Pixels and Pupils: A Comparative Analysis of Remote Sensing and Citizen Science Land Cover Data in Los Osos, CA

Abstract

Human impact is constantly reshaping land cover dynamics. Remote sensing technology is revolutionary in its ability to provide imagery and land cover data of local and global regions. However, the accuracy of localized remote sensing data remains uncertain. This validation study uses citizen science data to assess the accuracy of remote sensing datasets at a local scale. This evaluation was performed in Los Osos, California, an environmentally sensitive region currently assessing human impact and the health of habitats and endangered species. The null hypothesis was that there would be no differences between remote sensing and citizen science data sets. Research was conducted through the two citizen science platforms: the GLOBE Observer App and Collect Earth Online. These were compared to open-source remote sensing data sets obtained through the Earth Map platform. Visual analysis of data collected at each of the 37 sample units rejected the null hypothesis because of conflicting classification patterns. Citizen science data revealed that remote sensing disagreements were often associated with the incorrect consolidation of land cover elements or confusion between similar features. While remote sensing is generally more efficient, citizen science field observations were critical to the improved understanding of the land cover of local regions. These findings demonstrate citizen science data’s ability to validate remote sensing data sets and identify misclassifications. This emphasizes the need for the inclusion of field observations to improve local land cover maps.

.