354 International Forestry Review 5(4), 2003

Forest cover and deforestation in Jamaica: an analysis of

forest cover estimates over time1
O.B. EVELYN and R. CAMIRAND

Forestry Department, Trees for Tomorrow Project, 173 Constant Spring Road, Kingston 8, Jamaica

Email: [email protected]

SUMMARY

Forest cover and related deforestation data published in reputable international forest assessment documents during the past decade
have characterised the island of Jamaica as having one of the world’s highest rates of deforestation. An analysis of pre-1990 land use
data together with a detailed study of 1989 and 1998 LANDSAT TM imagery of the island revealed that there were fundamental
errors in the data used during the 1990s. A study by the Forestry Department and the Trees for Tomorrow Project found that the
annual rate of deforestation for the period 1989–1998 was 0.1%, much less than had been previously estimated. This paper seeks to
explain the variation in existing forest cover data, its change over time and to point to the need for more accurate and reliable data to
be developed internally, based on a rigorous forest cover definition, forest type classification and monitoring methodology.

Keywords: deforestation, forest type classification, forest cover, forest monitoring, Jamaica.

INTRODUCTION Although the information came mainly from institutions

within the countries themselves, there were fundamental
The island of Jamaica is located in the north-western problems associated with the data.
Caribbean Sea and is cantered on latitude 18° 15' N and These data collection problems were addressed at a
longitude 77° 20' W. It is the third largest of the Greater European Community/FAO sponsored workshop held in
Antillean islands and is approximately 230 km from east Trinidad in February 2000. The theme of the workshop
to west and 80 km wide with a total area of 10,900 km². centred on Data Collection and Outlook Effort for Forestry
The name ‘Jamaica’ is believed to be derived from the in the Caribbean, 16 Caribbean countries attended. A summary
Taino word, Xaymaca, which means ‘Land of Wood and of some of the problems encountered by the countries and
Water’. When Christopher Columbus landed on the island some of the recommendations made are as follows:
in 1494, most of it was densely covered with forests, except • The assessment of forest resources on a regular and
for scattered clearings occupied by the Tainos (Swabey 1945). systematic basis in most countries was not
This dense vegetative growth would have protected the soil institutionalised. Most assessments were carried out on
and supported the continuous surface flow of the island’s an ad hoc basis for specific purposes and with the support
many rivers and streams. After centuries of improper land of external assistance.
use, the island has lost much of its natural vegetation and, • While all countries have some capacity to conduct forest
as a consequence, is suffering major environmental problems assessments, institutional cooperation between agencies
such as flooding, soil erosion, destruction of wildlife and was not fully developed.
wildlife habitat, and decreased surface flows in streams and • In general, there is no organised system to collect,
rivers (Cunningham 1993). Better land use management is analyse and distribute forest resource data. However, in
needed to stop the vegetation loss. most countries, there exists informal research that needs
In order to develop policies and plans to better manage to be formalised.
the forests of Jamaica on a sustainable basis, timely, accurate • The structure of data gathering and organisation has to
and reliable data are needed on the extent and location of follow national requirements and capacities. Well-
forest cover and its changes over time. Since 1972, the Food defined national vegetation classification schemes need
and Agriculture Organisation of the United Nations (FAO) to be developed in order to facilitate data aggregation
has been the lead agency documenting areas of forest cover at regional and international levels.
and deforestation rates in Jamaica. In fact, the first major • Local definitions should very clearly state their
forestland use/cover study in Jamaica was undertaken by a
UNDP/FAO Project (Gray and Symes 1972). In recent years 1
A previous version of this text was presented at the Public Awareness
it has become evident that the statistics FAO published for Workshop, United Nations Convention on Desertification (30–31
many Caribbean countries, including Jamaica, were flawed. March 2000) in Kingston, Jamaica.
 Forest cover and deforestation in Jamaica 355

quantitative and qualitative characteristics to facilitate and what definitions can be used to estimate the areas of
regional and international data aggregations. trees outside forests. Table 1 lists some definitions used in
• FAO land use classes are not entirely compatible with the assessment of forests, particularly in Jamaica.
the forest vegetation classes that occur in the Caribbean. Similarly, there are many interpretations of the terms
As a result of this incompatibility, lands with similar ‘deforestation’ and ‘forest degradation’. The following are
woody vegetation were inconsistently assigned to the the FAO definitions (FAO 2001a):
FAO classes. • deforestation: the conversion of forest to another land
• FAO classes must be used with caution and in use or the long-term reduction of the tree canopy cover
conjunction with more detailed definitions of the forest below the minimum 10% threshold.
classes in each country. • forest degradation: changes within the forest that
negatively affect the structure or function of the stand
The above observations summarise some of the problems or site, and thereby lower its capacity to supply products
associated with the data FAO used in its publications. and/or services.
Deforestation data published in 1995 by FAO on Jamaica
contained some of these problems and has caused much After reviewing the above definitions, it is clear that any
debate and controversy within the country. This has pointed meaningful comparison of two different land use/cover
to the need for more accurate and reliable data to be studies requires using the same definition of forest types
developed. and the same criteria for the vegetation interpretation and
Factors such as the definition of ‘forest’ and classification. However, almost all of the land use/cover
‘deforestation’, the different techniques used in the various studies carried out in Jamaica used different definitions
land use/cover studies, and the purpose of these studies, for ‘forest’ and ‘forestland’. This has made it difficult to
have also led to a wide range of variations and inconsistencies compare the results from the various studies, as outlined
in the data on the rate of change in the forests of Jamaica. and analysed below.
This paper reviews the existing studies on the estimate
of forest cover in Jamaica and introduces a methodology
for systematic and sustainable evaluation and monitoring FOREST COVER CHANGE OVER TIME
of the Jamaican forests.
Forest cover change in Jamaica is relatively well
documented, but the results are highly variable and the
DEFINITIONS OF TERMS ‘FOREST’ AND estimates of annual deforestation rates range from between
‘DEFORESTATION’ 0.03 to 6.7% (Table 2). As hypothesised for Costa Rica’s
forest cover estimates by Kleinn et al. (2000), the reasons
Evelyn (1997) pointed out in his report on deforestation in for the differences in the forest cover estimates over time
Jamaica that the different definitions of the terms ‘forest’ and the related deforestation rate are as follows:
and ‘deforestation’ used in various land use/cover studies • the definition of forest and the forest classes included
in Jamaica have created problems when comparing results (forest and/or wooded land);
of studies carried out in earlier periods to results of later • the reference area (entire country or region);
studies. According to UNEP (1998), there are many • the reference year (photograph/image production year
definitions in use for what is classified as forested lands or study publication year);

TABLE 1 Definitions of forest

Definitions of forest Reference

All lands with a forest cover (including bamboo and palm); i.e. with trees whose FAO 1967. Definition used in the Gray
crown covers more than 20% of the area, and not used primarily for purposes other & Symes (1972) land use/cover study in
than forestry. Include areas temporarily unstocked and all plantations established for Jamaica.
forestry purposes. Exclude isolated tree groups smaller than 0.5 hectare.
Coniferous – Forest stands including both coniferous and broad-leaved evergreens. CRIES 1982. Definition used in the
The coniferous evergreens are commonly referred to as softwoods. The forest must Comprehensive Resource Inventory and
have a 75% or more crown closure and be in stands of at least 16 hectares or more. Evaluation System (CRIES) Project
Deciduous – Forest stands made of a mixture of various broadleaf species. The (1982) and the Rural Physical Planning
deciduous forest types must have a 75% or more crown closure and be in stands of Department (1988) land use/cover studies
at least sixteen hectares or more. in Jamaica.
Forest includes natural forests and forest plantations. The term is used to refer to land FAO 2001a. Definition used in the Forest
with a tree canopy cover of more than 10 percent and area of more than 0.5 ha. Forests Resources Assessment 2000.
are determined both by the presence of trees and the absence of other predominant
land uses. The trees should be able to reach a minimum height of 5 m. Young stands
that have not yet reached, but are expected to reach, a crown density of 10 percent and
tree height of 5 m are included under forest, as are temporarily unstocked areas.
356 O.B. Evelyn and R. Camirand

TABLE 2 Studies of forest cover change in Jamacia

Annual
Period/Region 1 rate (%) Reference
??? / Jamaica -3 FAO/UNEP 1981 In Eyre 1987b
??? / Jamaica -3 US Congress 1984 In Eyre 1987b
??? / Jamaica -2.05 Allen & Barnes 1985 In Eyre 1987b
??? / Jamaica -3 Aiken et al. 1986 In Eyre 1987b
1980–1986 / Jamaica -3.3 Eyre 1986, 1987b
1980–1986 / Jamaica (Rural districts) -4.3 Eyre 1986, 1987b
1980–1986 / Jamaica (Land altitude > 1000 m) -1 Eyre 1986, 1987b
CRIES Studies 1981–1987 / Cockpit Country -2.8 Eyre 1989 In Miller 1998
??? / Jamaica -5.3 WRI 1994 In Eyre 1994, 1996
1980–1990 / Jamaica -5.3 FAO 1995
1982–1993 / Hope & Upper Yallahs watersheds -1 Graaff de 1997
1990–1995 / Jamaica -6.7 FAO 1998b
1961–1991 / Cockpit Country (Primary rainforest) -0.09 Miller 1998
1961–1991 / Cockpit Country (Including secondary forest) -0.03 Miller 1998
1989–1998 / Jamaica -0.1 FD-TFT 1999
1990–2000 / Jamaica -1.5 FAO 2001a, 2001b
1
??? symbol indicates no information concerning the period.

550
• the type of study (mapping or field sampling or both);
• the precision of the estimates (photographs or satellite 500
y = 0.003x2 - 1.454x + 433.477
Forest Area (,000 ha)

imagery or field survey); 450 R2 = 0.637

• the information sources used (traditional forest
400
inventory or research plots); and
• the objectives of a particular study (agriculture, forestry 350

or conservation). 300

250
Forest cover estimation before 1950
200
0 100 200 300 400 500
It is difficult to determine the area of forestlands before
Period (year)
1950 because the means of assessment were very limited
and a clear definition of the term ‘forestlands’ or ‘tropical FIGURE 1 Trend of Forest Area Change in Jamaica from
rainforests’ was not stated by the authors. Table 3 lists some 1491 to 2000, excluding 1980 data
estimates of the probable area of forests in Jamaica before
1950. (number of weeks with rainfall > 50 mm/week). Therefore
Eyre (1996) estimated that the depletion of tropical the total forest area was probably higher.
rainforest in Jamaica was 224,350 ha over a period of 300 According to Eyre (1986, 1987b), there was a wave of
years (1491 to 1791), which is approximately 748 ha or forest clearance in the 18th century for plantation
0.17% per annum. The depletion of the forestlands area establishment in Jamaica. He also stated that ‘in the
was 104,530 ha between 1886 and 1943, or approximately plantation era many parts of the country were considerably
0.57% per annum (Hooper 1886, Swabey 1945) and the more bare of trees than presently’ (Eyre 1986). Also the
forested lands covered approximately 20% of the island. It mid to late 19th century seems to be the start of a period of
should be noted that Eyre’s figure for 1491 concerned only substantial regrowth of forests mainly because of the
the forest zone named ‘tropical rainforest’ at that time. This decrease of human economic usage of land (Eyre 1987b;
zone was based on an estimation of the mean annual Higman 1988; Satchell 1990), i.e. probably the decrease of
temperature (>23.8°C or <1,250 m in altitude) and rainfall sugar cane production. The trend of forest cover change in
Jamaica is shown in Figure 1. The same relative trend is
TABLE 3 Jamacia forest cover estimates before 1950 hypothesised by Watts (1987) for some other countries in
the West Indies.
Forest lands Tropical
Year (ha) rainforests (ha) Reference
Forest cover estimation between 1950 and 1990
1491 432,500 Eyre 1996
1791 208,150 Eyre 1996 This period was the beginning of systematic land use/cover
1886 323,760 Hooper 1886 studies using mainly remote sensing techniques. Four
1943 219,230 Swabey 1945 studies were carried out during this time using aerial
 Forest cover and deforestation in Jamaica 357

TABLE 4 Jamacia forest cover 1954–1985

Natural Forest Other wooded Total

Year forest (ha) (%) plantation (ha) land (ha) (ha) (%) Reference
1954 319,800 (29%) 319,800 (29%) DOS 1958 In Eyre 1986
1968 260,869 (24%) 4,131 226,252 491,252 (45%) Gray and Symes 1972
1980 504,453 (46%) 3,709 23,427 531,589 (48%) CRIES 1982
1985 329,114 (29%) 5,930 218,282 553,326 (48%) RPPD 1988

photographs taken in 1954, 1968, 1980 and 1985 have reported such a large change in forest cover for
respectively. The analysis of land use/cover changes has Jamaica in its 1990 report.
been undertaken by Eyre (1986) for the 1954 and 1980 data Another classification anomaly noted in the CRIES
(with some ground truth investigations), by the Rural study is that although it defines forest as ‘75% or more crown
Physical Planning Division (RPPD 1988) for the 1980 and closure and be in stands of 16 ha or more’ (see definitions
1985 data, and by Evelyn (1997) for the 1968, 1980 and in Table 1), the study identified the most forest area as shown
1985 land use/cover data. in Table 4. It would be expected that, given the more
The objectives and the methodologies for the analysis restrictive definition of forest used, the CRIES study should
of the land use/cover changes were different in each study. have reported less forest areas than the other studies.
It is therefore very difficult to compare these studies and Eyre (1986) noted, quite correctly, that the Gray and
to form any conclusion about the depletion of forest cover Symes (1972) study and CRIES (1982) study were totally
in Jamaica. FAO used the 1980 data (Comprehensive incompatible for a comparative analysis of the
Resource Inventory and Evaluation System (CRIES) deforestation rate in Jamaica. This is so because the Gray
Project 1982) and 1985 data (Rural Physical Planning and Symes study was designed to acquire forest data while
Division (RPPD) 1988) and applied ‘a modelling technique the CRIES study was oriented to acquiring general land
… which makes use of ancillary variables (e.g. population use knowledge of Jamaica. This is another cause for what
density and ecological zone) correlated with forest cover appears to be an increase in forest cover between the period.
area and its change in time’ to derive the figures for Jamaica In 1986, Eyre also carried out a field study in 24
(5.29% per annum of deforestation) published in its Forest Jamaican districts to verify the rate of change in the forests.
Resource Assessment 1990 (FAO 1993). Using the 1980 data (CRIES 1982) as a baseline, he sampled
Analysis of the data from 1980 and 1985 yielded a 687 km², about 13% of the forested area reported and
negative change (-175,339 ha) on the forest cover in derived an annual rate of deforestation of 3.3% for the
Jamaica, that is, 35% depletion of Jamaican forests for the period 1980 to 1986. However, as previously mentioned,
period (Table 4). Evelyn (1997) pointed out that the main the CRIES data contained errors and inconsistencies which
problem with this analysis is the misclassification of the needed to be accounted for in any analysis of forest cover.
‘Dry limestone forest’ areas (Portland Ridge in Clarendon, The above discussion of available forest cover data shows
Hellshire Hills in St. Catherine, Long and Dallas that, except for the 1980 CRIES figure (see Table 4), the
Mountains in St. Andrew, and Albion in St. Thomas). Gray changes in forested area of Jamaica over the past few
and Symes (1972) classified the dry limestone forest areas decades are not very large. The forested area varies between
as ‘Other wooded areas’, while CRIES (1982) used 24 to 29% of the country over the period of assessment. In
‘Deciduous forest’ (see definition in Table 1). A close fact, the percentage of natural forest for 1954 and 1985 (31
analysis of the FAO definitions of what constitute forests year period) is approximately the same, showing little change.
would put ‘Deciduous forest’, as defined by the CRIES
report, in the compilation of total forest areas while ‘Other Forest cover estimation since 1990
wooded’ areas would not be included. This explains why
the forested area reported for 1980 was so high; it included Since 1990, three land use/cover studies have been carried
these ‘Other wooded’ areas. out based on interpretation of 1989 and 1998 LANDSAT
In 1988, the RPPD reclassified the dry limestone forests TM satellite images: The Nature Conservancy (TNC)
as ‘Brush’, which is similar to the FAO definition of ‘Other (Grossman et al. 1992), the Forestry Department/Trees for
wooded’, and this time correctly excluded these areas from Tomorrow Project (FD-TFT) (Forestry Department 1999)
the reported forested area figure submitted to FAO. The and in 2000, a new estimation was undertaken by TNC using
‘change’ indicated between the CRIES and RPPD reports the FD-TFT delineation of the 1998 satellite images and a
was therefore not a real change in forested area but a re- new vegetation classification developed for the Caribbean
classification. This means that the forested area of 504,435 Atlas Project (Li et al. 2000). Although all the studies used
ha reported by CRIES in 1980 needed to be adjusted LANDSAT TM imagery, they nonetheless used different
downwards by approximately 175,000 ha or the RPPU methods to classify land use/cover and different software
figure adjusted upwards by the same amount when for the interpretation, which would require some regrouping
comparing the figures for the two periods. If this situation as that shown in Appendix 1 in order to compare the data.
had been brought to the attention of FAO, it would not The compilation and comparison of forested area is set out
358 O.B. Evelyn and R. Camirand

in Table 5. The difference in areas between the TNC classification systems, are outlined in Figure 2 and
(Grossman et al. 1992; Li et al. 2000) and the FD-TFT Appendix 1, respectively.
(1999) analyses is what would be expected and can be The new FD-TFT figure of forest cover has been
explained by the different systems that were used. Theses accepted by FAO. It was used in the regression analysis and
differences between the total forest area figures are not extrapolation for the 2000 estimates published in its
significant with only a 3% difference. publications Forest Resources Assessment 2000 (FRA 2000)
The FD-TFT assessment found that, because of the size and State of the World’s Forests 2001 (FAO 2001a; 2001b).
and mountainous nature of the island, the limitations in The areas of natural forest and forest plantation, presented
image processing and classification techniques of in these publications, represent 29.2% and 0.8% of the island
LANDSAT TM technology (Richards 1986; Schowengerdt of Jamaica, respectively (Table 5). The FD-TFT, TNC and
1983), LANDSAT TM could only be used for broad FAO-FRA 2000 forest area figures show a close correlation.
national forest classification. The more detailed A comparison between the total forest area estimates
classification from these broad types, which is needed for for the year 1990 presented in the FRA 1990 (FAO 1993,
operational and forest management plans, require the use 1995) and in the FRA 2000 (FAO 2001a, 2001b) shows a
of aerial photographs. The classification system that was difference of 49.2%, i.e. 254,000 ha (FRA 1990) versus
developed by the FD-TFT, the definitions for the classes, 379,000 ha (FRA 2000). This large difference in the
and how they correspond to both the TNC and the FAO estimation is explained by changes in methodology for

FIGURE 2 Diagram of land use/cover types hierarchical classification

TABLE 5 Jamacia forest cover 1988–2000

Natural forest Forest (ha) (%) Total forest

Year (ha) (%) plantation (ha) (%) Reference
1988/89 362,012 (33.0%) 3,232 (0.3%) 365,244 (33.3%) TNC/Grossman et al. 1992
1998 359,365 (32.8%) 10,227 (0.9%) 369,592 (33.7%) TNC/Li et al. 2000
1989 335,079 (30.6%) 8,856 (0.8%) 343,935 (31.4%) FD-TFT 1999
1998 332,016 (30.3%) 8,187 (0.7%) 340,203 (31.0%) FD-TFT 1999
1990 379,000 (35.0%) FAO 2001b (FRA 2000) 1
2000 316,000 (29.2%) 9,000 (0.8%) 325,000 (30.0%) FAO 2001b (FRA 2000)
1
In the FRA 1990 (FAO 1993, 1995): Total forest 1990 = 254,000 ha (239,000 ha + 15,000 ha).
 Forest cover and deforestation in Jamaica 359

generating forest area change figures for developing methodology needs a very careful analysis of each area
countries. The FRA 1990 area and forest area change value in time used to estimate the area change rate,
estimates were generated by means of a deforestation model, particularly the baseline value.
i.e. a forest area adjustment model using population density
and growth (FAO 1993). The use of the model was
discontinued for FRA 2000 after acknowledging that the A SYSTEM FOR MONITORING FOREST COVER
correlation between forest area change and the demographic
parameters used as model input was weak at the national To prevent future problems created by using different
level. The use of the model for forest area estimates may definitions of land use/cover types and methodologies, the
have overestimated deforestation figures for some tropical Forestry Department, assisted by the Trees for Tomorrow
countries for the 1980s (FAO 2001c). In their study covering Project, is establishing protocols to permit the systematic
six countries of West Africa, including Ivory Coast with a evaluation of changes in forest cover in Jamaica. The Forest
deforestation rate of 7.3% in 1981–85, Fairhead and Leach Act, 1996 charges the Department with the responsibility
(1998) suggest that twentieth-century deforestation rate in of preparing a National Forest Management and
these countries is probably only one-third of that suggested Conservation Plan, which should be revised at intervals of
by the international data in circulation. ‘not more than five years’. The first plan was completed in
For Jamaica, time series information showing high 2001 and contains ‘an inventory and description of
comparability between the observations (12 publications forestlands’ as the Act stipulated. The Department is
consulted), have been used for estimating the forest area therefore the agency mandated for the sustainable
change, i.e. the FRA 2000 area change has been estimated collection, analysis and distribution of forest cover data in
by means of extrapolation of existing data sets from two Jamaica. The steps being taken by the Forestry Department
or more points in time (FAO 2001c). That explains the in this direction are as follows:
difference between the area change percentages of the FD- • developed a national classification system for
TFT (2 area values in time, 1989–1998) versus the FAO- LANDSAT TM interpretation and aerial photographs;
FRA 2000 (a series of area values in time). The FAO • beginning in 1999, undertake vegetation interpretation

TABLE 6 Land use/cover change in Jamaica (1989–1998)

1989 1998 Difference Lost/Gain

Land use/cover types 1 (ha) (ha) (%) (%)
Forest land use/cover (>75%)
BB Bamboo 2,791 2,979 188 6.74
MG Mangrove 9,751 9,731 -20 -0.21
PF Closed Broadleaf 88,717 88,231 -486 -0.55
SF Disturbed Broadleaf 177,254 174,725 -2,530 -1.43
SL Short Open Dry 12,083 12,104 21 0.17
SW Swamp 2,358 2,247 -111 -4.69
WL Tall Open Dry 42,125 41,998 -127 -0.30
Sub-total 335,079 332,015 -3,064 -0.91
Mixed land use/cover (first class >50%; second class >25%)
BC Bamboo and Fields 29,818 29,047 -772 -2.59
BF Bamboo and Disturbed broadleaf 12,311 12,687 376 3.05
BS Bauxite and Disturbed broadleaf 1,590 2,960 1,370 86.13
CS Fields and Disturbed broadleaf 118,898 117,966 -932 -0.78
PP Fields/Disturbed broadleaf and Pine Plantation 8,856 8,187 -669 -7.56
SC Disturbed broadleaf and Fields 166,838 165,954 -884 -0.53
Sub-total 338,312 336,801 -1,511 -0.45
Non-forest land use/cover
BA Buildings / Other infrastructure 51,910 52,260 350 0.67
BE Bauxite extraction 1,193 4,922 3,729 312.47
BR Bare rock 867 934 67 7.72
FC Fields 273,176 274,479 1,303 0.48
HW Herbaceous wetland 10,914 10,914 0 0.00
PC Plantations 83,145 82,341 -804 -0.97
WA Water bodies 1,656 1,586 -70 -4.23
SI Small islands 164 164 0 0.00
Sub-total 423,025 427,600 4,574 1.08
TOTAL 1,096,416 1,096,416 0
1
Definition of land use/cover types in Appendix 1.
360 O.B. Evelyn and R. Camirand

of satellite images using the classification system clouds and shadows. Table 6 gives the 1989 and 1998 land
developed; use/cover areas derived from the analysis.
• undertake interpretation using satellite images every 5 The total change in the forestland use/cover classes over
or 10 years with the same definition of vegetation types the period is a decrease of 3,064 ha or 0.91%. This gives a
and the same methodology of interpretation; decrease of approximately 0.1% per annum (Table 6). The
• carry out periodic detailed analysis using aerial largest increase among all land use/cover classes is in the
photography for management planning purposes and bauxite areas, which shows an overall increase of 5,099 ha
the monitoring of localised changes; and (3,729 ha plus 1,370 ha). Approximately one-half of the
• monitor any destructive changes that are taking place new bauxite areas were probably originally agriculture fields
in the forests and develop mitigation measures to deal (non-forestlands). This is understandable due to the fact
with them. that bauxite is traditionally extracted in the valleys, covered
by pasture, citrus orchard, crops, evergreen forest or forest
As a first step, an analysis of forest cover change over the plantation, between the limestone hills (Scholten and
period 1989 to 1998 was carried out by the FD-TFT using Andriesse 1986). In addition to the increase of the bauxite
the following LANDSAT TM images: areas, the detailed study shows a net degradation of the
• full scene 012/047 (10 Dec 1989), full scene 011/047 (03 forest cover, i.e. an increase of the mixed forest type with
Dec 1989) and quarter scene 011/047 (12 Nov 1990); agriculture fields (Table 7). For instance, a significant area
• full scene 012/047 (07 April 1998), quarter scene 012/ changes from the ‘disturbed broadleaf forest’ type to the
047 (02 Feb 1998) and full scene 011/047 (04 Oct 1996). ‘disturbed broadleaf forest and fields’ mixed type (2,018
ha). There has also been a significant increase in the
The system developed for the LANDSAT TM interpretation proportion of agriculture fields in the mixed types, for
was limited to a mapping scale of 1:100,000. The smallest example, 1,301 ha of ‘disturbed broadleaf forest and fields’
area or polygon to be resolved was 25.0 ha. A supervised mixed type was re-classified to the ‘fields and disturbed
classification system was used and over 100 locations island- broadleaf forest’ mixed type.
wide were visited for ‘ground truthing’ purposes (and Over the last ten years a positive change in the forest
photographed) and the coordinates taken by GPS for areas has been noted in some Caribbean islands such as
verification purposes. Aerial photographs (1991–92 1:15,000 Cuba and Guadeloupe (FAO 2001a, 2001b). This has also
and 1999 1:40,000) were also used to verify seven large been observed in Jamaica in the Buff Bay/Pencar Rivers
blocks, which were very difficult to interpret because of watershed located in the north-eastern section of the island.

TABLE 7 Interchange matrix of land use/cover area in Jamaica (1989–1998)

Land use/cover 1989 Total Difference

Land use/cover Forest Mixed Non-forest Gain 1998-1989
types 1 BB MG PF SF SL SW WL BC BF BS CS PP SC BA BE BR FC HW PC WA SI (ha) (ha)
BB 3 139 46 188 188
MG 1 40 42 -20
PF 0 -486
Forest SF 396 106 7 62 571 -2,530
SL 227 227 21
SW 0 -111
WL 6 6 -127
Land use/cover 1998

BC 193 16 90 299 -772

BF 241 156 172 569 376
Mixed BS 229 230 911 1,370 1,370
CS 359 28 378 385 1,301 1 30 60 2,541 -932
PP 0 -669
SC 90 2,018 357 101 8 2,573 -884
BA 85 31 105 27 76 29 351 350
BE 1,725 250 1,754 3,729 3,729
BR 67 67 67
Non- FC 62 167 109 111 372 979 89 947 657 3,491 1,303
forest HW 0 0
PC 123 3 127 -804
WA 47 47 -70
SI 0 0
Total loss (ha) 0 62 486 3,101 206 111 133 1,071 193 0 3,472 669 3,457 1 0 0 2,188 0 931 117 0 16,197 0
1
Definition of land use/cover types in Appendix 1.
 Forest cover and deforestation in Jamaica 361

This area was selected as a pilot area by the Trees for EYRE, L.A. 1987b. Jamaica: a test case for tropical deforestation.
Tomorrow Project in which detailed land use and socio- Ambio, 16(6):336–343.
economic studies were undertaken. The area was mapped EYRE, L.A. 1989. Slow death of a tropical rainforest: the
at a scale of 1:10,000 using aerial photographs and the Cockpit Country of Jamaica, West Indies. In: LURIA, M.,
STEINBERGER, Y. and SPANIER, E. (eds.) Environmental
land use analysis showed a positive change in forest areas
quality and ecosystem stability – Vol IVA – Environmental
of 0.26% per annum for the period of 1991 to 1999
quality. ISEQS Publication, Jerusalem, Israel. pp. 599–606.
(Forestry Department 2001b). These changes are mainly EYRE, L.A. 1994. Jamaica’s deforestation rate is world’s highest.
attributable to an increase of Caribbean pine (Pinus Sunday Gleaner, December 18.
caribaea mor. var. Hondurensis Barr and Golf.) plantations EYRE, L.A. 1996. The tropical rainforests of Jamaica. Jamaica
(or areas regenerated) and an abandonment of cultivation Journal, 26(1):26–37.
fields (coffee and food crops). FAIRHEAD, J. and LEACH, M. 1998. Reframing deforestation
– Global analysis and local realities: studies in West Africa.
Routledge, London. 238 pp.
ACKNOWLEDGEMENTS FAO 1967. Report of the headquarters meeting of forest
inventory experts on UNDP/SF projects. Held in Rome, 11–
22 September 1967. 260 pp.
We thank Marilyn Headley, Conservator of Forests,
FAO 1993. Forest resources assessment 1990 – Tropical countries.
Forestry Department; John Latham, Senior Forest Advisor
FAO Forestry Paper 112, Rome.
and Susie Latham, Agro-economist, Trees for Tomorrow FAO 1995. National forestry action programmes – Update 32.
Project for their helpful comments on the manuscript. We NFAP Support Unit, FAO Forestry Department, Rome.
would also like to thank Claude Tremblay, Remote Sensing 369 pp.
Specialist, who did most of the LANDSAT image analysis FAO 1998a. FRA 2000 terms and definitions. Forest Resource
for the Trees for Tomorrow Project. The Trees for Tomorrow Programme, Rome, 4 November 1998. 18 pp.
Project is funded by the Government of Jamaica and the FAO 1998b. Latin America and the Caribbean national forest
Canadian International Development Agency (CIDA). programmes – Update 33. FAO Regional Office, Santiago,
Chile. 206 pp.
FAO 1998c. FRA 2000 guidelines for assessments in tropical and
sub-tropical countries. Forestry Department working paper
REFERENCES
2, Rome. 43 pp.
FAO 2001a. State of the world’s forests 2001. Rome. 181 pp.
AIKEN, K., GARRAWAY, E. and BOOTHE, R. 1986. A case
FAO 2001b. Global forest resources assessment 2000 – Main
for protected forested areas in Jamaica. Univ. of the West
report. FAO Forestry Paper 140, Rome. 479 pp.
Indies, Zoology Dept., Kingston.
FAO 2001c. Comparison of forest area and forest area change
ALLEN, J.C. and BARNES, D.F. 1985. The causes of
estimates derived from FRA 1990 and FRA 2000. Forest
deforestation in developing countries. Ann. Ass. Amer. Geogr.,
Resources Assessment Programme working paper 59, Rome.
75:163–184.
67 pp.
ARECES-MALLEA, A.E., WEAKLEY, A.S., LI, X., SAYRE,
FAO/UNEP 1981. Tropical forest resource assessment project.
R.G., PARRISH, J.D., TIPTON, C.V. and BOUCHER, T.
1999. A guide to Caribbean vegetation types: preliminary FAO, Rome.
classification system and descriptions. The Nature FORESTRY DEPARTMENT 1999. Land use/cover types areas
Conservancy / International Institute of Tropical Forestry– 1989 and 1998 per watershed management unit and
USDA Forest Service / Eros Data Centre – U.S. Geological protection status, Jamaica. Trees for Tomorrow Project
Service / USAID, Washington. 166 pp. technical report, Kingston. 69 pp., 1:100 000 and 1:250 000
BROWN, E.M. 1956. Jamaica – Statement by the Forest scale colour maps.
Department. Prepared for the British Commonwealth FORESTRY DEPARTMENT 2001a. National forest
Forestry Conference 1957, Kingston. 14 pp. management and conservation plan – Jamaica. Ministry of
COMPREHENSIVE RESOURCE INVENTORY AND Agriculture, Kingston. 100 pp.
EVALUATION SYSTEM (CRIES) PROJECT 1982. FORESTRY DEPARTMENT 2001b. Buff Bay Pencar
Jamaican resource assessment. Rural Physical Planning watershed management unit – Forest management plan.
Department-Ministry of Agriculture/Michigan State Ministry of Agriculture, Kingston. 151 pp.
University/USDA/ Ohio State University, Kingston. 75 pp. GRAAFF, J. de 1997. Evaluating the sustainable development
CUNNINGHAM, C.G. 1993. A brief report on the status of of the Kingston watersheds in Jamaica, 1980–1993.
the island’s watersheds with emphasis on Yallahs Valley, Wag Caribbean Geography, 8(1):46–56.
Water, Buff Bay, and Rio Grande. Kingston. 25 pp. GRAY, K.M. and SYMES, G.A. 1972. Forestry development
DOS 1958. Topographic and land use survey of Jamaica. and watershed management in the upland regions, Jamaica
Directorate of Overseas Surveys (UK), London. – Forest inventory of Jamaica. UNDP/FAO technical report
EVELYN, O.B. 1997. Deforestation in Jamaica: an analysis of 3, Kingston. 175 pp.
the data. Forestry Department, Kingston. 9 pp. GROSSMAN, D.H., IREMONGER, S.F. and MUCHONEY,
EYRE, L.A. 1986. Deforestation in Jamaica: its rate and D.M. 1992. A rapid ecological assessment of Jamaica – Phase
implications. Department of Geology, University of the West I: an island-wide characterisation and mapping of natural
Indies, Kingston. 29 pp. communities and modified vegetation types. The Nature
EYRE, L.A. 1987a. Deforestation in Jamaica. Journal of the Conservancy / Conservation Data Centre-Jamaica / Rural
Scientific Research Council of Jamaica, 6:17–24. Physical Planning Division, Arlington, Virginia. 44 pp.
362 O.B. Evelyn and R. Camirand

APPENDIX 1 Definition of land use/cover types (1:100 000 scale – LANDSAT TM 1998)

Definition Corresponding Corresponding

Type (code) (FD-TFT) TNC classes 1 FAO classes 2

Forest Land Use/Cover (> 75 %, Minimum unit: 25 ha)

Closed Broadleaf (PF) Closed primary forest with broadleaf trees Undisturbed
at least 5 m tall and crowns interlocking, Upper Montane Forest Closed Forest
with minimal human disturbance Lower Montane Forest
Semi-evergreen Moist
Disturbed Broadleaf (SF) Disturbed broadleaf forest with broadleaf Disturbed
Broadleaf Forest
trees at least 5 m tall and species-indicators Closed Forest
(Natural Communities)
of disturbance such as Cecropia peltata
(trumpet tree)
Bamboo (BB) Bambusa vulgaris (bamboo brakes) on the Disturbed Natural Forest Other Wooded Land
lower shale hills (disturbed forest) (Modified Communities)
Tall Open Dry (WL) Open natural woodland or forest with trees Deciduous/Semi-deciduous Open Forest
at least 5 m tall and crowns not in contact, Broadleaf Forest
in drier part of Jamaica with species- (Natural Communities)
indicators such as Bursera simaruba (red birch)
Short Open Dry (SL) Open scrub, shrub, bush or brushland with trees Thorn Forest Other Wooded Land
or shrubs 1–5 m tall and crowns not in contact, (Natural Communities)
in drier part of Jamaica with species-
indicators such as Prosopis juliflora (cashaw)
or Stenocereus hystrix (columnar cactus)
Swamp (SW) Edaphic forest (soil waterlogging) with a Freshwater Swamp Forest Disturbed
single tree storey with species-indicators (Natural Communities) Closed Forest
such as Symphonia globulifera (hog gum)
and Roystonea princeps (royal palm)
Mangrove (MG) Edaphic forest (areas with brackish water) Mangrove Disturbed
composed of trees with stilt roots or (Natural Communities) Closed Forest
pneumatophores, species-indicators such as
Rhizophora mangle (red mangrove)
Mixed Land Use/Cover
Fields or Disturbed Broadleaf >50% fields or Disturbed Broadleaf forest; Forest Pl antations Plantations
Forest and Pine Plantation (PP) >25% Pine plantation (Modified Communities)
Disturbed Broadleaf Forest >50% Disturbed Broadleaf forest; Disturbed Natural Forest 75% Other Wooded Land
and Fields (SC) >25% fields (Modified Communities) 25% Other Land
Bamboo and Disturbed >50% bamboo; >25% Disturbed Broadleaf forest Disturbed Natural Forest Other Wooded Land
Broadleaf Forest (BF) (Modified Communities)
Bamboo and Fields (BC) >50% bamboo; >25% fields Disturbed Natural Forest 75% Other Wooded Land
(Modified Communities) 25% Other Land
Fields and Disturbed Broadleaf >50% fields; >25% Disturbed Broadleaf forest Disturbed Natural Forest 25% Other Wooded Land
Forest (CS) (Modified Communities) 75% Other Land
Bauxite Extraction and Disturbed >50% bauxite extraction; Disturbed Natural Forest 25% Other Wooded Land
Broadleaf Forest (BS) >25% Disturbed Broadleaf forest (Modified Communities) 75% Other Land
Non Forest Land Use/Cover
Plantations (PC) Tree crops, shrub crops like sugar cane, bananas, Non-forest Land Cover Other Land
citrus and coconuts (Agriculture)
Fields (FC) Herbaceous crops, fallow, cultivated grass/ Non-forest Land Cover Other Land
legumes (Agriculture)
Herbaceous Wetland (HW) Edaphic vegetation (soil waterlogging) with Non-forest Land Cover Other Land
herbaceous plants (Natural Communities)
Water Bodies (WA) Lakes, rivers Water Bodies (Open Water) Inland Water
Small Islands (SI) Mostly sand/limestone, unvegetated small Non-forest Land Cover Other Land
islands (cays) (Natural Communities)
Bare Rock (BR) Bare sand/rock Non-forest Land Cover Other Land
(Natural Communities)
Bauxite Extraction (BE) Surface mining/bauxite Non-forest Land Cover Other Land
(Urban/Industrial)
Buildings and Other Buildings and other constructed features such Non-forest Land Cover Other Land
Infrastructure (BA) as airstrips, quarries, etc. (Urban Industrial)
1
Grossman, Iremonger and Muchoney (1992), Muchoney, Iremonger and Wright (1994).
2
FAO (1998a, 1998c).
 Forest cover and deforestation in Jamaica 363

HIGMAN, B.W. 1988. Jamaica surveyed: plantation maps and RICHARDS, J.A. 1986. Remote sensing digital image analysis:
plans of the eighteenth and nineteenth centuries. Institute of an introduction. Springer-Verlag, Berlin. 280 pp.
Jamaica Publications Ltd, Kingston. 307 pp. RURAL PHYSICAL PLANNING DIVISION 1988.
HOOPER, E.D.M. 1886. Report upon the forests of Jamaica. Quantitative and qualitative analysis of land use changes
Waterlow and Sons Ltd., London. 33 pp. related to forest degradation in Jamaica between 1979 and
KLEINN, C., CORRALES, L. and MORALES, D. 2000, 1985. Ministry of Agriculture, Kingston.
(Submitted) Forest area in Costa Rica: a comparative study SATCHELL, V.M. 1990. From plots to plantations: land
of tropical forest cover estimates over time. Paper presented transactions in Jamaica, 1866–1900. Institute of Social and
at the workshop/training course Designing Multipurpose Economic Research, University of the West Indies, Kingston.
Resource Inventories, held in Turrialba, Costa Rica, 21–25 197 pp.
February 2000. 21 pp. SCHOLTEN, J.J. and ANDRIESSE, W. 1986. Morphology,
LI, X., WALLER, W., SAYRE, R.G., BOUCHER, T.M., genesis and classification of three soils over limestone,
SOTOMAYOR, L., KERNAN, K., GOYERT, W.M., Jamaica. Geoderma, 39:1–40.
HELMER, E.H., RAMOS, O.M., DEL MAR LOPEZ, T., SCHOWENGERDT, R.A. 1983. Techniques for image processing
QUINONES, M., WOOD, E.C., NELSON, J.J., COAN, M. and classification in remote sensing. Academic Press, Orlando,
and ARECES-MALLEA, A.E. 2000. Mapping land use and Florida. 248 pp.
natural vegetation for the islands of Jamaica, Puerto Rico, SWABEY, C. 1945. Forestry in Jamaica. Forestry Bulletin No.
and Dominica. The Nature Conservancy / International 1, Forest Department, Kingston. 44 pp.
Institute of Tropical Forestry – USDA Forest Service / Eros UNEP 1998. Countries with low forest cover. Discussion paper,
Data Centre – U.S. Geological Service, Washington. 24 pp. Nairobi, Kenya. 12 pp.
and 1:250 000 scale colour maps. UNITED STATES CONGRESS 1984. Technologies to sustain
MILLER, D.J. 1998. Invasion of the cockpits: patterns of tropical forest resources. Office of Technology Assessment,
encroachment into the wet limestone rainforest of Cockpit Washington.
Country, Jamaica. In: McGREGOR, D.F.M., BARKER, D. WATTS, D. 1987. The West Indies: patterns of development,
and LLOYD EVANS, S. (eds.) Resource sustainability and culture and environmental change since 1492. Cambridge
Caribbean development. The Press University of the West Studies in Historical Geography 8, Cambridge University
Indies, Kingston. pp. 373–389. Press, Cambridge. 609 pp.
MUCHONEY, D.M., IREMONGER, S.F. and WRIGHT, R. WORLD RESOURCES INSTITUTE 1994. World resources
1994. A rapid ecological assessment of the Blue and John Crow 1994–1995. Oxford University Press.
Mountains National Park, Jamaica. The Nature Conservancy
/ Conservation Data Centre – Jamaica / Jamaica Agricultural
Development Foundation – Research Programme, Arlington,
Virginia. 90 pp.