Canopy Architecture, Light Extinction and Self-shading of a Prairie Grass, Andropogon gerardii.
ABSTRACT.--The three-dimensional characteristics of canopy structure and light environment of clones of a common tallgrass prairie species, Andropogon gerardii (big bluestem), were analyzed in native and reconstructed prairies near Ames, Iowa to determine if clones are limited in size by the effects of self-shading. Clones tended to be nearly circular, with a mean circumference of 158 cm and a mean cross-sectional area of 2060 [cm.sup.2]; clone height was consistently near 60 cm in early summer. Irradiance at a given height in the canopy, calculated as a fraction of radiation above the canopy, decreased rapidly during May and June as plants completed most of their seasonal growth, but was relatively constant in July and August. A vertical gradient in light dominated the spatial variation in light within clones. Horizontally through a clone, light levels were higher around the southern (sun-side) periphery. Spatial variation in light, in terms of the range of values found, was least at the bottom and upper pa rts of the canopy and greatest in the middle of the canopy. Leaf mass/area and total leaf area per increment of canopy height were highest at 30 cm below the top of the canopy and were lower at 15 cm and 45 cm below the top of the canopy. Mean leaf area index of clones was 6.58. Diameter of clones had a weak effect, if any, on the light environment within a clone. Only clones with diameters of [less than]20-30 cm are likely to have substantially less self-shading than larger clones. Competition with other plants or limitation by other morphological constraints, rather than limitation by self-shading, may control the upper size limit of clones.
Within vegetative canopies the amount of sunlight decreases from the top of the canopy down, due to absorption by leaves and other plant parts. Different types of canopies create different gradients of light extinction, and the growth form of each species in the canopy affects how much light it may receive (Mitchley and Willems, 1995). When a canopy contains several species the tall dominant species may absorb almost 75% of the light entering the canopy, leaving only 25% for the subordinate species (Hirose and Werger, 1995). In monospecific canopies growth form of the species determines the degree of self-shading, i.e., the arrangement of leaves and their spatial display determines where light is intercepted and how much penetrates to the base of a plant (Warren-Wilson, 1965). For plants that expand by the clonal production of multiple stems, as in the clone-forming grasses found in tallgrass prairies of the central United States, the form of a clone may also affect the light environment within the clone, i. e., degree of self-shading potentially may be a function of clone size and shape. Self-shading may have significant effects on plant growth and architecture (Ackerly and Bazzaz, 1995).
The amount of light available at a given point in the canopy often changes the way a plant grows and its physiological characteristics, and vice versa (Bjorkman, 1981; Chabot et al., 1979). The plant may change its leaf mass per leaf area, total biomass and leaf area, potential photosynthetic rate, etc., in different spatial positions in its canopy to compensate for the different amounts of light received at those positions. Patterns of leaf distribution and corresponding light absorption vary widely among different types of canopies. As examples of two extremes, in mixed-grass prairie canopies leaf biomass tends to be much greater in the lower parts of the canopy (0-15 cm aboveground) than in the upper parts of the canopy (30-45 cm aboveground) (Conant and Risser, 1974), whereas leaves in tree canopies tend to be found relatively higher in the canopy (Jurik et al., 1985). Physiological characteristics in turn vary through the canopy. For example, nitrogen content per unit leaf area, an important factor in p hotosynthetic rate, is typically higher in the upper layer of canopies where most of the light is absorbed and lower in the bottom layers where very little light penetrates (Hirose and Werger, 1987). This distribution allows a plant to increase its total, whole-canopy, carbon gain (Hirose and Werger, 1987). In species that largely create their own effective canopy environment, such as in grasses, canopy structure and physiological attributes may thus have a significant interaction with whole-plant carbon gain and may set limits on the size and shape of clones.
Most previous studies of availability of light in different parts of canopies have examined spatial variation in one dimension, i.e., vertical gradients (e.g., Baldocchi et al, 1984, 1986; Fliervoet and Werger, 1984; Werger et al., 1986; Turner and Knapp, 1996; and many others). A few studies have examined horizontal variation in light at one vertical position (e.g., Smith et al., 1992; Clark et al., 1996; Nicotra et al., 1999). However, there is less published information on three-dimensional spatial variation in canopies (but see Yoda, 1974; Cladwell et al., 1983; Canham et al., 1994; Ryel et al., 1994; Pearcy and Yang, 1996). One objective of this study was to characterize the seasonal time course of light at different levels in a tallgrass prairie canopy. A second objective was to characterize the three-dimensional spatial structure and light environment within clones of a dominant grass species of the tallgrass prairie and to examine the potential influences of the size and shape of a clone on its light environment. The dominant grasses in tallgrass prairies characteristically have canopies from 0.6 to 1.5 m tall (Great Plains Flora Association, 1986), but we know of no information on the range of clone diameters that can be achieved in a tallgrass prairie and the effects of different diameters on plant characteristics. By examining how light and plant characteristics vary in a three-dimensional structure we hoped to determine if self-shading limits size of clones.
Field sites.--The seasonal change in light availability at different heights aboveground was studied in 1987 in a 20+ y-old reconstructed prairie at Iowa State University's Curtiss Farm, near Ames, Iowa. A portion of the site dominated by Andropogon gerardii Vitman (big bluestem) was used.
Clone size and spatial variation in light environment were studied in June and July 1998. Big bluestem plants of unknown ages and various sizes in a small native prairie, the Richard Pohl Preserve, near Ames High School in Ames, and a 25+ y-old reconstructed prairie in McFarland Park, 8 km northeast of Ames, were used. For measurements of light and biomass (see below), clones in one of three size classes (35, 50 or 60 cm diameter), encompassing most of the range of clone sizes found in the field, were arbitrarily chosen. These clones had discrete boundaries, but were typically surrounded by moderately dense vegetation that was approximately 30-60 cm tall.
Size distribution.--The cross-sectional size distribution of clones was characterized at the McFarland Park prairie in 1998. Clones were sampled along a 50 m north-south transect. All identifiable discrete clones within 5 m of the transect were measured. The diameters of the clone along north/south and east/west axes were measured at 30 cm aboveground, which was half the height of most clones. This height matched the middle of our range of heights sampled (see below) and was representative of most of the canopy, since in early summer leaves and tillers were quite erect with little of the splaying outward that occurs in late summer. For analyses of horizontal size, clones were idealized as ellipses. The cross-sectional area of each clone was calculated as area = [pi].a.b, where a is the radius of the longest axis of the clone and b is the radius of the shorter axis. Circumference was calculated using the formula circumference = [pi](a + b).(1 + [m.sup.2]/4), where m = (a - b)/(a + b). Eccentricity of the elli pse was calculated as eccentricity c/a, where c is half the distance between the foci of the ellipse and [a.sup.2] = [b.sup.2] + [c.sup.2]. Eccentricity is the departure from circularity and ranges from 0 (=a circle) to 1 (a straight line).
Light extinction.--In the 1987 seasonal study of light, six sets of United Detector Technology F004E photodiodes (with each set having sensors at 3, 15, 30, 60 or 90 cm aboveground) were spaced at 0.5 m intervals through the canopy, irrespective of individual clonal boundaries. The photodiodes sensed visible light plus a fraction of the infrared range (400-900 nm); effects of this wavelength response are discussed by Jurik and Pleasants (1990). Essentially, the effect is conservative, such that differences among canopy positions are not as large as they would be using sensors that measure only visible light (400-700 nm). Readings were recorded once a minute for 1 to 10 d using a Campbell Scientific 21X micrologger. Total daily irradiance for each photodiode was integrated over each day and divided by the integrated value for sensors above the canopy to produce daily irradiance expressed as a fraction of light incident above the canopy. Values for different days and all sensors at a given height were then ave raged. Canopy height was measured to the tops of the leaves, disregarding flowering stalks late in summer.
To quantify the three-dimensional structure of light extinction through a canopy, photosynthetically active radiation (400-700 nm) was measured in late June and July 1998 at the Pohl Preserve prairie and the McFarland Park prairie. A Campbell Scientific 21x micrologger recorded photon flux density values from two quantum sensors. One of the quantum sensors was placed on a tripod 1.5 m to the side of the target clone to monitor incident sunlight above the canopy. The other sensor was placed on the end of a thin rod that was probed into a clone. A reading was taken every 5 cm along a horizontal transect through the canopy, starting just outside the clone and ending just outside the clone on the opposite side. Each reading was the average of nine measurements taken at 0.3 s intervals by the micrologger. Readings were taken along transects running north/south, east/west, northeast/southwest and northwest/southeast at each of three distances below the top of the canopy (15, 30 and 45 cm). Clones with average diam eters of 35 (n = 3), 50 (n = 4) and 60 (n = 10) cm were sampled; all clones were approximately 60 cm tall. Samples were collected between approximately 9 AM and noon (local solar time). Light at any point in the canopy (I) was divided by the value above the canopy ([I.sub.o]) for that measurement interval to obtain light as a fraction of incident radiation (I/[I.sub.o]).
Leaf mass/area.--To determine how leaf characteristics varied with light environment in 1998, samples for determining leaf mass/leaf area (LMA, g/[m.sup.2]) were collected in a spatial pattern similar to that used for the light extinction measurements. Leaves were collected from two 50 cm-diameter clones, one on each site, every 10 cm on north/south and east/ west transects through a clone. In the laboratory the area of 5-cm long segments from portions of the leaf corresponding to 13-18 cm, 28-33 cm and 43-48 cm below the top of the canopy were measured with a LICOR 3100 leaf area meter. The samples were then dried at 65 C for 24-48 h and weighed.
Leaf area index.--To measure the spatial density of grass blades in clones a 5 cm X 10 cm rectangular sample frame with three fixed sides and one removable side was slid into the canopy and rotated to a horizontal position, with the long axis perpendicular to a horizontal transect through the canopy. The number of leaves was counted inside the rectangle; sampling was repeated every 5 cm along the transect through the canopy. Measurements were taken on each of four transects at different angles through the canopy, as described above, at 15, 30 and 45 cm below the top of the canopy. In each rectangle, width of one leaf also was measured. Leaf area was measured on two clones in the McFarland Park prairie and one clone in the Pohl Preserve prairie. All clones were approximately 50 cm in diameter. To calculate leaf area index (LAI, leaf area per unit ground area) for a given height in the canopy, the number of leaves in each sample rectangle was multiplied by mean width of the leaf blades times a 1-cm height incr ement, to give leaf area per 50 [cm.sup.3]. The mean blade width for a sample rectangle was calculated by averaging the value recorded in a rectangle with the values from adjacent rectangles. Total LAI for a clone was calculated by first interpolating and then averaging all the values at a given height in the canopy (see below), to obtain a mean value (LAI per 1 cm height increment) at that height. Those values were then multiplied by a height increment (20, 15 or 15 cm, for the 15, 30 and 45 cm below top levels, respectively) and summed to give total unifacial canopy leaf area. This procedure assumed that the bottom of the canopy was at approximately 12 to 15 cm aboveground, which was the point where leaves unfolded from the leaf bases, and that the uppermost 15 cm of the canopy contained 20 cm lengths of leaves, since the upper ends of leaves tended to be less vertical than parts lower in the canopy. Values of LAI for different height increments were averaged over the three sample plants, for comparisons wi th light extinction values.
Contour calculations.--Clones were modeled as a 60-cm-tall cylinder with a circular cross-sectional shape and a diameter of 35, 50 or 60 cm, with measurements taken 15, 30 and 45 cm below the top of the canopy. For each height in the canopy, spatial position and values of light, LMA and leaf area were digitized along each sampling axis using a model diagram of the axes. Each diagram represented a 35, 50 or 60 cm diameter circle with scaled hatchmarks. The diagrams were placed onto a Summasketch digitizer tablet connected to a microcomputer. Solar altitude and azimuth were calculated based on the date and average time of measurement for each set of measurements (Gates, 1980). Diagrams for LMA and leaf area were oriented simply by compass direction. Diagrams of light extinction were adjusted so that the sun always appeared to be due south, at the bottom of the diagram. This procedure resulted in the original sample axes (N/S, E/W, NE/SW and NW/SE) for different plants all having somewhat different orientations , since solar azimuths at the times of sampling different plants varied. Sample data sets thus could be merged with solar azimuth held constant, although solar altitude varied from 46[degrees] to 68[degrees] among clones. Such merged data sets for light extinction consisted of 200-500 original data points with a reasonably uniform distribution over the cross-sectional area of a clone, although there were more data points near the center because of the radial arrangement of the sample axes.
For the digitized measurements of light and leaf area, three-dimensional contours were produced using Surfer software (Golden Software, Inc.), using merged data sets (i.e., original data for all clones of a given diameter combined, as described above). Surfer first created a uniformly spaced grid of data points by interpolation based on the sample data; subsequent contour calculations were based on the grid values. Interpolation was by kriging, with a nugget effect (Cressie, 1991) and a quadrant search using 10 values, i.e., the 10 values closest to a grid point in each of the four quadrants around that location were weighted by distance and combined to create the grid point value (see Surfer documentation). Mean values of a variable at a given height in the canopy were then calculated using integration techniques in Surfer, i.e., the integrated volume under the contours was derived. That volume divided by the cross-sectional area of the clone gave the spatially-weighted mean value of the variable for that h eight in the canopy. By choosing different base levels for the volume integration, the fraction of the clone cross-sectional area with a given range of values of the variable of interest could also be calculated; these fractions were then used to create histograms showing fraction of the clone with a given range of I/[I.sub.o], for example.
Statistical analyses.--Possible horizontal gradients in I/[I.sub.o] for each height in the canopy for each size clone were determined by trend surface analysis (Swan and Sandilands, 1995), in which a multiple linear regression model was fitted to each original data set (not the interpolated Surfer grid surface). The cubic regression model:
Z = [b.sub.1]X + [b.sub.2]Y + [b.sub.3][X.sup.2] + [b.sub.4][Y.sup.2] + [b.sub.5]XY + [b.sub.6][X.sup.3] + [b.sub.7][Y.sup.3] + [b.sub.8][X.sup.2]Y + [b.sub.9]X[Y.sup.2] + [epsilon]
was calculated by SAS (SAS Institute Inc., 1996); components of the model were required to have P [less than] 0.15 to be included. In this procedure, the Z variable (I/[I.sub.o]) was analyzed as a function of horizontal position, i.e., the XY coordinates of the data point. For our analyses, the origin (X = 0, Y = 0) of the coordinate system was taken to be to the southwest of a clone, where clones were plotted with north at the top, west at the left, etc. Thus, a regression showing a significant effect of Y with a negative coefficient, for example, would mean that I/[I.sub.o] decreased from south to north through the clone. A model with quadratic terms allows a surface to be curved, with one sense of curvature (i.e., allows one peak or one valley), whereas a model with cubic terms allows for one change in the sense of curvature (i.e., accounts for a peak and a valley).
Following the trend surface analyses, which typically indicated changes in I/[I.sub.o] in a north-south direction through each clone, subsets of the data for each clone were analyzed in one dimension to estimate rates of change in I/[I.sub.o] with horizontal position for the central portions of the north and south halves of each cross-section. Data were restricted to a 20 cm-wide strip running north-south through the center of each cross-section. Regressions of I/[I.sub.o] vs. horizontal distance were calculated for data from the south edge to the center and for the center to the north edge of each cross-section.
A regression of spatially-averaged I/[I.sub.o] vs. vertical distance through the canopy was calculated for each size clone using simple linear regression with the model forced through a point of I/[I.sub.o] = 1.0 at 0 distance into the canopy. Slopes of the regressions for different size clones were then compared using the General Linear Models procedure of SAS (SAS Institute Inc., 1996) by testing whether clone diameter, used as a classification variable, had a significant effect ([alpha] [less than] 0.05) on an analysis conducted with data for all clone sizes combined.
Size distribution.--In cross-section, clones were slightly elliptical with a mean major radius of 27.2 cm (SD = 6.6, n = 192) and a mean minor radius of 23.0 cm (SD = 5.6, n = 192). Mean eccentricity was 0.44 (SD = 0.26, n = 192). Thus, most clones sampled were approximately circular, with the major axis only 18% longer than the minor axis. The distribution of circumferences was approximately normal, ranging from 79 cm to 291 cm (data not shown). The mean circumference was 158 cm (SD = 36.4, n = 192). The distribution of cross-sectional areas was also approximately normal but apparently slightly skewed to the right, with areas ranging from 490 [cm.sup.2] to 6680 [cm.sup.2] (Fig. 1). The mean area was 2060 [cm.sup.2] (SD = 944, n = 192). The clone diameter categories of 35, 50 and 60 cm used for our studies of light, leaf area and mass had areas of 960, 1960 and 2830 [cm.sup.2], respectively, and thus spanned a range of areas that included about 80% of the clones found in the field; the 50 cm diameter categor y was very near the mean area of clones in the field.
Light extinction.--In 1987 the canopy grew rapidly from early May through late June (Fig. 2); growth then slowed, although height continued to increase until August. Fraction of incident light, calculated as a fraction of the daily total received, decreased rapidly during May and early June at 3 cm aboveground, but then decreased relatively little over the rest of the growing season. Values for other heights had similar patterns that were initially offset in time, depending on when the top of the canopy reached a given height (Fig. 2).
Spatial patterns of I/[I.sub.o] were generally the same for the three classes of clone diameters, with distance below the top of the canopy having the greatest effect, i.e., a vertical gradient being the most pronounced (Fig. 3). Calculation of the fraction of the cross-sectional area of a clone with a given value of I/[I.sub.o] showed little difference among size classes (Fig. 4). At 15 cm below the top I/[I.sub.o] ranged from 0.4 to 0.9 (Figs. 3, 4). At 30 cm below the top values were reduced; I/[I.sub.o] ranged from 0.15 to 0.85. At 45 cm below the top values of I/[I.sub.o] ranged from 0.01 to 0.35. The range of values of I/[I.sub.o] was greatest at 30 cm below the top of the canopy for all diameter classes (Fig. 4); values at the top or bottom of the canopy were less variable. As with the ranges of values, the spatially-weighted means of I/[I.sub.o] for different diameter classes at a given height in the canopy were very similar (Fig. 4). For all diameter classes combined the spatially-weighted means of I/[I.sub.o] for 15, 30 and 45 cm below the top of the canopy were 0.68, 0.33 and 0.12, respectively. A regression of spatially-averaged I/[I.sub.o] vs. vertical position in the canopy revealed a practically linear decline with distance from the top of the canopy (Fig. 5). The three diameter sizes had essentially identical patterns; slopes of the regressions were not significantly different (Fig. 5). Averaged over all three diameter classes the mean slope was --0.0199, i.e., a 10 cm vertical distance through the canopy resulted in an average decrease in I/[I.sub.o] of 0.199.
Besides the striking vertical gradients in I/[I.sub.o], there were horizontal ones as well. For all clone diameter classes the southern sides (Fig. 3, bottoms of the plots) at 30 and 45 cm below the top of the canopy had higher I/[I.sub.o], reflecting the standardization of solar azimuths to 180[degrees] (due south) and consequent greatest direct solar radiation from that direction, given that solar altitude ranged from 46[degrees] to 68[degrees]. Trend surface analysis (Table 1) showed significant horizontal trends in I/[I.sub.o] for all heights and clone diameters except for 15 cm below the top in the 60 cm diameter clone. In most cases, the trends were dominated by the Y coordinate which always had a negative coefficient, indicating a general decrease in I/[I.sub.o] from south to north. For 30 and 45 cm below the top in the 50 and 60 cm diameter clones, quadratic and cubic terms were also significant (Table 1), indicating changes in direction of curvature, i.e., a complex pattern of change in I/[I.sub.o]. Patterns at 15 cm below the top of the canopy in the 50 and 60 cm diameter clones generally were more complex than at other heights and could not be adequately described even by quadratic or cubic regression models, perhaps indicating a more dynamic light environment created by greater patchiness and movement of leaves. For all clones light levels were generally higher around the periphery particularly on the southern side. This resulted in the darkest areas being in the center and in the side away from the sun (Fig. 3). For the southern and northern halves of each cross-section considered separately, regressions of I/[I.sub.o] vs. horizontal distance along a south-north band through the central 20 cm of each clone showed differences in the rate of change of I/[I.sub.o] (Table 2). There was a significant positive trend (i.e., slope for I/[I.sub.o] greater than zero) for the northern half only for the 30 cm below the top cross-section of the 60 cm diameter clone (Table 2). In contrast, there were significant negative slopes (i.e., decreases in I/[I.sub.o] with northward distance) for all southern halves of cross-sections except for the 15 cm below the top cross-section of the 35 cm diameter clone. For the eight southern halves with significant slopes, the mean slope was -0.0125, i.e., a 10 cm horizontal distance through the canopy resulted in an average decrease in I/[I.sub.o] of 0.125.
Leaf mass/area.--Values of LMA. (spatial patterns not shown) at 15 cm below the top of the canopy ranged from 46 to 69 g/[m.sup.2]. At 30 cm below the top there was more variation from point to point, with values ranging from 75 g/[m.sup.2] in a few patches up to 183 g/[m.sup.2], although most of the values were between 90 g/[m.sup.2] and 135 g/[m.sup.2]. The pattern for 45 cm below the top was intermediate between 15 and 30 cm below the top, with respect to both absolute values and general uniformity. Mean values for 15, 30 and 45 cm below the canopy top were 55, 122 and 69 g/[m.sup.2], respectively.
Leaf area index.--Mean cross-sectional leaf area index ([cm.sup.2]/[cm.sup.2] per cm vertical increment of the canopy) was lowest (0.08) at 15 cm below the top of the canopy, highest (0.23) at 30 cm below the top and intermediate (0.15) at 45 cm below the top (Fig. 6). There were no clear horizontal trends in distribution of leaf area across a cross-section (Fig. 6).
Total leaf area index per clone was calculated for each of the three sample clones by multiplying the LAI per cm height values by a height increment (see Methods); values of LAI ranged from 4.5 to 10.6, with a mean of 6.58. By multiplying LAI by cross-sectional area or LMA, one can calculate that a "typical" clone of average cross-sectional size (0.206 [m.sup.2]) would have a total leaf area of 1.36 [m.sup.2] and a total leaf mass of 620 g per clone.
Since light absorption by leaves is a primary determinant of patterns of light extinction through a canopy, I/[I.sub.o] was plotted against cumulative LAI above a given point (data not shown). The relationship was nearly linear, with some slight concavity in the shape. Using the model light extinction formula of Monsi and Saeki (1953) of I = [I.sub.o][e.sup.-kLAI], the extinction coefficient (k) for the canopy was 0.28.
Approximately 85% of the clones of big bluestem were between 97 and 202 cm in circumference and had areas between 750 and 3250 [cm.sup.2] (Fig. 1). Clones larger than 4250 [cm.sup.2] were very rare. Our sampling method probably underestimated the number of young small clones ([less than]10 cm diameter, or 80 [cm.sup.2] area) that could have been easily overlooked. While it is possible that the clones in the McFarland Park prairie are simply not old enough to be larger, our casual observations suggested they were consistent in size with clones in native prairies. Our range of diameter classes (35-60 cm) used for light measurements and modelling of light extinction corresponded to circumferences of 110-190 cm, or areas of 960-2830 [cm.sup.2], and thus encompassed the great majority of sizes encountered in the field (cf., Fig. 1).
Leaf mass/area was highest in the middle of the canopy and lower at the top and bottom, thus showing no consistent relationship to light measured at the time of this study. Previous studies have found LMA to be related to integrated light energy on a leaf's surface during leaf development (e.g. Chabot et al., 1979). Since grass blades have basal meristems and produce new tissue at the base of the leaf, lower values for the shaded bottom part of the canopy are consistent with the previous generalization, but the lower values at the top of the canopy are not. Since the leaf tissue at the top of the canopy was produced at a different time of year (earlier), factors other than light may have affected LMA values.
Our leaf area measurements primarily revealed leaf density in the canopy, or how many leaves were in a given area, since width did not vary much at a given height in the canopy. The canopy was most dense in the middle section (30 cm below top) and less dense on the upper and lower ends (Fig. 6). Horizontally, distributions of leaf area were quite variable, perhaps reflecting the difficulties inherent in measuring positions of leaves that literally swayed in the wind, although even cross-sections low in the canopy where leaves moved less also showed considerable spatial variation. However, our sample size was relatively small. Thus no generalizations about the horizontal spatial pattern of leaf distribution are possible beyond the observation that distributions were not uniform.
Detailed models of light penetration through canopies typically distinguish diffuse and direct beam radiation, which may have different rates of absorption by the canopy. Furthermore, temporal variation in the light environment, on various time scales ranging from seconds to seasons, is often very important in plant growth. Also, the amount of light absorbed by photosynthetic surfaces at any point in a canopy is highly dependent on the angle of the surface relative to the direction(s) of the rays of light. All these factors may have significant effects on the response of a plant, or parts of a plant, to its local light environment. However, since our purpose was primarily to document spatial variation in light within a clone and to compare clones rather than to explore fully all the nuances of the light environment, we simply used fraction of visible light above the canopy as a general indicator of the light environment at a given point in the canopy. Our measurements of spatial variation in light were taken at times of day (9 AM to noon) with the highest solar elevations (note that patterns for noon to 3 PM should mirror those we measured). One might expect less variation both within a clone and among clones with lower solar elevations earlier or later in the day, when shading by surrounding vegetation would be greater. The 1987 study of seasonal change in light environment indicates that the light environment, as measured by integrated daily I/[I.sub.o], at a given height above ground changes rapidly during May and early June, but is relatively constant over the remainder of the growing season. Thus, our spatially-oriented measurements recorded for clones in 1998 should be representative of the light environment for much of late June, July and August.
One might expect the light environment within clones early or late in the day, with low solar altitudes, to differ from that at midday due to different angles between leaves and the solar beam plus greater degrees of shading both within a clone and by surrounding vegetation, etc. However, the similarity of our values of daily integrated I/[I.sub.o] at a given height in 1987 to the instantaneous values measured at midday in 1998 (cf., Figs. 2, 4) indicate that light levels early and late in the day are low and constitute a small fraction of the total daily photon flux. Any potential differences among clones in degree of selfshading at those times of day are not likely to create substantial differences in carbon gain among clones of different sizes because of the overall low photon flux densities, although detailed information on rates of photosynthesis at different light levels, temperatures, etc., is needed to determine the actual magnitude of any such effects. Shading by neighboring plants also becomes a la rger factor at low solar altitudes, and such shading is likely independent of clone size.
Vertical light extinction coefficients based on leaf area index of plant canopies generally range from 0.3 to 1.5, with those less than 1.0 characteristic of nonhorizontal, clumped leaf canopies, and those greater than 1.0 characteristic of horizontal leaves, for high solar altitudes (Jones, 1983). Our value for big bluestem, with mean solar altitude of approximately 55[degrees] was 0.28, or very close to the lowest value of 0.3 given by Jones. Low values imply vertical orientation of leaves, which certainly was the case for big bluestem. While we made no specific measurements, practically all leaves appeared to be within 15[degrees] of vertical. Such a low extinction coefficient for high solar altitudes implies that light penetrates a long way down into a canopy. In our canopies, the low value of the extinction coefficient probably was primarily due to verticality of leaves with most of the light reaching the bottom of the canopy having passed through most of the canopy above that point, rather than penetra ting from the side of the clone. Our measurements of light extinction as a function of distance through the canopy provide quantitative support for this conclusion. There was no consistent horizontal gradient in I/[I.sub.o] for the north half of clones, and the largest horizontal gradients in light, which occurred on the southern sides of clones, had slopes (mean of -0.0125) only about 2/3 the magnitude of the mean slope (-0.0199) of the vertical gradient averaged over the entire area of a clone.
Much of the light that enters a big bluestem canopy is absorbed by leaves and other plant parts fairly quickly. Light, in prairie canopies, is a discontinuous resource (Knapp et al., 1998). For all our clone diameters about two-thirds of the light was absorbed by the upper half of the canopy (in terms of distance). Net photosynthesis is highly dependent on the light environment. Photosynthetic saturation of big bluestem often occurs at about half of full sunlight, so even the upper middle layers of the canopy can still be maximally photo-synthetically productive (Schimel et al., 1991). In the shaded lower canopy leaves net photosynthesis is less than 30% of that in the upper canopy leaves (Schimel et al., 1991). For lower parts of a canopy, proximity to the periphery of a clone might have an appreciable effect on photosynthetic rate by increasing the amount of light. Thus, depending on the proportion of the leaves in the interior of a clone vs. near the periphery, and depending on the proportion of photosynt hesis that occurs in the upper part of a canopy vs. the lower part, a clone hypothetically could be limited to a size where the upper canopy's net photosynthesis can support the rest of the clone. One might expect smaller diameter clones to have a greater proportion of their leaves near the perimeter of the clone and thus to have more light, if light penetration from the side of the canopy has an appreciable effect on the light environment in a
clone. In contrast, larger clones would have proportionally more of their leaves in the interior of a clone. Contrary to these expectations, our spatially averaged values of light availability at the different levels in a canopy showed no significant effect of canopy diameter, for clones 35 to 60 cm in diameter. Thus, differential self-shading by clones, in terms of creating different horizontal gradients in light for different size clones, was not apparent. At least for high solar altitudes, the low vertical light extinction coefficient of clones means that an abundan ce of light penetrating from the top, rather than the sides, of a clone may swamp any differential horizontal gradients in different size clones. For any solar altitude, shading by surrounding vegetation may decrease any effect of the diameter of a clone on altering the light environment of that clone.
Our results indicate that the vertical gradients in light availability through a canopy outweigh horizontal gradients. While horizontal gradients do occur and create consistent spatial patterns in light availability, size of a clone seemed to have little effect on those gradients or on mean values of light at various heights in a clone for the range of clone sizes studied here. The magnitude and extent of the horizontal gradients suggest that only clones with diameters less than 20-30 cm would experience substantially greater light availability over most of their cross-sectional area due to decreased self-shading. This could allow rapid growth and, thus, potentially accounts for the general absence of small diameter clones in the field. Such small clones may have been overlooked in our sampling, but they may also be rare because they can grow more rapidly than larger clones until they reach the 35-60 cm diameter range. Small clones may simply not stay small very long and thus may not be common in the field. At the other extreme, the spatial patterns and similiarity of values for 35 to 60 cm diameter clones suggests that larger clones will not differ substantially in their light environments from 60 cm diameter clones.
If the light environment of most (i.e., 35-60 cm diameter) clones does not differ substantially, what then controls the maximum diameter of clones? Competition and resource limitation of some sort apparently do not often allow growth past a certain size, which seems to be circumferences of about 225 cm or cross-sectional areas of 4000 [cm.sup.2], in this study. In prairies, nitrogen and water are typically limiting and do not allow for unchecked growth (Knapp et al., 1998). In each of our field sites there was a moderately high density of other grass and forb species competing for the same resources. Further study of the interactions of such other factors with rates of clone growth are needed to determine what controls maximum size. In sum, while the light environment of a clone certainly is altered by the growth of the clone, self-shading does not seem to set a limit on diameter of clones.
Acknowledgments.--P art of this research was performed at Iowa State University during summer 1998 as part of a research internship by H. Kliebenstein sponsored by the Program for Women in Science and Engineering and the Department of Botany. Additional support was provided by the U.S. Environmental Protection Agency (Grant # R825796, T. W. Jurik and K. Moloney, principal investigators). We thank Randy Lueth and Engineering Plus, Inc. for use of digitization equipment.
(1.) Corresponding author: Telephone (515)294-5617; FAX (515)294-1337; e-mail: firstname.lastname@example.org
ACKERLY, D. D. AND F. A. BAZZAZ. 1995. Leaf dynamics, self-shading and carbon gain in seedlings of a tropical pioneer tree. Oecologia, 101:289-298.
BALDOCOHI, D. D., D. R. MATT, B. A. HUTCHISON AND R. T. MCMILLEN. 1984. Solar radiation within an oak-hickory forest: an evaluation of the extinction coefficients for several radiation components during fully-leafed and leafless periods. Agric. and Forest Meteorol, 32:307-322.
-----, B.A HUTCHISON, D. R MATT AND R T. McMILLEN. 1986. Seasonal variation in the statistics of photosynthetically active radiation penetration in an oak-hickory forest. Agric. and Forest Meteorol, 36:343-361.
BJORKMAN, O. 1981. Responses to different quantum flux densities, p. 57-107. In: O. L. Lange, P.S. Nobel, C. B. Osmond and H. Ziegler (eds.). Encyclopedia of plant physiology, Vol 12A. Physiological plant ecology: responses to the physical environment. Springer-Verlag, Berlin.
CALOWELL, M. M., T. J. DEAN, R S. NOWAK, R. S. DZUREE AND J. H. RICHARDS. 1983. Clonegrass architecture, light interception, and water-use efficiency: assessment by fiber optic point quadrats and gas exchange. Oecologia, 59:178-184.
CANHAM, C. D., A. C. FINZI, S. W. PACALA AND D. H. BURBANK. 1994. Causes and consequences of resource heterogeneity in forests: interspecific variation in light transmission by canopy trees. Can. J. For Res., 24:337-349.
CHABOT, B. F., T. W. JURIK AND J. F. CHABOT. 1979. Influence of instantaneous and integrated light-flux density on leaf anatomy and photosynthesis. Am. J. Bot., 66:940-945.
CLARK, D. B., D. A. CLARK, P. M. RICH, S. WEISS AND S. F. OBERBAUER. 1996. Landscape-scale evaluation of understory light and canopy structure: methods for application in a neotropical lowland rain forest. Can. J Forest Res., 26:747-757.
CONANT, S. AND P. G. RISSER. 1974. Canopy structure of a tall-grass prairie. J. Range Management, 27: 313-318.
CRESSIE, N. A. C. 1991. Statistics for spatial data. John Wiley and Sons, New York. 900 p.
FLIERVOET, L. M. AND M. J. A. WERGER. 1984. Canopy structure and microclimate of two wet grassland communities. New Phytol., 96:115-130.
GATES, D. M. 1980. Biophysical ecology. Springer-Verlag, New York. 611 p.
GREAT PLAINS FLORA ASSOCIATION. 1986. Flora of the Great Plains. University Press of Kansas, Lawrence, Kansas. 1392 p.
HIROSE, T. AND M. J. A. WERGER. 1987. Maximizing daily canopy photosynthesis with respect to the leaf nitrogen allocation pattern in the canopy. Oecologia, 72:520-526.
----- AND -----. 1995. Canopy structure and photon flux partitioning among species in a herbaceous plant community. Ecology, 76:466-474.
JONES, H. G. 1983. Plants and microclimate. Cambridge University Press, Cambridge. 323 p.
JURIK, T. W., G. M. BRIGGS AND D. M. GATES. 1985. A comparison of four methods for determining leaf area index in successional hardwood forests. Can. J For Res., 15:1154-1158.
----- AND J. M. PLEASANTS. 1990. The effects of light and plant size on fate of seedlings of the prairie compass plant (Silphium laciniatum L.). Functional Ecol., 4:661-665.
KNAPP, A. K.,J. M. BRIGGS, J. M. BLAIR AND C. L. TURNER. 1998. Patterns and controls of aboveground net primary production in tallgrass prairie, p. 193-221. In: A. K. Knapp, J. M. Briggs, D. C. Harmett and S. L. Collins (eds.). Grassland dynamics. Oxford University Press, Oxford.
MITCHLEY, J. AND J. H. WILLEMS. 1995. Vertical canopy structure of Dutch chalk grasslands in relation to their management. Vegetatio, 117:17-27.
MONSI, M. AND T. SAEKI. 1953. Uber den Lichtfaktor in den Pflanzengesellschaften und seine Bedeutung fur die Stoffproduktion. fap. J. Bot., 14:22-52.
NICOTRA, A. B., R. L. CHAZDON AND S. V. B. IRIARTE. 1999. Spatial heterogeneity of light and woody seedling regeneration in tropical wet forests. Ecology, 80:1903-1926.
PEARCY R W. AND W. YANG. 1996. A three-dimensional crown architecture model for assessment of light capture and carbon gain by understory plants. Oecologia, 108:1-12.
RYEL, R. J., W. BEYSCHLAG AND M. M. CALDWELL 1994. Light field heterogeneity among tussock grasses: theoretical distributions of light harvesting and seedling establishment in tussocks and uniform tiller distributions. Occologia, 98:241-246.
SAS INSTITUTE INC. 1996. SAS Proprietary Software Release 6.12. SAS Institute, Inc., Cary, North Carolina.
SMITH, A. P., K. P. HOGAN AND J. R. IDOL 1992. Spatial and temporal patterns of light and canopy structure in a lowland tropical moist forest. Biotropica, 24:503-511.
SCHIMEL, D. S., T. C. F. KITTEL, A. K. KNAPP, T. R SEASTEDT, W. J. PARTON AND V. B. BROWN. 1991. Physiological interactions along resource gradients in a tallgrass prairie. Ecology, 72:672-684.
SWAN, A. K. H. AND M. SANDILANDS. 1995. Introduction to geological data analysis. Blackwell Science Ltd., Oxford. 446 p.
TURNER, C. L. AND A. K. KNAPP. 1996. Responses of a C4 grass and three CS forbs to variation in nitrogen and light in tallgrass prairie. Ecology, 77:1733-1749.
WARREN-WILSON, J. 1965. Stand structure and light interception analysis of point quadrats. I. Applied Ecology, 2:383-390.
WERGER, M. J. A., E. M. DUSINK AND L. M. FUERVOET. 1986. Types of phytomass- and leaf area indexprofiles in grassland vegetation. Vegetatic, 65:39-45.
YODA, K. 1974. Three-dimensional distribution of light intensity in a tropical rain forest of West Malaysia. Jap. J. Ecology, 24:247-254.
|Printer friendly Cite/link Email Feedback|
|Author:||JURIK, THOMAS W.; KLIEBENSTEIN, HEATHER|
|Publication:||The American Midland Naturalist|
|Article Type:||Statistical Data Included|
|Date:||Jul 1, 2000|
|Previous Article:||Effects of the Exotic Invasive Shrub Lonicera maackii on the Survival and Fecundity of Three Species of Native Annuals.|
|Next Article:||An 11-year Ecological Study of a Rare Prairie Perennial (Asclepias meadii): Implications for Monitoring and Management.|