Font Size: a A A

Research On In-season Potential Yield Prediction And Nitrogen Utilization In Rice

Posted on:2018-07-01Degree:MasterType:Thesis
Country:ChinaCandidate:X K GeFull Text:PDF
GTID:2393330575967097Subject:Crop Cultivation and Farming System
Abstract/Summary:PDF Full Text Request
In crop growth,nitrogen is one of the most important elements involved in the synthesis of multiple substances.The appropriate nitrogen management strategy is essential for efficient crop production and sustainable agricultural development by reducing the negative impact of nitrogen fertilizer application.In this research,we studied the topics of"Predicting Rice Grain Yield based on Dynamic Change of Spectral Index in Early and Middle of the Growing Season" and "Relationship between Nitrogen Use Efficiency and Yield of Different Panicle Fertilizer Rates under Different Base Fertilizer Conditions".As a result,we set up the dynamic model of spectral index of rice in early and middle growing season and its grain yield predicting algorithm and found the relationship between nitrogen use efficiency and yield in different fertilization by analyzing the data systematically,which makes supply of nitrogen fertilizer on demand possible and benefit for solving the problem of excessive application of nitrogen fertilizer.We created different rice colonies with different varieties and eco sites in field experiments of nitrogen rates.Spectral indexes were collected frequently by RapidScan CS-45 before flowering.Relationships between each spectral index and agronomy parameters,including LAI,plant N accumulation and leaf N accumulation.The result showed that NDRE had best correlation with agronomy parameters among all indexes.Besides that,the dynamic change of NDRE index showed "slow-fast-slow" growth pattern.For this reason,we selected S-type curve for dynamically modeling with RAGDD as independent variable and NDRE as dependent variable.The equation of the model is y =A1-A2/1+e(x-x0)/dx+A2.The coefficient of determination of models in all experiments are bigger than 0.9 while RRMSE of those are smaller than 5%,indicatingthe correction and high accuracy of the model.The model includes three parameters,x0,BGV and FPV,corresponding to x0、A1、A2 in the formula.xo is the critical point of dynamic change.Rice is in early growing season before x0 with small NDRE while grow fast.Rice enter the middle of growing season after passing the xo when NDRE is bigger and grow slower.In the early stage of rice fertility,the spectral index value is low and fluctuates around a value,which we called the spectral index background value(BGV).BGV can reflect the growth status of rice in early period,but the correlation is low.In the middle of the growing season,the spectral index growth slows down,only increases or approaches the maximum,which is called feature platform value(FPV).FPV has a good linear relationship with agronomic parameters,and the coefficient of determination is above 0.7.Each treatment has only one FPV and it is suitable for the monitoring and diagnosis of nitrogen nutrition in rice and yield prediction.Based on FPV we set up the in-season potential grain yield predicting model,whose equation is y=36081x-1174.4.Tesing result showed that the accuracy of the model is more than 80%,which is significantly higher than that of the existing algorithms and it is can be used in japonica rice varieties of the middle and lower reaches of the Yangtze River region and Jianghuai region.By comparing nitrogen use efficiency of different panicle fertilizer rates under different base fertilizer conditions,we found that panicle fertilizer have an significant effect on the grain yield.With the increase in the amount of topdressing,yield increased by 10%-20%,resulting from the increase of effective number of spike per unit area and grains per spike.NUE is also significantly influenced by the rate of panicle fertilizer.Four types of NUE including partial factor productivity of nitrogen(NPFP),agronomy efficiency of nitrogen applied(NAE),nitrogen apparent recovery efficiency(NRE),and physiological efficiency of nitrogen(NPE)decreased with the increase of fertilizer rates,while the NUE panicle fertilizer was higher than that of base fertilizer.NPFP was the most closely related to yield among NUE,and NPFP was significantly different at different yield levels.Farmers can be based on the target yield or potential production corresponding NPFP reasonable set fertilizer,nitrogen fertilizer to achieve efficient use.
Keywords/Search Tags:rice, NDRE, dynamic change model, S-curve, grain yield prediction, nitrogen use efficiency, panicle fertilize
PDF Full Text Request
Related items