range as the occurrence data from outside
Pakistan for these plant species were not available; this should be
done in the future modeling efforts.
Species with known economic value experience dual pressure
due to: a) loss of habitat from rapid climate change and changes in
land use and land cover, and b) over-exploitation because of their
known utility. Both threats are serious if not managed properly.
Climate change is one of the greatest challenges to biodiversity, and
affects all organisms. Changes have already been observed in
different regions around the world, including impacts to change
27
threats to plant diversity in Europe (Thuiller et al., 2005), Mediter-
ranean habitat (e.g boreal tree species i.e Pinus sylvestris, Castro
et al., 2004) or eroding the geographical range of the Namib
Desert tree Aloe through population declines and dispersal lags (e.g
Foden et al., 2007). Climatic changes are resulting in drastic effects
on plant life cycles and distributions (Cavaliere, 2009), and increased
global temperatures are having a negative impact on woody plant
species (Allen, 2009). Ecological niche models are commonly used
for predicting species distributions under changing climates (e.g.,
Buisson et al., 2010; Sanchez et al., 2011), and these models may be
used to make recommendations for conservation practitioners to
deal with potential climate change (Sanchez et al., 2011).
4.1. Conservation strategies
Many areas predicted to have suitable habitat for these medicinal
plants might already be devoid of populations due to land use trans-
formations and human exploitation for medicinal purposes. Several
researchers proved this globally including Midgley et al. (2003) who
suggested that the combined effects of future land transformations
and climatic changes will increase unsuitable habitats in Cape Floristic
region even more than evaluated. Practically the entire study region is
transected by roads or within the vicinity of populated places, and
much of the suitable habitat for these plants has already been con-
verted into agriculture or urbanized (e.g., in the second year of this
two-year study, we found that almost 30% of the natural land had been
28 R. Khanum et al. / Acta Oecologica 49 (2013) 23e31

Fig. 3. Response curves showing the relationships between the probability of presence of a species and two top bioclimatic predictors of P. spiralis (a, b), T. hirsuta (c, d) and
V. arnottianum (e, f). Values shown are average over 10 replicate runs; blue margins show 1 SD calculated over 10 replicates.

transformed; personal observation, R. Khanum). Successful conser-
vation planning for human-dominated landscapes can be used to
address this issue and generate connectivity across protected areas
(Urbina-Cardona and Flores-Villela, 2010).
Another possible conservation strategy, especially for areas at high
risk of habitat loss (e.g., south-eastern Sindh for P. spiralis and
northern upper parts of Punjab for T. hirsuta) could be compensation
by ex-situ germplasm collection, as suggested by Sanchez et al. (2011)
for Adansonia digitata (African baobab tree). The areas with dense
populations (e.g., plain areas of Punjab for P. spiralis, areas of Gilgit
Biltastan for T. hirsuta) may contain interesting genetic pools for
future domestication of these medicinal plants. According to Hampe
and Petit (2005), it is important to conserve the population at the rear
edge of shifting ranges. Furthermore, seeds and follicles may be
preserved and will remain viable for at least a few years (Sacande
et al., 2006). Another potential in-situ conservation strategy in-
volves introducing a forestry law that will limit access rights to these
medicinal plants; however, this strategy appears to be unsuccessful in
Africa (Buchmann et al., 2010). While in circumstances of Pakistan
state control of the forests was never accepted by the local
population, particularly in those forests where traditional usufructs
rights had existed. This resulted in a conflictual relationship between
the government and the locals that has persisted over the years
(Hassan, 2001; Shahbaz et al., 2007). The most effective strategy for
protecting viable medicinal Asclepiads in the areas of suitable future
habitat is “conservation through utilization” (Sanchez et al., 2011). For
example, if people are convinced that they may reap higher economic
benefits from these plants in their natural habitat than if grown
domestically, conservation may become a priority. Recognizing the
value of medicinal Asclepiads and promoting their conservation
could also help to preserve the ecosystem where these plants live.
5. Conclusions
When considering the results in this study generated from
ENMs, there will be moderate to high impacts under future climate
scenarios on the distribution of three important medicinal Ascle-
piads in Pakistan. Vincetoxicum arnotianum showed the most vari-
ation, as it would lose habitats in its current range but gain habitat
in higher altitudes due to climate change. Plants from moderate
 R. Khanum et al. / Acta Oecologica 49 (2013) 23e31 29

altitude and climate (e.g., T. hirsuta) will suffer more habitat loss Appendix C
from climate change compared to others. Conservation action is, Bioclimatic profile of Tylophora hirsuta based on its occurrence data; SD is standard
therefore, immediately needed for T. hirsuta. Multiple strategies deviation.
could be used to preserve the medicinal Asclepiads and maintain Variable Minimum Maximum Mean SD
not only the pharmaceutical and income resources of local people
Annual mean temperature (
C; Bio1) 6.6 22.5 19.5 3.5
in Pakistan, but also the ecosystems. The options range from in-situ Mean diurnal range in temperature 8.2 14.3 12.8 1.7
conservation in protected areas with suitable habitats to ex-situ (
C; Bio2)
conservation in seed banks, and conservation through utilization of Isothermality (Bio3) 28 38 36.7 2.3
Temperature seasonality (SD  100) 6141.0 7816.0 7237.7 361.9
these medicinal plants.
(Bio4)
Maximum temperature of 20.7 40.7 36.4 4.7
Acknowledgments warmest month (
C; Bio5)
Minimum temperature of coldest 7.9 3.9 2.2 2.3
month (
C; Bio6)
We thank the Natural Resource Ecology Laboratory at Colorado
Temperature annual range (
C; Bio7) 26.6 38.0 34.2 2.9
State University for providing logistical support. We also thank Ms. Mean temperature of wettest quarter 9.6 29.8 25.6 5.6
Ghazala, Mustafa for helping us in field, and Shabir for helping in (
C; Bio8)
the collection and deposition of plants in herbarium (Pakistan Mean temperature of driest quarter 3.5 17.7 15.2 3.0
Museum of Natural History). We are grateful to Amanda West and (
C; Bio9)
Mean temperature of warmest 15.6 31.3 28.0 3.6
Aaron Sidder and two anonymous reviewers for their useful com- quarter (
C; Bio10)
ments and suggestions that improved the manuscript. Mean temperature of coldest 3.6 12.1 9.5 3.3
quarter (
C; Bio11)
Appendix A Mean annual precipitation (mm; Bio12) 547 1536 938.5 218.0
Precipitation of wettest month 80 326 198.8 54.8
Province based records of each species (mm; Bio13)
Precipitation of driest month 7 36 18.6 6.1
Sp. name KPK BALUCHISTAN PUNJAB SINDH FATA AJK (mm; Bio14)
Vincetoxicum 19 3 5 0 2 7 Precipitation seasonality (CV) (Bio15) 41 94 72.3 15.7
arnotinaum Precipitation of wettest quarter 194 717 470.5 122.3
Tylophora hirsuta 3 0 18 0 6 (mm; Bio16)
Pentatropis spiralis 7 10 45 34 0 0 Precipitation of driest quarter 34 134 83.1 22.0
(mm; Bio17)
Precipitation of warmest quarter 170 698 431.0 126.5
(mm; Bio18)
Precipitation of coldest quarter 83 293 173.2 45.4
Appendix B (mm; Bio19)

Bioclimatic profile of Pentatropis spiralis based on its occurrence data; SD is standard Appendix D
deviation.
Bioclimatic profile of Vincetoxicum arnotianum based on its occurrence data; SD is
Variable Minimum Maximum Mean SD
standard deviation.


Annual mean temperature ( C; Bio1) 5.1 27.9 23.4 4.4
Mean diurnal range in temperature 8.5 17.1 13.8 1.8 Variable Minimum Maximum Mean SD
(
C; Bio2) Annual mean temperature (
C; Bio1) 6.9 26.5 17.4 4.3
Isothermality (Bio3) 24.0 49.0 40.8 4.3 Mean diurnal range in temperature 8.1 16.7 11.7 2.1
Temperature seasonality (SD  100) 4116 8858 6664.8 1215.0 (
C; Bio2)
(Bio4) Isothermality (Bio3) 28 47 35.9 4.1
Maximum temperature of warmest 12.6 44.5 39.0 4.3 Temperature seasonality (SD  100) 4626.0 8633.0 7051.4 735.3
month (
C; Bio5) (Bio4)
Minimum temperature of coldest 22.7 13.1 5.31 4.7 Maximum temperature of warmest 21.0 40.7 33.4 5.2
month (
C; Bio6) month (
C; Bio5)
Temperature annual range (
C; Bio7) 22.6 39.3 33.7 4.5 Minimum temperature of coldest 9.4 12.8 1.2 3.7
Mean temperature of wettest quarter 6.4 33.4 28.2 6.4 month (
C; Bio6)
(
C; Bio8) Temperature annual range (
C; Bio7) 25.0 38.0 32.1 3.5
Mean temperature of driest quarter 2.2 33.8 20.0 4.4 Mean temperature of wettest quarter 6.9 30.7 21.0 7.0
(
C; Bio9) (
C; Bio8)
Mean temperature of warmest 5.9 35.5 30.9 3.7 Mean temperature of driest quarter 3.9 29.1 14.4 4.6
quarter (
C; Bio10) (
C; Bio9)
Mean temperature of coldest 16.9 19.8 14.0 5.2 Mean temperature of warmest quarter 15.9 32.2 25.8 4.3
quarter (
C; Bio11) (
C; Bio10)
Mean annual precipitation 96.0 1468.0 448.3 303.5 Mean temperature of coldest quarter 3.8 19.8 7.7 4.5
(mm; Bio12) (
C; Bio11)
Precipitation of wettest month 23 305 120.7 66.3 Mean annual precipitation (mm; Bio12) 102 1536 964.1 400.8
(mm; Bio13) Precipitation of wettest month 30 326 182.4 84.3
Precipitation of driest month 0 34 6.0 6.4 (mm; Bio13)
(mm; Bio14) Precipitation of driest month 1 36 21.2 9.7
Precipitation seasonality (CV) 45 165 103.7 28.0 (mm; Bio14)
(Bio15) Precipitation seasonality (CV) (Bio15) 41 98 62.3 11.0
Precipitation of wettest quarter 52 676 267.7 160.8 Precipitation of wettest quarter 61 717 431.5 192.2
(mm; Bio16) (mm; Bio16)
Precipitation of driest quarter 3 129 30.4 26.9 Precipitation of driest quarter 3 134 90.2 38.0
(mm; Bio17) (mm; Bio17)
Precipitation of warmest quarter 11 656 203.0 148.6 Precipitation of warmest quarter 19 698 393.7 199.7
(mm; Bio18) (mm; Bio18)
Precipitation of coldest quarter 7 284 74.5 58.3 Precipitation of coldest quarter 61 294 193.7 70.4
(mm; Bio19) (mm; Bio19)
30 R. Khanum et al. / Acta Oecologica 49 (2013) 23e31
Appendix E. Consensus among predictions across three
different Global Circulation Models (GCMs) for (a) Pentatropis
spiralis (predictions from only two models [CCCMA &
HadCM3] are shown here because CSIRO predictions looked
very unrealistic), (b) Tylophora hirsuta, and (c) Vincetoxicum
arnottianum. Cyan color shows unsuitable habitat (i.e. none of
the GCMs predicted suitable habitat for a species), Yellow
shows suitable habitat predicted by at least one of the models,
Green shows suitable habitat predicted by two models, and
Red shows suitable habitat predicted by all three models.
References
Ahmed, M., Husain, T., Sheikh, A.H., Hussain, S.S., Siddiqui, M., 2006. Phytosociology
and structure of Himalayan forests from different climatic zones of Pakistan.
Pakistan Journal of Botany 38 (2), 361e383.
Allen, C.D., 2009. Climate-induced forest dieback: an escalating global phenome-
non? Unasylva 60 (231/232), 43e49.
Araujo, M.B., New, M., 2007. Ensemble forecasting of species distributions. Trends in
Ecology & Evolution 22, 42e47.
Balmford, A., Bond, W., 2005. Trends in the state of nature and their implications for
human well-being. Ecology Letters 8 (11), 1218e1234.
Blood, P.R., 1996. Pakistan: A Country Study. DIANE Publishing, Washington.
Buchmann, C., Prehsler, S., Hartl, A., Vogl, C.R., 2010. The Importance of Baobab
(Adansonia digitata L.) in rural West African subsistence-suggestion of a
cautionary approach to International Market export of baobab fruits. Ecology
Food & Nutrition 49, 145e172.
Buisson, L., Thuiller, W., Casajus, N., Lek, S., Grenouillet, G., 2010. Uncertainty in
ensemble forecasting of species distribution. Global Change Biolgy 16, 1145e1157.
Castro, J., Zamora, R., Hodar, J.A., Gomez, J.M., 2004. Seedling establishment of a
boreal tree species (Pinus sylvestris) at its southernmost distribution limit:
consequences of being in a marginal Mediterranean habitat. Journal of Ecology
92, 266e277.
Cavaliere, C., 2009. The effects of climate change on medicinal and aromatic plants.
Herbal Gram (American Botanical Council) 81, 44e57.
de-Siqueira, M.F., Durigan, G., Junior, P.M., Peterson, A.T., 2009. Something from
nothing: using landscape similarity and ecological niche modeling to find rare
plant species. Journal for Nature Conservation 17 (1), 25e32.
Durrani, M.J., Hussain, F., 2005. Ethnoecological profile of plants of Harboi range-
land, Balochistan. International Journal of Biology & Biotechnology 2, 15e22.
Elith, J., Graham, C.H., Anderson, R.P., et al., 2006. Novel methods improve predic-
tion of species’ distributions from occurrence data. Ecography 29 (2), 129e151.
Elith, J., Kearney, M., Phillips, S., 2010. The art of modelling range-shifting species.
Methods in Ecology and Evolution 1, 330e342.
Elith, J., Phillips, S.J., Hastie, T., et al., 2011. A statistical explanation of MaxEnt for
ecologists. Diversity and Distributions 17, 43e57.
Encarta, 2011. www.filestube.com/e/encartaþ2011 (assessed 05.07.11.).
Endress, M.E., Bruyns, P.V., 2000. A revised classification of the Apocynaceae sl.
Botanical Review 66 (1), 1e56.
Ertug, F., 2000. An ethnobotanical study in central Anatolia (Turkey). Economic
Botany 54 (2), 155e182.
Evangelista, P.H., Kumar, S., Stohlgren, T.J., Young, N.E., 2011. Assessing forest
vulnerability and the potential distribution of pine beetles under current and
future climate scenarios in the Interior West of the US. Forest Ecology and
Management 262 (3), 307e316.
Fielding, A.H., Bell, J.F.,1997. A review of methods for the assessment of prediction errors
in conservation presence/absence models. Environmental Conservation 24, 38e49.
Foden, W., Midgley, G.F., Hughes, G.O., Bond, W.J., Thuiller, W., Hoffman, M.T.,
Kaleme, P., Underhill, L.G., Rebelo, A.G., Hannah, L., 2007. A changing climateis
eroding the geographical range of the Namib Desert tree Aloe through popu-
lation declines and dispersal lags. Diversity & Distributions 13, 645e653.
Google Earth, 2011. www.google.com/earth/download/ge/ (assessed 04.05.11.).
Grabherr, G., Gottfried, M., Pauli, H., 1994. Climate effects on mountain plants.
Nature 369, 448.
Graham, M.H., 2003. Confronting multicollinearity in ecological multiple regres-
sion. Ecology 84 (11), 2809e2815.
Guisan, A., Thuiller, W., 2005. Predicting species distribution: offering more than
simple habitat models. Ecology Letters 8 (9), 993e1009.
Hameed, M., Chaudhry, M.M., Man, M.A., Gill, A.H., 2002. Diversity of plant species
in LalSuhanra National Park, Bahawalpur. Pakistan Journal of Biological Sciences
2, 267e274.
Hampe, A., Petit, R.J., 2005. Conserving biodiversity under climate change: the rear
edge matters. Ecology Letters 8, 461e467.
Hassan, L., 2001. Analysing Institutional Set-up of Forest Management in Pakistan.
Pakistan Institute of Development Economics, Islamabad.
Hijmans, R.J., Cameron, S.E., Parra, J.L., et al., 2005. Very high resolution interpolated
climate surfaces for global land areas. International Journal of Climatology 25
(15), 1965e1978.
IPCC, 2007. In: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B.,
Tignor, M., Miller, H.L. (Eds.), Climatic Change 2007: The Physical Science Basis.
Contribution of Working Group I to Fourth Assessment Report of The Inter-
governmental Panel on Climate Change (IPCC). Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA.
Jan, S., Khan, M.A., Din, S.U., et al., 2008. Herbal Remedies used for gastrointestinal
disorders in Kaghan Valley, NWFP, Pakistan. Pakistan Journal of Weed Science
Research 14 (3e4), 169e200.
Jurgens, A., Dotterl, S., Liede-Schumann, S., Meve, U., 2009. Chemical diversity of
floral volatiles in Asclepiadoideae-Asclepiadeae (Apocynaceae). Biochemical
Systematics and Ecology 36, 842e852.
Khan, D., Shaukat, S.S., 2005. Above ground standing phytomass of some grass
dominated communities of Karachi: winter aspect. International Journal of
Biology & Biotechnology 2 (1), 85e92.
Kumar, S., Stohlgren, T.J., 2009. Maxent modeling for predicting suitable habitat for
threatened and endangered tree Canacomyrica monticola in New Caledonia.
Journal of Ecology and Natural Environment 1 (4), 94e98.
Kumar, S., Spaulding, S.A., Stohlgren, T.J., et al., 2009. Potential habitat distribution
for the freshwater diatom Didymosphenia geminata in the continental US.
Frontiers in Ecology and the Environment 7 (8), 415e420.
Kunstler, G., Thuiller, W., Curt, T., Bouchaud, M., Jouvie, R., Deruette, F., Lepart, J.,
2007. Fagus sylvatica L. recruitment across a fragmented Mediterranean land-
scape, importance of long distance effective dispersal, abiotic conditions and
biotic interactions. Diversity &. Distribution 13 (6), 799e807.
Menon, S., Soberon, J., Li, X.G., Peterson, A.T., 2010. Preliminary global assessment of
terrestrial biodiversity consequences of sea-level rise mediated by climate
change. Biodiversity and Conservation 19, 1599e1609.
 R. Khanum et al. / Acta Oecologica 49 (2013) 23e31
Meve, U., 2002. Species numbers and progress in asclepiad taxonomy. Kew Bulletin
57 (2), 459e464.
Midgley, G.F., Hannah, L., Millar, D., Thuiller, W., Booth, A., 2003. Developing
regional and species-level assessments of climate change impacts on biodi-
versity in the Cape Floristic Region. Biological Conservation 112, 87e97.
Midgley, G.F., Hughes, G.O., Thuiller, W., Rebelo, A.G., 2006. Migration rate limita-
tions on climate change induced range shifts in Cape Proteaceae. Diversity &
Distributions 12, 555e562.
Millennium Ecosystem Assessment, 2005. Ecosystems and Human Well-being:
Synthesis. Island Press, Washington, DC.
Moiseev, P.A., Shiyatov, S.G., 2003. The use of old landscape photographs for
studying vegetation dynamics at the tree line ecotone in the Ural Highlands,
Russia. In: Nagy, L. (Ed.), Alpine Biodiversity in Europe. Springer-Verlag, Berlin,
pp. 423e436.
Morisette, J.T., Jarnevich, C.S., Holcombe, T.R., Talbert, C.B., Ignizio, D., Talbert, M.K.,
Silva, C., Koop, D., Swanson, A., Young, N.E., 2013. VisTrails SAHM: visualization
and workflow management for species habitat modeling. Ecography 36. http://
dx.doi.org/10.1111/j.1600-0587.2012.07815.x.
Nasir, E., Ali, S.I., 1982. Flora of Pakistan. No. 142. Karachi University.
National Geographic Taylor software, 2003. http://home.hiwaay.net/wtaylorc/
toolbox/geodesy/datumtrans/. Chuck Taylor Web site are copyrightÓ 1997e
2003 by Charles L. Taylor.
Neilson, R.P., Pitelka, L.F., Solomon, A.M., Nathan, R., Midgley, G.F., Fragoso, J.,
Lischke, H., Thompson, K., 2005. Forecasting regional to global plant migra-
tion in response to climate change: challenges and directions. Bioscience 55,
749e759.
Pauli, H., Gottfried, M., Grabherr, G., 1996. Effects of climate change on mountain
ecosystems: upward shifting of mountain plants. World Research Review 8,
382e390.
Pearson, R.G., Dawson, T.P., 2005. Long-distance plant dispersal and habitat frag-
mentation: identifying conservation targets for spatial landscape planning
under climate change. Biological Conservation 123, 389e401.
Pearson, G.R., Dawson, T.P., Liu, C., 2004. Modelling species distributions in Britain:
a hierarchical integration of climate and land-cover data. Ecography 27 (3),
285e298.
Pearson, R.G., Raxworthy, C.J., Nakamura, M., Peterson, A.T., 2007. Predicting species
distributions from small numbers of occurrence records: a test case using
cryptic geckos in Madagascar. Journal of Biogeography 34, 102e117.
Phillips, S.J., Anderson, R.P., Schapire, R.E., 2006. Maximum entropy modeling of
species geographic distributions. Ecological Modelling 190 (3e4), 231e259.
Pimm, S.L., Russell, G.J., Gittleman, J.L., et al., 1995. The future of biodiversity. Science
269 (522), 347e350.
31
Reddy, B.K., Balaji, M., Reddy, P.U., et al., 2009. Antifeedant and antimicrobial activity
of Tylophora indica. African Journal of Biochemistry Research 3 (12), 393e397.
Reddy, B.K., Anjaneyulu, E., Ramgopal, M., et al., 2010. An efficient protocol for shoot
bud induction and regeneration of Tylophora indica, an endangered medicinal
plant. Current Trends in Biotechnology and Pharmacy 4, 922e929.
Sacande, E., Rønne, C., Sanon, M., et al., 2006. Adansonia Digitata L. Forest and
Landscape Denmark. Seed Leaflet. No 109.
Sanchez, A.C., Osborne, P.E., Haq, N., 2011. Climate change and the African baobab
(Adansonia digitata L.): the need for better conservation strategies. African
Journal of Ecology 49, 234e245.
Shahbaz, B., . Ali, T., Suleri, A., 2007. A critical analysis of forest policies of Pakistan:
implications for sustainable livelihoods. Mitigation and Adaptation Strategies of
Global Change 12 (4), 441e453.
Shiu-Ying, H., 1980. An Enumeration of Chinese MateriaMedica. Chinese University
of Hong Kong Press.
Srivpuri, D.N., Singhal, S.C., Parkash, D., 1972. Treatment of asthma with an alcoholic
extract of Tylophora indica e a cross-over, double-blind study. An Allergy 30 (7),
407e412.
Stockel, K., Stocklin, W., Reichstein, T., 1969. Die Glykoside von Vincetoxicum
hirundinaria. Medikus Helvitica Chemica Acta 52, 1175e1202.
Svenning, J.C., Harlev, D., Sørensen, M.M., Balslev, H., 2009. Topographic and spatial
controls of palm species distributions in a montane rain forest, southern
Ecuador. Biodiversity Conservation 18, 219e228.
Swets, J.A., 1988. Measuring the accuracy of diagnostic systems. Science 240 (4857),
1285e1293.
Thiruvengadam, K.V., Haranath, K., Sudarsan, S., 1978. Tylophoraindicain bronchial
asthma (a controlled comparison with a standard antiasthmatic drug). Journal
of the Indian Medical Association 71, 172e176.
Thuiller, W., Lavorel, S., Arau0 jo, M.B., Sykes, M.T., Prentice, I.C., 2005. Climate
change threats to plant diversity in Europe. Proceedings of National Academy of
Sciences 102, 8245e8250.
Udupa, A.L., Udupa, S.L., Guruswamy, M.N., 1991. The possible site of antiasthmatic
action of Tylophora asthmaticaon pituitary-adrenal axis in albino rats. Planta
Medica 57, 409e413.
Urbina-Cardona, J.N., Flores-Villela, O., 2010. Ecological-niche modeling and prior-
itization of conservation-area networks for Mexican Herpetofauna. Conserva-
tion Biology 24 (4), 1031e1041.
Wickens, G.E.,1982. The baobab e Africa’s upside-down tree. Kew Bulletin 37,173e209.
Wickens, G.E., Lowe, P., 2008. The Baobabs, Pachycauls of Africa, Madagascar and
Australia. Kluwer Academic Publishers Group, Dordrecht, The Netherlands.
Zaidi, M.,A., Crow, S.A., 2005. Biologically active traditional medicinal herbs from
Balochistan. Pakistan Journal of Ethnopharmacology 96 (1e2), 331e334.