Trends in vegetation productivity related to climate change in China’s Pearl River Delta

A reversal in the 3-decade trend of increasing growing season and peak productivity was observed between 1995 to 2004, accompanied by a period of decreased sunshine hours over PRD. This corresponds to a period of accelerated urbanisation and industrialisation, with aerosols contributing to solar dimming across the region, potentially with plant tissue damage from high pollutant concentrations. Along with this reduced radiation, higher summer tem- peratures during this period may explain the reverse productivity trend, as crops and forest species meet upper temperature thresholds. The main finding of this study is overall increased vegetation productivity during the last 3 decades over the PRD associated with winter warming. However, our observed broad-scale, decade-long decline in productivity and negative correlations of summer growth with temper- ature, also have implications for crop production, the health of natural ecosystems and global climate modelling. The study presents the trends, but more research is needed to thoroughly explain these trends affecting this important region. Supporting information S1 Fig. Annual average time series of temperature observations recorded at the Zhongshan and Zhuhai meteorological station in the PRD. (DOCX) S2 Fig. Temporal trend in the monthly NDVI times series based on annual trend analysis. (DOCX) S3 Fig. Temporal trend in the annual phenometrics. Seasonal Adjusted (SA) removes first the seasonal cycle from a time series and then computes the trend on the seasonal-adjusted

time series, Annual Aggregated Time Series (AAT) refers to trend based on annually aggre- gated time series, and the phenometrics trend is based on AAT as there one observation for each year. (DOCX) S4 Fig. Land use and land cover-based graphical summary of the phenometrics. F = Forest, SH = Shrubland, CL = Cropland, Mix = Mosaic of Forest and Cropland. (DOCX) S5 Fig. Trend in phenometrics. The metrics showing overall insignificant trends in most land cover classes of different land cover classes with loess fitting and confidence intervals; correla- tion coefficient (r), slope, significance (p) and total change (TC) is obtained by linear regres- sion analysis of the variables over time. (DOCX) S6 Fig. Maps of partial correlations of phenometrics and precipitation variables. (DOCX) S7 Fig. Maps of partial correlations of phenometrics and temperature variables. (DOCX) S1 Table. List of phenology dynamics metrics. (DOCX) Author Contributions Conceptualization: Sawaid Abbas, Man Sing Wong. Formal analysis: Sawaid Abbas. PLOS ONE Climate and vegetation trends in the Pearl River Delta PLOS ONE | https://doi.org/10.1371/journal.pone.0245467 February 24, 2021 16 / 20