Riparian Forest Restoration Drives the Recovery of Soil Chemistry, Microbial Community Structure, and Enzymatic Activity in the Itaipu Reservoir Protection Zone

TL;DR

Restoring riparian forests in Brazil improves soil chemistry, microbial communities, and enzyme activity, showing recovery toward natural forest conditions over time.

Contribution

Demonstrates the long-term recovery of soil properties and microbial communities through riparian forest restoration.

Findings

Soil chemical properties like organic carbon and cation exchange capacity improved with restoration age.

Microbial biomass carbon increased by 60% and enzyme activities related to C, P, and S cycling rose significantly.

Microbial community structure progressively converged toward native forest conditions with restoration duration.

Abstract

Riparian forests play a critical role in protecting soil and water resources and maintaining ecosystem stability. In this study, we evaluated the response of soil chemical and microbial attributes to different stages of riparian forest restoration in the protection zone of the Itaipu Reservoir (Brazil). Soil samples were collected during summer and winter from sites representing four restoration stages (initial, 3, 19, and 30 years), as well as from an adjacent agricultural field and a native forest used as reference systems. We assessed soil chemical properties, microbial biomass carbon, basal respiration, enzymatic activities, and the soil microbial community structure using 16S rRNA gene sequencing. Principal component analysis (PCA) revealed a clear restoration gradient, with older restored sites progressively converging toward the native forest condition. Soil chemical properties showed gradual recovery along the restoration trajectory, with increases in soil organic carbon, cation exchange capacity, and base saturation. In contrast, the availability of P, K, Ca, and Mg declined at early restoration stages and increased with restoration age. Microbial biomass carbon increased by approximately 60% from early restoration to native forest conditions, while metabolic quotients (qCO2) decreased, indicating greater microbial efficiency and reduced metabolic stress. Enzyme activities related to C, P, and S cycling increased by 1.5- to 3-fold with restoration age. Sequencing analyses indicated a progressive convergence of microbial community composition toward that of the native forest, driven by shifts in relative abundance and the enrichment of forest-associated taxa, such as Verrucomicrobia and Acidobacteria, at advanced restoration stages. Overall, long-term riparian forest restoration promoted substantial recovery of soil chemical fertility and microbial community structure and functioning, reinforcing the role of soil microbiota as a sensitive indicator of ecosystem resilience and restoration success.

Figures

Full text

1. Introduction

Riparian forest zones are transitional ecotones that connect terrestrial and aquatic ecosystems and provide essential ecosystem services, including hydrological regulation, sediment and nutrient retention, erosion control, stabilization of stream banks, and buffering of water bodies against diffuse pollution from surrounding landscapes [1,2]. However, these areas are among the most threatened ecosystems worldwide due to deforestation, agricultural expansion, and urbanization, which result in biodiversity loss, soil and water degradation, and reduced ecological resilience [3].

Land-use change from native riparian forest to agriculture or pasture alters soil physicochemical properties and the composition and functioning of soil microbial communities. In Australian riparian ecosystems, the conversion of forest to pasture reduced bacterial taxonomic diversity and shifted communities towards more copiotrophic phyla, whereas revegetation with native species promoted partial recovery of bacterial composition and activity [4]. These environmental consequences of land-use change have driven a global expansion of riparian restoration initiatives, especially in tropical and subtropical regions, where pressures from agriculture are intense and evidence on restoration outcomes remains limited [2].

During the restoration of riparian forests, the reintroduction of tree species and the re-establishment of vegetation promote the gradual recovery of flora, fauna, and aboveground ecological interactions. However, an equally fundamental process unfolds belowground, although it often remains overlooked. Unlike agricultural systems, where there is a continuous external supply of mineral nutrients, restored environments rely almost entirely on the ecosystem’s own capacity to re-establish biogeochemical fluxes through internal nutrient cycling, a process strongly regulated by the soil microbial community. As vegetation becomes established, increased inputs of litter, root exudates, and organic compounds stimulate and reorganize microbial populations, fostering the return of functional groups involved in decomposition, mineralization, and nutrient retention (C, N, P, and S) [5,6,7].

Recent studies show that as forest restoration progresses, microbial biomass, enzymatic activity, and both taxonomic and functional diversity increase, allowing the system to recover its autonomous ability to transform and recycle nutrients, thereby contributing to long-term ecosystem stability and resilience [8,9,10]. Nevertheless, most studies have been conducted in temperate ecosystems, and tropical riparian forests remain comparatively underexplored. Furthermore, few studies have applied integrated approaches that combine soil chemistry, enzymatic activity, and metagenomic analyses, despite their potential to reveal the multidimensional nature of soil recovery [10,11].

The Itaipu Hydroelectric Power Plant, located on the Brazil–Paraguay border, is one of the largest hydropower facilities in the world. Constructed between 1975 and 1984, it created a reservoir of 1350 km^2^, supplying approximately 8.6% of Brazil’s and 86.3% of Paraguay’s electricity demand [12]. Before flooding, large extensions of native riparian vegetation were removed to allow the establishment of pastureland and agricultural fields. The construction and operational infrastructure of the dam, left extensive portions of the reservoir margins unprotected. After inundation, agricultural expansion in the surrounding landscape progressively advanced toward these exposed areas, with cultivation occurring directly adjacent to the reservoir. This combination of vegetation removal and intensive land use accelerated sedimentation processes, threatening to shorten the reservoir’s projected lifespan of 184 years [13]. In response, Itaipu Binacional established one of the largest riparian restoration programs in South America, involving large-scale reforestation, spring rehabilitation, and the implementation of soil conservation measures. These actions aimed both at controlling water erosion in agricultural and pastoral areas and at the physical interception of surface runoff, sediment retention, and biogeochemical filtering, in order to reduce sedimentation, eutrophication, contamination, and pollution processes that threaten the reservoir’s lifespan Despite the magnitude of this initiative, little is known about how restoration age influences the trajectory of soil microbial diversity and functionality in this tropical setting.

Regardless of advances in riparian restoration research, major gaps persist regarding the recovery of belowground processes, particularly in tropical systems undergoing large-scale reforestation. In the Itaipu restoration landscape, where vegetation removal followed by agricultural encroachment substantially altered soil conditions, understanding how microbial communities reassemble through time is essential to evaluate restoration success. The objective of this study was to assess how soil chemical properties, microbial attributes, and bacterial community composition respond across different stages of riparian forest restoration along a chronosequence in the Itaipu reservoir protection zone. We investigated soil microbial community composition, enzymatic activities, and associated soil properties from initial restoration stages to 30-year-old forests. To ensure the robustness of the observed recovery patterns, soil sampling was conducted during both summer and winter, allowing the evaluation of whether restoration-driven trends were consistent across contrasting seasonal conditions. This study provides novel insights into the temporal dynamics of microbial recovery and establishes benchmarks for monitoring large-scale tropical riparian restoration.

2. Materials and Methods

2.1. Areas Description and Soil Sampling

This study was conducted in the protection zone of the Itaipu Hydroelectric Power Plant reservoir, located on the Brazil–Paraguay border in Paraná State, Brazil. The dam was built directly on the Paraná River, and the study sites correspond to riparian forests established along its margins and associated drainage areas. Since the 1980s, Itaipu Binacional has implemented extensive riparian restoration programs throughout the reservoir perimeter. Six sites representing distinct land-use types and restoration stages were selected:

1. Native forest (NF) used as a reference site (25°29′39″ S, 54°22′10″ W);
2. Restored sites at different ages: 30 years (R30; 25°21′40″ S, 54°28′27″ W), 19 years (R19; 25°19′17″ S, 54°25′45″ W), 3 years (R3; 25°21′48″ S, 54°28′37″ W), and an initial restoration stage previously used as pasture (RI; 25°19′23″ S, 54°26′00″ W);
3. An agricultural area (AA) under conventional soybean cultivation in summer and maize cultivation in winter (25°21′50″ S, 54°28′36″ W).

Restoration stages were classified according to the time elapsed since the initiation of riparian forest restoration, based on official records of planting year provided by Itaipu Binacional. Soil samples were collected at 0–10 cm depth during two seasonal campaigns: post-summer harvest (23–24 March 2017) and winter (10–11 August 2017). In each study area and restoration stage, five composite soil samples were collected from spatially distributed points, including one central point and four additional points located 50 m apart in the north, south, east, and west directions. Each composite sample point consisted of five subsamples collected approximately 0.5 m apart, homogenized to obtain ~1 kg of soil, composing 5 replicates of each area (n = 5). This sampling design was consistently applied in both seasons. Samples were sieved (<2 mm), stored at 4 °C, and transported to the Soil Microbiology Laboratory at IDR-Paraná for analysis.

2.2. Soil Chemical Analyses

Soil pH (CaCl_2_), organic matter (OM), exchangeable Ca^2+^, Mg^2+^, Al^3+^, and potential acidity (H^+^ + Al^3+^) were determined following Pavan et al. [14]. Phosphorus (P) and potassium (K^+^) were extracted with Mehlich-1 (0.5 mol L^−1^ HCl + 0.025 mol L^−1^ H_2_SO_4_) and quantified by colorimetry and flame photometry, respectively. Ca^2+^ and Mg^2+^ were extracted with KCl (1 mol L^−1^) and determined by EDTA titration. Al^3+^ was measured by NaOH titration, and cation exchange capacity (CEC) and base saturation (V%) were subsequently calculated.

2.3. Microbial Biomass and Enzymatic Activity

Microbial biomass carbon (MBC) was determined by the fumigation–extraction method [15], with a K_EC_ correction factor of 0.33 [16]. Basal respiration (BR) was measured by incubating 50 g of soil with 0.5 mol L^−1^ NaOH traps for 7 days, and the released CO_2_-C was quantified by titration [17]. The metabolic quotient (qCO_2_) was calculated as CO_2_-C per unit MBC [18]. The activities of arylsulfatase (EC 3.1.6.1), acid and alkaline phosphatase (EC 3.1.3), and β-glucosidase (EC 3.2.1.21) were determined according to Tabatabai [19], and expressed as µg p-nitrophenol h^−1^ g^−1^ dry soil.

2.4. Total DNA Extraction and Sequencing

Total DNA was extracted from 0.5 g of soil using the PureLink™ Microbiome DNA Purification Kit (Invitrogen, Carlsbad, CA, USA), per the manufacturer’s instructions. DNA integrity was confirmed via agarose gel electrophoresis, and DNA concentration was quantified using the Qubit dsDNA BR Assay Kit on a Qubit^®^ 2.0 fluorometer (Invitrogen). The 16S rRNA gene (V4 region) was amplified by PCR using primers described by Pereira et al. [20], pA primers (5′-AGA GTT TGA TCC TGG CTC AG-3′) and pc5B (5′-TAC CTT GTT ACG ACT T-3′). Libraries were prepared with the Nextera XT DNA Sample Prep Kit (Illumina, Inc., San Diego, CA, USA) and sequenced on an Illumina MiSeq platform (Illumina Inc., San Diego, CA, USA) using the MiSeq Reagent Kit V3 (Illumina, Inc., San Diego, CA, USA), in collaboration with the ESALQ-USP Soil Microbiology Laboratory at ESALQ-USP. Sequence quality was assessed using FastQC (v0.10.1, Babraham Bioinformatics, Babraham Institute, Cambridge, UK), read merging was done using FLASH (v1.2.7) [21], and filtering with Seqyclean (v1.3.12) [22]. Final sequences were annotated in MG-RAST (v4.0.3) [23], and taxonomic classification was done using the RDP database with 80% similarity cutoff.

2.5. Statistical Analysis

Soil chemical, microbial biomass, and enzymatic activity data were analyzed using ANOVA, followed by Tukey’s test (p ≤ 0.05) in R-software version 4.4.1 (R Core Team, R Foundation for Statistical Computing, Vienna, Austria) [24] using the ExpDes package [25] and data were analyzed using principal component analysis (PCA) to visualize patterns and group separations among the restoration stages. Metagenomic data were analyzed using STAMP (Statistical Analysis of Metagenomic Profiles) [26] to identify differentially abundant taxa. Group comparisons were performed using Welch’s t-test (p ≤ 0.05), with Benjamini–Hochberg false discovery rate correction applied to account for multiple testing.

3. Results

3.1. Soil Chemical Attributes

The chemical properties of the soil samples collected during summer and winter are presented in Table 1. Significant differences were observed among all land-use types, including the agricultural area (AA), riparian forests at different restoration stages—initial (RI), 3 years (R3), 19 years (R19), 30 years (R30)—and the native forest (NF). Although seasonal variation affected absolute values, the overall patterns were consistent across sampling periods.

Soils from areas at early restoration stages (RI and R3) exhibited chemical profiles closely resembling those of the agricultural area. Specifically, RI and AA displayed comparable values for pH, exchangeable aluminum (Al^3+^), potential acidity (H^+^ + Al^3+^), calcium (Ca^2+^), magnesium (Mg^2+^), sum of bases (SB), and base saturation (V%) in both seasons, indicating limited improvement in soil fertility during early successional stages.

Phosphorus (P) concentrations were highest in AA (41.9 mg dm^−3^ in summer; 22.4 mg dm^−3^ in winter), and declined progressively along the restoration gradient, reaching the lowest levels in NF (2.9 mg dm^−3^ in summer; 5.6 mg dm^−3^ in winter). This trend likely reflects residual fertilizer inputs in AA and reduced external nutrient inputs in less-managed systems. Notably, RI exhibited intermediate P levels, potentially due to its previous use as degraded pastureland, which may influence soil nutrient dynamics differently than former cropland.

Soil organic carbon (C) also followed a consistent pattern: lower levels in AA (17.9–21.5 g dm^−3^) and higher in NF (26.4–35.0 g dm^−3^). The gradual increase in soil carbon along the restoration chronosequence reflects the slow accumulation of organic matter over time. The carbon difference of over 3 g dm^−3^ in AA across three decades illustrates both the potential and challenges associated with restoring soil carbon stocks following land-use change.

Soil pH was highest in AA (5.4 in summer; 5.3 in winter), likely due to liming practices, and declined along the restoration gradient, with NF showing the lowest values (4.9–5.0). These data suggest that vegetation-driven nutrient and organic matter accumulation progressively buffer soil acidity. Cation exchange capacity (CEC) and base saturation followed trends similar to soil carbon. The highest values occurred in NF (CEC: 16.5 cmol_c_ dm^−3^ and V%: 68.7% in summer; CEC: 14.7 and V%: 78.5% in winter), highlighting the role of organic matter in maintaining soil fertility. In contrast, R3 and R19 exhibited lower CEC and V values, particularly in summer, likely reflecting limited litter input the still-developing root systems at intermediate restoration stages.

3.2. Microbial Biomass and Respiration

Microbial biomass carbon (MBC), microbial basal respiration (BR), and the metabolic quotient (qCO_2_) are presented in Table 2. MBC was consistently highest in NF during both seasons and increased progressively along the restoration gradient. Although MBC values fluctuated seasonally, the overall pattern indicated enhanced microbial biomass in mature forested systems.

BR and qCO_2_ were notably higher in AA and RI during the winter collection, suggesting that disturbed or recently restored areas exhibited higher microbial turnover and metabolic stress. The qCO_2_ was lowest in NF, indicating greater microbial efficiency under stable conditions.

3.3. Enzymatic Activity

The activities of alkaline phosphatase, acid phosphatase, arylsulfatase, and β-glucosidase are shown in Table 3. Enzymatic activity varied significantly across sites and seasons, with enzyme-specific response patterns.

Alkaline phosphatase activity showed marked seasonal variation, increasing from summer to winter in both AA (from 67.7 to 147.5 μg pNP h^−1^ g^−1^) and NF (from 212.7 to 359.8). This pattern may reflect enhanced microbial activity or increased substrate availability during the cooler season.

RI exhibited elevated enzyme activity in both seasons, particularly for acid and alkaline phosphatases, despite its early stage of restoration. Arylsulfatase activity showed minor seasonal variation but increased with restoration age, peaking in NF, suggesting a strong linkage with vegetation development and sulfur cycling.

β-glucosidase activity, related to carbon cycling and cellulose degradation, was generally higher in winter and most pronounced in RI and NF. This may indicate greater microbial processing of organic residues during cooler months. AA and early-stage restoration sites, such as R3, exhibited the lowest enzymatic activities, likely due to reduced organic matter inputs and lower microbial functional diversity.

The principal component analysis (PCA) of the summer dataset explained 75.9% of the total variation (F1 = 54.81%; F2 = 21.09%) and revealed a clear separation among land-use types (Figure 1). Agricultural soils (AA) and the most riparian site (R3) clustered in the negative region of F1, closely associated with high soil phosphorus and aluminum. In contrast, restored forests (R19 and R30) grouped near the center of the ordination, while the intermediate stage (RI) was positioned closer to microbial variables such as β-glucosidase, acid phosphatase, and microbial biomass. The native forest (MN) separated along F1’s positive axis, characterized by higher pH, CEC, sulfur, calcium, and basal respiration, reflecting a more mature soil biochemical functioning.

Similarly, the winter PCA captured 74.26% of the total data variability (F1 = 53.59%; F2 = 20.66%) and preserved the patterns observed in summer. Again, AA and R3 occupied the negative F1 region, indicating simplified biochemical profiles associated with disturbance. Restored sites distributed along the central axis, with RI remaining strongly associated with enzyme activities (β-glucosidase, acid and alkaline phosphatases), while R19 and R30 displayed intermediate profiles. The native forest maintained a distinct position, driven by chemical and biochemical attributes indicative of advanced ecosystem recovery. Together, these ordinations demonstrate that soil biochemical and microbial properties progressively differentiate along the restoration chronosequence, with mature forests showing a more complex and functionally integrated signature.

3.4. Soil Microbial Community Composition

Microbial community diversity, based on 16S rRNA sequencing, revealed distinct taxonomic profiles among the studied areas. At the phylum level, the most abundant bacterial groups were Firmicutes, Proteobacteria, and Verrucomicrobia. Sequences from Acidobacteria, Actinobacteria, Bacteroidetes, Cyanobacteria, Fusobacteria, and unidentified bacteria were also detected. Additionally, eukaryotic sequences included Arthropoda and Streptophyta (Figure 2).

Comparisons of bacterial taxa between studied areas and native forest (Figure 3) revealed that agriculture and early restoration sites exhibited significant shifts in bacterial community composition relative to NF. AA showed a large deviation, characterized by reduced abundance of Acidobacteria and Verrucomicrobia, groups typically associated with undisturbed, nutrient-poor forest soils. Conversely, AA had higher proportions of Bacteroidetes and unclassified bacterial groups, indicating disturbance-driven community restructuring. The initial restoration stage (RI) still retained signatures of degradation, with elevated relative abundances of Nitrospirae and Burkholderiales, but also showed gradual shifts toward NF profile, particularly in Actinobacteria and Firmicutes. R3 also exhibited marked shifts in bacterial composition, with several taxa remaining significantly depleted or enriched relative to NF, indicating that early-stage restoration has not yet re-established a microbial community structure comparable to that of mature riparian forest.

Intermediate (R19) and late-stage restoration (R30) demonstrated greater similarity to NF, with R30 exhibiting minimal differences in dominant phyla, suggesting substantial taxonomic recovery. Overall, restoration age was directly associated to increased similarity to native forest microbial profiles, confirming that riparian reforestation promotes the progressive reassembly of bacterial communities over time.

4. Discussion

Understanding the dynamics of soil chemical and microbial attributes during riparian forest restoration is essential for evaluating ecosystem recovery. Key changes involve shifts in nutrient input and internal cycling processes driven by vegetation structure and microbial activity.

The evaluated areas showed clear contrasts among soil chemical attributes along the restoration gradient, reflecting both the legacy of past land use and the inherent constraints of tropical soils. NF exhibited the highest levels of soil organic carbon, CEC, and base saturation (V%), consistent with advanced litter accumulation, deeper rooting systems, and efficient nutrient cycling typical of mature riparian forests [1,2]. In contrast, AA, RI, and R3 exhibited elevated nutrient concentrations, particularly P and K, largely associated to fertilization practices and low organic matter recovery. These patterns align with findings from tropical restoration chronosequences showing that early-stage forests retain soil properties strongly influenced by prior disturbance, particularly in highly weathered Oxisols where acidity, low phosphorus availability, and limited natural CEC prevail [27,28].

The progressive increase in soil C and CEC observed in R19 and R30, approximately 20–35% higher than the agricultural area, reflects the gradual buildup of litter and fine-root biomass, processes widely recognized as key drivers of chemical recovery in tropical riparian and secondary forests [6,29]. As organic residues accumulate and decompose, they contribute to the formation of organo-mineral complexes that enhance nutrient retention, buffer soil acidity, and support the development of a more functionally stable belowground environment. However, the trajectories shown in this study also confirm that soil chemical restoration is slower than vegetation recovery, a pattern commonly reported in tropical forest regeneration studies [8,9]. On highly weathered soils, where mineral surfaces have limited capacity to adsorb nutrients (composed primarily of kaolinite and Fe- and Al-oxides), improvements in fertility depend primarily on organic matter inputs and biological cycling, explaining the slower convergence of R19 and R30 relative to NF. Collectively, these findings reinforce that while vegetation structure may recover within decades, the full restoration of soil chemical functioning in tropical riparian ecosystems may require longer timeframes and sustained organic matter inputs.

The conversion of agricultural areas to forest systems reduces dependence on external nutrient inputs and promotes the retention of organic matter, establishing an internal nutrient cycling regime that directly shapes microbial biomass, activity, and community composition. Functional indicators, including microbial biomass, basal respiration, and enzyme activities, further supported the progressive microbial reassembly observed across the restoration gradient. Native forest soils showed higher microbial biomass carbon and lower metabolic quotients (qCO_2_), suggesting greater stability and efficiency in organic matter use [30]. In contrast, elevated qCO_2_ in agricultural and early restoration sites indicated microbial stress or inefficient functioning. Enzyme activities (acid phosphatase, arylsulfatase, and β-glucosidase) were also higher in advanced restoration stages, driven by organic inputs and improved microclimatic conditions [31].

Notably, despite its early stage of restoration, the RI site exhibited relatively high enzymatic activity for acid and alkaline phosphatases. This pattern may indicate a transient phase of microbial reorganization, where opportunistic microbial groups respond rapidly to organic inputs from initial revegetation. It may also reflect residual nutrient availability from past land use as pasture, temporarily stimulating enzymatic function before full microbial stabilization. This reinforces that soil enzyme activities can respond rapidly to changes in litter deposition even when taxonomic recovery remains incomplete [11]. Soil enzyme activity levels increased with restoration age, showing progressive convergence toward the native forest condition over time, a pattern also observed by Zheng et al. [32], who reported similar increases in microbial enzymatic activity when comparing Picea crassifolia plantations with natural forest soils.

Studies evaluating microbial community responses along forest or riparian restoration chronosequences consistently show that taxonomic reassembly follows a directional trajectory, shifting from disturbance-adapted, copiotrophic communities toward more diverse and functionally specialized assemblages typical of mature forests. In riparian systems, land-use conversion to agriculture generally reduces the relative abundance of oligotrophic groups such as Acidobacteria and Verrucomicrobia while favoring copiotrophic taxa, including Bacteroidetes and Proteobacteria. Revegetation, in turn, promotes the gradual recovery of forest-associated microbial taxa, defined here as bacterial groups consistently reported in mature forest soils and linked to oligotrophic life strategies and organic matter turnover [4]. Comparable patterns have been reported in wetland and forest chronosequences, where early restoration stages retain microbial signatures of degradation, whereas later stages show progressive increases in Acidobacteria, Actinobacteria, and other taxa linked to litter-driven nutrient cycling and stable soil organic matter pools [8,9]. Evidence from mangrove and fire-affected forests further indicates that restoration enhances not only taxonomic composition but also microbial network complexity and functional redundancy, underscoring the sensitivity of microbial communities as indicators of ecosystem recovery [10,11]. When contrasted with temperate riparian systems, our results highlight both general and biome-specific recovery patterns; although the direction of microbial reassembly is broadly consistent across biomes, recovery trajectories in tropical systems tend to be slower and more constrained by soil chemical limitations. Overall, the literature supports the conclusion that microbial taxonomic recovery is gradual but directional, with advanced restoration stages increasingly approaching the structural and functional attributes of native forests.

Our study demonstrated that long-term riparian restoration is essential for the recovery of soil chemical and microbial functioning. The progressive reestablishment of key forest-associated taxa, such as Acidobacteria, Actinobacteria, and Verrucomicrobia, observed in R19 and R30, reflects a clear trajectory toward the native forest community structure, while AA, RI, and R3 still retain signatures of disturbance. This pattern is consistent with PCA ordinations, which positioned restored sites along a gradient of increasing biochemical maturity aligned with higher MBC, enhanced enzymatic activity, and improved soil fertility. Soil chemical and microbial properties are subject to seasonal variation driven by temperature and moisture fluctuations. Although some variables differed in magnitude between summer and winter, the direction and relative separation of restoration stages remained consistent, suggesting that the observed patterns of microbial and biochemical recovery are robust across seasons. Together, these findings indicate that integrating chemical, microbial, and taxonomic indicators provides a robust framework for evaluating restoration success and guiding adaptive management in riparian ecosystems.

The results obtained in this study demonstrate the effectiveness of riparian forest restoration in gradually re-establishing critical soil functions, including nutrient cycling, organic matter accumulation, and microbial activity. Although it is difficult to specify an exact timeframe for the recovery of soil chemical properties and microbial communities, our findings reinforce evidence from previous studies indicating that such recovery generally occurs over decadal timescales. This study advances riparian restoration research by assessing soil chemical and microbial recovery in a reservoir-edge system developed on highly weathered tropical Oxisols. By integrating chemical, enzymatic, and microbial indicators across a 30-year chronosequence, our results show that recovery in tropical riparian zones follows extended, decadal trajectories, underscoring the need for long-term restoration strategies. These improvements are fundamental to ensuring long-term ecological resilience and water quality protection. For Itaipu Binacional, these findings reinforce the importance of ongoing restoration initiatives around the reservoir, as they contribute not only to biodiversity conservation but also to sediment control and the operational sustainability of the hydro-electric plant. By supporting the recovery of soil structure and function, riparian reforestation plays a strategic role in safeguarding ecosystem services essential for the region’s socio-environmental stability.

5. Conclusions

This study demonstrates the progressive recovery of soil chemical and microbiological attributes along riparian forest restoration stages in the Itaipu reservoir protection zone. Restored areas aged 19–30 years showed substantial improvements in soil organic carbon, nutrient balance, microbial biomass, and enzymatic activities, reaching approximately 70–85% of native forest levels for key indicators such as soil organic carbon and microbial biomass carbon, alongside lower metabolic quotients (qCO_2_) indicative of more efficient microbial functioning. These thresholds suggest that advanced restoration stages attain functional conditions increasingly comparable to native forests, providing practical benchmarks for monitoring restoration success. Our findings highlight the importance of long-term investment and monitoring in riparian restoration strategies that prioritize soil health as a foundation for sustainable landscape management. In the context of Itaipu Binacional, the recovery of riparian forests enhances biodiversity, regulates nutrient dynamics, supports water quality, and contributes to the long-term operational sustainability of the hydroelectric system.