Deployment Notes: Goalanda, Bangladesh, August 2023
We hit the water just before dawn, the air thick with a humidity that feels like a wet blanket. The Padma doesn't just flow here; it heaves. Standing on the deck of the survey boat near the Goalanda confluence (23.6°N, 89.5°E), the river looked more like liquid chocolate than water. This is the peak of the August monsoon crest, and the sheer volume of Himalayan silt being shoved toward the Bay of Bengal is staggering. It is a violent, churning slurry of organic debris and sediment that makes any one-dimensional measurement feel like a guess.
The conditions were chaotic. The current was screaming past us, and the water surface was a mess of eddies and boils. We were operating in a hydrodynamic environment where discharge can exceed 60,000 m³/s. For most engineers, this is a nightmare. For me, it is the only way to truly test if a sensor can handle a high suspended sediment concentration (SSC) without choking. The wind was pushing against the flood surge, creating a choppy, unpredictable surface that made positioning the ADCP a fight against the river itself.
What We Found
The data was wild. We saw apparent current speed jumps of 0.4 m/s in a matter of minutes. This wasn't a fluke or a sensor glitch; it was the salt wedge. As the freshwater surge from the Ganges slams into the denser saline waters pushing up from the coast, it creates a sharp halocline. This density interface refracts acoustic signals. It literally bends the beam. If you aren't accounting for the sound speed variation caused by these salinity gradients and temperature swings—which can shift 5-8°C over a few kilometers—your velocity calculations are essentially fiction.
Then there is the bathymetry. I've spent years in the field, but the Padma is uniquely volatile. We found deep pockets of 20 meters sitting right next to submerged sandbars that barely reached 3 meters. These features aren't permanent. They shift weekly. I suspect a channel we recorded as deep on Monday could be two meters shallower by Friday after a heavy rain event upstream in India. The river is a pulsing artery, scouring the bed and relocating tons of sediment in a single tidal cycle. It makes fixed-point monitoring a fool's errand because you are chasing a moving target in a river that actively wants to eat your equipment.
The most striking observation was the non-linear vertical velocity profile. In most rivers, you can extrapolate a bit from the surface. Not here. The shear stress at the Jamuna-Padma confluence is so intense that surface-float measurements are mathematically irrelevant. We saw 'dead zones' in the lee of shifting islands where the water was nearly stagnant, while just fifty meters away, the main thread of the current was ripping through the channel. This spatial variability is extreme. You cannot simply average the flow across a cross-section and call it a day.
Equipment Performance
We used a variety of transducers, and the results were telling. Standard flow meters were nearly useless here. The high SSC creates massive backscatter. Without aggressive gain adjustments, the signal saturates instantly. I call it the 'white-out' effect. The transducer cannot distinguish between the actual water column movement and the dense cloud of particles moving with it. Honestly, the 600kHz unit outperformed the higher-frequency models. It penetrated the slurry better and gave us a cleaner signal, though we still struggled with bin contamination near the bed. We had to discard about 15% of the bottom-most cells to avoid the noise created by the heavy bed-load transport. If you don't filter that noise, you're just measuring the movement of rolling pebbles, not water.
Recommendations for Future Deployments
If you are heading into the deltaic plains of Bangladesh, don't trust your default factory settings. You need to be aggressive with your configuration to survive the monsoon pulse.
- Dynamic Gain Control: Manually adjust gain settings to prevent signal saturation in high-silt zones.
- Sound Speed Correction: Use a real-time CTD (Conductivity, Temperature, Depth) sensor to correct for halocline refraction.
- Frequency Selection: Opt for lower-frequency transducers (e.g., 600kHz) to minimize attenuation in turbid water.
- Frequent Ground-Truthing: Re-verify bathymetry every 48 hours; the bed shifts too fast for old charts to be useful.
- Robust Mounting: Use heavy-duty bottom mounts with reinforced cabling to prevent scouring from ripping the unit out of the sediment.
Field report by Dr. Alistair Vance. Dr. Vance is a specialist in underwater acoustics and salt wedge modeling with over twenty years of experience in estuarine instrumentation.
Field Deployment Report: Velocity Profiling at the Goalanda Confluence