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


     

Published online 2 December 2005
Published in Soil Sci Soc Am J 70:153-162 (2006)
DOI: 10.2136/sssaj2004.0173
© 2005 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
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 Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Joshi, A. B.
Right arrow Articles by Johnson, A. H.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Joshi, A. B.
Right arrow Articles by Johnson, A. H.
Agricola
Right arrow Articles by Joshi, A. B.
Right arrow Articles by Johnson, A. H.
Related Collections
Right arrow Watershed-Scale Studies
Right arrow Nutrient Cycling
Right arrow Nitrogen

Forest, Range & Wildlands Soils

Litter Quality and Climate Decouple Nitrogen Mineralization and Productivity in Chilean Temperate Rainforests

Amishi B. Joshi, David R. Vann and Arthur H. Johnson*

Dep. of Earth and Environmental Science, The Univ. of Pennsylvania. 240 S. 33rd St., Philadelphia, PA 19104

* Corresponding author (drvann{at}sas.upenn.edu)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUMMARY
 REFERENCES
 
We measured litter quality, N mineralization, air and soil temperatures, soil moisture, and aboveground net primary production (ANPP) at 18 plots in three forest types typical of full-stature forests found on western slopes of the Cordillera de Piuchué (CP) on Isla de Chiloé in southern Chile. The primary objectives were to examine the range of net N mineralization across forest types that have different vegetation and litter composition, assess probable controls on mineral N production and leaching, and to determine if the relationship between N mineralization and ANPP was similar to that observed in cool temperate northern hemisphere forests. Average annual litter lignin/N ratios were high in the evergreen montane broadleaf forest plots (50:1) and very high in the montane conifer plots (80:1). Net N mineralization during the summer months was correspondingly low (1.2 and 2.2 kg ha–1, respectively). The high litter lignin content and associated low rate of N mineralization can explain the extremely low concentrations of mineral N in soils and upland streams of this region. In the broadleaf evergreen coastal forest plots, lignin/N ratio was lower (23:1) and N mineralized was considerably greater over the same measurement period than in the montane forest plots (14 kg ha–1). In contrast to findings in many cool temperate northern hemisphere forests, ANNP in the broadleaf forest type was high (average 11.1 Mg ha–1 yr–1) in spite of the low net N mineralization rates, and ANPP and net N mineralized were not correlated. Net N mineralization was best correlated with litter quality, soil temperature, and soil moisture content, and ANPP was best correlated with growing degree days.

Abbreviations: ANPP, aboveground primary production • CB, coastal mixed broadleaf • CP, Cordillera de Piuchué • DBH, diameter-at-breast-height (1.4 m) • GSDD, growing season degree days • MB, montane mixed broadleaf • MC, montane conifer forests • PMA, phenylmercuric acetate


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUMMARY
 REFERENCES
 
IN THE CP of Isla de Chiloé in southern Chile, thin soils, steep slopes, a cool, wet climate and difficult access have discouraged logging and agriculture, preserving old-growth (>550 yr) forests. In addition to the lack of direct human impact on these forests, atmospheric deposition of anthropogenic H+, N, and S are very low (Levy and Moxim, 1989). Low deposition rates of mineral N in southern Chile prompted comparisons of mineral N in CP stream water draining old-growth forests with that in similar forests in the northern hemisphere, which receive substantial amounts of atmospheric N deposition. The comparisons done by Hedin et al. (1995) showed lower inorganic N concentrations in stream water draining CP broadleaf and conifer forests than measured in streams draining any temperate old-growth forests in North America.

Low mineral N loss from forests is typically associated with very efficient N cycling and/or a lack of mineral N production and deposition. Like North American conifer forests that conserve mineral N, the broadleaf and conifer CP forests are evergreen with long-lived foliage. In the montane broadleaf CP forests, canopy turnover time is approximately 3 to 6 yr and in the montane conifer forests, canopy turnover time is approximately 10 to 15 yr (Vann et al., 2002). Long-lived foliage limits the annual demand for soil N so given low mineral N levels in soil and stream water, we hypothesized that N mineralization rates would be low and would, at least in part, account for the very low export of dissolved mineral N. We expected that foliar quality would be consistent with low N mineralization rates (e.g., high lignin/N values) and thus serve as a major control on mineral N production, availability to plants, and leaching losses.

In light of the potential for very low rates of mineralization, we also sought to examine the relationship between N mineralization and ANPP. Vann et al. (2002) reported substantial rates of aboveground biomass production in CP conifer and broadleaf forests (about 7 and 9 Mg ha–1 yr–1, respectively), with a corresponding demand for N from the soil of 21 to 49 kg ha–1 yr–1. Mineral N availability has been judged to be the most likely limitation to temperate forest productivity (e.g., Mitchell and Chandler, 1939; Vitousek and Howarth, 1991; Reich et al., 1997). In support of this, several temperate-region studies have shown a positive correlation between ANPP and N mineralization (Pastor et al., 1984; Nadelhoffer et al., 1985; Zak et al., 1989; Reich et al., 1997). Thus we wished to see if moderate-to-high ANPP levels were realized in the CP in spite of what we expected to be low N mineralization rates. For comparison, the data of Reich et al. (1997) indicate that ANPP in the 7 to 9 Mg ha–1yr–1 range in cool temperate forests of the North Central region of the USA is associated with measured N mineralization of 50 to 100 kg ha–1 yr–1.

We measured the rate of litterfall and seasonal variations in its composition and net N mineralization during the growing season in three undisturbed, cool temperate perhumid forests of the CP: a conifer forest and two different mixed-species broadleaf evergreen forests. We measured and evaluated litter N and lignin content and climate variables likely to influence net mineralization and ANPP and finally, we compare the relationship between N mineralization and ANPP with data from North American forests.

Site Description
Isla de Chiloé is located off the coast of southern Chile. Low mountains of the CP run the length of the island's west coast; the highest peaks reach approximately 800 m above sea level. Small streams are common in valleys and wetlands are common on flat watershed divides as well as adjacent to streams. Our research examined N dynamics in conifer and mixed-species broadleaf evergreen forests on the slopes of the CP and mixed-species broadleaf evergreen forests found at lower elevations (approximately 50 m a.s.l.) along the Pacific coast (Armesto et al., 1995). The study area contains no known exotic plant species (Hedin et al., 1995) and lies within and adjacent to the Parque National de Chiloé, a protected area of approximately 43 000 ha of old-growth forests. There is no history of commercial timber exploitation or agricultural clearing of these forests; human disturbance has been limited to some bark gathering and localized fires along access trails, and occasional removal of dead alerce trees (Fitzroya cupressoides Mol. Johnst.) (Hedin et al., 1995). Known natural disturbance agents in our study area include wind storms, icing events, and fire. Large-scale stand-leveling disturbances occur very infrequently and the most common type of disturbance is multiple tree falls that produce small gaps (Armesto and Fuentes, 1988; Donoso et al., 1993; Armesto et al., 1995). In the forests we studied previously, some individuals of F. cupressoides exceed 550 yr in age (Battles et al., 2002).

The study area is characterized by a cool, wet temperate climate with strong oceanic influence (Pérez et al., 1998). Weak seasonality is expressed as rainy and cool winters from June to August and warmer, less humid summers from December to March (Pérez et al., 1998). Mean annual precipitation measured approximately 650 m a.s.l. is about 5500 mm, with 10 to 15% falling during summer (Pérez et al., 1991, 1998). Average temperatures for the CP range from 8 to 11°C in the summer months to 3 to 6°C in the winter (Pérez et al., 1991, 1998, D.R. Vann, unpublished data, 1997–2000). Temperature extremes are moderated in the coastal forests due to their proximity to the Pacific and frequent fog. Dominant winds are westerly.

The bedrock of Isla de Chiloé is Precambrian schist (Pérez et al., 1991; Lopez-Escobar et al., 1995). Low levels of silica in drainage waters indicate that weathering of the bedrock does not play a significant role in the mobilization of nutrients in these mountains (Hedin and Hetherington, 1996). Available P and exchangeable base cations Ca, Mg, and K are strongly related to soil organic matter content (Zarin et al., 1998; Thomas et al., 1999). Pollen records show that montane ecosystems on the unglaciated western slopes of the CP have remained floristically stable throughout the Holocene (Heusser and Flint, 1977; Heusser, 1983).

Soils in the study forests fit the characteristics of Inceptisols as defined by the USDA Natural Resources Conservation Service (Joshi, 2001). In CP conifer forests, occasional sites have organic horizons sufficiently deep to qualify as Folists (A. Joshi, unpublished data, 1999). Montane soils typically develop on relatively steep slopes (30–60°), with average depth to bedrock ranging from 40 cm in the conifer forests to about 75 cm in the broadleaf forests. The soils of the conifer forests have particularly low bulk densities as they are principally organic mats (Zarin et al., 1998).

Soil chemical properties for broadleaf and conifer forests in the vicinity of our study sites have been tabulated (Zarin et al., 1998; Joshi, 2001) and Vann et al. (2002) and Battles et al. (2002) determined aboveground biomass and nutrient pool sizes and flux rates. Temperature and vegetation differences between the study forests produce some variation in percentage of soil C and pH, with conifer forests having more soil C and broadleaf forests having slightly warmer and less acidic soils. In general, the mineral horizons are classified texturally as loams and are well drained (Joshi, 2001). These forests have high basal area and biomass (up to 138 m2 ha–1 and 656 Mg ha–1, respectively, Battles et al., 2002) and quantities of N, P, K, Ca, and Mg that are similar in magnitude and distribution to those of high biomass forests of cool temperate regions of the northern hemisphere (e.g., Johnson and Lindberg 1992, DeAngelis et al., 1981). In the CP, steep slopes combined with high soil porosity result in substantial runoff; soils are saturated only after rains, remaining almost continuously aerobic (W.L. Silver and D.R. Vann, unpublished data, 1997).


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 SUMMARY
 REFERENCES
 
The study area is located along a west–east transect in the CP, from a low-elevation coastal location to a higher elevation inland location. The elevational gradient of approximately 600 m and the frequent fogs at the coastal site provide differences in climate. Three study sites were established at three elevations and in three different forest types: coastal mixed broadleaf (CB), montane mixed broadleaf (MB), and montane conifer forests (MC). Dominant species and other characteristics of these sites are summarized in Table 1.


View this table:
[in this window]
[in a new window]
  		Table 1. Summary of selected Isla de Chiloé study site characteristics.

 
At each site, six 100- or 50-m2 plots were established; all sampling took place within these plots (n = 18). At the CB site (0% slope), 10-m square plots were used; at the MB site (40–50% slope), rectangular plots (5 by 20 m) were used with the long axis parallel to the fall line. Upper-elevation conifer watersheds are small in areal extent and contain F. cupressoides, a protected species (Anonymous, 1977; U.S. Fish & Wildlife Service, 1979). Due to high stem density and to minimize disturbance, the MC site was sampled using smaller 5 by 10 m plots. The slopes of MC plots were about 5%. At each of the three sites, electronic data loggers placed in three of the plots recorded hourly air temperature at canopy level and soil temperature at the 5-cm depth. Soil water content measured via frequency-domain reflectometry (ThetaProbe mL-1, Delta-T instruments, Cambridge, UK) was also recorded electronically at noon and midnight.

At each plot, all trees were identified and diameter-at-breast-height (DBH = 1.4 m) was measured for all live and dead standing individuals greater than 1.4 m in height. Height was measured for the 5 to 7 tallest and/or largest-diameter trees in each plot. These trees were then cored through the center using a 4.5-mm increment borer. An estimate of the maximum age of the plot was then obtained by counting the rings in these cores. Smaller cores (about 3 cm long) were collected from a minimum of 10 and a maximum of all small canopy and/or subcanopy trees. Using all tree cores, the widths of the outermost ten rings were measured and used to calculate average annual diameter-growth rate.

We measured the following variables: total annual litterfall, C, N and lignin content of the litter and soil as well as monthly net N mineralization rate for the summer (December-March) months. To estimate litterfall, litter from four to six randomly located litter traps was collected and combined into one sample at each plot per sampling period. The combined litter collectors sampled approximately 2% of the plot area. Samples were collected monthly during the growing season (November 1997 to April 1998) and four times during winter 1998 (June, August, November, December). Samples were oven-dried at 80°C to constant weight and ground in a Wiley Mill to pass a 1.0-mm mesh and analyzed for C, N, and lignin.

Six randomly located soil core samples were collected monthly (sampling detailed below) to determine total C and N levels, soil water and dry matter content and bulk density in each plot during the 1997–1998 growing season. The samples were weighed fresh in the field and subsequently air-dried. An equal percentage of soil was then subsampled from each core in a plot and combined into a single sample, so as to form one soil sample per plot per incubation period. These subsamples were then pulverized using a mechanical mortar and pestle for C and N analysis. Soil samples from the beginning, middle, and end of the growing season were also analyzed for lignin content.

An elemental analyzer and standard procedures (Carlo-Erba NA 1500 C/N Analyzer, Fisons Instruments, Beverly, MA) were used to determine C and N for both soil and litter samples. Lignin content was measured at the University of Maine using a variant of the acid (H2SO4) detergent fiber analysis (Van Soest, 1979) with further H2SO4 digestion (B. Hoskins, U. Maine, personal communication, 1999).

Net N mineralization was measured from November 1997 to April 1998, but critical sets of samples were lost during air transit, compromising that data set. The N mineralization study and the temperature and moisture measurements were repeated from December 2000 through April 2001 and we report those data in this paper. While this resulted in a temporal decoupling of the litterfall and N mineralization data, there is no reason to believe that litter quality (the key variable in this study) was substantially different between years. The data we do have for the earlier mineralization study shows the same extractable N values as obtained in the latter period, and the relationships between net N mineralization and litter quality for the three forests are consistent with North American values summarized by Scott and Binkley (1997). As shown below, the important conclusions we make would not be different even if there were substantial (e.g., twofold) differences in N mineralization between years principally because it is so low in two of the forests. N mineralization in the top 10 cm of soil (this typically included both O and A soil horizons) at each site was estimated using the in situ resin core method (after DiStefano and Gholz, 1986). We used PETG plastic tubes (Visipak, Inc., Arnold, MO) 10 cm long and 5 cm in diameter. Ion-exchange resin bags (Na+–saturated cation and Cl––saturated anion exchange resins) were placed at the bottom of the cores to trap NH4 and NO3 leaching from the soil core.

At the beginning of each incubation period, in each plot, six pairs of tubes were installed at random locations, avoiding trees, large roots and rocks. Loose surface forest floor material was set aside and tubes were pressed into the soil. For each pair, one core was immediately removed from the ground and the soil collected to determine initial (extractable) inorganic N levels as well as the soil variables described above. The second core was gently removed, ion-exchange resin bags were placed at the bottom of the core, and the core was re-inserted into the ground. Forest floor material was replaced onto the top of the core. Each incubation period was approximately 4 wk long; there were a total of four incubation periods, December–March. These are the months during which Pérez et al. (1998) measured positive rates of net N mineralization in nearby forests. At the end of each period, we removed the incubated cores from the ground and collected the soil and ion-exchange resin bags.

All soil samples were processed in a field laboratory within 2 to 6 h of collection. Soil was passed through a 2-mm mesh sieve to remove rocks and coarse organic matter and to homogenize the soil. An 8- to 12-g (fresh weight) subsample of the sieved soil was extracted with 50 mL of 2 M KCl and filtered. One milliliter of 100 ppm phenylmercuric acetate (PMA) was added to each extract as a preservative. Extracts were shipped on ice to the USA for analysis at a University of Pennsylvania laboratory in Philadelphia, PA (the trip usually took 4 d). In Philadelphia, extracts were kept refrigerated and analyzed within 2 wk. The remainder of the sieved soil was used for the soil analyses described above.

Storage and handling of cation- and anion-exchange resin bags were similar to that used for soil samples. Resin bags were shipped to the USA along with the soil extracts. At the University of Pennsylvania, each resin bag was extracted with 30 mL of 2 M KCl; resin extracts were centrifuged (to remove escaped resin and residue soil) and refrigerated until analysis. Preservative (1 mL PMA) was added to each resin extract. These extraction procedures have been shown to recover essentially all of the ammonium and nitrate from soils and at least 80 to 85% of the ions from the resin bags (as determined from recovery of standard additions to resins by Binkley et al. [1994] and confirmed in our lab).

The NH4 and NO3 concentrations of all extracts were determined colorimetrically using a Technicon Auto Analyzer II (Tarrytown, NY) at the University of Pennsylvania using standard procedures (Technicon Industrial Systems, 1977, 1973).

Soil core results are reported in kilograms of inorganic N per hectare of surface to10-cm depth soil for both initial extractable and net production values for NH4 and NO3; resin bag results are reported as kilograms of NH4–N or NO3–N leached per hectare surface to 10-cm depth soil. Net N mineralization was calculated on a plot-by-plot basis as the average post-incubation quantity of NH4–N and NO3–N in both the soil and resin bags, minus the initial quantity of NH4–N and NO3–N in the pre-incubation soils. A negative difference between post-incubation and pre-incubation N concentrations indicates net N immobilization. It should be noted that gross N mineralization was not measured, as microbial immobilization of NH4 and NO3 within the core over the incubation period was not determined.

To examine relationships among climate, net N mineralization, and productivity, three derived variables were used. The annual aboveground wood production was estimated using our measured 10-yr average increment growth and allometric equations constructed for the major tree species in these forests (Vann et al., 1998; Johnson, 1999). Annual ANPP is the sum of wood production and the total annual litterfall. Mean daily air temperatures from December to April were combined into one index value, growing season degree days (GSDD), equal to the mean daily air temperature multiplied by the number of days during the growing season that the mean is >0°C. We note that the growing season may be longer at the coastal forest than at the higher elevations, but air and soil temperatures were not measured outside of December through April.

Statistical Techniques
Pearson product moment correlation analysis was used to explore correlations among variables. Linear regression analysis was used to test hypotheses concerning variables that could explain differences in N mineralization and productivity across the three forests. Stepwise multiple linear regression was used to examine the relative contribution of GSDD and net N mineralization to ANPP and its components as well as the effect of soil and litter parameters on net N mineralization. Note that the structure of analysis focuses on a climate gradient; we do not replicate vegetation type. As such, results for individual forest types are not reliably extendable to other individual sites, although we expect the overall pattern driven by climate to be consistent across the region.

Some analyses were performed on a plot by plot (n = 18) basis to assess small-scale effects at each sampling location. Certain variables were log10–transformed when needed to pass normality tests for regression analyses; when transformation was necessary, it is specified in the results.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 SUMMARY
 REFERENCES
 
Litterfall Quantity and Chemistry
With the exception of the first litter collection period in early spring, patterns in litterfall quantity were similar in the two broadleaf forests (Fig. 1A ), though absolute amounts of litterfall were higher in the MB than in the CB. All three forests shed their maximum amount of litter in autumn (April–June). The percentage of productivity represented by litterfall quantity was significantly higher in the MB compared with both the MC (Tukey pairwise multiple comparison test, p < 0.001) and the CB (Tukey Test, p = 0.002).



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 1. Annual pattern in litterfall: (A) quantity and (B) lignin/N ratio over a 1-yr period in three forests of Isla de Chiloé, Chile. Standard error bars are shown (n = 6 plots).

 
Changes in litter lignin/N ratios over the growing season and winter are shown in Fig. 1B. While the MB produced litter with a higher lignin/N ratio than did the CB, the pattern of change (or lack thereof) of lignin/N ratios was identical in both broadleaf forests. The conifer forest, on the other hand, produced litter with the highest lignin/N ratio of the three forests and also showed a strong seasonal pattern in litter quality, producing litter with the highest lignin/N ratio in late summer.

Soil Variables
Values for the soil variables measured are shown Table 2. We saw no changes in these values over the course of the measurement period, so the seasonal average is reported. These montane soils are nearly entirely organic matter, and have very low bulk densities. Similar bulk densities were reported by Zarin et al. (1998). Soil N does not appear to be related to soil C content. Soil lignin content was highest in the conifer-dominated watershed, consistent with the high lignin values in the litter input. Soil C was dominated by lignin residue, implying very little C available for microbial growth.


View this table:
[in this window]
[in a new window]
  		Table 2. Soil variables (0- to 10-cm depth) measured for three CP forests. Values are seasonal mean ± s.e. (maximum/minimum) and are on a soil dry weight basis.

 
Nitrogen Mineralization
Table 3 shows mean values of the productivity, net N mineralization and temperature data over the December through April period. We restricted our measurements to these months because they were the months that had positive N mineralization values in the study of Pérez et al. (1998). The MB and MC forests showed very low net N mineralization rates, 2.2 ± 0.8 kg N ha–1 (mean ± s.e.) and 1.2 ± 0.2 kg N ha–1 respectively; the CB forest mineralized a total of 14.2 ± 2.4 kg N ha–1. Seasonal patterns in inorganic N concentrations are shown in Fig. 2 and varied between forests. In particular, the relatively high rates (about 4–5 kg ha–1 mo–1) at the beginning and end of the sampling period in the coastal forest suggest that mineralization in excess of immobilization probably occurred both prior and subsequent to the period of measurement.


View this table:
[in this window]
[in a new window]
  		Table 3. Site variables determined in this study for three evergreen forests on Isla de Chiloé, Chile. Water content, mineralization, and nitrification values based on the top 10-cm soil. Values are mean ± s.e. or mean (seasonal maximum/minimum).{dagger}

 


View larger version (22K):
[in this window]
[in a new window]
  		Fig. 2. Change in soil inorganic N (NO3 and NH4) concentrations over the growing season in three forests on Isla de Chiloé, Chile, as measured by the in situ resin core method of estimating N mineralization. "Pre-incubation" values represent inorganic N in the soil just before installing incubations; "post-incubation" values represent the amount of inorganic N accumulated and leached from the incubation after approximately 4 wk. Standard error bars are shown. The difference between pre- and post-incubation concentrations represents the amount of N mineralized (positive difference) or immobilized (negative difference) during that period.

 
Nitrate production was very low in the montane forests (MB and MC); the maximum nitrification (10% of average total net mineralization) occurred in the MC between February and March (data not shown). In the CB, there was considerable nitrification; nitrate production remained fairly stable (range of averages for each period 1.3–1.6 kg ha–1) over the growing season after an initial (Dec to Jan) maximum, which yielded an average of 2.2 kg ha–1. As a percentage of mineralization, nitrification in the CB varied, ranging from 36 to 75%.

We evaluated the importance of our measured environmental variables and the soil and litter quality parameters as factors potentially regulating N mineralization and nitrification rates. The coefficients of determination and probabilities for these tests are shown in Table 4. Substrate quality variables (soil C/N and litter lignin/N ratios) explained more variation in N mineralization (R2 = 0.733 and 0. 621 respectively, p < 0.001) and nitrification (R2 = 0.605, 0.502 respectively, p < 0.001) than did any other variable. Some variation in net N mineralization could be attributed to soil temperature (R2 = 0.578) and moisture (R2 = 0.421), with decreasing net N mineralization in wetter (and cooler) soils. These soil temperature and litter quality parameters showed similar patterns of determination with nitrification, but with slightly lower, though still significant R2 values. Neither total litterfall nor litterfall N content explained a significant fraction of the variance in log10–transformed N mineralization (p = 0.975 and p = 0.170 respectively) or in log10–transformed nitrification (p = 0.750 and p = 0.095 respectively). No combination of factors in multiple regressions improved relations with net N mineralized or nitrified.


View this table:
[in this window]
[in a new window]
  		Table 4. Coefficients of determination and significance values from linear regressions between N mineralization and possible driving variables (based on annual values for 18 plots).GSDD, growing season degree days; SOM = soil organic matter. Litter lignin/N values were log-transformed before analysis.

 
Productivity
Aboveground net primary productivity in the broadleaf (MB and CB) forests was 12.2 and 10.0 Mg ha–1 yr–1 respectively; ANPP in the conifer (MC) forest was lower, 4.6 Mg ha–1 yr–1 (Table 5). In regression analyses, GSDD explained most of the variance in ANPP (R2 = 0.874, p < 0.001; Fig. 3A ). Residuals from the ANPP–GSDD relationship showed no significant relationship with any variable, including net N mineralization (Fig. 3B) or soil moisture (Fig. 3C). Aboveground net primary productivity also had a significant positive relationship with soil GSDD (R2 = 0.351, p < 0.001). In linear regressions, ANPP did not appear to be driven by net N mineralization (p > 0.38). Aboveground net primary productivity showed a negative relationship with soil moisture (R2 = 0.588, p < 0.001); drier soils were also warmer.


View this table:
[in this window]
[in a new window]
  		Table 5. Annual aboveground net primary productivity (ANPP) for forests located in different watersheds on Isla de Chiloé, Chile. The montane conifer forest (MC), montane mixed broadleaf (MB) and coastal mixed broadleaf (CB) were forests examined in this study (see Table 1) while the CPES-C and CPES-B were forests examined by Vann et al. (2002) in similar but unattached watersheds on Isla de Chiloé, Chile. CPES-C is a conifer forest, analogous to the MC in this study while CPES-B is a mixed broadleaf forest, analogous to the MB and CB in this study.

 


View larger version (24K):
[in this window]
[in a new window]
  		Fig. 3. (A) Comparison of aboveground net primary production (ANPP) on an annual basis with growing season degree days (GSDD). Line is a linear regression, R2 = 0.874, p < 0.001. (B) Regression of residuals from ANPP/GSDD relationship with net N mineralization (R2 = 0.010, p > 0.1) and (C) soil moisture (R2 = 0.010, p > 0.1).

 
Analyzing the components of ANPP (litterfall and wood production) separately, GSDD explains most of the variation in litterfall (R2 = 0.860, p < 0.001); addition of net N mineralization explains an additional approximately 6% of the variation (multiple R2 = 0.924, p < 0.01). Growing season degree day explains some 49% of the variation in woody growth (R2 = 0.487, p < 0.001). Soil GSDD explains a similar fraction of the variance in woody growth (R2 = 0.567, p < 0.001), whereas litterfall quantity was not predicted by soil GSDD. Wood increment was also the only measure of productivity that varied significantly with net N mineralization (R2 = 0.285, p < 0.05); litterfall quantity and net N mineralization were unrelated (R2 = 0, p = 0.975).


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
SUMMARY
 REFERENCES
 
Nitrogen Mineralization Rates and Controls
In this study, the two montane forests (MC and MB) showed very low rates of net N mineralization during the measurement period while the coastal forest (CB) was considerably higher. These findings contrast with earlier published values in Pérez et al. (1998), who reported annual mean net N mineralization values of about 20 and 35 kg ha–1 yr–1 in nearby conifer and broadleaf forests, respectively. Additionally, they found a higher proportion of the mineralized pool to be nitrate; 54 (±23)% in the conifer forest and, in the broadleaf forest, 62 (±24)%. The differences between the studies might be attributed to the following:

(1.) Pérez et al. (1998) used the buried bag incubation technique (Eno, 1960). While there have been studies showing that both incubation techniques yield similar estimates of N mineralization (Binkley et al., 1992; Zou et al., 1992; Subler et al., 1995), some research indicates that differences in these techniques may become more important in areas of fluctuating soil moisture (Binkley and Hart, 1989; Hart and Firestone, 1989; Zou et al., 1992; Gonçalves and Carlyle, 1994; Rovira and Vallejo, 1997). (2.) A major source of the difference is related to the soil bulk density values used to estimate per-hectare N mineralization rates. The Pérez et al. (1998) study used a single value, 100 Mg ha–1 (averaged from the total soil profile to a depth of 40 cm), to calculate values for both conifer and broadleaf forest soils. We used bulk density values determined for each soil sample. There was a large difference in bulk density between the forest types (Table 2). Recalculating the Pérez et al. (1998) rates with the lower bulk densities from Table 2 reduces the estimates substantially, though they remain higher than seen in this study. (3.) Pérez et al. (1998) transported unrefrigerated soil samples to Santiago, Chile for processing and analysis. More recent measurements by C. Pérez (personal communication, 2003) in the same watersheds show substantially lower rates, very close to those we report here.

Our nitrification results are similar to those of Perakis and Hedin (2001), who measured gross rates of production and consumption of NH4 and NO3 in a nearby CP montane mixed broadleaf evergreen forest using 15N pool dilution techniques. These authors found that NO3 production accounted for less than 10% of overall inorganic N production, a finding consistent with the results of this study.

While our estimates of net N mineralization in the CP forests are very low, they are consistent with the data of Scott and Binkley (1997). Figure 4 shows net N mineralization vs. litter lignin/N data compiled by Scott and Binkley (1997) along with the CP data. The Scott and Binkley (1997) data represent forests in a similar montane, cool-temperate (summer high temperature < 20°C) climate. This comparison indicates that the very low net N mineralization values reported here are appropriate for the litter quality available to the biota in the cool CP forest soils.



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 4. The relationship between net N mineralization and litter lignin/N ratio in broadleaf and needleleaf forests in the Cordillera de Piuchué and cool temperate (summer month high temperature < 20°C) North American sites from Scott and Binkley (1997) (R2 = 0.55).

 
It is also noteworthy that average litter lignin/N values for CP conifer forest plots are at least as high as those at any North American site, and that the values for the montane broadleaf CP plots (MB) fall in the upper half of the range of the North American conifer forests tabulated by Scott and Binkley (1997) and Moore et al. (1999). The lignin/N values reported here are similar to those of Patagonian conifers across the Andes in Argentina (Satti et al., 2003). The extremely high ratios seen during the growing season (Fig. 1) were probably due in part to increased N withdrawal rates before abscission throughout this period (about 60–65 vs. 50–55%). High foliar lignin content is likely related to the long canopy retention time in these forests. We found no relationship between total litterfall N and net N mineralization (p = 0.26), similar to findings summarized by Scott and Binkley (1997).

Rates of net N mineralization are also influenced by abiotic factors such as soil temperature and moisture, primarily through their influence on rates of microbial activity and chemical reactions in the soil (e.g., Moore et al., 1999). The positive relationship between net N mineralization rate and soil temperature we observed is consistent with many other studies (Matson and Boone, 1984; Powers, 1990; Boone, 1992; Kim et al., 1995; Stottlemyer et al., 1995; Sveinbjornsson et al., 1995; Reich et al., 1997). In contrast, Decker and Boerner (2003) reported that they found no climate effect in an elevational series of Nothofagus forests where species effects on litter dominated mineralization rates, however, their elevational gradient was about one-half of that of this study.

Although it was generally a small fraction of the total mineralization, net nitrification remained significantly correlated with net N mineralization; in addition, those factors that explained the most variance in net N mineralization also explained significant variance in net nitrification, indicating that the same variables probably govern both processes. Noteworthy nitrate production only occurred in the CB. Nitrification has been shown to be positively related to pH of the soil solution (Keeney, 1980; Tietema et al., 1992). Previous research at the same sites indicates CB soils have a higher pHH2O (4.9) than do soils of the MB or MC (4.5 or 4.3 respectively: Joshi, 2001). Additionally, across a wide range of North American forest/soil types, Ollinger et al. (2002) found significant nitrification to be present only when soil C/N ratio was below 22. In this study, only the CB soils were below this threshold (Table 2).

Overall, the mineralization of litter in these forests appears to be governed by the same factors observed in a wide spectrum of forests with low net N mineralization rates resulting primarily from high-lignin foliage and cool temperatures. It is likely that these are the main controls governing the very low concentrations of inorganic N in small streams of the region and the dominance of dissolved organic N over dissolved mineral N noted by Perakis and Hedin (2002) and Hedin et al. (1995).

Nitrogen Mineralization and Aboveground Primary Production
While N mineralization measured by the existing techniques provide indices of mineral N availability rather than absolute amounts, the difference between the rates we measured in the MB and MC and the annual N requirement are much larger than in comparable studies performed elsewhere. The annual difference between net N mineralization and aboveground N demand is >12 kg N ha–1 in the MC and about 50 kg N ha–1 in the CB and MB forests (see Table 5). The differences are sufficiently large that we believe plants are either obtaining N via routes other than N mineralization (through mycorrhizae, for instance), or else they are very effective competitors for N, rapidly obtaining most of the gross amount of N mineralized before microbes can obtain it. The data of Perakis and Hedin (2001) allow the latter possibility to some extent, but because they did not directly measure plant uptake of N, it must be inferred by difference. Figure 5 shows the CP ANPP vs. N mineralization values against a background of data from approximately 70 boreal and temperate forest plots in North America. The low net N mineralization rates combined with moderate to high productivity of the broadleaf CP forests contrast with northern hemisphere forests where mineral N availability and ANPP are correlated (Fig. 5).



View larger version (13K):
[in this window]
[in a new window]
  		Fig. 5. Relation between ANPP and N mineralization rate for three forests in southern Chile (black squares) and various North American forests. Open squares: Reich et al. (1997), Table 3. Closed circles: Joshi et al. (2003). Open circles: Van Cleve et al. (1983). Triangles: Pastor et al. (1984).

 

   SUMMARY
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUMMARY
REFERENCES
 
Litterfall lignin/N ratios in the montane CP coniferous forests are unusually high compared with North American coniferous forests, and similarly, the lignin/N ratio in the evergreen broadleaf-dominated montane CP forests are considerably higher than in temperate deciduous broadleaf forests studied in North America. Coupled with low temperatures and abundant water in the soil during the growing season, those biotic factors support only very low rates of net N mineralization. The coastal forest we studied has litter lignin/N ratios similar to that in broadleaf forests of the northern hemisphere, and mineralization rates are more comparable with those measured in North American broadleaf forests, especially considering the monthly patterns in Fig. 2, which indicate a high probability of N mineralization outside the summer months during which we measured N mineralization. The low rates of N mineralization in the montane forests are not surprising, but the relatively high productivity of these CP forests makes them distinctly different from North American forests where similar ANPP is associated with much higher N mineralization rates (Fig. 5). Unlike most studies in North America where N mineralization rates are at least somewhat similar to the annual N required from the soil, the disparity between mineral N production and N uptake in the montane CP forests is large enough to warrant careful study to determine how trees obtain their annual requirement of N.


   ACKNOWLEDGMENTS
 
We thank Dana Royer and Greg Brunkhorst for their long hours spent in soggy conditions hauling wet soil to the field lab. We gratefully acknowledge the Andrew Mellon Foundation for financial support. We thank C.O.N.A.F. Chile, and in particular park rangers José Nonque and José Ríos for their assistance in using the park.

Received for publication May 19, 2004.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 SUMMARY
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
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 Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Joshi, A. B.
Right arrow Articles by Johnson, A. H.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Joshi, A. B.
Right arrow Articles by Johnson, A. H.
Agricola
Right arrow Articles by Joshi, A. B.
Right arrow Articles by Johnson, A. H.
Related Collections
Right arrow Watershed-Scale Studies
Right arrow Nutrient Cycling
Right arrow Nitrogen

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Vadose Zone Journal Journal of Plant Registrations
Journal of Natural Resources
and Life Sciences Education		 Journal of
Environmental Quality