Hidden River in Kaziranga Reveals Surprising Rules of Water Flow

Deep inside the floodplains of Assam, where the forests of Kaziranga National Park meet the foothills of the Karbi Hills, a modest 31-kilometer river has delivered an outsized lesson in how rivers really work. The Kohora River, a southern tributary of the Brahmaputra and a lifeline for the wildlife of one of the world’s most celebrated conservation areas, has been the subject of an unusually detailed field investigation into hydraulic geometry, the science of how a river’s width, depth, and velocity respond to changes in discharge. The findings, published in the journal Discover Geoscience, show that this small Himalayan-fed stream obeys the grand theoretical rules of river behavior at individual cross-sections, yet breaks them in striking ways as water moves downstream.

Hydraulic geometry is one of the oldest and most powerful frameworks in fluvial geomorphology. Introduced by Luna Leopold and Thomas Maddock in 1953, it expresses channel width, mean depth, and mean flow velocity as power functions of discharge, each governed by an empirical coefficient and a scaling exponent. In an idealized rectangular channel, the exponents for width, depth, and velocity must sum to one, and the product of the coefficients must also equal one, a direct consequence of the continuity equation that governs all flowing water. Deviations from these tidy constraints tell geomorphologists how a particular river is adjusting to its landscape: whether it grows wider, digs deeper, or simply speeds up as more water pours through it.

To test these relationships in a monsoon-dominated tributary, researchers Snehasish Choudhury, Niranjan Bhattacharjee, and Dharitri Choudhury surveyed six cross-sections, labeled C1 through C6, along the length of the Kohora River during the pre-monsoon season of May 2022. Using measuring tapes, dumpy levels, and a current meter operated with the two-point 0.2/0.8 depth method, they recorded channel width, water depth, and flow velocity at each site, then computed discharge as the product of cross-sectional area and velocity. The river basin they studied covers roughly 55 square kilometers, receives an average annual rainfall of 2,220 millimeters during the southwest monsoon, and descends from steep hilly terrain through a narrow middle course to join the Brahmaputra floodplain. In the middle reaches the river spans 15 to 27 meters in width and reaches depths of up to 4.5 meters, while in the confined upper course it narrows to as little as two meters but can surge at velocities between 2.12 and 7.0 meters per second.

The at-a-station analysis, which examines how hydraulic variables fluctuate with changing discharge at a fixed cross-section, produced mean exponents of 0.334 for width, 0.538 for depth, and 0.134 for velocity, summing to 1.006, remarkably close to the theoretical value of one. The corresponding coefficients averaged 4.46, 0.56, and 0.40, with a product of 0.9905, again confirming internal hydraulic consistency. The standout number is the depth exponent. Because it exceeds both the width and velocity exponents, it indicates that the Kohora River primarily absorbs increases in discharge by getting deeper rather than wider or faster. This vertical adjustment is a signature of channels hemmed in by resistant banks, valley confinement, and vegetation reinforcement, conditions that prevail along much of the river’s forested course.

Beneath the averages, however, the six cross-sections told strikingly different stories. At station C1, set in the steep, confined upper basin where dense Sal and Teak forest stabilizes the valley walls, flow velocity responded strongly to discharge, with a velocity exponent of 0.7 and a correlation coefficient of 0.9 between velocity and discharge. The same site showed a strong positive correlation between depth and discharge, reflecting active downward erosion into the channel bed. Farther downstream, cross-sections C3 and C4 responded to higher flows through pronounced widening and deepening, signaling greater accommodation capacity, weaker bank resistance, and abundant sediment. Meanwhile, sections C3, C5, and C6 displayed weak and in some cases negative velocity responses, meaning that more water did not necessarily mean faster water. The researchers attribute this counterintuitive behavior to the emergence of mid-channel sand bars, rapid sedimentation, and gentle bank-side slopes in the middle and lower course, all of which spread the flow, increase friction, and damp velocity even as discharge climbs.

The downstream analysis, which compares hydraulic variables across successive cross-sections under comparable discharge, delivered the study’s most surprising result. The classical Leopold-Maddock model predicts downstream exponents of roughly 0.5 for width, 0.4 for depth, and 0.1 for velocity, reflecting the common observation that rivers grow wider faster than they grow deeper or quicker. The Kohora River defies this pattern. Its downstream exponents came out at 0.2 for width, 0.4 for depth, and 0.4 for velocity, with coefficients of 5.19, 0.57, and 0.35. In other words, as discharge increases downstream, the river deepens and accelerates far more than it widens. The team points to bank resistance, embankments, settlement encroachment, and other human pressures that constrain lateral expansion, combined with mid-channel bars that constrict the active channel and concentrate flow into narrow, faster threads. The steep upper-basin gradients, ranging from 21.93 to 39.64 degrees, inject additional kinetic energy that propagates through the downstream hydraulic adjustment.

To place these field measurements in a broader environmental context, the researchers turned to satellite remote sensing. Landsat 8 imagery with 30-meter spatial resolution, acquired on 8 May 2022 from the United States Geological Survey, was processed in ArcGIS 10.7.1 to derive two spectral indices. The Normalized Difference Water Index, which ranges from minus one to plus one and flags open water and saturated surfaces, ranged from minus 0.38 to 0.10 across the basin, with only limited slightly positive values marking confined zones of active surface flow. The lowest values corresponded to hilly forest and riparian vegetation, intermediate values traced transitional floodplains subject to periodic inundation, and the higher negative values indicated moist near-channel environments with shallow subsurface saturation. The Normalized Difference Vegetation Index, spanning 0.07 to 0.44, mapped a clear gradient of vegetation density. Bare sandbars and active channel zones scored lowest, signaling high erodibility and instability, while values between 0.28 and 0.44 marked dense deciduous riparian forest that strengthens bank cohesion and restrains lateral migration.

The coupling between these vegetation patterns and the hydraulic measurements is one of the study’s most practically valuable insights. Low-NDVI zones, where sandbars dominate and bank vegetation is sparse, correspond to wider, shallower, and less stable channel sections, while densely vegetated reaches maintain narrower, deeper, and more stable geometry. This spatial relationship demonstrates that riparian vegetation is not merely a passive backdrop to river dynamics but an active control on hydraulic geometry, modulating bank strength, sediment supply, and the very exponents that describe how the channel responds to flood flows. It also echoes a broader literature showing that dense bank vegetation narrows channels, bed vegetation raises flow resistance, and that hydraulic geometry coefficients vary regionally with Manning’s roughness, channel slope, and bank stability even when exponents remain comparatively consistent across regions.

The authors are candid about the limitations of their dataset. Only three discharge measurements were collected at each station during a single pre-monsoon survey, a constraint imposed by the small size of the channel and the fact that this seasonal tributary runs dry in winter. The derived relationships should therefore be read as short-term, site-specific characterizations rather than long-term statistical generalizations. Even so, the work establishes a quantitative baseline for a river that had previously escaped detailed hydraulic analysis, and it extends the hydraulic geometry framework to the small tributary systems of Northeast India, where most prior research has concentrated on the giants of the region, the Ganga, Brahmaputra, Godavari, and Narmada basins, whose dynamics are dominated by monsoon deluges, heavy sediment loads, and intense seismic and tectonic activity.

The practical stakes extend well beyond academic theory. The lower reaches of the Kohora River, with their gentle gradients and expanding channels, are acutely vulnerable to seasonal flooding that inundates settlements, farmland, transport infrastructure, and parts of Kaziranga National Park itself, imposing both socio-economic losses and ecological damage. The authors argue that building flood resilience in the basin will require nature-based solutions, including wetland restoration and riparian reforestation, to enhance natural flood storage, promote sediment retention, and strengthen ecosystem resilience, alongside integrated governance frameworks that prioritize community-led participatory planning and incorporate indigenous knowledge systems. In an era when climate change is intensifying monsoon extremes across the Brahmaputra basin, understanding precisely how a river like the Kohora accommodates rising water, whether by deepening, widening, or accelerating, is not an abstract exercise. It is the quantitative foundation on which flood forecasts, channel management, and the protection of one of the planet’s most treasured wild landscapes will depend.

Subject of Research: Hydraulic geometry analysis of the Kohora River, a Brahmaputra tributary in Assam, India

Article Title: Hydraulic geometry analyses of the Kohora River, Assam, India

Article References: Choudhury, S., Bhattacharjee, N., & Choudhury, D. (2026). Hydraulic geometry analyses of the Kohora River, Assam, India. Discover Geoscience, 4(1), Article 349. https://doi.org/10.1007/s44288-026-00719-8

Image Credits: AI Generated

DOI: 10.1007/s44288-026-00719-8

Keywords: hydraulic geometry, Kohora River, Assam, Brahmaputra tributary, fluvial geomorphology, flow discharge, at-a-station analysis, downstream hydraulic geometry, NDVI, NDWI, Landsat 8, Kaziranga National Park