Abstract
This analysis and discussion section presents the results of hot spot analysis and spatial autocorrelation using the Getis-Ord Gi* statistic and Global Morans 1 index, respectively, to identify patterns and relationships in river centerline data errors. The hot spot analysis confirmed the presence of hot spots with a high level of confidence for errors related to ‘not a river’ and ‘oxbow lakes’. The missing centerline errors were also analyzed, revealing hot spots in the northwest corner and central region of the region of interest. The results were consistent with the overall error density map. Spatial autocorrelation analysis showed a significant spatial relationship between centerlines ending abruptly and river width, as well as a strong correlation between missing centerlines and river width. The findings highlight the importance of considering various factors, such as river width, in remote sensing analyses. However, limitations due to small dataset size and other factors, such as cloud and tree cover, should also be taken into account. Overall, this analysis provides insights into the spatial patterns and relationships in river centerline data errors, contributing to a better understanding of the data quality and potential sources of errors in remote sensing analyses.
Please reference the full text below:
