Sensors and Data Storage and Retrieval

A key component of geographical analysis is the focus on spatial variations, or the tracking of phenomena as they vary across the surface of Earth. Many geographical scientists rely on remote sensing from satellites, a technology with high initial costs (billions of dollars have been invested in the “big iron” of NASA’s Earth Observing System program over the past two decades), but one that has yielded astonishingly productive results in the form of massive amounts of fine-resolution data that can be used to map and analyze ocean temperatures, land cover, urban growth, and a host of other phenomena. Sizeable investments have also been made in the technologies that allow for the processing, storage, and dissemination of information, allowing thousands of scientists worldwide to benefit from the power of space-based sensors. It is important for these investments to continue, with commitments to develop new sensors as the science of sensing progresses; to replace aging satellites as they fail, thus ensuring reliable longitudinal series; and to explore the capabilities of each new sensor once it is in orbit. The geographical sciences can contribute to, and benefit from, all of these areas; indeed they have a vested interest in doing so, given the importance of this source of data to many of the questions posed in Part II.

Recently, the scientific community has begun to understand the potential of ground-based remote sensing. Ground-based sensor networks are often envisioned as arrays of fixed, inert sensors distributed across the landscape, each one capable of measuring useful properties of its immediate environment, determining its own location, and transmitting these measurements and locations to a central service where data can be integrated and disseminated to the scientific community. Large projects to install networks of value to the geographical sciences have been proposed by ecologists (National Ecological Observatory Network), oceanographers (National Science Foundation’s [NSF’s] Ocean Observatories Initiative), hydrologists (the WATERS network), and others.

As discussed in Chapter 10, however, there are limits to what can be sensed remotely, whether from space or from land, creating an imbalance in the data supply for studies such as those in the geographical sciences that deal with social and environmental systems. Traffic sensors on highways can provide useful information, but more generally there is little prospect of providing the kinds of data required to support the geographical perspective on social systems or on the social dimensions of coupled natural–human systems. The U.S. Census, which used to provide highly detailed decennial snapshots of the spatial differentiation of the population, will in the future provide only the most basic demographic data; the more detailed socioeconomic questions will now be covered by the American Community Survey, a rolling monthly sample that will provide finer temporal resolution but much coarser spatial resolution.

There is a growing imbalance between environmental and social data infrastructures. It is important to understand environmental systems, of course, but solving many of society’s pressing problems requires an equivalent level of understanding of social systems and of interactions between the environmental and the social. Indeed, human activity is both the cause of much environmental change and the recipient of many of its impacts. A possible source of social data has already been discussed in Chapter 11: the potential of humans to act as sensors of important social variables through a form of citizen science. Although this approach has already proved valuable in many areas, systematic research is needed into quality assurance, techniques for integration and dissemination, and organizational structures if it is ever to achieve its apparent promise. Other options, such as facilitating access to the administrative and commercial records that are largely off-limits to scientists, hold promise only if the obvious difficulties of ownership and confidentiality can be addressed.

One approach to resolving issues of access might be to create tightly controlled environments in which scientists could work with sensitive data but leave only with results that protected the confidentiality of the original records. This approach has been tried with some success by the U.S. Census Bureau, which has established firewalled data centers at selected universities where researchers can make custom requests of individual census records. An NRC report (NRC, 2007b) argues that this approach could be implemented through appropriately designed software, allowing researchers to access a range of confidential databases remotely through the Internet, thus avoiding the



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement