| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a USGS-FRESC, 3200 Jefferson Way, Corvallis, OR 97331
b Dep. of Crop and Soil Sciences, Oregon State University, Corvallis, OR 97331
c Prineville Office, Bureau of Land Management, 3050 NE Third Street, Prineville, OR 97754
d Burns Office, NRCS, HC 74, 12858 Highway 20 W., Hines, OR 97738
e Burns Office, Bureau of Land Management, HC 74, 12533 Highway 20 W., Hines, OR 97738
* Corresponding author (David_Clausnitzer{at}usgs.gov)
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
Abbreviations: d subscript, dithionite-citrate extractable • NRCS, Natural Resource Conservation Service • NWCC, National Water and Climate Center • o subscript, oxalate extractable
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
The National Technical Committee for Hydric Soils (USDA-NRCS, 2002) has developed criteria for identification of hydric soils. Applicable criteria to our vernal pool are soils that are frequently ponded for a continuous period of at least 7 d during the growing season, with a >50% probability in any one year. The growing season is the portion of the year when soil temperatures are above biological zero in the upper part. Biological zero is 5°C at 50 cm below the soil surface. Soil redox potentials can be used to determine anaerobic conditions.
A conclusive hydric soil determination requires long-term observations of hydrology, climate, and redox potential. Alternatively, soil morphological indicators that reflect long-term conditions may be used to make field determinations of hydric soils. Redoximorphic features are morphological indicators that are formed by processes of reduction, translocation, or oxidation of Fe and Mn oxides (USDA-NRCS, 2002). Some soils that meet hydric criteria do not develop clear soil morphological indicators. This is often the case in semi-arid regions where low precipitation, high soil pH, low organic C content, and low Fe content inhibit the development of morphological indicators.
Such conditions are exhibited by a class of seasonally ponded depressions occurring in the sagebrush steppe of the southeastern Oregon High Lava Plains. These shallow depressions serve as collection points for local surface runoff. Because subsoil permeability is very slow, sufficient water can accumulate to create perched water tables and prolonged seasonal ponding (Holland, 1976; Thorne, 1981). During the dry season these sites appear nearly barren, but during the wet season they support unique vegetation communities. Locally these depressions are called upland playas, but the term ‘playa’ is probably inappropriate, as true playas are generally devoid of vegetation (Neal, 1975; Soil Science Society of America, 1997). In their vegetation, soils, geology, and hydrology, the Oregon pools resemble the Northern Basalt Flow Vernal Pools of California (Sawyer and Keeler-Wolf, 1995, p. 361).
Our objectives in this study were to determine if hydric soils were present in the pool, and how observed field indicators were correlated with hydrologic and chemical data. Based on the limitations of current field indicators, we suggest morphological features to be examined for development of new indicators for vernal pools in this region.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
The regional climate is semi-arid, with long severe winters and short dry summers. Annual precipitation is typically 250 to 500 mm, of which 55 to 75% occurs between 1 October and 31 March. Average January temperatures are below freezing. Average July temperatures approach 20°C (Franklin and Dyrness, 1988, p. 38).
The most conspicuous pool vegetation consists of silver sagebrush (Artemisia cana Pursh) and povertyweed (Iva axillaris Pursh). A variety of small, ephemeral herbs occur during late winter and early spring. The wetland indicator statuses (Reed, 1988) of the ephemeral herbaceous pool species (fine-branch popcorn flower (Plagiobothrys leptocladus (Green) I. Johnst.), water knotweed (Polygonum watsonii (= P. polygaloides ssp. confertiflorum) Meisn.), common downingia (Downingia elegans (Dougl. ex Lindl.) Torr.), sedge mousetail (Myosurus apetalus (aristatus) Benth. ex Hook.), and whitehead navarretia (Navarretia leucocephala Benth.)) range from facultative wetland to obligate wetland (Clausnitzer, 2002).
The pool soil is mapped as Swalesilver series (fine, smectitic, frigid Aquic Palexeralfs), very deep, somewhat poorly drained soils that formed in stratified alluvium with an admixture of loess and ash, and occurring in small closed basins on lava plateaus or on lake terraces. Swalesilver typically exhibits few fine distinct yellowish red (5YR 4/6) mottles in the surface horizon (USDA-NRCS, 1999). These are redox features associated with anaerobic conditions and Fe reduction, but because their abundance is low, they do not meet the criteria of any current hydric soil field indicator (Hurt et al., 1996). The surrounding upland is 2 to 3 m higher than the pool bottom, and consists of Coztur soil series (loamy, mixed, frigid Lithic Xeric Haplargids), shallow well-drained soils that formed in residuum from volcanic and tuffaceous rocks (USDA-NRCS, 1999).
Methods
Two sites were selected for monitoring, soil description, and sampling. Site 1, in a Swalesilver sandy loam phase at the pool's lowest elevation near the sloping pool edge, supports only small herbaceous plants. Site 2, in a Swalesilver silt loam phase at the pool's highest elevation near the center, is 0.07 m higher than Site 1 and contains silver sagebrush and herbs.
A backhoe trench at each site provided morphological descriptions and samples for physical and chemical characterization at the National Soil Testing Laboratory in Lincoln, NE. Tests specific to this study were pH (2:1 CaCl2 solution/soil), electrical conductivity (saturated paste), organic C (K2Cr2O7 method), oxalate-extractable Fe, dithionite-citrate extractable Fe and Mn, and particle-size analysis (Soil Survey Laboratory Staff, 1992).
Ground water observations were made at both sites using piezometers installed in triplicate at depths of 25, 50, and 100 cm, and singly at 200 cm. Piezometers were 2.54-cm (1-inch) diam. polyvinyl chloride (PVC) pipes with slits at the bottom covered by landscaping fabric. They were installed in auger holes, backfilled in approximate depth sequence with soil from the auger hole, and sealed at the soil surface with bentonite. Ponding depth was measured with a ruler. Redox potentials were measured at least 25 cm beyond any shrub canopy or shrub mound with Pt electrodes constructed and installed according to procedures outlined by Austin and Huddleston (1999). Electrodes were installed in triplicate at depths of 25, 50, and 100 cm at both sites. Data were obtained with a calomel reference electrode, and adjusted to the standard hydrogen electrode by adding 244 mV to each reading. Soil temperature was measured with a thermocouple thermometer and single thermocouples installed at depths of 25, 50, and 100 cm at both sites. Before the third field season, additional sets of three Pt electrodes and single thermocouples were installed at a 5-cm depth at both sites. Monitoring was done at approximately 2-wk intervals during winter and spring, and less frequently in summer and fall.
Pond-water samples for chemical analysis were collected at midmorning on four dates in the spring of 1998. Two samples were collected on each date, from within 25 m of each of the two sites. Water was filtered through 0.2-µm cellulose acetate membrane filters directly into 20-mL polyethylene scintillation bottles. Samples intended for analysis of Fe and Mn were acidified below pH 2 with 50 µL of concentrated HCl to impede oxidation of reduced species and to avoid adsorption of ions on the container walls (Buffle, 1988, p. 391).
Soil solution samples were obtained at the same locations and times as pond-water samples. We dammed a small area of ponded soil using the upper half of a plastic drywall compound tub, removed surface water with a cup and sponge, and excavated saturated surface (4–5 cm) soil. These soil samples were placed immediately in plastic freezer bags. Air was expressed from the bags, which were then sealed and placed in insulated coolers for transport to the laboratory. Soil was kneaded in the bag and extruded though a small hole directly into 50-mL centrifuge bottles. Each bottle was filled to the top and capped immediately. Soil solution was extracted by centrifuging at 33 000 x g for 90 min in a refrigerated centrifuge. The supernatant was decanted into scintillation bottles, some of which were acidified as above, and passed through 0.2-µm filters.
Both pond-water and soil-solution samples were analyzed for Mn (assumed to be Mn2+) with an inductively coupled photometer using commercial standards. Iron-2+ was determined with a Lachat colorimetric analyzer (Lachat Instruments, Milwaukee, WI). Solution pH was measured with a pH meter with a combination electrode calibrated to pH 7 and 10 standards. Electrical conductivity was measured with a conductivity meter.
Nodules from the four upper horizons of the vernal pool soils were wet-sieved and inspected with a binocular microscope at 15X and 30X magnification. Examples of nodules were photographed using scanning electron microscopy. Nodules were mounted using DUCO cement (Devcon, Danvers, MA), and coated with spectroscopic carbon to a thickness of about 30Å. Analysis of relative elemental composition was done with an x-ray energy spectrometer using a light element detector.
Precipitation data (1961–1990) for Burns Airport, OR and pan evaporation data (1961–1990) for Malheur Experiment Station, OR (Oregon Climate Service, 2002) were used to estimate the season of effective precipitation for ponding. Precipitation data (1968–1998) for Burns Airport, OR were used to estimate long-term rainfall probabilities by fitting the data to a 
distribution and calculate goodness-of-fit with a Chi-squared test using Statgraphics 4.0 (1999) software.
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
|
|
Hydrology
At no time did any of the piezometers contain water, indicating that the subsoils were never saturated. The hydrology appears to be controlled by direct precipitation, with a small contribution from runoff from surrounding uplands, interacting with evapotranspiration. Water entering the pond becomes perched above a shallow Bt horizon (Table 1). Many California vernal pools exhibit this type of hydrology (Hanes et al., 1990; Hanes and Stromberg, 1998; Keeley and Zedler, 1998).
|
Ponding coincided with soil temperatures exceeding 5°C for 7 wk in 1996, 3 wk in 1997, and 8 wk in 1998 (Fig. 1 and 2). The hydric soil criteria stipulate that this coincidence must occur with a probability exceeding 50% in any year over the long term. This probability is normally estimated by fitting long-term precipitation data to a distribution and computing critical percentile values. Analyses of data from many weather stations can be obtained from the NWCC (1999). As none of the nearby weather stations available in the NWCC analyses had complete data coinciding with our field observations, we analyzed precipitation data for 1968 through 1998 from the Burns Airport weather station.
Typically, vernal pools fill from precipitation during periods when the rate of water input exceeds the rate of water loss, primarily from evapotranspiration (Zedler, 1987). In our pool's region, pan evaporation greatly exceeds precipitation for much of the year (Fig. 3) . Evaporation data are not available for the winter months, when average monthly temperatures are near or below freezing. We consider that effective precipitation for ponding begins in November. Because observed soil temperatures at a 50-cm depth usually exceeded 5°C beginning by the middle of March, winter precipitation must be adequate through March to maintain ponding for at least 7 d coinciding with the defined growing season. Crowe et al. (1994) considered November through March to be the season of effective precipitation for filling vernal pools in eastern Washington, where the climate is similar to that in eastern Oregon.
|
distribution (p = 0.023, Chi-squared goodness-of-fit test), the bimodally distributed November through March data were not (p = 0.023, Chi-squared goodness-of-fit test) (Fig. 4) , and so were unsuitable for determining percentile probabilities. Therefore, an alternative way of looking at the precipitation data was necessary. The 137.5 mm of precipitation in the 1997/1998 wet season, the driest we observed, supported ponding from mid December until 16 April. This precipitation amount was equaled or exceeded 15 yr out of the 30 for which we had data. As the overlap of growing season and ponding was 8 wk long that year, it appears likely that these conditions would occur with a probability exceeding 50% over the long term.
|
|
Solution Chemistry
Conductivity of surface water in the pool increased from 0.05 to 0.11 dS m-1 between March and June 1998 (Table 3), indicating some seasonal concentration of solutes through evaporation. These values overlap the conductivity range of 0.04 to 0.07 dS m-1 typically found in many California vernal pools (Keeley and Zedler, 1998).
|
Small concentrations of reduced Fe and Mn were present in solution during the wet season (Table 3). These values are in the low range in the context of reduced soils (Meek et al., 1968), and are comparable with values measured by Vepraskas and Wilding (1983) in a periodically saturated soil with low organic C contents.
Redox Potentials
A review of several sources (Ponnamperuma, 1972; Gambrell and Patrick, 1978; Bohn et al., 1985, p. 254; Sposito, 1989, p. 108) suggested ranges of empirical EH values for significant reduction of various soil components. Based on these sources, at pH 7, oxygen should be substantially depleted when potentials drop to approximately +350 mV. Manganese and Fe should undergo significant solubilization near +200 and +150 mV, respectively.
Redox potentials at 25-, 50-, and 100-cm depths, where saturation was never detected, remained fairly static in the anaerobic range at both sites, fluctuating between 250 and 325 mV during the wet season (Fig. 1 and 2). Potentials appeared to drop slightly during periods of ponding and rise slightly after ponding disappeared, and were unresponsive to changes in soil temperature while ponded. Redox potentials never dropped below the empirical level at which substantial reduction of Fe and Mn would be expected.
The redox potential levels at 5-cm depth were much more variable than levels in deeper horizons (Fig. 1 and 2). Before the onset of ponding, redox potentials were nearly identical at 5 and 25 cm at both sites. After ponding, while EH at 25 cm remained between 275 and 300 mV, EH at 5 cm actually increased for a few weeks. We hypothesize that this increase was because of dissolved oxygen increases in the pond water caused by infusions of rainwater and by photosynthetic oxygen production. Active photosynthesis during the day can cause water to become supersaturated with respect to oxygen (Keeley and Morton, 1982), which may inhibit the reducing conditions that tend to develop in submerged soils (Zedler, 1987).
A subsequent drop in potentials at 5 cm coincided with rising soil temperatures that would encourage biological activity. The lowest redox potentials measured in the pool occurred at Site 2, which had higher organic C contents than Site 1, between late February and mid March. Potentials never dropped as low as those recorded by Hobson and Dahlgren (1998) at 5-cm depth in a vernal pool in central California that has higher temperatures, higher precipitation, and higher soil organic matter contents.
Redox potentials at 5 cm rose again near the end of March, despite increasing temperatures and persistent ponding that would have been expected to induce lower potentials or at least maintain them at a low level. We again attribute this to high daytime oxygen contents in the very shallow surface water. Redox potentials rose sharply in late April as surface ponding disappeared, and dropped during a brief restoration of wet conditions in May.
Summary of Hydric Conditions
It appears likely that the pool is ponded during the observed growing season for at least 7 d in most years. Redox potentials in the surface horizons were low enough to reduce oxygen. Redox levels never dropped low enough to induce substantial Fe solubilization. This is confirmed by the low (<1 mg L-1) Fe2+ concentrations we detected in interstitial solutions. There was sufficient overlap among spring ponding, soil temperatures above biological zero, and redox potentials well within the anaerobic range to satisfy the criteria set forth in the hydric soils definition. Formation of abundant or conspicuous hydric soil morphological features would be inhibited because of low solubilization of Fe and Mn.
Redoximorphic Features
Redox concentrations are bodies of apparent accumulation of Fe and Mn oxides that commonly appear in the soil as soft masses or pore linings (Hurt et al., 1996). The A1 and A2 horizons at Site 2 contained common fine distinct olive brown (2.5Y 4/4) soft masses (Table 1). The color of these masses makes it doubtful that they are redoximorphic concentrations, because they are much more yellow than would be expected for secondary ferrihydrite (Vepraskas, 1994). The A2 had few fine distinct 10YR 4/6 soft masses. Abundance of these soft masses increased to common in the upper 3 cm of the Bt1 horizon, and decreased to few in the remainder of the Bt1. Their 10YR hue suggests that they are redoximorphic goethite concentrations (Vepraskas, 1994). Site 1 exhibited few fine faint 10 YR 4/4 soft masses in the A1 and A2 horizons.
A depleted matrix is a volume of soil that exhibits low chroma and high value colors because of removal or transformation of Fe by reduction and translocation (Hurt et al., 1996). Munsell color of the A1 horizons of both sites (Table 1) exhibited sufficiently low chromas and high values to suggest the presence of depleted matrices. Because they are A horizons, however, other morphological features are required for classification as a depleted matrix. This is discussed in the next section.
The redoximorphic features described above reflect the hydrology, EH, solution chemistry, soil chemistry, and soil morphology data for the pool. Episaturation creates mildly reducing conditions and low concentrations of dissolved Fe near the soil surface. Water perches because of increasing amounts of clay at the A2–Bt boundary. A combination of the pH increase (Collins and Buol, 1970; Weitkamp et al., 1996), limited hydraulic conductivity in the clay, and contact of the surface interstitial solution with underlying unsaturated soil would favor Fe precipitation at or near this boundary. This situation is reflected in changes in Fed content and Feo/Fed ratios. One would expect the pool's chemical and textural conditions to produce scarce or inconspicuous redoximorphic features that indicate low levels of Fe solubilization near the soil surface and Fe reprecipitation at the perching layer.
Nodules and concretions are redox concentrations that appear as firm to extremely firm masses of Fe and Mn. As they may be relict or transported features, they are not considered to be reliable hydric soil indicators in most cases (Vepraskas, 1994; Hurt et al., 1996). Pedogenic Mn nodules have been reported in California vernal pool soils (Weitkamp et al., 1996), and are a feature of hydric soils in ponded High Plains Depressions (Indicator F16) (Hurt et al., 1996). Most of the horizons in our soils were described in the field as having few very fine soft nodules and concretions. High-Fe basaltic glass is common in this region (Walker, 1979), and rounded black nodules are known to weather out of the lower part of the volcanic duff parent materials in the Palomino Butte area (Johnson, 1992). Specimens examined under a binocular microscope could be consistently grouped into smooth and rough types. We examined one specimen of each type by scanning electron microscopy (SEM). The smooth nodule (Fig. 5) had a glassy, nearly featureless surface (Fig. 6) and high x-ray count peaks (data not shown) for both silica and Fe, suggesting that it is primary high-Fe basaltic glass. The rough nodule (Fig. 7) appeared to be a clastic particle containing feldspar crystals (Fig. 8) . Scanning electron microscopy analysis showed it to be very siliceous, with very low Fe and Mn peaks (data not shown), suggesting primary geologic origin rather than formation by cementation of fine particles by Fe and Mn.
|
|
|
|
The sandy loam soil (Site 1) fails to meet indicators F3, F9, and F8. Although the 5/1 matrix color of the A1 horizon (Table 1) suggests Fe depletion, insufficiently abundant and conspicuous redox concentrations disqualify the horizon as having a depleted matrix, and requirements for Indicators F3 and F9 are not met. Indicator F8 does not apply because none of the horizons in the upper 15 cm have at least 5% redox concentrations.
Both the A2 and the Bt1 horizons at Site 1 have matrix chromas higher than 2, disqualifying them from Indicators F3 and F9. Redox concentrations were insufficiently abundant, eliminating Indicator F8. Thus the sandy loam soil, although showing evidence of Fe depletion from surface horizons, does not meet the test of any of the existing indicators.
The silt loam soil (Site 2) also fails all three indicators. The A1 and A2 horizons together are entirely within the upper 15 cm (Table 1), and both have matrix colors of 4/2. To qualify as a depleted matrix, however, this combination of value and chroma requires common or many distinct or prominent redox concentrations as soft masses or pore linings. Although common distinct soft masses were present, their color casts doubt on whether they are true redox concentrations. Neither horizon meets the surficial depleted matrix criteria for F3 and F9, and neither possesses the redox concentrations required for F8.
The Site 2 Bt1 horizon also has a 4/2 matrix and distinct redox concentrations. However, they were most abundant in upper 3 cm of the horizon, which was not separated from the remainder of the Bt1 horizon during the initial morphological description in the field. The overall abundance of redox concentrations in the remainder of the Bt1 horizon as described would average <2%, too low to classify this soil as hydric based on Indicator F3.
We conclude that the evidence in support of positive confirmation of the soil at Site 2 as a hydric soil using existing indicators is impossible. However, observed morphological features do suggest that surficial horizons are depleted of Fe, and that Fe accumulates in the upper part of the perching layer.
|
CONCLUSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
One way to recognize the hydromorphic processes at work in these soils would be to relax the requirement that a depleted matrix A horizon also must have common distinct redox concentrations as soft masses or pore linings. Because these soils are also very low in organic C, waiving the requirement completely could lead to erroneous conclusions, but simply requiring evidence of Fe reduction and reprecipitation in the form of a few, faint soft masses or pore linings with hues of 10YR or redder would be sufficient. Such a criterion would allow identification of the sandy loam soil as a hydric soil, although it would not work for the silt loam soil.
Another way to generate an indicator is to recognize the nature of the hydrological control of saturation and reduction in these vernal pools. Perching of water above very slowly permeable clayey horizons, with unsaturated soil beneath thin saturated surface horizons, creates an environment in which Fe and Mn mobilized in the saturated surface soil can move down to the contact with unsaturated soil, where it oxidizes again to form redox concentrations. Very limited movement into the clayey layer, however, means that the zone in which these concentrations occur is necessarily very thin. Identification of such a thin zone containing redox concentrations at the perching layer near the surface would make it possible to confirm the silt loam soil in the center of the pool as a hydric soil.
The Palomino Butte pool can be defined as a wetland in that it displays hydric soils, wetland hydrology, and a distinctive wetland plant community (Clausnitzer, 2002). As the pool is just one example of many such pools in the region, verification of reliable hydric soils indicators would be useful for surveying their abundance and diversity. Development of new indicators would require a survey and categorization of eastern Oregon vernal pools, followed by examination of more vernal pool pedons.
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
Received for publication November 26, 2001.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. T. O'Geen, W. A. Hobson, R. A. Dahlgren, and D. B. Kelley Evaluation of Soil Properties and Hydric Soil Indicators for Vernal Pool Catenas in California Soil Sci. Soc. Am. J., May 1, 2008; 72(3): 727 - 740. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
of the two vernal pool soil phases.
