2.1 Study area characterization
The study area corresponds to the hydrographic basin of Jucu River, located in the Center-South region of Espírito Santo State as Fig. 1 indicates.
More precisely this hydrographic basin is located between parallels 20º10 ’and 20º40’ south and meridians 41º9 ’and 40º16’ west. The basin area is mostly occupied by agricultural land use, with watercourses being predominantly used for electricity generation, crop irrigation, fishing, tourism, water supply, as well as industrial development. JRHB has a drainage area of approximately 2,092 km², equivalent to 209,200 hectares, covering six municipalities.
The hydrographic region of Jucu River belongs to the Atlantic Forest biome, with the main tributaries being Jucu - Braço Norte and Jucu - Braço Sul rivers, whose source is located in Pedra Azul region, which is a place of strong tourism within state.
According to KÖPPEN classification, the region has two climatic types: hot and humid; and tropical in altitude (mesothermal), which is the most frequent within Espírito Santo State. The precise classifications are (Oliveira, 2011):
• Aw: Humid tropical climate with rainy season in summer and dry in winter;
• Am: Humid tropical climate with no pronounced dry season;
• Cfa: mild humid climate;
• Cfb: humid temperate climate;
• Cwa: mild dry winter climate;
• Cwb: Dry winter temperate climate.
The hydrographic basin presents great climatic diversity, due to its great extension and rugged relief, whose altimetry levels vary between 0 to 2000 meters.
The main soil types in the JRHB, along with their characteristics, are (Oliveira, 2011):
• Oxisols:
Soil group with advanced evolution and significant performance of the latolization (ferralitization) process, resulting in intense weathering of primary mineral constituents, and even less resistant secondary ones, with relative concentration of resistant minerals and / or iron and aluminum oxides as well as hydroxides, with inexpressive clay mobilization or migration, ferrolysis, gleization or plintitization (Santos et al. 2018).
• Cambiosols:
Poorly developed soils with an incipient B-horizon. They present little advanced pedogenesis, which is evidenced by the soil structure development, with alteration of the source material, which is expressed by the almost absence of rock structure or sediment stratification. In addition, there is higher chroma, redder hues or higher clay content than those in the underlying horizons (Santos et al. 2018).
• Neossolos:
Poorly evolved soils, without a defined B-horizon. They correspond to soils in process of formation, either due to incipient performance of the pedogenetic processes or due to characteristics inherent from the source material (Santos et al. 2018).
• Argisols:
Group of soils with textural B and low Clay activity, as well as high Clay activity when combined with low base saturation or with an aluminum character. Advanced evolution with incomplete performance of the ferralitization process, in connection with kaolinitic-oxidic or virtually kaolinitic or vermiculite with hydroxy-Al between layers, when there is the presence of clay mobilization from the most superficial soil parts with concentration or accumulation in the subsurface horizon (Santos et al. 2018).
2.2 Land use and occupation mapping and changing analysis between 2005 and 2018
The methodological steps (Figure 2) for analyzing changes were performed in the ArcGIS® 10.3, SAGA and TerrSet computer applications.
Two images, from 2005 and 2018, were selected from MapBiomas platform to identify and map the classes of land use and occupation within Espírito Santo State (Table 1). The images have been later cropped based on the study area boundaries.
According to Mapbiomas (2018), all annual coverage maps of land use and occupation are produced from pixel by pixel classification (30 x 30 meters) of images from Landsat satellite (30-meter spatial resolution). The entire process is done with extensive machine learning algorithms in Google Earth Engine platform, which offers immense processing capability in the cloud. To facilitate algorithms tuning and the organization of all processing steps, 556 letters of 1 x 1.5 ° (lat / long) from IBGE have been used.
The general accuracy was obtained on MapBiomas website (Figure 4), which is calculated from the cross tabulation of sampling frequencies concerning mapped and real classes. Then, pixels are sampled to generate the reference data, which is composed by approximately 75,000 samples, with sampling size being already pre-determined by statistical sampling techniques. Finally, an evaluation of the reference database is performed through metrics that compare the mapped classes with the manual evaluation done by technicians.
2.3 Delimitation of permanent preservation areas (PPAs)
For PPAs delimitation, the four classes mentioned in Table 2 along with the criteria established by Law No. 12,651, from May 25, 2012 (BRASIL 2012) were considered.
• PPA for watercourses:
To obtain the PPA concerning watercourses, it was necessary a watercourse database, in line vector format, obtained from GEOBASES website (https://geobases.es.gov.br/). Subsequently, the database was clipped to the study area extension. Then, a 30-meter buffer was applied in the watercourses in order to delimit the marginal strips for preservation in each margin (PPA), whose value was considered once watercourses width does not exceed 10 meters.
• PPA for springs:
Springs were manually marked using ArcGIS® 10.3 editor and the hydrography dataset generated in the previous step. Each headspring was marked with a point, generating a vector spatial feature with 5,083 springs. Then a 50-meter buffer radius was created around each spring, resulting in the PPA for springs.
• PPA for slope:
Slopes were calculated based on the Digital Elevation Model (DEM) concerning the JRHB. Then, NoData was assigned for slopes below 45º while slopes above that value received 1 value. Subsequently, the raster resulting from this process was converted into a vector file, generating the PPAs for slope.
• PPA for hilltop:
For delimiting PPA concerning hilltop, the automatic delimitation extension (Oliveira and Filho 2016) was used, with the redesigned DEM as input.
• Total PPA mapping:
By grouping the individually acquired data for each PPA class, the map containing all existing PPA in the hydrographic basin was produced.
The total PPA map was confronted with land use and occupation map through tabular crossing, using the overlapping technique. Thus, it was possible to quantify the percentage of occupation for each land use class in the total PPA of the JRHB.
All methodological steps for mapping PPAs were performed in ArcGIS® 10.3 computer application and shown in Figure 3.
2.4 Confrontation of land use and occupation with PPAs
Through clip and intersect tools, available in ArcGis® 10.3, it was performed a confrontation between the thematic maps of land use and occupation and PPA. Each type of PPA was evaluated in separately. In addition, rocky outcrop, forest formation, water bodies, dune, mangrove, salt flat and other non-forest natural formation, were considered as suitable (not conflicting) to PPA, as they already correspond to classes of natural use. Following the forest code, the remaining land use and occupation classes were considered as conflicting ones when occurring in PPA (Figure 5).