全球4km逐日无缝浮游植物类群数据集（1998-2023）

Continuously, globally covered, high-resolution monitoring products of phytoplankton functional types (PFTs) in both space and time are essential for a profound understanding of marine ecosystems and global biogeochemical cycles, and for fostering effective marine management. In this study, by integrating artificial intelligence (AI) with multi-source marine big data, we have developed a Spatial–Temporal–Ecological Ensemble model based on Deep Learning (STEE-DL), and then generated the first AI-driven global daily gap-free 4km PFTs product from 1998 to 2023 (AIGD-PFT), significantly enhancing the accuracy and spatiotemporal coverage of quantifying eight major PFTs (i.e., Diatoms, Dinoflagellates, Haptophytes, Pelagophytes, Cryptophytes, Green Algae, Prokaryotes, and Prochlorococcus). The input data include physical ocean, biogeochemical, spatiotemporal information, and the gap-filled ocean color data using discrete cosine transform with penalized least square (DCT-PLS) approach. The STEE-DL model employs an ensemble strategy of 100 ResNet models based on the Monte Carlo method, utilizing ensemble mean to estimate the optimal values of PFTs, while employing ensemble standard deviation to evaluate model uncertainty. Through the validation of multiple cross-validation (CV) strategies (i.e., standard, spatial-block, and temporal-block CV), combined with in-situ data from long time series, we comprehensively assessed the model's performance, demonstrating the excellent robustness and generalization of STEE-DL. The daily updates and seamless characteristics of the AIGD-PFT product demonstrate its capturing ability of complex dynamic processes in coastal regions. Finally, through a comparative analysis using a triple-collocation (TC) approach, we validated the competitive advantages of the proposed AIGD-PFT product over existing products.