Frontiers in Ecology

and the Environment

Can web crawlers revolutionize
ecological monitoring?
Victor Galaz, Beatrice Crona, Tim Daw, Örjan Bodin, Magnus Nyström, and Per Olsson

Front Ecol Environ 2009; doi:10.1890/070204

This article is citable (as shown above) and is released from embargo once it is posted to the
Frontiers e-View site (www.frontiersinecology.org).

Please note: This article was downloaded from Frontiers e-View, a service that publishes fully edited
and formatted manuscripts before they appear in print in Frontiers in Ecology and the Environment.
Readers are strongly advised to check the final print version in case any changes have been made.

esa
© The Ecological Society of America www.frontiersinecology.org
CONCEPTS AND QUESTIONS

Can web crawlers revolutionize ecological

monitoring?
Victor Galaz1*, Beatrice Crona1,5, Tim Daw3, Örjan Bodin1,4, Magnus Nyström1,2, and Per Olsson1

Despite recent advances, ecosystem service monitoring is limited by insufficient data, the complexity of
social–ecological systems, and poor integration of information that tracks changes in ecosystems and eco-
nomic activities. However, new information and communication technologies are revolutionizing the genera-
tion of, and access to, such data. Can researchers who are interested in ecological monitoring tap into these
increased flows of information by “mining” the internet to detect “early-warning” signs that may signal
abrupt ecological changes? Here, we explore the possibility of using web crawlers and internet-based informa-
tion to complement conventional ecological monitoring, with a special emphasis on the prospects for avoid-
ing “late warnings”, that is, when ecosystems have already shifted to less desirable states. Using examples from
coral reef ecosystems, we explore the untapped potential, as well as the limitations, of relying on web-based
information to monitor ecosystem services and forewarn us of negative ecological shifts.
Front Ecol Environ 2009; doi:10.1890/070204

T he combined impacts of global environmental

change and the complex behavior of ecological sys-
tems, create opportunities for major “ecological sur-
prises” at various spatial scales (Schneider and Root
1995; Gunderson 2003; Gordon et al. 2008). Ecosystems
provide many vital ecosystem services (ES), such as
water purification and food production, but rapid
changes due to, for instance, climate change and shifting
global markets, present serious challenges to their future
ability to deliver these life supporting services (MA
2005). Examples of such changes include collapsing fish-
eries at national and global scales (Berkes et al. 2006),
irreversible degradation of freshwater ecosystems and
coral reefs, and decreasing soil productivity (Scheffer et
In a nutshell:
• Steering away from catastrophic shifts in ecosystems is of prime
concern in an era of global environmental change
• Existing monitoring of ecosystem services is poor and frag-
mented, especially in developing countries
• Information and communications technology is revolutioniz-
ing the generation of, and access to, social, ecological, and eco-
nomic information
• Systematic “data mining” of such information through the
internet can provide important early warnings about possible
pending abrupt losses of ecosystem services
al. 2001; MA 2005).The situation is exacerbated by
national and international responses to such changes
that are either insufficient or non-existent. Restoration
may be difficult, because feedbacks in the system can act
to stabilize these new, undesirable ecosystem states
(Scheffer et al. 2001; Gordon et al. 2008). It is therefore
of primary importance to try and avoid crossing the
thresholds that lead to these outcomes.
Despite advances in monitoring technology (Clark et
al. 2001), it is evident that existing information on
changes in ES tends to be poor and contains serious gaps.
Furthermore, existing monitoring systems are unable to
capture the impacts of rapid demographic, economic,
and sociopolitical changes that result from economic
development and increasing global flows of information,
trade, and technology (MA 2005; Berkes et al. 2006;
Carpenter et al. 2006). The difficulties in quantifying
social and ecological uncertainty, the lack of expert
agreement on what indicators to monitor, poor-quality
existing data, and the costs associated with setting up
long-term monitoring programs (Walters 2007) all ham-
per our ability to steer away from, or to prepare for,
abrupt changes to ecosystems and the loss of related ES.
This is particularly true for countries that suffer from
poor governance and weak environmental institutions
(Danielsen et al. 2003; UNEP 2007).

1
Stockholm Resilience Centre, Stockholm University, Stockholm,
Sweden *(victor.galaz@stockholmresilience.su.se); 2Department of
Systems Ecology, Stockholm University, Stockholm, Sweden;
3
School of Development Studies, University of East Anglia,
Norwich, UK; 4Department of Government, Uppsala University,
Uppsala, Sweden; 5The Centre for the Study of Institutional
Diversity, Arizona State University, Tempe, AZ
! Information and communication technologies
The role of information and communication technology
(ICT) – for economic growth, education, and human
development – has been discussed elsewhere (Leach and
Scoones 2006). Meanwhile, the evolution of “web 2.0”
permits more interactive use of the internet and allows
users to post, edit, comment on, and provide information

© The Ecological Society of America www.fr ontiersinecology.or g

Web crawlers and ecological monitoring V Galaz et al.

in blogs and wikis or via podcasting, videoblogs, and Madagascar and Leach and Scoones [2006] on participa-
other networking tools. Globally, access to information tory geographic information system “citizen-maps” for
technology is very unequally distributed (IER 2005; hydrological monitoring).
Leach and Scoones 2006), but access to, and use of, the One primary difficulty, however, lies in designing mon-
internet is increasing rapidly in all regions of the world. itoring systems that are able to scan the internet continu-
For example, between 2000 and 2004, the number of ously for predefined ecological events and changes that
internet users in the developing world tripled, from 96 might signal emerging ecological vulnerabilities, and sub-
million to almost 333 million; in Africa alone, the num- sequently integrating that information with existing, offi-
ber of users increased more than five-fold, from 4.3 to cial monitoring data. Although the realization of such a
21.8 million, during that same period (IER 2005). system is far into the future, innovative uses of web
The rapid development of ICT has not only led to crawlers (software programs or automated scripts that
increased flows of information at a global scale, but also browse the World Wide Web in a methodical, automated
sets the stage for innovative uses of internet-based infor- manner) are likely to provide an important complement
mation – ranging from e-mail lists and local newspaper to conventional monitoring in the present. The case
articles to preprints of peer-reviewed journal articles – as study we highlight of the live reef fish trade is a clear
an important complement to conventional ecological example of the problems inherent in relying on official
monitoring. The potential of ICT is currently being data alone, and one where a creative application of inter-
explored in a number of contexts for ecology. Examples net-based information could provide a valuable resource
here range from the Resilience Assessment wiki (see Panel 1).
(http://wiki.resalliance.org), to online datasets such as The potential of web crawlers is illustrated by the success
those posted by the US National Center for Ecology and of the Global Public Health Intelligence Network
Analysis and Synthesis (www.nceas.ucsb.edu), to the use (GPHIN), an early disease detection system developed by
of the internet to coordinate citizen-science projects Health Canada for the World Health Organization
(Levitt 2002). In addition, Crowl et al. (2008) suggest (WHO). GPHIN gathers information about unusual dis-
the creation of a coordinated “cyber-infrastructure” to ease events by monitoring internet-based global media
facilitate prompt warnings of invasive alien species and sources, such as news wires, web sites, local online newspa-
infectious diseases. pers, and public health e-mail information services, in eight
Here, we explore the possibilities and limitations of languages, with non-English articles filtered through a
more systematic “data mining” of the internet, and the translation engine. The system retrieves approximately
potential for obtaining complementary information and 2000–3000 news items per day; roughly 30% are rejected as
early warnings – not only about discrete ecological events duplicative or irrelevant, but the remainder are sorted by
(eg a disease outbreak caused by invasive species), but GPHIN analysts and posted on GPHIN’s secure website
also changes in ecological drivers, and the impacts of (Weir and Mykhalovskiy 2006).
ecosystem change – to forewarn us of ES losses. The ability to trawl extensively for various signals, the
wide diversity of information sources, and the ability to
! Ecology on the internet identify alarming early-warning signs seem give the system
the flexibility and speed needed to detect unexpected
One example of how informal ICT informa-
tion can support ecological monitoring is the Panel 1. We b c r a w l e r s a n d t h e l i v e f i s h t r a d e
use of electronic mailing lists to disseminate Globalized markets have become important drivers for fisheries systems, dri-
and compile field observations tracking ving rapid development, overexploitation, and collapse of local fisheries,
global-scale coral bleaching during the before effective management can be established (Berkes et al. 2006). The live
1997–1998 El Niño event. The existence of reef fish trade (LRFT) supplying seafood to restaurants in Asia is a good exam-
an electronic mailing list for coral reef-asso- ple. This fishery has been characterized by a boom-and-bust pattern of
ciated news proved invaluable for prompt sequential exploitation of reefs and nations, and serial depletion of the most
assessments of the mass-bleaching event valuable species (Scales et al. 2006).
(Hoegh-Guldberg 1999), with reports rang- Although some Pacific Ocean nations have recognized the threat of LRFT
ing from “detailed accounts with accurate and have started to take precautionary actions, coordinated by the Secretariat
measures of bleaching and mortality, to brief of the Pacific Community (Sadovy et al. 2003), many in other areas, such as the
anecdotal reports obtained during a rapid Caribbean and the Western Indian Ocean, have not, and lack of data on the
site visit” (Wilkinson 1999; see WebPanel status of many small-scale reef fisheries has also been a severe impediment to
1). Information of this kind can, in princi- action. Socioeconomic and ecological signals, provided by web crawlers, could
potentially improve early detection of nations and regions at risk of being hit
ple, be easily associated with participatory
by the next sequential wave of LRFT. Examples of the types of signals that
ecological monitoring projects or citizen-sci-
could be used include trade advertisements, availability of products by area,
ence initiatives, provided that they are prices, number of suppliers, observations by non-state entities, such as envi-
posted on the internet (see Andrianandra- ronmental organizations, and newsletters.
sana et al. [2005] on wetland monitoring in

www.fr ontiersinecology.org © The Ecological Society of America

 V Galaz et al. Web crawlers and ecological monitoring

Driver signals Data type Data type Impact signals

Drivers of system change

Coral reef

Global
Global

Foreign investment

Trade volumes, trophic

levels, size grades
Development aid Government
investment fisheries stats

Regional
Regional

Disease
Subsidies for outbreak
exploitation

Local
Local

Coastal development Coral cover, fish

and construction community composition
Courtesy of B Crona and R Kautsky/Azote

Online news media (eg newspapers, radio and television, Published (online) reports and documents from
and online newsletters) government agencies, UN agencies, OECD, and similar

Biogs (eg dialogues conducted through dive clubs, interest Online accessible databases (eg trade statistics, landing
groups, NGOs, electronic mailing lists, purchase requests) statistics)

Figur e 1. Examples of drivers and impact signals regarding a coral reef social–ecological system that, in principle, could be detected
by a web crawler. “Driver signals” are key social, ecological, and economic factors that risk leading to loss of ecosystem services.
“Impact signals” are changes that may indicate pending loss of ecosystem services. Note that the list of signals is not exhaustive. Based
on Nyström et al. (2000); Berkes et al. (2006); Scales et al. (2006); and McCook et al. (2007). The analysis of these signals is
not necessarily carried out by a single entity or individual, but rather may include, for example, academia, UN agencies, NGOs,
government and citizen scientists, and military and diplomatic agencies.

disease outbreaks. For example, GPHIN currently captures
the first hints of about 40% of the 200–250 outbreaks subse-
quently investigated and verified by WHO each year.
GPHIN was also one of the first systems to obtain non-offi-
cial reports of a suspected influenza outbreak in mainland
China in 2002, which, 3 months later, was identified by
WHO as severe acute respiratory syndrome (SARS; Fidler
2005; Weir and Mykhalovskiy 2006).
! Web crawlers and ecosystems
Abrupt losses of ecosystem services are obviously difficult
to forecast with certainty, mainly because they result from
multiple changes at different scales (Clark et al. 2001).
However, research on coupled social and ecological sys-
tems over the past decade has identified several changes
that may provide early warnings of potential damage to
ecosystem services. For example, an abrupt transition from
a coral-dominated reef to an algae-dominated one may be
preceded by declining abundance of large herbivorous fish
(Nyström et al. 2000); a rapid transition from a clear to a
highly turbid and eutrophic state in a lake may be preceded
by increased fertilizer use on nearby farms (Gordon et al.
2008); and heavy investment in specific fishing gear and
technical equipment may precede the loss of certain key
species in marine fisheries (Berkes et al. 2006). Figure 1
uses the example of coral reef ecosystems to illustrate
diverse sources of internet-based information on both dri-
vers and ecosystem responses, to monitor and forewarn of
pending ecological shifts. Nonetheless, the collection and
presentation of signals need to be supplemented by expert
analysis, knowledge management approaches (see
McDermott [1999] for more information), and an under-
standing of local ecological and social conditions. Only
then will we be able to obtain robust estimates of possible
impacts and to evaluate the possible countermeasures or
adaptation strategies we might use in response (Crowl et al.
2008).

© The Ecological Society of America www.fr ontiersinecology.or g

Web crawlers and ecological monitoring V Galaz et al.

(a) responses before the ES are lost. Here, we suggest

three potential ways to using web crawlers to
forewarn of ecological shifts.
First, web crawlers can collect information on
the drivers of ecosystem change, and not just on
the resulting impacts. For example, if emerging
markets for high-value species are known to be
socioeconomic drivers that lead to the overex-
ploitation and collapse of a fishery (see Panel 1),
web crawlers can be designed to collect informa-
tion on rapid changes in prices, landings, or
investments in particular regions (Figure 2).
Meyerson and Reaser (2003), for instance,
report on a web crawler developed by the US
(b)
Department of Agriculture’s Animal and Plant
Health Inspection Service to search for, and
report on, sales of prohibited organisms over the
internet, in an attempt to address the threat of
(c) invasive alien species.
Second, future early-warning systems can
make use of the recent insight that variance
within ecosystems can increase in response to
Figure 2. Ecological information is often accessible in several languages and stress. For example, the variability of fish pop-
diverse settings on the internet. (a) This screen shot from a Chinese food ulations has been shown to increase in
market web page illustrates the type of information that can be retrieved. (b) response to exploitation (Hsieh et al. 2006).
Information about marine species for sale in the market, together with Carpenter and Brock (2006) argue that vari-
information about the highest, lowest, and average price. The last column ance within complex ecological systems gen-
provides price statistics for the chosen species. (c) A news section, which erally increases in advance of catastrophic
includes changes in access to specific marine species. First news item reads: shifts. Although web crawlers harvest infor-
“According to an integrated investigation of the coastal zone, both Chinese mation on discrete events, rather than provid-
prawn and little yellow croaker have returned in the Bohai Sea”, while the ing the time series needed to formally analyze
sixth news item reads: “Big stocks of little yellow croaker have re-emerged variance patterns, increases in variance are
after 30 years in the Yellow Sea”. very likely to result in an increased frequency
of what is perceived as “unusual events”,
Analysis and response are not necessarily organized which may make their way into local newspapers,
around a single – ie national government – entity. On blogs, or electronic mailing lists (Figure 3a).
the contrary, both might occur as a result of collabora- Nonetheless, the realization that increased variance
tions between, for example, state agencies and other indicates a pending ecological shift is a recent one,
expert analysts in the form of non-governmental organi- based on ecological modeling (compare with Oborny
zations, private companies, universities, and the general et al. 2005; van Nes and Scheffer 2007). Thus, whether
public. If the outputs are more widely available, analysis this approach is possible with web-crawler-based mon-
and response could even be the result of autonomous itoring systems needs to be explored further.
actions, assumed by independent organizations and Finally, a more clear-cut approach is one that builds
individuals. on the fact that ecological shifts at small scales often
precede similar shifts in other locations or, more seri-
! Warnings that come too late ously, larger-scale systemic changes. Examples include
outbreaks of invasive alien species (Meyerson and
There are important differences between monitoring for Reaser 2003), or the way in which resilience of ecosys-
loss of ES and disease outbreaks. Web-crawler-based tems, such as forest reserves and coral reefs, is thought
early-warning systems for disease epidemics rely on the to be dependent on surrounding refuge areas, which can
identification of discrete events (Weir and Mykhalovskiy aid the recovery from small-scale shifts through, for
2006), rather than on monitoring underlying social, eco- example, the movement of species and supply of larvae
nomic, or ecological changes. However, discrete events (Nyström et al. 2000; Bengtsson et al. 2003; see Figure
can, in principle, be used as early warnings of approach- 3b). Therefore, repeated small-scale shifts may not only
ing abrupt shifts in ecological systems. Given the poten- lead to a cumulative loss of “spatial resilience”, but can
tial for irreversible loss of ES, early warnings are impor- also provide early indications of large-scale systemic
tant in allowing the introduction of management losses of ES (Figure 3b).

www.fr ontiersinecology.org © The Ecological Society of America

V Galaz et al. Web crawlers and ecological monitoring

(a) (b)

Time series

Distributions
System 1 System 2 System 3

Healthy local ecosystem

Local ecosystem in degraded state
Probabilities Flows from refuge areas to assist recovery
for unusual events

Figure 3. (a) Increasing variance of a key system variable (eg fish abundance or nutrient concentration) (top). As variance
increases, the probability distribution changes (bottom), which could imply more frequent observations of “unusual” events (beyond
horizontal dotted lines). (b) Ecological shifts at smaller scales can provide warnings of impending changes to large-scale systems.
System 1: recovery from disturbances is assisted by multiple sources of “ecological memory”. System 2: higher frequency of local
shifts, which increases the risk of the system moving into a large-scale phase shift. System 3: majority of sites are degraded, making
recovery of both local sites and the large-scale system unlikely. Adapted from Nyström et al. (2008).

! Data management and lack of societal response
Despite the exciting possibility of using web crawlers for
ecological monitoring and in early-warning systems, we
recognize that crucial challenges need to be addressed
before these systems can contribute to the detection (and
possible avoidance) of abrupt ecosystem changes. There
is still a need to integrate, verify, and manage ecological
and socioeconomic data. Data integration, expert analy-
sis, and knowledge management have proven to be the
main obstacles to ecological monitoring (Carpenter et al.
2006), even among well-defined monitoring systems in
developed countries. For example, communicable disease
surveillance in the European Union (Amato-Gauci and
Ammon 2008) and invasive species monitoring in the
US (Meyerson and Reaser 2003; Crowl et al. 2008) illus-
trate the difficulties posed by fragmented or otherwise
insufficient social and ecological data, and the continu-
ing risk of creating “information junkyards” – ie increas-
ingly large collections of data with little or no practical
value (McDermott 1999) – instead of robust ecological
monitoring systems. Any web-crawler-based monitoring
system would therefore require the support of a coupled
knowledge management and expert judgment system.
Early warnings are never a guarantee of timely and appro-
priate remedial responses. The need for prompt responses to
outbreaks of Ebola hemorrhagic fever and avian influenza
(H5N1), for example, has gained increased social and polit-
ical support over the past few years, spurring the develop-
ment of new, international regulations and response opera-
tions. This is facilitated by a relatively strong international
organization for human health, with an international man-
date – the WHO (Fidler 2005). This development is in
stark contrast with global environmental governance,
which suffers from implementation deficits, serious coordi-
nation failures, and inadequate funding (Biermann 2002).
Responses to infectious disease (eg isolation, vaccination,
medical care) are also likely to be simpler and less politically
contentious than the responses to approaching ecological
shifts (eg fishing restrictions, restrictions on agricultural
activity, implementation of deforestation legislation).
The challenges of data integration and the current lack
of governmental response to ecological change should
not be underestimated. They do not, however, preclude
the need to explore innovative solutions to bridge the gap
between poor monitoring, and the rapid rate of
social–ecological change, with potentially serious reper-
cussions for human well-being. The use of web crawlers
should be explored further, in an attempt to prepare for
the ecological challenges of an uncertain future.
! Call for comments
The authors invite readers to discuss and comment on
this article at http://resilienceinnovation.blogspot.com
! Acknowledgements
This work was supported by the Stockholm Resilience
Centre, and by grants from the Foundation for Strategic
Environmental Research (Mistra). We thank F Westley
of the University of Waterloo and several other col-
leagues for inspirational discussions on this subject.
Assistance in translation and with the web search was
provided by G Han, Stockholm Environment Institute.
! References
Amato-Gauci A and Ammon A. 2008. The surveillance of com-
municable diseases in the European Union – a long-term strat-
egy. Eurosurveillance 13: 26.
Andrianandrasana HT, Randriamahefasoa J, Durbin J, et al. 2005.
Participatory ecological monitoring of the Alaotra wetlands in
Madagascar. Biodivers Conserv 14: 2757–74.
Bengtsson J, Angelstam P, Elmqvist T, et al. 2003. Reserves,
resilience and dynamic landscapes. Ambio 32: 389–96.
Berkes F, Hughes TP, Steneck RS, et al. 2006. Globalization, roving
bandits, and marine resources. Science 311: 1557–58.

© The Ecological Society of America www.fr ontiersinecology.or g

Web crawlers and ecological monitoring
Biermann F. 2002. Green global governance – the case for a World
Environment Organisation. New Econ 9: 82–86.
Carpenter SR and Brock WA. 2006. Rising variance: a leading
indicator of ecological transition. Ecol Lett 9: 311–18.
Carpenter SR, DeFries R, Dietz T, et al. 2006. Millennium
Ecosystem Assessment: research needs. Science 314: 257–58.
Clark JS, Carpenter SR, Barber M, et al. 2001. Ecological forecasts:
an emerging imperative. Science 293: 657–60.
Crowl TA, Crist TO, Parmenter RR, et al. 2008. The spread of
invasive species and infectious disease as drivers of ecosystem
change. Front Ecol Environ 6: 238–46.
Danielsen F, Mendoza MM, Alviola P, et al. 2003. Biodiversity
monitoring in developing countries: what are we trying to
achieve? Oryx 37: 407–09.
Fidler DP. 2005. SARS, governance and the globalization of dis-
ease. New York, NY: Palgrave McMillan.
Gordon L, Peterson GD, and Bennett EM. 2008. Agricultural mod-
ifications of hydrological flows create ecological surprises.
Trends Ecol Evol 23: 211–19.
Gunderson LH. 2003. Adaptive dancing: interactions between
social resilience and ecological crises. In: Berkes F, Colding J,
and Folke C (Eds). Navigating social–ecological systems: build-
ing resilience for complexity and change. Cambridge, UK:
Cambridge University Press.
Hsieh CH, Reiss CS, Hunter JR, et al. 2006. Fishing elevates vari-
ability in the abundance of exploited species. Nature 443:
859–62.
Hoegh-Guldberg O. 1999. Climate change, coral bleaching and
the future of the world’s coral reefs. Mar Freshwater Res 50:
839–66.
IER (Information Economy Report). 2005. Information economy
report. New York, NY and Geneva, Switzerland: UN
Conference on Trade and Development.
Leach M and Scoones I. 2006. The slow race: making technology
work for the poor. London, UK: Demos.
Levitt JN (Ed). 2002. Conservation in the internet age: threats and
opportunities. Washington, DC: Island Press.
MA (Millennium Ecosystem Assessment). 2005. Ecosystems and
human well-being. Washington, DC: Island Press.
McCook LJ, Folke C, Hughes TP, et al. 2007. Ecological resilience,
climate change and the Great Barrier Reef. In: Johnson JE and
Marshall PA (Eds). Climate change and the Great Barrier Reef.
www.fr ontiersinecology.org
V Galaz et al.
Townsville, Australia: Great Barrier Reef Marine Park
Authority.
McDermott R. 1999. Why information technology inspired but
cannot deliver knowledge management. In: Lesser EL,
Fontaine MA, and Slusher JA (Eds). Knowledge and commu-
nities. Boston, MA: Butterworth-Heinemann.
Meyerson LA and Reaser JK. 2003. Bioinvasions, bioterrorism, and
biosecurity. Front Ecol Environ 1: 307–14.
Nyström M, Folke C, and Moberg F. 2000. Coral reef disturbance
and resilience in a human-dominated environment. Trends Ecol
Evol 15: 413–17.
Nyström M, Graham NAJ, Lokrantz J, and Norström A. 2008.
Capturing the cornerstones of coral reef resilience: linking the-
ory to practice. Coral Reefs. 27: 795–809.
Oborny B, Meszéna G, and Szabó G. 2005. Dynamics of popula-
tions on the verge of extinction. Oikos 109: 291–96.
Sadovy YJ, Donaldson TJ, Graham TR, et al. 2003. While stocks
last: the live reef food fish trade. Manila, Philippines: Asian
Development Bank.
Scales H, Balmford A, Liu M, et al. 2006. Keeping bandits at bay?
Science 313: 612–13.
Scheffer M, Carpenter SR, Foley JA, et al. 2001. Catastrophic shifts
in ecosystems. Nature 413: 591–96.
Schneider SH and Root TL. 1995. Ecological implications of cli-
mate change will include surprises. Biodivers Conserv 5:
1109–19.
UNEP (UN Environment Programme). 2007. Global environment
outlook – GEO4. Nairobi, Kenya: UN Environment
Programme.
van Nes E and Scheffer M. 2007. Slow recovery from perturbations
as a generic indicator of a nearby catastrophic shift. Am Nat
169: 738–47.
Walters CJ. 2007. Is adaptive management helping to solve fish-
eries problems? Ambio 36: 304–07.
Weir L and Mykhalovskiy E. 2006. The geopolitics of global public
health surveillance in the twenty-first century. In: Bashford A
(Ed). Medicine at the border – disease, globalization and secu-
rity, 1850 to the present. New York, NY: Palgrave McMillan.
Wilkinson CR. 1999. The 1997–1998 mass bleaching event
around the world. Compilation of internet reports, global coral
reef monitoring network. Townsville, Australia: Australian
Institute of Marine Science.
© The Ecological Society of America
V Galaz et al. – Supplemental information
WebPanel 1. Selected excerpts from the Coral List from early 1998
Accessible at www.coral.noaa.gov/lists/archives.shtml.
Senders e-mail addresses have been removed, and region added by article authors.

Panama
Wed, 1 Oct 1997 14:49:26
Significant coral bleaching was observed on 17 September 1997 at Uva Island in the Gulf of Chiriqui, Pacific Panama.All zooxanthellate
scleractinian coral species were affected, at all depths (no corals present > 20 m). The most severely bleached (completely white)
colonies still had extended polyps and no signs of algal overgrowth, suggesting the event occurred relatively recently. Most colonies of
the hydrocoral Millepora intricata (the only common species of the genus remaining after the 1982–83 ENSO) were already dead and
covered with a thin algal film, suggesting they may have bleached earlier than the scleractinians.

Galapagos
Thu, 5 Jan 1998 21:11:06
As of Dec 18–30, bleaching was observed first hand in Galapagos. Roughly 20% of polyps of roughly 80% of the coral I saw was bleached
near the top (mostly a brown lumpy coral, I don’t know the name, anyone?) although I was only able to visit Santa Cruz, Bartolome,
Santa Fe, and Espanola; NOT the islands typically known for large coral assemblages (Devil’s Crown, Isabella).

Galapagos
Sat, 21 Jan 1998 12:32:09
FYI, a NOAA Press Release:
EL NINO CAUSING CORAL BLEACHING IN GALAPAGOS,
NOAA ANNOUNCES
El Niño’s extremely warm waters in the Pacific Ocean have caused coral bleaching in the waters around the Galapagos Islands, the
Commerce Department’s National Oceanic and Atmospheric Administration announced today.

Hawaii
Tue, 3 Mar 1998 20:46:37
The coral reef here is a bloody disaster.What isn’t dead is bleached so white from loss of algae that I think much of it will starve before
it comes good.The sea temp reached 33˚C at 15 meters depth at four mile reef last month.We are getting South easters, now bringing
in cooler water but it is still very hot.This is unprecedented. No one can remember anything like this happening before.

Western Samoa
Thu, 5 Mar 1998 13:20:18
A survey at Palolo Deep (a National Marine Park near Apia,Western Samoa) on 28 February revealed severe coral bleaching. Between
60 to 70% of all staghorn Acropora on the reef top was bleached.This has occurred with amazing rapidity (over a period of 5–6 days).
In deeper water, all seemed well.

© The Ecological Society of America w w w. f r ontiersinecology.or g