Study area

This research was carried out in three different ecosystems (pine, oak and pine-oak), which are located in the Pablillo ejido, Galeana municipality, state of Nuevo León (Figure 1), whose climate is arid, temperate, with over 18 % of the rainfalls per year occurring between summer and winter (^{García, 2004}). The predominant plant communities in the area are scrub and coniferous forests with the presence of Pinus pseudostrobus Lindl. and mixed stands with Pinus teocote Schltdl. & Cham., P. ayacahuite var. brachyptera Shaw, P. arizonica Engelm., P. cembroides Zucc., P. hartwegii Lindl., Pseudotsuga menziesii (Mirb.) Franco var. glauca (Mayr) Franco and the association of Pinus pseudostrobus with Quercus spp., as well as with Quercus sideroxyla Bonpl., secondly: P. culminicola Andresen & Beaman, Arbutus glandulosa M. Martens & Galeotti, Quercus crassifolia Bonpl., Alnus acuminata Kunth, and Juniperus deppeana Steud. var. robusta Martínez (^{Tapia y Návar, 2011}).

Figure 1 Location of the study sites in Galeana municipality, Nuevo León, Mexico. 

The altitudinal range in which the ecosystems under study develop is 2 267 masl for pine (geographic location: 24°34'44.34" N; 99°58'30.12" W), 2 313 masl for oak (geographic location: 24°34'28.32" N; 99°58'33.180" W), and 2 160 masl for pine-oak (geographic location: 24°35'27.66" N; 99°58'34.50" W).

Mulch degradation and harvesting frequency

Based on the litter decomposition bag method proposed by ^{Bocock and Gilbert (1957)}, the decomposition process of mulch in three temperate ecosystems was evaluated over a one year period under outdoor conditions. Mulch corresponds to all the material that is formed once the leaf litter falls to the ground forming an organic layer (^{Pérez-Vázquez et al., 2021}); it was handled uniformly in the three ecosystems, for this purpose, 10 g dry weight of a uniform mixture of the litter components (leaves, branches, bark, reproductive structures and other miscellaneous components), collected in each of the ecosystems one month before the start of the experiment so that it would be fresh material and placed in paper bags, it was then dried at 65 °C. The collection was carried out randomly at each site (^{López-Hernández et al., 2018}).

A total of 60 decomposition bags were used for each ecosystem type; 12 were randomly distributed in five 20 m × 20 m experimental plots adjacent to one other in order to cover the total area of the site in each ecosystem. The 25 cm × 20 cm bags were made of 1 mm nylon fabric (^{Rodríguez et al., 2019}), allowing invertebrate detritivores to access the interior but minimizing fragmentation losses (^{Martínez-Atencia et al., 2020}). Each bag was placed on the surface of the soil without altering the conditions of the decomposing substrate, in order to avoid modifying the natural conditions in which the decomposing microorganisms develop, in addition to maintaining contact with the organic matter of the soil (^{Rodríguez et al., 2019}).

The assessment was carried out during 12 sampling dates between February 2021 and January 2022; five bags were collected every 30 days from each plot and study ecosystem. The material was transported to the Chemistry Laboratory of the Graduate School of Forest Sciences of the Universidad Autónoma de Nuevo León, where the methodology proposed by ^{López-Hernández et al. (2018)} was applied to dry it in a FE-292AD Felisa^® forced-air oven at a constant 65 °C temperature, until a constant weight (g) was obtained. Several studies (^{Rivera et al., 2013}; ^{Tapia-Coronado et al., 2022}) consider that soil particles adhering to the plant material do not alter the results and can be regarded as negligible, however, in order to remove them, the subsamples were cleaned with a brush. The remaining weight (g) of the mulch from each repetition was obtained in order to estimate the degree of decomposition by the difference between the initial weight and the final weight of each sampling date, both were recorded with a LC 620 S Sartorius^® digital balance, with a 0.001 g resolution.

Using the Simple negative exponential model proposed by ^{Olson (1963)}, the decomposition rate of the uniform mixture of the mulch components (10 g) was estimated for each ecosystem. In addition to Olson's (1963) exponential model, and in order to estimate mulch decomposition rates, other mathematical models (simple linear, logarithmic, and power) were tested to determine which had a better fit in estimating the decomposition rate compared to the reference model:

Where:

H = Remaining mulch fraction (years)

T = Time (years)

nl = Natural logarithm

β ₀ = Percentage of initial weight of mulch

β ₁ = Mulch degradation rate

Statistical analyses

The averages of the remaining mulch fraction by sampling date and ecosystem were subjected to statistical analysis to test the assumptions of normality (Kolmogorov-Smirnov test with Lilliefors correction) and homogeneity of variances (Levene's test) (^{Ott, 2001}). Since both assumptions were demonstrated, an analysis of variance was performed according to an experimental design with a classification criterion (^{Steel and Torrie, 1980}) in order to detect the differences in the amount of remaining mulch fraction between forest ecosystems for each sampling date. A comparison of means of the remaining mulch fraction by date was performed using Tukey's HDS test (p=0.05) (^{Ott, 2001}). The evaluation of the goodness of fit of each utilized model for determining the mulch decomposition rate was carried out based on the graphs of the residues, the root mean square values of the error, and Pearson's coefficients of determination (R ² ) and correlation (r) (^{Ott, 2001}). All statistical analyses were performed with the Statistical Package for Social Sciences (SPSS) version 22.0 for Windows (^{IBM, 2016}).

Mulch decomposition

Table 1 shows the average remaining mulch fraction by ecosystem. The statistics and p-values of the Kolmogorov-Smirnov test and Levene's test (L), as well as the F and p-values of the analysis of variance by sampling date, are also included. According to Tukey's HDS test, most of the collection dates showed significant differences (p<0.05) between ecosystems, except for the fourth (MAY-21), sixth (JUL-21) and last (JAN-22) sampling dates.

Figure 2 shows the dynamics of the remaining mulch fraction for the three ecosystems during the study period. In the first three dates (FEB-21 to APR-21), the remaining fraction of mulch ranged between 0.913 (pine-oak ecosystem) and 0.958 (oak ecosystem). Between APR-21 and DEC-21, the fraction of remaining mulch in the pine ecosystem gradually decreased to 0.83 %, while in the pine-oak and oak ecosystems this fraction was 0.78 % and 0.69 %, respectively; decomposition was more noticeable in the oak ecosystem. With respect to the last collection date (JAN-22), a slight increase in the remaining fraction of mulch was observed in the three ecosystems, according to ^{Pinos et al. (2017)}, this phenomenon happens due to the particular climatic conditions that develop during the decomposition process, which act in a direct way, slowing or increasing the process. In the same sense, ^{von Arx et al. (2012)} state that the decomposition process is strongly influenced by climate characteristics such as humidity, temperature, and precipitation, which actively slow down decomposition during specific periods of time, and even stop it.

Each plotted value represents the mean of n _ij observations of ecosystem i on sampling date j, with i = 1, 2, 3; j = 1, ... , 12 ± Standard error of the mean.

Figure 2 Fraction of remaining mulch of the three ecosystem types. 

The results of the present study are congruent with the research conducted by ^{Xuluc-Tolosa et al. (2003)}, but differ from that documented by ^{Torres et al. (2018)}, who indicate that the rate of decomposition in the first months is higher and then the slope becomes slower and is significantly reduced, which largely responds to the type of ecosystem studied; the rapid mass loss in the first period is associated with the low presence of tannins and nitrogen released during the decomposition process of mangrove-associated species. On the contrary, temperate forest ecosystems may exhibit a slow decomposition process in the first phase due to the chemical composition of pine needles (^{Moreno et al., 2018}).

^{Rivera et al. (2013)} point out that the first stage is controlled by fragmentation by the soil fauna and is related to water-soluble compounds, while in the second stage it is more directly related to the lignin content of the mulch, which slows down decomposition. This may explain the differences in mulch decomposition in the pine ecosystem, which was slower than in the oak ecosystem (Figure 2).

This reasoning allows us to assume differences in the chemical composition of the mulches in the ecosystems studied. In this regard, ^{Bruno-Eutimio et al. (2022)} document that the chemical composition of the mulch is a determining factor that is reflected during the decomposition process. In the same sense, ^{Avendaño-Yáñez et al. (2020)} and ^{Berg and McClaugherty (2020)} conclude that the differences observed during the decomposition process are associated with the type of litter and its chemical composition, which can pause or drive the process in different ecosystems. The percentage of mulch decomposition at the end of the study was 14 % for the pine ecosystem, 22 % for the pine-oak ecosystem, and 23 % for the oak ecosystem.

According to the results of ^{Rodríguez et al. (2019)}, the loss of the remaining mulch fraction for pine, oak, and pine-oak ecosystems was 17, 27 and 23 %, respectively; which coincide with the intervals observed in the present study. Therefore, the decomposition process in pine, oak, and pine-oak ecosystems follows a trend, even when these ecosystems are located in different areas, with different soil types and particular characteristics, and these similarities can be explained at the local scale by the chemical composition of the mulch typical of mixed forests, which generally have low fertility (^{Reed et al., 2012}).

From the research study developed by ^{Soong et al. (2015)} under laboratory conditions at the Colorado State University, mulch fraction losses for pine (Pinus ponderosa P. Lawson & C. Lawson) is 28 %, and 58 % for oak (Quercus macrocarpa Michx.), with higher decomposition compared to the results of the present research for oak, results that are associated with some factors acting passively in the decomposition process, for example: chemical and biotic factors that play a fundamental role in the behavior of mulch mass loss, even when they are species of the same genus (^{Bangroo et al., 2017}).

An important factor to consider in mulch decomposition studies is the time to evaluate the process of mass loss, mainly associated with forest ecosystems, as suggested by ^{Prescott (2005)}, who indicates that decomposition patterns can be modified in long-term experiments, since the percentage of decomposition of mulch in the long term reaches a higher disintegration and thus a better understanding of the process is achieved. In this regard, ^{Horodecki and Jagodziński et al. (2019)} document that, in order to understand processes such as mineralization and humidification, it is necessary to fully analyze the process, whereby a more accurate understanding of nutrient release can be attained.

The study period for the present study was one year, considered convenient for temperate climate species associated with forest ecosystems (^{Edwards, 1977}; ^{Rocha-Loredo and Ramírez-Marcial, 2009}; ^{López-Hernández et al., 2018}), which compared to other studies in tropical climate ecosystems is sufficient to stabilize the decomposition process. For example, ^{Oliveira et al. (2019)} only analyzed leaf litter decomposition over a period of 136 days for Urochloa brizantha (Hochst. ex A. Rich.) R. D. Webster, and conclude that the residual mass of mulch after one day of decomposition adds to the mass of newly deposited mulch that continues to decompose day after day; thus, it is a continuous process that needs to be studied at continuous time intervals. ^{Rivera et al. (2013)} studied the process of disintegration of organic matter in a temperate ecosystem in the Selva El Ocote Biosphere Reserve for 223 days, including several species, among them: Enterolobium cyclocarpum (Jacq.) Griseb., Salix bonplandiana Kunth, Cecropia peltata L., Heliocarpus appendiculatus Turcz. and Mastichodendron capiri (A. DC.) Cronquist. The most determining factor in the decomposition process was the age of the individuals. ^{Tapia-Coronado et al. (2022)} in 210 days analyzed the decomposition behavior of leaf litter from forest species [Gmelina arborea Roxb. ex Sm., Tabebuia rosea (Bertol.) DC. and Acacia mangium Willd.] in tropical dry climate silvopastoral systems, whose decomposition process is similar to that of the present work, with a rapid initial loss of material that favors the active release of nutrients, and then becomes a slower process. ^{Zhang et al. (2008)} observed that variations in mulch decomposition rates are mainly due to differences in litter quality, microclimates, microbial community type and composition, and soil properties, with mulch quality being the main factor in the decomposition process.

In general, in broadleaved forests where precipitation and temperatures are high, the decomposition process is usually more accelerated than in ecosystems where temperature and precipitation are lower, for example, coniferous forests. Another determining factor is the acidity of the soils in coniferous forests, which, being more acidic than in other ecosystems, such as grasslands, limits microbial activities and, therefore, reduces the decomposition process of mulch (^{Gholz et al., 2000}).

Mulch decomposition models

The values of the fit statistics of the four decomposition models evaluated are described in Table 2; it is observed that the Coefficient of determination presented a fluctuation from 0.455 to 0.719, relatively lower values in relation to those documented by ^{Moreno et al. (2018)} in mixed ecosystems (0.965 to 0.069), and ^{Rodríguez et al. (2019)} who refer for this statistic an interval between 0.577 and 0.899 in forest ecosystems in the north of the country.

The values obtained from the fit of the models in the three ecosystems under study for the Correlation coefficient ranged between 0.675 and 0.848, which are low compared to those cited by ^{Del Valle-Arango (2003)} of 0.966, and ^{Ibarra et al. (2011)} of 0.946. The Simple negative exponential model (^{Olson, 1963}) presented a good fit according to the Coefficient of determination, ranging from 0.617 (pine-oak ecosystem) to 0.710 (oak ecosystem). Pearson's correlation coefficient was distributed between 0.785 and 0.843. This model is the most commonly used to describe the rate of mulch decomposition (^{Swift et al., 1981}; ^{Burghouts et al., 1998}; ^{Gama-Rodrigues et al., 2003}; ^{Tapia-Coronado et al., 2022}).

The Simple linear model had a Coefficient of determination (R ² ) ranging from 0.620 (pine-oak ecosystem) to 0.719 (oak ecosystem). The logarithmic and power models presented low determination coefficients (0.462 to 0.543), with respect to the exponential and simple linear models.

Mulch decomposition rate

The decomposition rate (constant k) of mulch resulting from the exponential and simple linear models, which were considered as a reference based on the highest values of the coefficients of determination, are presented in Table 2; it is observed that this constant for the three ecosystems was 0.352, 0.180, and 0.128, respectively, from which the theoretical time required for 25 %, 50 %, and 75 % of the mulch to degrade was estimated. The estimated values, according to the model proposed by ^{Olson (1963)}, ranged from 0.866 (oak ecosystem) to 10.455 years (pine ecosystem) (Table 3).

The values of the decomposition constant (k) of the Simple linear model ranged from -0.116 (pine-oak ecosystem) to -0.303 (oak ecosystem) (Table 2). Estimates of the time required to degrade mulch 25, 50, and 75 % ranged from 0.854 (oak ecosystem) to 6.080 years (pine ecosystem) (Table 3). These results indicate, in general, that the estimation of mulch decomposition is higher with the Olson model than when using the Simple linear model, as the average decomposition times are longer. It should be noted that the decomposition times of mulch at 25, 50, and 75 % in both models should be considered with reservations, since the environmental conditions ―biotic, abiotic, and chemical― do not remain constant.

In this sense, ^{Tresch et al. (2019)} point out that the process is influenced by the environmental conditions of each site, which modify the decomposition rates; ^{Correa-Araneda et al. (2020)} refer to the fact that the environmental conditions prevailing at specific time intervals control decomposition through various direct and indirect processes that are rarely constant, for example, ^{Bölscher et al. (2020)} mention that high temperatures and precipitation facilitate soil acidification, which probably increases the presence of decomposing agents, and therefore microbial activities, which can be altered by extraordinary climatic phenomena, and, consequently, decomposition rates are modified.

Figure 3 illustrates the dispersion of the data (observed and predicted) of the remaining mulch fraction considering the Exponential model and the Simple linear model during the study period in the three temperate forest ecosystems.

a = Predicted values obtained with reference to the Exponential model for each ecosystem under study; b= Predicted values obtained by reference to the Simple linear model for each ecosystem under study.

Figure 3 Observed and predicted values of mulch decomposition during the time (year) of the trial in each ecosystem. 

The values of k in comparison with those obtained for other ecosystems seem low. ^{Rocha-Loredo and Ramírez-Marcial (2009)} recorded in pine-oak, pine, and oak forests a k constant of 1.40, 1.44 and 1.74, respectively, while ^{Rivera et al. (2013)} estimated that the time required to degrade 50 and 99 % of the mulch in low deciduous forest is 75.9 and 504.3 days, respectively. ^{Tapia-Coronado et al. (2022)} documented that the decomposition rate for forest species is more accelerated than that of grasses, with k values of 1.77, which allows inferring that the decomposition rate of mulch is influenced by the composition of plant species and their distribution (^{Peña-Peña and Irmler, 2016}; ^{Djukic et al., 2018}). The present study did not consider the chemical analysis of the mulch components, therefore, it is possible to assume that the differences in the values of the Decomposition constant (k) between the ecosystems studied depend directly on the species composition and the successional stage of the ecosystems, as indicated by ^{Rocha-Loredo and Ramírez-Marcial (2009)}. On the other hand, the nutritional and structural components of the mulch are factors that act directly on decomposition rates and account for their differences, although these are not usually studied (^{Berg and McClaugherty, 2020}).