On the regional scale, the main land surface parameters acquired by quantitative
remote sensing include three types: land use data, vegetation physiology, and canopy
physical data—most of these parameters are closely related to crop growth and yield.
With decades of development, several quantitative remotely sensed products at
regional and global scale from multisource satellite data have been available.
Among these products, the crop-related products include LAI, PAR, FPAR, GPP,
NPP, FVC, ET, LST, SM, and LC. LAI is an important crop growth parameter,
which is defined as the total leaf area per unit of ground area. It is an important factor
for describing several crop growth processes such as photosynthesis, evapotranspiration, and yield. PAR, LST, and SM also directly affect the process of crop
photosynthetic production and evapotranspiration, which are the driving and material conditions of crop growth. ET participates in the process of crop water and
energy balance. GPP/.NPP is the cumulative state in different periods of crop
assimilation processes, which is the direct material basis for organ formation. SM
can be used to crop drought conditions monitoring. LC provides crop spatial
distribution information for crop growth monitoring. In addition to the spectral
indices, these yield related products also used for yield estimation, such as PAR is
used to predict sugar beet yield in Europe. The EOS satellite series of the United
States, the Sentinel satellite series of Europe, and the HJ series and high-resolution
GF series in China provide multisource data for crop growth monitoring and yield
forecasting and become possible.
The advantage of remote sensing techniques on crop growth monitoring is
obvious. Remote sensing has been widely used and is often used in crop yield
forecasting systems in serval counties. However, remotely sensed data cannot
describe the physical process of crop growth, and some improvement of data
accuracy is still a challenge. The mixed pixels were inevitable in regional crop
applications, and this problem is hard to overcome in the near future. Therefore, the
traditional crop modeling method is still needed when no remote sensing data are
available.
11.5 Data Assimilation
Simulation model and observation are two means to crop growth monitoring and
yield forecasting. Integration remote sensing and crop growth model give a better
description for crop growth and yield, water balance, and the carbon cycle. In the
past years, various data assimilation approaches were carried out for integrating
remote sensing data in crop growth models. There are mainly two data assimilation
strategies: calibration approach of model parameters and updating approach of
model state variables. Calibration data assimilation is adjusting the initial conditions
or model parameters through a minimum difference between the observed and
simulated state variables (Fig. 11.3a). In this strategy, initial conditions or model
214
H. Pan and Z. Chen
remote sensing include three types: land use data, vegetation physiology, and canopy
physical data—most of these parameters are closely related to crop growth and yield.
With decades of development, several quantitative remotely sensed products at
regional and global scale from multisource satellite data have been available.
Among these products, the crop-related products include LAI, PAR, FPAR, GPP,
NPP, FVC, ET, LST, SM, and LC. LAI is an important crop growth parameter,
which is defined as the total leaf area per unit of ground area. It is an important factor
for describing several crop growth processes such as photosynthesis, evapotranspiration, and yield. PAR, LST, and SM also directly affect the process of crop
photosynthetic production and evapotranspiration, which are the driving and material conditions of crop growth. ET participates in the process of crop water and
energy balance. GPP/.NPP is the cumulative state in different periods of crop
assimilation processes, which is the direct material basis for organ formation. SM
can be used to crop drought conditions monitoring. LC provides crop spatial
distribution information for crop growth monitoring. In addition to the spectral
indices, these yield related products also used for yield estimation, such as PAR is
used to predict sugar beet yield in Europe. The EOS satellite series of the United
States, the Sentinel satellite series of Europe, and the HJ series and high-resolution
GF series in China provide multisource data for crop growth monitoring and yield
forecasting and become possible.
The advantage of remote sensing techniques on crop growth monitoring is
obvious. Remote sensing has been widely used and is often used in crop yield
forecasting systems in serval counties. However, remotely sensed data cannot
describe the physical process of crop growth, and some improvement of data
accuracy is still a challenge. The mixed pixels were inevitable in regional crop
applications, and this problem is hard to overcome in the near future. Therefore, the
traditional crop modeling method is still needed when no remote sensing data are
available.
11.5 Data Assimilation
Simulation model and observation are two means to crop growth monitoring and
yield forecasting. Integration remote sensing and crop growth model give a better
description for crop growth and yield, water balance, and the carbon cycle. In the
past years, various data assimilation approaches were carried out for integrating
remote sensing data in crop growth models. There are mainly two data assimilation
strategies: calibration approach of model parameters and updating approach of
model state variables. Calibration data assimilation is adjusting the initial conditions
or model parameters through a minimum difference between the observed and
simulated state variables (Fig. 11.3a). In this strategy, initial conditions or model
214
H. Pan and Z. Chen