Merrick, Davis and Mahaney, William C.: Particle Size of three
Holocene Soil Profiles in South-Central Ontario. Geografisk
Tidsskrift 86: 63-69. Copenhagen, June 1986.

A chronosequence in the Rouge River Basin of south-central Ontario, were studied to investigate methods for describing the particle size distributions and to determine whether there are significant differences in the distributions for soil sola of different ages. These soils, forming in alluvium of mixed mineralogy, represent weathering and soil formation in the late Holocene (Site Rl3), mid-Holocene (Site Rl2) and early Holocene.

David Merrick. National Coal Board, The Vache, Chalfort St. Buckinghamshire, U.K. and William C. Mahaney, ass prof, at the Atkinson College, York University, 4700 Keele Street, North York, Ontario, Canada M3J 2R7.

Keywords: Chronosequence, partical size analysis, soils, Ontario,
Canada.

Assessment of the time factor in soil formation is achieved using the chronosequence concept of Jenny (1941, 1980). This is defined as a sequence of soils developed in similar parent materials and topographic settings, under the influence of non raying climatic and biotic factors whose different states can be attributed to lapse in time since the initiation of soil formation (time zero). The soil morphogenesis process involves weathering and leaching, which induces gains, losses, transformations and translocations of organic and inorganic constituents. These variations are reflected in the morphological, physical, chemical and mineralogical properties of the soils.

Soil formation often involves increases in finer texture (Birkeland, 1984), changes in structure (Mahaney, 1974, 1978, 1984), solum thickness and horizon development (Mahaney and Fahey, 1976), changes in surface and subsurface color (Crocker and Major, 1955). Many chronosequence studies reveal that, over time, organic matter increases, and soluble salts, basic cations and pH decrease (Crocker and Major, 1955; Dickson and Crocker, 1954; Franzmeier et al, 1963). The development of Fe oxides in the soil provides an index of time (Alexander, 1974; Cambell, 1971) and increases in total clay content reflect the time factor (Ahmad et al., 1977). The present paper considers, in particular, the size distributions of samples from soil profiles of different ages. A method of characterizing the size distributions is proposed and tentative conclusions on the differences between size distributions are drawn.

FIELD AREA

The valleys of south-central Ontario are characterized by alluvial terraces and floodplain deposits formed by postglacial stream activity. Fluvial sediments in the Rouge River Valley (Figure la) are derived from a wide range of glacial and nonglacial deposits, which have shale, limestone, granitic, and gneissic clasts incorporated in them. Stream incision has given rise to three distinct age surfaces, shown in Figure Ib. Sample site locations representing these surfaces are shown in Figure la. The deposits are named from oldest to youngest: Rouge, Twyn Rivers, and Highland Formations. Soils formed in these deposits are given the prefix 'post', to avoid a terminologic proliferation. The two oldest soils, post-Rouge and post-Twyn

Rivers, developed in pebbly loamy alluvium. These soils are found in very late Pleistocene/Holocene to late Holocene surfaces and are classified as an Orthic Regosol (post-Highland soil, R13); and Orthic Sombric Brunisols (post-Twyn Rivers, Rl2; and post-Rouge soils, Rls), Canada Soil Survey Comm., 1977. The post-Highland soil is estimated to be late Holocene in age (4000 radiocarbon years 8.P.) and undergoing continual redeposition of sediments by river flooding. The post-Twyn Rivers soil is estimated to be mid-Holocene in age while the post soil is likely early Holocene to very late Pleistocene in age.

Soils were sampled in terrace surfaces with less than 1-2° slope. The climate of the area is humid continental, cool summer, no dry season type described by Brown (1968) and Mahaney and Ermuth (1974). The average temperature ranges from 20° C in July to -7° C in January; extremes reach 40°C in July and -34° C in January. A frost-free period of 150 days lasts from mid-May to early Wind in the area is dominantly westerly, and mean annual precipitation is 850 mm.Soils were sampled in areas covered with sugar maple and beech stands (two highest surfaces) and willow stands (low surface). Some clearing for cultivation had occurred at site Rl2; sites Rl5 and Rl3 were undisturbed (Figure la for location).

METHODS

Duplicate soil samples were collected from each soil profile described in detail. Soil descriptions follow Canada Sou Survey Comm. (1977) and Birkeland (1984), while soil color was determined from the Standard Soil Color Charts of Oyama and Takehara (1970). Soil samples were air dried and passed through a 2 mm sieve. For particle size analysis samples were treated with H₂0₂ to remove organic constitutents, and with sodium pyrophosphate to achieve deflocculation. All samples were agitated with a Branson 350 cell dismembrator to separate clay constituents. Sands were wet sieved using 63 mm sieves and coarse grade sizes were determined after dry sieving. Fine grade sizes of silt plus clay were determined by sedimentation following Bouyoucos (1962) and Day (1965). The statistical analysis was carried out using standard linear regression techniques, for example see Moore and Edwards (1965).

RESULTS AND DISCUSSION

The three soils were sampled to different depths. Because the soils in each terrace are similar, results from one of each soil stratigraphic unit were presented. Only the morphologic and particle size parameters of importance to soil genesis are discussed below.

Morphology

The morphology of the three soils is presented in Tables 1,2 and 3. The color in the upper solum is fairly uniform between the three soils, while subsoil (subsolum) color varies with a trend from lighter 10YR 5 and 6 hues to darker 10YR 4 hues over time. These colors grade into 2.5Y and light 10YR hues in the parent materials (Cu horizons). A parallel trend is observed with topsoil structure which ranges from weak granular in the youngest soil to a stronger grade of granular development in the older pedons. Structure below the Ah horizons in the lower solum and subsoil ranges from depositional stratification in the youngest profile (post-Highland soil) to weak blocky aggregates in the B horizons of the older soils (post-Twyn Rivers and post-Rouge soils). The consistence of the surface soil horizons does not differ appreciably, but some differences in the lower horizons are discernible, presumably the result of particle size variations. Coatings on ped faces occur only in the B horizon of the oldest profile (post-Rouge soil). Increasing age of each soil is reflected by deeper pedons, greater solum thickness, and horizon differentiation (especially the development of B horizons). Exact differences in profile depth as a function of age are complicated by the presence of buried soil horizons in the post-Highland soil.

Particle Size

Data resulting from particle size analyses of the three soils are shown in Table 4. While the distributions of sand and silt varied somewhat, the values for clay increased slowly with depth in the post-Highland soil (Rl3). However, in the older post-Twyn Rivers soil (Rl2), silt and clay are higher in the solum than in the subsoil and parent material. A similar pattern occurs in the post-Rouge soil (Rls) where clay in the upper solum increases to 10 per cent. The trend towards greater clay content in surface soil horizons with increasing age suggests increased production of clay over time as a function of weathering. It is also possible that these variations occur as a function of paleohydrological changes in stream regimen, or as a result of airfall influx of material. However, no minerals that might have an allochthonous origin were discovered in the clay-silt grade sizes in the soil sola.

Particle size distribution

A statistical analysis was carried out to investigate whether the size distributions of the samples could be correlated satisfactorily by the Rosin-Rammler equation. This equation is widely used to describe the size distributions of crushed materials such as coal (for example, Rose and Cooper, 1977), and has the following simple analytical form: y=exp (-ax^b)') where yis the fraction of the sample with particle diameters greater than x microns, and a and b are constants.

It is usual to exclude measurements for which there is
less than 2 per cent oversize, as the Rosin-Rammler distribution
does not normally apply in such cases.

The Rosin-Rammler equation can be transformed into
a linear equation in the parameters log(a) and b, as follows:
b log x + log a = log log ('/y)2).

An ideal Rosin-Rammler particle size distribution
therefore gives a straight line if the log-log-reciprocal oversize
is plotted against the log particle size.

The parameters a and b were obtained for each of the samples described in Table 4 by carrying out a linear regression analysis (least - squares fit) on the transformed equation (2). The data used in the analysis were given in the table, except that the measurements for 7.8 microns and below were excluded.

For many of the samples, however, it would have been sufficiently accurate to plot the data on log-log-reciprocal v log graph paper (for example, Chart Well 5596) and fit a straight line 'by eye'. The parameters a and b can then be calculated by choosing two points on the line, substituting the values into equation 2, and solving the resulting simultaneous equations.

The experimental data are shown in Figures 3 to 9, together with the line corressponding to the 'best-fit', obtained by the statistical analysis descriped above. As illustrated in the figures, the Rosin-Rammler equation generally provides a reasonable describtion of the size distribution for particles greater than about 10 microns, the distribution below this size being significantly finer than would be expected by extrapolation from the coarser sizes.

The results are summarized in Table 5, together with the standard error of predicting the percentage oversize for the data points used in the analysis. As shown in the table, the standard error lies in the range 1.1 to 3.1 percentage points for particle sizes greater than (or equal to)

15.6 microns. Also shown in the table is the difference between the observed percentage of the material less than 1.95 microns and that predicted by extrapolation of the best-fit Rosin-Rammler equation. With the exception of Rls-Bml and Bm2, the differences are less than 3 percentage points.

The best-fit equation was used to estimate the median
size of the distributions, and the upper and lower quartiles;
these values are also given in Table 5.

DISCUSSION

The particle size distributions were compared in order to investigate the differences between the samples. A formal 'analysis of variance' approach was not employed because of the limited number of data points available. Three parameters were chosen to characterize the size distributions:

1) The median particle size (this gives a general indication
of whether the size distribution is 'coarse' or
'fine').

2) The parameter b of the Rosin-Rammler equation (a high value corresponds to a smaller spread of particle sizes, or equivalently a more 'peaky' size distribution, than a low value).

3) The excess material finer than 2 microns.

For the A horizons, these parameters are compared in Figure 10. Although there is no clear overall trend, it appears that the Rl3 size distribution is more 'peaky' than those for Rl2 and Rl 5. The data also suggest that Rl 5 has a lower median size than Rl3 and Rl2, and also contains more excess fine material less than 2 microns. For the B horizons, the trends are shown in Figure 11. The indications are that Rl5 has a somewhat more 'peaky' size

distribution, a lower median particle size, and more excess
fine material below 2 microns than Rl2.

In the present study, the limited amount of data makes it difficult to establish the statistical significance of these trends, although the standard test could be applied in situations where more data (including replicate samples) are available. Even so, the use of the Rosin-Rammler equation as described makes it easier to identify trends than when working directly from the 'raw' data in Table 4.

CONCLUSIONS

On the basis of the above, limited analysis, two conclusions
are tentatively proposed:

1) The Rosin-Rammler equation can provide a satisfactory description of the size distribution of soil samples taken from the Rouge River Basin for particles larger than about 10 microns. This therefore includes, in one continuous distribution, the sand and much of the silt.

2) The particle size distributions can be characterized by the three parameters median size, the parameter b of the Rosin-Rammler equation, and the difference between the measured and predicted amounts of fine particles (the 'excess material' below 2 microns). Comparison of these parameters for the present samples indicates that, for the A horizons, Rl3 has a smaller range of particle sizes than Rl2 and Rl5. For the B horizons, Rl5 has a somewhat smaller range of particle sizes, a lower median size and more excess fine material below 2 microns than Rl2.

Summary

Three soils of postglacial age, representing a chronosequence in the Rouge River Basin of south-central Ontario, were studied to investigate methods for describing the particle size distributions and to determine whether there are significant differences in the distributions for soil sola of different ages. These soils, forming in alluvium of mixed mineralogy, represent weathering and soil formation in the late Holocene (Site Rl3), mid-Holocene (Site Rl2) and early Holocene/late Pleistocene (Rls) surfaces. With increasing age, horizon differentiation and soil thickness increases, along with percent clay in the sola (A+B horizons) of the three soils show that the Rosin equation provides a reasonable description of the size distribution for particles greater than 10 microns. The distribution below this size (fine silt and clay) is finer

(as determined by experimental results) than would be expected by extrapolation from the coarser sizes. An analysis of parameters derived from the Rosin-Rammler equation permits two conclusions to be drawn on the differences in the size distributions:¹) The Rosin-Rammler equation provides a satisfactory description of the size distribution of soil samples for particles larger than about 10 microns, and 2) particle size distributions can be characterized by the three parameters - median size, parameter b of the Rosin-Rammler equation, and the difference between the measured and predicted amounts of fine particles.

ACKNOWLEDGEMENTS

This research was made possible by grants from the Natural Science and Engineering Research Council of Canada and York University. Larry Gowland assisted with the field and laboratory analysis. Research was carried out in the Geomorphology and Pedology Laboratory of Atkinson College, and at the National Coal Board, Cheltenham, U.K.

References

Ahmad, M., Ryan, J. and Paeth, R.C., Soil development as a
function of time in the Punjab River plains of Pakistan: Soil
Sei. Soc. Am. J., v. 41, p. 1162-1166.

Alexander, E.8., 1974, Extractable iron in relation to soil age on
terraces along the Truckee River, Nevada: Soil Sei. Soc. Amer.
Proc. 38, p. 121-124.

Birkeland, P. W. 1984, Sous and Geomorphology, New York, Oxford
Press, 372 p.

Bouyoucos, G.J., 1962, Hydrometer method improved for making
particle size analyses of soils: Agron. Jour., v. 54, p. 464-465.

Brown, D.M., et al, 1968, The Climate of Southern Ontario, Climatological
Studies no. 5: Dept. of Transport, Meteorological
Branch, Ottawa, Queen's Printer.

Cambell, 1.8., 1971, A weathering sequence of basaltic soils near

Dunedin, N.Z.: New Zealand J. Sou Sei., v. 14, p. 907-924.
Canada Soil Survey Comm., 1977, Soils of Canada, v. 1: Ottawa,

Agriculture Canada, 243 p.

Crocker, R.L. and Major, J., 1955, Soil development in relation
to vegetation and surface age, Glacier Bay, Alaska: J. Ecol., v.
43, p. 427-448.

Day, P., 1965, Particle fractionation and particle-size analysis, in
Black, C. A., ed., Methods of Soil Analysis: Madison, Wise.,
Am. Agron. Soc., p. 545-567.

Dickson, 8.A., and Crocker, R.L., 1954. A chronosequence of soil
and vegetation near Mt. Shasta, California III: J. Soil Sei., v. 4,
no. 2, p. 173-191.

Folk, R.L., 1968, Petrology of Sedimentary Rocks, Austin, Tex.,
Hemphill Press, 170 p.

Franzmeier, D.P., Whiteside, E.P., and Mortland, M.M., 1963, a chronosequence of podzols in northern Michigan: 111. Mineralogy, micromorphology, and net changes occurring during soil formation: Mich. Agric. Exp. Stn. Bull., no. 46, p. 37-57.

Hodson, J.M., ed., Soil Survey Handbook, Soil Survey Tech,
Monog. no. 5, Rothamsted Exp. Stn., Harpenden, Herts, U.K.,
99 p.

Jenny, H., 1941, Factors of Soil Formation, a system of quäntitative
pedology: New York, McGraw-Hill, 281 p.

Jenny, H., 1980, The Soil Resource: Origin and Behavior: New
York, Springer-Verlag, 377 p.

Mahaney, W.C., 1974, Soil stratigraphy and genesis of Neoglacial deposits in the Arapaho and Henderson Cirques, Central Colorado Front Range, in Mahaney, W.C., ed., Quaternary Environments: Proceedings of a Symposium: Geographical Monographs, no. 5, p. 197-240.

Mahaney, W.C., 1975, Soils of post-Audubon age, Teton Glacier
area, Wyoming: Arctic and Alpine Research, v. 7, no. 2, p.
141-153.

Mahaney, W.C., 1978, Late-Quaternary stratigraphy and soils in the Wind River Mountains, Western Wyoming, in Mahaney, W.C., ed., Quaternary Soils: Norwich, U.K., Geoabstracts Ltd., p. 223-264.

Mahaney, W.C., 3984. Superposed Neoglacial and late Pinedale (Wisconsinan) tills, Titcomb Basin, Wind River Mountains, Western Wyoming, Palaeogeography, Palaeoclimatology, Palaeoecology, v. 45, p. 149-163.

Mahaney, W.C., and Ermuth, H.F., 1974, The Effects of Agriculture
and Urbanization on the Natural Environment: Geographical
Monographs no. 7, 152 p.

Mahaney, W.C., and Fahey, 8.D., 1976, Quaternary Sou Stratigraphy of the Front Range, Colorado, in Mahaney, W.C., ed., Quaternary Stratigraphy of North America: Stroudsburg, Pa., Dowden, Hutchinson and Ross, p. 319-352.

Moore, P.G., and Edwards, D.E., 1965, Standard Statistical Calculations:
London, Pitman.

Oyama, M., and Takehara, H., 1970, Revised Standard Soil
Color Charts, Japan Research Council for Agriculture, Forestry
and Fisheries.

Rose, J. W., and Cooper, J.R., 1977, Technical Data on Fuel: British
Natl. Comm. World Energy Conf., London.