Sikandrabad, situated about 50 km from the NCR, New Delhi and 15 km from Greater Noida, Gautam Buddha Nagar, Uttar Pradesh, is a municipal town in Bulandshahr district in the state of Uttar Pradesh, northern India (Figure 1). The study area lies between Latitude 28° 24’ 14.73” N Longitude: 77° 32’ 33.87” E and an altitude is 200 msl. The administrative boundary of Sikandrabad is surrounded by district Ghaziabad in the north-east and Dadri town (Gautam Buddha Nagar district) in north-west. In the south are Khurja and Aligarh, districts and Haryana state in the south-west. The town has a population of 69,902; males constitute 52% of the population and females 48% (Census, 2011).
Climate: The climate of the study area is sub-humid and characterized by hot summer and bracing cold season. After February there is continuous increase in temperature till May which is generally the hottest month. In summer, from March-June, the weather remains hot and temperature ranges from a maximum of 45 °C to a minimum of 23 °C (Joshi, 2008-09). Temperature falls substantially down to as low as 3-4°C at the peak of winter. In December-January, a dense fog envelopes the study area reducing visibility. Monsoon season prevails during mid-June-September with an annual average rainfall of 93.2 cm.
Soil: The soil ranges from pure sand to stiff clays and including all combination of the two extreme litho units. The mixture of sand and clay in equal proportion forms loam, a good agriculture alluvial soil is widespread.
Agriculture: The major agricultural crops of Bulandshahr district are wheat, maize, barley, rice, pulses, pearl millet, potato and sugarcane. The existing cropping pattern of the district shows that out of 82.49 % gross cropped area under the nine major crops, wheat covers 38.15% of the area in the district followed by maize, rice and other crops (Khan and Khan, 2014). Singh and Islam (2010) have reported decrease in size of agricultural land holding in Bulandshahr district during 1991-2001.
Flora and fauna: The main tree species found in the study area are: Azadirachta indica A. Juss., Butea monosperma (Lam.) Taub., Cordia dichotoma G. Forst., Dalbergia sissoo DC., Ficus religiosa L., Holoptelea integrifolia Planch., Mitragyna parvifolia (Roxb.) Korth., Neolamarckia cadamba (Roxb.) Bosser., Prosopis cineraria (L.) Druce , P. juliflora (Sw.) DC., Phoenix sylvestris (L.) Roxb., Tamarix ramosissima Ledeb. And Terminalia arjuna (Roxb. ex DC.) etc. The study area is rich in the plant herbal diversity including economically and medicinally important plant species, such as Amaranthus viridis L., Achyranthes aspera L., Alternanthera sessilis (L.) R.Br. ex DC., Centella asiatica (L.) Urb., Bacopa monnieri (L.) Wettst., Calotropis gigantea (L.) Dryand., C. procera (Aiton) Dryand., Cannabis sativa L., Chenopodium murale L., Convolvulus prostrates Forssk., Croton bonplandianus Baill., Dysphania ambrosioides (L.), Digera muricata (L.) Mart., Mosyakin & Clemants, Heliotropium indicum L., Oxystelma esculentum (L.f.) Sm., Eclipta prostrata (L.) L., Phyllanthus niruri L., P. maderaspatensis L., Phyla nodiflora (L.) Greene, Stellaria media (L.) Vill. and Tribulus terrestris L. etc., (Alamand Anis, 1987; Singh and Ahmad, 2010; Aggarwal et al., 2012; Chaudhary and Narayan, 2013).
The study area has rich bird diversity including that of migratory birds viz., Asian Open bill Stork, White-necked Stork, Black-crowned Night Heron and Black-headed Ibis. Significantly, the study area supports the breeding of some rare birds like Bristled Grass bird (Chaetornis striata), Black-necked Stork (Ephippiorhynchus asiaticus) and Sarus Crane (Grus antigone). Apart from large population of avifauna, the study area has six species of mammals includes Nilgai (Boselaphus tragocamelus), Indian Grey Mongoose (Herpestes edwardsii), Indian Hare (Lepus nigricollis), black buck deer (Antilope cervicapra), Deer (Odocoileus virginianus), Leopard (Panthera pardus) and Golden Jackal (Canis aureus), etc.
2.2 Data used
Satellite images IRS P6 LISS-III of the study area for the year 2004 and 2010 were obtained from the National Remote Sensing Centre, ISRO, Hyderabad, India (Table 1). The satellite data were corrected geometrically using Landsat TM data with 30 m spatial resolution. The radiometric corrections viz., histogram equalization, dark pixel subtraction, contract enhancement and image starching, etc. were done on all the bands were combined into a single file using layer stacking techniques (Lillesand et al., 2007). The complete methodology is shown in Figure 2. The satellite images were converted into False Colour Composite (FCC) for identification of tonal characteristics (Figure 3 and Figure 4). For image classification satellite image processing steps were involved in starting from processing of IRS P6 LISS-III satellite images for radiometric and geometric corrections. The training data set taken in different places for supervised classification of satellite images of 2004 and 2010 were based on delineation of different land features viz., forest, agriculture, settlement, wasteland, wetland, grassland and water body. The classified 2004 and 2010 satellite images were checked in the field and based on the ground truth, land use/cover map of 2004 and 2010 prepared (Figure 5a and Figure 5b). The information provided by the satellites in combination with other sources to quantify the various parameters for efficient mapping of land use/cover of the basin was evaluated by applying various image processing steps using ERDAS Imagine ver. 9.3 and ArcGIS ver. 10.1.
2.2.1 Satellite data processing
Georeferencing of satellite imagery is essential for analyzing land use/cover pattern of a particular geographic area. The atmospheric corrections/radiometric corrections is an important part of remote sensing images, which are more pronounced in the shorter wavelength regions, which cause some additional contribution to spectral reflectance. In the present study, satellite data were geometrically corrected for the distortions and degradations caused by the errors due to variation in altitude, velocity of the sensor platform, earth curvature and relief displacement. The images of IRS P6 LISS-III satellite data were geometrically corrected and geo-coded to the WGS 1984 datum with Lambert Conformal Conic (LCC) projection coordinate system by using a reference image of Landsat TM satellite data with spatial resolution 30 m. A minimum of 15 regularly distributed ground control points were selected from the images. The georeferencing was performed using first order polynomial transformation, resampling using a nearest neighbour algorithm. The transformation with a Root Mean Square Error (RMSE) was 0.005. Image enhancement, dark pixel subtraction, histogram equalization, contrast stretching and false colour composites were also worked out (Jensen, 2005).
2.3 Change detection analysis
Land use/cover map was produced from the IRS P6 LISS-III satellite data employing supervised classification. The training sites and extraction of signatures from the images were taken and then classification of the images was attempted. Training data extraction was a critical step in supervised classification as these must be selected from the regions representative of land use/cover class under consideration. Thus, the satellite data were collected from relatively homogeneous areas consisting of these classes. After the training site areas were digitized, the statistical characterization of the information was created called signatures. Finally, the classification methods were applied. All the classification techniques like the Maximum Likelihood Classification (MLC), Parallelepiped and Minimum Distance to Mean classification were applied for the images and the best classification technique was then selected. It was observed that Maximum Likelihood Classification (MLC) gave good results as compared to the other two techniques.
To determine the accuracy of classification, a simple testing pixel was selected on the ground truth reference data. For assessing the temporal changes in the land use/cover map was prepared using IRS P6 LISS-III geocoded FCC of 2004 and 2010 data on 1: 50,000 scale. The vegetation areas were delineated from their red tone and contiguous pattern. These classes were identified from their red tone, coarse texture and scattered pattern. The agriculture and settlement were identified from the light reddish-brown tone and regular pattern and wetlands from dark and light blue tone. Necessary ground truth was carried out and correction was made at required places and the various classes’ viz., forest, wasteland, agriculture, grassland, settlement, water body and wetland were identified. Thus, a thematic layer depicting the various land use classes was generated. The primary land use/cover map was prepared based on field observation and image interpretation. Then, using software such as ArcGIS ver.10.1 and ERADS Imagine ver. 9.3 classified land use/cover map was prepared. Finally, the status of change detection was analyzed from 2004 to 2010 (Table 2). land use/cover maps were assessed by overlaying the maps of 2004 to 2010.
- Results and Discussion
The accuracy assessment of land use/cover showed overall classification accuracy of 87.50 % with Overall Kappa Statistics = 0.8522. For user accuracy of 200 reference points and 200 classified points, 175 correct points were checked in the field (Table 3). Higher producer accuracy was found in settlement (100.00%), followed by wasteland (96.43%) followed wetland (93.33) and waterbody (88.00), medium accuracy found in water body (88.00%), forest (86.49%) and low accuracy found in agriculture (81.58%) and grassland (79.41%). Similarly, high user accuracy was observed in forest (91.43%) and agriculture (91.18%), medium accuracy was accessed in grassland (90.00%) and water body (88.00%), low accuracy found in wasteland (87.10%), followed by wetland (87.50) and settlement (61.54%). In each land use/cover laid minimum five sample points for accuracy assessment and ground truthing. The present study demonstrates the capability of geospatial technology to capture the land use/cover categories in a semi-arid region of upper Gangetic plains of Uttar Pradesh, India, which is important from the point of view of sustainable utilization of natural land resources, biodiversity conservation and management planning.
The results showed that land use/cover classes were altered remarkably in the study area during the six year period from 2004-2010. Singh and Islam (2010) reported 16.74 % increase in population in Bulandshahr district during 2001-2011. In 2004, of the total of 850 km2 the maximum area was occupied by agriculture 521.86 km2 (61.04 %) followed by wetland 124.97 km2 (14.70 %), grassland 106.07 km2 (12.48 %), wasteland 51.59 km2 (6.07%), settlement 26.27 km2 (3.09%), forest 13.85 km2 (1.63%) and water body 5.39 km2 (0.63%). However, within a span of six years in 2010 the land use pattern showed drastic changes. Although, the maximum area remained under agriculture (498.25 km2) (58.62 %) albeit at reduced level by 2.32%. This was followed by settlement 125.92 km2 (14.81%), wasteland 80.38 km2 (9.46 %), wetland 61.44 km2 (7.23%), grassland 60.85 km2 (7.16%), water body 11.65 km2 (1.37%) and forest 11.51 km2 (1.35%) (Figure 6). This showed significant incremental changes in settlement by 11.72% and reduction in wetland by 7.47% and grassland by 5.32% followed by agriculture (2.79% and forest (0.28%). This change in different land use/cover classes has been as a result of vigorous anthropogenic activities, rapid urbanization and population increase in the study area. The land use/cover-wise change detection analysis brought out very interesting trends in change in land use pattern from 2004 to 2010.
Due to urbanization, area under settlements increased by adding 223.08 km2 (47.76%) at the expanse of agriculture (26.17%), wetland (57.17%), wastelands 23.67% and grassland (20.92%). Loss of agricultural land 121.32 km2 (26.67%) due to expansion of settlements got compensated by conversion of grassland 93.20 km2 (20.13%) wasteland 44.94 km2 (9.69%) and wetland 102.25 km2 (22.06%) into cultivated lands (Table 4). Hence, there was a gain in area under agriculture by 119.17 km2 (4.12%). The areas rich in wild biodiversity such as wetlands (34.39%), wastelands (4.80%) and grasslands (24.64%) and were greatly affected due to expanding settlements and agriculture showing a decline in the land area under these categories over the six year period (Figure 7).
Significant loss was noticed in the case of wetlands. A total of 159.42 km2 of wetlands were lost to agriculture 102.25 km2 (22.06%) and settlements 57.17 km2 (12.33%) accounting for a total loss of 34.39% area under wetland (Table 5). Krishna (2012) reported a decrease of water bodies by 50% during 1986-2011 in adjacent Gautam Buddha Nagar district. He ascribed the major cause for this decrease to wide spread urbanization. In the study area also, wetlands rich in aquatic and avian diversity, were the worst affected which is a serious concern. Wetlands are one of the most threatened of all biomes as they become the first land use victim of development (Panigrahy et al., 2012). Wetlands, also called “biological supermarkets” because of extensive food chain they support, play a crucial role not only in the hydrological cycle but in the ecosystem (Prigent et al., 2001; Varghese et al., 2008). Prasad et al., (2002) reported that rapidly expanding human population, large scale changes in land and improper use of watersheds have all caused a substantial decline of wetlands resources of the country. These significant losses have resulted from its conversion to agriculture, urbanization and other developmental activities similar to our findings in the study area. It is possible that excessive use of ground water through numerous unauthorized bore wells may also have resulted in drying of wells (Sikka, 2002).
The loss of wetland habitat will definitely affect the migratory birds in western Uttar Pradesh. Dadri wetlands, near Bil Akbarpur in Dadri harbor more than 220 migratory birds. A rare and endangered species bristled grass bird (Chaetornis striata) was recorded in 2010 with breeding nest in Dadri wetlands. Possibly it was the first record in India in recent times (Anon., 2010). As wetlands are destroyed some birds may move to other less suitable habitats putting survival as well as future migration in danger. Thus, many wetland birds are edging very close to extinction through disturbance and conversion of their habitats (Kumar et al., 2005).
Similarly, the grasslands, which also provide ecologically unique ecosystem were adversely affected and an area of 114.22 km2 (24.64%) was converted to agriculture 93.30 km2 (20.13%) and settlement 20.92 km2 (4.51%). Grasslands, mostly occupied by grasses such as Saccharum spontaneum L., S. bengalense Retz., and Phragmites karka (Retz.) Trin. ex Steud., are common in Gangetic plains with scanty populations of other weeds such as Parthenium hysterophorus L., Cannabis sativa L., Chenopodium album L., etc. Grasslands provide traditional livelihood to local people as they utilize these grasses for various household purposes as well as products for sale.
A total of 68.61 km2 (4.80%) of wastelands, which harbor rich wild plant diversity were also lost during the six year period as this land use was converted to agriculture 44.94 km2 (9.69%) and settlement 23.67 km2 (5.11%). Loss of wastelands will result in loss of rich wild plant diversity as a number of ethno-botanical studies on wild/weed flora of waste lands of Bulandshahr district reported a number of valuable medicinal plants which are being used in traditional medicines by local people (Mashkoor Alam and Anis, 1987; Singh and Ahmad, 2010; Aggarwal et al., 2012; Chaudhary and Narayan, 2013). Disappearance of herbal flora in wastelands will deprive people of these medicinal herbs occurring in the wild and practicing the traditional way to cure various ailments by local population.
Overall, this study revealed significant changes in various land use land cover classes in Secundrabad and surrounding areas in upper Gangetic plains of western Uttar Pradesh, India using remote sensed satellite images of 2004 and 2010 and GIS techniques. Secundrabad town (Bulandshahr district) and area of Dadri in Gautam Buddha Nagar district being closer to National Capital Region (NCR), especially townships of Greater Noida and Gurgaon (Krishna, 2012; Sharma et al., 2013) which are moving at a tremendous pace towards a massive urban expansion, it is essential that the future development is on a sustainable model without adversely affecting the natural resources especially biodiversity rich water bodies. There is already population increase reported by Singh and Islam (2010) for Bulandshahr district from 29.13 lakhs in 2001 to 34.99 lakhs in 2011 as well as population density which increased from 656 to 776 during the same period. To accommodate this growing population, expanding urbanization and other developmental activities it is imminent that in the near future the area will see a great change in terms of land use land cover due to conversion of land for various activities (Rahman et al., 2011). Such changes will cause environmental degradation with loss of valuable biodiversity and fertile agricultural land due to negative impacts of unplanned urbanization. With these in view future studies are strongly recommended to develop a sustainable model of development in the NCR region for sustainable utilization of natural resource and their conservation.