STUDY of WATER DEMAND and SUPPLY in NORTHERN IRELAND Volume 2 Ra I Suppor T in G St Udies 'ARCH August 1984

Depar t ment of t he Environment for Nort hern Ir STUDY OF WATER DEMAND AND SUPPLY IN NORTHERN IRELAND Volume 2 Ra i Suppor t in g St udies 'ARCH August 1984

on • ide n

R1/2 e ms f l- 441% _r 1 fl •

SIR ALEXANDER GIBB & PARTNERS in association with DELOITTE, HASKINS & SELLS MANAGEMENT CONSULTANTS I S 10 10 10 0 1 • • • M O M • • • • • • 1 • M a Depar t ment of t he Envir onment f or Nor t her n Ir elan d STUDY OF WATER DEMAND AND SUPPLY IN NORTHERN IRELAND

Vol u me 2 Su ppor t i n g St udi es August 1984

on onden y

• BELF A T

/ ,5"

SIR ALEXANDER GIBB & PARTNERS in associat ion wit h DELOITTE, HASKINS & SELLS MANAGE MENT CONSULTANTS MNI UMI ONI IINI 10 10 10 1 • 11M M I MI IM IIIM I M DEPA RT M ENT OF THE EN V IRONM ENT FO R NO RT HE RN IRELAND

ST UDY OF WAT ER 6EMAND AND SU PPLY IN NORT HERN IRE LA ND

V OLUM E 2

SUPPO RT ING ST UD IE S

TA BLE OF CO NTENT S

PA GE NO .

INT RO DUCT ION AND S UMMA RY (i )

C HA PT ER 1 SOURC E YIE LDS 1

1.1 In tro du ct ion 1

1.2 River F low Reco rd s 1

1.3 M eteo ro lo gica l Data 10

1.4 Reco rd Extension 14

1.5 Re g iona l Sto ra ge Y ield Relat io nship 2 1

1.6 Yie ld E st im atio n 27

1.7 R iv er Ab stract ions 35

1.8 Severity o f Period (1970- 198 3 ) 37

1.9 Lo ugh Neagh 4 0

1. 10 G roundwa ter 4 1

1. 11 Co nc lu sio ns and Re comm e ndatio ns 42

C HA PT ER 2 HOUSEHO LD SURVE Y 45

2 .1 F ind ings of the Su rvey 45

2 .2 Ho u seho ld Su rvey Qu estionnaire 45

C HA PT ER 3 DETA ILS OF M ET ERED D IST R ICTS 6 1 PAG E NO .

C HA PTER 4 A DJU ST M ENT OF M ET ERE D CO NS UMPT ION TO ACCO UN T FO R

NON -M ET ERED CONSUMERS 6 3

4 .1 Ag ricu ltu re 63

4 .2 Indu stry 65

4 .3 Co nstru c t ion 65

4 .4 Comm e rce and oth er Se rvic es 65

C HA PT ER 5 LEA KAG E IN SUPPLY SYST EMS 70

5 .1 Ob je ctive 70

5 .2 Ba se Year 70

5 .3 Method 70

5 .4 Night F low Ra tes 7 1

5 .4 . 1 Ge nera l 7 1

5 .4 .2 Ea stern Div ision 74

5 .4 .3 No rthern Div isio n 74

5 .4 .4 Sou th ern Div is ion 74

5 .4 .5 Weste rn D iv isio n 75

5 .5. Estim at ion o f Le aka ge 76

5 .5 .1 Dist rib utio n Sy stem 76

5 .5 .2 Tr unk Ma in Leaka ge 76

5 .5 .3 To ta l Leakage 79 L IST OF TABLES

PAG E NO .

TAB LE 1.1 Summ ary of F low Data from No rthe rn Ire land 3

T AB LE 1.2 Ava ilab le Re co rd s fo r An na long Ca tc hm ent

T AB LE 1.3 Red , Blu e , G reen and Contro l Monthly Ave rage

Ca tchment Ra in fa ll 7

TA BLE 1.4 .Re d , B lue , G re en a nd Contro l Month ly Ave ra ge

C atchme nt Run off 8

TA BLE 1.5 . Summ ary o f .F low Da ta from Re pub l ic o f Ire land 9

T ABL E 1.6 Summa ry o f Flow Da ta fr om South We st Sco tland 9

TAB LE 1.7 Annua l Catchm ent Ra in fa ll fo r No r the rn Ireland

Ca tchm ents 11 TAB LE 1.8 Ra in fal l Ratio s for Sile nt V a lley and Anna long

Ra ingau ge s 12

TAB LE 1.9 SAA R (194 1-70 ) fo r Silent Va lley a nd Anna long

Catchm ent 12

TA BLE 1. 10 De fin it ion o f Va riab le s used for Record Ex tension 16

TA BL E 1. 11 Corre la tio n Mat rix fo r An na long Record Ex te ns ion 17

TAB LE 1. 12 Correlation Ma trix for Woodbu rn Record Extension 17

T A BLE 1. 13 Sto ra ge Requ ire ment fo r 50 y ear Re turn Pe rio d 24

T AB LE 1. 14 Y ie lds of Re servo irs in No rthe rn Irela nd 28

T A BLE 1. 15 Yie ld of Rd ver Derg at T iev enn y • 38

TAB LE 1.16 Data an d Ra nk o f Highest De fic it in Pe rio d 1970- 1983 38

TAB LE 1. 17 Re turn Pe riod o f Ar magh Rainfa ll fo r Se le cted

Du rations • 39

T ABLE 2. 1 Own ersh ip of Wate r Us in g App li ance s 46

T AB LE 2 .2 Hou seho ld Cha ra cter is tics by T ensu re Stat us 47

TA BLE 4 .1 Co n centratio n of Ag ricu ltu ra l Water U sage 64 •

TAB LE 4 .2 Con ce nt ra tion o f Indu st rial W ate r Usa ge 66

TAB LE 4 .3 Co ncentra t ion of Co n stru ct ion W ate r U sage 67

TAB LE 4.4 Con cent ra tion of Serv ice Wate r Usage 68

TAB LE 5 .1 Me te red M inim um N ight Flow s in Se le cted A re a s 72 t ABLE 5 .2 M ete re d Le g it im ate Night F lows in Se lected A reas 7 3

TABLE 5 .3 Night F l o W Rates and Leakage 77 L IST OF F IGURE S

Fo llow in g Page No

F IG U RE 1. 1 Lo ca t io ns o f Gaugin g Stations us ed in Stu dy 1

F IG URE 1.2 Sile nt Va lley Sy stem 3

F IGU RE 1.3 Woo dbu rn Comp lex 5

F IG URE 1.4 Re la t ionsh ip Be tw ee n Potent ia l and A ctua l

Ev ap o ratio n 13

F IGU RE 1.5 Sto ra ge/Yield Diagra m s for An na lo n g Ca tchment

189 5- 19 79 2 1

F IG URE 1.6 Rela t io nship be tw een 10 ye ar Retu rn Pe rio d

Sto ra ge and 80 Pe rcentile Ex ce ed an ce D is charge

fo r Rep ub lic of Ire land Ca tchm ents 22

F IGU RE 1.7 Sensit iv ity o f Sto ra ge/Yie ld to Pe riod o f

Reco rd - A nna lo n g 23

F IGU RE 1.8 50 Yea r Re tu rn Pe rio d Storage/Y ie ld Relat ionship s 24

FIG URE 1.9 Re giona l De sign Cu rves fo r No rthe rn Ire la nd 25

F IGU RE 1. 10 Comp a r ison of Re giona l Sto rag e/ Yie ld Relationship s 25

FIG URE 1. 11 Yield/ Fa ilu re Re la t ions hip fo r Sile nt Va l ley 32

F IGU RE 1.12 Retu rn Pe riod o f Rationing fo r G ive n Yie ld s

and Proba bi lity of To ta l Fa ilu re 34 INT RO DUCT ION AN D SUMMA RY

Th is vo lume of the repo rt on the Study o f the W ater Dema nd and Sup p ly in No rthe rn Ire land comp rise s chap te rs w hich d escribe the wor k ca rried out on a numb er o f fundam enta l to p ic s w hich supp o rt the ma in study . These a re sum ma rised as fo llow s :-

1. SOU RCE Y IELDS

Th is ch ap ter compr is es a re v iew of th e data tha t is av a i lab le on w ate r re so u rc es , p art icu la rly su rface w ater sou rces , w hich a re p redom inant in th e Pro vince . A regiona l sto rage/yie ld re la tionsh ip is de ve lope d and u se d for the asse ssm ent of the re l iab le y ie lds of a ll reservo ire d ca tchments in No rth ern Ire la nd at retu rn pe rio d s of 1 in 20 and 1 in 50 y ears . For the S ile n t Va lley sys te m , an asse ssme nt is made o f th e ava i la b le yie ld tha t wou ld re su lt from th e = po s itio n of res tr ic t io ns in supp ly fo r a range of freq uenc ie s .

Rive r ab stract ion and gr oun dw ater a re also rev iew ed .

2 . HOU SEHO LD SURV EY

A su rv ey of 25 00 househo ld s wa s ca rried ou t to asce rtain th e ow nersh ip o f w ater us ing fa c ilities and app lian ce s . T he resu lts a re sum m a rize d by type of ho u se ho ld . A copy of the su rvey .form is app ended .

3 . SURVEY OF M ET ERED D IST RICT S

Details of the 76 m ete re d di s tric ts wh ich we re ide ntifie d for th e pu rp o se ,o f assessin g cu rrent w ater consum p tion a re p re se nted .

(i ) 4 . ADJUSTM ENT OF MET ERE D CONSUM PT ION FIGU RE S TO TA KE ACCOUNT OF NON-

M ET E RED C ONSUME RS

Th is ch apte r de sc ribes the ad ju stm e nts th at have bee n made to a llow for non -m ete red con sumers in agr icu ltu re , indu stry , construction , comm erce and othe r se rv ices .

5. LEA KA GE IN S UP PLY SYST EM S

An exp la na t ion is given of th e way in wh ich es t ima tes have be en made o f lea ka ge ra tes du ring the base y ear on a su b -divis io nal bas is . M inim um n igh t flow ra tes are ava ilab le from wastewate r mete rs or district mete rs for ov e r 8 0% of th e Prov in ce by su pp ly . Leg itim ate night flows a re dedu c ted from mi nim um night flow s to de rive net nigh t flow s and lea kage rates . Estim ates of tru nk m ain leaka ge a re m ade on th e bas is of th e resu lt s of a surv ey of da ta from UK W ate r Au tho rities . C HA PT ER 1

SOU RC E Y IE LD S

1.1 INT RO DUC T ION

Th e ob jec t iv e of th is aspect o f the stu dy is to assess the yie ld

of ex ist in g and p o te ntia l wate r so u rces i n No rthe rn Ire la nd . A lthou gh

groundwa te r re so urces 'are Inc luded , th e ma in em phas is of the wo rk is on

su rface w ate r re sou rces and pa rt icu la rly the y ie ld of impound ing re se rvo ir s

a nd lou ghs . Th e ap p ro ach to yie ld ca lcu la t ion fo r up la nd sou rce s has. bee n

to dev e lop a reg iona l s to rage/y ie ld re la tionsh ip fo r No rthern Ire land .

Fo r the Sile nt Va l ley sou rce yie lds have also be e n es tim ated by ca r ry ing

out a mo re deta iled simu la tion o f re se rvo ir beh av iou r . Estim a tes have also

bee n made o f th e yie ld of d irect riv e r ab stra ct io ns , the hydro logy o f Lo u gh

Ne agh and the g rou ndw ater re so urces of the p rov in ce . II

1.2 RIVE R FLOW RECO RD S

Th e Wa te r D ata Un it (N I ) has ope ra te d a to ta l o f 73 gaug in g sta tio ns ,

some of w h ich reco rd o nly leve ls w h ilst o th ers ha ve be e n disco nt inued .

Re co rd s of mea n da ily flow s . are arch ived fo r 30 s ta t ions with ICS Compu t ing

L im ited in Be lfast . G iven the t ime cons t ra ints o f the p rese nt stu dy on ly

the se arch iv ed re co rds we re co ns ide re d in de tai l : th ese are lis te d on Tab le

1.1. Mo st of th ese re co rds a re rated riv er sec tions w ith cablew ays and the

ove ra ll qua lity of the da ta is good .

D iscu ss io ns w ith W DU ide nt ified som e po ore r quality flow reco rds .

The se inc luded stations w he re the natu ral f low re g ime is signif icantly

in flue nced by art ific ia l fac to rs (su ch as rese rvo irs or cana ls ), sta t ions

whe re , as a re su lt o f an unstab le co ntro l , the re is w ide sca tter of po in ts

o n th e rat in g cu rve and the flow reco rd is o f low accu ra cy : and statio ns

ha v ing a be tte r qua l ity ga ug ing station e ithe r up or dow nstrea m . These

in fe r io r (fo r th e pu rpo se of the cu rre nt study ) f low reco rds am ount to 8 o f

th e 30 p ro cessed re cords and th eir station n umb e rs a re s hown in p a re n th es is

in T a b le 1.1. Figu re 1. 1 show s the loca t io n of ea ch of the gau ging

stat io ns used in the study . sc; )

• RESE RVOIRS • 2 0 4 0 0 1 . RIV ER FLOW GA UGI NG STATIO N o° 20 29 9 9 • 2 0 3 0 28 I ALT NA HEGLIS 0 RAINGAUGE H , . 20 3 0 11 • 2 0 10 0 7 20 ?•0 A 2 9 5 0 10 AS 20 30 2027 13° 21 . 20 3 0 27 20 5 9 9 5 6 20 10 0 8 WOODB URN 2 0 10 0 2 2 3 7 20 30 18 2 0 3 012 41, 2 0 10 0 5 LFAST '. 2 0 10 0 6 .7 0 50 0 5 . LOUGH ERNE 6 20 50 0 8 33 6 9 9 20 3 0 100 • 2 0 3 0 2 5 ARM AG H° 3 3 5 0 11 20 30 3 3. . 3 3 6 0 15 SILENT VAL L EY •• 2 0 6 9 95

3 26 0 11 30 6 0 2 1e • 30 60 2 3 . 6 3 2 6 0 19 • 30 70 0 6

3 0 90 0 1. DU BLIN

6 3 14 0 0 5

6 316 00 4

6 3 2 3 0 0 2

Dzz,

.2 9 L OCATIONS OF GA UGING ST A T IONS USED IN S TUD Y

Figur e 1. 1 It w ill be seen from Tab le 1.1 that all the standa rd hyd rometric sta t ions have sho rt reco rds : the ave rage reco rd length is 8 years w ith no da ta availab le be fore 1970 . Th is data was the re fore enhanced from th e fo llow ing sou rces :

Anna long gauge 1895- 1979 . (Figu re 1.2 ). This gauge is a rectangu lar p late weir with low flow notch , ma inta ined in good cond ition , w ith little leakage and clean straight app roach section . The ca libration o f the direct f l ow reco rde r is unknown as is the detailed history o f datum che cks on the we ir . The site lacks a sta ff gauge . Howeve r, it

is tho ught th at the qua lity of reco rd may we ll be equa l to that of othe r long flow reco rds in the Brit ish Isles and given that the

catchment feeds the ma jor sou rce fo r Belfast , it was conside red essentia l that an attemp t wa s made to ana lyse th is reco rd . Man ua l ab strac tions of flows from the large , 4 ft squa re , loga rithm ic chart

reco rds was ca rried out by WDU (NI ). These fl o ws were read at three

hourly inte rvals du ring rapid changes of river f l ow and at da ily

in te rva ls during pe riods of relative ly constant discharge . The data was then p rocessed by the Inst itute of Hydrology to produ ce a mean

daily fl o w reco rd from 189 5, although the re are ma ny miss ing pe riods (Tab le 1.2 ).

The S lieve Binnian tunne l was constructed in 1955 to divert the river Anna long flows to the Sile nt Va lley reservo ir and thereby increase its yie ld (Co lebrook , 1955 ). The intake to the tunne l is constructed as a mass concrete st illing pool w ith two orifices to the tunnel and an overflow weir discha rg ing wate r into the old river course . Flows

down the tunnel are measu red by a non- sta ndard d irect f low recorder using the se rec tangu lar orifices as a contro l. The deta iled calibration is not known . Ch arts are ava ilab le from 19 57 to 1979 and manual ab stractions of these charts has been carried ou t by the

Institute o f Hydro logy to prov ide a reco rd of mean daily flows .

To compute natura l catchment flow po st 1955 it is necess ary to sum the tunnel diversion d ischarge and the flows at the original Annalong

gau ge . Unfortunately this is not po ssib le because of gap s in the An na long reco rd . However , an exam ination of the tunnel charts • • IS • N M • • MIII • 0 • • ONI OM • UM • OM • •

1. Re co rds from station num be rs in b ra ckets to ) COC used .w ith c a utio n 2 . Mon th ly sta tio ns No 8E 1 - Q 95 an d Q80 fro m month ly da ta r - / / ". CATCH MENTS / \ . ...• BOUNDARY ...-• \ 7 \ ------". " \

\ \ \ s„ " I / b . N .

/ •

LOUGH BEN GROH • NAGH RESERVOIR

/ , • I / l \ N N \ \ 1 \ SILE NT VALLE Y - k \ ...... TUNNEL 1 \ RESERVOIR GAUGM G STATION : \ t k ...- / \ ess-:+0.--c - \ ...... / S‘P 4- 1: -. / ' #;: r4 N AL O N G .- % N GAUGING / / I ° STATION / / .. .. — ..- i e / i / \ I \ / 0 ...... - -- • . . ...• 15. 1 A 2 2 a

R A IN G A UG ES : a S ile n t V a l le y W a te r W o rk s

b S lic ve R e a rn a gh

Lo u gh S han n a gh

il S lie v e L araaga n

A nn a lo n g

S ILENT VALL EY SY ST EM

Figur e 1.2 T ABLE 1.2

AVA ILAB LE RECO RDS FOR T HE ANNA LONG CATC HMENT

M ONTHLY DA ILY

ANNA LONG GA UG ED FLO W S T UN NEL GA UGED CATC HM ENT A RMAGH

FLO W S RA IN FA LL RA IN FA LL

189 5- 189 7 in comp le te 1853

1898- 1899 co mp lete

1900 m iss in g

190 1 in co mp le te

1902- 1905 mi ss ing

1906-.1909 incomp lete

19 10- 19 11 com p lete

19 12- 19 16 incomp le te com plete

19 17 comp le te

19 18- 19 19 incom p le te

19 20- 1927 comp lete

1928 incomp le te

1929- 19 34 co mp le te

19 35 incomp le te

19 36- 1939 comp le te

1940 in com p le te 1939

194 1- 196 0 m issin g 1955- 19 58 incomp le te

19 59- 196 3 comp lete

196 0- 19 79 incomp le te 1964- 1966 in com p lete comp lete

1967- 197 1 mi ss ing

1972- 19 74 incom plete

1975- 19 79 comp le te 11 0 198 0 revealed that the tunne l diverts all hut the peaks of the extreme, ' floods , (it wås designed to divert 97% of ite catchment 's runoff ). • It vas therefore possible to eetimate flows at the Annalong site by multiplying tunnel flows by 1.347 to allow for .the difference in area... • and 'mean annUal effective rainfall of the partial .and total catchMent • areas .

(ii ) Altnaheglish catchme nt 1926- 1959. Despite extensive enquiries by WDU (NI ) the orig in of this weekly inflow record is UnknoWn . Month ly • flows have been calculated and ueed in the stora ge yie ld ina lysis ;

(iii i.Woodbu rn expe rimental. catchments . Four experimental catchments were instrumented by the Belfast City and District :Water Commissioners .to: studY "the effec ts of -afforestation on w ater runoff": .(Savill , 1974 ): The catchments , named the Red ; Blue , Green and Control areas , are .. shown in FigUre• 1.3. Direct flow Lei 'recorders were insti lled in 1959 together wIth .Weirs hav ing ,90 degree v-notChes :for low floWs, and rectan gular weirs for higher flows: A weighted mean of the four . monthly flow. records (Table 1.4) have been used to develoP a råinfall - • runoff relatiOnahip for the Woodburn Comp lex :4/11db has enabled . . . . -synthetic monthly flows to be eitiMaied .

(iv ) Republic of Ireland . .Discus sion with the Office of Pub lic Works , Dublin , identified 27 stat ions (from a netWork Of 200) with recOrd lengths in excise of 20 years; having natural :flows of accep table accuracy . :This data has been transferred to IH and was initially examined to identify which catchments had similar floW reg imes t o those of Northern Ire land . Eleven such records were identified , their ,locations .are shown in figure 1 .1 and the ir catchment deta ils are listed on Table 1.5. The Electricity Sup p l y Board , DU6lin , also operate gauging stations and have Made aVailable monthly gauged • outflows and inflowe to Lough Erne based on measurement at Cathleen , Falls P woer station ind ch anges in Lough level .

(v ) South West Scotland . In view of.the proximity o f south west Scotland

to parts of North ern Ireland it W M considered app ropriate to use . flow recOrds from the Solway Pnrification Board . The c atohmente Were selected -66 the basis of reco rd length and simi larity to the ' a n e mo n s o n — a a m a a m = o de

- - - ' --. RA I NC A UG ES : .. - ‘ .. ‘.. \ -.. / I 1) M o n th ly rn in g nu ge s \ '' '' --N % 2 ) (o r R e d , G re e n D lu c - - CATCHMEN1S 3 ) an d C o n tro l c a tc h - BOUNDARY 4 ) m e n ts . LOUGH 5 ) W o o d b u rn No r th MO(JRNE 6 ) W o o db u rn S o u th

L 7 ) D o r ls lan d

- --.. --- — N. **.. • v .. / -.. \ -...... _ ., — "\.." \ , NORTH WOODBURN \ 14 1 RESERVOIR COPELAND nEsErwoin 5 RED, BLUE, GREEN AND • conTROLCATCHMEN TS

MIDDLE r

A LOWEr ' \ woom unN 4 RESERVOIRS UPPEn

7 \ • CARRICKFERGUS DORISL AND RESERVOIR / ‘ / \ / / \ ..-- ___ reservoired catchments of Northe rn Ireland , that is, having high

annual ave rage rainfa ll and impermeab le geo logy . The deta ils of the se lected catchments are summa rised in Tab le 1.6 .

Values of ave rage flow (ADF ) in cumecs, base flow index (BFI )* , the 95 and

80 pe rcentile discharge from the flow duration curves (Q 95 and Q80 ), catchment area and standard annual average rainfa ll 194 1-70 (SAAR ) are shown in Tab le 1.1, 1.5 and 1.6. An inspection of the se data indicated that the variat ion in flow ind ices between catchm ents is sma ll in North ern Ireland in relation to the rest of the British Isles , that the catchme nts are genera lly impe rmeab le and that a single regional storage y ield re lationsh ip may be app rop riate to use throughout the province . This ana lys is of both flow data and catchment characte ristic s prov ided an object ive bas is for sele ct in g flow records from outside Northe rn Ire land which were simi lar to those w ith in the pro vince .

• The ratio of ba se flow di scha rge to total diScha rg e (Institute of Hydrology 1980 ). M I I S a M I M N W I OM - - = 11 M MIN S

TABLE 1.3

RE D , BLUE , G RE EN AND CONT ROL MONT HLY A VE RA GE CATC HM ENT RA IN FALL

YEAR J AN FEB MAR APR MAY JUNE JU LY A UG SE PT OCT NOV DEC TOT AL

1960 124 .2 99 .6 8 3 .6 94 .2 53 .8 6 3 .0 230 .4 142 .7 9 3 .0 176 .3 135 .1 79 .2 1375 .2

196 1 134 . 1 106 .2 48 .5 140 .7 93 .7 67 .8 69 .1 105 .9 120 .9 145 .8 67 .3 102 .9 1202 .9

1 19 62 132 .8 60 .2 48 .8 72 .9 73 .7 45 .7 68 .1 147 .3 ...1 194 .6 5 1.3 119 .6 130 .6 1145 .5 1 196 3 47 .5 39 .1 94 .2 68 .3 10 1.9 12 1.1 96 .0 134 .4 88 .1 149 . 1 190 .2 28 .4 1 158 .3

1964 5 3.3 25 .9 108 .2 66 .0 6 5 .0 86 .4 48 .8 128 .0 104 .6 169 .2 73 .7 128 .0 1057 .1

1965 109 .5 17 . 1 88 .0 94 .7 64 .3 65 .3 105 .2 84 .9 119 .0 76 .4 187 .0 14 1.3 1103 .3

1966 68 .2 17 3 .9 100 .5 107 .8 111.5 107 .5 77 .8 110 .8 102 .5 13 1.2 100 .1 184 .9 1376 .6

1967 119 .8 78 .1 100 .7 65 .5 84 .5 67 .0 76 .2 124 .2 168 .1 168 .1 14 3 .1 96 .1 129 1.7

19 68 14 2 .1 86 .1 52 .2 8 8 .8 98 .7 8 3 .4 73 .7 76 .9 133 .6 124 .7 168 .2 116 .2 1244 .5

1969 136 .9 82 .2 88 .2 103 .0 108 .3 83 .9 82 .6 49 .6 33 .6 60 .2 103.9 114 .5 1046 .6

19 70 12 1.3 9 1.6 119 .5 9 7 .6 68 .7 48 .0 83 .3 140 .5 158 .7 106 .3 144 .5 86 .0 1265 .8 MO I S OM 6• 11 M O M I 11• 1 - M 111 - UM IM

TA BLE 1.4

RE D, BL UE , G RE EN AN D CONT ROL CATCHMENT RUNOFF

YEAR JAN FEB MA R APR MA Y JUNE JULY AUG SEPT OCT NOV DEC TOTAL

1960 9 1.95 83 .82 74 .68 6 1.21 15 .75 10 .92 , 50 .29 65 .79 51.31 79 .25 129 .54 67 .31 781.82

196 1 107 .70 88 .90 30.48 82 .55 64 .0 1 16 .5 1 11.43 16 .00 35 .56 92 .20 5 1.31 100 .08 696 .72 1962 82 .80 56 .90 27 .43 45 .97 23 .11 11.43 10.67 33.02 99 .82 39 .88 82 .55 98 .55 6 12.14 1963 30 .99 62 .23 7 1.63 45 .72 49 .53 39 .12 42 .42 49 .53 46 .48 89 .92 144 .78 48 .77 72 1.11 col 1964 33 .02 28 .70 88 .14 40 .13 36 .07 32 .5 1 14 .48 32 .00 40.64 120 .90 40.64 116 .33 623.56 196 5 113 .53 38 .11 56 .74 78 .08 37 .03 25 .70 3 1.36 32 .7 1 46 .51 5 1.39 82 .48 162 .44 757 .08 1966 75 .21 125.8 1 108 .44 80 .09 68 .70 53 .5 1 34 .03 37 .16 34 .11 105 .90 70 .68 158 .06 951.70 1967 105 .43 76 .12 77 .25 37 .78 4 1.15 34 .29 20 .62 32 .14 83 .79 105 .57 124 .59 79 .54 8 18 .27

1968 133 .27 79 .04 43 .80 66 .54 58 .28 39 .73 29 .94 26 .59 48 .99 82 .00 134 .02 109 .10 8 51.30

1969 118 .95 94 .80 82 .80 79 .83 8 1.35 53 .74 33 .40 28 .83 16 .70 17.56 36 .6 0 69 .23 708 .89 1970 111.49 88 .79 108 .41 73.32 6 1.27 27 .10 22 .24 46 .41 66 .73 57 .30 99 .97 92 .74 855 .77 OM OM N M I = E MI • IIIMI NM 0 MII I0 I= I = • I= a • • 1• •

TABLE 1.5

SUMMA RY OF FLO W DATA FROM RE PUBL IC OF IRE LAN D

NUM BER ST AT ION PER IOD ADF 13F I Q 9 5 Q8 0 A RE A SAA R cam eos %A DF BA DF sq km mm

3 06 02 1 G ly de at Mansf ie ldtow n 1955- 1980 4 .857 0 .640 11.94 25 .45 3 12 92 1 306 023 Dee at D rum go o le stown 1954- 1980 3.8 37 0 .553 11.13 2 1.14 302 924 307006 Moy nalty at Fy anstown 19 56 - 1980 2 .9 54 0 .5 59 9 .11 22 .27 179 9 59 309 00 1 Ryew ate r at Leixlip 1956- 198 2 2 .56 1 0 .496 3 .94 12 .69 2 15 860 3 14005 Ba rrow at Po rtal in gt on 19 55- 198 1 6 .384 0 .5 13 7 .22 19 .85 398 966 3 16004 Su ir at Thu rle s 1954- 198 1 4 .157 0 .5 16 5 .84 16 .62 236 950 323002 Fea le a t L istow el 1946- 1980 22 .2 03 0 .309 5 .94 14 .62 646 128 1 3260 11 Breedo ge at Be lla Bridge 1953- 198 0 1.920 0 .365 4 .06 12 .24 112 1150 326 0 19 Caml in at M u lla gh 1953- 1980 4 .743 0 .503 8 .64 23 .55 26 0 968 33 50 11 Bo ne t at D romh air 1957- 198 1 9 .432 0 .343 8 .72 2 1. 34 294 1430 336 0 15 F inn at An lone 1956- 1979 2 .8 05 0 .363 3 .67 13 .69 175 1022 1 336 999 Ern e at Pow er Stat io n (in flo ws ) 1900- 198 3 96 .755 - 11.48 32 .48 4 349 1489 w 1 1. *month ly stat ions - No BFI - Q 95 and Q 8 0 from m onth ly data .

TAB LE 1.6

SUM.MARY OF FLOW DATA FROM SO UT H W EST SC OTLA ND

NUMBER ST AT IO N PER IOD ADF BFI Q 95 Q 80 AREA SAA R cum ecs %A DF liA DF sq km mm

79 002 N ith at Fria rs Ca rse 19 57- 1982 2 5.57 0 .380 10 .98 23 .06 .799 1598 79 003 Nith at Ha ll B r id ge 1959- 198 2 5.25 0.270 6 .72 15.32 155 1692 790 04 Scar W ater at Capenoch 19 63- 198 1 5 . 14 0 .3 10 6 .42 19 .29 142 1700 79 0 05 C lu den Wate r at Fidd le rs Fo rd 1963- 1982 7 .45 0 .370 6 .93 15 .44 238 14 07 79 006 N ith a t Drum la n rig 196 7- 1982 15 .27 0 .340 9 .13 18 .42 4 7 1 16 13 8 000 1 U rr at Da lb ea tt ie 19 6 3- 1980 5 .4 5 0 .350 4 .87 15 .50 199 132 1 8 1002 C ree a t New to n Stewa rt 1963- 1980 14 .85 0 .270 7 .03 18 .22 368 17 15 8 100 3 Lu ce at Ai ryhemm in g 1967- 1982 5 .7 5 0 .230 5 .10 11.86 17 1 1433 8 20 0 1 G irv an at Rob sto ne 196 3- 19 80 6 . 11 0 .34 0 9 .17 17 .90 24 5 14 35 1.3 METEO ROLOG ICA L DATA

(a ) Rainfa ll data

Annua l rainfalls for gau ged catchmen ts (Tab le 1.7 ) were provided by the Meteo ro logica l O ffice Be lfast and are used to estimate average annual losses for Northern Ire land . In addition standard annual avera ge rainfa ll 194 1-70 (SAA R ) for each of the reservoired catchm ents (except Silent

Valley ) were obta ined from Table C 3.7 o f Water Statist ics 1980 and are shown in Table 1.14.

Month ly ra infall for extend ing the Woodburn flow reco rds was based on th e North Woodbu rn ra ingauge from 1886- 1980. These we re calcu lated for the comb ined Red , Blue , Green and Contro l catchm ent areas by using a ratio of po int to catchment ra infa ll based on SAA R values . For the period o f concu rrent flow and rain fall data (1960- 1970 ) a mo re accurate month ly catchm ent rainfa ll was ca lculated by weigh ting the da ta from 6 raingauges in or near the catchmen ts . (Figure 1.3, Table 1.3 ).

Tab le 1.8 shows the rat io of the 194 1-58 average annual ra infa ll to the 194 1-70 annua l rainfa ll for five gauges in the Silent Valley area . Inconsistenc ies in the ratios for the Sile nt Valley waterworks gau ge (due to relocating the gau ge in 1958 ) and the Slieve Lamagan gauge were d iscussed w ith the Meteo rolog ica l Office (Be lfa st and B racknell ).

Fo llow ing the se discussions the Meteoro logical Office revised catchment

SAAR values based on 1:250,000 isohy etal ma ps (Table 1.9 ). Mo nth ly catchment rainfa ll for the Anna long and Silent Va lley catchments for the p eriod 1939 to 1980 w ere then dete rm ined by w eighting the three cons istent rain gauges by the approp riate gauge and catchm en t SAA R value .

(b ) A ctual eva oration

Fo r each rese rvoired catchmen t actual evaporation (Ea) is required to estima te the ave rage annua l discharge from ca tchment ra infall. Penm an potential evapo transp iration (Ep ) fo r short gra ss , adjusted fo r mea n catchment altitude , has been used as an estimate of ca tchment actual evaporation . An investigat ion of catchment losses show ed that the d ifference between mean Ea and mean Ep is v ery sma ll in Northern Ireland : the resu lts are desc ribed be low .

- 10- I 11E 1 NMI 11= I ON M O IMO M a Il l e ONO I O a I II I UM IM O = I= •

TA BLE 1.7

ANNUAL CAT C HM EN T RA IN FA LL FO R NO RT HERN IRE LA ND C AT C HM EN TS

NJ NJ NJ NJ NJ NJ N NJ NJ NJ NJ NJ NJ NJ 0 0 0 0 0 O 0 0 0 0 0 0 0 0 0 0 Li Li Li l a Li Li L i L I Li ie ; 0 O tcsj 0 0 5. 0 0 NJ NJ N NJ NJ NJ NJ Li NJ 0 Li co ,;;O p- a %.0 Li l a fl

0 IC g 0 0 0

Yea r

194 1- 1174 1170 1174 146 2 1005 14 16 1158 1238 1066 1113 1267 12 15 104 6 96 5 100 9 1299 1120 1056 1455 1239 100 7 935 9 15 1252 1020 70 19 7 1 78 8 119 3 907 1024 878 9 19 110 5 79 6 19 72 1222 983 12 39 992 108 2 9 54 104 8 1194 10 79 9 32 898 954 920 1330 1/15 19 73 116 5 1057 1079 89 5 1222 9 19 1052 89 0 926 1169 102 2 88 1 82 8 84 2 108 3 97 1 1125 84 2 75 0 77 7 1025 85 1 1974 1297 1132 1189 10 2 1 1322 1083 115 1 108 1 109 2 13 54 12 18 10 19 10 15 949 120 1 1157 111 1 1159 10 11 1040 922 1300 1053 197 5 10 1 1 84 2 88 7 723 10 39 779 929 8 17 79 7 1025 9 14 70 1 68 3 75 5 99 6 87 9 83 3 938 7 18 70 0 672 1005 75 5 19 76 1152 1086 1048 1495 95 5 124 5 111 1 1093 10 11 10 52 1324 1110 9 75 906 948 115 1 1028 10 12 1548 110 1 9 18 9 50 36 1 139 1 96 7 1977 12 18 114 3 1167 16 08 950 134 5 107 1 1182 1024 109 4 128 6 1194 102 0 90 1 88 3 125 0 1126 1024 148 1 116 5 953 93 0 875 1243 959 19 78 129 1 1205 122 1 17 18 844 14 16 1158 12 38 1098 118 3 13 30 119 1 994 89 7 9 58 1299 1064 115 1 148 4 1177 1098 10 30 9 72 12 39 94 9 19 79 1338 12 17 1256 17 18 10 15 134 5 118 1 126 3 114 1 1146 134 3 122 7 106 7 95 5 1080 124 7 1098 115 1 138 2 1227 106 7 96 7 93 5 13 39 104 0 198 0 1402 1158 1237 179 6 980 14 19 1119 1296 1138 1226 1293 1257 1056 996 1130 1335 1142 1162 146 1 1269 1057 98 0 94 1 128 0 10 13 TABLE 1.8

RA INFALL RAT IOS FOR THE SILENT VALLEY AND ANH ALONG RA INGAUGES

IRISH SAAR ' RAIN FALL RAT IO . GAUG E GRID 194 1-70 144 1-58, 1941 -70 - REFERENCE ram

( a ) Silent Valley 3 305 216 1 355 96 .0 W at er .WOrk e

(b ) Slieve Bearnagh 3 314 279 1770 98 .6

(c ) Lough Shannagh 3298 259 •1808

Sliev e Lamagon 3322 255 1787 101.2

(e ) Annalong 3 355 255 1240 98 .5

TABLE 1.9

SAAR (194 1-70 ) FOR SIL ENT VALLEY A ND 'ANNALONG CATC HMENTS

- • cATCHMENT • SAAR AREA (mm ) (ha )

.Si lent Valley (re sid ua l F atchment te low 1730 - J 4 18 :8 : Ben Cram exclud in g Annalong diversion )

Ben Crom 1845 8 10 .2

Silent Valley Plus Ben Crom 1772 • 2229 .0 excluding Annalong

-inhalOng abOve Tu nnel DiversiOni 1726 1011:3

Anna long above Mnu rne Conduit 1654 142 1.8

sileht valley„ sen crOm and Ann alen g abO've Tu nn el Diver sion 1758 324 0.3 I . The mean and sta ndard deviation of annual losse s (rainfall minus ru noff ) were ca lcu lated foi each gauged catchm ent . Ten catchments w ith eigh t or more years of flow data and a low standa rd dev iation of annua l .

losses were used for further analysis . The ave rage annual losses for this

group of catchme nts was 373 mm . The mean Ep va lues (Table C .3 .5.3 DOE , N . Ireland 1980) ca lculated for clima to log ical stations in Northern Ireland were ad justed to sea level using a 'lap se rat e ', provided by the Meteoro logical Office , o f 29 .3 mm/ 100 m . The average Ep o f 13 c lima te

stations was 44 0 mm ad justed to sea level . (T hese 13 stations exc lude d

coasta l sites , the Silent Valley Water Works site wh ich serious ly ove r- estima ted Ep , and sites with less th an 10 yea rs of data ). Th is value of I I 440 mm is used as the m ean sea level Ep throughout Northern Ire land , from which any catchm ent Ep can be calculated . A m ea n alt itude ad jus ted Ep of I . 397 mm was ca lcu lated for the ten gauged catchm ents and compa red w ith the actual losses from the catchment wa ter balance of 373 mm , giving a me an

difference between Ep and Ea of 24 mm . Th is as sum es that all losses may be attributed to evapo ration and this is suppo rted by the imperme ab le 1 catchme nt geology . A second and independent method of estima ting th e difference betw een i. Ep and Ea emp loyed a soil-mo isture defic it model described in the fo llowin g section . Th is model es tima ted daily Ea and was run fOr a range of Ep and

SAAR valties . The ra in fa ll w as distributed accor ding to the daily rainfalls at Armagh Observato ry over the pe riod 194 1-70. The resu lts are shown in Fi gure 1.4 where the d ifference between Ep and Ea can be estimated for a

site with a given value of SAAR and Ep . Entering Figu re 1.4 w ith a mean

SAAR of 1 146 mm and Ep o f 397 mm for the catchm ents in th e water ba lance study yields a diffe rence of 5 mm between Ep and Ea .

In v iew of the sma ll d ifferences betw een Ep and Ea from bo th me th ods and the fac t that most rese rvoirs have SAAR values in excess of 1000 mm ,

where any d iffe rences between Elo and Ea woU ld be sma ller than the water balance catchm ent sub se t , it was concluded that Ep cou ld be assumed to be equal to Ea . Fu rthermore , mo st reservoirs are located in upland areas w ith impermeab le so il and geo logy so the only losses would be evaporation .

Hence m ean annual runo ff can be estimated from SAAR Ep .

- 13- 8 0

6 0

E.:7.50 0

fl Er7.45

E 20

E.= 4 0 0

10 EFF 35

Ep= 30 0 60 0 800 10 0 0 120 0 14 0 0 160 0 m e a n a n nua l r a i nf a l l Im m l

REL ATIONSHIP BET W EEN POTENT IA L AND ACT UAL EVA PORAT ION Figur e 1.4 (c ) Soil Mo isture Deficit

The extension of the An nalong and Woodburn flow records also requ ire

estima tes of the so il moisture defic it (SMD ) appropria te to each catchment . •Th ere are no pub lished long term SM D data availab le for Northe rn Ireland and so values have be en est ima ted from clima tolog ical data . A recent study o f SMD models (Ca lder , 198 3 ) has shown that a reliable estimate of SMD .can

be ach ieved us ing a mean estimate of potential evaporat ion , da ily ra infall and a simple , 2 layer , so il molstu re extraction model .

One of the lon gest da ily rain fa ll records in Northern Ireland is

A rma gh Ob servatory (1853- 198 2 ). Armagh is approx imate ly 50 km from Silent Va lley and 65 km from the Woodbu rn comp lex and it has a mu ch low er SAAR of on ly 866 mm . However the daily catchm ent rainfall values can be estimated using the ratio o f (catchment SAAR/Armagh SAAR ) to prov ide a reasonable basis for SMD ca lcu lation . Th e me thod for estim ating mean ann ua l catchment potentia l evap oration has been described ea rlier ; values for the Silent Valley and Woodburn catchm ents were 325 mm and 38 0 mm respective ly .

Daily SMD va lues were ca lcu lated for each site using the simp le two layer mode l, daily catchment rainfa ll -and the sea Sona l distribut ion of mean annual potential evaporation . Th e 'start of the m onth ' SMD va lu es were extracted from the resu lts for use in the flow ext ension models described below .

1 . 4 RECORD EXT ENSION

(a ) Model development

It was necessary to in fill gaps in the Annalong reco rd and to ex tend the Woodbu rn flow sequ ence to p rovide additional data on which to base the re giona l sto rage yield analysis and to carry out a more detaile d simulation ' o f the Silent Valley reservo ir sy stem .

- 14- Concu rrent flow and rainfall data and lon g rainfa ll records we re available at bo th locations and a regression mod el was used for da ta .extension . Th is method provides ob jective pa rameter est imation and avo ids the need to ma ke sub jective judgem en ts abo ut ca tchment p rocesses which is necessary with most conceptual models . Regress ion ana lys is is used to construct simp le linear relationsh ips betw een any set of inputs and an oUtput variab le , such that the difference between the ob served and p red icted suns of squ ares is a minimum .

Data inpu ts for the regression were rain fa ll , evapo ration and soil mo istu re deficit (SMD) ; outputs were month ly flow s , w ith all va riab les exp ressed in mm . Mod el form u lation is selected pa rtly by expe rience but aided by an analysis of va riance and an inspe ction of the residuals (the d iffe rence between the ob served and predicted month ly discharge ). Th e GENSTAT statistical pa ckage was used to assist w ith the analysis . This program was de sign ed by Rothams tead Expe rimental Station to pe rform ana lys is of variance, and incorpo rate extensive model bu ilding aids such as trans forma tions , residual plotting and subse t search rou tines .

Tab le 1.10 lists the variab les used in developing the regression model. Fitting na tural and logarithmic flow s was carried out at each site based on the ra in fa ll and SMD data available . Examp les of the An nalong and

Woodb urn corre lation ma trices are shown in T ab le 1.11 and 1.12 . The fo llow ing Checks were ca rried out to ensure th at th eie w ere :-

(1 ) No major errors in a simp le wate r ba lance of the rain fall , evapo ration and flow data used t 6 calibrate the model .

(2 ) No tendency to over or under p re dict at high , medium or low flow s .

(3) No seasona l trend in the residua l flows (fitted - observed ).

(4 ) No tren d in the re siduals throu ghout the period of re cord .

- 15- TABLE 1.10

DEFI NI T I ON OF VARIABLES USED FOR RECO RD EXTEN SION

VA RIAB LE DESC RIPTION

FL 1, FL 2 The mean mo nthly flow from th e previou s 1, 2 , 3 and 4th month FL3 , FL4 respe ct iv ely .

RR The current month 's effective rainfa ll (ca tchment rainfa ll

less evapo ration ). If actual evapo rat ion ex ceeds rainfall , RR is set to Zero . This happ ened ra rely .

RRL 1, RRL 2, Effective rainfall (RR ) lagged by 1, 2 , 3 and 4 month s RRL 3 , RRL 4 respe ct ively . It was conside red un necessa ry to include

higher lag te rms because the Catchm ents are sm all , flashY and

imperme able .

SMD T he start of month soil moistu re de ficit de sc rib ed in sectio n

1 .3 (c ). •

• SMDL 1, S 4DL 2, The SMD la gged by 1, 2 , 3 an d 4 mon ths respe ctive ly . SM DL 3, SMD L4

EVA P Actual evapo rat io n dur ing the curren t month .

EVAP 1, EVAP2 Actual evapo rat ion lagg ed by 1 and 2 mo nth s respectively .

;•16- ‘‘ . M O OM M O M MN W I a IMMI a a a I S M I IS M NM M a

TA BLE 1.11

. CO RRE LA T ION MAT RIX FOR ANNAL O NG RECO RD EXT ENS ION

1 FLOW 1 1.0000 r -4 1 LOG FLO W 2 0 .9 256 1.0000

EVAP 3 -0 .3865 - 0 .406 1 1.0000

RR 4 0 .8658 0 .8 307 -0 .5 1 16 1.0000

RRL 1 5 0 .3 122 0 .3648 -0 .5 156 0 .2876 1.0000

RRL2 6 0 .252 2 0 .2808 -0 .4 15 1 0 .23 15 0 .3 142 1.0000

SMD 7 -0 .39 12 -0 .446 1 0 .4242 -0 .359 1 -0 .39 93 - 0 .3349 1.0000

SM D L 1 8 -0 .1198 -0 . 1562 0 .28 03 -0 .0973 -0 .34 58 -0 .4 39 8 0 .38 05 1.0000 SM D L2 9 0 .0794 0 .09 72 0 .0492 0 .13 9 1 - 0 .0 04 2 -0 .2793 0 .1700 0 .3737 1.0000

1 2 3 4 5 6 7 8 9 M I S I M S I OM M N • MN a MN a 111111 0 1 IS M O IMMI NM M I M N M I

TABLE 1.12

CO RRELAT ION MA T RIX FOR WOO DBURN COM PLEX RECO RD EXTEN SION

FLOW 1 1.0000 1 r 03 LOGFLOW 2 0 .9449 1.0000 1 EVAP 3 -0 .6638 -0 .6536 1.0000

RR 4 0 .8310 0 .798 1 -0 .5935 1.0000 • RRL 1 5 0 .5850 0 .6 135 -0 .6782 0 .4 118 1.0000

RRL 2 6 0.4067 0 .424 5 -0 .5636 0 .1799 0.4 126 1.0000 SMD 7 -0 .6040 -0 .6 173 0 .5894 -0 .4568 -0 .55 39 -0 .4800 1.0000

SMDL 1 8 -0 .45 13 -0 .4555 0 .4657 -0 .1910 -0 .4627 -0 .5576 0 .39 18 1.0000

SMDL2 9 -0 .12 18 -0 .0933 0.2153 0 .1123 -0 .1874 -0 .454 1 0 .1492 0 .3926 1.0000

1 2 • 3 4 5 6 7 8 9 (b ) A lication to the Woodbu rn Co lex

A regression was carried out on the 11 complete years (1960- 1970 ) of catchment runoff and rainfall data for the Red , Blue , Green and Control catchments (Tab le 1.3 and 1.4 ). The be st equation obtained , fitting the month ly flow Q in mm is :-

Q = 22 .2 + 0.497 RR + 0.098 BAL I - 0.4 10 SMD - 0 .588 SM DL 2 (1)

R2 = 8 1.5% , standard error o f Q = 15.1 mm .

A ll terms were sign ificant abo ve the 99% con fidence level .

Add itional evap oration , lagged ra infall and SMD terms were not sign ificant at the 95% leve l and hence they do not lea d to any imp rovement in the estimation of month ly flows . Equat ion (1) enables mon thly runo ff to be ca lcu lated for the Woodburn catchment from 189 5 - 1979 .

The av erage daily flow (ADF ) 194 1-70 of the reco nstructed sequence was compa red with the ADF calcu late d for the regiona l analysis . Th ese we re ve ry similar , being 58 .4 Ml /d and 6 1.2 Ml/d respe ctiv ely . The resu lts of a storage yie ld analysis will be heav ily reliant on the estimate of average flow . Therefo re , for consistency w ith the regional study, the synth etic sequence wa s ad justed to comp ly with a 194 1-70 A DF of 6 1.2 Ml /d .

(d ) A lication to the Annalon catchme nt

Th e mode lling based on ad justed tunnel flow s (Section 1.2 ) was carried out on the period of record January 1955 to De cember 1979, the only period for wh ich correspo nd ing catchment runoff and rain fall data are av ailab le (Table 1.2 ). Although a number of gaps remain the statist ical pa ckage used wa s ab le to make optimum use of all data .

The best fit equation obta ined on monthly data is given below (in mm ).

Q 32 .46 + 0.703 RR - 0.437 SMD (2 )

R2 = 75 .7% , standard error o f Q 29 .1 mm .

- 19- A ll the terms were sign ificant abov e the 98% con fidence level .

Inc lu s ion of add itional te rms were not sign ificant at the 95% level . Equation 2 enables runoff to be estimated from 1853 to 1979 for the Anna long catchment using ra infa ll and SM D data . Month ly catchment ra in fall based on the 3 local gau ges has be en used from 1939 to 1980 , but be fore

th is the Arma gh rain fall, transferred using a SAA R rat io of 1.9 1, has be en

used .

Th e ADF of the synthetic sequence (194 1-70 ) wa s compared with the ADF

from the regional study . Th ey were very simi la r , bein g 52.94 M l/d and

5 1.77 M l/d respe ctively . Th e synthet ic sequ ence was adjusted t o comp ly w ith an ADF of 5 1.77 Ml/d as was carried out for Woodburn. The ad justed

serie s was used to infill gap s (Tab le 1.2 ) and to extend th e Anna long

reco rd .

(d ) Co ariso n of Woodburn and Annalon model

It is interesting t o no te th at the structu re of both the Woodburn and Anna long mode ls is sim ilar ; in both cases natural rath er than loga rithm ic

flows were fitted best . From equation (2 ) it can be seen that the Annalong

flows are related on ly to the current month ly ra infa ll and SMD , wherea s equation (1) for Woodbu rn in cludes mo re lagged terms . Any physical

interp retation of the differences in the regression mode ls must be made w ith caution but the genera lly steep er and more imperm eable Annalong

catchm ent would be expected to have a 'flashier ' flow regime , with less

'hydro log ical mem ory ' than the more subdued top ography of Woodbu rn . Th is

is reflected in the ab sence of lagged terms in equatio n (2 ). The larger

error for the Annalong model may be due to th is flashier flow regime , t o a larger catchm ent area , or to the lower accuracy of catchment rainfa ll

(there were no raingauges on the ca tchment to the tunnel in contrast to Woodb urn where the re were 6 gauges on the catchment ). Howev er the

reconstructed flows for the Annalong catchm ent w ill p rov ide a mu ch smaller pe rcentage of the tota l record . W ithin the constraints of the current study it is thought that bo th compo site f lows adequate ly rep resent the histo rica l flows . Fu rthermo re the average flow over the period 194 1- 1970

for the two flow sequences is very close to that estim ated from rainfall- evaporation . Th is indepe ndent check suppo rts the resu lts of the evapora tion studies of SectiOn 1.3 .

-20- 1.5 REG IONAL STORAGE YIELD RELAT IONSH IP

In o rder to estimate the yield of a large number of sou rces a

genera lised stora ge yield relat ionsh ip ha s bee n developed which could readi ly be app lied to any re servoir in No rthern Ireland .

A storage yield analysis wa s carried out on each of the availab le flow re co rds by ca lcu lating the sto rage requ irement Si needed to maintain a

yie ld Y from :-

Si+ 1 = Si - Qi + Y

where Q i is the daily (or mo nthly ) discharge . The va lue of Si will

increase during a drou ght and decrease as 'reservoir ' inf lows exceed the y ield . Th e 'rese rvoir ' w ill spil l when Si is negat ive , in wh ich case Si is

re set to zero . The maximum value of Si, for any drought event , is thus the

stora ge needed to just main ta in the given yield . Th is simu lation is

carried out for the comp lete reco rd and a series of Sj, the annual ma ximum values of Si are extracted . These va lues are ranked from the sma llest

(j= 1) to the largest and the no n-exceeda nce probability Fj is ca lcu lated

usin g the Blom plotting po sition :-

- 0 .375 Fj - n + 0.25

where n is the numb er of years of data . The entire pro cedure is rep eated

for differen t Y va lues . Figure 1.5 shows a plot of the stora ge requ irement

Sj against pe rcentage exceedance on log-no rmal probab ility pape r for a number of different y ields . The y ie ld is expressed as a percenta ge of the

average daily flow (ADF ), and Sj as a pe rcenta ge of.th e annua l runo ff vo lume (ARV ). (Th is facilitates compar ison of storage yield re lationsh ips

for catchmen ts with different avera ge flow s ). A sm ooth curve ma y be drawn through the po ints on Figure 1.5, and the storage required to ma intain a

y ie ld for a given pe rcentage of yea rs without reservoir failure may be

estimated.

Th e above analysis wa s carried out on each of the follow ing flow sequences :-

-2 1- SNP MN I NII 0 0 0 1 IIIIII IS IMI 0 I • OM 0 IS 1NO M E

M r Z ° A di —I CI ... z- .1P10 M CO 1 1 CD 0 0 0 . 1 > C 1 9, . 5 89 SO 95 BO BO 10 GO 50 40 30 . 20 10 5 . 0 2 . 0 1 . 0 0 . 5 - 4, o . K PE RC EN TAGE OF TEAR S RE QU IR ING AN NU AL MAX IMUM GR EATER THAN Y (1) Short records ("10 years)from the N .I. hydrometric network , i.e . the stations included in Table 1. 1

(2 ) Longe r reco rds (20-40 yea rs ) from the Repub lic of Ireland shown in

Tab le 1.5 and South West Scotland (Tab le 1.6 ).

(3) Mo nth ly flow records for Lough Erne (1900- 1983 ) and Altneheglish

rese rvoir (1929- 1950).

(4 ) Month ly gau ged and reconstructed reco rds for Woo dburn (1886- 1980 ).

(5 ) Da ily gauged (189 5- 19 79 ), and mon thly gaug ed and reconstructed f lows

for An na long (189 5- 19 79 ). Th e pre 189 5 data we re not used for the ' final analysis be cau se of their poo rer accuracy due to estimatiOn from the Arma gh raingauge .

The relationship between the storage yield diagram and the catchm en t

flow regime (indexed by the 80 p ercentile exceedance discharge from the

flow durat ion curve ) was invest igated using an enlarged data set for the • Repub lic of Ire land . The 10 year return period storage for y ie lds o f 20%

ADF and 6 0% ADF we re plotted in Figu re 1.6 against the 80 percent ile

exceedance d ischarge , Q80. Although the figu re suggests that there may be

som e tende ncy for flashy , Impe rmeab le catchments with low values of Q80 to require la rger sto rages , the relationsh ip was poor . The conclusion of Section 1.2 w as that catchments in North ern Ireland are genera lly more

impe rm eable than tho se of the Republic and m oreov er that reservoired

catchments were among the most impermeab le areas of the North . Thu s desp ite the inco nclu sive results shown on Figu re 1.5, it was thou ght

desirable only to use the catchm ents from the Republic with simi lar values o f Q80 to those in the North . These elev en catchments were put into two

group s , tho se with a Q 80 between 20% and 25% ADF and the rema inder , having Q 80 betw een 12% and 16 % ADF . For each group of st ations the average sto rage frequency re lat ionship wa s found by avera ging individual station p lots for a given yie ld . The flash ier group of catchments show ed a

s ligh tly greate r sto ra ge requirement for a given yield and frequency and

th is curve w as used in subsequent comp arisons of storage yield diagrams . This averaging procedu re wa s rep eated on the impe rmeab le catchment data sets from North ern Ireland and Sou th West Scotland .

-22- • 4

4 4 I C N fl

• 03

• 4 • • <1 l tO . - 0 • • tx • 4

fl • U3 0 • • • 4

• fl •

fl 0 tO o 0) fl 0

( A ki V % p s• aIn b e l • 6 c 10 4s o o tu rti o A

RELAT IONSHIP BET W EEN 10 YEA R RET URN PERIOD ST ORAGE A ND 80 PERCENT ILE EXCEEDANCE DISCHARGE FOR REPUBLIC OF IREL A ND CAT CHMENT S Figur e 1.6 As a resu lt o f the numb er of severe drou ghts in th e pe riod 1972- 1979 (Section 1.8 ) it was thought that the sto rage frequency re lationsh ip

base d on the sho rt Northern Ireland reco rds would be biased . This wa s

checked by compa ring the storage y ield relat ionship for the An na long derived from the pe riod 1895- 1979 w ith that derived from the period 1972-

1979 . The resu lts are shown on Figure 1.7 wh ich illustrates the large

underestimation of yield for a given sto rage from the short record . (For y ie lds of 20 % ADF and in excess of 80% ADF there were insu fficinet events

to ma ke compa risons ). Th e poo led Northern Ire land and the Repu blic of

Ireland storage y ie ld d iagrams were thus adjusted to allow for this bias . 1./11,-.1 2 In the case of the Repub lic of Ire land the ad justm ent was based on the period 1956- 1979 wh ich w as typical of the reco rds used from the south .

A comparison of daily with month ly base d stora ge yie ld relationsh ips for the Anna long record show ed that the month ly analysis u nderestima tes the

50 year return pe riod sto rage requ irement for amounts ranging from 1% - 7% o f the annual runnoff vo lume , for yields o f 20 % and 95% of the average flow . Th is arises because daily data give a mo re accurate and larger estima te of w ith in month storage requirements . Th is adjustm ent w as made to the fina l storage yield relationsh ip for each of the month ly records . For each storage freque ncy cu rve the storage requ ired for 2% of the years w as est ima ted for each yield . This was also carried out for each of the individual long period records and the results are listed in Table 1.13 and p lotted on Figu re 1.8 . The figu re illustrates that there is good agreement betw een each of three small up land catchments ; Altnaheglish , Woodbu rn and An nalong and with the pooled curve from North ern Ire land (except at the lower y ields ). The Republic o f Ireland pooled curve p lots be low this group as does the Lou gh Erne analysis . The difference between the Lo ugh Erne and

Republic o f Ireland curv es compared w ith all th e No rthern Ireland cu rves may be due to :-

(1) Di ffere nces in flow regim e caused by difference in catchm ent soil,

physiography , annual average or se asonal distribution of ra infa ll . In co nsidering the Lough Erne resu lts it should also be noted that the ca tchm ent is very large (4349 km2 ), w ith an exten sive area of

Lough , and althou gh useful for assessing the frequency of extrem e ev en ts it is not rep resentative of the sma ll reservoired ca tchm ents of Northern Ireland .

-23- St o r a g e % A R V

SENSITIVIT Y OF STORA GE / YIELD TO PERIOD OF RECORD - ANNALONG Figure 1.7 MN a • OM I IIII M OM a a ONI IMI ON ONI O IS MIO NIN N M UM

T ABLE 1.13

STO RA GE RE Q U IRE MENT FO R 50 YEAR RET U RN PE R IOD OF FA ILU RE

1 Ad jus ted fo r da ily data 2 M ean of 5 f low records ad ju sted fo r sho rt length of re co rd 3 M ean o f 6 f low records a d ju sted for sho rt le n gth of re cord 4 M ean of 9 sta t io ns

A l l sto rage req u irem en ts exp ressed as a pe rcen ta ge of annua l runo ff vo lume . 0 10 20 30 4 0 5 0 6 0 7 0 8 0 9 0 10 0 S t o r a g e ARV

I .

1 5 0 YEAR RETURN PERIOD STORA GE / YIELD RELATIONSHIPS Figur e 1.8 (2 ) Diffe rences caused by the incidence and distribution o f ex treme

drou ghts , the erro rs in ad justing frequency cu rves based on short reco rds and in estimat ing storage requ irements from the invidual

station sto rage frequency p lots .

(3 ) Erro rs in the regression mode ls leading to an ov er est imate of low flows . The c lose agreem ent between the gauged Erne re co rd and the Anna lon g and Woodburn record s in ranking historica l events suggest

that is not a se rious problem . Fu rthermore , most of the notable

droughts are gau ged in the Anna lon g reco rd and thu s an y errors in the

reconstructed flows will have a m inimal influence on the sto rage

y ield re lationship .

In view of the long lengths of reco rd , simi larity of catchment type w ith other reservo irs in Northern Irela nd and consiste ncy of resu lts , the sto rage yield relationsh ips for the Altnehe glish , Anna long and Woodburn catchm ents are considered to be the mo st approp ria te for y ield estimat ion .

The difference between the three plots is small and a compo site curve based on . the largest of the three storage volumes for each y ield (Wo odburn for y ie lds less than 70% ADF and An nalong for highe r yie ld s ) has been used as the reg ional design cu rve for Northern Ireland . For the reservo irs with sto rage oE less tha n 5% A RV th e yie ld will be mo re dep endent on the cha racteristics of indiv idual catchm ents and the error in yield est imation is correspo nd ingly higher . The above app roach was repe ated to estimate the 20 y ear retu rn period sto rage yie ld relationsh ip : the two final cu rves are shown in Figu re 1.9 .

Figu re 1.10 compa res the regiona l design curve with the follow ing genera lised storage y ield relationsh ips .

(1) South West Scotland - based on the above ana lysis for a retu rn pe rio d

o f 50 years .

(2 ) No rth West Water Authority (NWWA , 198 1) - base d on the mo st

co nservative design from a m inimum ru noff ana lysis of four long flow reco rds . Retu rn pe riod is 50 yea rs .

-25- 10 0

9 0 Ma x i m um T r 20 1 : 5 0 De s ign Yi e ld

8 0

6 0

5 0

3 0

2 0

oo 10 2 0 3 0 4 0 5 0 6 0 70 8 0 9 0 10 0 S t o r a g e % A R V

REGIONA L DES IGN CURVES FOR NORTHERN IRELAND Figur e 1.9 COMPARIS ON OF REGIONAL ST ORAGE / YIELD RE LAT IO NSHIPS Figure 1. 10 (3) Lapworth curve - based on an annua l runoff sim ilar to th e Silent Va lley rese rvoir of 56" per annum and a fter adjusting for an increase

in yie ld by a factor of 1.15 . Th is adjustment accords w ith cu rrent design p ractice in Northern Irela nd .

Figure 1. 10 illustrates that for a given storage the y ie ld is highe r in N orth ern Ireland than in Sou th West Scotland and North West England . The re lationsh ip between these three regional curves is suppo rted by a recent stu dy of the coe fficient of variation of annual ra infall over Europe

(Tabo ny , 1982 ). Th is study wa S based on 185 raingauge records in Europe using data from 186 1- 1970 and over 2000 raingau ge reco rds in the UK dating from 1911. Th e resu lts ,indicated that the year to year variab ility of rainfa ll is mu ch lower over Northern Ireland than Great Britain and most of the Republic of Ire land . The coefficient of va riation varies from 11.5% in Northern Ireland to 13% in Sou th West Scotland and 14% in North West En gland .

Figu re 1. 10 shows that the Lapwo rth curve generally overestima te s yie ld compared w ith the Regiona l De s ign Cu rve . However the position of the Lapworth curve on Figure 1.10 depends on the annual runoff of the reservoired catchment , and for a site w ith a much lower runoff than Silent

Valley the Lapworth curve would be below the regional de s i gn curve for most y ie lds .

A fea tu re of the NWWA stu dy (NWWA 198 1) is that a max imum yie ld of ' 76% ADF, corresponding to a 5 year refill period , is set on all sources . Refill pe riods asso ciated with yields of 8 5% A DF for this study are in excess of 5 years and for yields o f 90% ADF they a re in excess of 8 years .

In view of these ve ry long refil l pe rio ds and the errors associated with estimating the average flow of each reservoired catchment , the m ax imum yield availab le for any sou rce has been set at 90% A DF for a return pe riod of failure o f 50 y ears and 90.5% ADF for a retu rn period of failure of 20 years . For such sources the scope for increa sing the yield by operating the source con junctively is obv ious ly lim ited .

-26- 1.6 YIELD EST IMAT ION

(a ) Re ional Stora e Yie ld Ana lysis

In order to app ly the regiona l storage yie ld relationsh ip to a

part icu lar reservoir , the av era ge flow of its catchment must be

ca lcu lated . The mean altitude of each rese rvo ired catchment was us ed to ma ke an altitude adjustm ent to the mean sea level va lue of pp for

Northern Ire land . Section 1.3 show ed that the difference between Ep and Ea w ere min imal and he nce the ave rage 'runo ff in mm cou ld be ca lcu lated from the difference betw een SAAR and Ep (Tab le 1.14 ).

Publish ed valu es (D.O .E. N .Ireland , 1980) of direct and indirect

catchm ent areas (Tab le 1.14 ) were us ed . W ith in the scope of the p re sent study it was not poss ib le to estim ate the efficienc ies of 0,tak .d.: ind ividua l catchwate rs and so their effective area was genera lly assumed to be 80 % of their actual area , unless m ore de ta iled ivd/ u 0 - t co,i-cseve-t e" in fo rma tion was availab le . Th is va lue was based on previous desi gn pfcas j p ractice in Northern Ire land . (In the case of the Annalong tu nnel

diversion to S ilent Valley rese rvoir , an ana lysis of the tunnel flows (.11,44- ind icated that a catchwater efficiency of 100% cou ld be used ). Tab le & e a- 1.14 also shows the ind irect area as a percentage of the tota l area -

re servo irs with a high percent age may warrant a more detailed inve stigation of their catchwaters . Th e total effective catchmen t

area was ca lculated wh ich ; to gether w ith the annual runo ff in mm , enab led the average daily flow (ADF ) to be calculated in Ml/d .

Pu blish ed values of usable volume were exp ressed as a pe rcenta ge of

the annua l runoff volume (ARV a ADF x 36 5 .25 ) from which the yield (% AD F ) cou ld be est imated from Figure 1.9 . For some catchments w ith mo re than one rese rvo ir the tota l vo lum e of storage has been used and treated as if it were one source . For yield calcu lation th is

• implicit ly assumes tha t the low er rese rvoir w ill only spill wh en all up stream reservoirs are fu ll. Any depa rtu re fram this w ill result in

a reduction in reservoir yie ld . Tab le 1.14 list s these yield s for a retu rn period of 50 and 20 yea rs exp re ssed in un its of Ml/d . From

these gross yie lds the compe nsation flow must be deducted to produce the net y ield . These are listed where pu blish ed comie nsation flow s

are available .

-27- I MI • M I a • M O a • • • • ONI • OM • • OM a OM •

TABLE 1.14

YIELDS OF RESERVO IRS IN NORT HE RN IRELAND

Source Ca tchment Ca tchment Ar ea (Km 2 ) Usable AD F G ross Yie ld Average Vo lume M l/d (M l/d ) Direct Indirect Total IA as SAAR Altitude Ep DA IA DA+0 .8IA % of M I % of AR V T=20 T=50 (mm ) (m ) (mm ) DA+ IA

(1) Sto rages of six rese rvoirs have been comb ined . Mo re accurate yie ld estima tes of 47 .57 m l/d and 45 .89 M l/d are giv en by con sidering separate re se rvoirs (2 ) Storage of Ben Crom and Silen t Valley reservoirs have be en comb ined . Catchw ater efficiency of 1.0 assumed fo r Tu nn el divers ion . No allowance for y ie ld of Mou rne Cond uit intake from River Annalong . Simu lation yie lds are 117 .1 M l/d and 112.8 m l/d ad justed for Daily data . (3) No allow ance fo r Pre scribed flow (PF ) of 0 .35 M l/d . (4 ) No al lowan ce fo r PF of 0 .25 M l/d . (5) Stora ge of Uppe r and Lower Con lig are comb ined . (6 ) Rese rvoir not currently in use . (7 ) No a llow ance fo r PF o f 0 .05 M ild . (8 ) Indirect ar ea is not known . M I S IM • • M I IW O UM

'TABLE 1.14 (CONT )

2 Source Catchment Catchment Area (Km ) Usab le AD F Gross Yield Average Volume. Ml /d (Ml/d ) Direct Ind irec t Total IA as SAAR A ltitude Ep DA IA DA +0 .81A It of M I % of ARV T=20 T= 50 (mm ) (m ) (mm ) DA+IA

(1) No allow ance for PF o f 0 .0 10 m l/d . (2 ) Very low sto ra ge :- yie ld will depe nd on river flow characteristic s . (3 ) Excludes area of Upp er Quolie . (4 ) Y ield exclude s pumpe d storage from Bann . (5 ) No allowance for compensation di scharge o f 3 .21 M l/d . (6 ) Inclu des catchment area and stora ge vo lum e of Ki llyla ne rese rvoir . Usab le stora ge o f 436 00 Mi u sed for G lenwhirry reservoir , Schem e II , stage 2. No allow nace made for proposed compensa tion flow 17 .5 M l/d . The catchment area from the consu ltants ' repo rt has been us ed but a rev ised SAAR has been eStima ted . a l la i 11. 11 - . 111

TABLE 1.14 (CONT )

Sour ce Catchm ent Cittchment Area (Km 2 ) Usab le AD F Gross Yie ld Average Vo lume M I/d (81/d ) Direct Indirect Total IA as SAAR Al titu de Pp DA IA DA +0.8IA % of 81 it of ARV T- 20 T= 50 (mm ) (m ) (mm ) DA+IA

(1) Excludes PF o f 1.14 M I/d . (2 ) Excludes Upper Al tmore . Very low storage :- y ie ld will depend on river flow characte ristics . (3 ) Excludes max imum PF of 9 .30 M l/d . (4 ) Excludes PF of 2 .27 M Ild and four annua l flushes of 45 .46 M l. (5) Ex cludes PF o f 9 .09 M I/d . (6 ) Catchmen t area and SAAR values from consultant rep orts have been used - in some cases these differ from those quoted in N .I . Wate r Statistics 1980 . (7 ) Excludes Fofanny rese rvoir and propo sed compe nsation flow of 7 .1 MI/d . Propo sed effective areas of ind irect catOhm ents used . (8 ) Excludes Epelga re serv oir and pro po sed compensation flow of 4 .06 M I/a . Proposed effect ive areas of ind irect catchm ents used . (9 ) Based on tota l usable sto ra ge , exclude s proposed compensation flow of up to 11.4 Mild . • OM IS OM IM IS 11. / in a in a OM

TABLE 1.14 (CONT i

Source Catchment Catchm ent Area (Km 2 ) • Usab le ADF Gross Yield Ave ra ge Volume M I/d (M l/d ) Direct Indirect Total IA as SAAR Altitude Ep DA IA DA +0 .8 IA 8 of M l % of ARV T=20 T=50 (mm ) (m ) (mm ) DA+IA

Weste rn Div ision :-

Al tnaheglish (1) 1489 366 330 7 .28 - 7 .28 1759 20 .8 23. 10 15 .96 15 .06 Ki llea 1234 20 1 380 1.68 - 1.68 76 5.3 3.93 1.38 1.14 Cre ggan 1210 131 400 2 .72 - 2 .72 - 582 26 .4 6 .03 4 .64 4 .34 Lo ugh Fingrean (2 ) 1250 192 385 1.13 1.81 2 .94 62 8 10 3 1.9 6 .96 5 .70 5.36 Lou gh Macro ry (3 ) 1100 183 385 2 .41 9 .47 9 .70 80 209 3.0 18 .99 4 .75 3 .70 Lough Bradan (4 ) 1408 2 10 380 2 .88 6 .52 8 .10 69 626 7 .5 22 .80 9 .63 8 .44 Glenco rd ia l (5 )) 1250 305 350 5.56 3 .50 8 .36 39 64 0.8 20.60 1.6 1 1.11 Ki lly fole 1075 150 395 4 .50 3 .48 7 .28 44 7 19 14 .5 13 .55 7 .89 7 .32 Ballydoo lagh Lou gh 110 1 143 400 1.70 2 .70 3 .86 61 6 18 22 .8 7 .4 1 5 .35 5 .02 (1) Exc lude Glenedra inta ke . (2 ) Revised direct (1.13 Km2 ) and indirect (1.8 1 Km 2 )-areas . Ca tchw ater efficiency of 1.0 used . 2 2 (3) Excludes Lou gh Fingrean . Rev ised direct (2 .41 Km ) and indirect (9.47 Km ) areas . Catchwater efficiency of (b ) Simu lation of Silent Valle Reservo ir - Fixed Y ie ld

The objectives of these additional ana lyses are to prov ide an indepe ndent check o n the regiona l analysis and , in the follow ing

section , to inv est igate the effect of rationing water supp lies on the frequ ency of fa ilure . The me thod adopted for the yield evaluation is that attributed to Go uld (McMahon and Me in , 1978 ). Th is method

requires tha t the reservoir is divided into several (N ) states of

equ al storage . Each year of the inflow data I s treated separately and is routed th rou gh the reservoir , on a month ly basis , starting th e

reservoir in each of th e N states and not ing the state in wh ich it fin ishes . Wh en this procedure has been repe ated for each year of

data the resu lts are collated in a transit ion ma trix. Th is expresses

the probab ility of ending in any of the N states , cond itiona l on the starting state . A tally of the failures wh ich occur is also kept and . a combinat ion of these two matrices enab les the probab ility of

failure to be ca lcu lated . Th is method w ill prov ide a more reliab le yield estimate tha n the regiona l analys is when it is ba sed on a long flow sequence at the site .

The frequency of failure is defined in the same way as in the regional study , that is the propo rt ion of years containing a total

reservoir failu re , the reciproc al of wh ich is th e return pe riod of

failure T Yields w ith a T of 20 and 5 0 yea rs are est imated F. for compa riso n with the regional ana lys is . Discharge from the Anna long f low record from 1895- 1979 were transferred to Silent valley

inflows by al low ing for differences in catchmen t area and effective rainfa ll . The Silent Valley system inclu des two reserv oirs Silent

Valley and Ben Crom . The ope ration of these two reservoirs involves filling Ben Crom be fo re sp illing any water to Silent Valley , and so

the two reservo irs can be lumped together and treated as one sto rage unit of 20634 M1 capacity .

Having estab lish ed the ana lys is technique , inf low sequen ce and

reservoir characterist ics , a series of re se rvo ir simulations were carried out with differen t yie lds . FigUre 1.11 shows the relationship betw een yield and retu rn pe riod from wh ich the 20 and 50 year TF yie lds of 119 .6 and 115.3 M l/d can be re ad . Th e simu lation

-32- On In On e I n a a M U M In a a a a IM O 11. 1 Probability of failure I proportion of years with a failure 1

RETURN PERIOD lyean I

100 50 20

•

• was based on month ly data wh ich would be expecte d to overestima te the

yield by about 2 .5 mil d (2% ADF ) compa red with the daily based regiona l resu lts . Th ese yields thus sup port the correspond ing yield s of 112 .12 an d 107 .67 Ml /d produced by the regional approach (Table

1.14 ).

(c ) Simu lation of Silent Valle Reservoir-W ith Rationin

The advantage of the simp le defin ition of fa ilu re (i .e . reservoir

empties once in 50 years with a constant yield ) is that it can readily be used for the assessment of a la rger number of sources with

different yie lds , sto ra ge and runo ff cha racter istics. How ever the me thod assumes that fu ll output is ma inta ined from a reservior until

it is emp ty . In prac tice output is of cour se reduced be fore the rese rvo ir be comes emp ty because one cannot be certain th at a pa rticu lar drought will not develop into one that is so re severe tha n

the design standa rd . Inevitably restriction will be introduced

which , after the even t , w ill appea r unnecessary , result ing in the frequ ency of conse rvation measures (rota cuts , standpipe s ) be ing

greater than the quoted frequ ency of failure as de fined by the above simp le criterion . Fu rthermo re , total failure rate will be less than the design failure rate .

An analysis w as carr ied out to determ ine :-

(a ) the inc rease in yield by introducing rationing with a given retu rn perio d T R and given retu rn period of total fa ilure

F .

(b ) the effect of rationing on the frequency of total failure T F

The level of sto rage (504 , 40% , 33% , 25% and 10% ) at which 'rationing

is instigated w as fixed for each simu lation to p rovide a range of

possible ration in g schem es . A reduction in yield of 20% was assumed for all rationin g periods .

- 33- Figure 1.12 summa rises the results of a numbe r of simu lations and shows the relationsh ip betw een the y ield, T R and TF . It is made up of five series of plots each de rived for a di fferent sto rage th reshold . Each se ries ind icates an inc reasing y ield as the p robability of ration ing T R increases with a cOrresponding increase in retu rn pe riod of to tal fa ilure TF . Contours of TF have been supe rimpo sed on Figure 1.12 an d the contour TF = 50 can be follow ed to show that a y ield o f 117 .5 Ml/d can be met by the Silent V alley re servoirs with a T o f 20 years . Th is is higher th an the yie ld of

115 .3 Mi/d shown on Figure 1.11 (section b ) with the same TF of 50 years but w ith no rationing .

The analysis can also be used to investigate the relationsh ip between the retu rn period of total failure T F with and w ithout rationing , but with the sam e yie ld . Fo r examp le the yield for TF = 50 and w ith no rationing is 115.3 Mild . From Figure 1. 12 it can be seen that with this same yie ld of 115.3 M l/d and assum ing that rationing is introduced when the reservo ir is 25% fu ll , then the retu rn period of rationing wou ld be 20 years ; whilst the retu rn period of tota l failu re is mu ch hi gher at 200 years . The figu re can thus be used to provide guidance on the sensitivity of yie ld , freg ency of rationing and frequency of failure to the in trodu ction of simp le rationing scheme s; which are dep en dent on rese rvo ir contents .

Th is approach cou ld be adop ted for other reservoirs w ith d ifferent storage characteristics wh ich may have a different sensitivity of yield to rationing . Figu re 1. 12 cou ld be extended to assess the effect of conjunctive use of this source . The probability of rationing is then the p robab ility of using the other source and the quantity of wat er required to make up the dema nd is contro lled by the reduction in y ield impo sed by ration ing . In this case a reduction of 20% was us ed bu t any other value cou ld easily be incorporated . Th is type of ana lysis cou ld also be carried out for the Woodburn com plex so tha t the benefits of con junctive uae of the two largest sources in Northern Ireland could be assessed .

-34- 7 I rf .SO T . 100 I • ...- --- - Tf . 200 .., , --- • ..-- - Tt • 10 5 • ---- / • ..... I / • / .....( • y --- Ti : 50 0 / • z e< t...„ r 7 o °Z l l / . ,/ ••-- • / - 0 j i 11/• Ti t 100e i • 3 / / I. " iff 7 0. , e z „ x / 6 6 x // 7 o • , / / • / 0 , 1 i s / x I 2 • / / / ; It til • / fk . 0 1 V . C / / • • / 0* (7 /

111 5 . / 4 / / 8 / • / / / / I / / • 0 / • at / / i , 7 ci 0 i t • / ci • 0 / -= • / / -0 7 I 2) ,/ ,i 0 • , , 7, , • / o- ii. / w 0 5 / / . • , / • / / / I c / / a / / x/ /

/ ./ / 1/ . / / / 1 x 1 • SO S S TORAGE THRE SHO LD .0 3 / / / / / ii / X / • 4 0 % / i • 3 3 % .0 2 / x / I • 2 5 % / x x 10 % / I

x / / 1 x / 0 1 I 110 115 120 125 130 Yie ld ( MI / d ay

RETURN P ERIOD O F RATIONING F OR GIVEN YIELDS AND P RO BABILITY OF TO TAL F AILURE Figure 1. 12 1.7 RIVE R ABSTRACT IONS

(a ) Bac k g r o u n d

A numbe r of sma ll abstractions with yields less than 2 M Ild

have been identified . It is consid ered uneconom ic to up grade

ma ny of these schemes and furthermore for many ab stractions , large and sma ll , the flow characteristics of the river do not

provide a const raint to ab st ractio n . Of the rema inder :-

• (i ) The Bann for Ba llinrees . Th e y ield is lim ited by the capacity of the intake , but abstraction ma y be

constrained by down stream re sidual flows . How ever in

view of the sma ll abst ract ion rate , relative to the flow

in the Lower Ba nn , th is seems unlik ely .

(ii ) The River Dou glas . Ow ing to the poor qua lity of the

wate r , treatme nt wou ld be ne cessary if it were to be further developed . Howeve r the quantity is too sma ll to • justify this economica lly .

(iii ) Altnaheglish . Othe r consu lt ants are workin g on a

detailed study of the merits of usin g the A ltnaheglish river and reservoir with the G le nedra river (and

po ssible reserv oir ) in a num ber of d ifferent comb inations . Detailed conjuctive use studi es wou ld be

ne cessary and can be carried out if requested .

(iv ) Fau ghan Riv er . Binn ie 's repo rt of Decembe r 1969 put this •

into pe rspective and there i B very little further to

add , until rese rvoir sites have be en more fully studied .

(v ) The Tievenny (113248 59 ) abstraction on the River Derg is th e on ly river source identified that requires yield

estima t ion , the ca lcu lations are described be low :-

-35- The yield of this source was estimated by calculatin g the 95 .

pe rce nt ile from the 10 day flow du ration cu rve , Q95 (10) and the 20 year and 50 year return pe riod 10 day annual minimum flow s .

The calcu lat ion s were ba sed on the techn ique s described in the

Low Flow Study Report (Inst itute of Hydro logy 1980 ) incorp orating the resu lts from an analysis of local flow data

from the gau ging station 5 k ilom ete rs up stream at Cast lederg (Station 20 1008, Figure 1.1) and from Lough Erne inflows .

Est imates we re made for the Cast lederg site and then transferred to the abstraction point .

(b ) Flow Du ration Curve

Using the Low Flow Stu dy Repo rt Numbe r 2.1 (LFSR 2.1) a value o f Q95 (10 ) o f 7 .4% ADF was calcu lated from the ob se rved Base

Flow Index and catchment SAA R (Tab le 1.1). This va lue com pares favourably w ith the value o f 6 .5% ADF from the short flow record from 1979- 1980 at Cast lederg .

(c ) Annual M inima

Using LFSR 2.2 and the sam e va lues of HF I and SAAR the 20 and 50 y ear retu rn period annual m inima were estima ted as 2.1% ADF

and 1.5% ADF respectively . An ana lysis of the flow s in May 1980, the month w ith the low est discharge in the Castlederg

record , enab led additiona l est imates to be made . Insp ection of

the long Lough Erne inflow reco rd (1900- 1983 ) show ed that this

month wa s the annual mi nimum discharge with a retu rn period Of app roximately 5 years . Al though the flow regime of th e Lcnigh

Erne inflows would be diffe rent from the sma ller Derg catchment , it is p rob ab le th at the frequ ency of drou ght events

on the two catchm ents would be sim ilar . It was thus assumed that Ma y 198 0 was also the 5 yea r 1 mon th , annual m inimum on

the De rg . Using multip lying facto rs from LFSR 2.2 app ropriate to the Derg ca tchment , it was possible to estimate the 1 m onth , 20 and 50 year return pe riod annua l minima from this 5 year

annual minimum . Fu rthermo re , relationsh ips betw een mo nth ly and

daily flow statistics (LFSR 2.2 ) enabled the 10 day ann ua l

mi nima to be calcu lated . Res u l t s from LFSR 2.2 indicated that

-36- the 10 day annual min im a wo u ld be 54% of the 1 month annual

m inima . The ratio of the lowest 10 day average flow to the

mean flow in May 198 0 has a sim ilar va lue of 50% . Using this m on th ly to daily adjustment o f 50% the 20 and 50 year return

period annua l mi nima we re est imated as 0 .77% ADF and 0 .56% ADF

respectively . These va lues are lower than tho se estimated from SRI and SAA R but as they are de rived from a greater use of

local data , they are the p re ferred values for yield

ca lcu lation .

(d ) T ievenny Abst raction Po int

Th e abov e sta tistics are exp ressed as a % A DF and we re

converted to M l/d using an est ima te of av erage discharge at the abstraction po int . Th is was affected by adjusting the 1979-80 gau ged average discharge at Cast lederg to the m ean 1941-70

discharge ba sed on the ratio of mean effective ra infa ll for the

two periods . Th is d ischarge was increased by a factor o f 1. 11 (to allow for the large r catchm ent area to the abst raction

po int ) to give an ADF o f 1237 Ml /d and used to convert the flow sta tistics to un its o f M Ild as shown in Tab le 1.15. Th e 20 and

50 year retu rn period annual minima est ima ted using LFSR are higher than the obse rved discha rge in May 1980 an d are considered to be over estimates . The data ba sed est ima tes are .

there fore the recommended design yields .

1.8 SEVER ITY OF PERIOD 1970- 1983 \ -

A lthough it is beyo nd the scope of this stu dy to assess the

retu rn period of individual drought events , some general resu lts concernin g the sever ity o f riv er flow and rain fall in the p eriod

1970-198 3 may assist in assessing the frequency of water re sou rce

shortages in recen t years . For each of the long flow sequences the date and volume of the max imum de ficit (for a given y ield ) in this

period was noted . Table 1.16 shows the date and rank of these max ima derived from compa riso ns w ith the deficits from earlier records . This tab le illustrate s that for the higher yields the period conta ined sone of the worst droughts in nearly 100 yea rs of reco rds .

-37- TAB LE 1.15

YIELD OF RIVER DERG ABSTRACT ION AT T IEVENNY

Ra nk shown in bracke ts relate s to long flow reco rd i.e. rank (1) for

An nalong is worst event from 1895- 1979 .

- A second app ro ach was to app ly a depth duration frequency

ana lysis (Tabon y 1977 ) to the Arma gh ra infa ll reco rd . For each

du rat ion , determ ined from the resu lts o f the reservoir simulation for

Loch Erne inflows , the retu rn pe riod of the ra in fa ll for a fixed

starting month wa s estimated (Table 1.17 ). These resu lts suppo rt the

ana lysis of the flow da ta w ith retu rn periods in excess o f 100 years

and some in excess of 500 yea rs be ing att rib uted to the rain fall

ev ents .

The mo st recent flow data currently availab le is for 'the period

up to Sep tembe r 198 3 for the Lo ugh Erne inflow se ries . Th is su ggests

that extreme low flows have also occurred in 1983. Ju ly 1983 is the lowest av erage mont h ly flow on reco rd and the period Ju ly/Au gust 198 3

is the lowest 2 month average flow on record , equalling the avera ge discharge o f June/Ju ly 1976 . The genera l conclusion is th at the

recent period has inc luded som e notab le drought events of both long and short du ration w ith retu rn pe riods equal to or in excess o f 100 years .

TABLE 1.17

RETU RN PERIOD OF ARMAGH RAINFALL FOR SELECT ED DU RAT IONS

+ RESERVOIR Y IELD FROM START OF RESERVOIR FROM STA RT OF RESERVO IR DE PLET ION , TO MAX IMUM DE PLET ION TO FULL DE PLET ION

eADF RETU RN PERIOD PERIOD ' RET URN PERIOD PER IOD

Start of dep letion is lagged by one month for rainfa ll return period ba sed on Lough Erne reservo ir simu lation .

-39- 1.9 LA UGH NEAGH

Th e water qu ality and quan tity aspe cts of the wate r resource s •

o f Lough Nea gh were con sidered in deta il by th e Lough Neagh Working

Group (197 1). Hydro logical aspects of the study were based on a reco rd from 1937- 197 1 of inflows into th e Lower Bann catchm ent upstream of Movanagher weir , estima ted from daily flow records at the we ir and changes in sto rage in Lough Neagh . This exe rcise could now

be updated us in g flow and level reco rds up to 1983 and could be

supp lem ented by a more detailed simu lat ion of yie ld draw down re lationsh ips . Howeve'r the 197 1 investigation highlighte d the

difficu lties of measuring lake lev els and th ese errors w ill rem ain in

any further hydro logical analysis . Furthermore there is ev idence

that the frequency ana lysis used resu lted in an underestim ation of the probability of occu rrence of cumu lative inflow s - th at is design

drawdow ns for a given dem and w ill recur less frequently than pred icted . Se ct ion 12.9 o f the repo rt conc luded "Amp le supp lies of water are availab le from Lough N eagh and th e Lower River Bann Basins

to meet all types of dema nd for the foreseeab le future . Large

quantities of water , up to about 450 Ml /d (100 mgd ) may be perma nently expo rted from the 2 Basins with ve ry litt le adverse

effect on other interests . Invest igations shou ld , however, be carr ied out over the next decade to qu an tify the effects on any in terest s resu lting from th ese exports" .

Although it is not possib le to oomment on the 'adverse effects ', it is considered that in re lation to the magn itude of the po tent ial yie ld a further hydro logical study wo uld produce little change to th e

above conclusion . A lthough outside the scope of the current study it ma y l e appropria te to review further water qua lity , fisheries or

gene ral env ironmen tal studies wh ich have been comp leted since 197 1.

In this contex t the severe droughts of the period 197 1-1976 may have prov ided fu rther evidence to .assess the env ironm ental impact of increased ab stractions . In this regard the general assessment of the

frequ ency of recent droughts (Section 1.8 ) may be of value in app ra isin g the Lo ugh Neagh schem e .

-4 0- 1. 10 GROUN DWATER

The groundwate r resources of No rthern Ire land ha ve been

considered by rev iew ing published ma terial and by discu ssion w ith the

Geo logical Su rvey of Northern Ire land . Th e comp lex solid and drift

geo logy of the prov ince has resu lted in a ll the aq uifers be ing small

in area and in yie ld . However the re has been considerable recent

groundw ater deve lopm ent w ith pumping capac ity increasing from 9 M l/d

in 1964 , to 69 M l/d in 198 0 and to app rox imately 100 M l/d in 1983 .

Th is rep resents 15% of the total pu blic supp ly but is not wholly

utilised .

Various aqu ife rs of Carbonife rous age wh ich occu r in the we st

of the p rovince give indiv idual well yields of u p to 3.5 Ml/d but relatively litt le development has taken place becaus e of severa l

factors , in cluding low demand , ava ilability of surface water and qua lity p rob lems from the chem istry of the groundwate r . Prod uctive so lid rock aq uifers inc lude th e Perm ian and Triass ic

sandstones , in the Lagan Valley and to the west of Lough Neagh with

recent ly comm is sioned bo rew ells yield ing a total o f 10 Ml/d . The Cre taceou s chalk is much thinner and of lower pe rmeability than its En glish counterpa rt and its value as a resource is restricted

p rima rily to the number of springs wh ic h issue from its base where it over lies impe rm eable st rata . Al thou gh th ere may be some scope for

developmen t of the Tert ia ry Basalts , this y ou ld be confined to the dev elopm ent of local bo reholes w ith y ie lds genera lly less than 1

Ml /d .

With the excep tion of up land areas the solid geo logy is cov ered by a vene er of dr ift deposits . Where this drift is compo sed of

boulder clay , recha rge to underly ing aqu ifers will be restricted . Bennett (1978 ) lists the ev idence for the very .low pe rmeab ility of th e boulder clay and this is sub stantia ted by the low values of the Ba se Flow Index (Tab le 1.14 ) for catchm ents w ith a high p rop ort ion Of

bou lder clay cover . In contrast , where the drift consists of fluv io glacia l sands and gravel (in th e Lagan valley and in the north and west of the region ), the sup erficia l depo sits prov ide a local

groundw ater re source . Indiv idual borew ell y ie lds are up to 4 .5 M l/d and one group of three we lls produCes 10 M l/d .

-4 1- In conclusion local dema nds pa rticula rly in the north and west

of the p rov ince ma y be met by further groundwater dev elopment from

either solid or superficial aquifers . However groundwater

development will not ma ke a significa nt ad dit ional contr ibution to

dema nds in the Belfast area .

1.11 CONCLUSIONS AND RECOMM ENDAT IONS

A regional storage yie ld relationsh ip has been deriv ed for

Northern Ire land and used to estima te the yie ld of a numbe r of reservoirs . The metho d is based prima rily on three long flow

records , A ltnaheglish , Woodburn and Annalong , which show consiste nt sto rage yield relationships . Furthe rmo re the curve is suppo rted by a regional storage yie ld re lationship derived indepe nden tly for th e NWWA area and from a curv e derived in this stu dy fo r South Wes t Scotla nd .

Erro rs in yield estima tion may arise from three sources . The

first is the extent to which individual ca tchm ent sto ra ge y ie ld relationship s depart from the design curve . This departure will be

propo rtiona te ly sma ll at high y ields , wh ich are co ntrolled by the annual variability of rainfa ll , but will be larger f6 r small yield s

and sto rages where the low flow characteristic of the river Wi l l assume grea te r impo rtance . The seco nd source of error is in estimat ing the av era ge d isc harge of the reservo ired ca tchment . This is dep enden t on the accuracy of pub lished va lues of catchment area and ra infa ll , to gether with the accuracy of estima ting losses . The

latter will again be dependent on the chara cterist ic s of indiv idua l catchments . For example , no allowance has be en ma de for catchments

wh ich are heav ily forested an d where the actual los ses may be hi gher

than estima ted values ; or for any groundwater leakage in the catchment . The third sou rce of erro r is in estima ting the eff iciency

of catchwaters and w ill only be aignificant for those reservoirs with large ind irect catchm en t area a .

-4 2- The estima ted yield of the Silent Valley reservoir , using the re gional app roach , has been con firmed using a fu ll simu lation of

rese rvoir behav iour but w ith a simp le de fin ition of failure . A m ore

realist ic sim ulation has also been ca rried out to exam ine the

sensitiv ity of yield to different frequencie s of rationing and tota l

fa ilure . It is recomm ended that a fu ll simulation is ca rried out for the other ma jor rese rvoirs including the Woodburn comp lex and at '

least one rese rvo ir with a relatively low storage volume . If requ ired this app roach could also be used for conjunctive use

studies .

,The main emphasis of the hydrolo gical aspe cts of the stu dy has been on rese rvoir y ield estima tion w ith only the -Fi ver De rg be ing considered for additional river abstractions . The hydrology of Lough Neagh has been reviewed and .this indicated that water quantity would no t be a constraint on future abst ractions . Finally the review of groundwater ab stract ions indicated its cu rrent impo rtance in the

Lagan va lley and for loca l supp lies elsewhere ; however its po tential for making a major contribution to meet increasing dema nds in the

Be lfast area is limited .

-43- ACKNOWLEDGEMENT S

The Inst itute of Hydrolo gy acknow ledge North West Wate r Autho rity 's pe rm ission to rep roduce the regional sto rage y ie ld re lation sh ip for no rth west Eng land .

REFEREN CES

Be nnett , J .R .P . (1978 ). Groun dwater development in Northern Ireland ,

Geo log ica l Survey of Northern Ire land , Open File Repo rt No . 59 .

Calder , I.R ., Hard in g R .J . and Ro sier P .T .W . (1983 ). An ob jective assessmento f Soil-Mo isture Deficit Mode ls , Journal of Hydrology 60 , 329- 355. Colebrook , C .F . (1955 ). Th e D iversion of the Anna long River into the Silent Valley Rese rvo ir , Works Construct ion Div is ion Meet ing , Paper No . 30 , 20 December 195 5. ICE . Depa rtmen t of the Environm ent (N .I .) (1980 ). No rthern Ireland Water

Stat istics .

Insitu te of Hydro logy , (1980 ). Low Flow Study . Lo ugh Neagh Working Group , (197 1). Adv isory Repo rt , Volume 2, Go vernment of Northern Ire land . McMa hon , T .A . and Me in , R .G . (1978 ). Reservoir Capa city and Yie ld . Developm ents in Wate r Science No . 9 . Elsev ier.

North West Water Au tho rity (198 1). Survey of Existin g Su rface Water

Sou rces .

Savill , P .S ., and Weatherup , S .R .C . (1974 ). Th e Effect of Afforestat ion on Water Ru noff in the Woo dburn Catchm ent Ar ea , Jnl . of Inst . of Fo restat ion

of Great Britain , Vo lum e 47, No . 7 . Tabony (1977 ). The variation of long du ration rain fall over Great

Br itain . Meteo ro lo gical Office Sc ien tifiC Pape r No . 37 . Tabo ny (1982). The Homogenisation and An alysis of European Rain fall

Re cords , Met 0 13 Branch Mem orandum 76 .

-44-, CHA PTER 2

HOUSE HOLD SURVEY

2 .1 F IND INGS OF THE SURVEY

The main househo ld survey conducted for this stu dy consisted of a

su rvey of a stru ctured sam p le o f 2500 addresses designed to obta in

informa tio n conce rn ing the ownership and use of wate r-u sin g facilities and

app liances by type of household .

Tables 2 .1 and 2 .2 summa rise findings from the surv ey . These statistics are simp le averages of su rvey responses and have not been reweighted to take accoun t of the samp le structu re . Tab le 3 .6 in the Ma in

Repo rt provides a comparison of N I ownersh ip of water usin g app liances and am enities w ith leve ls evident in G B .

Tab le 2 .1 shows simp le average figur es for hou sehold own ership of water using app liances and amenities togeth er with back ground informa tion .

In Tab le 2 .2 percenta ge figures for level of ownersh ip on type of app liance are given for each ty pe of househo ld, together w ith employm ent

statu s and income of householders and numb er of pe rsons pe r househo ld .

2.2 HOUSEHOLD SURVEY QUEST ION NA IRE

A rep rodu ction of the questionna ire u sed in the su rvey is prov ided In the pages follow in g Tab le 2 .2.

-45- TA BLE 2 .1

OW NE RSH IP OF W AT ER US ING APPL IANC ES AND AM EN IT IES

1 Ap p ro x im ate ra te pe r pe rso n 1.5 ba ths pe r w eek 0 .5 show ers pe r we ek

-46- TABLE 2 .2

1 HOUSEHOLD CHA RACT ER ISTICS BY TENU RE STAT US

1 Re sults have not been reweighted to take account of the samp le structure . -4 7- TABLE 2 .2 (CONT )

HOUSE HOLD C HARACT ER IST ICS BY TENURE STATUS

-48- TABLE 2 .2 (CONT )

HOUSE HOLD CHA RACTERIST ICS BY TEN URE STAT US

- Income month (of Hea d of Household )

-49- HOUSEHOLD SURVEY QUEST I ONNAI RE.

NORTHERN (HELM WATER USAGE SURVEY -

HOUSEHOLD SCHEDULE

• Househol d Ident i f i cat i on Number . HousehOl dNu er ( i f onl .y one i nt er I I )

• How many peopl e are t here i n your househol d, t hat i s, peopl e whose mai n resi dence t hi s i s anth ill° are cat er ed f or by t he same per son. as your sel f or shar e l i vi ng accommodat i On wi t h yOu1

HOUSEHOLD I NFORMATI ON MUST OE OBTAINED FROM SPOUSE OR NON.

Per . Sex I Age Fem. ' Cur r ent Empl oyment Age No Uni t Ful l - t i me St at us Fi ni shed Educat i on Work i ng Cont i nuous Pr i mary/ Fur t her / (Emp l oyee/ Sel f Regi st er ed At Ret i r ed Ful l - t i me Secondary Hi gher empl oyed) Unempl oyed Home EdUcat i on F ( Yr s) School +Educat i on F. T. P. T. ( Yr s ) s+ 01 1 2 1 2 3 4 5 . 6 • 7 L L_ 1 02 1 2 L _s _j L - 1 1 2 3 4 5 6 7 03 1 2 L - 1 1 2 3 4 5 6 7 L_S - 1 1 2 1_ 1_ 1 L- 1 1 2 3 4 5 6 7 L.S .1 OS 1 2 1 2 • 3 4 5 6 7 • _J - 71 06 1 2 L-1 1 2 3 4 5 6 +.7 1 3 4 07 1 2 L_ I _J L a 2 5 6 7 • os 1 2 1 •2 3 4 5 6' 7 09 1 2 L - L- J L- a 1 2 3 4 5 6 7 L- L - 1 10 1 2 L- 1- 2 L- 4 2 3 4 5 6 7 LS -1 11 1 2 L Å 1 2 3 4 5 6 7 L 1 12 1 2 L - 1 1 2 3 4 6 7 13 1 2 1 2 3 4 + 5 6 7 L_S _J 14. 1 2 L - 1 1 2 3 4 5 6 7 L_ L - 4 I S 1 2 L J 2 3 4 5 6 7

Tot al number of househol ds at address Tot al number of persons i n household Tot al number of adul t s ( 16+) i n househol d Interviewer c lete for all addresses

1. T e of buildin at/in which address is located

Whole house detached 1 Whole bungalow detached 2 Whole house/bungalow - semi-detached 3 -Terraced house or terraced bungalow 4 Purpose-built flat or mainsonette 5 Part of house/converted flat or maisonette/rooms in house6 Dwelling with business premises 7 Other (specify) Go to 132

2. Could you tell me about your household accommodation? When was the property built? Was it Before 1919 1 Between .1919 and 1939 2 Between 1940 and 1944 3 Between 1945 and 1964 4 After 1964 Don't know -1

I F °DON'T KE W probe for estimate. Only code °Don't know° (-1) if neither informant nor interviewer are able to give an estimate. Go to 173

3. How many bedrooms do you have? and livingrooms? and bathrooms? and kitchen? and other rooms?

Could I check, that makes a total of

IF ANY DISCREPANCY, resolve before proceeding

1M3TE: If 'bedsitter' accommodation please code main room as bedroom and write in below GO to ot

4. Could you tell me if you are — Owneroccupier 1 Renting N1HE/Other Public Authority 2 Renting Privately 3 Other (e.g. stied') 4 Don't know -1 Go to Q5

-51- S. Does this household have a garden? •11 Mil Yes 1 - (A) No 0 -06

(A) IF YES

Are you, or anyone else in this household, resOdIsible for maintainingany of this? 11 . . Yes No - 0%

(B) IF YES

Could you tell me the approximate size? length (feet) width (feet) IF •DON'T KNOW' code ' -1' for each dimension

so ot of

- 52- 6. May I aSk you some questions about some things inside the house which use water for personal or clothes washing or for cleaning?

(A) How many sinks and washbasins are there In this household? PROBE TO DISTINGUISH SINKS AND WASH BASINS - sinks mainly for 'Washing dishes Or clothes - washbasins for personal washing

How many fixed baths are there in this household?

RECORD IN GRID BELOW

Is this sink/washbasin/bath used only by your household or is it also used by another household living at this address?

RECORD IN GRID BELOW WHETHER OR HOT ANT FACILITIES ARE SHARED

MULTI-HOUSEHOLD ADDRESSES ONLY SINK(S) WASHBASIN(S) FIXED BAN S)

Used only by household 0 0 Also used by other households Go to 117

-53- e . • •• Yes 1 No Got0 09

8. In this house do you have a washing machine? Yes -(A) No -09

(A) IF tES

Is that an automatic, a twin tub or some other type?. Automatic Twin Tub 2 Other type 3 Don't know -1 Go toQ9

9. Do you have an electric dishwashing machine? Yes No 0 Go to 010

10. Do you have an electric waste disposal unit fitted into a kitchen sink? Yes 1 No 0 Go to 011

--5 4- 11. Do you have a fitted shower in this household? Yes 1 -(A) No 0 -Q12

PROBE AND EXPLAIN - Plumbed in, not just for washing hair at a sink

(A) IF YES

How Sany shower fittings do you have? Fill in number

(B) Of these how many are Inside a bath CODE ALL In a cubicle in a bathroom In a separate shower roan Other Go to 012

CHECX R OG ER ADDS TO TOTAL AT (A)

12. Nowadays, some houses have a bidet. Do you have one in this household? Yes No Go to 013

I would like to ask you some more general questions about your home

13. Do you have any form of central heating including electric storage heaters?

Full • All rooms (or almost all) heated from one central source. Partial • 2 or more rooms but not all rooms heated from one central source.

(8) What fuel does it use? 011 1 Gas 2 Electricity 3 Solid Fuel 4 Other 5 Don't know Go to Q14

-5 5- 14. Do you have a piped hot water system? Yes 1 -(A) No 0 -(C)

(A) IF YES

Which of the following method(s) is/are used to heat the water?

CODE ALL Back boiler/range Yes 1 No

Central heating boiler Yes 1 No 0

Immersion heater Yes 1 No ••••

District heating Yes 1

No 0

'Ascot' type gas heaterYes 1 No 0

(8) What is the main fuel used for water heating? Oil 1 Gas 2 Electricity 3 Solid Fuel 4 4 15 Other 5 Don't Know -1

What doyou do for hot water? electric geyser 1 electric kettle 2 heat wateron openfire range/stove 3 combination of above 4 other means(specify) 5 nothing 6 6o to 015

-56- 16. Now may I ask you sere questions about your use of water outside of the house?

As you probably know in Northern Ireland there is no licence required or extra charge for using a hosepipe or sprinkler

In this household do you use a hosepipe, for

example water the garden or when washing a car? `1• • Yes 1 No Go to Q17

17. Do you ever use a lawn or garden sprinkler? I mean one of the type that can be set on the ground and left - not attached to a hand held hosepipe •

19. Do you have a rainwater butt or rainwater collection tank? Yes 1 No 0 GO to Q20

-57- 20. Do you have any other private source of water (except for rainwater) such as a spring or well or do you get all of your water from the public mains? All from mains Private supply Go to Q21

To get a better idea of water usage,1 may ask you scale further questions?

21. ASK ONLY AT HOUSEHOLDS WITH BATH

(A) Although I know you may only be able to give me a rough answer can I ask you how many baths were taken In this household yesterday? ENTER NUMBER OR CODE ' -1' IF 'DON'T KNOW°, OR ' -2' IF MO ANSWER -(B)

(B) Roughly, how many baths are usually taken in this household at the weekend? (Friday 5 p.m. to Sundaymidnight) ENTER NUMBER OR CODE '-1' IF 'DON'T KNOW% OR '-2' IF MO ANSWER Go to Qb°2

NOTE - Exclude baths not taken in fixed bath (e.g.babies' baths). Count children's shared baths as one bath.

22. ASK ONLY AT HOUSEHOLDS W1TW SHOWER11

(A) Hcnomany showerswere taken in this household yesterday? ENTER NUMBER OR CODE ' -1' IF 'DON'T KNOW', OR ' -2' IF NO ANSWER

(B) Roughly how many showers are usually taken in this household at the weekend? (Friday 5 pm to Sunday midnight) ENTER NURSER CO CODE ' -1' IF 'DON'T KNOW°, OR ' -2' IF NO ANSWER

-58- 23. ASK ONLY IF WASHING MACHINE OWNED 0111

If respondent is not housewife, check with housewife if possible.

How many separate loads were washed in the washing machine here, yesterday?

ENTER NUMBER CR COCC ' -1'IF 'DON'T KNOW'. IF WASHING MACHINE NOT WORKING CODE '0' -(A)

(A) Thinking back over the last seven days how many separate loads can you remember washing in the washing machine during the last week?

ENTER NUMBER OR CODE ' -1 IF •DON'T KNOW'. IF WASHING MACHINE NOT WORKING ALL WEEK CODE '0'

(8) Roughly how many separate loads areusually washed here at the weekend? (Friday 5 pm to Sunday midnight)

ENTER NUDG EROR CODE ' -1' IF DON'T KNOW G° to 024

24. Could you give me some information about the head of this household?

(A) Head of Household is . employee self emp loyed 2 unemployed 3 retired 4 -(8) full time student 5 other 6 don't know

If sick but has job code as 1 or 2 as appropriate If sick and is unemployed code as 3

PLEASE COOE FOR ALL CATEGORIES IN (A)

Office Use Onl y (0 Occupation L_JL_J

Industry L_JL_J Go to 025

-59- 25. Could you tell me in confidence, taking everything into account - that is, all benefits and everyone's wages, how much comes into this household each week or month, after deductions and/or taxesand insurance?

26. Can you think of any other major water using facilities or appliances in this household which we have not covered in this survey?

INTERVIEWER PLEASE WRITE IN End Questionnaire

Thankyou verymuch for yodr help in carrying out this survey

LEAVE LEAFLET

-6o- CHA PTER 3

DETA ILS OF MET ERED D IST RICTS

The follow ing Schedu le giv es details of the metered districts which were surveyed for purpo ses of estima ting cu rrent leve ls of per cap ita co nsumption . (See Section 3 .1.2 of Main Repo rt ).

No .of DIV IS ION SUBDIV ISON TOWN ST RE ET Prop s

No rthern An trim Antr im Men in Rd . 16 Northe rn An trim Antr im Ma ssereen Ga rdens 14 Northern Antrim Antrim Muckamore Gdn .V llg .Terra ce 16 No rthern Antrim Ant rim Burlon Road 14 Northern Ba llym ena Ba llym ena Ba lly loughan Heights 19 Northern Ba llymena Ballym ena Wood land Av . & O ld Galgorm Rd . 8 Northern Ballymena Cookstow n Gorta l Low ry Park 90 Northern Ba llym ena Cooks town Sweep Rd . 3 Northe rn Sallym ena Cook stown Drum Rd . 6 No rthe rn Ballymena Cook stown Ki llymoon St . 19 Northern Ba llym ena Cookstown Fo rtv iew Terrace ,Chapel Rd . 22 Northern Co leraine Cole ra ine Ba llycranny Rd . 16 Northern Co lera ine Co le raine Jam es St . 20 Northern Coleraine Co leraine Wh eatsheaf He ights 14 Northern Colera ine Tam laght O 'Crilly Church Pa rk 20 Northe rn Co lera ine Dundaraue Bu shm il ls 84 So uthern A rma gh/Dung . A rma gh Cavancaw Park , New ry Rd . 28 Southern Arma gh/Dung . Ki llylea Old Fo rge 19 South ern Armagh/Dung . Castlecaulfie ld Drumr eaney Gardens 20 Southe rn Arma gh/Dun g . Coalisland Da rnagh Crescent 22 South ern Armagh/Dung . Du ngannon No rthland V illa ge , Killyman Rd . 46 Southe rn Armagh/Dung . Dungannon Belm nt .O ff Blvd r .Prk .Off Cunn iingh ams Ln e 24 Southern Armagh/Dung . M illtown Brook Street 4 Southe rn Cra igav on Po rtadow n Grantham Pa rk , Bridge Street 27 Southern C ra igavon Cra igavo n Ballyhannan Pa rk 49 Southe rn Cra igavon Po rtadown Russwood Pa rk 30 Southern Cra igav on Donacloney Blackskull 12 Southe rn Newry/Banbr . New ry Warren Hill , Wa rre n Point Rd . 30 Southern New ry/Banbr. Rostrevor Syanite Place 13 Southern New ry/Banb r . Seapatrick Recto ry Pa rk 12 South ern New ry/Banbr . Scarva Glenloughan Park 12 Easte rn Be lfast Be lfast Broadw ay Parade 50 Eastern Belfast Be lfast Ga wn Street 38 Eastern Belfast Be lfast Trigo Parade 30 Ea stern Belfast Belfast Chatsworth Street 42

-6 1- No .o f D IV IS ION SUBD IV ISON T OWN ST RE ET Prop s

Eastern Nth .Down/Do w np . Bangor Rosema ry Dr ive 27 Eastern N th .Down /Dow np . Bango r Ch ester Avenue 2 1 Eastern N th .Dow n/Dow np . Ba ngor C la nm o rris C lo se & A venue 20 Easte rn Nth .Do w n/Dow np . Ban go r Ma xw e ll Pa rk 2 1 Ea s te rn Nth .Dow n/Do w np . Bango r Linden Pa rk & Ga rdens 24 Easte rn Nth .Down /Dow np . Pineh ill C re scent .5 G arden s 16 Ea stern N th .Do w n/Down p . Ba llynah inch W indm il l Ga rrdens & H ill C res . 148 Eas te rn Nth .Down/Dow np . C rossga r Ki llw ood Pa rk , C res . & C lose 39 Eas tern Nth .Do w n/ Down p . Downp a t rick Meadow lands 158 Easte rn Downp a tr ic k New ca stle Schoo ne r Site , Dund rum Rd . 129 Ea stern Downp a tr ick Ki lk eel Mo ssva le Pa rk , Sc ro gg Rd . 39 Ea stern Do wn p atrick Downp a trick Spa Hou sing E state 46 Ea stern L isb urn L isburn A lanb ro oke Av enue 20 Eastern Lisbu rn Ba lly nado l ly Ki l lu lta gh Rd . G lenavy Rd . 33 Ea stern LiSbu rn Lisbu rn Be lv o ir Pa rk 14 Eastern Lisbu rn Lisburn C lon ia ra Pa rk 30 Ea stern Lisburn L isburn Co ol sara Park 27 Eastern L isbu rn L isbu rn F lo ren ce Cou rt 24 Easte rn L isbu rn L isbu rn G reenmount G ardens 19 Easte rn Lisburn Lisburn Ri chm o nd C ou rtnu e 58 We stern Ennisk illen Bal ly gaw ley Lisdda rt T e rrace 17 W estern Ennisk illen Au gh e r Knockview 16 W estern Enn is k illen Letterb reen Heathda le 8 W es tern Enn iski lle n D rumk een , Ba llinam a ll O ld Stat io n Pa rk 14 we ste rn Lo ndo nderry Lo ndo nde rry Knockw e l lam Prk . & Ca rd ilea Prk . 2 1 Weste rn Lo ndonde rry Londonderry Hin io n Park , Waterside , 17 W este rn Lo ndo nde rry Lo ndo nde rry He ron W ay , C lo on ey Est .W a tersid e 13 We ste rn Lo ndo nder ry Lo ndo nde r ry Melv in Cou rt , Lima v ady Ro ad 18 W estern Om agh Om a gh Tow nv iew Avenue So uth 30 Western Omagh Oma gh Cu lmo re G a rdens 26 Western Oma gh Oma gh Cam ow en T e rrace 6 We ster n Oma gh Om a gh Co ro natio n Cotta ges 9

-62- 1 CHA PTER 4

ADJUSTMENT OF METERED CONSUMPT ION F IGURES TO

TAKE ACCOUNT OF NON -METERED CONSUMERS

4 .1 AGR ICULTURE

Agricu ltura l consumption is current ly reco rded for three different 1 typ es of meter , 'agricultu ra l', 'intensive unit ', and 'cattle trou gh '.

Since 'agricultura l ' meters cover farmhouse consumption , as we ll as all other consump tion on the farm wh ich is not separate ly metered as 1 'intensive ' unit or 'cattle trough ', the numbe r of agricu ltural m ete rs should co rrespond c lose ly to the total numbe r' of farms .

Wate r charging records for 198 1 show 39 ,063 agricu ltural meters . Agricu ltu ra l Census reco rds for the same year identify 45 ,224 Census units , 1 includ ing 19 ,197 ho ldings wh ich we re class if ied as 'ins ign ifican t '. Th is su ggests that about 14% of a ll farm units may not be m etered . Since at least 4% o f rura l prope rties are not on the public supp ly (base d on Census figures ), it ma y be concluded that no more than 10 % of farm units draw ing wate r from public supp lies are not mete red .

1 Unm ete red farms are like ly to be of sma ller size . An alysis of agr icu ltura l me te red consumption by consumption band us ing water charging

reco rds for the first half o f 198 1 shows that usage is high ly concentrated among a small pro po rtion of very large use rs . Agricu ltura l census data on

farm size is summ arised in Tab le 4 .1 tog ether w ith figu res ob ta ined from wate r charg ing reco rds .

1 It may be assumed from this analysis that unmetered users have average consump tion app roxima te ly eq uiva lent to that of the lowest 53%, or perhap s the lowest 75% o f mete red use rs . Th is suggests that an addition of

2% or 3% respec tively to total metered farm usage (including 'catt le troughs ' and 'intensive un its ') is app rop riate to take account of non -

1 registe red farm use rs .

-6 3- M I MN I S M IM I M I MN M I M O M I M IO MII I M a M I a a 11111 • a •

TABLE 4 .1

CONCENT RAT ION OF AG RICULTU RA L 1 WAT ER USAGE

2 WAT ER CHA RGES RECORDS CENSUS FIGURE S Propo rtion of Livestock We ighted Consumption Proportion Propo rtion of Prop ortion of by Estimated Size Band of Meters 3 Consum tion Size Band ESU Censu s Units Wate r Usa e (% ) (% ) (% ) (% ) Cub ic metre s 4

1 Inc ludes on ly 'Agricu ltura l ' meters . 2 Based on Statistical Rev iew of NI Agricu lture 198 1. 3 European size un it . 4 Per approxim ate six month period . 4 .2 INDUSTRY

Mete r reco rds for 198 1 show a tota l of 2 ,745 m eters in the

'industrial ' and 'non-rateable ' catego ries . This compareS very closely

w ith the tota l of 2 ,674 m anu factu ring un its reco rded in the Census of

Emp loym ent . A numb er of large firms are known to have more than one meter ,

but the propo rt ion of indiv idua l manufacturing un its with more than one

meter is likely to be ve ry sma ll . On this basis it is reasonab le to assume that no mo re than 10% o f indu stria l users are unmetered .

Tab le 4 .2 shows the concentration of consumption am ong industrial

users ; 25% o f mete red indu strial users account for 97% of water used in this cate gory . Al so shown in Tab le 4. 2 is the high concentration of

emp loyment in a sMall propo rt ion of indu st rial firms .

1 If it is assum ed that the average consump tion of unm ete red users is sim ilar to that o f the lowest 75% o f mete red indust rial users , the addition

necessary to take a ccount o f 10% of industrial use rs being unmetered

1 am ou nts to less than 1% o f the tota l metered vo lume .

4 .3 CONSTRUCT ION

Only 1,562 (47% ) construction users we re me t e r e d in 198 1 as oompa red

w ith 3 ,3 12 Census units . Analysis of concentrat ion shows that the low est

778 o f me t e r e d const ruction users accounted for on ly 7% of consumption 1 (Tab le 4 .3). If the 53 % of users who are unm ete red are treated as equ ivalent to the lowest 77% of me t e r e d users , an addition to me tered use o f 14% is req uired .

4 .4 COMME RCE AND OTHER SERVICES

Commerce and 'other ' serv ices have been considered togethe r for the

purpo se of assessing metered usage . The class ifica tion in water charges

reco rds of oomme rc ia l and pub lic service estab lishments between these con sumption secto rs is unreliab le and it is only po ssible there fore to compa re meter numb ers with the number of cen su s un its at an aggregate

I I level. -65- M ON = 0 100 • 1111 0 1113 11• 1 11111 10 • 1111 11111 111111 • 1110 111111

TAB LE 4 .2

CONCENT RAT ION OF INDUSTR IAL WAT ER USAGE

1 WATER CHARG ES RECORDS CEN SU S FIGURE S Consumption propo rtio n Propo rtion of No of Propo rtion of Prop o rtion of Size Band of Meters Consum tion E lo ees Census Unit s Em lo ees (% ) (% ) (% ) (1 ) Cub ic metres 3

0 - 100 53 1 1-4 33 .53 2 101 - 200 10 1 5-9 18 .35 3 20 1 - 500 12 1 10- 14 11. 18 3 501 - 2500 14 6 15- 19 5 .86 2 250 1 - 5500 4 5 20-49 3 .65 2 5501 - 10000 2 6 25-49 11.06 8 1 m 1000 1 - 20000 2 7 50-99 7 .22 11 m 1 20001 - 40000 1 13 100- 149 2 .91 8 4000 1 - 80000 1 18 150- 199 1.94 8 over 80000 42 200-249 0 .85 4 250-499 2 .44 18 500-999 0 .62 10 1000- 199 9 0 .19 6 over 2000 0.19 15

Ba se : 2 ,753 Ba se : 2 ,674 (Tota l no . of meters ) (Total no . of un its ).

1 Industr ial includes the 'Industrial ' and 'Non Ra teable ' in the WCB 's classification . .2 Base d on 198 1 Census of Emp loym ent . 3 Per app rox imate six mon th p er iod . M I NM • • OM OM M O • IIIIII M O M a a a • • 11111 M O M I

TAB L E 4 .3

C ONCENT RA T ION OF CON ST RU CT ION WAT ER USA G E

W AT E R C HA RGES REC ORD S 1 C ENSUS F IGURE S Con sumptio n Pro po rt ion Propo rtio n of No of Pro po rtio n of Pr opo rtio n of Size Band of M ete rs Con sum t io n Em loyee s Censu s Un its Em loyee s (% ) (% ) (% ) (% ) Cub ic metre s 2

0 - 100 77 7 1-4 55 .22 14 10 1 - 200 10 6 5-9 2 1. 12 16 20 1 - 50 0 8 1 1 10- 14 9 .8 1 13 501 - 2500 4 .5 20 15- 19 4 .90 9 250 1 - 5500 0.5 10 20-49 2 .95 7 1 5501 - 10000 0.32 m 10 25-49 4 .14 16 -.1 1000 1 - 20000 0.19 10 1 50-99 1.06 8 20001 - 40000 0 .06 8 100- 149 0 .29 4 4000 1 - 80000 0.06 18 150- 199 0 .23 5 ov er 80000 200-249 0 .09 3 250- 499 0 .09 3 500-999 0 .03 2 1000- 1999 0 0 over 2000 0 0 Ba se : 1,560 Ba se : 3 ,312

1 Ba sed on 198 1 Census of Emp loyment 2 Pe r app rox im ate six month pe r io d . I ll• • MN • • M I • • M I OM N S SS • O M ON1

TABLE 4 .4

CO NCENTRAT ION OF SE RV ICE 1 WATER USAGE

2 WAT ER CHARG ES RECORDS CENSUS FIGU RES Consumption Propo rtion Proportion of No of Proport ion of Propo rt ion of Size Band of M eters Consum tion E lo ees Census Un its Em loyees (% ) (% ) (%) (% ) Cub ic metres 2

0 - 100 46 1 1-4 54 .11 8 101 - 200 13 2 5-9 20 .07 9 20 1 - 500 16 5 10- 14 8 .37 7 50 1 - 2500 20 2 1 15- 19 4 .48 5 250 1 - 5500 3 11 20-49 2 .73 4 1 5501 - 10000 1 8 25-49 5.80 14 m w 1000 1 - 20000 1 11 50-99 2 .55 12 1 20001 - 40000 1 9 100- 149 0 .76 7 4000 1 - 80000 1 16 150- 199 0 .31 4 ov er 80000 1 16 200-249 0.24 4 250-499 0.32 8 500-999 0 .17 8 1000- 1999 0.07 7 over 2000 0.01 3

Base : 8 ,462 Base : 24 ,129 (Total no . of mete rs ) (Tota l no . of un its )

1 Serv ice includes the 'Commercia l ' and 'Other ' cate go ries in the NCB 's classification 2 Based on 198 1 Census of Emp loyment 3 Per app rox ima ie six month pe riod . The 8 ,74 1 m eters in these two catego ries accoun ted for only 358 of the 24 ,129 Census units in 198 1. Metered consumption concentration is summarised in Table 4 .4 .

Aga in , users without mete rs can be exp ected to be sma ll concerns .

This is confirm ed by work undertaken in the cou rse of the recent study of waste levels in Belfast sub-div ision . This study identified non-domest ic consum ers without mete rs in ea ch waste-mete ring district . A small-scale rev iew and me tering exerc ise carried out by sub- div ision management has shown that a ma jority of the se consumers fall in to the commerc ial category and that consump tio n lev els are typ ically extremely sma ll .

If unmetered users are trea ted as equ iva len t: to the low est 75% of metered users , the app ropriate pe rcentage addition for the se catego ries is 6% to take accoun t of non-registration .

-69 - CHAPT ER 5

LEAKAGE IN SUP PLY SYSTEMS

5 .1 OBJ ECT IVE

Th e ob jective of th is section of the study is to estimate as direct ly as possib le the quantities of water which leak or are allowed to escape

from the supp ly system without ach iev ing a usefu l pu rpose .

5.2 BASE YEAR

The base year has been taken as 198 1 and va lues of m inium night flow

availab le for that year have been used where po ssib le . Elsew here the

init ia l va lues o f m in ium n ight flow measured in pa rticular area s , as close as possible to 198 1, have be en used .

5 .3 MET HOD

The method used to assess leakage is the tota l night flow rate metho d

recomend ed in DOE/NW C .Rep ortNo . 26 , 'Leak age Contro l Po licy and Practice ,'

as fo llow s :-

(MNF - LN F - 2 l/p /h ) x n x 20 M I/d 106 Wh ere L = Leaka ge in distribu tion system

MNF a Mi nimum night flow in lit re s/p rope rty/hour

LNF a Legitimate (metered ) night flow in litre s/p rope rty/hour 2 l/p /h = A llowance for night dom est ic us e

n = Number of p rope rties supp lied 20 - Recommended factor to convert hourly flow to da ily

flows making a llowances for higher night p re ssu res

Th e minimum and legitima te night flow rates for each sub-division have be en assessed as described be low .

-70- 5 .4 NIGHT FLOW RAT ES

5.4 .1 Gene ral

Values of minimum night flow are avai lab le , either from district

meter read ings or from waste m eter read ings , for areas representing som e

0 0 % of the Prov ince by supp ly . In mo st of Londonderry sub-div ision and about 50% of Oma gh sub-d ivis ion , d istrict meters have been use d for measuring min imum night flows . In the rest o f the Province , wa ste water

meters have been used . In orde r to emp loy the to tal night flow rate

method , it is necessary to exp ress these values in litres/prope rty/hour and

to establish legitimate night flow rates also in litre s/p rope rty/hou r .

In areas wh ere legitimate night flow s are no t av ai lab le , estimates have been made by comparison w ith sim ilar areas or from inv estigations in

samp le areas . Sample areas were selected throughout the Prov ince and the resu lts obta ined are reco rded and averaged for urban and rural areas . Tab le 5 .1 gives mi nimum night flows and Table 5 .2 gives legitima te night flows m ea sured for these samp le areas .

In rural areas where direct prope rty coun ts are not availab le , the numbe r of prope rties relating to specific areas cov ered by waste metering ha s been assessed us ing the most app rop riate Northern Ire land Census publications . Available accura te p roperty counts ind icated that it w as necessary to apply a facto r of 1.05 to the nuMbe r of pr ivate househo lds g iv en in the Census in order to allow for non-domest ic properties .

It is not po ss ib le to estima te accu rate ly popu lation s in Waste

D is tric ts in u rban areas from census pub lications . There fore , in o rder to estab lish flow rates for urban areas in each sub-d iv ision it was necessa ry to carry out deta iled p roperty counts in se lected towns covered by waste mete ring . Rate s obta ined in th is manner we re extended to cov er the urban areas in each sub-d iv is ion . Urban areas have been de fined as those w ith popu lations in excess of two thousand .

The me thods used to asse ss night flow rates in each Div ision are • descr ibed be low .

-7 1- 1 TABLE 5 .1

1 MET ERED m IN IMUM NIGHT FLOWS IN SELECTED AREAS 1 (i) Urban Areas

No . of P rop ert ies MN F

1 in Samp le A rea 1/p rop /hr

1 Po rtadown 7345 17 Lisburn/Dunmurry 20773 18 1 Downpatrick 98 02 27 Londond er ry 18 165 29

Om agh 4297 22 Ballymena 9 16 1 23 1 Avera ge 23 (ii ) Ru ral A reas

No .of Prop ert ies MN F

in Sample Area 1/p rop /h r

Ballymena 758 0 35

Antrim 8880 32

Colera ine 13072 38 Downpatric k 9582 45

Lisburn 7860 29 Londonde rry 74 15 70

Oma gh 11153 47 Enniskillen 5952 45 1 Craigavon 72 11 46 Newry 14885 37 Av erage 42

-72- I I TABLE 5 .2

M ET E RED LEG IT IMAT E N IG HT F LOW S IN SELECT ED A RE A S

U rb an Ar eaS

No . o f Prop ertie s LNF

in Samp le A rea l/p rop /h r

Londonde rr y 2686 2 .26

Oma gh 974 0 .8 5

Enn isk ille n 409 0 .46

Po rtadown 4874 1.4 0

Lisbu rn 2077 3 0 .93

Down pat ric k 9 802 1.36

Av e ra ge 1.00

(ii ) Ru ra l A re as

N o .of Prop e rties LNF

in Sam p le A rea l/p rop /h r

Londonde rry 889 2 .25

Po rtadown 607 0.24

Lisbu rn 2530 3.67

C ast le reagh 28 3 1 0.09

Downp at ric k 9 582 5 .95

Av era ge 4 .00

-73- 5 .4.2 Eastern Division

The who le of Be lfa st and Lisburn sub-d ivis ions are prov ided w ith w aste metering and p roperty counts in the waste districts are ava ilab le .

M in imum night flows and legitimate night flow s ha ve been monitored

period ically since 1976 and 198 1 records have been used .

In the Downpatrick sub-division app roximately 55% of the area is

covered by w aste m etering . Le gitim ate night f low and p rop erty coun ts are

availab le for som e waste districts . Fo r the pa rt of the sub -division not

covered by waste m etering, average values of mi nimum and leg itimate night

flow for sim ila r areas have been used . Wh ere nece ssary , prope rty counts have been estimated using NI Cen sus Publicat ions .

5.4 .3 Northern Division

In order to assess nigh t flow rates for the ur ban pa rt of the

Northern Div ision , a samp le area comp rising that pa rt of Ba llymena town covered by wa ste metering was monito red in orde r to estab lish legitima te

night flow . The numb er of p rop erties included in th is area w as count ed and the results we re used togeth er with 198 1 or other app ropriate record s of

mi nimum night flow to assess typ ical night flow rates for u rban areas

throughout the D iviSion .

In the rura l areas of each sub-div ision , large area s are covered by waste m etering but no leg itimate night flow rates or accu rate p roperty counts are availab le . Avera ge figur es for legitimate ni ght flow obtai ned in samp le ru ral areas have been used in these areas . Property counts for

the rural area s cov ered by waste metering and for the sub-d ivision as a who le were estim ated usin g NI Census Publications . Night f low rates in

litres/prope rty/hour calculated on the basis of these figures we re extended to cover each su b-div ision .

5.4 .4 Sou thern Divis ion

In order to assess nigh t flow rates in the urba n areas of Cra igavon sub-div ision , a samp le a ea in Porta down w as monitored to establish legit imate night flow . Prope rt ies we re counted over the area and

-74- re lated to 198 1 or other app ropriate reco rds of minimum night flow . Night

flow rates calculated on the basis of th ese figures were used for u rban

areas throughout the sub-div ision .

It was not po ssib le to se t up samp le urban areas in the New ry and Arma gh sub-d iv isions and average figures shown in Tab les 5.1 and 5.2 have

been used .

The tota l numb er of prope rties in the rura l areas of each sub -

div ision has be en estima ted using NI Census publications . The numbe r of

rura l prope rties covered by wa ste metering in the Cra igavon and New ry sub - div isions has also been estima ted in th is way and night flow rates hav e

been assessed using average va lues of legitimate ni ght flow .

In the rural areas of Arma gh sub-division average night flow rates obta ined for the rura l areas in the rest of the Prov ince have been used .

5.4 .5 Western Div ision

Mi nimum night flows in the Londonde rry and Omagh urban areas can be re lated to p roperty counts and flow rates ca lcu lated . Th ese values , together with legitima te night flows obtained in samp le areas , have been used to estimate net night flow rates for the u rban areas of Le ndonderry and Oma gh sub-div isions .

Ave rage figures for night flow rate s in urban areas have been used for the u rban areas of Enn isk illen sub-div ision .

N igh t flow rate s fo r rura l areas in each sub-division have been est ima ted using av ailable mi nimum night flows , average legitima te night flow s from sample areas and prope rty coun ts assessed us ing N I Census publications .

The N I Census publications have also been used to assess the total p rope rt ies in each sub-d iv ision .

-75- 5 .5 EST IMAT ION OF LEAKAGE

5.5 .1 Distribu tion S stem

Night flow rates asse ssed for each sub-divis ion have been used to estimate leakage using the method ou tlined in 5.3 above . The results

obtained are pre sented in Tab le 5 .3 .

5.5 .2 T runk M ain Leaka e

In genera l little information is avai lable regard in g leakage from

trunk mains throughout th e Provin ce . Trunk ma in .leakage has therefore been asse ssed using the result s of a survey of data from UK water authorities

given in the W RC Technica l Rep ort T R 154 .

Th e repo rt states that the ma jority of mains tes ted (73t ) had leaka ge of less than 500 litres p er hour per ki lom et re . It also states that o lder

trunk ma ins we re more likely to expe rience higher leakage than newer ones and that no relationship cou ld be found betw een leakage and the diam eter of

the ma ins .

In areas othe r than the Ea ste rn D iv ision , the trunk mains are

re latively new and a figu re of 500 l/km/ hr has been used to assess leaka ge . In the Eastern Div ision , where the mains are genera lly olde r , a slightly higher figu re of 750 1/kmihr has been used .

A detailed exami na tion carried out in the Downpa trick sub-division indicated th at approx ima te ly 10% of the mains in the sub -div ision are not

included in wa ste districts . In view of the resu lts of the expe rim ent and p rogramme refe rred to in Rep ort 26 which indicates th at in general trunk

main leakage is relatively sma ll it is considered sufficiently accurate to use th is pe rcenta ge for the purpose of estimating trunk ma in leakage in the

'other Div isions . Th e es timates of leakage from trunk ma ins in each sub - division are also given in Tab le 5.3.

-76- S I l a = I M I ON • M S M O IS • • M N IN• 0 0

TABLE 5 .3

N IGHT FLOW RATES AND LEAKA GE

Su b-Division Numbe r of Night Flows • Estimated Leak age M l/d Prop erties Litres/prop/h r Length of T runk Mains

Mi nimum Le gitimate Net KM Distribution Trunk Tota l System Ma ins

Belfast Urban 166 136 23 2 2 1 3 13 63 .1 4 .70 67.8 Harbour Estate 2 .1 Dow npatrick Urba n 28743 27 1 26 1 68 33 .0 1.0 -...; Rura l 26703 45 7 38 34 .0 -4 1 Lisbu rn Urba n 23964 18 1 17 36 10 .6 0.5 Rura l 7863 29 5 24 11.1

EASTERN DIVISION 253409 25 2 23 4 17 106 .7 6 .2 115 .1

An trim Urban • 13004 23 1 22 Rural 133 12 .9 1.3 14 .2 14765 • 32 4 28

Ballym ena Urban 1152 3 23 1 22 145 18 .0 1.5 Rural 14765 32 4 28 19 .5

Colera ine Urban 5255 23 1 22 142 16 .4 1.4 • Rural 22293 38 4 34 17.8

NORTHERN DIV ISION 87634 32 3 29 420 47 .3 4 .2 51.5 • INN • • • M I INO • = I OM I= 61111 IIMI • = I OM MN M I OM

TAB LE 5 .3 (CONT )

NIGHT FLOW RAT ES AND LEAKAGE

Sub— Division Numbe r of Nigh t Flows Estima ted Leakage M l/ d Properties Litres/p rop/hr Length of Trunk Mains

M inimum Le gitimate Net KM Distribution Trun k Tota l System Mains

Arma gh Urban 63 14 23 1 22 138 15 .0 1.4 16 .4 Rural 17346 42 4 38

Craigavon Urban 1696 3 17 1 16 174 11.3 1.7 13.0 Rural 8202 46 4 42

CO New ry Urban 879 1 23 1 22 179 17.3 1.8 19 .1 Ru ral 22273 37 4 33

SOUT HE RN DIV IS ION 79889 32 3 29 49 1 43 .6 4 .9 48 .5

Ennisk il len Urban 3200 23 1 30 110 12 .6 1. 1 13.7 Rural 13868 45 4

Lo ndonderry Urba n 20 769 29 2 27 206 23.0 2 .1 25 .1 Rural 9901 70 4 66

Oma gh Urban 7097 22 1 2 1 112 14 .7 1. 1 15 .8 Rural 146 18 47 4 43

- WE STERN DIV ISION 694 53 4 1 3 38 428 50 .3 4 .3 54 .6 3.5.3 Tota l Leakage

Leakage from the distribution system and trunk main leaka ge have been added to give the total leakage in each sub-div inion as set out in the final column of Tab le 5.3.

-79- 11111 1• 11 0 • 10 M 11111 0 11 1111 0 1 111• 0 0 111 0