Conventional single-stage static models cannot fully capture the cumulative effects of meteorological factors and canopy growth over the growing season. It is often required to accurately and rapidly predict summer maize yield. In this study, a rolling prediction framework was developed to integrate field meteorological observations, canopy hyperspectral information, and the extremely randomized trees (ERT) model. Yield estimates were dynamically accumulated as growth-stage information. A relatively reliable prediction window was identified to improve prediction accuracy and stability for yield assessment of summer maize. Field experiments were conducted with staggered sowing dates at the Gucheng National Agro-meteorological Field Scientific Observation and Research Station in Hebei Province, China, from 2018 to 2023. Meteorological and canopy hyperspectral data was collected at six growth status, including the three-leaf, seven-leaf, jointing, tasseling, milk, and maturity stages. Furthermore, 13 vegetation indices were calculated from canopy hyperspectral reflectance to characterize canopy structure, chlorophyll status, water condition, and photosynthetic activity. Two strategies were constructed for prediction: Single-Stage Prediction (SSP) and Rolling Prediction (RP). A nested cross-validation framework was adopted for the two-stage feature selection using variance inflation factor (VIF) and feature importance screening via node impurity reduction. The 2024 field experiment data was further used for validation. The results showed that: 1) The predictors differed between the SSP and RP models, but the number of retained features was similar. The SSP models retained 6-8 features, whereas the RP models retained 7-9 features, indicating that the two-stage feature selection strategy effectively reduced feature redundancy for the predictive information. 2) The RP models shared better prediction performance than the SSP ones. On the test set, the coefficient of determination (R2) of the RP models reached 0.88 with a root mean square error (RMSE) of 636.05 kg/hm2 and a mean absolute percentage error (MAPE) of 4.70%. From the tasseling to the maturity stage, rolling integration of earlier meteorological and later canopy spectra was reduced the prediction fluctuation caused by single-stage modelling. Moreover, 2024 data validation showed that the RP model was achieved in more stable performance after rolling to the milk and maturity stages, especially for the maturity stage with an R2 of 0.88, an RMSE of 616.34 kg/hm2, and a MAPE of 6.18%. The milk-to-maturity period was a relatively reliable window to predict the field-scale summer maize yield. 3) Feature importance analysis showed stage-dependent patterns. Before the maturity stage, yield prediction was dominated by temperature, precipitation, and their cumulative effects, which were closely related to source–sink establishment and dry matter accumulation. At the maturity stage, the importance of remote sensing vegetation indices, including the MERIS terrestrial chlorophyll index (MTCI) and red edge position index (REP), increased again, because leaf senescence caused chlorophyll degradation and canopy water loss, thereby altering canopy spectral responses. The rolling prediction framework with multi-source data and ERT model improved the accuracy and stability of summer maize yield prediction. The available stage-wise yield estimates can be updated after evaluation. The milk-to-maturity period was identified as a relatively stable window for final yield assessment, whereas earlier predictions were more suitable for preliminary risk warning with meteorological and canopy spectra.
- Article type
- Year
- Co-author
Based on the field staging experiments, this study revealed the different responses of winter wheat and summer maize sowing date adjustments in growth and development, photosynthetic physiological characteristics, grain filling, yield formation and quality to climate warming in northern area of North China, providing scientific basis for agricultural production measures to cope with climate change in North China Plain.
Different sowing date experiments of winter wheat and summer maize have been conducted at Hebei Gucheng Agricultural Meteorology National Observation and Research Station in northern area of North China from 2017 to 2023, which were set up in four sowing dates, including 10 d early sowing, 10 d late sowing, 20 d late sowing, and control. The growth process, above-ground dry matter accumulation and distribution, photosynthetic characteristics of leaves, grain filling rate, yield agronomic traits, and grain nutrients of winter wheat and summer maize were observed.
The whole growth period of winter wheat was shortened with the delay of sowing date, mainly because of the shortening of seedling stage before winter. There was a parabolic relationship between the whole growth period and sowing date of summer maize. The seedling stage was shortened by 1.3 d, and the flowering stage and grain formation-filling stage were extended by 1.5 d and 1.6 d for every 10 d delay of sowing date. The grain filling characteristics of winter wheat and summer maize were not sensitive to sowing date adjustments, and the grain filling rate of summer maize was little different during different sowing dates, but the grain formation period, the filling end date and the peak date were successively delayed due to the delay of sowing date, and the duration of filling days was shortened by 4.0 d for every 10 d delay of sowing date. Under the background of warm autumn and winter in northern wheat region, the sowing duration of winter wheat was extended, while its influence on the yield was obviously weakened. The delay of sowing date with the increase of sowing seed amount would result in yield increase slightly. The yield of summer maize decreased significantly with the delay of sowing date, and the decline rate of theoretical yield was 1 381.50 kg·hm-2 for every 10 d delay of sowing date, but the yield of winter wheat and summer maize showed a jumping decease for 20 d late sowing. The grain distribution rate increased by 1.67% for winter wheat, decreased by 1.57% for summer maize with every 10 d delay of sowing date. As a result, the harvest index increased by 0.017 for winter wheat, and decreased by 0.016 for summer maize with every 10 d delay of sowing date. The leaf photosynthetic rates (Pn) of winter wheat and summer maize were also different in response to sowing date, they were similar for winter wheat during different sowing dates, while decreased by 1.21 μmol·m-2·s-1 for summer maize for every 10 d delay after sowing date. Sowing date adjustments had no significant effects on grain quality of winter wheat and summer maize in northern area of North China.
Extending suitable sowing date range and sowing date delay of winter wheat in North China Plain were positive and effective measures to adapt to climate warming. The early sowing of summer maize in North China Plain might avoid the negative effects of high temperature and heat damage, and would promote the increase of yield.
Open Access
Research
Issue
Net primary productivity (NPP) in forests plays an important role in the global carbon cycle. However, it is not well known about the increase rate of China's forest NPP, and there are different opinions about the key factors controlling the variability of forest NPP.
This paper established a statistics-based multiple regression model to estimate forest NPP, using the observed NPP, meteorological and remote sensing data in five major forest ecosystems. The fluctuation values of NPP and environment variables were extracted to identify the key variables influencing the variation of forest NPP by correlation analysis.
The long-term trends and annual fluctuations of forest NPP between 2000 and 2018 were examined. The results showed a significant increase in forest NPP for all five forest ecosystems, with an average rise of 5.2 gC·m-2·year-1 over China. Over 90% of the forest area had an increasing NPP range of 0-161 gC·m-2·year-1. Forest NPP had an interannual fluctuation of 50-269 gC·m-2·year-1 for the five major forest ecosystems. The evergreen broadleaf forest had the largest fluctuation. The variability in forest NPP was caused mainly by variations in precipitation, then by temperature fluctuations.
All five forest ecosystems in China exhibited a significant increasing NPP along with annual fluctuations evidently during 2000-2018. The variations in China's forest NPP were controlled mainly by changes in precipitation.
京公网安备11010802044758号