Spatial-Temporal Distribution of Malaria Risk and Its Association With El Niño Southern Oscillation (ENSO)

_____________________________________________________________________________________________________ Sarawak recorded the second-highest number of cases since 2013 until 2017 after Sabah. Sarawak is the largest state in Malaysia and needs to provide spatial information, especially to the ministry of health. The objective of this study was to examine the impact of El Niño-Southern Oscillation (ENSO) on the distribution of malaria risk maps. To achieve the objectives of this study requires Oceanic Niño Index (ONI) data, Visible Infrared Imaging Radiometer Suite (VIIRS), daily temperature, and secondary data on the number of malaria cases in Sarawak. The results of the study clearly show that the occurrence of La Niña and El Niño affects the total distribution of Malaria risk maps. The number of malaria cases is also related to the ONI value. The lower the ONI value causes the malaria case value to decrease. The results of this study suggest that most of the hot spots in the forest, forest fringe, and inland areas of Sarawak. This clearly shows the lack of knowledge and knowledge causing the rural population to be prone to malaria. The Ministry of Health needs to focus on the interior in disseminating teachings and knowledge in dealing with malaria mosquitoes.


Introduction
By 2020, the World Health Organization (WHO) has identified that Malaysia is on the list of 21 potential countries where malaria can be epidemic to humans. Although malaria cases are reported to be declining, the need to increase measures to control and eradicate malaria is a major problem, especially in rural areas where the population lacks knowledge or education to deal with malaria, especially in Sarawak and Sabah compared to Peninsular Malaysia (Hussin et. al., 2020). Hussien et al., (2020) stated that the states of Sabah and Sarawak recorded the highest accumulated cases from 2013 to 2017 of 7150 in Sabah and 5684 in Sarawak representing 77.7% of all cases in Malaysia. This explains why the researcher chose the state of Sarawak as the study area. Furthermore, Hussin et al. (2020) strongly recommended that a complete geographical distribution map is very important. Besides, those responsible should understand the geographical distribution of Malaria is another key point to identify high-risk infection areas and plan appropriate control strategies and surveillance systems. It is also important to project areas where malaria transmission is most likely to occur even if human malaria has been controlled. The use of satellite technology that studies weather and climate is very important in the use of climate indicators to predict the risk of malaria. In African countries, the use of remote sensing technology in predicting malaria is widely practiced (Thomson et al, 2001;Teklehaimanot et al., 2004), but for other malaria-endemic areas, such technology is underdeveloped. For example, similar studies are rare in Asia and rarer in Malaysia (Husin et al., 2020). A large www.msocialsciences.com number of literature reviews as described below use predict malaria risk using GIS and remote sensing using weather and climate factor parameters especially rain and temperature. However, the study did not examine the effects of ENSO on malaria. In this study, researchers tried to predict the risk of malaria map during ENSO occurrence using VCI.

Literature review
The latest study and similar to this study are Hasyim et al., (2018) who conducted a study in South Sumatra Province, Indonesia. Their study of the spatial association between malaria occurrence and environmental risk factors used GIS Analysis ordinary least square (OLS) and geographically weighted regression (GWR) in producing six potential ecological predictors of malaria cases map. Their study found that independent variables such as altitude, distance from forest, and rainfall play a role in increasing malaria cases. The amount of rainfall plays an important role in the spread of malaria. The results of this study were also found to be similar to the study conducted by Alemu et al., (2010) in Jimma Town, South West Ethiopia. Excessive rainfall and temperatures ranging from 25-27 degrees Celsius will increase the life span of malaria mosquitoes and increase the distribution of malaria mosquito breeding sites (Yamana et al., 2013;Abiodun et al., 2016;Weaver et al., 2014). At 28 degrees Celsius and lack of moisture will shorten the life span of mosquitoes (Yamana et al., 2013). The first study in Asia using remote sensing technology was Adimi et al., (2010) who conducted a study on making malaria risk predictions in Afghanistan. Their study predicted malaria risk from 2004-2007 using Moderate Resolution Imaging Spectroradiometer (MODIS) and Tropical Rainfall Measuring Mission (TRMM). MODIS data provides temperature data, vegetation index, and TRMM rainfall data. Their study uses regression technique in building malaria cases model based on environmental predictors. The results of their study found that vegetation index through Normalized Difference Vegetation Index (NDVI) is the strongest predictor in the spread of malaria. Their study used a linear regression model in predicting the spread of malaria. Mushinzimana et al., (2006) studied landscape determinants of anopheline mosquito larva habitats and used remote sensing technology in identifying their habitats in western Kenya Highlands. Their study used aerial photos, IKONOS and Landsat Thematic Mapper 7. The study compared the effectiveness of the three data in identifying mosquito habitats. Based on their results, it was found that IKONOS images were able to identify 40.6% compared to aerial photos and 0 for Landsat. Midekisa et al., (2012) first study using seasonal autoregressive integrated moving average (SARIMA) models in measuring the relationship between malaria cases and remote sensing environmental variables namely rainfall. Land surface temperature and vegetation indices such as NDVI and EVI and lastly evapotranspiration. Their study found that these malaria cases were positively related to land surface temperature, rainfall, EVI, and transpiration. Dhimal et al., (2014) studied the spatial-temporal distribution of malaria and the relationship between malaria and climate factors and vector control interventions in Nepal. Their study found that malaria micro-stratification, mosquito repellents should be prescribed for those living in the forest, forest fringe, and foothill. Based on previous studies it was found that vegetation indices such as NDVI, rainfall factor, and temperature play an important role in the spread of malaria. However, there have been no attempts to study the relationship between ENSO and the spread of malaria using remote sensing. In addition to testing the VIIRS sensor in studying the spread of malaria is still lacking. This study also uses the Vegetation Condition Index (VCI) which was first tested in studying its relationship with malaria.

Method and study area
VIIRS satellites were used in this study. Data used from 2013 to 2020. However, this study only focuses on two ENSO incidents in La Niña in late 2017 and early 2018 and El Niño 2015/2016. Data has been downloaded from the National Center for Applications and Research, National Environmental Satellite Data and Information Services, NOAA, and its URL is https://www.star.nesdis.noaa.gov/smcd/emb/vci/VH/index .php. Apart from VIIRS data, this study also uses Oceanic Niño Index (ONI) data, daily and monthly temperatures from the Malaysian meteorological department (MMD). In-depth information on ONI data, malaria cases, and temperature will be discussed in the next chapter.

Study Area
The study area is the state of Sarawak. According to Zulfaqar et al., (2017), the temperature in the study area is hot throughout the year with an average daily temperature between 23 degrees Celsius in the early morning to 32 degrees Celsius during the day. Sarawak's climate is influenced by Madden Julian Oscillation, El Niño Southern Oscillation (ENSO), Indian Ocean Dipole (IOD), and Monsoon season (Dindang et al., 2013;Yik et al., 2015). The occurrence of the ENSO phenomenon plays an important role in shaping the daily weather in the study area. In 1997/1998 and 2015/2016 saw a prolonged drought and forest fires. Meanwhile, 2009 saw heavy rains and caused major floods in lowland areas in Sarawak. These extreme weather changes have aroused many researchers' interest in conducting analytical studies on meteorological data and remote meteorological remote sensing. Going back to this study, the question that arises is whether the occurrence of ENSO affects the risk of Malaria.

Vegetation Index
The Vegetation Condition Index (VCI) is the best parameter for plants to respond to total rainfall. Refers to VCI capable of capturing water-related stress.
Where NDVImax and NDVImin are the maximum and minimum values of NDVI. Kogan (1990) stated that VCI is more sensitive to the value of dynamic rainfall than NDVI.

Oceanic Niño Indeks (ONI)
The ONI is well established to identify events of El Niño and La Niña (Rojas et al., 2014;NOAA, 2020). The ONI index shows the development and intensity of El Niño or La Niña events in the Pacific Ocean. ONI is a three-month Sea Temperature (SST) anomaly in Niño region 3.4 5 ° N -5 ° S, 120 ° -170 ° W). The occurrence of El Niño is defined when the average value of three months at or above + 0.5 ° C anomaly, while the event of La Niña is defined as or under climate Anomaly (NOAA, 2020) −0.5 ° C (NOAA Forecast Center, 2019).
www.msocialsciences.com         Whereas if the ONI value is negative 1 eat the VHI value is 65. Based on the evidence from NOAA (2020) classify if the negative ONI value 0.5 and below is classified as La Niña. This explains why the La Niña incident also had a good effect on VHI. A more in-depth description of La Niña's impact on VHI will be explained through the VHI map below. ONI values have a significant effect on the distribution of VHI values. was due to the El Niño incident that hit Sarawak in that year. The El Niño incident led to changing rainfall and weather patterns which impact the vector bionomics (Copper et al., 2017). Things can be explained in more detail on the malaria risk map in the next chapter. Dhimal et al., (2014) reported that the effects of rainfall on malaria are complex. In some areas, the effects of heavy rainfall are providing breeding grounds and encouraging mosquito population growth. However, heavy rainfall can cause the breeding grounds to disappear and thus decrease the abundance of mosquitoes. For this study, researchers used the Vegetation Condition Index (VCI) because of the response to the amount of rainfall. VCI values range from 0 to 100. Classification of VCI values in classifying water content conditions and malaria risk as shown in the table below. Suwito et al., (2010) found that temperature plays an important role in determining the intensity of mosquito volume. The density of mosquitoes in the monsoon season where temperatures range from 26-26.5 degrees Celsius is the peak time of mosquito numbers. The number of malaria mosquitoes began to decline at 27 degrees Celsius. He also stated that air humidity plays an important role in contributing to the density of mosquitoes, especially in December in Indonesia. Their study also found that low humidity will shorten the life of mosquitoes and high humidity prolong the life of mosquitoes The table above shows the classification of water content and malaria risk using VCI. If the VCI value is 75 to 100 then it is classified as very high-water content and very high risk of malaria, 50 to 75 is normal malaria risk and moderate water content. While 25 to 50 means moderate low water content and risk of malaria and the last value of 0 to 25 means very low risk of malaria and water content.    Based on figure 9, it is found that the distribution of moderately high-risk malaria risk especially in December 2017 and January 2018 where the value of ONI is ONI is negative 0.9 and 1 almost all of Sarawak except 11 districts such as Sibu, Sarikei, Balingian, Belawai and Tanjung Manis. Based on figures 9 and 8 clearly shows a significant difference where during the La Niña incident that is figure 9 many green areas that have a high risk of malaria compared to the La Niña incident. This proves that VCI values play an important role in predicting malaria risk. Besides, it can be observed that the malaria hot spot risk area of most inland areas consists of forest areas. The diagram above clearly shows a lower VCI value during the El Niño incident which is 45.41 for land plants and 36.55 for cropland due to lack of water content due to lack of rainfall compared to during the La Niña incident which is 79.57 and 75.74 during the La Niña incident. VCI values are positively related to the total risk of malaria as well as the number of mosquitoes.

Results and discussion
In addition to previous studies also found that the humid conditions during the occurrence of La Niña will prolong the life of malaria mosquitoes compared to during the occurrence of El Niño (Hasyim et al., 2018).

Conclusion
This study succeeded in achieving the objective of the study which is to study the impact of ENSO on malaria risk distribution. This study clearly shows the occurrence of La Niña and El Niño affect the distribution of malaria risk in Sarawak. The occurrence of La Niña where the ONI value is low which causes the amount of rainfall to be more than usual causes the distribution of mosquito disease risk to be higher than during El Niño. Through the provision of information in this column, the health sector knows where the area needs to be focused in channeling aid, awareness campaigns, and knowledge to the locals.