You are on page 1of 23

Wetlands Inventory, Mapping and Land

Cover Index Assessment on Mauritius

Carne M.Mamoun, Rody Nigel &


Soonil D.D.V.Rughooputh

Wetlands
Official Scholarly Journal of the Society
of Wetland Scientists
ISSN 0277-5212
Wetlands
DOI 10.1007/s13157-013-0415-z

1 23

Your article is protected by copyright and


all rights are held exclusively by Society
of Wetland Scientists. This e-offprint is for
personal use only and shall not be selfarchived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com.

1 23

Author's personal copy


Wetlands
DOI 10.1007/s13157-013-0415-z

ARTICLE

Wetlands Inventory, Mapping and Land Cover Index


Assessment on Mauritius
Carne M. Mamoun & Rody Nigel &
Soonil D. D. V. Rughooputh

Received: 22 August 2012 / Accepted: 18 March 2013


# Society of Wetland Scientists 2013

Abstract Past studies conducted on wetlands of Mauritius


lead to the conclusion that half of the wetlands have been
backfilled for touristic and housing development and that the
ecological condition of the remaining wetlands is being seriously challenged by numerous threats, natural and anthropogenic. This research aimed to fill the information gap
concerning wetland type and distribution. For this, using
published maps and satellite imagery, wetlands were digitised
resulting in 144 wetlands and categorised into 8 ecological
units. Afterwards, a number of wetlands were ranked
according to their environmental condition based on a series
of Land Cover Indices (LCIs). These indices were derived by
analysis of land cover types and slope gradient within a 50 m
and a 950 m watershed-bounded buffer zone. Wetlands in
forested areas were the least disturbed, with LCI of typically
0.30, but potentially threatened by sediment accumulation due
to a high slope gradient (>20 %). Three wetlands in Mauritius
are classified as Ramsar wetland. One of them, the Blue Bay
Marine Park (BBMP) has moderately good condition
(LCI= 0.55). The other one, the Rivulet Terre Rouge
Estuary Bird Sanctuary (RTREBS) was among the most
heavily impacted (LCI=0.87) due to intensive urbanization.
Electronic supplementary material The online version of this article
(doi:10.1007/s13157-013-0415-z) contains supplementary material,
which is available to authorized users.
C. M. Mamoun : S. D. D. V. Rughooputh
Faculty of Science, University of Mauritius, Rduit, Mauritius
C. M. Mamoun
e-mail: carene.mamoun@gmail.com
S. D. D. V. Rughooputh
e-mail: sdr@uom.ac.mu
R. Nigel (*)
INRSCentre Eau Terre Environnement, 490, de la Couronne,
Qubec, (QC) G1K 9A9, Canada
e-mail: rody.nigel@ete.inrs.ca

Keywords Wetlands . Mauritius . Mapping . GIS .


Inventory

Introduction
Mauritius is an Island State in the Indian Ocean which is
surrounded by seawater and which contains 25 major river
basins (WRU 2003). The river systems combined with the
diverse topography of the island have given rise to numerous wetlands, particularly along the coast, and include estuaries, marshes, mangroves and floodplains. These wetlands
are often used for recreational and fishing activities by the
local population but also for religious purposes, such as the
Grand Bassin Crater Lake (Proag 1995; Saddul 1995). The
local Environment Protection Act 2002 recognizes wetlands
as forming an integral part of the coastal resources of
Mauritius, such that permission is required before any development work is to be performed in a 50-m buffer zone
around the wetland.
The history of wetland loss is closely linked to the
residential, industrial and agricultural development of the
country, where agriculture today occupies 55 % of the total
surface area of the island, out of which 98 % is sugarcane
cultivation (Saddul 2002). More recently, the increasing
demand in land capacity for the construction of hotels,
Integrated Resort Scheme (IRS) projects, business parks
and residential areas, is posing a serious threat to remaining
wetlands existence and condition (MOE 2005; Laurance et
al. 2012). Anthropogenic activities leading to point and nonpoint sources of pollution are also considered as threats to
wetland condition (MOE 2005). A study commissioned by
the Ministry of Environment concluded that the loss of 70 %
of wetland extent in a northern coastal region was directly
related to flooding problems in this area (MOE and GIBB
2002; NWFS et al. 2008).

Author's personal copy


Wetlands

In parallel, efforts are being made by local authorities to


conserve and protect the remaining wetlands. A concrete
example is the adherence of Mauritius to the Ramsar
Convention in September 2001, which has led to the designation of three Ramsar sites until now: the Rivulet Terre Rouge
Estuary Bird Sanctuary (RTREBS), the Blue Bay Marine Park
(BBMP) and the Pointe dEsny Wetland. The designation of
these three Ramsar sites have certainly helped in the international and national recognition of local wetlands.
Concerning local legislation, there are 4 Acts which
directly relate to wetland protection namely, The Wildlife
and National Parks Act 1993 under which the Rivulet Terre
Rouge Estuary Bird Sanctuary was declared nature reserve
in 1999, The Fisheries and Marine Resources Act 1998
(under this Act, two marine parks and six fishing reserves
have been declared marine protected areas), and The
Environment Protection Act 2002, where wetlands are cited
as forming part of the coastal resources of the island. The
latest act is The Planning and Development Act 2004, which
stipulates that any development within a 15 m buffer restriction zone around a wetland is prohibited. Past this restriction
zone, a 35 m buffer control zone exists where only certain
developments respecting specific environmental conditions
are allowed.
The aim of this paper is twofold. First, a wetland inventory and classification including detailed information about
wetlands climatic patterns, geological and topographic settings will be conducted using geographic information systems (GIS). With this primary information alone, a wetland
database can be set up which should act as a good starting
point for increased knowledge about local wetland types and
distribution across the country. Secondly, a wetland assessment methodology making use of GIS software and land
cover map will then be conducted to investigate wetland
condition based upon existing land cover patterns. Wetland
assessment methods making use of GIS, such as the synoptic approach developed by Leibowitz et al. (1992) and the
modified synoptic approach of Brooks et al. (2004) are
examples of how GIS can be used to prioritise wetland
restoration. However the former is heavily data-dependent
and requires information from hydrological gauging stations
while the latter uses a 1-km radius for wetland level-1
assessment. The latter approach, while making very broad
assumptions about land use impacts on wetlands, enables
results to be obtained that are comparable to more intensive
assessment and thus can be adopted as a first step toward
determining the condition of wetlands on a watershed basis
(Brooks et al. 2004).
Accordingly, for our wetland assessment methodology, we
will develop a method that makes use of available local
information, such as, the land cover map of Nigel and
Rughooputh (2010a) and soil and geological maps. A 1-km
buffer analysis which integrates the 50-m buffer zone

mentioned in local legislation will also be used as a primary


factor for the methodology. Such a land cover approach
should provide a good indication of the extent and nature of
the environmental stresses that individual wetlands are facing
as mentioned before.
Results will include a series of land cover indices to rank
wetlands from least impacted to most impacted. The main
advantage provided by this method is that wetland protection and conservation can be targeted effectively using minimum financial and human resources using GIS. Coupled
with field surveys at later stage, information obtained from
the inventory and land cover assessment can be updated and
improved on a regular basis by any wetland practitioner
such as the National Parks Conservation Service (NPCS).
Such a follow-up on changing land cover patterns may thus
help in developing an appropriate course of action in a
timely manner by local wetland managers. Information
gathered from this method may also help to develop comprehensive water quality monitoring programs and increase
public awareness regarding the wise-use of wetlands.
Study Area Description
Mauritius is situated in the southwest of the Indian
Ocean (Fig. 1) at latitude 2010S and longitude
5730E. The island has an elliptical shape with a major
axis 63 km, minor axis 43 km, surface area 1,859 km2
and the highest peak 828 m altitude. The island is of
volcanic origin with eruptions lasting 100.025 million years
before present. Two main soil groups exist: mature Latosols
originating from highly weathered basaltic lava rock and
immature Latosolic soils with minerals still in the process of
weathering (Parish and Feillaf 1965).
The climate is tropical maritime with two seasons, a rainy
summer from NovemberApril dominated by cyclone passage and a dry winter from MayOctober dominated by the
South-East Trade Wind and frontal systems. About 70 % of
mean annual rainfall is received during summer. February is
the wettest and hottest month, whereas October is the driest
month. Mean annual rainfall is ~2,000 mm equivalent to
~3,700 Mm3, of which annual evaporation is 30 %, surface
runoff 60 % and groundwater recharge 10 % (WRU 2003).
Surface runoff is confined within 213 river basins (Fig. 1),
which occupy 83 % of the island (Nigel and Rughooputh
2010b). Torrential flows with severe bank erosion and turbidity in the lagoon are common during intense rainfall
events (Arlidge and Wong You Cheong 1975).
Over time, most of the islands indigenous vegetation
(e.g., ebony forestsDiospyros tesselaria; bois dOlive
Elaeodendron orientale; and aloesFurcraea gigantea) had
been removed to make room for extensive sugarcane cultivation, which to-date occupies 55 % of the island and 98 %
of its cultivated lands. The other cultivations are tea,

Author's personal copy


Wetlands

Fig. 1 Study area

vegetables and fruit plantations. Only ~4 % of indigenous


vegetation remains. Scrub covers ~11 % of the island, forest
27 % (including indigenous vegetation) and urban areas
6 %. The remaining 1 % is water bodies, wetlands and
sandy beaches (Saddul 2002).

four broad classification systems namely: i) Estuarine systems, ii) Riverine systems, iii) Lacustrine systems and iv)
Palustrine systems (Breen et al. 1997). Dugan (1990) acknowledges seven main landscape units, based on the
Ramsars list of wetland types, which are: estuaries, open
coasts, floodplains, freshwater marshes, lakes, peatlands and
swamp forest.

Materials and Methods


The methodology adopted for gathering wetland inventory
data has been broken down into three parts: (1)Analysis of
land use map series of Mauritius; (2) Collection of digital
data layers related to wetlands physical attributes and distribution, (3) Grouping of wetlands using the SADC wetland classification and the landscape unit approach of
Dugan (1990). The SADC wetland classification recognizes

Analysis of Map Series and Satellite Imagery


for Determination of Local Wetland Types
To determine local wetland types, an analysis of the published land use (LU) map series of Mauritius at 1: 25,000
scale can be made. Also, freely available high resolution
satellite imagery viewable on Google Earth can be used.
The LU map was produced in 1991 and is updated yearly. In

Author's personal copy


Wetlands

the present case, we had the updated version up to 2005


(MHL 2005). The LU map represents the most official data
about wetland distribution for Mauritius, having been produced from a combination of aerial photography and field
surveys (GOM-OS 1991).
The legend of the LU map gives a comprehensive listing
of all water bodies mapped (they were grouped under the
heading features of interest in the legend). This heading
contained anything that could be linked to a water body,
namely, Watercourse, Waterfall, Rapids, Waterhole, Marsh
or Swamp, Mangrove, Coral and Reservoir.
However, it was noticed that some LU features present on
the map were not identified on the map legend, particularly
those with open water characteristics. A total of five such
features were identified, and thus were added to the
Features of interest list. These are: basin, pond, salt pans,
mudflats and rivers. Two more wetland types were added,
namely, estuaries which, by definition, are present wherever
freshwater from mainland, in this case a river, meets with
seawater and floodplains, which were represented as marshy
land on either side of a river channel on the LU map.
The hard copy LU paper maps were scanned and georeferenced in GIS and used along with the vector updates by
MHL (2005). However, not all of the Features of interest
were explicitly digitized and/or updated by MHL (2005).
Attention was mostly paid by MHL (2005) to digitize roads,
urban areas, tracks, etc. Therefore, both the digital LU maps
of MHL (2005) and the initial ones (scanned and georeferenced) were used to identify the various wetland types.
Apart from looking at wetlands on the map, the locations of
wetlands from the LU map were also re-viewed on Google
Earth (GE) satellite imagery, which for Mauritius dates
from 2000 to 2011. Wetlands with open water characteristics
are typically represented on GE satellite imagery by a light
green colour, while surrounding vegetation is shown by a dark
green colour. To verify a wetlands location on GE satellite
image, a particular land use reference features such as coastline geomorphology and/or road intersections are first identified on the land use map and then compared with those on the
satellite image. However, wetlands that can be identified with
this method include only those that were of reasonable size
and easily discernible. Hence, wetlands too small in size or
lacking presence of water cannot be verified with this method.
Based on the information gathered from the land use
map, a total of 8 ecological units were identified for the
wetland classification and digitization process. These are:
&

&

Marine coastsincluding coastal wetlands which, either


form part of the lagoon or have an open connection with
the sea, are not fed by inland freshwater and do not form
part of the other eco-units
Estuariesidentified from the LU map, and typically
are enclosed coastal body of brackish water that are fed

&
&
&
&
&
&

by inland freshwater from rivers, and with a free connection to the open sea
Mangrovesidentified from the LU map, and which are
the various types of medium height trees and shrubs that
grow in saline coastal sediment habitats
Marsh/Swamppresent on the LU map, and which
typically is a wetland with different vegetation structures, including woody, shrubby and herbaceous species
Depressionswhich typically include the basin and
pond features from the LU map
Floodplainsidentified from the LU map, and which are
areas of land adjacent to a stream or a river that experience
temporary flooding during periods of high river flow
Lakeswhich consist essentially of volcanic and natural (non-modified) lakes and identified from the LU map
Reservoirscategorised as man-made wetlands and
identified from the LU map

Therefore, a classification system inspired from the SADC


and Dugan (1990) classification systems is devised in this
work to create the 8 ecological wetland types above. The
Dugan (1990) classification system is devised for continental
and coastal areas and is based on the Ramsar classification
system. The SADC classification is meant for South African
countries (SADC) of which Mauritius forms part. These two
classification systems combined are capable of classifying all
of the wetlands of Mauritius that are visible on satellite imagery and published maps. This enabled creation of the above
8 ecological wetland types of Mauritius, such that although
these wetland types are based on classification systems devised elsewhere, they are considered useful for an oceanic
situation like that of Mauritius.
Apart from the SADC and Dugan (1990) classification
systems, no other classification system has been adopted in this
study, such as the hydrogeomorphic (HGM) approach (Brinson
et al. 1995), which is increasingly being used around the world.
However, the methodology and data requirement involved in
the HGM approach for wetland assessment was considered too
complex and thus, the ecological classification system using
the 8 wetland types above has been retained for this study.
Creation of a GIS Wetland Database
ArcGIS software, namely ArcInfo 9.3 was used for all
processing including analysing and manipulating data
layers. The datasets used in this study for compiling wetland
attribute data are
&
&
&

Geology map at 1:50, 000 scale by Giorgi et al. (1999),


Digital topographic contours map at 10 m interval from
MHL (2005) converted into a 25 m Digital Elevation
Model (DEM),
Moisture deficiency (humidity map) at 1:50,000 scale
from Halais and Davy (1969),

Author's personal copy


Wetlands

&
&
&

Soil map at 1:100,000 scale by Parish and Feillaf


(1965),
Land Use (LU) map of the island (13 sheets at 1:25,000)
from OS and GOM (1991), with digital map update
from MHL (2005), and
Isoyetal map showing the nine geographic regions into
which Mauritius is divided: North (N), East (E), SouthEast (SE), South (S), South-West (SW), West (W),
North-West (NW), Central Plateau-Leeward side (CPlee) and Central Plateau-Windward side (CP-win) available in hard copy format from WRU (2003).

For each digitized wetland, its attribute table was populated


with the following information: wetland name, eco-unit (Id),
mean elevation, underlying geology, soil type, climate type
(humidity regime), region (geographic location), and surface
area. The attribute eco-unit (Id) contains an integer (18), each
identifying a particular wetland type, namely, Marine coasts
(Id-1), Estuaries (Id-2), Floodplains (Id-3), Marsh/Swamp (Id4), Lakes (Id-5), Depressions (Id-6), Mangroves (Id-7) and
Reservoirs (Id-8). The attribute surface area was automatically calculated by the software by using the command
prompt calculate surface area.
Wetland Land Cover Assessment (WLCA)
A land cover assessment method was devised and applied to
a sample of 14 wetlands showing spatial variability. The
wetlands and their code names are: La Prairie (LPA), Mare
Sarcelle (MSA), Le Morne (MOR), Blue Bay Marine Park
(BBMP), Wolmar (WOL), Mare Camp Pav (MCP), Mare
Phedre (MPD), Grande Mare Longue (GML), Mare du Puits
(MPU), Mare Grand Maree (MGM), Mare Millet (MML),
Mare Pont Malette (MPM), Mare Dodo (MDO) and Rivulet
Terre Rouge Estuary Bird Sanctuary (RTREBS).
The wetland land cover assessment method adopted in
this study is described in detail in Appendix 1. In summary,
the method yields a Wetland Land Cover Index (WLCI) for
each of the above 14 wetlands, where the index value is
between 0 and 1. The WLCIs calculated are then grouped
into categories to represent the degree of impact. These are:
00.29 (no impact), 0.300.45 (low impact), 0.460.61
(moderate impact), 0.620.77 (high impact), and >0.77
(very high impact).

Results
A total of 144 wetlands were digitized (Fig. 2) while the
distribution of wetlands in regions (north, east, west, etc.) is
summarized in Table 1. The total surface area of wetlands is
~18.39 km2. By excluding reservoirs, natural wetlands occupy 6.39 km2, which is 0.34 % of the total surface area of

Mauritius, thus emphasizing the very small extent occupied


by wetlands in Mauritius. The predominant wetland class is
marshes and swamps, represented by 55 features and totalling a surface area of ~2.4 km2, or 38 % of the total (natural)
wetland surface area. The others are: Depressions (27),
Estuaries (22), Mangrove (16), Reservoirs (8), Floodplains
(7), Marine coasts (6) and lakes (3). From a geographic
point of view it can be observed that most wetlands are
situated in the northern, eastern and south-eastern coastal
plains with wetland distribution as follows: east (42 wetlands), north (31) and south-east (22).
Table 2 shows a sample of the GIS database created of the
144 wetlands digitized. Important information can be gathered
from the wetland inventory process such as the water regime
of a particular wetland. For example, marshes and swamps are
found predominantly along the eastern coast of the island
where Arlidge and Wong You Cheong (1975) describe this
region as suitable groundwater discharge areas where springs
are common features. It can thus be inferred that the main
hydrological inputs for wetlands located in these regions are
primarily subsurface flows coupled with precipitation occurring during the year. According to Winter (1988), wetlands
situated in such geographical settings and having such hydrological inputs can be assumed to have a mostly permanent to
seasonal water regime.
The land cover assessment methodology (WLCA) applied
on the 14 wetlands yielded a total of 14 wetland indices,
representative of wetland condition, from least impacted to
most impacted. The sample GIS attributes and locations of the
14 wetlands studied with WLCA are shown in Table 2 and
Fig. 2 respectively. In addition, the supplementary online
material provides 7 maps showing the 14 studied wetlands.
The wetlands Mare du Puits (MPU), Mare Grand Maree
(MGM), Mare Millet (MML) and Mare Pont Malette
(MPM), Mare Dodo (MDO) and RTREBS have high
WLCA with values ranging from 0.78 to 0.86. These 6 wetlands were thus ranked in the very high impact category. In
the other side, three wetlands (La Prairie, Mare Sarcelle and
Le Morne) have WLCA values 0.4 and thus were categorized
as low impact level wetlands (Table 3).
Results for percentage land cover types in the 50 m and
950 m buffer zones were grouped into 5 main categories
(Tables 4 and 5): Agriculture (Sugarcane, tea, food crops),
Forest, Natural vegetation (scrub, sparse vegetation), Urban
(Urban area and barren land) and Water bodies (wetlands and
water bodies). It can be observed that 3 wetlands have more
than 90 % of their 50 m riparian zone classified as agriculture
(Mare Millet, Mare Pont Malette and Mare Dodo, Table 4).
This is not a normal situation, as according to the law, a 15 m
restriction zone followed by a 35 m control zone must be
respected for a wetland. The same applies to the Ramsar site,
Rivulet Terre Rouge Estuary Bird Sanctuary (RTREBS),
where 84 % of the buffer zone consists of urban areas. A total

Author's personal copy


Wetlands

Fig. 2 Map showing the location of wetlands. The 14 wetlands studied in more details are labelled

of 6 out of the 14 wetlands have >50 % of their LC classified


as agriculture (Table 4).
Taking the example of wetlands with LCIs in the very high
impact category, we find that there is no particular correlation
between geographic location of these wetlands and the calculated LCIs. A common factor to the majority of these wetlands
however is the high percentage of sugarcane cultivation and/or
urban areas in both their 50 m and 950 m riparian zones
contributing to elevated LCIs and correspondingly degraded
condition. However, the Ramsar wetland, RTREBS differentiates itself from the rest by having approximately 84.92 % of
its riparian zone and 67.60 % of its upland buffer zone

classified as urban area, thus making urbanization the major


cause of its highly degraded condition with an LCI of 0.87.
A water quality monitoring program undertaken by
Mamoun (2006) at RTREBS concluded that the estuary
was being heavily impacted by both organic and nonorganic pollutants. The major sources of organic pollution
were identified as potentially originating from a nearby pig
farm and domestic sewage. Elevated concentrations of chloride ions (982 mg L1) and COD reaching 2,454 mg L1
were obtained at various sampling points in the estuary and
were observed within the 15 m restriction zone. It is therefore possible to infer that the LCI obtained from the present

Author's personal copy


Wetlands
Table 1 Distribution and
classification of the 144
wetlands inventoried in this
work. Distribution is a
geographic basis, with, north
(N), northwest (NW), east (E),
etc. Classification is into 8
ecological wetland types found
exhaustive to classify all of the
wetlands of the island

Ecological units

Marine coasts
Estuaries
Floodplains
Marsh/Swamp
Lakes
Depressions
Mangroves
Reservoirs
Total

Area km2

1.18
0.95
0.62
2.4
0.25
0.5
0.49
12
18.39

Region
N

NW

SE

SW

CP(L)

CP(W)

2
2

19
1
2
4
1
31

2
3
21

7
9

42

3
7

1
3

22

3
1
1
2
7

14

1
1

5
3
2

1
13

4
8

2
5

methodology corroborates to a large extent the actual situation faced by the estuary.
The mean slope and standard deviation for 50 m and
950 m buffer zones are shown in Table 6. For the 950 m
buffer zone, 2 wetlands have mean slope >8 %, which is
the critical threshold for areas with high susceptibility to
erosion (Nigel and Rughooputh 2010b, 2012). These two
wetlands are La Prairie and Le Morne wetlands with mean
slope 29.05 % and 24.78 %, respectively. Both La Prairie
(LCI=0.3) and Le Morne (LCI=0.4) were assessed as low
impact category wetlands based on the land cover assessment only. The 950 m buffer of La Prairie consists of
3.85 % agriculture, while that of Le Morne has 17.72 %
agriculture. Due to the steepness of the surrounding agricultural lands, there may be an enhanced surface run-off

during rainfall events, thus contributing to silt and nutrient


accumulation in these wetlands.

Discussion
In this paper, a wetland distribution map was produced,
showing the location and types of wetlands across the island
followed by the creation of a GIS wetland database
containing information about the physical attributes of 144
mapped wetlands. The second part of this work had for
objective to assess the condition of a sample of wetlands
across the island by developing a wetland land cover assessment method. This method however had to satisfy local
condition and resources available. These primary conditions

Table 2 Wetland attribute table of the 14 wetlands studied in more detail. The attributes given are wetland type (eco-unit), geographic region
located (southeast SE, north N, etc.), and elevation above mean sea level, amongst others
Wetland name

Eco-unit

Region

Elevation (m).

Area (km2)

Geologya

Soilb

Climatec

Mare Dodo
Mare Phedre
Grande Mare Longue
Mare Pont Malette
Mare Grand Maree
Le Morne
Mare Millet
Wolmar
La Prairie
Mare du Puits
Mare Camp Pave
Mare Sarcelle
RTREBS

Marsh/Swamp
Marsh/Swamp
Marsh/Swamp
Marsh/Swamp
Marsh/Swamp
Marine Coast
Marsh/Swamp
Marsh/Swamp
Marsh/Swamp
Marsh/Swamp
Marsh/Swamp
Marsh/Swamp
Estuarine

SE
N
N
N
N
SW
SE
W
SW
E
N
E
NW

8.02
3.15
4.01
18.43
15.08
1.15
4.65
3.28
3.95
6.32
4.32
3.47
0.00

0.030
0.166
0.135
0.044
0.172
0.063
0.027
0.370
0.316
0.120
0.060
0.231
0.266

DL1
DL2
DL2
DI2
DI2
BTP
DL1
S1
S1
DL2
DL2
DL3
N/A

P
T
T
L
L
C
P
C
C
T
T
T
N/A

H
SB
SB
SB
SB
H
H
SB
H
H
SB
H
SB

BBMP

Marine Coast

SE

0.36

0.847

N/A

N/A

SB

BTP old lavas, DL1 late lavas (1st flow), DL2 late lavas (2nd flow), DL3 late lavas (3rd flow), S1 sandy coast

C regosols, L low humic latosols, T lithosols, P latosolic red prairie soils

H humid, SB sub-humid

Author's personal copy


Wetlands
Table 3 Riparian buffer scores,
upland buffer scores and
calculated Wetland Land Cover
Indices (WLCI) of the 14
wetlands studied in detail. A
higher WLCI is indicative of a
wetland that is impacted more
according to its land cover type.
For the riparian buffer score, the
land cover types assessed are
within a 50-m buffer while the
upland buffer score is for an
additional 950-m buffer zone
around the wetland, but
delimited by the wetlands
watershed

Wetland code

Riparian buffer score

Upland buffer score

WLCI

Impact Category

LPA
MSA
MOR
BBMP
WOL
MCP
MPD
GML
MPU
MGM
MML
MPM
MDO
RTREBS

1.24
1.81
2.13
2.23
2.76
3.24
3.42
3.66
4.70
4.22
4.75
4.59
4.80
4.45

1.71
1.54
1.85
3.28
2.86
3.27
3.37
3.24
3.11
3.70
3.43
3.70
3.81
4.20

0.30
0.34
0.40
0.55
0.56
0.65
0.68
0.69
0.78
0.79
0.82
0.83
0.86
0.87

Low
Low
Low
Moderate
Moderate
High
High
High
Very high
Very high
Very high
Very high
Very high
Very high

were fulfilled by using as main input the land cover map and
a buffer approach including the 15 m restriction zone and
35 m control zone suggested by legislation.
A large amount of information was used in the attempt to
classify wetlands of Mauritius. The land use map of 2005 of
Mauritius produced by the Ministry of Housing and Lands
provided the basis for identifying potential wetlands. The
resulting classification confirms the presence of wetland types
such as floodplains, marshes, swamps, mangroves and estuaries typically present in countries like Mauritius which support a humid to sub-humid climate. Whilst the ecological
benefits of wetland types such as mangroves, swamps, estuaries are relatively well acknowledged, those of other natural

wetland types mentioned in this study such as depressions and


natural lakes remain unknown. This study also suggest that
geographic location influences wetland type in Mauritius,
where approximately 50 % of wetlands digitized in the north
and eastern part of the Island are classified as marshes and
swamps while the rest are distributed unevenly and to a
smaller extent in the remaining classes (see Table 1).
During the GIS mapping of the wetlands, field trips were
carried out on five occasions, where some thirty of the
wetlands mapped were visited and photographs taken.
Additionally, for some of these wetlands, a few parameters
were measured on-site, including pH, temperature and salinity (Mamoun 2011). Thus, although the exact number of

Table 4 Percentage distributions of land cover categories in the riparian


buffer zone (50 m). The wetlands are arranged in increasing order of
Wetland Land Cover Indices (WLCI) ranging from La Prairie (LPA) with

WLCI=0.3 (low impact) to Rivulet Terre Rouge Estuary Bird Sanctuary


(RTREBS) with WLCI=0.87 (very high impact). Agricultural and urban
land cover types in the buffer zone pose greater impact on the wetland

Wetland code

Agriculture (%)

Natural vegetation (%)

Forest (%)

Water bodies (%)

Urban (%)

LPA
MSA
MOR
BBMP
WOL
MCP
MPD
GML

3.11
0,00
0,00
5.56
4.56
42.13
42.12
52.16

27.52
84.95
36.52
32.10
34.39
39.54
42.33
26.52

54.23
8.74
44.10
44.55
20.13
10.58
3.15
13.56

15.14
6.31
0,00
1.76
9.24
5.03
7.22
3.13

0.00
0,00
19.38
16.03
32.68
2.72
5.18
4.63

MPU
MGM
MML
MPM
MDO
RTREBS

89.57
82.24
93.64
91.75
95.78
0,00

6.58
0,00
0,00
0,00
0,00
4.06

0,00
9.11
4.29
0,00
0,00
19.14

3.24
8.65
2.07
8.25
4.22
0,00

0.61
0,00
0,00
0,00
0,00
84.92

Author's personal copy


Wetlands
Table 5 Percentage distributions of land cover categories in upland
buffer zone (950 m). The wetlands are arranged in increasing order of
Wetland Land Cover Indices (WLCI) ranging from La Prairie (LPA) with

WLCI=0.3 (low impact) to Rivulet Terre Rouge Estuary Bird Sanctuary


(RTREBS) with WLCI=0.87 (very high impact). Agricultural and urban
land cover types in the buffer zone pose greater impact on the wetland

Wetland code

Agriculture (%)

Natural vegetation (%)

Forest (%)

Water bodies (%)

Urban (%)

LPA
MSA
MOR
BBMP
WOL
MCP
MPD
GML

3.85
0.00
17.72
67.44
15.72
51.26
59.11
46.77

45.75
37.84
20.64
5.15
34.87
24.40
21.83
27.73

44.66
54.29
57.69
20.73
18.92
4.23
5.23
7.50

1.64
2.96
0.41
1.31
2.97
6.02
3.23
3.06

4.10
4.91
3.55
5.37
27.52
14.10
10.59
14.95

MPU
MGM
MML
MPM
MDO
RTREBS

58.03
83.29
63.33
89.27
92.36
11.27

22.54
5.39
6.52
0.60
3.60
17.59

10.35
6.65
11.26
4.30
1.49
3.20

4.34
0.82
5.77
4.42
1.89
0.035

4.74
3.85
13.12
1.41
0.67
67.60

wetlands mapped was not ascertained through an exhaustive


ground truthing, the short field trips undertaken show that at
least some of the wetlands mapped was a good reflection of
the real situation on the ground.
Moreover, the high LCI of RTREBS directly related to a
high urbanization cover suggested that the estuary was being
potentially threatened by anthropogenic activities, which was
corroborated with the work of Mamoun (2006). Therefore

although field surveys were not undertaken for each sample


wetland, the assumption that the WLCA method reflects
reality on the ground to a non-negligible extent could be
made, hence validating the methodology devised.
Although the digitization process and compilation of wetland data in the database was relatively time-consuming, once
everything was set up, data can be easily added, removed or
updated at any time. For instance, the biological data on

Table 6 Slope mean and standard deviation for 50 m and 950 m buffer
zones. The wetlands are arranged in increasing order of Wetland Land
Cover Indices (WLCI) ranging from La Prairie (LPA) with WLCI=0.3
(low impact) to Rivulet Terre Rouge Estuary Bird Sanctuary (RTREBS)

with WLCI=0.87 (very high impact). A higher slope in the buffer zone is
indicative of higher runoff potential, thus higher risk of contaminant
loading, especially in the presence of agricultural practices or intensive
urbanisation

Wetland code

Slope statistics for 50 m riparian buffer zone

Slope statistics for 950 m upland buffer zone

Mean (%)

Standard Deviation (%)

Mean (%)

Standard Deviation (%)

LPA
MSA
MOR
BBMP
WOL
MCP
MPD
GML

3.35
0.81
2.45
8.90
1.42
0.93
0.95
0.77

2.92
0.16
2.19
7.34
0.99
0.19
0.12
0.13

29.05
1.25
24.78
2.92
3.31
1.31
1.14
1.28

21.04
1.45
23.06
3.03
3.32
1.01
0.90
1.02

MPU
MGM
MML
MPM
MDO
RTREBS

2.27
2.70
1.29
3.30
1.42
1.43

1.86
2.09
0.13
1.86
0.54
2.06

1.31
2.25
1.60
2.17
1.64
1.91

1.15
1.46
1.30
1.48
1.26
2.00

Author's personal copy


Wetlands

wetlands of Mauritius recently collected by Laurance et al.


(2012) can be added. Such a database can be of great help to
wetland practitioners such as the National Parks Conservation
Service (NPCS) and the scientific community alike, since for
the first time, information on wetland location, size and type
of Mauritius will become available and contained within one
single database. Additional data on the fauna and flora and
water quality of wetlands can also be added at any time to
satisfy the needs of all those concerned with wetland protection. Conferences can be organised, initiated by scientists and
environmental groups to disseminate information to the public
at large. The ministry of education may also find of interest to
encourage school teachers across the country to contribute to
wetland protection by raising childrens interest in the subject.
This could be done, for instance, as part of the Environmental
Science (EVS) subjects at the primary and secondary levels.
The wetland map and database produced in this paper can be
used to this end, which would contribute to raising
populations awareness which is one of the constraints in
effective wetland protection in Mauritius.
GIS is the primary tool for the manipulation and analysis
of spatial data over large geographical scales (U.S. EPA
2002). A GIS was used in the present work where it enabled
the manipulation of large amount of data covering the whole
island within a single database.
The Wetland Land Cover Assessment Methodology
yielded a series of land cover indices, representative of wetland condition. From the analysis of riparian scores, the riparian zones of La Prairie and Mare Sarcelle wetlands are seen to
be relatively well preserved, consisting mostly of forest and
natural vegetation. On the other hand, wetlands with the
highest LCIs, namely MDO, MPU, MML, MGM and MPM
are situated on sugarcane cultivated land belonging almost
exclusively to private land owners. The location of these
wetlands suggests that they are most certainly acting as natural
sinks for nutrient run-off during the irrigation period and rainy
season. More research must however be carried out to determine the extent of the nutrient retention function of these
wetlands and the effect of the accumulation of nutrients on
their ecological condition. The Ramsar wetland RTREBS,
situated in the vicinity of Port-Louis harbour is the most
highly impacted wetland (LCI=0.87) and is being threatened
mostly by anthropogenic activities linked to urbanization.
The riparian scores for the wetlands are a clear indication
that legislation is not being respected in some cases. The
buffer approach not only helps target priority action areas
but also helps pinpoint the type of environmental stress a
particular wetland is facing such as agriculture and/or urbanization through the land cover analysis. These wetlands having their riparian zones consisting of agriculture are mostly
located on private land belonging to sugarcane planters.
Private landownership is amongst one of the major obstacles regarding the conservation of wetlands in Mauritius. This

is so, because still nowadays, the wrongful mind set is that


wetland protection and conservation does not bring any financial gain and so is not worth the attention. In parallel, countries
located not too far from Mauritius such as Australia, the
United Kingdom and New Zealand have long since seen the
economic gain of using artificial wetlands to treat contaminated industrial water (Mitsch and Gosselink 2007). Local wetland practitioners such as the NPCS can thus use this
information to specifically target sugarcane planters for
information dissemination and encourage them towards
wetlands wise-use through meetings. The ministry of
environment can also use this information as proof of
the urgency in protecting wetlands and hopefully influence legislation regarding wetlands in a favourable way.
A considerable advantage of the WLCA method is that it
is highly reproducible and flexible since the scores are user
dependent and can be modified to adapt to any local situation. Therefore, other Island States or developing countries
similar to Mauritius may find useful to apply the WLCA
methodology presented in this paper to assess their own
wetlands. In an international context, non-governmental
organizations and governments may find data on wetlands
of an Island State like Mauritius which is so limited in
expertise and resource, of considerable value in international ecological debates and studies.
The principal stakeholderand being also the authority, funding agency and controlling body of other stakeholderson the matter of wetlands in Mauritius is the
Ministry of Environment. The ultimate aim of studies
like the present one is to get stakeholders involved in wetland
protection/conservation or wise use. The only way to achieve
this is to involve stakeholders (who might use the information)
in the process of conceptualising, collecting, analysing and
interpreting the data embedded in the GIS. In fact, participation is the key to change practices damaging wetlands. It must
be conceded that in the present study there is a lack of
participatory approach, and this can be problematic, as
command-and-control (top-down) approaches are unlikely to
be effective given the development-driven needs in Mauritius.
Further studies to the present one are envisaged for the mapping of the wetlands of Mauritius, notably, a more detailed
classification, an improved wetland assessment methodology
and the use of higher resolution and more up-to-date remote
sensing datasets. The participation of the Ministry of
Environment and other stakeholders in such future works
may be a means to help change practices damaging wetlands
in Mauritius.

Conclusions
The information gathered in this paper can be considered a
huge step towards increasing knowledge about local

Author's personal copy


Wetlands

wetlands and bridging the information gap. Both the wetland inventory and the wetland assessment have been shown
to be important in decision making regarding wetland management. Information will be used to increase population
awareness on the importance of wetlands in Mauritius and
the need to conserve and protect the remaining wetlands.
Results of the wetland assessment methodology can be used
as a preliminary step in the identification of priority action
sites for wetland protection and conservation but also to
improve legislation eventually. It is thus possible to say that
with such concrete information as presented in this paper
and its wise use, wetland protection and conservation in
Mauritius seems less utopic and more tangible.
Acknowledgments The authors acknowledge the Cartographic Section of the Ministry of Housing and Lands and the Mauritius Meteorological Services for the provision of data. This work was given support by
a Mauritius Tertiary Education Commission (TEC) scholarship for the
first author at the University of Mauritius. We thank the two anonymous
reviewers for their comments and suggestions made on earlier versions of
the manuscript.

References
Arlidge EZ, Wong You Cheong Y (1975) Notes on the land resources and
suitability map of Mauritius at 1:50 000 scale. Mauritius Sugar
Industry Research Institute (MSIRI) Occasional Paper 29:1137
Breen CM, Quinn NW, Mander JJ (1997) Wetlands conservation and
management in southern Africa: challenges and opportunities.
IUCN, South Africa
Brinson MM, Hauer FR, Lee LC, Nutter WL, Rheinhardt RD, Smith
RD, Whigham D (1995) A guidebook for application of
hydrogeomorphic assessments to riverine wetlands. Technical
Report WRP-DE-11. U.S. Army Engineer Waterways
Experiment Station, Vicksburg, NTIS No. AD A308 365
Brooks RP, Wardrop DH, Bishop JA (2004) Assessing wetland condition
on a watershed basis in the MidAtlantic region using synoptic
landcover maps. Environmental Monitoring and Assessment
94:922
Dugan PJ (1990) Wetland Conservation: a review of current issues and
actions. IUCN Switzerland
US Environment Protection Agency (US EPA) (2002) Methods for
evaluating wetland condition: landuse characterization for nutrient
and sediment risk assessment http://epa.gov/waterscience/criteria/
nutrient/wetlands/modules/17landuse.pdf. Accessed 22 June 2011
Giorgi LS, Borchiellini S, Delucchi L (1999) Geologic map
hydrogeological scheme, 1:50,000 scale. Water Resources Unit,
Rose-Hill
GOM-OS (1991) Map of Mauritius 1:25,000. Ed 6-OS 1991, series
Y881 (DOS 329). Ordnance Survey (OS) for the Government of
Mauritius. GOM, London

Halais P, Davy EG (1969) Notes on the 1:100,000 agro-climatic map of


Mauritius. MSIRI Occasional Paper 23:129
Hutchinson MF (1989) A new procedure for gridding elevation and
stream line data with automatic removal of spurious pits. Journal
of Hydrology 106:211232
Laurance SGW, Baider C, Vincent Florens FB, Ramrekha S, Sevathian
JC, Hammond DS (2012) Drivers of wetland disturbance and
biodiversity impacts on a tropical oceanic island. Biological
Conservation 149:136142
Leibowitz SG, Abbruzzese B, Adamus PR, Hughes LE, Irish JT (1992)
A synoptic approach to cumulative impact assessment: a proposed
methodology. EPA/600/R-92/167, U.S. Environmental Protection
Agency, Environmental Research Laboratory, Corvallis, OR
Mamoun CM (2006). Vulnerability assessment of RTREBS to surrounding sources of pollution. Unpublished BEng Dissertation.
University of Mauritius, Rduit, Mauritius
Mamoun CM (2011) Assessment of wetland condition in Mauritius
using a landscape characterisation method. Unpublished MPhil
Thesis. University of Mauritius, Rduit, Mauritius
MHL (2005) 10 m digital contours of Mauritius and 1:25,000 digital map
of Mauritius. Ministry of Housing and Lands (MHL), Port-Louis
Mitsch WJ, Gosselink JG (2007) Wetlands, 4th edn. John Wiley &
Sons, USA
MOE (2005) Mauritius: staking out the future. Ministry of Environment
& National Development Unit, Port-Louis
MOE and GIBB (2002) Study of environmental risks in grand Baie
area, vol:1. GIBB (Mauritius) Ltd in assoc. with ARCUSGIBB
for the Ministry of Environment, Port-Louis
Nigel R, Rughooputh SDDV (2010a) Mapping of monthly soil erosion
risk of mainland Mauritius and its aggregation with delineated
basins. Geomorphology 114:101114
Nigel R, Rughooputh SDDV (2010b) Soil erosion risk mapping with
new datasets: an improved identification and prioritisation of high
erosion risk areas. Catena 82:191205
Nigel R, Rughooputh SDDV (2012) Application of a RUSLE-based
soil erosion modelling on Mauritius Island. Soil Research
50:645651
NWFS, NC, MOE (2008) Survey report of Grand Baie Wetlands,
Grand Baie, Mauritius. Intermediate report. NWFS consultancy
and the Watershed Company (WC) for the Ministry of
Environment (MOE), Mauritius
Parish DH, Feillaf SM (1965) Notes on the 1:100 000 soil map of
Mauritius. Mauritius Sugar Industry Research Institute (MSIRI)
Occasional Paper 22:145
Proag V (1995) The geology and water resources of Mauritius. The
Mahatma Ghandi Institute Press, Mauritius
Saddul P (1995) Mauritius-a geomorphological analysis. The Mahatma
Ghandi Institute Press, Mauritius
Saddul P (2002) Atlas of Mauritius, 2nd edn. Edition de lOcean Indien
and George Philips Ltd, Mauritius
Skagen SK, Melcher CP, Haukos DA (2008) Reducing sedimentation
of depressional wetlands in agricultural landscapes. Wetlands
28:594604
Winter TC (1988) A conceptual framework for assessing cumulative
impacts on the hydrology of nontidal wetlands. Environmental
Management 12(5):605620
WRU (2003) Hydrology data book for the period 19951999. Water
Resources Unit (WRU), Rose-Hill

ONLINE SUPPLEMENTAL FILE


APPENDIX 1 PROCEDURE FOR WETLAND LAND COVER ASSESSMENT
Riparian buffers play an important role in improving the overall water quality of wetlands by intercepting
and filtering surface run-off containing nutrients and sediments (Skagen et al 2008). Two scores are
calculated, first a riparian score for a 50 m buffer width, starting from the wetland itself and representative
of the ability of its 50 m buffer strip to filtrate contaminants from the upland zone. This 50 m value is
chosen as the The Planning and Development Act 2004 says that no construction must be done at 15 m
buffer zone restriction with a further 35 m control zone. So, it is assumed that there will be mostly some
form of natural vegetation existing within the 50 m buffer strip. The second score is calculated for a buffer
width of 950 m, starting from the outer edge of the 50 m buffer (Note that the distance of each buffer is
bounded by the watershed of the wetland). This means that if a wetland has a watershed (i.e., an upslope
drainage area) of only 100 m2 and maximum length 15 m, then the buffer cannot go past this maximum 15
m length and in area, cannot exceed 100 m2. The watershed, in this case, is the limit to the surface water
flow to the wetland. Any land cover outside a wetlands watershed is thus assumed to not affect or
contribute to surface water flow to the wetland. The drainage basins, flow direction and flow accumulations
were obtained from Nigel and Rughooputh (2010b).
Assessing land cover types and their threats to wetlands
A land cover map of Nigel and Rughooputh (2010a) lists 11 land cover types obtained by partly digitizing
the LU maps and its updated versions, coupled with supervised image classification on a SPOT 10 m pan
sharpened multi-resolution image of 1994. The land cover map is in raster format with 10 m cell size. The
11 land cover (LC) types were assigned risk factors, indicative of the potential for a particular LC to either
improve or impact negatively on the overall condition of a wetland and ranges from 0 (no impact)
representing the lowest risk, to 5 (very high impact) or equivalent to highest risk. The 11 land cover types
and their associated risk factors for the 50 m buffer width and 950 m buffer width are: sugarcane (5 for 50
m, 4 for 950 m); sparse vegetation (2, 2); scrub (3, 2); barren land (4, 3); food crops (5, 4); tea (4, 4); forest
(1, 1); urban area (5,5). The remaining land cover types which are wetlands, water bodies and sand are
given a risk factor of 0 (no impact) as they are assumed to be of no threat to the wetlands themselves.
Generating Wetland Buffer Zones
The ArcGIS software, ArcInfo(9.3) was used to create wetland buffer zones using as main inputs 1) a
GIS data layer containing the 14 wetland polygons and 2) the land cover map of Nigel and Rughooputh
(2010a) at 10 m ground pixel size. In GIS, buffer widths can be automatically generated using the buffer
analysis tool. After selecting the wetland polygon of interest from the wetland layer map, two buffer
widths, one of 50 m, labelled riparian buffer zone and a second buffer width of 950 m (starting from the 50
m buffer zone) labelled upland buffer zone, were created for each polygon. A total buffer width of 1 km
was thus generated around each wetland. Adjustments to the 950 m buffer zone were needed in order to
account for watershed boundary and upslope contributing areas only. As an example, La Prairie wetlands
buffer zones were trim to keep only upslope contributing areas that are inside the wetlands watershed
boundary (Fig. 2). The river basins and watersheds of the island, which give the equivalent watershed
perimeters for the studied wetlands, were obtained from Nigel and Rughooputh (2010a).

Fig. 2 Buffer zones for La Prairie wetland


Analysis of slope
We also aim to determine the slope gradient of the buffer zones to assess their susceptibility to erosion,
which affects wetland condition through surface run-off. Nigel and Rughooputh (2010a) found that nearly
31% of the Islands surface area has slopes > 8% which by soil erosion standards, are areas highly
susceptible to erosion. The slope map used in this paper for slope analysis is at 25 m cell resolution, was
produced in ArcGIS software from a 25X25 m Digital Elevation Model (DEM). The 25 m DEM in turn
was produced in ArcGIS software with the topogrid tool, which uses the ANUDEM programme of
Hutchinson (1989). Input elevation data for the DEM creation were 10 m interval contours from MHL
(2005). The slope layers (in percent) for the 50 m and 950 m buffer zones were extracted for each 14
wetland using the Extraction by mask spatial analyst tool in ArcInfo(9.3), where in this case, the mask
corresponds to the buffer zone being used. Mean slope and standard deviation for each slope layer was then
obtained by enquiring after the slope layer statistics. Slope statistics are automatically calculated by the
software and include mean, standard deviation, maximum and minimum values of the extracted slope.

Calculating percentage land cover


The method used was to first extract the land cover pattern corresponding to a particular buffer zone using
the Extraction by Mask spatial analyst tool in ArcInfo 9.3. This step generates a land cover layer which
was saved as a layer file. The GIS software automatically generates an attribute table for each land cover
layer created which consists of the number of counts of cells (pixels) for each land cover feature, where
one count is equivalent to one pixel size of 10X10 m (or 100 m2). The attribute table was then exported to a
Microsoft Excel Sheet for calculation of percentage land cover within each buffer. To calculate the
percentage of a land cover class within a particular buffer zone, for example sugarcane, the total number of
counts for the buffer zone layer was first computed. The number of counts for sugarcane was then divided
by the total number of counts to obtain percentage sugarcane. This procedure was carried twice for each 14
wetlands, one for the upland buffer zone and the other for the riparian buffer zone.

Final computation of Wetland Land Cover Assessment (WLCA)


After obtaining all these information on land cover types, the following steps are carried out to calculate the
WLCI:
1.
Assign risk factors (05) to each 11 land cover types
2.
Calculate the percentage of each 11 land cover types (sugarcane, scrub, forest, urban area etc.) in
riparian buffer zone for wetland of interest
3.
A value is then obtained by multiplying percentage land cover with the corresponding risk factor.
This step is repeated for each 11 land cover types
4.
The 11 values obtained for the wetland are then added to obtain a wetland riparian score
(RIPSCORE)
5.
Steps 1 through 5 above are repeated for the upland buffer zone to obtain an upland buffer score
(UPLSCORE)
Addition of the UPLSCORE and RIPSCORE then yields a wetland score called WETSCORE which is then
divided by 10 to obtain an overall Wetland Land Cover Index (WLCI) value between 0 and 1 for a
particular wetland.

WLCI

UPLSCORE RIPSCORE
10

Equation 1

The LCIs calculated are then grouped into categories to represent the degree of impact. These are: 00.29
(no impact), 0.300.45 (low impact), 0.460.61 (moderate impact), 0.620.77 (high impact), and >0.77
(very high impact).
The WLCA methodology developed in this study however may need some adjustment for specific
wetland types such as floodplains and mangroves. This is because floodplains receive their water supply
from catchments linked to the stream network such that local land use around the floodplain is relatively
unimportant. In contrast, mangroves receive their water supply from streams and from the sea. It is thus
proposed that in a future study, the land-cover assessment methodology is extended to the watershed scale
for floodplains and not limited to the 1-km buffer. In the case of wetlands situated along the coastal plains
and which are directly influenced by the sea, such as mangroves, further studies including modelling of
current flow and contaminant movement in the lagoon is suggested to get a clearer picture of the overall
impact brought about by land and sea pollution.

SUPPLEMENTARY MATERIAL FOR ONLINE PUBLICATION


Below are seven (7) maps showing in more detail the 14 wetlands studied in this paper.

Map 1: Wolmar wetland (WOL) in the west of the island. Its eco-unit is Marsh/Swamp.
WOL is in the moderately impacted wetland category with a Wetland Land Cover
Indices (WLCI) of 0.56 (ranked 5th, see Tables 2 and 3)

Map 2: Le Morne (MOR) and La Prairie Wetlands (LPA) in the southwest of the island.
Their eco-units are Marine Coast (MOR) and Marsh/Swamp (LPA). Both MOR and LPA
are in the lowly impacted category with Wetland Land Cover Indices (WLCI) of 0.4
and 0.3, respectively (ranked 3rd and 1st, respectively, see Table 3)

Map 3: Rivulet Terre Rouge Estuary Bird Sanctuary (RTREBS) in the northwest of the
island. Its eco-unit is Estuarine. RTREBS is a Ramsar site and is in the very highly
impacted category with a Wetland Land Cover Indices (WLCI) of 0.87 (ranked 14th, see
Table 3)

Map 4: Grande Mare Longue (GML), Mare Phedre (MPD), Mare Camp Pave (MCP) in
the north of the island. All three wetlands have eco-unit of Marsh/Swamp and all three
are in the highly impacted category with Wetland Land Cover Indices (WLCI) of 0.69,
0.68 and 0.65, respectively (ranked 8th, 7th and 6th, respectively, see Table 3)

Map 5: Mare Grand Maree (MGM), Mare Pont Malette (MPM) and Mare Sarcelle
(MSA) in the north and east of the island. MGM and MPM have eco-unit of
Marsh/Swamp and are in the very highly impacted category with Wetland Land Cover
Indices (WLCI) of 0.79 and 0.83, respectively (ranked 10th and 12th, respectively). Mare
Sarcelle (MSA) has eco-unit Marsh/Swamp and is in the lowly impacted category with
a Wetland Land Cover Indices (WLCI) of 0.34 (ranked 2nd, see Table 3)

Map 6: Mare du Puits (MPU) in the east of the island. MPU has eco-unit Marsh/Swamp
and is in the very highly impacted category with a Wetland Land Cover Indices (WLCI)
of 0.78 (ranked 9th, see Table 3)

Map 7: Mare Millet (MML), Mare Dodo (MDO) and Blue Bay Marine Park (BBMP) in
the southeast of the island. MML and MDO have eco-unit Marsh/Swamp and are in the
very highly impacted category with Wetland Land Cover Indices (WLCI) of 0.82 and
0.86, respectively (ranked 11th and 13th). The Blue Bay Marine Park (BBMP) is a Ramsar
site and is in the moderately impacted category with a Wetland Land Cover Indices
(WLCI) of 0.55 (ranked 4th, see Table 3). The third Ramsar site, Point dEsny Wetland,
an eco-unit type Marine Coast, is also shown

You might also like