Temporal and spatial evolution of net primary productivity in the Three-River Headwaters Region under phenological changes and anthropogenic influence.

Understanding the spatiotemporal changes in net primary productivity (NPP) and the driving factors behind these changes in climate-vulnerable regions is crucial for ecological conservation. This study simulates the actual NPP (NPPA) and climate potential NPP (NPPC) in the Three-River Headwaters Region from 2000 to 2020. The Theil-Sen Median method and Mann-Kendall mutation analyses are employed to explore their spatiotemporal variation patterns, while geographic weighted regression and machine learning are used to investigate the influence of anthropogenic activities and climatic factors on NPPA, the results indicate that the average NPPA across the entire region over multiple years is 382.506 g C m - 2 yr - 1 , which is 0.132 times the average annual NPPC over the past 21 years, showing an overall distribution pattern of low in the northwest and high in the southeast. The annual increase in NPPA from 2000 to 2020 is approximately 1.034 g C m - 2 yr - 1 . The source region of the Yangtze River shows the largest improvement in vegetation, with 74.1% of the area showing improvement. Between 2002 and 2003, the annual NPPA in the Three-River Headwaters Region experienced a sudden change, lagging behind the NPPC change by 1 year, and after 2005, the upward trend in NPPA became more pronounced. The impact of anthropogenic activities on NPPA shifted from positive to negative to positive from 2000 to 2020, with significant impact areas mainly concentrated in the northeast and a few areas in the central and southern parts. The proportion of areas with extremely significant impact increased from 1.9% in 2000 to 3.7% in 2020. Over the past 21 years, the main factors influencing NPPA changes in the Three-River Headwaters Region have been soil moisture and precipitation, with the influence of different climate factors on NPP changing over time. Additionally, NPP is more sensitive to changes in altitude in low-altitude areas. This study can provide more accurate theoretical support for ecological environment assessment and subsequent protection efforts in the Three-River Headwaters Region.

Spectral characteristics of the correlation between elemental arsenic and vegetation stress in the Yueliangbao gold mining.

Some soils in the Yueliangbao gold mining area have been contaminated by heavy metals, resulting in variations in vegetation. Hyperspectral remote sensing provides a new perspective for heavy metal inversion in vegetation. In this paper, we collected ground truth spectral data of three dominant vegetation species, Miscanthus floridulus, Equisetum ramosissimum and Eremochloa ciliaris, from contaminated and healthy non-mining areas of the Yueliangbao gold mining region, and determined their heavy metal contents. Firstly, we compared the spectral characteristics of vegetation in the mining and non-mining areas by removing the envelope and derivative transformation. Secondly, we extracted their characteristic identification bands using the Mahalanobis distance and PLS-DA method. Finally, we constructed the inverse model by selecting the vegetation index (such as the PRI, DCNI, MTCI, etc.) related to the characteristic band combined with the heavy metal content. Compared to previous studies, we found that the pollution level in the Yueliangbao gold mining area had greatly reduced, but arsenic metal pollution remained a serious issue. Miscanthus floridulus and Eremochloa ciliaris in the mining area exhibited obvious arsenic stress, with a large "red-edge blue shift" (9 and 6 nm). The extracted characteristic wavebands were around 550 and 680-740 nm wavelengths, and correlation analysis showed significant correlations between vegetation index and arsenic, allowing us to construct a prediction model for arsenic and realize the calculation of heavy metal content using vegetation spectra. This provides a methodological basis for monitoring vegetation pollution in other gold mining areas.