"The RangeView application is the result of a collaborative, ongoing knowledge exchange between stakeholders. . .and researchers. . . .This two-way exchange educates researchers on user requirements while simultaneously educating users on how to make use of the application and interpret its data and value-added products."

• Introduction
• Problem statement
• The current functionality of the RangeView application
• The adaptive research and development model
• Capturing and meeting evolving needs: RangeView and the Kano model
• Conclusions
• References
• Author information
• Additional web resources

Introduction

(Back to top)
Geospatial tools such as geographic information systems (GIS), the global positioning system (GPS), remote sensing, and spatial modeling have proven invaluable to environmental research. Moreover, because geospatial tools help in pattern identification and prediction, they provide an effective medium for participatory appraisal where expert opinion and local knowledge are critical to natural resource management. While geospatial tools are most frequently used to facilitate participatory research on or management of natural resources, it is also possible to start from the opposite direction, using participatory methods to create geospatial tools for natural resource management.

Within a larger Raytheon initiative called "Synergy," prototype online "Infomarts" are being implemented to demonstrate how Earth Observing System (EOS) satellite imagery and other commercial satellite data can be used to address real-world problems confronting users on the ground. The EOS Data and Information System (EOSDIS) Synergy project is intended to enhance the efforts of NASA and university researchers around the United States who are investigating potential applications of remote sensing data. NASA sponsors these efforts under a contract with Raytheon, the prime contractor for the development of the EOSDIS Core System.

RangeView, one of these prototype Synergy Infomarts, involves collaboration between natural resource managers (public agency personnel, ranchers, resource conservationists, etc.) and researchers at the University of Arizona specializing in remote sensing, rangeland and forestry science, wildlife biology, and information science. RangeView (University of Arizona 2002) uses ongoing participatory research and development to create a geospatial application requested by natural resource managers to assist in managing vegetation dynamics through time and over large areas . The web site has shown a steady increase in user visits (that is, uniquely identified web-browsing sessions comprising multiple hits) from an average of approximately 18 per day in March 2002 to 120 per day in 2003. Many users are now learning about RangeView indirectly, in part due to a local newspaper article which was later picked up and distributed around the U.S. by the Associated Press (Umehara 2003).

The participation of natural resource managers in the development of the RangeView application has been a continual, iterative process that began in 2001. It is an adaptive research and development model that is driven by user demand for value-added information products created with geospatial technology. The systems requirements defined by end-users evolve with each iteration of the application. Each time, users identify the range of options (and limits) afforded them by using RangeView for natural resource management decisions and provide feedback that will drive the next iteration of the application.

Thus, the adaptive research and development approach used for RangeView is user-driven rather than technology-driven. Admittedly, this approach poses a major challenge to researchers and developers, who generally need to identify the research questions or specify system parameters prior to conducting an experiment or developing software. However, and more important, it also leads to the development of a product that is more relevant to natural resource managers, better reflects the needs of stakeholders, and more effectively transforms validated research findings into practical applications that can be used beyond the research community.

This article is divided into five sections. It opens with an examination of the problems faced by natural resource managers who hope to use geospatial technology. In the following section, we summarize the current functionality of RangeView to provide context. In the third section, we document the participatory model used by RangeView to assess user needs and encourage diffusion of innovation. Fourth, we provide a framework to track the evolving application attributes (utility and performance) from the perspective of natural resource managers using the web site. Finally, we summarize the participatory process used to create RangeView and make the application responsive to those evolving needs.

Problem statement

(Back to top)
Natural resource management aimed at long-term sustainable use of public land in the western U.S. requires communication and interaction among natural resource agency personnel, ranchers, developers, and other land users. Bridging the knowledge gap among these groups in terms of new tools and technologies is key to effective communication and management. Natural resource managers have been making requests for value-added, spatially explicit, easy-to-use products derived from remote sensing data since the launch of the first civilian satellite in 1972. Government officials (ARSC 2000) and natural resource managers (Marsh et al. 2001, 2002) in the agriculture sector (Moran, Inoue and Barnes 1997) have echoed the constraints to adoption of the products.

From the end user's standpoint, the adoption of geospatial technology is predicated on the regular availability of timely, quantitative, validated, operational, and site-specific information products, and the direct involvement of end users in product development, implementation and evaluation. Traditionally, the response to these demands has been on a project-by-project basis, often without the continuity necessary to generate broader-based demand. For sectors where geospatial expertise is often unavailable at the local level, the project-by-project approach has done little to transfer technology or expand the overall user base. Successful natural resource management applications have tended to be spatially explicit (classifications, inventories, assessments, etc.) or have focused on change detection. While these applications are useful to natural resource managers for a specific location or a specific time period, it has been difficult to provide frequent, operational access to satellite imagery and other geospatial data that could support the monitoring of natural resources and related phenomena through time.

The need for landscape-scale imagery available relatively frequently is great. Of the 2.26 billion acres of land area in the United States in 1992, 590 million acres were considered pasture/rangeland and 740 million acres were considered forest land (Daugherty 1995; NRCS 2000). Vegetation change on these lands has been significant over the past century, with 62 million acres undergoing afforestation and another 70 million undergoing deforestation (Birdsey and Lewis 2003). Management of private and public range and forest land requires a systems approach that integrates the objectives of all land users: traditional ranchers and private foresters for sustaining economic use; ex-urban land holders for preserving life style and conserving open space, water and biological resources; and public land managers, conservationists, recreationists and ecotourists. The potential for conflict among varying uses over larger areas of dynamic landscape conditions emphasizes the need for an integrated strategy to monitor vegetation effectively at scales which would not be practical with medium- or high-resolution spatial imagery.

The current functionality of the RangeView application

(Back to top)
RangeView concentrates on the design and delivery of information products to help inform the decision-making process of natural resource specialists and allow them to better manage natural resources. RangeView is a web-based information system that facilitates access to near real-time spatial data including:

a) vegetation indices (e.g. the Normalized Difference Vegetation Index or NDVI; see Sidebar 1) derived from satellite imagery acquired frequently across large areas from the NOAA Advanced Very High Resolution Radiometer (AVHRR; see Sidebar 2) and NASA Terra Moderate Resolution Imaging Spectroradiometer (MODIS; see Sidebar 3) satellite sensors;

b) spatially distributed climate data; and

c) ancillary vector data layers (i.e. roads, jurisdictional boundaries, township/range/section lines, contours) to help locate a particular area of interest. It also provides analytical functions such as interactive graphing associated with the spatially explicit time-series data.

Link to Fig. 1, RangeView home page (~54K)

The RangeView web site is centered on the application of geospatial technology, but it also includes extensive educational materials and reports concerning time series data continuity and the relationship between geospatial products and ground-based vegetation monitoring.

The RangeView application provides a range of visualization and analysis functionality:

Choose a satellite sensor

Select imagery from several satellite sensors based on your individual preferences, image resolution, range of historical data, and number of vegetation measurements.

Zoom in to your area of interest

Use the overlaid map features to locate your area of interest and then zoom in for a more detailed view of the image.

Add map features to the satellite imagery

Overlay map features onto the satellite images, such as roads, cities, county boundaries, and contour lines.

Link to Fig. 2, NDVI maps (~43K)

Compare maps for different time periods, different views of vegetation greenness, different satellite sensors, or different spatial resolutions.

View several map windows at once to compare vegetation greenness at different times, in different relationships to past greenness, or with different vegetation greenness measurements..

Link to Fig. 3, Rodeo-Chediski Fire NDVI data (~56K)

Decide whether to view vegetation greenness across landscapes or relative to past greenness

Compare vegetation greenness across landscapes and through time. Assess vegetation greenness relative to an average greenness for the past decade or relative to a previous time period.

Select a time period

Select the time frame for which you wish to view the satellite images. Current images are updated regularly, and some sets of imagery are available from 1989 to present.

Link to Fig. 4, simulation of animations (~40K)

Animate!

RangeView will generate an animation of the customized images based on the time period you selected. Browse through the images one at a time or view them continuously. The animations actually run like film strips that the user can control.

Link to Fig. 5, PixelGrabber graph (~16K)

Compare an image to a graph of greenness vegetation

Select a small area on the image and see a graph of the values of vegetation greenness in that area for the time period you selected. This function, termed "PixelGrabber," not only accesses images from multiple dates, but also returns the pixel values for a specific geographic coordinate or set of coordinates across those data in tabular or graphical form.

Link to Fig. 6, images and precipitation graph (~56K)

Compare an image to a graph of precipitation and climate

Select rain gauges in your area of interest, and see a graph of precipitation alongside the images for the time period you selected. Monitor the climate (El Niño/La Niña, etc.) for the time period.

In order to support future developments to accommodate the growing number and variety of end users, the application framework behind RangeView has been designed to be extensible. To achieve increased flexibility, the application involves a tiered architecture that separates application functionality into three independent layers, effectively decoupling the business logic from the presentation and the database in the software architecture. In this way, complex flows are separated into more manageable layers and changes in one layer do not affect other layers. A standard interface protocol maintains each layer of the system, facilitates communication among them, and controls process logic. By sharing a standard protocol, system components can be substituted and the overall system can be revised irrespective of the technology used in the front end (the web site), middleware (the software used to retrieve data from the database), or back-end (the database). Wherever possible, reusable components are developed, facilitating collaboration among projects with different goals or audiences but similar development needs. This development strategy encourages team programming, which increases the efficiency of the process and the effectiveness of each application element. It also permits an adaptive research and development approach that permits end-users more direct input regarding not only initial systems specifications but also enhancements and modifications throughout the life of the project.

The adaptive research and development model

(Back to top)
While the application of geospatial data and technology in RangeView is quite powerful, it is the "voice of the customer" that determines what information is most important and how it needs to be presented, displayed, explained, and disseminated. This requires participatory input from stakeholders at several junctures in the research and development process.

The RangeView application is the result of a collaborative, ongoing knowledge exchange between stakeholders (natural resource managers from government agencies, ranchers, resource conservationists, wildfire managers and information officers, educators, etc.) and researchers (including the web and database development team). This two-way exchange educates researchers on user requirements while simultaneously educating users on how to make use of the application and interpret its data and value-added products.

This model allows midstream adjustment in development through responsiveness to ongoing evaluation and user feedback. The strategy for technology transfer is based on a "high tech-high touch" approach that integrates a web-based system of product delivery with informal knowledge exchange both among potential users and between users and the research and development team. Futurist John Naisbitt (1982) demonstrated powerful insight when he suggested, "whenever new technology is introduced into society, there must be a counterbalancing human response -- that is, high touch -- or the technology is lost. The more high tech, the more high touch." Traditional methods for promoting the diffusion of innovation in Cooperative Extension are ideal for ensuring "high touch."

Link to Fig. 7, technology adoption cycle (~19K)

The mechanism we are using for diffusion of innovation is based on early research in Cooperative Extension, something now called the "technology adoption cycle" that was first formally documented by Ryan and Gross (1943). The concept is a fundamental element of how innovations are diffused (Rogers 1995). It is hypothesized that the pool of those adopting a new technology might be segmented into a theoretical bell curve according to the time taken to adopt, as follows:

1. "Innovators," comprising some 2.5% of all adopters, are the first to use the new technology;
2. "Early adopters" come next, comprising approximately 13.5% of all users;
3. "Early majority" members are a large group, comprising some 34% of all users;
4. "Late majority" users follow the early majority and also comprise about 34% of all adopters; and
5. "Laggards," comprising some 16% of all adopters, are the last to begin using the new technology.

Identifying these different groups allows for conscious facilitation of interactions among early adopters and later ones, speeding the overall dissemination process.

This model has been embraced by marketing experts in some of the most successful high-technology companies (Moore 1999). By dividing the market or set of potential users into segments associated with factors that encourage or constrain adoption, it is possible to tailor marketing efforts to each segment. This foundation serves as the information base from which diffusion can be encouraged. For example, "early adopters" tend not only to be progressive and willing to take a risk, but also to be highly visible in their communities. If such early adopters are quickly identified and made part of the product design and system specification process, it is more likely that the "early majority" will learn about the innovation and trust that it may be useful to them as well. In this way, innovations are diffused to a broader audience of users through relationships with early adopters or "lead users". This model is still being used by Cooperative Extension today. Stephens (1991) reported that the most effective adoption of new technologies by producers resulted from initial demonstration work with "benchmark" early adopters who were always watched, observed, and imitated by other potential users.

For RangeView, two interrelated activities made it possible to understand user needs. The first was a systematic marketing research assessment of user needs that began with the state, local and tribal government sector already using GIS (ARSC 2000) and later focused specifically on natural resource managers (Marsh et al. 2001, 2002). Second, University of Arizona (UA) Cooperative Extension and UA/NASA Space Grant teamed up to create a "Geospatial Extension Program" designed to bring geospatial research and technology to the public, with emphasis on natural resource managers (Orr, Hertzfield and Thwaits 2001).

The survey of state, local and tribal governments focused on identifying factors that would encourage or constrain adoption of RangeView (ARSC 2000). This research demonstrated that:
a) data applications need to cut across sectors;
b) potential users of remote sensing products lack necessary knowledge;
c) institutional challenges surrounding the perceived costs and risks of investing in this technology inhibit its adoption; and
d) local "networks" for data and technical support are critical both for adoption and for sustained use.

Barriers to the diffusion of innovation and the adoption of high-technology data products are not exclusively the concerns of applied research and development groups. A "whole product" marketing methodology suggests that approaching the problem from the perspective of the user and the technology simultaneously can enhance the potential to adopt.

In response to the survey results, the UA team has conducted a series of site visits with natural resource extension agents across Arizona to inventory preferred applications, to assess educational requirements, and to identify early adopters or "lead users" who now regularly assist in systems specification and in evaluating intermediate products. The UA has also held a series of workshops and seminars with lead users and other stakeholders (Marsh et al. 2001, 2002; Garfin and Morehouse 2001). Finally, the UA has partnered with Marketing Intelligence, LLC, a marketing research firm that is helping define a business model for the web-based dissemination of satellite remote sensing data, value-added land resource management products, and other geospatial data to help public and private land use decision-makers to determine the factors critical to achieving long-term sustainability (Jain, Klewer and Baker 2001; Jain, Baker and Rusu 2002).

Link to Fig. 8, photo of RangeView workshop (~38K)

Numerous presentations, workshops and training sessions have been held to identify lead users, introduce intermediate products, and educate end users on how these innovations can be used as a decision aid to enhance their current management practices.

The process began with background research about current methods for vegetation management, followed by a series of strategy meetings with marketing and development teams to help define user groups and hypothesize needs. While a very broad list of potential user groups was identified, it became evident that many groups such as educators and recreationists would not adopt the technology until it was proven useful by practicing natural resource managers, including government agency personnel, ranchers, and Cooperative Extension agents.

The in-depth interviews and a more quantitative survey revealed fundamental differences among these three core segments of the natural resource management community. Each exhibited different needs and constraints, and yet the adoption process within the community as a whole showed clear signs of their interdependence.

Agency personnel, who could be divided into those directly responsible for land management and those with specific resource responsibilities, were viewed as the most likely to adopt RangeView technology. However, their need for the application was dependent on the capacity of on-the-ground land managers to comfortably use and comprehend the technology and associated information products. In particular, agency personnel wanted ranchers, who act as stewards of the public land allotments that they manage, to understand and have access to these same information products. Cooperative Extension agents, tasked with bringing university research knowledge to the public, were viewed as the bridge between the RangeView R&D team and the end-user community as well as a link between government agency personnel and ranchers. The individual needs of and the interrelationships among these three core segments of the natural resource management community helped us to define and schedule RangeView activities to account for factors that would influence the pace of adoption unique to each segment.

Capturing and meeting evolving needs: RangeView and the Kano model

(Back to top)
The assessment of user needs and its influence on the development process is not a static process. When end users are exposed to spatial data, invariably the nature of the question changes. It was therefore essential to make the research and development process adapt both to requests for modifications on current products or functionality and to the identification of entirely new needs. To help differentiate features, prioritize needs and support product specification in an evolving market, RangeView incorporated a customer quality model developed by Dr. Noriaki Kano (1984).

Link to Fig. 9, Kano model (~62K)

The Kano model was developed as a means to help businesses and organizations differentiate their product and service offerings from those of competitors. The basic foundation of the model is to provide an understanding of the correlation between the ability of an entity's performance of certain attributes or features vis-à-vis the level of satisfaction provided to the consumer. Customer attributes are often depicted on an X,Y plot where the X axis depicts increasing product performance and the Y axis depicts increasing product utility. According to Kano, there are three types of attributes; Basic/Expected (Minimum Required), Performance/Linear and Exciter/Delighters.

Basic or minimum required attributes are those that consumers expect the product or service to contain, and are, from the perspective of the consumer, so obviously essential that they are hardly even noticed. Though high performance of a basic attribute will not increase satisfaction much, poor performance or the absence of basic attributes will keep consumers from trying or continuing to use a product. At the current stage of RangeView development, the temporal dynamic provided by time-series animations which provide a film strip of vegetation greenness over time for a preset (and unadjustable) location fit this category. Early in the research process, it was determined that this feature provided a starting point towards meeting the demands of the targeted user groups, but it provided little in terms of differentiation from alternative solutions. As a result of these findings, development resources were concentrated into other areas.

Performance attributes are those in which the relationship between the ability to perform and the resultant consumer satisfaction is linear. Increased performance will lead to increased satisfaction, the reverse of which is also true. Consumer or marketing research efforts directed toward performance attributes are fundamental in the quest to differentiate one product from another in the eyes of the consumer. They are the attributes that consumers are simultaneously most cognizant of and most concerned about. For RangeView, making the time-series animations spatially dynamic provided performance attributes. Research has consistently indicated that the easier it is for RangeView users to control time and space in looking at vegetation greenness, the more likely they are to adopt the new technology. Many lead users found the spatially static time-series animations to be very interesting and capable of providing basic knowledge, but the ability to zoom into specific parcels of land and add different map layers, while simultaneously controlling time, would be key to driving their adoption of RangeView. Because of the potential for such features to differentiate RangeView from other solutions offered, a great deal of our development resources are focused on spatially and temporally dynamic animations.

The third and final set of attributes is called delighters. These have the ability to excite the consumer because they are not in any way expected. Delighters are very difficult for users to find and thus their absence does not have a negative effect on satisfaction levels. However, the mere presence of a delighter provides a high level of satisfaction to the consumer. Because of this, exploratory research funds are allocated towards identifying the delighters, introducing them, and then developing them to best meet the needs of the users. Adding additional data sources such as precipitation to the dynamic animations is currently the RangeView delighter. While many different possible features for future development were discussed with lead users, a consistent priority was the ability to tie vegetation greenness to spatially explicit measures of rainfall. Demonstrated in prototype form, this attribute excited the different targeted user groups. Dynamically graphing precipitation while vegetation greenness images animate through time is a function that still under development, but it will likely help in the effort to cross the chasm from the early adopters to the early majority.

A final caveat to the Kano Model is the effect of time. Because consumer expectations are always on the rise, a feature that may be considered a performance attribute today is likely to be an expected basic attribute tomorrow. Likewise, time will shift delighters to performance attributes. This is why the participatory process of developing these geospatial tools is not limited to an initial assessment of user requirements. The partnership between RangeView and its lead users is ongoing and their feedback continues to influence the research and development process.

Conclusions

(Back to top)

Link to Fig. 10, diagram of RangeView sustainability (~52K)

Sustainability of RangeView requires success in four areas:

1. identification of user groups and a thorough understanding of their needs;
2. development of adequate information products and delivery mechanisms focused on user needs;
3. development and delivery of a value proposition that is compelling enough to foster high levels of user loyalty; and
4. the ability to sustain the delivery of such value through execution of suitable marketing and project management strategies).

The initial marketing research conducted by Marketing Intelligence, LLC, revealed a particular need for an application that could capture vegetation dynamics through time and over large areas. In response, the University of Arizona set up a web-based information system that provides user access to near real-time spatial data (i.e. vegetation indices derived from satellite imagery acquired frequently across large areas from the NOAA AVHRR and NASA Terra MODIS satellite sensors, spatially distributed climate data, ancillary vector data layers to help locate a particular area of interest, and analytical products associated with time-series data). The initial interface took a number of different forms early on as the developers attempted to provide functionality to the lead users, who in turn provided feedback on needed modifications and enhancements; their feedback became more and more specific as they came to understand the data and information products and the functionality options that could be made available through further development.

Involvement of user communities in product development from the beginning of the RangeView project and soliciting periodic feedback to aid in enhancing its services and products were also critical for a broader adoption of geospatial data and technology. The vast majority of end users initially had little or no experience in computerized maps and time series satellite imagery. Most only became comfortable in recommending system changes after attending a presentation or workshop. By integrating the continual assessment of user specifications with a systematic educational outreach program, we were able to refine our understanding of user requirements and incorporate the architectural flexibility in back-end systems necessary to handle the new requests that will inevitably come as natural resource managers become more familiar with these data products.

By pursuing a participatory research and development approach, RangeView has successfully developed relationships with lead users, identified challenges that will need to be overcome in "crossing the chasm" from the visionaries and early adopters to the mass user community, and, at each stage in the process initiated and implemented product revision plans.

References

(Back to top)
ARSC. 2000. NASA regional workshops for state, local and tribal governments: Findings and implications. Tucson, Ariz.: Arizona Remote Sensing Center, University of Arizona. Online: http://aria.arizona.edu/report.pdf (Accessed June 2003).

Birdsey, R.A. and G.M. Lewis. 2003. Current and historical trends in use, management and disturbance of United States forest lands. In The potential of U.S.. cropland to sequester carbon and mitigate the greenhouse effect, eds. R. Lal, J. Kimble, V.C. Cole and R. Follett, 15-34. Boca Raton (FL): CRC Press.

Daugherty, A.B. 1995. Major uses of land in the United States: 1992. Agriculture Economic Report (AER) 723. Washington, D.C.: USDA-Economic Research Service (ERS).

Garfin, G. and B. Morehouse. 2001. 2001 Fire and Climate Workshop Proceedings, February 14-16 and March 28, 2001. Tucson, Ariz.: CLIMAS-ISPE, University of Arizona.

Jain, K., C. Baker, and A. Rusu. 2002. Synergy II InfoMart marketing plan. Tucson, Ariz.: Marketing Intelligence, LLC.

Jain, K., J. Klewer, and C. Baker. 2001. Synergy II InfoMart marketing plan. Tucson, Ariz.: Marketing Intelligence, LLC.

Kano, N. 1984. Attractive quality and must-be quality. The Journal of the Japanese Society for Quality Control 14(2): 39-48.

Marsh, S.E., B. Orr, A-M. White, L. Baker, K. Mauz, C. MacDonald, G. Oldham, H. Rodriguez, C. Wallace, B.S. Hutchinson, A. Thwaits, L. Howrey, G.B. Ruyle, P.R. Krausman, A. Heydlauff, D. Schafer, E. Trobia, J. DeVos. 2001. Final report: An Infomart for natural resources: EOS science results and data for public land management in the western states. Tucson, Ariz.: Arizona Remote Sensing Center, University of Arizona.

Marsh, S.E., B.J. Orr, W. van Leeuwen, A-M. White, C.F. Hutchinson, L. Baker, M. Farah, W. Grünberg, S. Herrmann, C-Y. Huang, C. McDonald, G. Oldham, H. Rodriguez, C. Wallace, B.S. Hutchinson, A. Thwaits, L.D. Howery, G.B. Ruyle, P.R. Krausman, A. Heydlauff, D. Schafer, K. Jain, and C. Baker. 2002. Final report Synergy III: Extending and sustaining infomart tools and data for natural resource management. Tucson, Ariz.: Arizona Remote Sensing Center, University of Arizona.

Moore, G.A. 1999. Crossing the chasm: Marketing and selling high-tech products to mainstream customers. Rev. ed. New York: HarperCollins.

Moran, M.S., Y. Inoue, and E.M. Barnes. 1997. Opportunities and limitations for image-based remote sensing in precision crop management. Remote Sensing of Environment 61:319-346.

Naisbitt, J. 1982. Megatrends: Ten new directions transforming our lives. New York: Warner Books.

NRCS. 2000. Summary report 1997 National Resources Inventory. USDA Natural Resources Conservation Service.

Orr, B.J., M. Hertzfeld, and A. Thwaits. 2002. Geospatial Extension Program in Arizona. Online: http://arsc.arid.arizona.edu/extension/ (Accessed June 2003).

Rogers, E.M. 1995. Diffusion of innovations. 4th ed. New York: Free Press.

Ryan, B. and N.C. Gross. 1943. The diffusion of hybrid seed corn in two Iowa communities. Rural Sociology 8(1-4) 15-24.

Stephens, L.L. 1991. Information Sources and methods used by progressive farmers in the learning process. Ph.D. Dissertation, Montana State University.

Umehara, Y. 2003. RangeView web site helps survey land change. Arizona Daily Star, March 2, 2003.

University of Arizona. 2002. RangeView: Geospatial tools for natural resource management. Online: http://rangeview.arizona.edu/ (Accessed June 2003).