Student Research Reports
Open social media sharing options menu
Analyzing Local Land Cover Using Surrounding Data
Country:United States of America
Student(s):Yuvraj Sahu
Grade Level:Secondary School (grades 9-12, ages 14-18)
GLOBE Teacher:Cassie Soeffing
Contributors:Dr. Rusty Low, scientist, IGES
Peder Nelson, scientist, OSU
Dr. Erika Podest, scientist, NASA JPL
Dr. Becky Boger, scientist
Matteo Kimura, peer mentor
Report Type(s):International Virtual Science Symposium Report, Mission Mosquito Report
Protocols:Land Cover Classification, Mosquitoes
Presentation Video:
View Video
Presentation Poster:
View Document
Optional Badges:I am a Data Scientist, I am an Engineer, I am a STEM Storyteller
Language(s):English
Date Submitted:02/28/2022
Land cover can be predicted where it is not feasible to make observations. Using the Land Cover
tool within the GLOBE Observer app, data was collected and classified based on percent
coverage. Four factors were selected for testing - deciduous trees, evergreen trees, grass, and
urban. Using a weighted average formula, the predicted land cover at each possible location
within each possible rectangle was calculated. For all 4 corners, it used the 2 directions facing
the point being predicted. As the distance or angle difference increased, the weight decreased.
The mean absolute error, in percentage points, was then calculated for each of the factors. The
overall error was 16.6 for deciduous trees, 13.4 for evergreen trees, 17.5 for grass, and 38.5 for
urban. In general, the small rectangles had lower error than the large rectangles. This confirms
that, for certain factors, surrounding data can be used to predict land cover at a certain point.