20

Conservation Biology, Pages 20–23Volume 17, No. 1, February 2003

The Challenge of Measuring Global Change in Wild Nature: Are Things Getting Better or Worse?

Introduction

The Ministerial Declaration made atthe World Summit on Sustainable De-velopment in Johannesburg agreed toprotection and restoration of the in-tegrity of our planet’s ecological sys-tems and identified as a major prioritya significant reduction in the rate ofcurrent biodiversity loss at nationaland global levels. This goal is essen-tially a reformulation of that of theConvention on Biological Diversityagreed to in Rio de Janeiro 10 yearsago. For us to have any idea of wherewe are in relation to this goal, weneed information on current globaltrends in the status of natural systems.By 2005 it is envisaged that the Millen-nium Ecosystem Assessment ( MA)will have provided a comprehensiveaccount of global biodiversity from acompilation of existing data, but thatis a long time to wait. What can wesay now about trends in the state ofwild nature since the United Nations’1992 Conference on Environment andDevelopment in Rio de Janeiro? Here,we briefly review recent global trendsin habitat area in as many broadly de-fined natural habitats as possible andin indices of animal populations char-acteristic of those habitats. The infor-mation available indicates continuingdeclines in habitat area and species,but that data are extremely sparse.

Trend Estimates

We searched the published litera-ture and available databases for glo-

bal estimates of recent trends in theareal extent of largely unmodifiedhabitats in all the biome categoriesof Costanza et al

.

(1997) except rock,ice, and open ocean, which are usu-ally not converted by humans toother habitats, and except urban andcropland, which are already modi-fied fundamentally. We included anyestimate of global trend that wasbased on a series of studies begin-ning in 1970 or later and that in-cluded at least 5 years of data collec-tion after 1992. Because we foundestimates fulfilling these criteria foronly four habitats, we supplementedthese with biome-specific indicesbased on time-series data on popula-tions of wild vertebrates derived fromthe World Wildlife Fund (WWF)2000 Living Planet Index (LPI) (Lohet al

.

2000 ). We calculated annualpercent change in area or index (

a

)value by taking the values

a

1

,

a

2

in

the first year (

t

1

) and the last year(

t

2

) of the series under consider-ation and calculating 100*(1

�

(

a

2

/

a

1

)

(1/(

t

2

�

t

1 )

).

Tropical Forests

The Forest Resource Assessment2000 (Food and Agriculture Organi-zation [FAO] 2000) estimates a glo-bal net change of

�

7% in the area ofnatural moist and dry tropical forestfor the period 1990–2000, yieldingan annual decline of 0.8%. We recog-nize the technical difficulties and po-tential inaccuracies that arise from thisassessment’s use of a mixture of re-ports on forest inventory data from

national forest services and remotesensing. These and other problemshave led some authorities to suggestthat the 1990–2000 assessment un-derestimates the rate of decline (e.g.,Stokstad 2001). A more recent study,however, based entirely on remote-sensing measurement of changes intropical moist forest cover in a strati-fied random sample of areas compris-ing 6.5% of the humid tropics (Achardet

al. 2002), yielded a somewhatlower estimate of 0.4% for annualnet deforestation for this forest typeduring the period 1990–1997 andannual conversion of forest to “visi-bly degraded” areas of 0.2%. The LPIfor vertebrates in tropical forestsshowed a decrease of 26% between1970 and 1999, yielding an averageannual decline of 1.1%. This is some-what higher than the FAO figure (FAO2000) for change in area and notablyhigher than the estimates of Achardet al

.

(2002).

Temperate and Boreal Forests

The FAO estimates that temperateand tropical forests have increasedin extent by 1% during the period1990–2000, yielding an annual in-crease of 0.1%. The LPI for temper-ate forests showed a change of

�

4%between 1970 and 1999, yielding asimilar, small annual increase of0.1%.

Mangroves

Valiela et al

.

(2001) estimates on thebasis of a comprehensive assessment

Jenkins et al. Issues in International Conservation

21

Conservation BiologyVolume 17, No. 1, February 2003

of mangrove resources that at least35% of the global area of mangroveforests has been lost in the past twodecades. This yields an annual de-cline of at least 2.5%.

Coral Reefs

Although Bryant et al

.

(1998) reportthat around one-quarter of theworld’s reefs are believed to be athigh risk of degradation, there areno reliable global estimates for therate at which coral reefs are actuallybeing lost.

Sea Grass and Algal Beds

There are no global estimates for theextent of or rates of change in algalbeds. No comprehensive survey ofsea grass beds has been carried out,although it has been estimated thatthere may be between 500,000 and1,000,000 km

2

(M. Spalding, personalcommunication). Short and Wyllie-Echeverria (1996) have stated thatperhaps 900 km

2

of sea grass bedswere lost globally between 1985 and1995, although the basis for this esti-mate is not clear. Extrapolationwould give an annual decline of0.01–0.02%, although clearly littleconfidence can be attached to thisfigure.

Estuaries

We found no global assessment ofrates of loss of estuarine habitats.

Coastal Shelf and Open Ocean

The only measure of coastal-shelf habi-tat modification for which we foundglobal estimates was for disturbanceof the seafloor by bottom trawling. Itis not clear, however, what propor-tion of bottom trawling caused long-term habitat degradation, so we didnot use this estimate.

The marine component of theWWF LPI does not distinguish be-tween different marine biomes.Overall it indicates a 35% decline inabundance of marine fishes, mam-mals, birds, and reptiles over the pe-

riod of 1970–1999, yielding an aver-age annual decline of 1.5%. Furtherevidence for decline is provided byfitting a curve to FAO (2001) dataon changes since 1974 in the propor-tion of all the world’s marine fishstocks that are exploitable (i.e., stockswith status categories of overex-ploited, depleted, and recovering).Most fish stocks reduced to unex-ploitable levels show little evidence ofrecovery within 15 years of their de-cline ( Hutchings 2000) and so canbe regarded as effectively lost to ex-ploitation for the foreseeable future.The fitted curve suggests that exploit-able fish stocks have declined innumber at a rate of 1.0% per year.

Swamps, Floodplains, Lakes, and Rivers

There are no global estimates forrates of change in the extent ofthese habitats. The WWF LPI for in-land waters shows a decline of 51%between 1970 and 1999, yielding anaverage annual decline of 2.4%.

Grasslands, Rangelands, Deserts, and Tundra

There are no global estimates of ratesof change in the extent of these hab-itats or of overall changes in theircondition. The available data for ver-tebrate populations are currently in-adequate for developing a reliable LPIfor any of these biomes.

Discussion and Conclusions

Our review indicates that the con-version and modification of three ofthe four relatively unmodified natu-ral habitats for which we found dataon change in habitat area, usually foruse by humans, has continued intothe period since the United NationsConference on Environment and De-velopment in Rio de Janeiro in 1992.Declines in area have occurred fortropical forests, mangroves, and seagrass beds, whereas the area of tem-perate and boreal forests has in-creased. Unfortunately, there were

many habitats large in extent and im-portant for biodiversity and ecosystemservices for which we found no esti-mates of change in area. The list in-cludes estuaries, coral reefs, algalbeds, grasslands, savannahs, andfreshwater wetlands. Without glo-bal data on these important habitats,it is not possible to be certain thatthis pattern of decline applies towild nature as a whole. However,our estimates based on aggregatedtrends in the abundance of free-liv-ing vertebrates from the LPI showedaverage rates of change broadly simi-lar to those based on area. By includ-ing the LPI and FAO fisheries dataand averaging for habitats with morethan one estimate, we obtained areatrends for six habitats. Five of thesehabitats have declined (tropical forest,freshwater wetlands, mangroves, seagrass beds, and marine habitat) andone has increased (temperate andboreal forest). The mean of the sixrates of change is a decline of 1.1%per year (SE

�

0.47%).Comparison of estimates of change

in area and LPI changes is instruc-tive. In the case of temperate for-ests, the one habitat for which areaincreased rather than declined, bothapproaches showed a similar in-crease. For tropical forests, the rateof loss based on indices of verte-brate populations was considerablyhigher than either of the estimatesof net area loss. This is perhaps notsurprising, given the high levels ofexploitation of wild vertebrates forfood in tropical forests (Robinson &Bennett 2000; Fa et al

.

2002).The trend estimates for the marine

environment based on data for pop-ulations of all classes of free-livingvertebrates from the LPI and fromthe trend in exploitable stocks offish were broadly similar. We con-clude from this that indices based onanimal populations, although cur-rently rather crude and based on anonrandom set of vertebrates, haveconsiderable potential for assess-ment of the status of habitats. Furtherattention should be given to the sam-pling and selection of the taxa to in-

22

Issues in International Conservation Jenkins et al.

Conservation BiologyVolume 17, No. 1, February 2003

clude in the calculation of an indexand the statistical methods used toestimate a single trend from com-bined data on many taxa.

Ideally, an index of global trendfor a given habitat would be basedon information for a random sampleof species selected from withinspecified taxonomic and geographi-cal strata. To date, however, such anal-yses have necessarily tended to usewhatever population time-series infor-mation can be obtained from the sci-entific literature. Inevitably, the taxacovered are not representative ofthe specified subset of the biota ofthe selected habitat, and geographi-cal coverage tends to be strongly bi-ased. Furthermore, series varywidely in duration and in startingand ending points. These problemsmake the analytical methods used es-pecially important, although that mayseem at first to be a minor technicalissue. Differences in approach canlead to a wide divergence in inter-pretation of the same data series. Forexample, Houlahan et al

.

(2000) andAlford et al

.

(2001) analyzed thesame data set on trends in 936 am-phibian populations in differentways and reached markedly differ-ent conclusions about both globaltrends and regional variations (seealso Houlahan et al

.

2001).Although there is a clear need to

collect new data on populationsfrom a more widely representativeset of taxa and regions, it would be acounsel of despair to conclude thatgood use cannot be made of existinginformation. What is needed is fur-ther development, testing, and appli-cation of robust analytical tech-niques for combining time-series formultiple populations of differentspecies in ways that reduce theirsensitivity to bias from the usual lackof representative sampling.

We recognize that our primary ob-jective of finding estimates of thechange in area of broadly definedhabitats is itself insufficient. First,the apparent trend in a broadly de-fined habitat may be particularly sen-sitive to the details of the definition

used. Producing separate data forsubdivisions of a habitat and thenevaluating the sensitivity of thetrend in the aggregated area ofthem to changes in the componentsgrouped in different ways allowsthis possibility to be explored.

Second, information is also neededon the degradation of habitats as wellas changes in their area. For example,coral reefs affected by sediment depo-sition, or tropical forests depleted oflarge vertebrates by overexploitationof bushmeat, may still be the samehabitats by definition, but their stockof biodiversity, function, and servicedelivery may be very different. The LPIfigure for tropical forests indicatesthat vertebrate populations are be-ing depleted in these forests far fasterthan the forest itself is disappearing.The study by Achard et

al. (2002) isthe first we are aware of that attemptsa global assessment of degradation ina major biome, although even here itis unclear how the category of “visi-bly degraded” relates to ecosystemfunction and value for biodiversity. Itis evident that a combination of bothapproaches is likely to be requiredto build up any kind of comprehen-sive picture.

Our most striking finding is thedearth of information on trends at aglobal scale. A major source of infor-mation at a global level remains theFAO, which manages reasonably com-prehensive global data sets on for-estry and fisheries. As its full name im-plies, however, its primary focus is onfood security and the provision ofnatural resources for direct humanuse. Although the value of naturalhabitats to humans is high, it is fre-quently not recognized in economicdecisions about conversion, degrada-tion, and exploitation because of a lackof information, market failures, andintervention (Balmford et al

.

2002).Thirty years after the establishment

of the U.N. Environment Programme(UNEP) and 8 years after the entryinto force of the Convention on Bio-logical Diversity, there is no programfocused directly on documentingchanges in biodiversity and natural

habitats. The Millennium EcosystemAssessment will provide an account ofexisting data on biodiversity, but thereis a clear need to extend its missionto ongoing monitoring of globalchange in biodiversity based on sys-tematic and repeatable methods.There are, admittedly, many concep-tual and technical problems to beovercome, but these are far from in-surmountable. We found a range ofexcellent regional and local studies ofchanges in biodiversity that, given ad-equate investment and political will,could be used as models for a majorglobal monitoring effort. What seemsclearly lacking at present is the politi-cal will to implement such a program.Without it, the declaration of theWorld Summit on Sustainable Devel-opment is likely to continue to ringhollow and the target of reducingbiodiversity loss will remain elusive.

Acknowledgments

This study resulted from a workshopconvened by the Royal Society forthe Protection of Birds (RSPB) andsponsored by the RSPB and theUnited Kingdom Government’s De-partment for Environment, Food andRural Affairs. We thank A. Balmfordfor discussions.

Martin Jenkins

World Conservation Monitoring Centre, 219 Hunt-ingdon Road, Cambridge CB3 0DL, United Kingdom

Rhys E. Green

Royal Society for the Protection of Birds andConservation Biology Group, Department of Zo-ology, University of Cambridge, Downing Street,Cambridge CB2 3EJ, United Kingdom

Joah Madden

Royal Society for the Protection of Birds, TheLodge, Sandy, Bedfordshire SG19 2DL, UnitedKingdom

Literature Cited

Achard, F., H. D. Eva, H.-J. Stibig, P. May-aux, J. Gallego, T. Richards, and J.-P. Ma-lingreau. 2002. Determination of defor-

Jenkins et al. Issues in International Conservation

23

Conservation BiologyVolume 17, No. 1, February 2003

estation rates of the world’s humid

tropical forests. Science

297:

999–1002.Alford, R. A., P. M. Dixon, and J. H. K. Pech-

man. 2001. Global amphibian declines.Nature

412:

499–500.Balmford, A., et al. 2002. Economic reasons

for conserving wild nature. Science

297:

950–953.Bryant, D., L. Burke, J. McManus, and M. Spal-

ding. 1998. Reefs at risk. World ResourcesInstitute, New York.

Costanza, R., R. d’Arge, R. de Groot, S. Farber,M. Grasso, B. Hannon, K. Limburg, S.Naeem, R. V. O’Neill, J. Paruelo, R. G.Raskin, P. Sutton, and M. van den Belt. 1997.The value of the world’s ecosystem services

and natural capital. Nature

387:

253–260.

Fa, J.E., C. A. Peres, and J. Meeuwig. 2002.Bushmeat exploitation in tropical forests:an intercontinental comparison. Conser-vation Biology

16:

1–9.Food and Agriculture Organization (FAO).

2000. The state of the world fisheries andaquaculture. FAO, Rome.

Food and Agriculture Organization (FAO).2001. Global forest resources assessment2000: main report. FAO, Rome.

Houlahan, J. E., C. S. Findlay, B. R. Schmidt, A.H. Meyer, and S. L. Kuzmin. 2000. Quantita-tive evidence for global amphibian declines.Nature

404:

752–755.Houlahan, J. E., C. S. Findlay, A. H. Meyer, S. L.

Kuzmin, and B. R. Schmidt. 2001. Globalamphibian declines: reply. Nature

412:

500.

Hutchings, J. A. 2000. Collapse and recoveryof marine fishes. Nature

406:

882–885.Loh, J., et al. 2000. Living planet report 2000.

World Wildlife Fund, Gland, Switzerland.Robinson, J. G., and E. L. Bennett. 2000. Hunt-

ing for sustainability in tropical forests.Columbia University Press, New York.

Short, F. T. and S. Wyllie-Echeverria. 1996.Natural and human-induced disturbanceof seagrasses. Environmental Conserva-tion

23:

17–27.Stokstad, E. 2001. U.N. report suggests

slowed forest losses. Science

291:

2294.Valiela, I., J. L. Bowen, and J. K. York. 2001.

Mangrove forests: one of the world’sthreatened major tropical environments.BioScience

51:

807–815.