CR 82:161-176 (2021)  -  DOI: https://doi.org/10.3354/cr01626

ABSTRACT: Sunshine duration (SD) has recently received a great deal of attention as it relates to climatic research under global climate change. Here, we explored the spatiotemporal variability of SD in the Pearl River Delta (PRD) in mainland China during 1960-2018. A novel equation was proposed to assess the relative influence of 5 climatic variables on variability in SD based on random forest (RF) analysis. The dominant variables, their temporal evolution driving SD, and their spatial distribution are discussed. Results showed that SD in the PRD varied significantly, both at a temporal and spatial scale. At the spatial scale, SD decreased gradually from the southeast to the northwest; in most areas, SD exhibited a downward trend. Temporally, annual SD decreased significantly at a rate of 57.01 h decade^-1 during 1960-2018 (p < 0.01), suggesting a continuous dimming; an abrupt change occurred in 1981 with a main oscillation period of 3.0 yr. The annual mean SD showed a positive correlation with wind speed (WS) and a negative correlation with the other 4 climatic variables. Precipitation was determined to be the most influential variable, with a relative influence rate of 34.46%. The relative influence rates of the 5 examined variables changed over time, and their spatial distribution demonstrated significant diversity across different regions. Aerosols were found to be a factor driving SD in the PRD, but could not completely explain the change in SD in recent years. These findings provide a reference for studying the spatiotemporal variability of SD and detecting its influencing factors in highly urbanized areas.

KEY WORDS: Sunshine duration · Climate change · Spatiotemporal variation · Random forest · Pearl River Delta