deforestation and the spatio-temporal distribution of savannah and forest members of the simulium...

TRANSACTIONSOFTHEROYALSOCIETYOFTROPICALMEDICINEANDHYGIENE(2002)96,632-639

Deforestation and the spatio-temporal distribution of Savannah and forest members of the Simulium damnosum complex in southern Ghana and south- western Togo

M. D. Wilson’, R. A. Cheke**, S. P. J. masse *, S. Grist*, M. Y. Osei-Ateweneboana’, A. Tetteh-Kumah’, G. K. Fiasorgbor3, F. R. Jolliffe4, D. A. Boakye’, J.-M. Hougard3y6, L. Yameogo3 and R. J. Post’ ‘Noguchi Memorial Institute for Medical Research, University of Ghana, I? 0. Box 25, Legon, Ghana; ‘Natural Resources Institute, University of Greenwich, Central Avenue, Chatham Maritime, Chatham, Kent ME4 4TB, UK; 3 WHO Onchocerciasis Control Programme in West AjPica, B.P. 549, Ouagadougou, Burkina Faso; 4School of Computing and Mathematical Sciences, University of Greenwich, Maritime Greenwich University Campus, 30 Park Row, Greenwich, London SE10 9LS, UK; ‘Department of Entomology, The Natural History Museum, Cromwell Road, London SW7 5BD, UK; 61nstitut de Recherche pour le De’veloppement, 34032 Montpellier, Cedex 1, France

Abstract Spatio-temporal data on cytotaxonomic identifications of larvae of different members of the Simulium damnosum complex collected from rivers in southern Ghana and south-western Togo from 1975 until 1997 were analysed. When the data were combined, the percentages of Savannah blackflies (S. damnosum sensu strict0 and S. sirbanum) in the samples were shown to have been progressively increasing since 1975. The increases were statistically significant (P < O.OOl), but the rates of increase were not linear. Further analyses were conducted according to the collection seasons and locations of the samples, to account for possible biases such as Savannah flies occurring further south in the dry season or a preponderance of later samples from northern rivers having more Savannah flies. These analyses showed that the increasing trend was statistically significant (P < 0.0001) only during the periods April to June and October to December. The presence of adult Savannah flies carrying infective larvae (L3) indistinguishable from those of Onchocerca volvulus in the study zone was confirmed by examinations of captured flies. The percentages of Savannah flies amongst the human-biting populations and the percentages with L3s in the head were higher during dry seasons than wet seasons and the Savannah species were found furthest south (5”25’N) in the dry season. Comparisons of satellite images taken in 1973 and 1990 over a study area in south-western Ghana encompassing stretches of the Tano and Bia rivers demonstrated that there have been substantial increases in urban and Savannah areas, at the expense of forest. This was so not only for the whole images but also for subsamples of the images taken at 1, 2, 4, 8 and 16 km distant from sites alongside the River Tano. At every distance from the river, the percentages of pixels classified as urban or Savannah have increased in 1990 compared with 1973, while those classified as degraded or dense forest have decreased. The possibility that the proportionate increases in Savannah forms of the vectors of onchocerciasis, and hence in the likelihood of the transmission of Savannah strains of the disease in formerly forested areas, were related to the decreases in forest cover is discussed.

Keywords: onchocerciasis, Simulium damnosum cytospecies, transmission, forest, Savannah, deforestation, remote- sensing, Ghana, Togo

Introduction The symptoms and pathology of onchocerciasis dif-

fer between regions within West Africa. Patients from forested areas ;end to have fewer of the most severe ocular manifestations (blindness rates < 2%: WHO. 1987) than patients fro& the Savannah zones (blindness up to 15 %). Following a series of experiments in which it was shown that Savannah vectors preferentially transmitted the Savannah form of the disease, and forest vectors the forest form, DUKE et al. (1966, 1967) proposed the idea of disease-vector complexes. Sup- port for the existence of such complexes came with the demonstration that Simulium damnosum sensu lato, the vector in West Africa, is not a single taxonomic entity but a complex of sibling species (VAJIME & DUNBAR, 1975) and that each member of this complex has its own ecological characteristics. Some are forest-dwelling forms, while others prefer drier habitats or are adapted to exclusively Savannah areas.

The initial aim of the World Health Organization Onchocerciasis Control Programme in West Africa (OCP) was control of Savannah forms of the vector only, and hence only the more serious, blinding, savannah form of the disease. Thus. understanding the eui- demiology of onchocerciasis and hence de&ions bn control options for different areas, perhaps when bud- gets are limited, depends upon knowledge of the forms of both the parasite and the vector, or vectors, prevail- ing in a given area.

Thete is now evidence from various countries in

*Author for correspondence; phone +44 (0) 1634 880088, fax +44 (0) 1634 880066, e-mail [email protected]

Africa that the geographical distribution of Savannah forms of the vector is changing with increasing deforestation. The Savannah cytospecies S. sirbanum has been recorded further south fhan usual in Savannah enclaves within the forest belt in Liberia (GARMS. 1987; GARMS et al., 1991) and it penetrates further south than was previously known during dry seasons in Sierra Leone (BAKER et al., 1990; THOMSON et al., 1996). There is also evidence of S. sirbanum and S. dam&sum sensu stricto, both Savannah forms, appearing in previously forested parts of CBte d’Ivoire (BEATIN et al.. 19971 and Tono (CHEKE, 1993). Related nhenom- ena occui in eastern &ica &here s. damnosum s.1. has spread following deforestation in Malawi and Uganda (WALSH et al., 1993). However, in parts of Uganda, deforestation may have led to an amelioration of the disease with the decline of the principal vector S. neavei. This species is dependent on crabs upon which its immature- stages are phoretic, and it has been reolaced bv a less efficient vector. S. kilibanum, a member of-the S. damnosum complex (FISCHER et al., 1997; IQCJGER et al., 1999).

In this paper, we present evidence that Savannah forms of S. damnosum have been increasing since the 1970s in areas of southern Ghana and Togo subject to deforestation, and describe associated patterns of vegetation change based on analyses of remotely-sensed images. The use of satellite imagery has been widely used in studies of malaria, trypanosomiasis and other tropical diseases (HAY et al., 2000; ROGERS et al., 2002) but few studies have exploited the technology for investigations of onchocerciasis. One example has been an examination of the areas where onchocerciasis and

DISTRIBUTION OF SAVANNAH FLIES 633

loiasis are potentially in sympatry, with respect to adverse reactions associated with administration of the microfilaricidal drug ivermectin to onchocerciasis patients who also have loiasis (THOMSON et al., 2000).

Materials and Methods Study area

Southern Ghana has 2 main areas of forest. One is a large area in the south-west of the country that is drained by a series of sub-parallel river systems, which flow north-east to south-west directly into the Atlantic from an upland watershed, along a north-west to south-east line north of Sunyani - Kumasi - Koforidua (60 km due north of Accra) (Fig. 1). The uplands along the watershed consist of montane forest vegetation (AntiarisiChZorophora association). North of this is gui- nea Savannah, with rivers draining into the Volta river system which-are outside the area of this study. South of the highlands is moist deciduous forest (CeltislTriDlo- chiton association), which grades into tropical rain ior- est in the extreme south-west of the country. The second area of forest is in the Volta Region, -on the mountains east of the Volta Lake which lie along the Ghana-Togo international frontier. The vegetation there is montane forest (AntiarklChZorophora association), similar to that further west. All these zones are in various states of deforestation.

For the study of remotely-sensed images, a zone surrounding a section of the Tano and Bia rivers was chosen. The zone was defined by coordinates in its upper-left (7”33’58”N, 2”42’5”W), upper-right (7”28’ 9’N, 2”1’5”YQ, lower-left (6”42’57’YN, 2”49’52”W), and lower-right (6”35’51’%, 2”9’12”W) corners.

Cytotaxonomy of vectors The geographical distributions of the different cyto-

species of the S. damnosum complex occurring in southern Ghana and Togo were analysed using hata from larvae collected in rivers during 1995-97 and historical identifications. The latter were compiled from all pub- lished records for southern Ghana and Togo which included details of the numbers of larvae identified (VAJIME & DUNBAR, 1975; VAJIME & QUILL&&& 1978; MEREDITH et al., 1983, 1986; POST, 1986; CHEKE & DENKE, 1988; SURTEES et al., 1988; GARMS et al., 1989; VAJIME, 1989; CHEKE et al., 1992) and unpublished OCP cytotaxonomy reports from 1975 onwards written by C. G. Vajime, S. E. 0. Meredith, G. K. Fiasorgbor, D. A. Boakye, and A. Weber. Care was taken to avoid duplication of records from the 2 source categories. Cytotaxonomic data (sample sizes in brackets) were obtained from larvae collected in the following rivers: Bia (154), Tano (3708), Ankobra (27), Pra (3147), Offin (958), Dayi (3672), Wulubong (16), Asuakawkaw (6016), Todzie (548), Amou (2240), and Pawnpawn (800). Nomenclature followed that of VAJIME & DUNBAR (1975), as amended or added to by MEREDITH et al. (1983), POST (1986), and SURTEES et al. (1988). The classification of the S. sanctipauli complex was altered after the date when the first records were obtained and so all records were adapted to fit the revised scheme described by POST (1986), whenever possible. In practice this meant that most old records of S. soubrense were reassigned to S. sanctipauli, except where the Beffa form of S. soubrense was involved. However, as both S. sanctipauli and S. soubrense were classified as forest cytospecies in the ensuing analysis, this taxonomic complication has no material bearing on the results described and discussed.

Statistical analyses of cytotaxonomic data To test if any apparent trends in the combined data

of increasing percentages of Savannah flies with increasing time were significant, a test for a trend in proportions (A~ITAGE, 1955; MANTEL. 1963) was used. As the nimbers of _ blackfly varied *with &the season, a

separate test for trend was conducted on data from each of the 4 quarters January to March, April to June, July to September, and October to December. For each quarter, the data were first classified into 4 periods: 1975-80, 1981-86, 1987-92, and 1993-97. In the event of a significant trend in proportions, a further test on whether the trend was linear was performed.

To test for trends taking account of individual areas, a method described by DARBY & REISSLAND (1981) and ARMITAGE & BERRY (1994) was adopted. This uses the data divided into different seasons and periods, as before, but treats each river or section of a river as a separate stratum. Strata where there were no Savannah blackflies recorded at all, or where blackfly were observed within one period only, were ignored as they added no information about the trends. Details of all test statistics used are given in the appendix.

Captures and dissections of adultjlies During 1996 and 1997 captures of adult female S.

damnosum coming to bite humans were made at sites along the River- Tano’s upper stretches (Tanoso, 7”16’N. 2”15’W: Ntotoroso. 7”4’N. 2”19’W1. middle reaches’ (Kpoikrom, 6”13’N,- 2”28’w; Nsawa& 6”9’N, 2”36’W) and lower reaches (Jema, 5”25’N, 2”41’W); along 2 northern tributaries of the River Pra beside the River Ofin (Bontomuroso, 6”34’N, 2”4’W; Anniamoa, 6”33’N, 2”2’W; Pokukrom 6”l’N, l”49’W; Assin-Asa- man 5”54’N, l”3O’w) and beside the River Anum (Gyankobaa, 6”19’N, l”16’W); and at 2 sites on the lower River Pra (Twifo-Praso, 5”37’N, l”33’W; Se- kyere-Heman, 5”l l’N, l”35’W). The flies were caught according to standardized procedures (WALSH et al., 1978) during dry seasons in February to early May and wet seasons in late May to August. Subsamples of the catches were examined morphologically and members ofthe Savannah group (S. damnosum s.str. or S. sirbanum) were separated according to criteria described by GARMS & CHEKE (1985) and WILSON et al. (1994) and dissected when ‘still fresh to search for developing stages of Onchocerca volvulus.

Land use mapping with remote sensing The definition of land cover change between the

1970s and the 1990s was obtained by analysing and comparing 2 Landsat MSS (Multispectral Scanner) satellite images, incorporating the study area. One was recorded over Ghana on 25 November 1973 and the other on 31 December 1990. Images from the dry season were chosen deliberately to miminize the likelihood of areas being affected by cloud cover and to maximize variations in vegetation cover (during the wet season most vegetation shows up as intense chlorophyll activity, making discrimination between different types of ecosystem difficult from a single satellite image). Each scene was available in the following spectral bands: 0.5-0.6 ,LIXI (green); 0.6-0.7 p (red); 0.7- 0.8 urn (near infrared NIRl): and 0.8-1.1 um (NIR). wi& a spatial resolutidn of ‘& m. To ensurk a change map with minimum interference from ‘noise’, care was taken in coregistering the images, as well as applying a cloud mask by removing individual clouds, their sha- dows, and areas with atmospheric disturbances (through multithresholding, buffering and visual tech- niques) .

Land cover change is commonly assessed in Landsat imagery by visual interpretation and classification tech- niques, leading to reliable assessments of change, using minimum field validation (e.g. SKOLE & TUCKER, 1993; ALVES & SKOLE, 1996). Consequently, supervised classification (maximum likelihood method) combined with visual interpretation was used to classify the images into 4 land cover types: (i) urban (including areas of bare ground/soil/dry river beds); (ii) Savannah; (iii) degraded/secondary forest; and (iv) dense/primary forest. The authors used their extensive field know-

634

ledge, photographs and land use maps to define the training sites for the supervised classification, as well as ensuring sensible classification results. Dense forest in the study area is mostly confined to well-delimited protected areas, making visual interpretation appropri- ate. Particular care was taken to ensure consistency of the land cover classes between the 2 dates, although no comprehensive validation of the map was completed.

A change map was derived to indicate the changes in land cover between 1973 and 1990. This map revealed that 10% of the transitions were most unlikely, invol- ving shifts towards denser vegetation (e.g. Savannah to dense forest), probably consequences of some impreci- sion in the classification. Although some of these vegetation changes could have been genuine, e.g. the growth of sacred groves around new villages (FAIR- HEAD & LEACH, 1998), they were all eliminated from the final analysis (vegetation mask).

It is not clear whether what most influences a blackfly to select a breeding site is the habitat immediately surrounding a stretch of rapids in a suitable river or the surrounding environmental conditions. To see how habitats (classified by remotely-sensed images) varied in relation to distances from a river, 4 sites were selected at intervals along the River Tano. These were selected at random, but locations were restricted to those where a circle with a radius of at least 16 km could be sampled from around the site. Therefore the sites could not be too near to the edges of the image. The sample areas were extracted at intervals of 1, 2, 4, 8, and 16 km. The central locations of the 4 sites selected were as follows: site 1, 7”02’48”N, 2”19’09”W; site 2, 6”55’49”N, 2”19’43”W; site 3, 6”52’26”N, 2”22’31”W; and site 4, 6”48’22”N, 2”25’41”W. Although the images were identical in geographical size, the number of pixels at each location varied according to how many pixels had been eliminated by either the cloud mask or the vegetation mask. The most southern and northern sites (sites 1 and 4) overlapped the edges of the image fractionally in the 16 km samples, and in addition the southern site encompassed a high degree of cloud mask, which reduced the total number of pixels in the sample. At each location, data from within the circular areas were extracted from the 1973 classification, the 1990 classification and the change map.

Results Cytotaxonomy

Data for the River Tano were separated into those for the upper reaches (7”0’-7”28’N), middle (6”0’- 6”59’N) or lower sections (5”23’-5”59’N) and those for the River Pra were divided into data from upper reaches (5”52’-6”24’N) and the lower sections (south of 5”52’N). The percentages of larvae classified as Savannah (S. damnosum s.str. and 5’. sirbanum) or forest (S. sanctipauli, S. soubrense, S. yahense, and S. squamosum) recorded in different rivers, or parts of rivers as above, in different quarters (January-March, April- June, July-September, and October-December) according to period (1975-80, 1981-86, 1987-92, and 1993-97) were collated. Full details of the identifications are available from R. A. C. or R. J. P.

Using all the data from different rivers combined (Table l), there is a highly significant trend in the proportions in each of the 4 quarters (P < 0.001). Inspection of the data showed that the percentages of Savannah blackfly have increased with time, but this trend, although significant, is not linear in any quarter. Thus, overall, the percentages of Savannah blackflies have been progressively increasing since 1975 in southern Ghana and south-western Togo but at rates which are not linear. However, the results could have been biased by a preponderance of later samples from northern rivers where more Savannah flies might be expected, for instance, so a further test to take account of the

M. D. WILSON ETAL.


spatial and temporal heterogeneity of the different rivers sampled was conducted. This revealed that, after allowing for location, the trends were highly significant (P < 0.0001) in both the April to June and October to December quarters, but not during the January to March period (P= 0.82), with the July to September period also non-significant but less so (P = 0.075). So, although the combined data set showed that there has been a significant increase of Savannah flies since 1975 in all seasons, further analysis taking account of location of sampling has shown that this trend is significant only during the April to June and October to December quarters.

Identification of adults and infections in savannahflies during 1996-97

During the wet seasons, Savannah flies were recorded at rates of 5.58.5% of the fly catches along the upper and middle sections of the River Tano and in the upper River Pra zone (13.2%), but not at the most southern sites (Table 2). Infective Savannah flies were found in the upper River Pra zone at Gyankobaa during July and August but not elsewhere in the wet season. During the dry seasons, Savannah flies were recorded in all the study zones, comprising 0.2-68.4% of the biting populations (Table 2), and reached as far south as the lower River Tano at Jema and the lower River Pra at Twifo- Praso. Although infection rates in the Savannah flies were also low, 15 were identified with infective larvae (L3) indistinguishable from those of 0. VOZVUZUS in their heads.

Vegetation cover change between 1973 and 1990 Inspection of maps of vegetation cover for 1973 and

1990 (Fig. 2) clearly suggests that there were increases in the extent of urban areas and of Savannah, and decreases in dense forest and degraded forest between 1973 and 1990. This was confirmed by the numbers of pixels classified into the 4 groups in the 2 images (Table 3) indicating a 50% reduction in those classified as dense forest and a quadrupling in the extent of Savannah areas. While the total area of degraded forest was only slightly reduced (4.47%), detailed analysis of the changes in land cover (Table 4) indicated that nearly 26% of the degraded forest changed to savannah, whilst the Savannah areas lost 22% of their extent to urban cover, resulting in the net increase of Savannah of 16.6% (Table 3).

Habitat variation at diSferent distances from the riverside It is clear from the data in Table 5 that the trends of

increasing urbanization and Savannah encroachment at the expense of degraded and dense forest close to and at increasing distances up to 16 km from the river are consistent with the patterns recorded in the large scenes. At every distance from the river, the percentages of pixels classified as urban or Savannah have increased in 1990 compared with 1973, while those classified as degraded or dense forest have decreased. In each of the years, both the highest percentage of urban pixels and the lowest percentage of dense forest pixels were recorded closest to the rivers, the former decreasing and the latter increasing with increasing distance from the river’s edge. It is therefore difficult to conclude anything about the determinants of habitat choice by blackflies, which might have been possible had, for instance, deforestation only been occurring at the riversides.

Discussion The interpretations of the remotely-sensed images

clearly demonstrated that between 1973 and 1990 there was a substantial increase in urban and Savannah areas, at the expense of forest, in a zone of Ghana where onchocerciasis transmission is endemic. Coin- ciding with this deforestation has been an increase in

636 M. D. WILSON ETAL.

Fig. 1. Maps of Ghana and the study area.

Urban

Savannah

Degraded Forest

Dense Forest

Fig. 2. A remotely-sensed image of the study area in Ghana on 25 November 1973 following a supervised classification according to urban, Savannah, degraded forest, dense forest or unclassifiable pixels (left) and a similar image of the same area on 3 1 December 1990 (right).

the proportion of the Savannah forms of onchocerciasis vectors in southern Ghana and Togo according to analyses of the available cytotaxonomic data. When all the latter data are combined, the trend of increasing proportions of Savannah flies was statistically significant. More detailed analysis taking account of spatial and temporal variation, however, only demonstrated a statistically significant trend for the April to June and

October to December quarters. Nevertheless, this pro- vides firm evidence that Savannah members of the S. damnosum species complex have been occurring at higher levels in the south of Ghana and Togo than they did at the outset in 1974 of the vector control opera- tions mounted by the OCP. Although the causation is not demonstrated, the likelihood is that these incur- sions by S. damnosum s.str. and by S. sirbanum are


Table 3. Land cover classification of study sites in southern Ghana in terms of numbers of pixels in different categories in the 1973 and 1990 images and percentage changes for each category

Cover type

1973

No. pixels %

1990 Percentage

No. pixels % change

Urban 8134 1.01 62931 7.82 +6.81 Savannah 40046 4.97 173681 21.57 +16.60 Degraded forest 451306 56.06 415322 51.59 -4.47 Dense forest 305563 37.96 153115 19.02 -18.94 Total 805049 805049

Table 4. Change map statistics of study sites in southern Ghana for 1973 to 1990 land cover classification with percentages of pixels in each transition category

Land cover change from 1973 to 1990

1973 1990 % of 1973

No. pixels % Total classification

Unchanged urban 8134 Savannah urban 8809 Unchanged Savannah 31237 Degraded forest to urban 37195 Degraded forest to Savannah 117214 Unchanged degraded forest 296897 Dense forest to urban 8793 Dense forest to Savannah 25230 Dense forest to degraded forec it 118425 Unchanged dense forest 153115 Total 805049

1.01 100~00 1.09 22.00 3.88 78.00 4.62 8.24

14.56 25.97 36.88 65.79

1.09 2.88 3.13 8.26

14.71 38.76 19.02 50.11

Table 5. Numbers of pixels classified as urban, Savannah, degraded forest or dense forest in areas of radius 1, 2, 4, 8 and 16 km away from the River Tano, south-western Ghana in 1973 and 1990 and the percentage changes in the totals between 1973 and 1990

1973 1990

Habitat No. of pixels” % No. of pixels” Percentage

% change

1 !un from river Urban Savannah Degraded forest Dense forest Total

2 km from river Urban Savannah Degraded forest Dense forest Total




296 314

1871 439

2920

10.1 10.8 64.1 15.0

1154 556

1209 2

2921

39.5 +29.4 19.0 + 8.2 41.4 -22.7

0.1 -14.9

587 1004 7822 2196

11609

5.1 8.6

67.4 18.9

3330 2663 5443

173 11609

28.7 +23.6 22.9 +14.3 46.9 -20.5

1.5 -17.4

1421 3290

30688 10464 45863

3.1 7.2

66.9 22.8

9421 10848 24054

1540 45863

20.5 +17.4 23.7 +16.5 52.4 -14.5

3.4 -19.4

3373 11211

127344 39430

181358

1.9 6.2

70.2 21.7

28807 42863

101238 8450

181358

15.9 +14.0 23.6 +17*4 55.8 -14.4

4.7 -17.0

10323 44184

471817 164245 690569

1.5 6.4

68.3 23.8

79294 170715 396433

44127 690569

11.5 +10.0 24.7 +18.3 57.4 -10.9

6.4 -17.4

“Total from 4 different sites in each category

638 M.D.WILSON ETAL.

related to the deforestation which permits them to occupy breeding sites which were previously inimical to them. Similar chances in the distribution of other disease vectors, tsetse flies Glossina spp., in relation to habitat changes have been recorded in Burkina Faso (DE LA ROCQUE et al., 2001). Encroachment of agricul- tural fields in riverine areas has led to declines in tsetse flies, whereas increased human activity and higher densities of cattle have led to fly increases.

If the observed changes in the spatio-temporal distributions of the larvae of Savannah blackflies had been noticed during the time when the OCP’s activities relied on vector control, then it would have been necessary to treat rivers in Ghana and Togo much further south than was the routine. This point was emphasized by the finding of Savannah flies with infective larvae reaching as far south as S’25’N in 1996-97. Now, when much onchocerciasis control relies on distribution of the microtilaricidal drug ivermectin, the encroachment of Savannah vectors further south, pre- sumably carrying the more virulent form of the disease, is of less importance operationally but it does have important epidemiological implications. If onchocerciasis control breaks down, then the more blinding form of the disease will be prevalent in areas where previously only the more benign forest form was being transmitted. Consideration of the presence of Savannah flies in southern regions may also impinge on interpretations of current epidemiological patterns. Further- more, the results have wider implications if a link between deforestation and increases in Savannah flies is accepted. The study warns of the potential folly involved in a lack of stewardship for the protection of the environment as this could have serious consequences for public health as well as for environmental conserva- tion.

Acknowledgements We are grateful for suunort from the WHO Special Pro-

gramme fo; Research and *Training in Tropical D&ease Task force on Diseases and Environment (project nos. 940723 and 960456) and from the WHO Onchocerciasis Control Pro- gramme in West Africa.

References Alves, D. S. & Skole, L. L. (1996). Characterising land cover

dynamics using multi-temporal imagery. International Jour- nal ofRemote Sensing, 17, 835-839.

Armitage, I’. (1955). Tests for linear trends in proportions and frequencies. Biometrics, 11, 375-386.

Armitage, P. & Berry, G. (1994). Statistical Methods in Medical Research, 3rd edition. Oxford: Blackwell Science.

Baker, R. H. A., Guillet, I’., Seketeli, A., Poudiougo, I’., Boakye, D., Wilson, M. D. & Bissan, Y. (1990). Progress in controlling the reinvasion of windborne vectors into the western area of the Onchocerciasis Control Programme in West Africa. Philosophical Transactions of the Royal Society of London, Series B, 328,731-750.

Boatin, B., Molyneux, D. H., Hougard, J. M., Christensen, 0. W., Alley, E. S., Yameogo, L., Seketeli, A. & Dadzie, K. Y. (1997). Patterns of epidemiology and control of onchocerciasis in West Africa. Journal of Helminthology, 71, 9 1 - 10 1.

Cheke, R. A. (1993). Invasions by savanna Simulium damnosum s.1. into forest habitats and by forest forms into savannas in West Africa. British Simuliid Group Bulletin, 2, 9.

Cheke, R. A. & Denke, A. M. (1988). Anthropophily, zoophily and roles in onchocerciasis transmission of the Djodji form of Simulium sanctipauli and S. squamosum in a forest zone of Togo. Tropical Medicine and Parasitology, 39, 123- 127.

Cheke, R. A., Sowah, S. A., Avissey, H. S. K., Fiasorgbor, G. K. & Garms. R. f 1992). Seasonal variation in onchocerciasis transmission by &m&urn squamosum at perennial breeding sites in Togo. Transactions of the Royal Society of Tropical Medicine and Hygiene, 86, 67-71.

Darby, S. C. & Reissland, J. A. (1981). Low levels of ionizing radiation and cancer-are we underestimating the risk? Journal of the Royal Statistical Society, A, 144,298-331.

de la Rocque, S., Augusseau, X., Guillobez, S., Michel, V., De Wispelaere, G., Bauer, B. & Cuisance, D. (2001). The

changing distribution of two riverine tsetse flies over 15 years in an increasingly cultivated area of Burkina Faso. Bulletin ofEntomological Research, 91, 157-166.

Duke, B. 0. L., Lewis, D. J. & Moore, P. J. (1966). Onchocer- ca-Simulium complexes. I. Transmission of forest and Su- dan-Savannah strains of Onchocerca volvulus, from Cameroon, by Simulium damnosum from various West Afri- can bioclimatic zones. Annals of Tropical Medicine and Para- sitology, 60, 318-336.

Duke. B. 0. L.. Moore. P. T. & DeLeon. T. R. (1967). Onchocerca-Simulium complexes. Annals of Tropical Medicine and Parasitology, 61,332-337.

Fairhead, J. & Leach, M. (1998). Reframing Deforestation. London: Routledge.

Fischer, P., Garms, R., Buttner, D. W., Kipp, W., Bamuhiiga, J. & Yocha, J. (1997). Reduced prevalence of onchocerciasis in Uganda following either deforestation or vector control with DDT. East African Medical 7ournal. 74. 321-325.

Garms, R. (1987). &currence of-me savannah species of the Simulium damnosum complex in Liberia. Transactions of the Royal Society of Tropical Medicine and Hygiene, 81, 5 18.

Garms, R. & Cheke, R. A. (1985). Infections with Onchocerca volvulus in different members of the Simulium damnosum complex in Togo and Benin. Zeitschrijt fiir Angewandte Zoologie, 72,479-495.

Garms, R., Cheke, R. A., Fiasorgbor, G. K. & Walsh, J. F. (1989). Seasonal extension of the breeding range of Simu- hum sanctipauli from forest into savanna in eastern Ghana and Togo. Zeitschrtftfir Angewandte Zoologie, 76,457-467.

Garms, R., Cheke, R. A. & Sachs, R. (1991). A temporary focus of savanna species of the Simulium damnosum complex in the forest zone of Liberia. Tromcal Medicine and Parasitol- ogy, 42,181-187.

Hay, S. I., Randolph, S. E. & Rogers, D. J. (editors) (2000). Remote Sensing and Geographical Information Systems in Epidemiology. Advances in Parasitology, 47, special issue.

Kruger, A., Nurmi, V., Yocha, J., Kipp, W., Rubaale, T. & Garms, R. (1999). The Simulium dumnosum complex in western Uganda and its role as a vector of Onchocerca volvulus. Tropical Medicine and International Health, 4: 819-826. -

Mantel, N. (1963). Chi-square tests with one degree of free- dom: extensions of the Mantel-Haenszel procedure. Journal of the American Statistical Association, 58, 690-700.

Meredith, S. E. O., Cheke, R. A. & Garms, R. (1983). Variation and distribution of forms of Simulium soubrense and S. sanctipauli in West Africa. Annals of Tropical Medicine and Parasitology, 77, 627-640.

Meredith, S. E. O., Kurtak, D. & Adiamah, J. H. (1986). Following movements of resistant populations of Simulium soubrenseisanctipauli (Diptera: Simuliidae) by means of chromosome inversions. Tropical Medicine and I’arasitologv, 37,290-294.

Post, R. J. (1986). The cytotaxonomy of Simulium sanctipauli and Simulium soubrense (Diotera: Simuliidae). Genetica, 69. . _ 191-207.

Rogers, D. J., Randolph, S. E., Snow, R. W. & Hay, S. I. (2002). Satellite imagery in the study and forecast of malaria. Nature, 415, 710-715.

Skole. D. L. & Tucker. C. 1. (1993). Tropical deforestation and habitat fragmentation in the Amazon-satellite data from 1978 to 1988. Science, 260, 19051910.

Surtees, D. I’., Fiasorgbor, G., Post, R. J. & Weber, A. (1988). The cytotaxonomy of the Djodji form of Simulium sanctipauli. Tropical Medicine and Parasitology, 39, 120- 122.

Thomson, M. C., Davies, J. B., Post, R. J., Bockarie, M. J., Beech-Garwood, P. A. & Kandeh, J. (1996). The unusual occurrence of Savannah members of the Simulium damnosum species complex (Diptera: Simuliidae) in southern Sierra Leone in 1988. Bulletin of Entomoloeical Research, 86. 271-280.

Thomson, M. C., Obsomer, V., Dunne, M., Connor, S. J. & Molvneux, D. H. (2000). Satellite mapping of Lou loa prevalence in relation to ivermectin use inwest and central Africa. Lancet, 356, 1077-1078.

Vajime, C. G. (1989). Cytotaxonomy of Sirba form populations of the Simulium damnosum comvlex in West Africa. amendments to sex chromosomes and-sibling status. Tropi- cal Medicine and Parasitology, 40, 464-467.

Vajime, C. G. & Dunbar, R. I. (1975). Chromosomal identiti- cation of eight species of the subgenus Edwardsellum near and including Simulium (Edwardsellum) damnosum Theobald (Diptera: Simuliidae). Tropenmedizin und Parasitologic, 26, 111-138.

Vajime, C. G. & Quill&&e, D. (1978). The distribution of the


Simulium dumnosum complex in West Africa with particular reference to the Onchocerciasis Control Programme Area. Tropenmedizin und Parasitologic, 29, 473-482.

Walsh, J. F., Davies, J. B., Le Berre, R. & Garms, R. (1978). Standardization of criteria for assessing the effect of Simu- lium control in onchocerciasis control programmes. Transac- tions of the Royal Society of Tropical Medicine and Hygiene, 72, 675-676 [correspondence].

Walsh. 1. F., Molvneux. D. H. & Birlev. M. H. 11993). Defbr&tatibn: effects on vector-borne disease. Par&itolo& 106, S55-S75.

Appendix. Details of test statistics applied to data

Notation ni = total number of blackfly in the ith time period; N = Bni = total number of blackfly (forest and savan-

nah); ri = total number of Savannah blackfly in the ith time

period; R = Xri = total number of Savannah blackfly; xi is a score representing a time period (x1 = 1 for

1975-80, x2 = 2 for 1981-86, x3 = 3 for 1987-92, x4 = 3.9 for 1993-97).

Test for a trend in proportions The test statistic used was:

The calculated value of the statistic x2 is compared with a x2 distribution with 1 d.f.

Test for linear trend Consider classifying the numbers of blackfly by time

period and whether or not they are Savannah blackfly as in the 2-way contingency table below.

1975-80 1981-86 1987-92 1993-97

ri

WHO (1987). WHO Expert Committee on Onchocerciasis, Third Report. Geneva: World Health Organization.

Wilson, M. D., Mafuyai, H. B. & Post, R. J. (1994). Morpho- logical identification of sibling species of the Simulium damnosum (Diptera: Simuliidae) complex from Nigeria, Cameroun and Bioko. Proceedings of the Section Experimental and Applied Entomology of the Netherlands Entomological So- ciety (N.E.V.), 5, 181-185.

Received 16 November 2001; revised 20 March 2002; accepted for publication 27 March 2002

Let xi be the x2 statistic for testing the null hypoth- esis that there is no association between time period and whether or not the blackfly are Savannah. The difference between xi and the statistic for a trend in proportions gives a test for linear trend. In the application in this paper the difference is compared with a x2 distribution with 2 d.f.

Test for trend taking account of individual areas For each stratum (river or section of a river) k, a

statistic T, and its variance Var(&) were calculated as follows:

Tb = 2 xi(ri - e,) = 2 xiri - i=l i=l

where e, = E(rJ = (Rni)/N

,lN

where, as in T,, all sums are over the 4 time periods. Strata where every r, = 0 or where only one n, was

not zero were ignored (see text). The statistic (C T,)‘/C var(T,) is compared with a

x2 distribution with 1 d.f. Here the sums are over the different strata within each quarter.

Announcement

ROYAL SOCIETY OF TROPICAL MEDICINE AND HYGIENE Denis Burkitt Fellowships

The Denis Burkitt Fund was set up by his family in memory of Denis Burkitt, FRS, who died in 1993; it is administered by the Royal Society of Tropical Medicine and Hygiene.

One Fellowship (maximum value E7000) or two separate Fellowships (of &3500 each) are awarded annually for practical training, travel, or direct assistance with a specific project (preferably clinico-pathological, geographical or epidemiological studies of non-communicable diseases in Africa).

Applications must be made at least six months before the commencement of the proposed study (by 15 March in each year). A short report on the study should be submitted, within 3 months of the recipient’s return.

Application forms are available from the Administrator, Royal Society of Tropical Medicine and Hygiene, Manson House, 26 Portland Place, London, WlB lEY, UK; fax +44 (0)20 7436 1389, e-mail [email protected]

deforestation and the spatio-temporal distribution of savannah and forest members of the simulium...

Documents