LANDSCAPE VISIBILITY MAPPING: THEORY and PRACTICE', John P. Felleman 1979 School of Landscape Architecture State University of New York College of Environmental Science and Forestry Syracuse, New York This research was sponsored by the New York Sea Grant Institute under a grant from the Office of Sea Grant, National Oceanic and Atmospheric Administration (NOAA), US Department of Commerce. FORHARD: An undertaking of this magnitude, covering two and a half part-- time years, involves many participants. I would like to thank Dave Harper and John Harbach of the School of Landscape Architecture for their continued support. Students (present and former) who have assisted in this effort, and the Port Bay Visibility Case Study, include: Varda Hilensky, Rick Dumont, Holly Burgess, Pete Jackson, Doug Johnson, Dennis Jud, and Mark Holzman. In particular, this study moles its existence to the many practicing professionals and academics \1ho have created a new and exciting field out of whole cloth in a very short period. The logical structures and conclusions which have been developed to organize the material arE: my m·m. The data used is primarily from project reports, at best a Iisecondary" source. I apologize at the Qutset for occasionally bending square pegs into round holes, and excisi.ng information from its context. If it stimulates thought and discussion, the artistic license is worth\.Jhile. John Felleman Syracuse, Summer 1979 I. II. III. IV. V. VI. VII. Contents List of Figures Introduction Visibility Mapping - A I-Iorking Model Data Assembl;' Viewshed Deliniation Processes - Stationary Vietvshed Deliniation Processes - Hoving Output - Plan Views and Perspectives Conclusions Appendix A References ii 1. 2. 3. 4. 5. 6a. 6b. 7. 8. 9. 10. 11. 12. 13. Visibility in Environmental Land Management Visual Transmission Point Light Source Model Visibility Model Landscape Continuum Distance Zones Distance Zones - Perspective Visibility Factors - Open Water Solar Position Seasonal Light Conditions - General View "Of" and "From" Road Skylining Water Influence Zone Enclosure - Massachusetts 14a. Enclosure - Hudson Valley l4b. 15. 16. Enclosure - Ross County, Ohio Vista Points Viewer Positions 17. Viewer/Distance Distribution 18. 19. 20. Viewshed Limit Accuracy Plan/Perspective Detailed Field Sketch 21. Photo Reference Sketch - Plan and Section 22. Artist Photo Rendering 23. 24. 25. 26. 27. 28. 29. 30. 3la. 3lb. 33. 34. 36. 37. Digitized Elevations DMA Landform Truncation Digitizing Flowchart Terrain Facets USGS Legend LUNR - Area and Point Data NYS Functional Highway Class- ification View Position Types Observer Environment Conditions Complexity of Visual Field and Edge Field Mapping Accuracy Full Scale View Simulation Projectile Trajectory Plan/Section Military Crest Scenic Overlook 38a,b,c Viewshed from Sections iii 39. 40. 41. 42. 43. 44. 45. 46. Similar Triangles Topographic Model Illuminated Model Modelscope Photo Intermediate Sightline Points Effects of Distance Selected View Lines Extent of Views 47. Roadway Distance Zones 48. Threshold View Angles 49. Water Based View Angles 50. PREVIEW Terrain Plot 51. Slope Fitting 52. Relative Aspect 53. Driver View Cones 54. 55. 56. Weighted View Cone Times Seen Map Continuotis Road View Methodology 57. Highway Filmstrip 58. Roadway Visibility 59. 60. 61. 62. 63. 64. 65. 66. Roadside Visibility Site Visibility - Summer Site Visibility - Winter Field Sketch Panorama Panoramic Photographic Sequence View Orientation Vectors Overlapping Viewsheds Weighted Overlapping Viewsheds 67. Overlay Analysis Mapping 68. Illustrative Perspective and Section 69. 70. Orthographic Projection Highspeed Printer Tonal Map 71. Line Plotter Visual Map 72. 73. 74. Grid Perspective with Diagonal Enhancement Hidden Line Section Perspective perpindicular to Line-of-Sight. PREVIEW Surface Feature Plot 75. Computer Generated Structures iv 76. Highway Perspective 77. Outline Perspective 78. Fuel Break Computer Montage 79. Highway Computer Montage 80. Birdseye Project Cell Visibility 81. NY Sea Grant Visibility Test Composite Map v Vision plays a central role in man's environmental behavior. It has been estimated that approximately ninety percent of our sensory stimulation is visual. Throughout the evolution of culture, landscape visibility has been a major determinant of the location and physical form of human settlements. Examples include defense fortifications, dominant religious structures, navigational aids, and recreation site development. The comprehensive management of environmental resources encompasses critical stages of resource analysis, land planning and project design. As illustrated in Figure 1, each of the interfaces between these stages incorporates visibility information. These include: scenic assessment, project location, impact analysis, activity allocations, and performance criteria. Visibility deals with both the geographic extent of surfaces which can be seen, and the legibility of features which, in composite visibility mapping, provides the basis for human perception and cognition €:NVfRONMe:NTAL ·~eNE.RY ASSESSMENT: LAND MANAGEMENT _.sPECIFIC l..DGATfON AND IMPACT NAl...YSIS "G!:NE"RALIZED LDCATfOHS AND PERFOf'..MANC£ CRITfRIA PZZZZ/ll VISIBILIT,( RELP-TE'D Figure 1: Visibility in Environmental Land Management.' 1 of landscapes. Geographic extent of visibility is the primary emphasis of this monograph. Historically, development siting and design decisions utilized a limited, intuitive approach to resource analysis. Visibility information was often developed in-situ, by means of direct terrain observations. In the twentieth century, as accurate topographic maps and remote sensing information became available, more sophisticated, offsite methods of visibility mapping evolved. The recent momentum given to environmental studies, particularly aesthetic concerns, by the National Environmental Policy Act (N.E.P.A.) has led to the widespread use of visibility mapping technigues. In the context of coastal aesthetic research conducted for the New York Sea Grant Program, a wide range of theoretical studies, and project reports were reviewed. The author found that although a variety of methods were apparently being used by design and resource professionals to map visibility, the published documentation exhibited a widespread lack of clarity in both conceptual logic, terminology, and methodological approaches. Many of these studies appeared to be underfunded, resources expended did not reflect the significance of the information, and some were 2 clearly isolated from or "tacked on" to a more comprehensive study. In contrast, some excellent prototypical state-of-the-art applied visibility analyses are beginning to emerge from the public and private sectors. The purposes of this report are threefold: a. To develop a coherent, conceptual construct of landscape visibility mapping; b. To systematically articulate the alternative methods of data organization, and visibility mapping through the use of-selected illustrated examples; and c. To foster improved integration of visibility information into complex resource planning and project design. The body of this report is presented in six parts. Section II contains a working definition of a comprehensive visibility model. Section III includes a discussion of data assembly for the elements of the visibility model; while Section IV is focused on line-of-sight processing methods for stationary positions; and Section V for moving observers. Visibility study outputs, plan views and perspectives, are discussed in Sedtion VI, while some brief conclusions are identified in Section VII. 3 EIYVIROfYMEtiTAl MODEL Dafa -..(} lA."dForm ~ ----::---i-----+--f---l----I--- E: Cove-r ~ It:~-I----+---+--I----t---f I.(t H-:-~.,.--,..we.r-,----l---+---I--+--+----l EnvIron"",,,! Oire.c..t Oi ito.. \ 5/GHTL/t/E PROCf55[5 Plan Per h"vc OUTPUT 00 \:J03)O[QflLLOLfW ~~[fQr>D~ ~ &~D~cGJ~[S A. INTRODUCTION Many primitive peoples have conceived of sight as a physical process which emanates from the human eye, as shown in Fig. 2. Although modern physics has shown the reverse is true (light enters the eye from external sources) the primitive approach is ideal for understanding the geographic extent of visibility. (Note: A visibility analysis of a smoke stack could "look" at the stack from the adjacent environment, or "look" from the stack into the environment.) Consider building a scale three-dimensional model of a real landscape. The model is placed in a dark room, and a tiny light source is placed on the model's surface at the position:of ·an observer .. The surfaces which are directly illuminated represent the locus of all visible points, the "viewshed". In the model, the configuration of the surfaces blocks the light from reaching the dark (hidden) areas. This blocking is called "interposition". If we then project the illuminated viewshed vertically LI13ffT ,,0SOURer ~~ Figure 2: Visual Transmission. to a horizontal plane, we have constructed a scale "potential visibility map". This is shown in Fig. 3. The infinite number of light rays in the above example are analogous to "lines of sight" passing from observer to the environment. A major issue in visibility analysis is to selectively reduce the number of such lines investigated to a representative, manageable set. The word "potential" is used above to clarify the difference between the simulation and the complexities of the real environment. A more comprehensive model of visibility mapping is shown in Fig. 4. Each of the elements is discussed below. B. MACRO LANDSCAPE Landforms and surface features are the primary elements of interposition. They also provide the visual content which is the basis for scenery analysis. The role of landforms in the context of scenic evaluations has been extensively explored in previous N.Y. Sea Grant work (Felleman, 1977). Landforms - Landforms include terrain and surface water features. Visually significant characteristics include size, shape, distance from observer, and aspect (orientation relative to solar position and observer location). In large scale, rugged landscapes, landforms tend to provide Figure 3: Point Light Source Model. VII:W5HfO MAl' .:.';';~ ;:.':.:::...;·1:',...LAHDFORM IXCI1 LINE OF !SIGHT VIEWER TYf'! POSITION !l :SURFACE' fEA.TU1\e~ ATM05PH!IIt LI>.1'I0FOrlM Figure 4.' Visibility Model. 6 the majority of interposition determinants. A frequently used simplification of the visibility model utilizes only landform, observer, lines of sight, and viewshed record. The latter is properly called a "Potential Topographic Viewshed" to clarify its limited scope. Surface Features - Surface features include vegetation and built forms. Regional scenic studies have developed the general principle that as the scale of landforms decreases, the visual significance of "z surface features increases (see Figure 5) (Research Planning and Design ~o ~z "'=>< ~Q ~ Z ..Associates, 1972, p.N-19). => ~ "' In addition, field research regarding scale and distance, indicates ~ ..< that the significance of surface features will decrease as sight-line " u z '"3~ Q % 0 '" <« ~ distance increases from foreground, to midground, to background (see Figures 6a, 6b) (Litton, 1968). A New England highway study revealed that land development types could be identified at a maximum of 1 km (0.5 km mean) (Jacobs and Way, 1969). In a significant water related analysis, the combination of earth curvature and light refraction are shown to reduce the apparent height of water surface objects (See Figure 7) (Roy Mann Associates, July, 1975a, p.293). (See also discussion in Section IV-D). Figure 5: Landscape Continuum. Landforms are generally static within the time frame of a project oriented visibility study. A major earth moving project would be an 7 ~ V I • I IIff } >I SM 4M .1M 2M //1 ':7'I< 8ackJround l1iddlejroull ~ro"nd Midd~.."<1 '"':J-rJ ro,",,,d Figure 6a: Distance Zones. Figure 6b: Distance Zones Perspecti ve . i-________________~3~M~/=Lf~~~_______________i_ Rf"FRAcnOl! Of LIGHT ------ ---- - ------------..J"-M.-?-' CURVATURE OF rHE EART/! Figure 7: Visibility Factors - Open Water. 8 example of the exception to this statement. However, the temporal dimension is highly relevant in the treatment of surface features as the seasonal attributes of vegetation and land development changes. A comprehensive visual analysis would include a representative treatment of expected surface feature conditions which are relevant during a project's short-term implementation stage, and its long-term useful life. Atmosphere - Atmosphere characteristics are a continuously variable element in visibility studies. Lighting conditions, clouds and precipitation can all modify the potential topographic and surface feature visibility. This is important to both viewshed and scenic SID!! f~IiT .--------------...-0:.\ I ,,-;-- ... '1'~~ ............ a9cK ?~- --.,....~ , . ;0'1,1' I ----" : ~ 'u' analyses. A dramatic visual analysis of Boston from major highways in day and night conditions illustrates an extreme example of variable lighting conditions (Appleyard & Lynch, 1964). The significance of terrain aspect, sun angle, and observer position in viewing surface Figure 8: Solar position. features is illustrated in Figures 8 and 9 (U.S. Forest Service, 1972, p.l2). Aesthetic field research has shown that coastal haze and fog is a frequent factor modifying on-site visibility (Felleman, 1979). 9 SEASONAL LIGHT CONDITIONS _ GENEIl.iIL Color Hue Color Value Sun Angle Daylight Time Probable Light Variations Mean ShadO~1 Lengths J'Iinimum Maximum Lowest Shortest Medium Maximum transi tion period SPRING Maximum Medium Medium Increasing Increasing High Medium Decreasing SUJ>lh.i>R Hedium Medium Highest Longest Low Shor~est transition period F1ILL Potential Maximum Potential Maximum Medium Decreasing Decreasing High Medium Increasing Figure 9: Seasonal Light Conditions - General. 10 C. OBSERVER As noted earlier, a major consideration in designing a visibility study is to effectively select a representative set of 1ines-of-sight to be investigated. Particular views may be highly significant because of their frequency of occurrence, the unique content of the scene, or a combination of these factors. Frequency of occurrence is typically associated with concentrations of observers. In addition, for projects which will generate new viewers the analysis should include views of the project and from the projects. Figure 10 depicts "views of the road" and "from the road". Type and Quantity - Type and quantity of viewers is often interpreted from activity patterns such as residential clusters, recreation sites, and major vehicular routes. The U.S. Forest Service has identified three functional criteria for analyzing observers: number, view duration, and scenic concern (recreation, residential, other) (U.S. Forest Service, 1974, p.18). In an analysis of proposed cooling tower alternatives for a Hudson River power plant, analysts quantified and weighted residential, auto, rail, and boat viewers within the potential viewshed influence zone (Jones and Jones, 1975). In a transmission line study, the number of observers was factored by ..• "An attention analysis, (I NOl\MAL sumed to have a 360 0 potential viewshed, medium and high speed motion ~~:'.::""" ... ,- .... . ........t>- ...... has been shown to limit the normal cone of vision particularly for the driver. The concept of view cone is discussed in Section V. OBS~~V£r. SUPEIIJO/'l Figure 16: Viewer Positions. 15 Observer Environment - Observer environment includes both observer container (vehicle, building windows ••• ) if any, and the immediate natural and built landscape. Although conceptually these factors are a localized continuum of the Macro Landscape surface described above, it is analytically useful to differentiate immediate foreground objects. Both observer containers and immediate landscape are often design variables which may be studied in detail such as window orientation and screening plantings. Because these features may not be visually opaque or continuously solid, view "filtering" as well as interposition may occur. Many impact studies now incorporate seasonal "foliate" and "defoliate" visibility analyses. In addition, the accuracy and scale of data needs may be very different for immediate and macro landscapes. For example, a forest mass on a topographic map may correctly define a hillside midground skyline condition while the map may not have any indication of a single roadside hedge which effectively blocks or filters views from the route. D. PROCESSES Lines of Sight - All viewshed delineation methods make use of one or more line-of-sight techniques. These may be generally grouped into: field approaches, physical analogs, and numerical simulations. Field .51 'J. 5 MI. ~"'DCUGI!O . TABLE 15. Idealized Viewpoint Distribution: Natural Draft Cooling Tower Alternative (12 Final Viewscapes Required). Distance Foreground 0-1/2 Miles M1ddleground 1/2-5 ruOes Back.ground >5 mil es Observer Observer Observer Inferior Nanna1 Superior 6 2 Figure 17: Viewer/Distance Distribution. 16 approaches are the traditional in-situ "actual views". Modern adaptations include the use of airplanes, helicoptors and balloons, as well as photographic recording techniques to expand the scope and content of the method. Physical analogs primarily include interpretation of topographic maps by means of cross sections, vertical stereo air photo interpretation, and the use of terrain models utilizing periscope optics (model scope) or point light sources. The latter was briefly described in the introduction (see Figure 3). Numerical simulations utilize digital computers to "pass" line-of-sight vectors from the selected observer positions to intercept a numerical (x,y,z) approximation of the macro landscape. Recording - The locus of lines-of-sight must be recorded in a format which is compatible with the resource analysis, planning, or design. data needs. Limits of visibility may be recorded directly in the landscape by the placement of markers. More typically, plan view maps and perspectives are prepared which depict the viewshed limits and view content, respectively. It is important in processing data to articulate both the type and quality of view limit so that subsequent interpretations are properly founded. For example, recording should differentiate between moving and stationary views, the presence or absence of seasonal 17 vegetation considerations (such as filtering), and the geographic specificity (l1hardtl , "soft","ambiguous tl ) of the viewshed limit delineation. The latter is illustrated in Figure 18. Computerized resource studies typically require numerical inputs of visibility. These can range from a single +,- (visible, not visible) to sophisticated geographic matrices of "weighted scores" which incorporate the area of view, distance, slope/aspect, and number and types of observers (see Section IV-D). Although mapped visibility is useful to the interpreter, actual computer format is typically tabulated cards or tapes. Perspectives are highly useful in illustrating the content of scenes due to the ease of reader legibility (see Figure 19) (Roy Mann Associates, Dec., 1975, p.77). • ~ , .. "0• • • • _ PL..AN PERSPECTIVe Figure 19: Plan/Perspective. ~AR:D ~ .::oPT ····,:;:,ii/,·.: AM516tUOU~ \\\111/111 Figure 18: Viewshed Limit Accuracy. 18 EIYVIROfYMEI'ITAL MODEL Dafa 7Yr"'~ '"'~ t' ~ ~~ .~ ~ VAL/.EY~ /"oWf~ED RIIG~ Figure 24: DMA Landform Truncation. 26 In discussing computers, it is useful to incorporate the process sequence: input, analysis, and output. Since geographic information can be grouped into points, lines and areas (polygons), Figure 25 depicts the variety of approaches currently available for developing digital terrain model base data. Quaternary processing entails superimposing a grid on the data source information and either manually recording, or electronically digitizing corner point (or centroid) elevations. 5. Pentenary Pentenary data can be developed in various ways. "Randomlt points (either statistically random or selected) can be digitized in x,y,z coodinates and a numerical surface program run to create grid elevations (Sampson, 1978, p.91). Linear contours can be digitized (x,y coordinates along the contours, one z elevation associated with each linear string) with subsequent transformation into a numerical grid (Aerospace Corp., 1977, p.5-1). An analytically powerful means of representing a three dimensional surface is to approximate it with a finite number of 27 ~~-B <.. c: .~ ~'"t V")\-!2 5fereo Air Phofos •~--------------,,--------~ c.. (l ~ ~ 1.1 2.7 25 LlfJeo,;" .surf'tlce Facet Contours Ar.!!tls ~.~------or------------~ 28 29 28 2.7- U- U 21- 2.8 29 .30 n 2.6 26 26 27 26 28 {] Figure 25: Digitizing Flowchart. //yPUT DATA COt1PtlTER AIIALY5/5 OUTPIJT 28 facets, each with internally consistent surface characteristics. This approach is widely used in industrial design (automobile bodies), and computer graphic shading (Newman, Sproull, 1973, part IV). In land form analysis a growing utilization is being made for slope, aspect, and watersheds (see Figure 26). A computer-derived numerical data bank can be made by inputting the polygon or corner outline of facet areas and associating general surface curvature with each area (Wagar, 1977). C. SURFACE FEATURES The significance of terrain surface features, vegetation, and buildings, for interposition in the study area should be carefully considered at the project outset. As the scale of terrain features, and/or the distance to observer positions increase, the significance of surface features in defining macro landscape limits of visibility diminishes. 1. Primary Field sketching and field photography are generally an inefficient means of assembling comprehensive surface cover information. Major difficulties may be encountered in transcribing such information accurately to a topographic base map. In contrast, vertical air photos (particularly stereo pairs) provide the most significant data source. Note, however, that field checks Figure 26: Terrain Facets. 29 :1'Q ~ '""" ~ CI) '"CI) . Ib' § "" .~ Primary highway, hard surface ..... ',' " ..... ", ....... ______ Boundaries: National ............... , ...............______ Secondary highway, hard surface ....... : ............. ~=_=_ State ........................................___ _ Light-duty road, hard or improved surface .............. ===== County, parish, municipio .................. , , , ... _____ _ Unimproved road ........• ' .... " ............ : .... ; .• ========== Civil township, precinct, town, barrio ........... ,,' ..___ _ Road under construction, alinement known .'....... ' .... ______ Incorporated city, village, town, hamlet ......." ..... ___________ Proposed road ...................................._--- Reservation, National or State................... , .__ . __ . 'Dual highway, dividing strip; 25 feet or less.'..... ; . "'" ..'_____ Small park, cemetery, airport, etc................ . :Dual highway, dividing strip exceeding 25 feet.. .......... ===== Land grant .....................................__ .. __ .. I Xrail. ....... ., .................' ..........,....: .....,,-'--..;.-------- Township or range line. United States land survey .. i ... 'FraJlr~ad: single·track·and multiple track ....... :':-.. :~<, .:.=::::=: :Ra'ilroads in juxtaposition .............. ," ...... :.: ., ._....___ Narrow gage: single track and multiple track ...........::::::::::::::::: Railroad in street and carline ........'.............. , ...., =====Bridge,/road and railroad. , .........' ..... ',; ,... ,; .' .:.. :~.... . . . Drawbridge: road and railroad,', . , ...... : : ..... :'. ":'. .. . Footbridge. . . . . ... . . . .. : ... '~ ..... : , ..; .... : ..... =.:~L=:;:: . Tunnel: road and railroad ...._.... ~"""': '.;,'. '-'-".' ".:. :"'~~-r="=E-- Overpass and underpass .............. ',',. ,,~, .. " :, ',: '. .. I II ,11=t=. Small masonry or concrete dam ............... ; .".". ,. y ", Dam with lock............ " ......_..........'........ . Dam with road. : ......... : ... .- ..... ;. , ... ; . :"..... . ~.Lc:.l. i Can~1 with lock ...'......... ','" . :_.< ....,...<.........'==':=0<===. Buildings (dwelling; place of ~mployment, etc.) ..... , ...••1411 . rt~:Cem: School, church, and cemetery ........ , ........-.......n ~ ..~ :__....._~ Buildings (barn, warehouse, etc.) ............... ;', ... :. ac:::JI. Township or range line, approximate location .......... - _______ Section line. United States land survey ...............______ Section 'line, approximate location ...................._______. Township line, not United States land survey ..................................... Section line, not United States land survey ......................--.-........--.~ i Found corner: section and closing ....................-+ -- - T i i Boundary monument: land grant and other. . ....... 0 ..••_•...••. 0 'Fence or field line ................. ~· ............... _____ _ Index contour ........~ Supplementary contour Fill.......... ....... \\I/I! '\J Intermediate contour.. ____ Depression contours .. ~ Cut. . . . .'. . . . . . . . . . .. ,Y!ifiJ '" Levee ...............,....."...................... Levee with road ...... "''''''''''''''' Mine dum p. . . . . . . . . . ~,,:,:~ Wash . . . . . . . . . . . . . . . ~~?:W.\~:;\ ::/:~.;: Tailings .. '.......... J:i:·::t~J~"~(:S~':: '::",1 Tail ings pond ....... J~~~~~~~~~~~~~~~~~~~~~ . ~~'\",:"~i~~~:'::;lShifting sand ordunest/.~;;:)\.<;] 1°_._ 1 .''\" ( iIntrrcate surface .. _..... ~. '-.', /'" Sand area ..... . ~y; ;)'1',· ,;1 Gravel beach .........g~}}'!;:?~~::~:%@;.fN I Power transmission line with located metal tower...." ...•~___•____a____; . ~ I . ~ __ '.' ' . " .Perenmal streams .... ~, ntermlttent streams .. _'.. _........:...:Telephone line, pipeline,'etc. (labeled as to type) .....'.. ..___...:...___A: . d A d I Elevated aque uct.... que uct tunne ..... '__'P~""''''~!-- Wells-other than water (labeled as to type) ............. oOil. .... oGas Tanks: oil, water, etc. (labeled' only if water) ...........~_ •• @Water Water well"and spring.o ....... g.,..... 'Glacier........ . .-----. -·~-~-~_I/ Small rapids ......... ~ Small falls .......... ~ Located or landmark object; windmill ........... '.' '." '.' o.... ; ...... !. Large rapids ...... . ..;~ ~ Hf: Large falls. . . . . . . . . . .. ,~ Open pit, mine,"or quarry; prospect ...... . . . ~". ~ .. ': ..x ......... x Shaft and tunnel entrance .............._.............If ••••••••• .I Horizontal and vertical control station: Tablet, spirit level elevation ..... "..... . BMA5653 Other recoverable mark, spirit level elevation. A5455 Horizontal control station: tablet, vertical angle elevation VA8M t:.9519 Any recoverable mark, vertical angle or checked elevation tJ.3775 Vertical control station: tablet, spirit level elevation.. .. 8M X957 Other recoverable mark, spirit level elevation. . . . . . . . X954 Spot elevation .. Water elevation. . . x 7369 x 7369 670 670 Intermittent lake ..... C~~~::) Dry lake bed ........ . Foreshore flat ....... _..o';':~:;:).t~:r~~:\ Rock or coral r~ef .... _ r(/I''1.,., Sounding, depth curve. --.!!!.-..--------... Piling or dolphin .. Exposed wreck ... ..... Sunken wreck ..... . Rock, bare or awash; dangerous to navigation ..... . "'" '" 1- .....i1. Marsh (swamp)..... .t, -,,,,"-.;; Wooded mars.h ....... _ Woods or brushwood .. Vineyard ..... Land subject to controlled inunda tion Submerged marsh. Mangrove Orchard. Scrub ..... Urban area • .... • (t.:: 1'::;'::::::('::'::::1.~.: ..::..:::;.::. ~ 1 ,,,~.,,,,~.--,,,,,-.-.-..-"-::--~......~,~-- ,,-.~,~-..-..-.-,....-:-~--,...,.~...-....,..-..-'-.--.-.~-:-----,-,...~.-.....,--- ... -...-.--~..-- ..~....--.---.-...--- -.. -.~-~--:----.-~"---- are highly useful in developing a correct photo interpretation "key" for categories such as vegetation type and height (Reeves, 1975). An important use of stereo photos in New York is to update the L.U.N.R. map interpretations (see C3 below). 2. Secondary The U.S.G.S. topographic maps contain a rich spectrum of cultural and natural features. An example is shown in Figure 27 (U.S.G.S., 1972). The user should be cautioned as to the date and accuracy of this information which is noted in the map legend. In New York State the Department of Transportation has made a statewide update of political and cultural features at the identical U.S.G.S. 7 1/2 minute map series. Maps are available as planimetric or as overprints on the original U.S.G.S. topography from the Department's Map Information Unit in Albany. 3. Tertiary The New York State L.U.N.R. system is an excellent example of a rich surface feature data source that is increasingly available to the visibility analyst. L.U.N.R. is an automated data bank that was constructed in the late 1960's to provide an information base for multipurpose local, regional and state planning. 1968 and 1969. air photos 31 were interpreted for categories of land use and natural resources. The interpretations were manually transcribed to transparent overlays which fit the U.S.G.S. quads. However, visibility and visual character were not one of the system's application objectives. The L.U.N.R. overlays contain the outlines of photo interpretation for point, linear and aerial information types (see Figure 28) (N.Y.S. Office of Planning Services, 1974). This mapped data is avai1able in print or overlay form at the U.S.G.S. quad sheet scale. Both mapped data and numerical grid data are available to analysts. The former, although dated, continues to represent a major data source for many current impact assessments. The L.U.N.R. categories mapped were not developed for visual analysis. Thus, there may be a range of visual character types encountered in a single land use type such as single family residential. Study area field checks may be necessary to clarify this situation. Numerous surface type classifications are developed for national, state, and local planning and project purposes. The advent of a national land use and land cover system keyed to the U.S. Geological Survey base maps will set the framework for future analyses (9 general, 37 specific categories) (U.S. Geological Survey, 1978). - n// Xt/· P/ ~/ P r I '" ·3 P I . o~ rj f. ~ '-, i"'x @ if ~1<.1Ol "- Pl=;< f' P-fH-.... 'f).. KCOl"--' II , ./K~ " I "-- \ II Y ~ (,KG) f' / t- ~ f<~ / n p / i-V *Jl.f' / i./ __ :A-- _" 1~-- --{-'i ~ a\?-4 ~ j.. P~ d(o)® d@ . ,-,-,.- ----_ .. I,r I C. 1'~ f\ Ir I~ II 1\ ...-I'- -l- Ie;:: \'; r-. I( ~ " I '" , h I-- 17 b 1'-I, R I..... - v c J.-' II 1\ IV( 1/)- U II " , r-. c. , H - ...... 1\1 It' K~ "I-- 1><" '), -I" 1""':: " '- tFigure 28: L.U.N.R. - Area and Point Data. 1-11 I~ 1J 1- -v, 1,- - 't --_.... I--' s 'Ll CJl ... 1: 32 4. Quarternary The L.U.N.R. system described above was designed to provide fully automated data and analysis assembly. A statewide 1 kilometer square grid system was superimposed on this mapped data and quantitative information was stored for each cell, by area, length, or number (Figure 28). In contrast, some recent projects have incorporated in a multipurpose data bank, land use and surface cover categories that are integrally related to scenic analysis. Applied research conducted at the University of Massachusetts (Fabos, 1976) and Harvard (Steinitz, 1978) utilized prior field and photography preference tests in assembling data, and building interpretive models. The Harvard work is noteworthy in its dynamic synthesis of surface types and viewing distance (see previous discussion Section II-Surface Features). The 267 land use and landscape types which are potentially visible in foreground (200 meters) are aggregated into 30 types in the midground (300 meters +), and 13 groups at "far" distances (Steinitz, 1978, p.29). D. ATMOSPHERE This is one of the most complex elements of the visibility model due to the rapid rates of change inherent in climate. 33 1. Primary Field observations can be made under varying day/night, and weather conditions to gauge generalized visibility distances associated with a predefined set of significant climatic conditions. 2. Secondary Charts and tables of solar position can be used to map seasonal, potential sunlight. The U.S. Environmental Protection Agency keeps visibility (haze, smog tables ••• ) information for metropolitan areas and major industrial regions. 3. Tertiary Weather bureau and airport and coastguard data is highly sitespecific. Extreme care should be taken in extrapolation of cloud cover and visibility data to remote sites. E. OBSERVER TYPE AND QUANTITY The importance of visual features may vary among observer types. As noted above, the U.S. Forest Service, in its Visual Management System, differentiates between recreation and nonrecreation travellers. The quantity of viewers is used by analysts to select important line-ofsight locations, and to weigh the relative importance of various views. 34 1. Primary Field surveys are a frequent method used by recreation and transportation analysts to characterize and quantify user groups. These approaches (surveys, questionnaires •.• ) can be directly applied to visual studies. Most public parks maintain visitor count records. Recreation, (Shafer, 1966), land planning (Zube et a1, 1975) and other researchers have developed scenery evaluation approaches which involve direct field (or photo) evaluations. 2. Secondary Frequent use is made of highway traffic counts to quantify potential numbers of views from the road. This is done by multiplying vehicle counts ~Buch as computed Average Annual Daily Traffic, A.A.D.T.) by a selected occupancy rate, such as 2.5 people per car, and factoring for daylight hours. Such an approach does not deal directly with user types, except where special counts are available. State, county, and some municipal highway departments maintain traffic count data for facilities under their jurisdiction. Where data for precise numbers of travellers is not available or necessary, the Federal Aid Highway Program's Functional Classification System is a useful (and comprehensive) proxy. All routes in the country have been classified for both urban and rural areas (Bureau of Public Roads, 1969). In New York, the State Department of 35 Transportation has mapped these classifications on the 7 1/2 minute (1" = 2000') planimetric base (Figure 29). sceu:j.c. . A Federally mandated hlghway eva1uatlon was conducted by each state in the early 1970's. (Federal Highway Admin., 1973.) In addition, many counties and municipalities have designed scenic routes. These play an important role in developing impact hierarchies. (Wirth Associates, 1976, p.7-16). 3. Tertiary For urbanized areas, land use maps, census data, master plans, and zoning may be utilized to approximate existing and future number and type of viewers. Metropolitan transportation studies include industrial and commercial square footage which can be extrapolated to estimate users. This information is, at best, approximate and should be presented with clear explanatory notes. A common problem with quantification of viewer data is the misuse of significant figures, and the lack of provision of an expected statistical range. 4. Quarternary Land use and transportation computer models are frequently used in simulating future conditions to assist resource managers in decision making. These tools can be adapted to provide gross viewer type and quantity data for a geographic study area such as an urban traffic zone. I I HEW YO,,~'gT"'TE DEPAnlMiNT 0" T~"'Hs..(mTATION: . 198D HIGUWAV FUNCTIONAL' ClASSIFICATldN _.SYSTEM LEGEND ROAOS URBAN CLASSIFICATION' RURAL l:l"Ass,FlcAT.,ON Inlalll.le ======::: .'" !'linel,,_, Arl~rilt Ihl,,~,iw,yl COflnpclltla link _ _ _ _ _ __ Non-(tIhUcljn~ Link. ...... _ _ _ PrinCipii ~'Itriil (SIIl'tIr . . .' C'IMIII'f-tin!illnl ', __ .. ___ _ Minar Anw,ill NOh·tOnMclint LInk· ..._______• Minor Aru,~1 .' CoilettDI Major Ctln~elDr.. -...:.-.......;..-'_______.. 'Minor Cnllrttor lOClI . .oilftDAfly . '. ," .' ......~.~ -, ". -.' Figure 29: N.Y.S. Functional Highway Classification. 36 F. OBSERVER POSITION AND MOTION The selection of a finite number of viewing conditions, from the virtually unlimited number of possible views is a major challenge of study design. Studies may contain important stationary observation points and movement paths, as well as "proxy" positions from scenic SEI-ECfeo VIEWING f'OslflON!!> elements. View an,alysis positions may be functionally selected, regularly spaced, random or continuous (see Figure 30). "Landscape Control Points" (a concept researched by Litton and utilized by Jones and Jones, Zube and others) incorporates a few selected viewing positions which provide f>,f"GULM\LY SPI\J:.fD POSITION:) spatially extensive, representative views of a variety of landscape types (Litton, 1973). Regularly spaced positions are frequently used in a grid format for computer analysis of areas, and in evenly spaced (distance or time) points along roadway and travel corridors such as /. 2. ¥ElK J. is( --....,k~·~kc.....-- RANDOML r 5PAC.r;o fU51rlOl'/5 scenic rivers. Randomly generated points have been used to assess "typical" views in a landscape for areawide (Boster, 1976,p.92) and roadway contexts (Viohl, 1977) (Figure 30). The approach of "continuous" view positions is often used in the analysis of views along movement paths . \_- (Figure 30) • . ~--~"'-----:---: CONTINUOUS f'Of>lTiON!l Figure 30: View Position Types. 37 1. Primary The selection of visual control points should be made (or confirmed) in the field. Stereo air photos are frequently used with topographic maps to prescreen locations. 2. Secondary Topographic maps are the usual base for designation of regular, random and continuous positions. 3. Tertiary These sources are used to supplement topography in selecting control points. Examples include: maps of historic sites and natural features, and maps showing future concentrations of viewer activities, such as a proposed town or park master plan. 4. Quarternary Data bank models can screen visually sensitive locations. The U.S. Forest Service has utilized its VIEWIT program to identify highly visible project impact locations (Johnson, 1974). Cells with high visibility can be designated as significant viewpoints for subsequent analysis. 38 G. OBSERVER ENVIRONMENT ~~The importance of observer environment data is a direct function of PEEP CANOPY the observer positions selected, and the 1ine-of-sight method to be used. For example, if a sensitive site is to be analyzed for views from adjacent public roads then existing roadside conditions are crucial. In contrast, ~ta regional location search for a utility route may omit all observer DOUBLE CANOPY environment data until narrow study corridors have been designated. 1. Primary Air photo interpretation with subsequent field analysis is the most j bSINGLE CANOPY accurate means of establishing comprehensive, area wide observer environ- WII7I ONE liD6E ment conditions. The range of potential diver~ity is illustrated in Figure 3la(Kunit and Calhoun, 1973, p.110) and Figure 31b (Hornbeck and j ~ LOkerlund, 1973). Vegetative conditions are temporal, thus often necesOOUS!£ fOuE sitating seasonal (foliate and defoliate) checks (See also Section V). 2. Secondary U.S.G.S. topographic maps (1" = 2000') are generally unsuitable for txaccurately establishing local observer-environment conditions. The 5JHaL£ C4HOPY complexity of local sites including new structures, road signs, individual trees, roadside hedges and walls is not included on these maps. As they become available, new orthophoto maps should provide an excellent base for .. ~5J/tfLf. EDaE analysis. Figure 31a: Observer Environment Conditions. 39 Complexity of The Visual Field High Complexity . Many and diverse elements widely visible in the cone-of·vision template. Medium Complexity Some elements visible in the cone-of-vision template. Low Complexity Few elements visible in the cone-of-vision template. No Complexity Either completely open or completely enclosed. Impact Background 3 2 1 o Complexity of The Visual Edge High Complexity Many types of edge and high complexity of form. Medium complexity Some types of edge and some complexity of form. low Complexity Few types of edge and little complexity of form No Complexity No visual edge or completely enclosed. Impact Foreground 3 2 1 o Foreground 2 1 o o Background 2 1 o o ~ ~;' ... . -- --. ~ -&--"-" Figure 31b: Complexity of Visual Field and Edge. 40 3. Tertiary Some surface condition maps, such as New York's L.U.N.R., were not directly developed for the interpretation of line-of-sight screening and filtering. Care should be taken in their use. Some nonmilitary research has been conducted on visual penetrations of forest types (Way and Knode, 1969) but in general, precise standards for such interpretation do not exist. 4. Quarternary/Pentenary Due to the typically coarse grain of computer data bases (grid cells 500 feet - 1 kilometer square), precise viewer environment screening/filtering information is generally not available. However, programs such as EDAP (Landscapes Limited, 1973) and OCTVIEW (Steinitz, 1978) can identify the "potential" for such screening. (See Chapter VI.) This potential, if important, could then be clarified using a primary or secondary method. 41 EI'IVIROf'lMEfiTAL MODEL Dafa Iype!5 t"1 ~ t1 ~tI .~ ~tr ." !I .~ J .:: ~~ ~ •~ d --[! lAndform Plan ~ SurfClc-e ~ ~ Per6pecf,·vef2 Cove.r ~ Afmo5,here- OIJTPUTl.(£ ~ \/.. .~ ~~ Env,ro"me~f ~ ,;;."'"',.."{ " r.statiol'O.r;1 MoVJl'~.9 Direct . PhC{5'-co.l Ano.10!l Digito..l 5IGHTL/t/[ PROCf55[5 ~[g~lID~[Q) mrboI]ffifuuD~ m~~p ~~lTIJ(Q)lifl~ A. INTRODUCTION A variety of approaches is available for viewshed mapping and development of perspectives. The analyst must frequently select a related set of methods that most efficiently produce the desired product. Approaches can be grouped into three categories: Directfield analysis, Physical Analogs - map and photo analysis, and Digital Analogs. Field analysis approaches utilize actual lines-of-sight, either from selected viewer positions into the landscape, or from a proposed facility location back to potential viewer locations. Air photo and topography methods include intuitive interpretation, topographic cross sections, and three dimensional models. Numerical techniques utilize computer programs to develop interpreted visibility maps and perspective ,plots of landscape scenes. Due to the inherent differences between "stationary" and moving visibility, the latter will be dealt with in Section V. 42 B. DIRECT A field observer can record limits of visibility directly in the landscape by means of physical signs (flags, etc.). An alternative is to record the view in the field directly onto a map. Litton's comparative work clearly points out the potential inaccuracies inherent in establishing the actual location of view limits. In his study observers tended to map viewsheds to the highpoint of the interposing landforms, not the military crest, thus overestimating visible areas (see Figure 32) (Litton, 1973, p.11). A detailed field study in the Lake Ontario coastal zone revealed mid and background 10cationa1 accuracy problems in low relief terrain (Fellman, 1979). Extensive field testing in the Netherlands, established that field mapping accuracy was limited to a distance zone of 500-1200 meters within which "space defining elements" (surface, small landforms) can be perceived in stereo (Vander Ham and Iding, 1971). Innovative applications of field methods have replaced manual records with film media thus permitting subsequent interpretation at another location. Balloons, helicopters, scaffolds and other techniques have also been used to simulate full scale views to and from proposed facilities, such as timber harvest outlines, proposed cooling towers, and micro-wave antennas. (See Figure 33) (U.S. Forest Service, 1972, p.8s.) I I I I I I I I I , I , , , I I I I I I ! I ! I I I I I I \ ~ I I I I ~I I I I I I \ I • I I I , I f~--------------~ OBS/:/lVEi\' Cache Creek Landscape Control Paint ~ Field Plol c=J VlEWIT Plot ~ Section Plot , MILES Contour interval: 200 feel Figure 32: Field Mapping Accuracy. 43 . '," "'" '. f Note: Each grid unit represents 2.5 acres .;Figure 33: Full Scale View Simulation . of harvest area map 44 C. PHYSICAL ANALOGS With the advent of accurate topographic maps, a wide variety of offsite interpretation techniques have been developed to delineate viewsheds. The origin of modern visibility analyses can be associated with the French military engineering development of cross sections to ascertain the spatial extent of protection from projectiles that is provided by a fortification. (See Figure 34.) The term "defiled" means: ..• to arrange, plan and profile (section) of a fort so that their lines should be protected from••. fire" -Oxford Universal Dictionary 1. Topographic Sections A cross section is a graphic depiction of the vertical and horizontal relationships of a three dimensional form which occurs along a preselected 1Icutting plane". For maximum clarity, sections are drawn as viewed perpendicular to the cutting plane. (See Figure 35.) The French military use of cross sections has a direct analogy to viewshed construction with viewer positions and straight lines-of-sight replacing artillery placements and projectile trajectories. The concept of "military crest" describes positions that ·provide optimal observation and gun placements to command adjacent valleys (Figures 36) (Greitzer, 1944). These occur on hillsides where a steep slope tapers to a LINE OF ~IGHT PROT[c.TIOI'I ZONE Figure 34: Projectile Trajectory. fJ..AN :fEeT/0rt A- A' Figure 35: Plan/Section. MJLITIlR'f CJ!ES' Figure 36: Military Crest. /NVISIBt.£ "-0/'0 J1~ 45 terrace or crown. In domestic applications such sites are often choice locations for land use activities to make use of a panoramic vista such as residential buildings and roadside rest areas. (See Figure 37) (Hough Stansbury) . A basic training text used in World War II illustrates the analysis steps: topographic plan view, construction of cross sections, location of view limits on sections, transferring of limits to cutting planes Figure 37: Scenic Overlook. in plan view, and interpretive outlining of viewshed. Note, the cross section method includes all possible sight-lines in the vertical cutting plane. Exaggeration of vertical scale, (often up to lOX the horizontal), does not distort the line-of-sight analyses, and is frequently used to enhance the visual interpreta~ion (see Figure 38 a,b,c,) (Greitzer, 1944, p.112). When conducting a cross section analysis a primary concern is how many sections are necessary. This decision will determine the number of points that are ultimately connected to delineate the viewshed. A standard approach utilizes sections every 100 for the entire 3600 potential view cone (Greitzer, 1944). Other project studies have been conducted with even 50, 7~0, 150, 300 and 450 spacing. Ao alternative is for the trained analyst to individually select cross section locations based on 46 Figure 38a,b,c: Viewshed from Sections. p -~.---+---~y----- ~~===~~Q~====~,:,§~L____-iWL-____________________________~ -Finding defiladed areas. -Visibility diagram. 47 a review of terrain features (Litton, 1973, p.13). This approach appears to provide reasonable accuracy along with potential reduction of effort, as based on N.Y. Sea Grant test studies (Felleman, 1979). Line-of-sight cross sections may be adapted to incorporate all elements of the visibility model, including viewer environment, atmosphere, and surface features. This typically entails supplementary data in addition to that normally contained on topographic maps. For example, a field or air photo check of forest height could be used to interpret vegetative mass in the midground cross section. Detailed cross sections may be subsequently used to construct three dimensional, and block diagram views of the landscape. These are very useful in scenery content evaluation. If only potential topographic viewsheds are required, the work involved in constructing generalized sections can often be reduced through the simplifying trigonometric principle of "similar" (proportional) , triangles. This method examines only the critical limiting line-ofsight in a vertical cutting plane. In the accompanying figure 39, hill B will only be visible if the slope (tangent a=i ) of the sight line is positively increasing, that is the ratio. ,of Figure 39: Similar Triangles. 48 Using this relationship, approximate locus points at the boundary of a viewshed may be rapidly plotted. The U.S. Army used this technique in World War I, both as an algebraic relationship, and as an analog model ("rubberband" proportions). (Pearson, p. 62). Graphical proportions, plotted directly on the topo base map were recently used in a major power plant study MOO!!1. (Battelle, 1974). This approach can also be developed into an analog calibrated mechanical jig which is applied to the topographic base. 2. Topographic Models Scale Models have long served analysts as a means for visualiz- I'I.AN of HILL «Al...e 1"-"'=10'-0· ing three dimensional environments. By cutting layers of material cardboard, etc .•.. ) for each contour, a simple terrain model may be SEc.'nON A-A' constructed. Vertical exaggeration can enhance visibility analysis C~- stC;(jON as shown in Figure 40 (Salisbury, 1975). More elaborate sculptural techniques are available. Two general methods of simulating sight lines are used: a point light source at the. observer or object position(s), and direct viewing of the model through a model scope. Using a point light source, the bright area delineation is Figure 40: Topographic Model manually transcribed to a topographic basemap. A photograph may be 49 made of the illuminated model (see Figure 41). Theoretically it would also be possible to coat the model with a light sensitive emulsion and permanently record the visible area(s). A second approach is the use of a "model scope," a special magnifying periscope which allows the observer (or a camera) to view the model "approximately" as a site visit would permit. A probe is moved through the model tracing the limits of view. Recording of viewshed limits can be done on the model, on an adjacent topographic map, or by photographs taken through the model scope (see Figure 42). N.Y. Sea Grant research demonstrated that both techniques give quite acceptable results in complex terrain (Fe11eman, 1979). One of the most elaborate model simulation studies undertaken has been conducted at the University of California-Berkeley. In addition to topography this model includes scale vegetation, buildings and street furniture. A computer controlled model scope camera is used to simulate movies of auto trips throughout the- study area. Psychological tests based on field movies have yielded similar viewer reaction/ results to the simulated trips (Appleyard, et a1., 1973). 3. Air Photos Stero air photos can be manually interpreted to define landform surface cover, and contours (although optical and automated means are Figure 41: Illuminated Model. Figure 42: Modelscope Photo. 50 usually used to photogrametrica11y produce accurate maps). Stereo interpretation is an efficient means of locating scenic vista points (military crest type locations in rugged, and/or unmapped terrain. However, adapting the "floating dot" technique, used to establish contours on a horizontal plane to assess 1inesof-sight which are usually at a vertical angle is a complex undertaking. Researchers conducting a forest road study concluded: To determine whether or not impacts could be seen from a roadway, a "floating line" technique (same principle as the "floating dot" technique described in most elementary aerial photogrammetry texts) was tried on stereo paired photographs .•. this was found too time consuming ..• especia11y when the floating stereo line crosses more than one stereo pair (Potter and Wagar) • Analysts did find that this approach was useful in "checking" local viewer environment with 1:24,000 scale photos, and in large scale preliminary mapping with 1:250,000 scale imagery for their study area in the Pacific Northwest. 4. Inspection Often in the initial stage of viewshed mapping, it is necessary to roughly estimate the potential viewshed in order to efficiently select and utilize viewer positions and alternative 1ine-of-sight techniques. It is common for a trained professional or technician to 51 use stereo photos and topo maps in an informal manner to rapidly develop an approximate viewshed map and identify locations where a more detailed approach is needed D. DIGITAL ANALOGS With the recent advent of readily available computer hardware and software, many rapid developments are taking place in the area of digital terrain models, of which visibility is one topical area (American Society of Photogrammetry, 1978). The following is a brief highlighting of the basic concepts of automated simulation. A widely diverse group of problem solvers are concerned with utilizing computer analyses of three dimensional forms. For example, space scientists simulate complex rocket and satellite docking maneuvers, highway designers "test drive" a proposed road to check for unsafe visibility conditions, while architectural engineers design complex structural framing systems. As described above (II Data Assembly-Landform Quarternary), many digital line-of-sight programs utilize a matrix of elevations. These programs are generally known as "hidden line" algorithms (Newman, Sproull, 1973, Ch.14). 52 1. Sight Lines Generalizing, these programs efficiently compute, compare and store results of the visibility proportion (see Figure 39) between a designated viewer (or object) position and all "relevant" terrain grid points. In a large data base, the number of calculations and store requirements are significant enough to effect cost and hardware storage capabilities, particularly for mini or micro computers. One aspect of this problem, sight lines without intermediate points, is illustrated in Figure 43. A rigorous solution would entail interpolating between D and E to find elevation X, and then using the sight line proportion to ascertain the visibility of H. Numerous linear programming approaches have been developed to efficiently "scan" the terrain, and sequence the equations and temporary storage (Travis, et al., 1975; Tucker, 1976; Steinitz, 1978; Tomlin, 1978). Of particular interest are the questions of critical sight lines, and maximum view distances. The programs are typically only capable of analyzing cells, thus the maximum density of analytical coverage is a function of the cell size. Note there is an inverse relationship between plan view and perspective area in view cone distance zones (see Figure 44) (Landscape Figure 43: Intermediate Sightline Points~ 53 EFfECTS OF DISTANCE FaT distance 5O~. land area Middle distance 40% Land area Foreground 10,..Land area , Actual land area on ~8n viewpoint Views recorded on pfen and seen in perspective Far distance 10% yiewarea Middle distance _ 40 %view areB Foreground 5O~yiew area Apparent land ar88 is reduced due to perspective. Tone value and te;l(ture arB also reduced with the effect of distance The distant view is therefore less p4'orninent. Figure 44: Effects of Distance. Evaluation Research Project, 1976). One interesting approach to reducing computations is to limit sight line directions by selectively eliminating intermediate pOintssuch as "X" illustrated in Figure 45. This approach includes all cells adjacent to the viewer position, and a "sample" whose userselected density decreases as distance increases (see Figure 45) (Steinitz and Paulson, 1975, p.184). 54 053 052 051 g~3 048 047 046 045 044 043 042 041 040 039 0:38 037 036 035 034 033 032 031 g~3 028 8~l 025 024 023 g~f 8f3 8f'016 81~ 013 012 011 010 8§J006 005 004 003 002 001 oogoooooooooooooooooooooooooooooooooooooooooooooooooo 00 00000011111111112222222222333333333344444444445555 12345678901234567890123456789012345678901234567890123 , 3/26 3/9 , •••• 3/26 / / 5/26 • •••• •• • ••••.. ----'- 3/9 /9126 '- 5/26 9/26 " 00000000000000000000000000000000000000000000000000000 00000000011111111112222222222333333333344444444445555 12345678901234567890123456789012345678901234567890123 LB3END: X/Y X = Rays per Octant Y = Length of Ray Fi gure 4 Figure 45: Selected View Lines. 053 052 051 050 049 048 047 046 045 044 043 042 041 040 039 03<1 037 036 035 034 033 032 031 030 029 028 027 026 025 024 023 022 021 SfS018 017 016 015 o 4 013 012 011 010 oog 8B7006 005 004 003 002 001 55 2. Sight Distance A common method (both computer and manual)-of limiting-analysis is to place a maximum effective length on sight lines. This decision can either be based on the length of views that typically occur in a landscape, or the threshold cognition distance associated with a project scale. In low relief or high local enclosure environments the potential for distance views is slight. Dutch (Vander Ham and Iding, 1971) and British studies indicated that many analyses could take place within a 1 kilometer cell (see Figure 46) (Landscape Evaluation Research Project, 1976, p.84). 56 ~ / r-... ( '\ \... I 1\ L--I.J 1'1 ~ lkm :t o Area seen outside the grid square can be ® Area of the grid ••en fromwithin(trompointAJrnay ®ThO maximum extent of vision has greater than that s~en within it. be I.u thin that seen from outside the square an irregular boundary which can cover (frOfTI point B). A (~The 8xtent of vision can o\lorlap for views from tWQ squares. A&8 Figure 46: Extent of Views. ® The area. of de.d ground within a common field of vi.ion from two squares will be different 008ad ground from A visible from 8. . ClO.ad ground '!om B visible from A. 57 Transmission line studies, both manual and computer, have related structure size to significant view distance. In the Adirondacks a hierarchy of roadway distance zoneS was established (see Figure 47) (BRI, 1975, p. 30). In a low relief prairie landscape, view searches were limited to adjacent 1/4 km cells (Landscapes Limited, 1973); while in the northwest a maximum of 6-10 miles has been used (Jones and Jones, 1976, p.84). Figure 47: Roadway Distance Zones. • Impact reports have related the angular size of proposed objects in the visible field to minimum cognition (and impact) levels, thus establishing maximum analysis distances. o In a land-based study, 10 was used for both horizontal and vertical viewing angles (see Figure 48) (Steinitz, Rogers Assocs., 1977, p.2l8). For a pioneering water-based study, 100 was used for a horizontal threshold and 50 for a vertical threshold (see Figure 49) (Roy Mann Assocs., Inc., July, 1975b, p.294). This latter approach is consistent with psychological studies which have shown our increased perceptual sensitivity to vertical objects placed in horizontal fields. 59 AC = 1.2 miles = horizontal viewshed distance BC = .7 miles = vertical viewshed distance Therefore: 1.2 miles is the viewshed 'distance in this example. .3 B A Hori zonta1 View: Measurement .5 Viewshed Perimeter Structure .6 viewshed extends .3 miles vegetat' n Vertical View Measurement .t: t:/::'" .:./:f '.~.';.:~.: . '. .,:::':' .!.~ 1.2 (."::': ::..'.' :: ····C':. 1.1 ::':::.,. \.. 1.0 .::. ~:~ructu;.~··:::::·.. '::::" spoi 1. .::::::.... ....:.::.. .•..:..:.::.:.:.. •..••:::..:::...::•...•:....:... vegetation c Figure 48: Threshold View Angles. 60 " difttraction zone depth of view field line significant viewable surface ~;;::::::Jl::::~::~~~~~::~~~(~v~a~r~i~e~s~a~c~c~o~r~d~in~g~t~o~distanceu -----1'-- from observ"r) ,I view lim:l.t. c rea t int.·rmcdiate Ct"l1 t .....Inull.n\ll~ Aignificant sight line centroid 3. Lines and Surfaces disposal facility ---------j ~-------­~~minimum significant sight line Figure 49: water Based View Angles. ~---J Computer graphic systems can process points, lines, and areas. The above described search and comparison method results in the identification of visible or non-visible points. Three dimensional warped grid drawings, such as shown in Figure 50, incorporate a hidden line algorithm I section plan where each pair of adjacent points is checked. In the PREVIEW output Figure 50: PREVIEW Terrain Plot. 61 three conditions are processed: both points visible-plot line, one point visible-plot 1/3 of line from visible point, both points hidden-no plot (Myklestaa and Wagar, 1975). A more comprehensive approach than lines is surface facets. The widely used VIEWIT program associates a single elevation for each input grid cell. A wide range of internally generated (pentenary) data can be developed including slope, and aspect. These are computed by "fitting" a plane through the eight adjacent cells (see Figure 51) (Travis, et al., 1975, p.12). The "z angle" subroutine enables the user to specify a minimum vertical an!;\le with horizon "below which it is assumed the observer cannot see" (Travis, et aI., 1975, p.lO). Using the "ASPECT WEIGHTING" subroutine analysts can differentiate 10 different ranges of cell aspect relative to observer position based on visible size area of seen surface. This technique is particularly significant in identifying "visible" areas with low surface content information (see Figure 52) (Travis, et aI, 1975, p.16). Figure 51: Slope Fitting. OBSERVER· POINT CELL _._~ .~CELLA CELL EI Figure 52: Relative Aspect. p 62 · EIYVIROIYMEtiTAL MODEL Dafa -..f!. /.andform Plan ~ 1--::---;;------t--+----1f---+-_+___ Movil"l Direct 5IGHTL/t/[ PROCf55[J t2 ~O@M~IE[Q) ~[L[]W~u~ [}Vm@)tl:E~~ ~ ~(Q)m~@ In reality, all viewing is done by a moving sensory system as we constantly scan the environment with our eyes. Head and body activity increases the complexity of analyzing actual viewer behavior. It is generally accepted that pedestrian activities can be approximated by one or a set of stationary view points with a 3600 viewing potential. At the other extreme, years of driver behavior research has established that viewing is limited and focused for drivers at moderate and high speeds. Additional research remains to be done before we fully understand viewing phenomena at slow speed such as bicycles, urban traffic, boats, as well as for vehicular passengers at high speeds. A. VIEW CONE The "view cone" concept states that as speed increases: 1. The focal point moves away from the viewer; and 2. The effective cone of detailed vision narrows. This is shown in Figure 53 (U.S. Forest Service, 1972, p.112). Figure 53: Driver View Cones. 63 Each of the stationary 1ine-of-sight methods discussed in Section IV is applicable to the analysis of movement paths. A basic study design decision is to either approximate a continuous experience by means of a series of stationary points, or to attempt to directly assess a "continuous" experience. (See Figure 30.) Often a combination is de- sirable. Theories of spacing can be related to types of scenery, overlapping views, and speed of viewer. A study on Cape Cod incorporated stationary positions every 0.25 miles (Hornbeck and Okerlund, 1973). A manually conducted scenic river study used cross sections every 0.1 mile (Pitz, 1977, p.84), while a computer based river study used 450 points in 149 miles of river at major changes of river direction and side valley slope and at intermediate locations (VanDyke, 1978, p.13). A northwestern forest road study utilized evenly spaced points 0.2 miles apart. In testing a field photography technique with four photographs (77 0 ) approximating a 360 0 panorama, they computed that 1.4 acres were not "observed" between any two points (Potter and Wagar) • The pho to"' were interpreted in the office to establish the viewshed on a topographic base. 64 Views from the road are critical in the "Landscape Control Point" method (Litton, 1973). These were preselected at uneven intervals. As input to a scenic quality analysis for Jamaica Bay, random numbers were used to select viewing locations and directions along routes in a waterfront study area. The resultant views were "representations" of the auto experience (Viohl, 1977, p.46). In the Hornbeck study noted above, a weighted cone-of-vision template was used to identify the driver's central focal area (Hornbeck, 1973, p.115). The template, reproduced in Figure 54, is placed at preselected analysis points on the centerline and the view cone is transferred to a base map. This is repeated for both directions of travel. A stationary line-of-sight technique such as cross sections may be applied within the view cone area. NOTE: nollo scale; 60 mph (/('sign speed ---...-/ xl ----" .... ,1 --------" Centerline of Alignment " --------........ ~__~Fi.i",,'gri,,'i""::;:.d I 'Hddl,g""",~d-4--,._~,"B"C"k""="i"",=d;-~(2/3 F.O.) '--~/3F.D.) iTo Hod",,) Focusing Distance; 1800 It Figure 54: Weighted View Cone. Computer techniques such as VIEWIT can generate a composite "number of times seen" map which depicts the cumulative viewsheds from a series of points selected along a route. The cone of vision and maximum sight line distance may be specified for each point. (See Figure 55) (Travis, et al., 1975, p.29.) 65. I I 2 2 3 ; 5 0 5 0 5 0 , 52 51 • ·50 • • .-'8 .7 ••• 5 •• • •OJ .2 .1 00 30 Figure 55: Times Seen Map. l8 5 .31 5 5 l. I' 15 • • 1010 l' • • 5 1010 33 510 5 5to10 l2 10 510 11 5 5 1I) 10 10 510 5 5 2_ 5 5 51010 28 5 10 27 5 5 5 5 5 510 10 I' 2. 5 5 10101010 5 25 51010 5 to 5 5 5 101010 I' 2' S 5 5 5 5 5 51010 5 21 5 5 5 5 5 5 5 5101010101010 5 22 5 5 10 5 5 5 510tO tOIl) 5 21 5 5 5 5 5 5 510 5 5 510 510 10 510 20 5 5 5 5 10 5 510 510 5 '5 5 510 10 5 5 5 5 510 5 5 5 5101010101010 10 5 18 5 5 5 5 510 5 5 51010 51010 101010 17 5 5 5 5 5 5 10 5 5 5 510 5t010 101010 I. 5 5 5 5 5 5 5 5 5 5 5 5 5 51010 to 10 15 5 5 5 5 5 5 5 5 5 5 5 5 5 555 5 51010 I" 5 5 5 5 5 5 5 5 5 5 5 5 510 5 5 10101010 II 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 51010 12 5 5 5 5 5 5 5 5 5 5 5 5 5 5 510 5 5 5 5 5 5 5 0; II 5 5 5 5 5 5 5 5 5 5 5 5 5 510 5 5 5 5 5 5 5 5 5 5 10 5 5 5 5 5 5 5 5 5 5 5 5 lOt 0 5 5 5 5 0; '5 5 - 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 8 • • 5 5 5 5 5 5 5 5 5 7 5 5 5 5 5 5 5 10 5 5 5 5 5 5 5 • • • 5 5 5 5 5 5 5 5 • • 5 5 5 5 5 • • • 5 5 5 5 3 • • 5 5 2 • • 5 5 5 I 5 5 5 5 I 1 2 2 1 , 5 0 5 0 5 0 5 B. CONTINUOUS ANALYSIS A standard approach for establishing "view from the road" is to field record on a base map while driving the designated routes. A comprehensive procedure for this method, developed in the Sea Grant research is shown in Figure 56. Office preparation included enhancement of base maps, development of a set of symbols to be used in the field, and the Hornbeck view cone technique (see Figure 31), applied continuously. Each road was driven three times by a two-person team (driver and recorder). The first time was for orientation (exact observer location is a difficult problem) and to check the preliminary view cone viewshed. Then the route was driven once in each direction at a moderate speed, approximately 2/3 the posted speed limit. The recorder plots view focused on near shoulder (note: this gives a slightly wider view cone than from driver position). Where deciduous vegetation was a significant viewer environment and midground feature, the field work was repeated for foliate and defoliate seasonal conditions. (note: a 1" = 1000' base map scale was found to be more suitable than 1" = 2000' U.S.G.S. maps). Limits of views were plotted for fore and midground. Where background views occurred, topographic cross sections were used to delinate visibility limits. 67 {)3(J5 i6poJro.phiC &5'" Enlo.'Jed fo /"'" /000' Dist4'lf VIew Cros~ -SediollS Figure 56: Continuous Road View Methodology. Of"fICE PRfPARAT/(J/'I flELD I,JORK ()fflCE PRODIlCT5 68 A variation on this approach would be to video record the trips and then analyze their content. State highway departments are developing photo logs of their entire system for management programs (see Figure 57) (Kunit and Calhoun, 1975, p.81). Other continuous methods include model simulations using motor driven model scope cameras (Appleyard, et a1., 1973); and computer based animation. The latter are of growing importance in highway safety design. Examples of visibility from moving positions are shown in Figures 58 (Litton, 1968, p.51) and 59 (Wirth, 1976, p.SVII-12) and Appendix A. 69 50 1 501 501 50 1 Figure 57: Highway Filmstrip. 501 50 1 1001 iEstablish Correct :Entry Cell Frame .for each Module 70 •••• ••••••••••••••------..---;;;.....•• •••••• ••••••••• •••••• Figure 59: Roadside visibility • ••••••••••••• TAMARACK CREEK •••- ..- . •••- . Roadside Visibility - OPEN:' TRANSMISSION UNE WOU1.D BE SKYLIGHTED DUE TO LACK OF tOPOOnAPHIC !. ~. BACKGnOUNO INFLUENCE OR RIDGE LOCATIONS. ~ODIFIEO - i~~~~rz.iW:c~g~~~~~ ~~g~'~~~:~E~~~~:~ ~~~~ FROM HIGHWAY. _ NONV/srBLE • VIEWS TO TRANSMISSION LINE WOULD BE BLOCKED OY TOf'OQF\APHY OR VEGETATION Figure 58: Roadway Visibili ty . MODIFIED 71 EIYVIROIYMEtlTAL MODEL Dafa 6~.o,,,,,.J'r ---' 0;.;, d ~ t' Cl <:I<5 ~ 1 ijCii "'~o.!!! > > ~ ."-' --'-'. --'-'"- 1 ---+-- I I,······J7GTH r " a. eu .. ~ '""0 .=",e> .5 iO :;-u~ :!!~ -- ~ ·S.!!! ~0 "'--.c > ~ > -I --53 • == "0 "! ~ u E 2 -;,; ...c "0 0 "'e>"0_,. ~·E ;; ~ J ~~ ~ .. .!!!~:> >0 ~ (/) 0 W II: of 0 ""0 0 ~ 0 0 ex> .....Wo Wo ...... ~ o g \.-r 73 I Visual Conditions Construction: \ • \ 0. '. 0 .... u c· .. ~10 '"0"' "0 .=....< ._ "0 "'''~ ;;" .!: ii "3_ -- ", .' :2~ .- ~ "'~ .- 0 ~0 cdl ~- "'.-> > .D > .. > .~ • ~ • = "0 "• ..U E .s ;; 0 "0.D 0 "'"2:: "0- ~-E ~~ ..~K Ii.! .~-0.> >0 III 0 w a: '