Getting Started in Remote Sensing: Common Barriers and Bottlenecks
There was a remarkable convergence in the issues raised by the workshop presenters. In this chapter, the steering committee discusses the common problem areas identified by state, regional, and local government officials and looks at how public entities addressed them. Chapter 4 looks at how the federal government and the private sector facilitated the use of remote sensing data by state and local governments.
The workshop presentations emphasized that when public sector officials introduce remote sensing data and applications into existing or changing government operations, they must deal not only with significant technical issues but also with significant nontechnical issues. How these issues are addressed may determine whether a particular application of remote sensing in the public sector will succeed and whether remote sensing will be able to provide the information needed for management or decision making.
The issues fall naturally into several broad categories: (1) financial and budgetary constraints; (2) institutional, organizational, and political issues; (3) experience, skills, and training; (4) transitioning from photographic to digital data; and (5) licensing and data management.
A better understanding of the issues dealt with by governments that have already decided to adopt remote sensing may help transform the nonfederal public sector from what a workshop participant called one of the “largest untapped markets for remote sensing imagery” into a sophisticated, economically sustainable, technologically innovative sector for the use of remote sensing applications.
FINANCIAL AND BUDGETARY CONSTRAINTS
Financial constraints can have a major impact on state and local government operations. Workshop participants emphasized that fiscal constraints faced by jurisdictions in the nonfederal public sector are increasing. Many state and local governments are currently experiencing reduced revenues and even budgetary deficits. Because remote sensing is not often perceived as a must-have technology, the current fiscal situation in the nonfederal public sector could curtail or even eliminate any expansion of remote sensing activities in the sector.
The initial costs of developing an in-house remote sensing capacity can be considerable. There is, first, the cost of imagery itself. Though federal agencies provide data at a minimal cost and some types of data can be obtained free, the resolution of their data may not meet the needs of a public sector client that, particularly in urban and suburban areas, may require higher resolution. Such data must be purchased on the domestic or international market at commercial prices far in excess of the price of federal data.
In addition to data, however, there are other costs faced by state and local governments in developing an internal remote sensing capability, and these associated costs can add up to considerably more than the cost of the imagery. In the first report of this steering committee, Transforming Remote Sensing Data into Information and Applications, there was discussion of the many expenditures that must be assumed by a public entity intending to use remote sensing for the first time, and a recommendation that NASA should study the short- and long-term costs and benefits of developing remote sensing applications (see Appendix A). The steering committee reiterates the importance of this recommendation for informing state and local governments about the costs and benefits of using remote sensing data (Box 3.1).
These costs include, at a minimum, the initial cost and the cost of regularly updating hardware, software, and technical training for personnel. If trained personnel are not available on staff, introducing remote sensing may also include the cost of hiring new personnel with the appropriate technical and data skills or training existing GIS personnel. A viable alternative for many governments to establishing an in-house capability is to contract with a university or a private firm for remote sensing services. This approach may be expensive, however, and it does not build capacity within government for the horizontal diffusion of remote sensing applications.1
The contribution of universities in providing education and training to the remote sensing workforce and in creating new types of applications is discussed at length in the steering committee’s first report, Space Studies Board and Ocean Studies Board, National Research Council, Transform ing Remote Sensing Data into Information and Applications, National Academy Press, Washington, D.C., 2001, pp. 42-44. The executive summary of that report is reprinted here as Appendix A.
In its first report, the steering committee stated as follows:
The full, life-cycle cost of developing and using remote sensing data products goes beyond obtaining the data and includes, among others, staff for data processing, interpretation, and integration; education and training; hardware and software upgrades; and sustained interactions between technical personnel and end users. . . .
Recommendation: NASA’s Office of Earth Science, Applications Division, in consultation with other stakeholders . . . should mount a study to identify and analyze the full range of short- and long-term costs and benefits of developing remote sensing applications and the full costs of their implementation by public, nongovernmental, and other noncommercial users. In addition, NASA should support economic analyses to reduce the start-up costs of developing remote sensing applications.
SOURCE: Space Studies Board and Ocean Studies Board, National Research Council, Transforming Remote Sensing Data into Information and Applications, National Academy Press, Washington, D.C., 2001, pp. 3-4.
Capital or Operating Expense?
A related problem for a jurisdiction seeking to initiate or expand remote sensing activities is that it is not always clear how to budget for the purchase and use of remote sensing data. Remote sensing advocates in state and local government face problems in obtaining appropriate budgetary support for purchasing data. The Baltimore and Boulder County speakers at the workshop reported that they encountered uncertainty as to whether remote sensing data purchases should be considered a one-time capital expense or a routine operating cost. The problem could be exacerbated by the fact that the need for subsequent data for monitoring purposes or measuring change may not fit easily into the annual budget cycle of most public entities. Data may be needed annually or biannually or at some other interval that does not conform to or fit within the budget cycle. Public sector budgets are generally based on recurring costs; the purchase of a new type of data does not fit comfortably into a budget based on incremental deviations from the previous year’s expenses.
Finally, there is the issue of public sector procurement processes and their role in acquiring remote sensing data. Although public sector jurisdictions in the aggregate constitute a significant market for remote sensing data, the purchase of data is conducted by individual public sector units governed by complicated procurement regulations. There are 50 states and over 3,100 counties in the United States, all of which are potential customers with different data needs. There are many more towns and cities that might be interested in remote sensing data. Because the procurement of data for each of these units involves expenditure of public monies, it must be done according to processes that are transparent, fair, competitive, and based on principles that emphasize the lowest bid. Meeting such public sector contracting requirements can be a lengthy and costly process for both buyer and seller.
State and local governments have responded to these financial and budgetary problems in several ways. They have tried to justify the price of remote sensing data by comparing the costs of airborne remote sensing data with those of satellite data from the federal government and from the commercial sector and balancing cost estimates for specific tasks with anticipated uses of the data. Many public sector groups pool their data requirements and form cooperatives or centers that purchase remote sensing data for the use of all contributing member agencies or governments. This approach can reduce both the absolute cost and the transaction costs of obtaining data. This was the experience of cities in the Red River Valley, as well as Boulder County, Colorado, the state of Washington, and Portland Metro (see Chapter 2). Other governments, like that of Missouri, inventory remote sensing data that have already been purchased so that, if licensing arrangements permit, they can be used by the state for other purposes.
Procurement questions point to a more fundamental expectation about the use of remote sensing data in the nonfederal public sector. State and local governments may not anticipate new uses of the data beyond the specific project for which they were purchased. Though this one-time approach is encouraged by commercial licensing practices, one consequence is that public agencies do not then frame their data purchases in the broader context of multiple uses of the data.
A number of state and local governments, including Richland County, South Carolina, have reduced the costs of adopting remote sensing by building on the management capacity already in place for GIS and related geospatial data. By using existing staff and facilities, many jurisdictions can reduce the start-up costs of introducing remote sensing data. Not only do there appear to be economies of scale for the public sector entity that combines geospatial data activities into a single management unit, but there may also be benefits in being able to choose from among multiple types of geospatial data to select the most appropriate source of data for the information needed for management and decision making.
INSTITUTIONAL, ORGANIZATIONAL, AND POLITICAL ISSUES
Among the nontechnical constraints on the use of remote sensing data in state and local governments are (1) institutional issues related to the organization of remote sensing activities and (2) political and legal issues that can influence how the data are used. The workshop presentations repeatedly emphasized that state and local governments are subject to greater political pressures than the national government. It is not that party considerations govern public sector management. Rather, voters and elected officials are closer to operational decision making in states and localities than they are at the federal level. Elected officials are therefore more often able to influence expenditures on remote sensing, the uses to which the data are put, and the institutional structure within which remote sensing is managed.2
Use of Remote Sensing Data in the Courts
Problems with judicial acceptance of satellite remote sensing data as evidence in court cases can also influence the application and use of remote sensing in public sector entities. Because technological advances enable remote sensing satellites to identify objects smaller than a meter, the technology now permits detailed observation of personal movements, objects stored and used on private land, and even in some cases corporate production runs and inventories. This ability to observe has raised issues related to the right to privacy and protection from unreasonable searches. Perhaps because aerial remote sensing data have been available for a longer period of time, they have been upheld as evidence in legal cases, but space remote sensing data have not. Decisions about the use of satellite remote sensing data in court are expected to unfold as satellite imagery becomes more commonplace.3
One opportunity for managers to influence the use of remote sensing data is in response to Government Accounting Standards Board (GASB) Statement 34, which established in 1999 new financial reporting requirements for state and local governments in the United States. One requirement is that nonfederal governments are to report on the physical condition and maintenance level of such infrastructure as roads, highways, bridges, and sewers. According to one county representative, remote sensing information could be a valuable tool for conducting inventory surveys to comply with GASB Statement 34. Compliance affects a city, state, or county government’s bond rating and credit worthiness, an important factor for nonfederal governments that often finance infrastructure investments with municipal bonds. For information on the GASB and Statement 34, see <http://accounting.rutgers.edu/raw/gasb/repmodel/index.html>, accessed on August 6, 2002.
For more on the use of aerial photography in court cases, see the discussion in a note issued by the NASA Inspector General’s Office, “Remote Sensing and the Fourth Amendment: A New Law Enforcement Tool?” at <http://www.hq.nasa.gov/office/oig/hq/remote4.html> (accessed July 2002); Timothy J. Brennan and Molly K. Macauley, “Remote Sensing Satellites and Privacy: A Framework for Policy Assessment,” Law, Computers, and Artificial Intelligence, 4(3):233-248 (1995).
The workshop presentations emphasized the importance of having an internal advocate for remote sensing when developing a successful public sector remote sensing program. This could be a person with program responsibility who needs remote sensing data to do the job, or it could be someone with programmatic responsibility for geospatial data and technologies. Regardless of that person’s functional position in the government, his or her role as the internal advocate appears to be important both in advancing the use of remote sensing data and information and in fostering the building of capacity. Without such an advocate in-house, the initial budgetary, organizational, and technical barriers that must be overcome to establish the use of remote sensing data operationally can appear to outweigh the benefits. Moreover, the benefits themselves are usually not well understood. The internal advocate provides information to public sector managers and decision makers on what remote sensing can do and takes the initial steps to obtain data and the capacity to develop remote sensing applications.4
Organizational Location of Remote Sensing Expertise
Another institutional issue that must be addressed is where to place remote sensing expertise organizationally within the government. This issue affects the state or local government agency’s ability to share remote sensing information with other agencies, and in turn affects the overall visibility of remote sensing applications. There appear from the presentations at the workshop to be several viable options, depending on the nature of the applications needed, the long-term
strategies for using remote sensing, and existing government structures. It is necessary to address this issue even if a public sector entity decides not to conduct remote sensing work in-house but to contract it out to a local firm or university. In the latter case, there is a need for in-house remote sensing expertise for preparing, competing, and monitoring contracts. If an operational remote sensing capacity is set up within the public sector, remote sensing is likely to have a more visible and permanent role in both operations and policy discussion than when data are obtained through external contracts.
The presentations at the workshop suggested that state and local governments have responded to these institutional, organizational, and political issues in different ways. Most presentations emphasized the need to communicate regularly with elected officials, in addition to government managers and decision makers, about the benefits and uses of remote sensing data (a point emphasized also by the steering committee in its first report5). This is not only reasonable but also valuable in creating an interest in and support for remote sensing. It is even valuable to communicate with voters. Too often the contributions of this technology do not come to the attention of public sector leaders, or the role of remote sensing data is hidden or overlooked in substantive discussions of the information obtained from it. As one participant noted, no one ever lost an election because of remote sensing.
Some participants noted a tendency in speaking with elected officials and upper management to emphasize the pretty pictures produced by remote sensing at the expense of more utilitarian discussions of its applications. Although the images have an intrinsic appeal, these participants felt that support for remote sensing is stronger when elected officials understand the long-term informational benefits, including the economic benefits, of using remote sensing data and its role in the protection of life and property. Others felt that the images offered a way to attract the attention of elected officials and upper management but agreed that the story of remote sensing should not stop with the image. Highlighting the contribution of remote sensing by discussing successful applications was felt to be very useful.
Experience, Skills, and Training
Discussions at the workshop emphasized that there are few formally trained remote sensing experts in state and local government. Many of the people who work in this field have moved into remote sensing responsibilities from technical positions that require GIS training and experience, as did the geographic information officer of Richland County, South Carolina. Others have moved into the field of remote sensing from analytical positions in offices that require the type of data that can be obtained only through satellite remote sensing, as in the state of Missouri and the city of Baltimore. Not surprisingly, although many remote sensing staff in the public sector are highly trained, some who have remote sensing responsibilities are self-taught, as was pointed out in the Baltimore case study.
However, when a public sector entity does not have enough technically trained staff, it can be at a disadvantage in drawing on available remote sensing resources. For example, at the planning meeting for the workshop, National Oceanic and Atmospheric Administration (NOAA) National Environmental Satellite, Data, and Information Service (NESDIS) officials reported that some state and local governments did not participate in a remote sensing grants program directed specifically at the nonfederal public sector because they lacked sufficient remote sensing staff.6 In addition, the Baltimore and Portland Metro case studies illustrate the need for technically trained government workers who can manage remote sensing contracts that are outsourced to third parties.
To improve their remote sensing expertise, as was mentioned earlier, many state and local governments work closely with local university centers for remote sensing research and hire new staff from among their graduates. State and local officials at the workshop spoke of the advantage enjoyed by agencies that can draw on local universities with strong remote sensing programs, as is the case in South Carolina, Colorado, and Missouri.
Many workshop participants also mentioned a potential role for the private sector in remote sensing training and skills development. The role of the private sector in providing an extensive and widely available array of GIS training courses was seen as a viable and valuable model for the extension of remote sensing expertise.
In its earlier report, the steering committee recommended that remote sensing training be provided not only for technical personnel but also for managers and decision makers, noting that
Whether in the public or the private sector, an organization’s capacity to incorporate remote sensing applications into its operations depends on having either technical staff with the necessary skills and understanding to process the data and transform it into usable information, or knowledgeable staff who can manage contracts with external, value-adding service providers.7
Although the focus and purpose of the training would be different, it is useful for those who will be using information derived from remote sensing data to have some familiarity with the technology and some understanding of its potential and its limitations. Training or information sessions could be held at meetings that state and local managers and chief information officers are likely to attend, such as the National Association of Counties, the International City/County Managers Association, or the Environmental Consortium of the States.
TRANSITIONING TO DIGITAL DATA
The issues that must be addressed to improve and facilitate the adoption of remote sensing in state and local government are not all institutional. Some are technical. One of the most important of these is the need to develop digital mapping standards for remote sensing. State and local entities are governed by a number of statutes and regulations that dictate photogrammetric mapping standards, but data processing of digital remote sensing data is not yet regulated by any governmental standards. Algorithms are not standardized, nor are there standards for the acquisition of data. In part, this is because there are few standardized products or expectations for the products derived from remote sensing. In most cases, the products are designed for specific decision making or operational management activities in the jurisdiction obtaining the data. However, much of the growth in the application of remote sensing data in state and local government is likely to occur across jurisdictions. Moreover, throughout the federal government there is a growing emphasis on setting common standards for all types of digital data to facilitate the use of data from multiple sources for emergency management and homeland security.
Another issue raised at the workshop is that the federal government tends to work in the metric system and state and local governments use the traditional English system of feet and inches. This can complicate the use of remote sensing applications across federal and nonfederal agencies.
The standards for specifying the parameters for constructing maps, which were developed as a result of the advent of aerial photography, are based on the characteristics of aerial cameras that use photographic film. The products have high spatial resolution, and the geometric correction of these images is complex because of the characteristics of the camera’s optical system. As a result, the feature identification part of the prevailing standards is quite stringent, while the geometric standards allow for what is possible given the type of camera and the scale at which the data are acquired. The geometric correction of images acquired by digital sensors, on the other hand, is a relatively straightforward process compared with that of film images, but the resolution of digital imagery is limited by the size of an individual pixel when projected onto the ground. In general, a digital image that can meet the geometric specification of traditional mapping standards for a given scale is not able to meet the feature identification part of the standard for the same scale. The mapping standards based on digital imaging technology need to take into account the fact that there are fundamental differences between the characteristics of the spatial information carried in a digital image as compared with the corresponding film image.
Requirements for geospatial information products are often stated in terms of the mapping standards developed for the production of maps from aerial photography rather than in terms of digital standards suitable for introduction into modern computer-based GIS. This makes it difficult for commercial remote sensing companies to respond to requests for proposals. Photogrammetric mapping standards are based on traditional aerial photography (analog) technology. Although these data are now often processed digitally, the mapping standards in common use for operational mapping are based on the use of photographic film to acquire the original data. Digital sensors, such as those carried by spacecraft, have fundamentally different characteristics (Box 3.2), which are difficult to reconcile with the traditional standards. This can significantly inhibit the use of remote sensing data in state and local governments, according to workshop participants.
LICENSING AND DATA MANAGEMENT
The licensing of remote sensing data and related issues of data access and continuity are critical for the economic and even the legal viability of using remote sensing data in state and local government. A number of participants in
the workshop spoke of the need to amortize the cost of remote sensing data across multiple units, whether these be agencies within a state or locality, independent jurisdictions within a region, or partnerships of federal and nonfederal units. Although federal regulations regarding the use of remote sensing data, such as Landsat data, permit this type of data sharing, it is often prohibited by the licensing restrictions of private data providers. The barriers to sharing data that licensing restrictions present are one reason why many state and local governments continue to use aerial photography. Because the air photos are typically sold, not licensed, to the purchaser, they can be shared without penalty. Current data licensing agreements that prohibit relicensing and distribution are a disincentive to local governments interested in adopting products based on commercial satellite images.
Another problem raised by private sector licensing restrictions is legal. Many jurisdictions in the nonfederal public sector are subject to freedom of information requirements that they provide data to citizens requesting them. This problem was raised by several workshop participants, but there appeared to be no common response to it.
Data Sharing and Cooperatives
If the right to share data depends on licensing provisions, the capacity to share data is related to the technological skills, use of common standards, and organizational arrangements in state and local government. Remote sensing professionals in the public sector must have sufficient technical capacity to manage and provide access to data for multiple purposes and to combine them with other sources of spatial data such as GIS and GPS. Officials in Boulder County, Colorado, found that sharing data across jurisdictions is facilitated by having formal, written data-sharing agreements. These agreements are the basis for membership in BASIC. The state of Missouri found that having an inventory of remote sensing data purchased by state agencies and departments provided information about the status of state-owned data to individuals in other departments that might be interested in using the data. State and local universities also share nonproprietary data with local government agencies.
Finally, data continuity is a critical issue for state and local governments. If a nonfederal public entity invests scarce resources in the purchase and use of remote sensing data, managers want the assurance that they will be able to obtain comparable data in the future. The investment in remote sensing is occasionally based on a one-time need for data, but more often it is premised on the belief that more data will become available for identifying and monitoring changes over time. For example, in the Red River Valley, even small changes in contours and
elevation due to development and flood damage in earlier years can affect future flooding and water flow. New data will have to be obtained frequently, possibly annually.
If these data are obtained through a contract with an airborne remote sensing firm, the contract can be renewed each year and control over access to new data is in the hands of the state or local government agency. However, if the data for a remote sensing application are obtained from satellite sources, the issue of whether data will be available in the future depends on satellite availability and renewal. The decisions about these issues are made by federal agencies, private firms, and even international remote sensing data providers and are often based on long-term economic, policy, and strategic criteria rather than short-term needs for continuity in data for specific applications. Yet from the perspective of the current and future development of remote sensing applications and the development of a robust market for remote sensing data in state and local government, unanswered questions about future data availability and uncertainty about access to existing historic or heritage data can be a disincentive to investing in new applications.