drivers of wetland disturbance and biodiversity impacts on a tropical oceanic island
TRANSCRIPT
Biological Conservation 149 (2012) 136–142
Contents lists available at SciVerse ScienceDirect
Biological Conservation
journal homepage: www.elsevier .com/locate /b iocon
Drivers of wetland disturbance and biodiversity impacts on a tropical oceanic island
Susan G.W. Laurance a,⇑, Claudia Baider b, F.B. Vincent Florens c, Shyamduth Ramrekha d,Jean-Claude Sevathian e, David S. Hammond f
a Centre for Tropical Environmental and Sustainability Science (TESS) and School of Marine and Tropical Biology, James Cook University, Cairns, Queensland 4878, Australiab Mauritius Herbarium, R.E. Vaughan Building, MSIRI, Réduit, Mauritiusc Department of Biosciences, University of Mauritius, Réduit, Mauritiusd Stem Consulting, Port-Louis, Mauritiuse Mauritian Wildlife Foundation, Vacoas, Mauritiusf NWFS Consultancy, 4050 NW Carlton Court, Portland, OR, 97229, USA
a r t i c l e i n f o a b s t r a c t
Article history:Received 31 October 2011Received in revised form 8 December 2011Accepted 12 December 2011Available online 28 February 2012
Keywords:Edge effectsFillingFloodingLand useUrbanizationWetland biodiversityWetland conversionWetland fragmentation
0006-3207/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.biocon.2011.12.015
⇑ Corresponding author. Address: School of MarineCook University, Cairns, Queensland 4870, Australia. T7 4042 1390.
E-mail address: [email protected] (S.G.W
Wetlands are biologically important elements of landscapes and among the most threatened ecosystemson Earth. On the island of Mauritius, many remaining wetlands are being rapidly converted and frag-mented by intense land-use demands. We surveyed 209 coastal wetlands on Mauritius to assess theirbiophysical attributes, land-use activities, and patterns of disturbance, to help identify factors that threa-ten wetland biodiversity. Most wetlands exhibited severe edge-related disturbances and more than halfwere fragmented. Plant species richness was highest in larger, unfragmented wetlands and lower in wet-lands with degraded margins. Urban wetlands were smaller and more likely to be fragmented than thoseadjoining other land uses such as grazing and agriculture. Flooding of urban homes and infrastructurewas more likely to occur near fragmented than natural wetlands. Ongoing wetland decline in Mauritiusnot only contributes to the loss of local biodiversity but reduces the larger ecosystem role these habitatsplay in regulating surface water and protecting adjacent marine habitats.
� 2011 Elsevier Ltd. All rights reserved.
1. Introduction
Despite international recognition that wetlands are enor-mously important for biodiversity, ecosystem health and humanwell-being, they continue to be degraded and destroyed fasterthan any other terrestrial ecosystem. About half of the wetlandsworldwide have been lost, converted or degraded in the twenti-eth century (World Resources Institute, 2005). Many causal fac-tors have been associated with wetland loss and degradationincluding vegetation clearing and drainage for agriculture, infra-structure expansion, invasive species, pollution and global cli-mate change (Daniels and Cumming, 2008; Foote et al., 1996;Zelder and Kercher, 2004).
Biologically, wetlands are unique ecosystems where aquaticand terrestrial life-forms intermix (Mitsch and Gosselink, 1986).When permanently inundated, wetlands require specializedplant-forms (hydrophytic species) that can survive both inunda-tion and low-oxygen soils and aquatic fauna that can tolerate
ll rights reserved.
and Tropical Biology, Jamesel.: +61 7 4042 1237; fax: +61
. Laurance).
varying oxygen levels (Keddy, 2000). The wet to dry transitionalong wetland ecotones frequently supports a distinctive, produc-tive and diverse biological community that is not well-representedelsewhere in the landscape (Carter et al., 1994; Chapman et al.,1996). Further, wetlands provide crucial habitats for terrestrialspecies that have aquatic life-stages or require water for survival(Keddy, 2000).
In addition to their biological values, wetlands have keyhydrological functions. They capture surface flows, slowing watermovement, and may promote aquifer replenishment (Gosselinkand Turner, 1978). Reducing water velocity allows large particu-late matter to settle, reducing sedimentation and nutrient pollu-tion of coastal waters (Keddy, 2000). In Mauritius, as in manyother tropical regions, coastal waters support sensitive lagoonenvironments that include inshore coral reefs and seagrass beds.In many parts of the tropics, such environments are being de-graded by nutrient pollution and sedimentation arising fromexpanding land-uses in coastal regions (Laws, 1992; Rogers,1990).
Wetlands provide further environmental services. On average,Mauritius has one cyclone approach annually within 100 km ofits coasts (Jury, 1993), producing heavy rains that can overwhelmthe landscape, impacting mountain slopes, rivers, floodplains and
S.G.W. Laurance et al. / Biological Conservation 149 (2012) 136–142 137
human communities. Wetlands and low-lying areas provide arepository for excess water from overrun rivers and surface flows.When wetlands are lost or people colonize floodplains, severe rain-fall events can cause a major loss of property and loss of life (Ked-dy, 2000). In many parts of Asia, for example, annual flooding fromhigh rainfall affect millions of people each year (United Nations,2009).
It is only in the last decade that wetlands in Mauritius have re-ceived much societal recognition and even partial legal protection.The country signed the Ramsar Convention in 1997 and designatedits first protected marine wetland in 2001. Despite some level ofinstitutional recognition, wetlands are still being rapidly convertedin Mauritius, with a �60% loss of wetlands in the northern regionof the country since 1980 (Government of Mauritius, 2002). Theunrelenting pressure on the wetlands in Mauritius highlights ageneral lack of understanding of their biological importance as wellas poor government policy for their protection.
Here we examine the causal factors associated with wetlanddisturbance and loss in coastal regions of Mauritius, an island thathas already suffered a severe loss of its native vegetation (Safford,1997) and biodiversity (Cheke, 1987). We assess the distribution ofcoastal wetlands and describe their biophysical attributes, andthen identify land-use activities that are their key threats.
80
100
q km
)
2. Methods2.1. Wetland location
Wetlands can develop under various landscape conditionsdepending on topography and dominant water sources, such asfrom riverine flooding, landscape depressions, non-permeableedaphic features, and estuarine or lake fringes (Gosselink andTurner, 1978). For the purposes of this study we defined wetlandsas areas of brackish water not actively attached to the marine envi-ronment via surface flows and characterized by soils, plants andanimals whose distributions are affected by permanent or frequentinundation.
We located wetlands based on 1:25,000 topographic maps(circa 1990), local government planning maps at 1:5000 scale(circa 2006), and expert opinion. The data presented here includeonly smaller coastal wetlands (<30 ha in size) and not largerprotected wetlands. Wetland delineation was based on the identi-fication of hydric soils, hydrophytic plants and the presence ofpermanent standing water or frequent inundation. Hydric soilsare saturated during the plant-growing season and developanaerobic conditions that favour the growth and regeneration ofhydrophytic vegetation (Sprecher, 2001). Hydric soil features caninclude sulfidic odor, gleying, redoximorphic features (such as areduced matrix and zones of Fe–Mn oxides), and oxidized rootzones (Sprecher, 2001).
0
20
40
60
0 1 2 3 4 5 6
Wet
land
are
a (s
River density (km per sq km)
Fig. 1. The relationship between river density and wetland area in Mauritius. Apower curve is fitted to the data for 17 river catchments (river den-sity = 20.243 � [wetland area�0.971]). A linear regression comparing observed andfitted values was highly significant (F1,17 = 27.38, R2 = 64.61%, P = 0.0001).
2.2. Wetland biophysical attributes
We surveyed up to 209 coastal wetlands using rapid assessmenttechniques, recording floristic, structural and physical attributesfor each wetland. The area, shape and circumference of each wet-land were determined by delineating the wetland edge with a GPSunit. Vegetation structure and composition were recorded bywalking the circumference of each wetland and traversing itscentre with 3–5 parallel transects, depending on wetland size.The presence of terrestrial and aquatic vertebrate (mammals, birds,reptiles, amphibians, and fish) and selected invertebrate (butter-flies and mollusks) species was determined by visual sightings,acoustical cues, and other signs (burrows, herbivory patterns,footprints, droppings, and bones) during our daytime surveys.
Soil and hydrological characteristics were assessed by digging3–5 pits to a depth of 25–50 cm. Soil chroma and value were de-fined using Munsell Soil Colour Charts (Munsell, 1992). Salinityof standing water (when it occurred) was measured using a salinityrefractometer with values ranging from 0 to 33. We estimated algalcover of the wetland surface area and water clarity, and noted thewater source and periodicity.
2.3. Anthropogenic disturbances
We assessed five features of anthropogenic disturbance to mostwetlands: (1) Edge disturbance: The proportion of the wetland mar-gin disturbed by human activities such as agriculture, wetland fill-ing and infrastructure. (2) Fragmentation: Three fragmentationconditions were recorded: natural state, fragmented by road, frag-mented by land-use activities. (3) Filling: The presence and compo-sition of fill and the extent of the edge that was filled. (4)Surrounding land uses: Categorized as no active land use, rural graz-ing, rural farming (sugarcane or small farmer), urban withouthouses or urban with houses (or golf courses). (5) Flooding risk:Determined by examining obvious floodwater marks (not season-ally defined) on nearby houses and infrastructure (with the censusalso occurring during the wet season) with the following ranking:none, infrastructure, houses.
2.4. Data analysis
We examined wetland distribution across the island of Mauri-tius at different spatial scales. First, we investigated the relation-ship between topography and wetland distribution withinwatersheds using a regression analysis of wetland area and riverdensity. Second, we evaluated the spatial relationships betweenwetlands with nearest-neighbor comparisons of pairwise lineardistances among sites.
From >600 wetland transects, we identified major gradients inthe vegetation composition of wetlands using Nonmetric Multidi-mensional Scaling (McCune and Mefford, 1999). Monte Carlo ran-domization tests (100 runs) were used to determine whetherordination axes explained significantly more variation than ex-pected by chance. To improve ordination performance, singletonspecies (those recorded at only one wetland) were excluded fromanalyses. We used Spearman rank correlations to determinewhether species richness of floral and faunal species were
138 S.G.W. Laurance et al. / Biological Conservation 149 (2012) 136–142
associated with continuous environmental variables. Human land-use, fragmentation and filling activities were measured as categor-ical variables and analysed using v2 and Mann–Whitney U-tests.
3. Results
3.1. Wetland distribution and spatial correlation
Surviving wetlands on Mauritius cover an area of �22 km2 andare largely clustered in the northern and eastern parts of the island.This pattern is well illustrated by the negative and nonlinear rela-tionship between wetland area and river density (Fig. 1). Highmountains in the southern and western parts of the island ensurewater moves rapidly across the landscape via rivers whereas thenorthern and eastern areas offer little topographical relief andsurface water is more likely to form wetlands.
0
10
20
30
40
50
60
70
80
10 25 50 75 100
250
Nearest Neighbour Distan
No.
of W
etla
nds
0
1000
2000
3000
4000
5000
6000
7000
8000
50 75 100
250
500
750
1000
Inter-Wetland Distanc
No.
of W
etla
nd P
airs
(B)
(A)
Fragmentation
Fig. 2. Wetlands are clustered locally (2A) and spatially clumped (2B) across the is
We examined wetland clustering at local and landscape scalesacross the island (Fig. 2). At the local scale, most individual wet-lands occurred within distances of 250–1000 m from their nearestneighbour (Fig. 2A). An investigation of the pool of all wetlanddistance pairs showed three detectable aggregations of wetlanddistance in relation to spatial scale: (Fig. 2B): (1) small-scaleaggregations (<100 m inter-wetland distance), which reflectedfragmentation of once-contiguous wetlands; (2) intermediate-scale (500 m) clustering that is probably a topographical effect,created by the patchy distribution of lowland plains along thenorth and east coastal areas; and (3) large-scale clustering(5000 m) that results from island geography, reflecting the factthat most lowland wetlands in Mauritius are coastal. The sizes ofindividual coastal wetlands are log-normally distributed, rangingacross four orders of magnitude and with the largest being 5–10times larger than the mean wetland size of 2.1 ha.
500
1000
2500
5000
7500
1000
0
2500
0
ce (m) quarti-log10 intervals
2500
5000
7500
1000
0
2500
0
5000
0
7500
0
e (m) (interval maximum)
Topography
Coastal geometry
land of Mauritius based on nearest-neighbor and all-wetland distance values.
Typha domingensisPaspalum distichumMikania micrantha Ipomoea aquatica
Acrostichum aureumRhizophora mucronata
Acro
stic
hum
aur
eum
Pasp
alum
dis
tichu
mPa
spal
idiu
m g
emin
atum
Cyn
odon
dac
tylo
n
AXIS 1
AXI
S 2
More More
Mor
eM
ore
Fig. 3. An ordination of plant communities from the coastal wetlands fromMauritius. Site positions are markedly influenced by three main species associa-tions: (1) mangroves and Acrostichum ferns, (2) Typha, Mikania and Ipomoea, and (3)grassland species such as Paspalum.
Table 1Pearson correlations for eight common plant species with two ordination axesproduced by Nonmetric Multidimensional Scaling (only species exhibiting significantassociations with ordination axes are listed here). Axes 1 and 2 represent wetlandsthat are influenced by gradients in salinity and water inundation. Values in bold weresignificant using a Bonferroni-corrected alpha value (P = 0.0022).
Species Salinity gradient
Permanent wetlands(Axis 1)
Seasonally-inundatedwetlands (Axis 2)
Acrostichumaureum
�0.657 0.526
Cynodon dactylon 0.159 �0.234Ipomoea aquatica 0.254 �0.021Mikania
micrantha0.266 0.081
Paspalumdistichum
0.340 �0.386
Paspalidiumgeminatum
0.129 �0.669
Rhizophoramucronata
�0.454 �0.147
Typhadomingensis
0.781 0.168
Table 2Spearman rank correlations between the observed species richness estimates of wetland f
Species Algae Area Clarity Edge disturb
FaunaEndemic NS 0.200* NS �0.198*
Migrant 0.395*** 0.273** NS �0.224*
Endangered 0.205* NS NS NSVulnerable NS NS NS NSTotal fauna 0.229* 0.397*** NS NS
FloraWet-flora 0.350*** 0.244* NS NSDry-flora NS 0.235* 0.383*** �0.360***
Total flora 0.216* 0.343*** 0.207* �0.298**
Total species 0.258** 0.453*** NS �0.202*
NS: not significant.* P < 0.05.** P < 0.01.*** P < 0.001.
S.G.W. Laurance et al. / Biological Conservation 149 (2012) 136–142 139
3.2. Biophysical attributes of wetlands
In total, 88 plant species were recorded from 608 wetland tran-sects. We used an ordination (Nonmetric Multidimensional Scal-ing) analysis to examine vegetation structure and communitycomposition of all coastal wetlands, using 45 of the more-common(non-singleton) plant species (Fig. 3). Two ordination axes ex-plained most (59%) of the variation in the plant communities, col-lectively discriminating three main vegetation communitiesassociated with water permanence and salinity (Table 1). Axis 1,which captured 36% of the variation, described a gradient betweenplant species that prefer permanent, nutrient-rich and freshwaterwetlands, such as Typha domingensis, Mikania micrantha, and Ipo-moea aquatica, and those tolerant of saline conditions, such asthe aquatic fern Acrostichum aureum and mangrove Rhizophoramucronata (Axis 1 vs. salinity; rs = �0.499, P < 0.0001, n = 151;Spearman rank correlation). Vegetation cover was higher in wet-lands dominated by Typha, Mikania and Ipomoea, and lower in wet-lands with Acrostichum and Rhizophora. Water salinity influencedvegetation cover, with open wetlands dominated by Acrostichumand Rhizophora occurring with high levels of salinity and vegetatedwetlands dominated by Typha, vines and grasses with low salinity.
Axis 2 captured 23% of the total variation and described a veg-etation gradient affected by water inundation and salinity. Plantspecies significantly and negatively correlated with Axis 2 includethe grass species such as Paspalum distichum, Paspalidium gemina-tum and Cynodon dactylon, which occurred at sites seasonallyinundated with low salinity, whereas the salt-tolerant Acrostichumaureum, which was positively correlated with Axis 2, occurred inwetlands with permanent water (Axis 2 vs. salinity: rs = 0.160,P = 0.050, n = 151; Spearman rank correlation).
Water clarity, an index of suspended sediment in wetlands, wassignificantly correlated with vegetation composition and structure.Low levels of suspended sediment were detected in the open-water wetlands dominated by Acrostichum and high levels in thevegetated wetlands dominated by Typha, vines (Mikania and Ipo-moea) and grasses (Axis 1 vs. water clarity; rs = �0.243, P = 0.001;Axis 2 vs. water clarity: rs = 0.256, P = 0.0005; n = 181; Spearmanrank correlations).
Biodiversity was generally higher in larger wetlands. Wetlandsize was positively correlated with seven measures of speciesrichness including endemic, migrant and total faunal species (butnot endangered and vulnerable species); dry, wet, and total floralspecies; and total plant and animal species richness (Table 2). Edgedisturbance, however, was negatively correlated with six measuresof species richness, including endemic and migrant faunal species,
auna and flora and environmental variables for Mauritian wetlands (n = 108).
ance Fauna Flora Salinity Vegetation cover
– NS NS NS– 0.385*** NS �0.251**
– NS NS NS– NS NS NS– 0.249** NS NS
NS – NS NS0.193* – 0.269* �0.265**
0.249** – NS NS– – NS NS
Urban0
20
40
60
80
100
No backfillVolcanic rockConstruction debrisRubbish
Natural Grazing Farming
>6 ha<0.5 ha <2 ha <6 ha0
10
20
30
40
50
60
70
80
NaturalGrazingFarmingUrban
Severe >90%
0
20
40
60
80
100
120
Very low <10%
Low<25%
Moderate >25%
High>75%
>6 ha0
10
20
30
40
50
60
70
80
NaturalRoadsFragmented
<0.5 ha <2 ha <6 ha
No.
of W
etla
nds
(C) Adjacent landuse & wetland area
(A) Edge disturbance (% total edge)
(D) Wetland filling & adjacent landuse
(B) Fragmentation & wetland area
Fig. 4. The impact of human activities on Mauritian wetlands from edge disturbance (% of total edge) (A), fragmentation (B), adjacent landuse (C), and wetland filling (D).
140 S.G.W. Laurance et al. / Biological Conservation 149 (2012) 136–142
and dry plant, total plant, and total plant and animal species rich-ness (Table 2). Wetland communities showed relatively high fau-nal diversity with 57 native species recorded overall. Snails werethe most diverse taxon with 21 species detected. Although 14 spe-cies of migratory bird have been recorded previously for Mauritiuswetlands (Annotated Ramsar List: Mauritius, 2001), only three
migratory birds were detected in this survey, which occurred overthe austral winter. Total faunal species richness was positively cor-related with dry and total plant species but not with wet speciesrichness (Table 2), suggesting that the non-flooded habitat thatborders wetlands is an important resource for Mauritian faunalspecies.
S.G.W. Laurance et al. / Biological Conservation 149 (2012) 136–142 141
3.3. Human disturbance of wetlands
Approximately two-thirds of wetlands in Mauritius are underprivate ownership and these are typically 30% smaller in size thanwetlands owned by the State. The largest private ownership ofwetlands occurred in the northern region where multiple ownersof small parcels have dissected the wetlands. Outside of this region,private wetlands were typically under single ownership.
The most pervasive human impact on wetlands was edge dis-turbance, with half (105 of 209) of all wetlands showing severe(P90% of edge) disturbance and only 11 being largely intact(<10% edge disturbed) (Fig. 4A). About two-thirds (65%) of the wet-lands were fragmented by land use or roads (Fig. 4B). Wetlandsfragmented by land use (excluding roads) were significantly smal-ler than were natural wetlands (P = 0.0003, Mann–Whitney U-test),averaging 0.97 ± 2.07 ha for fragmented and 2.73 ± 3.24 ha for nat-ural wetlands. Urbanisation was the land-use activity most likelyto be associated with fragmentation (Fig. 4C; v2 = 65.83, df = 8,P < 0.0001; Chi-square test). Edge disturbance was also signifi-cantly higher in wetlands fragmented by land use (P < 0.0001;Mann–Whitney U-test), averaging 89.9 ± 12.9% of fragmented and50.8 ± 38.8% of the natural wetlands. Filling of wetland edges wasthe most damaging direct human activity, affecting 87% of all wet-lands. Although most wetlands had some historic filling from rockthat was removed from nearby agricultural lands, current fillingwas more apparent in urbanized areas and included more house-hold and construction wastes (Fig. 4D; v2 = 50.7, df = 9,P < 0.0001). Within the urban landscape we observed more flood-ing of roads and homes adjacent to fragmented than natural wet-lands (v2 = 11.67, df = 2, P = 0.003, n = 161; Chi-square tests).
Species richness was lower in fragmented wetlands for endemicfaunal (P = 0.003) and total faunal species (P = 0.02; Mann–Whit-ney U-tests). We observed no such effect for plant species.
4. Discussion
Our findings reveal that most coastal wetlands in Mauritius suf-fer from severe edge disturbances arising from adjoining humanactivities. Land conversion caused by wetland filling and fragmen-tation is reducing wetland area and leading ultimately to the lossof nearly all wetlands in urban areas. The rapid development ofwetlands in urban landscapes can cause a range of environmentaldisturbances (Azous and Horner, 2001), but ultimately it occurs be-cause the lands are readily available and cheap (Hall, 1988) and areactively being filled as sites for waste disposal (Jayawickreme,2011).
Coastal wetlands in Mauritius were concentrated in flat, low-ly-ing areas where river densities were low. We observed three aggre-gation patterns in wetland distribution that resulted from land-useactivities and natural island topography and landform. Wetlandswere closely clustered and patchily distributed around the coastalregions of the island. Small wetlands (60.5 ha) were likely to ad-join urban land uses whereas larger wetlands were more likelyto adjoin farming (0.5–2 ha) or grazing (2–6 ha). The positivehydrological functions of wetlands in an urban landscape are pre-dicted to diminish with increasing fragmentation and diminishingwetland size (Mitsch and Gosselink, 2000), because wetland func-tion tends to occur over a larger landscape scale (Johnston, 1994).With a lower wetland reservoir area and with housing increasinglyconstructed in low-lying areas, a higher flooding risk to propertyand infrastructure was regularly detected in these urban areas.
Larger wetlands appeared to support more faunal and floralspecies, including endemic and migrant fauna. These trends are ex-pected given the species-area relationship (Rosenzweig, 1995), butit was the discovery that the bordering shrub and woodland con-
tributed substantially to biological diversity in intact wetlands thatwe believe is important. This highlights the need to include bufferzones around wetlands to protect them against deterioration oftheir important ecotonal vegetation. Upland buffers around wet-lands are generally acknowledged to protect water (Lowranceet al., 1984) and wildlife resources (Spackman and Hughes, 1995)with a recent review of amphibians and reptiles recommendingprotection zones of >250 m from the waters’ edge (Semlitsch andBodie, 2003).
Wetlands in Mauritius vary in species composition. Our com-munity analysis identified a range of wetland vegetation typeswith species associations indicative of different environmentalconditions; for example, we surveyed saline-tolerant species suchas mangroves and Acrostichum ferns, seasonal wetlands with theirgrass-dominant communities and permanent wetlands dominatedby Typha and Mikania. These different vegetation communitiesshowed no clear association with wetland size although wetlandsdominated by Acrostichum appeared less fragmented than othercommunities. These saline-tolerant plants often occurred in low ly-ing areas close to the sea, which may reduce their desirability fordevelopment. Wetlands adjacent to grazing were frequentlygrazed themselves and were dominated by grassland species thatare tolerant to inundation. These wetlands were unique in thateither their ongoing grazing or private ownership protected themfrom being used for waste disposal.
The widespread destruction of wetland edges from filling hashad serious negative effects on faunal and floral species richness.The filling process promotes the loss of the non-flooded shruband woodland habitats that border wetlands and appear to en-hance their species richness (Castelle et al., 1994). In addition, wet-land edges are important nursery and refugia areas for endemicspecies such as native fish (Chapman et al., 1996).
The rapid conversion of wetlands globally is occurring for anumber of reasons including their perceived low economic value,inadequate regulations designed to protect wetlands and weakgovernance or enforcement (Jayawickreme, 2011). Mauritius isone of the most densely populated countries in the world (UnitedNations, 2011), with available land in high demand. Hence, it is notsurprising that urban conversion is the leading driver of wetlandloss and degradation in Mauritius, not unlike many other placesin the world (Radeloff et al., 2010). The diminishing area of wet-lands also reduces their important ecosystem functions such asflood protection, aquifer replenishment and nutrient and sedimen-tation control (Gosselink and Turner, 1978). In addition to the bio-logical values of the wetlands, the loss of these services could havelong-term consequences for the Mauritian environment and resi-dents. We encourage the Mauritian government to protect andrehabilitate its critically imperiled wetlands before they declineeven further.
Acknowledgments
We thank the Mauritian government staff and landholders fortheir assistance, access to data and study sites. W.F. Laurance com-mented on the manuscript and W. Edwards provided statisticaladvice.
References
Annotated Ramsar List: Mauritius, 2001. The Annotated Ramsar List of Wetlands ofInternational Importance: Rivulet Terre Rouge Estuary Bird Sanctuary,Mauritius, ed. Ramsar, Gland, Switzerland.
Azous, A.L., Horner, R.R., 2001. Wetlands and Urbanization: Implications for theFuture. Lewis Publishers, Boca Raton, FL, USA.
Carter, V., Gammon, P.T., Garrett, M.K., 1994. Ecotone dynamics and boundarydetermination in the great dismal swamp. Ecol. Appl. 4, 189–203.
Castelle, A.J., Johnson, A.W., Conolly, C., 1994. Wetland and stream buffer sizerequirements – a review. J. Environ. Qual. 23, 878–882.
142 S.G.W. Laurance et al. / Biological Conservation 149 (2012) 136–142
Chapman, L.J., Chapman, C.A., Chandler, M., 1996. Wetland ecotones as refugia forendangered fishes. Biol. Conserv. 78, 263–270.
Cheke, A.S., 1987. An ecological history of the Mascarene islands, with particularreference to extinctions and introductions of land vertebrates. In: Diamond,A.W. (Ed.), Studies of Mascarene Island birds. Cambridge University Press,Cambridge, pp. 5–89.
Daniels, A.E., Cumming, G.S., 2008. Conversion or conservation? Understandingwetland change in northwest Costa Rica. Ecol. Appl. 18, 49–63.
Foote, A.L., Pandey, S., Krogman, N.T., 1996. Processes of wetland loss in India.Environ. Conserv. 23, 45–54.
Gosselink, J.G., Turner, R.E., 1978. The role of hydrology in fresh water wetlandecosystems. In: Good, R.E., Whigham, D.F., Simpson, R.L. (Eds.), FreshwaterWetlands: Ecological Processes and Management Potential. Academic Press,New York, pp. 63–78.
Government of Mauritius, 2002. Study of Environmental Risks in Grand Baie Area,ed. Ministry of Environment, Port Louis, Mauritius.
Hall, R.S., 1988. Urban wetlands: are 404(b)(1) guidelines effective? In: Kusler, J.A.,Daly, S., Brooks, G. (Eds.), Proceedings of the National Wetland Symposium:Urban Wetlands. Association of State Wetland Managers, Berne, NY, USA, NewYork, pp. 231–233.
Jayawickreme, D.H., 2011. Threats to Sri Lanka’s urban wetlands. Frontiers Ecol.Environ. 9, 320–321.
Johnston, C.A., 1994. Cumulative impacts to wetlands. Wetlands 14, 49–55.Jury, M.R., 1993. A preliminary study of the climatological associations and
characteristics of tropical cyclones in the SW Indian Ocean. Meterol. Atmos.Phys. 51, 101–115.
Keddy, P.A., 2000. Wetland Ecology: Principles and Conservation. CambridgeStudies in Ecology, Cambridge, UK.
Laws, E.A., 1992. Aquatic Pollution. John Wiley and Sons, New York.Lowrance, R., Todd, R., Fail, J., Hendrickson, O., Leonard, R., Asmussen, L., 1984.
Riparian forests as nutrient filters in agricultural watersheds. Bioscience 34,374–377.
McCune, B., Mefford, M.J., 1999. PC-ORD: Multivariate Analysis of Ecological Data.Version 4. MjM Software Design, Gleneden Beach, Oregon, USA.
Mitsch, W.J., Gosselink, J.G., 1986. Wetlands. Van Nostrand Reinhold, New York.Mitsch, W.J., Gosselink, J.G., 2000. The value of wetlands: the importance of scale
and landscape setting. Ecol. Econ. 35, 25–33.Munsell, 1992. Munsell Soil Color Chart. Kollmorgen Instrument Corporation,
Baltimore.Radeloff, V.C., Stewart, S.I., Hawbaker, T.J., Gimmi, U., Pidgeon, A.M., Flather, C.H.,
Hammer, R.B., Helmers, D.P., 2010. Housing growth in and near United Statesprotected area limits their conservation value. PNAS 17, 940–945.
Rogers, C.S., 1990. Responses of coral reefs and reef organisms to sedimentation.Mar. Ecol. Prog. Ser. 62, 185–202.
Rosenzweig, M.L., 1995. Species Diversity in Space and Time. Cambridge UniversityPress, Cambridge.
Safford, R.J., 1997. A survey of the occurrence of native vegetation remnants onMauritius in 1993. Biol. Conserv. 80, 181–188.
Semlitsch, R.D., Bodie, J.R., 2003. Biological criteria for buffer zones around wetlandsand riparian habitats for amphibians and reptiles. Criterios Biológicos paraZonas de Amortiguamiento Alrededor de Hábitats de Humedales y Riparios paraAnfibios y Reptiles. Conserv. Biol. 17, 1219–1228.
Spackman, S.C., Hughes, J.W., 1995. Assessment of minimum stream corridor widthfor biological conservation: species richness and distribution along mid-orderstreams in Vermont, USA. Biol. Conserv. 71, 325–332.
Sprecher, S.W., 2001. Basic concepts of soil science. In: Richardson, J.L., Vepraskas,M.J. (Eds.), Wetland Soils – Genesis, Hydrology, Landscapes and Classification.CRC Press LLC, Florida, USA.
United Nations, 2009. Global Report on Human Settlements 2009: PlanningSustainable Cities. United Nations Habitat: For a Better Urban Future.
United Nations, 2011. World Population Prospects: The 2010 Revision, ed.Department of Economic and Social Affairs – Population Division, New York.
World Resources Institute, 2005. Millennium Ecosystem Assessment. Ecosystemsand Human Well-being: Wetlands and Water Synthesis. Washington, DC.
Zelder, J.B., Kercher, S., 2004. Causes and consequences of invasive plants inwetlands: opportunities, opportunists and outcomes. Crit. Rev. Plant Sci. 25,431–452.