Effects Of Water And Nitrogen Application Amount Under Fertigation On Nitrogen Utilization And Its Supply Decision In Apple Orchard

The planting area of apple in China ranked first in the world,The Loess Plateau is one of the world’s recognized apple advantageous producing areas.At present,there are many problems in apple production,such as excessive use of chemical fertilizer and the time mismatch between nutrient input and apple tree demand,resulting in the decline of apple quality,low rate of excellent fruit and increased the risk of environmental pollution.To solve these problems,this paper was taken the apple trees planted with high density dwarfing（row-by-stand spacing of 4 m by 2 m）as the study object,and employed fertigation method to supply water and nitrogen,which has the characteristics of water saving,fertilizer saving and efficiency enhancing.The field experiment of water and nitrogen application amount of apple trees under fertigation was conducted in Luochuan,Shaanxi Province,from October2017 to October 2020.Two upper irrigation limits（W1:80%θ_f、W2:100%θ_f）and four nitrogen fertilizer application levels(N1:0 kg?hm^-2;N2:120 kg?hm^-2;N3:240 kg?hm^-2;N4:3600 kg?hm^-2)were set in the experiment,resulting in 8 treatments in total.Dynamic monitoring of canopy hyperspectral reflectance of apple trees,leaf nitrogen content,canopy growth characteristics（spring shoot length;leaf area index,LAI）,yield,quality（appearance quality,internal quality）,soil nitrate nitrogen（NO₃^--N）distribution and residue at different growth stages.Based on these indices,the effects of water and nitrogen application amount of fertigation on nitrogen nutrition,canopy growth characteristics,yield and quality of apple trees were studied;the distribution and residual characteristics of NO₃^--N in apple orchard soil under different water and nitrogen application were analyzed;combined evaluation method was used to optimize the amount of water and nitrogen of fertigation in apple orchard based on apple tree growth-nitrogen nutrition dynamics-yield and quality-soil NO₃^--N residue;the method for estimating nitrogen content in apple tree leaves based on hyperspectral remote sensing was explored,on this basis,the canopy leaf nitrogen content prediction and the supply decision of water and nitrogen were studied based on hyperspectral remote sensing.Following are the main results of this thesis:（1）The effects of water and nitrogen fertilizer application amount of fertigation on the leaf nitrogen content and canopy growth dynamics characteristics of apple tree were found.The nitrogen content of apple tree leaves increased with the increase irrigation upper limit at the same nitrogen fertilizer application amount,but the difference was not significant.Compared with no nitrogen application,nitrogen application could significantly increase the leaf nitrogen content of apple trees;N4 had a significant effect on the leaf nitrogen content,but there was no significant difference between N2 and N3,but it was significantly higher than no nitrogen application.Affected by the nitrogen nutrition of apple tree storage,the apple orchard was changed to ditch fertilization mode for fertigation,the application of N2in the first year（2018）was beneficial to the growth of new shoots;the application of N3 in the second year（2019）was beneficial to the extension of new shoots.The growth law of spring shoots accorded with the characteristics of Logistic curve,and the model simulation showed that higher irrigation upper limit or nitrogen application could delay the emergence of the maximum spring shoot growth rate and prolong the spring shoot growth time.Increasing the irrigation upper limit was beneficial to increase LAI under the same nitrogen fertilizer application,but the increase was not significant.In the experimental year,the relationship between the relative size of LAI and the growth of spring shoots in early growth stage of apple trees was related.N2 in 2018 and N3 in 2019 were more conducive to the formation of LAI.N4 was more conducive to the delayed abscission of apple leaves after harvest.（2）The effects of water and nitrogen fertilizer application amount on apple yield,quality and water and nitrogen use efficiency were revealed.The apple yield was significantly affected by the single factor of irrigation upper limit or nitrogen fertilizer application amount.Nitrogen application could significantly optimize the yield components（single fruit weight,fruit number per plant）,and significantly improve the apple yield.The highest yield was obtained in N3 treatment,and N4 resulted in a small reduction in apple yield.The relationship between apple yield and nitrogen application amount was quadratic parabola（P<0.05）,and the best nitrogen application amount was 230～240 kg?hm^-2.Reasonable nitrogen application amount could significantly improve the vertical and horizontal diameters of apple,but the fruit shape index of apple was not significantly affected by water and nitrogen application amount.The results showed that,increasing the upper limit of irrigation could decrease the flesh firmness,soluble solids,soluble sugar,ratio of sugar to acid,ratio of solid to acid,and increase the contents of titratable acid and vitamin C,but these effects were not significant（P>0.05）.Increasing nitrogen fertilizer application amount would significantly reduce the flesh firmness and titratable acid content of apple,and significantly increase soluble solids,soluble sugar,vitamin C,ratio of sugar to acid,ratio of solid to acid.There was a significant correlation between internal quality indices of apple.Compared with the W1 irrigation upper limit,W2could improve water use efficiency and nitrogen agronomic utilization efficiency under the same nitrogen fertilizer application amount.The highest nitrogen agronomic utilization efficiency(75.49 kg?kg^-1)was obtained in the treatment of W2N2.The treatment of W2N3had significant effect on apple yield maintain,improve quality and achieving higher water use efficiency.（3）The effects of fertigation on the soil NO₃^--N distribution,residue and interannual variation characteristics in apple orchard after harvest were clarified.The content of NO₃^--N in 0-80 cm soil layer increased with the increase of soil depth under the treatments of N2,N3,N4,soil NO₃^--N mainly accumulated in the depth of 80-160cm and decreased gradually in 160-200 cm soil layer,indicate that NO₃^--N accumulated outside the main root distribution layer（0-80 cm）.The maximum value of NO₃^--N in horizontal direction appears at 0 cm from the tree row（under the drip pipe）.No nitrogen treatment（N1）soil NO₃^--N moved with water to the lower soil layer,the higher soil NO₃^--N content was near the depth of 180 cm;the maximum soil NO₃^--N content in the horizontal direction was 100 cm away from the tree row,and the soil NO₃^--N content was proportional to the distance from the tree row.The temporal（yearly）distribution characteristics of soil NO₃^--N content were mainly affected by nitrogen application amount.From 2017 to 2020,the NO₃^--N content in the root layer（0-200 cm）of N1 and N2 treatments decreased year by year,with the largest decline rate of 78.56%;there was no significant change in soil NO₃^--N content under N3 treatment;the highest increase rate of soil NO₃^--N content was 197.30%under N4 treatment.The residual amount of NO₃^--N in apple orchard soil after apple harvest was significantly affected by nitrogen application amount.After 3 years fertigation used,in 2018,there was no significant difference in soil NO₃^--N residue among the treatments;in 2019,the residual amount of soil NO₃^--N was significantly higher in nitrogen application treatment than that no nitrogen application treatment;in 2020,there were significant differences among different nitrogen application treatments.The soil NO₃^--N residue decreased year by year under the treatments of N1 and N2,and the reduction extent of N1 was significantly higher than that of N2;the variation range of soil NO₃^--N residue was about 10%in the treatment of N3,and the difference was not significant;nitrogen fertilizer application amount of N4 significantly increased soil NO₃^--N residue,the largest increase rate was81.05%.（4）The amount of water and nitrogen application of fertigation with the optimum comprehensive effects of high-density dwarfing apple orchard on Loess Plateau were put forward.Based on the experiment data of the year without meteorological disaster（2019）,the yield,quality,canopy growth,soil NO₃^--N and water/nitrogen use efficiency were used as evaluation indexes,the principal component analysis method,Technique for Order Preference by Similarity to Ideal Solution（TOPSIS）,grey correlation degree analysis method and membership function analysis method were performed to comprehensive evaluate the response of water and nitrogen application in apple orchard,the results were inconsistent.Furthermore,a fuzzy Borda combination evaluation model based on the evaluation results of four single evaluation methods was established,the results showed that the optimum amount of water and nitrogen fertilizer application was W2N3 during growth period.（5）An estimation method of leaf nitrogen content based on canopy scale hyperspectral reflectance of apple trees was explored.Different spectral preprocessing methods have different denoising ability for spectral curves.On the whole,the first derivative（FD）processing can improve the spectral signal-to-noise ratio,while the second derivative spectrum causes the signal-to-noise ratio to decline.The feature variable extraction methods,namely competitive adaptive reweighted sampling（CARS）,successive projection algorithm（SPA）,random frog（Rfrog）and partial least squares（PLS）greatly reduce the number of factors used for modeling,and the extracted wavelength variables are widely distributed in the VIS/NIR.The prediction accuracy of the nonlinear model based on the same data set is obviously better than that of the linear model.Overall,the standard normal transformation（SNV）combined with FD preprocessed the spectra,extracted variables with Rfrog and extreme learning machine（ELM）systematic modeling method（SNV-FD-Rfrog-ELM）,or Savitzky-Golay（SG）smoothing combined with FD preprocessed the spectra,extracted latent variables with PLS and ELM systematic modeling method（SG-FD-PLS（LVs）-ELM）has a good accuracy in estimating the canopy scale nitrogen content of apple trees（Fuji）in the Loess Plateau.（6）A supply decision model of water and nitrogen for apple trees under fertigation was established,based on hyperspectral remote sensing.The stacked ensemble（SE）model was adopted,the hyperspectral reflectance of apple canopy was used as the model input,and the extreme learning machine（ELM）;ELM optimized by differential evolution（DE）algorithm and ELM optimized by self-adaptive differential evolution（Sa DE）algorithm were employed as a sub-model modeling method,respectively;finally,the sub models were fused with the weights were given based on RMSE of sub-models or partial least squares（PLS）.PLS weighting strategy could provide the best weights in the process of fusion several sub models,and improve the problem of low weight prediction based on RMSE strategy.The overall accuracy of the model was:Sa DE＿ELM>DE＿ELM>ELM.The SE-Sa DE＿ELM model and PLS weighting strategy can resist the influence of outliers,and recommend for in-situ prediction of nitrogen contents in apple tree leaves with excellent accuracy under field conditions,the R_P²,RMSEP and RRMSE were 0.843,1.747 mg?g^-1 and 8.019%,respectively.The young fruit stage and the early stage of fruit expansion are the key periods of nitrogen supply for apple trees.Based on the regression model of top-dressing nitrogen amount-irrigation upper limit-leaf nitrogen content and the regression model of top-dressing nitrogen amount-irrigation upper limit-single fruit weight at the young fruit stage and early stage of fruit expansion;according to the target single fruit weight,the target amount of nitrogen application and the irrigation upper limit were obtained;refer to the results of spectral diagnosis of nitrogen content in apple leaves,the amount of nitrogen applied to apple trees was estimated,and the optimal amount of water and nitrogen supply was given.The model had achieved good results,there was little difference between the theoretical value and the actual value of water and nitrogen supply in young fruit stage and early stage of fruit expansion,the relative error of nitrogen top-dressing amount ranged from 1.67%to9.92%（except for 44.92%of a sample tree in young fruit stage）.Overall,the model has achieved good results.