Human Judgments of Visual Air Quality

The previous section discussed methodologies for establishing the change in atmospheric particulate loading required to be noticeable either as a layered haze or as a change in scenic quality. It should be emphasized that calculations of detection thresholds and JNCs are statements about changes in information content in an image. JNC changes in the appearance of an image are not necessarily good indicators of judged image quality. For instance, a change in 10 JNCs in a scene with low overall contrast may not be judged to have the same change in image quality as 10 JNCs in a high-contrast scene.

Studies by Malm et al. (1980, 1981), Latimer et al. (1980, 1983), Middleton et al. (1984), Stewart et al. (1983), and Hill (1990) have established relationships between judgments of image quality of natural scenes and various atmospheric and vista parameters such as mountain/sky contrast, solar angle, extinction coefficient, sky color, and percent cloud cover. Latimer et al. (1980) had observers judge scenic beauty (SBE) and visual air quality (VAQ) for a number of eastern and western national park vistas as they appeared under a variety of illumination and meteorological conditions. The results of their study were mixed and in some cases contra-

dietary. In Latimer et al. (1980, p. 113), they conclude that "to different extents for different vistas, ratings of VAQ and SBE both increase with increasing visual range." In Latimer et al. (1983, pp. 49-50), they conclude that "ratings of SBE of a given vista were independent of visual range unless there was a dominant distant landscape feature in the landscape scenery." Since the visual range calculation "normalizes" out specific unique characteristics of vistas, these results are not surprising. The Latimer studies did conclude that changes in illumination did have a considerable effect on SBE ratings. Middleton et al. (1984) also concluded that illumination was important to VAQ judgments and were able to show at one site that there is a good correlation between VAQ and ln(/)scat), where bsc.d[ is the atmospheric scattering coefficient. Additionally, Hill (1990) emphasizes that color is extremely important to judgments of scenic beauty.

Malm et al. (1980) examined the relationship between VAQ and vista contrast. They showed that, under fixed illumination and meteorological conditions, apparent vista contrast of the most distant vista element was a good prediction of VAQ judgments. The study also showed that changes in foreground color (due to change in illumination), addition of clouds, or snow cover caused the VAQ ratings to be higher but did not cause the sensitivity of VAQ to change in vista contrast change. Malm et al. (1980) also presented a model of human perception of VAQ. The model is based on the observation that ratings of VAQ are proportional to the sum of the fraction of each scenic element subtended by various landscape features multiplied by the atmospheric transmittance between that landscape feature and observer. It was shown that when a single landscape feature, void of color and textural detail, dominates the perceived change in visual air quality, the model predicts a linear relationship between VAQ and the apparent contrast of that landscape feature (contrast of form).

Several researchers have found that judgments of photographs can be used as surrogates for judgments made in the field provided the experiments have been properly designed. This is an important finding since one way to reduce the per-observation cost of obtaining judgment-based measurements of visual air quality is to use judgments of photographs rather than field observations. For example, Stewart et al. (1984) found that although visual air quality tends to be judged slightly worse in photographs than in the field, the relative differences among scenes are approximately the same whether visual air quality is judged from photographs or in the field.

The implication of the visual air quality perception research described in the preceding paragraphs is that there are a number of variables such as sun angle, cloud cover, and scene composition that are firmly integrated into judgments of aesthetic value of a scenic resource. Therefore, studies designed to assess social, psychological, or economical value associated with a given change in atmospheric particulate concentration must be designed in such a way that these confounding variables do not affect the outcome of the experiment. For instance, a number of experiments have been carried out using photographs of landscape features under a variety of air quality conditions as the stimulus. To avoid extraneous variables such as sun angle from affecting the study, it is essential that the study be carried out using photographs taken at the same time of day and under similar lighting conditions.

Was this article helpful?

0 0

Post a comment