| The vegetable planting mode represented by greenhouses has become the most important development direction of vegetable production in China.As a Man-made facility,greenhouses have obvious spectral characteristics,but they are often mixed with other Man-made buildings.In addition,the“High Water Fertilizer”and“High Multiple Cropping Index”planting patterns resulted in significantly higher N2O emissions from the facility vegetable fields than in traditional open-air vegetable fields.Therefore,how to accurately obtain information on the greenhouse area of a large area,estimate its N2O emissions,and analyze its impact on the environment has become one of the hotspots of current research on farmland ecological environment.This paper takes the Changzhou City and Taicang City in the Taihu Lake Basin as the research area,and carries out the research on the Greenhouse Index?GI?construction of the greenhouse,and establishes the estimation model of the greenhouse,which is verified in Changzhou.Based on the GI index,the area of greenhouses in Taicang City was obtained.The total N2O emissions from greenhouses in Taicang City were estimated by DNDC model.The spatial differences,seasonal differences and total emission differences of greenhouse N2O emissions were analyzed.A variety of emission reduction measures to explore the potential and feasibility of N2O emission reduction.conclusion as below:?1?Based on the medium resolution Remote Sensing image Landsat-8,combined with the cap transform?K-T transform?,Wetness Index?WI?,Green Vegetation Index?GVI?,Brightness Index?BI?;Normalised Difference Vegetation Index?NDVI?,Normalised Difference Built-up Index?NDBI?,Normalised Difference Bareness Index?NDBaI?,Modified Normalised Difference Water Index?MNDWI?,Enhanced Built-up and Bareness Index?EBBI?;eight kinds of Gray Level Co-occurrence Matrix?GLCM?texture feature quantities,spectrum Feature analysis,separability analysis,and texture feature analysis of greenhouses were carried out.Through spectral analysis,it is found that the spectral characteristics of greenhouse are similar to those of artificial land,and it is determined that the coverage of plastic film will enhance the surface reflectivity.Through the separability analysis,it is determined that the greenhouse and the arable area are highly classified,and the distinction between the bare land and the artificial building is low.According to the separation index M,comprehensively consider the separation between the greenhouse and the bare land,urban land,and cultivated land,and obtain multi-spectral data Coastal?Blue?Green?NIR,the Brightness Index,NDBI,EBBI are sensitive to greenhouse.Through the analysis of texture features,four kinds of texture feature quantities such as Contrast,Variation,Mean and Entropy are obtained,which are effective for distinguishing local objects.?2?Using the Logistic Regression Model,the above eleven bands and Remote Sensing Index were fitted and analyzed to construct a Greenhouse Remote Sensing Index?GI?to extract greenhouse area information;regression analysis of Greenhouse Fraction and Greenhouse Index,establishment of Greenhouse Fraction Model,the minimum can identify 18%of greenhouse coverage.Through training samples and verification samples,the GI obtained Kappa coefficients of 0.82 and 0.81 in the sample area respectively;Kappa coefficient of 0.78was obtained throughout Changzhou,and Kappa coefficient of 0.82 was obtained in Taicang City;Compared with the existing Greenhouse Remote Sensing Index,Achieving higher classification accuracy indicates that the newly proposed GI is more suitable for greenhouse extraction in the Taihu Basin.?3?Based on the measured data of N2O emissions from greenhouse vegetable fields,it is determined that the DNDC model has a good fitting result for the seasonal dynamics of N2O emissions and the total emissions,indicating that the DNDC model has been parameterized after different management measures.The approximate simulated N2O emission flux can be compared in terms of quantity and dynamics,and the N2O emission of soil in greenhouse vegetables can be estimated.Through sensitivity analysis,it is determined that soil pH value is the most significant factor affecting N2O emissions under the same climatic conditions.Soil Organic Carbon,Bulk Density,Fertilization Amount and Irrigation Amount have a greater impact on N2O emissions.?4?Under the 95%confidence level,the DNDC model calculates the average N2O emission per unit of cultivated land area greenhouse is 58.92kg N·hm-2 a-1,under the“celery—calachine—small green vegetables—lettuce”rotation system from 2017 to 2018?1 year?in Taicang City.According to the Independent and Identical Distribution Center Limit theorem,the data can represent the overall situation of Taicang City,and the parameter averaging method is used to estimate the N2O emissions of each township in Taicang City.Based on the GI index,the greenhouse area of Taicang City at the end of 2018 was about 6973 hm-2.Based on the DNDC model,the total N2O-N emissions from greenhouses in Taicang City were estimated to be 410.8t·a-1.According to the analysis of N2O emission load per unit of cultivated land area,it is determined that there are obvious regional differences and seasonal differences in greenhouse N2O emissions.Among them,summer is the peak period of N2O emissions,accounting for 80-90%of the total annual emissions;and different soil properties and fertilization management measures are the main reasons for the difference in N2O emissions in Taicang City.Based on the traditional open-air vegetable fertilization amount,the recommended fertilization amount of facility agriculture and the three emission reduction measures of controlled release fertilizer,Taicang greenhouses can reduce N2O emissions by 22.25%,35.54%and 90%,respectively;Through feasibility analysis,Determining the reduction of fertilization is a feasible scheme for N2O emission reduction in greenhouses in China at this stage. |
PDF Full Text Request