Construction and Validation of Spatial-Temporal Distribution Models on Cotton Fiber Quality Indexes
CHEN Bing-Lin,CAO Wei-Xing,ZHOU Zhi-Guo,ZHU Yan
DOI: https://doi.org/10.3321/j.issn:0496-3490.2007.05.011
2007-01-01
ACTA AGRONOMICA SINICA
Abstract:Construction of spatial-temporal distribution models on fiber quality indexes in cotton are highly important for forecasting fiber quality indexes in different ecological regions and guiding import and export of raw cotton and textile factory's cotton blending in China.Field experiments were conducted at Nanjing(32°03′N,118°47′E) in 2004 and Guanyun(34°18′N,119°14′E) in 1999.Based on the experiment data,the effects of genotypes,fruit branches and nodes,mean daily temperature,sunlight length,soil moisture and plant nitrogen content on cotton fiber quality indexes were quantified at flowering and boll-forming stages,and spatial-temporal distribution models on cotton fiber length,strength,micronaire value and length uniformity were further constructed by integrating existing eco-models on cotton fiber quality.Then the models were validated using the experiment data from different eco-sites,cultivars,N fertilizations and water conditions.The results showed that the values of the root mean square error(RMSE) between simulated and observed values on fiber length,strength,micronaire value and length uniformity were,respectively,0.15 mm,0.29 cN tex-1,0.18,and 0.36 for temporal distribution models,and 0.22 mm,0.60 cN tex-1,0.15,and 0.86 for spatial distribution models,with the actual errors all lower than 5%.Overall,the present models are accurate and applicable for predicting spatial-temporal patterns of cotton quality indexes under different conditions.