Estimation of cotton yield based on net primary production model in Xinjiang, China

Time series data of the China environment and disaster reduction satellite (HJ), TM images and improved Carnegie Ames Stanford Approach (CASA) model were used in this study to estimate cotton yield in 121 groups Xinjiang province. we used CASA model to calculate the cotton net primary production (NPP),(NPP=(SOL × FPAR × 0.5) × ε). Finally, yield was estimated through converted the NPP to biomass, then cotton yield were obtained and validated by the field data. For NPP model, the relative error between the predicted cotton yield and the actual yield HJ was -18.00%, and TM images was -16%. It was feasible to predict cotton yield by HJ satellite data for estimating cotton yields. But in this paper, the light use efficiency (ε) as the constant, we had not considered the influence of the temperature and precipitation of the space variability and climate condition, all of these needed further study.