Subscribe Now Subscribe Today
Research Article

Application of Two Way Indicator Species Analysis in Lowland Plant Types Classification

Yahya Kooch, Hamid Jalilvand, Mohammad Ali Bahmanyar and Mohammad Reza Pormajidian
Facebook Twitter Digg Reddit Linkedin StumbleUpon E-mail

A TWINSPAN classification of 60 sample plots from the Khanikan forest (North of Iran) is presented. Plant types were determined from field observations and sample plot data arranged and analyzed in association tables. The types were defined on the basis of species patterns of presence, absence and coverage values. Vegetation was sampled with randomized-systematic method. Vegetation data including density and cover percentage were estimated quantitatively within each quadrate and using the two-way indicator species analysis. The objectives of the study were to plant type`s classification for Khanikan lowland forest in North of Iran, Identification of indicator species in plant types and increase our understanding in regarding to one of Multivariate analysis methods (TWINSPAN). Five plant types were produced for the study area by TWINSPAN, i.e., Menta aquatica, Oplismenus undulatifolius, Carex grioletia, Viola odarata and Rubus caesius. Therefore, at each step of the process, the program identifies indicator species that show strongly differential distributions between groups and so can severe to distinguish the groups. The final result, incorporating elements of classification can provide a compact and powerful summary of pattern in the data set.

Related Articles in ASCI
Similar Articles in this Journal
Search in Google Scholar
View Citation
Report Citation

  How to cite this article:

Yahya Kooch, Hamid Jalilvand, Mohammad Ali Bahmanyar and Mohammad Reza Pormajidian, 2008. Application of Two Way Indicator Species Analysis in Lowland Plant Types Classification. Pakistan Journal of Biological Sciences, 11: 752-757.

DOI: 10.3923/pjbs.2008.752.757



Multivariate analysis methods assume that several random variables have interred relation between each other and that they are important for the researcher (Ajbilou et al., 2006). Classification and ordination are two main branches of multivariate analysis and in this study as classification method was used. Due to lack of statistical analysis, understanding the structure of plant species is associated with considerable mistakes, therefore, in the first step of each study, vegetation should be classified (Dollar et al., 1992).

TWINSPAN have designed on the basis FORTRAN computer program that use for study of ecological factors and species. Study of plant vegetation in natural sites with compare tables has known Dombois and Ellenberg (1974) handy method (Host and Register, 1991). Hill (1979) has modified Dombois and Ellenberg (1974) handy method to binary clustering analysis. Afterwards with progressive in computer program, this method had displayed with name of TWINSPAN.

Main idea of TWINSPAN is on the basis the first hypothesis of phytosociology that group of quadrates will distinct by group of different species. This species will settle in binary tables. Namely, quadrates have compared each other on the basis presence or absent of species and pseudo species. The quadrates that have more similar to each other are beside each other (White and Hood, 2004).

TWINSPAN simultaneously classify species and samples (Grant and Loneragan, 2001). The program is geared towards ecological data. At its core, TWINSPAN is based on dividing a reciprocal averaging ordination space. The procedure was described in accessible terms by Gauch and Whittaker (1981) and White and Hood (2004). One of the most useful features of TWINSPAN is the final ordered two-way table. Species names are arrayed along the left side of the table, while sample numbers are along the top. The pattern of zeros and ones on the right and bottom sides define the dendrogram of the classification of species and samples, respectively. The interior of the table contains the abundance classes are defined by pseudo species cut levels (White and Hood, 2004).

Currently TWINSPAN is probably the most frequently used procedure for the classification of community data sets (Sebald, 1964; Spies and Barnes, 1985; Kent and Coker, 1996; Mc Cune and Mefford, 1999; Mesdaghi, 2001). Fundamental to TWINSPAN is the concept of pseudo species. Although information on the abundance of species is retained, the pseudo species scaling will reduce is that it allows the occurrence and abundance of species to treat as essentially separate variables, recognizing that both can be important characteristics of an association (White and Hood, 2004).

Also, the most popular divisive method (and program of the same name) TWINSPAN (two-way indicator species analysis) was partially inspired be the classificatory methods of the classical phytosociology (use of indicators for definition of vegetation types). The idea of an indicator species is basically a qualitative one. Consequently, the methods work with qualitative data only. In order not to loose the information and the quantity of species, the concept of pseudo species and pseudo species cut levels was introduced. Each species can be present as several pseudo species, according to its quantity in the sample. The pseudo species is present, if the species quantity exceeds the corresponding cut level. Imagine that we selected pseudo species cut levels (Waite, 2000).

The present study is, however, the first comparatively extensive one on the botany of Khanikan lowland forest (North of Iran) which seems to be interesting from the biodiversity viewpoint. The present study, therefore, deals with the species composition and life form structure of this forest. It also aims at identifying the prevailing plant types using technique of multivariate analyses (TWINSPAN). The objectives of the study were to plant type`s classification for Khanikan lowland forest in north of Iran, identification of indicator species in plant types and increase our understanding in regarding to one of Multivariate analysis methods (TWINSPAN).


Study area: Khanikan forests are located in the lowland and midland of Mazandaran province in north of Iran with the area of 2807 ha. (Between 36° 33` 15", 36° 37` 45" latitude and between 51° 23` 45", 51° 27` 45" longitude). The maximum elevation is 1400 m and the minimum elevation is 50 m. Minimum temperature in December (7.5°C) and the highest temperature in June (24.6°C) are recorded, respectively. Mean annual precipitation of the study area were from 698.2 to 102.5 mm at the Noushahr city metrological station, which is 10 km far from the study area. This research was performed in summer of 2006.

Data collection: In lowland region 268.7 ha of this forest was selected. For investigation of tree and shrub covers sixty quadrate (20x20 m AR.) (Hill, 1979; Hedman and Graceland Ling, 2000; Mesdaghi, 2005) and sub quadrate (1 m2 AR.) in each quadrate for investigation of herbaceous covers (Monier and Abd, 2000; Mesdaghi, 2005), were taken by randomized-systematic method. Quadrate size was determined for each vegetation type using the minimal area method (Cain, 1959). Considering variation of vegetation, floristic list and canopy cover percentage were determined in each quadrate.

Data analysis method: Data matrix of quadrates number and vegetation type was made. The windows (Ver. 3.0) of PC- ORD (Mc Nab et al., 1999) were used for classification of plant types. Data were analyzed by multivariate techniques as the Two-way indicator species analysis (TWINSPAN). TWINSPAN analysis is a numerical method for classification of vegetation belonging to similar groups. This allows the investigator to recognize the homogenous groups. Two-way indicator species analysis using the computer program TWINSPAN (Waite, 2000), a commonly used program in ecological studies. It is a robust technique, fairly impermeable to sampling errors or noise (Grant and Loneragan, 2001).

Identification of indicator species: Indicator species are those that occur predominantly among the samples at one end of an ordination axis. Their distribution should reflect, i.e., indicate, the environmental characteristic of the samples at either end of the gradient depicted by the ordination axis. Samples along the ordination axis are divided into two groups about the centre of the axis. Those to the right of the centre are placed in one group, called the positive group (+); those to the left are placed in a different group, called the negative group (-). Based on the distribution between these groups, an indicator value is calculated for each species using the formula (White and Hood, 2004):

Ij = nj+/N+ –nj/N

where, Ij, the indicator value for species j; nj+, the number of samples in the positive (right-hand) group containing species j; nj, the number of samples in the negative (left-hand) group containing species j and N+, N, the numbers of samples in the positive group and the negative group.


In studied area 56 species of 36 families were recognized that the number of woody species and herbaceous were 14 and 42 respectively (Table 1). Life forms were determined by Raunkiaer system and according to the biological spectrum, phanerophytes and cryptophytes (35.71%) and hemicriptophytes (28/57%)

Table 1:
Species, life form, endemic and family of recorded species in study area
form1: Ph: Phanerophyte. Cr: Cryptophyte. He: Hemicriptophyte. Chorotype2: H: Hyrcanian. M: Mediteranian. It: Irano -Touranian. Pol: Poly zona

were dominant life forms of the studied area. Also, vegetation chorology showed hyrcanian elements with 55.35% was dominant chorotype of khanikan lowland forests. Number of 8 species (14.287%) was endemic of Iran Flores (Table 1).

Five plant types were identified after the application of TWINSPAN programs (Table 2, 3). They are named after the characterizing species as follows: Menta aquatica (I), Oplismenus undulatifolius (II), Carex grioletia (III), Viola odarata (IV) and Rubus caesius (V) (Table 3). The most number of plant species (36) is relation to III plant type and the least of it (15) is relation to I, IV plant types (Table 4). Also, III plant type and I plant type had the most (23.7) and the least (8.3) mean of cover (%), respectively (Table 4).

The plant types were defined for the Khanikan lowland forests in North of Iran. It was the first attempt to develop such plant types in this area, thus making it impossible to compare this study with other studies. Five plant types are generated, in the present study, after application of two-way indicator species analysis (TWINSPAN) to the cover estimates of 56 species in 60 quadrates (Table 2, 3). Cluster analysis has performed for data of species covering (Table 3).

The first division, 60 quadrates have divided to two clusters that in left direction exists 56 quadrates and exists no indicator species. In right direction exists 4 quadrates with Menta aquatica indicator species. The second division, 56 quadrates have divided to two clusters that in left direction exists 23 quadrates with Brachypodium pinnatum indicator species and right direction exists 33 quadrates with Hedra pustuchovii, Euphorbia amygdaloides and Ruscus hyrcanus indicator species. The third division, 23 quadrates have divided to two clusters that in left direction exists 4 quadrates with Rubus caesius indicator species and right direction exists 19 quadrates with Viola odarata indicator species. The fourth division, 33 quadrates have divided to two clusters that in left direction exists 17 quadrates with Carex grioletia indicator species and right direction exists 16 quadrates with Oplismenus undulatifolius indicator species.

For determination of indicator species in classified types have used the formula (1). Thus, five plant types are named as follows: I plant type includes Menta aquatica with Carpinus betulus woody indicator species. II plant type includes Hedra pustuchovii, Euphorbia amygdaloides and Ruscus hyrcanus with Oplismenus undulatifolius. III plant type includes Hedra pustuchovii, Euphorbia amygdaloides and Ruscus hyrcanus with Carex grioletia. IV plant type includes Viola odarata, Brachypodium pinnatum with Parottia persica and Cratagus sp. Woody indicator species. V plant type includes Rubus caesius, Brachypodium pinnatum with Parottia persica, Cratagus sp. and Quercus castanifolia.

Result of this research have showed that III plant type had the most number of plant species (36) and I, V plant types had the least of it (15) (Table 4). Also, have

Table 2:
TWINSPAN of the vegetation cover in 60 quadrates and 56 species

showed that III plant type and I plant type had the most (23.7) and the least (8.3) mean of cover (%), respectively (Table 4).

The use of quantitative values (coverage values) was essential in defining and using plant types. There was a certain degree of overlap among the types; in almost all cases more than one type occurred in a given ecosystem. Such over lapping was observed in Northern Michigan ecosystem studies (Kashian et al., 2003) and in West Germany (Dahdouh et al., 2002). Nevertheless, some plant types were more characteristic than others of certain types of ecosystems.

Ecological classification and grouping of forest habitats was the main subject of forest management since

Table 3:
TWINSPAN indicator species for the first three level of division

Table 4:
Number and mean cover of species in plant types

of 1980 decade (Barbour, 1999). Many methods had used in order to classification of forest habitat (Bairley, 1978; Barnes, 1982) but, they couldn`t show, the relation ecosystem components very well. Because, the most of they, have been used of one component similar to soil or plant vegetations, alone (Dahdouh et al., 2002).

Using of two-way indicator species analysis in order to plant type`s classification have been dealt with in numerous works. Among the main ones are : Host and Register (1991), Dollar et al. (1992), Zahedi Amiri (1998), McNab et al. (1999), Witte (2002), Picard and France (2003) and Kashian et al. (2003).

Now days, multivariate methods have been using in systems of ecological classification, extensively (Witte, 2002; Picard and France, 2003). Traditional systems of classification, had used on the basis dominant species or species similarity that its basis was subjective (Verlinden and Dayton, 2005) but, the new systems of classification is on the basis objective methods that has high precision and researchers reaches to some results with this systems Ajbilou et al., 2006).

1:  Ajbilou, R., T. Maranon and J. Arroyo, 2006. Ecological and biogeographical analysis of Mediterranean forests of Northern Morocco. Acta Oecol., 29: 104-113.
Direct Link  |  

2:  Bairley, R.G., 1978. Nature of land and resource classification. J. For., 76: 650-655.

3:  Barbour, M.G., 1999. Terrestrial Plant Ecology. 3rd Edn., An Important of Addison Wesley Longman Incorporation, London, pp: 649.

4:  Barnes, B.V., K.S. Pregitzer, T.A. Spies and V.H. Spooner, 1982. Ecological forest site classification. J. For., 80: 193-198.
Direct Link  |  

5:  Cain, P.D., 1959. Manual of Vegetation Analysis. Harper and Brothers, New York.

6:  Dahdouh, F., J.G. Kairo, L.P. Jayatissa, S. Cannicci and N. Koedam, 2002. An ordination study to view vegetation structure dynamics in disturbed and undisturbed mangrove forests in Kenya and Sri Lanka. Plant Ecol., 161: 123-135.
Direct Link  |  

7:  Dollar, K.E., S.G. Pallardy and H.G. Garrett, 1992. Composition and environment of floodplain forests of Northern Missouri. Can. J. For. Res., 22: 1343-1350.
CrossRef  |  Direct Link  |  

8:  Dombois, D.M. and H. Ellenberg, 1974. Aims and Method of Vegetation Ecology. John Wiley and Sons Publisher, New York, Sydney, Torento.

9:  Gauch, Jr. H.G. and R.H. Whittaker, 1981. Hierarchical classification of community data. J. Ecol., 69: 537-557.
CrossRef  |  Direct Link  |  

10:  Grant, C.D. and W.A. Loneragan, 2001. The effects of burning on the under story composition of rehabilitee bauxite mines bin Western Australia: Community changes vegetation succession. For. Ecol. Manage., 145: 255-277.
Direct Link  |  

11:  Hedman, C.W., S.L. Grace and S.E. Ling, 2000. Vegetation composition and structure of Southern coastal plain pine forests: An ecological comparison. For. Ecol. Manage., 134: 233-247.
Direct Link  |  

12:  Hill, M.O., 1979. TWINSPAN: A FORTRAN Program for Arranging Multivariate Data in an Ordered Two Way Table by Classification of Individual and Attributes Ecology and Systematics. Cornell University, Ithaca, NY., USA., Pages: 90.

13:  Host, G.E. and K.S. Pregitzer, 1991. Ecological species groups for upland forest ecosystems of northwestern Lower Michigan. For. Ecol. Manage., 43: 87-102.
CrossRef  |  Direct Link  |  

14:  Kashian, D.M., B.V. Barnes and W.S. Walker, 2003. Ecological species groups of landform level ecosystems dominated by jack pine in Northern Lower Michigan. Plant Ecol., 166: 75-91.
Direct Link  |  

15:  Kent, M. and P. Coker, 1996. Vegetation Description and Analysis, A Practical Approaches. John Wiley and Sons, London.

16:  McCune, B. and M. Mefford, 1999. Multivariate Analysis of Ecological Data Version 4.17. M.J.M Software. 1st Edn., Glenden Beach, Oregon, USA., pp: 233.

17:  McNab, W.H., S.A. Browning, S.A. Simon and P.F. Fouts, 1999. An unconventional approach to ecosystem unit classification in western North Carolina, USA. For. Ecol. Manage., 114: 405-420.
CrossRef  |  Direct Link  |  

18:  Mesdaghi, M., 2001. Vegetation Description and Analysis: A Practical Approach. 1st Edn., Jehad Daneshgahi of Mashhad, Mashhad, pp: 161-179.

19:  Mesdaghi, M., 2005. Plant Ecology. Jehad Daneshgahi of Mashhad, Mashhad, pp: 184.

20:  Monier, M. and G. Abed, 2000. Floristic and environmental relations in two extreme desert zones of Western Egypt. Global Ecol. Biogeogr., 9: 499-516.
Direct Link  |  

21:  Picard, N. and A. France, 2003. Are ecological groups of species optimal for forest dynamic modeling? Ecol. Model., 163: 175-186.
Direct Link  |  

22:  Sebald, O., 1964. Okologische artengruuppen fur den wuchsbezirk Obere Nektar. Mitt. Vereins f. forstl. Standortsk. U. Forstpflz, 14: 60-63.

23:  Spies, T.A. and B.V. Barnes, 1985. Ecological species groups of upland Northern hardwood-hemlock forest ecosystems of the Sylvania recreation area Upper Peninsula, Michigan. Can. J. For. Res., 15: 961-972.
CrossRef  |  Direct Link  |  

24:  Verlinden, A. and B. Dayton, 2005. A comparison between indigenous environmental knowledge and a conventional vegetation analysis in North central Namibia. J. Arid. Environ., 62: 143-175.
Direct Link  |  

25:  Waite, S., 2000. Statistical Ecology in Practice: A Guide to Analyzing Environmental and Ecological Field Data. Pearson Education Limited, USA.

26:  White, D.A. and C.S. Hood, 2004. Vegetation patterns and environmental gradients in tropical dry forests of the Northern Yucatan Peninsula. J. Vegetation Sci., 15: 151-160.
Direct Link  |  

27:  Witte, J.P.M., 2002. The descriptive capacity of ecological plant species groups. Plant Ecol., 162: 199-213.
CrossRef  |  Direct Link  |  

28:  Zahedi, G.H., 1998. Relation between ground vegetation and soil characteristics in a mixed hardwood stand. Ph.D. Thesis, University of Ghent Belgium, Academic Press, London.

©  2021 Science Alert. All Rights Reserved