Modeling the Influence of Climate Factors on Malaria Transmission Dynamics in North Kordofan State, Sudan

Background: Despite great efforts by the government to control malaria in Sudan, the disease is the most significant human disease and was widespread in North Kordofan State. Morbidity and mortality of the disease are increas-ing in the State. Usually, the disease reached its peak after rainy season. This study aims to estimate the role of climate factors on malaria transmission dynamic by modeling the relationship between malaria cases and climatic variables, such as rainfall, relative humidity, and temperature, in Kordofan State. Methods: We used Pearson correlation coefficient and an ordinary least square method to assess this relationship. Results: The results show that there are statistically significant associations between malaria cases and rainfall, relative humidity, and minimum temperature (P-value < 0.001). The regression analysis results suggest that the appropriate model for predicting malaria incidence includes malaria cases lagged by one month, maximum temperature, and minimum temperature. This model explained 72% of the variance in monthly malaria incidence. Conclusion: The results of this study suggest that climatic factors have potential use for malaria prediction in the State.


Background
Malaria is caused by parasites of the genus of plasmodium and is transmitted through a bite of female anopheles mosquitoes [1]. Malaria is world's most important vector-borne disease and is the leading cause of death among children under the age of five, a major cause of adult morbidity, and a major cause of loss of these cases were in Africa and India [2]. Variation in climatic circumstances, such as temperature, rainfall patterns, and relative humidity, has a profound effect on the survival of mosquitoes and on the growth of malaria parasites in mosquitoes. Consequently, these climatic factors can also influence the transmission of malaria [3] [4] along with public health services, insecticide usage, drug resistance rates, and human population movements [5]. Nutrition has also been shown to affect malaria transmission, especially in children and pregnant women [6] [7].
Several studies have shown the influence of variations in these climatic factors on malaria transmission. Associations between climate factors and the number of malaria cases have been reported recently in many areas in Africa [8] [9].
Conversely, some studies have reported that there is no significant association between climatic trends and the number of malaria cases in East Africa where cases of malaria increase every year [10]. Some studies found that only minimum temperature affects malaria transmission [11]. Huang et al. found a positive effect of rainfall on malaria transmission [12] while Mabaso et al. suggested a combination of rainfall, minimum and maximum temperature affect malaria transmission [13].
Environmental changes such as deforestation and housing structure could also influence malaria transmission as they enhance mosquito breeding conditions [14]. Ye

Malaria Transmission
Malaria spreads after a mosquito becomes infected with the disease by biting an infected human. The infected mosquito then transmits the disease when it bites an uninfected human and the malaria parasites move to the liver. When a parasite matures, it leaves the liver and infects the red blood cells [19]. Temperature is one of the climatic factors, known to play an important role in the development of the parasite [20] [21] [22]. The transmission of malaria in Sudan is seasonal, with the peak during autumn rainy season, depends on climatic conditions which can powerfully affect the number and survival of mosquitoes [18].
Therefore, it is possible for malaria epidemic outbreaks to occur in areas with vulnerable populations having little or no immunity to malaria. In addition, unexpected meteorological phenomenon and human activities affecting the environment, such as, deforestation and agricultural irrigation could affect malaria transmission and lead to epidemic outbreaks [23]. The forecasting of the consequences of climate change in malaria transmission is of great importance to achieving better malaria control and avoid epidemic outbreaks. The dynamic process of malaria transmission includes numerous interlinked variables, from uncontrollable environmental conditions to man-made natural disturbances [24]. In this paper we study the linkage of malaria transmission with climatic factors (temperature, relative humidity, and rainfall) in Kordofan State where most people are in extreme poverty.

Study Area
This study was carried out in North Kordofan State, Sudan. This state lies in the central-western part of country at the northern edge of the savannah belt. The area of Kordofan is about 244.700 km 2 ; with a population of 2,920,000 representing approximately 7.5% of the total population Sudan. The State is divided into 9 localities 45 units. El-Obeid is the capital of the state (see Figure 1).
The economic structure based largely on agriculture, focusing on growing crops and livestock breeding. However, most of the population are living in poverty or far below reasonable standards of living as urban deprivation is widespread.
Rainfall in the State is sporadic in occurrence with a rate between 150 mm in the north and 850 mm in the south of the State. There are recurrent droughts at varying length and severity (e.g., 1968-1974; 1983-1985 and 1990-1991, 2000, 2003, and 2009   The VIF for the jth predictor could be defined as follows:

Statistical Analysis of Data
where 2 j R is the coefficient of determination obtained by regressing the jth predictor on the remaining predictors.
The variables we have in the regression model are: Y t , the number of malaria cases at time t; max (˚C), the maximum temperature; min (˚C), the minimum temperature; R.h (%), relative humidity; and rainfall (mm). We used a stepwise regression method to calculate all possible regression models using a set of our climatic variables. This method is used to determine which explanatory variables are relevant from a set of candidate explanatory variables. Then we chose the appropriate model which overcomes the multicollinearity problem (low VIF j ) with a high value of R 2 .       Table 2 illustrates the results of the Spearman correlation coefficient analysis, which examines the relationship between climatic variables and malaria cases.

Results
The results show that there was a significant positive relationship between malaria cases and minimum temperature, relative humidity, and monthly total rainfall (P-value < 0.001). There was no significant relationship between malaria cases and maximum temperature. To study the relevance lags of accumulated number of malaria cases, we shifted one month behind the monthly malaria cas- The time shift revealed there is a significant serial correlation of the number of malaria cases (P-value < 0.001).
The relationships between malaria and climatic variables were further checked by linear regression analyses. Multiple regression analysis was used to predict the number of malaria cases based on climatic factors. Table 3  The P-values for all independent variable were less than 0.0001, which suggests that the three independent variables account for a significant part of the variation in t Y .   Variables

Discussion
This study investigated the influence of climatic factors on malaria parasite transmission, in North Kordofan State, Sudan. It is well known that the relationship between malaria outbreaks and climatic factors is indirect and complex.
The climatic variability that affects mosquito vectors, malaria parasites, and intrinsic of malaria transmission dynamics contributes significantly to malaria transmission [26].
These results suggest that increased rainfall increases the number of malaria cases. This effect can be explained by the large amount of swamp water in the Kordofan State during an autumn season. The changes of temperature in the Kordofan justify the significant relationship between malaria and minimum temperature [27]. These results suggest that climatic factors have potential use for malaria prediction in the State.
The significant serial correlation of the number of malaria cases ( Table 2, r = 0.644) indicates that the variable "the number of cases in this month" affects malaria transmission in next month. This means that malaria in any month could be a potential variable for malaria prediction and future outbreaks. Our results showed that monthly minimum temperatures were the strongest predictor of malaria incidence, followed by relative humidity. The strong positive correlation between total rainfall and relative humidity shown in our data (

Conclusions
In this paper, we model the influence of selected climatic variables on the incidence of malaria in Kordofan State, Sudan. We employed correlation coefficient and regression analyses. Most of the malaria cases in the study area are reported to have occurred during rainy months compared to the dry season ( Figure 3).
The results revealed that the minimum temperature, relative humidity, and For the government to reduce malaria transmission, the WHO goal of malaria control should be carried out through prevention interventions such as distribution of insecticide-treated nets, intermittent preventive treatment of malaria and Indoor residual spraying.
There may be a limitation of this research that some bias of the prediction model as rainfall is significantly correlated with the number of malaria cases, but due to the multicollinearity problem of regression model, the best model is Model 2 in which rainfall is excluded.