AquaSustain-Bench: A Standardised Multi-Farm Benchmark Dataset and Evaluation Framework for Aquaculture Water Quality Prediction, Disease Risk Detection, and Autonomous Sustainability Control

AquaSustain-Bench is the first open benchmark dataset and evaluation 
framework for commercial aquaculture precision intelligence. It provides 
47.8 million validated sensor-timestep observations from 12 commercial 
partner farms across Egypt, Saudi Arabia, and Bangladesh, covering three 
aquaculture system types (Nile tilapia earthen ponds, whiteleg shrimp 
biofloc RAS, catla/rohu freshwater polyculture ponds) and 551 IoT sensor 
nodes. The dataset includes 228 documented disease and anomaly events with 
ground-truth labels at 15-minute resolution, complete actuator command 
histories, and digital twin state vectors.

Eight standardised benchmark tasks are defined: (1) 72-hour water quality 
forecasting, (2) event detection and early warning, (3) autonomous 
sustainability control, (4) cross-farm transfer learning, (5) new-farm 
onboarding speed, (6) digital twin state estimation, (7) CPS pipeline 
latency, and (8) federated multi-farm learning. Seven published baselines 
and five AquaFarm stack model implementations are included.

GitHub repository: https://github.com/Drwae/AquaSustain-Bench