Irrigation and the risk of Rift Valley fever transmission: A
case study from Kenya Bernard Bett, International Livestock Research Institute
One Health for the Real World: Zoonoses, Ecosystems and WellbeingLondon, UK
17–18 March 2016
Acknowledgements
Mohammed Said1, Rosemary Sang2, Salome Bukachi3, Johanna Lindahl1, Salome Wanyoike4, Ian Njeru5, Delia Grace1
1. International Livestock Research Institute, Nairobi2. Kenya Medical Research Institute, Mbagathi Way, Nairobi3. Institute of Anthropology, Gender and African Studies, Nairobi4. Department of Veterinary Services, Ministry of Agriculture, Livestock & Fisheries, Nairobi5. Division of Disease Surveillance and Response, Ministry of Public Health, Nairobi
Dynamic Drivers of Disease in Africa REF:NE/J001422/1”
• RVF:• Mosquito-borne viral
zoonosis• High and persistent rainfall
• Would irrigation promote endemic RVF?
• Irrigation and trade offs in ecosystem services
Water and food Risk of vector-borne diseases
Irrigated site with stagnant water in the drainage canals – source of water for people but also breeding grounds for mosquitoes
Rift Valley fever case study
• The study site:• Arid/semi-arid region in
northeastern Kenya• Two irrigation schemes and
adjacent pastoral areas• Studies:
o Ecological/GIS analyses –Entomological surveys
oParticipatory studies and socio-economic surveys
o Sero-epidemiological surveys in livestock and people
• Support to policy makers to improve disease surveillance and response
Methods
Study site in Kenya, GIS team, ILRI
20 0 20 40 60 80 Kilometers
N
Open shrubs (65-40% crown cover)Very open shrubs (40-15% crown cover)
Closed herbaceous vegetation onpermanently flooded landOpen to closed herbaceous vegetationon temporarily floodedOpen to closed herbaceous vegetation
Irrigated land / Cropland
CloudsTana River-Waterbodies
Urban and Rural Settements
Open trees on temporarily flooded landTrees and shrubs savannahVery open trees (40-15% crown cover)Open trees (65-40% crown cover)Closed trees
Legenda) 1975 b) 2010
Ecological analyses: Land cover changes between 1975 and 2010
Activities – Field sites• Mosquito sampling
o 6pm-6am for 3 consecutive days/site
• Livestock and human samplingo Blood samplingo Serum extraction and storageo Sample screening using ELISA kits
• Data analyzed using geostatistical models to account for spatial effect
Field surveys
Animal sampling (photo credit: ILRI/Bernard Bett)
CDC light trap for mosquitoes (photo credit: ILRI/Bernard Bett)
Participatory and socio-economic surveys
Services- Water- Food - Income
Dis-services- Diseases (malaria,
bilharzia)- Exposure to agro-
chemicals
Land use change and disease transmission
Aedes spp Anopheles spp
Culex spp Mansonia spp
1
10
100
1000
10000
irrigated area
non-irrigated area
Villages
Mosquito species
Log
num
ber o
f mos
quito
es
Aedes spp Anopheles spp
Culex spp Mansonia spp
1
10
100
1000
10000irrigated area
non-irrigated area
Farms
Mosquito species
Log
num
ber o
f mos
quito
esAedes spp Anopheles
sppCulex spp Mansonia
spp
1
10
100
1000
10000
irrigated areanon-irrigated area
Villages
Mosquito species
Log
num
ber o
f mos
quito
es
Aedes spp Anopheles spp
Culex spp Mansonia spp
1
10
100
1000
10000
irrigated area
non-irrigated area
Farms
Mosquito species
Log
num
ber o
f mos
quito
es
I
Fallo
w p
erio
d Irr
igati
on se
ason
Results: Apparent densities of mosquitoes trapped
Variable Levels All mosquitoes trapped Primary RVF vectors Mean SD Credible interval Mean SD Credible interval
2.50% 97.50% 2.50% 97.50%Land use Irrigation 1.23 0.38 0.46 1.94 1.47 0.19 1.10 1.85
Other 0.00 0.00 Rain 0.03 0.00 0.02 0.03 0.03 0.00 0.02 0.03Hyper-parameters Theta 1 -3.03 1.97 -6.79 0.95 -3.53 3.16 -9.75 2.68Theta 2 1.87 1.53 -1.23 4.75 2.26 3.16 -3.95 8.46DIC 1099.57 641.39
Outputs of a regression model used to analyse the effects of rainfall and irrigation on mosquito densities
Analysis of sero-prevalence data from people
Variable Level Rift Valley fever sero-prevalence Odds Ratio P> |Z | Estimate 95% CI Fixed effects Gender Male 1.85 1.28 – 2.66 0.00 Female 1.00 Age (years) <9 - 9 - <18 0.10 0.02 – 0.48 0.00 >18 - <30 0.64 0.42 – 0.98 0.04 >30 1.00 Occupation Farmer 0.44 0.21 – 0.92 0.03 Pastoralist 1.00 - Student 0.32 0.05 – 2.03 0.23 Other 0.85 0.47 – 1.54 0.60 Household size <10 1.00 - >10 1.81 1.20 – 2.73 0.01 Site Irrigated 1.77 0.85 – 3.92 0.12 Riverine 1.83 0.85 – 3.92 0.11
Pastoral 1.00 Random effects ICCc: Household | Village Log likelihood -343.87
Discussion• Irrigation – increased food production but more habitat
fragmentation and less biodiversity
• Primary vectors of RVF found in drainage canals. This implies increased risk of RVF
• Seri-prevalence in livestock and people– higher in irrigated area but not significant. Surveillance for active infections required
• To manage vector-borne diseases -- better irrigation technologies instead of flood irrigation should be considered
This work, Dynamic Drivers of Disease in Africa Consortium, NERC project numberNE-J001570-1, was funded with support from the Ecosystem Services for Poverty Alleviation (ESPA) programme. The ESPA programme is funded by the Department for International Development (DFID), the Economic and Social Research Council (ESRC) and the Natural Environment Research Council (NERC).