Identification of Non-Point Source Pollution Risk Using GIS and Remote sensing Techniques
N.D.K. Dayawansa
Department of Agricultural Engineering, university of Peradeniya, Sri Lanka
J.P.Delsol, H.Andrianasolo and V.V.N.Murthy
STAR Program, AIT,P.O. Box 4, Klong Luang, Pathumthani 12120, Thailand
Abstract
Non-point source pollution is a major problem in most of the agricultural catchments where sediments, plant nutrients, pesticides and animal constitute the pollutant loads. In this study, Agricultural Non-point Sources (AGNPS) pollution model was used to simulate runoff. Sediments and nutrient transport from a sub catchments. The land use data was derived form the supervised classification of IRS LISS II imagery and the modeling resolution for input data was set ot 40 acres. Results were obtained for the individual rainfall events and also for the total rainfall days in the rainfall year 1992/93. the reduction of pollutant loads due to the application of sloping Agricultural Land Technology 9 SALT) is also verified through model simulations.
Introduction
With population growth, the demand for land and water resources hs been ever increasing all over the world. The over exploitation of these resources leads to a number of environmental problems including pollution of streams and water bodies. Non- point source agricultural pollution is one of the major factor in polluting surface waters in agricultural areas. Non-print sources are those which discharger into a catchment in a way that they depend upon the routes of the hydrological cycle to transport them to the stream system. They differ from te point sources due to their unidentifiable nature and the difficulty in control and management . the non-point agricultural pollutants are organic and inorganic materials including sediments, plant nutrients, pesticides and animal wastes entering surface and ground waters form non-specific or undefined sources in sufficient quantities to contribute to the problem of pollution .
The magnitude and the extent of non-point source pollution can be evaluated long term on-site monitoring and simulation modeling. Due to the resource constraints associated with on-site measurements, simulation have frequently been relied upon to provide guidelines in developing and implementing pollution control programme. There are several no-point source pollution assessment models and Agricultural Non-point source (AGNPS) pollution model selected for the study is a widely used single storm event based model. It simulated runoff. Sediments and nutrient transport form agricultural catchments . the model consists of three components namely, hydrology, erosion and sediments, and chemical transport and operates at the user-defined cell resolution .
Non-point source pollution modeling requires as extensive knowledge about the land use, soil, slope, agricultural activities and the socio-economic background of the area. The remote sensing and Geographical information system (GIS) with its spatial analysis capabilities facilitate handling these large quantities of spatially varying data required for pollution modeling in catchments.
Approach
This study was carried out at an important sub catchment called Nilambe in the central part of Sri Lanka. This area is highly degraded and consists of very steep slopes and eroded soils . there are three major development products on domestic water supply, hydropower generation and irrigation implemented at Nilambe. The success of these projects mainly depend on the yield and the quality of water . hence, the importance of identifying the potential non-point sources of pollution to the water bodies is understood and the need for creating awareness among the people and implementing remedial measures to protect te land, water and human resources is emphasised.
The main objective of this study was to apply AGNPS model to the Nilambe catchment through the integration of remote sensing and GIS in order to identify the non-point source pollution risk locations. There are 22 input parameters required by the model and among them, five impute parameters and directly dependent on ht eland use and cover. Tehrefoere, it is important to identify the existing conditions of land use and cover for model simulations and also to make provision for updating land use and cover to represent the dynamic status with time. Indian Remote Sensing Satellite (IRS LISS II) imagery acquired in March, 1992 was used to derive the land use and cover map of the area using the supervised classification techniques. The nest there band combination for display purposes was identified using optimum Index Factor (OIF) approach ( Dwivedi and Rao, 1992) which involves the variation and correlation among different bands. Six land use and cover categories were included in the legend for the land use and cover map derived from the imagery ( figure 01 ). Five model parameters namely, surface condition constant, Manning's roughness coefficient, runoff curve number, cover and management practices factor in universal soil loss equation and Chemical Oxygen Demand (COD) factor were derived based on the land use and cover. The slope, slope aspect and slope length were derived using 1;50,000 topographic map sheets. Soil textural information was extracted using 1:10,000 soil map along with the associated information obtained through field investigation.
Figure 1 Land use and cover map produced from supervised classification of IRS LISS II imagery
Figure 2 Identified zones based on sediment production
