| The predictions of crop yield, quality, and physiological and biochemical indicators using remote sensing technology have been frequently reported. This technology has been extensively used to predict the production of some crops, which have been proved very successful. The current study was carried out on the research bases of weak- and medium-gluten wheat in Jiangsu Province, including five coastal cities and counties and the areas along the lower reaches of Yangtze River, adopting the method of randomizing sampling sites. The relationships between wheat morphological and physiological parameters, satellite remote sensing parameters, and grain yield and quality at jointing, booting and flowering stages were studied. The parameters obtained using remote sensing were extracted. And the models based on satellite remote sensing were developed to predict the morphological and physiological parameters, grain yield and quality of weak- and medium-gluten wheat. Based on these studies, we expected to explore the agronomical mechanism of monitoring grain yield and quality of weak- and medium-gluten wheat using satellite remote sensing technology and to provide references for further establishment of grain yield and quality models suitable for multi-year and multi-environment. The primary results were as follows:1. Due to the shortage of distinguishable plant characters of weak- and medium-gluten wheat in the same year, the differences of wheat plants in prior and middle growth period were less significant. Owing to the differences in production environment, the performance of monitoring of grain yield and quality varied in different years.2. Using factor analysis with varimax rotation, the plant physiological parameters at jointing, booting, and flowering stage were classified into plant biomass, plant pigment and nitrogen content, and plant sugar content; the satellite remote sensing parameters at flowering stage were classified into the parameters reflecting the status of plant pigments, the parameters reflecting the status of plant pigments and nitrogen, the parameters reflecting the coverage of green vegetation and crop growth status, and the parameters reflecting the plant and soil water content. Each type of these remote sensing parameters reflected primarily one of the plant physiological parameters respectively.3. At jointing stage, the models using a single plant parameter were not sufficient to predict grain yield and quality. However, the appropriate leaf area index, biomass, and plant nitrogen content could be used as references for the regulation and control of plant population. At booting stage, the nitrogen content of green leaves and plant could be used to predict the contents of protein and wet gluten. At flowering stage, SPAD reading and plant nitrogen content could be used to predict starch content and sedimentation value. Owing to the differences in plant characteristics(LAI, SPAD) at booting and flowering stage, grain yield and quality were significantly different.4. Remote sensing parameters at jointing stage, such as NDVI,OSAVI,SAVI,NRI, could be used to predict SPAD reading and the nitrogen content of green leaves; remote sensing parameters at flowering stage, such as NDVI,OSAVI,SAVI,NRI, could be used to predict SPAD reading and the nitrogen content of green leaves, and B6, WI, NDWI, NDWI2 could be used to construct models of predicting grain protein content, wet gluten content, and sedimentation value. The above-mentioned models were better in predicting grain protein content, wet gluten content, and sedimentation value than predicting starch content.5. Grain yield, protein yield, and starch yield were closely related to the accumulation and translocation of dry matter of prior to and post- anthesis. Protein content, wet gluten content, and sedimentation value were primarily related to nitrogen accumulation of prior to anthesis. The relationship between starch content and the accumulation and translocation of dry matter didn't reach a significant level. Therefore, it was feasible to construct a model of multiple variables to predict grain yield by incorporating the accumulation of dry matter and translation of leaf nitrogen of prior to and post- anthesis. The accumulation of nitrogen of prior to anthesis and the contribution rate of nitrogen accumulation of post-anthesis could also be incorporated to build a model of multiple variables to predict protein content.6. The predictive effects of multiple variables models using remote sensing parameters reflecting plant biomass, plant pigment and nitrogen content, and soil water content and temperature were much better than those of the single variable models. These multiple variables models included: 1) grain yield model using LAI and NRI at the jointing stage; 2) protein content model using LAI and B1 at jointing stage; 3) starch yield model using LAI and NRI at jointing stage; 4) grain yield model using LAI and B1 at flowering stage; 5) protein content model using LAI, B4, and B6 at flowering stage; 6) starch yield model using LAI and B1 at flowering stage; 7) wet gluten content model using NDVI, NRI, and NDWI2 at flowering stage; 8) Zeleny model using NDVI, B1, and B7 at flowering stage.7. The comparison of different models showed: 1) At booting and flowering stages, the predictive effects of grain yield models using single plant parameter reflecting biomass were much better than those of models using single remote sensing parameter; 2) At flowering stages, the predictive effects of grain protein content and wet gluten content models using single remote sensing parameter were much better than those of models using single plant parameter; 3) The predictive effects of models at flowering stages were much better than at booting stages; 4) The predictive effects of multiple variables models were much better than those of single model; 5) The predictive effects of multiple variables models basing on carbon-nitrogen metabolism were much better than those of other models.
|
PDF Full Text Request