Home Articles GIS in forestry: A bottom to top approach for high accuracy...

GIS in forestry: A bottom to top approach for high accuracy and large-scale integration of geo-spatial data”

M. K. Yadava
Working Plan Officer, Upper Assam Circle 500
Jorhat – 785004

The forest personnel in the field have to discharge police, magisterial and quasi judicial functions. An error in the GIS database may prove very costly at times in a field situation.

For more than a century in the Indian sub continent, Forestry management has evolved around principles of rudimentary GIS, and map base data has become an integral part of working and management of the forests, The advent of sophisticated computer hardware and software has, however, immensely widened the usage of the front-line technology of GIS-GPS-RS (Geographical Information System – Global Positioning System – Remote Sensing).

Though the technology is being employed extensively in various types of studies by researchers, it is yet to be effectively used for decision making in the field level. The field applications have remained largely confined to preparation of Working Plans and preparation of some thematic maps. However, its other uses are fast emerging. Satellite imageries are being used as evidence in the court of law against encroachers. In addition monitoring of afforestation and plantation schemes, corridor mapping, habitat mapping, land capability mapping are also being taken up at the field level.

The use could be further widened to selection of sites for plantation, monitoring of encroachment at Division level, regulation of burning and jhum, expansion of infrastructure and communication network, forest village management, personnel management, corridor planning for animal migration, survey and demarcation to suggest a few.

Another potential area for application of this technology is forest -protection. Forests are shrinking at a fast rate. The demand on the forest land is tremendous. Forest protection is a major challenge today. The GIS technology can play a great role here, provided accurate information can be obtained remotely. To make GIS descend to the field level, one has to ensure high level of accuracy in the data and also to be able to organize geo-spatial data in a meaningful way for different levels of decision making. The forest personnel in the field have to discharge police, magisterial and quasi judicial functions. While discharging such sensitive functions, accuracy in data becomes a prominent factor in use of the technology. An error in the GIS database may prove very costly at times in a field situation.

Sources of Error
The author is of the opinion that errors creep into the geo-spatial databases right at the time of base map preparation for satellite data interpretation. Survey of India topographic sheets on 1:50,000 and 1:250,000 scales have been widely used as base maps in most of the remote sensing applications in the country. The Survey of India publishes these sheets after a long interval of time. For example, in the North-East India,, the latest sheets available are those surveyed during 1964-66 and published in 1971. The sheets pose three major problems in forestry application in a typical field situation in the North-East India. Some of the areas demarcated as Reserved Forests on the map have incorrect boundaries. The forest areas brought under reservation subsequently, do not appear on the map (naturally !). Though in principle, ground position can be transferred on map, yet in reality it is a time consuming process requiring a series of trangulations. In effect, this acts as an impediment and only increases the “Idle Time” by days and even months. The third greatest impediment in making correct base maps, especially in the areas having international border such as Assam, is non-availability of Survey of India topographic sheets for the bordering areas.

Moreover, an error may not appear at all on 1:50,000 scale, but may become very prominent in a digital map. Such maps containing error may only serve the purpose of researchers and scientists. However, they are not fit for use in the field situation. An error of 10m in on the ground would be represented by 0.2 mm on a 1:50,000 scale map. The drafting pens normally employed for paper maps are of dimension 0.1 to 0.2 mm. Therefore, such an error may escape un-noticed on a paper map. If that error were to run for 40 cm on the map along a line, it would represent 20 ha of actual error on ground. Estimated loss that may occur to the exchequer for such an error could be well above Rs. 10 crores assuming 60% stocking of the area.

The need for such a high level of accuracy did not arise earlier as because administrative decisions were not based on high end technologies. Error in mapping is largely technology related. Error in paper maps may be termed tolerable, but not so at all in digital maps, An inference drawn from a computer based GIS which is not free from error, may trigger a series of undesirable actions in the executive hierarchy of governance.

An Iterative Algorithm for Error Free Digital Mapping The advantage of working with digital maps is that there is always a room for improvement and correction, involving very little cost and time. Such flexibility is not available with paper maps. Data from various sources may be used to correct digital maps.

Modern cartography largely depends upon aerial photography and satellite imageries. It is, therefore, possible to develop in-house cartographic skills to meet one’s mapping and accuracy requirements. In the paragraphs below is briefly given a sketch of mapping requirements from forestry point of view and also how digital maps are being prepared in Assam for various Forest Divisions.

The management of Reserved Forests and other forest areas is based on compartments which are nothing but small sub-divisions of the forest, the boundaries of which are actually laid down on the ground, Each compartment has a unique number. The complete data base of a compartment is maintained in Compartment History. History of a particular compartment would show not only its geology, rock, soil and forest type but also management principles applied to it from time to time. It also contains records of plantation, incidents of fire, any dereservation/disforestation, alteration in boundaries and any other significant incident pertaining to that compartment. These histories are updated during compilation of Working Plans at fixed interval of time.

To build forestry GIS for field application, compartment is the right mapping unit to start with. During the early British period of conservancy, the compartments were duly mapped by Survey of India on 4″ = I Mile scale. Now original prints of such maps are hard To come by. Repeated tracing over the past several years have rendered the copies of the maps available today unusable, as they have come to contain many errors.

While preparing digital maps for the Divisions, care was taken to verify the map data from various sources, including field visits. As 4″ = I Mile maps were not available or were unusable, there was no option but to use topographic sheets on 1:50,000 and 1:250,000 scale. It is noteworthy to mention here that these scales, in general are considered unsuitable for forest mapping. Reliable old stock maps were used wherever possible.

As Assam has a long international border with Bangladesh, many of the topographic sheets have been frozen by Survey of India. Wherever old sheets were available on 1″ = I Mile and I”= 4 Mile, they were also utilized. To put together maps on various scales in order to produce a single map, digital mapping was the only way out.

Digital maps so prepared were subjected to refinement by using IRS IB/IC False Colour Composites. Details such as changed river courses, channels and streams, water bodies, highways and railway lines were lifted from the imageries and incorporated in the digital maps. In case of ambiguity, even IRS IC PAN images were also used to modify the maps further. In cases where ground survey data was available for the Reserved Forest areas, the bearings and distances were plotted in a separate file and subsequently inserted in the main maps. In certain cases where boundaries were not natural, ground survey was resorted to.

Digital maps thus modified were, then, used for subsequent processing of satellite data. Thus, errors were either eliminated (in few cases) or reduced considerably. However, it is strongly felt that survey is required to place the forest areas on map accurately. Survey in forest areas, especially in hilly terrain, is a time consuming and costly process, In order to produce accurate digital maps, one has to use state of the art technology.
 
Use of GPS in Forest Traverse
In forest areas, in almost all the states in the country, survey is still carried out in the traditional manner with chain and compass. As already stated, the process is not only man-power intensive and time consuming, but also prone to large subjective and objective error, especially in hilly terrain. Use of sophisticated survey equipment such as EDM is yet to dawn in this sector. In the present scenario, GPS appears to be a viable alternative for accurate forest traverses.

The use of GPS in forested terrain has its own sources of error. In closed canopy conditions, either the satellite radio signal may not reach the GPS antenna, or the signal may get attenuated. The canopy effect may lead to loss of lock. In such circumstances, the idle time may increase considerably. However, as per available literature, performance of DGPS in static mode is very satisfactory. While using DGPS in static mode for accurate boundary description, canopy lifting/opening may be resorted to so as to enable the GPS antenna to lock to minimum four satellites.

The advantages of GPS traverse over conventional survey are many. Operational time, manpower and cost are reduced considerably. Closing error which is so common in, compass and chain survey, does not exist at all in static GPS traverse, as because one directly gets the nodes of the polygon. In case of a large traverse, time is further saved by downloading the data into computer.

Researchers at Ohio State University have developed the concept of GPSVan, a mobile mapping system. Such concepts can be easily put to utility in forest areas without involving any overhead costs. GPS can be mounted on a vehicle and forest areas such as roads, fire lines, compartment and reserve boundaries can be surveyed and mapped. Mapping of streams and channels can be made possible by mounting the instrument on a motor/country boat, Thus, with the help of GPS, locations of watch towers, camp sites, depots, administrative units, plantation sites, bio-diversity hot-spots, migration route of animals also can be mapped with great accuracy and little operational cost.

By using DGPS in static mode, it is also possible to reconstruct forest boundaries in terms of absolute co-ordinates (say in WGS 84). These co-ordinate pairs, then, can be directly used to position the reserves accurately on the map. Thus, ‘With the help of GPS one to one mapping is established from map to ground and vice-versa. Accuracy of the digital maps will increase further.

In order to use GPS more effectively, it is necessary to have Ground Control Points (GCPs) whose co-ordinates in WGS 84 datum (or in any other accepted system) are known with a high degree of accuracy. In order to facilitate quick mapping, there should be established at least one such GCP in every district. The onus of maintaining such GCPs may be vested in the Survey of India.

For such detailed information, the process of digital mapping must be that of bottom to top approach. Mapping at the ground end would generate large volumes of geo-spatial data, which necessitates the concept of seamless integration of digital maps.

bCan Maps be Viewed in Isolation?
The earth’s surface is a vast continuum of various geographic features. Maps represent projections of these features for a particular area on a particular scale. A series of maps may be produced to map the earth’s entire surface. But, paper maps have inherent limitations. To illustrate the case, let us take the example of Majuh, the biggest river island of the world, situated in Jorhat district of Assam. Approximately, it has an area of 1,000 sq. krn. The Survey of India mapped the island on 1:15,840 scale in 1968. For grid referencing, the whole area was divided into grids referenced horizontally as BO to BZ, and vertically I I to 22. Thus, a total of 144 topographic sheets were produced for the entire island, which is a large number.

Maintenance, storage and indexing of such large number of paper maps is not possible in normal office environments. Shelf life of these maps is also very short as they are subject to mutilation and soiling. Therefore, there is no doubt that switching over to digital maps is the only solution. However, digital maps also suffer from certain limitations, Size of these maps is directly related to computer memory (RAM), storage space (disk drive and back up capacity) and processor speed (chip type and its clock speed). A digital map file of size 100MB may take from several minutes to more than a hour just only to load, depending upon the computer configuration. Larger file sizes increase idle time considerably. Thus, digital maps too have limitation on size just as paper maps. A large number of map files of smaller sizes, say I to 30 MB need to be generated. What if these maps cannot be tiled or mosaiced with one another for want of memory and processor speed? The very purpose of digital mapping would be defeated.

 
Virtual Mosaics
The way out is to employ effective drawing management. A powerful drawing management technology has been provided in the AutoCAD Data Extension (ADETm) which has the capability to access, query and display simultaneously multiple maps. Maps may also be nested.Thus, the forest map (originally digitized from paper maps on 4 1 Mile) along with all object data becomes available for the entire state to be queried, viewed, analyzed and saved as a single file in a drawing session. It is interesting to note that digitization/vectonization may be carried out for all the maps and data saved in a number of individual files, each representing the smallest management unit. At higher levels of nesting, maps can be viewed by simply attaching connected files in a drawing session. To effectively use this scheme of drawing management, maps are to be planned in advance to draw a group of related geographic entities and objects in a single file.

However, as the mapped area increases, the map co-ordinate system needs to be changed. The drawings at Level I to 3 can be based on individual polyconic projections as followed by Survey of India. However, in state level maps, the co-ordinate system may be changed to UTM or any other acceptable datum. The drawing management technology of ADETM allows automatic switeliing over and co-ordinate transformation from one system to another in attached drawings. Thus, seamless virtual mosaics of maps is achieved from bottom to top in digital maps.

The discussion in the preceding paragraphs of the section mainly concentrated on integration of vectors. But digital maps are more than a collection of vectors in different colours, they also contains object data and attributes. Not only each type of object data should preferably be placed on a uniquely defined layer with its own colour, line, thickness and pattern type, but also map should contain attribute and data attached to map objects either internally as a part of the drawing or externally through a DBMS. It is essential from the point of view of ease in working, to evolve standard codes to define layers, data tables and object data fields. Else a great chaos would result by nesting a large number of files. Such standards of data input need to be evolved by all concerned in the field of GIS. Standardization of object and data codes for digital maps requires development of a detailed system of nomenclature which should be comprehensive (to save memory and storage space) and at the same time be applicable to all geographical regions.

Map indexing schemes have little relevance in digital maps. Maps from different regions may be attached in a drawing session of ADETM , and they would appear as a single map. Using this multiple file access technology, GIS users can incorporate geo-spatial data of other regions in their drawing. For example, a GIS user in Bangladesh studying effect of deforestation on the floods caused by the Brahmaputra and the Ganges rivers, may incorporate geo-spatial maps of the North-East India, Tibet, Nepal and North-West India into his drawing session. However, such a state of affairs would require international understanding and regional co-operation.

Beyond the Political Boundaries
Regional co-operation is required especially in the mountain eco-systems where a delicate balance exists between man and nature. As because disturbance in one region may adversely affect the eco-systems elsewhere, mountain regions need to be viewed from a broader perspective. Forestry management, wildlife management, bio-diversity conservation and many such issues are inter-related in a mountain system. There is already a welcome move in the direction of transboundary wildlife and bio-diversity conservation in the Indo-Bhutan region. In the changing scenario, GIS technology can play a meaningful role in conservancy and sustainable use of natural resources, only if there is free exchange of geo-spatial database among members of the region, Institutions such as ICIN40D should take lead in development of accurate database, and their exchange, especially in the forestry, wildlife and ecology in the Hindu Kush Himalayan region.

References

  • M. K. Yadava, IFS , Working Plans – The Main-stay of GIS in Forestry, MapIndia 98, 7-9 October, 1998, New Delhi,
  • CI Liu, Richard (Dick) Brantigan Using Differential OPS for Forest Traverse Surveys, Can. I For. Res., 25:1795-1805 (1995)
  • Kurt Novak , Mobile Mapping TechnologyfOTGIS Data Collection, Photogrammetry Engg. & Remote Sensing, Vol. 61, No. 5, May 1995, pp.493-501.
  • K.R. Srinivasan, GIS/GPS Technology: The Road Ahead, [email protected],development, May-June, 1998, p.7
  • Ajay Rastogi, Transboundary Co-operation for Bio-diversity Conservation and Management, ICIMOD Newsletter., No. 31, Autumn, 1998, p 10.
  • AutoCAD Map Reference Manual.