DROUGHT ASSESSMENT USING REMOTE SENSING AND CIS TECHNIQUES

Abstract

Drought has a varying frequency from once in two years to once in eight years. Generally it is observed that one part ofthe country is facing floods while another is in the grip ofa severe drought. This study highlights the problem ofdrought in India in the past in general. In India, every one part or the other faces the problem of the drought except some pockets of the north-eastern region and Kerala. Drought usually occurs because ofthe deficient rainfall causing water scarcity in the region. Out ofthe total gross cultivated area ofthe country, 56 million ha are subjected to inadequate or highly variable rainfall. Frequent droughts in the country result in misery, erode livelihoods, damage integrity of natural eco-systems and cause diseases or deaths due to poor-quality of water and hunger. The drought in the year 2002 is one of the severest in the past 130-year where 56% of the geographical area and the livelihoods of 300 million people in 18 states including Karnataka was affected. To cope with drought, people have over exploited groundwater mainly for food production and the situation is aggravating further, as rainfall is insufficient to recharge the falling groundwater levels. In the present study, Gubbi taluk (tahsil) of Tumkur district, Karnataka state which is drought prone has been selected for drought assessment. Three methodologies have been developed for drought assessment that uses rainfall data, remotely sensed data and meteorological data. The objectives of this research work are (i) to identify the parameters influencing drought phenomena, (ii) to assess drought based on rainfall analysis, (iii) to identify and assess agricultural drought village-wise and season-wise prevailed during 1996-2001 from Drought Severity Index (DS1) that uses remotely sensed data, (iv) to create a database in Geographical Information System (GIS) involving drought influencing parameters for drought analysis and (v) to develop a GIS based drought model to generate the agricultural and meteorological drought map season-wise at village level. The dataset utilised in this study comprises of both spatial and non-spatial data. Spatial dataset includes topographical maps from Survey of India, revenue map from Director, Land Survey and settlements, Govt, of Karnataka and Bangalore Publishers, agro-climatic map from Karnataka State Landuse Board, climatic map published by Central Building Research Institute, Roorkee, rainfall map from Drought Monitoring Cell and remotely sensed data during 1996 to 2001 which is procured from National Remote Sensing Agency, Hyderabad, India. Non-spatial dataset comprises village-wise statistics and information extracted from Census Report, 1991 and 2001. The details of water tanks such as water-spread area, live storage, alchkat (cultivable land coming under the command of a particular water tank) coming under each tank is extracted from the Report of Minor Irrigation. The Information about village amenities is extracted from the Report of village development works by Zilla-parishad, Tumkur, 2001. The daily rainfall data pertaining to the study area during 1950-2001 is procured from Directorate ofEconomics and Statistics. Drought assessment has been carried out using the above dataset. In the first method, the drought assessment is made from departure analysis using rainfall data as suggested by Indian Meteorological Department (IMD). This is a traditional one and do not depict the actual drought since, the rainfall spread is assumed to be uniform over an area. In the second method, season-wise agricultural drought assessment has been attempted by using remotely sensed dataset. This is carried out by evolving DSI that utilizes cropland and barrenland derived from landuse/Iand cover map that are generated from remotely sensed dataset. The third method helps to identify both agricultural and meteorological drought. Here, the data used includes meteorological and remotely sensed datasets. The drought assessment in general is carried out by the creation of the database with 17 drought parameters that are recognized based on drought literature review. Since drought is a weather related phenomenon, the exact number of parameters and their inter-play which is highly difficult to identify. A database is created in GIS environment using remotely sensed dataset, meteorological dataset and other published information on drought. The database creation involves the generation of thematic layers of drought parameters that are created using spatial dataset in ERDAS IMAGINE software. The landuse/land cover maps are prepared by adopting supervised classification technique that uses Maximum Likelihood Classifier (MLC). These layers are exported into Arc View for creating iii database in GIS. The layers are ranked using Saaty's pair-wise comparison method. This helps to remove subjectivity in assigning ranking. GIS based drought severity models are developed to generate drought severity maps by integrating the thematic layers in spatial modeler (a built-in model in Arc View). From rainfall analysis, the frequency of occurrence of drought at all raingauge stations has been obtained. It helps to identify drought years into 4 drought classes as per IMD. It has been observed that the area under high drought probability did not coincide with the classification of drought. Kadaba region has highest probability of drought occurrence but did not account for severe drought condition while Ankasandra region is ranked second least in drought probability but has experienced more severe drought. Agricultural drought has been assessed season-wise by using DSI. It helps to classify agricultural drought into 5 drought classes and thereby villages are grouped. Having identified the villages, the extent of agricultural drought is assessed. Agricultural and meteorological drought is identified season-wise from the drought severity map generated from GIS based drought models. Drought assessment is made village-wise by overlaying the village boundary map over drought map. The methods followed in this study for drought assessment are unique in their dataset and methodology adopted. This study has demonstrated the integrated use of remotely sensed data and GIS technique to identify extent and severity of drought. The drought severity maps thus generated are helpful to combat the problem of drought in providing timely reliefto the immediate suffering communities at large. iv