Effective Decision Support Systems

A decision-support system includes the individuals, organizations, networks, and institutions that develop decision-relevant knowledge, as well as the mechanisms to share and disseminate that knowledge and related products and services (NRC, 2009g). Agricultural or marine extension services, with all their strengths and weaknesses, are an important historical example of a decision-support system that has helped make scientific knowledge relevant to and available for practical decision making in the context of specific goals. The recent report Informing Decisions in a Changing Climate (NRC, 2009g) identified a set of basic principles of effective decision support that are applicable to the climate change arena: "(1 ) begin with users' needs; (2) give priority to process over products; (3) link information producers and users; (4) build connec tions across disciplines and organizations; (5) seek institutional stability; and (6) design processes for learning."

Effective decision-support systems work to both guide research toward decision relevance and link scientific information with potential users. Such systems will thus play an important role in improving the linkages between climate science and decision making called for both in this report and in many previous ones (e.g., Cash et al., 2003; NRC, 1990a, 1999b, 2009g). Research on the use of seasonal climate forecasts exemplifies current understanding of decision-support systems (see Box 4.4).

The basic principles of effective decision support are reasonably well known (see, e.g.,

BOX 4.4 Seasonal Climate Forecasts

For the past 20 years, the application of seasonal climate forecasts for agricultural, disaster relief, and water management decision making has yielded important lessons regarding the creation of climate knowledge systems for action in different parts of the world at different scales (Beller-Sims et al., 2008; Gilles and Valdivia, 2009; NRC, 1999b; Pagano et al., 2002; Vogel and O'Brien, 2006). Successful application of seasonal climate forecasting tends to follow a systems approach where forecasts are contextualized to the decision situation and embedded within an array of other information relevant for risk management. For example, in Australia, users and producers of seasonal climate forecasts have created knowledge systems for action in which the forecasts are part of a broader range of knowledge that informs farmers' decision making (Cash and Buizer,2005; Lemos and Dilling, 2007). In the U.S. Southwest, potential flooding from the intense 1997-1998 El Niño was averted in part because the 3- to 9-month advance forecasts were tailored to the needs of water managers and integrated into water supply outlooks (Pagano et al., 2002).

The application of seasonal climate forecasts is not always perfect. Seasonal forecasts have proven useful in certain U.S. regions directly affected by El Niño events but may have limited predictive skill outside those regions and outside the extremes of the El Niño-Southern Oscillation cycle (see Chapter 6). There is evidence that too much investment in climate forecasting may crowd out more sustainable alternatives to manage risk or even harm some stakeholders (Lemos and Dilling, 2007). For example, even under high uncertainty, a forecast of El Niño and the prospect of a weak fishing season give companies in Peru an incentive to accelerate seasonal layoffs of workers (Broad et al., 2002). More recent efforts to apply the lessons from seasonal climate forecasting to inform climate adaptation policy argue for the integration of climate predictions within broader decision contexts (Johnston et al., 2004; Klopper et al., 2006; Meinke et al., 2009). In such cases, rather than "perfect" forecasts, the best strategy for supporting decision making is to use integrated assessments and participatory approaches to link climate information to information on other stressors (Vogel et al., 2007).

NRC, 2009g). However, they need to be applied differently in different places, with different decision makers, and in different decision contexts. Determining how to apply these basic principles is at the core of the science of decision support—the science needed for designing information products, knowledge networks, and institutions that can turn good information into good decision support (NRC, 2009g). The base in fundamental science for designing more effective decision-support systems lies in the decision sciences and related fields of scholarship, including cognitive science, communications research, and the full array of traditional social and behavioral science disciplines.

Expanded research on decision support would enhance virtually all the other research called for in this report by improving the design and function of systems that seek to make climate science findings useful in adaptive management of the risks of climate change. The main research needs in this area are discussed in Informing Decisions in a Changing Climate (NRC, 2009g), Informing an Effective Response to Climate Change (NRC, 2010b), and several other studies (e.g., NRC, 2005a, 2008g). A recent review of research needs for improved environmental decision making (NRC, 2005a) emphasized the need for research to identify the kinds of decision-support activities and products that are most effective for various purposes and audiences. The report recommended studies focused on assessing decision quality, exploring decision makers' evaluations of decision processes and outcomes, and improving formal tools for decision support.

The key research needs for the science of decision support fall into the following five areas (NRC, 2009g):

• Information needs. Research is needed to identify the kinds of information that would add greatest value for climate-related decision making and to understand information needs as seen by decision makers.

• Communicating risk and uncertainty. People commonly have difficulty making good sense and use of information that is probabilistic and uncertain. Research on how people understand uncertain information about risks and on better ways to provide it can improve decision-support processes and products.

• Decision-support processes. Research is needed on processes for providing decision support, including the operation of networks and intermediaries between the producers and users of information for decision support. This research should include attention to the most effective channels and organizational structures to use for delivering information for decision support; the ways such information can be made to fit into individual, organizational, and institutional decision routines; the factors that determine whether potentially useful information is actually used; and ways to overcome barriers to the use of decision-relevant information.

• Decision-support products. Research is needed to design and apply decision tools, data analysis platforms, reports, and other products that convey userrelevant information in ways that enhance users' understanding and decision quality. These products may include models and simulations, mapping and visualization products, websites, and applications of techniques for structuring decisions, such as cost-benefit analysis, multiattribute decision analysis, and scenario analysis.

• Decision-support"experiments." Efforts to provide decision support for various decisions and decision makers are already under way in many cities, counties, and regions. These efforts can be treated as a massive national experiment that can, if data are carefully collected, be analyzed to learn which strategies are attractive, which ones work, why they work, and under what conditions. Research on these experiments can build knowledge about how information of various kinds, delivered in various formats, is used in real-world settings; how knowledge is transferred across communities and sectors; and many other aspects of decision-support processes.

Was this article helpful?

0 0

Post a comment