H. G. WILM
Associate Dean, State University College of Forestry, Syracuse, New York
As a feature of its current program of work, the Forestry Division of FAO is preparing for publication a study on the influence of the forest on water, soil and climate, and their bearing on land-use policy. This article, is an extract from the concluding chapter by Dr. Wilm. Other contributors to the study are: Professor Joseph Kittredge, Dr. R. J. van der Linde and Dr. Marvin D. Hoover: a general introduction is given by Professor Aldo Pavari.
IF any broad generality can be stated about the problem of forest influences in the and tropical humid zones, perhaps it is this: that these influences tend to be more striking and pronounced than in more temperate zones; and that, therefore, the need for wise and careful management of forest vegetation is more acute, so as to achieve optimum regulation of water and stability of soil. If this is true, it means that the forest land manager has a particular need to understand the principles underlying forest influences, so that he can know how to protect or manipulate the vegetation on his land to attain the desired objectives. As background for the present development of forest land-management policies and procedures, these principles will be reviewed and summarized to indicate the relation of forest vegetation to climate and especially precipitation; to surface runoff and land stability; and to the regimen of streams. Particular reference will be made to the effects of forest manipulation, as a guide to the management of water shed land.
Forest influences
One of the earliest and most widespread conjectures on the relation of forests to water has been their influence upon climate. As suggested elsewhere (Kittredge, 1948), this influence is hard to prove or disprove in terms of general regional climate. The areas compared are necessarily large, and inherent variations tend to obscure any regional effect of the presence or absence of forest vegetation. At least in certain cases, however, there seems to be a strong rational basis for arguing a regional effect. Where forest vegetation occurs in areas exposed to warm maritime air masses, it seems quite likely that the presence of the forest exerts a measurable effect upon total precipitation and perhaps upon regional climate. Illustrations are the redwood forests along the Pacific coast of the United States, and certain forest areas in West Africa.
Entirely aside from conjectural regional effects, forest vegetation is known to exert substantial effects upon local climate within and adjacent to its own area. Universally, as can be testified by anyone who has taken shelter under trees or in the lee of shelter belts, the presence of forest vegetation tends to ameliorate local climate. Maximum temperatures are lowered and minima raised; relative humidity is a little higher; and wind velocities may be reduced a little or a great deal. These are generally beneficial influences, especially in and regions. But of greater interest to the watershed land manager is the influence of forests upon local precipitation, runoff, soil movement, and water losses.
As precipitation falls on forest cover, the first influence is interception. Depending on the character of precipitation and the density of the vegetation, it variable amount of water is held by the forest canopy and evaporated into the atmosphere without reaching the ground. Experimental data obtained for a variety of forest types and climates indicate that the amount of precipitation intercepted by a forest canopy is it purely mechanical function of the storage space avail able on vegetation surfaces when the rain begins. A heavy forest of trees with long, dense crowns can there fore be expected to intercept maximum amounts of water from any storm; minimum interception would be expected from a thin cover of sparse vegetation.
Aside from the effects of the kind and density of vegetation, however, interception tends to become less as a storm progresses and the forest canopy becomes wetter. Toward the end of a large long storm, on], % enough additional water is retained on the crowns to offset any evaporation from the leaf surfaces that occurs during the storm. Thus the greatest possible losses of water could be expected in climates that are characterized by a large number of light, small storms, separated by periods of clear weather. But in any event, interception means a loss of precipitation through evaporation.
Aside from this loss, forest vegetation also consumes water by evapo-transpiration. After precipitation has passed through the forest canopy, some or all of it moves into the soil. Some fills up storage space available in the "pores" of the soil: filling the capillary spaces and then, if a sufficient volume of water is applied, the larger, non-capillary spaces. As the storm proceeds to its end, water in non-capillary storage passes on by gravity; either into deeper storage, to ground-water tables, or through subsurface flow into streams. The remaining water is held against gravity in the soil and some or all of it is drawn back out into the atmosphere by direct evapo-transpiration or transpiration through the roots and canopies of vegetation. The volumes of water consumed in this fashion vary with temperature, humidity, and wind movement; with the volume of soil-water in storage, with the extent and depth of rooting systems, and with the amount of time which elapses between storms.
The relative amounts of water consumed by direct evaporation and by transpiration through vegetation can be reasoned out, although they are not easily separable in actual measurement. Direct evaporation occurs from water standing on the soil surface or in the interstices of the soil. Rates of evaporation are greatest at or near the surface and decrease rapidly with depth, reaching negligible rates at depths of perhaps one-half-meter or so in most soils. It may occur at greater depths, of course, if the soil mantle is deeply cracked and fissured. Then, in effect, the area of land surface exposed to air is increased and projected into the soil mantle.
As to transpiration, the total volume of water removed depends upon several inter related factors: not only the volume of water available in soil storage, but also the depth and completeness of root systems, and the hydrologic depth of the soil. The greatest amounts of water are consumed by transpiration through deep and complete systems of roots in deep soil, penetrated to its full depth by stored water. On the other hand, transpiration may be decidedly limited if the root systems are restricted by low density of vegetation, or by a shallow hydrologic depth of the soil: where the penetration of water into soil is restricted, either by the presence of a dense layer such as rock or clay a relatively short distance below the land surface, or by deficient precipitation which never penetrates beyond a relatively shallow depth.
Transpiration naturally decreases with reductions in vegetation density, and evaporation tends to increase. Thus, in shallow soils the sum of evaporation and transpiration would tend to remain constant even with variations in the density of vegetation - as observed by Wilm and Dunford in Colorado (U.S.A., 1948). Where the soil is deep, vegetation dense, and precipitation abundant, as in the Southern Appalachian Mountains of United States, on the other hand, the removal of vegetation reduces evapo-transpiration and increases water yield to a pronounced degree (Hoover, 1944).
Perhaps these thoughts deserve further amplification. According to Thornthwaite's (1955) concept of " potential evapo-transpiration ", a calculable amount of water will be moved into the atmosphere from a land surface for any given set of conditions - and with a continuously ample supply of water in the soil.
With an abundant water supply, plants tend to transpire in proportion to their receipt of solar radiation. But as soil-water supply decreases, plant stomata close and transpiration decreases even though solar radiation remains high.
Thus the actual losses of water from a soil mass are commonly lower than the "potential" quantities because of deficiencies in soil water supply. In and climates, the total precipitation may be only a small fraction of the potential consumption; and this potential amount can be attained only in areas of continuously high ground-water (Wilm, 1952). Hence it is only in such areas - along streams or at oases -where abundant vegetation can develop. In some and localities, as in the southwestern United States, such vegetation causes serious losses of water yields.
Where ground water is not available, on the other hand, water consumption in and areas is only the -amount of water that penetrates the soil from precipitation. After surface runoff is subtracted, this consumption is only a few inches or centimeters per year. And, because this water penetrates only a short distance into the soil before being consumed, it is all lost by either evaporation or transpiration, whether vegetation is present or absent. Under such conditions, vegetation has little or no influence on water consumption; but it still serves an important function in soil stabilization.
Extension of these ideas suggests that actual evapo-transpiration may be expected to increase with increases in available water supply, up to a maximum equalling potential evapo-transpiration. This seems to be generally true; but two examples will illustrate that the underlying principle should not be oversimplified.
First, it is often inferred that actual evapo-transpiration is independent of variations in vegetation, as long as the ground surface is well covered. This might be true if the available supply of water to vegetation did not change. such variations in vegetation. Such a condition would be found in areas of constantly high water-table or of shallow soils and root systems. But consider an area of deep, well-drained soils and abundant, but well-distributed precipitation with sunny periods between rains. If no vegetation existed or if it had only shallow root systems - say, less than 0.5 meter deep - the total water available for evapo-transpiration must certainly be far less than if the vegetation were a mature forest with deep, fully-developed root systems occupying a column of soil perhaps two meters deep. The effect of removing such a forest and replacing it with a dense young sprout-hardwood forest is graphically shown by experiment 8 in the southeastern United States (Hoover, 1944, Kovner, 1945). There, the removal of a mature, hardwood forest decreased annual evapo-transpiration from about 39 inches (99 centimeters) to about 20 inches (51 centimeters). As the sprout forest grew up, the annual water loss slowly increased again. Even 12 years after the clear-cutting, however, the remaining loss was still several inches smaller than before treatment.
Second, it is often considered that evapo-transpiration opportunity increases directly with annual precipitation. This may be generally true when total annual precipitation does not greatly exceed potential evapo-transpiration. But in humid areas with excessive precipitation, it seems logical that evapo-transpiration opportunity may actually decrease with further increases in precipitation. This seems to be illustrated by precipitation-runoff relations in the Luquillo Mountains of Puerto Rico, where precipitation is around 200 inches (500 centimeters), and approximate runoff data indicate total evapo-transpiration losses around 20 inches (50 centimeters) or less. Incidentally, this particular area is characterized by a dwarf "rain forest", which might also indicate suppressed transpiration rates.
Aside from its influence on the volume of evapo-transpiration, forest vegetation exerts other extremely important influences upon the disposition of water. After precipitation has passed through the forest canopy, not all of it may enter the soil, passing into soil storage or ground water and then moving in a tranquil manner into streams. Depending upon the rate of precipitation and the density of upper soil, smaller or larger quantities of water may run off over the land surface. Precipitation rates cannot be controlled by land-management activities; but the density of the upper soil layers may be affected tremendously by the growth, manipulation, or deterioration of vegetation. As an undisturbed forest vegetation develops, it tends to build a more and more favorable environment for the formation of permeable soil. Organic litter is first deposited upon the land surface. As it begins to decompose, organic decomposition products move into the upper soil, providing favorable conditions for bacteria and other plant and animal life to go to work in building soil structure. Also, a network of " feeder " roots penetrates throughout the topmost soil layers, providing mechanical stability as well as creating myriads of underground channels as they die and rot.
Thus, over a period of time the development of forest cover builds optimum conditions for the penetration and storage of water. On the other hand, the removal of vegetation - even without mechanical disturbances - tends to cause degenerative changes to a greater or less degree. Removal of the sheltering canopy exposes litter and humus to increased oxidation, to the impact of rain, and to removal by wind. Progressive deterioration of microfauna and flora, the loss of feeder roots without replacement, and the battering of rain all tend to compact the surface layers until, at some critical stage of deterioration, surface runoff begins to occur. If, then, the recovery of vegetation is not accomplished, a progressive cycle of further deterioration develops. Surface runoff stirs up sediment, which passes into the upper soil interstices; this means lowered porosity, which means further increased surface runoff, and so on. The cycle ends in a relatively stable condition of low infiltration and storage capacities of the watershed, and excessive rates of surface runoff. High floods occur in the streams during rainy periods, and dry channels between the storms.
Such watershed degeneration is aggravated by mechanical soil compaction caused by the cutting of skid trails and the construction of roads; by the movement of logs; by the pressure of heavy logging machinery-, and by the trampling of animals while working or grazing. On pastured and range lands especially, trampling of animals can be an important cause of watershed deterioration. So also can trampling by human beings, in areas of high concentration.
From this discussion it can be seen that the whole object of watershed management or protection is to avoid causing deterioration of forest cover to a point beyond the critical stage of degeneration. In some areas, with relatively porous soils and low rates of precipitation, this may permit quite heavy use of the vegetation without serious damage. In other areas, with high rate of precipitation and easily damaged soils, it may require complete protection from any kind of use.
In summary, then, several statements can be made regarding forest influences on water supply, soil stability, and stream regimen. As a generality, vegetation consumes water in varying amounts, depending upon the climate and character of vegetation. Ordinarily such water consumption is an important factor in watershed management; although in extremely and or wet climates or where the soil is hydrologically shallow, the presence or absence of vegetation may make little difference in the total yield of water from precipitation. Its presence may, however, exert a tremendous effect in stabilizing soil and regulating streams.
If, therefore, the primary object of land management is flood regulation and soil stabilization, the aim should be to maintain the densest possible cover of vegetation. If, on the other hand, the object is to obtain maximum water yields compatible with soil stability and flood regulation, the land manager would want to maintain minimum densities of vegetation above the critical level of deterioration.
It will be noted that these concepts introduce policies for management and protection which differ from policies developed for timber production alone, and may be in conflict with them. The characteristic idea in forestry policies is the production of maximum volumes of timber per unit area. In contrast, the watershed land manager wants to maintain densities of forest cover which will provide the desired results in water-yield production and the stabilization of streamflow and soil. In some cases, this means complete protection of the vegetation, which yields no wood products at all. In other cases, where water yields are important, it means the maintenance of forest stands substantially lower in density than desired by silviculturists for maximum production of wood, Necessarily, in all cases these differing objectives must be harmonized so as to provide the greatest long-time returns to the greatest number of people concerned.
Application of principles to problem regions
Arid zones. As regards forest influences, the and zones of the world may be considered to have two outstanding characteristics. First is an obvious shortage of water which, as civilizations develop, leads to intense demands. Second is a characteristically slow and sparse development of vegetation, correspondingly serious damage from exploitation, and slow recovery. These factors call for the most careful designing of forest land management to achieve optimum water production and watershed stability. Typically in the drier zones, any substantial encroachment upon the forest means a real setback in watershed conditions: one which takes a long time to overcome, and often goes beyond the critical stage into a severe cycle of deterioration which may end in permanent or semipermanent desert. In West Africa, for example, the cutting and grazing of the forest in Northern Nigeria and the Sudan country has resulted progressively in the degradation of the deciduous forest, the development of " savannah " types, and the southward advance of the Sahara Desert (Stebbing, 1939). In some cases, however - probably where the soil is more pervious - the degradation of the deciduous forest may not result in deteriorated watershed conditions:
"A degraded type of deciduous forest in which the fire-resisting species naturally predominate, becomes, in the course of decades or centuries of this treatment, of poorer height growth with contorted stems and an open cover. All this was represented today, the general average of height growth being about 25 to 30 feet (7.5 to 9 meters): though we got into patches, usually with dense elephant grass, of taller trees, and a dense canopy, even intact in a few places.The general character of the forest was what interested me. It is not yet past praying for. The water in this region, as I have noted, is still plentiful in the streams, and it is obvious that reservation here would retain and preserve the existing conditions, which are good. It is equally obvious that, if the people are allowed to continue firing the forest in the present fashion, the existing forest must gradually deteriorate to such a point that water supplies will be affected and that the present apparent prosperity of this region will be reduced." (Stebbing, 1939, page 41).
Stebbing suggests the establishment of belts of timber in an east-west direction to arrest the southward progress of the Sahara Desert: one across the Haute Volta and across to Lake Chad, and the other through the Gold Coast and Nigeria in the approximate latitude of Kintampo. He also seems to favor, in general, the preservation of timbered areas from public use as a means to preserving watershed conditions. Wherever shifting cultivation is practiced, he suggests the taungya method of planting trees immediately in the cleared areas, while they are being used for cultivation, as a means of land restoration. The taungya method should be combined with the regulated rotation of shifting cultivation, so as to give maximum opportunity for land recovery.
In the semiarid areas of the United States such as the Front Range of the Rocky Mountains and areas in Arizona and New Mexico, large tracts of land have been seriously deteriorated by generations of cutting, repeated fires, and overgrazing. The result in some cases has been depletion or removal of the forest cover; in others, invasion of grazing land by pinyon pine and juniper. Here, one solution has been the establishment of national forests in which timber cutting, grazing, and fire are under the regulation of the Federal Government. Occasionally, however, a consequence has been further overburdening of lands outside the national forests.
From these examples and others which might be cited it seems evident that, in the and zones, the management of watershed lands calls for special care. In some cases it may be possible to remove timber to a considerable degree, if it is replaced by good stands of grasses and other forage plants and if grazing can be carefully regulated. Otherwise, too often the consequence is a prolonged cycle of serious deterioration.
Tropical humid zone. In these great areas of the world the climate and vegetation present a completely different set of problems. Typically, the precipitation is excessive - even to the point that evapo-transpiration may be suppressed below the theoretical maximum. Vegetation develops rapidly except in the wettest areas, and forms almost impenetrable cover which more than adequately protects the Boil against erosion, After the forest is removed by clearing or burning, its return is so rapid that the soil is again completely covered in a surprisingly short space of time. At the same time, surface erosion from cleared land is severe; and shifting cultivation has been many times cited as the source of sediment, floods, and deteriorated soil. It is also associated with the existence of impoverished populations.
All of these statements are doubtless true at least to some degree; it is suspected that some of them have been overdone. Certainly, as a primitive form of agriculture, shifting cultivation is accompanied by corresponding waste of human effort and impoverishment of land. And erosion does accompany this kind of agriculture on steep slopes and in areas of torrential rainfall. But its overall magnitude and long-term implications are, perhaps, a different question.
In Puerto Rico, for example, shifting cultivation has been cited as a serious cause of sedimentation into valuable reservoirs. But a quantitative analysis by the United States Soil Conservation Service (Noll, 1963) of silting into the Caonillas Reservoir in Puerto Rico indicates an annual loss of only about 0.024 inch (0.065 centimeter) of soil on the watershed. This rate of erosion would mean an ultimate reservoir life of approximately 800 years- longer than can be expected in many areas of more gentle climate and less spectacular cultivation problems. There, as elsewhere, the most serious erosion per acre came from clean-cultivated lands. At the time of the survey these occupied only 3 percent of the watershed, but produced 23 percent of the soil loss. Coffee lands, occupying 26 percent of the area, produced 40 percent of the loss. Noll suggests that this erosion can be greatly reduced by maintaining the vegetative cover underneath the coffee trees. Unimproved pastures, occupying 37 percent of the watershed area, produced 21 percent of the soil loss; and brushy forests, occupying 26 percent of the area, produced 14 percent of the loss. In order to minimize soil erosion and land deterioration, Noll recommends improved management of the coffee lands, pasture improvement by fertilization and other means; conservation treatment of the cultivated land; a grassland type of farming, and increases in forest area. As far as possible, shifting cultivation needs to be placed under control wherever it occurs; perhaps in combination with reforestation of the cleared areas by some system such as the taungya method.
Summary
In both the and tropical humid zones, uncontrolled forest exploitation sets up obvious dangers. In and areas, it means serious deterioration of land, the production of flashy floods, and lower summer flows. In humid areas, it means excessive rates of erosion and lowered productivity of land. Both consequences imply the need for definite policies designed to control forest exploitation: in some cases, to provide complete protection of watershed lands; in others, to permit management of forest and grazing lands to a degree which will not cause deterioration below the critical level. In and regions, the primary aim of such protection and management should be the control of water yields and the attainment of maximum soil stability. In tropical humid regions, the corresponding objectives should be the enhancement of land productivity, together with the reduction of soil erosion to a reasonable level.
In both zones, wherever watershed problems are of sufficient importance, policies for forest protection and management should be dominated by the ideas of optimum forest density. Where flood control and soil stabilization are of crucial importance, this optimum density may be the maximum obtainable through complete protection or the most careful manipulation of forest cover. Where soils are stable, rainfall and water yields are low, and the attainment of maximum water yields is of critical importance, the optimum density of forest vegetation may be substantially lower than that desired for maximum timber production. There the objective would be to maintain the thinnest possible stand of timber compatible with maintaining a stable watershed. In some areas, this objective might even be reached by complete removal of the timber and substitution of a good cover of range forage or of shrubs.
It should be particularly emphasized, however, that much of this fabric of conclusions and recommendations is based upon extrapolation of existing knowledge, obtained largely in the temperate zones. Throughout the and tropical humid zones of the world there, is a great need for quantitative research and pilot-plant projects designed to work out the optimum methods of management for forest and range watershed lands.
Until such research has been successfully carried through, it will continue to be difficult for a land manager or a policy-maker to assess the relative values of protect-ion as compared to reasonable management, or even to decide what constitutes reasonable management.
HOOVER, MARVIN D. - Effect of removal of forest vegetation, on water yields. Trans. Amer. Geographical Union. Part IV, pp. 969-975. 1944.
KITTREDGE, JOSEPH - Forest Influences. 394 pp. illus. McGraw-Hill Book Co., Now York-Toronto-London. 1948.
KOVNER, J. L. - Changes in streamflow and vegetation characteristics of a southern Appalachian watershed brought about by forest cutting and subsequent natural regrowth. Ph. D. thesis, typed, 245 pp. State University College of Forestry at Syracuse University, Syracuse, New York. 1955.
NOLL, JOHN J. - The silting of Caonillas Reservoir, Puerto - Rico. 22 pp. mimeo, illus. Soil Conservation Service, U. S. Dept. of Agric. Tech. Paper 119, 1953.
STEBBING, E. P. - The Forest of West Africa and the Sahara. 245 pp. illus., W. and R. Chambers, Ltd. 1937.
THORNTHWAITE, C. W. - The water balance. Publ. in Climatology, Vol. VIII, No. 1, 104 pp. The Drexel Inst. of Technology Laboratory of Climatology, Centerton, N. J. 1955.
WILM, H. G. - The relation of different kinds of plant cover to water yields in semiarid areas. Proceedings, 6th Intern. Grassland Congress, pp. 1046-1050. 1952.
WILM, H. G., and E. G. DUNFORD. Effect of Timber Cutting on Water Available for Stream Flow from a Lodge. pole Pine Forest. U. S. Dept. of Agric. Tech. Bull. 968. 43 pp. illus. 1948.
