Early warning and impact duration assessment of green tide disasters in marine ranches based on satellite remote sensing

While green tides threaten marine ranches on China's Shandong Peninsula, most studies focus on remote sensing monitoring and mechanistic analysis, leaving a gap in precise early warning products specifically designed for marine ranch prevention and control operations. This study established methodologies for forecast impact time (FIT) and impact duration assessment (IDA) of green tide in marine ranches, utilizing spatial distribution data from Gaofen-1/6 (GF-1/6) and HuanjingJianzai-2 (HJ-2 A/B) satellites along with drift speed and direction predictions. These methodologies form the foundational framework for developing early warning products against green tide impacts in marine ranches. Application in five marine ranches showed 2022-2023 FIT estimation errors within 5 days; mean/median prediction errors (-2.50/-2.00 days) indicated FIT systematically predicts earlier arrival. IDA mean relative error was -2.69 %, and forecast outcomes were: 50 % underestimated (≤8 days), 40 % overestimated (≤8 days), and 10 % exact. Bootstrap resampling (10 original samples, 10,000 resamples) quantified median errors: FIT had stable 1.00-4.00 day lead time (95 % CI: [-4.00,-1.00] days), while IDA had a highly uncertain, bidirectional error range of -5.50 to +4.50 days (95 % CI: [-5.50, 4.50] days), suggesting that FIT offers reliable advance warning while IDA requires accommodating substantial temporal variability. The findings enable the provision of customized green tide disaster mitigation products for marine ranches, improving prevention efficiency and supporting precise management by fisheries departments.