parameters could obtain from quantitative remote sensing data including leaf area
index (LAI), canopy water content (CWC), fractional vegetation coverage (FVC),
soil moisture (SM), evapotranspiration (ET), and fraction of photosynthetically
active radiation (FPAR). These satellite data have been widely used in crop research
and management (Brown et al. 2012; Glenn et al. 2011; Qu et al. 2014; Zelitch
1982). Spectral indices are the main method used for crop growth monitoring and
yield estimation. Many researchers develop various indices for crop yield estimation
using satellite data such as normalized difference vegetation index (NDVI),
enhanced vegetation index (EVI) (Kouadio et al. 2014), and temperature vegetation
condition index (TVDI) (Holzman et al. 2014). Food and Agricultural Organization
(FAO) using the NDVI to estimate regional crop yield from NOAA satellite data
with a daily 1.1 km resolution (Hielkema and Snijders 1994). To meet the demand of
precision agriculture, more high-spatial-temporal-resolution remote sensing been
used for crop growth modeling and yield forecasting such as GF, Sentinel, and
UAV (unmanned aerial vehicle) data. So far, many remote sensing data sources have
been used in the crop growth monitoring and yield forecasting system, such as
monitoring agricultural by remote sensing (MARS) crop yield forecasting system
developed by Joint Research Center in Italy (de Wit and van Diepen 2008) and
China’s agricultural remote sensing monitoring system CHARMS (Chen et al.
2012). The spatial-temporal characteristics of remote sensing make crop growth
monitoring and yield forecasting available at regional scale. However, many
methods using remote sensing for crop growth and yield research are empirical
models and cannot describe the mechanism of crop growth process.
Crop growth models and remote sensing become the main tool for crop growth
monitoring and yield estimation. At regional scale, the input variables of the crop
growth model are usually poorly known, such as grid meteorological data, initial
field conditions, sowing date, and soil conditions. The regional field status and crop
biophysical parameters can be estimated using remote sensing data. The successful
application of data assimilation in the land surface model gained increasing attention
in the agricultural community; research found that assimilate remote sensing data
can improve the performance of crop growth model (Claverie et al. 2009; de Wit and
van Diepen 2007; Fang et al. 2008). Numerous research suggested that data assimilation of the crop growth model and remote sensing has become an effective tool for
crop growth monitoring and yield forecasting (de Wit and van Diepen 2007; Huang
et al. 2015; Ines et al. 2013; Wu et al. 2011). Several crop data assimilation schemes
with different data assimilation algorithms, crop growth models, and remote sensing
variables have been developed and evaluated during the last decade, and the results
suggested that they have tremendous potential for improving the simulation performance of crop growth dynamic, water balance, and regional crop yield. There are
mainly two schemes based on the data assimilation algorithm. The first is variational
assimilation, which minimizes the difference between crop variables estimated from
the crop growth model and remote sensing by adjusting the model parameters
(Huang et al. 2015). The second is sequential assimilation with crop model state
variables update by minimizing the uncertainty of model and observations when the
remote sensing data available (Wu et al. 2011). As more terrestrial observation plans
11 Crop Growth Modeling and Yield Forecasting
207
index (LAI), canopy water content (CWC), fractional vegetation coverage (FVC),
soil moisture (SM), evapotranspiration (ET), and fraction of photosynthetically
active radiation (FPAR). These satellite data have been widely used in crop research
and management (Brown et al. 2012; Glenn et al. 2011; Qu et al. 2014; Zelitch
1982). Spectral indices are the main method used for crop growth monitoring and
yield estimation. Many researchers develop various indices for crop yield estimation
using satellite data such as normalized difference vegetation index (NDVI),
enhanced vegetation index (EVI) (Kouadio et al. 2014), and temperature vegetation
condition index (TVDI) (Holzman et al. 2014). Food and Agricultural Organization
(FAO) using the NDVI to estimate regional crop yield from NOAA satellite data
with a daily 1.1 km resolution (Hielkema and Snijders 1994). To meet the demand of
precision agriculture, more high-spatial-temporal-resolution remote sensing been
used for crop growth modeling and yield forecasting such as GF, Sentinel, and
UAV (unmanned aerial vehicle) data. So far, many remote sensing data sources have
been used in the crop growth monitoring and yield forecasting system, such as
monitoring agricultural by remote sensing (MARS) crop yield forecasting system
developed by Joint Research Center in Italy (de Wit and van Diepen 2008) and
China’s agricultural remote sensing monitoring system CHARMS (Chen et al.
2012). The spatial-temporal characteristics of remote sensing make crop growth
monitoring and yield forecasting available at regional scale. However, many
methods using remote sensing for crop growth and yield research are empirical
models and cannot describe the mechanism of crop growth process.
Crop growth models and remote sensing become the main tool for crop growth
monitoring and yield estimation. At regional scale, the input variables of the crop
growth model are usually poorly known, such as grid meteorological data, initial
field conditions, sowing date, and soil conditions. The regional field status and crop
biophysical parameters can be estimated using remote sensing data. The successful
application of data assimilation in the land surface model gained increasing attention
in the agricultural community; research found that assimilate remote sensing data
can improve the performance of crop growth model (Claverie et al. 2009; de Wit and
van Diepen 2007; Fang et al. 2008). Numerous research suggested that data assimilation of the crop growth model and remote sensing has become an effective tool for
crop growth monitoring and yield forecasting (de Wit and van Diepen 2007; Huang
et al. 2015; Ines et al. 2013; Wu et al. 2011). Several crop data assimilation schemes
with different data assimilation algorithms, crop growth models, and remote sensing
variables have been developed and evaluated during the last decade, and the results
suggested that they have tremendous potential for improving the simulation performance of crop growth dynamic, water balance, and regional crop yield. There are
mainly two schemes based on the data assimilation algorithm. The first is variational
assimilation, which minimizes the difference between crop variables estimated from
the crop growth model and remote sensing by adjusting the model parameters
(Huang et al. 2015). The second is sequential assimilation with crop model state
variables update by minimizing the uncertainty of model and observations when the
remote sensing data available (Wu et al. 2011). As more terrestrial observation plans
11 Crop Growth Modeling and Yield Forecasting
207