Observation operators for assimilation of satellite observations in fluvial inundation forecasting

Abstract

Abstract. Images from satellite-based synthetic aperture radar (SAR) instruments contain large amounts of information about the position of flood water during a river flood event. This observational information typically covers a large spatial area, but is only relevant for a short time if water levels are changing rapidly. Data assimilation allows us to combine valuable SAR-derived observed information with continuous predictions from a computational hydrodynamic model and thus to produce a better forecast than using the model alone. In order to use observations in this way a suitable observation operator is required. In this paper we show that different types of observation operator can produce very different corrections to predicted water levels; this impacts on the quality of the forecast produced. We discuss the physical mechanisms by which different observation operators update modelled water levels and introduce a novel observation operator for inundation forecasting. The performance of the new operator is compared in synthetic experiments with that of two more conventional approaches. The conventional approaches both use observations of water levels derived from SAR to correct model predictions. Our new operator is instead designed to use backscatter values from SAR instruments as observations; such an approach has not been used before in an ensemble Kalman filtering framework. Direct use of backscatter observations opens up the possibility of using more information from each SAR image and could potentially speed up the time taken to produce observations needed to update model predictions. We compare the strengths and weaknesses of the three different approaches with reference to the physical mechanisms by which each of the observation operators allow data assimilation to update water levels in synthetic twin experiments in an idealised domain.

A Mutual Information‐Based Likelihood Function for Particle Filter Flood Extent Assimilation

“…Flood extents simulated by the “truth” model were used to generate the synthetic observations, at the time steps corresponding to actual SAR acquisitions, using the approach proposed by Cooper et al. (2019). Backscatter distributions of flood and nonflood classes were assumed to follow the form of Gaussian Mixture Models.…”

“…(2018) assimilated flood extent maps; and Cooper et al. (2019) directly used backscatter values to reduce flood forecast uncertainty.…”

“…Thus, developing a flood extent-based likelihood to isolate the best performing ensemble members which can efficiently drive the assimilation is a significant scientific challenge (Lai et al, 2014). For instance , Cooper et al (2019) proposed the conversion of modeled binary flood extents into synthetic SAR observations, and inter compared the wet and dry backscatter observations at the flood boundary with actual SAR images. While directly utilizing backscatter values reduces the processing time, it discards any information contributed by the addition of texture (Dasgupta et al, 2018b) or coherence (Chini et al, 2019), which may reduce uncertainties.…”

On the Impacts of Observation Location, Timing, and Frequency on Flood Extent Assimilation Performance

“…These were subsequently converted into binary flood extents using a minimum water depth threshold of 1 cm, chosen to maximize extent variability between time steps. The truth model simulated binary flood extents were then processed into synthetic SAR images using the approach proposed by Cooper et al (2019); an example of this process is illustrated in Figure 1. For each pixel in the binary "truth" model simulated flood maps, a backscatter value was sampled from the distributions of flood and non-flood classes obtained from typical flooded image histograms.…”

Automated Water Segmentation and River Level Detection on Camera Images Using Transfer Learning

“…Gauges are expensive to install and maintain, and their measurements can be unreliable when the river goes out-of-bank during a flood. Satellite data from Synthetic Aperture Radar (SAR) can provide information when the river goes out-of-bank, but the frequency of satellite overpasses is limited (currently at most once in each 12 h period) [ 5 , 17 ].…”