Development of 2010 national land cover database for the Nepal
Graphical abstract
Introduction
In the last few decades the Hindu Kush Himalayas (HKH) has undergone rapid economic, social, and environmental changes. However, there is a lack of cohesive information on these changes and how they are impacting on land cover and land cover change. Nonetheless, it is clear that land cover change in the HKH is driving change in ecosystems and their services (Koschke et al., 2012). The HKH region extends over 3500 km encompassing all or parts of eight countries: Afghanistan, Bangladesh, Bhutan, China, India, Myanmar, Nepal, and Pakistan. The region contains 10 of Asia's largest river systems, which provide water and ecosystem services to the 210 million people living in mountain areas, as well as the 1.3 billion people downstream (Molden and Sharma, 2013). The region is extremely fragile in terms of land cover diversity and its association with variable terrain, climate, and socio–demographic interactions. The HKH region is significantly rich in terms of biodiversity, but is also one of the least studied in the world (Sharma and Chettri, 2005). The Intergovernmental Panel on Climate Change (2007) has recognized the HKH region as a ‘data-deficit area’. Although scientists and institutions are attempting to fill some of the gaps, reasonable and reliable sources of data for the development of accurate land cover maps for the HKH are scarce. The available data in the region are sporadic, inconsistent and inaccessible (Bajracharya et al., 2010).
Nepal has a high-level of diversity and complexity in terms of altitude, terrain, biodiversity, and socio-demography and is broadly representative of the land cover diversity in the HKH region (Bhattarai et al., 2009). There is a need to understand the interactions between these diversities to support land resources use, development, and conservation (Zomer and Susan, 2001). Climate Change impacts, habitat fragmentation, and high population density are changing in the way people in Nepal (and the HKH) are using land and causing land use conflicts. These multiple drivers of change and the interactions between them need to be understood so that policy makers and planners can better manage Nepal's natural resources.
According to the 2011 census, Nepal has a total population of 26.5 million, with a population growth rate of 1.35% per annum. The overall literacy rate (for the population aged 5 years and above) has increased from 54.1% in 2001 to 65.9% in 2011 (National Population and Housing Census, 2012). Nepal's current forest policy and legislation classifies the country's forests mainly according to tenure or control over the land as government-managed, community-managed, leasehold, religious, private, and protected forest (Acharya, 2002). According to the Food and Agriculture Organization of the United Nations (2010) country report, the forest living biomass (above and below ground biomass) is 484 million metric tonnes (359 million metric tonnes above and 126 million metric tonnes below).
Satellite remote sensing is as an important tool in providing reliable historical and current land cover information at the local, national, regional, and global levels (Foley et al., 2005). At the global level, numerous efforts have been made to provide satellite-based land cover and forest cover information (Reis, 2008, Schweik et al., 1997, Xian et al., 2009), including GLOBCOVER (Arino et al., 2007, Bontemps et al., 2011), annual MODIS land cover (Friedl et al., 2002, Friedl et al., 2010), MODIS VCF (DiMiceli et al., 2011), the rescaling of MODIS VCF at 30 m (Sexton et al., 2013) and detection forest cover changes provided by Global Forest Watch (World Resources Institute, 2013). Public domain satellite data and online visualization tools like Google Earth and BHUVAN allow end users to assess the accuracy of land cover data based on very high resolution satellite images and observations.
In Nepal, institutional-level national land cover assessments were conducted in 1963, as part of a forest resources survey (FRS) using aerial photography, and in 1986 for a land resources mapping project (LRMP) using satellite data. Since then, no national-level land cover assessments have been conducted and subsequent assessments have focused only on national forest cover mapping, for example, the multi-stakeholder forestry programme (MSFP) and national forest inventory (NFI). A number of individual researchers have also tried to fill the land cover data gap in their own capacity at various scales (Bhattarai et al., 2009, Carson et al., 1986, Gautam et al., 2002, Jackson et al., 1998, Niraula et al., 2013), but none of these have produced land cover maps with national coverage using standardised classification scheme.
In Nepal, landscape changes and social change patterns have been observed as a function of land use change, and these have implications for social and ecosystem functions and services (Millette et al., 1995). Nepal's community forestry programme is acknowledged to be one of the most successful forest conservation initiatives in the world (Niraula et al., 2013). However, despite the success of this programme and the importance of forests in supporting livelihoods of the people of Nepal and providing ecosystem services to those downstream, there has been little research on land cover and land cover change. A comprehensive understanding of the changing patterns of land cover over the last two decades and its drivers at the national and sub-national level is lacking. This lack of data and information has been one of the major limitations on policy and decision makers in addressing regional environmental issues including the development of greenhouse gas (GHG) inventories, the evolution of reducing emissions from deforestation and forest degradation (REDD) mechanisms, and the assessment of land degradation, as well as optimal land use planning (Dangi, 2012). This study is expected to be useful in addressing such regional issues and informing initiatives in relation to Nepal's national and global commitments, such as its communications to the United Nations Framework Convention on Climate Change (UNFCCC).
The present study on land cover assessment of 2010 is taken up as part of regional initiative on developing consistent and harmonized temporal land cover databases over HKH region. At the initial level, study was conducted using public domain Landsat TM data of 2010 and 2011 by adoption of geographic object based image analysis (GEOBIA) classification technique. The land cover product validation system is developed as part of the study using online web based tool. The assessment of land cover patterns in relation to historical trends and implications over natural resources management over different physiographic regions and potential application for different national and global commitments initiatives also described. Considering the number of global land cover datasets and studies are available, we compared our land cover product with global product of Gong et al. (2013) to explore the possible adoption of global algorithms for national monitoring systems. In this study the forest fragmentation and edge effects was calculated by dividing the land cover into forest and non-forested areas. An online crowd source-based validation tool was developed to collect and analyse feedback from voluntary participant.
Section snippets
Study area
The study area covers the whole of Nepal, which falls between latitudes 26°22′N to 30°27′N and longitudes 80°04′E to 88°12′E and shares an international border with China to the north and India to the south, east, and west. With a total land area of 147,181 km2. Nepal is divided into five physiographic regions: High mountain, Middle mountain, Hill, Siwalik and Tarai (Fig. 1). Administratively, Nepal has 75 districts and 4057 village development committees (VDCs). These 75 districts are divided
Data and software used
For land cover mapping, Landsat TM satellite images of 30 m spatial resolution of 2009, 2010 and 2011 were used (Table 1). Altogether, 11 scenes (185 × 185 km each) of Landsat TM were used covering the entire study area. In some cases, alternative images of different season of the same row and path were deployed to ensure the accurate identification of land cover features. The satellite images of November 2010 to February 2011 period were chosen representing mostly green to semi deciduous
Land cover map
The study produced land cover statistics and a land cover map of Nepal showing the LCCS-based 12 classes (Table 2 and Fig. 2). The overall accuracy of the classification was estimated at 85.13% with producer's and user's accuracies at 81% and 82%, respectively. Overall, a 0.82 kappa statistics value was obtained, indicating a high level of accuracy. The classification accuracy of needleleaved open forests was found to be relatively low and there was a low level of separability between open and
Discussion
The adaption of the eCognition-based classification method used in this study has helped in integrating ancillary information, developing scene-specific standards and knowledge, and achieving better classification accuracy. This study limited the classification scheme to broad land cover classes, without delineating sub-classes, such as plantations, orchards, settlements, or a higher number of forest density classes, which are amenable at Landsat based TM resolution. This means that the
Conclusion
The main value of Nepal land cover 2010 is its ability to provide a complete, consistent, and harmonized national land cover map. The broad classification system used allows the classification of freely available Landsat TM images from previous years to enhance our understanding of the drivers of land cover and land use changes in Nepal. The land cover information generated can be of great use in informing policy makers and planners working in natural resource management.
The study revealed two
Acknowledgements
This study was funded by USAID NASA (SERVIR Himalaya). The authors are grateful to NASA, USAID, and the MENRIS-Geospatial team of ICIMOD. We would also like to express our sincere gratitude to remote sensing specialist, Faisal Mueen Qamer, and Khurram Shehzad for their intermediate validation and feedback. Thanks are also due to Salman Asif Siddiqui and Manish Kokh for their contribution during the initial stages of this work. The findings reported stand as scientific study and observations of
References (61)
- et al.
Determinants of deforestation in Nepal's Central development region
J. Environ. Manag.
(2009) - et al.
Placing land cover pattern preferences on the map: bridging methodological approaches of landscape preference surveys and spatial pattern analysis
Landsc. Urban Plan.
(2013) - et al.
Detecting trends in forest disturbance and recovery using yearly landsat time series: 2. timesync — tools for calibration and validation
Remote Sens. Environ.
(2010) - et al.
Global land cover mapping from MODIS: algorithms and early results
Remote Sens. Environ.
(2002) - et al.
MODIS collection 5 global land cover: algorithm refinements and characterization of new datasets
Remote Sens. Environ.
(2010) - et al.
Geo-Wiki: an online platform for improving global land cover
Environ. Model. Softw.
(2012) - et al.
The characterization and measurement of land cover change through remote sensing: problems in operational applications
Int. J. Appl. Earth Observ. Geoinform.
(2003) - et al.
Land use dynamics and landscape change pattern in a mountain watershed in Nepal
Agric. Ecosyst. Environ.
(2003) - et al.
A comparison of three image-object methods for the multiscale analysis of landscape structure
ISPRS J. Photogramm. Remote Sens.
(2003) - et al.
A multi-criteria approach for an integrated land-cover-based assessment of ecosystem services provision to support landscape planning
Ecol. Indic.
(2012)