| 其他摘要 | The periodic outbreak of green tide has become one of the most serious ecological disasters in the southern Yellow Sea, causing huge social and economic losses to the coastal regions. In the later stage of green tide dissipation, large amounts of Ulva prolifera was either landing in coastal regions, or sinking into the sea. Abundant types of nutrients will be released into the sea during the decomposition of green tide, altering the physical-chemical environment of the sea water. In addition, the landing of green tide will lead to serious impacts on the landscape ecological environment and human activities in coastal regions. Therefore, further studies should be conducted on the dissipation of green tide to clarify its dissipation characteristic, so as to provide theoretical basis for evaluating its impact on marine ecology and formulating the corresponding green tide prevention and control strategies.
Multiple satellite data including MODIS, HJ-1A/1B, GF-1 and Sentinel-2 during the period from 2007 to 2019 were used in this study. Firstly, the solutions were proposed to improve the accuracy of the estimation area of the green tide limited by the cloud and spatial resolution. Then, the daily dissipation rate and spatiotemporal variation of green tide were analyzed based on the monitoring results of remote sensing image. Finally, the possible deposition regions were preliminarily speculated. The main conclusions are as follows:
(1) Cloud cover and spatial resolution affect the estimation accuracy of green tide. To eliminate the impact of cloud, a method for estimating the green tide area under cloud was proposed based on cloud and green tide distribution. The percentage of the green tide area covered by the cloud (P_cloud_GT) ranged between 0.47% and 28.17% to the total green tide area during the period from 2007 to 2019.
To reduce the impact of spatial resolution, based on semi-simultaneous multi-source images, two solutions were proposed. The first one is that we constructed transformation models of green tide areas (Area) estimated by multi-source images. Applying these models, the relative deviations of corresponding patch area from GF-1, Sentinel-2 and HJ-1A/B images decreased from 22.77% and 36.40% to 14.44% and 24.38%, respectively. The relative deviations of corresponding patch area from MODIS and Sentinel-2, GF-1 and HJ-1A/1B images decreased from 229.09%, 153.06% and 89.17% to 23.97%, 18.69% and 23.25%, respectively. The second one was that we proposed a green tide sub-pixel coverage (ACCM) estimation model to reduce the impact of coarse spatial resolution. Applying this model, the maximum relative error of ACCM between MODIS and high-resolution image was 13.59%, which was much smaller than the relative error of Area (the minimum is 55.85%).
(2) The daily dissipation rate (DR) of green tide during the period from 2007 to 2019 was analyzed from two aspects. One was the DR between two neighboring dates (TND), and the other was the DR from different starting dates to the minimum area date (DSE). It is found that the DSE exhibits an increasing trend within a year (except 2010), while the TND show much difference for different year. In interannual scale, the overall trend of the maximum of DSE and TND is increasing, which is consistent with the maximum of green tide area, with the corresponding R2 of 0.60, 0.41 and 0.34, respectively. A prediction method of the green tide dissipation duration has been proposed based on the daily dissipation rate variation during the period from 2007 to 2017. When using this method to 2018 and 2019, the predicted values were very close to the actual values, especially in 2018 where the relative error between them was only 6.67%.
(3) The dissipation characteristic of green tide during the period from 2007 to 2019 were analyzed from two aspects. One was the landing sequence of green tide in the southern part of Shandong Peninsula, and the other was the distribution density of green tide in the southern Yellow Sea. Two types of landing sequence were found. One type was following the order of Rizhao, Qingdao, Rushan and Haiyang, the other type was in the reverse order. The distribution density of green tide had significant differences in the southern Yellow Sea. The most seriously affected region was mainly located in the area of 34º50'N~36º1'N, 120º32'E~121º33'E. While concerned in southern Shandong Peninsula coastal water, the most seriously affected region was coastal water near Qingdao, while the lightest affected region was coastal water near Rushan. Based on the dissipation characteristic of green tide, it can be speculated that the coastal regions of Shandong peninsula and the center of southern Yellow Sea(33º31'N ~ 35º22' N,
120º14 'E ~ 122º8' E) were most likely to the possible deposition regions of green tide. |
修改评论