MAPPING (EXTENSION AND CHARACTERIZATION) OF RIVERINE AND COASTAL PLUMES (ALSO IN RELATION TO RIVER INPUTS)

Federico Falcinia, Mario Benincasaa,b, Jaime Pitarcha

a Institute of Marine Sciences, National Research Council of Italy (CNR-ISMAR), Rome (Italy)

b RHEA Group, Frascati (Italy)

Research theme aims

In this research theme we show a remote sensing approach that estimates suspended sediment concentration of specific grain-size populations, i.e., the ones that effectively contribute to coastal morphodynamics. For this task we will use multispectral ocean colour data along with a complete set of in-situ marine observations. Our strategy focused on: i) the retrieval of GSD (Grain-Size Distribution) from particle backscattering coefficient spectrum bbp(λ) [1]. This is made off the Tiber River mouth (Tyrrhenian Sea); ii) for a larger scale analysis, the TSM (Total Suspended Matter) statistical trends over the Adriatic coast, to provide an assessment of sediment starvation pattern for this basin. Finally we apply a new methodology that combines observational data with numerical modelling. The aim is to pair satellite measurements of suspended sediment with velocity fields from numerical model for marine currents, to obtain an estimation of the sediment flux and its divergence along the Adriatic coast.

Materials and Methods

For the Tyrrhenian Sea, we used data from the Moderate Resolution Imaging Spectro-radiometer (MODIS) since it provides the longest, single sensor, data record that includes the November 2012 flood event. We downloaded AQUA L1A images over the area for the period 2002 to 2014. L2 data were computed with the Seadas 7.0 software by means of l2gen command. The cloud threshold was set to 0.012. The Remote Sensing Reflectances (Rrs) were extracted from a ~ 20 x 20 km box centered around the Tiber River mouth. To track the river’s turbid water off the river mouth, we selected the maximum Rrs(667) values for each day (in the practical implementation, we chose the sampled pixel as the closest to the 99.5th percentile, to avoid outliers). We thus obtained a time series of daily Rrs spectra (at 443, 488, 555 and 667 nm) over the same period of the collected water discharge data. The monthly average Rrs spectra were then used to retrieve monthly particle backscattering spectral slope (h) (eq.1).

                          

                  (Eq.1)

The next step requires the GSD estimation from the particulate backscattering coefficient bbp(l). Kostadinov et al. (2009) [1] describes the mathematical procedure. In particular, the relationship between h and x is:

                   (Eq. 2)

In particular, here we defined the fine sediment that between the limits 3.9 µm and 62.5 µm, the medium-sized sediment between 62.5 µm and 125 µm and the gross sediment between 125 µm and 250 µm.

For TSM trend estimation in the Adriatic basin we coupled the Mann-Kendall test and the Sens’ method, which are here applied to a de-seasonalized monthly time series as obtained from the X-11 technique. The dataset covers the period spanning from 2003-01-04 to 2012-04-07, with a daily temporal resolution and a spatial resolution of 300 m. Because the seasonal component can mask small movements in the trend signal, we remove the seasonal signal from Coastcolour TSM dataset (http://www.coastcolour.org/) before determining the trend. Finally, to estimate the divergence of sediment flux we combine a remotely sensed TSM field from Coastcolour with a velocity field from an oceanographic model. This allows us to estimate the sediment erosion and deposition patterns on the West Adriatic coastal plume. In detail, we make use of a MITgcm run (Massachusetts Institute of Technology General Circulation Model; http://mitgcm.org) that has been setup to investigate coastal upwelling and downwelling processes in the Adriatic Sea during a strong dense water event that occurred in winter 2012 [2]. The model data are available every 3 hours while the satellite data are available with a variable period, roughly close to 24 h, depending on the satellite overflying times. We chose to pick one single model result per day, i.e., the closest to the satellite observation.

Results and Discussion

Off the Tiber River Delta, we found a relationship between flow stage characteristics (i.e., flow erraticity) and the GSD. By plotting the monthly particle backscattering spectral slope (h) versus the coefficient of variation (CV). we find that an increase of CV (i.e., more erratic river discharges) correlates to more negative values of h, which indicate coarser particles in suspension. On the contrary, more persistent flow discharges (i.e., low CV) are related to finer particles in suspension (Fig.1). Besides delivering coarser sediment fraction off the river mouth, erratic floods seem to be also efficient in spreading sediment along the coast, thus providing a potential contribution to coastal growth [3,4]. Indeed, by mapping the relative amount of sediment volume of the 125-250 µm fraction with respect to the total sediment volume (Fig.1), we find that erratic floods (high CV) show a much wider coastal plumes when compared with low-CV floods. This is primarily due to the high-water discharge that characterizes flash floods.

Figure 1: Relative concentration (in %) of the amount of sediment volume for the 125-250 µm fraction respect to the total sediment volume. Arrows indicate where the events are located within the central plot. Water discharge (Qw) is indicated for each event. The central panel shows Scatter plot of the coefficient of variation of monthly river discharge (CV) vs. the monthly particle backscattering spectral slope (h). Dot diameter is proportional to the number of cloud-free satellite images. Best linear fit is also overlapped (black line). Months with less than three valid observations were discarded, as calculation of a standard deviation is not possible for them. Dot color indicates the month (see color bar).

At sub-basin scale, for the Adriatic Sea, we focused on the whole suspended sediment budget that feed the Italian coastal area. By using satellite-based statistical trend analysis, we were able to assess the along shore distribution of suspended sediment from a climatic point of view. In particular, we found a general positive trend off the Po River Delta, due to sediment input of the river runoff, which confirms the coastal geomorphological balance in this area. However, the North and Central portion of the Italian coasts is marked by a negative trend, which indicates sediment starvation and, in turn, may relate to coastal erosion at large spatial and temporal scale (Fig.2).

Figure 2: On the left the statistical trend for TSM concentration over the period 2003-2012. On the right the statistical significance of the TSM trend. White pixels mark significant values of TSM concentration trend (left map).

In figure 3 we show monthly averages of sediment flux divergence fields: December 2011, January 2012, February 2012 and March 2012. A pixel with positive divergence (a yellow or a red pixel) means a sediment erosion point, or, more correctly, a point where we have sediment from the bottom layer resuspended   into the water flow. Conversely a pixel with negative divergence (a blue or light blue pixel) means a sediment deposition point, or a point where suspended sediment is precipitating on the bottom layer.

Figure 3: Monthly averages of sediment flux divergence: (a) December 2011; (b) January 2012; (c) February 2012; and (d) March 2012.

Our results provided a satellite-based tool for a synoptic investigation of river plume sediment characteristics. In particular, we find an emerging relation between grain-size distribution and flow stages (erratic vs. persistent floods), also relating this to seasonality. This analysis demonstrate the geomorphological effectiveness of specific floods, i.e., those erratic flows that are able to spread along shorelines coarser suspended load. Due to a decrease in the Po River runoff, observed from the 2003 [5], we can argue that the sedimentary plume debouching from its mouth dramatically reduced its ability to redistribute the suspended sediment along the whole Adriatic coast. This would explain the two TSM trends we found, i.e., a positive trend immediately off the river mouth and a negative trend all along the North- and Central-East coast.

Finally, through the coupling of TSM remote sensing data and numerical modelling for marine currents, it was possible to investigate the mass conservation law, linked to the erosive-depositional balances of sediment along the coast on a regional and basin scale.

Future perspectives

The ongoing work is on exploring the capability of remote sensing in characterizing suspended sediment in terms of grain size distribution (GSD) is becoming a crucial topic in fluvial geomorphology. However, GSD of suspended particles from river plumes is still seldom quantified and insufficiently documented. The relationship between the spectral beam attenuation coefficient (c_p) and the GSD has been predicted theoretically and verified experimentally. However, those findings had no application to remote sensing data since the beam attenuation coefficient is not retrievable from space. Fortunately, after commercial optical backscattering meters became available, research is providing evidence that the spectral particle backscattering coefficient (bbp) and the GSD are also related in a very similar fashion as c_p and the GSD. The advantages of this finding are based on the fact that bbp can be retrieved from ocean color observations. All this sets the base for future investigations on flow-stream variability, river plume patterns, and GSD of suspended sediment from satellite analyses.

The coupling between remote sensing data of TSM and marine currents from numerical models has finally set the stage for the understanding of the erosive-depositional processes of sediment along the coast, the recognition of areas of under-feeding of sediment and, therefore, the diagnostics of any areas of potential coastal erosion.

Publications and Presentations

  • Pitarch, J., Falcini, F., Nardin, W., Brando, V. E., Di Cicco, A., & Marullo, S. (2019). Linking flow-stream variability to grain size distribution of suspended sediment from a satellite-based analysis of the Tiber River plume (Tyrrhenian Sea). Scientific reports, 9(1), 1-10.
  • Falcini, F., Pitarch, J., Nardin, W., Brando, V., Di Cicco, A., & Marullo, S. (2019, December). Linking flow-stream variability to grain size distribution of suspended sediment from a satellite-based analysis. In AGU Fall Meeting Abstracts (Vol. 2019, pp. EP11E-2071).
  • Benincasa, M., Falcini, F., Adduce, C., Sannino, G., & Santoleri, R. (2019). Synergy of satellite remote sensing and numerical ocean modelling for coastal geomorphology diagnosis. Remote Sensing, 11(22), 2636.
  • Benincasa, M., Falcini, F., Adduce, C., Santoleri, R., & Sannino, G. (2018, October). Remote sensing and coastal morphodynamic modelling. In 2018 IEEE International Workshop on Metrology for the Sea; Learning to Measure Sea Health Parameters (MetroSea) (pp. 1-6). IEEE.
  • Falcini, F., Di Cicco, A., Pitarch, J., Marullo, S., Colella, S., Volpe, G., … & Santoleri, R. (2016, April). Remote sensing analysis of the Tiber River sediment plume (Tyrrhenian Sea): spectral signature of erratic vs. persistent events. In EGU General Assembly Conference Abstracts (pp. EPSC2016-8466).
  • Falcini, F., Di Cicco, A., Pitarch, J., Colella, S., Lai, A., Marullo, S., … & Volpe, G. (2015, December). Variability of spatial patterns of total suspended matter in the Tyrrhenian Sea coasts from remote sensing data. In AGU Fall Meeting Abstracts (Vol. 2015, pp. EP21D-01).

References

  1. Retrieval of the particle size distribution from satellite ocean color observations. Journal of Geophysical Research: Oceans, 114(C9).
  2. Investigation of model capability in capturing vertical hydrodynamic coastal processes: a case study in the North Adriatic Sea. Ocean Science Discussions, 12(4).
  3. D’Emidio, M. (2012). Linking the historic 2011 Mississippi River flood to coastal wetland sedimentation. Nature Geoscience, 5(11), 803.
  4. Po River discharges: a preliminary analysis of a 200-year time series. Climatic Change, 89(3-4), 411-433.