Spatial Extrapolation Of In Situ Observations of Suspended Sediment Via Remote Sensing Imagery in the Mekong Delta Region

In typical field experiments in situ observations occur at very specific spatial locations over long time scales. Unfortunately it is difficult to know what might be changing in nearby locations during the time of the in situ collection. This is where remote sensing imagery can play a useful role by providing a large field of view snapshot of the same observations that are being made with the in situ equipment, thereby giving spatial context for the observations.

In this study we collected satellite multi-spectral imagery (MSI) and synthetic aperture radar (SAR) imagery over a region of the Mekong Delta where in situ observations of suspended sediment concentration and river flow were made over two periods: Fall of 2014 and Spring of 2015. We are developing algorithms to estimate suspended sediment concentration from the MSI data, and in particular the higher resolution (5 meter) imagery from the Rapid Eye Sensor, using ratios of the various bands and comparing the results to in situ observations when they were coincident and within the image field of view. The river flow measurements are used to convert spatial changes in suspended sediment along the river from the MSI imagery into temporal changes at the in situ measurement site so that they can be compared directly. We then generate spatial maps of suspended sediment concentration, particularly over the river outflow plume, to generate the overview to provide context for the in situ measurements.

Initial results from the Fall 2014 experiment show that we can generate suspended sediment concentration from the MSI imagery that is within ±0.03 grams/liter of the coincident in situ observations, and can image flow patterns in the sediment maps within the river outflow region. In the talk we will present final algorithm results for both the Fall 2014 and Spring 2015 experiments, as well as a time series of river flow sediment patterns, compared to in situ observations. We intend this to be a demonstration of the utility of remote sensing imagery to provide larger-scale spatial patterns of the very local in situ observations typically performed in field experiments.