Trends, patterns and determinants of biodiversity conservation outcomes in Buxa Tiger Reserve, West Bengal, India

ABSTRACT This paper analyses the trends, patterns and determinants of biodiversity conservation in the Buxa Tiger Reserve (BTR), India. Temporal remote sensing data from 1990 to 2020 shows a loss of 27.69 km2. The results show that the total forest area has seen a significant decline from 1990 to 2020, whereas non-forest and degraded forest areas have been on the rise. The decline of forest area is observed more in fringe and lower altitude areas where accessibility is easy for humans to extract forest resources. The secondary data shows a decline in the wildlife population including the flagship species, the Bengal tiger. The decline in natural resources due to human activities in the BTR is likely to continue unless a participatory biodiversity conservation programme is established. The establishment of the Joint Forest Management Committee (JFMC), a formal local institution where local communities and forest departments (FDs) jointly share the rights and responsibility towards the use and management of the forest, seems to be effective in reducing negative forest activities. The sustainability of the BTR is possible if the local people abandon the ‘tragedy of the commons’ activities and work together, with government guidance for the promotion of livelihoods and biodiversity conservation.


Introduction
Conservation of biological diversity is considered critical for human well-being as these biological resources provide a variety of ecosystem services such as food, fibre, fuel, medicines, and other forms of nourishment. The Millennium Ecosystem Assessment Report classified these ecosystem services into four categories: provisioning, supporting, regulating and cultural services (MEA 2005). However, it is extensively documented that biodiversity is seriously threatened due to climate change, global warming, overexploitation, invasive species, pollution and land use change (MEA 2005;Mazor et al. 2018;Caro et al. 2022). The increasing loss of biodiversity has largely impacted human well-being, particularly that of the people who rely on ecosystem services for their survival (Díaz et al. 2006).
The increasing biodiversity loss and its adverse consequences for human society have made several countries around the world realize the importance of conserving biodiversity for sustainable development. This has influenced India to create protected areas (PAs) in order to preserve flora and fauna including threatened and endangered species (Naughton-Treves, Holland, and Brandon 2005;Soliku and Schraml 2018). But the establishment of PAs is not free from challenges, as large numbers of people living in and around PAs depend on the forest for their survival. Reconciling the needs of the local people, who rely on the forest for their livelihood, and the desire to conserve biodiversity is complex and this region has struggled to find effective management strategies.
Interestingly, although the extent and coverage of PAs have increased over the years (Naughton-Treves, Holland, and Brandon 2005), their effectiveness in conserving biodiversity is highly controversial and contested in academic research. Some researchers argue that PAs have been found to be effective in protecting biodiversity (Naughton-Treves, Holland, and Brandon 2005); while others have raised serious concerns on this issue (Guadilla-Sáez et al. 2019). Guadilla-Sáez et al. (2019) did not find any significant difference in species composition between PAs and non-protected forest commons. Leberger et al. (2020) observed that PAs having the highest level of protection experienced less forest loss than those PAs that permit more human intervention; however, the opposite finding was observed in other regions. Their findings also confirmed that deforestation is accelerating in PAs across all International Union for Conservation of Nature (IUCN) categories. Bruner et al. (2001) observed that many parks are effective at stopping land clearing and to a lesser extent successful at reducing logging, hunting and grazing in tropical countries. However, these authors argued that the effectiveness of conservation of a park is dependent on basic management activities like enforcement, direct compensation to local communities and boundary demarcation. Similarly, Struhsaker, Struhsaker, and Siex (2005) observed that the conservation of biodiversity in Africa's rainforests is effective in PAs. However, it should be noted that rising population and the influx of migrants are found to threaten biodiversity in the PAs (Struhsaker, Struhsaker, and Siex 2005). According to Rahman and Islam (2021), while overall PAs are helping to reduce deforestation in four PAs in Bangladesh, the long-term sustainability of forests in the PAs remains uncertain.
In India, higher dependency on forest resources, larger numbers of human-wildlife conflicts, habitat loss and fragmentation, poaching and illegal trade of wild animals, and conflicts between the forest department (FD) and local people are frequently observed in PAs (Ghosh-Harihar et al. 2019).
The goal of this study is to understand whether the creation of a PA has been successful in conserving the biodiversity in Buxa Tiger Reserve (henceforth BTR) located in the Indian state of West Bengal and to identify and analyse the factors that affect the conservation outcomes. Nad, Roy, and Roy (2022) studied the changes in land use and land cover (LULC) using a remote sensing and Geographic Information System (GIS) approach in the areas adjoining the BTR. Sam (2022) showed the effects of natural and anthropogenic disturbances on forest health in the BTR but failed to analyse temporal changes in LULC. Both of the previous studies on the BTR have lacked data regarding the wildlife populations and determinants of conservation outcome. To fill these knowledge gaps in the existing literature, this study aims to examine the temporal change of forest cover in the BTR, analyse the temporal change in specific wildlife populations in the BTR, and identify and analyse the factors affecting conservation outcomes.

Evolution of forest policies in the BTR
During the British rule in India, homogeneous Sal and Teak plantation activities were undertaken in the Buxa forest division. Plantations established in this time transformed the natural vegetation. The Buxa Timber and Trading Company acted both as a planter of Sal and a collector of other plant timber to cater to the increased demand for timber products in local markets in the northern part of Bengal to produce tea packing boxes (Ghosh and Ghosal 2019). Large-scale harvesting of timber from the BTR forest was also carried out to support domestic and European markets for the production of railway slippers, house construction, furniture for the Royal British navy and other purposes. The British also initiated an agro-forestry system known as Taungya cultivation, where domestic crops are initially grown as tree plantations regrow. To support the Taungya cultivation new forest villages were created, bringing tribal families (Santhals, Madesi, Munda, Rabhas, Garos, and Bhutia) from the Madhya Pradesh, Jharkhand, and Chotonagpur plateau regions. These new workers were allowed to settle in the forest region to work the plantations. In return, the forest labourers had certain rights to use the forest services for their families (Das 2009;Ghosh and Ghosal 2019). The existing forest dwellers, like the Rabhas, were deprived of legal and formal rights to the forest. The legitimate right to collect dried wood was insufficient to support the forest dwellers so they cut trees illegally (Karlsson 2000) and allowed their livestock to graze there (Ghosh and Ghosal 2019).
Since 1991, participatory PA management has been implemented in the fringe villages bordering the BTR. Later, the Forest Protection Committee (FPC) and the Eco-development Committees (EDCs) were created in the BTR by the FD. These committees have been redesigned as Joint Forest Management Committees (JFMCs). According to data provided by the FD, currently, there are 63 JFMCs actively functioning in and around the BTR. However, conflicts continue to arise related to land use, resulting in the BTR facing 'tragedy of the commons' issues.
In West Bengal, the official implementation of the Forest Rights Act (FRA 2006) came into effect on 31 December 2007 to recognize the rights of the forest dwellers over their land (Ghosh and Ghosal 2019). But conflicts continue to arise since individual rights have been granted but community rights have not. In light of the institutional and policy complexity, it is highly imperative to examine the pattern and trends of biodiversity conservation in the BTR nature park.

Study site and its biological importance
The BTR is an important tiger reserve in the eastern Indian state of West Bengal, located in the Alipurduar district. It is between longitudes 89°20′ and 89°55′E and latitudes 26°30′ and 26°55′N, covering an area of approximately 760 km 2 , in which 390.58 km 2 constitutes a core area or critical tiger habitat and 370.28 km 2 is defined as a buffer zone. In 1983, Buxa was proclaimed the fifteenth 'tiger reserve' in the country. The BTR is comprised of two divisions (East and West), 14 territorial ranges and 48 territorial beats (TCP 2015-24) ( Figure 1). The term 'beat' is used to refer to the smallest unit of forest area managed by forestry staff officers under the government FD.
The fragile 'Terai ecosystem' is a part of the PAs within the Buxa Tiger reserve. The annual temperature fluctuates from 15 to 39°C, and the annual rainfall ranges from 3570 to 5600 mm. The BTR is split by several rivers including the Jayanti, Sankosh, Raidak, Phashkhawa, Churnia, Turturi, Nonani and Dima, which flow in a north-south direction across this PA and enrich the exquisite natural beauty of the park. The altitude varies across the region from 125 to 1750 m; this significant altitudinal gradient supports great amounts of biodiversity in flora and fauna. The elegance and beauty of nature in the BTR also draws the attention of people and make it an attractive place for tourists.
There are approximately 37 forest villages located within the reserve, 46 fringe villages and 34 tea gardens. In addition, there are four fixed demand holding villages (FDHV) 1 in the BTR. The BTR is occupied by many tribal people who belong to diverse ethnic groups. The major tribal communities are Rabha, Oraon, Garo, Santhal, Bhutia, Mechia, Neapli, Rajbanshi and Bengalis.

Remote sensing data
Remote sensing data is frequently used by researchers to show the temporal change of LULC (Islam et al. 2019(Islam et al. 2018. To show the decadal change in forest cover of the BTR, the latest high-resolution images were downloaded from the US Geological Survey Earth Explorer satellite (https://earthexplorer.usgs.gov/). Satellite imagery of two different time periods, 1990 and 2020, were used to estimate LULC. The spatial resolution of both Landsat 5 Thematic Mapper (TM) and Landsat 8 Operational Land Imager (OLI) and Thermal Infrared Sensor (TIRS) sensors was 30 m (Table 1). Images having cloud cover could significantly decrease the accuracy of the classification (Islam et al. 2018). The cloud cover was assessed in the metadata and it was determined that in India the mean total cloud cover varies according to season, with monsoon season presenting significantly more cloud cover than the winter season (Jaswal 2017). Therefore, all images used for analysis were from the winter season.

Image processing
ERDAS Imagine14 software was used for image processing and layer stacking. Three bands were converted into a single layer; the three bands are 5 (Near Infrared i.e. NIR), 4 (red), and 3 (green) for Landsat 8 and 4 (NIR), 3 (red), and 2 (green) for Landsat TM. This process was followed for both types of imagery. The study area was masked using the sub-set tool from the layer file, using the area of interest (AOI) layer. Then the sub-set image was re-projected to Universal Transverse Mercator (UTM) zone 45 N.

LULC classification
ERDAS Imagine14 and ArcGIS 10.5 v software were used to prepare an LULC map of the study area. A supervised classification method was adopted to classify the images. The selection of training sites is an important aspect of this method. Some pixels with similar pixel colour tones were assigned as training sites, which were made by drawing polygons. For each land use class, 60 polygons were drawn to train the dataset and saved as a signature file. These signature files were used during supervised classification to identify land cover features. The researchers have widely applied the maximum likelihood classification (MLC) method during supervised classification, which is reliable for land use change assessment (Islam et al. 2018). Based on the MLC method, a classified image was generated, and the same rules were followed for the next image. After classifying the image, it was recorded into the desired number of classes, and then the final LULC map of the BTR was prepared. A total of five land use classes have been identified over both image sets, i.e. forest, degraded forest, scrub land, water bodies, and non-forest.

Accuracy assessment
After classifying and recoding the images, an accuracy assessment was executed. For this study, 100 sample locations were chosen at random for each of the two years, and the samples were then cross-validated using Google Earth data. Then the points were matched, and for accuracy assessment, kappa statistics was generated. It should be noted that accuracy of more than 80% is commonly found in the literature and considered acceptable (Islam et al. 2018).

Calculation of year-wise percentile share of LULC
The percentile share of each land use category in the respective year was calculated by applying the following equation: where AELC = area of each land use category in a year, and TAP = total area of PA in that year.

Overall rate of LULC change
The map of overall change in land use was prepared using the following equation: where AELC 2020 = area of land use category in the year 2020, and AELC 1990 = area of land use category in the year 1990.

Data on wildlife population
Changes in major wildlife species have been shown using secondary data collected from the Annual report

Dependent variable
The health of forests is considered an important indicator of biodiversity conservation as forests house both flora and fauna. Forest conservation outcomes can be measured in terms of various indicators such as forest growth, change in wildlife populations, or travel time to collect forest resources (Dash and Behera 2015). However, conservation outcomes measured in terms of forest growth are widely used by researchers (Dash and  Behera 2015). Using the remote sensing and GIS data for 1990 and 2020, the forest growth for about 48 forest beats in the BTR was calculated. Interestingly, there is a variation in forest growth rate across the beats, which allows us to use it as a dependent variable.

Independent variables
The size of the population or community residing in and around the forested areas is important as they significantly influence the forest conservation outcome. People with heavy reliance on forests for their livelihood and energy requirements may cause greater degradation of forest resources in the absence of an appropriate institutional mechanism that can regulate the users' behaviour (Heltberg 2001). Bera, Saha, and Bhattacharjee (2020) stated that high population density is one of the principal factors leading to deforestation. On the contrary, Dash and Behera (2015) found that the size of the community positively affects the growth rate of forest cover. Therefore, the effect of size of human population on conservation outcome is ambiguous. Another important factor is the size of the livestock population, which directly affects forest growth because grazing practices and the demand for livestock feed put pressure on the forest ecosystem and hence hamper its growth. Dash and Behera (2015) found that when communities possess larger livestock population or greater wealth endowment, this leads to a lower growth rate of forest cover. Therefore, the size of the livestock population is expected to have a negative impact on the forest growth.
Effective functioning of local institutions promotes positive conservation outcomes (Dash and Behera 2015;Heltberg 2001). Institutions created by villagers out of their customs, beliefs and rituals are more likely to generate positive forest outcomes than formal institutions imposed from without (Dash and Behera 2015). Therefore, the availability of local institutions may have a positive impact on forest growth. In our study area, the JFMC is where local communities and FDs jointly share the rights and responsibility towards use and management of forest. Therefore, in the model estimation, we have considered the presence of JFMCs in the concerned area.
Location of the forest area is another crucial factor impacting conservation outcome. The beats located in the core region do not have access to markets, and thus are less likely to facilitate illegal tree felling and timber trading. Buffer area faces high anthropogenic pressure compared to core area in the BTR (Sam 2022). Therefore, the location of a beat is assumed to be negatively linked with forest growth.
The development of a road network has a negative impact on forest growth. Bera, Saha, and Bhattacharjee (2020) found a positive relationship between distance from road and deforestation. In the study area, an extensive road network has been constructed to ease transportation and communication, resulting in significant deforestation and the facilitation of illegal timber trading (TCP 2015-24). Hence, it is assumed that road length is expected to have a negative impact on forest growth.
Forest fire is another important factor that is expected to have a significant impact on forest growth. Every year, forest fire destroys many seeds and several plant species, and their regeneration is adversely affected in the BTR (TCP 2015-24). Hence, the number of forest fires is assumed to have a negative relationship with forest growth.
In addition, a dummy variable (D 1 ) on division of FD has been used to see the effect of division on forest area change. Table 2 describes the variables used in the econometric model, including their expected effects on forest conservation outcomes. Table 3 shows the summary statistics of variables used in the econometric model. As the dependent variable is continuous, an ordinary least squares (OLS) regression model has been employed to estimate the relationship between dependent and independent variables mentioned above. The OLS regression model estimation on the determinants of forest growth has the following functional form: Forest growth ¼α þ β 1 Population density þ β 2 Livestock density þ β 3 JFMC þ β 4 Beat location þ β 5 Division þ β 6 Road density þ β 7 Forest fire þ ε 1 (3)

Accuracy assessment
The kappa statistics were 0.88 and 0.81 for 2020 and 1990, respectively. The overall accuracy was 96% and 93% for 2020 and 1990, respectively. The high accuracy assessment statistics indicate the analysis is robust and accurate. Figure 2 presents the land use pattern in the BTR and the extent of its changes over the period from 1990 to 2020. In 1990, among the five LULC categories, the highest proportion of land (692.62 km 2 , 91.20%) was covered by forest, spreading from higher to lower altitudes in the region. The area of degraded forest was 3.78 km 2 (0.50%). The prevalence of scrub land area was 14.54 km 2 (1.91%). The area of water bodies was 2.58 km 2 (0.33%). The non-forest area covered about 45.99 km 2 (6.06%). According to the LULC map of 2020, the forest area was 664.93 km 2 (87.56%). Non-forest area covered 67.32 km 2 (8.86%). Degraded forest was spread over 11.80 km 2 (1.55%). Scrub land was found over an area of 12.16 km 2 (1.60%), and 3.23 km 2 (0.43%) of area was identified as water bodies (Table 4).

Land use and land cover
A comparison of the land use pattern data between 1990 and 2020 shows a significant decline in forest cover, which is estimated to be about 4% of the total forest area. Interestingly, the share of non-forest area has increased by more than 46% in 30 years inside the BTR. The extent of degraded forest land has increased significantly within the same time period (Figure 3). Therefore, the data show the overall degradation in forest cover inside the PA. It should be noted that the area under scrub land and water bodies in 2020 is more or less equal to the respective area in 1990.

Change in forest area
Out of total 48 beats, the forest area has decreased in 34 beats and increased in 14 beats. The greatest change in the percentage of total forest area is found in the Barnabari beat (−23.91%), followed by Mainabari (−23.23%). The highest increase in total forest area was recorded in Chnuiajhora beat (15.46%). The beats that are found to have only minor changes in forest area are mostly located in high-altitude and inaccessible regions, like Chuniajhora, Bhutri, Santrabari, etc. The beats that are located at lower altitudes and in close proximity to human habitat, like Barnabari, Mainabari, Rangamati, Sankosh, and Godamdabri, have witnessed higher rates of forest degradation ( Figure 4). It should be noted that the adverse consequences of degraded forest in the BTR are also reflected in the conservation of wildlife -especially the Bengal tiger, as discussed below. Subsequently, in the section of change of wildlife population below to understand the factors that are likely to influence the forest conservation outcomes using the beat-level data.

Change of wildlife population
The BTR was established primarily to conserve Bengal tiger, which is the flagship species of the PA. Hence, it was made an extreme priority to conserve this carnivorous species of the tropical ecosystem. Before the implementation of a participatory biodiversity conservation programme, the population of tiger used to fluctuate significantly (15-33 individuals) but with the PAs it was expected that the tiger population would stabilize. However, the population continues to fluctuate, suggesting   Non-forest includes settlement, flood plain, and agricultural land.  that the PAs are not supporting tiger conservation ( Figure 5 and Table S1). Interestingly, the number of leopard has sharply increased, from eight in 1984 to 149 in 2002 ( Figure  6 and Table S2). Like the carnivore population, the herbivore population has fluctuated within the BTR. For instance, the density 2 of barking deer and bison increased, while the density of spotted deer, hog deer, monkey, wild boar and sambar decreased significantly, in the BTR from 1994 to 2013. The density of barking deer was 1.45 in 1994, and reached 2.753 in 2013. During this period, the overall increase of barking deer was more than 89%. In contrast, the density of spotted deer decreased from 1.26 in 1994 to 1.026 in 2013. In 1994, the density of the bison population (an endangered species) was 0.59; however, strong protection measures led to a quick recovery, with the population up to 1.823 in 2013. Within approximately two decades, the population of bison increased by more than 200%. Unlike the bison population, the hog deer population was reduced by 72%, down to a density of 0.128 in 2013 from 0.47 in 1994. In 1994, the density of the monkey population was 20.13, the highest among the mentioned herbivore population. Although it still possessed the topmost position in 2013, there is cause for concern as the monkey population density has declined to 8.569, a decrease of 57%. A negative trend was observed in density of the wild boar population, which became 3.292 in 2013, down from 3.75 in 1994. Therefore, the BTR witnessed a ~12% reduction in the density of the wild boar population. Similar to the wild boar population, the sambar (a species of deer) population has also decreased at a significant rate. The data of 1994 indicated that the density of sambar population was 0.46. However, the 2013 data showed that its density had fallen to 0.28. Thus, the data show that the overall loss of the sambar population was >39% in the BTR (Figure 7 and Table S3). Table 5 reports the results of the estimation of the OLS regression model on factors affecting forest growth in the BTR. These variables were tested for multicollinearity, and the Variance Inflation Factor (VIF) and tolerance values were used to look for multicollinearity. Due to the fact that all VIF and tolerance values were <10 and >0.1, all independent variables taken into consideration by the study were free from multicollinearity problems. The overall model has an R 2 of 40.76% and the model is highly significant (F 7, 40 = 4.66, Prob > F = 0.007). Four out of seven explanatory variables used in the model turned out to be significantly related to forest growth rate: livestock density, JFMC availability, beat location, and number of forest fires.

Factors affecting conservation outcome
Forest density is negatively associated with human population density, which suggests that forest beats with a higher population density are likely to register less forest growth. However, the effect is not statistically significant. The variable livestock is found to be negatively and significantly (at the level of 5%) associated with the change in forest growth, which indicates that forest beats with larger livestock populations tend to have less forest growth. It is an empirically verified fact that livestock grazing in forest areas poses a significant threat to forest growth and leads to forest degradation (Wangchuk 2002;Dash and Behera 2015). Grazing of livestock is prevalent in the study area, which results in damage to both existing and newly planted trees. The effect of local participation (in a JFMC) is found to be positively related to change in forest growth and the effect is significant (at the 5% level), signifying that beats with JFMCs are more likely to be associated with an increase in forest area and vice versa, implying that the presence of local institutions tends to regulate forest use by villagers and increase the protection of forest areas, and hence may contribute towards growth of the forest. Location of the forest beat has turned out to be negatively and significantly (at the 5% level) linked with growth of forest area, which implies that beats located on the outskirts are less likely to be associated with an increase in forest area. This may be due to several facts. First, the villagers of these communities benefit from easy access to markets, increasing their interest in collecting fuelwood and other forest resources for income. Second, surrounding revenue villages and tea gardens also extract a significant amount of forest resources. As expected, the variable forest fire turned out to be negatively and significantly (at the 1% level) associated with forest growth, which suggests that an increase in the incidence of forest fire tends to reduce forest growth.

Discussion
The results of the study suggest that the forest area of the BTR has been consistently declining, due to several anthropogenic reasons. It is observed that the maximum forest loss was reported in the isolated portion situated southeast of the BTR, which is well connected by roads and transport. The density of forest and area covered with forest decreased in the northwest and southwest areas, where human activities are intensive, with well-developed communities. Except for the northern area, where accessibility is difficult due to high mountain cover, all other sides of the BTR boundary have been observed to be a gateway to deforestation by woodcutters and smugglers. The reason behind the higher degraded forest area in 1990 can be ascribed to massive illegal destruction of forest during the 1980s (Karlsson 2000). The flagship species of the park, the  Bengal tiger, has decreased over the study period and is now threatened. From the researchers' interactions with villagers and fuelwood collectors during field visits, there were no reported sightings of the Bengal tiger in the BTR. One tiger was, however, recently spotted by FD in the BTR.
The regression analysis has shown that livestock population, JFMC availability, beat location and forest fire are some of the key factors that are found to have influenced the forest area in the BTR. These findings have important policy implications. The effect of human settlements and the resulting anthropogenic activities like livestock grazing, easy accessibility to the BTR through an elaborate network of roads, and the lack of lucrative alternative economic livelihood activities, are the main drivers of forest degradation in the BTR. The fringe areas of the park are found to be more vulnerable to forest degradation compared to the core or inaccessible areas. Similar results are reported by Sam (2022). The local institution (JFMC) was set up to conserve forest in a more sustainable manner, and it appears to be working as the results of this study shows that JFMC involvement has a positive effect on forest area. This result is similar to other important, related empirical studies (Dash and Behera 2015;Heltberg 2001). This suggests that these institutions are fulfilling the objective for which they were established. Hence, adequate effort should be invested to make these vital local institutions actively functional, and required financial support should be made available. Forest fires, which are mostly man-made, have negatively affected forest growth in BTR. Hence, a more effective forest fire management strategy should also be implemented.
However, during our field visits, it was observed in few cases that JFMCs were largely dysfunctional because there is a lack of mutual trust between local people and the FD. Rising demand for cooking fuel by the settled human population in and around the BTR has also contributed significantly to the decline of forest growth. Provision of alternative cooking fuel sources such as cooking gas, solar cookers, and efficient smokeless cooking stoves would reduce the biomass pressure on the forest.

Conclusion
The Buxa Tiger Reserve is important for preserving biodiversity in India. This study has explored trends, patterns and determinants of biodiversity (flora and fauna) conservation in the BTR, by comparing data from 1990 and 2020. With the total forest area declining by 27.69 km 2 (−4.00%) over the past 30 years, there is cause for concern. Not surprisingly, the Bengal tiger, the flagship species of the park, has also decreased over the same period. To ensure conservation becomes a priority, participatory biodiversity conservation strategies that incorporate the local inhabitants and empower them to be responsible for sustainable management is essential. Further, the findings strongly suggest that the BTR management authority should expand the scope of management focus from core villages to other fringe villages that are located around the BTR. If a holistic and comprehensive eco-development programme was designed and implemented covering all the villages located in and around the BTR, this might produce a winwin situation for the local people and for the forest and wildlife in terms of achieving both livelihoods and conservation outcomes.

Notes
1. The FD was in need of a continuous supply of labourers for mining and other non-forestry activities. Because of the scarcity of labourers in the local area, FD issued passes to numerous labourers working in such companies in accordance with the yearly renewal. Every year the labourers with passes would pay tax to the forest department. This type of village is referred to as a fixed demand holding village (Sarkar and Das 2012

Data availability
Data can be obtained from the corresponding author on request.