| With the rapid development of the economy,the ecological environment in the world is becoming increasingly severe.China's current eco-environmental problems include serious soil erosion,grassland degradation,forest sharp decline,biodiversity reduction.These problems often involve large spatial and temporal scales,complex processes,and many driving factors.It is difficult to deal with using traditional means.The rapid development of satellite remote sensing technology provides a wealth of data sources for large-scale long-term eco-environmental assessment and early warning.In recent years,the rapid development of big data technology provides a powerful analytical method for the in-depth analysis of massive eco-environmental spatio-temporal big data.Therefore,how to apply big data analysis methods and integrate multi-source heterogeneous spatio-temporal big data such as satellite remote sensing,meteorology,ground survey,etc.to monitor,evaluate and warn the ecological environment in a wide range and serve the eco-environmental protection and governance in China is of great significance.This paper takes the Qilian Mountain Ecological Environmental Protection Area in Qinghai Province as research area,and integrates satellite remote sensing,meteorology,ground survey and other multi-source heterogeneous spatio-temporal big data to carry out regional eco-environmental big data early warning methods and applications.The application research includes:(1)This paper established a warning indicator system for the Qilian Mountain Ecological Environment Protection Zone in Qinghai Province.Based on the principle of grading the eco-environmental early warning indicator system and referring to the "Technical Methods for Monitoring and Early Warning of Resources and Environmental Carrying Capacity(Trial)",this paper focused on the characteristics of the ecological environment in the study area,reduced the duplication between indicators,and increased the portability of the indicator system.Based on the first-level classification of natural potential factors and human disturbance factors,a three-level early warning index system for the research area was constructed to provide model support for the realization of regional eco-environmental early warning decisions;(2)This paper extracted the eco-environmental elements of the study area over the years,and completed the spatio-temporal analysis and trend analysis of each eco-environmental element.Based on the established early-warning indicator system,this article collected and organized multi-source heterogeneous spatio-temporal big data(such as remote sensing big data,meteorological big data,etc.),and extracted regional eco-environmental factors from 1989 to 2017,including land cover,meteorological indicators,vegetation coverage,and biodiversity,etc.This paper applied big data analysis techniques to mine the space distribution characteristics and time-varying rules of each ecological element,and analyzd the changing trend of regional ecological environment in units of grids,in order to provide quantifiable index information for the realization of regional eco-environmental early warning decisions;(3)This paper achieved regional eco-environmental assessment in the study area over the years.This article first completed special evaluations in units of counties based on industry standards,including environmental evaluations,ecological evaluations,and theoretical stocking calculations.Then the paper selected the comprehensive index evaluation method for regional eco-environmental evaluation based on the portability of the evaluation method.The entropy weight method was used to determine the index weights to reduce human interference and take into account the relationship between phases.Each index weight can represent the importance of the indexes participating in the evaluation.At the same time,in order to take into account the areas that have deteriorated rapidly in the study area,the evaluation results were corrected in combination with the resource depletion index,providing an operable evaluation method for realizing regional eco-environmental early warning decisions;(4)This paper established an eco-environmental big data prediction model for the study area.Based on the quality of the index and the length of the time series,considering the accuracy and efficiency of the model,this paper chose BP artificial neural network to realize the eco-environmental prediction and reduce the possibility of overfitting.This article separately predict the indicators of single ecological elements and the results of comprehensive regional evaluation,and obtained the ecological environment status of the study area in the next 3-5 years.The accuracy of prediction was verified by the full sample of the index extraction values in 2017,making recommendations for protection and development of Qilian Mountain Ecological Environment Protection Zone. |
PDF Full Text Request