Modelling African Catfish Fingerling Survivability with a Seven-Layer IoT Architecture: The Integrated Cyber-Ecological IoT Survivability Framework

Abstract

African catfish (Clarias gariepinus Burchell, 1822) is a valuable aquaculture species in Africa, but fingerling mortality driven by poor water-quality and cannibalism remains a key constraint on production. Drawing on a mixed-methods survey of twelve catfish farmers and forty-eight IoT practitioners in Kenya, this paper proposes a seven-layer Internet of Things (IoT) architecture for environmental and behavioural management of the fingerling ecosystem, integrating four complementary theories with real-time sensing and threshold-based automated actuation. The model was evaluated through a simulation-based computational experiment, comparing an automated control scenario against an uncontrolled baseline. Modelled mortality fell from 59.7% to 22.5% under the automated control scenario relative to the uncontrolled baseline. Because no live fish were reared or observed, these figures describe the model’s behaviour under author-specified virtual conditions and should be read as an upper-bound estimate rather than a demonstrated biological outcome. Multiple linear regression identified dissolved oxygen and ammonia as the strongest predictors of the modelled survivability index (R² = 0.846, RMSE = 0.128). The seven-layer architecture is proposed as a scalable, low-cost design for resource-constrained aquaculture, with an estimated capital cost of approximately KES 52,000 (USD 400) for a reference single-pond configuration. The study introduces the Integrated Cyber-Ecological IoT Survivability Framework (ICEISF), synthesising General Systems, Water-Quality, Fish Behaviour and Physiology, and Cyber-Physical Systems theories; only the water-quality construct currently has simulation-based quantitative support. The model contributes a testable, cost-specified architecture to precision fish farming and digital aquaculture.