**3. Results**

The sample of images gathered in the investigation includes forested lands (**Figure 2**), agricultural lands (**Figure 3**), residential environments (**Figure 4**) (known as urban savanna), downtown-like environments (**Figure 5**) (known as cliff detritus), industrial sites (**Figure 6**), and open water (**Figure 7**) [32].

A Visual Quality Prediction Map for Michigan, USA: An Approach to Validate Spatial Content http://dx.doi.org/10.5772/intechopen.79490 119

**Figure 2.** Image of sample number 67 of a forested landscape in the upper peninsula of Michigan (visual score of 38.548).

**Figure 3.** Image of sample number 80 of a farmland landscape in north-east of the lower peninsula of Michigan (visual score of 44.09).

regarded as close to unanimous in their agreement. The W test statistic approximates a Chisquare distribution with n−1 degrees of freedom (Eq. (3)). If computed values for Chi-square (results of (Eq. (3)) are greater than significant values in a Chi-square table for n−1 degrees of freedom (in this case 29 = 30–1), then there is a high level of agreement/concordance—the

**Variable Score** A. Purifies Air +1 0–1 B. Purifies Water +1 0–1 C. Builds Soil Resources +1 0–1 D. Promotes Human Cultural Diversity +1 0–1 E. Preserves Natural Resources +1 0–1 F. Limits Use of Fossil Fuels +1 0–1 G. Minimizes Radioactive Contamination +1 0–1 H. Promotes Biological Diversity +1 0–1 I. Provides Food +1 0–1 J. Ameliorates Wind +1 0–1 K. Prevents Soil Erosion +1 0–1 L. Provides Shade +1 0–1 M. Presents Pleasant Smells +1 0–1 N. Presents Pleasant Sounds +1 0–1 O. Does not Contribute to Global Warming +1 0–1 P. Contributes to the World Economy +1 0–1 Q. Accommodates Recycling +1 0–1 R. Accommodates Multiple Use +1 0–1 S. Accommodates Low Maintenance +1 0–1 T. Visually Pleasing +1 0–1

X<sup>2</sup> = m(n − 1)W (3)

The sample of images gathered in the investigation includes forested lands (**Figure 2**), agricultural lands (**Figure 3**), residential environments (**Figure 4**) (known as urban savanna), downtown-like environments (**Figure 5**) (known as cliff detritus), industrial sites (**Figure 6**),

predicted scores and the actual scores are in agreement.

**Table 1.** Variables for the environmental quality/health index in Eq. (1).

**3. Results**

Total score

**Health index**

118 Land Use - Assessing the Past, Envisioning the Future

and open water (**Figure 7**) [32].

**Figure 4.** An image of a residential landscape (urban savanna with visual quality score of 46.587), sample number 12 in the northwest of the lower peninsula.

**Table 2** presents the rankings of the images from the study. The predicted ranks are scores generated from the expected a land-use scores and applying the expected score to a land-use map of the study area. The actual scores are values taken and measured from random sites in the study area. Kendall's Concordance analysis revealed a Chi-square score of 54.267. The

A Visual Quality Prediction Map for Michigan, USA: An Approach to Validate Spatial Content

http://dx.doi.org/10.5772/intechopen.79490

121

**Property Predicted ranking Mean expected score Actual score Set 2 ranking**

 63.06312 55.94164 15 63.06312 46.58693 8 63.06312 70.73992 20 63.06312 68.34915 18

 81.56982 84.30117 24 81.56982 86.72108 27 81.56982 82.76631 23 81.56982 84.77721 25

 55.92384 57.74200 16 55.92384 45.25820 7 55.92384 44.09459 6 55.92384 53.27078 13

 90.91938 94.63968 28 90.91938 96.94017 29 90.91938 78.77838 22 90.91938 97.67801 30

 42.67716 46.98920 9 42.67716 36.70580 2 42.67716 36.70580 2 42.67716 38.58380 4

 44.99797 47.78978 11 44.99797 35.62803 1 44.99797 47.27917 10 44.99797 41.39820 5

Residential l 18 63.06312 73.69797 21

Downtown 23 81.56982 69.28333 19

Farmland 13 55.92384 59.12320 17

Industrial 28 90.91938 86.56068 26

Forested 3 42.67716 54.40120 14

Water 8 44.99797 52.89468 12

**Table 2.** Comparison of ranks between the average expected score and actual site photographs.

**Figure 5.** An image of sample number 73 of a downtown environment (cliff detritus with a visual score of 82.766).

**Figure 6.** An image of sample number 77 of an industrial environmental (with a visual score of 78.778).

**Figure 7.** An image of sample number 46 of a primarily open water environment of Lake superior (with a visual score of 41.498).

**Table 2** presents the rankings of the images from the study. The predicted ranks are scores generated from the expected a land-use scores and applying the expected score to a land-use map of the study area. The actual scores are values taken and measured from random sites in the study area. Kendall's Concordance analysis revealed a Chi-square score of 54.267. The


**Figure 5.** An image of sample number 73 of a downtown environment (cliff detritus with a visual score of 82.766).

120 Land Use - Assessing the Past, Envisioning the Future

**Figure 6.** An image of sample number 77 of an industrial environmental (with a visual score of 78.778).

**Figure 7.** An image of sample number 46 of a primarily open water environment of Lake superior (with a visual score

of 41.498).

**Table 2.** Comparison of ranks between the average expected score and actual site photographs.

this category. The study revealed numerous landscapes with scores in the 40s and 50s. Such landscapes are often modestly preferred environments. Scores in the 70s are less preferred and scores near or above 100 are not preferred [23]. Across the state, least preferred environments were rarely encountered, primarily in the southeastern portion of the state. The 95% confidence interval for any scores is ±5 points [33]. Thus, it takes a separation of 10 points for

A Visual Quality Prediction Map for Michigan, USA: An Approach to Validate Spatial Content

http://dx.doi.org/10.5772/intechopen.79490

123

The average expected score (the sum of the products between the number of grid cells by their expected score, divided by the total number of grid cells) for the whole state is approximately 47.4. This score suggests that collectively the whole state of Michigan may not be the most beautiful of all environments; on the other hand, the environmental score is quite respectable

To understand the context of the average expected score in Michigan, it is useful to explore the land-uses, published perceptions, recreational activities, and agriculture within the state. In an article published on the 29th of June, 2015, the Detroit Free Press reported that Thrillist who ranked all of the states in America, placed Michigan as the top state [34]. While the results of the list do not definitively demonstrate that Michigan is at the top, it does indicate that the environment of Michigan merits consideration as a noteworthy place in regard to visual quality and may be similar in score to numerous rural mixed agrarian and woodland environments around the world, such as in Poland, Romaina, or Hubei Province, PR China. It may be reasonable to consider the extensive shorelines, vast expanse of national forest lands (3 national forests), state forest lands, national parks, 99 state parks, a national lakeshore, wildlife refuges, woodlots, and agrarian landscapes assist in maintaining a relatively preferred environment [35]. According to the map (**Figure 8**), the impact of the urban and industrial development is not widespread in the state and has not yet affected the state with large megalopolis expanses. Without attempting to be promotional, the state remains a big fishing (salmon, trout, walleye, northern, pan-fish), hunting (black bear, elk, white-tailed deer, and wild turkey), recreation state (camping, boating, hiking, cross country trails, snowshoeing trails, biking trails, snowmobiling trails, and horseback-riding—actually thousands of miles of trails) and in the Upper Peninsula there are 150 waterfalls. Michigan has also been rated as one of the top places in the world for watching sunrises on earth [36]. Such features are promoted in Michigan's 'Pure Michigan' tourist campaign [37]. The state has the most diverse agricultural economy after California [36]. Michigan produces cherries, apples, blueberries, many vegetable crops, nursery plants, surgarbeets, a strong wine producing industry, potatoes, dairy, cattle, hogs, chickens, turkeys, timber, hardwoods for flooring and furniture, and paper pulp, as well as the staple corn, soybeans, and winter wheat. While the state is associated with the famed mid-west rust-belt and the failures of Detroit, the overall impact upon the state's extensive forested lands and agrarian landscape is relatively minimal [38]. In addition, visitors to the Detroit metropolitan area often are surprised with the activity and prosperous nature of the Detroit metropolitan area. The Detroit metropolitan area is a distributed urban environment with no dominant central district, the opposite of many other large cities which are more concentrated in the core such as a city like Shanghai, P.R. of China. The state is a major constituent of the third coast, the longest coastline associated with the United States and Canada (excluding the Arctic) [39]. The results of this study reflect the impression that

any pairs of images to be notably different as perceived by respondents.

and viewed by respondents as at least somewhat scenic.

**4.2. Applications of the map**

**Figure 8.** A map of the predicted environmental/visual quality of the Michigan landscape. Green areas are environments with good visual quality. Blue, magenta, red, and orange indicate successively less preferred environments. Yellow areas are the least preferred.

number found in a standard table for such a Chi-square is 52.336 (a 99.5% confidence level (p ≤ 0.005) for 29 degrees of freedom. Since 54.267 is larger than 52.336, the predicted scores and the actual scores are in agreement at a 99.5% confidence level (p ≤ 0.005). These results suggest that it is possible to construct an environmental/visual quality map of Michigan that is relatively statistically reliable (**Figure 8**). In other words it is possible to predict the visual quality of any site in Michigan by knowing the land-use and to accurately predict the expected visual quality correctly 199 times in 200 attempts for any one attempt.

## **4. Discussion**

#### **4.1. Understanding the map**

To interpret the scores, Burley notes that scores in the 30s indicate highly preferred environments [23]. From randomly selected sites across the study area, no landscape scored in this category. The study revealed numerous landscapes with scores in the 40s and 50s. Such landscapes are often modestly preferred environments. Scores in the 70s are less preferred and scores near or above 100 are not preferred [23]. Across the state, least preferred environments were rarely encountered, primarily in the southeastern portion of the state. The 95% confidence interval for any scores is ±5 points [33]. Thus, it takes a separation of 10 points for any pairs of images to be notably different as perceived by respondents.

#### **4.2. Applications of the map**

number found in a standard table for such a Chi-square is 52.336 (a 99.5% confidence level (p ≤ 0.005) for 29 degrees of freedom. Since 54.267 is larger than 52.336, the predicted scores and the actual scores are in agreement at a 99.5% confidence level (p ≤ 0.005). These results suggest that it is possible to construct an environmental/visual quality map of Michigan that is relatively statistically reliable (**Figure 8**). In other words it is possible to predict the visual quality of any site in Michigan by knowing the land-use and to accurately predict the expected

**Figure 8.** A map of the predicted environmental/visual quality of the Michigan landscape. Green areas are environments with good visual quality. Blue, magenta, red, and orange indicate successively less preferred environments. Yellow areas

To interpret the scores, Burley notes that scores in the 30s indicate highly preferred environments [23]. From randomly selected sites across the study area, no landscape scored in

visual quality correctly 199 times in 200 attempts for any one attempt.

**4. Discussion**

are the least preferred.

122 Land Use - Assessing the Past, Envisioning the Future

**4.1. Understanding the map**

The average expected score (the sum of the products between the number of grid cells by their expected score, divided by the total number of grid cells) for the whole state is approximately 47.4. This score suggests that collectively the whole state of Michigan may not be the most beautiful of all environments; on the other hand, the environmental score is quite respectable and viewed by respondents as at least somewhat scenic.

To understand the context of the average expected score in Michigan, it is useful to explore the land-uses, published perceptions, recreational activities, and agriculture within the state. In an article published on the 29th of June, 2015, the Detroit Free Press reported that Thrillist who ranked all of the states in America, placed Michigan as the top state [34]. While the results of the list do not definitively demonstrate that Michigan is at the top, it does indicate that the environment of Michigan merits consideration as a noteworthy place in regard to visual quality and may be similar in score to numerous rural mixed agrarian and woodland environments around the world, such as in Poland, Romaina, or Hubei Province, PR China. It may be reasonable to consider the extensive shorelines, vast expanse of national forest lands (3 national forests), state forest lands, national parks, 99 state parks, a national lakeshore, wildlife refuges, woodlots, and agrarian landscapes assist in maintaining a relatively preferred environment [35]. According to the map (**Figure 8**), the impact of the urban and industrial development is not widespread in the state and has not yet affected the state with large megalopolis expanses. Without attempting to be promotional, the state remains a big fishing (salmon, trout, walleye, northern, pan-fish), hunting (black bear, elk, white-tailed deer, and wild turkey), recreation state (camping, boating, hiking, cross country trails, snowshoeing trails, biking trails, snowmobiling trails, and horseback-riding—actually thousands of miles of trails) and in the Upper Peninsula there are 150 waterfalls. Michigan has also been rated as one of the top places in the world for watching sunrises on earth [36]. Such features are promoted in Michigan's 'Pure Michigan' tourist campaign [37]. The state has the most diverse agricultural economy after California [36]. Michigan produces cherries, apples, blueberries, many vegetable crops, nursery plants, surgarbeets, a strong wine producing industry, potatoes, dairy, cattle, hogs, chickens, turkeys, timber, hardwoods for flooring and furniture, and paper pulp, as well as the staple corn, soybeans, and winter wheat. While the state is associated with the famed mid-west rust-belt and the failures of Detroit, the overall impact upon the state's extensive forested lands and agrarian landscape is relatively minimal [38]. In addition, visitors to the Detroit metropolitan area often are surprised with the activity and prosperous nature of the Detroit metropolitan area. The Detroit metropolitan area is a distributed urban environment with no dominant central district, the opposite of many other large cities which are more concentrated in the core such as a city like Shanghai, P.R. of China. The state is a major constituent of the third coast, the longest coastline associated with the United States and Canada (excluding the Arctic) [39]. The results of this study reflect the impression that the state is still predominantly a rural environment and is modestly beautiful in a 'low-key' manner. In other words, the average expected score and the existing landscape condition are mutually similar in expectations.

The map resulting for this study is an image representing the spatial perceptions of the respondents concerning the quality of the landscape across the state. As the land is managed and developed, the scores can be recomputed to estimate the perceived changes (both improvements and degradation) for the state. With additional research, the map could be considered a metric of the state's general perceived environmental quality. Across the globe, people are concerned about the impacts of human's transforming the environment; yet, comparatively, the compiled total environmental quality has not appreciatively changed since the arrival of Europeans, Africans, and Asians (a score predicted to be around 43 between the years 1816 and 1858 to a score of 47.4 in the year 2001, which is not significantly different in perceived quality) [40]. **Figures 9**–**11** are images from the landscape with scores near the current expected mean for much of the state. Notice none of the images are spectacular; however, none of the images age dismal either.

The results of this study and related studies indicate that land cover type is a strong predictor of visual/environmental quality. In planning and design circles, the nuances of the built environment are heavily debated with small details carefully examined by experts. Yet respondents evaluate these cover types relatively uniformly. In other words the refined changes and differences observed by experts and taught in planning and design schools are not necessarily observed/detected by the public. For experts the differences between communities can be quite distinct. As Charles Jencks indicated, there are at least two levels of cognition: the expert's and the public's [41]. For example, architects debate the merits of buildings; however, the research suggests that a poor building in a warehouse district or a noted piece of architecture such as Frank Lloyd Wright's Falling water house are perceived as simply structures and

A Visual Quality Prediction Map for Michigan, USA: An Approach to Validate Spatial Content

http://dx.doi.org/10.5772/intechopen.79490

125

**Figure 11.** An image of a great lakes coastline, generating a score of 47.8.

the less of them comprising the view, the more the environment is preferred [23].

Similarly, landscape architects debate the merits of various plant material and landscape settings; yet, the public sees a noble red pine tree (*Pinus resinosa* Sol. Ex Aiton.) or a weedy boxelder tree (*Acer negundo* L.) as basically the same thing—a tree. It is only when the plant has ornamental flowers, is there an increased appreciation during the period of flowering. Within any cover type, there is variation (scores can range up or down), but the expected mean within the cover type is quite consistent when measured repeatedly across time, across locations, and among respondent groups in America and Europe. In other words, the most important characteristic that the planner or designer makes concerning visual/environmental quality may be in determining the cover type for a parcel of land, not the refined details of a design. In the future, designers may develop improved techniques to mask human intrusions and abundantly incorporate preferred features such as wildlife, flowers, and views of distant landscapes, creating strong variation in the range of scores possible for a given cover type. Then the covariation of land cover type with visual/environmental quality may no longer hold true. The relationships between the perception of environmental quality and cultural settings are explored in the book From Eye to Heart: Exterior Spaces Explored and Explained [42]. This book begins with discussing expectation values concerning the environment. Then the book examines some of the many interpretations for planning and designing across the globe and

**Figure 9.** An image of a modest rural residential setting collected in the study area with a score of 46.6.

**Figure 10.** An image of a typical agricultural landscape found in much of Michigan. This image produced a score of 44.1.

**Figure 11.** An image of a great lakes coastline, generating a score of 47.8.

the state is still predominantly a rural environment and is modestly beautiful in a 'low-key' manner. In other words, the average expected score and the existing landscape condition are

The map resulting for this study is an image representing the spatial perceptions of the respondents concerning the quality of the landscape across the state. As the land is managed and developed, the scores can be recomputed to estimate the perceived changes (both improvements and degradation) for the state. With additional research, the map could be considered a metric of the state's general perceived environmental quality. Across the globe, people are concerned about the impacts of human's transforming the environment; yet, comparatively, the compiled total environmental quality has not appreciatively changed since the arrival of Europeans, Africans, and Asians (a score predicted to be around 43 between the years 1816 and 1858 to a score of 47.4 in the year 2001, which is not significantly different in perceived quality) [40]. **Figures 9**–**11** are images from the landscape with scores near the current expected mean for much of the state. Notice none of the images are spectacular; however,

**Figure 9.** An image of a modest rural residential setting collected in the study area with a score of 46.6.

**Figure 10.** An image of a typical agricultural landscape found in much of Michigan. This image produced a score of 44.1.

mutually similar in expectations.

124 Land Use - Assessing the Past, Envisioning the Future

none of the images age dismal either.

The results of this study and related studies indicate that land cover type is a strong predictor of visual/environmental quality. In planning and design circles, the nuances of the built environment are heavily debated with small details carefully examined by experts. Yet respondents evaluate these cover types relatively uniformly. In other words the refined changes and differences observed by experts and taught in planning and design schools are not necessarily observed/detected by the public. For experts the differences between communities can be quite distinct. As Charles Jencks indicated, there are at least two levels of cognition: the expert's and the public's [41]. For example, architects debate the merits of buildings; however, the research suggests that a poor building in a warehouse district or a noted piece of architecture such as Frank Lloyd Wright's Falling water house are perceived as simply structures and the less of them comprising the view, the more the environment is preferred [23].

Similarly, landscape architects debate the merits of various plant material and landscape settings; yet, the public sees a noble red pine tree (*Pinus resinosa* Sol. Ex Aiton.) or a weedy boxelder tree (*Acer negundo* L.) as basically the same thing—a tree. It is only when the plant has ornamental flowers, is there an increased appreciation during the period of flowering. Within any cover type, there is variation (scores can range up or down), but the expected mean within the cover type is quite consistent when measured repeatedly across time, across locations, and among respondent groups in America and Europe. In other words, the most important characteristic that the planner or designer makes concerning visual/environmental quality may be in determining the cover type for a parcel of land, not the refined details of a design. In the future, designers may develop improved techniques to mask human intrusions and abundantly incorporate preferred features such as wildlife, flowers, and views of distant landscapes, creating strong variation in the range of scores possible for a given cover type. Then the covariation of land cover type with visual/environmental quality may no longer hold true.

The relationships between the perception of environmental quality and cultural settings are explored in the book From Eye to Heart: Exterior Spaces Explored and Explained [42]. This book begins with discussing expectation values concerning the environment. Then the book examines some of the many interpretations for planning and designing across the globe and through time. The book concludes with the relationship between environmental science and the built/managed landscape. This book provides some greater and broader context to the visual mapping study presented in this investigation.

**5. Conclusions**

of additional areas.

**Author details**

, Chung Qing Liu2

1 Namk Kemal University, Tekirdag, Turkey

3 Michigan State University, USA

\*Address all correspondence to: burleyj@msu.edu

2 Jiangxi Agricultural University, Nanchang, PR China

Academy of Sciences. 2015;**112**(28):8567-8572

Rüya Yilmaz1

**References**

Predictive, respondent based models have been constructed to measure environmental and visual quality. This work is based upon over 50 years of research by investigators in the social, recreational, and planning and design disciplines/profession. The attributes of the landscape can be measured to form reliable maps of environmental/visual quality, providing a metric to assess landscapes, including urban landscapes. We were able to produce such a metric map for Michigan. We believe our approach allows investigators to evaluate these visions and assess, measure, and quantify environmental perceptions. Furthermore, we believe the methods are reproducible, allowing investigators around the world to produce similar maps

A Visual Quality Prediction Map for Michigan, USA: An Approach to Validate Spatial Content

http://dx.doi.org/10.5772/intechopen.79490

127

and Jon Bryan Burley<sup>3</sup>

[1] Martin B, Ortega E, Otero I, Arce RM. Landscape character assessment with GIS using map-based indicators and photographs in the relationship between landscape and

[2] Bratman GN, Hamilton JP, Hahn KS, Daily GC, Gross JJ. Nature experience reduces rumination and subgenual prefrontal cortex activation. Proceedings of the National

[3] Dupont L, Antrop M, Van Eetvelde V. Does landscape related expertise influence the visual perception of landscape photographs? Implications for participatory landscape

[4] Filova L, Vojar J, Svobodova K, Sklenicka P. The effect of landscape type and landscape elements on public visual preferences: Ways to use knowledge in the context of landscape planning. Environmental Planning and Management. 2015;**58**(11):2037-2055 [5] Meagher BR, Marsh KL. Testing an ecological account of spaciousness in real and virtual

planning and management. Landscape and Urban Planning. 2015;**141**:68-77

roads. Journal of Environmental Management. 2016;**180**:324-334

environments. Environment and Behavior. 2015;**47**(7):782-815

\*

#### **4.3. Limitations**

Additional research should be conducted to refute, validate, or refine the findings presented here. However, it may be surprising that with only 60 images from across the state, the selected images facilitate a significant statistical result and make a reliable map for a whole state. In some studies, weakly developed experiments with large data sets and numerous observations may produce unremarkable results. The results presented in this paper follow a philosophy that relies upon methodologies that have yielded significant results (such as employing Q-sort techniques as opposed to Likert scales), an exploration of reliable predictors, and non-parametric statistical procedures. We would like to believe that well-grounded, focused, simple experiments often yield results that some expensive and elaborate studies fail to yield. The senior investigator in this study had spent over 30 years, carefully plotting each step, conducting the next logical step/experiment before proceeding. So each study is simple (for example the statistics were done on a spreadsheet), yet yielding meaningful results. This philosophical approach has served the team well and we encourage others to have this level of insight and commitment when formulating environmental quality studies. Investigators sometimes believe that technology, huge sample sizes, and big money make impressive research.

The results produced in the map are dependent upon the quality of the land cover type map. Many cover maps concentrate upon the great variation of vegetation and naturalistic cover types. Urban and suburban cover types are rarely produced with the same level of sensitivity. Brady et al. at the University of Waterloo developed an excellent example concerning the classification of human disturbed areas, based upon morphological and ecological features [32]. Yet this level of description has not yet permeated many land cover maps. We believe that classification systems that adopt ideas embedded in work by investigators such as Brady et al. will produce higher quality maps in urban areas than the maps that are currently produced [32].

In the landscape there are several types of cover types that exist in the landscape such as large sand dunes, mud flats, and bare rock that may be beyond the predicative capabilities of this study. These more rare landscape types were neither studied in the prediction models nor in creating the map presented in this study. Cover types such as these would need to be studied in detail to make a more comprehensive and complete map. This is computed to be a substantial area of land, approximately 1700 km<sup>2</sup> . Yet this area is only about 0.68% of the land of Michigan.

The resolution of the map is relatively coarse (2.207 km2 of land). Large maps with finer resolution will yield more refined results. A grid cell of that size will certainly have variation in it. The score represents the mean expected value within the cell.
