LAUSR

Daily measurements of photosynthetically active radiation (PAR) and other meteorological parameters at 16 stations in China during 2005–2015 were used to develop the all-weather Adaptive-Network-based Fuzzy Inference System (ANFIS) PAR models. The ANFIS models were compared with other semi-empirical and semi-physical methods, and the results showed that ANFIS models using global solar radiation (G) as input parameter provided more accurate estimates than other models. Daily PAR values during 1961–2016 were reconstructed using meteorological parameters at 908 meteorological stations, and the spatiotemporal characteristics of PAR in China were analyzed. The daily PAR ranged from 20.03 to 36.29 mol m−2 day−1 with annual mean value being about 29.03 mol m−2 day−1 in China during 1961–2016. The PAR values were generally higher in Qinghai-Tibet Plateau (QTP) and Inner Mongolian Plateau (IMP) and lower in Northeast and Southeast China, especially in the Sichuan Basin. PAR decreased in most part of China, especially in Northeast China, with the minimum value being about − 0.11 mol m−2 day−1 year−1 during 1961–2016. The increasing trend only appeared in Southwest China with the maximum value being about 0.02 mol m−2 day−1 year−1. Annual mean daily G, PAR fraction (Fp, the ratio of PAR and G), sunshine duration (SSD), relative humidity (RH), and water pressure (MP) were further analyzed to find the influencing parameters for PAR. The results showed that G, Fp, and SSD were the major factors for influencing PAR in most part of China, but in some areas, RH and MP influenced the PAR more effectively. Finally, the monthly variations of aerosol optical depth (AOD) in the Yangtze River Basin gathered from the Moderate Resolution Imaging Spectroradiometer (MODIS) Level 3 Atmosphere Product during 2002–2016 were analyzed, which was one important factors influencing the PAR values in summer and winter months.