Science Cafe Posts

Identifying Environmental Attributes Characterizing Mosquito Habitats- A Regression Analysis: Guest Scientist Blog by Eric C., 2020 NASA SEES Intern

Cases of mosquito-borne disease are increasing in the United States. Because these diseases are directly correlated with the abundance of mosquitoes, finding what environmental variables impact mosquito populations is essential to prevent the transmission of disease. For the SEES 2020 Mosquito Mapper research project, our group decided to  focus on identifying the statistical correlations exhibited between mosquito abundance and topographic, underground, and climate variables. The analysis used GLOBE Observer Mosquito Habitat Mapper data. The GLOBE database was very simple to access and provided my group important data in a time-constrained period. Because the mosquito data is the independent variable, all the explanatory variables had to be dependent on the longitude and latitude coordinates of where the in-situ mosquito data was collected. AppEEARS and NASA Worldview provided data on these explanatory variables: elevation, soil moisture, vegetation, temperature, GPP (Gross Primary Productivity) , NEE (Net Ecosystem Exchange), RH (Relative Humidity), and SOC (Soil Organic Carbon).

Results:

Table 1: Primary Results and Statistics for Variables

Through utilizing single regression line graphs, we were able to find the statistical relationship between the abundance of mosquitoes and a myriad of variables. The table above displays our results. The direction of the coefficient explains the type of relationship that the variable has with the mosquito larvae count. A positive coefficient indicates a positive relationship, a negative coefficient indicates a negative relationship, and coefficient zeroing indicates an insignificant relationship. Thus, we can conclude that variables like NDVI had a positive relationship with the mosquito larvae count while variables like elevation had an insignificant relationship. Moreover, the p-Value demonstrates the statistical significance of our conclusions.

 

Regarding the model as a whole, the regression model had a f-value of 0.0138 and an r-squared value was 0.22, indicating that 22% of the sample data was accounted for in the regression.

Future Plans:
My group wasn’t able to use all of GLOBE Observer’s features to its fullest extent. Thus, my group plans on extending our research through adding more variables.

Acknowledgments:
The analyses presented in this blog are the outcomes of our 2020 SEES Team Project, Analysis on the Relationship Between the Prevalence of Mosquitoes and Topographic, Underground, and Climate Variables. Research contributions by Jessica W., Eric Q., and Kaveh S. are gratefully acknowledged.

Eric C. is a high school student from California who is working on a research project this summer using the GLOBE Observer Mosquito Habitat Mapper. His virtual internship is part of a collaboration between the Institute for Global Environmental Strategies (IGES) and the NASA  Texas Space Grant Consortium (TSGC) to extend the TSGC Summer Enhancement in Earth Science (SEES) internship for U.S. high school (http://www.tsgc.utexas.edu/sees-internship/). He shares his experience this summer in this guest blog post.

 

More Blog Entries