Banko Janakari, Vol 27 No. 2, 2017

12

Nepal has an appreciative history on conservation of rhino with remarkable proportions 
of Asian rhino population protected and managed within the protected areas. Poaching 
risk map promises early warning and a way to target preventive action, which can 
safeguard both human and ecosystem. This study was designed to identify and map 
risk areas of rhino (Rhinoceros unicornis) poaching within and around eastern sector 
of Chitwan National Park (CNP). A multi criterion GIS method was used to analyze and 
derive the risky zone. A binary logistic regression and expert consultation were done 
to finalize variables and risk rating, then weighted sum index model was performed 
in ArcGIS to derive risk zonation map. Presence/poaching and pseudo absence data 
were dependent variables and distance to guard post, settlement and road network 
from poaching events, land cover, slope and elevation were predictor variables for 
logistic regression model. Poaching events were observed to be spatially distributed 
around the park except in the south part. Among the seven predictable variables, five 
variables except terrain (slope and elevation) were statistically significant at 10% level 
of test (p<0.1). The poaching risk map indicates that areas near to roads, far from 
the guard post, and densely populated area of grasslands are high risk zone areas 
for rhino poaching. The GIS based maps will be practical and strategical to wildlife 
managers in CNP to facilitate decision making on intervention programmes and how 
best to direct law enforcement patrol within and around the park.

Key words: Pseudo absence data, Rhinoceros unicornis, weighted sum index model

Identification and mapping of risk areas of rhino 
poaching; a geospatial approach: A case study from 

eastern sector of Chitwan National Park, Nepal
M. Subedi1* and R. Subedi2 

Among the natural resource management 
programme in Nepal wildlife conservation 

has been steadily budgeted since the late 1970s 
(Poudyal et al., 2012). Primarily this is true for 
greater one-horned Indian rhinoceros, which is 
protected within the protected areas in the low-
lying Terai region (Poudyal and Knowler, 2005). 
Chitawan National Park (CNP) provides prime 
habitat for the most of the rhinoceros in Nepal, 
and the preservation of rhinos in this park is 
impressive conservation success stories (Poudyal, 
2005). Due to the high value on illegal markets, 
endangered species are often targeted species of 
poachers.

The Greater one horned rhinoceros, also known as 
Indian rhinoceros is an odd-toed ungulates of the 
family Rhinocerotidae. The greater one-horned 

rhinoceros is the largest of three types of Asian 
rhinos and, together with African white rhino, is 
the largest of all rhino species (WWF, 2017). It 
is listed in appendix I of CITES (Bhattarai and 
Rupakheti, 2015), globally vulnerable (IUCN, 
2017) and nationally endangered (Jnawali et al., 
2011) and protected by National Park and Wildlife 
Conservation Act 1973. Rhino population has 
been increasing in Nepal for few years (WWW, 
2017).

Poaching and illegal trade in endangered species 
and products made from them are considered 
foremost serious problem in biodiversity 
conservation, hence poaching is the biggest 
challenge in the biodiversity conservation 
(Aryal, 2002). According to the NBSAP (2014) 
illegal hunting and trade of important wildlife 

1 Department of forest research and survey, Kathmandu, Nepal, * Email: subedi.iof68@gmail.com
2 Institute of forestry, Central Campus, Pokhara, Nepal



Banko Janakari, Vol 27 No. 2, 2017

13

species is a major threat in the management of 
protected area’s biodiversity which has affected 
more severely to those vertebrates whose demand 
of products in international market is very high. 
Poaching is one of most challenge faced by the 
management in CNP (Acharya, 2006). As long as 
there is demand for oriental medicine-prepared 
from wildlife products like rhinoceros’s horns, 
there is always risk of poaching (Aryal et al., 
2009). Despite the success in preserving the 
rhinos in Nepal, substantial number have been 
poached within and outside these protected areas 
since the establishment of national park.

According to Treves et al. (2011), identification 
of the spatial distribution (e.g., extent, location) 
of poaching activities is utmost for managers to 
mobilize limited resources appropriately to the 
concentrated areas where poaching severity is 
high. Monitoring, enforcement and deterrence are 
difficult for poaching due to its illegal nature. To 
control poaching significant human and financial 
resources are needed (Keane et al., 2008) and 
manager most prioritize the monitoring and 
assessment activities relative to the other natural 
resources based on economic analysis (Sheil, 
2001). Thus, it is necessary to investigate the 
relationship between accessibility, habitat and 
control factors with poaching events to identify 
high poaching risk zone.

During the last three decades, Remote Sensing 
(RS) and Geographical Information System (GIS) 
technologies are emerging as new tools assisting 
in resolving land use conflict and management of 
natural resources (Brown et al., 1994) and also have 
made significant contributions in the management 
of natural resource and also for environmental 
monitoring (Zaman, 2012). GIS technology is 
a powerful tool for managing, analyzing, and 
visualizing wildlife data to intended areas where 
interventional management practices are required 
to monitor their effectiveness (ESRI, 2010). This 
study was carried out with an integrated approach 
using RS and GIS techniques together with 
ancillary data for poaching risk mapping of rhino.

Population density and catch per unit effort of 
poachers is inversely proportionate which means 
if rhino population density increases, the effort 
required to find a rhino to poach will decreases 
(Metzger et al., 2010). In the study area, 
population density of rhinoceros is very high in 

comparison to other protected areas of Nepal 
which has created more favorable environment 
for poachers. Regular monitoring of population 
is therefore essential to guide protection efforts 
and management decisions (Subedi et al., 2013). 
Mapping can be used to prioritize conservation 
efforts and to minimize wildlife poaching risk 
which is helpful to manage wildlife and to 
delineate the most vulnerable area of poaching 
risk for a specific species (Sanches et al., 2008). 
This study had two main objectives: (i) to establish 
the relationship between accessibility habitat and 
control factors with spatial and temporal pattern 
of poaching events of rhino and (ii) to derive a 
poaching risk map. The results of this study will 
be useful for concerned stakeholders to conserve 
rhino in and around the CNP.

Materials and methods

Study area

The study was confined to the eastern sector 
of world heritage listed Chitwan National Park 
(Fig. 1). Geographically, the study area extends 
from 27° 25’ 30’’ N to 27° 39’ 30’’ N latitude and 
from 84° 23’ 30’’ E to 84° 45’ 03’’ E longitude. 
The eastern sector covers around 407 sq. km 
of core area and 138 sq. km of buffer zone 
and spreads into the Parsa, Makawanpur and 
Chitwan districts. The eastern sector consists of 
very good habitat for rhinoceros and has good 
population density. Since the last 12 years 30 % 
of total poaching incidents were found in this 
sector (CNP, 2015).

Fig 1: Map showing the location of study area

Subedi and Subedi



Banko Janakari, Vol 27 No. 2, 2017

14

Methodology

The overall methodology used in this study is given in figure 2.

Fig. 2: Overall methodology for the risk mapping

Data use

The Rapid Eye image dated February 2010, having 
5 m spatial resolution of the study area was used for 
the land use land cover map preparation. ASTER 
Digital Elevation Model (DEM) of the 30-m 
spatial resolution dated October 2011 was also 
downloaded from USGS (https://earthexplorer.
usgs.gov) and used for the slope and elevation 
map preparation. Digital topographic maps of 
the study area were purchased from Department 
of Survey, Kathmandu Nepal and used for the 
road network and settlement map preparation. 
The location of the guard/security post, poaching 
and population information, habitat distribution 
from 2003 to 2015 were used and the data were 
supplied by the CNP.

Data analysis

Generation of spatial layer and land cover map

Spatial layers of Land Use Lan Cover (LULC), 
settlement, slope, elevation, guard post, road, 

habitat distribution and poaching events of 
rhino were generated using ArcGIS. Object 
based image analysis (OBIA) techniques were 
used for the land cover map using Classification 
and Regression Tree (CART) approach (Fig. 3) 
and eCognition Developer version 8 was used 
to produce the LULC map. First of all, images 
were divided into object segments and then using 
the test sample, image objects were classified 
into samples. Finally, the tuning parameters of 
different classifiers were adjusted to generate 
high classification accuracy. Altogether, 206 
sample points were used for the training sample 
and 60 sample points were used for the accuracy 
assessment. The mean values, standard deviation, 
brightness, max. diff. (max. intensity difference), 
NDVI (Normalized Difference Vegetation Index) 
and NDWI (Normalized Difference Water Index) 
were also chosen for classifications.

Subedi and Subedi



Banko Janakari, Vol 27 No. 2, 2017

15

Fig. 3: Flow chart of image classification using 
CART Pre-processing of poaching events data

Poaching data and poaching absence data were 
combined and processed in logistic regression 
analysis. Poaching data were given name to 
presence data and poaching absence data were 
generated from the study area where poaching 
events were absent assuming there is no poaching 
and gave name to pseudo absence data. Fifty 
pseudo absence data were generated using QGIS 
and merged with presence data and got the 
presence- absence dataset which were later used 
as a dependent variable in regression analysis. 
Presence data were labelled 1 and pseudo absence 
data were labelled 0 during the logistic regression 
analysis.

Logistic regression model

Logistic regression is one of the Generalized 
Linear Model (GLM) which is distinguish with 
other statistical model since it is not influenced 
by the supposition of variance inequalities across 
the groups, and is appropriate to use whenever the 
dependent variable is binary in nature (Hosmer 
et al., 2000). In the logistic regression model 
presence absence data as a dependent variable and 
distance to road network, distance to settlement, 
distance to guard post, land cover and slope as 
predictor variables, were used to explain the 
relationship with poaching events.

Distance to road, settlement, guard post as 
descriptive variables were determined from 
presence absence dataset using the near distance 
function in ArcGIS. The attributes and raster 
value of LULC map, slope and elevation were 
analyzed and extracted using ArcGIS and used 
in regression model. Statistical Package for 
Social Science (SPSS) were used for the logistic 
regression analysis.

The poaching risk mapping

After establishing the relationship between the 
bio-physical factors and poaching events, the 
results were discussed with the concerned expert 
and their weightage was fixed. Only significant 

variables in logistic regression analysis were 
discussed with expert. This method was done 
because poaching spatial factor and field 
scenario may be different so only statistical test 
is not sufficient. After finalizing criteria all maps 
according to the risk rating were prepared and 
finally used to derive poaching risk zonation 
map of the study area by using multi-criterion 
weightage sum index modelling in raster GIS 
environment.

Results and discussion

Relationship of poaching events with variables

The logistic regression model perfectly predicted 
34 pseudo absence (non-poaching) events out 
of 50 sample points and 25 presence (poaching) 
events out of 29 poaching events. 85.5 % overall 
accuracy was provided by the full model.

To explain the rhino poaching within and around 
the eastern sector of CNP, six predicted variables 
were examined for their significance using 
stepwise logistic regression. Out of six predictor 
variables, only four variables such as distance to 
guard post (p=0.030), distance to road (p=0.010), 
distance from settlement (p=0.021) and land 
cover (p= 0.074) were remained in the model 
with significant negative relationship of distance 
to road network and distance from settlement and 
positive relationship of distance to guard post and 
land cover with poaching events (p<0.1). Logistic 
regression analysis table is given annex 1.

In the scenario of natural resource management, 
roads make easier to people’s movement in 
formerly unreachable areas. If the area is easier 
to reach; then poacher can go in the area at short 
time for poaching (Toxopeus,1996). Similarly, 
in this study most of the poaching events were 
found to be occurred in areas within one to two 
kilometers.

Security guard post plays the most important 
role in bio-monitoring of illegal activities. The 
wildlife conservation history showed that more 
the guard posts less the poaching. This study 
showed negative relationship between security 
guard posts and illegal activities, as the increase in 
distance from guard post enhanced the likelihood 
of poaching.

Subedi and Subedi



Banko Janakari, Vol 27 No. 2, 2017

16

Some local inhabitant adjoining to the protected 
areas, illegal hunting comprises the part of 
their livelihood. Ouko (2013) found the direct 
relationship between the incident of poaching and 
level of income of local people. In this study, the 
relationship between distance from the settlement 
and poaching incident was found negative which 
means nearer the area from settlement, higher the 
risk of poaching.

Rhinoceros preferred the alluvial plain grasslands 
and swampy area. They have definite spots for 
dropping their excreta (Thakur et al., 2014). 
Grasslands and water bodies are the potential 
area for rhino poaching. Most of the poaching 
incidents were in grassland area, it could be 
because of habitat limitation in other areas.

Spatial distribution of poaching events

The results showed that during a period of 12-years 
from 2003 to 2015, 72 % of the poaching events 
occurred in the grasslands, 7 % in forestlands, 
14 % in water bodies, 4 % in cultivated area and 
remaining 4 % were in the sandy/river cutting 
area (Fig. 4). About 35 % poaching events were 
found at a distance of 3000 m to 4500 m away 
from the guard post and very less at a distance of 
less than 1000 m (Fig. 5).

Fig. 4: Distribution of poaching events with 
LULC

Fig. 5: Poaching incident and distance from 
guard post

Most of the poaching events were occurred at a 
distance of less than 1000 m from the road (Fig. 
6) and 2 – 4 km away from the settlement. No 
poaching events were found at a distance of 3000 
m away from road and 8 km away from settlement 
(Fig.7)

Fig. 6: Poaching events with distance from 
road

Fig. 7: Poaching events with distance from 
settlement

Subedi and Subedi



Banko Janakari, Vol 27 No. 2, 2017

17

Finalized criteria for risk mapping

The variables were classified into five classes. The 
very high-risk area, high risk area, medium risk 
area, low risk area and very low risk area were 
rated as 5 to 1, respectively. Finalized variables 
and their weightage is given in annex 2.

Risk zonation map

Figure 8 depicts the risk zone areas of rhino 
poaching within and around eastern sector of the 
CNP. The high-risk zone areas for rhino poaching 
are those areas close to roads, far from guard post 
and higher populated area of grasslands with red 
to green tones. This map is a result of combining 
the guard post distance risk map, road distance 
risk map, settlement distance risk map and land 
cover risk map.

Fig. 8: Rhino poaching risk zonation map

About 34. 6 km2 area was found as very higher 
risk area, 106.23 km2 as high risk, 132.92 km2 as 
medium, 161.88 km2 as low and 109.66 km2 as 
very low risk area of poaching (Fig. 9).

Fig. 9: Area covered by different risk zone

Conclusion

The findings from this study indicate that the 
multi-criterion GIS based weightage sum index 
model presented in this research identified and 
mapped risk areas of rhino poaching within 
and around eastern sector of the CNP. Poaching 
events occurred all, except in the southern part 
of eastern sector of CNP. Frequency of poaching 
was concentrated more in western site of the 
sector. The average rhino poaching location was 
found to be 792 m away from road, 3106 m away 
from settlement and 3424 m from guard post. 
Grasslands and water bodies were more likely to 
exhibit poaching events. The increase in distance 
from the guard post increased the likelihood of 
poaching, but the increase in distance from road 
and settlement reduced the likelihood of poaching. 
Most poachers avoid long distance travelling 
inside the park boundary. Nearby village of the 
CNP has also created the proxy environment for 
the poaching. About 45 % incidences were found 
at the distance between 4—6 km from settlements.

Hence for the effective control of rhinoceros 
poaching, more security guard posts should be 
established and the area of responsibility (AoR) 
of existing guard posts should be increased 
(maximized). This study recommends to prepare 
the similar risk map to other sectors of the CNP 
taking into account social factor in addition to 
the habitat, control and accessibility factors and 
the park should be arranged the anti-poaching 
activities according to risk zonation map.

Acknowledgements

This paper is based on a part of the author’s 
B.Sc. Forestry thesis submitted to the Institute of 
Forestry, Pokhara Campus, Nepal. We are obliged 
to the Project Coordination Unit, National Trust 
for Nature Conservation (NTNC) for financial 
support to conduct the study. We are deeply 
indebted to whole CNP family for providing the 
important data, their support and guidance during 
field work. We are grateful to the Department 
of National Parks and Wildlife Conservation 
(DNPWC) for allowing permission to conduct 
the research in the study area. We express our 
cordial thanks to Mr. Binod Prasad Heyojoo, Mr. 
Yajna Prasad Timalsina, Mr. Navin Kumar Yadav, 
Kamal Jung Kuwar, Amul Kumar Acharya, Shiva 
Khanal and Buddi Sagar Poudel for their critical 

Subedi and Subedi



Banko Janakari, Vol 27 No. 2, 2017

18

suggestions and guidance during the study.

References

Acharya, D. 2006. A Report about Rhino 
Poaching in Chitwan National Park, 
Nepal. Media Consultancy, Nepal.

Aryal, R. S., Box, P. O., Lane, N. P. 2009. Report 
on the Facts and Issues on Poaching of 
Mega Species and Illegal Trade in Their 
Parts in Nepal. Transparency International 
Nepal. Review Literature and Arts of The 
Americas 1—61.

Aryal, R.S. 2002. Wildlife trade in Nepal. 
Environment 7 (8): 1—5.

Bhattrai, G. and Rupakheti, N. 2014. Flora and 
Fauna of Nepal in CITES Annexes. 
Kathmandu: Department of National Park 
and Wildlife Conservation. Retrieved 
from https://drive.google.com/file/d/0B_
AvMj98dT2hWDQ4VVNFbXJHUmc/
view?usp=sharing

Brown, S., Schreier, H., William, A., Thompson 
and Vertinsky, I.1994. Linking Multiple 
Accounts with GIS as Decision Support 
System to Resolve Forestry/Wildlife 
Conflicts. Journal of Environmental 
Management 42: 349—364.

CNP. 2015. Chitwan National Park (CNP). 
Poaching Incident Trend. Unpublished 
manuscript.

ESRI. 2010. Environmental Systems Research 
Institute :GIS Based Practices in Wildlife 
Management. Retrived from: https://
www.esri.com/library/bestpractices/
wildlife-management.pdf

Hosmer, D.W., Lemeshow, S. and Sturdivant, R. 
X. 2013. Applied Logistic Regression. 
3rd edition. John Wiley and Sons, New 
York, USA.

IUCN. 2017. The IUCN Red List of Threatened 
Species. Version 2017-1. <www.
iucnredlist.org>. Downloaded on 08 July 
2017.

Jnawali, S.R., Baral, H.S., Lee, S., Acharya, K.P., 
Upadhyay, G.P., Pandey, M., Shrestha, 
R., Joshi, D., Laminchhane, B.R., Griiths, 
J., Khatiwada, A. P., Subedi, N., and 
Amin, R. 2011. The Status of Nepal 
Mammals: The National Red List 
Series, Department of National Parks 
and Wildlife Conservation Kathmandu, 
Nepal.

Keane, A., Jones, J. P. G., Edwards-Jones, G. 
and Milner-Gulland, E. J. 2008. The 
sleeping policeman: understanding 
issues of enforcement and compliance 
in conservation. Animal Conservation 
11: 75–82. doi:10.1111/j.1469—
1795.2008.00170.x

Metzger, K. L., Sinclair, A. R. E., Hilborn, R., 
and Mduma, S. A. R. 2010. Evaluating the 
protection of wildlife in parks : the case of 
African buffalo in Serengeti. Journal of 
Biodiversity Conservation 19:3431—3444. 
Retrieved from http://doi.org/10.1007/
s10531-010-9904-z

NBSAP. 2014 . National Biodiversity Strategy 
and Action Plan. Ministry of Forests and 
Soil Conservation, Kathmandu, Nepal.

Ouko, E. O. 2013. Where , When and Why are 
there Elephant Poaching Hotspots in 
Kenya ?

M.Sc. Thesis, University of Twente, ITC, 
Netherlands.

Poudyal, M. 2005. A Study of the Reasons for an 
Increase in Poaching of the One-horned 
Indian Rhinoceros in Royal Chitwan. 
M.Sc. Thesis, University of Simon Fraser, 
Burnaby, BC, Canada.

Poudyal, M. and Knowler, D. 2005. Economic 
incentives and poaching of the one-
horned Indian rhinoceros in Nepal. PREM 
Working paper 05/07 . Retrieved from 
http://www.rhinoresourcecenter.com/pdf_
files/117/1175857241.pdf

Poudyal, M., Rothley, K., and Knowler, D. 
2012. Ecological and economic analysis 
of poaching of the greater rhinoceros ( 

Subedi and Subedi



Banko Janakari, Vol 27 No. 2, 2017

19

Rhinoceros unicornis ) in Nepal. Journal 
of Ecological Applications 19 (7):1693—
1707.

Sánchez-Mercado, A., Ferrer-Paris, J. R., Yerena, 
E., García-Rangel, S. and Rodríguez-Clark, 
K.M. 2008. Factor affecting poaching risk 
to vulnerable Andean bears Tremarctos 
ornatus in the Cordillera de Mérida, 
Venezuela: space, parks and people. 
Journal of Flora and Faunna Oryx 42 (3): 
437—447. DOI: https://doi.org/10.1017/
S0030605308006996

Sheil, D. 2001. Conservation and Biodiversity 
Monitoring in the Tropics: Realities, 
Priorities, and Distractions. Conservation 
Biology 15 (4): 1179—1182. doi:10.1046/
j.1523-1739.2001.0150041179.x

Subedi, N., Jnawali, S., Dhakal, M., Pradhan, 
N., Lamichhane, B., Malla, S., and Jhala, 
Y. 2013. Population status, structure and 
distribution of the greater one-horned 
rhinoceros Rhinoceros unicornis in 
Nepal. Fauna & Flora International, 
Oryx 47 (3): 352—360 doi:10.1017/
S0030605313000562.

Thakur, S., Upreti, C. R., and Jha, K. 2014. Nutrient 
analysis of grass species consumed by 
greater one-horned Rhinoceros (Rhinoceros 
unicornis) in Chitwan National Park, Nepal. 
International Journal of Applied Sciences 
and Biotechnology 2 (4): 402—408.http://
doi.org/10.3126/ijasbt.v2i4.11119

Toxopeus, A. G. 1996. Ism, an Interactive Spatial 
and Temporal Modelling System as a 
Tool in Ecosystem Management : with 
Two Case Studies: Cibodas Biosphere 
Reserve, West Java, Indonesia, Amboseli 
Biosphere Reserve, Kajiado District, 
Central-Southern Kenya. ITC, United 
States.

Treves,A., Kerry A. Martin, Adrian P. Wydeven, 
Jane E. Wiedenhoeft; 2011. Forecasting 
Environmental Hazards and the Application 
of Risk Maps to Predator Attacks on 
Livestock, BioScience; 61 (6) : 451—458. 
https://doi.org/10.1525/bio.2011.61.6.7

WWF. 2017. Greater One horned Rhino. 
Retrieved from http://wwf.panda.org/what_we_
do/endangered_species/rhinoceros/asian_rhinos/
indian_rhinoceros/ on 7/8/2017.

Zaman, M. A. 2012. “GIS and Remote Sensing 
Applications in Natural Resources Management 
in Bangladesh” Paper presented at the World 
Conference on Computers in agriculture, 
Federation for Information in Agriculture, Taipei, 
Taiwan.

Subedi and Subedi



Banko Janakari, Vol 27 No. 2, 2017

20

Annex 1: Logistic regression analysis table

Variable β df Sig. Exp (B)
Distance to guard post .068 1 .030* 1.001
Distance from settlement -.001 1 .021* .999
Distance to road -.053 1 .010* .997
Land cover 3.054 1 .074** 21.195
Slope 2.984 1 .598 7.25
Elevation 6.34 1 .895 15.43
Constant -2.882 1 .293 .056
* significant at 0.05 and 
** significant at 0.10

Annex 2: Variables in forest poaching risk area modeling, their ratings and poaching Occurrence 
and β coefficient in regression analysis

Variable Class Poaching
Occurrence

Risk
Rating β Statistically 

significant

Land cover

Forestland 3 3

3.054
(p=0.074) 
Significant at α=10%

Grassland 21 5
Water bodies 3 4
Sandy areas 1 2
Cultivated area 1 1

Distance from 
guard post (m)

<1000 1 1 0.680

(P=0.030) 
Significant at α=5%

1000 — 2000 5 2

2000 — 3000 6 3

3000 — 4500 10 5

>4500 7 4

Distance from 
road (m)

<1000 19 5

-0.053 (P=0.010) 
Significant at α=5%

1000— 2000 8 4

2000—3000 2 3

3000— 4500 0 2

>4500 0 1

Distance from
Settlement (m)

<2000 8 2

-0.001 (P=0.021) 
Significant at α=5%

2000— 4000 13 5

4000—6000 6 4

6000— 8000 2 3

>8000 0 1

Subedi and Subedi