RapidSCAN And Unmanned Aerial Vehicle Remote Sensing-based Rice Nitrogen Status Diagnosis And Precision Management In Cold Region

According to previous research results and recommended cultivation techniques by local agricultural management departments,regional optimum rice management system(RORM)is used to be conducted with consistent fixed fertilization rates and management in our study area(Jiansanjiang,Heilongjiang Province,Northeast China).However,due to the variability of crop varieties,soils,topography and meterorological conditions within the region,the potential of further increasing yield and nitrogen(N)use efficiency(NUE)is still existed.Precision crop management aims to manage spatial and temporal variability associated with key aspects of crop production for optimum profitability,sustainability,and protection of the environment.Currently,there is an urgent need for remote sensing-based precision rice management(PRM)system for increasing both grain yield and NUE in Northeast China.This study evaluated different approaches to non-destructive rice N status diagnosis and developed precision N management(PNM)strategies using RapidSCAN CS-45 active canopy sensor and unmanned aerial vehicle(UAV)-based multi-spectral camera Mini-MCA across different N rate treatments,management systems,sites and years.At the same time,remote sensing-based PRM strategies by integrating other high yield and high efficiency management technologies was also developed,and systematically evaluated.The main results were concluded as follows:1)The most practical and stable method of using the RapidSCAN sensor for rice N status diagnosis is to calculate N sufciency index(NSI)with the sensor default vegetation indices and then to estimate N nutrition index(NNI)non-destructively(R2 = 0.50-0.59).This semi-empirical approach achieved a diagnosis accuracy rate of 59-76%at the key growth stages.2)The most practical and stable method of using UAV-based multi-spectral camera Mini-MCA for rice N status diagnosis is to calculate NSI with the top vegetation indices and then to estimate NNI non-destructively(R2 = 0.65-0.70).This semi-empirical approach achieved a diagnosis accuracy rate of 74-82%at the key growth stages.3)The recommended N rate of PNM decreased significantly with the increase of rice N status,and there was no significant difference with the in-season economically optimal N rates(EONR).Compared with the regional optimum N management(RONM),three-year average yield and net return margin of NNI-based PNM were increased by 6-10%and 20-32%under N deficient,and partial factor productivity of applied N(PFPN)increased by 15-31%without reducing net return margin under N surplus,respectively.Compared with NNI-based PNM,N fertilizer optimization algorithm(NFOA)-based PNM had better and more stable performance.4)The PRM system had potential to further increase N agronomic efficiency by 7-8%and net return margin by 496-612 ? ha-1 over the RORM system in three years of fixed field experiments.Furthermore,the precision high yield management(PHM)system incorporated cup-seedling management further increased rice yield by 5-7%with high NUE across three years and two varieties.5)Compared with RORM system across ten sites and two years,RapidSCAN sensor-based PRM system increased yield by 11%,net return margin by 2686 ? ha-1,and PFPN by 21-26%.UAV-based PRM system further increased yield by 9%,net return margin by 2210 ? ha-1,and PFPN by 12-25%.Meanwhile,this PRM system also decreased greenhouse gas emissions by average of 252 kg CO2eq ha-1.Considering the large scale of rice farming,this study recommended UAV-based PRM system in Northeast China.