In tropical humid Africa, sandy soils under periodically burnt herbaceous savannas exhibit generally low carbon (C) content and nitrogen (N) availability. Savanna afforestation may overcome these limitations through changes in soil functioning, yet these processes still need to be explored. In this study, we investigated whether changes in the composition of soil micro-food web may explain soil C and N cycling increases following savanna afforestation. We conducted a 7-year experiment in Congo including Eucalyptus and N2-fixing Acacia monocultures and Eucalyptus-Acacia mixtures established on former herbaceous savannas. We assessed in each of these modalities the soil attributes: organic C and N, pHH2O, nitrate, ammonium, net C and N mineralization and nitrification rates, along with the abundances of bacteria, fungi, nematodes, ammonia-oxidizing archaea (AOA) and bacteria (AOB) in the top 10 cm layer. Afforestation of savannas with Eucalyptus for timber production increased soil C by 1.7 times, soil net N mineralization rates by 1.9 times and soil inorganic nitrogen by 2.5 times. Mixed AcaciaEucalyptus and Acacia monoculture plantations further improved the rate of net nitrogen mineralization by a factor of 1.4 and soil inorganic N by a factor of 2.3 compared with Eucalyptus monocultures. These changes were associated with a gradual increase in AOA abundance from savanna to Eucalyptus monoculture, Eucalyptus-Acacia mixtures and finally Acacia monoculture. Savanna afforestation resulted in a significant increase in the absolute abundance of bacterial-feeding nematodes by 678 %, but to a decrease in the abundance of both fungal-feeders, and omnivores and predators. Increase in N cycling was positively associated with both AOA and nematode bacterial-feeder abundances, underlining the importance of monitoring micro-food web structure to understand better how land use changes affect soil biogeochemical cycling in the context of tropical afforestation.

The decomposition of crop residues plays a crucial role in nutrient biogeochemical cycles, contributing to the maintenance of soil fertility over successive replanting cycles in tropical perennial crops. In rubber plantations, the main source of natural rubber worldwide, the role of crop residues (hereafter referred to as logging residues) in nutrient cycling has been little studied. This study aimed to gain insight into the dynamics of nutrient release during the decomposition of logging residues in intensively managed rubber plantations, until their complete decay. The impact of contrasting practices of logging residue management on soil nutrient stocks was determined over 3.5 years after clearcutting. The experiment was replicated at 2 sites with contrasting soil conditions in Côte d'Ivoire. Four treatments (replicated in four blocks at each site) were set up with increasing amounts of logging residues at planting. Initial stocks of nutrients, and then decomposition rates and nutrient releases were monitored over 3.5 years after clearcutting, for leaves, twigs, branches and trunks. Soil nutrient stocks were determined in the 0-30 cm layer from clearcutting (i.e., 6 months before planting) up to three years after planting. About 94 % of the total nutrients in logging residues were released over the first three years after replanting rubber trees. All nutrient stocks in the topsoil increased drastically in the treatments with large amounts of residues. Recovery rates were calculated to evaluate soil efficiency in retaining nutrients released from logging residues: for all nutrients (except N), recovery rates were below 100 % at both sites, indicating possible nutrient losses. This indicates that further studies are necessary for a thorough understanding of the fate of nutrients originating from logging residues, particularly regarding deep drainage losses and nutrient uptake by trees. This study highlights the importance of improving practices to maximize nutrient storage and uptake such as i) uniformly spreading logging residues across the plot, ii) sowing legume cover crops and iii) planting trees as early as possible after clearcutting. Leaving logging residues on-site, either partially or fully, could reduce dependence on mineral fertilizers and promote more sustainable natural rubber production.

Description of the subject. In Madagascar, afforestation plays a significant role in providing wood energy for the local population. However, reforestation projects often face low success rates due to biophysical constraints. Objectives. The aim of this study was to evaluate the impact of topographic position on the physico-chemical properties of potential reforestation sites in the central highlands of Madagascar. Method. The study covered 16 sites and analyzed 192 soil samples. The results were compared with deficiency thresholds established for Ferralsols. Results. The soil pHwater varied slightly, from 5.3 to 5.5. The 0-10 cm depth showed significantly higher levels of organic carbon (Corg), total nitrogen (Ntot), available phosphorus (Pdispo), and exchangeable potassium (Kéch) than the 10-40 cm depth. Soils on the slope were significantly poorer in Ntot and Kéch than those on the bottom of the slope, with mean values of 1,3 g N·kg-1 ± 0,07 and 0,09 cmolc·kg-1 ± 0,01 compared to 1,5 g N·kg-1 ± 0,08 and 0,12 cmolc·kg-1 ±0,01, respectively. Of the 16 sites, eight had to be discarded for future reforestation. Of the remaining eight, only two showed soils with little or no Ntot, Pdispo, and Kéch deficiency over 0-10 cm, but both were deficient in Pdispo. Conclusions. This study highlighted the importance of prior soil characterization, choice of topographic position and recommends starter fertilization for successful reforestation. The results of this study are a contribution to the choice of the most appropriate afforestation sites in Madagascar.

Description of the subject. In Madagascar, afforestation plays a significant role in providing wood energy for the local population. However, reforestation projects often face low success rates due to biophysical constraints. Objectives. The aim of this study was to evaluate the impact of topographic position on the physico-chemical properties of potential reforestation sites in the central highlands of Madagascar. Method. The study covered 16 sites and analyzed 192 soil samples. The results were compared with deficiency thresholds established for Ferralsols. Results. The soil pHwater varied slightly, from 5.3 to 5.5. The 0-10 cm depth showed significantly higher levels of organic carbon (C-org), total nitrogen (N-tot), available phosphorus (P-dispo), and exchangeable potassium (K-ech) than the 10-40 cm depth. Soils on the slope were significantly poorer in N-tot and K-ech than those on the bottom of the slope, with mean values of 1,3 g N.kg-1 +/- 0,07 and 0,09 cmolc.kg-1 +/- 0,01 compared to 1,5 g N.kg(-1) +/- 0,08 and 0,12 cmolc.kg(-1) +/- 0,01, respectively. Of the 16 sites, eight had to be discarded for future reforestation. Of the remaining eight, only two showed soils with little or no N-tot, P-dispo, and K-ech deficiency over 0-10 cm, but both were deficient in P-dispo. Conclusions. This study highlighted the importance of prior soil characterization, choice of topographic position and recommends starter fertilization for successful reforestation. The results of this study are a contribution to the choice of the most appropriate afforestation sites in Madagascar.

Soil erosion on agricultural land is a major threat for food and raw materials production. It has become a major concern in rubber plantations introduced on sloping ground. Alternative agroecological crop management practices must be investigated. One aim of our study was to assess the ability of logging residues (i.e., trunks, branches, leaves and stumps of clearcut stands) and of legume cover to mitigate nutrient losses through runoff and soil detachment in a young rubber plantation. The other aim was to investigate the relationships of these nutrient losses with soil structure and soil macrofauna diversity. Runoff and soil detachment were monitored for 3 years using 1-m2 microplots under different practices as regards the management of logging residues and the use or not of a legume cover crop. The planting row, where soil was bare, was the hotspot of soil erosion, with an average annual runoff of 832 mm and annual soil detachment of 3.2 kg m-2. Sowing a legume in the inter-row reduced runoff and soil loss by factors of 8 and 48 respectively, compared to bare soil. Spreading logging residues as well as growing a legume cover almost eliminated runoff and soil detachment (19 mm y-1 and 4 g m-2 y-1 respectively). Nutrient losses were negligible as long as the soil surface was covered by a legume crop, with or without logging residues. Total N losses from soil detachment ranged from 0.02 to 0.2 g m-2 y-1, for example. Spreading logging residues in the inter-rows significantly improved soil structure and soil macrofauna diversity compared to bare soil. Nutrient losses from runoff and soil detachment were negatively correlated with improved soil structure and soil macrofauna diversity. We recommend reducing the width of planting rows from 2 m to 1 m and investigating alternative ways to manage planting rows.

Soil erosion on agricultural land is a major threat for food and raw materials production. It has become a major concern in rubber (Hevea brasiliensis) plantations introduced on sloping ground. Alternative agroecological crop management practices must be investigated. One aim of our study was to assess the ability of logging residues (i.e., trunks, branches, leaves and stumps of a clearcut plantation) and of legume cover (Pueraria phaseoloides) to mitigate N, P and K losses through runoff and soil detachment in a young rubber plantation. The other aim was to investigate the relationships of these nutrient losses with soil structure and soil macrofauna diversity. Runoff and soil loss were monitored for 3 years using 1-m2 plots under different practices as regards the management of logging residues and the use or not of a legume. The monitoring started when rubber trees were one-year-old. The planting row, where soil was bare, was the hotspot of soil erosion, with an average runoff of 832 mm y-1 and soil loss of 3.2 kg m-2 y-1. Sowing a legume in the inter-row reduced runoff and soil loss by 88 % and 98 % respectively, compared to bare soil. Spreading logging residues as well as growing a legume cover almost eliminated runoff and soil detachment (19 mm y-1 and 4 g m-2 y-1 respectively). Nutrient losses were negligible as long as the soil surface was covered by a legume crop, with or without logging residues. Total N loss from soil detachment ranged from 0.02 to 0.2 g m-2 y-1, for example. Spreading logging residues in the inter-rows significantly improved soil structure and soil macrofauna diversity compared to bare soil. Nutrient losses from runoff and soil detachment were negatively correlated with improved soil structure and soil macrofauna diversity. We recommend investigating alternative ways to manage planting rows.

Planting nitrogen (N) fixing species can provide substantial levels of N for tree growth, but it can also result in large nutrient losses through deep drainage, threatening soil fertility. Nutrient losses through deep leaching have been little studied in tropical forest plantations and comprehensive studies are needed before planting these species on a large scale. We assessed nutrient fluxes in soil solutions collected in monocultures and mixed plantations of Acacia mangium and Eucalyptus grandis in Sao Paulo state (Brazil) on Ferralsols fertilized with potassium, phosphorus and lime. Soil solution sampling began after replanting the trees in an experiment that had already been conducted for six years in a first rotation. We collected soil solutions beneath the forest floor and at depths of 30, 100 and 300 cm, and determined nutrient concentrations monthly for 3.5 years. Drainage fluxes were calculated at the depth of the lysimeters by modeling the water fluxes with Hydrus 1D. The N concentrations in topsoil solutions were generally higher in Acacia than in Eucalyptus monocultures, confirming that the introduction of Acacia increases soil N availability. However, these differences were no more observed at 100 cm depth and total N leaching at 300 cm depth was low in all the treatments, with an average of 4.8 kg N ha- 1 yr -1, probably due to the rapid root system development and high nutrient requirements of these fast-growing trees. Leaching fluxes of K+, Cl-, Ca2+ and Mg2+ peaked in the upper soil layers after fertilizer appli-cations, but the fluxes drastically decreased in deep soil layers. The low fluxes of dissolved nutrients in deep gravitational solutions in our study suggest that the risks of nutrient leaching losses are low in fast-growing plantations established on deep Ferralsols. Our results show that planting N-fixing trees can enhance the N availability for plants without producing large leaching losses, which confirms the interest of mixed plantations with N-fixing trees.

We hypothesized that the nitrogen-fixing tree Acacia mangium could improve the growth and nitrogen nutrition of non-fixing tree species such as Eucalyptus. We measured the N-mineralization and respiration rates of soils sampled from plots covered with Acacia, Eucalyptus or native vegetation at two tropical sites (Itatinga in Brazil and Kissoko in the Congo) in the laboratory. We used a bioassay to assess N bioavailability to eucalypt seedlings grown with and without chemical fertilization for at least 6 months. At each site, Eucalyptus seedling growth and N bioavailability followed the same trends as the N-mineralization rates in soil samples. However, despite lower soil N-mineralization rates under Acacia in the Congo than in Brazil, Eucalyptus seedling growth and N bioavailability were much greater in the Congo, indicating that bioassays in pots are more accurate than N-mineralization rates when predicting the growth of eucalypt seedlings. Hence, in the Congo, planting Acacia mangium could be an attractive option to maintain the growth and N bioavailability of the non-fixing species Eucalyptus while decreasing chemical fertilization. Plant bioassays could help determine if the introduction of N2-fixing trees will improve the growth and mineral nutrition of non-fixing tree species in tropical planted forests.

Sustainable management of highly productive eucalypt plantations requires the application of fertilizers to balance nutrient exports associated with biomass removal every 6-7 years. Although deep leaching after clear-cutting is an important component of input-output budgets of nutrients in forest soils, accurate quantifications are rare in tropical plantations. Our study aimed to assess the consequences of management practices in highly productive eucalypt plantations on nutrient losses by deep leaching in two hillslope positions with contrasting soil types. Soil solutions were continuously collected using lysimeters down to a depth of 3 m, in the last year before the clear-cutting, then in the first 3.5 years after replanting. Concentrations of the main cations and anions were determined monthly and fluxes of gravitational solutions at the depths where the lysimeters were installed were estimated using the Hydrus 1D model, calibrated in situ. Stand productivity was high in both landscape positions with a mean basal area at harvest of 35 m2 ha-1 at the bottom of the slope and 27 m2 ha-1 at the top of the slope. Soil solution chemistry in the topsoil was highly responsive to management practices with sharp increases in ionic charges after clearcutting and fertilizer applications. While leaching fluxes of mobile ions (especially N-NO3-, K+ and Mg2+) reached values greater than 30 kg ha-1 yr-1 at a depth of 1 m after clear-cutting, they remained lower than 3 kg ha-1 yr-1 at a depth of 3 m both in the upper and lower hillslope positions throughout the rotation. Our study suggests that splitting fertilizer applications may not be necessary in tropical Eucalyptus plantations established in very deep tropical soils.

A sustainable management of highly productive eucalypt plantations requires the application of fertilizers to balance nutrient exports associated with biomass removal every 6-7 years. Although deep leaching after clear-cutting is an important component of input-output budgets of nutrients in forest soils, accurate quantifications are rare in tropical plantations. Our study aimed to assess the consequences of management practices in highly productive eucalypt plantations on nutrient losses by deep leaching comparing two Ferralsols with contrasting clay contents (40% vs 20% clay) in the same commercial plot. Soil solutions were continuously collected using lysimeters down to a depth of 3 m, in the last year before the clear-cutting, then in the first 3.5 years after replanting. Concentrations of the main cations and anions were determined monthly and fluxes of gravitational solutions at the depths where the lysimeters were installed were estimated using the Hydrus 1D model, calibrated in situ. Stand productivity was high in both soil types with a mean basal area at the harvest of 35 m2 ha-1 in the clayey soil and 27 m2 ha-1 in the sandy soil. Soil solution chemistry in the topsoil was highly responsive to management practices with sharp increases in ionic charges after clearcutting and fertilizer applications. While leaching fluxes of mobile ions (especially N-NO3-, K+ and Mg2+) reached values greater than 30 kg ha-1 yr-1 at a depth of 1 m after clear-cutting, they remained lower than 3 kg ha-1 yr-1 at a depth of 3 m in both soil types throughout the rotation. Our study suggests that it is not necessary to split fertilizer applications in tropical Eucalyptus plantations established in very deep tropical soils, which could reduce the cost of silvicultural practices by greatly reducing tractor use in the plots.

Eucalyptus is the tree most widely planted in tropical countries to satisfy growing demand for wood products, but high yields require high fertilizer inputs. Introducing N2-fixing trees (NFT), such as Acacia mangium, has been proposed to improve soil fertility and aboveground tree biomass in Eucalyptus plantations. In addition to N inputs, NFT species may increase plant P nutrition through increased rates of organic P (Po) cycling. However, the positive effect of acacia on soil P availability and plant P nutrition was found to vary substantially between sites. The ability of acacia to improve P bioavailability might mainly depend on Po sequestration in microbial biomass, preventing Po mineralization by phosphatases and efficient Po recycling. This hypothesis was tested at two tropical sites, Itatinga (Brazil) and Kissoko (Congo) by measuring inorganic phosphate (Pi), Po and enzyme-labile Po in bicarbonate extracts from the topsoil collected from plots with Eucalyptus, acacias, or native vegetation. We used bicarbonate enzyme-labile Po after soil autoclaving as an indicator of microbial Po, and a Eucalyptus bioassay to measure the actual P bioavailability for Eucalyptus seedlings. At Itatinga, bicarbonate-Pi was very low, while Po was the main P form. Enzyme-labile Po was very weak in intact soils and high in autoclaved soils, indicating high immobilization in microbial biomass. At Kissoko, Po was highly enzyme-labile in both intact and autoclaved soils, especially from acacia plots, suggesting very low Po immobilization in microbial biomass. Growth and P accumulations in Eucalyptus seedlings were low in all soils at Itatinga and were the highest in Eucalyptus plants grown in acacia soils at Kissoko. Our results highlight the potential of acacia trees for improving P bioavailability for other tree species if labile Po enrichment in the soil provided by this N2-fixing tree is not locked into the microbial biomass.

Soil erosion causes major problems of land degradation in agricultural systems leading to losses of soil fertility. Rubber tree is one of the main tropical perennial crops with about 13 million hectares of plantations worldwide in 2018. In the early stage of a rubber plantation, soil is especially vulnerable to degradations given the low canopy cover and heavy soil surface disturbance related to clear-cutting of the previous plantation. This study aims at assessing runoff and soil losses as well as understanding the main soil factors influencing soil erosion in a young rubber plantation in Côte d’Ivoire. We intensively measured soil runoff, soil detachment, soil structure maintenance and soil macrofauna for 2.5 years under different managements of logging residues and the use or not of a legume cover crop. The results showed that the restitution of logging residues has reduced runoff by 6 and soil losses by 14 compared to plot without logging residues, over the study period. The planting line where soil is kept bare was by far the most critical area in term of soil erosion. The restitution of logging residues significantly improved soil structure maintenance as well as soil macrofauna diversity. We found strong relationships between runoff, soil losses, soil structure and soil macrofauna diversity. These results evidence that the restitution of logging residues and the sowing of cover crop are appropriate agroecological practices in young rubber plantations. Our results suggest that keeping a cover in the planting line could be the most relevant lever to limit soil erosion in the context of the study.

Soil health is defined as the soil's capacity to deliver ecosystem functions within environmental constraints. On tree plantations, clear-cutting and land preparation between two crop cycles cause severe physical disturbances to the soil and seriously deplete soil organic carbon and biodiversity. Rubber, one of the main tropical perennial crops worldwide, has a plantation life cycle of 25 to 40 years, with successive replanting cycles on the same plot. The aim of this study was to assess the effects of clear-cutting disturbance on three soil functions (carbon transformation, nutrient cycling and structure maintenance) and their restoration after the planting of the new rubber crop, in two contrasting soil situations (Arenosol and Ferralsol) in Côte d'Ivoire. In this 18-month diachronic study, we intensively measured soil functions under different scenarios as regards the management of logging residues and the use or not of a legume cover crop. We investigated the relationship between soil macrofauna diversity and soil heath. At both sites, clear-cutting and land preparation disturbed carbon transformation and nutrient cycling significantly and, to a lesser extent, structure maintenance function. When logging residues were applied, carbon transformation and structure maintenance functions were fully restored within 12 to 18 months after disturbance. By contrast, no restoration of nutrient cycling was observed over the study period. A legume cover crop mainly improved the restoration of carbon transformation. We found a strong relationship (P ≤ 0.001; R2 = 0.62–0.66) between soil macrofauna diversity and soil health. Our overall results were very similar at the two sites, despite their contrasting soil conditions. Keeping logging residues in the plots and sowing a legume in the inter-row at replanting accelerated the restoration of soil functions after major disturbance caused by clear-cutting and land preparation. Our results confirm the necessity of taking soil macrofauna diversity into account in the management of tropical perennial crops.

Rubber trees are the main source of natural rubber (NR). The area occupied by rubber plantations rose from 3.9 million ha in 1961 to 12.5 million ha in 2018. Both the expansion of rubber plantations in marginal zones (prone to biotic and abiotic stress), and long-term rubber tree cultivation in traditional areas, raise questions about the sustainability of NR production in a context of climate change. Our study set out to gain insights into the biogeochemical cycles in rubber plantations, for a better matching of fertilizer inputs to the dynamics of nutrient demand throughout rubber tree growth. Nutrient accumulation in tree biomass is a major component of the biological cycle in tree plantations. We studied the dynamics of biomass and nutrient accumulation in two chronosequences covering the whole lifespan of a plantation in Ivory Coast managed on a sandy soil at the SAPH site, and one on a clayey soil at the SOGB site. In total, 56 trees were destructively sampled in 2-, 5-, 20- and roughly 40-year-old stands. While the use of allometric relationships is common for estimating nutrient stocks in planted forests, this study was the first to provide allometric equations predicting nutrient stocks in rubber tree components. Allometric models were applied to the inventory of 4 commercial stands, for each age at each site, to estimate stand biomass and nutrient stocks. The current annual increments of nutrient stocks in tree biomass peaked between 2 and 5 years after planting. They reached 80 kg ha(-1) yr(-1) for N, 14 kg ha(-1) yr(-1) for P and 34 kg ha(-1) yr(-1) for K at SAPH (53, 7, and 39 kg ha(-1) yr(-1) respectively at SOGB), which highlighted the importance of an appropriate fertilization schedule for young rubber trees. At the clear-cut age (38-40 years), the amounts of nutrients accumulated in tree biomass were 970 kg N ha(-1), 188 kg P ha(-1), 366 kg K ha(-1), 941 kg Ca ha(-1) and 255 kg Mg ha(-1) on the sandy soil at SAPH (907,118, 629 1499, and 375 kg ha(-1) respectively on the clayey soil at SOGB). Contrasting soil properties and management practices at the two sites had a much greater effect on the amounts of P, K, Ca and Mg accumulated in the trees than on N accumulation. Logging practices in rubber plantations can lead to considerable nutrient exports on poor tropical soils. Harvest residues should be distributed uniformly in the plots so that the roots of young trees can quickly gain access to the nutrients released during decomposition.

Potassium (K) is essential for a wide range of physiological functions in plants, and a limiting element for wood productivity in numerous forest ecosystems. However, the contribution of each of the K-sensitive physiological processes to the limitation of wood productivity is poorly known. In trees, K deficiency acts both on the source and the sinks of carbon making it difficult to disentangle its effects on wood productivity. Here, we review the literature dealing with the influence of K-limitation on tree physiological processes. Results from extensively studied tropical Eucalyptus plantations are used to illustrate the physiological processes the most impacted by K deficiency. We identify the main processes that limit the availability of K to the trees and influence the circulation of K ions in the ecosystem. Then, we describe the influence of K bioavailability on carbon assimilation, the water economy of trees, and carbon partitioning. We conclude this review by identifying the main priorities towards the process-based modelling of the influence of K on the carbon and water cycles in forest ecosystems. For each process modelling priority, we identify options that could be used in the current conceptual framework of most eco-physiological models.

The determination of the mechanisms and extent of soil mineral weathering can be challenging, and the caveats reside in 1) difficulty identifying minerals that are actually involved in weathering reactions, 2) non-stoichiometric release of cations during weathering processes due to coupled dissolution, precipitation and transformation reactions and, 3) impact of vegetation activity on elemental cycles in upper soil horizons.To better characterize mechanisms controlling mineral weathering in soils and trace the evolution of B concentration and isotope ratios during chemical weathering, quantitative mineralogical analyses were coupled to B isotopes in a group of minerals (biotite, muscovite, K-feldspar and albite). Samples were selected along an Alocrisol (Alumic Cambisol, WRB FAO) soil profile from the bedrock (at 130 cm depth) up to 20 cm depth, developed on granitic bedrock in the Breuil-Chenue forest (France). The samples consist of residual primary minerals associated with weathering secondary phases (vermiculite, kaolinite…) in varying proportions.The B isotopic compositions of the most pristine minerals span a very narrow range of values (around −31‰), whereas all secondary phases point to a much heavier value (around −16‰), regardless of mineralogy. Our results also show a mineral-dependent evolution of B concentration or isotopic composition as weathering progresses: no variation is observed during dissolution of K-feldspars; B behaves like a very mobile element in micas (biotite and muscovite), whereas it concentrates in weathered products derived from albite. However, rates of B concentrations and changes in isotopic compositions appear to be much faster than those inferred from mineralogy or major element concentrations determined by XRD and bulk chemical analyses, respectively. These results indicate that B is involved in very early weathering reactions and raises the question of its actual location in the structure of the various soil minerals as well as its pathway to solution.

Many forest ecosystems are developed on acidic and nutrient-poor soils and it is not yet clearly understood how forests sustain their growth with low nutrient resources. In forestry, the soil chemical fertility is commonly defined, following concepts inherited from agronomy, as the pool of plant-available nutrients in the soil at a given time compared to the nutritional requirement of the tree species. In this two-part study, Part 1 (Hansson et al., 2020) showed, through the compiled dataset of 49 forest ecosystems in France, Brazil and Republic of Congo, the limits of this definition of soil chemical fertility in forest ecosystem contexts. In this study (Part 2), we investigated the nutrient pools and fluxes between the different ecosystem compartments at 11 of the 49 sites in order to better characterize the role of the biogeochemical cycling of nutrients in the chemical fertility of forest ecosystems, and in particular the roles of the biological and geochemical components of biogeochemical cycling. The analysis of our dataset shows different types of biogeochemical functioning. When the geochemical component (inputs through mineral weathering and/or atmospheric inputs, capillary rise) is predominant, sufficient nutrients are provided to the plant-soil system to ensure tree nutrition and growth. Conversely, when the geochemical component of the cycle brings too few nutrients to the plant-soil system, the biological component (litterfall, plant internal cycling) becomes predominant in tree nutrition and growth. In the latter case, forest production may be high even when pools of nutrients in the soil reservoir are low because small but active nutrient fluxes may continuously replenish the soil reservoir or may directly ensure tree nutrition by bypassing the soil reservoir. This study highlights the necessity to include biogeochemical cycling and recycling fluxes in the definition and diagnosis methods of soil chemical fertility in forest ecosystems. We show that the chemical fertility is not only supported by the soil in forest ecosystem but by the sum of all the ecosystem's compartments and fluxes between these pools.

Forest soil fertility can be defined as a combination of physical, chemiCal and biological factors characterising the biomass production capacity of the soil. However, numerous ecological variables affect tree growth and the aim of the present study was to investigate the specific influence of soil chemical properties on tree productivity at 49 acidic forest sites. A standardized tree productivity index based on tree height expressed as dominant height of the studied stand divided by maximum tree height observed at the same age for the same species in the same climatic region was firstly computed at each site. This index is assumed to limit the influence of species, ages and climate. A soil database was also compiled with data on soil properties from 47 temperate (France) and two tropical (Congo, Brazil) sites. Data included seven tree species, varying in age from 1 to 175 years. Commonly used indicators such as C:N ratio, soil pH, as well as available and total pools of soil nutrients were compared to the standardized tree productivity index, to find the most reliable indicator(s). Nutrient pools at fixed mineral soil depths (down to 100 cm) were used, as well as (for 11 stands) the depth comprising 95% of fine roots. Our results show that none of the common soil chemical parameters tested in this paper could individually explain stand productivity. Combinations of different parameters were also tested using PCA and they could better explain the variability of the data set but without being able to separate the sites according to their standardized tree productivity index. Moreover, random Forests performed on our dataset were unable to properly predict the standardized tree productivity index. Our results reinforce the idea that the influence of the soil chemical fertility on stand productivity is complex and the soil chemical parameters alone (individually or combined) are poor predictors of tree productivity as assessed by the Ho:H. index. In this paper we focused on static soil chemical indicator and more dynamic indictors, such as nutrient fluxes involved in the biogeochemical cycles, could better explain stand productivity. A companion paper (Legout et al., 2020) focuses on the connection between productivity and different components of the biogeochemical cycle, using data from 11 of the stands presented in this paper.