Quantitative remote sensing of land surfaces
By S. Liang. John Wiley & Sons Inc., , 2004. ISBN 0 471 28166 2. 160 x 243 
mm. xxvi + 534 pages, plus CD-ROM. Illustrated. Price 8250 hardback.

Review by J. Hogg

This new book presents the principles of optical remote sensing and 
selected algorithms for quantitatively estimating land surface variables 
from remotely sensed observations. According to the author, it evolved 
from lecture notes he prepared for a graduate course on remote sensing. 
The structure, content and style of the book reflect this. It has an 
explicit structure: it contains neat descriptions of context, terms, 
concepts, principles and complex models; it cites relevant scientific 
literature, and the style facilitates and encourages the study of 
quantitative methods of optical remote sensing of land surfaces.

The book is essentially in two parts, with introductory and concluding 
chapters. The first part covers Chapters 2 to 4, and the second part 
Chapters 5 to 12.

In Chapter 1, the author explains concepts of optical remote sensing of 
land surface variables. Quantitative models for estimating land surface 
variables from remotely sensed data fall into three categories: 
statistical, physical and hybrid. A concise review of concepts of optical 
remote sensing of land surfaces describes digital numbers, radiance and 
its directional dependence through solid angles, irradiance, 
bi-directional reflectance, albedo and extra-terrestrial solar irradiance. 
The components of modelling systems for remote sensing include scene 
generation, scene radiation modelling, atmospheric radiative transfer 
modelling, navigational modelling, and mapping and locating the results of 
modelling on the earth's surface. A forward modelling system predicts what 
remote sensing data will be, under a given set of remote sensing and 
environmental conditions. A physically-based inversion scheme determines 
various land surface geophysical and biophysical variables. The author 
aptly lays the foundation for the rest of the book.

The first part of the book describes methods of physical modelling of the 
earth's atmosphere, vegetative canopies, soil and snow. Chapter 2 
introduces radiative transfer in the atmosphere. It describes the 
radiative transfer equation in the solar spectrum, outlines analytical 
models of the bi-directional reflectance distribution function (BDRF) of 
land surfaces, to specify the lower boundary conditions of the radiative 
transfer equation, and links the quantities in this equation to 
atmospheric properties. It includes various numerical and approximate 
solutions, and approximate forms of radiative interactions between 
atmosphere and surface. Chapter 3 introduces physically-based canopy 
reflectance models for linking canopy properties with radiance measured by 
a remote sensor. It describes canopy radiative transfer formulae that are 
appropriate for dense vegetative canopies, introduces leaf-optical models 
that provide necessary leaf-optical properties for most canopy reflectance 
models, outlines geometric optical models that are used for sparse 
vegetative canopies, and discusses Monte Carlo ray tracing and radiosity 
simulation models. Chapter 4 introduces soil and snow reflectance models, 
including radiative transfer and geometric optical models and highlighting 
similarities between modelling BDRFs for snow and ice. It introduces the 
scattering properties of single particles of snow and soil, optical 
properties of soil, multiple scattering solutions for angular reflectance 
from snow and soil, including approximate and numerical solutions, 
geometric optical modelling and inversion of snow parameters, and 
discusses practical issues. The systematic analytical presentation of key 
concepts, theories and methods is impressive, for both its clarity and its 
brevity.

The second part of the book describes selected algorithms for estimating 
land surface variables from optical remotely sensed data. It provides a 
synthesis of a wealth of scientific information and forms the heart of the 
book. Chapter 5 explains the importance of radiometric calibration of 
satellite sensors, presents concepts and principles, describes current 
post-launch methods of sensor calibration, and shows calibration results 
from Landsat 4 and 5 Thematic Mapper and NOAA AVHRR. Chapter 6 presents 
algorithms for atmospheric correction of optical remote sensing images, 
describing the context, representative algorithms for correcting single- 
and multiple-viewing angle images, and presenting selected algorithms for 
estimating water vapour content of the atmosphere. Chapter 7 is on 
applying topographic corrections to remotely sensed images. It explains 
semi-empirical and physically-based methods, and reviews current methods 
for generating digital elevation models, because of their critical 
importance in applying topographic corrections. Chapter 8 is on estimating 
land surface biophysical variables from optical remote sensing. It 
presents concisely and informatively selected vegetative indices, and 
their functional relationships with biophysical and biochemical models, 
both for multispectral and hyperspectral remote sensing systems. It 
explores spatial signatures, in terms of variograms, and outlines 
procedures for estimating leaf-area index and fractional 
photosynthetically active radiation from AVHRR and techniques for physical 
inversion, using traditional optimisation methods in conjunction with a 
canopy radiative transfer model, generic algorithms, table lookup and 
hybrid algorithms. Chapters 9 and 10 neatly describe methods of estimating 
surface radiation budgets, using broadband albedo and long-wave radiation, 
respectively. Chapter 11 is on four-dimensional data assimilation, a 
relatively new technique for providing physically consistent estimates of 
spatially distributed environmental variables. The estimates are derived 
by using a dynamic model from heterogeneous, irregularly distributed and 
temporally inconsistent observation data, acquired by different optical 
remote sensing systems, with different levels of spatial, spectral and 
temporal resolution. After presenting basic concepts and mathematical 
descriptions of dynamic models for data assimilation, representative 
algorithms are briefly outlined from meteorology and oceanography, 
hydrology and crop-growth models. Chapter 12 describes methods of 
calibrating models, and validating products derived from models by field 
measurement on the ground. As spatial resolution becomes coarser, advanced 
methods of spatial scaling must be developed to determine the accuracy of 
products. The author summarises background information, then explains the 
need for, and explores methods of validating, land surface products, as 
well as comparing different algorithms and products.

Earth observation is driven by scientific questions and practical 
applications, according to the author. Having reviewed techniques for 
estimating land surface variables, the author sets out, in the final 
chapter, to show how the derived products can be used to solve practical 
problems, by reviewing ways in which to integrate remote sensing with 
ecological models, agriculture, urban-heat islands, global carbon cycles, 
and improving our understanding of land-atmosphere interactions.

This book represents a tour de force by Liang. It provides a systematic 
and contemporary review of the theory and methods of estimating land 
surface variables from optical remote sensing. The concise synthesis of 
concepts, theories, methods and applications is not only stimulating to 
read, but also invaluable as a comprehensive reference on the subject. 
Written in an economical style and without much discussion, the structure 
and argument is coherent and commendably clear, the illustrations are 
helpful and the compact disc contains relevant computer programs and sets 
of data. With over 1300 references to the scientific literature, readers 
can delve further into most topics. The book is extremely well produced, 
with a good layout, clear type, a useful list of acronyms and a 
comprehensive index. It is timely and refreshing, fills a gap in the 
market, will make a valuable addition to university libraries, and is 
highly recommended to students, academics and scientists.

J. Hogg
  
The Photogrammetric Record
Volume 19 Issue 108 Page 413 - 415, December 2004