Ecological niche modeling for reintroduction and conservation of Aristolochia cathcartii Hook.f. & Thomson (Aristolochiaceae), a threatened endemic plant in Assam, India

: Aristolochia cathcartii Hook.f. & Thomson is a medicinal plant species native to Assam (India). Karbi people have traditionally used the plant to treat a variety of ailments. The population stock of this species has been rapidly depleting in its natural habitats due to over-utilization, habitat fragmentation, and other anthropogenic activities. Extensive field surveys were carried out to investigate the population status of A. cathcartii in various forest areas of Assam’s Karbi Anglong district. In 20 km of transects, a total of 36 quadrats were observed. A. cathcartii density, frequency of occurrence, and abundance were recorded to be 0.65, 17.8, and 3.81, respectively. Ecological niche modelling was used to identify suitable habitat for the reintroduction and conservation of this plant in Assam in order to prevent its extinction in the future. The maximum entropy distribution modelling algorithm was used to identify suitable areas and habitat for the species’ reintroduction and conservation. Primary data on the occurrence of A. cathcartii was gathered from the natural habitat of Karbi Anglong district, Assam, for modelling. The model identified various forest areas in northeastern India that have suitable climatic conditions for plant reinforcement.


INTRODUCTION
Aristolochia cathcartii, belonging to the family Aristolochiaceae, is a large climber. Traditionally, A. cathcartii has been used by the Karbi community of Assam to treat cholera, stomach pain, fever, and poisonous bites (Sarma et al. , 2017. Overexploitation, climate change, habitat fragmentation and loss, and rapid urbanization cause gradual depletion of this medicinally important plant from its natural habitat. Species reinforcement is the best technique for the restoration of depleted species populations and degraded habitats and ecosystems (Leaper et al. 2006;Martinez-Meyer et al. 2006;Kuzovkina & Volk 2009;Ren et al. 2009;Rodríguez-Salinas et al. 2010;Polak & Saltz 2011). Ecological niche modeling helps in identifying sites of species occurrence and also helps to spot other suitable habitats for reintroduction. Ecological niche modeling (ENM) is a tool in geographic information system (GIS) software that uses occurrence data of a species across landscapes and correlates them with digital raster GIS coverage to develop a model of environmental conditions that meet ecological requirements and identify the suitable environment of the species (Guisan & Zimmermann 2000;Elith et al. 2006;Kozak et al. 2008). ENM facilitates interpolation as well as the extrapolation of species distributions in geographic space across different periods and it helps to prepare habitat distributional maps by spotting areas suitable for reintroduction of threatened species (Irfan-Ullah et al. 2006;Kumar et al. 2009;Ray et al. 2011). For conservation strategy, it is essential to identify areas which bear appropriate environmental conditions suitable for the species persistence. Therefore, the present work was undertaken to study the population distribution status of A. cathcartii and to model the habitat distributional map in its native range.

Plant material
A. cathcartii Hook.f. & Thomson belonging to the family Aristolochiaceae is a large climber, with corky furrowed bark; young branchlets, and petioles villous. Leaves 5.5-10 by 3.5-6.5 inch, broadly ovate, sometimes ovate-lanceolate, acute or acuminate, entire; base cordate, sometimes slightly lobed along the sinus, 3-nerved or pedately 5-nerved, thinly coriaceous, pubescent along the midrib and larger nerves especially towards the base, otherwise glabrous above, clothed, often felted with long silky hairs beneath; lateral nerves excluding the basal 5-6 on either side; petioles 1.5-4 inch long, sometimes twining. Flowers usually in short brown villous cymes from axils of existing or fallen leaves; pedicels 0.6-1 inch long, villous. Perianth yellowishwhite, with purple veins clothed with spreading hairs outside, 2.5-3 inch long along the bends, sac bent near the short neck, mouth square, densely purple papillose along the edge and the recurved lip. Capsule about 6.5 by 1.7 inch, linear-oblong, bluntly apiculate, softly tomentose, 6-ribbed, grooved between the furrows; seeds about 0.4 inches long, not winged, obovate, acute at the base, margins slightly incurved on the inner face, dorsally more or less truncate and margined (Kanjilal & Bor 1940). This plant is native to Assam, Bangladesh, China south-central, eastern Himalaya, Myanmar, Nepal, and Tibet (Plants of the world online, Royal Botanic Gardens, Kew).

Survey of the plant species and its population status
A frequent field visit was carried out to record the population status of A. cathcartii in Dhansiri, Kalioni, Nambar, Lahorijan, and Matipung Reserve Forest of KarbiAnglong district, Assam (India). The forest of KarbiAnglong is moist semievergreen and moist mixed deciduous type. The total population of A. cathcartii was calculated through a direct count method for all individuals. The grid size was taken 250 Ï 250 m and individuals were categorized as seedlings (<1 m height), saplings (>1 m height), and matured individuals (≥1.37 m height). The density, frequency, and abundance of the plant species were calculated with the following formulae: Ecological niche modeling Primary locations of the species were collected through field surveys. To record the coordinates of occurrence points of the species global positioning system (GPS) was used to an accuracy of 10-40 m. Then the coordinates were translated to decimal degrees to be used in habitat distribution modeling software (Adhikari & Barik 2012). For ecological modeling different types J TT of environmental datasets are available in public domain websites. In our study, the index of normalized difference vegetation (NDVI) was used to model the distributional pattern of A. cathcartii in northeastern India ( Table 2). The NDVI was obtained from Global Land Cover Facility (GLCF, University of Maryland). All the analyses were conducted at the spatial resolution of 250 m.

Validation of model robustness
For habitat modeling of A. cathcartii, the NDVI and the maximum entropy modeling (MEM) was used to develop the model (Adhikari & Barik 2012). MaxEnt uses presence-only data to predict the geographic location of a species based onthe principle of maximum entropy (Phillips et al. 2006;Elith et al. 2011). For the calibration, we used the presence and background data locations where 75% of the records were used for training the model and 25% for the test (Adhikari & Barik 2012). We conducted 20 replicated model runs and the replicated run type was cross-validation with a 10-percentile threshold rule of training presence to validate the model robustness (Adhikari & Barik 2012; Sarma et al. 2018). Since the program is already calibrated, therefore, other parameters were set as default (Adhikari & Barik 2012). Replicated runs generated average, maximum, minimum, median, and standard deviation. Quality of the model was assessed based on area under curve (AUC) value and the model was classified according to Thuiller et al. (2007) as very good (0.95 < AUC < 1.0), good (0.9 < AUC < 0.95), fair (0.8 < AUC < 0.9), and poor (AUC < 0.8).

Population status vis-à-vis model thresholds
Extensive field visits were executed to investigate the robustness and relevance of the model in predicting the population status of A. cathcartii in each occurrence area as predicted under various model thresholds. The total population of the species was calculated by direct count of all individuals of seedlings, saplings, and mature individuals in each 250 × 250 m grid of occurrence within the predicted localities. The population data of A. cathcartii in each occurrence area was then correlated with the corresponding threshold level of the distribution models to check whether regions fell under higher threshold level sustain higher populations thus favoring improved habitat conditions for species establishment and vice versa.

Analysis of habitat status and recognition of areas for reintroduction
We analyzed the habitat type in the occurrence areas of the species as well as the predicted potential areas through repeated field surveys.

Population distribution status of Aristolochia cathcartii
The population distribution status of a species indicates its importance in conservation. Species with a limited range of distribution needs to be protected more than a wide range of distribution. Considerable field surveys were conducted to explore the population status of A. cathcartii in each occurrence area. A total of 36 numbers of quadrats were observed along 20 km of transects. The density, frequency of occurrence, and abundance of A. cathcartii are shown in Table 1. The observation tabulated below depicted the mean density of A. cathcartii as 0.65, frequency of occurrence 17.77, and abundance concerning other associated species as 3.81.

Calibration of models
The model calibration test for A. cathcartii yielded satisfactory results (AUC test= 0.96 ± 0.002).

Response curves
The response curves ( Figure 1) reflect the dependence of predicted suitability both on the selected variable and on dependencies induced by correlations between the selected variable and other variables. The curves show the mean response of the 20 replicate Maxent runs (red) and the mean +/-one standard deviation (blue, two shades for categorical variables).

Analysis of variable contributions
The table 2 gives estimates of relative contributions of the environmental variables to the Maxent model. importance. The environmental variable with the highest gain, when used in isolation, is eu5_1_eur (May), which therefore appears to have the most useful information by itself. The environmental variable that decreases the gain the most when it is omitted is eu4_1_eur (April), which therefore appears to have the most information that isn't present in the other variables. Values shown are averages over replicate runs.

Population status vis-à-vis model thresholds
A total of 589 number of individuals were recorded within the area of occurrence spread over 25 250 x 250 m grids. Of these 345 numbers of individuals were adults, 187 numbers of individuals were sapling and 57 numbers of individuals were seedlings (Table 3). The analysis of population structure at each locality revealed that the highest number of adult individuals were in Dhansiri (78), Daldali (75), Lahorijan (67), Matipung (65), and Nambar (60). The population size including all adults, saplings, and seedlings was larger in the areas under the high suitability threshold category followed by amedium to low category (Table 3). Areas predicted as a medium to high suitable classes represent 84% of the total population followed by a low threshold. This establishes the strong correlation between population size and level of the model threshold. Of the 25 localities, nine localities fell under high class, 11 localities under medium, and five localities fell under low habitat suitability class.
Saplings were poorly represented in most of the areas. The number of seedlings was also very poor even absent in some areas. The number of seedlings was highest in Daldali with 13 seedlings, followed by Matipung with 12 seedlings, Nambar, and Lahorijan with 11 seedlings each, and Dhansiri with 10 seedlings. Similarly, the number of saplings also highest in Daldali with 48 individuals, followed by Nambar with 40 individuals, Lahorijan with 35 individuals, and Dhansiri and Matipung with 32 individuals each. The population structure based on a seedling, sapling, and adult individuals revealed that good regeneration takes place in the moist semi-evergreen habitat followed by mixed deciduous habitat whereas in other habitats it depicted poor regeneration (Table 3).

Habitat status assessment and identification of areas for reintroduction
Field surveys for assessing the habitat type of A. cathcartii in the predicted potential areas revealed that the species occurred in moist semi-evergreen and mixed deciduous forests. Superimposition of the predicted potential habitat distributional map of the species on Google Earth Pro, satellite imageries showed that the areas with high habitat suitability for the species were moist semi-evergreen and evergreen forests. The areas with medium habitat suitability were mixed deciduous forests and grasslands. The areas with low habitat suitability were degraded open forests and homestead gardens ( Table 4).
The superimposition of predicted potential habitat distribution map on Google Earth Pro imageries identified different forest areas of northeastern India, viz., KarbiAnglong (Rangapahar, Bokajan) district of Assam, foothills of Assam-Nagaland border (Mokokchung, Wokha, Kohima), Meghalaya (West Khasi Hills, Ri Bhoi), Arunachal Pradesh (East Siang, Papumpare) (Image 1). These areas could be used as in situ conservation and reintroduction of A. cathcartii in the wild.

DISCUSSION
A. cathcartii, is best known among the Karbi community of Assam for its high medicinal value. Locally this plant is called ChongaLota. Due to overexploitation of this plant by the local community, and other natural,

J TT
and secondary surveys using Google Earth Pro satellite imageries established that the predicted potential areas of the species fell under all suitability threshold levels, i.e., low to high suitability. Within 25 250 x 250 m grids, 589 individuals were counted, of which 345 were adults, 187 saplings, and 57 seedlings. The number of saplings and seedlings were very poor in most of the occurrence areas of the species. Areas identified as medium to high suitable classes represent 84% of the total population and it establishes a strong correlation between the population size and the model thresholds. In the present study, evergreen, moist semievergreen, and mixed deciduous forests offer potential habitats at higher levels of probability. Hence, for in situ conservation and reintroduction of A. cathcartii, such forest areas could serve as suitable habitats. The present study demonstrates that habitat distribution modeling serves as an important tool in identifying the potential habitats for the reintroduction of threatened species. The areas identified in the present study for reintroduction would help in the improvement of the conservation status of species population of A. cathcartii. Therefore, the results would be quite helpful in the management of this species in its natural habitat and conservation of overall biological diversity in the region.

CONCLUSIONS
We present an ecological niche model of Aristolochia cathcartii Hook.f. & Thomson, a potential medicinal plant found in some forest pockets of Assam's KarbiAnlong district. We were able to create a distributional map of A. cathcartii using our modelling approach. The areas identified in this study for reintroduction would aid in improving the conservation status of the A. cathcartii species population. As a result, the findings would be extremely useful in the management of this species in its natural habitat as well as the conservation of the region's overall biological diversity.

www.threatenedtaxa.org
The Journal of Threatened Taxa (JoTT) is dedicated to building evidence for conservation globally by publishing peer-reviewed articles online every month at a reasonably rapid rate at www.threatenedtaxa.org. All articles published in JoTT are registered under Creative Commons Attribution 4.0 International License unless otherwise mentioned. JoTT allows allows unrestricted use, reproduction, and distribution of articles in any medium by providing adequate credit to the author(s) and the source of publication.