Aspects of the phosphorus cycle in Hartbeespoort Dam (South Africa) Phosphorus loading and seasonal distribution of phosphorus in the reservoir

Hartbeespoort Dam, a hypertrophic, warm monomictic impoundment in South Africa, receives extremely high phosphorus loads (14.6–25.9 g m^–2 a^–1) that are dominated by point source discharges from municipal wastewater treatment works. The reduced state of the phosphorus discharged from the works has led to the dominance of the dissolved phosphorus pool by low molecular weight orthophosphates which are analytically detectable as soluble reactive phosphorus (SRP; 60% of total phosphorus pool). Seasonality in the in-lake total phosphorus pool is regulated by a combination of abiotic and hydrological processes; biotic processes appear to play a minor role. Mass balance calculations indicate that between 62 and 77% of the annual total phosphorus inflow load is retained within the impoundment each year.     

The phosphorus status of sediments in a hypertrophic impoundment (Hartbeespoort Dam): implications for eutrophication management

The phosphorus status and distribution of sediments in a hypertrophic water supply reservoir (Hartbeespoort Dam) were investigated, with a view to assessing the role of sediments in counteracting the effects of reduced external phosphorus loading as a restoration measure. In comparison with similar water bodies in South Africa, the sediments in Hartbeespoort Dam contained high levels of both total and potentially mobile phosphorus. The potentially mobile fraction constituted about 60% of the total phosphorus content of the sediments, compared with about 11% in other reservoirs. The excessive eutrophication of Hartbeespoort Dam is clearly reflected in the phosphorus status of the sediments. Sediment distribution in the impoundment was found to be extremely heterogeneous, due to the combined influences of morphometry, hydrology and an imbalance in the nutrient loads entering via rivers at remote points in the water body. It is concluded that sufficient mobile phosphorus has accumulated in the sediments to prolong the response time of the impoundment to phosphorus load reductions. Since phosphorus release from sediments is dependent on dynamic processes not addressed in this study, the extent of the delays in trophic response to load reduction cannot be estimated.     

Effect of rainstorms on heterotrophic bacterial activity in a hypertrophic African lake

Hartbeespoort Dam is a hypertrophic man-made lake which is located in the Transvaal Province of South Africa. This region has recently experienced its most severe drought of the century. However, on three occasions in the summer rainy seasons of 1984 and 1985, major rainfalls (> 50 mm) occurred which caused large inflows to the lake. Inflowing river water entered as a density current causing marked silting of the water. Within the epilimnion (0–10 m) prior to these rainfalls there was usually no variation of bacterial numbers with depth, but heterotrophic bacterial activity (glucose uptake) decreased with depth concomitant with primary production. With the increased river inflow bacterial numbers did not increase but bacterial activity at the bottom of the epilimnion (10 m) increased to as high as 2.7 µg C l^–1 h^–1 in January 1985, reversing the depth profile of bacterial activity within the epilimnion. This resulted in decreased glucose concentrations (K_t + S_n) and turnover times. Heterotrophic activity per cell increased by between 2.5 and 5 times. These data demonstrate that storm events are important phenomena causing short-term changes in the metabolic activity of planktonic heterotrophic bacteria in lakes.     

A new biomonitoring protocol to determine the ecological health of impoundments using artificial substrates

The colonisation potential of macroinvertebrate families was measured using artificial substrates. These substrates were tested on three highveld impoundments. Bronkhorstspruit Dam was an amictic, mesotrophic impoundment, while Hartbeespoort and Roodeplaat Dams were both monomictic, eutrophic impoundments. Bronkhorstspruit Dam had the best water quality and the smallest algal population, comprising mainly diatom species. Hartbeespoort Dam had the most saline water (TDS 96–400 mg/l; NH₄ <0.004–0.218mg/l; NO₂+NO₃ 0.109–2.776mg/l; PO₄ <0.04–0.06mg/l), while Roodeplaat Dam had the highest nutrient concentrations (TDS 200–373mg/l; NH₄ <0.04–0.933mg/l; NO₂+NO₃ <0.04–2.310mg/l; PO₄ 0.041–0.472 mg/l). Both Hartbeespoort and Roodeplaat Dams had large algal populations that resulted in very alkaline water during summer. Microsystis dominated the phytoplankton population in these two impoundments during summer and autumn, while green algae and diatoms were dominant during winter and spring. Ecological health was determined using two different biotic indices. Two modifications of the Belgian Biotic Index and a modification of SASS4 (SASSD) were tested. Three health classes were suggested for SASSD scores and ASPT values. It is concluded that artificial substrates with SASSD (and ASPT) as a biotic index can be used to determine the biological health of impoundments, but that further refinement of the index is needed.     

Influence of Cyanobacterial Hyperscum on Heterotrophic Activity of Planktonic Bacteria in a Hypertrophic Lake

The response of the planktonic heterotrophic bacterial community to the buildup and breakdown of a semipermanent, crusted, floating cyanobacterial mat, or hyperscum, that covered 1 to 2 ha was studied in a hypertrophic lake (Hartbeespoort Dam, South Africa). The initial response of bacteria in the main basin to the release of dissolved organic carbon (DOC) from the hyperscum 1 km away was an increase in activity per cell from 35 × 10⁻¹² to 153 × 10⁻¹² μg of C cell⁻¹ h⁻¹ for total cell counts, while activity per cell for metabolically active cells increased from 19 × 10⁻¹¹ to 85 × 10⁻¹¹ μg of C cell⁻¹ h⁻¹. No major population growth occurred at this stage. Later, with the continuous supply of DOC from the hyperscum, total bacterial numbers increased from 6.6 × 10⁶ to 20 × 10⁶ cells ml⁻¹, while the activity per cell declined. Metabolically active bacteria followed the same trend. Shorter-term DOC increases caused only increases in bacterial activity per cell. The data from Hartbeespoort Dam demonstrate an interesting and little-documented mechanism by which aquatic bacteria respond to increased DOC concentration and which may be universal for aquatic systems.     

Health and chemical burdens of fish species from polluted and hyper-eutrophic freshwater ecosystems in South Africa

Three aquatic ecosystems in South Africa, the Hartbeespoort, Klipvoor and Bospoort Dams, are classified as hyper-eutrophic, because of high nutrient loads and chemical pollution. Water and two fish species, Clarias gariepinus and Cyprinus carpio, were collected from these dams to assess the impact of eutrophication and chemical pollutants on their health status. Water and muscle samples were analysed for organic and inorganic chemicals. Condition factor was determined and a necropsy performed to note any macroscopic abnormalities. A histology-based fish health assessment was done on the liver, kidney, gills and gonads. A number of fish from the three dams exhibited livers with fatty change and focal discoloration, skin lesions and parasites within the visceral cavity. The prevalence and severity of histopathology in the liver resulted in higher liver index values than the index values for kidneys and gills. Aluminium, silicon and chromium were detected in the water and muscle tissue. The DDT metabolite p,p’-DDE was present in both species, as well as in fish from the reference site, Marico-Bosveld Dam. Only C. gariepinus from Hartbeespoort Dam had p,p’-DDD levels higher than 5 µg g^?1 per edible portion. Water from hyper-eutrophic dams adversely affects the health of freshwater fish.     

Variety specific relationships between effects of rhizobacteria on root exudation,growth and nutrient uptake of soybean

Aims

It has been increasingly recognized that only distal lower order roots turn over actively within the <2 mm fine root system of trees. This study aimed to estimate fine root production and turnover rate based on lower order fine roots and their relations to soil variables in mangroves.

Methods

We conducted sequential coring in five natural mangrove forests at Dongzhai Bay, China. Annual fine root production and turnover rate were calculated based on the seasonal variations of the biomass and necromass of lower order roots or the whole fine root system.

Results

Annual fine root production and turnover rate ranged between 571 and 2838 g m^?2 and 1.46–5.96 yr^?1, respectively, estimated with lower order roots, and they were increased by 0–30 % and reduced by 13–48 %, respectively, estimated with the whole fine root system. Annual fine root production was 1–3.5 times higher than aboveground litter production and was positively related to soil carbon, nitrogen and phosphorus concentrations. Fine root turnover rate was negatively related to soil salinity.

Conclusions

Mangrove fine root turnover plays a more important role than aboveground litter production in soil C accumulation. Sites with higher soil nutrients and lower salinity favor fine root production and turnover, and thus favor soil C accumulation.

     

Heterotrophic bacterial activity and primary production in a hypertrophic African lake

The relationship between heterotrophic bacteria and phytoplankton in the epilimnion (0–10 m) of hypertrophic Hartbeespoort Dam, South Africa, was examined by statistically analyzing three years of parallel measurements of heterotrophic bacterial activity (glucose uptake) and phytoplankton particulate and dissolved organic carbon production. Algal biomass ranged between 4.0 and 921.1 mg Chl a m^-3 at the surface. Primary production varied between 69.5 and 3010.0 mg C m^-2h^-1 while algal production of dissolved organic carbon (EDOC) ranged from 2.5 to 219.2 mg C m^-2h^-1. Bacterial numbers reached a summer peak of 44.23 × 10⁶ cells ml^-1 in the first year and showed no depth variation. The maximum rate of glucose uptake, V_max, reached a peak of 5.52 g C l^-1h^-1. V_max, maximum glucose concentration (K_t + S_n) and glucose turnover time (T_t) were usually highest at the surface and decreased with depth concomitant with algal production. At the surface, V_max was correlated to EDOC (r = 0.59, n = 67, p < 0.001) and primary production (r = 0.71, n = 70, p < 0.001). At 5 and 10 m, V_max was correlated to integral euphotic zone (~ 4 m) algal production and bacterial numbers. Glucose turnover time was inversely related to integral algal production (r = -0.72, n = 70, p < 0.001) and less strongly to bacterial numbers. The data indicated that although bacterial numbers and biomass were low relative to algal biomass in this hypertrophic lake, the heterotrophic bacteria attained high rates of metabolic activity as a result of enhanced algal production of available organic carbon.     

Environmental factors favouring the formation of Microcystis aeruginosa hyperscums in a hypertrophic lake

Hyperscums are crusted buoyant mats of densely packed cyanobacteria, often decimeters thick, that persist for periods of weeks to months at the same site. In Hartbeespoort Dam, a hypertrophic lake in South Africa, hyperscums of the cyanobacterium Microcystis aeruginosa that cover more than a hectare and contain up to 2 tons of chlorophyll a typically form in winter and persist for 2–3 months. This paper reports on the environmental conditions that favour hyperscum formation.Reynolds & Walsby (1975) postulated that cyanobacterial bloom formation depended on the coincidence of three preconditions: a pre-existing population, a significant proportion of the organisms having positive buoyancy, and turbulent mixing that is too weak to overcome the tendency of the cells to float. This model of bloom formation is evaluated in the context of hyperscums, based on a case-study from Hartbeespoort Dam. We examine the occurrence of hyperscums and the dynamics of their formation and breakdown in relation to diurnal and seasonal changes in the wind regime and in relation to the population dynamics and buoyancy of Microcystis. We conclude that Reynolds and Walsby's preconditions are essential but not sufficient to explain hyperscum formation. The additional preconditions are prolonged low speed wind regime, suitable lake morphometry, large cyanobacterial standing crops, and high insolation. The rare co-occurrence of these conditions make hyperscums an uncommon phenomenon, but with increasing eutrophication worldwide the frequency and distribution of hyperscum occurrence are likely to increase.     

Seasonal changes in the dissolved free amino acid and DOC concentrations in a hypertrophic African reservoir and its inflowing rivers

Dissolved free amino acid (DFAA) concentration and composition and dissolved organic carbon (DOC) concentration were measured over 16 months at three depths in hypertrophic Hartbeespoort Dam, South Africa and in its two perenially inflowing rivers. The range of DFAA concentrations in the reservoir and both rivers were similar with dominant DFAA consisting of serine, glycine, alanine and ornithine in all three systems. The range of DOC concentrations in the rivers was 1.5–11.1 mg l^–1, the major river (Crocodile) having about twice the DOC concentration of the Magalies River. The DFAA/DOC ratios ranged between 0.02–1.1% in the Crocodile River and 0.13–3.7% in the Magalies River. DFAA and DOC concentrations were positively correlated to the Magalies River flow, but for the Crocodile River, which received domestic and industrial effluents, DOC was inversely correlated to flow. The source of DFAA in both rivers was mainly terrestrial, in contrast to the main DOC source in the Crocodile River which was the effluents. The DFAA load of the Crocodile River ranged between 0.22 and 208 kg C d^–1.DOC (5.0–24.8mg l^–1) in Hartbeespoort Dam generally decreased with depth but DFAA (15–4800 nmol l^–1) concentration showed no clear trend. The DFAA/DOC ratios varied between 0.02 and 2.9%. DFAA concentrations were correlated (r = 0.3, n = 30, p = 0.04) with bacterial numbers at 0 and 10 m only while no significant correlations were found with bacterial production, chlorophyll a concentration and phytoplankton primary and EDOC (extracellular DOC) production at any depth. The rate of bacterial utilization of DFAA was low compared with data from other lakes. Diurnal phytoplankton production of DFAA in the euphotic zone of the whole lake was calculated to vary between 268 and 30 780 t C d^–1 indicating autochthonous DFAA sources were dominant to allochthonous DFAA sources. The autochthonous production of DFAA was > 2 × gross bacterial production of the euphotic zone indicating that although DFAA concentrations were frequently < 10 g C l^–1, the rate of DFAA production exceeded bacterial requirements.     

DIGESTION OF MICROCYSTIS AERUGINOSA BY OREOCHROMIS MOSSAMBICUS

SUMMARY

The mean assimilation efficiency of aquarium acclimatized Oreochromis mossambiaue fed on a diet of Microcystis aeruginosa collected from Hartbeespoort Dam was determined as 50,8% for total organic matter, 63,7% for protein and 75,5% for phosphorus. Transmission electron microscopic examination of faeces of fish fed on M. aeruginosa, revealed that most Microcystis cell walls had become permeable allowing cell contents to leach out. Further digestion resulted in the break down of the cell wall structure. Up to 25% of the cells, however, appeared intact after passing through the fish. Fish fed on a diet of M. aeruginosa lost mass initially, but after 21 days showed a slight gain in mass. The high protein content of M. aeruginosa nay have inhibited efficient metabolism and would have led to reduced growth in fish.     

Phosphorus transformations in the epilimnion of humic lakes: biological uptake of phosphate

SUMMARY. 1. The hypothesis that dissolved humic material (DHM) stimulates bacterial involvement in phosphorus transformations and may thus lead to decreased accessibility of phosphorus to algae was investigated by studying three small forest lakes in southern Finland representing a wide range of concentrations of DHM. 2. Other chemical differences between the three lakes were slight, although the most humic lake exhibited higher concentrations of total phosphorus and of molybdate-reactive phosphorus. Bacterial biomass did not differ significantly between the lakes, but algal biomass was significantly lower at higher DHM concentrations. Consequently the ratio of algal biomass to bacterial biomass was significantly lower in the most humic lake. 3. Uptake of phosphorus from added ³³PO₄ was partitioned between algal and bacterial size fractions by differential filtration. No significant variation between lakes was found in the proportion of particulate ³³P recovered from the algal fraction. 4. Turnover times for phosphate were significantly longer in the most humic lake and also showed lower variability. In general turnover times were long in comparison with values reported from many other lakes. Only briefly in mid summmer did turnover times in two of the lakes shorten to values which would indicate that demand for phosphate was outstripping supply. 5. Short-term storage of samples from the most humic lake stimulated biological incorporation of ³³P, but additions of nitrogen and iron had little effect on phosphate uptake. 6. In these small forest lakes it is probable that no single nutrient consistently limits plankton development. Since no evidence was found that DHM shifts the balance of plankton phosphate uptake away from algae towards bacteria, the influence of DHM on phosphorus transformations may rather be through chemical regulation of free phosphate availability.     

Population dynamics and nutrient fluxes in an aquatic microcosm

An aquatic microcosm, consisting of three spatially separated yet mutually dependent trophic levels, was established in the laboratory and monitored for 310 days. A three-fold research approach evaluates the experimental potential of this large, multicompartmental microecosystem. Realistic biological and chemical features and nutrient fluxes parallel identifiable patterns observed in natural aquatic ecosystems as well as in published laboratory observations. Two successional patterns developed in the autotrophic community: a sequential change in species composition and a progression from a one-compartment planktonic situation to a two-phased planktonic-attached system. Although the microcosm was initially seeded with an axenic culture of Cryptomonas ovata var. palustris Ehr, contamination by Chlorella, Scenedesmus, Closterium, and Anabaena occurred within 41 days. The appearance of attached algae, noted on day 5, marked the transition from a planktonically-based ecosystem to a heterogeneous system. Crashes in the cladoceran population occurred on days 103 and 202. The second collapse was final. Repeated attempts to reestablish Daphnia middendorffiana failed. Mineralization and nutrient cycling are recognizable properties of the microcosm. Ammonification, nitrification, and nitrogen assimilation occurred predominantly in the decomposer tank as did the regeneration of inorganic phosphorus. A peak on day 205 in the ammonia input to the algal tank drawn from beneath the bacterial filter bed followed a peak in total Kjeldahl nitrogen (TKN) (day 135) and preceded peaks in nitrate (day 219) and TKN (day 233). Although levels in the algal tank were undetectable after three weeks, dissolved orthophosphate was actively regenerated in the decomposer bed, recycled to the autotroph unit, and rapidly assimilated by the algae. Characteristic patterns of radiotracer circulation also were evident. Sequential movement of ³²P from the dissolved compartment to phytoplankton to attached algae was proposed. Conversely, ¹⁴C was steadily incorporated into the phytoplankton compartment; filtrate activities fluctuated. Tracer behaviors in the cladoceran compartment were superficially cyclic. Carbon turnover times in the algal and zooplankton compartments were 17 and 11.11 hours, respectively. Indicative of the greater biological mobility of phosphorus, respective turnover times of 2.50 and 2.44 hours were similarly calculated for phosphorus. Unlike dissolved carbon which had a turnover time of 625 hours, dissolved phosphorus was rapidly cycled into the algae (turnover time = 0.58 h).     

MODELS OF ECOSYSTEMS AND AQUATIC ENVIRONMENTS: DEVELOPMENT AND APPLICATION

SUMMARY

An historical treatment of a number of principal events is given. This leads to an assessment of the current research approach in hydrobiology and its success in the development of process orientated ecosystem component models. With the contemporaneous development of similar models in hydrodynamics it is possible to link the two disciplines into an interactive framework. The recent work in the Hartbeespoort Dam project, Lake le Roux and Burrinjack Reservoir, Western Australia shows how valuable this approach has become, particularly as regards defining, with greater precision, the effects of variation in stability in the uppermixed layer of reservoirs upon the accumulation and dispersion of algal communities.     

Phosphorus uptake and turnover by periphyton in the presence of suspended clays

We investigated the effect of suspended bentonite and kaolinite clays on phosphorus uptake and turnover by lotic periphyton in laboratory microcosms. Clays were characterized for their phosphorus affinity using laboratory batch experiments. Periphyton cultivated on glass microscope slides was subjected to a 0.02 mg L⁻¹ radiolabeled soluble reactive phosphorus solution in which a 200-mg L⁻¹ clay load was suspended. A 1-h uptake experiment was followed by a 10-day turnover experiment. Biomass normalized phosphorus uptake, and turnover rates were described by mean rate constants ranging from 0.14 to 0.17 min⁻¹ for uptake and 0.04–0.07 days⁻¹ for turnover. Mean phosphorus concentrations were compared among treatments using repeated measures analysis of variance (ANOVA). Mean phosphorus concentrations among treatments were compared using one-way ANOVA. No significant differences were found among treatments for either analysis. Under laboratory conditions, these clays appear to have little or no short-term influence upon phosphorus uptake or turnover by periphyton.     

THE USE OF TROFIC AS AN AID FOR THE MANAGEMENT OF EUTROPHIC LAKES

SUMMARY

As a result of eutrophication studies that have been performed by the National Institute for Water Research for many years it was decided in 1980 that a detailed investigation into this area take place. Hypertrophic Hartbeespoort Dam was selected as the study site. A simulation model, named TROFIC, of the ecosystem was developed and updated over a three-year research program. Different management strategies were performed to attempt to reduce the size of phytoplankton standing stock and to modify the phytoplankton species composition so that blue-green algae are no longer dominant. In this paper the applicability of each strategy to the model is examined by means of results obtained and conclusions drawn from these results.     

Fate and dynamics of recently fixed C in pasture plant–soil system under field conditions

The flow of photosynthetically fixed C from plants to selected soil C pools was studied after ¹³CO₂ pulse labeling of pasture plants under field conditions, dynamics of root-derived C in soil was assessed and turnover times of the soil C pools were estimated. The transport of the fixed C from shoots to the roots and into the soil was very fast. During 27 h, net C belowground allocation reached more than 10% of the fixed C and most of the C was already found in soil. Soil microbial biomass (C_MIC) was the major sink of the fixed C within soil C pools (ca 40–70% of soil ¹³C depending on sampling time). Significant amounts of ¹³C were also found in other labile soil C pools connected with microbial activity, in soluble organic C and C associated with microbial biomass (hot-water extract from the soil residue after chloroform fumigation-extraction) and the ¹³C dynamics of all these pools followed that of the shoots. When the labelling (2 h) finished, the fixed ¹³C was exponentially lost from the plant–soil system. The loss had two phases; the first rapid phase corresponded to the immediate respiration of ¹³C during the first 24 h and the second slower loss was attributable to the turnover of ¹³C assimilated in C_MIC. The corresponding turnover times for C_MIC were 1.1 days and 3.4 days respectively. Such short turnover times are comparable to those measured by growth kinetics after the substrate amendment in other studies, which indicates that microbial growth in the rhizosphere is probably not limited by substrate availability. Our results further confirmed the main role of the soil microbial community in the transformation of recently fixed C, short turnover time of the easily degradable C in the rhizosphere, and its negligible contribution to more stable soil C storage.     

NITROGEN TRANFORMATIONS AND THE NITROGEN BUDGET OF A HYPERTROPHIC IMPOUNDMENT (HARTBEESPOORT DAM,SOUTH AFRICA)

SUMMARY

The major components of the nitrogen balance of hypertrophic Hartbeespoort Dam were studied at weekly intervals from October 1980 to September 1984. The Crocodile River contributed over 80Z of the hydraulic load and over 97X of the annual total nitrogen load to the impoundment. More than 802 of this nitrogen load was composed of nitrate. Monthly measurements of nitrification and denitrification rates carried out within the impoundment over one annual cycle revealed that the rates of both processes were highest at the ammonia-oxygen chemocline during summer stratification and at the sediment-water interface during winter isothermy. Mass balance calculations showed that between 35 and 49% of the annual total nitrogen load were lost via denitrification each year; the data are compared with those in the literature. Conditions in the lake will favour a change in phytoplankton species composition from Microcystis to a nitrogen-fixing species if the present trend of decreasing inorganic nitrogen concentrations and N: P ratios is maintained.     

APPARENT PREDATION OF MICROCYSTIS AERUGINOSA KüTZ. EMEND ELENKIN BY A SAPROSPIRA-LIKE BACTERIUM IN A HYPERTROPHIC LAKE (HARTBEESPOORT DAM,SOUTH AFRICA)

SUMMARY

The cyanobacterium Microcystis aeruginosa is the dominant autotroph in hypertrophic Hartbeespoort Dam, South Africa. Tn early summer (November) of 1983 and 1984 Microcystis colonies were heavily colonized by a gram-negative, spiral-shaped organism whose filaments appeared as si.ngle cells under both the light and scanning electron microscope. DNA staining with 4'6 diamidino-2-phenylindole (DAPI) revealed that each filament consisted of a number of bow-shaped cells. The organism has been tentatively identified as a bacterium, Saprospira albida, on the basis of it morphology. The Microcystis cells at the site of colonization were apochlorotic suggesting lysis by Saprospira.     

N/P imbalance as a key driver for the invasion of oligotrophic dune systems by a woody legume

Oligotrophic ecosystems, previously considered to be more resilient to invasive plants, are now recognised to be highly vulnerable to invasions. In these systems, woody legumes show belowground ecosystem engineering characteristics that enable invasion, however, the underlying processes are not well understood. Using a Portuguese primary dune ecosystem as an oligotrophic model system, belowground biomass pools, turnover rates and stoichiometry of a native (Stauracanthus spectabilis) and an invasive legume (Acacia longifolia) were compared and related to changes in the foliage of the surrounding native (Corema album) vegetation. We hypothesized that the invasive legume requires less phosphorus per unit of biomass produced and exhibits an enhanced nutrient turnover compared to the native vegetation, which could drive invasion by inducing a systemic N/P imbalance. Compared with the native legumes, A. longifolia plants had larger canopies, higher SOM levels and lower tissue P concentrations. These attributes were strongly related to legume influence as measured by increased foliar N content and less depleted δ¹⁵N signatures in the surrounding C. album vegetation. Furthermore, higher root N concentration and increased nutrient turnover in the rhizosphere of the invader were associated with depleted foliar P in C. album. Our results emphasize that while A. longifolia itself maintains an efficient phosphorus use in biomass production, at the same time it exerts a strong impact on the N/P balance of the native system. Moreover, this study highlights the engineering of a belowground structure of roots and rhizosphere as a crucial driver for invasion, due to its central role in nutrient turnover. These findings provide new evidence that, under nutrient‐limited conditions, considering co‐limitation and nutrient cycling in oligotrophic systems is essential to understand the engineering character of invasive woody legumes.