Simulation And Verification Of COTTON2K Model Under Different Nitrogen Application Rates And Salt Stress In Southern Xinjiang

Soil salinization is one of the important factors that restrict the level of agricultural production in Xinjiang.COTTON2 K model is more suitable for application in salinized areas because the influence of soil salinity on cotton growth is considered in the design.This paper simulates the growth of cotton under the conditions of mild soil salt,moderate soil salt and different nitrogen application levels.Firstly,the sensitivity of 46 genetic variety parameters of COTTON2 K model is analyzed by using the Extended Fourier amplitude test method,and then the sensitive genetic variety parameters are optimized by using NSGA-Ⅱ multi-objective genetic algorithm.The COTTON2 K model was calibrated by using the observation data under mild soil salinity and high nitrogen application in 2018.Finally,the simulation accuracy of the model was verified by using the observation data of the remaining five treatments.The specific research results are as follows:(1)Parameter sensitivity analysis.1)Yield: under the conditions of mild saline soil and moderate saline soil,there are more globally sensitive genetic varieties with more than 20 parameters,of which three genetic varieties have the same parameters,which are varpar26(vertical trunk growth),varpar01(effect of density on growth)and varpar42(temperature effect,yield);2)Dry matter: under the two soil salinity conditions,two globally sensitive genetic variety parameters are the same,which are varpar01(effect of density on growth)and varpar36(fruit development);3)Plant height: varpar21(vertical trunk growth),varpar01(effect of density on growth)and varpar26(vertical trunk growth)are sensitive to plant height under two soil salt conditions.4)Boll opening stage: under the two soil salinity conditions,there are more globally sensitive genetic varieties with more than 15 parameters,of which two genetic varieties have the same parameters,which are varpar40(temperature effect,cracking)and varpar01(effect of density on growth).(2)NSGA-Ⅱ genetic algorithm was used to correct the parameters of 20 genetic varieties with strong sensitivity.Compared with the default value,the parameter values of 8 genetic varieties increased and 12 genetic varieties decreased.The variation of genetic variety parameters is related to soil conditions and cotton varieties.(3)After optimizing the parameters,under two kinds of saline soil conditions and different nitrogen application conditions,the model simulation results were the best for the simulation of seed cotton yield,with an average are of 10.32% and 7.72% respectively.Among them,the are under the mild saline soil condition and high nitrogen application treatment was 0.5%,and the yield simulation results under this treatment were the best.RMSE was 689.12kg/hm2 in mild saline soil and 367.32kg/hm2 in moderate saline soil.Nrmse was less than 13% in both soil conditions;The plant height simulation results were the second,the average are were 8.49% and 18.1% respectively,the nrsmes were less than 20%,and the RMSE was between 6.38 cm and 11.57 cm.The plant height simulation results were the best under medium and mild saline soil conditions and medium nitrogen application rate;The dry matter simulation result is the worst.The nrmse under mild saline soil is 24.70%,and the reliability of the result is medium.The nrmse under moderate saline soil is 42.27%,and the reliability of the result is poor.This may be due to the excessive impact of cotton2 k model on the accumulation of dry matter under moderate saline soil conditions.In the simulation of growth period date,the simulation results of first bud day and first flower day are better,while the simulation results of 60% boll opening day are worse.This study provides a scientific theoretical basis for the application of COTTON2K model in southern Xinjiang.