| Wheat is one of the main food crops in the world,and also an important food crop in China.The yield of wheat is affected by many factors,such as soil nutrient,irrigation,cultivation and so on.The traditional fertilization method has the problems of low nutrient utilization rate and environmental pollution.We should use modern agricultural technology to implement variable rate fertilization to maximize productivity and limit the impact of excessive fertilization on the environment.The main path of nutrient cycling in wheat field is"input soil wheat output".Wheat,as an important link,provides key information for nutrient monitoring.In this study,the wheat fields of Shandong Province,Huantai experimental station of ecological and sustainable development of China Agricultural University and Huantai agricultural standardization demonstration area(agricultural standard area)of Shandong Province were selected as the research objects.From the perspective of spatial and temporal change laws,combined with the measured data of wheat and soil nutrients,the methods of statistical analysis,principal component analysis and geostatistical analysis were comprehensively used,The characteristics of winter wheat nutrients(above ground)and soil nutrients(below ground)were discussed.Based on the multi spectral data of UAV and ground sampling data,the main nutrient sensitive spectral index of wheat field soil crop coupling was selected,and the estimation model of wheat field soil nutrients was constructed.From the two scales of size,the variable fertilization model of wheat field was constructed,including ground,UAV and satellite remote sensing,It provides scientific basis for precision fertilization in wheat field and fertilization and production management in wheat area.The main research work and results are as follows:(1)A systematic study was carried out on the nutrient status of winter wheat and the multi spectral image of UAV to determine the nutrient changes and corresponding spectral status during the growth period of winter wheat.The results showed that the SPAD value of Winter Wheat in ecological station and agricultural standard area increased with the growth period of winter wheat.The differences of N,P and K in wheat plants in agricultural standard areas were small in the same period,and the variability of nutrients was slightly different.The change of multi spectral characteristics of wheat UAV in the ecological station is consistent with the change of winter wheat nutrients.The NIR and red bands contain the main information of wheat canopy during the growth period.During the growth period of winter wheat,the correlation coefficient of NIR and red band decreased from 0.45 to 0.08,that of green and red band increased from 0.87 to 0.97,and that of NIR and reg band decreased from 0.90 to 0.40.(2)Based on geostatistics and exploratory analysis tools,the temporal and spatial changes of available N,P and K in wheat field were studied.The results showed that the spatial variability of available N and P in the soil nutrients of wheat field was larger than that of available K.There are obvious trend and autocorrelation of soil nutrients in wheat field,and the trend and autocorrelation will change with the growth period of winter wheat.During the growth period of winter wheat,the soil nutrients in the wheat field are strongly affected by crops,and the overall trend is declining,and the decline speed in the region with higher nutrient level is faster than that in the region with lower nutrient level.The model of soil nutrient change during the growth period of winter wheat was constructed.The linear equation could explain the trend of soil nutrient change better,and the model was y=-20.007x+114.72.(3)Based on the analysis of Wheat Nutrient Data and soil nutrient data,the sensitive index screening analysis of canopy multispectral data and nutrient modeling,a coupling model of soil crop nutrient was constructed.The results showed that there was a close relationship between soil nutrients and crop nutrients.In space,the greater the variance of nutrients,the more obvious the relationship between the two;in time,the overall correlation between the two increased with the growth of wheat.The optimum period of topdressing is from the turning green stage to the jointing stage.In the whole growth period of nitrogen,the sensitive vegetation index was MCARI2,the phosphorus sensitive vegetation index was MCARI2,EVI2 and GNDVI,and the potassium sensitive vegetation index was MCARI2,CVI and GNDVI.In the growth period,the sensitive index of crop nutrient changes with time,and the semi-structured index type MCARI2 which can adjust the soil effect is the best.The resolution of J-M in each nutrient of MCARI2 index was close to 2,but the difference of J-M between nutrients was not significant.MCARI2 histogram can directly reflect the nutrient level of crops,and its distribution pattern has the potential of nutrient classification.Using MCARI2to model SPAD of winter wheat,the secondary model is selected in the growth period of winter wheat,the model is y=-220.77x~2+95.43x-0.39,while the linear model of other periods is better,R~2 is above 7.0.The models were y=-0.84+35.12x,y=-2.94+5.98x,y=-0.85+3.76x,respectively.(4)Based on the yield of winter wheat,spectral sensitivity index of canopy and soil nutrients of wheat field,a variable rate fertilization model of winter wheat at field scale was constructed.The results are as follows:the decision-making model of winter wheat demand for nitrogen,phosphorus and potassium based on the yield and fertilizer amount of the wheat field is constructed respectively.The model of yield and nitrogen amount of winter wheat:when n?240 kg/hm~2,y=4.55+0.0127x;when n>240 kg/hm~2,y=7.41 t/hm~2,the recommended nitrogen amount is 240kg/hm~2.Based on the sensitive vegetation index of soil crop coupling and screening,the model of canopy sensitive index and winter wheat yield was established:y=-294.46MCARI2~2+163.79MCARI2-12.47.The recommended model of coupling winter wheat nitrogen fertilizer(N_r):N_r=23185.82MCARI2~2-12896.82MCARI2-1580.15,and the look-up table of nitrogen application amount and MCARI2 index was calculated.Finally,the variable fertilization map is generated by using the variable fertilization model of IDL,which is consistent with the actual investigation.In addition,based on the method of fuzzy clustering and soil nutrient data,a regional variable fertilization model based on the variation of soil nutrient in wheat field was constructed.(5)Based on the GEE platform,the remote sensing extraction of large-scale wheat field information and variable fertilization were studied.The MODIS-NDVI time series data set is constructed by using the double star data.The winter wheat summer maize planting area in Shandong Province is extracted by using the random forest classification method through the linear interpolation and Hants smoothing,and removing the influence of cloud and other factors.The results show that the area extraction of Winter Wheat Based on time series curve has a high accuracy.The average user accuracy of area extraction over the years is 94.7%,and the mapping accuracy is 90.4%.The area determined by remote sensing and Yearbook is 0.96,and the area distribution is consistent with the actual situation.Based on the wheat field area data,the differences between MCARI2 index of winter wheat satellite image and canopy spectral index of ground UAV were analyzed in large scale,forming a large-scale wheat field variable fertilization decision model.The results showed that the MCARI2 values of wheat field in the middle and low areas of Shandong Province were quite different from January to March(about 0.5).After March,the MCARI2 values of the two areas were similar.The modified model of ground satellite MCARI2 index is y=-0.00004x~2+0.0044x+0.0671.The decision-making model of variable fertilizer application in large-scale wheat field was obtained by using the index correction model,which was N_r=23185.82x~2-21707.43x+4867.54.The model was applied to the satellite index map of provincial scale to draw the variable fertilizer application map in large-scale wheat field.In this study,based on two scales of region and field,from the acquisition of Wheat Nutrient information,around the spatio-temporal coupling of soil crop nutrients,the wheat field variable fertilization model,the soil crop nutrient change law model and the large-scale wheat field variable fertilization decision-making model were constructed,which provided a scientific basis for the study of precision fertilization. |
PDF Full Text Request