Mapping of Trees Outside Forests

Subhash Ashutosh
Deputy Director, Forest Survey of India, Dehradun
[email protected]

P S Roy
Dean, Indian Institute of Remote Sensing, Dehradun
[email protected]

Tree resources outside the forest serve a number of ecological and economic functions. Socio-economic settings in the tropics are integrated with these resources. In addition, they play a very important role in carbon sequestration. Under Inter-governmental Panel on Climate Change (IPCC) regime, special emphasis has been laid on TOF’s role in carbon budgeting. Compared to other natural resources, however, little is known about this resource on the large area basis. If it is seen in the context of regionally decreasing forest cover and increasing forest fragmentation, importance of tree resources outside the forests will further grow in future.

‘Trees outside forests – i.e. trees and tree systems found on agricultural land, on meadow and grazing lands, on unproductive lands and in cities and other human settlements, among other places- have numerous, often essential roles and functions.’

• they make critical contribution to sustainable agriculture, food security and diversification of household economies;
• they supply many products (including wood for fuel and construction, fruits, barks and food products) and services (e.g. biodiversity, habitat for wildlife, microclimate stabilisation);
• they protect crops and the soil against water and wind erosion, thus combating drought and desertification and protecting water resources;
• they provide shade and mark property boundaries; and
• rich in significance from a cultural and social perspective.

Trees outside the forest are a highly heterogeneous natural resource and because of that generally no coherent or regional planning is done to mange it. In India, most of the TOF resource is on agricultural lands. Apart from those planted under farm forestry, horticulture or under social forestry plantations, TOF are not major object of management.

However, with increasing realisation of significance of TOF, there is need for policy initiative to include trees outside the forest in natural resources planning and consider it as an integral component of landscape inventories. One of the first steps is mapping of the resource to allow quantitative and qualitative characterisation.

Methodology
Bijnor district of Uttar Pradesh has been chosen as the study area. The study area falls between 78° & 78° 56’ latitudes and 29° & 29° 46’ longitudes. IRS 1C PAN and LISS III data of the period December, 1998 has been used in the study. ERDAS IMAGINE 8.4.1 and ARC/INFO software on WINDOWS NT workstation platform were used for the study. Both the images were geometrically rectified with the help of Survey of India toposheets on 1:50,000 scale. PAN and LISS III images were fused using Brovey’s algorithm (Fig 1). During the reconnaissance survey, signatures of different tree formations, water bodies and other areas appearing dark on the PAN image like burnt agriculture fields, marshy areas etc. were picked up and several points on agricultural fields and tree cover were marked on the hard copy PAN image. A hierarchical approach of classification was adopted to pick up features of interest (Fig 2). The LISS III image was classified into tree cover, agriculture, settlements, water bodies and other dark appearing surface areas using maximum likelihood

Fig. 1: Fusion on Multi spectral and Panchromatic images

Fig. 2
algorithm of supervised classification. The PAN image was masked to remove water bodies and dark surface features using the classified image (pseudo) of LISS III. The PAN data was opened as pseudo image and using the ground

Fig.3: Map showing forest and TOF in Bijnor district

Fig. 5: Spatial characterisation of the TOF layer
reference data threshold grey value of PAN image was determined to identify slice of DN values corresponding to tree cover. The resulting classified image was visually analysed alongwith fused image of LISS III and PAN for further refinement of the classification. Editing was done to remove the misclassified patches and also to include the omissions. Further, for removing noise from the scene and for better cartographic presentation, isolated clusters up to 4 pixels were clumped and eliminated. Classified raster image was finally converted to vector for enabling GIS analysis of the layer.

Result & Discussion
The digital image processing and GIS analysis followed in this study provides a methodology for using multi spectral and panchromatic data for mapping of trees outside forests (TOF). The flow chart depicting methodology is shown in Fig 2. The mapping includes TOF patches of size greater than 0.014 ha (i.e. area of 4 pixels of PAN data). Ground validation of the classified output on 60 points/patches has shown complete agreement with the ground situation. However, omission of few patches was noticed in the mapping. On analysis it was found that small patches of size >>0.2 ha with low canopy density (< 0.2) could not be picked up in the mapping.

Qualitative characterization of TOF layer in vector form (polygons) was done with the help of GIS. The TOF polygons were grouped into different polygon size classes and also their distribution along roads and canals was mapped using buffer analysis (50m on the either side). This type of characterization of the resource may help in creating different stratum for field inventory purpose. For example, the entire TOF can be divided into five strata based on size of patches (three size intervals) and their distribution along roads and canals. Extent of TOF in the study area has been assessed as 2360.2 ha. Figure 3 shows Forest / Tree Cover in the district. Figure 4 shows enlarged views of parts of the TOF map. Figure 5 shows characterization of TOF layer into five components: a) stratum based on TOF patch size b) along the road side and c) along the canal.

Acknowledgement:
The author (1) gratefully acknowledges guidance and support given by Dr. Devendra Pandey and Dr. J.K. Rawat, the former and the present Director, FSI, respectively.

References

• Kleinn,C. 2000. On large-area inventory and assessment of trees outside forests. Unasylva, 200(51):3-10
• FSI 2000. State of Forest Report 1999. Dehradun
• Kleinn,C. 1999 Tree Resources Outside the Forest In Central America : A regional assessment approach based on remote sensing and field survey.