Water and cation movement in a tropical rainforest environment.
I. Objectives, experimental design and preliminary results.
S . N o r t c l l f f ( * ) J . B. T h o r n e s ( * * )
Abstract
The paper briefly outlines w o r k being done on t h e movement of cations under a Palmetum-type rain forest near M a n a u s , Amazonas. It is suggested that t h e l o w cation content of river w a t e r s I s concerned w i t h t h e nature and dynamics of w a t e r m o v e m e n t in relation to soil moisture and tension as much as lack of availa bility of cations in the s o i l . Preliminary results support the idea of a "by-pass' mechanism w h e r e b y the smaller pore spaces, saturated w i t h w a t e r , only drain and provide cation-rich w a t e r late in t h e dry season.
INTRODUCTION
A commonly presented paradox of t h e Amazonian tropical r a i n f o r e s t is that soils are excessively d e f i c i e n t in available n u t r i e n t s w h i l s t they support some of t h e m o s t dense, luxuriant f o r e s t in t h e w o r l d . Richards (1966) states that " I t can thus be seen that in mature soil t h e capital of plant n u t r i e n t s i s mainly locked up in t h e living vegetation and t h e humus layer, between w h i c h a v e r y nearly closed cycle is s e t u p " . This theme has been elaborated upon by Sioli (1966) and Stark
(1971). Stark described t h i s as t h e 'direct nutrient cycling hypothesis', w h e r e t h e nutri ents in t h e Amazonian f o r e s t are held mainly
in t h e organic phase, in l i v i n g roots, w o o d , bark, and leaves and in dead organic m a t t e r . There are considered t o be f e w n u t r i e n t s available in the heavily leached soils w h i c h are legarded as having high percentages of s i l i c a . Stark sup ports the hypothesis and suggests that t h i s state of affairs is of long standing because t h e deposits on w h i c h t h e present f o r e s t is de veloped w e r e originally nutrient d e f i c i e n t a l luvium exposed during t h e M i o c e n e . Leaching from heavy rains is thought t o have depleted
t h e soil nutrients f a s t e r than additions f r o m t h e atmosphere and by w e a t h e r i n g could de liver fresh s u p p l i e s .
A closer inspection of t h e literature reveals that statements concerning the n u t r i e n t status of t h e soil are based only rarely upon obser vation of t h e soil as a w h o l e . Rather, they t e n d to be based on ancillary measurements only tenuously linked t o soil s t a t u s . For example W e n t & Stark (1968) c o m m e n t on t h e d i s t r i bution of feeder r o o t s , essentially in t h e t o p 15 c m of predominantly organic material, together w i t h w h a t they describe as w h i t e o r grey s t e r i l e sands or poor y e l l o w clays, as indicative of a nutrient d e f i c i e n t s o i l . Further conclusions relating t o t h e poor status o f Amazonian soils have been d r a w n f r o m t h e analyses of stream and river samples con sidered t o be carrying t h e leaching products of t h e s o i l s ( S i o l i , 1968, 1975) . A s t h e solute concentrations of many Amazonian streams and rivers have been found t o be so l o w
( e . g . Edwards and Thornes, 1970), i t has been generally concluded t h a t t h e soils had l o w sources of leachable m a t e r i a l s . The arguments
underlying these conclusions together w i t h those f o r a 'nearly closed c y c l e ' are presented by Jordan & Kline ( 1 9 7 2 ) . O f t h e l i m i t e d number of soil analyses available f o r t h e region around Manaus, those of Brinkmann & Nascimento (1973) w e r e f o r only t h e t o p 20 cm of t h e t w o groups of s o i l s , y e l l o w latosols and hydromorphic s o i l s , b u t they concluded that t h e soils as a w h o l e generally have a low f e r t i l i t y . The only detailed descriptions and analyses of w h o l e soil p r o f i l e s readily available are those of IPEAN (1969) but these are f e w in number because of t h e very l o w sampling d e n s i t y .
( * ) — D e p a r t m e n t of Geography, King's College, Strand. London WC2R 2 L S .
It is w i t h t h i s background that the current study was i n i t i a t e d . The broad aim is t o ex amine in greater detail the n u t r i e n t movement through the soil s y s t e m , endeavouring if possi ble to complete some of the links in the sub system between inputs at the organic surface and output in the stream w a t e r s . To do t h i s , w e have adopted a s i m p l e model of spatial and seasonal variations of soil w a t e r concen trations and m o v e m e n t w h i c h is described in the f o l l o w i n g s e c t i o n . In turn t h i s w a s used t o develop an empirical design for the experiment which takes the f o r m of a detailed investigation of a slope segment, established as a f i e l d labo ratory, in a small c a t c h m e n t . The w a t e r movements into and t h r o u g h the slope segment have been m o n i t o r e d , and analyses of w a t e r quality w e r e made for incoming rainfall, the soil w a t e r in the saturated and unsautrated zones on the slope, and of the outgoing stream w a t e r . In addition detailed studies have been made of t h e soil over t h e whole slope length, though as yet no analytical results are available for t h i s part of the investi g a t i o n .
GENERAL M O D E L
The extent to w h i c h s o i l materials appear in stream channels relates to the relative solubility of the oxides or e l e m e n t s , t h e i r rela tive abundance in the soil and the movement of w a t e r through the s o i l , provided that the t i m e required for the equilibration is relatively short compared w i t h the residence t i m e of t h e water in the s o i l . As w e mentioned earlier, it is the second of t h e s e t h a t has generally dominated hypotheses about nutrient cycling in the humid tropical e n v i r o n m e n t . In this work the soils are t o be examined in detail for the relative availability of nutrients but ap preciable effort is also directed t o the amounts and conditions of w a t e r movement.
In a system of pore spaces, the order of evacuation of of pores w i t h drainage goes f r o m the relatively large t o the relatively f i n e pores. W i t h abundant w a t e r content, the finer pores may undergo l i t t l e activity
as i n f i l t r a t i n g w a t e r bypasses t h e m through the coarser part of the s y s t e m . This may
be t h e case in tropical soil s y s t e m s . As the f i n e r pore spaces are bypassed concen trations of elements are higher than in the coarser pore spaces and w e a t h e r i n g occurs at a lower r a t e . The pore s y s t e m may be likened to a cup w i t h ions d i s t r i b u t e d about the inner s u r f a c e . When the cup is f u l l , addition of f u r t h e r w a t e r leads to o v e r f l o w and there is turnover of w a t e r near the b r i m but not a t d e p t h . Concentrations near the brim are lower and decrease more rapidly w i t h t i m e than at the b o t t o m of the cup because of the higher frequency of e m p t y i n g and f i l l i n g ( F i g . 1 a ) . This could be regarded as the w e t season situation. In the dry season the cup is gradually e m p t i e d , o u t w a r d diffusion is less, because the concentration gradient is l o w e r and only w i t h large ' f l u s h i n g ' rains, towards the end of the season are the elements in the finer pore spaces evacuated. This occurs as a result of o v e r f l o w w i t h the f i r s t heavy r a i n s . Base saturated w a t e r at the b o t t o m of the cup is redistributed over the entire depth and t h i s o v e r f l o w s . During the w e t season, if only o v e r f l o w f r o m the cup is examined w e may expect concentrations t o be relatively l o w . Plants apply considerable stress and therefore can draw f r o m the b o t t o m of the c u p . To know w h a t elements are available to plants and in w h a t quantity it is not enough, t h e r e f o r e , t o examine only freely draining soil w a t e r .
In the profile as a w h o l e ( F i g . 1b) t h e movement of w a t e r depends on the d i s t r i b u t i o n of potential and the saturated and unsaturated c o n d u c t i v i t i e s . The latter are heavily de pendent on soil moisture w h i c h in turn is related to the pore space d i s t r i b u t i o n in t h e soil as w e l l as the recent history of i n f i l t r a t i o n into and drainage f r o m the s o i l . During the w e t season it is likely that the upper levels of the soil experience more rapid t u r n o v e r and higher f l o w rates than a t depth because of t h e natural d i s t r i b u t i o n of potential due to head (as opposed to matric potential) and, on steeper slopes, a greater tendency to lateral m o v e m e n t . The lower layers are close to s a t u r a t i o n , rates of movement are slow and turn-over of elements may be expected to be l e s s . Conse quently one may anticipate s i g n i f i c a n t vari ations in the vertical sense in t h e concen t r a t i o n s of elements in soil w a t e r . These
effects should also contrast seasonally being stronger in the w e t than in the dry s e a s o n .
Finally, significant variations in the concen trations of solutes may be expected on the hillslope scale ( F i g . 1c) on the basis of h i l l -slope hydrology. In the w e t season one might reasonably expect the floodplain sector t o be comparable t o the top of the c u p . Rapid and frequent response t o rainfall may lead t o a shunting to-and-fro of w a t e r b e t w e e n the
floodplain and the c h a n n e l . Further back, away f r o m the saturated wedge at the foot of the slope, groundwater m o v e m e n t is s l o w e r and dictated in part by receipts f r o m the overlying soil as w e l l as the gradient of the piezometric surface i t s e l f . An exception to t h i s occurs w h e n a groundwater ' h u m p ' moves downslope carrying w i t h it the recent ' o v e r f l o w ' w a t e r . Such humps may be expected to have ap preciably l o w e r concentrations than the w a t e r
(a) ' ' ; » ' . i , 1 1 . ' • i * i
+
+
•I+
+
+
(b)m i
k • •
vegetation layer saturated zone unsaturated saturatedbelow the seasonally stationary piezometric surface. A hump of this type is shown schematically in F i g . 1c. A g a i n , t h e concen trations of solutes in t h e stream channel may be a relatively poor indicator of their availa bility t o plants on t h e hillslope since t h e immediate response of t h e river, especially in the w e t season, may be largely d e t e r m i n e d by flood plain storage and r e s p o n s e .
These considerations led us t o t h e con clusion that a much better understanding of the relationship between soil and plants could be obtained via t h e soil w a t e r c o m p o s i t i o n . Contrary to being in direct opposition t o t h e closed nutrient cycle concept, w h e t h e r ex pressed in t e r m s of indirect or d i r e c t c y c l i n g , it might preclude t h e need f o r it, at least as far as certain elements are c o n c e r n e d . If this w e r e t r u e , more conventional processes of
nutrient supply could be envisioned and t h i s might have some implications for t h e manage ment of soil resources in t h e Manaus a r e a .
AREA O F S T U D Y
v
The f i e l d area chosen f o r the investigation is in the National Forest Reserve, Reserva Ducke, 26 km east of Manaus on t h e road t o Itacoatia-ra (Fig. 2). The area is underlain by Tertiary deposits of t h e Barreiras Series and is t h e centre of considerable research from the Ins tituto Nacional de Pesquisas da A m a z o n i a . This research provided a useful source of background i n f o r m a t i o n . In addition t o t h e general INPA research, closely related research by Dr. W . Franken (Max Planck Institute f o r Limnology) w i l l provide complementary data on the hydrochemical budget of t h e Barro Branco catchment as a w h o l e .
The research area has three major forest types, described by Brinkmann & Santos
(1973); (1) Riverine Forest (2) Carrasco Forest (3) Terra Firme Rain Forest. The experimental site w a s located in the second of these types w h i c h is intermediate between t h e other t w o . It has canopy heights f r o m 2 2 - 3 2 m , is quite heterogenous, w i t h an understory of numerous seedlings and saplings and has some herbaceous plants and p a l m s . On t h e slope segment studied the soils correspond broadly w i t h those found by Brinkmann and Santos that
is, y e l l o w and y e l l o w / b r o w n latosolic soils w i t h some sand throughout t h e p r o f i l e and
generally of m e d i u m porosity, w i t h gleyed sands at t h e foot of the s l o p e . Full details of our o w n studies of the plants and soils on t h e s i t e w i l l be published l a t e r .
A summary of t h e c l i m a t o l o g i c a l data f r o m Reserva Ducke has recently been published by Ribeiro (1975) In t h e period 1965-73 t h e range of temperature w a s f r o m 3 7 - 1 4 . 3 ° C . t h e average maximum daily temperature being about 33°C. A precipitation intensity of 6 0 m m / hour appears t o occur w i t h an annual return period, t h e w e t t e s t months being March and A p r i l (about 400 m m ) t h e driest being Sep tember w i t h about 5 0 m m .
E X P E R I M E N T A L D E S I G N A N D M E A S U R E M E N T
The basic design comprises a hillslope section f r o m the channel t o the local divide partitioned into six approximately equal segments and five d e p t h s . The design has been based on the need t o e s t i m a t e vertical and downslope variations in w a t e r m o v e m e n t and is s i m i l a r t o those t y p i c a l l y used in hillslope hydrology.
i) W A T E R M O V E M E N T. Following t h e ideas
of Freeze (1972) and Weyman (1970) t h e w a t e r fluxes have t o be e s t i m a t e d f r o m po tential differences, resulting from head and matric suction v a r i a t i o n s . The latter w e r e d e t e r m i n e d using a standard design of tensio-m e t e r (Webster, 1966) and cotensio-mpensating f o r tube size and depth in t h e usual f a s h i o n . Originally it w a s planned t o determine soil m o i s t u r e using a neutron probe, but due t o malfunctioning t h i s i n s t r u m e n t w a s not availa ble and soil m o i s t u r e had t o be e s t i m a t e d by conventional auger and g r a v i m e t r i c t e c h n i q u e s . The locations of t h e t e n s i o m e t e r s are shown in F i g . 2 The s i x sites and five depths w e r e also used as t h e basis f o r soil moisture d e t e r m i n a t i o n s . The t e n s i o m e t e r cups w e r e set at depths of 20, 36, 66, 96 and 120 c m below t h e surface and the t e n s i o m e t e r s w e r e in spected at f r e q u e n t intervals during and after storms (shortest interval, 10 m i n . ) but some storms occurring at night w e r e not o b s e r v e d . On the days of no rain t h e t e n s i o m e t e r s w e r e
read on a 2 hour basis. Because of the need to preserve the sites, augering could only be done on a w e e k l y basis, and t h i s has led t o some serious loss of d e f i n i t i o n of the charac teristic curve between tension and m o i s t u r e .
Steady state i n f i l t r a t i o n has been d e t e r m i ned using the standard double ring in-f i l t r o m e t e r w i t h a small constant head ( H i l l s ,
1970) under both saturated and unsaturated conditions for the underlying p r o f i l e . These w e r e carried out over prolonged periods and have resulted in c o n s i s t e n t results in t h e various e n v i r o n m e n t s . Run-off w a s measured from a natural piot four metres in length and two mestres w i d e . This was bounded by plastic edge to a depth of 15 cm into the s o i l , and measured at the outlet by a tipping bucket raingauge.
On the floodplain and to its hillslope edge a set of eight s i m p l e w e l l s w e r e installed for the measurement of the piezometric surface in that r e g i o n . These comprise plastic, 2 cm diameter tubes of 2 m l e n g t h . These w e r e measured four t i m e s daily and at more frequent intervals during and after major s t o r m s . Un fortunately all instruments w e r e not read all weekends because of problems of l o g i s t i c s . However since our concern is essentially on water movement this is not a serious problem since a storm-based sampling scheme is more m e a n i n g f u l .
ii) W A T E R C H E M I S T R Y. Water samples w e r e collected for the analysis of concentrations of the cations C a +X, M g + + , N a + and K + and
the samples were obtained f r o m stream-water, rainwater .unsaturated zone in the soil and w e l l s (saturated z o n e ) .
The stream w a t e r samples w e r e initially collected at hourly intervals during the day. w i t h more frequent sampling during and after storm events, but after analyses of the early samples it became evident that the stream water was relatively invariant both in storms and throughout the day. Subsequent stream water samples w e r e collected once or t w i c e per day w i t h some increase in sampling intensity during and after s t o r m e v e n t s .
The rain w a t e r samples w e r e analysed for one randomly selected rain gauge w i t h i n each of the five sampling blocks (see b e l o w for description) for seven collecting p e r i o d s . In
addition to the collections for chemical analy sis f i e l d observations w e r e made of the colour and particulate matter in the s a m p l e .
Soil w a t e r samples ( f r o m the unsaturated zone) w e r e collected at t w o positions w i t h i n the segment: ( F i g . 2 ) . Porous ceramic cup samplers w e r e used at depths of 23, 33, 63 and 102 cm for both s i t e s , w i t h additional samples at 180 cm at the lower site and 7 cm at the upper s i t e . The samplers c o n s i s t of a tube w h i c h can be evacuated to a known s u c t i o n , tipped w i t h a porous ceramic cup w h i c h takes in the soil w a t e r . A standard suction of 600 cm was established and the samplers left for approximately 180 minutes (exact t i m e s w e r e noted) before the accumu lated samples w e r e e x t r a c t e d . Full recognition w a s taken of the c o m m e n t s of Hansen & Harris
(1973), but it w a s considered that a standard suction was more apropriate for testing the proposed model than a variable suction related to the ambient matric s u c t i o n . The soil w a t e r samples obtained by the porous cup method should be more representative of the soil solution than those obtained by l y s i m e t e r s or zero tension lysimeters (Jordan. 1968) for the reasons outlined earlier.
Well water samples w e r e extracted f r o m w e l l s located in the flat, frequently inundated area at the foot of the slope s e g m e n t . Three w e l l s w e r e sampled: Wl — located d i r e c t l y at the foot of the experimental slope segment, some 2 m f r o m the break of slope, WF4 and WF5. located a small distance upstream on the same broad flat area at the base of the slope. During the period of sampling, the local groundwater level w a s above the b o t t o m of the sampling tubes ( w e l l s ) , consequently these samples are representative of the soil moisture in the saturated z o n e ,
iii) I N T E R C E P T I O N. Interception of rainfall by
t h e forest canopy has long been recognised as important in the overall hydrological balance of tropical rainforests (Richards, 1966) . Clegg 1963 investigated interception amounts in relation to forest s t r u c t u r e and storm i n t e n s i t y . In this study an a t t e m p t was made to estimate the relative amounts of i n t e r c e p t i o n in the
experimental site by randomly locating four standard raingauges (50 cm above the ground) w i t h i n each of five blocks located on the slope
parallel t o that used f o r t e n s i o m e t e r measure m e n t s . The f i v e blocks w e r e approximately 10m w i d e a n d broadly correspond t o positions on t h e main slope in t h e f o l l o w i n g m a n n e r : A — b e t w e e n t h e stream and t e n s i o m e t e r site I, B — b e t w e e n t e n s i o m e t e r sites I and III, C — between t e n s i o m e t e r sites III and IV, D between t e n s i o m e t e r sites IV and V and E between t e n s i o m e t e r s i t e s V and V I . A f t e r each rainfall collection the gauges w e r e re located r a n d o m l y . The rainfall received at each site w a s expressed as a percentage of t h e open site r a i n f a l l .
iv) R U N O F F. Dr. Franken has kindly p e r m i t t e d access t o his run-off data for t h e Barro Branco for t h e period in q u e s t i o n . These are based on stage records f o r a broad-crested w e i r and we w e r e able t o rate these using t h e O t t current m e t e r . The recording station is about 150m d o w n s t r e a m f r o m our o w n site and stages there correlate perfectly w i t h those a t the main gauging s t a t i o n .
v) C L I M A T O L O G I C A L D A T A. A full range of
climatologlcal data, but especially rainfall data is recorded at the Reserva Ducke and w e have been able t o abstract this data f o r t h e observation p e r i o d . Rainfall is measured by standard tipping bucket raingauge w i t h a 24 hour chart w h i c h is changed at 0700 hours. From this rainfall totals f o r 10 minute periods for t h e w h o l e period of observation (3 months) have been o b t a i n e d .
PRELIMINARY R E S U L T S
(i) H Y D R O L O G Y
We present here a f e w general results in the belief that they w i l l be of immediate
interest t o c o w o r k e r s , especially those at INPA. They are intended as a background t o a more detailed analysis and separate papers w h i c h w i l l f o l l o w l a t e r . During t h e peroid of observation, f r o m M a r c h t o late M a y , 1977 there w a s a gradual shift f r o m very w e t t o relatively much drier conditions, w i t h associ ated trends in soil tension and runoff. The overall period w a s marked by four very pro nounced s t o r m s , on t h e 23rd M a r c h , on 3rd A p r i l , 15-16th A p r i l and again on t h e 6th M a y .
These rainfalls ( f i g . 3a) provided important pulsed inputs, the earliest into very w e t s o i l ,
the latest into relatively dry s o i l . In f i g . 3b w e indicate three storms of comparable volume but of quite d i f f e r e n t t i m e d i s t r i b u t i o n s . Early during the observation period sotrms w e r e more sustained and of generally lower intensi t y than during t h e latter part of t h e p e r i o d .
Throughout, w e w e r e surprised at t h e very sharp response of t h e Barro Branco to r a i n f a l l :
lag t o peak t i m e s w e r e almost invariably w i t h i n an hour of t h e cessation of rainfall and t i m e t o rise w a s almost instantaneous ( e . g . Fig. 3 c ) . We consider these results to indicate the very fast response of the saturated floodplain areas both here and upstream f o r they coincide w i t h extensive inundation of the f l o o d p l a i n . The recession limbs of the hydro-graphs w e r e more varied and a f u l l e r analysis by Dr. Franken w i l l appear later. However most revealed a considerable attenuation w h i c h m i g h t be attributable t o s l o w e r drainage of t h e lower hillslopes or at least of the floodplain areas. A l t h o u g h t h e w e l l data has yet t o be plotted, it is evident that towards the back of the floocpalin t h e w e l l s s h o w e d a rapid rise and fall throughout individual s t o r m s . This probably reflects the g r o w t h and decay of a saturated wedge at t h e slope base after t h e fashion described by Weyman (1970) and o t h e r s . The rapid response of this zone may, itself, be a further indication of t h e by-pass m e c h a n i s m .
During t h e entire period t h e overall drying produced a corresponding s h i f t of t h e mean tensions w i t h i n t h e p r o f i l e s . ( F i g . 4 ) . Site 1 was almost continuously saturated at all depths below 20 c m so that t h e diagram of average matric suction ( F i g . 4) shows that positive pressures, increasing w i t h depth as a result of hydrostatic head, prevailed through out t h e e n t i r e p e r i o d . Site 2 shows the influ ence of p r o x i m i t y t o t h e piezometric surface during a w e t period ( 1 1 - 1 6 / 0 5 / 7 7 ) . The otner sites 3 and 5 show almost vertical matric suction p r o f i l e s , indicating that the main driving f o r c e in w a t e r movement is differences in head p o t e n t i a l . This can largely be attributed to t h e relatively high moisture c o n t e n t . In sites 4 and 6 however,, the latest period shows strong evidence of t h e effects of drainage.
Fig. 5 shows t h e d i s t r i b u t i o n of total hydraulic potential w i t h reference t o the t o p
70 60 SO E 40 E 30
Î
2 0I
10 12 10 E 8 E 6 -Ii
March April 10 -- 1 1r-16
14- 121 0 -E 8" E6
I1
2"
(X 0 6 0 0 1200 1800 3 1 0 230Ï
150 a. O ls
May B 1 2 0 - 0 3 - 7 7 lotal 56.8mm - 1 1 r B 2 0 3 - 0 4 - 7 7 Total 68.4 mm - i — î — î — î î B 3 0 6 - 0 5 - 7 7 Total 64.1 mm 2 4 0 0r-J
• t .
1 1 1 60 5 0 40 3 3 0 § 0 4 - 0 5 - 7 7 0 5 - 0 5 - 7 7 0 6 - 0 5 - 7 7 0 7 - 0 5 - 7 7 0 8 - 0 5 - 7 7Fig. 3 — (a) Plot of daily rainfall totals during period of observation, (b) Plot of hourly rainfall totals for three different storms to show distribution t y p e s , (c) Hydrograph for storm of 6th June 1977. Barro Branco.
5 0 Pressure —i | | i i i i • i • Tension 5 0 i i 1 1 1 1 — i — i — i — i — i — i — - 1 0 0 SZ ' SITE 1 A 100 150 A 1 9 - 2 3 / 0 3 / 7 7 B 1 1 - 1 6 / 0 5 / 7 7 C 1 9 - 2 7 / 0 5 / 7 7 SITE 2 B C t Tension H 1 1 — 50 + ioo 4-50 100 —I 1 1 1 1 1 1 1 1 h
t
+ TensionS
i Q 50 + 100 4-SITE 3 A B C 50 100 —I 1 1 1 1 1 — 150 SITE 4 T Tension 50 100 4-50 100 H 1 1 1 1 1 h 150Fig. 4 — Distribution of matric potential at d e p t h . Averages calculated for selected p e r i o d s . Horizontal bars show one standard deviation.
profile indicating that during periods A and B the potential w a s d i s t r i b u t e d more or less according to head ( d e p t h ) , the matric potential being constant. In period C the effects of drainage in producing high matric potentials in the upper horizons are much more evident. A preliminary inspection of the moisture data reveals similar f a i r l y u n i f o r m p r o f i l e s . The tension-moisture curves are a l m o s t v e r t i cal near the saturation l i m i t , there is a w i d e range of tension values for moisture contents near 2 7 c m3/ c m3, the average s a t u r a t i o n . These
characteristics are t o be expected of f a i r l y permeable m a t e r i a l . Steady i n f i l t r a t i o n rates ranged from 0 . 0 9 7 6 m m / s e c on the latosols to 0.0384 for the f l o o d p l a i n . Average rates on the root mat at the top of the soils w a s 0.3152 m m / s e c w h i c h is very h i g h . A comparable order of magnitude is observed for migration of the w e t t i n g f r o n t under saturating con d i t i o n s . Regression of the data for t i m e s w h e n the soil was already very w e t yield an average of 0.075 m m / s e c . so that typically a f t e r a lag of up to 45 minutes for saturation of the rootmat, i n f i l t r a t i o n into the latosols w a s very s w i f t . This, together w i t h the very fast river response, appear to c o n f i r m the by-pass mechanism described earlier and should help to explain the low concentrations of river water mentioned e a r l i e r .
ii) C H E M I C A L A N A L Y S E S
Water samples w e r e c o l l e c t e d during each week and analysed at the end of each w e e k using a Perkins Elmer A t o m i c A b s o r p t i o n Spectrophotometer (AAS) . Selected examples were examined after longer storage periods and showed no significant changes in recorded concentrations. At this stage, results pre sented are mean values, w i t h some intro ductory c o m m e n t s on the variance character istics of the t h r o u g h f a l l , further analyses of variance characteristics and patterns related to climatic conditions and antecedent soil moisture conditions w i l l be presented later. a. River Samples. The river samples col lected throughout the period show exception-nally low concentrations in respect of all f o u r cations examined (Table 1 ) . The mean value for Ca + + w a s 0.01 m g / 1 , w i t h a m a x i m u m
value of 0.04 and m i n i m u m of 0 . 0 0 . Given the s e n s i t i v i t y of the A A S , it may be more correct to consider all these values as zero or very close to zero. The mean value for fvig + + is slightly higher at 0.03 m g / 1 , w i t h a range from 0.10 to 0.00. but as w i t h the Ca + + a relatively low variance. Further analysis of the changes in M g + + is progress in relation to stream discharge and long and short t e r m antecedent climatic c o n d i t i o n s . N a + concentrations are considerably higher than for the other three cations, but are low w h e n compared w i t h published r e s u l t s . The mean concentration for N a + during the period w a s 0 . 1 9 m g / 1 w i t h a range f r o m 0.48 t o 0 . 0 0 . Values for K + are again very low w i t h a mean of 0 0 6 m g / 1 and a range of 0.45 t o 0 . 0 0 .
W i t h the exception of the low mean for N a + these results agree in general w i t h those obtained for the same catchment by Dr. Franken and are w i t h i n the range of values presented in other studies (see Sioli. 1968) . The intensive sampling at the start of the observation period both during a day and throughout s t o r m s suggested no clear patterns, but further analysis is in progress to investi gate relationships at longer t i m e s c a l e s .
b. Throughfall Samples. Analyses of through-fall samples show markedly higher concen trations for all four cations than those obtained for stream w a t e r samples (Table 1 ) . The overall mean concentration for Ca-*- + is 0.32 m g / 1 w i t h a range of 0.88 to 0.10. M g + + as 0.20mg/1 w i t h a range of 0.40 to 0 . 0 4 . N a + as 0 . 4 2 m g / 1 w i t h a range of 1.58 t o 0 . 0 5 , and K + as 1 3 3 m g / 1 w i t h a range of 7.47 to 0.22. Preliminary tabulation of mean concentrations by rainfall events and by blocks foi all rainfall events is given in Table I I .
TABLE I — M e a n Chemical Results : Barro Branco and Rainfall. M a r c h — M a y 1977 ( m g / 1 )
C a + + M g + + N a + K +
Barro Branco 0 . 0 1 0 . 0 3 0 . 1 9 0 . 0 6
Total hydraulic potential considered from top of profile (150cms)
50 100 150 200 — i 1 1 1 1 1 1 1 1 i i i i\
A 1 9 - 2 3 / 0 3 / 7 7 B 11 - 1 6 / 0 5 / 7 7 C 1 9 - 2 7 / 0 5 / 7 7Fig. 5 — Distribution of total hydraulic potential with reference to top of profile.
100 -50 • • Site 2 a Site 3 • Site 4 d Site 5 • ^ y'u m Site 6 ^ a CD Q I — i — i — i — i — i — i — i — i i — i — i — i — i — i — i — i i i i 1 1 1 1 1
50 100 150 2 0 0 Time
Fig. 6 — Travel times for w e t t i n g fronts plotted against depth; the regression line is computed only for the cases w h e r e the soil was initially at or very close to saturation.
TABLE II — M e a n s for chemical analysis of throughfall samples ( m g / 1 ) (1) By Events Date Ca M g Na K 220477 0.33 0 . 1 4 0 . 4 0 1.08 250477 0 . 5 9 0 . 3 0 0 . 5 9 1 81 280477 0.40 0 . 2 5 0 . 4 5 1.52 060577 0.13 0 . 0 7 0 . 1 5 0 . 8 4 110577 0 . 2 8 0 . 1 6 0 . 2 0 0 . 6 7 160577 0.24 0.23 0 . 7 1 2 . 2 9 270577 0 . 2 6 0 . 2 1 0 . 4 3 1.10 (2) By Blocks Block Ca M g Na K A 0 . 3 5 0 . 1 7 0 . 3 5 1.22 B 0 . 3 5 0 . 1 7 0 . 2 0 0.81 c 0 . 3 7 0 . 1 8 0 . 2 3 0 . 9 3 D 0 . 2 8 0 . 2 6 0 . 7 8 2 . 7 7 E 0 . 2 5 0 . 1 9 0 . 5 3 0 . 9 2 Overall M e a n 0 . 3 2 0 . 2 0 0 . 4 2 1.33
and K + , Lower Site 33 and 6 3 c m . The relation ship between depth and concentration are d i f f e r e n t for the t w o sites and w i t h respect to d i f f e r e n t c a t i o n s . Further analyses relating soil w a t e r concentration to antecedent rainfall and matric suction are in progress to elucidate any r e l a t i o n s h i p s . One s i g n i f i c a n t feature of the results is w i t h respect to the 180cm sample at the Lower Site w h i c h , f r o m f i e l d obser vations, appeared to be sampling the saturated zone during the period of observation, but the concentrations are c o n s i s t e n t l y higher than those obtained for the W e l l Samples (Table V I ) , also considered to be sampling the satu rated zone.
Means for all samples and for the four sample depths common to both sites are given in Table V . Taking results for all samples, there are d i s t i n c t differences w i t h respect to C a + + and M g + + for the t w o s i t e s , w i t h c o n s i s t e n t l y higher values at the Lower S i t e . When the samples f r o m 180cm (Lower Site)
Initial investigations to elucidate patterns in the data has been undertaken using a 2-way analysis of variance model w i t h o u t r e p l i c a t i o n . The sources of variation incorporated in the model w e r e (i) the s t o r m e f f e c t (variations between throughfall c o l l e c t i n g periods for all b l o c k s ) , (ii) the block effect (variations between blocks for all c o l l e c t i n g periods) and
(iii) the interaction e f f e c t . Mean squares are computed for each of the effects and t h e existence of s t o r m or block effects are t e s t e d by dividing their mean squares by the inter action mean squares and comparing w i t h the F-distribution for the appropriate degrees of f r e e d o m . Initial results are given in Table III, where the significance level of the e f f e c t is listed ( N . S . indicates not s i g n i f i c a n t ) . These results suggest that C a + + and M g + + inputs vary in relation to s t o r m intensity and duration w h i l e Na-^- and K + inputs are more strongly influenced by spatial location, probably, as a result of variations in the vegetative cover, c. Soil Samplers. The preliminary results for samples collected f r o m the soil samplers are given in Table IV, by site and d e p t h . In general the" values are higher than for the throughfall and stream samples, the through-fall exceptions being C a + + , Lower Site 33cm
TABLE III — Summary of analysis of variance : throughfall chemical d a t a .
Cation Blocks Storms
C a — NS 0 . 0 0 1
M g * * NS 0 . 0 0 1
N a * 0.001 0 . 0 1
K * 0 . 0 1 NS
TABLE IV — M e a n values of chemical data by depths for soil s a m p l e r s , ( m g / 1 ) Uper Site Depths Ca M g N a K 7 cm 0 . 3 4 0 . 4 7 6 . 0 0 2 . 8 5 23 c m 0 . 8 5 0 . 7 5 2 . 6 8 1 3 7 33 cm 0 . 6 8 0 . 4 4 2 . 6 3 1 3 2 63 c m 1.08 0 . 6 6 3.54 2 . 3 9 102 c m 0 . 5 7 0 . 4 8 5 . 4 7 2 . 2 8 Lower Site Depths C a M g N a K 23 cm 0 . 4 9 0 . 5 3 1.13 1.30 33 c m 0 . 2 5 0 . 3 0 0 . 5 7 0 . 6 9 63 cm 0.56 0.21 0 . 6 5 0 85 102 c m 0.84 0 63 1.96 1.53 180 cm 11.45 3 5 . 5 1 3 . 1 3 4 . 8 7
TABLE V — M e a n C o n c e n t r a t i o n s : Soil S a m p l e r s . ( m g / 1 )
C a + + M g + + N a + + K +
All Samples : Uper Ü 77 0 57 3 72 1 .90 Lower 1 64 4 05 1 26 1 46 Four Depths : Uper 0. 78 0 58 3 63 1 86 (23. 33. 63, 102) Lower 0 50 0 40 1 04 1 05
TABLE VI — M e a n values of chemical data for w e l l s a m p l e s , ( m g / 1 ) C a M g N a K W1 0 . 3 1 0 . 1 1 2 . 8 3 1.85 WF4 0 . 8 4 0 . 0 8 0 . 7 4 0.86 WF5 0 . 4 8 0 . 0 7 0 . 6 1 0 . 9 4 All Wells 0 . 5 5 0 . 0 9 1.33 1.19
TABLE VII — Raw Interception data
Date Time Off Rf. m m % Rain Received
2 9 . 0 3 0820 8 . 0 7 3 . 0 % 3 1 . 0 3 0833 3 5 . 5 8 0 . 5 % 0 4 . 0 4 1200 6 9 . 7 8 1 . 9 % 06.04 1200 3 . 3 1 0 7 . 0 % 11.04 0825 2 7 . 5 9 6 % 11.04 1505 5 . 0 4 2 . 8 % 12.04 1005 2 . 2 3 5 % 14.04 0850 19.65 7 5 % 14.04 1530 4 . 4 5 5 2 . 7 % 20.04 0840 142.2 6 7 % 20.04 1400 8 . 4 8 2 . 1 % 2 2 . 0 4 1500 1 3 . 9 6 4 . 0 % 25.04 1200 16.7 9 4 . 1 % 27.04 1035 5 . 4 5 3 . 5 % 28.04 1200 1 0 . 3 2 7 3 . 4 % 02.05 0830 3 9 . 6 9 9 . 5 % 03.05 1455 3 . 2 5 5 8 . 5 % 06.05 1300 6 2 . 3 7 8 . 9 % 09.05 0940 2 4 . 6 8 2 . 4 % 09.05 1340 2 . 5 7 2 . 2 % 10.05 1515 7 . 1 8 4 . 4 % 11.05 1535 12.7 7 5 % 13.05 0950 6 . 0 8 2 . 7 % 16.05 1400 2 7 . 4 5 1 0 2 . 9 % 24.05 0950 10.45 4 5 . 0 % 26.05 0820 4 . 2 8 7 2 . 9 % 27.05 0840 1 0 . 1 3 8 5 . 7 % Mean Rainfall received 7 4 . 7 % .
and 7cm (Upper Site) are not considered, the pattern is r e v e r s e d . W i t h respect to both Na + and K + the Top Site has higher average concentrations for both data s e t s .
d . Well Samples. The mean concentrations f o r w e l l samples are given in Table V I . The values obtained f r o m the three w e l l s are w i t h i n the range of values obtained for the soil samplers w i t h the exception of M g + + , w h i c h has recorded concentrations w i t h i n the range of values obtained for incoming rainfall and outgoing s t r e a m w a t e r . A n a l y s i s is in progress to relate the concentration at W l to the t h r o u g h f l o w of w a t e r f r o m the slope segment, and at all sites to changing stream discharge c o n d i t i o n s .
Mi) I N T E R C E P T I O N
Interception results are presented in Table VII and VIII, in t e r m s of the percentage of open site rainfall received at the f o r e s t s i t e s . Table VII gives the summary results for each event (means of 20 g a u g e s ) , w i t h detailed results for selected rainfall events presented in Table V I I I . Initial analysis of the variation in rain
TABLE VIII — Percentage rain received by Blocks
1104 1505 5.0 m m A 1 8 . 5 % B 4 3 . 0 % C 2 5 . 5 % D 4 6 . 0 % E 8 1 . 0 % 2504 1200 16 7 m m A 9 2 . 1 % B 2 3 0 . 7 % C 4 9 . 4 % D 6 1 . 5 % E 3 6 . 8 % 0905 0940 24.6 mm A 1 2 3 . 4 % B 8 9 . 5 % C 8 8 . 2 % D 5 0 . 2 % E 6 0 . 9 % 0605 1300 62.3 m m A 7 5 . 2 % B 7 3 . 1 % C 1 3 4 . 4 % D 4 6 . 6 % E 6 5 . 3 %
TABLE IX — Percentage of rainfall received related to rainfall amounts (mean for all events)
Rainfall % Rain received
0 — 10 m m 6 8 . 1
10 — 20 m m 7 2 . 9
20 m m + 8 6 . 0
received suggests a broad relationship w i t h rainfall amounts as indicated in Table IX, intensity and d u r a t i o n . A two-way analysis of variance t e s t has been undertaken t o investi-gate t h e existence of s t o r m and block e f f e c t s on the i n t e r c e p t i o n v a l u e s . The e f f e c t s w e r e found t o be not s i g n i f i c a n t .
ACKNOWLEDGEMENTS
This work is being carried o u t under a grant f r o m The Royal S o c i e t y . During our f i e l d work w e w e r e greatly assisted by t h e f a c i l i t i e s provided by t h e Instituto Nacional de Pesquisas da Amazonia both in Manaus and at Reserva Ducke. Dr. W . Franken kindly provided access to the stage records f o r t h e Barro Branco and Dona Nazaré G . Ribeiro provided access t o unpublished rainfall d a t a .
Resumo
Relata-se um trabalho feito sobre o movimento d e cations sob uma floresta úmida do tipo " P a l m e t u m " perto de Manaus, A m a z o n a s . Sugere-se que o baixo teor de cátions das águas dos rios está ligado à natu-reza e dinâmica do movimento da água e m relação ao teor e tensão da umidade do solo, assim como à falta de disponibilidade de cátions no solo. Os resultados preliminares apoiam a hipótese de haver um mecanis-mo tipo "by pass" pelo qual os micropóros, saturados com água, somente drenam e fornecem água rica de cátions no fim da estação s e c a .
BIBLIOGRAFIA C I T A D A
BRINKMANN, W . L . F . & NASCIMENTO, J . C .
1973 — The e f f e c t of slash and burn agriculture on plant nutrients in t h e Tertiary region of central Amazonia. Acta Amazonia, 3(1) : 55-61.
BRINKMANN, W-. L . F . & SANTO:, A .
1973 — Natural w a t e r s in A m a z o n i a . VI Soluble calcium p r o p e r t i e s . A c t a Amazônica, 3(24 : 33-40.
CLECG. A C .
1963 — Rainfall interception in a Tropical Forest Caribbean Forester, 24(2) : 75-79.
HOWARDS, A . M . C . & THORNES, J . B
1970 — Observations on the dissolved solids of the Casiquiare and Upper Orinoco. April-June, 1968. Amazoniana, 11(3) : 245-256.
FREEZE. R . A .
1972 — The role of subsurface flow in generating surface runoff. 2 . Upstream source areas. W a t e r Resources R e s . , 8. 5. 1272-1283. HANSEN. E . A . & HARRIS, A . R
1975 — Validity of soil w a t e r samples collected w i t h porous ceramic cups. Proc S o i l . S o c . Soc A m e r . , 3 9 ( 2 ) : 528-536
HILLS, R . C .
1970 — The determination of the infiltration capacity of field soils using the cylinder infiltrometer. Technical Bulletin, 3 British Geomorphologi-cal R e s . Group, Geo Abstracts. N o r w i c h . I P E A N
1969 — Os Solos da Area Manaus • Itacoatiara JORDAN, C . F .
1968 — A simple tension-free l i y s i m e t e r . Soil S c i . , 105 : 81-86.
JORDAN. C . F . & KLINE, J . R .
1972 — Mineral cycling : Some basic concepts and their application in a tropical rain f o r e s t . A n n . R e v . Ecol. S y s t . ( 3 ) : 33-51.
RIBEIRO, M . DE N . G .
1976 — Aspectos climatológicos d e M a n a u s . Acta Amazónica. 6(2) : 229-233.
RICHARDS, P. W .
1966 — The Tropical Rain Forest. Cambridge U n i v . Press, Cambridge
SIOLI, H .
1966 — General features of the delta of t h e Amazon. In : U N E S C O , Scientific problems of t h e humid tropical zone deltas and their impli-cations : Dacca Symposium, p. 381-390. 1968 — Hydrochemistry and Geology in t h e Brazilian
Amazon Region. Amazoniana, 1(3) : 267-277 1975 — Tropical rivers as expressions of their
terrestial environments in : Tropical Ecologi-cal S y s t e m s , E d . F. B. GOLLEY a n d . E ME-D I N A . Springer V e l a g , N e w Y o r k .
STARK. N .
1971 — Nutrient cycling : I Nutrient distribution In some Amazonian soils. Tropical Ecology,
12(1) : 24-50. WEBSTER, R .
1966 — The measurement of soil w a t e r tension in the f i e l d . N e w Phytologist. 65 : 249-258. W E N T . F . N . & STARK. N .
1968 — M y c o r r h i z a . Bioscience, 1 8 ( 1 1 ) : 1035-1039. WEVMAN, D . R .
1970 — Throughflow on hillslopes and its relation to the s t r e a m hydrograph. Bull. I n t . A s s o c . S c i . H y d r o l . , 15(2) : 25 33
