International Research Journal of Plant Science

Assessing land use pattern and species diversity of a tropical dry deciduous forest ecosystem using geospatial techniques

Vandenhawii Chawlkha¹, Lalji Singh^1*, A.O. Varghese², Ravi Kant Chaturvedi³, Shalini Toppo¹ and Tarun Kumar Thakur^{4 1}Department of Forestry, Indira Gandhi Krishi Vishwavidyalaya, Raipur- 492012, India
²Regional Remote Sensing Centre-Central, NRSC, ISRO, Nagpur- 440 010, India
³Center for Integrative Conservation, Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, Menglun, Mengla, Yunnan 666303, P.R. China
⁴Department of Environmental Science, Indira Gandhi National Tribal University, Amarkantak, Amarkantak, MP, India-484887, India
^*Corresponding Author:
Lalji Singh, Department of Forestry, Indira Gandhi Krishi Vishwavidyalaya, India, Email: singh.js1@gmail.com

Received: 30-Nov-2023, Manuscript No. IRJPS-23-121714; Editor assigned: 02-Dec-2023 Reviewed: 15-Dec-2023 Revised: 20-Dec-2023 Published: 27-Dec-2023, DOI: http:/dx.doi.org/10.14303/irjps.2023.48

Introduction

In the last decade of the 20^th century, human activities such as deforestation, overharvesting, conversion of forest into non-forest, urbanization, fires, burning of fossil fuels and other land-use modifications had rapidly increased which caused the changes in atmosphere for instance increase a large amount of carbon dioxide (CO₂) and other greenhouse gases. The forest plays remarkable role in balancing global climate and sustaining global carbon cycle due to rapid increase in atmospheric CO₂, most scientists believe that it has implications on global warming.

The LULC pattern has vital role in global climatic condition. The increase in number of population and socio-economic requirements builds a force on transformation of land cover. That pressure creates unplanned and mismanagement changes on LULC. The LULC alteration leads to deforestation, global warming, biodiversity loss and extreme environmental crisis like landslides, drought, floods, etc. Hence, available data on LULC provides important contribution to execute the management of environment and planning for use in future. Moreover, geospatial technology has been used for monitoring and mapping of LULC in various tropical region of the world (Zhang et al. 2014; Wu et al. 2018 and Mishra et al. 2019). Vegetation analysis is an important tool to study species composition and phyto-sociological structure of the plant community. It helps to quantify various land, conservation management of endangered species, soil and water. Vegetation analysis plays a very important role in adaptation of plants to future climate change.

Geospatial technology provides reliable and unbiased information on LULC and species diversity analysis, their spatial distribution (Park and Lee 2016; Aslami and Ghorbani 2018). The estimation also helps to interpret the resources and modification in the ecosystem of forest, because forest can be influenced by several reasons like deforestation, pests & diseases, fires, grassland conversion which results in the changes of forest ecosystem. The present investigation helps in knowing the relationship between vegetation indices and structural attributes of forest and to estimate vegetation index that can correlate with the Shannon index. Vegetation indices (VIs) models are the most widely used models for estimation of the LULC in several studies (Geist and Lambin 2001; Foody et al. 2002; Yuan et al. 2005; Chetan et al. 2017). VIS are the arithmetical alteration of the innovative spectral reflectance which can be used to employ in knowing the vegetation types and cover (Rahman et al. 2003; He et al. 2006; Patel et al. 2007). The theory of this model helps in understanding the vegetation indices derived from vegetation, and vegetation has a high NIR wavelength due to the reason of spreading by leaf cells and a low red reflectance. The chlorophyll pigments absorption can be used in several satellite based vegetation index models for evaluating the vegetation parameters like biomass, leaf area and other activities (Baret & Guyot 1991; Verrelst et al. 2008). In order to understand the relationship between altitude, land cover and vegetation cover, remote sensing satellites and ground based techniques are used (Schlerf & Alzberger, 2005).

The conventional method derived from ground measurements is the most accurate but complicated to extend big areas and confirmed that it is very uneconomical as it requires lots of time and labour. In the recent years, RS and GIS techniques are most widely used for inventory, surveying, monitoring and mapping of the vegetation (Boyd et al. 1999; Lu et al. 2004; Ingram 2005; Maynard et al. 2007). RS and GIS are powerful tools to get precise information on the distribution of LULC transformation over large areas. A number of researchers employed the use of geospatial techniques for assessing LULC and vegetation in tropical country (Thakur et al. 2014; Soha & El-Raey 2019; Mishra et al. 2019).

Materials and Methods

Study area

The vegetation of study sites is tropical dry deciduous forest ecosystem in Pench National Park (as Tiger Reserve) and surrounding environment. The total study area was 398653.25 ha, of which forests occupied more than 60 % area. The study area, Pench landscape lies between 20º35’ to 21º44’ N lat and 78º15’ to 79º40’ E long and shared both the boundaries of Madhya Pradesh and Maharashtra. It is also declared as Pench Tiger Reserves. The location of study area and the details of sample points are illustrated in (Figure 1).

The average annual rainfall is approximately 1400 mm with the south-west monsoon accounting for most of the rainfall in the region. The mean rainfall for the dry period was 59.5 mm. The temperature varies from 25ºC to 47ºC in the months of April to June. The temperature varies from 6ºC to 31ºC in winter. The temperature varies from the minimum of 0ºC in winter and maximum 47ºC in summer. Relative humidity normally lies between 20 to more than 80 percent. Relative humidity is lowest in the month of May and drops below 25 percent. During monsoon, the relative humidity reaches more than 80 percent. Geological survey of India has done mapping of the geological information (Griffiths et al., 2010). Lithologically, the area is divided into nine distinct geological zones namely Gneisses, Basalt, Chorbaoli formation, Lameta formation, Pench river, Lohangi formation, Laterite, Manganese and Amphibolites.

Materials

LULC, structure and species diversity of a dry deciduous forest were analyzed using geospatial techniques along with ground measurements. The data is digitally analysed for determination of various LULC pattern and vegetation structure of different forests of Pench National Park to serve as primary stratification units (PSUs). The LULC and structural attributes investigation was first employed in these strata. Landsat 8 OLI of 144 path and 45 row, December 2017 and Sentinel 2A data of 2018 were used. Landsat 8 OLI Image (28 December 2017) was procured from open source earth explorer and Sentinel 2A data image of 16 October 2018 was procured using Python software. The data sets used in the study area (Pench National Park and surrounding) are given in (Table 1).

The digital analysis of data was performed on ERDAS Imagine (Version 9.1) and the subsidiary data together from SOI topomaps was determined in ARC-GIS (Version 10.3) in personnel computer. The geospatial analysis was carried out at Regional Remote Sensing Centre, Indian Space Research Organisation (ISRO), Nagpur, India.

Ancillary data

Location map of study area was prepared from survey of India toposheets 55/K 13,14,15, 55/N 11,12 and 55/O 1,2,3,5,6,7,10,11,14,15 (1:50,000) corresponding to the year 1967 and it would be used for rectification of satellite images. The location map was also used as reference map for precisely locating ground sample plots in the study area.

Stacking, Clipping and extraction of study area

Landsat 8 OLI of 144 path and 45 row, December 2017 and Sentinel 2A data of October 2018 were used. Landsat 8 OLI image (28 December 2017) was taken from open source earth explorer and Sentinel 2A data image of 16 October 2018 was procured using Python software. All the visible, near infrared and shortwave infrared bands was stacked in Landsat 8 OLI whereas all the visible and near infrared bands was stacked in Sentinel 2A. Later, the study area was exported in ARC-GIS after clipped and overlaid operations of the boundaries on two composition bands using ARC-GIS (Version 10.3).

Classification of satellite images and generation of vegetation indices

The supervised classification scheme was performed and the classification was made under ERDAS IMAGINE digital environment. Maximum Likelihood Algorithm Classifier was employed for classifying the LULC pattern and structural attributes of various dry deciduous forests. The generation of Vegetation indices such as EVI, SR, VDVI, NDVI and TNDVI map were performed using ERDAS Imagine (Version 9.3) software. Sentinel 2A data was used to generated Vegetation indices map. The following vegetation indices are derived from different spectral bands presented in (Table 2).

Analysis of vegetation structure and Plants diversity analysis

The detailed methodology adopted for the characterization of vegetation in the study area is depicted in (Figure 2). The coordinates of the sample points in the field were taken with the help of Handheld GPS (Global position System). 15 sample plots were selected for ground survey and 5 sample plots in each forest types. The sample points used in the study area are given in (Table 3).

The phyto-sociological analysis in each forest types has been carried by stratified random sampling by laying plots of 30 x 30 m, 2 x 2 m and 0.5 x 0.5 m for trees, shrubs and herbaceous vegetation, respectively. In each sample plot, vegetation was enumerated for their girth at breast height (GBH). The GBH of individual tree was measured at 1.37 m, while shrubs were measured at 15 cm above ground level by using measuring tape. The vegetation data in each forest/strata was analyzed for the determination of various structural attributes and diversity parameters for trees, shrubs and herbs components were determined following (Thakur, 2018).

Relationship between vegetation indices and Shannon index

The relationship between vegetation indices (viz. EVI, SR, VDVI, NDVI and TNDVI) and structural attributes (basal area, density and species diversity) were developed (Hurcom & Harrison,1998). The best fitted module was employed on the basis of regression coefficient (r²) and t- values. Best fitted module has been used for characterzing the structural attributes and Shannon index in a tropical dry deciduous forest of Pench National Park, India.

Results and Discussions

Spatial distribution pattern of land cover classes

The land cover classification of study area was performed using maximum likelihood algorithm. Before classification, all the visible and infrared bands of Landsat OLI were stacked and thereafter the polygon boundary of pench landscape was overlaid and the study area was extracted in ARC-GIS. Standard False Colour Composite (SFCC) was generated with the band combinations of visible and near infrared bands. An overview of Standard False Colour Composite (SFCC) of study area is shown in (Figure 3).

Supervised classification scheme using maximum likely hood algorithm wasemployed for classification of land cover types in ERDAS digital image analysis software. Six land cover types viz., water bodies, scrubland, open forest, dense forest, very dense forest and agriculture land were delineated. The results on spatial distribution pattern of LULC classes are presented in (Table 4) andalsodepictedthrough(Figure 4).

The Water bodies occupied an area of 7286.85 ha, scrubland 91314.3 ha, open forest 75704.8 ha, dense forest 123206 ha, very dense forest 41490.9 ha and agriculture land 59650.4 ha. The total study area comprises of 398653.25 ha. The forest occupies more than 60 percent of the total area and remaining 40 percent covered by water bodies, scrubland and agriculture. The classified image (Figure 5) shows the spatial distribution pattern of LULC class categories. Dense forest area occupied largest area, which accounted 30.9 per cent area of total geographical area, while water bodies occupied smallest area accounted only 1.83 percent of total area. The other classes i.e. scrubland, open forest, agriculture land and very dense forest, which covered 22.9%, 18.9%, 14.9% and 10.4% of total area, respectively. The overall classification accuracy varied from 91% to 93% for different land cover classes.

Remote sensing techniques were used by several workers for analyzing land use and vegetation in different surroundings through adopting various classification algorithms and was comparatively better classification technique appropriate for heterogeneous environments (Saxena et al. 1992; Sudhakar et al. 1994; Krishna et al., 2001). Results of present study are compare with the (Sugumaran et al.,1994), (Sehgal and Dubey, 1997) and (Mahajan et al.,2001), Similar observations have been also reported by (Manugul et al., 2017) for 96% overall accuracy. The current study also performed maximum accuracies for scrubland, dense forest and agriculture land. This might be due to distinctive spectral behavior of respective forests, which could be simply distinguishable in near infrared bands and also achieving better accuracies compared to other classes. The present findings (single date satellite data) could be one of the possible reasons for the achieving classification accuracy below 93%. Higher spatial resolution with multi temporal data help in improving classification accuracies, however the present accuracies for different land cover types observed to be over 90%, which is quite satisfactory. The use of better spatial resolution observed maximum accuracy (100%) compared to single date satellite data (Kachhawaha 1993; Ravan et al. 1996). This result is in agreement with the finding of (Zhou, 2013 for 85% classification accuracy.

Overall vegetation composition of tropical dry deciduous forest

The tropical dry deciduous forests of the area are rich in species composition and variety of flora was found in three distinct canopy layers. Forty-one species were found in tree layer, seven species in shrub layer and nine species in herbaceous layers. In tree layer, the prominent species were  Tectona grandis, Terminalia tomentosa, Largerstroemia parviflora, Anogeissus latifolia, Dalbergia paniculata, Chloroxylon swietenia and Cleistanthus collinus, whereas  Diospyros melanoxylon, Bauhinia racemosa, Acacia catechu and Mitragyna parvifolia were common co-dominant species, while Pterocarpus marsupium, Boswellia serrata, Adina cordifolia, Semecarpus anacardium and Cassine glauca as suppressed communities. Lantana camara, Woodfordia furticosa, Holorrhena antidysentrica, Grewia hirusta, Asparagus racemosus, Nycthanthes arbor-tristis and  Carrissa carandus  were found in shrub layer and Achyranthes aspera, Aristida funiculata, Tridax procumbens, Chrysopogon fulvus, Tephrosia hamiltonii, Mimosa pudica, Cyndon dactylon, Cyperus rotundus and Euphorbia hirta were recorded in herbaceous layer (Table 5-7).

Structure analysis of dry deciduous forest

Results on density (trees ha^-1) and basal area (cross sectional area of stems ha^-1) extent maximum forest areas are illustrated in (Figure 6 and Figure 7). The density ranged from 391 to 936 stems ha-1. Very dense forest reported thehighest density followed by dense forest and lowest in open forest. Basal area ranged from 36.69 to 64.32 m² ha^-1 in different forest types. Very dense forest observed maximum BA followed by dense forest. The minimum basalarea was recorded in open forest Figure 7.

Diversity analysis of dry deciduous forest

The different diversity indices were determined for different forest types to compare the difference in structural and diversity parameters among various forest types of Pench National Park. The results on species diversity of Pench National Park are given in (Table 8).

Shannon index among different forest types ranged from 2.53 to 2.97 in tree layer, 0.65 to 1.12 in shrub layer and 0.93 to 1.51 in herbaceous layer. Simpson index varied from 0.06 to 0.13 in tree layer, 0.37 to 0.54 in shrub layer and 0.26 to 0.51 in herbaceous layer. The Margalef’s richness varied from 4.25 to 4.96 in tree layer, 0.77 to 4.77 in shrub layer and 4.81 to 6.82 in the herbaceous layer. Beta diversity in tree layer varied from 1.98 to 2.80, 1.40 to 2.33 in shrub layer and 1.28 to 1.80 in herbaceous layer. Interestingly, the present results on species diversity parameters are similar with others workers in the tropical environment (Singh and Singh 1991; Gupta & Shukla., 1991, Prasad and Pandey 1992; Singh et al. 2005; Thakur 2019). Prasad and Pandey reported the ranged from 0.32 to 3.76 andconcentration of dominance from 0.07 to 0.63 (Prasad and Pandey, 1992). The species diversity values of dry deciduous forests in present study were moderately worse than data reported by (Singh et al.,1984) and (Swamy, 1998).

Correlation between structural, diversity parameters and vegetation indices

Correlation analysis was performed to study the relationship between structural attributes, diversity and diversity indices of tropical dry deciduous forests and finding are illustrated in Table 9. Result indicates that Shannon index was significantly correlated with NDVI, SR, VDVI and TNDVI, while EVI exhibited negative correlation. Among the different parameters studied, NDVI, SR and TNDVI, EVI were strongly correlated with structural and diversity parameters, whereas VDVI showed non-significant correlation (Table 9). The NDVI, EVI, SR, VDVI and EVI map were depicted in (Figure 8). The best and maximum correlation coefficient (R²) between species diversity and vegetation indices was found in NDVI with R2 value with 0.869 followed by SR, TNDVI, EVI, VDVI and with R² value were 0.789, 0.698, -0.737 and 0.151, respectively. The present study indicated a positive and significant correlation between NDVI and Shannon index for different vegetation types of study area which matches with reports of earlier workers who found NDVI is key variable strongly correlated to vegetation analysis of (Thakur et al., 2014). Presents finding confirms the result of previous worker (Zheng et al., 2004).

Conclusions

The study indicated that RS and GIS techniques were found quite useful techniques for the characterization of LULC pattern, species diversity analysis of a tropical dry deciduous forest of Pench National Park, located in the part between Madhya Pradesh and Maharashtra, India. The study also demonstrated that it is feasible to accomplish desirable classification accuracies (91- 93%) for various LULC classes in tropical dry deciduous forest using supervised classification through MLA. However, data should be free from clouds and preferably taken during October-December. Further, it is suggested to use multi date satellite data for enhancing the accuracy interval of classification and even more precisely delineating Very dense tropical dry deciduous forest. The utilization of multi temporal and high resolution satellite data will give an opportunity to understand the community level organization in different forests.

The study reveals that tropical dry deciduous forests of Pench National Park are biologically as rich in terms of structure, diversity and biomass. But due to increased anthropogenic interferences these forests are degrading and affecting the density and also decreasing the number of trees in bigger diameter classes resulting poor above ground biomass in the forests. The study recommends adopting agri-silviculture practices and intensive conservative measures, silvipastural systems should be developed in natural open grasslands by planting MPTs and other palatable grasses to protect the forests from overgrazing and browsing. The study revealed that tropical dry deciduous forests of Pench National Park have highly possible intended for carbon sequestration and are moderately young and immature state, since most of the standing trees are in lower diameter class. The conclusion and suggestion emerged from the study are essential for the sustainable management of tropical dry deciduous forests of Pench National Park.

Acknowledgement

Thanks are due to Indira Gandhi Krishi Vishwavidyalaya, Raipur for the financial support to Vandenhawii Chawlkha in the form of scholarship and to Regional Remote Sensing Centre – Central, NRSC, ISRO, Nagpur for the providing the training to Vandenhawii Chawlkha during her attachment under MoU.

References

Aslami F, Ghorbani A (2018). Object-based land-use/land-cover change detection using Landsat imagery: a case study of Ardabil, Namin, and Nir counties in northwest Iran. Environ Monit Assess.190:376.

Indexed at, Google Scholar, Cross Ref

Baret F, Guyot G (1991). Potentials and limits of vegetation indices for LAI and APAR assessment. Remote Sens Environ. 35, 161–173.

Indexed at, Google Scholar, Cross Ref

Boyd DS, Foody GM, Curran PJ (1999). The relationship between the biomass of Cameroonian tropical forests and radiation reflected in middle infrared wavelengths. Int J Remote Sens. 20(5), 1017-1023.

Indexed at, Google Scholar, Cross Ref

Chetan MA, Dornik A, Urdea P (2017). Comparison of object and pixel-based land cover classification through three supervised methods. J Geodesy Geoinformation sci.

Indexed at, Google Scholar, Cross Ref

Foody GM (2002). Status of land covers classification accuracy assessment. Remote Sens Environ. 80:185–201.

Indexed at, Google Scholar, Cross Ref

Geist HJ, Lambin EF (2001). What Drives Tropical Deforestation? A Meta Analysis of Proximate and Underlying Causes of Deforestation Based on Sub-national Case Study Evidence. Louvain-la-Neuve (Belgium): LUCC International Project Office, LUCC Report Series no. 4.

Indexed at, Google Scholar

Griffiths P, Hoster P, Grubner O, van der, Linden S (2010). Mapping megacity growth with multi sensor data. Remote Sens Environ.114, 426–439.

Indexed at, Google Scholar, Cross Ref

Gupta OP, Shukla RP (1991) The composition and dynamics of associated plant communities of sal plantations. Trop Ecol. 32: 296–309.

Indexed at, Google Scholar

He YH, Guo XL, Wilmshurst J (2006) Studying mixed grassland ecosystems: suitable hyperspectral vegetation indices. Can J Remote Sens. 32, 98-107.

Indexed at, Google Scholar, Cross Ref

Hurcom SJ, Harrison AR (1998). The NDVI and spectral decomposition for semi-arid vegetation abundance estimation. Int J Remote Sens, 19 (16), 3109- 3125.

Indexed at, Google Scholar, Cross Ref

Ingram JC, Dawson TP, Whittaker RJ (2005). Mapping tropical forest structure in southern Madagascar using remote sensing and artificial neural networks. Remote Sens Environ. 94, 491-507.

Indexed at, Google Scholar, Cross Ref

Kachhawaha TS (1993). Temporal and multisensar approach in forest/ vegetation mapping and their identification for effective management of Rajaji National park Uttar Pradesh, India. Int J Remote Sens.  14(17): 3105-3114.

Indexed at, Google Scholar, Cross Ref

Krishna NDR, Kmaji A, Murthy KYVN , Rao BSP (2001). Remote Sensing and Geographical Information System for Canopy Cover Mapping, J Indian Soc Remote Sens. 29(3): 108-113

Indexed at, Google Scholar, Cross Ref

Lu D, Mausel P, Brondizio E, Moran E (2004). Relationships between forest stand parameters and Landsat TM spectral responses in the Brazilian Amazon Basin. For Ecol Manag. 198, 149-167.

Indexed at, Google Scholar, Cross Ref

Mahajan S, Panwar P, Kaundal D (2001). GIS application to determine the effect of topography on land use in Ashwani khad watershed. J Remote Sens. 29 (4): 243-248.

Indexed at, Google Scholar, Cross Ref

Manugul SS, Sagar M, Nihar KS, Reddy KA (2017). Digital classification of land use/land cover by using remote sensing techniques, Int J Innov Eng. 8(2) 149-156.

Indexed at, Google Scholar, Cross Ref

Maynard CL, Lawrence RL, Nielsen GA, Decker G (2007). Modeling vegetation amount using bandwise regression and ecological site descriptions as an alternative to vegetation indices. Geosci Remote Sens. 44(1), 68-81.

Indexed at, Google Scholar, Cross Ref

Mishra PK, Rai A , Rai S C (2019). Land use and land cover change detection using geospatial techniques in the Sikkim Himalaya, India, Egypt J Remote Sens Space Sci.

Indexed at, Google Scholar, Cross Ref

Park J, Lee J. (2016). Detection of landuse/landcover changes using remotely-sensed data. J For Res.27:1343-50.

Indexed at, Google Scholar, Cross Ref

Patel NK, Saxena RK and Shiwalkar A (2007). Study of fractional vegetation cover using high spectral resolution data. J Indian Soc Remote Sens. 35: 73-79.

Indexed at, Google Scholar, Cross Ref

Rahman AF, Gamon JA, Sims DA, Schmidts M (2003). Optimum pixel size for hyperspectral studies of ecosystem function in southern California chaparral and grassland. Remote Sens Environ. 84: 192-207.

Indexed at, Google Scholar, Cross Ref

Prasad R, Pandey RK (1992). An observation on plant diversity of Sal and Teak forest in relation to intensity of biotic impact of various distances from habitation in Madhya Pradesh. A case study. J Trop For. 8(1): 62-83.

Indexed at, Google Scholar

Ravan SA (1994). Ecological analysis of vegetation from satellite remote sensing at Madhav National Park Sivapuri (M.P.). Ph.D. Thesis, HNB Garhwal University, Srinagar, India.

Google Scholar

Ravan SA, Roy PS, Sharma CM (1996). Accuracy evaluation of digital classification of Landsat TM data – An approach to include phenological stages of tropical dry deciduous forest. Int J Ecol Environ Sci. 22: 33-43.

Google Scholar

Saxena KG, Tiwari AK, Porwal MC , Menon ARR (1992). Vegetation maps, mapping needs and scope of digital processing of Landsat Thematic Mapper data in tropical region of South-West India. Int J Remote Sens. 13(11):2017-2037.

Indexed at, Google Scholar, Cross Ref

Schlerf M, Alzberger C, Hill J (2005). Remote sensing of forest biophysical variables using HyMap imaging spectrometer data. Remote Sens Environ. 95, 177-194.

Indexed at, Google Scholar, Cross Ref

Sehgal VK , Dubey RP (1997). Evaluation of potential usefulness of IRS-1C WIFS simulated data for dryland rabi sorghum crop discrimination. J Indian Soc Remote Sens. 25(3): 137-143.

Indexed at, Google Scholar, Cross Ref

Singh JS, Singh SP, Saxena AK , Rawat YS (1984). India"s silent valley and its threatened rain forest ecosystem. Environ Conserv. 11: 223-233.

Indexed at, Google Scholar, Cross Ref

Singh L Singh JS (1991). Species structure, dry matter dynamics and carbon flux of a dry tropical forest in India. Ann Bot, 68: 263-273.

Indexed at, Google Scholar, Cross Ref

Soha A M, El-Raey E M (2019). Land cover classification and change detection analysis of Qaroun and Wadi El-Rayyan lakes using multi-temporal remotely sensed imagery. Environ Monit Assess.191:229.

Google Scholar, Cross Ref

Sudhakar S, Kumar A, Arrawatia ML, Sengupta S (1994). Forest Cover Mapping of East District, Sikkim using IRS-1A LISS II Satellite Data. J Indian Soc Remote Sens.Vol. 22, No. 3.

Indexed at, Google Scholar, Cross Ref

Sugumaran R, Sandhya G, Rao KS, Javed RN, Kimothi MM (1994). Delineation of social forestry plantation under various afforestation programme using satellite digital data. J  Indian Soc Remote Sens. 22(4): 245-249.

Google Scholar, Cross Ref

Swamy SL (1998). Estimation of net primary productivity (NPP) in an Indian tropical evergreen forest using remote sensing data. Ph.D. Thesis, Jawaharlal Nehru Technology University, Hyderabad.

Google Scholar

Thakur TK (2018). Diversity, composition and structure of understorey vegetation in the tropical forest of Achanakmaar Biosphere Reserve, India. Environ Sustain. 1 (2): 279-293.

Google Scholar, Cross Ref

Thakur T, Swamy SL, Nain AS (2014). Composition, structure & diversity analysis of dry tropical forest of Chhattisgarh using Satellite data. J For Res, Vol. 25 (4) Pp 819-825.

Google Scholar, Cross Ref

Thakur TK, Swamy SL, Bijalwan A (2019). Assessment of biomass and net primary productivity of a dry tropical forest using geospatial technology. J For Res. 30 (1): 157-170.

Indexed at, Google Scholar, Cross Ref

Verrelst J, Schaepman ME, Koetz B, Kneubu¨ hler M (2008). Angular sensitivity analysis of vegetation indices derived from CHRIS/PROBA data. Remote Sens Environ. 112, 2341-2353.

Indexed at, Google Scholar, Cross Ref

Wu T, Luo J, Fang J, Ma J, Song X (2018). Unsupervised object-based change detection via aWeibull mixturemodel-based binarization for high-resolution remote sensing images. IEEE Geosci Remote Sens Lett. 15, 63–67.

Google Scholar, Cross Ref

Yuan F, Sawaya KE, Loeffelholz B, Bauer ME, (2005). Land cover classification and change analysis of the Twin Cities (Minnesota) Metropolitan Area by multitemporal Landsat remote sensing. Remote Sens Environ. 98, 317–328.

Indexed at, Google Scholar, Cross Ref

Zhang P, Lv Z, Shi W (2014). Local spectrum-trend similarity approach for detecting land-cover change by usingspot-5 satellite images. IEEE Geosci Remote Sens Lett. 11, 738–742.

Google Scholar, Cross Ref

Zheng D, Rademacher J, Chen J, Crow T, Bresee M, et al., (2004). Estimating aboveground biomass using landsat 7 ETM + data across a managed landscape in northern Wisconsin, USA. Remote Sens Environ. 93 (3): 402-411.

Indexed at, Google Scholar, Cross Ref

Zhou W (2013). An object-based approach for urban land cover classification: integrating LiDAR height and intensity data. IEEE Geosci Remote Sens. 29 (11), 3119-3135.

Indexed at, Google Scholar, Cross Ref