1 CHAPTER 2 FOREST COVER 2.1 Introduction Forest cover, as explained in the previous chapter, includes all lands more than 1 ha area having tree canopy density of 1 percent and above. The basic data for forest cover is obtained from the remote sensing satellite whose sensor captures reflectance of sun light from the tree canopy in multiple bands. In such data, with present technique and skills, no distinction with respect to the tree species has been attempted. Moreover, no cognizance of the type of land ownership or land use or legal status of land was taken as such information was neither felt necessary nor possible to collect at country level. Thus, all species of trees (including bamboos, fruits or palms, etc.) and all types of lands (forest, private, community or institutional) satisfying the basic criteria of canopy density of more than 1 percent have been delineated as forest cover while interpreting satellite data. The minimum area of 1 ha for forest cover has been kept because of technological capability of the satellite/sensor and the technique of data interpretation used. Methodology employed for assessment of forest cover using digital technique and various steps involved are described in the following sections. 2.2 Satellite Data and its Period The present assessment is based on digital interpretation of satellite data for the entire country. The satellite data was procured from the National Remote Sensing Agency (NRSA), Hyderabad in digital form. For the present assessment LISS-III sensor data of IRS-1C satellite with a resolution of 23.5 m has been used. Data for nearly all the states pertained to the period from October-December 2. Only in case of Madhya Pradesh and Maharashtra, the satellite data of October-December 1998 was used. While procuring data, only those scenes were selected where cloud cover was less than 1 percent. 2.3 Digital Interpretation Using Digital Image Processing (DIP) software, data was downloaded on the computer. Radiometric and contrast corrections were applied for removing radiometric defects and for improving visual impact of the false colour composites (FCC). Geometric rectification of the data was carried out with the help of scanned SOI toposheets. Based on tone and texture the forest cover areas were delineated. Interpretation of forest cover for the whole country was done at 1:5, scale in the present assessment. Areas less than 1 ha were excluded by clustering pixels. Density classification of forest cover was done by Normalized Difference Vegetation Index (NDVI) transformation. Shadow areas in the scenes were treated separately. Classification of forest cover was done as dense forest and open forest. Mangroves, conspicuous on the imageries due to their
2 characteristic tone, texture and location were delineated separately. Mangroves were further classified into dense forest and open forest, based on NDVI transformation and were added up in their respective forest cover classes. The interpretation was then followed by extensive ground verification that was followed by incorporation of corrections. The methodology has been shown schematically in Figure 2.1. All the classified scenes of a state were mosaiced and state wise area was extracted by using state boundary. Vector layer of forest area boundaries, obtained by digitizing the boundaries of forest and green wash areas from SOI toposheets, were overlaid on classified image to determine area of forest cover within and outside the digitised forest area. Similarly, vector layer of district boundaries were used to compute district-wise area of forest cover. Figure 2.1: Mapping Flow Chart Showing Methodology of Forest Cover 2.4 Advantages of Digital Interpretation over Visual Interpretation Digital Image Processing (DIP) technique offers a more objective assessment of forest cover at a larger scale and has better cartographic presentation, thereby overcoming the limitations of visual interpretation to a large extent. Some important advantages are highlighted below:
3 (i) Since visual interpretation is based on individual interpretation skill and experience, there is always a possibility of subjectivity in the assessment while in computer based digital interpretation, individual subjectivity is minimized to a great extent. (ii) Method of computation of area is different in visual interpretation from digital interpretation. In the former, it is done using dot-grid templates by manual calculation, while in digital method, it is done by computer software and is based on number of pixels in each class. In visual method, considering the number of FCC sheets that could be handled manually, FCC at 1:25, scale were interpreted. Due to cartographic limitations, the smallest area on sheet that could be interpreted visually is 2x2 mm that represents an area of 25 ha on the ground. Thus, it was not possible to delineate isolated patches of forest cover having an area less than 25 ha. Similarly, blanks of less than 25 ha in forested areas also could not be seen. On the other hand, due to adoption of digital image processing technique at 1:5, scale in the present assessment, isolated patches of forests or blanks up to 1 ha in extent (represented by 2x2 mm on screen at this scale) could be delineated. Advantages of the forest cover mapping on larger scale would be evident from Figure 1.2 which shows that the details, not discernable on 1:25, scale, are highlighted on 1:5, scale. Due to change in scale of interpretation alone, the area of forest cover in large forested regions may be reduced as the openings which were not discernible on the smaller scale are easily picked up on the larger scale. Conversely, in other areas, the area of forest cover may show increase as small patches of forests are picked up on the larger scale, which were not discernible on smaller scale. (iii) In digital method, the computer calculates the area under a particular class by counting number of pixels in that class. In visual method, on the other hand, area computation is done by counting the grids on a dot-grid plate manually and there are chances of human error in this operation. 2.5 Limitations of Remote Sensing Technology However, there are still certain limitations with remote sensing technology when used for assessment of forest cover. Some of the major ones are listed below: Since resolution of data from LISS-III is 23.5 m, the linear forest cover along roads, canals, bunds and rails of width less than the resolution are generally not recorded. Young plantations and species having less chlorophyll contents in their crown do not give proper reflectance and as a result are difficult to be interpreted correctly. Considerable details on ground may be obscured in areas having clouds and shadows. It is difficult to interpret such areas without the help of collateral data.
4 Variation in spectral reflectance during leafless period poses problem in interpretation. Gregarious occurrence of bushy vegetation, such as lantana, sugarcane, etc., often poses problems in delineation of forest cover, as their reflectance is similar to that of tree canopy. 2.6 Forest Cover: 21 Assessment Results of present assessment (21) of forest cover are summarized for the country in Table 2.1 and shown in a pie-chart in Figure 2.2. It gives dense and open forest cover in the country. As mentioned earlier, dense and open forest cover also include mangrove cover of the corresponding density class. The total forest cover of the country as per 21 assessment is 675,538 km² and this constitutes 2.55 percent of the geographic area of the country. Of this, 416,89 km² (or percent) is dense forest cover while 258,729 km² (or 7.87 percent) is open forest cover. The non-forest includes scrub estimated to cover an area of 47,318 km². The over all accuracy of classification, as assessed by creation of error matrix, is 95.9 percent (See Annexure II). The country map of forest cover is given in Figure 2.3. Forest Cover Assessment 21 Dense Forest Open Forest Non Forest Total Forest Cover = 2.55 % Table 2.1: Figure 2.2: Forest Cover Status of Forest Cover in India Class Area (km²) Percent of Geographic Area Forest Cover a) Dense 416, b) Open 258, Total Forest Cover* 675, Non-forest Scrub 47,
5 Total Non-forest 2,611, Total Geographic Area 3,287, * Including 4,482 km 2 under mangroves (.14% of country s geographic area) 2.7 State/UT wise Forest Cover The State/UT wise forest cover in the country is shown in Table 2.2 and as bar chart in Figure 2.4. It shows that Madhya Pradesh with 77,265 km² has the maximum area under forest cover, followed by Arunachal Pradesh (68,45 km²) and Chhattisgarh (56,448 km²). Considering proportion of geographic area under forest cover, the Union Territory of Lakshdweep, due to inclusion of coconut plantations within forest cover, has the maximum percentage (85.91 percent). It is followed by Andaman & Nicobar Islands (84.1 percent), Mizoram (82.98 percent), Arunachal Pradesh (81.25 percent) and Nagaland (8.49 percent). The ranking of States/UT according to these criteria is given in Table Forest Cover Open Forest (in ' sq.km.) Dense Forest 2 1 Figure 2.4: Forest Cover in States and UTs Table 2.2 Forest cover in States/UTs in India in km²) (Area
6 State/UT Andhra Pradesh Arunachal Pradesh Assam Bihar Chhattisgarh Delhi Goa Gujarat Haryana Himachal Pradesh Jammu & Kashmir Jharkhand Karnataka Kerala Madhya Pradesh Maharashtra Manipur Meghalaya Mizoram Nagaland Orissa Punjab Rajasthan Sikkim Tamilnadu Geo-graphic Area 275,69 25,827 83,743 53,932 78,438 15,83 94,163 3, ,191 37,88 1, ,72 1, ,22 8,673 44,212 1,139 55,673 1, ,236 11,848 79,714 11, ,791 26,156 38,863 11,772 38,245 44,384 37,713 3,894 22,327 5,71 22,429 5,681 21,81 8,936 16,579 5, ,77 27,972 5,362 1, ,239 6,322 7,96 2,391 13,58 12,499 Forest Cover Dense Open Total Percent 18,81 14,113 11,884 2,348 18, , ,931 9,389 1,85 1,835 3,788 32,881 16,588 11,216 9,93 8,558 7,952 2, , ,983 44,637 68,45 27,714 5,72 56, ,95 15,152 1,754 14,36 21,237 22,637 36,991 15,56 77,265 47,482 16,926 15,584 17,494 13,345 48,838 2,432 16,367 3,193 21, Scrub 9, , , , ,452 6, , , ,18
7 Tripura Uttar Pradesh Uttaranchal West Bengal Andaman & Nicobar Chandigarh Dadra & Nagar Haveli Daman & Diu Lakshdweep Pondicherry Total 1,486 3,463 7,65 3,62 24,928 8,965 13,746 4,781 53,483 19,23 23,938 4,915 88,752 6,346 1,693 4,347 8,249 6,593 6, ,287, ,89 258, , , Additional Analysis based on Forest Cover Information for the Country Once the spatial and statistical information on forest cover for the country, States/UTs and districts is available, it can be analysed to generate several other kinds of information. Three such analysis have been attempted in this report, viz. (i) forest cover in the hill districts of the country, (ii) forest cover in the tribal districts of the country, and (iii) distribution of forest cover within and outside the recorded forest areas. These data for the country and all the States and UTs are presented in the following sections. 2.9 Forest Cover in Hill Districts The National Forest Policy (1988), aims at having a minimum of one third of geographic area of the country under forest and tree cover and enjoins maintaining two third of the area in hills under forest cover in order to prevent erosion and land degradation and also to ensure maintenance of ecological balance and environmental stability. It is therefore felt desirable to know the extent of forest cover in the hill districts in the country. With this objective FSI started assessing forest cover in the hill districts of the country since The classification of hill states, districts and talukas as adopted by the Planning Commission is based on the criterion of an area having an elevation of more than 5 meters above mean sea level. The Planning Commission has applied this criterion for Hill Areas and Western Ghats Development Programmes. Since forest cover assessment is done taking district as a unit, only those districts have been categorised as hill districts where the total area of hill talukas exceeds 5 percent of the geographic area of a district. The abstract is
8 given in Table 2.3. The information on forest cover in these hill districts (marked H in District wise forest cover tables) pertaining to the present assessment is given in Chapter 6. There are 123 districts in the country that can be classified as hill districts on the basis of the criterion explained above. The total forest cover in the hill area of the country is 271,326 km² constituting percent of the geographic area, against the goal of 66 percent as laid down in the National Forest Policy Out of total 123 hill districts, only 51 districts have forest cover more than 66 percent. Of the rest, 33 hill districts have forest cover less than 66 percent but more than 33 percent and the remaining 39 districts have even less than 33 percent forest cover (including 11 districts having less than 1 percent forest cover). Table 2.3: State/UT State/UT wise Forest Cover in Hill Districts Number of Hill Districts Geographic area in Hill Districts Forest Cover (Area in km²) Percent Total Forest Cover Dense Forest Open Forest Arunachal Pradesh 13 83, Assam 3 19, Himachal Pradesh 12 55, Jammu & Kashmir , Karnataka 6 48, Kerala 1 29, Maharashtra 7 69, Manipur 9 22, Meghalaya 7 22,429 5,681 9,93 15, Mizoram 8 21, Nagaland 8 16,579 5,393 7,952 13, Sikkim 4 7,96 2, , Tamil Nadu 5 22, Tripura 3 1,486 3,52 3,563 7, Uttaranchal 13 53,483 19,23 4,915 23, West Bengal 1 3,149 1, , Total , ,771 95, , Forest Cover in Tribal Districts Importance of forests in tribal economy is well known, as they are a source of subsistence and livelihood for the tribal communities. It is commonly believed that the tribal communities live in harmony with nature and protect forests. Assessment of forest cover in tribal areas therefore acquires a special significance. Since the 1997 assessment, FSI is regularly providing information on forest cover in districts identified as tribal districts under the Integrated Tribal
9 Development Programme of the Government of India. In addition, all the districts of Arunachal Pradesh, Manipur, Meghalaya, Mizoram, Nagaland, Sikkim, Tripura, Dadra & Nagar Haveli and Lakshdweep have also been included in the list of Tribal districts owing to high tribal population. The abstract is given in Table 2.4. Information on forest cover situation in these tribal districts (marked T in District wise forest cover tables) for the current assessment is given in Chapter-6. Out of 593 districts in the country, 187 districts have been identified as tribal districts. The present assessment reveals that the total forest cover in these tribal districts is 44,87 km². It constitutes percent of the total geographic area of the tribal districts. Of the 187 tribal districts, forest cover is over 66 percent in 53 districts, between 33 and 66 percent in 43 districts and less than 33 percent in the remaining 91 districts (with less than 1 percent in 28 districts). The forest cover in the tribal districts constitutes 59.8 percent of the total forest cover of the country whereas the geographic area of 187 tribal districts forms only 33.6 percent of the total geographic area of the country. It demonstrates that tribal districts are generally rich in forest cover, and hence forest resources. Enhanced investments in forestry activities can be used as an instrument for rapid economic development of tribal communities. Table 2.4: State/UT State/UT wise forest cover in Tribal Districts Number of Tribal Districts Geographic area in Tribal Districts Dense Forest Forest Cover Open Forest (Area in km²) Percent Total Forest Cover Andhra Pradesh 8 87,9 17,62 8,339 25, Arunachal Pradesh 13 83,743 53,932 14,113 68, Assam 16 5,137 7,233 5,73 12, Chhattisgarh 9 9,134 27,852 13,322 41, Gujarat 8 48,65 5,85 2,486 7, Himachal Pradesh 3 26,764 2,12 1,23 3, Jharkhand 8 44,413 7,826 5,83 13, Karnataka 5 26,597 1,9 2,419 12, Kerala 9 27,228 9,274 3,42 12, Madhya Pradesh ,448 27,883 13,935 41, Maharashtra ,272 18,656 1,126 28, Manipur 9 22,327 5,71 11,217 16, Meghalaya 7 22,429 5,681 9,93 15, Mizoram 8 21,81 8,936 8,558 17, Nagaland 8 16,579 5,393 7,952 13, Orissa 12 86,124 19,8 13,832 32, Rajasthan 5 38,218 2,343 3,79 6, Sikkim 4 7,96 2, , Tamil Nadu 6 3,72 3,198 2,87 6, Tripura 3 1,486 3,52 3,563 7, Uttar Pradesh 1 7,68 1, ,
10 West Bengal 11 69,43 6,18 4,22 1, Andaman & Nicobar 2 8,249 6, , Dadra & Nagar Haveli Daman & Diu Lakshdweep Total 187 1,13, ,48 147,39 44, Forest Cover vis-à-vis Forest Area It is noticed that a common reader generally does not distinguish between forest cover and forest area whereas these are two different entities. A land may be recorded as forest area and under management of forest department but may not have any discernible forest cover. On the other hand, all wooded lands or plantations, delineated as forest cover from satellite data may not be legally recorded as forest area as these could be private plantations or institutional wood lots. Although, majority of forested lands happen to be within legally recorded as forest areas, all the changes taking place in the forest cover is not necessarily due to changes in the forests managed by the forest departments. Therefore, it is important from policy and planning point of view to know the extent and quality of forest cover within recorded forest areas and outside it. An exercise can be attempted to make such estimation. This information will be important and useful for the concerned forest department, civil administration and others. With availability of GIS tools, such an exercise would be very convenient if latest geo-referenced forest maps for the whole country showing the latest boundaries of recorded forest areas were available at 1:5, or 1:25, scales. However, the forest departments generally have cadastral forest maps, which are not geo-referenced. Can some alternative approach be used for this exercise to overcome this problem? The Survey of India (SOI) toposheets show forest areas with double dotted lines and wooded lands as green wash patches. Most of the green wash patches are within forest boundaries but some patches that are not notified as Reserved or Protected Forests may not be surrounded by such boundaries. It can be presumed that these patches would mostly represent Unclassed Forests where the legal status of the forest land is yet to be finalised. However, the SOI toposheets are usually based on surveys done several decades back and may be at variance with the current status of forest boundaries. So the question is, do the SOI maps provide nearly complete information on recorded forest areas and can these boundaries be used as proxy for current forest boundaries? If yes, then data layer of these forest boundaries can be overlaid on the forest cover maps and information on forest cover within and outside the recorded forest area can be generated using GIS tools. When this exercise was attempted and forest areas within the boundaries digitised from SOI toposheets at 1:25, was computed, it was found that for many states and union territories it differed significantly from the area of recorded forests obtained from the concerned forest departments. For instance, in case of Delhi state, the SOI maps did not show any forest boundaries or green wash areas whereas its recorded forest area is 85 km 2. In case of Himachal Pradesh, the area
11 digitised from SOI toposheets was only one third of the recorded forest area. Apparently, boundaries of many protected forests were not shown on the maps. In case of Haryana, Punjab and Uttar Pradesh where strips along roads, canals, bunds, etc. are notified as protected forests but were not shown in the toposheets as such, the digitised area was much below the recorded forest area. In case of West Bengal and Maharashtra also the area as per SOI was lower than recorded forest area. In West Bengal, partly the reason is perhaps the way mangrove forest area in the Sundarbans is recorded. The Forest Department includes the rivers and creeks between islands as forest area but these are not shown as such on the toposheets where only the mangrove islands are shown with green wash. In Maharashtra, many forest areas were simply not shown in the SOI maps. On the other hand, in several states, such as Meghalaya, Arunachal Pradesh, Mizoram, Nagaland, Orissa, Kerala, Goa, etc. the forest area assessed from SOI maps substantially exceeded the actual recorded forest area. The main reason being that certain green wash patches without forest boundaries, presumed to capture unclassed forest, are in fact institutional forest lands managed by village councils or panchayats. Another reason could be that some of the forest areas may have been de-recognised or diverted for other uses after the SOI maps were made. Due to these problems, use of SOI toposheets as proxy for actual forest maps would not provide correct information for many states and the country. The data generated on forest cover inside and outside recorded forest areas by this exercise would be erroneous for many states and UTs. However, in order to reemphasize the importance of such information, indicative figures obtained for a few states can be quoted here for illustration purposes. For example, the area estimated from SOI maps for the states of Tamil Nadu, Karnataka and Assam were reasonably close to the respective recorded forest areas of the states. Presuming that SOI maps correctly captured the actual forest areas, it was found that, respectively, about 31%, 27% and 25% of the forest cover in Tamil Nadu, Karnataka and Assam lay outside the boundaries of forest areas. These figures highlight the contribution and role of trees outside forests. This can also be seen as indicator of success of afforestation projects such as social forestry and wastelands development. Another interesting information that can be obtained from this exercise is to know how much of the recorded forest area is without any discernible forest cover. For the states of Tamil Nadu, Karnataka and Assam it is found that, respectively, about 35%, 3% and 23% of the recorded forest area is has no forest cover. Even though these are only approximate figures, these provide insight into the status of forest and forest cover in these states. Similar situation may exist in other states and union territories as well. These indicate that there are large chunks of land inside departmentally managed forests that require protection, regeneration, afforestation and restocking. In order to get more accurate information on forest cover within and outside forest area, statistically as well as spatially, it is necessary for all the State/UT Forest Departments to prepare, on their own, the latest geo-referenced forest maps at 1:5, scale showing all categories of recorded forest areas. With
12 the advent of Geographical Positioning System (GPS) and easy availability of these gadgets, such an exercise in forest mapping would not be very time consuming or expensive. It may also be necessary for the Forest Departments to have forest cover maps at divisional or sub-divisional levels (up to 1:5, scale) to know the distribution of forest cover and to identify blanks or degraded areas inside forest boundaries. FSI can assist the Forest Departments in producing such maps.