Observations of Needle Ice Growth and Potential for Accelerated Erosion on the Georgia Piedmont

Southeastern Geographer Vol. XX, No. 1, May 1980, pp. 31-41 OBSERVATIONS OF NEEDLE ICE GROWTH AND POTENTIAL FOR ACCELERATED EROSION ON THE GEORGIA PIEDMONT* Vernon Meentemeyer and Jeffrey Zippin Needle ice is a frost form which grows as an accumulation of fine, bristle-like columns of ice protruding from the soil surface. (J) Field and laboratory investigations of this form of segregation ice have shown that needle ice forms only within narrow microclimatic and soil moisture and texture ranges. It forms often enough on the Southern Piedmont to be recognized as an agent capable of enlarging gullies. (2) Field studies of the microclimatic and soil conditions under which needle ice will grow on the Georgia Piedmont were performed as part of a study of the soil texture and moisture limits of needle ice growth. Measurements included needle ice numbers per unit area, height, and mass and weight of soil displaced for two level sites and two sloping locations in Athens, Georgia. This paper reports the observations and measurements made and presents an estimate of the potential for accelerated erosion by a typical needle ice event. THE GROWTH OF NEEDLE ICE. For needle ice to grow, a freezing plane must develop at or just below the soil surface. Thin ice filaments grow molecule by molecule from a soil pore and lift whatever soil, stony debris, or organic matter is available above the freezing plane (Fig. 1). Soil water must be able to ascend rapidly to this freezing plane to provide water for growing ice and to keep the plane from descending deeper into the soil. The stable freezing plane from which the ice grows is maintained by a delicate balance between the rate of cooling and the rate of heat released as the latent heat of fusion. (3) Actual movement of water to the freezing plane is caused by the thermal gradients created * Funds for this study were provided by the Georgia Forest Research Council , Macon, Georgia. Laboratory space was provided by the University ofGeorgia at the Riverbend Research Laboratory. Dr. Meentemeyer is Assistant Professor of Geography at the University of Georgia in Athens, GA 30602. Mr. Zippin is Physical Scientist, U.S. Corps ofEngineers, Washington, D.C. 20315. 32 Southeastern Geographer Fig. 1. Artist's rendition of needle ice based on authors' photographs taken at the experimental site and elsewhere in the Athens, Georgia area. Drawn by Leila Oertel of the Cartographic Services Laboratory, University of Georgia. by surface cooling and suction gradients caused by water removal during ice formation. If the freezing plane stabilizes at lower levels in the soil, other forms of segregation ice such as ice lenses can form; or, if the freezing plane descends too rapidly the soil and soil water will freeze without producing ice crystals. In Georgia the type of cooling necessary to produce ice crystals is usually the result of passage of a vigorous cold front. The front may first produce copious rainfall followed by a decrease in temperature and by clearing weather. The rainfall provides the large amounts of soil moisture needed, and the clear night skies following frontal passage permit radiative cooling of the soil surface. If the wind speed is too great, however , advective cooling may proceed so rapidly that the freezing plane descends quickly into the soil without producing needle ice. Needle ice will never grow on a frozen soil. Thus a large diurnal temperature range favors ice formation by keeping soil thawed by day and causing rapid soil cooling at night. Vol. XX, No. 1 33 p o J&U________H **«<*_ 16' ADDITIONAL? STATION"---^l/í¿^ ROCK OUTCROP Fig. 2. Cross-sectional view of the study site for an observer looking due south. Cloudiness and high humidity can inhibit radiation loss and thereby prohibit the cooling necessary to grow needle ice. (4) A light shrub or grass cover can also inhibit radiative cooling of soil and thereby effectively keep the freezing plane from forming at the soil surface. Considering the delicate combination of conditions required for needle ice to grow, it is remarkable that needle ice grows as often and profusely as it does on the Georgia Piedmont. MATERIALS AND METHODS. The observation site consisted of colluvial material and in situ soil lying above and below a rock outcrop. Soils here are of the Pacolet Series (Typic Hapludults) of sandy clay loams normally found on eroded slopes of 10 to 15 percent. The site was generally covered with sod, but the observation stations were essentially bare. A long slope above the site was covered with mixed diciduous forest; it did not interfere with radiative cooling. Four primary stations were selected at this site based upon previous observations of likely locations for needle ice growth (Fig. 2). These stations were: I. a level area inside a small depression; II. a small hill next to the level area; III. a slope (38°) situated below the level area and hill, and above a rock outcrop; and IV. a slope (21°) below the elevation of the rock outcrop. Supplementary data were collected on a 16° slope at an elevation approximately equal to that of Station III. The evening of 27 January 1976 was selected to monitor needle ice 34 Southeastern Geographer TABLE 1 MICROCLIMATIC DATA COLLECTED AT STATION I ON 27-28 JANUARY 1976 Time Net Radiation Iy cmr2 min"1 Screen Height Temperature + 1.5 m Soil Surface Temperature Om SoilSoilSoil Tempera-Tempera-Tempera-Relative ture ture tureHumidity — 13 mm—25 mm—76 mm% 1800 1900 2000 2100 2200 2300 2400 0100 0200 0300 0400 0500 0600 0700 0800 -.04 -.12 -.10 -.06 -.06 -.06 -.05 -.06 * -.05 -.06 -.05 -.06 -.05 -.02 7.8 6.7 3.9 3.9 2.8 2.2 1.1 1.7 0 0 0 -0.3 -0.6 -0.8 0 3.3 1.1 0 .6 -0.3 -1.1 -1.1 -0.6 -0.8 -1.1 -1.4 -1.4 -1.7 -2.2 -1.4 3.3 1.7 0.8 0.8 0.8 0.8 0.6 0.6 0.6 0.6 0 0 -0.6 0 3.3 2.2 0.8 1.1 1.1 0.8 0.6 0.6 0.6 0.6 0.6 0.6 0.6 0.6 2.2 2.2 1.7 1.7 1.4 1.4 1.4 1.1 1.1 1.1 1.1 1.1 1.1 1.1 20 15 45 100 Data missing or not collected. growth. Most of the southeastern United States was dominated by continental high pressure and clear skies. Precipitation had occurred two days earlier, and by 1430 hours (EST) of 27 January the air temperatures had risen to 55°F (13°C) and relative humidity had fallen to 14 percent. Net radiation became negative at sunset at rates of loss up to —0.12 calories cm~2min"^1. Temperatures at various elevations, as well as humidity and net radiation, were monitored throughout the night. Detailed measurements of temperature were made at points 1.5 meters above the soil surface, at the surface, and 13 mm, 25 mm, and 76 mm below the soil surface using mercury in glass thermometers. Humidity at various elevations was measured using a Bendix motor driven Psychrometer, and net radiation was monitored using a miniature net radiometer. Readings were taken on the half hour until midnight and hourly thereafter to provide a permanent record of conditions during ice development. Detailed measurements were confined to Station I, the level site, because of difficulty in moving the instruments. The hourly data for Station I are presented in Table 1. Vol. XX, No. 1 35 TABLE 2 SOIL MOISTURE AND TEXTURE VALUES OF SOIL SAMPLES TAKEN FROM FIELD STATIONS ON 27 JANUARY 1976 SoilSoil Texture Moisture (% DrySandSiltClay StationWeight)% % % I. Level area inside small depression 23.650.221.928.8 II. Level area on small hill top 25.053.727.518.8 III. Slope below depression and above 22.161.213.425.4 rock outcrop IV. Slope below rock outcrop 23.652.813.234.0 Before and after needle ice growth soil samples were collected and bagged. Moisture percentages were determined later in a laboratory using the gravimetric method. Texture classes of these samples were determined using the method outlined by Bouyoucos. (5) Despite the topographic variations in the sites, the soil moisture percentages were similar, ranging from 22.1 to 25.0 percent by weight (Table 2). When needle ice growth ceased the number ofneedles per unit area was counted , and the quantity of soil uplifted and mass of ice extruded per unit area were sampled. OBSERVATIONS OF ICE GROWTH AND MICROCLIMATE. By 2130 hours of 27 January the soil surface temperature at Station I had fallen to 0°C, but needle ice did not begin to form until slightly before midnight (Fig. 3). After 2100 hours relative humidity at the soil surface remained near 100 percent and net radiation stabilized at a rate of approximately —0.05 cal. cm^min^1 (Table 1). When needle ice began to grow, no abrupt changes in net radiation or temperature could be observed , and humidity readings were discontinued at the soil surface because the wet bulb began to freeze. Needle ice continued to grow at a fairly steady rate until 0600 hours on 28 January. One characteristic of needle ice which is readily visible is the number of needles per unit area. The more level areas (I and II) grew fine, bristle-like needles, often clustered, and averaged nearly eight needles per square centimeter. On the sloping areas (III and IV) the number was diverse, ranging from slightly less than one to nearly four needles per square centimeter. It was also apparent, however, that the needles 36 Southeastern Geographer E E ? O UJ T UJ O 4Or 30 20 10 r9 10 Il 12 I MlONlGHT 2 3 4 5 6 7 8 EASTERN STANDARD TIME Fig. 3. The growth of needle ice observed on the evening of 27-28 January 1976 in Athens, Georgia on a level site. produced on slopes were thicker (some as much as 6 mm in diameter) and displayed various prismatic shapes. To test the hypothesis that all stations produced the same volume of ice, but arranged in differing thicknesses and needle frequencies, samples were taken from precisely measured and bounded areas at each station. Ice samples were placed in containers and then weighed and oven dried in the laboratory. Because the sampled ice included a considerable amount of soil and debris, oven drying was required to determine the weight of ice produced per unit area. Four samples were taken from the level stations (I and II) and four from the slope stations (III and IV) (level stations in grams cm2: 0.462, 0.682, 0.490, 0.487; slope stations: 0.304, 0.423, 0.412, 0.206). The MannWhitney U Test was applied to these two data sets and revealed a significant difference (at 0.05) in water volume. The level stations, contrary to expectations, produced more total ice and also produced taller needles. To remove the ice-height bias, a second test was performed in terms of ice produced per cubic centimeter (including air space) rather than per square centimeter. This test showed no difference among the level and slope stations, indicating that the number of ice crystals compensated for crystal thickness. Vol. XX, No. 1 37 TABLE 3 DOWNSLOPE TRANSPORT OF SOIL BY NEEDLE ICE ON 27 JANUARY 1976 Slope Angle Slope Area3 in cm-2 Grams cm Lifted" Downslopec Movement Grams of" Soil Moved One Centimeter 38° 16° 21° 1,161 774 3,341 1.12 0.85 1.05 2.25 cm 1.40 cm 1.57 cm 2.52 1.19 1.65 a Area of bare soil exposed. b Dry weight of soil lifted by needle ice. c Calculated according to the "gravity fall" model suggested by Higashi and Corte, 1971. d Calculated as the product of grams per cm"' lifted by the downslope movement in cm. The causes for these differences between slope and level stations are difficult to decipher. Throughout much of the Piedmont needle ice is most often seen growing along the spring lines of road cuts and on the walls of gullies. Yet, soil moisture was similar or slightly higher at the level stations (Table 1). Furthermore, the level stations were more exposed to the atmosphere and perhaps for this reason grew taller crystals . Apparently the ultimate control of needle ice growth is not slope but soil moisture and the ability of the soil to transfer this water to the freezing plane, together with the presence of favorable microclimatic conditions. NEEDLE ICE INDUCED EROSION. The lifting of soil, stones, and litter by ice needles has been documented by several authors; however, most studies of the erosional aspects of needle ice have concentrated on first observing needle ice and later measuring the sediment load by runoff over bare ground where needle ice grew. (6) Apparently little information is available on the quantities of soil lifted prior to the measurement of sediment loads, perhaps because it is a tedious task. In a separate analysis the weight of soil and debris lifted by the crystals was measured by gathering crystals and their attached soil/debris from known areas. These samples were oven dried to ascertain water content and residue load. It appeared that more soil had been lifted on the slope stations than the level stations despite the lower number of crystals per unit area on slopes. Four samples were taken from the level stations and four from the slope stations (level stations in 38Southeastern Geographer grams cm2-. 0.192, 0.816, 0.247, 0.268; slope stations: 1.211, 1.065, 0.513, 0.834). The Mann-Whitney U Test verified that significantly more soil was lifted at the slope stations. Several explanations account for the greater soil lifting on the slopes. Rough surfaces covered by loose and often dry debris apparently are more conducive to the formation of a deeper freezing plane and the lifting of large quantities of soil. Visually the slopes appeared to have more loose soil on the surface, probably caused by needle ice on previous evenings. The depth at which the stable freezing plane forms controls the quantity of soil lifted. For the slope stations the freezing plane appeared to have formed later in the evening and deeper in the soil than on the more exposed level stations. Regardless of the reasons for the variations in measured quantity of soil lifted, it is clear that the downslope movement of soil in only one needle ice event is massive. The downslope movement of soil by needle ice has been described by many authors. Of the many terms used in these descriptions the term congelifluction, coined by Dylik, seems especially appropriate. (7) The term congelifluction recognizes the frost origin of downslope movement and distinguishes it from ordinary solifluction. Essentially three types of congelifluction have been identified. Needle ice tends to grow vertical to its slope, so the first type of movement that can be defined is one of simple gravity fall. On level ground, lifted soil would be dropped back to its origin, but on slopes the dropped soil would be moved downslope . The second type is caused by the tumbling of needles down the slope upon melting. Tumbling accentuates soil movement on steep slopes. A third type produces even more movement. For this type there is ice tumbling as in the second type plus an undetermined length of "slide" by lifted debris. Higashi and Corte believe the slide events to be more common than what is reported in the literature. (8) On many occasions we have observed this sliding type of congelifluction on the Georgia Piedmont, especially on the steep slopes of gullies, roadcuts, and construction sites. These events are occasionally spectacular, but so ephemeral that they are seldom observed. To answer the question of magnitude of soil movement by needle ice, ice heights and soil weight lifted were measured at three surfaces below level Station I. These three surfaces correspond to stations III (38°) and IV (21°) but with additional measurements made at a lateral slope of 16° between stations III and IV. Slope angles were measured Vol. XX, No. 1 39 using a Brunton compass. The surfaces were sufficiently smooth so that a uniform slope angle could be assumed. Using the method of Higashi and Corte, the measured needle ice heights were used to calculate downslope movement of debris for the three slopes. In order to produce a conservative estimate of erosion, the first type of movement (gravity fall) was assumed. (9) Table 3 summarizes these measurements and calculations based upon data collected on 27 January 1976. On the first slope (38°), 1.12 grams of soil per square centimeter were lifted and moved by gravity fall (calculated) 2.25 cm. On this steep slope the equivalent of 2.52 grams of soil were moved one centimeter. The remainder of the table may be similarly interpreted. Soil movements do not seem impressive when calculated for small units such as a square centimeter. Yet, consider the 21° Station IV, which had an area of 3.34 square meters of bare soil on which needle ice grew. On this area and on this occasion, an estimated 5.51 kilograms of soil were moved one centimeter down slope. On one hectare with a 21° slope this rate is equal to 16.5 metric tons moved down slope one centimeter . Of course one does not normally find one-hectare plots with bare soil exposed on 21° slopes. Nevertheless, these estimates must be conservative, because only gravity fall of lifted soil was assumed rather than the tumbling or tumble plus slide models. Soons and Rayner, and Schumm found a significant increase in sediment yield from runoff over areas in which needle ice had been observed growing, indicating that needle ice may be a major contributor of soil loss, at least for bare exposed areas. (10) In winter especially, the cycles of rain followed by clear cool weather, but with afternoon temperatures above freezing, suggest that the Georgia Piedmont is an ideal location for needle ice induced erosion. However, needle ice erosion takes another less prominent form which may have effects more severe than loosening and transporting soil; the overturning and dislodging of vegetation. Cryoturbation is the accepted term for the overturning and mixing of soil by frost action. (11) This term has also been applied to the loosening of grassy clods by needle ice action; however Troll preferred the name turf exfoliation for overturning and dislodging by needle ice. ( J2) Turf exfoliation was at one time thought to be an eolian process wherein deflation of bare ground undermined adjacent turf cover. Needle ice is now recognized as the undermining process with eolian erosion a secondary process. 40Southeastern Geographer Observations of lateral penetration of turf cover and stripping of grass were made following several needle ice events at the study site in Athens, Georgia. Exfoliation occurs when the freezing plane descends and needle ice grows beneath an exposed turf edge and cleaves some portion of the grass from the main part. An entire strip of turf can be so removed. This is a positive feedback mechanism, because the bare soil exposed is usually moist, encouraging more needle ice growth and additional turf exfoliation. Gradwell and Troll have described large areas of bare soil with some turf outliers which developed due to needle ice action in the Alps of New Zealand and mountain ranges of Europe, respectively . (13) Our measurements indicated a relationship between needle ice height and penetration into the turf; however, our measurements of exfoliation were too imprecise to be able to provide a quantitative relationship . Clearly turf exfoliation can occur with every needle ice event, but the controls of the process are complex and require further work. SUMMARY AND CONCLUSIONS. Erosive land use on the Georgia Piedmont is legendary as is the inherent ease with which these deeply weathered soils can be eroded. Trimble examined this accelerated erosion on the Piedmont, and from his discussion it is easy to see that needle ice must also contribute to keeping gully walls and bare areas free ofvegetation. (14) Ultimately the soil lifted and loosened by needle ice will be transported by fluvial processes. Our observations suggest that the climate on the Piedmont is ideal for needle ice, and the soil moistures and textures which are favorable to needle ice growth are also present. (J 5) When these soils are exposed to the forces which create needle ice as well as to the forces of running water, the result is selfevident . All of the measurements made in this study for needle size and water content, soil moisture and texture, and microclimate necessary for ice growth are in agreement with available literature. That needle ice can expand the perimeters of bare areas was visually verified in this study. Furthermore, during the period 27 January 1976 to 24 February 1976, a period of relatively favorable weather for needle ice, six measurable needle ice events were observed. During the winter of 1978-79, 22 events produced at least visually present ice crystals at a construction site where conditions for ice growth were favorable. Half ofthose events Vol. XX, No. 1 41 were judged to have disrupted the soil surface. The closely monitored event on 27 January 1976 suggests that even when a conservative estimate of soil movement is calculated, as much as 16.5 metric tons of soil per hectare will be transported one centimeter down a 21° slope. The observations and measurements conducted in this study suggest that the contribution by needle ice to accelerated erosion on the Piedmont is massive and has probably been underestimated. (1)A. L. Washburn, Periglacial Processes and Environments (New York: St. Martin's Press, 1973). (2)H. A. Ireland, C. F. S. Sharpe, and D. H. Eargle, The Principles of Gully Erosion in the Piedmont ofSouth Carolina, U.S. Department ofAgriculture, Technical Bulletin No. 633. Washington, D.C., GPO, 1939. (3)S. I. Outcalt, "An Algorithm for Needle Ice Growth," Water Resources Research, Vol. 7 (1971), pp. 394-400. (4)R. Geiger, The Climate Near the Ground. Translated by M. N. Stewart. (Cambridge, Mass.: Harvard University Press, 1950). (5)G. J. Bouyoucos, "Directions for Making Mechanical Analysis of Soils by the Hydrometer Method," Soil Science, Vol. 42 (1936), pp. 225-230. (6)S. A. Schumm, "Seasonal Variations of Erosion Rates and Processes on Hillslopes in Western Colorado," Zeitschrift für Geomorphologie, Suppl. Band 5 (1964), pp. 215-238; J. M. Soons and J. N. Rayner, "Micro-Climate and Erosion Processes in the Southern Alps, New Zealand," Geografiska Annaler, Vol. 50-A (1968), pp. 1-15; C. Troll, Structure Soils, Solifluction, and Frost Climates of the Earth (U.S. Army, Corps of Engineers, Snow, Ice and Permafrost Research Establishment. Translation 43). Wilmette, 111., 1958. (7)J. Dylik, "Soliflucxion, Congelifluxion and Related Slope Processes," Geografiska Annaler, Vol. 49-A (1967), pp. 167-177. (8)A. Higashi and A. E. Corte, "Solifluction: A Model Experiment," Science, Vol. 171 (1971), pp. 480^482. (9)Higashi and Corte, footnote 9. (10)Footnote 6. (11)K. Bryan, "Cryopedology—The Study of Frozen Ground and Intensive Frost-Action with Suggestions on Nomenclature," American Journal of Science , Vol. 244 (1946), pp. 622-642. (12)C. Troll, "Rasenabschälung (Turf Exfoliation) als periglaziales Phänomen der subpolaren Zonen und der Hochgebirge," Zeitschrift für Geomorphologie , Suppl. Band 17 (1973), pp. 1-32. (13)M. W. Gradwell, "Physical Properties and Instability in South Island High Country Soils," Proceedings: New Zealand Society of Soil Science, Vol. 5 (1962), pp. 18-21; C. Troll, footnote 12. (14)S. W. Trimble, Man-Induced Soil Erosion on the Southern Piedmont, 17001970 . (Ankeny, Iowa: Soil Conservation Society of America, 1974). (15)V. Meentemeyer and J. P. Zippin, "Soil Moisture and Texture Controls of Selected Parameters of Needle Ice Growth," Earth Surface Processes (forthcoming ). ...