Estimating Water Content and Grain Size of Intertidal Flat Sediments Using Visible to Shortwave-Infrared Reflectance and Sentinel 2A Data: A Case Study of the Red River Delta, Vietnam

Abstract

Sediment properties such as water content (WC) and grain size (GS) are essential to characterize the environmental conditions of tidal flats. This article aimed to develop appropriate models to estimate the WC and GS of surface sediments for an intertidal flat on the Red river delta (Vietnam) using Sentinel 2A (S2A) images. The spectral reflectance, WC, and GS of 96 sub-samples from 12 sediment samples collected on December 17, 2017 were measured to clarify their relationships. The WC was highly correlated with the reflectance ratio of two shortwave-infrared bands, R(2190)/R(1610) (R2 = 0.93). The median GS (D50) at 0%, 15%, and 20% of WC was significantly correlated with the reflectance ratio of the near-infrared band (842 nm) versus the visible-green band (560 nm) (R2 > 0.78). Next, D50 was estimated from a multivariate regression model using this band ratio, the visible-red band (665 nm), and WC. The accuracy of the models was verified by comparisons with WC and D50 from 20 samples collected on March 12th 2019 (RMSE of both WC and D50 < 15%). Then, the WC and sediment type distributions were mapped by applying these models to two S2A scenes. The maps showed high WC (>30%) in very fine sediments (silts), which is consistent with other intertidal flats with similar sediment types. This article was limited to fine sediment samples. Therefore, our next step is to incorporate coarse sediments into the models to provide more universal mapping of WC and sediment types.

Keywords

• Hyperspectral imaging
• land surface
• remote sensing
• sediments
• soil properties