SISTEMA DE GESTION DE CALIDAD MEJORAMIENTO CONTINUO
CICLO PHVA
The zero crossing technique does not reveal any preference for positive or negatively skewed bedforms or any association between form asymmetry and slope. Without any site scale trends, the frequency distribution of the asymmetry coefficients could be used in much the same manner as the frequency distribution of bedform spacing. Even if only the modal asymmetries were used this would improve the degree to wiiich natural alluvial river form is replicated. Unfortunately, as these asymmetry characteristics are derived without the use of a tolerance value (due to the preferential removal of bedforms, Figure 3.21), they include the asymmetries of bedforms which could be termed morphologically and hydraulically 'insignificant' in relation to larger pool and riffie bedforms.
The adapted bedform differencing technique, which does apply a tolerance limit, reveals an overall asymmetry preference for positively skewed bedforms. Most of these positively skewed bedforms are preferentially located upstream, with the bedforms in the downstream reaches not showing any asymmetry preference. These trends are presented in Figure 4.28 in relation to slope. A simple summation of these characteristics is that most bedforms on slopes > 0.008 are positively skewed with asymmetry ratios between 1.5 and 4 (i.e. the lee or downstream side is 1.5 to 4 times longer than the stoss or upstream side). Most bedforms on slopes < 0.008 show no positive or negative asymmetry preference. Modal asymmetry ratios of ± 1.5 are present on these on these gentler slopes, with values ranging from ± 5. In addition to bedform shape, the variation in longitudinal bedform asymmetry reflects another means by which bedforms adjust to the imposed conditions. 0.012 0.010 0.008 0.006 w 0.004 0.002 0.000 -0.002 I ▲ Centreline • Thalweg A • A • A • A A • • A m A • • A -3 -2 -1 Asymmetry
Figure 4.28 Reach scale averages of longitudinal bedform asymmetry characteristics in relation to centreline and thalweg channel slope. Adapted bedform differencing technique applied to the centreline and thalweg profiles of the Clyne River.
4.4 Summary
Investigation of the natural alluvial Clyne River has revealed an extremely diverse morphology which varies both at the reach and sub reach-scale. Some of the form characteristics identified in the Clyne reaches are also revealed in other studies conducted on other natural non-alluvial river systems. These can provide a basis by which current templates can be adapted. Unfortunately, studies of pool-riffle morphology in natural (or indeed any) river systems is sparse (Carling and Orr, 2000) and that which does exist has been derived by techniques and associated methodologies which are not explicitly stated. Consequently, only a few form characteristics are outlined in Table 4.11.
Some of the form characteristics identified in the Clyne reaches are not similar to those identified in other river systems (Table 4.12). This variability may reflect form adjustment to factors such as local geology and sediment calibre (Wohl et al, 1993), sediment supply (Lisle, 1982), the hydrological regime, the energy of the environment (possibly prompting Heys’ recommendations of pool-riffle wavelength variations with slope) or the presence of large woody debris (Montgomery et al, 1995; Gumell and Sweet, 1999). The transferability of the morphological characteristics identified in the Clyne is therefore unknown. Further evaluation of naturally developed river systems is therefore required with a greater attention paid to catchment as well as more localised factors which may influence the morphology.
Form characteristic
Design recommendation Other examples
Slone Linear or near linear slope at the reach scale Shepherd (1985) Long profile Natural divisions in the long profile are evident.
These segment the river into relatively short sections (100m to 300m) of uniform character.
Profile divisions evident in studies relating to tributary input (Shepherd,
1985), geology, tectonic activity.
Bedform spacing
Modal spacing at 3wb
Significant positively skewed frequency
distribution. Implemented spacing characteristics should conform to Table 4.10
Individual bedform spacing range: 1 Wb to 19 Wb
Keller and Melhom, 1978 Milne, 1982a
Clifford, 1993b Grant et a/., 1990 Range: 1 Wb to 19 Wb Keller and Melhom, 1978 Cross-sectional
asvmmetrv
Pools should be asymmetrical and the riffles symmetrical with the deepest point in the centre of the channel.
Einstein and Shen, 1964 Keller, 1975
Thome and Hey, 1979 Hey, 1986
Thompson, 1986 Table 4.11 Recommended design criteria to enable rehabilitation schemes to replicate rural
Form characteristic
Morphological characteristics in the Clvne River
Evidence found in other studies Bedform
snacine
Average spacing 4 - 5wy on steeper slopes (>0.006) and 3 - 5wb on gentler slopes (<0.006)
No clear reach-scale trends
Opposite trends to Hey who advises average bedform spacing on steeper slopes of 4wy and 8wy on gentler slopes.
No evidence from other studies. Bedform
amolitudes
Increase with distance downstream.
Significantly positively skewed frequency distribution for pools, but normal
distribution for riffles. No clear reach-scale trends.
Wohl et a l, (1993) also identifies bedform amplitudes to increase with distance downstream.
No evidence from other studies.
No evidence from other studies. Bedform
aspect ratio
On average, riffle heights should be 0.011
to 0.02i on steeper slopes (>0.006), and 0.02i to O.OSi on gentler slopes (<0.006). Aspect ratios derived from the centreline profile (bedform differencing).
Significantly positively skewed frequency distribution on steeper slopes,
normalising on shallower slopes. Individual aspect ratio range: 0.001, to 0.07,
No clear reach-scale trends.
Opposite trends to Carling and Orr, (2000) who observe aspect ratios to decrease with distance downstream.
No evidence available from other studies.
No evidence from other studies through bedform differencing. No evidence from other studies Cross-
sectional asvmmetrv
The deepest point in the pool trough should be around 20% to 40% greater than the pool depth in the centre of the channel (measured from the low flow water surface) on slopes > 0.006. The deepest point in the pool should be around 5% to 15% greater than the pool depth in the centre of the channel on slopes < 0.006.
No evidence available for asymmetry trends although if sinuosity increases downstream and the bed becomes more erodeable, opposite trends are expected.
Longitudinal asvmmetrv
Slopes > 0.008 should mainly contain positively asymmetric forms with a modal value of 1.5 and range from 1 to 5 (i.e. the lee or downstream side is 1 to 5 times longer than the stoss or upstream side). Slopes < 0.008 should include a 'mix' of asymmetry characteristics, with modal values of ±1.5 and range from ±1 to ±5).
Ranges similar to Carling and Orr, (2000) although there is no evidence for asymmetry/slope trends from other studies.
Longitudinal shape
Bedforms should generally reflect a 'dome up' shape. Riffle lengths {not
wavelengths) as determined by the distance along the natural gradient (trendline) should be 10 to 15% longer than pool lengths.
No evidence available from other studies.
Table 4.12 Variation in pool-riffle characteristics derived from the River Clyne (mainly through bedform differencing) in relation to characteristics derived from other studies.