Greening Across Global Tropical Rainforests During the Dry Season? Observed Discrepancies from Passive and Active Remote-Sensed Vegetation Indexes (VIs)
Feng Tian, Shuaijie Li, Yanan Guo, Litao Zhou, Jianhua YangThe future trajectory of tropical vegetation greening remains highly controversial, and the dry-season dynamics of tropical rainforests represent a critical knowledge gap. To advance understanding, this study integrates a suite of multisource satellite vegetation indexes (VIs) and process-based ecosystem models, aiming to comprehensively assess long-term trends in both canopy greenness and photosynthetic capacity during the dry season, and to critically evaluate the inter-product discrepancies in trend detection. Furthermore, we applied partial least squares regression (PLS) to isolate the relative contributions of major climatic factors to dry-season vegetation change. Results showed that great discrepancies in trends of vegetation greenness and productivity from different VIs, presenting that, dry-season canopy greenness (leaf area index, normalized difference vegetation index), and vegetation photosynthesis (sun-induced chlorophyll fluorescence), consistently displayed an increasing trend in tropical rainforest areas during the dry season over the past four decades at the average rate of 5.0 × 10−3 m3·m−3·yr−1, 0.3 × 10−3 yr−1, and 0.9 × 10−3 W·m−2·nm−1·sr−1·yr−1, however, vegetation productivity (gross primary productivity, and vegetation optical depth with C, X, and Ku bands) showed large decreasing trends at a rate of −5.0 × 10−3 gC·m−2·yr−1, −0.7, −0.1, and −0.7 × 10−3 yr−1, respectively, and it was occupied with great spatial disparities that the initial increasing GPP in Congo weakened over time, shifting around the early 2000s from a positive trend (1982–2002) to a negative trend thereafter (2002–2018), while this trend was overwhelmingly contrary to GPP in Southeast Asia rainforest. Multisource remote-sensed VIs did not consistently support a significant greening tendency across the tropical rainforests during the dry season. However, process-based models consistently predicted a robust increasing trend of LAI and GPP in the dry-season tropical rainforests throughout the 21st century at the rate of 0.9 × 10−3 m3·m−3·yr−1 and 2.4 × 10−3 gC·m−2·yr−1. CO2, vapor pressure deficit (VPD), temperature, and solar radiation, defined as the heat-related variables, played a largely negative role in contributing to vegetation growth in dry-season tropical rainforests from different VIs, and the water-related factors, including PRE, SM, and ET, positively contributed to vegetation growth. In terms of contributions, rising CO2 concentration, temperature, and VPD were detected as three of the most important drivers of changes in tropical rainforests, with a contribution degree of 45%, 25%, and 20%, respectively. This study emphasizes that a comprehensive evaluation framework for data assimilation between multisource remote-sensed data and process-based model data in the tropics, which helps reduce uncertainties induced by intensifying climate change and the absence of observation stations, needs to be further improved and refined.