Segmentation of the Apodi-Mossoró river and pedogenetic changes in soils in alluvial environments

ABSTRACT The Apodi-Mossoró River Basin is important for the west region of the state of Rio Grande do Norte (RN), Brazil. The rapid increase in population and consequent need for space has resulted in anthropogenic actions, such as the segmentation of the river course within the municipality of Mossoró, RN. The objective of this study was to assess the effects of this segmentation on soils in alluvial environments. Soil collection points were defined in natural and anthropized areas in the Mossoró urban segment; soil profiles were described and collected; and samples of the soil 0-10 and 10-20 cm were collected. The samples were subjected to physical and chemical analyses to determine granulometry, density, aggregates, pH in water, available phosphorus (P), exchangeable bases, potential acidity, total organic carbon, nitrogen, and micronutrients (Cu and Zn). Maps of location and land use and cover were developed. The soil found in the natural area was classified as Neossolo Fluvico Psamitico eutrico, presenting typical physical and chemical characteristics of this soil class, with variable texture and P contents along the soil profile. The soil in the anthropized area was classified as Planossolo Haplico Carbonatico vertissolico, with Ap Horizon presenting a sandy loam texture, since it is a horizon formed due to anthropogenic activity in the environment. All horizons and profiles presented high pH. The evaluated areas formed different groups regarding soil physical and chemical attributes, with overlapping points that can be explained by anthropogenic actions, even in the environment considered natural.


INTRODUCTION
The municipality of Mossoró, Rio Grande do Norte (RN), Brazil, is in the region of the Apodi-Mossoró River Basin, which is the second largest river basin in the state, drains an area of approximately 15,500 km 2 , and has a significant importance for the regional economy, mainly for oil extraction and sea salt production activities; land use for irrigated agriculture and fruit production and extensive livestock; and limestone mining (CARVALHO; KELTING; SILVA 2011).
The occupational structure of Mossoró often caused severe floods in areas near the river during rainy periods.Thus, the municipal government carried out works between 1976 and 1986 that altered the course of the Apodi-Mossoró River within the urban area of this municipality (LIMA, 2007), dividing it into segments, a process called segmentation, to reduce flooding in the urban zone (MOURA, 2014).The segmentation project involved deepening of existing channels and opening secondary branches in the river, specifically at points surrounding Santa Luzia, Rincao, and Coroa islands.The management of urban waters requires to consider not only land use and occupation but also the fluvial evolution of watercourses.Thus, urban planning of occupation along these water courses should consider fluvial dynamics in decision-making processes (BENINI; ROSIN, 2017).
The growth of cities has strong impacts on soils, as materials from buildings forms layers that do not connect with the original material in these areas (LADEIRA, 2012).Environmental impacts from human activities are significant and can result in soil changes, generating several natural and social problems.Probably, anthropogenic actions have caused changes in soil attributes of alluvial environments along the Apodi-Mossoró River.
Human activities make environments subject to changes, including in soils.Soil changes resulting from anthropogenic activities can be morphological, chemical, physical, and mineralogical (TEXEIRA; LIMA, 2016).Natural soils are modified through disposal of solid waste on them, combined with other mechanical processes in urban environments that cause losses of identification horizons (ANTONIO; METTERNICH; TOMMASELLI, 2017).
Intense anthropogenic actions in alluvial environments around the Apodi-Mossoró River may have altered chemical and physical characteristics of soils in these environments.Thus, the objective of this study was to understand anthropogenic effects on soils in alluvial environments caused by the segmentation of the river course within the urban zone of Mossoró, RN, Brazil.

Study area characterization
The municipality of Mossoró, state of Rio Grande do Norte, Brazil, is in the West Potiguar mesoregion; its territorial area is 2,099.33Km2, and population of 264,577 in habitants (IBGE,2023).The climate of the region was classified as BSh, dry hot, according to the Köppen classification, with a mean annual temperature of 26.5 °C and irregular rainfall, with mean annual depths lower than 650 mm, concentrated from February to June; Caatinga vegetation is predominant in the region (ALVARES et al., 2013).
The geological formations of the Mossoró region include rocks from the crystalline basement and sediments from the Potiguar Basin and the Barreiras Group (JACOMINE et al., 2015).

Description and collection of soil profiles and samples
The study was conducted in the urban area of Mossoró, which is affected by the segmentation of the Apodi-Mossoró River course that crosses the urban perimeter.Points around the river were defined (Table 1) and trenches were opened in selected locations for morphological description of soil profiles (P1 and P2), following the Field Soil Description and Collection Manual (SANTOS et al., 2015), and soil classification up to the 4 th categorical level of the Brazilian Soil Classification System (SANTOS et al., 2018).P1: Profile 1; P2: Profile 2; A1, A2, and A3: anthropized areas; N1, N2, and N3: natural areas.
Profile 1 was in the river trichotomized branch and Profile 2 was in the river natural branch.Three points along the natural course (N1, N2, and N3) and three points along the anthropized course of the Apodi-Mossoró River were selected (A1, A2, and A3) for collecting soil samples (Figure 1).
Simple soil samples were collected from the 0.0-0.10 and 0.10 -0.20 m layer in each point.The samples were air-dried and sieved through a 2.0 mm mesh sieve to obtain the bulk soil.
According to the soil attributes, undisturbed soil samples were collected for soil density analysis using the clod method.

Physical analyses
Physical analyses were carried out in triplicate, using the bulk soil.Granulometry was determined using the method proposed by Teixeira et al. (2017).Soil aggregates were analyzed using the wet sieving method.Soil samples were sieved in the field using 4.76 mm and 2.00 mm mesh sieves, taken to the laboratory, placed on a set of sieves (2.0, 1.0, 0.50, and 0.25 mm meshes), immersed in water, and agitated in a mechanical vibration shaker for four minutes.

Chemical analyses
Chemical analyses were carried out in triplicate, using the bulk soil, according to the manual of soil analysis method of the Brazilian Agricultural Research Corporation (Embrapa).The pH in water was measured in a soil-liquid suspension at a ratio of 1:2.5 (TEIXEIRA; CAMPOS; SALDANHA, 2017b); electrical conductivity (EC) was measured in a water extract solution using a conductivity meter (TEIXEIRA; CAMPOS; PIRES, 2017).
Total organic carbon (TOC) was determined according to the method proposed by Mendonça and Matos (2017) and adapted from Yeomans and Bremner (1988), through oxidation of organic matter by 0.2 mol L -1 dichromate of potassium in a sulfuric acid medium, followed by titration with 0.2 mol L -1 ammonium ferrous sulfate.
Nitrogen (N) was determined using the distillationtitration method (Kjeldahl) (BALIEIRO; ALVES, 2017), in which the ammonium (NH 4 + ) produced during digestion with sulfuric acid (H 2 SO 4 ) is distilled in a strongly alkaline medium.The condensed NH 4 + is collected in a boric acid (H 3 BO 3 ) solution and titrated with a hydrochloric acid (HCl) solution.

Cartographic products
Location and land use and cover maps were developed using the open-source GIS software QGIS 3.16 Hannover.A mosaic of Planet satellite images (Dove satellites), with a spatial resolution of three meters, and the cartographic basis of the Brazilian Institute of Geography and Statistics (IBGE, 2020) were used for manual vectorization.Coordinates of sampling points in the study area were collected in the field.

Statistical analysis
Statistical analysis of the evaluated soil profiles consisted of multivariate analysis using the software Statistica 7.0 (STATSOFT, 2004), based on the Pearson's method (p ≤ 0.05) for variables to ensure that the attributes had minimal correlations for their use in the data matrix.The correlation matrix established a pattern for analytical results for the application of principal component analysis (PCA) (HAIR et al., 2009).
Principal components with eigenvalues higher than 1 were extracted in the factorial analysis; the factor axes were rotated using the Varimax method.A loading factor value of 0.70 was set to consider significant loadings (HAIR et al., 2009).
Two diagrams (Factors 1 and 2) for physical and chemical attributes were generated for PCA.According to these data, a two-dimensional diagram was developed to distinguish the areas and a vector projection diagram was developed to determine the most sensitive soil attributes for the differentiation of the study area (HAIR et al., 2009).

Soil profiles
The soils found in the natural and anthropized alluvial areas had distinct classifications (Table 2).The soils in areas along the river course subjected to segmentation were classified as Planossolo Haplico Carbonatico vertissolico (Figure 2a).The soil along the natural course of the Apodi-Mossoró River were classified as Neossolo Fluvico Psamitico eutrico (Figure 2b).
The distinction of soil classes is based on pedogenetic processes occurring in the fluvio-marine area where the Mossoró region is located.Clay illuviation and gleying were the predominant processes identified in the anthropized areas, resulting in the formation of Neossolos Fluvicos only along the riverbanks (natural areas), through the deposition of alluvial sediments.
Rev. Caatinga, Mossoró, v. 36, n. 4, p. 843 -856, out. -dez., 2023 847 This denotes that the segmentation of the river course exposed these areas to continuous flooding for a period, contributing to the gleying process.Accumulation of alluvial sediments was found in anthropized areas, with formation of Ap horizon presenting physical and chemical characteristics different from the underlying horizons.
Regarding the morphological characterization, Profile 1 presented a single-grain structure in the surface horizon (Ap) and prismatic structure in subsurface, which breaks down into angular blocks and less frequently into subangular blocks.This type of hydromorphic environment in semiarid climates is typical of the pedogenetic process of ferrolysis, which causes an abrupt textural change resulting from destruction of silicate clays on the soil surface due to redox cycles of iron and can cause the complete removal of iron oxides, resulting in soil discoloration (KAMPF; CURI., 2015) Moisture was found throughout Profile 1 in all horizons, except the Ap horizon; Btg1, Btg2, and Btg3 horizons presented high plasticity and stickiness, preventing assessments of dry soil consistency, which is characteristic of Planossolos.Profile 2 presented typical morphological variability of Neossolos Fluvicos, with variations in color, structure, and consistency along the AC and C1 horizons (Table 2).
Regarding the physical characterization, the Ap horizon of Profile 1 presented a sandy loam texture; this horizon is formed from alluvial sediments probably deposited due to the anthropogenic action of segmentation of the river course.It differed from other horizons, which presented sandy clayey loam (Btg1) and clay (Btg2 and Btg3) textures.The textures of the horizons throughout the Profile 2 varied, mainly from sandy loam to sand (Table 3).
The Ap horizon of Profile 1 showed higher soil density than the others horizons, confirming the sandy texture of the recently deposited material due to anthropogenic intervention in the area.The soil density in Profile 2 varied, which can be explained by the genesis of Neossolos Fluvicos, in which fluvial deposits are the essential source material (PINHEIRO et al., 2020).Aggregate stability can be used to evaluate soil quality under different land uses and managements (SOUZA, 2015); it tends to be higher under higher input of organic matter.Therefore, the organic matter content has a direct and significant correlation with the distribution of larger-sized aggregates (RIBON et al., 2014).
Profile 1 presented geometric mean diameters (GMD) higher than 1 in all horizons and smaller weighted mean diameters (WMD), except for the Btg3 horizon (Table 4).The higher values found for the anthropized area may be due to the added plant material, which contributes to increases in organic matter and, consequently, promotes soil aggregation.Profile 2 presented smaller GMD and WMD, which varied from 0 to higher than 1 throughout the profile.WMD increases as the percentage of coarse aggregates retained on the sieve increases.GMD represents an estimate of the size of the most common aggregate class.Smaller WMD and GMD are a consequence of the breakdown of soil aggregates due to anthropogenic interventions (LOSS et al., 2015).
Overall, both profiles presented variable texture, mainly the anthropized area, which is typical of fluvial environments: alkaline, low Na + , and rich in P, TOC, and exchangeable bases (Ca 2+ and Mg 2+ ) (Table 4).Currently, the Apodi-Mossoró River presents a strong eutrophication, therefore, these results are important to explain the contamination potential of water, mainly affecting the areas where the segmentation was carried out.This confirms that the segmentation resulted in land use and occupation of previously non-existent areas, which was the main responsible factor for the environmental contamination.
Soil physical and chemical attributes change by anthropogenic intervention processes, whose effects cause damages to surface water sources in river basins due to sediment runoff, which can directly contribute to eutrophication processes in these systems (MEDEIROS, 2016).
All horizons and profiles presented high pH, which is consistent with the high contents found for cations, mainly Ca 2+ and Mg 2+ , and low contents found for H+Al (Table 4).These results may be explained by the flooding process that occurs in these soils, which decreases the redox potential and, consequently, increases pH.Water availability is connected to increases in soil pH (SILVA et al, 2012).Similarly, all profiles and horizons presented low EC and Na + levels (Table 4).
High P contents were found in the Ap and Btg1 horizons of Profile P1 (Table 4), which is characteristic of soils with anthropic A horizon (SANTOS et al., 2018).However, P contents varied throughout Profile 2, which is characteristic of Neossolos Fluvicos.According to Medeiros (2016), the use of alluvial soils increases P and N contents, causing decreases in water quality and ecological functioning of aquatic systems, i.e., increasing eutrophication.
TOC contents were higher in the Ap horizon of Profile 1 (Table 4).The profiles presented a high C to N ratio (C/N) (Table 4).Higher C/N in soils is associate with the hydromorphic aspect, which hinders carbon degradation and facilitates N losses by leaching (SANTOS, 2019).Table 3. Physical attributes of soil profiles evaluated in anthropized and natural areas around the Apodi-Mossoró River.

Random soil sampling points evaluated in the anthropized and natural areas of the Apodi-Mossoró river
The soil sampling conducted at random points in the natural and anthropized river course areas indicated changes in the soil physical and chemical characteristics throughout the studied alluvial environment.
The granulometric composition of the anthropized area showed a predominance of the total sand fraction in surface and subsurface layers, as well as in the natural area (Table 5).All areas presented small WMD and GMD close to 1, except in the 0.10-0.20 m soil layer of the natural area, which presented a value close to 2. All means of soil density were close to 1.0 g cm -3 .Hor: horizon; EC: electrical conductivity; H+Al: acidity potential; Na + : sodium; K + : potassium; P: phosphorus; Ca2 + : calcium; Mg 2+ : magnesium; BS: base saturation; TOC: total organic carbon; N: nitrogen; C/N: C to N ratio; Cu: copper; Zn: zinc.The natural area, i.e., without segmentation, presented higher TOC and P contents than the anthropized area after segmentation (Table 6), differing from the results found for soil profiles.One hypothesis that could explain these results is the absence of actual natural areas, as the riverbanks have been improperly used, for example, for waste dumping, husbandry, and discharge of domestic effluents.
Regarding the soil samples from the anthropized areas, sodium contents in the 0.0-10.0layer stood out, exceeding 26 cmol c Kg -1 , resulting in an ESP of 25%, classifying this layer as sodic (SANTOS et al., 2018).

Land use and cover
An extensive urban area can be observed around the Apodi-Mossoró River (Figure 3).Additionally, a large area of bare soil was found, reinforcing that these areas are constantly used for disposal of materials or other anthropogenic activities.Araújo et al. (2012) reported that areas with bare soils, corresponding to deforested areas, are mostly in peripheral zones o the Mossoró, specifically between the East-West bridge and the Barrocas dam.Furthermore, the main sources of pollution found along the riverbanks were sewage systems, a brick industry, sand extraction, agricultural activities, waste dumping, and animal rearing.The natural area (before segmentation) and the anthropized area (after segmentation) formed distinct groups, which tended to be different even between depths (Figure 5).Different factors were significant for distinguishing the areas, despite human activities along the alluvial environment through the intense land use and occupation in the areas adjacent to the river.
Considering the projection vectors of factors 1 and 2 (Figure 4), the attributes WMD, CEC, ESP, Na + , Mg 2+ , TOC, and N affected the distinction of the first sampling point in the anthropized area on the surface (A1) from the other areas.The factors that had most affected the 0.10-0.20 m soil layer of the anthropized area were pH, clay, silt, BS, and Ca 2+ .
Three factors were generated, considering eigenvalues higher than 0.70 a significant; factor 1 contributed with 65.75% for distinguishing the areas, by the effects of the attributes EC, Na + , P, Ca 2+ , Mg 2+ , BS, CEC, ESP, C/N, Zn, TS, silt, clay, WMD, and Ds (Table 7).The natural area (before segmentation) and the anthropized area (after segmentation) formed distinct groups, which tended to be different even between depths (Figure 5).Different factors were significant for distinguishing the areas, despite human activities along the alluvial environment through the intense land use and occupation in the areas adjacent to the river.
Considering the projection vectors of factors 1 and 2 (Figure 4), the attributes WMD, CEC, ESP, Na + , Mg 2+ , TOC, and N affected the distinction of the first sampling point in the anthropized area on the surface (A1) from the other areas.The factors that had most affected the 0.10-0.20 m soil layer of the anthropized area were pH, clay, silt, BS, and Ca 2+ .
Regarding the natural areas, the attributes that most affected the surface layer (0.0-0.10 m) were EC, P, Zn, H+Al, Ds, TS, and Ds.The 0.10-0.20 soil layer m was significantly affected by C/N, differing from the other areas.
Considering the projection diagram of vectors (Figure 6) and respective correlation coefficients, the attributes TOC, P, N, and Na + affected the distinction of the Ap horizon from the other horizons.The attributes that most affected the C1 horizon of Profile 1 were GMD, WMD, Mg 2+ , and H+Al.The C2 and C3 horizons of Profile 1 and the A horizon of Profile 2 were grouped together, with higher effect of CEC, Ca 2+ , pH, K + , clay, and silt.The C1, C3, C5, and C6 horizons of Profile 2 were grouped together, with higher effect of the attributes TS, Ds, C/N, ESP, BS, and Zn; the C2, C4, C7; and the C8 horizons of Profile 2 were grouped with higher effect of EC and Cu.
Five factors were generated to distinguish the profiles and horizons, explaining 87.72% of the cumulative variance.Factor 1 accounted for 36.37% of the variance; the discriminant variables were Ca 2+ , CEC, C/N, TS, and silt (Table 8).
R. N. S. LIMA et al.

Figure 1 .
Figure 1.Map of location of soil profiles and soil sampling points in the study area.

Figure 3 .Figure 4 .
Figure 3. Map of land use and cover in alluvial environments around the Apodi-Mossoró River.

Figure 5 .
Figure 5. Order diagram of Principal Components (PC) for the different studied areas for PC1 and PC2.

Figure 6 .
Figure 6.Projection diagram of soil attributes vectors for the evaluated soil profiles for the Principal Components 1 and 2.

Figure 7 .
Figure 7. Order diagram of Principal Components (PC) for the evaluated soil profiles for PC1 and PC2.

Table 1 .
Location and geological units of soil profiles and random sampling points in the urban area of the Apodi-Mossoró River.

Table 2 .
Morphological description of soil profiles evaluated in anthropized and natural areas around the Apodi-Mossoró River.

Table 4 .
Chemical attributes of soil profiles evaluated in the anthropized and natural areas of the Apodi-Mossoró River.

Table 5 .
Means of soil physical attributes in random sampling points in the anthropized and natural areas of the Apodi-Mossoró River.

Table 6 .
Means of soil chemical attributes in random sampling points in the anthropized and natural areas of the Apodi-Mossoró River.

Table 7 .
Correlation coefficients of principal components (Factors 1, 2, and 3) for soil attributes of the different evaluated areas.