Uncertainty analysis of multiple global GPP datasets in characterizing the lagged effect of drought on photosynthesis. (June 2020)

• Record Type: Journal Article
• Authors: Xie, Xinyao; Li, Ainong; Tan, Jianbo; Lei, Guangbin; Jin, Huaan; Zhang, Zhengjian
• Is Part Of: Ecological indicators. Volume 113(2020)
• Publication Date: 2020-06
• Subjects: Global GPP datasets -- Lagged effect of drought -- Photosynthesis -- LUE-based models -- ML-based models -- Process-based models; Environmental monitoring -- Periodicals; Environmental management -- Periodicals; Environmental impact analysis -- Periodicals; Environmental risk assessment -- Periodicals; Sustainable development -- Periodicals; 333.71405
• DOI: 10.1016/j.ecolind.2020.106224
• Languages: English
• ISSNs: 1470-160X

Abstract

Highlights: 26 global GPP datasets from LUE-, ML-, and process-based models were compared. Lagged effect of drought on photosynthesis from various GPP datasets was discussed. Tropics presented lower reliabilities of lagged effect than other regions. Our study highlights the importance of GPP datasets in drought assessment. Abstract: Understanding the lagged effect of drought on photosynthesis is essential in global climate change research. Various gross primary productivity (GPP) datasets have been used to assess the drought effect on photosynthesis. However, discrepancies have been found in these GPP datasets, and whether the GPP discrepancies can cause uncertainties in understanding the lagged effect of drought on photosynthesis remains unclear. Here, twenty-six global GPP datasets from light use efficiency (LUE)-, machine learning (ML)-, and process-based models during 2001–2010, were used to evaluate the role of GPP discrepancies in characterizing the lagged effect of drought on photosynthesis. Based on probability analysis, a relatively reliable pattern about the lagged effect of drought on global photosynthesis was derived from multiple GPP datasets. Results showed that these 26 GPP datasets existed obvious discrepancies across the globe, with a standard deviation (SD) value of 42 g C m −2 month −1 . Moreover, the area presenting the lagged effect of drought on photosynthesis was found to range between 32% and 69% of the global vegetated lands, confirming the obvious differences in drought patterns derived from different GPP datasets. Our results also indicated that the tropical region (20°N–20°S) presented lower reliabilities of lagged effect than other regions, indicating that the assessment of drought effect on photosynthesis in the tropical region should be more cautious. Furthermore, the probability-based lagged effect of drought on global photosynthesis showed that the lagged effect of drought on photosynthesis existed in 50% of the vegetated lands, with dominant lagged months being less than 8 months, suggesting that the water deficiency in preceding months probably influences vegetation growth. Our study highlights the need to better constrain the carbon modelling under tropical climate and demonstrates that uncertainties caused by GPP datasets should be considered when assessing drought effect on photosynthesis.