Different methods for ground-based assessment of trees outside forests

During the workshop, the participants deliberated on various issues related to the ground-based assessment of TOF. Various preferential rating scales were designed and discussed (see section on “Parameters, criteria and indicators”). A review of methodologies for TOF assessment is given below for a quick reference.

Review of methodologies for assessment of TOF

Assessment of TOF in Haryana in North India by Chaturvedi (1990)

Chaturvedi (1990) classified trees planted on non-forested lands as (i) homestead planting (ii) commercial tree planting, and (iii) tree planting on farmlands.

In homestead planting, all trees with breast girth of 15 cm or more were enumerated. Plants of lower dimensions were ignored. Measurement of girth at breast height was carried out by semi-skilled enumerators. The distribution in girth classes was very wide. No method of sampling could, therefore, deliver accurate results with acceptable limits. Since the trees were concentrated in a small area, it was convenient to carry out total enumeration rather than sampling. Apart from number of trees, this assessment provided data on growing stock in cubic meters, which was quite useful for planning raw material supply to industries, village artisans, etc.

Commercial tree planting was not found in the study site at Dhanawas, but 10 planting units were observed in Berka Alimuddin. The areas of these sites varied from 0.05 ha to 2.12 ha. Therefore, a total count of trees by girth classes was carried out.

Sampling of tree planting on farmlands was based on three independent samples of 30 drawn from random numbers taken from the Fisher and Yates Tables (1963).

This was reportedly the first study of its kind where three villages in Haryana State in North India were selected for assessment of TOF in three types of landscapes. However, the criteria used to select the three villages out of a total of 6,988 villages were not described. In India where there are over 650 000 villages, total enumeration may not be feasible, as it would require a great amount of time and resources to complete the work. This study also did not consider other landscapes such as sides of roads, canals, drains, railways, ponds, and other community areas with substantial tree cover. Thus, poor representation of all landscapes is a major weakness in this study. The classification may be applicable on a small scale but not for a national-level measure of TOF.

Study of TOF by Kerala Forest Research Institute

Krishnankutty (1992) adopted a stratified three-stage sampling procedure. The percentage of dry land area under agricultural use to the total area under agricultural use and population density was calculated from the data available from the State Land-use Board and the 1981 census report, respectively. In all, 15 strata were formed on the basis of five classes of percent dry land area under agriculture use and three population density classes. The villages were then classified under the different strata accordingly. Revenue villages in each stratum were treated as first stage units of sampling. The villages in each stratum were chosen at random.

Although the study recorded the precise number of trees in homesteads, it failed to acknowledge the importance of trees planted on roadsides and those growing on barren lands. Furthermore, trees on lakesides and backwaters were also not included in the assessment. However, this was reportedly the first such effort to systematically enumerate wood resources in the homegardens of Kerala. This study provided the background of methodology and accorded adequate importance to the tree resources growing in the homegardens outside the public forests. This assessment provided good account of the contribution of homestead gardens in meeting domestic requirements of fuelwood, small timber, and some inexpensive construction material for households. Thus, the study not only enumerated trees, but also calculated the volume from recorded tree dimensions. This information on raw material availability can have significant influence in planning sustainable management of natural forests. It is also relevant for estimating the potential of homestead gardens for domestic use of timber and non-timber products and cash income in the event of emergency such as marriage of children and other ceremonies.

The Karnataka experience of Dr Ravindranath and Dr Somasekhar (Indian Institute of Science, Bangalore)

Ravindranath and Somasekhar (1995) adopted methodologies to appraise the distribution of tree species on various farm sizes in a semi-arid Ungra in Tumkur District of Karnataka. They studied tree diversity and density on farms, the location-specific nature of the tree species, the rationale for the choice of species and the possible end-uses. To investigate the impact of landholding size and economic status on tree-growing practices, a stratified random sample of 30 households was selected to include 6 farmers in each of the following categories: landless, marginal (<1 ha), small-scale (1.01 to 2 ha), medium size (2.01 to 3 ha), and large-scale (>3 ha).

This methodology gives a better sampling technique to enumerate TOF in rural settings. However, urban areas and the inventory of urban trees are excluded. In addition, the methodology does not consider railway and roadside trees, which are also important TOF resources in India.

The emphasis of this study was mainly on tree diversity and density in terms of species and number of trees. The study did not collect information on height, girth, and therefore, did not assess the volume of trees on farmlands.

The FSI Experience

The FSI began a regular inventory of TOF in 1991 based on stratified random sampling. A district (or a group of districts) in a state was treated as a stratum and villages as a final sampling unit. The number of sample villages to be surveyed in a state was decided by a pilot study. The boundary and area of each of the randomly selected villages were obtained from the Revenue Department and all trees standing in the selected villages were physically counted and measured. The team leader was provided with the list of sample villages to be inventoried, along with maps of their locations (at 1:50 000 scale). The entire village was divided into suitable angular quadrants in such a way that the inventory of each quadrant could be completed in one working day. Generally, living trees above 10 cm in diameter were measured and counted but in some states trees up to 5 cm in diameter were also measured. In Haryana State, out of a total of 7 000 villages, 290 villages were selected for TOF inventory.

The inventory would have taken four years to complete. To accelerate the process, the methodology was modified in 1999. Analysis of the completed inventory of non-forested areas gave significant positive correlation between village area and number of trees and also between the population size of villages and number of trees. This relationship has been utilized to fix the sample size of other states and dispense with the additional task of conducting a pilot survey. States are being divided into agro-ecological zones and each zone is being treated as a stratum. The number of sample villages in each zone is decided on the basis of proportional allocation. The precision level is slightly reduced because of reduction in the sample size. Further, the proportion of trees enumerated in a village was also reduced by 50, 25 and 10 percent if their number exceeded 2 000 and lies between 2 000 to 5 000, 5 000 to 10 000 and more than 10 000, respectively. With the modification, the inventory in four states - Andhra Pradesh in the south, Gujarat, and Rajasthan in the west and West Bengal in the east - is nearing completion. The number of villages selected in these states were 110, 142, 141 and 167 out of a total of 27 999, 18 509, 39 810 and 40 890 villages, respectively.

This sampling method covers a wide array of landscapes and tree types, but the information on the inventory and sampling method for urban areas needs further elaboration. The amount of time taken for the exercise is also a constraint. Ground verification is needed as the trees in rural areas might already be harvested for fuel and fodder by the time the survey took place in the village. It is virtually impossible for one agency to enumerate the TOF in the whole of India. The size of the country and time constraint are the biggest hurdles for large-scale inventories.

This study has been more elaborate than other studies reported in terms of sampling intensity, area coverage, and number of the tree units per village.

Study of TOF from the socio-ecological point of view by Malhotra and Kumar (1987)

Malhotra and Kumar (1987) studied the socio-ecological aspects of the avenue trees in Calcutta, particularly tree density, species diversity and associated human activities. The methodology involved the selection of a 4.6 km stretch of road in Calcutta and a complete census (species and number of trees and the related human activity) of all the planted trees separately on both sides of the road. The human activities associated with each tree were also recorded. The authors found that the avenue trees in urban centers not only performed important ecological role, but also played significant socio-economic and religious roles in the lives of the urban dwellers. Several hawkers/vendors set up their stalls selling fruits, vegetables, sweetmeat, etc. under the shade of trees growing in the urban areas. The authors emphasized the need for a systematic collection of data on trees along the roads, parks and residential areas in other cities also.

The study stresses on the social aspect of the avenue trees only, and fails to extend the inventory to include an array of sites and urban landscapes apart from roadsides (such as trees on institutional lands and household trees). Therefore, the method that has been tested on roadside trees cannot be adequately applied to trees on farm bunds and plantations. Moreover, the study focused on the socio-ecological functions of the TOF and therefore has restricted value for its replicability. The study does not provide the total volume of trees as present along the roads.

Assessment of TOF through sampling and secondary information by the Indian Institute of Forest Management

The IIFM started the work on a national assessment of TOF as a collaborative project with over 30 NGOs from across the country in June 2000, based on secondary information and sample field survey. It was considered desirable in view of limitation of resources and lack of well-defined methodology for sampling. These NGOs and partner institutions were selected on the basis of their fieldwork on the subject in the respective areas.

Secondary information was collected from the following sources:

¨ Land Record Office
¨ District Statistical Office
¨ District Rural Development Agency (DRDA) offices
¨ District Collectorate
¨ Municipal Corporation offices (parks superintendent)
¨ Research institutions in the area
¨ Village-level interviews
¨ Tourism Department (for maps)
¨ Various plantation companies
¨ Public sector coal companies who raised compensatory afforestation
¨ NGOs engaged in tree plantation and watershed works
¨ District Forestry Office (social forestry/research and extension)

1. The State of Madhya Pradesh was broadly divided into three categories: urban, semi-urban (including district headquarters and small towns), and rural areas.

2. The State was divided into five main zones: North, East, West, Central, and Southeast.

3. Five major cities in the urban areas were chosen from each of the following zones: Gwalior (North), Jabalpur (East), Indore (West), Bhopal (Central), Bilaspur (Southeast).

4. One district headquarter in the semi-urban areas (near the selected city, i.e. the nearest possible semi-urban area was taken irrespective of the distance) was selected for each zone.

5. Four villages within 10 km from the district urban center, one each in the North, East, South and West of the district urban center, were selected in each zone.

The target urban area is divided into four main categories depending on the ownership pattern as follows:

¨ Landscapes were broadly divided into residential, institutional, roadside, and garden/pond/park/temple.

¨ Zones: 4 quadrants were laid in each zone of the selected city, i.e. East, West, North, South, Central.

¨ Data collected include number of trees, species, girth classes, and height classes.

a. Residential

Sample plot size: 100x100 m quadrant was taken in each zone of the city, and the trees were totally enumerated in the selected quadrant along with an estimate of the number of households.

b. Institutional

1. If an institution has an area of less than 3 ha, total enumeration was to be carried out.

2. In institutions with more than 3 ha, 20x5 m quadrants were laid in the institutions’ precincts where the trees were planted.

3. Zonal criteria: 3 institutions in each zone (total of 15 in a city) i.e. East, West, North, South, Central.

The Cardinal Grid Method (CGM^©) was applied and tested in Gwalior City, which was divided in three zones, namely Gwalior City, Lashkar and Murar. The method was discussed at length during a brainstorming session.

The study appeared to be based on systematic sampling, which could give rise to biased results. Extrapolation of results should be considered. Participants emphasized the need for random sampling in the initial stage to minimize potential errors. The selected grids were based on the spatial layout of Gwalior, and are interchangeable. The city was divided into directional grids, the center being the densest populated area. The importance of selection of samples through the ring method by giving examples leading to extrapolation of the number of trees was discussed. The relationship could be assessed through a semi-variogram model. The proposed CGM^© method needed to be further refined based on the field experiences and suggestions of the participants.

Garden/pond/parks/temple side

These were selected on a case-by-case basis depending on the occurrence of trees.

Sampling unit: 20x5 m, four quadrants per hectare of the target land

Striplands

Table 1. Available striplands in India

Land category	Potential (million km)
Roads	1.2
Railway lines	0.6
Drains	0.2
Canals	1.5
Around cultivated lands on bunds and boundaries	60.0

Source: Govt of India, 1982.

Four transects of 10 × 100 running meters were laid in each direction (North, East, West, South) on both sides of the city’s state highway roads.

Canals

One transect of 10 × 100 running meters were laid every 3 km length on both sides of the canal.

In rural areas, there are a number of unclassified landscapes with trees. The procedure to inventory this kind of vegetation is important, but is not included here. The CGM^© is applicable only to urban areas.

Analysis of the methodologies

The methodologies provide various ways for assessing the TOF resources. The first method justifies the stratification of landscapes but fails to include the roadside and railside trees. Each of the methodologies described above provides a unique dimension to the enumeration process. Malhotra and Kumar considered roadside trees but not trees on farmlands. The FSI methodology suggested total enumeration of trees in the study area, although it is clearly not feasible to inventory a diverse and vast country like India completely. Therefore the ideal method for TOF assessment would be a judicious mixture of methodologies that would cover both urban and rural areas, be representative enough to reduce errors, and be cost and time effective.