In Mongolia, fire is a major factor which determines spatial and temporal dynamics of forest ecosystems. I also drives the trend of forest formation, varying with altitude. Out of the total of ca. 17 million ha of forest land, 4 million ha are disturbed to different levels either by fire (95%) or by logging (5%). Logged areas have increased drastically for the past 20-25 years. 600,000 ha of cuts have not yet recovered (Korotkov et al. 1990, Gunin et al. 1992). With consideration of the above data, the goal of our study was to investigate spatial and temporal forest fire patterns and conditions of their occurrence, as well as to reconstruct fire history over several centuries by dendrochronological methods.
A number of dendrochronological investigations of a variety of specific site conditions have been conducted in the last decades. Data are now available on annual increment of larch (Larix sibirica) and pine (Pinus sylvestris) in some regions of the country (Davaazhamz and Lovelius 1974, Bitwinskas et al. 1987). Annual tree increment was established to be strongly dependent on winter and spring air temperatures and on summer precipitation, with the influence of the latter varying with month. 11- and 22-year tree increment fluctuations were identified related with solar and geomagnetic field activity (Lovelius et al. 1992). Although there is quite a number of publications addressing Mongolian forests, there is still a lack of data on the influence of climate on tree increment changes. Also, there is no information available on fire frequency in forest ecosystems of this region.
Fire Frequency and Distribution in Mongolian Forests
Forest land comprises about nine percent of the total area of Mongolia. The highest fire danger is characteristic of low-mountain pine and larch stands growing on seasonally freezing soils. These stands are distributed on Khentey, East Khentey and Khubsugul foothills. Climate here is of an extreme continental type. During a year, air temperature fluctuations can amount to 90° C, with the summer maximum being +40° C. Annual precipitation ranges 250 to 350 mm. In exceptionally dry years, this value does not exceed 200 mm in forest regions. Precipitation is unevenly distributed during a year. Snow cover is not more than 10-15cm deep. Summer is the season of the lowest precipitation that usually occurs as heavy showers (Isaev et al. 1992).
Mongolian forests are characterized by high fire danger. Fire occurrence and extent are controlled by several factors, such as geographic location, climate, vegetation patterns, and human activities. The forest fire statistics for the period 1963 to 1997 are given in Table 1. The majority of fires burned within the central and eastern parts of the forested area. This can be attributed to the predominance of highly fire susceptible (highly flammable) pine and larch stands. Moreover, economic activity is much higher here as compared to other parts of the region. Extreme fire seasons are induced by long droughts. Fires burn from April to July under such conditions. Large fires occur on the background of mass fires. 1992, 1996, and 1997 were extreme fire years. In Mongolian forests, fire seasons are usually discontinuous, i.e. they have two peaks of fire danger. One peak is observed during long dry spring (from March to mid June) and accounts for 80 per cent of all fires. The other fire danger peak falls within a short period in autumn (September-October) and it accounts for 5-8 percent of fires. In summer, fires occur very rarely (only 2-5% of the total) because of heavy rains.
1985-1994 fire distribution data provided by seven forest protection air bases for Khanngai and Trans-Baikal forest zones are shown in Table 2. In these zones, fire activity is the highest in May and April – 48.1% and 33.3% of their total number in a fire season, respectively. Fires start in late March and early April, immediately after snow melt when forest fuels are drying rapidly on southern- and western-facing slopes. Strong winds of a continental-cyclonic character, whose average speed amounts to 5m/s in spring time, also contribute to fast drying of forest fuel.
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 stands on these elevated dry landscape elements and in those where herbs and small shrubs form a loosely compacted living ground cover layer. The number of fires reaches its maximum in May and June which are the hottest and driest months. In summer, abundant green vegetation reduces the fire start risk considerably. In exceptionally dry years, however, fires remain active during summer period.
Tab.1.Forest fire statistics of Mongolia 1963-1997. Source: Ministry of Environment and Nature, Mongolia.
Burnt forest (ha) (multiply by 1000)
Burnt pasture (ha) (multiply by 1000)
Number of fire
first half of 1998
Tab.2. 1985-1992 Forest fire distribution in Mongolia by month
Numerous fires in the forest-steppe and subtaiga zones are induced mainly by steppe fires that invade forest stands under certain weather conditions. In the mountain forest belt, especially in the high elevations, lightning fires are most common. Lightning storm activity increases considerably at the end of May and in early June. High fire danger is to a big extent due to prevalence of light-needled conifers in stands adjacent to steppe areas. These are mainly pine stands with mixed herb ground cover, which are characterized by high fire danger in spring and autumn. Steppe vegetation and surrounding pine stands attain high flammability practically simultaneously. Fire occurrence depends on forest type, precipitation distribution, and availability of fire sources. Fires are frequent in pine and larch stands of the forest-steppe and subtaiga zones, while they occur much rarer in larch and Siberian pine stands of the mountain taiga.
Fig.1. Fire scars at the lower part of standing old growth of pines and larches are common in Mongolia.
Study Area and Methods
We analyzed forest fire frequency in Mongolia using dendrochronological data plus official fire records and forest fund information for the period 1985 to 1994. In order to perform dendrochronological analysis and forest fire dating, we collected tree-ring samples in subtaiga pine forests of East Gubsugul highland and Selenga river valley. Sample sites are characterized in Table 3. These are mostly pine stands with mixed herb ground cover growing on rocks. Methods of reconstructing fire histories in forest communities from tree-ring analysis of forest age structures and fire scars (Fig.1) are well known, as are methods for reconstructing past climatic variations from tree-ring measurements (Swetnam 1996). Our fire scar dating project was conducted in the East Gubsugul highland where we collected chronologies on six pine sites (Fig.2). We identified scars of 56 fires which burned during the last two centuries. The biggest number of fires was recorded in pine stands on southern and western slopes (e.g., 16 fires were recorded by fire scars in a pine dry site with Carex/mixed herb ground cover).
In order to build forest fire chronologies, we collected full cross-sections from fire scarred trees, both alive and dead. Fire dates were established on each cross-section from the years when fire scars occurred. As a result, we obtained a general fire chronology, which was based upon to calculate fire intervals and fire frequency for different periods of time (Dieterich and Swetnam 1984, Fritts and Swetnam 1989). For dating fires, we used the method of cross-dating with regional and local tree-ring chronologies (Shiatov 1986). These chronologies were built using cores from old trees which had no fire scars. We used 24 P.sylvestris full cross sections.
Tab.3. 1985-1994 Dendrochronological sample sites in the East Gubsugul highlands, Mongolia.
SeriesSiteLocationForest typeNotes 1 Barum Khairlsteingold river valley 940 m a.s.l. – 49° 37′ N 104° 16′ E 25-30° SW slope Pine-Carex-mixed herbs (dry site) Exposed rocks 2 Barum Khairlsteingold river valley 940 m a. s.l. – 49° 38’N 104° 6’E 20° SE slope Pine-Carex-mixed herbs Exposed rocks Burn 1990 3 Barum-Khairlsteingold river valley 1040 m a.s.l. – 49° 33’N 104° 02′ E 25° SE slope Pine-Carex-mixed herbs on stony soil Exposed rocks Burn 1990 4 Barum-Khairlsteingold river valley 1100 m a.s.l. – 49° 38’N 104° 13′ E 15-20° NW slope Pine-Carex-mixed herbs on stony soil Exposed rocks 5 Barum-Khairlsteingold river valley 1100 m a.s.l. – 49° 38’N 104° 13′ E 15° W slope Pine-mixed herbs Exposed rocks 6 Selenga river valley 820 m a.s.l. – 49° 37′ N 104° 21′ E Flat site Pine-mixed herbs (in a valley) Old Burn
Forest Fire Frequency
Fire frequency varies in pine stands with forest type, slope aspect, and level of recreation. The shortest fire interval (4 years) was found for the pine stand with Carex/mixed herb ground cover in dry sites (Tab.4). Fire activity was the highest in spring, and autumn fires accounted for only 5% of the total number of fires recorded. Mean fire interval was found to range from 13.9 to 18.8 years in mountain subtaiga pine stands and 22.8 years in pine stands located in valleys. High fire frequency in spring is due to dry conditions, frequent lightning storms, and predominantly human activities, especially by cattle herders.
Our fire history reconstruction for the past 250 years showed 31 years with fires in the 19th century, while their number decreased to 22 in the current century. This can be due to a considerable progress in forest fire fighting that has been achieved in Mongolia over the past half century. Forest fire fighting, however, presents a big problem in Mongolia. Fires spread very fast in mountains, and the use of ground technical resources to suppress fires is limited and often ineffective because of long distances, restricted access and steep slopes. This makes fire prevention the key point. The goal should be to prevent steppe fires from invading adjacent forest ecosystems (see also contribution by Wingard and Erdenesaikhan, this volume).
Fig.2. Forest fire chronologies for pine stands of the East Gubsugul highlands.
From conducted investigation the following can be concluded: Fire seasons are usually discontinuous in Mongolia, with 80% of fires occurring in spring and only 5-8% in autumn. The development of abundant green grass layers together with heavy dragged out rains minimize fire danger in summer. However, fires can be active during summer in exceptionally dry years. Most forest fires are induced by steppe fires invading adjacent forest stands under certain weather conditions. Lightning fires are common in mountain taiga belt because of increasing storm activity in late May and early June. Extreme fire seasons occur every thee years in Mongolia. These seasons account for almost half the number of fires and 1/3 of the total area burned over the past decade. The mean fire interval varies from 9 to 22 years depending on forest type, slope aspect, and human factor.
Bitvinskas, T., V.Balchunas, I.Kairaitis, E.Gira and D.Shagdarsuren. 1987. First steps and prospects of dendrochronological investigations in Mongolia. In: Temporal and Spatial Climate Changes and Tree Rings. Part 3. Kaunas (7-38).
Davaazhamz, Z., and N.V.Lovelius. On conifer increment variability at the southern tree line. 1974. In: Structure and Dynamics of Steppe and Desert Ecosystems in Mongolia. Leningrad, Nauka Publ., 116-119.
Dieterich, J.H., and T.W.Swetnam. 1984. Dendrochronology of fire scarred ponderosa pine. For. Sci. 13, 238-247. Fritts, H.C., and T.W.Swetnam. 1989. Dendrochronology: A tool for evaluating variations in past and present environment. Adv.Eco.Res. 19, 111-189.
Girs, G.I. 1977. Physiological parameters of conifer plant resistance to fire. In: Forestry Problems in Siberia. Moscow. Nauka Publ., 148-159.
Gunin, P.D., E.A.Vostokova, Z.Davaazhamz, A.Ya.Druk, Yu. G. Yefstifeyev, Yu.N.Krasnoshchekov, I.A.Korotkov, E.I.Rachkovskaya, G.S.Samoilova, D.A.Timofeyev, and N.Uzlijkhutag. 1992. Ecological assessment of the environment and criteria for building maps of anthropogenic ecosystem disturbances in Mongolia. Ecology and Natural Resource Use in Mongolia. Pushchino, 11-25.
Isaev,A.S., E.S.Petrenko, E.N.Savin, Yu.N.Krasnoshchekov, I.A.Korotkov, and Ch. Dugarzhav. 1992. Functional characteristics and problems of sustainable use of forest ecosystems at the sothern tree line of Eurasia. In: Ecology and Use of Natural Resources in Mongolia. Pushchino, 49-57.
Korotkov, I.A., Yu.N.Krasnoshchekov, Yu.S.Cherednikova, G. Zedendash, V.N.Gorbachev, and V.A. Kuzmin. 1990. Current state of Mongolian forest ecosystems. In: Methodological Issues of Mongolian Environment Assessment. Pushchino, 58-59.
Lovelius, N.V., Z.Davaazhamz, and P.D.Gunin. 1992. Dendroindication of environmental changes in Mongolia and possibilities of their prediction. In: Ecology and Use of Natural Resources in Mongolia. Pushchino, 32-49.
Melekhov, I.S. 1948. Influence of Fire on Forest. Moscow- Leningrad, 126 pp.
Molchanov, A.A. 1976. Dendroclimatic Principles of Weather Forecasting. Moscow. Nauka Publ., 168 pp.
Shiatov, S.G. 1986. Dendrochronology of the Upper Tree Line in the Urals. Moscow. Nauka Publ., 136 pp.
Swetnam, T.W. 1996. Fire and climate history in the Central Yenisey Region, Siberia. In: Fire in ecosystems of boreal Eurasia (J.G. Goldammer and V.V.Furyaev, eds.), 90-104. Kluwer Academic Publ., Dordrecht, 528 pp.
From: E.N. Valendik and G.A. Ivanova and Z.O.Chuluunbator Address: V.N. Sukachev Institute of Forest Russian Academy of Sciences, Siberian Branch Akademgorodok, Krasnoyarsk 660036 RUSSIA Institute of Biology Mongolia and