Insights Into Land Use/Cover And Human Activities Pattern For Explanation Of Plague Infection Risks In Western Usambara Mountains, Tanzania

ABSTRACT

Land cover, land use, and human activities pattern have been reported to be important determinants of vector borne diseases transmission including plague. Plague, still occurring in different parts of the world, has been a threat in the Eastern Africa region including the Western Usambara Mountains, Tanzania. Plague is a severe, rodent associated, bacterial zoonosis caused by Yersinia pestis. Literature suggests that factors influencing the critical contact between rodent hosts, flea vectors, and humans as well as human behaviours that may enhance or diminish this contact are not well understood in many areas particularly in the Eastern Africa region. Hence, studies that link a complete geographic perspective including land use and land cover, host, vector and human activities pattern dimensions are important. Understanding the influence of the landscape factors on small mammals like rodents and flea abundance and their spatial distribution as well as the human exposure risks is vital and can assist in formulating prevention and surveillance mechanisms.

Studies carried out in East Africa and elsewhere report a wide variety of potential health impacts arising from land use and land cover and terrain factors most of which are not well studied in the Western Usambara Mountains. Hence, the current research aimed to contribute to efforts of giving an insight into the roles of land cover, land use and human activities pattern in plague infection risks in the Western Usambara Mountains, Lushoto District, Tanzania. Specifically the study aimed: (i) to map land cover and terrain attributes and determine their association with plague hosts and vectors at landscape level (ii) to identify land use and land management practices associated with abundance and distribution of plague hosts and vectors at farm level (iii) to model people’s movements and activities pattern in order to determine chances of their exposure to plague in space and time.

The study was carried out in three selected landscapes in the Western Usambara Mountains in Lushoto District, Tanzania. Study sites were selected to reflect a geographic gradient in plague incidence for the period 1980-2004, based on results from previous research on rodents, fleas and plague casualties in the area. The findings from those studies concur in distinguishing high, medium, and low incidence zones. Within this gradient three representative landscapes were selected differing in terms of (i) the incidence of plague, (ii) diversity in land use and associated human activities, (iii) landform characteristics, and (iv) climatic conditions. The selected landscapes are named Shume (high plague incidence), Lukozi (medium plague incidence) and Mwangoi (low plague incidence). In the context of the current study, Landscape is defined as “an area of land covering two to three villages with a repetition of similar relief types or an association of dissimilar relief types (valleys, plateaus, mountaineous hilly relief types), more or less homogeneous land use/cover types (natural forest, cultivated land, plantation forest and built up areas) and having specific historical plague incidence rate”.

Twenty four observation sites (quadrats) of 100 x 100 m were established per sample landscape area. A Stratified random sampling procedure based on broad land cover types and topography was used to locate the observation sites in each sample landscape area. Decision on the number of observation sites considered representative sample size, time and human resources availability. Data collection was done in the wet season (April-June 2012) and the dry season (August-October 2012). The study used a geospatial approach to examine the influence of land cover/use and terrain factors on the abundance and spatial distribution of plague hosts (small mammals) and plague vectors (fleas). The approach included use of remote sensing and Geographic Information System (GIS), field observational survey and household questionnaire survey for mapping land cover/use, human movements and activities pattern, and, trapping of small mammals. Isolation and counting of fleas was done from the trapped rodents. During field survey, various visible indicators of land use were mapped and quantified within the quadrats. Two major categories of land use were defined: (a). Land management practices e.g. terrace, (b). Crop types and their associated elements e.g. maize farming. These land uses are named as ‘individual land use types’ in the current study. Each quadrat was also classified based on its aggregated land uses. For example whenever the quadrat was dominated by both annual and perennial crops, the aggregated land use type for that particular quadrat was classified as ‘Mixed annual perennial crops’.

Data analysis on land cover/use and terrain factors was done using remote sensing image processing tools and GIS. The approach used to determine the human activity spaces was kernel density estimation. Interpolation and zonal statistics GIS analyses were used to determine human-flea co-occurrence. Analysis of Variance (ANOVA) was used to evaluate differences in the data (Aggregated land use types, small mammal and flea abundance, human-flea co-occurrence) whenever the data passed normality and homogeneity tests and in case of non-success the non-parametric ANOVA on Medians (Mood’s Median test) was used. The Boosted Regression Tree (BRT) modeling technique was used to clarify the relationships between the individual land use types, land cover types, and terrain attributes, with small mammal and flea abundance.

Results indicate that elevation positively influenced the presence of small mammals (plague hosts). This could be attributed to the increased resource availability (water and food) as one moves from low to high altitude on the landscape. The presence of fleas (plague vectors) was clearly influenced by land management features such as miraba which tended to increase in intensity with increase in slope gradient. Miraba is an indigenous land management practice with grass strips surrounding crop fields in a rectangular shape in the Western Usambara Mountains, Tanzania (a unique indigenous soil erosion control practice in the Usambara Mountains). Medium to high resolution remotely sensed data and field collected data integrated in GIS have been found to be quite useful in studying plague infection risks. These findings contribute to efforts on plague surveillance and awareness creation among communities on the probable risks associated with various landscape factors during epidemics. The identified land cover/use and terrain characteristics integrated in the expert GIS engine provide future potential analysis and understanding of the association of plague risk indicators including human behaviour variables at farm scale. The results also show that there was a significant variation (p ≤ 0.05) of small mammal abundance among land use types. Plantation forest with crop farming, natural forest and fallow had higher populations of small mammals than the other aggregated land use types. Plantation forest with crop farming, and fallow which is mainly surrounded by agricultural fields, offer conducive environment for small mammals in terms of food and shelter. Natural forest also provides food, water and shelter for small mammals. Shelter and food are important factors in breeding, recruitment and survival of rodents. Both miraba and fallow tended to favour small mammals’ habitation whereas land tillage practices had the opposite effect in dry season. Tillage of land could have resulted in the destruction of rodents burrows and mounds, destruction of nest sites, alteration of microclimate, and removal of vegetation some of which comprise food sources and shelter for the rodents. In addition, during the wet season crop types such as potato and maize appeared to positively influence the distribution and abundance of small mammals which was attributed to both shelter and food availability.

A significant variation (p ≤ 0.05) of flea indices in different land use types was also identified. Fallow and natural forest had higher flea indices whereas plantation forest mono-crop and mixed annual crops had the lowest flea indices among the aggregated land use types. The observed variations of flea index among aggregated land use types could be attributed to the impact of land use practices on flea habitat structure. Fallow structure, which in most cases is also surrounded by agricultural fields, provides conducive microclimate for fleas on one hand and a supply of both food and shelter for rodents on the other. The influence of individual land use types on flea indices was variable with fallow having a positive effect and land tillage showing a negative effect. Fallow fields have also been associated with plague cases in many countries including Uganda and hence findings from this study further lend credence to the hypothesis that plague infection risk could be associated with fallow in the study area. This is because previous studies showed that flea index could be used as an indicator of plague infection risk. Tillage of land which destroys surface and subsurface microclimate could be detrimental to flea survival. This observation is of practical significance with regard to the need of clearing surroundings of homesteads and avoiding long fallow cycles.

The results also demonstrated a seasonal effect, part of which could be attributed to different land use practices such as application of pesticides. These findings suggest that land use factors have a major influence on rodent flea abundance which could be taken as a proxy for plague infection risk. The results further point to the need for a comprehensive package that includes land tillage and crop type considerations on one hand and the associated human activities on the other, in planning and implementation of plague control interventions.

The results indicate further that, the degree of spatial co-occurrence of potential plague vectors (fleas) and humans in Lushoto focus differs significantly (p ≤ 0.05) among the selected landscapes for both dry and wet seasons. For the dry season, the Mood’s Median test indicated that Shume had the highest median average flea index (Median = 0.983) followed by Lukozi (Median = 0.575) and Mwangoi (Median = 0.380). For the wet season, the ANOVA means also followed the gradient of plague incidence rates i.e. 0.54 for Shume, 0.50 for Lukozi and 0.24 for Mwangoi. The study suggests that plague surveillance and control programmes at landscape scale should consider the existence of plague vector contagion risk gradient from high to low incidence landscapes due to human presence and intensity of activities. The current study has demonstrated the importance of land use/cover and human activity spaces in the study of plague infection risks. Based on the findings from the current study the following conclusions can be drawn:

i. The relationship between land cover and terrain attributes on one hand and small mammals and fleas as potential hosts and vectors of plague, has been well elaborated by remote sensing and GIS integration of geodatabase at different spatial scales and resolutions. Hence a geomatic approach using remote sensing data and GIS technologies is valuable in studying plague infection risks.

ii. Small mammals and fleas abundance and distribution is influenced by the specific land use and land management types namely Fallow, Miraba, Tillage, Plantation forest with farming, Natural forest and Woodlot. Tillage has a negative influence whereas the other five land use and management types have positive influence.

iii. Small mammal presence in different land use types can influence abundance of fleas. These findings therefore, make a significant contribution towards efforts in the control of plague risk factors in space and time.

iv. Spatial co-occurrence of a potential disease vector and humans differs significantly among the plague incidence landscape areas and follows the established plague incidence gradient of high, medium and low for both dry and wet seasons. This trend gives a coarse indication of the possible

ix

association of the plague outbreaks and the human frequencies of contacting environments with fleas.

v. The findings from this study are of public health relevance because they may guide plague surveillance, prevention and control programmes at fine scales by providing information to health workers to focus control measures on land use/cover and landscape units with high concentration of rodent fleas, especially during epizootic periods.

vi. Systematic trapping of small mammals and collection of rodent fleas for surveillance should target miraba, fallow land, plantation forest with farming, natural forest and woodlot.

The following recommendations are made in the light of gaps revealed from the findings of this study so as to provide further insights into the plague disease.

i. Land management practices including tillage of land and crop types and the associated human activities should be included in the general scheme of plague control and management.

ii. Future efforts to predict and map spatial and temporal human plague infection risk at farm scale should consider the role played by land use on small mammals and rodent fleas abundance and distribution.

x

iii. The study suggests that plague surveillance and control programmes at landscape scale should consider the existence of plague vector contagion risk gradient from high to low incidence landscapes due to human presence and intensity of activities.

iv. Further studies should be conducted to investigate how land use practices influence surface and subsurface microclimate conditions of various small mammals and flea species

v. Outdoor application of insecticides to control flea abundance has been found to be an effective measure against plague. However, further studies on timing of applications during epizootics vis-à-vis crop type should be considered