Mineral Sources and Vertical Distribution of Nutrients in Extremely Acidic Pit Lakes: Impact on Microbial Ecology

Abstract

Nutrient cycling has barely been studied in acidic environments and may have an important influence on the evolution of the microbial communities. In this research, we studied nutrient sources and fluxes in acidic metal-mine pit lakes to evaluate their relationship with the lakes’ microbial ecology. Nutrient concentrations (including phosphorus, nitrogen, and dissolved inorganic carbon) increase with depth in all the studied pit lakes. Phosphorus comes mainly from the leaching of the host rock and is rapidly scavenged from the aqueous phase in the oxygenic and Fe(III)-rich mixolimnion due to adsorption on ferric precipitates (schwertmannite, jarosite), which leads to an important P-limitation in the photic zone. Below the chemocline, however, the sum of phosphorus inputs (e.g., settling of algal biomass, desorption from the ferric compounds, microbial reduction of Fe(III)-sediments) sharply increases the concentration of this element in the anoxic monimolimnion. Nitrogen is very scarce in the host rocks, and only a limited input occurs via atmospheric deposition followed by N-uptake by algae, N-fixation by acidophilic microorganisms, sedimentation, and organic matter degradation in the sediments. The latter process releases ammonium to the anoxic monimolimnion and allows some nitrogen cycling in the chemocline. Soluble SiO₂ in the mixolimnion is abundant and does not represent a limiting nutrient for diatom growth. Differences in phytoplankton biomass and extent of bacterial sulfate reduction between relatively unproductive lakes (San Telmo) and the more fertile lakes (Cueva de la Mora) are likely caused by a P-limitation in the former due to the abundance of ferric iron colloids in the water column. Our results suggest that phosphorus amendment in the photic zone could be an efficient method to indirectly increase acidity-consuming and metal-sequestering bacterial metabolisms in these lakes.

1. Introduction

The microbial ecology of lake and river ecosystems depends on the abundance and bioavailability of essential nutrients. Among the most important nutrients, phosphorus and nitrogen are essential for the development and growth of aquatic life [1]. Nitrogen is a basic constituent of proteins and amino acids, while phosphorus is specifically associated with structural (DNA, RNA) and energetic (ATP, ADP) molecules [2]. The concentration of these two elements largely controls the metabolism of primary producers (typically including algae, cyanobacteria, and higher plants), being therefore a pivotal parameter determining the diversity and structure of lake ecology [1]. In addition, dissolved inorganic carbon (DIC, comprising CO₂, HCO₃^−, and CO₃²⁻) is the primary source of carbon for photosynthesis and therefore fuels organic productivity [3]. Silica also exerts some control on algal productivity in lakes, as it is required for frustule synthesis in diatoms, which can be an important component of phytoplanktonic and phytobenthic communities [1].

Lakes formed in abandoned coal and metal mines usually have low pH values and high concentrations of toxic metals [4,5,6,7]. This combination of high proton and metal contents, along with a low nutrient concentration (which is commonly the case), results in food chains having far fewer levels and species than those of natural freshwater lakes [8]. However, despite their chemical harshness, some pit lakes formed in Spanish mining districts such as the Iberian Pyrite Belt (IPB) in Huelva and Seville (SW Spain) or La Unión in Sierra de Cartagena (SE Spain) show an important development of phytoplankton, which is believed to fuel planktonic and benthic bacterial activity [9,10,11,12,13,14,15]. The biodiversity of these algal communities is comparatively small with respect to freshwater lakes, but the biomass is comparable to, or even higher than, that usually measured in circumneutral lakes [9].

The extent of phytoplanktonic growth, however, differs notably from one pit lake to another. In the biologically more productive lakes, deep chlorophyll maxima (DCM) with associated peaks of dissolved oxygen concentration develop several meters below the lake surface [14]. This photosynthetic oxygen production favors the bacterial oxidation of Fe(II) in the surrounding environment. In addition, the organic exudates released by these acidophilic microalgae are believed to provide fresh organic carbon for heterotrophic and mixotrophic microorganisms (including sulfate reducers and iron reducers) inhabiting the transitional layers and the deep waters of the monimolimnion [11,12,13,14,15,16,17]. These acidophilic microorganisms drive the biogeochemical cycling in these lakes, including the removal of metals from the water column. Therefore, understanding the sources and dynamics of essential nutrients fueling microbial activity in these environments is essential for the development of in situ bioremediation strategies [11,15,16,17,18].

While the mobility of metals is well established, limited knowledge still exists on the sources and geochemical behavior of basic nutrients in these extremely acidic and perennially stratified lakes. Thus, this study aimed to determine the sources, vertical distribution, and fluxes of essential nutrients (especially phosphorus and nitrogen, but also DIC, total organic carbon (TOC), and silica) in different acidic pit lakes formed in abandoned and flooded metal mines. For this purpose, we selected four pit lakes in Spain (Cueva de la Mora –CM–, Herrerías-Guadiana –HE-G–, San Telmo –ST–, and Brunita –BRU-) with different sizes, water chemistry, and phytoplankton abundance. We first discuss the mineral sources of nutrients in these lakes, and then we present vertical profiles of nutrient concentration in waters, suspended particulate matter, and bottom sediments. The final goal was to evaluate the impact of nutrient dynamics on the microbial ecology of these lakes.

2. Materials and Methods

2.1. Study Sites

Detailed limnology, microbiology, and aqueous geochemistry of the selected pit lakes (CM, HE-G, ST, and BRU) can be found in previous works by the authors [19,20,21,22,23,24,25]. All the selected pit lakes are meromictic (permanently stratified), and some of them are physically connected to associated subhorizontal mine galleries [24]. These pit lakes, however, show differences in size, stratification pattern, water chemistry, nutrient content, and phytoplankton development, which roughly cover the spectrum displayed by the whole set of metal-mine pit lakes studied to date in Spain [10,19].

The three pit lakes studied in the IPB (CM, HE-G, ST) are located in remote areas with very little or no agricultural, industrial, or urban activities (Figure S1). Local soils are usually strongly degraded mining soils and are not amended with fertilizers. These soils are poorly developed, anthropogenic regosols and leptosols [26]. The vegetation is limited to a few plant species (e.g., Eucalyptus sp., Pinus sp., Rosmarinus sp., Cistus sp., Erica andevalensis), and the external sources of nutrients are those related to plant biomass degradation and weathering. The rocks surrounding the pit lakes include volcanic and sedimentary lithologies such as tuffs and lavas of rhyolitic to dacitic composition, basic volcanics, shales, and greywackes [27,28]. With a maximum depth of 130 m and a volume of around 8 × 10⁶ m³, ST is the largest among the studied pit lakes. This pit lake usually shows very low chlorophyll a (chl-a) concentrations (<1 µg/L) and a rare presence of algal biofilms in shore sediments [14]. On the other hand, CM and HE-G are smaller pit lakes (maximum depths of 40 and 70 m, respectively, and surface area ~17,000 m²) that show more intense and widespread DCM near the chemocline, where concentrations up to 240 μg/L chl-a have been measured [14,22]. These two lakes show evidence of microbial sulfate and/or ferric iron reduction in their water columns [9,11,12].

The pit lake of Brunita mine is situated at 2 km south of the mining town of La Unión, in the province of Murcia (SE Spain) (Figure S1). The mineral paragenesis of the rocks hosting the mineralization includes quartz, chlorite, and carbonate-containing minerals like siderite, dolomite, and minor calcite [13]. Some rare phosphate minerals like vivianite [Fe₃(PO₄)₂·8H₂O] and ludlamite [(Fe, Mg, Mn)₃(PO₄)₂·4H₂O] are also present in the host rocks [29,30]. The pit lake has a surface area of 45,000 m² and a maximum depth of 22 m. Eukaryotic photosynthetic microorganisms are abundant in this lake [13].

2.2. Water and Sediment Sampling

Sampling of waters and sediments, and chemical analyses presented in this study were carried out in the course of different projects and in different seasons between 2008 and 2020. The water samples for nutrient analyses were taken from different depths with a Van Dorn^® sampling bottle (KC Denmark) operated from a boat. Sediment traps (HYDRO-BIOS, GmbH, Kiel-Holtenau, Germany) were installed at different depths in the lakes for systematic collection, examination, and chemical analyses of suspended sediments. These sediments were thoroughly washed with MilliQ water, dried at room temperature, weighed, and stored under ambient conditions until analyzed. Bottom sediments were also sampled in different parts of the lakes with a gravity corer (UWITEC, GmbH, Mondsee, Austria) for the study of sediment and pore-water chemistry.

2.3. Chemical Analyses

Water samples were analyzed for ammonium, nitrate, and total nitrogen, phosphorus as orthophosphate, and total organic and inorganic carbon by UV-VIS spectrophotometry with a HACH DR2800 spectrophotometer (Hach, Loveland, CO, USA) using cuvette tests LCK 304 (NH₄-N, measurement range of 0.05–2.0 mg/L), LCK 339 (NO₃-N, 0.23–13.50 mg/L), LCK 138 (N_T, 1–16 mg/L), LCK 348 (PO₄-P, 0.05–5.0 mg/L), LCK 385 (TOC, 3–30 mg/L), and LCK 388 (TIC as CO₂, 55–550 mg/L), following analytical procedures by Hach Lange GmbH (Düsseldorf, Germany). The bottom sediments and the suspended particulate matter (SPM) collected in the traps were chemically analyzed for organic and inorganic carbon (elemental analyzer Eltra CS-800; ELTRA GmbH, Haan, Germany), total nitrogen (Kjeldahl method), and total phosphate (X-ray fluorescence, XRF). Inorganic carbon was measured after previous elimination of organic carbon by calcination to 550 °C for 2 h. Organic carbon was calculated as the difference between total and inorganic carbon. Scanning Electron Microscopy (SEM) coupled with Electron Dispersive Spectrometry (EDS) was carried out on SPM with a JEOL JSM-6400 microscope (JEOL Ltd., Tokyo, Japan) at the SGIker facilities (UPV/EHU). The carbon isotopic signature (δ¹³C_PDB, in ‰) of SPM was analyzed at Universidad Autónoma de Madrid (UAM, SIdi facilities, Stable Isotopes Lab; Madrid, Spain).

3. Results and Discussion

3.1. Sources and Chemical Species of Essential Nutrients

3.1.1. Dissolved Inorganic Carbon (DIC)

The freshwater courses in the IPB mining area (pristine creeks not affected by acid mine drainage pollution) act as a source of DIC to the acidic pit lakes. These freshwater courses show some degree of natural alkalinity, which is primarily present in the form of bicarbonate ion (6–113 mg/L HCO₃⁻ [31]). Carbonate minerals (calcite, dolomite, ankerite) represent an important source of DIC in the pit lakes studied through acid digestion. At the pH range of 2.2–4.5 (which covers the water column of all the studied pit lakes), DIC is mainly present as dissolved carbon dioxide (CO₂ aq.) [32,33].

3.1.2. Total Organic Carbon (TOC)

These lakes receive very little input of surface water. Therefore, the major source of organic carbon in the lakes is autochthonous and formed by primary production. However, visual examination of the acidic lake sediments has revealed the presence of tree leaves and plants growing on the pit walls and surrounding areas, which may partly contribute to the TOC budget. Remains of dead insects (e.g., acid-tolerant corixids and chironomids, which have been observed in the shallow sediments of all lakes) and excrements from birds and other animals may represent additional minor sources of organic carbon. Black shales with organic matter contents between 0.1 and 1.5 wt.% and 2–4 wt.% have been reported in ST, CM, and HE-G [28,34] and may also contribute to the dissolved carbon content in the acidic lakes. Primary production by photosynthetic algae and autotrophic bacteria is thus envisaged as a major input of carbon into the food web of these lakes [17].

Comparison of TOC concentrations measured in filtered (0.45–0.1 μm) and unfiltered samples from different depths in the CM pit lake (

Table 1. Comparison of total organic carbon, total nitrogen, and total phosphate concentration measured in filtered (0.1 and 0.45 µm) and unfiltered water samples from the CM pit lake.

Depth	Total Organic Carbon				Total Nitrogen				Phosphate
(m)	Unfiltered	Filtered	Filtered		Unfiltered	Filtered	Filtered		Unfiltered	Filtered
		(0.45 µm)	(0.1 µm)			(0.45 µm)	(0.1 µm)			(0.1 µm)
10	3.36	2.84	2.12		4.80	4.80	2.12		-	-
19	2.27	-	2.11		13.70	12.40	12.40		2.68	1.28
35	10.70	8.44	7.98		24.10	23.10	22.00		3.03	0.68

) suggests that between 60% and 90% of the TOC content is likely present as dissolved organic carbon (DOC). The remaining portion of TOC is particulate carbon in the form of either dead algal biomass and/or organic acids adsorbed onto iron mineral particles [9,20].

There is still little information on the types, vertical distribution, and degradation pathways of organic compounds in these acidic pit lakes. Dissolved organic carbon (DOC) in a German pit lake was reported to consist of polymeric fulvic acids [35]. The photochemical reduction of Fe(III) (which is a widespread phenomenon in surface waters of these lakes [36,37]) can transform high-molecular-weight compounds (which initially make up most of the bulk DOC in these lakes) to low-molecular-weight organic acids such as acetate, formate, or pyruvate, which represent more suitable organic carbon sources for utilization by some heterotrophic bacteria [38]. The scarce information on the sources and sinks of TOC and DOC in acidic metal-mine pit lakes highlights the need for detailed studies on this topic.

3.1.3. Phosphorus

Compared to other major nutritional components (carbon, nitrogen), phosphorus is less abundant and commonly limits biological productivity [1]. Inorganic phosphorus has been identified as a clear limiting factor for primary production and biomass yield in acidic mine pit lakes [39,40,41]. The major source of inorganic phosphorus in fresh waters is usually found in mineral phases of rocks and soils [1]. The phosphate content of rocks in the IPB is highly variable [28] (see Table S1 in Supplementary Materials).

Sedimentary rocks show phosphate contents typically below 0.1%wt. P₂O₅ (i.e., around 435 ppm P). Felsic volcanic rocks show lower phosphate contents (0.02 wt.%–0.04 wt.% as P₂O₅, ≈90–180 ppm P). In contrast, some basic igneous rocks may contain higher P₂O₅ contents of 0.11 wt.%.–0.33 wt.% (≈480–1440 ppm P), and other rocks may locally possess important phosphate concentrations (e.g., 0.56 wt.% in gabbros or 1.08 wt.% P₂O₅ in dacites; Table S1). Overall, these data point to an important phosphate pool in the host rocks, which is likely present in the mineral apatite and/or bound to clays or ferric hydroxides. The geological context of Brunita is different and includes uncommon phosphates like vivianite and ludlamite [13], which are both readily soluble at low pH.

The acid-soluble phosphate content of volcanic and sedimentary rocks of the IPB has been confirmed by acid-digestion experiments carried out with rock samples collected from HE-G and CM mine pits (Table S2 in Supplementary Materials). Digestion of these rocks in HCl-acidified distilled water (pH 1.5) for five weeks resulted in variable phosphate release depending on the rock sample: 363–454 μg/L PO₄³⁻ for shales, 728 μg/L PO₄³⁻ for rhyolites, and 914–945 μg/L PO₄³⁻ for a combined sample with dacites and basic volcanic rocks. Chemical analyses of acidic mine waters emanating from waste piles and mine adits of the area revealed significant PO₄³⁻ contents (e.g., 0.48 mg/L P-PO₄³⁻;

Table 2. Total nitrogen, ammonium nitrogen (N-NH₄), phosphate phosphorus (P-PO₄³⁻), and total organic carbon (TOC) were measured in unpolluted water of the Olivargas River and in two acidic mine effluents of the Iberian Pyrite Belt. All concentrations are in mg/L.

Site	Effluent Type	Total N	N-NH4+	P-PO43−	TOC
Olivargas	Unpolluted river	21.20	b.d.	0.04	3.06
Olivargas	Unpolluted river	6.27	b.d.	b.d.	5.09
Río Tinto	Waste-rock pile	6.24	1.58	0.48	4.30
La Zarza	Mine portal	6.39	0.13	0.12	4.48

Abbreviations: b.d., below detection.

), while background phosphate contents in local unpolluted streams have been reported to be relatively lower (<50 to 200 μg/L PO₄³⁻ [31]). These data indicate an enhanced dissolution of acid-soluble phosphate at low pH conditions from the mine host rocks.

The most common form of dissolved inorganic phosphorus (DIP) in fresh waters is orthophosphate (PO₄³⁻). In acidic mine waters, geochemical calculations with PHREEQC indicate that orthophosphate is usually complexed with aqueous metal cations like Fe³⁺, Fe²⁺, Mg^2+, or Ca²⁺ to form different chemical species. In the oxidizing and Fe³⁺-rich mixolimnion, the main phosphate-bearing species are FeHPO₄⁺, H₂PO₄⁻, FeH₂PO₄^2+, and H₃PO₄⁰, with a minor amount of orthophosphoric acid (H₂PO₄). In contrast, under the reducing conditions prevailing in the anoxic, Fe²⁺-rich monimolimnion, the major phosphate species are FeH₂PO₄⁺, H₂PO₄^−, and MgH₂PO₄⁺.

A major geochemical factor limiting the abundance and bioavailability of phosphorus in acidic mine waters is its strong affinity for iron [14]. DIP can be bound to ferric iron in several ways, including (1) adsorption to ferric mineral colloids (e.g., schwertmannite, jarosite, ferrihydrite) formed in the water column during bacterial oxidation of Fe(II) and (2) direct precipitation of ferric phosphate (e.g., strengite). Chemical analyses of ferric precipitates formed in different streams and lakes of the IPB revealed that schwertmannite is an efficient phosphate adsorbent and may adsorb up to 0.1 wt.%–0.2 wt.% P₂O₅ [20].

Comparison of total phosphorus measured in unfiltered and filtered (0.45 μm, 0.10 μm) deep-water samples from CM lake (Table 1; Figure S2 in Supplementary Materials) allows for the differentiation between particulate and truly dissolved phosphorus (measured as total orthophosphate). Between 50% and 80% (depending on the sampling depth) of total PO₄³⁻ (defined by chemical analyses using spectrophotometry, see Section 2.3) in unfiltered samples is actually solid-phase phosphorus. SEM studies of SPM collected in sediment traps suggest that solid-phase phosphorus may include: (1) phosphate minerals directly formed in the water column (Figure S3). (2) PO₄³⁻ adsorbed to schwertmannite (Figure S4), and (3) settling algal biomass.

3.1.4. Nitrogen

Nitrogen in fresh waters occurs in both organic and inorganic forms. Dissolved inorganic nitrogen (DIN) includes molecular nitrogen gas (N₂), ammonia (NH₄⁺), nitrites (NO₂⁻), and nitrates (NO₃⁻). Dissolved organic nitrogen (DON) may be present as different compounds such as amino acids, amines, proteins, urea, or humic compounds [1,42,43,44,45]. Common sources of nitrogen to lakes usually include: (1) precipitation onto the lake surface, (2) pelagic and benthic nitrogen fixation, and (3) inputs from surface and groundwater [1]. Acid leaching of igneous rocks present in the pit lakes could represent an additional source of nitrogen, since these rocks can contain significant amounts of this element (e.g., 10–100 ppm as total N, basically present as ammonia, NH₄⁺) [46]. Ammonium can also be significantly adsorbed into clay minerals [47,48,49,50] and black shales of the area (e.g., 60–700 ppm N [34]) and could be released during water/rock interaction.

The direct atmospheric input of nitrogen in the studied acidic lakes is unknown, but concentrations of 12–15 mg/L N₂ (g) have been inferred in the mixolimnion of the studied lakes by geochemical calculations [25], which are approximately consistent with the theoretical solubility of this gas in surface waters at 1 bar and 20 °C [1].

Chemical analyses of local unpolluted freshwater streams show that background nitrogen contents are highly variable (Table 2). Total nitrogen concentrations exceeding 21 mg/L have been measured in the Olivargas River before this stream enters the CM mining area. Nitrate-nitrogen (NO₃⁻-N) and ammonium-nitrogen (NH₄⁺-N) were usually below detection (Table 2) so most of this nitrogen is considered to be organic.

Regarding the biological component of the nitrogen cycle in these lakes, recent omics studies have reported nitrogen fixation and nitrate reduction in the chemocline of CM, where Coccomyxa assimilates nitrogen in the upper layer and chemocline [16]. The activity of nitrogen-fixers and nitrate-reducers in the chemocline and deep layer may explain increasing concentrations of ammonium with depth in CM [16]. In the deep layer of this lake, genes involved in dissimilatory nitrate reduction, denitrification, and/or nitrogen fixation were also present and affiliated with Actinobacteria, Chloroflexi, Firmicutes, Nitrospirae, and Proteobacteria [16].

Nitrogen measurements conducted in acidic effluents seeping from mine adits and waste-rock piles found in the IPB reveal that these mine waters may have high N concentrations (e.g., around 6 mg/L N_T; Table 2). Since only a small part (<10%) of this nitrogen is actually DIN (NO₃⁻ + NH₄⁺), most N_T must consist of organic compounds with unknown composition. Comparison of total nitrogen measured in unfiltered and filtered (0.45 μm, 0.10 μm) samples from CM (Table 1) also suggests that nitrogen is present either as dissolved or colloidal organic matter. This is in good agreement with freshwater rivers worldwide, where organic nitrogen usually accounts for 20 to 95% of total nitrogen [51,52].

An additional N source, which may be worth considering in the studied mines and resulting pit lakes, is the acid digestion of explosives. Nitrogen-based explosives such as 2,4-Dinitrotoluene (DNT or Dinitro, CH₃C₆H₃(NO₂)₂) and Ammonal (made up of ammonium nitrate, 2,4,6-trinitrotoluene -TNT- and aluminum powder) were widely used in the former mines. In CM, the annual consumption of these compounds was 6 Tm, with a ratio of 190 g of explosives per Tm of mined ore [53]. TNT can dissolve in water up to maximum solubility concentrations around 150 ppm [54]. Red waters produced during TNT purification have a complex composition containing many different aromatic compounds and nitroaromatics. TNT transformation occurs through sequential reduction of nitro groups by several nitroreductases to form nitroso-, hydroxylamino-, and amino-derivation [55,56]. Explosive residues such as TNT, 2,4-dinitrotoluene, and 2,6-dinitrotoluene (2,4-DNT and 2,6-DNT, respectively) or hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) have relatively high aqueous solubilities and weak binding affinity for soil [57]. As a result, these munitions constituents could be leached and transported to groundwater [58]. In addition, some microbes can biodegrade TNT [59,60,61], which, together with the N leached during TNT and DNT dissolution, could serve as a slow but continuous source of N in the pit lakes. The presence, composition, and bioavailability of these N compounds need further investigation.

3.1.5. Silicon

Dissolution of the common aluminosilicates (clay minerals, micas, feldspars) present in volcanic and sedimentary rocks is the primary source of silica in mine waters and acidic pit lakes. The abundance of dissolved silica in local unpolluted freshwater courses (9–26 mg/L SiO₂ [31]) is only slightly higher than the world average (13 mg/L SiO₂ [1]) and within the range reported for common surface and groundwater (0.5–54 mg/L SiO₂ and 0.8–63 mg/L SiO₂, respectively [62]). On the other hand, dissolved silica concentration in the acidic mine pit lakes is notably higher (20–128 mg/L;

Table 3. Concentration of dissolved inorganic carbon (DIC), total organic carbon (TOC), phosphate phosphorus (PO₄³⁻-P), total nitrogen (N_T), nitrate nitrogen (NO³⁻-N), ammonium nitrogen (NH⁴⁺-N), and soluble silica (SiO₂) measured at different depths in pit lakes of southern Spain. A complementary table with pH, Eh, and metal concentrations for the same lakes and depths is given for interested readers in the Supplementary Materials (Table S3).

Pit Lake	Water Layer Type	Depth	DIC	TOC	PO43−-P	NT	NO3−-N	NH4+-N	SiO2
Units		m	mg/L	mg/L	µg/L	µg/L	µg/L	µg/L	mg/L
San Telmo	Mixolimnion, oxygenic	0	40	1.61	b.d.	704	n.a.	27	n.a.
	Mixolimnion, oxygenic	10	48	1.20	b.d.	b.d.	16	30	70.4
	Mixolimnion, oxygenic	20	45	2.22	n.a.	1070	n.a.	31	n.a.
	Monimolimnion, anoxic	40	88	0.91	b.d.	1330	n.a.	47	n.a.
	Monimolimnion, anoxic	95	104	0.61	b.d.	1620	16	40	69.2
Cueva de la Mora	Mixolimnion, oxygenic	0	57	0.54	b.d.	b.d.	410	24	116
	Mixolimnion, oxygenic	4	56	0.57	b.d.	b.d.	420	25	116
	Monimolimnion, anoxic	11	310	2.70	b.d.	4420	n.a.	455	128
	Monimolimnion, anoxic	19	640	2.11	2676	11,000	n.a.	399	121
	Monimolimnion, anoxic	24	768	1.70	3210	13,050	25	352	112
	Monimolimnion, anoxic	35	1270	10.70	3030	22,200	25	580	80
Herrerías-Guadiana	Mixolimnion, oxygenic	0	71	0.90	b.d.	b.d.	601	53	20.7
	Mixolimnion, oxygenic	6	83	1.00	115	536	572	56	78.6
	Mixolimnion, oxygenic	14	67	1.10	403	512	637	84	140
	Monimolimnion, anoxic	20	1100	3.04	b.d.	3720	b.d.	137	70.8
	Monimolimnion, anoxic	40	2600	19.10	1310	29,600	b.d.	118	101
	Monimolimnion, anoxic	55	4962	14.20	1810	26,800	b.d.	317	20.2
Brunita	Mixolimnion, oxygenic	0	n.a.	1.74	40	n.a.	1200	30	75
	Mixolimnion, oxygenic	12	n.a.	2.87	40	n.a.	2040	66	71
	Monimolimnion, anoxic	16	n.a.	1.58	97	n.a.	1890	77	36
	Monimolimnion, anoxic	20	n.a.	1.87	484	n.a.	1190	494	20.1
	Monimolimnion, anoxic	24	n.a.	3.05	946	n.a.	1040	682	22.5

Abbreviations: b.d., below detection; n.a., not analyzed.

) due to enhanced silicate dissolution at low pH. Silica may be present in dissolved or particulate forms. Dissolved silica at low pH is traditionally represented as silicic acid (H₄SiO₄), though the most abundant forms are monomeric (Si(OH)₄) and polymeric silica (e.g., Si₂O₂(OH)₅⁻ [63]). Particulate silica may be present in several forms, including (1) fine-grained silicates (clays), (2) silica adsorbed into iron and aluminum hydroxides, (3) silica complexed with organic acids, (4) silica in diatom frustules, and (5) amorphous silica [1].

3.2. Nutrient Concentrations in the Water Column: Vertical and Seasonal Variations

3.2.1. Dissolved Inorganic Carbon

Dissolved inorganic carbon (DIC) concentrations measured in the studied lakes at several depths are shown in Table 3, and their concentrations are relatively low near the lake surface (40–48 mg/L in ST, 56–57 mg/L in CM, and 67–83 mg/L in HE-G). As already discussed, under low pH conditions prevailing in these lakes, DIC is chiefly present as carbon dioxide (CO₂) and diffuses as a dissolved gas phase through the mixolimnion. The concentrations of CO₂ largely exceed the theoretical equilibrium (10^−3.5 atm), so these lakes are net CO₂ contributors to the atmosphere. The higher CO₂ concentrations are always found in the monimolimnion of these lakes, suggesting a deep source for the dissolved CO₂.

The vertical variation in DIC concentration and the DIC isotopic composition in CM and HE-G lakes are shown in Figure 1. Dissolved inorganic carbon concentrations sharply increase at depth and reach values of thousands of mg/L near the lake bottom (e.g., ~1300 mg/L CO₂ in CM, ~5000 mg/L CO₂ in HE-G; Table 1, Figure 1), i.e., between 30 and 70 times those observed at the lake surface. The case of HE-G is exceptional and represents considerable concern due to its potential risk of a gas outburst [25]. The source of this CO₂ is mainly inorganic and results from the dissolution of acid carbonates in host rocks and connected mine tunnels. The isotopic signature of this inorganic carbon (δ¹³C_DIC = −10 to −14‰) is rather similar to those measured in carbonate minerals present in the host rocks and in dissolved bicarbonate of local freshwater courses [25]. In addition, the carbon isotopic composition also suggests a subordinate contribution of microbial CO₂ (more negative values of δ¹³C_DIC between −16 and −20‰), which probably comes from the degradation of organic matter in anoxic waters and sediments.

3.2.2. Total Organic Carbon

Total organic carbon (TOC) concentrations measured in the studied lakes are given in Table 3, and the vertical evolution of TOC in CM and HE-G is illustrated in Figure 2. TOC contents differ greatly between lakes and also vary significantly with depth within a given pit lake. TOC in the upper meters of the water column is consistently low in all the studied lakes (0.54–1.61 mg/L), most probably as a result of the low external carbon inputs from the drainage basin. These TOC contents are much lower than those found in freshwater rivers worldwide, where TOC content is ~9–10 mg/L TOC on average [51]. TOC in the lake’s deeper waters is related to nutrient availability and phytoplankton biomass in the photic zone [9,10,11,12]. In the relatively unproductive ST pit lake, a low nutrient content in the mixolimnion (see below) has resulted in poor phytoplankton development, as revealed by low chlorophyll a concentrations [10]. This situation leads to a low organic carbon flux (via settling of dead microalgal biomass) to the lake bottom. In contrast, in more biologically productive pit lakes such as CM and HE-G, higher nutrient concentrations lead to greater phytoplankton development near the chemocline and, consequently, greater fluxes of organic carbon input to the deeper levels (Figure 2). Maximum TOC concentrations measured at depth in CM and HE-G are 10 mg/L and 19 mg/L, respectively. This TOC includes particulate carbon and “dissolved” (<0.1 µm) organic carbon, possibly derived from organic matter degradation. In CM, particulate organic carbon represents about one-third (~37%) of the total carbon below the DCM at 10 m depth (Table 1). This percentage is reduced to one-fourth (~25%) near the lake bottom (35 m) and only represents 7% at intermediate depth (19 m), far from the most reactive zones (chemocline and lake bottom).

3.2.3. Phosphorus

Total phosphorus (measured as PO₄³⁻-P) is strongly variable between lakes and with depth (Table 3, Figure 3). As discussed above (see Section 3.1.3.), phosphate shows a strong affinity for binding (via adsorption) to ferric compounds such as schwertmannite, and the concentration of this nutrient in the oxygenic, Fe(III)-rich waters of the lakes is very low (e.g., <30–50 µg/L PO₄³⁻-P in the photic zone of ST and CM; Table 3). Since these values include particulate phosphorus, the dissolved phosphate content is even lower and corresponds to oligotrophic conditions [1]. These low P contents in the photic zone represent a strongly limiting factor for algal growth in the uppermost layers of the lakes. However, in the deep, anoxic, and Fe(III)-free waters of CM, HE-G, and CB, phosphate is sharply increased to values of 3210, 1810, and 2100 µg/L PO₄³⁻-P, respectively (Table 3, Figure 3). This difference in phosphate concentration between the mixolimnion and the anoxic monimolimnion (Figure 3) favors ionic diffusion through the chemocline and can exert a strong impact on the microbial ecology of these lakes. For example, previous studies have shown that certain photoautotrophic acidophiles (e.g., Coccomyxa) can migrate to deeper levels to benefit from a higher phosphorus availability immediately above the chemocline, where dissolved phosphorus concentrations are higher due to diffusion from the monimolimnion [14]. This concentration of algal biomass at deeper levels forms the DCM, where much of the phosphorus is present as particulate matter, as discussed above (Section 3.1.3, Table 1). In CM, at a depth of 10 m (i.e., below the productive zone with conspicuous phytoplankton growth during the summer), more than half (52%) of the phosphorus is particulate and mostly present as dead algal biomass. This contribution of solid-phase phosphorus is even higher (78%) near the lake bottom, where several inorganic forms of phosphate (e.g., phosphate minerals, adsorbed phosphate on settling ferric precipitates) may be combined with the organic phosphorus (Figures S3 and S4). This proportion is similar to that observed in freshwater rivers worldwide (95% P present as particulate phosphorus, of which 40% is organic [51]). The phosphate content may also display some seasonal fluctuations in these pit lakes as a result of varying rates of settling phytoplankton biomass (Figure 3).

3.2.4. Dissolved Inorganic Nitrogen

Concentrations of nitrate and ammonium nitrogen measured in the three lakes at different depths and in several seasons are also given in Table 3. Due to low nitrite concentration (below detection limit, <20 μg/L NO₂⁻) in these waters, the sum of nitrate and ammonium is here considered to be equivalent to the DIN content. There is an important difference in DIN between ST (27–46 µg/L DIN) and the microbially more productive lakes of CM and HE-G (430–795 µg/L DIN). Nitrate is mainly present in the mixolimnion (434–721 µg/L NO₃⁻-N) and nearly absent (below detection, <20 μg/L NO₃⁻) in the monimolimnion (Table 3; Figure 4), while reversed trends are often observed for ammonium (24–84 µg/L NH₄⁺-N vs. 118–580 µg/L NH₄⁺-N). The case of Brunita pit lake is exceptional, since inorganic nitrogen can exceed concentrations of 2000 µg/L in the mixolimnion (Table 3; see also Figure S5). The source of this nitrogen is currently unknown, but atmospheric deposition seems a plausible option considering the high nitrate concentrations reported in local rainwater (4–23 mg/L NO₃⁻ [64]). The vertical trends of ammonium concentration (Figure 4; see also Figure S5) suggest that this compound is released from the sediments during the degradation of organic matter. These ammonium concentrations are higher than in most natural lakes (80–200 µg/L NH₄⁺-N [65]). Nitrification is inhibited below pH 3.0 [66], but the pH of the redoxcline and sediments of CM and HE-G (3.0–4.5; Table S3) could allow nitrification and/or some microbial oxidation of ammonium.

3.2.5. Total Nitrogen

Total nitrogen (N_T) concentrations increase notably with depth in all lakes (Table 3; Figure 5). The measured concentrations of total nitrogen are usually much higher than the DIN content, so that the remaining portion of nitrogen present in these acidic waters must be organic. The filtration results in the chemocline of CM suggest that 56% of this nitrogen is present in the particulate fraction (>0.1 µm; Table 1). As with TOC and phosphate, this nitrogen is mostly present as dead microalgal biomass coming from the photic zone [11]. However, this percentage of particulate nitrogen is notably reduced at depth, and only accounts for around 10% of total nitrogen in the monimolimnion (Table 1). The remaining 90% of this nitrogen could be present as various forms such as: (i) organic colloids (i.e., with particle diameters < 0.1 µm), (ii) organic compounds resulting from organic matter degradation (e.g., amino acids, proteins, amines, urea), and/or (iii) other compounds released during the acid dissolution of explosive remnants present in the mines.

Another important aspect about the total nitrogen content is that, independently of its bioavailability, it is far more abundant than phosphate at all depths. The N/P molar ratios are usually equal or much higher than the value of [N/P]_m = 10, which indicates phosphorus as the limiting nutrient in the studied lakes [67,68]. These calculations evidence a clear P-limitation for primary production in these lakes.

3.2.6. Soluble Silica

Maximum concentrations of soluble silica in the studied lakes range from 20 mg/L SiO₂ in BRU to 128–140 mg/L SiO₂ in CM and HE-G (Table 3). In the latter, silica increases with depth but then decreases again near the lake bottom, which has been attributed to adsorption onto Al compounds [69]. In any case, enough silica exists in all lakes for diatom growth requirements.

3.3. Nutrient Concentrations in Shallow and Deep Sediments

A summary of the available data on nutrient concentrations in bottom sediments taken from different depths in the studied lakes is given in

Table 4. Concentration of phosphate (as P₂O₅), total organic carbon (C_total), inorganic carbon (C_inorg), organic carbon (C_org), and total nitrogen (N_total) measured in sediment cores taken at different depths (and in different layers within a given sediment core) in pit lakes of the Iberian Pyrite Belt.

Pit Lake	Depth	Layer	P2O5	Ctotal	Cinorg	Corg	Ntotal
	(m)	(cm)	(%)	(%)	(%)	(%)	(%)
Guadiana	20	0–2	0.26	1.53	0.03	1.50	0.02
Guadiana	60	0–2	0.31	n.a.	n.a.	n.a.	n.a.
Guadiana	60	2–4	0.27	1.04	0.03	1.01	<0.01
San Telmo	0	0–2.5	0.06	0.87	0.02	0.85	n.a.
San Telmo	30	0–2.5	0.05	0.60	0.01	0.59	n.a.
San Telmo	40	0–1	0.10	2.04	0.02	2.02	n.a.
Cueva de la Mora	5	0–2.5	0.23	9.31	0.20	9.11	n.a.
Cueva de la Mora	5	3.5–4.5	0.21	10.27	0.16	10.12	n.a.
Cueva de la Mora	5	5–6	0.23	16.63	0.78	15.85	n.a.
Cueva de la Mora	35	0–2	0.21	1.23	0.02	1.21	n.a.
Cueva de la Mora	37	0–3	0.08	1.40	0.01	1.39	n.a.
Cueva de la Mora	38	8–10	0.12	1.30	0.12	1.18	n.a.
Cueva de la Mora	38	13–15	0.12	1.81	<0.10	1.86	n.a.

Abbreviations: n.a., not analyzed.

. A general feature shown by most sediments of all lakes is a low phosphate and total carbon concentration. The phosphate content ranges from the lowest values measured in sediments from ST (0.05 wt.%–0.10 wt.% P₂O₅) to the highest values found in sediments of HE-G pit lake (0.26 wt.%–0.31 wt.% P₂O₅). With respect to the organic carbon content, there is no significant difference among the deep sediments of all lakes (0.6 wt.%–2.0 wt.% in ST, 1.0 wt.%–1.5 wt.% in HE-G, 1.2 wt.%–1.9 wt.% in CM). However, an important difference exists between shallow and deep sediments in the CM pit lake. The high TOC content in the shallow sediments (9.1 wt.%–15.8 wt.%; Table 4; Figure 6) is the result of the accumulation of benthic microalgae, which can still conduct primary production through photosynthesis in the photic zone. This is in contrast to the carbon-deficient sediments of the deep monimolimnion of all lakes, where the only carbon source is external and originates from sedimentation from the upper levels. This difference in the availability of organic carbon between mixolimnetic and monimolimnetic sediments implies an important distinction regarding microbial activity and biogeochemical cycling of sulfur, iron, and carbon in these lakes [9,12] (see Section 3.5 and Section 3.6).

The few data available on the nitrogen content of these sediments reflect a very low concentration of this nutrient (Table 4). In HE-G, this element was very close to the detection limit (0.01 wt.% N_T). In CM, total nitrogen concentrations around 0.5 wt.% and 0.2 wt.% have been reported in mixolimnetic and monimolimnetic sediments, respectively [9], which is in good agreement with the nitrogen data measured in suspended particulate matter analyzed in this study (see Section 3.4).

3.4. Vertical Fluxes of Nutrients Across the Water Column

The TOC, nitrogen, and phosphorus concentrations of SPM recovered from sediment traps installed at different depths in the ST pit lake, along with their respective carbon isotopic composition (δ¹³C), and their corresponding [C/N] molar ratios, are given in

Table 5. Concentration of total organic carbon (C_T), total nitrogen (N_T), total phosphorus (P_T), [C/N] molar ratio, and carbon isotopic signature (δ¹³C) measured in suspended sediments collected from different depths in ST pit lake and sampled at variable time intervals of 3–6 months between 2011 and 2014.

Sample	Depth	Sampling		CT	NT	PT	[C/N]m	δ13CPDB
	m			wt.%	wt.%	wt.%		‰
ST-10	10	October 2014		2.37	0.42	0.04	6.64	−25.79
ST-40	40	October 2014		2.04	0.36	0.03	6.51	−26.28
ST-40	40	April 2011		1.28	0.20	0.02	7.47	n.d.
ST-40	40	May 2012		15.24	n.d.	0.09	-	n.d.
ST-100	100	October 2014		1.92	0.33	0.03	6.80	−26.04

Abbreviation: n.d., not detected.

. These data serve as a proxy to evaluate the downward flux of solid-phase nutrients to the bottom of these lakes. The material recovered from the sediment traps was mostly composed of ferric chemical precipitates (schwertmannite, jarosite) formed by precipitation of dissolved Fe(III), which is present in the whole water column [12]. However, part of the SPM consisted of dead microalgal biomass settling from the photic zone of the lake, and therefore, these sediments transport a significant part of organic carbon, nitrogen, and phosphorus to the deep part of the lakes. SPM from ST shows TOC contents of 1.28 wt.% to 2.37 wt.% (with the exception of 15.24 wt.% measured at 40 m depth soon after an algal bloom), 0.20 wt.% to 0.42 wt.% N, and 0.02 wt.%–0.09 wt.% P. The [C/N] molar ratios (ranged between 6.5 and 7.5), and the carbon isotopic composition (δ¹³C = −25.8‰ to −26.3‰) are similar to those of planktonic microalgae from freshwater lakes (Table 5), and suggest negligible external (i.e., terrestrial) carbon inputs to the lake. A similar conclusion has been obtained for CM (Figure 7), where the TOC content of SPM was considerably higher (up to 8 wt.%.–10 wt.% [11]).

Measurements obtained in the sediment traps have provided sedimentation rates ranging between 0.2 and 0.6 cm/year in CM and around 1 cm/year in ST (

Table 6. Annual sedimentation rates and annual fluxes of organic carbon (C), nitrogen (N), and phosphorus (P) were calculated for different depths in ST and CM pit lakes.

San Telmo Pit Lake		10 m	40 m	100 m
Annual sedimentation rate (1)	cm/year	1.0	1.1	1.1
Total sediment thickness (2)	cm	23.2	23.5	23.8
C flux (3)	g m−2 year−1	8.0	8.5	10
N flux (3)	g m−2 year−1	1.3	1.4	1.6
P flux (3)	g m−2 year−1	0.03	0.03	0.04
Cueva de la Mora pit lake		11 m	21 m	38 m
Annual sedimentation rate (1)	cm/year	0.2	0.4	0.6
Total sediment thickness (2)	cm	8.3	14.5	24.9
C flux (3)	g m−2 year−1	33	36	24
N flux (3)	g m−2 year−1	0.8	1.2	1.2
P flux (3)	g m−2 year−1	0.2	0.1	0.1

⁽¹⁾ Annual sedimentation rates were calculated by averaging all measurements of thickness obtained in sediment traps installed at these depths between 2011 and 2014. ⁽²⁾ The total sediment thickness is a theoretical calculation of the thickness of the sediment layer formed in the pit lake bottom, considering the provided sedimentation rates and the corresponding age of each lake (around 25 years in San Telmo and 42 years in Cueva de la Mora). ⁽³⁾ The annual nutrient fluxes were calculated by considering the average weight and elemental composition of these sediments (as given in Table 5).

). These values vary seasonally and between years due to climatic and hydrological factors (e.g., higher detrital sediment input due to more abundant precipitation and runoff in wetter seasons). In ST, the amount and composition of sediments collected at different depths are rather homogeneous and reflect a sedimentation regime dominated by autochthonous production of chemical precipitates of ferric iron. The large size and surface area of this lake minimize the relative proportion of detrital sediments compared to the internal (chemical, biological) production. In the smaller lakes, the effect of runoff from the drainage basin is more visible, and the relative proportion of detrital sediments (composed of quartz, feldspars, and clay minerals) is more important than the autochthonous production. The theoretical sediment thickness calculated for these two lakes (~23–25 cm) is more or less consistent with that commonly observed in sediment cores taken from the deep part of these lakes (Figure S6 in Supplementary Materials).

Considering the compositional data on C, N, and P of the SPM settling to the lake bottom (Table 5) and the reported sedimentation rates, we have calculated average annual carbon and nutrient fluxes to the deep pit lake waters (Table 6). The calculated fluxes of organic carbon and phosphorus in CM (24–36 g C m⁻² year⁻¹ and 0.1–0.2 g P m⁻² year⁻¹) are notably higher than those of ST (8–10 g C m⁻² year⁻¹ and 0.03–0.04 g P m⁻² year⁻¹), while those of nitrogen are very similar (0.8–1.2 g m⁻² year⁻¹ N vs. 1.3–1.6 g m⁻² year⁻¹ N).

3.5. Influence of Nutrient Distribution on Phytoplankton Dynamics

The main sources and transformation pathways of nutrients identified in this study have been schematically illustrated in Figure 8. Several photosynthetic microalgae, including Zignematales, Chlamydomonadales, Ochromonadales, Coccomyxa spp., and different diatom species, have been observed in the mixolimnion of CM, HE-G, and ST [22]. These eukaryotes are typically found in acidic pit lakes [8,39,71,72,73] and can migrate to deeper levels to obtain higher amounts of nutrients like phosphorus or nitrogen, which are more abundant below the chemocline (Figure 3, Figure 4 and Figure 5). The concentration of phytoplankton communities in deeper levels forms DCM, as evidenced in vertical profiles of chlorophyll a concentration (Figure S7) [14]. In CM, these DCM are mostly composed of Coccomyxa sp. and Chlamydomonas acidophila [14], two photoautotrophs known for being highly sensitive to nutrient limitation (Chlamydomonas acidophila is very sensitive to inorganic phosphorus limitation [73], while Coccomyxa sp. is also sensitive to nitrogen [14]). Since pH and toxic metal concentrations show very little variation in the mixolimnion of these lakes (Table S3), the formation of DCM is considered to be mainly driven by the vertical distribution of nutrients described above. Further, the planktonic microalgae present in the DCM play a pivotal role in the bacterial activity of these acidic lakes since (1) they provide O₂ to allow Fe(II)-oxidizing metabolisms, and (2) their organic exudates are an important carbon source to sustain Fe(III)- and sulfate-reduction below the chemocline [11].

3.6. Influence of Nutrient Distribution on Bacterial Ecology

Despite the general belief that extreme acidity and concentration of different toxic metals are what limit microbial abundance and diversity in acidic mining lakes, microbial activity in these systems has been observed to be more closely associated with organic carbon production [74]. The studied pit lakes are known to host many species of acidophilic and metal-resistant bacteria and archaea [12,13,14,15,16,17,18], so that the slight differences in pH and metal concentrations (e.g., Fe, Al, Cu, Zn, As) existing between different depths in the studied lakes (Table S3) cannot explain the observed variations in the extent of certain microbial metabolisms throughout the water column. However, the microbial communities in these lakes include many heterotrophic microorganisms (e.g., Fe(III)- and sulfate-reducers), whose metabolisms are highly dependent on organic carbon availability [12,13,15,16]. Some organic carbon sources are related to microalgae, which produce certain metabolites. For example, both Dunaliella sp. and Pseudococcomyxa sp. can produce glycerol under certain stress conditions [75]. Thus, the production of glycerol or other organic compounds by microalgae may act as an important carbon source for the microbial community inhabiting these environments [18]. Therefore, the vertical distribution of nutrients (which in turn controls the migration of certain microalgae to deeper levels, as explained above) exerts an indirect control on bacterial activity in the water column of the studied pit lakes.

In addition to carbon availability, the quality of the available carbon compounds can also influence microbial activity. For example, in the upper photic levels of these lakes, the intense UV radiation causes the photoreduction of Fe(III) [37]. This photoreduction is usually coupled to photo-oxidation of refractory dissolved organic carbon [36,76,77]. Thus, light can stimulate the conversion of high-molecular-weight organic compounds to low-molecular-weight organic acids such as acetate, formate, or pyruvate [38]. This may be an additional factor contributing to higher microbial activity in the upper levels of these lakes compared to the deeper part [9,12,21].

The calculated carbon, nitrogen, and phosphorus fluxes (Table 6) should be sufficient to sustain heterotrophic microbial growth. However, the rates of ferric iron and sulfate reduction, and thus, the extent of biogeochemical cycling of iron and sulfur, have been reported to be considerably higher in shallow, mixolimnetic sediments than in deep, monimolimnetic sediments [9]. This has been ascribed to the differing amount and quality of the organic carbon in both types of sediment, as the deep basin sediments do not have the continuous supply of fresh organic carbon from benthic microalgae (Table 4). The same applies to the intense sulfate reduction activity detected in the chemocline of CM, which is mainly fed by organic exudates released by the phytoplanktonic communities forming the DCM [11]. This contrasts with the absence of bacterial sulfate reduction in the rest of the water column and in other pit lakes (e.g., ST) where phytoplankton biomass is much less abundant.

4. Conclusions

Our study has provided a comprehensive report on the aqueous concentration, ionic speciation, mineral sources, vertical distribution, and solid-phase fluxes of essential nutrients (phosphorus, nitrogen, carbon) in some of the world’s most acidic and metal-enriched pit lakes. The study suggests an important control of nutrient availability on the microbial ecology of the studied acid pit lakes. The relative abundance of acid-soluble phosphate in the host rocks does not represent a clear advantage because the leached phosphorus in the oxygenic and Fe(III)-rich mixolimnion is sequestered by ferric compounds formed in the water column, which leads to an important P-limitation in the photic zone. This P-limitation is consistent with observations in other acidic pit lakes of the world. However, the release of phosphorus during degradation of settling microalgal biomass, desorption from ferric compounds, and microbial reduction of Fe(III)-sediments causes a strong enrichment of this element in the monimolimnion, which may also reach the chemocline and sustain, through ionic diffusion, the growth of deep phytoplankton communities near the bottom of the mixolimnion.

Nitrogen is scarce in the rocks hosting the lakes, and only a limited input occurs via atmospheric deposition, followed by nitrogen uptake by algae, sedimentation, and organic matter degradation in the sediments. The latter process releases ammonium, which diffuses upward through the water column and reaches the chemocline. Ammonium-oxidizing archaea have been detected in sediments of the CM pit lake, which could mediate in the nitrogen cycle. High organic nitrogen contents have also been detected at depth in CM and HE-G, though the sources, nature, and significance of this nitrogen remain unknown and should be addressed in future studies. DIC (solely present as CO₂ aq.) and soluble SiO₂ are abundant in the mixolimnion and do not represent a limitation for microalgae and diatom growth.

The differences observed in the extent of phytoplankton development and bacterial activity between the relatively unproductive lakes (e.g., ST) and the more productive lakes (e.g., CM, HE-G, BRU) cannot be explained by differences in the concentration of toxic metals (e.g., As, Cr, Pb, Cu, Zn, Al), which show comparable contents in the photic zone of these lakes. These differences are most likely caused by a marked P-limitation existing in the former lakes due to the abundance of P-binding ferric iron colloids in the water column. In contrast, the observed differences in the intensity of several microbial metabolisms (e.g., sulfate and iron reduction) among different zones of a given pit lake (chemocline vs. monimolimnion, deep sediments vs. shallow sediments) seem to be mostly related to organic carbon availability for the heterotrophic microorganisms.

As a final remark, the presented data suggest that phosphorus amendment could be a reasonable and cost-effective measure to indirectly increase (by enhancement of primary production) natural rates of alkalinity-producing and metal-sequestering microbial metabolisms in these lakes. With more available phosphorus in the photic zone (e.g., by accumulation of P-rich waste from nearby agricultural or industrial activities), the volume and biomass of phytoplankton could be notably increased in the mixolimnion. This could also provoke higher amounts of settling phytoplankton biomass to deeper levels, which could in turn lead to higher rates of bacterial sulfate reduction below the chemocline, which are beneficial because this metabolism (1) increases pH and reduces acidity and (2) removes toxic metal(oid)s (e.g., As, Cu, Zn, Cd, Fe) through precipitation of insoluble metal sulfides (As₂S₃, CuS, ZnS, CdS, FeS). The technical and economic considerations of this artificial eutrophication technique are beyond the scope of this paper and deserve future studies.