The Salinity-Driven Acoustic Refraction of the Mahakam Estuary
A fluvial velocity of 1.2 m/s during the Northwest Monsoon is a beast. But in the Mahakam River around Samarinda, the real enemy isn't the current speed; it's the density stratification. I've observed salinity gradients that shift violently within a single tidal cycle, creating a salt wedge that pushes deep into the freshwater runoff. This creates a stratified environment where sound velocity changes rapidly over a few vertical meters. If you ignore this, your ADCP data is essentially a guess.
Most technicians treat the speed of sound as a constant 1,480 m/s. In Samarinda, that's a rookie mistake. The interaction between the Makassar Strait's saline intrusion and the river's discharge creates a refractive lens effect. Acoustic pings don't travel in straight lines here; they bend. This leads to massive errors in bin depth and velocity calculations. I've seen discharge estimates swing by 15% simply because the operator failed to perform a proper CTD (Conductivity, Temperature, Depth) cast to calibrate the sound velocity profile.
The turbidity adds another layer of chaos. The water looks like chocolate milk. This high suspended sediment concentration (SSC) doesn't just block light—it absorbs acoustic energy. We call it signal attenuation. In the Mahakam, the attenuation coefficient is volatile. You might get a clean signal at the surface, but by the time the ping hits the benthos, the return is buried in noise. It makes ground-truthing the bottom depth a constant struggle.
The Bathymetric Volatility of the 0.5° S to 1.0° S Corridor
The stretch between 0.5° S and 1.0° S is a bathymetric nightmare. I've mapped sections where the riverbed drops from 5 meters to 22 meters over a distance of barely thirty meters. This isn't a stable channel; it's a shifting landscape of silt and scour holes. The thalweg—the deepest part of the channel—moves unpredictably. I remember one deployment where the main current axis shifted 15 meters in a single semi-diurnal tidal cycle. Fixed-point monitoring is useless here because the river literally moves beneath the sensor.
Tidal ranges fluctuate between 0.5m and 1.5m. While that sounds modest, the volumetric displacement in a channel this wide is staggering. The semi-diurnal regime forces a dense saline layer upstream, which then 'sloshes' back toward the Makassar Strait. This creates a complex shear zone. We see velocities drop to 0.3 m/s during the dry season, but the tidal influence remains dominant, often reversing the flow direction entirely in the lower water column while the surface continues to push freshwater seaward.
Acoustic Propagation Challenges in This Environment
Optical sensors are dead on arrival in Samarinda. I've seen turbidity meters bottom out within minutes of deployment. Laser-based flow meters? A waste of budget. The suspended silt is so dense that the mean free path of a photon is negligible. We rely entirely on acoustics, but even that is a fight. The primary technical battle is bin contamination. When the vertical resolution isn't tight enough, the high-velocity surface layer leaks into the lower bins, skewing the volumetric discharge total.
Temperature also plays a role, though salinity is the primary driver of sound velocity errors here. The tropical heat of East Kalimantan keeps surface waters warm, but the incoming salt wedge is cooler and denser. This temperature-salinity coupling bends the acoustic beam. Without a real-time sound velocity profile, you're just guessing where your pings are actually hitting. I've found that relying on the ADCP's internal temperature sensor is insufficient. You need an external SV-probe to get a sanity check on the data.
600kHz Frequency Selection and Deployment Logic
I chose a 600kHz ADCP for these surveys. Some of my colleagues pushed for 300kHz to get deeper penetration, but that was a mistake. In the Mahakam, 600kHz is the sweet spot. It provides the vertical resolution necessary to isolate the salt wedge's boundary layer. If you go too low in frequency, the bins are too wide, and you lose the ability to detect the shear between the freshwater and saline layers. The 600kHz unit gave me a clean signal in the upper 30 meters, which covers the vast majority of the flow profile in the Samarinda reach.
Deployment is where most people fail. You can't just drop a transducer and hope for the best. We used a stabilized mounting frame to minimize vessel motion noise. Even a slight pitch or roll in these currents can introduce artificial velocity components into the data. I prefer a towed configuration with a heavy sinker to keep the transducer vertical, though the shifting thalweg means you have to constantly adjust your transects to ensure you're capturing the full cross-section of the flow.
Data Interpretation and Field Findings
The data reveals a river in constant conflict. During the peak Northwest Monsoon (December to March), the fluvial transport is dominant. We recorded peak velocities of 1.2 m/s, but the flow is remarkably non-uniform. The center of the channel carries the bulk of the volume, while the margins are plagued by eddies and recirculating zones. When we plotted the discharge curves, the asymmetry was glaring. The ebb tide doesn't just reverse the flow; it accelerates the freshwater runoff, creating a 'flush' effect that clears out some of the suspended silt (temporarily).
Conversely, the dry season data is a mess of tidal oscillations. We saw the flow drop to 0.3 m/s, at which point the saltwater intrusion becomes the primary driver of the river's hydraulics. The 'sloshing' effect I mentioned earlier creates a zero-velocity point—a null zone—that migrates up and down the river. Mapping this null zone is critical for understanding how pollutants or sediments move through the Samarinda region. If you ignore the tidal phase, your average daily discharge will be wrong by a significant margin.
Operational Implications for East Kalimantan
These findings have immediate consequences for local infrastructure. The shifting thalweg means that dredging schedules in Samarinda are often based on outdated data. If the riverbed is moving 15 meters in a tidal cycle, a dredge map from three months ago is a fantasy. For flood monitoring, the interaction between the monsoon surge and the high tide is the danger zone. When the saltwater wedge is pushed back by a massive freshwater pulse, the resulting turbulence can destabilize riverbank reinforcements.
From a monitoring perspective, the Mahakam demands an active approach. You cannot 'set and forget' equipment here. The sediment load is so aggressive that transducers can get fouled or pitted over time. Frequent calibration and manual sound-velocity corrections are the only way to maintain a reliable dataset. For any agency attempting to manage the water levels in Samarinda, the lesson is simple: trust the CTD, verify the thalweg, and never assume the speed of sound is constant.
About the author: Dr. Kenji Sato. A specialist in underwater acoustics with over 20 years of experience designing instrumentation for high-turbidity estuarine environments. He focuses on the intersection of acoustic signal processing and fluvial geomorphology.
Mitigating Sound Velocity Errors in the Mahakam River Salt Wedge at Samarinda