ISSN : 1016-0582

BANKO JANAKARI jgsf] hfgsf/L A JOURNAL OF FORESTRY INFORMATION FOR

Vol. 26, No. 1 2016 Banko Janakari is published by the Department of Forest Research and Survey (DFRS), Ministry of Forests and Soil Conservation, Government of Nepal. Any opinions expressed in this publication are those of the authors and should not be taken to represent official policy.

Editorial policy is made by the Advisory Committee of the Department of Forest Research and Survey. The members are :

P. Mathema, Director General R. Shakya, Deputy Director General D. K. Kharal, Deputy Director General Y. P. Pokharel, Deputy Director General

Editorial Board Members of Banko Janakari are:

B. S. Poudel, Co-ordinator R. Malla, Member S. Khanal, Member K. K. Pokharel, Member Secretary

Contributions and correspondence should be addressed to:

Member Secretary Department of Forest Research and Survey PO Box 3339 Babar Mahal Kathmandu, Nepal E-mail: [email protected]

© Department of Forest Research and Survey Produced by the Information Section of DFRS

Previous issues can be retrieved from http://www.dfrs.gov.np Banko Janakari A Journal of Forestry Information for Nepal Role of trees and forests in disaster risk reduction and mitigation Disaster, as defined by the United Nations Office for Disaster Risk Reduction, is a serious disruption of the functioning of a community or a society involving widespread human, material, economic or environmental losses and impacts, which exceeds the ability of the affected community or society to cope using its own resources. In recent years, the frequency and severity of both natural and man-made disasters have increased resulting in significant human casualties and loss of property. According to the International Emergency Disasters Database (EMDAT), 22,773 people were killed and 98.6 million people affected in 346 disasters reported worldwide in 2015. The economic damage totaled USD 66.5 billion in the year 2015. Floods, storms and droughts were the three most frequent natural disasters in 2015. Other common disasters were: landslides, earthquake and tsunami, wildfires and extreme temperatures. Nepal is susceptible to various kinds of disasters like earthquake, flood, landslide, fire (including forest fire), avalanche, epidemic, extreme rainfall, wind-storm and Glacial Lake Outburst Flood (GLOF). Every normal year, about 500 people are killed and many people get injured due to such disasters. According to the data of the Ministry of Home Affairs, the total property loss was USD 12 million in 2012. As per the Post Disaster Needs Assessment (PDNA) Report prepared under the leadership of National Planning Commission, the earthquake of April 2015 and its aftershocks caused a total of 8,790 human casualties, over 22,300 injuries, and property damages and loss worth USD 7 billion. The EMDAT reported this as the biggest natural disaster of the world in 2015. There are various preventive and curative mechanisms against natural and man- made disasters. Most of them are highly technical, externally designed, needing huge investments and difficult to implement. In this context, the conservation and management of forests can provide locally available, relatively less expensive but significant contribution in protecting against disasters and in reducing the negative impacts of disasters. Trees and forests can contribute in disaster risk reduction and mitigation mainly in four ways. First, forests and trees can offer physical barrier against natural disasters such as floods, landslides and tsunamis, thereby preventing loss of lives, property and livelihoods. Trees provide protection against floods, landslides and windstorms. Roots of the trees reduce soil erosion and land degradation by binding soils and Banko Janakari, Vol. 26, No. 1 soil nutrients. By reducing the speeds of the wind and water, forests and trees can diminish the magnitudes of disasters. In this context, bamboos, due to their fast growth and extensive root system, have a powerful role in controlling soil erosion and protecting landscapes. The physical barriers created by mangroves against cyclone and tsunami are extensively reported in literature. Likewise, tree plantations and conservation along the rivers have been considered as a measure to prevent river scouring and protection against floods. The role of trees and forests in water storage and recharge is widely regarded as the good strategy against drought. Trees on our hills have also been found to prevent human casualties and injuries by stopping stone falls. Second, forests provide building material for shelter and infrastructures in the recovery, rehabilitation and reconstruction phase to build back better. Forests provide the most convenient and locally available building materials for quick construction of shelters to the affected households and communities. The uprooted or damaged trees can be harvested to build temporary and permanent houses. Forests, therefore, have significant roles in the rehabilitation of communities affected by natural disasters. Third, forests support people by providing various kinds of foods before, during and after disasters. Several forestry products can be used as safety nets in the case of emergency and food shortages. Fourth, in the context of climate change contributing to increased frequency and severity of disasters, conservation of forests, reforestation/afforestation programmes, and sustainable forest management can minimize disaster risk by mitigating climate change through the reduction of greenhouse gases emissions and through enhanced carbon sequestration. In Nepal, an estimate of Department of Forests showed that 52 million cubic feet of logs are required for the reconstruction of infrastructures destroyed by the 2015 earthquake. The Government of Nepal has also eased the process of transportation of timber salvaged from old and damaged houses within the 31 earthquake affected districts of Nepal. Moreover, the Government has been planning to increase the production and manage the supply of timber for post- disaster reconstruction in Nepal. In this direction, prevalent community-based forest management regimes (community forestry, leasehold forestry, and collaborative forest management) should also be managed with ample emergency provisions for providing wood in any likely event of disaster. In addition, the forestry sector should have an emergency plan for responding promptly to any future disasters. Furthermore, there is a need for research on the role and effectiveness of trees and forests in preventing and mitigating various kinds of disasters in Nepal.

2 Effects of the environment on species richness and composition of vascular plants in Manaslu Conservation Area and Sagarmatha region of Nepalese Himalaya

S. K. Rai1,2,3, S. Sharma2, K. K. Shrestha2, J. P. Gajurel2,3, S. Devkota3, M. P. Nobis3 and C. Scheidegger3

This study analyzed how the environmental conditions constrained the species richness and composition in the four river valleys of Central Nepal i.e. two from Manaslu Conservation Area (MCA) and two from Sagarmatha region. Topographical, bioclimatic and measured variables were used to analyze their effects on the vascular plant diversity along elevation and land use gradients. Altogether, 148 plots were established at five elevation levels between 2,200 m and 3,800 m above the mean sea level. Four land use types namely crop field, meadow, exploited forest and natural forest were sampled at each elevation level. Altogether, 790 species of vascular plants belonging to 114 families were recorded; Asteraceae had the highest number of species (84) followed by Rosaceae (52) and Poaceae (50). Explorative data analysis of species composition by canonical correspondence analysis (CCA) showed that the topographical variables explained the composition better than both the bioclimatic set of variables and the logger data. However, all groups of variables revealed significant effects on species composition. Generalized Linear Model (GLM) also revealed significant effects of elevation, land-use types, slope angle, aspect, temperature and precipitation on species richness.

o Canonical correspondence analysis, elevation, generalized linear model, land use types, multivariate analysis, species richness

pecies diversity patterns are governed by 1992; McCain and Grytnes, 2010), for example, Sa varied set of biotic and abiotic factors. in mammals (McCain, 2007), birds (Island, 2012) Keeping biotic interactions at one end, the abiotic and vascular plants (Trigas et al., 2013). However, environmental drivers of species distribution both latitude and elevation alone cannot elucidate has gained much attention in recent studies all the causal biological factors, instead they are (Guisan and Zimmermann, 2000). There are proxy for numerous variables such as temperature, several environmental relationships that can be moisture energy and so on that change along the used to describe patterns of species distributions elevation (Körner, 2007), topography (Hofer et as well as species richness. Changes of species al., 2008) and latitude (Carpenter, 2005). Land distributions along the latitudinal and elevation use and geographic factors such as aspect and gradients are well known since the advent slope also play important roles in distribution of of modern biogeography (Lomolino, 2001; species in any area (Sanders and Rahbek, 2012). Colwell et al., 2004). The effect of latitude on species richness has been known for a long time In the Himalaya of Nepal and adjoining countries, (Pianka, 1966; Stevens, 1989). Stevens (1989) the species richness along the elevation gradients has compiled the published literatures showing have shown the mid-elevation peaks for vascular the effect of latitudinal gradients in the species plant species (Vetaas and Grytnes, 2002; Bhattarai richness at regional as well as local scales. and Vetaas, 2003), ferns (Bhattarai et al., 2004), Species richness and their distribution are also bryophytes (Grau et al., 2007), lichens (Baniya affected by the elevation gradients (Stevens, et al., 2010) and reptiles (Chettri et al., 2010). Those studies have often focused on elevation

1 Department of Plant Resources, Kathmandu, Nepal. E-mail: [email protected] 2 Central Department of Botany, Tribhuvan University, Kathmandu, Nepal 3 Swiss Federal Institute for Forest, Snow and Landscape Research, Switzerland

3 Banko Janakari, Vol. 26, No. 1 Rai et al. pattern in the species richness taken as proxies in (Fig. 1). The study was of changes in temperature, energy and water conducted during 2011 to 2013. availability (Bhattarai et al., 2004). In contrast, the topographical variables such as slope angle, aspect or regional differences were rarely analyzed in the Himalayan region (Paudel and Vetaas, 2014).The same hold for microclimates such as point temperature and water availability which might affect upon the species distribution (Geiger et al.,1995). In addition, different land use types also indicate different species communities with varying species richness and pattern. The settlements in the mountains of the Fig. 1: Map of Nepal showing the study districts Himalaya chiefly rely on agro-pastoral system. Gorkha at the center and Solukhumbu in the The shifting and open grazing system is practiced east in the mountain areas. Besides crop farming, the mountain people keep herds of cattle for the In Manaslu Conservation Area (MCA), two supply of food and economic needs. Their energy river valleys viz. the Nubri and the Tsum (Fig. source is mainly the firewood collected from the 2a) were studied. The Nubri valley starts from nearby forests. All the above activities can lead the confluence of Budhi Gandaki river and Siyar to the degradation of the natural habitats which khola (river) near Lokpa. This valley runs along affect upon the species diversity in different the Budhi Gandaki river upwards in north-west ways (Cousins, 2009; Honnay et al., 2005). In direction. Our study area started from Gap (2,200 most of the cases, the species diversity declines m) to Samagaun (3,700 m) located between in the degraded area due to the fragmentation of 28˚31’48.9” N and 28˚35’22.5” N latitude and the natural forests (Tilman et al., 1997; Maitima between 84˚38’29.6” E and 84˚49’51.9” E et al., 2009). These losses are linked with the longitude. The vegetation on the bank of the river disturbances and changes in the nutrient cycling near 2,200 m is broad leaved consisting of species processes such as organic carbon in the soil such as Benthamidia capitata, Michelia kisopa, (Maitima et al., 2009), and available nitrogen (Li Pinus wallichiana and Quercus semicarpifolia. et al., 2006). Above 2,500 m altitude, there is a dense forest of Tsuga dumosa, and above 3,000 m altitude, This study aims to find out the effects of most the forest is changed into larch forest (Larix widely used environmental variables such as himalaica). At 3,400 m altitude near Shyala, exists temperature, precipitation and topography at local a dense forest of Abies spectabilis associated as well as regional scales. We have also selected with Hippophae salicifolia and Cotoneaster spp. four land use types with an aim to show that Similarly, the Tsum valley is oriented towards species distribution pattern are also the function the north-east along the Siyar khola after the of land use types. The principle research questions confluence with the Budhi Gandaki river. Our are: (i) How the species richness and composition study area is located between 28˚26’19.3” N and vary along the altitudinal, precipitational and 28˚36’56.2” N latitude and between 84˚54’44.3” E other topographical indicators?, (ii) How the and 85˚06’40.4” E longitude. The lower elevation species are distributed in the different land use consists of alder (Alnus nepalensis) and pine types? and (iii) Which types of environmental Pinus wallichiana) forests. They are replaced variables are most suitable to explain the species by hemlock (Tsuga dumosa) and Himalayan fir richness and composition in the Himalaya? Abies spectabilis) at around 3,000 m altitude. The north facing slope of the valley harbors dense Materials and methods vegetation. Larix himalaica forest is dominant at Study area around 3,400 m altitude near Rachen Gumba. The north facing slopes possess more vegetation The study was conducted in the four river valleys cover than the south facing slopes. Betula utilis i of the two regions of Nepal: Manaslu Conservation found upto 3,800 m altitude near Kalung. Most of Area in and Sagarmatha region the south-facing slopes consist of open meadows

4 Rai et al. Banko Janakari, Vol. 26, No. 1 intersected by small human settlements such as located between 27˚40’18.1” N and 27˚49’48.3” Chumling, Gho, Chhekampar and Nile. N latitude and between 86˚42’3.2” E and 86˚44’25.2” E longitude. The plots located at 2,200 m at Surke and Nakchung and those at Muse and Sengma at 2,600 m elevation are outside the Sagarmatha National Park whereas the rest of the plots are within the boundaries of the Park. The vegetation of the site starts from Schima- Castanopsis and alder (Alnus nepalensis) at 2,200 m and is replaced by Pinus-Rhododendron at mid elevation (3,000 m) and is further replaced by Silver fir-birch-rhododendron at Khumjung (3,800 m). The study area at Dudhkund valley is located between 27˚30’39.9” N and 27˚39’49.1” N latitude and between 86˚34’34.5” and 86˚37’01.6” E longitude, and lies towards the west of Dudhkoshi river valley; the two valleys are separated by a chain of mountains. The Dudhkund valley does not fall inside the Sagarmatha National Park area. The plots, laid at 2,200 m and 2,600 m elevation, are near the Fig. 2a: Map of Gorkha district with plots settlements and the forests are managed by the overlaid on Nubri river valley on the left local Community Forest User Groups (CFUGs). and Tsum river valley on the right The forests above 3,000 m elevation are managed by the Government as national forest. The crop fields are not found at and above 3,000 m altitude except one at Taksindu. The study started at Boldok-Kholaghari (2,200 m). Going upwards from Phera (2,600 m), Taksindu (3,000 m) and Sarkaripati (3,400 m), our highest plot was located near Sasarbeni (3,800 m). The vegetation at 2,200 m is Schima-Castanopsis-Alnus, Pinus and then followed by Pinus-Quercus-Rhododendron a mid-elevation. Abies spectabilis forest can be noticed at Sasarbeni (3,800 m).

Study design

Five elevation levels were investigated with a regular elevation interval of 400 m starting from 2,200 m to 3,800 m. At each elevation level, four land use types were considered viz. (i) natural forest, (ii) exploited forest, (iii) meadow and (iv) Fig. 2b: Map of Solukhumbu district with crop field (Scheideggeret al., 2010). The category plots overlaid on Dudhkunda river valley on of the land use types were based on the visual the left and Dudhkoshi river valley on the right observation in accordance with the methods of FAO (Gregorio and Jansen, 2000). The crop In Sagarmatha region, we studied the Dudhkoshi fields are cultivated areas where the vegetative and the Dudhkunda (Fig. 2b) river valleys. cover is created by anthropogenic activities, The region is famous for the world’s highest and so become bare during off-crop season. mountain, Sagarmatha (Mt. Everest, 8,848 The meadows are isolated patch or wide area of m) and the Sagarmatha National Park. The grazing land where the tree species are less than Dudhkoshi river valley runs northwards along the 20%, and they are also affected by anthropogenic bank of Dudhkoshi river. The studied plots are activities such as livestock grazing and grass

5 Banko Janakari, Vol. 26, No. 1 Rai et al. collection. The natural forests are far from the Data source human settlements which are rarely intervened Plant species records as response variable by anthropogenic activities. The exploited forests comprise the vegetation not planted by humans Most of the flowering plant species were but influenced by their actions. This does not identified in the field by using the books written require human activities to be maintained in the by Polunin and Stainton (1984) and Stainton long-term as compared to the crop fields. (1988). The specimens unidentified in the field were identified at the National Herbarium and All the four land use types were assessed for Plant Laboratories (KATH), Godawari, Lalitpur. species records on both sides of the river. Two The voucher specimens were submitted to the sample plots (25 m x 2.5 m) were selected KATH Herbarium. randomly per land use type at each elevation level (e.g. 2,200±50 m) on the one side of the river, and For nomenclature of the species, we followed the same number were replicated on the another the Angiospermic Phylogenetic Group (APG III) side of the river (Scheidegger et al., 2010). Each system (Chase et al., 2009). In the case of the plot was divided into 5 m x 2.5 m sub-plots unresolved names (according to APG III), the for species record. Thus, each elevation level nomenclature of Press et al. (2000) was adopted. consisted of eight sample plots (Fig. 3). Crop On the other hand, the nomenclature of Iwatsuki fields were not found at the elevations of 3,400 (1998) and Fraser-Jenkins (2008, 2011) were m and 3,800 m except a few in some valleys. A used in the case of pteridophytes. The individual total of 148 plots were sampled during the study species’ presence/absence data in each studied period of 2011 – 2013. plot were used as the response variable in the current study.

Environmental variables as predictor variables

The following sets of environmental variables were selected as predictor variables (Table 1).

I. The first set of predictor variables included the microclimate (temperature and humidity) data recorded by the logger installed in the field, from 2011 to 2013. The HOBOs (Onset Computer Corporation, Bourne, MA 02532, USA) were used to record air humidity and Fig. 3: Schematic diagram of sampling plot air temperature 2 m above the ground level in design of the study per land use type in all each plot. The HOBOs recorded data in every elevation (C= crop field, E = exploited forest, 30 minutes interval. The soil temperatures M = meadow, F = forest and the straight line at were recorded at 10 cm below the ground the center represents the river) level using Button (Maxim Integrated, San Jose, CA 95134, USA) in each plot. The soil All the species within each plot were recorded. If temperature data were recorded after every 3 the same species occurred in the next plot, it was hours. The mean, minimum and maximum recorded as “1” (in the presence of the species). values of the year-round data were derived The species recorded in the first plot but not in the using the recorded data afterwards (Table 1). second plot were recorded as “0” (in the absence The non-available (NA) values of the data of the species). Two replicate plots of the same were replaced by the mean of the respective land use type were later merged into one. Each variables so that there would be no loss of plot was visited twice in order to record as many data rows in the data frame. species as possible. To reduce the sampling bias caused by spatial auto-correlation, the replicate of II. The second set of predictor variables included each plot was established at least 50 m away from the bioclimatic variables extracted from the the first plot (Magurran, 2004). Worldclim-Global climate data (Hijmans et al., 2005). The data were obtained in 30 arc

6 Rai et al. Banko Janakari, Vol. 26, No. 1

seconds (0.93 km x 0.93 km= 0.86 sq. km.) variables and all the others were taken as the resolution. The latitude and longitude of ratio variables (Table 1). each plot recorded with the help of Garmin 60S GPS were supplied in the DIVA GIS ver. The above set of variables contained large 7.5.0. The software extracted the interpolated number of variables. The Hmisc (Harrell et al., values of the bioclimatic variables from the 2016) Package was used to check the collinearity WORLD CLIM database for each plot. Out among the environmental variables. The Pearson of the 19 bioclimatic variables as defined correlation coefficient was used to describe the by the USGS Data Series 691 (O’Donnell relationships between the variables. The highly and Ignizio, 2012), only 10 less correlated correlated variables (r 0.7) were taken for variables were chosen for further analysis analysis (Dormann et al., 2013). (Table 1). Data analysis III. The third set of data contained the information Initial data recording and management were done of the topography of the studied area, and using MS Excel and MS ACCESS. The further were directly recorded in the field. Garmin analyses were performed on the R ver. 3.1.2 (R GPS 60S was used to record the elevation Core Team, 2015). of the plots. Brunton Compass was used to record the aspect while Clinometer was R-package vegan (Oksanen et al., 2015) was used to record the slope angle of the sample used for the multi-variate ordination analysis. plots. The land use types, the regions and the Detrended Correspondence Analysis (DCA) was valleys were considered as the categorical performed for the species data (Hill and Gauch,

Table 1: The list of environmental variables selected from three sets Set Variable acronym Contained information (1) Loggers’ data MaxT.H maximum air temperature recorded by HOBO MaxT.iB maximum soil temperature recorded by iButton MeanT.iB mean soil temperature recorded by iButton MinT.iB minimum soil temperature recorded by iButton MaxH.H maximum air humidity recorded by HOBO meanH.H mean air humidity recorded by HOBO MinH.H minimum air humidity recorded by HOBO (2) Bioclimatic BIO1 annual mean temperature BIO3 Isothermality of temperature BIO5 maximum temperature of warmest month BIO6 minimum temperature of coldest month BIO8 mean temperature of wettest quarter BIO10 mean temperature of warmest quarter BIO14 precipitation of driest month BIO15 precipitation seasonality BIO17 precipitation of the driest quarter BIO19 precipitation of the coldest quarter (3) Spatial REG two regions (Manaslu and Sagarmatha) VAL four valleys HABIE exploited forest HABIF natural forest HABIM meadow ALTG recorded elevation ASP aspect SLOP slope angle

7 Banko Janakari, Vol. 26, No. 1 Rai et al.

1980) showing the gradient length of the first analyzed using CCA. The species data were ordination axis higher than 2.5 standard units. further constrained separately by the logger data, Therefore, we used the Constrained Ordination bioclimatic data and topographical variables for Method, the unimodal model of the Canonical CCA analysis. The performances of the variables Correspondence Analysis (CCA) (Ter Braak, are presented in Table 2. 1986). The CCA plots show the effect of the environmental The inertias of all the predictors were compared variables on the species composition (Fig. 5, 6 among each other in order to find out the amount and 7). The distribution of the species were found of variances explained by them. The diversity to be affected by the temperature along the CCA indices like Shannon-Wiener, Simpson and axis-1 and the humidity along the CCA axis-2 Inverse Simpson indices were calculated using (Fig. 5) This clearly showed that the temperature “vegan” R Package (Oksanen et al., 2015). and humidity were controlling environmental factors for the distribution of the species (Table Generalized Linear Model (McCullagh and 2). In terms of percentage, the variation explained Nelder, 1989) with quasi-poisson distribution for by the CCA axis-1 and the CCA axis-2 were counts were used to evaluate the relationships ~41.8% and ~27.7%, respectively; thus, 69% of between the species richness as response variable the variation were explained by the two CCA and different environmental predictors. The axes (Table 3). model was fitted against the null model to check for its robustness and performance. The second order polynomial function was also tested, but Fisher’s alpha was not significant. Thus, we proceeded with the first order linear model. Results and discussion

The study revealed 790 vascular plant species of 337 genera within 114 plant families. The highest number of species were recorded for Asteraceae (84 spp.) followed by Rosaceae (52 spp.), Poaceae Fig. 5: CCA plot showing the species (50 spp.) and Fabaceae (38 spp., Fig. 4). composition constrained by humidity and soil temperature; the crosses indicating the species, the circles indicating the plots and the arrows showing the predictors

The precipitation seasonality (BIO15) possesses the longest gradient length to shape the species distribution. Isothermality (BIO3) refers to the percentage of the mean diurnal range divided by the annual temperature. Thus, the growing days and length of the days which shape the temperature pattern has also significant contribution for species distribution. Precipitation of the driest Fig. 4: Bar diagram showing the representative month (BIO14) is another contributor for species families, number of species on the Y-axis and distribution. Annual mean temperature (BIO1), families on the X-axis (families representing mean temperature of coldest month (BIO6), mean more than 10 spp. are included) temperature of the warmest quarter (BIO10) and mean temperature of the wettest quarter Species composition (temperature combined with the precipitation, The Detrended Correspondence Analysis (DCA) BIO8) were found to have the significant effect of the species values against the plots studied was on the species composition in the study areas performed. All of the DCA axes were more than (Table 2 and Fig. 6). 2.5 standard units; therefore, the data were further

8 Rai et al. Banko Janakari, Vol. 26, No. 1

Table 2: The test statistics expressed by the environmental variables while constraining the species composition (by “margin” i.e. each marginal term analyzed in a model with all other variables) Significance Variable Set Code Df Chi Square F Pr(>F) codes Loggers MeanT.iB 1 0.1385 1.7665 0.001 MinT.iB 1 0.1450 1.8498 0.001 MaxH.H 1 0.1187 1.5135 0.002 MinH.H 1 0.1303 1.6617 0.001 Residual 143 11.2110 Bioclimatic BIO1 1 0.1259 1.6474 0.001 BIO3 1 0.1664 2.1761 0.001 BIO6 1 0.1078 1.4100 0.001 BIO8 1 0.1132 1.4808 0.001 BIO10 1 0.1185 1.5506 0.001 BIO14 1 0.1617 2.1150 0.001 BIO15 1 0.1458 1.9071 0.001 Residual 140 10.7033 Spatial REG 1 0.2408 3.2757 0.001 VAL 1 0.2368 3.2217 0.001 HABI 3 0.5225 2.3695 0.001 ALTG 1 0.3770 5.1288 0.001 ASP 1 0.1242 1.6903 0.001 SLOP 1 0.1109 1.5089 0.001 Residual 139 10.2174 Significance codes: ‘***’ forP =0.001, ‘**’ for P=0.002

The CCA axes of the bioclimatic variable were The results obtained by constraining species found to have performed less than the CCA axes with the annual temperature and precipitation obtained from the logger data. The CCA axis- mean and their derivatives show that not only the 1 was found to have explained 24.39% of the mean, minima and maxima of the temperature variation followed by the CCA axis-2 (20.19%), and precipitation are important but also their the CCA axis-3 (14.82%) and the CCA axis-4 combined effect are equally important to shape (12.24%). Thus, a total of 72% of the variation the distribution of the species in the given was found to be explained by these four axes environmental hyper-volume (Hutchinson, 1957). (Table 3). The predictor variables constructed with the derivatives of temperature and precipitation alone and combined have the physiological role in the germination, growth and proliferation (Wright et al., 2006). Soil temperatures are important for the physiology of the cell, water availability and nutrient uptake from the soil (Korner, 2003). Temperature is related with the energy balance as well (Scherrer et al., 2011). Topographical variables also show significant effect upon the species composition (Fig. 7, Table 2).

Fig. 6: CCA plot showing the species composition constrained by the bioclimatic variables; the crosses indicating the species, the circles indicating the plots and the arrows showing the predictors

9 Banko Janakari, Vol. 26, No. 1 Rai et al.

The CCA axis-1 explains ~25% followed by the CCA axis-2 (~21.8%) and the CCA axis-3 (~15.2%) (Table 3).

The valleys (VAL) were also found to be significant for species composition (Table 2). The Sagarmatha region receives more annual precipitation (average 1,640.95 mm) as compared to the MCA region (average 545.36 mm, Hijmans et al., 2005). The valleys in the Sagarmatha region are geographically nearer to the Bay of Bengal, the origin of the monsoon rain system, and are Fig. 7: CCA plot showing the species less rain-shadowed by the high mountains. In composition as shaped by the topographical contrast, the MCA valleys are geographically variables; the crosses indicating the species, farther from the Bay of Bengal and rain-shadowed the circles indicating the plots and the arrows by Mt. Ganesh (7,422 m). showing the predictors Species richness Elevation (ALTG) was found to be one of the significant variables for the species composition For each environmental variable, annual model in our study (Table 2). It is a surrogate of a number was first created and was tested with the first order of environmental factors,e.g. temperature, which Generalized Linear Model (GLM). Transect-wise in turn stands for energy, water etc. Slope angle species richness was taken as response variable (SLOP) and aspect of the plots were also found which regressed against different environmental to be significant contributors for the species variables as predictor. These included land use composition (Table 2). More than 60% of the types (LUT), elevation (ALTG), precipitation variation was found to be explained by the three seasonality (BIO15), annual precipitation CCA axes produced by constraining species (BIO12), slope angle (SLOP) and aspect (ASP) composition with the topographical variables. of the plots. These developed models were tested

Table 3: Percentage of variation explained by the CCA axes when species richness were constrained with the predictor variables Constrained Percentage Cumulative Data set CCA axes Eigenvalues inertia variation explained variation % Loggers' Set 0.750 CCA1 0.3132 41.77 CCA2 0.2076 27.68 69 CCA3 0.1223 16.31 86 CCA4 0.1072 14.30 100 Bioclimatic Set 1.258 CCA1 0.3068 24.39 CCA2 0.2539 20.19 45 CCA3 0.1864 14.82 59 CCA4 0.1539 12.24 72 CCA5 0.1512 12.02 84 CCA6 0.1081 8.60 92 CCA7 0.0975 7.75 100 Spatial Set 1.744 CCA1 0.4344 24.97 CCA2 0.3780 21.72 47 CCA3 0.2643 15.19 62 CCA4 0.2277 13.09 75 CCA5 0.1937 11.13 86 CCA6 0.1073 6.17 92 CCA7 0.0846 4.86 97 CCA8 0.0542 3.11 100

10 Rai et al. Banko Janakari, Vol. 26, No. 1

Table 4: Test statistics of the generalized linear model (GLM) of species richness against the individual environmental variables Resid. Resid. Significance Code Predictors Deviance F Pr(>F) df dev. codes LUT Land Use Types 144 1460 75445 2608.9 < 2.20E-16 ALTG Elevation 146 1357 75547 8574.1 < 2.20E-16 BIO15 Precipitation Seasonality 146 1373 75532 8460.7 < 2.20E-16 BIO12 Annual Precipitation 146 1462 75443 7934.5 < 2.20E-16 SLOP Slope Angle 146 1464 75441 7928.0 < 2.20E-16 ASP Aspect 146 1467 75437 7895.5 < 2.20E-16 Significance codes: ‘***’ forP 0.001 among each other by using “F” statistics. Over- dispersed residual of errors were standardized after application of “quasipoisson” family of distribution of error. The significant environmental variables with deviance and “F” values are indicated in Table 4. The graphics of some more interpretable and statistically significant variables are shown in Fig. 8a–8d.

Fig. 8c: Species richness versus precipitation seasonality

Fig. 8a: Species richness versus elevation of the plots

Fig. 8d: Species richness versus annual precipitation

Note: In Fig. 8b, C = crop field, E = exploited forest, F = natural forest and M = meadow; in Fig. 8c, units are precipitation coefficients and in Fig. 8d, precipitation is in mm.

The species richness increased with the increase in the elevation of the plots studied. The previous studies in Nepal showed the unimodal richness pattern with elevation (Baniya et al., 2010; Grau et al., 2007; Vetaas and Grytnes, 2002). Those Fig. 8b: Species richness versus land use types studies analyzed long elevation gradients whereas this study considered relatively short elevation gradient between 2,200 m and 3,800 m above the

11 Banko Janakari, Vol. 26, No. 1 Rai et al. mean sea level. The short gradient in our study also significantly affected by the aspects in the was not sufficient to test the species richness mountain environments at point-scale (Kroner, humps. However, there are studies which show 2003; Parker, 1991). The variation in the slope the plateau of species richness of birds at high and aspect, thus, result in the variation of the elevation (Patterson et al., 1998). An elevation soil moisture, nutrient cycling and availability of limit of species occurrence is expected for high energy dissipation (Mohammad, 2008) resulting mountains e.g., the Himalaya, always covered in different composition and richness (Carmel with snow and the permafrost. The hump shaped and Kadmon, 1999). unimodal distribution of species richness are expected for such restriction in the absence of any Conclusion environmental gradients (Colwell and Lees, 2000; Colwell et al., 2004) or isolation from other zonal Altogether 790 vascular plant species belonging communities (Lomolino, 2001). However, hump to 114 families were recorded from six river is a union of linear segments at local scale. The valleys studied. Asteraceae (84 spp.) was the result was obtained from only 1,600 m elevation. most dominant family among them. The three Thus, the result from this study could be a sets of environmental variables were used to local phenomenon rather than the large-scaled study their effect on the species composition unimodal pattern found by the earlier researchers. and species richness of vascular plants. The This interpretation resembles quite similar to that loggers recorded the microclimate data of each of Baniya et al. (2012). plot. Soil temperature and humidity of the plots affected the plants composition significantly. Out Four land use types namely (i) crop field, (ii) of 19 bioclimatic variables only seven showed meadow, (iii) exploited forest, and (iv) natural significant effect on the plant composition. forest were studied. The exploited forests were Annual mean temperature (BIO1), isothermality more species-rich, followed by the meadow, the of the temperature (BIO3), minimum temperature natural forest and the crop field. The soil use of the coldest month (BIO6), mean temperature of intensity and fragmentation are thought to be the wettest quarter (BIO8) were the temperature loss of biodiversity (Cousins, 2009; Honnay et related variables. Precipitation of the driest al., 2005; Maitima et al., 2005). This explains month (BIO14) and precipitation seasonality the less richness in the crop field. The species (BIO15) also were significant variables. The richness in the exploited forest is described by topography of the plots (elevation, aspect and the intermediate disturbance hypothesis (Connell, slope) affected the vascular plant composition 1978) and some empirical studies (Townsend and significantly. Nearly 50 percents of the variations Scarsbrook, 1997). were explained by two axes of the CCA in all three sets of environmental variables. Four land In our study, the species richness was found to use types were considered during the study. These have increased significantly with the increase in land use types also affected the species richness the annual precipitation and seasonality (Fig. 8c and composition significantly. The results of the and 8d). Precipitation seasonality is the coefficient study are in accordance with the previous studies. of variation of the monthly precipitation. The However, the unimodal hump of the species four valleys studied have different precipitation richness distribution was not revealed due to seasonality, which is explained by this study. shorter elevation gradient in this study. The different valleys receive varying degree of precipitation shaping different scale of species Acknowledgements richness and their pattern (O’Brien, 1993; Pauses and Austin, 2001). We are grateful to the CDB-WSL Project run by the Central Department of Botany, Tribhuvan The species richness and composition pattern University, Nepal for providing the fund and are also affected by the slope and aspect of the logistics for the field trip. We are also thankful sampling plots (Nuzzo, 1996). The south-facing to the Swiss Federal Institute for Forest, Snow and steeper slopes are drier than the north-facing and Landscape Research, WSL, Switzerland for slopes, and more number of species is expected providing us scholarships for data management towards the wet areas (Kassas and Zahran, 1971; and processing. This study was funded by Pook and Moore, 1966). The temperature is the Swiss National Science Foundation (JRP

12 Rai et al. Banko Janakari, Vol. 26, No. 1

IZ70Z0_131338/1 to CS). The first author Phylogeny Group classification for the orders is indebted to the Ministry of Forests and and families of flowering plants : APG III, Soil Conservation, the Government of Nepal 105–121. for granting him a three-year study leave to accomplish the study. Our sincere thanks go to Chettri, B., Bhupathy, S. and Acharya, B. K. Dr. Keshab Raj Rajbhandary, Dr. Khem Raj 2010. Distribution pattern of reptiles along an Bhattarai and Rita Chhetri for identification eastern Himalayan elevation gradient, India. of plants at the National Herbarium and Plant Acta Oecologica 36 (1): 16–22. Laboratories (KATH), Godawari. The employees Colwell, R. K. and Lees, D. C. 2000. The mid- of KATH and Wilder Places Treks, Kathmandu domain effect: geometric species richness, 15 are duly acknowledged. We appreciate the (2): 70–76. contribution of Ms. Laxmi Sankhi, Ms. Srijana Shah and Mr. Hem B. Katuwal in data collection Colwell, R. K., Rahbek, C. and Gotelli, N. J. during the field-work. We acknowledge Dr. Chitra 2004. The mid-domain effect and species Baniya for sharing his ideas and analysis methods richness patterns: what have we learned so while developing this paper. far? The American Naturalist 163 (3): E1–23.

References Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199: 1302– Baniya, C. B., Solhøy, T., Gauslaa, Y. and Palmer, 1310. M. W. 2010. The elevation gradient of lichen species richness in Nepal. The Lichenologist Cousins, S. A. O. 2009. Extinction debt in 42 (1): 83–96. fragmented grasslands: paid or not? Journal of Vegetation Science 20: 3–7. Baniya, C. B., Solhøy, T., Gauslaa, Y. and Palmer, M. W. 2012. Richness and composition of Dormann, C. F., Elith, J., Bacher, S., Buchmann, vascular plants and cryptogams along a high C., Carl, G., Carré, G. and Lautenbach, S. elevation gradient on Buddha Mountain, 2013. Collinearity: a review of methods to Central Tibet. Folia Geobot 47: 135–151. deal with it and a simulation study evaluating their performance. Ecography 36 (1): 27–46. Bhattarai, K. R. K. and Vetaas, O. R. 2003. Variation in plant species richness of different Fraser-Jenkins, C. 2011. Nepal’s little known life forms along a subtropical elevation pteridophytes, the hidden work of David gradient in the Himalayas, east Nepal. Global Don, and the geography and distribution of Ecology and Biogeography 12 (4): 327–340. Indo-Himalayan ferns. www.groups.yahoo. com/group/Indian-Ferns accessed on 13 Bhattarai, K. R., Vetaas, O. R. and Grytnes, J. A. June, 2012. 2004. Fern species richness along a central Himalayan elevation gradient, Nepal. Journal Fraser-Jenkins, C. 2008. Taxonomic Revision of Biogeography 31 (3): 389–400. of Three Hundred Indian Subcontinental Pteridophytes with a Revised Census-list. Carmel, Y. and Kadmon, R. 1999. Effects Bhisen Singh Mahendra Pal Singh, Dehra of grazing and topography on long term Dun, India. vegetation changes in a Mediterranean Ecosystem in Israel. Plant Ecology 145: Geiger R., Aron, R. H. and Todhunter, P. 1995. The 239–250. Climate near the Ground. Friedr. Viewegund. Sohn Verlagsges. mbH, Braunschweig, Carpenter, C. 2005. The environmental control Wiesbaden, 327–406. of plant species density on a Himalayan elevation gradient. Journal of Biogeography Grau, O., Grytnes, J. and Birks, H. J. B. 2007. 32 (6): 999–1018. A comparison of altitudinal species richness patterns of bryophytes with other plant Chase, M. W., Fay, M. F., Reveal, J. L., Soltis, groups in Nepal, Central Himalaya. Journal D. E., Soltis, P. S., Peter, F. and Kenneth, of Biogeography 34 (11): 1907–1915. J. 2009. An update of the Angiosperm

13 Banko Janakari, Vol. 26, No. 1 Rai et al.

Gregorio, A. D. and Jansen, L. J. M. 2000. tokyo.ac.jp/publish_db/Bulletin/no31/ Land Cover Classification System (LCCS): no31018.html accessed on 14 June, 2013. Version 1.0. FAO. www.fao.org. accessed on 01 March, 2011. Kassas, M. and Zahran, M. A. 1971. Plant life on the costal mountains of the Red Sea, Egypt. Guisan, A. and Zimmermann, N. E. 2000. Journal of Indian Botanical Society 50a: Predictive habitat distribution models in 571–586. ecology. Ecological Modelling 135 (2–3): 147–186. Körner, C. 2003. Alpine Plant Life. Springer- Verlag, Berlin Heidelberg GmbH, Germany. Harrell, F. E. Jr, with contributions from Charles Dupont and many others. 2014. Hmisc: Harrell Körner, C. 2007. The use of “altitude” in Miscellaneous. R package version 3.17–2. ecological research. Trends in Ecology and https://CRAN.R-roject.org/package=Hmisc Evolution 22: 569–574. accessed on 10 July, 2014. Li, Y., Xie, D. and Wang, S. 2006. Impact of land Herzog, S. K., Kessler, M. and Bach, K. 2005. cover types on the soil characteristics in karst The elevation gradient in Andean bird species area of Chongqing. Journal of Geographical richness at the local scale: a foot hill peak Sciences 16 (2): 143–154. and a high elevation plateau. Ecography 28: Lomolino, M. V. 2001. Elevation gradients of 209–222. species-density: historical and prospective Hijmans, R. J., Cameron, S. E., Parra, J. L., Jones, views. Global Ecology and Biogeography P. G. and Jarvis, A. 2005. Very high resolution 10: 3–13. interpolated climate surfaces for global land Magurran, A. E. 2004. Measuring Biological areas. International Journal of Climatology Diversity. Blackwell Science Ltd, USA. 25: 1965–1978. Maitima, J. M., Mugatha, S. M., Redi, S. R., Hill, M. O. and Gauch, H. G. 1980. Detrended Gachimbi, L. N., Majule, A., Lyaruu, H., correspondance analysis: an improved Pomery, D., Mathai, S. and Mugisha, S. 2009. ordination technique. Vegetation 42: 47–58. The linkages between land use changes, land Hofer, G., Wagner, H. H., Herzog, F. and Edwards, degradation and biodiversity across East P. J. 2008. Effects of topographic variability Africa. African Journal of Environmental on the scaling of plant species richness in Science and Technology 3 (10): 310–325. gradient dominated landscapes. Ecography McCain, C. M. and Grytnes, J. A. 2010. Elevation 31: 131–139. Gradients in Species Richness. John Wiley Honnay, O., Jacquemyn, H., Bossuyt, B. and and Sons Ltd, Chichester, UK. Hermy, M. 2005. Forest fragmentation effects McCain, C. M. 2007. Area and mammalian on patch occupancy and population viability elevation diversity. Ecology 88 (1): 76–86. of herbaceous plant species. New Phytologist 166 (3): 723–736. McCullagh, P. and Nelder, J. 1989. Generalised Linear Models. 2nd edition. Chapman and Hutchinson, G. E. 1957. Concluding remarks. Hall, London, UK. Cold Spring Harbor Symposium on Quantitative Biology 22: 415–427. Mohammad, A. 2008. The effect of slope aspect on soil and vegetation characteristics in Island, H. 2012. Bird species richness along an Southern west Bank. Bethlehem University elevation gradient in a forest at Jianfengling, Journal 27: 11–27. Hainan Island, China. Zoological Studies 51 (3): 362–371. Nuzzo, V. A. 1996. Structure of cliff vegetation on exposed cliffs and the effect of rock climbing. Iwatsuki, K. 1998. An enumeration of the Canadian Journal of Botany 74: 607–617. Pteridophytes of Nepal. http://www.um.u-

14 Rai et al. Banko Janakari, Vol. 26, No. 1

O’Brien E.M. 1993. Climatic gradients in Press, J. R., Shrestha, K. K. and Sutton, D. woody plant species richness: towards A. 2000. Annotated Checklist of the and explanation based on an analysis of Flowering Plants of Nepal. The Natural southern Africa’s woody flora. Journal of History Museum, London, UK. Biogeography 20: 181–198. R Core Team. 2015. R: A language and O’Donnell, M. S. and Ignizio, D. A. 2005. environment for statistical computing. Bioclimatic predictors for supporting R Foundation for Statistical Computing, ecological applications in the Conterminous Vienna, Austria. URL https://www.R-project. United States. US Geological Survey Data org. Series 691, 10. Sanders, N. J. and Rahbek, C. 2012. The patterns Oksanen, J., Blanchet, F. G., Kindt, R., Legendre, and causes of elevation diversity gradients. P., Minchin, P. R., O’Hara, R. B., Simpson, Ecography 35 (1): 1–3. G. L., Solymos, P., Stevens, M. H. H. and Wagner, H. 2015. vegan: Community Scheidegger, C., Nobis, M. P. and Shrestha, K. K. Ecology package. R package version 2.3–0. 2010. Biodiversity and livelihood in land-use https://CRAN.R-project.org/package=vegan. gradients in an era of climate change – outline accessed on 13 March, 2016. of a Nepal-Swiss research project, 7–17.

Parker K. C. 1991. Topography, substrate and Scherrer, D., Körner, C., Korner, C., Körner, C., vegetation patterns in the northern Sonoran Ko, C. and Korner, C. 2011. Topographically desert. Journal of Biogeography 18: 151–163 controlled thermal-habitat differentiation buffers alpine plant diversity against climate Patterson, B. D., Stotz, D. F., Solari, S., Fitzpatrick, warming. Journal of Biogeography 38 (2): J. W. and Pacheco, V. 1998. Contrasting 406–416. patterns of elevation zonation for birds and mammals in the Andes of south-eastern Peru. Stainton, A. 1988. Flowers of the Himalaya: A Journal of Biogeography 25: 593–607. Supplement. Oxford University Press, New Delhi, India. Paudel, S. and Vetaas, O. R. 2014. Effects of topography and land use on woody plant Stevens, G. C. 1989. The latitudinal gradient in species composition and beta diversity in geographical range: how so many species an arid Trans-Himalayan landscape, Nepal. coexist in the tropics. The American Journal of Mountain Science 11 (5): 1112– Naturalist 133 (2): 240–256. 1122. Stevens, G. C. 1992. The elevation gradient in Pauses, J. D. and Austin, M. P. 2001. Patterns altitudinal range: an extension of rapoport’s of plant species richness in relation to latitudinal rule to altitude. The American different environments: an appraisal. Naturalist 140 (6): 893–911. Journal of Vegetation Science 12: 153–166. Ter Braak, C. J. F. 1986. Canonical correspondance Pianka, E. R. 1966. Latitudinal gradients in analysis: a new eigenvector t technique for species diversity: a review of concepts. multivariate direct gradient analysis. Ecology American Naturalist 100: 33–34. 67 (5): 1167–1179.

Polunin O. and Stainton A. 1984. Flowers of the Tilman, D., Knops, J., Wedin, D., Reich, P., Himalaya. Oxford University Press. New Ritchie, M. and Sieman E. 1997. The influence Delhi, India. of functional diversity and composition on ecosystem processes. Science 277: 1300- Pook, E. W. and Moore, W. E. 1966. The influence 1302. of aspect on the composition and structure of dry sclerophyll forest on Black Mountain, Townsend, C. R., and Scarsbrook, M. R. 1997. The Canberra. Australian Journal of Botany 14 intermediate disturbance hypothesis, refugia (2): 223–242. and biodiversity in streams. Limnology and Oceanography 42 (5): 938–949.

15 Banko Janakari, Vol. 26, No. 1 Rai et al.

Trigas, P., Panitsa, M. and Tsiftsis, S. 2013. Wright, I. J., Reich, P. B., Atkin, O. K., Lusk, C. Elevation gradient of vascular plant species H., Tjoelker, M. G. and Westoby, M. 2006. richness and endemism in Crete-the effect Irradiance, temperature and rainfall influence of post-isolation mountain uplift on a leaf dark respiration in woody plants: continental island system. PloS One 8 (3): evidence from comparisons across 20 sites. e59425. accessed on 10 June, 2013. The New Phytologist 169 (2): 309–19.

Vetaas, O. R., and Grytnes, J. A. 2002. Distribution of vascular plant species richness and endemic richness along the Himalayan elevation gradient in Nepal. Global Ecology and Biogeography 11 (4): 291–301.

16 Mapping deforestation and forest degradation using CLASlite approach in Eastern Churia of Nepal

S. Khanal1* and A. Khadka1 Monitoring deforestation and forest degradation is essential for forest conservation and sustainable management. Those activities have become more relevant in order to get reference emission level required for Reducing emissions from deforestation and forest degradation (REDD) initiative. The study aimed to assess forest degradation and deforestation in the Churia region of Eastern Nepal using CLASlite approach. This approach is based on Spectral Mixture Analysis and provides highly automated technique for forest cover, deforestation and forest degradation mapping. The Landsat imageries of 2002 and 2013 were processed for estimation of deforestation and forest degradation. The validation of results based on the high-resolution multi-temporal Google Earth imageries and the field sample plots indicated that CLASlite approach could be feasible approach to monitor forests for deforestation and degradation. The results can be further improved by including more frequent time-series observation from Landsat.

o Churia, CLASlite, deforestation, forest degradation, spectral mixture analysis he importance of forest resources has been this area is the youngest mountain range in the Twell understood because of their important Himalayas and have high potential of erosion services to the society such as supporting rural hazards. The significance of Churia conservation livelihood, carbon storage, climate change has been highlighted for quite some time. For mitigation, and biodiversity conservation. The instance, Gurung and Khanal (1986) reported issue of forest conservation is more directly the significant land use change in Churia, and relevant in the case of Nepal, as forest resources recommended for detailed study of forestry are significant for ecosystem balance and people’s among other sectors. According to the latest forest livelihood (Birch et al., 2014). Further, because of resource assessment (2010–2014) results, this pertinent issue of deforestation and degradation region has 73.0% (1,384,445 ha) of the total area (Panta et al., 2008) timely monitoring forest under forests; majority of the forest area (76%) resources is very essential. Successful design and falling outside the protected areas with the annual implementation of programs to reduce emissions rate of forest cover change of –0.21% between from deforestation and forest degradation and the period 2001–2010 (DFRS, 2014). promote reforestation (REDD+) would require periodic monitoring of deforestation and forest Till now, much focus has been given to quantify degradation (Goetz et al., 2015; Herold and and monitor deforestation, but the understanding Johns, 2007). as well as mapping of spatial distribution of forest degradation has remained far behind. Nepal is divided into five physiographic regions: Furthermore, for deforestation monitoring, High Himalaya, High Mountain, Middle several standard methods have been tested and Mountain, Siwaliks or Churia and Terai (LRMP, are available for user community. However, 1986). Churia is a belt of hilly region stretching getting reliable estimates of forest degradation from East to West in the entire length of Southern remains a challenge (DeFries et al., 2007). Nepal. This region is ecologically diverse, and Forest degradation is generically defined as the as it provides several ecosystem services to the “reduced capacity of a forest to provide goods areas downstream, it has direct influence on the and services” (FAO, 2002). Some major causes of quality of the environment. However, the Churia forest degradation can be one or the combination Hills are structurally weak (Khanal, 1989) since of processes such as selective logging, conversion

1 Department of Forest Research and Survey, Kathmandu, Nepal. *E-mail: [email protected]

17 Banko Janakari, Vol. 26, No. 1 Khanal and Khadka of land cover and natural disturbances (such as landslide, fire, flood etc.). Thus, the time-scale of processes leading to forest degradation can range from few years to a few decades. Grainger (1993) has defined forest degradation as a process leading to a “temporary or permanent deterioration in the density or structure of vegetation cover or its species composition”. The ultimate consequence of degradation is, therefore, reduced productivity of forests due to impact of disturbances. Reduction in the canopy cover is observed to indicate forest Fig. 1: Map showing the Churia in the Eastern degradation while the estimates of forest canopy Development Region of Nepal cover is estimated on the basis of remotely sensed observations (Wang et al., 2005; Souza et al., Method for forest degradation monitoring 2003). Different methods have been tried for forest This study was conducted under the broad theme degradation monitoring ranging from use of of the Department of Forest Research and Survey LiDAR (Jubanski et al., 2013), RADAR (Ryan (DFRS) to map and monitor Nepal’s forests et al., 2012) and Landsat time series analysis with advanced forest monitoring technology. (Healey et al., 2005). Given the challenges associated with mapping forest degradation, it is very essential to validate The Spectral Mixture Analysis (SMA) approach and apply the available forest degradation has been developed to extract information at and deforestation monitoring approaches. The sub-pixel level. It decomposes mixed pixels into outcome of such studies has a high relevance to fractions of end-members, and has been proposed decide on the robust approach that can be applied to overcome the mixed pixel problem found in for other parts of Nepal and also to integrate forest degraded forests (Souza et al., 2003). The SMA assessment systems for periodic monitoring of approach (Equation 1) has been applied with deforestation and forest degradation. CLASlite SPOT Satellite data (Souza et al., 2003) and has been demonstrated as an applicable tool Landsat data (Asner et al., 2005; Souza et al., to support REDD+ initiative for reliable forest 2005; Negron-Juarez et al., 2011) with varying cover estimation in order to support sub-national success for purposes ranging from canopy gap to national reference levels (Reimer et al., 2015). detection and biomass estimation to delineation However, most of the studies on applicability of logged areas. of CLASlite approach has been conducted on ecosystems much different than in Nepal (Allnutt ...... (1) et al. 2013; Bryan et al. 2013; Carlson et al. 2012). This study examined the applicability of CLASlite Where, R = Reflectance for each band b in the image, approach for assessment of deforestation and b forest degradation in Nepal’s Churia forest. N = No. of end-members, fi = Fraction of end member “i”, Materials and methods Ri,b = Reflectance of end-member “i” in band b, and Study area b = Unmodeled residual. The study area encompasses the Churia in the Souza et al. (2005) developed Normalized Eastern Development Region (EDR) of Nepal Difference Fraction Index (NDFI) to combine (Fig. 1). The major portions of the Churia belt in the information from SMA significant for forest the EDR are located in Siraha, Udayapur, Saptari, degradation monitoring into one band, and Sunsari, Morang, Dhankuta, Jhapa and Ilam attained 94% accuracy in detecting forest areas districts. with canopy damage. Intact forests have high NDFI values because of higher green vegetation while non-photosynthetic vegetation as well as soil fraction increase as forests are degraded, and thus NDFI values become lower.

18 Khanal and Khadka Banko Janakari, Vol. 26, No. 1

One of the tools that support the implementation and also to identify larger areas that underwent of SMA is CLASlite program. It includes a set through changes so that the implementation of tools that automate radiometric correction agencies could identify the areas requiring and use Monte Carlo Unmixing (MCU) to restoration interventions with the help of the produce estimates of the percentage cover of outputs obtained. soil, photosynthetic vegetation (PV), and non- photosynthetic vegetation (NPV) in every image Table 1: Details on the Landsat images used pixel (Asner et al., 2009). Thus, the output Path- Cloud Satellite Sensor Date raster would have three fractional cover classes: Row Cover i) Photosynthetic vegetation (PV) i.e. Live Landsat 7 ETM+ 139-041 2002-12-16 8.62 vegetation ii) Non-photosynthetic vegetation 140-041 2002-12-23 3.88 (NPV) i.e. Dead or senescent vegetation and iii) Landsat 8 OLI /TIRS 139-041 2013-12-06 5.86 Bare substrate (S) i.e. Soil. We used CLASlite version 3.2 package (CLASlite Team, 2014) for 140-041 2013-11-11 5.22 the purpose of our study. Default threshold values were used for the estimation of forest as well as detection of forest cover loss (deforestation) and degradation (areas of persistent forest disturbance).

Datasets used Landsat images were selected in two time periods for detecting deforestation and forest degradation. The available scenes were filtered to be within close months and have the least cloud cover as much as possible. The details on the Landsat Fig. 2: Footprint of Landsat scenes over the scenes used in the analysis are presented in Table study area. The images shown are fractional 1, while the footprints of those scenes are shown cover derived from Landsat 8 of 2013 in Fig. 2. Results and discussion The resulting deforestation and degradation outputs were generalized by filtering the pixel The outputs from the analysis provided the map groups that were less than 0.5 hectare. This of the areas that had deforestation and forest was done in order to avoid the scattered pixels degradation between 2002–2013 (Fig. 3).

Fig. 3: Map showing the deforested and degraded areas in the studied Churia region between 2002–2013

19 Banko Janakari, Vol. 26, No. 1 Khanal and Khadka

The district-wise area of deforestation and degradation in the Churia region was calculated. For comparison, the data from Hansen et al. (2013) was used which has analyzed time series Landsat images to estimate global-level forest extent and change from 2000 to 2013. The result for the study area extent is presented in the Table 2, so as to make general comparison. The last column is the deforestation estimated based on Hansen et al., 2013. However, it is important to note that the analysis was done at global scale, and thus requires local-level validation. was found to have the highest area of deforestation and forest degradation (Table 2). Table 2: District-wise area of forest degradation and deforestation estimated Deforestation Deforestation Degradation District (ha) (Hansen (ha) (ha) et al. 2013) Dhankuta 27.36 17.01 1.44 Ilam 95.22 111.51 137.07 Fig. 4: Example of a deforestation site mapped Jhapa 28.71 61.47 36.81 in Sukrebaray area of on the Morang 28.35 228.15 36.09 Google Image of Nov 2002 (upper panel) and Saptari 242.64 254.61 352.35 Dec 2013 (lower panel). The polygon with Siraha 921.33 500.31 412.83 white boundary in the lower panel shows the Sunsari 96.39 171.54 13.68 mapped deforested area. Udayapur 2,904.48 1,420.11 680.31

The degradation and deforestation map produced was validated in the field using 40 sample sites. Those plots were generated randomly over the detected deforestation and degradation sites. The deforested areas were also cross-examined using the available high-resolution multi-temporal imageries. Several examples of interesting observations of deforestation were made. One of the examples is presented in Figure 4. The forest cover loss was more clearly detected (Fig. Fig. 5a 5a,b,c) as compared to the forest degradation as it required more field observation to identify why a certain pixel was recorded under the category of degraded areas. The observations indicated that the areas identified as degraded had some major characteristics of disturbance such as the areas with lopped-trees, the areas with selective logging and the areas covered by dense invasive shrubs (Fig. 5a,b,c). Some natural disturbances on those sites included landslide, fire impact and river deposition. Fig. 5b

20 Khanal and Khadka Banko Janakari, Vol. 26, No. 1

examples of changes, the approach seems to be promising and would need further in-depth analysis. The results can be potentially further improved by including more frequent time-series observations which are becoming more practical option with the availability of newer images including Landsat. The information generated through this study are expected to be useful for identifying the deforested areas and or degraded forest areas on the one hand and for identifying the areas to be reforested or planted on the other hand. Fig. 5c References

Fig. 5a,b,c: Degraded forest observed at the Allnutt, T. F., Asner, G. P., Golden, C. D. North of Bengari in . The and Powell, G. V. 2013. Mapping recent area in the Fig. 5b (Apr, 2014) and the white deforestation and forest disturbance in north polygons show the mapped degraded sites eastern Madagascar. Tropical Conservation since May, 2003 (Fig. 5a). Field observation Science 6 (1): 1–15. indicated trees removed and the area covered with dense invasive shrubs (Fig. 5c). Asner, G. P., Knapp, D. E., Cooper, A. N., Bustamante, M. M., and Olander, L. P. Spectral libraries derived from extensive field 2005. Ecosystem structure throughout the databases and satellite imageries have been used Brazilian Amazon from Landsat observations to represent tropical forests ranging from lowlands and automated spectral unmixing. Earth to mountain ecosystems in the newer version Interactions 9 (7): 1–31. of the CLASlite Software Package (CLASlite Team, 2014). This could be the reason behind the Asner, G. P., Knapp, D. E., Balaji, A., and Páez- interesting performance of the package in spite of Acosta, G. 2009. Automated mapping of being originally developed for different ecosystem tropical deforestation and forest degradation: than a mountainous country like Nepal. However, CLASlite. Journal of Applied Remote Sensing future studies should test and identify proper 3 (1): 033543. threshold of parameters and select the best ones based on the validation results. Further, there are Birch, J. C., Thapa, I., Balmford, A., Bradbury, always challenges associated with shadows in the R. B., Brown, C., Butchart, S. H., Gurung, case of a mountainous country like Nepal. These H., Hughes, F. M., Mulligan, M., Pandeya, issues can be potentially addressed through more B. and Peh, K. S. H. 2014. What benefits do frequent time-series observations of cloud-free community forests provide, and to whom? satellite imageries. A rapid assessment of ecosystem services from a Himalayan forest, Nepal. Ecosystem Conclusion Services 8: 118–127.

The findings from this study indicate that Bryan, J. E., Shearman, P. L., Asner, G. P., Knapp, CLASlite-based analysis can be one potential D. E., Aoro, G., and Lokes, B. 2013. Extreme approach to monitor forests for deforestation differences in forest degradation in Borneo: and forest degradation. In this study, some areas comparing practices in Sarawak, Sabah, and that underwent through such changes could not Brunei. PloS One 8 (7): e69679. be accounted for due to the lack of information induced by clouds and shadows in the Carlson, K. M., Curran, L. M., Ratnasari, D., mountainous regions. This issue can be possibly Pittman, A. M., Soares-Filho, B. S., Asner, minimized by incorporating cloud-free image G. P., Trigg, S. N., Gaveau, D. A., Lawrence, composites. However based on validation with D. and Rodrigues, H. O. 2012. Committed high-resolution multi-temporal Google Earth carbon emissions, deforestation, and Imageries and as observed in some interesting community land conversion from oil palm

21 Banko Janakari, Vol. 26, No. 1 Khanal and Khadka

plantation expansion in West Kalimantan, Krankina, O. N., 2005. Comparison of Indonesia. In Proceedings of the National Tasseled Cap-based Landsat data structures Academy of Sciences 109 (19): 7559–7564. for use in forest disturbance detection. Remote Sensing of Environment 97 (3): 301–310. CLASlite Team, 2014. CLASlite Forest Monitoring Technology: Version 3.2 User Herold, M. and Johns, T. 2007. Linking Guide. Carnegie Institution for Science, requirements with capabilities for Stanford, CA. http://claslite.carnegiescience. deforestation monitoring in the context of the edu accessed on 23 Dec, 2015. UNFCCC-REDD process. Environmental Research Letters 2 (4): 045025. DeFries, R., Achard, F., Brown, S., Herold, M., Murdiyarso, D., Schlamadinger, B., Jubanski, J., Ballhorn, U., Kronseder, K., Franke, and de Souza, C. 2007. Earth observations J. and Siegert, F. 2013. Detection of large for estimating greenhouse gas emissions above-ground biomass variability in lowland from deforestation in developing countries. forest ecosystems by airborne LiDAR. Environmental Science and Policy 10 (4): Biogeosciences 10: 3917–3930. 385–394. Khanal, N.R. 1989. Morphometric properties DFRS. 2014. Churia Forests of Nepal. Forest of the drainage basins in the Chure range, Resources Assessment Nepal Project, Nepal. Himalayan Review 20: 59–65. Department of Forest Research and Survey (DFRS), Kathmandu, Nepal. LRMP. 1986. Land System Report. Land Resource Mapping Project (LRMP): Kenting FAO. 2002. Proceedings: Second Expert Meeting Earth Science, Canada. on Harmonizing Forest-related Definitions for Use by Various Stakeholders. 11-13 Negron-Juarez, R. I., Chambers, J. Q., Marra, September 2002, Rome. http://www.fao.org/ D. M., Ribeiro, G. H. P. M., Rifai, S. W., docrep/005/y4171e/y4171e00.htm accessed Higuchi, N. and Roberts, D. 2011. Detection on 23 Dec, 2015. of subpixel treefall gaps with Landsat imagery in Central Amazon forests. Remote Goetz, S., Hansen, M., Houghton, R. A., Sensing of Environment 115: 3322–3328. Walker, W., Laporte, N. T. and Busch, J. 2015. Measurement and Monitoring for Panta, M., Kim, K. and Joshi, C. 2008. REDD+: The Needs, Current Technological Temporal mapping of deforestation and Capabilities, and Future Potential. Center for forest degradation in Nepal: applications Global Development Working Paper, 392. to forest conservation. Forest Ecology and Management 256 (9): 1587–1595. Grainger, A. 1993. Controlling Tropical Deforestation. Earthscan Publications, Reimer, F., Asner, G. P., and Joseph, S. 2015. London, UK. Advancing reference emission levels in subnational and national REDD+ initiatives: Gurung, H. and Khanal, N. 1986. Landscape a CLASlite approach. Carbon Balance and processes in the Chure Range Central Nepal. Management 10 (5). doi: 10.1186/s13021- Himalayan Review 17: 1–39. 015-0015-8.

Hansen, M. C., Potapov, P. V., Moore, R., Hancher, Ryan, C. M., Hill, T., Woollen, E., Ghee, M., Turubanova, S. A., Tyukavina, A., Thau, C., Mitchard, E., Cassells, G., Grace, J., D., Stehman, S. V., Goetz, S. J., Loveland, T. Woodhouse, I. H. and Williams, M. 2012. R., Kommareddy, A., Egorov, A., Chini, L., Quantifying small-scale deforestation and Justice, C. O. and Townshend, J. R. G. 2013. forest degradation in African woodlands High-resolution global maps of 21 -century using radar imagery. Global Change Biology forest cover change. Science 342: 850–853. 18 (1): 243–257.

Healey, S. P., Cohen, W. B., Zhiqiang, Y. and Souza Jr., C., Firestone, L., Silva, M. L., Roberts

22 Khanal and Khadka Banko Janakari, Vol. 26, No. 1

and D.A. 2003. Mapping forest degradation Sensing of Environment 98: 329–343. in the Eastern Amazon from SPOT 4 through spectral mixture models. Remote Sensing of Wang, C., Qi, J., and Cochrane, M. 2005. Environment 87: 494–506. Assessment of tropical forest degradation with canopy fractional cover from Landsat Souza Jr., C., Roberts, D. A. and Cochrane, M. ETM+ and IKONOS imagery. Earth A. 2005. Combining spectral and spatial Interactions 9 (22): 1–18. information to map canopy damage from selective logging and forest fires. Remote

23 A comparative study on carbon stock in Sal (Shorea robusta) forest in two different ecological regions of Nepal H. P. Pandey1 and M. Bhusal2

Estimation of total biomass and carbon sequestration in any forest is crucial as it gives ecological and economic benefits through various environmental services. With an aim to quantify the carbon stock densities in the two different ecological regions–the Hills and the Terai, two Community Forests (CFs) having the dominance of Shorea robusta were selected from Gorkha (in the Hills) and Chitwan (in the Terai) districts for the purpose of the study. Systematic random sampling with 1% sampling intensity was used to collect necessary data. The total carbon stock in the CFs of the Hills and the Terai were found to be 234.54 t ha-1 and 479.29 t ha-1, respectively. The biomass carbon stock density in the CF of the Terai was found to be higher (384.20 t ha-1) than the one in the Hills (123.15 t ha-1). Carbon densities of different carbon pools such as tree; sapling; leaf litter, grass and herbs were significantly higher (P<0.05) in the Terai than in the Hill forest whereas dead wood and stumps and the soil organic carbon density were found to be not significantly different in these regions. Similarly, the highest amount of soil organic carbon (SOC) was found in the uppermost soil horizon in the forests of both the regions. These results revealed that the biomass carbon stock density was higher in the Terai S. robusta forest than in the Hill S. robusta forest. However, the SOC obtained was in inverse relation to that of the biomass carbon stock in both the ecological regions. It would not be biased if different ecological regions with similar forest types are intervened with different management strategies for having more carbon stocks and for the conservation of biodiversity in the days to come.

o Biomass carbon, carbon stock density, ecological regions, Shorea robusta, soil organic carbon

arbon stock is the absolute quantity of carbon (Banskota et al., 2007), and so it has become Cheld within a pool at a specified time whereas a least cost-solution in the world. Secondly, carbon sequestration is the process of increasing managing forests sustainably in tropical areas will the carbon content of a carbon pool other than the greatly reduce carbon emissions as it is estimated atmosphere (FAO, 2011). In other words, carbon that global deforestation alone accounts for about sequestration is understood as the removal of 17.4% (IPCC, 2007) of the global greenhouse carbon from the atmosphere by storing it in the gases emissions. Thirdly, there is a high potential biosphere (IPCC, 2000). Forests play a profound for enhancing the carbon sequestration in the role in reducing ambient carbon dioxide (CO2) vegetation and soils of the Himalayan region levels as they sequester 20–100 times more through improved management of degraded lands carbon per unit area than croplands (Brown and (Upadhyay et al., 2005). Pearce, 1994). Hence, vegetation and soils play role of viable sinks of atmospheric carbon and Soil organic carbon (SOC) stocks display a high may significantly contribute mitigation of global spatial variability (Cannell et al., 1999). In fact, climate change (Bajracharya et al., 1998; Lal, most of the studies concern only the topsoil (e.g. 0–30 cm), although carbon sequestration or loss 2004). Biological sequestration of CO2 by forest has numerous benefits to other emission reduction may also occur in deeper soil layers (Bird et al., technologies. Firstly, it is considered to be more 2002; Fontaine et al., 2007). SOC is an important effective than other carbon sequestration methods index of soil quality because of its relationship to

1 REDD Implementation Centre, Kathmandu, Nepal. Email: [email protected] 2 Nepal Herbs and Herbal Products Association, Kathmandu, Nepal

24 Pandey and Bhusal Banko Janakari, Vol. 26, No. 1 crop productivity (Lal et al., 1997). In this regard, Nepal so as to generate carbon revenues as well as SOC sequestration has numerous economic as non-carbon benefits for the nation and her people well as ecological implications in terms of trading where preliminary estimates show that REDD+ carbon credits, improving quality of soil and may bring between $20-86 million per year to water resources, and achieving food security. Nepal (UN-REDD, 2014). To provide somewhat base-line information regarding REDD+ and to Sal (Shorea robusta) is the most dominant species improve the carbon sink of the forests (REDD- of the tropical and subtropical broadleaved IC, 2015), this study could be a reference forests of Nepal (Jackson, 1994). The Shorea document since forest reference level is yet to robusta forest in Nepal is confined to the Hills be finalized in the national scenario. Therefore, and Terai ecological regions. It shares the highest the first step is to estimate the amount of carbon 3 tree volume i.e., 109.4 million m (28.2% of the stocks in the forests of different conditions and total tree volume) (Amatya and Shrestha, 2010). types for getting credits from forest carbon in S. robusta forests not only have higher economic days to come. Hence, the present study aims to value, but also serve as an important ecological quantify and compare the carbon stock densities benefit in the form of abating global warming and and to compare the soil carbon densities with at climate change through conserving atmospheric different depths in the community forests having CO2 (Shrestha, 2008). dominance of S. robusta in the two different ecological regions of Nepal. About 1.45 million households or 35 percent of the population of Nepal are involved in Materials and methods community-based forest management program through forming 17,685 Community Forest User Study area Groups (CFUGs) who are managing 1,652,654 Located between 27o15’ and 28o45’ N latitude and hectares of national forest handed over to them between 84o27’ and 84o57’ E longitude, Gorkha (DOF, 2015). Among these CFUGs, few of them District lies in the hilly region of Central Nepal. have got reward for the carbon enhancement in The total area of the district is 3,610 km2 with an their forests (Adhikari, 2016) as pilot programs. altitudinal variation of 228–8,163 m above the There is still a hope that the financial resources mean sea level (amsl) (CBS, 2006). The district required to assist the developing countries in exhibits wide range of climate, from sub-tropical undertaking mitigation, and adaptation activities in the south to alpine in the north, with 23.1oC will become more and more significant in the mean annual temperature and 1,500.20 mm future (UNFCCC, 2015) including Nepal. To average annual precipitation (DHM, 2015). Two ensure the carbon credit getting from one or other community forests having the dominance of S. programs, it is very crucial to account carbon robusta were selected from Gorkha (in the Hills) accumulation in forest to make the communities and Chitwan (in the Terai) districts for the study get benefited from carbon credits and biodiversity (Fig. 1). conservation. In spite of the protection of forests from local communities for 33 years, only a little One of the two CFs selected for the purpose was attention has been paid so far to forest and soil Ghaledanda Ranakhola CF situated in the Bunkot inventories considering carbon stocks. Hence, Village Development Committee (VDC) of amount of soil and biomass carbon sequestrated Gorkha District. The CF selected for the purpose is unknown (Shrestha, 2008) in these forests. In lies in the sub-temperate region with an area Nepal, most of the studies have been conducted of 181.63 ha and with 750–1,000 m altitudinal in forests for their tangible economic benefits variation amsl (CBS, 2006). Most of the land is whereas a very few studies have been done on plain consisting of clayey loam soil; the soil is intangible benefits such as carbon sequestration reddish under the S. robusta cover and blackish and biodiversity conservation. Information on near the water holes. The CF is managed by a total carbon stocks in different forest ecosystems in of 459 households with 3,014 family members. Nepal is still lacking (Shrestha and Singh, 2008). The forest is consisted of medium-sized trees and dominated by S. robusta associated with Schima The participation in Reducing Emissions from wallichii, Castanopsis indica and Lagerstroemia Deforestation and Forest Degradation plus parviflora. (REDD+) mechanism has brighter prospect for

25 Banko Janakari, Vol. 26, No. 1 Pandey and Bhusal

Fig. 1: Location map of two study areas in Chitwan and Gorkha districts of Nepal Another CF selected for the purpose was sampling technique with 1% sampling intensity Nawajyoti Buffer Zone CF situated in the was used. For the purpose of plot allocation, Barandabar Corridor Forest of Geetanagar VDC both the forests were stratified on the basis of in . Located between 27o21’ N species composition, and then, on the basis of and 27o52’ N latitude and between 83o54’ E and proportion allocation, concentric circular sample 84o48’ E longitude, Chitwan District lies in the plots (CCSPs) were allocated randomly in each Inner–Terai region of Central Nepal; the total area stratum. A total of 75 and 20 CCSPs with the radii of the district is 2,205.90 Km2 with an altitudinal of 8.92 m (for measuring trees and poles), 5.64 variation of 141 m to 1,947 m amsl (CBS, 2006). m (for measuring saplings), 1 m (for measuring Moreover, subtropical type of climate prevails in seedlings) and 0.56 m (for taking the samples the district with the yearly average temperature of of the leaf litter, herbs, grass and soil) (Fig. 2) 26oC and 1,512 mm average annual precipitation were laid out in the CFs of Gorkha and Chitwan (DHM, 2015). The total area of the CF is 44.7 ha Districts, respectively using the method described with an altitudinal variation of 175 m to 250 m by ANSAB (2010). amsl, and is managed and conserved by the Buffer Zone Community Forest User Group (BZCFUG) with a total of 302 households with 1,791 family members. Most of the land is plain consisting of clayey loam soil. The soil is reddish under the S. robusta cover and blackish near the water holes, similar to the one in the CF of Gorkha District. Likewise, there are yellowish fertile alluviums in some parts of the forest. The community forest is primarily dominated by S. robusta with a number of broad-leaved associate species including Fig. 2: Layout of a CCSP in the field Sapium insigne and Bombax ceiba. Measurement of sample plots

Forest sampling and plot allocation The diameter (at breast height) and height of all For forest carbon inventory, systematic random the trees, poles and saplings were measured using D-tapes and Suunto-clinometers, respectively

26 Pandey and Bhusal Banko Janakari, Vol. 26, No. 1 while the seedlings were counted in the respective in GoN, 2011). The carbon stock density was sample plots and all the herbaceous and woody calculated by summing the carbon stock densities vegetation inside the innermost 0.56 m–radius of the individual carbon pools. In order to plots were clipped and collected separately to compare the carbon stock densities between the take the fresh weights; the representative sub- two ecological regions, t-tests were performed samples of 300 m of all the herbaceous plants, at 5% level of significance with the help of R- woody vegetation, leaf litters and grasses were Statistical Package Version 2.11.1 (RDCT, 2012). taken to the soil laboratory of the Institute of The graphs and tables were constructed using Forestry (IOF), Pokhara for oven drying. MS-Excel 2010.

Soil sample collection Results and discussion

The soil samples from various depths (0–20 cm, Biomass carbon stock density 20–40 cm, 40–60 cm, 60–80 cm and 80–100 cm) were collected using the methodology adopted The total carbon stock density of the forest by Awasthi et al. (2005). All the samples were vegetation including carbon in the trees, saplings, bagged, labeled and sent to the soil laboratory leaf litters, herbs and grass together with the dead for further analysis. The pit method was used wood and stumps was found to be 123.15 t ha-1 for determining the bulk densities (Pearson et and 384.20 t ha-1 in the Shorea robuta forests of al., 2005) of the soil samples. Besides, several the Hills and the Terai, respectively. The above published and unpublished information were also ground biomass carbon stock density comprised collected from the secondary sources so as to of 91% among the total biomass carbon stock fulfill the objectives of the study. density and the rest 9% belonging to the below ground (root) in both the ecological regions Data analysis (Table 1). The bulk amount of carbon was found to be stored in the trees (including poles) The biomass and carbon stock density from followed by the leaf litters, herbs and grasses each site were analyzed using the allometric in both the regions. The t-test showed that there equation and the guidelines mentioned in “Forest was significant difference (P<0.05) in the carbon Carbon Inventory Directives, 2011” prepared stock densities of these pools with respective by the REDD Implementation Centre under the to the ecological regions; the amount of carbon Ministry of Forests and Soil Conservation of the density was found to be significantly higher in the Government of Nepal (GoN, 2011). Terai forest than in the Hill forest. On the other The analysis of the soil samples were performed, hand, the dead wood and stump carbon stock adopting Walkey and Black Method (Jackson, density had no significant difference (P>0.05) 1958), in the Regional Soil Testing Laboratory with respect to the ecological regions (Table 1). of Pokhara. The aggregates of all the sources This signifies that same sorts of rules prevailing of carbon pools were added to obtain the total in the CFs that the users extract forest biomass carbon stock. The carbon content was assumed one or other ways for fulfilling their daily needs. to be 47% of dry biomass (IPCC, 2006 cited

Table 1: Biomass carbon stock densities in the study areas Carbon pools df Hill CF Terai CF Mean P-value Remarks Tree carbon density (t ha-1) 114.004 115.22 359.69 237.455 0.000031 Sapling carbon density (t ha-1) 36.298 0.28 0.39 0.335 0.002886 Herbs, grass and leaf litter(t ha-1) 14.977 7.51 23.92 15.715 0.000016 Dead and stump carbon(t ha-1) 32.151 0.14 0.20 0.170 0.135100 Total of biomass carbon density 123.15 384.20 253.675 Soil organic carbon (t ha-1) 5.678 111.39 95.09 103.240 0.558200 Total carbon density 234.54 479.29 356.915 Remarks: * = significant atP <0.05 Carbon on branches, leaves, and roots is included in their respective pools

27 Banko Janakari, Vol. 26, No. 1 Pandey and Bhusal Soil organic carbon density (Table 1). The total organic carbon in the Shorea robusta dominated forest was found to be The soil organic carbon (SOC) of Shorea robusta 234.54 t ha-1 in the Hill S. robusta forest while it dominant CF was found to be higher in the Hill S. was 479.29 t ha-1 in the Terai S. robusta forest. A -1 robusta forest (111.4 t ha ) as compared to the one study conducted in the Kusumdanda Community -1 in the Terai S. robusta forest (95.1 t ha ) (Table Forest in which is also situated in 1). The SOC contained in the topmost layer of the Hill ecological region of Nepal, showed the soil (0–20 cm depth) in the Hill forest was found carbon stock density in a Shorea robusta forest to -1 -1 to be highest (28.45 t ha ) and lowest (16.9 t ha ) be 186.95 t ha-1 (Nepal, 2006) and later, 235.95 within the horizon of 60–80 cm depth (Fig. 3). t ha-1 Shrestha (2008), which are almost similar Similarly, the SOC of the topmost layer of soil to the results of this study conducted in the same in the Terai forest was found to be highest (36.6 t ecological region. Another study conducted by -1 -1 ha ) and lowest (9.4 t ha ) within the end horizon ANSAB in the same CF (Ghaledanda Ranakhola of 80–100 cm. Even though the cumulative value CF in Gorkha) in 2011 estimated 211.184 t ha-1 of the SOC is quite different in these community of carbon stock whereas the SOC was calculated forests, there was no significant difference only up to 30cm depths (ANSAB, 2011). The P>0.05) in storing organic carbon per unit area total organic carbon in the Terai (479.29 t ha- with respect to the ecological region (Table 1). 1) was found to be far greater than the average The overall trend of the SOC showed the gradual carbon stock of the tropical forests (285.0 t decrease in carbon density with the increase in the ha-1) of the world (Jina et al., 2008) although soil depth in the Hill forest but there was no such lower than the mean carbon stock estimation significant trend in the Terai forest. However, the of the CF in Nepal (Acharya et al., 2009). This SOC density was found to have decreased from might be the reason of the availability of greater top to bottom with the increase in the soil depth average diameter and height classes of the trees in both the ecological regions (Fig. 3). in the S. robusta dominated forest in the Terai as compared to the one in the Hills. The carbon stock density estimated in this study was higher than the average carbon stock in Nepalese forest (161.1 t ha-1) estimated by the Global Forest Resource Assessment Report of FAO (2006). The reason behind might be because of the use of the projected Forest Inventory data of 1994 by the FAO in its report.

However, the SOC in the Terai S. robusta forest was found to be lesser than that in the Hill S. Fig. 3: Ecological region-wise SOC density at robusta forest. Besides, the SOC content was also different soil depths found to be low as compared to the biomass carbon content in the Terai forest whereas it was almost Although the trend line was found to have equal in the forest situated in the Hill region. In decreased drastically within the lowest horizon the case of SOC, it is highly dependent on bulk (80–100 cm depth), it was quite fluctuating within density, content of organic matters, type of soil the middle horizons (40–60 cm and 60–80 cm and parent material since bulk density depends on depths) in the Terai forest. This clearly indicates several factors such as compaction, consolidation the seepage of SOC from the water holes in the S. and amount of carbon present in soil (Morisada, robusta forest of the Terai ecological region. The et al., 2004 and Leifeld et al., 2004 cited in main reason of difference in the SOC density is Ranabhat et al., 2008). Also, this clearly indicates because of the difference in their bulk densities the higher collection of leaf litter and plant and the alluvial deposition in the soils of the Terai residues by the community forest users from the forest. Terai S. robusta forest, or the alluvial deposition Total carbon stock density in the soils of the Terai S. robusta forest, because of which the organic matter content in the soils is The combined carbon stock density is the sum of significantly decreased. biomass carbon stock density and the SOC stock

28 Pandey and Bhusal Banko Janakari, Vol. 26, No. 1 In the case of the Hill S. robusta forest, both the found to be significantly higher in the Terai CF above ground biomass and the SOC were found (479.3 t ha-1) than in the Hill CF (234.5 t ha-1). to possess 48% of carbon and the below-ground These results reveal that the communities have biomass (4%) whereas in the case of the Terai S. been conserving remarkable amount of carbon in robusta forest, more than two third of the carbon their forests through proper forest management. was found to be accumulated in the above- However, replications of similar types of research ground biomass (72%) followed by SOC (21%) on both the ecological regions are crucial to get and the least (7%) in the below-ground biomass valid inferences for the generalization of the (Fig. 4). The above-ground biomass carbon in result throughout the nation. the Terai forest was almost 1.5 times more than that in the Hill forest whereas the SOC content Acknowledgements was in reverse order as the Terai S. robusta forest The authors are thankful to Eng. Kausal Raj contained less than half proportion of SOC than Gnyawali for his GIS work and the anonymous the Hill S. robusta forest. reviewer for his/her constructive feedbacks in the preliminary version of the manuscript. References

Acharya, K. P., Dangi, R. B., Tripathi, D. M., Bushley, B. R., Bhandary, R. R. and Bhattarai, B. 2009. Ready for REDD ? Taking Stock of Experience, Opportunities and Challenges in Nepal. Nepal Foresters’ Association, Kathmandu, Nepal. Adhikari, S. 2016. Growing money from carbon. Kathmandu. http://communityredd.net cited on 16 February, 2016. Amatya, S. M. and Shrestha, K. R., 2010. Nepal Forestry Handbook. Nepal Foresters’ Association, Kathmandu, Nepal. ANSAB. 2010. Forest Carbon Stock Measurement: Guidelines for measuring carbon stocks in community-managed forests. Asia Network for Sustainable Agriculture and Bio-resources (ANSAB), Fig. 4: Different carbon pools in the Hill and Federation of Community Forest User Groups Terai S. robusta forests and International Centre for Integrated Conclusion Mountain Development, Kathmandu, Nepal. ANSAB. 2011. Forest Carbon Stock of The total biomass carbon density was found to Community Forests in Three Watersheds be significantly higher in the Terai S. robusta (Ludikhola, Kayarkhola and Charnawati). -1 dominated CF (384.2 t ha ) than in the Hill S. Asia Network for Sustainable Agriculture -1 robusta dominated CF (123.2 t ha ). However, and Bio-resources (ANSAB), Federation -1 the SOC in the Hill CF (111.4 t ha ) was found of Community Forest User Groups and to be more than the one in the Terai CF (95.1 t International Centre for Integrated Mountain -1 ha ). It was found that the SOC decreased with Development, Kathmandu, Nepal. the increase in soil-depth in both the ecological regions. The SOC contributed 48% of the total Awasthi, K. D., Singh, B. R. and Sitaula, B. K. carbon pool in the Hill S. robusta forest whereas 2005. Profile carbon and nutrient levels and it accounted for 21% in the case of the Terai CF. management effect on soil quality indicators Similarly, the total carbon stock density was in the Mardi watershed of Nepal. Acta

29 Banko Janakari, Vol. 26, No. 1 Pandey and Bhusal

Agriculture Scandinavia Section B-Soil and Report. Food and Agriculture Organization Plant 3 (55): 192–204. of the United Nations (FAO), Rome, Italy. Bajracharya, R. M., Lal, R. and Kimble, J. M., FAO. 2011. FAO Glossary of Terms. Food 1998. Soil organic carbon distribution in and Agriculture Organization of the United aggregates and primary particle fractions as Nations (FAO), Rome, Italy. http://www.fao. influenced by erosion phases and landscape org/faoterm accessed on 2 February, 2016. position. In Soil Processes and the Carbon Fontaine, S., Barot, S., Barre, P., Bdioui, N., Mary, Cycle (eds) Lal R., Kimble J., Follett. R B. and Rumpel, C. 2007. Stability of organic and Stewart B.A. CRC Press, Boca Raton, carbon in deep soil layers controlled by fresh Florida, USA, 353–367. carbon supply. Nature 450: 277–280. Banskota, K., Karky, B. S. and Skutsch, M. GoN. 2011. Forest Carbon Measurement 2007. Reducing Carbon Emissions Directives. (In Nepali version). The through community-managed Forests in REDD Forestry and Climate Change Cell, the Himalayas. International Centre for Government of Nepal (GoN), Ministry of Integrated Mountain Development, Nepal. Forests and Soil Conservation, Kathmandu, http://www.bookicimod.org accessed on 17 Nepal. February, 2016. IPCC. 2000. IPCC Special Report: Land Use, Bird, S. B., Herrick, J. E., Wander, M. M. and Land-Use Change, and Forestry (Summary Wright, S. F. 2002. Spatial heterogeneity of for Policymakers). A Special Report of the aggregate stability and soil carbon in semi- Intergovernmental Panel on Climate Change. arid range land. Environmental Pollution 11 (IPCC), Geneva, Switzerland. (116): 445–455. IPCC. 2007. Climate Change 2007: The Brown, K. and Pearce, D. 1994. The economic Physical Science Basis – Summary for value of non-timber benefits of tropical Policymakers. Intergovernmental Panel on forests: carbon storage. In The Economics of Climate Change, Geneva, Switzerland. Project Appraisal and the Environment; New Horizons in Environment Economics (ed) Jackson, M. L. 1958. Soil Chemical Analysis. Weiss, J., and Elgar E. Aldershot Publishing, Prentice hall, New York, USA. Chelttenham, UK, 102–123. Jackson, J. K. 1994. Manual of Afforestation Cannell, M. G. R., Milne, R., Hargreaves, K. in Nepal (Volume I). 2nd edition. Forest J., Brown, T. A. W., Cruickshank, M. M., Research and Survey Center, Kathmandu, Bradley, R. I., Spencer, T., Hope, D., Billett, Nepal. M. F., Adger, W. N. and Subak, S. 1999. National inventories of terrestrial carbon Jina, B. S., Sah, P., Bhatt, M. D. and Rawat, Y. S. sources and sinks: The UK experience. 2008. Estimating carbon sequestration rates Climatic Change 26 (42): 505–530. and total carbon stockpile in degraded and non-degraded sites of Oak and Pine forest CBS. 2006. Environmental Statistics of of Kumaun Central Himalaya. Ecological Nepal. Central Bureau of Statistics (CBS), Society, Nepal. Ecoprint 15: 75–81. Kathmandu, Nepal. Lal, R., Kimble, J. M. and Follett, R. 1997. DOF. 2015. Status of Community Forest Users’ Land use and soil carbon pools in terrestrial Groups. Department of Forests (DoF), ecosystems. In Management of Carbon Kathmandu, Nepal. http://dof.gov.np/dof_ Sequestration in Soils (eds) Lal, R., Kimble, community_forest_division/community_ J. M., Follett, R. CRC Press, New York, USA. forestry_dof accessed on 2 February, 2016. Lal, R. 2004. Soil carbon sequestration to mitigate DHM. 2015. Meteorological Data of Nepal. climate change. Geoderma 123 (1–2): 1–22. Department of Hydrology and Meteorology (DHM), Kathmandu, Nepal. Nepal, S. 2006. A comparative study on carbon sequestration from two forest types in FAO. 2006. Global Forest Resource Assessment community forestry system (A case study

30 Pandey and Bhusal Banko Janakari, Vol. 26, No. 1

from coniferous and broad-leaved forests Agroecosystems 81: 179–191. in Palpa District). B.Sc. Thesis, Tribhuvan Shrestha, B. P. 2008. An Analytical Study of University, Institute of Forestry, Nepal. Carbon Sequestration in Three Different Pearson, T. R., Brown, S. and Ravindranath, Forest Types of Mid-Hills of Nepal. M.Sc. N. H. 2005. Integrating Carbon Benefit Thesis, Tribhuvan University, Institute of Estimates into GEF projects: Capacity Forestry, Pokhara, Nepal. Development and Adaptation Group UNFCCC. 2015. Paris COP 21: Information Guidelines. Global Environment Facility, Hub. http://unfccc.int/focus/climate_finance/ United Nations Development Program, New items/7001.php accessed on 16 February, York, USA. 2016. Ranabhat, S., Awasthi, K. D. and Malla, R. 2008. Upadhyay, T. P., Sankhayan, P. L. and Solberg, Carbon sequestration potential of Alnus B. 2005. A Review of carbon sequestration nepalensis in the Mid-Hills of Nepal: a case dynamics in the Himalayan region as a study from . Banko Janakari 18 function of land use change and forest/soil (2): 3–9. degradation with special reference to Nepal. RDCT. 2012. R Statistical Packages. R Agriculture, Ecosystems and Environment Development Core Team (RDCT). http://r- 105: 449–465. project.org accessed on 2 January, 2012. UN-REDD. 2014. Understanding Drivers REDD-IC. 2015. REDD+ Strategy for Nepal and Causes of Deforestation and Forest (Draft). REDD Implementation Centre Degradation in Nepal: Potential Policies and (REDD-IC), Ministry of Forests and Soil Measures for REDD+. Discussion paper. Conservation, Kathmandu, Nepal. http://www.tinyurl.com/nepal-drivers-redd accessed on 23 March, 2014. Shrestha, B. M. and Singh, B. R. 2008. Soil and vegetation carbon pools in a mountain watershed of Nepal. Nutrient Cycling in

31 Resource assessment of Bel (Aegle marmelos) and potentiality to establish its processing enterprise in of Nepal

K. Baral1 and B. R. Upreti2

Participatory resource assessment of Aegle marmelos was carried out in six community forests (CFs) of the Jamune VDC of Tanahun district to find out potentiality for establishing community-based A. marmelos processing enterprise in the locality. Circular sample plots of 500 square meters were laid down taking 5% sampling intensity. The DBH (Diameter at Breast Height) and height of trees on the sample plots were measured and classified them into three DBH classes (10–20 cm, 20–30 cm and more than 30 cm). The number of the fruits on the three branches of each A. marmelos tree – one lower branch, one middle branch and one top branch were counted, the average number of fruits on the three branches calculated and multiplied with the number of the branches of the tree to find out the average number of fruits per tree. By calculating the weighted mean of the three different DBH-class trees, the total number of the fruits was estimated. The study found 54,830 kg of harvestable amount of A. marmelos fruit in the studied six CFs which can produce 19190.5 kg of pulp per year. Thus, establishment of juice processing enterprise was found to be feasible in the locality. Nevertheless, some shortcomings related to the management of A. marmelos resource, such as lacking of information and management options for A. marmelos trees in the Operation Plans (OPs), lower regeneration status of A. marmelos trees, higher incidences of forest-fire and open grazing in the CFs, were also recorded. The study suggests for carrying out awareness-generating activities targeted to the CF user group members, revision of OPs and incorporation of A. marmelos resource information and management options and preparation and implementation of regeneration protection plan against forest fire and grazing.

o Aegle marmelos, enterprise, regeneration protection

egle marmelos (L.) Correa is a medium- altitude (Shrestha and Shrestha, 2005). It prefers Asized deciduous tree. Its local names are Bel comparatively drier and sunny or warmer aspect (Nepali), Vilva, Biranab (Sanskrit) Bengal queen with well-drained loamy soil. It is found growing (English) etc. It belongs to the Rutacae family naturally in the mixed stands of Shorea robusta, and Aurantioideae sub-family. Its branches are Terminalia tomentosa, Adina cordifolia and so on thorny and the bark is gray in color. The leaves in the Tarai, Bhabar and the Mid-hills. It copes are trifoliate with numerous oil glands. Flowers with a wide range of soil conditions (pH range of A. marmelos are greenish white and bisexual 5–10), is tolerant to water-logging and has an in nature. Flowering starts during April-June, unusually wide temperature tolerance (from 7°C fruiting occurs from April to July of next season to 48°C). It requires a pronounced dry season to (Parajuli et al., 1998; Kunwar, 2006; Pathak et give fruit (Bhattrai, 2001; Poudel, 2005). al., 2015). A. marmelos trees are generally found in the outer Himalayas, Siwaliks and Tarai with A. marmelos has both medicinal and religious altitudes up to 1500 m. In Nepal, A. marmelos values. The ripe fruit is taken scientifically as is distributed abundantly in the Siwalik region, brain tonic and energetic, and it improves eternal Inner Tarai, and lower valley region mostly power and longevity. Moreover, it is considered at riverside having sandy soil at 150–1,220 m as a very good medicine for the patients suffering

1 District Forest Office, Baitadi, Nepal. E-mail: [email protected] 2 District Forest Office, Tanahun, Nepal

32 Baral and Upreti Banko Janakari, Vol. 26, No. 1 from constipation. Unripe A. marmelos fruit and the local people are energetic to establish A. is prescribed for curing cholera, diarrhea, marmelos processing enterprises to support the worms and other stomach diseases because livelihood of the members. In this regard, the of its antibacterial, antiviral and anthelmintic current study was carried out in 2014 with the properties. The leaves are excellent medicine objectives of carrying out resource inventory for the diabetic patients (Shrestha, 2003). of the A. marmelos trees, finding out the total Furthermore, A. marmelos fruit contains high- number and annual fruiting trees, assessing the valued chemical compounds including alkaloids, total and annual harvestable weight of fruits and coumarone, steroids, mucilage pectin, sugar, and finding out the potential to establish A. marmelos tannin whilst the leaves produce an essential oil. processing factory at Jamune VDC of Tanahun The fruit has high nutritional value containing district by estimating pulp production potential of tonics, vitamins, carbohydrates, proteins, fats and the six CFs. a range of medicinal substances (Bhattrai, 2001). Materials and methods Juice, jam, candy, sweets and other food products Study site can be made from the fruit in addition to a range of Ayurvedic medicines. No substance poisonous Tanahun is a hilly district with altitude ranging to humans has been found in A. marmelos fruit from 187–2,325 m above the mean sea level. (Poudel, 2005). Juice of the ripen A. marmelos This district is located between 27o36’ - 28o05’ fruit has very good market in Nepal and India. North latitude and between 83o57’ and 84o34’ E A. marmelos is identified as a major high- valued longitude. The area of the district is 1,560 square non-timber forest product (NTFP) which has kilometers out of which the forest area occupies great potentiality to improve the livelihood of 50.5% (788 sq. km). Altogether, 545 CFs have the Community Forest User Groups (CFUGs) been handed over to 52,989 households covering members in Nepal. Considering this fact, a few 37689.71 hectares (ha) of forest land in the district A. marmelos juice factories have been established (DFO, 2014). in different parts of the nation in the initiation of District Forest Offices (DFOs), the Federation of Jamune VDC lies almost at the centre of the Community Forestry Users Nepal (FECOFUN), district. Till now, 12 CFs have been handed over Community Forests (CFs) and other organizations. to the 1951 households covering 1,105.21 ha of The Tamakoshi Bel Juice Processing Company of forest land in this VDC (DFO, 2014). Among Eastern Nepal and the Nabadurga Community them, the six CFs selected for the study purpose Bel Juice Factory of in Mid- (Table 1) were reported to have higher number western Nepal are some successful examples of of A. marmelos trees, and thus have ample A. marmelos enterprises run by the CFUGs (Baral potentialities to run the A. marmelos processing and Khadka, 2007). enterprise. These CFs lie in a cluster which is approximately 15 km far from Damauli, the The CFs of the Jamune Village Development district headquarters (Fig. 1). Committee (VDC) of Tanahun district is reported to have substantial number of A. marmelos r ,

Table 1: Brief descriptions of the six studied CFs Total area Number of S.N. Name and address of CFs Population Reference (ha) households 1. Poseli CF, Jamune-1 48.75 93 489 CFOP 2008a 2. Barchyang CF, Jamune-2 160.00 93 596 CFOP 2008b 3. Jantang Pandhera CF, Jamune-3 104.90 134 815 CFOP 2008c 4. Bhirpani CF, Jamune-4 124.00 156 930 CFOP 2008d 5. Uma Chwok CF, Jamune-5 94.50 181 996 CFOP 2008e 6. Siddha Batasan CF, Jamune-6 115.7 90 495 CFOP 2008f Total 647.85 747 4,321

33 Banko Janakari, Vol. 26, No. 1 Baral and Upreti

Plot measurement I. A. marmelos-dominant blocks

First of all, 5% sample out of the whole area of the blocks was taken, and a total of 117 circular sample plots of 500 square meters were laid. Transect lines were drawn on the map of blocks and first sample plot was laid down on the starting point of transect line. Remaining sample plots were taken on the transect line and plot to plot distance was fixed by following the NTFP Inventory Guidelines published by Department of Forest (2013). Then, the diameters (at breast Fig. 1: Map showing the location of the height) and heights of all the A. marmelos r study area were measured and recorded; the A. marmelos trees were classified into three diameter (DBH) The forests in all the studied CFs are, moreover, classes viz. 10–20 cm, 20–30 cm and more than sub-tropical mixed-hardwoods with S. robusta, 30 cm. After that, the number of the fruits, on the T. tomentosa, A. cordifolia, Acacia catechu, three branches (of each A. marmelos tree)- one Bombax ceiba, Schima wallichii and Castanopsis lower branch, one middle branch and one top indica as the major tree species. Apart from A. branch, were harvested, counted and weighed marmelos, other important NTFPs found in the separately as per the above mentioned DBH CFs are Asparagus racemosus, T. bellirica and classes. Then the pulp content of all the harvested T. chebula. The major wildlife species found in fruits were weighed and the average number of the CFs are leopard, jackal, fox, monkey, rabbit, fruits and their pulp content per tree within each squirrel, civet and jungle cat. of the aforementioned three DBH classes were Sampling design calculated. Finally, the total number of the fruits and pulp content of the blocks were found out. A participatory resource inventory was carried out involving the members of the CFUGs and II. Non-A.marmelos-dominant blocks the Tanahun DFO field staff. The blocks of the studied CFs were stratified into the A. marmelos- In the case of the non-A. marmelos-dominant dominant blocks and non- A. marmelos-dominant blocks, all the A. marmelos trees were first counted blocks; all the blocks having more than 50% and classified into the aforementioned three DBH A. marmelos trees out of the total trees were classes. The classified numbers of the trees were categorized as A. marmelos-dominant while then multiplied with their corresponding average the blocks having less than 50% A. marmelos figures per DBH class calculated earlier in the trees were categorized as non- A. marmelos case of the A. marmelos-dominant blocks so as to -dominant blocks. In course of the stratification find out the total amount of fruits and their pulp. of the forests into such blocks, full observation of the blocks and consultation with the concerned Results and discussion CFUG members were done rigorously. In the Assessment of the fruiting A. marmelos trees A. marmelos-dominant blocks, inventory was carried out following the rules mentioned in the Altogether, 7,832 A. marmelos trees were found NTFPs Inventory Guidelines, 2013 (DoF, 2013). in the six CFs. Out of this number, only 5,483 In the case of the non-dominant blocks, all the A. trees (70%) were found to be fruiting, and the marmelos trees counted and their measurements rest 2,349 (30%) were non-fruiting (Fig. 2). This were taken. result is similar to those of Baral and Khadka (2007) who found 24% non-fruiting A. marmelos trees in the CFs of Bardiya district. Most of the trees located under the Sal canopy and steep sloppy area were found to be non-fruiting.

34 Baral and Upreti Banko Janakari, Vol. 26, No. 1

Availability of A. marmelos fruits and pulp

The weighted mean of A. marmelos fruits in different DBH class was found 150. Hence, a total number of 822,450 A. marmelos fruits are estimated in the six studied CFs which are equal to the 137,075 kg at the average rate of 6 fruits per kg weight. Likewise, average production of fruits was calculated 212 kg per hectares (Table 3).

Fig. 2: Presence of fruiting and non-fruiting Harvesting amount of A. marmelos fruits and trees in the studied CFs production of pulp

Out of the total area of 647.85 ha of all the six The total estimated numbers of A. marmelos CFs, only 117.0 ha (18%) had the A. marmelos fruits available in the six CFs was 822,450. trees both in the A. marmelos-dominant and Nevertheless, due to the difficult terrain and non-A. marmelos-dominant blocks (Table 2). In inaccessibility, large quantity of fruits could not the case of the dominant blocks, the entire blocks be harvested. In the fruiting season, monkeys were incorporated as effective areas; however in and bears used to damage the fruits. On the the case of the non-A. marmelos-dominant blocks, other hand, some fruits should be left in the trees consultation with the concerned CFUG members for full ripening and dispersal so as to promote and field observation were done so as to estimate natural regeneration. The CFUG members were their areas. In this regard, the Siddha Batasan CF found to have long experience on A. marmelos had highest number of A. marmelos trees (1,738) fruit harvesting. Based on the rigorous discussion whereas the Poseli CF had the least (778 trees). with the CFUG members who have been actively On an average, 8 fruiting A. marmelos r r participating in its harvesting, we concluded that found per ha in the CFs. The maximum number only 40% of the total fruits could be harvested. of fruits found in a A. marmelos tree was recorded Hence, the total amount of harvestable A. as 450 and the minimum as 50. The number of marmelos fruits was estimated to be 328,980 seedlings and saplings together was found to be (54,830 kg). Different literatures (Kunwar, 2006; almost half the number of mature trees, and only Baral and Khadka, 2007) show that the pure pulp 33 seedlings and saplings together per hectare is only 35% (on an average) of the total weight of were observed (Table 2). This indicated the very the A. marmelos fruits while the remaining 65% poor regeneration status of A. marmelos trees in belong to the bark, fibers, seeds and the wastage. the studied CFs on the one hand and higher threat Thus, pulp production potential of the six CFs to the conservation as well as sustainability of the was found to be 19,190.5 kg per year (Table 3). species on other hand in this locality.

Table 2 : Description of A. marmelos trees in the studied CFs No. of A. No. of No. of No. of Total area Effective S.N. Name and address of CFs marmelos fruiting fruiting seedlings + (ha) area (ha) trees trees trees /ha saplings 1. Poseli CF, Jamune-1 48.75 9.0 778 545 11 910 2. Barchyang CF Jamune-2 160.00 32.0 1,340 938 6 520 3. Jyantang Pandhera CF Jamune-3 104.90 18.0 1,330 931 9 610 4. Bhirpani CF Jamune-4 124.00 20.0 1,423 996 8 615 5. UmaChock CF Jamune-5 94.50 15.0 1,223 856 9 460 6 Siddha Batasan CF Jamune-6 115.70 23.0 1,738 1217 11 850 Total 647.85 117.0 7,832 5,483 8 3,965

35 Banko Janakari, Vol. 26, No. 1 Baral and Upreti

Table 3: Total and harvestable amount of A. marmelos fruits and pulp Harvestable amount of A. marmelos Total amount of A. marmelos fruits Name and address fruits S.N. of CFs Total no. Weight of Production No. available Weight of Pulp content of fruits fruits (kg) (Kg/ha) for harvesting fruits (kg) (kg) 1. Poseli CF, 81,750 13,625 279 32,700 5,450 1,907.5 Jamune-1 2. Barchyang CF 140,700 23,450 147 56,280 9,380 3283 Jamune-2 3. Jyantang Pandhera 139,650 23,275 222 55,860 9,310 3,258.5 CF Jamune-3 4. Bhirpani CF 149,400 24,900 201 59,760 9,960 3486 Jamune-4 5. Uma Chock CF 128,400 21,400 226 51,360 8,560 2996 Jamune-5 6. Siddha Batasan CF 182,550 30,425 263 73,020 12,170 4,259.5 Jamune-6 Total 822,450 137,075 212 328,980 54,830 19,190.5

Information deficiency in the CFOPs marmelos; some literatures (Baral and Khadka, 2007; Poudel, 2005) also support this fact. The OPs of the studied CFs were found to have resource assessment summary of different tree Perception of different stakeholders species. However, most of the OPs were found The perception of the major stakeholders was to have information gap on A. marmelos and found to be extremely positive towards the other NTFPs although the CFs were reported to conservation and sustainable use of A. marmelos possess a plenty of these resources. The reason resource to support the livelihood of the CFUGs. behind was that A. marmelos and other NTFPs In this regard, the Tanahun DFO together with were ignored and higher priorities were given the FECOFUN/CFUGs and all the other local to Acacia catechu, Shorea robusta, Terminalia organizations were found to be very much tomentosa and other timber-yielding species committed to playing active role in conservation during field inventories. and sustainable use of A. marmelos resource in Conservation threats the district. Poor regeneration due to the higher incidences of Conclusion forest fire and grazing were observed in the CFs, which were found to be serious threats for the The CFs of the Jamune VDC were found to be conservation of this species. Besides, the thorny rich in the A. marmelos resource; however, there nature of this species creates difficulty in carrying was no proper utilization of the resource. The six out its tending and harvesting operations; so, studied CFs were found to have possessed 54,830 some CFUGs were found to have even tendency kg annual harvestable amount of A. marmelos to remove A. marmelos trees from their CFs. fruits with the potential of 19,190.5 kg of pulp production per year. With this annual production Availability of A. marmelos trees in the amount, the CFUGs can establish a small-scale adjoining CFs and VDCs A. marmelos processing enterprise in the locality. The study was concentrated in the CFs of Jamune Moreover, other CFs in the adjoining VDCs VDC. However from the discussion with the were also reported to have possessed massive CFUGs, field staffs of DFO and field observation, numbers of A. marmelos trees. Thus, a higher we came to know that plenty of A. marmelos r potentiality was observed for the development were also found in the CFs of the adjoining VDCs A. marmelos processing enterprise in the locality such as , , etc. of so as to support the livelihood of the CFUG the district. According to the CFUG members, members in the Jamune VDC of Tanahun district. every alternate year is a good seed-year for A. However, further resource assessment of the

36 Baral and Upreti Banko Janakari, Vol. 26, No. 1 adjoining CFs is necessary to find out the actual CFOP. 2008e. Community Forestry pulp production potential so as to sustain the Operational Plan (CFOP), Uma Chowk proposed A. marmelos processing enterprise. Community Forest, Jamune VDC-5, On the other hand, higher threat to conservation Tanahun, Nepal. of this species has been noticed due to the poor regeneration status of this species caused by CFOP. 2008f. Community Forestry forest fire and open grazing in the studied six Operational Plan (CFOP), Siddha CFs. Therefore, awareness generating activities Batashan Community Forest, Jamune among the concerned CFUGs should be carried VDC-6, Tanahun, Nepal. out as soon as possible. Besides, preparation and DFO. 2014. Annual Report (2014). District implementation of regeneration protection plan Forest Office (DFO), Tanahun, Nepal. with forest fire and grazing control activities are extremely necessary for management and DoF. 2013. Non-Timber Forest Products sustainable use of A. marmelos resource in the Inventory Guidelines (2013). Department Jamune VDC of Tanahun district. of Forest (DoF), Kathmandu, Nepal. Acknowledgments Kunwar, R. M. 2006. Non-Timber Forest Products of Nepal: A Sustainable We are grateful to Mr. D. R. Gautam together with Management Approach. Centre for the CARE Nepal, Hariyo Ban Program for their Biological Conservation Nepal and financial support to conduct this study. Similarly, International Tropical Timber Organization, we are thankful to Mr. R. B. Poudyal (District Japan. Forest Officer of Tanahun, DFO), Mr. G. Pandey, Assistant Forest Officer of Tanahun DFO, B. K. Parajuli, D., Gyanwali, A. R. and Shrestha, B. M. Shreshta, Chairman of the Siddha Batasan CF and 1998. Manual of Important Non-Timber the members of the CFUGs for their kind support Forest Products of Nepal. International during the study period. Tropial Timber Organization/Institute of Forestry, Pokhara, Nepal, 6. References Pathak, L. N., K. C., R. and Chaudhary, C. Baral, K. and Khadka, D. 2007. Resource L. 2015. Cultivation Practice of Major Inventory of Bel (Aegle marmelos) in Tropical Non-Timber Forest Products of Bardiya District. Council for Commerce and Nepal. Food and Agriculture Organization Industry (CCI), Bardiya, Nepal. of the United Nations, Kathmandu, Nepal, 121–127. Bhattrai, D. R. 2001. Jadibuti Manjari. (Text in Nepali). Shubhash Printing Press, Poudel, D. 2005. Including the Excluded: A Pro- Kathmandu, Nepal. Poor Bel Fruit Juice Making Enterprise in Nepal. International Tropical Timber CFOP. 2008a. Community Forestry Operational Organization (ITTO), Community Forestry Plan (CFOP), Poseli Community Forest, Training Center for Asia and the Pacific Jamune VDC-1, Tanahun, Nepal. (RECOFTC), Forest Trends, RRI, Thailand, CFOP. 2008b. Community Forestry Operational 10–19. Plan (CFOP), Barchyang Community Shrestha, R. 2003. Importance of Aegle marmelos Forest, Jamune VDC-2, Tanahun, Nepal. in Nepal. Nepal Foresters’ Association, CFOP. 2008c. Community Forestry Kathmandu. The Nepal Journal of Forestry Operational Plan (CFOP), Jantang XII (2): 68–69. Pandhera Community Forest, Jamune Shrestha, U. B. and Shrestha, S. 2005. Major VDC-3, Tanahun, Nepal. Non-Timber Forest Products of Nepal. CFOP. 2008d. Community Forestry (Text in Nepali). Bhundi Puran Publication, Operational Plan (CFOP), Bhirpani Kathmandu, Nepal. Community Forest, Jamune VDC-4, Tanahun, Nepal.

37 Forest type mapping using object-based classification method in , Nepal

A. K. Chaudhary1*, A. K. Acharya1 and S. Khanal1 In the recent years, object-based image analysis (OBIA) approach has emerged with an attempt to overcome limitations inherited in conventional pixel-based approaches. OBIA was performed using Landsat 8 image to map the forest types in Kapilvastu district of Nepal. Systematic sampling design was adopted to establish sample points in the field, and 70% samples were used for classification and 30% samples for accuracy assessment. Landsat image was pre-processed, and the slope and aspect derived from the ASTER DEM were used as additional predictors for classification. Segmentation was done using eCognition v8.0 with the scale parameter of 20, ratios of 0.1 and 0.9 for shape and color, respectively. Classification and Regression Tree (CART) and nearest neighbor classifier (k-NN) methods were used for object-based classification. The major forest types observed in the district were KS (Acacia catechu/ Dalbergia sissoo), Sal (Shorea robusta) and Tropical Mixed Hardwood. The k-NN classification technique showed higher overall accuracy than the CART method. The classification approach used in this study can also be applied to classify forest types in other districts. Improvement in classification accuracy can be potentially obtained through inclusion of sufficient samples from all classes.

o Landsat, machine learning algorithm, object-based classification

emote sensing provides a useful source of through automatic image classification technique. Rdata from which land-cover information In the recent years, Ikonos imagery has been can be extracted for assessing and monitoring frequently used for vegetation mapping using vegetation changes. In the past several decades, pixel-based image classification methods (Wang air-photo interpretation has played an important et al., 2004a; Wulder et al., 2004; Metzler and role in detailed vegetation mapping (Sandmann Sader, 2005; Souza and Roberts, 2005). Pixel- and Lertzman, 2003), while applications of based method, however, has constraints with medium spatial resolution satellite imagery VHR image classification because of decrease such as Landsat Thematic Mapper (TM) and in classification accuracy due to high-spectral SPOT high-resolution visible (HRV) alone variability within classes (Yu et al., 2006; Lu and have often proven insufficient or inadequate for Weng, 2007). It also ignores the context and the differentiating species-level vegetation in detailed spectral values of adjacent pixels (Townshend et vegetation studies (Harvey and Hill, 2001). al., 2000; Brandtberg and Warner, 2006). Various image classification techniques have been Recently, object-based image analysis (OBIA) developed such as object-based, textural, and approach has been widely utilized for remote contextual image classifications in order to reduce sensing studies as an alternative and also the limitations associated with VHR images (Guo comparatively better classification approach to et al., 2007; Lu and Weng, 2007). the conventional pixel-based image classification techniques. Successful launch of very high- The geographic object-based image analysis resolution (VHR) commercial imaging satellites (GEOBIA) technique emerged since the late in the late 1990s are helpful for resource 1990s, to overcome human interpreters’ ability to inventory and monitoring (Ehlers et al., 2003; identify and delineate features of interest (Benz Ehlers, 2004). The VHR imagery is anticipated to et al., 2004; Meinel and Neubert, 2004). The be an alternative option to aerial photographs for GEOBIA technique could be useful to solve the characterization of forest structure and dynamics problems of high-spectral variability within the

1 Department of Forest Research and Survey, Kathmandu, Nepal E-mail: [email protected]

38 Chaudhary et al. Banko Janakari, Vol. 26, No. 1 same land-cover classes in VHR imagery (Yu et al., 2006; Lu and Weng, 2007). To overcome the high-resolution problem and salt-and-pepper effect, it is useful to analyze groups of contiguous pixels as objects instead of using the conventional pixel-based classification unit. This will reduce the local spectral variation caused by crown textures, gaps, and shadows. In addition, with spectrally homogeneous segments of images, both spectral values and spatial properties, such as size and shape, can be explicitly utilized as features for further classification. The basic idea of this process is to group the spatially adjacent pixels into spectrally homogenous objects first, and then conduct classification on objects as the minimum processing units.

The object-based classification procedure includes image segmentation, training sample selection, Fig. 1: Map of study area classification feature selection, tuning parameter setting and, finally, algorithm execution. The Preliminary work accuracy of image classification is influenced by The field crew members were trained on the segmentation quality (Dorren et al., 2003; Meinel collection of Global Positioning System (GPS) and Neubert, 2004; Addink et al., 2007). Dorren location of sample points, basal area calculation, et al. (2003) stated the importance of image crown cover measurement and forest type object-size in forest classification and mapping. signature collection in the field through various There are no specific guidelines to take optimal trainings. Consultation was done with District segmentation size and it is a matter of trial-and- Forest Office (DFO), Sector Forest Offices and error methods which influence segmentation Ilaka staffs for delineation of major forest types. quality (Definiens, 2004; Meinel and Neubert, 2004). Kim et al. (2008) emphasized spatial Data autocorrelation analysis to determine optimal segmentation size for forest stands. In the recent Multi-spectral satellite imagery of Landsat 8 was years, pixel-based classification with texture obtained from United States Geological Survey information has been employed to improve the (USGS). The characteristics of the image are accuracy of forest/vegetation mapping (Ferro presented in Table 1. The ASTER DEM of 30 and Warner, 2002). Therefore, this study was m resolution was obtained from USGS (2015) carried out to map forest types using object- and terrain parameters (slope and aspect) were based classification technique and recommend calculated which were used in Classification and the appropriate classification technique for other Regression Tree (CART) analysis. districts. Table 1: Characteristics of Landsat 8 Image Materials and methods Satellite Sensor Path-Row Date Band Study area Landsat 8 OLI and 13 Feb, 142–41 2–7 Kapilvastu district is situated in TIRS 2014 Landsat 8 OLI and 19 Jan, of Western Development Region of Nepal. 143–41 2–7 Geographically, it extends from 27o25’ N to TIRS 2014 27o84’ N latitude and from 82o75’ E to 83o14’ E longitude (Fig. 1). It spreads ranging from 93 to 1,491 m above sea level. The district enjoys Sampling design tropical and sub-tropical climate. Kapilvastu Systematic sampling design was employed to district covers 1,738.00 km2 land representing establish sample points in the field. A forest mask with forest cover area of 63,438.42 ha.

39 Banko Janakari, Vol. 26, No. 1 Chaudhary et al. for the study area was taken from the recent Image analysis and mapping Forest Resource Assessment (FRA) of the Terai (DFRS, 2014). Since, FRA forest cover was The Landsat 8 images acquired were pre- done on physiographic region scale, there were processed (layer stacking, image enhancement some minor discrepancies. Those were manually and mosaic king). Before image segmentation and edited using high resolution Google Earth classification, non-forest areas such as agriculture, Image. A systematic grid at the interval of 500 grassland, built up, river etc. were masked out m was generated within the district and all of the since the main concern of this proposed study was generated sampled points in regular grid (n= 213) to focus on forest types. Spectral bands (Band were visited in the field using GPS and dominant 2–Band 7) of Landsat 8 along with slope and species in the plot identified based on the species aspect derived from ASTER DEM were used in basal area. the segmentation. Segmentation was done using eCognition Version 8.0 with a scale parameter Point sampling of 20. The values of 0.1 and 0.9 were chosen for the ratios of shape and color, respectively. Horizontal point sampling was used to estimate Spectral signatures of individual forest types basal area for forest type mapping purpose. In this were extracted from the different bands of the sampling a series of sampling points were selected masked image by using training data and then systematically distributed over the entire area to be classification was performed by standard nearest inventoried. Trees around this point were viewed neighbor classifier (k-NN) and Classification and through any angle-gauge at breast height and all Regression Tree (CART) method which takes into trees forming an angle bigger than the critical consideration of spectral parameters and ancillary angle of instruments were counted. The basal area data (Definiens, 2004). The CART algorithm is per hectare was calculated by multiplying Basal one of the most commonly used decision trees Area Factor (BAF) of instrument with number of that works as a binary recursive partitioning tally trees to identify the forest types in the field. procedure by splitting the training sample set into The particular species representing basal area subsets based on an attribute value (set) and then greater than 60% corresponds to the same forest by repeating this process on each derived subset. type as the species. Based on the dominance of The tree-growing process stops when no further species basal area, the three major forest types splits are possible for subsets. The maximum namely Khair/Sissoo (KS-Acacia catechu/ depth of the tree is the key tuning parameter in the Dalbergia sissoo), Sal (S-Shorea robusta) and CART, determining the complexity of the model. Tropical Mixed Hardwood (TMH) (DFRS/FRA, In general, a larger depth can build a relatively 2014) were found in the district. The sample plot more complex tree with potentially higher overall distribution according to their categories is shown classification accuracy. Therefore, in this study in Figure 2. The total sample points (n=213) were the tree depth was set at 10. The k-NN algorithm divided into training data sets (70%) for forest uses an instance-based learning approach and type classification and 30% sample points for does classification by assigning class based on evaluating classification accuracy. the class attributes of its K-nearest neighbors. The CART technology is recently applied in ecology; it provides a low-cost, high quality alternative to approximate the human learning process and make accurate generalizations concerning the relationships of input variables and the value of the target feature, without such difficulties (Maniezzo et al., 1993).

Accuracy

The overall accuracy was calculated for summary measures (Gong et al., 1992), which can be used to compare individual class difference between distinct classifications (Coburn and Roberts, Fig. 2: Field sample points 2004).

40 Chaudhary et al. Banko Janakari, Vol. 26, No. 1 Results and discussion classification. The results from this study were, however, in contrast which might be due to the Forest types less number of field samples as well as limited number of predictor variables (forest types). The forests of Kapilvastu district were classified into three major forest types namely Khair/Sissoo The areas of different forest types were calculated (KS), Sal (S) and Tropical Mixed Hardwood using the CART and the k-NN classification (TMH). Forest types classification results based methods. The areas of KS (2,865.4 ha) and on CART and k-NN nearest neighborhood S (11,427.7 ha) calculated using the CART classification methods are presented in Fig. 3 and classification method were found to be higher as 4. compared to those (KS 1,944.9 ha and S 9,365.1 ha) obtained using the k-NN classification method. On the contrary, the area of the TMH (52,128.4 ha) computed using the k-NN classification was higher than the one (49,145.3 ha) worked out using CART method (Table 2).

Table 2: Area of three forest types calculated using CART and k-NN methods

Forest CART k-NN S.N. type Area (ha) Area (ha) 1. KS 2,865.42 1,944.90 Fig. 3: Forest type classification using CART 2. S 11,427.70 9,365.13 method 3. TMH 49,145.30 52,128.39

Accuracy assessment

The accuracy assessments of the two classification methods were accomplished to assess the qualities of the classified map products. The overall accuracy (69.7%) using the CART classification method was found to be slightly lower than the one (72.7%) obtained using the k-NN classification method whereas the user’s accuracies for Sal (20.0%) and TMH (84.3%) were recorded higher in the CART classification Fig. 4: Forest type classification using k-NN than those (S: 14.3% and TMH: 80.7%) recorded method in the k-NN classification (Table 3). On the contrary, the user’s accuracy for KS (50%) using Blaschke (2003) observed that the optimal the k-NN classification method stood higher as size of segmentation is critical and challenging compared to the one (20%) obtained using the task in GEOBIA. Therefore, as the optimum CART method. On the other hand, the producer’s segmentation size increases the classification accuracy for TMH (82.7%) based on the CART accuracy, it is likely that the classification results method was found to be lower than the one can be further improved by evaluating and (88.5%) based on the k-NN method whereas the selecting the optimal size. producer’s accuracy for ‘S’ based on the CART method was found to be exactly two times more Qian et al. (2015) evaluated and compared the (18.2%) than the one (9.1) based on the k-NN performance of four machine-learning classifiers method. Both the classification methods gave namely Support Vector Machine (SVM), Normal the same result of 33.3% producer’s accuracy for Bayes (NB), CART and k-NN using an object- KS. The classification accuracy as reported by based classification procedure, and found Czaplewski and Patterson (2003) was only 40% that CART method was superior to the k-NN

41 Banko Janakari, Vol. 26, No. 1 Chaudhary et al.

Table 3: Accuracy assessment of CART and k-NN methods Ground-truth field samples CART k-NN User's Error User's Error Classes KS S TMH Total accur. of com. KS S TMH Total accur. of com. (%) (%) (%) (%) KS 1 3 1 5 20 80 1 1 2 50 50 S 2 8 10 20 80 1 6 7 14.28 85.72 TMH 2 6 43 51 84.31 15.69 2 9 46 57 80.70 19.30 Total 3 11 52 66 3 11 52 66 Producer's accur. (%) 33.33 18.2 82.69 33.33 9.09 88.46 Error of 66.67 81.8 17.31 66.67 90.91 11.54 omiss. (%) Overall 69.69 72.72 accur. (%) or less for thematic information extraction at the ready information. ISPRS Journal of species-level based on the Landsat TM and SPOT Photogrammetry and Remote Sensing 58: HRV Images. The results of this study however 239–258. higher accuracy although there were only three classes. Blaschke, T. 2003. Object-based contextual image classification built on image segmentation, Conclusion Proceedings of the 2003 IEEE Workshop on Advances in Techniques for Analysis The proposed methods offer a reasonably of Remotely Sensed Data, 27–28 October, accurate forest type classification approach. Out Washington D.C., USA, 113–119. of the two classification algorithms, the k-NN classification technique showed higher overall Brandtberg, T. and Warner, T. 2006. High resolution accuracy than the CART method. The approach remote sensing. In Computer Applications in combining image segmentation and machine Sustainable Forest Management (eds.) G. learning method can be applied for mapping the Shao and K. M. Reynolds, Springer-Verlag, forest types in other Terai districts and potentially Dordrecht, Netherlands, 19–41. in other areas as well. More detailed classification can be potentially obtained through inclusion of Coburn, C. A. and Roberts, A. C. B. 2004. adequate number of samples in more classes and A multiscale texture analysis procedure also inclusion of smaller patches of forests by for improved forest stand classification. adjusting the sampling approach so that they are International Journal of Remote Sensing 25 included in the training and test samples. (2): 4287–4308.

References Czaplewski, R. L. and Patterson, P. L. 2003. Classification accuracy for stratification with Addink, E. A., de Jong, S. M. and Pebesma, E. remotely sensed data. Forest Science 49 (3): J. 2007. The importance of scale in object- 402–408. based mapping of vegetation parameters with Definiens, 2004. eCognition User Guide 4, hyperspectral imagery. Photogrammetric Definiens AG, Germany. Engineering and Remote Sensing 72 (8): 905–912. DFRS. 2015. State of Nepal’s Forests. Department of Forest Research and Survey Benz, U. C., Hofmann, P., Willhauck, G., (DFRS), Kathmandu, Nepal. Lingenfelder, I. and Heynen, M. 2004. Multi-resolution, object-oriented fuzzy DFRS/FRA. 2014. Standard Guidelines for analysis of remote sensing data for GIS- Forest Cover and Forest Types Mapping.

42 Chaudhary et al. Banko Janakari, Vol. 26, No. 1

Technical Report No. 3. Department of for Knowledge- driven Remote Sensing Forest Research and Survey, Forest Resource Applications (eds) Blaschke, T., Lang, S. and Assessment Nepal Project, Kathmandu, Hay, G. J. Springer-Verlag, Berlin, 291–307. Nepal. Lu, D. and Weng, Q. 2007. Survey of image Dorren, L. K. A., Maier, B. and Seijmonsbergen, classification methods and techniques for A. C. 2003. Improved Landsat-based forest improving classification performance. mapping in steep mountainous terrain using International Journal of Remote Sensing 28 object-based classification. Forest Ecology (5): 823–870. and Management 183: 31–46. Maniezzo, V., Morpurgo, R. and Mussi, S. 1993. Ehlers, M., Gaehler, M. and Janowsky, R. 2003. D-KAT: A Deep Knowledge Acquisition Automated analysis of ultra high resolution Tool. Expert Systems 10 (3): 157–166. remote sensing data for biotope type mapping: New possibilities and challenges. Meinel, G. and Neubert, M. 2004. A comparison ISPRS Journal of Photogrammetry and of segmentation programs for high-resolution Remote Sensing 57: 315–326. remote sensing data. Commission VI in Proceeding of XXth International Society Ehlers, M. 2004. Remote sensing for GIS for Photogrammetry and Remote Sensing applications: New sensors and analysis (ISPRS) Congress, 12–23 July, Istanbul, methods. In Remote Sensing for Environmental Turkey, unpaginated (CD-ROM). Monitoring, GIS Applications, and Geology III (eds) Ehlers, M., Kaufmann, J. J. and Metzler, J. W. and Sader, S. A. 2005. Model Michel, U. Proceedings of SPIE, Bellingham, development and comparison to predict Washington, USA, 1–13. softwood and hardwood per cent cover using high and medium spatial resolution imagery. Ferro, C. J. S. and Warner, T. A. 2002. Scale International Journal of Remote Sensing 26 and texture in digital image classification. (17): 3749–3761. Photogrammetric Engineering and Remote Sensing 68 (1): 51–63. Qian, Y., Zhou, W., Yan, J., Li, W. and Han, L. 2015. Comparing machine learning classifiers Gong, P., Marceau, D. J. and Howarth, P. J. 1992. for object-based land cover classification A comparision of spatial feature extraction using very high resolution imagery. Remote algorithms for land-use classification with Sensing 7: 153–168. SPOT HRV data. Remote Sensing 40: 137– 151. Sandmann, H. and Lertzman, K. P. 2003. Combining high-resolution aerial photo- Guo, Q., Kelly, M., Gong, P. and Liu, D. 2007. graphy with gradient-directed transects to An object-based classification approach in guide field sampling and forest mapping in mapping tree mortality using high spatial mountainous terrain. Forest Science 49 (3): resolution imagery. GIS science and Remote 429–443. Sensing 44 (1): 24–47. Souza, C. M. and Roberts, D. 2005. Mapping Harvey, K. R. and Hill, G. J. E. 2001. Vegetation forest degradation in the Amazon region with mapping of a tropical freshwater swamp in the IKONOS images. International Journal of Northern Territory, Australia: A comparison Remote Sensing 26 (3): 425–429. of aerial photography, Landsat TM and SPOT satellite imagery. International Journal of Townshend, J. R. G., Huang, C., Kalluri, S. N. Remote Sensing 22 (15): 2911–2925. V., Defries, R. S., Liang, S. and Yang, K. 2000. Beware of per-pixel characterization of Kim, M., Madden, M. and Warner, T. A. 2008. land-cover. International Journal of Remote Estimation of optimal image object size Sensing 21 (4): 839–843. for the segmentation of forest stands with multispectral Ikonos imagery, Object- USGS. 2015. NASA EOSDIS Land Processes based Image Analysis - Spatial Concepts DAAC, USGS Earth Resources Observation

43 Banko Janakari, Vol. 26, No. 1 Chaudhary et al.

and Science (EROS) Center, Sioux Falls, identification of individual trees with local South Dakota. http://www.gdem.aster.ersdac. maxima filtering. International Journal of or.jp/ accessed on 1 May, 2015. Remote Sensing 25 (11): 2225–2232.

Wang, L., Sousa, W. P. and Gong, P. 2004a. Yu, Q., Gong, P. N., Clinton, G., Biging, M. K. and Integration of object-based and pixel-based Shirokauer, D. 2006. Object-based detailed classification for mapping mangroves with vegetation classification with airborne high Ikonos imagery. International Journal of spatial resolution remote sensing imagery. Remote Sensing 25 (24): 5655–5668. Photogrammetric Engineering and Remote Sensing 72 (7): 799–811. Wulder, M. A., White, J. C., Niemann, K. O. and Nelson, T. 2004. Comparison of airborne and satellite high spatial resolution data for the

44 Ecological status and diversity indices of Panchaule (Dactylorhiza hatagirea) and its associates in Lete village of , Nepal C. B. Khadka1, A. L. Hammet2, A. Singh3, M. K. Balla3 and Y. P. Timilsina3

This paper focuses on the ecological status and diversity indices: Simson’s Index (C), Simson’s Index of Dominance (D) and Shannon-Weaver Index (H) of Dactylorhiza hatagirea and its associates- Rheum australe and Rumex nepalensis in Lete village of Mustang District within the Annapurna Conservation Area. The study was conducted during the monsoon season (June/July) of 2013 in the Lete VDC of Mustang District. The study site possessed an area of 4.5 ha. Altogether, 100 circular plots, each with 25 m2 area, were laid out purposively within the study area; the sampling intensity being 5.55%. The relative frequency, the relative density, the abundance, the relative coverage and the Important Value Index of the species were found to be 61.11, 53.91, 1,061.54, 72.2 and 187.24 respectively. Similarly, the Simson’s Index (C), the Simson’s Index of Dominance (D) and the Shannon-Weaver Index of the species were found to be 0.41, 0.59 and 3.27 respectively, indicating relatively even and relatively diverse community. The study showed relatively higher values of all the parameters of D. hatagirea as compared to its associates indicating good ecological value. However, threats remain due to the illegal harvesting of this valuable orchid and overgrazing in the study site.

o Biodiversity, community, conservation, medicinal plants

epal constitutes a unique and enormous the development and commercialisation of Ndiversity of flora and fauna within a relatively Medicinal and Aromatic Plants (MAPs) as a small geographical area due to variations in priority programme for poverty alleviation. topography, altitude and climate. In spite of being This shows the commitment of the government a small country, it possesses around 7,000 species for conservation and management of medicinal of vascular plants with 2,000 species of medicinal plants in the nation. Rare and high-priced plants (Shrestha and Shrestha, 1999). Baral and medicinal herbs are on the top priority for Kurmi (2006) have compiled and described 1,792 their domestication, research and cultivation, medicinal plants. According to Bhattarai and processing and marketing. Ghimire (2006), 49% of the traded medicinal plants are herbs, 29% trees, 14% shrubs and 8% MAPs of high altitudes are invaluable resources climbers. So, Nepal is a veritable treasure trove of not only to the local communities and the nation, medicinal plants (Phoboo et al., 2008). but also to the global community at large. They have high ecological as well as economical During the last 10 years, a great interest has been values, and so the poor rural communities are given for the promotion of Non-Timber Forest highly dependent on them for their livelihoods. Products (NTFPs) throughout the world. Huge sums have been invested in exploring the potential Out of many MAPs, Dactylorhiza hatagirea of NTFPs (Wollenberg, 1999). Nepal is also not (D. Don) Soo under the family of Orchidaceae far from this condition. The Master Plan for commonly known as “Panchaunle” in Nepali Forestry Sector (1988), Forest Policy (2015) and has been listed in Appendix II by the Convention the Thirteenth Plan (2013–2016) had emphasized on International Trade in Endangered Species

1 Chitwan National Park, Chitwan, Nepal. Email : [email protected] 2 Virginia Polytechnic Institute, Virginia, USA 3 Institute of Forestry, Pokhara, Nepal

45 Banko Janakari, Vol. 26, No. 1 Khadka et al. of Wild Fauna and Flora (CITES), vulnerable hatagirea is one of them (DPR, 2006). According species listed by the Conservation Assessment to Forest Act 1993 and Forest Regulation and Management Plan (CAMP) and threatened 1995, the Government of Nepal has banned the species by the International Union for collection, trade and processing of the rhizome of Conservation of Nature (IUCN) (Samant et al., D. hatagirea. 2001). This Himalayan endemic medicinal orchid is found in Hindu Kush Himalaya range (IUCN, As with many other terrestrial orchids, the 2004). It occurs in the sub-alpine and the alpine populations of D. hatagirea have decreased zones between 2,800 m – 4,200 m altitude above due to their habitat loss. Another threat to D. the mean sea level (IUCN, 2004). Other than hatagirea is the collection of their tubers to make Nepal Himalayas, it occurs in the same altitudinal salep (paste made to cure wounds), used as food ranges of India, Pakistan, Bhutan and China too. and medicine. This is a particularly important threat in the Himalayas (Srivastava and Mainera, It is a terrestrial-ground-dwelling perennial 1994), where D. hatagirea is judged critically herb. Its stem is erect, hollow and obtuse, and endangered (CAMP, 1998) due to their over- bears palmately lobed and lanceolate leaves with collection. Thus, several species of D. hatagirea sheathing leaf-base. The cylindrical and terminal are declining, and some are already protected at spike bears rosy purple flowers with green bracts a national scale, e.g., in Belgium, Luxembourg, (Fig. 1). Flowers are 1.7–1.9 cm long with curved Nepal, and the United Kingdom (Pillon et al., spur. The inflorescence consists of a compact 2005). raceme with 25 to 50 flowers developed from axillary buds. The dark purple spotted lip of the The concept of diversity, including biodiversity flower is rounded and lobed (1 to 5). The plants itself as well as the narrower concept of species store a large amount of water in their tuberous diversity, is a human reflection without any roots to survive in arid conditions (Chaurasia et unique mathematical meaning. The simplest al., 2007). measure of species diversity is species richness, but a good case can be made for giving some weight to evenness as well. Diversity indices are mathematical functions that combine richness and evenness in a single measure, although usually not explicitly. Although there are many others, the most commonly used diversity indices in ecology are Shannon Diversity, Simpson Diversity, and Fisher’s Alpha. Both Shannon and Simpson diversities increase as richness increases, for a given pattern of evenness, and increase as evenness increases, for a given richness, but they do not always rank communities in the same order. Simpson Diversity is less sensitive to richness and more sensitive to evenness as compared to Shannon Diversity, which, in turn, is more sensitive to evenness than is a simple count of species richness (S) (Colwell, 2009).

There are very few studies conducted especially regarding ecological status and diversity of D. hatagirea. There is a lack of management and conservation plan from the government side. Fig. 1: D. hatagirea plant Similarly, lack of awareness of importance regarding D. hatagirea among the rural villagers The Government of Nepal has prioritized 30 is leading towards the extinction of this valuable important medicinal plants for the purpose of species. Although it is a banned species, its unwise their research and management. Among those, 12 harvesting, unscientific use and illegal trading plants have been selected for agro-technology. D. are in practice which in turn is resulting in the

46 Khadka et al. Banko Janakari, Vol. 26, No. 1 reduction of net income of the primary collectors Sampling Design and the government revenue. The objectives of the research study were: The inventory was carried out in 4.5 ha during the monsoon season (June/July) in 2013. A 21 m x 21 i) to document the ecological status of D. m grid (plot to plot distance) was laid on the map hatagirea and its associates in the study site, and, altogether, 100 plots were established using and the Arc Map Geographical Information System ii) to calculate diversity indices of D. hatagirea (GIS) Software of version 10; the sampling and its associates in the study site. intensity being 5.55% (the greater intensity was due to the small area of the study site). Among the 100 plots, the plants of D. hatagirea were found Materials and methods in 34 plots only. The sample plots were chosen Study area purposively so that the plots where the plants of D. hatagirea and its associates were found The study was conducted in Paplekharka, a grass would not be left during the field inventory as the land situated in the Lete Village Development topography in some plots was either rocky or stiff Committee (VDC) of Mustang District (Fig. 2) or with dense forest cover or barren (Fig. 3). which lies in the Annapurna Conservation Area (ACA).

Mustang District lies between 28o24’ N to 29o20’ N latitude and 83o30’ E to 84o15’ E longitude. The altitudinal range varies from 1,372 m to 8,167 m above the mean sea level representing sub- tropical, temperate and alpine types of climate (Ranapal, 2009). The Lete VDC lies within the Jomsom Unit Conservation Office (UCO) of the Lower Mustang which is a transition between the Trans-Himalaya and the Inner Himalaya. The Fig. 3: Google Image (20-11-2014) of the Lete VDC receives 1,545 mm annum rainfall (ACAP, VDC showing the location of the sample plots 2009). It consists of deep gorges made by the for the inventory of D. hatagirea Kaligandaki River. All the plants of D. hatagirea found in 26 plots were measured. The number of plots in which D. hatagirea was absent was 15 (in which other herbs were also present) excluding the plots with barren, forest, rocky and stiff topography. Similarly, out of the total 100 plots, 20 plots included barren area, 16 plots forest area and 23 plots rocky and stiff area in which inventory was not possible.

Study methods Primary data collection

The primary data was collected through reconnaissance survey and herb inventory. First of all, a reconnaissance survey was carried out for general field observation, rapport building with the local people and the ACAP personnel about the trail of the location where D. hatagirea a found. Besides, a sketch map was also prepared Fig. 2: Landuse map of the study site for carrying out the field inventory smoothly.

47 Banko Janakari, Vol. 26, No. 1 Khadka et al.

Generally, 1 m × 1 m sample plots are used for No. of species in all plots x 100 D) Relative Density = inventory of herbs. However, as D. hatagirea Total no. of individuals of all species is a low abundant herb, circular sample plots of 25 m2 were established as recommended No. of species in all plots x 10,000 m2 E) Abundance = by Ravindranath and Premnath (1997). All the No. of plots in which a species occurs x area of a plot plants of D. hatagirea and its associate species Area occupied by a species x 100 were counted and their mean height, mean collar F) Coverage (%) = diameter (5 cm above the ground) and mean age Area of a sample plot were measured within the sample plots. The mean Coverage of a species x 100 age of D. hatagirea was calculated by summing up G) Relative Coverage (%) = all the ages of D. hatagirea plants and dividing by Total coverage of all species the number of plots where D. hatagirea occurred. H) Important Value Index (IVI)=Relative Frequency+Relative Similar process was repeated to calculate the Density+Relative Coverage mean height and the mean collar diameter of I) Simson’s (1949) Index (C) = ∑ (P )2 and the species. A Global Positioning System (GPS) i J) Simson’s Index of Dominance (D) = 1 - ∑ (P )2, set was used to locate the plots, and a 20–meter i

Reel Tape was used to measure the radius of the Where, Pi is the proportion of the important value th plot. Similarly, Vernier Calliper was used to find of the i species (Pi = ni/N, ni being the IVI of the out the collar diameter of the herb whereas a 5– ith species and N being the Important Value of all feet Steel Tape was used to detect the height of species) the herb. An Inventory Sheet was developed to record the details for calculating the ecological The Simpson’s Index values range from 0 to 1. status and the diversity indices of the species. The closer the value of Simpson’s Index to 0, the more diverse the plot will be. A plot with only one Secondary data collection species would have a Simpson’s Index value of 1. Trends are opposite to those found for Shannon- The secondary data were obtained with the help Weaver values since Simpson’s Index values of the annual reports, newsletters, bulletins, decrease with increased diversity (Reich et al., journals, dissertations, publications, maps and so 2001). In practice, the values below 0.5 indicate a on available in the ACAP Libraries as well as at relatively even community, while high values are the Department of Forest Research and Survey indicative of communities dominated by one or a (DFRS), the Institute of Forestry (IoF) and the few species. International Centre for Integrated Mountain K) Shannon-Weaver (1949) Index (H) = – ∑ (p ) x ln (p ), Development (ICIMOD). Additional information (i=1) i i were also acquired through internet. Where, H = index of species diversity, and th Data analysis and interpretation Pi = the proportion of the important value of i

species (Pi = ni/N, ni is the important value index The information obtained from the herb inventory of the ith species and N is the important value of was analyzed using the Statistical Package for all species) Social Science (SPSS) Software version 19. The results were then presented in the form of tables, Due to its logarithmic nature, the Shannon- graphs and charts. The quantitative data was Weaver Index is sensitive to uncommon plant analyzed as follows: species and less sensitive to very common species (Krebs, 1989). The Shannon-Weaver Index can, No. of plots where species occcurs x 100 A) Frequency = theoretically, range from zero (a community Total no. of plots with only one species, which is technically just a “population”) to infinity. In practice though, a Frequency of species x 100 B) Relative Frequency = value of 7 indicates an extremely rich community Sum of all frequency while values under 1 suggest a community with No. of species in all plots x 10,000 m2 low diversity. Often, values above 1.7 are taken C) Density = Total no. of plots x area of a plot to indicate a relatively diverse community. The qualitative data was analyzed descriptively.

48 Khadka et al. Banko Janakari, Vol. 26, No. 1 Results and discussion has indicated the occurrence of D. hatagirea in Paplekharka as 71%. Similarly, Rheum australe Distribution of D. hatagirea was found to have an occurrence of 8% as compared to 65% found by Ranapal (2009). A total of two scientifically identified herbs viz. On the other hand, Rumex nepalensis had an Himalayan rhubarb (Rheum australe) and Nepal occurrence of 20% (Fig. 4). dock (Rumex nepalensis) were found in the project site with dominance of D. hatagirea (Table 1). The altitudinal range of habitat distribution of D. hatagirea in the study site was from 3,200 m to 3,600 m above sea level, which was similar to the study done by Ranapal (2009). The aspect of habitat distribution of D. hatagirea in the study site was South-West.

Mean height, collar diameter, number of leaves and age of D. hatagirea

The height, collar diameter and age of each D. hatagirea plant were measured in each plot in which a total of 69 D. hatagirea plants were found in 26 measured plots. Fig. 4: Frequency of D. hatagirea and its associates From the inventory, it was found out that the mean height of D. hatagirea was 91.08 cm which was Relative frequency of D. hatagirea and its greater than 41.97 cm found by Ranapal (2009) associates and 60 cm by Dutta (2007). Ranapal (2009) used three individual plants (tall, medium and short) Relative frequency is the frequency of a species to calculate the mean height of D. hatagirea. in relation to other species. The relative frequency Similarly, the diameters (at 5 cm above the of D. hatagirea was high (69%) as compared to ground) of D. hatagirea plants were measured, its two associates. However, according to Ranapal and the mean diameter was calculated. The mean (2009), it was 17%. The species having the lowest frequency were Rumex nepalensis (22%) and the collar diameter at 5 cm above the ground was Rheum australe (9%). found to be 1.63 while the mean age was found to be of 2 years and the mean number of leaves Density of D. hatagirea and its associates was found to be 5. The difference in height might be due to the methodology used as well as the This study showed the highest density of 276 age/topographic/soil/climate factors. The greater per ha of the Himalayan orchid (D. hatagirea) number of D. hatagirea was found in the South- as compared to its associates (Table 2). This west aspect. was comparatively far lower than the one (1,671 per ha) indicated by Ranapal (2009). The big Frequency of D. hatagirea and its associates difference might be due to the smaller area (4.5 ha) of the project site taken during the research The study shows 72% occurrence of D. hatagirea period. The least number was that of Nepal dock in the sample plots. However, Ranapal (2009) with 80 per ha which might be due to the least Table 1: List of identified herbs in the sample plots S.N. Local name English name Scientific name Family Nature of plant 1. Panchaunle Himalayan Orchid Dactylorhiza hatagirea Orchidaceae Herb (D. Don) Soo 2. Padamchal Himalayan Rhubarb Rheum australe D. Polygonaceae Herb Don 3. Halhale Nepal Dock Rumex nepalensis Polygonaceae Herb Spreng.

49 Banko Janakari, Vol. 26, No. 1 Khadka et al. distribution of this species in the plot and heavy Coverage and relative coverage of D. hatagirea grazing pressure. The density of D. hatagirea a and its associates reported to be 0.2 individuals per m2 in Samar Lek of the Upper Mustang (Chhetri and Gupta, D. hatagirea was found to have the highest 2006). The next reported density of D. hatagirea coverage (26%) and highest relative coverage was 2.66 per m2 in the grazed sites and 3.2 per m2 (72.2%) as compared to those of Rheum australe in the ungrazed sites at Tungnath, India (Nautiyal and Rumex nepalensis (Table 4). The highest et al., 2004). The low density in the unprotected coverage of D. hatagirea was due to the highest areas might be due to heavy grazing pressure. number of plots in which it occurred as compared to its associates. Table 2: Density of D. hatagirea along with its associates Table 4: Coverage and relative coverage of D. hatagirea along with its associates S.N. Species Density per ha S.N. Species Coverage Relative 1. D. hatagirea 276 (%) Coverage (%) 2. Rheum australe 156 1. D. hatagirea 26.00 72.22 3. Rumex nepalensis 80 2. Rheum australe 7.00 19.45 3. Rumex 3.00 8.33 Relative density of D. hatagirea and its nepalensis associates Total 36.00 100.00 Relative density is the density of a species with respect to the total density of all species (Ranapal, Important Value Index of D. hatagirea and its 2009). In the study site, D. hatagirea was found associates to have the highest relative density (53.91%) as compared to those of its associates (Table 3). D. hatagirea was found to have the highest IVI However, the study conducted by Ranapal (2009) (187.24) as compared to its two associates (Table indicated the relative density of D. hatagirea to 5), indicating its dominance in the study site. be quite low, only 9% or 0.09. The big difference Table 5: Important Value Index of D. hatagirea in the relative density of this species might be along with its associates because of the smaller area (4.5 ha) of the project site taken during the research period. S.N. Species IVI 1. D. hatagirea 187.24 Table 3: Relative density of D. hatagirea along with its associates 2. Rheum australe 69.35 3. Rumex nepalensis 32.29 S.N. Species Density per ha Total 288.88 1. D. hatagirea 53.91 2. Rheum australe 30.47 Diversity indices for D. hatagirea and its 3. Rumex nepalensis 15.62 associates Total 100.00 The Simson’s Index (C) and the Simson’s Index of Dominance (D) were found to be 0.41 and 0.59, Abundance of D. hatagirea and its associates respectively (Table 6), indicating relatively even Rumex nepalensis was found to have the highest community and higher dominance of one species abundance of 2,666.67 per ha followed by Rheum i.e. D. hatagirea in the study site. Similarly, the australe with 2,228.57 per ha with the Himalayan Shannon-Weaver Index (H) was found to be 3.27, orchid having the least abundance of 1,061.54 per indicating relatively diverse community in the ha. However, the abundance of D. hatagirea a study site. The difference was due to the presence reported by Ranapal (2009) was higher (2,367 per of the greater number of D. hatagirea individuals ha) than the one found in this study; the difference as compared to its two associates. The diversity might be due to the least number of Himalayan of the species in the study site was, therefore, not orchid plants found in the study area. satisfactory.

50 Khadka et al. Banko Janakari, Vol. 26, No. 1

Table 6: Simson’s index of dominance and for their necessary support, and Department of Shannon-Weaver Index for D. hatagirea and National Parks and Wildlife Conservation for its associates allowing permission to conduct the research in the study area. Diversity Index Value Remarks Simsons' Index 0.41 Relatively even References (C) community Simsons' Index 0.59 Higher dominance ACAP. 2009. Management Plan of Annapurna of Dominance of one species i.e. Conservation Area (2009–2012). Annapurna (D) D. hatagirea Conservation Area Project, Pokhara, Nepal. Shanon-Weaver 3.27 Relatively diverse Baral, S. R. and Kurmi, P. P. 2006. A Compendium Index (H) community of Medicinal Plants in Nepal. Kathmandu, Nepal. Conclusion Bhattarai, K. R. and Ghimire, M. D. 2006. Commercially important medicinal and The study revealed that frequency, relative aromatic plants of Nepal and their distribution frequency, relative density, relative abundance pattern and conservation measure along the and relative coverage of D. hatagirea r elevation gradient of the Himalayas. Banko higher as compared to its two associates- Rhuem Janakari 16 (1): 3–13. australe and Rumex nepalensis. This indicated CAMP. 1998. Executive Summary Report the good ecological status of D. hatagirea in the of the Conservation Assessment and study area. However, the value of Simson’s Index Management Plan (CAMP) of the (C) indicated the relatively even community and Biodiverity Conservation Prioritisation the Simson’s Index of Dominance (D) indicated Project on Selected Medicinal Plants of the dominance of one species i.e. D. hatagirea Northern, North-eastern and Central while the value of the Shannon-Weaver Index (H) India. [http://msubiology.info/vesna/nauka/ indicated relatively diverse community i.e. the pillon2006.pdf accessed on 29 March, 2014. plant diversity was found to be not satisfactory, suggesting for necessary actions for the Chaurasia, O. P., Ahmed, Z. and Ballabh, B. conservation of the diversity of D. hatagirea in 2007. In Ethnobotany and Plants of Trans- the study area. Further research is recommended Himalaya. Satish Serial Publishing House, on an annual basis so as to maintain database Delhi, India. on the population dynamics and the harvesting Chhetri, H. B. and Gupta, V. N. P. 2006. NTFP level of this valuable medicinal plant. Research potential of Upper Mustang, a Trans- on genetic diversity using Molecular Marker Himalayan region in western Nepal. Scientific Technique is also recommended to compare the World 4 (4): 38–43. genetic diversities of the populations of this orchid at different locations of Nepal. Although the Colwell, R. K. 2009. “Biodiversity: Concepts, occurrence of D. hatagirea was found to be higher Patterns and Measurement”. In The Princeton than its two associates, the illegal harvesting of Guide to Ecology (ed) Levin, S.A. Princeton this valuable orchid and overgrazing in the study University Press, Princeton, 257-263. http:// site are likely to bring it to extinction. Therefore, press.princeton.edu/chapters/s3_8879.pdf awareness programmes about the in-situ and ex- accessed on 23 August, 2014. situ conservation of the endangered medicinal DPR. 2006. Prioritized Medicinal Plants orchid D. hatagirea should be conducted in the for Economics Development of Nepal. study area. Government of Nepal, Department of Plant Resources (DPR), Kathmandu, Nepal. Acknowledgements Dutta, I. C. 2007. Non-Timber Forest Products We are obliged to the Rufford Small Grants of Nepal: Identification, Classification, Foundation, United Kingdom for financial Ethnic Uses and Cultivation. Hill Side support to conduct the study. We are grateful to Press, Kathmandu, Nepal. the research assistants, the community people of IUCN. 2004. National Register of Medicinal Ghasa and Annapurna Conservation Area Project and Aromatic Plants. International Union

51 Banko Janakari, Vol. 26, No. 1 Khadka et al.

for Nature Conservation Nepal, Kathmandu, Reich, P. B., Bakken, P., Carlson, D., Frelich, Nepal. L., Friedman, S. K. and Grigal, D. 2001. Influence of logging, fire and forest type on Krebs, C. J. 1989. Ecological Methodology. biodiversity and productivity in southern Harper Collins College Publishers, New boreal forests. Ecology 82 (10): 2731–2748. York, USA. Samant, S. S., Dhar, U. and Rawal, R. S. 2001. Nautiyal, M. C., Nautiyal, B. P. and Prakash, V. Himalayan Medicinal Plants- Potential and 2004. Effect of grazing and climatic changes Prospects (eds) Samant, S. S., Dhar, U. on alpine vegetation of Tungnath, Garhwal and Palni, L. M. S. Gyanodaya Prakashan, Himalaya, India. The Environmentalist 24: Nainital, India, 166–184. 125–134. Shanon, C. E. and Weaver, W. 1949. The Phoboo, S., Devkota, A. and Jha, P. K. 2008. Mathematical Theory of Communication. Medicinal Plants in Nepal- An Overview. University of Ilinois Press, Urbana, Ilinois, Medicinal Plants in Nepal, an anthology of USA. contemporary research. Ecological Society, Kathmandu, Nepal. Shrestha, G. L. and Shrestha, B. 1999. An Overview of Cultivated Plants in Nepal; Pillon, Y., Fay, M. F., D. S., Shipunov, A. and Proceedings of National Conference, the Chase, M. W. 2005. Species diversity versus Green Energy Mission/Nepal, Kathmandu, phylogenetic diversity: a practical study in the Nepal. taxonomically difficult genus Dactylorhiza (Orchidaceae). Biological Conservation 129: Simpson, E. H. 1949. Measurement of diversity. 4–13. Nature 163: 68. Ranapal, S. 2009. An Assessment of Status and Srivastava, R. C. and Mainera, A. K., 1994. A Antibacterial Properties of Dactylorhiza note on Dactylorhiza hatagirea (D. Don) hatagirea in Annapurna Conservation Area Soo: an important medicinal orchid of (A Case Study of Paplekharka, Lete VDC, Sikkim. National Academy of Science Letters Mustang). B.Sc. Forestry Research Thesis 17: 129–130. submitted to Tribhuvan University, Institute Wollenberg, E. 1999. Methods for assessing of Forestry, Pokhara, Nepal. the conservation and development of forest Ravindranath, S. and Premnath, S. 1997. Biomass products. In Incomes from the Forest (ed) Studies: Field Method for Monitoring Wollenberg, E. and Ingles, A. CIFOR. Biomass. Oxford and IBH Publishing Co. Indonesia, 1–16. Pvt. Ltd., India.

52 Identification of land reclamation area and potential plantation area on Bagmati river-basin in the Terai region of Nepal A. K. Acharya1*, A. K. Chaudhary1 and S. Khanal1 Utilization of land reclamation area offers the potentiality of increasing greenery as well as providing forest products. This study refers to the identification of the land reclamation areas and potential plantation areas on the Bagmati river-basin in the Terai region of Nepal, and recommends appropriate species for plantation in order to rehabilitate such areas. Multi-temporal Landsat Satellite Images (Landsat 7 and Landsat 8) were acquired for 2002 and 2014. Object-based Image Classification method was used to classify the land cover classes into four broad categories: i) Water, ii) Sand and gravel, iii) Plantation potential (open areas suitable for plantation) and iii) Others (forest, agriculture, built-up areas etc.). The Mean Normalized Difference Water Index (NDWI) values and Mean Brightness values were found to be helpful in identifying the water and sand & gravel areas from the other land cover classes. The overall classification accuracy was 0.97 with a kappa coefficient of 0.89 in the case of the 2014 Image classification. In this study, the land reclamation area referred to the areas occupied by water, sand & gravel on the river-beds that were converted into plantation potential and other classes between 2002 and 2014. Similarly, the potential plantation area referred to the summation of the area of reclaimed land, the area of ‘Others’ class converted into ‘Plantation potential’ class and the area that remained to be plantation potential on the bed of the Bagmati River and its tributaries between 2002 and 2014. Altogether, 4,819.10 ha land was reclaimed in the study area, and a total of 5,395.10 ha land was found to be potential for plantation within the study area. o Bagmati river basin, land reclamation, object-based image classification, potential plantation area, Terai

and use and land cover data are essential for Land reclamation area is, in general, the land Lplanners, decision makers and managers for created by river after changing its course. In natural resources management. Up-to-date land Nepal, most of the rivers debouch into the Terai use and land cover information are required for plains at the foot hills of the Churia, and provide monitoring and analysis of natural resources water for livelihood of the people living in the to support their sustainable management Terai. During the monsoon months from June (Xian, 2009). Remote Sensing and Geographic to September, all these rivers get inundated Information System (GIS) techniques have with bank-full discharges, and cause flooding in been recognized as an effective tool for the several parts of the Terai. Most of the rivers in the classification of land use and land cover data and Terai are prone to change their course frequently. assessment of trend, rate, nature, location and In many cases, these rivers find new paths, and magnitude of the changes (Adeniyi et al., 1999). enter into the cultivated lands leaving the old The remote sensing technology has offered a courses (Adhikari, 2013). wide variety of satellite imagery that covers most of the earth’s surface. The multi-sensor and multi- Most of the flooded river-side areas are in temporal satellite data is a promising tool for unutilized state due to low availability of nutrients producing accurate land cover maps (Yoon et al., in the sandy soil. Plantation could be a better 2004). Water resources assessment and coastal option to manage such land reclamation areas. It management (Xu, 2006) is one of the many field will be helpful to fulfill the growing demand for of applications of remotely sensed imagery. forest products of the local people. It will not only

1 Department of Forest Research and Survey, Kathmandu, Nepal Email: [email protected]

53 Banko Janakari, Vol. 26, No. 1 Acharya et al. provide goods but also several ecosystem services such as landscape beauty, carbon sequestration, water table recharge, creation of wildlife habitat, prevention of river-bank erosion and so on.

The Government of Nepal has recently promulgated “Forest Policy 2015” which emphasizes on plantation program in forest as well as public and private lands. The Ministry of Forests and Soil Conservation (MoFSC) expends its fund in plantation activities through its departments. Recently, the MoFSC has declared “forest decade program (2014–2023)” Fig. 1: Map showing the location of the study for growing greenery through massive plantation. area Thus, land reclamation area will be potential sites for plantation to support the forest decade Methods program. The methodology used in this study is highlighted So far, no adequate number of GIS-based studies in Fig. 2. The data analysis steps are explained in that deal with identification and mapping of the the subsequent sections. land reclaimed areas have been conducted in Nepal. Therefore, study on land reclamation has great importance for future plantation activities. The objectives of this study were to identify the land reclamation areas and the potential plantation areas on the Bagmati river-basin in the Terai region, and recommend appropriate species for plantation in order to rehabilitate such areas. Materials and methods Study area The study was confined to the Bagmati river- basin in the Terai physiographic region (Fig. 1). Geographically, the study area extends from 26°44’30” N to 27°12’22” N latitude and from 85°15’59” E to 85°34’02” E longitude. It includes 34 Village Development Committees (VDCs) of and 37 VDCs of , and covers 61,279.78 ha area.

The Bagmati River originates from the Mid- hills (Shivapuri hill situated in the north of the Kathmandu Valley), and drains the Gangetic plain flowing across the Mahabharat range and the Churia hills. It covers an area of 3,640 km2. The basin of this river, thus, transects three distinct latitudinal physiographic zones viz. Mountain, Siwalik and Terai of Nepal (Paudel, 2001). The study was conducted in February to June 2015. Fig. 2: Flow chart of methodology

54 Acharya et al. Banko Janakari, Vol. 26, No. 1

Data heterogeneity of image-objects for a given resolution (Su et al., 2009). The multi-temporal satellite imageries of Landsat 7 (ETM+) and Landsat 8 were used The Normalized Difference Water Index (NDWI) for the purpose of the study (Table 1). Both the developed by Mcfeeters (1996) was used to Imageries were stacked in layers with the help of delineate water features in the study. NDWI ArcMap Software. The stacked Imageries were is derived using the principles similar to the merged with higher-resolution panchromatic Normalized Difference Vegetation Index (NDVI), Image (Band 8 for Both Landsat 7 and 8) so as and is defined as follows: to create a single higher-resolution (15 m) color NDWI = (GREEN – NIR)/(GREEN + NIR), image. The Bagmati river-basin boundary and the LRMP physiographic boundary layers were also Where, GREEN is green band and NIR is near used to define the study area extent. infrared band.

Object-based image classification The reason behind the selection of these bands is to: (i) maximize reflectance of water by Object-based image analysis (OBIA) is a using green wavelengths; (ii) minimize the low technique developed to overcome the problem reflectance of NIR by water features and (iii) take of traditional pixel-based image analysis. Pixel- advantage of the high-reflectance of NIR from based image analysis is based on the information vegetation and soil features. As a result, water in each pixel whereas the object-based image features have positive values whereas vegetation analysis is based on information from a set of and soil usually have zero or negative values similar pixels called objects or image-objects. (Mcfeeters, 1996). OBIA reduces the local spectral variation caused by crown textures, gaps, and shadows. In addition For both the Landsat Imageries of 2002 and 2014, to this, both spectral values and spatial properties, the object-based image analysis was executed such as size and shape, can be explicitly utilized using the eCognition Developer 8.7 Software. as features for further classification with The thematic layer of the study area was used spectrally homogeneous segments of images. In in chessboard segmentation. Multi-resolution this process, spatially adjacent pixels are grouped segmentation was performed at scale parameter into spectrally homogenous objects first, and then of 20 with shape 0.5 and compactness 0.5. Blue, conduct classification on objects as the minimum Green, Red, Near Infrared, Short-wave Infrared-1 processing units (Yu et al., 2006). Basically, there and Short Wave Infrared-2 bands were used in are two steps involved in OBIA. They are: (i) multi-resolution segmentation. The segmented image segmentation to produce image objects (or images were classified into four land cover classes segments) that are the relatively homogeneous viz. i) Water ii) Sand and gravel, iii) Plantation groups of pixels, and (ii) image classification potential (open area suitable for plantation) and based on these image-objects (Dorren et al., iii) Others (forest, agriculture, built-up areas etc.) 2003; Meinel and Neubert, 2004; Addink et al., using the following parameters: 2007). Two image segmentation methods viz. a) Mean brightness values, Chessboard Segmentation Method and Multi- resolution Segmentation are frequently used b) Mean NDWI, and in OBIA. Chessboard Segmentation splits the c) Mean relation border to neighbors. pixel domain or an image-object domain into square image-objects whereas Multi-resolution In the first step, sand and gravel were separated Segmentation Method is an optimization using the mean brightness values of the Images. procedure which locally minimizes the average NDWI was used to separate water bodies. All the water bodies outside the Bagmati River and

Table 1: Characteristics of Landsat image Satellite Sensor Path-Row Date Resolution (m) Band Landsat 7 ETM+ 141–41 Nov 5, 2002 30, 15 (Band 8) Band 1 to 8 Landsat 8 OLI and TIRS 141–41 Nov 28, 2014 30, 15 (Band 8) Band 2 to 8

55 Banko Janakari, Vol. 26, No. 1 Acharya et al. its tributaries were categorized into ‘Others’ samples from the river reclamation areas were class. ‘Plantation potential’ was separated using also verified through field observation. mean relation border to neighbors. In this study, ‘plantation potential’ referred to the open area Social survey for plantation near the river (near water and sand Focus-group discussion is one of the widely used and gravel classes). The remaining areas were methods to collect data in social survey. In this categorized into “Others”. The classified images study, focus-group discussion was organized for were exported for further analysis. identification of suitable species to rehabilitate Visual interpretation and change detection the land reclamation areas through plantation. A total of four local-level focus-group discussions, To improve the classification, the classified 2 in Rautahat district and 2 in Sarlahi district, and polygons were visually interpreted with pan- one district-level focus-group discussion were sharped Landsat images of 15 m resolution in organized in Rautahat district. ArcMap 10. ‘Union’ function was applied on classified data to find out the changes between Results and discussion 2002 and 2014. The Landsat Images of 2002 and 2014 were classified into four types of land cover classes Accuracy assessment viz. i) Water, ii) Sand and gravel, iii) Plantation Using new and innovative methodology which potential and iv) Others. The land cover maps of represents new area of ecological data, it is the study area for the years- 2002 and 2014 are important to test accuracy of data being produced presented in Fig. 3. (Mathieu et al., 2007). A greater effort should be made for checking how much ‘misclassification’, e.g. omitted classification has occurred in course of land use classification. In the case of our study, the user’s accuracy, the producer’s accuracy, the overall accuracy as well as the kappa coefficient were used to test the accuracy assessment result.

Training-samples were selected systematically to assess the accuracy of classification. Point-grids at a spacing of 500 m x 500 m were generated systematically over the entire study area. These point-grids were spatially joined using the already classified polygons. From each class, 10% sample was selected; altogether, 243 sample points (6 Fig. 3: Land cover map of the Bagmati river- from ‘Water’, 6 from ‘Sand and gravel’, 11 from basin in the Terai region in 2002 and 2014 ‘Plantation potential’ and 220 from ‘Others’) were verified with the help of high-resolution The change detection matrix showing the regional Image i.e. Google Earth. Some representative change in hectares between 2002 and 2014 are given in Table 2.

Table 2: Change detection matrix showing the change in the areas of the classes (in ha) between 2002 and 2014 Plantation 2002\20014 Water Sand and gravel Others Total potential Water 628.84 347.99 495.60 1,114.60 2,587.42 Sand and Gravel 685.82 949.10 1,227.30 1,981.60 4,843.81 Plantation 42.79 90.60 229.29 952.65 1,315.34 Potential Others 298.01 203.84 346.66 51,685.09 52,533.60 Total 1,655.45 1,591.53 2,299.25 55,733.94 61,279.78

56 Acharya et al. Banko Janakari, Vol. 26, No. 1 In this study, the land reclamation area referred This study was confined to identification of land to the area previously occupied by ‘Water’ and for plantation on the Bagmati river-basin in the ‘Sand and gravel’ classes in 2002 and later Terai region of Nepal. Only the land reclamation converted into ‘Plantation potential’ and ‘Others’ study during the aforementioned period (2002– classes in 2014. Altogether, 4,819.1 ha land was 2014) is not enough to assess the actual area reclaimed within the study area over the last 12 suitable for plantation, the existing plantation years (2002–2014), out of which 2,708.8 ha and potential area should also be detected. In this 2,110.3 ha lands were reclaimed in Rautahat study, the potential plantation area referred to the and Sarlahi districts. A total of 1,114.6 ha and summation of the area of the reclaimed land, the 495.6 ha areas under ‘Water’ were converted area of ‘Others’ class converted into ‘Plantation into ‘Others’ and ‘Plantation potential’ classes, potential’ class and the area that remained as respectively (Table 2). Similarly, 1,981.6 ha and ‘Plantation potential’ on the basin of the Bagmati 1,227.3 ha ‘sand and gravel’ areas were converted River and its tributaries between 2002 and 2014. into ‘Others’ and ‘Plantation potential’ classes, Over the period of 12 years, a total of 5,395 ha respectively. The map of land reclamation area is land was found to be potential for plantation presented in Fig. 4. (Table 3). Out of this, 3,054.12 ha area was found

Fig 4: Map of the land reclamation area (2002–2014)

57 Banko Janakari, Vol. 26, No. 1 Acharya et al. to be potential for plantation in Rautahat district in the selection of suitable species for plantation whereas 2,340.93 ha was found to be potential for in such land reclamation areas. the same in Sarlahi District. Accuracy of the classification results Table 3: Potential plantation areas in the two districts The overall accuracy in the case of the Classified Landsat 2014 Image was found to be 97.9% Potential plantation with a kappa coefficient of 0.89 (Table 4). In the area District River classification, the producer’s accuracy ranged Total area Area (ha) from 75% in the case of ‘Sand and gravel’ class to (ha) 100% in the case of ‘Water’ and ‘Others’ classes Rautahat Bagmati 2,383.51 whereas the user’s accuracy was just the opposite, Paurahi 572.46 3,054.12 ranging from 83.3% in the case of ‘Water’ class to 100% in the case of ‘Sand and gravel’ class Chadi 98.15 (Table 4). Sarlahi Bagmati 2,340.93 2,340.93 Total area 5,395.05 5,395.05

Table 4: Accuracy assessment of the Classified Landsat 2014 Image Ground Truth User's Error of Classes Sand and Plantation Water Others Total accuracy commis- gravel potential (%) sion (%) Water 5 1 6 83.33 16.67 Sand and gravel 6 6 100.00 0.00 Plantation potential 1 10 11 90.91 9.09 Others 3 217 220 98.64 1.36 Total 5 8 13 217 243 Producer's accuracy (%) 100.00 75.00 76.92 100.00 Error of omission (%) 0.00 25.00 23.08 0.00 Overall accuracy (%) 97.94 Kappa Coefficient 0.8904

Conclusion Suitable species for plantation in the land reclamation area This study indicates that the rivers in the Terai region change their courses over time periods, Focus-group discussions were organized among resulting in significant land reclamation. Over the local people for identification of suitable the last 12 years (2002–2014), a total of 4,819.1 species for the purpose of plantation in the potential ha land area was found to be reclaimed on the land reclamation areas. Most of the participants Bagmati river-basin within the study area. Such of the focus group discussions preferred Khayer area could be planted with the suitable tree species Acacia catechu), Sissoo (Dalbergia sissoo), along with the possible cash crops such as water Babul (Acacia nilotica), Gutel (Trewia nudiflora) melon, pea-nut and gourd. This will, no doubt, together with Eucalyptus and Bambusa spp. for fulfill the local demand for fuel-wood and fodder plantations in such land reclamation areas. The on the one hand and provide income to the local identified potential land reclamation areas were people to some extent. However, further research either public- or private-owned. In the private should be done to determine suitable species and lands, the participants desired to adopt agro- appropriate agro-forestry model before executing forestry model by growing watermelon, pea- plantations over large land reclamation areas. nut and gourd together with trees. However, the Nevertheless, the proposed method is promising, ecological characteristics of tree species and the and can be replicated to other river-basins too for need of the local people should also be considered

58 Acharya et al. Banko Janakari, Vol. 26, No. 1 mapping the potential reclamation areas. These Paudel, A. 2001. Environmental Management results can be further improved by using high- of the Bagmati River Basin. In UNEP EIA resolution satellite images together with sufficient Training Resource Manual: Case studies field validation. from developing countries, 269–279. References Su, W., Zhanx, C., Zhu, X. and LI, D. 2009. A hierarchical object oriented method for land Addink, E. A., de Jong, S. M. and Pebesma, E. cover classification of SPOT 5 Imagery. J. 2007. The importance of scale in object- Wseas Transactions on Information Science based mapping of vegetation parameters with and Applications 6 (3): 437–446. hyperspectral imagery. Photogrammetric Xian, G., Homer, C. and Fry, J., 2009. Updating Engineering and Remote Sensing 72 (8): the 2001 National Land Cover Database land 905–912. cover classification to 2006 by using Landsat Adeniyi, P. O. and Omojola, A. 1999. Landuse/ imagery change detection methods. Remote landcover change evaluation in Sokoto-Rima Sensing of Environment 113 (6): 1133–1147. basin of north-western Nigeria on archival Xu, H. 2006. Modification of normalised remote sensing and GIS techniques. Journal difference water index (NDWI) to enhance of African Association of Remote Sensing of open water features in remotely sensed the Environment 1: 142–146. imagery. International Journal of Remote Adhikari, B. R. 2013. Flooding and inundation Sensing 27 (14): 3025–3033. in Nepal Terai: issues and concerns. Hydro Yoon, G. W., Cho, S. I., Chae, G. J. and Park, J. Nepal 12: 59–65. H. 2004. Automatic land-cover classification Dorren, L. K. A., Maier, B. and Seijmonsbergen, of Landsat images using feature database A. C. 2003. Improved Landsat-based forest in a network. International Archives of mapping in steep mountainous terrain using Photogrammetry Remote Sensing and Spatial object-based classification. Forest Ecology Information Sciences 35 (2): 564–568. and Management 183: 31–46. Yu, Q., Gong, P., Clinton, N., Biging, G., Kelly, Mathieu, R., Freeman, C. and Aryal, J. 2007. M. and Shirokauer, D. 2006. Object-based Mapping private gardens in urban areas using detailed vegetation classification with object oriented techniques and very high- airborne high spatial resolution remote sensing resolution satellite imagery. Landscape and imagery. Photogrammetric Engineering and Urban Planning 81: 179–192. Remote Sensing 72 (7): 799–811. Mcfeeters, S. K. 1996. The use of normalized difference water index (NDWI) in the delineation of open water features. International Journal of Remote Sensing 17 (7): 1425–1432. Meinel, G. and Neubert, M. 2004. A comparison of segmentation programs for high-resolution remote sensing data. International Archives of Photogrammetry and Remote Sensing 35 (B): 1097–1105.

59 Estimating tiger and its prey abundance in Bardia National Park, Nepal

J. B. Karki1, Y. V. Jhala2, B. Pandav2, S. R. Jnawali3, R. Shrestha4, K. Thapa5, G. Thapa5, N. M. B. Pradhan6, B. R. Lamichane7 and S. M. Barber-Meyer8 We estimated tiger and wild prey abundance in the Bardia National Park of Nepal. Tiger abundance was estimated from camera trap mark recapture in 85 days between December, 2008 to March, 2009 by placing 50 camera trap pairs in 197 trap locations with a sampling effort of 2,944 trap nights. We photo captured 16 individuals (≥1.5 year old) tigers identified on the basis of their unique stripe patterns. The number and density (per 100 km2) of tiger was 19 (SE 3.3) and 1.31 (SE 0.32), respectively. Distance sampling was used to assess the prey abundance on 170 systematically laid line transects between May–June, 2009. The density of all the wild prey (individuals/km2) was 56.3 (SE 6.5). The density (individuals /km2) of Chital was 29.3 (SE 4.3). The density of barking deer, wild pig and sambar were in higher to medium, medium and medium to low range as compared to other protected areas in South Asia respectively. The study indicated decline of tiger in Bardia National Park even though the existing level of the prey population appears to be adequate to support higher tiger numbers. There is hope of meeting the ambitious goal of doubling the tiger population by 2022 set by the Tiger Range Countries which was evident in 2014 with 50 tigers in Bardia National Park and Khata Corridor. The tiger habitats outside the protected areas should be managed with the local community- based initiatives to ensure the acceptance of low density tiger movement.

o Bardia, camera trap, density, line transect, tiger, wild prey erai Arc Landscape (TAL) Nepal encompasses Subcontinent (Mondol et al., 2009) and probably Tan area of 23,199 km2, covering 14 Terai by the same percentage through the rest of the districts from Rautahat in the east to Kanchanpur tiger’s range (Seidensticker, 2010). in the west, and consists of over 75% of the remaining forests of the Terai and the foot hills The current global tiger population is comprised of Churia. The protected areas (PAs) are part of of <5% of what was estimated just a century ago the global tiger conservation landscape and are (Dinnerstein et al., 2007) with the current adult source to maintain the wildlife. The corridor and number estimated to be mean 3643, distributed in connectivity within and between the countries are Bangladesh 440, Bhutan 75 (67–81), Cambodia vital for the long-term maintenance of wildlife. 10–30, China 45 (40–50), India 1,411 (1,165– Thus, the regular monitoring of the forest resources 1,657), Indonesia 325 (250–400), Lao PDR 17 and wildlife is important for the management of (9–23), Malaysia 500, Myanmar 85, Nepal 155 the wildlife. The Bardia National Park (BNP) (124–229), Russia 360 (330–390),Thailand 200 has been listed as category II tiger conservation and Vietnam 10s (estimated) (GTIS, 2011). landscape in global tiger conservation scenario Historically, tigers were distributed continuously (Dinerstein et al., 2007). across the lowland Himalayan forests in Nepal but Over the past 200 years, wild tiger populations the surveys, between 1987 and 1997, documented have declined by more than 98% in the Indian only three isolated tiger populations (Smith et al., 1998); BNP being one.

1 Kathmandu Forestry College, Kathmandu, Nepal. E-mail: [email protected] 2 Wildlife Institute of India, Dehradun, India 3 WWF Nepal Hariyoban Program 4 WWF Canada 5 WWF Nepal 6 Bird Conservation Nepal 7 National Trust for Nature Conservation Nepal 8 US Geological Survey, ELY MN 55731, USA 60 Karki et al. Banko Janakari, Vol. 26, No. 1 In Nepal, the oldest population estimates of tiger Materials and methods come from Chitwan National Park (CNP). The estimates until the mid 1990’s were mainly based Study area on either radio-telemetry (Sunquist, 1981; Smith, 1993; Smith et al., 1999) or pugmark surveys The study was conducted in the BNP situated in the (McDougal, 1999). Although they provide a Terai plains and the Siwaliks of the Mid-Western minimum estimate, these methods face the issues Nepal. Established in 1969 and extending over of incomplete spatial sampling of the area of an area of 968 km², the Park is located between interest and incomplete detection of animals even 28°15’ N and 28°35.5’ N latitude and between within the area that is sampled. Thus, population 80°10’ E and 81°45’ E longitude. The terrain of sampling approaches that explicitly deal with the Park ranges from 152 m to 1,440 m from the these two problems by employing appropriate mean sea level. Most of the Park area is occupied statistical models are essential for robust by the lowland flood plains and the inner valley; estimation of animal abundance (Seber, 1982; about 80% of the Park area is covered by forests. Williams et al., 2002; Thompson, 2004). This Field methods study uses the spatially explicit capture-recapture likelihood approach. Camera trap survey (Karanth and Nichols, 1998 and 2002; DNPWC, 2005 and 2008; Dhakal et Chital (Axis axis), sambar (Cervus unicolor), al., 2014) was conducted in 20 blocks of 50–100 swamp deer (Cervus duvauceli duvauceli), wild km2; altogether, 197 traps were located (Fig. 1) at pig (Sus scrofa), hog deer (Axis porcinus) and different points covering a total area of 1,456 km2 barking deer (Muntiacus muntjak) are major prey (½ Mean Maximum Distance Moved: MMDM species of tiger in the BNP. The quantification area) in 2,944 trap nights. Each block was camera of these prey species is of utmost importance in trapped for 15 days between December, 2008 these PAs that are supporting different carnivores and March, 2009 by employing Stealth Cam and species including tiger, leopard (Panthera pardus) Moultrie passive camera traps placed around and wild dog (Cuon alpinus). 16:00 hours and removed after 09:00 hours to In this paper, we have described the use of avoid theft. camera-trap mark-recapture method to obtain the Camera traps were rotated between blocks to abundance estimate of tigers, and line transects to cover the entire area. At each site, paired cameras obtain the density of the tiger wild prey. were deployed using 15-day sampling period in

Fig. 1: Map showing the line transects and the location of camera traps within the study area (BNP)

61 Banko Janakari, Vol. 26, No. 1 Karki et al. each of the camera locations. The trap distance the issue of geographical closure using tiger between the two trapping stations was 1.5 km. habitat. Care was taken not to leave any potential gaps in the sampling area of interest. Tiger wild-prey was first analyzed as one group for the whole Park along a total of 559.2 km distance Transects were laid out systematically using within the three distinct strata viz. i) the Karnali DISTANCE Software (Thomas et al., 2009) with flood plains (KFP, along 211.9 km distance), ii) the random start option for tiger’s wild prey. We the Foot hills (FH, along 273.1 km distance) and determined minimum two temporal replicates iii) the Babai Valley (BV, along 74.2 km distance) and 170 spatial replicates (127 in the Karnali and followed by species having more than 40 flood plains and the Churia foot hills and 43 in observations afterwards. For selecting the best the Babai Valley) (Fig. 1). Computer-generated model (or models) to use for generating density transect points were laid on the map and uploaded estimates, model robustness, relative Akaike on the GPS. Information Criterion (AIC) values, various goodness of fit tests, relative estimate precision During May–June after the burning heat, two and the detection function shape (wide shoulder observers, on elephant-back, moved between near the y axis) were considered. The more robust 06:00–09:00 hours and 16:00–19:00 hours when group approach (Buckland et al., 2001) prior to the prey-animals were most active along the line analyses was performed in case of spiked data. transect recording all the prey species, the number of individual animals, the radial sighting distance Results and discussion to the animal (or the centre of the animal cluster) and the sighting angle between the transect line Tiger abundance and the animal or the centre of the cluster of the animals observed (Buckland et al., 2001). In the 197 trap locations throughout the BNP, 16 individual tigers (5 male, 8 female and 3 gender Data analysis unknown) were identified. About 70% of the total individual tigers were recaptured more than Photographic capture-recapture analysis once with a mean maximum distance between the (Karanth and Nichols, 1998; Pollock et al., 1990) two capture events of 8.9 km (SD 10). No new was undertaken to estimate tiger population tigers were trapped after the 10th night (pooled parameters. Capture histories (X matrices) across blocks), while the total number of captures were developed on individual tigers identified increased steadily until the 14th night (pooled on the basis of the stripe pattern on the body across blocks). Nearly 65% of the total individual flanks, legs and face (Karanth, 1995; McDougal, tiger captures were made during the first 5 days of 1977; Schaller, 1967; DNPWC, 2005). Data camera trapping. were analyzed using the CAPTURE 2 Interface Program (Otis et al., 1978; Rexstad and Burnham, The estimated tiger number was 19 (SE 3.3 with 1991; White et al., 1982) for estimation of the range 17.2–36) from the model Mh-Jackknife number. from CAPTURE. The density was 1.31 (SE 0.32) and 0.87 (SE 0.28) tigers/100 km2 from ½ MMDM We used spatial density analysis (Maximum and MMDM of Program DENSITY (Table 1). The Likelihood Spatially Explicit Capture Recapture, density from Spatial Explicit Capture-recapture DENSITY Software (Efford, 2009) to overcome of Maximum likelihood provided 0.61(SE 0.15) Table 1: Number and density in tigers in Bardia National Park, Nepal Camera Pop. D (SE) from ½ D (SE) D(SE)ML SECR Best Model Trap Estimate MMDM ETA (MMDM) CAPTURE Score nights N (SE) (Km2) ETA (Km2) All area Mask Mh-Jacknife 0.98 2,944 19 (3.3) 1.31 (0.32) 0.87 (0.28) 0.61 (0.15) 0.94 (0.23) 1,456 2,182 Note: Pop. Estimate N (SE) = Population Estimate Number (Standard Error); D (SE) MMDM ETA = Density (Standard Error) Mean Maximum Distance Moved Effective Trapping Area; ML SECR = Maximum Likelihood Spatially Explicit Capture Recapture.

62 Karki et al. Banko Janakari, Vol. 26, No. 1 ranging from 0.37–0.99. Habitat mask was used 2010; Karki et al., 2008; Chundawat et al., 2011). and density was estimated at 0.94 (SE 0.23) 95% The current Government’s effort to reinforce CI 0.58–1.52 with the area of 1,896 km2 from the the protection of the Park and trans-boundary DENSITY Program. initiative is very positive and the Government’s commitment to make the 2010 tiger population For density analysis, the likelihood approach double by year 2022 could be achieved provided (Efford et al., 2004) seems to be appropriate these areas are supplemented with additional being comparatively less sensitive to buffer width prey species particularly in the Babai Valley of as it is directly based on parameters estimating the Park. The doubling of the tiger population by density unlike Bayesian and is faster, and both 2022 (T x 2) is possible from the population of spatial methods (Royle et al., 2009) have not 37 adult breeding tigers (15 male and 22 female) shown any significant difference in terms of in the BNP (BNP, 2012) and 50 (45–55) in BNP density estimation (Kalle et al., 2011). Thus, and Khata Corridor (forest) (Dhakal et al., 2014). the likelihood approach was interpreted for Dhakal et al. (2014) found the density of tiger to discussion. be 3.38 /100 km2, which was quite higher than the one found in 2009 (0.9/100 km2) in the BNP. During 1990, there were 28 tigers estimated based on 1994–96 (Basnet et al., 1998). The drastic The improvement of tiger prey density as well decline in the tiger population in the BNP was as the tiger population in the BNP indicates the consistent till this study period from 42 (Bhatta et success of control of poaching and illegal wildlife al., 2002) to 18 in 2010/011 (Table 2). trade including the control of poaching and illegal wildlife trade (Dhakal et al., 2014). This study strongly shows that the decline is not due to prey loss as compared to the dense tiger The candidate species for introduction in the Babai bearing PAs (Table 3) but may be due to poaching Valley are wild water buffalo and swamp deer to of tiger and its prey (Check, 2006; Gopal et al., supplement the prey-animals and rhino to build

Table 2: Population of tiger in BNP Nepal during the period of 1998–2013 Density/100 km2 Number Year 19872 1998/993 2000/013 19981/(area km2) 1999/2000 2005 2009 2010 2011 2013 BNP 2.7 2.08 2.18 25/50 32–40 32–40 18 18 37 50 1Smith et al., 1998; 2Smith et al., 1987; 3Wegge et al., 2009

Table 3: Number and densities of tiger and their wild prey in the PAs of Nepal and India Tiger number Tiger density (D)/100 Name of PA D_Prey/Km2 (SE) km2 (SE) ML SECR Corbett TR* 109 (5.4) 16.23 (1.63) 72.4 Ramnagar FD* 27 (1.5) 13.8 (2.74) 72.4 Kaziranga NP* 69 (0.5) 12.63 (1.5) 56.1 Kishanpur (Dudwa, TR)* 19 (7.31) 4.64 (1.11) 25 Katerniaghat (Dudwa, TR)* 20 (2.61) 4.82 (1.19 25 Dudhwa NP (Dudwa, TR)* 21 (5.47) 4.79 (1.28) 25 Pilibhit FD* 12 (0.17) 3.78 (1.17) 25 Chitwan NP (Karki et al., 2013) 126 (21) 2.30 (0.31) 51.7 Valmiki TR* 8 (2.1) 1.12 (0.52) BNP, 2009 PS 19 0.9 (0.23) 56.3 Suklaphanta WR, 2009 PS (Karki et al., 2015) 7 2.1 (0.8) 144.8 Parsa WR-PS (Karki, 2011) 4 0.61 (0.32) 6.6 Jhala et al., 2011 Note: TR = Tiger Reserve; FD = Forest Division; NP = National Park; WR = Wildlife Reserve and PS = Present Study

63 Banko Janakari, Vol. 26, No. 1 Karki et al. viable population in the BNP. A regular habitat 6.5); 50.5 (SE 8.4); 21.8 (SE8.4) and 19.2 (SE management for ungulates by cutting grass in the 5.2) for the entire BNP, the Karnali flood plains, early winter (Karki, 1997; Peet, 1997), and control the Churia foot hills and the Babai Valley of the burning to regulate succession in the relocated BNP, respectively (Table 4). The Half Normal villages are essential. The current prey abundance Model was found to best fit for the data of the in the BNP can support about 100 tigers assuming BNP, the Karnali flood plains and the Babai the removal of the current abundance of 10% per Valley while the Uniform Cosine was best for year (the annual removal of 50 ungulates/yr/tiger the data of Churia foot hills. The average density ranging from rhesus to wild elephant in size). estimates (chital/km2) based on the Model were The past highest tiger abundance did not cover 29.3 (SE 4.3), 50.5 (SE 8.4), 21.8 (SE 8.4) and the Babai Valley for estimation. It links with 19.2 (SE 5.2) for the whole BNP, the Karnali Katerniaghat (India) via. Khata Corridor, and flood plains, the Churia foot hills and the Babai with the Suhelwa Wildlife Sanctuary (India) via. Valley, respectively. Besides, the Half Normal Banke National Park which further supports tiger Model was also found to be best fit for the data in this landscape. The number of tiger was found for chital. to have increased in BNP (DNPWC, 2013; Dhakal 2 et al., 2014); the new tigers probably entering The density estimates (number/km ) for sambar, from the south eastern part. One tigress was wild pig, barking deer, langur, rhesus macaque found to have regularly used the Khata Corridor and barking deer and hog deer combined were while another tiger was found to have routinely 3.07 (SE 0.7), 2.4 (SE 0.6),1.4 (SE 0.3), 9.2 (SE visited the Corridor from the flood plains of the 2.3), 10.6 (SE 2.8) and 2.3 (SE 0.58) barking deer BNP (BNP, 2012). and hog deer, respectively based on the global detection function and cluster size (Table 4). We Ungulate estimates could not have sufficient data for the swamp deer points for the BNP due to their narrow distribution The density estimates for the wild-prey of tiger range in the Karnali flood plains. (individuals/km2) in the year 2009 were 56.3 (SE

Table 4: Density of tiger’s prey species (individuals/km2) in the Karnali Flood plains, Churia foot hills and Babai Valley of the BNP, Nepal Encounter Cut Species ESW Cluster size DS (±SE)/ D (±SE)/ Total Species rate (±SE/ point L, model (SE) (±SE) km2 km2 effort km) R (m) BNP_T. Half N. 43.6 (1.4) 5.6 (0.3) 12.9 (1.3) 56.3 (6.5) 1.1 (0.1) 559.16 88, 0.2 KFP Half N. 47.0 (1.9) 6.0 (0.3) 21.0 (2.2) 103.7 2.0 (0.2) 211.93 85, 0.2 (12.8) Foot Hills Unif. Cos. 37.6 (2.7) 7.4 (1.1) 3.7 (0.8) 22.2 (6.5) 0.3 (0.06) 273.08 65.3 B. Valley Half N. 47.7 (3.5) 2.4 (0.2) 16.8 (2.7) 37.3 (6.6) 1.6 (0.2) 74.16 100 Chital Half N. 49.4 (2.4 ) 7.0 (0.4) 5.4 (0.7) 29.3 (4.3) 0.22 (0.03) 559.16 91 Ch-KFP Half N. 50.0 (3.0) 6.5 (0.5) 9.7 (1.4) 50.5 (8.4) 1.0 (0.1) 211.93 87 Ch-FH Unif. Cos. 39.5 (6.1) 10.5 (1.6) 2.2 (0.7) 21.8 (8.4) 0.2 (0.05) 273.08 80 Ch-BV Half N. 42.0 (5.4) 3.2 (0.5) 6.6 (1.5) 19.2 (5.2) 0.6 (0.1) 74.16 82 Sambar Half N. 41.7 (4.9) 2.3 (0.2) 1.3 (0.3) 3.0 (0.7) 0.1 (0.02) 559.16 67 Wild pig Half N. Cos. 40.6 (6.8) 2.2 (0.3) 1.0 (0.2) 2.4 (0.6) 0.08 (0.02) 559.16 98.1 Bk. Deer Unif. Cos. 31.8 (3.4) 1.2 (0.07) 1.1(0.2) 1.4 (0.3) 0.07 (0.01) 559.16 54 Langur Unif. S.P. 53.6 (3.2) 7.6 (0.8) 1.4 (0.3) 9.2 (2.3) 0.1 (0.02) 559.16 77.7 R. macaque Unif., Cos. 44.8 (3.6) 8.5 (1.0) 1.2 (0.2) 10.6 (2.8) 0.1 (0.02) 569.12 77.7 B + H. deer Half N. 32.8 (3.5) 1.5 (0.1) 1.6 (0.3) 2.3 (0.5) 0.1 (0.02) 569.16 65.4 Note: BNP_T. = Total of the BNP; KFP = Karnali Flood Plains; B. Valley = Babai Valley; Ch-KFP = Chital Karnali Flood Plains; Ch-FH = Chital Foot Hills; Ch-BV = Chital Babai Valley; Bk. deer = Barking deer; R. macaque = Rhesus macaque; B + H. deer = Barking and Hog deer combined; N. = Normal; Unif. = Uniform; Cos.=Cosine; S.P. = Simple polynomials ; ESW = Effective Strip Width, SE = Standard Error; DS = Group Density, D = Density; L = Left; R = Right

64 Karki et al. Banko Janakari, Vol. 26, No. 1

The overall density of the Park had increased from (10.7) was found to be slightly higher than the 56.3 animals/km2 in 2008 to 92.6 animals/km2 in one (16.7±6.6) found out by Malla (2009). 2013 (Dhakal et al., 2014), which is in the higher range as compared to those (5.3–107 animals/ In both the Karnali flood plains and the Babai km2) in some PAs of the South Asia region. The Valley of the BNP, the wild prey density was improvement was found to have been contributed comparable with that of Malla (2009), but that mainly by Chital. of chital was slightly lower in the Karnali flood plains. However, this study had covered large Nominal decrease in the densities of barking area in the Karnali flood plains as compared to deer, sambar and wild pig was because of the the studies conducted by Wegge et al. (2009) larger area covered in current study compared to and Malla (2009). The density (individuals/km2) the earlier studies conducted in the Karnali flood of Chital (29.3) was similar to the one found by plains (Wegge and Storaas, 2009 Wegge et al., Malla (2009) in the Babai Valley. However, the 2009; Dinerstein,1980). The combined density density of Chital in the entire BNP was found (19.9/km2) of langur (9.2) and rhesus macaque to in the moderate range as compared to the one

Table 5: Density of prey-animals (individuals/km2) in the PAs of South Asia D_Prey Density of Barking PA/Prey habitat D±SE Sambar Wild pig Tot Chital deer Bardia NP, 2009 PS 56.3 56.3 (6.5) 29.3 (4.3) 3 (0.7) 2.4 (0.6) 1.4 (0.3) Bardia NP, 2014 (Dhakal et al., 2014) 4.45 1.97 92.6 92.6 (8.8) 53.99 (10.3) 4.8 (0.5) (0.8) (0.5) Babai, BNP, 2009 (Malla, 2009) 1.2 2.5 Karnali, BNP, 2009 (Malla, 2009) 50.5 3.1 3.1 Karnali, BNP, 1976 (Dinerstein, 1980 ) 33.9 3.5 4.2 1.7 Karnali, BNP, 1993 (Wegge et al., 2009) 1 2.6 Chilla Range, Rajaji NP, 2005 and 2006 76.2 6.5 ± 4.1 43.5 19.6 (Harihar et al., 2006) Chitwan NP, 1982 (Tamang, 1982) 16.8 2.7 6.6 Chitwan NP, 2008, 2009, 2010 (Thapa, 113.8 113.8 86.3 8 10.5 4.1 2011) Chitwan NP, 2009 (Karki, 2011) 51.7 52.88 ±4.9 32.3 3.5 3.4 2.1 Dudwa,Valmiki, Pilibhit, Katerniaghat 24.92 13 0.14 1.99 0.72 24.92 2010 (Jhala et al., 2011) (3.75) (2.17) (0.02) (0.55) (0.23) Gir LS, 1997 (Khan and Vohra,1997) 50.8 2 2.1 Kanha NP, 1987 (Newton, 1987) 55.5 57.3 ± 4.07 3.2 0.9 0.5 0.4 Kaziranga (Jhala et al. 2011) 56.1 58.1 ± 6.51 Melghat (Jhala et al. 2011) 5.3 5.3 ± 0.76 Nagarhole, 2092 (Karanth and Sunquist, 52.9 56.1 ± 3.95 50.6 5.5 4.2 4.2 1992) Parsa WR, 2009 (Karki, 2011) 6.6 6.6 ±1.1 Pench TR, 2000 (Karanth and Nichols, 51.3 9.6 2000) Rajaji-Corbett, 2010 (Jhala et al., 2011) 46.71 7.49 72.4 72.4 ±13.0 (13.25) (1.78) Ranthmbore, Sariska, 2010 (Jhala et al., 107.7 31.62 8.24 107.7 4.86 (7.7) 2011) ±10.0 (10.4) (1.8) Suklaphanta WR, 2009 (Karki, 2011) 144.8 144.8 79 ±22.8 Melghat, Pench, Tadoba, 2010 (Jhala et 107.74 37 5.34 5.83 0.61 107.74 al., 2011) (9.95) (6.06) (0.57) (1.11) (0.15)

65 Banko Janakari, Vol. 26, No. 1 Karki et al.

(5.3–107) in the Indian PAs (Table 5), indicating Kathmandu, Nepal. the availability of adequate prey-animals of tiger in the Park. On the other hand, the density of Bhatta, S. R., Bajimaya, S. and Jnawali, S. R. sambar (3) in the Park was in the lower range as 2002. Status, Distribution and Monitoring compared to the one (0.14–19.6) in the Indian of Tigers in Protected areas of Terai PAs. Similarly, the density of wild pig (2.4) in Arc Landscape-Nepal. A photographic the Park was in the medium range as compared Documentation of camera-trapped tigers. to that (0.5–5.8) in the Indian PAs. Likewise, the WWF Nepal, NTNC and DNPWC, density of barking deer (1.4) in the Park was in Kathmandu, Nepal. the medium range as compared to that (0.4–4.2) BNP. 2012. Tiger and Prey Base Monitoring in in the Indian PAs. Bardia National Park and Khata Corridor, The density of swamp deer could not be Nepal. Final Progress Report submitted determined due to the limited data points, but to DNPWC, Bardia National Park (BNP), significantly preferred (Hayward et al., 2012; Nepal. Wegge et al., 2009) by tiger owing to large- Buckland, S. T., Anderson, D. R., Burnham, K. P., bodied wild prey-animals (sambar and nilgai) in Laake, J. L., Borchers, D. L. and Thomas, L. the Karnali flood plains of the BNP. 2001. Introduction to Distance Sampling: Conclusion Estimating Abundance of Biological Populations. Oxford University Press, The population of tiger in the BNP was found to London, UK. have increased from 18 in 2009 to 50 in 2013. The reason behind is the habitat management of Check, E. 2006. The tiger’s retreat. Nature 441: wild prey-animals of tiger in the Park and control 927–930. in poaching and illegal trade of wild animals Chundawat, R. S., Habib, B., Karanth, U., from the Park. Therefore, habitat management Kawanishi, K., Ahmad Khan, J., Lynam, T., of wild prey-animals of tiger together with the Miquelle, D., Nyhus, P., Sunarto, S., Tilson, control in poaching of wild animals and illegal R. and Wang, S. 2011. Panthera tigris. In wildlife trade will result in further increase in the IUCN Red List of Threatened Species, Version population of tiger. 2011.2. www.iucnredlist.org accessed on 2 In order to support the tiger doubling aim of March, 2012. Nepal by the end of 2020, it is recommended Dinerstein, E. 1980. An ecological survey of to improve the prey-base. One of the ways to the Royal Karnali-Bardia Wildlife Reserve, improve the prey-base in the BNP is to increase Nepal: Part III: ungulate populations. the number of swamp deer in the Babai Valley, Biological Conservation 18 (1): 5–37. study the feasibility of introducing gaur and wild water buffalo in BNP. Dinerstein, E., Loucks, C., Wikramanayake, E., Ginsberg, J., Sanderson, E., Seidensticker, Acknowledgements J., Forrest, J., Bryja, G., Heydlauff, A., Klenzendorf, S., Leimgruber, P., Mills, J., We are grateful to Save the Tiger Fund (STF), O’Brien, T. M., Shrestha, M., Simons, R. and US Fish and Wildlife Services (USFWS), World Songer, M. 2007. The fate of wild tigers. Wildlife Fund (WWF) US, WWF-UK, WWF- Biosciences 57 (6): 508–514. International for providing financial support to accomplish this study. We would also like to Dhakal, M., Karki, M., Jnawali, S. R., Subedi, N., thank all those involved in the fieldwork of this Pradhan, N. M. B., Malla, S., Lamichane, B. study. R., Pokhrel, C. P., Thapa, G. J., Oglethorpe, J., Subba, S. A., Bajracharya, P. R., and Yadav, References H. 2014. Status of Tiger and Prey in Nepal. Department of National Parks and Wildlife Basnet, K., Shrestha, K. M., Sigdel, R. and Conservation, Kathmandu, Nepal. Ghimire, P. 1998. Bardia Extension Area: Biodiversity Survey. WWF Nepal Program,

66 Karki et al. Banko Janakari, Vol. 26, No. 1

DNPWC 2005. National Report Status of Tiger Kalle, R., Ramesh, T. Qureshi, Q. and Sankar, in Nepal. Department of National Parks and K. 2011. Density of tiger and leopard in a Wildlife Conservation, Ministry of Forests tropical deciduous forest of Mudumalai Tiger and Soil Conservation, Government of Nepal, Reserve, southern India, as estimated using Kathmandu, Nepal. photographic capture–recapture sampling. Acta Theriologica 56 (4): 335–342. DNPWC. 2008. Tiger Monitoring protocol, Nepal. Department of National Parks Karanth, K. U. and Sunquist, M. E. 1992. and Wildlife Conservation (DNPWC), Population Structure, density and biomass Kathmandu, Nepal. of large herbivore in the tropical forests of Nagarhole, India. Journal of Tropical DNPWC. 2013. Annual Progress Report (FY Ecology 8: 21–35. 2069–70). Department of National Parks and Wildlife Conservation (DNPWC), Karanth, K. U. 1995. Estimating tiger Panthera Kathmandu, Nepal. tigris populations from camera-trap data using capture—recapture models. Biological Efford, M. G., Dwson, D. K. and Robbins, C. Conservation 71: 333–338. S. 2004. DENSITY: software for analyzing capture data from passive detector arrays. Karanth, K. U., and Nichols, J. D. 1998. Estimation Animal Biodiversity and Conservation 27 of tiger densities in India using photographic (1): 217–225. captures and recaptures. Ecology 79: 2852– 2862. Efford, M. 2009. DENSITY 4.4 Spatially explicit capture-recapture 4.4.1.2 4/19/2009. Karanth, K. U. and Nichols, J. D. 2000. Ecological Zoology Department, University of Otago, Status and Conservation of Tigers in India. New Zealand. WCS, US Fish and Wildlife Service, Center for Wildlife Studies, Banglore, India. Gopal, R., Qureshi, Q., Bharadwaj, M., Singh, RKJ., and Jhala, Y. V. 2010. Evaluating the Karanth, K. U., and Nichols, J. D. 2002. status of the endangered tiger Panthera tigris Monitoring Tigers and Their Prey: A and its prey in Panna Tiger Reserve, Madhya Manual for Researchers, Managers and Pradesh, India. Oryx 44: 383–389. Conservationists in Tropical Asia. Centre for Wildlife Studies, Banglore, India. GTIS. 2011. Global Tiger Recovery Program 2010–2011. Global Tiger Secretariat. Karki, J. B. 1997. Effects of Grazing, Utilization Washington DC, USA. and Management on Grasslands of Royal Bardia National Park, Nepal. M.Sc. Harihar, A., Pandav, B. and Goyal, S. P. 2006. Dissertation, Wildlife Institute of India, Monitoring Tiger and its Prey in Chilla Saurastra University, Rajkot, India. Range, Rajaji National Park. Uttaranchal, India. Karki, J. B., Poudel, B. S. and Shrestha, P. B. 2008. Threat from poaching and illegal trade Hayward, M. W., Jedr zejewski, N. and in tiger, its parts and derivatives and actions Jedrzekewska, B. 2012. Prey Preferences of taken/required to combat the same in Nepal. the tiger (Panthera tigris). Journal of Zoology In Biodiversity Coservation Efforts in Nepal: 286: 221–231. a special issue published on the occasion of 12th Wildlife Week (eds) Manandhar, L. P. Jhala, Y. V., Qureshi, Q., Gopal, R. and Sinha, P. Department of National Parks and Wildlife R. 2011. Status of the Tigers, Co-predators Conservation, Kathmandu, Nepal. and Prey in India, 2010. National Tiger Conservation Authority, Govt. of India, Karki, J. B., Pandav, B., Jnawali, S. R., Shrestha, New Delhi and Wildlife Institute of India, R., Pradhan, N. M. B., Lamichane, B. R., Dehradun, India. Khanal, P., Subedi, N. and Jhala, Y. V. 2013. The largest tiger population in Nepal:

67 Banko Janakari, Vol. 26, No. 1 Karki et al.

estimating tiger from flood plains and from capture data on closed animal Churiya, Chitwan National Park. Oryx 49 populations. Wildlife Monograph 62: 1–135. (1): 15–156. Pollock, K. H., Nichols, J. D., Brownie, C and Karki, J. B., Barber-Meyer, S. M., Jhala, Y. V., Hines, J. E. 1990. Statistical inference for Pandav, B., Jnawali, S. R., Shrestha, R., capture-recapture experiments. Wildlife Thapa, K., Thapa, G., Pradhan, N. M. B., Monograph 62: 135. Lamichane, B. R. and Dhakal, M. 2015. Estimating the abundance of tigers and their Peet, N. B. 1997. Biodiversity and management prey in Suklaphanta Wildlife Reserve of of tall grasslands in Nepal. Ph.D. Thesis, Terai Arc Landscape, Nepal. In Biodiversity University of East Anglia, Norwich, UK. Conservation Efforts in Nepal: a special Rexstad, E., and Burnham, K. P. 1991. User’s issue published on the occasion of 20th Guide for Interactive Program Capture. Wildlife Week (eds) Dhakal, M. and Shrestha, Abundance Estimation of Closed Animal R. Department of National Parks and Wildlife Populations. Colorado State University, Conservation, Kathmandu, Nepal, 41–56. Colorado, USA. Karki, J. B. 2011. Occupancy and abundance Royle, J. A., Karanth, K. U., Gopalaswamy, A. M. of tigers and their prey in the Terai Arc and Kumar, N. S. 2009. Baysian inference in Landscape, Nepal. PhD. Dissertation, FRIU, camera trapping studies for a class of spatial Dehradun, Uttarakhand, India. capture-recapture models. Ecology 90 (11): Khan, J. A. and Vohra, U. 1997. Estimation of 3233–3244. ungulate densities by line transect method Seidensticker, J. 2010. Saving wild tigers: A case in Gir forest, India. Tropical Ecology 38 (1): study in biodiversity loss and challenges to 65–72. be met for recovery beyond 2010. Integrative McDougal, C. 1977. The Face of the Tiger. Zoology 5: 285–299. doi: 10.1111/j.1749- Rivington Books, London, UK. 4877.2010.00214.x accessed on 18 March, 2016. McDougal, C. 1999. You can tell some tigers by their tracks with confidence. In Riding Schaller, G. 1967. The Deer and the Tiger. the Tiger: Tiger Conservation in Human- University of Chicago Press, Chicago, USA. dominated Landscapes (eds) Seidensticker, Seber, G. A. F. 1982. The Estimation of Animal J., Christie, S. and Jackson, P. Cambridge Abundance and Related Parameters. 2nd University Press, Cambridge, UK. edition. Macmillan, New York, USA. Malla, S. 2009. Estimating the Status and Impact Smith, J. L. D., McDougal, C. and Sunquist, of Hunting on Tiger Prey in Bardia National M. E. 1987. Land tenure system in female Park, Nepal. M.Sc. Thesis, Saurastra tigers. In Tigers of the World: the Biology, University, Gujrat, India. Biopolitics, Management and Conservation Mondol, S., Karanth, K. U. and Ramakrishnan, of an Endangered Species (eds) Tilson, U. 2009. Why the Indian subcontinent holds R.L. and Seal, U.S. Noyes Publication, Park the key to global tiger recovery? PLoS Genet Ridge, New Jersey, USA, 464–474. 5 (8): doi:10.1371/journal.pgen.1000585 Smith, J. L. D. 1993. The role of dispersal in accessed on 18 March, 2016. structuring the Chitwan tiger population. Newton, P. 1987. The social organization of Behavior 124: 165–195. forest hanuman langur (Presbytes entellus). Smith, J. L. D., Ahearn, S. C., and McDougal, C. International Journal of Primatology 8: 1998. Landscape analysis of tiger distribution 199–232. and habitat quality in Nepal. Conservation Otis, D. L., Burnham, K. P., White, G. C., and Biology 12 (6): 1338–1346. Anderson, D. R. 1978. Statistical inference

68 Karki et al. Banko Janakari, Vol. 26, No. 1

Smith, J. L. D., McDougal, C., Ahearn, S. C., Joshi, of St. Andrews, UK. http://www.ruwpa.st- A. and Conforti, K. 1999. Metapopulation and.ac.uk/distance/ accessed on 18 March, structure of tigers in Nepal. In Riding the 2018. Tiger: Tiger Conservation in Human- dominated Landscapes (eds) Seidensticker, Thompson, W. L. 2004. Sampling Rare or J., Christie, S. and Jackson, P. Cambridge Elusive Species: Concepts and Techniques University Press, Cambridge, UK. for Estimating Population Parameters. Island Press, Washington D.C., USA. Surquist, M. E. 1981. The social organization of tigers (Panthera tigris) in Royal Wegge, P. and Storaas, T. 2009. Sampling tiger Chitwan National Park, Nepal. Smithsonian ungulate prey by the distance method: lessons contribution to Zoology 336: 1–98. learned in Bardia National Park, Nepal. Animal Conservation 12: 78–84. Tamang, K. M. 1982. The Status of Tiger Panthera tigris) and its Impact on Principal Wegge, P., Odden, M., Pokhrel, C. P. and Storaas, Prey Populations in Royal Chitwan National T. 2009. Predator prey relationships and Park, Nepal. PhD Dissertation. Michigan responses of ungulates and their predators State University, Michigan, USA. to the establishments of protected areas: a case study of tigers, leopards and their prey Thapa, T. 2011. Habitat Suitability Evaluation in Bardia National Park, Nepal. Biological for Leopard (Panthera pardus) using Remote Conservation 142 (1): 189–202. Sensing and GIS in and around Chitwan National Park, Nepal. PhD Thesis. Saurashtra White, G. C., Andrson, D. R., Burnham, K. P., University, Rajkot, India. and Otis, D. L. 1982. Capture-recapture and Removal Methods for Sampling Thomas, L., Laake, J. L., Rexstad, E., Strindberg, Closed Populations. National Laboratory, S., Marques, F. F. C., Buckland, S. T., Los Alamos, New Mexico, USA. Borchers, D. L., Anderson, D. R., Burnham, K. P., Burt, M. L., Hedley, S. L., Pollard, J. Williams, B. K., Nichols, J. D. and Conroy, M. J. H., Bishop, J. R. B. and Marques, T. A. 2009. 2002. Analysis and Management of Animal “Distance 6.0. Release “x”. Research Unit for Populations. Academic Press, San Diego, Wildlife Population Assessment, University California, USA.

69 Impact of roads on biodiversity: a case study from Karekhola rural road in of Nepal

J. K. KC1 and A. P. Gautam2 Loss and degradation of biodiversity is continuing despite the past conservation efforts in Nepal. Out of many potential causes, this study strives to investigate the effects of a road project on biodiversity in the Middle Hills of Nepal. Information about floristic composition was collected from the adjoining community forests using group of 30 circular sample plots, each located at 50 m and 20 m far from the edge of the road. Results provide evidence that rural road projects are contributing to reduction of biodiversity which may be due to the removal of low-yielding timber species near the road-edge. The study also suggests that proximity to road-edge reduces under- storey vegetation which will lead less capable forest to sustain its original biodiversity. However, silvicultural operations have potential to minimize the indirect loss of biodiversity caused by road projects.

o Community forests, Middle Hills, rural road, woody plant species’ diversity

iversity of species plays an important role consequences cannot be neglected only foreseeing in ecosystem functions and services. Nepal economy. Unplanned and wrongly designed possesses a disproportionately high diversity of economic development can cause destabilization flora and fauna at genetic, species and ecosystem of the natural environment, which is evidenced levels due to its unique geographic position by many past efforts in Nepal and elsewhere. and altitudinal variations. The Government of Earlier studies (WWF, 2013; MFSC, 2014) have Nepal (GoN) is committed to the protection and found that unplanned rural roads constructed by management of biological resources and their the local governments are one of the major threats diversity on a sustainable basis (MFSC, 2014). to environment and biodiversity. The Department As a signatory country of the Convention on of Roads estimates that around 25,000 kilometers Biological Diversity (CBD), the GoN has revised rural road tracks had been opened in Nepal by the country’s National Biodiversity Strategy and 2010, most of which have been constructed Action Plan (NBSAP) in 2014. In addition, the without any environmental safeguard (DOR, GoN has put a lot of efforts in the past regarding 2010). implementation of International agreements as well as formulation of strategies to include the Forest roads are termed as “ecosystems” as they local communities in biodiversity conservation. occupy ecological space (Hall et al., 1992) and However, the efforts made so far mainly relate provide habitat for associated plants and animals to reducing poaching, trade and illegal activities (Lugo and Gucinski, 2000). Road infrastructure within the protected areas. Other threats like causes direct and indirect loss in forest ecosystem. unplanned road projects, the key causes of Direct loss refers to the reduction of forest area habitat loss and fragmentation, are usually and indirect loss of roads refers to fragmentation underestimated. The problem is not restricted and degradation of the ecosystem (Geneletti, to motorways. However, narrow country roads 2003). One of the major effects of roads relates occupy less area per kilometer, and are more to its edge effects, which can be defined as the frequent than motorways, so their combine effect alternation to habitat quality due to proximity upon the landscape can be considerably larger to edge. It can cause indirect loss of habitat by (Seiler, 2001). changing species composition, temperature, moisture, light availability and wind speed and, Road provides a basis for long-term development therefore, alteration in original biodiversity in the rural areas, but the environmental (Gysel, 1951). The effect of edge on plant

1 Building Climate Resilience of Watersheds in Mountain Eco-regions Project, Dadeldhura, Nepal. E-mail: [email protected] 2 Kathmandu Forestry College, Kathmandu, Nepal 70 KC and Gautam Banko Janakari, Vol. 26, No. 1 diversity can occur up to 30 m far from the road The road is the main pillar in the development of or even beyond (Seiler, 2001). Jarbuta VDC area. All kinds of traffic are being used to transport essential goods to meet livelihood The study was, moreover, a descriptive research requirements of the people of Jarbuta VDC. The limited to investigating the impacts of road road passes through Neware Community Forest projects on diversity of woody plant species. In (CF) in the northern part and Devis than CF in order to assess the impacts of roads on diversity the southern part. The adjoining CFs are managed of woody plant species, the effects mainly upon by the local Community Forest User Groups the species diversity and the structural diversity (CFUGs) according to the respective operational were investigated. These indirect effects of plans. The FUGs have deployed the locally-hired road projects were assessed by comparing and forest guards for protection of the forests. analyzing woody plant species distribution in between two effect-zones (20 m and 50 m far Methodology from road-edge). Systematic sampling was used for the collection Materials and methods of primary data on diversity of woody species in the CFs. First of all, reconnaissance survey was Study area carried out and the Karekhola Road was surveyed with the help of GPS (Global Positioning System) The study area is located between 28°36’14” N Device. Preliminary data analysis was done using and 28°36’20” N latitude and between 81°38’19” GIS (Geographical information System) Software. E and 81°38’58” E longitude in the adjoining The forest area lost due to road construction was forests of the Karekhola Rural Road. The road determined by multiplying the length of the road connects Jarbuta Village Development Committee with its width. Similarly, the spacing between the (VDC) and Birendranagar Municipality of two successive plots (45 m) was determined on Surkhet District situated in the Middle Hills (Fig. the basis of different factors, such as length of the 1). The earthen road is 5 m wide and 1.46 km road (1.46 km), maximum coverage of the road in long. The study was conducted in 2012. the study area and the possibility of intersection of the area of the sample plots on the road bends.

Then from the starting point (on the road), 45 m distance was marked on the road-length with the help of a Measuring Tape. Making perpendicular to the road length, concentric circular sample plots (CCSPs) with 1 m, 3 m and 10 m radii were laid out at 20 m and 50 m distances from both the edges of the road (Fig. 1 and 2) with the help of the Measuring Tape. The sampling protocol developed and used by the International Forestry Resources and Institutions Research Network (IFRIRN) was used for assessing the forest conditions (IFRI, 2013). This research protocol has been widely used by the researchers in the past (Gautam, 2002; Gautam, 2006). In the innermost circle of the plot (1 m radius), all the woody seedlings were identified and counted. In the next circle (3 m radius), all the shrubs, saplings, and climbers were identified and counted, and also the diameters and the heights of the woody stems having diameter at breast height (DBH) in 2.5–10 cm class were recorded. In the largest circle (10 m radius), all the stems Fig. 1: Map showing the location of the study with 10 cm or greater DBH were counted, and area in Western Nepal their diameters and heights measured. The same

71 Banko Janakari, Vol. 26, No. 1 KC and Gautam process was repeated from the opposite edge of Where, the road. Again, at the distance of 45 m from the S1 = Total no. of species found in the CCSP previous location, the same process was repeated. located at 20m distance from the road edge, In this way, altogether 30 circular sample plots S2 = Total no. of species found in the CCSP were laid at 20 m and 50 m distances from the located at 50 m distance from the road edge, and road-edge throughout the study area. C = No. of species common to both the effect- zones.

Structural diversity

In order to compare the structural diversity of the two effect-zones, the DBH and height distribution classes were prepared. The diameters (at breast height) were categorized into the classes of 0–5 cm, 5–10 cm, 10–15 cm, and above 15 cm. Then, the DBH class distributions between the two zones were compared, and analyzed. Similarly, the heights were categorized into the classes of 1–5 m, 5–10 m, and above 10 m, and the heights between the two zones were also compared and analyzed. Fig. 2: A concentric circular sample plot Results and discussion

The woody plant species’ diversity was assessed Plant species found in the two types of effect- and compared using the Simpson’s Diversity zones Index (D) and the Sorenson’s Similarity Index (SSI). Altogether, 23 plant species including 18 tree species, 3 shrub species and 2 woody climber Simpson’s Diversity Index (D): It gives the species were found in the study area. A total of 19 probabilities that the two randomly chosen plant species were found in the effect-zone at 20 individuals drawn from a population belong to the m distance while a total of 16 plant species were same species. Higher the probabilities that both found in the effect-zone at 50 m distance. Seven the individuals belong to the same species, lower tree species, three shrub species and two woody the diversity. For finite communities (where all climbers were common in both the effect-zones members have been counted), (Table 1). The difference in species richness between the two zones could be due to the Simpson’s Diversity Index (D) = ∑ pi² edge effects, which often results higher species (Baral and Katzensteiner, 2009), richness and greater numbers of exotic species Where, pi is the proportional abundance of the at the edges (Ranney et al., 1981), and potential ith species i.e. the proportion of individuals of a ecosystem processes and productivity function given species relative to the total no. of individual alters (Laurance et al., 1997). in an effect-zone (i.e. forest stand). The calculated Sorenson’s Similarity Index Sorenson’s Similarity Index (SSI): It is a very (SSI) value of the two effect-zones was found simple measurement of beta diversity. The SSI to be 0.69 (near to value 1) which indicated value ranges from 0 where there is no species that the species found in both the effect-zones overlap between the effect-zones to 1 when were more or less similar. Shorea robusta a exactly the same species are found in both the found to be the dominant tree species in both effect-zones. the effect-zones. Other common tree species noticed were Dalbergia sissoo, Terminalia alata and Buchanania latifolia. Similarly, Argemore Sorenson’s Similarity Index (SSI) = 2c / (S1 + S2) (Magurran, 1988), maxicana was found to be the principal shrub species in both the effect-zones.

72 KC and Gautam Banko Janakari, Vol. 26, No. 1 Table 1: List of the plant species found in the two effect-zones S.N. Local Name Botanical Name 20 m distance 50 m distance 1. Sal (T) Shorea robusta √ √ 2. Sissoo (T) Dalbergia sissoo √ √ 3. Jamun (T) Syzigium cumini √ √ 4. Khirro (T) Wrightia arborea √ 5. Tate (T) Sapindus mukorossi √ 6. Ranisalla (T) Pinus roxbughii √ 7. Khannyu (T) Ficus semicordata √ 8. Pyar (T) Buchanania latifolia √ √ 9. Tilka (T) Wendlendia appendiculata √ √ 10. Bhorla (W) Bauhinia vahlli √ √ 11. Bhalayo (T) Rhus wallichii √ 12. Bot dhanyero (T) Largerstromia parviflora √ 13. Gaitihare (T) Inula cappa √ 14. Saj (T) Terminalia alata √ √ 15. Imili (T) Tamarindus indica √ √ 16. Kyamuno (T) Syzigium cerasoides √ 17. Mauwa (T) Madhuca indica √ 18. Amba (T) Psidium guajava √ 19. Latimauwa (S) Engelhardia spicata √ √ 20. Kutmero (T) Litsea monopetala √ 21. Mainfalkada (S) Catuna regamspinosa √ √ 22. Badulpate (W) Cissampelos pareira √ √ 23. Thakkal (S) Argemore maxicana √ √ Note: T = Tree, S = Shrub and W = Woody climber

Species diversity of trees diversity as compared to the effect-zone at 20 m distance due to the road-edge effects. Altogether, 11 plant species were found to be at tree stage in the effect-zone at 20 m distance Table 2: Simpson’s Diversity Index values in while a total of 10 plant species were found to be the two effect-zones at that stage in the effect-zone at 50 m distance Forest stand Stage of the Simpson's S.N. (Table 1). For the higher plant species diversity, (effect-zone) plants Index (D) the number of plant species present in an effect- 1. At 20 m distance Tree 0.7113 zone is not so important, but the even distribution of each individual plant species within the zone is Sapling 0.7187 important. In the effect-zone at 50 m distance, the Seedling 0.4449 plant species were found to be evenly distributed 2. At 50 m distance Tree 0.6057 as compared to the one in the effect-zone at 20 Sapling 0.6850 m distance. The Simpson’s Diversity Index (D) Seedling 0.3881 was found to be 0.6057 in the effect-zone at 50 m distance while it was 0.7113 at 20 m distance Species diversity of saplings (Table 2). Thus, the value of D was found to be slightly less within the effect-zone at 50 m Altogether, 7 plant species at sapling stage were distance as compared to the one within the effect- found in the effect-zone at 20 m distance while a zone at 20 m distance, which showed that the total of 9 plant species at that stage were found in effect-zone at 50 m distance was rich in plant

73 Banko Janakari, Vol. 26, No. 1 KC and Gautam the effect-zone at 50 m distance. A total number Structural diversity of 167 and 225 saplings of S. robusta r recorded in the effect-zones at 20 m distance and Moreover, dominant trees with 10–15 cm DBH 50 m distance, respectively. S. robusta was found class and 5–10 m height class were found to be to be unevenly distributed in the effect-zone at distributed in the study area. Trees with higher 20 m distance than in the effect-zone at 50 m DBH classes (10–15 cm and above 15 cm) and distance; other plant species were found to be in higher height classes (5–10 m and above 10 m) very few numbers. On the contrary, other species were found to be distributed more in the effect- were found to be evenly distributed in the effect- zone at 20 m distance than in the effect-zone at zone at 50 m distance in spite of the dominancy 50 m distance (Fig. 3 and 4). On the contrary, of S. robusta. The Simpson’s Diversity Index plants with lower DBH classes (0–5 cm and 5–10 was found to be 0.7187 in the effect-zone at 20 cm) and lower height class (0–5 m) were found m distance and 0.6850 in the effect-zone at 50 to be distributed more in the effect-zone at 50 m m distance, indicating a little bit higher plant distance than in the effect-zone at 20 m distance. diversity in the effect-zone at 50 m distance than The study indicated that the under-storey in the effect-zone at 20 m distance. The result also vegetation in the effect-zone at 20 m distance showed that the forest stand (effect-zone) at 50 m was comparatively lesser than that in the effect- distance was richer in species diversity at sapling zone at 50 m distance. This reveals that the roads stage as compared to the one at 20 m distance. affect not only upon plant species’ diversity but also have potential impact on structural diversity. Species diversity of seedlings This also reveals that proximity to road edge reduces under-storey vegetation and results less Altogether, 10 plant species were found at sustainable forest. seedling stage in the effect-zone at 50 m distance while 11 plant species were noticed at that stage in the effect-zone at 20 m distance. A total of 242 seedlings of S. robusta were found to be distributed in the effect-zone at 20 m distance while a total of 205 seedlings of this species were frequency found in the effect-zone at 50 m distance. The Simpson’s Diversity Indices were found to be 0.4449 and 0.3881 in the effect-zones at 20 m and 50 m distances, respectively (Table 2). The result showed that the effect-zone at 50 m distance DBH class possessed more plant diversity at seedling stage as at tree and sapling stages than the effect-zone Fig. 3: Plant distribution in terms of DBH at 20 m distance due to the proximity to the road- classes in the two effect-zones edge.

In the community-managed forests, removal of bigger and older trees is carried out during silvicultural operations so as to provide space for preferred species of younger trees (Baral frequency and Katzensteiner, 2009). Due to the removal of older trees, appropriate environment is created to regenerate new crop and also establishment of the younger ones which leads the forest towards more diverse in undergrowth. Suding (2001) Height class carried out one of the several studies which also Fig. 4: Plant distribution in terms of height documented proportionate relationship between classes in the two effect-zones species richness and light availability on the forest floor. Several studies show that silvicultural practices can have a positive or neutral effect on under- storey plant species richness (Jenkins and Parkers,

74 KC and Gautam Banko Janakari, Vol. 26, No. 1 1999). So, proper silviculture practices can, to some extent, reduce the effects of road-edge on the forests. Nevertheless, the number of plant species is not only one component of biological diversity that should be considered; under-storey species composition, spatial scale, number of endemic species and taxonomic singularity of the elements must also be taken into consideration (Ojeda et al., 1995; Zavala and Oria, 1995).

Distribution of major plant species Shorea robusta

S. robusta, T. alata, Syzigium cumini, D. sissoo, Fig. 5a: Distribution of S. robusta at different A. maxicana and Engelhardia spicata r stages in the two zones categorized as major species and the rest were categorized as others for the purpose of the study. At tree stage, S. robusta was found to be unevenly distributed in the effect-zone at 20 m distance than in the effect-zone at 50 m distance (Fig. 5a), and other major species were either almost equally distributed (e.g. T. alata) in both effect-zones or more distributed (S. cumini and D. sissoo) in the Syzigium cumini Terminalia alata effect-zone at 50 m distance (Table 5b). Dalbergia sissoo Argemore maxicana At sapling stage, all the major plant species were found to be distributed higher in the effect-zone at 50 m distance than in the effect-zone at 20 Fig. 5b: Distribution of major plant species m distance. The distribution of other species at except S. robusta at different stages in two sapling stage was higher in the effect-zone at 20 zones m distance. At sapling stage, only Catunaregam Communities are highly promoting and protecting spinosa was found to be distributed in the effect- timber yielding trees like S. robusta even in mixed zone at 50 m distance, but fodders like Ficus S. robusta forest (Ojha and Bhattarai, 2001; semicordata and Listea monopetala were not Acharya, 2003) at the expenses of low quality detected at sapling stage in the effect-zone at timber-yielding species and shrubs (Kandel, 20 m distance. At seedling stage, S. robusta, T. 2007 cited by Shrestha et al., 2010; Acharya et alata, S. cumini and C. spinosa were found to be al., 2007; Shrestha, 2005), which may be one of in higher distribution in the effect zone at 20 m the causes behind the less woody plant species’ distance than in the effect-zone at 50 m distance. diversity in the effect-zone at 20 m distance. On the other hand, all the three shrub species Proximity to road-edge makes easy to remove viz. E. spicata, C. spinosa and A. maxicana r other valuable species from the forest. found to be equally distributed at their seedling Conclusion stage in both the effect-zones. On the other hand, the two species of woody climber viz. Cissampelo The study indicates that roads bring about adverse spareira and Bauhinia vahlli were also found impacts upon the woody plant species diversity to be less distributed in the effect-zone at 20 m in the adjoining forests. Proximity to road-edge distance. reduces species diversity due to removal of low- yielding timber species. The findings of the study also reveal that the effects of road-edge cause reduction in under-storey vegetation. However, the removal of over mature, dead, dying, diseased and deformed trees can have positive effects upon the species diversity in the forest. Therefore, silvicultural operations should be carried out

75 Banko Janakari, Vol. 26, No. 1 KC and Gautam in the forests nearby roads so as to mitigate the IFRI. 2013. International Forestry Resources adverse impacts caused by the roads. and Institutions (IFRI) Network: Research Methods. http:/ www.ifriresearch.net References (accessed on: 12 December, 2015).

Acharya, K. P. 2003. Conserving biodiversity Jenkins, M. A. and Parker, G. R. 1999. and improving livelihoods: the case of Composition and diversity of ground- community forestry in Nepal. Paper layer vegetation in silvicultural openings presented in International Conference on of Southern Indiana Forests. The American Rural Livelihood, Forests and Biodiversity, Midland Naturalist 142: 1–16. Bonn, Germany, 9–13. Laurance, W. F., Laurance, S. G., Ferreira, L. Acharya, K. P., Gautam, K. R., Acharya, B. K. and V., Merona, J. M. R., Gascon, J. M. R. and Gautam, G. 2007. Participatory assessment Lovejoy, T. E. 1997. Biomass collapse in of biodiversity conservation in community Amazonian forest fragments. Science 278: forestry in Nepal. Banko Janakari 16: 46–56. 1117–1118.

Baral, S. K. and Katzensteiner, K. 2009. Diversity Lugo, A. E. and Gucinski, H. 2000. Function, of vascular plant communities along a effects and management of forest roads. disturbance gradient in a central mid-hill Forest Ecology Management 133: 249–262. community forest of Nepal. Banko Janakari 19 (1): 3–10. Magurran, A. E. 1988. Ecological Diversity and its Measurement. Princeton University DOR. 2010. Statistics of Strategic Road Press, Princeton, USA. Networks (2009/10). Department of Roads (DOR), Kathmandu, Nepal. MFSC. 2014. National Biodiversity Strategy and Action Plan. Ministry of Forests and Gautam, A. P. 2002. Forest Land Use Dynamics Soil Conservation (MFSC), Kathmandu, and Community-based Institutions in a Nepal, 20–28. Mountain Watershed in Nepal: Implications for Forest Governance and Management. Ojeda, F., Arroyo, J. and Maranon, T. 1995. Ph.D. Dissertation, Asian Institute of Biodiversity components and conservation of Technology, Thailand, xi–174. Mediterranean heathlands in Southern Spain. Biological Conservation 72: 61–72. Gautam, A. P. 2006. Combining geomatics and conventional methods for monitoring Ojha, H. R. and Bhattarai, B. 2001. Understanding forest conditions under different governance community perspectives of silvicultural arrangements. Journal of Mountain Science 3 practices in the middle hills of Nepal. Forests, (4): 325–333. Trees and People Newsletter 40: 55–61.

Geneletti, D. 2003. Biodiversity Impact Ranney, J. W., Bruner, M. C. and Leenson, J. Assessment of roads: an approach based B. 1981. The importance of edges in the on ecosystem rarity. Environment Impact structure and dynamics of forest islands. In Assessment Review 23 (2003): 343–365. Forest island dynamics in a man-dominated landscape (eds) Burgess, P. L. and Sharp, D. Gysel, W. L. 1951. Borders and openings of beech- M. Springer-Verlag, New York, USA, 67–95. maple woodlands in southern Michigan. Journal of Forestry 49: 13–19. Seiler, A. 2001. Ecological Effects of Roads: A Review. Introductory Research Essay No. Hall, C. A. S., Stanford J. A. and Hauer, R. 1992. 9. Department of Conservation Biology, The distribution and abundance of organisms Swedish University of Agricultural Science, as consequence of energy balances along Upsalla, Sweden. multiple environmental gradients. Oikos 65: 377–390. Shrestha, B. B. 2005. Fuelwood harvest, management and regeneration of two

76 KC and Gautam Banko Janakari, Vol. 26, No. 1 community forests in Central Nepal. WWF. 2013. Chitwan-Annapurna Landscape: Himalayan Journal of Science 3: 75–80. Drivers of Deforestation and Forest Degradation. Hariyo Ban Programme, World Shrestha, U. B., Shrestha, B. B. and Shrestha Wildlife Fund Nepal, Kathmandu, Nepal. S. 2010. Biodiversity conservation in community forests of Nepal: rhetoric and Zavala, M. A. and Oria, J. A. 1995. Preserving reality. International Journal of Biodiversity biological diversity in managed forests: and Conservation 2 (5): 98–104. a meeting point for ecology and forestry. Landscape and Urban Planning 31: 363– Suding, K. N. 2001. The effects of gap creation on 378. competitive interactions: separating changes in overall intensity from relative rankings. Oikos 94: 219–227.

77 Plant diversity and stand structure comparison of Mikania micrantha invaded and non-invaded tropical Shorea robusta forest

S. Basnet1*, D. B. Chand1, B. H. Wagle2 and B. Rayamajhi1

Mikania micrantha is considered to be the most problematic in terrestrial ecosystem in Eastern and Central Nepal. Despite the current situation of the Mikania invasion, quantitative data on the impacts and scale of the problems are lacking for the country. Due to the lack of information regarding scale of invasion, the stakeholders have not put forwarded the proper control mechanism of the species. This paper has made an attempt to analyze the scale of invasion through the comparison of plant biodiversity and basal area per hectare as the significance of stand structure between Mikania- invaded and non-invaded tropical Shorea robusta forest areas. This study was conducted in Barandabhar Buffer Zone Forest of Chitwan National Park. Sampling and measurement was conducted in both the invaded and non-invaded forest areas. The stand structure of both the invaded and non-invaded areas were compared in terms of different parameters like seeding density, sapling density, pole basal area per hectare and tree basal area per hectare. Statistical analysis showed that there is significant impact ofMikania on plant diversity at seedling and sapling stages. There is negative effect of Mikania on stand structure of the forest. Hence, there is urgent need to control the invasive weed so as to control further invasion and to conserve biodiversity and productivity.

o Basal area, invasion, Mikania micrantha, plant diversity

epal is well known for its diverse flora climber, commonly called mile-a-minute weed, Nand many plant species are endemic to because of its vigorous and rampant growth the country. However, some of them have been habit. It has been reported to grow to 27 mm a introduced unintentionally owing to the land- day (ISSG, 2005). It is a pernicious weed in crops linked situation of the country. In a country like such as rubber, cacao, oil palm, coconut, banana, Nepal, an infestation of plant invasive species pepper and tea, and usually grows profusely in makes rural livelihoods more vulnerable because places receiving high rainfall or humid habitats economy of farming community heavily depends (Holm et al., 1977). upon forest resources (Adhikari et al., 2004), and an introduction of such invasive plants are likely In Nepal, Mikania was first reported in 1963 in to influence upon the native ecosystem. the eastern part of Nepal (Tiwari et al., 2005). Later on, it started spreading towards the western In Nepal, altogether 166 alien plant species are part, and now it has been recorded in the 20 invading different ecosystems including forest, Terai districts (Rai et al., 2012). M. micrantha grassland, agricultural land and wetlands (Tiwari is assessed as one of the six high-risk-posed et al., 2005). Mikania micrantha (Kunth), a invasive alien species in Nepal (Tiwari et al., tropical plant belonging to the family Asteraceae, 2005). In the Chitwan National Park (CNP), M. is a perennial, sprawling vine with a wide micrantha was found to be the most serious weed distribution in the Neo-tropics, which extends among the eight invasive alien species (IAS) from Mexico to Argentina (Holmes, 1982). It in terrestrial ecosystem (Sapkota, 2006). The is one of the top 10 worst weeds in the world species is considered as the most problematic (Holm et al., 1977). It is a fast growing, perennial terrestrial invasive species in the tropical parts of

1 International Union for Conservation of Nature, Lalitpur, Nepal. *E-mail: [email protected] 2 Institute of Forestry, Hetauda, Nepal

78 Basnet et al. Banko Janakari, Vol. 26, No. 1

Nepal (Poudel et al., 2005; Siwakoti, 2007). geographical area of the Park as well as the surrounding forest and the shrub-land, Mikania In Nepal, this plant is known by different names is being economically serious in and around the such as Lahare Banmara and Lahare, and is found study area (Rai et al., 2012). up to 1,300 m altitude. The plant has very low use values except as fodder during the lean period; Sampling design cattle only consume it if nothing else is available, and people collect edible ferns that grow under Stratified systematic sampling method was the canopy of Mikania (Baral, 2004). adopted for data collection. First of all, almost equal two distinct strata of Mikania-invaded and At present, the weed has vigorously invaded the non-invaded areas were delineated visually with core and buffer zone of the Chitwan National the help of GPS hand receiver. Then, systematic Park, threatening to biological diversity and sample plots were allocated on both the strata ecosystem. The plant spreads appallingly fast for the collection of sample. Both the strata were in the moist part of the park, becoming dense of almost similar soil condition and stand age. within 8–10 years (Tiwari et al., 2005). Mikania Sample plots were established in both invaded is spreading freely without any hindrance in the and non-invaded areas for comparing the stand lowland of Nepal. structure and plant diversity at species level. The study area comprises the tropical Shorea robusta Despite the current situation of the weed, forest. The major tree species found are Shorea quantitative data on the impacts and scale of the robusta, Terminalia tomentosa, Bombax ceiba, problems are lacking for the country. Therefore, Dalbergia sissoo, Trewia nudiflora and so on. this study aimed at comparing the effects of Mikania invasion on understory and over story Data collection plant diversity. The information regarding impact level could be useful to the stakeholders for Fourteen nested circular sample plots were common understanding. So, this paper also aims established in each stratum for the measurement of to create stakeholders’ concern in such a serious tree, pole, sapling and regeneration (Fig. 1). Thus, 2 biodiversity threat in the nation. a circular plot of 500 m area (12.6 m radius) was established for tree (a) within which nested plots Materials and methods of 100 m2 (5.6 m radius), 25 m2 (2.8 m radius) and 10 m2 (1.8 m radius) were established for pole Study area (b), sapling (c) and regeneration (d), respectively The study was carried out in Barandabhar Buffer (DoF, 2004). Species-wise diameters at breast Zone Forest of Chitwan National Park (CNP) height (DBHs) and heights were measured and which is located at 27o37’02” N latitude and recorded in the case of trees and poles. Similarly, 84o26’15” E longitude. The CNP is renowned DBH was measured and recorded in the case of for its unique diversity of flora and fauna. saplings whereas species-wise plant numbers Recognizing its unique ecosystem of International were recorded in the case of regeneration. significance, the UNESCO declared it as world heritage site in 1984 (MFSC, 2014). Barandabhar Forest Corridor is the only remaining natural forest with an area of 161 km2 that connects the CNP and the Mahabharat Range. The human population around the Corridor is over 50,000 (CBS, 2011), imparting excessive pressure on its ecosystems. Retention and restoration of such d ecological corridors, linking protected areas, are considered essential in maintaining and restoring wildlife populations across the landscape.

The invasion is reported to have invaded the Park and the adjoining area in the 1980s and within the last 45 years, Mikania has colonized in large Fig. 1 : A nested circular sample plot

79 Banko Janakari, Vol. 26, No. 1 Basnet et al.

Data analysis the non-invaded area than in the invaded area at 5% level of significance P( 0.001; N=14). The data collected from the field were analyzed by assessing the regeneration status together with The coverage of M. micrantha was found to have the sapling diversity and density. The basal areas reduced the growth as well as germination of of the trees and poles per hectare was computed the tree species, resulting in the poor density of to assess the over-story tree density. For the the seedlings and saplings of the tree secies. The calculation of plant diversity of the study area, lower seedling density in the invaded area showed Simpson’s Index (D) was adopted. the adverse effects of Mikania on the growth and regeneration of forest crops; the invasion of 2 Simpson’s Index (D) = ∑ (Pi) , Mikania was found to have created shade and lack of openings on the ground, resulting in the Where, Pi = Proportion of individual species in low density of the seedlings. the community i. Sapling density Species distribution and tree basal area were calculated as follows: The per hectare density of the saplings was found to be higher in the non-invaded area (149/ha) 1. Species density No. of regeneration = x 10000 than in the invaded area (60/ha). This showed (N per ha) Area of a plot that the growth of the understory was severely affected by the invasion of Mikania. Independent πd2 2. Individual tree basal area (BA) = t-test revealed that the saplings per hectare was 4 significantly higher in the non-invaded area than Where, d = diameter at breast height. in the invaded area at 5% level of significance P = 0.015; N=14). After the calculation of various attributes (density, basal area per ha) of growing stock, the two strata The difference could be due to the effect of the were compared using various statistical tests invasion of Mikania on the germination of the (t-test, independent sample test). plants as well as the growth of the seedlings, resulting in the lower density of saplings in Results and discussion invaded strata. Plant diversity at seedling and sapling stages Basal area of pole per hectare The Simpson’s Index values of the Mikania The basal area of the poles per hectare was more invaded and non-invaded strata at regeneration than three times higher in the non-invaded (5.6 i.e. seedling level were found to be 0.1806 and m3/ha) area than in the invaded area (2.2 m3/ha). 0.0681, respectively while those at sapling level Also, it was found significantly higher in the non- were found to be 0.3406 and 0.3192, respectively. invaded area than in the invaded area at 5% level As Simpson’s Index values are inversely related to of significance (P = 0.002; N=14). The diameter the plant diversity, the plant diversities at sapling growth and basal area formation was found to and seeding stages were found to be higher in the be significantly less in the invaded area than in non-invaded area. the invaded area. The invasion of Mikania might The climber and shoots of Mikania was found to have affected upon the phenological process of be totally obstructing the opening to ground level the plant species, its development and cambium which might have restricted germination and formation might have been also affected. That growth of its saplings and seedlings. is why the density of poles and its individual size was found to be comparatively lower in the Seedling density invaded area.

The per hectare density of the seedlings in the non- Tree basal area per hectare invaded area (19,000/ha) was found to be almost six times more than that in the invaded area (3,214/ The non-invaded area had more than two times ha). Independent sample t-test revealed that the more basal area per hectare than in the invaded 3 3 seedings per hectare was significantly higher in area (invaded: 2.2 m /ha; non-invaded: 5.6 m /

80 Basnet et al. Banko Janakari, Vol. 26, No. 1 ha). It was also found to be significantly higher in Biodiversity Conservation Effort in Nepal. the non-invaded area at 5% level of significance International Centre for Integrated Mountain P = 0.024; N=14), indicating significantly higher Development, Lalitpur, Nepal. basal area growth in the non-invaded area than in CBS. 2011. Nepal Population Census Report, the invaded area. 2011. Central Bureau of Statistics (CBS), As in the case of the pole basal area, the invasion Kathmandu, Nepal. might have great impact upon the phenological DoF. 2004. Community Forest Inventory process of the trees, resulting in the slow growth Guidelines. Government of Nepal, Ministry in the overall basal area of the forest stand in the of Forests and Soil Conservation, Department invaded area. of Forests (DoF), Kathmandu, Nepal. Conclusion Holm, L. G., Plucknett, D. L., Pancho, J. V. and Herberger, J. P. 1977. The World’s The study has revealed the current scenario of Worst Weeds: Distribution and Biology. plant diversity and stand structure in Mikania- University Press of Hawaii, USA. invaded and non-invaded areas. It this study, the plant diversity and composition at undergrowth Holmes, W. C. 1982. Revision of the old level was found to be significantly affected world Mikania (Compositae). Botanische from the invasion whereas the non-invaded area Jahrbücher 103 (2): 211–246. had better growth condition. The invasion of ISSG. 2005. Ecology of Mikania micrantha. Mikania had created serious ecological problem Invasive Species Specialist Group Database. in forest management, both in terms of diversity http://www.issg.org/database/species// and productivity. The findings of the study have, ecology.asp?si=42 accessed on 4 April, 2015. therefore, clearly indicated the adverse impact of MFSC. 2014. Nepal Biodiversity Strategy and Mikania invasion on forest growth. Therefore, Action Plan (2014–2020). Government regular treatment and control mechanism should of Nepal, Ministry of Forests and Soil be developed so as to control invasion of Mikania Conservation (MFSC), Kathmandu, Nepal. and other invasive species in a forest invaded by such invasive species and to improve the forest Poudel, A., Baral, H. S., Ellison, C. A., Subedi, condition. K., Thomas, S. and Murphy, S. 2005. Mikania micrantha Weed Invasion in Nepal. Acknowledgements A Summary Report of the First National Workshop for Stakeholders held on 25 We sincerely thank Chitwan National Park and November, Kathmandu, Nepal. Baradhabar Buffer Zone Forest Management Committee for their support in data collection Rai, R. K., Scarborough, H., Subedi, N. and during field works. We would also like to thank Lamichhane, B. 2012. Invasive plants - do Md. Yusuf Ansari, Campus Chief, Institute of they devastate or diversify rural livelihoods? Forestry, Hetauda Campus and Mr. Rajiv Kumar Rural farmers’ perception of three invasive Jha for their guidance and recommendation. plants in Nepal. Journal for Nature Conservation 20 (3): 170–176. References Sapkota, L. N. 2006. Invasive Alien Species in Chitwan National Park, Nepal. M.Sc. Thesis, Adhikari, B., Di Falco, S. and Lovett, J. C. Institute of Forestry, Tribhuvan University, 2004. Household characteristics and forest Pokhara, Nepal. dependency: evidence from common property forest management in Nepal. Ecological Siwakoti, M. 2007. Mikania weed: a challenge Economics 48 (2): 245–257. for conservationists. Our Nature 5: 70–74. Baral, H. S. 2004. Mikania micrantha Weed Tiwari, S., Adhikari, B., Siwakoti, M. and Subedi, Invasion in Nepal. Paper presented at K. 2005. An Inventory and Assessment of First National Stakeholders’ Workshop, Invasive Alien Plant Species of Nepal. Kathmandu, Nepal, 25 Nov, 2004. IUCN-The World Conservation Union, Kathmandu, Nepal. Bhatta, S. R. 2006. Grassland Management in Royal Chitwan National Park:

81 Microbial inoculant influences the germination and growth of Albizia lebbeck seedlings in the nursery B. M. Khan1*, M. A. Kabir2, M. K. Hossain1 and M. A. U. Mridha3

Microbial inoculants (MI), a biofertilizer, composed of many different beneficial microorganisms has positive role on seed germination and growth of plants. In the present study, its efficacy on seed germination and seedling growth ofAlbizia lebbeck in the nursery was studied. The seeds were sown in polybags filled with a mixture of forest soil and cow dung (3:1) and treated with 0.1%, 0.5%, 1%, 2%, 5% and 10% concentrations of MI. Most of the parameters studied (seed germination, shoot and root lengths, dry weights of shoot and root, collar diameter, leaf number etc) were found maximum in 2% of MI . Although the highest vigor index, volume index and quality index (7053, 3738 and 1.106, respectively) were found in 2% MI, but the highest sturdiness (65.95) was found in 1% MI solution. The nodule number was higher at a very low (0.5%) concentration of MI but it normally decreased with the increase of concentration. Total pigment content in leaf was recorded highest (112.86 mg.L-1) in 2% of MI. Therefore, MI influences seed germination and seedling growth of A. lebbeck and the low concentration (2%) of the inoculant can be recommended for getting maximum seed germination and seedling growth of the species studied.

o Albizia lebbeck, germination, microbial inoculant, seedling growth

lbizia lebbeck (L.) Benth [Kalo sirish in conditions (Luna, 1996; Qadri and Mahmood, ANepali, Kala koroi in Bengali], is a moderate 2005; Mishara et al., 2010; Bobby et al., 2012; to large deciduous tree with a straight bole and Shaikh et al., 2014). The plant has already been broad crown under the family Leguminosae proven successful for afforestation, reforestation, (Mimosoideae). The species is widely spread social forestry and agroforestry programs in in the world, and is native to Asia, Africa and Bangladesh (Zabala, 1990; Dey, 2006). Northern Australia. It grows naturally in Nepal, Bangladesh, Myanmar and Pakistan and has been To fulfill the high demand, many organizations cultivated in tropical and subtropical regions in are producing A. lebbeck seedlings in the nursery Northern Africa, the West Indies, South America, in Bangladesh to supply those in the plantation and south Asia. Extensive plantations had been programs. Because the plants are grown mostly established in Nepal and in South India (Luna, in unfavorable soil conditions, beneficial soil 1996; Kumar et al., 2010; Elzaki et al., 2012; microorganisms can play a significant role in early Missanjo et al., 2013). The wood is excellent establishment and better growth of the inoculated for furniture and timber for general uses, post, seedlings under field conditions. piles, fuel wood and charcoal. The tree is used The microbial inoculant (MI) in this study, with the as ornamental and nurse tree for tea gardens, commercial name “Effective Microorganisms” coffee and cocoa orchards (Mishara et al., 2010; or EM was developed at the University of Missanjo et al., 2013; Shaikh et al., 2014). It is a Ryukyus, Okinawa, Japan, in the early 1980s soil improver because of its inherent ability to fix by a distinguished professor of horticulture, nitrogen (Qadri and Mahmood, 2005). The leaves Dr. Teruo Higa (Kyan et al., 1999). The main are excellent fodder for livestock. The species is species comprising MI are lactic acid bacteria also very useful for the rehabilitation of degraded Rhodopseudomonas spp.), photosynthetic lands because of its rapid growth, ability to bacteria (Lactobacillus spp., Streptococcus fix nitrogen and tolerance for a range of soil

1 Institute of Forestry and Environmental Sciences, University of Chittagong, Chittagong-4331, Bangladesh. *E-mail: [email protected] 2 Department of Agroforestry, Patuakhali Science and Technology University, Patuakhali-8602, Bangladesh. 3 Plant Production Department, King Saud University, Kingdom of Saudi Arabia.

82 Khan et al. Banko Janakari, Vol. 26, No. 1 spp.), yeast (Saccharomyces spp., Candida spp.), Actinomycetes (Streptomyces spp.) and beneficial fungi (Aspergillus spp., Penicillium spp.). The microorganisms are added into the inoculant in the manufacturing process and can survive in the inoculant liquid at pH 3.5 or below.

The density of most of the above mentioned microbes is in the range of 1 × 106 to 1 × 108 mL-1 (Xu, 2000). MI can be applied as inoculant to increase the microbial diversity of soils. It has been used with considerable success to improve soil quality and yield of crops, particularly in nature farming and organic farming systems (Xu, 2000). Inoculation of MI culture can improve Fig. 1: Map showing the location of the photosynthesis and fruit yield (Xu, 2000; Wang et nursery of IFESCU (Institute of Forestry al., 2000). Although A. lebbeck is used for wide and Environmental Sciences, University range of purposes and even planted intensively of Chittagong) in Bangladesh where the in the field, the initial growth potential under the experiment was conducted. influences of MI was not studied. Therefore, the aim of this study was to observe the effectiveness Treatment design of MI on germination of seed and the growth of seedlings of A. lebbeck and also to find out the The experiment was conducted in 03 March to best concentration of MI solution for ensuring 02 August 2015. There were seven treatments maximum seedling development in the nursery. including control and 25 replications for each treatment. Seeds were sown in polybags (filled Materials and methods with soil and cow dung) with no added MI but water only for control treatment. Other treatments Collection of seeds and soil included; sowing the seeds in polybags with the The experiment was carried out in the nursery concentrations of 0.1%, 0.5%, 1%, 2%, 5% and of the Institute of Forestry and Environmental 10% of MI, respectively. For preparing 0.1% of Sciences; University of Chittagong, Bangladesh MI solution, 0.1 mL stock solution of MI was (lies approximately at the intersection of 91o50´E added with 99.9 mL of water while for preparing and 22o30´N) (Fig. 1). The seeds of A. lebbeck 0.5% of MI solution, 0.5 mL stock solution of were collected from the seed orchard division MI was added with 99.5 mL of water and the of Bangladesh Forest Research Institute (BFRI) same formula was used for preparing the other where the source of seed was a mother tree of 25 solutions. For each treatment, 50 mL of MI years old. The soils collected from the degraded of required concentration was poured in each hills of the University Campus was sieved well polybag (filled with soil and cow dung) before a (<3 mm) and mixed thoroughly with decomposed week of sowing the seeds while another 50 mL cow dung in a ratio of 3:1. The brown hill soils was poured after a week of sowing the seeds. Five (Rashid, 1991), are sandy loam to sandy clay seeds were sown in each polybag to observe the loam, moderately to strongly acid and poorly influence of MI on germination in the nursery o fertile with pH <5.5, organic matter <2.0%, CEC conditions (temperature, 28 C; humidity, 75%). <10 me/100 g, BSP <40% (Osman et al., 2001). After completion of germination, only one The white polybags of 15 cm x 10 cm in size were seedling (the best one) per polybag was managed filled with the prepared mixture and a thin layer to observe growth performance and nodulation of coconut husk was added to each bag as top status of the seedlings. Partial shade and cover layer to reduce evaporation and to supply organic was ensured using polythene sheet on the nursery matter. roof to protect the seedlings from strong sunlight and rains.

83 Banko Janakari, Vol. 26, No. 1 Khan et al.

Growth measurement carefully, avoiding the fragmented plant tissues. The process was repeated until the leaf fragments Germination was recorded daily from the date became colorless. Finally, the volume was made up of seed sowing to the last of germination. The to 25 mL with fresh acetone and the measurement seedlings were allowed to grow altogether for five was taken immediately after preparation of the months from the time of seed sowing. After five solution. The measurement of chlorophyll-a, months, five representative seedlings from each chlorophyll-b, and carotenoid were made at 662 treatment were selected for measuring growth nm, 644 nm, and 440.5 nm respectively, with a parameters. The recorded parameters were shoot spectrophotometer (Spectronic-20). The pigment and root lengths, collar diameter, leaf number, contents in the extract were calculated following fresh shoot and root weights, dry shoot and root the formula of Wettstein (1957). weights, and nodulation status. For recording dry weights, shoots and roots were oven dried at 80oC (9.784E662 – 0.99E644) x V x d C = (1) for 48 hr. To assess the seedling vigor, total height a 1000 F (from the soil surface to seedling tip) of each w seedling in each sub-plot was measured using (21.426E644 – 4.650E662) x V x d a ruler to the nearest 0.1 cm. Vigor index was Cb = (2) calculated according to Abdulbaki and Anderson 1000 Fw (1973) as germination percent x seedling total length i.e. total shoot and root length. Volume [4.695E440.5 – 0.268(C +C E662)] x V x d C = a b (3) index was obtained by multiplying shoot height c 1000 Fw or shoot length (cm) with the square of collar 2 diameter (mm) of the seedling. Quality index where, Ca is the chlorophyll-a (mg.L-1); Cb -1 was developed following Dickson et al. (1960) the chlorophyll-b (mg.L ); Cc the carotenoid to quantify seedlings morphological quality. The (mg.L-1); V the total volume (25 mL); d the formula for calculating quality index is as follow: dilution factor; FW the fresh weight of leaf disc (g); and E is the absorbance at a particular H S dw wavelength (440.5, 644 and 662 nm). QI = T dw / ( + ) D c R dw Statistical analysis where, QI is quality index, T dw is total dry weight All data were analyzed statistically using computer (g), H is seedling height or shoot length (cm), D c software SPSS (Version 20, SPSS Incorporation, is collar diameter (mm), S dw is shoot dry weight Chicago, USA). Possible significant variations (g), R dw is root dry weight (g). Sturdiness was among the treatments were explored by Duncan’s obtained by dividing shoot height or shoot length Multiple Range Test (DMRT). Prior to statistical (cm) with collar diameter (cm) of the seedling. analysis, the normality of each data set was tested using an Anderson-Darling test, in case Measurement of pigment contents a transformation was necessary. Log base 10 transformation was done where required. Means The pigment contents (chlorophyll-a, were separated by MS Excel (MS Office 2010). chlorophyll-b, and carotenoid) were determined from the fresh leaves of seedlings in different Results and discussion treatments (Wettstein, 1957; Khan, 2012). Ten leaf discs were cut with a cork borer (inside Germination and seedling growth diameter of 5 mm), weighed immediately after The highest (77%) seed germination was cutting, and dipped in 100% acetone in 5 mL in recorded in 2% of MI, while the lowest (61%) test tube with stopper. After 24 hr of incubation, was recorded in control treatment (without the supernatant-colored solution from the top was inoculation). Shoot (54.0 cm) and root (37.6 cm) decanted carefully into a 25 mL volumetric flask. lengths were also highest in 2% of MI (Table 1). The leaf discs were then crushed with a blunt With the application of 2% of MI, shoot length glass rod gently and 5 mL fresh acetone was increment was 33% compared to control (Fig. 2). added to the test tube and left for 15 min. Then Collar diameter was maximum (8.32 mm) in 2% the supernatant-colored solution from the top was of MI and was significantly (P=0.015) different again decanted to the same volumetric flask very

84 Khan et al. Banko Janakari, Vol. 26, No. 1 Table 1: Influence of microbial inoculant on germination, shoot and root lengths, collar diameter and leaf number of Albizia lebbeck in the nursery Concentration of Germination Length (cm) Collar dia. Number of MI (%) (%) Shoot Root Total (mm) leaf

Control 61c 40.6b 27.8b 68.4b 6.41b 16.8b

0.1 65b 45.3ab 31.2ab 76.5ab 6.95b 17.6b

0.5 72a 48.4a 33.5a 81.9a 7.42a 18.0a

1 74a 52.3a 35.7a 88.0a 7.93a 19.6a

2 77a 54.0a 37.6a 91.6a 8.32a 20.2a

5 68b 47.1ab 32.3ab 79.4ab 7.47a 18.4a

10 65b 42.9b 29.4b 72.3b 6.94b 17.2b P value <0.001 0.002 <0.001 <0.001 0.015 0.006 F value 137.18 16.37 74.82 104.51 6.35 9.21 Note: a-c = Mean values with different lowercase superscripts in a column are significantly different atP <0.05, according to Duncan’s Multiple Range Test (DMRT). from the control. Seedlings treated with MI had g, respectively) in 2% of MI and were significantly more leaf compared to the control (Table 1). P<0.001 and P=0.026, respectively) varied from control (Table 2). Though maximum vigor index, volume index and quality index (7053, 3738 and 1.106, respectively) were found in 2% of MI (Table 2 and Fig. 3), the maximum sturdiness (65.95) was found in 1% of MI treatment (Fig. 4). The total dry biomass increased gradually with the increase of the concentrations of MI up to 2% while decreased gradually from its maximum point as the concentrations of MI increased above 2%. Increased biomass production might be due to the better root development in the treated seedlings (Khan et al., 2011, 2014). Such promotion might also be due to the biological active substances in Fig. 2: Influence of microbial inoculant on the inoculant (Lim et al., 1999), such as indole shoot length increased to decreased (%) acetic acid (IAA) and gibberellins produced by with respect to control in Albizia lebbeck. Lactobacillus spp., Rhodopseudomonas spp., Cont. = Control. Aspergillus spp. and Saccharomyces spp. which enhance plant growth (Chowdhury et al., 1994). Both fresh (21.4 g) and dry (6.95 g) shoot weights However, germination rate was below 80% which were maximum in 2% of MI. Both fresh and dry was probably due to the quality of seeds as well root weights were also maximum (9.1 g and 3.01 Table 2: Influence of microbial inoculant on fresh and dry weights of shoot and root, vigor index and volume index of Albizia lebbeck Fresh weight (g) Dry weight (g) Total dry Index Concentra- biomass tion of MI Shoot Root Total Shoot Root Total increment Vigor Volume (%) (%)

Control 14.9b 5.9b 20.8b 4.48b 1.89c 6.37b 00.00 4172c 1668c

0.1 15.7b 6.7b 22.4b 4.72b 2.25b 6.97b +9.42 4973b 2188b

0.5 16.5b 7.3a 23.8ab 4.65b 2.37b 7.02b +10.20 5897a 2665a

1 19.3a 8.0a 27.3a 5.88a 2.58a 8.46a +32.81 6512a 3289a

2 21.4a 9.1a 30.5a 6.59a 3.01a 9.60a +50.71 7053a 3738a

5 20.7a 8.8a 29.5a 6.29a 2.92a 9.21a +44.58 5399b 2628a

10 17.6b 7.8a 25.4ab 5.16a 2.57a 7.73a +21.35 4700b 2066b P value <0.001 <0.001 <0.001 0.017 0.026 0.021 -- <0.001 <0.001 F value 146.12 114.09 127.65 6.19 4.83 5.27 -- 57.65 88.34 a-c = Mean values with different lowercase superscripts in a column are significantly different atP <0.05, according to Duncan’s Multiple Range Test (DMRT). 85 Banko Janakari, Vol. 26, No. 1 Khan et al. as the environmental factors regulating seed many other countries of the world. Application germination. of MI can play a role in enhancing germination, growth, and yield of various agricultural crops and vegetables (Vongprachanch, 1995; Zacharia, 1995; Iwaishi, 2000; Chowdhury et al., 2002). MI with organic fertilizer and other chemicals is also reported to enhance germination, growth, and yield of different grains (Ahmed et al., 1995; Anuar et al., 1995; Xu, 2000).

But the influence of MI on forest crops has not been studied widely (Khan et al., 2006, 2011). From this study, it has been observed that soil amended with different concentrations of MI can also improve the seedling growth of forest crops. Mridha (2005), Khan (2012) and Khan et al. (2011, Fig. 3: Influence of microbial inoculant on 2014) also reported enhanced seed germination quality index of Albizia lebbeck. Cont. = rate and seedling growth with the application of Control. low concentrations of MI. Specially, applying 2% of MI enhanced seed germination and seedling growth at an optimum level which might be due to the presence of microbial population in such a density within this concentration which was cable to produce optimum level of growth enhancing hormones and other chemicals. However, as the concentration of MI increased, seed germination and seedling growth was depressed, which may be due to the perniciousness exerted by the higher concentrations of the inoculant (Mridha, 2005; Khan et al., 2014).

Nodulation status

Fig. 4: Influence of microbial inoculant on Although most of the parameters were highest sturdiness of Albizia lebbeck. Cont. = Control. in 2% of MI, the highest (134) nodule number Microbial inoculants are being applied in Japan, was in 0.5% of MI while the lowest (95) was in China, Thailand, United States, France, Brazil and 10% of MI (Table 3). Both fresh and dry nodule weights were maximum (2.07 g and 0.67 g,

Table 3: Influence of microbial inoculant on nodule number and their fresh and dry weights of Albizia lebbeck Nodule Concentration Weight (g) Weight increased or decreased (%) of MI (%) Number Fresh Dry Fresh Dry

Control 127a 1.77a 0.56ab 0.00 0.00

0.1 131a 2.01a 0.65a +13.56 +16.07

0.5 134a 2.07a 0.67a +16.95 +19.64

1 121a 1.89a 0.61a +6.78 +8.93

2 116b 1.75a 0.54ab -1.13 -3.57

5 104b 1.57b 0.51ab -11.30 -8.93

10 95c 1.42b 0.44b -19.77 -21.43 P value <0.001 0.013 0.019 -- -- F value 51.42 6.68 5.3 -- -- a-c = Mean values with different lowercase superscripts in a column are significantly different atP <0.05, according to Duncan’s Multiple Range Test (DMRT)

86 Khan et al. Banko Janakari, Vol. 26, No. 1 respectively) at 0.5% of MI and lowest (1.42 g Content of leaf pigments and 0.44 g, respectively) were in 10% of MI. The rate of nodule increment was positive in case of Effects of MI on the content of leaf pigments 0.1% and 0.5% of MI, while negative for all other (Chlorophyll-a, chlorophyll-b, and carotenoid) treatments compared to control (Fig. 5). The were also determined (Table 4). Chlorophyll-a -1 nodule dry weight increment rate was positive for was highest (56.12 mg.L ), in 2% of MI and -1 0.1%, 0.5%, and 1% of MI while negative for all lowest (37.39 mg.L ) in the control treatment. -1 other treatments, with respect to control (Table 3). Chlorophyll-b was highest (15.67 mg.L ) in -1 The number of nodule in soybean root was also 2% of MI, followed by 14.98 mg.L in 5% and -1 not changed significantly due to application of low 13.59 mg.L in 1% of MI, and was significantly concentrations of MI (Thach et al., 1999) while P<0.001) varied from control. Carotenoid was -1 decreased in higher concentrations in Dalbergia maximum (41.07 mg.L ) in 2% of MI. Total -1 sissoo and Acacia auriculiformis (Khan et al., pigment was recorded highest (112.86 mg.L ) in 2011, 2014). The higher concentrations of MI 2% of MI and was significantly (P<0.001) different solution in the substratum may affect the growth from control. The results are in agreement with of plants because of toxic effect of secretion of the findings of Xu (2000), Wang et al. (2000), toxic metabolites. Mridha et al. (2002), Khan et al. (2006, 2011) and Khan (2012) that low concentrations of MI Table 4: Influence of microbial inoculant on pigment contents in fresh leaves of Albizia lebbeck Concentration of Pigments concentration of leaf (mg.L-1) Total pigment MI (%) Chlorophyll-a Chlorophyll-b Carotenoid Total increment (%)

Control 37.39c 10.34b 28.41c 76.14c 0

0.1 41.51b 11.75ab 32.97b 86.23b +13.25

0.5 43.34b 12.06ab 34.02b 89.42b +17.44

1 50.02a 13.59a 37.18a 100.79a +32.37

2 56.12a 15.67a 41.07a 112.86a +48.23

5 54.17a 14.98a 40.84a 109.99a +44.46

10 47.49b 12.32ab 31.49b 91.30b +19.91 P value <0.001 <0.001 <0.001 <0.001 -- F value 38.26 49.58 91.63 61.47 -- a-c = Mean values with different lowercase superscripts in a column are significantly different atP <0.05, according to Duncan’s Multiple Range Test (DMRT)

with organic fertilizer promote root growth and enhance photosynthetic efficiency and yield of seedling. Conclusion Microbial inoculant influences seed germination and seedling growth of A. lebbeck and the low concentration (2%) of the inoculant can be recommended for getting maximum seed germination and seedling growth of the species in the nursery conditions that may also be effective for seedling development in the field. References Fig. 5: Influence of microbial inoculant on nodule number increased to decreased (%) Abdul-Baki, A. and Anderson, J. D. 1973. Vigor with respect to control in Albizia lebbeck. determination in Soybean seed by multiple Cont. = Control. criteria. Crop Science 13: 630–633. Ahmed, R., Hussain, T., Jilani, G., Shahid, S. A., Akhtar, S. N. and Abbas, M. A. 1995. Use

87 Banko Janakari, Vol. 26, No. 1 Khan et al.

of effective microorganisms for sustainable and quality of paddy-rice varieties. Journal crop production in Pakistan. In Effective of Crop Production 3 (1): 269–273. Microorganisms (EM) (eds) Sharifuddin, Khan, B. M., Hossain, M. K. and Mridha, M. A. H. A. H., Anuar, A. R. and Shahbudin, U. 2014. Improving Acacia auriculiformis M. F. Proceedings of the 2nd International seedlings using microbial inoculant Conference on Effective Microorganisms (Beneficial Microorganisms). Journal of (EM). Kyusei Nature Farming Center, Forestry Research 25 (2): 359–364. Saraburi, Thailand, 15–27. Khan, B. M., Hossain, M. K. and Mridha, M. Anuar, A. R., Sharifuddin, H. A. H., Shahbudin, A. U. 2006. Effect of microbial inoculants M. F. and Zaharah, A. R. 1995. Effectiveness on Albizia saman germination and seedling of effective microorganisms (EM) on maize growth. Journal of Forestry Research 17 (2): grown on sandy tin tailings. In Effective 99–102. Microorganisms (EM) (eds) Sharifuddin, H. A. H., Anuar A. R. and Shahbudin, Khan, B. M., Hossain, M. K. and Mridha, M. A. U. M. F. Proceedings of the 2nd International 2011. Nursery practice on seed germination Conference on Effective Microorganisms and seedling growth of Dalbergia sissoo (EM). Kyusei Nature Farming Center, using beneficial microbial inoculants.Journal Saraburi, Thailand, 42–54. of Forestry Research 22 (2): 189–192. Bobby, N., Wesely, E. G. and Johnson, M. 2012. Khan, B. M. 2012. Environmental Biology: HPTLC profile studies on the alkaloids of Growth Improvement of Forest Seedlings Albizia lebbeck. Asian Pacific Journal of using Beneficial Microbial Inoculant Tropical Biomedicine 1–3:1–6. (Effective Microorganisms). Lap lambert academic publishing, Germany. 160 pp. Chowdhury, A. R, Hussain, M. M., Mia, M. S., Karim, A. M. M. S., Haider, J., Bhuyan N. Kumar, S., Bansal, P., Gupta, V., Sannd, R. I. and Shifuddin, K. 1994. Effect of organic and Rao, M. 2010. The clinical effect of amendments and EM on crop production Albizia lebbeck stem bark decoction on in Bangladesh. In Kyusei Nature Farming Bronchial Asthma. International Journal of (eds) Parr, J. F., Hornick S. B. and Simpson, Pharmaceutical Sciences and Drug Research M. E. Proceedings of the 2nd International 2 (1): 48–50. Conference on Kyusei Nature Farming, Kyan, T., Shintani, M., Kanda, S., Sakuria, M., Piracicaba, Brazil, 155–163. Ohashi, H., Fujisawa, A. and Pongdit, S. Chowdhury, M. H. U, Mridha, M. A. U., Khan, B. 1999. Kyusei Farming and the Technology M. and Xu, H. L. 2002. Effects of effective of Effective Microorganisms. International microorganisms on seed germination and Nature Farming, Bangkok, Thailand and seedling growth of Oryza sativa L. Nature Research Center, Atami, Japan and Asia Farming and Environment 3 (1): 23–30. Pacific Natural Agriculture Network. Dey, T. K. 2006. Useful Plants of Bangladesh. Lim, T. D., Pak, T. W. and Jong, C. B. 1999. The Ad. Communication, Anderkilla, Yields of rice and maize as affected by Chittagong, Bangladesh. effective microorganisms. In Kyusei Nature Farming and Effective Microorganisms Dickson, A., Leaf, A. L. and Hosner, J. F. 1960. (eds) Senanayake, Y. D. A. and Sangakkara, Quality appraisal of white spruce and white U. R. Proceedings of the 5th International pine seedling stock in nurseries. Forestry Conference on Kyusei Nature Farming and Chronicle 36: 10–13. Effective Microorganisms for Agricultural Elzaki, O. T., Khider, T. O., Omer, S. H. and and Environmental Sustainability. Bangkok, Shomeina, S. K. 2012. Environment friendly Thailand, 92–98. alkaline pulping of Albizia lebbeck from Luna, R. K. 1996. Plantation Trees. International Sudan. Nature and Science 10 (4): 76–82. Book Distributors. Dehra Dun, India. Iwaishi, S. 2000. Effect of organic fertilizer and Mishara, S. S., Gothecha, V. K. and Sharma, effective microorganisms on growth, yield A. 2010. Albizia lebbeck: a short review.

88 Khan et al. Banko Janakari, Vol. 26, No. 1

Journal of Herbal Medicine and Toxicology Microorganisms (eds) Senanayake, Y. D. A. 4 (2): 9–15. and Sangakkara, U. R. Proceedings of the 5th International Conference on Kyusei Nature Missanjo, E., Maya, C., Kapira, D., Banda, H. and Farming and Effective Microorganisms Kamanga-Thole, G. 2013. Effect of seed size for Agricultural and Environmental and pretreatment methods on germination of Sustainability. Bangkok, Thailand, 254–260. Albizia lebbeck. ISRN Botany 1: 1–4. Vongprachanch, B. 1995. The use of EM Mridha, M. A. U, Khan, B. M., Hossain, M. fermented materials in farm trials. In Effective K., Tasnim, H. and Rahman, M. F. 2002. Microorganisms (EM) (eds) Sharifuddin, Effective microorganisms (EM) for neem H. A. H., Anuar A. R. and Shahbudin M. tree. APANews 20: 5–6. F. Proceedings of the 2nd International Mridha, M. A. U. 2005. Growth improvement Conference on Effective Microorganisms of seedlings of a tropical forest species, (EM). Saraburi, Thailand, 90–94. Khair (Acacia catechu Willd.) using Wang, R., Xu, H. L. and Mridha M. A. U. beneficial microbial inoculant (effective 2000. Effect of organic fertilizer and EM microorganisms). Saudi Journal of Biological inoculation on leaf photosynthesis and fruit Sciences 12 (1): 67–73. yield and quality of tomato plants. Journal of Osman, K. T., Rahman, M. M. and Barua, P. Crop Production 3 (1): 173–182. 2001. Effects of some forest tree species Wettstein, D. 1957. Formula of chlorophyll on soil properties in Chittagong University determination. Exp. Cell Research 12 (3): Campus, Bangladesh. Indian Forester 127 427–489. (4): 431– 442. Xu, H. L. 2000. Effect of a microbial inoculants Qadri, R. and Mahmood, A. 2005. Ultrastructural and organic fertilizer on the growth, studies on root nodules of Albizia lebbeck photosynthesis and yield of sweet corn. (L.). Pakistan Journal of Botany 37 (4): 815– Journal of Crop Production 3 (1): 183–214. 821. Zabala, N. Q. 1990. Silviculture of Species. Rashid H. 1991. Geography of Bangladesh. Development of the Professional Education University Press Ltd. Dhaka, Bangladesh. in the Forestry Sector, Bangladesh. UNDP/ Shaikh, F. K., Gadge, P. P., Shinde, A. A., Padul, M. FAO BGD/85/011, Field Document No. 14. V. and Kachole, M. S. 2014. Characterization 174 pp. of the A1T113 protein from Indian siris Zacharia, P. P. 1995. Studies on the application Albizia lebbeck) that inhibits the growth of of effective microorganisms (EM) in paddy, cotton bollworm (Helicoverpa armigera). sugarcane and vegetables in India. In Effective Journal of Asia-Pacific Entomology 17 (3): Microorganisms (EM) (eds) Sharifuddin, 319–325. H. A. H., Anuar, A. R. and Shahbudin, Thach, N. Q., Long, C. A., Liet, V. V., Trung, M. F. Proceedings of the 2nd International N. V., Thanh, N. X., Dich, T. V., Duong, Conference on Effective Microorganisms N., Tuna, N. K., Xuan L. T. H. and Dan, P. (EM). Kyusei Nature Farming Center, V. 1999. Preliminary results of effective Saraburi, Thailand, 31–41. microorganisms (EM) application in Vietnam. In Kyusei Nature Farming and Effective

89 Drivers and dynamics of land use land cover in Ambung VDC of , Nepal

D. Pandey1*, B. P. Heyojoo1 and H. Shahi2 Land use and land cover change has immense impact on the global environment and ecosystem. Geospatial technologies are very important for monitoring these changes. This research aims to find out the land use land cover dynamics and drivers of Ambung VDC, Tehrathum district. The Landsat images of the year 1990 and 2013 were used for quantifying the changes. Household survey, key informant interview, focus group discussion, training samples collection and direct field observations were carried out to gather socio-economic and bio-physical data. Supervised classification was performed to prepare land cover maps. Change on land use was calculated by using post classification change detection. During 1990–2013, forest cover was found to have increased by 6.6%, agriculture decreased by 5.9% and others (barren, settlement, grass, rock and water bodies) decreased by 0.7%. The VDC was found to have severe problem of rapid drying of water resources in spite of the increase in forest cover, and so research should be carried out to find out the reason and solve the problem before it is too late.

o Forest cover, geographic information system, remote sensing, supervised classification and use and land cover (LULC) dynamics (Awasthi, 2004). Satellite images and aerial Lare widespread, accelerating and significant photos are useful for quantitative evaluation of process driven by human action which produces LULC changes over time (Balla et al., 2007). some changes that impact humans (Agarwal et al., 2001). Land cover corresponds to the There are regional variations in terms of changes physical condition of the ground surface, such as in forest conditions, e.g. the forest area in the Terai forest, grassland, agriculture land and so on while was found to have decreased with the annual rate land use reflects human activities like the use of of 1.3 percent during 1978/79-1990/91 (DOF, the land for different purposes such as industrial 2005). A recent assessment of forest cover by zones, residential zones, and agricultural fields. the Forest Resource Assessment (FRA) Project This definition shows that there is a direct link showed that the Terai forest had decreased by between land cover and the actions of people in 0.44 percent and 0.40 percent per annum during their environment, i.e. land cover change may the periods of 2001–2010 and 1991–2010, result from land use (Phong, 2004). Changes in respectively (FRA/DFRS, 2014). The FRA land-use and land cover have impacts on soil and findings indicate a declining rate of forest loss in water quality, biodiversity, and global climatic the Terai region in more recent years (MoFSC, systems and, thus, have important consequences 2014). Human intervention in forest environment on natural resources (Awasthi et al., 2002). is generally accepted as the main trigger behind forest conversion (Phong, 2004). Moreover, Remote sensing and geographic information the mountain region of Nepal is subjected to system (RS/GIS) technologies can greatly deforestation and agriculture expansion in the facilitate the collection, analysis and presentation marginal lands (Awasthi et al., 2005). Some of resource data (Gautam, 2007). Repeated programs, such as community forestry have satellite images and aerial photos are useful for carried out exemplary work on conserving forest visual assessment of natural resource dynamics resources, but there are also some activities occurring at a particular time and space, physical responsible for the dwindling of forest resources features such as land use, soils, vegetation, stream in the country (UNEP, 2001). In this context, it networks, and landforms at different time scales is important to understand the status of land use,

1 Institute of Forestry, Pokhara Campus, Nepal, *Email: [email protected] 2 District Forest Office, Pyuthan, Nepal 90 Pandey et al. Banko Janakari, Vol. 26, No. 1 measures undertaken to manage the forests, and programs laid down for the future by the national government. The available land use and forest cover data are both scanty and scattered (UNEP, 2001) most of which is not updated; a very few data have been reported for middle mountain region of Nepal (Awasthi, et al., 2005). Keeping this into consideration, this study was carried out in the Ambung VDC of middle mountain region which lacks studies on the land use dynamics and its driving factors. Materials and methods

Study area Fig. 1: Map of study area The study was conducted in the Ambung VDC situated in the northern part of Tehrathum District Data collection of Koshi Zone of Eastern Nepal (Fig. 1). The VDC lies within the Middle Mountain ecological Both primary as well as secondary data were region, and is at a distance of 3 miles from collected for the purpose of the study. The socio- Myanglung, the district headquarters. It is located economic data and information pertaining to between 27˚6’ and 27˚56’ N Latitude and between research issues were collected through field 87˚25’ and 87˚45 E Longitude. The terrain rises observation, household survey, key informant from 600 m to 2,900 m above the mean sea level; interview and focus group discussions conducted grassland and pasture are found at the higher with key persons and different groups of people altitudes, forests at the middle and settlements like community forest user groups (CFUGs), at the lower altitudes. The district exhibits sub- disadvantaged groups and mother groups. Bio- tropical type of climate at the lower altitude to physical data, satellite images (Landsat-5 TM of temperate at the higher altitudes with an average 1990 and Landsat-8 of 2013), topographic map temperature of 15˚– 25˚C and 1,650 mm annual (at the scale 1:25000) of the study area, climatic average rainfall. data and statistical data were collected from other sources such as the Institute of Forestry, This VDC is an important part of biologically very the District Forest Office, Tehrathum, the important Tinjure-Milke-Jaljale forest, which is Federation of Community Forest Users of Nepal, considered as the capital of Rhododendron. The the Department of Forest Research and Survey, major portion of the VDC is covered by forests with the Central Bureau of Statistics and the Rural major tree species like Alnus Nepalensis, Schima Reconstruction Nepal, Tehrathum. wallichi, Pinus roxburgii, Pinus wallichiana, Quercus lanata, Quercus floribunda, Castonopsis Purposive sampling with 10% intensity i.e. 80 tribuloides and Rhododendron arboretum. households incorporating different aspects of socio-economic condition especially the people The total population of the VDC is 3,613 with 1,681 including gender, ethnicity, education and males and 1,932 females; the total households geographic location were selected for the purpose being 790 (CBS, 2012). Majority of the people in of household survey. The ground-truth data the VDC are Limbu followed by Gurung, Magar, were collected with the help of a GPS set. Data Kami, Brahmin, Chhetri and Newar. Most of the analysis was done using SPSS and MS Excel people depend upon the integrated system of 2013 Packages while the Satellite image analysis farming, forests and livestock for their livelihood was performed with the help of ArcGIS 10 and as in other hilly areas of Nepal. However, in the ERDAS IMAGINE 8.4 Softwares. recent years, especially the younger people are gradually leaving their traditional occupation and Digital image processing and classification are attracted to foreign employment; particularly, Mongolian communities are attracted towards The study area was separated from the whole national as well as foreign security services. scene of the Landsat Satellite images of both

91 Banko Janakari, Vol. 26, No. 1 Pandey et al. the dates (1990 and 2013) using VDC shape file. The “extract by mask tool” of ArcGIS was used for this process to separate area of Interest (AOI) for the study. Ten training sites for each class were collected and merged to give the best representation of the class spectral reflectance. The ancillary data were converted from shape file format to Arc/Info format which is readable in remote sensing software. These files were then displayed over the Landsat scenes to aid in picking representative training sites.

Supervised classification with Maximum likelihood classifier was utilized for image classification and for the preparation of base maps for change detection (Lillesand et al., 2004). The data of different classification items (land use types) obtained from the field study were used as training samples for supervised classification of 2013 image and that of the topographical map were used for supervised classification of 1990 Image. The land use classes used for image classification were forest land, agriculture land, Fig. 2: Classified map of the study area in 1990 and others (grassland, rocks, settlement, and Similarly, the Landsat TM 2013 scene showed barren land and water bodies). that forest was the major land use with 1,195.02 Land use change detection ha (61%) followed by others 337.59 ha (22%), and agriculture 443.61 ha (17%) (Fig. 3). The raster grids of 1990 and 2013 images were overlaid using “spatial analyst” of ArcGIS 10 Software. The land use changes in terms of areas were detected using the raster calculator. Finally, the areas converted from one class to another (class) were computed. The analysis and interpretation of different aspects of the numeric data of land use change were performed with the help of Microsoft Excel 2013.

Social data analysis

The analysis and interpretation of different aspects of the social data were performed using SPSS 14 and Microsoft Excel 2013. Results and discussion Land use land cover change detection

The classification of the Landsat TM 1990 Scene showed that forest was the major land use covering 1,066.23 ha (54%) followed by others Fig. 3: Classified map of the study area in 2013 (barren, settlement, grass, rock and water bodies) 457.83 ha (23%) and agriculture 455.04 ha (23%) The study revealed that forest cover had increased in the study area (Fig. 2). whereas agriculture and others had decreased during the period of 23 years between 1990

92 Pandey et al. Banko Janakari, Vol. 26, No. 1

Table 1: Land use land cover change Land use/ Landsat 1990 Landsat 2013 Increase Decrease cover type Area (ha) % Cover Area (ha) % Cover Area (ha) % Cover Area (ha) % Cover Forest 1,066.23 53.88 1,195.02 60.47 128.79 6.6 - - Agriculture 455.04 22.99 337.59 17.08 - - 117.45 5.91 Others 457.83 23.13 443.61 22.45 - - 14.22 0.68 Total 1,979.10 100.00 1,976.22 100.00 128.79 6.6 131.67 6.59 and 2013. The change was found to be highest with a total of 3,329,885 trees grown in 2,361 ha in forest, from 53.9% to 60.5%, an increase by of private land (MFSC, 2014). 6.6% (Table 1). Gautam et al. (2002) and Balla et al. (2007) also observed similar results in their Table 2: Dynamics on land use change between studies conducted at the Upper Rosi Watershed 1990 and 2013 in the Middle Mountain region of Nepal and the Agri- Total, Land use Forest Others Galaudu and Pokhare Khola watersheds in the culture 1990 Mid-hill region of Nepal, respectively. Agriculture was found to have decreased by 5.9%, from 23.0% Forest 917.19 131.58 142.83 1,191.60 to 17.1%. Similarly, others were found to have Agriculture 75.69 155.07 105.21 335.97 decreased by 0.6%, from 23.1% to 22.5%. The Others 68.13 165.15 208.89 442.17 forests in the Middle Mountains are, in general, Total, 2013 1061.01 451.80 456.93 1,969.74 better managed and in many places forest cover have increased in recent years mainly due to the community forestry programme (Gautam et al., 2002; Niraula et al., 2013).

Land use change map

The Table 2 and Figure 4 below highlight the dynamics on land use change in the VDC within the period of 23 years. The agriculture cover was found to have decreased in the study area due to less favorable climatic condition and shortage of labor as most of the young people migrated to cities or foreign countries. This led people to shift the land use from agriculture to planting trees like Alnus nepalensis in private land, resulting increase in forest cover; this result corresponds with the Fig. 4: The LULC change map of the study previous studies. The National Biodiversity area between 1990 and 2013 Strategy and Action Plan (2014–2020) states that private forest has increased throughout the The results in Table 3 below indicate that the country. Currently (as of August 18, 2013), there performance of the classification methodology are 2,458 registered private forests in the country is quite good, reaching an overall accuracy of

Table 3: Accuracy assessment of classification Producer’s accuracy User’s accuracy Classes 1990 2013 1990 2013 Forest 97.14 100.00 98.55 93.33 Agriculture 74.42 77.78 94.12 89.74 Others 93.48 90.24 76.79 86.05

93 Banko Janakari, Vol. 26, No. 1 Pandey et al.

89.94% and 90.14% for the 1990 and 2013 were found to be dependent upon remittance and images respectively and overall Kappa statistics pension (after retirement from service). Out of the of 0.85 for both the Images. Both the accuracy total respondents, 62.5% were illiterate and only assessments indicate high Cohen’s Kappa values 37.5 % were literate. Majority of the respondents which are close to one (1). Considering the (77.5%) were dependent on fuel-wood while the individual classification accuracies, the forest in rest (22.5%) had used LPG as major sources of 2013 was most accurately classified with 100% energy. for producer’s accuracy and 93.33% for user’s accuracy. The higher accuracy of the maps could Change in forest cover and its causes be possibly due to small study area as well as Majority of the respondents (90%) believed that purposively selected smaller number of validation the forest cover was in increasing order in the sites. VDC as compared to the previous decades. About The forest cover in the study area was found to 7.5% respondents felt that it was in decreasing have increased annually at the rate of 0.5% during order while the rest (2.5%) believed that it was the period of 1990–2013 (Table 4). On the other intact. hand, the agriculture land and the others lands Majority of the respondents (52%) believed were found to have gone down by 1.3% and 0.1% that community forestry programme was the per annum. principal cause for the increase in forest cover. Table 4: Rate of LULC change of the study The community forestry programme has been area during the period of 1990–2013 generally successful in controlling or reversing the trends of deforestation and forest degradation S.N. Land use/cover Rate of change (%) in the Middle Mountains where 66.5 percent of 1. Forest 0.50 Nepal’s CFUGs are managing 910,379 ha (53.5% of the total community forest area), whereas the 2. Agriculture -1.29 programme is less successful in the Terai and 3. Others -0.14 High Mountain regions (MFSC, 2014).

(1/n) About one third (28%) of the respondents Rate of Change (%) = ((a2/a1) -1) x 100 (FAO, 2000), believed that the change in consumption pattern was responsible for the increase in forest cover.

Where, a2 = end-year data, a1 = base-year data and According to them, previously the local people n = numbers of years. had used forest as major source of fuel-wood and fodder, but now they had shifted their Social characteristics of the respondents consumption pattern and had started growing trees on their own land mainly for producing Out of the 80 respondents, 60% respondents were and selling products like timber and NTFPs and male and 40% were female. The respondent’s secondly for producing fuel-wood and fodder for categories included the middle-aged (40-60 their household consumption. Some (13%) of the years) with 70% of the total sampled population respondents believed that awareness had helped followed by 15% young-aged (below 40) and 5% in increasing the forest cover. Nowadays, people old-aged (over 60 years) belonging to all ethnic are more aware of the ecological and economical groups such as Limbu, Tamang, Gurung, Chhetri, benefits that forests provide, and thus the activities Brahmin, Newar and Kami). The study was that can damage forests are avoided by the local focused on the accumulation of past experience of people. almost 30 years back from the present, therefore the sampled population was rather purposive as A few (7%) of the respondents believed that the middle-aged category was focused. population growth (mainly due to influx of people from the higher parts to the lower of the VDC) The questionnaire survey revealed that 42.5% of had caused for decrease in the forest cover in the respondents were involved in farming (crop some parts of the VDC. and livestock), 20% in business, 12.5% in wage labor and 7.5% involved in others; the rest 17.5%

94 Pandey et al. Banko Janakari, Vol. 26, No. 1

Land use change and driving factors migration. During the period of 23 years (1990- 2013), the land use of the Ambung VDC of Majority (75%) of the respondents claimed that Tehrathum district was found to have an increase the land use of the VDC was changing and the of 128.8 ha (6.6%) cover in forest cover but a rest (25%) of the respondents did not have any decrease of 117.5 ha (5.9%) in agriculture and idea about it. Respondents were asked about 14.2 ha (0.7%) in others (barren, settlement, the possible drivers responsible for causing the grass, rock and water bodies). The results of LULC change. The result showed that 55% of the study showed that unfavorable climatic the respondents believed that migration from the condition, drying of water resources and lack VDC to city areas or abroad for several purposes of labor owing to out-migration from the VDC had led to change in land use. Many youths had resulted in diverting the local people of the migrated to other areas so as to earn more money VDC from agriculture to planting trees like A. resulting in scarcity of working manpower in the nepalensis in their agriculture lands as well as village. So, most of the households changed their in other lands within the VDC. The VDC was agricultural lands from cultivating agricultural found to have a severe problem of rapid drying of crops to growing trees like Alnus nepalensis (Utis) water resources in spite of the increase in forest and Cardamom (Alainchi) under the tree cover, cover, and so research should be carried out to resulting in the increase in the forest cover and find out the reason and solve the problem before decrease in the agriculture area within the VDC. A it is too late. Besides, the use of Remote Sensing slightly more than half (52%) of the respondents and GIS should be increased and diversified for believed that infrastructure development like road monitoring natural resources for better results and was responsible for the change in the LULC of easy updates. Further, the local people especially the VDC. Nearly half (45%) of the people felt that the local youths should be provided with income- policies like introduction of community forestry generating activities for their livelihood so as to programme, energy-efficient stoves, stall feeding retain them in their villages, which in turn might system and so on in the VDC were responsible for cease drastic land use changes in the rural areas the change in the LULC. Similarly, almost half causing adverse effects. (41%) of the respondents claimed that climate change was directly or indirectly responsible for Acknowledgements the change in the LULC of the VDC while some (10%) of the respondents believed that increase This paper is based on a part of the author’s in population of the area was responsible for the B.Sc. Forestry Thesis submitted to the Institute of change in the LULC of the VDC. Forestry, Pokhara Campus, Nepal. We are thankful to Rural Reconstruction Nepal/Multi Stakeholder Analyzing the population of three decades, it was Forestry Program (RRN/MSFP) and WWF, Nepal found that there had been decrease in the total for providing Financial and technical support to population of the VDC in the recent years. Both conduct the study. We express our cordial thanks the male and female populations had increased to Dr. Krishna Raj Tiwari, Mr. Yajna Prasad from 1981 to 2001, but the total population Timalsina, Mr. Navin Kumar Yadav for their had decreased from 2001 to 2011 as there was critical suggestions during the study. Similarly decrease in the male population from 1,973 in DFO- Tehrathum, FECOFUN, Forest Action 2001 to 1,681 in 2011. It was verified by the local and RRN, Tehrathum and all the respondents of people that many young males had left the village Ambung, Tehrathum are also duly acknowledged in search of jobs and opportunities. for their cooperation in accomplishing the study. Conclusion References

Out-migration, infrastructure development, Agrawal, A. and Ostrom, E. 2001. Collective climate change and policy are the major driving action, property rights, and decentralization forces for the change in land use and land cover. in resource use in India and Nepal. Politics However in the case of the Ambung VDC of and Society 29 (4): 485–514. Tehrathum district, the main reasons for the same were unfavorable climatic condition, drying of Awasthi, K. D., Sitaula, B. K., Singh, B. water resources and lack of labor owing to out- R., Balla, M. K., Bajrachrya, R. M., and

95 Banko Janakari, Vol. 26, No. 1 Pandey et al.

Dhoubhadel S. P. 2005. Analysis of land use Gautam, A. P., Webb, E. L. and Eiumnoh, A., structure in two mountain watershed of Nepal 2002. GIS assessment of land use-land cover using FRAGSTATS. Forestry 13: 1–18. changes associated with community forestry implementation in the Middle Hills of Nepal. Awasthi, K. D. 2004. Land Use Change Effects on Mountain Research and Development 22 (1): Soil Degradation, Carbon and Nutrient Stock 63–69. Sand Greenhouse Gas Emission in Mountain Watersheds. PhD Thesis, Agricultural Gautam, A. P. 2007. Land Use Dynamics University of Norway, Norway. and Landscape Change Pattern in a Mountain Watershed in Nepal. http:// Awasthi, K. D., Sitaula, B. K., Singh, B. R. www.gisdevelopment.net/application/ and Bajrachrya, R. M. 2002. Land-use environment/overview/envo007.htm change in two Nepalese watersheds: GIS accessed on 18 Oct, 2014. and geomorphometric analysis. Land Degradation and Development 13: 495–513. Lilisand, T. M. Kiefer, R. W. and Chipman, J. W. 2004. Remote Sensing and Image Balla, M. K., Awasthi, K. D., Singh, B. K. and Interpretation. John Wiley and Sons, New Pradhan, B. M. 2007. Land use changes York, USA. and geomorphometric analysis in Galaudu and Phokhare Khola watersheds in mid-hill MFSC. 2014. National Biodiversity Strategy region of Nepal. International Journal of and Action Plan (2014–2020). Government Ecology and Environmental Sciences 33 (2– of Nepal, Ministry of Forests and Soil 3): 171–182. Conservation (MFSC), Kathmandu, Nepal.

CBS. 2012. National Population and Housing Niraula, R. R., Gilani, H., Pokharel, B. K. and Census 2011 (Village Development Qamen, F. M. 2013. Measuring impacts of Committee/Municipality). Central Bureau community forestry programme through of Statistics (CBS), Kathmandu, Nepal, 2. repeat photography and satellite remote sensing in of Nepal. Journal DoF. 2005. Forest Cover Change Analysis of Environment Management 126: 20–29. of the Terai Districts (1990/91-2000/01). Department of Forests (DoF), Kathmandu, Phong, L. T. 2004. Analysis of Forest Cover Nepal. Dynamics and their Driving Forces in Bach Ma National Park and Buffer Zone using FAO. 2000. Global Forest Resources Remote Sensing and GIS. MSc Thesis Assessment 2000. Food and Agriculture submitted to the International Institute Organization of the United Nations, Rome, for Geo-information Sciences and Earth Italy. Observation, Enschede, The Netherlands.

FRA/DFRS. 2014. Terai Forests of Nepal (2010 UNEP. 2001. Nepal: State of the Environment – 2014). Forest Resource Assessment Nepal 2001. United Nations Environment Program, (FRA), Department of Forest Research and Regional Resource Centre for Asia and the Survey (DFRS), Kathmandu, Nepal. Pacific, Thailand.

96 Short Note Comparison of forest cover mapping results of two successive forest resource assessments of Nepal

S. Khanal1*, B. S. Poudel1, P. Mathema1, Y. P. Pokharel1 and D. K. Kharal1

n Nepal, the first national-level forest inventory FRA, it is difficult to conclude that the forest area Iwas carried out in the 1960s (FRS, 1967). Since has increased between the two assessments. then, several forms of forest resource assessment activities have been carried out in different This paper aims to briefly highlight the problems periods, each different in terms of purpose, scale, associated with comparison of results between the scope, design and technology used. Six national- two assessments. One key issue is the difference level forest cover assessments were carried out in in materials and methodologies used. With the the last four decades (DFRS, 1999; 2015). development of science and technology, newer methods are being developed and are improving Results of nation-wide forest resource assessment the accuracy of forest parameters estimation. (2010–2014) of Nepal were recently published The latest assessment applied a set of materials (DFRS, 2015). As per the assessment, Foresta and methods more advanced than the previous covers 5.96 million ha (40.36%), Other Wooded one. The comparison of the results in terms of Landb covers 0.65 million ha (4.38%) and Other forest resource estimates is problematic due to Landc covers 8.16 million ha (55.26%). Forest differences in methods, materials, duration of and OWL together comprise 44.74% of the total assessments as well as validation approach used. area of the country. The previous nation-wide One analogous example includes a recent global forest resource assessment (NFI, 1994) was estimate of total tree number as 3.04 trillion which done in 1990s (DFRS, 1999). The forest area as is more than seven times the previous estimate estimated by NFI (1994) was 29% (4.27 million done in 2008 (Crowther et al., 2015). Certainly, ha) and shrub 10.6%, making a total of 39.6% of this doesn’t necessarily imply increase in tree the geographical area of the country. Both of these number, but could be attributed to improvement nation-wide forest resource assessments were in the methodology adopted. conducted by the Department of Forest Research and Survey, with support from the Government of The FRA of 2010s has some key differences as Finland. One of the key interests after successive compared to the NFI of 1990s. All the sample assessments is the change in forest parameters plots measured in the recent FRA have geo- between the assessment periods. In this context, referenced locations, and are set up as permanent the forest area estimated by the recent FRA sample plots (PSPs). Besides, the recent FRA is more than that of the NFI of 1990s which is more comprehensive with a scope of re- may be attributed to three factors: (1) higher assessment. Furthermore, the recent FRA is a mapping resolution of FRA (ii) abandonment multi-source forest resource assessment, as it of agricultural land, which in turn changed to included additional variables (soil characteristics, forested land, and (iii) the community forestry soil carbon, litter and dead wood, stump and interventions (DFRS, 2015). However, given disturbance) in addition to tree parameters. the methodological differences between NFI and Unlike the previous assessment, which excluded

1 Department of Forest Research and Survey, Babarmahal, Kathmandu, E-mail: [email protected] a Forest is defined as an area of land at least 0.5 ha in area and a minimum width/length of 20 m with a tree crown cover of more than 10% and tree heights of 5m at maturity; b Other Wooded Land (OWL) includes: (i) the land not classified as forest spanning over more than 0.5 ha, having at least 20m width and with 5-10% tree canopy cover, (ii) the land with less than 5% tree canopy cover, but the combined cover of more than 10% shrubs, bushes and trees; and (iii) the areas of shrubs and bushes where no trees are present; and c Other Land (OL) refers to all other land areas that are not classified as Forest or Other Wooded Land

97 Banko Janakari, Vol. 26, No. 1 Khanal et al. the protected areas (PAs), this assessment had pixel-based and the object-based image analysis. sample plots across the entire country including Further, point sampling using aerial photographs the PAs. potentially offers quick and cost effective method for area calculation, but the accuracy depends on The key differences between the two national appropriate sample design. On other hand, the forest assessments of 1990s and 2010s can be hybrid approach used in the recent assessment has summarized under materials, methods, duration been recommended as simple, robust and cost- of assessment and verification approach (Table effective (GOFC-GOLD, 2008). The compilation 1). Scale of the base data used for mapping is one of results from different sources in the 1990s NFI important determinant of the comparability of posed yet another challenge to compare results output products. Since the resolution of the image against a uniform approach applied in the recent (Landsat TM) used in the 1990s for a portion of FRA (Table 1). Consistency in data collection and Nepal is 30m, the smaller forest patches could analysis may be an important issue in assessment have been excluded from forest cover mapping. involving compilation from different sources. A However, high-resolution image offers several Minimum Mapping Unit (MMU) is defined as advantages over low-resolution images for forest “the smallest size areal entity to be mapped as cover mapping, e.g. the ability to map smaller a discrete entity” (Lillesand et al., 2014). As the patches. On the other hand, the season of image MMU determines the extent of detail in the map acquisition has impact on the detectability of (Saura, 2002), it is one of the most critical issues vegetation. One potential issue with the aerial in comparing the two assessments. The FRA had photographs acquired in December–January in a much smaller MMU than the NFI. The higher the 1990s could be snow cover that could affect mapping resolution (or lower MMU) generally the interpretation of forest cover especially results in increased forest area as compared to the in the High Mountains and the High Himal lower mapping resolution since small patches that physiographic regions, while the issue with are not visible on lower-resolution images can be March–April image used in the 2010s assessment mapped on higher-resolution ones. This could be could be the defoliation of some deciduous one of the reasons why the estimate of forest area tree species that makes forest cover mapping was more in the latest assessment. challenging. Despite the difficulty in direct comparison with Method of forest mapping is probably the most previous assessment, the latest assessment did important issue when we want to look for changes. establish a baseline for a range of forest resource The results of the two assessments are not directly assessment parameters. comparable due to the differences between the Table 1: Comparison of forest assessments in the 1990s and 2010s NFI 1990’s FRA 2010’s Description Materials Base data for Combination of aerial photos and Wall-to-wall coverage The differences in data sets forest area satellite image (Landsat TM); Aerial of Rapideye images used to map forest cover mapping and photo covered 83.7% of Nepal’s area makes comparison difficult. interpretation while satellite image covered 16.3% Scale Aerial photo at 1:50000 scale and 5 m spatial resolution The resolution determines remaining area covered by Landsat the ability to map smaller TM Satellite Image of 30 m spatial patches. resolution a a December–January March–April Season of image acquisition acquisition affects the detectability of month vegetation. Methods Forest area Point sampling using aerial photos Wall-to-wall mapping The methods used for forest estimation and visual interpretation of grid through integration of area estimation are entirely system for 51 districts. advanced object-based different.

98 Khanal et al. Banko Janakari, Vol. 26, No. 1

image classification; classification and regression tree (CART) and extensive visual interpretation using high-resolution Google Earth Image. Sources of data Compilation of work done by different FRA Nepal single work Consistency in data organizations (Satellite image analysis: collection and analysis would NFI- 16.3% of area, District Forest pose an important issue in Inventory- 8.5% of area, Churia comparison. forest inventory- 3.1% of area and the remaining 51 Hill Districts- 72.1% of area) Minimum Varied based on methods used, ranging 0.5 ha Marked differences in MMU mapping unit from 1 to 25 ha. Satellite Image will have implications on (MMU) Analysis: >9 ha (assumed 10 pixels); the mapping output, leading District Forest Inventory: 6.25 ha and to difficulty in comparison 25 ha as 1:25000 and 1:50000 scale between the assessments. aerial photos were used; Churia Forest Inventory: 6.25 ha; Remaining Hill Districts: 1 ha. Duration Assessment 1987–1998 (12 years) 2011–2014 (4 years) Changes might have occurred period due to the long duration of assessment in the case of NFI. Verification Field Not done The mapping results Field verification is important verification were validated against to assure validity and field data for Terai, reliability of any mapping Churia, Middle work; it was lacking in Mountains; for High NFI (1994), was conducted Mountains and High Field verification in 3 Himal, verification physiographic regions in done by using high FRA (2010–2014). resolution images in Google earth. Accuracy Done for hilly area (72.09 % of total The forest cover The confidence limit of assessment area with 7,685 grids; the forest and mapping accuracy was forest statistics could not be shrub area combined estimate being evaluated in terms determined due to wall-to- 37.7% with a 95% confidence limit of of overall accuracy wall mapping in the FRA 1.1%; thus, in terms of percentage, the (85.16%) and kappa (2010–2014). area can vary from 36.6% to 38.8%. (0.72 with standard error of 0.0175). This assessment, however, doesn't provide confidence limits of area estimates. Verification Not done; the results reported from the The results were also The estimates from the with point sampling only. obtained by using latest assessment seem independent independent method reliable as the mapping approach through visual interpre- was verified with extensive tation of regular grids visual interpretation and field of points (>55,358) at 4 verification (DFRS, 2015). km interval throughout the country.

99 Banko Janakari, Vol. 26, No. 1 Khanal et al.

Subsequent future assessments will produce FRS. 1967. Forest Statistics for the Terai and information on forest cover change. However, to Adjoining Regions. Publication No. 4. measure changes and track the impact of forestry Forest Resources Survey (FRS), Kathmandu, sector policy and programme interventions, it is Nepal. necessary to compare forest cover with consistent GOFC-GOLD. 2008. Reducing greenhouse methods and datasets in at least smaller areas, if gas emissions from deforestation and not possible for the whole country. degradation in developing countries: a References sourcebook of methods and procedures for monitoring, measuring and reporting, Crowther, T. W., Glick, H. B., Covey, K. R., GOFC-GOLD Report version COP13- Bettigole, C., Maynard, D. S., Thomas, S. 2. Global Observation for Forest Cover M., Smith, J.R., Hintler, G., Duguid, M. and Land Dynamics Project Office, Natural C., Amatulli, G. and Tuanmu, M. N. 2015. Resources Canada, Alberta, Canada. Mapping tree density at a global scale. Nature 525 (7568): 201–205. Lillesand, T., Kiefer, R. W. and Chipman, J. 2014. Remote Sensing and Image Interpretation. DFRS. 1999. Forest Resources of Nepal (1987– John Wiley and Sons, New York, USA. 1998). Department of Forest Research and Survey (DFRS), Kathmandu, Nepal. Saura, S. 2002. Effects of minimum mapping unit on land cover data spatial configuration DFRS. 2015. State of Nepal’s Forests. Forest and composition. International Journal of Resource Assessment (FRA) Nepal, Remote Sensing 23 (22): 4853–4880. Department of Forest Research and Survey (DFRS), Kathmandu, Nepal.

100 Acquisition Banko Janakari, Vol. 26, No. 1

Central Forest Library Acquisition List No. 64, 2016

APFISN. 2013. Invasive Alien Plants in the ICIMOD; RSPN (2014). An Integrated Forests of Asia and the Pacific. Asia-Pacific Assessment of the Effects of Natural and Forest Invasive Species Network, FAO, Bangkok. Human Disturbances on a Wetland Ecosystem: 213p. A Retrospective from Phobjikha Conservation Area, Bhutan. International Centre for Integrated ACIAR. 2015. Partners in Research for Mountain Development/ Royal Society for Development Linking Farms to Markets the Protection of Nature, Bhutan. 57p. Value of Bio-secure Produce Market-Driven Innovation. Australian Centre for International IUCN. 2014. IUCN Nepal Review. International Agricultural Research, Australia. 31p. Union for Conservation of Nature, Kathmandu. 24p. Bhatt, D. P. 2015. Ecotourism in Nepal: Concepts, Principles and Practices (2nd ed.). MFSC. 2015. Strategy and Action Plan 2015- Anju Bhatt, Kathmandu. 258p. 2025 Terai Arc Landscape, Nepal. Ministry of Forests and Soil Conservation, Kathmandu. 78p. CBS. 2013. National Sample Census of Agriculture Nepal, 2011/2012 National Report. MSFP. 2014. Synthesized Forestry Sector National Planning Commission Secretariat, Gender Equality and Social Inclusion (GESI) Central Bureau of Statistics, Kathmandu. 113p. Policy BRIEF. Multi-stakeholder Forestry Programme (MSFP), Kathmandu. 18p. Chettri, N., Bubb, P., Kotru, R., Rawat, G. and Ghate, R. 2015. Long-term Environmental and MSFP. 2014. Assessment of Implementation Socio-ecological Monitoring in Transboundary Status of Forestry Sector Gender Equality Landscapes: An Interdisciplinary and Social Inclusion Strategy 2065. Multi- Implementation Framework. ICIMOD, stakeholder Forestry Programme, Kathmandu. Kathmandu. 24p. 52p.

Devkota, N. R. 2014. Research Reports MOSTE. 2015. Nepal Earthquake 2015 Rapid (Volume I). Agriculture and Forestry University. Environmental Assessment. Ministry of Science, Directorate of Research and Extension, Chitwan. Technology and Environment, Kathmandu. 84p. 270p. NTNC. 2014. Profiles of the Greater One- Dogra, P. 2011. CSWCRTI Vision 2030. Central horned Rhinoceros (Rhinoceros unicornis) Soil and Water Conservation Research and of Bardia National Park and Shuklaphanta Training Institute, Dehradun. 36p. Wildlife Reserve, Nepal. National Trust for Nature Conservation, Kathmandu. 56p. Dukpa, K. 2011. Forestry Development in Bhutan: Policies, Programmes and Institutions. Mandal, R. A., Dhakal, S. R. and Hamal, N. 2015. Department of Forests and Park Services, Bhutan. Proceedings of the Sixth National Community 125p. Forestry Workshop. Department of Forests, Kathmandu. 490p. Fraser-Jenkins, C. R., Kandel, D. R. and Pariyar S. 2015. Ferns and Fern-Allies of Nepal (Vol. NPQP. 2014. Technical Guideines for Detection 1). National Herbarium and Plant Laboratories, Survey of Plant Pests in Nepal. National Plant Lalitpur. 492p. Quarantine Programme, Lalitpur. 34p.

ICIMOD. 2014. An Integrated Assessment of Ning, W., Ismail, M., Joshi, S., Qamar, F. M., the Effects of Natural and Human Disturbances Phuntsho, K., Weikang, ?, Khan, B., Shaoliang, Yi., on a Wetland Ecosystem: A Retrospective Kotru, R. and Sharma, E. 2014. Understanding from Koshi Tappu Wildlife Reserve, Nepal. the Transboundary Karakoram-Pamir International Centre for Integrated Mountain Landscape. ICIMOD, Kathmandu. 84p. Development, Kathmandu. 75p.

101 Banko Janakari, Vol. 26, No. 1 Acquisition

Piketty, T. 2014. Capital in the Twenty-First and Forest Carbon in Community Forests of Century. Harvard University Press, London. Terai, Bhawar and Chure (A Case Study from 685p. ). IOF/ComForM Project, Pokhara. 22p. Pradhan, N. and Dhakal, M. 2014. Adapting to Climate Change for Sustainable Agribusiness Takayuki, K., Ryuichi, T., Satoshi, T., Masahiro, in High Mountain Watersheds: A Case Study and K. March 2015. Bulletin of the Forestry and from Nepal. ICIMOD, Kathmandu. 47p. Forest Products Research Institute. Forestry and Forest Products Research Institute, Japan. PCTMCDB. 2015. Chure Conservation Area 72p. (An atlas of 36 districts). President Chure-Terai Madhesh Conservation Development Board, Kotru, R., Choudhary, D., Fleiner, R., Khadka, Kathmandu. 144p. M., Kant, P. and Appanah, S. 2013. Guidelines for Formulating National Forest Financing Pathak, R., Gautam, N. P., Shrestha, H. L. Strategies. ICIMOD, Kathmandu. 164p. and Khadka, A. 2014. Analysis of Adapation Strategies to Climatic Hazards (Case Study Pradhan, R., Khakurel, B., Chapagain, N., from Barpak VDC, Gorkha). IOF/ComForM Basnet, R., Lamichchane, D. and Jha, L. 2014. Project, Pokhara. 35p. Plant Resources (A Scientific Publication). Department of Plant Resources, Kathmandu. Pyakuryal, B. 2013. Nepal’s Development 132p. Tragedy: Threats and Possibilities. Fine Prints, Kathmandu. 339p. . Ramakrishnan, P. S., 2010. Primer on  Notice  Characterising Biodiversity: Trans- Disciplinary Dimensions. National Book Trust, The Central Forest Library requests India. 66p. all researchers, academicians and Rasul, G., Hussain, A., Khan, M. A. and Jasra, A. students to send in a copy of their W. 2014. Towards a Framework for Achieving research papers, theses, books and Food Security in the Mountains of Pakistan. ICIMOD, Kathmandu. 22p. reports related to forestry, wildlife, botany, soil conservation, socio- Siwakoti, M. and Rajbhandary, S. 2015. economic studies, medicinal plants, Taxonomic Tools and Flora Writing. Tribhuvan University, Kathmandu. 210p. environment, biodiversity and related topics to this library. SFC. 2012. Proceedings of Workshop on Stakeholder Engagement to Enhance Development and Productivity of Trees l o Outside Forests (TOF) Colombo, Sri Lanka 14–16 November 2012. SAARC Forestry Centre, Department of Forest Research and Survey, Thimpu. 110p. P.O. Box: 3339, Kathmandu, Nepal E-mail: [email protected] Shah, K. B., Ale, S. B. and Adhikari, A. 2015. Snow Leopard Environment Camp Held in [email protected] Upper Mustang. Snow Leopard Conservancy Tel. No. 4220482/4269491 (Nepal Program) and National Trust for Nature Conservation, Lalitpur. 39p.

Sankaran, K.V. and Suresh, T. A. 2013. Invasive Alien Plants in the Forests of Asia and the Pacific. FAO/UN, Bangkok. 213p.

Yadav, Y., Dutta, I. C., Mandal, R. A. and Acharya, N. 2014. Comparison of Bio-Diversity

102 Key Word Index to Vol. 26, No. 1, 2016

Key word Page no. Key word Page no. Aegle marmelos 32 Land use types 3

Albizia lebbeck 82 Landsat 38

Bagmati River Basin 53 Line transect 60

Bardia 60 Machine learning algorithm 38

Basal area 78 Medicinal plants 45

Biomass carbon 24 Microbial inoculant 82

Biodiversity 45 Middle Hills 70

Camera trap 60 Mikania micrantha 78

Canonical Correspondence Analysis 3 Multivariate analysis 3

Carbon stock density 24 Object-based classification 38

Churia 1 7 Object-based Image Classification 53

CLASlite 17 Plant diversity 78

Community 45 Potential plantation area 53

Community forests 70 Regeneration protection 32

Conservation 45 Remote sensing 90

Deforestation 17 Rural road 70

Density 60 Seedling growth 82

Ecological regions 24 Shorea robusta 24

Elevation 3 Soil organic carbon 24

Enterprise 32 Species richness 3

Forest cover 90 Spectral Mixture Analysis 17

Forest degradation 17 Supervised classification 90

Generalized Linear Model 3 Terai 53

Geographic Information System 90 Tiger 60

Germination 82 Wild Prey 60

Invasion 78 Woody plant species’ diversity 70

Land reclamation 53

103

Guidelines for contributors

Banko Janakari is a refereed journal. Research articles, short notes, or book reviews may be submitted. The submissions are accepted on the understanding that the editor may shorten or alter them in an appropriate way. In the case of research articles, authors will be consulted as far as possible depending on the difficulties of communication. Please state whether an article submitted to Banko Janakari, relates to work that has been commissioned by any government or other organization, or done in any connection with a course of study at a university or other institution. The author should also declare that the submitted article has not been published elsewhere, or submitted to another journal for publication.

Articles should be of interest to the broad categories of forestry professionals in general and be understandable to non-specialist also. Articles should be written in plain and concise language and jargons should be avoided. Technical terms that may be unfamiliar to readers should be defined when they appear for the first time. Footnotes should be avoided as far as possible. Manuscript should be in English, typed in double space and submitted in duplicate. However, manuscript in electronic format or as e-mail attachment is preferred to hard copies. The length of the manuscript should be between 3000 and 4000 words for longer articles, and 1000 and 2000 for short notes. However, shorter or longer articles may be considered in certain cases. Articles should follow the journal’s format: a) Abstract b) Introduction c) Materials and methods d) Results and discussion e) Conclusion and f) References.

The first page of the article should provide the full name, title and complete address of the author(s) including e-mail address. Subsequent pages should be numbered sequentially. The article should commence with a concise and informative abstract in one paragraph without reference to text or figures. It should be about 200 words. The abstract should be followed by four to six keywords. Authors are requested to provide supporting illustrative materials (tables, graphs, maps and drawings) with manuscripts. Such materials must be numbered and supplied on separate sheet. Photographs should be of high resolution and good contrast. All measurements should be given in the metric system. Exchange rate of Nepali Rupees with US dollars is a must where monetary figures are supplied. All the tables more than half of A4 size paper and those other than results should be kept as annexes. Hatching and patterns should indicate different divisions of maps and graphs instead of different colours.

References should be given at the end of the article on separate sheet of paper. Reference cited in the text must be listed alphabetically. References used in the text should indicate the name of the author(s) and date of publication at appropriate points (e.g. Sarkeala, 1995; Dennis et al., 2001), with the full reference given in a separate list at the end of the article. The term et al. should be used in the text when there are three or more authors. However, complete list of the authors should be named in references. For the convenience of authors, some examples of citation of different publications are given below. Please also note the case, italicization and bold typeface etc. of the referencing style given below. When information from the Internet is cited, full address of the website and the date accessed should be included, and listed under References.

Journal articles Dennis, R., Hoffman, A., Applegate, G., Von Gemmingen, G. and Kartawinata, K. 2001. Large-scale fire: creator and destroyer of secondary forests in Western Indonesia. Journal of Tropical Forest Science 13 (4): 786-799.

Books Draper, N. R. and Smith, H. 1981. Applied Regression Analysis. 2nd edition. Wiley and Sons, New York, USA.

Edited books/Proceedings Tsuchida, K. 1983. Grassland vegetation and succession in eastern Nepal. In Structure and Dynamics of Vegetation in Eastern Nepal (ed) Numata, M. Chiba University, Chiba, Japan, 47-87.

Reports EPC. 1993. Nepal Environment Policy and Action Plan: Integrating Environment and Development. Environment Protection Council, Kathmandu, Nepal.

Paper (Seminar/Workshop) Alder, P. and Kwon, S. 1999. ‘‘Social Capital: The Good, The Bad and The Ugly’’ paper presented at the Academy of Management, Chicago, USA.

Thesis Shrestha, S. M. 1993. Comparison of Different Sampling Techniques in Forest Inventory in Southern Nepal. M.Sc. Thesis, University of Joensuu, Joensuu, Finland. 1 Banko a nakari, ol. , No. 1

Contents

Editorial 1 Articles S. K. Rai, S. Sharma, K. K. Shrestha, J. P. Gajurel, S. Devkota, M. P. Nobis and C. Scheidegger o ni ron n on i ri n an o o i ion o a l ar l an in ana l on ra ion r a an aa r a a r i on o N a l i alaa

S. Khanal and A. Khadka ai n or a ion an or r aa ion in li ar oa i n a rn r ia o N a l 1

H. P. Pandey and M. Bhusal o a ra i on aron o k in al Shorea robusta or in o i r n oloi al r i on o N a l 2

K. Baral and B. R. Upreti or a n o B l Aegle marmelos an o n iali o al i i r o in n rr i in ana n i ri o N a l 2

A. K. Chaudhary, A. K. Acharya and S. Khanal Forest type mapping using object-based classification method in Kapilvastu district, Nepal

C. B. Khadka, A. L. Hammet, A. Singh, M. K. Balla and Y. P. Timilsina oloi al a an i r i i ni o an a l Dactylorhiza hatagirea an i a o ia in Lete village of Mustang district, Nepal

A. K. Acharya, A. K. Chaudhary and S. Khanal Identification of land reclamation area and potential plantation area on Bagmati ri r a in in rai r i on o N a l

J. B. Karki, Y. V. Jhala, B. Pandav, S. R. Jnawali, R. Shrestha, K. Thapa, G. Thapa, N. M. B. Pradhan, B. R. Lamichane and S. M. Barber-Meyer Estimating tiger and its prey abundance in Bardia National Park, Nepal 6

J. K. KC and A. P. Gautam Impact of roads on biodiversity: a case study from Karekhola rural road in Surkhet i ri o N a l

S. Basnet, D. B. Chand, B. H. Wagle and B. Rayamajhi lan i r i a n an r r o a ri on o Mikania micrantha ina a n non ina roi al Shorea robusta or

B. M. Khan, M. A. Kabir, M. K. Hossain and M. A. U. Mridha Microbial inoculant influences the germination and growth ofAlbizia lebbeck l in in nr r 2

D. Pandey, B. P. Heyojoo and H. Shahi Drivers and dynamics of land use land cover in Ambung VDC of Tehrathum district, Nepal

Short Note S. Khanal, B. S. Poudel, P. Mathema, Y. P. Pokharel and D. K. Kharal o a ri on o or o r ai n r l o o i or r or a n o N a l

Central Forest Library Acquisition List 1 1

Key word Inde 1