999
RETRIEVALS OF CANOPY BIOPHYSICAL
VARIABLES USING MULTI-TEMPORAL REMOTE SENSING DATA
Zhiqiang Xiao a ' b *, Jindi Wang a,b , Tiegang Tong c
State Key Laboratory of Remote Sensing Science, Jointly Sponsored by Beijing Normal University and Institute of
Remote Sensing Applications, 100875, China
Research Center for Remote Sensing and GIS, School of Geography, Beijing Key Laboratory for Remote Sensing of
Environment and Digital Cities, Beijing Normal University, 100875, China - zhqxiao@bnu.edu.cn
c School of Info-Physics and Geomatics Engineering, Central South University, Changsha, 410083, China
KEY WORDS: Remote Sensing, Retrieval, Leaf area index, Multi-temporal
ABSTRACT:
The integrated application of multi-source and multi-temporal remote sensing data is the trend of remote sensing application
research, and it is also the practical need to solve the inversion problem of remote sensing. In this paper, a method is developed to
retrieve canopy biophysical variables using multi-temporal remote sensing data. The inherent change rules of biophysical variables
are introduced into the retrieval methods by coupling the radiative transfer model with land process model to simulate time series
surface reflectances. A cost function is constructed to compare the reflectances simulated by the coupled model with time series
reflectances measured by sensors and the canopy biophysical variables with the available prior information. And an optimization
method is used to minimize the cost function by adjusting the values of input canopy biophysical variables such as the temporal
behaviour of the reflectances simulated reaches the best agreement with the multi-temporal reflectances measured. Retrieval of leaf
area index from MODIS surface reflectance data (MOD09) at the Bondville site was performed to validate this method. The
experimental results shows that the use of multi-temporal remote sensing data can significantly improve estimation of canopy
biophysical variables.
1. INTRODUCTION
Remote sensing data have been widely used to estimate the
canopy biophysical variables which are applied to large area
water and carbon cycle simulation, climatic modelling and
global change research. Therefore, it is very important to
precisely estimate these variables from remote sensing data at
the regional or global scale. Currently, there are many methods
to estimate biophysical variables from remote sensing data
(Liang, 2004; Weiss, 1999). And they can be roughly divided
into following classes: by statistical relationship between LAI
and spectral vegetation indices, by physical model inversion
and by other nonparametric methods. These methods have their
own advantage and disadvantage. Since the model inversion
methods is physically based and can adjust to a wide range of
situation, radiative transfer models are more and more used in
the inverse mode to estimate the canopy biophysical variables
(Kuusk, 1991; Jacquemoud, 1993).
It is well known that the inverse problem is by nature an ill-
posed problem mainly because of the not unique solution and
the measurement and model uncertainties (Combal, 2003).
Currently, there are two sorts of ways to solve the ill-posed
problem. One is dependent on some kinds of hypothetical
condition, which may reduce the accuracy of remote sensing
data products retrieved by radiative transfer model, and the
other is to use the prior information of biophysical variables in
model inversion.
With the development of remote sensing technique, many new
types of sensors have been developed, and there are large
numbers of remote sensing data with different temporal, spatial
and spectral resolution obtained from space-borne or air-borne
sensors. The integrated application of multi-source, multi
temporal remote sensing data is the trend of remote sensing
application research, and it is also the practical need to solve the
inversion problem of remote sensing. Noted that most of the
biophysical variables, such as LAI, are time-dependent and
possess inherent change rules along with time which are often
represented by process models such as crop growth models, it is
an important way for us to introduce the inherent change rules
of biophysical variables into the retrieval methods to add the
amount of information needed to retrieve biophysical variables
and improve the accuracy of remote sensing products. In this
paper, we will present our research on the issue of integrating
multi-temporal remote sensing data to retrieve biophysical
variables.
Radiative transfer model are coupled with land process model
to simulate time series surface reflectances. And a cost function
is constructed, according to the posterior probability formula
defined by Tarantola, to compare the reflectances simulated by
the coupled model with time series reflectances measured by
sensors. And an optimization method is used to minimize the
cost function by adjusting the values of input canopy
biophysical variables such as the temporal behaviour of the
reflectances simulated reaches the best agreement with the
multi-temporal reflectances measured. By introducing the
physical constraints from the land surface model, the method
can integrate multi-temporal remote sensing data to retrieve
biophysical variables.
* Corresponding author.