Many of Earth's most biodiverse and biogeochemically active aquatic ecosystems-including groundwater karst systems, turbid estuaries and the deep ocean-are perpetually dark and hydraulically complex, making long-term, high-resolution monitoring technologically challenging. Conventional optical and acoustic sensors suffer rapid signal attenuation and high energy demand in these conditions. Cavefishes of the genus Sinocyclocheilus, which inhabit lightless subterranean waters, have evolved distinctive cranial morphologies-a duckbilled head, dorsal horn and hump-hypothesized to enhance hydrodynamic perception. Here we show, by combining vital staining of neuromasts with validated computational fluid dynamics simulations across a morphological series of Sinocyclocheilus species, that these structures dramatically amplify differential pressure signals (by up to 429.8%) and near-wall velocity gradients (by up to 69.2%) while extending perceptual range. Regions of maximal hydrodynamic variation predicted by the models closely match the observed distribution of canal and superficial neuromasts, revealing a clear biomimetic design principle: sensors should be positioned where flowfield gradients are strongest. These findings establish a quantitative, evolution-guided framework for optimizing artificial lateral line (ALL) sensor arrays, enabling autonomous underwater vehicles to perform energy-efficient, high-fidelity monitoring in some of the planet's most sensitive and datascarce aquatic environments. (c) 2026 The Authors. Published by Elsevier B.V. on behalf of Chinese Society for Environmental Sciences, Harbin Institute of Technology, Chinese Research Academy of Environmental Sciences. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).