FOREST COVER MAPPING AND GROWING STOCK ESTIMATION OF INDIA S FORESTS GOFC-GOLD Workshop On Reducing Emissions from Deforestations 17-19 April 2007 in Santa Cruz, Bolivia Devendra PANDEY Forest Survey of India, Dehradun, INDIA www.fsi.nic.in
Forests constitute a very important natural resource for India. It is a major land-use and occupy about 23% land area of the country (77 mn ha) but the actual forest cover is only about 20.64% (67.8 mn ha) India s population being more than a 1 billion the per capita forests is too low (600 m2) against world s average of 6000 m2 The pressure on forests is therefore high also because more than 70% India s population has rural based economy
SATELLITE IMAGERY INTERPRETATION LAB FOREST SURVEY OF INDIA, DEHRADUN
Forest Cover Assessments Over The Years in India Cycle Year of Assessment Satellite & Sensor Resolution Scale I 1987 LANDSAT MSS 80m x 80m 1:1million II 1989 III 1991 LANDSAT TM 30m x 30m IV 1993 V 1995 IRS-1B LISS-II 36m x 36m 1:250,000 VI 1997 VII 1999 IRS-1C LISS-III 23m x 23m VIII 2001 IRS-1C/1D LISS-III 23m x 23m 1:50,000 2003 IX IRS-1D, LISS-III 23m x 23m 1:50,000
AGENCIES USING R.S and GIS IN FORESTRY IN INDIA Forest Survey of India (FSI)- main agency for wall to wall mapping for the entire country every 2 years Some State Forest departments NRSA and Regional Remote Sensing Centers and SAC of Deptt of Space Private agencies
Forest Cover Assessment INPUTS Satellite data of the entire country from National Remote Sensing Agency (NRSA) IRS ID/IRS-P6 (23.5m spatial resolution) SOI Topographic sheets - 1: 50,000 METHODOLOGY Digital Interpretation/visual Ground Verification Minimum map able area is I ha OUTPUTS Forest cover maps on 1:50,000 scale in digital or hard copy form showing following forest cover classes: CATEGORY CANOPY DENSITY Very Dense Forest More than 70% canopy Moderately Dense Forest 40-70% Open Forest 10-40% Scrub Less than 10% in forest lands Mangroves
SATELLITE DATA (IRS ID/IC- LISS III) GEOMETRIC CORRECTION CONTRAST ENHANCEMENT SOI TOPOSHEETS SCENE SUBSETTING DENSITY SLICING UNSUPERVISED CLASSIFICATION / NDVI TRANSFORMATION MASKING FOREST &NON-FOREST AREAS EDITING POST CLASSIFICATION CORRECTION ACCURACY ASSESSMENT GROUND REFERENCE DATA LAND COVER MAP ( STATE WISE) OVERLAY OF BOUNDARIES FLOW CHART FOR FOREST COVER MAPPING
Forest Cover Assessment -2005 LATEST RESULTS Satellite data Resource sat-1 1 (IRS P6) LISS-III III Data period Nov/Dec 2004/ Feb 2005 Loss of Forest cover 728 Km2 Reasons of major losses - Tsunami (Dec 2004) -Diversion of forests for Irrigation project-dam and reservoir -Removal of root stock for fuelwood and charcoal - Human induced Forest fire- in the North Eastern States- slash and burn/ shifting cultivation practice
HIT BY TSUNAMI KATCHALL-ISLAND(Andaman & Nicobar Is) IRS-P6 AWiFS DEC. 21, 2004 IRS-P6-AWiF FEB.16,2005 PRE-TSUNAMI POST-TSUNAMI Loss in Forest Cover- 2,589 ha (DF- 2567 ha, OF- 22 ha)
HIT BY TSUNAMI TRINKAT-ISLAND (Andaman & Nicobar) IRS-P6-AWiFS DEC.21, FEB.16, 04 04 IRS-P6-LISS III JAN.4, 05 PRE-TSUNAMI POST-TSUNAMI POST-TSUNAMI Loss in Forest Cover- 287 ha (DF- 285 ha, OF- 2 ha
Accuracy assessment (Forest Cover) Accuracy assessment of forest cover is done by using Field inventory data - high resolution satellite data (5.8m) Out of the 8000 points inside forest, 3509 points were selected to provide spatial representation of the whole forest area for preparation of error matrix. Ground truth data for these points giving land use class at each points were recorded with the help of field inventory data and high resolution satellite data in an area of 1.0 ha. Over all accuracy level found is 92.03% in 2005 assessment.
Growing Stock estimation Inventory of forests Inventory of trees outside forests (TOF) -Inventory of trees in rural areas using high resolution satellite data (5.8 m) multi-spectral -Inventory of trees in urban area
Methodology of NFI since 2001 The basic goal is to estimate a national level growing stock (wood volume) on a two year cycle and improve the estimate in subsequent cycles. For this purpose, the country has been stratified into 14 physiographic zones- based on climate, vegetation, physiography Ten percent (60) districts are covered in a two year cycle. India has about 600 civil districts. The districts are selected randomly from each zone with probability proportion to size. Along with the Forest inventory, vegetation survey of herbs and shrubs is also carried out. Measurement of soil and litter carbon is also carried.
Physiographic Zones of India
Physiographic Zones on Forest Cover
Randomly Selected 60 Sample Districts for Inventory
Topographic sheets on scale 1:50,000 (15' 15' Grid) divided into 5 ' 5' Sub Grids
Marking of Plots 5 X 5 QUADRANT IS DIVIDED INTO FOUR GRIDS OF 2½ X 2½ INTERVAL EACH 2½ X 2½ GRID IS FURTHER DIVIDED INTO 1¼ X 1¼ INTERVAL GRID
Methodology of NFI since 2001 At grid centre a plot of 0.1 ha is laid out all parameters pre-decided decided for forest inventory are measured and recorded Four sub plots of 1 sq. m are laid out at all corners of the sample plot of 0.1 ha to collect sample for litter and humus and soil carbon Nested quadrates of 3x 3 m and 1x1 m for enumeration of shrubs and herbs are laid at 30 m distance from the center of 0.1 ha plot in all four corners to assess biodiversity. On average 4000 temporary sample plots are laid out every year in forests.
Parameters of Forest inventory Species and diameter class wise trees Crown Diameter Soil & leaf litter for estimation of carbon Listing Herbs & Shrubs for Biodiversity Indices Regeneration status Fire incidence Grazing incidence Presence of weeds Crop injury Information on bamboo
Inventory of Trees Outside Forests (TOF) TOF resource has become most important in today s s context in India as most of timber requirement of industries have to met from TOF In the present methodology, high resolution satellite data (5.8 m) is used to identify TOF patches and stratify the same into -block, -linear and -scattered strata After the stratification appropriate sample plots are laid on the ground for field inventory. On average 4000 sample plots are laid every year.
24 BLOCK PLANTATION OF EUCALYPTUS
TREES GROWING IN AGRICULTURAL FIELDS
AILANTHUS EXCELSA & ACACIA NILOTICA ON FARM BUNDS 29
Tamrind cultivated with Amla
METHODOLOGY FOR ASSESSMENT OF TREE OUTSIDE FOREST USING REMOTE S ENSING DATA Satellite Data India Map with District boundaries LISS III Data PAN Data District Boundary Toposheets Mosaic Geometric and Radiometric Correction & fusion Digitization of Green wash Fused Data Green Wash Map Green wash Masking out of Green Wash from Fused Data Classification of fused data without Green wash Trees in Group Scattered Water Bodies Linear Block Elimination of area < 0.1 ha. Classified Map Block Linear Scattered Block Linear Field data collection Generation of Random Points Data Analysis & Report Generation Scattered
Plot Size & Number of Samples in rural TOF per district Strata Block Linear Scattered Scattered (Hill) Plot size No. of Samples 0.1 ha 35 10x125 m 50 3.0 ha 50 0.5 ha 95 - Random points for block, linear & scattered stratum along with coordinates communicated to field units for survey - Sample points in field are approached by using GPS & data recorded in prescribed formats
UFS Block Block No 8 UFS Block Map
Field Inventory of Urban TOF Urban trees have mainly environmental functions- In India urban areas are categorized into 5 classes (strata) based on population Urban Frame Survey (USF) blocks of National Sample Survey Organization (NSSO) are taken as sampling units by FSI Optimum number of UFS blocks are selected in each district for the survey as follows If UFS blocks < 500 10 % selection min 20 blocks UFS blocks > 500 5 % selection min 50 or max 60 blocks Data is collected on the designed formats on various parameters
CLASSIFIED MAP- Trees outside Forests BLOCK LINEAR SCATTERED Overall accuracy of classification = 92%
CONCLUDING REMARKS There exists many gaps in the methodology aspect of inventory of forests and TOF as well as forest cover assessment. One important challenge is to integrate RS data with field inventory so that reliable estimate on wood volume/ biomass/ carbon stock can be made rapidly. Capacity building and continuous skill up-gradation to adopt new methods and technology.
THANKS