IMPACT OF LANDUSE
LAND COVER CHANGE ON STREAMFLOW OF UPPER BAITARANI RIVER BASIN USING SWAT
Abstract: Change in land use and land cover pattern influences the hydrology of
a watershed by changing its stream flow and groundwater characteristics. The
main objective of this study was to assess the land use land cover change
impact on stream flow of upper Baitarani River Basin using SWAT (Soil &
Water Assessment Tool). SWAT is a helpful tool to determine the different
hydrologic characteristics under different conditions. In this work, the study
area was delineated based on Anandpur Hydrological Observation Station and a
hydrologic model was set up using SWAT software. After Sensitivity Analysis,
Calibration and Validation of SWAT model were done using observed daily stream
flow at Anandpur gauging station with the help of SUFI2 algorithm in SWAT-CUP
software for 1995, 2004 and 2013. The results represent SWAT model performs
satisfactorily in simulating runoff according to NSE (Nash Sutcliffe
Efficiency), PBIAS (Percent bias) and R2 (Coefficient of Determination) values.
After calibration from 1979 to 1990 daily stream flow NSE, PBIAS and R2
and 0.64 respectively. Similarly, after validation NSE, PBIAS and R2
were 0.57, 14.2
and 0.60 respectively. The result
showed forest, agricultural area, and barren land were decreased, but the
built-up area was increased significantly. As a consequence discharge of
Baitarani River was decreased.
Keywords: LU/LC change pattern, SWAT modelling,
SUFI 2 algorithm.
2. Materials and methods
2.1 STUDY AREA
The study was conducted for Baitarani Basin. Originating
from the Guptaganga hill of Keonjhar District of Odisha at an elevation about
900 m above Mean Sea Level the Baitarani River crosses a length of 360 km and
meets with Bay of Bengal. The Brahmani River flows on the South and West of the
basin, the Subarnarekha River on the North and the Budhaganbalanga and the Bay
of Bengal on the east. About 6.7% and 93.3% area of Baitarani basin falls in
Jharkhand and Odisha respectively with total catchment area of 10,982 sq. km.
The upper reach of Baitarani River is up to
Anandpur hilly region. The study area lies approximately between east longitudes
85°10′ to 86°23′ and north latitudes 20°53′ to 22°15′ covers area about 8,619
sq.km. Anandpur is situated at Keonjhar district of Odisha (Figure-1). The climate
is Tropical with average annual rainfall of 1442.53 mm. The hilly areas face
lesser temperature variations compare to plain areas during the years.
Generally December and yearly January are the coldest months with minimum
temperature of 12° C whereas the average annual maximum
temperature of 37.67° C and annual minimum temperature is 20.32° C.
April and May are the hottest months of the year when the temperature varies
between 35° C to 38°C. The wind velocity is higher in months of April,
May and June whereas it’s lesser in December and January (CWC 2014).The basin contains loam soil, sandy soil and sandy
clay loam soil. Forest, agriculture and built-up area covers the major part of
Figure-1: location of study area
2.2 INPUT DATA
2.2.1 METEOROLOGICAL DATA
parameters (precipitation, maximum and minimum temperature, relative humidity,
solar radiation, wind speed etc.) influence the hydrologic cycle greatly.
That’s why to create SWAT model this data is very compulsory. The
meteorological data from 1979 to 2013 was obtained from The National
Centers for Environmental Prediction (NCEP) (https://globalweather.tamu.edu).
The data was collected from six stations inside the study area. Figure
2.2.2 GIS DATA
flows from high elevation to low elevation terrain data is necessary for
hydrological modelling. The DEM (Digital Elevation Model) of Baitarani river
basin was obtained in the form tiles from CIGAR-CSI website (http://srtm.csi.cgiar.org) with 90 by 90 m DEM resolution.
These tiles were mosaiced into a single tile in ArcGIS 10.1. The study area was
automatically delineated with the help of the outlet point. DEM was also used
to depict the drainage pattern, flow accumulation and stream flow networks of
the basin area. Figure
The soil texture plays a very important role in SWAT model as different kind soils in the
world have different physical and chemical characteristics .Harmonized world
soil data of 1:50,00,000 scale in vector format was acquired from FAO (Food and
Agricultural Organisation) soil portal (http://www.fao.org). The soil map of study area was clipped
by the shape file in ArcGIS 10.1. With the help of Harmonized world soil
database the type of soil in the study area was obtained. The upper Baitarani
river basin contains I-Ne or loam soil (28.38%), I-bc or loam soil (6.65%), Nd50-2b or sandy
clay loam (64.92%) and Lf95-1a or sandy loam soil (0.05%).
LAND USE LAND COVER DATA
Land use is
defined as expenditure of land to produce goods and services. Hence, land use
is dependent on the purpose for which the land is being used whereas Land cover
means natural physical cover, as seen by eyes or remote sensing (FAO, 1997).Land
use land cover map were obtained from USGS LANDSAT satellite imagery (https://earthexplorer.usgs.gov/).
Using path 139 and 140 and row 45 LANDSAT 4, 5 and 8 satellite images for 1995,
2004 and 2013 having spatial resolution 30 m were attained. Then in ArcGIS 10.1
the images of 1995, 2004 and 2013 were mosaiced and clipped respectively. Each
clipped image were classified into five categories of land use-cover using
maximum likelihood algorithm in supervised classification tool of ArcGIS. The
five classes were water body, forest, agriculture, barren land and built up
area (Figure-1995,04,2013). Table.1 represents the quantified
change of land use-cover categories.
2: DIFFERENT LAND USE LAND COVER AREA IN UPPER BAITARANI BASIN
LAND USE LAND COVER SPECIFICATION
2.2.3 HYDROLOGICAL DATA
discharge data (1979-2013) of Baitarani River at Anandpur stream gauge station (21°12’40”N, 86°7’14”E)
was obtained from Water
Resources Information System of India (http://india-wris.nrsc.gov.in/wris.html). This data was further used for
calibration and validation purposes of the SWAT model.
2.3 SWAT MODEL SET UP AND SIMULATION
SWAT is a comprehensive, Semi distributed, hydrologic model
which runs at a continuous time-step (Arnold et al., 2012). It was developed by
Dr. Jeff Arnold for USDA-ARS to determine the impact of land management
practices on water as well as sediment, agricultural yields in different
complex watersheds under varying land use and soil conditions (Neitsch et al.,
2009).SWAT uses spatial and temporal data as input. Spatial data includes DEM
(Digital Elevation Model), land-use map and soil map. Temporal data includes
hydrological data (stream flow) and weather data (precipitation, temperature,
wind speed, relative humidity and solar radiation). SWAT divides a basin into
multiple sub basins depending on the topographic criteria and further sub
divided into a number of HRUs (Hydrologic Response Units) on the basis of
homogeneous land use, soil and slope combination characteristics. Simulation
can be done for different kinds of hydrological (runoff, water quality,
evapotranspiration etc.), agricultural variables (crop yield, nutrients
etc.).Based on the requirement and available data user can easily simulate the variables
in SWAT. Hydrological simulation of the basin was carried out using ArcGIS
extension of SWAT called ArcSWAT (Arnold, 2013).
study, ArcSWAT v2012.10.1.18 (http://swat.tamu.edu/software/arcswat) was used. To
set up the SWAT model the
study area was automatically delineated with the help of the outlet at Anandpur
stream gauge station (21°12’40”N, 86°7’14”E) (CWC, 2014).The
upper Baitarani basin had produced only one sub basin and that was further
subdivided into 18 HRUs based on
inputted soil and land use/cover data. Then the model was simulated from 1979
to 2013 on daily time step with a warm period of five years individually using
1995, 2004 and 2013 land use/cover map.
Figure of simulation for 1995, 2004 and 2013
2.4 SENSITIVITY ANALYSIS, CALIBRATION AND VALIDATION OF SWAT