CS(AR)-23/96-97 FLOOD PLAIN MAPPING OF PHULBARI AREA USING SATELLITE DATA (ASSAM/MEGHALAYA) a ¥ ” NATIONAL INSTITUTE OF HYDROLOGY JALVIGYAN BHAWAN ROORKEE: 247 667 1996-9710 20 “0 List oF ras List oF Flares List OF PLATES aeview 2 Penures of Flood Plain 2.2 Techniques of Flood Plain eapping 2.3 Flood Plain taping Using Area! Motos 2-4 Flood Plain Mapping Using Sattelite beta 2.3 Visual Method of analysing R.S.Date 26 colour Photezrarty 2 Thenatie mapper vs Ms 2s Flood Plain Mapping Using Digital vara 25 amore Sensing Sateliices Zha'case seudres Geeation Rainfall Temperature and tunidity agricultore forests ‘Soll characteristics opus von. Statonent of the Probiew vemooo.ecy FT Satellite Dace £2 Toposheets AUALYSIS AD RESEETS RECOMMENDATIONS REFERENCES 2 ”List oF TABLES: Land eat Flood Plain indicators Suitability of spectral Banas for Flood Inundst fon Mapping Detaite of IRS Data Progucts Location Map Stuay Aree map Post Monsson Flood #1 of Phuibers Area, 1988 in inundation Map Different Stages of Rivers Jinjiran and the Branmaputra at Phulbert Benkline Migration of the Branmaput ra Land Use/Land Cover of Phulbari Area Flood inundation Map of Phulbari,1992 Details trom Ground Teun Verification List oF PLATES Partia) View of channet-1 confluence of Old & New Channe! at Phulbary Bank Protection et Phulbari Joint Piscussion buring Ground Truth verification 26 a 2 ” a 38Flood 18 @ natural calamity in wnich aost parts of the North Eastern States are ravaged. Hanagenent of the problen in these disastrous flood affected areas often requires toad plain mapping for pratection and resource develoenant. Tate has Iostly Being dene on the easis of ground baved survey which 1s enote sensing tethods especially with the advent of sa Vite era, provide exceTlent periodic infomation in real tine tor these areas. In fact, data obtained fron satellite has now became the only relfable weans for macping the problen areas. Prulbari, one of such problem areas in Assan & Moghalaya within flood plain of the mighty Brarmaputra has been urderao- {ng continuous erosion by the recurrent floods. To conceive protection scheme for Phulbari, flood plain mapping has long been warranted. The roport attenpte to mio the flood plain of the araa with the renote sensing dota for which IFS-IA & 18, 150000 False Colour Conposite have been used. Analysis hes. been carried cut trom the dolineatien of pre & gost wenscon periods of 1968, 1992, and 1996. Survey of India toposhest (195) hae been coppared with the Tatest satelTite data for the study of bankline migration of the Brahmaputra at Phulbari ares, Extensive site visit wns eade for greund truth vorition: tion and interpretations trom inagertes were supplemented,Flood plain is on aree adjoining a mein river course, stream or other water ‘course which gets flooded during periods of high water when stream flow exceeds the carrying capecity of normal channsl uring the process of river bed defor~ imation, it results from bank erosion and deposition of sediments transported by the stream, Natural floodplains provide a temporary storage for attenuating floods significantly. These generally are excellent arees for aquacultural, agricultural and forestry production, pleasing settings for homes, and ready nde trans~ portation corridors that make them attractive for intense human activities, Due to increasing human encroachments, capacity of flood plains to store water tenporerily gots reduced cesviting increased flood peaks on downstream and necessitating costly protection works efore taking up any protection works detailed maps depicting extent of roodplains and water features are required, Also, to assure the effectiveness of ‘existing flood protection works investigations are to be carried out at regular Intervals. Such studies help in detecting possible changes in the terrain condi- tion, land cover ete, due to developmental activities In the flood plains and to conceive suitable steps to deal with the problems In india, an area of about 40 million hectares is prone to floods, out of which 30 qillion could be protected economically according to National Flood Ccommission,1985, The area affected annusily on an average is 7.4 milion hectares. ‘The loas when covered in terns of money is Rs.640 erores. Major flood problem in india fs mostly in the Brahnaputea and ite tributaries and other rivers ike the which env Derek in Ascam. North East of India experiences heavy saint frequent overflowing of the rivers, erosion of the river banks and inundation of towns and agricultural lands. Phulbari area in west Garo Hills of Meghalaye is one ‘areas severely affected by large scale erosion. Here the river of the prot wnJnjiram with its frequent changing course outfalls into the Srahmaputra on its south bank between Fakirgan) and South Saimara of Dhubri District of Assam. Under the guidance of CWC, the Irrigation Dept.Govt of Meghalaya has undertak- fn erosion control project in the area, ‘To supplement the field survey already carried out for the project, flood isin mapping from satellite imageries for different periods have been made. 1 thes been further updated with ground truth verification, ‘The conventional approsches to collect such information have been ground besed surveys which are both uneconomical and time consuming. Remote sensing nta is used with advantage over conventional methods because of ite eapability {0 provide broad synoptic coverage which helped in studying the intonsated effects of various aspects of the ecosystem. It is now possibie to correlate the cause effects of changes which are being monitored and plan for sustainable20. REVIEW : 2.1 Featuree of Flood Plain : Mapping of flood plains of a river system Is based on the {dentifiestion of Various features connected with floodplains. These features are river course, ‘drainage pattern, alluvial fans, oxbow lakes, marsh ereas and back swamps, point bar deposits, river levees and flood plain deposits 2.44 River Course Three distinct patterns of river course are i fentifid viz; (i) braided scream (i) straight stream end (il) meandering stream. 4 braided stream ox be defined as one which flows in two oF mote channels around alluvial islands. Bralded pattern develops after local deposition of coarser material which can not bbe transported under local condition existing in the reach. Straight reach is rarely found in streams for a long steetch. Rivers flowing through alluvial Imaterial follow a zigzag path. This phenomenon ia known as meandering. These atreans follow # more of less sinusoidal path. 2.1.2 Drainage Pattern The drainage pattern may be defined as the arrangement of surface flow channels covering an area, in complete details of thelr density, orientation, ‘uniformity ete. The surface drainage pattern is probably the most consistently reliable indicator of ground conditions available. 2.43 Alluvial Fans When there ia a reduction either in the discharge or in the slope of aa ‘equilibrium stream, the stream can not transport the material carried by it and the excess materials will be deposited. These deposits depending upon the modeof deposition are called es alluvial fans 214 Oxbow Lakes As meanders grow, s narrow neck of land is often cut through from two 8 the stream {0 straighten its course, The ends of the neander that have been cut off are then likely to be chocked with sediments Water fom the maln stroam seeps in to this meander and forms what is commonly known as Oxbow Lake. 2.43 marshy Areas and Backswamps Low lying areas that get inundated during floods turn io to marshes when Hoods recede, Sediments of marshes consists of organic clays. Burrows and plant roots are also its features. Backswamps are developed behind the levees sehen these are overtopped by flood waters. 2.1.6 Point Bar Deposits Bed forms having lengths of the same 0.der as the channel width or grest= er and heights comparsble t the mean depth of flow are known as bars. A bar cecurring on the inside of the bed in al Wisk channel is called point bar. 2.1.7 Flood Plain Deposits During floods, river water rises above the banks and it moves away from the current, its velocity is reduced. As a result a lerge part of the “deposite are called flood plain depost 2.18 River Levees River levees are formed by deposition of sedinents when flood waters ‘overtop the slver bank. The velocity Is reduced causing deposition of much ofthe suspended sediments near the channel, The maximum height of « levee indie cates the water level reached during the highest flood. 2.2 Techniques of Hood Plain Mapping ‘The techniques of flood plain mapping can be classified in two categories. hey are 1) Conventional Technique and 48) Remote Sensing Technique 2.2.1 Flood Plain Mapping by Conventional Technique 1 available for mapping of floodplains by conven- 1 variety of methods tional technique. The appropriateness of @ particular technique can be presume- bly be judged by the extent to which the flood plain information satisfies the objective of the user. Methods of mapping areas subjected to flood can be grouped es + physiography, pedology, vegetation, regional flood of selected frequency, Slood profile snd back water curves. Physiographic Mapping: The concept of physiographic mapping of flood prone on of specific topographic features with flood ie areas is based on the correl ischarge of known frequency. Thus the flood plain, often the lowest topogs bench near a river, has been found to lle at an olevation that is overtopped enerally once in every 1 to 2 years. Where successive levels of stepped topor~ raphy exist, these river terraces or deposits on them may be correlated with particular flood events. Once such a correlation ie established large areas of a sociated sith floods of known frequency and elevation. The method will not ofcourse permit distinction of local variations of topography due to scour and erosion _Soll Mapping: Mlapping of flood areas by correlation of soll type, steatifiation of oposite, or drsinage characteristics with known flood levele requires the samekind of estoblished relationships as those sequired for flood zone mapping based ‘on topographic form. In the field soil characteristics are often associated with topographic features and hence one might be used to strengthen the other. Neggsation Mapping: Many observers have noted an sociation of vegetation with Presunes flood levels, Specific assemblages of plants way sometimes may correlat~ ‘ed with specific water levels. studies have established between susceptibility of species of damage snd flood depesition and duration of inundation. Such physio- logical evidence and the distinet age differences in the flood plain trees pro- stuced by destructive floods suggest the possibility of designating sasiaus Mood heights oF minimus land elevations subject t0 flood from assoclations between fooding and vegetation. ‘exiong)_ Flood Mapping: Experience shows that la a given physiogrephic or hydrologic region, flood heignts of chosen frequencies may be mupved on a rewional basis from records at selected localities. The method is based on obser- ‘ations at stations in the region of the heights above the channel bed attained by foods of different magnitudes. By relating flood heights of @ifferent return Periods to parameters such as drainage arca and mean snnual flood discherse, curves ean be drawn that permit flood heights to be determined st vngnuged sites of known drainage area within the resion. ‘ood Profile and Back-Water Curve: In this method flood lines are delineated by ‘defining flood profile from high water marks or detailed hydraulic computations. To achieve high degree of accuracy detailed topographic information is reavired {or the purpose which may be obtained either by engineering ground surveys or serial photography combined with 2.22 Flood Plain Mapping Throwgh Remote Sensing Technique For food plain mapping two types of information are of primary concer ‘Timely information of charceter and extent of flooding and lidAccurate floodplain delinestion Conventional ground based methods are well established but often time consuming and hazardous to be undertaken before the recession of flood. Hydrol fopists as well as water cesources planners and decision makers often felt handi- about flooding extent and other related ata cil capped in absence of timely da the advent of remote sensing techniques to acquire such date To meet the need of quick and el ble information, modern techaigue of remote senting can be used 19 map the extent of flood pling, to monitor floods in progress and also to predict the occurrence of floods. As in the cate of most renote sensing applications, the best use of satellie capabilities is achieved in the context of well designed multifaccted observing system which makes the best scientific and economic use of available data, The optinum mi and svsface data is situstion specific, 1 will depend upon the size of the basin boing considered, the availablity and relative costs of verious types of data and the analysis of such deta. Remote sensing has been gaining wide spread applic ‘don in post important fields of Hydrology, environment, agriculture, flood, mete collection abiuties. orology ete, due to Its repetitive and real time dat ‘Two basic approsches for flood plain delineation by remotely sensed data 2 available. The dynamic method records floods as they actually occur oF soon after the high waters have receded. It takes the advantage of the fact thet visi- ble evidence of inundation in the near infra-red region of the spectrum remains for upto two of more weeks after the flood, This is significantly reduced noar Infrared reflectivity in the flood areas caused by the presence of increased surface soil moisture, moisture stressed vegetation end Isolated pockets of stand Real Tine Flood Plain Mapping by NRSA: Real tine daca on flood indicating its extent was prepared by NRSA and can be procured at ary time. with only time required for mapping, the outline of the flood plain can be drawn. NRSA prepares such maps on requisition, on cloud covered days it fs difficult tosensing dats which wakes browsing essential before ordering for the data. On tine browsing facilities are not available st the state remote sensing application centres, however from 1998 July taped information shall be provided by NRSA ‘through INTERNET. Preparation of real tine flood maps will then be much easier. ‘the second method referred to as the static approach, utilizes the fect that food plains can be recognized with remote sensing beceuse of permanent or lone term features caused by historical floods. These indicators (Table-2.1) have been ‘enumerated by Rango and Anderson, “TABLE:2.A LANDSAT FLOODPLAIN INDICATORS plana physioncaphy. ‘Ratershed characteristic such ‘density et 2. agree of abandonment of natural levees. 4. Geturrence of stable sand dunes on river terraces, 3. Ghannel coofiguration nd floviat eeomorphic: charac & —— Backswanp areas, 5. Soiismoustore avail & Sell ditterence 5. Vegetation differences. 1 —_Landuse boundaries. 1, -Agnicuttural development. 12 Flood alleviation measures on the floodplain tn remote sensing two types of platforms- Aircraft and satellite are most componly used for earth observations, The data obtained from these platforms are termed as aerial photos and inageries respectively. Aerial photography is found to give satisfactory results especially when it is possible to obtain & series of successive photographs during the flooding and ation becomes difficult in the areas where recession period. However the interpi bodies are covered by bushes end forest. In a wide and poorly forested flood plain, water edges can be easily identified. Stereoscopic viewing is helpful inarent in which interpretation based on tonal variations fs difficult. Black end white infrared flim is superior to the black and white panchromatic fiIm. The ‘Addition of colour aids interpretation in both the visual and near infrated wave lengths where heavy vogetation is encountered colour infrared film appears to be the best available film for inundation mapping(tiayer et ah1978. Alsborne multispectral scanner data has also been sed for flood plain rapping by many researchers. Using aircraft multispectral date and using oaturel Indicators such ao vegetation types, soll types, moisture eifferences and geolosi- cal variations, accuracy of the flood plain delinention wns improved in pristine areas and reduce the costly field surveys. Remote sensing delinestions of Mood Non of various natural and artificial Indicators. alzborne surveying of feod~ Plaine requires favorable coincidence of the Inundation level needed for flooding sod suitable weather for taking photographs. These difficulties considerably al the use of erin! photographs for studying flood-plain inundation over fone river resches. ly after the launching of Landsat series from 1972, Because of the unique and inctive property of water to absorb energy at reflected infrared, it becomes ery convenient to use reflected or near infrared data obtained from Land fanp flood inundation. It may be observed that water as significant lower se Alectance then terrestrial features throughout the near infrared(0.?~ 12 um) portion of the electromagnetic spectrum, Because of this, surface water appears black om rellective infrared images fexcopt for cases of high turbidity. The serial spread of water bodies is relatively easily and accurately delineated from neat infrared date Measurements of covered by & water spread Is function of the pixel size, Suitability of uitferent‘bands for flood plain mapping is given in Table: 2.2. TABLE 2.2: SUITABILITY OF DIFFERENT SPECTRAL BANDS "POR FLOOD INUXDATION MAPPING ‘SpecinAL —_—-Appeanance OF —___“Siavanmiiy oF BAND FLOODED AREA FLOOD MAPPING an Lighter Pair (028 0 0456 um) (048 to 0.57 un) Lighter to Darker very poor MSS-S.Red Lighter to Darker Poor Mss-7.in Datker Goo Colour Tk Composite Greented @ IF Bands) Mottled redysrading to Very Good sreen,blue or aay. lue an IR fal loured as blue and More blue to pusple Excellent 25° Visual Method of Analyzing Remote Sensing Data ‘Aerial photos and imageries represent the permanent record of terrein information at the tine of exposure. Different terrain features reflect different sount of electro“ yenctic energy due to variation in their chemical composition fand physicel condition. This variation in amount of reflected enersy, which is recorded on imagery, forms the basis of its identification, The process of extract ‘ng qualitative information from the interpretation. snery is termed a2 inas ‘There are wight elenents in image interpretation which help in identifying the features of the object of interest. These ared}shape, site, si)patterm, iv)shadow, wone, viltextur9, visite and inlassociation ‘Shape: This is an important single factor sn recogsizing objects from thelr Shape relates to the general form, configuration oF outline of an individual object. A railway can be distinguished from & road because of its shape consists of long straight tangents and gentle curves when compared to the curved shape of highway. WBE: Size of an object ts an usable clue to identify it. By measuring enunknown object In an imagery the interpreter can inete from considera ‘Hon groups of possible classification, ‘iypaern: Pattern relates to spatial acrangement of objects, The repetition of certain general forms or relation shipe is characteristic of many objects both aatural and manmede and gives objects « pattern which aids the inter- preter in recognizing them, iv)Shadaw: Shadows are of important to interpreters in two opposing re~ spects. The outline of a shadow affords & profile view of objects withia shadows reflect litle ight and are difficult to interpret. vyTong: Tone refers to colour oF relative shades of the imagery. It i related to reflectance of light from the obiects. Water which absorbs nearly all incident light appears black where as cement concrete highway reflects a ‘high percentage of light tone without tonal differences, shapes, patterns and textes of objects can not be distinguished. vidTexure: Texture is crested by tonal reflection in groups of objects that fare often too small to be identified as individual objects. It gives a visue Impression of roughness or smoothness. The texture of an image may be coarse, medium of fine. Enlarge seale photographs for example broad leaf tees species can de distinguished from small leaf species on the basis of vitlassocitions ‘The location of objects with respect to terrain features and fothe. objects is often helpful, The open flat dark vegetation is esi ienti- fied as swamp of f2h Jn vegetated area once It ts known that the area of 26 Colour Photography Both Panchromatie (BAW) and colour photography can be used effectively to conduct photographic interpretation, There are special applications where colour Photography can be very valuable. These include geological mapping, geochemical scobotanical studies, forestry studios, detailed Innduse studies where differencein growth pauterns of the time of exposure are of significance, In certain river, harbor and coastal studies colour photo much in use ‘The strong contrast petween water and surrounding sustaces in Landsat band 7 MSS imagery is of much use for delineation of surface waterbosies. Sur- face water features have 4 vary dark tone in this band. However, some times if bisck soils are metwith, s would pose difficulty in delinesting the water bodies fs both have almost the same reflectance, In such eases band 5 imagery can be referred to a False Colour Composite (FCC) Which is of best use in such interps tation difficulties. 27 Themntic Mapper (Ti) ve Multi-Spectral Scanning (MSs) Duta for Flood: studies = ‘The Thematic mapper (TM) data has a high spatial resolution and regarding napping of surface water resources, it renulls in improved surface mater detec ‘on copability both in terms of location and water sprend fluctuations. The tiner evolution inl 8 smaller percentage of boundary pixels for the same water Spread area compared to MSS, leading to improved aren calculations. The increase from 4 bands in MSS to 7 bands in TM sensor, theoreticelly doesn't add to Improved surface water mapping end monitoringusing visual interpretation tech ‘ques, since band 4 of TH enalogs to band 7 of MSS is one of the best for this activity In general the black and white images trom near infrared wave lengths provided best visual interpretation capabilities. FCC products provide additional Information on surface water ity which are not easily obtainable from black fang white imagery. Studies have shown thet TM bend combinations 345 83.47 fre better than 2.3.4 combination, for visual delineation of water features, Dieit Imapping as indicated that band 7, in combination with MSS anslogs bands prom vides better results, Surface water mapping Is possible upto 1:$0,000 acale using ‘currently available TM 4: enlargements Inthe case of food studies, embankments, spurs and other control worksfare more clearly discernible on TM imagery as compared to MSS imagery because of the finer resolution of TM date, Recause of more number of TM bands, better ifferentiation between wet/soist, dry areas, standing water damage scares, sand fast areas etc is possible, Anong the FCC band combination 4 $ 7 combination is 1s from other inundated very useful especially for differentiating sandcast « arena, Mai nding ‘areas are enhanced although different classes within the water are sot Identifiable. While TH FCC with band combination 2.3 4 fs compara Dle with Landsat MSS FCC except for higher contrast in the former, TM FCC with band combination 34 5 is more useful to delinente dry and wet crops. TM FCC 3 4/5 is also useful to delineate command areas under tanks and inundated aces under crops. General flood inundation mapping can be done at 1:250,000 seale but finer details tke breaches in embankments end detalled river configuration stud- Jes are possible at 180,000 scale. It Js alto possible to distinguish more aumber of classes of damaged groups and partially damaged crops by digital analysis fone of the distinct sdvantage of Landsst TM date i boundary delineation of afferent strata of inundation due to increase in resolu~ ion. The individual bands bring out certain wniave aspects of vegetation cover as manifestation of physiological and phrenoiogial variations. Considering the spectral separeblity between land water and vegetation is more pronounced in the visible to middle reflected region of the clectso~aagnetic spectrum, the different multispecteal band combinations of Landsat TM increase the scope for ientification of more number of land cover classes. Disital tech higues ore more useful for differentiating between overlapping classes like forest 428 Flood Plain Anelysie Using Digital Rete = Use of digital date enables an eutomatic classifier to be developed. There sre (wo basic approaches that use pixel by pitel analysie to map water bodies.This method is fest and easy to set on computer but it may result in serious Iinclasification if there are ares adjacent to the flooded area that have sinilar Jom reflectances. The second approsch is to develop a classifier from training set fata This approach is to minimize the misclassification by taking advantage of reflectance data from more then one spectral band, In addition, this approach my De the only reliable way to differentiate sediment laden water from adjacent land, Digital mapping has indicoted thet bend 7, in combination with MSS snalogs nds provides better results. Digital tochniques are more useful for differentiat= ing between overlapsing classes like forest cover and plantation, plantation and crops ete. where interpretation cannot discriminate, 29 Remote Sensing Satellites: ‘The satelites im the space are capable of providing stable platform with Sistortion free images covering # larger area under uniform illumination condition 1 Tciltate easy recognition of major features of the earth surface. Since satel lites orbit around the escth/rith the rotation of the earth, it has added advan {ase that it Js possible to obtain repetitive coverage at periodic interval Siferent seasonal and illumination condition. Based on the purpose and objec lives, the satelites are elassifieg into: L.Meather satelites/meteorological xatllites (etsst) TIROS-I/MIMBUS, NOMA, INSATAD 2Renote sensing satellites — IRS, LANDSAT, SPOT, ERS Marine resources satellites - SEASAT, MOS specific purpose oriented satellites - RADARSAT, SPY satelites ‘nother txpe of classification Which i¢ common with the satelites syst ‘based on their orbital characteristics viz, Polar oxbiting satellites and Geostation- ary sstellites, Polar orbiting satelites circle around the earth from north to south and the orbit is near polar. Most of the Remote sensing satslves are of‘his nature which facilitates ear scanning of the earth surface at periodic inter~ ‘yale, Example: 1RS, CANDSAT, SHOT, NOAA, ete. The Geostationary Satellites fare positioned in the space in such a way hat the satellite orbit is synchronized sith the spood of the earth sesulting in continuous abservations of the same spot fof the earch. Example: INSAT 2.9.4 Indiana Remote Sensing Satelite (KS): In India, Department of Space is engaged in space research activities Goving the past decades, One of the off shoot of space activities is the desian development and management of Remote Sensing satellites. The first Indian ‘experimental satelite for remote sensing, designed and developed by Department lof Space was BHASKARAT, Inunched from Soviet Cosodrome in 1979. The second satelite, named BHASKAR-IC was launched in 198i. These satll an altitude of $25 km with circular orbiting characteristics carrying sensors composed of oo television cameras and theee microwave radio meters. Parallely. OHNE geries of satellites designed nd fabeicated by the Department of Space were also launched during 1981 and 1963. The experience gained with the lounching of experimental satelites such as BHASKARA and ROME gave enough confidence to desian and develop indian Remote Sensing satellites (IRS). The IRS-IA which is a representative of the first series of remote sensing satellites for resources survey ond monitoring waa laueched on March 17, 1988, This se lite & sun-syachronous, polar orbiting at an altitude of about 900 kx. wi repetitive cycle of 22 days with equatorial crossing time of 10:00 AM, This satel lite is expected to carey a pavlosd consisting of one low resolution (72.5 m) and two meslium resolution (36.25 m) LINEAR IMAGE SELF SCANNING exmeras utitz~ ing solid state near arrays operating with a "PUSH BROOM" sean mode. The moltispecteal data in the IRS-1 aystem is collecting data i four different. spec- ‘rat regions ranging from 042 m to 0.86 m. The spectral bands chosen in IRS-A fare closer to those of the first four bands of thematic mapper in Landsat-5 and also those provided in the French remote sensing satellite, Spot. Second Indianemote sensing satellite IRS-18 was launched in AvEvst, 1991 since the life of INS-IA was to expire in 1991 though it still wocking. So IRS-I8 wos ltunched 95 4 complimentary to IRS-1A both having similarity in various spectral and ‘orbital characteristics, For aeguisition, storage, retrieval evaluation, dissemine- \Yon and training for utlization of IRS data, XRSA has been identified to play the key sole. Encouraged by the past experience Degartment Of Space took up the ambi- tious challenge of developing next generation satellite namely IRS-10 with in proved -ASOF and coverage capabilities to mest the graving needs. The populati- ty of satellite based remote sensing has eretted @ need for providing data with Deter resolution, coverage and revisit. 1RS-IC was conceived to meet these de- ‘mands, Two satellites IRS-HC and 1RS-LD with similar payloads each with a life Inission of thees yours were also planned, Principle components of the mission are ‘ Uhroe axis stabilized polar aun synchronous satelite with these sensors, ground system for in-orhit satelite control and the ain objective of the mission is to develop new arcas of user applications to take full advantage of the enhanced resolution and capability of IHS~IC/D sensors. The pay load platform accommo ‘lates the Panchromatic camera (PAN), Linear Imaging and Self Seanning sensor UL (L1SS-I11) and Wide Field Sensor (WIFS) cameras. In addition it accommodstes Earth sensors and star sensors RS-P2 is nlso placed in sun-synchronous orbit at an altitude of about 817 Km on October 15,1994 by PSLV-D2 from Shriharikota vith » ground tracing veloe= Hy of 6467 Km/s. The 804 kg satelite carvies two cameras LISS-IIA and LISS=118, wortding images in four spectral bands similar to IRS-1A/18 without 1188-1 sensor. INS-P2 sensors are configured on board with ew ecbsicharge Couple Device} placed in the focal plane of 4 single optics of each band thus sesulting jn omy one electro-optic module, This provides the working capability of L155: ‘of IKS-IA/1B with reduced volume and weight. 2.10 case Stodies:Remotely sensed data for flood plain mapping have been used in india by various investigators, Using satellite data (Landeat 1 and 2) taken during 1972 land 1975, Dhanju (1976) studied the shifting meandering and flooding of Kost iver. From the imagery taken in the flooding period he was able to map che inundated areas flood boundaries and other associated features. Chakraborty (1978) has observed that courses of interlacing channels of the river can be ‘etsily delineated through band 7 imagery lone. Sand § anelysis of airborne multispectral Seanner dats was found vo be useful in obtaining detailed tnforma- tion regarding flood protection works. Further applications of remote sensing methods on Gangs flood plain mapping indicate the usefulness of satellite data for elineation of the features like inundated areas, backswemps, oxbow lakes, river levees ete Lewatly, aerial photogcaphs of 1:30000 to 1:60,000 are used. Some of the ties dealing ith flood mapping with aerial photographs oF with data collected with airborne svrveys are mentioned. Bhattacharya end Mankhsnd(1979) have delineated geomorphic units in the flood plains of Gangs Gomti in Aramgarh and Ghesipur districts In Uttar Pradesh. chopra(i9#0) had Gelinented an abandoned meander belt close to the Ranganedi river in disteict Lakhinpur, Ascom. Such recognitions are important as these indicate the channel niigration pattern and susceptibility to floods. Ae the instance of Ministry of Water Resources, Govt. of india part of the anga in the vicinity of Farakka Barrage was surveyed in May 1981. Aerial Photomosaics and river configuration maps were prepared and supplied to Farak- ke project to take timely flood protection and anti erosion meatuces and to under take model study at CWPRS. In order to provide data produets to Ganga Flood Control Commission scanner end photogsaphic survey of Gandak river below International border to ite conflvence with Ganga was wndertaken in December 1982. This survey showed that the roach of Gandak which was surveyed has Uundergone considerable changes {fom what has been shown in the toposheets of ane1st. Nations! Remote Sensing agency arrenged 2 umber of aerial fights both With modular multispectral scanner end serial photography over toed prone rivers i the country (Ramamoorthi, 1993). At the instance of Central Water Commission, an aerial photography and acenning of certain reaches of the Brak Imaputea river was undertaken in August, 1978. The Brahmaputes is normally at its highest stage in August. Aerial photographs and scanner visicorder photo ‘outputs covering the Brahmaputra river and adjoining area near Jogighopa-Goat- pera reach inciuding the photo mosaic were prepared. computer analysis of igital data of four different channete were also done. ‘The thematic photo out PUIS provided significant hydrologic information about the flsad plain Features, Submerged areas and situation of river when it waa in high stage, ‘The survey was repented in November , December, 197# for a stretch of 60 Ke from Sediva to Dhubri, This survey was dene to provide information to the Brahmaputra Flood Control Commission for etfeetive flood control planning. Dhenju (1976) made visual Interpretation of Kosi River Flood Plein with Jandsat imagery. Areas under inendation by floods during August/September 1975 were delineated by Dhanje using Oct 1975 Landsat Imagery. In October some areas were stil under water due to low gradient and poor drainage. Dhanju delineated flood plain features.of the Gangs basin Iving between the latitudes of 29° W to 27° N and Ionpitudes 63° E to 85° E by visual interpreta ‘ion of Landsat Imagery. Interestingly this study inferred that frequent inund: on of Ganga Kosi Area might result In Kosi joining the Ganga above Bhagalapur bend instead of {ts present point at Kursela, distance between then being about 50 Km. This forecast may or may not materialize but it demonstrates the potent lites of such studies Multi stage remote sensing dats has been used to study the Mooded costal ‘areas of Andhra Pradesh during Nov. 1977 by Narain and Petel. Landsat imagery was used at the stage-l and the acrisl imagery at stage-2. Diazo esl ‘te and biack and white paper prints were used for deliestions employing visualinterpretation. This study has 1 avick assessment of damage cevsed by flood. It has been estimated that such study would cost Ra.t5/- per Sqm, Study of Hamamoorthi and Subba Rao (1963) Is among the few where both Gisital and visual interpretation techniques were employed in flood plain mapping, tn this study, dealing sith the Sahibi river flood in Delhi Harayans in Aug, 19 outputs were produced based on ieproved contrasts between water/wet surfaces and surrounding dry ares, Rand rotstion and contrast stretching were resorted to, Colour thematic outputs showing the pfe flood and pest flood conditions of the Mood affected areas wore prepared in 1:280,000 scale. However, it is worthy to note that there is not such difference between the aerial extent of the flooded areas delineated from satelite data and from photo Chaturvedi (1983) delineated the flood inundated areas in parts of Southern and castern Uttar Pradesh during the peok floods of September, 1982 on the basis, of sharp tonal contrast between the water spread and the adjacent arens, 1 was also possible to delineate areas from where the water had just receded. es and ponds was particularly noticesble ‘The impounding of pre-existing dry Is in the vienity of fonded rivers. Notwithstanding, the limitations on account of cloud cover and resolution of present day satellites, it was stil possible to make accurate assessment of flood affected arens ‘share et ol (1985) visually analyzed the band 7 infra-red landsat for mapping flood plains and allied features in the Ganges between Allahabad to Cchapra (thar). The features were identified in the 1:250,000 scale enlargement of bend 7 imagery The serial extent of the flood plain area in the study reaion was also computed which came out to be 3800 Sa.km, It has been observed in the imagery that flood plain of river Ganga increases immensely st downstream of Veranesi end reaches ite maximum at chapra. The river Gomti which meets with ver Ganga about 40 Km. downstream of Varanasi is iceelf famous for its devas- tating Moods in the study regions.3.0 STUDY AREA + 2.1 Location: The study area Is geographically located between 25°S0'end 26°0" of north- fern Initode and 90 snd 90°13" of eastern longitude. Its situated in West Garo Wills of Meghalaya ang Dhubri district of Assan and at s distance of 250 km from Guwahat. Location of study area and ite IRS imagery index is shown in Fig 31 & a2, 3.2 Raintat Average annual rainfall in the Jinjirem sub-basin is about 3226,70 mi November to February are the dry months of the year, Spring stins are fre~ quenty high which About 26% of the annual rainfall June to August are the wettest months of the yesr and rainfall during this period is about 62% of the fotel annual rainfall Rainfall and related date regarding the Jinjiram river and the basin collected from CWC, Phulbact at the time of ground truth verification Maximum daily eaifall = S140 om Minimum daily rainfall = 0.256 mm Total monthly rainfall (ax) ~ 818.730 mm (September 1925) ‘Total monthly rainfall (Mini) 8.12 mm (anuery 1996) Maximum water level = 2350 m (uly 1995) Mininum waterlevel ~ 2312 w (February 1995) Mexinun discharee = 3864.40 m/sec (suly 1986) Minimum discharge = 4.46 m2/see iMazeni996)Tig B. LOCATION OF S1UDY AMER ON BRARMAPUTRA DNBITS NG IAGERY IDE2.2 Temperature and Humidity: Wis seen that in the sub-basin annual average temperature ranges be- (ween 295°C and 19,7%C, The temperature starts rising from the beginning of March and reaches maxieum in July and Avgust. Maxinum & minimum temperatures are 28.5 and 19.7", [Relative humidity observations indicate high aoisture content in this repion for which the wenther is very oppressive. Relative humidity in the tainy season (atay to September) is betwee 128 to 85K. February to Apri are relatively dry with relative humidity ranging between SOX to 79%. 24 Ariculeare: plains paddy is the wain agriculture crop. Most common variety of paddy is Sai, Abe ete, and other products aro pulses, potatoos, maize, wheat, sugarcane and Dbamanas. tn the portion of Garo lll district paddy is the main product other ‘haa orange, banana and cotton. 35 Forests: Above 608 of the sub-basin is covered with thick to mild forests. It is Very thick in the hilly areas of the state of Megha a mhere numerous wild ‘animals lke clephant, tiger, Jecksls end deer are found yundantly. Forests ere covered with vory valuable timbers like Sal, Chegun, Gamari ete, Some people Deve taken forestry as their maln occupation. Afforestation is taking place in the slain arens of the basin for environmental reasons besides protecting the soil from erosion due to heavy rainfall. Forestry is mainly carried out by the Social Forestry. 26 soll characteristice: Eventhough detaled soil survey in the Jinjicam sub-basin has not been -2280under taken, soll of the sub-basin may be classified as New Alluvin, Old aliuviel and Red Loam, New alluvial type of soil is predominant in the plains while red loam & old alluvisl soils are seen In the hilly areas of the Jinjiram sub-basin. It Is necessary to under take soil survey to ascertain broadly the soll characteris- ‘os to agcertein soil groups end finally to arrive at a scientific cropping pattern 3:7 Population = Based on the figures of 1991 census the plain areas have high population Gensity while the hilly areas are thinly popblated. The density of the basins 300 persons per ag:km. in 1981 and 467 porcons per sq.km in 1981. These are agske in the respective census years. Percentage krowth of population of the basin between the census years 1971 and 1991 ia estimated to be above SOx. 2.8 statement of the Problem ‘The Pholbari aren where the river Jinfiram outfall into the mighty Brah- raaputee has been subjected to severe erosion and floods since long past. The aren at present is ers ‘erosced by number of new channels/spill channels chang- ing its courses frequently end in the process eroding away new areas on one side and creating sand chars on the other. Protection of the Phulbari township has been the major concern of the people and the Government. The problem has been examined in details in Chapter the ‘Analysis & Results.40 wetnopoLoay : in the present study visual interpretation technique of satellite imageries vss employed to delineate food plains and inundated areas at Pub fes from Indian Renote Sensing Satellites (IRS) were obtained fsom HRSA, iyo ‘erabad. Comparison of pre-flood and post-food date helped in better identifica tion of the flood plain features, In visual interpretation, tone & texture variation forms the basis of identification of flood plain and inundated areas, Flood plain maps prepared from imageries were found to have slight geometric distortions ‘and were corrected by optical projection techniques with reference to the base ‘aap prepared from survey of India topographic maps. Flood plain maps prepared for the years 1988, 1989, 1995, and 1996 were studied for detecting various changes that has occurred in the study ares round truth verification was conducted with the help of CWC, Silchar Division. ence materials also used, ‘Toposheets, field survey maps and other available ret sltes were visited and cccurrence of features verified wherever more transperan- Cy was felt for daring interpretation of inageries. 41 Satelite Data + ‘The false colour composite (FCC) of IRS, LISS-IT data generated on two ‘ifferent scales were used for delineation. Details of IRS dota product used are‘TABLE NO.t.1: DETAILS OF IRS DATA PRODUCT USED pate Seale” P baie PFath Row me Eee ev LIRS= IA -29.488 1350000 FECT = 80 2 RS 1A CLONES 180,000 FC 17 ~ 50 SIRS = 1174292 11800000 Film A. ns —18 09s 1000 Fe SD IRs = 1B 04.02.96 1390000 Feo 17 - 50 42 Toposhects ‘The study aree has been’? ineated'from 78 K/1 (1:80,000) survey of Indie0 ANALYSIS AND RESULTS Flood plain of Brahmaputra nd ite tributary Jinjiram extends up to the foot of Garo fille st Phulbsri Area where elevation rises due to starting of tills after an extent of flood plnin (Fig.S1). The minimum and maximum distance be- tween Brehmeputra and the flood plain is 2 kn and 11 km at Phulbari end Pushkarnipara respectively. Unlike other parte of the country's riverine areas, features of tow lying flood pining of the Brahmaputea are extensively visible uring the post monsoon period Initially Jinjieam river was flowing through Masipes channel (Pig:5.2) etarced flowing along ite course through Mochhumsil The proc fess of formation of new channel has taken place between 1988 and 1992. tn the process, it has occupied » considerable portion of human habitation of Phulbari township and village Mochhus was completely engulfed by erosion. Phulbaci town and its surrounding areas which ie between the old and new channel of injicam totally semain under submerged condition for a period of ninimun two Imonths after the fived and thus exesting enormous hardship to the inhabitants of the aren. The sime condition prevalls for nesrty 70% of the study area. ‘The Arehnapstra is joined by sinjiram siver (Fig.S.2} at tho pointe during the period 1990 to 1992, The first one wes by washing away the towns Sabet Sororbatarl and destroying the horn shape of the Jinjiram river, The second ‘was through an old alluvial channel which was active long back ego and to pass rough south Solmara. Ia the processes of Jinjleam mer ing into the Brahamapu- tra, a fertile land of 4.0 aqcke has Deon ozeupled and the horn shape of the river which was present at the North of Phulberi town started disappearing uring this period rhe following interpcetation (5.9) has been carried out with the help of 5.0.1, saps compared with the imageries1. Shifting of river Sanks during 1970-96 was studied. It reveals that the ranmaput has extended its ban ine to a distance of 24 kim at Phulberi by ashing away the town South Selmera. At Dhelemara, Sorfamont and Pho! shifted a distance of 3.5 kia, 7.0 kum, and 7.5 kim respectively. tn this period the Brahaputea has eaten away vast fertile land which includes tno bis towns Fakirkanj, Golaberiage and nearly fifty villages that were existing in the study 2 is soon that (Hg.5.3) 1958 bankline was extended to m sixinum distance of 12 km. towards south from the 1989 bank line, This was due to high flood (25600 w3/sec) occurred tn 1968, 5. The Beahmaputea water carries huge amount of suspended soll. tn due course of tine altuvi deposits take place in the low lying areas where water stagnates for a considerable period, Alluvial deposits thet occurred in large arce 4m and around Pholinori, show that these arcas were under su wer208 condition Tore long pesiod after the flood, It Is seen that from {992 to 1996 ores under submerged condition exlended to @ considerable portion than the period 1988 to 992, Im this process the bow shaped portion of river Jinjiram at Pholimori has been covered under sand which induced Jinjiram 9 divert ite flow 4 dis-ance of 1.25 to away from the former 4. Ie is seen thar the sranmaputra river benk line has been shifted by about 1.2 km 41 km, in five years at Nemashaln and Pholimant aren respectively, This necessitated bank protection measures in this reach, 5. Villages Paikarpara, Janadarpora, Natanbasti, Sasnlya, and Tharangaibar Ines been identified as flood risk places 6 Compariaon of yest sunsenn seasuns Of 1988 (rig.5.4) end 1D CHB.) revels that 218, 8%, 35% and 75% 8, 10% of the study area ie under seversiys vroderately and partially inundated condition respectively. The alarming rate of Increase in severely inundated condition is dve to the fact that the Brahpeputra hes occupied more lend surface by extensively migeating its bankline whichfnturn caused drainage congestion. In the pre-monsoon season (FBSA) nearly TOR and 15% of the area was under fallow and vegetation condition, Flood inunda- tion during post monsoon period of 1992 Is mapped as In Fig.S.S. 7 It in seen that 20 kin. length of embankment existing along the Bratmapy- ‘tra was rooted out in 1988 due to the high flood occurred uring the period. 8, At the time of ground eruth verification (Fig.S.6) st is seen that PHD. Meghalaya has Initiated construction of embankment across new channel and slong the down stream of Jinjlram to protect Phulbarl from high flood. Locetion of deflectors has been proposed In such a way that Incoming water from the Upstream of Jinsiram join its course with the Branmaputra. A stream from the Deahaaputra parses towards south end joins with Jinfiram named as channel Another stream from little down stream of the Brahmaputra and perpendicular to ‘the channel-l digged its way into Jinjiram which is termed as channel 2 by the PWD officials for cartographical purpose and its identification. Development of these channels which are of the size of large rivers has further complicated the problem,moetmy vr rarer ge eg nt ing rear hyCD coe bt eet BB vga a360 RECOMMENDATIONS + 1 The bankline studies for m mighty river lke the Brahmaputca is very diti~ ett I8 can be cartied out by continuously updating the chenges occurred with ‘he Rep of the now techniques of remote sensing using Mulidate suluitenpor satellite imagery. 2 Using satelise imagery aunerous riverine istends, locally known oa “Chars san be located and the deflection of river flow duc to these chars can be visual feo. This helps in planning and effective positioning of river training measures such as spurs, bank revetments pitched islands, surface and botion pang ete % Satelite images ond other remote sensing data may be analyzed yearly before and after monsoon season for monitoring the fuve-norphologieal ehangca 70 arrest bank erosion effective anti-erosion measures may be undertaten with the help of mathomatical modeling studies with date input trom remote sens- ing techniques, specially for areas having steep slopes 5 Future studies can be made to correla Fiver turbidity and erosion poten Aint in affected areas using satelite data, & _Afforestetion end scil conservation measures aay be taken up in priority Sania Jn the arons for which & trend i established by renote sensing techniques 7 Fermanent solution for protection of the area is however lies only in mmedersting the flood peaks of the Brahmeputra by construction of series of Storage reservoir projects. caseade system of reservoirs aa practiced in river Danube may be cost festive for which sites are to be thoroushly investigated 5 The ongoing protection messures taken up bY the Gort, of Meghalaya for a vrnell crsich abould be 6 gcod example and experience to be extended further, However, the aren being intesostate Govt of Assam may also join hands for a comprehensive protestion planning % Watershed mangement mensuses however are regional propositions ond are stective if eacrled out at upper reaches of Beahmaput This may call for inter tote collaboration much more effectively than at present mas9.0 REFERENCES ‘Abidoo A. (1976) Satellite survey of Particulate Distribu- ion pattierne in Lake Kainjitewote Sensing Environment, vol. 5. yp 109-123, Agarval C$. and axK.kishra (1987}:"Visual Iatespretation ofree Satelite Date for Chana! Migration and Water Lob ine conditions. Along Ghaghra s Terni Rivers in Part of Disst. donde and ahasich, U-PsPRotonirvechae Vol. 13, Not, pp 198, Ahmed Lanth thesis for Mester of Engineering, At.T Fenskot, 1983, lor 6.3 and Searpace #1 (197) "improving stream Flow Estimates through uce of Landsat”, Satellite Hydrology, 3th hnnal Willnm #-Pecora Memmorial Symp. on Remo.Sen.r Amer, Suter Reso.Arso,, Minseapalis, pp 288-291. Balakrishnan o.(19H6) ISRO report ANEMS TH-67-86 Bangalore. Barret £.C.und Curtis Li. (1882)Introduetion to Environ~ fmentel Remote Sensing, Chapman and Hall, New York Brekke TJ. end Kanenosu E.7, (1981), Insolation Estimation from sattcite Measurements of Reflected Radiation, Remote Sensing Envins Vol. ir pp 187-18? Berg, Pu DR wiesnet, Me Matson (1979), "Assessing the, Red aver of the North Plooding from NOAA Satellite Data", Satelite AydroloRy, pp 308-315. Bisarin M.S.» (1986),"Review of Flood Dereeges in the Country Proce: of Workshop on flod damage Assesement, Deli Burgess L.0 (1967),Inter.Confe-water for Peace 1(4),867-881. Ccontral Hoard of irrigation & Powe Gamage assessment-rrocedings-NewDell rorkshop on flood 386. Chakraborty A. (1979), "study of Kosi River characterist te using Aitborn/space Orbital Multlspectzal Scanner Data", Symposium on the conribution of space Observations to Water Resources studies and Management of these Rescurces,COSPER, CChaturvedt B.S. and Mohan #.(1983), National Symposium on emote Sensing Developenent and Management of Water Resour cee, Inter Remote Sensing; DeheaDun, pp 79°87. Choudhary AM Study of floods in Bangladesh and India with the help of meterological satellites 1980. Das 5. and R.tee (1988)," nonctraditional methodology for food ataze-damage calculation", Water Resources Bulletin,2. 2. 2s. 26. 2. ». Journal of awa, vol. 2 avis Dav, (1979}-"Gomprehensive flood plain studies using Spatial data management techniques",Technical paper Not? HEC, california, Deutsch, M. aod F. Ruggels (1974, "Optional Dete Processing tnd. Applications of the ERTS-! Imagery convering the 1972 Mississipi River Vally Floods" Water Resources, Bulletin, vl 10, No. 5. Deutsch, M. and F#LRuggles (1978), “Hydrologic Applications Of Landsat Imagery Used in the Study of the 1973 Indus River Flooa, Pakistan, Water Resources Bulletin, Vol 14 Now 2, pp 261-273, Dhanju ats. (1976). Remote Sensing’, studies of tos! aivor Flood Plains by /arology Review, Voli. No. & pp 43-8 hove Resources",COSPER,Bangnlore, pp 215-218, ElesAbi Mand J. Rouse Stage oF Discharge (1383)/"Flood Plain Managenent journal of Hysslogy, ASCE,VO! 640.3 El-Jabi N. and J, Rousselle (1987)," A Flood Damage Mode! for Flood Plain Studies”, Weter Resources Bulletiay Jounal of awh, ol 23 ‘gman BP. and Gurney RJ. (1991) Remote Sensing In Hydrolo fy, Chapma end tell, New York. France M.J. and Hedges P.D. (1985) tiydrological Conpari- Sion of Landsat TH, Landsat MSS, Black and white Aerial Photography ‘Symp. of Lenote Sensing for esoress Develop ‘nent and Enviconmental Management, Enschede, The Netherlands pe 71-72. Franeworth RK Barrett E.C., and Dhanju M.S. (1964) Appl Cation of Remote Sensing to Hydrology including Ground Water UNESCO THP Project 41S, pp 122. Genge Flood Control Commission (1986),Classification of Flood. Damages, Proc. Workshop on Flood Damage Assessment, Newoethi. Gautier Cy Diak G. and Massie 5. (1980) "A simple Model to fstinate incident solar radiation at the surface from GOES Satelite Data", JAppIMeteorol, Vol. 19, Bp 1005-1001. (Ho (1982) Glossary of Hydrological Terminalogy with defini toes, Berne, Grade RJ. and Ranga kaju K.6., Mechanics of sediment trans port and alluvial stream probloms, craw Hill Book Company.a 2 as. a 2. 6. Haralick RM Wang S.and Shapiro 1.0. (1985) “Extration of Drainage Network by Using @ Consistant Labelling Technique Remote Sensing Envirn., VoL 18, pp 163-175. Hatifield J.L.(1983),Advances in Irrigation, Vol. 2, Acadam- fe Prese, Mew York. Hellman J.L. and Moor D.C. (1981) Satellite Hydrology, Amer. Water Resor. Associa pp A46=A54, rst P.t.werdenkanp DB. (1966) Journal Hydrology Vol, pp 26-272, Jackson RD-(1981) Int Geoscl, Rem. Sensing Symposium, aigton Bit BP 465 Jackson 4, (1980) Jour ‘lon, ASCE (IR2) pp 81-92 of terigaton and Drainage divi- Jhonson RW. (1980) Renote Sensing Environment, Yolo, No, pp 197-209. Kelubarme w.a.,Sabal 8, and Rapat WV (1983) “Remote sensing Of waterlogged and Salt affected soils in Mani Command area” Proc. of Nat. Symp. on Renote Sensing in Development and Management of Water Resources,Ahmedabady pp 182-190. Khorrem S. (1981) international Journal of Remote Sensing, Wal 2 Nod, pp 145-153. Khorran$, (1985) Photogrametry and Remote Sensing, VoLS1, Now ly pp $362. Kyros J MDeutsch, P.t. Hanson (1979) "Hood Applications Of Satelite Imagery", Satellite uydrology, American Water Resources Association, pp 293-201 Linsely R.x.Kohler M.A.(1582) Hydrology for Engineers, McGraw hill Book Cow NewDelhi. NRSAnRemote Sensing of channel changes of Brehnaputes iver, Hyrabed, 1988, Meyer Wand Weleh Ri (1975) Manual of Romote Sensing, ‘Aperican Society of Photogrametry, Virsina.Plate ~ 1 Partita! view of Cnannel~1 ang erosion ‘at South Salmara Piate - 2. Conflunce of até and new chennel at Phulbari thosonstruction of embankmer Plate - 4, Scientists and Engineers in discussion while ground truth verification, ~tAl t=