Inversion diffuse attenuation coefficient of photosynthetically active radiation based on deep learning.

Accurate estimation of the diffuse attenuation coefficient of photosynthetically active radiation, K(PAR), is critical for understanding and modeling key physical, chemical, and biological processes in waters. In this study, a deep learning model (DLKPAR) was developed for remotely estimating K(PAR). Compared to the traditional empirical algorithms and semi-analytical algorithm, DLKPAR demonstrated an improvement in the model's stability and accuracy.