Usable Science VIII
Early Warning Systems: Do's and Don'ts
20-23 October 2003 - Shanghai, China

What Constitutes an Early Warning System?

An early warning system (EWS) is made up of several components and is not well represented only by the issuance of a warning. Components are the formulation of the warning, the issuance of the warning, and the reception of and response to the warning -- each of which has to be considered in evaluating an EWS. A weakness in any part of this process of events from warning preparation to responses can render the early warning system ineffective, and an early warning system that does not warn will not be taken seriously. Also, part of an effective EWS is the feedback loop that must always be in place so that those responsible for the warnings can better determine the value to at-risk people of specific warnings as well as of the system (a hindcasting activity).

Basic questions must be addressed before setting up or evaluating the effectiveness of an EWS. For example, do people agree on what is meant by early? What constitutes a warning? Who is to be warned? What seems obvious may prove to be not so obvious; a warning to one person may not be considered a warning by another. Does everyone (e.g., the government, the media, the public, the military) need to be warned at the same time and in the same way and by the same warning system?

How is a warning affected by the reliability and credibility of the sources of information on which it is based? Are there levels or degrees of warning? Is the warning to be an on/off warning (e.g., high alert followed by no alert), or are there levels of warning such as those we have for traffic lights: green, yellow, and red? Or are the levels of warning parallel to those that are used for weather phenomena such as tornadoes and hurricanes, i.e., a watch, a warning and an alert?

In fact, there are numerous EWS's in any given society, in addition to those in hydrometeorology. They are designed for famine, flash floods, infectious disease epidemics, heat, drought, fires, etc. The meteorological community has been using EWS's for a long time to deal with all kinds of extreme hydrometeorological events. Thus, there is considerable expertise to draw on when constructing a new EWS for the first time.

There is an EWS for just about every conceivable event of concern to individuals, societies, and governments. They exist at the global scale (e.g., for global warming or ozone depletion) at the national level (e.g., regional drought, desertification, famine, or large-scale flooding) and at the local level (e.g., infectious disease outbreak, nutritional changes).

The truth of the matter is that EWS's are constantly being created for newly identified threats (terrorism, West Nile virus, SARS), are being revised for changes in existing threats (food insecurity, invasive species), and critiqued for shortcomings (just about every system). There are examples of truly successful early warnings that led to the saving of lives and livelihoods and the protection of properties.

For the most part, EWS's are under constant scrutiny with each half-generation (every 10 years or so) trying to create the perfect system. The task is daunting. Existing systems are constantly being challenged by Nature. Making the difficult task of early warnings even more difficult is the fact that societies are changing as well as are ambient environmental conditions. Thus, the event or process that is being warned about is embedded in a set of confounding events or processes, and this synergy frequently leads to what has been labeled complex humanitarian crises.

* * * *

By coincidence, this workshop was convened almost ten years to the day after our first Usable Science Workshop on "Usable Science against Famine: Food Security, Famine Early Warning and El Nino" held in 1993 in Budapest, Hungary. While that meeting and publication (Glantz, 1994) focused on food security in sub-Saharan Africa, this meeting assessed early warning systems in general but with possible lessons for specific warning systems.

Glantz, M.H. (Ed.), 1994: Usable Science I: Food Security, Early Warning, and El Nino. Proceedings of workshop held 25-28 October 1993 in Budapest, Hungary. Boulder, CO: Environmental and Societal Impacts Group, National Center for Atmospheric Research. 250 pp.

Last Updated: 9 September 2003

The Environmental and Societal Impacts Group is part of the National Center for Atmospheric Research, which is managed by the University Corporation for Atmospheric Research and is sponsored by the National Science Foundation.