HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Iwata, Y. Articles by Hirota, T. Search for Related Content PubMed Articles by Iwata, Y. Articles by Hirota, T. GeoRef GeoRef Citation Agricola Articles by Iwata, Y. Articles by Hirota, T. Related Collections Hydraulic Conductivity Flow Infiltration

ORIGINAL RESEARCH

Comparison of Snowmelt Infiltration under Different Soil-Freezing Conditions Influenced by Snow Cover

^a National Agricultural Research Center for Hokkaido Region, Sinsei, Memuro, Hokkaido, 082-0081 Japan
^b Dep. of Geoscience, Univ. of Calgary, Calgary, AB T2N 1N4, Canada
^c National Agricultural Research Center for Hokkaido Region, Hitsujigaoka 1, Toyohira-ku, Sapporo, Hokkaido, 062-8555 Japan
^* Corresponding author (iwatayuk{at}affrc.go.jp).

All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.

Received 9 May 2007.

	ABSTRACT

TOP ABSTRACT INTRODUCTION Study Site and Methods Results and Discussion Conclusions REFERENCES

 
The depth of soil frost is decreasing in cold regions around the world as a result of climate warming. To evaluate the potential impacts of the reduction in frost depth on the hydrologic cycle, it is necessary to understand snowmelt infiltration processes in frozen soils. A field study was conducted at an agricultural site characterized by volcanic ash soil in Tokachi, Hokkaido, Japan, where frost depths have decreased significantly in the last 20 yr. Soil temperature, water content, matric potential, snow cover, and meteorological parameters were monitored to quantify snowmelt infiltration flux for four winters that had different snow and soil conditions. When snowmelt began, the soil frost was 0.1 to 0.2 m thick in two winters and was absent in two other winters, providing a unique opportunity to compare snowmelt infiltration under frozen and unfrozen conditions. Most of the snowmelt water infiltrated into the soil under both frozen and unfrozen conditions, indicating that the frozen soil layer did not impede infiltration. The lack of flow impedance in the frozen soil was partly due to relatively high air temperature and an absence of freeze-back events during the snowmelt period. Furthermore, the temperature of the frozen soil layer was close to 0°C when the melt started, meaning that very little meltwater refroze in the soil before the temperature reached 0°C. The thick (>1 m) snow cover insulated the soil surface, allowing the frozen soil layer to warm up with the upward conduction of heat from the unfrozen layer below. These results indicate the importance of the interaction between snow cover and soil, which can be significantly affected by climate change.

Abbreviations: SWE, snow water equivalent • WCR, water content reflectometer

	INTRODUCTION

TOP ABSTRACT INTRODUCTION Study Site and Methods Results and Discussion Conclusions REFERENCES

 
The depth of soil frost and the length of the frozen period are decreasing in cold regions around the world as a result of climate warming (e.g., Cutforth et al., 2004; Frauenfeld et al., 2004; Hirota et al., 2006). The reduction of frost depth and the frozen period has important implications for the transport of soil water and dissolved mass in winter and early spring because the condition of the frozen soil layer strongly affects the amount and timing of snowmelt infiltration.

Soil hydraulic conductivity under frozen conditions is generally lower than under unfrozen conditions, which reduces snowmelt infiltration rates and generates runoff (Johnsson and Lundin, 1991; Gray et al., 2001; Bayard et al., 2005). When the soil has a relatively high intrinsic permeability and the amount of snowmelt is relatively small, however, snowmelt water may infiltrate into frozen soil completely without generating runoff (Granger et al., 1984; Stähli et al., 1999). Depending on snowpack thickness and air temperature before snow accumulation, the soil may become frozen or unfrozen in different years (Bayard et al., 2005). Thus, snowmelt infiltration processes are affected by many factors including soil temperature, frost depth, pre-winter water content, snowpack thickness, and their complex interactions (Stähli, 2005). Due to this complexity, it is difficult to predict how changes in climate and soil-freezing conditions may affect environmentally significant problems such as soil erosion and the transport of nutrients. Models to describe snowmelt infiltration have been developed by researchers working in North America (e.g., Gray et al., 2001) and northern Europe (e.g., Stähli et al., 1999), but the applicability of these models to other regions with different climates and soils needs to be carefully evaluated.

Seasonally frozen soil has occurred widely in the Tokachi district of northern Japan (Fig. 1 ), which is considered one of the most important crop production areas of Japan; however, the frost depth has decreased significantly during the past 20 yr (Hirota et al., 2006). The soil of the Tokachi district is dominated by volcanic ash soils (Andisols), characterized by high porosity and hydraulic conductivity (Hasegawa et al., 1994). As a result, the Tokachi soils have different infiltration characteristics than heavy-textured soils in continental cold regions such as the North American prairies (e.g., Gray et al., 2001). The climate of Tokachi is also different from the continental climate in that >1 m of snow accumulates in the plains area due to its proximity to the sea.

View larger version (25K): [in this window] [in a new window]   FIG. 1. Location and soil characteristics of the study site. Soil color is indicated by the Munsell color code. For soil texture, SiL is silt loam, CL is clay loam, C is clay, and SL is sandy loam.

Snowmelt infiltration into frozen soils has been quantified by percolation tests using ring infiltrometers (Kane and Stein, 1983; van der Kamp et al., 2003), monitoring of the total soil water content (Granger et al., 1984; Gray et al., 2001), and monitoring of runoff (Stadler et al., 1996; Bayard et al., 2005). Dye tracing methods have also been used to examine the flow pattern in frozen soils (e.g., Stähli et al., 2004). Percolation tests are useful for measuring infiltrative ability in the frozen soil, but they do not represent the natural condition because snow cover needs to be removed and an artificial ponding condition needs to be imposed. Monitoring total soil water content is effective for observing infiltration under natural conditions; however, the infiltration front may penetrate below a normal monitoring depth (e.g., 1 m) in regions with a large amount of snowmelt, making it impossible to quantify the infiltration rate from the water content data alone.

To overcome these problems, we established a soil and snow monitoring site in an agricultural field in Tokachi, equipped with tensiometers, soil temperature and water content sensors, and meteorological instruments in the fall of 2001 (Iwata and Hirota, 2005a,b; Hirota et al., 2005). This study examined the data from the winters of 2001–2002 and 2002–2003 when the soil was frozen to a maximum depth of approximately 0.2 m (hereafter frozen winter), and of 2003–2004 and 2004–2005 when the soil was not frozen at the time of snowmelt (hereafter unfrozen winter). The interannual variability of frost depth was primarily caused by the variability in the condition of the snow cover, which gave us a unique opportunity to compare the dynamics of snowmelt infiltration under frozen and unfrozen conditions at a single site. The objectives of this study were to quantify snowmelt infiltration rates in each year, evaluate the effects of frozen soil on infiltration, and determine the factors controlling the infiltration processes.

	Study Site and Methods

TOP ABSTRACT INTRODUCTION Study Site and Methods Results and Discussion Conclusions REFERENCES

 
The study was conducted at an experimental plot operated by the National Agricultural Research Center for Hokkaido Region, located in the town of Memuro in the central part of Tokachi district, Hokkaido, Japan (northernmost 42.59, southernmost 42.53, easternmost 143.05, westernmost 143.04; Fig. 1). The 1979 to 2000, mean annual precipitation and mean annual air temperature, recorded at Memuro meteorological station, located 2.5 km west of the site (Japan Meteorological Agency, 2007), are 970 mm and 6°C, respectively. The soil of the study site is derived from volcanic ash, which accumulated between 10,000 and 1000 yr ago (Kikuchi, 1981) and has the texture and horizons shown in Fig. 1. The soil is classified as a Typic Hapludand (Soil Survey Staff, 2006). The water table in this area is located approximately 8 m below the ground surface (Oka, 2000). Due to high porosity and saturated hydraulic conductivity (Fig. 1), the study site is considered well drained. The site has very little topographic relief (<1% slope) and has not been cultivated since the study started in 2001. Herbicides were used to keep the soil surface devoid of vegetation during the study period.

Soil matric potential was measured with tensiometers specifically designed for monitoring the potential of unfrozen soil below the frozen soil layer (Iwata and Hirota, 2005a,b), installed at depths of 0.5, 0.6, and 0.7 m. Liquid soil water content was monitored by Campbell Scientific (Logan, UT) CS615 water content reflectometers (WCRs) installed at a depth of 0.05 m and every 0.1 m from 0.1 to 1 m. The WCRs were installed horizontally and located approximately 5 m from the tensiometer nest. Soil temperature was monitored by thermocouples at depths of 0, 0.02, 0.05 m and subsequently every 0.1 m from 0.1 to 1 m. Thermocouples, pressure transducers for the tensiometers, and the WCRs were continuously monitored at 10-min intervals by a Campbell Scientific CR-10X datalogger. We continuously monitored snow depth, using an ultrasonic snow-depth gauge (Kaijo Corp., Tokyo, SL-340), air temperature at 1.9 m above the ground (Vaisara, Finland, HMP45A), and precipitation by an overflow-type tipping-bucket rain gauge with a heated water reservoir and a windshield (Yokogawa Electric Corp., Tokyo, RT-4). Frost depth, snow depth, and snow water equivalent (SWE) were measured manually once or twice a week. Frost depth was measured using a frost tube consisting of a 25-mm external diameter acrylic tube filled with 0.03% methylene blue solution, which has a dark blue color under unfrozen conditions and turns colorless on freezing. The frost depth was indicated by the boundary between the blue and colorless parts of the tube (e.g., Kinoshita et al., 1967). The acrylic tube was suspended in a 38-mm external diameter polyvinyl chloride (PVC) casing installed in the soil. The SWE was measured using a 50-mm internal diameter aluminum snow survey tube.

Soil porosity, bulk density, and saturated hydraulic conductivity of each soil horizon (Fig. 1) were determined for triplicate soil cores (50 mm in length and 51 mm in diameter). To calibrate WCRs for the specific soil of the site, large volumes of soil samples were air dried and repacked in cylindrical PVC containers (83 mm internal diameter and 340 mm in length) to approximately the same bulk densities as those of each soil horizon. The WCRs were then installed in these samples and the soil water content in each sample was adjusted to a desired value by adding distilled water. Because the WCR measures the apparent dielectric constant of the soil, which is most sensitive to liquid water, it was assumed that the WCR-measured water content in the frozen soil is the amount of liquid water remaining in the frozen soil (Flerchinger et al., 2006). The liquid water content measured by WCR at 0.05-m depth quickly dropped as the soil started to freeze in December 2002, and reached a minimum value of 0.19 m³ m^–3 (Fig. 2e ) when the soil temperature reached the lowest value of –4°C (Fig. 2d). This represents the residual liquid water content in the frozen soil. The value was consistent with the residual liquid water content in the frozen soil of 0.2 m³ m^–3 measured on a similar volcanic ash soil sampled in Tokachi using time domain reflectometry (Suzuki et al., 2002), suggesting that the WCR gives reasonable values of liquid water content in frozen volcanic ash soils.

View larger version (32K): [in this window] [in a new window]   FIG. 2. Time series of (a) daily mean air temperature, (b) snow cover thickness, (c) frost depth, (d) daily mean soil temperature, and (e) daily mean liquid water content in the winter of 2002–2003. Snow cover and frost depth were observed at 0900 h each day. The data gap in soil temperature was caused by an instrument problem.

Since evaporation rates from the snow surface at the study site were very small (usually <1 mm d^–1) compared with snowmelt rates (Hayashi et al., 2005), daily snowmelt (M) was calculated by the following:

[1]

where SWE₁ and SWE₂ are consecutive measurements of snow water equivalent (SWE₁ measured earlier than SWE₂), days₁₂ is the number of days between the two SWE measurements, and P is the amount of precipitation during the time interval.

	Results and Discussion

TOP ABSTRACT INTRODUCTION Study Site and Methods Results and Discussion Conclusions REFERENCES

 
Interannual Variability of Soil Freezing and Snow Accumulation
In the winters of 2001–2002 and 2002–2003, the soil froze to depths of 0.21 and 0.17 m, respectively, and was still frozen at the time of snowmelt (Table 1 ). In contrast, the soil froze to a depth of only 0.04 m in 2003–2004 and 0.05 m in 2004–2005 and was unfrozen at the time of snowmelt. The maximum snow-cover thickness was slightly smaller and average air temperature during the soil-freezing period was slightly lower in the years that had deeper frost (Table 1); however, relatively minor differences in snow-cover thickness and air temperature cannot explain the large difference in frost depth (Hirota et al., 2006). To gain insights into the differences between the years with shallower and deeper soil frost, the winters of 2002–2003 and 2004–2005 are compared in the following analysis. The soil was frozen during the snowmelt period in 2002–2003 (Fig. 2) and unfrozen in 2004–2005 (Fig. 3 ). Similar trends were observed in the other frozen and unfrozen winters. In these figures, the first vertical line indicates the beginning of the snowmelt period and the second line indicates the day when a steady-state condition was reached in the top 1.05 m of the soil profile, as described below.

View larger version (30K): [in this window] [in a new window]   FIG. 3. Time series of (a) daily mean air temperature, (b) snow cover thickness, (c) frost depth, (d) daily mean soil temperature, and (e) daily mean liquid water content in the winter of 2004–2005. Snow cover and frost depth were observed at 0900 h each day.

In 2004–2005, the soil stopped freezing on 3 December, when the frost depth was only 0.05 m (Fig. 3c), due to an early development of thick snow cover (Fig. 3b). The soil thawed quickly after this date, during which time the temperature in the top 0.3 m increased. After the thawing episode, the soil temperature gradually decreased at all depths (Fig. 3d) but never refroze to any significant depth (Fig. 3c). Soil water content did not change appreciably throughout the winter (Fig. 3e), as the effects of soil freezing was minimal and significant snowmelt did not start until mid-March.

The comparison of the two winters shown in Fig. 2 and 3 demonstrates the importance of the timing of snow cover development in soil-freezing processes (Hirota et al., 2006). Soil temperature and water content responded rapidly to snowmelt infiltration in both years (Fig. 2 and 3, after the first line) as described in detail below.

Comparison of Snowmelt Infiltration Dynamics under Frozen and Unfrozen Soil Conditions
In both years examined above, the main phase of snowmelt started when daily mean air temperature exceeded 0°C (the first vertical line in Fig. 4a and 4b ). Snow water equivalent decreased almost monotonically after these dates (Fig. 4b) except during a brief period from 28–31 Mar. 2005. The air temperature stayed relatively high during the whole episode of snowmelt (Fig. 4a), indicating that significant freeze-back did not occur at this site in either year.

View larger version (22K): [in this window] [in a new window]   FIG. 4. Time series of (a) daily mean air temperature, (b) snow water equivalent (SWE) at 0900 h, (c) daily mean soil temperature at the soil surface (0 m) and 0.3 m, (d) daily mean value (solid circle) and hourly mean value (gray line) of liquid soil water content at 0.05 m, (e) daily mean matric potential head () at 0.5 m, and (f) the cumulative changes in liquid water stored in the 1.05-m soil profile (S) and cumulative snowmelt (M) from 1 March to 15 April.

In contrast to 2003, the soil was unfrozen at the beginning of the melt in 2005 (Fig. 3c). Soil temperature was already above 0°C at the surface and at 0.3-m depth before the melt started (Fig. 4c, right panel) and slowly decreased during the early stages of the melt (between the two lines in Fig. 4c). The pre-melt liquid water content at 0.05-m depth was much higher in 2005 than in 2003 because the soil water remained unfrozen and therefore existed as a liquid (Fig. 4d). Liquid water content gradually increased during the melt (Fig. 4d, right panel) in response to snowmelt infiltration.

The soil matric potential head at 0.5 m before the melt was much lower in 2003 than in 2005 (Fig. 4e), indicating the drier soil condition in 2003 due to the upward flow of water from the unfrozen zone to the frozen zone. The matric potential head at 0.5 m started to increase on the second day after the melt started in 2003 (Fig. 4e, left panel) and the first day after the melt started in 2005 (Fig. 4e, right panel), indicating that the infiltration front had reached this depth very quickly.

The cumulative change in the total amount of liquid water stored in the 1.05-m soil profile (S) was calculated from the WCR data, using 1 March as the start date for each year. In both 2003 and 2005, S was nearly constant and started to increase rapidly as the snowmelt infiltration started (the first line in Fig. 4f); S reached a steady value after approximately 1 wk (the second line in Fig. 4f) as the infiltration front went below the bottom of the soil profile monitored by the WCR. The increase in S during the first week of melt (between the vertical lines in Fig. 4f) was larger in 2003 than in 2005, because a larger amount of meltwater (M in Fig. 4f) infiltrated in 2003 and also because the soil was drier to start with in 2003 (Fig. 2e and 3e). The snowmelt period is divided into the early period characterized by the continuous increase of S (between the lines in Fig. 4) and the late period characterized by the steady S.

Estimation of Snowmelt Infiltration Rate
The S at the end of the early snowmelt period (second line in Fig. 4f) was 95 mm in 2003 and 44 mm in 2005, which were comparable to the cumulative snowmelt (M) of 88 mm in 2003 and 39 mm in 2005 during the same period (Fig. 4f). The S in 2003 may include the liquid water added by the phase change within the soil. To estimate the amount of water added by the phase change, we note that the thickness of the frozen layer was 0.09 m in the beginning of the early snowmelt period (Fig. 2c), and that the liquid water content at 0.05 m increased from 0.26 to 0.46 during the early snowmelt period (Fig. 4d, left panel). The increase was due to the combination of the phase change and snowmelt infiltration. If the increase (0.46 – 0.26 = 0.20) was entirely caused by the phase change, the amount of water added by the phase change in the 0.09-m frozen layer would be 18 mm. The actual amount added by the phase change was probably <18 mm. Thus, S resulting from infiltration had values very similar to M in both years and hence, almost all of the snowmelt water infiltrated into the soil during the early snowmelt period in both the frozen year (2003) and the unfrozen year (2005). The similar slopes of S and M in Fig. 4f during the early snowmelt period indicate that the infiltration rate was essentially determined by the snowmelt rate, which is similar to the cases of unlimited infiltration reported by Gray et al. (2001).

During the late snowmelt period, S was no longer equal to the infiltration rate because significant amounts of water were leaving the bottom of the 1.05-m soil profile. It is useful to estimate the late snowmelt infiltration rate, however, particularly in frozen years, to determine if unlimited infiltration continues even when the snowmelt rate increases toward the end of the melt period (Fig. 4b). Using the tensiometer data from the 0.6- and 0.7-m depths to calculate the hydraulic gradient between the two, the infiltration rates were estimated from the Darcy equation. To estimate unsaturated hydraulic conductivity K() as a function of liquid water content (), we noted that at 0.65 m, computed as the average of the measured at 0.6 and 0.7 m, was nearly constant and very close to 0.55 during the late snowmelt period in all 4 yr of the study (Table 2 ). Considering that the K– relationship is unaffected by hysteresis in most soils (e.g., Topp, 1969; Hasegawa and Sakayori, 2000), it is reasonable to determine K() in one year and use the same value for other years.

View this table: [in this window] [in a new window]   TABLE 2. The duration of the late snowmelt period; mean and standard deviation of soil matric potential head (), hydraulic gradient, and soil water content at 0.65 m; cumulative value of soil water flux at 0.65 m (q0.65) and snowmelt (M); and the q0.65/M ratio.

[2]

where S_0.65 is the change in soil water stored between the surface and 0.65 m. The Darcy equation calculates q_0.65 from K(_0.65), where _0.65 is the water content at 0.65 m and I_0.65 is the magnitude of the hydraulic gradient at 0.65 m:

[3]

It follows from Eq. [2] and [3] that

[4]

Equation [4] was used to calculate K(_0.65) during the late snowmelt periods of the unfrozen years (2004 and 2005) for those days on which the topsoil was clearly unsaturated and the hydraulic gradient at 0.65 m was stable. Equation [4] cannot be used for frozen years because the phase change may have significant effects on S_0.65. The average K(_0.65) was 1.3 x 10^–7 m s^–1 in both 2004 and 2005. Therefore, this value was used to represent K(_0.65) in all years.

For the late snowmelt periods of 2002 and 2003 (frozen years), q_0.65 was calculated using Eq. [3] for the entire period. For the late snowmelt period of 2004 and 2005 (unfrozen years), q_0.65 was calculated using Eq. [2] when the topsoil was clearly unsaturated and Eq. [3] when the topsoil water content was close to porosity, implying the possibility of runoff. Examples of daily data are shown in Fig. 5 for 2003 and 2005 as cumulative flux (q_0.65) starting on the first day of the late snowmelt period (second vertical line in Fig. 4). The late snowmelt period was much shorter in 2003 than in 2005, as the snow cover in 2003 was thinner to start with (see Fig. 4b, second line) and melted faster. Also shown in Fig. 5 are the cumulative change (S) in the total amount of liquid water stored in the 1.05-m soil profile and the cumulative snowmelt (M).

View larger version (24K): [in this window] [in a new window]   FIG. 5. Cumulative values of snowmelt (M), soil water flux at 0.65-m depth (q0.65), and the change in soil water storage in the 1.05-m profile (S) in the late snowmelt period of 2003 (top) and 2005 (bottom).

Similarly, q_0.65 and M increased at comparable rates for the entire late snowmelt period of 2003 (Fig. 5), when the former was calculated independently from the latter using Eq. [3]. These results suggest that the frozen soil did not impede snowmelt infiltration even during the late snowmelt period when the topsoil was wetter (Fig. 4d, left panel) and the snowmelt rate was faster (Fig. 4b, left panel) than during the early snowmelt period.

Using the same methods, the total soil water flux (i.e., the final value of q_0.65) and snowmelt during the late snowmelt period were calculated for all 4 yr of the study, along with the mean and standard deviation of the matric potential head and hydraulic gradient during the late snowmelt period (Table 2). The total soil water flux was 80 to 90% of the snowmelt in 2002 and 2003 (frozen soil), and 76 to 78% in 2004 and 2005 (unfrozen soil), suggesting that the presence of the frozen layer did not have a major effect.

Effects of the Snow-Cover Insulation and Air Temperature on Soil Freezing and Thawing
Bayard et al. (2005) compared snowmelt infiltration in frozen and unfrozen winters and found that the refreezing of snowmelt water at the soil surface and in the frozen layer reduced snowmelt infiltration at an alpine site having a thick (>0.9 m) snow cover generating a large amount of meltwater. In contrast to their results, a significant finding of this study was that the frozen soil layer did not impede snowmelt infiltration. This is partly due to reasonably high air temperatures during the snowmelt period (Fig. 4a) preventing freeze-back, which would cause the formation of an ice sheet at or near the soil surface (Stadler et al., 1996; Bayard et al., 2005). Another important factor is the insulation provided by the thick snow cover, which kept the temperature in the frozen layer very close to 0°C in 2003 (Fig. 2). To demonstrate the effects of frozen soil temperature, a simple energy balance calculation is presented below.

After the large snowfall event on 4 Jan. 2003 (Fig. 2b), the temperature of the frozen layer gradually increased to between –0.1 and 0°C just before the snowmelt period (Fig. 2d and 4c). When the soil temperature is below the freezing point, the infiltrating meltwater refreezes, releasing latent heat and warming the frozen layer toward 0°C (Zhao et al., 1997). If the initial soil temperature is sufficiently low, the amount of ice formed by refreezing may be large enough to reduce the infiltrative ability of the frozen soil. The volume of ice can be estimated from the initial soil temperature as follows.

The volumetric heat capacity (C) of frozen soil is given by

[5]

where C_s, C_w, and C_i are the specific heats of soil solids, water, and ice, respectively; _b is the dry bulk density of the soil; f_w and f_i are the volumetric fractions of water and ice, respectively; and _w and _i are the densities of water and ice, respectively. Using the data for volcanic ash soil measured by Kasubuchi (1977), C_s was estimated to be 837 J kg^–1 K^–1, whereas C_w = 4217 J kg^–1 K^–1, C_i = 2117 J kg^–1 K^–1, _w = 1000 kg m^–3, and _i = 917 kg m^–3 at 0°C (Dorsey, 1940; Budavari, 1996). From the soil samples obtained from the surface layer, _b was estimated to be 930 kg m^–3. Using the WCR data at 0.05-m depth just before the soil froze and at the beginning of snowmelt (Fig. 4d, left panel), f_w and f_i were estimated to be 0.23 and 0.22, respectively. Substituting these numbers into Eq. [5], C was calculated to be 2.2 x 10⁶ J m^–3 K^–1. Assuming a meltwater temperature of 0°C and the latent heat of fusion to be 333.6 x 10⁶ J m^–3 (Dorsey, 1940), only a very small amount of meltwater (0.0007 m³ m^–3) would be required to refreeze to raise the temperature in the 0.09-m-thick frozen layer from –0.1 to 0°C. If the temperature was –2°C and the frozen layer was 0.17 m thick, however, as observed on 2 Jan. 2003 (Fig. 2c and 2d), a much greater refreezing of meltwater (0.012 m³ m^–3) would be required to raise the soil temperature to 0°C, which could reduce the infiltrative ability of the frozen soil. These calculations demonstrate that the warming of the frozen layer due to snow-cover insulation was partly responsible for the lack of flow impedance in the frozen soil.

	Conclusions

TOP ABSTRACT INTRODUCTION Study Site and Methods Results and Discussion Conclusions REFERENCES

 
In the volcanic ash soil with high porosity and hydraulic conductivity, most of the snowmelt water infiltrated into the soil, both in the years of no soil frost and the years when the soil was frozen to 0.17- to 0.21-m depth. During the early snowmelt period, steady-state conditions were reached in the top 1 m of the soil profile within 1 wk, and a substantial amount (78–161 mm) of meltwater percolated through this layer during the late snowmelt period. This is in contrast to the observations commonly made in other cold-region agricultural areas around the world, where soil freezing reduces hydraulic conductivity and impedes snowmelt infiltration. The lack of flow impedance in the frozen soil was partly due to relatively high air temperature during the snowmelt period preventing any freeze-back events and also because the temperature of the frozen soil layer was close to 0°C when the snowmelt started, meaning that very little meltwater refroze in the soil before the temperature reached 0°C. The frozen soil temperature was close to 0°C because a thick (>1 m) snow cover insulated the soil surface, allowing the soil to warm up with the upward conduction of heat from the unfrozen layer below. The timing of thick snow cover development has become earlier in recent years, possibly due to climate change, causing the annual maximum frost depth to decrease gradually to a point where the soil does not freeze at all in some years. In the agricultural lands of the Tokachi district, characterized by the same soil and meteorological conditions as the study site, a large amount of snowmelt infiltration can transport nutrients deeper into the soil column in early spring, which may have significant implications for soil and environmental management.

	ACKNOWLEDGMENTS

 
We thank Syuichi Hasegawa for useful advice on monitoring pressure head below the frozen layer; Shinji Suzuki for helpful suggestions for our research; Kunihiko Katoh and Seiichi Yasui for assistance in determining the soil types; Kaoru Marutani for useful advice on the groundwater level at our research site; Akira Yorisaki and other staff of Climatec Inc. for help in site instrumentation; and Greg Langston for editorial comments on an earlier draft. The technical assistance of Masamitsu Fujiwara, Norihiro Hirai, Yoshikazu Sato, Yuji Kato, and other members of the Field Operation Section of the National Agricultural Research Center for Hokkaido Region is gratefully acknowledged. The study was partially funded by the Global Environment Research Coordination System Grant from the Japanese Ministry of Environment and an Invitation Fellowship to M. Hayashi from the Japan Society for Promotion of Science.

	REFERENCES

TOP ABSTRACT INTRODUCTION Study Site and Methods Results and Discussion Conclusions REFERENCES

 

• Bayard, D., M. Stähli, A. Parriaux, and H. Flühler. 2005. The influence of seasonally frozen soil on the snowmelt runoff at two Alpine sites in southern Switzerland. J. Hydrol. 309 :66–84.[CrossRef]
• Budavari, S. (ed.). 1996. The Merck index. 12th ed. Merck Publ. Group, Rahway, NJ.
• Cutforth, H., E.G. O'Brien, J. Tuchelt, and R. Rickwood. 2004. Long-term changes in the frost-free season on the Canadian prairies. Can. J. Plant Sci. 84 :1085–1091.
• Dorsey, N.E. 1940. Properties of ordinary water substances. Reinhold, New York.
• Flerchinger, G.N., M.S. Seyfried, and S.P. Hardegree. 2006. Using soil freezing characteristics to model multi-season soil water dynamics. Vadose Zone J. 5 :1143–1153.[Abstract/Free Full Text]
• Frauenfeld, O.W., T. Zhang, and R.G. Barry. 2004. Interdecadal changes in seasonal freeze and thaw depths in Russia. J. Geophys. Res. 109(D05101), doi:10.1029/2003JD004245.
• Granger, R.J., D.M. Gray, and G.E. Dyck. 1984. Snowmelt infiltration to frozen prairie soils. Can. J. Earth Sci. 21 :669–677.
• Gray, D.M., B. Toth, L. Zhao, J.W. Pomeroy, and R.J. Granger. 2001. Estimating areal snowmelt infiltration into frozen soils. Hydrol. Processes 15 :3095–3111.[CrossRef]
• Hardy, J.P., P.M. Groffman, R.D. Fitzhugh, K.S. Henry, A.T. Welman, J.D. Demers, T.J. Fahey, C.T. Driscoll, G.L. Tierney, and S. Nolan. 2001. Snow depth manipulation and its influence on soil frost and water dynamics in a northern hardwood forest. Biogeochemistry 56 :151–174.
• Hasegawa, S., S. Osozawa, and H. Ueno. 1994. Measurement of soil water flux in Andisols at a depth below a root zone of about 1 meter. Soil Sci. Plant Nutr. 40 :137–147.
• Hasegawa, S., and T. Sakayori. 2000. Monitoring of matrix flow and bypass flow through the subsoil in a volcanic ash soil. Soil Sci. Plant Nutr. 46 :661–671.
• Hayashi, M., T. Hirota, Y. Iwata, and I. Takayabu. 2005. Snowmelt energy balance and its relation to foehn events in Tokachi, Japan. J. Meteorol. Soc. Jpn. 83 :783–798.[CrossRef]
• Hirota, T., Y. Iwata, T. Hamasaki, R. Sameshima, and M. Hayashi. 2005. Micrometeorological conditions and the thermal and moisture characteristics of seasonally frozen soil in eastern Hokkaido. J. Agric. Meteorol. 60 :673–676.
• Hirota, T., Y. Iwata, M. Hayashi, S. Suzuki, T. Hamasaki, R. Sameshima, and I. Takayabu. 2006. Decreasing soil-frost depth and its relation to climate change in Tokachi, Hokkaido, Japan. J. Meteorol. Soc. Jpn. 84 :821–833.[CrossRef]
• Iwata, Y., and T. Hirota. 2005a. Development of tensiometer for monitoring soil-water dynamics in a freezing and snow covered environment. J. Agric. Meteorol. 60 :1065–1068.
• Iwata, Y., and T. Hirota. 2005b. Monitoring over-winter soil water dynamics in a freezing and snow covered environment using thermally insulated tensiometer. Hydrol. Processes 19 :3013–3019.[CrossRef]
• Japan Meteorological Agency. 2007. Normals 1971–2000. Available at www.data.jma.go.jp/obd/stats/data/en/normal/normal.html (verified 13 Dec. 2007). Jpn. Meteorol. Agency, Tokyo.
• Johnsson, H., and L.-C. Lundin. 1991. Surface runoff and soil water percolation as affected by snow and soil frost. J. Hydrol. 122 :141–159.[CrossRef]
• Kane, D.L., and J. Stein. 1983. Water movement into seasonally frozen soil. Water Resour. Res. 19 :1547–1557.[CrossRef]
• Kasubuchi, T. 1977. Thermal properties of soil. (In Japanese.) Soil Phys. Cond. Plant Growth Jpn. 35 :29–34.
• Kikuchi, K. 1981. Interpretative classification of the soils in Tokachi district. (In Japanese, with English abstract.) Rep. Hokkaido Prefectural Agric. Exp. Stn. 34 :1–118.
• Kinoshita, S., Y. Suzuki, K. Horiguchi, K. Tanuma, and M. Aota. 1967. Frost heave in Monbetsu (1966–1967). (In Japanese.) Low Temp. Sci., Ser. A 25 :229–232.
• Lindström, G., K. Bishop, and M.O. Löfvenius. 2002. Soil frost and runoff at Svartberget, northern Sweden: Measurements and model analysis. Hydrol. Processes 16 :3379–3392.[CrossRef]
• Nyberg, L., M. Stähli, P.-E. Mellander, and K.H. Bishop. 2001. Soil frost effects on soil water and runoff dynamics along a boreal forest transect: 1. Field investigations. Hydrol. Processes 15 :909–926.[CrossRef]
• Oka, T. 2000. Geological features and it's explanation at Middle Tokachi Plain. (In Japanese.) Geol. Surv. of Hokkaido, Sapporo, Japan.
• Soil Survey Staff. 2006. Keys to soil taxonomy. 10th ed. Available at soils.usda.gov/technical/classification/tax_keys/ (verified 27 Nov. 2007). NRCS, Washington, DC.
• Stadler, D., H. Wunderli, A. Auckenthaler, H. Flühler, and M. Bründl. 1996. Measurement of frost-induced snowmelt runoff in a forest soil. Hydrol. Processes 10 :1293–1304.[CrossRef]
• Stähli, M. 2005. Freezing and thawing phenomena in soils. p. 1069–1076. In M.G. Anderson (ed.) Encyclopedia of hydrological sciences. Vol. 2. John Wiley & Sons, New York.
• Stähli, M., D. Bayard, H. Wydler, and H. Flühler. 2004. Snowmelt infiltration into alpine soils visualized by dye tracer technique. Artic Antarctic Alpine Res. 36 :128–135.[CrossRef]
• Stähli, M., P.-E. Jansson, and L.-C. Lundin. 1999. Soil moisture redistribution and infiltration in frozen sandy soils. Water Resour. Res. 35 :95–103.[CrossRef]
• Suzuki, S., J. Kashiwagi, S. Nakawaga, and K. Soma. 2002. Mechanism of hysteresis in thermal conductivity of frozen soils between freezing and thawing processes. (In Japanese, with English abstract.). Trans. Jpn. Soc. Irrig. Drain. Reclam. Eng. 218 :97–105.
• Topp, G.C. 1969. Soil-water hysteresis measured in a sandy loam and compared with the hysteretic domain model. Soil Sci. Soc. Am. Proc. 33 :645–651.
• van der Kamp, G., M. Hayashi, and D. Gallén. 2003. Comparing the hydrology of grassed and cultivated catchments in the semi-arid Canadian prairies. Hydrol. Processes 17 :559–575.[CrossRef]
• Zhao, L., D.M. Gray, and D. Male. 1997. Numerical analysis of simultaneous heat and mass transfer during infiltration into frozen ground. J. Hydrol. 200 :345–363.[CrossRef]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Iwata, Y. Articles by Hirota, T. Search for Related Content PubMed Articles by Iwata, Y. Articles by Hirota, T. GeoRef GeoRef Citation Agricola Articles by Iwata, Y. Articles by Hirota, T. Related Collections Hydraulic Conductivity Flow Infiltration