基于模糊评价的福建沿海水质卫星遥感监测模型

Water Quality Satellite Remote Sensing Monitoring Model of Fujian Coastland Based on Fuzzy Evaluation

  • 摘要: 卫星遥感技术对于开展连续、实时、大尺度的海洋水质状况监测具有明显优势。该文充分利用卫星遥感的标准算法模型和半分析算法模型分别反演得到叶绿素a、颗粒状有机碳、黄色物质和透明度等海洋生态参数,并选取与福建沿海海水受污染程度密切相关的溶解氧、化学耗氧量、无机氮和活性磷酸盐作为海洋水质的评价因子,通过建立海洋生态参数与海洋水质评价因子两者之间的统计关系模型,在海洋水质综合评价中引入模糊综合评价法,最终建立一套基于卫星遥感和模糊评价的海洋水质监测模型,并利用2009—2013年福建沿海同步获取的海洋水质现场实测数据对模型的反演精度进行验证。结果表明:使用该监测模型开展对福建沿海海洋水质卫星遥感监测是可行的,监测准确率为81%,具有较好的业务化应用前景,由于模型对于Ⅳ类海洋水质监测的准确率明显高于Ⅱ类和Ⅲ类海洋水质,因此,比较适合于福建近岸海域的海洋水质监测。

     

    Abstract: As the development of satellite remote sensing technology of the high resolution and multi-survey spectrum channel, it brings an obvious advantage for ocean monitoring of water quality situation for continuous, real-time, large scale. With the observed samples of 12 marine water quality monitoring sites in the marine sector in Fujian coastal water, the remote sensing satellite standard algorithm and semi-analysis algorithm model is adopted, and marine ecological parameters are retrieved such as chlorophyll-a, particulate organic carbon, yellow substance and transparency. Evaluation factors of sea water closely related to the marine pollution degree in Fujian coastland are selected as dissolved oxygen, chemical oxygen demand, inorganic nitrogen and active phosphate, the statistical relationship model is established between marine ecology parameters and the evaluation factor of marine water quality. The evaluation of sea water quality is conducted by the fuzzy comprehensive evaluation method, and finally a set of marine water quality monitoring model is finished based on the satellite remote sensing and fuzzy evaluation model. The inversion precision of the model is verified by the synchronization acquisition field measurement of the marine water quality of Fujian coastal water in 2009-2013. Results show that, for the computed yellow substance by semi-analytical algorithm, the average relative error is 34%, the root mean square error is 0.179; for the computed transparency the average relative error is 26%, the root mean square error is 0.332, so the inversion result is ideal. For dissolved oxygen, the average relative error of the statistical model is the minimum, while the active phosphate model is the maximum, all four models have good agreement with measured values, and the correlation coefficient of the regression equation is more than 0.58. Results of the satellite remote sensing monitoring show that in most of the harbor and rivers into the sea along the coast of Fujian, the water quality barely meets class Ⅳ standard, the ocean water quality gradually improves with the offshore distance, and the harbor pollution trend is greater inside bay than that of the outside. From 304 samples of marine water quality monitoring, the satellite monitoring results of 245 samples are consistent with the ground observations, and the accurate rate of monitoring is 81%. The use of the monitoring model is feasible for the satellite remote sensing monitoring of sea water quality in Fujian coastland, so the prospect of business application is better. Because the accurate rate of the model is obviously higher to the marine water quality monitoring of Ⅳ than that of Ⅱ and Ⅲ , the marine water quality monitoring model is more suitable for the offshore sea area in Fujian.

     

/

返回文章
返回