MEASUREMENT OF FROST DEPTH by D. K. McCool¹ and Myron Molnau²
Proceedings Western Snow Conference 52:33-42.1984
Sun Valley, Idaho
James Meiman, General ChairmanINTRODUCTION
The northern tier in the United States experiences frozen soil of varying depth every year. Knowledge of frost depth is necessary for projecting water lines and designing roadbeds to prevent frost heaving. Hydrologists in the Pacific Northwest are particularly interested in the ground because of the area's winter precipitation regime which often results in flood occurring on frozen soil (Johnson and McArthur, 1973; Molnau and Bissell, 1983) or severe erosion caused by rain on thawing soils (McCool et al., 1976; Formanek, 1983). Numerous studies have been conducted in the Pacific Northwest on the effect of frozen soil on erosion (McCool, et al., 1982; Yen and Molnau, 1982; Zuzel et al., 1982; Formanek, 1983). These studies have shown that thawing soils have very low shear resistance and are very susceptible to soil movement from even minimal rainfall or snow melt. In order to design tillage systems that can minimize these problems and also to accurately predict erosion in the area, knowledge of the number of times the soil freezes and thaws during the winter and the depth of soil frost is essential. To obtain this knowledge, some means of measuring the occurrence and depth of frost is necessary.
The objectives of this manuscript are to evaluate various methods for determining frost depth and to report results of tests to determine use and performance of a simple and inexpensive gage for determining frost depth.
BACKGROUND
The classic method of determining the depth of frost is to dig a pit and observe the occurrence of ice crystal. This requires a great seal of time and effort and destroys an experimental area, so indirect methods have been devised to infer the location of a frost line. Garstka (1944), Harrold and Roberts (1960), and Burgess and Hanson (1979) used solid moisture blocks to detect the frost line. Conductivity approaches zero when moisture in the blocks freezes. Soil moisture blocks are an accurate\ means of determining frost depth but provide only point measurements. However, the blocks and instrumentation may be installed at remote sites and visited infrequently. Sartz (1967) reported the results of his tests of three indirect indicators of frost depth. These were moisture blocks, thermistors, and a frost meter, a tube containing a series of solution-filled glass bottles; the lowest broken bottle indicated frost depth. He concluded that the moisture blocks were more accurate that thermistors when compared to a penetrometer test.
The low cost, ease of fabrication, and ease of use by untrained personnel have led to numerous experiments to develop and improve the tube-type frost gages. Harris (1970) reported on tests of two water-dye filled tubes. These were a gage using a 0.01 percent solution of green fluorescein dye with medium-size clean sand in a polyethylene tube, similar to a gage developed by Gandahl (1957), but as modified at the Cold Region Research and Engineering Laboratory (CRREL), and a similar tube using a Kool-aid solution. He conclude the modified Gandahl gage performed better during the critical periods of fast freezing a thawing.
Rickard and Brown (1972) compared the performance of two frost tubes. The first was a modifies CRREL tube, as reported by Harris (1970), constructed of a 19.0 mm ).D. inner tube of polyethylene filled with sandblasting sand and a 0.1 percent fluorescein dye solution. The second tube was filled with a 0.1 percent methylene blue solution (no sand). They found that when the frost was penetrating, the modified CRREL tube responded to frost penetration faster, with the indicated frost depth within 20 mm of the 0 degrees Celsius isotherm. The methylene blue tube consistently lagged the modified CRREL tube by as much as 50 mm. During thawing periods, they found the methylene blue also lagged. They concluded that the modified CRREL sand-filled tube with green fluorescein was superior to the methylene blue tube because the added water mass of the methylene blue tube slowed the response time and the greater volume expansion of the liquid filled tubes resulted in physical damage. Also the blue solution tended to give unclear edges.
In 1976, Ricard at al. reported on a field assembled frost gage. This was referred to by one of the co-authors as a CRREL-Gandahl gage. This gage used a polyethylene inner tube of 12.7 mm I.D. and 15.9 mm O.D. and was filled with a 0.05 percent solution of methylene blue. A nylon string up the center of the tube anchored the ice and prevented floating during thawing periods. The outer casing was of rigid polyethylene or PVC pipe. It was specified that the inner diameter of the casing should be no more that 6 mm greater than the outer diameter of the inner tube. They recommended that the gage should be at least 0.30 m longer than the expected frost depth to minimize freezing point depression as dye migrates from the freezing point.
FROST GAGE DEVELOPMENT IN THE PACIFIC NORTHWEST
In 1974, personnel of the USDA-ARS at Pullman, Washington and of Agricultural Engineering Department of the University of Idaho began to search for a simple and inexpensive methods by which frost depth could be measured. A tube filled with fluorescein solution and sand as described by Harris (1970), but with a rigid inner tube, was first used by ARS during the winter of 1974/75. This design was also used by the University of Idaho in the winter of 1975/76. The design seemed to work fairly well for researchers and technicians acquainted with the idiosyncracies of the gage. However, many cooperative observers could not differentiate the color change from yellow-green to grey as the liquid froze. Also, super-cooling was a problem in that some tubes did not begin to freeze until exposed to very cold temperatures. There was also occasional difficulty with the rigid inner tube freezing and breaking is the gages were exposed to cold temperatures prior to installation. Thus, in the fall of 1978. Both USDA-ARS and University of Idaho adopted the CRREL-Gandahl design. The line between frozen and unfrozen material was much easier to locate, and if necessary, the ice blocks could be located by squeezing the tube.
There seemed to be some problems with freezing and thawing of the solution in the tubes lagging behind freezing and thawing of liquid in the soil. Thus, in the fall of 1979, 4 mm diameter glass beads were added to the methylene blue solution in the tube. The intent was to decrease the quantity of liquid in the tubes and hence thermal lag. However, the beads seemed to offer little advantage and were discontinued. Table 1 lists the gage types and their period of use.
TABLE 1. Frost Gage Development in the Pacific Northwest
¹ Harris, 1970
Period of use
Frost Gage type
1974/75 through 1977/78
Harris¹
1978/79
CRREL-Gandahl²
1979/80
CRREL-Gandahl²(glass beads)
1980/81 through 1983/84
CRREL-Gandahl²
² Ricard et al., 1976
Two main types of tubing have been used in the CRREL-Gandahl gages. The early CRREL-Gandahl tubes were constructed of Tygon tubing S-50-HL.³ More recently the tubes have been constructed of Nalgene 8000 tubing. Performance has been similar. Because the methylene blue stains the tube walls, it has been necessary to replace the tubes on an annual basis. Thus cost and availability are factors in tube selection.
CONSTRUCTION, INSTALLATION, AND USE OF FROST GAGES
Construction
The first frost gages are constructed of 12.7 mm I.D., 15.9 mm O.D. clear plastic tubing of about 1 m length, with plastic plugs in each end (Figure 1). A nylon string is stretched between the plugs, with a rubber band at the upper end keeping the string in tension. Approximately 0.90 m of the tubing is filled with a 0.04 to 0.05 percent by weight solution of methylene blue. An air gap of 50 to 100 mm is left at the top of the tube. An eye screw is placed in the upper plug and a string is attached for pulling the tube from the outer casing. A strip of self-adhesive metric tape is placed with zero at the liquid level. The outer casing is 23.3 mm I.D., 27.0 mm O.D. PVC pipe, sealed at the bottom and extending far enough above ground so that snow will not cover the tube. The casing is measured and a ground line mark is placed so that inserting the casing into the soil to that depth will place the liquid level of the inner tube at ground line when the bottom of the inner tube rests on the plug at the bottom of the casing. A self-adhesive metric tape is attached to the outside of the s\casing so that the snow depth can be quickly and easily measured.
Installation
The frost gage should be placed so that the top of the liquid in the tube is just at the soil surface. This is necessary to ensure accurate measurement of frost depth. An auger is used to make the hole for the casing. If the auger is appreciably larger than the casing, a mixture of sand and bentonite cane be used to fill the annulus between the casing and the soil. This will prevent air gaps, entry of surface water, and erratic results. If the auger is only slightly larger than the casing, silicone spray can be used to prevent the casing from sticking as it is inserted into the soul. If the hole is too deep soil can be placed below the casing to raise it. If the casing is accidentally raised while placing the sand and bentonite mixture around the casing, it may be necessary to abandon the hole and start over, as it is nearly impossible to auger excess sand from the hole.
The value of the frost depth data will be enhanced if daily maximum/minimum air temperatures are collected at or near the tube. These temperature data will permit calibration or verification of frost depth models. Thermometers should be placed so the are not in direct sunlight.
Use of Frost Gages
Frost depth observations should be taken daily when minimum temperature drops below 0 degrees Celsius or when the ground is thawing. Readings may be taken less frequently when the average daily temperature is greater than 0 degrees Celsius and conditions are stable. This is to ensure that the tubes are functional and are not leaking. Reading should be taken at the same time each day.
Even if frost depth is not read daily, it is desirable to read temperature and snow depth daily. Otherwise, it is difficult to use the data with frost depth models.
FROST GAGE PERFORMANCE
In the fall of 1980, and investigation was started to evaluate the performance of the frost gages. The experiment used soil moisture blocks as the standard against which the frost gage reading would be compared. A total of 10 moisture blocks were installed; the first five at 40 mm intervals from 40 to 200 depth; the next four at 50 mm intervals from 200 to 400 mm depth; and the last one at 500 mm depth. Two frost gages were installed about 1500 mm down slope from the blocks. Performance was tested under three ground covers; bare soil, standing stubble, and chisel/disked stubble. The experimental are was on a south-facing slope of 20 percent steepness. Resistance readings wee recorded automatically 8 times per day. Frost gage readings were taken once a day throughout the work week during periods of freezing and thawing. Moisture blocks were spaced closer together in the fall of 1983 since it was felt that this would result in a better correlation with frost gage readings. All 10 blocks were installed at 30 mm intervals fro 30 to 300 mm depth. Resistance readings were reduced from 8 to 4 times per day. A typical plot layout, frost gage, and soil moisture block horizonal and vertical spacing for the 1983/84 season are shown in figure 2 and figure 3.
During the winter of 1983/84, three separate cold periods caused appreciable frost penetration. One of these events, from mid-January to early February, will be used to illustrate frost gage performance. Plotting of data from the standings and bare soil plots are presented in figures 4a and 4b, respectively. Because the soil moisture blocks are installed at 30 mm depth intervals with readings at 6-hr intervals, the data appear as a step function. The frost gage readings are plotted as discrete observation points.
During freezing, frost depth readings obtained with the two methods are quite similar. Maximum frost depths are also quite similar. Because the maximum moisture block depth was 300 mm, no comparisons can be made below that depth. The lag of the frost gages becomes apparent during the thaw. The soil thawed rapidly from the surface and more slowly from beneath. The gages showed from 2 to 4 days of lag when thawing from beneath. Maximum depth lag was approximately 50 mm, similar to that reported by other investigators (Gandahl, 1957).
The difference in frost depths between the standing stubble and bare soils is appreciable; maximum frost depth on the standing stubble was 180 mm whereas that on the bare soils was 330 mm. There was no appreciable snow cover on either plot during this event.
DATA COLLECTION NETWORKS
The first frost gage data were collected by USDA-ARS at Pullman during the winter of 1974.75. The gages were installed at weather stations near plots used in runoff-erosion studies. Additional gages were constructed at Pullman and sent to USDA-ARS at Boise to be compared with instrumentation already in place. In the fall of 1975 additional gages were constructed. The University of Idaho constructed and installed gages in tillage plots near Moscow, Idaho and the USDA-ARS worked with USDA-SCS to establish a small cooperative network of observers in Whitman County, Washington. The data collected was expanded in the winter of 1977/78 with establishment of a frost depth network of 15 locations in 8 states as part of a Western Agricultural Experiment Stations Climate project. This experiment was discontinued following the winter of 1979/80, although some states are continuing their portion of the network because of the usefulness of the data. In the fall of 1978/79, frost gages were added to the USDA-ARS tillage plots near Pullman, Washington. An expanded network of cooperative observers was established in the fall of 1980. This network includes from 85 to 100 tubes in eastern Washington, a small number of tubes in northern Idaho, and a small number of tubes in southeaster Utah. Data collected have not been fully exploited but have proven useful for preliminary work with frost depth and erosion mapping. A summary of the data collection networks is presented in Table 2.
SUMMARY AND CONCLUSIONS
Two tube-type frost gages have been used to determine the depth of soil frost in experiments in the Pacific Northwest. The first tube used fluorescein dye solution in a sand-filled tube. At the dye concentration used in experiments, the tube was difficult for many observers to read. Later, a methylene blue dye solution was used in a tube filled with liquid. This tube has been easy to read but exhibits lag during rapid temperature changes, especially during thawing periods. Performance in indicating maximum frost penetration during extended freezing periods has been adequate.
Tube-type frost gages can provide good frost depth information at low cost. Goos temperature and snow depth records are essential to the data set.
¹ Agriculture Engineer, Land Management and Water Conservation Research Unit, USDA- RS, Pullman Washington
² Professor, Agricultural Engineering Department, University of Idaho, Moscow, ID
³ Trade names are included for the benefit of the reader and do not imply endorsement or preferential treatment of the product by the U.S. Department of Agriculture.
TABLE 2. Data Collection Networks in the Pacific Northwest
1974/ 75
75/76
76/77
77/78
78/79
79/80
80/81
81/82
82/83
83/83
Pullman Weather Station
x
x
x
x
x
x
x
x
x
x
Reynolds Creek Watershed
x
x
x
x
x
x
x
x
x
x
Moscow STEEP Plots
x
x
x
x
x
x
x
x
Whitman County Cooperative Network
x
x
x
x
Regional Agric. Exp. Station Project
x
x
x
Small Regional Cooperative Network
x
Expanded Regional Cooperative Network
x
x
x
x
Individual State Experiment Stations
x
x
x
x
Pullman STEEP Plots
x
x
x
x
x
x
REFERENCES
Burgees, M.D. and C.L. Hanson. 1979. Automatic soil-frost measuring system. Agricultural Meterology 20:313-318.
Formanek, G.E. 1983. Shear strength of thawing silt loam soil: An approach to erodibility. Master of Science thesis. Washington State University, Pullman.
Gandahl, R. 1957. Determination of the depth of soil freezing with a new frost meter. (Text in Swedish) Rapport 30, Stattensaginsititut Stockholm (SIP 16347) Grunforbattring m. 1, pp. 3-15. Cited by Rickard, W. and J. Brown. 1972. The performance of a frost-tube for the determination of soil freezing and thawing depths. Soil Science 113(2):149-154.
Garstka, W.U. 1944. Hydrology of Small watersheds under winter conmditions of snow-cover and frozen soil. Transactions, American Geophysical Union, Vol. 25, Part VI, Section of Hydrology Papers, pp. 838-871.
Harris, A.R. 1970. Direct reading frost gage is reliable, inexpensive. Research Note NC-89. North Centeral Forest Experiment Station, Forest Service, U.S. Department of Agriculture.
Harrold, L.L and R.L. Roberts, Jr. 1960. Winter runoff from snow and frozen gound. Michigan Quarterly Bulletin 43(1):154-169.
Johnson, C.W. and R.P. McArthur. 1973. Winter storm and flood analysis, Northwest interior. Proceedings of the 21st Annual Hydraulics Division Speciality Conference. Hydraulic Engineering and the Environment, ASCE, Bozeman, Montana, August 15-17, pp. 359- 369.
McCool, D.K., Myron Molnau, R.I. Papendick, and F.L. Brooks, Jr. 1976. Erosion research in the dryland grain region of the Pacific Northwest: recent developments and needs. In Soil Erosion: Prediction and Control, SCSA, Ankeny, Iowa.
Molnau, Myron and V.C. Bissell. 1983. A continous frozen ground index for flood forecasting. Proceedings, Western Snow Conference, Vancouver, WA, pp. 109-119.
Ricard, J.A., W. Tobiasson, and A. Greatorex. 1976. The field assembled frost gage. Technical Note. Cold Regions Research and Engineering Laboratory, U.S. Army Corps of Engineers, Hanover, New Hampshire.
Rickard, W. And J. Brown. 1972. The performance of a frost-tube for the determination of soil freezing and thawing depths. Soil Science, Vol 113, No. 2, pp. 149-154.
Sartz, R.S. 1967. A test of three indirect methods of measuring depth os frost. Soil Science, Vol. 104, No. 4, pp. 273-278.
Yen, E.S. amd Myron Molnau. 1982. Frozen soil effects on the erosion hazard. Research Technical Completion Report. Project A-070-IDA, Idaho Water and Energy Resources Research Institute. University of Idaho, Moscow, Idaho, 66 pages.
Zuzel, J.F., R.R. Allmaras, and R. Greenwalt. 1982. Runoff and soil erosion on frozen soils in northeastern Oregon. Journal of Soil and Water Conservation 37(6):351-354.
