I I I I I I I t I I I I I I t I t I I I N A T U R A LE N V I F O N M E NRTE S E A R CC HO U N C I L INSTITUTEof HYDROLOGY REPORTNo 17 JULY 1972 A CONCEPTUAL RUNOFFMODELFORTHECAMCATCHMENT by w. T. DtcKtNSON and J. R. DOUGLAS I N S T I T U TO EF H Y D R O L O G Y HOWBERY PARK WALLINGFORD BERKS HI R E I I I T t I I I I I I T I I I T t I I I t t I CONTENTS t I I I I I I I I I I I I I I I I Page t. INTRODUCTION 2 1.1 Purpose at1d objectil€s 2 settiig The lhlsio8laphic 2 1.3 Data networks 5 2, HYDRO',oCI CA.IJPNOCESSES 7 2.1 General vofuEtric 2.2 Mo}pholEtri 7 8 q^; r€spcmse c in ali ces e+^Pqr. B Grounah,rater1eve1s 10 Gromalvater storage and outf[ov 13 I F^i c+r,!@ 2.6 Cdrcl-usion frord process observatlods 3, DIVXIoPMEtrT0I coNCrPTUAl IIODEL A furEpednodel optiDisation of par€.neter€ 2T 29 Interilependence of par&@ters 3.b Prediction test L. CO,\C],L5IONSNND COMMENTS 5. REIERENCES 39 43 I I I I t I I I I FIOURES Pege l. The Can catchnoent above Der'nford Mill- 2. Soifs snd su.rface drainage catchlrent 3. Estimated field capacity piofiles observatio"r s.iLes l+. waler ta.o.Lecontours ror l). J.oy 5. Ch€rrges in vater cat chnent 6. Groundva.ter alischarge region 7. Con4)arati1/e lrater table ccntour3, titted and quartic suxface, fox 25.3.69 8. Ivbsn vater table level computed by i.ntegration of quartic surface, pLotted aeainst the arithrnetic average vefl levef I 2 I T I I I I I I t table on the Can c@ditions for soil at southen of the Estination of stora€e coefficient tater level and outflov data by eye, 11. Felationsbip betreen basellov anal nean uater 12. Relationship bet{een baseffov and fagged rater 13. Growah,rater storage 6nal baseflov l\. Three storage 1r. Graphica.l representation of nodel to soil noistu.e store 16. Sanple recoral of obserl€d snd predicted 17'22 Error function 23. Pr€dicted hoi?c fmction table 20 le\.ef table 2\ level of recharge the C&n catchmeni; functions runoff values betlreen fo}^ 1966-1969 ilischalges 28 relating s u.rfaces shoeing interdelenatence nf h ^dal , -_-_-!aratreters s'lal observeal monthly 18 grormih-ater Sanp.leseasonaf Datterns of soil wate! ccotent, percolat j. on t o grounabrater, grorm dlwater s tor6,ge a.nd baseflolr vari6,'< 15 16 Ca& catchmerlt fron noite l for 1ll enal of 10. conceltu€"f l1 raoistur€ 12 gradient as a linear 3 6 36-38 l+l TABIES 1. Mor-?honetri c paralleters 2. le charge estinates 3. Recharge fates required to yiefal estinateal for storage coefficient of 2.lr percent f+. Fr.mctions used in various 5. Sets of optindsed larareter vatues ! vith indices of efficiency anil vo-tune error a. Goodness of fit of the Csln catchnent from grounalsater indices ncdel 9 balance volunes 22 for.lB respective 33 - 1 INSTITUTE OF HYDROLOGY Report No. 17 J u l y 1972 A CONCEPTUAL RUNOFF mDEL FOR THE CAM CATmMWT W,T, Dickinson* and J.R. Douglas ABSTRACT The hy&ological response of t h e 200 km2 catchment o f the River Cam above Dernford M i l l , Cambridgeshire, is shown to be cansiderably more uniform than would be expected for an area half covered by boulder clw and h a l f exposed chalk. Percolation to the grounhater system is found to take place over p r a c t i c a l l y t h e whole area. As a consequence, a 'lumpedv conceptual model was thought adequate to simulate t h e catchment response to p r e c i p i t a t i o n . The resulting model simulates the groundwater sys tern and i t s contribution to streamflaw adequately, although t h e modelling of surface runoff events proves more d i f f i c u l t , due to t h e a r t i f i c i a l controls present in the catchment. *School of Engineering, University of Guelph, Canada 1. 1.1 INTRODUCTION Purposes and objectives The simulatim of runoff processes of a catchment is an important a i d i n f o r e c a s t i n g river flows and f o r managing t h e water resources of river basins. Ideally, the parameters of such a s i m u l a t i a model should be determined objectively and should be closely r e l a t e d to measurable or p r e d i c t a b l e physical bas.in c h a r a c t e r i s t i c s . Differences ir, parameter values m e y then be examined i n l i g h t of p h y s i c a l changes on a basin, natural or man-made, or changes in b a s i n characteristics from one catchment to another. The C a m catchment, d r a i n e d by the River C a m above Dernford M i l l , h a s been s e l e c t e d for t h e purpose of developing a runoff simulation m d e l permeable basin. for a The o b j e c t i v e s of t h e s t u d y have been ( i ) to hypothesize a conceptual r u n o f f model f o r the C a m catchment, and (ii) to t e s t t h e s u i t a b i l i t y of t h e model f o r s e l e c t e d periods of measured i n p u t data. 2 The physiographic s e t t i n g Located approximately t e n miles south of Cambridge, t h e Cam catchment comprises 197 km2 of predominantly agricultural countryside. The basin is s i t u a t e d on the margin of the Fens ; in the headwaters and f u r t h e r to the s o u t h e a s t l i e t h e t i l l covered Chalk uplands and to t h e northwest lies the w e s t e r n plateau. Tne c e n t r a l branch of t h e r i v e r rises in t h e Chalk uplands s o d h of Saffron Walden, and flows s l i g h t l y west of north to the gauging s t a t i o n near J k r n f o r d fill. A map of t h e catchment and its l o c a t i o n are shown in Figure 1, The Chalk uplands are characterized by a low Chalk plateau with a variable, and often considerable, thickness of Boulder C l a y as a s u p e r f i c i a l d e p o s i t (Sparks, 1957). The p l a t e a u is bounded to the northwest by a gentle and irregular escarpment. The Boulder C l q , deposited to n t a x i m u m depths on t h e emtern and w e s t e r n topographic d i v i d e s o f the catchment, stretches almost to t h e present v a l l e y bottom at a number of locations. The general. topography ( F i g u r e 1) in c o n j u n c t i o n w i t h the d i s t r i b u t i o n of Boulder Clay (Figure 2 ) suggests t h a t the valleys are largely pre-Boulder Clay features , as is the es carpmen t . . FIGURE 1 THE CAM CATCHMENT ABOVE DERNFORD MILL The Chalk formation rests on the Cambridge Greensand, a concentrate of coarse material derived from c l a y (Steers, 1 9 6 5 ) . The Greensand and the u n d e r l y i n g G a u L t Clay e x h i b i t a low regional d i p t o t h e southeast, causing outcrops i n the Chalk to form b e l t s l e n d i n g n o r t h e a s t to southwest. Of p a r t i c u l a r interest i s a h a r d , relatively impermeable band o f Greensand which o u t c r o p s in t h e v i c i n i t y of Dernford M i l l , The presence of this band is l i k e l y t o prevent t h e occurrence of much subsurface seepage northwards past t h e r i v e r gauging s t a t i o n . In t h e mass, the Chalk is very porous, h o l d i n g water in about one t h i r d o f i t s total volume when saturated (Woodland, 1946). However, the pores are small and w a t e r i s held l a r g e l y by c a p i l l a r y a c t i o n . 4,s a result, t h e water c a r r y i n g capacity of the Chalk depends on the e x i s t e n c e and c h a r a c t e r of fissures. ( ~ u n t eB r lair, The e f f e c t i v e p o r o s i t y , thought to be about 1.5 to 2.0 percent 1965 ; Steers, 1965), is l i k e l y t o be dependent therefore upon t h e d i s t r i b u t i o n and s i z e of the fissures. The Chalk is so porous and w e l l f i s s u r e d t h a t ~ r t u a l l ya l l water f a l l i n g upon it p e r c o l a t e s , r e s u l t i n g i n l i t t l e , if any, d i r e c t m o f f . Only in t h e v a l l e y bottoms, where t h e forma- t i o n nay be saturated to the surface, does surface r u n o f f occur. On t h e other hand, the Boulder C l a y deposit is thought to be consider- 1946; Thomasson, 1969). T y p i c a l l y , it yields l i t t l e water; however, it may become very pebbly in places, or contain patches a b l y less permeable (Woodland, of rubbly c h a l k or lqrers o f s a n d and gravel, l e a d i n g to t h e presence of small if unreliable water supplies. Whereas direct measurements o f p ~ r c o l a t i o nhave been conducted in t h e C n a l k (Woadland, 1946), y i e l d i n g annual amounts of 150 t o 180 v a l u e s are available for the Boulder Clay. m, no measured Woodland ( 1946) and Hunter B l a i r (1965) assumed a p e r c o l a t i o n rate for t h e deposit of $0 mms p e r year. From the f i s s u r e d nature of the Chalk and the variable composition of the Boulder C l a y , it seems l i k e l y tha* p e r c o l a t i o n rates Tor both may vary considerably over the catchment. Such v a r i a b i l i t y is noted in t h e surface drainage c h a r a c t e r i s tics given in Figure 2. The presence of a b u r i e d channel beneath t h e c e n t r a l branch of the river has been r e p o r t e d a n d documented (woodland, 1946; Steers, 1965). '&li S t a r t i n g at ttles ford, t h e channel deepens southward, departing f r o m t h e present course of the Cam near Quendon. It is deep, narrow, steep-bided and filled t o depths of 60 m with a chalky silty till. Due to the nature of t h i s f i l l material, and to t h e observation t h a t buried channels in the region are not graded t o a base l e v e l b u t rather undulate unevenly in a series of rack b a s i n s , it appears u n l i k e l y t h a t much undergromd seepage occurs aut of the catchment by w a y of the buried channel. The p r e c i p i t a t i o n pattern over the area is fairly uniform, with an average r a i n f a l l of about 635 mms. Woodland ( 1946) has noted the seasonal d i s t r i b u t i o n for t h e period 1877 t o 1944, with a low i n March of 4 1.5 mas, and the w e t t e s t months of July and October y i e l d i n g 68.8 mms and 65.6 ms respectively. The annud p o t e n t i a l transpiration f o r Cambridgeshire is approximately 530 mms ( b b i s t r y of Agriculture, Fisheries and Food, 1967). The seasonal pattern is very marked, with 460 mnas during the summer period f r o m April to September, m d 70 mrns i n the winter months. 1,3 Data.networks The network of hydrological instruments located in and around t h e C a m catchment during t h e interval 1966 to 7970 is s h m in Figure I . The data collected from the various f i e l d instruments have been checked and u t i l i z e d for t h e compukation of (i) river discharge, (ii) mean basin p r e c i p i t a t i o n , (iii) p o t e n t i a l evaporation arid t r a n s p i r a t i o n , and (iv) mean b a s i n s o i l moisture status. The river discharge has been determined for two hourly i n t e r v a l s from t h e continuous r i v e r s t a g e record and a stage : dischare r e l a t i o n s h i p developed by the Great Olrse River ~ u t h o r i t y - f i a i l y mean basin precipitation h a s been computed as the Thiessen average of the gauge network. Hourly p r e c i p i t a t i o n was obtained by d i s t r i b u t i n g t h e basin mean i n time aceoraing to t h e c o n t i n u ~ u s _I P o t e n t i a l daily evaporation and gauge record obtained at the climate station. transpiration estimates have been determined f r o m t h e equations presented by Penman ( 1948) X~rorn each s e t of s o i l moist- condition has been determined. data, a. mean b a s i n s o i l moisture The above data are stored on magnetic tapes *... .. .... . . Cakanous gloy at; Chdky Bouldmr Clay ; Poor to free drainage. 8,- . .. . . .... * . a m uluP.wr d and Brolrvn earth; Chalky drlft Ww Chalk; Frw drainam. Imperfrct to poar drainage. Calcanous gamy mil; Alluvium, locully overlyi nu peat; Imperfect to poor dra~nage. ml 0 1 Towns and viblag=. 1 FIGURE 2 Seal. in krn 2 3 4 rn 5 SOILS AND SURFACE DRAINAGE CONDITIONS ON THE CAM CATCHMENT ( after Thomasson, 1969) and are r e a d i l y available for computer analysis. 2. HYDOLOGICAL PROCESSES Examinaticol of t h e s e t t i n g of t h e Cam Catchment reveals two aspects of t h e hydro1ogica.l f l o w regime which warrant some i n v e s t i g a t i o n prior t o the development of a simulatim mdel. Firstly, the role which t h e Boulder C l a y plws i n d e t e h n i n g d i r e c t runoff and recharge t o the groundwater system requires some c l a r i f i c a t i o n . SecbndJy, it would be helpful i f t h e storage characteristics o f t h e Chalk and t h e i r areal variability within the catchment were b e t t e r understood. The data available for t h e experimental catchment has allowed study of these aspects, 2 1 General volumetric response Preliminary analysis of the runoff hydrographs for the period January 1966 to December 1970 reveals t h a t an average of only 30 p e r c e n t of t h e water f a l l i n g as p r e c i p i t a t i o n discharges i n t o the river during a 80 percent of this runoff occurs as baseflow, or as slow response. Data collected between 1952 and 1962 ( ~ u n t e rB l a i r , 1965) also year. Further, show t h a t mly 25 to 30 percent of the r a i n f a l l runs off above Demford M i l l . Such figures are i n d i c a t i v e of an extiremely permeable catchment. Monthly runoff percentages pertaining to rapid or direct r u n o f f volumes also denote h i g h permeability. Values determined by d i v i d i n g the estimated d i r e c t runoff by t h e p r e c i p i t a t i o n during each monthly interval range between 0.4 and 19.0 p e r c e n t for t h e p e r i o d 1966 to 1970. me d i s t r i b u t i o n of t h e s e values is p o s i t i v e l y skewed, t h e mean being 6.0 percent and the median 4.0 percent. Compared with similarly computed values in the literature ( ~ i c k i n s o nand Whiteley, 1970), these index values for the Cam catchment confirm that the basin does n o t generate much rapid runoff at any time of' t h e year. These i n i t i a l analyses s u g g e s t rather strongly t h a t mst of the area of t h e catchment which is covered by Boulder C l a y deposits must be relatively permeable material since if it were t i g h t and impermeable considerable surface runoff w o u l d be generated, resulting in h i g h e r annual and monthly runoff volumes. I - 8 - 2.2 B 8 I Morphometric indices A s m m q y of some morphometric parameters o f presented in Table 1. the Cam catchment is Regions e x h i b i t i n g t h e most impermeable d r a i n a g e I c h a r a c t e r i s t i c s i n c l u d e Debden Water, Fulpen Slade , and t h e headwaters of Wenden Brook (see Figure I ) . The remaining major portion of the catchment I d i s p l a y s tributary p a t t e r n s exemplary o f extremely l i m i t e d surface runoff conditions. I The proportian of e f f e c t i v e l y impermeable area r e f l e c t e d by the This geomorph~logicpattern i n d e x e d above appears t o be seven to ten p e r c e n t . area is concentrated i n the headwaters above Saffron Wdden , w i t h a small p o r t i o n on t h e western watershed. In general, such a p a t t e r n i s i n accord w i t h t h e volumetric response observea in t h e preceding section. 2.3 Soil moisture storage The data o b t a i n e d monthly from t h e various soil moisture observation s i t e s have been e x m i n e d with regard to v a r i a b i l i t y in time and space. As the s i t e s were i n i t i a l l y selected to be representative of the mspped soil conditions (see Figures 1 and 2 ) , and l o c a t e d under similar grassed v@geFtation,it w a s a n t i c i p a t e d t h a t similarities and/or differences i n soil moisture storage conditions might be observed. Fox each of the seven sites for which soil moisture measurements were c o l l e c t e d d u r i n g t h e i n t e r v a l 1967 t o 197'0, estimated. a f i e l d capacity p r o f i l e h a s been These profiles ( ~ i g u r e3) were determined from f i e l d measurements o b t a i n e d during the w i n t e r months of December to March before which no p r e c i p i - t a t i o n had been o b s e r v e d f o r at least five days. The v a x i a b i l i t y between such s e t s o f measurements at each s i t e w a s found generally to be w i t h i n sampling e r r o r , and the r e s u l t i n g p r o f i l e s are thought to be representative of f i e l d capacity c o n d i t i o n s . Figure 3 reveals t h a t , except for sites A and G, the estimated profiles are remarkably similar. S i t e A, l o c a t e d in the c h a n n e l alluvium, m i & t expected t o behave uniquely. be On t h e other h a n d , the d m s t constant " s h i f t " of s i x to eight p e r c e n t moisture volume f r a c t i o n e x h i b i t e d by site G i s 1 8 8 t I 1 1 I J I m 8 thought to be peculiar to the site and not necessarily representative of As the neutron probe has n o t been c a l i b r a t e d for each i n d i v i d u a l s i t e , it is possible for estimated moisture volumes to be the soils in the area. in error due to i n s t a l l a t i u n or site peculiarities. Considering the v a r i a b i l i t y of soil conditions d i s c u s s e d in t h e I n t r o d u c t i a n , t h e similarities in Figure 3 seem remarkable. Although t h e seven soil moisture observatim s i t e s represent a small s m p l e of the soils in t h e catchment, nevertheless the above results f'wther suggest that some o f t h e Chalk and Boulder C l a y areas s t o r e moisture in a similar way. Correlations between observations at the various sites and an areal w e i g h t e d mean f o r t h e catchment reflect the similarities observed above. A l t h o w the Mole Hall location, s i t e E , w a s found t o be the best c o r r e l a t e d , t h e r e was very l i t t l e to choose between sites B, D, E, and F, located in both Chalk and Boulder C l a y . 2.4 Groundtwater l e v e l s Observations discussed in the p r e c e d i n g sections suggest t h a t , from a h y d r o l o g i c a l respmse p o i n t of v i e w , much of t h e area of the Cam catchment which is characterized by a s u p e r f i c i a l Boulder C l a y deposit may behave similarly to regions of bare Chalk, That is, t h e manner in which regions generate d i r e c t runoff and store soil moisture appears similar. Consideration is now given to t h e groundwater system in order to study the recharge, storage, md outflow patterns from t h e saturated zme. Water t a b l e contour maps ( ~ i g u r e4 ) have revealed t h e possible e x i s t e n c e of a r e g i o n a l groundwater flow system and allowed the following observations to (i) the regional water t a b l e d i v i d e generally coincides with t h e surface topographic d i v i d e ; (ii) water table gradients over much of t h e area are be made: similar; and (iii)the growhater system discharges along the c e n t r a l river course. With the possible exception of t h e southern edge of the catchment, the water table d i v i d e corresponds to t h e physiography. F u r t h e r , t h e flow l i n e s t e n d to converge on t h e o u t l e t s e c t i o n n e a r Demford M i l l . A t the southern end, t h e water t a b l e gradient becomes v e l y flat and t h e grounhater divide FIGURE 3 ESTIMATED FIELD CAPACITY PROFILES FOR SOIL MOISTURE OBSERVATION SITES Water table contour metres 0.0. -6- 0 1 Scale in km 2 3 ? FIGURE 4 5 \ / I WATER TABLE CONTOURS FlTTf D BY EYE FOR DATA OF MARCH 25th, 1989 8 Data from wells i n the vicinity suggest t h a t the d i v i d e mhy in fact s h i f t its position seasonally (Figure 5). However, as the gradients ill-defined. are particularly l e v e l , seepage into or out of the catchment is l i k e l y t o be of n e g l i g i b l e volume. A similarity in water table gradients over much of the catchment area implies t h a t recharge to the groundwater system and transmissibility throw t h e saturated zme are somewhat uniform. In small areas, where there are local i r r e g u l a r i t i e s i n surface conditions, soils and the Chalk, this statement is undoubtedly invalid. The appearance of s t e e p gradients in t h e northwest of t h e catchment A similar concentration of low transmissibility coincides w i t h t h e escarpment. has been observed i n a l i k e location t o t h e east of the Cam catchment (Woodland, 19461, It is suggested that such zones mi&t have been the result of t h e w a y i n which t h e Chalk s t r a t a were folded. Examination of t h e water table contour maps i n conjunction w i t h t h e topographic map reveals t h e major groundwater discharge region ( ~ i g u r e6 ) . This regim appears to be continuous along the main stream channel, implying that all t r i b u t a r i e s to this channel are likely to become dxy during the summer months. As is t o be expected, the discharge r e g i o n of Figure 6 corresponds closely to khe region of maximum water well y i e l d mapped by Woodland ( 1946). In addition to t h e i r variability in space, t h e w e l l levels w e r e c o n s i d e r e d with regard to t h e i r f l u c t u a t i o n s in time. Although the range of seasonal f l u c t u a t i o n can vary s d s t a n t i a l l y from one observation site t o another, a l l w e l l s generally peak at t h e same t i m e in March or A p r i l , and reach a low together in October or Noveniber. No l a g between wells near t h e divide and those in the valley is apparent, nor between those beneath Boulder Clay and bare Chalk. Recharge to t h e water table, therefore, seems to occur similarly in time t h r o ~ o u t h e b a s i n . The w e l l s exhibiting the greatest f l u c t u a t i o n s in water l e v e l are l o c a t e d in t h e zone o f low t r a n s m i s s i b i l i t y . 2.5 Groundwater storage and out flaw, In order that estimates could be made of the valme of groundwater & Y - 1.5. Gradient toward catehrmnt 5 o; N Gradient out of the catchment A 4 $ 1.0- e # 0 L*1 K Z m 2 D rn I #"i I I YI A 0 t 2-0.5- I I 0 K W I I 1 -1.0- I I I I I I '1 FIGURE 5 I, I I I I IYI 1 I t I t ,8' ' 4' I I1 :: ; I I 4 t w I I ' I I 1 #\ ' '' I I a '' , ,4' r' I I I II II II It I t -1.5- A r l 1968 CHANGES IN WATER TABLE GRADIENT AT SOUTHERN END OF CAM CATCHMENT Scale in km FIGURE 6 GROUNDWATER DISCHARGE REGION OF THE CAM CATCHMENT Eye fittinp, contours in m. Quartic function fitting, oontours in m. - \ \ \ Weighting coefficient x 100. a -4.12 FIGURE 7 '. f' P COMPARATIVE WATER TABLE CONTOURS FITTED BY EYE AND A QUARTIC SURFACE FOR MARCH 25 th,I969 8 s t o r e d in the catchment, a quartic surface was f i t t e d to each s e t of water level measurements. A mean water t a b l e l e v e l was then deterwined a t t w o weekly intervals, employing a s e t of w e i e t i n g coefficients f o r 38 well sites to f a c i l i t a t e the volumetric integration (Chidley and Keys, 1970). A comparison of t h e q u a r t i c and eye f i t t i n g s t o a s a t of data is shown in Figure 7, along with t h e weighting coefficients which ma^ be applied to t h e appropriate well levels to. determine a basin mean water table level. A comparison of the b a s i n m e a n water table l e v e l estimated f r o m t h e quartic surface and from an a r i t h m e t i c awrage is presented in Figure 8. Observation points appear to be s u f f i c i e n t J y numerous and w i d e l y d i s t r i b u t e d in area t h a t differences in l e v e l from date to date are essentially i d e n t i c a l , Storage changes estimated by i r r e s p e c t i v e of computation procedure. arithmetic mean tend to be four percent l a r g e r than those y i e l d e d by t h e i n t e g ~ a t i o nmethod. Changes i n t h e mean water table level, itself an index of t h e groundwater storage condition over t h e catchment , w e r e considered together with estimated groundwater runoff volurnt3s during t h e summer months to determine an e f f e c t i v e porosity, or storage c o e f f i c i e n t , for t h e saturated zone. selected (i.e. For the data during periods of little or no groun&ater recharge) the base flow v o l w should approximate t h e change in grounChater storage. The s l o p e of Figure 9, t h e r e f o r e , is an estimate of the s t o r a g e c o e f f i c i e n t . If it could be assumed t h a t there w e r e no seepage or evaporation losses during the time i n t e r v a l s considered, and t h a t t h e estimates of storage change and r m o f f were w i t h o u t e r r o r , then the line used for determining the storage c o e f f i c i e n t would be a lower envelope of t h e measured points. However, in t h e l i g h t of the data and computations used, t h e line was sketched as shown. The storage coefficient having been estimated to be 2.4 percent, a water balance of t h e groundwater system was performed, for which c o n t i n u i t y was expressee, GPR - GRO = GS where GPR is the recharge volume to groundwater in a given ti= interval, 2 E k 53- f sz- E A A A Y 3 2 51- d f n sox Ez 49 - I I A 49 47 l8 i0 ARlTHmTlC MEAN WELL LEVEL, METRES FIGURE 8 MEAN WATER TABLE LEVEL COMPUTED BY INTEGRATION OF QUARTIC SURFACE PLOTf ED AGAINST THE ARITHMETIC AVERAGE WELL LEVEL C H A N L IN GROUNDWATER STORAGE, MMS FIGURE 9 ESTIMATION OF STORAGE COEFFICIENT FROM GROUNDWATER WATER LEVEL AND OUTFLOW DATA BASEFLOW, MMS/DAY GROUNDWATER* STORAGE, MMS PERCOLATION TO GROUNDWATER, MMS / DAY SOIL WATER CONTENT, MMS GAO is the groundwater outflow volume in the same i n t e r v a l , as estimated fmm the streamflow hydrograph, GS is t h e change in groundwater storage, estimated from the mean water table Levels and t h e storage c o e f f i c i e n t . The patterns of recharge, storage and outflaw determined in the balance are i l l u s t r a t e d i n Figure 10. Soil nmietum observations have been included for carsparison, along w i t h recharge, or percolation, estimated from a storage c o e f f i c i e n t of four percent. Figure 10 reveals two items of particular i n t e r e s t . Firstly, an increase in s t o r a g e coefficient to four percent changes t h e pattern o f seasonal recharge, but does n o t alter t h e annual volume. groundwater out f l a v hydrograph anticipates rather than 1- storage condition. E a c h i t e m is cansidered -her Secondly, t h e the groundwater below. The use of a storage c o e f f i c i e n t of 2.4 percent results in the estimation of essentially no recharge during the summer months of 1967, 1969 and 1970, w i t h cmtinuous recharge throughout 1968. An increase in storage c o e f f i c i e n t t o 4.0 percent reduces the sunner recharge i n 1968 t o a n e g l i g i b l e amaunt, but results in a requirement for negative recharge during t h e o t h e r three summers. Such negative values could he meaningful either if seepage o u t of t h e c a t c h m n t into adjacent bagins w e r e to occur at such times, or if s u b s t a n t i a l evaporation w e r e to take place from the groundwater system. It seems l i k e l y t h a t if any seepage occurs across the catchment b o u n d q during the summer seasons, it is i n t o rather than out of the Cam Catchment (see Figure 5). Further, as t h e water table is a cmsiderable distance below t h e ground surface in a l l but the immediate discharge region of the main stream, large amounts of evaporation from the groundwater system seem u n l i k e l y . Therefore, m t h e basis of seasonal =charge pattern, a storage coefficient of 2.4 percent appears r e a l i s t i c . For values of storage coefficient in the range o f those alrea* mentioned, annual recharge totals are very similar, as revealed in Table 2. For such patterns of recharge, the annual v o l m is n o t a sensitive indicator of an a p p r o p r i a t e storage c o e f f i c i e n t . It is also u s e f u l t o consider at this point t h e relatim -itude of . RECHARGE ESTIMATES DETERMINED TABLE 2. FROM GROUNDWATER BAUN CE Recharge to groundwater ( m over catchment) Storage c o e f f i c i e n t of 2.4 percent Time interval Storage coefficient of 4.0 percent 6th J m 1967 to 3rd Jan 1968 136 139 4th Jan 1968 to 3rd Jan 1969 1 86 20 1 3rd J m 1969 to 8th Jan 1970 150 133 124 143 9th Jan 1970 to 26th TABLE 3. me 1970 F3CHARI;E RATES REQUIRED TU YIELD ESTIMATED VOLUMES M)R STORAGE: COEFFICIENT OF 2.4 P E R W T Recharge rate t o groundwater ( n m over catchment axea noted) Time interval Area of 91 h.2 Area of 155 kn? Area of 197 3 6th Jan 1967 to 3rd Jan 1968 295 172 136 Jan 1969 40 4 236 186 3rd Jan I969 to 8th Jan 1970 326 190 150 9th Jm 1970 to 26th dune 1970 269 157 124 4 t h Jan 1968 to 2nd the annual recharge volumes. Table 3 shows the annual recharge rates required if uniform percolation w e r e assumed to take place over 9 1 km2, 155 km2 and 197 km2, and catchment. no percolation existed over t h e remainder of t h e The first a E a is approximately equivalent to that of t h e bare Chalk; t h e second mw be obtained by considering the t o t a l area less 18 km2 of groundwater discharge fegion and less 23 km2 of e f f e c t i v e l y impermeable soil; the f i n a l figure is t h e catchment area. The recharge values presented, when regarded w i t h previous estimates of 150 t o 180 mms percolating through bare Chalk annually (Woodland, 1944) , t e n d to confirm t h e notion that at l e w t 155 k$ of the c a t c h m e n t behave as permeable recharge area. The manner in which groundwater outflow precedes, or a n t i c i p a t e s s t o r a g e is curious! , This r e l a t i o n s h i p is a l s o demonstrated in Figure 1 1 by t h e strong clockwise hysteresis e f f e c t . This hysteresis is essentially remved if the base flow is p l o t t e d against the mean w e l l levels lagged by four weeks, as shown in Figure 12. A second approach to r e m o ~ n gt h e hysteresis i s t o c a s i d e r the storage to be a l i n e a r combinatkn of recharge and baseflow, s h a m in F i g u r e 13. The "backward lag" between groundwater storage and baseflow has a l s o been observed by Hunter B l a i r ( 1965). H i s explanation of t h i s phenomenon (i) begin to rise before the main catchment storage had begun to i n c r e a s e , and (ii begin Lo w a s based on a bank storage hypothesis, causing f l o w to decline before storage had achieved i t s peak. As t h e water t a b l e w e l l levels d i s c u s s e d e a r l i e r revealed n o l a g tendency in space, the above hypothesis appears as acceptable as any, altho@ it can n e i t h e r be proven nor dispraven w i t h the av&lable data. Figure 12 a l s o reveaLs that, d t h o u g h a l i n e a r r e l a t i o n s h i p may e x i s t between recharge, storage and baseflcru, the storage axis may s h i f t If such s h i f t s do occur in tirw , p o s s i b l y r e f l e c t i n g the f i l l i n g up and depletion of secondaxy storage during excessively wet or dry years. systems , modelling of the grom&waterstorage system becomes extremely tenuous 2.6 . Conclusion f'rom process observations Information on general volumetric response , geomrpholoery , soil MEAN WATER TABLE LEVEL. M FIGURE 11 RELATIONSHIP BETWEEN BASEFLOW AND MEAN WATER TABLE LEVEL Watw table levels are lagged 4 weeks behind mean daily basef lows MEAN WATER TABLE LEVEL, M FIGURE t 2 RELATIONSHIP BETWEEN BASEFLOW AND LAGGED WATER TABLE LEVEL I I t I &4 0 H 0 GROUNDWATEP STORAGE (RELATIVE TO MEAN LEVEL OF 51 M), MblS 0 a d 0 N 0 W 0 I L 0 misture and g r o u n h a t e r suggests that the Cam catchment is predominantly a highly permeable area. There i s n o indication that the Boulder Clay d e p o s i t as a whole behaves d i f f e r e n t l y from regions of bare Chalk. Further, t h e c h a r a c t e r i s t i c s of the Chalk appear uniform over much of t h e c a t c h m n t . As water storage and transmissim properties may be similar in many areas of t h e b a s i n , it i s cmcluded t h a t t h e Cam catchment may be considered as a s i n g l e homogeneous region for t h e purpose of modelling hydrologic response. 3. 3.1 DEVELOPMENT OF CONCEPTUAL M2DELS A lumped model On the basis of t h e preceding analysis, a lumped model ( i .e. lumped in space) has been hypothesized for t h e Cam catchment. The model is conceived as three storage systems, as i l l u s t r a t e d in Figure 14. The n a t u r e of each system and the c h a r a c t e r i z a t i o n of fluxes i n t o , between and away from each system are d i s c u s s e d below. a)' Surface s t o r e The surface storage component accepts t h e incoming r a i n f a l l and p o t e n t i a l evaporative demand and c a l c u l a t e s e f f e c t i w r a i n and a residual evaporative demand to be passed rm to t h e soil moisture s t o r e , after s u b t r a c t i n g evapora- tion from the surface store. The p r e c i p i t a t i o n entering the store, R A I N (11) is described by 6-hourly p r e c i p i t a t i o n t o t a l s obtained f'rom the raingauge network ( s e e Section 1.3). The p o t e n t i a l daily pan evaporation, EVAP (11), is e s t i m a t e d from t h e climate s t a t i o n data and apportioned i n t o 6-hourly i n t e r v d s with the w e i g h t i n g factors 0.697, 0.138, 0.0, and 0.165 for t h e p e r i o d s 09- 15, 15-2 1 , 2 1-03 and 03-09 hours respectively. Evaporation from t h e surface s t o r e , ES, is assunted to be a f'unction of t h e potential evaporation, EVAP (111, and t h e e f f e c t i v e r a i n discharged t o the s o i l moisture store, ERAIN, a Function of the incoming p r e c i p i t a t i o n , and t h e surface storage moisture s t a t u s , SS-DS. b) S o i l moisture store Three hydralogical processes are conceived t o be associated with t h e s o i l - 11 - - I I I I. SURFACE STORE r r r r - - - - r r - r r e I L SOIL MOISTURE STORE DCL 1 GROUND WATER STORE FIGURE 14 THREE STORE CONCEPTUAL MODEL FOR CAM CATCHMENT GS moisture storage comgonent . These are a c t u a l t r a n s p i r a t i o n , direct runoff md percolation to t h e groundwater store. A c t u a l t r a n s p i r a t i o n is assumed to be a f'unction of the evaporative demand on t h e s o i l moisture s t o r e , EEVAP, and the moisture status of the store, DC. t h a t not ~ a t i s f i e dby t h e surface s t o r e . The evaporative demand is Direct runoff, ROFF, and p e r c o l a t i o n to groundwater, GRP, are a l s o characterized as simple Arnctions of the soil moisture storage. c) Groundwaterstore Storage in t h i s component of the system is assumed to be a function of input to t h e s t o r e , GPR, and output, GRO. A s i m p l e r form r e l a t i n g only storage and outflow has also been considered. The moisture s t a t u s of the s t o r e is monitored as ES. d) Routing of runoff values : The direct runoff volume, ROW, is routed through a single l i n e a r reservoir of storage, S , and lagged by a time i n t e r v a l , DEL, t o produce the The r o u t e d d i r e c t runoff, R , is then added to d i r e c t r u n o f f contribution. t h e groun8water outflow component, GRO , which be lagged by the interval GDEL, to yield the estimated six-hourly outflow Q. The various functional forms utilized in the model are o u t l i n e d in Table 4. Some of t h e relationships are presented graphically in Figure 15. 3.2 Optimisatian of parameters The error f u n c t i o n , or residual variance, has been defined as the cumulated sum of squares of difference between computed and observed six-hourly runoff increments. That is, @ n (Q' - Q)', eh?. efficiency w a s considered as the proportion of the "no model" variance accounted f o r by the model, R~ = (F$ - + ) / F O ~ , where Fo2 was t h e "no-model" residual variance ash and S u t c l i f f e , 1970). The four years of r e c o r d , 1966 t o 1969, were divided into two equal. periods of two years' duration. Various forms o f t h e model w e r e optimised on t h e TABLE 4 FUNCTIONS USED IN VARIOUS mDEL F O M Model 1 1 Effective r a i n f a l l ERAIN=RAIB(II) + m-ss 2 Evaporation from surface ES=FS. EVAP(II) 3 Residual evap. demand EEVAP=EVAP(TI) - ES Parameters: FS, SS C o n s t r a i n t s : FS= I . 0 4 A c t u a l transpiration ET=ECP. EEVAP where E CP=FC ( DCT-DC) ( DCT-DCS - 30 Model. 2 Model 3 bdel 4 + + + + + * j + + - + + 5 Direct runoff generation ERAIN ROF-ROP where . ROP=RC/ ( EIS+DC) + 6 Percolation to groundwater GPFPGLR ( 1.0 - zM nc ) + constrained + Constraints : BOP s ROE4 + + Parameters : C S M Parameters : GSM, GF 8 S=DRK.R an d R delwed by Dm GRO delwe d by GDEL Parameters : DRK , DEL, GDEL . + No delay of GRO GDEL deleted . + + GS=GSM [GF.GPR *( 1-GF) GR3 ] * T1OP=RC. e-FS DC _C, GS=GSM. GRO + - - Parameters: FC, DCT, DCS, RC, RS, ROM, GLR, GDM 7 Flaw from groundwater i. FS n o t + - ROM deleted fro c o n t r a i n t s _CS + + + j + * j _e, a) Actual transpiration b ) Ditoct runoff ROM ROP e ) Percolation to groundwater GLR GPR 0 GDM DC FIGURE 1 5 GRAPHlCAL REPRESENTATION OF MODEL FUNCTIONS RELATING TO SOIL MOISTURE STORE first two y e a r s ' record, and t h e r e s i d u a l variance, e f f i c i e n c y , and o p t i m i s e d parameter values were noted. A f i n a l form of the model was then a p p l i e d w i t h a s e t of' optimm values to the next two years of data to o b t a i n a second and independent assessmnt of residual varimce and efficiency. The o p t i m i s a t i o n process used w a s t h a t of 0' Carznell et a2 ( 1970). Instead of optimising all parameters simultaneously, groupings of up t o e i g h t parameters were o p t i m i s e d while the remainder w e r e f i x e d a t seemingly appropriate or previously optimised values. In t h e l i g h t of e x i s t i n g interdependence among p a r a t e r s , this approach, u t i l i z e d i n conjunction with t h e mapping of t h e error f u n c t i o n for various combinations, proved e f f i c i e n t and economical. The various f u n c t i o n a l forms of the model presented in Table 4, optimised for the 1966-1967 period of record, yield the results shown in Table 5 . The s e t of parameter values l i s t e d for %&el 5 w a s subjectively s e l e c t e d , aFter i n s p e c t i o n of the e a r l i e r optimised values and mappings of the e r r o r function. Regarding t h e general "goodness of fitf'of predicted flows to t h o s e observed, t h e higher n u d e r e d mdels appear to describe grotmawater f l o w conditions adequately b u t to encounter difficuley w i t h the direct runoff. Further, the problem w i t h direct runoff appears to be associated m o r e with t h e time d i s t r i b u t i o n than w i t h the occurrence and volume o f flow. G e n e r a l l y , the observed hydrograph is more peaked than t h e p r e d i c t e d f l o w record. These observations may be noted in t h e sample period of observed and p r e d i c t e d f l o w s sketched i n Figure 16. 3+3 Interdependence of parameters A s noted by Mandeville e t aZ, (1970), t h e o p t i m i s a t i o n and utility of p a r t i c u l a x model parameters can most e a s i l y be studied if the para3lleters are o t h o g o n d ( i. e .a change in one does n o t involve a change in another) . Unfortunately, such o r t h o g o n a l i t y is not g e n e r a l l y experienced in h y d r o l o g i c a l conceptual models and interdependence of parameters ~ 0 m 3 n l y i n c r e a s e s with t h e number of parameters. TABLE 5 SETS OF OPTIMISED PARAMETER VALVES, WITH FQSPECTIVE I N D I C F S OF EFFICIENCY AND VOLUME EAROR Parmeter Model 1 Model 3 Model 2 Model 4 Model SS 14.3 14.3 9.29 9.5 9.5 FS 1.0 1.0 2.42 1.4 1 FC 0.65 0.65 0.65 0.65 0.60 Dm 16.86 16.86 16.86 16.86 DCT 128.60 128.60 104.5 ------- 104.5 - - - -- - RC I1S 4.40 4.40 4-5 18.92 18.92 I 5. 0 ROM 0.30 0.30 0.27 0.27 -5 17.0 100.0 0.24 0,24 0.029 0.20 0.029 - 0.27 0.27 0.28 57.54 57.54 - ------GLR GDM 50.0 57-54 65. 0 ------GS M 100.0 - GF GEL 600. 0 539.5 -0.15 -0.13 - -0.13 - - - 4.0 539.5 560. 0 -0.075 ------DEL 2-32 2 - 32 2.32 2.32 2.32 DRK 4.85 4.85 4.85 4.63 4.55 --Efficiency as a. percentage E r r o r in runoff volme as a per c e n t age 61 + 4.2 72 + 1.8 76 - 5 0.19 77 - 0.66 78 + 1.5 FIGURE 16 - - -P z SAMPLE RECORD OF OBSERVED AND PREDICTED RUNOFF VALUES -- --- --- -- *---- - 4 - + ---- ---- Ad- --* --* - - ---+ ----* 1- 4 i 5 & +.a . OBSERVE0 nrOROGRAPH PREMCTZD H-APH --- B ; $ a 1 a h E rx 0.5 + z I S s z w A \ ("J - Y Y -\ --% --- 1.0 1 ' 2 ' 3 ' 4 ' I ' 6 ' 7 ' 6 ' 9 ' 1 0 ' 1 1 ' 1 t ' 1 3 ' 1 4 r 1 l ' 1 6 ' 1 7 ' I 8 ' W ' ~ ' 2 ' 2 ~ ' 2 3 ' U ' 2 5 ' ~ ' n ' t l ~ 1 ' ? ' 3 ' 4 ' 5 ' 6 ' 7 ' 8 ' 9 ' * 0 ' FEBRUARY In68 MARCH 1866 A considerable range of interdependence conditions w a s found to e x i s t in various portions of t h e model forms cansidered above. S-le mappings of t h e error f u n c t i o n , p r e s e n t e d i n Figures 1 1 to 22, reveal t h e situation. Such mappings w e r e found to be very u s e m as guides to the selection of an optimum s e t of parameter values. The h i g h e s t interdependence was observed among the parameters concerning t h e surface s t o r e and evaporation from the s o i l moisture stom. The parameters are not only dependent within each of these storage phases, but a l s o between phases. For a catchment in which 70 percent of the incoming annual precipitation is evaporated, it is perhaps to be expected that a l l parameters i n v o l v i n g evaporation should be highly interdependent. The parameters describing the surface s t o r e and a c t u a l transpiration are conceived t o g e t h e r as modelling the t o t a l evapotranspiration process. For t h e p r e s e n t , they seem d i f f i c u l t t o i n t e r p r e t physically. What a surface s t o r e of approximately t e n mms capacity physically characterizes, arid w h e t h e r evaporation from agricultural surfaces occurs at one and a half times the p o t e n t i a l r a t e , remain unanswered questions. FS and SS parameters In fact, the merely d e s c r i b e a black box which suitably disposes of moisture for t h e rest of t h e model. The parameter DCT t e n d s to assume a value such that evaporation *om the soil moisture store is possible at all times. That is, DCT c o i n c i d e s w i t h the 1argest soil moisture d e f i c i t experienced in t h e p e r i o d of record. Whether such a condition indicates continuous evaporation *om the soil, or whether it r e f l e c t s the need f o r a modelling mechanism t o evaporate considerable volumes of water, is open to speculation. The direct runoff, percolation and groundwater storwe parameters are reasonably independent and more amenable to physical i n t e r p r e t a t i o n . Althoueh, the e r r o r f u n c t i o n surface for the d i r e c t runoff parameters, RS and RC, is r e l a t i v e l y flat, t h e r e is a c l e a r l y defined optimum. These parameters appear tk be f a i r l y stable when o t h e r parameter values are modified. The r a g e of r u n o f f percentages y i e l d e d by the FE and RC values obtained are comparable w i t h those i d e n t i f i e d in S e c t i o n 2.1. Figures 2 1 and 22, showing error function contours f o r percolation ERROR FUNCTION SURFACES FOR THE EVAPORATION PARAMETERS OF THE M O M 1 FIGURE 17 FIGURE 18 THE FS/SS SURFACE THE FS/DGf SURFACE I I 1.2 1.3 I rn FS 1.4 1.5 FIGURE 19 ERROR FUNCTION SURFACE SHOW l NG I NTERDEPENOENCE BETWEEN THE TWO EVAWRATJON REOUCT!ON FACTORS FIGURE 20 INTERDEPEND€NCE BETWEEN THE SURFACE RUNQFF GENERATION PARAMETERS (ROP= R C / ~ D C x R S ) ERROR FUNCTION SURFACES FOR VARIOUS GROUNDWATER PARAMETE~S 0.21- 0.27- GLR 0.25- 0.2%- I 1 49.54 1 53.54 1 1 1 57.54 61.54 65.54 GDM FIGURE 21 THE GLR/GDM SURFACE SHOWING INTERDEPENDENCE BETWEEN PARAMETERS IN THE PERCOLATION MODEL FIGURE 22 THE GF/GSM SURFACE SHOWS THE LOW SENSITIVITY OF THE GROUNDWATER RESERVOIR COEFFICIENT, GSM and groundwater storage parameters, are both relatively flat. In the former an optimum is r e l a t i v e l y clear; in the latter, the optimum of GF is clear, while the e r r o r function is less sensitive to GSM. A suitable value of GSM may be obtained from cansideration of well l e v e l and baseflow data such as that of Figure 13. The parameter GLR may correspond t o a physically based maximum rate of percolation through the upper hori zans . Prediction test Having obtained an optimum s e t of parameter values from t h e 1966- 1967 period of record (i.e. h d e l 5 ) , the model w a s used to pre&ct f l o w s given rainfall and evaporation, for 1968 and 1969. Indices of the goodness of fit are given in Table 4 and a comparative p l o t of discharges is presented in Figure 23. The results of such a prediction t e s t the u t i l i t y of t h e model, As the flow during 1968 and 1969 was cmsiderably more variable than in 1966 and 1967, as shown i n the f'no-m~del" variance terms, a f i n a l optimisatim of t h e d i r e c t runoff parameters RC, IE and DRK was performed an t h e t o t a l four y e a r s of record. The resulting values were 0.267, 0.026 and 0.152 respectively. As noted above, the RC and RS parameters remained stable ; however, the DFK parsmeter changed s i g p i f i c a n t l y changes in e f f i c i e n c y are shown in Table . Tbe corresponding 6. The model has predicted baseflaw in 1968 and I969 as w e l l as it did in 1966 and 1967. This is reassuring, as som earlier results (Section 2.5) suggested the possibility of a somewhat unstable groundwater system. As a consequence of an adequate basefluw fit, errors in volume remain fairly small. On the o t h e r hand, p r e d i c t i o n o f s, suitable distribution of d i r e c t runoff is d i f f i c u l t t o achieve, I t is worth n o t i n g t h a t the l o w e f f i c i e n c y obtained in 1968, and t h e change i n DRK resulting i n an improved efficiency r e l a t e t o a sin@.e major f l o ~ devent in September of t h a t year. The remainder of t h e record is i n s i m i f i c a n t l y affected. * Period of record Annual rTno-mdel,, v a r i an c e T e s t r u n I* Ann u d efficiency T e s t run 2%' Pe rcent age error in runoff volume Annul efficiency Percent age error in run off vol m e - 4.6 75.6 5.47 77.9 78. t + 8.1 78.1 - 2-95 + 9-10 15.83 67.2 - 2.0 75 4 - 13.05 73.9 + 3.7 74.0 + 3.83 1966 8-79 1967 1968 1969 . 1.58 "Model optimised an 1966-1967 record and u t i l i s e d for p r e d i c t i o n during 1968- 1969. **Par=ters reCOT d. RC, RS, and DFK optimised on the f o u r years of FIGURE 23 PREDICTED AND OBSERVED MONTHLY DISCHARGES FOR 1966 -1969 points lie within thew bounds) PREDICTED MONTHLY DISCHARGE, MMS 4. CONCLIEIOHS AND COMMENTS The Cam catchment has been observed to behave hydrologically as an e x t r e m e l y permeable area. Regions o f s u p e r f i c i a l Boulder Clay deposit and bare Chalk appear t o behaw in s u r p r i s i n g l y similar f a s h i o n s w i t h regard to t h e i r generation of d i r e c t runoff, storage of moisturn and a b i l i t y to recharge t h e groundwater system. The b e h a v i o u r of t h e s a t u r a t e d f1.m regime i n d i c a t e s t h a t much of the Chalk s t o r e s and transmits moisture r a t h e r uniformly, exhibiting a storage c o e f f i c i e n t of approlcimately 2.5 percent. The lumped conceptual model hypothesized for the catchment predicts the b e s e f l o w response both in volume and time d i s t r i b u t i m q u i t e adequately d u r i n g wet and dry periods. This r e s u l t is most encouraging for the u t i l i t y of t h e model for general water resources management. F u r t h e r , i f more d e t a i l e d s t u d i e s of t h e groundwater regime w e r e to be considered, the model might be. used t o generate a uniformly d i s t r i b u t e d recharge i n p u t . For t h e purpose of p r e d i c t i n g the time d i s t r i b u t i o n of f l o o d e v e n t s , the model is less s a t i s f a c t o r y . Whether problems in this area are associated w i t h the behaviour of t h e surface s t o r e component and/or the routing portion is n o t c l e a r l y understood. These two aspects of t h e model, also the m o s t d i f f i c u l t to i n t e r p r e t physically, warrant further thought and modification. I I - t I I t t I I I I t I I I I t I I t I slld IGYS, K.M. 19?0. J. Hfdrol., iainfaLl. 12, - X E F E I E N CE S 6. dlIDLnY, T.i.E. qJ A .epid nethod of coeputing areal 15-2\. DIctINsoN, W.T. and WHITS,EY, ir. 1970. water€heal ar:eas contributing to SUnp, on the tesults of reeea?ch on ?epresentatiue @rd etperinental b($'i|ts, tlel l,ington! N.2., vaL I, IASH/UNESCA,IASH PubL 96! 12-26, llrLoff, HUN1ER BLAIR, A. 1965. The hydiologr Queens Uni1€rsity of the River Cajo. M.Sc. Ihesis, of Belfast. M,A1',IDEYILLE r A.N., orCoNNIltL,F.f.,, 'UTCLITFI, J,V. srrd NASH,J.E. 19?0. niver flal. forecasting through coneeptuaf noalels. Part III - Ihe Rqy catchnent at Grendon Lhlderarood. J. Hldnol. 11, 1A9-128. IISHEffES lXD I00D. 1967. Pa tei tiaI, t?@rspi ta rt on. M .TISTRYOF AGRICUI.,TUnE, Tech.su1l. 16, HI4SO. NASH, J.E. moilels, snd SUTCLM'E, J.V- Parl I - e discussion 19?0. Ri!.er flotf forecastin8 of principles, J. IfAdvoT'' through conceptusl 10' 282-290. P.8., NASH,J.E. and IAXR!!L, J.P. 1970. Ri!.er flov folecasting OTCONNXLL, through colceptus,l rdodefs. Part II - the Srosna catchhent at I'erbsne. J. HUdt'aL., 10, 317-329, PENIM, H.L. 19118, Natural evaporation flon oI)en vater, P|oc. RoV. Soc, Lqnd. (D 193, Pa-1\r. SPnnKS,B.W., 195?. {he evofution of the relief Ceog". J., bare 6oi1 anal grass ' of the Caft Valley. 123, 188-207. STEEFS,J.A. 1965. fhe Cafuddge Region. Britis}] Association, caDbridge ( UniYercity Pr€ss), lltoMAssoN,A.J, 1969. soiL6 of the So,ff?da alden Dlsttict, Iiarpenalen. speciaf surv. 2, I t WOoDLAND, A.ht. 19t5. Ipslrich Ust?.ict. I'late" steplA fton undetgz,owtd ooLnces of, C@rby.Ld,ge Ceol., Sufv. Wa"tine palphlet 20. Further Reaalinc HoDGX,C.A.H. and SE.AIE, R.6. 1966. The eoils Men. SoiI Surv. Gt Br., I{arpentlen. of the diet?ict a"owd Cafu?iage. MrI{ISTRY0I't{OUSIlrc .Al,tDLoCAL GOVEXT{MENT. 1960. Rive" eteat juae HUdtolog'i&L Sun eU. H}'fC\ VJOFSSA.I4, B.C. snd lAYLOR, J.H. Inst. Geol. Sci.,l4ee. 1969. geol. Surv. Gt 5eolory of the colmtrV arolutd C@th?idge. Br., H!6o. I I I t t t I I I I I I I I I I I I