JEQ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (6)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kelly, J.M.
Right arrow Articles by Mays, P.A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kelly, J.M.
Right arrow Articles by Mays, P.A.
Agricola
Right arrow Articles by Kelly, J.M.
Right arrow Articles by Mays, P.A.
Related Collections
Right arrow Forest Soils
Right arrow Ecological Risk Assessment
Right arrow Other Models
Right arrow Nutrient Management
Journal of Environmental Quality 30:427-439 (2001)
© 2001 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America

TECHNICAL REPORT
ECOLOGICAL RISK ASSESSMENT

Hardwood Seedling Root and Nutrient Parameters for a Model of Nutrient Uptake

J.M. Kellya, J.D. Scarbroughb and P.A. Maysb

a Dep. of Forestry, Iowa State Univ., Ames, IA 50011-1021
b Tennessee Valley Authority, P.O. Box 920, Norris, TN 37838-0920

Corresponding author (jmkelly{at}iastate.edu)

Received for publication February 17, 2000.

   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Use of mechanistic models is an increasingly accepted way to evaluate complex processes. The Barber-Cushman model provides a means to simulate nutrient uptake once information on root system characteristics, nutrient uptake, and soil nutrient supply are developed. Objectives of this study were to determine during a growing season: (i) root growth for 1-yr-old black cherry (Prunus serotina Ehrh.), northern red oak (Quercus rubra L.), and red maple (Acer rubrum L.) seedlings; (ii) net plant increase in N, P, K, Ca, and Mg; (iii) soil solution and solid phase nutrient concentrations; and (iv) the influence of root growth and soil nutrient supply changes on nutrient uptake using the Barber-Cushman model. Seedlings were grown in pots containing A horizon soil from two forest sites. Measurements were made on five occasions during the growing season. Root growth averaged 41.5 cm d-1 for red maple compared with 28.0 and 16.7 cm d-1 for cherry and oak, respectively. Seventy-five percent of root growth occurred at the end of the growing season. Total plant N showed the greatest change (25–58%) due to soil source. Model simulations underestimated observed uptake by 31 to 99%. A clear relationship between soil solution nutrient concentration and plant uptake, an important assumption of the model, was not observed. Results indicate care will need to be exercised in the development and use of root growth and nutrient supply values in mechanistic models.

Abbreviations: b, buffer power • hsd, honestly significant difference • ICP, inductively coupled plasma


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
EFFORTS to assess impacts of environmental stress on forests have been hampered by the inherent complexity of forest ecosystems and the fact that forests respond slowly and inconsistently to stresses (Hallet and Hornbeck, 1997). There is a growing body of research that indicates increased levels of atmospheric CO2 (Will and Teskey, 1997), chronic ozone exposure (Anderson et al., 1991; Kelting et al., 1995), and changes in precipitation amount (McLaughlin and Downing, 1995) lead in varying degrees to changes in root growth and nutrient availability. The relative impact of a change in root growth on overall nutrient uptake and subsequent plant productivity will be dependent in part on the availability of nutrients and the pattern of nutrient uptake. If root growth is reduced but nutrient supply is high and all other environmental factors are favorable, the change will be of little or no consequence. However, if supply is less than plant needs or other environmental factors are less favorable, then additional C will need to be allocated to root growth or overall plant growth will be reduced by nutrient limitation. Because of the complexity of the situation created by the potential interaction of a number of independent variables, Hogsett et al. (1993)( 1997) have proposed the use of mechanistic tree process models as one means to assess impacts of environmental stress on forest growth and health.

The Barber-Cushman nutrient uptake model, as described by Oats and Barber (1987), has been used successfully to describe nutrient uptake in a variety of woody species (Van Rees et al., 1990; Gillespie and Pope, 1990; Kelly et al., 1992; Smethurst and Comerford, 1993; Kelly et al., 1994; Kelly and Kelly, 2001). Equations that describe both root growth and nutrient availability/movement are combined in the Barber-Cushman model and provide a means to mechanistically describe the influence of changes in root growth, root competition, and soil nutrient supply on nutrient uptake. To use this and other mechanistic models, information on soil nutrient supply, root morphological characteristics, and root uptake kinetics are required for the soil and plant species of interest.

While there is information in the literature on seedling, sapling, and mature tree root growth (Brundrett and Kendrick, 1988; Hendrick and Pregitzer, 1992), parallel information on nutrient uptake, soil solution nutrient concentrations, and soil nutrient supply patterns are generally not available from the same studies. It is important to understand how these values might change during the growing season and whether, as discussed by Kelly et al. (1994), instantaneous or seasonal mean values should be used in the model.

Objectives of the study were to use a simplified system provided by a pot study to determine over the course of a growing season: (i) root growth rates for black cherry, northern red oak, and red maple seedlings; (ii) the net increase in N, P, K, Ca, and Mg in the study plants; (iii) the soil solution and solid phase nutrient concentrations; and (iv) use the Barber-Cushman model as a means to assess the influence of changes in root growth and soil nutrient supply on plant nutrient uptake.

This study was undertaken as part of an ongoing comprehensive study of the impacts of ozone in the Great Smoky Mountains National Park in southeastern Tennessee (Samuelson and Kelly, 1997). The three tree species were chosen for study because they occur at the Twin Creeks and Cove Mountain intensive study sites and have a broad geographic distribution (Fowells, 1965) as well as demonstrated sensitivity to biospheric ozone (Davis and Skelly, 1992; Samuelson, 1994; Samuelson and Kelly, 1997). Nutrient analyses were focused on N, P, K, Ca, and Mg because of their status as macronutrients and because of anticipated similarities and differences in concentration levels between the soils at the two study sites (Kelly and Mays, 1999).


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Experimental Design
The study used all combinations of three species and two soil sources to provide the experimental material from which all data were collected. Pots were grouped by species with individual pots being randomly located within a species grouping. Sixty-five pots were prepared for each species x soil source combination so that 12 pots could be harvested on each of five dates with a few spares if needed. Pots to be harvested from each species x soil source combination on each sampling date were designated at the start of the study using a random number table. A total of 390 pots were prepared.

Plant Culture Techniques
Soil used for this study was collected from the A horizon of a loamy-skeletal, mixed, mesic Umbric Dystrochrept (Twin Creeks) and a loamy- skeletal, siliceous, mesic, Typic Haplumbrept (Cove Mountain) during the early spring of 1996. Further details on vegetation and soils at the collection site are available in Samuelson and Kelly (1997), Chappelka et al. (1997), and Kelly and Mays (1999).

In preparation for soil collection, O horizon material was removed from the underlying mineral soil. Soil was carefully removed from the A horizon and passed through a 2 by 2 cm mesh sieve to remove large roots and stones. The soil was then thoroughly mixed in bulk and placed in 7.5-L pots and taken to an outdoor nursery area on the Tennessee Valley Authority Reservation at Norris, TN. Samples of field moist soil were collected as the pots were filled so that soil weights could be expressed on a dry-weight basis. Each pot contained approximately 7.5 kg (oven-dry basis) of soil at an average bulk density of 0.78 Mg m-3. Bulk density was calculated using the soil dry weight and the volume of the pot occupied by soil. To prevent roots from passing from the bottom of the pot into the nursery soil, a sheet of 6-mil plastic was placed on the soil surface at the nursery site to serve as a barrier. One hundred thirty pots were seeded for each species using local half-sib seed provided by the Park Service. After germination, the seedlings were allowed to grow for one season and overwintered in the pots in the field. Natural rainfall was supplemented with additional irrigation water to promote seedling survival and growth during the first growing season.

At the beginning of the second growing season, a harvest program was initiated to follow root growth and nutrient uptake by the seedlings as well as changes in soil nutrient levels. The first harvest occurred on 13 Mar. 1997, before the initiation of leafout. The second and subsequent harvests were coordinated with the completion of a growth flush or the end of the 1997 growing season (Table 1). During the second growing season, rainfall ranged from 47 to 201 mm mo-1 with a total of 891 mm for the period March through September 1997. Irrigation was used as needed to keep the pot soil in a well-watered condition. Mean monthly temperatures during the 1997 growing season were 10.8, 10.5, 15.3, 21.1, 24.4, 22.5, and 19.7°C for March through September, respectively. Temperature extremes during the 1997 growing season ranged from a low of 2.8°C in March to a high of 31.1°C in July.


View this table:
[in this window]
[in a new window]
  		Table 1. Harvest dates for black cherry, northern red oak, and red maple seedlings during the 1997 growing season and the number of days elapsed (d) between each harvest

 
Processing of Seedlings
On each harvest date, 12 seedlings from each tree species x soil source combination were chosen at random for harvest. Seedlings were carefully separated from soil to avoid breaking of roots. Root systems were separated from the shoot just above the root collar. Broken roots remaining in the soil, were collected by spreading the soil in a thin layer and systematically searching for root pieces. All roots were placed on a 2-mm mesh screen and washed with tap water. Root systems were weighed and placed in a plastic bag and frozen at -10°C to await further processing. Frozen root samples were subsequently used to determine total root length values for each plant using a modified line intercept method (Tennant, 1975). Total length of all roots, independent of root diameter, and root fresh weight were used to calculate mean root radius according to the method described by Mackay and Barber (1985). Following measurement of length, the root samples were oven-dried for at least 72 h at 65°C and the dry weight determined.

Immediately after harvest, the shoot portion of each seedling was oven-dried at 65°C for at least 72 h and an individual dry weight determined for each plant. To reduce analytical costs, the oven-dried shoot and root materials from three whole plant samples per soil source were combined into a single sample for chemical analysis. Seedlings in the composite sample were cut into small segments, mixed thoroughly, and then ground in a Wiley Mill to pass a 1-mm screen. All composite samples included leaves except those from the 13 March harvest that occurred before leafout. Compositing seedlings resulted in a total of four samples for chemical analysis for each harvest date x species x soil source combination. Plant samples were analyzed for total N by micro-Kjeldahl (Bremner and Mulvaney, 1982). Phosphorus, K, Ca, and Mg were determined by dry ashing followed by inductively coupled plasma (ICP) analysis (Greensburg et al., 1995). Analytical values from the composite samples were then multiplied by the average weight of the three composite plants to develop estimates of average plant nutrient content.

Soil Analysis
In conjunction with each plant harvest, a soil sample representative of the entire pot profile was collected from each pot. As with the plant samples, soil samples were combined in groups of three to form four composite samples for each species x soil source combination. The same pots were used to create both the soil and plant composites. These soil samples were used in both the solution and solid phase nutrient determinations. Solution phase nutrient samples were collected using the displacement procedure described by Adams (1974). Solution samples were analyzed for NO3–N (auto Cd reduction), NH4–N (auto phenate), P, K, Ca, and Mg (ICP) using methods described in Greensberg et al. (1995). For the solid phase determinations, soil from each composite sample was air-dried and exchangeable P, K, Ca, and Mg determined using the dilute double acid (0.5 M HCl + 0.025 M H2SO4) extraction of Helmke and Sparks (1996) followed by ICP analysis as described by Issac and Kerber (1971). Nitrate and NH4–N concentrations were determined following extraction with 0.1 M KCl using the method described by Greensberg et al. (1995). In both the plant and soil analyses, calibration standards were run before and after each sample set and were required to fall within 2% of the known value. Standards and rechecks constituted approximately 10% of each sample set.

Data Processing and Statistical Analysis
Destructive samples were collected on each of five harvest dates with 12 randomly designated pots taken from each treatment combination. Measurements of root length, root radius, and total plant mass are based on 12 observations per species x soil source combination. Measurements of soil solution and solid phase nutrient concentrations, as well as plant nutrient concentrations and contents, are based on four composite observations for each species x soil source combination. Analysis of variance was used to test for differences in root length, root radius, plant nutrient content, solution phase nutrient concentration, and solid phase nutrient concentration as a function of species, soil source, and harvest date. Mean separations, where appropriate, were performed using Tukey's honestly significant difference (hsd) procedure (SAS Inst., 1990). The 0.05 level of probability was used as the decision level for the acceptance or rejection of statistical significance in all analyses.

Modeling Nutrient Uptake
The Barber-Cushman nutrient uptake model as described by Oats and Barber (1987) was run using root morphology and soil supply parameters developed from the data reported from this study. Uptake kinetics values (Imax, Km, Cmin) were obtained from companion studies previously reported (Kelly et al., 2000; Kelly and Kelly, 2001) or calculated with the model as described by Kelly et al. (1994) based on observed changes in nutrient content. Values used to conduct the initial model runs are growing season means and are listed in Table 2. Uptake of N, P, and K by red maple was chosen as the focus of the modeling efforts because more information was available from previous studies. The growing season was defined as 176 d for the P and K simulation based on the length of the red maple study. The N simulation had to be shortened to 30 d.


View this table:
[in this window]
[in a new window]
  		Table 2. Nitrate, P, and K soil supply estimates; root morphological parameters, and uptake kinetics values used in the Barber-Cushman model to simulate 1-yr-old red maple seedlings growing in the Cove Mountain soil. Values reported are means based on all five harvests unless indicated otherwise

 

   RESULTS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Root Parameters
Root length values increased from fourfold to sixfold (Fig. 1) with most of the increase occurring at the end of the growing season. Statistical analyses indicated a difference in total root length among the three species. However, there were no differences in root length for any of the three species attributable to the two soils evaluated. Red maple roots exhibited the greatest change with an increase in length in both soils in excess of 7000 cm. Black cherry increases were slightly less than 5000 cm, whereas northern red oak seedlings averaged about 3300 cm for the two soils. Mean separation analysis indicated that only the fifth harvest root length values for black cherry and red maple were different from values observed in the earlier harvests. Root length values from the last two northern red oak harvests were significantly greater than those observed in the first three harvests (Fig. 1).



View larger version (34K):
[in this window]
[in a new window]
  		Fig. 1. Changes during the second growing season in mean root length (cm) for black cherry, northern red oak, and red maple seedlings grown in pots containing A horizon soil from two forest sites, Twin Creeks and Cove Mountain. Means within a species x site combination marked with the same letter are not different based on Tukey's HSD test (P = 0.05)

 
Root radius differed significantly by species, harvest, and soil source. Mean separation indicated a variable response of root radius over time (Fig. 2) . No statistical differences in root radius were observed for any of the three species when grown in the Twin Creeks soil (Fig. 2); the root radius means for all three species varied as a function of harvest for seedlings grown in the Cove Mountain soil. No species x soil source, soil source x harvest, or species x harvest interactions were observed in either the root radius or root length response.



View larger version (46K):
[in this window]
[in a new window]
  		Fig. 2. Changes during the second growing season in mean root radius (cm) for black cherry, northern red oak, and red maple seedlings grown in pots containing A horizon soil from two forest sites, Twin Creeks and Cove Mountain. Means within a species x site combination marked with the same letter are not different based on Tukey's HSD test (P = 0.05)

 
Plant Biomass
Differences were observed among the three species in the total amount of plant biomass produced by the time of the fifth harvest. Total plant biomass at the end of the study period was also altered by the soil in which the seedlings were grown (Fig. 3) . Seedlings grown in the Twin Creeks soil consistently had a greater total mass (Fig. 3). The greatest growth generally occurred between the fourth and fifth harvest with the northern red oak seedlings being the largest of the three species (Fig. 3).



View larger version (29K):
[in this window]
[in a new window]
  		Fig. 3. Mean total plant weight (g) on each of five harvest dates for black cherry, northern red oak, and red maple seedlings grown in pots containing A horizon soil from two forest sites, Twin Creeks and Cove Mountain. Means within a species x site combination marked with the same letter are not different based on Tukey's HSD test (P = 0.05)

 
Plant Nutrient Content and Uptake
Differences in nutrient content among the three species were observed with black cherry seedlings (Fig. 4a) typically having a lower nutrient content compared with northern red oak (Fig. 4b) and red maple (Fig. 4c) seedlings. Comparison of the first harvest nutrient content values (Fig. 4) indicates that, in most cases, seedling nutrient content had not been influenced substantially by the soil source during the first year of growth. Red maple N and Ca are the exceptions with reduced contents of both nutrients already evident in seedlings grown in the Cove Mountain soil. Part of this difference may be attributable to the difference in plant weight. However, black cherry with approximately the same biomass difference did not differ in N or Ca. However, during the second growing season, soil source did have a significant impact on both seedling total N and Mg content (Table 3). As might be anticipated, there were differences in nutrient content as a result of date of harvest. At the fifth harvest, nutrient content had increased severalfold and differences attributable to soil source had developed. Nitrogen values at the end of the growing season were 25 to 58% higher in the seedlings grown in the Twin Creeks soil compared with those grown in the soil from Cove Mountain. Plant content of other nutrients were also generally higher in the seedlings grown in the Twin Creeks soil, but not to the degree observed for total N (Fig. 4). Species x soil source and species x soil source x harvest date interaction terms were not significant.



View larger version (36K):
[in this window]
[in a new window]
  		Fig. 4. Mean plant total N, P, K, Ca, and Mg content (mmol) on each of five harvest dates for (a) black cherry, (b) northern red oak, and (c) red maple seedlings grown in pots containing A horizon soil from two forest sites, Twin Creeks and Cove Mountain. Means within a species x site combination marked with the same letter are not different based on Tukey's HSD test (P = 0.05). Note differences in the content scale for each species

 

View this table:
[in this window]
[in a new window]
  		Table 3. Plant nutrient content P values for main treatment response of N, P, K, Ca, and Mg

 
Over the course of the growing season, seedlings of all three species generally exhibited the greatest net gain in nutrients when rooted in the Twin Creeks soil (Table 4). Total N differed most in net gain (Table 4) with seedlings grown in the Twin Creeks soil acquiring 1.5 to 2.8 times more total N than seedlings grown in Cove Mountain soil.


View this table:
[in this window]
[in a new window]
  		Table 4. Mean net gain in plant nutrient content (mmol) of N, P, K, Ca, and Mg by black cherry, northern red oak, and red maple seedlings at the end of the study period. Seedlings were grown in pots containing A horizon soil from Twin Creeks (TC) or Cove Mountain (CM)

 
Solution Phase Soil Nutrient Concentrations
Solution phase nutrient concentrations for most nutrients in both soils peak in the samples from the first or second harvest in the cherry (Fig. 5a) and oak (Fig. 5b) pots, while peak values in the red maple (Fig. 5c) pots generally occurred at the fourth harvest. This observation implies peak nutrient availability during the late April to early May time frame in the cherry and oak pots and later in the season for red maple. Both species and soil source had significant impacts on soil solution NO3–N, P, K, Ca, and Mg (Table 5). Essentially no solution phase P was detected in any of the samples collected, with the exception of the first northern red oak harvest sample. However, even this value is just above the detection limit of the analytical technique. Extremely low solution values for NH4–N were also observed (Fig. 5).



View larger version (40K):
[in this window]
[in a new window]
  		Fig. 5. Solution phase NO3–N, NH4–N, P, K, Ca, and Mg concentrations (µmol) on each of five harvest dates from (a) black cherry, (b) northern red oak, and (c) red maple pot soils from Twin Creeks and Cove Mountain. Means within a species x site combination marked with the same letter are not different based on Tukeys HSD test (P = 0.05)

 

View this table:
[in this window]
[in a new window]
  		Table 5. Solution and solid phase P values for main treatment response of NO3–N, NH4–N, P, K, Ca, and Mg concentrations

 
Solid Phase Soil Nutrient Concentrations
Tree species, soil source, and harvest date had a significant effect on solid phase NO3–N, P, K, and Ca concentrations (Table 5). Lowest concentrations of NO3–N (Fig. 6) were observed in samples collected at the time of the third harvest (5–16 June) with peak concentrations generally observed approximately 6 wk later in the fourth sample (16–28 July). No clear pattern was evident in the solid phase P and K concentrations as a function of harvest date (Fig. 6). However, there was a general tendency for the Cove Mountain soil to have higher solid phase P and K concentrations.



View larger version (53K):
[in this window]
[in a new window]
  		Fig. 6. Solid phase NO3–N, NH4–N, P, K, Ca, and Mg concentrations (µmol) on each of five harvest dates from (a) black cherry, (b) northern red oak, (c) and red maple pot soils from Twin Creeks and Cove Mountain. Means within a species x site combination marked with the same letter are not different based on Tukey's HSD test (P = 0.05)

 
Solid phase NH4–N concentrations differ with species and harvest date, but not soil source. Lowest NH4–N values were observed at the time of the second harvest and extended in most cases into the third harvest (Fig. 6). Peak NH4–N values were observed at the time of the fourth harvest for all three species. Magnesium concentration did not respond to species, but did respond to soil source and harvest date. As a general trend, solid phase Mg concentrations increased as the growing season progressed (Fig. 6), although not always supported by the mean separation analysis.

Model Simulations
Based on the initial values for N, P, and K presented in Table 2 model estimates of red maple uptake were 6078 µmol for N, 1.53 µmol for P, and 920 µmol for K. All model estimates are substantially less than the corresponding net increase values reported in Table 4 for each nutrient. For comparative purposes, the net increase values for red maple growing in the Cove Mountain soil were 8810, 830, and 1890 µmol, respectively, for N, P, and K. These values provide a lower limit estimate of uptake since they do not account for any nutrients taken up and subsequently lost to foliar leaching or tissue mortality. However, as a simplifying assumption for modeling purposes, the net increase values presented in Table 4 will be assumed to approximate total uptake. Note that the length of the simulation period for N uptake had to be reduced to 30 d due to the relatively low buffer power value. A buffer power value in excess of 7.5 was needed for the model to run for the full study period. Simulated N uptake using a buffer power value of 7.5 exceeded observed uptake by >10-fold (99385 µmol).


   DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
The length of roots present at the beginning of the study period was quite similar (~1100 cm) for both black cherry and northern red oak (Fig. 1). This is in contrast to red maple seedlings that averaged 2078 cm across both soil sources. Since all plants were started from seed, these values give a first approximation of the net amount of root growth that occurred in the first growing season and the initial differences in root growth among the species. The first harvest data reflect not only aggressive root growth by red maple (Fig. 1), but also indicate by inference that a larger portion of fixed C was allocated to roots than shoots when total plant weights (Fig. 3) are compared to northern red oak. At the end of the study period, black cherry roots had a mean net increase in length across both soil sources of 5050 cm. Northern red oak had a mean net addition of 3371 cm of length while the red maple mean net increase was 7998 cm. Converting these increases to daily root growth rates averaged over the growing season provides a range of values from 16.7 cm d-1 for northern red oak, to 28.0 cm d-1 for black cherry, to 41.5 cm d-1 for red maple. It should be noted that these values represent a minimum estimate since no allowance is made for any root mortality that may have occurred.

Because most of the net increase in root length occurred at the end of the growing season, using either seasonal averages or values from a single harvest for modeling purposes would overestimate root growth for most of the growing season and substantially underestimate length increase and uptake at the end of the season. For example, using the growing season mean values for K listed in Table 2 gave an estimated uptake of 920 µmol. If all other values remain constant, and uptake is estimated using the sum of the average root growth rate for the first 129 d of the study period (8.98 x 10-5 cm s-1) added to the estimate obtained using the root growth rate for the final 47 d (1.53 x 10-3 cm s-1), uptake is estimated to be 284 plus 249 for a total of 553 µmol. This value is 58% of the original model output and only 28% of the observed increase. For model output to approximate observed uptake, an average root growth rate of 1.19 x 10-3 cm s-1 would be required.

Since nutrient uptake is strongly influenced by root surface area, the potential exists within the Barber-Cushman model for a disconnect between root growth patterns and any seasonality in nutrient availability patterns (Fig. 4). Using plant N content for all species and harvest dates as a surrogate of uptake, Fig. 7 illustrates the relationship between root surface area and plant N content. Certainly there is a degree of auto correlation between root surface area, total plant weight, and plant nutrient content at each harvest. Nevertheless, Fig. 7 does illustrate the importance of root growth changes, as reflected in root surface area, and plant nutrient content.



View larger version (13K):
[in this window]
[in a new window]
  		Fig. 7. Relationship between seedling N content and root surface area across all species, soil sources, and harvest dates. A linear function provided the best fit to the data

 
Therefore, use of a seasonally derived root growth average or an instantaneous value would be problematic if the desired modeling outcome was a reasonably accurate representation of the nutrient uptake pattern within a single growing season. If, however, the modeling goal involved an analysis of the cumulative impact of several growing seasons, use of growing season averages for all parameters might be less problematic. Models such as those developed by Weinstein et al. (1991) and Yanai (1994) do not place any limitations on the pattern of root growth during a simulation. This is much better than the fixed value approach used in the Barber-Cushman model. However, to do so requires a more complex model and additional data inputs. Although not the focus of this study, there is also reason to question how consistent the observed pattern of root growth might be. For example, Cote et al. (1998) in a 2-yr study of fine root growth in a maple stand observed that production peaked in May and then declined steadily for the remainder of the growing season in Year 1. In the second year, growth peaked in July with a second peak occurring in September. Price and Hendrick (1998), in a study of a sweetgum coppice, found that total root length increased at a fairly linear rate from April through mid-September. Hendrick and Pregitzer (1992)(1997) found maximum fine-root length production in the northern hardwood forests they studied to occur in spring and early summer. These studies combined with the present work point to the need for more than one season of observation to establish the most probable pattern of growth for a particular species given that variation in both biotic and abiotic factors can interact to create a variety of potential root growth patterns. Previous studies (Eissenstat, 1991; Price and Hendrick, 1998) have generally found root radius changes to be smaller than root length changes. Root radius values in this study clearly vary with species and even exhibited differences within a species in the Cove Mountain soil (Fig. 2) as a function of harvest date. The actual root radius values reported are quite comparable to the fine root means reported by Hendrick and Pregitzer (1993) and Price and Hendrick (1998). It should be noted that the studies just cited are examples from forest stands in contrast to the potted seedlings used in the current study. It should be further noted that roots of all sizes were used in our mean radius calculations, while those <2 mm are generally used in most field studies. In our case we do not believe the inclusion of >2-mm roots represents an important bias because qualitative evaluations indicated that <5% of the roots from our plants were >2 mm in diameter.

Canham et al. (1996), in a greenhouse study of 2-yr-old seedlings, found an increase in red maple shoot growth when soil resources were plentiful and an increase in root growth when soil resources were low. Northern red oak, on the other hand, exhibited only small changes in allocation to root growth in response to soil resource availability (Canham et al., 1996). In this study, northern red oak seedlings, with the lowest observed root length values, produced the greatest amount of total plant mass (Fig. 3). Red maple, on the other hand, with more than twice the root length of northern red oak had 40% less total biomass production. Based on the root and shoot growth patterns identified by Canham et al. (1996), observations from this study indicate that soil nutrient levels were adequate for northern red oak and less than adequate for red maple. However, if red maple had a higher rate of root turnover than northern red oak, this observation would not hold. Since we have no estimates of root mortality and loss, we cannot resolve this uncertainty.

As with any pot study, there are concerns with how representative these data may be of the forest situation. At a larger scale, there are demonstrated physiological differences between seedlings and trees (Grulke and Miller, 1994; Samuelson and Edwards, 1993; Samuelson and Kelly, 1997). A large pot (7.5 L) was used in this study to provide a greater soil volume for the root system to explore. Roots were well distributed throughout the pot, but in the end, roots did occur along the pot–soil interface. Conceptually, the occurrence of roots along the pot–soil interface is less problematic from the perspective of developing model input parameters if one is willing to assume that soil solution nutrient content under well watered conditions is in a dynamic equilibrium throughout the pot and that roots at either the center or outside edge of the pot have similar access to nutrients through the soil solution. Bakker et al. (1999) found relatively little variation between the bulk soil solution and rhizosphere soil solution when solution ionic strength was increased by either liming or the addition of rain water. However, when deionized water was added to the soil, significant differences between bulk and rhizosphere concentrations were observed. Work by Bakker et al. (1999) does lend some support to the pot solution equilibrium assumption because both rain and irrigation water contained dissolved nutrients.

Unfortunately, our ability to study root systems in the field is greatly complicated by our inability to measure roots accurately using available techniques, since all accepted approaches have limitations (Vogt et al., 1998). Root estimates from pot studies, while not perfect, do provide data from a simplified system. Pot studies also provide a greater degree of control for root recovery than is possible in most field studies.

The substantial differences among the species in root length and root growth rate should reflect a greater ability to obtain nutrients if uptake is largely a function of root surface area as implied in Fig. 7. Red maple seedlings grown in the Twin Creeks and Cove Mountain soils had at the time of the final harvest mean root surface areas of 2514 and 2073 cm2, respectively. Black cherry seedlings were at the low end of the observed range with 1603 and 1257 cm2, while northern red oak values were intermediate at 1790 and 1657 cm2 for Twin Creeks and Cove Mountain soils, respectively.

Comparison of these root surface area estimates with the nutrient net gain values in Table 4 indicates that total uptake is not always directly related to total root surface area alone. Estimates of net N increase on a daily basis presents a slightly different picture with black cherry having a rate of 0.057 µmol cm-2 d-1, followed by red maple at 0.055 µmol cm-2 d-1, and northern red oak at 0.043 µmol cm-2 d-1. The daily values illustrate the fact that there are species differences in mean uptake rate per unit of root surface area. Black cherry with 10% less root surface area than northern red oak seedlings, accumulated 10% more N. This lack of direct correlation is also illustrated by comparison of the K, Ca, and Mg values in Table 4 where northern red oak with an intermediate amount of root surface had uptake levels sometimes in excess of twice the black cherry and red maple levels. Increases in root length are very important to the acquisition of nutrients such as P that diffuse slowly in the soil (Clarkson, 1985), while uptake of more mobile ions such as NO-3, NH+4, and K+ appear to be less responsive to root elongation (Robinson and Rorison, 1983). This may explain in part why root growth has been widely referenced as responding positively to soil-P fertility and negatively to soil-N fertility (Cote et al., 1998).

It is an established fact that species differ in their nutrient requirements (Reuter and Robinson, 1997), so the species differences observed are not surprising. Clarkson (1985), summarizing the work of several authors, notes that plant uptake of nutrients is influenced by prior nutrient status of the plant, plant age, and the inherent vigor of the plant. Data from this study point out that any attempt to simplify the modeling process by focusing on root surface area as the only determinant of uptake may lead to substantial overestimates or underestimates of uptake. Previous work by Kelly et al. (1994)(2000) suggests that the potential exists for genotypic differences in the kinetics of nutrient uptake, even within a species. Therefore, a good understanding of the range in uptake kinetics values within a species may also be important to successful model representations.

Mechanistic nutrient uptake models of the type described by Barber and Cushman (1981), Smethurst and Comerford (1993), and Yanai (1994) rely on the simplifying assumption that all nutrient uptake occurs through the solution phase. Comparison of the solution phase concentrations of NO3–N for all three tree species (Fig. 5) indicates that although a difference in liquid phase concentration was observed (Table 5), the magnitude of the difference in the solution concentration between the two soils seems inadequate to explain the more than twofold difference observed in plant content at the end of the study (Fig. 4). Peak solution concentrations for most of the nutrients evaluated occurred at the second harvest (Fig. 5), while the peak period in plant growth (Fig. 3) and plant nutrient content (Fig. 4) is generally reflected in the fifth harvest. The largest growth increments and increases in nutrient content generally occur between the fourth and fifth harvest for all species. While an increase in solution concentration was observed at the time of the fourth harvest, it is unclear if the magnitude of this increase is sufficient to account for the level of change in plant content observed. These observations suggest some degree of uncertainty in the direct relationship between the solution phase and plant growth and nutrient content. An increase in plant mass without a parallel increase in nutrient content has the effect of diluting plant nutrient concentration. This phenomenon has been observed in many plant species where a period of uptake and high nutrient concentration is followed by a period of rapid growth and reduction of nutrient concentration (Marschner, 1995). This may in part explain the observed species differences as well as the lack of a clear-cut relationship between solution concentration and seedling growth and nutrient content within an interval. An alternative explanation may be that we are attempting to describe a cumulative and dynamic process with a series of instantaneous measures that may or may not give a clear picture of nutrient uptake and availability.

Comparison of the solution phase P concentrations and P uptake by all three species argues against the solution phase assumption. In this study, solution phase concentrations were below the detection limit and thus effectively zero, yet plant uptake of P was observed to occur throughout the growing season. Earlier work (Kelly and Mays, 1999) with the same two soils produced similar observations on field soil solution P concentrations, so the low P levels do not appear to be an artifact created by using potted soils. At this point, it is our assumption that the apparent disconnect is more a result of solution phase extraction methodology than an error in model assumptions. Previous work on P availability, as summarized by Barber (1995), has not identified this situation in the higher P availability environment associated with agronomic crops, although as observed by Bakker et al. (1999) there is some reason to question this in low supply situations. Clarkson (1985) observed that while uptake estimates provided by models of the type developed by Barber and Cushman (1981) and others have been most impressive in high fertility situations, their predictive powers deteriorate when nutrient concentrations are low.

Solid phase concentrations (Fig. 6) do not appear to resolve these questions either. In the case of N, both NO3–N and NH4–N solid phase concentrations generally peak in the fourth harvest samples and support the major increase in plant N content between the fourth and fifth harvest. Phosphorus, K, Ca, and Mg solid phase concentrations were generally highest at the time of the third harvest (Fig. 6), while solution phase concentrations for the same nutrients, excluding P, peaked at the fourth harvest (Fig. 5). Although plant N content exhibited the greatest increase during the interval between the fourth and fifth harvest, the content of other nutrients increased more between the third and fourth harvest. This again may reflect a plant strategy to collect a reserve of nutrients in preparation for the flush of growth observed in the fifth harvest (Fig. 3).

Comparison of the soil solution and solid phase concentrations for all nutrients indicated a very low level of direct correlation between the two. Correlation coefficients for all nutrients except K were <0.10. Figure 8 illustrates the relationship between solid and solution phase K across all species for the Twin Creeks soil. While this figure depicts the strongest relationship observed, it is still rather weak.



View larger version (12K):
[in this window]
[in a new window]
  		Fig. 8. Relationship between solution phase and solid phase K concentration across all species and harvest dates for the Cove Mountain soil. A power function provided the best fit to the data

 
In the model, the relationship between the solution phase and solid phase is modulated through the buffer power (b). The simulated response of NO3 uptake (Table 2) illustrates the importance of the buffer power. In this simulation the buffer power was exhausted by the end of 30 d. One explanation for this is that the solid phase determination used to calculate the buffer power does not measure the amount of N made available during the growing season by biological processes. Solid phase N, as measured, represents a relatively small segment of the plant-available N and this accounts for the relatively small b value. Simply increasing the b value by one order of magnitude from 1.35 to 13.5 and holding all other values constant results in an increase in simulated uptake from 6078 µmol in 30 d to 179992 µmol in 176 d. The excessive level of simulated uptake compared with the observed value is due in large part to the fact that in reality other model values would not remain constant. For example, the higher b value allows the simulated plant to take up N at the maximal rate throughout most of the simulation. This observation has implications for model simulations where nutrient availability is arbitrarily set at nonlimiting levels.

Ingestad (1979) and colleagues have for some time asserted that plant growth and nutrient content should be addressed on a more holistic basis and that more attention should be focused on the ratios of nutrient supply rather than absolute amounts of individual nutrients. Some of our inability to explain observed responses in this study may in part be due to a failure to understand the complex interactions among nutrient availability, nutrient uptake kinetics, the influence of prior nutrient status of the plant, and root response to nutrient availability. Forest tree species exhibit varying degrees of plasticity in their capacity to deal with nutrient availability, natural and anthropogenic stresses, and other site differences. The three species used in this study are widely distributed species occurring across most of the eastern half of the USA and into southeastern Canada (Fowells, 1965). This wide distribution speaks strongly for a high degree of plasticity and opens the door for a variety of interactions and compensation mechanisms that allow these species to be successful across such broad ranges. Thus, modeling even simple systems such as seedlings grown in pots, may not be straightforward.

In summary, the results of this study indicate substantial differences among the three species in the length and growth rates of roots. In general, the results of this study support the observation of Price and Hendrick (1998) that the overall effects of soil resource availability on root system dynamics are unclear. While total root surface area plays an important role in nutrient uptake (Clarkson, 1985), its relative importance varies with species and the nutrient in question. These differences may be in part due to some degree of plasticity in the uptake kinetics of the various nutrients. These observations point to a need to allow root growth rate to vary across the growing season rather than remaining a fixed value as is the Barber-Cushman approach. Doing so will minimize mismatches between nutrient supply and root surface area available for uptake. Likewise, nutrient supply functions need to reflect the changing pattern of supply during the growing season rather than being fixed at nonlimiting levels as is the case in many model scenarios. The results of this effort do not support a clear relationship between soil solution concentration and plant nutrient availability, an important assumption of the Barber-Cushman model. Although not addressed as part of this study, part of this lack of correspondence may be due to either the method of measurement or the generally lower levels of nutrient availability in forest soils, or both. To address this uncertainty a closer evaluation of the relationship between bulk soil solution and rhizosphere solution chemistry needs to be undertaken at the relatively low concentrations typical of many forest soils. As noted in the introduction, nutrient uptake even in simple systems, is a potentially complex process that integrates and responds to both plant and soil processes. At a minimum the results of this study indicate considerable care will need to be exercised in both the development and application of root growth rate, root length, root surface area, and soil nutrient supply values used in assessment models.


   ACKNOWLEDGMENTS
 
The authors express appreciation to J.K. Kelly and G. Zames for assistance with root measurements, sample processing, and data summary. Appreciation is also expressed to Dr. N.S. Nicholas and Dr. L.J. Samuelson for their constructive reviews of the draft manuscript. The authors also express appreciation to the National Park Service, Great Smoky Mountains National Park, for site access and logistical support.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Journal Paper no. J-19090 of the Iowa Agriculture and Home Economics Experiment Station, Ames, IA, Project no. 3440, supported in part by McIntire-Stennis and State of Iowa funds as well as a grant from the Tennessee Valley Authority (TV-96584V).


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (6)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kelly, J.M.
Right arrow Articles by Mays, P.A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kelly, J.M.
Right arrow Articles by Mays, P.A.
Agricola
Right arrow Articles by Kelly, J.M.
Right arrow Articles by Mays, P.A.
Related Collections
Right arrow Forest Soils
Right arrow Ecological Risk Assessment