GIS Project Proposal




"Using GIS to Facilitate Management of the Nariva Swamp,

Trinidad and Tobago."






Gerard Alleng

Center for Energy and Environmental Policy,

University of Delaware.



May, 1998.



1.0 Background

The Nariva wetland is the largest wetland system in Trinidad and Tobago covering an area of approximately 6000 ha. It is located on the east coast of Trinidad between 10º 30’N and 10º 22’N latitude and 61º 01’W and 61º 06’W l ongitude (Figure 1). It lies within a large basin or depression, the average elevation of which is at or below sea level.

Hydrologically, the system is controlled by a number of rivers and streams particularly along its western boundary. These inlets drain a very large catchment basin of the eastern slope of the Central Mountain Range of Trinidad. The flow of water in the wetland is directed out of the system by the large Nariva River. There are very few areas within the wetland where open water predominates and these are mainly restricted to the drainage canals which traverse the wetland.

The diversity of plants and animals within the system is extremely high, as the system is environmentally diverse and ecologically complex. Bacon et al (1979) records 319 species of plants and 614 species of animals ut ilizing the area. The wetland is principally a freshwater system but has a large saltwater component. The vegetation pattern is divided into four major wetland types, (a) mangrove swamp, (b) freshwater swamp wood, (c) palm swamp and (d) freshwater marsh (Marshall, 1934; Bacon et al, 1979). There are also small po ckets of tropical rainforest inhabiting upland areas or inliers within the wetland.

In addition to the high biodiversity of plants and animals in the wetland, some are also vulnerable or endangered species. The system is known to support a small population of West Indian manatees (Trichechus manatus). The an aconda (Eunectes murinus) and other reptiles (e.g. Caiman sclerops) also use the area, including the iguana (Iguana iguana), which has an established breeding site in the wetland. A number of vulnerable waterfowl such as the Black-bellied whistling duck (Dendrocygna autumnalis) also utilize the system. As a result of the uniqueness of the wetland, a section of it was declared a wildlife sanctuary in 1968 by the Government of Trinidad and Tobago (Tikasingh, 1972).

Direct economic use or consumption of the components of the wetland include the removal of the palmiste palm (Roystonea oleracea), the freshwater conch (Pomacea urceus), mangrove oysters (Crassostrea rhizophorae), crabs (Cardisoma guanhumi and Ucides cordatus) and fish particularly the cascadura (Hoplosternum littorale). It has been estimated that this fishery in 1979 was worth TT $ 1,000,000 dollars (Bacon et al 1979). In spite of the enormous ecological and economic importance of the wetland, some sections have been severely damaged by human impact through conversion for agricultural purposes, mainly that of rice farming.

Between 1986 and 1992, over 650 ha of predominately freshwater marsh in the western section of the wetland had been converted to rice paddies (Alleng, 1994). This conversion altered the hydrological regime of the wetland through channel ization and damming, but it is uncertain how this has affected the integrity of the wetland.

In response to the human impact on the wetland, the area was declared a prohibited area in 1993 by the Government of Trinidad and Tobago, through The Forests (Prohibited Areas) Order, Citation 1993 (Figure 2). This was in addition to the area being designated in 1992, as Trinidad and Tobago’s first site fo r the "List of Wetlands of International Importance" under the Ramsar Convention. The boundaries used for the demarcation of the prohibited area were taken from the boundaries of a proposed national park suggested by Thelen and Faizool, (1980). However, i t is uncertain how these were chosen especially since they exclude a significant portion of the freshwater component of the wetland. Furthermore, over the past 16 years, there have been significant changes in the wetland since the plan was devised. The fa ilure to recognize and use appropriate ecological boundaries, could lead to the degradation of protected areas (Salm and Clark, 1984). Therefore, there is the need to revise and modify the proposed boundaries of the area. It has been strongly recommended that this revision be undertaken, together with the preparation of a detailed management plan for the area, using various categories of protection/management (RAMSAR, 1996).


2.0 Literature Review

The spatial nature of wetland areas is an ideal testing ground for the applications of GIS technology. Squires (1993) provides examples of wetland indicators and their associated measurements, which all require some form of spatial analysis (Table 1). These indicators are essential components in the evaluation of the conditions of a wetland system.


Table 1. Examples of wetland indicators and associated measurements

(adapted from Squires, 1993).

 Indicator Group

Example Measurements

Wetland extent

Wetland acreage


Fragmentation; ratio of vegetation: open water; corridors; miles of drainage ditches/canals

Water regime

Tidal amplitude; depth of flooding; duration of flooding; frequency of flooding

Sediment deposition

Annual rate of sediment deposition

Land management practices

Acres cultivated




An integral part of the determination of the conditions of a wetland is the undertaking of an inventory of its abiotic and biotic features. This inventory would comprise the gathering of data on the location, size and quality of the se resources and give insight into their temporal dynamics. Using GIS, Aschbacher et al (1994) were able to take an inventory of wetland habitats in the Phangnga Bay of southern Thailand. The GIS was used to support the application of remote sensin g techniques, in order to discriminate mangrove forests from non-mangrove classes or areas. The incorporation of the GIS into the analysis, improve the overall classification performance of the remote sensing data and was able to help with the monitoring of changes in the mangrove forests. Zalidis et al, (1997) were also able to successfully test a wetland habitat inventory and classification system for wetland areas in northern Greece, using a GIS database. Nevertheless, there was a problem with t he application of the GIS, because of a lack of adequate information on the wetlands, such as a current list of vegetation species, soils and hydrologic data.

The functions or roles of wetlands can also be investigated using GIS technology. One role in particular is the ability of wetlands to act as nutrient, sediment or chemical sinks. Greiner and Hershner (1998) were able to use a GIS t o assist in the analysis of total phosphorus retention by wetland systems near West Point, Virginia, USA. They investigated the assumption that landscape position of a wetland and land-use patterns surrounding it, were related to wetland phosphorus retent ion. Their results indicated that total phosphorus retention did not vary with landscape position of a wetland. However they suggest that this may have been the function of the landscape parameters chosen, which were too similar to detect any significant differences.

3.0 Hypotheses of study

The questions to be answered by the study are:

  1. Is it possible to delineate wetland boundaries on the basis of biological features (vegetation types), soil type and hydrology using GIS?
  2. Is it possible to delineate habitat boundaries of threatened species (e.g. West-Indian manatee, Iguana and Red-bellied macaw) using GIS?
  3. It is possible to model the movement of pesticides and nutrients from rice paddies and poultry farms, through the wetland?



4.0 Data

The data sets to be used are as follows:

(a) Topographic maps - hard copy format


Three topographic maps at a scale of 1:25,000.


(b) Soils maps - hard copy format


Scale of 1:25,000.


(c) Aerial photographs - hard copy format


Scale of 1:12,500, taken in 1996.


(d) Satellite Data - digital format


 Satellite data

Date of acquisition

Landsat 5 TM

May 1986

Landsat 5 TM

August 1989

Landsat 4 TM

May 1992

Radarsat radar scene

April, October, 1997


(e) GPS data - digital format



4.1. Data import

The importation of the data into the GIS will consist of direct import of the digital data (satellite images) and by the digitizing of the non-digital data (aerial photographs, soils, streams, roads).


4.2. Data structure

The data will be structured both as raster and vector formats. The satellite data, aerial photographs and soils data will be in raster formats, whereas streams and roads will be in vector format.



5.0 Methodology

The analysis of the data for the study will be conducted under a GRASS 4.0 GIS.


5.1. Digitizing of data

Data will be digitized using digitizing modules (r.digit and v.digit) for the raster and vector data structures. This procedure for the non-digital data will be time consuming and prone to errors since paper formats are being us ed.


5.2. Image processing

Image processing will have to be performed on the satellite data because these are raw spectral bands. This processing will involve digital enhancement (manipulating the contrast between objects) and spatial filtering (to detect the edges between features thereby defining boundaries), in order to improve the interpretation of these images.

Image processing will also entail the geo-referencing of the satellite and aerial photographs. All images will be spatially referenced to UTM Projection Zone 20. Ground control points (GCPs) will be collected using the GPS, w hich are to be used to geo-correct the images. It is necessary to use a GPS to collect GCPs because of the lack of many distinct and permanent landmarks in wetland areas.

Classification of the images will have to be undertaken, in order to separate the spectral data into different categories. This is with particular reference to the various vegetation types in the study area.


5.3. Boundary determination

In order to determine the boundary limits of various features such the wetland limits and the critical habitat areas for threatened species, buffers will have to be constructed around these features. This will entail the use of data from the biotic elements (e.g. vegetation classification) and abiotic elements (e.g. roads, slopes values, streams, soils). The land use/land cover arrangement in the wetland will also have to be determined and used in the analysis.


5.4. Liquid flow modeling

In order to assess the movement of nutrients and chemicals across the wetland, GRASS Watershed modeling tool will be used to simulate the flow of these substances. Elevation and stream data will be used to generate the model.

5.5. Field verification

Sites visits have to be made to the wetland and surrounding area to ground-truth the data.



6.0 Anticipated Results

The expected results of the study are as follows:

  1. Improved demarcation of boundaries of the prohibited area.
  2. Delineation of critical habitat for threatened species.
  3. Classification of vegetation types and land use patterns within and surrounding the wetland.
  4. Identification of possible liquid flow courses and areas affected, if and when nutrients and chemicals move through the wetland.


7.0 Policy Applications

Wetlands management has become an environmental policy issue in the last few years in Trinidad and Tobago. It is expected that improved delineation of boundaries in the Nariva wetland will facilitate the implementation of related policy regulations. Presently, if is difficult to effectively enforce, legally defend or resolve conflict regarding these regulations because of the uncertainly of where various boundaries are. In addition, the development of the liquid flow mode l is expected to improve emergency response and mitigatory action regarding adverse effects of nutrient enrichment and pesticide use.


8.0 Budget

The following are the cost anticipated for the project:


    1. Software - GRASS 4.1
    2. No cost, as it is free under a public domain provision.


    3. Hardware - Sun Microsystem Ultra 10 system
    4. 256 MB ECC Ram, 4.3 GB EIDE Hard Drive

      cost = $7,880.00


      GTCO Digitizer Super LX cost = $1,995.00

      sub-total = $9,875.00

    5. Labor - 5 man hours/day for 6 months (1 month = 4 weeks) @ $25.00/hour.
    6. sub-total = $15,000.00


    7. Reporting (5% of project hardware and labor costs)

sub-total =$1,244.00


TOTAL =$26,119.00

9.0 Timeframe

The study will be for a 6 month duration and will consist of the following components:


1. Digitizing - 4 months


2. Field verification - 1 month. Sampling is to take place during the wet and dry seasons.


3. Analysis and reporting - 1 month










ALLENG, G.P., 1994. Advanced application of remote sensing with specific reference to Caribbean coastal environments. Final report submitted to the Caribbean Development Bank, St. Michael’s, Barbados; 51 pp.


ASCHBACHER, J., PRASAD GIRI, C., OFREN, R. S., TIANGCO, P. N., DELSOL, J., SUSELO, T. B., VIBULSRESTH, S. AND CHARUPAT, T., 1994. Final Report: Tropical Mangrove Vegetation Mapping Using Advanced Remote Sensing and GIS Technology. Asian I nstitute of Technology, National Research Council of Thailand, Royal Forest Department, UNEP/GRID, Austrian Academy of Sciences and the Austrian Association for Development and Cooperation; 90 pp.


BACON, P.R., KENNY, J.S., ALKINS, M.E., MOOTOOSINGH, S.N., RAMCHARAN, E.K. AND SEEBERAN, G.S.B., 1979. Studies on the biological resources of Nariva Swamp, Trinidad. Occasional Papers No.4, Zoology Department, University of the West Indies, St. Augustine, Trinidad; 455 pp.


GREINER, M. AND HERSHNER, C., 1998. Analysis of wetland total phosphorus retention and watershed structure. Wetlands, 18(1); 142-149.


MARSHALL, R.C., 1934. Physiology and vegetation of Trinidad and Tobago. Oxford Forestry Memoire, 17; 1-56.


RAMSAR, 1996. Ramsar Convention Monitoring Procedure: Final Report- Nariva Swamp, Trinidad and Tobago. Gland, Switzerland; 102 pp.


SALM, R.V. AND CLARK, J.R., 1984. Marine and Coastal Protected Areas: A guide for planners and managers. International Union for Conservation of Nature and Natural Resources, Gland, Switzerland; 302 pp.


SQUIRES, L., 1993. A research strategy to develop ecological indicators of wetland condition. Pp. 778-791. In M.C. Landin (Ed.), 1993. Wetlands: Proceedings of the 13th Annual Conference of the Society of Wetland Scientists, New Orle ans, LA. South Central Chapter, Society of Wetland Scientists, Utica, MS, USA 39175-9351; 990 pp.


THELEN, K.D. AND FAIZOOL, S., 1980. Plan for a System of National Parks and other Protected areas in Trinidad and Tobago. Technical Document, Forestry Division, Ministry of Agriculture, Lands and Fisheries, Trinidad and Tobago; 106 pp.


TIKASINGH, E.S., 1972. Bush Bush Wildlife Sanctuary. Pp. 67-71. In Wildlife Conservation Committee (ed.). The Wildlife Sanctuaries of Trinidad and Tobago. Ministry of Agriculture, Lands and Fisheries, Trinidad and Tobago; 80 pp.


ZALIDIS, G.C., FITOKA, E.N AND MANTZAVELAS, A.L., 1997. Habitat inventory on two Greek wetlands. Wetlands, 17(4); 439-446.