Forest Vegetation and Dynamics Studies in India

Forests across the globe have been exploited for resouces, and over the years the demand has increased, and forests are rather exploited instead of sustainable use. Focussed research on vegetation and forerst dynamics is necessary to preserve biodiversity and functioning of forests for sustanence of human life on Earth.This article emphasis that the India has a long history of traditional knowledge on forest and plants, and explorations from 17th century on forests and provided subsequent scientific approach on classification of forests. This also explains the developments of quantitative approach on the understanding of vegetation and forest diversity. Four case studies viz., Mudumalai, Sholayar, Uppangala, Kakachi permanent plots in the forests of Western Ghats has been explained in detail about their sampling methods with a note on the results of forest monitoring. In the case of deciduous forests, the population of plant species showed considerable fluctuations but basal area has been steadily increasing over time, and this is reflecting carbon sequestra-tion. In Sholayar, a total of 25390 individuals of 106 woody species was recorded for < 1 cm diameter at breast height in the first census of the 10 ha plot in the tropical evergreen forest. In Uppangala, 1) a 27- year long investigation revealed that residual impact of logging in the evergreen forests and such forests would take more time to resemble unlogged forests in terms of composition and structure; 2) across a similar temporal scale, the unlogged plots trees < 30 cm gbh showed a more or less similar trend in mortality (an average of 0.8% year-1) and recruitment (1%). The Kakachi plot study revealed that 1) endemic species showed least change in stem density and basal area whereas widely distributed species showed greater change in both; 2) The overall recruitment of trees was 0.86 % per year and mortality 0.56% per year resulting in an annual turnover of 0.71% ; 3) majority of the gap species had high levels of recruitment


Introduction
Forests across the globe have been utilized and many times exploited by humans ever since life style changed from nomadism to settled agri-based system. Forests used to supply many resources including fuel wood, medicine, timber, food etc. But over the years the demand has increased and forests are rather exploited instead of sustainable use.
Historically Indian forests attracted traders for its diversity in spices. The discovery of sea route to India resulted in exodus of European traders. The Portuguese, English, Dutch, French and Germans arrived in India in their quest for the plant species that were important as spices. Records are available that Romans and Arabs were briskly trading with Kings especially along the Western Ghats for variety of spices such as pepper, ginger, cardamom and other condiments.

Vegetation studies in India
India is richly endowed with climatic, edaphic and orographic gradients. Geographical extent of India ranges from tropical latitude to temperate latitude with tropic of Cancer passing through India dividing into almost half with one part predominantly tropical nature while the other is subtropical and temperate in nature. Hence the natural vegetation of India is also ranging from rainforests in the Western Ghats and Eastern Himalayas to desert in Rajasthan.
Ancient description of vegetation largely confined to composition of species primarily of medicinal use and utilization in rituals. Ancient India had a rich tradition of life sciences. There are reviews relating to the knowledge of plant sciences in ancient India [1]. The study of plants in ancent India was mainly under two heads namely, 1. Plants utilized for medicinal purposes and 2. Plants relating to agriculture [1]. However, the study on description of plants and animals was not popular. Takshashila University encouraged the collection, identifying and description of plants found around Takshashila. The traditional medicine in India which is primarily plant based has description of several plant species occurring in different forest types. Though, the descriptions largely confined to medicinal properties and parts of the plants used.

Studies on plant systematics
Significant contribution to botany of India was made during the colonial period with a strong pursuit of harvesting botanical resources. There were several European botanists and explorers contributed to the knowledge of botanical resources of India. Van Rheede H. A. (17th Century) the then Dutch Governor of Cochin made an effort to scientifically document the wealth of plants and indigenous medical knowledge with the native Malabaris. He produced a well written book, 'Hortus Malabaricus' published during 1678-1693 [2]. This book contained scientific description and life size illustrations of about 742 native plants in Malabar (Kerala) and of which about 650 are of medicinal value. This book was used by several European botanists such as Linnaeus, De Candole, De Jussieu, Adanson, Blume and Wight to describe many species from India and Asia [3]. Some of the important names include, Johann Gerhard Koenig (1768) who made extensive collections from India, Robert Kydd (1787) who was instrumental in starting the Royal Botaical Gardens at Calcutta and Sir Joseph Hooker who wrote monumental book "Flora of British India" and Gamble and Fischer (1915-36) who wrote a comprehensive flora for the region of the presidency of Madras [4]. There were several Europeans who made collections and described many plant species from different parts of India. The development of science under colonial period is well described by Kochhar [4,5]. Important component in development of plant science in India under colonial rule was establishment of botanical gardens not only in Calcutta but in several places across India with major focus on breeding and maintenance of economically important species from different colonial parts of southeast Asia.

Indian forest types
First systematic classification of Indian forest types was by Sir. H.G. Champion in 1936 which was later revised in the year 1968 by Champion H.G. and S.K. Seth [6]. They identified 16 major forest types based on rainfall and temperature ( Table 1). Contemporaneously, the French Institute of Pondicherry (IFP; http://www.ifpindia. org) produced vegetation maps at the one million scale for peninsular India (by publishing 12 sheets between 1959 and 1973) in collaboration with Indian Council of Agricultural Research. Subsequently published six vegetation maps (scale 1: 250000) covering south and central Western Ghats region in collaboration with state forest departments of Karnataka, Kerala and Tamil Nadu [7][8][9][10][11][12]. These maps were produced considering bioclimate of the region, floristic series (dominant species based on climax species, structural and floristic composition) of the forest types, limits of forest types, altitudinal zonation, degree of degradation of forests, relationships between different stages of succession and the potentiality of a disturbed forest to return to the climax. Since then there are studies to improve the classification of forest types considering the feasibility for the forest managers to manage their forests. Attempts have been made by the Forest Survey of India (FSI) to revisit different forest types and reassign the forest types based on ground survey [13] ( Table 2). In 2014, Indian Council of Forestry Research and Education (ICFRE) has revisited Champion and Seth [6] forest type classification by rapid assessment mode.

Major developments in understanding vegetation of India
There was a discernible change in describing vegetation in India. The trend was not to describe species found in any vegetation type but quantitatively describe the vegetation of a locality. There were several studies on quantitative description of vegetation in India. It was initiated by Rai [16], who inventoried all trees ≥10 cm dbh in four plots of 2.7, 2.7, 2.63 and 1.09 hectares respectively at Devimane, Malemane, Kodkani and Katlekan areas of the Western Ghats. Most such studies in the Indian evergreen forests have been conducted during the last decade of the 20th century, and many of them are once census plots. Interest in tree mortality and forest dynamics has increased recently because forest dynamics is thought to be involved in determining tree species diversity [17], and also thought to be related to global climate change, in particular [18]. Phillips and Gentry (l.c.) concluded that tree turnover rates have increased in tropical forests during the latter part of the 20th century. This has proved to be a controversial finding (e.g. [19,20]

Forest dynamics study by Indian Institute of Science (IISc), Bengaluru, Karnataka, India
India's first biosphere reserve the Nilgiri Biosphere Reserve (NBR) was established in 1986 and the responsibility of conducting research was given to Center for Ecological Sciences (CES), Indian Institute of Science (IISc). As a principle institute responsible in setting NBR, IISc has a commitment towards ecological research in the biosphere. The climatic and altitudinal gradient in NBR harbours different vegetation types ranging from dry thorn forests to rainforests. The altitudinal range has dry forests in the lower elevation to high altitude montane forests. Hence there is a tremendous variation in species composition across both climatic and altitudinal gradient.
When IISc began its research in NBR there were several issues regarding the choice of study area. Based on both logistical and academic reasons, IISc decided to join the international network of 50-ha forest dynamics plots promoted by Prof. Hubbell [21]. CES selected species poor deciduous forests of Mudumalai for variety of reasons. Firstly, Mudumalai would complement plot at Barro Colorado Island (BCI), Panama (tropical semi-evergreen forest, neotropics) and Malaysian plot,  [14,15].
FRIM, Malaysia (equatorial rainforest). Secondly the factors influencing the dynamics in dry forests are totally different from factors influencing dynamics of forests at both Panama and Malaysia [22].

Choice of the site
IISc has selected 50 hectare area in 17th compartment of the Kargudi range in Mudumalai Tiger Reserve as because, 1. The area is relatively free of anthropogenic disturbances as settlements are far off, 2. This area was selectively felled during late 1960s and we could identify the trees that were removed from the plot. They could identify the species of the stumps left behind and map them spatially in the plot, 3. This area lies in the transition zone between dry and moist deciduous vegetation and has both elements represented in the plot.

Methods
Establishment of plot involved two steps a. gridding and b. enumeration of the plot. Gridding involves dividing the plot into blocks of 20 X 20 metres after correcting for slope. Correction for slope is important to give equal opportunity for all individuals to compete for resources.
Enumeration involves measuring of all woody individuals and mapping them. Block of 20 X 20 meters is further divided into blocks of 10 X 10 meters temporarily by laying ropes. All woody individuals >1 cm dbh (diameter at breast height) are identified, measured for size, marked with unique number and mapped by taking X and Y coordinates. Point of measurement (POM) was marked where size measurement was made.

Results
There were 25,929 stems duing the first census belonging to 71 species. Most abundant species was understorey tree Kydia calycina (Malvaceae) with 5175 individuals accounting for 20% of total abundance. Dominant canopy species such as Anogeissus latifolia (2280), Lagerstroemia microcarpa (3980). Tectona grandis (2143) and Terminalia crenulata (2776) accounted for 55.9% of total abundance. The genus Ficus (Moraceae) had five species and followed by Terminalia (Combretaceae) with three species. Relative abundance cumulative abundance of top ten species is listed in the Table 3. Top ten species accounted for 87.52% of the total abundance. There were 21 species with less than 10 individuals in the plot and 7 species had one individual in the entire 50 hectare plot ( Table 3).

Forest dynamics (1988-2016)
Population of woody species in the plot has shown considerable fluctuations over different census periods. Population has grown from 25,935 individuals >1 cm dbh in 1988 to 48,360 individuals in 2016 (Figure 1)

Patterns in mortality and recruitment
Entire plot (50 hectares) was annually censused for mortality and recruitment till 2008. Since 2009, annual census was done in sample plots of 40 meters X 40 meters inside the plot. There were 100 such randomly placed plots accounting for little more than 1/3rd of the total area. The reports on annual mortality for the plot from sample plots from 1989 to 2016 were published [23].
Mean mortality rates due to different causes across census periods is tabulated in the Table 6. Mean rate of mortality due to fire during the census period between (2004-2008) and (2012-2016) was zero suggesting fire did not result in the mortality of any individual. There was a considerable variability in mortality rates across other census periods ( Table 6). Elephant related mortality rate was high during the first census period (1989-1992) and 5th census period (2004)(2005)(2006)(2007)(2008) owing to abundance of elephant favoured species such as Kydia calycina and Helicteres isora. The mortality rate inflicted by other causes also showed considerable fluctuation with high variability during the 7th census period (2012-2016).

Basal area changes and biomass across census periods
The basal area in the plot has been steadily increasing over time (Figure 2). Above Ground Biomass (AGB) and hence carbon stock also shows a similar trend (Figure 3), with both the native woody vegetation and invasive ground vegetation showing increment. Basal area changes do not necessarily translate to AGB changes: for instance, the slight decline in basal area during 1992-1996 is not reflected in AGB, which may be partly due to differences in wood densities (e.g. hardwoods growing more than softwoods). Native woody vegetation biomass in 2004 shows a slight reduction owing to a severe drought in the preceding years and a large fire in 2002. However, the invasive Lantana camara L. increased substantially following the drought, and therefore the total biomass remained at the 2000 census level. Large increment in biomass were seen in all subsequent censuses: 2008, which followed a period of higher than average precipitation and no fires, 2012, despite a fire in 2010 and only 807 mm precipitation in 2012 (compared to the long-term average of 1260 mm), and 2016, which follows a fire-free census interval. The estimates for basal area and biomass for the census period 2016 to 2020 is based is based on first 10 Ha.

Long-term monitoring programme of Kerala Forest Research Institute (KFRI), Peechi, Thrissur, Kerala
Being a part of Western Ghats range of mountains, one of the global biodiversity hotspots, Kerala has bestowed with diverse forest ecosystem with high degree of endemism. Kerala Forest Research Institute (KFRI) currently having more than 40 permanent plots across the state representing all major forest ecosystems (Figure 4) and more plots are coming up as a part of various ongoing research projects. KFRI Long-term monitoring programme represents all major ecosystems like mangrove, moist deciduous, dry deciduous, wet evergreen, montane shola forests and grasslands. As of now, the programme covers 50,309 woody individuals of more than 350 species. Majority of our plots are smaller in size (≤1 ha) in which survey would   summarized as Table 7. Complete inventory of woody individuals ≥1 cm dbh were done and each individual was permanently tagged with sequentially numbered aluminium tags. In the 10-ha plot, a total of 25,390 individuals of 106 woody species  were recorded [25]. These individuals were belonging to 44 families and 81 genera.

A case study: Uppangala Permanent sampling plots
Since the early 1980s, the French Institute of Pondicherry has been in collaboration with the Forest Department of Kerala and Karnataka to explore structure and diversity of wet evergreen forests of the Western Ghats. In 1979-80, a total of 147 trees ≥ 30 cm girth at breast height was monitored untill 1982 for growth (with a precision of 0.02 mm) at monthly intervals in a 0.2 ha plot at Attapadi. Monitoring the plot in the region was stopped for logistic reasons. Subsequently, IFP has established two sets of sample plots in low elevation wet evergreen forest in Kadamakkal RF, Sampaje Range, Kodagu (ca. 12°32 0 15″N, 12°33' N, 75°39'4″E; Figure 1a). Currently this area comes under the Pushpagiri Wildlife Sanctuary in Kodagu district. The study area, the Uppangala was subjected to selective logging, between 1974 and 1983 [26]. During the logging operation, the forest was divided into compartments of 28 ha each, 237 to 359 large trees (stems ≥ 180 cm) of medium wood (> 0.5 but ≤ 0.72 g cm À3 ) Dipterocarp species viz., Dipterocarpus indicus and Vateria indica were logged per compartment. An average of 8 to 13 dipterocarp trees ha À1 were logged manually and hauled using elephants locally, a method that causes much less damage than mechanized skidding. A few patches of forest remain unlogged. The elevation ranges between 400 and 600 m a.s.l. It belongs to the Dipterocarpus indicus-Kingiodendron pinnatum-Humboldtia brunonis type of wet evergreen forests and is a part of the West Coast Tropical Forests of Champion and Seth's classification. Uppangala receives slightly more than 5100 mm per year and the dry season lasts 4.5 months.
The first set of sample plots was installed in 1984 to study the post-logging effects on the forest dynamics of a once logged 30-ha compartment (Figure 5b). It consists of 14 plots of 600 m 2 (20 x 30 m). All trees ≥ 10 cm girth at breast height (gbh) were recorded during the first census. All the plots were recensused (except 4 plots, which were recorded as burnt) in 1988 and 1993 for recruitment and mortality. In 1989, a second set of sample plots was established in another 30-ha compartment (Figure 5c), which had escaped logging operation due to the ban on selective felling from 1987 in the forest of Western Ghats.
The unlogged compartment probably represents the last example of old-growth low-elevation Dipterocarp forest in the entire Western Ghats. Five north-south oriented transects (viz., A, B, C, D and E; Figure 5c)  long and 100 m apart from each other were established to inventory trees ≥30 cm gbh. Collectively they represent a 3.12 ha À1 systematic sample. Subsequently, additional rectangular sampling plots viz., H, R and S, which overlap with sampling area of the transects and represent an additional area of 1.95 ha, were established between 1990 and 1993 to study the forest dynamics according to topography (slope and more or less flat terrain). Totally 3870 trees were identified, mapped and installed with dendrometric belt (precision of 0.2 mm) for growth monitoring, which has no equivalent in any other tropical forest in the world. The sampling area has been monitored annually for recruitment and mortality. In 2010-2011, fifteen 1 ha plots was established to appraise allometric relationship of tree diameter and tree crown (for trees ≥ 30 cm gbh) and to estimate above ground biomass. These 1 ha plots were sampled randomly in both the logged and unlogged forests. Of these, four plots were selected to understand the residual impact of logging on species composition, population structure and biomass. In 2013-2014, the sampling area of the unlogged compartment has been increased to 9.9 ha and all trees ≥ 30 cm gbh were inventoried within a 330 x 300 m 2 area (Figure 5d). All the trees were identified, girth measured and mapped.

Tree density and diversity
The systematic sampling plots of logged compartment was recorded with a total of 2748 trees ≥ 10 cm girth at breast height (gbh) during the first census in 1985 [27]. Similarly, a total of 1981 trees ≥ 30 cm gbh were recorded with 91 species in the first census of 3.14 ha area of the unlogged compartment in 1990 [28]. Pronounced species hierarchy is another characteristic feature of the forest. Just 10 most abundant species contributed 71% for the forest stand ( Table 9). Subsequent additional sampling area allow us to monitor more number of trees in the unlogged compartment. Totally 3870 trees were enumerated in the 5.07 ha during 1994 census and all those trees fitted with stainless dendrometric belts for measuring growth with a precision of 0.2 mm. At present we are monitoring 6672 trees (of which 3127 trees were installed with dendrometer bands) representing 111 species in the unlogged forest plot. The forest is characterized by high tree density and basal area (661 stems ha À1 ; 43 m 2 ha À1 ). Pronounced species hierarchy is another characteristic feature of the forest. Just four species namely, Dipterocarpus indicus (emergent layer), Vateria indica (upper canopy and emergent), Myristica dactyloides (intermedidate) and Humboldtia brunonis (understorey) dominate the forest stand, and they collectively account for greater than 50% of density and basal area of the forest.

Impact of logging on tree diversity
A decade long monitoring of logged and unlogged forest for trees ≥ 30 cm gbh revealed the logged compartment had 347 trees and 54 species in 0.6 ha (1986) whereas the unlogged compartment had 1891 trees and 88 species in 3.12 ha in 1990 [29]. Initial stand density and basal area of the trees were slightly lower in the logged forest (578 stems ha À1 ; 34.8 m 2 ha À1 ) than in the unlogged forest (606 stems ha À1 ; 39.3 m 2 ha À1 ). Mean density and basal area for the 20 Â 30 m 2 samples of the two compartments displayed no significant difference (t-tests, P>0.25). The mortality rate was more or less similar for the compartments (0.89% for logged and 0.87% for unlogged), which is lower than the rates observed in other tropical forests. Annual recruitment rate of logged (1.68%) and unlogged forests (1.34%) were not significantly different. Mean diameter increment was 2.1 mm and 2.9 mm yr. À1 for unlogged and logged compartments. In the logged forest, Antiaris toxicaria, Aphanamixis polystachya, Beilschmiedia  Garcinia morella (Gaertn.) Desr. Clusiaceae 44 2 71 Table 9.
Top-ten dominant species in the unlogged compartment plot during the first census 1990.
guttata and Vitex altissima. In the same time, unlogged plots showed no disappearance of species and appearance of three new species namely, Agrostistachys borneansis, Clerodendrum viscosum and Syzygium hemisphericum. This decade long investigation suggested that the logged compartment gradually recovered and resemble unlogged forest within 20 years. However the recent inventories of the logged compartment at 1 ha scale shows the residual impact of logging even after 27 years (Table 10; [30]). Logged plots had low floristic similarity between them (0.45 to 0.56%) and also with the unlogged plots (0.41 to 0.43%). Mantel and partial Mantel tests proved that logging was the main driver for the species composition rather than the elevation and spatial distance. Higher abundance of species belonging to canopy, intermediate and light wood categories and lower density of emergent, understory and medium wood types were recorded in the logged plots. As compared to unlogged plots, logged plots had 20-59% less above ground biomass (AGB) due to paucity of large trees, especially in the emergent and medium wood types. However, the logged plots had higher AGB in canopy and hardwood categories. These findings indicated that the compositional shifts has occurred in the logged patches and the recovery process may depend on the resurgence of emergent and medium wood categories (Figure 6).

Forest dynamics in unlogged forest
In the unlogged forest, over the study period of 1990-2016, mortality rates ranged from 0.7 to 1.2% yr. À1 with an average of 0.8% yr. À1 while the recruitment ranged from 0.4 to 1.2% yr. À1 with an average of 1% yr. À1 (Figure 7). The basal area of the stand showed a loss of 13.8% due to tree death and an addition of 21.6% basal area by growth of trees. Overall, it shows an increment by 7.8% of the stand basal area. During the period of 26 years, four species Memecylon wightii Thwaites, Goniothalamus cardiopetalus (Dalz.) J. Hk. & Thoms., Clerodendrum viscosum Vent. and Walsura trifolia (A. Juss.) Harms. disappeared by tree deaths and one species Diospyros assimilis Bedd. appeared by new recruits. A total of 73 species have registered either recruitment and/or mortality, while population density of the remaining 18 species was unchanged during 26 year period. Of these, 44 species showed decline in population density. Notable among them includes Myristica dactyloides, Humboldtia brunonis, Palaquium ellipticum and Knema attenuata, lost more than 10 individuals during the study period. Nineteen species showed increase in population density. They include Kingiodendron pinnatum, Holigarna nigra, Diospyros bourdillonii and Leptonychia caudata each was recorded with increase in population density of 5 individuals.
Where D is the diameter at breast height in cm, H is total height in m and ρ is wood density in g cm À3 [32]. The estimated value ranged from 268 to 491 Mg ha À1 for those plots in the logged compartments and 611 to 649 Mg ha À1 for the unlogged compartment Table 10. The AGB ha À1 of unlogged plots of the present study is high compared to the available data on Indian forests and the other tropical forests across  Summarized results of a decade long monitoring study of logged and unlogged plots [31].
the continents: a mean of 287.8 AE 105.0 Mg ha À1 for South America, 393 AE 109.3 Mg ha À1 in Asia and 418 AE 91.8 Mg ha À1 in Africa for trees AE 10 cm dbh [33].
In summary, the continued monitoring of the plot will enhance our capacities to understand the forest dynamics in space and time, and response of the forest to the influence of climate change.

Forest dynamics plots in Kakachi forest, Kalakad Mundanthurai Tiger Reserve, Western Ghats
Small scale forest dynamic plots were established in the wet evergreen forest at Kakachi in Kalakad Mundanthurai Tiger Reserve (hereafter KMTR) (8 o 33' N. Lat. 77 o 23'E. Long, Figure 8). It covers an area of 887 km 2 along the eastern slopes of Agasthyamalai range. The altitude ranges from 100 to 1890 m with generally steep slopes and deep valleys. KMTR supports large stretches of evergreen forests, which are contiguous to the rest of the WG, and endowed with large number of endemic and rare plant species, and provide habitats for rare animals such as Lion Tailed Macaque, Nilgiri langur, Tiger, Elephants etc. KMTR receives both South-West and North-East monsoons and being a major watershed, seven major rivers originates from the forest. These rivers meet the water requirements of the arid regions of south Tamil Nadu. Kakachi is located at 1300 m amsl and receives an annual rainfall of over 3500 mm. The rainfall is spread over 8 to 10 months in a year. The spread out of the rainfall in the study site is due to Southwest monsoon and Northeast monsoon rains. Mean maximum temperature is 24 o C and minimum about 16 o C [34]. The terrain is highly undulating and is traversed by numerous mountain streams. The vegetation is characterized by three dominant tree species, Cullenia exarillata, Palaquium ellipticum and Aglaia bourdillonii [34]. Between 1993 and 1994 three 1 ha forest dynamic plots were established in an undisturbed wet evergreen forests at Kakachi, Kalakad Mundanthurai Tiger Reserve (KMTR) of Agasthyamalai range.
The principal objective of this study was to measure the changes in diversity, structure, recruitment and mortality of tree species compared to other forests within WG as well as globally. Following are the specific objectives: (1) Determine the diversity and population structure of trees at Kakachi (2) Compare diversity and density of endemic with the widely distributed species in the site (3) Determine the overall recruitment, mortality and turnover rates of tree species.
Three 1-ha plots of 250 m x 40 m dimension were established within the wet evergreen forest during 1993 and 1994. A minimum distance of 1 km, spatially separated these plots. The plots were permanently marked using PVC pipes and all trees above 10 cm dbh at 1.3 m above ground were enumerated and tagged.

Floristics and species diversity
A total of 68 tree species >10 cm dbh were recorded from the 3 ha. The sixtyeight tree species belonged to 52 genus and 31 families. The most species rich family was Lauraceae with 12 species followed by Euphorbiaceae (7 sp.) and Myrtaceae (6 sp.). Seventeen families had only one species. Syzygium was the most common genus with 6 species followed by Litsea with 4 species. Genus with single species represented over 90% of the total genus. Shannon diversity index was 2.79 (=4.02 log 2 ) and the evenness index was 0.66. The number of species recorded per ha was 46.

Stem density
A total of 2116 live stems were encountered in the 3 ha at an average of 705 stems ha À1 . Three species Agrostistachys borneensis (19%), Cullenia exarillata (16%) and Palaquium ellipticum (13%) represented 45% of the stems, while the other 65 species accounted for the remaining 55%. The species abundance relationship shows that majority of the species had a density between 4 and 8 individuals per 3 ha. Seventy percent of the species had less than 10 individuals in the 3 ha plot and only 10% (7 species) had over 100 individuals. Over 3.5% of the stems were dead during the first enumeration.

Basal area
Total basal area of all the trees was 60.51 m 2 ha À1 . Two dominant species Cullenia exarillata and Palaquium ellipticum accounted for over 58% of the total basal area, while all other species individually accounted for less than 5%. Mean basal area of individual trees was 0.0769 m 2 ha À1 and ranged from 0.026 m 2 ha À1 to 0.3233 m 2 ha À1 .

Life forms
Of the 68 tree species 42 were canopy trees and 23 (35%) were understorey trees. Maximum height of canopy trees was 40 m. The tallest species is Cullenia exarillata with a mean height of 22 m. Canopy trees were at a higher density than understorey trees. Many of the common canopy trees were dominant component of the stand. Canopy species such as Cullenia exarillata, Aglaia bourdillonii and Palaquium ellipticum accounted for 66% (808) of the canopy trees. Among the understorey tree species Agrostistachyis borneensis, Gomphandra coriacea, and Epiprinus mallotiformis accounted for over 80% (663) of the total stems. Canopy trees were also larger in girth besides being more abundant, therefore contributed to higher total basal area 49.04 m 2 ha À1 compared to understorey trees 8.4 m 2 ha À1 .

Habitat
In terms of habit preferences there were 36 closed forest species and 33 gap species. Nineteen of the closed forest trees were canopy species and remaining 17 were gap invaders. Similarly in the understorey 14 were closed canopy species and 15 were gap species. The closed forest species were 11 times (1840) more abundant than gap species (144) and majority of the gap species contributed to less than 7 individuals per ha. Basal area of closed forest species was 22 times greater than gap species.

Species
Thirty-three of the total 68 identified tree species (49%) in the plots were endemic to the Western Ghats. The endemic species richness increases from localised endemics to more widely distributed endemic species. Greater proportion (76%) of the species were endemic to the entire Western Ghats (EWGE, Entire Western Ghats Endemic), 18% to southern Western Ghats, (SWGE Southern Western Ghats Endemic comprising of Nilgiris and south of the Palghat gap) and 6% to Agasthyamalai (AGME Agasthyamalai Endemic) region alone (localized endemic). Some of the common endemic species are Palaquium ellipticum-endemic to whole of Western Ghats, Litsea keralana -restricted to southern Western Ghats and Aglaia bourdillonii -endemic to Agasthyamalai.

Density
Endemic species of the Western Ghats accounted for 51% (1079) of the total stems encountered in the 3 ha. EWGE were the most numerous, and accounted for 83% (893) of the stems followed by 16% (172) for AGME species and only 1.3% (14) to SWGE. The density of trees under the 3 endemic gradients is significantly different (KW = 9.84, p < 0.01). The WGE species were significantly more abundant than SWG species (Dunn's test p < 0.01). The median density value was 8 for EWGE species and 2 for SWGE species. Localized AGME species such as Aglaia bourdillonii was at high density. Contrary to species richness, density is high for local endemic species and highest for EWGE species but very low for SWGE species.

Basal area
Basal area of endemic species accounted for 94% of the total basal area, of which 95% were EWGE, 0.6% SWGE endemic and 5% AGME. Though there were only two species endemic to Agasthyamalai, one of them Aglaia bourdilonii was a highdensity species but accounted for only 3.3% of the basal area. Trend in basal area was also similar to density; EWGE highest followed by AGME and finally SWGE.

Changes in recruitment, mortality and turnover of tree species over time
Endemic species showed least change in stem density and basal area whereas widely distributed species showed greater change in both. The overall recruitment of trees was 0.86% per year and mortality 0.56% per year resulting in an annual turnover of 0.71% (Table 11). Thirty-three species did not show any recruitment and mortality. Forty species showed no recruitment and 37 species no mortality. The dominant species such as Cullenia exarillata, Palaquium ellipticum, Agrostistachys borneensis and Aglaia bourdillonii had low recruitment and mortality rate.
Majority of the gap species had high levels of recruitment and mortality resulting in a high turnover. Some closed forest and canopy species such as Nageia wallichiana (Podocarpus), Elaeocarpus tuberculatus and understorey species such as Antidesma menasu, Syzygium mundagam and Miliusa wightiana showed high levels of recruitment. Gap species had higher mortality and recruitment than closed forest species. Recruitment and mortality was not significantly different between canopy and understorey species. In general gap species was the major contributor to the turnover in the forest.

Way forward
Long-term data is essential for undestanding vegetation dynamics. Vegetation dynamics is directly related to climate variability that an ecosystem experience. Extreme events such as floods, drought and snowfall forms part of long-term variability in climate. Vegetation response to such extreme events depends upon the type and intensity of an event. Government of India has initiated two major national projects to undestand and combat the impacts of varibility in climate through understanding natural vegetation dynamics.  kind in India to address climate change issues at national and international level and helps to trace footprints of climate change impacts through vegetation and also reveals to what extent our forests are resilient to change in the climate. Further it will also address the issues flagged by UNFCCC, IPCC, NAPCC, SAPCC etc. Major objectives of the programme includes establishment of Permanent Preservation Plots (PPP) to observe and understand the changes in species diversity, composition and growth pattern due to climate change over a period of time. The methodology for selection and laying of sample plots, assessment, identification and tagging of plants is based on Centre for Tropical Forest Science (CTFS) protocol. Aimed at precision, uniformity, and large scale of international acceptance, it was decided to laydown country wide permanent plots (preferably 10 Hectare size) in major forest types of the country wherein woody individuals >1 cm diamter at breast height (DBH) would be monitored for vital parameters such as recruitment, mortality and growth in relation to climate. The study also includes dendrochronology, edaphic factors, survivality, regeneration, invasive species, dynamics of soil microflora, phenological studies, insect-pest incidence, disease infection, pollinators etc in the pemanent plot and surrounding forest area.
Complementary to this initiative Indian Government launched a new pan India research program-Indian Long Term Ecological Observatories (ILTEO) with a larger goal of assessing the influence of climate change on the biodiversity at national level. To address issues related to climate change the Government of India has set up Indian Network for Climate Change Assesment (INCCA) to provide frame work to monitor impacts of climate change, assess the drivers of climate change and to develop decision support system. It is been recognzed that climate change of one of major drivers, Long-term ecological monitoring is required to identify pattern and drivers of change. Moreover long term monitoring is required to frame the national policies and signing international conventions such as United Nations Framework Convention on Climate Change (UNFCC). There are several isolated programs monitoring the changes. However, there is a need for unified multidisplinery national level program to address the issues of climate change. All India Coordinated Research project under ICFRE is one such national level effort to address encouragement and research to climate change.