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

temperature regimes. Cropland shows a higher response to temperature (r = 0.77) and shorter lag time (30 days), than rainfall (r = 0.48) and 60 days lag time. The natural ecosystems also show a higher response to temperature, with a 75 days lag time for forest (r = 0.63) than to rainfall with 120 days for forest (r = 0.43). Shrubland’s response is approximately between those of forest and cropland in both the level of correlation as well as lag time. The natural eco- systems show almost no response to present rainfall ie. zero lag time, (r < 0.15), but their response to temperature at zero lag time is higher, ranging r between 0.2 and 0.4. Discussion This study demonstrates significant medium-term changes in vegetation productivity over the region incorporating the world’s largest megacity, the PRD region of southern China. We observed increasing productivity in both natural and agricultural ecosystems over the 34-year period 34 1982–2015, especially in the winter season. However, a reversal in the upward trend for seasonal climate indicators MGS and Peak for the middle period between 1995 and 2004 suggests the operation of factors other than temperature. The observed negative correlations between cropland and summer-time temperatures accord with warnings [2,45] that rising global temperatures may impose stress on tropical crops grown in summer such as rice, which occupies 76% of agricultural land in PRD. The overall greening trend may also be associated with other factors such as CO2 fertilisation or intensified

farming. However, the observed increased in warming (Fig 2) and its negative influence on cropland greening and positive association with natural vegetation (Figs 4 and 6) is profound and corresponds with other global and regional studies [2,45–47]. For vegetation growth and high grain yield, rice needs daytime air temperatures between 25˚C and 32˚C, but lower tem- peratures between 20˚C and 25˚C for grain filling and ripening [48]. However, summer daytime temperatures in the region have recently exceeded this regularly, as mean daily temperatures for June to September in Hong Kong are now around 28˚C [49]. Rice is Asia’s foremost staple food crop, and climate change predictions for rice production vary according to models used as well as with or without CO2 fertilization effects [50]. Wassmann et al. (2009) [51] explained that in many parts of Asia including China, current temperatures in July and August are already approaching critical levels for rice production during the critical growth stages, and this is sup- ported by Tao et al. (2013) [52], who found that rice yield in eastern China was reduced by 3.4% due to increase in mean temperature during the vegetation growth period. Furthermore, Tao and Zhang (2013) [47] explained that rising temperatures accelerate the growth of rice, leaving less time for grain development. Lesk et al. (2016) [53] found that extreme temperature stress during the growing season had reduced global cereal yields by 9–10% between 1964 and 2007, thus our finding should not be unexpected. It is important to note that the crop greening or

extension of growing season may not necessarily show a direct correlation with crop yield, espe- cially under mixed or heterogenous cropping system [54]. Crop yield may vary non-linearly with increasing NDVI or greenness indicators as NDVI may not indicate a direct relation with crop biophysical parameters [55]. For example, the relationship between NDVI and Leaf Area Index (LAI) becomes non-linear in crops with highly dense canopy [56]. Besides croplands, some forest and shrubland regions in the northern parts of Huizhou and Guangzhou prefectures also show a negative correlation with temperature (r = -0.6), and the dissociation between temperature and forest greening has been noted in other studies. For example, deciduous needle-leaf forest in Russia showed negative response to warming [46] which is counter-intuitive to the observed increase in plant growth in northern high latitudes due to rising temperatures [46,57–59]. Guo et al. (2018) [57] showed that between 1982 and 2012, only 20% of the non-tropical northern hemisphere showed a strong positive relationship PLOS ONE Climate and vegetation trends in the Pearl River Delta PLOS ONE | https://doi.org/10.1371/journal.pone.0245467 February 24, 2021 13 / 20