GISdevelopment.net ---> AARS ---> ACRS 1991 ---> Land Use

Integration of Remote Sensing and Geographic Information System to study temporal changes in Land Use

Mrs. P. Venkatachalam, C.V.S.S.B.R. Murty and S. Chowdhury
Centre of studies in Resources Engineering
Indian Institute of Technology, Bombay - 400076, India


Abstract
Integration of Remote Sensing and geographic information systems is essential for effective resources management. By combining Remote Sensing and GIS, the resources managers can have timely and accurate view of the temporal changes occurring on the technology is helping to look Remote Sensing as a spatially oriented data base and an important component of operational resources-based GIS. In the centre for studies in Resources Engineering, Indian Institute of Technology, Bombay, a micro based GIS Package (NRDMS - GIS) has been developed which has the capability to handle vector cum raster data, analyze digital Remote Sensing data and integrate the Remote Sensing data plane with other map planes using geo-registration. Using this package a study has been carried out to understand the temporal changes in land use pattern for a part of Goa. India. The land use maps of 1969, 1987 and prepared using topographical maps, aerial photographs and thematic mapper data respectively have been overlayed using GIS. The areas where forests have been depleted over 2 decades and changed into open scrubs or mining or agriculture or urban area have been brought out. The study shows the fast environmental changes occurring in the hilly train of Goa and the depletion of forestry.

1. Introduction
World-wide interest in improved management of natural resources stems from the recognition that resources are no longer plentiful. As the supply of commodities dwindles and the number of environmental incidents grow, it becomes evident that better information is needed to enable men to use their resources wisely. The need for resources data has out-paced the capabilities of conventional survey and monitoring techniques. This need is particularly important in developing countries. One method of meeting this challenge is through an integrated information system for resources management that could provide higher quality information in a more timely fashion than current systems. A geographic information integrated with a Remote Sensing image analysis system can help in integrating and analyzing a wide variety of spatial information as a support to mapping, decision making, planning and monitoring (Ehlers, 1990).

The data inputs in a geographic information system are usually spatial and consist of thematic maps derived from a combination of existing maps, aerial photographs and manual interpretation of remotely sensed imagery. The thematic map such as land use map becomes obsolete quickly and it is essential to update these maps periodically in the GIS. Remote Sensing data can become the most cost effective source for these updates. In this paper, an attempt has been made to study the changes occurred in land use over the past 2 decades in part of Goa, India. For this purpose, the land use maps of 1969, 1977 and 1987 have been prepared using topographical maps, aerial photographs and thematic mapper data respectively and have been integrated using an indigenously developed GIS package NRDMS - GIS.

Study Area
An area of 38.69 sq. kms south of Dicholi in North Goa has been selected for this study. The area falls within 73°56' and 74° E longitudes and 23°36' and 23°40'N latitudes. The confluence point of rivers Valvot, Kudana and Dicholi with Mandovi falls within the study area. The has rolling topography in South Western parts, flood plains of Mandovi in central and Southern parts and small hills north Western parts. The area shows good forest cover and very little mining area in the landuse map of 1967 activity has increased and the forest has given way to cashew plantations.

Software Overview
An important component of a GIS is to handle large quantities of spatial and nonspatial (attribute) data which must be efficiently manipulated and integrated (Gorte, 1988). Hence NRDMS has been built with five major module namely input module, analysis module, terrain module, image processing module and print module. Input module helps in transformation of the analogue maps input computer compatible form through digitization editing of maps to remove the digitization errors, creation of vector database with topological relations, polygon formation, acceptance of thematic unit for each polygon and rasterisation. Attribute data for each polygon can be entered through the keyboard to create the geometrically corrected Remote Sensing data or results interpreted form Remote Sensing data can be directly appended to raster data base.

Many of the analytical functions such as map overlays, classification, proximity calculations, mathematical and logical operational can be easily implemented in a raster domain. Hence the analysis module supports various manipulation functions that can be performed using raster data planes. Terrain module helps in manipulating contour/elevation data for slope, aspect, relief etc of given surface.

Remote Sensing data are an important source of information for monitoring temporal changes and updating the map planes. Original Remote Sensing data is normally in raster form where for each pixel, there is a list of associated reflectance values in multispectral bands. Using proper image processing procedures, these data can be transformed in to thematic information for landuse. Soils, vegetation etc. image processing module provides the facilities for preprocessing, image display , enhancement, edge detection, classification and geometric correction.

Print module provides capabilities for displaying results in both graphic and tabular forms and hard copy can be obtained using inkjet printer.

Help facility is provided for each function describing its use and the syntax for execution. The software is arranged in a hierarchical structure with submenus at different levels. Figure 2 shows the data flow structure in the system and a suitable procedure can be selected depending on the data type and the application.

Analysis
The land use maps of 1967 (topographic maps), 1977 (aerial photographs) and 1987 (thematic mapper imagery) have been digitized. The temporal; change detection both spatially and quantitatively has been accomplished with the help of GIS map manipulation functions as they provide a greater freedom for map handling. The maps have been compared for all the landuse types from 1967 to 1987 and a table (table 1) showing the changes in terms of area percentage for each ladnuse type has been prepared.

Table 1
Landuse Type Area %(1967) Area % (1987)
Agriculture
Marshy
Open Scrub
Forest
Builtup land
Mining
33.76
4.16
9.51
35.87
3.14
0.14
44.998
7.32
16.79
13.34
5.92
3.04

It is clear from the table that there has been an increase in each land use except in case of forests which show a considerable decrease. So the forest area from the two maps has been analysed separately to find out land use practices in 1987 in the deforested areas. The final results have been summarized in Table 2.

Table 2
Landuse change in forest cover (1967-1987) Area %
Forest to agriculture and plantations 21.59%
To open scrub 3.42%
To built-up land 1.15%
To mining .84%
Total depleted forest area in % 27.00%

Conclusion
Appraisal of the resources and obtaining timely and accurate knowledge of the resources is very essential to plan an optimal resource management strategy for a country's development. As the date on resources are being collected from several decades it becomes important to make use of traditional and modern source of information. Geographic information system with a Remote Sensing component is the ideal tool for integrating the basic information sources such as maps, imagery and statistical data for effective resources management. (Goodenough, 1988). The capability of a GIS to convert an analogue maps into digital maps and the ability to merge both vector and rater data sets opens up a whole new area of analysis potential (Barker 1988). For the current study on detecting the land use changer in a span of 20 years both spatially and quantitatively integration of Remote Sensing data with GIs has been found extremely useful.

References
  • Barker, G.R., (1988) "Remote Sensing: the underalded component of Geographic Information Systems" Photogrammetric Engineering and Remote Sensing vol. 54, no. 2, 195-199.
  • Elhers. M. (1990) "Integration of Remote Sensing Image Analysis and GIs ; Status and Trends", Presented at the Tenth INCA Conference, Bombay.
  • Good enough, D.G, (1988) "Thematic Mapper and SPOT integration with a geographic information system' photogramemtirc engineering and Remote Sensing Vol. 54, No. 2, 167-176.
  • Gorte, B., Liem, R and Wind J. (188), The ILWIS Software Kernel, ITC Journal, No 1, P. 15-22.