proudly sponsored by
|
archiving of SOWACS proudly sponsored by |
|
NOTE: To get off this list, send email to majordomo@aqua.ccwr.ac.za with the body of the message containing the line: unsubscribe sowacs James Sorry for this late reply. In some of the later papers on fingering flow there were instances of this "fingering" flow developing in relatively uniform sandy soil profiles. Papers were in "Journal Soil Sci Soc Amer" and "Soil Science" around late '80's I think. Perhaps it might be worth following up to see if you soil texture is similarly coarse? regards Martin owner-sowacs@aqua.ccwr.ac.za wrote: > NOTE: To get off this list, send email to majordomo@aqua.ccwr.ac.za > with the body of the message containing the line: > unsubscribe sowacs > > Martin, > > Thanks a lot for your comments. The soil in question is very uniform down > the profile, all the way to the declining water table at 2.5-3.5 > metres. pF curves throughout the profile also indicate a uniform water > holding capacity, so I don't think a difference in texture is the cause of > the problem? > > James > > At 10:11 4/3/01 +0000, you wrote: > > >Hello James > > > >Is this a possible explanation? What you are seeing is simply an interface > >effect, > >at a sharp change of soil texture. There is a good diagram of this effect > >in Don > >Nielsen's book "Soil Water", taken from the paper by Srinilta et al > >(1969). Your > >capillary fringe is not indicating upway movement. It is the expected > >potential > >gradient above a water table within the lower coarse textured layer. > > > >There is a more detailed description of this phenomena given beneath. It > shows > >that you should read the literature written more than 30 years ago! > > > >best regards > > > >Martin Parkes > >Wuhan University (WUHEE) > >8 Dongh Nanlu, Wuchang > >430072 WUHAN > >P R CHINA > > > >5 Flow in layered media > > > >Review of profile development (Interface effects) > > > >Anomalous behaviour occurs in laboratory measurements, when there is > >unsaturated > >flow from fine to coarse pored media. Measurements and analysis of > >unsaturated > >flow from fine to coarse-pored soil media afford some insights. Alway and > Dole > >(1917) showed that a layer of sand beneath a fine textured soil restricted > >drainage from above. Similarly Day and Luthin (1953) found that soil > >moisture in > >the fine material at the interface of layered soil was held under suction, > >despite > >ponded conditions above. The coarser the underlying material, the greater > >is the > >interface tension and the greater the fall in infiltration rate through the > >overlying material (Miller and Gardner, 1962). So through-flow is dictated > >by the > >conductivity at the interface, depending in turn on the coarseness of the > >underlying material and it's lower boundary condition. Miller and Bunger > >(1963) > >found that water, retained in a 0.6 m soil depth by restricted drainage, was > >independent of the thickness of the underlying sand strata, for a > thickness of > >0.076-0.91 m. The hydraulic gradient across the interface changed after one > >day, > >from unit hydraulic gradient promoting downward flow to zero gradient after > >2-3 > >days then to a positive gradient indicating upward flow after 6 days. > >During study > >of drainage from saturation of a silt loam to sand sequence, Eagleman and > >Jamieson > >(1962) recorded interface hydraulic gradients of -0.25, -1.25, +1.2, +1.0, > >-1.0,-1.0 for 1,7,14,19,22 and 26 days respectively following drainage from > >saturation. Estimates of hydraulic conductivity at these times were 0.95, > >0.07, > >0.01, 0.01, 0.001, 0.001 cm/day respectively, assuming unit negative > hydraulic > >gradients in place of positive gradients. > > > > Takagi (1960) and Zaslavasky (1964) analysed steady flow behaviour > >for such > >a sequence. Takagi considered a two-layer sequence with water ponded on the > >surface of the fine pored media and with water flowing through the coarse > >pored > >media to a stationary water table. He showed that if a zone of constant > >negative > >pressure (suction) exists within the profile then it must begin at the > >interface > >between the 2 layers. Furthermore a transition zone of varying moisture > >content > >must exist in the upper fine-pored material which changes from saturated > >conditions to an unsaturated condition at the interface. A similar > >transition zone > >must also exist within the coarse pored material between the zone of constant > >negative pressure and the water table. As the depth of fixed water table in > >reduced so the zone of constant negative pressure is eliminated and the > >profile > >has a saturation zone above two transition zones, which intersect at the fine > >coarse interface. Zaslavsky argued that steady flow could not be preserved > for > >such a flow sequence unless atmospheric pressure is introduced at the > >interface. > > > > Srinilta et al (1969) undertook laboratory investigations to examine > the > >predictions of Takagi and Zaslavsky. They confirmed many of the features of > >earlier experimental work. Specifically through-flow was sensitive to topsoil > >thickness, with increasing thickness reducing flow and making the interface > >tension slightly more negative. A reduction of 0.25 to 0.05 m depth of > ponding > >from reduced the through-flow 20-30%, with little change in the interface > >tension. > >Water flux values were un-affected by subsoil thickness of 0.62 to 0.92 m, > >though > >the interface pressure appeared to drop about 5 cm. An important > >observation was a > >4 fold reduction of through flow, when the lower boundary condition of the > >lower > >layer was increased from 0.45 m to 1.05 m positive pressure. > > > >Description of vertical flow within a fine/coarse sequence > > > > The eventual steady state condition for ponded flow through a > > fine/coarse > >soil sequence depends on the lower boundary condition of the lower layer. > >Consider > >this to be equivalent to a fixed water table so that the final boundary > >condition > >will be zero tension. With no flow, above the water table the soil water will > >exist in a condition of zero hydraulic gradient. So soil tension will > increase > >with increasing height above the water table. With a lower soil depth of > >1.20 m > >then the soil tension at the interface will be 1.20 m of water. However, > >initially > >the lower layer is considered uniformly dried to a moisture content > >corresponding > >to 1 bar tension or more. > > > > Imagine that the upper 0.3 m layer of fine soil is also dried to a > >similar > >moisture content before infiltration. Water ponded on the surface begins to > >infiltrate so that a wetting front moves down to the interface. The wetting > >front > >is temporarily held up at the interface as water accumulates in the upper > >layer, > >so reducing the tension to a point were the coarse pores fill and begin to > >conduct > >water. During infiltration all air is not entirely displaced from the soils. > >Entrapped air contains water vapour at vapour pressures dictated by the > >radius of > >curvature of the air/vapour/water interface. > > > > In moving from a fine to a coarse layer the radius of curvature of the > >air/vapour water interface must change from small to large. Consequently a > >vapour > >pressure gradient exists across the interface which could assist flow of > >water at > >the wetting front. Vapour will only condense to assist the liquid flow > >provided > >sufficient tension exists in wetted soil behind the wetting front. If > >conditions > >exist to set up such a vapour flow then it is maintained by water vaporising > >within fine pores above the interface. However flow through the upper layer > >will > >be unable to match saturated flow in the lower layer so sites of vapour > >condensation become incipient fingers. Once fingering flow has begun then > >pores > >conducting vapour to the wetting front maintain the forward progress of the > >fingers. > > > > Once the exit point is reached, a zero tension boundary condition is > >created. > >Then the lower layer wets from below to create uniform flow directly > >beneath the > >interface, changing to a moisture gradient corresponding to zero hydraulic > >gradient, immediately below the exit. Since the vapour conductivity of the > >upper > >layer increases as tension increases so the upper layer loses water to > >provide a > >net water flux closer to the saturated hydraulic conductivity of the lower > >layer. > >A major part of the upper soil layer becomes unsaturated and displays a steep > >potential gradient. > > > >Occurrence of instability (Agreement with observed fingering phenomena) > > > > Miller and Gardner (1962) investigated infiltration into a fine/coarse > >soil > >sequence. They noticed that water passing into an initially dry lower sand > >layer > >wets the sand in only a few places with the remainder staying dry. Subsequent > >liquid movement in the lower sand layer is restricted to the water filled > >channels. Hill and Parlange (1972) demonstrated the nature of fingers > >intruding > >into the coarse sand of a fine coarse sand sequence with a pore size ratio of > >around 1:10. A random distribution of wetting fingers developed 0.03-0.06 m > >below > >the interface and moved at a speed corresponding to the saturated hydraulic > >conductivity of the lower sand. Fingers were not initiated at > >discontinuities of > >the interface and fingers formed preferentially along side the walls of > >Plexiglas > >containers. Fingers were of approximately uniform width and were not equally > >spaced. > > > > All reported findings appear compatible with the hypothesis that > >fingering > >flow is a consequence of vapour assisted flow under a pressure gradient set > >up by > >differences of radius of curvature of air/vapour/water interfaces. > >Differences are > >set up by a fine coarse sequence, by an opposing air pressure gradient, > and by > >water fronts with different contact angles at the advancing and receding > >fronts, > >in media that is sufficiently coarse. Sufficient tension must exist behind > the > >wetting front to allow capillary condensation. > > > > 6 Conclusions > > > > When moisture flow through fine-pored soil meets an interface adjoining > >coarse-pored soil then instability can occur. A necessary condition for > >instability is that the tension of the coarse-pored material is > >sufficiently high > >that the associated hydraulic conductivity is less than that of the > fine-pored > >material. This commonly corresponds with tensions exceeding those at which > >flow by > >capillary condensate is initiated in the coarse-pored soil. > > > >owner-sowacs@aqua.ccwr.ac.za wrote: > > > > > NOTE: To get off this list, send email to majordomo@aqua.ccwr.ac.za > > > with the body of the message containing the line: > > > unsubscribe sowacs > > > > > > Could anyone explain something to me? > > > > > > I have some suction measurements from Kazakhstan last summer, using > > > laboratory constructed mercury tensiometers. When the data is plotted as > > > potential it indicates that moisture was flowing upwards from the water > > > table. Indeed, the capillary fringe was about 1.30m, and the actual > > > moisture from the capillary fringe was 'pulled' much further upwards, > > > possibly another 60cm. > > > > > > However, the shallowest tensiometer at 30cm shows that moisture was slowly > > > draining from the top 30cm of the soil profile, or being drawn > downwards by > > > the plants. TDR measurements at the same depths show moisture contents > > > close to field capacity, and at maximum 1 bar suction. > > > > > > What I can conclude from this, as has been found in previous studies in > the > > > area, is that a plough 'pan' has formed, and limits the irrigation waters > > > downward, or drainage flow. TDR results hardly indicate an irrigation took > > > place, yet over 100mm was applied at this time. Could the plough 'pan' be > > > restricting drainage flow (hence the equipment did not register the > > > irrigation), so that a very slow drainage downwards took place, at the > same > > > rate of wetting, as that of drying, therefore actual volumetric moisture > > > content didn't change? If this is the case, the pan could almost prevent > > > soil evaporation, as it is restricted in downward movement by the physical > > > barrier of the 'pan', but actually slowly pulled down by the plant roots > > > (at a higher rate or force than bare soil evaporation)? > > > > > > If this is the case, am I really only seeing transpiration, and not > ETo? If > > > so, how can I calculate transpiration alone, or at least compare an > > > empirical method to my soil moisture measurements (i.e. page 135-136, FAO > >56)? > > > > > > A reply to Wenceslau, > > > > > > I have also experienced the same problem, TDR and field samples giving > > > different results. Soil variability will of course play a part, but in > the > > > end we used very thin walled sample rings, originally tubing from a hang > > > glider, so it was of aircraft quality. These were very slowly inserted > > > into the soil in a trench, using a modified car jack to slowly push the > > > ring in. Yet differences were still apparent. In the end we concluded it > > > was due to soil density, and possible TDR insertion errors, or calibration > > > errors. > > > > > > Thanks for any help > > > > > > James Dalton > > > James A. Dalton > > > Research Assistant > > > Institute of Irrigation and Development Studies > > > Department of Civil and Environmental Engineering > > > University of Southampton > > > Highfield > > > Southampton > > > SO17 1BJ > > > Hampshire > > > United Kingdom > > > Tel: +44(0)23 8059 2746 > > > Fax: +44(0)23 8067 7519 > > > > > > > > > > > > > > > > James A. Dalton > Research Assistant > Institute of Irrigation and Development Studies > Department of Civil and Environmental Engineering > University of Southampton > Highfield > Southampton > SO17 1BJ > Hampshire > United Kingdom > Tel: +44(0)23 8059 2746 > Fax: +44(0)23 8067 7519