| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Forestry, 3108 Jordan Hall, College of Natural Resources, North Carolina State Univ., Raleigh, NC 27695
b Adirondack Watershed Institute, Paul Smith's College, Routes 86 & 30, P.O. Box 265, Paul Smiths, NY 12970
* Corresponding author (keltind{at}paulsmiths.edu)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: DAP, diammonium phosphate
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
One of the largest gains in biomass production in southern pine plantations comes from fertilization (Stanturf et al., 2003). Fertilization may increase soil C as a beneficial result of enhancing peak leaf area index (LAI) which results in increased biomass production and return of litter. Studies investigating loblolly pine plantations in the U.S. South have found LAI to double with fertilization (Albaugh et al., 1998; Lai et al., 2002). In addition, fertilization has been proven to enhance coarse root biomass production (Albaugh et al., 1998; Retzlaff et al., 2001). This large increase in foliage and coarse root biomass production with fertilization translates into increased inputs to the forest floor and soil, respectively. As these larger quantities of forest floor and root biomass material decompose, corresponding increases in the level of soil C should be expected.
The ability of the soil to sequester the additional C fixed in a fertilized loblolly pine plantation is largely dependent on the soil's capacity to retain and protect the C from respiration and/or leaching loss. Retention of C is largely a function of soil texture, mineralogy, and drainage class. On well-drained upland sites, finer textured soils (clayey soils) have a higher C retention capacity than coarse textured soils (sandy soils) (Sparks, 1995; Jastrow and Miller, 1998; Hassink et al., 1997; Amelung et al., 1998). Organic matter can be stabilized (turnover time substantially increased) by complexation with clay minerals on charged surfaces and physical protection within very small pores (Oades et al., 1989). In addition, there may be a greater potential for C storage at greater depths in finer textured soils, with the organic matter potentially being adsorbed by clay minerals in deeper horizons, a beneficial outcome for long-term soil C storage.
Since soils serve as the largest terrestrial reservoir of C (Pennock and van Kessel, 1997; Percival et al., 2000) and forest soils contain 50 to 63% of all soil C (Kimble et al., 2003), identifying and applying management strategies to intensify the storage of C in this pool could have major impacts on future terrestrial C retention. Fertilization and soil textural classes are two possible factors that may function independently or in combination to positively affect soil C sequestration via enhanced biomass production and the ability for long-term retention of C, respectively.
The objective of this study was to determine the effect of N + P fertilization at planting on whole ecosystem C storage on two upland sites with contrasting soil textures. It was hypothesized that fertilization would increase whole ecosystem C storage and the increase in soil C with fertilization would be greater for the site with the finer-textured soil.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The site in Alabama was a well-drained clayey soil of the Greenville (Fine, kaolinitic, thermic Rhodic Kandiudults) series, which was formed in clayey marine sediments. The site in Mississippi was a well-drained sandy soil of the Eustis (Siliceous, thermic Psammentic Paleudults) series, which was formed in coarse-textured fluvial sediments.
The previous stands were 28- and 26-yr-old slash pine (Pinus elliottii Engelm.) plantations in Alabama and Mississippi, respectively. Both sites were clear-cut harvested in winter of 1989, followed that summer by mechanical site preparation via a three-pass shear, rake, and disk. Following mechanical site preparation, the experiment was installed as a randomized complete block design with four blocks. Two 40 m x 40 m treatment plots, with 20 m x 20 m measurement plots centered within each, were established in each block for a total of 8 plots per site. The plots were randomly located within each block with a minimum buffer of 10 m between each plot.
Herbicide (3.5 L Velpar ha–1; Velpar [C12H20N4O2], DuPont, Wilmington, DE) was applied to all plots on both sites in April 1990, followed by planting loblolly pine seedlings at a density of 1850 trees per hectare. Two fertilization treatments were randomly assigned to the plots in each block: no fertilization (control) versus fertilization at planting with 45 kg N and 50 kg P per hectare. The fertilizer was applied as a broadcast application of 250 kg ha–1 of DAP. No further silvicultural operations were done throughout the duration of this study.
Field Collection
Pretreatment soil samples, herein referred to as Year 0 soil samples, were collected from the surface 0- to 10-cm and 10- to 20-cm mineral soil depths at 10 random locations within each plot using a 2-cm i.d. push tube in April 1990. These subsamples were composited at time of collection for one sample per plot, air-dried, and archived by the North Carolina State Forest Nutrition Cooperative in Raleigh, NC.
Eleven years after stand establishment, soil cores (Year 11) were collected at five random locations within each plot from the surface 0 to 50 cm of mineral soil using a 6-cm i.d. Ruark soil coring sampler (Ruark, 1985) in May 2001. These 50-cm cores were divided into five 10 cm increments (i.e., 0- to 10-cm, 10- to 20-cm, etc.).
Forest floor samples were collected in December 2001 using a 35 cm x 35 cm grid at six random locations within each plot. The samples were divided into the Oi (fresh litter or L layer) and Oe+Oa (fragmented litter and humus or F+H layer) layers at the time of collection and composited by plot. Five subsamples of an additional 1 cm of mineral soil were collected and composited from the six random locations beneath the Oe+Oa layer within each plot using a 2- cm i.d. push tube in December 2001 for a total of 30 subsamples that were mixed to yield one composite sample per plot.
Sample Preparation and Analysis
Soil samples were air-dried and passed through a 2-mm sieve to separate the fine and coarse fractions. The coarse fractions (roots and rocks) were removed, dried, and weighed to be accounted for in the calculation of bulk density for the fine earth fraction (<2 mm). Subsamples of the fine earth fractions were ground using a 60 mesh (1 mm) grinder for analysis.
Forest floor samples were air-dried for 2 d, oven-dried 80°C for 3 d, and weighed. Subsamples were ground using a 1-mm grinder for analysis. The loss on ignition procedure (Nelson and Sommers, 1996) was used to correct for mineral soil contamination within the forest floor samples.
All ground subsamples (soil and forest floor) were analyzed for total C using a CHN elemental analyzer (CE Instruments–Model NC2100, CE Elantech Inc., Lakewood, NJ). Total C concentration data were converted to a content basis (Mg ha–1) using bulk density for soil C and mass for forest floor C. Bulk density data for pretreatment samples (Year 0) were not available; therefore, bulk density values for samples collected in May 2001 (Year 11) were used in estimating total C content at Year 0. The total C data for the five soil cores collected in May 2001 were averaged to yield one value per plot to be compared with pretreatment samples.
Tree Volume and Biomass
Tree diameter at breast height (DBH = 1.3 m) and total height were measured every 2 yr at both sites, and volume was estimated using an equation from Smalley and Bower (1968). Allometric equations from the NUTREM 2.0 (Ducey and Allen, 2001) model were used to predict tree biomass components (fine and coarse root biomass, branch biomass, and foliage biomass) over the 11 yr of stand development based on total height and DBH measurements. Carbon was estimated as 50% of biomass. The model estimates foliar biomass based on annual stemwood increment calculated from the inventory data, essentially running backward to estimated the amount of leaf area needed to grow the wood. A key assumption of the model is growth efficiency (stemwood production/leaf area), which we assumed did not vary between the sites. Though this assumption may be invalid, it is not critical for our purposes as it only affects the estimates of foliage and fine root mass, which are small tree biomass components in relation to the others in an 11-yr-old plantation.
Soil Temperature
Soil temperature was monitored on an hourly basis for a year (April 2001–April 2002) at an 18-cm depth (approximately 3 cm below the A horizon of at each site) in the soil profile using a data logger (Temperature Data Logger–Model StowAway Tidbit XT, Onset Computer Corp., Bourne, MA) for recording temperature. One probe was installed in one of the replicated control plots within each site. In addition to the continuous measurements, temperature was also measured at the same soil depth in each plot on a monthly basis. The continuous data were used to adjust the monthly measurements to a common hour.
Statistical Analysis
The study was analyzed as a randomized complete block design with replicates nested within sites. Mean values from the four replicated plots within treatments within each site were evaluated for comparisons between treatments and sites after establishing no significant block differences.
An analysis of variance model (ANOVA) was used to evaluate the statistical significance of various effects (i.e., site, fertilization, interactions, etc.) on specific response variables. Effects were considered significant at the 0.10 probability level. The general linear model (GLM) procedure within SAS statistical software (SAS Institute, 2001) software was used for all analysis and mean estimates.
The site effect was evaluated using Type III mean sum of squares (MS) for Rep(Site) as an error term for all models. The treatment effect was referred to as fertilization since there was only one level of fertilization evaluated in this study. The fertilization and site x fertilization effect were evaluated using Type III MS for site x rep x fertilization as an error term for some models.
Since each depth increment was related to the surrounding depth increment(s), it was evaluated as a numeric, continuous variable in the ANOVAs. This was true when evaluating depth interactions as well.
Since pretreatment soil samples (Year 0 samples) were only collected for the 0- to 10-cm and 10- to 20-cm depths, these were the only depths to be compared with soil samples collected 11 yr later (Year 11 samples) in the same plots. To evaluate site and treatment effects on soil C over the 11 yr of stand development, the difference (soil C Year 11 – soil C Year 0) between the means for soil C for each of the 16 plots for Year 11 and Year 0 were compared, and t tests were used to test if the within treatment change in soil C was significantly different from zero.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
Mineral Soil at Year 11
Carbon concentration and content were significantly different between sites across the 0- to 50-cm soil depths, with the depth effect varying by site (Table 3). Based on Year 11 data without respect to pretreatment samples, fertilization had a significant effect on C concentration and content. There were no statistically significant fertilization x site, fertilization x depth, or fertilization x site x depth interactions. No statistically significant effects of fertilization on C concentration or content were observed below the 0- to 10-cm layer (results not shown for depths >20 cm).
Averaged across both treatments, the clayey site contained 27.5 Mg ha–1 (62%) more C in the 0- to 50-cm of mineral soil compared with the sandy site (Table 4). On an absolute basis, the clayey site contained 22.6 Mg ha–1 (72%) more C in the surface 20 cm of mineral soil, with both sites having similar C levels below 20 cm. Both sites had the same distribution of C on a relative basis, with the surface 0- to 10-cm of mineral soil containing about 50% of the C content for the 0- to 50-cm layer. The surface 1 cm of mineral soil contained about 10% of the C content.
Averaged across both sites, the fertilized plots had 7.6 Mg ha–1 greater C in the surface 0- to 50-cm of mineral soil, representing a 14% increase over control plots (Table 4). However, this greater C in the surface 0- to 50-cm for the fertilized plots was only significant for the sandy site (P = 0.09). The average increase in the 0- to 10-cm depth was 6.0 Mg C ha–1, or about 80% of the total increase measured for the fertilized plots.
Change in Soil Carbon from Year 0 to 11
There were no statistically significant effects of site, fertilization, or the site x fertilization interaction on the change in soil C from Year 0 (pre-treatment) to Year 11 (posttreatment) for the surface 0- to 10-cm and 10- to 20-cm of mineral soil (result not shown). There was, however, a significant reduction in C in the 10- to 20-cm depth for the control plots on the sandy site, where C decreased by 12 Mg ha–1 (Table 5). Interestingly, though not statistically significant, the trend for change in C from Year 0 to 11 was negative for both sites, with the exception of the 0- to 10-cm depth for the fertilized plots on both sites.
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The large growth response to fertilization clearly indicates that both sites are nutrient deficient. In fact, the 5.9 m3 ha–1 yr–1 annualized growth gain with fertilization on the clayey site is consistent with growth responses for severely P-deficient sites (Allen et al., 2001). The smaller annualized growth response to fertilization of 2.9 m3 ha–1 yr–1 on the sandy site is probably a function of several limiting factors (e.g., low available water, N, and micronutrients), as repeated fertilization with a combination of macro- and micronutrients on excessively well drained sands has resulted in loblolly pine growth gains in excess of 10 m3 ha–1 yr–1 (Albaugh et al., 1998).
Neither site is achieving its C storage potential, as the fertilization effect on stand growth was probably nearly gone by Age 11. Evidence for this can be seen in the forest floor, where there was no fertilization effect on the mass of the Oi layer. The similarity in Oi layers is a good indication that control and fertilized plots are producing similar amounts of leaf area, which is also supported by visual observation of individual tree crowns in the field. Since leaf area is strongly related to current productivity (Albaugh et al., 1998), the trees on both control and fertilized plots are now probably growing at similar rates. To maintain high growth rates it is necessary to carry out repeated fertilizations (Allen et al., 2001). With repeated fertilizations in combination with other silvicultural tools, growth rates exceeding 25 m3 ha–1 yr–1 are possible (Allen et al., 2001).
When evaluating the Year 11 mineral soil C results without respect to the pretreatment (Year 0) samples, it appears that fertilization increased soil C storage by 7.6 Mg ha–1 (14%). Based on the increases in growth measured with fertilization and without evaluating the pretreatment samples, it would be convenient to state that the change in soil C is explained by increased C inputs with higher productivity.
When considering the Year 11 samples with respect to the pretreatment (Year 0) samples, the analysis of the change in soil C from Year 0 to 11 showed that there were no significant changes in soil C in the first 11 yr of stand development, with the exception of the control plots on the sandy site which actually lost 12 Mg C ha–1. This refutes the "fertilization effect" that was based solely on the Year 11 data.
Our results for mineral soil C are consistent with the findings of Harding and Jokela (1994), who reported no change in soil C 25 yr after fertilization at planting even though fertilized plots had 110% greater biomass accumulation. Canary et al. (2000) also found no significant changes in mineral soil C with multiple applications of urea over a 16-yr period in Douglas-fir (Pseudotsuga menziesii) stands that did exhibit significant increases in biomass. In their meta analysis of fertilization effects on soil C, Johnson and Curtis (2001) found that fertilization in general increased soil C, but some studies in the analysis showed no response.
The lack of change in mineral soil C with fertilization found in our study and reported by others may be explained by rapid decomposition of C inputs. Richter et al. (1999) found that of the 38 Mg C ha–1 estimated to have been input to the mineral soil over 40 yr of development of a loblolly pine plantation, only 1.45 Mg C ha–1 (4%) had accumulated in the mineral soil, demonstrating rapid turnover of C inputs. Rapid turnover of C inputs may explain why studies have found soil C to remain stable over long periods in forest ecosystems (e.g., Richter et al., 1995; Trettin et al., 1999). So, given the likelihood of rapid turnover of C inputs, the potential to increase the size of the mineral soil C pool with intensive management appears to be low.
An interesting result was the loss of mineral soil C from control plots on the sandy site. In fact, it was this loss that made it appear that fertilization increased mineral soil C at Year 11. The loss in mineral soil C is consistent with the model of change in soil C over time following harvest presented by Henderson (1995), wherein soil C declines for a period of time following harvest as C outputs via respiration exceed C inputs in young stands. The pattern of initial C loss followed by accumulation was shown empirically by Richter et al. (1999), who tracked soil C over 40 yr of stand development in a loblolly pine plantation. In their study, the surface 0- to 7.5-cm layer of mineral soil lost C for at least the first 6 yr, and it was 15 to 20 yr before mineral soil C was back at the stand establishment level. Why was significant C loss not observed above the 10- to 20-cm layer on the sandy site in our study? The 0- to 10-cm layer probably did show significant C loss early on, but this is also the zone where root activity and forest floor C transfer to mineral soil would be greatest, allowing more rapid recovery of C in this layer.
The significant loss of mineral soil C from Year 0 to 11 in control plots on the sandy site may be attributed to site and treatment effects on the loss and build up of soil C. Both sites were prepared by raking off most surface debris followed by disking. Since most debris was raked off of the sites, very little organic matter was incorporated into the mineral soil with disking. Thus, all plots would have experienced C loss from the mineral soil for some period following site preparation. However, soil C loss was probably lower on the clayey site owing to greater physical protection in the finer-textured soil, also soil C probably recovered more rapidly on the clayey site given its higher productivity (71% greater stand volume at Age 11 compared with the sandy site) and likely greater C inputs from root turnover and litterfall. The clayey site also had lower surface soil temperature, with the lower surface soil temperature being attributed to a combination of higher leaf area, greater forest floor mass, and probably higher soil moisture. The magnitude of difference in surface soil temperature in the earlier years was probably even greater than the 2°C difference measured between the sites during the summer in Year 11. Thus the combination of higher soil temperatures, lack of organic matter incorporation, and tillage on the sandy site would have slowed soil C recovery and accumulation on this site. Adding fertilizer on the sandy site helped to ameliorate these negative effects by promoting more rapid canopy closure and forest floor accumulation compared with the non-fertilized control plots. We don't have the temperature data from the early years to support this explanation, so the explanation offered is a hypothesis. However, the effect of fertilization on soil temperature and mineral soil C loss was documented for loblolly pine plantations growing on a similar soil (Maier and Kress, 2000), where fertilization resulted in lower soil temperature, correspondingly lower heterotrophic soil respiration, and positive net ecosystem productivity: non-fertilized plots had higher soil temperature, correspondingly higher heterotrophic soil respiration, and negative net ecosystem productivity.
Biomass is the most significant pool in terms of opportunity to increase C sequestration with fertilization. The majority of the aboveground pool will be exported from the site through thinning and final harvest, leaving the belowground pool behind. At Year 11, the belowground pool was estimated to range from 3.5 to 10.6 Mg C ha–1 depending on site and treatment. This pool will continue to accrue with time, as studies in older loblolly pine plantations have reported belowground C pools ranging from 12.8 to 17.0 Mg ha–1 (Johnson and Lindberg, 1992; Richter et al., 1995). Thus, the fate of the belowground C pool has important implications for longer term sequestration. In a chronosequence study of loblolly pine root systems, Ludovici et al. (2002) found that large roots persisted for 35 to 60 yr following harvest. With such low decomposition rates, and considering that plantations are managed on a 25-yr rotation, several cohorts of large roots could be present in the soil profile, contributing significantly to long term C sequestration.
|
ACKNOWLEDGMENTS |
|---|
Received for publication September 5, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. A. Khan, R. L. Mulvaney, T. R. Ellsworth, and C. W. Boast The Myth of Nitrogen Fertilization for Soil Carbon Sequestration J. Environ. Qual., October 24, 2007; 36(6): 1821 - 1832. [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 | |||
