A recent paper out in Frontiers in Ecology and the Environment (Galaz et al 2009) identifies novel and fascinating ways on how to capture looming ecological crises.
The basic problem addressed by the authors is this: The six billion people on Earth are changing the biosphere at unprecedented rates. Ecosystems tend to respond to such change in unpredictable ways; collapsing fisheries and sudden phase shifts observed in freshwater ecosystems and coral reefs are good examples of such phenomena. The challenge is that existing ecological monitoring systems are not in tune with the speed of social, economical and ecological change and early warnings of pending ecological crisis are to a large extent limited by insufficient data, and geographical gaps in official monitoring systems.
So how do we deal with this situation? Look to the internet for guidance! Not quite so simple, but the researchers from the Stockholm Resilience Centre and the University of east Anglia, explore the possibilities of using information posted on the Internet to detect ecosystems on the brink of change.
Much of the pioneering work in this type of Internet surveillance has come in the public health field, where software programs that search the Internet in methodical and automated manners, web crawlers, are used to track disease.
The potential of web crawlers is illustrated by the success of the Global Public Health Intelligence Network (GPHIN), an early disease detection system developed by Health Canada for the World Health Organization (WHO). GPHIN gathers information about unusual disease events by monitoring internet-based global media sources, such as news wires, web sites, local online newspapers, and public health e-mail information services, in eight languages, with non-English articles filtered through a translation engine. The system retrieves approximately 2000–3000 news items per day; roughly 30% are rejected as duplicative or irrelevant, but the remainder are sorted by GPHIN analysts and posted on GPHIN’s secure website.
Web crawlers could be designed to complement conventional ecological monitoring. The authors use coral reef ecosystems to illustrate how such a process could progress. Data-mining the internet for information on potential drivers of coral ecosystem change (e.g. heavy investment in fish gear that can precede heavy exploitation of key reef organisms) and ecosystem responses (changes in coral cover, fish community composition) can be the basis for early warning assessments of ecological change.
Addtionally, by searching the internet for reports of local scale coral reef degradation can provide early indicators of large scale systemic collapses of reef systems. The success of such web-crawlers will be highly dependent on information becoming rapidly accessible online via”web 2.o” applications such as blogs, wikis and other networking tools such as electronic mailing lists (Coral-List is highlighted as an example).
I guess that a problem, and one highlighted by the authors, is that fragmented and insufficient data from several sources, could lead to information junkyards instead of robust ecological monitoring systems. Any web crawler based monitoring system would therefore need to be plugged into a coupled knowledge management and expert judgement system. Would that slow the process down to the extent of nullyfying any gains made through the rapid information sweeps generated by the web crawler? In any case, its a refreshing approach and a fascinating read.