Remote Sensing Inversion Of Planting Area,Growth Potential And Yield Of Sugarcane In Guangxi And Its Influencing Factors

Guangxi is the largest sugarcane planting and sucrose industry base in China,but it’s the average per mu yield is not high due to long-term natural disasters.Timely acquisition of Spatio-temporal dynamic information of planting area,growth trend,phonology,and yield for many years and its related response characteristics can provide important scientific support for regional sugarcane planting optimization,disaster risked management,and adjustment of sucrose industrial structure.This paper takes sugarcane crops in Guangxi as a research object and is based on LANDSAT8 OLI,interpreting the planting area distribution of sugarcane in Guangxi from 2014 to 2018 with random forest classification method;based on MODIS-NDVI data,construct growth potential difference monitoring model to identify the growth trend of sugarcane stem in this period;reconstruct MODIS-LAI time-series data,and realize the identification of the spatial and temporal characteristics of the key phenological period of sugarcane in this period by dynamic threshold method;utilize key growth and development which can characterize the yield of sugarcane in Guangxi.The period index was used as a comprehensive influencing factor to establish the sugarcane yield estimation model,and the geographic detector and geographically weighted regression model were used to further explore the effects of risk factors on the spatial differentiation of sugarcane growth and yield in Guangxi.The paper draws the following main conclusions:(1)The interpreting method used in the paper has a good effect.The overall classification accuracy of the sugarcane planting area in Guangxi is higher than 92%,the Kappa coefficient is more than 0.8,and the 5-year mean area relative error is-10.7%.Sugarcane planting in Guangxi is mainly in Chongzuo,Nanning,and Laibin,and the planting area in the whole area is distributed locally and generally dispersedly.(2)NDVI difference monitoring model can reflect the interannual Spatio-temporal differences of stem elongation and ten-day fluctuation characteristics in Guangxi.Due to the comprehensive influence of meteorology,soil,and field management,the growth trend and area distribution of sugarcane are generally characterized by alternating changes of good,normal,and poor grades,and can better respond to the water and soil moisture content of drought and flood alternating,low hills and karst development in Guangxi.Comprehensive features such as changing circumstances.(3)The phenological extraction method used in the paper can effectively eliminate the unstable fluctuations and singularity values of LAI sequences in Guangxi,among which the fitting results of the Asymmetric Gauss and Double Logistic method better compensate for the volatile changes of local curve values in Savitzky-Golay filtering processing,and can simulate the growth process of sugarcane in Guangxi more accurately.The RMSE and the average error of the beginning time of the key phenological period of sugarcane in Guangxi were all within ±15d according to the Asymmetric Gauss method.(4)Remote sensing estimation results of sugarcane yield based on the BP neural network model are better than those of univariate and multivariate statistical models.R2 and RMSE in BP1 and BP2 reliability tests are 0.9243,1.8064 million t,0.9272,1.7461 million t respectively,and BP2 simulation results are better than BP1.(5)Regardless of the meteorological,land type,or social and economic factors,its influence on planting area,growth trend,and yield of sugarcane is not independent,but the combined action of multiple factors.For example,the underlying surface conditions such as plain,platform,iron-bauxite in Central Guangxi are the advantageous environment for large-area planting of sugarcane;for example,longitude,latitude,slope,and other land factors are the sugarcane growth,space-time difference of yield.The average contribution rates of dissimilation were 29.17% and 12.94% respectively,and the effect of two factors on growth and yield of sugarcane was greater than that of a single factor,showing the relationship between non-linear and two-factor enhancement.The influence mechanism of spatial differentiation of planting area,growth potential,and yield of sugarcane in Guangxi was analyzed through influence factor research.The research results are of great scientific significance to the development of the "double-high" sugarcane industry in Guangxi.