Chemical and Physical Denudation Rates in the Poços de Caldas Alkaline Massif, Minas Gerais State, Brazil

Abstract

Chemical and physical denudation rates have been assessed in areas with different lithologies. Surprisingly, there are no studies that attempt to document these rates in the Poços de Caldas Alkaline Massif (PC), the largest alkaline magmatism in South America and an important Al supergene deposit in Brazil. Therefore, the chemical and physical denudation rates were assessed and explained in the PC. Surface water and rainwater samples were collected at the Amoras Stream basin, covering one complete hydrological cycle (2016). All samples were analyzed for dissolved cations, silica, anions, total dissolved solids (TDS), and total suspended solids (TSS). The results reflected the seasonal variation on discharge, water temperature, and electrical conductivity in the Amoras Stream, with most of the cations, anions, silica TDS, and TSS being carried in the wet season. Partial hydrolysis and silicate incongruent dissolution are the main water/rock interactions in the PC, with an atmospheric/soil CO₂ consumption rate of 1.6 × 10⁵ mol/km²/a. The annual fluxes of Cl⁻, PO₄³⁻, NO₃⁻, and Al³⁺ were significantly influenced from rainwater. Chemical and physical weathering rates were 4 ± 0.8 and 3.0 ± 0.6 m/Ma in the PC, respectively, indicating that under the current climatic condition, the weathering profile is in dynamic equilibrium.

1. Introduction

Chemical and physical denudations play an important role in the geomorphological modeling of the Earth’s surface [1]. The chemical denudation occurs due to chemical weathering processes, which are important mechanisms of atmospheric CO₂ removal and the consequent deposition of carbonates Ca²⁺ and Mg²⁺ (and smaller amounts of Fe²⁺ and Mn²⁺) in oceans, moderating the terrestrial climate [2,3,4]. On the other hand, the physical denudation is associated with the soil removal by erosion processes, promoting an increase or decrease in the weathering profile thickness. Both dissolved and suspended materials originate during chemical and physical denudation, respectively, and are released into the drainage and subject to the transportation and sedimentation processes [1]. Therefore, quantifying chemical and physical denudation rates can be obtained using the fluvial geochemistry.

Studies of chemical and physical denudation rates, involving a mass balance in large or small watersheds, have focused mainly on fluvial geochemistry, especially in granitic and basaltic watersheds under different climates, relief, vegetation, and land use conditions [5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23]. These studies conclude that temperature and runoff are the main parameters controlling the water/rock interactions and soil removal worldwide. In Brazil, several studies have been carried out related to chemical and physical denudation for silicate rock watersheds [24,25,26,27,28,29,30,31,32,33,34,35,36,37,38], showing that the transfer of dissolved and suspended material from the lithosphere to the hydrosphere is controlled by local conditions, such as lithology and climate, associated with topography and soil cover. Fernandes et al. [39] studied the human influences on chemical and physical denudation rates of the Paulista Peripheral Depression (PPD), São Paulo State, and indicated that those rates in the PPD were approximately 4- and 7.5-fold higher than the natural denudation rates, respectively, evidencing the complexity of the human–landscape systems. Recently, Conceição et al. [40] explained the main chemical and physical denudation processes to produce supergene P, Ti, Nb, and REE deposits in the Tapira and Catalão I alkaline-carbonatite complexes.

Southeastern Brazil’s passive margin is marked by the presence of Mantiqueira and Tocantins orogenic systems, where there are outcrops of igneous and metamorphic rocks, and part of the Paraná Sedimentary basin and São Francisco Craton [37]. The Poços de Caldas Alkaline Massif (PC) is located in the Tocantins Orogenic Belt (Figure 1), Minas Gerais State, and is the largest alkaline magmatism in South America (800 km²) [41]. PC comprises a suite of volcanic and plutonic rocks, mainly tinguaites, nepheline syenites, and phonolites, associated with the Upper Cretaceous alkaline magmatism, with the ages ranging from 75 to 86 Ma and the main magmatic episodes related to a short age span (1 to 2 Ma) at ~79 Ma [42]. During the water/rock interactions, the chemical denudation generated an important Brazilian Al supergene deposit [43]. In addition, PC also hosts the Osamu Utsumi U mine, currently disabled, and the Morro do Ferro Th-REE deposit; both deposits are located in hydrothermally altered areas [44]. Surprisingly, there are no studies assessing the chemical and physical denudation rates in the PC, limiting the understanding of the interplay between chemical and physical denudations on this area hosting a significant Al supergene deposit.

Here, we assess the chemical and physical denudation rates in the PC using as the study area the Amoras Stream basin, one small watershed not affected by domestic and industrial wastes and/or agricultural processes, avoiding complexities associated with investigating larger rivers draining a variety of distinct rock types. Therefore, it only required correcting the rainwater influences on the annual flux of elements or compounds during the fluvial transport. After the necessary corrections, the chemical denudation rate was quantified and explained, and the atmospheric/soil CO₂ consumed during the water/rock interaction was modeled. To determine physical denudation, the annual flux of dissolved TSS due to soil removal was measured and the balance between physical and chemical denudation rates was discussed. Our results provide new insights into the geological and climatic factors that control the current landscape evolution in the PC, generating useful information to complement global knowledge about the chemical and physical denudation rates in similar alkaline rocks elsewhere in the tropics or under different climate conditions in future studies.

2. Study Area

Geomorphologically, the PC stands out as a remarkable topographic feature, with an outer rim composed by tinguaites, especially prominent in the northern and southern segments, with the highest elevations over 1200 m or even over 1600 m (Figure 2). The basement rocks around the PC crop out as a dissected plateau with irregular topographic ridges and peaks, prevailing altitudes ranging from 900 to 1000 m. The landscape evolution in the PC was briefly defined in six evolutionary stages [44]: (1) regional uplift attributes to activity in an underlying magma chamber; (2) period of volcanic activity; (3) caldera subsidence caused by intense tectonism; (4) renewed uplift, with intrusions along the radial and circular fractures; (5) major ring-fracture volcanism; and (6) intrusions accompanied by the hydrothermal activities.

Due to the presence of Al supergene deposits [45], open pit mining areas have occurred in the PC for bauxite exploitation, with production capacity varying from 10,000 t/year to 1,000,000 t/year [46]. In addition to the open pit mining areas, activities related to the aluminum production chain are located in the PC, producing metallic aluminum by the Bayer process by the digestion of crushed bauxite in a concentrated caustic substance (NaOH) at elevated temperatures. During this industrial process, a large volume of red mud is generated, composed of hematite (Fe₂O₃), goethite (FeO(OH)), quartz (SiO₂), gibbsite (Al(OH)₃), calcite (CaCO₃), sodalite (NaAlSi₄O₁₂Cl), kaolinite (Al₆Si₂O₂(OH)₄), and rutile (TiO₂) [47].

The study area is a small watershed, without anthropic activities, located in the northwest PC, called Amoras Stream, with an area of 18 km² (Figure 3). Located on the left bank of the lower course of Antas Stream, the drainage flows into the Bortolan Reservoir. The predominant climate is Cwb [46], i.e., humid temperate with dry winters and moderately hot summers, with an annual average temperature of 17 °C [48]. In the historical series from 2003 to 2013, the annual average rainfall in the study area was 1910 mm [48]. In 2016, the annual rainfall was 1760 mm, according to rainfall data collected at the rainfall station located at the Federal University of Alfenas (UNIFAL), Campus Poços de Caldas. The rainy season (October–March) was responsible for 82.4% of the total rainfall in 2016.

The predominant rocks outcropping in the Amoras Stream basin are nepheline syenites, tinguaites, and phonolites, occupying ~70% of its territory, while volcanoclastic deposits represent ~30% (Figure 1). These rock types have a similar chemical and mineralogical composition, being characterized by the presence of alkali feldspar [orthoclase—KAlSi₃O₈ and sanidine—(K_0.75,Na_0.25)AlSi₃O₈], followed by nepheline [(Na_0.75,K_0.25)AlSiO₄], and clinopyroxenes [aegirine—NaFe³⁺(Si₂O₆) and aegirine–augite—(Ca_0.75,Na_0.25)(Mg_0.5,Fe²⁺_0.25,Fe³⁺_0.25)(Si₂O₆)] [41]. The range of volcanoclastic deposits is composed of breccias, agglomerates, and volcanic tuffs, and may contain phenocrysts of orthoclase and clinopyroxene, with a matrix of clinopyroxene and nepheline, and phlogopite may also occur [KMg₃AlSi₃O₁₀(OH)₂] [41].

3. Sampling and Analytical Techniques

Twelve surface water samples were collected in the Amoras Stream for hydrochemical analyses, covering one complete hydrological cycle (from January 2016 to December 2016). The dates were selected because they have the largest variation in rainfall, with the minimum values in July–August and the maximum ones in January–February [48], and the consequent discharge. Discharge was calculated using the float method [35] (Equation (1)). Rainfall samples were collected during the same period (n = 18), using a “bulk” (dry and wet deposition)-type collector, with each rain event representing the precipitation that occurred during a 24 h period. The total rainfall volume was measured using a pluviometer.

$$Q=\frac{A·D·C}{T}$$

(1)

where Q is the discharge (m³/s), A is the cross-section of the river (m²), D is the distance to estimate the water velocity in the river (m), C is the empirical coefficient (0.9), and T is the time the floating body will take to travel the distance D (s).

The concentrations of dissolved chemical species were quantified from 1000 mL of surface water and rainwater samples, which were stored in labeled polyethylene bottles at 4 °C until chemically analyzed. The parameters pH, water temperature (only for surface waters—Temp in °C), and electrical conductivity (EC in μS/cm) were characterized directly at the sampling site, using a multiparametric probe, YSI 556. The pH electrode was calibrated using high-purity standards (4.00 and 7.00 ± 0.01 to 25 ± 0.2 °C). The conductivity meter was calibrated using a standard KCl solution (1.0 mmol/L) of known conductivity (147 μS/cm at 25 °C).

Filtered surface waters (0.45 μm Millipore membrane, Burlington, MA, USA) were used for analyzing dissolved cations (Na⁺, K⁺, Ca²⁺, Mg²⁺) and anions (Cl⁻, PO₄³⁻, SO₄²⁻, and NO₃⁻) by ion chromatography (Dionex ICS-90, Thermo Scientific, Waltham, MA, USA), calibrated with Dionex standard solutions and equipped with the analytical columns (IonPac^® CS12A 4 × 250 mm and IonPac^® AS14A 4 × 250 mm) (Thermo Scientific, Waltham, MA, USA). The detection limit was 0.001 mg/L [37]. The HCO₃⁻ was measured using the Gran method [49]. Filtered samples were also used in the quantification of dissolved SiO₂ (0.2 to 100 ± 1 mg/L) and Al³⁺ (0.01 to 0.80 ± 0.01 mg/L), using a Hach DR 2800 spectrophotometer (Hach, Loveland, CO, USA). The sum of dissolved cations, anions, and silica was considered as the total dissolved solids (TDS), while the total suspended solids (TSS) were quantified by gravimetry.

To validate the results obtained for surface waters and rainwaters, the normalized ionic mass balance (NICB) was used, considering the ratio between the sum of cations (Σ⁺) and anions (Σ⁻). According to Mosello et al. [50], the ionic equilibrium must be between ±10%. All samples showed NICB less than 10% for all surface waters and rainwaters collected, with variation between 5.3 and 5.5%. Therefore, all samples were considered in this study using a NICB. For the surface waters, the discharge weighted average element/compound concentrations (C_E in mg/L) were obtained though Equation (2). The rainfall weighted averages for the study period (P_E—in mg/L) were also calculated using Equation (2), using rainfall volume in the event instead of instantaneous discharge.

$$C_E=\frac{{\sum }_{i=1}^n C_i·Q_i}{{\sum }_{i=1}^n Q_i}$$

(2)

where C_i is the ion concentration for the ith measurement (mg/L); Q_i is the discharge of the stream on the day of the ith measurement (m³/s).

4. Results

4.1. Surface Waters

All data obtained for surface waters are in

Table 1. Data measured in the Amoras Stream.

Sampling	Q 1	pH	EC 2	T 3	Na+	K+	Mg2+	Ca2+	Al3+	SiO2	HCO3−	NO3−	PO43−	Cl−	SO42−	TDS 4	TSS 5
	(m3/s)		(µS/cm)	(°C)	(mg/L)
01/26	1.92	6.2	22	27.0	3.0	1.6	1.0	2.0	0.18	1.0	14.4	0.5	0.06	1.4	0.6	25	10
02/16	1.81	6.0	23	25.0	3.1	0.7	0.7	2.1	0.20	4.0	15.7	0.2	0.07	0.8	0.4	28	12
03/08	1.27	6.3	22	24.0	2.9	1.2	1.2	2.1	0.60	3.5	16.6	0.4	0.04	1.3	0.6	30	15
04/05	0.48	6.4	21	24.0	2.9	1.0	1.5	2.4	0.43	4.4	19.1	0.3	0.05	1.1	0.3	33	3
05/09	0.30	6.8	19	25.0	2.7	1.4	1.4	2.6	0.50	2.5	19.4	0.6	0.03	1.0	0.4	33	4
06/17	0.47	6.3	14	24.0	3.1	1.7	1.7	2.3	0.30	1.4	21.3	0.6	0.06	1.7	0.7	35	5
07/17	0.20	6.1	14	26.0	3.2	1.7	1.7	2.9	0.19	1.8	22.9	0.6	0.02	1.5	0.6	37	3
08/12	0.19	6.7	12	20.0	3.7	1.9	1.9	3.0	0.38	3.1	24.6	0.7	0.03	1.6	0.8	42	3
09/15	0.08	5.9	18	33.0	3.3	1.8	1.8	3.2	0.40	1.8	23.7	1.0	0,04	1.2	0.5	39	4
10/13	0.30	6.3	20	35.0	3.3	2.0	2.0	3.5	0.46	1.9	25.1	1.2	0.08	1.2	0.7	41	4
11/04	0.18	6.7	21	33.0	3.0	1.9	1.5	3.2	0.45	2.1	23.5	0.9	0.05	1.3	0.8	39	6
12/20	0.58	6.4	22	23.0	2.9	1.0	1.0	2.8	0.41	1.7	21.0	0.4	0.01	0.9	0.6	33	5
CE	0.65	6.2	19	27.1	3.1	1.5	1.4	2.7	0.37	2.4	20.6	0.6	0.05	1.2	0.6	33	7

¹ Discharge; ² electrical conductivity; ³ water temperature; ⁴ total dissolved solids (TDS); and ⁵ total suspended solids (TSS).

. The seasonal influence in the Q values occurred in the Amoras Stream, with the Q values in the wet season (October to March) being higher than the dry season (April to September). The highest and lowest Q values were characterized in January (1.92 m³/s) and September (0.08 m³/s), respectively, with an annual average of 0.65 m³/s. The annual pH average was 6.2, ranging from 5.9 (September) to 6.8 (May), with no significant seasonal variability. The annual EC average was 19 μS/cm, with an EC amplitude of ~ 5 μS/cm between the dry and wet seasons. The annual T averages ranged from 20.0 °C (August) to 35.0 °C (October), with a seasonal influence well defined. In relation to the dissolved concentration of ions and silica, the lowest and highest values were obtained in wet and dry seasons, respectively. The dissolved cation with the highest C_E was Na⁺ (3.1 mg/L), followed by Ca²⁺ (2.7 mg/L), K⁺ (1.5 mg/L), Mg²⁺ (1.4 mg/L), and Al³⁺ (0.37 mg/L). The C_E for SiO₂ was 2.4 mg/L, varying between 1.0 and 4.4 mg/L. The predominant dissolved anion was HCO₃⁻, with a C_E of 20.6 mg/L, with a C_E of 0.6, 0.05, 1.2, and 0.6 mg/L for NO₃⁻, PO₄³⁻, Cl⁻, and SO₄²⁻, respectively. The discharge weighted averages of TDS and TSS were 33 and 7 mg/L, respectively.

To verify the seasonality on TDS and TSS concentrations due to tropical climate in the PC, the empirical power-law function, [TDS] or [TSS] = a.Q^b, was used [51]. This power-law function is utilized to analyze the concentration–runoff (C-Q) relationship in watersheds, where a and b are constants and b can be used to indicate the [TDS] and [TSS] behavior due to hydrologic changes [52]. The relationship between the [TDS] and the Q was inverse (Figure 4a), indicating the highest [TDS] during the dry season, with a b value of −0.17, representing the chemostatic behavior, with [TDS] being stable in relation to changes in discharge [53]. The [TSS] was directly related to the discharge, with a b value of 0.61 indicating an enrichment behavior (Figure 4a), where the concentration increases along with discharge. The higher [TSS] occurred in the wet period due to soil erosion after rainfall events. Figure 4b suggests that the majority of the TDS and TSS were transported during the summer period, in accordance with most world rivers [51,54,55,56]. In Brazil, the inverse relationship between Q and TDS was observed in tropical watersheds composed of different lithologies [30,31,32,33,34,35,36,37,38,39,40,57,58].

4.2. Rainwaters

Table 2. Precipitation (P), pH, electrical conductivity (EC), dissolved chemical species, total dissolved solids (TDS), and total suspended solids (TSS) measured in rainwater sampled in the Poços de Caldas Alkaline Massif.

Sampling	P	pH	EC	Na+	K+	Mg2+	Ca2+	Al3+	SiO2	HCO3−	NO3−	PO43−	Cl−	SO42−	TDS	TSS
	(mm)		(µS/cm)	(mg/L)
01/25	16.6	5.6	19	1.6	1.3	0.3	1.2	0.20	<0.02	3.7	0.5	0.03	2.5	0.2	11	2
02/15	26.2	5.7	8	1.4	0.1	0.1	1.0	0.21	<0.02	3.2	0.3	0.04	0.4	0.2	7	2
02/17	8.2	5.3	9	1.2	0.9	0.1	0.7	0.18	<0.02	3.3	0.8	0.02	0.5	0.4	8	1
02/21	22.6	5.7	6	0.7	0.4	0.1	1.0	0.17	<0.02	2.8	0.5	0.04	0.3	0.2	6	<1
02/23	18.6	5.8	5	1.4	0.2	0.1	0.9	0.15	<0.02	3.8	0.4	0.04	0.3	0.1	8	<1
02/27	9.4	5.6	20	1.4	1.7	0.2	1.4	0.18	<0.02	3.6	0.6	0.02	2.3	0.2	12	<1
03/03	52.6	5.6	5	0.9	0.9	0.1	0.8	0.20	<0.02	1.2	0.5	0.03	1.0	0.4	6	<1
03/10	8.6	6.2	8	1.2	0.8	0.2	1.0	0.18	<0.02	3.3	0.2	0.01	1.0	0.7	8	<1
03/16	27.4	5.9	5	0.9	0.4	0.1	0.6	0.21	<0.02	2.4	0.3	0.03	0.5	0.1	6	<1
05/31	24.4	5.8	7	0.7	0.2	0.2	0.8	0.12	<0.02	3.1	0.9	0.01	0.3	0.5	7	<1
06/02	40.0	5.1	15	1.4	0.8	0.1	1.0	0.24	<0.02	1.4	0.7	0.06	1.2	0.3	8	<1
06/06	20.2	5.6	10	1.0	0.6	0.4	1.0	0.12	<0.02	2.9	1.1	0.03	1.1	0.1	9	<1
06/07	29.0	6.3	17	1.3	1.2	0.2	0.9	0.23	<0.02	3.1	1.3	0.03	1.9	0.3	10	<1
08/21	34.2	5.1	5	1.2	0.2	0.3	1.1	0.20	<0.02	2.3	1.5	0.01	0.4	0.5	8	<1
10/24	22.4	5.7	11	1.5	1.3	0.2	1.1	0.21	<0.02	2.4	1.0	0.08	1.2	0.7	10	2
10/25	38.8	6.4	7	1.4	0.2	0.2	1.0	0.20	<0.02	1.8	0.9	0.01	0.2	0.5	6	1
11/05	75.0	6.1	6	0.9	0.3	0.1	0.9	0.18	<0.02	2.1	0.4	0.09	0.1	0.1	5	2
12/10	22.8	5.3	2	1.3	0.1	0.1	0.9	0.17	<0.02	2.1	0.8	0.01	0.2	0.3	6	2
PE	27.6	5.7	9	1.1	0.6	0.1	0.9	0.19	<0.02	2.7	0.7	0.03	0.8	0.3	8	1

presents the values of precipitation, pH, EC, and the dissolved concentration of chemical species, TDS and TSS. The pH values of the samples collected vary between 5.1 and 6.4. The P_E calculated for pH was 5.7, indicating “clean” rainwater, with pH values representing the partial solubilization of CO₂ in rainwater (H₂O) to form H₂CO₃ [59]. The measured EC had an average of 9 μS/cm, indicating relatively low values of EC in the PC. The P_E are highly variable, with Na⁺ and HCO₃⁻ being responsible for 40% and 60% of the total cation and anion sum, respectively, suggesting that they are the most abundant ions in the rainwater. P_E concentrations showed the following tendency for ion concentrations: Na⁺ > Ca²⁺ > K⁺ > Al³⁺ > Mg²⁺ for cations and HCO₃⁻ > Cl⁻ > NO₃⁻ > SO₄²⁻ > PO₄³⁻ for anions.

PC is located in southeast Brazil, with the Atlantic Ocean being ~ 250 km to the east. Thus, the chemical composition of rainwater samples in this study should be less influenced by sea salt than rainwater samples from coastal areas, due to the distance of the sampling sites from the sea. This is confirmed by the smaller P_E concentration of Na⁺ (51.7 μeq/L) measured in the PC, being lower than those values obtained for different southern Brazilian cities located close to the Atlantic Ocean, e.g., Ilha Grande (142.2 μeq/L), Rio de Janeiro [60]; Niterói (62.6 μeq/L), Rio de Janeiro [61]; and Cubatão (82.0 μeq/L), São Paulo [62]. The lack of marine influence can be confirmed comparing the ionic ratios measured in this work (Ca²⁺/Na⁺ = 0.95, Mg²⁺/Na⁺ = 0.26, K⁺/Na⁺ = 0.29, Cl⁻/Na⁺ = 0.41, SO₄²⁻/Na⁺ = 0.12, and HCO₃⁻/Na⁺ = 0.79) and in the rainwater collected at Cubatão [62], indicating important terrestrial or anthropic elements and compounds in the rainwaters of PC. The acidification potential of rainwater is due to the presence of SO₄²⁻ and NO₃⁻, as well as other organic acids, with NH₃ and Ca²⁺ being the main responsible parties for the neutralization process [63,64]. The linear correlation coefficients (

Table 3. Relationship (p ≤ 0.01) among all ions measured in rainwater sampled in the Poços de Caldas Alkaline Massif (n = 18).

	H+	Na+	K+	Mg2+	Ca2+	Al3+	HCO3−	Cl−	NO3−	PO43−	SO42−
H+	1.00
Na+	−0.10	1.00
K+	−0.05	0.39	1.00
Mg2+	−0.11	−0.01	0.47	1.00
Ca2+	−0.12	0.53	0.26	0.47	1.00
Al3+	0.10	0.46	0.31	0.34	0.08	1.00
HCO3−	0.14	0.21	0.23	−0.34	0.26	−0.37	1.00
Cl−	−0.04	0.47	0.89	0.44	0.58	0.29	0.32	1.00
NO3−	−0.18	0.10	0.06	0.48	0.15	−0.01	−0.12	0.08	1.00
PO43−	−0.26	0.26	0.13	0.06	0.11	0.27	−0.14	0.05	0.23	1.00
SO42−	0.08	0.14	0.14	0.04	0.07	0.14	−0.19	−0.03	0.29	0.45	1.00

) indicated that the contribution of these ionic species was not significant for the acidification/neutralization of rainfall in the PC.

5. Discussion

5.1. Natural and Anthropogenic Inputs into Annual Fluxes of Dissolved Elements

The contribution of different sources to fluvial fluxes was calculated using a mass balance model expressed in Equation (3), modified from White and Blum [12].

F_w = F_R − F_A

(3)

where F_w is the annual flux of dissolved ions, TDS and TSS (t/km²/a); F_R is the river annual flux calculated using Equation (4) (t/km²/a), with $\overline{Q}$ being the average discharge of the study period (m³/s) and A the surface area of the watershed (km²); and F_A is atmospheric inputs obtained through relation P_E (Table 2) and total precipitation in the study period (1760 mm) (Equation (5)) (t/km²/a).

$$F_R=\frac{C_E·\overline{Q}}{A}$$

(4)

$$F_A=\frac{P_E·P}{1000}$$

(5)

The values of F_R, F_A, and F_w are in

Table 4. Mass balance (t/km²/a) in the Poços de Caldas Alkaline Massif.

Flux	Na+	K+	Mg2+	Ca2+	Al3+	SiO2	HCO3−	Cl−	NO3−	PO43−	SO42−	TDS	TSS
FR 1	3.5	1.7	1.6	3.0	0.4	2.7	23.4	1.4	0.7	0.1	0.7	39.3	7.0
FA 2	2.1	1.1	0.3	1.7	0.3	0.0	4.7	1.5	1.3	0.1	0.6	13.3	1.0
Fw 3	1.4	0.6	1.3	1.3	0.1	2.7	18.7	−0.1	−0.6	0.0	0.1	26.0	6.0

¹ Annual flux of dissolved ions, TDS and TSS; ² annual river flux; and ³ atmospheric inputs.

. Figure 5 shows the contributions of different sources in the PC. Positive mass balance values were found for Na⁺, K⁺, Mg²⁺, Ca²⁺, SiO₂, HCO₃⁻, TDS, and TSS. Local bedrock composition clearly controls the chemical weathering, where all samples are located as nearly silicate end-members (Figure 6). Plots of surface waters and rainwaters in the PC (Figure 7) also confirm that surface waters are controlled by local water/rock interactions, according to those proposed in the Gibbs diagram [65].

The weathering index (R_E—Equation (6)) [6] indicates the main chemical weathering process in different watersheds, with R_E ≈ 0 characterizing the total hydrolysis process and R_E ≈ 2 and R_E ≈ 4 related to the partial hydrolysis of minerals, with kaolinite montmorillonite formation, respectively [30].

$$R_E=\frac{3·F_K+3·F_{Na}+2·F_{Ca}+1.25·F_{Mg}-F_{Si{O}_2}}{0.5·F_K+0.5·F_{Na}+F_{Ca}+0.75·F_{Mg}}$$

(6)

where F_K, F_Na, F_Ca, F_Mg, and F_SiO2 are dissolved fluxes of K⁺, Na⁺, Ca²⁺, Mg²⁺, and SiO₂ related to chemical weathering (mol/a).

The R_E of 2.47 was obtained, revealing that the chemical weathering is the main water/rock interactions occurring in the PC under tropical climate, with partial removal of SiO₂ from the profile and complete leaching of Na⁺, K⁺, Mg²⁺, and Ca²⁺.

Table 5. Main water/rock interactions in the Poços de Caldas Alkaline Massif.

Hydrolysis

2 KAlSi3O8 (orthoclase) + 2 CO2 (aq) + 11 H2O (liq) → Al2Si2O5(OH)4 (kaolinite) + 2 K+ (aq) + 2 HCO3− (aq) + 4 H4SiO4 (aq)

2 (K0.75,Na0.25)AlSi3O8 (sanidine) + 2 CO2 (aq) + 11 H2O (liq) → Al2Si2O5(OH)4 (kaolinite) + 1.5 K+ (aq) + 0.5 Na+ (aq)+ 2 HCO3− (aq) + 4 H4SiO4 (aq)

Incongruent Dissolution

(Na0.75,K0.25)AlSiO4 (nefeline) + 4 CO2 (aq) + 4 H2O (liq) → 0.75 Na+ (aq) + 0.25K+ (aq) + Al3+ (aq) + 4 HCO3− (aq) + H4SiO4 (aq) NaFe3+(Si2O6) (aegirine) + 4 CO2 (aq) + 6 H2O (liq) → Na+ (aq) + Fe3+ (aq) + 4 HCO3− (aq) + 2 H4SiO4 (aq) (Ca0.75,Na0.25)(Mg0.5,Fe2+0.25,Fe3+0.25)(Si2O6) (aegirine-augite) + 4 CO2 (aq) + 6 H2O (liq) → 0.75 Ca2+(aq) + 0.25 Na+ (aq) + 0.5 Mg+3 (aq) + 0.25 Fe2+ (aq) + 0.25 Fe3+(aq) + 4 HCO3− (aq) + 2 H4SiO4 (aq) KMg3AlSi3O10(OH)2 (phlogopite) + 10 CO2 (aq) + 10 H2O (liq) → K+ (aq) + 3 Mg2+ (aq) + Al3+ (aq) + 10 HCO3− (aq) + 3 H4SiO4 (aq)

Forming Kaolinite and Goethite

3 Al3+ (aq) + H2O (liq) → 3 H+ (aq) + Al(OH)3 (aq)

2 Al(OH)3 (aq) + 2 H4SiO4 (aq) → Al2Si2O5(OH)4 (kaolinite) + 5 H2O (liq) 4 Fe2+ (aq) + O2 (aq) + 6 H2O (aq) → 4 FeOOH (goethite) + 8 H+ (aq) Fe3+ (aq) + 2 H2O (liq) → FeOOH (goethite) + 3 H+ (aq)

summarizes the main chemical weathering processes in the PC. Hydrolysis of orthoclase and sanidine forms kaolinite. Silicate incongruent dissolution (nefeline, aegerine, augirine–augite, and phogopite) forms kaolinite and goethite. Interestingly, the intensity of the current chemical weathering process in the PC is lower than that occurring since the Upper Cretaceous, where bauxite profiles were formed, due to changes in climatic factors over geological time [43].

The F_w values close to 0 indicate that annual fluxes for Al³⁺, Cl⁻, and SO₄²⁻ are only associated with atmospheric inputs. The negative mass balance value for NO₃⁻ is attributed to local atmospheric fallout. Burnt fossil fuels from vehicles are significant anthropogenic sources of Cl⁻, SO₄²⁻, and NO₃⁻ associated with atmospheric pollution. Although the annual Al³⁺ flux is close to zero, there is an important annual flux of Al³⁺ being transported by Amoras Stream. Under the current chemical weathering process, and with the pH values for surface waters, Al³⁺ should not be released from the weathering profiles, being incorporated into the secondary minerals formed during the water/rock interactions. Therefore, the atmospheric input of Al³⁺ due to the aluminum production chain, which releases significant amounts of particulate matter, fluorine compounds, and Al³⁺ into the atmosphere [66], is the main responsible party for the elevated annual flux of Al³⁺ in the PC. Dissolved Al³⁺ is potentially toxic to aquatic biota, interfering with gill respiration processes and promoting fish mortality events [67], as reported in 1983 and 1985 at Bortolan Reservoir, due to change in the pH values and the presence of Al³⁺ and other ions [68].

5.2. Chemical Denudation Rate in the Poços de Caldas Alkaline Massif

Using the [TDS], the F_w value was 26 t/km²/a (Table 4). Despite the small size of the Amoras Stream, it is useful to compare the F_w values obtained in this study with values measured in alkaline watersheds under different climatic and tectonic settings in Brazil or elsewhere (

Table 6. Annual flux (F_w) in basaltic watersheds.

Area	Fw (t/km2/a)	Temperature (°C)	Runoff (mm/a)	Reference
Poços de Caldas Alkaline Massif	26	17.0	1139	This study
Paraná CBF province	30	22.5	565	[35]
Tapira	29	20.7	512	[40]
Catalão I	32	22.6	541	[40]
São Miguel	35	16.0	730	[14]
Deccan traps	37	25.0	460	[17]
Iceland	36	2.0	1883	[23]
Réunion island	102	18.7	2430	[13]
Java island	326	26.6	4050	[13]

). Temperature and runoff are the most important parameters controlling the F_w values in watersheds composed of alkaline rocks [17]. The F_w value for PC is slightly lower than those obtained for the Paraná CBF province and Tapira and Catalão I alkaline-carbonatite rocks, areas located under tropical climate in Brazil or in São Miguel, Deccan Traps, and Iceland. In addition, it is lower than the F_w values obtained for Réunion and Java islands due to elevated runoff characterized on these islands (2430 and 4050 mm/a, respectively). Interestingly, the relative low flux of dissolved cations and anions out of the system suggests that local relief and internal processes dominate water/rock interaction in the PC.

The relationship F_CO2 = F_Na + F_K + 2F_Ca + 2F_Mg [2,30] allowed us to determine the atmospheric/soil CO₂ consumption rate (F_CO2 in mol/km²/a) in the PC. The F_CO2 was 1.6 × 10⁵ mol/km²/a in the PC, a value practically similar to the world continental average of F_CO2 (1.61 × 10⁵ mol/km²/a) [39]. However, this rate is higher than the Amazon (0.3 × 10⁵ mol/km²/a) [30] and Orinoco (0.7 × 10⁵ mol/km²/a) watersheds [2]. Considering the tropical climate in southeastern Brazil, the F_CO2 in the PC is lower than those calculated for the Paraná CBF province (0.4 × 10⁶ mol/km²/a) [35] and igneous and metamorphic rocks belonging to Ribeira Belt (0.2 × 10⁶ mol/km²/a) [37]. In addition, this rate is lower than in carbonate areas (0.8 × 10⁶ mol/km²/a) [69] and basaltic watersheds, such as São Miguel (0.6 × 10⁶ mol/km²/a) [14], Deccan Traps (1.3 × 10⁶ mol/km²/a) [17], Iceland (0.7 × 10⁶ mol/km²/a) [23], and Réunion island (2.3 × 10⁶ mol/km²/a) [13].

As previously discussed, the partial hydrolysis assumes that all silicate minerals are transformed into kaolinite during the water/rock interaction process in the PC, allowing the use of mass balance models based on the concentration of Na⁺, K⁺, Mg²⁺, and Ca²⁺, and SiO₂. Here, a mass balance model based on SiO₂ will be used to quantify the chemical denudation rate (C_D in m/Ma) in the PC (Equation (7)) [15], considering the weathering into saprolite is isovolumetric, as well as the SiO₂ flux being conservative and only due to the weathering of alkaline rocks.

C_D = F_SiO2/(S₀ − S_S)

(7)

where F_SiO2 is the annual flux of SiO₂ (t/km²/a); S₀ is the SiO₂ content in the bedrock (t/m³); and S_S is the SiO₂ content in the saprolite (t/m³).

The annual flux of dissolved SiO₂ in the Amoras Stream, after atmospheric corrections, was 2.70 t/km²/a at PC (Table 4). SiO₂ contents in bedrocks and saprolites were 52.48 and 40.41 wt%, respectively [43]. When these values are associated with a density of 2.95 t/m³ for bedrocks and 1.85 t/m³ for saprolites [40], the S₀ and S_s values of 1495 and 747 kg/m³ can be obtained, respectively. Therefore, the chemical weathering rate for the alkaline rocks in the PC was 4 ± 0.8 m/Ma, considering the uncertainty arising from the number of samplings [61]. Using the same uncertainty from the other alkaline watersheds in Brazil, this rate is practically similar to the chemical denudation rates obtained for the Paraná CFB province (6 ± 1.2 m/Ma) [35] and for Tapira (4 ± 0.8 m/Ma) and Catalão I (5 ± 1.0 m/Ma) alkaline-carbonatite rocks [40], important P, Ti, Nb, and REE deposits.

5.3. Physical Denudation Rate in the Poços de Caldas Alkaline Massif

The F_w value associated with soil removal was 6 t/km²/a (Table 4) or 16 kg/km²/day. According to Meybeck et al. [70], the F_w value of 16 kg/km²/day obtained in the PC can be classified as low daily TSS flux (from 10 to 50 kg/km²/day) due to flat relief occurring within the PC. The physical denudation rate (P_D in m/Ma) of 3.0 ± 0.6 m/Ma was easily calculated from the TSS transport and the density of soils (1.85 t/m³) in the PC. The difference between the chemical and physical denudation rates (D = C_D − P_D) can be used to quantify if the weathering profile thickness is increasing (D > 0) or decreasing (D < 0) [15]. The values of C_D (4 ± 0.8 m/Ma) and P_D (3 ± 0.6 m/Ma) can be considered similar in the PC due to the uncertainties, indicating that under the current climatic condition, the weathering profile is in dynamic equilibrium. Doranti-Tiritan et al. [71] studied the landscape evolution of the Poços de Caldas region based on thermochronology data and 3D thermokinematic modeling and indicated a modern physical denudation rate of ~17 m/Ma in areas with higher altitudes and slopes, and that of practically 0 in flatter areas at low altitudes. The rate obtained here suggests that the current denudation rate in areas at low altitudes in the PC is not 0, but 3.0 ± 0.8 m/Ma.

It is evident that the land use/land cover changes (LULCC) can increase the soil removal on the local, regional, and global scale. For example, in São Paulo State, the LULCC was directly related to increasing rates of soil removal, i.e., 0.03, 3.5, and 12.6 t/ha/a, for the natural conditions and current land use and with the expansion of sugar cane crops [72]. The current land use in the Amoras Stream basin is divided into pastures (areas of livestock breeding) (32%), reforested areas (29%), urban areas (14%), natural vegetation (13%), and Riparian Forest and water bodies (12%) (Figure 3), showing clearly that human–landscape systems affect physical denudation in the PC. With the continuous LULCC, the physical denudation rates will probably be higher than that obtained in this study, increasing the natural process of soil removal, directly altering the dynamic equilibrium and, consequently, the geomorphic responses in the PC.

PC is located in the Tocantins Orogenic Belt, and, unfortunately, there are no studies in small or large watersheds in this geological province reporting on the physical weathering rates. Conceição et al. [73], using a combined ⁴⁰Ar/³⁹Ar (Mn oxyhydroxides) and (U/Th)-He (goethites) geochronology, estimated physical denudation rates in other alkaline-carbonatite complexes in the Tocantins Orogenic Belt, e.g., Araxá (3.4 m/Ma—Minas Gerais State) from 66 Ma to the present and at Catalão I (4.7 m/Ma—Goiás State) from 45 Ma to the present (see location of both alkaline-carbonatite complexes in Figure 1). Interestingly, the long-term physical denudation rates in the Tocantins Orogenic Belt match the current physical denudation in the PC, suggesting analogous landscape evolution histories. The long-term physical denudation rates in the Mantiqueira Orogenic Belt are higher than those obtained in the Tocantins Orogenic Belt, with the rates increasing towards the Atlantic margin [74].

6. Conclusions

PC is the largest alkaline magmatism in South America and an important Al supergene deposit in Brazil. This study intended to determine the chemical and physical denudation rates occurring in the PC, using the fluvial transport dynamics, allowing the following conclusions:

• The concentration of dissolved cations, anions, and silica in surface waters increased during the dry period in relation to the wet period. The same behavior is observed for pH, EC, temperature, TDS, and TSS. The relationship between the [TDS] and the Q was inverse, representing the chemostatic behavior. The [TSS] was directly related to the discharge, indicating an enrichment behavior. The higher [TSS] occurring in the wet period is due to soil erosion after rainfall events.
• The pH values in rainwaters vary between 5.1 and 6.4, with a weighted average of 5.7, indicating that the pH values are close to the “clean” rainwater (5.6) due to the partial solubilization of carbon dioxide to form carbonic acid (H₂O + CO₂ → H₂CO₃). Na⁺ and HCO₃⁻ are responsible for 40% and 60% of the total cation and anion sum, respectively, suggesting that they are the most abundant ions in the rainwater. The chemical composition of rainwater samples should be less influenced by sea salt than rainwater sampled from coastal areas, due to the distance of PC from the Atlantic Ocean.
• Positive mass balance values are found for Na⁺, K⁺, Mg²⁺, Ca²⁺, SiO₂, HCO₃⁻, TDS, and TSS in the PC. The R_E of 2.47 reveals that the chemical weathering is the main water/rock interactions occurring in the PC under tropical climate, with partial removal of SiO₂ from the profile and complete leaching of Na⁺, K⁺, Mg²⁺, and Ca²⁺. Hydrolysis of orthoclase and sanidine forms kaolinite. Silicate incongruent dissolution (nefeline, aegerine, augirine–augite, and phogopite) forms kaolinite and goethite. The F_w values that are negative or close to 0 indicate that annual fluxes for Al³⁺, Cl⁻, NO₃⁻, and SO₄²⁻ are associated with atmospheric pollution (aluminum production chain and burnt fossil fuel from vehicles).
• The F_w value due to [TDS] was 26 t/km²/a, a similar value to those obtained for the Paraná CBF province and Tapira and Catalão I alkaline-carbonatite rocks, areas located under tropical climate in Brazil. The F_CO2 associated with chemical weathering was 1.6 × 10⁵ mol/km²/a in the PC, also similar to the F_CO2 world continental average. The chemical weathering rate was 4 ± 0.8 m/Ma, this rate being practically similar to the chemical denudation rate obtained for the Paraná CFB province and Tapira and Catalão I alkaline-carbonatite rocks.
• The F_w value associated with soil removal was 6 t/km²/a or 16 kg/km²/day, which can be classified as low daily TSS flux due to flat relief occurring within the PC. The physical denudation rate was 3.0 ± 0.6 m/Ma in the PC. The difference between the chemical and physical denudation rates indicated that under the current climatic condition, the weathering profile is in dynamic equilibrium. The LULCC are responsible for the increase in the soil removal in the PC, showing clearly that the human–landscape system affects physical denudation in the PC.

Future studies comparing the present physical rates with Cenozoic long-term soil removal rates, using ⁴⁰Ar/³⁹Ar and (U/Th)-He geochronology, must be carried out in the PC, allowing us to assess whether the physical denudation rates were constant or episodic during the Cenozoic. In addition, the weathering geochronology associated with a more extensive and systematic use of ⁴⁰Ar/³⁹Ar geochronology in all alkaline lithologies and low-temperature thermochronometers, such as apatite fission-track (AFTT) or (U-Th)/He (AHe) thermochronology, will help complement the history of the magmatism, exhumation, weathering, erosion, and paleoclimate in the PC.