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


     

Published online 6 May 2005
Published in Soil Sci Soc Am J 69:906-914 (2005)
DOI: 10.2136/sssaj2004.0130
© 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) 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 Web of Science
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 Web of Science (16)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bautista-Cruz, A.
Right arrow Articles by del Castillo, R. F.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Bautista-Cruz, A.
Right arrow Articles by del Castillo, R. F.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Bautista-Cruz, A.
Right arrow Articles by del Castillo, R. F.
Related Collections
Right arrow Ecosystem Management
Right arrow Tropical Soils
Right arrow Biogeochemical Processes
Right arrow Tropical Soil Management
Right arrow Forest Soils

Forest, Range & Wildland Soils

Soil Changes During Secondary Succession in a Tropical Montane Cloud Forest Area

Angélica Bautista-Cruz* and Rafael F. del Castillo

Dep. de Recursos Naturales, CIIDIR IPN Oaxaca. Hornos 1003, Xoxocotlan 71230, Oaxaca, Mexico

* Corresponding author (mbautistac{at}ipn.mx)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 SITE DESCRIPTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The development of second-growth forest after abandonment of agricultural fields in tropical montane cloud forest (TMCF) areas is common in southern Mexico, but little is known about how such changes affect soil properties. We evaluated the changes in soil properties at the 0- to 20- and 20- to 40-cm depths during this conversion in three chronosequences, each composed of a cornfield, and adjacent forests of {approx}15, {approx}45, {approx}75, and ≥ 100 yr after abandonment at El Rincón, Oaxaca, Mexico. All studied soils were acidic, rich in organic C and total nitrogen (TN), and had low levels of plant-available P, exchangeable Ca, Mg, Na, and K, and high levels of exchangeable Al. Most of the soil properties analyzed changed significantly with the age after abandonment, but in most cases the patterns of response varied with the chronosequence, usually {approx}15 to 45 yr after abandonment. In all chronosequences, soil pH and N/P ratio decreased, and the thickness of the O horizon increased, during the first 100 yr of forest development. The highest rates of soil C sequestration and the highest drop in exchangeable K, Mg, and Ca concentrations took place the first 15 yr of forest development. Most of soil changes can be associated with nutrient retention by vegetation and litter, the concentration of exchangeable Al in soil, and the role of soil pH in mineralization rates, ion solubility, and rock weathering. The different patterns of response found among chronosequences illustrate the importance of having replicates before making general statements about changes in soil properties after disturbance.

Abbreviations: BS, base saturation • DB, soil bulk density • EA, exchangeable acidity • ECEC, effective cation-exchange capacity • SMBC, soil microbial biomass carbon • SOC, soil organic carbon • SOM, soil organic matter • TMCF, tropical montane cloud forest • TN, total nitrogen


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
SITE DESCRIPTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
AFTER A LAND IS CLEARED FOR AGRICULTURE or other purposes and later abandoned, ecosystems undergo a series of directional changes, a process called secondary succession. The most conspicuous of such changes are those related with vegetation composition and structure, but other ecosystem properties are also affected (Peet, 1992; Buol, 1994). Soil processes, by being closely correlated with plant changes, are expected to be associated with successional dynamics (Bever, 1994; Willis et al., 1997; Amiotti et al., 2000, Woods, 2000), and influence other important ecosystem processes such as primary productivity (Oren et al., 2001), and C sequestration (Post and Kwon, 2000). The latter is particularly relevant, as it reverses some of the effects responsible for soil organic carbon (SOC) losses when the land was converted to agricultural fields of annual crops in areas originally occupied by perennial vegetation (Post and Kwon, 2000).

Documenting the changes in soil properties during succession has been an active area of research in both temperate (Marrs, 1993; Compton and Boone, 2000; Knops and Tilman, 2000) and lowland tropical areas (Reiners et al., 1994; Hughes et al., 1999). Other ecosystems, however, remain little studied. One example is the TMCF. This kind of ecosystem has a persistent, frequent or seasonal cloud cover at vegetation level. Compared with lowland tropical rain forests, TMCFs usually occur at higher elevation, and the stature of trees is lower. Studies of soil properties in TMCFs are meager. No overriding soil chemical factor characterizes TMCFs (Bruijnzeel and Proctor, 1995). However, soils are usually acidic and, consequently, decomposition of soil organic matter (SOM; Hafkenscheid, 1997) and soil concentration of exchangeable bases and base saturation (BS) are usually low (Bruijnzeel and Proctor, 1995). Nevertheless, high contents of K and cation-exchange capacity have been reported in a TMCF of western Mexico (Meave et al., 1992). The TMCFs are among the most endangered ecosystems in the world due to land clearing mostly for agriculture or cattle raising (Churchill et al., 1995; Webster, 1995; Rzedowski, 1996; Aldrich and Hostettler, 2000; Bruijnzeel and Hamilton, 2000). After being used, the land may be abandoned due to human emigration, changes in economic activities, or the slash-and-burn method of cultivation, which requires a fallow period (Jolivet et al., 1997; Paustian et al., 1997; Knops and Tilman, 2000). In TMCFs of southern Mexico, the first forest to appear after abandonment is dominated by species of pine. If disturbance is low, pine trees are substituted by hardwood species (González-Espinosa et al., 1991; Sánchez-Velásquez and García-Moya, 1993; Cordova and del Castillo, 2001).

In other parts of the world, soil processes change widely when conifers are substituted by or replace other species of plants (Amiotti et al., 2000; Lilienfein et al., 2000). Substantial acidification, depletion of nutrients, and disruption of biogeochemical cycles are common soil changes associated with pine afforestation (Scholes and Nowicki, 1998). Therefore, important changes in soil properties are expected during secondary succession in areas originally occupied by TMCFs and replaced by pine-dominated forests. The few studies available on secondary succession in TMCF areas have shown important consequences to both plants and soils. The levels of foliar N, P, and K declined significantly in a successional gradient in Costa Rican TMCFs (Kappelle and Leal, 1996). Changes in soil pH and vegetation structure have been detected in a study of anthropogenic disturbance in TMCFs in Chiapas, Mexico (Ramírez-Marcial et al., 2001).

Land use has dramatically changed in areas originally occupied by TMCFs in El Rincón, Sierra Norte, Oaxaca, southern Mexico, during the last 100 yr as a result of factors such as shifting cultivation and emigration. Such changes have affected epiphyte abundance (Cordova and del Castillo, 2001), vegetation (Blanco, 2001), soil genesis, and mineralogy (Bautista et al., 2003). Umbric and cambic horizons were found at early successional stages, whereas folistic and sombric horizons were detected at late stages. Minerals, such as chlorite and muscovite, decreased during forest recovery, probably as a result of changes in soil pH (Bautista et al., 2003). Because of such changes, soil properties, including those associated with soil fertility and ecosystem processes such as C accumulation, are expected to experience important modifications during succession. More specific, as the stand ages, pH changes are expected to affect soil fertility, decomposition rates, and SOC sequestration. However, no study documenting such changes is available.

The objectives of this study were (a) to characterize the soils during the substitution of cornfields by secondary vegetation in TMCF areas in El Rincón, including their potential for C sequestration; and (b) to assess the differences in soil properties related with soil fertility during this recovery. To accomplish such goals, we relied on the chronosequence method (Jenny, 1941) by performing comparisons of local soils with the same parent material and topographic conditions but differing in age after abandonment. Studies using this indirect approach may be confounded with other factors besides time that affect soil properties, in particular those changing locally (Pickett, 1989). Of particular interest in this study is the fact that we were able to analyze three distinct chronosequences in the same region, composed of groups of similar age after abandonment, and therefore to consider in our analyses the importance of age after abandonment relative to that of other local sources of environmental variation.


   SITE DESCRIPTION
TOP
 ABSTRACT
 INTRODUCTION
 SITE DESCRIPTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The study area, El Rincón, is located between 17°19' to 17°23' N and 96°16' to 96°22' W, in Oaxaca, state of Mexico, in the Sierra Norte mountain range (Fig. 1). Here, elevation ranges from 1400 to 2300 m. We conducted our study at 1850 ± 150 m, at the windward side of the mountains, where TMCF is the primary vegetation. Topography is usually steep (15–64%). The mean annual precipitation at the nearest meteorological station ({approx}16 km from the study site) is 1719 mm yr–1, with a rainy season in summer and a dry season in winter. The mean annual temperature ranges between 20 and 22°C (Anonymous, 1999). The climate is temperate-humid to subhumid (Comisión Nacional de Biodiversidad, 2002). The soils underlain on a bedrock of schist of Mesozoic age (Castillo and Castro, 1996). Soil moisture regime is udic (sensu Van Wambeke, 1987). The soils from forested areas were classified as Inceptisols, and those of crop fields as Entisols (Bautista et al., 2003).



View larger version (30K):
[in this window]
[in a new window]
  		Fig. 1. Location of the three chronosequences studied: Juquila, Tanetze, and Yotao, at El Rincón, Oaxaca, Mexico. I = cornfield, II = incipient forest, III = young forest, IV = mature forest, and V = old-growth forest.

 
At the study site, the original vegetation is an upper TMCF (sensu Webster, 1995). Billia hyppocastanum Peyr, Cinnamomum zapatae Lorea-Hern, Oreopanax flaccidus Marchal, Podocarpus matudae Lundell, Quercus spp., Q. corrugata Hook., Q. laurina Humb. & Bonpl, Symplocos coccinea Humb. & Bonpl, and Ternstroemia oocarpa (Rose) Melchoir are among the most conspicuous species (Cordova and del Castillo, 2001). Part of the original forest has been cut down and burned for corn (Zea mays L.) cropping, in which Pteridium feei (Schaffner ex Fée) Faull is the most common weed along with several herbaceous species in the Asteraceae, Melastomataceae, Phytolacaceae, Poaceae, Rubiaceae, and Smilacaceae families. Cornfields are typically cultivated during each wet season. According to landholders, fertilizers were not applied to any of the cornfields sites sampled in this study. Corn cropping is typically short-lived, 3 to 5 yr, after which time a fallow period follows. As a result, cropping moves to another, usually adjacent, piece of land. Because of this system of shifting cultivation, cornfields can be considered as random samples of the studied area. Because of human emigration, and the introduction of coffee plantations at lower elevations, corn cropping was discontinued in certain areas. Thus, the landscape is a mosaic of crop fields and forests of different ages. We selected three chronosequences, Tanetze, Juquila and Yotao, to conduct our study. Each chronosequence consisted of an ordered series of five stands of differing age after abandonment, ranked I to V, with the same climate, parent material, and similar topography, including a cornfield (Stage I), and adjacent forests of {approx}15 (incipient forest, Stage II), {approx}45 (young forest, Stage III), {approx}75 (mature forest, Stage IV), and ≥100 (old-growth forest, Stage V) yr after abandonment (Fig. 1, Table 1).


View this table:
[in this window]
[in a new window]
  		Table 1. Description of the chronosequences of Tanetze, Juquila and Yotao, at El Rincón, Oaxaca, Mexico (after Blanco, 2001), showing age rank, approximate site age after abandonment, tree density, and basal area.

 
We based our ranking on (i) the estimated age of the shade-intolerant pioneer tree Pinus chiapensis (Martinez) Andresen, using tree cores; (ii) the floristic composition and vegetation structure of the stands, in particular, the abundance and size of tree species typical of primary TMCF (see below); and (iii) the opinion of the local people regarding the age of the stands. The three procedures gave the same ranking to the stands, but the exact age could not be determined. The age of P. chiapensis from tree cores gives an approximate estimate of the age of the stands, as tree establishment usually does not take place immediately after abandonment. Tree growth rates estimated from wood cores are usually consistent with the annual growth rates derived from an ongoing demographic study in the same area (unpublished data). Annual tree rings in P. chiapensis are likely explained by the uneven distribution of annual rainfall, as only 5% or less takes place during the winter months in the study area (Anonymous, 1999). Firewood extraction was not detected in any stage or chronosequence, except for Stage V of Tanetze.

The first forest to appear, approximately 15 yr after abandonment, is dominated by P. chiapensis. Other important species are: Clethra lanata M. Martens & Galeotti, Gaultheria acuminata Schldl. & Cham., Liquidambar macrophylla Oerst. (= Liquidambar styraciflua L.), and Phyllonoma laticuspis (Turcz.) Engl. At {approx}45 yr after abandonment, the stands reached the highest density and the stem area at 130-cm height, an estimate of tree biomass, began to decrease significantly. The most common species are Bejaria mexicana Benth., Clethra kenoyeri Lundell, P. laticuspis, and Vaccinium leucanthum Schldl. After this stage, self-thinning takes place, and hardwoods eventually replace pine trees. Thus, the previous species, and Persea americana Mill., Quercus spp., Rapanea spp., and Ternstroemia hemsleyi Hochr. are common in forests of {approx}75 yr old. Old-growth forests 100 yr old or older also have Beilschmiedia ovalis (Blake) C.K. Allen, Freziera spp., Osmanthus americanus (L.) Benth. & Hook. f. ex A. Gray, and Quetzalia occidentalis (Loes. ex Donn. Sm.) Lundell, and a species composition similar to adjacent primary TMCFs (Blanco, 2001; see also Cordova and del Castillo, 2001, for further details of the study site).


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 SITE DESCRIPTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Soil Sampling and Processing
The three chronosequences studied were treated as experimental blocks and 0.4-ha plots were installed within forests and cornfields of each block. We attempted to select soils on similar landforms, typical of the study area. The plots within blocks had no visual indications that they had been different from the forest sites before clearing. In addition, the proximity of the plots (<2 km) within chronosequences decreased the risks of having large variations of other factors within each chronosequence that may affect our results (Fig. 1).

Soil sampling followed the procedures described by Dick et al. (1996), Petersen and Calvin (1996), and Boone et al. (1999). At each stand, we chose a central point from which we set four more sampling points separated 20 m from the central point, at each cardinal direction, during the fall of 2000. After discarding the entire O horizon, soil was sampled at the 0- to 20- and 20- to 40-cm depth. Soil samples were kept in sealed plastic bags. Samples were air-dried and passed through a 2-mm screen before chemical and physical soil analysis. Fresh soil subsamples for soil microbial biomass carbon (SMBC) determination were stored at 4°C until analysis to minimize artifacts incurred during long-term storage. The SMBC and bulk density (DB) were determined from three replicates in each soil plot. Soil DB was estimated using a 47.6-mm-diam. soil auger.

Soil Analysis
Standard chemical and physical analysis of the soils followed the procedures described in the Soil Survey Laboratory Methods Manual (Soil Survey Staff, 1996). Soil organic C was analyzed by dry combustion in a Shimadzu TOC-5050A C analyzer in finely ground dry samples. Plant-available P was extracted by using the Bray and Kurtz solution; the extracts were analyzed for P by the chlorostannous-reduced molybdophosphoric blue color method in HCl. Total N was determined by the Kjeldahl method. Exchangeable Ca, Mg, K, and Na were extracted with 1 M ammonium acetate (pH 7) and quantified by atomic absorption spectrophotometry. Effective cation-exchange capacity (ECEC) was estimated from the sum of extracted base cations and exchangeable acidity (EA), and BS was estimated from the quotient of the sum of exchangeable base cations and ECEC. Exchangeable Al and EA were extracted with 1 M KCl using 0.05 M NaOH titration. Soil pH was measured using a 1:2 soil/water mixture and a digital pH meter. Soil DB was determined using the core method (Blake and Hartge, 1986). The thickness of the O horizon was measured with a metric tape. Soil microbial biomass C was measured by the fumigation-extraction method. Organic C rendered extractable to 0.5M K2SO4 by fumigation was determined with a Shimadzu TOC-5050A C analyzer. Finally, SMBC was calculated as the difference between extractable C in chloroform-fumigated and unfumigated samples using a correction factor Kc = 0.25 ± 0.05 (Vance et al., 1987; Voroney et al., 1993; Horwath and Paul, 1994). The amount of soil C accumulation per area at the 0- to 20- and 20- to 40-cm soil depths were calculated from our SOC estimates after adjusting for soil DB. Throughout soil analyses, at least 10% of the samples were randomly replicated for quality monitoring. The difference between replicates was usually <5% (>90% of samples) and never >10%. An internal soil reference sample was run along each set of determinations.

Statistical Analysis
General linear models, and the GLM procedure of SAS (v. 8.01, SAS Institute, 1990), were used to analyze the changes in the soil property y associated to the age of the stand a, locality (chronosequence) c, soil depth d, and interactions using the model y = ßo+ ß1a+ ß2c + ß3d + ß4bc + ß5ac + ß6ab + ß7abc + {epsilon}, where ßi is a fitted coefficient of the model, ßo is the general mean, and {epsilon} is the error term which includes all the other sources of variation not included in the model. The age of the stand was treated as an ordinal variable, as explained above, and was decomposed in linear and curvilinear (quadratic, cubic, and quartic) components using contrasts (Montgomery, 1985).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 SITE DESCRIPTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Soil Properties and Carbon Sequestration
Soils of both cornfields and forests were acidic, rich in SOC and TN; had low levels of plant-available P, exchangeable Ca, Mg, Na, and K, and high levels of exchangeable Al. All these properties, and ECEC, were significantly higher at the 0- to 20-cm than at the 20- to 40-cm soil depth within each site (P ≤ 0.01). The exchange complex was dominated by EA, mainly Al. The forest soils studied appear to be important reservoirs of C and, after adjusting for DB, the greatest rates of soil C sequestration per year took place during the first 15 yr after abandonment at both soil depths (429 g C m–2 yr–1, 0–20 cm; and 168 g C m–2 yr–1, 20–40 cm). Soil C decreased at a rate of –168 g C m–2 yr–1, from {approx}15 to 45 yr of forest development, and increased afterward at a rate of 54 g C m–2 yr–1 at the 0- to 20-cm soil depth in the three chronsequences. At the 20- to 40-cm, and after {approx}15 yr, soil C varied among chronosequences, but no significant changes associated with time after abandonment could be detected.

Changes in Soil Properties during Forest Recovery
Time after abandonment appears to be closely associated with changes in soil properties in areas originally occupied by TMCF in El Rincón, as only four variables, TN concentration (0.45% ± 0.13, mean ± SE), C/N ratio (20.01 ± 7.6), exchangeable Na content (0.04 cmolc kg–1 ± 0.02), and SMBC (820.53 µg g–1 ± 513.54), did not change significantly (P ≥ 0.84) with the age of the stand after abandonment. However, only three variables showed consistent trends across the three chronosequences: the thickness of the O horizon significantly increased (P = 0.0001, Fig. 2), and soil pH (P = 0.0001) and N/P ratio (P = 0.0001) significantly decreased with the age of the stand (Fig. 3). Most of the other soil properties had similar patterns of response at the two soil depths analyzed, depending of the studied chronosequence, as will be discussed below.



View larger version (19K):
[in this window]
[in a new window]
  		Fig. 2. Patterns of response to the age of stand (mean values) for the thickness of the O horizon in three chronosequences of tropical montane cloud forest at El Rincón, Oaxaca, Mexico.

 


View larger version (17K):
[in this window]
[in a new window]
  		Fig. 3. Patterns of response to the age of stand (mean values) for pH at the 0- to 20- and 20- to 40-cm soil depths and N/P ratio at the 0- to 20-cm soil depth in three chronosequences of tropical montane cloud forest at El Rincón, Oaxaca, Mexico.

 
The response to age in most of the soil variables analyzed (SOC, plant-available P, exchangeable Ca, Mg, Al, ECEC, and EA) was different depending on the chronosequence, as was statistically evidenced by the significant age x chronosequence interaction effect (P ≤ 0.001). The largest changes in soil characteristics with age, and the most consistent across chronosequences, usually took place {approx}15 to 45 yr after abandonment, after which time the patterns of change with age became more inconsistent among chronosequences (Fig. 4 and 5). Tanetze usually had a different response to age than Juquila and Yotao chronosequences, which, in turn, had similar responses despite the fact that Tanetze is geographically closer to Juquila (Fig. 1). In addition, Tanetze had lower concentrations of SOC, exchangeable Ca, Mg, K, and Al compared with Juquila and Yotao (Fig. 4 and 5, P ≤ 0.001). The SOC concentration significantly decreased in Tanetze (P = 0.0001), but considerably increased during the first 45 yr of forest development in Juquila and Yotao (P = 0.0001). In the three chronosequences studied, the highest concentrations of exchangeable Mg, Ca and BS were found in cornfield soils, which also had the lowest acidity (Fig. 4 and 5). By contrast, the highest EA was found at intermediate stages, either in young (Juquila and Yotao) or mature forest soils (Tanetze, Fig. 5). In most of the chronosequences and soil depths, exchangeable Ca, Mg, and K concentration had a sharp drop during the first 15 yr of forest development and remained relatively stable afterward or have an increase in the old-growth forest stage (Fig. 4). Exchangeable Al concentration (P = 0.167) did not change significantly with the age of the stand in Tanetze, but changed significantly in Juquila and Yotao (P ≤ 0.01, Fig. 4). In these two chronosequences, exchangeable Al concentration increased with the age of the stand. Effective cation-exchange capacity changed significantly with the age of stand in the three chronosequences studied (P = 0.01). The lowest concentrations of ECEC were found in cornfield soils, reached a maximum at young or mature forest soils, and decreased or remained stable in old forest soils (Fig. 4). Base saturation had a curvilinear response to time after abandonment in Tanetze and Yotao chronosequences (P = 0.0001). In Juquila, BS decreased with the age of the stand (P = 0.0001). As the age of the stand increased, plant-available P concentration had a significant tendency to increase in Juquila and Yotao (P = 0.0001) during the first 100 yr of forest development at both soil depths, and in Tanetze, during the first 45 yr, at 0 to 20 cm (P = 0.003, Fig. 5). Finally, DB significantly decreased in Yotao at the 0- to 20-cm soil depth (P = 0.001, Fig. 5). Except for the previous cases, no other significant changes were detected for plant-available P and DB.



View larger version (30K):
[in this window]
[in a new window]
  		Fig. 4. Patterns of response to the age of stand (mean values) for exchangeable Ca, Al, K, and Mg, and effective cation-exchange capacity (ECEC) at the 0- to 20- and 20- to 40-cm soil depths in three chronosequences of tropical montane cloud forest at El Rincón, Oaxaca, Mexico.

 


View larger version (26K):
[in this window]
[in a new window]
  		Fig. 5. Patterns of response (mean values) for soil organic carbon (SOC), plant-available P, soil bulk density, exchangeable acidity (EA) and base saturation (BS) to the age of stand at the 0- to 20- and 20- to 40-cm soil depths in three chronosequences of tropical montane cloud forest at El Rincón, Oaxaca, Mexico.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 SITE DESCRIPTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
In general, the studied soils were acidic, rich in SOC, with high levels of exchangeable Al, but with low contents of exchangeable base cations and plant-available P, as has been reported in soil studies of TMCFs areas in other parts of the world (Meave et al., 1992; Bruijnzeel and Proctor, 1995). Conversion of croplands to secondary forest in areas originally occupied by TMCF results in substantial gains of SOC sequestration the first 15 yr of forest development at El Rincón. However, after 15 yr, SOC sequestration followed different paths depending on soil depth and the time interval. Post and Kwon (2000) reviewed the reported trends in C sequestration during forest establishment after agricultural use, and could not find a common trend. Our results show a great variation in SOC sequestration even within the same locality.

Forest development in TMCF areas used for corn cropping and later abandoned were associated with changes of soil properties. Of these, one of the most important is a natural progression of soil acidification. At least five processes may account for such an increase: (i) The uptake of nutrient cations, and (ii) the release of acidic litter by the growing vegetation. (iii) The weathering of parent material favored by soil acidification. Indeed, minerals, such as muscovite and chlorite, had a sharp decrease as the forest aged at our study area (Bautista et al., 2003). As a result (iv), Al and other cations are released to the soil (Pappe and Legger, 1994). Low pH increases Al solubility (Brady and Weil, 1999). In fact, at our study sites, exchangeable Al was negatively correlated with soil pH (r = –0.447, P < 0.0001). Finally (v), a further loss of base cations is expected due to the higher affinity of the exchange complex by H and Al ions over base cations. Similar changes in soil pH during secondary succession have been found in other humid areas associated with fire. In Chiapas highlands, early successional soils of pine forests were less acidic than late-successional soils under hardwoods (Galindo-Jaimes et al., 2002).

The increase in thickness of the litter layer with stand age in the three chronosequences is expected given the increasing acidity detected, which inhibits decomposition rates (Brady and Weil, 1999), and the overall increase in plant biomass (Table 1). Aluminum is another factor explaining this result, as it plays an important role in SOM stabilization by lowering soil mineralization rates (Boudot and Choné, 1985; Abadín et al., 2002), and, at our studied sites, SOM content was positively correlated with soil Al concentration (r = 0.625, P < 0.0001). The pattern of litter accumulation with stand age was similar to those found in a forest ecosystem of Mauna Loa, HI (Raich et al., 1997), but contrasts with that found in temperate forests, as those in the Upper Coastal Plain of the eastern United States, where the litter layer increases during early successional stages, but decreases at late stages (Switzer et al., 1979). A shift in species composition during succession altering soil pH is the most likely explanation. Early successional stages are dominated by conifers, whose litter contains large amounts of phenolic compounds and lignin yielding acid residues, lowering soil pH and slowing decomposition rates (Scholes and Nowicki, 1998). The hardwoods that replace the conifers produce less-acidic litter favoring SOM decomposition, and decreasing the thickness of the litter layer. During secondary succession in TMCF areas of southern Mexico, a similar pattern occurs with respect to species change: conifers dominate early successional stages and, eventually, are replaced by hardwoods (e.g., Cordova and del Castillo, 2001; Galindo-Jaimes et al., 2002). However, the thickness of the litter layer does not decrease when hardwoods replace conifers. There is scarce information on the patterns of litter accumulation during secondary succession (Facelli and Pickett, 1991), but our results show that extrapolations of the patterns of litter accumulation during succession, observed in other kinds of humid forests, are unwarranted to TMCFs.

In contrast with other kinds of humid forests showing the opposite trend (Peet, 1992), the N/P ratio decreased significantly as forests develop in areas originally occupied by TMCFs in southern Mexico. This relation is expected to change during succession due to the contrasting differences in biogeochemical cycles of N and P. Nitrogen is derived primarily from the atmosphere by microbial fixation, while P is derived primarily from rock weathering. Thus, N is nearly absent in young soils, but, with time, the invasion of N fixers is expected to increase the biological availability of N in the system due to atmospheric fixation (Gorham et al., 1979). By contrast, the amount and availability of P is expected to decline during long-term soil development due to the demands of the growing vegetation, in such a way that P is becoming largely bound to secondary minerals or SOM, leading to extremely P-deficient soils (Aerts and Chapin, 2000). However, primary TMCFs are not particularly rich in species harboring N-fixing symbionts. The absence of legumes is characteristic of upper TMCFs (Webster, 1995); and, at our study site, this family was virtually absent except for few herbs detected in incipient forests. Other species known to harbor N-fixing symbionts, found at our study sites, are Myrica cerifera L. and Alnus acuminata Kunth (Alexander, 1994). The former species is more abundant at early successional stages, whereas A. acuminata was only detected in Yotao chronosequence (Blanco, 2001). Nonsymbiotic N fixation probably increases in older stands, as the greater humidity of the forest is likely to enhance the establishment and growth of N fixers at both the forest floor and in epiphytes. Such results may explain the lack of significant changes of N with time.

On the other hand, P was found to be very scarce in ecosystems originally occupied by TMCFs of southern Mexico. At the study sites, the lack of essential minerals containing P may explain this result (Bautista et al., 2003). However, P concentration in the soil significantly increased with the age of the stand in Juquila and Yotao chronosequences during the first 100 yr of forest development, and in Tanetze during the first 45 yr. These results may contribute to explain the decrease in N/P ratio observed in the three chronosequences, and suggest that other forms of P uptake, such as dilute aerosol deposition, become more important during forest development. Older forests in TMCF areas are probably more efficient at capturing atmospheric nutrients due to the greater canopy of the trees and the higher cover of epiphytes, in particular bryophytes (Cordova and del Castillo, 2001). The latter are well known for trapping airborne particles and holding substantial amounts of water (see Pcks, 1980). Other forms of P input cannot be ruled out. In a chronosequence in Krakatau, Indonesia, Schlesinger et al. (1998) found an excess of accumulation of organic P relative to the losses of inorganic P, probably as a result of plant uptake from lower soil depths. Higher N/P loss ratios in Hawaiian TMCFs were explained by an efficient P recycling and high N throughputs (Hedin et al., 2003). In any case, our results do not support the generalization that N is limiting during earliest stages and P during late stages of soil development (Walker and Syers, 1976).

The major and most consistent changes in soil properties across chronosequences took place the first 15 to 45 yr after abandonment. Such changes coincide with the most dramatic changes detected in vegetation structure during succession, namely the highest increases in tree density and basal area (Table 1), supporting the general notion that vegetation is a major determinant of soil properties (Jenny, 1941). The highest concentrations of exchangeable K, Ca, and Mg were found mainly in soils of crop fields, probably as a consequence of the slash-and-burn system of cultivation, which releases essential nutrients to the soils (Adedeji, 1984). These results are consistent with the information provided by local people regarding the recent origin of the crop fields studied (<3 yr); otherwise, these cations are expected to be lost by leaching given the heavy rains, steep slopes, and low vegetation and litter cover, as it happens in lowland tropical areas (Ewel et al., 1981). The decrease in K, Mg, and Ca concentrations from cornfields to incipient or young forests in all chronosequences could be explained by the high uptake of the growing vegetation during the first 45 yr after abandonment. Such results also support the notion that soil fertility is one of the major factors limiting production in TMCFs areas of the world (Grubb, 1977; Bruijnzeel and Veneklass 1998). The hypothesis that soil fertility is a factor that limits production seems to find additional support by the fact that, of all chronosequences studied, Tanetze had the lowest concentrations of SOC, exchangeable K, Ca, Mg, and ECEC, and the lowest basal area, particularly {approx}45 yr after abandonment (Table 1, Fig. 4 and 5). After {approx}45 yr, self-thinning and the slow pace of vegetation growth may allow certain recovery of the levels of these cations in soil in Yotao and Juquila chronosequences. This may not happen in Tanetze probably because of its lower levels of soil fertility and the continuous extraction of firewood in the old-growth forest, not found in Juquila and Yotao. Significant losses of essential elements via harvesting of wood take place in other tropical ecosystems (Ewel et al., 1981). The differences in change in soil properties with time among chronosequences point for more studies and the need of including several chronosequences within a single area before making general statements about soil changes during secondary succession.


   ACKNOWLEDGMENTS
 
We thank Javier González-Cruz for his invaluable field assistance, Raúl Rivera for his help in the field and cartographic work, and Jorge Etchevers for lab facilities. M.A. Hajabbasi, W.H. Schlesinger, and four anonymous reviewers provided valuable comments to the manuscript. The first author enjoyed a scholarship through CONACyT during postgraduate studies at Colegio de Postgraduados, Mexico. This research project was funded by grants from the Darwin Initiative for the Survival of Species (United Kingdom), the European Community INCO IV programme (BIOCORES project contract no. ICA4-CT 2001-10095), by CONACyT, Sistema de Investigación Benito Juárez, and Instituto Politécnico Nacional.

Received for publication March 31, 2004.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 SITE DESCRIPTION
 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 Web of Science
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 Web of Science (16)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bautista-Cruz, A.
Right arrow Articles by del Castillo, R. F.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Bautista-Cruz, A.
Right arrow Articles by del Castillo, R. F.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Bautista-Cruz, A.
Right arrow Articles by del Castillo, R. F.
Related Collections
Right arrow Ecosystem Management
Right arrow Tropical Soils
Right arrow Biogeochemical Processes
Right arrow Tropical Soil Management
Right arrow Forest Soils

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