Remote Sensing and Global Environmental Change

Contents

Cover

Half Title Page

Title Page

Copyright

Preface

Acknowledgments

Chapter 1: Introduction

1.1 Key concepts

Chapter 2: Remote sensing basics

2.1 Electromagnetic waves

2.2 The electromagnetic spectrum

2.3 Reflectance and radiance

2.4 Atmospheric effects

2.5 Multispectral feature recognition

2.6 Resolution requirements

2.7 Key concepts

Chapter 3: Remote sensors and systems

3.1 Introduction

3.2 Remote sensors

3.3 Remote sensing platforms

3.4 The NASA Earth observing system

3.5 Global Earth observation systems

3.6 Existing image archives

3.7 Key concepts

Chapter 4: Digital image analysis

4.1 Image data format

4.2 Image pre-processing

4.3 Image enhancement and interpretation

4.4 Image classification

4.5 Image band selection

4.6 Error assessment

4.7 Time-series analysis and change detection

4.8 Field sampling using GPS

4.9 Use of Geographic Information Systems

4.10 Key concepts

Chapter 5: Monitoring Changes in Global Vegetation Cover

5.1 EM spectrum of vegetation

5.2 Vegetation indices

5.3 Biophysical properties and processes of vegetation

5.4 Classification systems

5.5 Global vegetation and land cover mapping programmes

5.6 Remote sensing of vegetation as a monitor for global change

5.7 Remote sensing of wetlands change

5.8 Fire detection

5.9 Key concepts

Chapter 6: Remote Sensing of Urban Environments

6.1 Urbanization

6.2 Urban remote sensing

6.3 Microwave sensing of subsidence

6.4 Textural metrics

6.5 Monitoring city growth

6.6 Assessing the ecology of cities

6.7 Urban climatology

6.8 Air quality and air pollution

6.9 Climate change as a threat to urbanization

6.10 Key concepts

Chapter 7: Surface and ground water resources

7.1 Remote sensing of inland water quality

7.2 Remote sensing sediment load and pollution of inland waters

7.3 Remote sensing non-coastal flooding

7.4 Bathymetry of inland waters

7.5 Mapping watersheds at the regional scale

7.6 Remote sensing of land surface moisture

7.7 Remote sensing of groundwater

7.8 Key concepts

Chapter 8: Coral reefs, carbon and climate

8.1 Introduction

8.2 The Status Of The World’s Reefs

8.3 Remote Sensing Of Coral Reefs

8.4 Light, Corals And Water

8.5 Passive optical sensing

8.6 Sensor-down versus reef-up sensing

8.7 Spectral unmixing

8.8 Image-derived bathymetry

8.9 LiDAR

8.10 Sonar

8.11 Sub-bottom acoustic profiling

8.12 Radar applications

8.13 Class assemblages and the minimum mapping unit

8.14 Change detection

8.15 Key concepts

Chapter 9: Coastal impact of storm surges and sea level rise

9.1 Predicting and monitoring coastal flooding

9.2 Coastal currents and waves

9.3 Mapping beach topography

9.4 LiDAR bathymetry

9.5 Key concepts

Chapter 10: Observing the oceans

10.1 Introduction

10.2 Ocean colour, chlorophyll and productivity

10.3 Hazardous algal blooms and other pollutants

10.4 Sea surface temperature

10.5 Ocean salinity

10.6 Physical ocean features

10.7 Ocean Observing Systems

10.8 Marine Gis

10.9 Key Concepts

Chapter 11: Monitoring Earth’s atmosphere

11.1 The status of Earth’s atmosphere

11.2 Atmospheric remote sensing

11.3 The ‘A-Train’ satellite constellation

11.4 Remote sensing atmospheric temperature

11.5 Atmospheric remote sensing of ozone

11.6 Atmospheric remote sensing of carbon dioxide

11.7 Remote sensing atmospheric dust

11.8 Clouds

11.9 Forecasting Earth’s atmosphere

11.10 Atmospheric models and reality

11.11 Hurricanes

11.12 Key concepts

Chapter 12: Observing the cryosphere

12.1 Introduction

12.2 The history and status of the polar ice sheets

12.3 Ice and sea level

12.4 Ice and climate

12.5 Present ice loss in context

12.6 Remote sensing of the Earth’s ice sheets

12.7 Ice sheet mass balance

12.8 Remote sensing permafrost

12.9 Key concepts

Chapter 13: Effective communication of global change information using remote sensing

13.1 Global environmental change as an interdisciplinary issue

13.2 Effective communication through accessibility of data

Chapter 14: Looking ahead: future developments

14.1 Emerging technologies

14.2 The near future

14.3 The more distant future

14.4 Advanced image analysis techniques

14.5 Looking ahead at a changing Earth

References

Index

REMOTE SENSING AND GLOBAL ENVIRONMENTAL CHANGE

This edition first published 2011, © 2011 by Samuel Purkis and Victor Klemas

Blackwell Publishing was acquired by John Wiley & Sons in February 2007. Blackwell’s publishing program has been merged with Wiley’s global Scientific, Technical and Medical business to form Wiley-Blackwell.

Registered Office

John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

Editorial Offices

9600 Garsington Road, Oxford, OX4 2DQ, UK

The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

111 River Street, Hoboken, NJ 07030-5774, USA

For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com/wiley-blackwell

The right of the author to be identified as the author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.

Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought.

Library of Congress Cataloguing-in-Publication Data Purkis, Samuel J.

Remote sensing and global environmental change / Samuel Purkis and Victor Klemas.

p. cm. Includes index.

ISBN 978-1-4443-3935-2 (cloth) – ISBN 978-1-4051-8225-6 (pbk.) 1. Global environmental change–Remote sensing. 2. Environmental monitoring–Remote sensing.

I. Klemas, V. II. Title. GE149.P87 2011

550.28’4–dc22

2010043279

A catalogue record for this book is available from the British Library.

This book is published in the following electronic formats: eBook 9781444340266; Wiley Online Library 9781444340280; ePub 9781444340273

Set in 10/12.5pt Minion by SPi Publisher Services, Pondicherry, India

Preface

This book is intended to provide the reader with a broad grounding in the science of Earth observation (EO) of our changing planet. It contains a comprehensive sequenced discussion covering the significant themes of global change, their causes and how they can be monitored through time. In doing so, it represents a good source of basic information while providing a general overview of the status of remote sensing technology. The text will serve as an invaluable reference for managers and researchers, regardless of their specialty, while also appealing to students of all ages.

The scope of the work yields a reference book that presents the science of EO through a series of pertinent real-world environmental case studies. It offers excellent background material for a course curriculum, being tuned in terms of its length and material covered to fit a variety of teaching scenarios.

The book has been written with students from both bachelors and masters degree programs in mind. For the former group, it contains sufficient material to support a full-semester course; for the latter, the work is intended to serve as a user-friendly introductory text for newcomers to a master’s program in environmental or climate change where the role of EO is considered. The book is thus aimed at a broad audience concerned with the application of remote sensing in Earth science, biological science, physical geography, oceanography and environmental science programs. It is designed for a reader who does not require an in-depth knowledge of the technology of EO, but who needs to understand how it is used as a key tool for mapping and monitoring global change. As such, the book is also intended to serve as an easily approachable text for professionals already working in fields such as ecology, environmental science and engineering, land use planning, environmental regulation, etc., who come across remote sensing in their work and would benefit from learning more about its practical uses, but who are disinclined to take another master’s course or delve into the mathematical tomes on the subject.

The first four chapters of the book introduce the fundamentals of EO, available platforms and the basic concepts of image processing. This will provide an introductory treatment of remote sensing technology to readers who have not been previously exposed to the field. Chapters 5 through 12 each present an important environmental application that:

i is relevant to global change or the status of the biosphere; and

ii lends itself to remote monitoring. Each case study offers insights into new EO techniques.

The work presents the fundamental mechanisms of environmental change and the technology underpinning each sensor relevant for its detection and measurement. An in-depth mathematical treatment of the science is purposely avoided, which should make the text especially appealing to many students and professionals with a non-numerate science background. The final chapter provides a look into the near and more distant future of EO as a tool for monitoring global change, before closing with a sombre but pragmatic look at how climate change has become the defining issue of our time.

The book’s framework is based on using examples from the recent literature, via a gap-bridging cross-disciplinary approach, rather than presenting a classical ‘textbook’ that simply projects the author’s understanding or perception of the use of remote sensing for the study of global environmental change. This inevitably leads to the citation of numerous references that may be of considerable value to those readers desiring to pursue a topic in more detail, and also ensures that these sources will be duly credited and that readers have direct access to them.

Given its emphasis on transmitting concepts rather than techniques, the book does not include problems or exercises. At the end of each chapter, however, a series of ‘key concepts’ are presented that summarize the most crucial points covered. These provide appropriate material that can be developed into challenging exercises, even for advanced courses, if the instructor chooses to elaborate on the themes with a more mathematical underpinning of the technologies illustrated.

Samuel J. Purkis

National Coral Reef Institute

Nova Southeastern University

Dania Beach, Florida, USA

Victor V. Klemas

College of Earth, Ocean and Environment

University of Delaware

Newark, Delaware, USA

Acknowledgements

We wish to thank Dr. Ian Francis for the opportunity to undertake this project and his guidance of the text from conception to birth. Invaluable comments, for which we are extremely grateful, were provided by two anonymous reviewers. Throughout this endeavour, Sam Purkis was supported by the National Coral Reef Institute and Nova Southeastern University’s Oceanographic Center. Similarly, support was provided to Vic Klemas by the College of Earth, Ocean and Environment, University of Delaware. We are indebted to Chris Purkis for his patient and unwavering assistance with the artwork. The text draws upon large numbers of images and illustrations from the work of others, and we appreciate the generosity of the many individuals and publishers who have made available source materials and gave permission for their use.

Our children and grandchildren, Isis, Grace, Andy, Paul, Tom, John Paul and Asta, provided an oft-needed distraction and perspective on the priorities of life. Writing this book demanded considerable time and energy, a burden that was shared as much by family as by the authors. We thank our respective wives, Lotte Purkis and Vida Klemas, for their support, which stretched far beyond the call of duty.

It is our hope that the publication of this book will provide stimulation to a new generation of students and researchers to perform in-depth work and analysis of our changing Earth using remote sensing.

Chapter 1

Introduction

The Earth’s climate is now clearly changing and getting warmer. Many components of the climate system are changing at rates and in patterns that are not natural and are best explained by the increased atmospheric abundances of greenhouse gases and aerosols generated by human activity during the 20th century. These changes include temperatures of the atmosphere, land and ocean; the extent of sea ice and mountain glaciers; the sea level; the distribution of precipitation; and the length of the seasons (AGU, 2008).

The Intergovernmental Panel on Climate Change (IPPC), made up of hundreds of scientists from 113 countries, reached a consensus in 2007 that, based on new research concluded in the last decade, it is 90 per cent certain that human-generated greenhouse gases account for most of the global rise in temperatures. The IPPC was very specific, predicting that, even under the most conservative scenario, the global increase of temperature will be between 1.1 C and 6.4 C by 2100, and the sea level will rise between 18 cm and 58 cm during that same time period (Collins et al., 2007; IPCC, WMO/UNEP, 2007).

The Earth has warmed consistently and unusually over the past few decades in a manner that can be explained only when a greenhouse process is overlaid on orbital variation, solar variation, volcanic eruptions and other natural disturbances. Observational evidence, complex modelling and simple physics all confirm this. Whatever the proportion of human-induced rise in global temperature versus natural rise, there is no doubt that the temperature and the sea level are rising, the Greenland and Antarctic ice sheets are disintegrating, and major weather patterns and ocean currents are shifting (Figs. 1.1 and 1.2.). The Earth’s warming is already causing severe droughts and flooding, major vegetation transformations in deserts and forests, massive tundra methane releases and the degradation of the Amazon rainforest and Saharan vegetation (Gates, 1993). It is also starting to impact the Indian Ocean Monsoon, the Atlantic Conveyor Belt and El Niño weather patterns. The economic impacts of droughts in the USA alone cause $68 billion in losses per year (Chagnon, 2000).

Figure 1.1 The projected range of global averaged sea level rise, re-plotted from the IPCC 2001 Third Assessment Report for the period 1990 to 2100.

Figure 1.2 The atmospheric concentration of carbon dioxide versus the occurrence of temperature anomalies from 1950 to 2001 (IPCC, 2007). An anomaly, in this case, is expressed as the difference between the observed annual global land-surface air temperature and the 1961 to 1990 mean.

Decision-makers and scientists need reliable science-based information to make informed judgements regarding policy and actions in order to be able to address the risks of such changes and variability in climate and related systems. To have any hope of mitigating or adapting to these mostly undesirable changes, we must be able to monitor them continuously and over large global regions. Ship and field observations have provided important data on these phenomena in the past and will do so in the future. However, to be able to observe environmental changes globally, it is necessary to use remote sensors on satellites and aircraft in order to extend local measurements to large areas.

For example, without satellite remote sensing we could not have mapped accurately the changes in the Antarctic ozone hole or the disintegration of ice sheets in Greenland and other areas. In fact, it was the ability to view these global changes with remote sensors from satellite altitudes that brought the enormity and severity of these environmental changes to the attention of scientists, politicians and the general public. Many such important datasets are now available in near-real time at no cost, through web portals such as Google Earth and Google Ocean, allowing ‘citizen’ scientists to accomplish research objectives and promoting public engagement with science in general.

Remote sensing is now a mature enough technology to answer some of the fundamental questions in global environmental change science, namely:

1. How and at what pace is the Earth system changing and what are the forces causing these changes?

2. How does the Earth system respond to natural and human-induced changes?

3. How well can we predict future perturbations to the Earth system and what are the consequences of change to human civilization?

Ice sheets, ocean currents and temperatures, deserts and tropical forests each have somewhat different remote sensing requirements. For instance, ocean temperatures are measured by thermal infrared sensors, while ocean currents, winds, waves, and sea level require various types of radar instruments on satellites. Most ocean features are large and require spatial resolutions of kilometres, while observations of desert or forest changes may require resolutions of tens of metres and many bands within the visible and near-infrared region of the electromagnetic spectrum. Monitoring of coral reefs demands even finer spatial resolution and multiple bands pooled in the short-wavelength visible spectrum.

Fortunately, by the turn of this century, most of these requirements had been met by NASA (the National Aeronautics & Space Administration) and NOAA (the National Oceanic & Atmospheric Administration) satellites and aircraft, the European Space Agency (ESA) and the private sector (Jensen, 2007). Furthermore, new satellites are being launched, carrying imagers with fine spatial (0.64 m) and spectral (200 narrow bands) resolutions, as well as other environmental sensors. These provide a capability to detect changes in both the local and the global environment even more accurately. For the first time, constellations of satellites are being launched with the sole aim of quantifying aspects of the Earth’s climate synergistically. With such technology available, governments are no longer alone in being able to monitor the extent of tropical forests and coral reefs, the spread of disease and the destruction caused by war.

Advances in the application of Geographical Information Systems (GIS) and the Global Positioning System (GPS) help to incorporate geo-coded ancillary data layers in order to improve the accuracy of satellite image analysis. When these techniques for generating, organizing, sorting and analyzing spatial information are combined with mathematical climate and ecological models, scientists and managers can improve their ability to assess and predict the impact of global environmental changes and trends (Lunetta & Elvidge, 1998).

To handle the vast quantities of information being generated by today’s Earth observation programmes, there have been significant advances made in the use of the Internet to store and disseminate geospatial data to scientists and the public. The Internet is set to play an even greater role in the handling of products delivered by future missions.

This book is intended to provide the reader with a broad grounding in the science of Earth observation of our changing planet. It contains a comprehensive sequenced discussion that covers the significant themes of global change, their cause, and how they can be monitored through time. In doing so, it represents a good source of basic information and a general overview of the status of remote sensing technology.

The text will serve as an invaluable reference for managers and researchers, regardless of their specialty, while also appealing to students of all ages. The scope of the work yields a reference book that presents the science of remote sensing through a series of pertinent real-world environmental case studies. It offers excellent background material for a course curriculum, being tuned in terms of its length, and the material covered, to fit a variety of teaching scenarios. Each chapter presents an important environmental phenomenon that:

i is relevant to global change or the status of the biosphere; and

ii lends itself to remote monitoring.

Each case study offers insights into new remote sensing techniques. The work presents the fundamentals of the technology underpinning each sensor type and delivers sufficient detail for the reader to grasp the mode of operation of the instrument and how it can be used to detect and measure the environmental parameters at hand. Thus, the book is aimed at a broad audience concerned with the application of remote sensing in Earth science, biological science, physical geography, oceanography and environmental science programmes. It is designed for a reader who does not require an in-depth knowledge of the technology of remote sensing, but who needs to understand how it is used as a key tool for mapping and monitoring global change.

1.1 Key concepts

1. The Earth’s climate is getting warmer and the patterns of the weather and ocean currents are changing. Severe droughts and flooding are becoming more prevalent and the ice sheets of Greenland and at the poles are disintegrating.

2. The global sea level is rising by about 2 to 3 mm per year, threatening to inundate many coastal areas by the end of this century.

3. Remote sensors on satellites offer an effective way for monitoring environmental trends on a global scale. They can detect physical and biological changes in the atmosphere, in the oceans and on land. Satellite systems have become the defining technology in our ability to quantify global change.

4. The accuracy and applicability of satellite imagery is constantly improving due to technological advances, such as finer spectral/spatial resolution, more powerful computers, the Global Positioning System (GPS) and Geographical Information Systems (GIS).

5. When these techniques for generating, organizing and analyzing spatial information are combined with mathematical and environmental models, scientists and managers have a means for assessing and predicting the impact of global environmental changes.

Remote Sensing and Global Environmental Change, First Edition. Samuel Purkis and Victor Klemas. © 2011 Samuel Purkis and Victor Klemas. Published 2011 by Blackwell Publishing Ltd.

Chapter 2

Remote sensing basics

Remote sensing is primarily concerned with collecting and interpreting information about an object or landscape from a remote vantage point. The platform can be anywhere, ranging from a balloon just above the surface of the Earth to a satellite hundreds of kilometres away in space (Figure 2.1). Examples of remote sensing include aerial photography, satellite imagery, radar altimetry and laser bathymetry. Coupled with ground measurements, remote sensing can provide valuable information about the surface of the land, the oceans and the atmosphere.

Figure 2.1 Remote sensing platforms.

Techniques for acquiring aerial photographs were already developed in the 1860s; however, colour and colour-infrared (CIR) aerial photographs were not widely used until the 1940s and 1950s. The 1950s and 1960s marked the appearance of remote sensing applications for airborne radar and video technologies.

A significant event in terms of land remote sensing was the 1972 launch of the first Landsat satellite, originally called ERTS-1. The satellite was designed to provide frequent broad-scale observations of the Earth’s land surface. Since 1972, additional Landsat satellites have been put into orbit. Other countries, including France, Japan, Israel, India, Iran, Russia, Brazil, China, and perhaps North Korea, have also launched satellites whose onboard sensors provide digital imagery on a continuous basis.

2.1 Electromagnetic waves

Electromagnetic (EM) energy refers to all energy that moves with the velocity of light in a harmonic wave pattern. A harmonic pattern consists of waves that occur at equal intervals in time. EM waves can be described in terms of their velocity, wavelength and frequency. All EM waves travel at the same velocity (c). This velocity is commonly referred to as the speed of light, since light is one form of EM energy. For EM waves moving through a vacuum, c = 299,793 km sec¹ or, for practical purposes, c = 3 × 10⁸ m sec¹.

The wavelength (λ) of electromagnetic waves is the distance from any point on one cycle or wave to the same position on the next cycle or wave (Figure 2.2). The units of measurement usually used in conjunction with the electromagnetic spectrum are the micrometre (μm), which equals 1 × 10⁶ metres, or the nanometre (nm), which equals 1 × 10⁹ metres. This book adopts the nanometre as its unit of wavelength.

Figure 2.2 An electromagnetic wave, including a sinusoidal electric wave E and a similar magnetic wave M at right angles, both being perpendicular to the direction of propagation.

Unlike velocity and wavelength, which change as EM energy is propagated through media of different densities, frequency remains constant and is therefore a more fundamental property. Electronic engineers use frequency nomenclature for designating radio and radar energy regions, but this book uses wavelength rather than frequency in order to simplify comparisons among all portions of the EM spectrum.

Velocity (c), wavelength (λ), and frequency (f) are related by c = λf. The direction of the E field vector determines the polarization of the wave. Thus the EM wave in Figure 2.2 is vertically polarized.

2.2 The electromagnetic spectrum

The electromagnetic spectrum is the continuum of energy that ranges from nanometres to metres in wavelength, travels at the speed of light and propagates through a vacuum such as outer space. All matter radiates a range of electromagnetic energy such that the peak intensity shifts toward progressively shorter wavelengths as the temperature of the matter increases. Figure 2.3 shows the EM spectrum, which is divided on the basis of wavelength into regions. The EM spectrum ranges from the very short wavelengths of the gamma-ray region (measured in fractions of nanometres) to the long wavelengths of the radio region (measured in metres).

Figure 2.3 The electromagnetic spectrum.

Remote sensors have been refined over the past several decades to cover the visible, reflected infrared, thermal infrared and microwave (radar) regions of the EM spectrum (Figure 2.3). Besides photographic and video cameras, digital satellite sensors are used for mapping vegetation, general land cover, and ocean features. As we will see in Chapter 3, some of these sensors are passive, like infrared scanners and microwave radiometers, while others are active, such as radar and laser altimeters.

2.3 Reflectance and radiance

Electromagnetic waves that encounter matter, whether solid, liquid, or gas, are called incident radiation. Interactions with matter can change the following properties of the incident radiation: intensity; direction; wavelength; polarization; and phase. Remote sensors detect and record these changes. We then interpret the resulting images and data to determine the characteristics of the matter that interacted with the incident electromagnetic energy (Sabins, 2007).

During interactions between electromagnetic radiation and matter, mass and energy are conserved according to basic physical principles. Therefore, the incident radiation can only be:

transmitted i.e. passed through the substance. Transmission of energy through media of different densities, such as from air into water, causes a change in the velocity of electromagnetic radiation.

absorbed giving up its energy largely to heating the matter.

emitted by the substance, usually at longer wavelengths, as a function of its structure and temperature.

scattered that is, deflected in all directions. Relatively rough surfaces which have topographical features of a size comparable to the wavelength of the incident radiation produce scattering. Light waves are scattered by molecules and particles in the atmosphere whose sizes are similar to the wavelengths of the incident light.

reflected that is, returned from the surface of a material with an angle of reflection equal and opposite to the angle of incidence. Reflection is caused by surfaces that are smooth relative to the wavelength of incident energy. As shown in Figure 2.4, the reflectance of different land covers varies significantly as a function of wavelength over a wide range of values.

Figure 2.4 Spectral reflectance of various land cover types in the visible spectrum.

Emission, scattering, and reflection are called surface phenomena because these interactions are determined primarily by properties of the surface, such as colour and roughness. Transmission and absorption are called volume phenomena because they are determined by the internal characteristics of matter, such as density and conductivity. The particular combination of surface and volume interactions with any particular material depend on both the wavelength of the electromagnetic radiation and the specific properties of that material. These interactions between matter and electromagnetic waves are recorded on remote sensing images, from which one may interpret the characteristics of matter (Sabins, 2007).

Remotely-sensed imagery is usually digitized so that it can be stored, analyzed, and displayed in digital form on a computer. Although imagery derived from camera and video sensors can be digitized, modern satellite sensors produce data that is already in digital form. Like the images on a television screen, a digital image is composed of an array of picture elements, or ‘pixels’, which represent the smallest part of a digital image. Each pixel contains reflectance information about the features being imaged. Reflectance is what allows us to distinguish the features in an image. It is a measure of the amount and type of energy that an object reflects (rather than absorbs or transmits).

Reflectance is given the notation R and is unitless, meaning that it is represented by a scale between 0 and 1 or 0 and 100 per cent. Reflected wave intensity (power) modified by the atmosphere between the ground and the sensor is called radiance (L). Radiance is what the sensor measures, and it may be very different from the ground reflected intensity because of haze and other substances which scatter light (Lachowski et al., 1995). Radiance has specific units and is typically quoted in watts per steradian per square metre per nanometre (W sr¹ mr² nm¹). These terms will be revisited in more detail in Chapter 8, where we consider the flux of photons in the shallow waters atop coral reefs.

2.4 Atmospheric effects

To an observer, the atmosphere seems to be essentially transparent to light, and we tend to assume that this condition exists for all electromagnetic energy In fact, the gases of the atmosphere absorb electromagnetic energy at specific wavelength intervals called absorption bands. Figure 2.5 shows these absorption bands, together with the gases in the atmosphere responsible for the absorption.

Figure 2.5 Atmospheric absorption and transmission bands. Note change of wavelength units from nanometres (nm) to centimetres in the microwave spectrum. 1 nm = 1.0 × 10⁷cm.

The most efficient absorbers of solar radiation are water vapour, carbon dioxide and ozone. Wavelengths shorter than 300 nm are completely absorbed by the ozone (O3) layer in the upper atmosphere (Figure 2.5). This absorption is essential to life on Earth, because prolonged exposure to the intense energy of these short wavelengths destroys living tissue. Clouds consist of aerosol-sized particles of liquid water that absorb and scatter electromagnetic radiation at wavelengths less than about 0.1 cm. Only radiation of microwave and longer wavelengths is capable of penetrating clouds without being scattered, reflected or absorbed (Sabins, 2007).

Atmospheric scattering is caused by particles in the atmosphere deflecting EM waves in all directions and can be of three types: Rayleigh scatter, Mie scatter and non-selective scatter.

Rayleigh scatter dominates when radiation interacts with atmospheric molecules and other tiny particles which have a much smaller diameter than the wavelength of the EM radiation. Rayleigh scatter is inversely proportional to the fourth power of the EM wavelength; therefore short wavelengths are scattered much more than long wavelengths. The blue sky colour is caused by Rayleigh scatter, since it scatters blue light more than the longer wavelength colours such as green and red. Rayleigh scatter is one of the primary causes of haze in imagery. Haze can be minimized by using camera filters that do not transmit the short wavelengths (Figure 2.6).

Figure 2.6 Relative scatter as a function of wavelength for various levels of atmospheric haze.

Mie scatter dominates when the atmospheric particle diameters are approximately equal to the radiation wavelength. Water vapour, dust and various aerosols are the main causes of Mie scatter. Mie scatter affects all EM wavelengths, including long ones.

Non-selective scatter occurs when the diameter of the scattering particles is much larger than the EM wavelength. Thus, water droplets, having diameters from 50 to 1000 nm, scatter visible, near-IR and mid-IR wavelengths nearly equally.

To minimize image degradation by the various types of scatterers in the atmosphere, a wide range of atmospheric correction techniques have been developed, including ‘dark-object subtraction’ and aerosol modelling. This is particularly important for remote sensing from satellites and high-altitude aircraft (Jensen, 2007; Lillesand & Kiefer, 1994).

2.5 Multispectral feature recognition

Two objects that may be indistinguishable in one portion of the electromagnetic spectrum, because their reflectance is similar in this portion, may be highly separable in another, where their reflectance differs more markedly. To help capture the spectral uniqueness of ground objects, many sensors are designed to collect information in specific regions of the electromagnetic spectrum (Figure 2.3).

Figure 2.7 shows the four spectral bands of a QuickBird satellite image, including their mutually-registered pixels. A pixel or picture element is the smallest part of a digital image, and in most cases the pixel size represents the spatial resolution of the image. Most digital sensors collect information this way, and the acquired multispectral data are stored in spectral bands, where each pixel has a unique intensity value expressed as a digital number (DN). The DN represented by each pixel is an averaged reflectance value for the corresponding objects on the ground.

Figure 2.7 The anatomy of a multispectral QuickBird image containing unique information in four spectral bands.

Typically, changes in land cover (e.g., vegetation) will correspond to changes in the digital numbers of pixels in an image. If the image is multispectral (i.e. it contains fewer than ten or so broad spectral bands), the changes occurring in each band can help differentiate one land cover type from another. Figure 2.8 shows typical reflectance curves for some common ground covers. Since remote sensing platforms, such as satellites and aircraft, are moving rapidly and their sensors are scanning a wide swath of ground, there is inadequate time for a sensor to also scan the entire spectrum for each point on Earth. As a result, most multispectral scanners have a limited number of spectral bands resulting in discontinuous spectral curves, as shown in Figure 2.8 for a 4-band system. The 4-band spectrum in Figure 2.8 is a poor representation of the continuous spectrum, yet quite adequate for discriminating between major types of land cover.

Figure 2.8 Representative spectra for a terrestrial landscape as acquired using the hyperspectral AVIRIS sensor aboard the ER-2 platform (black lines), compared to the crude spectral capability of a 4-band multispectral satellite (gray bars). In both cases, the three land cover types are separable. Most airborne sensors point continuously downward along the local vertical.

2.6 Resolution requirements

With the wide variety of remote sensing systems available, choosing the proper data source for observing land cover, the oceans or the atmosphere can be challenging. Characteristics often used to describe and compare these analogue and digital systems are grouped into four different types of resolution: spatial; spectral; radiometric; and temporal. Resolution is commonly attributed to an image and the sensor that provides the image data.

Spatial resolution is a measure of sharpness or fineness of spatial detail. It determines the smallest object that can be resolved by the sensor, or the area on the ground represented by each picture element (pixel). For digital imagery, spatial resolution corresponds to the pixel size. Spatial resolution is often represented in terms of distance (e.g. 30 metres, 1 km, etc.) and describes the side length of a single pixel. Thus, the smaller the distance, the higher the spatial resolution (the finer the ‘grain’) of the image (Lachowski et al., 1995). For instance, some weather satellites have resolutions of 1 km, with each pixel representing the average brightness over an area that is 1 km × 1 km on the ground.

Spectral resolution is a measure of the specific wavelength intervals that a sensor can record. For example, while normal colour photographs show differences in the visible region of the electromagnetic spectrum, colour infrared photographs and the majority of digital sensors can provide information from both visible and infrared (IR) regions of the spectrum. For digital images, spectral resolution corresponds to the number and location of spectral bands, their width, and the range of sensitivity within each band (Jensen, 2007).

Radiometric resolution is a measure of a sensor’s ability to distinguish between two objects of similar reflectance. Radiometric resolution can be thought of as the sensor’s ability to make fine or ‘subtle’ distinctions between reflectance values. For example, while the Landsat Thematic Mapper (TM) has a radiometric resolution of 256, the Moderate Resolution Imaging Spectrometer (MODIS) has a radiometric resolution of 4,096. This means TM can identify 256 different levels of reflectance in each band, while MODIS can differentiate 4,096. Thus, MODIS imagery can potentially show more and finer distinctions between objects of similar reflectance (Campbell 2007). One can also say that TM has 8-bit resolution (8 bit = 2 raised to the power of 8 = 256) and MODIS has 12-bit resolution (12 bit = 2 raised to the power of 12 = 4,096).

Temporal resolution is a measure of how often the same area is visited by the sensor. Unlike the three types of resolution discussed above, temporal resolution does not describe a single image, but rather a series of images that are captured by the same sensor over time. While the temporal resolution of satellite imagery depends on the satellite’s orbit characteristics, aerial photography obviously requires special flight planning for each acquisition. Temporal resolution for satellite imagery is represented in terms of the amount of time between satellite ‘visits’ to the same area (e.g. two days for SeaWiFS, 16 days for Landsat TM, 26 days for SPOT, 35 days for IKONOS).

The temporal and spatial resolution requirements for detecting, mapping and monitoring selected terrestrial, oceanic, and atmospheric features and processes are shown in Figure 2.9. For instance, urban infrastructure and emergency response both require high spatial resolution, yet urban infrastructure does not change rapidly and needs to be observed much less frequently than emergency situations such as when hurricanes make landfall. On the other hand, monitoring of the weather and climate both require less spatial resolution, yet climate changes slowly whereas weather conditions can change very rapidly.

Figure 2.9 The temporal and spatial resolution requirements vary widely for observing terrestrial, oceanic, and atmospheric features and processes. Modified after Jensen (2007) and Phinn et al. (2010).

As shown in Figure 2.9 and will be seen in later chapters, each application of remote sensors has its own unique resolution requirements (Bissette et al., 2004; Mumby & Edwards, 2002). Since none of the existing remote sensing systems will perfectly meet all resolution requirements, one must always make trade-offs between spatial resolution and coverage, spectral bands and signal-to-noise ratios, etc.

Demanding the highest spatial resolution available may not always result in better image classification accuracy (Wang, 2010). Factors such as computer processing time must also be considered when choosing types of imagery; an increase in resolution will increase the processing time required for interpreting the information. Also, smaller pixels require more storage space on computers for a given area. For example, an image with 10-metre pixels would require significantly more storage space and time to process than a lower-resolution 20-metre pixel image, since the amount of data bits to be processed will increase at least four-fold.

2.7 Key concepts

1. Remote sensors on satellites and aircraft operating at various altitudes use electromagnetic (EM) waves effectively to detect and map features and changes on the surface of the land, sea and in the atmosphere.

2. Remote sensing systems can provide digitized data in the visible, reflected infrared, thermal infrared and microwave regions of the electromagnetic spectrum.

3. Differences in reflectance allow remote sensors to distinguish between objects and features on the ground.

4. The atmosphere absorbs and scatters EM waves. Remote sensors are designed to avoid absorption bands and they use filters or atmospheric correction techniques to diminish the effects of scattering.

5. When planning a remote sensing campaign, one must define the spatial, spectral, radiometric and temporal resolution requirements for the project.

Remote Sensing and Global Environmental Change, First Edition. Samuel Purkis and Victor Klemas. © 2011 Samuel Purkis and Victor Klemas. Published 2011 by Blackwell Publishing Ltd.

Chapter 3

Remote sensors and systems

3.1 Introduction

Aerial photography started approximately in 1858 when the famous photographer, Gaspard Tournachon, obtained the first aerial photographs from a balloon near Paris. Since then, aerial photography has advanced, primarily during wartime, first to include colour infrared films (for camouflage detection) and later to use sophisticated digital cameras. Aerial photography and other remote sensing techniques are now used successfully in agriculture, forestry, land use planning, fire detection, mapping wetlands and beach erosion, and many other applications. For example, in agriculture it has been used for land use inventories, soil surveys, crop condition estimates, yield forecasts, acreage estimates, crop insect/pest/disease detection, irrigation management and, more recently, precision agriculture (Jensen, 2007).

Since the 1960s, ‘remote sensing’ has been used to describe a new field of information collection that includes aircraft and satellite platforms carrying electro-optical and antenna sensor systems (Campbell, 2007). Up to that time, camera systems dominated image collection and conventional photographic media dominated the storage of the spatially varying visible (VIS) and near-infrared (NIR) radiation intensities reflected from the Earth.

Beginning in the 1960s, electronic sensor systems were increasingly used for collection and storage of the Earth’s reflected radiation and satellites were developed as an alternative