Acoustic and optical methods to infer water transparency at the Time Series Station Spiekeroog , Wadden Sea ( southern North Sea )

Water transparency is a key indicator of optical water quality that is driven by suspended particulate and dissolved material. In this study we carried out an intercomparison of observations related to water transparency, determine correlations among the measured parameters and demonstrate the utility of both acoustic and optical tools in monitoring water transparency. The data set used here is from the operational Time Series Station Spiekeroog located at a tidal inlet of the Wadden Sea. An Acoustic Doppler Current Profiler was used to obtain acoustic measurements in the water column. Optical observations were 5 determined using a set of three radiometers above water to collect radiometric quantities and a turbidity sensor within the water column. Bio-fouling was identified as a source of anomaly in turbidity measurements. We observed significant correlations between in-situ optically measured turbidity and derived turbidity from above water color sensing and acoustic backscattering strength. These findings underline that both optical and acoustic measurements can be reasonable proxies of water transparency with the potential to mitigate gaps and increase data quality in long-time observation of marine environments. 10


Introduction
Over the last years, the importance of water quality and environmental awareness has risen for the public, scientists and policy makers (WFD, 2000;OECD, 1993;Borja et al., 2013).In order for us to better understand water quality, it is important to gather environmental information from different tools on platforms at varying temporal and spatial coverage (Zielinski et al., 2009;Pearlman et al., 2014).The information from different tools ought to be comparable for a better understanding of environmental dynamics that drive water quality.However, water quality is an abstract term that is used widely to describe the state of water according to its ecological, chemical, optical or morphological properties; each of these having a set of related variables that are used as indices.One of the key indicators for water quality is the water transparency (CRWN, 2012).Two main aspects of water clarity (Davies-Colley and Smith, 2001) or water transparency (Wilson, 2010) are the light penetration and the visual clarity (Kirk, 1988).Both are strongly affected by dissolved and suspended material within the water (Kirk, 1985) and the optical properties of the water.

1
Ocean Sci.Discuss., doi:10.5194/os-2016Discuss., doi:10.5194/os- -24, 2016 Manuscript under review for journal Ocean Sci.Published: 30 May 2016 c Author(s) 2016.CC-BY 3.0 License.Optical water transparency is determined from optical observations that involve using the human eye as a tool or methods that replicate the human eye sensing approach (Moore et al., 2009).In general, water transparency has been determined for decades as the tools needed are easy to use, fast, inexpensive and robust.Typical optical observations provide information about the light availability in the water column which can be translated into water transparency.Turbidity is one such measurements which refers to a relative index of water cloudiness influenced by the inherent dissolved and particulate material (Kirk, 1985;Moore, 1980).Remote sensing reflectance (R RS ), an essential climate variable, is a proxy for the intrinsic color of water and optically active constituents of water (Garaba et al., 2015;Garaba and Zielinski, 2013;GCOS, 2011;Watson and Zielinski, 2013).The natural color of water, driven by the optically active constituents of water and environmental conditions, can be distinguished using a standard Forel-Ule color scale.The Forel-Ule color scale numerically categorizes natural waters from 1 (indigo-blue) to 21 (cola brown), and this information can also be derived from R RS information (Garaba et al., 2014;Wernand and van der Woerd, 2010;Garaba et al., 2015).
In more recent years, acoustic backscatter has been used to estimate abundance and distributional patterns of suspended matter (Thorne and Hanes, 2002;Deines, 1999;Thorne et al., 1991).Acoustics is one of a number of technologies advancing our capabilities to probe sediment processes (Thorne and Hanes, 2002;Voulgaris and Meyers, 2004).Acoustic backscatter signals provide information about the suspended material in a given water body and enable to record the changes over a long time scale.An Acoustic Doppler Current Profiler (ADCP) measures non-intrusive and three-dimensional, making it a very powerful tool for examining small-scale sediment transport processes (Thorne and Hanes, 2002;Schulz et al., 2015).
In coastal and estuarine regions, the composition and concentration of suspended material is highly variable (Winter et al., 2007;Fugate and Friedrichs, 2002).Fragile flocculants change their characteristics over short and over long time scales due to hydrodynamic forcing conditions among them currents, turbulence and tides (Vousdoukas et al., 2011;Fugate and Friedrichs, 2002;Burchard and Badewien, 2015).Depending on whether optical methods (White, 1998;Sutherland et al., 2000) or acoustic methods (Voulgaris and Meyers, 2004;Fugate and Friedrichs, 2002) are used, the scattering properties of sediment differ.
Optical measurements are more sensitive to fine sediment, while acoustic measurements are more sensitive to coarser material, depending on the operating frequency (Gartner, 2004).Thus Winter et al. (2007) concluded that the combination of different instruments reveal different aspects of SPM dynamics.Comparisons of acoustic and optical sensors for measurements of suspended sediment concentrations performed under laboratory conditions showed that most in-water sensors have a linear response under bimodal and randomly sorted suspended sediments (Vousdoukas et al., 2011).
In this work, we therefore aim to find out if the information from optical water quality variables is related to the backscatter signal.The assumption is based on the fact that all these observations provide information about the inherent suspended particulate and dissolved material and therefore can be practical indicators of water transparency.We also evaluate to which extent combining acoustic and optical techniques are suited for environmental monitoring.We discuss additionally, whether qualitative and quantitative information gathered at operational long time series observatory platforms helps identifying indicators of change within natural waters.Comparing different tools for analysing water quality may also prove valuable for assessing precision as well as accuracy of measurements.At the same time, this approach also helps, understanding associations among the observed variables.This in turn may enable to close gaps in relevant information concerning dynamics of suspended material in the water column when individual instruments fail.
2 Materials and methods

Study area
The Time Series Station Spiekeroog (TSS, Fig. 1) is an operational multidisciplinary autonomous observatory located (53 • 45.016' N, 007 • 40.266'E) in a tidal channel between the islands of Langeoog and Spiekeroog (Badewien et al., 2009).These islands are part of an island barrier system in the East Frisian Wadden Sea, southern North Sea, which belongs to the UNESCO World Natural Heritage sites.The region is part of an extended North Sea tidal flat system with shallow water depths ranging from 0 to 20 m.The typical water depth is 13.5 m with tidal range of about 2.7 m at the site of the Time Series Station Spiekeroog (Holinde et al., 2015).Usual current velocities during ebbing and flooding periods reach values of up to 2 m/s.
The distribution of suspended particulate and organic material is strongly influenced by the semi-diurnal tidal currents as well as waves driven by wind (Bartholomä et al., 2009;Reuter et al., 2009;Badewien et al., 2009).These strong and rapid dynamics around TSS make it a valuable study area for biogeochemical and physical investigations in transitional waters from land to the open sea.

Sampling and analysis
Three radiometers (TriOS, Germany) continuously measure with a 5 minute interval, hyperspectral radiance (2 x RAMSES ARC) and irradiance (1 x RAMSES ACC) to assess remote sensing reflectance (R RS ).The radiometers are mounted to the TSS at 24 m above the seafloor (approx.12 m above NN).The radiometric quantities are corrected for environmental perturbations (Busch et al., 2013;Garaba et al., 2015Garaba et al., , 2012) ) and are transformed into Forel-Ule color indices that can be matched to the intrinsic color of water (Wernand, 2011).Detailed information on the processing of these measurements is presented in an earlier study (Garaba et al., 2014).A submerged ECO FLNTU (WETlabs, USA) sensor continuously measures turbidity at 12 m above the seafloor with 1 minute interval.The ECO FLNTU sensor samples turbidity data with optical backscattering at a wavelength of 700 nm.The sensitivity of the sensor is 0.01 N T U and the typical range of turbidity is between 0 N T U and 25 N T U .Because of problems associated with bio-fouling, turbidity measurements at the TSS might not be at all times reliable.
In order to compare optical data to other variables, data were taken from a period directly after cleaning the ECO FLNTU sensor (low water, 28 August 2013) from bio-fouling.Figure 2 shows the turbidity data in N T U over a time period from 28 August 2013 until 02 September 2013.The blue points show the in-situ raw data, limited between 0 N T U and 25 N T U .The red points represent quality checked turbidity data.These turbidity data were measured near the surface of the water column at a fixed height above the seabed, therefore these data can directly used for the comparison.
where D is the distance from ADCP to the surface in [m], and φ is the angle of the beam relative to vertical [ • ].The resulting blank space near the surface reached values between 3.0 m to 3.5 m (see Fig. 3).In order to fill this gap, we extrapolated the acceptable backscatter data from R max -depth to the sea surface layer.
In different tidal phases (flood, ebb, slack water) the backscatter signal shows different profiles over depth.Figure 4 displays that the profiles at slack water (high / low water -light blue, orange, violet) have similar shapes.Also, the profiles at maximum velocity (flood / ebb -green, blue, dark yellow, brown) show similar shapes.
To extrapolate the backscatter data to the surface layer, various curve fitting techniques using MATLAB R2015a, Curve Fitting Toolbox (MathWorks, USA) were implemented.These methods were (i) the exponential fitting method ('exp', equation: ), (ii) the polynomial fitting method ('poly', equation: p1 • x + p2), (iii) the power fitting method ('power', equation: a • x b ) and (iv) constant extrapolation (using the last reliable data value as the surface value).Comparisons between these methods are shown in Fig. 5.The top left graphic shows a profile during low water (violet in Fig. 4) and during flood (blue in Fig. 4) with different extrapolations to the surface layer.The top right graphic shows the extrapolation results.Here, the data derived in the near bottom range up to 4 m are excluded since it is highly influenced by strong currents during ebb and flood.
We assume that the distribution of suspended matter near the surface layer is homogeneous because of turbulence and wind influence.
The best results for extrapolation over different existing data sets (extrapolation with complete (wwc) and reduced water column (rwc)) are given by the exponential extrapolation ('exp').The R 2 values over the entire period are good with R 2 > 0.99 (see Fig. 5, bottom).Figure 6 shows the resulting signals at the surface layer.To compare backscatter data with the other parameters, the exponential fitting method and the extrapolations with constant values (BS Ex,exp and BS Ex,const ) were used.
The latter was used to fulfil the assumption of homogeneity in the top layer of the water column.
The acoustic backscatter signal is used to quantitatively determine suspended matter and therefore relates to turbidity (Deines, 1999;Lohrmann, 2001;Schulz et al., 2015).
One aim of this study is the inter-comparison of measurements from different tools to understand correlations among the observed variables and to mitigate gaps in relevant information concerning changes in suspended material or water transparency in the water column, e.g. when instruments fail (Oehmcke et al., 2015).Vousdoukas et al. (2011) found a linear response between SPM data obtained by optical and acoustic sensors in laboratory conditions under various suspended sediment conditions.Our early investigations estimated a linear relation between turbidity and Forel-Ule-Index (Garaba et al., 2014)  a polynomial function of second order between turbidity and acoustic backscatter signal (Schulz et al., 2015).For this study, we consider for the first time all data sets of the different measurement approaches (acoustic and optical, in-water and above water) to evaluate their utility for environmental in-situ monitoring of information on water transparency variables.In order to confirm correlations, we apply the Spearman rank correlation test.

Results and discussion
A time series of the acoustic and optical measurements from the 29.08.2013 is presented in Fig. 7. Visual inspection of the data suggested that there was a moderate correlation.We assume it is because of the position of the sensors, above the sea surface, submerged near the sea surface and submerged near the seafloor (an overview of the fields of view is shown in Fig. 8).
Additionally, the lack of a strong correlation may be due to the different scatter behaviour of suspended sediment (White, 1998;Sutherland et al., 2000;Voulgaris and Meyers, 2004;Fugate and Friedrichs, 2002) depending on whether optical or acoustic methods are applied.Optical measurements are more sensitive to fine sediment and are wavelength dependent, while acoustic measurements are more sensitive to coarser material, depending on the operating frequency (Gartner, 2004).Both depend on the amount of suspended sediment in the water column.signal.Furthermore, we presume these instruments provide a reasonable proxy for the suspended material which is comparable.
In order to confirm the correlations we applied the Spearman rank correlation test (for two time periods of different length).
Results of the Spearman rank correlation are shown in table 1.To compared the data at nearly the same sampling target, we extrapolated the backscatter signal towards the sea surface area.Two of these extrapolated variables (BS Ex,const and BS Ex,exp ) are used for the Spearman rank correlation test.The correlation coefficient between the data sets increases from moderate (ρ Spearman > 0.4 and ρ Spearman < 0.6) to strong (ρ Spearman > 0.6 and ρ Spearman < 0.8).The correlation between Forel-Ule-Index F U I and turbidity T RB is even very good (ρ Spearman > 0.8).For further investigations, we use the constant extrapolated backscatter signal BS Ex,const .
The Forel-Ule color scale is an inexpensive and longstanding tool used to determine intrinsic color of water (Garaba et al., 2015;Garaba and Zielinski, 2015).In a previous study it was shown that the Forel-Ule-Index (FUI) can be used to accurately derive turbidity (Garaba et al., 2014).We therefore evaluate its potential in providing information about suspended material, which in turn can be compared to information derived from backscatter signals (Fig. 9).We made comparisons with data obtained from different tidal phases (ebb, flood and slack water periods) and observed correlations between ρ Spearman = 0.34 (for high tide) and ρ Spearman = 0.81 (low tide) (all results are shown in table 2). Figure 10 shows a comparison between the backscatter signal and the turbidity data, which are separated into the tidal phases: ebb (green), flood (red), high waters The correlations between backscatter data BS Ex,const and turbidity T RB are weak at high tide and flood and otherwise strong (ρ Spearman > 0.6 and ρ Spearman < 0.8).

Conclusions
The results of this study show that bio-fouling decreases the data quality of in-water optical measurements of turbidity within short time periods.Hence, it is important to find an approach to improve the monitoring over time and increase the robustness of the turbidity results.This study demonstrates that bottom-mounted ADCP measurements, which are hardly influenced by bio-fouling, is a suitable alternative to overcome the problem.We found that using the backscatter signal and Forel-Ule-Index both yield reliable results, broadening the work of Garaba et al. (2014).Our results regarding the correlation between backscatter signal and turbidity agree well with investigations of Schulz et al. (2015).The linear responses between the different sensors types found in this study agrees with previous results from laboratory investigations (Vousdoukas et al., 2011).
We have also shown that data sets from different measurement principles (optical and acoustic) are comparable and complementary.This is despite of the position of the sensors, above the sea surface, submerged near the sea surface and submerged near the seafloor and the well-known fact that the scattering properties of particles derived from both methods differ (White, 1998;Sutherland et al., 2000;Voulgaris and Meyers, 2004;Fugate and Friedrichs, 2002).
Thus, these methods can be utilized as an affordable tool and from different platforms to monitor environmental processes.On a qualitative level, using the Forel-Ule-Index, derived from radiometer measurements, is a powerful tool as much as data sets derived from ADCP measurements.Our investigations also underline that long term observatories are key in understanding the marine environment because short-term studies only allow for a limited view on the considered dynamics.
Summarizing we found out that the information from optical water quality variables is related to the acoustic backscatter signal.We also evaluated the utility of acoustic and optical technology in environmental monitoring to gather qualitative and quantitative indicators of change within natural waters taking advantage of operational long time series observatory platforms.
The goals of this study was to perform an inter-comparison of measurements from different tools, to understand correlations among the observed variables, and to develop methods geared at mitigating gaps in relevant information about changes in water transparency in the water column when instruments fail.
A bottom-mounted (1.5 m above the seafloor), upward looking 1200 kHz ADCP (Teledyne RD Instruments Workhorse Sen-Ocean Sci.Discuss., doi:10.5194/os-2016-24,2016 Manuscript under review for journal Ocean Sci.Published: 30 May 2016 c Author(s) 2016.CC-BY 3.0 License.tinel, USA) estimates the current velocity using the Doppler effect in three dimensions.The ADCP is installed at a distance of about 12 m towards North-north-west from the pole of the station.We receive data over the entire water depth with a vertical resolution of 0.20 m (bin size) and a temporal resolution of 5 minutes (measurements are averaged over 45 pings in 5 minute bursts).Because the ADCP has a beam angle of 20 • and a tilted orientation with pitch ∼ 19.39 • and roll ∼ 17.96 • , the maximal range R max [m] of acceptable data is given by and Ocean Sci.Discuss., doi:10.5194/os-2016-24,2016 Manuscript under review for journal Ocean Sci.Published: 30 May 2016 c Author(s) 2016.CC-BY 3.0 License.
Badewien et al. (2009) measured a range of particle sizes from 1.25 to 26.9 µm (radius) at the same site.Therefore, mainly fine sediment concentrations are expected.A modelling study by Stanev et al. (2007) showed different sediment concentrations and dynamics for fine SPM (mud, d mud = 63 µm) and sand (d sand = 200µm) for this Wadden Sea area.Depending on the specific location, the dynamics of the different sediment types (fine or coarse) acted differently dependent on the tidal signal.The dynamics of all data sets correspond well to the observed tidal

(
dark blue), low waters (light blue).The Fig. shows also the comparison between the backscatter signal from the ADCP which Ocean Sci.Discuss., doi:10.5194/os-2016-24,2016 Manuscript under review for journal Ocean Sci.Published: 30 May 2016 c Author(s) 2016.CC-BY 3.0 License.was constantly extrapolated to the sea surface BS Ex,const versus Forel-Ule-Index F U I on the left and BS Ex,const versus turbidity data T RB on the right.Spearman rank correlation tests were also applied to signals at different tide phases (ebb, flood and slack water periods) and results are presented in table 2. The correlations between backscatter data BS Ex,const and Forel-Ule-Index F U I are weak at high tide (ρ Spearman = 0.34) and mostly moderate (ρ Spearman > 0.4 and ρ Spearman < 0.6).

Figure 1 .Figure 2 .Figure 3 .Figure 4 .Figure 5 .
Figure 1.Schematic of the Time Series Station Spiekeroog showing the position of the radiometers (24 m), turbidity meter (12 m) and ADCP (1.5 m above the seafloor).The typical water depth is 13.5 m with tidal range of about 2.7 m between mean high and low water.The insert shows the location of the Time Series Station where the colors indicate the water depth at high water.

Table 1 .
Spearman rank correlation results of the backscatter data from the ADCP BSEX,const and BSEX,exp for one day and for a longer time period (29.08.2013 -02.09.2013) as well as the estimated Forel-Ule-Index F U I and the turbidity T RB.Ocean Sci.Discuss., doi:10.5194/os-2016-24,2016 Manuscript under review for journal Ocean Sci.Published: 30 May 2016 c Author(s) 2016.CC-BY 3.0 License.

Table 2 .
Spearman rank correlation results of the backscatter data from the ADCP BSEX,const, the estimated Forel-Ule-Index F U I and the turbidity T RB -separation tidal phases.Ocean Sci.Discuss., doi:10.5194/os-2016-24,2016 Manuscript under review for journal Ocean Sci.Published: 30 May 2016 c Author(s) 2016.CC-BY 3.0 License.