AD-A284 059 N PAG For Approv

04(8.B~ NaA.qo. CC 20i43

- I. GEN'1. REPORT TYPE AND DATES COvERED1.AIIAugust 1994 gratfc Paper4. TITLE AND SUBTITLE S. FUNDING NUMBERS

The Use of GIS and Remote Sensing in Groundwater

Exploration for Developing Countries

6. AUTHOR(S)

Timothy Minor, Jerome Carter, Matthew Chesley,

Robert Knowles, and Per Gustafsson

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) II. PERFORMING ORGANIZATION

U.S. Army Topographic Engineering CenterREOTNMR

ATTN: CETEC-PAO7701 Telegraph RoadAlexandria, VA 22315-3864 R-236

9. SPONSORING/ MONITORING AGENCY NAME(S) AND ADORESS(ES) 4ACCA4 POT UME

NTIS CRAMIDTIC T ABUnannounced Li

xnT C Ju'stiticcltion.................11. SUPPLEMENTARY NOTES OF- G

By .................Dist ibuition I

112a. DISTRIBUTION/I AVAILABIUTY STATEMENT 12bi a. 0o$"MWLtja" 01s

Avail arnd/orApproved for public release; Dist Specialdistribution is unlimited.

13. ABSTRACT (Ma Nimum 20 oosis)

The Damer Research Jedvm (DRI) is developing anatepes diat Wumert various data typesto disracmini grounewaftr P a -ir, a for die Wehfcazama of cand~ismt wel locations in centralGhwana Won Affis. Study anes, were aellecied oae which remote nseing data of various scaleswene aembled along with all available bydrogeodopc information. Waw sanAmes for cbeoncalanallysis wat coll~ected and uainnrou field obseavazions were recorded diuing the survey of 200wet and dry boreboles uaing Global Panosiwnng Systm (O3PS) receivers. AMl data collected wase; 1 o1 aduad insgapazed into a Geographic Information Systm (GIS). Relatimships betwee foa-ances mkou in -m uenog dama boraohes necordls and odur hydopoiloecharaceristics embeing aowndl manrus uwo deveilapemet efilcienmiin. The (IS inmhing results am beingtasted with an ongoing dullng p mi Pbm - Preftizniny resulo indicatea a high corellation between drll-ling succes and prdimiy to Landsat-derived lineammnts.

14. SUBJECT TERMS IS. NUMBER OF PAGES

12

Groundwater Remote Sensing, and Geographic Information I6. PRICE CODESystem (GI15 ___________

17. SECURITY CLASSIFICATION I15L SECURIJTY CLASSIFICATION 19. SECURITY CLASSIFICATION 20. LIMITATION OF ABSTRACTOF REPORT OF THIS PAGE OF ABSTRACT

unclassified I unclassified I unclassified___________NSN 7540-01-280-5500 Standard Form 298 (Rev. 2-89)

pnqm.g4b" ov Ifu U0E Z39-16

296-102

b -

GENERAL INSTRUCTIONS FOR COMPLETING SF 298

The Report Documentation Page (RDP) is used in announcing and cataloging reports. It is importantthat this information be consistent with the rest of the report, particularly the cover and title page.lnrtructions for filling in each block of the form follow It is important to stay within the lines to meetoptical scanning requirements.

Slaock 1. Agency Use Only (Leave blank). Block 12a. Oistribution/Availability Statement.Denotes public availability orlimitations. Cite any

Block 2. Report Date. Full publication date availability to the public. Enter additionalincluding day, month, and year, if available (e.g. I limitations or special markings in all capitals (e.g.Jan 88). Must cite at least the year. NOFORN, REL, ITAR).

Black 3. Type of Report and Dates Covered. DOD - See DoDD 5230.24, DistributionState whether report is interim, final, etc. If Statements on Technicalapplicable, enter inclusive report dates (e.g. 10 Documents.nJun 87 - 30 Jun 88). DOE - See authorities.

Block 4. Title and Subtitle. A title is taken from NASA - See Handbook NHB 2200.2.

the part of the report that provides the most NTIS - Leave blank.

meaningful and complete information. When areport is prepared in more than one volume, Block 12b. Distribution Code.repeat the primary title, add volume number, andinclude subtitle for the specific volume. Onclassified documents enter the title classification DOD Leave blank.

in parentheses. DOE - Enter DOE distribution categoriesfrom the Standard Distribution for

Block 5. Funding Numbers. To include contract Unclassified Scientific and Technical

and grant numbers; may include program Reports.

element number(s), project number(s), task NASA - Leave blank.

number(s), and work unit number(s). Use the NTIS - Leave blank.

following labels:

C Contract PR Project Block 13. Abstract. Include a brief (MaximumIG Grant TA Task 200 words) factual summary of the mostPE - Program WU Work Unit significant information contained in the report.

4 Element Accession No.

Block 6. Author(s). Name(s) of person(s) Block 14. Subiect Terms. Keywords or phrasesresponsible for writing the report, performing identifying major subjects in the report.the research, or credited with the content of thereport. If editor or compiler, this should followthe name(s). Block 15. Number of Paqes. Enter the total

number of pages.Block 7. Performing Organization Name(s) andAddress(es). Self-explanatory. Block 16. Price Code. Enter appropriate price

Block 8. Performing Organization Report code (NTIS only).Number. Enter the unique alphanumeric reportnumber(s) assigned by the organization Blocks 17 -.19. Security Classifications. Self-

m texplanatory. Enter U.S. Security Classification inBlock 9. Sponsoring/Monitoring Agency Name(s) accordance with U.S. Security Regulations (i.e.,and Address(es). Self-explanatory UNCLASSIFIED). if form contains classified

information, stamp classification on the top andBlock 10. Sponsorinq/Monitoring Aglency bottom of the page.Report Number. (If known)

Block 11. Suoolementary Notes. Enter Block 20. Limitation of Abstract. This block mustinformation not included elsewhere such as: be completed to assign a limitation to thePrepared in cooperation with...; Trans. of...; To be abstract. Enter either UL (unlimited) or SAR (same

published in.... When a report is revised, include as report). An entry in this block is necessary ifa statement whether the new report supersedes the abstract is to be limited. If blank, the abstractor supplements the older report. is assumed to be unlimited.

Standard Form 298 Back (Rev. 2-89)

Timothy B. Minorl, Jeromne A. Carterl, Matthew M. Chesleyl, Robert B. Knowles2, Per GusafUOU

t Desert Research Institute, 7010 Dandinl Blvd., Renm, NV 895122 U.S. Army Corp of Egneers,'lbpographtic Engineering Center, Alexandria, VA 22310

3Chtainers University of Technology, S-4129% Goteborg, Sweden

ABSTRAC'r

The Desert Research Institute (DRI) is developing strategies that integrate various data typesto characterize Iwaterresources for the identification of candidate well locations in centralGhana, West Africa. Study area were selected over which reot data of various calmwere assembled along with all available hydrogeologic information. Wersamples for chemnicalanalysis wer collected and mnuerous field observations were recorded during the survey of 200wet and dry boreholes using Global Positioning System (GPS) receivers. All data collected wrprocessed and integrated into a Gegahcifomto pe (GIS). Relationships between fea-tunes identified in remote sensing data. borehole rcrsai ~ t hydrogeologic characteristics amebeing examined to maximize water development efficicies. The GMS modeling results are beingtested with an ongoing drilling project. Preliminary results indicate a high correilation between dril-ling success and proximity to Landsat-derived lineaments.

1.0 INTRODUCrION

Lack of adequate potable and agricultural water supplies inhibits the progress of developing ouisandis the cause of considerable hardship to humans worldwide. A thorogh hydrogeologic understanding is often necessary for cast-effetve wate development. In many developing countries, remote sensin data often comprise the mostaccurate and comprehenIv spatial information available. Establishing relationships between remote seingm dataand hydrogeologic phenomena may maximie the efficiency of water developmnent projects. The Desert ResearchInstitute (DRI) is attemnptng to identify those strategies that effectively charateie groundwater resources in thecountry of Oham using a Geographic Information System (GIS). For selected target areas, the strategies developedby DRI will recommend a combination of analytical methods, data types and data feature which can be economicallyutilized to select well sate with the highest probability of success. While an emphasis will he placed on detetnminingwhether features mapped using remote sewing can be correlated with hydrogeologic phenomena, all methodsanddats,types available will be tasod in conjunction with the ongoing World Vision International (WVJ) Ghana Rmml WhateProject (GRWP). Ti~s pprdocumients steps undertaken in constructing a GIS model fir increased hytirogelogic

un Wstning of study areas within Ghana, and presents preliminary results of the relationship observed between lin-cament derived remote senning and the location of wet and dry boreholes.

2.0 BAOCGROUIND

Mhe Wst Affican country of Ghana experiences a lack of safe and adeuate water for its inhabitants; 25 to50 percent of the population do not have acess to safe water supplies (Lean at al., 1990). Potable water short~agea xitdue to waterbormu diesease and awe exacerbated as sources desiccate during the dry season. A major objective of tieOas-ian govauinent and noni-governmant organizations (NGOs) is to improve the availability of safe and reliablewater resources by drilling wells and installing hand pumps. However, throughout much of Ghana, including the studyareas for this project. wet well success rate ame below 60 percnt Wet weils for these hand pump installation programsare defined as having yields equal to or greater than 10 liters pv minute. Thbe major factor limiting success rate isthe low primary permneability of crystalline and consolidated sedimentary rocks. Secondary fracturing in these rocks,however, may a= to control the occurrence, movement and storage of groundwater. If fracture can be mapped andcorrelated with productive grwantwater zones, wail site selection may be improved.

9428125

Fracture traces and invas have traditionally been mapped through remote sensing methods and havebeen differentiated based an Untman, (19:58) defined fracure traces as natural lHuma features expressed contin-uously for less than one mi -le lineaments aregreater than one mile in length. While fracture traces and lines-ments have been shown to be barriers to groundwater flow (Taylor et a&L, 1992), most studies have correlated themwith zones of increasd fracture concentratino. which can act as cooimt for the transport and storage of goumndwawe(L~attman and Parirek. 1964; Parizek, 1976; Parijek et &1. 1990). Previous stuidies by Boeckh (1992), Gustafama(1993a) and Peter et al., (1988), have shown the utility of multiscale remote sensing strategies for increased hydrogeo-logic understanding in a variety of terrains and geological settngs.

The integration of remote sensing daata and GIS teclinology may provide a greater hydrogeologic understand-ing; by combining physioprapbic. geologic. hydrogeologic and geochemical data in a spatially referenced moodaL 013use in water resource evaluation has rcnowdy expanded with inizeasing emphasis in surface and subsurface applica-tionas (Maidment, 1991; Moore et al., 1991). Furthernore, the combination of remote sensing and GIS has shownproma-ise for groundwater development in other regions --, Africa (Guazaftson, 1993a).

,- S- )YAREA

A study are was developed to encompass ~v IL.nctgeologic tezmins within central Ghana (Figure 1). TheTease area lies entirely within the southern part of the VoIta~n Sedimentary Basin, while the Nkawkaw area to thesouthwest stradlca the contac bdetwe the Voltaan and the F.-rainbdan basemeiet rocks. These arma wereselecteddue tothe presencof an active driling progtramwith a larg percetag of unsuccessul wells. Tease is locsten ihathe Aft=a Plains, a geneally flat, low-lying region of the Voltas!&i Bvtin adjacent to Lake Volta. an 8500 kin 2 lakeformed by the damnming of the Volta River in 1966.

The Nbltaia Basin is a synclinal structure composed of relativwely flat-lying Paleo'~ic sandstoom, shahm,and conglonmerates that form outward-facing escarpments around the periphery of the bamii (I~esse. 1985). Thin see-tions of quartz areuito msauoaes W=k nea Tease and Fradaka in die Afma Plaim indicate izat clean ummedsand has beeni compFacted. Grain conacots am pnxdormna~ndy couave-convex mndicatinz a Iigb degree of couqmcdonby deformation of ductie grains. Consolidation and cementation have reduced the rock volume, destroying muc ofthe primary porosity and permneability. The stable mineral composition of the rocks permits little dev~lopmaen of seo-ondary permeability that might result from chemical weathering of ionegranuilar cementing materials. The geineallyimpervious nature of near-surface unweathered irockIs inimire direct infiltraton. reduare said weatherns whichame focused along joint and fractur systems. Groundwater flow in the basin has two primary components: a dominanttvertical flow recharging fissure at depth and a moan subdued aubhorizonal flow parallel with stream gradients. Theriver systems, therefore generally follow fractures where weathering and erosion have beeni enhanced and posentialgroundwater recharge is highest (Bauemamsn, 1988).

Them eow-weportion of the Nkawkaw -m consists ofnimoomuphic and granitic 1racks of the Birwswy.-tamn at the eastern edge of the Precambrxian Wee Africani shield. The Birimtian rocks have been folded, mtaiuwbaasdWand in places assimilated by granitoid bodies. Folding is intens with dips comimonly on the crder of 300 to 900 aloniga northeast-southwest aids. Faulting and jointing is most commonly parallel and perpendicular to the strike of the folds(Kesse, 1985). As with the Vataian, driling data in the Birintian indicates discontinuous aquifers with primary stoagelocated along fraceture zones, lithologic contacts and honrion of weathered rock masses (Asomaning, 1993). A lithe-feldapathic aniente from the Vatama portion of the Nkawkaw area contain plagioclase. potassium feldspai and nota-morphic and aedimmantary rock fragments that am normaelly unstable during the weathering processes. however theafragments =r unweathixd in th sample, indicating a lack of diageussis (Gin, 1994).

The rainfall pIt o n the Ahrm Plains consists of a wet season extending fiom. March to October followedby a four-month dry season. The annual avenage rainfall is approximately 140 centimeters. The natural vegetslinof the AftamPlaicnsists of Guina savanna. The mjajr cove found in the Guinea savanna zone consists al a Eft-contrlolldtissava cm onmmy ofiroad-leafd decidmuou i.densely distiuted ins cotinouscoveroipuum-nial bunch grasses arid fobs (USAID, 1962). The Nkawkaw area normally receives more rainfall than the Afram Plainsregion, avenaging 160 cn annually. The natura vegetation of the Nkawkaw area consists of thick deciduous foumwithin a semi-deciduous climatic region (Dush et sl.. 1993).

1-169

I4

GhanaRural Water

Prja

Project

I a IN

Figure 1. Arec Study Area

4.0 MM~ODS

loIumpedon aitasgie we being developed thatmheategrimmpod-Imud and remotely aoused data in a GISmodel. Qitcal steps in this development effot hincde: 1) the acquisition of .11 pertienem dats; 2) dafta pcessanxand 3) databue dweulopment. In evalusatig th study ara, powanda souros and types of dafta necessary to form ftUK S database were idemitfiud Dita won evaluated with respect to their poteanta uaefulzaimm cost, and availabilitywitin -rjc time - I -u Dam types deemed necessay hinde pulogc hydroolgic, -oorsb vept-don, and soil maps; aeial photography and staellift hosagry, gophysical lops and surveys; barahal location andrelated hydrologic information suc as wel yield war level, wrater quality, aquifer test data, (hydrulic condutivity.ransmnisivity, porosity, etc.); and reports addressing local and regional hydrogeologic investiptons

1-170

4.1 DATA ACQUISITION

Data collection effors were generally divided among three types of information: 1) reports. maps and relateddata readily available from the United States and abroad; 2) data that would require coUection within Ghana; and 3)remote semin data. General background information through water exploration program reports specific to our studyara were collected through standard literature search methods and various mapping agencies. Similar searches wereconducted fbr resources which discussed the use of aerial photography and other remote sensing methods for investi-gating geologic and hydrologic phenomena with an emphasis on fracture trace and lineament analysis. Many of theresources collected referenced other work and reports only available within Ghana.

4.1.1 Data Colle.ctd in Ghana

A trip was made early in the project to gather information identified through literature searches and othersources. to identify new information sources pertinent to the project, and to arrange for future data transfer with WVIand other agencies. Copies of topographic and geologic maps were obtained along with reports and any available welllogs. The release of data inside and outside Ghana often required appropriate agency permission.

A second field trip was necessary to accurately survey a significant number of boreholes for the GIS to deter-mine and test relationships between identified features and borehole characteristics. Three factors created large unw-tainties in borehole locations: 1) borehole locations an drill logs were identified only with respect to the commumitiesto be served; 2) some boreholes were located as far as 2 km from their associated village; and 3) many commuiinmpostdated the topographic maps. These factors, combined with cost consuaints and accuracy requirements. dicumedthe use of sub-20 m Global Positioning System (OPS) receivers to accurately survey borehole positions. Logisti'cconsideration suggested using two teams, each with a OPS receiver in stand-alone mode to survey the maximun umm-ber of well sites possible. To circumvent the single-mode, 100--m limitation imposed by selective availability (SA),arrangements were made through the U.S. Army Corps of Engineers to use two Magnavox MX 7100 receivers withthe PPS code enabling 10 to 20 m accuracies.

While borehole locations were the primary objective of the second in-country visit, numerous other data ob-servations were recorded during the well site surveys. These data included elevation, water level, soil type, surhcialgeology, topographic setting, hydrologic indicators and classification of surrounding vegetation and land use at 50 and500 m. Sketch maps, operating stats and water samples for gross chemistry and isotopic analysis were collected alongwith temperature, EC and pH. Whenever possible, site selection criteria were obtained from the Ghanaian hydrompio-gist responsible for siting wells. Well positions were also plotted on Liandsat imagery and aerial photographs usingscale positioning software available on Trimble's Scout GPS receiver. Because borehole elevations and water leveldata were necessary for constructing accurate potentiomnetric maps, additional steps were taken to improve elevationmeasmrements. As GPS-detemined altitudes are generally limited to 1.5 times the horizontal accuracy (15 to 30 m),elevations were determined by plotting their horizontal positions on 50-ft contour maps. Improving this further reda-tive well altitudes were recorded with baromeruic altimeters in the field using time and base station barometric caoec-tions.

4.1.2 Image Data Agqoiition

Initial data searches revealed a limited availability of usable Landsat Thematic Mapper (TM) and SPOT daftdue to cloud cover during the extended wet season and attemation by clouds, haw, dust, and smoke during the dryseason. Quality con:erms and time comntrainta dictated that two TM scenes be acquired during the dry season, ol•yafter viewing the actual data sets. Due to a lack of available archived SPOT panchromatic data. a Special AcquisitionRequest (SAR) was made in January 1993 for the collection of 11 panchromatic SPOT scenes. Of 16 SPOT stemacquired since the request, none have been acceptable for analysis over the study areas due to atmospheric conditionsand semn adjustments. Given the difficult atmospheric conditions and vegetation cover, attempts were made to ac-quire ERS- 1, Seasat and Shuttle Imaging Radar data. but were unsuccessful. Digitally scanned data sets of declmssifiedRussian MK--4 photography at 7.5 in resolution were identified in an archive in Moscow. but closer examination foundthe data to be corrupt. Data from the KVA-1000 camera system for the Aftam Pains were not available and couldnot be acquired within project time constraints. 1972-73 black and white infrared stereo photography of the study sites

1-171

aU 1:40.000 scale was collected only after obtaining permission from the Ghana Survey Departmnt.L in addition to die

1.40,000 fromm 1:10.000 mzlsrgemuazmt of selected areas wore also acquired.

4.2 DATA PROCESSING AND INTERPRETATION

Mhe ease and functionality of UPS have allowed the opportunity to mix dama from differen coordinate systemsand datums within a GIS. While coordinate an the thread that tie all darn together, many GISs have limited coordinatesystm and datum, transformation routnes. Thsis. it a citical that coordinates be transferred to the same coordinatesystem and daumiefr insertion into any GIS. UIPS-deteirmined coordinates taken with respect to the WGS84 datumnwere converted into the "Accr datum" using the "Geographic Calculato." to minimi~ze database inaccuracies. Coordi-nafte wene trandmed using the Molodeasky Method which shifted positions by up to 350 m. Bearing an distancemeasurements were used to determine UIPS offset locations with the National Geodetic Survey (NOS) INVPWD pro-

Level 0 Landsat TM data were precision-corrected using a cubic convolution resampling method and1:5,00 scale topographic maps. The Root Mean. Square (BIAS) error of fte georectification piroc ess Wasam~ pixelor 30 mn. 7No adjacent Landsat scenes (Path/Row 193M05 and 1941055) were digitally mosaikewd using Bit Mapperimage procssin software. Various color compouites were generated to determine the best band combbnaions foranalysis of liniemsme fracture, traces and vegetation anomalies. The Landsat data contained bad scan. lines which.,with the added attnuation contributed by smoke and dust in the atmosphere. prevented the effective use of TM bandsI and 2 in the construction of composites. Therefore, band combinations 453 and 743 in RGB were used to interpretthe imagery. Each color composite was edge-enhanced using a 3x3 edge sharpening filter. In addition wo geneatingedge-enhanced color composits. directonally filtered images were constructed using TM band 4 in a multi-suep pro-cedure similar to that described by Moone and 'Waltz (198). Thie resultant directionally enhanced componet imaerepresented fourr azimuth hearngs: East-West; North-South; Northwest-Southeast; Northeast-Soushwest. PrincipalComponent (PC) trasfomations were performed on bends 543. 743 and 745 to produce thre-band PC compositesin ROB. In addition. PCI ficm bands 743 and 543 were used as greyscale imags.

U geunera interpretation approach has been to digitize linear features on screen using the annotation toolin ER Mapper. Interpolation and extrapolation of features have been avoided to produce a more -objective" lineamexismap, for mibsequsis tectoni interpretation. Lineament interpretation can be carried out in different, ways dependingupon t image type availale ground iuthing requiredl. and aim of the interpretation Ifth aim of the interpretationis to reproduce the, general fluctr trend of the area, the positional accuracy of fth lineaments is not as importantas the orientation and correct frequency in each interval. If. on the other hand, the position of the inferred fluctue zoneis to be used in comparisons with other spatial dafta such as boreholes, the positional accuracy is more imporetan Shu-man (1991), determined that ineament frequency decreases as lineanent size increases, while a plot of average linea-inent length agains imagery scale shows a positive polynomial relationship. The higher spectral resolution of TMverues SPOT dam is more importaut for lineamen analysis, in aras of dwe vetation. Therefre, both spatial andspectral image resolution ha a greas impact on image interpretation results.

lIn this studA a systeaticmappig effort was conducted to perfrm the lineament and fracturetroce analysisbased on techniques adapted from Whteeler and Wis (1983).7Thi effort had several objectives: 1) consistently identifylniniment and Frac Ptureace features; 2) classify the liniaments based on a rating system related to vector length,orienttion, and feature promninence; and 3) eliminate non-structural and "false lineaments. The interpretations were

peformed;wih no prior knowlede of wet or dry well locations.

7UThe Tenoae Landinat scen was mubsectioned into overlapping 1: 100,000 scale images on she worksttationscemen. The study -m were covered using 50 percent overlapping windows to avoid overlooking features at window

A_

1-172

edges (Oustafison. 1993•b and 1994). The strategy employed evaluated three sees of image dam- color composites,directionally filtered imae and principal component composites. Three ope:ors independently mapped all linu-ments. Individual inetn results were digitally combined using different colored vectors to evaluate nappingresults, compare class types, and eliminate "false" lineaments. 11m Nkawkaw area interpretation was carried out ata larger scale than in the Tern aream due to the high demity of lineamens adjacent to the Voltaian escarpzen. Theinterpretation scale an scrm was appeoximately 1:45,000. enabling moe accurate positioning of linear features.

4,2.A Photo IntMMafm

The aerial photography was interpred stereoscopically on 1:40.000 scale prints with 2x and lIx maSnifica-don to enable better positioning of small or narrow linear features. The aerial photos were analyzed by several inser-preter, focusing on photo linears within a 2-kin radius of existing wells and villages. This constraint was implemeaedbecause 2 km was the maximum distance a well would be placed from a village by WVI. The interpretaton werperforned using acetate overlays mad the identified photo linesm were classified into three categories accrding totheir prominence. Upon completion of the independent analyses, the mapping results were compared. "false" fracturetraces were eliminated, and features checked for redundancy.

4.3 DATABASE DEVELOPMENT

The GIS database continues to be developed using Arc/INlf software. To dae, base map, well location to-pogranpic, field observato and satellite renote sensing dart have been fully automated and integrated into ft CIS(Figure 2). GIS display and geor•lational capabilities allow for efficient analysis of both raster and vector dao types.All integrated data will be managed within a coherent spatial reference frame for development of strategies to identifydata types and analysis techniques most useful for siting well locations.

4.3. BIUMWm

The project study =a were subdivided for reference into a series of 15-minute quadrangles. Comr tim ofthese quadrangles serve as registration control points for all base map automation. The GIS uses a Umn, asal Tram-verse Mercator (UTM) coordinate system and is based on the 1:50,000 scale topographic map series produced by theGhana Survey Department

The 1:50,000 map series provides essential base map data, incuding surface hydrography, transportation fee-tures. and village locations. Surface hydrographic features are valuable for understanding groundwater flow, whilevillage locations and transportation features are important from a well-siting perspective. The source maps were avail-able oan paper media only, thus map registration error were corrected to account for map shrinkage. Data layers inthe GIS were appropriately coded with as many descriptive attributes as could be derived from the base maps. Drainagefeatmr wre coded a perennial or ephmneral tranportation features were coded to reflect level of use and semealreliability; and commities wene coded to capture variaions in spelling.

4.3.2 We Locations

Understanding the spatial distribution of existing wells is clearly an important first step in developing a wll-siting strategy; a pteliminar saeening of productive versus 'dry' wells can provide a valuable initial assessment ofgroundwater conditions. Furhermor accurate locational (and elevation) data for these wells are critical for maximiz-ing use of borehole information collected during drilling. Prio to the GPS survey, a preliminary well location damnlayer was developed, using the spatial tools of the GIS to place the wells within the village to which they were rfer-enced. The preliminary well location data layer gave project hydrologista a provisional set of data from which to devel-op early working hypotheses. Upon completion of the OPS survey, all borehole positions in the GIS were updated toreflect their new positions.

Hydrogeologic data in the form of drafted borehole logs are generated as pan of WVIs ongoing water welldrilling program. Borehole logs include hydrogeologic and lithologic descriptions of rock units encountered during

1-173

Ghana Rural Water Project

Water Table Elevationsand Well Locatons

Afram PlainsWOWa~ 1d" sa"= 0 ". "wbu Usd ewdV-

__ d keewe~mik pb n MNW j-i

&M12 us" kbwq82N1t

7 11eM 1M14aa

mm WA 7lu 40.iamaiam M- ~ iL... .in.J76 17.?IUAI 74.66.166M 4 08d I

_____&MA mm 77mu 2ua.ian~n

701 76.212.84 73OA46 d AdkM 2

1-740 "76I*7SM

drilling, well construction information describing casing and scre en intervals, and development history of the well.in addition. the lop contained water chemistry and well yield information, vital to the chiaracterization of groundwaterflow in the study areas. Data entry templates were created for efficient entry of large quantities of well log infornatimnIn all, 14 separate well log data filies were created, each containing from eight to 40 descriptive attibutes. To expediftthis process for' fummr borebolms WV! has bee~n proivided with logging software, so that subsequent well lop can becreated and transferred digitally. In the 015, all well log files ame linked relationully to the well location attrbute table.so that any combination of well consuuocton, Lithologic. geochemical, or hydrogeologic characteristics. from anynumber of wells, can be viewed in spatial contex with one another.

Topographic data are a vital in determining water table elevations and, subsequently, developing pounio-metric maps. As no digital topographic data of sufficient resolution were available. paper copies obtained in Ghanawere scanned to produce a Digital Elevationz Model (DEM) and contour plot files. These data will allow three-dimain-sional viewing, watershed calcuilations, and tectoni interpretation, and permit refinement of well elevations and po-tenniometric maps.

4-3.113~

The lineament and fracture traces Identified in the digital analysis of L.andsat TM data were automasd andincorporated into the OIS. Thie Landsat-dwived lineaments and fracture traces were imported as vectors fim ERMlapper. As a data layer in the 015, these features were classified based on length and orientation and msessed for duairproximity to extisting wells. The fracture tiaces derived from stereoscopic analysis of 1:40,000 scale black-and-whiftaerial photography will be digitized after being transferred to rectified 1:40,000 scale photomosaics. Once inseguutedinto the. GIS. spatial analyses will be conducted, messsing fracture trace properties relative to existinig wells and theirhydrogeologic characeristics.

5.0 RESULTS AND DISCUSSION

5.1 DATA COLLECMiON

Initial literature searchbes and an early tr- ip to Ghana provided the means to collect most exist-ing data over the study area. Project emphasis an fracure trace analysis. coupled with previous studies having diownthe relatively narrow nature of fracture zones (Parizek et al.. 1990), required the collection of momr accurate well posi-tions. The resulting OPS fheld effoct surveyed over 200 boreholes in 82 different communities. Numerous other fieldobservations were simultaneously collected in addition to water samples from 50 sites. Gross chemistry and isotopicanalysis is underway, while all other data have been entered into the 015.

5.2 REMOTE SENSING DATA

Image analysis tesults of the Landsat data indicated that the most useful band combination for discriminatinglineatments in the Team area was the PC composite from TM bands 543. While the 453 and 473 color composites pro.-vided some utility, the PC composite better disriminsnted soil brightness and vegetation 0ifferusces, with reducednoise levels. Thie directionally filtered images contained too many image artifacts to be of use in this study. This wasdue in paz to the bending effects in the Landiat imagery as well as, the relatively fiat terrain and uniform landcveof the study ares. Most of the lineamnenta are related to structurally contrldW drainages in valley bottoms. Amalysisof die predominant lineamrent orientation indicated a maxima between 270-275*. with a less dominant orientatonfound between 330-335* (Frigure 3). A total of 900 lineamemt were identified in the Tease area. The linearime nxt luhvaried between 84 and 5938 mn with a mean length of 1377 mn. The orientations are consistent with regional strucaciralprocesses and fractures observed in the field. The dry season acquisition dates of the Landsat imagery permitted theeffective use of riparia and tap root vegetation species as indicators of fracture. systems. In some cases. riparian ape-des in drainage basinis helped accentuate lIneainents.

The most useful ILAndsat bend combinations for lineament interpretations in the Nkawkaw ares were die PCcomposite from TM bands 543 and PCI fr-om bands 743 in greyscale. The interpretation resulted in 1037 lineamients

1-175

0

46. 24. 12. 36. r. 8

Flguz 3. Fhxemque and Oriczatedaz ofactuare Trnces and lines-m inthaeTease Area.

0

270 9036. 24. 12. S s 0 2

Figuue 4. Pa Iunc and Orientadan aiFh~e a ma.'fu and flineamentm in the Nkawka Area.

1-176

identified in die Nkawkaw area. with the majority of lineamuenta found in a 10-km-wide section along the northweu-3outbe5It-V00diflg escrment. The lineamnent lengths vanied frwn around 100 to to 10 km with an average around1200 m. Mae predominan lineament orientations indicate maximas at 3250,3150 and 3050. with less dominant oinra-.tioxa found between 290-300 (Figure 4).

AlthOug photeuupik contrast is poor (due primarily to attmospheric haze). preliminary results of air photoinUMIpM0aao indicate that, the larger scale anod stereoscopic characteristic allow the discrimination of fracture tracesnot diacernible on the satellite imagery. Subtle fracture traces linked to chainges in tree heights and soil tonal align-mens were identified. In several canes. fracture trace intrsections were found near villages. providing new potentialtargets for future drilling progrms. Once integrated into the GIS model, spatial analysis operations can be conductedto determine relatouships between photo-erived fracture traces and other features, in the database.

5.3 GIS MODELING RESULTS

While (US umodling efforts are only preliminary, initial proximity analysis of fracture traces and lineanuentsderived fran Landast 'IM imagery and borehole locations indicate pronrisig results. Almost 95 percent of all wellssurveyed in the Teaae ame are located within 2kim of identiid frature traces and lineaments. This suggests the avail-ability of potential targets within the present project drilling constraints. Of greater significance. 100 percent of allwells within 100 m of identified fracture traces and lineament am wet, with a decrease; in wet well percentage as thedistance fran a fracure inr~eame (Figure 5). Furthermore, lesa than seven percent of wells within 250 m were diry.The initial results aubm that 30-mn Landat TM damta y be an effebctv tool for siting wells. These fmtlings alsosupport previous observations (Bamrmann, 1988) that fracture zones in the area act as primary pathways for storageand nmoement of groundw- -z. Further analysis of fracture orientation and length, proximity to well lIncation, fracwreweil yield relationships. ama lithology will be conducted in hopes of determining other factors critical for improvingdrilling success rates and well yield.

Cunrulative Number of Wells,

100- 5 if' 18 24 29 32 41 66 88 122 139 144

60--

20--

0-- - - -

<25 <50 <100 <150 <200 <250 <500 <750 <1000 <1500 <2000 <3000

Distance in Meters

Figure 5. Distribution of Wet and Dry Wells vs. Distance from Nearest Lineament or Fracture Trace.

1-177

6.0 SUMMARY AND CONCLUSIONS

Prefiminary results indicate that satellite imagery, when integrated with other hydrogeologic data, may bea valuable tool in OIS-baed determination of potential wel targets. The synoptic view provided by satellite imageryallowed a better understading of the regional hydrogeologic settings in this project. The full potential of remotelysensed data, however. has not been realized due to the limited availability of other types of imagery over the studyareas. The lack of SPOT data and Russian satellite-based imagery, as well as the absence of radar coverage, has inhib-ited the application of additional remote sensing techniques for the evaluation of hydrogeologic features in the studyarea. Efforts to acquire image data, specifically SPOT pandhromatic scenes, shall continue through the end of the 1994dry seaon in hopes of improving the hydrogeologic GIS model. The stereoscopic, high resolution aerial photographs,though poor in quality, allowed the identification of subtle fracture traces not discemable on the Landsat data. If similarrelationships are found between photo-deived liiears and borehole success, it may provide new drilling targets closerto the communities in need.

GIS database development and data integration in a developing nation present a variety of unique challenges.Uncertainty regarding the availability and format of numerous data resulted in many delays in database developmentactivities. Remote site data collection efforts require thorough logistical considerations, contingency plans, and spareequipment, especially in areas environmentally hostile to modem dam collection instruments. It is critical that the GISoperator be completely aware of the geo-referencing systems, datums, and elevation surfaces associated with the vari-ous data types collected, so as not to mix incomptible coordinates.

Although the results of this paper and the project are preliminary, several significant relationships betweenfeatures derived from the Landsat TIM data and hydrogeologic characteristics of the Tease area were observed. Addi-tional proximity analysia. integration and analysis of fractur traces derived from the aerial photography, and correla-tion of well yields to lineament length, orientation and geology, will be conducted as the project progresses, the ulti-mate goal being the selection of well sites with the highest probability of success in different geologic settings.

REFERENCES

Asomaina, G., 1993, Proceedings of the 19th WEDC (Water Engineering and Development Centre) Confer-ence, Edited, J. Pickford, P. Garter, A. Coad, M. Ince, R. Shaw, B. Skinner and M. Smith Acra Ghana,September 6-10, p.44-47.

Bannerman, R.. 1988, Progress in Groundwater Resources Assessment in the Voltaian Sedimentary Basin ofGhana. Unpublished Report, p. 2.

Boeckh, E, 1992, An Exploration Strategy for Higher-Yield Boreholes in the West African Crystalline Baae-meat From Wright, . P. and W. 0. Burgess (editom), Hydogeology of Crystalline Basement Aquifersin Africa. Geological Society Special Publication No. 66, pp 87-100.

Duah, A. A., S. A. Larmie and J. 0. Aduakye, 1993. Evaluation of Groundwater Use in Rural Ashanti. Pro-ceedings of the 19th WEDC 'Water Engineering and Development Centre) Conference, Edited, J. Pick-ford, P. Gare-, A. Coed, M. oce, R. Shaw, B. Skinner and M. Smith Accra, Ghana, September 6-10,p.44-47.

Gin, O.d., personal communication, 1994.

Guatfson, P., 1993a, High Resolution Satellite Data and 0IS as a Tool for Assessment of the GroundwaterPotential of a Semi-Arid Area. Proceedings of the Ninth Thematic Conference on Geologic RemoteSensing, Vol 1, pp. 609-619.

Gustafson, P., 1993b, Satellite Data and GIS as a Tool in Groundwater Exploration in a Semi-arid Area. The-sis for Licentiate Degree in Groundwater Engineering. Department of Geology, Chalmern Universityof Technology. Publ. A 74.

A..

1-178

Guatafuson, P.. 1994. SPOT Satellite Data For Exploration of Fractred Aquifers in a Semi-arid Area in

Southeastern Botswana. Applied Hydrogeology 2/94 (in press)

Kesse. 0.0., 1985. The Mineral and Rock Resources of Ghana. A.A. Balkema, Rotterdam, 610 p.

ILanman, L H., 1958, Technique Of Mapping Geologic Fracture Trace and Lineaments on AerialPhotographs. Photogrammetric Engineering, Vol. 19, No. 4, pp. 568-576.

Lattman, L HL and R. R.. Parizek. 1964. Relationship Between Fracture Traces and the Occurrence ofGroundwater in Carbonate Rocks. Journal of Hydrology. Vol. 2. pp. 73-91.

Lean, G., D. Hlnrichsen and A. Markham. 1990. Atla of the Environment. Prentice Hall Press. New York.pp. 30-31.

Maidment, D. R,. 199 1. GIS and Hydrologic Modeling. Proceedings from the First International ConferenaceWorkshop on Integrating Geographic Information Systems and Environtmental Modeling. Boulder, CO.

Moore, 0. K. and F. A. Waltz. 1983, Objective Procedures for Lineament Enhancement and Extraction.Photoprammetric Engineering and Remote Sensing, Vol. 49, No. 5 pp. 641-647.

Moore L D,. A. IC. Turner. J. P. Wilson. S. K. Jenson and L E. Band. 199 1, CIS and Land Surface-SubsurfaceProeI Modeling. Proceedings from the First Internaonal Coafmrnce/ Workshop on Integrating Geo-graphic Information System and Environmental Modeling. Boulder. CO.

Parizek, R. R.. 1976, Lineaments and Groundwater, in Interdisciplinary Applications and Interpretations ofEREP Data within the Susquehanna River Basin. ORSER-SSL Technical Report 2-76, SKYLABEREP Inv. 475, May 1, 1972-November 30,.1975, pp. 4.59-4.86

Parizek. R. R, Lavin, P.M., Greenfleld. R. J., Weiss, R. A., Schuamn, C. A. and Moran. M. L. 1990, Natureand Hydrologic Significance of Fracture Tracms Lineaments, and Fracture Zones Related to Groundwa-ter Monitoring. U.S. Environmental Protection Agency, Office of Research and Development. Environ-mental Monitoring Systems Laboratory. Technical Document CR8 13660-01-0, 165p.

Peter, K. D., K. E. Kolm, J.5S. Downey and T. C. NITchols, Jr., lS,88, Lineaments: Significance, Criteria forDetermination, and Varied Effects on Groundwater Systems- A Case History in the Use of RemoteSensing. in Johnson. A. 1. and C. B. Petterson (editors), Geotechnical Applications of Remote Sensingand Remote Data Trantsmission. ASTM Special Publication 967, pp. 46-88.

Shum.. C. A., 1991, Joints, Fracture Traces, and Lineametux A Multiple-Scale Analysis of Fracture Fea-tures Across the Appalahinm Plateau in New York. Proceedings of the Eighth Thematic Conference onGeologic Remote Sensing, Vol. 2, pp. 1229-124 1.

Taylor. K., M. Widmer and M. Chesiey, 1992, Use of Transient Electromnagnetics to Define Local Hydrogeol-ogy in an Arid Alluvial Environment. Society of Exploration Geophysics, Vol. 47. No. 2, pp.343-352.

Whaser, R.L and DXV Wise, 1983, Linesmanship and the Practice of L inear Cow-Art. Discussion and Reply.Geological Society of America Bulletin, Vol. 94, pp. 1377-1379.

U.S. Agency for International Development, 1992. Report on Soil and Agricultural Survey of Sene -ObosumtRiver Basins, East Brong-Abafo and Ashanti Regions. Ghana, Team Leader George K. Smith, pp.99-107.

1-179