- Home
- Documents
*Annex 1 Estimating soil hydrological characteristics from ... Estimating soil hydrological...*

prev

next

out of 106

View

2Download

0

Embed Size (px)

115

Annex 1

Estimating soil hydrological characteristics from soil texture and structure

It is possible to derive rough estimates of the hydraulic conductivity (K) and the drainable pore space (μ) from observations of the soil profile. This is because these soil hydraulic qualities depend on soil texture and structure. Table A1.1 average presents μ values, compiled by FAO (1980) and based on data from the USBR (1984), together with K values estimated from the μ/K relationship. For soils with distinct horizontal layers, the vertical K may be taken as being at least 10 and on average 16 times lower than the horizontal one.

As these estimates may be imprecise, more realistic K values are obtained through field measurements, as described in Annex 3.

However, interpreting the soil structures mentioned in Table A1.1 may not be easy. It should be done through observations of soil profiles, but shallow groundwater levels often prevent excavation of soil pits. Moreover, soil texture and structure should be evaluated when the soil is moist throughout.

However, in special cases, it is possible to estimate drain spacings directly from the visual aspects of the soil profile, as was done by people with detailed local experience in the Zuiderzee polders, the Netherlands, where it was the only possible method – drain spacings of 8, 12, 16, 24, 36 and 48 m were distinguished and the choice between possibilities was possible.

For pure sands (almost without clay and silt), an estimate is:

where: K = permeability (m/d). m50 = median size of grains above 50 µm. Half of the weight is above this size, half

below.

1 C: clay; L: loam; S: silt; s: sand. Source: Adapted from FAO, 1980, with further elaboration.

TABLE A1.1 K and µ values according to the soil texture and structure Texture (USDA)1 Structure µ K

(m/d)

C, heavy CL Massive, very fine or fine columnar 0.01–0.02 0.01–0.05

With permanent wide cracks 0.10–0.20 > 10

C, CL, SC, sCL Very fine or fine prismatic, angular blocky or platy 0.01–0.03 0.01–0.1

C, SC, sC, CL, sCL, SL, S, sCL Fine and medium prismatic, angular blocky and platy 0.03–0.08 0.1–0.4

Light CL, S, SL, very fine sL, L Medium prismatic and subangular blocky 0.06–0.12 0.3–1.0

Fine sandy loam, sandy loam Coarse subangular block and granular, fine crumb 0.12–0.18 1.0–3.0

Loamy sand Medium crumb 0.15–0.22 1.6–6.0

Fine sand Single grain 0.15–0.22 1.6–6.0

Medium sand Single grain 0.22–0.26 > 6

Coarse sand and gravel Single grain 0.26–0.35 > 6

Guidelines and computer programs for the planning and design of land drainage systemss116

The presence of silt (< 50 µm) and especially clay (< 2 µm) will lower this value considerably. Therefore, this formula should not be used for such soils.

REFERENCES FAO. 1980. Drainage design factors. FAO Irrigation and Drainage Paper No. 38. Rome. 52 pp. USBR. 1984. Drainage manual. A Water Resources Tech. Publication. Second printing. Denver,

USA, US Dept. of Interior, Bureau of Reclamation. 286 pp.

117

Annex 2

Statistical analysis of extremes

GUMBEL’S METHOD The Gumbel distribution can be used for extrapolating from a limited number of extreme values (Gumbel, 1954 and 1958). The basic data appear in groups, such as the daily rainfall in August (31 days per year), or the water levels in a river per year (365/366 days). The highest value in such a group is the extreme. The groups should contain at least ten elements, and the minimum number of extremes (often years) is at least ten.

The method assumes that the underlying process remains constant. This supposition is doubtful because of recent climate changes, which also influence data such as river flows. These changes are especially noticeable in the extreme values. Therefore, the method should be used with care.

Extreme values are obtained as follows: ÿ Select the highest (sometimes lowest) value in a group, e.g. the highest autumn

rainfall or the highest river discharge in a year. Each group should contain at least ten values.

ÿ These extremes are sorted according to their magnitude in order to prepare for further analysis.

The probability that a certain value x does not exceed a limit x0 is:

with (1)

where: P = probability; n = number of extremes; u = constant (shift); x = values of the extremes. The average is the standard deviation is sx; x0 = limiting value; y = reduced Gumbel variable, with average c and standard deviation sy. For y and

for a very large number of observations, c = 0.57722 = Euler’s constant; α = constant (slope). The probability that x exceeds x0 is:

(2)

The return period T is the number of groups in which the limit x0 is exceeded. If there is one group per year, T is in years (as in the above examples). T is defined as:

(3)

For the x values, the procedure is:

Guidelines and computer programs for the planning and design of land drainage systems118

standard deviation of y.

Table A2.1 shows the values derived by Kendall for a smaller number of observations.

The line y = α(x - u) has two parameters: the slope α, and the shift u. They can be found by plotting on Gumbel probability paper, usually with the return period T on the horizontal axis, the value of the extremes on the vertical. The line may be drawn visually through the points to allow extrapolation. In this way, the once-per-century rainfall or the river discharge can be estimated. This is even possible for much longer return periods.

The program GUMBEL calculates the parameters automatically and provides estimates for the extremes to be expected with a certain return period.

For agricultural drainage design, a return period of 2–10 years is often taken, 2– 5 years for field drainage and even 10 years for crop systems with high planting costs, and 5–10 years for the main system where it does not affect inhabited places.

By extrapolation, a prediction can be given over much longer periods of time in order to obtain estimates for values to be expected once in 100 years (the once-per-century value) and even for much longer times. However, the uncertainty of the estimates becomes very large for such longer return periods. Moreover, for such periods (and even for a century), the basic data series cannot be considered as constant, owing to human and geological influences.

Nevertheless, such a prediction is valuable for engineering purposes, e.g. the height of a river embankment able to withstand a “100-year flood”. This will almost certainly

not occur 100 years later, but it has a chance of 1 percent of occurring next year.

The influence of climate changes can be analysed by comparing data from the last 10–20 years with earlier ones (where available), and it is wise to employ the worst prediction. Where not different, the basic data include recent changes already.

REFERENCES Gumbel, E.J. 1954. Statistical theory of extreme values and some practical applications. Applied

mathematics series 33. U.S. Dept. of Commerce, National Bureau of Standards. Gumbel, E. J. 1958. Statistics of extremes. New York, USA, Columbia Press.

TABLE A2.1 Values of c and sy as a function of n

n c sy

10 0.495 0.950

15 0.513 1.021

20 0.524 1.063

25 0.531 1.092

30 0.536 1.112

40 0.544 1.141

50 0.548 1.161

119

Annex 3

Field methods for measuring hydraulic conductivity

INTRODUCTION The K value can be measured directly in the soil layers situated below the groundwater level using the methods described below. Less reliable methods are used to estimate the saturated hydraulic conductivity above this level. For well-moistened granular soils, the soil permeability for saturated flow can be estimated from the capillary hydraulic conductivity of the unsaturated zone. However, this is not the case in well- structured soils where this permeability is caused by cracks, holes or other macropores. Infiltrometer or inverse auger-hole methods are often used as a compromise. They measure conductivity under “almost saturated” conditions.

The field methods for determining K are based on a basic principle: water flows through a volume of soil, whose boundary conditions are known, and the discharge is measured; the K value is calculated by applying an equation derived from Darcy’s Law applied to the specific geometry of the soil volume.

The following paragraphs review the suitability of the field methods most commonly used to measure the soil hydraulic conductivity (auger-hole, piezometer, and inverse auger-hole). The methods are different according to the groundwater depth at the time of measurement. Details on these methods can be found in the bibliographic references (Van Hoorn, 1979; USBR, 1984; Oosterbaan and Nijland, 1994; Amoozegar and Wilson, 1999).

AUGER-HOLE METHOD FOR DETERMINING SOIL PERMEABILITY The auger-hole method (Van Beers, 1983) is the most suitable way of measuring the K value of saturated homogeneous soils down to a depth of about 3 m. It is based on the relationship between the K value of the soil surrounding a hole and the rate at which the water level rises after pumping. The method measures the saturated permeability in a rather large volume, which is an advantage in view of the large variability in natural soils.

Method This method for determ