Abstract

Multispectral remote sensing data have proven to be useful in deriving tidal flat topography. However, limited satellite observations over a certain period have large uncertainties. In this study, we used MODIS time-series data with a high observational frequency to generate accurate tidal flat topography in Nantong, China based on the relationship between the T_Tide-derived tidal levels and the MODIS-derived inundation frequency map. First, 8-day MOD09Q1 data from 2007 to 2008 were used to perform the land-water classification. Second, 92 land-water maps were stacked to generate the inundation frequency map of the tidal flat. Then, the T_Tide package was applied to calculate the tidal levels at Lvsi Tide Gauge Station. Finally, the inundation frequency map and the tidal levels were integrated to derive the tidal flat topography, which agreed well with the in-situ elevation data (RMSE = 0.40 m, r = 0.89) and the Landsat-based elevation data (RMSE = 0.18 m, r = 0.98). In addition, the derived slopes agreed well with the slopes from the in-situ elevation data (RMSE = 1.00‰, r = 0.85). We highlighted the necessity of using all MODIS data for deriving an accurate tidal flat topography. Our proposed method has a potential to derive tidal flat topography and temporal changes over the past 20 years from MODIS data.