ADP Versatility in San Felipe, Mexico Deployment

Introduction

This report presents data from the deployment of a SonTek Acoustic Doppler Profiler (ADP) in the Gulf of California near San Felipe, Mexico on April 4 and 5, 1996. The instrument was deployed from a small boat working with the R/V Francisco Ulloa, operated by CICESE (Centro de Investigacion Cientifica y de Educacion Superior de Ensenada) in Ensenada, Mexico. The scientist coordinating the deployment of the ADP was Luis Gustavo Alvarez from CICESE; Craig Huhta and Matt Curry from SonTek were present to oversee instrument operation and assist in deployment and recovery.

This data report is divided into the following sections.

• Deployment Location and Configuration - A description of the ADP, instrument configuration, deployment location, and data collection parameters.
• Tidal Elevation - ADP surface level measurements used to estimate tidal elevation.
• Water Currents - ADP velocity data including mean current speed and direction, contour plots showing temporal and spatial variation, and current change at slack water.
• Near Surface Velocity Measurements - A detailed look at the ability of the ADP to make measurements near the boundary.
• Signal Strength - Signal strength data from the ADP, and their relation to suspended sediment.

Deployment Location and Configuration

A 1.5-MHz Stand-Alone ADP was deployed for 16 hours starting at 2:40 p.m. (Mountain Standard Time) on April 4, 1996. The deployment site was at 31º 07.88’ N, 114º 40.96’ W, approximately 16 km northeast of San Felipe, Mexico. Water depth at high tide was about 18.5 m; tidal variation during the deployment was more than 5 m. Figure 1 shows a map of the area and the deployment site.

Figure 1. San Felipe ADP Deployment Location

The instrument included a 20-MB internal recorder, compass and 2-axis tilt sensor, temperature and pressure sensors. The internal recorder was used as memory buffer; data were collected on a portable computer using the ADP real-time data acquisition software. The compass and 2-axis tilt sensor allowed the instrument to report velocity data in earth coordinates (East-North-Up) regardless of the orientation of the instrument. The temperature sensor was used for sound speed corrections (handled automatically by the ADP); the pressure sensor was used to estimate deployment depth and tidal variations.

Figure 2 illustrates the deployment configuration. The ADP was mounted on a small platform using lead shot as ballast; the frame was deployed on the bottom with the instrument looking upwards. A small vessel was secured nearby using a 3-point anchor; a multi-conductor cable from the vessel to the instrument supplied power to the ADP and real-time communication. The primary deployment and recovery cable was attached to a surface float located 30-40 m away; a secondary (safety) recovery line was run to the monitoring vessel.

Figure 2. ADP Deployment Configuration

Power was supplied from a marine battery with an inverter supplying 110 VAC. The battery was changed after 15 hours; a data gap of about 4 minutes occurred during the change. The deployment site was chosen to match previous measurements in this area and to provide a water depth maximizing the effectiveness of the ADP. With a maximum water depth about 18 m, the ADP measured the entire velocity profile with a vertical resolution of 0.5 m. The ADP recorded 50 cells per profile, using a 0.5-m cell size, ensuring that the recorded profile included the entire water column and surface reflection; the ADP recorded the mean velocity profile each 5 minutes. Velocity data were recorded in earth (East-North-Up, relative to magnetic north) coordinates; at this location, magnetic north is 11º 46’ east of true north.

Tidal Elevation

The ADP used for this deployment has two means to estimate distance to the surface, and hence total water depth. The first method is the optional pressure sensor installed in the ADP. While not intended for precise tidal measurements, the pressure sensor provided excellent resolution of the large tidal variations in this deployment. The second method to estimate surface elevation uses the profile of return signal strength. Sample plots of ADP signal strength versus range are shown in Figure 3; in this plot, the x-axis gives range from the transducers (in m) and the y-axis gives return signal strength (in internal units called counts).

Figure 3. Measuring Surface Range using the ADP Signal Strength Profile

Range to the surface is estimated from the location of the spike in signal strength data associated with the reflection of the acoustic pulse from the surface. For the profile shown in Figure 3, this peak occurs in a cell whose center is located 18.4 m from the ADP transducers. Since the transducers are 0.3 m off the bottom, this gives an estimated water depth for this profile as 18.7 m; the corresponding estimate from the pressure sensor was 18.2 m.

Figure 4 compares estimates of surface elevation from the pressure sensor and signal strength over the entire deployment. Data collection began just after high tide, continued through the next low and high tides, and ended just before the following low tide. The minimum water depth is estimated at 13 m, maximum depth is about 18.5 m. Data from the pressure sensor and signal strength showed very consistent results, with a mean offset of about 0.2 m (the pressure data consistently estimate lower water depth than signal strength). The expected accuracy of the pressure sensor is ±0.5 dBar (equals ±0.5 m), while the accuracy of the signal strength is equal to the cell size (0.5 m).

Figure 4. Estimated Tidal Elevation

Water Currents

The primary interest in ADP data is the measurement of water velocity. This section shows several different presentations of velocity data to illustrate the information contained in the ADP data and to highlight interesting features in this data set.

The deployment covered two periods of maximum ebb and one of maximum flood current. Current speeds near the surface reached a maximum of 75 cm/s; mean currents over the entire water column reached 60 cm/s at maximum flow. Figure 5 and Figure 6 show plots of the mean current speed and direction over the course of the deployment. The mean velocity values are taken as all cells in the profile to within approximately 1 m of the surface (see the next section for a discussion of near surface data). The center of the first cell is located 1.2 m above the bottom; the last good cell is specified such that its center is at least 1 m below the surface level estimated by the ADP pressure sensor.

Flood current is shown as positive while ebb is negative. Ebb current showed a direction estimated as 133°, while flood current showed a direction of 315°. There was no variation of current direction over the entire water column, nor was there any significant variation in flow direction during ebb or flood. See the vector plots later in this section for details about the current change at slack water.

Figure 7 is a color contour plot of current speed versus time and depth. This presents the temporal and spatial variations in water speed over the entire deployment. The color shown at each depth cell and each time indicates current speed by the scale on the right side of the plot. No editing was done to the current data for this plot; the values were filtered using a 1-cell, 4-point linear interpolation to improve the color presentation. (The value shown at each cell is an average including data from the same profile in adjacent cells, and the same cell in adjacent profiles.) The variation in the height of the contour plot is based upon the surface elevation with changing tide. In addition to the mean tidal fluctuations, we see significant variations in velocity shear during the deployment. These variations in shear may be related to changes in wind speed, and the relative direction of wind and current.

Figure 7. ADP Tidal Current Speed Contour

One interesting feature in the current data is the behavior during slack tide. Figure 8 and Figure 9 show vector plots of the water current at each period of slack tide. These plots show an arrow for each data point, with the length and orientation of the arrow determined by current speed and direction. For clarity, these plots were generated using every other profile and every other range cell (no averaging or filtering done). Particularly in the first plot (at low tide), we see that the bottom water changes direction considerably sooner than the water at the top of the profile. A similar effect, although less pronounced, is seen in the current change at high tide.

Near Surface Velocity Measurements

One special consideration when using a current profiler in shallow water is the validity of near boundary data; how close to the surface (or bottom) is the instrument able to make accurate measurements? There are two sources of interference: the reflection of the acoustic pulse (along the axis of the beam) from the boundary, and side lobe interference. The first occurs when the center of the acoustic beam hits the boundary; the region affected by this interference is a function of the length of the acoustic pulse. The ADP uses an acoustic pulse size equal to the depth cell size. For a stationary boundary, direct pulse reflection will affect the last two cells immediately before the boundary; the center of the last good cell is located two cell sizes from the boundary. For this deployment, using 0.5-m cells, we expect the center of the last good cell to be located 1.0 m from the surface. When waves are present, this distance will be measured from the lowest water level.

While ADP transducers are constructed to concentrate the majority of the acoustic energy in a narrow cone (the half power beam width of the ADP is 1.4°), some energy is transmitted in all directions. Some portion of this energy will travel a direct path to the boundary; the reflection of this "side lobe" energy can contaminate velocity data while the axis of the beam is still some distance from the boundary. The region potentially affected by side lobe interference is determined by the beam-mounting angle. ADP transducers are mounted 25° off vertical so side lobe interference may affect the last 10% of the velocity profile (this is 10% of the profile below the cells affected by the direct reflection of the pulse). The ADP transducers have been designed to minimize side lobe levels, thus interference may not be a significant factor depending upon the acoustic conditions. Without an external reference for velocity measurements, there is no absolute means to determine the presence of side lobe interference.

Figure 10 and Figure 11 show the mean current speed and signal strength profiles from two different periods during this deployment. In each plot, current speed (in cm/s) is plotted using a "*", while signal strength is plotted as signal-to-noise ratio (in dB) using a "o". Figure 10 is an average of 15 minutes of data during ebb flow shortly after high tide. Figure 11 shows the average of 15 minutes of data during ebb flow just before low tide.

The profile of SNR decays with distance from the transducers, with a large spike corresponding to the reflection from the surface. The location of the peak of this spike corresponds to the range to the mean surface level. The current profiles show the measured current speed at each cell; these estimates show large variations starting just before the surface reflection.

From the SNR profiles in Figure 10 and Figure 11, we see increased signal strength in two cells below the surface peak; this is caused by the direct reflection of the acoustic pulse from the surface, and we expect the velocity data from each of these cells are contaminated. There is a potential for side lobe interference to affect 10% of the profile before these cells; thus, the previous 2-3 cells may also see interference.

In the data from high tide, the third cell from the peak shows significant interference with a velocity estimate of 22 cm/s compared with 48 cm/s seen in the main portion of the profile. The contamination in this cell may be caused by a lower effective water level (due to waves), or by side lobe interference. The next two cells in the profile show a decrease of 5-10% (44-45 cm/s versus 48 cm/s) from the rest of the profile. This decrease in velocity may reflect the true motion of the water (decrease in surface velocity caused by the effect of wind and waves) or may indicate limited side lobe interference. Without an external reference, it is not possible to determine whether these cells reflect the true water motion. Based upon data parameters from the ADP, we would discard the data from the three cells prior to the surface peak, and would mark the next two cells as suspect for minor interference.

For the profile taken at low tide, there is similar, although far less pronounced, bias in velocity data in two cells prior to the surface peak; these cells show an increase of 5-10% over the main portion of the profile. There is nothing in the data for cells below these to indicate any surface reflection or side lobe interference. We would discard the two cells prior to the surface as contaminated by direct reflection, and would consider the remainder of the profile to reflect true water motion.

Based upon the observations above, we have set a standard editing criteria for determining the last good cell within each profile. For all profiles, we would discard the two cells prior to the surface as contaminated; in some cases, we have also observed contamination in the third cell. For this report, we have discarded any cell whose center is located within 1 m of the surface level as estimated by the ADP pressure sensor. While this criteria leaves a small amount of potentially contaminated data, we feel this is best for preliminary analysis; additional editing, if necessary, can be done at a later time.

Signal Strength

ADP signal strength is a measure of the strength of the acoustic reflection from the water. In shallow water applications (using high acoustic frequencies), this return is normally dominated by reflection from suspended sediment. The relationship between signal strength and sediment concentration is a function of sediment size, type, and concentration. When comparison measurements of sediment concentration are available, signal strength can be converted to concentration with a reasonably high degree of accuracy and excellent spatial and temporal resolution.

Several processing steps are required when analyzing ADP signal strength data: converting internal units, correcting for range losses, and setting a relative scale value. The first step involves multiplying the ADP signal amplitude data (in internal units called "counts") by 0.43 to convert to dB. Next, the data must be corrected for the effects of geometric spreading and absorption to allow comparison of data from different portions of the profile. Spreading and absorption cause a decay in signal strength with increasing range from the transducers. This decay can be predicted by the following formula.

DECAY = -20 * log₁₀(R_beam) - 2 * a * R_beam

where:

DECAY = decrease in signal strength as a function of range (in dB)
R_beam = along beam range (equals the vertical range divided cos(25°))
a = sound absorption (for 1.5 MHz at salinity 35 ppt = 0.68 dB/m).

To make signal strength data independent of range, we subtract the range correction for each depth cell from the signal strength (after converting signal amplitude to dB). The final step is to choose a reference level; for this report, we have set the lowest values of range corrected signal strength to 0 dB, with all other values shifted relative to this. The choice of a reference level is arbitrary, as the relative scale is most interesting for analysis.

With precise measurements of transducer output power and receive sensitivity, it is possible to provide an absolute measure of the return signal strength. This absolute measure is referred to as volume scattering strength; it is measured as the strength of the return reflection relative to the strength of the incident signal. Calibration allows the comparison of acoustic from multiple instruments. This type of calibration is difficult and typically results in an overall accuracy of no better ±3 dB; it is useful only in situations utilizing multiple acoustic instruments and precise calibration of the acoustic signal.

Figure 12 shows a contour plot of signal strength from the ADP over the entire deployment. As with the velocity contour, the scale on the right shows the signal strength as plotted for each depth cell. The ADP data have been corrected for range losses and converted to an arbitrary reference as described above. As with the velocity contour plot, data were filtered using a linear interpolation from each adjacent point.

Figure 12. ADP Signal Strength Data Corrected for Range Loss

Figure 12 shows more than 20 dB variation in corrected signal strength during the deployment. Increased signal strength reflects greater sediment concentrations towards the bottom at all times during in the deployment, and increased sediment concentrations through most of the water column during periods of high current. We also see increased signal strength throughout the water column towards the end of the deployment (during a period of high wind and waves). The increased signal strength at the top of the water column may reflect sediment suspension or bubbles caused by wind and breaking waves.

ADP signal strength can be theoretically related to sediment concentration only in certain situations. Sediment must be the only significant source of acoustic scattering, and the size distribution of sediment must be assumed to be constant. Under these conditions, signal strength is directly proportional to sediment concentration; an increase in signal strength of 3 dB will correspond to an increase in sediment concentration by a factor of 2 (10*log₁₀(2) = 3 dB). When analyzing signal strength data, it is important to remember there may be other sources of scattering, particularly biological creatures or suspended air bubbles. If comparison measurements of sediment concentration are available, signal strength can often be related to sediment concentration even with non-uniform particle sizes and types.

During this project, hourly profiles of suspended sediment concentrations were taken using an optical backscattering sensor (OBS). Figure 13 shows sediment concentration as measured by the OBS versus signal strength data from the ADP. The comparison is made by matching a single ADP profile (closest in time to the suspended sediment profile) and depth cell (closest in depth to the estimated location of the OBS measurement) to the suspended sediment data. Figure 13 shows a scatter plot of all comparisons available over the entire deployment period. We see excellent correlation between the data at higher concentrations, with larger variations at lower levels. The larger variations at lower levels have many potential sources including the presence of non-sediment scattering material (biological matter, bubbles, etc.), and different sensitivities of the different measurement techniques.

Figure 13. ADP Signal Strength Calibration using Preliminary Concentration Data

The mean slope of the correlation is considerably steeper than expected based upon the theoretical relation discussed above (i.e., a change in concentration of a factor of 10 should give a change in scattering strength of 10 dB). This is caused by changes is the particle size distribution with depth; higher concentrations are typically towards the bottom of the water column, where particle sizes will be larger. For sediment sizes smaller than about 200 mm, the 1.5-MHz ADP has significantly greater sensitivity (signal strength for a given concentration) to larger particles.

Using the comparison data shown in Figure 13, we can make a best-fit relation between ADP signal strength and suspended sediment concentration. This would allow the contour plot shown in Figure 12 to be converted to absolute concentration units. In this instance, the ADP provides an estimate of sediment concentration with each profile (every five minutes) and at each depth cell (every 0.5 m). Based on the scatter seen in Figure 13, these estimates of suspended sediment concentration should be accurate to approximately a factor of two.

• ADP