| dc.description.abstract | This study employed Ecological Niche Modeling (ENM), a technique that encompasses a suite
of tools that relate known occurrences of species or phenomena to raster geographic information
system layers that summarize variation in several environmental dimensions. The spatialtemporal distributions of the main malaria vectors in Kenya were quantified using BIOCLIM and
DOMAIN models to determine the relationship between vector distribution and climate change.
The biological data used was from published sources (Okara et al., 2010 and MARA/ARMA,
1998), comprising of point samples for geo-referenced malaria vector occurrences. The climate
data used was maximum temperature, minimum temperature and precipitation for current climate
(1950-2000) and climate projection for HADCM3, CCCMA and SCIRO models of IPCC
projected future climate under the A2a scenario by the years 2020, 2050 and 2080. The climate
data was acquired in grid format from WorldClim global climate data which was further
processed to generate 19 bioclimatic variables for Kenya.
The predictions showed that by the year 2020, the suitability areas for malaria vectors in Kenya
will start to change from the current ecological suitability. Most areas where the malaria vectors
are thriving currently will still remain suitable ecologies. New suitability zones will emerge in
most counties ranging from low to very high suitability as shown by the predictions. By the year
2050, areas of suitability will expand at an alarming extend. The year 2080 has been predicted to
show that the suitable ecologies will start to revert to the original areas of suitability as in the
current climate. Therefore, climate change in Kenya will adversely affect the environment at an
alarming rate by 2050, but beyond that there will be a level of stabilization, where further change
will trigger reversal to the past climate.
For instance, BIOCLIM True or False prediction from HADCM3 by the year 2050 showed wide
spread of malaria in counties like Narok, Kajiado, Kitui, Makueni, Machakos, Meru, Marsabit,
Isiolo, Samburu, Baringo, West Pokot. Turkana county and Mandera among a few others will
have some emerging isolated malaria hot spots. ENM prediction with HADCM3 future climate
showed that Laikipia County will become unsuitable malaria ecology by the year 2050 and the
case remains the same by the year 2080.
Validation results for prediction model performance showed that all the models used had errors
in prediction as none of them had kappa =1 or AUC=1. The highest kappa (k = 0.909) and Area
under ROC Curve (AUC=0.954) values were achieved from DOMAIN model with CCCMA
projection by the year 2020. The lowest model performance values of k = 0.427 and AUC =
0.714 were obtained from BIOCLIM True or False model with HADCM3 projections by the
year 2020.
The following conclusions were drawn from the Ecological Niche Modeling done using
BIOCLIM, BIOCLIM True or False and DOMAIM prediction models: There is correlation
between climate change as an explanatory variable and the distribution of main malaria vectors
in Kenya. The spatial-temporal distribution of the main malaria vectors in Kenya varies under
different IPCC future climate projections which are HADCM3, CCCMA and CSIRO. The future
ecological niches for malaria vector occurrence in Kenya will extend from the current niches in
most endemic areas, new hotspots will emerge and some suitable ecology will become
unsuitable, resulting in varying areas from current climate predictions to projections by the year
2020, 2050 and 2080 under IPCC A2a scenario. Intervention strategies such as indoor or outdoor
residual spraying, distribution of insecticide-treated mosquito nets (ITNs) and long-lasting
insecticide-treated nets (LLINs) should be diversified in new emerging areas for disaster risk
reduction and increase adaptive capacity and resilience among local communities. | en_US |
