This research represents an assessment of the current state of pollution from polychlorinated biphenyls (PCBs) in China. Various environmental media including sediment, water, organism and soil were analyzed and the spatial character of PCB pollution in each environmental medium was determined. On a national basis, PCB levels in all environmental media were relatively low, with little evidence of major contamination in ChinaÕs main regions and rivers. However, there were some locations with high PCB concentrations. Sediments in Pearl River (83.1 ng/g) and its estuary (58.9 ng/g), Dalian Bay (58.1 ng/g) and Songhua River (36.8 ng/g) had relatively high level of PCBs. There were also some areas, which were mainly the industrial pollution sites or PCB equipment storage locations, remained seriously polluted with the highest PCB residue level of 150 000 ng/g. In addition, Minjiang Estuary (985.2 ng/g) and Taihu Lake (631 ng/g) had high levels of water pollution, while Pearl Estuary (635.7 ng/g) and Jiaozhou Bay (273.3 ng/g) had relatively high PCB levels in organisms. PCB pollution in soil was limited to a few special pollution areas with the highest PCB level of 4.5448 · 10 6 ng/g. Point source pollution was the common pattern of contamination, influenced primarily by local geographic, economic and historical factors. Analysis of PCB concentrations from the 1980s to 1990s shows an increasing trend, possibly due to the improper disposal of and leakage from PCB containers, chemical transfers, and the general rise of industrial pollution.

Based on monthly AVHRR NDVI data and weather records collected from 160 stations throughout China for 10 years, the relationships between NDVI and two ecoclimatic parameters (growing degree-days (GDD) and rainfall ) were analysed. The results indicate that a signi cant correlation exists between NDVI and the two ecoclimatic parameters; the NDVI/GDD correlation was stronger than the NDVI/rainfall correlation. The NDVI/rainfall correlation coe Y cient exhibits a clear structure in terms of spatial distribution. Further, the results indicate that for natural vegetation, the NDVI/rainfall correlation coe Y cient increases in order from evergreen forest, to deciduous forest, to shrubs and desert vegetation, to steppe and savanna. The correlation coe Y cients associated with cultural vegetation type depend on a number of factors including annual rainfall, seasonal variation in precipitation, type and intensity of irrigation practice and other environmental factors.

No abstract