Grassland-woodland transitions over decadal timescales in the Terai-Duar Savanna and Grasslands of the Indian subcontinent.

Grassland-woodland transitions over decadal timescales in the Terai-Duar Savanna and Grasslands of the Indian subcontinent. | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Grassland-woodland transitions over decadal timescales in the Terai-Duar Savanna and Grasslands of the Indian subcontinent. Subham Banerjee, Dhritiman Das, Robert John This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-1398899/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract The decline of grasslands and savanna due to woodland encroachment is a matter of global concern. In the Terai-Duar Savanna and Grassland ecoregion (hereafter terai), located at the base of the Himalayas in the Indian subcontinent, there are indications that grasslands and savanna are being lost, with unknown consequences for biodiversity and ecosystem function. We assessed large-scale vegetation changes in terai eight large protected terai habitats over three decades (1989-2019) and quantified the environmental and anthropogenic drivers of the observed changes using Bayesian Conditional Autoregressive spatial models. Notably, very little grasslands remain and the initial extent of 1417 km2 declined to 923 km2 (34.4%) in 30 years. Woodland area increased from 3235 km2 to 3516 km2 (8.7%). Dry season grass fire had the strongest influence on grassland persistence, followed by anthropogenic impacts. Terai ecosystems also experience significant threats from climatic changes and increasing human footprint, particularly in India. Grassland-Woodland Transitions Landcover Classification Habitat Loss Anthropogenic Fire Bayesian Model Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Despite their global significance as a reservoir of biodiversity, tropical grassy biomes (including grasslands and savanna) are being degraded and lost across many regions of the world 1–4 . Multiple factors, including woodland encroachment, conversion to croplands, the spread of invasive species, climatic changes, grazing by domestic livestock, altered fire regimes, and bio-resource exploitation, are known to be important drivers of grassland loss globally 2 . However, the decline of tropical grassland and savanna has not attracted the same level of conservation attention as the loss of tropical forests 2,5–7 . This is partly due to the misconception that tropical grassy biomes are anthropogenically derived through the degradation of forests and that forests hold greater ecological and economic value than grassland-dominated ecosystems 1,6 . Encroachment by woody plant species is a significant cause of the transformation of grassland and savanna globally 8–11 . The conversion of mixed tree-grass plant communities to tree-dominated woodland may profoundly influence the productivity, water balance, and nutrient cycling of these ecosystems. Woody cover appears to be determined by four major environmental variables - water availability, resource availability, fire, and herbivory – factors that vary over a wide range of climatic conditions 10,12 . In sites where low water availability alone does not limit tree growth, woody cover may still be maintained well below the climatic potential due to a complex interplay between multiple processes such as fire, herbivory, and resource competition 10,13,14 . When environmental changes modify these processes, grasslands and savanna may be transformed to woodland or may even be created by tree cover loss in woodland areas. Such transitions between grassland and woodland dominance influence plant and animal diversity, primary productivity, carbon and nutrient cycling, hydraulic conductivity, and below-ground processes 9,11,15–17 . In a global-scale empirical study 11 of 224 dryland sites located in all continents (except Antarctica), Soliveres et al. found that plant diversity and ecosystem multifunctionality peaked at intermediate levels of woody cover in these drylands, a relationship that became stronger in wetter sites. However, anthropogenic effects, including overgrazing, changing rainfall patterns, and increased atmospheric CO 2 concentrations, drive increases in woody species cover and abundance in the world’s grassland and savanna regions 9,10 . Woody cover is anticipated to increase further in future environmental change scenarios, with significant consequences for ecosystem structure and function 18 . Although woody encroachment is mainly due to hardy shrubs across arid to sub-humid ecosystems, wetter sites may be experiencing encroachment by trees with a corresponding loss of native grasslands 11 . Here we study the transitions between grass and woody species dominance over decadal timescales in the Terai-Duar Savanna and Grasslands ecoregion 19 (hereafter terai) in the Indian subcontinent. The terai is a unique assembly of alluvial grassland, savanna, and forest ecosystems, which occurs in a narrow belt extending east to west at the base of the Himalayas in the Indian subcontinent 20 . The tropical to subtropical climatic conditions (>1200 mm rainfall per annum) support high productivity and should favour tree dominance. However, there are extensive areas of grasslands and savanna, with varying densities of tree cover, ranging from treeless grassland to savanna (discontinuous tree cover) and dense forest. However, terai savanna and grassland habitats are experiencing woody encroachment and anthropogenic disturbance, threatening terai ecosystems 21,22 . Woody species density varies naturally over a wide range of values in mesic environments. For example, in a study 23 of sites across Africa, Australia and South America, Hirota et al. report the presence of three attractors – treeless grassland, savanna, and forest, with varying tree density. Savanna (with 5 to 60% tree cover) occurred over a wide range of mean annual precipitation, and the frequency distribution of tree cover across all sites was strikingly trimodal. The occurrence of multiple stable states implies that systems can undergo transitions in response to climatic changes and anthropogenic disturbance 23 . In the terai, all three states occur at large and small spatial scales, and any directional shifts towards grass or tree dominance may be driven by environmental changes or disturbance. To study the transitions between grass and woody species dominance in the terai ecoregion and to test the environmental drivers of the changes, we chose to investigate landcover changes in the Protected Areas (PAs) in the terai in which grasslands and savanna, and woodlands are known to occur. We found eight such sites - four each in India and Nepal - which together account for a total land area of nearly 5000 km 2 (Fig. 1 and Table 1). Although the terai ecoregion is much larger, native terai vegetation is mainly restricted to protected areas where human activities are not permitted by law. Using remotely sensed data, we analysed landcover and landcover changes observed over a 30-year period (1988-1989 through 2018-2019) to characterize the grassland-woodland transitions in these sites. We found that grassland areas declined substantially in the majority of the sites. However, woodland was always the dominant vegetation and accounted for greater area overall (>66%), and the persistence of grasslands (about 29% of the area) appears precarious in the face of encroachment by woodland and conversion to cropland by people. So, more than half the initial grassland area was lost in three sites and over a quarter in two sites (Fig. 2). Grassland area increased in three protected areas (18 to 34%), but of the overall grassland area of 1417 km 2 in 1988-1989 only about 42% remained as grassland in 2018-2019 (Fig. 2). This decline of grasslands was accompanied by increase in woodland area, and of the 826 km 2 of grassland that was lost, about 68% was converted to woodland. The remainder was accounted for by conversion to sand banks and open land, to water bodies (meaning inundated), and about 95 km 2 was converted to cropland. Our main goal was to test the influence of a series of environmental factors on the stability or persistence of grasslands and grassland-woodland transitions over the 30-year interval (1988-1989 to 2018-2019). Considering 500 m spatial resolution, we tested the influence of four classes of spatial environmental factors – fire, topography, habitat neighbourhood, and anthropogenic influence – on the persistence of grassland in each 500m cell. The conversion of grassland areas to woodland has been a matter of concern and in order to test whether any environmental changes drive this conversion, we carried out Bayesian Conditional Autoregressive spatial analyses, with the persistence (or not) of grasslands in each 500m cell over the 30-year interval as the response variable and environmental variables as predictors. We found that the frequency of fire occurrence over the 30-year interval was the most important predictor – greater fire frequency maintains grasslands. Regression models also show that grasslands were better preserved when they were at greater distance from woodland or cropland sites, which reflect the advancement of woodland at the grassland-woodland ecotone, and that many grassland sites have been converted to cropland by people in nearby settlements, respectively. Topography does significantly influence the persistence of grasslands in the face of change driven by woodland encroachment and other factors. Grasslands persisted better on sites with greater slope and in sites with lower potential for hydrological saturation (lower values of the Topographical Wetness Index), but topographic ruggedness did not affect it. We chose to restrict our study to protected areas in order to focus on the natural drivers of changes to grasslands. Yet we found that protected areas in this region are not free of anthropogenic threats that affect the habitat and vegetation. First, Himalayan and sub-Himalayan sites are predicted to experience greater rates of climate change than lowland sites. We found that all the protected areas that we studied have seen increases in mean temperatures (range 0.535 - 0.726°C) in the last five decades. In the last decade all sites have had some (up to 20%) dry months (<100mm rain per month) even during the monsoon season. A human footprint index, which is based on range of anthropogenic variables - human settlements, croplands, pasture lands, human population density, night-time lights, railway and road networks, and navigable waterways - increased by about 50% in some Indian protected areas and up to 18% in sites in Nepal during 1993 to 2009. Although we have not been able to link these changes to changes in vegetation, these environmental stressors are potential threats to terai ecosystems. These results provide the impetus for further research on the drivers of grassland-woodland transitions in the terai ecoregion. Results Landcover classification and transitions: Using supervised landcover classification for all eight sites for three time points (1988-1989, 2003-2004, and 2018-2019) in the 30-year period, we found that our classification or prediction accuracy was greater than 90% for all the twenty-four classifications (8 sites × 3 time points). Grassland and woodland habitats were classified with overall prediction accuracy of 92.2% and 94.3%, respectively. Woodland was the largest landcover type at all times in the terai, and it increased in extent during the last three decades. Including all the protected areas, the woodland cover was 3235 km 2 in 1988-1989, which increased to 3515 km 2 in 2018-2019 (Table 2). Grassland was the second-largest landcover across all the protected areas, but it declined steadily in extent. In 1988-1989, grasslands covered an area of 1417 km 2 across the eight protected areas, but declined to 1209 km 2 by 2003-2004, further reducing to 923 km 2 by 2018-2019 (Table 2 and Supplementary Table 1). However, these changes to grassland extent were not similar across all the PAs. Grassland areas declined in five sites, while there were smaller increases in three sites. The overall decline of grassland area by 34.8%, therefore, does not fully reveal the high dynamism in the individual sites. Thus, in the five sites where the grassland area declined, more than half the grassland area was lost in three sites, Shukla-Phanta NP (51.6%), Dudhwa NP (59.8%), and Chitwan NP (65.6%), and more than a quarter in two sites, Manas NP (40.2%) and Jaldapara NP (26.2%) (Supplementary Table 1). The increases in grassland area in three sites ranged from 18.2% to 34.5%, but two of these sites were smaller in absolute area (Supplementary Table 1). In the five sites where grassland areas declined, four showed increases in woodland area. However, the increase in a woodland area in each case was smaller than the decline in grassland area. The differences were mainly due to grassland conversion to cropland, but also due to inundation by riverine action (Supplementary Table 2). In Jaldapara NP, both the grassland area and the woodland area declined during this period. In Bardiya NP and Koshi Tappu WS, grassland areas increased during this period, and the increases were comparable to the declines in the woodland area. However, in Valmiki TR, which also showed increase in grassland area (33.4 km 2 ), the increase was much smaller than the decline in woodlands (-91 km 2 ). So, there are significant differences among sites in the transition from grassland to other land cover types. While Shukla-Phanta, Dudhwa, Chitwan, Jaldapara, and Manas NPs have seen sharp declines in grassland cover in the last three decades (Table 2), in Bardiya NP and Valmiki TR, the extent of grassland increased from 1988-1989 over the following fifteen years but declined in the last fifteen years of our study period (Supplementary Table 1). In Koshi-Tappu WS, grassland cover increased steadily throughout the study period (Supplementary Table 1). We found that only Dudhwa NP had cropland (15% of its area or 118.0 km 2 ) at the beginning of our study (Table 2), but by 2018-2019, all the four protected areas in India had some agricultural land within the park boundaries, accounting to a total of 213 km 2 (Table 2). None of the protected areas in Nepal have seen any conversion to cropland during the entire study period. Human settlements were present only in Dudhwa NP in India, which appears to have been founded during 2004-2005 but grew to cover an area of 32.7 km 2 by 2018-2019 (Fig. 3). Landcover classification maps of the eight protected areas in three periods are presented in figures (Fig. 3 and Fig. 4). Influence of anthropogenic and environmental factors on grassland persistence: First, we computed the proportion of grassland area in each 500m cell using landcover classification at 30-m spatial resolution. Using a nonspatial binomial regression model of the proportion of grassland area that persisted in each 500m cell (where 0 = no grassland remaining, and 1 = all grassland remaining) over the 30-year period, we found that fire frequency (FIRE_OCCURR) was the most decisive predictor of grassland persistence. Greater fire occurrence had a strong positive influence on the maintenance of grasslands (β = 0.638, p < 0.001) (Table 3). The proportion of grassland that remained was also significantly affected by general anthropogenic influence captured as a distance to human settlements, croplands, or the park boundary (DIST_ANTH), whichever was nearest. Greater distances led to lower persistence of grasslands, as evident in the significant negative binomial coefficient (β = - 0.256, p < 0.001) (Table 3). Being located further away from woodland sites (DIST_WOOD), however, helped maintain grasslands (β = 0.082, p=0.0006), indicating greater rates of transformation at the ecotone between these habitats (Table 3). Topography may influence fire and hydrology and we found that grasslands were more persistent at sites with higher slopes (SLOPE) (β = 0.0970, p=0.0025) and with lower potential for hydrological saturation, where the latter was detected as a significant negative influence of the topographical wetness index (TWI) (β = - 0.0860, p=0.0094) (Table 3). Another topographical feature, ruggedness (RUGG), had no significant influence on grassland persistence (β = - 0.020, p=0.4193) (Table 3). The Moran's I statistic for residuals of the nonspatial binomial regression was 0.622, and the Monte Carlo permutation test for the presence of spatial autocorrelation in the proportion of grassland conserved per cell was statistically significant ( H 0 : I = 0, p < 0.001). Therefore, we included spatial autocorrelation as a conditional autoregressive prior in a Bayesian spatial binomial regression model of the proportion of grassland persisting in each 500m cell with the same set of environmental predictors (but now spatial at 500m resolution) to compute the posterior estimated median values and the corresponding 95% credible intervals for the coefficients of all the environmental predictors (Table 4). We found that every single factor that was statistically significant in the non-spatial model was also significant in the Bayesian spatial regression. So, FIRE_OCCURR, SLOPE and DIST_WOOD had significant positive influence on the proportion of grassland conserved in a cell, and DIST_ANTH and TWI had a significant negative influence. The median values of the posterior distributions of the regression parameters show the positive and negative influence of the predictor variables, and statistical significance is indicated where the 95% credible intervals that do not include zero (Table 4). RUGG did not have a statistically significant influence on the proportion of grassland conserved in a cell. As in the non-spatial model, FIRE_OCCURR showed the highest influence (median value 0.6193) on the proportion of grassland conserved per cell (Table 4). Assessment of threats to the Protected Areas: Observed environmental changes already constitute a threat to the PAs in the terai ecoregion, but due to the lack of high-resolution spatial data we were unable to quantify their influence on grassland-woodland transitions. All the PAs have experienced an increase in mean temperature (range 0.535 - 0.726°C) over a span of five decades, with Bardiya NP recording the highest increase (Table 5). In the last decade (2011-2020), all the protected areas have experienced some dry months (<100 mm rain per month) even during the monsoon. Valmiki TR and Chitwan NP, which occur adjacently across the boundary between India and Nepal, have seen drought in 20% (8 out of 40) of the monsoonal months in the last decade (Table 5), while others have had fewer (2 to 6) dry months during the same decade. Examining other human impacts using the human footprint index - based the extent of applicable anthropogenic variables including human settlements, croplands, pasture lands, human population density, night-time lights, railway and road networks, and navigable waterways - we found that between 1993 to 2009, anthropogenic effects increased by more than 50% in Dudhwa and Manas NPs (Table 5). The smallest impact with just a 9% increase in human footprint was seen in Bardiya NP. Protected areas in Nepal experienced comparatively smaller increases (17.8% on average) in human footprint compared to the average 40.9% increase in human footprint in the Indian protected areas (Table 5). Discussion The decline of grassy biomes due to encroachment by woody species is a problem of global significance 24,25 . The trend towards increased shrub and tree abundance in grassland sites have been widely documented since decades - in North America 26,27 , South America 28 , Africa 29 , Australia 30 , India 31 , but most reports leading up to more recent studies are about shrub encroachment in dryland sites 16,17,32,33 . We studied grassland-woodland transitions in the unique Terai-Duar Savanna and Grassland ecoregion in the sub-Himalayan region of the Indian subcontinent, which lies in the sub-humid to humid zones but nevertheless hosts significant areas under grassland. While the presence of grasslands in this region is itself an intriguing ecological pattern that is not well understood, observations of decline in grassland habitat in the face of changing environmental conditions and increasing human impacts has become a matter of increasing concern. Although the climatic conditions prevalent in the region are expected to favor woody vegetation, grassland and savanna support much of the biodiversity, particularly of the endemic species. Grassland areas declined overall by about 35% of its original extent during the 30-year period, but the losses occurred in five of the eight sites that we studied, while it increased marginally in three sites. This suggests that grassland decline is perhaps driven by variation in local environmental conditions. The presence of dry season grass fires was the most important factor that affected the persistence of grasslands across all the sites, and fire regimes do vary significantly among sites depending on local conditions 34,35 . In general, the decreases in grassland areas were accompanied by increases in the woodland areas, which was the most common transformation of grasslands in the terai. We would need to examine historical changes in the entire terai ecoregion in a more comprehensive study to test whether woodland encroachment has been the dominant trend across this ecoregion. But our analyses of best protected conservation areas in the terai already suggest that woodland encroachment is a threat to grassland habitat in this region. The greater persistence of grasslands in spatial proximity to anthropogenic influence may appear counter-intuitive, but it may be caused by the influence of human activities on fire occurrence. Anthropogenic influence, through climatic changes and direct impacts (measured as a human footprint), have been increasing across all sites in the terai, but we lack high-resolution spatial-temporal data on climatic changes to quantify their influence on grassland-woodland transitions. Our landcover classification was highly accurate, and we found that woodland (mostly forest, but also early-successional woodland) is the dominant landcover class overall. The dominance of woody vegetation is expected given the climatic conditions, but the presence of a high biodiversity of grassland specialist plant and animal taxa suggest the long-term importance and prevalence of grassland habitats in this ecoregion. The smaller extent of grasslands and the highly variable woodland versus grassland proportions that we observed in the eight sites, may be due to more recent anthropogenic influence within the last century or more. For example, Manas NP, had twice as much grassland area compared to the woodland area as recently as 1988-1989 and it is among the wetter sites. Currently, Manas NP and Shukla-Phanta NP are the only parks with more grassland than woodland. The trends that we see in the last 30 years may have persisted for longer, and may continue into the future. It is noteworthy that despite considering all the protected habitats which have substantial grasslands in the terai ecoregion, we found only 1417 km 2 of grassland habitat in 1988-1989, which further reduced to 923 km 2 by 2018-2019. This loss of grassland has not drawn adequate conservation attention and needs to be addressed. In sites that saw declines in grassland, the reduction was large, with more than 50% loss in three sites, and this occurred within three decades. Grassland is therefore being lost rapidly in key terai sites, and there is a clear need to investigate the drivers of these changes at each site in detail. The causes of woodland encroachment are often broken down into three hierarchical scales of drivers: (1) global-scale drivers such as elevated CO 2 , (2) regional-scale climatic drivers (e.g., precipitation and temperature changes), and (3) local-scale drivers such as land management history, changes in fire frequencies, land fragmentation, and removal of native herbivores 25,36–39 . Globally, each biome that is undergoing woody encroachment may have a suite of these interacting drivers that influence the rate of woody encroachment. We tried to examine the influence of a set of regional and local drivers on the persistence of grasslands in the terai. However, we did not have the high spatial resolution in the time series data on temperature and rainfall, nor did we have a large number of sites, to quantify the influence of regional climatic drivers on the differences in vegetation changes we observed. Mean temperature increase and changes in the drought index (SPI) were different among sites, but since we had just eight sites, we could not reliably address how the observed climatic changes within the Terai region have influenced grassland persistence. We did not also detect any consistent relationship with the total precipitation or seasonality and the proportion of grassland in the site. For example, Shukla-Phanta NP and Manas NP had the lowest and highest mean annual rainfall, respectively, but they also had a greater proportion of grassland area than other sites (Table 1 and Supplementary Table 1). Bardiya NP, Valmiki TR, and Chitwan NP had mean annual rainfall in the range 1310 to 1946 mm, and all had low proportions of the area under grassland. It follows, therefore, that discerning the influence of future climatic changes on the vegetation in this region is likely to be complex. At these scales, the ecological mechanisms behind the dominance of grasses or woody species are known to be associated with the interactions between precipitation amount, seasonality, and soil properties; however, all these factors may interact with the occurrence of fire, herbivory, or hydrological attributes in a particular site. Despite high annual rainfall, the presence of a long and strong dry season can promote fires and influence vegetation physiognomy, and this appears to be the dominant force shaping terai vegetation. Local hydrology can influence the fire regime and soil water profiles in each site, thereby affect vegetation physiognomy at a site 23,40 . The levels and duration of inundation due to monsoon floods varies within terai habitats and may create relatively swampy or dry edaphic conditions. Soil infiltration rates vary between fine-textured (clayey) and coarse-textured (sandy) soils, affecting the soil water profile and thereby the balance of woody versus grassy vegetation 41,42 . Some parts of the terai certainly have swampy conditions that some grasses can tolerate but woody plants may fail to establish. Here, due to the wet conditions that may persist through the dry season, even fire occurrence may be low compared to dry-grassland sites 35 . In dry-grassland sites, regular dry season fires are more probable, and the occurrence of fire and persistence of grassland extent appears to be involved in a positive feedback cycle 35 . The occurrence of fires in terai ecosystems is largely due to anthropogenic influence emanating from land use activities by local people and management interventions by the forest department. We found that greater anthropogenic proximity (DIST_ANTH) led to greater persistence of grassland over time. We surmise that anthropogenic proximity has the greatest influence on fire. Since these are protected areas, grazing by domestic livestock and resource harvest is not legally permitted, but such activities may be occurring at low intensity in some sites and may affect encroachment by woody vegetation 43,44 . To a small extent, grassland loss has also been due to conversion to cropland, and cropland area has expanded from the edges of human settlements, contributing to the influence of anthropogenic proximity. Such anthropogenic effects appear to be greater at the Indian sites, where population density and pressure is greater compared to that in Nepal. Our study does not explicitly consider several important factors that are known to affect grassland and woodland composition. While fire may be sufficient to maintain grasslands in sites where the climatic conditions favour woody vegetation, other factors such as the actions of large mammalian herbivores, hydrological attributes, and soil properties may also be important in any site. We did not have such detailed data for the Terai region to carry these tests, but future studies in each site must consider the entire suite of variables that determine grassland presence in terai ecosystems. Fire may well be critical for maintaining grasslands in the terai, and proper fire management is needed to ensure the desired distributions of habitats for wildlife species, particularly the region's endangered and endemic grassland specialists. Nevertheless, the increasing fragmentation of terai ecosystems may also be altering the influence of hydrology and long-term ecological processes that maintain the diversity and complexity of terrain vegetation. Methods Study Area: The extensive lowland region of floodplain grasslands, woodland-grassland mixtures, and forests located at the southern base of the Himalayan Mountains in India and Nepal is known as the Terai-Duar Savanna and Grasslands ecoregion 21 . It is a narrow belt of alluvial grasslands and savannas with patches of forests stretching from Yamuna River in the west to the Brahmaputra River in the east 20 . Typically, the alluvial grassland is dominated by tall grasses like Themeda arundinacea, Saccharum narenga, Phragmites karka, Imperata cylindrica, Chrysopogon zizanioides and several other species. The terai is prime habitat for the flagship species like the tiger ( Panthera tigris ), the Asian elephant ( Elephas maximus ) and critical habitat for the greater one-horned rhinoceros ( Rhinoceros unicornis ) 20 , and several grassland specific endangered and endemic species like the hispid hare ( Caprolagus hispidus ), pygmy hog ( Porcula salvania ) and Bengal florican ( Houbaropsis bengalensis ) 45,46 . This region has a monsoon-influenced sub-humid to humid subtropical climate with significantly more rain in the wet summer months (April – September) than in the dry winter months (October – March) 47 . While the western part of the terai receives a mean annual rainfall of about 1200-1300 mm, the eastern part is much wetter with over 2400-2500 mm of rain. The mean annual temperature range is 20 - 28 °C, with mean monthly minimum and maximum temperatures (30-year average) being 5.5 °C and 40.5 °C, respectively 47 . Terai grasslands are highly threatened, and most of them have already been converted to other land uses. The largest areas of natural habitat are confined to a few protected areas 20 . These eight protected areas that we study here are prominent sites in India and Nepal and represent typical terai sites with extensive grassland and grassland-woodland mixtures (Fig. 1 and Table 1). Landcover Classification: The vegetation across the eight protected areas is typically a mosaic of forests, woody savanna, and pure grasslands, of which moist deciduous forest and alluvial grassland are most notable. The tree species in forests are characteristic of both semi-evergreen forests and moist deciduous forests. Grasslands are broadly two distinct types - the swampy and dry grasslands. The woody savanna formations occur mostly at the ecotone of forest and grassland with a few early successional trees such as Bombax ceiba , Dillenia pentagyna , and Lagerstroemia parviflora 48 . All the eight protected areas that we studied exhibit a composite of the vegetation formations mentioned above. However, to keep the classification simple, and because of the difficulty of distinguishing the forest from woody savanna accurately through remote sensing, we considered only two broad vegetation types: grassland, which represent pure grassland and perhaps sites with very low tree density (<10% canopy cover) and woodland as representing woody savanna and forest. For completeness, we needed to consider four additional landcover types: water bodies, sand and open land, cropland, and human settlements. Remotely-sensed data: Landsat imageries are convenient for detecting land-cover change 49 and forest-cover dynamics in grasslands and forested areas 50–52 . We acquired Landsat 4-5 Thematic Mapper (TM), Landsat 7 Enhanced Thematic Mapper Plus (ETM+), and Landsat 8 Operational Land Imager (OLI) images at 30m spatial resolution from the U.S. Geological Survey. After the thorough screening, we selected the post-monsoon dry season (December-March) images to ensure cloud-free coverage. For all three times points (1988-1989, 2003-2004, and 2018-2019), one single Landsat tile was sufficient to cover the full extent of Shukla-Phanta, Dudhwa and Bardiya NPs. However, we needed two tiles for Valmiki Tiger Reserve and Chitwan National Park also for the 2003-2004 data for Manas NP. Landsat 7 ETM+ products have scanline errors in post-May 2004 imageries, so the Landsat 7 ETM+ images we used in the study were selected with the acquisition date before May 2004. The detailed information of the obtained Landsat images used for the landcover classification is shown in Supplementary Table 3. Training and Validation Sample Data : We mostly collected the training and validation data through visual inspection from Google Earth imagery, but we also used real ground verification data for Manas NP in India. The practice of generating training data for landcover classification through manual visual interpretation of high spatial resolution images from Google Earth imageries is widely applied and reported in the literature 53–55 . We focused on Google Earth imagery separately for the three periods, 1988-1989, 2003-2004 and 2018-2019, and selected a set of point locations to serve as training data for six land cover classes. For the 2018-2019 period, we also collected ground data through extensive field surveys at Manas NP in India. Using a GPS unit, we gathered ground truth information for at least 20 locations for each landcover type located within the park boundary. The combined data were then used as training inputs for the classifier. The number of training and validation samples for each protected area was different, depending on the land cover class and the size of the protected area, but we ensured sufficient sample size in each case. Supervised Landcover Classification: We performed supervised land-cover classification using 'Random Forests' (RF), an ensemble-based decision-tree classification algorithm 56 . RF is a well-known classification method where several regression trees are generated, and a consensus tree is derived by averaging the predictions of the individual trees 57 . The RF algorithm has been widely preferred for land cover classification using satellite imagery because of its higher classification accuracy 58,59 . Although other classification algorithms are also available, we reasoned that RF was most appropriate for our objective. For classification, we built the predictive regression model using eight predictor variables. These include six Landsat bands (Blue, Green, Red, Near Infrared, Short-wave Infrared I and Short-wave Infrared II) and two vegetation indices (Normalized Difference Vegetation Index and Normalized Difference Moisture Index). For each classification, we used the 'confusion matrix' reports for accuracy assessment. After accuracy assessment using the validation data, we derived the classified images from the predictions of the RF model and then used the images for quantifying landcover transitions. Influence of anthropogenic and environmental factors on Grassland Habitat: Our landcover analysis was constrained to start at 1988-1989 due to the availability of satellite data. We identified the grassland area for 1988-1989 from the classified images and generated the baseline grassland polygons for all eight sites. We then placed a 500 × 500 m 2 grid to produce non-overlapping areas over the grassland polygons and overlaid the 30-m landcover classification map of 2018-2019. This allowed us to compute the percentage of grassland area that persisted as grassland per grid cell at the end of 30 years. We also note that because of the irregular shape of the grassland polygons and the park boundary, not all grid cells were exactly 500 × 500 m 2 in size and had different numbers of pixels. Therefore, quantifying the percentage of pixels transformed within each grid cell was a robust measure for spatial data analyses. More generally, non-overlapping areas (cells) of any shape can be placed on the study area in such analyses. We thus derived spatial data for grassland transitions as the proportion of pixels in a cell that remained as grassland at the end of the study period. To understand the environmental determinants of the observed spatial pattern of grassland transitions, we derived several independent spatial environmental variables for testing their influence on grassland persistence across the terai sites. Using ArcMap 10.8 software, we assembled the following variables: (i) Fire occurrence: The near real-time Moderate Resolution Imaging Spectroradiometer (MODIS) provides 'active fire' data (with data code MCD14DL), which gives the point location of active fire with the exact time and date of detection 60 . These data are available from October 2000 and can be obtained from NASA's automated Fire Information for Resource Management System (FIRMS). Here we used the fire occurrence data of 19 years (October 2000 to September 2019). We delineated a rectangular area around the geographical boundary of each protected area, with a buffer of 1 km from the park boundary and extracted all the active fire incidents within this rectangular area for 19 years. Then, we counted the number of fire events (FIRE_OCCURR) within each grassland grid cell to derive the fire counts per cell in 19 years. (ii) Topographic features : We obtained the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) Global Digital Elevation Model Version 3 (GDEM 003) elevation data with a spatial resolution of 30 m for the Terai region. Using the 'spatial analyst' and 'raster calculator' tools in ArcMap, we generated the Slope (SLOPE), Topographic Wetness Index (TWI) 61 and Ruggedness (RUGG) rasters following standard protocols and upscaled them to derive the mean value for each 500 m grassland cell. We use these topographical attributes because they known to affect soil water retention and the potential for hydrological saturation, thus influencing small-scale variation in water availability. (iii) Woodland proximity: Human-induced alteration of grazing and fire regimes may promote the replacement of grassland by woodland 62 . However, the scope of ecological succession may not be uniform everywhere. A higher richness of soil seed bank and a favourable physical environment is expected in grassland sites near the forest than sites farther from the forest edge 63 . So, we considered the distance to the nearest woodland patch (DIST_WOOD) as a potential factor. The 1988-1989 period woodland coverage was considered for the proximity calculation. (iv) Anthropogenic proximity: Transformation of grassland to agriculture has been occurring across the world. WE expect that the prospect of grassland conversion near existing farmland and human settlement is greater than for distant grassland areas. So, we attempted to compute the distance to the nearest cropland area for each grassland cell, but cropland did not exist in some protected areas in 1988-1989. In the absence of the cropland within the boundary, we considered the distance to the human settlement or boundary of the protected area as the proxy for anthropogenic proximity. We obtained anthropogenic proximity (DIST_ANTH) for each grassland cell by calculating the distance from the cropland or human settlement or the protected area boundary (depending on which one is the nearest). Conditional Autoregressive Bayesian Spatial model: We calculated the proportion of grassland that persists ( Y k ) for a 500-m grassland cell based on the number of 30-m pixels ( n k ) that were transformed within the cell. If all the pixels in a cell remain as grassland, then it gets a maximum value of 1, and if the cell is completely transformed, it is assigned a minimum value of 0. When only a subset of pixels is transformed, the corresponding proportional value is assigned. As environmental predictors of grassland persistence, we used six variables - FIRE_OCCURR, SLOPE, TWI, RUGG, DIST_WOOD, and DIST_ANTH. These variables are either relatively stable over time (e.g., SLOPE) or were cumulative values over the study period (e.g., FIRE_OCCUR). Considering the grassland cells in all the protected areas together, we first fitted a nonspatial binomial model that incorporates covariate effects at the cell scale and assessed spatial autocorrelation by examining the residuals of the regression. In the statistical framework, we assume that the study region is partitioned into k non-overlapping areal cells (500 × 500 m 2 in our case), which are linked to a corresponding set of responses Y 1 , . . . , Y k and a vector of known offsets O 1 , . . . , O k . As explained above, each Y i ( i =1, . . . , k ) represents the proportion of pixels in the i th cell that persists as grassland at the end of the study period. The covariate values were scaled before fitting the model. In this binomial model, the number of pixels in cell k ( n k ) is the number of trials for the k th cell, while θ k is the probability of success in a single trial 64 . The model therefore was: Where, is the transpose of the matrix of the explanatory spatial covariates. A spatial structure component ( ψ k ) is included to model any spatial autocorrelation present in the data. The vector of regression parameters is denoted by β = (β 1 ,...,β k ) 64 . We first performed the nonspatial binomial regression using the 'glm' function in the R statistical software's 'stats' package 65 . To quantify spatial autocorrelation in the residuals, we used a Monte Carlo test and compared the observed Moran's I with the null distribution of Moran's I values obtained under no autocorrelation. This null distribution of Moran's I was computed after randomly permuting the grassland proportion data across the grid cells in the landscape 1000 times and computing Moran's I for each permutation. Finally, we performed Bayesian spatial modelling of the proportion of grassland that persisted per cell using a Conditional Autoregressive (CAR) prior, which induces spatial autocorrelation through the adjacency structure of the areal units, and implemented this in the R package 'CARBayes' 64 . The linear predictor was the set of environmental variables indicated above for the binomial response and was treated as random effects. We implemented the function 'S.glm' within the R package CARBayes to fit the Bayesian regression model with autocorrelation specified as shown above. S.glm fits the model with no random effects and is an appropriate conditional autoregressive prior for the binomial variables 64 . We used a Markov Chain Monte Carlo (MCMC) based simulation of the posterior distributions of model parameters. The simulation involved 10,000 MCMC samples generated from a single Markov chain, executed for 300,000 iterations with a burn-in of 100,000 and then thinned by 20 to reduce Markov chain autocorrelation. Assessment of threats to the Protected Areas: We computed the threat level in three categories for our eight protected areas of interest: habitat loss, anthropogenic influence, and climatic stressors: (i) Habitat loss: We focused only on the threats to the grassland habitat that resulted in grassland loss. Here, we calculated the change in the grassland cover over 30 years, starting with the grassland extent observed in 1988-1989. (ii) Anthropogenic influence: To measure the extent of anthropogenic influence on the protected areas, we obtained the global map of human footprint (HFP) from https://wcshumanfootprint.org. The human footprint map is prepared based on human settlements, croplands, pasture lands, human population density, night-time lights, railway and road networks, and navigable waterways 66 . It captures the cumulative human pressure on the environment. The human footprint maps of two time points (1993 and 2009) are available at present. We obtained the maps of both periods and computed the change in human pressure within each protected area over the 16-year period. We could not cover the entire 30-year study period due to the lack of data. (iii) Climatic stressors: We examined the average temperature change and rainfall anomalies to understand the climatic threat for the protected areas. We obtained climatic data for 60 years (1961-2020) from the Climate Research Unit (CRU TS v. 4.04) database ( http://www.cru.uea.ac.uk/data ) 67 . We calculated the average temperature for two decadal periods, 1961-1970 and 2011-2020, to measure climatic changes over half a century. We also computed a drought index called the Standardized Precipitation Index (SPI) using monthly rainfall data cumulated over two months 68 . The terai receives ample rainfall between June to September, but climate change may have impacted rainfall distribution. For any month, an SPI value less than -1.5 indicates severe drought. Here, we calculated the number of monsoonal months in the last decade (total forty months, four months for ten years, 2011-2020) that have scored below -1.5 to understand the rainfall anomalies. Declarations Acknowledgements This work was supported by Grant No. EMR/2017/002769 awarded to RJ by SERB-DST, Government of India and Rufford Small Grant No. 21980-1 awarded to SB by Rufford Foundation. We thank the Assam Forest Department for support on field visits. We also want to thank Mr Swapnil Bhowal for helping us to obtain the historical images from Google Earth to carry out the landcover classification of the past. Author contributions Subham Banerjee: Conceptualization, Methodology, Data curation, Investigation, Writing- Original draft preparation. Dhritiman Das: Conceptualization, Data curation. Robert John: Supervision, Validation, Writing- Reviewing and Editing. Competing interests The authors declare no competing interests. References 1. Bond, W. J. & Parr, C. L. Beyond the forest edge: Ecology, diversity and conservation of the grassy biomes. Biological Conservation 143 , 2395–2404 (2010). 2. Bardgett, R. D. et al. Combatting global grassland degradation. Nat Rev Earth Environ (2021) doi:10.1038/s43017-021-00207-2. 3. Gibbs, H. K. & Salmon, J. M. Mapping the world’s degraded lands. Applied Geography 57 , 12–21 (2015). 4. Gang, C. et al. Quantitative assessment of the contributions of climate change and human activities on global grassland degradation. Environ Earth Sci 72 , 4273–4282 (2014). 5. Overbeck, G. E. et al. Conservation in Brazil needs to include non-forest ecosystems. Diversity Distrib. 21 , 1455–1460 (2015). 6. Parr, C. L., Lehmann, C. E. R., Bond, W. J., Hoffmann, W. A. & Andersen, A. N. Tropical grassy biomes: misunderstood, neglected, and under threat. Trends in Ecology & Evolution 29 , 205–213 (2014). 7. Murphy, B. P., Andersen, A. N. & Parr, C. L. The underestimated biodiversity of tropical grassy biomes. Phil. Trans. R. Soc. B 371 , 20150319 (2016). 8. Van Auken, O. W. Shrub Invasions of North American Semiarid Grasslands. Annu. Rev. Ecol. Syst. 31 , 197–215 (2000). 9. Schlesinger, W. H. et al. Biological Feedbacks in Global Desertification. Science 247 , 1043–1048 (1990). 10. Sankaran, M., Ratnam, J. & Hanan, N. Woody cover in African savannas: the role of resources, fire and herbivory. Global Ecol Biogeography 17 , 236–245 (2008). 11. Soliveres, S. et al. Plant diversity and ecosystem multifunctionality peak at intermediate levels of woody cover in global drylands. Global Ecology and Biogeography 23 , 1408–1416 (2014). 12. Frost, P. et al. Responses of Savannas to Stress and Disturbance. Biology International, Special Issue 10. International Union of Biological Sciences, Paris (1986). 13. Sankaran, M. et al. Determinants of woody cover in African savannas. Nature 438 , 846–849 (2005). 14. Bond, W. J., Midgley, G. F. & Woodward, F. I. What controls South African vegetation — climate or fire? South African Journal of Botany 69 , 79–91 (2003). 15. Ratajczak, Z., Nippert, J. B. & Collins, S. L. Woody encroachment decreases diversity across North American grasslands and savannas. Ecology 93 , 697–703 (2012). 16. Quero, J. L., Maestre, F. T., Ochoa, V., García-Gómez, M. & Delgado-Baquerizo, M. On the Importance of Shrub Encroachment by Sprouters, Climate, Species Richness and Anthropic Factors for Ecosystem Multifunctionality in Semi-arid Mediterranean Ecosystems. Ecosystems 16 , 1248–1261 (2013). 17. Eldridge, D. J., Soliveres, S., Bowker, M. A. & Val, J. Grazing dampens the positive effects of shrub encroachment on ecosystem functions in a semi-arid woodland. J Appl Ecol 50 , 1028–1038 (2013). 18. Kulmatiski, A. & Beard, K. H. Woody plant encroachment facilitated by increased precipitation intensity. Nature Clim Change 3 , 833–837 (2013). 19. Olson, D. M. et al. Terrestrial Ecoregions of the World: A New Map of Life on Earth. BioScience 51 , 933 (2001). 20. Semwal, R. L. The Terai arc landscape in India: Securing protected areas in the face of global change . (Forests & Biodiversity Conservation Programme, World Wide Fund for Nature …, 2005). 21. Dinerstein, E. An ecological survey of the royal Karnali-Bardia Wildlife Reserve, Nepal. Part I: Vegetation, modifying factors, and successional relationships. Biological Conservation 15 , 127–150 (1979). 22. Sharma, B. Wildlife habitat mapping by using Geographic Information Systems (GIS) in the Karnali floodplain of Royal Bardia National Park at lowland Nepal. (1999). 23. Hirota, M., Holmgren, M., Van Nes, E. H. & Scheffer, M. Global Resilience of Tropical Forest and Savanna to Critical Transitions. Science 334 , 232–235 (2011). 24. Archer, S. R. et al. Woody Plant Encroachment: Causes and Consequences. in Rangeland Systems (ed. Briske, D. D.) 25–84 (Springer International Publishing, 2017). doi:10.1007/978-3-319-46709-2_2. 25. Stevens, N., Lehmann, C. E. R., Murphy, B. P. & Durigan, G. Savanna woody encroachment is widespread across three continents. Glob Change Biol 23 , 235–244 (2017). 26. Archer, S. Tree-grass dynamics in a Prosopis-thornscrub savanna parkland: Reconstructing the past and predicting the future. Écoscience 2 , 83–99 (1995). 27. Hobbs, R. J. & Mooney, H. A. Community changes following shrub invasion of grassland. Oecologia 70 , 508–513 (1986). 28. Adamoli, J., Sennhauser, E., Acero, J. M. & Rescia, A. Stress and Disturbance: Vegetation Dynamics in the Dry Chaco Region of Argentina. Journal of Biogeography 17 , 491 (1990). 29. Vegten, J. A. Thornbush invasion in a savanna ecosystem in eastern Botswana. Vegetatio 56 , 3–7 (1984). 30. Burrows, W. H., Carter, J. O., Scanlan, J. C. & Anderson, E. R. Management of Savannas for Livestock Production in North-East Australia: Contrasts Across the Tree-Grass Continuum. Journal of Biogeography 17 , 503 (1990). 31. Singh, J. S. & Joshi, M. C. Ecology of the Semi-Arid Regions of India with Emphasis on Land-Use. in Developments in Agricultural and Managed Forest Ecology vol. 7 243–275 (Elsevier, 1979). 32. Barger, N. N. et al. Woody plant proliferation in North American drylands: A synthesis of impacts on ecosystem carbon balance. J. Geophys. Res. 116 , G00K07 (2011). 33. Eldridge, D. J. et al. Impacts of shrub encroachment on ecosystem structure and functioning: towards a global synthesis: Synthesizing shrub encroachment effects. Ecology Letters 14 , 709–722 (2011). 34. Murthy, K. K., Sinha, S. K., Kaul, R. & Vaidyanathan, S. A fine-scale state-space model to understand drivers of forest fires in the Himalayan foothills. Forest Ecology and Management 432 , 902–911 (2019). 35. Banerjee, S., Das, D. & John, R. Grassland vegetation and roads have dominant influence on decadal-scale spatial-temporal patterns of fires in a species-rich protected Terai habitat in northeastern India. Agricultural and Forest Meteorology 304–305 , 108411 (2021). 36. Archer, S., Schimel, D. S. & Holland, E. A. Mechanisms of shrubland expansion: land use, climate or CO2? Climatic Change 29 , 91–99 (1995). 37. Van Auken, O. W. Causes and consequences of woody plant encroachment into western North American grasslands. Journal of Environmental Management 90 , 2931–2942 (2009). 38. Venter, Z. S., Cramer, M. D. & Hawkins, H.-J. Drivers of woody plant encroachment over Africa. Nat Commun 9 , 2272 (2018). 39. Wigley, B. J., Bond, W. J. & Hoffman, M. T. Thicket expansion in a South African savanna under divergent land use: local vs. global drivers? Global Change Biology 16 , 964–976 (2010). 40. Sala, O. E. & Maestre, F. T. Grass-woodland transitions: determinants and consequences for ecosystem functioning and provisioning of services. J Ecol 102 , 1357–1362 (2014). 41. Sala, O., Lauenroth, W. & Golluscio, R. Plant functional types in temperate semi-arid regions. Plant functional types: their relevance to ecosystem properties and global change 1 , 217 (1997). 42. Saxton, K. E. & Rawls, W. J. Soil Water Characteristic Estimates by Texture and Organic Matter for Hydrologic Solutions. Soil Sci. Soc. Am. J. 70 , 1569–1578 (2006). 43. Goheen, J. R. et al. Conservation lessons from large-mammal manipulations in East African savannas: the KLEE, UHURU, and GLADE experiments: Large mammal manipulations in Laikipia, Kenya. Ann. N.Y. Acad. Sci. 1429 , 31–49 (2018). 44. Ward, D. Clipping frequency but not nutrients affect the architecture and non-structural carbohydrates of a browsing lawn. Plant Ecol 217 , 21–29 (2016). 45. Collar, N., Garcia, E. & Sharpe, C. Bengal Florican (Houbaropsis bengalensis). in Birds of the World (eds. Billerman, S. M., Keeney, B. K., Rodewald, P. G. & Schulenberg, T. S.) (Cornell Lab of Ornithology, 2020). doi:10.2173/bow.benflo2.01. 46. Oliver, W. L. R. & Deb Roy, S. The pygmy hog (Sus salvanius). Pigs, Peccaries, and Hippos: Status Survey and Conservation Action Plan (Oliver, WLR ed.). IUCN, Gland, Switzerland (1993). 47. Fick, S. E. & Hijmans, R. J. WorldClim 2: new 1‐km spatial resolution climate surfaces for global land areas. Int. J. Climatol 37 , 4302–4315 (2017). 48. Johnsingh, A. T. Status and conservation of the tiger in Uttaranchal, Northern India. AMBIO: A Journal of the Human Environment 35 , 135–137 (2006). 49. Wulder, M. A. et al. Current status of Landsat program, science, and applications. Remote Sensing of Environment 225 , 127–147 (2019). 50. Achard, F. et al. Determination of tropical deforestation rates and related carbon losses from 1990 to 2010. Glob Change Biol 20 , 2540–2554 (2014). 51. Hu, Y. & Nacun, B. An Analysis of Land-Use Change and Grassland Degradation from a Policy Perspective in Inner Mongolia, China, 1990–2015. Sustainability 10 , 4048 (2018). 52. Keenan, R. J. et al. Dynamics of global forest area: Results from the FAO Global Forest Resources Assessment 2015. Forest Ecology and Management 352 , 9–20 (2015). 53. de Sousa, C. et al. Cloud-computing and machine learning in support of country-level land cover and ecosystem extent mapping in Liberia and Gabon. PLoS ONE 15 , e0227438 (2020). 54. Midekisa, A. et al. Mapping land cover change over continental Africa using Landsat and Google Earth Engine cloud computing. PLoS ONE 12 , e0184926 (2017). 55. Phan, T. N., Kuch, V. & Lehnert, L. W. Land Cover Classification using Google Earth Engine and Random Forest Classifier—The Role of Image Composition. Remote Sensing 12 , 2411 (2020). 56. Breiman, L. Random forests. Machine learning 45 , 5–32 (2001). 57. Cutler, D. R. et al. Random forests for classification in ecology. Ecology 88 , 2783–2792 (2007). 58. Maxwell, A. E. et al. Large-Area, High Spatial Resolution Land Cover Mapping Using Random Forests, GEOBIA, and NAIP Orthophotography: Findings and Recommendations. Remote Sensing 11 , 1409 (2019). 59. Tamiminia, H. et al. Google Earth Engine for geo-big data applications: A meta-analysis and systematic review. ISPRS Journal of Photogrammetry and Remote Sensing 164 , 152–170 (2020). 60. Davies, D. K., Ilavajhala, S., Min Minnie Wong & Justice, C. O. Fire Information for Resource Management System: Archiving and Distributing MODIS Active Fire Data. IEEE Trans. Geosci. Remote Sensing 47 , 72–79 (2009). 61. Mattivi, P., Franci, F., Lambertini, A. & Bitelli, G. TWI computation: a comparison of different open source GISs. Open geospatial data, softw. stand. 4 , 6 (2019). 62. Hibbard, K. A., Schimel, D. S., Archer, S., Ojima, D. S. & Parton, W. Grassland to woodland transitions: integrating changes in landscape structure and biogeochemistry. Ecological Applications 13 , 911–926 (2003). 63. Duncan, R. S. & Duncan, V. E. Forest Succession and Distance from Forest Edge in an Afro-Tropical Grassland. Biotropica 32 , 33–41 (2000). 64. Lee, D. Carbayes version 4.6: An r package for spatial areal unit modelling with conditional autoregressive priors. University of Glasgow, Glasgow (2017). 65. R Core Team. R: A language and environment for statistical computing. (2021). 66. Venter, O. et al. Sixteen years of change in the global terrestrial human footprint and implications for biodiversity conservation. Nat Commun 7 , 12558 (2016). 67. Harris, I., Osborn, T. J., Jones, P. & Lister, D. Version 4 of the CRU TS monthly high-resolution gridded multivariate climate dataset. Sci Data 7 , 109 (2020). 68. World Meteorological Organization. Standardized precipitation index user guide. (2012). Tables Table 1 Sl No Protected Area Area (km 2 ) Country Latitudinal Range (°N) Longitudinal Range (°E) Elevation Range (m) Temperature Range (°C) Annual Precipitation (mm) Effective Rainfall (mm) A Shukla Phanta National Park 376 Nepal 28.76 ­ 29.05 80.10 - 80.39 143 - 1337 6.75 - 40.45 1258 -248.7 B Dudhwa National Park 746 India 28.32 - 28.70 80.46 - 80.93 115 - 233 6.50 - 39.65 1266 -304.1 C Bardiya National Park 851 Nepal 28.29 - 28.67 81.20 - 81.71 119 - 1574 5.72 - 37.83 1310 -220.1 D Valmiki Tiger Reserve 874 India 27.16 - 27.52 83.83 - 84.65 60 - 860 7.72 - 39.00 1693 165.5 E Chitwan National Park 1137 Nepal 27.34 - 27.69 83.88 - 84.75 89 - 784 6.72 - 37.20 1949 484.5 F Koshi Tappu Wildlife Sanctuary 143 Nepal 26.57 - 26.73 86.92 - 87.08 42 - 110 6.12 - 33.20 1536 -44.1 G Jaldapara National Park 221 India 26.52 - 26.86 89.25 - 89.42 11 - 841 8.20 - 32.15 2586 1124.8 H Manas National Park 519 India 26.60 -26.82 90.81 - 91.24 6 - 305 7.22 - 31.57 2368 817.3 Table 1 Sl No Protected Area Area (km 2 ) Country Latitudinal Range (°N) Longitudinal Range (°E) Elevation Range (m) Temperature Range (°C) Annual Precipitation (mm) Effective Rainfall (mm) A Shukla Phanta National Park 376 Nepal 28.76 ­ 29.05 80.10 - 80.39 143 - 1337 6.75 - 40.45 1258 -248.7 B Dudhwa National Park 746 India 28.32 - 28.70 80.46 - 80.93 115 - 233 6.50 - 39.65 1266 -304.1 C Bardiya National Park 851 Nepal 28.29 - 28.67 81.20 - 81.71 119 - 1574 5.72 - 37.83 1310 -220.1 D Valmiki Tiger Reserve 874 India 27.16 - 27.52 83.83 - 84.65 60 - 860 7.72 - 39.00 1693 165.5 E Chitwan National Park 1137 Nepal 27.34 - 27.69 83.88 - 84.75 89 - 784 6.72 - 37.20 1949 484.5 F Koshi Tappu Wildlife Sanctuary 143 Nepal 26.57 - 26.73 86.92 - 87.08 42 - 110 6.12 - 33.20 1536 -44.1 G Jaldapara National Park 221 India 26.52 - 26.86 89.25 - 89.42 11 - 841 8.20 - 32.15 2586 1124.8 H Manas National Park 519 India 26.60 -26.82 90.81 - 91.24 6 - 305 7.22 - 31.57 2368 817.3 Table 3 Variable Estimate Std. Error z value Pr(>|z|) (Intercept) -0.443 0.023 -19.139 <0.001 FIRE_OCCURR 0.638 0.028 22.741 <0.001 TWI -0.086 0.033 -2.596 0.0094 SLOPE 0.097 0.032 3.02 0.0025 RUGG -0.02 0.025 -0.808 0.4193 DIST_WOOD 0.082 0.024 3.448 0.0006 DIST_ANTH -0.256 0.027 -9.609 <0.001 Table 4 Variable Median 2.50% 97.50% n.effective Geweke.diag (Intercept) -0.435 -0.480 -0.390 8515.9 0.3 FIRE_OCCURR 0.627 0.573 0.682 7001.2 -0.2 TWI -0.099 -0.165 -0.033 4756.9 -0.9 SLOPE 0.089 0.027 0.151 5022.2 0.1 RUGG -0.010 -0.060 0.038 7634.9 1.1 DIST_WOOD 0.062 0.015 0.110 8662.0 -0.9 DIST_CROP -0.256 -0.308 -0.205 7809.0 -0.8 Table 5 Protected Area Mean Temperature change (°C) Dry Monsoon Months [SPI <-1.50] Human Footprint change (%) Grassland Cover change (%) Shukla-Phanta National Park 0.607 2 21.92 -51.65 Dudhwa National Park 0.652 3 54.33 -59.82 Baradiya National Park 0.726 5 9.1 34.57 Valmiki Tiger Reserve 0.656 8 39.9 18.12 Chitwan National Park 0.674 8 27.3 -65.59 Koshi Tappu WLS 0.577 3 12.98 28.97 Jaldapara National Park 0.684 6 18.48 -26.16 Manas National Park 0.535 4 51.05 -40.25 Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryTable1.docx Supplementary Table 1 SupplementaryTable2.docx Supplementary Table 2 SupplementaryTable3.docx Supplementary Table 3 Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-1398899","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":88215935,"identity":"fcde4a1e-810d-43ea-8077-ea4f00ed3ddd","order_by":0,"name":"Subham Banerjee","email":"","orcid":"https://orcid.org/0000-0002-9212-6000","institution":"Indian Institute of Science Education and Research Kolkata","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Subham","middleName":"","lastName":"Banerjee","suffix":""},{"id":88215936,"identity":"56efff4f-2a57-4128-ba93-6a23e2851a3b","order_by":1,"name":"Dhritiman Das","email":"","orcid":"","institution":"Pygmy Hog Conservation Programme, Durrell Wildlife Conservation Trust","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Dhritiman","middleName":"","lastName":"Das","suffix":""},{"id":88215937,"identity":"2f2312c3-58d6-4eda-a252-ec7762330c6d","order_by":2,"name":"Robert John","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABBklEQVRIie3SMWuDQBTA8SeCWa51fRJJvsJJQJfgZ3nhIJNDxhaKdUqX2NmhH+Yk4JQ6ZxT6BZqlpFBKrzY3FMy1Y6H3R4R7+OPuQACb7U/G+rcLCE5HvB848nPkGQn1xOUnAr8ioIiHeiBNx0ruyvrp+ggjf1w2V90qnySbrZRwk8LluBgk4a4V0U4dLHhol3vi21n4uCYJjQAvHN4MMYuDQhG+z2JF5KLyGZfgqQfpHEleNVkRz2+/yLuRxI4mQNwlvNhw6awNhLUiKJboBlU2Q3WXqGINycW9YGfJqKwPxXwufMyiw/EtnyITdff8kk6m1TDRUnxfk/4rDKU/fWCz2Wz/uA+u/FI9emjotQAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0002-2848-9413","institution":"Indian Institute of Science Education and Research Kolkata","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Robert","middleName":"","lastName":"John","suffix":""}],"badges":[],"createdAt":"2022-02-26 14:05:43","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-1398899/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-1398899/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":19304489,"identity":"539d8af2-23bf-4ef5-8b36-ea4d64174050","added_by":"auto","created_at":"2022-03-16 18:35:39","extension":"tif","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":11179919,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version\u003c/p\u003e","description":"","filename":"Fig.1.tif","url":"https://assets-eu.researchsquare.com/files/rs-1398899/v1/efd90fcbddf608a531696c56.tif"},{"id":19304490,"identity":"ff252379-9f5a-4c1d-9ddb-d85635978b52","added_by":"auto","created_at":"2022-03-16 18:35:39","extension":"tif","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":168059,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version\u003c/p\u003e","description":"","filename":"Fig.2.tif","url":"https://assets-eu.researchsquare.com/files/rs-1398899/v1/1d739cee5c309bcdc521ff9c.tif"},{"id":19304493,"identity":"28115118-25fa-4ab3-99e6-20f517d5a1c8","added_by":"auto","created_at":"2022-03-16 18:35:39","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":2376164,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version\u003c/p\u003e","description":"","filename":"OnlineFig.3.png","url":"https://assets-eu.researchsquare.com/files/rs-1398899/v1/bd002e00d1493d075ef7deb8.png"},{"id":19304494,"identity":"2abcf4d4-a7a7-4b3a-be8e-70eb7c6819b7","added_by":"auto","created_at":"2022-03-16 18:35:39","extension":"tif","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":15012832,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version\u003c/p\u003e","description":"","filename":"Fig.4.tif","url":"https://assets-eu.researchsquare.com/files/rs-1398899/v1/9a0cb659806b6e4797aa6c05.tif"},{"id":19304549,"identity":"332a2ad3-edef-45fa-9ea8-dee790948e2a","added_by":"auto","created_at":"2022-03-16 18:35:57","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":33549702,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-1398899/v1/75da08e0-54b1-458e-b591-9d1ef35de6af.pdf"},{"id":19304491,"identity":"6aabe181-edac-47c8-b1f6-59770627369a","added_by":"auto","created_at":"2022-03-16 18:35:39","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":15187,"visible":true,"origin":"","legend":"Supplementary Table 1","description":"","filename":"SupplementaryTable1.docx","url":"https://assets-eu.researchsquare.com/files/rs-1398899/v1/43c75d949eca67b4cf6138b9.docx"},{"id":19304492,"identity":"3a11647f-0478-4432-a1e8-6cba039c0546","added_by":"auto","created_at":"2022-03-16 18:35:39","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":28764,"visible":true,"origin":"","legend":"Supplementary Table 2","description":"","filename":"SupplementaryTable2.docx","url":"https://assets-eu.researchsquare.com/files/rs-1398899/v1/019dae8996407e18ae7698f9.docx"},{"id":19304488,"identity":"10671889-e3df-48cc-b272-9a51c054b64c","added_by":"auto","created_at":"2022-03-16 18:35:39","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":14284,"visible":true,"origin":"","legend":"Supplementary Table 3","description":"","filename":"SupplementaryTable3.docx","url":"https://assets-eu.researchsquare.com/files/rs-1398899/v1/27262334fe94e91d59c56f4a.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Grassland-woodland transitions over decadal timescales in the Terai-Duar Savanna and Grasslands of the Indian subcontinent.","fulltext":[{"header":"Introduction","content":"\u003cp\u003eDespite their global significance as a reservoir of biodiversity, tropical grassy biomes (including grasslands and savanna) are being degraded and lost across many regions of the world\u003csup\u003e1\u0026ndash;4\u003c/sup\u003e.\u0026nbsp;Multiple factors, including woodland encroachment, conversion to croplands, the spread of invasive species, climatic changes, grazing by domestic livestock, altered fire regimes, and bio-resource exploitation, are known to be important drivers of grassland loss globally\u003csup\u003e2\u003c/sup\u003e. However, the decline of tropical grassland and savanna has not attracted the same level of conservation attention as the loss of tropical forests\u003csup\u003e2,5\u0026ndash;7\u003c/sup\u003e. This is partly due to the misconception that tropical grassy biomes are anthropogenically derived through the degradation of forests and that forests hold greater ecological and economic value than grassland-dominated ecosystems\u003csup\u003e1,6\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eEncroachment by woody plant species is a significant cause of the transformation of grassland and savanna globally\u003csup\u003e8\u0026ndash;11\u003c/sup\u003e. The conversion of mixed tree-grass plant communities to tree-dominated woodland may profoundly influence the productivity, water balance, and nutrient cycling of these ecosystems. Woody cover appears to be determined by four major environmental variables - water availability, resource availability, fire, and herbivory \u0026ndash; factors that vary over a wide range of climatic conditions\u003csup\u003e10,12\u003c/sup\u003e. \u0026nbsp;In sites where low water availability alone does not limit tree growth, woody cover may still be maintained well below the climatic potential due to a complex interplay between multiple processes such as fire, herbivory, and resource competition\u003csup\u003e10,13,14\u003c/sup\u003e. When environmental changes modify these processes, grasslands and savanna may be transformed to woodland or may even be created by tree cover loss in woodland areas. Such transitions between grassland and woodland dominance influence plant and animal diversity, primary productivity, carbon and nutrient cycling, hydraulic conductivity, and below-ground processes\u003csup\u003e9,11,15\u0026ndash;17\u003c/sup\u003e. In a global-scale empirical study\u003csup\u003e11\u003c/sup\u003e of 224 dryland sites located in all continents (except Antarctica), Soliveres et al. found that plant diversity and ecosystem multifunctionality peaked at intermediate levels of woody cover in these drylands, a relationship that became stronger in wetter sites.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eHowever, anthropogenic effects, including overgrazing, changing rainfall patterns, and increased atmospheric CO\u003csub\u003e2\u003c/sub\u003e concentrations, drive increases in woody species cover and abundance in the world\u0026rsquo;s grassland and savanna regions\u003csup\u003e9,10\u003c/sup\u003e. Woody cover is anticipated to increase further in future environmental change scenarios, with significant consequences for ecosystem structure and function\u003csup\u003e18\u003c/sup\u003e. Although woody encroachment is mainly due to hardy shrubs across arid to sub-humid ecosystems, wetter sites may be experiencing encroachment by trees with a corresponding loss of native grasslands\u003csup\u003e11\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eHere we study the transitions between grass and woody species dominance over decadal timescales in the Terai-Duar Savanna and Grasslands ecoregion\u003csup\u003e19\u003c/sup\u003e (hereafter terai) in the Indian subcontinent. The terai\u0026nbsp;is a unique assembly of alluvial grassland, savanna, and forest ecosystems, which\u0026nbsp;occurs in a narrow belt extending east to west at the base of the Himalayas in the Indian subcontinent\u003csup\u003e20\u003c/sup\u003e. The tropical to subtropical climatic conditions (\u0026gt;1200 mm rainfall per annum) support high productivity and should favour tree dominance. However, there are extensive areas of grasslands and savanna, with varying densities of tree cover, ranging from treeless grassland to savanna (discontinuous tree cover) and dense forest. However, terai savanna and grassland habitats are experiencing woody encroachment and anthropogenic disturbance, threatening terai ecosystems\u003csup\u003e21,22\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWoody species density varies naturally over a wide range of values in mesic environments.\u0026nbsp;For example, in a study\u003csup\u003e23\u003c/sup\u003e of sites across Africa, Australia and South America, Hirota et al. report the presence of three attractors \u0026ndash; treeless grassland, savanna, and forest, with varying tree density. Savanna (with 5 to 60% tree cover) occurred over a wide range of mean annual precipitation, and the frequency distribution of tree cover across all sites was strikingly trimodal. The occurrence of multiple stable states implies that systems can undergo transitions in response to climatic changes and anthropogenic disturbance\u003csup\u003e23\u003c/sup\u003e. In the terai, all three states occur at large and small spatial scales, and any directional shifts towards grass or tree dominance may be driven by environmental changes or disturbance.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo study the transitions between grass and woody species dominance in the terai ecoregion and to test the environmental drivers of the changes, we chose to investigate landcover changes in the Protected Areas (PAs) in the terai in which grasslands and savanna, and woodlands are known to occur. We found eight such sites - four each in India and Nepal - which together account for a total land area of nearly 5000 km\u003csup\u003e2\u003c/sup\u003e (Fig. 1 and Table 1). Although the terai ecoregion is much larger, native terai vegetation is mainly restricted to protected areas where human activities are not permitted by law. Using remotely sensed data, we analysed landcover and landcover changes observed over a 30-year period (1988-1989 through 2018-2019) to characterize the grassland-woodland transitions in these sites. We found that grassland areas declined substantially in the majority of the sites. However, woodland was always the dominant vegetation and accounted for greater area overall (\u0026gt;66%), and the persistence of grasslands (about 29% of the area) appears precarious in the face of encroachment by woodland and conversion to cropland by people. So, more than half the initial grassland area was lost in three sites and over a quarter in two sites (Fig. 2). Grassland area increased in three protected areas (18 to 34%), but of the overall grassland area of 1417 km\u003csup\u003e2\u003c/sup\u003e in 1988-1989 only about 42% remained as grassland in 2018-2019 (Fig. 2). This decline of grasslands was accompanied by increase in woodland area, and of the 826 km\u003csup\u003e2\u003c/sup\u003e of grassland that was lost, about 68% was converted to woodland. The remainder was accounted for by conversion to sand banks and open land, to water bodies (meaning inundated), and about 95 km\u003csup\u003e2\u003c/sup\u003e was converted to cropland.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOur main goal was to test the influence of a series of environmental factors on the stability or persistence of grasslands and grassland-woodland transitions over the 30-year interval (1988-1989 to 2018-2019). Considering 500 m spatial resolution, we tested the influence of four classes of spatial environmental factors \u0026ndash; fire, topography, habitat neighbourhood, and anthropogenic influence \u0026ndash; on the persistence of grassland in each 500m cell. The conversion of grassland areas to woodland has been a matter of concern and in order to test whether any environmental changes drive this conversion, we carried out Bayesian Conditional Autoregressive spatial analyses, with the persistence (or not) of grasslands in each 500m cell over the 30-year interval as the response variable and environmental variables as predictors. We found that the frequency of fire occurrence over the 30-year interval was the most important predictor \u0026ndash; greater fire frequency maintains grasslands. Regression models also show that grasslands were better preserved when they were at greater distance from woodland or cropland sites, which reflect the advancement of woodland at the grassland-woodland ecotone, and that many grassland sites have been converted to cropland by people in nearby settlements, respectively. Topography does significantly influence the persistence of grasslands in the face of change driven by woodland encroachment and other factors. Grasslands persisted better on sites with greater slope and in sites with lower potential for hydrological saturation (lower values of the Topographical Wetness Index), but topographic ruggedness did not affect it.\u003c/p\u003e\n\u003cp\u003eWe chose to restrict our study to protected areas in order to focus on the natural drivers of changes to grasslands. Yet we found that protected areas in this region are not free of anthropogenic threats that affect the habitat and vegetation. First, Himalayan and sub-Himalayan sites are predicted to experience greater rates of climate change than lowland sites. We found that all the protected areas that we studied have seen increases in mean temperatures (range 0.535 - 0.726\u0026deg;C) in the last five decades. In the last decade all sites have had some (up to 20%) dry months (\u0026lt;100mm rain per month) even during the monsoon season. A human footprint index, which is based on range of anthropogenic variables - human settlements, croplands, pasture lands, human population density, night-time lights, railway and road networks, and navigable waterways - increased by about 50% in some Indian protected areas and up to 18% in sites in Nepal during 1993 to 2009. Although we have not been able to link these changes to changes in vegetation, these environmental stressors are potential threats to terai ecosystems. These results provide the impetus for further research on the drivers of grassland-woodland transitions in the terai ecoregion.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eLandcover classification and transitions:\u0026nbsp;\u003c/strong\u003eUsing supervised landcover classification for all eight sites for three time points (1988-1989, 2003-2004, and 2018-2019) in the 30-year period, we found that our classification or prediction accuracy was\u0026nbsp;greater than\u0026nbsp;90%\u0026nbsp;for all the twenty-four classifications (8 sites \u0026times; 3 time points). Grassland\u0026nbsp;and\u0026nbsp;woodland\u0026nbsp;habitats were classified with overall\u0026nbsp;prediction accuracy of 92.2% and 94.3%,\u0026nbsp;respectively.\u0026nbsp;Woodland\u0026nbsp;was the largest landcover type\u0026nbsp;at all times in the terai,\u0026nbsp;and\u0026nbsp;it\u0026nbsp;increased in extent\u0026nbsp;during the last three decades. Including all the protected areas, the woodland cover was 3235 km\u003csup\u003e2\u003c/sup\u003e in 1988-1989, which increased to 3515 km\u003csup\u003e2\u003c/sup\u003e in 2018-2019 (Table 2). Grassland was the second-largest\u0026nbsp;landcover\u0026nbsp;across all the protected areas,\u0026nbsp;but it declined steadily in extent. In 1988-1989, grasslands covered an area of 1417 km\u003csup\u003e2\u003c/sup\u003e across the eight protected areas, but declined to 1209 km\u003csup\u003e2\u0026nbsp;\u003c/sup\u003eby 2003-2004, further reducing to 923 km\u003csup\u003e2\u003c/sup\u003e by 2018-2019 (Table 2 and Supplementary Table 1). However, these changes to grassland extent were not similar across all the PAs. Grassland areas declined in five sites, while there were smaller increases in three sites. The overall decline of grassland area by 34.8%, therefore, does not fully reveal the high dynamism in the individual sites. Thus, in the five sites where the grassland area declined, more than half the grassland area was lost in three sites, Shukla-Phanta NP (51.6%), Dudhwa NP (59.8%), and Chitwan NP (65.6%), and more than a quarter in two sites, Manas NP (40.2%) and Jaldapara NP (26.2%) (Supplementary Table 1). The increases in grassland area in three sites ranged from 18.2% to 34.5%, but two of these sites were smaller in absolute area (Supplementary Table 1).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn the five sites where grassland areas declined, four showed increases in woodland area. However, the increase in a woodland area in each case was smaller than the decline in grassland area. The differences were mainly due to grassland conversion to cropland, but also due to inundation by riverine action (Supplementary Table 2). In Jaldapara NP, both the grassland area and the woodland area declined during this period. In Bardiya NP and Koshi Tappu WS, grassland areas increased during this period, and the increases were comparable to the declines in the woodland area. However, in Valmiki TR, which also showed increase in grassland area (33.4 km\u003csup\u003e2\u003c/sup\u003e), the increase was much smaller than the decline in woodlands (-91 km\u003csup\u003e2\u003c/sup\u003e). So, there are significant differences among sites in the transition from grassland to other land cover types. While Shukla-Phanta, Dudhwa, Chitwan, Jaldapara, and Manas NPs have seen sharp declines in grassland cover in the last three decades (Table 2), in Bardiya NP and Valmiki TR, the extent of grassland increased from 1988-1989 over the following fifteen years but declined in the last fifteen years of our study period (Supplementary Table 1). In Koshi-Tappu WS, grassland cover increased steadily throughout the study period (Supplementary Table 1).\u003c/p\u003e\n\u003cp\u003eWe found that only Dudhwa NP had cropland (15% of its area or 118.0 km\u003csup\u003e2\u003c/sup\u003e) at the beginning of our study (Table 2), but by 2018-2019, all the four protected areas in India had some agricultural land within the park boundaries, accounting to a total of 213 km\u003csup\u003e2\u003c/sup\u003e (Table 2). None of the protected areas in Nepal have seen any conversion to cropland during the entire study period. Human settlements were present only in Dudhwa NP in India, which appears to have been founded during 2004-2005 but grew to cover an area of 32.7 km\u003csup\u003e2\u003c/sup\u003e by 2018-2019 (Fig. 3). Landcover classification maps of the eight protected areas in three periods are presented in figures (Fig. 3 and Fig. 4). \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInfluence of anthropogenic and environmental factors on grassland persistence:\u0026nbsp;\u003c/strong\u003eFirst, we computed the\u0026nbsp;proportion of grassland area in each 500m cell using landcover classification at 30-m spatial resolution. Using a\u0026nbsp;nonspatial\u0026nbsp;binomial\u0026nbsp;regression model of the proportion of grassland area that persisted in\u0026nbsp;each 500m cell\u0026nbsp;(where 0 = no grassland remaining, and 1 = all grassland remaining) over the 30-year period,\u0026nbsp;we found that fire frequency (FIRE_OCCURR) was\u0026nbsp;the\u0026nbsp;most decisive\u0026nbsp;predictor of grassland persistence.\u0026nbsp;Greater fire occurrence had a strong positive influence on the maintenance of grasslands (\u0026beta; = 0.638, p \u0026lt; 0.001) (Table 3). The\u0026nbsp;proportion of grassland that remained\u0026nbsp;was also significantly affected by general\u0026nbsp;anthropogenic influence captured as a distance to human settlements, croplands, or the park boundary\u0026nbsp;(DIST_ANTH), whichever was nearest. Greater distances led to lower persistence of grasslands, as evident in the significant\u0026nbsp;negative binomial coefficient (\u0026beta; = - 0.256,\u0026nbsp;p\u003cem\u003e\u0026lt;\u003c/em\u003e0.001) (Table 3). Being located further away from woodland sites (DIST_WOOD), however, helped maintain grasslands (\u0026beta; = 0.082, p=0.0006), indicating greater rates of transformation at the ecotone between these habitats (Table 3). Topography may influence fire and hydrology and we found that grasslands were more persistent at sites with higher slopes (SLOPE) (\u0026beta; = 0.0970, p=0.0025) and with lower potential for hydrological saturation, where the latter was detected as a significant negative influence of the topographical wetness index (TWI) (\u0026beta; = - 0.0860, p=0.0094) (Table 3). Another topographical feature, ruggedness (RUGG), had no significant influence on grassland persistence (\u0026beta; = - 0.020, p=0.4193) (Table 3). \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe Moran\u0026apos;s \u003cem\u003eI\u0026nbsp;\u003c/em\u003estatistic for residuals of the nonspatial\u0026nbsp;binomial\u0026nbsp;regression was 0.622,\u0026nbsp;and the Monte Carlo permutation test for the presence of spatial autocorrelation in\u0026nbsp;the\u0026nbsp;proportion of grassland conserved\u0026nbsp;per cell was statistically significant (\u003cem\u003eH\u003c/em\u003e\u003csub\u003e0\u003c/sub\u003e: \u003cem\u003eI\u003c/em\u003e = 0, p\u003cem\u003e\u0026lt;\u003c/em\u003e0.001). Therefore, we included spatial autocorrelation as a conditional autoregressive prior in a Bayesian spatial binomial regression model of the proportion of grassland persisting in each 500m cell with the same set of environmental predictors (but now spatial at 500m resolution) to compute the posterior estimated median values and the corresponding 95% credible intervals for the coefficients of all the environmental predictors (Table 4). We found that every single factor that was statistically significant in the non-spatial model was also significant in the Bayesian spatial regression. So, FIRE_OCCURR, SLOPE and DIST_WOOD had significant positive influence on the proportion of grassland conserved in a cell, and DIST_ANTH and TWI had a significant negative influence. The median values of the posterior distributions of the regression parameters show the positive and negative influence of the predictor variables, and statistical significance is indicated where the 95% credible intervals that do not include zero (Table 4). RUGG did not have a statistically significant influence on the proportion of grassland conserved in a cell. As in the non-spatial model, FIRE_OCCURR showed the highest influence (median value 0.6193) on the proportion of grassland conserved per cell (Table 4).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAssessment of threats to the Protected Areas:\u0026nbsp;\u003c/strong\u003eObserved environmental changes already constitute a threat to the PAs in the terai ecoregion, but due to the lack of high-resolution spatial data we were unable to quantify their influence on grassland-woodland transitions. All the PAs have experienced an increase in mean temperature (range 0.535 - 0.726\u0026deg;C) over a span of five decades, with Bardiya NP recording the highest increase (Table 5). In the last decade (2011-2020), all the protected areas have experienced some dry months (\u0026lt;100 mm rain per month) even during the monsoon. Valmiki TR and Chitwan NP, which occur adjacently across the boundary between India and Nepal, have seen drought in 20% (8 out of 40) of the monsoonal months in the last decade (Table 5), while others have had fewer (2 to 6) dry months during the same decade. Examining other human impacts using the human footprint index - based the extent of applicable anthropogenic variables including human settlements, croplands, pasture lands, human population density, night-time lights, railway and road networks, and navigable waterways - we found that between 1993 to 2009, anthropogenic effects increased by more than 50% in Dudhwa and Manas NPs (Table 5). The smallest impact with just a 9% increase in human footprint was seen in Bardiya NP. Protected areas in Nepal experienced comparatively smaller increases (17.8% on average) in human footprint compared to the average 40.9% increase in human footprint in the Indian protected areas (Table 5).\u0026nbsp;\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe decline of grassy biomes due to encroachment by woody species is a problem of global significance\u003csup\u003e24,25\u003c/sup\u003e. The trend towards increased shrub and tree abundance in grassland sites have been widely documented since decades - in North America\u003csup\u003e26,27\u003c/sup\u003e, South America\u003csup\u003e28\u003c/sup\u003e, Africa\u003csup\u003e29\u003c/sup\u003e, Australia\u003csup\u003e30\u003c/sup\u003e, India\u003csup\u003e31\u003c/sup\u003e, but most reports leading up to more recent studies are about shrub encroachment in dryland sites\u003csup\u003e16,17,32,33\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eWe studied grassland-woodland transitions in the unique Terai-Duar Savanna and Grassland ecoregion in the sub-Himalayan region of the Indian subcontinent, which lies in the sub-humid to humid zones but nevertheless hosts significant areas under grassland. While the presence of grasslands in this region is itself an intriguing ecological pattern that is not well understood, observations of decline in grassland habitat in the face of changing environmental conditions and increasing human impacts has become a matter of increasing concern. Although the climatic conditions prevalent in the region are expected to favor woody vegetation, grassland and savanna support much of the biodiversity, particularly of the endemic species. Grassland areas declined overall by about 35% of its original extent during the 30-year period, but the losses occurred in five of the eight sites that we studied, while it increased marginally in three sites. This suggests that grassland decline is perhaps driven by variation in local environmental conditions. The presence of dry season grass fires was the most important factor that affected the persistence of grasslands across all the sites, and fire regimes do vary significantly among sites depending on local conditions\u003csup\u003e34,35\u003c/sup\u003e. In general, the decreases in grassland areas were accompanied by increases in the woodland areas, which was the most common transformation of grasslands in the terai. We would need to examine historical changes in the entire terai ecoregion in a more comprehensive study to test whether woodland encroachment has been the dominant trend across this ecoregion. But our analyses of best protected conservation areas in the terai already suggest that woodland encroachment is a threat to grassland habitat in this region. The greater persistence of grasslands in spatial proximity to anthropogenic influence may appear counter-intuitive, but it may be caused by the influence of human activities on fire occurrence. Anthropogenic influence, through climatic changes and direct impacts (measured as a human footprint), have been increasing across all sites in the terai, but we lack high-resolution spatial-temporal data on climatic changes to quantify their influence on grassland-woodland transitions.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Our landcover classification was highly accurate, and we found that woodland (mostly forest, but also early-successional woodland) is the dominant landcover class overall. The dominance of woody vegetation is expected given the climatic conditions, but the presence of a high biodiversity of grassland specialist plant and animal taxa suggest the long-term importance and prevalence of grassland habitats in this ecoregion. The smaller extent of grasslands and the highly variable woodland versus grassland proportions that we observed in the eight sites, may be due to more recent anthropogenic influence within the last century or more. For example, Manas NP, had twice as much grassland area compared to the woodland area as recently as 1988-1989 and it is among the wetter sites. Currently, Manas NP and Shukla-Phanta NP are the only parks with more grassland than woodland. The trends that we see in the last 30 years may have persisted for longer, and may continue into the future. It is noteworthy that despite considering all the protected habitats which have substantial grasslands in the terai ecoregion, we found only 1417 km\u003csup\u003e2\u003c/sup\u003e of grassland habitat in 1988-1989, which further reduced to 923 km\u003csup\u003e2\u003c/sup\u003e by 2018-2019. This loss of grassland has not drawn adequate conservation attention and needs to be addressed. In sites that saw declines in grassland, the reduction was large, with more than 50% loss in three sites, and this occurred within three decades. Grassland is therefore being lost rapidly in key terai sites, and there is a clear need to investigate the drivers of these changes at each site in detail.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;The causes of woodland encroachment are often broken down into three hierarchical scales of drivers: (1) global-scale drivers such as elevated CO\u003csub\u003e2\u003c/sub\u003e, (2) regional-scale climatic drivers (e.g., precipitation and temperature changes), and (3) local-scale drivers such as land management history, changes in fire frequencies, land fragmentation, and removal of native herbivores\u0026nbsp;\u003csup\u003e25,36\u0026ndash;39\u003c/sup\u003e. Globally, each biome that is undergoing woody encroachment may have a suite of these interacting drivers that influence the rate of woody encroachment. We tried to examine the influence of a set of regional and local drivers on the persistence of grasslands in the terai. However, we did not have the high spatial resolution in the time series data on temperature and rainfall, nor did we have a large number of sites, to quantify the influence of regional climatic drivers on the differences in vegetation changes we observed. Mean temperature increase and changes in the drought index (SPI) were different among sites, but since we had just eight sites, we could not reliably address how the observed climatic changes within the Terai region have influenced grassland persistence. We did not also detect any consistent relationship with the total precipitation or seasonality and the proportion of grassland in the site. For example, Shukla-Phanta NP and Manas NP had the lowest and highest mean annual rainfall, respectively, but they also had a greater proportion of grassland area than other sites (Table 1 and Supplementary Table 1). Bardiya NP, Valmiki TR, and Chitwan NP had mean annual rainfall in the range 1310 to 1946 mm, and all had low proportions of the area under grassland. It follows, therefore, that discerning the influence of future climatic changes on the vegetation in this region is likely to be complex. At these scales, the ecological mechanisms behind the dominance of grasses or woody species are known to be associated with the interactions between precipitation amount, seasonality, and soil properties; however, all these factors may interact with the occurrence of fire, herbivory, or hydrological attributes in a particular site. Despite high annual rainfall, the presence of a long and strong dry season can promote fires and influence vegetation physiognomy, and this appears to be the dominant force shaping terai vegetation. \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Local hydrology can influence the fire regime and soil water profiles in each site, thereby affect vegetation physiognomy at a site\u003csup\u003e23,40\u003c/sup\u003e. The levels and duration of inundation due to monsoon floods varies within terai habitats and may create relatively swampy or dry edaphic conditions. Soil infiltration rates vary between fine-textured (clayey) and coarse-textured (sandy) soils, affecting the soil water profile and thereby the balance of woody versus grassy vegetation\u003csup\u003e41,42\u003c/sup\u003e. Some parts of the terai certainly have swampy conditions that some grasses can tolerate but woody plants may fail to establish. Here, due to the wet conditions that may persist through the dry season, even fire occurrence may be low compared to dry-grassland sites\u003csup\u003e35\u003c/sup\u003e. In dry-grassland sites, regular dry season fires are more probable, and the occurrence of fire and persistence of grassland extent appears to be involved in a positive feedback cycle\u003csup\u003e35\u003c/sup\u003e. The occurrence of fires in terai ecosystems is largely due to anthropogenic influence emanating from land use activities by local people and management interventions by the forest department. We found that greater anthropogenic proximity (DIST_ANTH) led to greater persistence of grassland over time. We surmise that anthropogenic proximity has the greatest influence on fire. Since these are protected areas, grazing by domestic livestock and resource harvest is not legally permitted, but such activities may be occurring at low intensity in some sites and may affect encroachment by woody vegetation\u003csup\u003e43,44\u003c/sup\u003e. To a small extent, grassland loss has also been due to conversion to cropland, and cropland area has expanded from the edges of human settlements, contributing to the influence of anthropogenic proximity. Such anthropogenic effects appear to be greater at the Indian sites, where population density and pressure is greater compared to that in Nepal.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Our study does not explicitly consider several important factors that are known to affect grassland and woodland composition. While fire may be sufficient to maintain grasslands in sites where the climatic conditions favour woody vegetation, other factors such as the actions of large mammalian herbivores, hydrological attributes, and soil properties may also be important in any site. We did not have such detailed data for the Terai region to carry these tests, but future studies in each site must consider the entire suite of variables that determine grassland presence in terai ecosystems. Fire may well be critical for maintaining grasslands in the terai, and proper fire management is needed to ensure the desired distributions of habitats for wildlife species, particularly the region\u0026apos;s endangered and endemic grassland specialists. Nevertheless, the increasing fragmentation of terai ecosystems may also be altering the influence of hydrology and long-term ecological processes that maintain the diversity and complexity of terrain vegetation.\u0026nbsp;\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eStudy Area:\u0026nbsp;\u003c/strong\u003eThe extensive lowland region of floodplain grasslands, woodland-grassland mixtures, and forests located at the southern base of the Himalayan Mountains in India and Nepal is known as the Terai-Duar Savanna and Grasslands ecoregion\u003csup\u003e21\u003c/sup\u003e. It is a narrow belt of alluvial grasslands and savannas with patches of forests stretching from Yamuna River in the west to the Brahmaputra River in the east\u003csup\u003e20\u003c/sup\u003e. Typically, the alluvial grassland is dominated by tall grasses \u003cem\u003elike Themeda arundinacea, Saccharum narenga, Phragmites karka, Imperata cylindrica,\u0026nbsp;\u003c/em\u003e\u003cem\u003eChrysopogon zizanioides\u003c/em\u003e and several other species. The terai is prime habitat for the flagship species like the tiger (\u003cem\u003ePanthera tigris\u003c/em\u003e), the Asian elephant (\u003cem\u003eElephas maximus\u003c/em\u003e) and critical habitat for the greater one-horned rhinoceros (\u003cem\u003eRhinoceros unicornis\u003c/em\u003e)\u003csup\u003e20\u003c/sup\u003e, and several grassland specific endangered and endemic species like the hispid hare (\u003cem\u003eCaprolagus hispidus\u003c/em\u003e), pygmy hog (\u003cem\u003ePorcula salvania\u003c/em\u003e) and Bengal florican (\u003cem\u003eHoubaropsis bengalensis\u003c/em\u003e)\u003csup\u003e45,46\u003c/sup\u003e. This region has a monsoon-influenced sub-humid to humid subtropical climate with significantly more rain in the wet summer months (April \u0026ndash; September) than in the dry winter months (October \u0026ndash; March)\u003csup\u003e47\u003c/sup\u003e. While the western part of the terai receives a mean annual rainfall of about 1200-1300 mm, the eastern part is much wetter with over 2400-2500 mm of rain. The mean annual temperature range is 20 - 28 \u0026deg;C, with mean monthly minimum and maximum temperatures (30-year average) being 5.5 \u0026deg;C and 40.5 \u0026deg;C, respectively \u003csup\u003e47\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eTerai grasslands are highly threatened, and most of them have already been converted to other land uses. The largest areas of natural habitat are confined to a few protected areas\u003csup\u003e20\u003c/sup\u003e. These eight protected areas that we study here are prominent sites in India and Nepal and represent typical terai sites with extensive grassland and grassland-woodland mixtures (Fig. 1 and Table 1).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLandcover Classification:\u0026nbsp;\u003c/strong\u003eThe vegetation across the eight protected areas is typically a mosaic of forests, woody savanna, and pure grasslands, of which moist deciduous forest and alluvial grassland are most notable. The tree species in forests are characteristic of both semi-evergreen forests and moist deciduous forests. Grasslands are broadly two distinct types - the swampy and dry grasslands. The woody savanna formations occur mostly at the ecotone of forest and grassland with a few early successional trees such as \u003cem\u003eBombax ceiba\u003c/em\u003e, \u003cem\u003eDillenia\u003c/em\u003e \u003cem\u003epentagyna\u003c/em\u003e, and \u003cem\u003eLagerstroemia parviflora\u003c/em\u003e\u003csup\u003e48\u003c/sup\u003e. All the eight protected areas that we studied exhibit a composite of the vegetation formations mentioned above. However, to keep the classification simple, and because of the difficulty of distinguishing the forest from woody savanna accurately through remote sensing, we considered only two broad vegetation types: grassland, which represent pure grassland and perhaps sites with very low tree density (\u0026lt;10% canopy cover) and woodland as representing woody savanna and forest. For completeness, we needed to consider four additional landcover types: water bodies, sand and open land, cropland, and human settlements. \u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRemotely-sensed data:\u0026nbsp;\u003c/strong\u003eLandsat imageries are convenient for detecting land-cover change\u003csup\u003e49\u003c/sup\u003e and forest-cover dynamics in grasslands and forested areas\u003csup\u003e50\u0026ndash;52\u003c/sup\u003e. We acquired Landsat 4-5 Thematic Mapper (TM), Landsat 7 Enhanced Thematic Mapper Plus (ETM+), and Landsat 8 Operational Land Imager (OLI) images at 30m spatial resolution from the U.S. Geological Survey. After the thorough screening, we selected the post-monsoon dry season (December-March) images to ensure cloud-free coverage. For all three times points (1988-1989, 2003-2004, and 2018-2019), one single Landsat tile was sufficient to cover the full extent of Shukla-Phanta, Dudhwa and Bardiya NPs. However, we needed two tiles for Valmiki Tiger Reserve and Chitwan National Park also for the 2003-2004 data for Manas NP. Landsat 7 ETM+ products have scanline errors in post-May 2004 imageries, so the Landsat 7 ETM+ images we used in the study were selected with the acquisition date before May 2004. The detailed information of the obtained Landsat images used for the landcover classification is shown in Supplementary Table 3.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTraining and Validation Sample Data\u003c/strong\u003e\u003cstrong\u003e:\u0026nbsp;\u003c/strong\u003eWe mostly collected the training and validation data through visual inspection from Google Earth imagery, but we also used real ground verification data for Manas NP in India. The practice of generating training data for landcover classification through manual visual interpretation of high spatial resolution images from Google Earth imageries is widely applied and reported in the literature\u003csup\u003e53\u0026ndash;55\u003c/sup\u003e. We focused on Google Earth imagery separately for the three periods, 1988-1989, 2003-2004 and 2018-2019, and selected a set of point locations to serve as training data for six land cover classes. For the 2018-2019 period, we also collected ground data through extensive field surveys at Manas NP in India. Using a GPS unit, we gathered ground truth information for at least 20 locations for each landcover type located within the park boundary. The combined data were then used as training inputs for the classifier. The number of training and validation samples for each protected area was different, depending on the land cover class and the size of the protected area, but we ensured sufficient sample size in each case.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupervised Landcover Classification:\u0026nbsp;\u003c/strong\u003eWe performed supervised land-cover classification using \u0026apos;Random Forests\u0026apos; (RF), an ensemble-based decision-tree classification algorithm\u003csup\u003e56\u003c/sup\u003e. RF is a well-known classification method where several regression trees are generated, and a consensus tree is derived by averaging the predictions of the individual trees\u003csup\u003e57\u003c/sup\u003e. The RF algorithm has been widely preferred for land cover classification using satellite imagery because of its higher classification accuracy\u003csup\u003e58,59\u003c/sup\u003e. Although other classification algorithms are also available, we reasoned that RF was most appropriate for our objective. For classification, we built the predictive regression model using eight predictor variables. These include six Landsat bands (Blue, Green, Red, Near Infrared, Short-wave Infrared I and Short-wave Infrared II) and two vegetation indices (Normalized Difference Vegetation Index and Normalized Difference Moisture Index). For each classification, we used the \u0026apos;confusion matrix\u0026apos; reports for accuracy assessment. After accuracy assessment using the validation data, we derived the classified images from the predictions of the RF model and then used the images for quantifying landcover transitions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInfluence of anthropogenic and environmental factors on Grassland Habitat:\u0026nbsp;\u003c/strong\u003eOur landcover analysis was constrained to start at 1988-1989 due to the availability of satellite data. We identified the grassland area for 1988-1989 from the classified images and generated the baseline grassland polygons for all eight sites. We then placed a 500 \u0026times; 500 m\u003csup\u003e2\u003c/sup\u003e grid to produce non-overlapping areas over the grassland polygons and overlaid the 30-m landcover classification map of 2018-2019. This allowed us to compute the percentage of grassland area that persisted as grassland per grid cell at the end of 30 years. We also note that because of the irregular shape of the grassland polygons and the park boundary, not all grid cells were exactly 500 \u0026times; 500 m\u003csup\u003e2\u003c/sup\u003e in size and had different numbers of pixels. Therefore, quantifying the percentage of pixels transformed within each grid cell was a robust measure for spatial data analyses. More generally, non-overlapping areas (cells) of any shape can be placed on the study area in such analyses. We thus derived spatial data for grassland transitions as the proportion of pixels in a cell that remained as grassland at the end of the study period.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo understand the environmental determinants of the observed spatial pattern of grassland transitions, we derived several independent spatial environmental variables for testing their influence on grassland persistence across the terai sites. Using ArcMap 10.8 software, we assembled the following variables:\u0026nbsp;(i) \u003cem\u003eFire occurrence:\u003c/em\u003e The near real-time Moderate Resolution Imaging Spectroradiometer (MODIS) provides \u0026apos;active fire\u0026apos; data (with data code MCD14DL), which gives the point location of active fire with the exact time and date of detection\u003csup\u003e60\u003c/sup\u003e. These data are available from October 2000 and can be obtained from NASA\u0026apos;s automated Fire Information for Resource Management System (FIRMS). Here we used the fire occurrence data of 19 years (October 2000 to September 2019). We delineated a rectangular area around the geographical boundary of each protected area, with a buffer of 1 km from the park boundary and extracted all the active fire incidents within this rectangular area for 19 years. Then, we counted the number of fire events (FIRE_OCCURR) within each grassland grid cell to derive the fire counts per cell in 19 years.\u003c/p\u003e\n\u003cp\u003e(ii) \u003cem\u003eTopographic features\u003c/em\u003e: We obtained the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) Global Digital Elevation Model Version 3 (GDEM 003) elevation data with a spatial resolution of 30 m for the Terai region. Using the \u0026apos;spatial analyst\u0026apos; and \u0026apos;raster calculator\u0026apos; tools in ArcMap, we generated the Slope (SLOPE), Topographic Wetness Index (TWI)\u003csup\u003e61\u003c/sup\u003e and Ruggedness (RUGG) rasters following standard protocols and upscaled them to derive the mean value for each 500 m grassland cell. We use these topographical attributes because they known to affect soil water retention and the potential for hydrological saturation, thus influencing small-scale variation in water availability.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(iii) Woodland proximity: Human-induced alteration of grazing and fire regimes may promote the replacement of grassland by woodland\u003csup\u003e62\u003c/sup\u003e. However, the scope of ecological succession may not be uniform everywhere. A higher richness of soil seed bank and a favourable physical environment is expected in grassland sites near the forest than sites farther from the forest edge \u003csup\u003e63\u003c/sup\u003e. So, we considered the distance to the nearest woodland patch (DIST_WOOD) as a potential factor. The 1988-1989 period woodland coverage was considered for the proximity calculation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(iv) Anthropogenic proximity: Transformation of grassland to agriculture has been occurring across the world. WE expect that the prospect of grassland conversion near existing farmland and human settlement is greater than for distant grassland areas. So, we attempted to compute the distance to the nearest cropland area for each grassland cell, but cropland did not exist in some protected areas in 1988-1989. In the absence of the cropland within the boundary, we considered the distance to the human settlement or boundary of the protected area as the proxy for anthropogenic proximity. We obtained anthropogenic proximity (DIST_ANTH) for each grassland cell by calculating the distance from the cropland or human settlement or the protected area boundary (depending on which one is the nearest).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConditional Autoregressive Bayesian Spatial model:\u0026nbsp;\u003c/strong\u003eWe calculated the proportion of grassland that persists (\u003cem\u003eY\u003csub\u003ek\u003c/sub\u003e\u003c/em\u003e) for a 500-m grassland cell based on the number of 30-m pixels (\u003cem\u003en\u003csub\u003ek\u003c/sub\u003e\u003c/em\u003e) that were transformed within the cell. If all the pixels in a cell remain as grassland, then it gets a maximum value of 1, and if the cell is completely transformed, it is assigned a minimum value of 0. When only a subset of pixels is transformed, the corresponding proportional value is assigned. As environmental predictors of grassland persistence, we used six variables - FIRE_OCCURR, SLOPE, TWI, RUGG, DIST_WOOD, and DIST_ANTH. These variables are either relatively stable over time (e.g., SLOPE) or were cumulative values over the study period (e.g., FIRE_OCCUR).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eConsidering the grassland cells in all the protected areas together, we first fitted a nonspatial binomial model that incorporates covariate effects at the cell scale and assessed spatial autocorrelation by examining the residuals of the regression. In the statistical framework, we assume that the study region is partitioned into \u003cem\u003ek\u003c/em\u003e non-overlapping areal cells (500 \u0026times; 500 m\u003csup\u003e2\u003c/sup\u003e in our case), which are linked to a corresponding set of responses \u003cem\u003eY\u003csub\u003e1\u003c/sub\u003e, . . . , Y\u003csub\u003ek\u003c/sub\u003e\u0026nbsp;\u003c/em\u003eand a vector of known offsets \u003cem\u003eO\u003csub\u003e1\u003c/sub\u003e, . . . , O\u003csub\u003ek\u003c/sub\u003e\u003c/em\u003e. As explained above, each \u003cem\u003eY\u003csub\u003ei\u003c/sub\u003e\u003c/em\u003e (\u003cem\u003ei\u003c/em\u003e =1, . . . , \u003cem\u003ek\u003c/em\u003e) represents the proportion of pixels in the \u003cem\u003ei\u003c/em\u003e\u003csup\u003eth\u003c/sup\u003e cell that persists as grassland at the end of the study period. The covariate values were scaled before fitting the model. In this binomial model, the number of pixels in cell \u003cem\u003ek\u003c/em\u003e (\u003cem\u003en\u003csub\u003ek\u003c/sub\u003e\u003c/em\u003e) is the number of trials for the \u003cem\u003ek\u003c/em\u003e\u003csup\u003eth\u003c/sup\u003e cell, while \u003cem\u003e\u0026theta;\u003c/em\u003e\u003cem\u003e\u003csub\u003ek\u003c/sub\u003e\u003c/em\u003e is the probability of success in a single trial\u003csup\u003e64\u003c/sup\u003e. The model therefore was:\u003c/p\u003e\n\u003cp\u003e\u003cimg src=\"data:image/png;base64,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\" style=\"width: 597px;\"\u003e\u003c/p\u003e\n\u003cp\u003eWhere,\u003cem\u003e\u0026nbsp;\u003cimg src=\"data:image/png;base64,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\" style=\"width: 28px;\"\u003e\u003c/em\u003e \u003cem\u003e\u0026nbsp;\u003c/em\u003eis the transpose of the matrix of the explanatory spatial covariates. A spatial structure component\u003cem\u003e\u0026nbsp;\u003c/em\u003e(\u003cem\u003e\u0026psi;\u003csub\u003ek\u003c/sub\u003e)\u003c/em\u003e is included to model any spatial autocorrelation present in the data. The vector of regression parameters is denoted by \u003cem\u003e\u0026beta; = (\u0026beta;\u003csub\u003e1\u003c/sub\u003e,...,\u0026beta;\u003csub\u003ek\u003c/sub\u003e)\u003c/em\u003e \u003csup\u003e64\u003c/sup\u003e. We first performed the nonspatial binomial regression using the \u0026apos;glm\u0026apos; function in the R statistical software\u0026apos;s \u0026apos;stats\u0026apos; package \u003csup\u003e65\u003c/sup\u003e. To quantify spatial autocorrelation in the residuals, we used a Monte Carlo test and compared the observed Moran\u0026apos;s \u003cem\u003eI\u0026nbsp;\u003c/em\u003ewith the null distribution of Moran\u0026apos;s \u003cem\u003eI\u0026nbsp;\u003c/em\u003evalues obtained under no autocorrelation. This null distribution of Moran\u0026apos;s \u003cem\u003eI\u0026nbsp;\u003c/em\u003ewas computed after randomly permuting the grassland proportion data across the grid cells in the landscape 1000 times and computing Moran\u0026apos;s \u003cem\u003eI\u0026nbsp;\u003c/em\u003efor each permutation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFinally, we performed Bayesian spatial modelling of the proportion of grassland that persisted per cell using a Conditional Autoregressive (CAR) prior, which induces spatial autocorrelation through the adjacency structure of the areal units, and implemented this in the R package \u0026apos;CARBayes\u0026apos; \u003csup\u003e64\u003c/sup\u003e. The linear predictor was the set of environmental variables indicated above for the binomial response and was treated as random effects. We implemented the function \u0026apos;S.glm\u0026apos; within the R package CARBayes to fit the Bayesian regression model with autocorrelation specified as shown above. S.glm fits the model with no random effects and is an appropriate conditional autoregressive prior for the binomial variables \u003csup\u003e64\u003c/sup\u003e. We used a Markov Chain Monte Carlo (MCMC) based simulation of the posterior distributions of model parameters. The simulation involved 10,000 MCMC samples generated from a single Markov chain, executed for 300,000 iterations with a burn-in of 100,000 and then thinned by 20 to reduce Markov chain autocorrelation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAssessment of threats to the Protected Areas:\u0026nbsp;\u003c/strong\u003eWe computed the threat level in three categories for our eight protected areas of interest: habitat loss, anthropogenic influence, and climatic stressors:\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e(i)\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eHabitat loss: We focused only on the threats to the grassland habitat that resulted in grassland loss. Here, we calculated the change in the grassland cover over 30 years, starting with the grassland extent observed in 1988-1989.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(ii) Anthropogenic influence: To measure the extent of anthropogenic influence on the protected areas, we obtained the global map of human footprint (HFP) from https://wcshumanfootprint.org. The human footprint map is prepared based on human settlements, croplands, pasture lands, human population density, night-time lights, railway and road networks, and navigable waterways\u0026nbsp;\u003csup\u003e66\u003c/sup\u003e. It captures the cumulative human pressure on the environment. The human footprint maps of two time points (1993 and 2009) are available at present. We obtained the maps of both periods and computed the change in human pressure within each protected area over the 16-year period. We could not cover the entire 30-year study period due to the lack of data. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(iii) Climatic stressors: We examined the average temperature change and rainfall anomalies to understand the climatic threat for the protected areas. We obtained climatic data for 60 years (1961-2020) from the Climate Research Unit (CRU TS v. 4.04) database (\u003ca href=\"http://www.cru.uea.ac.uk/data\"\u003ehttp://www.cru.uea.ac.uk/data\u003c/a\u003e)\u003csup\u003e67\u003c/sup\u003e. We calculated the average temperature for two decadal periods, 1961-1970 and 2011-2020, to measure climatic changes over half a century. We also computed a drought index called the Standardized Precipitation Index (SPI) using monthly rainfall data cumulated over two months\u003csup\u003e68\u003c/sup\u003e. The terai receives ample rainfall between June to September, but climate change may have impacted rainfall distribution. For any month, an SPI value less than -1.5 indicates severe drought. Here, we calculated the number of monsoonal months in the last decade (total forty months, four months for ten years, 2011-2020) that have scored below -1.5 to understand the rainfall anomalies.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by Grant No. EMR/2017/002769 awarded to RJ by SERB-DST, Government of India and Rufford Small Grant No. 21980-1 awarded to SB by Rufford Foundation. We thank the Assam Forest Department for support on field visits. We also want to thank Mr Swapnil Bhowal for helping us to obtain the historical images from Google Earth to carry out the landcover classification of the past.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSubham Banerjee:\u003c/strong\u003e Conceptualization, Methodology, Data curation, Investigation, Writing- Original draft preparation. \u003cstrong\u003eDhritiman Das:\u003c/strong\u003e Conceptualization, Data curation. \u0026nbsp;\u003cstrong\u003eRobert John:\u003c/strong\u003e Supervision, Validation, Writing- Reviewing and Editing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003cp\u003e1.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Bond, W. J. \u0026amp; Parr, C. L. Beyond the forest edge: Ecology, diversity and conservation of the grassy biomes. \u003cem\u003eBiological Conservation\u003c/em\u003e \u003cstrong\u003e143\u003c/strong\u003e, 2395\u0026ndash;2404 (2010).\u003c/p\u003e\n\u003cp\u003e2.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Bardgett, R. D. \u003cem\u003eet al.\u003c/em\u003e Combatting global grassland degradation. \u003cem\u003eNat Rev Earth Environ\u003c/em\u003e (2021) doi:10.1038/s43017-021-00207-2.\u003c/p\u003e\n\u003cp\u003e3.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Gibbs, H. K. \u0026amp; Salmon, J. M. Mapping the world\u0026rsquo;s degraded lands. \u003cem\u003eApplied Geography\u003c/em\u003e \u003cstrong\u003e57\u003c/strong\u003e, 12\u0026ndash;21 (2015).\u003c/p\u003e\n\u003cp\u003e4.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Gang, C. \u003cem\u003eet al.\u003c/em\u003e Quantitative assessment of the contributions of climate change and human activities on global grassland degradation. \u003cem\u003eEnviron Earth Sci\u003c/em\u003e \u003cstrong\u003e72\u003c/strong\u003e, 4273\u0026ndash;4282 (2014).\u003c/p\u003e\n\u003cp\u003e5.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Overbeck, G. E. \u003cem\u003eet al.\u003c/em\u003e Conservation in Brazil needs to include non-forest ecosystems. \u003cem\u003eDiversity Distrib.\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 1455\u0026ndash;1460 (2015).\u003c/p\u003e\n\u003cp\u003e6.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Parr, C. L., Lehmann, C. E. R., Bond, W. J., Hoffmann, W. A. \u0026amp; Andersen, A. N. Tropical grassy biomes: misunderstood, neglected, and under threat. \u003cem\u003eTrends in Ecology \u0026amp; Evolution\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 205\u0026ndash;213 (2014).\u003c/p\u003e\n\u003cp\u003e7.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Murphy, B. P., Andersen, A. N. \u0026amp; Parr, C. L. The underestimated biodiversity of tropical grassy biomes. \u003cem\u003ePhil. Trans. R. Soc. B\u003c/em\u003e \u003cstrong\u003e371\u003c/strong\u003e, 20150319 (2016).\u003c/p\u003e\n\u003cp\u003e8.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Van Auken, O. W. Shrub Invasions of North American Semiarid Grasslands. \u003cem\u003eAnnu. Rev. Ecol. Syst.\u003c/em\u003e \u003cstrong\u003e31\u003c/strong\u003e, 197\u0026ndash;215 (2000).\u003c/p\u003e\n\u003cp\u003e9.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Schlesinger, W. H. \u003cem\u003eet al.\u003c/em\u003e Biological Feedbacks in Global Desertification. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e247\u003c/strong\u003e, 1043\u0026ndash;1048 (1990).\u003c/p\u003e\n\u003cp\u003e10.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Sankaran, M., Ratnam, J. \u0026amp; Hanan, N. Woody cover in African savannas: the role of resources, fire and herbivory. \u003cem\u003eGlobal Ecol Biogeography\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 236\u0026ndash;245 (2008).\u003c/p\u003e\n\u003cp\u003e11.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Soliveres, S. \u003cem\u003eet al.\u003c/em\u003e Plant diversity and ecosystem multifunctionality peak at intermediate levels of woody cover in global drylands. \u003cem\u003eGlobal Ecology and Biogeography\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 1408\u0026ndash;1416 (2014).\u003c/p\u003e\n\u003cp\u003e12.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Frost, P. \u003cem\u003eet al.\u003c/em\u003e Responses of Savannas to Stress and Disturbance. Biology International, Special Issue 10. \u003cem\u003eInternational Union of Biological Sciences, Paris\u003c/em\u003e (1986).\u003c/p\u003e\n\u003cp\u003e13.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Sankaran, M. \u003cem\u003eet al.\u003c/em\u003e Determinants of woody cover in African savannas. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e438\u003c/strong\u003e, 846\u0026ndash;849 (2005).\u003c/p\u003e\n\u003cp\u003e14.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Bond, W. J., Midgley, G. F. \u0026amp; Woodward, F. I. What controls South African vegetation \u0026mdash; climate or fire? \u003cem\u003eSouth African Journal of Botany\u003c/em\u003e \u003cstrong\u003e69\u003c/strong\u003e, 79\u0026ndash;91 (2003).\u003c/p\u003e\n\u003cp\u003e15.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Ratajczak, Z., Nippert, J. B. \u0026amp; Collins, S. L. Woody encroachment decreases diversity across North American grasslands and savannas. \u003cem\u003eEcology\u003c/em\u003e \u003cstrong\u003e93\u003c/strong\u003e, 697\u0026ndash;703 (2012).\u003c/p\u003e\n\u003cp\u003e16.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Quero, J. L., Maestre, F. T., Ochoa, V., Garc\u0026iacute;a-G\u0026oacute;mez, M. \u0026amp; Delgado-Baquerizo, M. On the Importance of Shrub Encroachment by Sprouters, Climate, Species Richness and Anthropic Factors for Ecosystem Multifunctionality in Semi-arid Mediterranean Ecosystems. \u003cem\u003eEcosystems\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 1248\u0026ndash;1261 (2013).\u003c/p\u003e\n\u003cp\u003e17.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Eldridge, D. J., Soliveres, S., Bowker, M. A. \u0026amp; Val, J. Grazing dampens the positive effects of shrub encroachment on ecosystem functions in a semi-arid woodland. \u003cem\u003eJ Appl Ecol\u003c/em\u003e \u003cstrong\u003e50\u003c/strong\u003e, 1028\u0026ndash;1038 (2013).\u003c/p\u003e\n\u003cp\u003e18.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Kulmatiski, A. \u0026amp; Beard, K. H. Woody plant encroachment facilitated by increased precipitation intensity. \u003cem\u003eNature Clim Change\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, 833\u0026ndash;837 (2013).\u003c/p\u003e\n\u003cp\u003e19.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Olson, D. M. \u003cem\u003eet al.\u003c/em\u003e Terrestrial Ecoregions of the World: A New Map of Life on Earth. \u003cem\u003eBioScience\u003c/em\u003e \u003cstrong\u003e51\u003c/strong\u003e, 933 (2001).\u003c/p\u003e\n\u003cp\u003e20.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Semwal, R. L. \u003cem\u003eThe Terai arc landscape in India: Securing protected areas in the face of global change\u003c/em\u003e. (Forests \u0026amp; Biodiversity Conservation Programme, World Wide Fund for Nature\u0026nbsp;\u0026hellip;, 2005).\u003c/p\u003e\n\u003cp\u003e21.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Dinerstein, E. An ecological survey of the royal Karnali-Bardia Wildlife Reserve, Nepal. Part I: Vegetation, modifying factors, and successional relationships. \u003cem\u003eBiological Conservation\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 127\u0026ndash;150 (1979).\u003c/p\u003e\n\u003cp\u003e22.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Sharma, B. Wildlife habitat mapping by using Geographic Information Systems (GIS) in the Karnali floodplain of Royal Bardia National Park at lowland Nepal. (1999).\u003c/p\u003e\n\u003cp\u003e23.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Hirota, M., Holmgren, M., Van Nes, E. H. \u0026amp; Scheffer, M. Global Resilience of Tropical Forest and Savanna to Critical Transitions. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e334\u003c/strong\u003e, 232\u0026ndash;235 (2011).\u003c/p\u003e\n\u003cp\u003e24.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Archer, S. R. \u003cem\u003eet al.\u003c/em\u003e Woody Plant Encroachment: Causes and Consequences. in \u003cem\u003eRangeland Systems\u003c/em\u003e (ed. Briske, D. D.) 25\u0026ndash;84 (Springer International Publishing, 2017). doi:10.1007/978-3-319-46709-2_2.\u003c/p\u003e\n\u003cp\u003e25.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Stevens, N., Lehmann, C. E. R., Murphy, B. P. \u0026amp; Durigan, G. Savanna woody encroachment is widespread across three continents. \u003cem\u003eGlob Change Biol\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 235\u0026ndash;244 (2017).\u003c/p\u003e\n\u003cp\u003e26.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Archer, S. Tree-grass dynamics in a Prosopis-thornscrub savanna parkland: Reconstructing the past and predicting the future. \u003cem\u003e\u0026Eacute;coscience\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, 83\u0026ndash;99 (1995).\u003c/p\u003e\n\u003cp\u003e27.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Hobbs, R. J. \u0026amp; Mooney, H. A. Community changes following shrub invasion of grassland. \u003cem\u003eOecologia\u003c/em\u003e \u003cstrong\u003e70\u003c/strong\u003e, 508\u0026ndash;513 (1986).\u003c/p\u003e\n\u003cp\u003e28.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Adamoli, J., Sennhauser, E., Acero, J. M. \u0026amp; Rescia, A. Stress and Disturbance: Vegetation Dynamics in the Dry Chaco Region of Argentina. \u003cem\u003eJournal of Biogeography\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 491 (1990).\u003c/p\u003e\n\u003cp\u003e29.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Vegten, J. A. Thornbush invasion in a savanna ecosystem in eastern Botswana. \u003cem\u003eVegetatio\u003c/em\u003e \u003cstrong\u003e56\u003c/strong\u003e, 3\u0026ndash;7 (1984).\u003c/p\u003e\n\u003cp\u003e30.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Burrows, W. H., Carter, J. O., Scanlan, J. C. \u0026amp; Anderson, E. R. Management of Savannas for Livestock Production in North-East Australia: Contrasts Across the Tree-Grass Continuum. \u003cem\u003eJournal of Biogeography\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 503 (1990).\u003c/p\u003e\n\u003cp\u003e31.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Singh, J. S. \u0026amp; Joshi, M. C. Ecology of the Semi-Arid Regions of India with Emphasis on Land-Use. in \u003cem\u003eDevelopments in Agricultural and Managed Forest Ecology\u003c/em\u003e vol. 7 243\u0026ndash;275 (Elsevier, 1979).\u003c/p\u003e\n\u003cp\u003e32.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Barger, N. N. \u003cem\u003eet al.\u003c/em\u003e Woody plant proliferation in North American drylands: A synthesis of impacts on ecosystem carbon balance. \u003cem\u003eJ. Geophys. Res.\u003c/em\u003e \u003cstrong\u003e116\u003c/strong\u003e, G00K07 (2011).\u003c/p\u003e\n\u003cp\u003e33.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Eldridge, D. J. \u003cem\u003eet al.\u003c/em\u003e Impacts of shrub encroachment on ecosystem structure and functioning: towards a global synthesis: Synthesizing shrub encroachment effects. \u003cem\u003eEcology Letters\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 709\u0026ndash;722 (2011).\u003c/p\u003e\n\u003cp\u003e34.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Murthy, K. K., Sinha, S. K., Kaul, R. \u0026amp; Vaidyanathan, S. A fine-scale state-space model to understand drivers of forest fires in the Himalayan foothills. \u003cem\u003eForest Ecology and Management\u003c/em\u003e \u003cstrong\u003e432\u003c/strong\u003e, 902\u0026ndash;911 (2019).\u003c/p\u003e\n\u003cp\u003e35.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Banerjee, S., Das, D. \u0026amp; John, R. Grassland vegetation and roads have dominant influence on decadal-scale spatial-temporal patterns of fires in a species-rich protected Terai habitat in northeastern India. \u003cem\u003eAgricultural and Forest Meteorology\u003c/em\u003e \u003cstrong\u003e304\u0026ndash;305\u003c/strong\u003e, 108411 (2021).\u003c/p\u003e\n\u003cp\u003e36.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Archer, S., Schimel, D. S. \u0026amp; Holland, E. A. Mechanisms of shrubland expansion: land use, climate or CO2? \u003cem\u003eClimatic Change\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 91\u0026ndash;99 (1995).\u003c/p\u003e\n\u003cp\u003e37.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Van Auken, O. W. Causes and consequences of woody plant encroachment into western North American grasslands. \u003cem\u003eJournal of Environmental Management\u003c/em\u003e \u003cstrong\u003e90\u003c/strong\u003e, 2931\u0026ndash;2942 (2009).\u003c/p\u003e\n\u003cp\u003e38.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Venter, Z. S., Cramer, M. D. \u0026amp; Hawkins, H.-J. Drivers of woody plant encroachment over Africa. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 2272 (2018).\u003c/p\u003e\n\u003cp\u003e39.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Wigley, B. J., Bond, W. J. \u0026amp; Hoffman, M. T. Thicket expansion in a South African savanna under divergent land use: local vs. global drivers? \u003cem\u003eGlobal Change Biology\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 964\u0026ndash;976 (2010).\u003c/p\u003e\n\u003cp\u003e40.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Sala, O. E. \u0026amp; Maestre, F. T. Grass-woodland transitions: determinants and consequences for ecosystem functioning and provisioning of services. \u003cem\u003eJ Ecol\u003c/em\u003e \u003cstrong\u003e102\u003c/strong\u003e, 1357\u0026ndash;1362 (2014).\u003c/p\u003e\n\u003cp\u003e41.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Sala, O., Lauenroth, W. \u0026amp; Golluscio, R. Plant functional types in temperate semi-arid regions. \u003cem\u003ePlant functional types: their relevance to ecosystem properties and global change\u003c/em\u003e \u003cstrong\u003e1\u003c/strong\u003e, 217 (1997).\u003c/p\u003e\n\u003cp\u003e42.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Saxton, K. E. \u0026amp; Rawls, W. J. Soil Water Characteristic Estimates by Texture and Organic Matter for Hydrologic Solutions. \u003cem\u003eSoil Sci. Soc. Am. J.\u003c/em\u003e \u003cstrong\u003e70\u003c/strong\u003e, 1569\u0026ndash;1578 (2006).\u003c/p\u003e\n\u003cp\u003e43.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Goheen, J. R. \u003cem\u003eet al.\u003c/em\u003e Conservation lessons from large-mammal manipulations in East African savannas: the KLEE, UHURU, and GLADE experiments: Large mammal manipulations in Laikipia, Kenya. \u003cem\u003eAnn. N.Y. Acad. Sci.\u003c/em\u003e \u003cstrong\u003e1429\u003c/strong\u003e, 31\u0026ndash;49 (2018).\u003c/p\u003e\n\u003cp\u003e44.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Ward, D. Clipping frequency but not nutrients affect the architecture and non-structural carbohydrates of a browsing lawn. \u003cem\u003ePlant Ecol\u003c/em\u003e \u003cstrong\u003e217\u003c/strong\u003e, 21\u0026ndash;29 (2016).\u003c/p\u003e\n\u003cp\u003e45.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Collar, N., Garcia, E. \u0026amp; Sharpe, C. Bengal Florican (Houbaropsis bengalensis). in \u003cem\u003eBirds of the World\u003c/em\u003e (eds. Billerman, S. M., Keeney, B. K., Rodewald, P. G. \u0026amp; Schulenberg, T. S.) (Cornell Lab of Ornithology, 2020). doi:10.2173/bow.benflo2.01.\u003c/p\u003e\n\u003cp\u003e46.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Oliver, W. L. R. \u0026amp; Deb Roy, S. The pygmy hog (Sus salvanius). \u003cem\u003ePigs, Peccaries, and Hippos: Status Survey and Conservation Action Plan (Oliver, WLR ed.). IUCN, Gland, Switzerland\u003c/em\u003e (1993).\u003c/p\u003e\n\u003cp\u003e47.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Fick, S. E. \u0026amp; Hijmans, R. J. WorldClim 2: new 1‐km spatial resolution climate surfaces for global land areas. \u003cem\u003eInt. J. Climatol\u003c/em\u003e \u003cstrong\u003e37\u003c/strong\u003e, 4302\u0026ndash;4315 (2017).\u003c/p\u003e\n\u003cp\u003e48.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Johnsingh, A. T. Status and conservation of the tiger in Uttaranchal, Northern India. \u003cem\u003eAMBIO: A Journal of the Human Environment\u003c/em\u003e \u003cstrong\u003e35\u003c/strong\u003e, 135\u0026ndash;137 (2006).\u003c/p\u003e\n\u003cp\u003e49.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Wulder, M. A. \u003cem\u003eet al.\u003c/em\u003e Current status of Landsat program, science, and applications. \u003cem\u003eRemote Sensing of Environment\u003c/em\u003e \u003cstrong\u003e225\u003c/strong\u003e, 127\u0026ndash;147 (2019).\u003c/p\u003e\n\u003cp\u003e50.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Achard, F. \u003cem\u003eet al.\u003c/em\u003e Determination of tropical deforestation rates and related carbon losses from 1990 to 2010. \u003cem\u003eGlob Change Biol\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 2540\u0026ndash;2554 (2014).\u003c/p\u003e\n\u003cp\u003e51.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Hu, Y. \u0026amp; Nacun, B. An Analysis of Land-Use Change and Grassland Degradation from a Policy Perspective in Inner Mongolia, China, 1990\u0026ndash;2015. \u003cem\u003eSustainability\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 4048 (2018).\u003c/p\u003e\n\u003cp\u003e52.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Keenan, R. J. \u003cem\u003eet al.\u003c/em\u003e Dynamics of global forest area: Results from the FAO Global Forest Resources Assessment 2015. \u003cem\u003eForest Ecology and Management\u003c/em\u003e \u003cstrong\u003e352\u003c/strong\u003e, 9\u0026ndash;20 (2015).\u003c/p\u003e\n\u003cp\u003e53.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;de Sousa, C. \u003cem\u003eet al.\u003c/em\u003e Cloud-computing and machine learning in support of country-level land cover and ecosystem extent mapping in Liberia and Gabon. \u003cem\u003ePLoS ONE\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, e0227438 (2020).\u003c/p\u003e\n\u003cp\u003e54.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Midekisa, A. \u003cem\u003eet al.\u003c/em\u003e Mapping land cover change over continental Africa using Landsat and Google Earth Engine cloud computing. \u003cem\u003ePLoS ONE\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, e0184926 (2017).\u003c/p\u003e\n\u003cp\u003e55.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Phan, T. N., Kuch, V. \u0026amp; Lehnert, L. W. Land Cover Classification using Google Earth Engine and Random Forest Classifier\u0026mdash;The Role of Image Composition. \u003cem\u003eRemote Sensing\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 2411 (2020).\u003c/p\u003e\n\u003cp\u003e56.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Breiman, L. Random forests. \u003cem\u003eMachine learning\u003c/em\u003e \u003cstrong\u003e45\u003c/strong\u003e, 5\u0026ndash;32 (2001).\u003c/p\u003e\n\u003cp\u003e57.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Cutler, D. R. \u003cem\u003eet al.\u003c/em\u003e Random forests for classification in ecology. \u003cem\u003eEcology\u003c/em\u003e \u003cstrong\u003e88\u003c/strong\u003e, 2783\u0026ndash;2792 (2007).\u003c/p\u003e\n\u003cp\u003e58.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Maxwell, A. E. \u003cem\u003eet al.\u003c/em\u003e Large-Area, High Spatial Resolution Land Cover Mapping Using Random Forests, GEOBIA, and NAIP Orthophotography: Findings and Recommendations. \u003cem\u003eRemote Sensing\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 1409 (2019).\u003c/p\u003e\n\u003cp\u003e59.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Tamiminia, H. \u003cem\u003eet al.\u003c/em\u003e Google Earth Engine for geo-big data applications: A meta-analysis and systematic review. \u003cem\u003eISPRS Journal of Photogrammetry and Remote Sensing\u003c/em\u003e \u003cstrong\u003e164\u003c/strong\u003e, 152\u0026ndash;170 (2020).\u003c/p\u003e\n\u003cp\u003e60.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Davies, D. K., Ilavajhala, S., Min Minnie Wong \u0026amp; Justice, C. O. Fire Information for Resource Management System: Archiving and Distributing MODIS Active Fire Data. \u003cem\u003eIEEE Trans. Geosci. Remote Sensing\u003c/em\u003e \u003cstrong\u003e47\u003c/strong\u003e, 72\u0026ndash;79 (2009).\u003c/p\u003e\n\u003cp\u003e61.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Mattivi, P., Franci, F., Lambertini, A. \u0026amp; Bitelli, G. TWI computation: a comparison of different open source GISs. \u003cem\u003eOpen geospatial data, softw. stand.\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, 6 (2019).\u003c/p\u003e\n\u003cp\u003e62.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Hibbard, K. A., Schimel, D. S., Archer, S., Ojima, D. S. \u0026amp; Parton, W. Grassland to woodland transitions: integrating changes in landscape structure and biogeochemistry. \u003cem\u003eEcological Applications\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 911\u0026ndash;926 (2003).\u003c/p\u003e\n\u003cp\u003e63.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Duncan, R. S. \u0026amp; Duncan, V. E. Forest Succession and Distance from Forest Edge in an Afro-Tropical Grassland. \u003cem\u003eBiotropica\u003c/em\u003e \u003cstrong\u003e32\u003c/strong\u003e, 33\u0026ndash;41 (2000).\u003c/p\u003e\n\u003cp\u003e64.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Lee, D. Carbayes version 4.6: An r package for spatial areal unit modelling with conditional autoregressive priors. \u003cem\u003eUniversity of Glasgow, Glasgow\u003c/em\u003e (2017).\u003c/p\u003e\n\u003cp\u003e65.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;R Core Team. R: A language and environment for statistical computing. (2021).\u003c/p\u003e\n\u003cp\u003e66.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Venter, O. \u003cem\u003eet al.\u003c/em\u003e Sixteen years of change in the global terrestrial human footprint and implications for biodiversity conservation. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 12558 (2016).\u003c/p\u003e\n\u003cp\u003e67.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Harris, I., Osborn, T. J., Jones, P. \u0026amp; Lister, D. Version 4 of the CRU TS monthly high-resolution gridded multivariate climate dataset. \u003cem\u003eSci Data\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 109 (2020).\u003c/p\u003e\n\u003cp\u003e68.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;World Meteorological Organization. \u003cem\u003eStandardized precipitation index user guide.\u003c/em\u003e (2012).\u003c/p\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellpadding=\"0\" cellspacing=\"0\" width=\"1011\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003e\u003cstrong\u003eSl No\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003e\u003cstrong\u003eProtected Area\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e\u003cstrong\u003eArea (km\u003csup\u003e2\u003c/sup\u003e)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003e\u003cstrong\u003eCountry\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e\u003cstrong\u003eLatitudinal Range (\u0026deg;N)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e\u003cstrong\u003eLongitudinal Range (\u0026deg;E)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e\u003cstrong\u003eElevation Range (m)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e\u003cstrong\u003eTemperature Range (\u0026deg;C)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e\u003cstrong\u003eAnnual Precipitation (mm)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e\u003cstrong\u003eEffective Rainfall (mm)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eShukla Phanta National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e376\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eNepal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e28.76 \u0026shy; 29.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e80.10 - 80.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e143 - 1337\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e6.75 - 40.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e1258\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e-248.7\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eB\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eDudhwa National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e746\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eIndia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e28.32 - 28.70\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e80.46 - 80.93\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e115 - 233\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e6.50 - 39.65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e1266\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e-304.1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eBardiya National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e851\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eNepal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e28.29 - 28.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e81.20 - 81.71\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e119 - 1574\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e5.72 - 37.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e1310\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e-220.1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eValmiki Tiger Reserve\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e874\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eIndia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e27.16 - 27.52\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e83.83 - 84.65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e60 - 860\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e7.72 - 39.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e1693\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e165.5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eChitwan National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e1137\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eNepal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e27.34 - 27.69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e83.88 - 84.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e89 - 784\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e6.72 - 37.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e1949\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e484.5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eKoshi Tappu Wildlife Sanctuary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e143\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eNepal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e26.57 - 26.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e86.92 - 87.08\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e42 - 110\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e6.12 - 33.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e1536\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e-44.1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eJaldapara National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e221\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eIndia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e26.52 - 26.86\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e89.25 - 89.42\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e11 - 841\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e8.20 - 32.15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e2586\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e1124.8\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eManas National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e519\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eIndia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e26.60 -26.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e90.81 - 91.24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e6 - 305\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e7.22 - 31.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e2368\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e817.3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 1\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellpadding=\"0\" cellspacing=\"0\" width=\"1011\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003e\u003cstrong\u003eSl No\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003e\u003cstrong\u003eProtected Area\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e\u003cstrong\u003eArea (km\u003csup\u003e2\u003c/sup\u003e)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003e\u003cstrong\u003eCountry\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e\u003cstrong\u003eLatitudinal Range (\u0026deg;N)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e\u003cstrong\u003eLongitudinal Range (\u0026deg;E)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e\u003cstrong\u003eElevation Range (m)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e\u003cstrong\u003eTemperature Range (\u0026deg;C)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e\u003cstrong\u003eAnnual Precipitation (mm)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e\u003cstrong\u003eEffective Rainfall (mm)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eShukla Phanta National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e376\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eNepal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e28.76 \u0026shy; 29.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e80.10 - 80.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e143 - 1337\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e6.75 - 40.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e1258\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e-248.7\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eB\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eDudhwa National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e746\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eIndia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e28.32 - 28.70\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e80.46 - 80.93\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e115 - 233\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e6.50 - 39.65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e1266\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e-304.1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eBardiya National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e851\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eNepal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e28.29 - 28.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e81.20 - 81.71\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e119 - 1574\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e5.72 - 37.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e1310\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e-220.1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eValmiki Tiger Reserve\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e874\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eIndia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e27.16 - 27.52\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e83.83 - 84.65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e60 - 860\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e7.72 - 39.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e1693\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e165.5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eChitwan National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e1137\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eNepal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e27.34 - 27.69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e83.88 - 84.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e89 - 784\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e6.72 - 37.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e1949\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e484.5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eKoshi Tappu Wildlife Sanctuary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e143\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eNepal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e26.57 - 26.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e86.92 - 87.08\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e42 - 110\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e6.12 - 33.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e1536\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e-44.1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eG\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eJaldapara National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e221\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eIndia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e26.52 - 26.86\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e89.25 - 89.42\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e11 - 841\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e8.20 - 32.15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e2586\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e1124.8\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" width=\"4.055390702274975%\"\u003e\n \u003cp\u003eH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"22.05736894164194%\"\u003e\n \u003cp\u003eManas National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"6.824925816023739%\"\u003e\n \u003cp\u003e519\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.9129574678536105%\"\u003e\n \u003cp\u003eIndia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e26.60 -26.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"11.374876360039565%\"\u003e\n \u003cp\u003e90.81 - 91.24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"8.308605341246292%\"\u003e\n \u003cp\u003e6 - 305\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.385756676557863%\"\u003e\n \u003cp\u003e7.22 - 31.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"10.187932739861523%\"\u003e\n \u003cp\u003e2368\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" width=\"7.517309594460929%\"\u003e\n \u003cp\u003e817.3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellpadding=\"0\" cellspacing=\"0\" width=\"471\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.74468085106383%\"\u003e\n \u003cp\u003e\u003cstrong\u003eVariable\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.51063829787234%\"\u003e\n \u003cp\u003e\u003cstrong\u003eEstimate\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e\u003cstrong\u003eStd. Error\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.574468085106382%\"\u003e\n \u003cp\u003e\u003cstrong\u003ez value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e\u003cstrong\u003ePr(\u0026gt;|z|)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.74468085106383%\"\u003e\n \u003cp\u003e\u003cstrong\u003e(Intercept)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.51063829787234%\"\u003e\n \u003cp\u003e-0.443\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e0.023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.574468085106382%\"\u003e\n \u003cp\u003e-19.139\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.74468085106383%\"\u003e\n \u003cp\u003e\u003cstrong\u003eFIRE_OCCURR\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.51063829787234%\"\u003e\n \u003cp\u003e0.638\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e0.028\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.574468085106382%\"\u003e\n \u003cp\u003e22.741\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.74468085106383%\"\u003e\n \u003cp\u003e\u003cstrong\u003eTWI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.51063829787234%\"\u003e\n \u003cp\u003e-0.086\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e0.033\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.574468085106382%\"\u003e\n \u003cp\u003e-2.596\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0094\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.74468085106383%\"\u003e\n \u003cp\u003e\u003cstrong\u003eSLOPE\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.51063829787234%\"\u003e\n \u003cp\u003e0.097\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e0.032\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.574468085106382%\"\u003e\n \u003cp\u003e3.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0025\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.74468085106383%\"\u003e\n \u003cp\u003e\u003cstrong\u003eRUGG\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.51063829787234%\"\u003e\n \u003cp\u003e-0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e0.025\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.574468085106382%\"\u003e\n \u003cp\u003e-0.808\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e0.4193\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.74468085106383%\"\u003e\n \u003cp\u003e\u003cstrong\u003eDIST_WOOD\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.51063829787234%\"\u003e\n \u003cp\u003e0.082\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e0.024\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.574468085106382%\"\u003e\n \u003cp\u003e3.448\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.0006\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"25.74468085106383%\"\u003e\n \u003cp\u003e\u003cstrong\u003eDIST_ANTH\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.51063829787234%\"\u003e\n \u003cp\u003e-0.256\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e0.027\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"19.574468085106382%\"\u003e\n \u003cp\u003e-9.609\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.085106382978722%\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eTable 4\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellpadding=\"0\" cellspacing=\"0\" width=\"555\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"23.56115107913669%\"\u003e\n \u003cp\u003e\u003cstrong\u003eVariable\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e\u003cstrong\u003eMedian\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e\u003cstrong\u003e2.50%\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e\u003cstrong\u003e97.50%\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.007194244604317%\"\u003e\n \u003cp\u003e\u003cstrong\u003en.effective\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.805755395683452%\"\u003e\n \u003cp\u003e\u003cstrong\u003eGeweke.diag\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"23.56115107913669%\"\u003e\n \u003cp\u003e\u003cstrong\u003e(Intercept)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e-0.435\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e-0.480\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e-0.390\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.007194244604317%\"\u003e\n \u003cp\u003e8515.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.805755395683452%\"\u003e\n \u003cp\u003e0.3\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"23.56115107913669%\"\u003e\n \u003cp\u003e\u003cstrong\u003eFIRE_OCCURR\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e0.627\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e0.573\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e0.682\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.007194244604317%\"\u003e\n \u003cp\u003e7001.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.805755395683452%\"\u003e\n \u003cp\u003e-0.2\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"23.56115107913669%\"\u003e\n \u003cp\u003e\u003cstrong\u003eTWI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e-0.099\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e-0.165\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e-0.033\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.007194244604317%\"\u003e\n \u003cp\u003e4756.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.805755395683452%\"\u003e\n \u003cp\u003e-0.9\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"23.56115107913669%\"\u003e\n \u003cp\u003e\u003cstrong\u003eSLOPE\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e0.089\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e0.027\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e0.151\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.007194244604317%\"\u003e\n \u003cp\u003e5022.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.805755395683452%\"\u003e\n \u003cp\u003e0.1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"23.56115107913669%\"\u003e\n \u003cp\u003e\u003cstrong\u003eRUGG\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e-0.010\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e-0.060\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e0.038\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.007194244604317%\"\u003e\n \u003cp\u003e7634.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.805755395683452%\"\u003e\n \u003cp\u003e1.1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"23.56115107913669%\"\u003e\n \u003cp\u003e\u003cstrong\u003eDIST_WOOD\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e0.062\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e0.015\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e0.110\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.007194244604317%\"\u003e\n \u003cp\u003e8662.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.805755395683452%\"\u003e\n \u003cp\u003e-0.9\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"23.56115107913669%\"\u003e\n \u003cp\u003e\u003cstrong\u003eDIST_CROP\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e-0.256\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e-0.308\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"14.20863309352518%\"\u003e\n \u003cp\u003e-0.205\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"16.007194244604317%\"\u003e\n \u003cp\u003e7809.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.805755395683452%\"\u003e\n \u003cp\u003e-0.8\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eTable 5\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003ctable border=\"0\" cellpadding=\"0\" cellspacing=\"0\" width=\"624\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" width=\"31.41025641025641%\"\u003e\n \u003cp\u003e\u003cstrong\u003eProtected Area\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" width=\"17.147435897435898%\"\u003e\n \u003cp\u003e\u003cstrong\u003eMean Temperature change (\u0026deg;C)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" width=\"18.75%\"\u003e\n \u003cp\u003e\u003cstrong\u003eDry Monsoon Months \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; [SPI \u0026lt;-1.50]\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" width=\"15.064102564102564%\"\u003e\n \u003cp\u003e\u003cstrong\u003eHuman Footprint change (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" width=\"17.307692307692307%\"\u003e\n \u003cp\u003e\u003cstrong\u003eGrassland Cover change (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"0.32051282051282054%\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd height=\"23\" style=\"width: 8.739%;\" width=\"0%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" width=\"100%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd height=\"35\" style=\"width: 8.739%;\" width=\"0%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"31.41025641025641%\"\u003e\n \u003cp\u003eShukla-Phanta National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.147435897435898%\"\u003e\n \u003cp\u003e0.607\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.75%\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.064102564102564%\"\u003e\n \u003cp\u003e21.92\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.307692307692307%\"\u003e\n \u003cp\u003e-51.65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"0.32051282051282054%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd height=\"20\" style=\"width: 8.739%;\" width=\"0%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"31.41025641025641%\"\u003e\n \u003cp\u003eDudhwa National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.147435897435898%\"\u003e\n \u003cp\u003e0.652\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.75%\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.064102564102564%\"\u003e\n \u003cp\u003e54.33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.307692307692307%\"\u003e\n \u003cp\u003e-59.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"0.32051282051282054%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd height=\"20\" style=\"width: 8.739%;\" width=\"0%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"31.41025641025641%\"\u003e\n \u003cp\u003eBaradiya National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.147435897435898%\"\u003e\n \u003cp\u003e0.726\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.75%\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.064102564102564%\"\u003e\n \u003cp\u003e9.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.307692307692307%\"\u003e\n \u003cp\u003e34.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"0.32051282051282054%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd height=\"20\" style=\"width: 8.739%;\" width=\"0%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"31.41025641025641%\"\u003e\n \u003cp\u003eValmiki Tiger Reserve\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.147435897435898%\"\u003e\n \u003cp\u003e0.656\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.75%\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.064102564102564%\"\u003e\n \u003cp\u003e39.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.307692307692307%\"\u003e\n \u003cp\u003e18.12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"0.32051282051282054%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd height=\"20\" style=\"width: 8.739%;\" width=\"0%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"31.41025641025641%\"\u003e\n \u003cp\u003eChitwan National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.147435897435898%\"\u003e\n \u003cp\u003e0.674\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.75%\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.064102564102564%\"\u003e\n \u003cp\u003e27.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.307692307692307%\"\u003e\n \u003cp\u003e-65.59\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"0.32051282051282054%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd height=\"20\" style=\"width: 8.739%;\" width=\"0%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"31.41025641025641%\"\u003e\n \u003cp\u003eKoshi Tappu WLS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.147435897435898%\"\u003e\n \u003cp\u003e0.577\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.75%\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.064102564102564%\"\u003e\n \u003cp\u003e12.98\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.307692307692307%\"\u003e\n \u003cp\u003e28.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"0.32051282051282054%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd height=\"20\" style=\"width: 8.739%;\" width=\"0%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"31.41025641025641%\"\u003e\n \u003cp\u003eJaldapara National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.147435897435898%\"\u003e\n \u003cp\u003e0.684\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.75%\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.064102564102564%\"\u003e\n \u003cp\u003e18.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.307692307692307%\"\u003e\n \u003cp\u003e-26.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"0.32051282051282054%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd height=\"20\" style=\"width: 8.739%;\" width=\"0%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"31.41025641025641%\"\u003e\n \u003cp\u003eManas National Park\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.147435897435898%\"\u003e\n \u003cp\u003e0.535\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.75%\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"15.064102564102564%\"\u003e\n \u003cp\u003e51.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"17.307692307692307%\"\u003e\n \u003cp\u003e-40.25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"0.32051282051282054%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd height=\"20\" style=\"width: 8.739%;\" width=\"0%\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Grassland-Woodland Transitions, Landcover Classification, Habitat Loss, Anthropogenic Fire, Bayesian Model","lastPublishedDoi":"10.21203/rs.3.rs-1398899/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-1398899/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"The decline of grasslands and savanna due to woodland encroachment is a matter of global concern. In the Terai-Duar Savanna and Grassland ecoregion (hereafter terai), located at the base of the Himalayas in the Indian subcontinent, there are indications that grasslands and savanna are being lost, with unknown consequences for biodiversity and ecosystem function. We assessed large-scale vegetation changes in terai eight large protected terai habitats over three decades (1989-2019) and quantified the environmental and anthropogenic drivers of the observed changes using Bayesian Conditional Autoregressive spatial models. Notably, very little grasslands remain and the initial extent of 1417 km2 declined to 923 km2 (34.4%) in 30 years. Woodland area increased from 3235 km2 to 3516 km2 (8.7%). Dry season grass fire had the strongest influence on grassland persistence, followed by anthropogenic impacts. Terai ecosystems also experience significant threats from climatic changes and increasing human footprint, particularly in India.","manuscriptTitle":"Grassland-woodland transitions over decadal timescales in the Terai-Duar Savanna and Grasslands of the Indian subcontinent.","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2022-03-16 18:35:37","doi":"10.21203/rs.3.rs-1398899/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"3aee3668-d60b-4668-9fb0-bef160e6e818","owner":[],"postedDate":"March 16th, 2022","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2022-03-16T18:35:37+00:00","versionOfRecord":[],"versionCreatedAt":"2022-03-16 18:35:37","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-1398899","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-1398899","identity":"rs-1398899","version":["v1"]},"buildId":"-HB7Z8yhvgn0wM9Nzuekk","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}