Asia Center go to the web site of The Academy of Natural Sciences

Asia Center of The Academy of Natural Sciences

Tree Growth and Fire Study on Hovsgol Forest

Baymbasuren Oyunsanaa, Clyde E. Goulden, Nachin Baatarbileg and Batchuluun Gantsetseg

The growth dynamics of forest ecosystems is affected by a complex of environmental factors which include climate, insect infestations, fire, competition among trees, soil characteristics, and others. Trees respond to these influences with corresponding changes in their annual growth rings. Therefore, a record of past influences on growth is retained in tree-rings and can be unraveled through dendrochronological techniques (Fritts and Swetnam, 1989). At the end of the twentieth century, fire ecology has enjoyed a new wave of interest, for several reasons. Fire situation fire is a major agent that initiates and terminates vegetation succession; controls age structure, species composition and physiognomy of the vegetation. This produces the vegetation mosaic on the landscape; and influences nutrient cycles, energy flows, productivity, diversity, and stability throughout the ecosystem (Swetnam, 1993).

Tree-ring study can give long-term growth and historical fire regimes of forest trees. We hypothesized that different growth rates of tree are depends as a result of natural or human caused disasters, such as fire events. In this study we determined growth rate and fire events in the six valleys of eastern shore of Lake Hövsgöl.

Study Area

The study area is located in the eastern shore of Lake Hövsgöl and six valleys (Turag, Shagnuul, Noyon, Sevsuul, Dalbay and Borsog) of the Lake Hövsgöl were studied and monitored. It is about 90 km from Khatgal soum, Hövsgöl province of Mongolia.

Field sampling

A mixed-aged Siberian larch (Larix sibirica) forest was studied to examine long-term growth patterns, and size distribution, fire history and seasonality. In the Borsog valley, a transect line from south facing slopes to north facing slopes and other five valleys only in the north facing slopes was set up. The transect width was 100m by 1200m in length. Every 20m were flagged on two sides of transect dividing into 20x50m quadrates. A random number table was used to select quadrates. Two strata per slope were considered and divided by density of trees, plant community, soil moisture, and altitude (Jayaraman, 2000).

Within the quadrates sample plots of 5.6m circular plots were set up. Wooden posts were set up at each center of plot and coordination recorded. Permanent marker on a small area of polished bark numbered all trees. In all stands, diameter at breast height (DBH; measured at 1.3m) of trees (stems3;10cm DBH), height, and spatial distribution of trees within plots were recorded. Totally 84 sample plots were set up. Tree growth condition was recorded for each tree with the category dead, diseased, defective and normal growth.

Increment cores and disks were collected from larch trees for the growth and fire study. The cores (2 cores per tree) were extracted at breast height (1.3 m) or 50 cm with an increment borer. A total of 360 core samples and 96 disk samples from 84 plots were collected.

Ring-width chronologies

The core and disk samples were prepared and surfaced to enhance ring boundaries before measurement. Cores were mounted on grooved wooden boards and sanded by hand with a series of sandpaper grits up to 1000. Ring width measurement was carried out using Velmex measuring system (Velmex, Inc). The precision of measurement was 0.01 mm.

The cross dating was aided by the presence distinctive narrow rings and the quality of cross-dating was examined by the program COFECHA (Holmes, 1983). The age-related growth trend within each ring-sequence was removed using the program ARSTAN (Grissino-Mayer et all. 1993) in which the detrending curve selected was a negative exponential curve, horizontal line, or a straight line with negative slope. Deviations from this curve were standardized to produce a set of annual growth indices for each sample. Ring width chronology for each series was derived by averaging the growth indices for each year among different trees (Fritts, 1976). Also, for the statistical analysis of ring-growth, the program TSAP (Time Series Analysis Program) was used.

The ring width chronologies of Siberian larch are describing the seven chronologies were developed. The Turag Valley forest had longest chronology (over 440 years) due to its longevity and resistance to fires as a result of its thick bark or tolerance. Some of Valley trees are susceptible to pathogenic and decay fungi infection; therefore, it had the shortest chronology length. There are several factors should be considered for the fire resistance of trees; fire distribution, size and intensity. The several decades of growth reduction is occurred in the most of study valleys forest during the 1720s to 1770s, 1860s to 1890s and 1940s to 1970s suggested a state of increased tree to tree competition as recruitment increased overall tree density in the area sampled. Last 15 to 30 years growth indices had increased except Noyon Valley.

Fire dynamics

Reconstruction of the fire history based on tree-ring data provide valuable information on reference conditions of fire regimes prior to widespread Mongolian forests and its associated disturbances. In this study FHX2 is DOS-based software were used and that facilitates the analysis of past fire history from fire scars and other fire-related injuries found in the annual growth rings of trees. FHX2 provides basic summary information for each sample and performs a number of statistical tests and functions to help evaluate the fire history.

Long-term and short seasonal frost and permafrost sites are characteristic features mountain forest sites in Northern Mongolia. This zone is characterized by continental cold and dry climate. Moisture supply for the vegetation is secured by rainfall and additionally by the moisture that is set free by melting seasonal frost that will supply the needed moisture for the root system of plants and woods. Occurrence and impacts fires entering forest sites with seasonal frost depend on the season. During the period June to August the melting frost supplies sufficient moisture to the upper soil, the organic and litter layers and the understory vegetation, thus reducing the spread and intensity of fires. Under the cold spring and autumn conditions the surface fuels are dry and support the spread of fires and the development of fires with intensities that are higher than the summer fires thus more damaging.

Extremely large fires occurred in some years. In 1996-1997 13.4% of the total forest fund area had been affected by fire and greatly altered the characteristic of fire hazard. However, not every forest fire influences the forest ecosystem negatively. Surface fires that consume raw humus and surface fuels in low-productivity forests of the tundra and taiga zones influence the frost regime of the soils and lead to better growth conditions and formation of high-quality forests.

In our study area most recent fire occurred in 1997s. In each area, fire history was reconstructed up to the last 440 years; the length of some reconstruction was limited by tree longevity. Fire frequency varies in larch stands with forest type, slope aspect, and level of degradation. High fire frequency in spring is due to dry conditions, frequent lightning storms, and predominantly human activities.

Fire seasonality

We cross-dated 86 samples and recorded 102 fire scars (years) while identifying each fire scar season under microscope. After cross-dating we have made fire seasonality analysis by FHX2 software for analyzing temporal and spatial patterns in fire regimes from tree rings (Grissino-Mayer, 2004).The growing seasons of trees not indicated in the other study around the Lake Hövsgöl forest. Fires occurring after radial growth stops would produce dormant season fire scars, which means in late autumn. The early earlywood scars by fires occurring late May-early June, middle earlywood scars occurs in middle June, latewood scar by fire occurring July to August. Intensive solar radiation removes thaw water from the topsoil by evaporation, and the remaining thaw water flows from elevated sites downhill and accumulates in depressions because it cannot penetrate deeply into frozen soils. Spring fires are thus most common in our study area on these elevated dry landscape elements and in those where herbs and small shrubs form loosely compacted living ground cover layer.

The earliest fire recorded in the scars was in 1559, while the most recent scars were dated 1997 in the all six Valley forests. During the last century most of fire occurred in spring, either small number of fire occurred before 1880s.


The research result shows that radial growth rates of Siberian larch exhibited different response characteristics to environmental factors; therefore, comparisons of growth anomalies among different locations are providing useful information on past major climate events and forest disturbances such as fire.

From this fire seasonality study, we conclude that earlywood fires (spring) were dominating in our study area. More studies are necessary to characterize fire severity pattern, and more importantly, to understand the cause of this pattern and its effect on post-fire vegetation response. Also there is no long-period historical precedent of fires cited; therefore we cannot compare long-term fire years with other places in the country. However, this study produced valuable tree growth and fire chronologies around 1534-2005 years in northern Mongolia.

top of page

Asia Center at The Academy of Natural Sciences
1900 Benjamin Franklin Parkway, Philadelphia, PA 19103 |