Temporal evolution of litterfall and potential bio-element return in a successional forest sequence of the Espinal Ecoregion, Argentina.

Aim of study: The aim of this work was to assess the litterfall contribution and the return of bioelements of a successional forest sequence from the Mesopotamian Espinal (Argentina) which was associated with livestock production. Area of study: Mesopotamian Espinal, Argentina. Material and methods: Litterfall samples were taken and a chemical characterization of their fractions was determined in three stages: a) in the initial successional stage ( IF ); b) in an intermediate secondary forest ( SF ); and c) in a mature forest ( MF ). Main results: The litterfall contribution of the three forests was 1140 ±98, 2947 ±154, and 2911 ±57 kg DM ha -1 yr -1 ; respectively. The IF showed a seasonal pattern of contribution with a peak occurring during summer (528 ±85 kg DM ha -1 yr -1 ), then decreasing during autumn, winter, and spring (241 ±30, 165 ±27, and 207 ±29 kg DM ha -1 season -1 ,respectively). The SF showed a rather constant seasonal pattern (about 750 kg DM season -1 ). The MF showed significant differences among seasons, the maximum and minimum contributions ranging between 846 ±29 and 598 ±33 kg DM ha -1 season -1 in summer and spring, respectively. The litterfall leaves/branch ratio decreased as ecological succession advanced, being lower as the forest gets more mature. As a consequence, this ratio can be used as an indicator of maturity in the sequence. The potential return of bio-elements of the successional forest sequence was proportional to the litterfall input, with a maximum amount of N in the Fabaceae species.  Research highlights: The litterfall assessment and the leaves/branch ratio allowed the characterization of the successional stages in Xerophytic forest used for livestock production.  Keywords: Semi-xerophytic trees; tree production pattern; plant organ contribution; leaf/branch ratio; return of bio-elements; tree nutrients.

ce the release rate of nutrients in each forest ecosystem (Aceñolaza et al., 2009), determining its nutritional status.
Nutrient inputs to the soil through litterfall, is known as the potential return of bioelements (PRB), since they enter the soil after litter mineralization (Martín et al., 1996;Gallardo et al., 1989). These authors also found that, in general, temporal variations of bio-element return to the soil (via litterfall) follow an evolution similar to the pattern of litterfall contribution. Analyzing the different fractions of litterfall allows us to find some indicators of forest maturity. Martín et al. (1993) found that, as forest mature, the relationship between leaf and branch contribution tends to drop, even the leaf litter contribution increase, the branch production grows with a higher ratio.
The Mesopotamian Espinal has been recently modified in several areas of Entre Ríos province, due to the advance of the agricultural frontier, producing an important reduction in forest area (Aceñolaza, 2000;Muñoz et al., 2005;Arturi, 2006;Maldonado et al., 2012) and leading to an increasingly heterogeneous land mosaic (Maldonado et al., 2012). Inappropriate livestock management or forest exploitation for timber production (mainly Prosopis nigra) or firewood (Acacia caven and Prosopis affinis) had early transformed the native and primary forests into degraded or secondary forests. Deforestation and land abandonment have led to the development of a sequence of successional forest stages, with an initial stage dominated by a woodland of A. caven, followed by its colonization by P. affinis, and leading to a mature forest dominated by P. nigra, and thus, to the reestablishment of a stable forest (Aceñolaza, 2000;Lewis et al., 2006).
In this study, we firstly hypothesize that the plant succession, as an ecological process, has a direct influence on the litterfall production and, as a result, on the potential nutrient return (PRB) in the forests of the Mesopotamian Espinal; secondly, that leaf/branch ratio (L/B), as a characteristic of the litter production, increases as the ecological succession advances.
This work shows data about the contribution of litterfall and quantification of PRB in Mesopotamiam Espinal forests, little-known issue until now. This subject has also a practical interest, particularly because the forests studied are associated with extensive livestock production. Therefore, the aims of this work were to: A) Estimate the inputs of litterfall in a degraded/ restored successional sequence of the Argentine Mesopotamian Espinal eco-region. B) Describe the seasonal pattern of litterfall production in these forests. C) Quantify the potential return of bio-elements per species and forest in the successional sequence considered.

Material and methods
The Espinal is an Argentine eco-region located between 28°and 40°S latitude, to the South of the Chaqueño Park, covering approximately 330,000 km 2 (Lewis et al., 2006) and forms an arch that surrounds the Pampas Grassland eco-region. The Mesopotamian Espinal, in Entre Ríos province, comprises forests belonging to the Espinal phytogeographic Province, especially in Ñandubay district (Cabrera, 1994).

Study area
The study was conducted in the Villaguay Department (covering an area of 1,600 km 2 in Entre Ríos province, Argentina; Fig. 1).
The landscape was shaped by morphogenetic processes, and is currently a pen-plain, ranging from slightly undulating to plain relief, with evidences of characteristic gilgai micro-relief (De Petre & Stephan, 1998). Mean height is 40 m a.s.l. (INTA, 2000).
The climate in Villaguay is humid temperate. Annual mean precipitation is 1,000 mm yr -1 , with an important inter-annual variability (INTA, 2000), but with peaks of rainfall in autumn and spring. Mean annual temperature is about 16°C, with the mean temperature of the coldest (July) and hottest (February) months being 11 and 25°C, respectively. Temperature changes gradually from season to season; however, there may be days with minimum temperatures below 10°C in summer and with maximum values of 30°C in winter. The ombro-thermic diagram for the Villaguay Department, constructed with mean values of monthly precipitation and temperature (from 1980 to 2010), shows a drought period in February (Mendoza et al., 2012).
The present study was carried out in three different forests located in similar environmental (soils and topography) and land use conditions. Considerations were taken to circumscribe the successional pattern as the main source of variation, choosing stands with similar topography, soil, and land use (kind and intensity). All stands, at the present time, have extensive livestock grazing (in rotation), with 0.4 and 0.5 units ha -1 , and have no logging activity. They are situated in a 20 km range and represent different successional stages ( Fig. 1; INTA, 2000).
In each forest stand, four permanent plots of 800 m 2 were established to evaluate its structure and composition. All stands, in each forest category, were placed between 100 and 200 m of each other, looking for similar ecological/environmental conditions (e. g., soil, topography, and land use) and avoiding closeness.
The forests selected had the following characteristics: a) Initial forest: Monospecific forest of A. caven (IF), originated by colonization of this species in a cropland abandoned in 1998 (approximately 10 years at the beginnings of this study). b) Secondary forest: Mixed forest (SF) dominated by P. affinis, with the presence of other characteristic species, such as A. caven and C. ehrenbergiana, which includes trees of between 50 and 70 years old. c) Mature forest: Mixed old-growth forest (MF), dominated by P. nigra, with the presence of A. caven, P. affinis, and C. ehrenbergiana. This forest, even it presents signals of degradation (overgrazing, selected logging and fire), does not come from a clearcut area.

Climatic data
In the period when the samples were taken, temperature and rainfall were measured with a weather sta-

Litterfall sampling
In the three forests selected, four randomly stands (20 × 20 m 2 ) were evaluated as described before by Mendoza et al. (2012). Within each stand, all individual trees were identified to the species level.
Tree diameter at breast height (DBH > 5 cm) was measured, using a tape. Tree heights were calculated using a clinometer device. Tree canopy cover was estimated using the mean between N-S and E-W diameters of crowns. The main forest characteristics of the species present in MF, SF, and IF are showed in Table 1.
Litterfall samples were taken from the selected forest stands during 2009-2010, following the method proposed by Gauch (1982) and Aceñolaza et al. (2006). A total of 32 production boxes, with a 0.25 m 2 surface, were placed in December 2008. The criteria used to place boxes within the stands were under same species and only dominant woody species of each stand. In this sense, 4 boxes were placed under A. caven in IF, 12 boxes under SF (4 under C. ehrenbergiana,4 under A. caven,and 4 under P affinis), and 16 under MF (4 in each of the 4 named species).
Individual boxes were randomly located under trees in the selected areas of each stand. Production boxes were placed on the soil, but separated 3 cm from soil and the bottom covered by a catching net. They were monthly sampled, beginning in January 2009 and continuing for two years.
Litterfall samples were collected from each box in different paper bags that were previously labeled at the lab. Later they were dried in a 70°C heater until a constant weight was reached; then these samples were classified and weighed, separating the following fractions: a) leaves; b) branches; c) flowers; d) fruits; and d) 414 C. A. Mendoza et al. / Forest Systems (2014)  others (other plant material from other tree, shrub or herb species, excluding animal feces). Results of annual litter production are expressed in kg of dry matter (DM) ha -1 yr -1 .
Total litterfall, litterfall fraction, and moisture content of each production box were determined per species and per forest as a function of time. Total litterfall was calculated as the average of the annual accumulation of litterfall production.
Collecting trays were placed under individual species. For referring the litterfall production to ha, an adjustment of litter input values by tree should be performed, to f it the values measured to the actual coverage of each species in the area. For this, the relative tree cover on each species (Mendoza et al., 2012) was multiplied by the litterfall (resulting kg MS ha -1 y -1 ). This allowed us to estimate the real contribution by species and site. In the forests more than one species are found, total contributions is obviously the sum of the relative contributions of each species.

Chemical analyses of litterfall samples
All plant samples were ground using a ball mill (Retsch, model MM301). These samples were then weighed using a precision balance (± 0.001 g) and burned in a muffle furnace (at least 500°C) for 5 h.
Ashes were digested with concentrated HCl and then diluted with distilled water to a known volume. Aliquots were taken to determine P, using the vanadomolybdate yellow procedure (Chapman & Pratt, 1979) and a spectrophotometer (Genesys, model 20). Another aliquot of the same acidic solution was used to determine Ca, Mg, K, S, Na, Fe, Al, Zn, Mn, and Cu by means of inductively coupled plasma emission spectrometry (ICP-OES).
After the acid digestion of plant samples, N content was determined using a flow auto-analyzer (Bran+ Luebbe AA3).

Quantification of potential return of bio-elements (PRB)
The amount of bioelements that potentially can return to the soil through litterfall was determined by multiplying the pooled mean annual litterfall production estimated during the study period by the mean chemical composition of each plant residue. Then, PRB was calculated for each fraction per species and per forest.

Statistical analyses
Temporal analysis included average litterfall production accumulated by month and season (summer: January-March, autumn: April-June, winter: July-September, and spring: October-December).
Data were analyzed using a linear mixed model analysis of variance (McCulloch & Searle, 2001) with two between-subject factors in a hierarchical design (forest and species nested within forest) and three within-subject factors of repeated measures (year, season, and month nested within season). The comparison between means was done by using Fisher's Least Significant Difference (LSD) test.
For each fixed forest, year, and season, covariance among months is constant for a single species and equal for all species. Therefore, the structure of variances of the mixed linear model was estimated using the restricted maximum likelihood method (Bartlett, 1937).
The model proposed allowed us to estimate the actual contribution of the different forests as a function of the entire forest tree cover and the relative cover per species in the three stands representative of the Mesopotamian Espinal (Mendoza et al., 2012).

Total litterfall annual production
Mean forest litterfall contribution (Table 2) ranged between 1,140 kg DM ha -1 yr -1 in IF and 2,947 kg DM ha -1 yr -1 in SF, showing no significant differences with MF (2,911 kg DM ha -1 yr -1 in MF, p > 0.05). However, both SF and MF differed statistically from IF (p < 0.05).

Total litterfall seasonal production
Comparison of forests seasonal patterns (Table2) showed signif icant differences among forests in summer (p < 0.05); SF and MF exhibited the greatest production, without significant differences between both them (p > 0.05). The same pattern was observed Variation in litterfall in a forest successional sequence in autumn, SF and MF having the maximum production in summer and autumn.
IF, SF, and MF showed signif icant differences (p < 0.05) in total litterfall production in the three forests in winter and spring. In winter, IF, SF, and MF produced 165, 637, and 764 kg DM ha -1 season -1 , respectively; however, in spring, SF showed maximum production, with values of 207, 755, and 598 kg DM ha -1 season -1 for IF, SF, and MF, respectively (Table 2). The analysis of the temporal pattern exhibited a maximum contribution in summer for IF (528 kg DM ha -1 season -1 ), with signif icant differences (p < 0.05) from the other seasons (Table 2). SF shows constant production (p > 0.05). Finally, MF presented significant differences among seasons (p < 0.05), with the maximum and the minimum seasonal litterfall production ranging between 846 and 598 kg DM ha -1 season -1 (summer and spring, respectively; Table 2).

Total litterfall production by species
Comparing the total litterfall production by species, the maximum contributions were found in A. caven (IF), C. ehrenbergiana, and P. nigra (MF), without significant differences among them (p > 0.05); even that they were different compared to the minimum contribution of A. caven in the MF (Table 3).
Temporal litterfall pattern was analyzed considering mean contribution per species (Table 3); these data are referred to monthly means, and did not show a definite pattern.
As already mentioned, A. caven showed a decreasing seasonal production pattern in IF, with a maximum value of 528 kg DM ha -1 season -1 in summer, which differed significantly from the other three seasons (p <0.05).
Production of P. affinis was constant and did not differ significantly among seasons in SF (p >0.05), whereas A. caven and C. ehrenbergiana showed opposite contribution patterns among seasons compared to P. affinis; this is the reason why, when both are considered in combination, the total production follows an annual decreasing rate (Table 3).
Following the same trend, the contribution patterns of P. nigra and C. ehrenbergiana were complementary in MF, i. e., they were opposite among seasons; the same pattern was also observed between C. ehrenbergiana and A. caven in SF. Whereas P. nigra had the ma-    Table 3). The comparison between productions of a single species among forests showed that A. caven (both in MF and IF) also had a decreasing pattern (with a maximum in summer); additionally, the maximum for A. caven in SF was in spring/summer. P. affinis showed a constant production pattern in both MF and SF; however, the recorded contribution was lower in MF, probably due also to the competition of dominant species in these forests (P. nigra and C. ehrenbergiana).
C. ehrenbergiana showed a production pattern typical of deciduous species in both SF and MF, with peaks in autumn in both ecosystems.
Overall, comparison of litterfall productions shows signif icant differences among species and among seasons (Table 3).

Litterfall production by fractions
The comparison of litterfall production per fraction showed that leaves contributed with unusually low values in all cases, exhibiting maximum values for C. ehrenbergiana in MF (mean of 652 kg DM ha -1 yr -1 ) and minimum values for A. caven in MF (mean 22 kg DM ha -1 yr -1 ), with significant differences between them (p < 0.05; Table 4), ranging from 52 to 15% of the litterfall.
Branches showed very dissimilar contribution values, with a maximum of 710 kg DM ha -1 yr -1 produced by P. nigra and a minimum of 76 kg DM ha -1 yr -1 by A. caven Variation in litterfall in a forest successional sequence 417 Table 4. Total mean and litterfall contribution considering the different components (leaves, branches, flowers, fruits, and "others"), per forest and species (kg DM ha -1 yr -1 ±standard error) and leaves/branches relationship (L/B). The Initial forest (IF), secondary forest (SF) and mature forest (MF) are indicated. The relative percentages (below, in bold and italics) are also indicated. Different letters indicate significant differences (p < 0.05) between lines. Fisher's LSD test was used for comparisons of means, n = number of sample (both in MF), with significant differences between them (p < 0.05; Table 4). These contributions represented 59 and 52% of the total production, respectively. Flowers were residual, with the most important production belonging to C. ehrenbergiana in SF (26 kg DM ha -1 yr -1 ), and differing statistically from the contribution of the other species to the total; this maximum corresponded to the highest representation of the flower fraction among species (3%).
The highest fruit production corresponded to A. caven in IF, with a mean of 330 kg DM ha -1 yr -1 . Fruits of A. caven represented 29, 7.7, and 1.9% in IF, SF, and MF, respectively (Table 4).
Finally, contribution of the fraction "other" was highest in P. affinis in SF, C. ehrenbergiana in MF, and A. caven in IF and SF (p > 0.05;

Correlations between litterfall production and climatic parameters
The current analysis was conducted in forests of the Mesopotamian Espinal and no significant correlations were observed between monthly production and climatic factors, such as precipitation (r 2 < 0.01; p > 0.72) and temperature (r 2 = 0.01; p > 0.12).

Total potential return of bioelements (PRB)
Macro-elements (N, Ca, K, Mg, P and S) and micro-elements (Na, Fe, Al, Zn, Mn and Cu) were proportional to the litterfall production of the successional sequence. Total PRB values were 37.1, 140, and 150 kg ha -1 yr -1 in IF, SF, and MF, respectively, whereas micro-elements were 0.5, 1.8, and 1.9 kg ha -1 yr -1 in IF, SF, and MF, respectively ( Table 5). The order of quantitative importance of total bioelements returned by each forest was the following: IF: N > Ca > K > Mg > P > S >> Na > Fe > Al > Zn > Mn > Cu; SF: Ca > N > K > Mg > P > S >> Na > Fe > Al > Mn > Zn > Cu; and MF: Ca > N > K > Mg > P > S >> Na > Fe > Al > Mn > Zn > Cu.

Potential return of bioelements (PRB) per species in each forest
Maximum total PRB was observed for C. ehrenbergiana in MF, with a total of 82.6 kg ha -1 yr -1 of bioelements, whereas the minimum value was recorded for A. caven also in MF, with only 4.2 kg ha -1 yr -1 . In the remaining species, PRB ranged between 11.1 and 68.6 kg ha -1 yr -1 (Table 5).
PRB for A. caven (Table 5)  elements, respectively), but showed the same descending order of contribution of the following elements to the total: N > Ca > K > Mg > P > S >> Na > Fe > Al > Zn > Mn > Cu P. affinis (Table 5) returned 39.6 kg ha -1 yr -1 in SF and 11.1 kg ha -1 yr -1 in MF of total bioelements; however, there were differences in the amounts of nutrient return and in the order of importance of the total return. The descending order of element contribution in SF was: N > Ca > K > Mg > P > S >> Na > Fe > Al > Mn > Zn > Cu Whereas the descending order in MF was: Ca > N > K > Mg > S > P >> Na > Fe > Al > Mn > Zn > Cu The establishment of the primary forests generated displacement of N by Ca concerning the contribution of these two bioelements to the total PRB. C. ehrenbergiana (Table 5) had maximum contribution in SF and MF (82.6 and 68.6 kg ha -1 yr -1 , respectively) and the elements followed the same descending order of contribution in both forests: Ca > N > K > Mg > P > S >> Na > Fe > Al > Mn > Zn > Cu Finally, P. nigra (Table 5) produced 54 kg ha -1 yr -1 in MF, with the bioelement contribution following descending order: N > Ca > K > S > P > Mg >> Na > Fe > Al > Mn > Zn > Cu

PRB considering litterfall fractions and distribution among forests
The maximum contribution of the leaf fraction to the PRB was found in SF, with 86.4 kg ha -1 yr -1 , whereas maximum PRB contribution of branches was found in MF, with only 60.6 kg ha -1 yr -1 (Table 6). PRB contribution of flowers to the PRB was negligible, ranging from 0.8 to 3.5 kg ha -1 yr -1 in IF and SF, respectively, MF showing intermediate values. Fruits had a maximum contribution to PRB in IF (with 12.3 kg ha -1 yr -1 ), followed by SF and MF (with 9.0 and 6.7 kg ha -1 yr -1 , respectively).
The analysis of PRB values among species showed that the minimum and maximum N return by leaves was found in the leaf fraction in MF (0.6 and 17.1 kg N ha -1 yr -1 in A. caven and C. ehrenbergiana, respectively); obviously, the dominant factor seems to be production.
Branches had a minimum return of 1.2 kg N ha -1 yr -1 (A. caven in MF), reaching a maximum of 14.0 kg ha -1 yr -1 (P. nigra also in MF), according to the dominance of the latter species.
Maximum N in flowers was found for P. nigra, the remaining species having a negligible contribution.
Finally, fruits had a maximum contribution of 6.7 kg N ha -1 yr -1 for A. caven in IF.
In general, P return exhibited very low values (ranging between 0.7 and 0.1 kg P ha -1 yr -1 ; Table 6). Maximum contribution of K, Ca, and Mg was provided by the leaf fraction of C. ehrenbergiana (both in SF and MF). K return of the fruit fraction was more than double the contribution of the leaf fraction for A. caven in IF; branches contributed proportionally more Ca, whereas Al and Na were important in the flower fraction.
P. affinis in SF showed an important Al return through the branch and flower fractions, whereas Cu, Na, and Zn were only important in flowers, and Mn in branches; the remaining elements were proportional to the DM amount returned. In addition, in MF this species made significant contributions of Al, Cu, Fe, Na, and Zn through flowers (Table 6).

Annual forest litterfall contribution by species
The litterfall production in forests of the Mesopotamian Espinal studied are below the mean ones recorded for warm-temperate forests, although higher than those found in some meso-thermal subtropical forests. Annual forest litterfall was somewhat lower than the mean global estimation (5,600 kg DM ha -1 yr -1 ; Brinson et al., 1980) for warm-temperate forests. However, the three forests studied exceed the value of 300 kg DM ha -1 yr -1 indicated by Carnevale and Lewis (2001) for other meso-thermal subtropical forests studied in northern Argentina.
In this study, we explored the hypothesis that litterfall contribution increases as ecological succession progresses. A significantly greater annual production, both in SF and MF, than in IF was recorded. Maximum production was expected in MF, but the value found did not differ significantly from SF, probably because the canopy cover has become constant, with little variation between SF and MF. This fact partly conf irms our hypothesis, since productivity of ecosystems of the Mesopotamian Espinal became stable at the SF stage. tion in summer is associated greater evapotranspiration, increasing water stress; this result is in agreement with results reported by Santa Regina et al. (1997).
In agreement with Jeffrey et al. (2007), we observed that litter contribution is a result of complementary addition between P. nigra and C. ehrenbergiana in MF, and between A. caven and C. ehrenbergiana in SF. Overall, individual production of species is responsible for a constant contribution of organic residues throughout the annual cycle in SF and MF, although some significant differences among seasons were found.
These results show that maximum contribution in intermediate and advanced successional stages could be associated to the summer water deficit and species phenology characteristics (Mendoza et al., 2012).

PRB in the forest succession
As the chrono-sequence progressed, Ca replaced N in order of importance, due to the decrease of the leguminous species in the colonizing process after the IF stage, and the contribution to return of Ca by C. erhenbergiana in both SF and MF; then, the influence of the amount produced is much more important that the leaf concentration. The decline of Zn in SF and MF is noticeable, which is displaced by Mn, possibly because of presence of soil Ca (Table 1) and the limitation for plant absorption of Zn (Vogel et al., 2012) in IF. Another interesting process is the displacement of P by S in the MF; this fact indicates the importance of P availability in the first stages of the forest establishment.

PRB per species
In all three forests, A. caven showed the same order of importance of the elements found in the litterfall (Table 5).
According the order indicated for PRB in forests, N is the element of greatest potential return and shows the importance of the N-fixing capacity of the species of the family Fabaceae (Vogel et al., 2012).
The same order of importance of the bio-elements was observed for P. affinis in SF (Table 5), except that N was displaced by Ca, and P by S (under similar soil conditions; Table 1). As indicated above, this change in order could be associated with limitations imposed by dominant species, which conditions N biological fixation and P uptake.
C. ehrenbergiana in both SF and MF showed maximum contribution of Ca (Table 5), which might be attributed to the presence of Ca carbonate phytoliths (Fernández et al., 2005) in leaves; therefore, leaf accumulation of cations means higher return. P. nigra (Table 5) contributed with a maximum value of N in MF, which was consistent with its biological fixing capacity. Another important factor observed ( Table 5) was the high potential return of S and P (displacing Mg), whereas the remaining elements showed the same order of importance described for A. caven. The higher potential returns of S and P might be attributed to a strong absorption of these elements and the limitations for the other species (Vogel et al., 2012), as observed in P. affinis in MF.

PRB considering the litterfall fractions
PRB values found considering the litterfall fractions showed that the leaf fraction was not important for all the species. As observed, leaf fractions have higher contributions on PRB of litterfall in some species than in others (Table 6).
Hence, here L/B ratio could also considered an index because it is related to the PRB values. The greatest leaf/total production ratio values were 0.55 and 0.52 in C. ehrenbergiana (SF and MF, respectively), which is consistent with production of a deciduous species. By contrast, A. caven (a semi-deciduous species) showed limited growth and development in MF, but only contributed with 15 % of leaves (Aceñolaza et al., 2010).
The highest PRB values pertaining to the leaf fraction of C. ehrenbergiana in SF; also a high value for branches was found in P. nigra in MF (Table 6). This shows complementarity not only in the temporal contribution pattern between those species, but also in the types of fractions and could be used as a strategy for bio-element return (Patrício et al., 2012). As mineralization in leaves is more rapid than in branches, bioelements in leaf fraction of C. ehrenbergiana can be easily released, unlike those included in the branch fraction of P. nigra (that are released at a lower rate), with consequences in the nutrition of both species.
Based on our data and analyses, we conclude that: -The productivity of the ecosystems of the Mesopotamian Espinal is rather low and becomes stable when the SF successional stage is established. -Seasonal evolution of litterfall production in the forest successional sequence is determined by the production rate of the dominant species, decreasing in IF, in general, after summer, and showing a constant pattern in SF, and some sporadic significant differences in MF.
-The L/B ratio in forests seems to be a good indicator of the successional stage of semi-deciduous forests of the Mesopotamian Espinal; i. e., the lower the L/B ratio, the higher the forest maturity.
-The contribution of flowers represents a very low percentage of the total annual litterfall production (below 3 %).
-The highest litterfall production by fruit is associated with a colonization strategy (as observed for A. caven in IF).
-Total PRB was proportional to litterfall production in the successional sequence studied, with maximum N values found when leguminous species are dominant; meanwhile C. ehrenbergiana shows the highest contributions of Ca.