Back to Forests of the Central Appalachians | Feedback

Soil Temperature and Forest Type, II

RF Mueller, March 2003

Abstract

    For the second year soil temperatures are related to several forest types and environmental conditions in the Central Appalachian Mountains. As during the previous year ( Mueller, 2002 }, the White Pine - Hemlock - Hardwood forest with a boreal component at Virginia's Ramsey's Draft had growing season soil temperatures lower than those of Oak - Hickory Ridge and Mesic Slope forests in the Shenandoah Valley. Soil temperatures as well as associated spring and stream water temperatures reflected the markedly warmer 2002 summer temperatures when compared with those of 2001.

    Accompanying the largely warmer 2002 temperatures was a considerable variation in precipitation, with an early season abundance giving way to a severe summer drought. The resulting soil moisture variation produced secondary effects on the soil temperatures. For example, soil temperatures in the Mesic Slope forest, dominated by northerly aspects, were markedly cooler than in 2001 during the early part of the season as a consequence of the greater soil moisture. This effect was also apparent at Ramsey's Draft but not in the Oak - Hickory Ridge forest. All three forest types exhibited a temporary slight rise in soil temperatures in September after temperatures had already declined. This late season rise in temperature is interpreted as a consequence of exothermic respiration of pent up organic matter as a result of decreased respiration during the preceding drought. It is thought to have been initiated by an observed slight increase in soil moisture, possibly as a result of the shortened daylight period.

    The 2002 Deep - source Spring temperatures at Ramsey's Draft were of the order of one degree C higher than in 2001 for most of the summer. However, during late summer they fell well below those of 2001, an effect attributable to the truncation by the drought of a shallow water component of the spring water.

    Soil temperature determinations at several other locations in the mountains fit the pattern of 2001, with high elevation/ cool habitat data falling on or below the Ramsey's Draft trend, while those of the North River Valley lay above. Especially impressive, but not unexpected, is the low soil temperature registered at Dolly Sods in West Virginia's Allegheny Mountains.

    A particular advantage of soil and water temperatures that emerges from the expanded data base of the two seasons is their capacity to integrate and faithfully reflect several climatic components such as air temperatures, moisture availability and, possibly, forest energy conversions.

Introduction

    In the my treatment of the 2001 data it was opined that soil and air temperatures were relatively independent of each other in the short term, While this may be true on a scale of hours or even days, it is only apparently so, as a consequence of heat exchange lag. The nature of the major response of soil to air temperature is well established (Armson, 1979) and is too obvious to require discussion here. However, small scale responses to a number of climatic perturbations are not only complex, but of considerable importance to forest ecology. In the 2001 data the primary relations observed were the topographic and elevational effects and the correlations between soil temperature and

    forest type. Secondary effects were short term soil temperature variations and inferred associated heat release and absorption. Also of interest was the smooth curve obtained for the temperature of the water that issued from a prominent spring, as well as the time lag of this temperature relative to the soil temperature, features which led the writer to dub this a "Deep - source Spring."

    The 2002 season, in contrast to that of 2001, was characterized by abundant early season precipitation and some cool temperatures followed by a summer of unusually high temperatures and prolonged and severe drought. While the 2002 soil and water temperature data generally lie above those of 2001, and the forms of the distributions broadly similar, there are important differences in detail which, by comparison of the two years, allow, at least tentatively, identification of the mechanisms that are operative in these perturbations.

    A characteristic common to both the 2001 and 2002 data is the small variation in soil temperature between stations. This is particularly true of the Ramsey's Draft data which, in 2002, except for one date, show a spread of only 1.5 deg C over a distance of nearly a mile ( 1.6 km ) and the time span of the several hours it took to conduct the survey. There is also no systematic difference between individual or even widely separated stations. Even in the Mesic Slope Forest, in which the stations extend through different elevations and aspects, the soil temperature spread, with the exception of a single date, are equal or less than 2.0 deg C. However, in this case the two ridge top stations with essentially flat aspects, registered systematically higher temperatures than the others ( with northerly aspects ). This uniformity of recorded soil temperatures in a given forest type lends credence to the idea that a fundamental component of the ecosystem is being measured. The importance of soil temperatures is supported by the data of Hadley and Foster ( undated ) who stated that "Soil temperature is the most important single environmental variable affecting hemlock forest carbon balance during the warm half of the year." There is no reason to believe that it is less important in mixed or deciduous forests.

    For a description of the instrument and methods employed in the temperature measurements, as well as their precision, the reader is referred to the previous report ( Mueller, 2002 ).As in our treatment of the 2001 data. it should always be kept in mind that, while some temperature distributions, including part of the Ramsey's Draft soil temperature trend, appear to define a smooth curve, other distributions indicate that this may be a fortuitous consequence of the large time intervals between measurements. However, from the perspective of two season's data, it appears likely that the water temperatures of the Deep - source Spring do define a smooth curve as a consequence of the greater continuity inherent in the subsurface hydrology.

Ramsey's Draft

    Descriptions of the Ramsey's Draft area as well as the Oak - Hickory Ridge and Mesic Slope Forests were provided in the previous report ( Mueller, 2002 ). However that report had a serious omission, namely that the Canada Hemlock ( Tsuga canadensis ), which forms an important component of the canopy in parts of the study area at Ramsey's Draft, has been seriously impacted by the Hemlock Wooly Adelgid ( Adelges tsugae ), with substantial loss of canopy density. Since it is likely that the thermal regime of the Valley is largely imposed by subsiding cool air from nearby highlands, it is not clear what impact the loss of Hemlock would have. However, although the Valley climate may in part have and external source, the retention of the cool air is likely to be influenced by the Hemlock.

    Table 1 lists median values of the soil temperature, soil temperature spreads and single value water temperatures for the stream and two springs, in particular the Deep - source Spring.

    Date  Temperature
    (Degrees C)
    Temperature Spreads
    (Degrees C)
    4-11-02soil
    spring
    stream
    8.5
    6.5
    8.5
    8 - 9.5
    4-26-02soil
    spring
    stream
    9.5
    9.0
    9.0
    8.5 - 10.0
    5-15-02 soil
    spring
    stream
    11.5
    11.0
    11.0
    11.0 - 11.5
    5-23-02 soil
    spring
    stream
    9.0
    11.0
    10.5
    8.0 - 10.0
    6-15-02soil
    spring
    stream
    16.0
    14.0
    17.0
    16.0 - 17.5
    7-2-02soil
    spring
    stream
    18.0
    15.0
    19.5
    17.5 - 19.0
    7-27-02soil
    spring
    stream
    19.0
    16.5
    19.5
    18.5 - 20.0
    8-21-02soil
    spring
    stream
    19.0
    16.5
    20.0
    18.5 - 20.0
    9-6-02soil
    spring
    adj. spring
    stream
    17.5
    15.5
    16.0
    17.5
    17.0 - 18.0
    9-19-02soil
    spring
    adj. spring
    stream
    18.5
    dry
    17.0
    17.5
    18.0 - 18.5
    9-25-02soil
    spring
    stream
    17.0
    dry
    17.0
    17.0 - 17.0

    Table 1: Median temperatures and temperature spreads for seven soil temperature stations and single value water temperatures representing the Stream and two springs, including in particular the Deep - source Spring, all at Ramsey's Draft. The "adj. spring" refers to a spring adjacent to the Deep - source Spring. For the dates 7 - 2 - 02 and 9 - 25 - 02 only five soil temperature stations are represented.

    Figure 1: 2002 soil and water temperatures for Ramsey's Draft, Virginia. Circles indicate median soil temperatures, rectangles single value water temperatures of a Deep-source Spring.

    It should be noted that the Deep - source Spring ceased flowing some time after 9 - 6 - 02, the date of the last recorded temperature. However, an adjacent spring, which registered a temperature 0.5 deg C higher than the Deep - source Spring on that date, was still flowing on 9 - 19 - 02. Given the higher temperature value of this spring, it is likely that it had a shallower source, although probably not connected to subsurface stream flow.

    The median soil and Deep - source Spring temperatures of Table 1 have been plotted in Figure 1. While both distributions show regularity, there is a marked dip in the soil temperature during May, as well as a curious short - lived rise in September. The May dip in temperature is most easily explained as a result of a cool rainy period ( writer's weather notes ) and is consistent with the data of Davidson et al ( 1998 ). The September rise is of greater interest and best left for later comparison with the other forest types. However, very pertinent here is the observation of an associated slight increase in soil moisture as observed during the temperature measurements. Similarly, the detailed characteristics of the Deep - source Spring distribution are best left for comparison with those of 2001.

    As in 2001, stream temperatures match the soil temperatures rather closely, with those for six out of eleven dates differing only 0.5 deg C from the median soil temperatures, and only three by as much as 1.5 deg C.

Oak - Hickory Ridge Forest

    The median soil temperatures and temperature spreads for the five stations as well as two water temperatures are listed in Table 2. The median soil temperatures have been plotted in Figure 2, which also shows the soil temperature distribution for Ramsey's Draft. As was true of the 2001 data, the Oak - Hickory Ridge Forest distribution is characterized by rather abrupt changes and lies well above that of Ramsey's Draft. Note also that the small September temperature rise present at the latter location, occurs here as well.

    DateTemperature
    (Degrees C)
    Temperature Spreads
    (Degrees C)
    4-1-02 10.08.5 - 12.0
    4-9-02
    marsh water
    12.5
    15.0
    9.5 - 12.5
    4-24-0212.010.0 - 14.0
    5-4-02 13.011.5 - 14.0
    5-12-0216.014.0 - 17.0
    5-26-0216.014.0 - 18.0
    6-5-0220.018.0 - 21.0
    6-17-0217.016.0 - 19.5
    7-6-02 20.519.0 - 21.5
    7-28-0221.020.5 - 22.0
    8-23-0222.020.5 - 22.0
    9-11-0219.018.5 - 19.5
    9-22-02
    hoof print in marsh
    20.0
    20.0
    19.0 - 20.5

    Table 2: Median soil temperatures and temperature spreads and several water temperatures for the Oak - Hickory Ridge Forest. The "marsh water" temperature refers to standing water among marsh vegetation, while the "hoof print in marsh" refers to water that filled an animal hoof print at the wetland edge near the soil temperature station. Both values were obtained at sites close to the soil temperature station at the Ridge base.

    Figure 2: 2002 median soil temperatures for a Oak - Hickory Ridge Forest in the Shenandoah Valley, Virginia (in black). The Ramsey's Draft soil temperature trend (in grey) is shown for comparison.

Mesic Slope Forest

    The median soil temperatures and temperature spreads are listed in Table 3 for the seven stations of the Mesic Slope Forest. The median soil temperatures have been plotted in Figure 3, which also shows the Ramsey's Draft distribution.

    DateTemperature
    (Degrees C)
    Temperature Spreads
    (Degrees C)
    4-2-02 8.07.5 - 9.5
    4-15-0212.011.5 - 13.0
    4-30-0211.010.5 - 12.0
    5-8-02 13.513.5 - 14.5
    5-19-0211.511.0 - 13.0
    5-29-0215.015.0 - 16.0
    6-16-0216.016.0 - 17.0
    7-9-0220.018.5 - 21.0
    7-31-02 21.020.0 - 21.5
    8-26-0219.519.0 - 21.0
    9-8-0218.017.0 - 19.0
    9-20-0218.5 18.5 - 20.0

    Table 3: Median soil temperatures and temperature spreads for the Mesic Slope Forest

    Figure 3: 2002 median soil temperatures for a Mesic Slope Forest in the Shenandoah Valley, Virginia (in black). The Ramsey's Draft soil temperature trend is shown for comparison (in grey).

    The distribution is characterized by sharp early season fluctuations, One of these fluctuations, the sharp temperature drop in May, appears to coincide with that in the Ramsey's Draft data when the time spacing is considered. As mentioned earlier, this feature may be attributed to a cool rainy period, which included the entire region. Explanation of the abrupt September increase is best left to our summary of all forest types.

Additional Temperatures From the Mountains

    Table 4 and Figure 4 exhibit soil and water temperature data from four locations in the mountains of Virginia and West Virginia. All data represent single value measurements. As in the other figures, the Ramsey's Draft soil temperature trend is shown for comparison. As in the case of the 2001 data, the values obtained are consistent with the forest types and the environments represented. The North River soil and water temperatures fall above the Ramsey's Draft trend but a little below that of the Mesic Slope Forest (Figure 3).The data from Bear Mountain (see our section on this area) are represented by a 1330 meter elevation ridge crest value, a 1250 meter south slope value and a 1180 meter value from the Spruce forest on a valley flat. That the soil and water temperature values all fall close to but a little below the Ramsey's Draft trend is consistent with the elevations and topographic positions, although somewhat lower values might have been expected.

    Date Location Elevation
    (meters)
    Aspect Temperature
    (Degrees C)
    5-24-02 N. River 647 flat 11.0
    5-24-02 N. River ~746 flat 10.5
    5-24-02 N. River ~746 spring flat 11.0
    5-24-02 N. River ~746 flat 11.5
    6-7-02 Bear Mtn. 1330 flat 13.0
    6-8-02 Bear Mtn. 1180 flat 12.5
    6-8-02 Bear Mtn. 1180 stream flat 13.0
    6-9-02 Bear Mtn. 1250 S 13.5
    7-17-02 Bl. Springs 525 N 18.0
    8-8-02 Dolly Sods 1186 flat 16.0

    Table 4: Single value soil and two water temperatures from four locations in the Central Appalachian Mountains

    Figure 4: 2002 single value soil and water temperatures at various locations in the Central Appalachians (Table 4). The Ramsey's Draft soil temperature trend is shown for comparison.

    The single soil temperature from Blowing Springs is somewhat atypical for the area, but is consistent with the northern aspect from which it was obtained. This aspect at an elevation of only 515 meters asl is marginally tolerated by a few boreal species found there. These include Millet Grass (Milium effusum) and a single occurrence of Canada Mayflower (Maianthemum canadense). Consequently the position of the temperature value a little below and near the Ramsey's Draft trend is not beyond the realm of expectation.

    The impressively low value of the Dolly Sods soil temperature is consistent with what is known of the site ( see our section on the area ). Although not very high at 1180 meters asl, it falls at the edge of a large plateau that has a configuration especially suited to cold air accumulation, and additionally falls in the area of the Allegheny cloudy - day maximum ( Reifsnyder and Lull, 1965 ). Reflecting these conditions is the occurrence here of as many high elevation and boreal species as at any location in the Central Appalachians.

Comparison of the 2002 and 2001 Data and Summary

    It is informative to compare the 2002 data on soil and water temperature distributions with those of 2001.Figures 5 and 6 compare the median soil temperatures and single value Deep - source Spring temperatures for Ramsey's Draft, while Figures 7 and 8 compare the distributions for the Oak - Hickory Ridge and Mesic Slope Forests. These figures at once illustrate the fundamental similarity of the 2002 and 2001 distributions for each forest type, and the effects of variations in air temperatures and precipitation. We have already mentioned the role of early season precipitation in the formation of the observed spikes of low soil temperatures at Ramsey's Draft and in the Mesic Slope Forest. In the case of the latter we can see in Figure 8 an incipient early April correspondence of rapidly increasing temperatures in the leafless forest for the two years. In 2002 this was soon undone by heavy rains that resulted in the first downward spike. However, as the forest in late April was still largely leafless, soil temperatures again began to climb toward the second maximum. Once full leafing out occurred in early May of 2002, and as the rains continued, the second sharp downward spike, shared by Ramsey's Draft, occurred.

    Although hinted at in Figure 7 for the Oak - Hickory Ridge Forest, the variations so clearly shown for the Mesic Slope Forest, are virtually absent in the data of the former. It is possible that this is a consequence of the moderating effect of the associated wetland cold trap ( Mueller, 2002 ) or of the prevailing southerly slopes in the case of the Oak - Hickory Ridge Forest. In any case, the differences between the early season distributions for the two years in that forest fall within the uncertainty of the temperature measurements.

    The observed cooling effect of high precipitation has important implications in floristics. Davidson et al (1998) have shown that the relations between such factors as soil moisture, temperature and rate of respiration are very complex, and that soil respiration at first rises with increasing moisture at low moisture contents, but declines, at least under high moisture conditions. We should also remember that respiration of organic matter in soil is exothermic, as is clearly shown by such human activities as composting, in which high temperatures may be reached. The cooling effect of high precipitation, as indicated in Figures 1,3 and 8, may be offset in part or entirely by the heat generated by respiration, but at high moisture content the cooling process may dominate. It may be that this interplay lies behind the occurrence of some boreal species south of their normal ranges and in southern forest types. An example is the frequent trace of Yellow Birch (Betula alleghaniensis) found by Braun (1950) in the moist Cumberland Mountain mixed mesophyte forest, and the greater amount of this species in such forests that were rich in Hemlock. Similarly. Newell et al (1997 and 1998) found Yellow Birch to be a component of low elevation alluvial forests of the Joyce Kilmer - Slickrock and Linville Gorge Wilderness Areas of North Carolina. While a factor in the latter occurrences is the subsidence and accumulation of cool air from surrounding high mountains on flat bottomland, the relatively high precipitation of this Southern Appalachian region may also be involved.

    We have already referred to the abrupt September rise in soil temperatures in the three intensively studied forest types. Although not easily subject to proof, we may speculate that this rise is attributable to the same factors discussed above and related to regional conditions. As this rise was preceded by a long period of severe drought in the entire region, we should expect in late summer a pent up store of dead soil organic matter that is the normal consequence of herb and woody plant senescence and dormancy. Ordinarily, in times of normal soil moisture, this material would have been consumed by soil respiration . As daylight hours shortened, we should expect first some cooling as displayed in the figures, and with this an increased soil moisture, as was also observed at this time. Under these circumstances we might expect a rapid transient rise in soil temperature as a consequence of increase exothermic respiration of the pent up organic matter.

    The effect of the drought on springs, and most particularly on the Deep - source Spring was, we may infer, of an equally interesting nature, As shown in Figure 6 the 2002 distribution descends rapidly below that of 2001 as the drought attains its greatest severity. This cooling effect of the drought on spring water may be understood if we consider that this water is likely composed of components with varying residence times at different depths and, that during a drought, the shallow, an hence warmer components, are less likely to be replenished.

    It is also of interest to compare the Deep - source Spring temperatures with soil temperatures at a depth of 50 cm, as obtained by Crews and Wright (2000) in the Fernow Experimental Forest and vicinity in West Virginia's Allegheny Mountains. These authors attempted to determine the extent of the occurrence of "frigid soils", or those soils with a mean annual temperature in the range of 1 to 7 deg C at a depth of 50 cm and with the difference between mean winter and summer temperatures in excess of 5 deg C. While only one soil was classified as frigid for two consecutive years, and that on a summit at 1174 meters asl, several other soils at lower elevations and other aspects yielded mean summer temperatures comparable to the Deep - source Spring data. For example, a 1996 mean summer temperature value of 14.3 deg C was obtained at a 50 cm depth at an elevation of 762 meters and an E/SE aspect on Fork Mountain. In the same year and on the same mountain and elevation but on a SW aspect, a value of 14.2 deg C was obtained for the mean summer temperature. If the mean Deep - source Spring water temperature is calculated for the 2001 June, July and August data points (Mueller, 2002), a value of 14.2 is obtained , which almost identical to the cited 50 cm data. However, if the mean is determined for the comparable 2002 data (Table 1) the obtained value of 15.5 deg C clearly reflects the unusually high air temperatures of that year. These results are however in general agreement with the postulated deep source of this spring water.

    Interesting, but somewhat unexpected, is the close correspondence of the median soil temperatures and the stream water temperatures at Ramsey's Draft, especially inasmuch as the latter have their origins in part at higher elevations and in deeper source springs as well as near surface waters. While exposure to sunlight and air warming may be factors, these waters also flow out of sight beneath the stream's cobble bedload for substantial distances, especially during the very common summer dry periods.

Figure 5: Comparison of the median soil temperature trends at Ramsey's Draft for the years 2001 (grey line) and 2002 (black line).

Figure 6: Comparison of water temperature trends of the Deep - source Spring at Ramsey's Draft for the years 2001 (grey line) and 2002 (black line).

Figure 7: Comparison of the median soil temperature trends of the Shenandoah Valley Oak - Hickory Ridge Forest for the years 2001 (grey line) and 2002 (black line).

Figure 8: Comparison of the median soil temperature trends of the Shenandoah Valley Mesic Slope Forest for the years 2001 (grey line) and 2002 (black line).

Note: All graphs except Figure 6 are plotted directly from XML data using a generic Flash MX plotting movie. This was extensively expanded by Gus Mueller from a version found on 4GuysFromRolla.com. If you'd like to use this new version for your own data-plotting purposes, the .FLA source file is available for download. Questions about this code should be directed to Gus Mueller.

Acknowledgements

    The writer gratefully acknowledges the many fruitful experiences in the field as well as numerous stimulating discussions with Robert Hunsucker and Dorothy Simkins. The assistance of Gus Mueller with the figures and as web master is also appreciated, as is the aid provided by Patagonia Corp. and Save America's Forests. Finally, thanks are due to Ted Green and Jim Morrow of the U. S. Forest Service at Ashville, North Carolina for providing a document referenced in this study.

References

    Armson, K. A. (1979) Forest Soils: properties and processes, University of Toronto Press, Toronto.

    Braun, E. Lucy (1950) Deciduous Forests of Eastern North America. Macmillan Pub. Co. Inc., New York

    Crews, Jerry T. and Linton Wright (2000) Temperature and Soil Moisture Regimes in and Adjacent to the Fernow Experimental Forest. Research Paper NE - 713, Northeast Research Station, USDA Forest Service.

    Davidson, Eric A., Elizabeth Belk and Richard D. Boone (1998) Soil water content and temperature as independent or confounded factors controlling soil respiration in a temperate mixed hardwood forest. Global Change Biology 4, 217 - 228.

    Hadley, J. and D. Foster, "Dates since 1997," Site Prospect Hill Tract, A Physiological Model of CO2 Exchange by Hemlock Forest, contact J. Hadley. http://lnternet.edu/hfr/data/as007.html

    Mueller, R. F. (2002) Soil Temperature and Forest Type. Forests of the Central Appalachians Project. Virginians for Wilderness Web Site.

    Newell, Claire L., Robert K. Peet and Jonathon C. Harrod (1997) Vegetation of Joyce Kilmer - Slickrock Wilderness, North Carolina. Report to the U. S. Forest Service, Dept. of Agriculture, National Forests of North Carolina.

    Newell, Claire L. and Robert K. Peet (1998) Vegetation of Linville Gorge Wilderness, North Carolina. Castanea 63 (3), 275 - 322.

    Reifsnyder, William E. and Howard W. Lull (1965) "Radiant Energy in Relation to Forests." Technical Bulletin 1344, US Dept. of Agriculture, Forest Service, Washington, DC.

Back to Forests of the Central Appalachians | Feedback