section 2 : basic principles

Control of solar radiation in agroforestry practice

W.E. Reifsnyder

Yale University School of Forestry and Environmental Studies
New Haven, Connecticut, USA

Abstract

One of the major controls of microclimate and growth in agroforestry practice is solar radiation. It is also one that is subject to considerable control by man. This chapter reviews basic principles of solar radiation and discusses the ways solar radiation climate in a tree/crop combination can be controlled and/or modified.

These ways relate primarily to the tree component which is the relatively stable and continuing part of the system. Interactions between the trees and solar geometry produce the particular solar climate of a tree/crop system. These interactions and effects include: interception of radiation by tree stands of various densities; effect of canopy structure; effect of row orientation and spacing; effect of latitude and time of year on solar paths; shade from single tree crowns; and spectral quality of sunlight under partial shade.

Implications for agroforestry practice are discussed briefly.

Introduction

In the early days of agriculture, trees must have primarily been a nuisance: clearing away large trees to prepare an area for the planting of a crop was certainly a difficult and never-ending task. However, in some parts of the world where sunshine was abundant, perhaps over-abundant, the value of the shade provided by an overarching tree must have become apparent. And so by repeated experience, systems of food and fiber production were developed that involved some combination of tree and crop (King 1987).

At one time, tobacco grown in Sumatra under partial tree shade was highly prized for its qualities as a wrapper for fine cigars. Connecticut tobacco growers successfully mimicked the shade of the Sumatra trees by covering their fields with acres of cheesecloth, producing the thin, smooth leaf that was similar to the Sumatra leaf. In a study of the microclimate inside a tobacco shade tent, Waggoner, Pack and Reifsnyder (1959) found that solar radiation, evaporation and wind were all reduced substantially as compared with conditions outside. They concluded that the major effect of the tent was in the reduction of solar radiation, mimicking the reduction of solar radiation by the Sumatra trees. Thus an agroforestry system developed indigenously was transferred to another region through artificial control of the microclimate.

A similar use of artificial shade is reported by Stathers and Bailey (1986), in the growing of ginseng (Panax quinquefolium). This perennial herbaceous plant normally grows in the shade of deciduous trees in eastern North America. Commercial growers simulate the radiation climate with wooden lath or woven black polypropylene cloth. Measurements of downcoming solar radiation above and below a canopy of the cloth showed that the radiation was reduced to approximately 25% of the above-canopy value (Figure 1). The absorption of much of the radiation by the canopy produced a double maximum of temperature at the canopy at the canopy and soil surface (Figure 2). Other microclimatic modifications at plant level included a higher average air temperature; higher relative humidity; higher dewpoint temperature, decreased wind speed; and decreased evapotranspiration. These modifications permit successful growing of ginseng in the arid interior of British Columbia, Canada.

Figure 1 Solar radiation above and below a canopy of black polypropylene cloth, after Slathers and Bailey (1986).

Conklin (1953) has pointed out that the Hanunoo of the Philippines practiced a complicated agroforestry system in which clearing for rice cultivation involved the leaving of selected trees to provide a partial canopy of foliage to prevent excessive exposure to sunlight. The trees were not only utilized for their protective value but also for the production of wood, medicines, cosmetics and other products.

Despite the historical and, indeed, obvious role of trees in controlling the radiation climate in agroforestry systems, relatively little specific research has been undertaken, especially in the tropics where agroforestry is often the dominant cropping system. It seems likely that the main reason for this is that research in crop microclimate developed largely in temperate regions where large-scale agriculture has been dominant. In the past hundred years, much has been learned about forest and plant microclimate (Geiger 1957; Yoshino 1975). Much of this information may be transferable to agroforestry systems and practices, at least for a first approximation.

The objective of this chapter is to survey what is known about solar-radiation climate and its conscious control by vegetation manipulation. I hope that this information will suggest ways that knowledge of radiation climate might be used in agroforestry practice; and indicate gaps in our knowledge that can be filled by appropriate research.

Figure 2 Temperature profiles above and below shade canopy, clear summer day (Slathers and Bailey 1986).

Solar radiation

The amount and spectral distribution of sunlight reaching the earth's surface is dependent primarily on 'obstructions' that remove or modify the solar beam in some manner. Thus extraterrestrial sunlight shining through clear sky is reduced in total amount and selectively reduced by various molecules and aerosols (Figure 3; Gates 1980). Most of this selective reduction is in the near-infrared region. Light scattered by 'pure air' molecules is rich in short wavelengths; thus the sky appears blue while the solar beam is depleted in these wavelengths. Light transmitted and scattered by the water droplets in clouds spans the visible spectrum and thus appears white.

By contrast, light under a plant canopy is rich in near- infrared wavelengths (Figure 3). This enhancement of infrared light is the result of the spectral properties of green leaves which transmit and reflect highly in the near-infrared region (Figure 4; Ross 1975). In the far-infrared region (beyond about 3000 nanometers), leaves are nearly complete absorbers, but there is very little of this radiation in sunlight.

Figure 3 Solar irradiance, cloud light, sky light and shortwave radiation beneath vegetative canopy (from Gates 1980).

Figure 4 Shortwave transmissivity, absorptivity and reflectivity of green leaf (Ross 1975).

Radiation profiles in a forest stand

Transmission of solar radiation through a forest canopy depends on stand density; but the relationship is not linear (Reifsnyder and Lull 1965). This relationship can be looked at as function of crown closure or of stem density (Figure 5; Miller 1959). Thus a stand with 50% crown closure will transmit less than 20% of the incident solar radiation. With only 10% crown closure, certainly a rather open stand, radiation reaching the ground is reduced by about 25%. Thus it might be expected that even sparse stands would offer considerable protection from excessive radiation loads. On the other hand, a crown closure of only 1/3, a rather open stand, would reduce the solar radiation beneath by two-thirds, which might result in too little radiation for some crops.

Figure 5 Transmissivity of tree stands as a function of crown closure and stem density (Miller 1959).

However, it should be noted that most of the data on which Figure 5 was based come from mid-latitudes where even at noon, the sun is far from the zenith. I would expect that in the tropics where the sun is close to the zenith at noon, the relationship between transmission and crown closure would be closer to linear. This would especially be true for broad-leaved stands in which the upper canopy is more-or-less like a single layer of horizontal leaves.

The foregoing data are for direct and diffuse solar radiation from a clear sky. On cloudy days, when only diffuse radiation is present at the top of the canopy, transmission is greater than on clear days (Trapp 1938). That is, the diffuse light from the sky can find many more holes to come through than can the direct beam of sunlight. As Figure 6 shows, at every level in crown and trunk space of a dense broadleaf stand, the percentage of transmitted light was greater on an overcast day as compared with a clear day.

Figure 6 Vertical profiles of light in a beech stand on sunny and overcast days (Trapp 1938).

Effect of canopy structure 

The amount and spectral distribution of sunlight reaching the ground beneath a tree canopy depends not only on the crown closure (as indicated in the previous section), but also on the way that the canopy is structured. Young provides an illustrative example of this (Section 1), showing how a 25% tree cover can be arranged in different ways. Thus a plot of land can have the trees in a block, encircling the plot along its boundaries; as wide strips; as more numerous narrow strips (both of which may be oriented differently with respect to the sun's path); as small trees randomly distributed; or as fewer large trees also randomly distributed. Perhaps the most obvious effect of such distributions is in the control of the solar radiation at the ground. Unfortunately, relatively little information is available on these effects; and what there is, is scattered widely in the literature.

Single-layered canopy vs deep canopy

A deep canopy with more-or-less uniform distribution of vegetation elements has often been characterized (with respect to the attenuation of direct-beam radiation) as a uniformly scattering medium (Figure 7). In this case, the solar beam is attenuated according to the negative exponential equation:

I ( below) /I ( above) = exp ( - kx)

where x is the path length through the medium and k is an attenuation coefficient, a constant for a particular medium. If the canopy consists of a single layer of horizontal leaves, the solar beam is not attenuated, and the same proportion of direct-beam sunlight is transmitted at all solar angles:

I (below) /I ( above ) = constant

Figure 7 Direct-beam solar radiation penetrating single-layer canopy (upper diagram) and canopy composed of uniformly-scattering vegetation elements. 

These relationships can be tested in canopies that approximate the two conditions. Reifsnyder, Furnival and Horowitz (1971/72) compared the direct-beam radiation above and below canopies of two kinds: a conifer canopy representing a uniform scattering medium; and a broad-leaved canopy representing a single, horizontal layer of leaves (Figure 8). In the conifer stand, the extinction coefficient should be constant throughout the day (Figure 8a); in the broadleaf stand, the ratio of direct beam radiation below to that above should be constant (Figure 8b). The fall-off of the ratio at the beginning and end of the solar day can be attributed to the presence of tree trunks and other vegetation that destroys the single-layer assumption at low solar angles.

Figure 8 Direct-beam radiation above and below forest canopies of two types: a broadleaf stand (HWD) and a conifer stand (RP); a) extinction coefficient; b) ratio of direct beam radiation below the canopy (CAN TD) to that above (OPEN D). Numbers at each hour indicate the solar altitude at the time of observation (Reifsnyder, Furnival and Horowitz 1971/72).

Clumped vs dispersed crowns

As pointed out by Young (this volume, Section 1), the nature of clumping of tree crowns (few large crowns vs many small crowns) will affect the microclimate beneath. One effect of such clumping will be to control the pattern of light and shade on the ground beneath. Underneath a broadleaf canopy with many scattered vegetation elements and many gaps in the canopy, one should expect many small, short- duration sunflecks. A canopy in which the vegetation elements are clumped, the sunflecks should be larger and illuminate a particular ground spot for a longer time. Measurements in the two stands referred to above confirm this assumption (Reifsnyder and Furnival 1970). Figure 9 shows that there were many more short-duration sunflecks beneath the broad-leaf canopy than under the conifer canopy even though the overall canopy density was similar in the two stands.

How does the sunfleck distribution affect the total direct-beam radiant energy received for a day? At least for the two stands studied, it appears that there is not much difference in the daily totals (Table 1). Furthermore, most of the energy in both stands was in long-duration sunflecks. Very little of the total energy was contained in the more numerous sunflecks under the broadleaf canopy.

Figure 9 Number of sunflecks beneath conifer and hardwood stands, in terms of their duration at fixed sensor location (Reifsnyder and Furnival 1970).

Table 1 Distribution of direct-beam energy in sunflecks, as percentage of total in all sunflecks.

Source: Reifsnyder and Furnival 1970.

Effect of row orientation and spacing

Consider next an agroforestry system in which low-crop areas are separated by rows of trees, with the width of the open space nearly equal to the width of the tree crowns. In an area near the equator, the solar radiation reaching the center of the open row will depend on the orientation of the rows. Figure 10 shows diagrammatically the total solar radiation expected with N-S and E-W row orientations and with a random tree distribution of the same crown density.

With the N-S rows, the center of the open space will receive direct-beam sunlight for a period in the middle of the day, the exact duration depending on the ratio of the tree height to the width of the rows. With E-W rows, the center spot will typically be exposed to nearly full sunlight throughout the day, missing only the skylight obscured by the nearby trees. By contrast, a random orientation of tree crowns will result in an irregular pattern of sunfleck and shade, as shown in the last part of Figure 10.

Of course, rows can be oriented in any direction, can be on a slope of any inclination and azimuth, and can vary in the ratio of strip width to tree height, all of which will change the course of solar radiation received by a particular spot.

Figure 10 Diagrammatic representation of solar flux (F) above (solid line) and below (broken line) three types of canopies, for sun passing close to zenith. Top, tree rows oriented north-south; middle, tree rows oriented east-west; bottom, trees oriented randomly.

Effect of latitude and time of year

Thus the actual pattern of sun and shade will depend on the complex relationships of the solar path to the specific plot. This in turn depends on the latitude and time of year. Sun-path diagrams showing these relationships graphically are available in the Smithsonian Meteorological Tables (List 1958), for latitudes ranging from the equator to the poles. These relationships can be interpreted in terms of the solar radiation reaching the surface.

Harrington (1984) describes a computer algorithm that calculates the solar radiation in an open strip between trees for any strip orientation, slope inclination and azimuth, day of year, and latitude. The algorithm also permits the inclusion of elevation, atmospheric turbidity, cloud amount and cloud type to provide an estimate of daily total solar radiation as a function of position across the strip.

Shade from a single tree crown

Not all agroforestry practices involve strips; some involve growing crops under more-or-less-isolated trees. In contrast to the sunfleck that courses over a shaded floor, the isolated tree casts a shadow that moves over a sunny ground. Halverson and Smith (1974) list a FORTRAN program that will calculate the length of the shadow cast by a tree on any slope and azimuth. Quesada, Somarriba and Vargas (this volume, Section 2) describe a program that will plot the distribution of shadows from a specified plot of trees.

Figure 11 (Grant 1985) illustrates the radiance distribution of a clear sky with the shadow of a spherical crown superimposed. With a clear sky, the maximum sky radiance is centered on the sun; the minimum at right angles to the sun. Thus a tree blocks the brightest part of the sky. The percentage of blocked sky radiation (i.e., not including direct-beam radiation) is shown in Figure 12 for a range of zenith angles. Here, X/S is the relative distance from the center of the shadow to its edge. Although the overhead sun is not shown, it would be close to the top three curves. The percent reduction at the center of the shadow ranges from about ten percent for the shadow cast by a low sun, to about 35% for the sun near the zenith.

With a cloudy sky, the reductions are less than ten percent because of the nearly uniform radiance of the overcast sky.

Figure 11 Simulated clear-sky radiance with superimposed shadow from an isolated spherical-crowned tree (Grant 1985).

Figure 12 Reduction of flux density in shadow of crown, for various solar zenith angles (z), as a function of the ratio of distance from center of shadow (X) to the half-width (S) of the shadow (Grant 1985).

Spectral quality of sunlight under partial shade

Shortwave radiation reaching the ground surface beneath a partial canopy consists of at least four components: direct beam radiation coming through gaps in the canopy; diffuse radiation from the reflection and transmission of the direct beam by leaves and other vegetation elements; sky radiation transmitted through canopy gaps; and radiation reflected off vegetation elements (Figure 13; Reifsnyder, Furnival and Horowitz 1971/72). In this scheme, multiple reflections between canopy and ground surface are ignored; for most canopies these components are small.

In addition to reducing the amount of energy received by a sunny spot, the canopy changes the spectral distribution in sunfleck and shade because of the unique spectral properties of vegetation. In full shade under a broadleaf canopy, light is enriched in the near infrared regions (because of the high reflectivity of leaves in this spectral region); and diminished in the blue and red regions, as compared with green wavelengths (Figure 14; Horowitz 1975). Light in sunflecks is enriched in the red, and reduced slightly in the blue region of the spectrum, relative to green wavelengths.

Figure 13 Model of penetration of solar radiation through a canopy. OPEN TOT is the total solar irradiance at top of canopy, consisting partly of direct beam radiation (OPEN D) and partly of diffuse radiation (OPEN I); CAN PD is the penetrating direct-beam component; CAN RD is the component reflected and transmitted by leaves; CAN PI is the diffuse radiation penetrating canopy gaps; and CAN RI is the diffuse radiation reflected and transmitted by leaves (Reifsnyder, Furnival and Horowitz 1971/72).

Figure 14 Spectral distribution of light above a temperate-zone broadleaf forest, in the center of a sunfleck, and in full shade beyond the edge of the sunfleck (Horowitz 1975). Canopy in full leaf, clear sky.

Of course, sunflecks do not have a hard edge, largely because the sun is not a point source of radiation. There is thus a penumbra effect which produces not only a gradual change in the energy from the center of the fleck to full shade; but also a change in the spectral distribution of that energy. Figure 15 (Horowitz 1975) shows the change in canopy transmissivity of green light (550 nm) and red light (675 nm) from the center of a large sunfleck to full shade (indicated by the time elapsed as the sunfleck coursed over the radiometer). Although the curves are not smooth because the canopy gap producing the sunfleck had various vegetation elements in it, the trend is clearly from a red-rich center to a green-rich shade. The red/far-red ratio beneath the canopy also changes from relatively high values in the sunfleck to low values in full shade.

Figure 15 Transmissivity (calculated as the ratio of light of indicated wavelength incident above stand to that at ground level, as function of distance from center of sunfleck; and the red/far-red ratio at ground level, also as function of distance from sun-fleck center (Horowitz 1975).

Implications for agroforestry

It must be admitted that there is very little 'hard' information on control of solar radiation in agroforestry practice. Most of the available information concerns the extensive field or extensive forest; edge and transition conditions have been little studied. Yet these transition zones are at the heart of much agroforestry practice, certainly in alley-cropping and similar methods.

Because of the geometric niceties of solar radiation, it is relatively easy (especially in these days of microcomputers) to estimate radiation conditions resulting from various cultural practices. Predicting the effects of such radiation control on microclimate and consequent crop (and tree) growth is nevertheless still problematic. The bottom line, of course, is how well do plants A and B grow under particular radiation regimes and the microclimate produced by the plant configuration. We might try to perform field trials using all possible combinations, but such would soon become prohibitively expensive and difficult if not impossible to interpret.

An alternative might be to determine the radiation requirements of plants currently used in various agroforestry systems. Then with our knowledge of feasible radiation manipulation, we could plan field trials to test what seem to be the most likely combinations. Of course, radiation manipulation is only part of this: microclimate and soil moisture, as well as various cultural practices must be included in the experimental design.

It may also be necessary and appropriate to look at the problem in the reverse sense: given a radiation regime proscribed by solar geometry and certain cultural practices that may be difficult to modify drastically, what plant combinations are likely to succeed?

In both situations, the information most likely to be missing is that relating to specific radiation requirements of the plants. A first step might be to assemble, analyze and interpret known information on radiation requirements of plants currently used in agroforestry systems. This would point out the gaps in our knowledge and would form the basis for appropriate growth chamber and field experimentation.

In the meantime, agroforestry field experiments should certainly include careful consideration of the radiation climate — both by calculation and, where possible, by concomitant radiation measurements and visual observations.

References

Conklin, H.C. 1957. Hanunoo agriculture. Rome: United Nations Food and Agriculture Organization. Quoted in King 1987, p. 3.

Gates, D.M. 1980. Biophysical ecology. New York: Springer-Verlag.

Geiger, R. 1965. The climate near the ground. Cambridge: Harvard University Press.

Grant, R.H. 1985. The influence of the sky radiance distribution on the flux density in the shadow of a tree crown. Agric. For. Met. 35: 59-70.

Halverson, H.G. and J.L. Smith. 1974. Controlling solar light and heat in a forest by managing shadow sources. USD A Forest Service Research Paper PS W-102.

Harrington, J.B. 1984. Solar radiation in a clear-cut strip - a computer algorithm. Agric.. For. Meteorol. 33: 23-40.

Horowitz, J.L. 1975. The quantity and quality of incident light in forest sunflecks. PhD diss., Yale University, New Haven, Connecticut.

King, K.F.S. 1987. The history of agroforestry. In H. A. Steppler and P.K.R. Nair (eds.) Agroforestry, a decade of development. Nairobi: International Council for Research in Agroforestry.

List, RJ. (ed.) 1958. Smithsonian meteorological tables. Washington, DC: Smithsonian Institution.

Miller, D.H. 1959. Transmission of insolation through pine forest canopy as it affects the melting of snow. Schweiz. Anst. f. forstl. Versuchsw. Mitt. 35: 57-79.

Reifsnyder, W.E. and H.W. Lull. 1965. Radiant energy in relation to forests. USDA Forest Service Technical Bulletin 1344. Reprinted by AMS Press, New York.

Reifsnyder, W.E. and G.M. Furnival. 1970. Power-spectrum analysis of the energy contained in sunflecks. In Proc. Third Forest Microclimate Symposium, Kananaskis Forest Experiment Station, Seebe, Alberta, 23-26 September 1969. Canadian Forestry Service.

Reifsnyder, W.E., G.M. Furnival and J.L. Horowitz. 1971/72. Spatial and temporal distribution of solar radiation beneath forest canopies-Agric. Meteorol. 9: 21-37.

Ross, J. 1975. Radiative transfer in plant communities. In J.L. Monteith (ed.) Vegetation and the atmosphere. Vol. 1. New York: Academic Press.

Stathers, RJ. and W.G. Bailey. 1986. Energy receipt and partitioning in a ginseng shade canopy and mulch environment. Agric. For. Meteorol. 37:1-14.

Trapp, E. 1938. Untersuchung iiber die Verteilung der Helligkeit in eimem Buchen-bestand. Biokl. B. 5:153-158.

Waggoner, P.E., A.B. Pack and W.E. Reifsnyder. 1959. The climate of shade. A tobacco tent and a forest stand compared to open fields. Connecticut Agric. Exp. Sta. Bull. 626.