Impact of Weekly Climatic Variables on Weekly Malaria Incidence throughout Thailand: A Country-Based Six-Year Retrospective Study

Purpose. This study aimed to evaluate climatic data, including mean temperature, relative humidity, and rainfall, and their association with malaria incidence throughout Thailand from 2012 to 2017. The correlation of climatic parameters including temperature, relative humidity, and rainfall in each province and the weekly malaria incidence was analyzed using Spearman's rank correlation. The results showed that the mean temperature correlated with malaria incidence (p value < 0.05) in 44 provinces in Thailand. These correlations were frequently found in the western and southern parts of Thailand. Relative humidity correlated with malaria incidence (p value < 0.05) in 35 provinces. These correlations were frequently found in the northern and northeastern parts of Thailand. Rainfall correlated with malaria incidence (p value < 0.05) in 38 provinces. These correlations were frequently found in the northern parts and some western parts of Thailand. The impacts of the mean temperature, relative humidity, and rainfall were observed frequently in specific provinces, including Chiang Mai, Chiang Rai, Trat, Kanchanaburi, Ubonratchathani, and Si Sa Ket. This is the first study to report areas where climatic data are associated with malaria incidence throughout Thailand from 2012 to 2017. These results can map out the climatic change process over time and across the country, which is the foundation for effective early warning systems for malaria, public health awareness campaigns, and the adoption of proper adaption measures that will help in malaria detection, diagnosis, and treatment.


Introduction
Climate changes are alternations in one or more climate variables, including temperature, relative humidity, rainfall, wind, and sunshine. ese changes may impact the survival and reproduction of vectors and the transmission of vectorborne diseases [1]. Malaria, caused by Plasmodium spp., is one of the most climate-sensitive mosquito-borne diseases. Climate changes can directly affect malaria transmission by shifting the vector's geographic range and increasing the reproductive and biting rates and by shortening Plasmodium spp. incubation period [2]. A previous study successfully developed a system to forecast the probability of malaria incidence using a seasonal timescale multimodel of climate prediction and applied it for the prediction of malaria risk in Botswana [3].
ailand is a tropical country at the center of the Indochina Peninsula in Southeast Asia, with a total area of approximately 513,000 km 2 , and is located between latitudes 5°37′N to 20°27′N and longitudes 97°22′E to 105°37′E. ailand is bordered to the north by Myanmar and Laos, to the east by Laos and Cambodia, and to the south by the Gulf of ailand and Malaysia. ailand is divided into six regions defined by the National Geographical Committee in 1978, which include northern ailand's 9 provinces, northeastern ailand's 20 provinces, western ailand's 5 provinces, central ailand's 8 provinces, eastern ailand's 7 provinces, and southern ailand's 14 provinces. e climate of ailand is under the influence of monsoon winds of a seasonal character, such as the southwest monsoon, which usually starts in mid-May and ends in mid-October, and the northeast monsoon, which normally starts in mid-October and ends in mid-February [4]. Because of the effect of the location and monsoons, ailand has a relatively high temperature and humidity all year-round, which is ideal for the reproduction of Anopheles mosquitoes.
Exploring the relationships between climate change and malaria transmission in ailand is the first step in developing effective early warning systems for malaria because malaria has been endemic for more than 10 consecutive years in some areas of ailand. e objective of this study is to clarify the relationship between climate variables and malaria incidence dynamics throughout the 77 provinces of ailand between 2012 and 2017.

Malaria Incidence in ailand.
Weekly malaria incidences between 2012 and 2017 were retrieved from the Bureau of Vector-Borne Diseases, Ministry of Public Health, ailand [5], which is available to the public at http://www. thaivbd.org/n/home. Some missing data from the website were requested directly from the Bureau of Vector-Borne Diseases. ese data contained the weekly incidence of malaria from all 77 provinces throughout ailand (52-53 weeks per year).

Climatic Data of ailand.
e daily climatic data were derived from measurements of mean temperature, relative humidity, and rainfall from all weather stations across 77 provinces of ailand between 2012 and 2017. e climatic data were retrieved from the ai Meteorological Department. e climatic data included mean temperature (°C), relative humidity (percent), and rainfall (milliliter) in all 77 provinces. e daily climatic data were then transformed using Microsoft Office Excel (Microsoft Corporation, Redmond, VA, USA) to weekly climatic datasets to analyze with weekly malaria incidence data provided by the Bureau of Vector-Borne Diseases.

Correlation of Climatic Parameters and Malaria Incidence in ailand.
e correlation of climatic parameters, including temperature, relative humidity, and rainfall in each province, and the weekly malaria incidence was performed by Spearman's rank correlation. Any province correlated with climatic data was mapped by color in the ailand map according to the number of years it correlated with malaria incidence. For example, provinces with malaria incidence for more than six years were indicated with a dark color and provinces with malaria incidence for only one year were indicated with a light color.
All statistical analyses were performed using SPSS Statistics for Windows, version 17.0 (SPSS Inc., Chicago, IL, USA). e results are considered statistically significant when p values are less than 0.05. e figures were created using Microsoft Office Excel and Microsoft Office Power-Point (Microsoft Corporation, Redmond, VA, USA).

Malaria Incidence.
According to reports from the Bureau of Vector-Borne Diseases for 2012-2017, the largest malaria outbreak occurred in 2012 with 45,413 cases, and incidence rates continuously decreased year to year to 11,179 cases in 2017. Figure 1 shows the number of malaria cases in ailand between 2012 and 2017. Malaria incidence rates in each year began to rise in the 15 th week of the year and peaked in the 22-28 th weeks ( Figure 2). It is evident from this figure that malaria incidence and the amount of rainfall are correlated.

Correlation between Mean Temperature and Malaria Incidence in ailand.
e results from the correlation test showed that the mean temperature correlated with malaria incidence rates in 44 provinces (p value < 0.05). e frequency correlation between the mean temperature and malaria incidence between 2012 and 2017 is shown in Figure 3. Among those 44 provinces, Ratchaburi Province, located in the western part of ailand, showed a correlation with mean temperatures and malaria incidence rates for six consecutive years between 2012 and 2017 (p value < 0.0001). Mean temperatures in Kanchanaburi, Phetchaburi, and Prachuap Khiri Khan in western ailand and Narathiwat in southern ailand correlated with malaria incidence rates over five consecutive years between 2012 and 2016 (p value < 0.05). Mean temperatures in Chiang Rai, located in the northern part of ailand, also correlated with malaria incidence rates over five consecutive years between 2013 and 2017 (p value < 0.05).

Correlation between Relative Humidity and Malaria Incidence in
ailand. e results of the correlation test showed that relative humidity correlated with malaria incidence in 35 provinces (p value < 0.05). e frequency correlation between relative humidity and malaria incidence rates in 2012-2017 is shown in Figure 4. Among those provinces, Ubonratchathani, a province located in the northeastern region of ailand, showed a correlation with relative humidity and malaria incidence over five consecutive years between 2012 and 2017 (p value < 0.0001). Relative humidity in Ratchaburi, a province located in western ailand, correlated with malaria incidence rates over four consecutive years between 2014 and 2017 (p value < 0.05). Relative humidity in Narathiwat, a province located in southern ailand, correlated with malaria incidence rates in 2012, 2013, 2015, and 2016 (p value < 0.05).

Correlation between Rainfall and Malaria Incidence in
ailand. e results of the correlation test showed that rainfall correlated with malaria incidence in 38 provinces (p value < 0.05). e frequency correlation between 2012 and 2017 is shown in Figure 5. Among those provinces, Tak, a province located in western ailand, showed a correlation with rainfall and malaria incidence rates over five consecutive years between 2013 and 2017 (p value < 0.05).

Discussion
An association between malaria incidence rates and climatic factors has been established. is study examined the impact of climatic variables, including mean temperature, relative humidity, and rainfall, on malaria incidence throughout 77 provinces in six regions of ailand.
e initial findings showed that the mean temperature correlated with malaria incidence rates in 44 provinces in ailand. ese correlations were frequently found in western and southern ailand. is was because the temperatures in southern ailand are generally mild throughout the year as a result of being in the coastal area [4], and western ailand usually has long periods of warm weather because of its inland nature and tropical latitude [4]. ese specific locations with suitable temperatures can affect the spatial-temporal distribution of malaria vectors. Temperature is a key driver that determines transmission intensity, including the development and biting rates of Anopheles mosquitoes, as well as the development and survival rates of the malarial parasites within the mosquitoes [6,7]. A previous study also indicated that rising temperatures in low-latitude regions may lead mosquitoes to find new habitats in mid-or high-latitude regions, leading to geographical expansion or shifts in disease transmission [1]. Moreover, a study reported an association between interannual variability in temperature and malaria transmission in the African highlands [8]. Another study in Southeast Iran showed that as the mean, maximum, and minimum monthly temperature increased, the incidence rate of malaria increased significantly [9]. A study in China indicated that mean temperature was also associated with malaria cases over long periods of time [10]. Studies conducted in Tibet and Ethiopia showed strong positive and significant correlations between malaria occurrence and relative humidity, rainfall, and temperature [11,12].
Both Anopheles and Plasmodium are sensitive to temperature. eir life stage is dependent on temperature both in development and mortality rates [13].
e previous finding showed that temperatures exceeding 33°C-39°C may limit the development of P. falciparum and P. vivax [2]. Moreover, a study found that the extrinsic incubation period (EIP) of P. falciparum is reduced from 26 days at 20°C to 13 days at 25°C [14]. is means a high development rate of Plasmodium and biting rate of Anopheles may frequently be found.
Recent research found that Anopheles mosquitoes and malarial parasites were influenced not only by the average temperature but also by the temperature fluctuation of the low mean temperatures during the day, which helps to speed up biological processes, whereas fluctuation around the high mean temperatures acts to slow processes down [15]. P. falciparum transmission is limited by temperatures below 16°-19°C, whereas P. vivax development can occur at temperatures as low as 14.5°-15°C. Parasite development cannot occur with temperatures above 33°-39°C for P. falciparum and P. vivax [16].
e second result of this study showed that relative humidity correlated with malaria incidence (p value < 0.05) in 35 provinces. ese correlations were frequently found in the northern and northeastern provinces of ailand. Humidity was found to affect malarial parasite development in Anopheles mosquitoes [17]. Relative humidity affected malaria transmission by impacting the activity and survival of mosquitoes. A study reported that mean monthly relative humidity under 60% causes a shortened    lifespan in malaria vector mosquitoes, which results in low malaria transmission rates [18], and a relative humidity of less than 10% is fatal [19]. e results of this study correlated with those of a previous study, showing that relative humidity was positively associated with malaria incidence rates from the same month [20,21]. A previous study in China showed that relative humidity was associated with P. vivax and P. falciparum over 8-10 weeks and 5-8 weeks, respectively [10]. Moreover, a study conducted in Yongcheng, China, between 2006 and 2010 demonstrated temperature and relative humidity as the main drivers of malaria transmission [22]. e third result of this study showed that rainfall correlated with malaria incidence (p value < 0.05) in 38 provinces.
ese correlations were frequently found in northern parts and some western parts of ailand. e upper parts of ailand usually have rainy weather because of the onset of the southwest monsoons that lead to intensive rainfall from mid-May until early October in this region. e correlation between rainfall and the prevalence of malaria in ailand has been reported previously; however, the rainfall data were transformed from the geographical data and the prevalence of malaria was transformed from the overall infection rate of malaria, which was not based on real data of malaria cases and real climatic variables from weather stations across ailand [23]. A previous study reported that monthly rainfall was negatively correlated with the proportion of patients with malaria hyperparasitemia and the proportion of gametocyte carriers among P. falciparum cases [24]. e results of this study concurred with a study in Africa that showed that the correlation of persistent malaria transmission was associated with a higher level of rainfall [25]. A previous experimental study showed that rainfall may affect vector populations at the larval and adult stages. Since Anopheles mosquitoes breed in small natural pools of clean water, droughts usually result in decreases in vector populations and malaria transmission by limiting the number and quality of vector breeding sites [26]. However, a study in China indicated that rainfall had a decreasing effect on P. vivax [10]. is contrary effect may be due to excessive rainfall affecting the mosquito population because strong rains may sweep away their breeding sites [27].
e fourth result of this study showed that more than one climatic variable was associated with malaria incidence. Mean temperature and relative humidity affected malaria incidence rates in some provinces of northeastern, eastern, western, and southern ailand. However, mean temperature and rain affected malaria incidence rates in the northeastern, eastern, western, southern, and northern regions of ailand. is means that the alteration of both mean temperature and rainfall has a higher impact on malaria transmission in ailand than the alteration of both mean temperature and relative humidity. erefore, northern ailand, including Mae Hong Son in 2014 and Chiang Rai in 2015, was not affected. e findings also showed that relative humidity and rainfall affected malaria incidence rates in some provinces in all six regions of ailand, including northeastern, eastern, western, southern, northern, and central ailand.
is means that the alteration of both relative humidity and rainfall has a higher impact on malaria transmission in ailand than that of mean temperature and relative humidity or mean temperature and rainfall. Because of this, malaria incidence rates among provinces in central ailand were also affected. e affected provinces of central ailand included Kamphaeng Phet in 2013, Phetchabun in 2014, Phitsanulok in 2016, and Bangkok in 2014 and 2016. Lastly, the results also showed that alteration of all three climatic parameters, including mean temperature, relative humidity, and rainfall, affected malaria incidence rates in specific provinces in all 6 regions of ailand. However, the provinces were different from the two climatic parameters. Similar observations of mean temperature and relative humidity affected malaria incidence rates in the Guangzhou area, China, where both temperature and relative humidity were positively associated with malaria incidence [20]. A study in Kenya reported that hospital admissions for malaria have been associated with rainfall and high maximum temperature [28]. However, high maximum temperature and malaria incidence were not analyzed in this study.
e main limitation of the study was the lack of other climatic data, such as wind, length of sunlight, evaporation, and visibility. However, using one or more climatic variables, including mean temperature, relative humidity, and rainfall, could help to map out the climatic change process through time and across the country, which is the foundation for effective early warning systems for malaria, public health awareness campaigns, and the adoption of proper prevention measures that will help with malaria detection, diagnosis, and treatment.

Conclusion
is was the first study to report the areas where climatic data were associated with malaria incidence throughout ailand from 2012 to 2017. ese results can map out the climatic change process through time and across the country, which is the foundation for effective early warning systems for malaria, public health awareness campaigns, and adoption of proper prevention measures that will help with malaria detection, diagnosis, and treatment.

Data Availability
e datasets used during the current study are available from the corresponding author upon reasonable request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.