Tree and Building Influences on UV 12431 Measurements

Along with protective clothing and sunscreen use, seeking shade is a primary means recommended by public health agencies to avoid excess exposure to UV radiation (Environmental Protection Agency, 1995; Parisi and Kimlin, 1999; Saraiya et al., 2004). However, tree effects on UV may differ significantly depending on the view of the sky and to some extent, the trees species. Ultraviolet shadows generally differ substantially from the shadows of visible light (Heisler et al., 2003 a); with UV shadows being more "fuzzy" and much less sharp edged than visible shadows. Most research on shading structures and trees has been carried out under the premise that minimizing UV exposure at all times is most beneficial to people (Grant, 1997; Moise and Aynsley, 1999; Gies and Mackay, 2004); however, this view may change with increased value placed on the solar production of vitamin D.

The distribution of radiance across the sky for different wavelengths is important for irradiance in urban areas. With distributions such as the PAR, where most of the sky radiance emanates from regions close to the sun, shadows are distinct, and irradiance at points within a shadow is minimally enhanced by radiance from regions of the sky farther from the sun (Grant et al., 1996; Grant and Heisler, 1997). For the UV-A, and more so for the UV-B (Grant and Heisler, 1997; Grant et al., 1997a, b), the radiance is more evenly distributed across the sky, so that in shadows, irradiance may be considerably augmented by sky radiance if large areas of the sky are in view. This effect is enhanced by the greater proportion of total irradiance in the open that comes from the sky in the UV bands (Grant and Gao, 2003).

One effect of shade is the increase in the diffuse fraction (ratio of diffuse to direct-beam radiation). Parisi et al. (1999) found the average full-sun diffuse fraction for UV-B radiation in the open was 0.39, 0.26, and 0.46 during the morning, noon, and afternoon, respectively. In tree shade, the diffuse fractions were increased so that they averaged 0.60 to 0.61 for any of the time periods.

Available information on tree influences on UV irradiance is largely derived from essentially anecdotal measurements under a few trees with broadband sensors (e.g., Grant and Heisler, 1996; Moise and Aynsley, 1999). Spectral measurements at tree and in open have been taken simultaneously because of the limited availability of only one spectroradiometer (Parisi et al., 2001b). Part of the reason for the scarcity of data is the surprising difficulty in finding individual trees located where irradiance near or under the tree is not influenced by buildings or other nearby trees. The possibility for replications of measurements for full-sized trees of the same species is especially unlikely.

Results from a survey of tree influences on UV radiation (shown in Table 12.2) (Grant and Heisler, 1996; Heisler et al., 2003a), illustrate the effect of mature deciduous street trees, typically found in older suburban neighborhoods, on the PAR and UV-B irradiance at points of pedestrian height with significant sky views. Irradiances measured with UV-B and PAR sensors close together on one tripod were averaged over half-hour time periods at a mid-western US location (latitude 40.5°N). The sensor that measured the UV-B (an International Light SED240/ UV-B/W) had a response from 280 nm to 320 nm, cutting off sharply before 320 nm, unlike erythemal sensors that extend into the UV-A (Fig. 12.1). Skies were clear for all measurements. Comparable irradiances measured at a rural field provided an open-condition reference. Upward-facing, hemispherical-view photos from each measurement point (Fig. 12.2) showed that the percent of effective sky view varied from 34% to 60%. The examples include configurations where the shade was provided by trees only (Fig. 12.2(a), (b), (c)), and also by trees and buildings (Fig. 12.2(g)).

In the examples shown in Table 12.2, for in-leaf trees in the shade, PAR was as low as 15% of irradiance in the open (Point b), and UV-B was 44% of open. Conversely, at locations near in-leaf trees, but out of their visible shadow, PAR was not appreciably reduced, while UV-B measured at Point e was only about 59% of that in the open. Trees with only bare branches and twigs can cause substantial reductions in irradiance (Heisler, 1985). Table 12.2 also shows that the 44% of UV-B in the shade of leafless trees was only 7 percentage points more than the relative irradiance in the shade of in-leaf trees. The difference in relative irradiance in the UV-B was much less between shade and sunlit points (averaging 24 percentage points for in-leaf trees vs. 16 percentage points for leafless trees) than in the PAR (averaging 81 percentage points for in-leaf trees vs. 67 percentage points for leafless trees).

The UV-B relative irradiances in Table 12.2 are representative of tree effects, though irradiance will differ somewhat with a greater range of SZA and turbidity of the sky. The important contrast is between the relative irradiance in the UV-B

Table 12.2 Average UV-B and PAR irradiance below a street-tree canopy as a percent of irradiance in the open, away from any obstructions (Grant and Heisler, 1996; Heisler et al., 2003a). Column a is solar zenith angle. Sky views are expressed in percent; they would be converted to decimal fractions (0 to 1) in equations

Percent of irradiance in open Percent of view

Percent of irradiance in open Percent of view

Table 12.2 Average UV-B and PAR irradiance below a street-tree canopy as a percent of irradiance in the open, away from any obstructions (Grant and Heisler, 1996; Heisler et al., 2003a). Column a is solar zenith angle. Sky views are expressed in percent; they would be converted to decimal fractions (0 to 1) in equations

Point

UVB

PAR

Buildings

Trees

Sky

Skyeff*

In-leaf

Shade

a

21

16

32

68

57

57

b

44

15

46

54

47

47

c

47

18

53

47

52

52

Sunlit

d

74

97

45

55

45

45

e

59

96

33

67

51

51

Out-of- leaf

Shade

f

44

27

41

59

34

34

g

30

53

23

17

60

53

53

Sunlit

h

69

96

30

70

36

36

i

56

96

43

57

45

45

j

41

95

31

13

56

60

60

* Skyeff is sky view weighted according to the proportional contribution to irradiance on the horizontal from each 10° zenithal band of the sky.

* Skyeff is sky view weighted according to the proportional contribution to irradiance on the horizontal from each 10° zenithal band of the sky.

Figure 12.2 Fisheye photos from locations where below-canopy irradiance was measured as listed in Table 12.2. Frames (a), (b), (c), and (g) indicate points listed in the table and the visible (represented by the PAR band). Most of the in-leaf measurements were made in early September with SZAs ranging from 47° to 57°. Relative irradiances, particularly in the UV-B, would be expected to be smaller because of less scattering with smaller zenith angles and smaller diffuse fractions during mid-day in midsummer.

Differences in sky diffuse fraction are primarily responsible for the differences between UV-B and visible relative irradiance, although differences in reflectance, particularly for sunlit points, can also contribute, especially near building walls. At Point g (Table 12.2 and Fig. 12.2) where a building with a sunlit red brick wall made up 23% of the view, reflection from the wall led to a greater relative irradiance for PAR at a tree-shaded point than when no wall was present. In the UV-B waveband, the wall reduced the relative irradiance because reflection in the UV-B from the brick surface was less than the irradiance that would have come from the sky had the building not been present. Measurements showed that the brick building wall with some windows reflected about 18% of incident visible radiation, whereas according to calculations, it reflected only about 3% of the UV-B.

The effect of reflected radiation from tree crowns on irradiance near the tree is less apparent. As noted in Section 12.4.2, leaf reflectance in the UV-B is about 5% for most species, and PAR reflectivity may be three times larger. Thus, tree crowns contribute little scattered UV-B to adjacent points. However, reflection from trees would be significant in the PAR, and this could partly explain the small reduction of 3% for PAR in sunlit points near trees (Table 12.2).

Figure 12.2 Fisheye photos from locations where below-canopy irradiance was measured as listed in Table 12.2. Frames (a), (b), (c), and (g) indicate points listed in the table

Other studies have shown results comparable to those in Table 12.2. Where the sun is only obscured by a small tree crown leaving a large portion of the sky in view, the contrast between reductions in the visible and UV wavelengths is even more pronounced. For example, with a tree blocking only 20% of the sky view, UV-B relative irradiance was 63% compared to a visible relative irradiance of about 10% (Grant, 1997).

The potential effect of differences in tree species on UV-B irradiance below their crowns has not been well quantified (Heisler et al., 2003a). This is due to such factors as: (1) the importance of diffuse sky radiation in determining irradiance below tree crowns; (2) the difference in crown density with tree size and pruning regimes, and (3) the considerable difficulty of sampling irradiance effects of individual tree crowns when no other nearby trees and buildings have any influence. These tree and building influences are especially important in the UV-B.

Reductions in UV-B can almost be complete where trees obscure most of the sky. In tropical Australia, UV-B irradiance with clear sky conditions and a range of SZAs was reduced to an average of only 3% by the "dense foliage" of a fig tree, though sky view was not specified (Moise and Aynsley, 1999). This was a greater reduction than provided in seven other urban shade structures that ranged from a school grandstand to a concrete walkway cover. There are also large reductions of UV-B under dense forest canopies. Where the sky is nearly completely obscured and UV-B irradiance is reduced to essentially negligible levels, relative PAR penetration will be greater than UV-B penetration (Lee and Downum, 1991; Brown et al., 1994). This might be expected due to the low transmittance of UV radiation through leaves (Grant et al., 2003). Even in forest canopies with thin but horizontally uniform leaf distributions, UV-B is attenuated more than the PAR (Yang et al., 1993).

In order to more precisely evaluate the influence of tree shade on human health, the tree shade influence on various action spectra for health effects, (i.e., pre-vitamin D and erythema) can be modeled if the irradiance spectra are known. Spectroradiometer measurements at Toowoomba, Australia (27.5°S latitude) showed a linear increase in average relative irradiance with decreasing wavelength from 400 nm to 300 nm in the shade of five Australian trees. Relative irradiance at 300 nm was almost double that at 400 nm (Parisi and Kimlin, 1999). Similar measurements were made by Parisi et al. (2001b) for a camphor tree (Cinnamomum camphora) with a "medium dense" canopy that was 6.4 m tall and had a 4.2 m-wide canopy that extended to only 0.4 m from the ground. Interpretation of irradiance ratios in shade to irradiance in the open (see Fig. 12.2, Parisi et al., 2001b) yielded the shade ratios as shown in Fig. 12.3 for irradiance in the shade below the edge of the tree crown. A polynomial (shade ratio = 1.04 x 10-5 X2 - 0.0105 X + 2.778, X = wavelength in nm) fit the interpreted points closely (R = 0.995) as shown by the curve in Fig. 12.3. The measurements of Parisi et al. (2001b) below the tree crown, but half way from the edge of the crown to the tree trunk, had similar shade ratios. However, at the trunk, essentially all of the sky was blocked from the spectroradiometer's view and the shade ratios varied less by wavelength, ranging from about 0.15 at 305 nm to 0.10 at 395 nm. Figure 12.3 also shows a typical UV spectrum in the open for Lauder, New Zealand at 45°S from Fig. 2.5 in McKenzie and Liley (Chapter 2, this volume) and a predicted spectrum in the tree shade by applying the polynomial curve fit to the Parisi et al. (2001b) shade ratios.

-Irrad. in shade

-Irrad. in shade

300 320 340 360 380 400

Wavelength (nm)

Figure 12.3 Ratio of spectral irradiance on horizontal surfaces in tree shade to irradiance on the horizontal in the open (Parisi et al., 2001b) = square points; polynomial fitted to the shade ratio points = dashed line; spectra of UV irradiance in the open at Lauder, New Zealand on the summer solstice from McKenzie and Liley (Chapter 2, this volume) = heavy black line; and irradiance in tree shade predicted by the Parisi et al. (2001b) shade ratios = thin line

300 320 340 360 380 400

Wavelength (nm)

Figure 12.3 Ratio of spectral irradiance on horizontal surfaces in tree shade to irradiance on the horizontal in the open (Parisi et al., 2001b) = square points; polynomial fitted to the shade ratio points = dashed line; spectra of UV irradiance in the open at Lauder, New Zealand on the summer solstice from McKenzie and Liley (Chapter 2, this volume) = heavy black line; and irradiance in tree shade predicted by the Parisi et al. (2001b) shade ratios = thin line

The ratio of vitamin-D-weighted irradiance to erythema-weighted irradiance in tree shade can be higher than in the open. Figure 12.4 shows the irradiances in the open and in the tree shade from Fig. 12.3, but with weighting for erythema (UVEry) and pre-vitamim D production (UVVitD) using the CIE spectrum for pre-vitamin D (Bouillon et al., 2006). Integrating across wavelength, the ratio of UVVitD/UVEry is 1.95 in the open and 2.03 in the tree shade. In the example here, the difference is small between UVV;tD/UVEry in the open and in tree shade, and may not be of practical significance, although this ratio will generally be higher in tree shade than in the open. The UVVitD/UVEry ratio in tree shade would likely be somewhat higher for a location in the shade of a small isolated tree crown, where a large portion of the sky would be in view.

Was this article helpful?

0 0

Post a comment