| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Forestry (0324), Virginia Polytechnic Institute and State University, Blacksburg, VA 24060 USA
b USDA Forest Service, Pacific Southwest Research Station, 2400 Washington Ave, Redding, CA 96001
c USDA Forest Service, North Central Research Station, Lincoln University, Foster Hall, Room 208, Jefferson City, MO 65102
d MeadWestvaco Corp., Forest Science Lab, Box WV, 180 Westvaco Rd. Summerville, SC 29483
* Corresponding author (csiegel{at}vt.edu)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODS AND MATERIALSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
b) and volumetric water content (
v) to better understand soil compaction effects on seedling growth of: (i) ponderosa pine (Pinus ponderosa var. ponderosa Dougl. ex Laws) grown on Dome and Cohasset soils; (ii) shortleaf pine (Pinus echinata Mill.) on a Clarksville soil; and (iii) loblolly pine (Pinus taeda L.) on an Argent soil. Models of root length density (RLD) were developed using multiple regression. The general model of RLD = b0 + b1v + b2 b + b32v described rooting response for the Clarksville-shortleaf and Argent-loblolly soil–species combinations (p = 0.005). However, the ponderosa pine RLD response on Cohasset soil was linear and there was an interaction between v and b in the Dome soil. Shoot mass of seedlings growing within the least limiting water range (LLWR) was greater than those growing outside the range for all soil–species combinations except the Argent-loblolly pine (p = 0.05). The loblolly pines had greater shoot mass at v above the upper LLWR limits (aeration limiting). Least limiting water range has potential as a soil quality indicator, but seedling response was not always associated with LLWR. Root length density (RLD) response surface models in conjunction with seasonal site water data have potential for determining compaction-induced soil limitations for tree growth, but need to be field tested and calibrated for both soil and species.
Abbreviations: LLWR, least limiting water range • LTSP, USDA Forest Service Long-Term Soil Productivity Study • NLWR, nonlimiting water range • OWC, optimum soil water content • RLD, root length density • SWRC, soil water retention curves • b, bulk density • v, volumetric water content
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODS AND MATERIALSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Although reduced growth from compaction has been reported across many regions, the effect on individual tree growth parameters is variable, and overall growth response varies for different species and soil types. For example, shoot and root weight of Pseudotsuga menziesii var. glauca (Beissn.) Franco and Pinus monticola Dougl. ex D. Don seedlings were not affected by compaction after one growing season, but root volume was 41% less for the P. menziesii (Mirb.) Franco seedlings and seedling height was 6% greater for P. monticola (p = 0.05) (Page-Dumerose et al., 1998). Corns (1988) reported that P. contorta root weight, shoot weight, stem diameter, and stem height declined due to compaction on all four soils tested, but Picea glauca growth on two of the soils did not decline or increased twofold. P. contorta shoot weight decreased 64% on a silty clay soil when b increased from 1.2 to 1.5 Mg m–3, while shoot weight decreased 86% on a clay loam soil compacted to 1.5 Mg m–3. Wasterlund (1985) also reported species differences with Picea abies growth being more impeded by compaction than Pinus sylvestris L. growth. On several California sites, Gomez et al. (2002) reported that compaction effects on 4-yr-old ponderosa pines varied with soil texture and soil water regime. Stem volume on compacted soils was less, the same, and higher on clayey, loamy, and sandy loam soils, respectively. Compaction also favored Picea mariana (Mill.) B. S. P. and Pinus banksiana Lamb. growth on coarse textured soils classified as humo-feric podzols in northwestern Quebec (Brais, 2001). Growth increases on these soils were linked to harvest traffic compaction causing a more favorable pore-size distribution, which improved the balance between aeration porosity and available water holding capacity, similar to the findings by Gomez et al. (2002).
The persistence of compaction effects from forest harvest operations on tree growth and soil properties varies over time for different sites and species. Two studies with similar experimental designs in coastal Washington and inland Oregon report these results (Miller et al., 1996; Heninger et al., 2002). Bulk density increases on skid trails ranged from 2 to 40% for different soil types and persisted 8 yr after the logging operation (Miller et al., 1996). However, the effect of compacted skid trail soil on tree height lasted only one to two seasons for P. menziesii, but persisted more than 2 yr for Tsuga heterophylla (Raf.) Sarg. in coastal Washington (Miller et al., 1996). In comparison, Heninger et al. (2002) reported that P. menziesii growing on skid trails had height growth reductions that persisted for 8 to 10 yr on inland Oregon sites where soil textures were finer and the climate was drier. After 10 yr, trees growing on skid trails were 10, 14, and 29% less in height, diameter, and volume, respectively.
Several researchers have developed models to elucidate the roles that key soil properties such as soil strength, water, and aeration and their interactions have on plant growth. Greacen and Sands (1980) developed a conceptual model, which shows that compaction increases soil bulk density, which modifies both soil strength and porosity. These factors are further moderated by water content, and their combined interactions influence root growth. The concept of the non-limiting water range (NLWR), introduced by Letey (1985), combined the effects of several soil properties critical to plant growth into a single variable. The NLWR was defined as the range in which water availability is non-limiting to plants, generally bounded by field capacity and wilting point. As bulk density increases, the NLWR becomes narrower, with mechanical resistance becoming limiting at the dry end and poor aeration becoming limiting at the wet end.
Childs et al. (1989) used soil density and porosity data from a compaction study by Reicosky et al. (1981) to develop a generalized model relating soil physical conditions to root growth similar to Letey's NLWR. They also hypothesized that ideal root growth conditions were diminished with increasing soil density due to excessive soil strength at low water contents or inadequate aeration under wet soil conditions. In their model, ideal growth is depicted within a "root growth window" bound by non-specified water contents.
Da Silva et al. (1994) furthered these conceptual ideas by evaluating the NLWR as an index of the structural quality of soil. They used the term LLWR to recognize that plant response occurs along a continuum of water contents rather than as a step function. The critical limits defining the LLWR were: (i) volumetric water content (v) at field capacity and permanent wilting point (potentials of –0.01 and –1.5 MPa, respectively); (ii) air-filled porosity <10%; and (iii) soil strength >2.0 MPa (da Silva et al., 1994). All of these conceptual models attempt to integrate various soil property effects that, alone, do not fully account for root growth in a given environment. In a management context, one would want to maintain or improve those soil conditions that created the largest NLWR, LLWR, or root growth window. However, we need to determine if generalized models adequately reflect growth potential or if soil- and species-specific models need to be developed.
For a given region, soil type, and tree species, forest productivity is a function of v as it varies with climate across the growing season, and a function of the interrelated factors of b, soil strength, and porosity. Root growth has been found to be a more sensitive indicator of soil disturbance than shoot growth (Singer, 1981; Heilman, 1981). Additionally, reductions in root growth occur long before extreme soil strength or moisture conditions are reached (Eavis, 1972; Voorhees et al., 1975; Russell, 1977; Simmons and Pope, 1987). Developing soil- and species-specific root growth responses for a range of soil water, aeration, and b conditions would be valuable for assessing potential productivity declines due to compaction.
The USDA Forest Service Long-Term Soil Productivity Study (LTSP), composed of large-scale field experiments located at sites across the USA, was developed to assess the effects of soil compaction and surface organic matter removal on site productivity across a range of forest sites (Powers et al., 1990). Similar projects on industry lands have also been developed. To better understand the management implications of compaction, we used soils and associated tree species from a spectrum of LTSP study sites to test the hypothesis that best growth would occur at low b and moderate v, while as density increased, aeration would become limiting to growth on wetter soils, and soil strength would become limiting for dryer soils.
Our specific research objectives were to: (i) develop a response surface describing tree seedling root growth as a function of soil b and v; (ii) examine seedling growth using the LLWR; and (iii) based on our response surface models, determine if generalized models adequately reflect growth potential or if soil- and species-specific models are needed.
|
METHODS AND MATERIALS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODS AND MATERIALSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Experimental Design
A seven by seven factorial greenhouse experiment was performed to assess root growth as a function of b and v. A series of soil compaction tests determined the optimum technique for uniformly compacting soils in PVC cylinders that were used to assess compaction effects on various soil physical properties (Siegel-Issem, 2002). Polyvinyl chloride (PVC) cylinders with dimensions of 8 x 15 cm were packed at seven compaction levels with surface soil from each LTSP site. Compaction levels were assigned based on the range between the minimum and maximum b determined for each soil (Table 2). A gradient of seven levels of v was established to cover the range from permanent wilting point to near-saturation for each soil (Table 2). Tree seedlings of species typically associated with each soil type were planted in the soil columns compacted at each of the seven b levels. Water contents at the seven different v levels were maintained as closely as possible throughout the growing period. Approximately every 3 d during the growth period, all pots were weighed and watered as necessary to maintain the v within a range of 10 to 15% of target v. If weight was below target, water was added to achieve the target v; conversely, water was not added if the pot was too wet or within the range.
|
b (Siegel-Issem, 2002). Subsequently, all soils were compacted at their OWC.
Compactive effort for each soil was determined as a variation of the ASTM compaction standard effort tests to assess differences in each soil's b range (ASTM, 1996). This procedure was used to determine the compactive effort (number of hammer blows) needed to achieve target b for the compacted soil columns used in subsequent analyses. Sieved and moistened soil was added to the PVC cylinder, the surface settled and smoothed, and then compacted in several lifts. Each lift of soil received a set number of blows (1, 2, 4, 8, 16, 32, or 64 blows) to relate a range of compaction hammer blows and b, similar to the work done by Howard et al. (1981). Soil volume, mass, and v were measured and oven dry weight and b determined for each soil column. Minimum and maximum b for each soil were determined from graphs depicting b as a function of compactive effort, with maximum b being defined as the asymptote of the curve. Regression analyses on log transformed data were used to determine the relationship between number of hammer blows and b, and thus the number of compaction hammer blows needed to achieve each target b for the soil columns used in our experiment (Siegel-Issem, 2002).
Soil Strength
Soil strength was measured in each soil column at the end of the experiment with a lab pocket penetrometer (BSE Model S-170, Durham Geo-Enterprises, Stone Mountain, GA). All soil columns were near their targeted v when measured. The column was placed on its side and the outer PVC cylinder cut lengthwise in several places. The PVC segments were removed and triplicate soil strength measures taken. The flat-tipped pocket penetrometer was fitted with a smaller-than-standard tip to measure the high strengths of some soils. Volumetric water content was determined for each soil column at the time of measurement by measuring gravimetric water content and adjusting for soil bulk density.
Soil Porosity and Air–Water Balance
Soil water retention curves (SWRC) were developed for all compaction levels within each soil type. Soil v at soil water potentials (
w) of –0.005, –0.01, –0.03, –0.1, and –1.5 MPa were determined for each soil using standard tension table and plate techniques (Klute, 1986). Compacted soil columns (7.7 by 10 cm) were used to determine w for tensions up to –0.1 MPa, and 5 by 2.5 cm soil columns were used to determine v at a w of –1.5 MPa. Soil porosity at several key w ranges for all four soils along their b gradient was calculated from the SWRC: aeration porosity (w between 0.00 to –0.01 MPa), available water (w between –0.01 and –1.5 MPa), and unavailable water (permanent wilting point) (w less than –1.5 MPa).
Least Limiting Water Range
The LLWR, as used by da Silva and Kay (1997), was developed for each soil using our experimental data. The upper LLWR limit is the lesser v of field capacity (FC) or aeration porosity <10% (AP), while the lower limit is the greater water content associated with either wilting point (WP) or soil strength >2.0 MPa (SS). Soil water retention curve data from each of the seven density levels were used to determine FC and WP critical limits. AP was defined as total porosity minus 10%. The SS limit was determined using Busscher's (1990) regression model as selected by da Silva and Kay (1997) describing the relationship of strength as a function of bulk density and water content:
 | [1] |
Seedling Establishment and Growth
A 1-cm diam. hole was drilled in the center of each packed soil column to within 3.5 cm of the bottom. At planting time, this hole was back-filled with washed silica construction sand. The sand channel allowed rooting during seedling establishment, which was important in the highly compacted soil columns. Furthermore, the sand channel allowed water access to the soil column center along its depth, resulting in a more uniform v, and minimizing high surface density impact on water infiltration. The volume of this channel is 1% of the total soil column volume and, therefore, a small fraction of the volume roots would eventually utilize. A fine mesh plastic screen was attached to the bottom of each PVC cylinder to prevent the loss of soil and allow water drainage. Soil columns were placed on a metal mesh greenhouse bench throughout the experiment.
Ponderosa pine was grown on the Dome and Cohasset soils from California, shortleaf pine on the Clarksville soil, and loblolly pine on the Argent soil. Seed stock appropriate to the areas from which our soils were collected was used. Seeds were planted in trays in a potting soil and sand mixture and were set in the greenhouse to germinate and grow. After 28 d, the most vigorous seedlings of approximately equal size were selected and planted in the center of each pot. Approximately 0.5 cm of washed silica sand was added to the top of the soil to prevent soil surface disturbance from the watering treatments and to prevent the sand-planting channel from clogging with soil. The seedlings were grown for 6 wk with regular watering and nutrition to establish root growth before applying water stress.
After the establishment period, seedlings were grown for the experimental period of approximately 13 wk. We allowed the shortleaf pines on Clarksville soil to grow an additional 8 wk because these seedlings were still very small and roots had not fully exploited the soil volume after 13 wk. The average weekly minimum temperature during the study period was 18°C and the average maximum temperature was 33°C. Humidity ranged from 50 to 93% with a weekly average of 56%. A commercial fertilizer (15-30-15) nutrient solution was foliar-applied periodically to provide adequate nutrition throughout the experiment. At each fertilizer application, each seedling received 4.5 mL of fertilizer solution containing 710 mg L–1 N, 610 mg L–1 P, 590 mg L–1 K, 7 mg L–1 Fe, 3 mg L–1 Cu, 3 mg L–1 Zn, 2 mg L–1 Mn, 0.9 mg L–1 B, and 0.02 mg L–1 Mo.
All seedlings survived the establishment phase. However, after watering treatments were applied, there was significant ponderosa pine seedling mortality on the Dome and Cohasset soils. Several shortleaf pines on the Clarksville soil died. No loblolly pines growing on the Argent soil died.
Plant Analyses
After the growing period, seedling height, and root collar diameter were measured. Each core was then deconstructed and root systems separated from the soil by washing with water. Root length was determined for each seedling's entire root system using a computer imaging analyzer (Delta T, Delta T Devices, LTD, Cambridge, UK) and RLD (length of roots per volume of soil) was determined. Shoots and roots were oven dried at 70°C and both above and belowground biomass were measured.
Model Development and Statistics
Multiple regression techniques were used to model root growth as a function of b and v. We hypothesized that root growth would increase linearly with decreasing b (Foil and Ralston, 1967; Heilman, 1981; Mitchell et al., 1982). Furthermore, we hypothesized that root growth would be less both at the wet and dry ends of the soil water spectrum, with optimum growth occurring at moderate v. Therefore, we hypothesized that this relationship could be depicted mathematically as a quadratic function. The basic model
 | [2] |
b and v. Terms were then added or deleted, based on their significance in the model, to reflect the observed data for each soil. Regression diagnostics (Cook's D and leverage analysis) were used to examine the influence of outliers on the model shape. Plots of the residuals were evaluated to assess model fit. Seedling growth in and out of the LLWR was compared with a t test. All statistical analyses were performed using the SAS statistical software (SAS Institute, 1999). |
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODS AND MATERIALSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
v below 0.25 cm3 cm–3, soil strength became excessive at higher b, often exceeding 2.0 MPa, for all soils except the Argent soil (Fig. 1)
. However, the b at which soil strength increased significantly or exceeded 2.0 MPa was soil specific. The most compact and dry soil columns exceeded this limit at b above 1.13, 1.33, and 1.4 Mg m–3 for the Cohasset, Dome, and Clarksville soils. However, no Argent soil columns exceeded 2.0 MPa, but they did have the potential to exceed this value when b was above 1.55 Mg m–3 at v dryer than we measured. At the lowest densities for all soils, strength was generally <0.05 MPa and not affected by v.
|
b and reduced total and aeration porosity for all four soils (Fig. 2)
. The greatest total and aeration porosity reductions occurred for the Cohasset soil, which initially had the highest overall total porosity, available water, and aeration porosity of the four soils. At or above b of 1.13, 1.42, 1.44, and 1.55 Mg m–3 aeration porosity dropped below 10% for the Cohasset, Dome, Clarksville, and Argent, respectively (Fig. 2). Water contents above 0.35 cm3 cm–3, in combination with high b, created poorly aerated conditions, which limited root growth of shortleaf pines in Clarksville soil, and loblolly pines in Argent soil. For all soils, available water increased slightly with increasing compaction.
|
|
|
The LLWR did not define loblolly pine growth on the Argent soil as predicted (Table 3; Fig. 3). The shoot weight of loblolly pine seedlings within the LLWR was less than those outside the range despite the fact that they grew at <10% aeration porosity. Mean shoot weight of seedlings growing within the LLWR was 0.44 g while it was 0.72 g for seedlings growing out of the LLWR. The LLWR underestimates the ability of loblolly pine to grow across a wide range of soil moisture conditions in the Argent soil. The LLWR did not correspond to the "best" growth of these seedlings; the v contents associated with the standard limits are too low for the Argent soil growing loblolly pines.
Root Length Density Models
Root length density of shortleaf pine growing in the Clarksville soil and loblolly pine growing in the Argent soil responded to the soil water and b gradients as predicted (Fig. 4A,B)
. The RLD residuals plotted as a function of the predicted values for each soil were well distributed, indicating that there was no reason to believe that other terms would improve the model fit. Additional regression diagnostics determined that no individual points were outliers or had undue influence on the model shape. Together, v and b had a significant effect on root growth, explaining 33 and 61% of the variation in RLD for the Clarksville and Argent soils, respectively. The influences of b and v on RLD were independent of each other. Root length density decreased linearly with increasing b and decreased as v became wetter or dryer than 0.25 and 0.30 cm3 cm–3 for the Clarksville and Argent soils, respectively. Bulk density had a greater influence on shortleaf pine growing in Clarksville soil than loblolly pine in Argent soil. At higher b, the v range in which growth occurred was narrower for the Clarksville-shortleaf pine than the Argent-loblolly pine; at a b of 1.6 Mg m–3, shortleaf pine roots grew within a v range of 0.13 to 0.43 cm3 cm–3, while loblolly pine grew between 0.15 and 0.58 cm3 cm–3. Best growth occurred when v was near 0.25 cm3 cm–3 for the Clarksville-shortleaf and between 0.30 and 0.35 cm3 cm–3 for the Argent soil (Fig. 4A,B).
|
v and b; therefore, the interaction term v x b was added to the general model. With the expanded model, b and v explained 81% of the variation in RLD for the Dome soil. Root length density decreased with increasing b; however, that effect was moderated by v. The significant interaction of v and b had the effect of lowering the v at which best growth occurred as b increased, while decreasing growth to a greater extent on the wet end of the water gradient. Predicted best RLD of ponderosa pine on the Dome occurred in the v range of 0.25 to 0.35 cm3 cm–3.
In contrast, the 2v term was not significant for the Cohasset-ponderosa pine; the RLD response surface was planar (Fig. 4D). Bulk density and v had a significant effect on RLD and explained 77% of the variation in RLD. As b increased, RLD decreased. However, increasing v improved growth along the b gradient.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODS AND MATERIALSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
v. The best growth occurred across a broader range of v when the b was low. As density increased, v at either the dry or wet end of the spectrum interacted with b to create either high soil strength or poorly aerated conditions, thereby diminishing the range in which normal growth occurred. The general regression model describing RLD as a linear function of b and quadratic function of v was significant for two of the four soil–species combination and explained much of the variation in RLD. It is clear that root growth response is soil and species specific.
Although we cannot differentiate the exact causes of root growth limitations, it appears that soil strength and poor aeration, and combinations thereof are the primary causes of growth limitations at high b. Soil strength in excess of 2.0 MPa can significantly limit growth (Atwell, 1993; Greacen and Sands, 1980). An aeration porosity of 10% is often considered a critical limit for growth (Grable and Siemer, 1968). Eavis (1972) attempted to separate the effects of soil aeration, soil strength, and moisture stress on pea seedling growth and found that, generally, soil strength affected root growth in the w range of –0.01 to –0.1 MPa, and water stress was the main factor at w greater than –0.35 MPa. Voorhees et al. (1975) found that between w of –0.01 to –0.1 MPa, pea seedling root elongation was more sensitive to aeration when soil strength was low and that RLD increased with increasing strength. Our data generally agree with these findings. At the dry end of the water spectrum, when w was between –0.01 and –1.5 MPa, all soils except Argent, had soil strengths >2.0 MPa. At low b, inadequate water and poor aeration were the most likely cause of growth limitations. Furthermore, poorly aerated soils can cause physiological imbalances that lead to nutrient deficiencies. Although we fertilized our seedlings throughout the experiment, many seedlings grown at the highest water/highest densities were chlorotic, suggesting nutrient deficiencies. Nitrogen and other minerals were deficient in shoots of Pinus contorta growing in compacted, remolded soil cores (Conlin and ven den Driessche, 1996).
The four soils used in this study were formed from various parent materials and had different organic matter contents. Three of the soils had sandy loam textures (Dome, Cohasset, and Argent), yet the combination of various soil physical properties caused each to respond differently to compaction. For example, at similar water contents, soil strength values were as high as 3.5 MPa for the Cohasset soil at a b of 1.21 Mg m–3, while the Argent soil never exceeded 2.0 MPa even at b as high as 1.61 Mg m–3. These soil differences created water and air dynamics variations which subsequently affected seedling growth response.
The LLWR is being used as an indicator to assess soil physical quality for a range of agricultural and forest soils (da Silva and Kay, 1996; Tormena et al., 1999; Betz et al., 1998; Zou et al., 2000). It can also be used to determine the amount of time that seasonal soil water conditions are ideal for growth. Da Silva and Kay (1996) found a strong correlation between corn shoot growth and the percentage of time v fell outside the LLWR. Kelting (1999) determined the percentage of time that predicted daily v were within the LLWR for a southeastern loblolly plantation but did not relate that directly to plant growth responses. We found significant differences for several growth responses of tree seedlings growing within the LLWR and those growing outside the range, but the results varied with parameter measured and species. Based on our results, the LLWR was not applicable for loblolly pine on Argent soil without modification. Nonetheless, the LLWR has good potential for evaluating soil quality, and in conjunction with species–specific growth models, may help predict potential productivity declines due to forest management impacts.
Compaction and low and high v explained the least RLD variation for the Clarksville-shortleaf pine, compared with the other soil–species combinations. Root and shoot growth variability was high. Although there were up to 33% decreases in growth due to compaction they were not statistically significant. Shortleaf pine is a species that is found across a broad range of sites due to its tolerance for a wide range in soil conditions; however it does best on soils with silt loam and fine sandy loam textures (Lawson, 1992). Our soil, also a silt loam, had a wide LLWR allowing for less limited growth of this adaptable species across a wider range of water contents.
The Argent-loblolly pine combination appears to be the least affected by compaction and poor aeration of the four soil-species combinations we tested due to a combination of Argent soil properties and loblolly pine species adaptations. Increasing soil density decreased growth; however, the v had much less influence on loblolly pine RLD. The Argent soil, a fine sandy loam, had relatively low soil strengths, even at high b. We attribute this to the nature of the rounded, fine sand particles we observed and the clay mineralogy causing low shear strength. Low friction of these rounded, uniform particles, combined with the clay fraction's ability to hold water, probably allowed roots to move more easily through the soil. Furthermore, loblolly pines are adapted to poorly aerated soils and can tolerate occasional flooding with root anatomy changes that allow O2 to diffuse from the stem to the roots (Schultz, 1997). These changes include development of aerenchyma cells and intercellular spaces and formation of lenticels around the root collar (McKevlin et al., 1987; Topa and McCleod, 1986).
Ponderosa pine growth decreased with increasing b on the Dome soil and was affected by inadequate aeration porosity when b were above 1.42 Mg m–3 and v were above 0.30 cm3 cm–3. The seedlings in these pots were much smaller and were chlorotic for most of the growth period. This soil and species are from a Mediterranean climate with little rainfall and rapidly draining soil; therefore, inadequate soil aeration would seldom be a problem. The very dry conditions normally encountered, and subsequent increases in soil strength, could be detrimental to growth. Gomez et al. (2002) found enhanced ponderosa pine volume growth due to compaction on a similar sandy loam soil. On their site, b increased from 1.13 to 1.33 Mg m–3 in the top 30 cm and the resulting porosity change effectively increased available water by up to 10% on this typically droughty site. The non-compacted and compacted densities they found are comparable to the densities we created in our soil columns; however, we did not find the same growth increases with compaction. We attempted to maintain our soils at consistent v and so the benefit of increased available water holding capacity was not evident.
Inadequate water and high soil strength appeared to be the prime factors causing poor ponderosa pine growth and high seedling mortality on the Cohasset soil. Fifty percent of the trees that died were from the two lowest water levels. Even though aeration porosity was less than 10% when b exceeded 1.13 Mg m–3, infiltration and drainage were fairly rapid for this soil throughout the density range; therefore, we were not able to maintain near-saturated conditions over time that might have led to limiting aeration. Of the seedlings growing under the wettest conditions, we did not observe any hypoxic characteristics such as the chlorosis noted on the Dome-ponderosa pine seedlings. Aeration limitations due to low macroporosity from compaction may not occur for soils that are rarely saturated (Aust et al., 1998).
A discontinuity in RLD of ponderosa pines growing on Cohasset soil that corresponded with a sharp increase in soil strength was evident above a b of 1.0 Mg m–3. Root growth decreased dramatically at the higher b. This is interesting given that we would generally consider this to be a low or even ideal b. However, for this soil, this density was very compacted. Forest soils such as Cohasset, with high organic matter contents, high porosities and andic properties, may be very compact even at low densities (Howard et al., 1981; Gomez et al., 2002).
In contrast, Gomez et al. (2002) found no stem volume differences for 5-yr-old ponderosa pines growing on compacted Cohasset soil from the same LTSP site from which we collected our loose Cohasset soil. The compacted field density they measured, 0.95 Mg m–3, falls below the threshold b at which we found large soil strength increases. In fact, aeration porosity at that b exceeds 0.2 cm3 cm–3 and available water is not affected (Fig. 2). Compared with a clay and sandy loam soil, the loam Cohasset soil had the greatest increases in soil strength due to compaction, a finding similar to our strength results on compacted soil columns. Based on the LLWR we determined for this soil and data collected periodically from May to September by Gomez et al. (2002), the Cohasset LTSP site was within the LLWR during this period. Although we show that growth reductions due to poor aeration or high strength are possible for this soil, if moderate v contents are present during most of the growing season, these factors will have little effect on growth.
The applicability of studies conducted under greenhouse conditions is limited without field validation. Conditions that exist in the greenhouse soil column are not often found in the field. Forest soils have much greater spatial and temporal heterogeneity due to the actions of rocks, roots, animals, and climate in modifying the rooting environment. However, data published by Gomez et al. (2002) presented a chance to compare lab and field results for two of our soils. Their Blodgett site is the same California LTSP field site from which our Cohasset soil was taken, while their Rogers site is the Chaix soil series—a different series than our Dome, but formed from the same parent material and taxonomically very similar (coarse-loamy, mixed, superactive, mesic Typic Dystroxerepts). We used the b's Gomez et al. (2002) reported for the control and compacted field plots, their v for spring (May) and summer (July) 1999 field measurements (0–15 cm), and their stem volume increments for ponderosa pine in that year. We applied equations in Fig. 4 to their soil data to estimate RLD for their two sites and converted these to potential shoot weights from linear functions correlating shoot growth with root growth. Predicted shoot weights were transformed to relative shoot growth by setting growth for the control treatment to 1.00 (Table 4). Similarly, stem volumes measured in the field also were transformed to relative stem growth in 1999. Predicted relative growth from our models was compared against measured relative growth in the field.
|
v and overestimated field growth based on July v. This indicates that our model has potential for predicting growth in a field setting but further calibration is needed. Using integrated seasonal volumetric water content with the growth models, rather than point-in-time samples, would better test the models and their ability to predict growth.
Our models are the first step in the process of determining the potential root growth for trees growing in these soils. Root growth under various field and management conditions could be estimated by dynamically applying seasonal water content variations in conjunction with seasonal rooting patterns and the proportion of time that ideal water contents for growth are present. Spatial heterogeneity of field b and subsequent rooting patterns will also influence the ability of the model to predict productivity losses due to compaction. We used soil from the top 20 cm of the profile, which is the depth at which most roots are found and where compaction is often the greatest (Kozlowski, 1999). However, rooting is not restricted to the top 20 cm of soil, and roots will preferentially use any channels created by old roots or soil biota, thus reducing effects of compaction. In a naturally regenerated stand of loblolly pine, root density was greater in decomposing root systems at depths > 20 cm then in the soil matrix (Van Lear et al., 2000). Nambiar and Sands (1992) found that roots proliferated in perforations simulating natural soil channels of subsoil compacted zones, mitigating the effects of compaction on tree growth. Determining the percentage of time soil moisture is adequate and the percentage of roots utilizing root channels versus bulk soil matrix would help improve applicability of our models for assessing management impacts on root growth opportunity.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODS AND MATERIALSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
Received for publication May 22, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODS AND MATERIALSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  H. Blanco-Canqui and R. Lal Regional Assessment of Soil Compaction and Structural Properties under No-tillage Farming Soil Sci. Soc. Am. J., September 28, 2007; 71(6): 1770 - 1778. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||
