Rivers look like accidents of geology: water finds a path downhill, carves a channel, and the result is a sprawling, irregular network of tributaries. Yet hidden in that apparent chaos is one of the most persistent numbers in geomorphology. Across continents, climates, and rock types, the length of a river scales with the area it drains to the power of about 0.6. The relationship, known as Hack’s law, has held up since it was first measured in 1957, and no one has fully explained why. A study published this spring on the cover of Science found the same number operating in river deltas, where water does the opposite of gathering and spreads out.
The finding closes a symmetry. In a river basin, tributaries converge: many small streams merge into fewer large ones. In a delta, distributaries diverge: one river splits into many channels carrying sediment toward the sea. The two networks are mirror images of each other. The new analysis, led by Tian Dong of the University of Texas Rio Grande Valley with colleagues at UC Irvine, UT Austin, and the University of Zaragoza, measured more than 30 deltas from satellite imagery, roughly 6,000 points in total, distinguishing land from water. They found that the length of a delta’s channels scales with the area supplying them sediment to the power of 0.60, statistically indistinguishable from the 0.6 of Hack’s law.
A law discovered in the 1950s
John Hack, a USGS geologist, noticed the relationship while studying stream profiles in Virginia and Maryland. Plot stream length against drainage area and the points fall on a curve whose slope is about 0.6, regardless of the local geology. The exponent was surprising from the start. If basins were self-similar, the expected value would be 0.5, the square root of area. A value of 0.6 means something subtler: large basins are systematically elongated relative to small ones. MIT geophysicist Daniel Rothman has described small basins as short and squat and large basins as long and thin.
The elongation matters for how landscapes are organized. Long, thin basins pack together more tightly than round ones, and they give rivers a directionality that points toward the sea. The effect is visible in the shape of continents: the largest river basins are the most stretched.
What produces the 0.6? Two families of explanation have competed for three decades. The first, developed in the early 1990s by Andrea Rinaldo and colleagues, holds that river networks are optimal channel networks: among all possible branching patterns, the ones that dissipate the least frictional energy win, because inefficient arrangements erode away over geological time. Simulations showed that networks obeying the 0.6 exponent indeed dissipate less energy than alternatives. The second explanation works from the bottom up. In landscape evolution models, water flows downhill and erodes rock; channels that happen to capture more runoff erode faster, deepen, and attract even more water, a positive feedback that lets a few winners swallow their neighbors. Over thousands of years, the surviving network converges on the same statistical shape. Gravity supplies the energy; erosion dissipates it; the laws of probability do the rest.
The two explanations are not mutually exclusive, but they are not identical either, and the field has never fully settled which is fundamental. Hansjörg Seybold of the Austrian Academy of Sciences, who has spent years trying to derive the exponent from first principles, describes Hack’s law as the field’s big open question.
The same number, backwards
Deltas are a genuinely new testing ground. Their channel networks form under the opposite dynamics: instead of runoff converging, sediment-laden water diverges and deposits. The fact that the scaling exponent matches tributary networks suggests that whatever mechanism produces 0.6 is not tied to the direction of flow or the details of erosion, but to something deeper about how branching networks fill space.
The study’s authors are careful about the word apparent in their title. The scaling holds globally, but composite deltas show a break: near the apex, channels fill space densely; near the coast, they become quasi-linear. The local geometry is not uniform, even though the global average obeys Hack’s law.
Angles of a pentagon
The same research community has found other signs that river networks are governed by surprisingly clean mathematics. In 2012, Devauchelle, Petroff, Seybold, and Rothman predicted that when a stream bifurcates, the two branches should open at an angle of about 72 degrees, exactly one-fifth of a circle, and confirmed it in measurements of thousands of junctions in a Florida groundwater field: 71.9 degrees, plus or minus 0.8. Fivefold symmetry cannot tile the plane, a fact that links the angles to quasicrystals, the exotic atomic arrangements with pentagonal order that never quite repeat. Whether the connection is a genuine deep analogy or a coincidence remains speculative, but the number keeps appearing.
Not every river obeys these rules perfectly, and the caveats matter. Hack’s exponent varies across basins, from about 0.45 to 0.7, and some large basins show a smaller exponent at continental scale. A 2018 study argued that basin shape is controlled mainly by the law’s coefficient rather than its exponent, complicating the picture. Climate, tectonics, vegetation, and human engineering all distort the idealized statistics.
Still, the new delta result extends a remarkable pattern: a single exponent, 0.6, describing how branching networks of water scale, whether they gather or spread. It is the kind of number that makes geomorphology feel, at moments, like a branch of mathematics that happens to be written in stone and sediment. And it gives researchers a practical tool: if delta size follows predictable scaling, then predicting how deltas will respond to sea level rise, or how much land they can build, becomes a problem with a rule of thumb, not just a case study.
Sources
1. Tian Dong, Lawrence Vulis, Hongbo Ma, Alejandro Tejedor, Timothy A. Goudge, et al., “Apparent Hack’s law in river deltas,” Science 392(6797): 493-498, April 30, 2026. DOI: 10.1126/science.ady6805.
2. Natalie Wolchover, “Why are rivers so mathematical?,” Quanta Magazine, August 10, 2026. https://www.quantamagazine.org/why-are-rivers-so-mathematical-20260810/
3. J.T. Hack, “Studies of longitudinal stream profiles in Virginia and Maryland,” USGS Professional Paper 294-B, 1957. DOI: 10.3133/pp294B.
4. O. Devauchelle, A.P. Petroff, H.F. Seybold, and D.H. Rothman, “Ramification of stream networks,” PNAS 109(51): 20832-20836, 2012. DOI: 10.1073/pnas.1215218109.
5. Science (ScienceAdviser), “Hacking the shape of deltas,” May 5, 2026. https://www.science.org/content/article/scienceadviser-hacking-shape-deltas
6. UC Irvine, “Researchers uncover hidden rules governing growth of river deltas,” May 28, 2026. https://engineering.uci.edu/news/2026/5/researchers-uncover-hidden-rules-governing-growth-river-deltas