GISdevelopment.net ---> AARS ---> ACRS 1990 ---> Forestry

Vegetation classification system using remotely sensed data for Viet Nam's condition

Phan Phu Bong
Centre of Geography and Natural Resources, NCSR
Nghia Do, Tu Liem, Ha Noi- Viet Nam.


Abstract
Viet Nam is situated in the tropical monsoon zone of south Asia . The physiographic conditions are convenient for the plants development. On the other hand, Viet name is a place of crossing region of many migratory lines of plant from Indonesia, Malaysia, Myanmar and China. That's reason, the flora and vegetation of viet name is abundant and varied forms. At the present time, natural vegetation of Vietnam have been destroyed and have changed into a complex mosaic of secondary plant replacement communities. In this report, a vegetation classification system using Remotely Sensed data has been proposed for Viet name conditions. Based on the Remotely Sensed data has been proposed for Viet names conditions. Based on the Remotely Sensed data and ground vegetation of Viet name were classified as follow, was presented:
  • Closed evergreen forest
  • Deciduous forest
  • semi-deciduous forest
  • Mangrove forest
  • Woodland
  • Scrub
  • Herbaceous Vegetation
  • Vegetation on coastal sand dunes
  • Cultivated vegetation, forest plantations
  • and others
Also author will introduce a fragment of vegetation map of Viet name applying this classification system.

Introduction
Alarming data about the effect of human impact on natural vegetation, including the degraded and changed seriously tropical forest areas, therefore, make vegetation inventory and monitoring and more and more important. On the other hand, it is very necessary to inventory and to assess of natural conditions and natural resources for purpose proposing a rational utilization of natural resources. Regional planners are sometimes required to give quick or immediate answers to solve problems of planning and rational use of natural resources. They, therefore, need to know what vegetation types are present in an area. First of all, it is the most important to have a vegetation map of the area. Remotely sensed data are a major element in vegetation classification and mapping. It has been observed that the major physiognomic types, such as forest, woodland, scrub, grassland, mixed vegetation, could be distinguished using remote sensing data, which readily record the spectral characteristics of such vegetation cover.

In this report, based on the remotely sensed data which have been used in Viet nam, a classification system of vegetation using remotely sensed data has been proposed for Viet nam's condition and a primary result applying this one to compile vegetation map of VN.

Existing methods of vegetation classification
Plant communities have many characteristics features, that it is not possible to use them all in devising a method to classify vegetation. In the last about 100 years, various systems have been developed to describe and to classify vegetation cover. Schimper's vegetation classification system was the first one (Schimper, 1898). During the last century, many classification systems of vegetation have been given to describe and to classify vegetation on the world as well as regions and countries. However, it would be impracticable within the framework of this report to discuss them in detail. Here follows the list of the main vegetation classification systems which have been given during the last century: Warming (1909) , Tansley and Chipp (1926) Kuchler (1949), Elenberg (1956) Densereau (1957), Fosberg (1958, 1961) Trochain (1957, 1961), Gaussen (1959, 1967) Holdrige (1967) Elenberg and Mueller-Dombois (1967, 1974)............etc and at the global level -UNESCO'S International classification and mapping of Vegetation (UNESCO, 1973) provides a comprehensive framework for the more important categories to be used in vegetation maps at scale of 1/1,000,000 or smaller . In their work, P.Legris,F.Blasco (CNRS, University of Toulouse, France) and K.D.sing (FAO,Rome,1985)have summarized of existing classification systems for vegetation in Asia and Pacific Region.

In Vietnam, we can mainly refer to the vegetation classification systems by Chevalier (1918) , P.Maurand (1943) Rollet (1962) M.Schmid (1956,1974) Thai Van Trung (1963, 1970, 1978) , Tran Ngu Phuong 1970 and Phan Ke Loc (1984).

However, such classification can not be used in interpreting ACRS D-7-3 Remotely sensed data. It can provide general indicative information about the nature of vegetation types.

Based on Raunkiaer's life-forms, Howard (1969) developed and classification based on remote Sensing to describe vegetation formations and sub-formations in Eastern Australia andEast-Africa using Serial photographs. Howard,J.A. and Shade, J. (1982) introduced a standardized hierarchical classification of vegetation for remote sensing. Howard , J.A.,Z.D. Kalensky and F. Blasco (1985) presented conee-concepts for Global Mapping of woody vegetation using Remote sensing data. Anderson and other (1976) , Maling Greau (1977) introduced land cover/ land use classification use with Remote Sensing data in Indonesia…….

However, these vegetation classification are not suitable for Viet nam's conditions.

Vegetation classification system of Viet Nam
Viet nam is situated in the East of the Indochina peninsula, in the tropical monsoon area of South East Asia. The annual average rainfall varies from one region to another, from 1,500 mm to 2,500 mm(min. about 600 mm, and max about 45 00 mm ) The landform is complex . Average temperature changes from 23 oc to 27 oc . Because of its intermediate position , Viet nam contains elements of the floras of Indonesia, Malaysia, Myanma, China and India. This is results in an abundant and varied flora and vegetation in the country. However, after the prolonged destruction of the war and human impact, the natural ecosystem of Vietnam have been destroyed and have changed former vegetation into a complex mosaic of secondary vegetation replacement communities .The application of the classification systems mentioned above proved unsuitable for the classification and the mapping of actual vegetation of Viet nam.
  1. General principle classification of vegetation using Remotely Sensed data
    In general, principle classification of vegetation using remotely Sensed data is based on the spectral nature of remote sensing data. The spectral nature of remote sensing data is one of the major element for their application to vegetation studies. Remote sensing is usually concerned with the characterization of plant formations in terms of stand height and crown cover. The height of the stand is considered to be most consistent physiognomic parameter to describe a formation (Raunkiaer,1934; Kuchler, 1967). When remotely sensed data are the principal data for mapping purposes which is normally the case for large area mapping programmes, the information about land cover that can be directly derived from the imagery should form the basis for a classification system. A vegetation classification system should be designed according to the following principle:

    • The classification system should have categories that could be directly interpreted from imagery, taking advantage of the ground information as well as the ancillary information available.
    • The level of interpretation accuracy should be high.
    • The results can be repeated by a number of interpreters using imagery of different seasons

    The distinctions based on height classes are as follows (FAO, 1982) :

    >7m Forest
    0,5cm to 7m Scrubs

    The crown cover classes are listed below :

    >40 percent Closed forest
    10-40 percent open forest
    <10 Per cent Non-forest

    The each unit of classification of vegetation was described by its physiognomic structural characteristics with some supplementary ecological information .The units of vegetation were interpreted has based on the different intensities of the colour tone of the image and each one was described according to the results of the field check.

  2. Classification system
    Based on the remotely sensed data which have been used in Vietnam including : LANDSAT MSS, SPOT image, and SOYUZ photos, a classification system of vegetation using remotely sensed data has been proposed for Viet nam's conditions . The units of vegetation are mainly delineated through visual interpretation of remote sensing data and some supplementary data from field surveys:

    Classification system of Vegetation :

    Forest: -Closed Evergreen Forest.
    -Deciduous Forest.
    -Semi-deciduous forest.
    -Closed Evergreen Forest on limestone
    -bamboo mixed with evergreen Forest.
    -mangrove Forest.
    -Melaleuca Forest.
    -Pine Forest
    Woodland(Clear Forest): -Evergreen Wood land
    -Deciduous Woodland
    Scrubs: -Evergreen Scrubs
    -Deciduous Scrubs
    Herbaceous Vegetation: -Herbaceous Vegetation on coastal sand dunes
    -Herbaceous and Scrubs Vegetation
    Cultivated Vegetation: -Plantations Forest
    -other Plants Cultivated Vegetation

    Because, the landform of Viet nam is complex and strongly divided, vegetation changed as with altitude zones and latitude zones. Applying this classification system, therefore, interpreters must be payed attention to features of landform and characteristics of climate in the humid tropical moonson region.
Fig.1: Shows Vegetation Map of coastal zones of lower basin Mekong delta (Vietnam) compiling by using this classification system.


Fig. 1.: Vegetation Map of coastal zone of lower basin Mekong delta using remotely sensed data.


Conclusion
This is the first classification system of vegetation using remotely sensed data for Viet nam's conditions. This classifications system is set up based on the remotely sensed data which have been used in Vietnam. The units of vegetation in classification system are delineated based on the different intensities of colour tone of the image through visual interpretation. When applying this classification, interpreters must be payed attention to features of land form and characteristics of climate of Viet nam and results of the field check. The limitations found in this system were that the units of vegetation were delineated only through visual interpretation and that processing was done without digital analysis.

References:
  • Howard, J.A and J.Schade. Towards the global remote sensing of vegetation change -hierarchical classification(Draft paper, FAO,Rome, Italy 1981.)
  • J.A.Howard, Z.D.kalensky and F.Blasco. Concepts for global mapping of woody vegetation using remote sensing data. FAO, Rome , Italy , 1985.
  • UNESCO International Classification and Mapping of vegetation UNESCO, Paris,France,1973.
  • P.Legris, F.Blasco and K.D.Singh. Classfication and vegetation mapp Mapping in Asia and Pacific Region . FAO Rome Italy 1985.
  • RRSP/ ESCAP and SEAMEO-BIOTROP. Remote Sensing in vegetation studies , (Report of the ESCAP-BIOTROP). Remote sensing techniques applied to vegetation studies) Bogor, Indonesia, 1985.
  • Thai Van Trung . Tham Thuc Vat rung Viet nam,Hanoi, Viet nam, 1963 , 1978.
  • A. Singh. A Forest cover classification system using remotely sensed data. Asian Pacific Remote sensing Journal , Vol,1,N02 Jan 1989 Bangkok Thailand
  • Kuchler, A.W. VEGETATION MAPPING, The Ronald Press Company , New york, 1967.
  • Raunkiaer, C.The life forms of plants and statistical plant geography . Clarendon Press , Oxford,U.K.1934.