1

REMOTE SENSING IN FOREST MANAGEMENT Introduction Remote sensing technology has emerged as an important tool for assessing the forest resource rapidly. The satellite provides the best synoptic view of the forest landscape. Through satellite data, it is possible to assess the forest cover broken down to density classes and the other land cover classes accurately. Using multi-data scenes supplemented with ground data, it is also possible to find broad types of forests (deciduous or evergreen). The remote sensing satellite data therefore has become very handy for stratifying the forests into homogenous sections which make the field inventory efficient, accurate and cost effective. Using very high resolution satellite data, it is also possible to discern isolated patches of trees and other tree resources outside forests. India is currently carrying out biennial monitoring of forest cover using satellite data on 1:250,000 scale. However the system of forest management in India is almost 120 years old and remained as a state subject. The policy formulation and strategic planning is at the central level. For effective management of forest resources the various basic requirements and priorities are essentially of three tier system i.e., State level District level and Micro level information requirements. The priorities are 1. Estimation of forest growing stock 2. Forest inventory and volume estimation 3. Forest Type Mapping and Assessment of Distribution 4. Preparation of stock tables and yield calculation 5. Preparation of treatment/ zonation areas 6. Assessment of ecological and bio-diversity 7. Forest change and conversion studies 8. Forest damage assessment due to forest fires

In the first State of Forest Report of FSI 1987, the satellite data of LANDSAT having 80 m resolution was visually interpreted on 1:1 million scales. The minimum mappable area was 400 ha. In the present cycle, the data of Indian Remote Sensing Satellite RESOURCESAT- I (IRS-P6 sensor LISS-III) having resolution of 23.5 m has been digitally interpreted on 1:50,000 scale where minimum mappable area is 1 ha. This has helped in making the assessment more objective, and the analysis of the changes more accurate. The gaps/blanks in forests having more than 1 ha area which were earlier included in forest cover have been excluded. On the other hand, isolated forested patches of more than 1 ha area which was not captured earlier is included now. Further, maps being in digital form, it has become easy and fast to generate maps on any desired scale, of any specific area, for different users including State/UT Forest Departments (SFDs). The quality of satellite data (IRS- P6, LISS III), which is available since 2003 is also found to be superior in terms of radiometry and shadow compared to that of IRS 1C/1D which FSI was using until 2003. Digital interpretation has also made it possible to add one more canopy density class in the mapping of forest cover. Against the two density classes (dense forest between 40- 100% and open forest between 10-40%) earlier, it is now possible to classify forest cover in three density classes (very dense between 70-100%, moderately dense 40-70% and open 10-40%) since 2003. After reviewing the technological inputs in the forest cover assessments of the other countries, it appears that the spatial resolution and scale of interpretation currently used in India at the national and at the State level is optimal. Brazil currently uses 25 m spatial resolution whereas 250 m spatial resolution has been used in the project level assessments in other countries. It is hoped that with the application of GIS and GPS tools, generation of such maps would be convenient and geo referenced forest maps for the whole country showing the latest boundaries of recorded forest areas would be available on 1:50,000 scale in due course of time. Harmonization of Forest Cover of Previous Assessments With the adoption of new technology in 2001 when the assessment for the entire country was done digitally on 1:50,000 scale, a large number of additional forested patches were

captured. Since these patches were outside the areas of the previous assessments, they needed detailed ground verification. But the same could not be done comprehensively due to paucity of time and lack of human resource. As a result, many patches were wrongly interpreted. It has been possible to rectify these errors with detailed ground truthing in subsequent assessments, i.e., of 2003 and 2005. In addition, some rectification of the errors was done on account of radiometric and geometric errors of satellite data. The quality of satellite data (IRS- P6, LISS III), which is available since 2003 is also found to be superior in terms of radiometry and shadow compared to that of IRS 1C/1D which FSI, was using until 2003. 1) Forest Cover Monitoring/ Surveillance The entire process of assessment from procurement of satellite data to finalization of the Report rectification, interpretation, ground truthing, validation by State Forest Departments - takes almost two years. In the present assessment, the satellite data of Nov.-Dec. 2004/Feb. 2005 have been used. The change in the forest cover has been analysed by comparing forest cover of the current assessment with that of the year 2003. Forest cover includes all lands which have a tree canopy density of 10 percent and above with area 1 ha or more. The Minimum mapped area of 1 ha of the forest cover corresponds to the cartographic limit (a polygon of the size 2mmx 2mm) on a map at 1:50,000 scale. This definition is based on the scale of interpretation (1:50,000), the optimum size of cluster of pixels, resolution of digital satellite data used for the mapping (pixel size 23.5m x 23.5m) and the technique employed for image interpretation. Satellite Data and its Period The satellite data for the entire country was procured from the National Remote Sensing Agency (NRSA), Hyderabad in digital form. It was a multispectral (LISS-III sensor) data of IRS P6 satellite with a resolution of 23.5 m. One scene of LISS III covers an area of about 20,000 km (140 km x 140 km). There are considerable overlaps (15 to 20 percent) among adjacent scenes. Also at the border of the country or for islands, the whole scene has to be procured even though the area of

interest may be small. Thus a total of 321 scenes covering the entire country were procured. The period of satellite data is of utmost importance. The reflectance from the forest is dependent on the crown foliage and its chlorophyll content. A deciduous forest would, therefore, not give proper reflectance in leafless period. Thus, data of the spring-summer season for such forests is not suitable for interpretation. Further, during the rainy season, it is difficult to find cloud-free data; moreover, agricultural and like lands give similar reflectance as forest cover during this period. The satellite data of the period October to December is therefore the most suitable for forest cover mapping of our country. During this period, deciduous trees, which largely constitute India's forests, bear adequate foliage. This enables capturing of true reflectance from the forests by the satellite sensors. While procuring the data, only those scenes were selected where cloud cover was less than 10 percent. Methodology Using Digital Image Processing (DIP) software, the satellite data in digital form was downloaded on the workstations from the CDs procured from the NRSA. 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 and georeferenced Survey of India (SOI) toposheets on 1:50,000 scale. The methodology of interpretation involved a hybrid approach in which unsupervised classification (ISODATA algorithm) aided on-screen visual interpretation of forest cover was done. Normalized Difference Vegetation Index (NDVI) transformation was used for removing non-vegetated areas from the scene. Areas of less than one hectare, whether classified as forest cover within non-forest areas or blanks within forest cover, were excluded by appropriate DIP techniques. Degraded forests with tree canopy density less than 10 percent have been classified as scrubs, which do not form part of the forest cover. Shadow areas in the scenes have been treated

separately. Shadow regions on the images are highlighted using band ratio techniques. Forest cover classification of the totally obscure areas due to shadow or cloud cover has been done using the ground truth information. Mangrove forests have characteristic tone and texture on the satellite image. Their presence on the coastal areas makes them even more conspicuous. They have been, therefore, separately classified. Interpretation was then followed by extensive ground verification which took more than six months. All the necessary corrections were subsequently incorporated. Reference data collected through ground truth and field experience of the interpreter played an important role in delineating the forest cover patches and classifying them into three canopy density classes. Very Dense Forest All lands having tree cover with canopy density more than 70% Moderately Dense All lands having tree cover with canopy Forest density between 40% and 70% Open Forest All lands having tree cover with canopy density between 10% and 40% Scrub All forest lands with poor tree growth mainly of small or stunted trees having canopy density less than 10 percent Non-forest Any area not included in the above classes

Accuracy Assessment Accuracy assessment is an independent exercise under which randomly selected sample points on the satellite imagery were verified on the ground and tallied with interpreted data. In this assessment 4,291 points randomly distributed over the entire country were selected and the verification was done using field inventory data as well as high resolution satellite data (5.8 m) of these points. Out of the total 4,291 sampling points, the classification of 3,949 was correct. The overall accuracy level of the assessment therefore comes to 92.03%. Limitations of Remote Sensing Technology Since resolution of data of LISS-III sensor is 23.5 m, the linear strips of forest cover along roads, canals, bunds and railway lines of lesser width are generally not captured. Young plantations and species having less chlorophyll contents in their crown cannot be delineated as forest cover. 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 or ground truth.

Gregarious occurrence of bushy vegetation and certain agricultural crops, such as sugarcane, cotton, lantana, etc., often poses problems in delineation of forest cover, as their reflectance is similar to that of tree canopy. Status of forest cover in India Class Area (km) Forest Cover a) VDF b) MDF c) Open Total Forest Cover* Non-forest Cover Scrub Non-forest** Total Geographic Area 38,475 2,571,700 3,287,263 1.17 78.23 100.00 54,569 332,647 289,872 677,088 1.66 10.12 8.82 20.60 Percent of G e o g r a p h i c Area

* Includes 4,445 km under mangroves ** Excludes scrubs and includes water bodies

Fig. Forest Cover of India

Pictorial illustration of different classes of forests

2) Change in the Forest Cover The change in forest cover is analyzed with respect to forest cover of previous assessment. Besides the change between forests to non-forest and vice versa, the change within the forests, between different canopy density classes has also been analyzed. The main reasons of loss in the forest cover are: Nagaland - shifting cultivation and gregarious flowering of muli bamboo A&N Islands - destruction due to tsunami Manipur - shifting cultivation Madhya Pradesh - submergence of forest area due to construction of dams

Chhattisgarh - submergence of forest area due to construction of dams, felling of trees Gujarat - large scale felling and uprooting of Prosopis juliflora trees from private and revenue lands by local people. Assam - illicit felling in insurgency affected areas and due to shifting cultivation Similarly the main reasons for gains in the forest cover in the States mentioned below are : Mizoram - re-growth in the abandoned shifting cultivation areas in all the districts Arunachal Pradesh - re-growth in the abandoned shifting cultivation areas and partly due to plantations Meghalaya - re-growth in the abandoned area of shifting cultivation Tamil Nadu - plantations of miscellaneous tree species in many districts

10

11

12

3) Forest cover and Forest area The term 'Forest Area' (or recorded forest area) generally refers to all the geographic areas recorded as forest in government records. Recorded forest areas largely comprises Reserved Forests (RF) and Protected Forests (PF), which have been constituted under the provisions of Indian Forest Act, 1927. Besides RFs and PFs, the recorded forest area may include all such areas, which have been recorded as forests in the revenue records or have been constituted so under any State Act or local laws. On the other hand, the term 'Forest Cover' as used in the 'SFR' refers to all lands more than one hectare in area, having a tree canopy density of more than 10%. Thus the term 'forest area' denotes the legal status of the land as per the government records, whereas the term 'forest cover' indicates presence of trees over any land. Although most of the recorded forest area has vegetation cover on it, yet there are blanks and areas with density less than 10% within it. On the other hand, there are areas outside the recorded forests with tree cover of more than 10% canopy density. Such areas also constitute forest cover and are included in the forest cover assessment of FSI. Therefore, the changes taking place in the forest cover is not necessarily due to changes within the recorded forest area but because of changes outside recorded forest area also. It would be, therefore, quite useful if the area of forest cover within and outside recorded forest area is known. This will not only facilitate policy, planning and management of forest areas by the State Forest Departments but also benefit other agencies of Government, NGOs, private landowners and village communities in tree planting efforts. Mapping of forest cover within and outside the forest areas can be done only if the geocoded boundaries of the recorded forests are available at an appropriate scale (say 1: 50,000). There is no proper account of forest cover outside recorded forest area. Since the SFDs are the custodians of the details of the recorded forest areas, so if the boundaries of such areas are made available, then the forest cover within and outside the recorded forests can be determined. FSI has approached all the SFDs for providing the boundaries of recorded forests under their jurisdiction but only a few States/UTs namely, Andhra Pradesh, Kerala, Meghalaya and

13

Dadra & Nagar Haveli could provide boundaries that could be used in GIS mode. Area statistics of forest cover along with the changes within and outside the recorded forests have been generated for these States/UT using GIS. Depiction of forest covers within and outside recorded forest area

4) Forest Type Mapping and Assessment of Distribution METHODOLOGY: Methodology followed by FSI is GIS based. It involves two major steps(i) Preparation of Forest Type Reference Maps using existing thematic maps and other data. (ii) Preparation of final maps depicting spatial distribution of forest types by collecting new information through ground verification and using recent satellite data and incorporating the same on Reference Maps.

14

(i) Preparation of Forest Type Reference Maps FSI has huge data on forest resources in form of thematic maps (based on interpretation of aerial photographs and on 1:50,000 scale), field inventory reports and field forms, vegetation type maps (prepared in 1981-83 on 1:2 million scale) and forest cover maps. It is envisaged to integrate these information in a single unit area (grid of appropriate size) to present the most likely forest type in a particular grid. For this purpose the whole country is divided into a geo-referenced vector layer of grids using Arc Info software (8.2version). The grid size of 2 x 2 which on an average encompasses 20 km2 of area, has been chosen as the most appropriate size. In the conventional field inventory of FSI, a systematic random sampling has been used where the area is divided into grids of 2 x 2 and all the ground information is collected from the sample plots laid out in this grid. The information of thematic maps available on 1:50,000 scale can be conveniently overlaid on these grids containing 36 grids of size 2 x 2. Firstly spatial database at country level is created having vector coverage comprising 171,028 grids (Fig. 34). A unique ID is given to each grid so as attach attribute data to it. The description of the attribute data attached to each grid as follows:

15

a. Forest Inventory of FSI: Since its inception in 1965 as PISFR (Pre- Investment Survey of Forest Resources), FSI has carried out field inventory of forest resources. In addition to enumeration of trees, information on soil, species composition, density etc. of each grid in the inventory area is also available in the field forms. FSI has covered more than 80% of countries forest area under field inventory. b. Thematic maps prepared by FSI: Till late nineties, FSI had been preparing thematic maps on 1:50,000 scale based on interpretation of aerial photographs. These maps in all show 48 forest land use classes and 15 non-forest land use classes. Under thematic mapping FSI has covered approximately 72% of the countrys forest area. Though these maps are old but they had provided valuable base line information on species composition as usually there is not appreciable change in species composition of area over a period of time. c. Forest species composition maps: For a study taken up by FSI recently, FSI had prepared forest species composition maps integrating information of thematic maps and inventory records. d. Vegetation type maps: FSI prepared vegetation type maps on 1:2million scale in 1981-83. These maps were partially based on aerial photographs (thematic maps) and partially on inputs received from State Forest Departments. These maps also have provided valuable information in preparation of forest type reference maps. e. Latest ground truth information: In accordance with its mandate, FSI carries out assessment of forest cover on a 2-year cycle and generates forest cover maps. These maps are based on interpretation of satellite data and extensive ground truthing is done before preparing final maps. In each cycle about 2500 points in the forest areas are visited for ground truthing. In the ground truth field forms, information on species composition is also collected. f. Soil Maps: Soil maps prepared by National Bureau of Soil Survey and Land Use Planning on 1:1million scale has been procured. These maps have been scanned and are being used to delineate forest types where soil composition is playing a significant role in classification of forest types.

16

g. Spatial Information on Climate: Information on temperature and rainfall on a 10 x10 grid size has been obtained from the Centre for Ecological Sciences, Indian Institute of Science, Bangalore. The mean annual values have been used as attribute climate data for each grid. h. Forest type maps prepared by other organisations: In the year 2002, IIRS (Indian Institute of Remote Sensing) produced Biodiversity characterization maps of parts of the country like Andaman & Nicobar Islands, North Eastern region etc., Similarly, mapping work has been done by the French Institute of Pondicherry in the Western Ghats region. These maps will be used to update the reference map. Fig. 35 describes methodology of Stage 1 in brief.

(ii) Preparation of final maps using new data: The grid based forest type reference maps are being used for extensive ground verification using GPS. During ground verification, ancillary information from the State Forest Departments as available in working plans, stock maps, species composition, forest types, soil, climate, past treatment etc are also being collected. The latest satellite data of IRS 1D (LISS-III)

17

with a spatial resolution of 23.5m X 23.5m for the year 2002 will be used for forest type classification. Multitemporal satellite data including that of dry season and high-resolution data will also be used in specific cases. Digital image processing is being done using ERDAS Imagine Professional software (latest version). Following steps are being used in the digital image processing: 1. Geometric corrections: Images (False Colour Composites-FCC) in digital forms are first downloaded onto the computers and then registered geometrically using appropriate numbers of GCPs (Ground Control Points) of corresponding Survey of India (SOI) toposheets on 1:50,000scale. The area of interest (i.e., the District) is then extracted by overlaying digitised boundary of the district. 2. Digital Image Processing: FCC of the extracted area is then digitally interpreted using a hybrid approach consisting of unsupervised as well as supervised classification using maximum likelihood classifier algorithm. Ortho-rectification and NDVI (Normalized Density Vegetation Index) programmes are also run wherever needed. Rule based classification has been used in areas where topographical features and soil play dominant role in deciding forest type. The NDVI is used not only to detect spatial pattern of biodiversity but also the distribution of biophysical parameters as difference in vegetation type are primarily due to variation in soil, moisture, temperature and rainfall. 3. Editing of Classified Forest Type Maps: Classified forest type maps will be further edited on the basis of ground information, data on soil, climate, altitude etc Based on this post classification editing and refinement, final forest type maps will be produced on 1: 50,000 scale. Fig 36 shows methodology of Stage 2 in brief.

18

(iii) Validation of Forest Type Maps: These maps will then be sent to all concerned State/UT Forest Departments for verification and validation. In case of variation, joint verification with the Forest Department will be carried out to correct the error if any. Since these maps are generated using a GIS based methodology, it will be possible to incorporate any change even if found at a later stage. 5) Forest Inventory and Sampling The forest enumeration and surveying is generally carried out at the district level for detailed working plan inputs preparation generally on 1:15,840 (4' = 1 mile) on predecided percent area enumeration (5% , 10% etc.) The working plan enumeration inputs is based on chain surveys across the forests on a systematic way. The conventional way of systematic sampling suffer due to accounting various strata and burdened with intense field work & expensive procedures, The satellite data facilitate in generation of primary stratification units at

19

the district level either through use of forest density maps or forest type classification it is possible to stratify the area into homogenous forest strata. Based on the prior knowledge the minimum sample size can be selected and the field inventory can be accomplished by suitable proportional allocation of samples in all the strata The stratified sampling procedures provide a reliable quantification of forest resources depending on the objective set during the inventory. 6) Forest Volume Estimation and Generation of Volume Tables The satellite based multi phase approach forest inventories with defined objectives provide ample data for further processing and computation of volume and yield tables. The enumerated data during the inventory could be systematically organised strata wise ad a suitable local and stand regressions could be generated for further computation of species wise volumes. The predominant species volume equations thus generated would form the base for computation of total standing volume based on the plot enumeration data. The standing volume information through inventory data analysis would form as a baseline data to bring out correlation with ground crown density maps. This in a way to say once the relationships are established which are generally local specific with respect to volume and density it is possible to estimate the total growing stock of the area by generation of stack maps and conversion through established volume functions. The experience in generation of such stock to volume estimations in western Gats has shown promising to explore to avoid cumber some, field inventories. The study also discussed in detail the importance of cull factor for estimating merchantable yield. 7) Stock Tables Prepartion and Yield Calculation From the stratified sampling and inventory data analysis it is feasible to calculate and generate stock tables with respect to number of species diameter and height wise distribution. Such information with the available knowledge of minimum girth area for extraction it is possible to calculate the yield and also to compute the annual coupes based on equiproductive areas. The technique of data analysis and generation of stock tables developed in parts working

20

plans project have revealed greater scope for extending towards management information system specifically for the working plans at district level on stock tables and yield calculations. 8) Ecological Consideration and Zonation of Forests With the increased pressure on forest all over the world as well as in India the growing concern of forest management has shown shift in its priority from production forestry to conservation Forestry. In the light of the the scientific management of forest and categorization of the area into different zones have been adopted by the Karantaka State forest department as part of Western Ghats Eco-restoration project. Towards this NRSA and Karnataka Forest Department jointly carrying out a project in a area of 10,280 sq. km in Uttar Kanara forest circle by generation of multi-thematic information on 1:25,000 scale using aero-space techniques. The thematic information generated are forest density and height maps, forest type maps, slope and aspect maps, drainage map upto first order channels, volume class maps besides consultation of contour maps with the available topographical sheets. The availability of such an enormous data base when feed into the Geographic Information System enhances interpretability of the data and to classify the area into different zones required for effective forests management practices on a scientific footing. Accordingly the forests are proposed for categorization of zones for specific management practices, the zonation is the process of describing the physical consequences of the management plans, which have been understood for sustainable forests resources in the area. It is essentially link between management practices. The zonation in the process of describing the physical consequences of the management plans which have been understood for sustainable forests resources in the area. It is essentially link between management objectives and physical operations on the ground.

21

The below given process com The study carried out in the parts of Western Ghats showed adequacy of the tool and approach. Forest Zonation Matrix RF D Zone 1 + Zone 2 + Zone 3 + Zone 4 + 4.5 3.5 2.3 H(M) Slope% Elevation (M) Village >20 >20 >20 >50 High Nil+ Type Managroves Grassland

>35-50 Medium >35-50 Medium >35 Rolling

Sparse/Nill Any Type Inhabited Habitated Any Type -DoW. L

<0.25 All

Zone 5 Reve All Lands Falling in "C" and "D" Lands

9) Remote Sensing and Biodiversity Studies The forest management while through its forest inventory and zonation of the area can revolve ecologically sensitive and diversity wish rich areas through use of satellite data applications in conjunctive use of other ancillary data discussed as above in the zonation. It is estimated that the tree diversity is estimated to around 200 species in the northern parts as Western Ghats exhibited a great diversity between the plots and along the gradients. The extended application of forest inventory and enumeration including shrub layer and herb layer would enhance forest management system for preservation of ecologically rich zones and to account diversity of the areas. The initial probabilistic diversity zones should be narrow down using satellite data and the consistency of patchiness as a function of landscape dynamics could be used as an element for bio-diversity studies.

22

10) Forest fire Forest fires are among the major causes of degradation of forests in India. As per an estimate by the Forest Survey of India (the State of Forest Report, 2005) which is based on field inventory, about 53% of Indias forest are prone to fire; of this 9% forest area is affected by frequent fire while in 44% have occasional fires. The ecological and socio-economic consequences of forest fires in India include loss of timber, bio-diversity, wildlife habitat, wood and other forest products, damage to water and other natural resources, loss of natural regeneration etc. Bahuguna (2002) estimated that average annual loss due to forest fires in country is about Rs. 440 crore (US$ 100 millions). Effective forest fire control measures become difficult in India due to lack of timely information on forest fire occurrence. With the availability of some satellites in recent past, it has now become possible to detect active forest fires in near real time and take remedial measures. MODIS (Moderate Resolution Imaging Spectroradiometer) is one such satellite system launched by the NASA (USA) which detects active forest fires. FSI has recently (since November 2004) taken initiative in active forest fire detection. Methodology: The methodology involves following steps: 1. Acquisition of Data from the website (www.maps.geog.umd.edu): The attribute information of active forest fire is collected from the website (Web Fire Mapper). Main attributes include geo-coordinates, temperature, confidence level etc. 2. Processing of the point Data: The points taken from the website are saved in the excel sheets. The points showing zero confidence limit are removed from the database and then converted into text (.txt) file. The txt file is converted into ASCII (American Standard Code for Information Interchange) format for further conversion into the arc coverage (point coverage) using Erdas Imagine 8.7 software. 3. Geometric correction: The point coverage so obtained is then geo-referenced into Geographic Lat\Long projection system.

23

4. Joining of Attributes: The points are joined with other attributes such as state, district, and SOI toposheet number using Arc Info 8.2 Software. These points are overlaid on the forest cover map prepared by FSI. Points, which are not falling over the forested region, are removed and points that are falling on the forest are retained for further processing. 5. Dissemination of information: Details of the points falling within the forest cover like date of occurrence, geo-coordinates, state, district, and SOI toposheet no. are transferred to the excel sheet. The information so generated is disseminated to the concerned States through FAX/e-mail for taking remedial measures.

24

References MoEF (2005) The state of Forest Report 2005 GOI. Loetsch, F and Haller, K.E. (1964) Forest Inventory, BLV Verlagsesselschagt, Munich. ECE (1986) The ECE Timber Committee year book -XL (1) United Nations Economic Commissio for Europe. Champion, M.G. and Seth, S/K. (1968) Revised Survey of Forest types of India , Govt. of India publication. Gadgil, M., Sinha M. Pillai J. (1989) A Biomass budget. Final Report of the study group on Fuelwood and Fodder, Planning Commission Govt. of India. Unni, N.V.;M. (1990) Space and Forest Management. Procedings of space and forest managemt, Special Current Even Session 41st, IAF Congress, Dresden, Germany. Weigand , C.L and Richardson A/J. (1990) Use of spectral vegetation Indices to inter leaf area, ET and yield : 1 Rationale Agron J. 82: 623 - 629- Annexure -1 Hush, Bertram Miller , C I and Thomas W. Beers (1982) 3rd edition "Forest Mensuration" John Wiley & Sons. Inc. KFD (1977) working plan for sirsi and Siddapur Karnataka State Govt. publication. FAO (1999) Remote sensing application to Land Resources. 14th UN/FAO Int. Trg, Course material. SAC (1990" Mannual of procedures for Forest Mapping and Damage Detection using satellite date ISRO - SAC-TR-88-90) Howard, John A, (1991) "Remote Sensing of Forest Resources". Chapman and Hall Edn. P.S. Roy (1994) "Space Remote Sensing For Forest ECO_SYSTEM Analysis" Paper presented at 3rd ISY-Regional Remote Sensing Seminar on Tropical Eco-system management held during 23-25th August, 1994 at Bali (Indonesia)