BO OLOFSSON N GU - BU L L 4 3 9 , 2 002 - PAG E 15
Estimating groundwater resources in hardrock area s - a water balance approach
BOOLOFSSON
Olofsson, B. 2001: Estimating groundwater resources in hardrock areas - a water balance approach. Norgesqeoio giskeundersekelseBulletin 439,15-20.
Among decision makers in counties and municipalities, the re is a great need for simple but hydrogeologically cor rect estimations of the groundwater resources.Today's methods are ofte n based on extensive field investigations and compli cated mathe matical flow modell ing or too simpli fied annually based water balances. In this paper, a groundw ater balance meth od is presented that is especially adapted to small and com plex aquife rs, such as in hardrock, in wh ich the storage capacity of the aquife rs gives the extraction limit s.The balance is focused on the development over ti me of the storage, to which groundwater recharge and withdrawal from well users are added and drawn . A compu terised system has been develop ed as an expert system for users, who are not professional hydrogeo logists.The balance can be used for calculating th e maximum acceptable numb er of consumers in an area or for the set of a maximum acceptab le sanitary standard, based on available ground water resources.The balance method hasso far been used asa planning tool within several areasaround Stockholm, Sweden.The results are rea sonable,compared to experiences of t he real groundwater situations in the areas,although a tru e validat ion has not yet been carried out.
Bo Oloisson, Dept.of Land and Water ResourcesEngineering.Roya/ln stituteof Technology.SE-lOO44Stockholm, Sweden, e-mail:[email protected]
Introduction t ive methods previously used for t he estimation of ground Many densely populated areas in th e Scandinav ian coun wate r qu ant it ies and vuln erability.The qu alitativ e meth ods trie s are located in areas w ith limited groundwater sources. do not give absolute values of discharge possibil ities but Some of the biggest citi es, such as Stockho lm, Goth enburg, instead give a relative measure.Such met hod s include local Helsinki and Bergen are located in hardrock terrain with a expe riences and some variable methods, th e DRASTIC large surface area of bare outcrop and in t he lower parts method (AIIer et al. 1987) and th e LeGrand method (LeGrand soils, mainly consisting of till, clay and in places sand and 1983), w hich are vulnerability met hod s developed in th e gravel. The outer areas, which commonly have a mi xture of USA for comparison between different areas. A specifi c vari summer housing and permanent hou sing, generally rely on able meth od, adapte d to Swedi sh cond itions, th e Risk a small-scale water supply based on private dri lled we lls. Variable method (t he RV-met hod), has been deve lop ed and There is also a hard exploitati on pressure on those areas, tested for t he occurrence of salt groundwater in Sweden w hich are usually located in attractive surroundings. The using GIS in Norrtalje municipality, Stockholm coun ty areas are slowly converting from pu re summer hou sing to (Lindberg & Olofsson 1997). Although the qualita tive vari perma nent housing. In addition, the sanitary standards in able methods can sometimes give valuable information, th e the areas are ste adily improving, which in turn increases the results are not always easily understandable, which can lead specific demand for fresh water and increases th e amount of to inconsistent decisions. Quant itative methods, on t he sewage water produced.A too high groundwater ext raction other hand, give calculated values for th e possib ilities of sometimes leads to saltwate r intrusion or other groundwa extracti on of groundwater. The met hod s are commonly ter quality problems, such as increased nit rate content and based on direct measurements of physical prope rt ies, often bacteria. The municipal and county administ rations often hydraulic measurements in we lls and mo re or less complex have to take decisions wheth er an increased exploitation and sophi sticated calculations using water balances or can be accepted, based on the available water sources and math ematical modelling .The latter has commonly been car t he possibilities to handl e the sewage wat er.There are acute ried out during big con struction projects in some areas, e g., quality prob lems in some areas in municipal ities along the wit hin th e nuclear waste repository program me in Sweden Swedish eastern coastline, such as in Varmdo, Nacka, (for examp le, Follin 1995).However, math ematical modellin g Osteraker and Norrta lje municipal iti es in th e county of requires an exte nsive amo unt of data, especially in hydroge Stockholm. ologically heterog eneou s areas, and can usually only be car ried out by experts. Therefore, th e costs of such investiga Basis for municipal decisions t ion s are quite high. The ground water balances t hat have There isa la rge number of different qualitati ve and quant ita- been carried out, on t he oth er hand,have gen erally been too NGU-BU LL 439, 2002 - PAGE 16 BOOLOFSSON
simple, and have usually not taken th e arti ficial and time under fully saturated conditions. The flow from soil to rock is dependent discharge as well as the heterogeneous rock an even more uncertain variable.It can only occur at specific characteristics into consideration . hydrogeologically con ditions where a conductor in the rock Groundwat er balances have been used asa decision tool is hyd raulically connected to a conductive reservoir in the in mun icipal planning in Swed en since th e beginning of th e soil (Olofsson 1994). 1980s (Sund & Bergman 1980).The balances were orig inally In vegetation-covered areas of southern and centra l based on a simplified recharge-d ischarge approach in whi ch Sweden and Finland, groundwater recharge is assumed as t he annu al groundwater recharge (strict ly formed from pre small or negligible during the summer period, because th e cipitation) in an area was compared to t he annua l wi t h potent ial evapotranspiration greatly exceeds the precip ita drawal of ground water by house hold consumption. tion (Fig. 1).Therefo re, groundwater recharqe in those areas usually occurs during the spring and autumn. P'C 'A=Q 'n ±r (1) During winter, a considerable part of the precipitation is P=precipitation C=coefficient of infiltration A=area sometimes stored in snow and added to the infil tration dur Q=specific withdrawal n=number of persons r=rest ter m ing th e snow -melting period. Although th e potential The part played by precipitation, w hich cou ld infiltrate groundwater recharge in the county of Stockholm is as high into th e ground, the coefficient ofinfiltration (C), is assumed as 270 mm per year, t he amount of groundwater that can be to vary between different geological terrains (usually a frac stored in t he hardrock is mu ch lower in reality. In practice, tion of 0.OS-0.2). If t he artificial discharge from the area (by maximum discharge from t he hardrock is lim ited by the pumping) was equal or higher t han the recharge,water sup kinematic porosity of the geological material, since a hetero ply problems could not be avoid ed. The method was fre geneous fractu red rock and a heterogeneous till comprise quent ly used during th e 1980s (Sund & Bergman 1983, drainable as well as non-drainable pores (Fig. 2). Eriksson & Tilly 1984, among others).
The limitations of recharge and extraction The recharge and discharge of groundwater are highl y vary ing functions of the spatiall y and time-dependent distribu t ion of precipita tion, geo logical conditions, topography, typ e of land use and vegetation, housing and sanitary stan dards. Methods for recharge estimations in the most com mon soils in Swed en (in till) are given by Johansson (1987). Very little research, however, has been carried out regarding the recharge to rock. Bergman (1972) measured the loss of
water during art ificial sprink ling of water on bare hardrock Fig. 2. Fractures in hardrock:A= drainable,flowing.B= drainable,no flow outcrop s, which only can give an approximate measure C= not drainable, no flow (Olofsson et al. 2000).
160 140 120 100 E E 80 60 40 20 o
• Precipitation (mm) fZj PoLevapotranspiration (mm) 0 PoLgrw.recharge (mm)
Fig. 1. Monthly available amount of water for infiltration in Stockholm county (based on data from Swedish Meteorological Hydrologica l Institute, SMHI). BOOLOFSSON NGU-BU LL 439, 2 0 0 2 - PAGE 17
Fig.3. A new ground water balance approach; important factors. Q1-36S(withdrawal , permanent) _ Q1-36S(withdrawal , summer) tt -
c=J Rock Ilffijjfijgm...... = rmI _ Clay c:=J Sand
iura' discharge ....
The kinematic porosity in hardrock is very small, usually The balance can be mathematically formulated as: 365 m m much less than 0.05% (Carlsson & Olsson 1981, Olofsson ~S g, i (2) I n:N, ' l{Jg • Ag • Tg = Qi - n,± 2: 1991); therefore, in practice, only a minor part of the poten 1= 1 9=1 9=1 tial recharge can be stored. In soil, however, th e kinematic N,= Pi- ET; + R,(if Ni
ings, a typical stratigraphy is built up representing a stratig Withdrawal of groundwater raphy which is common in the surrounding areas (a hydro Many exploitational areas,especially summe r housing areas, geological type setting). In eastern central Sweden, for show a sign ificant seasonal variation of groundwater extrac example , clay usually covers a coarser soil, such as sandy tion. Whereas our knowledge of water consumption in per silty till , which in turn is resting on the fractured hard crys manent housing areaswith a high sanitary standa rd is quite tall ine rock. Each stratigraphical unit is given typical values good, there is a signi ficant lack of information on consump of kinematic porosity taken from the literature,a typical thick tion in summ er cottages with varying sanitary standards. A ness, based on drillholes or general experiences and a typical distinct ion is made between the wat er consumption in sum depth to the groundwater level. However, all units beneath mer cottages and in perman ent houses, based on literature the uppermost unit are usually assumed as fully saturated . values and local experience (eq. 5). The homogeneity factor is a rough measure of the percent age of stored groundwater, which in reality can be re-cap A computerised system tured. Some of the groundwater will usually be lost due to The computer program consists of an 'expert system' in the natural groundwater discharge. Decreasing homogene which hydrometeorolog ical data and typical hydrogeologi ity means that it is much more uncertain whether a specific cal information for the particular region has been set by an infiltrated amount of water really can be re-used for water 'expert; based on existin g databases and general experi supply.The factor depends on the geological material and is ences.The operator of th e system, usually within the mun ici usually given values between 50 and 100%, the latter for a pal admini stration, only has to add the distribution of soils homogeneous sand and gravel. For hardrock the value is and outcrop, and informati on regarding th e population and also affected by the fracture pattern, and several fracture the sanitary standard of th e actu al housin g areas. sets in various orientations increase the homogeneity factor. The program calculates th e develop ment of the storage When the storage isfilled up, no further recharge occurs and dur ing the year. A specific balance modelling can also be when it is empty no discharge is possible. carried out in order to calculate the maximum number of houses in a specific area,or the acceptable degree of perma Recharge to the storage nent housing and sanitary standard with respect to the The recharge to the storage is given by the precipitation actu al groundwater resources. when it is exceeding the potential evapotranspiration. Therefore, in principal th ere is no recharge in areas with veg Examples from eastern Sweden etation during the summer season.However,it is not impos The groundwater balance mod el has so far been used within sible that some recharge may occur during the summer in several areas in three municipalities in Stockholm county, hardrock outcrop areas, if the fractures are locally saturated Sweden (Fig.4).Some basic informatio n and results from the during summer rains. Recharge occurs as long as the tot al calculations are shown in Table 1,com pared to the real expe storage is not filled up. When a state of full saturation is riences of th e groundwater situation in the areas. reach, recharge may cont inue but on ly with th e speed of the natura l and induced groundwater flow. In large parts of th e Ekskogen Scandinavian countries, surface runoff is usually of mino r 2 importance except in hardrock outcrop areas, especially on Alga steep slopes and on natural discharge areas, e.g. fens. 3 Ska lsma ra However, after intensive rainfall and snow melting, the sur face runoff may be considerable. Rodhe (1987) has shown 4 that most of the water found in minor streams in Sweden emanates from groundwater. Hence, the infiltration factor given in the water balance (eq. 2), usually varies between 60% of th e potential infiltration for a terrain with some parts consisting of clay,and 100% for areas consisting of sand and gravel. Different infiltration factors are usually set for various soils depending on the local conditions. The values of precipitation and potential evapotranspi ration are obtained from hydrometeorological data bases as mean monthly values, but also values representing special N dry cond itions, e.g.,20 or 50 years return time,are used. Furthermore, there is a possibility to include artificial Fig.4. The location of some examples recharge into the balance, such as collection of precipitation of groundwater and percolation systems as well as re-infiltration of treated balances in the _ =_:=::::=::=J*" sewage water. Stockho lm area. Okm 20km 40km 60km 80km BOOLOFSSON NGU-BULL 439, 2002 - PAG E 19
Table 1. Results fro m gro undwater balance calculatio ns wi thin some areas in the vicin ity of Stockh olm .
Variable Algo Ekskogen Saltaro Skiilsm ara
Reference Olofsson (2000) Jacks et al (2000) VaiVa-projekt(2000) (unpubl. 1997) Municipality Nacka Vallentu na Varmd o Varrnd o Area (ha) total (netto) 200 (160) ha 160 (160) ha 510 (400) 350 (200) No of houses 540 269 763 387 No of Person s per hou se 2.S 2.3 2.3 2.5 Degr ee of perm anent housing (%) 35% 16% 28% 40% Specific w it hdrawal (m') Summ er 100 L/ p,d Summer 100 U p,d Summer 150 U p,d Summer 150 Up,d Perm . 180 L/p,d Perm. 150 L/ p,d Perm. 200 L/p,d Perm. 200 L/ p,d Geology Rock 80% Rock 70% Rock 57% Rock 61% Sand y silty till 10% Sandy silty till 10% Gravelly sandy till 9% Sand 1% Clay 10% Clay 20% Clay 32% Clay 33% Org.soils 2% Org.soils 5% Top og raphy 0-50 rn.a.s.l 30-70 rn.a.s.l, 0-30 m.a.s.1. 0-30 m.a.s.1. fairl y rough moderate moderate mod erate Mi nimum fill of storage Normal year (dry year) 39% (20%) 78%(73%) 78% (73%) 69% (60%) Minimum fill of sto rage at 100% 0% (0%) perm anent hous ing mod el stops at 81% Normal year (dry year) (51%) perm.hou ses 51% (31%) 60%(44%) 46% (25%) Balance result (ann.) (Recharge/withdrawal) 1.17 1.51 1.52 1.31 Groun dwater qu alit y sit uation Mod erate qualit y No qualit y Small and local Mo derate quality today problem s today prob lem s toda y qu alit y problems problems today (Cl",microbe s) (Cl", N0 3", PO.3") «n (Cl", microbes)
Although a true validation of the balance estimations One area, Alga, is already tod ay very close to overex has not,and probably cannot be carried out,the results from ploitation.The balance modelling at Alga indic ates th at th e the balance calculations clearly correlate with the actual groundwater resources can also sustain specific dry years groundwater situation within the four presented areas.If the (with a recovery time of 20 years) with th e current housing storage situation is not taken into consideration, an overesti situation (65% summer houses). However, th e proximi ty to mation of the groundwater resources is the likely outcome . Stockholm favou rs an increasing tendency towa rds perma Such calculations using the simp lified balance equation (eq. nent housing, which will definitely lead to stress on the 1) at Alga have given much higher recharge values and wat er resources. Using a withdrawal value of 180 lit res per extraction possibilities (VIAK 1989).The HBV-model (Sanner person per day,the groundwater situation may collapse du r 1995) at Ekskogen gave a theoretical grou ndwater recharge ing specific dry years, and already when th e permanent of 210 mm compared with the calculated available gro und housing reaches 50% (Fig.5).The situation can th en only be water recharge of on ly 41 mm using the above balance solved by artificially increasing the storage, decreasing th e approach (eq.2-5). Still, the extraction ofwater from th e area water consumption per person, desalination of seawater or was less than 15% of the potential recharge (Jacks et al. connecting to the main land district water supply. The two 2000). last-mentioned methods are very expensive since th ey
Actual month Iy storage situation (0.10) ~ Normal values ~ Extreme values Fig.5. Results fro m water balance calculations at Algo. Development over tim e of the groun dwater storage for norm al years and specific dr y years. 'jil The sto rage is emp ty when th e perm anent J a nu a ~ Mareh May September November housing reaches 81% February April June August October December (norm al years) and 51% (dry years). NGU -BULL 439, 2002 - PAGE 2 0 BOOLOFSSON
require construction of a distribution network, mainly References through rock blasting, instead of relying on today 's private AIIer,L., Bennet,T., Lehr, J.H. & Petty, RJ. 1987: DRASTIC- a standardised wells. system for evaluation of grou ndwater pollution potential using Re-infiltration of sewage water (grey water ) was not con hydrog eolo gical settings. Report. U.S.Environmental Protection sidered here since withdrawal of groundwater was made in Agency. drilled wells and infiltration is usually carried out in the most Bergman, G. 1972: Bestarnninq av infilt rations-koeffi cienter far bergytor och perko lationsbanor i jo rdlage r (Estimation of infiltration coeffi superficial and unsaturated parts of the soil. Microbe analy cients in the bedrock and percolation pathways in soil).University of ses on groundwater from Ekskogen and Alga indicate that Stockholm, Dept.oi Quaternary Geology, Final report STU 69 there are no shortcuts between infiltrated sewage water 519/ U386. (In Swedish) and the wells (Jacks et al. 2000). Carlsson, A.& Olsson,T. 1981:Hydraulic properties of a fractured granitic rock mass at Forsmark, Sweden. Sveriges geologiska undersokn inq SerC 783, Uppsala. Discussion and conclusions Eriksson, U. & Tilly, L. 1984:bversiktlig bedornninq av risken for saltvat The groundwater balance approach presented above gives tenintranqni nq pa Djur o, Varmd o komm un (Estimation of the risk reasonable values of groundwater recharge and withdrawal for saltwater intr usion at Dj uro, municipality of Varmdo ). Report, K konsult. (In Swedish) possibilities in areas with limited groundwater resources, Follin, S. 1995: Geohydrological simulatio n of a deep coastal repository. although a true validation is very difficult to carry out. The Swedish Nuclear Fuel and Waste Management Co. Technical Report simplicity of the balance makes it easily understandable and 95-33. it requires a limited amount of locally measured data, since LeGrand,H.E. 1983:A standardized system for evaluation waste-disposal sites.Report.National Water WellAssociation, second edit ion. much of the information needed can be collected from Jacks, G., Forsberg, J., Mahgo ub, F. &Palmqvis t, K. 2000: Sustainability of hydrometeorological databases, maps,general literature val local water supply and sewage system - a case study in a vulnera ues and from local experience .Therefore, it can be used as a ble environment. Ecological Engineering 1S, 147-153. tool in municipal planning but only as an aid in decision Johansson, P-O. 1987: Estimation of groundwater recharge in sandy till with two different methods using groundwater level fluctuations. making . Journal ofHydrology 90, 183-198. There are, however, several uncertainties, especially Lindberg,J.& Olofsson, B.1997:Risk for salt grundvatten - en studie med related to representative values of the homogeneity and hjalp av GISover delar av Norrtalje kommun (Risk for saline ground infiltration factors. Further experiences from application of water - a study using GIS in Norrtalje mun icipal ity ). Research report. RoyalInstitute ofTechnology (KTHj, Norrta lje municipality and the water balance in various geological terrains will improve SwedishGeologicalSurvey.(In Swedish) the selection of representative values. Olofsson, B. 1991: Groundwater condi tions whe n tunnelling in hard Considerable emphasis should also be given to the typi crystallin e rocks - a st udy of water flow and water chem istry at cal geological sections for each type of soil. Stratigraphy wi ll Staverhult, t he Bolmen tu nnel, S Sweden. BeFo, Research report 160: 4/91. be the most important single facto r, because the kinematic Olofsson, B. 1994: Flow of groundwater from soil to crystalline rock. porosity varies with a factor of up to 10' between different Applied Hydrogeology,vol.2, no 3, 71-83. geological materials. The stratigraphy must therefore be Olofsson, B. 2000: Om VA-fragan pa Alga (About water and sewage defined by an experienced hydrogeologist, and all available water on Alga). Summa ry of presentation at Nacka municipality 2000-04-05. Report, Division of Land and Water Resources, Royal drillholes and previous experiences should be imple Institu te ofTechnology(KTHj,Stockholm.(In Swedish) mented. Another important factor, which requires detailed Olofsson, S., Jacks, G., Knutsson, G. & Thunvik, R. 2000: Groundwater in hydrogeological knowledge, is the delineat ion of the hard rock - a literature review. In Nuclear waste state-of-the-art recharge area. The hydrologically defined drainage basins reports 2001. Swedish National Council for Nuclear Waste (KASAM). SOU2007:35,115-191. are generally dissimilar to the recharge areas of the specific Rodhe,A. 1987:The origi n of streamwater traced by oxygen-18 .Doctora l wells, especially in hard rock areas with well-developed frac dissertation, UniversityofUppsala, Division ofHydrology, Report A 4 1. ture zones. In many cases, the study area should also be fur Sann er, 1995: Bette r decision basis with effecti ve precipitation. ther subdivided into sub-areas based on changes in the Grundvatten 2/95. Geological Sveriges geologiske undersokning, Uppsala.(ln Swedish) geological conditions. In all th e examples presented above, Sund, B.& Bergman, G. 1980: Saltvatte nintrangning i bergborrade brun no such subdivision was made since the terrain type s were nar.(Saltwater intrusion in drilled wells). Swedish Environmental rather similar within the different areas. Research Institute (lVLj, Report B584.(In Swedish) Sund,B.& Bergman,G. 1983:Furusund - gru ndvattenprognos (Furusund - groundwa ter prog nosis).Swedish Environmental Research Institut e (lVL j, Report to the mu nicipality of Norrtatje. (In Swedish) Acknowledgements VaiVa-projekt, 2000: Saltvatte nutredn ing for Saltaro (Saltwater investi The work was fina nced by internal fu ndi ng at KTH and carried out in co gation at Saltaro). Report to varmdo municipality 2000-07-31. (In operation with the Nacka, Vallentuna and Varrndo municipalities in Swedish) Sweden aswell aswith VaiVa-project AB(Cristina Frycklund, now at VIAK VIAK, 1989: Balansberakninq av grundvattenbildning och uttag ino m AB). Prof. Gert Knutsson, KTH, has prov ided helpful and valuable com Algo (Estimation of groundwater recharge and discharge at Alga ). ments on th e manuscript. Report to Nacka municipality, 7989-04-10.VIAK AB.(In Swed ish)