WORLD METEOROLOGICAL ORGANIZATION OPERATIONAL HYDROLOGY REPORT No. 43

CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

by A. Rango and A.I. Shalaby

WMO-No.884

Secretariat of the World Meteorological Organization - Geneva - Switzerland 1999 Cover: NOAA-AVHRRfalse color composite image illustrating extent offload waters on the Red River ofthe North in Minnesota and North Dakota, USA during April 1997. Theflood waters are shown in white, remaining snow cover in red, clouds in yellow and orange, and open land in blue. The image was produced by the National Operational Hydrologic Remote Sensing Center, National Weather Service; National Oceanic and Atmospheric Administration (courtesy ofT. Carroll and D. Cline)

© 1999, World Meteorological Organization

ISBN: 92-63-10884-6

NOTE

The designations employed and the presentation of material in this publication do not imply the expression of any opinion whatsoever on the part of the Secretariat of the World Meteorological Organization concerning the legal status of any country, territory, city or area, or ofits authorities, or concerning the delimitation ofits frontiers or boundaries. CONTENTS

Page

FOREWORD . v

SUMMARY (English, French, Russian, Spanish) . vii

CHAPTER 1. INTRODUCTION . 1 1.1 Literature on applications of remote sensing in hydrology , . 1 1.2 Citations ofoperational applications of remote sensing in hydrology ", .. 1 1.3 Remote sensing for hydrologists . 1 1.3.1 Background . 1 1.3.2 Definitions . 1 1.3.3 Theory . 1 1.4 Definition ofoperational applications of remote sensing in hydrology ,. 2

CHAPTER 2. APPLICATIONS OF REMOTE SENSING IN HYDROLOGY ••••.•••..•••..•...•...... 3 2.1 Technical and economic aspects . . 3 2.1.2 Economic consideration . . 3 2.2 Remote s.ensing measurements of hydrological variables . 3 2.2.1 Precipitation . 3 2.2.1.1 Visible and infrared rainfall monitoring techniques . 4 2.2.1.2 Passive microwave rainfall monitoring techniques . 4 2.2.1.3 Multispectral/microwave sounding unit rainfall monitoring techniques .. 5 2.2.1.4 Ground-based radar rainfall monitoring techniques . . 5 2.2.1.5 Satellite radar rainfall monitoring ...... 8 2.2.1.6 Water vapor imagery rainfall monitoring techniques . 8 2.2.2 Soil moisture and groundwater ,. 8 2.2.2.1 Soil moisture ...... 8 2.2.2.2 Groundwater...... 10 2.2.3 Evapotranspiration...... 11 2.2.3.1 Semi-empirical methods ...... 12 2.2.3.2 Analytical methods ...... 12 2.2.3.3 Numerical models ...... ,. 13 2.2.4 Snow...... 13 2.2.4.1 Snow cover distribution ,. 14 2.2.4.2 Snow accumulation ...... 16 2.2.4.3 Snow metamorphism...... 17 2.2.4.4 Snowpack ablation and snowmelt runoff . . 17 2.2.5 Ice on land ...... , . 18 2.2.6 Surface water ...... 19 2.2.6.1 Surface water area ...... 20 2.2.6.2 Surface water stagellevel ...... 22 2.2.6.3 Surface water depth...... 22 2.2.6.4 Surface water temperature...... ,. 22 2.2.6.5 Surface water salinity and mineralization , ,...... 23 2.2.6.6 Surface water suspended sediments .. 23 2.2.6.7 Surface water chlorophyll . 23 2.2.6.8 Other pollutants in surface water . 23 2.2.6.9 Surface water vegetation biomass 24 2.2.7 Basin characteristics . 24 2.3 Remote sensing inputs for hydrological models 27 2.4 Current operational applications .. 28 2.4.1 Precipitation . 28 2.4.2 Soil moisture and groundwater . 29 2.4.3 Evapotranspiration . 29 2.4.4 Snow . 29 2.4.5 Icc on land . 29 2.4.6 Surface water . 29 2.4.7 Basin characteristics 30 2.5 Future trends ... 30 iv CONTENTS

Page

CHAPTER 3. NEW REMOTE SENSING ADVANCES...... •.. ...••.•...... •...... •...... 31 3.1 Recent state-of-the-art remote sensing technology advances and ongoing research 31 3.1.1 Satellites ...... 31 3.1.2 Aircraft...... 31 3.1.3 Ground-based methods and radars ...... 32 3.2 Potential new applications in operational hydrology _...... 32

CHAPTER 4. PROBLEMS IN APPLICATIONS OF REMOTE SENSING IN OPERATIONAL HYDROLOGY ., 34 4.1 Current limitations ofoperational utilization ofremote sensing _...... 34 4.1.1 Status in developed countries ...... 34 4.1.2 Status in developing countries _.. _ _. 34 4.2 Problem areas preventing the operational utilization of remote sensing in hydrology 34 4.2.1 Hydrological variables ...... 34 4.2.2 Hydrological models 35 4.2.3 Remote sensing technology _. . __ _. . . 35 4.2.3.1 Limitations of the state-of-the-art ,._ , , 35 4.2.3.2 Consistency of data and need for calibration ofnew sensors ,...... 36 4.2.4 Availability ofdata ...... 36 4.2.4.1 Geographical or climatological regions where remote sensing data are most applicable 36 4.2.4.2 Data acquisition and distribution capabilities of countries _, _ _. _. . . 36 4.2.5 Cost of data...... 37 4.2.6 Computer data-processing capability .. _.. __ _...... 37 4.2.7 Data presentation/publication/dissemination _ ,...... 37 4.2.8 Human resources and training capability , ,...... 37

CHAPTER S. A STRATEGY FOR ACCEPTANCE OF REMOTE SENSING IN OPERATIONAL HYDROLOGY 39 5.1 Past studies ofoperational hydrological observational needs and existing and future satellites...... 39 5.2 Development or modification ofhydrological models based on remote sensing infonnation 39 5.3 Demonstration ofsignificant technical and economic benefits _ ,,.,.... 39 5.4 Targeting major hydrological services 40

CHAPTER 6. ELEMENTS OF SUCCESSFUL REMOTE SENSING IN OPERATIONAL HYDROLOGY •..••.•• 41 6.1 Financial resources ...... 41 6.2 Regional remote sensing centers , "...... 41 6.2.1 Expert advice _...... ,...... 41 6.2.2 Training of personnel . __ ...... _...... 42 6.3 Ground receiving stations ...... 43 6.4 Groundtruth networks . , _...... 43 6.5 Key ancillary support systems ,,...... _...... 43

CHAPTER 7. RECOMMENDATIONS 4S 7.1 Key approaches to include in developed countries , , ,,,_ 45 7.2 Key approaches to include in developing countries ,...... 45

REFERENCES 46

APPENDIX 1. Cited literature on applications of remote sensing in hydrology •...•..•••••••...... •••••.•..••• 58

APPENDIX 2. Hydrological observational requirements and existing and planned satellites and sensors for hydrological applications. •...... •...... •...•..••••.....•.•...•.....•.•....•.•.•.•.... 60

APPENDIX 3. Specific sensors which might be carried by EOS 6S

APPENDIX 4. StateMofatheaart of satellite data requirements and required capabilities ofnear-polar orbiting and geostationary satellites .....••.•.•.•...... •.••.... •..•...... •••.•••...... 66

ACRONYMS 71 FOREWORD

Over the years, WMO has taken a particular interest in the potential of remote sensing for operational use in hydrology. Accordingly, WMO has issued a number of reports on the subject with the aim of advising national Hydrological Services of the current state of the art. In the past, the emphasis was mainly on the use of remote sensing for research, partly because the techniques had not been fully developed for routine application and partly because the satellites themselves were often designated as being for research with no guarantee of future continuity. This situation has now changed and remote sensing is being widely used in hydrology for a range of purposes. The fact that the opera­ tional community does not publish its results so frequently makes it even more valuable that WMO, through its Commission for Hydrology, maintains its role of monitoring and reporting on developments. At its ninth session in 1993, the Commission appointed Mr A. Rango (United States of America) as its Rapporteur on Applications of Remote Sensing and entrustred him with the task of preparing a report on "current applications of remote sensing in hydrology". To complete this task, Mr Rango invited Professor A.I. Shalaby of Howard University to assist him in putting together the report. It was then recommended for publication by the Commission for Hydrology at its tenth session. I should like to place on record the gratitude of th,e World Meteorological Organization to Drs Rango and Shalaby for the time and effort they have devoted to the preparation of this valuable report.

(G.O.P. Obasi) Secretary-General

SUMMARY

This report provides information on the most recent advances in identifies applications of remote sensing in each subdiscipline, and operational use of remote sensing in hydrology and water identifies those applications that qualify as operational. New resources management. The application of remote sensing data advances in remote sensing are briefly surveyed in Chapter 3. The can be termed operational if at least one of two conditions is various problems that hinder operational applications are covered fulfilled: 1) the application produces an output on a regular basis; in Chapter 4. Chapter 5 sets forth a strategy for acceptance of or 2) the remote sensing data are used regularly on a continuing remote sensing in operational hydrology, while Chapter 6 lists the basis as part of a procedure to solve a problem or make decisions. elements needed for successful operational remote sensing appli~ The report consists of seven chapters. The first chapter cations. Chapter 7 includes recommendations and key approaches introduces the report and makes reference to operational remote to use in both developed and developing countries. sensing applications and various definitions of operational appli· An extensive list of references as well as a listing of cations. Chapter 2 surveys the subdisciplines of hydrology, relevant literature are also included.

Ce rapport [ournit des informations sur les taus demiers progres de la teledetection dans chacune d'eHes et detennine celles que l'on accomplis dans I'utilisation operationnelle de la t616d6tection aux fins peut qualifier d'operatioIUlelles. 1£ Chapitre 3 expose brievement les de I'hydrologie et de la gestion des ressources en eau. L'application faits nouveaux survenus dans Ie domaine de la t6ledetection, apres quai des donnees de tel6detection pelit ctre qualifiee d'operationnelle si le Chapitre 4 traite des problemes auxquels se heurtent les applications rune au moins des deux conditions ci-apres est remplie : 1) l'appli­ operatioIUlelles. Le Chapitre 5 decrit une strategie pour l'adoption de la cation genece regulierement des produits; ou 2) les donnees de t61edetection en hydrologie operationnelle, tandis que Ie Chapitre 6 tel6d6tection sont utilisees de fa<;on suivie pour resoudre un probleme dOIUle la liste des elements areuoir pour que les applications opera­ ou prendre des decisions. tionnelles de la t616detection soient couronn6es de succes. Le Chapi­ Le rapport se compose de sept chapitres. Le premier sert tre 7 contient des recommandations et donne des exemples de d'introduction, 6voque les applications operationnelles de la teledetec­ demarche asuivre dans les pays developpes et ceux en developpement. tion et en donne diverses definitions. Le Chapitre 2 recense les Le rapport contient une liste tres complete de references sousRdisciplines de I'hydrologie, indique quelles sont les applications ainsi qu'une bibliographie de la litreranrre pertinente.

PE310ME

BHaeTOJIUJ;eM OT'IeTe npe)J;oCTaBillIeTCJI HH¢opMawul 0 caMblX nOCJIe.n;­ 30HlI,HpOBaHI1H B KmKj:(oM no.npa3;:J;ene, a TaIOKe orrpe;:J;eJIHlOTCJl Te npH­ HHX )J;oenDKeHIDIX B onepaTHBHoM HCUOlIb30BamfH .lUiCTaHl.\HOHHOro MeHeHHH, KOTophle KBaJuHIHIIJ;IipYIoTeJl KaK orrepaTHBHbIe. HOBbIe 30llJUlpOnaHIDI n o6nacTIl rHApOJJOrHH H HCnOJIh30naHHH BO)J;HblX pe~ ;:J;ocTH)I{eHHJl B JlHCTaHI..\HOHHOM 30H;:J;HpOBamm KpaTKo paCCMaT­ cypCOB. ITpHMeHeHHe AaHHblX J]JlCTaJrn;HOIIHOfO 30HAHpoBaHIDI MO)KHO pHBaIOTCJl B maBe 3. PaSJIH'lHble np06JIeMbI, KOTopble MellialOT Ha3BaTb onepaTHBHblM, eCJlli BbIII01IHJleTCJI no KpaMHeH Mepe OARO H3 BHeJlpeHHIO onepaTHBHbIX npHMeHeHH:ti, OXBaTbIBalOTCll B maBe 4. B AByX ycnomlli: 1) npHMeHeHJfe AueT pe3YJIbTaT Ha peryJUlpHO:l1: oeHOBe; maBe 5 YCTaHaBJIHBaeTCH cTpaTemJl rrpHeMneMocTH ,lUieTaHUHOHHoro HJlli 2) AaHHble AHeTanUJfOIlHoro 30HJlHpOBaHHJI peryJlllpHO HCn01lb­ SOHj:(HpoBamm B orrepaTIlBHO:l1: rH.nponorI-rn, B TO BpeMH KaK B maBe 6 3YlOTCn. Ha HenpephlBHoH oeROBe B Ka'leeTBe l.£aeTH npOUenypbI, rrepeqllCJIillOTCJl3J1eMeHTbl, Heo6xoj:(HMbre JlllJI ycnellIHhlX orrepaTHB­ npellHa3HaQeHHOH wm perueHlHl npo6neMblMH q,OpMHpOBaHHJI peruemlH. HhIX npHMeneHH:ti JlHCTaHUMoHHoro 30H;:J;HponaHHH. B many 7 OT'IeT COCTOHT H3 ceMH maB. ITepBaJI maBa lIBllileTeJ'I BBe­ BKnIOqeH:bI peKoMeH,naI..\HJl, a TaK)Ke KJIIO'IeSble rrO}J.XOj:(bI K HCnOJIb30­ JleUHeM B on:eT, B HeH nepe'lJICJImOTeH OnepaTJIBHble rrpHMeHeHHH BaHillO ,!lHCTaHUHOHHoro 30HJlMpOBaHIDl B pa3BHTbIX H pa3BHBaJOIUHXCJI JlHCTaHUHOHHoro 30HJlHpOBamIH, a TaK)Ke COAep)KaTCH pa3JIH'IHbIe CTpauax. onpeJleJIeHIDI onepaTHBHblX rrpHMeHeHH::it B maBe 2 npJIBOJ]JlTCH 0630P B OT'IeT BKnIOqeH 0611IHPHbl:ti nepeqeHb CCblJIOK, a TaK­ nOApa3AenoB nmpOJIOrHH, onpe.nellilIOTCll npHMeHeHJlll JlHCTaHll;IlOHHoro )fCe crrHCOK cooTBeTcTByww;efi JIJiTepaTypbI.

SUMARIO

En este informe se facilita infonnad6n sabre los progresos mas hidrologia, se sefialan las aplicaciones de lateledetecci6n encarla una recientes en el usa operativo de la te1edetecci6n en hidrologfa y de elias y se indican las que se consideran operativas. En el Capftulo gesti6n de recursos hidricos. La aplicaci6n de datos obtenidos por 3 se examinan brcvemente nuevos avances en materia de teledetec­ teledetecci6n puede calificarse de operativa si al menos se cumple una ci6n. Los diversos problemas que obstaculizan las aplicaciones de las dos condiciones siguientes: I) la aplicaci6n produce resultado operativas se tratan en el Capitulo 4. En el Capitulo 5 se presenta una con caracter regular; 0 2) los datos obtenidos por teledetecci6n se estrategia para la aceptaci6n de la teledetecci6n en hidrologia operaR utilizan regularmentc en forma continua como parte de un proce­ tiva, en tanto que en el Capitulo 6 se enumeran los elementos dimiento para resolver un problema 0 para tomar decisiones. necesarios para aplicar con 6xito la teledetecci6n operativa. EI EI informe consta de siete capitulos. En el primer capftulo Capitulo 7 comprende recomendaciones y m6todos esenciales para se presenta el informe y se hace referencia a las aplicaciones de telede­ utilizarlos en los paises desarrollados y en desarrollo. tecci6n operativas, con diversas definiciones de las aplicaciones Tambien se incluye un amplia lista de referencias y una practicas. En el Capitulo 2 se analizan las subdisciplinas de la enumeraci6n de la literatura de interes.

CHAPTER 1

INTRODUCTION

The science and practice of hydrology includes managing, assess­ 1.3 REMOTE SENSING FOR HYDROLOGISTS ing and forecasting the quantity and quality of water. Both historical and real-time hydrological data are collected, stored and This section contains information regarding background, theory, analysed. The resulting information is provided to the decision· and definitions of remote sensing for hydrologists. makers to manage the water resources and to mitigate floods, droughts, pollution incidents and similar water-related hazards. An important prerequisite is the availability of accurate and reliable 1.3.1 Background data. The International Conference on Hydrology (UNESCOI WMO/ICSU, 1993) recognized that a universal difficulty encoun­ The application of remote sensing techniques in hydrology is tered in the science and practice of hydrology is associated with becoming increasingly important mainly because of several unique data. aspects of remote sensing. First, remote sensing techniques have Up until a decade ago, hydrologists relied mostly on the ability to measure spatial information as opposed to point data. conventional data network systems. However, characterized by Second, remote sensing teclmiques have the ability to measure the insufficient spatial and temporal coverage of the Earth's surface, state of the Earth's surface over large areas. State variables include the conventional networks are being supplemented by remotely soil moisture, surface temperature, degree of frozen soil and sensed data network systems. surface energy balance. Finally, remote sensing techniques, espe­ The most important areas of application in hydrology cially those which utilize satellite sensors, have the ability to and water management include the following: assemble long-term data (Engman and Gurney, 1991). (a) precipitation; Furthermore, although the expense associated with remotely (b) soil moisture, groundwater and irrigation; sensed data may seem to be -high at ftrst glance, upon close inspec­ (c) evapotranspiration; tion the cost is very low when compared to acquiring the same (d) snow; spatial information using conventional means (Rango, 1994b). (e) ice on land; (j) surface water; and (g) basin characteristics. 1.3.2 Definitions Although existing techniques for remote sensing appli­ cations in these areas are generally either preliminary and/or Remote sensing is the observation of a target by a sensor without dependent upon data much less well suited to them than is desir­ physical contact. The basis for all remote sensing is the electro­ able (Barrett et aI., 1990), remotely sensed data are now being magnetic spectrum (EMS) illustrated in Figure 1. Remote sensing used operationally by a number of agencies for rainfall and snow takes advantage of th~ unique interaction of radiation from the monitoring, flood forecasting and water balance assessments. specific regions in the spectrum and the Earth (Engman and Gurney, 1991). There are four basic components of a radiation­ based remote sensing system, given as follows: 1) radiation source 1.1 LITERATURE ON APPLICATIONS OF REMOTE (e.g. the sun, radar, laser); 2) transmission path (e.g. atmosphere, SENSING IN HYDROLOGY vegetation canopy, shallow soil surface); 3) target (e.g. river, soil, snowpack); and 4) sensor (e.g. multispectral scanner, thermal The existing literature on the applications of remote sensing in sensor, microwave sensor). Remote sensing is based on interpret­ hydrology are numerous and may be found in various types of ing measurable variations in spectral, temporal and spatial documents including textbooks, technical reports, proceedings, characteristics of the Earth. Spectral characteristics (or signatures) journals and others. Only a representative collection of such docu­ of the target are the unique spectral reflectances for specific earth ments will be referenced in this report. Textbooks are listed in features (Engman and Gurney, 1991). Appendix I, A. Technical reports, proceedings, journals and others are listed in Appendix I, B, C, D and E, respectively. 1.3.3 Theory

1.2 Citations of operational applications of Remote sensing techniques measure the reflective, thermal or remote sensing in hydrology dielectric properties of the Earth's surface, within the EMS (see Figure 1). Different sensors can provide unique information about Many applications of remote sensing in hydrology have been the properties of the surface or upper layers of the soil (Engman made, but true operational applications are limited. Authors and Gurney, 1991). Aerial photography/cameras use the visible, reporting either operational or potentially operational applications near-infrared (IR) and thermal regions of the spectrum; photo­ of remote sensing for a specific area in hydrology are listed in the graphic images provide information on sediment plumes, erosion box in Appendix I by specific area of application. features, discharges from pipes and spills (Engman and Gurney, 2 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

1991). Multispectral scanners (MSS) measure the reflected solar used to estimate high-density topographic information, canopy radiation; the data are used to estimate land cover, land use, vegeta­ height, vegetation cover, stream valley cross-sections, air pollution tion biomass, soil type, vegetation type, snow cover, water area, and chlorophyll (Engman and Gurney, 1991). impervious area and various water quality parameters (Engman and Gurney, 1991). Thermal scanners (TSC) directly measure energy emitted from the Earth's surface/surface temperature; the data are 1.4 DEFINITION OF OPERATIONALAPPLICATIONS used to estimate evapotranspiration, soil moisture, drainage OF REMOTE SENSING IN HYDROLOGY patterns. groundwater seepage zones, canopy temperatures, and thermal plumes from thermoelectric power plants or industrial Based on Applications of remote sensting by satellite, radar and sources (Engman and Gurney, 1991). Gamma ray spectrometers other me/hods /0 hydrology (WMO-No. 804) (Rango, 1994b), a (radiometers) measure the gamma radiation flux by soil water or follow-up study has been undertaken of the current operational appli­ snow; data are used to estimate snow water equivalent and change cations of remote sensing in hydrology. According to Rango (l994a, in soil moisture. Microwave radiometers are passive sensors, which 1994b), the application of remote sensing in hydrology may be work in the same way as thermal infrared (TIR) radiometers, and termed "operational" if: 1) the application produces an output on a can directly measure energy emitted from the Earth's surface and regular basis; or 2) the remote sensing approach is used regularly and the dielectric properties; data are used to estimate soil moisture, on a continuing basis as part of a procedure to solve a pr9blem. vegetation type, snow water equivalent, condition of snowpack, However, according to Blyth (1993), for an application to be classi­ frozen soil and sea ice. Furthermore, active microwave remote fied as operational, it is not essential that the technique has already sensors or radars such as Side-looking Airborne Radar (SLAR), been demonstrated operationally, but simply that current sensors and Synthetic Aperture Radar (SAR), altimeters and scatterometers knowledge are adequate to make incorporation into an operational measure the reflected or backscattered radiation (energy) propa­ system feasible to a level where an improvement to that system can gated towards the Earth's surface by the radar (radiation source); be expected. As such, there are numerous examples of operational such data are used to measure rainfall (Engman and Gurney, 1991). applications ofremote sensing in hydrology. Examples falling under And, finally, laser remote sensors measure the energy reflected by both definitions will be presented in the following chapters. The the Earth's surface from a laser beam (radiation source); data are reader can make the final decision on operational readiness.

HERTZ I I IIII I I II I I I I I Frequency '9 16 15 14 13 12 '0 9 7 6 10'0 10 10" 10" 10 10 10 10 10 10" 10 10 10' 10 10 (cycles per I I I second) 00 W ZW ~ ~ «'" Q I~I~I ~~ 3: l.L.I.1.. <;;; ;1i 0 5J:~ >-'" '" =>z " "x o~ u'" '" " (IZI W« ~ ~~

I I METRES Wavelength

1nm lmm 1m

Relative 100 ~ Transmission 1 through transmission 0...------...... -.. atmosphere Total x-ray Atomic Imaging Radiometry, Passive Electro gamma imaging absorbtion single- and spectrometry, microwave magnetic Principal ray spectro­ multi-lens thermography radiometry, sensing counts, photometry cameras radar techniques gamma ray mechanical various film imaging for remote spectrometry line emulsions, scanning multispectral sensing photography

Figure 1-An illustration of the electromagnetic-spectrum showing the relationship between wavelength and frequency, the common names for spectral bands, and the relative atmospheric transmission. The principal techniques used in remote sensing are also shown for their respective spectral region (after Barrett and Curtis, 1976) CHAPTER 2

APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

2.1 TECHNICAL AND ECONOMIC ASPECTS instruments may be developed only after a careful consideration of the basin characteristics. Remote sensing has proven to yield Remote sensing has shown promise for the observation of hydro­ necessary comprehensive map information not represented on logical change over time and space. The hydrological variable is conventional maps (Ringrose and Matheson, 1991; Baker et at.• indirectly measured~ the electromagnetic radiation which is sensed 1991). Higuchi and Iozawa (1971) and Dozier (1989) illustrated is related to the hydrologic variable either empirically or via some that an optimum design of snow sampling regimes is possible form of transfer function. when remote sensing is used to monitor snow accumulation and The hydrological observational requirements may be melt patterns. Remote sensing of other variables such as glacier viewed in more than one way. A summary of hydrological obser­ areas and landcover is quite useful in mountainous regions vational requirements in relation to satellite monitoring because of restricted access to these regions. Collins and Van opportunities is given in Appendix II, Table 1 (after Barrett et aI., Genderen (1978) illustrated the use of remote sensing for 1990). The hydrological observational needs in relation to a hypo­ mapping, navigation and site selection in developing countries thetical platform HYDROSAT is given in Appendix II, Table 2 where conventional maps may be poor. According to Blyth (1993), (after Onnsby and Engman, 1993). Furthermore, WMO (1993a) "considerable savings can be made during groundwater explo­ has highlighted the relationship amongst the required geophysical ration by the targeting of test boreholes in areas having high parameters, observing technologies and candidate sensors, given groundwater storage potential as identified by remote sensing" in Appendix II. Table 3. A more detailed set of similar tables is (Gurney et al., 1982; Szekielda, 1988). presented in Remote sensing for hydrology - Progress and In an effort to economize through remote sensing. the prospects (WMO-No. 773) (Kuittinen, 1992). Russian Federation has developed the ALMAZ programme. Although most of the existing and planned sensors cover Recent practices have shown that Earth remote sensing from space the visible through the thermal region of the spectrum (see is the most effective method of providing data necessary to Appendix II, Table 4), the use of microwave sensors has been develop the mineral resources and raw materials ofSiberia and the shown to overcome cloud cover, a major factor limiting the use of Far North and other economy-related tasks (Viter, 1993). remote sensing. Blyth (1993) has summarized key satellites carry­ "Microwave radar data will be of unique value in the ing microwave sensors suitable for hydrological monitoring (see tropics or mountainous areas where cloud cov~r is usually persist­ Appendix 11, Table 5). Blyth (1993) has classified application ent, or in arid areas where its sensitivity to surface roughness areas of microwave remote sensing as being either operationally enables spectrally similar surfaces to be differentiated" (Blyth, feasible, near-operational, or requiring development and multifre~ 1993). Low frequency radar has the unique ability to penetrate quency sensors (see Appendix 11, Table 6). Finally, Blyth (1993) in forest canopies (Chauhan and Lang. 1991; Kimura and Kodaira. Appendix II, Table 7 identified the most appropriate microwave 1992); this may facilitate the retrieval of soil and soil moisture sensor/frequency combination likely to be required for the hydro­ infonnation, wetland and floodwater extent. logical variables. Furthermore. it was proposed that the HYDROSAT platform would have only microwave sensors (Ormsby and Engman, 1993). 2.2 REMOTE SENSING MEASUREMENTS OF HYDROLOGICAL VARIABLES

2.1.1 Economic consideration While some of the examples may be termed operational as defmed by Rango (1994a), others may have some operational potential Numerous studies have demonstrated how remote sensing can be according to Blyth (1993). used operationally and economically to replace. supplement. inter­ pret and standardize data provided by conventional techniques. The Institute for Remote Sensing Applications (lRSA) (1992) has 2.2.1 Precipitation reported success in four categories of application. The first is an agricultural study to provide rapid estimates of changes in acreage Rainfall is one of the most important processes in the hydrological and potential yield using high-resolution satellite data at sample cycle. Precipitation data is an important input to many environ­ sites scattered over Europe. The second is the operational land mental monitoring systems. The spatial and temporal depth of cover mapping of Europe with a high degree of accuracy. The rainfall is necessary for the management of water resources such third is investigation of the role of the ocean in the global system as rivers and reservoirs, irrigation, and weather forecasting. as well as on a local scale. The fourth is the monitoring of changes Conventional raingauges provide direct yet only point rates and in the forests of the tropical belt. depths ofrainfall. According to Blyth (1993), the use of remote sensing Remote sensing techniques provide the capability of will yield the most benefit when used at the earliest opportunity in observing precipitation in real- or near-real-time over relatively any study of hydrological catchments, hydrological basin large areas and thus complementing the conventional more accu­ processes or water resources requiring the selection of instrument rate point measurements. Useful data can be derived from sites. For instance, an optimum strategy for siting hydrological satellites used primarily for meteorological purposes. including 4 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY the polar-orbiting National Oceanic and Atmospheric The Finnish Meteorological Institute is applying a modified Administration (NOAA) and Defense Meteorological Satellite neural network algorithm developed by the Swedish Meteoro­ Program (DMSP) sateIIItcs, and the geostatIonary GOES, GMS logical and Hydrological Institute to classify clouds using NOAA and Meteosat (Engman and- Gurney, 1991). The three types of Advanced Very High Resolution Radiometer (AVHRR) data. electromagnetic radiation which have proven to be of special value Precipitation estimates are made in the operational phase (WMO, are the visible (VIS), thermal infrared and microwave. While the 1992a). passive microwave radiometers have been operational to date, • The Polar-orbiting Effective Rainfall Monitoring Interactive plans exist for the use of active microwave rainfall radars on future Technique (PERM11) is used operationally in Pakistan. Daily satellites, e.g. the Tropical Rainfall Measuring Mission (TRiVIlVl) rain/no-rain boundaries are determined using lR imagery. (Barrett, 1993). Also, the recent use of water vapour imagery (6.7 PERMIT was originally developed for use over the Western mm) shows significant promise for the estimation of potential Sahel in Africa (D'Souza et at., 1990). precipitation amounts (Thiao et af., 1993). The Martin-Howland (Grid History) technique (1982) has been designed more specifically for the estimation ofdaily rainfall in 2.2.1.1 VISible and infrared rainfall monitoring techniques the tropics and warm season mid-latitudes. Ithas been applied to estimate daily rainfall over the South China Sea, Arabian Sea, Visible and infrared techniques use measurements of cloud top tropical North Atlantic Ocean, Amazonia, and for monsoon rain­ reflectance and temperature to provide indirect estimates of rain· fall situations along the east coast oflndia. WMO (l992a) reports rate. Operational applications of VIS and IR remote sensing that Japan applies this technique to Geostationary Meteorological techniques in rainfall monitoring are given as follows: Satellite~4 (GMS-4) IR data to analyse monsoon activities, inter­ The NESDIS (NOAA Environmental Satellite Data and annual variations such as the El Nino/Southern Oscillation Information Service) interactive precipitation estimation tech­ (ENSO), inter-seasonal variations in the tropics and activities of nique (also known as the Interactive Flash FloodAnalyser (IFFA) the frontal zone around East Asia. technique) is one of the first and few techniques that can Kilonsky and Ramage (1976) derived monthly estimates over the currently be described as being in continuous operational use. It data-sparse Pacific Ocean using a hard-copy NOAA archive of originates from the Scofield-Oliver technique (1977). In this final polar-orbiting satellite Mercator mosaics - one per day. This form it has been taught to forecasters and has been used opera­ method was later used to provide rainfall maps over the GATE tionally for the real-time precipitation estimation of heavy (GARP Atlantic Tropical Experiment) area and the Atlantic rainfalls using geostationary cloud images over the USA for Ocean. Barrett et al. (1988a) used a similar technique to provide nearly a decade (D'Souzaetal., 1990). NESDIS (NOAA, 1994) a useful Atlas ofPotential Rainclouds over Oman. used the IFFA technique to make rainfall estimates resulting in In the design and implementation of new operational satellite­ the Great Flood of 1993 ofthe Midwest US. based visible/infrared rainfall estimation techniques over large The Food and Agriculture Organization (FAO) of the United regions, it is clear that no single technique would be able to Nations method was developed by the Remote Sensing Unit of provide rainfall estimates at all the spatial and temporal scales the University of Bristol and applied over Africa to assist the required. Thus, it seems appropriate that integrated or "hybrid" FAD's desert-locust and drought monitoring programme. techniques should be made available (D'Souza et al., 1990). As Secondary data user's station data were analysed interactively an example, NESDIS is hoping to implement the Hierarchical to produce maps of "probably precipitating clouds" (Barrett Objective Procedure (HOP) designed to guide objectively the and Hanison, 1986). Resulting montWy maps are being opera­ ~election ofgeographical windows within which the IFFA tech­ tionally used by FAO for locust invasion warning (locusts are nique should be applied. HOP incOlporates elements of several attracted by moist soil) (WMO, 1993a). National personnel other rainfall estimation techniques. have been trained for the fully operational programme which is being implemented in a new remote sensing office in Algiers. 2.2.1.2 Passive microwave rainfall monitoring techniques A new objective and internally self-calibrating geostationary infrared method called "B4" (Bristol/Barrett, Beaumont, While all of the VIS/IR tedmiques suffer from the disadvantages of Bellerby) is now being operationally tested over West Africa. indexing methods (i.e. indirectness), it is fortunate that imagery Raingauge data from real-time reporting Global Telecommuni­ from selected passive microwave bands can be exploited to provide cation System (GTS) stations are being used to set rain/no-rain more direct information of alternative or additional value for rain­ boundaries and rain rates (Barrett, 1993). fall monitoring purposes (Kidd and Barrett, 1990). The greatest An operational approach similar to the B4 method is being devel· constraints on the use of passive microwave techniques stem from oped for the Nile Forecasting System (NFS) using rainfall the fact that the data are available only once or twice daily in most satellite imagery from the European Space Agency's (ESA) tropical and mid-latitudinal areas of the world; data are currently Meteosat geostationary satellite (NOAA, 1993). obtained from military satellites. The spatial resolution of the Lovejoy and Austin (1979a, b) developed a technique using real­ microwave footprints is much coarser than that of the VIS and IR time radar data and infrared imagery for rainfall area estimation. sensors. However, improvements in resolution continue; the latest It has been used quasi-operationally in two real-time (so-called sensor, the Defense Meteorological Satellite Program-Special "nowcasting") rainfall estimating schemes: the large-scale Sensor Microwave/Imager (DMSP·SSrvIll), has a footprint of 15 x RAINSAT in Canada and the FRONTIERS system in the United 13 km in the 85.5 OHz region (Kidd and Barrett, t 990). Further Kingdom (UK) (D'Souza et a!., 1990). WMO (1992a) reports into the future, the rainfall research community awaits a passive that RAINSAT2 has been an operational system for processing microwave sensor on a geostationary platform (Kidd and Barrett, and displaying the integrated data in Quebec since 1991. 1990). Additionally, the PRIRODA instrument package on the CHAPTER 2 - APPLICATIONS OF REMOTE SENSING IN HYDROLOGY 5

Russian MIR space station included a 37 GHz band with an 8 km regions: Mexico, Amazonia, Western Africa and the western resolution and an 85 GHz band with a 5 km resolution. equatorial Pacific Ocean (TOGA COARE area). Since 1978, NESDIS has been operationally supplying Wilheit ef al. (1991) used operationally available SSMII data satellite estimates and outlooks of heavy precipitation to the to derive monthly rainfall indices. Results compared National Weather Service (NWS) Forecasting Offices and River favourably with the GATE radar observations. Forecast Centers. New additional products include: 1) Infrared Histograms of the GOES Precipitation Index; 2) Global Cloud 2.2.1.3 Multispectral/microwave sounding unit rainfall Liquid Water from the Special Sensor Microwave/Temperature monitoring techniques (SSMff) - over water only; and 3) Global Precipitation from the SSMJ] over land and water. A trispectral precipitation product is Simultaneously, capitalizing on the ability of passive microwave expected over the next 3-5 years (Scofield, 1991). algorithms and the temporal frequency of IR data, significant It is generally assumed that a frequency algorithm progress has been made in using the TIROS~N Operational Vertical should be most appropriate over land while a polarization algo­ Sounder (TOVS) retrievals in rainfall forecasting. The Hungarian rithm over water. The Remote Sensing Unit of the University of Meteorological Service successfully used atmospheric soundings Bristol is currently conducting research, based upon SSMII data from the Microwave Sounding Unit (MSU) radiometer on board the from the DMSP Block 5D-2 satellite, aimed to evaluate rainfall TIROS-N to forecast ultra-short-range precipitation in the over both land and sea simultaneously. The initial test area is the Carpathian Basin. This applied procedure can be incorporated into British Isles and the surrounding sea areas (Kidd and Barrett, the current operational rainfall forecasting procedure (Csistar and 1990). A combination of frequency and polarization algorithms is Bonta, 1993). WMO (1992a) reported that the improvement of being explored in order to establish a physically-determined TOVS retrievals has been a particular focus ofmany ofthe research rain/no-rain threshold. For the areas of simultaneous SSMII and efforts found in the country progress reports. Examples of radar coverage, the correspondence between both the areas and the operational applications exist in the Russian Federation, Japan, intensities of rainfall in the radar and the radarMcalibrated satellite Finland and Italy. Furthermore, Randall and Tjemkes (1991) rainfall maps are very good. favourably compared the results from a model which combines Applications of passive microwave remote sensing tech- SMMR and SSM!I data with actual TOVS data to estimate niques in rainfall monitoring arc given as follows: precipitable water. Barrett (1991) used passive microwave data from the DMSP SSM!I over northwest Europe, calibrated by FRONTIERS 2.2.1.4 Ground-based radar-rainfall monitoring techniques radar with good results. He also processed NIMBUS-7 Scanning Multichannel Microwave Radiometer (SMMR) Ground-based radar is conceptually similar tp spaceborne radar data along with SSM!I data to provide historic (climatic) except that the radar is stationary on the land surface and its area analyses of rainfall in the North Sea. of measurement is limited to a circle with a radius up to about 130 • Ferraro et al. (1986) utilized both the frequency algorithm kIn (Kuittinen, 1992). A recording raingauge is desirable for cali­ techniques and temperature thresholding to classify several bration. Thus, the rain rate, R, can be estimated according to geophysical parameters for a North American winter climate. (1) Results proved good enough to distinguish sea ice, dry snow, flooded land, rain, dry land and open water. where Z is the measured radar reflectivity and a and b are calibra­ Kidd (1988) compared frequency and polarization algo­ tion parameters. rithms. Three months (July 1979, and October and November According to D'Souza et ai. (1990), groundMbased 1983) of surface rainfall data and microwave data were weather radars have been used operationally for more than 20 analysed on a rain/no-rain basis. The polarization algorithm years in some countries, mostly in combination with raingauge providcd the most reliable results in the month of July. It was networks which are often used to calibrate them. Rainfall esti­ concluded, however, that this algorithm was seasonally mates made by weather radars arc sometimes more useful than dependent. those made by raingauges because they are continuous in time and Prabhakara et al. (1992) used microwave data from the space and provide areal coverage. However, there are problems SMMR and the SSMII which revealed information on rain associated with them dealing with backscatter, attenuation, and precipitation~sizcd icc; rain rates compared favourably absorption and reflection, particularly in areas of varying relief with radar observation. and signal calibration. Italy is investigating the possibility of using microwave Although the weather radar networks operated in coun­ images from the SSMII on board the DMSP (possibly linked tries such as Europe and North America by the national to the Multifrequency Imaging Microwave Radiometer meteorological agencies adequately service the primary meteoro­ (MIMR) data) on an operational basis mainly for the identifi­ logical requirement of daily weather reporting and forecasting, the cation of precipitation nuclei and sea state (WMO, 1992a). need for quantitative estimates of rainfall to support applications Adler ef al. (1991) combined hourly IR data from the in hydrology and water resources, especially flood forecasting, has Japanese OMS geosynchronous satellite and microwave data not been so well serviced (Moore, 1993a). The visual images of from the DMSP SSM!I to produce IRlmierowave monthly the spatial extent and propagation of storms provided by radars are rainfall estimates. now a familiar feature of weather reporting. Furthermore, the most Negri et al. (1994) used a scattering-based algorithm ofAdler Widespread of all quantitative applications is the use of radar for et ai. (1991) to derive climatologies of convective precipita­ rainfall forecasting. The majority of methods in operational use tion from microwave observations from the SSMII for four are based on an underlying advection model. Others include 6 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY feature tracking· and human/machine interface methods. All of subsequent linear evaluation for nowcasting. The application of these methods are used particularly for flood warning in both rural algorithms for determination ofrainfall intensity and amount in and urban areas. Radar is in widespread use as an informal means an operative procedure will be investigated (WMO, I992a). of initial alert of impending flooding utilizing the moving images Although the infonnal use of radar for flood warning is of storm systems that it provides. However, quantitative use of widespread, only two or three countries make quantitative use of radar data is much less common and is constrained by accuracy the data as part of a flow forecasting system. Even then, the radar limitations, especially in mountainous areas and at times when is used in a complementary way with additional information being bright band effects are apparent (Moore, 1993a). provided by networks of raingauges. First, the radar data are Some applications of radar data in rainfall reporting and corrected for attenuation and other effects, then a raingauge cali­ forecasting arc given as follows: bration is applied to improve accuracy prior to use for flood • The Ballistic Research Laboratory in the US conducted forecasting. The development of distributed flood forecasting experiments to measure the properties of rain backscatter at models (specially tailored to utilize the radar grid data through a 9.375, 35, 70 and 95 GHz. The experiment verified the Mie model grid formulation) are still at a pre-operational stage. Theory calculations; it was verified that there is a decrease in Furthermore, real-time implementation should only be contem­ rain backscatter coefficient above 70 GHz. The 95-GHz radar plated following detailed off-line assessment and proven improved "saw" the target in the rain better than the lower frequency performance relative to simpler and more conventional lumped radars (Richard et al., 1988). models (Moore, 1993a). Below are some applications of real-time Goldhirsh et al. (1987) compared radar-derived rain rate rainfall data from operational radars for flash flood forecasting. results of a dual-polarization radar located at the National Joss and Lee (1994) reported that the Swiss Meteorological Aeronautics and Space Administration (NASA) Wallops Institute was to install during 1993/1995 three new radar Flight Facility with in situ measurements of rain rate obtained stations on three mountain tops. These installations were with raingauges. The radar system gave relatively accurate completed in 1995. The new equipment should give opera­ measurements of rainfall for three rain days. tional flood warning in the Swiss Alps and improved • Hogg (1989) proposed the design of a joint radar/radiometric quantitative information in precipitation rate to be used in a system that can prove viable for quantitative operational national research programme on climate changes and natural measurement of rainfall. Important design factors which must hazards. be considered in providing accurate rain measurements • Walsh (1993) reports that weather radar has been actively with operational radars such as the next generation radar researched and developed in the UK since the 1960s. One of (NEXRAD) are discussed. the major indications leading to the weather radar hydrologi­ Smith and Krajewski (1991) used products of the NEXRAD cal application was the Dee Weather Radar and the Real-time system to estimate hourly rainfall over Norman, Oklahoma. Hydrological Forecasting Project which demonstrated Walton et ai. (1990) discuss the ability of the three-stage successful rainfall measurement. This was followed by the NEXRAD system to issue more accurate forecasts ofprecipi­ North West Weather Radar Project (NWWRP) starting in tation and their consequences, and to routinely measure 1978, which installed the first unmanned C-band radar and rainfall with accuracies never before achieved over large integrated it with an operational flood forecasting and areas in real-time. warning service (Collinge and Kirby, 1987). The NWWRP According to WMO (1992a), the Republic of Korea has involved meteorological, hydrological, operational and undertaken a study on the development of the short-range research agencies and recognized the need to continue the precipitation forecasting system using satellite and radar initiatives on hydrological applications of weather radar. observations". Furthermore, a need has been identified for an experimental In the Russian Federation (WMO, 1992a) a manual is being . basin, Hydrological Radar Experiment (HYREX), on which a prepared for the diagnosis and prediction of dangerous and number ofbasic studies may be conducted. hazardous precipitation (hail and squalls) using data from The CASCADE system of the UK National Rivers Authority meteorological radar and satellites (WEFAX images); Thames Region (Haggett et at., 1992), which is one of the methods of applying data from the OKEAN side-looking most advanced flood warning systems in terms of its use of radar have been tested and improved. weather radar in Europe, is operated in support of flood The UK has a short-term radar rainfall forecasting network, warning over the London metropolitan area and surrounding FRONTIERS, which is used in the context of flood forecasting more rural areas. It utilizes high-resolution (2 km, 5 minute) (Walsh, 1993). The concept combines radar and Meteosat data from the C-band London weather radar at Chenies as display imagery with an experimental forecaster and develops well as wider coverage radar (5 km, IS minute) in the form of software tools to support the forecaster in using the radar the UK radar network composite image. The system incorpo­ imagery to make rainfall forecasts (Moore, 1993a). Further­ rates the Institute of Hydrology's local rainfall forecasting more, ongoing developments aim to speed up the forecasting software as well as receiving the FRONTIERS national radar procedures through greater automation, including considera­ rainfall forecasts. Furthermore, a range of derived products tion of expert system techniques and object programming useful to flood warning is generated, including displays of methods (Conway, 1989). Ryall and Conway (1993) reviewed radar-derived catchment average rainfalls. These can be input the progress on the new automated scheme which eventually to flood forecasting models to obtain forecasts at a variety of replaced FRONTIERS in 1995 and is called NIMROD. locations to support flood warning decisions (Moore, 1993a). Italy is developing a technique to merge satellite and radar data • Unlike the CASCADE system, which can make model fore­ in order to identify heavy convective clouds, and will develop casts for specific catchments, the River Flow Forecasting CHAPTER 2 - APPLICATIONS OF REMOTE SENSING IN HYDROLOGY 7

System (RFFS) (Moore and Jones, 1991; Moore, 1993b) has precipitation (PMP) estimates for catchment areas of interest; the been developed to coordinate the construction of forecasts PMP is a concept that is often used in assessments of reservoir down the river network. Thus. the RFFS is widely regarded safety. Such an approach capitalizes on radar's ability to delineate as the most advanced flow forecasting system in use in the storms in space and time. Furthermore, a programme called UK. An interface to the HYDrological RADar system RADMAX implements the procedure and incorporates visualiza­ (HYRAD) (Moore, 1993b), allows the RFFS to utilize data tion of the stann transposition step (Moore, 1993a). Collier (1993) for model input from one or more radars. in preprocessed, suggested the use of radar (and satellite data for cruder estimates) recalibrated and forecast form for grid square or catchment to support PMP estimation by the stann model approach. Work is areas (Moore, 1993a). ongoing in the UK to assess this stonn model approach utilizing • The Wessex Flood Forecasting in the UK prominently radar data to support model parameterization. features weather radar. This is based on the Wessex Radar According to Cluckie and Tyson (1989), the true poten­ Information Project (WRIP) system and the STORM system tial for the use of quantitative radar data in urban drainage systems for radar data acquisition and display, both developed at the has yet to be fully evaluated and consequently developed. Moore University of Salford. Furthermore, a river network form of (1993a) states that this statement is still true. However, the use of the system is under development (Moore, 1993a). weather radar in urban applications is distinct because of the fine Moore et al. (1993) studied the utility of the radar-derived spatial and temporal scale of interest. Furthennore, the potential rainfall forecasts in the rainfall-rainoff models used opera­ for real-time control of urban drainage systems, supported by tionally for flood forecasting over London. Based on using radar rainfall measurements and forecasts, has provided a particu­ five types of rainfall input to the Probability Distributed larly important stimulus for application. Verwom (1993) provides Moisture (PDM) model, the best overall performance was a recent state~of-the~art review on urban applications of weather obtained using the locally recalibrated radar forecasts radar. Cluckie and Tyson (1989) conducted a review of research obtained from HYRAD. activities concerned with the use of weather radar for urban Fattorelli, et at. (1995) report that the countries of the drainage system management. These included three European European Union depend upon the Flood Forecasting Warning studies in France, Germany and Britain (Moore, 1993a). The and Response Systems (FFWRS) for managing flood hazards French project is the Seine~Saint-Denis region of northeast Paris. within the European Union. Radar and satellite estimation of A C-band radar at Trappes is used for rainfall measurement and rainfall are integrated within the FFWRS and flood forecast­ forecasts are produced from them using the SCOUT method. A ing within a real-time framework is performed through similar project in Germany concerns the city of Bremen on the conceptual models. River Bremen where storage in the drainage system will be "Radar data have not yet been used to their full potential utilized to minimize uncontrolled overflows to rivers. In the upper in hydrological design studies" (Stewart, 1993). According to reaches of the River Emscher an X-band radar called Hydrological Moore (1993a), weather radar data are less immediately suited to Emscher Radar Project (HERP) is being deployed in combination applications for design rainfall estimation, compared to their use with distrometers to support real-time control of urban drainage for flood forecasting, on account of the shortness of radar records (Moore, 1993a; Kanuner, 1991). Rohlfing (1993) used data from relative to the storm periods to be inferred. However, Moore an X-band radar at a resolution of 600 m to provide rainfall esti­ points out this is in part compensated for by the complete spatial mates to conduct a sensitivity analysis of an optimal urban coverage that radars provide. Furthermore, the advantages of the drainage control strategy for part of the city of Essen to estimate high temporal and spatial resolution of radar data ought to be of spatial variability in rain. In Britain much attention has been given particular importance for short duration design estimates. to the use of weather radar to manage the drainage system for The accuracy of radar data for very severe storms (often St Hellier in the Channel Islands. A case study involving Greater associated with convective activity) can lead to considerable debate. Manchester utilized radar data from the Hameldon Hill C-band When compared with raingauge estimates, it can lead to confusion radar.

(Moore, 1993a). However, the accuracy problem in design applicaH The interest in X-band for urban applications, as exem­ tions ofradar data has been helped in the UK by the development of plified in the two German case studies above, is common in the PARAGON data-processing system (May, 1988). TIlls system several other European countries and in other parts of the world. created quality-controlled hourly rainfall values on a 5 km grid by This interest is a result of their higher resolution and greater combining radar with daily raingauge records. Aerial reduction resilience to clutter over a restricted scanning range and to their factors and design temporal profiles are derived. Furthermore, the lower cost relative to the conventional network C-band radars. timing accuracy associated with weather radar (data is commonly Furthermore, the use of X-band continues to be a contentious available at five-minute intervals) can be used to good effect in issue and the subject of a number of research investigations across obtaining better estimates of time-to-peak involved in design calcu­ Europe. These include the studies of the IMG group in France lations (Moore, 1993a). Cluckie et al. (1987) report a (Delrieu et aI., 1989, 1992) and the Wageningen group in the depth-area-duration (DAD) analysis of extreme events using hourly Netherlands (Moore, 1993a). According to Moore (1993a), a prac­ radar rainfall totals for 5 km grid squares using the Clee Hill UK tical issue in the use of weather radar in urban drainage models is network radar. The need to fIrst correct and calibrate radar data is that often the software, and possibly the model configuration, emphasized. DAD analysis is perfonned on individual stonns as a require modification to make best use of these distributed data. means of classifying the flood producing potential. However, Cluckie and Tyson (1989) report progress on the devel­ Cluckie and Pessoa (1990) have used radar data from opment of a front-end to the WASSP drainage model so that it can northwest England to characterize actual storms which have then accommodate distributed storm data from radar and apply recali­ been maximized and transposed to obtain probable maximum bration using local raingauges if required. 8 CURRENT OPERATIONAL APPLICATIONS OFREMOTE SENSING IN HYDROLOGY

Moore (1993a) reports that the real-time control of Mesoscale (250 km) Convective Systems (MCSs) that produced combined sewer overflows is of increasing concern as environmen­ EHR. While they utilized GOES water vapour (WV) imagery in tal pressures become more acute. Specifically, there is clearly an their studYl they have started using the DMSP-SSMII derived total important need to reduce surface flooding by contaminated waters precipitable water fields over the ocean areas to analyse the depth and to reduce the impact of llllcontrolled overflows on the quality of moisture under the WVP. WMO (1992a) reported that Bulgaria of receiving streams. Two case studies concerned with sewage is conducting a study aiming at investigating the thermal advection disposal control are as follows. Anderson and Anderson (1992) field on the basis of analogue Meteosat imagery in the WV presented- the study undertaken by the Swedish Meteorological and channel. This method is used to improve weather analysis and Hydrological Institute. The Kastrup Airport C~band weather radar forecasting, especially in the cases of cyclone developments over at Copenhagen was used for the efficient real-time operation of the the Mediterranean and North Africa, the regions characterized by Malmo Water and Sewage Works sewage systems. Ermolin (1992), the lack of a satisfactory number of conventional meteorological on the other hand, working in Moscow, Russian Federation, observations. presented an algorithm for the operational control of an urban head~and-gravityflow sewage disposal system, but merely alludes to the prospect for radar application. 2.2.2 Soil moisture and groundwater

2.2.1.5 Satellite radar rainfall monitoring The subsurface component of the hydrologic cycle comprises two zones: 1) the zone of aeration; and 2) the zone of saturation. The Engman and Gurney (1991) stated that radar measurement of rain­ zone of aeration comprises three subzones, namely soil water, fall is based on the Rayleigh scattering caused by the interaction intermediate and capillary. Water in the zone of aeration is called of rain and the radar signal. Assuming the rainfall uniformly fills soil moisture and it varies both spatially and temporally. The water the radar pulse volume, the power of the radar pulse returned to in the zone of saturation is called groundwater and tends to be less the antenna can be expressed as: variable than soil moisture. Information on soil moisture is important to agricultural (2) water management and hydrological runoff forecasts. Information where Pr is the average power returned from the precipitation at on groundwater is needed primarily for water supply purposes and distance r, the radar target range, C1 is an instrument parameter, agricultural water management. The importance of groundwater Cz is related to the dielectric constant of the precipitation (approx­ will increase in the future due to the demands for better quality imately a constant), K is an attenuation factor and D is the drinking water. raindrop diameter. The radar response to rainfall is wavelength Remote sensing techniques are currently used to estimate dependent according to D6/A, where A is the wavelength. Because soil moisture properties at or near the surface. This information may the radar interacts primarily with the rain, it can identify the pres­ be used to infer about soil moisture profiles down to several metres. ence of rain below the clouds and theoretically, at least, provide an Currently, there are no direct remote sensing techniques opportunity to measure rainfall rates without several restrictive to map areas of groundwater. However, indirect information can assumptions. Spaceborne radar has not yet been used for estimat­ be obtained from remote sensing sources. ing precipitation for the simple reason that there is no existing satellite radar with suitable frequencies for rainfall. 2.2.2.1 Soil moisture Further into the future, the scientific community will await the proposed TRMM (currently conceived as a joint US/ Remote sensing of soil moisture can be accomplished using Japanese low altitude, circular orbiting satellite). It will be equipped visible, infrared (near and thermal), microwave and gamma data with: 1) two microwave radiometers; 2) a cross-track scanning (Engman and Gurney, 1991). However, the most promising tech­ precipitation radar; and 3) an AVHRR to provide co-registered niques are based on the passive and active microwave data. The visible and infrared data (including a nocturnal visible sensor). visible and near-infrared techniques, which are based on the meas­ The TRMM satellite is planned for an operational dura­ urement of reflected solar radiation, are not particularly viable tion of at least three years beginning in the late 1990s. A strong because there are too many noise elements that confuse the validation effort is planned with several key ground sites to be interpretation of the data. The thermal infrared techniques are instrumented with calibrated multiparameter rain radars (Simpson based on the relationship between the diurnal temperature cycle et ai., 1988). Wilson et al. (1989) developed a dual frequency, and soil moisture which depends upon soil type and is largely scanning Airborne Rain Mapping Radar (ARMAR) system. This limited to bare soil conditions. A main problem associated with radar system will support the design and development of the thermal infrared techniques is cloud interference. Microwave tech­ TRMM rain mapping radar system. ARMAR is designed to fly on niques for measuring soil moisture include both passive and active the NASA Ames DC-8 and the high-altitude ER-2 aircraft. microwave approaches; each has distinct advantages. Microwave techniques are based on a large contrast between dielectric proper­ 2.2.1.6 Water vapour imagery rainfall monitoring ties of liquid water and dry soil. The variation of natural terrestrial techniques gamma radiation can be used to measure soil moisture because gamma radiation is strongly attenuated by water. It appears that According to TIllao et al. (1993), the water vapour plume (WVP) operational remote sensing ofsoil moisture will involve more than characteristics associated with extreme heavy rainfall (EHR), over one sensor. Furthermore, both active microwave and thermal the US during a three-year summer season, appeared to be an infrared applications need much additional research before they important component for intense convection and development of can be used to extract soil moisture information. CHAPTER 2 - APPLICATIONS OF REMOTE SENSING IN HYDROLOGY 9

The reflection from bare soil. in the visible and near­ humid and semi-arid watersheds in the US, scientists conducted an infrared parts ofthe electromagnetic spectrum, can only be used under intercomparison between passive and active microwave remote limited conditions to estimate soil moisture. The accuracy of this sensing of soil moisture (Wood et al., 1993; O'Neill et al., 1993). method is poor and absolute values of soil moisture cannot be They concluded that the choice is based on the intended applica­ obtained. More spectral bands and a much higher geometrical resolu­ tions and information that is available. tion in the visible/near-infrared (VISINIR) range are needed for soil The passive microwave region has been exploited the moisture and agricultural purposes than that available from Landsat, most so far. At present, microwave radiometers capable of measur­ SPOT and the NOAA satellites. Soil moisture has been estimated by ing soil moisture are available only on aircraft. These are being using precipitation indices; operational applications have been devel­ used in both research and a few operational applications. Available oped by FAO using geostationary imagery over intertropical regions satellite data (e.g. NIMBUS-SMMR or DMSP-SSM/I) have (WMO, 1993a). Furthermore, WMO (1993a) reports that with the frequent overpasses but have a rather low spatial resolution (50 advent ofthe International Geosphere-Biosphere Programme (IGBP) km) at the best channel available in space for soil moisture meas­ the need for high-resolution data is increasing. urement, Le. the C-band on NIMBUS-SMMR. The desired Thermal infrared techniques have been successfully resolution for large-scale studies is about 10 lan. Recent research used to measure the surface few centimetres of soil moisture (Jackson and Schmugge, 1989; Kostov, 1992; Ijjas, 1992; Shutko, (Engman and Gurney, 1991). A limitation to the thermal approach 1986; and Pampaloni et a!., 1990) using passive microwave for is that it cannot effectively be applied to surfaces with vegetation observing soil moisture suggests the optimal wavelength falls in cover. Idso et al. (1975) found that the volumetric moisture the L-band range (20-30 em). The soil depth contributing to L­ content for soil layers between 2-4 cm thick was linearly related band measurements is primarily the top 5 em. However, it may be to the amplitude of the diurnal soil temperature. Van de Griend et possible to determine the soil moisture through a depth of one at. (1985) demonstrated a relationship between night-time infrared metre if two microwave lengths are used in conjunction with a data and soil moisture and indicated that there was little sensitivity priori information on local soil properties (Reutov and Shutko, to soil moisture when a vegetation cover is present. Seguin et a!. 1986). Example research applications include preplanting soil (1993) used thermal IR NOAA AVHRR data in parallel with the moisture condition assessment, irrigation management, water Normalized Differential Vegetation Index (NDVI) to assess the management, and water balance studies (Jackson and Schmugge, extension and severity of water shortage on cultivated crops in 1989; Shutko, 1985, 1987). Kostov (1993) presented a brief review France and Algeria. Attempts have been made to evaluate the soil of the current understanding of passive microwave remote sensing moisture through observation of the apparent thermal inertia using of soil moisture and highlighted efforts towards operational use. both AVHRR data from Landsat and SPOT and geostationary Several experiments have begun to demonstrate the potential for images; applications have been more pilot projects than opera­ operational application (Kondratyev et al., 1977; Jackson et aI., tioual (WMO, 1993a). 1987a; Kustas et aI., 1991). The first operational programmes Microwave techniques have shown a lot of potential for utilizing aircraft microwave radiometers were initiated in the measuring soil moisture but still need varying amounts ofresearch to Russian Federation in the early 1980s for agricultural, hydrometeo­ make them operational (Blyth, 1993). In order to progress to opera­ rological and land reclamation applications. The feasibility of these tional soil moisture monitoring by remote sensing techniques, capabilities has been widely demonstrated in countries such as multifrequency and multipolarization satellite data will be required; Bulgaria, Poland, Hungary, Germany, Viet Nam and Cuba (Shutko, such data are needed to quantify different swfaces and thus reduce the 1985). In recent years, the Russian radiometers have been tested in amount ofground truth required. Jackson and Schmugge (1989) found the US in the large-scale hydrology experiment called Monsoon 90 that for large river basin and global studies, satellite passive microwave (Jackson et al., 1991). Scientists in the Russian Federation have soil moisture measurements are closer to operational applications than taken the advances in the passive microwave remote sensing of soil SAR because of their relative insensitivity to surface roughness. moisture and developed a system for irrigation scheduling (Reutov Furthermore, with the operational Earth Observing System (EOS), and Shutko, 1986). Using· algorithms developed in large area field more suitable spatial and temporal passive microwave satellite data experiments, microwave radiometer data can be converted to areal may become available. surface layer moisture maps (Jackson et al., 1993). Only in the microwave region is there a direct physical There are several aircraft systems available for obtaining relationship between soil moisture and the reflection or emission of radar data, primarily for research. Satellite radar data for soil radiation. A unique advantage of using the microwave region is that moisture measurement include SAR data from short-term satel­ at long wavelengths the soil moisture measurements can be made lites and Space Shuttle radar missions, and more recently SAR through clouds. A passive microwave radiometer measures the instruments on board the ERS-l, J-ERS-l and ALMAZ satellites. intensity of emission, which is proportional to the microwave Most research in the active microwave region has focused on the brightness temperature, TB' from the soil surface; while an active C-band range (5-IO em) (Dobson and Ulaby, 1986), where only microwave approach measures the radar backscatter, St' comprised the uppermost 5 crn of the soil surface can be observed (Kuittinen, of backscatter from vegetation and soil, and the attenuation caused 1992). These recent SAR instruments provide quasi-operational by the vegetation canopy. At short wavelengths, vegetation affects radar data at resolutions of 15-30 m. the microwave measurement through its biomass and structure. Soil moisture information at a depth of several metres However, biomass corrections may be made using indices based on can be obtained from short pulse radar (wavelengths of 5-10 em) VIS/NIR data or a crop calendar. Myneni and Choudhury (1993) techniques (Finkelstein et aI., 1987). In the Russian Federation, have illustrated the synergistic use of optical and microwave data in this aircraft-based method is used for soil moisture measurements agrometeorological applications. During the MACHYDRO'90 in forested areas and for detecting zones of saturation down to a experiment, which was a remote sensing aircraft campaign for depth of 5-10 m. 10 CURRENTOPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

The use of gamma radiation is potentially the most Rango (1992) analysed Landsat thematic mapper (TM) data over accurate of the remote sensing methods developed for soil mois­ the Nile Valley and Delta to assist in various phases of ground­ ture measurement. The attenuation of gamma radiation can be water development in Egypt, including the production of a used to detennine changes in soil moisture in the top 20-30 em of groundwater master plan for Egypt. They combined Landsat TM the ground. "This technique requires that some field measure­ maps with conventional maps to derive a hydrogeological map of ments of soil moisture be made during the measurement flight, the study area. Kaufman et al. (1986) used high-resolution because it does not give the absolute values of soil moisture" Landsat TM data to define the karats drainage patterns in the (Kuittinen, 1992). According to Schultz and Barrett (1989), a Peloponnesus area of Greece. The result was a structural map used major advantage of this technique is that in contrast to microwave to identify major flow paths. Gurney et ai. (1982) used Landsat measurements, gamma radiation is not affected by vegetation. data to assess groundwater resources in a poorly mapped area of which means that the technique is also applicable in forested areas the Kalahari Plateau in Botswana. Zall and Russell (1981) used (Jones and Carroll. 1983). Because air strongly attenuates gamma Landsat imagery to identify linear features, fracture, faults and radiation, measurements must be made from aircraft altitudes of drainage patterns for the Bobo-Dioulasso area in Burkina Faso. 50-150 m, which is a big disadvantage. The NWS maintains an El-Shazly el al. (1983) interpreted Landsat MSS imagery of the operational airborne gamma radiation snow survey programme in Wadi Araba area of Egypt to provide criteria for locating the upper Midwest in which flights are also made to calculate real­ exploratory wells. Ahmed and co-workers (1984) combined satel­ time areal soil moisture values (Jones and Carroll, 1983). The data lite derived lineament maps and drainage maps for the Sudan to are used to support the NWS hydrologic runoff forecasting suggest groundwater potential target areas for further research. programme (Carroll and Schaake, 1983). A study including Dutartre el al. (1993) illustrated the use of SPOT MSS and aerial gamma radiation soil moisture measurements was conducted by photography to identify a fractured basement area in the Sahelian Peck et al. (1990) as part of the First International Satellite Land climatic zone of West Africa. Their result was discovery of Surface Climatology Project (lSLSCP) Field Experiment (FIFE). aquifers providing significant amounts of groundwater. Dutartre et 01. (1993) used SPOT imagery to supplement existing geophysical 2.2.2.2 Groundwater data to extract information on the groundwater potential over a sedimentary intermontane basin in the Andes. El-Baz (1993) Remote sensing techniques used to map areas of groundwater analysed Landsat TM images of the Arabian Desert and discov­ include aerial and satellite imagery in the visible, infrared and ered a junction of two known ancient wadis (a stream course or microwave regions of the electromagnetic spectrum. In particular, valley in a desert). "This ancient riverbed in Saudi Arabia and satellite imagery enables a view of very large areas and achieves a Kuwait may indicate the presence of aquifers" (EI-Baz, 1993). perspective not possible from ground surveys or even low-level Infrared images are valuable for mapping soil type and aerial photography. Although remote sensing is only one element vegetation surface features used in groundwater exploration. of any hydrogeological study, it is a very cost-effective approach Springs can best be detected using infrared and thennal imagery. in prospecting and in preliminary surveys. Due to the intervening According to Guglielminetti et al. (1982), even underwater springs unsaturated zone of soil, most remote sensing data cannot be used can be detected by this method. Furthermore, through temperature directly, but require substantial interpretation. As a result, infer­ differences, thermal infrared imagery has the potential to deduce ence of location of aquifers is made from surface features. information on subsurface moisture and perched water tables at Important surface features include topography, morphology and shallow depths (Heilman aud Moore, 1981a aud 1981b; vegetation. According to Engman and Gurney (1991), ground­ Salomonson, 1983; van de Griend et 01., 1985). water information can be inferred from landforms, drainage Shallow groundwater tables can be measured using patterns, vegetation characteristics, landuse patterns, linear and passive microwave radiometry. A dual frequency radiometer has curvilinear features, and image tones and textures. Structural been used on an aircraft to measure water table depths of 2 m in features such as faults, fracture traces and other linear features can humid areas and 4 m in arid areas (Shutko, 1982, 1985, 1987). indicate the possible presence of groundwater. Furthermore, other Radar has an all-weather capability and can be used to features, such as sedimentary strata or certain rock outcrops, may detect subtle geomorphic features even over forested terrain (Parry indicate potential aquifers. Furthermore, shallow groundwater can and Piper, 1981). Radar is also capable of penetrating dry sand be inferred by soil moisture measurements, changes in vegetation sheets to disclose abandoned drainage channels (McCauley et ai., types and patterns. and changes in temperature (Engman and 1982, 1986), and can also provide information on soil moisture Gurney, 1991). Groundwater recharge and discharge areas within (Harris et ai., 1984). Radar images may be used to detect water drainage basins can be inferred from soils, vegetation and shallow which is some decimetres below the ground surface in arid areas, or perched groundwater (Engman and Gurney, 1991). due to the increase in soil moisture near the surface. According to Engman and Gurney (1991) state that airborne explo­ Finkelstein et ai. (1987), near-viewing short pulse radars installed ration for groundwater has recently been conducted using on mobile ground or aircraft platforms provide information on the electromagnetic prospecting sensors developed for the mineral depth to a shallow water table down to 5-50 10. Engman and industry. This type of equipment has been used to map aquifers at Gurney (1991) illustrated the potential of radar imagery to pene­ depths greater than 200 m (paterson and Bosschart, 1987). trate the dense tropical rainforest and rainfall over Indonesia and Aerial photography supplemented by satellite data from produced a geological map for use in groundwater exploration. Landsat or SPOT are widely used for groundwater inventories, McCauley et al. (1982, 1986) illustrated the use of SIR-A data primarily for locating potential sources of groundwater. This tech­ taken over the eastern Sahara; the radar imagery revealed a vast. nique permits inferences to be made about rock types, structure previously unchartered network of valleys and smaller channels and stratigraphy (Engman and Gurney, 1991). Zevenbergen and buried by the desert sands. Furthermore, this drainage network had CHAPTER 2 - APPLICATIONS OF REMOTE SENSING IN HYDROLOGY II been obscured from the Landsat visible imagery. However, Ea is the net energy advected into the surface, Ee is the energy Dutartre et at. (1993) successfully merged SIR-A radar imagery utilized for evaporation/transpiration; H is the sensible heat flux with SPOT visible imagery over the arid Kutum area northwest of (loss from the surface to the atmosphere); and LE is latent heat HI Fasher, Sudan for use in hydrogeological prospecting, and flux (change in energy storage within the surface). For this recommended this approach for similar environments in the discussion ET and LE are synonymous. The components of "Sahelian Belt". Equation 4 represent the surface energy balance in terms of the soil surface temperature and a set of meteorological measurements which at least in principle are measured or estimated routinely. 2.2.3 Evapotranspiration Furthermore, the energy and moisture fluxes are intimately related because the LE gives the amount of water evaporated and tran­ Evapotranspiration (ET) is the combination of the evaporation and spired. Although this approach is used operationally, only a few transpiration hydrologic processes. Evaporation is the loss of meteorological stations (research sites) provide the necessary data water from open water and moist soil surfaces to the atmosphere. in primary fonn (Bailey, 1990). Transpiration is the loss of water from the airspaces in living plant Remote sensing observations combined with ancillary leaves to the atmosphere. Relative to the water balance (both local meteorological data have been used in obtaining indirect estimates and regional), ET is a significant and often the dominant water of ET over a range of temporal and spatial scales (Kustas et ai., flux at the Earth's land surface (Engman and Gurney, 1991). 1989; Choudhury, 1989). Recently there has been much progress Evapotranspiration is sensitive to the following causative factors: in the remote sensing of parameters including: 1) incoming solar I) solar radiation; 2) air temperature; 3) wind velocity; 4) vapour radiation; 2) surface albedo; 3) vegetative cover; 4) surface pressure gradients; 5) soil moisture; 6) soil temperature; and 7) temperature; and 5) surface soil moisture. However, there has been surface characteristics (Kuittinen, 1992). McCuen (1989) stated little progress in the direct remote sensing of the atmospheric that based on a number of studies, the first four of these factors are parameters which affect ET such as: 1) near-slUface air tempera­ believed to be the most important. Furthermore, he emphasized ture; 2) near-surface water vapour gradients; and 3) near-surface that the vapour pressure is a function of air temperature, and that winds (Engman and Gurney, 1991). Furthermore, remote sensing the solar radiation is highly correlated with air temperature. has a potentially important role because of its areal coverage in the Estimates of the ET component are required for various spatial extrapolation process of ET. purposes, including irrigation scheduling, reservoir losses, water Remote sensing of several important parameters used to and energy balance calculations, lUnoff predictions, inputs to soil estimate ET is made by measuring the electromagnetic radiation in a erosion models, and meteorological and climatological studies particular waveband reflected or emitted from the Earth's surface. (Engman and Gurney, 1991; Bailey, 1990). The incoming solar radiation can be estimated ~rom satellite obser­ The assessment ofET may be conducted by either direct or vations of cloud cover primarily from geosynchronous orbits using indirect measurements. Directmeasurements which include evapora­ MSS in the visible, NIR and thenna! IR parts of the EMS (Brakke tion pans and Iysimeters are subject to a wide range of errors. Thus, and Kanemasu, 1981; Tarpley, 1979; Gautier et al., 1980). The in practice empirically derived correction factors are used for pan surface albedo may be estimated for clear-sky conditions from meas­ measurements in an attempt to derive a "true" evaporation figure urements covering the entire visible and near-infrared waveband, (Bailey, 1990; Ward, 1975). As a result, meteorologists have attempted while measurements at narrow spectral bands can be used to deter­ to use indirect methods to assess ET. The two broad categories ofindi­ mine vegetative cover (Jackson, 1985; Brest and Goward, 1987). The rect methods are the diffusion approach (BlUtsaert, 1982) and the surface temperature may be estimated from MSS measurements at energy balance approach. The diffusion approach is generally thermal infrared wavelengths of the emitted radiant flux (Engman regarded as being most appropriate on the local scale, while the energy and Gurney, 1991). The soil moisture may be estimated using the balance approach is considered to be appropriate to larger areas than measurement of microwave properties of the soil (microwave emis­ the diffusion methods. sion and- reflection or backscatter from soil). However, there are The diffusion approach is based on Dalton's diffusivity uncertainties in such soil moisture estimates due to previously law, and the following simplified aerodynamic equation is used to mentioned factors such as surface roughness and vegetative cover represent the latent heat flux density (LE), i.e. the ET: (Engman and Gurney, 1991). The most practical remote sensing approach for the (3) future will include repetitive observations at the visible, near­ wherej(v) is an airspeed function, eo and ea are the vapour pressures infrared and thermal infrared, and microwave lengths (Engman of the slUface and air, respectively. Because the measurement of the and Gurney, 1991). Components for detennining the sensible heat variables makes considerable technical demands, the operational use flux will be measured by the EOS instruments. The latent heat of this fonn is difficult (Bailey, 1990). flux cannot be measured directly but EOS instruments will The energy balance approach applied at the surface does provide some sampling capability. Furthermore, the future not rely on knowledge of the details of the evaporation process programmes such as EOS should provide the necessary data for and is given as follows: evaluating ET on local, regional and global scales (Engman and Gurney, 1991). (4) Assuming that the surface energy balance equation is where Rn is the net radiation flux (Rn == Rt - Rr - Rt ), Rt is the the primary boundary condition to be satisfied in computing ET, total solar radiation incident to the surface, Rr is the reflected solar there are three techniques to solve Equation 4. The first technique radiation, RI is the net longwave radiation exchange between the uses semi-empirical methods, the second employs analytical atmosphere and the surface; G is the soil heat flux (G == Ea - Ee), methods and the third utilizes numerical models. 12 CURRENT OPERATIONAL APPUCATIONS OF REMOTE SENSING IN HYDROLOGY

2.2.3.1 Semi.empirical methods Reginato et al. (1985) combined remotely sensed reflected solar radiation and surface temperatures with ground­ The semi-empirical-methods represent an effort- to obtain a manage­ based meteorological data (incoming solar radiation, air able model to estimate ET. 1besemodem operational approaches are temperature, wind speed and vapour pressure) to estimate the net derived mainly from Penman's original formulation (penman, 1948), radiation flux (Rn) and the sensible heat flux (ll) of Equation 4. The which is a combination ofthe diffusion and energy balance approaches soil heat flux (G) was estimated to be a fraction of the net radiation (Bailey, 1990). Penman (1948,1956) used a simplification of the and the height of the plant canopy through an empirical equation. energy balance equation (Equation 4) for estimating ET. He further Using the energy budget approach of Equation 4, instantaneous reducedits complexity by making simplifying assumptions, fuld then values of ET were calculated. Jackson et at. (1987b) extended this calibrated it to account, in part, for the loss ofaccuracy due its simpli­ method by collecting the reflected solar and emitted thermal energy fications. Furthermore, empirical evaluations are used to fit the with low-level aircraft flights. Clothier et al. (1986) developed a coefficients for a particular location and set of conditions (McCuen, relationship for estimating G from a crop height measure and a 1989). One particular condition is that ET estimates are for periods of spectral vegetation index. Furthermore, Kustas and Daughtry (1989) about one day. The modified form ofthe Penman formulation, referred proposed a method for computing G based on multi-spectral data to as the Penman-Monteith equation or the combination equation, is that should be useful for regional energy budget studies. widely used in the calculation ofETfor operational purposes (Bailey, Caselles and Delegido (1987) developed a procedure for 1990; Monteith, 1965). Modification of Penman's original formula­ estimating ET for different crops. They used satellite-measured tion resulted from doubts that underlie the basic assumptions of the global radiation and maximum air temperature. Values for the conventional meteorological methods. For instance, research by maximwn ET for a given crop are a function of a crop coefficient. Seginer (1971) and Thorn and Oliver (1977) illustrated that: I) the Bausch and Neale (1987) developed a procedure for estimating the effect of radiation has been over-emphasized and that of wind real~time crop coefficient from reflectance properties of a crop/soil underestimated; and 2) the regional ET is not independent ofvegeta­ scene. They also used a linear transformation of the normalized tionheight (and thus surface roughness) in areas ofsubstantial rainfall. difference vegetation index (Tucker et at., 1979) to produce a The following are current applications of remote seasonal curve. Caselles and Delegido (1987) applied their results sensing assuming the semi-empirical methods to estimate ET. to the Valentian region ofSpain. Jackson et aI. (1977) proposed- the Jackson model and Kotoda et al. (I983a, 1983b) used airbome MSS data to later it was evaluated using empirical and theoretical results by map surface temperatures for a complex land cover in Japan. Using Seguin and Itier (1983). The energy balance model of Equation 4 the Priestly-Taylor (1972) relationship (an empirical solution of is integrated over a 24-hour period and thus assumes that G is Equation 4), they employed the remotely sensed temperature data to

negligible. Furthermore, observations by !tier and Riou (1982) and estimate the vapour pressure/temperature ratio and the Rn term. more recently by BruneI (1989) suggest the daily ratio of HlRn Dunkel and Vadasz (1993) used a simple statistical can be approximated by HlRn estimated near midday under clear method to calculate the daily ET in Hungary during the growing sky conditions. With some further approximations Equation 4 can season. The vegetation index and the surface temperature, derived be recast as: from the NOAA AVHRR images, were used to characterize the regional distribution of the surface state. The ET is proportional to (5) the difference between surface temperature and air temperature where Ts is the surface temperature estimated remotely, say from a measured at 2 m height. However, for operational use and for satellite-based thermal infrared sensor, Ta is the near-surface air more precise monitoring of daily evaporation several improve­ temperature obtained from a nearby weather station, the subscript i ments remain to be done. represents the "instantaneous" observation by the satellite over the Menenti (1993) used remote multispectral satellite area of interest, and A and B constants which Caselles and measurements for inputs to semi-empirical models to study and Delegido (1987) report as varying with location. In practice, map land surface ET in Egypt and Argentina. however, A and B vary with a wide range of both meteorological and surface factors (Bailey, 1990). This expression and derivatives 2.2.3.2 Allalytical methods of it have been tested and shown to produce reasonable estimates of daily ET (Brunei, 1989; Kerr et al., 1987; Nieuwenhuis et al., The analytical methods seek to analytically evaluate the individual 1985; Rambal el al., 1985; Thunuissen and Nieuwenhuis, 1990; elements of the energy balance equation. More physically-based Riou et al., 1988). Although Equation 5 is characterized by low approaches are used over a 24-hour period to produce an analyti­ demands for data provision and ease of operation, it is also charac­ cal solution (Pratt el al., 1980; Price, 1980). Both energy and terized by limited spatial and temporal areas of application together moisture fluxes are modelled in estimating ET. with poor accuracy especially in the presence of cloud when using The following are current applications of remote sensing satellite thermal infrared methods to obtain T, (Bailey, 1990). assuming the analytical methods to estimate ET. Jackson et al. (1983) proposed another form of the In an effort to overcome problems presented by incom­ Jackson model where ET was assumed to follow the temporal plete canopies in estimating ET, Choudhury and Monteith (1988) trend in solar radiation throughout the daylight period. Remotely and van de Griend and van Boxe1 (1989) have used analytical sensed (low-level aircraft) reflected solar and emitted thermal methods to evaluate the components of Equation 4 in multi-layer radiation data were used with ground-based measurements of models; soil and vegetation are treated separately using remotely wind,. vapour pressure and incident solar radiation to estimate ET. sensed data. Although their models yielded good results with The estimated ET rates compared favourably with rates measured measured values, they may be difficult to use operationally with Bowenratio instrumentation (Jackson et aI., 1987b). (Engman and Gumey, 1991). CHAPTER 2 - APPLICATIONS OF REMOTE SENSING IN HYDROLOGY 13

Gash (1987) investigated and fonnalized the analytical 2.2.4 Snow framework necessary to evaluate the spatial extrapolation of ET values obtained at a reference point using remotely sensed Snow is the occurrence of precipitation in the solid form. The water horizontal variations in surface temperature and reflected radia­ stored in snow is held in cold storage on a basin for an extended tion. In an effort to seek operational applications. Gash makes period of time.before it enters into the runoff process. During this several simplifying assumptions regarding some negligible terms. varying time period, the snow in storage can be augmented by addi­ Bernard et al. (1981) and Soares et al. (1988) used radar tional snow accumulation, redistributed by wind, changed in its estimates of soil moisture as the upper boundary condition in a internal structure and affected by meteorological factors (Rango, water balance model that simulated flow using the Richards equa­ 1993). The snowpack is removed from the basin by ablation tion in order to calculate ET rates. processes usually lasting several weeks to several months depending Price (1980) assumed the 24-hour value of G to be on the amount of snow and location. The established snowpack negligible. He used the Fourier expansion of the angle of solar undergoes various transformations or metamorphic changes with incidence and linearized the Stefan-Boltzmann equations for time. This includes a general decrease of surface albedo, an increase incoming and outgoing longwave and the bulk resistance equation in snow grain size and an eventual increase in snowpack liquid for H. Price (1982) used this method with NOAA AVHRR data to water until snowmelt occurs (Rango, 1993). Other changes associ­ estimate a 24-hour average ET, and obtained favourable evalua­ ated with these metamorphic changes are a progressive warming of tions when compared with standard meteorological and pan the snowpack, an increase in snow density, a settling or decrease in evaporation data. snow depth (if unaffected by new snow) and an increase in internal snowpack strength at least up to the appearance of liquid water in 2.2.3.3 Numerical models the snowpack. The snowmelt process takes place at various rates depending on the aspect, slope, elevation, vegetation cover and The numerical models simulate the flow ofheat and water in the soil atmospheric conditions (Rango, 1993). and use the surface energy balance as the boundary condition. The Predicting snowmelt runoff is required for various advantage of this approach is that only initial conditions are needed to purposes including water supply, flood control, hydropower gener­ detennine model parameters. Then, the temporal trace ofthe evapora­ ation and water allocation for irrigation (Rango, 1994b). The tive flux can be simulated and periodically updated (Rango, 1994b). hydrologist generally wants to know how much water is stored in The following are current applications of remote sensing a basin in the form of snow (Engman and Gurney, 1991). assuming the numerical models to estimate ET. The melting of snow is controlled by the net energy flux In an effort to provide ET estimates on a routine and exchange between the snowpack and the environment. The rate of timely basis, Soer (1980), Rosema et al. (1978), Carlson et al. snowmelt is largely controlled by the rate at w_hich the net energy (1981), Camillo et al. (1983), Raffy and Becker (1985) and flux is received by the snowpack. To place the various components Taconet et al. (1986) numerically modelled Equation 4 and its of the energy balance in perspective, the formulation by Dozier components. The moisture balance is inferred from the heat flow (1987) is used. in the soil. Furthermore, the analytical expressions for the Q=R+H+LE+G+M (6) components of Equation 4 can be manipulated to obtain an

expression for the surface energy balance in tenus of the surface where Q =: the net energy available for decreasing or increasing temperature and a set of routinely measured meteorological the snowpack temperature or, if the temperature is already DOC, measurements. Initial estimates of the soil/crop surface tempera­ melting the snowpack; R =: net radiation of incoming solar and

ture, Ts' are optimized using an iterative numerical technique to emitted longwave radiation; H =: convective or sensible heat flux; ensure that the boundary condition of Equation 4 is satisfied. LE =: latent heat flux; G =: heat flux to or from the soil; and M =: Taconet et al. (1986) illustrated the feasibility of this approach heat flux by advection (Rango, 1993). Most snowpack energy with NOAA-7, AVHRR infrared data over an area of the Beauce balance studies have shown that the net radiation is the dominant (France). More recently, Taconet and Vidal-Madjar (1988) term (Rango, 1993). Temperature, operating through the sensible included geostationary satellite data (Meteosat) to increase the flux, is also an important driving force. The latent heat flux can stability of model inversion and atmospheric correction to the also be an important melt factor at certain times. Combined, radia­ satellite observations. Elkington and Hogg (1981) suggested that tion and temperature are the causative factors for the vast majority a simplification in the initial and the boundary conditions would of melt. In general, the radiation effect prevails if temperatures are make this numerical modelling approach to estimate ET poten­ only slightly above DOC, whereas sensible heat gains importance tially operational. Furthermore, Engman and Gurney (1991) point when temperatures are high (Rango, 1993). As net radiation is out that the estimate of actual ET from remotely sensed measure­ difficult to measure, and temperature is relatively easy to measure, ments becomes feasible when it is possible to take initial temperature has been used as an index in many models. The most conditions from remote sensing measurements and boundary commonly used index is the degree-day value. The degree-day conditions from either routine meteorological observations or method (as outlined by Linsley, 1943) has been used extensively remotely sensed measurements. since and is the snowmelt algorithm employed in most detenninis­ According to WMO (1992a), Germany is utilizing tic models for producing snowmelt (Rango, 1993; Kustas et al., NOAA AVHRR data for input into numerical evaporation models 1994). The degree-day method is a temperature index approach in small-scale agricultural areas. Satellite data include vegetation, that equates the daily decrease of the water equivalent of the land-surface temperature gradients, soil moisture, diurnal tempera­ snowpack to a temperature difference between the daily average ture variations and solar irradiance. Extrapolation of the model air temperature and a base temperature (usually DOC) (Rango, results are to be tested (WMO, I992a). 1994b). The snowmelt rate is computed as follows: 14 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

M = a (T, - Tb) (7) tenns of water equivalent (Rango, 1994b). For many basins, there is where a is a degree-day factor (cmrC/day), Ta is average daily a very goodrelationship between runoff and snow cover area (Engman temperature cae), Tb is the base temperature COC), and M is the and Gurney, 1991). "For operational runoff forecasts, however, the snowmelt rate (em/day) (Kustas et aI., 1994). Because the above water equivalent must also be detennined" (Rango, 1994b). Use ofthe equation for snowmelt has to account for all terms of the energy so-called modified depletion curves (the S measured is plotted against budget that affeGt the mass balance of the snowpack, ex has a high the accumulated melt depth or HW) enables a snowmelt runoff spatial and temporal variability, and is a function of snow density, approach to become a forecasting model as opposed to simply a simu­ average daily air temperature range and mean wind speed (Kustas lation model (which uses conventional depletion curves). Furthermore, et al~, 1994), Another snowmelt algorithm is the radiation-based the indicators ofsnow which include S, HW, as well as the snow approach (also known as the surface energy balance model), and is condition, are generally difficult to measure and are likely to vary used in some deterministic snowmelt runoff models (Leavesley considerably from point to point, especially in mountainous terrain and:Stannard, 1990). The snowmelt depth is computed as follows: (Engman and Gurney, 1991). "Remote sensing offers a new valuable tool for obtaining snow data for predicting snowmelt runoff' (Engman M=m I1Q (8) Q and Gurney, 1991). Remote sensing of snow can be accomplished using where I1Q = Rn +H+LE (9) gamma rays, visible and near-infrared, thermal infrared and where I1Q is the energy available for melt, Rn is the net radiation, microwaves (Rango, 1993). Tables 1 and 2 give an overview of the His the sensible heat flux and LE is the latent heat flux, all in relative sensor band responses to various snowpack properties. It watts per square metre; and mQ converts .6.Q to M, the snowmelt. is evident from Tables 1 and 2 that the microwave band has the According to Kustas et at. (1994), a statistical analysis by Zuzel greatest overall potential followed by the visible and near-infrared and Cox (1975) indicated that daily snowmelt estimates improved band. The gamma ray portion is extremely limited by the fact that by incorporating net radiation, vapour pressure and wind in model the sensing must be carried out with low altitude aircraft, and to a formulations rather than air temperature alone. lesser extent that it is really only sensitive to a finite snow water Snowmelt runoff procedures have followed two distinct equivalent. Thennal infrared is also limited in potential, but it can paths: an empirical approach and a deterministic modelling be used from space in night-time situations (Rango, 1993). approach. The choice of approach depends upon both the avail­ According to Engman and Gurney (1991), "different approaches ability of data to quantify the snowpack and the extent of the detail for determining snow area, water equivalent and snow properties required of the output (Engman and Gurney, 1991). In order to have been developed. These have been driven for the most part by make accurate estimates of snowmelt runoff, hydrologists need to the availability of data from existing satellites or from experimen­ quantify the snowmelt as follows: 1) the areal extent of the snow tal aircraft and truck programmes." According to Rango (1993), (snow cover area, S); 2) the snow water equivalent (HW); and 3) remote sensing data are currently being used operationally in snow the condition/properties of the snow (such as depth, density, grain cover and snow water equivalent assessments, and seasonal size and presence of liquid water) (Engman and Gurney, 1991). A snowmelt runoff forecasts. "The potential of satellites to provide list of snowpack properties and properties affecting albedo and useable infonnation on snowpack dynamics is now widely recog­ emissivity of snow are given in Tables I and 2, respectively. nized and today many schemes exist which employ satellite­ "The meltwater volume is a product of the melt depth and derived snow measurements for snowmelt-runoff prediction" the snow covered area" (Rango, 1992). The gradually decreasing areal (Lncas and Harrison, 1990). extent is a characteristic feature ofthe seasonal snow cover. It should be noted that for day-to-day simulations of snowmelt runoff, whether 2.2.4.1 Snow cover distribution an empirical approach (using historical data) or a deterministic approach is used, it is sufficient to know the snow cover area in the According to Rango (1994b), only satellites enable seasonal snow basin each day without knowing the initial accumulation of snow in cover to be monitored periodically, efficiently and on a sufficiently

Table 1-Relative sensor band responses to various snowpack properties (Rango, 1993)

5emorband

5nowproperty Gamma rays Vi5iblel Thermal Microwaves near~infrared infrared

Snow cover area Low High Medium High Depth Medium If very shallow Low Medium Water eqUivalent High If very shallow Low High Stratigraphy No No No High Albedo No High No No Uquid·water content No Low Low High Temperature No No Medium Low Snowmelt No Low Low Medium Snow-soil interface Low No No High Additional factors - All weather capability No No No Yes - Current best spatial resolution Not possible 10 m 100 m 25 km passive - from space platform 10 km active CHAPTER 2 -APPLICATIONS OF REMOTE SENSING IN HYDROLOGY 15

Table 2 - Properties affecting albedo and emissivity of snow (Warren, 1982)

Snow property Visible Near-infrared Thermal infrared Microwave reflectance reflectance emissivity emissivity

Grain size • Yes No Yes Zenith (or nadir) angle No Yes Yes Yes Depth Yes No No Yes Contaminants Yes No No No liquid water contact No • No Yes Density No No No Yes Temperature No No No Yes

* Only for thin snowpack or if impurities present large scale. Significant remote sensing data for operational snow AVHRR; although characterized with a much higher frequency of mapping are available from satellites such as SPOT, Landsat, coverage (every 12 hours as opposed to every 16-18 days), the NOAA, GOES, Sky1ab and DMSP (Lncas and Harrison, 1990). problem with the NOAA AVHRR data is that the resolution of 1 km According to Martinec (1987) and Rango et at. (1983), the choice (in the visible red band (0.58-0.68 rom)) may be insufficient for of satellite for snow mapping depends upon the smallest partial snow mapping on small basins. area of the region to be monitored. While the accuracy of the Despite the spatial and temporal resolution problems snowpack delineation and snow area estimation depends upon the associated with visible aircraft and satellite imagery, these have spatial resolution of the sensors involved, operational snow proven to be very useful for monitoring both the buildup of snow mapping schemes are rarely necessary for such small areas (Lucas cover and the disappearance of snow cover area in the spring and Harrison, 1990). As a result, the Landsat TM sensor is usually (Rango, 1993). The NWS National Operational Hydrologic applied within the context of research projects (Lucas and Remote Sensing Center (NOHRSC) in Minneapolis uses NOAA Harrison, 1990); and in some cases, aerial photographs may be AVHRR data to operationally provide periodic snow cover extent preferred to TM imagery as these can be collected for selected maps on about 3 000 basins in North America (Carroll, 1990). The cloud-free days and for similar sized areas (Rango et al., 1983). maps and the map statistics are electronically. sent in near·real­ The data from the Landsat and NOAA satellites are time to a wide variety of users, mostly for hydrological forecasting characterized with a lower temporal resolution than, for instance, purposes (Rango, 1993; Carroll, 1990). Another operational the GOES satellites. According to Rango et al. (1983), a high product by the NWS is the monthly mean .snow cover charts temporal resolution is important, particularly for monitoring which are used for larger scale studies of regional and global changes in snow extent, due to melt or accumulation, and in the hydrology and climate (Rango, 1993; Dewey and Heim, 1981). provision of regular snow cover estimates between station­ METEOR (which has been used to delineate snow/no snow lines recorded observations, such as for runoff forecasting. Rango for river basins and other areas in the Russian Federation) and (1985) has recommended that the optimum frequency of observa­ NOAA data were combined to map snow cover area in basins tion of snow cover during depletion would be once a week. ranging from 530 to more than 12 000 km2 (Shcheglova and Furthelmore, "multitemporal observations of a scene may allow Chemov, 1982). the identification of cloud and snow surfaces as clouds may alter Ostrem et at. (1981) developed a method using NOAA in form and position over time whereas temporal variations in and TIROS data for measuring the remaining snow and to predict snow cover are generally less extreme" (Lucas and Harrison, the corresponding snowmelt runoff volume for a number of 1990; Meier, 1975). Norwegian high mountain basins. This work in Norway was "The areal extent of the snow cover is mapped opera­ viewed with great interest because the country produces 99 per tionally in many countries using weather satellite data" (Kuittinen, cent of its electricity from hydropower with greater than 50 per 1992). Although snow cover can be detected and monitored with a cent of the precipitation in the catchment areas occurring as snow. variety of remote sensing devices, the greatest application has been The largest hydropower company in Norway, Statkraft, used snow found in the visible and the near-infrared region of the EMS cover extent maps from NOAA AVHRR on an operational basis as (Rango, 1993). The reason is that "the reflectance of snow in the input to its hydropower production planning (Andersen, 1991). visible and near-infrared parts of the EMS is much greater than that Ramamoorthi (1983, 1987) has used snow cover data of any other natural material on the ground and thus snow can easily from NOAA AVHRR in a regression model for operational fore­ be detected and the extent of snow cover detennined" (Kuittinen, casts of seasonal snowmelt runoff on the Sutlej River Basin 1992). The reflectivity (albedo) depends upon snow properties such (43230 km2) in India; these forecasts have been extended to addi­ as the grain size and shape, water content, surface roughness, depth tional basins. Kumar et al. (1991) have used satellite snow cover and presence of impurities (Engman and Gurney, 1991). In particu· data as input to snowmelt runoff models (SRM) for use in opera­ 1ar, the visible red band (0.6-0.7 rom) of the MSS on the Landsat tional forecasts of daily snowmelt runoff on the Beas and Parbati has been used extensively for snow cover mapping because of its Rivers in India. strong contrast with snow-free areas (Rango, 1993). Rango (1993) Although snow can be detected in the near-infrared points out that although Landsat and SPOT may provide adequate band, the contrast between a snow and a no-snow area is consider­ spatial resolution for snow mapping, their inadequate frequency of ably lower than with the visible region of the EMS. However, the coverage hinders their snow mapping capabilities. As a result, many contrast between clouds and snow is greater in the Landsat TM users have turned to the NOAA polar-orbiting satellites with the Band 5 (1.57-1.78 mm). Thus the near-infrared band, when 16 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

available, serves as a useful discriminator between clouds and maps are currently produced using geophysical algorithms on the snow (Engman and Gurney, 1991). Harrison and Lncas (1989) and data from satellite microwave radiometers such as DMSP SSM/I. Lncas and Harrison (1989a, 1989b) nsed VIS/NIR difference data These maps are most reliable over large flat regions with little or from NOAA-9 imagery of the UK to locate areas of complete or low-lying vegetation when snow is dry." According to Rango partial snow cover and identify melt and accumulation zones. (1993), "the resolution problem can potentially be solved with the Daily snow area maps were produced and were subsequently use of high-resolution active microwave sensors. Unfortunately, composited to generate weekly estimates of snow distribution. few, if any; experiments with the short wavelength region of the "TIlls technique is currently being considered for operational use microwave spectrum (around 1 cm wavelength) that is sensitive to in the UK and elsewhere" (Lucas and Harrison, 1990). The recent snow have been reported." Davis and Shi (1993) found that using inclusion ofa 1.6 mm channel (near-infrared) on the NOAA radar, mapping snow in alpine regions by using single~polarization AVHRR sensor should allow more reliable derivations of snow SAR requires topographic information. Matzler and Schanda area; this may permit direct comparison with visible and/or near­ (1984) used X-band airborne radar data to map wet snow cover in infrared data, and subsequent calibration of data for the evaluation the Swiss Alps. RoU (1986) conducted some studies with 3 cm of snow condition (Schultz and Barrett, 1989). radar over snow with applications limited to a wet snowpack at Thermal infrared data have limited importance for snow this wavelength. According to Kuittinen (1992), ''no space-borne mapping and measuring properties because they are hindered by radar exist at present and the usefulness of the planned SARs for cloud cover, and the surface temperature of snow is not always that operational snow mapping will be limited by the lack of suitable much different from the surface temperatures of other surfaces wavelengths, narrow swath, high processing costs and very (Rango, 1993). However, thermal infrared data can be useful for restricted imaging time per cycle." helping identify snow/no-snow boundaries, and for discriminating between clouds and snow with AVHRR data because the near­ 2.2.4.2 Snow accumulation infrared band has not been available on this sensor. Furthermore, Kuittinen (1992) stated that the best result in snowline mapping can "Even more important than snow extent and location for various be achieved by combining the information of the thermal emission snowpack processes is the vertical dimension of the snowpack. and the reflectance in the visible part of the EMS. This vertical dimension essentially provides the information Although the airborne gamma ray spectrometry needed for estimating snow volume which relates directly to the approach is a very accurate remote sensing method for measuring potential for snowmelt runoff' (Rango, 1993). There is a lot of HW, its previously mentioned drawbacks limit its use. However, potential in the capabilities of remote sensing of snow depth or airborne gamma ray data and weather satellite data together water equivalent. provide good possibilities for operational snow cover mapping According to McGinnis et ai. (1975), visible satellite (Carroll, 1990; Kuittinen, 1989). imagery can be used to estimate snow depth in areas of relatively ''Although there are currently many problems with using shallow snow accumulation of up to 25 cm; they found a direct microwave sensing for mapping snow cover, one major advantage correlation between increasing brightness and increasing snow of the microwave approach is the ability to penetrate cloud cover depth in areas without tall vegetation cover in the southeastern and map snow extent" (Rango, 1993). Due to its cloud penetration USA. Furthermore, this approach is only possible if the area is or all-weather capability, the microwave wavelength at about relatively flat and has a fairly short vegetation cover (Rango, 1.0 em has the greatest overall potential for snow mapping. 1994b). According to Rango (I994b), the combination offreqnent However, the current major drawback is the poor passive visible satellite imagery and the accurate aerial gamma ray spec­ microwave resolutions from space (about 25 km) so that only very trometry can improve the knowledge of snow cover distribution. large areas of snow cover can be detected (Rango, 1993). As reso­ provide good areal averages for HW and offer real-time snow data lutions of passive microwave sensors improve (a Russian oyer large areas. In addition, it is possible to include the use of a microwave radiometer at about 0.8 cm wavelength with approxi­ Geographical Information System (GIS) to estimate the HW for mately 8 km resolution was launched in 1996), all-weather each pixel. Martinec and Rango (1981) used visible satellite snow capability will increasingly be exploited. The final advantage of cover data, a grid system of the basin and air temperatures for the microwave spectrum is that night~time measurements are each grid unit to estimate HW at the beginning of the melt season easily made because of the reliance on emitted microwave radia­ (e.g. I April). tion as opposed to reflected visible radiation (Rango, 1993). "Airborne gamma ray spectrometry can be used to deter­ According to Kuittinen (1992), "emittance and backscattering of mine the snow-water equivalent values, because snow attenuates microwave radiation are affected by almost all snow parameters, the terrestrial gamma radiation" (Kuittinen, 1992). "Background which complicates the measurement of the most needed para­ gamma radiation of the soil is obtained before snowfalls, and meters: water equivalent, areal extent, amount of free water." subsequent flights are flown to measure the gamma radiation Hallikainen (1984) and Rott and Knnzi (1983) have through the attenuating snow cover. The degree of attenuation is studied the data of NIMBUS-7 SMMR radiometer, bnt its poor related to the snow water equivalent through various calibration spatial resolution and lack of real-time data service do not allow graphs" (Rango, 1993). According to Carroll and Vadnais (1980), any operational applications (Kuittinen, 1992). The Japanese NOAA's National Weather Service has an operational programme MOS-l satellite may change this situation, but its low repeat cycle to obtain such data on the shallow snowpacks of the high plains of and poor spatiai resolution may limit its use (Kuittinen, 1992). the upper midwest USA. This operational programme currently Chang et al. (1987) lIsed NIMBUS-7 SMMR data to produce a flies over 1 400 flight lines in the USA and Canada (Carroll and snow extent map of the whole of the Northern Hemisphere. Carroll, 1989). According to Vershinina (1985), the maximum According to Davis and Shi (1993), "large scale snow cover extent depth of snow water equivalent is limited to about 30-40 em and CHAPTER 2 - APPLICATIONS OF REMOTE SENSING IN HYDROLOGY 17 interpretation of the data can be difficult if the background soil influenced by the snowpack energy balance, snow and air temper­ moisture changes during the season. However, Carroll and Vase atures and vapour pressures, liquid water content and impurities (1984) reported measurements in a forest environment with a snow deposited on the surface of the snow (Rango, 1993). Important water equivalent of 48 em. The water equivalent data generated by snow properties associated with metamorphism include grain size, this method are being used to calculate snowmelt (Vershinina, albedo, layering, surface temperature and snowpack temperature. 1985) and update forecasting models (Bergstrom and Brandt, "Remote sensing monitoring of these changes can be accom­ 1985). Furthermore, the technique also has significant importance plished to a certain degree by selecting the appropriate wavelength for the calibration of areal water equivalent measurements using bands" (Rango, 1993). Currently, grain sizes in the snowpack are microwave sensors mounted on aircraft on space platfonns (Rango, difficult to measure by remote sensing. Although microwave emis­ 1993). In addition to the US and Canada, Norway, Finland and the sion is strongly influenced by the grain size (Chang et al., 1982), Russian Federation also use a gamma ray spectrometer to opera­ there is no dependable way to quantify grain size with microwave tionally measure snow water equivalent (Rango, 1994b). measurements (Rango, 1993). Although the largest sensitivity of According to Lei et al. (1990) and Srivastav and Singh snow reflectance to grain size occurs in the near- infrared wave­ (1991), "good relationships have been established between snow lengths (1.0-1.3 mm), no current satellite sensors make depth and microwave emission and backscatter for snowpacks that measurements (Dozier, 1989). Furthermore, the capability of grain are dry and uniform with little evidence of layering" Blyth (1993). size measurement should increase with the use of future sensors Blyth (1993) further states that "such relationships are not so clear such as the high-resolution imaging spectrometer onboard the once the snowpack has been subjected to thaw and re~freeze EOS platform (Rango, 1993). The albedo of the snow snrface cycles, whilst the presence of unfrozen water anywhere within the changes in time as grain sizes change and surface impurities snowpack results in marked changes in microwave response. In collect on the snow. Albedo is made up of light reflected in all general, the use of microwave radiometry appears more reliable directions off the snow over all portions of the EMS. Albedo than radar for this type of measurement." Chang et al. (1990) used decreases in the visible region as more impurities reach the snow NIMBUS SMMR data to divide the Northern Hemisphere snow surface and in the near-infrared as the size of the individual grains map into a number of depth categories. These authors also derived or clusters of grains increases. As of yet, remote sensing tech­ regional relationships between microwave brightness temperature niques have not come up with the total bidirectional reflectance and snow and vegetation parameters in order to map HW on a distribution albedo measurements (Rango, 1993). However, regional scale. "The sensitivity of microwave brightness tempera­ Dozier et at. (1988) have conducted some research on albedo ture to HW was found to decrease with increasing vegetation measurements. The detection of snowpack layering is possible cover, with forests causing the greatest problem" (Blyth, 1993). using small portable microwave radar systems mounted on skis "Additional research needs to be performed to correct for varia­ and pulled over or beneath the snow surface (Rango, 1993). tions in vegetation cover and snow grain size" (Rango, 1994a), Furthermore, similar active microwave wavelength measurements because snow grain size also affects microwave measurement of from space should be possible because of the inherent high resolu­ HW. According to Goodison et al. (1990), the new SSMII data are tion of radar (Rango, 1993). The surface temperature of the snow being used operationally to produce HW maps of the Canadian can be measured with thermal infrared wavelengths from various prairies; these maps are now supplied operationally to Canadian altitudes; required measurements are made with NOAA AVHRR users (Goodison and Walker, 1993). at least twice every 24 hours. When the surface temperature Various researchers, including Stiles et at. (1981), persists at DOC throughout the diurnal cycle, it can be inferred that Matzler and Schanda (1983) and Rou (1986) have indicated that the snowpack is at or near an isothermal condition (Rango, 1993). the active microwave region has a potential similar to the passive Microwave techniques can be used to identify the initiation of microwave region. However, Rango (1993) points out that not melt metamorphosis in the snowpack. A passive microwave only are active microwave observations of the snowpack very radiometer measures a pronounced jump in the brightness temper­ sparse and almost non-existent, but the analysis of active ature when even a small amount of liquid water is produced in a microwave data is more complex than passive data because of the dry snowpack (Rango, 1993). Finally, it may be noted by a glance confusion caused by the effect of surface characteristics (including at Table 2 that in most cases simultaneous measurements may be soils) and geometry considerations on the reflected radar wave. needed at several wavelengths to determine specific snow proper­ However, the higher resolution (10 m from space) of the active ties. The combination of visible, near-infrared and microwave microwave is a considerable advantage over passive microwave. measurements at various spatial scales as planned for the HOS The major problem, as Rango (1993) points out, is the lack of mission will provide much of the necessary data for sophisticated sensors at about 0.8 em wavelength for experiments on any kind snow measurements and analysis of the snowmelt process of platform. Ullriksen (1985) has conducted research on measur­ (Engman and Gurney, 1991). iug the thickness and water equivalent of snow from a helicopter using short-pulse radar. Although satellite SAR can provide high­ 2.2.4.4 Snowpack ablation and snowmelt runoff resolution data, current single-frequency systems such as ERS-l are likely to be limited to the recognition of the onset of melt and As described above, a combination of these wavelengths should the delineation of wet snow extent (Blyth, 1993). allow a good estimate of the time when the snowpack is ready to transmit melt water from the surface to lower layers (ripe condi­ 2.2.4.3 Snow metamorphism tion) and to eventually produce runoff at the base of the snowpack (Rango, 1993). The first empirical approach to estimate snowmelt The process of metamorphism occurs continually during the snow runoff using remote sensing was developed by Rango et at. accumulation and ablation seasons. Snow metamorphism is (1977); they used satellite observed snow cover data in empirical 18 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

regression models developed for the Indus and Kabul Rivers in the important factor of the future climate is the areal extent ofseasonal Himalayas. Rango et at. (l975) used Landsat snow cover data for snow cover. The SWCC concluded that the most important inputs some basins in the Wind River Mountains in Wyoming, USA to of climate change will be the effects on the hydrological cycle and develop an empirical curve to relate snow cover on a given date to water management systems along with attendant increases in the seasonal runoff. Martinec et al. (1983), Martinec et al. (1994), extreme hydrological events (Askew, 1991). Rango and Martinec and Martinec and Rango (1986) used remotely sensed snow extent (1994) have used satellite areal snow cover extent data to evaluate data from satellites to construct snow cover depletion curves for whether a given climate change scenario will speed up or slow use in the deterministic SRM for snowmelt runoff simulation in down the seasonal decrease of snow cover area. They have the Rio Grande hasin. Martinec and Rango (1987) and Rango and developed a computer program which is now operational for use on van Katwijk (1990) used remotely sensed HW and temperature real basins and is demonstrated on the Rio Grande basin in data to construct modified snow cover depletion curves for use in Colorado and the Illecillewaet River basin in British Columbia. the SRM for snowmelt forecasts in the Rio Grande basin. Kawata Rango (1992) reported on a worldwide testing of the SRM using et al. (1988) used Landsat MSS data with SRM to predict runoff remote sensing inputs with applications for predicting the effects of and manage water levels in the Sai River dam, near Kanazawa, climate change. His initial results show some potentially serious Japan. The deterministic complex subalpine model, developed by problems involving water supply, flooding and drought. the US Forest Service to simulate daily streamflow (Leaf and Brink, 1973) during the snowmelt season is updated by real-time snow telemetry data and snow cover area data. Satellite snow 2.2.5 Ice on land cover data are being used operationally in the deterministic streamflow synthesis and reservoir regulation (SSARR) model in Ice on land may exist in the fonn of glaciers, lake and river ice, and the Pacific north-western US (Engman and Gurney, 1991). The permafrost as well as other more transitory fonns of ice like hoar­ NWS maintains a hydrologic simulation model in which a vital frost and rime ice. Land ice, a key component of the global component is the snow accumulation and ablation model (Carroll hydrologic cycle, represents both short- and long-teon storage of et al., 1993); it uses observed temperature and precipitation data to freshwater. According to Kuittinen (1992), glaciers consist of multi­ simulate snow cover conditions. The simulated model states are year ice which partly melts in spring and increases the streamflow. updated throughout the snow season using snow water equivalent In mountainous areas, long-term storage of water may result when estimates obtained from airborne and ground-based snow water the snow does not melt completely over a number of years, and equivalent data (Carroll et at., 1993). Leavesley and Stannard turns to fim and, with further compaction and movement, to glacier (1990) used remotely sensed snow cover extent to improve their ice. Glaciers cover about 10.8 per cent of the Earth's land surface hydrograph generation in their energy balance based deterministic but only contain about 2.15 per cent of the Earth's water. However, model PRMS (precipitation runoff modelling system). Marks glacier ice contains about 99 per cent of the Earth's surface freshwa­ (1988) used remotely sensed albedo values in his energy balance ter (Baumgartner and Reichel, 1975). Global and long-tenn glacier snowmelt runoff model. Finally, as Rango (1993) noted, "it is not area studies are important for the detection and monitoring of surprising that more of these snowmelt runoff models are turning climate change. Furthermore, in some areas glaciers are an impor­ to remote sensing input data. The remote sensing approach tant source of water for irrigation, water supply and hydroelectric supplies the kind of areal input that the deterministic, distributed power generation. Farnsworth et at. (1984) point out that ice cover models were originally designed to utilize." on lakes and rivers can significantly affect flows in the rivers. Ice In mountainous snow regions, snowmelt becomes an cover and the breaking up of icc cover impacts people living or important component of runoff, making up usually greater than working along the banks when ice jams cause flooding or when 50 per cent of the total streamflow. In some mountain basins, flowing blocks of ice cause mechanical damage to bridges and other snowmelt makes up 95 per cent of the runoff. Remote sensing is structures. Ice cover also affects commercial and recreational traffic usually very successful in mountain regions, especially when the in lakes and rivers. Similarly, hydroelectric power generation is objective is the mapping of snow cover. This snow cover mapping influenced by ice, which at times controls the flow of water, and by is only hindered in regions with very dense forest cover. the damage that may be done to generating equipment. Kuittinen Engman and Gurney (1991) state that new models that (1992) states that the presence of ice cover on lakes and rivers is are developed to use remote sensing data will also improve the important to know when estimating the evaporation from lakes. snow hydrology predictions. Engman and Gurney (1991) further Lake ice in high latitudes may be of particular interest as an indica­ state that the merging of remote sensing data with digital elevation tor of regional and even global climate because it can be monitored modelling and GISs enables different types of data to be combined for date of fonnation and melt each year. The dates of freeze-up and objectively and systematically. Digital elevation modelling is used break-up of lakes and lake ice thickness are very sensitive to air to normalize imagery by using the elevation of the sun and the temperature (Barry, 1984; Palecki and Barry, 1986). Long-term slope, aspect and elevation of the terrain (Baumgartner, 1988; storage of water may occur in permafrost regions if the permafrost Miller et al., 1982). GISs are helpful for combining vegetation contains water in the form of ice. However, permafrost may be masks with satellite imagery (Keller, 1987). comprised of any material (e.g. rock, soil) that remains below ODC In order to properly address the problem of climate for two consecutive years or longer. According to Kuittinen (1992), change, as highlighted by the Second World Climate Conference the soil frost existing in winter in high latitudes and altitudes and the (SWCC) in Geneva, 1990, several important factors must be permafrost in the Arctic regions cover about 20 per cent of the land considered (Rango and Martinec, 1994). While the air temperature area of the Earth. Frost and pennafrost affect the infiltration, evapo­ will gradually increase in the next decades (Schneider, 1989), the transpiration and moisture conditions of the ground. Permafrost also future precipitation change is less predictable. In addition, an prevents groundwater movement (Ron et at., 1985). CHAPTER 2 - APPLICATIONS OF REMOTE SENSING IN HYDROLOGY 19

"The important characteristics of ice are the extent, relatively high-resolution sensors such as Landsat MSS and 'I'M or amount, temperature and melting rate. All these give valuable SAR sensors (Sehneider ef aI., 1981). Gatto (1990) used photo­ information, e.g. for predicting the streamflow. The area and interprelation of Landsat MSS, TM and Return Beam Vidicom thickness of the frozen ground are important to know when (RBV) images to produce ice conditions on the Allegheny, surface runoff and infiltration are studied and estimated," Monongahela and Ohio Rivers in the central USA for the period (Kuittinen, 1992). 1972-1985. Landsat-derived ice observations compared Remote sensing techniques for detecting the characteris­ favourably with available ground and aerial observations 64-80 tics of ice and frost are at experimental or pre-operational stages. per cent of the time. Lake and river ice is easy to detect when it is The extent and concentration of lake and river ice can be moni­ snow covered, but during periods of thaw and freezing up, the tored operationally if needed using air- and space-based methods. interpretation of ice is very difficult (Sclmeider et al., 1981). This Glaciers can also be monitored, because the changes are not rapid is due to low reflectance of new ice and snow-free wet ice in the and thus clouds do not prevent imaging too often. WMO (l993a) visible and near-JR regions. Ice can melt rapidly and thus only a reports that well consolidated applications are reported in the use weather satellite can provide data frequently enough for the of AVHRR in ice surveillance and snow monitoring. Apart from hydrologist. The medium-grade spatial resolution ofpolar-orbiting obvious contour detection, ice and snow maps are used to estimate weather satellite data limits the use of remote sensing to the water reservoirs. Careful modelling of the hydrological basin is largest lakes and rivers if air-based methods are not used requested, as well as physical models of the relationships between (McGinnis and Schneider, 1978). Cloudiness, however, spoils the snow temperaturelbrightness and total-column water. The success images in many cases, thus limiting the application of operational of these applications is very dependent on the particular basin, and ice monitoring methods. Large-swath radars would greatly sometimes models can be very simple, e.g. in Scandanavia, and improve the operational ice-monitoring potential (Kuittinen, sometimes very complex, e.g. in the Alpine region. For large-scale 1992). Ice thickness has been measured using passive microwave ice surveillance, e.g. over Canada or in the polar caps, the SSM/I radiometers in helicopters and aircraft (Hall et aI., 1981) and pulse resolution is sufficient to provide age and type classification. The radar techniques (Finkelstein et al., 1987). The method is not in most commlidated use of SAR images is for ice surveillance in operational use because its accuracy greatly depends on the ice polar and sub-polar regions, particularly to monitor navigation type measured. The movement of ice can be detected only with conditions in certain seas during the frosting and de-frosting frequent observations. Weather satellite data can be used, but the seasons. WMO (1993a) reports that scientific works are dated relatively low spatial resolution and cloud cover limit the useful­ from the early 1960s, based on global cloud mapping and polar ness of these data. WMO (l993b) reports that charts of ice ice extents monitored by means of multi-temporal image analysis, conditions based on METEORMderived data are used in the generally calibrated against ground truth. The ISLSCP has estab­ Russian Federation to refine forecasts on ice. formation and ice lished a long-standing database intensively used for evaluating the drift on large lakes and reservoirs. An all-weather sensor radar performances ofclimate simulation models. would produce ideal data, but the amount of these data is and will The areal extent and snow cover ofglaciers may be detected continue to be limited due to the narrow swaths and small imaging in the visible, infrared and the microwave regions of the EMS. The time per orbit (Kuittinen, 1992). areal extent of glaciers can be mapped best using satellite imagery. The areal extent of permafrost may be detected in the Today both weather satellites and earth observation satellites can offer visible, near-infrared and the microwave region of the EMS. good multispectral scanner data for this purpose. Ifparticularly accu­ According to Kuittinen (1992), frost and permafrost cannot be rate results are needed, stereographic images from space (SPOn or detected directly by space-based remote sensing because in most aerial photographs can be used (Kuittinen, 1992). Air- and space­ cases vegetation, snow and ice prevent observation of the ground borne SAR images have also occasionally been used (Hall and itself and sufficiently low microwave frequencies cannot be used. Ormsby, 1983). Determination ofthe snow covered area ofthe glacier From low-flying aircraft, however, the detection of permafrost is based on the greater reflectance ofsnow than of ice in the visible during summer is possible with low frequencies (489 MHz) using and near-infrared region. Snow can also be detected when it is wet short-pulse radar (Hall, 1982). This is especially useful in the using radar (Ron et af., 1985). However, inmany cases, clouds inter­ discontinuous zone of permafrost. Vegetation, land forms and fere with the imagery and snow covered ice is almost impossible to watercourses are used as indicators of permafrost when multi­ distinguish from clouds (Kuittinen, 1992). The amount of ice in spectral scanner data are used (Hall and Ormsby, 1983). glaciers can be estimated only if the area and the thickness can be Shuman et al. (1993) used SSMII brightness tempera­ measured. At present there is no remote sensing method for measur­ ture trends and field observations to identify multiple hoar­ ing the thickness of ice in glaciers, but future space-borne radar development episodes in Central Greenland. altimeters should be able to measure the variation of the thickness of glaciers. It is also possible to use stereoscopic images for measuring the shape of the glacier, if the glacier is small (Kuittinen, 1992). 2.2.6 Surface water Satellite imagery is used in the Russian Federation to estimate the areal extent ofglaciers, their movement, traces ofavalanches, and the Surface water which may occur in the fonn of lakes, ponds, reser­ extent and location ofriver icings (Desinov, 1980). voirs, rivers, streams and areas temporarily covered by water is a The area and thickness of lake and river ice may be major hydrological basin storage component. Important measure­ detected in the visible, thermal and the microwave region of the ments regarding surface water include water quantity and water EMS. Remote sensing measurements of ice cover have been quality. Quantitative measurements include surface area, volume, performed routinely for large rivers and lakes. In general, stage/level and depth. Qualitative measurements include tempera­ remote sensing of lake and river ice can be accomplished using ture, salinity and mineralization, suspended sediments, 20 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY chlorophyll, vegetation biomass and other man-made pollutants. effects last for some time after inundation and may de detected for The quantitative measurements play an important role in water up to two weeks or longer after the passage of a flood, the need to delivery and drainage to both agriculture and urban areas and obtain data exactly during the flood "peak may not be necessary flood control. The qualitative measurements play an important (Engman and Gurney, 1991). Flood damage may be estimated role in describing whether or not the water is usable or whether or from multiple images of the flooded area. Sediment-laden water not the surrounding environment may be endangered by pollutants may give a considerably higher reflectance than clear water and in the water (Engman and Gnrney, 1991). In the past, virtually all result in confusion when delineating the land-water boundary. of the surface water measurements have, by neGessity, been esti· Thus, Landsat may be a valuable source of data when no ground mated via point measurements. Point measurements can be made surveys are available. very accurately and have the ability to sample the variables Corbley (1993) used Landsat TM data for monitoring extremely well in the temporal domain (through the use of elec­ the Great Flood of 1993 in the Midwest of the USA. NOAA tronic data logging, five-minute sampling rates are not (1994) highlights that high-resolntion satellite data from SPOT uncommon) (Blyth, 1993). While these point measurements may and Landsat are extremely useful in assessing the extent of be well suited to the needs of process studies and small basin damage caused by natural disasters. EOSAT (1993c) reports that, studies where relatively dense instrument networks can be during the 1993 flood in the Midwestern US, EOSAT and ERDAS supported, as the area of interest increases, they are unable to Inc. set up an emergency situation room at EOSAT headquarters to sample spatial variability sufficiently. Although remote sensing analyse and interpret Landsat flood-zone imagery for the US can never approach the point accuracy of groundmeasurements, Anny Corps of Engineers and Federal Emergency Management remote sensing techniques provide the most efficient means of Administration (FEMA). Staff members from EOSAT and sampling the spatial domain. Furthermore, remote sensing tech­ ERDAS input the Lands"at data into a GIS where these were niques have the ability to monitor and inventory water bodies compared with archived imagery to create classification maps of synoptically in areas or situations that are difficult to monitor the flood areas. The Lands-at data provided comprehensive assess­ using conventional methods, and multitemporal images can detect ment of the devastation and provided a valuable resource to changes in the extension of water bodies. utilities, insurance companies, engineering and consulting firms, as well as government agencies for evaluating the damage and 2.2.6.1 Surface water area making plans to rebuild. EOSAT (1993a) reports that Landsat TM imagery helped monitor and quantify the extent of flooding of the Mapping water surrace area is an important operational technique Fitzroy River in Australia, and was an invaluable information required for two main objectives. One objective is to delineate the source for relief services and planners. Wiesnetand Deutsch land-water boundaries for the purpose of flood monitoring, meas­ (1985) nsed the NIMBUS-7 coastal zone colonr scanner (CZCS) uring flooded areas, floodplain mapping and wetland mapping. A data to analyse the Parana River flood. They concluded that in the second objective is to delineate the surface water bodies for the absence of Landsat data, CZCS imagery provides a very suitable purpose of surface water supply inventories. According to Hockey alternative for delineating large floods even though its resolution et ai. (1990), while delineation of water surface areas in· theory is does not approach that of Landsat, although somewhat better than relatively straightforward, in practice the poor availability of suit­ that of the NOAA AVHRR (825 m for the CZCS as compared able imagery, and the lack of empirical calibrations for local with 1 100 m for the NOAA AVHRR). In spite of the coarser reso­ conditions, make operational use of this variable unlikely in the lution (1 100 ill versus 80 m for Landsat), the NOAA satellite near future. Furthermore, data on the required accuracy and spatial thermal infrared sensor has proved effective in measuring areas of resolution and temporal sampling frequency for remote sensing flood innndation (Berg et al., 1981; NOAA, 1994). Fnrthermore, presented in Appendix II, Table 1 as compared to the specifica­ the NOAA satellites have the advantage of more frequent coverage tions for satellite sensors in the foreseeable future presented in the (twice~daily average versus IS-day coverage for Landsat). Ali et table in Appendix III make this prospect clear. According to al. (1987) have nsed AVHRR Channel 2 data to delineate flooded Salomonson (1983), a rule of thnmb is that to be reliably detected areas in Bangladesh. Blasco et al. (1992) used SPOT data to esti­ in an image, a water body should have dimensions double that of mate the extent of floods in Bangladesh. McKim et ai. (1972) used the pixel size. ERTS-1 imagery to map flood waters in the national programme The delineation of land-water boundaries depends on the for the US for the inspection of dams. India used data from IRS as relative spectral characteristics of soil, vegetation and water well as Landsat, SPOT and NOAA satellites on a routine basis to (Engman and Gurney, 1991). Wavelengths in the near-infrared obtain near-real-time infonnation on damages due to floods. Flood region provide the best data for delineating the land-water inter­ maps showing areas in standing water and wet areas and crop face because of the very low reflections of water in this region of damage information are being provided in India within a week to the EMS. Microwave data are also ideal for identifying land-water district authorities entrusted with the task of providing relief (Rao, boundaries because the dielectric constant for water is consider­ 1989). ably greater than that for soil or vegetation-covered soil (Engman All-weather flood mapping is possible with active or and Gurney, 1991). The areal extent of flooding has successfully passive microwave data. Microwave data are ideal for this applica­ been mapped using black and white photography, infrared photo­ tion because cloud cover frequently present during floods does not graphy, thermal infrared data, multispectral scanner data and radar. obliterate the target area and in most cases a sharp land-water For most of these remote sensing techniques, determining areas of bonndary can be defined. The NOAAlNESDlS scientists (NOAA, inundation depends upon measuring reductions in reflectivity 1994) used passive microwave data from the DMSP F-10 and caused by standing or flowing water, high soil moisture, moisture­ F-ll series of polar-orbiting satellites to monitor the Great Flood stressed v6getation and temperature changes. Because these of 1993 in the Midwest USA. Under similar situations, radar CHAPTER 2 -APPLICATIONS OF REMOTE SENSING IN HYDROLOGY 21 systems are capable of higher spatial resolution than passive includes the area and volume of the water body, the boundaries, systems. Although the NIR part of the EMS is most suitable for vegetation types and patterns, and water movement directions delineating land-water boundaries, the dynamics of the land-water (Reinhold and Linthurst, 1975). Remote sensing is a very useful boundary can best be monitored with cloud-penetrating radar. tool to carry out this type of inventory because wetland studies are Blyth (1993) states that the most important factor determining the generally conducted for large and relatively inaccessible areas. ability of airborne or satellite radar to detect inland water bodies is Most large-scale applications involve delineation of the wetlands wind-induced roughness of the water surface. Furthermore. in tenus of the upland boundary and mean high-water mark, as smooth water behaves as a specular surface feature which reflects well as the major species of vegetation. Generally, the vegetation most of the radar energy away from the antenna and results in the is used to identify the boundaries since the species of vegetation is water body appearing black in the radar image. In most situations, usually indicative of the depth and duration of inundation (Carter this is in sharp contrast to the surrounding land surface which is and Schubert, 1974). Because of the importance of vegetation rougher and results in much higher backscatter. Current satellite identification, most applications require the use of seasonal multi­ ground resolution is about 25 m which restricts applications for spectral data that include visible and near-infrared measurements major river systems or poorly monitored areas (Blyth, 1993). Of (Salomousou, 1983). Low-altitude black and white, panchromatic, the satellite SARs planned for the near future, RADARSAT may colour, infrared and thermal photography have all been used be the most useful for operational purposes using the optional successfully in wetlands mapping. The resolution of current satel­ Scansar mode to provide complete coverage of the Northern lite data limits some of their applications to wetlands (Engman Hemisphere north of 49°, with a three-day return period (Ahmed and Gurney, 1991). In many cases, Landsat data will not yield as et al., 1990). Imhoff et al. (1987) used the Shuttle imaging radar high an accuracy as can be achieved with aerial photography (SIR-B) to map flood boundaries and assess damage over areas of (Werth and Meyer, 1981; Gammon er al., 1981). Furthermore, Bangladesh. The digitally processed and geometrically rectified Ackleson et al. (1985) concluded that the higher-resolution SPOT radar data were compared with earlier Landsat data to estimate the data delineated water masses with a high degree of accuracy. areas of inundation subsidence. Research in this area is still Landsat data have been used to make estimates of wetlands water needed to understand how vegetation shadows, depressions, etc. volumes on a monthly basis by combining depth/stage relation­ affect the radar response before it can be used on a satellite system ships with surface water area (Higer et al., 1974). Finally, an (Engman and Gurney, 1991). However, a study by Ormsby et al. additional aspect of wetlands management involves the extent of (1985) has shown that L-band radar data acquired by SEASAT dredging, lagoons drainage and other man·induced changes that and SIR-A can detect flooding under forest canopies. Richards et have an impact on the natural environment. Furthennore, remote al. (1987) proposed a backscatter model for L-band HH polarized sensing is well suited to operationally monitor these changes and data that explains the increased backscatter observed from flooded to make preliminary estimates of the enviro~ental impact. The forests. According to Engman and Gurney (1991), it appears that temporal aspect of Landsat data allows changes to be observed radar will be extremely useful for monitoring floods and land­ over time and in some cases predevelopment baseline data to be water boundaries under vegetative canopies, but there is still a obtained from early scenes (Engman and Gurney, 1991). great deal of research needed to define all the system and target The delineation of surface water bodies and the inven­ parameters necessary for a useful system. tory of surface water supplies, including lakes, ponds and EOSAT (1993b) and RADARSAT Internatioual, Inc. reservoirs, have historically been done using photointerpretation (RSI) (COSPAR, 1993) have illustrated the power of merging techniques, but recently digital multispectral data have also been SAR data from ERS-1 radar satellite with multispectral sensor used. Multispectral data can be used with an automated analysis data from Landsat and SPOT, respectively, to monitor the 14 July procedure to achieve repetitive, rapid results that in many cases 1993 floods in St Louis, MO. EOSAT (l994b) reports that during also meet the accuracy requirements. In general, the accuracy of the late December 1993 torrential rains in France, Germany, detection and measuring water bodies has been largely a function Belgium and the Netherlands, which triggered the worst flood in of proper identification of water and the sensor spatial resolution. Europe in more than 60 years, the ERS-I SAR image recorded the Identification problems involve confusion with areas with similar flooding as it occurred in the German cities of Cologne, Bonn and appearance, such as cloud shadows, dark soils and urban areas. Dusseldorf. Furthermore, the ERS-l SAR/Landsat TM merged However, aerial photograph interpretation can be used to minimize image provides important infonnation for studying the effects of these errors and to check the initial results. Therefore, for the swollen Rhine in the region south of Dusseldorf. extremely accurate work, aerial photographs provide the best Floodplains have been delineated using remotely sensed sources of data. Satellite data also provide good sources for deter­ data to infer the extent of the floodplain from vegetation changes, mining morphometric parameters (such as length, width and soils or some other cultural features commonly associated with surface water area for different elevations) if the resolution is suit­ floodplains. Landsat-delineated floodprone area signatures were able for the specific use. From a remote sensing perspective, water found to have, as yet, a relatively unexplained correlation with the has a relatively low reflectance, especially in both the near­ 100-year flood boundaries. Furthermore, although Landsat TM infrared and visible portions of the EMS. This fact will help to and SPOT data give better results and more accurate delineation of separate urban areaS, fields and sometimes cloud shadows, and floodprone areas, for many legal requirements it is necessary to ambiguities produced by variations in atmospheric transmission map floodprone areas from high-resolution aerial photography (Engman and Gurney, 1991). The use of TM data with its 30 m (Engman and Gurney, 1991). The environmental importance of nominal resolution will increase these accuracies. Furthermore, wetland mapping has led to an increased awareness of this the data from SPOT will be even better, with the expected accu­ resource, and a primary requisite to management and protection of racy being improved (Engman and Gurney, 1991). "Remote these areas is an accurate map and inventory. Any inventory sensing, for the most part, can only determine the surface of the 22 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

water and cannot measure the volume directly_ Mapping surface 2.2.6.3 Surface water depth water area in reservoirs can be llsed to estimate the volume of water in storage. The procedures that are used to estimate lake Surface water depth measurements are based on the penetration of volumes depend on an empirical relationship between surface area light into the water and reflectance of this light from the bottom to or shoreline length and volume. Either an area/volume relationship the sensor. The best part of the EMS is between 0.5 to 0.7 mm may be developed or topographic features can be used to estimate (Kuittinen, 1992). Saxena (1982) has studied the use of aerial the water stage in the reservoir and then relate the stage to water photography, multispectral scanners and lasers in this context. The volume. Government agencies can use this approach not only to main difficulties in mapping water depth are the variations in supplement data they obtain for the reservoirs they manage, but bottom types and water turbidity; these both interfere with the also to maintain an awareness of reservoirs they do not control but calibration of the optical methods (Kuittinen, 1992). According to which may affect their own management strategies under extreme Kuittinen and Sucksdorff (1984), using aerial photography, water conditions such as major flooding. These techniques clearly depths up to several metres can be measured in lakes and the demand high .spatial resolution data except under extreme flood errors are generally about 0.5 m. Ibrahim and Cracknell (1990) conditions" (Engman and Gurney, 1991). used Landsat scanner data and aerial photography in the 0.4-0.6 mm spectral range to measure the water depth near 2.2.6.2 Surface water stageRevel Penang Island in Malaysia. Lyons et al. (1992) used airborne multispectral data and models of the light interaction with water "Radar altimeters, as shown on the SEASAT and GEOSAT satel­ for general assessment of river bottom soil types and water depths lites, although designed for use over the oceans, have also in their study area in St Mary's River, Michigan. Penney et al. provided useful information on inland water level and extent" (1989) measured water depths up to 50 m using a dual-laser (Blyth, 1993). Such measurements from satellites require accu­ system. For the dual-laser system, one wavelength (usually rate information on the satellite position and atmospheric between 0.6 and 1.0 mm) is reflected by the surface water while conditions. These satellite measurements are only applicable for the second wavelength (usually between 0.4 and 0.6 mm) pene'­ large water surfaces' due to spatial resolution. Mason et aI. (1990) trates the water and is reflected by the bottom sediments. The were able to monitor lake changes to +25 em. Cudlip et al. (1990) difference between the two laser measurements is water depth. found that in wetland regions, changes in water level could be Furthermore, water clarity (colour, turbidity, suspended sediments measured with a precision of 10-15 em if adjacent areas of fixed and chlorophyll) will limit the application of optical techniques or known surface elevation were available to reduce the effect of for measuring water depth (Rango, 1994b). Atcording to satellite orbit error. According to Blyth (1993), preliminary Kuittinen (1992), the use of radar which gives only an indication results from the ERS-l radar altimeter suggest that water level of water depth can sometimes be useful in turbid waters. The precision better than 10 em should be possible over approxi­ measurement of the bottom topography by radar is based on mately 700 large lakes (>100 km2) worldwide and over many observation of changes in waves and currents at the water surface, smaller lakes and wetland areas to a precision which will vary which are both affected by the bottom topography. according to the size and location of the water body. Furthermore, Mason et al. (1990), Birkett (1994) and Birkett (1995) state that 2.2.6.4 Surface water temperature through global monitoring of such water bodies (and especially 'closed lakes') using radar altimeters, it may be possible to obtain "Surface water temperatures have been detennined operationally a proxy measure of climate change. Morris (1993) reports that for many years, mainly using weather satellites and airborne temporal changes in the altimeter-derived lake levels are actually scanner data" (Kuittinen, 1992). For instance, Ibre et al. (1979) a combination of real lake-level variation and altimeter error. used satellite infrared data to measure the daily surface water Morris developed a technique to overcome this problem and temperatures of Lake Ontario in Canada. Remote sensing measure­ applied it to data collected over the Great Lakes by GEOSAT ments of emitted thermal infrared and microwave radiation can be (11/86-11188, 17-day repeat cycle) and Topex/Poseidon used to estimate surface water temperatures. The surface tempera­ (9/92-7/93, 10-day repeat cycle). Lake levels obtained from the ture seldom represents the water temperatures farther below the Great Lakes Division, NOAAJNational Ocean Service, were used surface and, furthermore, the temperature profile cannot be as ground truth. The GEOSAT results have a root mean square obtained by the remote sensing techniques. However, Engman and (RMS) error of 11.1 em, while the Topex/Poseidon provides a Gurney (1991) point out that thermal maps of surface temperatures substantial improvement with an RMS error of about 5 em. Not have been useful for siting discharge pipe from factories and only is the annual cycle of the lakes accurately measured, but thermoelectric generating stations. The use of thermal infrared subtle changes in lake level are also detected. Furthermore, the radiation requires calibration which is performed with data from TopexIPoseidon is in place over more than 45 inland bodies of another channel to correct the attenuation effects of the atmos­ water, which suggests that this satellite could provide a signifi­ pheric moisture. Gibbons et al. (1989) used thermal infrared cant database for hydrological research. Laser profiling radiation measurements from Landsat TM and the NOAA AVHRR techniques from airplanes can be used to measure water storage to measure surface water temperature. Lathrop and Lillesand accurately. With accurate knowledge of the altitude of the (1987) used Landsat thennal TM to map surface temperature distri­ airplane using Global Positioning System (GPS) navigation, butions in the Great Lakes. Gagliardini and Karszenbaum (1987) accurate water stage can be measured. If GPS navigation is not used NOAA AVHRR to map surface temperature in La Plata available, then water stage can be detennined with laser profiling estuary. Aircraft measurements of thermal infrared and microwave by measuring a site of known elevation and relating water stage radiation data have been used to measure surface water temperature elevation toJhe known elevation (Rango, 1994b). and to map thermal flows from dams and power plants {Rango, CHAPTER 2 - APPLICATIONS OF REMOTE SENSING IN HYDROLOGY 23

1994b). Although microwave estimates of water surface tempera­ post-monsoon seasons. SPOTLIGHT (1993) reports that using ture are less dependent on atmospheric conditions, they provide spar data and GIS technology, the state of Maryland is monitor­ coarser resolution than thermal infrared data (Shutko, 1985, 1986). ing sediment and nutrient runoff loads from agricultural areas Furthermore, Kuittinen (1992) reports that although the scanning within the Chesapeake Bay watershed. microwave radiometer of the satellite NIMBUS-7 has been used in studying water surface temperatures, the very poor spatial resolu­ 2.2.6.7 Surface water chlorophyll tion will continue to prevent the use of microwave radiometers in mapping inland water temperatures. Concentration of chlorophyll can be estimated using aircraft or satellite scanner data (Gitelson et aI., 1988; Shutko, 1990; 2.2.6.5 Surface water salinity and mineralization Strumph and Taylor, 1988). Calibration data are needed to quan­ tify the relationship between chlorophyll and reflectance. Shutko (1985, 1986, 1987) has shown by theoretical studies and Estimates of chlorophyll in oceans and estuaries have been made demonstrated by experiments in laboratories and on aircraft plat­ using data from CZCS, AVHRR and other satellite scanners. forms that microwave radiometry can be used to measure salinity Current satellite scanners provide only broad wavelength data that and general mineralization of water. This application makes use of may be useful to determine chlorophyll in surface waters with low the sensitivity of microwave emissivity to water conductivity vari­ « 25 mgll) suspended sediments. At higher suspended sediment ations caused by changes in concentrations of salts, acids and concentrations, the reflectance from suspended sediments will other chemicals including waste water outflows. In addition to mask the chlorophyll reflections in these broad wavelengths measurements of microwave brightness temperature measure­ making estimates of chlorophyll very difficult (Ritchie et at., ments, actual physical water temperatures must be estimated by 1992). Future satellite scanners with higher spectral resolution thennal infrared techniques (Rango, 1994b). For obtaining quanti­ may make it possible to determine chlorophyll in the presence of tative estimates of salinity and mineralization, measurements have suspended sediments. Using the same principles, macophytes and to be conducted at microwave lengths of 10 em or longer (Shutko, other aquatic vegetation in surface water can be mapped 1985, 1986, 1987). Furthermore, this approach works best where (Ackleson and Klemas, 1987). Furthermore, laser-induced fluor­ there is a strong contrast in the salinity, such as in coastal regions. escence has been used for observation of chlorophyll in terrestrial water bodies (Brislow et al., 1985). 2.2.6.6 Surface water suspended sediments 2.2.6.8 Other pollutants in surface water The determination of the amount of sediment in water is based on the reflectance of radiation in the visible and infrared parts of the Oil spill detection and monitoring on surface waters using scanner, EMS (Kuittinen, 1992). In general, reflection is a nonlinear func­ laser or radar data are operational programmes (Cross, 1992). Lion of the concentration of suspended sediments with maximum Remote sensing programmes are used routinely to follow the fate reflectance dependent on wavelength and suspended sediment of oil spills in surface water. Changes in water colour due to concentration. Because turbidity and suspended sediments are natural organic chemicals may change reflectance (Carder et at., closely linked in most water bodies, estimates of turbidity can also 1991; Davies-Colley et al., 1988). Remote sensing of these colour be made. Curran and Novo (1988) used satellite and aircraft changes requires high spectral resolution data. Laser-induced scanner data to estimate suspended sediments. Scanner data could flourescence has been applied mainly in the marine environment, be used to develop a monitoring programme to follow changes in for example for the detection of oil pollution (Camagni, 1988). suspended sediments. A limitation on the use of this technique is Huhnerfus (1986) points out the inadequacies of airborne radar the need to collect field data to calibrate the relationship between and infrared radiometers and/or a LIDAR system in discriminating suspended sediments and reflectance. Ritchie and Cooper (1991) between crude oil spills and monomolecular sea slicks. report that calibration curves are not time dependent and may be Furthermore, Vane and Goetz (1988) state that the availability of applicable to regional areas. Furthermore, scanner data can be increasingly sophisticated devices may permit the identification of used without calibration data to map relative suspended sediment more pollutants and it will be interesting to see what the airborne concentrations in river plumes and draw conclusions about sedi­ imaging spectrometers have to offer in this regard. ment deposition patterns in lakes and estuaries. Some examples of Observations of colour are often made to evaluate the suspended sediment measurement from airplanes up to 50 mg/l amounts of living and non-living substances in the water. Plankton are given by Gitelson et al. (1988) and Shutko (1990). According algae are commonly associated with specific colours: blue-green to Engman and Gurney (1991), water bodies that have high algae are linked to dark-green colours; diatoms impact a yellow or suspended sediment concentrations would pinpoint upstream yellow-brown hue; some zooplanktons impact reddish hues; watersheds where major erosion problems are present. Thus, whereas humas can cause water colours to vary from green to dark conservation agencies could use this information to plan erosion brown. Although dissolved chemicals can impact a specific colour control strategies and focus their efforts on the most seriously to a water solution, these changes are frequently too subtle or affected areas. Soliman and Zaki (1986) observed the propagation obliterated to other sources of colour in natural water and these of silts with the advancement of the Nile flood in the Aswan Dam are generally not detectable by common remote sensing tech­ with the aid of Landsat imagery. Manikiam et at. (1993) used niques. Furthermore, inexact corrections for atmospheric multi-date data from the IRS-IA satellite to analyse sedimentation absorption of reflected light also inhibit the measurement of true levels and dispersal patterns for the coastal region of Mangalore in colour by remote sensing techniques (Engman and Gurney, 1991). Southern India. Digital techniques were used to bring out the vari­ The NIMBUS-7 CZCS and the Landsat and SPOT series can alions in sedimentation patterns (up to 14 levels) for the pre- and provide valuable spatial and temporal maps of specific COIOUTS that 24 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

can be related to specific water quality indicators (Engman and basis as needed. Data from satellites like Landsat TM andspar with Gurney, 1991). Feldman et al. (1984) used NIMBUS-7 data to their high spatial resolution are of particular value. Furthennore, with track changes in colour that they associated with phytoplankton their aid, it is even possible to identify seasonal changes of model concentrations. These data documented a major redistribution of parameters giving rise to different hydrological behavior (e.g. phytoplankton around the Galapagos Islands. Viollier et al. (1978) winter/summer infiltration behavior in rainfall-runoffmodelling). performed a series of ocean colour measurements with special Therefore, newer distributed system models based on Landsat and airborne radiometers to estimate the chlorophyll content over the SPOT data allow hydrological computations with higher resolution Gulfof Guinea. in space, thus rendering higher accuracy (Schultz and Barrett, 1989). WMO (1993a) reports that observations of water quality Important qualities of remotely sensed data to be used for basin have been attempted with AVHRR images, supported as practi­ characteristic inventories include spatial and spectral resolution, areal cable by Landsat and SPOT imagery. Direct observation of sea coverage, cost and timelines (Roller, 1984). However, Kuittinen pollution, e.g. oilspills and large algae extensions, is being (1992) points out that all other infonnation available, e.g. maps and reported. The_results appear very occasionally, and are more suit­ possibly some field records, should also be used in the inventory to able for scientific publications than for practical purposes. obtain the best possible result. Because the use of remote sensing to Convincing applications have been based on the CZCS data, detennine basin characteristics requires a significant amount of data which are now only available from archives. The ability to processing and interpretation in addition to various types ofmaps, the measure chlorophyll concentration and suspended sediments, optimal benefit of future inventories of basin characteristics will be either inorganic or of an organic nature, has been well proven with achieved if a GIS is used. Use of a GIS will allow the flexible use of CZCS. Future instruments should possibly combine the colour and infonnation and combination ofall geocoded data with- the data on temperature measuring capabilities for a more comprehensive basin characteristics. The combination ofGISs with satellite imagery observation of coastal waters. Kondrat'ev (1992) has scoped fonns the basis of detailed infonnation on hydrological systems and scientific priorities as a basis for developing a global ecology thus more accurate modelling. The ultimate value of a OIS, in regards observing system in the US. Global monitoring of the land and to the watershed, lies in its ability to organize and input data for world ocean bioproductivity (phytoplankton and forest dynamics) models that simulate hydrologic processes. Such systems or models is characterized as a high-priority task. can be used for analysing trends or for predicting consequences of planning decisions. For instance, the model parameters based on older 2.2.6.9 Surface water vegetation biomass imagery will differ from those after landuse changes have occurred. Thus, the effect oflanduse changes on hydrological processes can be Airborne microwave radiometers, operating at centimetre and quantified with the aid ofhydrological models; the model parameters decimetre wavelengths, may be used to detennine some character'­ are calibrated with the remote sensing data before and after these istics of vegetation above the surface of various water bodies. changes (Schultz and Barrett, 1989). Although remote sensing inputs Biomass and density of the vegetation can be determined in river do not represent the primary source of data, they do represent a deltas, marshy wetland areas and rice fields using Landsat TM potential source of new data for the OIS. Therefore, the efforts to imagery (Shutko, 1986, 1990; Gross ef aI., 1987). Freshwater modify or design hydrological models to be compatible withremote lakes and reservoirs in the southeastern US often produce large sensing data is easily justified on the basis of potential improved beds of persistent and non-persistent aquatic macrophytes. Jensen simulation accuracies and a better understanding of physical et al. (1993) analysed multiple date SPOT panchromatic data processes. Furthermore, because remote sensing methods are (10 x 10 ill spatial resolution) obtained in April and October of sometimes expensive, they canbe justified only if good hydrological 1988 through 1990 to inventory the spatial distribution of cattail models are available and these models have been proven to be and water lily beds in a freshwater reservoir located on the sufficiently useful (Kuittinen, 1992). The full advantage of remote Savannah River Site in North Carolina. sensing lies in the ability to provide up~to-date landuse and basin characteristics data for inclusion in distributed hydrological models (Blyth, 1993). 2.2.7 Basin characteristics Remote sensing techniques used to estimate basin char­ acteristics include aerial and satellite imagery in the visible, Physical basin characteristics include basin boundary (shape), infrared and microwave regions of the EMS. Also, laser remote area, drainage density; length, slope (topography), land uselland­ sensing is a rapidly expanding research area with potential appli­ cover, soil type, vegetation canopy and channel cross-sectional cations for estimating several basin characteristics: Although properties. Basin characteristics are required for the estimation of conventional aerial photography allows mapping of basin charac­ hydrological model parameters. All hydrological models that are teristics on the necessary scales, the use of aerial imagery will be to be applied at the basin scale require parameters that link the limited to some extent because of the high-resolution satellite hydrological processes describing the input, storage and output of imagery. Furthennore, the spectral resolution attainable with scan­ water to the physical characteristics of the basin. Most of the basin ners makes their data more suitable than cameras for various characteristics are either fixed or change only seasonally. mapping and inventory purposes. Electro-optical sensors will be Fiuthennore, conventional estimates of basin characteristics have the main satellite-based sensors available for mapping purposes. often been from inaccurate or obselete maps. Secondary to basin Scanning sensors with wavelengths ranging from shortwave characteristics are phenomena such as desertification, erosion, visible data to longwave microwave data are either currently in sedimentation and forest fires. orbit or being launched. The all-weather, day-night capability of Remote sensing techniques, especially satellite based, offer microwave data is especially advantageous and results of experi­ promising potentials of mapping basin characteristics on a repetitive mental inventories using radar (active microwave) data are quite CHAPTER 2 - APPLICATIONS OF REMOTE SENSING IN HYDROLOGY 25 favourable (Brisco et af., 1983; Foster and Hall, 1981). But a for its coarser resolution (30 m). Drainage density, which is the better understanding of how microwaves respond and react to number of stream channels per unit area, can be extracted different kinds of media is needed before these data can be adequately in those areas where the landscape is sufficiently utilizied in an operational way. According to Blyth (1993), the dissected to permit detailed drainage mapping. The drainage monitoring of basin characteristics using remote sensing is an area system is controlled by the slope of the surface and the type of which has seen slow progress in the past, mainly because of the underlying rock (Travaglia, 1989). Drainage patterns can provide lack of suitable satellite monitoring systems. However, this situa­ useful information concerning rock and soil relationships and tion should improve as new microwave sensors come into use. landform associations. The largest watercourses can be detected Furthermore, most basin characteristics could be supplied for even using weather satellite images and the good quality of operational purposes using remote sensing from aircraft because Landsat TM and SPOT data allows very accurate mapping of the repeat coverage is not a prerequisite. Using satellite-based electro­ water-covered area (Kuittinen, 1992). Foster and Hall (1981) optical systems, scales of up to 1:50000 may be obtained (Doyle, concluded that when weather conditions were optimum, visible 1984). Due to the stereoscopic capabilities of the SPOT satellite and near-infrared imagery from Landsat RBV data provided more and the high spatial resolution of its sensors (10 m), the smallest information on drainage density of the Wind River Range area in errors achieved are now 5-20 m in both the vertical and horizontal Wyoming than the SAR data. Furthermore, the inspection of planes. For economic reasons, however, the most suitable scales seasonal data over a given basin increases the chances that are those of I: 100 000 to 1:250000. It should be stressed that detailed.drainage network infonnation can be extracted. The basin increasing spatial resolution means increased amounts of data and length is usually used in computing a time parameter, which is a processing costs, and for each task the spatial resolution must be measure of the travel time of water through the basin. Topography selected carefully. For watersheds larger than 10 000 km2, data is required for the estimation of model parameters controlling from polar-orbiting satellites having reduced resolution capabili­ precipitation, vegetation, evapotranspiration and runoff. Petrie and ties are more practical (Tarpley, 1984). EI Niweiri (1992) used six types of satellite imagery including Conventional aerial photography is still the most suit­ Landsat MSS, RBV, and TM sensors, the MOMS scanner, the able remote sensing technique for measurement of channel ESA Metric Camera and NASA's Large Format Camera for topo­ cross-sectional properties/dimensions including channel erosion graphic mapping at 1: 100 000 scale over a test area established in and profile, water depth and even flow rate; additionally, overland the Red Sea Hills area of the Sudan. Further interpretation tests flow and basin boundaries in flat areas. This requires large-scale were carried out over the area of Khartoum and the Gezira. photographs, preferably stereoscopic images. Therefore, satellite Landuse inventories are needed for the estimation of images are less suitable for channel dimensions measurement due model parameters controlling infiltration, evaporation, intercep­ to the picture element presentation of the images (Kuittinen, tion and runoff. There are many landcover types and the accuracy 1992). Furthermore, although the best accuracy for landcover of inventories using remote sensing depends greatly on the data mapping can be achieved using aerial photography, the cost of the used. Also, landuse can change greatly in a short time and thus data is much higher than with satellite imagery which, however, affects the water balance of the area (Kuittinen, 1992). According gives poorer results (Kuittinen, 1992). Kienegger (1992) devel­ to Blyth (1993), in the majority of situations, conventional aerial oped a new system concept and method of implementation for photography orVISINIR satellite imagery such as from Landsat or film-based data capture and updating of geographic information. SPOT is most appropriate for landuse inventories when the timing Use is made of a novel film scanning concept that permits the of data acquisition is not critical. The Landsat TM is currently the complete integration of aerial photographs into the data acquisi­ most widely used satellite mapping system for landuse applica­ tion and maintenance task. The system was employed tions, primarily because its seven spectral bands are well chosen operationally for the collection of impervious surface information for this purpose (Blyth, 1993). Mathematical techniques like to build the basis for a billing system for the waste water service maximum likelihood estimates in a multidimensional feature charge of the City and County of Denver, Colorado. space (based on simultaneous information in several spectral The visible and near-infrared region of the EMS is bands) allow a rather accmate landuse classification or estimation widely used for mapping basin shape, area, drainage density, of vegetation indices which are needed in modelling hydrological length, topography, landcover, soil type, vegetation canopy and processes (Schultz and Barrett, 1989). Two of the most important channel sinuosity. This is mainly due to its relative ease in inter­ landuse types are floodplains and wetlands. The identification and pretation since the energy that is sensed is reflected energy mapping of flood-prone areas or floodplains are critical for emer­ (Salomonson, 1983). The basin shape (boundary) supposedly gency preparedness and for flood insurance evaluation. reflects the way that runoff will arrive at the basin outlet. Divides Flood-inundation mapping using visible data has been successful between adjacent basins demarcate differences in drainage direc­ on many rivers. Wetlands are an essential component of the hydro­ tion which can be readily observed from most earth resources-type logic system. Researchers at the National Remote Sensing Agency satellites. The basin area, which is probably the single most in Hyderabad used Landsat MSS and IRS-IA data over a period important basin characteristic for hydrologic design, reflects the of 20 years to follow the development of the spit being created volume of water that can be generated from rainfall. The drainage from sediments deposited from the Ballaleru River in Andhra area is the easiest basin characteristic to identify using satellite Pradesh. The data show that vegetation cover has increased over imagery. Of the earth resources satellites which are currently oper­ the years as the spits become broader and more stable. If these ating, the SPOT satellite system with its stereoscopic viewing processes continue, a lagoon will develop (EOSAT, 1994d). A capabilities and superior resolution (10 m) provides excellent third important landuse type is impervious areas; urban areas mapping of drainage area and density. The Landsat TM has affect the hydrology at local scales by concentrating and enhanc­ greater spectral range than the SPOT system which compensates ing runoff. Only data of good spatial resolution can be used 26 CURRENT OPERATIONAL APPLICATIONS OP REMOTE SENSING IN HYDROLOGY because urban (and agricultural) areas cannot be detected with include slope, aspect, elevation, hydrology, location of research sufficient accuracy using weather satellite data. EOSAT (1994c) and inventory plots, crown closure, size class/stand structure, reports that the state of Sooth Carolina nsed Laudsat TM data to species, current vegetation type polygons and other information. develop a detailed landcover map that encompasses 19 classes The mapping was performed using field measurements, aerial versus 10 classes in an earlier map produced using SPOT data. photography aud Laudsat TM satellite data. WMO (1993a) reports IRSA (1992) reported that the main emphasis of the work in the that AVHRR imagery has been applied in the discovery and moni­ Environmental Mapping and Modelling Unit concerns using toring of forest fi'res. Pilot projects have developed convincing Landsat TM data to map the various components of the European models of the overall warning system, within the framework of the landscape and 011 development of methods for applying remotely International Decade for National Disaster Reduction. Roberts et sensed data and GISs to management and protection of the al. (1993) analysed SPOT aud aircraft imagery to determine the European environment. The British Institute of Terrestrial Ecology distribution of the various vegetation types within the Balquhidder (lTE) produced the first computer-compatible digital landcover catchment in the UK for water balance studies. Channel cross­ map of Great Britain using Landsat TM data inApril 1993 (Fuller, sectional properties are needed for the computation of runoff. 1993). The map is integrated with other data in Decision Support The thermal infrared region of the EMS is valuable for Systems and GISs, such as the "Countryside Infonnation System" mapping topography. Digital infrared data have been analysed to (Fuller. 1993). Corbley (1994) reports that the Forest Survey of correlate satellite-derived temperatures with local topography India is using IRS data to map the entire country's forest land each (Schneider et at., 1979). Flood-inundation mapping using thennal year. India has also' input IRS data into a GIS to classify its 20 per infrared imagery has been successful on many rivers. cent wasteland into 13 categories, many of which will be targeted Passive microwave systems such as the SMMR on the for reclamation. Pulich and Hinson (1992) used Landsat TM NIMBUS-7 spacecraft cannot currently achieve adequate resolu~ imagery for classifying landcover in Texas coastal areas. The tion for mapping watershed features. But, however, for basins Russian KVR satellite photograph of Cairo, Egypt with a resolu­ larger than 10000 km2, the snow cover area extent and snow tion of 2 m has been used for geological mapping of lineaments, water equivalent can be estimated. This information is useful for regionallandcover identification, water resources mapping, and forecasting streamflow in the spring for mountainous watersheds urban aud regional p1auuiug applications (EOSAT. 1994a). Soil where seasonal snowpacks form. However, Choudhury (1993) type inventories are needed for the estimation ofmodel parameters used passive microwave SMMR and SSMII global data on board controlling infiltration, groundwater flow and runoff. According to the DMSP satellite to document seasonal and interannual land Kuittinen (1992), rock and soil types can be roughly mapped using surface changes at spatial scales of 50 km or larger in Saharan remotely sensed, data (especially multispectral scanner data) if Africa; changes which would be expected due to known seasonal good field mapping can be carried out to help the interpretation of vegetation growth aud drought. Sippel et al. (1993) nsed SMMR the images. Vegetation canopy inventories are needed for the esti­ NIMBUS-7 satellite data to reveal the complex spatial and tempo­ mation of model parameters controlling interception, ral variation in inundation patterns in large wetlands in Brazil. evapotranspiration, infiltration and runoff. Vegetation which for Seasonal changes were mapped for the Amazon River floodplain the most part absorbs visible light but reflects near-infrared light and the Paraguay River floodplain. can be detected and classified on multispectral imagery. Because Active microwave remote sensing alone may not be the of the importance of vegetation identification, most applications best method of measuring basin characteristics. but it is likely to require the use of seasonal multispectral data that include visible be used in conjunction with data from other sensors. Evans and and near-infrared measurements (Salomonson, 1983). WMO Smith (1991) found that a combination of airborne SAR and (l993a) reports that the most popular application of NOAA visible data such as Landsat TM enabled the backscattering effects AVHRR imagery is vegetation monitoring by means of the NDVI of vegetation and soils to be more easily separated than with SAR derived from the VIS and NIR channels. alone. The single most useful frequency for delineating basin In a number of countries around the world, the national characteristics is likely to be the L-band because it is less affected meteorological services have introduced the NDVI map routinely by vegetation than higher frequencies. However, many extensive on a monthly basis in order to monitor the seasonal regularity of radar surveys have been carried out with airborne X- and K-band vegetation development over large areas; the Sahel is a well docu­ systems and these have proven to be extremely useful for mapping mented case. According to Kuittinen (1992), if temporal data are hydrological features such as drainage networks (Trevett, 1986). used, the accuracy of vegetation mapping increases and it is poss­ The high resolution of SAR and radar altimeters can be used for ible to obtain, for instance~ vegetation indices for hydrological and detailed watershed mapping. Radar images, if available, can be climatological models. For large areas, weather satellite data can used to determine the density of the drainage network. The ability be used, whereas local differences are in most cases detectable of SAR to penetrate clouds and snow, and its day or night data using earth resources satellite data. A quasi-operational use of a acquisition ability, complements the use of VIS and NIR imagery temporal NDVI was used by Rangoonwala et al. (1993), Gupta et for mapping drainage density. Drainage density detail is good on at. (1993) and Menenti et al. (1993) to monitor vegetal- cover and SAR imagery because individual streams show up well due to crop growth in Pakistan, India and Zambia, respectively. reparian vegetation causing higher radar reflections which result NOAA/NESDIS: recently developed the Vegetation Condition from the "rough" surface which vegetation creates (Foster and Index (VCI), using NOAA AVHRR data, for use in detecting and Hall, 1981). Radar is very sensitive to changes in surface rough-' tracking droughts (Kogan and Sullivan, 1993). Congalton et at. ness and is thus capable of providing a qualitative measure of (1993) incorporated the use of remotely sensed data to produce a roughness for stream channels and slopes required for rainfall­ GIS database and map of old growth forest lands on national runoff models (Blyth, 1993). Multifrequency, multipolarization forest and p.ark lands in Oregon and Washington. The GIS layers and multi-incidence angle radar data are required to help separate CHAPTER 2 - APPLICATIONS OF REMOTE SENSING IN HYDROLOGY 27 roughness groups (Oh el at., 1992) and a combination of L-, C­ of Agriculture's Agricultural Research Service Walnut Gulch and X-band would generally cover a wide range of naturally experimental watershed near Tombstone, Arizona. Ritchie et al. occurring roughnesses. Such data are currently only available (1993a) used a laser altimeter mounted in a small twin-engine from aircraft. Leber! (1983) showed that radar steroscopy using aeroplane to measure surface properties of the landscape in Texas either aircraft or satellite SAR could provide topographic informa­ and Arizona. These airborne laser measurements were analysed to tion that would be useful for hydrological modelling. Zebker et at. provide infonnation on topography, vegetation canopy, and stream (1992) described an interferometric airborne SAR which is being and gully cross-sections. While conventional ground-based tech­ developed with the aim of providing topographic contouring at niques may be used to make all these measurements, airborne 2-3 m accuracy, and this has also been demonstrated to be feasible laser altimeter techniques allow the data to be collected in a quick from space. According to Blyth (1993), microwave radar data will and efficient way over large and inaccessible areas. Ritchie et al. be of unique value in the tropics or mountainous areas where (l993b) demonstrated the potential of digital airborne laser data to cloud cover is unusually persistent, or in arid areas where its measure canopy structure and properties for large areas quickly sensitivity to surface roughness enables spectrally similar surfaces and quantitively. Ritchie et al. (1994) used a laser altimeter to to be differentiated. At low frequencies, radar has the unique measure channel and gully morphology. The data are valuable for ability to penetrate forest canopies (Chauhan and Lang, 1991; measuring channel and gully cross-sections and roughness in rela­ Kimura and Kodaira, 1992) which may facilitate the retrieval of tion to the surrounding landscape, for assessing soil loss from soils, soil moisture information, wetland and flood water extent. gullies and channels, and for providing input to the understanding Richards el al. (1987) fonnd that river flooding could be readily of gully and channel dynamics in the landscape. detected beneath a forest canopy using SIR-B, L-band HH-polar­ ized radar as flown on the Space Shuttle in 1984. Radar images have also been successfully used to map flood-inundated areas on 2,3 REMOTE SENSING INPUTS FOR many rivers (EOSAT, 1993b; COSPAR, 1993). Although the NIR HYDROLOGICAL MODELS part of the EMS is generally regarded as the most suitable for delineating land/water boundaries, the hydrological interest lies in All hydrological models, whether simple or complex, that are to be the dynamics of the land/water boundary which can best be moni­ applied at the basin scale, require parameters that link the hydro­ tored with cloud-penetrating radar, the resolution of microwave logical processes ("primary hydrological variables") describing the radiometers being too coarse for this application. Even for landuse input, storage and output of water to the physical characteristics of mapping, the cloud independence of satellite radar may prove to the basin. Physical basin characteristics include basin area, channel be of critical importance in the future. Leak and Venugopal (1990) length, slope, land use and soil type. Hydrological basin inputs reported that the success of Landsat and SPOT satellite imagery include rainfall and snow. Hydrological basin ~torage components for mapping urban landuse has been limited by the spectral include surface water, lying snow and ice, vegetation moisture, soil overlap of materials commonly occurring in the urban environ­ moisture and groundwater. Hydrological basin outputs include ment. Fraysse and Hartl (1985) found that more categories of streamflow, groundwater flow and evapotranspiration. Model urban landuse were depicted on radar images of SEASAT than on parameters are estimated using the physical basin characteristics. Landsat MSS (78 m resolution) imagery of the City of Cologne. For instance, land use is an important physical basin characteristic Therefore, Blyth (1993) snggests that while the high spatial and which is required for the estimation of model parameters control­ spectral resolution of an airborne multispectral scanner is the most ling infiltration, evaporation, interception and runoff. Furthennore, suitable for detailed urban landuse classification, the mapping of "secondary hydrological variables," which are required mainly for Landsat TM or SPOT with future satellite radar data on digital the calculation of evaporation and for snow studies, include solar image processing systems may provide a powerful method of radiation, albedo, sulface temperature, air temperature, air humid­ urban landuse mapping. ity, atmosphere pressure; wind speed and direction, and water Laser remote sensing measures the reflected energy of a qnality (Blyth, 1993). laser beam or by the Earth's surface. Airborne lasers have been Many of the physical basin characteristics are slow used to collect data which can be used to estimate high~density changing and are conventionally extracted from maps. These topographic information, canopy height, vegetation cover and maps are now periodically updated with the use of remote stream valley cross-sections. Application of these estimates may sensing techniques. include topographic changes that may delineate erosion sources Although the conventional point measurements of the and sinks, erosion and gulley studies, and defining channel geo­ hydrological variables Eboth.primary and secondary) are higWy metry (Engman and Gorney, 1991). Menenti and Ritchie (1994) accurate and have a high temporal sampling frequency, as the presented a new method to estimate the effective aerodynamic basin area increases, they are unable to sample spatial variability roughness of a complex landscape. High-resolution elevation sufficiently. Remote sensing, on the other hand, provides the most profiles measured with an airborne laser altimeter have been used efficient means of sampling the spatial domain, but has in the past to compute geometrical parameters of three landscape elements of sampled the time domain somewhat poorly; this is due mainly to the Walnut Gulch experimental watershed, Arizona. Laser profIles low satellite repeat coverage. yielded geometrical parameters including mean crop height and In the application of remote sensing of hydrological spacing of taller shrubs and trees, mean amplitude and wavelength variables, it is most advantageous to combine both ground-based of hillrocks and ridges. Weltz el a1. (1994) favourably compared and aerial measurements. Remote sensing of the hydrological vari­ the laser-measured properties of plant height and canopy cover ables provides an indirect measurement. The electromagnetic (> 0.3 m) with field measurements made using the line-intercept radiation is sensed and is related to the hydrological variable method at seven of the eight sites evaluated on the US Department either empirically, or preferably via some form of transfer 28 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

function. However, in order for remote sensing to be successful, allow accurate calibration and interactive and real-time implemen­ the physical basis of the transfer function must be well under­ tation of the model (NOAA, 1994). McKim ef al. (1993) stood. Although remote sensing can never approach the point illustrated how the established US Army Corps of Engineers accuracy of measurements derived from the ground, this inherent (COE) SSARR hydrological model was adapted to create a inaccuracy of remote sensing is offset by the ability to measure dynamic model and thus receiving real and near-real-time data spatial variabilities that cannot otherwise be observed. Ground from remote sensors. Kouwen et at. (1993) introduced a method data are used to adjust or calibrate the transfer function to ensure for distributed hydrological modelling which makes maximum use convergence of the remotely sensed and point data. The remotely of remotely sensed data such as rainfall from weather radar and sensed data are then used to extend the point knowledge to land cover characteristics from Landsat or SPOT satellites. The provide information on variability between the points. method eliminates the need for small computational areas while The full advantage of remote sensing lies in the ability maintaining the requirement of computing runoff for homoge­ to provide up-ta-date basin characteristics and the spatial varia­ neous watershed subareas. Their method consists of grouping

bility of hydrological data for inclusion in distributed parameter hydrologic response units that have similar response characteris R hydrological models as opposed to lumped parameter hydrological tics on the basis of satellite classified landcover maps. Because models. Applications requiring simple transfer functions are likely model parameters are unique to individual landcover classes, this to achieve true operational status more rapidly than those having reduces the need for model calibration and allows for the transfer complex relationships between the remotely sensed signal and the of model parameters in both time and space. Two recent radar hydrological variable. Furthermore, most physical basin character­ grid-square model formulations have been developed for real-time istics could be supplied for operational purposes using remote flood forecasting in Italy (Chandler and Fattorelli, 1991) and the sensing for aircraft because repetitive coverage is not a prerequi­ UK (Moore and Bell, 1992; Moore, 1993b). Both models take as site (Blyth, 1993). their starting point the translation of excess runoff from a grid The spatial and temporal characteristics of remotely square using a linear channel, equivalent to a time delay from each sensed data have been successfully used to improve both model grid to the basin outlet. Results obtained for the Italian model on formulation and model performance in hydrology. Additionally, the 1 408 km2 Bachiglione River demonstrated superior perform­ remotely sensed data acquisition capabilities have been paralleled ance to an equivalent lumped model when the flood-generating by equal advancements in digital array processing and GISs, rainfall was not spatially uniform (however, use was made of rain­ which allow the effective extraction of both temporal and spatial gauge data due to lack of available radar data). However, the UK information. As a result, the trend in hydrological modelling has study presented results for two basins served by two different C­ shifted from simple lwnped, linear, deterministic models to more band radars. While the basin located in hilly terrain in northwest complex distributed, non-linear, stochastic models. Papadakis England showed variable results with both radar and rainfall et al. (1993) illustrated the use of multispectral and multitemporal perfonning best in different storms, the performance of the grid satellite imagery to estimate monthly area precipitation which is model using radar data proved consistently superior for the basin then transformed into monthly runoff values with the aid of the in southern England (Moore, 1993a). non-linear Voterra series model. On the basis of satellite data alone, the very short time series of hydrological data can be extended considerably, thus allowing an estimate of the future 2.4 CURRENT OPERATIONALAPPLICATIONS performance of water projects. Ottle et al. (1989) modified a hydrological model to simulate both the hydraulic and energy According to the more stringent definition of operational applica­ budgets at the soil surface. They estimated surface temperature tions presented in 1.4 above, one of two criteria must be met. The from thermal infrared measurements to be used to derive real application must produce an output on a regular basis, or the evaporation and soil moisture. Manavalan et at. (1993) developed remote sensing approach must be used regularly and on a continu­ the Watershed Response Analysis (WARA) model to study soil ing basis as part of a procedure to solve a problem. Although erosion and hydrological responses of a river basin in India. They many remote sensing applications for the various hydrologic vari­ used IRS-LISS 2 satellite data for developing the vegetation ables were presented in Section 2.2 above, only a few qualify as density and soil texture maps. Further, they converted the drainage operational applications if this more stringent definition is used. density and slope angle maps (generated from existing topo­ Some of the applications termed operational by this approach are graphic maps) in the form of digital layers and used an integration listed in 2.4.1 to 2.4.7 below. technique to provide quick appraisals on soil erosion, surface runoff and infiltration potentials of various land units within the watershed. 2.4.1 Precipitation Moore (1993a) reports that the availability of radar rain­ fall distributed on a square grid has provided a stimulus to develop Most of the operational satellite applications for precipitation new rainfall-runoff models configured to make better use of data utilize visible or infrared imagery. Furthennore, there are a few in this fonn. However, the prospects for improvement of model near-operational applications which utilize microwave imagery. performance utilizing more distributed rainfall-runoff formula­ Operational applications using visible or infrared imagery include tions need more investigation. The NWS Dynamic Wave the following: the NESDIS IFFA technique (D'Souza ef aI., 1990) OPERational (DWOPER) model is a physically based, distributed, provides real-time precipitation of heavy rainfalls and has been hydraulic routing model that simulates flow along a river. The used operationally in the USA for nearly a decade for flash flood high-resolution precipitation data expected from NEXRAD and warnings. The FAO method (Barrett and Harrison, 1986) provides the advanced techniques and procedures provided by GISs will rainfall maps which are used operationally in the assessment of CHAPTER 2 - APPLICATIONS OF REMOTE SENSING IN HYDROLOGY 29 locust hazard in northwest Africa. The B4 method which is opera­ Dutartre et al. (1993) used radar imagery and SPOT imagery for tional (Barrett, 1993) uses real-time rainfall data from satellites to hydrogeological prospecting over the Kuturn area northwest of El set rain/no-rain boundaries and rain rates over West Africa. The Fasher, Sudan. Because many Indian villages suffer a severe short­ B4 is also used to provide rainfall input for the Nile Forecasting age of drinking water, India has doubled its success rate of System (NOAA, 1993). Two real-time (so-called "nowcasting") obtaining drinking water from new wells through the use of satel­ rainfall estimation schemes (D'Souza et al., 1990; WMO, 1992a) lite mapping of IRS data. India creates 1:50 000 hydrogeomorphic which use real-time rainfall data include the large-scale maps with IRS data (Cmbley, 1994). RAINSAT2 in Canada, and the FRONTIERS system in the UK (these have been used quasi-operationally). The Finnish Meteorological Institute is operationally applying a Swedish algo­ 2.4.3 Evapotranspiration rithm (WMO, 1992a) to obtain precipitation estimates. PERMIT (D'Souza et al., 1990) is used operationally in Pakistan to obtain Thus far, there appear to be no operational applications for evapo­ daily rain/no-rain boundaries. Near·operational applications using transpiration. passive microwave imagery include the following: NESDIS has been operationally supplying precipitation products (including global precipitation from passive microwave (PMW) imagery) to 2.4.4 Snow the NWS Forecasting Offices and River Forecast Centers since 1978 (Scofield, 1991). Italy (WMO, 1992a) is using PMW There are numerous operational applications for snow cover, snow imagery on an operational basis for the identification of precipita­ water equivalent assessments and seasonal snowmelt runoff fore­ tion nuclei. Adler et at. (1991) combined IR data with PMW data casts. The NWS Office of Hydrology in Minneapolis uses NOAA to produce monthly rainfall estimates over Japan. Negri et at. AVHRR and GOES imagery to operationally provide periodic (1994) used the technique derived by Adler et al. (1991) to derive snow cover extent maps on about 3 000 basins in North America c1imatologies of convective precipitation for the TOGA COARE (Carroll, 1990). Kumar et al. (1991) used satellite snow cover data area. The Hungarian Meteorological Service combines IR data as input to SRM for use in operational forecasts of daily snowmelt with PMW data (TOVS data) to operationally forecast ultra-short runoff in the Beas and Parbati Rivers in India. Harrison and Lucas range precipitation in the Carpathian Basin (Csistar and Bonta, (1989) used NOAA imagery of the UK to operationally produce 1993). Other operational applications of TOVS data for rainfall daily and weekly snow area maps. Andersen (1991) reports on monitoring exist in the Russian Federation, Japan, Finland and using NOAA AVHRR snow maps for operational planning of Italy (WMO, I992a). Operational ground-based radars such as hydropower generation in Norway. The NWS has an operational NEXRAD provide real-time rainfall data for flash-flood warnings airborne gamma ray spectrometry progra~me to obtain HW (Hogg, 1989; WMO, 1992a; Walton et aI., 1990). Additionally, values for the upper Midwest USA and Canada (Carroll and three new radar stations in three mountain tops should give opera­ Carroll, 1989). Furthermore, Norway, Finland and the Russian tional flood warnings in the Swiss Alps (Joss and Lee, 1994). Federation also use gamma ray spectrometers to operationally Furthermore, the Northwest Radar Project in the UK integrated measure HW (Rango, I994b). Goodison and Walker (1993) use the installed radar with an operational flood forecasting and PMW imagery to operationally produce HW maps of the warning service (Collinge and Kirby, 1987). Canadian prairies. Satellite snow cover data are being used opera­ tionally in the SSARR model in the Pacific northwestern US (Engman and Gurney, 1991). The NWS updates the snow accumu­ 2.4.2 Soil moisture and groundwater lation and ablation model component of its hydrologic simulation model using airborne and ground-based HW data (Carroll et al., Only a few soil moisture applications have been operational. The 1993). India has developed computer models which use IRS satel­ FAO (WMO, 1993a) has used VISINIR imagery to estimate soil lite data to determine how much water will melt from the seasonal moisture through precipitation indices on an operational basis. snowcover. This information is used to predict water use for irriga­ Russian scientists have used PMW soil moisture imagery to tion and power generation (Corbley, 1994). develop an operational system for irrigation scheduling (Reutov and Shutko, 1986). The NWS hydrologic runoff forecasting programme (Carroll and Schaake, 1983) uses real-time aerial soil 2.4.5 Ice on land moisture values obtained from the NWS operational airborne gamma radiation snow survey programme in the upper Midwest Remote sensing techniques for detecting the characteristics of ice USA (Jones and Carroll, 1983). and frost are at experimental or pre-operational stages. Several semi-operational applications for groundwater have been accomplished. Zevenbergen and Rango (1992) combined Landsat TM imagery with conventional maps for 2.4.6 Surface water groundwater development in the Nile Valley and Delta as input to a groundwater master plan for Egypt. Dutartre et al. (1993) used There are operational applications for surface water. The India SPOT MSS and aerial photography to map aquifers in the Department of Agriculture uses IRS, Landsat and SPOT Sahelian climatic zone of West Africa. EI-Baz (1993) used data to monitor the impact drought is having on crops, Landsat TM imagery to discover potential aquifers in the Arabian soils and surface water resources, and issues weekly drought Desert via a junction of two known ancient Wadis. Sabins (1983) reports to local governments. The objective is to detect the prob­ used radar imagery for groundwater exploration in Indonesia. lems in advance and take steps to minimize resource losses 30 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

(Corbley, 1994). McKim el al. (1972) operationally used Associated Restructuring (MAR) provided major benefits during the ERTS~1 imagery to map flood waters in the national programme flood event, including flash flood warnings. However, maximum for the US for the inspection of dams. NOAA (1994) uses passive use ofWSR-88D data for hydrologic forecast and warnings awaits microwave data from the DMSP series satellite to monitor the completion of the NEXRAD (now called the WSR-88D) network Great Flood of 1993 in the Midwest USA. Furthermore, EOSAT over the upper Mississippi River basin. In addition, Advanced (1993b) and Radarsat Intemationallnc. (COSPAR, 1993) merged Weather Interactive Processing System (AWIPS), under MAR, is SAR data from ERS-l with multispectral data from Landsat and needed at the River Forecast Centers (RFCs) to process and mosaic SPOT, respectively, to monitor the 14 July 1993 floods in St the infonnation from multiple radars and other areas of forecast Lonis, MO. responsibility (NOAA, 1994). One of the most signifIcant changes will be in the way point precipitation observations will be merged and processed with precipitation estimates from multiple WSR-88D 2.4.7 Basin characteristics radars and information received from sateIlite observations. The result will be frequently updated, multisensor, high-resolution There are several operational applications for basin characteristics. precipitation estimates. It is anticipated that RFCs will have these The Forest Survey of India is using IRS data to map the entire high-resolution data sets for their entire areas of responsibilities. country's forest lands each year; India has experienced a severe The availability and- use of these data sets will change the way deforestation problem in the last 20 years. Furthennore, India has hydrologists interact and use hydrologic forecast models. input IRS data into a GIS to classify its 20 per cent wasteland into Forecasters will change their "mind-set" from executing forecast 13 categories, many of which will be targeted for reclamation. systems at six-hourly intervals, to running the model interactively Also, India has turned to satellite imagery to plan new infrastruc­ in near-real-time. RFCs will also have access to gridded quantitative ture such as roads to meet the demands of the growing cities precipitation forecast (QPF) estimates for use in their forecast (since most of its cities are growing too fast for urban planners to systems (NOAA, 1994). The high-resolution precipitation data will keep up) (Corbley, 1994). The British Institnte of Terrestrial allow hydrologists to re·examine how models are implemented. Ecology (lTE) produced the fIrst computer-compatible digital land Rainfall-runoff models and runoff distribution models will cover map of Great Britain using Landsat TM data in April 1993 eventually change from lumped parameter models to distributed (Fuller, 1993). The ITE has a wide range ofusers of the land cover parameter models based on gridded data. There is a major challenge map which include organizations concerned with environmental in order to implement distributed, physically based impact assessments, pollution control and water- resources hydrologiclhydraulic models that take maximum advantage of the management. For instance, the map data are used to evaluate the new observation systems (such as WSR-88D radars). The main throughput of nitrates and pesticides to groundwater. Furthennore, reason is the massive amount of time and effort needed to assemble the map is integrated with other data in Decision Support Systems the information required to calibrate these models. However, it is and GISs such as the "Countryside Infonnation System" (CIS). imperative that resources be found to accomplish the transition from The administrators use the CIS for the assessment of quantitative statisticalJempirical modelling to modelling the relevant physical information and thus in the decision-making process (Fuller, processes. Otherwise, the great progress made in observation 1993). Roberts el al. (1993) analysed SPOT and aircraft imagery systems, communication links and computer power provided by to determine the distribution of the various vegetation types within MAR will never be fnlly realized by the NWS hydrology the Balquhidder catchment in the UK for use in water balance progranune (NOAA, 1994). studies. The Russian KVR satellite photograph of Cairo, Egypt The development of Hurricane Andrew in August 1993, with a resolution of 2 m has been used for geological mapping of which has been described as one of the most destructive disasters lineaments, regional landcover identification, water resources of the 20th century, was monitored via satellites and research mapping, and urban and regional planning applications (EOSAT, aircraft by NOAA starting from its infancy to a more mature stage. 1994a). These developments were monitored on computer screens at the National Hurricane Center in Coral Gables, Florida and the National Severe Stonns Forecast Center in Kansas City, Missouri. 2.5 FUTURE TRENDS The NWS broadcast of the hurricane warnings and projected storm tracks saved thousands of lives and millions of dollars in The great Flood of 1993 of the enormous Mississippi River basin property damage by making early evacuations possible (Becker, was one of the first American emergencies in which remote sensing 1993). technologies were used to create "pictures" of the flood while the The late December 1993 torrential rains in France, flood was happening. Although the satellite data coupled with new Germany, Belgium and the Netherlands triggered the worst flood­ airborne technology gave scientists an opportunity to monitor and ing in Western Europe in more than 60 years. Authorities relied on assess the extent of the flood's progress, it is presently unclear if satellite imagery for assessing this massive natural disaster, and these data can function as tools of instant crisis management (i.e. thereby assisting in disaster relief, because of its large-area repeti­ help predict flooding) for the flood-type of catastrophic event, the tive coverage (EOSAT, 1994b). main reason being the inability of the data to provide coverage on In February and March of 1993 one of the largest rivers demand of the flood's continuing progress. However, this kind of in Australia flooded its banks. Landsat TM imagery helped data can identify precise flooded areas to prevent future disasters monitor and quantify the extent of the flooding of the Fitzroy (Becker, 1994). The installed base of Weather Surveillance Radar River and was an invaluable information source for relief services 88 Doppler (WSR-88D) systems under the Modernization and and planners (EOSAT, 1993a). CHAPTER 3

NEW REMOTE SENSING ADVANCES

3.1 RECENT STATE·OF·THE·ART REMOTE was upgraded with regard to a sounding mission in 1998 and will SENSING TECHNOLOGY ADVANCES AND be totally updated in 2001 with NOAA 0, P, and Q to be flown up ONGOING RESEARCH to 2010. The NOAA a.m. series will continne up to 2003, at which time it will be replaced by a European series METOP up to 2015. Remote sensing platforms can vary from ground~based systems to Japan's MOS series and I-ERS series are for ocean monitoring aircraft to satellites. In most cases, truck-mounted ground-based and land observation. Noteworthy of mention are the active and aircraft systems are used in sensor development to verify microwave systems on board both J-ERS-I and ERS-l which design characteristics and to learn how the sensor responds to the provide all-weather observation capabilities. Japan's ADEOS-I target characteristics (Engman and Gurney, 1991) (e.g. system, launched in 1996, complemented ESA's ENVISAT and NEXRAD). Thus, the ultimate goal in most cases is to mount the NASA's EOS eIforts; a second ADEOS model was planned for sensor on free-flying satellites. 1999. The Russian Federation's METEOR-3M series is in prepa­ ration for flight in 2000. France's SPOT-4 was upgraded in 1997. USA's Landsat-6 and -7 are approved to cover np to 2000. The 3.1.1 Satellites NASA programme is based on a number of single- research missions, some having relevance to hydrology: TOPEXIPOSEI~ Up to now, there has been no launch of a hydrological satellite. DON, in cooperation with France, for altimetry, and TRMM, in However, many past, present and future meteorological and earth cooperation with Japan, for tropical rainfalL Finally, countries resource satellites have provided data and sensors for use in planning or developing satellite observing systems include India hydrology (see Appendix 11, Table 4). Notable among the meteor­ (IRS for earth resources), Brazil (MECB for data collection), ological satellites are the geostationary GOES series, GMS series, China (FY-1 for meteorology), Canada (RADARSAT for multi­ Meteosat series, INSAT series and GOMS; and the polar-orbiting purpose imagery), and the Russian Federation (OKEAN for ocean NOAA series, DMSP, UARS and METEOR series. Notable and RESURS for land observations). among the earth resources satellites are the polar-orbiting Landsat The EOS/Columbus Project started in 1998 and will series, SPOT series, J-ERS-1, MOS-1, ERS-1, METEOR· continue up to 2010; major space nations will work together to PRIRODA and ALMAZ. provide a network of earth orbiting platforms (most important will At present, satellite sensors offer good resolution in be the polar-orbiting platform in conjunction with a manned space space or in time, but not both simultaneously. Sensors with high station). In an effort to meet the many hydrological requirements spatial resolution on board polar-orbiting satellites (both meteoro­ as listed in Appendix II, Table 1, the hydrological community is logical such as DMSP and NOAA series and earth resources such seeking specific sensors which might be carried by EOS (see table as Landsat and SPOT) provide information on hydrologic in Appendix Ill). Furthermore, the TRMM results are needed to processes such as snowmelt, ice, land use and other hydrologic complement the much more complex BOS mission. parameters. Sensors with high temporal resolution on board Ormsby and Engman (1993) have proposed a hydrology geostationary satellites (meteorological, such as GOES and specific series of sensors to fill the gaps not covered by the current Meteosat) allow the monitoring of more dynamic processes such and planned systems. They have called this hypothetical platform as rainfall, runoff and floods. Furthermore, picture loops made HYDROSAT. Assuming that the instruments planned on board from the geostationary satellite images provide information on EOS will be in orbit by the year 2000, HYDROSAT wonld only moisture and temperature fluxes for forecasting, detecting and need to carry instruments to complement the existing scheduled tracking severe storms (Ormsby and Engman, 1993). sensors. Specifically, it is proposed that HYDROSAT combine Recent advances have been made regarding continuity passive microwave and a pointable SAR to meet the hydrology of coverage of geostationary satellites. EUMETSAT's Meteosat-6 needs. The chosen orbit would be a three-day repeat polar orbit. and Meteosat Transition Programme (MTP) are committed up to Although a geostationary orbit would be ideal for operational 2000-2002. A second generation of Meteosat, namely MSG, will hydrology, it would not provide adequate global coverage. be developed by ESA and EUMETSAT and launched in 2000 and will be flown by EUMETSAT up to the year 2012. USA's NOAA started a new series, GOES-NEXT in 1994 and will be used up to 3.1.2 Aircraft 2005. Japan's GMS-5 is improved flight model was launched in 1994. India's INSAT-ll system has been upgraded and is planned Although most aircraft sensors are used to test or calibrate satellite to be continued through the 1990s. The Russian Federation's sensors, there are numerous instances where aircraft remote GOMS was lannched in 1993. And, finally, China's FY-2 was sensing provides the only, or a more accurate source of, remotely launched in June 1997. sensed hydrological data. Recent plans for near-polar-orbiting satellites are given NASA has a high altitude research aircraft which makes as follows. ESA's ERS series was to be upgraded to focus on measurements of the microwave brightness temperature for deter­ ozone and atmospheric chemistry. A new series, ENVISAT, was to minations of rainfall rate profiles. replace ERS. ENVISAT-I was being prepared for launch in 1999; There have been a number of aircraft and only a few a second model is planned for 2004. USA's NOAA p.m. series satellite instruments that have demonstrated the feasibility of 32 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY measuring soil moisture by remote sensing. At present, microwave one of the most important sources of precipitation data for the radiometers capable of measuring soil moisture (ideally at L-band future. Approximately 120 of these modem Doppler radars have or 21 em) are available only on aircraft. The available soil moisture been installed across the US (Hudlow, 1988). The National gamma radiation techniques use low-flying aircraftsnil moisture Operational Hydrologic Remote Sensing Center is currently devel­ measurements. Furthermore, at present the- only operational soil oping and refining precipitation products from NEXRAD; which moisture techniques are the aircraft based gamma ray and passive will provide radar precipitation estimates far superior to estimates microwave approaches (Engman and Gurney, 199-1). currently available. Aerial photography has been used for many years for Ground-based· radars serve the purpose for calibration preliminary slln'eys of groundwater, and has recently been supple­ and validation of satellite sensors such as the TRMM, ERS:..1 and mented by satellite data (especially from Landsat and SPOT). J-ERS-I. A strong validation effort is planned for TRMM with Recently, airborne exploration for groundwater has been conducted several key ground sites to be instrumented with calibrated multi­ using electromagnetic prospecting sensors developed for the parameter rain radar. Key ground sites for the TRMM include mineral industry. This equipment has been used to map aquifers at preliminary ground truthing at Kwajalein in the Marshall Islands depths greater than 200 m (Paterson and Bosschart, 1987). and in Darwin, Australia, and a prototype established in the vicin­ Evapotranspiration is estimated using a combination of ity of Cape Canaveral, Florida (Simpson et ai., 1988). As for the ground:"based and remote sensing data. Low-level aircraft flights ERS-l scatterometer radar, calibration and' validation activities are used to collect reflected solar and emitted thermal radiation. have been carried out at European Centre for Medium-range 'This remote sensing data is combined with ground-based measure­ Weather Forecasts (ECMWF) complementary to the "Halten­ ments of wind, vapour pressure and incident solar radiation to banken" field campaign held off the coast of Norway (Stoffelen estimate evapotranspiration (Jackson et al., 1983). and: Anderson, 1993). In order to assess the J-ERS-l satellite esti­ Snow cover is readily identifiable from aerial photo­ mates ofrainfall (apreliminary product of the Global Precipitation graphy; The water equivalent of snowpacks can be measured with Climatology Project), they were compared with ground~based low-flying aircraft carrying sensitive gamma radiation detectors. observations from a high-density radar network over Japan Satellite imagery of snow cover and low-flying aircraft gamma ray (Janowiak, 1992). spectrometry have been combined to monitor short melt seasons. Furthermore, many microwave measurements of snow properties have been conducted with both truck and aircraft experiments; 3.2 POTENTIAL NEW APPLICATIONS IN passive microwave estimates of snow depth using aircraft are a OPERATIONAL HYDROLOGY viable alternative to poor resolution satellite sensors. Laser profIling techniques from airplanes can be used to While hydrological applications have exploited remotely sensed accurately measure the stage of surface water. Aircraft measure­ data from the visible through the thennal range, similar exploita­ ments of thermal infrared and microwave radiation data have been tion awaits the data from the microwave. Microwave data, used to measure surface water temperature. Aircraft microwave especially, will play an important role in hydrologic remote radiometers can be used to measure salinity and general mineral­ sensing applications in the immediate future. Notable are the ization of water and to detect some characteristics of vegetation microwave remote sensors on board the ERS-I, J-ERS-l, above the surface water. Special airborne radiometers/scanners RADARSAT, ALMAZ and the Space Shuttle-based systems have been used to estimate the chlorophyll content and suspended SIR-C and X-SAR. A number of active and passive microwave sediments in surface water. Finally, aircraft scanners and cameras sensors will be included within the EOS. In particular, satellite are used routinely to follow the fate of oil spills in surface waters, SAR promises the greatest long-tenn advances in surface-oriented map the aerial extent of flooding, wetland mapping and map hydrological monitoring because of its unique ability to penetrate erosion. cloud and to achieve high spatial resolution if required (Blyth, There are aircraft sensor systems used to test and cali­ 1993). brate sateI-lite s.ensor systems. Examples are as follows. The A reference to Appendix 11, Table 6 (after Blyth, 1993) ARMAR is a radar system (designed to fly on the NASA Ames highlights hydrological applications that can use readily available DC-8 and the high altitude ER-2 aircraft) developed to support the microwave data or data that will become available in the near design and development of the TRMM satellite rain mapping future, According to Blyth (1993), applications requiring simple radar system. Both the FRONTlERS radar system (designed by transfer functions are likely to achieve true operational status more the UK) and the GATE radar system have been used to calibrate rapidly than those having complex relationships between the and validate the DMSP SSM/I passive microwave imagery and remotely sensed signal and the hydrological variable. Microwaves algorithms used to estimate precipitation. at low frequencies can penetrate clouds, rain, dry snow and vege­ tation to provide infonnation on the underlying soil (Eagleman, 1975), while at high frequencies, precipitation can be measured 3,13 Ground-based methods and radars (Barrett et al., 1988b} Blyth (1993) indicates that because satelIile-based Although satellite data are most useful in providing information multifrequency SAR systems were unlikely to be available before on the spatial distribution of potential rainfall, ground-based radar the end of the century, applications requiring them have been has proven to be the most useful remote sensing technology for precluded from achieving operational status even though the estimating rainfall rates and locating regions of heavy rain. methodology may have been demonstrated using airborne SAR. Notable is NEXRAD, which is a state-of-the-art data acquisition Furthermore, Aluned el al. (1990) indicate that of the satellite SARs system aime.d primarily at the severe storm warning problem. It is planned for the near future, RADARSAT may be the most useful CHAPTER 3 - NEW REMOTE SENSING ADVANCES 33 for operational purposes using the optional Scansar mode which NIMBUS-7 and the DMSP SSM/I. However, the temporal will provide complete coverage of the Northern Hemisphere with a sampling of rainfall distribution will be improved through the three-day return period. Zebker et al. (1992) describe an TRMM so as to enable the true potential of this type of sensor to interferometric airborne SAR being developed with the aim of be achieved for input to climate models. providing topographic contouring at 2-3 III accuracy; this has also According to Engman and Gurney (1991) some of the been demonstrated to be feasible from space. new space initiatives will provide the maximum amount and type While microwave remote sensing alone may not be the of remote sensing data needed for further advances in estimating best method of measuring hydrologic variables, when used in the global evapotranspiration fluxes. conjunction with data from other sensors it may become invalu­ The Advance Techniques Unit of the lRSA (1992) is able. "Microwave sensors will enable temporal changes in surface currently active in outlining new and potential applications of features to be studied in a way which has previously been imposs­ remote sensing. Among the progress are included the following: 1) ible, but the full benefit of remote sensing is likely to come from the preparation and implementation of a new programme in canopy the integration of knowledge gained from a range of sensor types" chemistry and canopy structure by means of optical spectrometry (Blyth, 1993). Evans and Smith (1991) found that a combination (splitting the visible and infrared light into high spectral resolution of airborne SAR and visible data (e.g. Landsat TM) enables the channels); 2) intensive work on the first set of ERS-l SAR images backscattering effects of vegetation and soils to be more easily over Europe and Africa test sites to include support in monitoring separated than with SAR alone. According to Blyth (1993), the ofenvironmental damage to the North Sea and in monitoring of the most useful application for satellite SAR is river and coastal flood deforestation in regions, including the tropical belt; 3) use intensive mapping. Once again, no single sensor can fully assess the loca­ methods (interferometric SAR data correlation) for the estimation tion, severity or scope of impact of most natural disasters (because of forest biomass in flat and mountainous regions; 4) synergy of clouds, smoke and rain commonly interfere with the reception of high-resolution airborne multiwsensor data with spaceborne SAR many sensors). Both EOSAT (l993b) and RADARSAT data sets from ERS-I and J-ERS-l for improved information on International Inc. (RSI) (COSPAR, 1993) have illustrated the ecological units at different scales (complemented by the evalua­ power of merging SAR data from ERS-I radar satellite with tion of the potential advanced optical spectrometer methods); 5) multispectral sensor data from an optical satellite (from the use of the first geophysical radar signal processor for land applica­ Landsat TM and SPOT, respectively) to study the 14 Iuly 1993 tions; and 6) using the airborne time resolved LIDAR fluorosensor floods in St Louis, MO. The SAR radar can penetrate haze, light data to monitor the marine environment. rain and clouds, and is effective for terrain analysis, while the Thiao et al. (1993) have illustrated the use of GOES multispectral sensor (Landsat TM and SPOT) allows identification water vapour imagery in predicting heavy precipitation. They and classification of surface features with a high degree of accu­ concluded that flash flood producing MeSs ~e a multi-scale and racy. Similarly; the authorities in France, Germany, Belgium and concatenating event. Therefore, a better understanding of the the Netherlands relied upon ERS-I SAR and Landsat TM data multi-scale nature of heavy precipitation will be attainable with during and after the late December 1993 torrential rain produced the merging of the WSR-88D radar data with the various GOES floods, to assess situations and get reScue, clean up and recon­ multi-spectral images, GOES-derived sounding data and SSMII struction priorities (EOSAT, 1994b). data (Thiao et aI., 1993). Satellites carrying passive microwave sensors offer the In an effort to optimize on the utilization of remotely best opportunity of obtaining repetitive observations of snow sensed observations in hydrology, a number of programmes have extent and snow depth. Chang et al. (1987, 1990) used the SMMR been organized to study hydrological processes in conjunction data from NIMBUS-7 to produce a snow extent map of the whole with remote sensing. Examples of such programmes include the Northern Hemisphere. Global Energy and Water Cycle Experiment (GEWEX) and the The most useful sensors for rainfall estimation have Tropical Ocean Global Atmosphere (TOGA) programme. TRMM been the multifrequency scanning radiometers on board results are needed for conducting GEWEX. CHAPTER 4

PROBLEMS IN APPLICATIONS OF REMOTE SENSING IN OPERATIONAL HYDROLOGY

4.1 CURRENT LIMITATIONS OF OPERATIONAL transmission systems, computer hardware and software, and a UTILIZATION OF REMOTE SENSING highly developed set of procedures for data collection and quality assurance (UNESCOIWMOIICSU, 1993). However, organiza­ Most of the problems limiting the operational application of tional problems hinder the operational utilization of remote remote sensing in hydrology are either financial or organizational sensing in hydrology. in natme. Financial problems result in limitations of funds neces­ sary to: 1) further the technology of remote sensing for hydrologic applications; 2) acquire, process, distribute and utilize remote 4.1.2 Status in developing countries sensing. data; and- 3) train and maintain personnel. Organizational problems result in: 1) a gap between the scientist (collecting the Many developing countries do not have the financial resources data) and the decision maker (using the data); 2) a lack of an inter­ andlor the organizational capabilities to benefit from the available disciplinary approach; 3) a lack of inter-agency collaboration; and remote sensing technology. There is a competition for scarce 4) barriers between national government agencies and international financial resources in some countries which are struggling to organizations (UNESCOIWMOI1CSU, 1993). provide the basic amenities of life to their people In general, there is a problem of the lack of awareness of (UNESCOIWMOI1CSU, 1993). According to UNESCOIWMOI remote sensing among managers and decision makers. According to ICSU (1993), many developing countries do not have accessibility Vaughan (1994), even in Western Europe, people in high posts often to new remote sensing technologies because of their high initial do not appreciate the usefulness ofmaps and spatial data. According cost, lack of high-level skills (needed to process and apply them) to Konecny and Vetrella (1994), the organizational problems stem and lack of personnel training capabilities. According to WMO from the fact that government and users have an organizational struc­ (1993a), the "reason" is the lack of efficient and reliable informa­ ture that does not take into account the uses ofremote sensing. Thus, tion systems such as global telecommunications and the related an organizational change needs to be introduced. For a long time, the use of electronic mail, electronic bulletin boards and computer-to­ drive in remote sensing has come from the providers of data rather computer file transfers. Other problems which contribute to the than the users. Although this situation is slowly changing, it is still inability to benefit from the available remote sensing technology dependent on the support of governments. Furthermore, the funding include language difficulties, insufficient dissemination of ofinformation technology and data is a problem. UNESCO and WMO publications to end-users (UNESCOIWMOI According to Cndlip (1994), the absence of high qnality ICSU, 1993) and organizational problems. An example of the last training in the management of information is holding up the devel­ problem is that operational hydrology is carried out in several opment of remote sensing. Therefore, it can be difficult for agencies with little or no coordination between agencies. decision makers to introduce remote sensing into their For instance, review of the remote sensing training programmes. Furthermore, remote sensing will often be just a needs of Africa suggests that the amount of remote sensing is small part of a project manager's job. However, if future managers extremely variable among the different countries, and that there can be trained in the use of Management Information Systems has been a variable amount of cooperation with European organi­ (MIS), then this will automatically generate the demand for zations. In the past, remote sensing has been presented too much remote sensing systems· and stimulate the value-added industry as a tool while operational applications have not been stressed (Cndlip, 1994). As Konecny and Vetrella (1994) and Cudlip sufficiently. The decision makers are also often not aware of the (1994) point out, remote sensing is not a discipline but rather it is economic usefulness of remote sensing and they need to be a technique or tool which produces information for applications. targeted in training programmes just as much as technicians. Cudlip (1994) points out that although the number of Furthermore. the cost of imagery inhibits the wider use of remote people being trained in remote sensing is about right, the field of sensing (Kabbaj and Mehrez, 1994). training does not always match the current shortages. There is a definite shortage of trained technicians for routine data-processing tasks. In particular, the staffing of satellite receiving stations is 4.2 PROBLEM AREAS PREVENTING THE proving difficult. Although several stations have been built in OPERATIONAL UTILIZATION OF REMOTE developing countries, they are not being used due to the lack of SENSING IN HYDROLOGY suitably trained personnel. Thus, a special remote sensing training effort is necessary in developing countries (Demerliac, 1994). 4.2.1 Hydrological variables

Although hydrological studies are still needed to improve the 4.1.1 Status in developedconntries understanding of the physical hydrological processes, operational or near-operational applications of remote sensing techniques to Most developed countries have the necessary resources and capa­ measure hydrological variables have been successful in varying bilities to collect, store, analyse, distribute and operationally apply degrees. According to UNESCOIWMOIICSU (1993), there is a remotely sensed data for hydrological purposes. Available to these clear and serious need for improved understanding of hydrologic countries a!e the latest in remote sensing technology, data processes everywhere in the world but particularly in nations of the CHAPTER 4 - PROBLEMS IN APPLICATIONS OF REMOTE SENSING IN OPERATIONAL HYDROLOGY 35

"South". Furthennore, the area concerning groundwater in both the rainfall) is available only at one point (raingauge) within an area unsaturated and saturated zones still requires a great deal of of, say, 1 000 km2 while the hydrological model works on a basis intensive studies. The application of remote sensing techniques to of Landsat pixels (30 x 30 m), then in such a case a simple date has been more limited for the measurement of both lumped system model would do the job as well. groundwater and evaporation. However. operational applications of "Model input" data can also be estimated with the aid of remote sensing techniques have been most successful for remote sensing. For instance, for the computation of snowmelt, the measurement of precipitation, soil moisture, snow and basin snow cover area is a relevant model input; for the estimation ofevapo­ characteristics. Operational applications of remote sensing transpiration, radiation and temperature values are relevant model techniques to measure ice and surface water characteristics have inputs; and for the estimation of rainfall (as input to rainfall-runoff also been progressing. models), cloud top temperature values are relevant model inputs. In the modelling of hydrological processes there is usually never a direct mathematical relationship between the elec­ 4.2.2 Hydrological models tromagnetic signal and the hydrological variable. Usually, a more complex transformation of the electromagnetic signals is neces­ Although the application of remote sensing data in hydrology sary,often combined with terrestrial information in order to offers numerous advantages over conventional data, the advent of estimate hydrological variables. Sometimes data from different remote sensing caused some confusion among the modellers since channels or from different platforms and sensors have to be used the remotely sensed data could not be used directly in existing in combination with terrestrial data. models. The reasons were that: I) the input data consist of electro­ As in the case ofconventional hydrological modelling prac­ magnetic information instead of hydro-meteorological data; 2) the tices, the structure of the hydrological model using remote sensing resolution in time and space is sometimes higher, sometimes lower data is usually a function of the scale ofthe system under considera­ than necessary due to the available sensors; and 3) there was no tion. In the case ofmicroscale « 100 km2) hydrological models, the information available in order to transform the electromagnetic structure ofthe model should subdivide the hydrological systems into signals into hydrological information with the aid of a model area elements ofpixel size, ifthe available spatial resolution ofremote (Schultz, 1993). A comprehensive study by Peck et al. (1981) sensing data is to be fully utilized. It is advisable to combine the showed that most existing hydrological models were not suitable remote sensing information with other data derived from Digital for use of remotely sensed data. However, during the last two Elevation Models (DEM), for instance. Inthe case ofmesoscale (100 decades one can observe the development of hydrological models to 10 000 km2) hydrological models, file number of area elements has designed for acceptance and processing of remotely sensed data to be reduced considerably as compared to the microscale models. (Schultz, 1988). "The structure of hydrological models using There are many ways to aggregate small are~ elements, such as remote sensing data depends to a certain degree on the scale of the Landsat pixels or grid cells of a DEM to larger units. Currently there system under consideration" (Schultz, 1993). are many operational models available for the microscale and the The water balance equation contains several terms mesoscale and some of which use remote sensing information are which can be estimated with the aid of remotely sensed data: under development. However, no physically satisfactory models are available for the macroscale (> 1M km2) except for more lumped P(t) - E(t) - Q(t) = dS(t)/dt (10) system approach models (Schultz, 1993). where P is precipitation, E is evapotranspiration, Q is runoff, S is There is an increasing interest particularly in macroscale water storage, and t is time (Schultz, 1993). There are various modelling. There is a need to couple atmospheric General Circulation techniques available which try to compute P or E with the aid of Models (GCMs) with ocean models and hydrological models for the remote sensing data. The runoff term Q is particularly difficult to simulation ofthe global water cycle. Since the OCMs work on a grid estimate using remote sensing data because it is basically a meas­ basis with grid sizes of 500 to 200 km in length, there is at present urement at a point on the stream. The tenn dSldt representing the some pressure on hydrologists to develop continental hydrological change of water storage on and in the earth contains several models creating information relevant for areas represented by such hydrological variables: infiltration, soil moisture, groundwater grid cells. The GEWEX organized by WMO and ICSU is particularly recharge, snow and ice. For these variables there are also models interested in such combined atmospheric-ocean-hydrology models. In available to compute their values with the aid of remote sensing this- context, a OEWEX Continental Scale International Project data. Furthermore, it is important to distinguish between the esti­ (OCIP) was initiated, in the framework of which such a combined mation of model parameters on the basis of remote sensing data, model for the total area ofthe Mississippi River catchment (3 M km2) and the estimation of model input with the aid of remote sensing shall be developed. In this modelling effort itis intended to use remote information (Schultz, 1993). sensing information in order to estimate meteorological, oceano­ According Lo Schultz (1993), good hydrological models graphic and hydrological parameters at continental and global scales have a structure which does not change from region to region. (Schultz, 1993). Only the "model parameters" are different in various regions and have to be recalibrated for each region separately. In the case of physically-based distributed system hydrological models, many 4.2.3 Remote sensing technology model parameters depend on the characteristics of the hydrologi­ cal system, e.g. the basin characteristics, aquifer characteristics 4.2.3.1 Limitations ofthe state-of-the-art and river reach characteristics. There must be a correspondence between the resolution in space of the distributed system type According to WMO (l993a), in order for the remote sensing data model and the resolution of the input data. If the model input (e.g. to have a real impact on the application, data quality has to be 36 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

specified (and fulfilled) at the same time as data requirement. Data 4.2.4.2 Data acquisition and distribution capabilities of quality has to be specified in terms ofhorizontal resolution, verti­ countries cal resolution (if applicable), frequency of the observation and accuracy. Because quality parameters are different for the VariOllS The World Weather Watch (WWW) has become the basic scales of application, two figures are generally quoted, for global programme of WMO and is of vital importance to the success of scale, G, and regional scale, R, respectively. Appendix IV, Table 1 other WMO programmes (WMO, 1992b). The Global Observing gives the Executive Council Panel of Experts on Satellites System (GOS) is one of the main elements of the WWW. The (ECSAT) list of satellite data requirements for upper air and for GaS is divided into the surface-based portion and the space-based surface. respectively. This table also lists the' user category of the portion. The environmental satellites (near-polar-orbiting and requirement. Furthermore, these satellite requirements are geostationary) constitute the space-based subsystem of the GaS, expressed in tenus of geophysical parameters, which are what the with the major goal of augmenting the information provided by end-user needs. These requirements are different from the required the surface-based subsystem (conventional data network) to capabilities of near-polar-orbiting satellites and of geostationary complete the global coverage (WMO, 1992b). The two types of satellites (given in Appendix IV, Tables 2 and 3, respectively), environmental satellites are complementary. The ground segment which are much closer to the original measurements used to derive of the space-based subsystem of the GOS is composed of the the geophysical parameters. satellite-data ground receiving stations and these have two major functions. The first function is to provide the reception of the 4.2.3.2 Consistency ofdata and needfor calibration signals from satellites containing qualitative and quantitative ofnew sensors information, including observations from data collection platfonns and other similar systems (e.g. for the ARGOS system). The Given that the ECSAT requirements of Appendix IV, Tables 2 and second function is to process, format, display and distribute the 3 are what WMO reasonably expects from satellites (and thus information received, either by direct broadcast via the satellites satellite-systems planners) in the near- and medium-term, WMO themselves, or over the GTS in pictorial or alphanumeric form to (1993a) has converted these data requirements (expressed in terms meet the global, regional and national requirements ofthe WWW. of geophysical parameters in Appendix IV, Table 1) into mission According to WMO (1992b), there are approximately requirements, as expressed in terms of observing techniques. The 700 established satellite receivers operated by WMO Members user requirements specified in terms of observing missions are throughout the world. The large number of receivers is a result of grouped as follows: 1) multi-purpose- imagery; 2) atmospheric the widely varying geographical location ofthe countries and their sounding; 3) sea-surface observation; 4) radiation budget and meteorological regimes. Furthermore, another factor contributing atmospheric chemistry; 5) high-resolution imagery; and 6) to this large number of receivers is the rapid technological telecommunications. Each mission has specific objectives, speci­ advances in satellites and in the equipment available for receiving, fied observation requirements, as well as specification of typical processing and presenting the quantitative and qualitative satellite instrumentation/sensors either operational at present or proposed data. in order to fulfil the mission; these are given in Appendix IV, A survey couducted in 1990 by WMO (I992b) through Table 4. The main objectives of the multi-purpose imagery, national Meteorological Services revealed information concerning supported by the atmospheric sounding, the earth radiation budget the status of satellite receiving equipment within WMO Regions. and the high-resolution imagery missions, are mainly driven by There are six Regional Associations (RAs): RA I (Africa), RA II operational hydrology. The corresponding observation require­ (Asia), RA III (South America), RA IV (North and Ceutral ments allow for the applications (objectives) to be implemented. America), RA V (South-West Pacific) and RA VI (Europe). Four Furthermore, the telecommunications mission supports the global categories of satellite receiving equipment were surveyed: low­ d-ata acquisition, direct broadcast of data and distribution of data resolution polar-orbit data (APT); high-resolution polar-orbit data from all geostationary and near-polar-orbiting environmental (HRPT); low-resolution geostationary data (WEFAX); and high­ satellites of interest for WMO programmes. In Appendix II, resolution geostationary data (RR). The goal for the WMO Table 3 is a summary table highlighting the relationships amongst Regions is that each WMO Member should be equipped with at the required geophysical parameters, observing technologies and least one polar-orbiting satellite data receiver and one geostation­ candidate sensors. ary satellite data receiver. However, the Voluntary Cooperation Programme of WMO coordinates the needs of WMO Members from developing countries with available resources from WMO 4.2.4 Availability of data Members from developed countries. A fact highlighted in the survey is the lack of satellite ground receiving equipment registra­ 4.2.4.1 Geographical or climatological regions tion with the national telecommunication administrations and the where remote sensing data are most applicable International Telecommunication Union's (lTV) International Frequency Registration Board (lFRB). Registration is advised in Because most techniques attempt to measure surface features, order to protect and preserve the present allocation. Furthermore, remote sensing works best in areas with little to obscure the surface the present allocation is limited and new technologies are (Rango, 1994a). Remote sensing tends to work best in arid and constantly vying for frequency allocation. semi-arid regions because vegetation is the most common feature Results of the survey indicated the geographical distri­ obscuring the surface. Furthermore, in the area of snow hydrology, bution of equipment as shown in Table 3 (WMO, 1992b). snow mapping- applications also tend to work best where there is Since it is recognized that the HRPT imagery is most little vegetation to obscure the snowpack (Rango, 1994a). useful to operational meteorological and hydrological forecasting, CHAPTER 4 - PROBLEMS IN APPLICATIONS OF REMOTE SENSING IN OPERATIONAL HYDROLOGY 37

Table 3 - Geographical distribution of equipment expires, only national meteorological centres connected to the GTS will continue to receive data free of charge. WMO Regional Association Members II III IV V VI 4.2.6 Computer data-processing capability Total 55 30 13 25 16 37 Per cent of Members Impressive developments have occurred in the field of image equipped with: handling systems (WMO, 1993a). The traditional hardcopy­ APT receivers 80 80 85 76 38 57 oriented approach has been replaced by software. HRPT receivers 13 27 15 20 31 46 Hardware/software systems have been developed which allow for WEFAX receivers 46 25 12 25 37 46 image enhancement, multi-channel image handling, multi-tempo­ ral image sequencing, geofrequency in appropriate projections and HR receivers 11 15 4 5 21 28 merging of images overlaid with other information (radar images, bulletins, maps, etc.) (WMO, 1993a). Geographical Information increasing the coverage is a first priority to enhance the implemen­ Systems are used to supply all necessary background information, tation of the WWW (WMO, 1992b). It may be noted that RA VI including DEMs. Furthermore, modern technology provides (Europe) has the best percentage of Members equipped with high­ reasonably inexpensive workstations able to support these applica­ resolution imagery receivers for both the polar-orbiting (HRPT) tions. WMO and national and international aid-to-development and the geostationary (RR) satellites. programmes are active in supporting developing countries for the The availability and access of remote sensing data acquisition and maintenance of processing facilities. Furthermore, needed for operational hydrological applications is difficult except the National Resources Institute (NRI) in the UK provides a when these applications arc run at large meteorological services comprehensive service for the interpretation and analysis of which have responsibilities in those fields (WMO, 1993a). High­ remotely sensed imagery, from both satellite and airborne plat­ rate digital acquisition stations are required for near-real-time forms. The latest technology including VAX, Sun and PC-based applications, otherwise data are obtained from archives, and thus image processors in conjunction with GIS software to address a only for delayed-time applications. High geometrical resolutions variety of environmental issues is used. are required for operational hydrological applications, and are available from operational satellites such as LANDSAT and SPOT (using the VISINIR range). However, the technical complexity of 4.2.7 Data presentation/publication/dissemination acquisition and distribution of these images seriously limits the access to these data beyond applications such as pilot projects. In order to communicate between the scientist collecting and Furthermore, passive microwave radiometry, which could be processing the data, and the decision maker using the data, the extremely valuable in hydrology, is not available with the appro­ data must be published, presented and otherwise properly dissemi­ priate resolution (WMO, 1993a). nated. WMO (1993a) points out that there is "difficulty in identifying true end-users from others whose primary interest is to transform information to be handed over, with added value, to 4.2.5 Cost of data somebody who, in reality, has not been involved which makes it difficult to get authoritative long-term validation of applications The financial aspects of remote sensing data include the direct cost theoretically demonstrated". of acquiring data, cost trends and policy, and thus the capabilities of countries to purchase the data. In the developed countries, funding is provided mainly by national institutions and interna­ 4.2.8 Human resources and training capability tional grants. However, there is a need for better coordination in order to make the best use of the funds available UNESCOIWMO/ICSU (1993) has identified a number of prob· (UNESCOIWMO/ICSU, 1993). In many developing countries, lems affecting the human resources and training capabilities of there is a severe problem regarding financial constraints and it has various countries. They are: 1) inadequate number and quality of become worse over the last 10 years (UNESCOIWMO/ICSU, qualified hydrologists and technicians; 2) the "brain drain" affect­ 1993). Sufficient funds are not available to support either hydro­ ing trained staff is too high; 3) technology is still unbalanced with logical research or adequate data collection. Thus, there is a need skills of the users; and 4) UNESCO and WMO publications do not for support of sustained hydrological research and data collection, reach a sufficient number of end-users. However, particularly in the humid, semi-arid and arid parts of the tropical UNESCOIWMO/ICSU points ont that the developed world has world (UNESCOIWMO/ICSU, 1993). According to FAO (1989), tried to make adjustments in the educational system to adapt to the the developing countries stand to benefit most from this remote rapid and unpredictable changes taking place; this is reflected in sensing teclmology. They urgently need reliable resource invento­ the increasing variety and contents of training courses and ries and maps to provide a basis for more rapid, geographically improved training methods. Furthermore, UNESCO/WMO/ICSU balanced and ecologically sound economic development (FAD, points out that in some developing countries, education is still 1989). According to WMO (1993a), most experimenters have conventional and emphasizes knowledge rather than developing received data free of charge, in connection with an Announcement skills and professional attitudes. There is an inadequacy in staff of Opportunity process, and most of them are financially and teaching facilities and thus poor results are attained from supported by their national space agencies. When this support education and training courses. However, the training 38 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

opportunities and documents resulting from UNESCO (IHP)- and was generally accepted in Europe that basic' education and WMO CORP) programmes have greatly contributed- to the coun­ research should be paid by the State. However, the trend now is tries' human resources and training capabilities. Furthermore, there towards education training and infonnation in specialized fields in is a need for an educational network which would put basic material remote sensing as the responsibility of the economic agents, Le. and dataproducts at the disposal of educational entities (Fea, 1994). users, industrial and service companies (Demerliac, 1994). According to Vaughan (1994), not only is the According to WMO (1995), the comprehensive strategic geographic and thematic demand for remote sensing education plan for systematically improving the utilization of satellite data and training dispersed, but so is its present and future provision. endorsed by the WMO Executive Council has started to work well The requirements and stail.dards for remote sensitig education and and is still valid for ongoing implementation over the next five to training vary, qualifications are not universally recognized and few 10 years. The strength of the plan is that it provides a coordinated training programmes exist. However, the European Association of approach to making more effective use of the large number of Remote Sensing Laboratories (EARSeL), which has been involved satellite education and training activities. These activities have in education and training programmes for many years, has so far been progressing independently at different rates within a large taken the main initiatives on the European scale by identifying number of different organizations. Further, WMO (1995) reports gaps in the existing provision, analysing needs and developing that cooperation, the poolings of existing resources and a commit­ training materials (Konecny and Vertrella, 1994). Cudlip (1994) ment to embrace the overall strategic goals of the plan was starting recognized that training courses tend to concentrate too heavily on to establish the critical mass of expertise and supporting infra­ technology and image processing. Rather, there needs to be an structure to make real progress in improving the application of improvement in the integration of remote sensing as a tool into satellite data in meteorology and hydrology. GISs and the various application disciplines. The aim should be to Two of the key issues for the WMO Congress in 1995 raise the general level of awareness of remote sensing technology (WMO, 1995)were: I} to actively support the strengthening of the across the disciplines and to provide specialist training for scien­ Specialized Satellite Application Training Centres at the Regional tists and engineers (Mather, 1994). Furthennore, a prerequisite for MeteorologiCal Training Centres (RMTCs) at Nairobi, Kenya and any proper training is the need to firstly instruct the instructor Niamey, Niger in association with the sponsoring satellite operator (Fea,. 1994} (EUMETSAT); and 2) to actively promote WMO's satellite educa­ The assessment of the costfbenefit aspects of remote tion and training strategy in RAs II, III, IV and V and, in sensing education is difficult because the return on investment is particular, to try to obtain sponsorship for the establishment of at not easy to evaluate (Torlegard, 1994). The cost of remote sensing least one specialized satellite applications training RMTC in each images is a problem for many universities. For instance, so far it of these Regions. CHAPTERS A STRATEGY FORACCEPTANCE OF REMOTE SENSING IN OPERATIONAL HYDROLOGY

In order to optimize the use ofremote sensing on an operational basis, 5.2 DEVELOPMENT OR MODIFICATION OF we must identify the hydrological needs regarding data collection, HYDROLOGICAL MODELS BASED ON modelling and expected benefits. Furthennore. major hydrological REMOTE SENSING INFORMATION services ofagiven country must be made the focal point of this strategy. WMO (l993a) highlights that many applications of Mathematical hydrological models started from simple linear lumped satellite data ended when a paper was published or with the report deterministic models and have developed towards more complex of a successful pilot project. However, the application never models containing non~linear and stochastic components as well as arrives for the end-user to utilize operationally. high resolution in space represented by distributed system models (Scholtz, 1993). Hydrological models based omemote sensing should use a distributed system model with a high resolution in space in which 5.1 PAST STUDIES OF OPERATIONAL the basin area will be subdivided into a large number ofarea elements. HYDROLOGICAL OBSERVATIONAL NEEDS Furthermore, we canexpect that future models will at certain times and AND EXISTING AND FUTURE SATELLITES for certain applications require a high resolution in time (e.g. one hour time increments). In this case it will be necessary that a transfOlmation There is a definite need to extend the use of satellites for provision of ofmodel input into model output be achieved by the model in such a hydrological data sensors for the purpose ofhydrological data acqui­ way that for each time increment and each area element in the system, sition (UNESCOIWMOIICSU, 1993). There is alsn a need to the relevant mathematical computations have to be camed out. implement the World Hydrological Cycle Observing System Furthermore, the model input will not consist ofhydrometeorological (WHYCOS) to facilitate access to global data and snpport hydrolog­ data (e.g. rainfall, snowmelt, etc.) but rather in the form of electro­ ical services in need (Rodda et aI., 1993), and to improve forecasting magnetic information obtained from various sensors. As a result, the system effectiveness by incorporating weather radar information. mathematical structure of a model being able to execute all these The wide range of operational hydrological observational numerous computations will be rather complex. Therefore, the model requirements and mission requirements presentedin4.2.3 above can only builder is faced with a problem to reduce the tremendous amount of bemetby a greatnumber ofsatellites and instnunents and thus aneed for information to the necessary quantities, e.g. by aggregation of similar new technological developments. Because oflimited resources, ECSAT area elements. Furthermore, the historical structures of mathematical has developed a "top-level statement" which attempts to indicate WMO models need new dimensions. Since hydrological distributed models priorities and suggested a stnltegy. ECSAT (WMO, 1993a) has made the consider the areal diversity of the hydrological process, these models following recommendations: 1) the continuation ofthe present/core oper­ are preferred over lumped models if remotely sensed data are used. ational mission, including those improvements under way; 2) the Although almost all hydrological processes are non-linear, many such immediate addition, to the baseline, of observations which have proven processes are not highly non-linear, whichoften allows the mathemat­ sufficiently mature for transition to an operational status; and 3) to under­ ical representation of the process by a linear model. Finally, in the take technological developments on the important data requirements for process of the development ofnew hydrological models using remote WMO, which cannot be met with present operational or demonstrated sensing information, it is advisable to test which model components instruments. The core mission for the space-based subsystem of the have to be deterministic and which model components can be repre­ Global Observing Systemis based primarily on molti-purpose and high­ sented by a stochastic approach. resolution imagery and atmospheric sounding plus related subsystems for "There is an interdependency between the structure of the data transmission to the ground together with the ancillary information hydrological models and the hydrorneteorological information avail­ needed for data processing. The most important requirements for imme­ able to be used in such models. Therefore, it is necessary - in order diate addition are: 1) to improve monitoring and understanding of to make a statement on future developments - to look at both; rnath~ atmospheric convection (from geostationary orbits); 2) to improve the ematical structures ofpresently available models using remote sensing performance of atmospheric sounding through high spectral resolution data as well as new remote sensing platforms and sensors which we lR spectroscopy (from ncar-polar orbits); 3) to extend the molti-purpose can expect in the near future, from which new types ofhydrologically imagery mission into the microwave spectral domain; 4) to have access relevant data can be obtained" (Schultz, 1993). Satellite imagery to high-resolution active microwave images (from SAR); 5) to improve allows recalibration of model parameters according to the relevant the present shortwave and longwave instnnnentation; and 6) to study the seasons of the year. Additionally, future remote sensing information chemical processes in the troposphere. As for high-priority longer-term opens even wider and more promising fields of hydrologic activity requirements, these are as follows: 1) high-vertical resolution frequent within the framework of global research programmes such as IGBP soundings using IR spectrometry (from geostationary-orbits); 2) wind or GEWEX (Scholtz, 1988). profiling (from near-polar-orbits); 3) use a backscatter LIDAR-based imagery to drastically improve the accuracy of soundings; 4) use a low­ inclination orbit wilh a rain-radar on board to improve global 5.3 DEMONSTRATION OF SIGNIFICANT measurements of precipitation rates in lhe tropics (namely the TRMM TECHNICAL AND ECONOMIC BENEFITS mission launched by the US andJapan in 1997); and 5) use DIALLlDAR and very-high-resolution spectroscopy in the 02A-band for measurement The development of sound basic data is a fundamental element in ofsUIface air pressure. understanding hydrological systems (UNESCOIWMOIICSU, 40 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

1993). Ftrrthennore, the information/data gathered through remote and assimilation of satellite-focused education and sensing is vital to the resolution ofmany serious conventional data training within the framework of the current local education and problems including the following. Hydrological data is oftentimes training organization should be considered. scarce and the quality may be unreliable. The data are not usually According to Torlegard (1994), remote sensing technol­ collected with sufficient attention to protocol and are not stored in ogy developments and education should be more demand driven such a way as to facilitate retrieval by subsequent users in the future, but it is difficult for clients/users to formulate (UNESCOfWMO/ICSU, 1993). The degradation in hydrological requirements. This is an interactive process. The information networks is due to insufficient funds to maintain and repair requirements of the users have to be established rather than the damaged stations. Whereas remote sensing from satellites, tele­ remote sensing requirements. Users should not have to be experts metry of data, weather radar, solid-state data loggers which can be in the field of remote sensing (Cudlip, 1994). When designing downloaded directly into computers, and state-of-the-art computer training programmes, the needs of the clistomers should be care­ software can provide opportunities for acquiring information more fully considered. For instance, courses organized in institutions cheaply and efficiently, and to higher standards of quality and universities in Europe may not correlate with the level of (UNESCOfWMO/ICSU, 1993). activity within, for example. individual African countries. What Remote sensing plays an important role in collecting up­ may be needed instead are well targeted, well prepared, in situ to-date information and data in both the spatial and temporal courses. These need to be designed to suit both the equipment domain. According to UNESCO/WMO/ICSU (1993), the most available and the particular problems of the users (Vaughan, severe resource management issues exist under extreme condi­ 1994). On the other hand, sometimes it is necessary to teach train­ tions: in time, during floods and droughts; and in space, in arid, ing courses in developed countries. At that time, the developing semi-arid and tropical areas and in coastal areas. The use of nations must make sure to send the end-users who need the hands­ remote sensing techniques is vital in the areal estimation espe­ on experience and not use the training course as a reward for cially of precipitation and soil moisture. Remote sensing seniority. The attendee must truly be a user of the data. techniques enhance the seasonal forecasting capability and aid in Mather (1994) suggests that remote sensing is best the development ofstorm surge forecasting, drought and low flow taught not as a stand-alone technology but as a part of a wide forecasting and improve the risk management practices. perspective which includes social, environmental, economic and political aspects of environmental management, as well as GIS techniques. Remote sensing should be established as a basic 5.4 TARGETING MAJOR HYDROLOGICAL component of GISs. Thus, the concept of geo-managementcan be SERVICES regarded as a discipline, and GIS the tool with which it can be perfonned (Cudlip, 1994). Therefore, new management concepts In order to bridge the gap between scientists and decision makers, need to be introduced into management schools. Educators need an important strategy is to integrate the remote sensing informa­ to address the problem of information management. Furthermore, tion through a national archive to support the decision-making the design of remote sensing tools (e.g. GIS) needs to be user­ process and to facilitate the education of decision makers driven and not produced by the manufacturers. The skills and the (UNESCOfWMO/ICSU, 1993). Another important strategy is to tools must fit the problems and not vice versa (Cudlip, 1994). provide effective mechanisms for coordination among decision A workshop held in Strasbourg in 1993 by EURISYon making agencies. An additional strategy is to establish joint "Training in Remote Sensing: A European Proposal?" resulted in programmes among the agencies for publications, meetings and two proposals. The first proposal was for the Integration of training (UNESCOfWMO/ICSU, 1993). Furthermore, there needs Environmental Remote Sensing in European Secondary Schools. to be an awareness of the needs for interdisciplinary collaboration The plan is to establish five to seven schools in a number of coun­ in hydrology and water resources so as to avoid overlaps and tries and to provide training materials including distance learning, duplication of activities between agencies and disciplines. To where instructors may transmit training modules over telephone accomplish this strategy requires overcoming financial constraints lines to remote learning locations. The second was related to a and modification of current organizational and institutional European Network of Tertiary Education Centres in Remote arrangements at both national and international levels. Sensing and was proposed by EARSeL. The project would In order to improve the use of satellite data particularly emphasize needs and standards, and would involve a network of in developing countries, WMO (1993a) recommends the follow­ about 10 universities to develop a curriculum and to establish a ing strategies. First, well designed pilot projects should be pilot system. The pilot project would concentrate on applications promoted, involving the end-user from the onset since it is the as opposed to hardware and software, teaching a common stan­ end-user that will have to ensure the continuation of the activity if dard of understanding across disciplines (Mather, 1994). Mather the pilot project is successful. Second, improved information serv­ (1994) suggests that in the teaching of remote sensing in Europe, ices should be developed to allow users, particularly from the role of national and European professional societies (EURISY. developing countries, to be able to survey existing information and EARSeL, British/French/ItalianiSpanish Remote Sensing and know where it is and how to access it. Data from different sensors Photogrammetry Societies) should be considered. Although it is or from more than one satellite can be combined with a GIS to not essential that everyone be trained in remote sensing, the prin­ produce new products and applications. Third, education and ciples should be included as part of general education at school training should reach the true end-users instead of stopping at an levels so that in the future, decision makers in particular have intermediate level (often due to language difficulties). Databases sufficient background knowledge to understand the tools being of training events, distribution of computerized training packages made available to them (Cudlip, 1994). CHAPTER 6

ELEMENTS OF SUCCESSFUL REMOTE SENSING IN OPERATIONAL HYDROLOGY

]n order for the smooth and rapid transfer of new developments in 6.1 FINANCIAL RESOURCES remote sensing into operational status, strategies for anticipating future trends are essential (WMO, 1993a). The remote sensing In order to maintain applications of remote sensing on a long-term instruments planned to be on board the new generation of opera~ operational basis, it is necessary to have sufficient financial tional satellites scheduled up to the early 2000s have already been resources available. International agencies need to provide funds decided and some are currently being tested on satellites in an and leadership to assist developing nations in establishing and experimental mode (WMO, 19930). The high rate of development maintaining long-term remote sensing data networks in the communication and computing technologies continues to (UNESCOIWMO/ICSU, 1993). However, WMO (l993a) reports have significant benefits for satellite data utilization; these offer that programmes are being severely pruned and launch schedules the capability to receive, process. disseminate and display increas­ significantly deferred now in almost all satellite operating coun­ ingly large amounts of data faster and faster, while at the same tries due to economic pressure for the resources to be directed time reducing costs and system size. Rapid advances are also elsewhere (for instance to social welfare programmes). WMO being made in education and training systems and technologies further reports that the new expensive earth resources satellites, with the experimental use of computer-assisted learning (CAL) with their output focus on climate and environment, are now systems. CD-ROM and video disc technology are now being used competing for funds from the funding pool which was once the for satellite data archival and for production of satellite application exclusive domain of the meteorological satellites, whose output models. These systems have the capacity to store and replay a focus is essentially on day-te-day weather forecasting. large amount of satellite imagery at relatively low cost and can be Furthermore, the development trend for the future of the connected to PCs to enable interaction with local applications meteorological satellite programme itself has flattened consider­ software. WMO (1993a) further reports that small but quite ably in the past few years and has certainly not kept pace with powerful and affordable desktop workstations are now in general trends in the communications satellite programme or in the trans­ use in the more developed countries and are penetrating into the fer of new technologies into satellite processing, applications and least developed countries. The Zeam Meteo Disc system, training systems (WMO, 19930). produced by the Freie Universitat Berlin, is being used in New Zealand and will be introduced soon for training in Australia. The new concept of Meteorological Data Distribution (MDD) system 6.2 REGIONAL REMOTE SENSING CENTRES works and it has shown great potential in dissemination of weather bulletins via the Meteosat satellite in Africa and the Middle East~ The establishment of regional remote sensing centres is an inter­ these have greatly enhanced meteorological data availability in national effort which aims to ensure that all countries have Africa. A project called MOSAIC has been initiated by EUMET­ sufficient remote sensing capabilities by transferring technology SAT and is being developed by the University of Oxford for and providing the appropriate technical assistance. Furthermore, training on the use of MDDdata with specific emphasis on appli­ the technology transfer process includes demonstration projects cations in devcloping countries. Data Collection Platfonns (DCP) and establishment of international databases to support projects on and Data Retransmission Systems have also shown an improve­ regional and global scales (WMO, 1993a). According to ment in the availability of data. However, WMO (l993a) points UNESCO/WMO/ICSU (1993), capacity-building programmes out that the introduction of Electronic Bulletin Boards for provid­ should be flexible in approach and give due consideration to ing information on satellite-related activities and digital differences between countries and their corresponding needs in infonnation networks is beyond the capacity of many of the devel­ time and space. oping countries in the short term. According to WMO (l993a), the new generation of earth resources satellites which carry active sensors, such as SAR, 6.2.1 Expert advice open up new horizons for applications in land use, oceanography, hydrology and the general environmental impacts of climate The report by UNESCOIWMO/ICSU (1993) suggests that the change; these offer substantial benefits for monitoring in these promotion of the development and operational application of new fields in the developing countries. These satellites deliver more remote sensing technology involves the cooperation of several data (orders of magnitude) than the current generation of meteoro­ types of experts (i.e. a multidisciplinary approach) such as opera­ logical satellites and thus require much larger and more expensive tional hydrologists/researchers, systems analysts, statisticians, ground stations and extremely expensive data-processing, archival geographers, remote sensing and GIS experts and computer and data access facilities. Furthennore, they open up a whole new specialists. WMO has provided a list of experts (Appendix II of field of education and training much of which will require very Annex II (WMO, 19930) report and Annex V (WMO, 1994) advanced knowledge such as a thorough understanding of the report) in the use of satellite data for meteorology and operational complexities of interpreting remotely sensed radar imagery and hydrology. These experts could be used as consultants for devel­ associated data. WMO (l993a) further states that it will take a oping special training modules or conducting special training long time to impart such knowledge and interpretive skills to large courses for RMTC inslructors. UNESCOIWMO/ICSU (1993) numbers of operational meteorologists and hydrologists. highlights that important organizational and institutional matters 42 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

requiring urgent consideration include the following: 1) effective education and training activities. The first priority should be to coordination between United Nations (UN) agencies involved in focus as much funding as possible to upgrading the satellite the water field; and 2) improved documentation elements of the education and training facilities and capabilities of the six Hydrological Operational MnWpurpose System (HOMS) RMTCs. According to WMO (1995), EUMETSAT has actively programme. investigated the possibility of co-sponsoring two Specialized Satellite Applications Training Centres (SSTC), one at Nairobi in Kenya and the other in at Niamey in Niger. These two SSTCs 6.2.2 Training of personnel would operate for trial and evaluation and serve as a basis for concentrated efforts to extend the concept to other parts of the According to WMO (1993a), ECSAT studied techniques to world. improve the use of satellite data through more effective education According to WMO (1995), good progress is being and training. Central to the techniques is the concept of training of made in implementing other components of the satellite education instructors at the RMTCs in order to provide "on-the-spot train­ and training strategy within WMO and in special organizations in ing." ECSAT has proposed a strategy for education and training in various parts of the world. The Japan Meteorological Agency satellite matters with the strategic goal to systematically improve sponsored a special international training seminar in Tokyo in the use of satellite data for meteorological and hydrological appli­ November 1994 on the utilization of GMS satellite data. Most of cations over the next 10 years in all Member countries, with a the attendees were from developing Asia-Pacific countries. The focus on meeting the needs of developing countries. In order to Australian Bureau of Meteorology coordinated a follow-up achieve this goal, there are three major strategic objectives. The seminar in Melbourne which was held in November 1996. WMO first objective is to build on the existing infrastructure (includes conducted a major satellite training event in Costa Rica in October functional responsibilities, organizational structure, management 1995. The China Meteorological Administration is planning to principles and data quality control) in a way that the initiatives are have an international satellite applications training seminar in consistent with the users' capabilities to absorb and sustain them Beijing to focus on the utilization of FY-2 satellite data about a independently in their own operational environment. With regard year or two after its launch. Furthermore, EUMETSAT hosted a to implementing new technology in developing countries, three Users' Forum at the Niamey RMTC in April 1995 during the week main areas requiring training include: 1) the operation and mainten­ following a WMO training event for RA I. ance of the facilities; 2) the use of the information; and 3) the The workshop held in Strasbourg in 1993 on Training in infrastructure supporting the facilities. The second objective is to Remote Sensing: A European Proposal? revealed additional focus on the developing countries with particular attention to successful education and training accomplishments. The UN systematically improving the level of expertise of instructors at all Economic and Social Commission for Asia and the Pacific RMTCs in the use of satellite data. ESA offers a variety of satel­ (ESCAP), based in Bangkok, provides various training courses lite application training activities which could be used by RMTC with remote sensing content for Asian and Pacific countries instructors. NOAA/NESDIS has proposed some initiatives for (Cudlip, 1994). Willemont (1994) presented and demonstrated satellite technology and applications training programmes and France's interactive remote sensing training course on a PC station seminars for RMTCs over a five-year period. The third objective is with a CD-ROM reader. He also presented PROMETE, a series of to anticipate future trends in satellite data applications and. in multimedia and multi-targeted programmes. These modular education and training techniques, so that new developments can programmes are designed to meet the training needs of specialists flow through to operational users quickly and efficiently. The in remote sensing worldwide as well as non-specialists for whom Executive Council Pane1JCommission for Basic Systems (CBS) remote sensing is becoming an important tool. Mather (1994) Working Group on Satellites has drafted a comprehensive set of presented the Combined Higher Education Software Team WMO satellite observation requirements up to the year 2010. (CHEST) arrangement in the UK, which provides access for In order for ECSAT to implement the above proposed higher education teaching purposes to an archival of up-to-date strategy, it has highlighted two important points. First is the estab­ Landsat TM images of the UK for an annual fee. Furthermore, lishment of specialized satellite application training centres at six various universities in Europe, such as Dundee University in the RMTCs (strategically located around the globe). Second is that UK, offer short courses in remote sensing for specific audiences. each satellite operator participating in the space-based subsystem These short courses have been valuable in terms of continuing of the Global Observing System cooperates with at least one of professional development of specialists working in industry and the six specialized RMTCs with regard to satellite training government (Mather, 1994). programmes, facilities and expertise required. Furthennore, while Significant progress has been made in two additional the strategy aims to serve the full spectrum of needs, from the areas of the WMO satellite education and training strategy. They least developed to the most developed countries, it is intended to both involve the introduction ofnew technology and they are: 1) the be applied in such a way that the focus is always on the needs of development of CAL modules for satellite education and training; the developing countries (the least developed will benefit more and 2) the development of electronic mail and bulletin board than the others). Thus, the gap in satellite data utilization capabil­ systems which open up worldwide access to satellite education and ity between the least developed and the more developed countries training infonnation (WMO, 1995). CAL activities are now being should become continually narrower. Additionally, the proposed heavily promoted at various international seminars and are being overall funding strategy is based on a cost minimization strategy coordinated by various bodies including Meteo-France and the and a cooperative funding mechanism which strives for more effi­ Standing Committee of Heads of Training Institutes of national cient use of the existing funded projects and funding sources by Meteorological Services which has strong links with the WMO focusing them all on a few clearly identified high-priority satellite Education and Training Programme. The German Foreign Aid CHAPTER 6 - ELEMENTS OF SUCCESSFUL REMOTE SENSING IN OPERATIONAL HYDROLOGY 43

Programme (GTZ) has agreed to fund a project on CAL generation. The Local Applications of Remote Sensing development within the goals and objectives of the WMO satellite Techniques (LARST) systems allow reception of Meteosat and education and training strategy. The expansion of electronic mail, NOAA satellite data locally, enabling users to interpret data electronic bulletin board systems and Internet facilities rapidly using personal computers. incorporating items such as the World Wide Web is now opening up In an effort to maintain the polar-orbiting and geo­ easy access to huge amounts of satellite training information and stationary satellite systems to ensure the continuity of operation, training modules. WMO Headquarters is now connected to Internet and the data dissemination and distribution services of those satel­ and has developed a series of useful home pages which provide lite systems, WMO (I993a) outlined the critical mission information via the World Wide Web. NOAAlNESDIS has set up a requirements: 1) the contingency plans include coverage of those Web home page containing three satellite interpretation tutOliaIs. regions of the world where severe weather conditions (e.g. EUMETSAT and several other organizations also have information cyclones, tornadoes; etc.) develop; 2) the importance of direct available for electronic access in this manner. The major immediate broadcast services such as APT, WEFAX, HRPT continuity should problem is to put in place the basic communications and computing be considered; and 3) to ensure the continued availability of high­ infrastructure to enable the specialized satellite RMTCs to connect resolution data, standardization of transmission links and fonnats to these systems. This involves items of funding, technical know­ should be considered. how and consideration of the problems of improving the basic communications technology in many developing countries. A US company, AT&T, has begun marketing a new 6.4 GROUNDTRUTH NETWORKS video-teleconferencing service which could make distance learn­ ing easier by enabling instructors to transmit CAL modules over In the application of remotely sensed data for hydrological meas­ telephone lines from desktop computers equipped with small urement problems, the advantages of both ground-based and aerial video-conferencing cameras. Although this system is still in the measurements are best combined. "Ground data are used to adjust development stages and would not be available immediately to or calibrate the transfer function to ensure convergence of the developing countries because of cost and communications infra­ remotely sensed and point data" (Blyth, 1993). For instance, the structure factors, once established such a system would be an UK developed procedures for the calibration of weather radar efficient, cost-effective way to train people who are geographi­ based on fitting multiquadratic surfaces to the calibration factor cally dispersed (WMO, 1995). values, conventionally defined as the ratio of raingauge to coinci­ dent weather radar grid~square estimates of rainfall. Using a gauge density of one gauge per 120 km2 area, the application of simple 6.3 GROUND RECEIVING STATIONS raingauge calibration to the radar data improves the accuracy by 15 per cent and the use of the more sophisticated surface fitting For a successful data acquisition system, there is a need for auto­ method increases this further to 22 per cent, on average. GISs are mated reliable, accurate field stations, telemetered for real-time essential to relate the spatial and temporal remotely sensed data to infonnation where appropriate and simple manual observation ground validation information. networks where appropriate (UNESCO/WMO/ICSU, 1993). According to WMO (1993a), there are certain data circulation requirements regarding both the near-polar-orbiting and geosta­ 6.5 KEY ANCILLARY SUPPORT SYSTEMS tionary satellites and thus their respective ground receiving stations. For the near-polar-orbiting satellite there is a two-fold In order for the application of remote sensing to be successful, key requirement of direct broadcasting capability and global data ancillary support systems must be available. Beyond the initial acquisition. Direct broadcasting benefits local processing (e.g. computer data processing- of the digital remote sensing imagery, nowcasting) and other real-time applications. Global data acquisi­ GISs are used to supply all other important information (spatial, tion is required without gaps in coverage or in time. Data may be temporal and statistical) and further integrate and analyse the available for applications within two hours of observation; this is remote sensing information. A GIS is an organized collection of achieved by on-board storage and successive transmission to computer hardware, software and geographic data designed to Command and Data Acquisition stations or by using Data Relay efficiently capture, store, update, manipulate, analyse and display Satellites. As for the geostationary satellites, real-time services are all forms of geographically referenced information (Johnson et al., one of their essential features, which are primarily used for 1992). The information pertaining to various spatial or temporal nowcasting. The portions of the image for regional use must be features is stored typic-ally as attributes in tabular files linked to available within a few minutes from the observation. the feature, often in special database management systems Geostationary satellites must provide a relay service for data (DBMSs). It is essential that efficient synergistic processing tech­ collection from command-based DCPs. These satellites could also niques be developed to cope with the large multisensor data be used for telecommunications, such as EUMETSAT's volumes and to allow efficient GIS integration. Remote sensing Meteorological Data Distribution. Furthermore, as a telecommuni­ technology and GISs are both tools for managing spatially distrib­ cation facility, a geostationary satellite could be used to transfer uted infonnation in large quantities and at a variety of scales. Both image data from other geostationary satellites and near-polar­ increase the capabilities of human decision makers and planners to orbiting satellites out of the acquisition range for local direct grasp relationships at larger scales and in more complex settings read-out (WMO, 1993a). The NRI in the UK, in conjunction with than has been possible before. In order for remote sensing and other British organizations, has developed low-cost, simple-to~use GISs to be integrated truly, several technical impediments still and robust systems for local satellite data reception and product need to be overcome. A major problem is the raster/vector (remote 44 CURRENT OPERATIONAL APPLICATIONS OF REMmE SENSING IN HYDROLOGY sensing/GIS) dichotomy; this is caused by the difference in the than that of remote sensing and so remote sensing can be incorpo­ structures to acquire and store the data. The most promising areas rated within it (Cudlip, 1994). Many preliminary efforts to unite that have emerged to date can be associated with the automated remote sensing and GISs into one technology for information generation of DEMs, change detection and database production of management have been made over the past few years, both in the GISs (Johnson et at., 1992). A characteristic of most successful commercial and the academic worlds (Ehlers, 1992). Very reason­ integrated systems is that they are confined to very specific prob­ ably priced and versatile GIS software is now available. The NRI lems and applications and specific case studies. Remote sensing in the UK provides a comprehensive service in the use of GISs for data from the Landsat, Shuttle and SPOT programmes have been the input, storage, retrieval, analysis and output of geographically­ used successfully in a GIS environment for base map production, referenced spatial data. automated DEM extraction, terrain visualization and map revision. Other important key ancillary support systems dealing Areas which require special attention include development of new with data quality, acquisition, distribution and processing were fundamental geometric data structures (especially for multitempo­ outlined by UNESCO/WMO/ICSU (1993) and are given as ral and three-dimensional data) and the development of an follows: 1) a data quality control and assurance system; 2) accessi­ integrated DBMS. Furthermore, the unification of these tecnolo­ ble integrated (multi-agency and multivariate) data banks; 3) gies will inevitably lead to a synergistic integration of spatial data coordinated international data banks with free exchange of data handling (Johnson ef al., 1992). for international projects and provision of analytical methods for The GIS market is currently growing very quickly. use by countries; 4) improved delivery of accurate, scientifically Satellite-derived infonnation can be a prime source of data for validated infonnation to decision makers, and 5) design methodol­ GISs if properly packaged. The GIS market is ten times larger ogy for cost-effective multi-purpose networks. CHAPTER 7 RECOMMENDATIONS

There are several key recommendations for both developed and information to produce -new products and applications; 3) educa­ developing countries with the goal to optimize the operational tion and training of the true end-users through databases of training utilization of remote sensing in operational hydrology. Although events, distribution of computerized training packages; and 4) most of the developed countries are not suffering from financial assimilation of satellite-focused education and training. Fifth, there constraints as badly as some of the developing countries, both is a need to provide effective mechanisms for coordination among share some organizational problems. decision-making agencies. The second and fifth recommendations each may be accomplished through joint programmes among their respective agencies for publications, meetings and training. 7.1 KEY APPROACHES TO INCLUDE IN DEVELOPED COUNTRIES 7.2 KEY APPROACHES TO INCLUDE IN There is a need for modification of current organizational and insti­ DEVELOPING COUNTRIES tutional arrangements at both national and international levels. First, there is a need for bettcr coordination and removal of barriers In addition to the organizational and institutional modifications between the national government agencies and international organ­ recommended above for the developed countries, there is also a izations that provide the funding for hydrological remote sensing need for continued international financial assistance and modifica­ and operational hydrology. Second, there is a need for inter-agency tion of national financial arrangements. First, there is a need for and interdisciplinary collaboration in the research ofhydrology and international agencies to continue to provide funds and leadership water resources so as to avoid overlaps and duplication of activities to assist developing nations in establishing and maintaining long­ between agencies and disciplines. 1bird, there is a need to bridge term remote sensing data (acquisition and distribution) networks. the gap between the scientists (including satellite systems planners Second, there is a need to facilitate the accessibility to new remote collecting the data) and the decision makers (end-users specifying sensing technologies by developing efficient and reliable informa­ data requirements) by integrating remote sensing information/capa­ tion systems especially in relaying satellite-related activities. bilities through a national archive. Fourth, there is a need for the Third, there is a need for national and international aid-to-develop­ research hydrologist to continue development of hydrological ment programmes to continue to actively support the acquisition models based on remote sensing inputs, demonstrate technical and and maintenance of processing facilities. Fourth, there is a need to economic benefits from using remote sensing, and to improve the continue effective education and training capability progranuness design of pilot projects (involving applications of remote sensing). and documents such as those resulting from UNESCO (IHP) and An improvement in the design of pilot projects could be accom­ WMO (OHP) programmess, thus encouraging the developing plished by involving the true end-users (decision makers) from the nations to emphasize developing skills and professional attitudes. onset since they will have to give authoritative long-term validation Fifth, there is a need to continue the establishment and develop­ of applications theoretically demonstrated and thus ensure the ment of regional remote sensing centres around the world, which continuation of the activity (i.e. operational application) if the pilot is an international effort to ensure technology transfer (through project is successful. This may be accomplished through: demonstration projects and establishment of databases to support 1) improved information services in order to allow the end-users to projects on regional and global scales), to provide the appropriate know where the information is located and how to use it; 2) utiliza­ technical assistance and to train instructors at the RMTCs to tion of GISs to combine remote sensing data with other provide "on-the-spot training." REFERENCES

Ackleson, S.G. and Klemas, V, 1987. Remote sensing of submerged Barrett; E.C. and Harrison, A.R., 1986. Rainfall monitoring by aquaticvegetation in- Lower Chesapeake Bay: A comparison of Meteosat in Africa, Consultant's Report, AGRT, FAO, ~ome, Landsat MSS to TM imagery, Remote Sensing a/Environment; 16pp, 22, pp. 235"248. Barrett, E.C, Herschy, RW. andStewart, J.B., 1988b, Satelliteremote Ackleson, S.G., Kkmas, V., McKim, H.L. and Merry, CT., 1985. A sensing requirements for hydrology and water managementfrom comparisonofspar simulationdatawithLandsat MSS imagery the mid~ 1990s, in relation to the Columbus Programme of the for delineating watermasses in Delaware Hay, Broadkill River, European Space Agency, Journal ofHydrological Sciences, 33 and adjacentwetlands, Photogram. Eng; Remote Sensing, 51, pp. (1)(2), pp. 1-17.

1123-1129' Barry, R.G., 1984. Possible CO2-produced wanning effects on the Adler, R.E, Negri, A.J"J and Hakkarinen, I.M., 1991. Rain estimation cryosphere, in Climate Changes on a Yearly to Millennial Basis, from combining geosynchronous IR and low-orbit microwave Edited by N.A. Morner and W. Karlen, D. Reidel, Hingham, data, Global and Planetary Change, 4(1/3), pp. 87-92. Mass., pp. 571-604. Ahmed, E, Andrawis, A.S. andHagaz, YA., 1984. Landsat model for Baumgartner, A. and Reichel, E, 1975. DieWeitwasserbilanz (World groundwater exploration in the Nuba Mountains, Sudan, Adv. waterbalance), Oldenbourg, Munich. Space Res., 4 (11), pp. 123 131. Baumgartner, M.E, 1988. Snowmeltrunoffsimulation based on snow Ahmed, S., Warren, M.D., Symonds, M.D. and Cox, RP., 1990. The cover mapping using digital Landsat-MSS and NOAAlAVHRR RADARSAT system, IEEE Transactions on Geoscience and data, USDA-ARS, Hydrology Lab. Tech. Rep. Remote-Sensing, 28, pp; 598-602. Bausch, W.e. and Neale, CM.D., 1987. Crop coefficients derived Ali, A., Quadir, D.A. and Huh, O.K., 1987. Agricultural, hydrologic from reflected canopy radiation: a concept, Trans. AnL Soc. and oceanographic studies in Bangladesh with NOAA AVHRR Agric. Eng., 30, pp. 701-709. data, International Journal ofRemote Sensing, Tech. Note, 8, pp. Becker, T.W., 1994. Remote sensing and theArnerican flood of 1993, 917-925. Earth Space Review, 3 (4), pp. 8-11. Andersen, T., 1991. AVHRR data for snow mapping in Norway, Becker, T.W., 1993. The Hurricane Andrew scenario, Earth Space Proceedings ofthe 5th AVHRR Data User Meeting, Tromsoe, Review, 2(4), pp. 7-10. Norway; Berg, c.P., Matson, M. and Wiesnet, D.R., 1981. Assessing the Red Anderson, T. andAnderson, B., 1992. Radar rain forecasts for sewer­ River of the North 1978 flooding from NOAA satellite data, age systems, 2nd Int. Symp. on Hydrological Applications of Satellite Hydrology, AWRA, Minneapolis, MN, pp. 309-315. Weather Radar, Univ. ofHanover, September 7-10, 1992, Paper Bergstrom, S. and Brandt, M., 1985. Snow mapping and hydrologi­ N1,lOpp. cal forecasting by airborne-ray spectrometry innorthern Sweden, Askew, A.J., 1991. Climate and water - a call for international action, HydrologicalApplications ofRemote Sensing and Remote Data Hydrological Sciences Journal, 36, pp. 391-404. Transmission (Proceedings ofthe Hamburg Symposium), lARS Bailey, lo., 1990. The potential value ofremotely sensed data on the Pub!. No. 145, pp. 421-428. assessment of evapotranspiration and evaporation, Remote B"emard, R., Vauclin, M. and Vidal-Madjar, D., 1981. Possible use of Sensing Reviews, 4 (2), pp. 349"377. active microwave remote sensing data for predictionofregional Baker, J.R, Briggs, SA, Gordon, V., Jones, A.R., Settle, J.J., evaporation by numerical simulation of soil water movement in Townsend, lR.G. and Wyatt, B.K., 1991. Advances in classifi­ the unsaturated zone, Water Resow: Res., 17, pp. 1603-1610. cation for land cover mapping using SPOT HRV imagery, Int. 1. Birkett, C.M., 1994. Radar Altimetry: A new concept in monitoring Remote Sens., 12, pp. 1071-1086. lake level changes, EOS Transactions, AGU, Vol. 75, No. 24, pp. Barrett; E.C., 1991. Diagnostic, historic, and predictive analysis of 273-275. rainfall using passive microwave image data, Global and Birkett, e.M., 1995. The global remote sensing oflakes, wetlands and Planetary Change GPCRE 4, 4 (113), pp. 99-106. rivers for hydrological and climate research,lEEE, 0-7803-2567, Barrett, E.e., 1993. Precipitation measurement by satellites: towards pp.1979-1981. community algorithms. Adv. Space Res., 13(5), pp. (5)119­ Blasco, E, Bellan, ME and Choudhury, M.U., 1992. Estimating the (5)136. extent offloods in Bangladeshusing SPOT data, Remote Sensing Barrett, E.C, Beaumont, MJ., andD'Souza, G., 19880. A system for ofEnvironment, 39, pp. 167-178. operational rainfall monitoring by Meteosat, Proceedings, VIIth Blyth, K., 1993. The use of microwave remote sensing to improve Annual Meteosat Scientific Users Meeting, Madrid, Spain, 28­ spatial parameterization of hydrological models, Journal of 30, Sept. 1988. Hydrology, 152, pp. 103-129. Barrett, E.C., Beaumont, MJ. and Herschy, R.W., 1990. Satellite Brakke, T.W. and Kanemasu, E.T., 1981. Insolation estimation from remote sensing for operational hydrology: present needs and satellite measurements of reflected radiation, Remote Sensing future opportunities, Remote Sensing Reviews, 4(2), pp. 451­ Environ., II, pp. 157-167. 466. Brest, c.L. and Goward, S.N., 1987. Deriving surface albedo meas­ Barrett, E.C. and Curtis, L.E, 1976. Introduction to environmental urements from narrow band satellite data, /1llernational Journal remote sensing, Chapman and Hall, London. ofRemote Sensing, 8, pp. 351-367. REFERENCES 47

Brisco, B., Ulaby, RT and Dobson, M.C., 1982. Spaceborne SAR data Caselles, Y. and Delegido, l, 1987. A simple model to estimate the for land·cover classification and change detection, IGARSS daily value of the regional maximum evapotranspiration from 1982, Munich, Germany. satellite temperature and albedo images, International Journal "Bristow, M.P.E., et al., 1985. Airborne laser florosensor survey ofRemote Sensing 8 (8), pp. 1151-1162. of the Columbia and Snake Rivers: simultaneous Chandler, S. and Fattorelli, S., 1991. Adaptive grid-square-based measurement of chlorophyll, dissolved oxygen and optical geometrically distributed flood-forecasting model, J.D. Cluckie attenuation, lmemational Journal ofRemote Sensing, 6, pp. and C.G. Collier (Eds.), Hydrological Applications ofWeather 1707-1734. Radar, Chapter 38, pp. 424-439, Ellis Horwood. BruneI, J.P., 1989. Estimation of sensible heat flux from measure­ Chang, A.TC., Foster, J.L. and Hall, D.K., 1987. NIMBUS-7 derived ments ofsurface radiative temperature and air temperature at two global snow cover parameters, Ann. Glaciol., 9, pp. 39-44. meters: application to determine actual evaporation rate. Chang, A.T.C., Foster, J.L. and Hall, DK, 1990. Effect ofvegetation Agricultural and Forest Meteorology, 46, pp. 179-191. cover on microwave snow water equivalent estimates, Proc. Int. Brutsaert, 1982. Evaporation into the atmosphere. Theory, history and Symp. on Remote Sensing and Water Resources. Enschede, applications (Dordrecht. Boston, London: Reidel). August 1990, Int. Assoc. Hydrogeologists and Netherlands Soc. Camagni, P., 1988. Diagnostics ofoil pollution by laser-induced fluo­ for Res. Sen., Enschede, pp. 137-145. rescence, IEEE Transactions on Geoscience and Remote Chang, A.T.C., Foster, J.L., Hall, D.K., Rango, A. and Hartline, B.K., Sensing, GE-26, pp. 22-26. 1982. Snow water equivalent estimation by microwave radiom~ Camillo, P.J., Gurney, RJ. and Schmugge, T.J., 1983. A soil and etry, Coid Regions Res. Teclmol., 5, pp. 259-267. atmospheric boundary layer model for evapotransporation and Chauhan, N.S. and Lang, RH., 1991. Radar modelling of a boreal soil moisture studies, Water Resources Research, 19, pp. 371­ forest. IEEE TransactiollS on Geoscience and Remote Sensing, 380. 29, pp. 627-637. Carder, K.L., Hawes, S.K., Baker, K.A., Smith, RC., Stewart, RG. Choudhury, BJ., 1989. Estimating evaporation and carbon assimila­ and Mitchell. B.G., 1991. Reflectance models for quantifying tion using infrared temperature data: vistas in modeling, in chlorophyll a in the presence of productivity degradation prod­ Theory and Applications ofOptical Remote Sensing (G. Asrar, ucts, Journal ofGeophysical Research, 96, pp. 20599-20611. Editor), John Wiley and Sons, New York, 734 pp. Carlson, TN., Dodd, J.K., Beujamin, S.G. and Cooper, J.N., 1981. Choudhury, B.J., 1993. Comparing SMMR and SMMII 37 Ghz Satellite estimation ofthe surface energy balance, moisture avail­ Observations for Monitoring Surface Change, EOS, AGU Fall ability and thermal inertia, Journal ofAppliedMeteorology, 20, Meeting. pp.67-87. Choudhury, B.J. andMonteith, J.L., 1988. A four-layer model for the Carroll, S.S. and Carroll, T.R, 1989. Effect of forest biomass on heat budget of homogeneous land surfacf?s, Q.J.R. Meteorol. airborne snow water equivalent estimates obtained by measur­ Soc., 114, pp. 373-398. ing terrestrial gamma radiation, Remote Sensing Environ., 7, pp. Clothier, B.E., Clawson, K.L., Pinter, Jr., P.I., Moran, M.S., 313-320. Reginato, R.J. and Jackson, R.D., 1986. Estimation of soil Carroll, S.S., Day, G.N., and Carroll, TR., 1993. Incorporating heat flux from net radiation during the growth of alfalfa, airborne data into the spatial model used to estimate snow water Agricultural and Forest Meteorology, 37. pp. 319-329. equivalent, Geographic Information Systems and Water Cluckie, J.D. and Pessoa,M.L., 1990. Dam safety: an evaluation of Resources, Journal ofthe Amer. Water Resources Association, some procedures for design flood estimation, Hydrol. Sci. J., 35 March, pp. 259-264. (5), pp. 547-569. Carroll, T.R., 1990. Operational remote sensing of snow cover in the Cluckie, I.D., Ede, P.E, Owens, M.D. and Bailey, A.C., 1987. US and Canada, Proceedings of the 1990 National Conference Some hydrological aspects of weather radar research in the on Hydraulic Engineering, American Society ofCivil Engineers, United Kingdom, Hydrological Sciences Journal, 32, 3, 9, San Diego, CA. pp. 329-346. Carroll, T.R. and Schaake, Jr., lC., 1983. Airborne snow water equiv­ Cluckie, J. and Tyson, J.M., 19.89. Weather radar and urban drainage alent and soil moisture measurement using natural terrestrial systems, Weather Radar and the Water Industry, British gamma radiation, Proceedings of SPIE - The International Hydrological Society Occasional Paper No.2, Institute of Society for Optical Eugiueering, 414, pp. 208-213. Hydrology, pp. 65-73. Carroll, T.R. and Vadnais, K.G., 1980. Operational airborne meas­ Collier, CG., 1993. Estimating Probable MaximwnPrecipitation (PMP) urement of snow water equivalent using natural terrestrial using a stonn model and remotely-sensed data, British gamma radiation, Proceedings ofthe 48thAnnual Western Snow Hydrological Society 4thNational Hydrology Symposium, Cardiff, Confereuce, Laramie, WY, pp. 97-106. September 13-16, 1993,Institute ofHydrology, pp. 5.33-5.38. Carroll, T.R. and Vose, G.D., 1984. Airborne snow water equivalent Collinge, Y.K. and Kirby, C. (Eels.), 1987. Weather Radar and Flood measurements over a forested environment using terrestrial Forecasting, J. Wiley, 296 pp. gamma radiation, Proc. 41st Annu. Eastern Snow Conf., New Collins, W.G. and Van Genderen, J.L. (Editors), 1978. Remote sens­ Carrollton, MD, Eastern Snow Conference, p. 19. ing applications in developing countries, Remote Sensing Carter, Y. and Schubert, 1., 1974. Coastal wetlands analysis from Society, Reading, 10 pp. ERTS MSS digital data and field spectral measurements, Congalton, R.G., Green, J., Teply, J., 1993. Mapping old growth Proceedings of the Ninth International Symposium on forests in National Forest and Park lands in the Pacific northwest Remote Sensing of Environment, Environmental Research from remotely sensed data, Photogrammetric Engineering and Institute of Michigan, Ann Arbor, MI, pp. 1241-1260. Remote Sensing, 59 (4), pp. 529-535. 48 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

Conway, B.l., 1989. Expert systems on weather forecasting, Met. Dozier, J., 1987. Recent research in snow hydrology, Rev. Geophys., Mag., 118, pp. 23-30. 25, pp. 153-161. Cooper, C.F.. 1965. Snow cover measurement, Photogr. Engin., 31, Dozier, J., 1989. Spectral' signature of alpine snow cover from the pp.611-619. Landsat Thermatic Mapper, Remote SellS. Environ., 28, pp. 9-22. Corbley, K.P., 1994. EOSAT -India partnership broadens interna­ Dozier, J., Davis, RE., Chang, A.T.C. andBrown, K., 1988. The spec­ tional remote sensing market, Earth Space Review, 3 (4), pp. tral bidirectional reflectance of snow, Spectral Signatures of 20-26. Objects in Remote Sensing, Fourth International Colloquium, eorbley, KP., 1993, Remote sensing and GIS provide rapidresponse ESA SP-287, European Space Agency, Paris, pp. 87-92. for flood relief, Earth Observation Magazine, September, pp. D'Souza, G., Barrett, B.C. and Power, C.H., 1990. Satellite rainfall 28-31. estimation techniques using visible and infrared imagery, Remote CaSPAR, 1993. Observations from space of the Mississippi Flood, Sel1sil1g Reviews, 4(2), pp. 379-414. Infonnation Bulletin, 128,45 pp. Dunkel, Z, andVadasz, v., 1993. Estimation of regional evapotran­ Cross, A.M., 1992. Monitoring marine oil pullution using AVHRR spiration over Hungary combining standard and satellite data, data: observations off the east coast of Kuwait and Saudi Arabia Adv. Space Res., 13 (5), pp.(5)257-(5)260. during January 1991, International Journal ofRemote Sensing, Dutartre,P. Coudert, J.M. and Delpont, G., 1993. Evolution in the use 13, pp. 781-788. ofsatellite data for the location and development of groundwa­ Csistar,1. and Bonta, I., 1993. Use ofsatellite soundings in monitor­ ter, Adv. Space Res., 13 (5), pp. (5)187-(5)195. ing mesoscale precipitation systems, Adv. Space Res., 13(5), pp. Eagleman, J.R., 1975. Detection of soil moisture and snow charac­ (5)137-(5)141. teristics from Skylab. Final- Report, 239-23, EREP Cndlip,. w., Ridley, 1.K. and Rapley, C.G., 1990. The use of satellite No. 540-A2, Contract NAS 9-13273, Kansas University radar altimetry for monitoring wetlands, Remote Sensing and Centre for Research. NASA; Lyndon B. Johnson Space Gl9bal Change, Proc. 16th Annual Conf. ofthe Remote Sensing Centre, Houston, TX, 313 pp. Society, University College Swansea, September 19-21, Remote Ehlers, M., 1992. Remote sensing and geographic information Sensing Society, Nottingham, pp. 207-216. systems: image-integrated geographic information systems, Int. Cudlip, W., 1994. Rapporteur's summary of session 2: Manpower Symp. on Mapping and Geographic Information Systems, requirements and qualifications, International Journal ofRemote ASTM Special Technical Publication n. 1126, ASTM, Sensing, 15 (15), pp. 3045-3047. Philadelphia, PA, pp. 53-67. Curran, P.l. and Novo, E.M.M., 1988. The relationship between EI-Baz, F.,1993. TMreveals Arabian Desert secrets,EOSATLandsat suspended sediment concentration and remotely sensed,spectral Data Users Notes 8(2), Snmmer 1993, pp. 3. data: Areview, Journal ofCoastal Research, 4, pp. 351-368. Elkington, M.D. and Hogg, J., 1981. The characteristics ofsoilrnois­ Davies-Colley, R.J., Vant, W.N. and Wilcock, R.I., 1988. Lake ture content and actual evapotranspiration from crop canopies water color:. Comparison of direct observations with under­ using thermal infrared remote sensing, Proc. Remote Sensing water spectral irradiance, Water -Sesources Bulletin, 24, pp. Soc., Reading. 11-18. EI Shazly, E.M., EI Raikaiby, M.M. and El Kassas, l.A., 1983. Davis, RE. and Shi, le., 1993. Microwave measurements of snow Groundwater investigation of Wadi Araha area, Eastern Desert cover: progress towards operational hydrology, EOS, AGU Fall of Egypt, using Landsat imagery, Proc. 17th Symp. on Remote Meeting. Sensing of the Environment, Ann Arbor, Michigan, May 9-13, Delrieu, G., Creutin, J.D. and Andrieu, H., 1989; Validation ofashort 1983, pp. 1003-1013. range X-band radar for rainfall measurement over an urban area, Engman, B.T. and Gurney, R.I., 1991. Remote sensing in hydrology, Proc. Int. Symp. on Hydrological Applications ofWeather Radar, Chapman and Hall, 225 pp. University ofSalford, UK. EOSAT, 1993a. Monitoring Australian Floods, EOSAT Notes, 8 (2), Delrieu, G., Creutin, J.D. and Andrieu, R., 1992. Analysis ofvarious p.7. algorithms for correcting attenuation effects; 2nd Int. Symp. on EOSAT, 1993b. Landsat and ERS-I data fusion, EOSAT Notes, 8 Hydrological Applications of Weather Radar, September 7-10, (3/4), p. 12. 1992, University ofRanover, Germany, 15 pp. EOSAT, 1993c. Government and Industry Team in an Emergency, Demerliac,Y., 1994. Rapporteur's summary ofsession 5: Cost/bene­ EOSAT Notes, 8 (2), p. 4. fit aspects ofremote sensing education, International Journal of EOSAT, 1994a. Russian Photo Covers Cairo at 2 Meteors, EOSAT Remote Sensing, 15 (15), pp. 3119-3121. Notes, 9 (I), p. 7. Desinov, L.V, 19.80. Use of satellite information in glaciology, EOSAT, 1994b. Floods Visit Western Europe, EOSAT Notes, 9 (I), Prnceedings of GGI, 276, Leningrad, Russian Federation, pp. p.4. 59-65. EOSAT, 1994c. EOSAT's statewide purchase plan keeps South Dewey; K.F. and Heim, Jr., R, 1981. Satellite observations of varia­ Carolina residents in the know, EOSAT Notes, 9 (I), pp. 3 tions in northern hemisphere seasonal snow cover, NOAA and 7. Technical ReportNESS 87, NOAA, Washington, DC, 83 pp. EOSAT, 1994d. Ocean to lagoon to marshland, EOSAT Notes, 9 (I), Dobson, M.e. and Ulaby, EE, 1986. Active microwave soil moisture p.1O research, IIEEE Transactions on Geoscience and Remote Ermolin, Y.A., 1992. Operational control of urban sewage system of Sensing, GE-24, No. I, pp. 23-36. Moscow, 2nd Int. Symp. on Hydrological Applications of Doyle, EJ., 1984. Surveying and mapping with space data, ITC Weather Radar, Univ. ofHanover, September 7-1 0, 1992, Paper Joumal, 4, pp. 314-321. N3, 9 pp. REFERENCES 49

Evans, D.L. and Smith, M.D., 1991. Separation of vegetation and rock Gitelson, A.A., Dubovitsky, G.A., Keydan, G.P. and Lopatchenko, signatures in Thematic Mapper and polarimetric SAR images, L.L., 1988. Evaluation of water ecosystem quality by its radia­ Remote Sens. Environ., 37, pp. 63-75. tion in the visible spectrum (in Russian), Ecological Farnsworth, RK, Barrett, E.C. and Dhanju, M.S., 1984. Application standardization and modeling ofanthropogeneous influence on of remote sensing to hydrology including groundwater, water ecosystem, Rostov-on-Don, Hydrochemical Institute, International Hydrological Programme, UNESCO, Technical Leningrad, pp. 101-135. Documents in Hydrology, IHP-II project A.1.5, Paris, 122 pp. Goldhirsh, 1., Rowland, J. and Musiani, B., 1987. Rain measurement Fattorelli, S., Casale, R., Borga, M and DaRos, D., 1995./lltegrating results derived from a two-polarization frequency-diversity S­ radar and remote sensing techniques ofrainfall estimation in band radar at Wallop Island, VA" IEEE Transactions on hydrological applications/orflood hazard mitigation, European Geoscience and Remote Sensing, GE-25 (6), pp. 654-661. Commission for Directorate General XII, Science, Research and Goodison, B. and Walker, A.E., 1993. Canadian development anduse Development, Environment Programme, 72 pp. ofsnow cover infonnation from passive microwave satellite data, Fea, M., 1994. Technological means to assist remote sensing training, Proceedings of the ESAINASA International Workshop on International Journal ofRemote Sensing, 15 (15), pp. 3079­ Passive Microwave Remote Sensing Research Related to Land­ 3093. Atmosphere Interactions, 81. Lary, France. Feldman, G., Clark, D. and Halpern, D., 1984. Satellite color obser­ Goodison, B., Walker, A.E. and Thirkettle, E, 1990. Determination of vations of the phytoplankton distribution in the Eastern snow water equivalent on the Canadian prairies using near real­ Equatorial Pacific during 1982-1983 EI Nifio, Science, 226, pp. time passive microwave data, Proceedings of the Workshop on 1069-1071. Application of Remote Sensing in Hydrology, National Ferraro, R.I., Grody, N.C. and Kogut, J.A., 1986. Classification of Hydrology Research Institute, Saskatoon, Saskatchewan, pp. geophysical parameters using passive microwave satellite meas­ 297-316. urements, IEEE Transactions Oil Geoscience and Remote Gross, M.E, Hardisky, M.A., Klemas, Y. and Wolf, P.L., 1987. Sensing, GE-24, 6. Quantification of biomass of marsh grass Spartina alterniflora Finkelstein, M.l, Mendelson, VL. and Kutev, VA., 1987. Radar sens­ Loisel using Landsat thematic mapper imagery, Photogrammetric ing of stratified Earth's covers (in Russian), Soviet Radio Engineering andRemote Sensing, 53, pp. 1577-1583. Publishing House, Moscow, 174 pp. Guglielminetti, M., Boltri, R., Morino, C.M. and Lorenzo, S., 1982. Food and Agricultural Organization (FAD) of the United Nations, Remote sensing techniques applied to the study of fresh-water 1989. Remote sensing applications to water resources, Remote springs in coastal areas of southern Italy, Proc. 16th Int. Symp. Sensing Centre, RSC Series 50, Report of the 13th Remote Sensing of Environment. E.R.I.M., Ann Arbor, UN/FAOIUNESCO International Training Course in cooperation Michigan. with the Government ofltaly, Rome, Italy, September 12-30 and Gupta, R.K., Prasad, S., Nadharn, T.S.Y. and Rao, G.M., 1993. Sassari, Jtaly, Octoher 2-7, 1988,393 pp. Relative sensitivity of district mean RVI and NDVI over an Foster, J.L. and Hall, D.K., 1981. Multisensor analysis of hydrologic agrometeorological zone, Adv. Space Res., 13 (5), pp. (5)261­ features with emphasis on the SEASAT SAR, Photogrammetric (5)264. Engineering and Remote Sensing, 47 (5), pp. 655-664. Gurney, R.I., Dehney, A.G.P. and Gordon, M.R., 1982. The use of Fraysse, G. and Hartl, P., 1985. Remote sensing correlation, CEC, Landsat data in groundwater study in Botswana, 1. Appl. Joint ResearchCentre-IspraEstab. Rep. S.A.I.04.E2.85.32, Joint Photographic Eng., 8, pp. 138-141. Research Centre, Ispra, pp. 81-98. Haggett, C.M., May, B.C. and Crees, M.A., 1992. Advances in oper­ Fuller, R., 1993. The land cover map of Great Britain, Earth Space ational flood forecasting in London, 2nd Int Symp. on Review, 2 (4), pp. 13-18. Hydrological Applications ofWeather Radar, Dniv. ofHanover, Gagliardini, D.A. and Karszenbaum, H., 1987. The importance ofthe September 7-10,.1992, Paper N2, 11 pp. NOAA-AVHRR data in resources inventories and environmen­ Hall, D.K., 1982. A review of the utility ofremote sensing inAlaskan tal monitoring inArgentina and neighboring countries, Advances permafrost studies, IEEE Transactions on Geoscience and in Space Research, 7 (3), pp. 7-10. Remote Sensing, GE-20 (3), pp. 390, 394. Gammon, PT., Rohde, w.G. and Carter, Y., 1981. Accuracy evalua­ Hall, D.K., et al., 1981. Freshwater ice thickness observations using tion ofLandsat digital classification in the Great Dismal Swamp, passive microwave sensors, IIEEE Transactions on Geoscience Satellite Hydrology, AWRA, Minneapolis, MN, pp. 463-473. andRemote Sensing, GE-19 (4), pp. 189-193. Gash, lH.C., 1987. An analytical framework for extrapolating evap­ Hall, DK and Ormsby, J.P, 1983. Use of SEASAT synthetic aper­ oration measurements by remote sensing surface temperature, ture radar and Landsat multispectral scarmer system for Alaskan Intemational Journal ofRemote Sensing, 8. pp. 1245-1249. glaciology studies, J. ofGeophysical Research, 88 (3), pp. 1597­ Gatto, L.W., 1990. Monitoring river ice with Landsat images, Remote 1607. Sensing ofEnvironment, 32, pp. 1-16. Hallikainen, M., 1984. Satellite microwave radiometry of snow Gautier, C., Diak, G. and Masse, S., 1980. A simple model to estimate cover, Proc. of the 18th Int. Symposium on Remote incident solar radiation at the surface from GOES satellite data, Sensing of Environment, October 1-5, Paris, France, pp. J. Appl. Meteorol., 19, pp. 1005-1011. 1405-1414. Gihbons, D.E., Wukelic, G.E., Leighton, J.P. and Doyle, M.J., 1989. Harris, R., Hobbs, A.J. and Munday, T.J., 1984. Study of microwave Application ofLandsat thematic mapper data for coastal thermal remote sensing techniques for land surface and hydrological plume analysis at Diablo Canyon, Photogrammetric Engineering application, Report prepared for NERC under contract and Remote Sensing, 66, pp. 903-909. F60/G6/08, Durham University. 50 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

Harrison, AR. and Lucas, R.M., 1989. Multi·spectral classification of Jackson, R.D., 1985. Evaluating evapotranspiration at local and

snow using NOAA AVHRR imagery, International Journal of regional scales j IEEE Transactions on Geoscience and Remote Remote Sensing, 10, pp. 907"916. Sensing, GE-73, pp. 1086-1095. Heilman, J.L. and Moore, D.C.. 1981a. Groundwater applications of Jackson, R.D., Hatfield, lL.; Reginato, R.I., Idso, S.B. and Pinter; Jr.• the Heat Capacity Mapping Mission, Deutch, Weisnet and Rango PJ., 1983. Estimates ofdaily evapotranspiration from one time (Eds.), Satellite Hydrology, pp. 446-449. ofday measurements, Agricultural Water Management, 7, pp. Heilman, IL. and Moore, D.C., 198th. Soil moisture applications of 351-362. the Heat Capacity Mapping Mission, Deutsch, Weisnet and lackson, RD., Moran, M.S., Gay, L.w. and Raymond, L.H., 1987b. Rango (&Is.), Satellite Hydrology, pp. 371-376. Evaluating evaporation from field crops using airborne radiom­ Herschy, R.W., Barrett, E.C. and Roozerkrans, J.N., 1988. Remote etry and ground~based meteorological data, Irrig. Sci., 8, pp. sensing in hydrology and water management. Final Report, 81-90. ESA Contract No. 5769/A84/DIJS (Sc), EARSeL, lackson, RD., Reginato, R.I. and Idso, S.B., 1977. Wheat canopy Strasbourg, 288 pp. temperature: a practical tool for evaluating water Herschy, R.W., Barrett, E.C. and Roozerkrans, J.N., 1985. T1le world's requirements, Water Resources Research, 13, pp. 651-662. water resources - a major neglect. ESA. BR-40 Paris. lackson, T.J., Hawley, M.E. and O'Neill, P.E., 1987a. Preplanting soil Higer, A.L., Coker; A.E. and Cordes, E.H., 1974. Watermanagement moisture using passive microwave sensors, Water Resources models in Florida using ERTS-l data, Proc. 3rd Earth Resources Bulletin, 23 (I), pp. 11-19. Tecbnology Satellite-I Symp., NASA SP-351-VI, pp. 1071­ Jackson, T., Levine, D., Griffis,A., Goodrich, D., Schmugge, T., Swift, 1088. C. and O'Neill, P., 1993. Soil moisture and rainfall estimation Higuchi, K. and Iozawa, T., 1971.Atlas ofperennial snow patches in over a semiarid environment with the ESTAR microwave Central Japan. Report ofWater Research Laboratory, Nagoya radiometer, IEEE Transactions on Geoscience and Remote University, Japan, 81 pp. Sensing, 31 (4). Hockey, B., Richards, T. and Osmaston, H., 1990. Prospects for oper­ Jack;mn, T.J. and Schmugge, T.J., 1989. Passive microwave remote ational remote sensing of surface water, Remote Sensing sensing system for soil moisture: some supporting research, Reviews, 4 (2), pp. 265-283. IEEE Transactions on Geoscience and Remote Sensing, 27 (2), Hogg, D.C., 1989. Rain, radiometry, and radar,IEEE Transactions on pp. 225-235. Geoscience andRemote Sensing, 27 (5), pp. 576-585. Jackson, T.J., Shutko, A.M., Shiue, J.e., Schmugge, T.J., Davis, Hudlow, M.D., 1988. Technological developments in real-time oper­ D.R., Parry, R., Haldin, A., Reutov, E., Novichikhim, E., ational hydrologic forecasting in the United States, Journal of Liberman, B., Kustas. W.P., Goodrich, D.C., Bach, L. and Hydrology, 102, pp. 69-92. Ritchie, le., 1991. Soil moisture observations using multi­ Huhnerfus, H., 1986. Discrimination between crude oil spills and frequency passive microwave sensors in an arid rangeland monomolecular sea slicks by airborne radar and infrared environment: a cooperative US-USSR experiment, radiometer - possibilities and limitations, International Journal Proceedings of the American MeteorologiCal Society Special ofRemote Sensing, 7 (8), pp. 1001-1013. Session on Hydrometeorology, Salt Lake City, pp. 174-177. Ibrahim, M. and Cracknell, A.P., 1990. Bathymetry using Landsat Janowiak, J.E., 1992. Tropical rainfall: a comparison of satellite­ MSS data of Penang Island in Malaysia, International Journal derived rainfall estimates with model precipitation forecasts, ofRemote Sensing, 11, pp. 557-560. climatologies, and observations. Monthly Weather Review, 120 !bre, G.I., Morcette, 1.1. andHogg, W.D., 1979. Surface water temper­ (3), pp. 448-462. ature of lakes from satellite infra-red data corrected for Jensen, J.R, Narunalani, S., Weatherbee, 0., Mackey, H.E., Jr.. 1993. atmospheric attenuation, Canadian Climate Center, Report No. Measurement of seasonal and yearly cattail and waterIily 79-7, Canada. changes using muItidate SPOT panchromatic data, Idso, S.B., Schmugge, TJ., Jackson, R.D. and Reginato, RI., 1975. Photogrammetric Engineering and Remote Sensing, 59 (4), pp. The utility ofsUlface temperature measurements for remote sens­ 519-525. ingof surface soil water status, J. Geophys. Res., 80, pp. lohnson, A.I., Petterson, C.B. and Fulton, I.L. (Eds.), 1992. 3044-3049. Geographic Information Systems (GIS) and mapping ­ Ijjas, G., 1992. Passive microwave remote sensing of soil moisture Practices and standards, ASTM Publication STP 1126, 346 pp. from aircraft in Hungary, International Journal of Remote Jones, W.K. and Carroll, T.R., 1983. Error analysis ofairborne gamma Sensing, 13 (3), pp. 471-479. radiation soil moisture measurements, AgriculturalMeteorology, Imhoff, M.L., Vermillion, C., Story, M.H., Choudry, A.M., Gafoor, 28, pp. 19-30. A. and Polcyn, E, 1987. Monsoon flood boundary delin­ Joss, J. andLee, R., 1994. Weather Radar: Operational Processing for eation and damage assessment using space borne imaging Nowcasting and Precipitation Estimation, Paper submitted to radar and Landsat data, Photogram. Eng. Remote Sensing, lAM, Feb. 21, 1994. 53, pp. 405-413. Kabbaj, M.M., and Mehrez, M.B., 1994. Remote sensing and devel­ Institute for Remote Sensing Applications (IRSA), 1992. Joint oping countries, training and technology transfer Europe-Africa, Research Centre, Commission of the European Communities, International Journal ofRemote Sensing. 15 (15), pp. 3003­ Annual Report. 3016. ltier, B. and Riou, e.R., 1982. Vne nouvelle methode de detennina­ Kammer, A., 1991. A low cost X-band radar system designed for use tion de I' evapotranspiration reele par thermographic JR, Journal in urban hydrology, Proc. 25th Int. Conf. on Radar Meteorology, de Recherches Atmospheriques, 14, pp. 113~ 125. Paris,lune 1991, pp. 844-847. REFERENCES 51

Kaufman, H., Reichart, B. and Hotzl, H., 1986. Hydrological research Workshop on Remote Sensing ofSnow and Evapotranspiration, in Peloponnesus Karst area by support and completion of Honolulu, Hawaii, NASA CP-2363, pp. 99-114. Landsat-thematic data, Proc. !GARSS '86 Symp., Zurich, Kouwen, N., Soulis, E.D., Pietroniro, A., Donald, 1. and Harrington, September 8-11,1986, ESA SP-254, 1, pp. 437-441. RA., 1993. Grouped response units for distributed hydrologic Kawata, Y. Kusaka. T. and Vena, S., 1988. Snowmelt runoff estima­ modelling, Journal of Water Resources Planning and tion using snow cover extent data and its application to optimum Management, 119 (3), pp. 289-305. control ofdam water level, Proc. !GARSS '88 Symp. Edinburgh, Kuittinen, R, 1989. Detennination of snow equivalent using NOAA Scotland, ESA SP-284, pp. 439-440. satellite images, gamma ray spectrometry and field measure~ Keller, M., 1987. Ausaperungskartierung mit Landsat-MSS Daten zur ments, IAHS Pub!. No. 186. Erfassung oekologische Einflussgnessen in Gebirge, Remote Kuittinen, R, 1992. Remote sensing for hydrology - progress and Sensing Series, No. 10, PhD Thesis. University ofZurich. prospects, Operational Hydrology Report No. 36, WMO-No. Kerr, YH., Assad, E., Freleaud, J.P., Lagourde, J.P. and Seguin, B., 773, Prepared for WMO, Geneva, Switzerland. 1987. Estimation of evapotranspiration in the Sahelian zone by Kuittinen, R. and Sucksdorff, Y., 1984. Detemrination ofwater depths the use of METEOSAT and NOAA AVHRR data, Advances in using aerial photography, Pub!. of the Water Research Institute Space Research, 7, pp. 161-164. No. 60, Helsinki, Finland. Kidd. C., 1988. Passive microwave rainfall monitoring over land. PhD Kumar, V.S., Haefner, H. and Seidel, K., 1991. Satellite snow Thesis, University ofBristol, UK, June 1988. cover mapping and snowmelt runoff modelling in Beas Kidd, C. and Barrett, E.C., 1990. The use of passive microwave Basin. Snow, Hydrology and Forests in High Alpine Areas imagery in rainfall monitoring, Remote Sensing Reviews, 4 (2), (Proceedings of the Vienna Symposium), IAHS Publication pp.415-450. No. 205, pp. 101-109. Kilonsky, B.I. and Ramage, C.S., 1976. A technique for estimating Kustas, WP., Cboudhury, B.I., Moran, M.S., Reginalo, R.I., Jackson, tropical open-ocean rainfall from satellite observations, Journal R.D., Gay, L.W. and Weaver, H.L., 1989. Determination ofsensi­ ofAppiied Meteorology, 15, pp. 972-975. ble heat flux over sparse canopy using thermal infrared data, Kienegger, E.H., 1992. Assessment of a Wastewater Service Charge Agricultural and Forest Meteorology, 44, pp. 197-216. by Integrating Aerial Photography and GIS, Photogrammetric Kustas, WP. and Daughtry, C.S.T., 1989. Estimation of the soil beat Engineering and Remote Sensing, 58 (11), pp. 1601-1606. flux/net radiation ratio from spectral data, Agricultural Forest Kimura, H. and Kodaira, N., 1992. Antenna pattern and penetration Meteorology, 49, pp. 205-233. measurements using spectrum analysers, IEEE Transactions on Kustas, WP. Rango, A. and Uijlenhoet, R., 1994. A simple energy Geoscience and Remote Sensing, 30, pp. 1153-1158. budget algorithm for the snowmelt runoff model, Water Kogan, F. and Sullivan, I., 1993. Development of global drought­ Resources Research, 30 (5), pp. 1515-1527. watch system using NOAAlAVHRR data, Adv. Space Res., 13 Kustas, WP. (and 36 co-autbors), 1991. An interdisciplinary field (5), pp. (5)219-(5)222. study of the energy and water fluxes in the atmosphere-bios­ Kondrat'ev, KYa, 1992. Earth observing system (EOS); ecological phere system over semiarid range lands: descriptions and priorities and observation planning, Issledovanie Zemli iz some preliminary results, Bulletin of the American Kosmosa, n. 3, May-June, pp. 96~106. Meteoroiogical Society, 72 (11), pp. 1683-1705. Kondratyev, KY., Melentyev, V.V, Rabiinovich, Y.I. and Shulginc, Lathrop, R.G. and Lillesand, T.M., 1987. Calibration of Thematic E.M., 1977. Passive microwave remote sensing ofsoil moisture, Mapper thermal data for water surface temperature mapping: Proceedings ofthe Eleventh Symposium ofRemote Sensing oj case study in the Great Lakes, RSE 22, pp. 297-307. Environment, pp. 1641-1661. Leaf, C.F. and Brink, G.E., 1973. Hydrologic simulation model of Konecny, G., and Vetrella, S., 1994. Earth observation in education Colorado SUbalpine forest, USDA Forest Service, Res. Pap. RM­ and training - EARSeL's viewpoint, International Jaurnal of 107, Fort Collins, CO. Remote Sensing, 15 (15), pp. 2970-2971. Leak, S.M. and Venugopal, G., 1990. Thematic Mapper thermal Kostov, KG., 1992. Passive microwave remote sensing of soil mois­ infrared data in discriminating selected urban features, Int. J. ture-experimental and modeling results, Presented XXIX Remote Sens., 11, pp. 841-857. Meeting ofthe Committee on Space Research, Washington, DC, Leavesley, G.H. and Stannard, L.G., 1990. Application of remotely August 28-8eptember 5, 1992. sensed data on a distributed parameter watershed model, Kostov, KG., 1993. Passive microwave remote sensing of soil mois­ Proceedings ofthe Workshop on Applications ofRemote Sensing ture-experimental and modeling results, Adv. Space Res., 13 (5), in Hydrology, Editors: Kite, G.W. and Wankiewicz,A., National pp. (5)105-(5)114. Hydrology Research Institute, Saskatoon, Saskatchewan, pp. 47­ Kotoda, K, Nakagawa, S., Kai, K, Yoshino, M.M., Takeda, K 68. and Seki, K., 1983a. Application of equilibrium evaporation Leberl, F., 1983. Photogrammetric aspects of remote sensing with model to estimate evapotranspiration by remote sensing tech­ imaging radar, Remote Sensing Reviews, 1 (1), Harwood nique in remote sensing of snow and evapotranspiration, Part Academic, New York. 11, Proc. 2nd US/Japan Workshop on Remote Sensing of Lei, G., Moore, R.K. and Gogineni, S.P., 1990. A method for meas­ Snow and Evapo-transpiration, Honolulu, Hawaii, NASA uring snow depth using fine-resolution radars, Remote Sens. CP-2363, pp. 115-123. Environ., 30, pp. 151-158. Kotoda, K, Nakagawa, S., Kai, K, Yoshino, M.M., Takeda, K. and Linsley, R.K., 1943. A single procedure for day-to-day forecasting of Seki, K, 1983b. Estimation ofregional evapotranspiration using runofffrom snowmelt, Trans. Am. Geophys. Union, Part III, pp. remotely sensed land surface temperature, Proc. 2nd US/Japan 62-67. 52 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

Lovejoy, S. and Austin, G.L., 1979a. The delineation of rain areas International Journal ofRemote Sensing, 15 (15), pp. 3073­ from visible and IR satellite data for GATE and mid-latitudes, 3075. Atmosphere-Ocean, 17, pp. 77-92. Matzler, C. and Schanda, E., 1983. Snow mapping with active Lovejoy, S. and Austin, G.L., 1979b. The sources of error in rain microwave sensors, January Report, University of Berne, amount estimating schemes for GOES visible and IR satellite Institute ofApplied Physics, Berne. data, Monthly Weather Review, 107, pp. 1048-1054. Matzler, C. and Schanda, E., 1984. Snow mapping with active Lucas, R.M. and Harrison, A.R., 1989a. A multispectral snow area microwave sensors, Int. 1 Remote Sens., 5, pp; 409-422. algorithm for operational 7-day snow cover monitoring, Remote May, B.R., 1988. Progress in the development o[PARAGON, Met. Sensing and Large-Scale Global Processes, Proceedings of the Mag., 117, pp. 79-86. lAHS ThirdInternational Assembly, Baltimore, MD, May, 1989, McCauley, J.F., Breed, e.S., Schaber, G.G. and Issawi, B., 1986. 1AHS Pub!. No. 186, pp. 161-166. Palaeodrainages of the eastern Sahara-shuttle imaging radar Lucas, R.M. and Harrison, AR., 1989b. Snow detection and descrip­ implications for a mid-Tertiary Trans African drainage system, tion using NOAA AVHRR imagery, Remote Sensing for Paper given at 20th Int. Symp. on Remote Sensing of Operational Applications. Proceedings of the 15th Annual Environment, Nairobi, Kenya, 4-10 Dec. 1986. Conference ofthe Remote Sensing Society, University ofBristol, McCauley, J.F., Schaber, G.G., Breed, C.S., Grotier, MJ., Haynes, Sept. 13-15, 1989, pp. 263-268. C.Y., Issawi, 8., Elachi, C. and Blom, R.. 1982. Subsurface Lucas, R.M. andHarrison, A.R., 1990. Snow observation by satellite: valleys and geoarcheology of the eastern Sahara revealed by areview, Remote Sensing Reviews, 4 (2), pp. 285-348. Shuttle radar, Scielice, 218, pp. 1004-1020. Lyons, J.G., Lunetta, R.S. and Williams, D.C., 1992. Airborne multi­ McCuen, R.H., 1989. Hydrologic analysis and design, Prentice Hall, spectral scanner data for evaluating bottom sediment types and 867pp. water depths ofthe S1. Marys River, Michigan, Photogrammetric McGinnis, D.E, Pritchard, l.A. and Wiesnet, D.R., 1975. Engineering andRemote Sensing, 58 (7), pp. 951-956. Determination of snow depth and snow extent from NOAA-2 Manavalan, P., Krishnamurthy, J., Manikiam, B., Adiga, S., satellite very high resolution radiometer data, Water Resources Radbakrishnan, K. and Chandrasekhar, M.G., 1993. Watershed Research, 1I (2), pp. 897-902. response analysis using digital data integration technique, Adv. McGinnis, D.F. and Schneider, S.R., 1978. Monitoring river ice Space Res., 13 (5), pp. 5(177)-5(180). breakup from space, Photogramm. Eng. andRemote Sensing, 44 Manikiam, B., Gowda, H.H., Manavalan, P., Jayaraman, Y. and (3), pp. 1597-1607. Chandrasekhar, M.G., 1993. Study of sediment dynamics using McKim, H.L., Cassell, E.A. and LaPotin, P.J., 1993. Water resource satellite remote sensing, Adv. Space Res., 13 (5), pp. (5)75-(5)78. modelling using remote sensing and object-oriented simulation, Marks, D., 1988. Climate, energy exchange, and snowmelt on Emerald Hydrological Processes, 7 (2), pp. 153-165. Lake watershed, Sierra Nevada, Unpublished PhD Dissertation, McKim, H.L., Marlar, T.L. and Anderson, D.M., 1972. The use of University ofCalifornia, Santa Barbara, CA, 149 pp. ERTS-l imagery in the national program for inspection ofdams, Martin, D.W. and Howland, M.R., 1982. Rainfall over the Arabian Sea CRREL Special Report 183, US Army Cold Region Research during the onset ofthe 1979 monsoon, Nature, 300, pp. 628-630. and Engineering Laboratory, Hanover, NH. Martinec, J., 1987. Importance and effects of seasonal snow cover, Meier, M.P., 1975. Applications of remote sensing techniques to the Large scale effects of seasonal snow cover. Proceedings of the study ofseasonal snowcover, Journal ofGlaciology, IS, pp. 251­ Vanconver Symposium, August, 1987. 1AHS (Int'!. Assoc. of 265. Hydrological Sci), Pub!. No. 166, pp. 107-120. Menenti, M., 1993. Understanding land surface evapotranspiration Martinec, 1. and Rango, A., 1981. Areal distribution of snow water with satellite multispectral measurements, Adv. Space Res., 13 equivalent evaluated by snow cover monitoring, Water Resources (5), pp. (5)89-(5)100. Research, 17 (5), pp. 1480-1488. Menenti, M., Azzali, S., Verhoef, W and van Swol, R., 1993. Mapping Martinec, 1. and Rango, A., 1986. Parameter values for snowmelt agroecological zones and time lag in vegetation growth by means runoffmodeling, J. Hydrol., 84, pp. 197-219. of fourier analysis of time series of NDVI images, Adv. Space Martinec, 1. and Rango, A., 1987. Interpretation andutilization ofareal Res., 13 (5), pp. (5)233-(5)237. snow cover data from satellites, Ann. Glacial., 9, pp. 166-169. Menenti, M. and Ritchie, J.c., 1994. Estimation of effective aerody­ Martinec,1. andRango, A. andMajor, E., 1983. The snowmelt-runoff namic roughness ofWalnut Gulch watershed with laser altimeter model (SRM) user's manual, NASA Ref. Pub!. 1100, measurements, Water Resources Research, 30 (5), pp. 1329­ Washington, DC. 1337. Martinec, J. Rango, A. and Roberts, R., 1994. Snowmelt-runoffmodel Miller, W.A., Shasby, M.B., Rhode, WG. and Johnson, G.R., 1982. (SRM) user's manual, Geographica Bernesia, P29, Department Developing inplace data bases by incorporating digital terrain of Geography, University of Bern, 65 pp. data into the Landsat classification process. Place Resource Mason, I., Harris, A., Birkett, C. Cudlip, Wand Rapley, C.G., 1990. Inventories: principles and practices, Proc. Workshop, August 9­ Remote sensing of lakes for the proxy monitoring of climate 14, 1981, University of Maine, Orono, Sponsored by the change, Remote Sensing and Global Change, Proc. 16thAnnual American Society ofPhotogrammetry. Com. of the Remote Sensing Society, University College Monteith, IL., 1965. Evaporation and environment, Symp. Soc. Exp. Swansea, September 19-21, Remote Sensing Society, Biol.1g., pp. 205-234. Nottingham, pp. 314-324. Moore, R., 1993a. Applications ofWeather RadarData to Hydrology Mather, P.M., 1994. Rapporteur's summary of session 3: Towards andWater Resources, WMO, Regional AssociationVI (Europe), common standards for training European education projects, Working Group on Hydrology, 26 pp. REFERENCES 53

Moore, R.J., 1993b. Real-timefloodforecasting systems: perspectives Pampaloni, P, Paloscia, S., Chiarantini, L., Coppo, 1'.., Gagliani, S. and and prospects, British-Hungarian Workshop on Flood Defense, Luzi, G., 1990. Sampling depth of soil moisture content by Budapest, September 6·10,1993,51 pp. radiometric measurement at 21 em wavelength: some experi­ Moore, R.J. and Jones, D.A., 1991. A river flood forecasting mental results, international Journal ofRemote Sensing, II (6), system for region-wide application. In: MAFF Conference of pp. 1085-1092. River and Coastal Engineers, Loughborough University, Papadakis, 1., Napiorkowski, J. and Schultz, G.A., 1993. Monthly England, July 8-10, 1991, Ministry of Agriculture Fisheries runoff generation by non-linear model using multispectral and and Food, 122 pp. multitemporal satellite imagery, Adv. Space Res., 13 (5), pp. Moore, R.J., Austin, R. and Carrington, D.S., 1993. Evaluation of (5)181·(5)186. FRONTIERS and Local Radar Rainfall Forecasts for use in Parry, D.E. and Piper, R., 1981. Application of remote sensing to Flood Forecasting Models, R&D Note 225, Research hydrogeology, A Survey of British Hydrogeology 1980, The Contractor: Institute of Hydrology, National Rivers Authority, Royal Society, London, pp. 61·73. 156 pp. Paterson, N.R and Bosschart, RA., 1987. Airborne geophysical Moore, R.I. and Bell, v., 1992. A gridsquarerainfall-runoffmodelfor exploration for groundwater, Groundwater, 25, pp. 41-50. use with weather radar data, Proc. Int. Workshop, November lI­ Peck, E.L., Carroll, T.R and Lipinski, D.M., 1990. Airborne gamma B. 1991, Lisbon, Portugal, Commission for the European radiation soil moisture measurements over short flight lines, Communities, in press, 9 pp. Symposium on the First 1SLSCP Field Experiment, February 7­ Morris, C.S., 1993. Measurement of Lake Level Variation Using 9, 1990. Satellite Altimetry, EOS, AGU Fall Meeting. Peck, E.S., Keefer. T.N., and Johnson, E.R, 1981. Strategies for using Myueui, R.B. and Choudhury, B.J., 1993. Synergistic nse of optical remotely sensed data in hydrologic models, NASA Rep. No. and microwave data in agrometeorological applications, Adv. CR·66729, Goddard Space Flight Center, Greenbelt, MD. Space Res., 13 (5), pp. (5)239-(5)248. Penman, H.L., 1948. Natural evaporation from open water, bare soil National Oceanic and Atmospheric Administration, US Department and grass, Proc. R Soc. Land., Ser. A., 193, pp. 120-145. ofCommerce, 1994. The Great Flood ofi993, Natural Disaster Penman. H.L., 1956. Estimating evaporation, Trans. Amer. Geophys. Survey Report. Union, 37(1). National Oceanic and Atmosphetic Administration, US Department Penney, M.E, Billard, B. and Abbot, R.H., 1989. LADS - the ofCommerce, 1993. Nile forecasting system - version 2, Edited Australian Laser Airborne Depth Sounder, International Journal by Edward van Blargan. ofRemote Sensing, 10, pp. 1463-1479. Negri, AJ., Adler, R.E, Nelkin, E.J. and Huffman, GJ., 1994. Petrie, G. and El Niweiri, A.E., 1992. The applicability of space Regional rainfall climatologies derived from special sensor imagery to the small-scale topographic mapping of developing microwave images (SSM/]) data, Bulletin of the American countries. A case study - the Sudan, ISRRS J. of Meteorological Society, 75 (7). Photogrammetry and Remote Sensing, 47, pp. 1-42. Nieuwenhuis, G.l., Smidt, E.H. and Thannissen, H.A.M., 1985. Prabhakara, e., Dalu, G., Liberti, G.L., Nucciarone, lJ. and Suhasini, Estimation of regional evapotranspiration of arable crops from R., 1992. Rainfall estimation over oceans from SM:MR and thermal infrared images. International Journal of Remote SSMII microwave data, Joumal ofApplied Meteorology, 31(6), Sensing, 6, pp. 1319-1334. pp.532-552. Oh, Y., Sarabandi, K. and Ulaby, ET., 1992. An empirical model and Pratt, D.A., Foster, SJ. and Ellyet, e.D., 1980. A calibration proce­ an inversion technique for radar scattering from bare soil dure for fourier series thennal inertia models, Photogrammetric surfaces, IEEE Transactions on Geoscience and Remote Sensing, Engineering and Remote Sensing, 46, pp. 529-538. 30, pp. 370-381. Price, le., 1980. The potential of remotely sensed thermal infrared O'Neill, P.E., Jackson, TJ., Chauhan, N.S. and Seyfried, M.S., 1993. data to infer surface soil- moisture and evaporation, Water Microwave soil moisture estimation in humid and semiarid Resources Research, 16, pp. 787-795. watersheds,Adv. Space Res., 13 (5), pp. (5)115-(5)118. Price, J.e., 1982. Estimation ofregional scale evapotranspiration through Ormsby, J.P., Blauchard, BJ. andBlanchard,AJ., 1985. Detection of analysis of. satellite thermal infrared data, IEEE Transactions on lowland flooding using active microwave systems, Photogram. Geoscience and Remote Sensing, GE-20, pp. 286-292. Eng. Remote Sensing, 51, pp. 317-328. Priestley, c.H.B. and Taylor, R.I., 1972. On the assessment ofsurface Ormsby, J.P. and Engman, E.T., 1993. Hydrosat: an instrument plat­ heat flux and evaporation using large-scale parameters, Monthly form for hydrology, Adv. Space Res., 13 (5), pp. (5)101-(5)104. Weather Review, 100, pp. 82-92. Ostrem, G., Anderson, T. and Odegaard, H., 1991. Operational use of Pulich, w., Jr. and Hinson, J.M., 1992. Methodology for classifying satellite data for snow inventory and runoff forecast, Satellite landcover and habitats from Landsat TM imagery of the Texas Hydrology, American Water Resources Association, coastal zone, Proceedings of SPIE, the Int. Society for Optical Minneapolis, MN, pp. 230-234. Engineering, 1930 (2), Bellingham, WA., pp. 855-868. Ottle, e., Vidal-Madjar, D. and Girard, G., 1989. Remote sensing Raffy, M. and Becker, E, 1985. An inverse problem for remote sens­ applications by hydrological modeling, Journal ofHydrology, ing in the thermal infrared bands and its solution, Journal of 105, pp. 369-384. Geophysical Research, 90, pp. 5809-5819. Palecki, MA and Barry, RG., 1986. Freeze-up and break-up oflakes Ramamoorthi, A.S., 1983. Snowmelt runoff studies using remote sens­ as an index oftemperature changes during the transition seasons: ing data, Proc. ladian Acad. Sci., 6, pp. 279·286. A case study for Finland, Journal of Climate and Applied Ramamoorthi, A.S., 1987. Snow cover area (SCA) is the main factor Meteorology, 25, pp. 893-902. in forecasting snowmelt runoff from major basins, Large Scale 54 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

Effects ofSeasonal Snow Cover (Proceedings ofthe Vancouver Richard, VW., Kammerer, lEo and Wallace, H.B., 1988. Rain Symposium), IAHS Pub!. No. 166, pp. 187-198. backscatter measurements at millimeter wavelength, IEEE Rambal, S., Lacaze, B., Mazurek, H. and Debussche, G., 1985. Transactions on Geoscience and Remote Sensing, 26 (3). Comparison ofhydrologically simulated and remotely sensed Richards, JA, Woodgate, P.w. and Skidmore,A.K., 1987. An expla­ actual evapotranspiration from some Mediterranean vegetation nation ofenhanced radar backscattering from flooded forests, Int. fonnulations, International Journal of Remote Sensing, 6, pp. J. Remote Sens., 8, pp. 1093-1100. 1475-1481. Ringrose, S. and Matheson, W., 1991. A Landsat analysis of range Randall, D.A. and Tjemkes, S., 1991. Clouds, the Earth's radiation conditions in the Botswana Kalahari drought, Int. 1. Remote budget, and the hydrologic cycle, Global andPlanetary Change, Sens., 12, pp. 1023-1052. 4 (1/3), pp. 3-9. Riou, e., Itier, B. and Seguin, 8., 1988. The influence of surface Rango, A., 1985. The snowmelt runoff model, Proceedings ofthe ARS roughness on the simplifiedrelationship between daily evapora­ National Resources Modeling Symposium, USDA-ARS-30, tion and surface temperature, International Journal ofRemote Pingree Park, CO, pp. 321-325. Sensing, 9 (9), pp. 1529-1533.

Rango; A. t 1992. Worldwide testing of the snowmeltrunoff model Ritchie, l.e. and Cooper, C.M., 1991. An algorithm for using Landsat with applications for predicting the effects of climate change, MSS for estimating surface suspended sediments, Water Nordic Hydrology, 23, pp. 155-172. Resources Bulletin, 27, pp. 373-379. Rango, A., 1993. Snow hydrology processes and remote sensing, Ritchie, J.C., Evans, D.L., Jacobs, D., Everitt, l.H. and Weltz, M.A., Hydrological Processes, 7, pp. 121-138. 1993b. Measuring Canopy Structure with an Airborne Laser Rango, A., 1994a. Application ofremote sensing methods to hydrol­ Altimeter, Trans. ofthe ASAE, 36 (4), pp. 1235-1238. ogy and water resources, Hydrological Sciences Journal, 39 (4), Ritchie, J.e., Grissinger, E.H., Murphey, J.B. and Garbrecht, J.D., pp.309-320. 1994. Measuring channel and gully cross-sections with an Rango, A., 1994b. Applications ofremote sensing by satellite, radar airborne laser altimeter, Hydrological Processes, 8, pp. 237-243. and other methods to hydrology, Operational Hydrology Report Ritchie, J.e., Jackson, T.J., Garbrecht, J.D., Grissinger, E.H., No. 39, WMO-No. 804, WMO, Geneva, 33 pp. Murphey, J.B., Everett, J.H., Escobar, D.E., Davis, M.R and Rango, A and Martinec, 1, 1994. Areal extent ofseasonal snow cover Weltz, M.A., 1993a. Studies using an airborne laser altimeter to in a changed climate, Nordic Hydrology, 25, pp. 233-246. measure landscape properties, Hydrological Sciences Journal, Rango, A., Martinec,J., Foster, 1 and Marks, D., 1983. Resolution in 38 (5), pp. 403-416. operational remote sensing of snow cover, Hydrological Ritchie, J.C., Schiebe, P.R, Cooper, e.M. and Harrington, J.A., Jr., Applications of Remote Sensing and Remote Sensing Data 1992. Landsat MSS studies of chlorophyll in sediment domi­ Transmission, Proceedings of Hamburg Symposium, August, nated lakes, Proceedings of IGARRS '92 Symposium, Vo!' II, 1983. IAHS Publ. No. 145, pp. 371-382. Clear Lake City, TIC, pp. 1514-1517. Rango, A., Salomonson, V.V.·and Foster, J.I.., 1975. Employment of Roberts, G., France, M., Johnson, R.e., and Law, J.T., 1993. The satellite snow cover observations for improving seasonal runoff analysis ofremotely sensed images ofthe Balquidder catchments estimates, Operational Applications of Satellite Snow Cover for estimation of percentages of land cover types, Journal of Observations, NASA SP-391, Washington, DC, pp. 157-174. Hydroiogy, 145, pp. 259-265. Rango, A., Salomonson, V V. and Foster, J.L., 1977. Seasonal stream­ Rodda, J.e., Pieyns, SA, Sehmi, N.S. and Matthews, G., 1993. flow estimation in the Himalayan region employing Towards a world hydrological cycle observing system, meteorological- satellite snow cover observations, Water Hydroiogical Sciences Journal, 38 (5), pp. 373-378. Resources Research, 13, pp. 109-112. Rohlfing, R., 1993. Radar data based optimization of control strate­ Rango, A. and V van Katwijk, 1990. Development and testing of a gies in an urban drainage system, Advances in Radar Hydrology, snowmelt-runoff forecasting technique, Water Resources Proc. Int. Workshop, November 11-13,1991, Lisbon, Portugal, Sufietin, 26, pp. 135-144. Commission for the European Communities, in press. Rangoonwala, A., Ahmed, S., Nasir, A., Raouf, A. and Shahab, H., Roller, N.E., 1984. Effective integration of data sources for optimiz­ 1993. Quasi-operational use ofNOAAJAVHRR vegetation index ing reSOurce inventories, 18th Int. Symp. on Remote Sensing of to monitor vegetal cover and crop growth, Adv. Space Res., 13 Environment, Paris, France. (5), pp" (5)265-(5)268. Rosema, A, Bijleveld, l.R, Reiniger, P., Tassone, G., Blyth, K. and Rao, V.R., 1989. Space and Flood Management, Special Current Gurney, R.I., 1978. TELL-US, a combined surface temperature, Event Session, International Astronautical Federation, 40th soil moisture and evaporation mapping approach, Proceedings Congress, Malaga, Spain, 7 pp. of the 12th International Symposium on Remote Sensing of Reginato, RJ., Jackson, R.D. and Pinter, Jr., P.J., 1985. Environment, Ann Arbor, Michigan, pp. 2267-2276. Evapotranspiration calculated from remote multispectral and Rott, H., 1986. Prospects of microwave remote sensing for snow groud station meteorological data, Remote Sensing Envbvn., 18, hydrology, Hydrologic Applications of Space Technology pp.75-89. (Proceedings ofthe CocoaBeachWorkshop), IAHS Publication Reinhold, RJ. and Linthurst, RA., 1975. Use of remote sensing for No. 160, pp. 215-233. mapping wetlands, Am. Soc. Civil Eng. 1. Tramp. Eng., 101, TE­ Rott, H., Domik, G., Metzler, C., Miller, H. and Lenhart, K.G., 1985. 2, pp. 189-198. Study on use and characteristics of SAR for land snow and ice Reutov, E.A. and Shutko, A.M., 1986. Prior knowledge based soil applications, ESA Contract Report No. 5441183/D/IM/(SC). moisture determination by microwave radiometry, Soviet Journal Rott, H. and Kunzi, T., 1983. Properties of the global snowcover and ofRemote Sensing,S (I), pp. 100-125. snow-free terrain from the NIMBVS-7 SMMR. First year data REFERENCES 55

set. Proc. of the Specialist Meeting on Microwave Radiometry Shutko, A.M., 1985. Radiometry for farmers, Sciellce ill the USSR, 6, and Remote SensingApplications, March 1-2, 1983, Rome, Italy. pp.97-113. Ryall, G. and Conway, BJ., 1993. Automated precipitation Ilowcast­ Shutko, A.M., 1986. Microwave radiometry of water surface and ing at the UK Met Office. First European Conference on grounds, Publishing House 'Nauka', Moscow, Russian Applications of Meteorology, September 27-october 1, 1993, Federation, 192 pp. (English translation). University of Oxford, UK, 4 pp. Shutko, A.M., 1987. Remote sensing of the waters and lands via Sabins, EE, Jr., 1983. Geologic interpretation of space shuttle radar microwave radiometry (the principles ofmethod, problems feasi­ images ofIndonesia. AmericanAssociation PetroleHm Geology ble for solving, economic use), in: Remote Sensing and Its Bulletin, 67, pp. 2076. Impact on Developing Countries, Pontificia Academia Salomonson, v.v. (compiler), 1983. Water resources assessment, Scientiarum, Scripta Varia-68, Vatican City, pp. 413-441. Colwell, R.N. (ed.) Manual ofRemote Sensing, 2nd Edition, Shutko, A.M., 1990. Offer on hardware, software and services on survey American Assoc. Photogrammetry, Ch. 29, pp. 1497-1570. of soil, vegetation and water areas - from aircraft, Institute of Saxena, N., 1982. Hydrologic surveys of the 1990s, Allgemeine Geoinfonnatics, Nongovernmental Center for Research and Vennessungsnachrichten, 6, pp. 227-233. Institute of Radioengineering and Electronics, Academy of Schneider, S.R, McGinnis, D.E andPritchard, J.A., 1979. Delineation Sciences, Russian Federation, IGI, IRE, Moscow, 16 pp. of drainage and physiographic features in North and South Simpson, J., Adler, RE and North, G.R, 1988. A propused tropical Dakota using NOAA-5 infrared data, Satellite Hydrolagy (Edited rainfall measuring mission (TRMM) satellite, American by Deutsch, D. Wiesnet and A. Rango), AWRA, Minneapolis, Meteorological Society, 69(3), pp. 278-295. Minnesota, pp. 324-329. Sippel, S.l., Hamilton, S.K., Melack, J.M. and Choudhury, B.J., 1993. Schneider, S.A., 1989. Global wanning -are we entering the green­ Passive Microwave Satellite Observations of Inundation in the house century? Sierra Club Books, San Francisco, 317 pp. PantanalWetland ofBrazil, EOS, AGU Fall Meeting. Schneider, S.R., McGiunis, D.F. aud Gatlin, J.A., 1981. Use of Smith, lA and Krajewski, W.E, 1991. Estimation of the mean field NOAAlAVHRR visible and near-infrared data for land remote bias ofradar rainfall estimates, Journal ofAppliedMeteorology, sensng, NOAA Technical Report NESS 84, Washington, DC 30 (4), pp. 397-412. 20233, USA, 48 pp. Soliman, M.M. and Zaki, A.M., 1986. Verification of Lake Aswan Schultz, G.A., 1993. Hydrological modeling based onremote sensing High Dam sediment model by Landsat imagery, AJ. Johnson informatiou,AdvancedSpace Research, 3 (5), pp. (5)149-(5)166. (Editor), Hydrologic Applications ofSpace Technology, lAHS Schultz, GA., 1988. Remote sensing in hydrology, Journal of Pub!. No. 160, pp. 245-248. Hydrology, 100, pp. 239-265. Soares, J.v., BernardR, Taconet, 0., Vidal-Madjar, D. andWeill, A., Schultz, G.A. and Barrett, E.C., 1989. Advances in remote sensingfor 1988. Estimation ofbare soil evaporation fn?m microwave meas­ hydrology and water resources management, International urements, J. Hydrolagy, 99, pp. 281-296. Hydrological Programme, UNESCO Technical Documents on Soer, GJ.R., 1980. Estimation ofregional evapotranspiration and soil Hydrology, IHP-llI Project 5.1, Paris, 102 pp. moisture conditions using remotely sensed crop surface temper­ Scofield, R.A., 1991. Operational estimation of precipitation from atures, Remote Sensing ofEnvironment, 9, pp. 27-45. satellite data, Global and Planetary Change, 4(1/3), pp. 79-86. SparLIGHT, 1993. Bay Watch from Space, The SPOT Image Scofield, R.A. and Oliver, V.l, 1977. A scheme for estimating convec­ Corporation Newsletter, p.7. tive rainfall from -satellite imagery, NOAA Technical Srivastav, S.K. and Singh, R.P., 1991. Microwave radiometry ofsnow­ Memorandum, NESS-86, Washington, DC, 47 pp. covered terrains, Int. J. Remote Sens., 12, pp. 2117-2131. Seginer, I., 1971. Wind effect on evaporation rate, Journal ofApplied Stewart, E.J., 1993. Application of weather radar to hydrological Meteorology, (4). design studies, Advances in Radar Hydrology, Proc. Int. Seguin, B., Courault, D. and Guerif, M., 1993. Satellite thennal Workshop, November 11-13, 1991, Lisbon, Portugal, infrared data applications in agricultural meteorology, Adv. Space Commission for the European Communities, in press. Res., 13 (5), pp. (5)207-(5)217. Stiles, W.H., Ulaby, F.T. and Rango, A., 1981. Microwave measure­ Seguin, B. and !tier, B., 1983. Using midday surface temperature to ments ofsnowpack properties, Nordic Hydrol., 12, pp. 143-166. estimate daily evaporation from satellite thennal IR data, Stoffelen, A. and Anderson, D.L.T., 1993. Wind retrieval and ERS-1 International Journal ofRemote Sensing, 4 (2), pp. 371-383. scatterometer radar backscatter measurements, Adv. Space Res., Shcheglova, G.P. and Chernov, V. Yu, 1982. Seasonal snow line within 13 (5), pp. (5)53-(5)60. the Fergana Basin and possibilities for using it in hydrological Strumph, R.P. and Taylor, M.A., 1988. Satellite detection of bloom forecasting. Izucheniye Gidrologicheskoga Tsikla and pigment distribution in estuaries, Remote Sensing of Aerokomicheskimi Metodami (Eds. K. Ya Kondratyev and Yu V. Environment, 24, pp. 385-404. Kurilova), Academy of Sciences, Russian Federation, Soviet Szekielda, K.B., 1988. Satellite moniton'ng ofthe Earth, John Wiley Geophysical Conunittee, Moscow, pp. 36-41. & Sons, New York, pp. 258-265. Shuman, C.A., Fitzpatrick, J., Mershon, G., Dibb, l. and Stearns, C.R, Taconet, 0., Bernard, R. and Vidal-Madjar, D., 1986. Evapo­ 1993. Identification of multiple hoar-development episodes in transpiration over an agricultural region using a surface Central Greenland with SSM!I brightness temperature trends and flux/temperature model based on NOAA-AVHRR data, Journal field observations, EGS, AGU Fall Meeting. ofClimate and Applied Meteorology, 25, pp. 284·307. Shutko, A.M., 1982. Microwave radiometry oflands under natural and Taconet, O. and Vidal-Madjar, D., 1988. Applications of a flux algo­ artificial moistening, IEEE Transactions on Geoscience and rithm to a field-satellite campaign over vegetated area, Remote Remote Sensing, 0E-20 0), pp. 18-23. Sensing ofEnvironment, 26, pp. 227-239. 56 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

Tarpley, J.D., 1979. Estimating incident solar radiation at the surface tropical waters in the Gulf of Guinea, Remote Sensing Environ., from geostationarysatellite data, J. Appl. Meteorol., 18, pp. 1172­ 7, pp. 235-248. 1181. Viter, V.v., 1993. The ALMAZ program and the development of Tarpley, lD., 1984. Monitoring global vegetation from meteorologi­ remote sensing techniques to study the Earth from space, Space cal satellite data, Proc. 18th Int. Symp. on Remote Sensing of Builetin, 1(1). Environment, Paris, France. Walton, M.L., Smith, JA. and Shedd, R.C., 1990. Remote sensing of Thiao, W., Scofieid, R.A. and Robinson, J., 1993. The reiationship rainfall with NEXRAD, Hydraulics/Hydrology ofArid Lands, between water vapor plumes and extreme rainfall events during American Society ofCivil Engineers, New York,pp. 304-310. the summer season, NOAA Tech. Report, NESDIS-67, Walsh, PD., 1993. New Directions in RadarHydrology, WMO RAVI Washington, DC. Workshop on Requirements and Applications ofWeather Radar Thorn, A.S. and Oliver, HR., 1977. On Penman's equation for esti­ Data in Hydrology andWater Resources, 13 pp. mating regional evapotranspiration, Quarterly Journal Royal Ward, R.c., 1975. Principles ofhydrology, 2nd Ed., McGraw Hill, MeteorologicalSociety, 103, pp, 345-357. London, 367 pp. Thunnissen, H.A.M. and Nieuwenhuis, a.I.A., 1990. A simplified Warren, S.G., 1982. Optical properties of snow, Rev. Geophys. Space method to estimate regional 24-h evapotranspiration from ther­ Phys., 20, pp. 67-89. mal infrared data, Remote Sensing of Environment, 31, pp. Weltz, M.A., Ritchie, J.e. and Fox, HD., 1994. Comparison of laser 2II-225. and field measurements of vegetation height and canopy cover, Torlegard, K., 1994. Costlbenefit aspects ofremote sensing education, Water Resources Research, 30 (5), pp.1311-1319. International Journal ofRemote Sensing, 15 (15), pp. 3101­ Werth, L.P. and Meyer, M.P., 1981. A comparison of remote sensing 3109. techniques for Minnesota wetlands classification, Satellite Travaglia, C, 1989. Drainage analysis, Remote Sensing Applications Hydrology, AWRA, Minneapolis, MN, pp. 492-497. to Land Resources, Report of the 14th UNIFAO International Wiesnet, D.R andDeutsch, M., 1985. A new application ofNIMBUS­ Training Course, in cooperation with the Government of Italy, 7 CZCS: Delineation of the 1983 Parana River flood in South Rome, Italy, RSC Series 54, pp. 101-106. Amelica, Proc. ofthe American Society ofPhotogrammatry, 45th Trevett, lW., 1986. hnaging Radars for Resources Surveys, Remote Annu. Meet., Washington, DC, pp. 746-754. Sensing Applications Services, Chapman & Hall, London, pp. Wilheit, T.T., Chang, A.T. and Chin, L.S., 1991. Retrieval of montbly 145-149. rainfall indices from microwave radiometric measurements using Tucker, C.I., Elgies, J.H., McMnrtrey, ill, I.E. and Fran, c.J., 1979. probability distribution functions, Journal ofAtmospheric and Monitoring com and soybean crop development with hand-held Oceanic Technology, 8 (1), pp. 118-136. radiometric spectral data, Remote Sensing Environ., 8, pp. 237­ Willemont, J., 1994. PROMEfE, a. multimediatraining programmein 248. remote sensing, International Joumal ofRemote Sensing, 15 (15), Ullriksen, P., 1985. Profiler genom snotacket registrerade i mars med pp.3091-3093. LTH's, 900 IvIhz impulsradarsystem, Vannet i Norden, 4, pp. 47­ Wilson, w.J., 1m, E., Li, F. and Parks, G.S., 1989. Deveiopment of a 70. dual-frequency airborne rain radar, International Geoscience and UNESCOIWMOIICSU, 1993. Report of the UNESCOIWMOIICSU Remote Sensing Symposium, 3, pp. 1496-__. International Conference on Hydrology, Towards the 21st Wood, BF., Lin, D.S., Mancini, M., Thongs, D., Troch, P.A., Jackson, Century: Research and operational needs, March 22-26, 1993, TJ., Famiglietti, J.S., and Engman, E.T., 1993. Intercomparisons Paris. between passive and active microwave remote sensing, and van de Griend, A.A., Camillo, P.J. and Gnmey, R.I., 1985. hydrological modeling for soil moisture, Adv. Space Res., 13 (5), Discrimination of soil physical parameters, thermal inertia and pp. (5)167-(5)176. soil moisture from diurnal surface temperature fluctuations, Water World Meteorological Organization, 1991. Executive Council Panel of Resources Research, 21, pp. 997-1009. Experts on Satellites, Ninth Session. Geneva, March 18-22, 1991, van de Griend, A.A and van Boxel, lH., 1989. Water and surface Final Report. energy balance model with a multilayer canopy representation for World Meteorological Organization, 1992a. Application of satellite remote sensing purposes, Water Resour. Res., 25, pp. 949-971. technology, Annual Progress Report, SAT-1O, Technical Vane, G. and Goetz,A.F.H, 1988. Terrestrial imaging spectroscopy, Document, WMOrrD No. 569. Remote Sensing ofEnvironment, 24 (I), pp. 1-30. World Meteorological Organization, 1992b. Satellite ground receiving Vaughan, RA., 1994. Rapporteur's summary of session 1: Remote equipment in WMO regions, Status Report SAT-I I, Tecbnical sensing education and training in Europe today, International Document, WMOrrD No. 576, 26 pp. Joumal ofRemore Sensing, 15 (15), PP. 3017-3019. World Meteorological Organization, 1993a. Executive Council Panel Vershinina, L.K., 1985. The use ofaerial gamma surveys ofsnowpack ofExperts on Sateilites, Final Report, Geneva, March 9-10. for spring snowmelt runoffforecasts, Hydrological Applications World Meteorological Organization, 1993b. Application of Satellite of Remote Sensing and Remote Data Transmission, Proc. Tecbnology, Annnal Progress Report, SAT- 12, WMO/TD No. Hamburg Symp.,August 1985, lAHS Pubi. No. 145, pp. 4II-420. 628. Verwom, H., 1993. Urban applications of weather radar, Advances in World Meteorological Organization, 1994. CBS Working Group on Radar Hydrology, Proc. Int. Workshop, November II-D, i991, Satellites, First Session, Geneva, March 7-11, 1994, Final Lisbon, Portugal, Commission for the European Communities, in Report. press. World Meteorological Organization, 1995. Progress Report on Action Viollier, M., Deschamps, P.Y. and Lecomte, P., 1978. Airborne remote From CBS WGSAT-1: Education and Training Rapporteurs, sensing ofchlorophyll content under cloudy sky as applied to the Marcb 1995. REFERENCES 57

Zall, L. and Russell, 0., 1981. Groundwater exploration programs in Zevenbergen, A.W. and Rango, A., 1992. Applying Landsat Africa, Satellite Hydrology, American Water Resources imagery for groundwater development in Egypt. Geocarto Association, Minneapolis, MN, pp. 416-425. lntemational-A Multidisciplinary Journal ofRemote Sensing, Zebker, H.A., Madsen, S.N., Martin, 1., Wheeler, K.B., Miller, T., Lou, 7 (3), pp. 41-5 I. Y., Alberti, G., Vetrella, S. and Cuci,A., 1992. TheTOPSAR inter­ Zuzel, 1.E and Cox, L.M., 1975. Relative importance ofmeteorolog­ ferometric radar topographic mapping instrument, IEEE ical variables in snow melt, Water Resources Research, 11, pp. Transactions on Geoscience and Remote Sensing, 30, pp. 933~940. 174-176. APPENDIX I CITED LITERATURE ON APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

A. RECENT TEXTBOOKS ONTHE APPLICATIONOF REMOTE SENSING IN HYDROLOGY

1. Remote Sensing in Hydrology (Engman and Gurney, 1991) 7. Remote Sensing ofIce and Snow (Hall and Martinec, 1985) 2. Remote Sensing and Geographical Tnformation Systems for 8. The Use of Satellite Data in Rainfall Monitoring (Barrett. Resource Management in Developing Countries (Edited by 1981) Belward and Valenznela, 1991) 9. Applications of Remote Sensing in AgriCltlture (Edited by 3. Microwave Imaging Techniques (Bernard and Subbaram, Steveu, 1990) 1991) 10. Introduction to Environmental Remote Sensing, 3rd ed. 4. Satellite Remote Sensing for Hydrology and Water (Barrett and Curtis, 1992) Management (Barrett, 1990) 11. Spaceborne Radar Remote Sensing: Applications and 5. Processing and Analysis ofCommercial Satellite Image Data Techniques (Elachi, 1987) of the Nuclear Accident Near Chernobyl, USSR (Sadowski 12. Manual of Remote Sensing, 2nd ed. (Edited by Colwell, and Covington, 1987) 1983) 6. Microwave Radiometry of Water Surface and Grounds (Shutko, 1986)

B. TECHNICAL REPORTS ONTHEAPPLICATION OF REMOTE SENSING INHYDROLOGY

1. Applications ofRemote Sensing by Satellite, Radar, and Other 15. Hydrological Assessment ofSubSaharan Africa. TheWorld Bank Methods to Hydroiogy. WMO-No. 804 (Rango, 1994) Reports. 2. Applications ofRemote Sensing to Hydrology. WMO-No. 773 16. 1986/7 Sudan Crop Estimation Assisted by Remote Sensing (Kuittineu, 1992) Techniques. US AID (Hassan, 1988) 3. Application ofSatellite Data for Estimation ofPrecipitation. 17. Nile Forecasting System. NOAA Technical Manual (Edited by WMOfID-300 (Kuittinen, 1992) Van Blargan, 1993) 4. Executive Council Panel of Experts on Satellites. WMO Final 18. The Relationship Between Water Vapor Plumes and Extreme Report (1993) Rainfall Events During the Summer Season. NOAA Tech Report 5. Executive Council Panel of Experts/CBS Working Group on NESDIS 67 (Thiao €I al., 1993) Satellites, Tenth Session. WMO Final Report (1992) 19. Satellite Derived Rainfall Estimates and Propagation 6. Executive Council Panel ofExperts on Satellites, Ninth Session. Characteristics Associated with Mesoscale Convective Systems WMO Final Report (1991) (MeSs). NOAA Tech Memo NESDIS 25 (Iuying and Scofield, 7. Remote Sensing for Hydrology - Progress and Prospects. 1989) WMO-No. 773 (Kuittinen, 1992) 20. A Geographic lnfonnation System to Support State-wide 8. Application ofSatellite Technology. Annual Progress Report Hydrologic and Nonpoim Pollution Modeling. Maryland DOT WMO SAT 10 (1992) (Ragan, 1991) 9. Satellite Ground Receiving Equipment in WMO Regions. Status 21. Feasibility ofUsing Remote Sensing Platfonns as anAidto Water Report WMO SAT 11 (1992) Quality Monitoring in the Tennessee Valley (Dierberg, 1992) 10. Advances in Remote Sensingfor Hydrology and Water Resources 22. Applying Remote Sensing and GIS Techniques in Solving Management. UNESCO IHP-Ill Project 5. 1 (Schultz and Barrett, Rural County Infonnation Needs. Final Report (Johannsen etal., 1989) 1992) 11. Water and Environment Key to Africa's Development. llIE 23. A Remote Characterization System for Subsurface Mapping of Report Series 29 (1993) Buried Waste Sites (Sandness et al., 1992) 12. Institute for Remote SensingApplications. Joint ResearchCentre 24. Space-based Remote Sensing ofthe Earth: A Report to Congress. Aunual Report (1992) NOAA, USDC, NASA, US AID (1987) 13. Remote Sensing Applications to Water Resources. FAO RSC 25. Scientific, Military, and Commercial Applications ofthe lAND­ Series 50 (1989) SAT Program. Conunittee on Science, Space, and Technology 14. Natural Resource and Environmental InfomJationfor Decision­ (1991) making. The World Bank Report (Edited by Hassan and Hutchiuson, 1992)

C. PROCEEDINGS OF CONFERENCES/SYMPOSIA ONTHEAPPLICATION OF REMOTE SENSING IN HYDROLOGY

1. Remote Sensing for Oceanography, HydrologyAndAgriculture. 3. Eastern Snow Conference. CRREL Special Report 90-44 Advances in Space Research. COSPAR. (Edited by Gower, (Carroll, 1990) Salomoson, Engman, Onnsby and Gupta, 1993) 4. HydrauIicsIHydrology ofArid Lands. HY and IR Div.!ASCE(1990) 2. Satellite Hydrology. AWRA (Edited by Deutsch, Wiesnet and 5. ISLSCP Field Experiment. American Meteorological Society Rango, 1979) (1990) APPENDIX I 59

6. Western Snow Conference (1989,1987,1983) 15. Remote Sensing Water Resources. IAH 7. Hydrometerology. American Meteorological Society (1990) 16. International Symposium on Mapping and Geographic 8. Large Scale Effects of Seasonal Snow Cover. Vancouver Information Systems. ASTM Special Tech. Publ. No. 1126, (IAHS Publ. No. 166) ASTM, USGS (1992) 9. SPIE. The International Society for Optical Engineering 17. Irrigation and Drainage: Saving a Threatened Resource - In (1992, 1983) Search ofSolutions, ASCE (1992) 10. American Congress on Surveying and Mapping. American 18. International Conference on Engineering, Construction, and Society for Photograrnmetry and Remote Sensing (1987) Operations in Space 111. ASCE (1992) 11. Remote Sensing Applications for Water Resources 19. Civil Engineering Applications of Remote Sensing and Management. US Army Corps of Engineers (1985) Geographic Information Systems. ASCE, US Army Corps of 12. International Conference on Hydrology. Towards the 21st Engineers (1991) Century: Research and Operational Needs. UNESCO! 20. Earth and Atmospheric Remote Sensing. SPIE, CREOL! WMOIICSU (1993) University of Central Florida (1991) 13. Weather Analysis and Forecasting/Flash Floods. American 21. Remote Sensing, Global Monitoring for Earth Management. Meteorological Society (1993) IGARSS (1991) 14. Hydrological Applications of Remote Sensing and Remote 22. Optical Remote Sensing of the Atmosphere. Optical Society Data Transmission. Hamburg (lAHS Pub!. No. 145) (1983) ofAmerica, American Meteorological Society (1991)

D, JOURNALS CONTAINING ARTICLES ONTHEAPPLICATION OF REMOTE SENSING IN HYDROLOGY

1. Earlh Space Review 17. Photogrammetric Engineering andRemote Sensing 2. Space Bulletin 18. International Journal ofRemote Sensing 3. Remote Sensing Reviews 19. Water Resources Planning andManagement, ASCE 4. Nordic Hydrology 20. Soviet Journal ofRemote Sensing, Translation ofIsslendovannie 5. Geographic Information Systems andWater Resources, AWRA ZemU iz Kosmosa 6. Remote Sensing ofEnvironment 21. Akusticheskii Zurnal 7. Hydrological Sciences 22. Journal ofAppliedMeteorology 8. Agricultural Meteorology 23. Global andPlanetary Change 9. IEEE Transactions on Geoscience andRemote Sensing 24. South African Journal ofMarine Science 10. Water Resources Bulletin, AWRA 25. Journal ofAtmospheric and Oceanic Tec1mQlogy 11. Hydrological Processes 26. Meteorological Magazine 12. Geocarto International 27. National Technical Infonnation Service 13. Nature andResources 28. Bulletin ofthe AmericanMeteorological Society 14. Geophysical Research Letters 29. Monthly Weather Review 15. Journal ofHydrology 30. HydraulicslHydrologyofAridLands, ASCE 16. National Weather Digest, NWA, Satellite 31. Water Resources Research

E. OTHER DOCUMENTS CONTAININGARTICLESIINFORMATION THEAPPLICATION OF SENSING INHYDROLOGY

1. SPOTLight. The SPOT hnage Corporation Newsletter 4. National Operational Hydrologic Remote Sensing Center, NOAA 2. EOSATNotes. Earth ObservationSatelIite Company (Newsletter) 3. EOSATLandsat Data Users Notes 5. COSPAR Information Bulletin

AUTHORS REPORTING OPERATIONAL OR POTENTIALLY OPERATIONALAPPLICATIONS OFREMOTE SENSING IN HYDROLOGY

a - Precipitation b - Soil Moisture and c - Evapotranspiration e - Ice on Land I. Adler et al., 1991 Groundwater I. Dnnkel and Vadasz, 1993 I. WMO,1993b 2. Barrett, 1993 I. Carroll and Schaake, 1983 2. Menenti, 1993 3. Barrett and Harrison, 1986 2. Corbley, 1994 f - S1l1jace Water 4. Csistar and Bonta, 1993 3. Dutartre el a!., 1993 d-Snow I. Corbley, 1994 5. D'Souza el a!. , 1990 4. EI-Baz, 1993 I. Carroll, 1990 2. COSPAR, 1993 6. Hogg, 1989 5. Engman and Gurney, 1991 2. Carroll and Carroll, 1989 3. EOSAT, 1993a 3. Carroll el al., 1993 7. Negri el al., 1994 6. Jones and Carroll, 1983 4. NOAA, 1993 4. Corbley, 1994 8. NOAA, 1993 7. Reutov and Shutko, 1986 5. Engman and Gurney, 1991 g - Basin Characteristics 9. Scofield, 1991 8. WMO,1993 6. Goodison and Walker, 1993 I. Corbley, 1994 10. Walton el of., 1990 9. Zevcnbergen and Rango, 7. Harrison and Lucas, 1989 2. EOSAT, 1994a 11. WMO,1992a 1992 8. Kumar et al., 1991 3. Fuller, 1993 9. Rango,1995 4. Roberts el al., 1993 APPENDIX II

HYDROLOGICAL OBSERVATIONAL REQUIREMENTS AND EXISTING AND PLANNED SATELLITES AND SENSORS FOR HYDROLOGICAL APPLICATIONS

Table 1 - A summary ofhydrology and water management observational requirements in relation to satellite monitoring opportunities (see Herschy, Barrett and Roozekrans, 1985 a1U11988)

Parameter Resolution Frequency Accuracy max. min. opt max. min. opt max. min. opt

Precipitation 100m 110 kml 1 km 5min ~ ~ 10% 30% 20% Snow depth 30m 110 kml 1 km 12 h 13~J1J 2cm 10cm ~ [1MJ 1 l L5.. Ice cover 10 m 11 km I 25m 12 h I 7;q 24 h 1% 1 20%1 [10% I

Glaciers ~ ~oml 1~5j1lJ [iL] ~ [IY] 1% I 5% I I 2% I Surface water areal exlent 10m 100 m 30m 12 h @] 24 h 1% 5% 3% Groundwater - 50m 1 km 100m 1 y 10m aquifer maps 5y 3y 5m 30m

Evaporation 100 m [lOkml ~ [iill ~ D:TI 10% 130% I 120% I Water quality - turbidity 30 m 300m 100 m 3 h 6h 10% 50 [ 20%-1 1 l --- ~ --- 1 % I Drainage - drainage areal 0 m 1100 ml 20 m QL] ~ Qi] 0.1% 1% 0.5%

Frequency: h =hour; d =day; M = month; y =year

Feasibility: CJ = Requirement can be generally met by existing satellite(s) -- = Requirment should be generally met by near future satellite(s) Note: Where a value is neither boxed nor underlined the observational requirement cannot generally met either by existing or firmly expected future satellites, given the present state of the art.

Table 2 - Hydrological observational needs (Ormsby andEngman, 1993)

Parameter Approach Spatial resolution Observation frequency Soil moisture ~ wave (passive/active) 1-10 km/30c 50 m 2 days

Vegetation • Identification Visible, NIR, TIR, ~ wave 1 km 7 days • Areal extent Visible, NIR, TIR, ~ wave 30m 30 days • Condition (stress) Visibie, NIR, TIR, ~ wave 30m 3 days

Snow • Water equivaient ~wave 1 km 7 days • Areal extent Visible, .~ wave 1 km 7 days • Thickness ~wave 1 km 7 days

Radiation • Short and long wave Visible, TIR 100 km2 6 hours

Precipitation J...l wave or in situ 1 km Daily

Evapotranspiration TIR, visible, ~ wave, modeiling 1 km Daily

Runoff TIR, visibie, ~ wave, modelling Daily

Wetland aeral extent Visible, NIR, TIR, ~ wave 30-100m Monthly APPENDIX II 61

Table 3 - WMO satellite data requirements (WMO, 1993a)

Measurement Instrument type Candidate sensor Wind Geostationary imaging radiometer VISSR, MVIRI Wind profile (undetermined) (undetermined) Temperature profile Atmospheric sounder (IR, microwave) HIRS, AIRS,IASI, MSU Humidity profile Atmospheric sounder (IR, microwave HIRS, AIRS, IASI, MSU, MHS Liquid water, total and Imaging multi-spectral (microwave) SSM/I, TMI, MIMR precipitation rate radiometer Cloud top temperature Imaging multi-spectral (VIS, IR) radiometer AVHRR, MODIS, MERIS, HIRS, AIRS, Atmospheric sounder (IR) IASI Cloud type and amount Imaging multi-spectral (VIS, IR) AVHRR, VISSR, MVIRI, SEVIRI, radiometer MODIS, MERIS Ozone total and profile Atmospheric chemistry spectrometer TOMS, GaME, GOMOS, SCIAMACHY, MOPITT, IMG, ILAS, HIRDLS, TES Radiation, net Earth radiation radiometer CERES, SCARAB Aerosol Imaging multi-spectral (VIS, IR) radiometer AVHRR, MODIS, MERIS Multi-purpose imagery Imaging multi-spectral (VIS, IR) radiometer AVHRR, SEVIRI, MODIS, MERIS Pressure, sea surface (undetermined) (undetermined) Sea-surface temperature (SST) Imaging multi-spectral (VIS, IR) radiometer AVHRR, aCTS, MODIS, MERIS Atmospheric sounder (IR) AIRS, lAS!, HIRS Multi-directionai radiometer (lR) ATSR Temperature, surface Imaging multi-spectral (VIS, iR) radiometer AVHRR, MODIS, MERIS Multi-directional radiometer (IR) ATSR Ocean surface wind vector Wind (microwave) scatterometer AMI, NSCAT, STIKSCAT Soil moisture (undetermined) (undetermined) Earth surface albedo Imaging mUlti-spectral (VIS, IR) radiometer MODIS, MERIS, AVHRR, AVNIR Ocean wave height Radar altimeter ALT, RA Wave spectrum Mapping radar (SAR) AMI Ice coverage, edge Imaging multi-spectral (VIS, IR, microwave) AVHRR, MODIS, MERIS, SSM/I, radiometer MIMR Ice thickness (undetermined) (undetermined) Snow coverage, edge Imaging multi-spectral (VIS, microwave) AVHRR, MODIS, MERIS, SSM/I, radiometer MIMR Snow depth, snow water equivalent (undetermined) (undetermined) 62 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

Table 4 - Existing and plannedteJilote sensing satellites/platforms and sensors relevant to hydrological applications (Kuittinen, 1992; Schulz andBarrett, 1989; Blyth, 1993; Engman and Gurney, 1991; WMO, 1991; WMO, 1992a, hj WMO,1993a)

Satellite (year) Orbit Sensors

• GOES series GfW VISSR, VAS, DCS, SEM (NOAA/USA, 1975-1990S) ~ GOES-NEXT GIW (NOAA/USA, 1994, 2005) • GMS series GJW VISSR (Japan, 19n, 1989, IB94, 1995) • Meteosat -series GJW VISSR (ESA, 1978, 1991, 1993, 1'995,2000) • MSG '(2nd Generation Meteosat) GIW (ESA and EUMETSAT,2000, ;2012) • INSAT series ,GfW VHRR (India, 1990, 1992, future) • GOMS 'OIW (Russian P"deration, 1991, 1993, mid-t990s) • FY-2GIW VISSR (China, mid-1990s) • NOAA series/POESITIROS·N PM AVHRR, TOVS/HIRS and MSU, SSU, DCS, (USA, 1978-2000) SEM,SBU\I, ERSE, SAR, AMSU, ACZCS • METOP (to repiace NOAA series) PM (ESA and EUMETSAT,2010) • DMSP PIW OlS, MIR, MTS, SES, SSM/I, SMR, AMSU (US Mil ital)', 1970s-1990s) • UARS PIW CLAES, HALOE, HRDI, ISAMS, MlS, PEM, (NASA/USA, 1991) SOLSTICE, SUSIM, WINTER, ACRIM, .SSUV, MEPS, AXIS, MAG, HEPS • Meteor series PIW . VISSR, TOMS, SCARAB (Russian Federation, 1977, 1990, 1991, 1996, future) • TRMM PIW SMR,RAR, AVHRR (USAand Japan, mid-1990s) • FY-l PIW CAVHRR,CSPD (China, 1990, 1991) • landsat series PIE MSS, TM, ETM (NASA/USA, 1972-1990s) • SPOT series PIE HR\I, DORIS, VMS, HRVIR (CNES/France, 1986, 1990,1992) • J-ERS-l PIE SAR,VNIR (NASA/USA and Japan, 1992) • MOS-l PIE SMR, MESSR, VTIR, MSR (Japan, 1991) • ADEOS PIE SMR, OCTS, AVNIR, NSCAT, TOMS, (NASDA/Japan, 1996) POLDER, IMG, ILAS, RIS • ERS series PIE AMI/SAR, ATSR, PRARE, lR RA, AlT (ESA, 1991, 1994, 1998) • ENVISAT (replacing ERS series) PIE (ESA, 1998 +) • METEOR-PRIRODA PIE MSU-M, MSU-S, MSU-SA, MSU-VA (Russian Federation, 1991) APPENDIX II 63

Satel/ite (year) Orbit Sensors

• ALMAZ PIE SAR,RS (Russian Federation, 1991) • RADARSAT PIE SAR (Canada, 1994-2000)

• IRS PIE L1SS, L1SS-1, L1SS-2 (India, future) • RESURS series/COSMOS PIE MSU-E, MSU-K, MSS (Russian Federation, 1988, 2000)

• LAGEOS PIE (USA and Italy, future) • OKEAN series P/Ocean MSU-S, MSU-M (Russian Federation, 1991, 1995) • N-ROSS P/Ocean SSM/I, SCAn, ALT, LFMR (NASA/USA, 1989) • TOPEx/POSEIDON P/Ocean ALT, LRA, IMR, GPS (USA and France, 1989, 1992) • Sea Star P/Ocean (USA, future) • TOMS/Earth Probe P/Ocean (USA, future) • EOS/Columbus Project P&G HIRIS, MODIS-T, MODIS-N, SMR, SSAR, (USA, Europe, Japan and Canada, mid-1990s) AVHRR, TM, OCM, ESTAR, CERES, MIMR • X-SAR/SIR-C series/and L1DAR Space SAR (USA, Germany and Italy, 1996) shuttle • MECB P/ (Brazil, future) • ARABSAT (future)

P := polar-'orbiting G geostationary W:= weather satellite E Earth resources satellite

Table 5 - Key satellites carrying microwave sensors suitable for hydrological monitoring (Blyth, 1993)

Satellite Planned launch date Sensor type Waveband Ground resolution

MOS-1B February 1991 SMR K,Ka 32,23 km DMSP December 1990 SMR K,Q,W 43-13 km ALMAZ March 1991 SAR S 15 m ERS-1 July 1991 SAR C 12 m + 3 years ALT Ku < 10 cm (relative ht) J-ERS-1 February 1992 SAR L iBm SIR-C1 November 1993 + 7 days SAR L,C;X 60-10 m RADARSAT 1995 + 5 years SAR C 100-8 m ERS-2 End 1994 + 3 years SAR C 12 m SIR-C2 November 1994 + 7 days SAR L,C,X 60-10 m ADEOS 1995 SMR Ku 25 km TRMM mid-1990s SMR K,O,W 10 km RAR K 4km SIR-C3 1996 + 7 days SAR L,C,X 60-10 m NASAEOS} 2000 Not yet finalized POEM

SAR := Synthetic aperture radar RAR := Real aperture radar SMR := Scanning microwave radiometer 64 CURRENT OPERATIONAL APPLICATIONS OF, REMOTE SENSING IN HYDROLOGY

Table 6 - Status of miCrowave remote sensing of hydrological-variables (Blyth, 1993)

Operational status Variable Platform Sensor

Operationally feasible Catchment physical caracteristics Watershed boundaries A SAR Stream networks A SAR River channei slope A Stereo"SAR Slope elements A Stereo"SAR Catchment variables Surface water extent S SAR Snow extent S Passive Limited { Snow depth S Passive • Snow water equivalent S Passive. Rainfall over oceans S Passive

Near operational Catchment variables Rainfall over land S Passive + SAR

Requiring development Catchment physicai characteristics Impervious surfaces A&S SAR Catchment variables Surface roughness A&S M"freq SAR Soil moisture S Passive + SAR Vegetation moisture­ S Passive + SAR Snowextent S Passive + SAR Snow melt S Passive + SAR Snow water equivalent S Passive- + SAH S = Satellite; A = Aircraft

TABLE 7 - Key microwave frequencies for hydrological studies (Blyth, 1993)

Hydrological variable Pas_sive Active Frequency (GHz) 104 6.6 10.7 18 21 37 90 0.2 OA 1.5 4.2 5.8 10:936.0 Band L XX KK Q W P S L C X K Satellite resolution 18 10 6 5 3 1.5 18 8 20 byyear2000 km km km km km km m m m Catchment physical characteristics Watershed boundaries • 0 Stream networks 0 Slope elements '0 River channel slope and roughness • • 0 Impervious surfaces • • Catchment variables Soil moisture below: - bare-so.il!short veg. .-' 0 a a • - crops .- 0, -·trees, 0 n Vegetation- moisture 0 • 0 Snow: -extent • • a a • 0' - onset ofmeit .- • 0 • _. water equivalent 0 0 0 0 0 0 Frozen ground 0 0 Surface water extent • 0 0. Rainfall 0 •• •

• =very useful; 0 =useful; 0 =helpfui APPENDIX III

SPECIFIC SENSORS WmCH MIGHT BE CARRIED BY EOS

Specification and justification for satellite sensors deemed likely to be of special value for hydrology and water management in the foreseeable future (from Barrett, Herschy and Stewart, 1988b).

Satellite type Orbit Sensor Spectral Spatial Imaging Principal type regions resolution frequency applications

Environmental Geostationary Imaging 0.5 - 0.9 ~m 0.5 km' 48 d-l Meteorology, cloud Radiometer 5.7-7.1 ~m 4.0 km 24 d-l systems, rainfall, flash (GIR) 10.5 -12.5 ~m 0.5 km' 48 d-1 floods, thermal inertia, surface water, ice and snow cover, aridity, drought

Environmental Near-poiar Scanning 90 GHz 5 km' At least Cloud liquid water (Dual Microwave 37 GHz 10 km' 2 d-l ' contents, rain rates, polarized) Radiometer 25 GHz 10 km' 4 d-l snow cover, snow (SMR) 19 Ghz 10km' preferabie' water equivalent, 5 GHz 15 km' relative snow depth, snow melt, sea ice and poiymas

Environmental Near-polar Imaging 0.55 - 0.7 ~m 0.33 km' At least 2 d-' Meteorology, cloud Radiometer 0.725 -1.3 ~m 0.33 km' (0.3.15 h) systems, rainfall, (AVHRR) 3.55 - 3.95 ~m 0.33 km' 45 d-' surface temperatures, 10.5-11.5 ~m 0.33 km' preferable ice and snow cover, (0.3,09, major floods, aridity 15,21 h) and drought, vegetation indices

Environmental Near-polar Synthetic X-band 30m Every Extent of wet snow, Aperture C-band 30m 7-10 days: lake and river ice, Radar (SAR) L-band 30m 5 days glaciers, surface preferable' water, soil mOisture, measurement and monitoring

Environmental Near-polar Ocean 0.400 ~m 0.25 km Every three Water colour, quality Colour 0.445 ~m 0.25 km days* and turbidity, Monitor 0.520 ~m 0.25 km suspended sediments, (OCM) 0.565 ~m 0.25 km pollutants, phylo- 0.640 ~m 0,25 km plankton, surface 0.685 ~m 0.25 km temperatures, iand 0.785 ~m 0.25 km applications including 1.020 ~m 0.25 km vegetation and land 1,600 ~m 0,25 km use 3.700 ~m 0.25 km 8.500 ~m 0.25 km 10.800 ~m 0.25 km 12.000 ~m 0.25 km

Earth Near-polar Thermatic 0.45 - 0.52 ~m 30m Morphology, land resources Mapper (TM) 0.52 - 0.60 ~m 30m cover, surface water 0.63 - 0.69 ~m 30m extent and quality, 0.76 - 0.90 ~m 30 m 4-5 days' giacier, lake and river 1.55 - 1.75 ~m 30 m ice, snow cover, 2.08 - 2.35 ~m 30m ground water, etc. 10.4 -12.5 ~m 120 m

Space shuttle/ Various Modified Visiblet c. 10 m Occasional Relief drainage, Space lab aerial survey Infraredt c. 10 m c. every topographic mapping cameras False colourt c. 10 m ten years (LFC)

Represents significant advance on capabilities of earlier/present sensor systems t Depending on films and filters used APPENDIX IV

STATE-OF-THE-ART OF SATELLITE DATA REQUIREMENTS AND REQUIRED CAPABILITIES OF NEAR·POLAR-ORBITING AND GEOSTATIONARY SATELLITES

Table 1 - ECSAT list ofsatellite reqlirements - upper air and surface (WMO, 1993a)

Requirement Horizontai resolution Vertical resolutIon Frequency Accuracy Use

UPPER AIR

Wind profile R 25 km 50 Ly 6 hr 1 m/s A G 100 km 15 Ly 12 hr 1 m/s A Temperature profile R 25 km 50 Ly 6 hr 1.0 K A G 25km 15 Ly 6 hr 1.0 K A,B Humidity profile R 25 km 10 Ly 6 hr 5% A G 50 km 10 Ly 6 hr 5% AB Precipitation rate R 5 km 30 min 1 mm/d A B, C G 25 km 6 hr 3mm/d A,B,C Liq uid water, total G 25 km 5 Ly 6 hr 0.1 « 100 nm) 2 (> 100 nm) A,B Cloud top height R 5km 1 hr .2km A G 10 km 3 hr .2km A Cloud top temperature R 5 km 1 hr 1 K A G 5 km 3 hr 1 K A Cloud type and amount R 5km 10 Ly 1 hr 10% A, B G 5km 10 Ly 6 hr 10 % A, B Heightof tropopause R 25km 6 hr 1 km A Ozone total and profile G 50km 6 hr 5% B Radiation, net R 10 km 12 hr 5w/m2 A,B,C G 50 km 12 hr 2.5 w/m2 A,B,C Aerosol G .5 km 1 km. 6 hr 10% B

SURFACE

Multi-purpose imagery R 10 m 12 hr A,B,C G .4 km 6 hr A,B,C Pressure, sea surface R 20km 1 hr .5mb A G 25 km 6 hr .5mb A Temperature, sea surface R 1 km 6 hr .1 K A G 1 km 6 hr .1 K A,B Temperature, surface R .5km 12 hr .1K A,B,C G 1 km 12 hr .5K A B, C Sea surface wind R 10 km 6 hr 3 kts, 10 deg A G 25 km· 6 hr 2 kts, 10 deg A Soil moisture R 20km 24 hr 6mm A,B,C G 25 km 24 hr 6mm A,B,C Albedo, visible R 1 km - 6 hr 1% B,C G 5 km 24 hr 1% B,C Aibedo, near IR G 2 x2 deg 1 mo 3% B,C Ocean wave height G 10 km 1 hr .3m A Wave spectrum G 10 km 1 hr 10 deg (dir) .5 s (period) A Ice coverage R 10 km 1 d 2% A,B,C Ice edge R 10 km 1 d 2% A,B Ice thickness R 25 km 1 d 10% B,C Snow coverage R 1 km 1 d 2% A,B,C G 1 km 1 d 10% A B, C Snow edge R 1 km 1 d 2% A B, C Snow depth R 1 km 1 d 5mm A B, C G 1 km 12 hr 5mm A,B,C Snow water equivalent R 1 km 12 hr 2mm C

User category codes: A = operational meteoroiogy R = regional B = climate and environmental monitoring and change G= global C = hydrology/hydrometeoroiogy/agrometeorology APPENDIX IV 67

Table 2 - Reqnired capabilities from near-polar-orbiting satellites {WMO, 1993a)

• On the surface and the Planetary Boundary layer (PBl): Humidity profile Air pressure Uquid water, total and profile Winds in the PBl Height of the tropopause Atmospheric instability Aerosol load and profiie Height of the PBl Ozone and other trace gases, total and profile Sea-surface temperature Clouds: coverage, brightness, top height, ice~iquid Sea state: wave height and spectrum Sea ievel, currents, fronts, eddies • Climate related information: Sea salinity -Most of the parameters listed above land temperature Radiation budget large-scale soil moisture - Relevant sun/earth interactions large-scale vegetation index Ice and snow characteristics • Telecommunication functions: Precipitation rate and accumulated rainfall Globai acquisition of the total volume of data Broadcasting of data to local users in real-time • Observations in the free atmosphere: Data collection and location Temperature profile Support to the Search and Rescue mission - Wind profile

Table 3 - Reqnired capabilities from geostationary satellites (WMO, 1993a)

• Observations from geostationary orbit (primary): large-scale vegetation index Ear[y detection of cyclone-genesis Precipitation index Convection development monitoring Temperature profile Winds as derived from motion of clouds or other Humidity profile tracers Tota[ ozone Atmospheric instability (temperature/moisture variation) Clouds: coverage, brightness, top height Sea-surface temperature variations in coastal waters Radiation budget Diurnal variations of land temperature Bi-directiona[ reflectance properties of clouds and land • Telecommunication functions: Diurnal variation of earth radiation budget Broadcasting of data to loca[ users in near-real-time Data collection available all time (for alert service) • Observations complement[ng those from near-polar orbit: Meteorologica[ data distribution Oceanic sea-surface temperature Continuous support to Search and Rescue mission - Continental [and temperature 68 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

Table 4 _ Observing missions and their corresponding objectives, observation requirements and typical instrumentation (WMO, 1993a)

1. THE MULTI-PURPOSE IMAGERY MISSION

The objectives' of this mission are (the term "inference" indicates that additionai processing is needed to get the information): direct observation of cloud patterns inference ofcloud properties inference of precipitation rate and indexes inference of land and sea-surface temperatures inference of vegetation index and soil moisture direct observation of ice and snow cover inference ofice and snowparameters inference of c1oud~motionwinds Inference of atmospheric instabiiity

The observation requirementsto allow for these applications to be implemented are as follows: spectral bands: atmospheric windows and specific absorption bands (particularly for water-vapour) in the visible (VIS), infrared (IR) and microwave geometrical resoiution: 0.5 to 5 km (meteorology, oceanography, ice), 10 to 50 km (climatology) radiometric accuracy: signal-to-noise (SIN) ~ 10 at 1% albedo (short-waves), noise equivalent delta temperature (NED1) = 0.1 Kat 300 K (long-waves) spectral channels: 3-5 channels in VIS/NIR, 3-10 channels in IR, 4-8 channels In microwave repeat cycle: 10-30 min (geostationary), 6 hours (polar)

Typical existing or proposed instruments to fulfil the mission are: AVHRR, MODIS, for medium-resolution VIS/IH MIMR, for microwave imagery Imagers from all geostationary satellites

2. THE ATMOSPHERIC SOUNDING MISSION

The objectives of this mission are mainly driven by operational meteorology. They are: vertical temperature profile vertical humidity profile vertical wind profile landwater (total and possibly profile) atmospheric discontinuities (height of PBL and tropopause). It is understood that all of these parameters need intensive processing for extraction and require support from the imagery mission and other-than-satellite information.

The observation requirements to allow for these applications to be implemented are as follows: spectral bands: absorption bands of CO2,O2, water vapour and nearby atmospheric windows in IR and microwave; selected VIS/NIR and IR lines exploited in an active mode (L1DAR) horizontal resolution: 50 km global, 25 km local vertical resolution: for profiles 1 km low-medium troposphere, 2 km medium-high troposphere, 3 km above the troposphere; for discontinuities 100 m PBL, t km tropopause spectral channels: 20 to 1000s channels in IR, 10 to 20 channels in microwave, 2-3 channel L1DAR in VIS/NIR, 1 channel Doppler L1DAR in IR repeat cycle: 6 hours

Typical existing or proposed instruments to fulfil the mission are: HIRS, MODIS for coarse vertical resolution in IR IASI, AIRS for high vertical resolution in IR AMSU-A and AMSU-B for all-weather sounding in microwave

Space L1DARs for wind profile and atmospheric discontinuities are not planned to be flown before 2000. APPENDIX IV 69

3. SEA-SURFACE OBSERVATIONS

The objectives of this mission account for meteorological, oceanographic and ciimate appiications. They are: accurate sea-surface temperature wind stress and vector wave height and spectrum, sea state sea levei, currents, fronts, eddies chlorophyll concentration and suspended sediments air pressure accurate precipitation rate

The sea-surface measurement requirements for these appiications to be implemented are: spectral bands: atmospheric windows in VIS/NIR, in TIR and in microwave; selected microwave channels exploited in active mode (radar) geometrical resolution: 25 km (global), 5 km (local) spectral channels: 5-10 in VIS/NIR, 3-5 in TIR, 5-8 in passive microwave; 2-3 in active microwave repeat cycle: 6 hours

Typical existing or proposed instruments to fulfil the mission are: ATSR, MODiS, for VIS/NIR and IR MIMR, for passive microwave MERiS, SeaWIFS, aCTS, for high spectral resolution VIS/NIR ASCAT, NSCAT, for scatterometry in active microwave RA-2, for altimetry In active microwave ASAR, for active microwave imagery.

There are no plans to fly instruments for air pressure or accurate precipitation rate observations on the operational or near­ operational satellites. Therefore, they have not been considered in this review.

4. RADIATION BUDGET AND ATMOSPHERIC CHEMISTRY

The objectives of this mission are driven by climatological and environmental monitoring and study. They are: short-wave components of the Earth radiation budget long-wave components of the Earth radiation budget aerosol load and profile accurate water vapour profile in high troposphere greenhouse trace gases (total and rough profile) ozone total and profile stratospheric trace gases active on the ozone life cycle tropospheric trace gases hazardous to humans and bio-masses solar-terrestrial interactions

The observation requirements to allow for these applications to be implemented are as follows: broad-band radiometry, accurately caiibrated, of all significant bands for energy budget: ultraviolet (UV), VIS, near, thermal and far IR scanning modes allowing multi-viewing angles to account for bi-directional reflectance properties and polarization high-resolution spectroscopy across all fields of Uv, VIS, IR and sub-millimetre waves; nadir-scanning for total-column observation and tropospheric profiling, and limb-scanning for high vertical resolution in the upper atmosphere, also exploiting sun and star occultation multi-platform observation (Including geostationary satellites) to account for diurnal variations monitoring of solar particles and space environment

Typical existing or proposed instruments to fulfil the mission are: ERBE, SCARAB, CERES, for broad-band UVNIS/IR MERIS, POLDER, MISR, for reflectance properties SBUV, GaME, TOMS, SCIAMACHY, for nadir-viewing UVNlS/NIR GOMOS, SAGE, SCHIMACHY, ILAS, for limb-viewing UVNlS/NIR IASI, AIRS, IMG, MOPITT, for nadir-viewing IR spectroscopy MIPAS, ILAS, for iimb-viewing IR spectroscopy SEM, for space environment monitoring 70 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

5. HIGH-RESOLUTION IMAGERY

The objective of this mission, limited to certain fields of WMO interest is, in general, to complement the multi-purpose imagery mission by more accurate observation of surtace details on specific sites critical for global climate monitoring and study. Some of these observations are: deforestation in tropical areas, desertification land water, runoff, floods, coastal erosion ice edge, thickness and type in the polar caps snow edge, depth and water equivalent in strategic basins coastal~zone water dynamics (eddies, currents, ...j

The observation requirements to allow for these applications are generally met by systems which have been designed prima­ rily for commercial applications (resources discovery, monitoring and management). For WMO purposes, they are: specfral bands: VIS and NIR atmospheric windows, microwave in active mode (SAR) horizontal resolution: 10 to 100 m spectral channels: 3-10 in VIS and NIR, 1-3 in TIR, 1-3 in active microwave (SAR) with two polarizations repeat cycle: one week to one month

The typical existing or proposed instruments to fulfil the mission are: lANDSAT TM, in VIS, NIR and TIR SPOT HRV, in VIS and NIR with stereo capability ADEOS-1 VNIR, in VIS and NIR EOS ASTER, in VIS, NiR and TIR with stereo capability ENVISAT-1 ASAR, in C-band, with dual polarization

6. TELECOMMUNICATION FUNCTIONS

The objectives of this mission are: global data acquisition from all geostationary and near-polar-orbiting environmental satellites of interest for WMO programmes direct broadcast to local read-out stations of all data useful for operational meteorology data collection from in situ automatic station and location of the platform if mobile meteorological data distribution through geostationary satellites in support of the GTS support to Search and Rescue

Most of the existing and planned systems either provide or will provide these services. The operational instrument for data collection and location from NOM satellites is called ARGOS. The Search and Rescue support is being provided by NOM and METEOR satellites by a system called SARSAT within the framework of an international programme called COSPAS. ACRONYMS

ACRIM Active cavity radiometer irradiance monitor ERR Extreme heavy rainfall ACZCS Advanced coastal zone colour EMS Electromagnetic spectrum ADEOS-! Advanced Earth Observing System (Japan) ENSO El Nino/Souther Oscillation AIRS Advanced infrared sounder ENVISAT ESA enviroment satellite series ALMAZ Newest Russian remote sensing satellite, means EOS Earth Observing System "diamond" EOSAT Earth Observation Satellite Company ALT Altimeter ERBE Earth Radiation Budget Experiment AMI Active microwave instrument ERS Earth remote sensing satellite (ESA) AMSU Advanced microwave sounding unit ERS-! First European remote sensing satellite APT Low resolution polar-orbit data ERTS Earth resources satellite ARABSAT Arabian satellite ERTS-! First earth resources satellite ARGOS Data relay and platform location system ESA European Space Agency ARMAR Airborne rain mapping radar ESCAP Economic and Social Conunission for Asia and ASAR Advanced synthetic aperture radar the Pacific (United Nations) ASCAT Advanced scatterometer ESTAR Electronically steered thinned array radiometer ASTER Advanced space-borne thermal emission and ET Evapotranspiration reflection radiometer ETM Enhanced thermatic mapper ATSR Along-track scanning radiometer EUMETSAT European Organization for the Exploitation of AVHRR Advanced very high resolution radiometer Meteorological Satellites AVNIR Advanced visible and near-infrared radiometer EURISY European Association for the International Space AWIPS Advanced weather interactive processing system Year AXIS Atmospheric X-ray imaging spectrometer FAO Food and Agriculture Organization of the United B4 Internally self-calibrating geostationary infrared Nations method FEMA Federal Emergency Management Administration CAL Computer-assisted learning (USA) CASCADE Advanced flood warning system (UK) FFWRS Flood forecasting warning an9- response system CAVRRR Sensor onboard FY-l FIFE First ISLSCP Field Experiment CBS Commission for Basic Systems (WMO) FRONTIERS Forecasting rain optimized using new techniques CD-ROM Compact disc peripheral device of interactively enhanced radar and satellite data CERES 1\\10 broadband scanning radiometer FY-l, FY-2 Geostationary meteorological satellites (China) CHEST Combined higher education software team (UK) GARP Global Atmospheric Research Programme CIS Countryside information system GATE GARP Atlantic Tropical Experiment CLAES Cryogenic limb array etalon spectrometer GCIP GEWEX Continental Scale International Project CNES Centre national d'etudes spatiales (France) GEOSAT Geodesic satellite (USA) COE Army Corps of Engineers (USA) GEWEX Global Energy and Water Cycle Experiment COSMOS Russian Federation satellite GIR Geostationary imaging radar COSPAR Committee on Space Research (ICSU) GIS Geographical Information System COSPAS International programme which supports search GMS Geostationary meteorological satellite (Japan) and rescue GOES, Geostationary operational environmental CSPD Sensor onboard FY-I GOES-NEXT satellite (USA) CZCS Coastal zone colour scanner GOME Global Ozone Monitoring Experiment (ESA) DAD Depth-area-duration GOMOS Global ozone monitoring by occulaLion of stars DBMS Database management system GOMS Geostationary operational meteorological DCP Data collection platform satellite (Russian Federation) DCS Data collection system GOS Global Observing System DEM Digital elevation model GPS Global Positioning System DIAL Differential absoption lidar GTS Global Telecommunication System DMSP Defense Meteorological Satellite Program GTZ Agency for Technical Cooperation (Germany) DORIS Dual Doppler receiver HALOE Halogen Occultation Experiment DWOPER Dynamic wave OPERational model (NWS) HEPS High energy particle spectrometer EARSeL European Association of Remote Sensing HERP Hydrological Emscher Radar Project Laboratories HIRDLS Upper-atmosphere sounding instrument ECMWF European Centre for Medium-range Weather HIRlS High-resolution infrared sounder Forecasts HIRS High-resolution imaging spectrometer ECSAT Executive Council Panel of Experts on Satellites HOMS Hydrological Operational Multipurpose System (WMO) HOP Hierarchial Objective Procedure 72 CURRENT OPERATIONAL APPLICATIONS OF REMOTE SENSING IN HYDROLOGY

HR High-resolution geostationary data MHS Microwave humidity sounder HRDI High-resolution Doppler imager MIMR Multifrequency imaging microwave radiometer HRPT High-resolution polar-orbit data MIPAS Michelson interferometer passive atmospheric HRV High-resolution visible imager sounder HRVIR Improved high-resolution visible imager MIR Multispectral infrared radiometer HW Snow water equivalent MIS Management information system HYDROSAT HYDROlogical SATellite System MISR Multi-angle imaging spectro-radiometer HYRAD HYdrological RADar System MLS Microwave limb sounder HYREX Hydrological Radar Experiment MODIS Moderate resolution imaging spectrometer IASI Atmospheric sounder for measuring humidity MOMS Modular optoelectronical multispectral scanner profile MOPIIT Measurements of pollutants in the troposphere ICSU International Council for Science MOS Maritime Observation Satellite (Japan) lFFA Interactive flash flood analyser MOSAIC Mcteosat operational service for data aquisition lFRB International Frequency Registration Board and interchange (EUMETSAT) (lTU) MSG Meteosat second generation lGBP International Geosphere-Biosphere Programme MSR Microwave scanning radiometer (ICSU) MSS Multispectral scanner lHP International Hydrological Programme MSU Microwave sounding unit (UNESCO) MTP Meteosat Transition Programme !LAS Improved limb atmospheric spectrometer MTS Microwave temperature sounder lMG Interferometric Monitor for Greenhouse Gases MVIRI Geostationary imaging radiometer to measure lMR TOPEX microwave radiometer wind lNSAT Indian national satellite (India) NASA National Aereonautics and Space Administration IR Infrared (USA) IRS Indian remote sensing satellite (India) NASDA National Aeronautics and Space Development IRSA Institute for Remote Sensing Applications Agency ISAMS Improved stratospheric and mesopheric sounder NDVI Nonnalized Differential Vegetation Index ISLSCP International Satellite Land Surface Climatology NEDT Noise equivalent delta temperature Project NERC National Environment Research Council (UK) ITE British Institute ofTerrestrial Ecology (UK) NESDIS NOAA Environmental Satellite Data and ITU International Telecommunication Union Infonnation Service J-ERS Earth resources satellite (Japan) NEXRAD Next generation radar J-ERS-l First Japan earth resources satellite NFS Nile Forecasting System KVR Satellite (Russian Federation) NlMBUS-7 Coastal satellite LAGEOS Laser geodynamics satellite NIMROD New scheme that replaced FRONTIERS Landsat NASA/USA polar~orbiting earth resources satel­ NIR Near infrared lite NOAA National Oceanic and Atmospheric LARST Local applications of remote sensing techniques Administration (USA) LFC Large format camera NOAA-# NOAA weather satellites LFMR Low frequency microwave radiometer NOHRSC National Operational Hydrologic Remote LHMA Latvian Hydrometeorology Agency Sensing Center (NWS) LIDAR Light detection and ranging NRI National Resources Institute (UK) LISS Linear imaging self scan N-ROSS Navy remote ocean sensing satellite LR Laser retroreflector NSCAT Naval scatterometer LRA Laser retroreflector array NWS National Weather Service (USA) LRIT Low rate infonnation transmission NWWRP North West Weather Radar Project MAG Fluxgate magnometer OCM Ocean colour monitor MAR Modernization and associated restructuring OCTS Ocean colour and temperature sensor MDD Meteorological data dissemination OHP Operational Hydrology Programme (WMO) MECB Future near-polar-orbiting satellite (Brazil) OKEAN Oceanic satellite (Russian Federation) MEPS Medium-energy particular spectrometer OLS Operational linescan system MERIS Medium resolution imaging spectrometer PARAGON Data-processing system (UK) MESSR Multispectral electronic self-scanning radiometer PBL Planetary boundary layer METEOR Polar-orbiting weather satellite (Russian PC Personal computer Federation) PDM Probability distributed moisture METEOR­ Environmental satellite (Russian Federation) PEM Particle environment monitor PRIRODA PERMIT Polar-orbiting effective rainfall monitoring inter­ METEOSAT Geosynchronous weather satellite (ESA) active technique METOP Near-polar-orbiting meteorological satellite PMP Probable maximum precipitation (ESA and EUMETSAT) PMW Passive microwave 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ACRONYMS 73

POEM NASA-EOS satellite SSARR Streamflow synthesis and reservoir regulation POES Polar-orbiting operational environmental satellite SSMII Special sensor microwave/imager POLDER Polarization and directionality of reflectances SSMIT Special sensor microwave/temperature PRARE Precision Range and Range Rate Experiment SST Sea-surface temperature PRIRODA Environmental satellite (Russian Federation) SSTC Specialized Satellite Applications Training Centre PRMS Precipitation runoff modelling system SSU Stratospheric sounding unit PROMETE Series of multimedia and multi-targeted STIKSCAT Wind microwave scatterometer programmes SUSIM Solar-UV spectral irradiance monitor QPF Quantitative precipitation forecast SWCC Second WiJrld Climate Conference RA Regional Association (WMO) TES Tropospheric composition instrument RA Radar altimeter TIR Thermal infrared RADARSAT Active and passive microwave satellite (Canada) TIROS Television and infrared observation satellite (USA) RAINSAT Quasi-operational real-time (nowcasting) rainfall TM Thematic mapper estimating scheme in Canada TMI Thermal multifrequency imager RAR Real aperture radar TOGA Tropical Ocean Global Atmosphere Programme RBV Return beam vidicom TOGA COARE TOGA Coupled Ocean-Atmosphere Response RESURS Near-polar-orbiting earth resources satellite Experiment (Russian Federation) TOMS Total ozone mapping spectrometer RFFS River flow forecasting system TOPEXI Ocean Surface Topography Experiment RIS Retroreflector in space POSEIDON (USAIFrance) RMS Root mean square TOVS TIROS-N operational vertical sounder RMTC Regional Meteorological Training Centre TOVSIHIRS TOVS ozone imagery (WMO) TRMM Tropical Rainfall Measuring Mission RS Radiometric scanner TSC Thermal scarmer RSI RADARSAT International Inc. UARS Upper atmospheric research satellite (NASA) SAGE Stratospheric aerosol and gas experiment UK United Kingdom SAR Synthetic aperture radar UN United Nations SARSAT Search and rescue satellite-aided tracking UNESCO United Nations Educational, Scientific and SBUV Solar backscatter UV radiometer Cultural Organization S Snow cover area US, USA United States ofAmerica SCARAB Scanner for earth's radiation budget UV Ultraviolet SCATT Scattometer VAS VISSR atmospheric sounder SCIAMACHY Scanning imaging absorption spectrometer for VCI Vegetation Condition Index atmospheric cartography VHRR Very high resolution radiometer SEASAT Earth-orbiting remote sensing satellite VIS Visible SeaWIFS Sea-viewing wide field sensor VISSR Visible and infrared spin-scan radiometer SEM Space environmental monitor VMS Vegetation monitoring sensor SES Space environmental sensor VNIR Visible and near-infrared radiometer SEVIRI Spinning enhanced visual infrared imager VTIR Visible and thermal infrared radiometer SIR-A, SIR-B Shuttle imaging radar WARA Watershed response.analysis model SIR-C Space Shuttle-based microwave remote sensor WEFAX Weather. facsimile (now LRIT) SLAR Side-looking airborne radar WHYCOS World Hydrological Cycle Observing System SMMR Scanning multichannel microwave radiometer WINTER Wind measurement in the mesosphere SMR Scanning microwave radiometer WMO World Meteorological Organization SIN Signal to noise ratio WRIP Wessex Radar Information Project SOLSTICE Solar Stellar Irradiance Comparison Experiment WSR Weather surveillance radar SPOT Systeme probatoire d'observation de la terre WV Water vapour (France) WVP Water vapour plume SRM Snowmelt runoff model WWW World Weather Watch SSAR Steerable synthetic aperture radar X-SAR Space Shuttle-based microwave remote sensors