Hack's law¶
An empirical power-law fit between a drainage network's longest upstream flow-path length and its contributing area over a stated measurement range.
Core Idea¶
Hack's law describes an empirical relationship between a drainage network's upstream contributing area A and the longest upstream channel or routed flow path L: L=C A^h over a stated range. Hack reported L=1.4 A^0.6 for measured streams in Virginia and Maryland using miles and square miles. The coefficient and exponent describe a fit, not constants for every basin.[^ref-daec30c640a0]
Scope of Application¶
Use the law when many network points can be paired with their contributing areas and longest upstream paths under one measurement rule. State the network, units, scale interval, and fitting method. A natural stream channel and a flow path computed from topography can support comparable length–area measurements without being the same physical channel.[ref-daec30c640a0][ref-1afd5b14d0c6]
Clarity¶
A is the area draining to a selected point; L follows the longest upstream path rather than the straight-line distance or total length of all channels. C changes when units change. h is estimated from measurements within a chosen range. A single basin does not establish h, and deviations can limit where the fit applies.[ref-daec30c640a0][ref-759e5fb40ccc]
Manages Complexity¶
The fitted relation summarizes many paired network measurements with a coefficient and exponent. It makes comparisons possible while keeping the underlying path definition and observed deviations visible. Higher map resolution alone does not guarantee that approximate scaling extends indefinitely.[^ref-759e5fb40ccc]
Abstract Reasoning¶
Define the drainage or routed-flow network, calculate each point's upstream area and longest upstream path, then estimate C and h across a declared interval. Examine departures from the fit before comparing a second network. In Cheraghi's laboratory experiment, h≈0.54–0.60 came from ratios of conditional upstream-length moments over a finite interval, not a simple regression of every L on A.[ref-daec30c640a0][ref-1afd5b14d0c6]
Knowledge Transfer¶
Hack's mapped streams and an unchanneled laboratory sediment surface ask the same measurement question. The laboratory team scanned the surface and used D8 routing to infer paths. That transfer concerns the area–path statistic, while the network's physical carrier and extraction method differ. A river-delta study calls a distributary-length versus nourishment-area relation apparent Hack's law; nourishment area is not an upstream tributary catchment.[ref-daec30c640a0][ref-1afd5b14d0c6][^ref-c94287567b45]
Example¶
Hack measured Virginia and Maryland stream localities, including the tributaries above each locality, and reported a regional power fit. Cheraghi and colleagues studied a 2 m × 1 m sediment surface under rainfall with no visible rills, then derived flow paths from scans. Both pair contributing area with longest path, with different carriers and bounded fitted results.[ref-daec30c640a0][ref-1afd5b14d0c6]
Neighborhood in Abstraction Space¶
Hack's law sits in a sparse region of the domain-specific corpus (98th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Unclustered & Miscellaneous (2551 abstractions)
Nearest neighbors
- Watershed — 0.80
- Drainage system (geomorphology) — 0.78
- Mesohabitat Simulation Model — 0.75
- Network Communication Efficiency — 0.74
- Playfair's law — 0.74
Computed from structural-signature embeddings · 2026-10-08
Not to Be Confused With¶
Hack's law does not require h=0.6 everywhere and is not any length–area scatterplot. Stream Power Law models incision from drainage area and channel gradient; Hack's law describes geometric path length. The approved DAG placement is an unparented domain-specific root because the available broader Prime identities impose conditions this context-dependent fit does not establish.[ref-daec30c640a0][ref-4943e8fe9533][^ref-759e5fb40ccc]
References¶
[^ref-daec30c640a0]: John T. Hack, Studies of Longitudinal Stream Profiles in Virginia and Maryland (1957), DOI 10.3133/pp294B, U.S. Geological Survey Professional Paper 294-B, printed p. 47 for A and L definitions, printed p. 63 for Eq. 3 and the regional L=1.4 A^0.6 fit, and printed p. 64 for Figure 25 in Relation of Stream Length to Drainage Area. https://pubs.usgs.gov/pp/0294b/report.pdf
[^ref-1afd5b14d0c6]: Mohsen Cheraghi et al., Catchment Drainage Network Scaling Laws Found Experimentally in Overland Flow Morphologies (2018), Geophysical Research Letters 45, 9614–9622, §2 printed p. 2 for the laboratory setup, §3 printed pp. 3–4 and Figure 4 for the conditional-length-moment analysis and finite exponent range. DOI 10.1029/2018GL078351. https://infoscience.epfl.ch/server/api/core/bitstreams/c90ded1d-cb95-4f82-9eee-80f0b046c95f/content
[^ref-759e5fb40ccc]: Peter S. Dodds and Daniel H. Rothman, Geometry of River Networks I: Scaling, Fluctuations, and Deviations (2001), Physical Review E 63, 016115, abstract for restricted approximate scaling and deviations not removed by resolution. DOI 10.1103/PhysRevE.63.016115. https://arxiv.org/abs/physics/0005047
[^ref-c94287567b45]: Dong et al., Apparent Hack's Law in River Deltas (2026), Science 392, 493–498, publisher/PubMed abstract for global apparent length–nourishment-area scaling and local uniform/composite pattern differences. DOI 10.1126/science.ady6805. https://pubmed.ncbi.nlm.nih.gov/42060736/
[^ref-4943e8fe9533]: Gregory E. Tucker and Kelin X. Whipple, Topographic Outcomes Predicted by Stream Erosion Models: Sensitivity Analysis and Intermodel Comparison (2002), DOI 10.1029/2001JB000162, Journal of Geophysical Research: Solid Earth 107(B9), 2179, §2.1 for the generalized stream-power incision law with drainage area and channel gradient. https://doi.org/10.1029/2001JB000162