**5. Conclusions**

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

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

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

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

lution will yield more refined results. A grid cell of that size will certainly have variation in it.

. Yet this area is only about 0.68% of the land

of land). Large maps with finer reso-

visual mapping study presented in this investigation.

126 Land Use - Assessing the Past, Envisioning the Future

**4.3. Limitations**

impressive research.

currently produced [32].

of Michigan.

substantial area of land, approximately 1700 km<sup>2</sup>

The resolution of the map is relatively coarse (2.207 km2

The score represents the mean expected value within the cell.

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 of additional areas.
