Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Forest influences on precipitation

The comparison of the water balance of forested and treeless areas needs firstly reliable estimates of precipitation. The differences in the amounts of precipitation falling on and stored in the forest as compared to the amounts in the adjacent, treeless areas (meadows, ploughland, etc.) are usually accounted for by the differences in the mechanism of formation, fall, and accumulation of precipitation due to: (a) dynamic roughness of the underlying surfaces; (b) condensation factors; (c) interception by vegetation; (d) snow-pack formation factors. In the USSR many investigations have been carried out to estimate the influence of each of the above factors, separately and in combination, on the total quantity of liquid and solid precipitation in the forest and in the open.


The forest creates additional roughness for air masses moving in the lower atmosphere, slows down their movement and causes turbulence, which leads to the formation of ascending air fluxes, air cooling, cloud formation, and, consequently, greater precipitation on forested areas. In the USSR the idea was first proposed by A. M. Voeikov, at the end of the nineteenth century, and later supported by M. A. Velikanov, G. P. Kalinin, and D. L. Sokolovsky. This idea is basic to many recent studies of forest influences on precipitation. However, the effect may be explained by additional roughness only for flat terrain; in hilly and mountainous areas this factor can hardly be significant.

According to O. A. Drozdov, the increase cannot reach more than 3 to 5% of the total annual precipitation. However, substantially higher estimates have been obtained using the data from several dozens of meteorological stations and river basins and correlating either the measured precipitation with the percentage of forest area around each station within a radius of 25 to 30 km (or within 20 square kilometres) or the average precipitation with the forest area percentage of the river basin. The data were corrected for wind effect. According to Kuznetsova's (1957) findings for the middle Volga region, Voronezh and Tambov districts, each 10% increase of forest area results in a 2% increase of precipitation on average. For the regions of Moscow, Kuibyshev, Perm, and the Upper Volga basin to Gorky, Rakhmanov (1962, 1971) obtained the rather large increases of 8 to 10 mm in the annual totals for each 10% increase of the forest area for the period 1945-1965. For the south Siberian region Lebedev (1982) estimated that for each 10% increase of forest area the increment of annual totals was equal to 12-13 mm; Opritova (1978), using 40 drainage basins in Primorie, found the increase of annual precipitation for each 10% of added forest area to be about 25 mm. This increase, however, may be due partly to more densely forested watersheds being usually at higher elevations. Studies by Fedorov (1977) of forest influence on the water balance for a number of years at the Valdai branch of the State Hydrological Institute led him to conclude that in the north-west of the European USSR annual totals for forest are 35 to 44 mm (or 4-5% of the annual total) greater than those above treeless areas due to the greater dynamic roughness of the former; the precipitation increase is confined mainly to the warm season and does not depend on the type of forest. This was confirmed from studies of water balances of river basins. For instance, Vodogretsky (1979) found that annual precipitation above forests usually increases by 5 to 8% in the forest and forest-steppe zones of the European USSR.

Thus, statistical estimates from large numbers of stations in various regions of the USSR show that the amount of precipitation falling on forests is somewhat greater than that above treeless areas. However, this fact is far from being universally recognized. As the increase is rather small (4 to 6% according to the latest findings) and the accuracy of precipitation measurements is rather low and varies for precipitation gauges installed in the forest and in the field due to different wind protection, many scientists in the USSR (Bochkov 1970; Bulavko 1971; Krestovsky 1969a; Subbotin 1978) do not accept these conclusions and do not take the phenomenon into consideration when calculating water balances and developing runoff forecasting methods. The problem needs further critical investigation.

Horizontal Precipitation

The hydroclimatic role of forests often includes condensation, dew, hoarfrost, etc., and horizontal precipitation (mist catching). Reliable quantitative estimation of this kind of precipitation is difficult and is not included in standard observations of meteorological stations (save for stating its occurrence). There are published, however, quantitative assessments of horizontal precipitation, indicating its importance in some regions. Such data are quite insufficient to make reliable comparisons of the quantities of this kind of precipitation in forests and in treeless areas.

According to Voronkov (1970), based on four years' observations, condensation under the canopy yields, on average, 10 to 15 mm during the warm period and 16 mm during the cold period, or 25 to 30 mm per year (4% of the annual precipitation). In cedar forests of the western Sayan Mountains liquid horizontal precipitation penetrating the canopy was estimated by Protopopov (1975) at about 7 to 10% of the total precipitation. According to Lebedev's data (1982), in mixed cedar-fir stands in the same region, liquid condensation makes up 15 to 30 mm, or about 2 to 3%, of the annual total; estimates obtained for fields with a vegetative cover are about the same. Condensation during the winter period is included in the snow-storage estimates. In the virgin steppes of northern Kazakhstan (Gidromet. 1966) maximum condensation is observed during the summer-autumn season and reaches 30 to 40 mm (20-30% of summer-autumn precipitation, or 9 to 11% of the annual total).

According to data summarized by B. L. Sokolov (unpublished) condensation plays an important role in the water balance of the permafrost regions of Central Yakutia and Kolyma, reaching during the warm season 80 to 100 mm on the average for a whole river basin. Condensation values may have a wide range depending on the land surface; that is, from 50 mm to 300 mm, maximum values being observed on gravel and rocks without a vegetative cover. Condensation leads to the increase of specific river discharges in the permafrost zone equal to 0.5 to 2.0 litres km-2 s-1 and occasionally to 4 litres km-2 s-1 in small watersheds. Large condensation values amounting to 100-150 mm during the warm period are observed in the forests of Primorie. Still greater condensation values are recorded in many other countries of the world, in mountainous coastal regions in particular (see the review by Rakhmanov 1981). Condensation in such regions amounts to about 20 to 25%, and sometimes half, of the total annual precipitation.

In the plains of the European USSR, with a continental climate, condensation seldom exceeds 3 to 5% of the annual total and seems to be the same on average for forests and treeless areas (Bulavko 1971; Gidromet. 1966). There are two prerequisites for condensation: high air humidity and a large daily air temperature variation. In forests air humidity is higher, while large diurnal air temperature changes are typical of the fields; hence these two factors are believed to compensate each other and thus cause the similar condensation yields.

Maximum condensation amounting to 20 to 30 mm, 3 to 4% of the annual total, is observed at the borders of forests, while minimum values, of 5 to 10 mm, or 1% of the annual total, are recorded at distances of 100-150 m and further into the forest. Winter condensation values are 10 mm and 5 mm (5% and 2% of winter precipitation) at the borders and deep in the forest respectively. In fields annual condensation is about 20 to 30 mm, or 3 to 4% of the anual total. Thus there is hardly any reason to believe that condensation, or horizontal precipitation, in the forest is greater than that in treeless areas.

Interception Loss

For water balance studies one also needs to carefully assess the portion of precipitation that is intercepted by the canopy and lost by evaporation without reaching the ground. Comparing observations from precipitation gauges installed in the open and under the forest canopy, Molchanov (1960) gives the following interception values: 34 to 46% of the annual total in spruce forest; 24 to 27% in pine forest; 24% in birch; and 22% in oak stands. Similar interception estimates were obtained by Voronkov (1970), Pobedinsky (1979), and Idzon, Pimenova, and Tsyganova (1980) for the Moscow region and by Bulavko (1971) for Byelorussia. In the Middle Urals and the northern part of the European USSR, Bratsev and Bratsev (1979) and E. P. Galenko et al. (Bratsev 1982) found that spruce intercepts 29 to 46%, pine 25 to 30%, and deciduous species 15 to 24% of total precipitation either during the growing season or a year. Interception values tend to be greater (40-60%) in dark fir and spruce forest in the Carpathian and the Caucasus mountains (Kaliuzhny, Pavlova, and Popov 1979; Rakhmanov 1981). For the forests of the Asian part of the USSR, data by V. V. Protopopov, L. K. Pozdnyakova, T. I. Tarakanova, A. P. Klintsov, and others from different types of forests in Yakutia, the Sayan Mountains, Primorie, and Sakhalin Island are summarized in V. V. Rakhmanov's 1981 review. The data show that interception values depend on forest type and density and the precipitation type (solid or liquid, amount, intensity, and duration).

In the USSR predictive equations for interception based on forest type and precipitation were suggested by A. I. Gribov, V. V. Protopopov, V. D. Chernyshev, and L. P. Kharitonov. In many formulae the main parameters of forest stands are the socalled leaf area index (i.e. the total area of all leaves divided by the area occupied by the forest) and the water-holding capacity of the canopy or a tree. The leaf area index, according to different authors, ranges in the temperate zone from 4 to 15 for deciduous forests, from 6 to 18 for pine forests, and from 20 to 40 for spruce and fir stands. As a rule, the greater the leaf factor the greater is the interception.

The water-holding capacity of the canopy or separate tree crowns is the maximum amount of moisture held by completely wet trees. It depends on the forest type and the tree age. For deciduous trees it ranges from 0.5 to 1.2 mm; from 0.9 to 1.5 mm for pine trees; and from 2.8 to 4.6 mm for spruce and fir, occasionally reaching 6 to 8 mm (Rakhamanov 1981). It should be noted that the water-holding capacity cannot serve as a direct indication of intercepted amounts since a considerable portion of precipitation is shaken off by the wind. These estimates of the water-holding capacity of trees clearly demonstrate that total interception depends not only on the type and density of forest but also on storm characteristics: the greater the proportion of low-intensity precipitation (which is almost entirely intercepted by all species) the greater is the total interception.

The most complete and detailed data on precipitation interception of both snow and rain by forests were collected at the Valdai branch of the State Hydrological Institute by Fedorov (1977) and Krestovsky (1969a) for stands of different species, density, and age. Krestovsky (1969a, 1969b) and Krestovsky and Sokolova (1980) also analysed observations made between 1900 and 1980 within the USSR area and in some countries of Western Europe. They showed that in some cases data are overestimates, evidently due to storm duration and intensity not being taken into account and due to reference precipitation gauges being in small clearings amidst high trees where precipitation is usually greater than above the forest. For reliable measurement of interception loss one should observe for a longer period, perhaps ten to thirty days for liquid precipitation and for the whole winter period for solid precipitation, starting from the first snowfall to the start of intensive spring snow melt when all the snow on the trees disappears. Precipitation gauges should be installed in clearings surrounded by rather low, not dense, preferably deciduous, forests.

Generalized data fulfilling the above requirements are given in table 1. It shows the portion of precipitation reaching the soil surface in forest stands of various type and density regardless of the height and age of the trees. The large seasonal and annual interception values are typical of spruce forest, reaching on the average 25% (forest stand density being 0.8). It should be noted, however, that in very dense mature spruce forest (with density 0.9 - 1.0) and in dense young spruce forest, interception amounts to 30-35%, whereas in dense young pine forest it is 25% of precipitation during both the warm and the cold periods (Voronkov 1970; Krestovsky 1969b; Fedorov 1977). Since forest type and density tend to change with time, the interception of precipitation by the canopy also changes considerably. Because the interception value depends on the crown volume, which is related to the weight of the foliage, interception should relate to the phytomass of the forest stand; such correlation has recently been established by the junior author of this review (see fig. 3).

TABLE 1. Ratio of precipitation reaching the soil surface to precipitation amount falling on forest.

Forest type Average forest density Portions of precipitation reaching the surface
Oct.-Apr. May-Sept. Year
Spruce 0.8 0.75 0.75 0.75
0.4 0.80 0.80 0.80
Pine 0.8 0.80 0.80 0.80
0.4 0.90 0.90 0.90
Pine-spruce 0.8 0.75 0.75 0.75
0.4 0.85 0.85 0.85
Mixed 0.8 0.92 0.80 0.85
0.4 0.97 0.85 0.90
Deciduous 0.8 1.00 0.85 0.90
0.4 1.00 0.90 0.93
Deciduous brushwood withconiferous undergrowth 1.0 0.95 0.80 0.85
Second growth of 10-15 years on felled areas 0.5 1.00 0.85 0.90
Open deciduous and marshy dwarf forest 0.1-0.3 1.00 0.95 0.97

Sources: Krestovsky 1969a; Krestovsky and Sokolova 1980

To estimate the hydrological role of forests one also needs estimates of precipitation interception by adjacent fields and grassland. Unfortunately, there are few reliable data on interception by different types of grass and crops, due to experimental difficulties. Therefore, interception by grass coenoses is often estimated indirectly, by using the leaf area index and water-holding capacity. According to different findings summarized by Rakhmanov (1981), the leaf area index of grasses and crops ranges from 10 to 85, being, for instance, 26 for clover, 80 for a , and from 20 to 50 for various grasses. These values are 1.5 to 2 times greater than those for forest stands, mentioned earlier. This does not mean, however, that grasses and crops intercept greater amounts of precipitation: firstly, the vegetative period of grass and crops is much shorter than that of deciduous species, to say nothing of the conifers. Secondly, the rate of evaporation from tree crowns and from herbaceous vegetation is very different. Detailed experiments on interception by herbaceous vegetation were carried out by Bulavko (1971). He used a special design of ground precipitation gauge and has studied for several years the interception of precipitation by various crops and grasses in the regions of Byelorussia. He also thoroughly reviewed the relevant data available in the USSR. He finds that interception values depend both on precipitation (type, rate, distribution in time and amount) and on the vegetative cover (type, density, and developmental phase). Interception by herbaceous vegetation, including crops, was observed only during the growing season, approximately five to six months. During the period of maximum plant development interception values may reach 30 to 37% of the monthly precipitation. During this six-month period rye, wheat, and barley intercept 8 to 9% of precipitation, while potatoes and perennial grasses intercept 13 to 18%; the respective annual values appear to be half as much, from 4 to 9% of annual precipitation. Interception estimates for deciduous forest are approximately 1.5 to 2 times greater and those for conifers 2 to 3 times greater (see table 1).

Snow Storage and Forest

The forest influence on snow storage is another aspect of its effect on precipitation. This aspect has great practical significance, especially for the plains of the USSR where streamflow is due mainly to the spring snowmelt. The main method of studying the forest influence on snow accumulation is by special and standard snow surveys in the forest and in the open.

Special snow surveys made by Molchanov (1961) and by Sabo (1956) in the Moscow region showed that maximum snow storage is in forest clearings. If one assumes storage here to be 100%, snow storage in pine stands of different ages and densities ranges from 75 to 93%, in birch and aspen forest from 84 to 98%, in spruce forest from 67 to 71%, while in open areas it is between 67 and 82%. Thus in the open snow accumulation is less than in all types of forest except spruce forest.

Subbotin (1966), using the data of the Podmoskovnaya water balance station for a 20year period and the data of routine snow surveys at 135 meteorological stations in the northern and central regions of the European USSR, showed that snow storage in conifer forests is 8 to 28% greater than in the open, and for deciduous forests the difference is 25 to 32%, reaching 42% in the south. Much lower estimates of snow storage in forests were obtained by Rakhmanov (1962). He used the data of snow surveys in the open and in the forest conducted at 125 stations of the western, central, and eastern regions of the forest zone of the European USSR. He concluded that the amount of snow accumulated in the forest is on average 17% greater than in the open. He also found that 27% of the snow surveys showed that snow stored in the forest appeared to be 5 to 10% less than that in the field. Later Rakhmanov (1971) studied snow cover formation in the Volga basin (drainage area: 229,000 km2) using the data of snow surveys at 170 meteorological stations during a 14-year period. It appeared that in coniferous forests snow storage is 10% greater than that in the open, while in deciduous and mixed stands the difference amounts to 27-28%. On average, in the forested area of the drainage basin snow storage is 20% greater than that in the open terrain.

There are, however, findings that do not support the above estimates. For instance, according to Voronkov (1970), based on 19 years of observations in experimental areas, snow storage in spruce and spruce-deciduous forests is usually less (approximately by 9% on the average) than in the nearby fields; occasionally in some years it may be quite the reverse. In the Middle and South Urals (Pobedinsky 1979) snow storage in mature spruce forests is also less than in treeless areas, whereas snow storage in pine and deciduous forests is usually the same as in the open.

In the forest zone of Siberia and the Far East snow surveys generally show that snow accumulation in forests, especially in mountains, is much greater than in the open, treeless areas. Protopopov (1975) found that for mixed and dense conifer mountain forests of Siberia the snow-storage coefficient is from 1.18 to 1.32; for deciduous and open cedar stands it is 1.16 to 1.22; in dense cedar and pine forests the coefficient is 0.80 to 1.10. Detailed studies of snow accumulation in the forest and in fields in different Siberian natural vegetation zones were conducted by Lebedev (1982). The snow storage coefficients in the forest zone were 1.1 to 1.2, rising to 1.5-1.7 in the steppe-forest zone and to 1.7-2.5 in the steppe zone. The very large values in the foreststeppe and steppe zones refer to small forest stands that accumulate additional snow carried by the wind over vast open areas. Thus wind is an important factor since it redistributes snow between field and forest. In the regions of East Siberia, the Far East, and Sakhalin, forests accumulate substantially higher snow amounts than fields (Klintsov 1973; Rakhmanov 1981). For instance, summaries of numerous snow surveys in the Baikal region showed that snow storage by forests at all elevations is approximately 30% greater than that in fields. This can be attributed primarily to the big differences in evaporation of snow in the open and in the forests, especially under conditions of clear skies during the anticyclones, which are frequent in these regions.

The main results of intensive surveys of snow storage in forest and open sites carried out since 1970 for different regions of the forest and forest-steppe zones of the European USSR, and in the steppes of Kazakhstan, are presented by A. P. Bochkov (1970); Vershinina (1972, 1979); Vershinina and Volchenko (1974); Gidromet. (1966); Krestovsky and Sokolova (1980); Ouryvaev et al. (1965). According to these studies the long-term ratios of maximum snow storage in the forest to that in the fields are as follows:

North-west of the European USSR 1.20-1.30
Central regions of the forest zone of the European USSR 1.20
North-east of the European USSR 1.10-1.20
Forest-steppe zone of the European USSR 1.30-1.80
Steppes of Kazakhstan 1.30-2.0

In the case of small forest stands surrounded by vast open space, the ratios are 1.30 to 1.50 for the forest-steppe zone; 1.5 to 2.0 for forest-steppe and steppe zones of the European USSR; and 2.0 to 3.0 for the steppe regions of Kazakhstan.

Detailed studies on the maximum snow storage in forests of different species and density and in different sizes of open areas have been carried out recently in the forest zone of the European USSR (Krestovsky and Sokolova 1980; Krestovsky 1980). The results are shown in table 2. The maximum amounts of snow are accumulated in the clearings in the middle of the forest as demonstrated by numerous experiments (Vodogretsky and Krestovsky 1975; Krestovsky 1969a; Fedorov 1977; Fedorov et al. 1981). In general, snow storage in forests is about 10-20% greater than that in open terrain, though in dense spruce forest snow storage is close to that in fields.

TABLE 2. Snow storage in forest and fields by the start of the spring snowmelt (longterm average relative snow storage for the forest zone of the European USSR)

Field and forest types Relative snow storage
Mixed (50% conifer, 50% deciduous) species 1.00
Deciduous 1.04-1.10
Spruce, normal density (0.3-0.6) 0.90-0.97
Spruce, high density (0.7-1.0) 0.80-0.90
Low mixed forest, dwarfed species 0.98-1.02
Brushwood and deciduous low forest 1.04-1.08
Large felled areas 1.04-1.08
Small felled areas and clearings surrounded by high forest 1.05-1.15
Mossy forest (moss bogs) 0.95-1.00
Field (ploughland, grassland, stubble, slopes of various orientation and flat areas): average 0.75-0.92
Fields, northwestern regions 0.75-0.79
Fields, central regions 0.85
Fields, northeastern regions 0.87-0.92

The ratios may change markedly from year to year. During mild winters with thaws they may reach 50% but fall to 5% during very cold frosty winters. The occurrence of different numbers of mild and frosty winters during a long-term period largely explains the various values of snow storage in fields of the north-west, central, and north-east regions of the European USSR (see table 2). It also explains the contradictory estimates obtained by different authors for the same regions. However, even a considerable reduction of snow storage in the field during mild winters would not necessarily cause the same reduction of water yield since melt water saturates the soil and reaches the aquifers.

Contents - Previous - Next