Evaluation and Fusion of MODIS Data to Enhance the Reliability of GK2A/GK2B Geostationary Satellite Vegetation Indices

This study aimed to implement fused data and enhance the reliability of the Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI) over the Korean Peninsula by synergizing the high temporal resolution of geostationary satellites with the proven stability of polar-orbiting satellites. The time-series characteristics of two geostationary satellites, the Geostationary Korea Multi-Purpose Satellite-2A (GK2A) and Geostationary Korea Multi-Purpose Satellite-2B (GK2B), were quantitatively evaluated using a long-term validated Moderate Resolution Imaging Spectroradiometer (MODIS) as the baseline. The triple collocation (TC) method was used to quantify the error variance and root mean square error of each dataset across various land cover types, enabling the assessment of relative uncertainties in the absence of ground reference data. Subsequently, a high-quality fusion framework was developed by applying weights derived from the TC-based error variances, thereby maximizing the strength of each sensor and mitigating its limitations. Although all three datasets captured seasonal variability in vegetation, distinct disparities in quality were observed. TC analysis identified GK2A as the most stable source, comparable to MODIS, whereas GK2B exhibited higher uncertainty, which was attributed to underestimation and unstable seasonal fluctuation. The resulting fused NDVI and EVI synthesized via error-variance weighting demonstrated significant improvements in spatiotemporal continuity and accuracy, reflecting the high fidelity of the GK2A. This study is significant for establishing a systematic methodology to evaluate relative uncertainty and fuse multi-satellite vegetation indices without ground truth. The proposed approach provides a scientific foundation for the high-frequency, high-density monitoring of vegetation dynamics, which is essential for effective environmental assessment and hazard mitigation on the Korean Peninsula.