Arid Lands Newsletter (link)No. 54, November/December 2003
Fire Ecology I

Anthropogenic fires in India: A tale of two forests

by Sonali Saha and Ankila Hiremath

"If dry tropical deciduous forests were protected from annual anthropogenic fire throughout Asia we think that their characters would be quite different. We would predict much more diverse forests dominated by seedlings and sprouts of sexually reproducing species adapted for environmental change, not clonal species taking advantage of suppression of most of the flora by human agency..."


(Back to top)
Tropical and subtropical deciduous forests occur between 23° north and south latitudes, are characterized by a 4-7 month dry period, and receive up to 2500 mm of precipitation during 5-8 wet months (Murphy and Lugo 1986a). Deciduous forest plants show a variety of adaptations to rainfall seasonality and dry-season drought. Seed dispersal and germination are cued to rainfall, leaves are shed during the dry season, and percentage of biomass allocated to roots is greater than in less seasonal forests (Holbrook et al.1995; Reich 1995), presumably to maximize uptake of soil moisture.

In most such forests, anthropogenic fires were set to clear land for agriculture or silviculture and promote fodder for livestock grazing; however, systematic fires to facilitate collection of non-timber products are of relatively recent origin (Gadgil and Meher-Homji 1985; Bowman 1998). Vegetation community responses to such practices may confound plants' pre-human adaptations to climate and other natural forces, or may themselves bring about fundamental changes in vegetative composition. The challenge is to distinguish the long-term adaptive signal from mostly anthropogenic noise.

Link to Saha & Hiremath Fig. 1
Link to Fig. 1, ~37K

India is of interest because tropical deciduous forests originally covered 75% of land area there but only a minor fraction remains as secondary forest today. Fires, common in these forests, are widely recognized as primarily anthropogenic (Gadgil and Meher-Homji 1985; Singh et al. 1985). Annual low-intensity surface fires affecting the forest understory characterize the anthropogenic fire regime in India. The fuel-load comprises that year's accumulation of senesced and dry leaves and grassy or herbaceous groundcover; where fires are not present, and in contrast to temperate ecosystem deciduous forests, litter completely decomposes during the Indian monsoons, leaving bare earth (Sundarapandian and Swamy 1999; Saha 2002). Sprouting from damaged shoots (i.e. clonal sprouting) is increasingly prevalent in these forests as clonal species replace non-clonal ones under the influence of recent and contemporary anthropogenic practices (Saha and Howe 2003). Sprouting, a general adaptation to damage from herbivory, drought, windthrow, fire, or other causes (Trabaud 1987), is evidently much more prevalent now than only decades ago.

There are no known primary deciduous forests remaining in India (Puri et al. 1983). Moreover, there are no systematic studies on community-level fire effects in Indian deciduous forest ecosystems. Most existing studies were conducted by British foresters interested in improving timber yield and therefore focusing on timber species, not the overall community. Their conclusions contradict arguments emerging from observational studies of the forests and experimental studies on timber species (Dutt 1907; Rodger 1907; Sengupta 1910; Walker 1910; Osmaston 1935; Thomas 1938). Some foresters supported fire exclusion, considering fires detrimental to tree juveniles and overall forest health; others held that low-intensity fires were beneficial, eliminating “unwanted” understory species. Thus no clear fire management legacy was left to post-colonial Indian foresters, nor did the post-colonial regime develop any unified policy towards anthropogenic fires. This has led to policy based on incomplete information.

Here, we hypothesize that that fires in dry deciduous Indian forests are leading to a loss of tree diversity. Fires promote the growth of clonal species over species that sprout from seed alone, or only from basal shoots. In support of this hypothesis, we review results of two studies that examined effects of fire on regeneration, growth, and diversity of tree juveniles and seedlings in the forest understory of two Indian tropical deciduous forests.

Mendha forest (central India)

(Back to top)
A 1999-2001 study examined the components of the fire regime in the study area and hypothesized that fires both stunt juvenile understory by reducing the stature of tree juveniles and, more important, reduce diversity of forests by favoring clonal-sprouting species (Saha 2000; Saha and Howe 2003).

The study area was in dry deciduous forests of the lower plateau region in the village of Mendha, Gadchirolli District (between 18° 40' and 20° 48' N, 79° 58' and 80°44' E), Maharashtra state. Maharashtra has 54,030 km2 of forested land, with approximately 30% in the Gadchirolli District. Mean annual rainfall is 1600 mm, with a range from 800 mm to 1800 mm; rainfall is concentrated from June to October and sometimes extending to November, with occasional showers in February and April. Temperatures in May and June range as high as 44°C. Topography is gentle to moderately sloping and ranges between 150-300 m above sea level. Soils are moderately deep, well-drained vertisols with high water-holding capacity (Challa et al. 1997).

Link to Saha & Hiremath Fig. 2
Link to Fig. 2, ~54K

The dry forests here are dominated by Terminalia tomentosa (Combretaceae), Pterocarpus marsupium (Fabaceae) and Anogeissus latifolia (Combretaceae). The canopy is 18-20 m high with a middle story dominated by 10-12 m tall trees and understory of herbs and shrubs. The forests were selectively logged until 1980. Presently no commercial logging except bamboo cutting is permitted. The forests are managed by indigenous Gond tribals living in adjacent villages. Ground fires occur annually during fire season, (February to April). Virtually all fires are anthropogenic in the study area, but they are not necessarily set by the people of Mendha. Leaf litter is the main fuel. People burn at this time to facilitate collection of Madhuca longifolia var. latifolia flowers ("Mahua," used for cooking) and Diospyros melanoxylon leaves (an economically important product used for rolling tobacco to make "bidis," a locally produced cigarette); they also credit fire with enhancing the growth and yield of these leaves.

The fire regime in the study area was described by monitoring fuel loads in small plots outside of the main experimental plots, aboveground fire intensity as measured by fire-sensitive tablets that were placed in the experimental plots, and flame height measured near the corners of the experimental plots.

Twenty-four plots of 9m2 each were established; of these, twelve were experimental plots that were burned and twelve were control plots protected from fire. The experimental plots were allowed to burn by locally set fires in March in each year, some experimental plots that escaped fire were burned by us . The control plots were protected from fire by regular removal of leaf litter. Tree juveniles (individuals > 1 year old and up to 1.5 m in height) and seedlings (< 1 year old) were marked and monitored.

In this study, variation in leaf-shedding patterns of tree species led to variation in fuel-load deposition, leading in turn to spatial variation in fire intensity. This affected seedling establishment due to seedlings' susceptibility to fire and consequent heavy mortality. This spatial variation, coupled with the lack of fuel accumulation between seasons, probably makes fire effects in these forests heterogeneous. This heterogeneity is responsible for maintenance of tree species that cannot tolerate fires in their seedling stage.

Link to Saha & Hiremath Fig. 3
Link to Fig. 3, ~7K

Within the burned plots of 9 m2 where fire intensity and juvenile response were monitored, the maximum temperatures monitored by temperature-sensitive tablets varied ten-fold. Juvenile die-back was about 66% overall in burned plots. However, several small juveniles escaped fire, did not undergo stem die-back, and not surprisingly exhibited height and growth patterns similar to unburned juveniles (Saha 2002). Small juveniles up to 10 cm in height were burned even when exposed to less than ten seconds of direct flame in this study. Conversely, this flame height and exposure period were not sufficient to cause burning and die-back in juveniles of 1.5 m. The majority of juvenile die-back occurred among individuals up to 75 cm tall (Figure 3).

Plots allowed to burn in 1999 and 2000 averaged 2.1 +/- 0.5 species of seedlings; those protected from burning averaged 3.3 +/- 0.4 species. Seedling diversity was thus significantly higher (28%) in fire-excluded plots than in burned plots. Fire-induced mortality was examined directly in seedlings of four common species, two clonal and two non-clonal. Seedling mortality in the two clonal species, T. tomentosa and D. melanoxylon, was not significantly different between burned and unburned plots. The non-clonal species, Pterocarpus marsupium (80% in control against 20% in experimental plots) and Dalbergia paniculata (60% in control and 13% in experimental plots) showed significantly greater seedlings survived in the control plots.

In contrast, juveniles and adults of all species showed no mortality in low-intensity fires in 1999 and 2000. All juveniles sprouted successfully after shoot die-back, but clonal-sprouting juveniles produced significantly higher numbers of post-fire stems per plot between 1999 and 2001, while unburned plots experienced reduced numbers of stems per plot: clonal stems increased 13% in burned plots, while dropping 24% in unburned plots. These results further suggest that fire promotes dominance of clonal species.

Biligiri Rangan Hills (South India)

(Back to top)
The Biligiri Rangan Hills (BR Hills) are part of an eastern spur of the Western Ghats, a mountain range that runs along India's west coast and is a global hotspot of biodiversity. The Biligiri Rangaswamy Temple wildlife sanctuary (11-13°N, 77-78°E) is a 540 km2 protected area located on the border between Karnataka and Tamil Nadu states.

The sanctuary ranges in elevation from about 600 to 1800 m above sea level. A north-south ridge running its length divides the region into two distinct rainfall zones: a western zone receiving rain largely from the south-west monsoon, and an eastern zone receiving rain from the north-east monsoon. In both zones, rainfall increases with elevation; thus rainfall in the region as a whole varies from about 900 mm to more than 1600 mm annually. Most rain falls from May to October, tapering off in November and December. January to April is the warm dry season. Soils are generally shallow to moderately shallow, well-drained gravelly clays, with moderate erosion on hills and ridges, and are classified as typic ustropepts (NBSSLUP 1996).

The BR Hills are home to the tribal Soligas, traditionally hunter-gatherers and shifting cultivators. The area was designated a wildlife sanctuary in mid-1970. Hunting and shifting cultivation were banned; the Soligas were settled in hamlets and allocated land to practice settled agriculture. They retained sole rights to non-timber products (e.g., honey, and fruits of Phyllanthus spp. and Terminalia spp.) and continue to graze their livestock in the sanctuary, setting dry-season fires there in large part to encourage fresh fodder for their stock. Evidence documents that this practice has existed for well over a century (Sanderson 1882), and in reality such fires have probably been around for millennia. But it is also likely that fire-return times have decreased significantly as local population densities have increased in the last four or five decades.

The sanctuary encompasses various forest types, from lower-elevation dry scrub forest to higher-elevation dry deciduous forest, lower-montane evergreen forest at mid-to-upper elevations, and a mosaic of cloud forests (also known as shola forests, from the Tamil word “shola” meaning "dense shade") and grasslands at high elevations. The scrub and dry deciduous forests together cover more than 70% of the sanctuary, and in turn are dominated by Anogeissus latifolia and several species of Terminalia. The deciduous forest, with its 15-20 m canopy, is taller than the 10-15 m scrub. Both forests have a mixed grass-shrub understory, sparser in the scrub than in the deciduous forest. The difference between the two forest types could largely be due to abiotic factors, although the species overlap between them has led to some suggestion that the scrub is an anthropogenically degraded form of the original deciduous forest (Uma Shaanker et al. 1998). These two forest types are most prone to annual burning.

A study following the 2001 fire season examined community-level fire effects on both these forest types and their regeneration. Transects measuring 20 x 50 m were established in burned, and neighboring unburned plots in each of four study sites. The transects were further subdivided into 10 subunits, each measuring 10 x 10 m; within each of these, all regenerating vegetation (whether tree, shrub, climber, or forb) was enumerated in a 5 x 5 m quadrat. Thus, a total area of 250 m2 was surveyed per burn-treatment, per site. Regenerating vegetation was further classified as juvenile (> 1 year old), or the current year’s seedlings or resprouts.

At the community level, there was a marked reduction in number of juvenile individuals in deciduous forest (from 7.2 +/- 3.3 to 3 +/- 3.6 individuals/m2), and only a marginal reduction in scrub forest (from 6.1 +/- 2.2 to 4.5 +/- 5.4 individuals/m2). Across life forms, post-fire regeneration was dominated by clonal resprouting; seedling recruitment made a very minor contribution to the total number of regenerating individuals in both forest types. (8% and 18% in dry deciduous and scrub forests, respectively). The number of resprouts in burned deciduous forest (6.8 +/- 1.5 individuals/m2) was almost five-fold the number in unburned deciduous forest (1.4 +/- 1.3 individuals/m2); conversely, resprouts in burned scrub (3.2 +/- 0.8 individuals/m2) were only about three-fold the number in unburned scrub (1.1 +/- 1.0 individuals/m2).

The differential effects of burning in dry deciduous and scrub forest are likely due to differences in fuel load. The understory is sparser in the scrub forest than in the deciduous forest, so scrub fires are likely to burn cooler than forest fires. Although these differences have not been quantified, additional evidence for this contention comes from other work focusing on two local, important non-timber forest product species, Phyllanthus emblica, and its congener, P. indofischeri (Ganesan and Setty, in review). These two species yield similar fruits collectively known as “nelli” or “amla” and not distinguished in trade (Ganesan 2003). However, they have non-overlapping distributions, with P. emblica largely restricted to the deciduous forest, and P. indofischeri growing in the scrub. Size-class distributions of these two species show that P. indofischeri, in the scrub forest, has the inverse J-shaped size-class distribution characteristic of healthy populations--that is, many individuals in the smaller size classes, with the number of individuals tapering off with increasing size. The deciduous P. emblica forest, however, shows an interrupted inverse J-shaped distribution, with a very depauperate population of individuals in the smaller size classes, akin to a population bottleneck. This may result from hotter fires, rather than more frequent fires, in the deciduous forest compared with the scrub forest. If this is the case, then further study is warranted to examine whether hotter fires further increase the tendency of clonal-species domination. The increased dominance of South Indian deciduous forests by clonal species such as Kydia calycina, Lagerstroemia microcarpa, and Anogeissus latifolia (e.g., Sukumar et al. 1998), and a reduction in numbers of regenerating individuals of basal resprouting species (Srivastava 2000) over the last century, provide observational support that this may in fact be the case.


(Back to top)
Admittedly, these studies have limitations, such as the short duration of the Mendha forest study and the observational nature of the BR Hills study; nevertheless, they yield some concrete results, interesting patterns, and more important, questions for further investigation.

These results suggest that fires negatively affect species diversity in tropical deciduous forests in India. Our data show that fires favor species able to resprout clonally over species that cannot sprout, or that produce basal sprouts only. Trends from long-term monitoring of a 50 ha plot in Mudumalai wildlife Sanctuary, in South India (Sukumar et al. 1998), and dominance of clonal sprouts reported in disturbed deciduous forests by Pandey and Shukla (2001), further suggest that the dominance of clonal-sprouting species is the rule, not the exception in Indian deciduous forests.

This clonal domination, and attendant low diversity among repeatedly burned plots, is less obvious in less-disturbed dry lowland tropical forests, indicating that it is not a "natural state" for many or most tropical deciduous forests. Other studies also suggest that forest recovery by basal sprouting typically becomes an important persistence strategy in disturbed tropical ecosystems (Bond and Midgely 2001), for example in tropical deciduous forests of Puerto Rico (Murphy and Lugo 1986b), Mexico (Lott et al. 1987), Madagascar (Sussman and Rakotozafy 1994), and Ghana (Swaine 1992). But nowhere is clonal sprouting and clonal conversion as important a persistence strategy as among annually burned tropical deciduous forests.

Our results suggest that fires retard the growth of tree juveniles and trigger regressive succession of tropical deciduous forests, the first step of which is stunting of the forest understory (Gadgil and Meher-Homji 1985; Puri et al. 1983; Singh et al. 1985; Murphy and Lugo 1986a; Roth 1999). This also occurs with repeated burning in fire-adapted monsoonal savannas of Africa and Australia (Trapnell 1959; Gill et al. 1990). Stunting is also known from other studies on fire in tropical deciduous forests of India (Dutt 1907; Osmaston 1935; Srivastava 2000). Juveniles exist in an arrested stage for a few years before they reach a size that is unaffected by fire.

Comparison of absolute growth rates indicates compensatory response of juveniles to biomass removal because absolute growth rates were significantly higher among burned plots than fire-excluded plots. Even though the absolute growth rates are higher, more important to the community structure is overall fire-caused reduction in plant height in burned compared to non-burned plots. These results suggest that fire promotion and fire exclusion in tropical deciduous forests lead to two different scenarios. Repeatedly burned forests have stunted understories and homogeneous size classes, while unburned forests have juveniles in different size-classes that reflect the intrinsic rates of growth in different species.

If dry tropical deciduous forests were protected from annual anthropogenic fire throughout Asia we think that their characters would be quite different. We would predict much more diverse forests dominated by seedlings and sprouts of sexually reproducing species adapted for environmental change, not clonal species taking advantage of suppression of most of the flora by human agency (see Williams 1975). A worthy goal of restoration would be enhancement of species diversity by regulation or elimination of artificial fires, thereby facilitating the establishment of diverse seedlings and ultimately mature communities. We believe that such a change in land use would prevent ongoing conversion of tropical deciduous forests into low-diversity systems heavily dominated by a few widespread clonal species. Where enough mature trees of non-clonal sexual species still exist to allow total domination of the seedling community, as is the case in our site in central India, a change in land-use involving fire suppression now would, we predict, rapidly improve the fortunes of non-clonal species and the high diversity that they represent.


(Back to top)
Bond, W. J. and J. Midgley. 2001. Ecology of sprouting in woody plants: The persistence niche. Trends in Ecology and Evolution 16: 45-51.

Bowman, D. M. J. S. 1998. Tansley Review No. 101: The impact of Aboriginal landscape burning on the Australian biota. New Phytologist 140: 385-410.

Challa, O., S, Wadivelu, and J. Sehgal. 1997. Bulletin of the soils of Maharashtra. Publication No. 54. Nagpur, India: National Bureau of Soil Survey and Land-Use Planning (NBSSLUP).

Dutt, S. 1907. Fire protection and natural regeneration. Indian Forester 33: 456-460.

Gadgil, M. and V. M. Meher-Homji. 1985. Land use and productive potential of Indian savannas. In Ecology and management of world’s savannas, ed. J.C. Tothill and J.C. Mott, 107-113. Canberra: Australian Academy of Science.

Ganesan, R. 2003. Identification, distribution and conservation of Phyllanthus indofischeri, another source of Indian gooseberry. Current Science 84: 1515-1518

Ganesan, R. and S. Setty. In review. Regeneration of amla (Phyllanthus emblica and P. indofischeri), an important NTFP, in a managed forest in southern India. Conservation and Society.

Gill, A. M., J. R. L. Hoare, and N. P. Cheney. 1990. Fires and their effects in wet-dry tropics. In Fire in tropical biota: Ecosystem processes and global changes, ed. J.G. Goldammer, 159-178. Berlin: Springer-Verlag.

Holbrook, N. M., J. L. Whitebeck, and H. A. Mooney. 1995. Drought responses of neotropical dry forest trees. In Seasonally dry tropical forests, ed. S. H. Bullock, H. A. Mooney, and E. Medina, 243-276. Cambridge: Cambridge University Press.

Lott, E., S. H. Bullock, and J. A. Solis-Magallanes. 1987. Floristic diversity and structure of upland and arroyo forests of coastal Jalisco. Biotropica 19: 228-235.

Murphy, P. G. and A. E. Lugo. 1986a. Ecology of tropical dry forest. Annual Review of Ecology and Sytematics 17: 67-88.

Murphy, P. G. and A. E. Lugo. 1986b. Structure and biomass of a subtropical dry forest in Puerto Rico. Biotropica 18: 89–96.

National Bureau of Soil Survey and Land-Use Planning (NBSSLUP). 1996. Map of Karnataka Soils. Nagpur, India: the Bureau.

Osmaston, F. C. 1935. The effect of burning on rate of growth of medium quality sal coppice. Indian Forester 61: 311-313.

Pandey, S. K., and R. P. Shukla. 2001. Regeneration strategy and plant diversity status in degraded forests. Current Science 81:95-102.

Puri, G. S., R. Puri, V. M. Meher-Homji, and S. Singh. 1983. Forest ecology. New Delhi: Oxford & IBH.

Reich, P.B. 1995. Phenology of tropical forests: Patterns, causes and consequences. Canadian Journal of Botany 73: 164-174.

Rodger, M. 1907. The effects of fire in teak forests. Indian Forester 33: 17-18.

Roth, L. C. 1999. Anthropogenic change in subtropical dry forest during a century of settlement in Jaiqui Picado, Santiago Province, Dominican republic. Journal of Biogeography 26: 739-759.

Saha, S. 2002. Anthropogenic fire regime in a tropical deciduous forest of central India. Current Science 82:101-104.

Saha, S., and H. F. Howe. 2003 Species composition and fire in a dry deciduous forest. Ecology 84: 3118-3123.

Sanderson, G.P. 1882. Thirteen years among the wild beasts of India: Their haunts and habits from personal observation; with an account of the modes of capturing and taming elephants. Repr., New Delhi: Asian Educational Services, 2000.

Sengupta, B. 1910. Fire conservancy in Indian forests. Indian Forester 34:132-145.

Singh, J.S., Y. Hanxi, and P.E. Sajise. 1985. Structural and functional aspects of Indian and Southeast Asian savanna ecosystems. In Ecology and management of world’s savannas, ed. J.C. Tothill and J.C. Mott, 34-51. Canberra: Australian Academy of Science.

Srivastava, R. 2000. Forest fires causing poor stocking of Santalum album and Terminalia chebula in Southern India. International Forest Fire News 22: 28-30.

Sukumar, R., H. S. Suresh, H. S. Dattaraja, and N. V Joshi. 1998. Dynamics of a tropical deciduous forest: population changes (1988 through 1993) in a 50- ha plot at Mudumalai, southern India. In Forest biodiversity, research and modeling: Conceptual background and old world case studies, ed. S. Dallamier and J. A. Comiskey, 495-506. Washington, DC: Smithsonian Research Institution.

Sundarapandian, S. M. and P. S. Swamy. 1999. Litter production and leaf-litter decomposition of selected tree species in tropical forests at Kodayar in the Western Ghats, India. Forest and Ecology Management 123:231-244.

Sussman, R. J., and A. Rakotozafy. 1994. Plant diversity and structural analysis of a tropical dry forest in southwestern Madagascar. Biotropica 26: 241-254.

Swaine, M. D. 1992. Characteristics of dry forest in West Africa and the influence of fire. Journal of Vegetation Science 3: 365-374.

Thomas, A. R. 1938. Effect of slash burning on soil and succeeding vegetation. Indian Forester 64: 362-364.

Trabaud, L. 1987. Fire and survival traits of plants. In The role of fire in ecological systems, ed. L. Trabaud, 65-90. The Hague: SPB Academic Publishing.

Uma Shankar, R., K. S. Murali, R. Uma Shaanker, K. N. Ganeshaiah, and K. S. Bawa. 1998. Extraction of non-timber forest products in the forests of Biligiri Rangan Hills, India. 4. Impact on floristic diversity and population structure in a thorn scrub forest. Economic Botany 52: 302-315.

Walker, H. C. 1910. Fire conservancy in Indian forests. Indian Forester 36: 679-686.

Williams, G. C. 1975. Sex and evolution. Princeton, NJ: Princeton University Press.


bar denoting end of article text

Author information

(Back to top)
Sonali Saha received her Ph.D. from the Dept. of Biological Sciences, University of Illinois at Chicago. Her address for correspondence is:
3079 Biological Laboratories
16 Divinity Avenue,
Cambridge MA 02138
phone: 617-496-3580
fax: 617-496-5854

Dr. Ankila Hiremath is a fellow of the Ashoka Trust for Research in Ecology and Environment, India. Her address for correspondence is:
Ashoka Trust for Research in Ecology and the Environment
659, 5th A Main Rd
Hebbal, Bangalore 560024

Additional web resources

(Back to top)
Ashoka Trust for Research in Ecology and Environment
The Ashoka Trust for Research in Ecology and the Environment (ATREE) was founded in 1996; its aim is to address the environmental challenges facing India. ATREE's efforts are focused on the Western Ghats and the Eastern Himalayas.

About the Arid Lands Newsletter

Link to ALN home page Link to index page for back web issues Link to index page for pre-web issue archive Link to ALN #54 table of contents