**2. Materials and methods**

#### **2.1 Description and characteristics of the basins and areas studied**

The study was conducted in two small paired basins (**Figure 1**), Cumaru (4,127 ha) and São João (2,518 ha), both of which are tributaries of the Maracanã River Basin, located in the municipality of Igarapé-Açu, Pará, Brazil. The two watersheds are located between latitudes from 1°12<sup>0</sup> 00"S to 1°16<sup>0</sup> 00"S and longitudes from 47° 32<sup>0</sup> 00"W to 47°34<sup>0</sup> 00"W. Igarapé-Açu is located in the Northeast Pará Mesoregion and Bragantina Microregion, about 120 km from the state capital, Belém. The main access routes are federal highway BR-316 and state highway PA-127. The two basins have easy access by dirt roads from the municipal seat, an important logistical factor for choosing them. These areas were selected because they have the predominant land-use systems in the region. They have highly permeable soils, predominantly by acrisols and ferralsols and gently rolling topography, cut by shallow channels (igarapés), with the presence of floodable marginal areas occupied by riparian forest stands (called igapós). The coverage mainly consists of secondary forested areas with different ages (capoeiras), small farm plots, and pastures with varying dimensions [9, 10]. Burning releases chemical compounds into the soil, making it temporarily fertile for crops. However, these compounds are rapidly lost to groundwater through leaching or carried away by runoff into the igarapés due to the physical characteristics *Soil Solution Chemistry in Different Land-Use Systems in the Northeast Brazilian Amazon DOI: http://dx.doi.org/10.5772/intechopen.101856*

#### **Figure 1.**

*The area was studied using soil solution extractors.*

of the soil [11]. In a previous study, the pH of the soil in the Cumaru microbasin was found to vary from 4.8 to 4.9 at the depths studied, lower than the pH levels of the soils in the São João watershed, which ranged from 5.2 to 5.6 [7].

The predominant soils at the higher elevations of the Cumaru basin, where the relief is gently rolling, are classified as Xanthic Acrysol epicarenic, associated with a smaller proportion of Xanthic Ferralsol endoarenic, in addition to the occurrence of Petric Plinthosols. At lower elevations, the predominant soils are arenosols with hydromorphic characteristics, associated with small occurrences of Gleysols with indiscriminate texture in a narrow strip in ravines and other areas of frequent flooding of the drainage network. In the São João basin, the soils are very similar to those in the Cumaru basin, also with a predominant class of Xanthic Acrisolsepiarenic, associated with a smaller proportion of Xanthic Ferralsols endoarenic, along the soil with occurrences of lateritic concretionary horizons (Petric Plinthosols) at the higher elevations, with gently rolling terrain. At lower elevations, arenosols also prevail, with hydromorphic characteristics of sand or loamy sand texture in a very narrow strip near the streambed, widening slightly near the outlet of Igarapé São João into the Maracanã River [7]. They have acid pH, as mentioned, and low cation-exchange capacity (CEC), between 4.27 and 4.37, as described by Da Silva et al. [7].

The land-use characteristics of the areas studied are presented in **Table 1**. For each watershed, the location of the sampling points and the respective land use and agricultural management are identified, including the code assigned to each area. The two microbasins, although having similar land uses, differ regarding the intensity of these uses.

The riparian vegetation in Cumaru is near the beds of watercourses, which are classified as first and second order until reaching the Maracanã River, which is a


#### **Table 1.**

*Location of the areas studied with different land-use classes.*

third-order watercourse. The original vegetation of this basin was mainly equatorial forest, of which there are only a few remaining areas associated with the hygrophilous forest of alluvial plains around and along springs, streams, and rivers, where also exist hygrophilous floodplain forests. In the São João watershed, which also empties into the Maracanã River, the current prevailing forest regime is a secondary latifoliate forest in various development stages, resulting from clearance of the original equatorial forest, together with remnants of hygrophilous forest in alluvial plains along watercourses [7]. Riparian forest stands protect these igarapés and the fluvial water quality, and contribute to the groundwater stock, unlike what happens in areas subject to high surface runoff, mainly in unmanaged pastures, which in rainy periods suffer large losses of nutrients due to burning and the consequent absence of biomass from leaves and roots in the surface soil.

The pastures in both basins are unmanaged, without specific treatment. During the period studied, the pasture area in the Cumaru watershed contained about 1,200 head of cattle, while the pasture in the São João watershed contained only about 600 animals. For the slash-and-burn system, in both microbasins we prepared areas covering 0.5 ha in which cassava was planted. With respect to the secondary vegetation and agroforestry systems, these were already established in the two basins, so we demarcated areas of

0.5 ha. Finally, for the chop-and-mulch system, we also prepared areas of 0.5 ha, where we cut the secondary vegetation and left the biomass on the ground, composed of trunks, branches, and leaves. The ages of the secondary forest areas are given in **Table 1**.

#### **2.2 Collection of samples and laboratory analyses**

Soil solution extractors were installed six months before the monthly sampling campaigns, which occurred from March 2014 to April 2015. These extractors were installed in the Cumaru and São João watersheds in 12 areas measuring 0.5 ha in each basin, representing the land-use systems—riparian vegetation, secondary forest up to 20 years of recovering, slash-and-burn agriculture, chop-and-mulch agriculture, agroforestry system, and pasture. All told, we installed 96 extractors at depths of 30 and 60 cm (four at each depth) in the land plots in each use system.

The soil solution extractors consisted of porous capsules connected to amber glass collector jars (capacity of 250 mL) for analyses of DIC (dissolved inorganic carbon), DOC (dissolved organic carbon), and DON (dissolved organic nitrogen); and clear glass bottles (capacity of 1000 mL) for analyses of cations and anions—chloride (Cl), nitrate (NO3 ), phosphate (PO4 ), sulfate (SO4 ), sodium (Na<sup>+</sup> ), ammonium (NH4 + ), potassium (K+ ), magnesium (Mg<sup>+</sup> ), and calcium (Ca<sup>+</sup> ). The collector jars were washed after each sample collection with running water and 1% hydrochloric acid (HCl) solution, followed by Milli-Q water, and then dried and replaced.

We first measured the volume of water in the extractor jars. Then the samples were transported to the laboratory of Embrapa Amazônia Oriental in the city of Belém, for filtering and storage at 4°C until conducting the analyses. For the analysis of DOC and DON, the samples were filtered through glass microfiber membranes (porosity of 0.7 μm), while for the other analyses, cellulose acetate membranes were used (porosity of 0.45 μm). To investigate the cations and anions, the samples were conserved in thymol (C10H14O) and submitted to ion chromatography with a chromatograph (Dionex DX120, USA).

The physical and chemical data of the soil samples from the region were obtained and described by Da Silva et al. [7].

### **2.3 Statistical analyses**

We performed comparisons of the average concentrations in mg L<sup>1</sup> of the inorganic ions, cations, and anions, and organic and inorganic carbon dissolved in the soil solution samples from sites with different land uses over time. We also calculated correlations (Pearson correlation for parametric data and Spearman correlation for nonparametric data).

We used the Minitab 16.0 software to compute descriptive statistics and tests for normal distribution (Shapiro–Wilk test*,* p > 0.05) and comparisons of the means and variances. The parametric (normally distributed residues and homocesdatic variances) data were submitted to analysis of variance (ANOVA) for one factor (F test of statistical equality between means, Tukey test), while for interpretation of nonparametric data (not normally distributed), the Kruskal-Wallis test was used. After the mathematical transformation of the data, DON was the only parameter for which the values after this procedure had a normal distribution. The reciprocal transformation (1/Y) was used to stabilize the variance, in the sense of minimizing the effect of possibly very high values of Y. With this test, there was no evidence against the normality of the residuals because the points were all closely distributed along a

straight line. In this case, one-way ANOVA was applied for a pairwise comparison of the means between land uses of the two microbasins. On the other hand, for the variables NO3 , NH4 + , and DOC, the Kruskal-Wallis test was applied to compare the means of the nonparametric data. **Table 3** reports the mean values of the concentrations of the variables measured in relation to each land-use class.
