Mosquitos are often cited as one of the deadliest animals in the world, causing up to one million deaths per year (WHO, 2020; CDC, 2021). They can carry and transmit a variety of diseases, including malaria, West Nile virus, dengue fever, and Zika virus; transmitting illness across the globe (Figure 1). To help decrease the burden of disease resulting from mosquitos, researchers are utilising satellite data and remote sensing models to better predict where mosquito breeding grounds may occur in the future. The ultimate goal is to ensure that high-risk areas are equipped with the necessary public health equipment and personnel to handle an outbreak should it occur. Oftentimes, remote communities have limited resources to manage intense disease outbreaks when they occur unexpectedly or with little warning. As a result, public health resources can become quickly overwhelmed at the beginning of an outbreak, ultimately limiting the level of care for each patient. With the help of space-based technologies, governments and health authorities can be armed with the knowledge of where outbreaks may occur in the short-term future, therefore allowing them to make more informed decisions on which regions and communities to focus aid towards before the first illness occurs. 

Figure 1: A mother holds her child at a malaria treatment hospital in Sudan (Medecins Sans Frontieres, 2020).
Figure 1: A mother holds her child at a malaria treatment hospital in Sudan (Medecins Sans Frontieres, 2020).

Solutions from space: tracking mosquito-borne disease outbreaks

Satellite data and remote sensing models can provide critical information about environmental factors that influence how mosquito populations survive and reproduce. Precipitation, for example, influences the hydrology within specific regions and determines whether suitable pools of water may be present for mosquito egg-laying (Smith et al., 2020).  Weather events like heavy rainfall can saturate soils and allow puddles to form within the environment, creating breeding grounds, and flooding may cause riverbanks to overflow and for water to subsequently stagnate. Temperature and humidity are also critical in determining mosquito survival, with warmer air temperatures and higher humidities being more conducive to life (Figure 2). Anthropogenic factors can also influence mosquito-borne diseases through changes in land use, human settlement patterns, and human movement patterns (Wimberly et al., 2021). To track all these factors over time, long-term satellite data is critical for monitoring trends and highlighting changes that may be impossible or prohibitively expensive to track from Earth directly. 

Figure 2: Modelling mosquito-borne diseases is a complex process. A multitude of factors must be considered by researchers to develop an accurate model for predicting disease outbreaks (Wimberly et al., 2021).
Figure 2: Modelling mosquito-borne diseases is a complex process. A multitude of factors must be considered by researchers to develop an accurate model for predicting disease outbreaks (Wimberly et al., 2021).

 

What type of data are these satellites actually measuring? A major contributor to our understanding of where mosquitos may flourish is from the measuring of electromagnetic radiation. Vegetation on Earth absorbs radiation from the sun and reflects and/or emits a percentage of it back outwards. The percentage of radiation that is refracted in certain specific spectral bands, such as near-infrared, red, and short-wave infrared, varies with plant health and is an indicator of vegetative health and water stress (Figures 3 and 4). Sensors on-board satellites capture the emitted electromagnetic radiation and analyze it to provide indices of plant health. For example, the normalized difference moisture index (NDMI) is indicative of vegetative moisture stress (Wimberly et al., 2021). Because plant health is dependent on temperature and precipitation, the NDMI can provide an indirect measurement of environmental factors that also impact mosquito vitality. Another spectral index, known as the normalised difference water index (NDWI), is used to detect open water bodies, which may provide breeding pools for mosquitos to lay larvae. Scientists can analyse indices like the NDMI, NDWI and normalised difference vegetative index (NDVI) for particularly wet or warm conditions which are more conducive to mosquito breeding (Figure 4). 

Figure 3: The health of a plant can be determined from the type of electromagnetic radiation that it is emitting (Antognelli, 2018).
Figure 3: The health of a plant can be determined from the type of electromagnetic radiation that it is emitting (Antognelli, 2018). 

 

Figure 4: Another spectral index, known as the normalized difference vegetative index (NDVI) is indicative of healthy plant matter. Plants require sufficient water to grow healthily but may also imply that there is enough water within the environment to provide breeding grounds for mosquitos. (Tucker, 1979; Reich, 2016).
Figure 4: Another spectral index, known as the normalised difference vegetative index (NDVI) is indicative of healthy plant matter. Plants require sufficient water to grow healthily but may also imply that there is enough water within the environment to provide breeding grounds for mosquitos. (Tucker, 1979; Reich, 2016). 

 

There are several different satellites that are important for capturing information relevant to mosquito-borne diseases. For instance, NASA’s Terra and Aqua satellites both contain a moderate resolution imaging spectroradiometer instrument, also known as MODIS. These sensors can produce images every one to two days at a resolution of 250m-1000m (NASA, 2023). MODIS data has been captured for decades, providing consistent, long-term records of environmental change. Scientists can study such data to determine the impact that environmental changes have on public health and disease transmission. Another option for remote sensing is through the Landsat satellites. These satellites can provide higher resolution images than MODIS, measuring optical data at 30m resolution and thermal data at 60-120m (Figure 5). The increased resolution that the Landsat satellites can provide is hugely beneficial in identifying areas of high-risk at a level that is more practical. For example, specific areas of high-risk may be distinguished from the community or region as a whole. The Landsat satellites, however, have a longer re-visit time than the MODIS satellites, having the potential to produce an image every 16 days. Because of this, Landsat imagery is more suitable for measuring changes on a seasonal and annual timescale, rather than daily. 

Figure 5: A comparison of imagery from the MODIS and Landsat satellites. Although MODIS has a coarser resolution, in this case, the circled features were still visible in both sets of imagery (Sibandze et al., 2014).
Figure 5: A comparison of imagery from the MODIS and Landsat satellites. Although MODIS has a coarser resolution, in this case, the circled features were still visible in both sets of imagery (Sibandze et al., 2014). 

 

In the case of mosquito-borne diseases, however, it is often useful to have imagery at a higher resolution than either the MODIS or Landsat satellites can provide. House-hold level interventions require house-hold level imagery and the ability to see geographic features like individual ponds or temporary water bodies. To achieve this, scientists use very high resolution (VHR) satellite imagery has resolutions of <1m to 5m (Wimberly et al., 2021). Because individual water bodies, such as puddles or small ponds, can serve as habitats for mosquito larvae, VHR imagery can be useful for interventions on the smallest scale, such as an individual household or neighbourhood (Valle et al., 2013). Producing VHR imagery on a broad scale, however, comes at significant expense, and therefore may not be the most practical method of remote sensing for widespread mosquito-borne disease analysis. 

Case study: Forecasting malaria outbreaks in the Peruvian Amazon

The deadliest of all mosquito-borne diseases is malaria, accounting for approximately 619,00 deaths in 2022 (World Health Organization, 2022). Within South America, most malaria cases occur within the Amazon region (Recht, 2017). Interestingly, there is a disparity of malaria cases between the western and eastern parts of the Amazon. In the western region, including portions of Peru, Colombia, and Ecuador, the number of reported cases of malaria have tripled between the years 2011 and 2017, despite many vector-control campaigns (Reiny, 2017). It is hypothesised that this region may be particularly vulnerable to mosquito-borne diseases due to unique hydrological conditions stemming from periodic occurrences of El Niño and related flooding (Pan et al., 2017). By contrast, the number of cases of malaria within the Brazilian Amazon steadily decreased during the same time period (Figure 6) (Recht et al., 2017; Reiny, 2017). 

Figure 6: The prevalence of malaria within both Peru and Brazil during the years 2011-2015. Although the number of cases in Brazil fell, the cases in Peru steadily increased (Recht et al., 2017).
Figure 6: The prevalence of malaria within both Peru and Brazil during the years 2011-2015. Although the number of cases in Brazil fell, the cases in Peru steadily increased (Recht et al., 2017).

 

An ongoing study to better understand how to forecast where malaria hotspots are and when malaria outbreaks are likely to occur has been taking place in the Peruvian Amazon. Led by Dr. William Pan and Dr. Zaitchik, the research team developed a four-component Early Warning System to estimate ¬where malaria hotspots and when malaria outbreaks will occur in the future. Their modelling system allowed them to forecast malaria outbreaks 12 weeks in advance, with approximately 90% sensitivity (Pan, 2020). A breakdown of each model component is described below: 

  1. Land Data Assimilation System (LDAS): this system is used by researchers to monitor temperature, precipitation, humidity, soil moisture, solar radiation, and vegetation on Earth. This system requires input from several NASA satellites, including Landsat, Terra, and Aqua, which produce spectral indices. As a modelling component, the data from these satellites is used to measure the distribution of water on Earth’s surface, as well as make estimations of where the water may be in the future (Rodell and Holland, 2022).
  2. Seasonal Human Population Model: this model is being used to estimate where human populations are located, particularly in relation to high-risk transmission areas (Reiny, 2017). 
  3. Sub-Regional Statistical Model: this model is used to locate areas that have exceeded the expected levels of malaria in the past (Samadoulougou et al., 2014; Pan et al., 2017). In other words, it aims to highlight past disease hotspots that may be problematic again in the future even if environmental conditions may not suggest it as a hotspot at present.  
  4. Agent-Based Model (ABM): this model is used “to integrate human, environmental, and entomological transmission dynamics” (Pan et al., 2017). In other words, it studies the influence of small-scale human movements on the formation of malaria hotspots. An interesting highlight from the ABM is that many cases of malaria are contracted by farm workers during the daytime (Pizzitutti et al., 2018). This is in contrast to the historical assumption that most malaria transmissions occur at night, given that Anopheles mosquitos typically have nocturnal biting habits (Chavasse, 2002). 

While it is, perhaps, too early in the study to see the efficacy of this technology over the long-term, achieving 90% sensitivity is a remarkable feat. It underscores the fact that disease prediction is becoming increasingly possible through the integration of space-based technologies within epidemiology. This study also provides several initial learnings for other researchers to implement in other geographic settings. For example, it is highlighted within Dr. Pan’s research how critical the importance of forming strong government and academic relationships within Peru has been. For novel technologies like this to be implemented and adopted long-term, it has been shown to be essential to create and foster strong working partnerships at the regional level.

Benefits and Limitations 

The potential of space-based technologies in reducing the spread of mosquito-borne diseases is clear. Remote sensing images and predictive models can help identify areas that may be at high-risk for outbreaks several months into the future. The ultimate intent is for governments to provide health resources and interventions to areas deemed most vulnerable before an outbreak even occurs. Dr. Pan states that “Instead of treating 100 percent of the community, we could focus vector control on certain households or specific areas of the community. It’s a targeted strategy that can achieve the same reduction in malaria, but at potentially lower costs and with a more rapid response”. In addition, perhaps the most powerful aspect of using space-based technologies is the ability to apply these models worldwide. The sheer scope of monitoring would be otherwise impossible to achieve without remote sensing. 

Despite the clear advantages that remote sensing can provide in terms of predictive health monitoring, there remain limitations in their usage beyond academia. Governments and health authorities generally have not had access to current data analysis and therefore have been unable to incorporate such information into decision making. Within many malaria-endemic countries, health data is managed and analysed through a platform known as District Health Information Software 2 (DHIS2). Until recently, there has been limited incorporation of Earth observation data within the DHIS2 platform (Beck, 2021). As researchers continue to integrate Earth observation data within the DHIS2 platform, there is hope that this data will be utilised more frequently within health ministries. Given that external sources of foreign aid are not unlimited, it is crucial for decision makers to make the most informed choices, with the inclusion of Earth observation data, of where to distribute precious resources. 

Space technology applications in other vector-borne diseases surveillance

Many of the recent large-scale studies that use Earth observation data to track mosquito-borne diseases have focused on the most prevalent diseases such as malaria, dengue fever, or West Nile virus. Scientists have recognised, however, that these same models may also be applicable to less common mosquito-borne diseases with minimal modifications (Figure 7). This is because a single genus of mosquito is often capable of transmitting a variety of different diseases. For example, both malaria and lymphatic filariasis are commonly spread by the Anopheles mosquito within rural regions of Africa (CDC, 2019). The Aedes genus is responsible for the transmission of Dengue fever worldwide but can also transmit Rift Valley Fever and Zika virus (ECDC et al., 2022). Because of this, it has been estimated that specific early warning systems may be applied to a variety of diseases within the same geographic area, given that they are spread by the same species of mosquito (Parselia et al., 2019). 

Figure 7: Mosquitos are a problematic vector because they can cause a wide variety of diseases (Boukhatem, 2017). Please note that this list is not inclusive of all mosquito-borne diseases.
Figure 7: Mosquitos are a problematic vector because they can cause a wide variety of diseases (Boukhatem, 2017). Please note that this list is not inclusive of all mosquito-borne diseases.

 

In addition to mosquito-borne diseases, Earth observation technologies are also being applied to other vector-borne diseases such as Lyme disease or schistosomiasis (Kotchi et al., 2021). Although the ecology, reproduction cycles, and transmission cycles of each specific vector may vary, the underlying principles and technologies remain the same. For example, schistosomiasis is a disease transmitted via freshwater snails. Several large-scale studies are currently ongoing across Africa and Asia that utilize space-based technologies and data to accurately model snail breeding grounds with the intention of limiting human exposure (Liu et al., 2022; Xue et al., 2021). Although the habitats and lifestyles of various vectors may differ substantially to those of mosquitos, satellite observation data can still be used in a similar fashion to allow for predictive modelling of breeding grounds. As scientists track the ecology and breeding patterns of different vectors more accurately, the potential to limit human exposure and prevent outbreaks increases. 

Conclusion

The use of space-based technologies to aid in decision making within health care is an emerging and exciting relationship that can yield immense, positive results.  Remote sensing holds the potential to detect early outbreaks across numerous vector species which would allow for focused and precise response in areas where it is most effective - allowing finite resources to be optimized. Managing mosquito-borne diseases is a constantly changing and evolving task.  With further research, support, development and implementation, remote sensing technology could help decrease illness and promote health, thereby saving millions of lives globally.

Sources

Antognelli, S., 2018. NDVI and NDMI vegetation indices: instructions for use | Agricolus [WWW Document]. AGRICOLUS. URL https://www.agricolus.com/en/vegetation-indices-ndvi-ndmi/ (accessed 2.18.23).

Beck, J.M., 2021. Improving Malaria Decision Support with Earth Observations. The University of Alabama in Huntsville.

Boukhatem, M.N., 2017. Major mosquito-borne diseases, pathogens and vector mosquito(es). | Download Table [WWW Document]. nternational Journal of Pharmacology, Phytochemistry and Ethnomedicine. URL https://www.researchgate.net/figure/Major-mosquito-borne-diseases-patho… (accessed 12.22.22).

CDC, 2021. CDC - Parasites - World Mosquito Day 2021: CDC’s Efforts to Control the World’s Deadliest Animal—Photo Essay [WWW Document]. Global Health, Division of Parasitic Diseases and Malaria. URL https://www.cdc.gov/parasites/features/world_mosquito_day_2021_photo_es… (accessed 12.9.22).

CDC, 2019. Lymphatic Filariasis - General Information - Vectors of Lypmhatic Filaraisis. Parasites - Lymphatic Filariasis.

Chavasse, D.D., 2002. Know your enemy: Some facts about the natural history of Malawi’s Anopheles mosquitoes and implications for malaria control. Malawi Med J 14, 7.

ECDC, Semenza, J., Menne, B., 2022. Vector-borne diseases. European Centre for Disease Prevention and Control. https://doi.org/10.1186/1476-072X-6-40

Kotchi, S., Bouchard, C., Brazeau, S., Ogden, N., 2021. Earth Observation-Informed Risk Maps of the Lyme Disease Vector Ixodes scapularis in Central and Eastern Canada. Remote Sens 13, 524.

Liu, Z.Y.C., Chamberlin, A.J., Tallam, K., Jones, I.J., Lamore, L.L., Bauer, J., Bresciani, M., Wolfe, C.M., Casagrandi, R., Mari, L., Gatto, M., Diongue, A.K., Toure, L., Rohr, J.R., Riveau, G., Jouanard, N., Wood, C.L., Sokolow, S.H., Mandle, L., Daily, G., Lambin, E.F., de 

Leo, G.A., 2022. Deep Learning Segmentation of Satellite Imagery Identifies Aquatic Vegetation Associated with Snail Intermediate Hosts of Schistosomiasis in Senegal, Africa. Remote Sensing 2022, Vol. 14, Page 1345 14, 1345. https://doi.org/10.3390/RS14061345

Medecins Sans Frontieres, 2020. Tackling TB in Sudan and South Sudan [WWW Document]. MSF. URL https://www.msf.org/msf-tackles-severe-malaria-outbreak-western-sudan (accessed 12.22.22).

NASA, 2023. MODIS Moderate Resolution Imaging Spectoradiometer [WWW Document]. NASA. URL https://modis.gsfc.nasa.gov/about/ (accessed 2.19.23).

Pan, W., 2020. AN EARLY WARNING SYSTEM FOR VECTOR-BORNE DISEASE RISK IN THE AMAZON NASA PROJECT NNX15AP74G. Health & Air Quality Applications Program Review.

Pan, W.K., Zaitchik, B.F., Pizzitutti, F., Berky, A., Feingold, B., Mena, C., Janko, M., 2017. Forecasting Malaria in the Western Amazon. AGUFM 2017, NH53A-0136.

Parselia, E., Kontoes, C., Tsouni, A., Hadjichristodoulou, C., Kioutsioukis, I., Magiorkinis, G., Stilianakis, N.I., 2019. Satellite Earth Observation Data in Epidemiological Modeling of Malaria, Dengue and West Nile Virus: A Scoping Review. Remote Sensing 2019, Vol. 11, Page 1862 11, 1862. https://doi.org/10.3390/RS11161862

Recht, J., Siqueira, A.M., Monteiro, W.M., Herrera, S.M., Herrera, S., Lacerda, M.V.G., 2017. Malaria in Brazil, Colombia, Peru and Venezuela: current challenges in malaria control and elimination. Malaria Journal 2017 16:1 16, 1–18. https://doi.org/10.1186/S12936-017-1925-6

Reich, L., 2016. Understanding your Aerial Data: Normalized Difference Vegetation Index NDVI - Geoawesomeness [WWW Document]. Geoawesome. URL https://geoawesomeness.com/eo-hub/understanding-aerial-data-normalized-… (accessed 2.18.23).

Reiny, S., 2017. Using NASA Satellite Data to Predict Malaria Outbreaks. NASA.
Rodell, M., Holland, R., 2022. Land Data Assimilation System. NASA. https://doi.org/10.5194/ESSD-14-3115-2022

Samadoulougou, S., Maheu-Giroux, M., Kirakoya-Samadoulougou, F., de Keukeleire, M., Castro, M.C., Robert, A., 2014. Multilevel and geo-statistical modeling of malaria risk in children of Burkina Faso. Parasit Vectors 7, 1–13. https://doi.org/10.1186/1756-3305-7-350/TABLES/4

Sibandze, P., Mhangara, P., Odindi, J., Kganyago, M., 2014. A comparison of Normalised Difference Snow Index (NDSI) and Normalised Difference Principal Component Snow Index (NDPCSI) techniques in distinguishing snow from related land cover types. South African Journal of Geomatics 3, 197. https://doi.org/10.4314/SAJG.V3I2.6

Smith, M.W., Willis, T., Alfieri, L., James, W.H.M., Trigg, M.A., Yamazaki, D., Hardy, A.J., Bisselink, B., de Roo, A., Macklin, M.G., Thomas, C.J., 2020. Incorporating hydrology into climate suitability models changes projections of malaria transmission in Africa. Nature Communications 2020 11:1 11, 1–9. https://doi.org/10.1038/s41467-020-18239-5

Tucker, C.J., 1979. Red and Photographic Infrared Linear Combinations for Monitoring Vegetation. Remote Sens Environ 8, 127–150.

Valle, D., Zaitchik, B., Feingold, B., Spangler, K., Pan, W., 2013. Abundance of water bodies is critical to guide mosquito larval control interventions and predict risk of mosquito-borne diseases. Parasit Vectors 6, 1–2. https://doi.org/10.1186/1756-3305-6-179/TABLES/1

WHO, 2020. Vector-borne diseases [WWW Document]. WHO. URL https://www.who.int/news-room/fact-sheets/detail/vector-borne-diseases (accessed 12.22.22).

Wimberly, M.C., de Beurs, K.M., Loboda, T. v., Pan, W.K., 2021. Satellite Observations and Malaria: New Opportunities for Research and Applications. Trends Parasitol 37, 525. https://doi.org/10.1016/J.PT.2021.03.003

World Health Organization, 2022. World Malaria Report 2022. World Health Organization.

Xue, J.B., Wang, X.Y., Zhang, L.J., Hao, Y.W., Chen, Z., Lin, D.D., Xu, J., Xia, S., Li, S.Z., 2021. Potential impact of flooding on schistosomiasis in Poyang Lake regions based on multi-source remote sensing images. Parasit Vectors 14, 1–13. https://doi.org/10.1186/S13071-021-04576-X/FIGURES/7