IEEE Geoscience and Remote Sensing Magazine - March 2014 - 35

systems (GLDAS) that integrates satellite and ground measurements, using land surface modelling and data assimilation techniques. The main aim is to generate optimal
fields of land surface parameters such as soil moisture, soil
temperature and evapotranspiration. The chapter provides
also plenty of details about how to access data and how to
use on-line visualization tools to have a quick view of the
data before downloading. Among the others, one of the
most versatile is called Giovanni (http://disc.sci.gsfc.nasa.
gov/giovanni/) where data can be visualized and analyzed
on line.
The Part 2 of the book is dedicated to the modeling aspects introducing two main classes of models: the Land
Surface Model and the Hydrological Model.
Land Surface Models (LSMs) are introduced in chapter 4
where the focus is on dry lands and high-elevation regions.
These are particularly sensitive to climate change and moreover they deserve a specific attention because the land-atmosphere interactions are not clearly understood yet. In
fact this chapter addresses main open issues in modeling
such as thermal coupling between land and atmosphere in
dry lands, and soil stratification beneath alpine grassland.
For these issues, the latest developments are presented and
the improvements in land surface models are discussed.
Chapter 5 presents a review of parameterization and parameters estimation of hydrological models. Hydrological
models have largely improved in the last decades especially
towards advance parameter estimation methods. The chapter illustrates the basic concepts as well as the latest trends
of hydrological modeling. The core part of the chapter is
the review of parameter estimation methods where different global optimization algorithms for calibrating hydrological model are presented including latest developments
on distributed hydrological models. These latest models
need new methods to deal with high-dimensional parameter space.
The Part 3 of the book deals with data assimilation techniques and related issues such as error estimation in land
data assimilation systems. Part 3 is made up of 5 chapters
and represents the main focus of the book.
Chapter 6 is dedicated to an overview of theories and
methods of data assimilation and specifically to applications in land surface studies. With data assimilation methods, observations are continuously inserted in model states
by taking advantages of constrains of physical models. The
advantages and disadvantages of different techniques such
as recursive Bayesian filter, Kalman and Ensemble Kalman
filters are presented. Two case studies on assimilation for
soil temperature are introduced as applications. Moreover, the chapter concludes indicating that thanks to the
increasing observations availability, developments of new
and effective methods are required in order to improve the
reliability of systems and achieve a multi-scale information fusion.
Chapter 7 addresses a key point in data assimilation
methodologies, the estimation of model and observation
march 2014

ieee Geoscience and remote sensing magazine

errors. The authors illustrate what could be the impact of
poor error assumption on the performance of land data
assimilation systems. After the presentation of a theoretical background of the problem, the chapter reviews the
recent development of adaptive filtering systems which
try to estimate modeling and observation error covariance information.
Chapter 8 deals with the adaptive inflation scheme for
adjusting the forecast error covariance matrix and prior
observation error covariance matrix in ensemble Kalman
filter assimilation. This method is a specific algorithm to
estimate the inflation factor of forecast error covariance
matrix by optimization of the likelihood function of innovation (that is observation minus forecast residuals). The
cases of time-dependent and time-independent inflation
are discussed. All the inflation adjustment methods discussed require linear or tangent linear observation operator. The use of different operators is highlighted as a next
step in this topic.
Chapter 9 provides an overview of error estimation in
Land Data Assimilation Systems. As already pointed out in
the previous chapters error estimation is a key point for improving the performances of data assimilation systems. The
review addresses three main parts, model input estimation
error, model parameter error estimation and model structural error estimation by indicating different approaches
such as multiplicative inflation and additive inflations.
Moreover, a new method is proposed based on evolutionary concepts in particular cross-over principles. When
compared with other existing methods, this approach can
determine improved results. Future investigations will be
in the direction of using such methods in real land data assimilation systems in order to solve the assimilation problems with real observations.
Chapter 10 introduces a framework named Multi-scale
Kalman Smoother-based (MKS) that is a modification of
the traditional Kalman filter. This method is used to estimate the probability distribution of hydrological variables
given model predictions, observations and MKS parameters. One of the most important applications in the context
of data assimilation can be obtained when observations
from different scales are available. The estimation of the
MKS parameters is obtained through an Expectation-Maximization (EM) algorithm. This framework is presented as
well through a numerical example.
The Part 4 of the book is dedicated to applications and is
made up of 4 chapters.
Chapter 11 provides an overview of the North America
Land Data Assimilation System (NLDAS), a system that
runs multiple land surface models (LSMs) such as the Noah,
Mosaic, Sacramento Soil Moisture Accounting (SAC-SMA)
and the Variable Infiltration Capacity (VIC) models over
the continental USA to generate long-term hourly, 1/8th
degree hydrological and meteorological products. These
LSMs have generated land surface products including water fluxes (evaporation, runoff/streamflow), energy fluxes
35

http://disc.sci.gsfc.nasa

Table of Contents for the Digital Edition of IEEE Geoscience and Remote Sensing Magazine - March 2014