Influence of Land Use on Trophic State Indexes in Northeast Brazilian River Basins

Environ Monit Assess (2019) 191:77 https://doi.org/10.1007/s10661-019-7188-7

Influence of land use on trophic state indexes in northeast Brazilian river basins

Olandia Ferreira Lopes & Felizardo Adenilson Rocha & Lucas Farias de Sousa & Daniela Mariano Lopes da Silva & Andrique Figueiredo Amorim & Ronaldo Lima Gomes & André Luiz Sampaio da Silva Junior & Raildo Mota de Jesus

Received: 7 August 2018 /Accepted: 2 January 2019 # Springer Nature Switzerland AG 2019

Abstract Eutrophication is a natural process within the from 2008 to 2015, at thirteen (13) sampling sites, with ecological succession of aquatic ecosystems that results quarterly collections. The results showed that, among from nutrient inputs to water bodies, especially limiting three basins analyzed, Cachoeira River basin presented elements such as phosphorus and nitrogen. However, the worst values for trophic state index (TSI) due to the the anthropogenic activities in river basin influence high level of anthropization, while best results were areas accelerate the eutrophication process of water found in Una basin. It was verified that land use exerted bodies. Eutrophication is a global problem and consid- a significant influence on the water quality of bodies of ered one of the most relevant reasons of aquatic envi- water evaluated. ronments’ degradation. In this context, watercourses that make up the Eastern Water Planning and Manage- Keywords Eutrophication . Degradation . ment Region (RPGA) receive high pollutant contribu- Anthropogenic action . Effluents tions due to release of wastewater and agriculture diffuse sources from cities located in influence area. The present study aims to evaluate the land use effect in trophic state Introduction of the water bodies in Eastern RPGA basins. The Carlson Trophic State Index in 1977, adjusted by Eutrophication is a natural process of succession in Lamparelli 2004, was used to determine the eutrophica- bodies of water. This phenomenon occurs due to the tion degree of the three river basins (Almada, Cachoeira, entry of nutrients especially the most limiting ones such and Una) located in the Eastern RPGA. The nutrient and as phosphorus (P) and nitrogen (N) into aquatic envi- chlorophyll a data were obtained from the Monitoring ronments. However, anthropogenic actions in river ba- Program (Monitora) of Environment and Water Re- sins (agricultural practices, the launching of sanitary, sources Institute of Bahia (INEMA), covering the period agroindustrial, and industrial wastewater without

O. F. Lopes : L. F. de Sousa : D. M. L. da Silva : Bahia 45201-570, Brazil R. L. Gomes : A. L. S. da Silva Junior : R. M. de Jesus (*) Universidade Estadual de Santa Cruz (UESC), Rodovia Jorge F. A. Rocha Amado, km 16, Bairro Salobrinho, Ilhéus, Bahia CEP: 45662-900, Instituto Federal de Educação, Ciência e Tecnologia da Bahia Brazil (IFBA), Av. Sérgio Vieira de Mello, 3150 - Zabelê, Vitória da e-mail: [email protected] Conquista, Bahia, Brazil

O. F. Lopes : A. F. Amorim R. M. de Jesus Instituto Federal de Educação, Ciência e Tecnologia da Bahia INCT de Energia e Ambiente, Universidade Federal da Bahia, (IFBA), John Kennedy, s/n - Loteamento Cidade Nova, Jequié, Salvador, Bahia 40170-280, Brazil 77 Page 2 of 14 Environ Monit Assess (2019) 191:77 previous treatment) accelerate this process. These activ- determine TSI adapted by Lamparelli (2004) are total ities cause the introduction of excessive amounts of phosphorus and chlorophyll a, considering that the these nutrients to the watercourse (Lundberg 2013; values of transparency generally do not represent the Vieira et al. 2013). According to Meng et al. (2015), trophic degree of the river, because high concentrations phosphorus runoff causes the eutrophication phenome- of turbidity present in watercourses may interfere in its non, resulting in serious environmental problems in results (Maia et al. 2015). river basins. Increased availability of elements such as The study of water quality of the rivers in the Eastern P and N causes numerous changes in aquatic environ- Water Planning and Management Region (RPGA) is ment, such as reduction of water bodies’ biodiversity, important for the population in Southern Bahia, espe- high pH, fish mortality, anoxic conditions, and produc- cially in municipalities such as Uruçuca, Una, Itabuna, tion of toxins that can be harmful to aquatic organisms Almadina, Barro Preto, Coaraci, and Itajuípe. Conser- and human health (Whitehead et al. 2011; Ouyang vation of these watercourses is essential, as the region 2012; Privette and Smink 2017). depends on this water for drinking, agricultural irriga- Land use in the watershed area plays a relevant role tion, fishing, and supplying urban as well as rural com- in the entry of sediments and nutrients into the riverbeds munities. This RPGA presents a significant level of of water bodies (Halstead et al. 2014; Shi et al. 2017). degradation due to the predominance of industrial and According to Castilla et al. (2015), aquatic environ- sanitary effluents associated with agriculture and live- ments suffer anthropogenic pressures mainly due to stock activities dumped without pretreatment into its urban expansion and increases in agricultural crops tributaries. In some stretches of rivers, aquatic macro- productions. Deforestation, agricultural activities, and phytes, silting, and vegetation growth are present in the urbanization often modify the characteristics of the riverbed. These factors demonstrate the serious situation earth’s surface, triggering changes in the volume of of water resources’ degradation. surface runoff, changes in water temperature, increases The present study aimed to evaluate the degree of in algae production, and decreases in oxygen concentra- eutrophication of the river basins located in Eastern tion in watercourses (Ding et al. 2015; Pilgrim et al. RPGA (BA), by different methodologies, relating the 2014). In this sense, monitoring the water bodies’ qual- different scenarios of land use with possible changes in ity is required in order to understand the scenario in water quality. relation to anthropogenic influences, as well as to assist in water resources planning and management. Consid- ering that water quality monitoring encompasses many Materials and methods variables, the development of quality indexes has the objective of transforming data generated through mon- Study area itoring into information that is more accessible and more easily understandable way by people involved in the The present study was developed in the Eastern RPGA management of this resource and the populations that (Fig. 1), which consists of the basins of Bahian Rivers use these watercourses. There are numerous indexes that flow into the Atlantic Ocean, bounded from the applied to water monitoring that simplify the data sets north and northwest by the Contas River RPGA and obtained in a single or a few values. Among them, it is from the south to the southwest by the Pardo River possible to use the trophic state index (TSI), which aims RPGA (INEMA 2016). to frame water bodies in different degrees of trophy, Eastern RPGA covers an area of 9507 km2 (1.68% of pointing out strategies for recovery and conservation the state), with a population of around 682,652 inhabi- of water resources (CETESB 2015). tants. The basins of the Almada, Cachoeira, and Una The potential eutrophication of water body can be River stand out. The biome is the Atlantic Forest, with evaluated through some indexes; the most classic is Seasonal Decidual and Semidecidual Forest formations. Carlson’s(1977) TSI developed for lentic environments The soil classes identified in the Eastern RPGA are located in temperate climates, which was modified by argisols, quartzeneic neosols, chernosol, and spodosol Toledo Jr. et al. (1983) for tropical environments and (INEMA 2016; CPRM 2010). then adjusted by Lamparelli (2004) for lotic water bod- According to Koppen’s classification (1936), the ies located in tropical regions. The variables used to boundaries of the Eastern RPGA comprise two climatic Environ Monit Assess (2019) 191:77 Page 3 of 14 77

Fig. 1 Land use coverage and location of study sample points typologies. In the coastal zones of the study area, which sensors (spatial resolution of 30 m) dated February 2008 are the northeastern and southern pieces, the tropical and January 2017, orbit point 284, and they are available forest climate (Afi) predominates, characterized by rain- in the digital image catalogs of the Space Research Na- fall indexes of approximately 2200 mm. These values tional Institute (INPE) and the Geological Survey Earth are the result of maritime processes, as it is a region with Resources Observation and Science Center (USGS- significant rainfall history in all months of the year. On EROS) (Clark et al. 2017). From the total number of the other hand, the central and western zones of the spectral bands that make up the image metadata, bands RPGA under study exhibit a tropical climate with a 1, 2, 3, 4, and 5 were used to identify different vegetation defined dry season that is characterized by a decrease cover and anthropogenic systems, as observed by of precipitation in the winter months between June and Lourenço and Landim (2004). In addition, the data repre- August. The cities of this zone present annual average sented in SEI (2004) and GlobCover (2009) database rainfall values of around 850 mm, with summer as the were added to geoprocessing of information in order to season with highest intense rains (Alvares et al. 2013). superimpose and delimit analytically the vegetation cover classes that compose the landscape of basins under study. Land use mapping of basins located in the Eastern ArcGIS software 10.3 was used for digital image RPGA processing. Technical geoprocessing procedures were initiated with atmospheric correction of imported images; Identification of the land use classes that make up the composition of scenes in RGB (red, green, blue); image watershed landscapes of the Eastern RPGA was carried cropping for the Cachoeira, Almada, and Una basins out with 1:50,000 scale mapping. Those classes were through the command extract by mask; and finally, the classified from satellite images of Landsat 5 and 8 TM imageresolutionrefinementfrom30to15mbyadding 77 Page 4 of 14 Environ Monit Assess (2019) 191:77 the panchromatic band of image metadata through pan- Program (PROCLIMA), while the flow data for the chromatic sharpening command (Chavez et al. 1991). period from 2008 to 2015 were obtained from National Polygons that represent (1) urban areas, (2) cacao- Water Agency (ANA). cabruca (agroforestry system that cacao tree grows under thinned forest canopy), (3) forest remnants, (4) Trophic state index restinga, (5) wetlands, (6) mangroves, and (7) agricultural areas were arbitrarily delimited. The accuracy of The TSI (PT) and TSI (CL) methodologies adapted for these data, for the scale adopted, was guaranteed tropical lotic environments performed by Lamparelli through the vegetation cover information conjuncture (2004) and currently used by CETESB (2015) were collected in cartographic data geoprocessing by the used for the trophic state index (TSI) determination of Mineral Resources Research Company (CPRM) and the Eastern RPGA, according to Eqs. 1 and 2. the SOS Mata Atlântica Institute (2015) and the mapping works observed in SEI (2004), Silva et al. (2015a), and Gomes et al. (2017). In addition, control points were TSIðÞ¼ PT 10ðÞ 6−ðÞðÞ0:42−0:36ðÞ lnPT =ln2 −20 ð1Þ defined on sensor scene images in order to distinguish some clippings with proximity of pixel values between the classes of forest/cacao-cabruca system and restinga/ TSIðÞ¼ CL 10ðÞ 6−−ðÞðÞ0:7−0:6 Â ðÞlnCL =ln2 −20 ð2Þ agriculture in supervised classification. The term restinga is generally used in Brazil to refer to the land- where, TSI (PT) is the trophic state index in relation scape with presence of coastal sandy soils and charac- to the total phosphorus variable for lotic environments; μ −1 teristic vegetation (Da Silva et al. 2010), with herba- PT is the total phosphorus concentration ( gL ); and ceous species and shrubs (Alves et al. 2007). CL is the chlorophyll a concentration measured at the water surface (μgL−1). Water quality data and climatic variables For the calculation of the Lagoa Encantada collection point, we used Eqs. 3 and 4 (Lamparelli 2004)appliedto The total phosphorus (PT), total nitrogen (NT), and lentic aquatic ecosystems of tropical areas. Mean TSI chlorophyll a data were collected from INEMA Moni- (PT + CL) was obtained through the arithmetic mean of toring Program online database, between 2008 and total phosphorus and chlorophyll a variables. 2015, with periodicity of quarterly samples collected, distributed in thirteen (13) monitoring stations through- TSIðÞ¼ CL 10ðÞð 6−ðÞðÞ0:92−0:34ðÞ ln CL =ln 2 3Þ out the Eastern RPGA, totaling 260 samples of each variable analyzed. The monitoring stations were the following: A1, upstream of the urban area on Almada TSIðÞ¼ PT 10ðÞð 6−ðÞ1:77−0:42ðÞ ln PT =ln 2 4Þ River; A2, water catchment point for the public supply on Almada River; C1, urban area on Colônia River; C2, where, TSI (PT) is the trophic state index in relation downstream of the urban area on Colônia River; L, pond to the total phosphorus variable for lentic environments; in Almada basin; CH, urban area on Cachoeira River; PT is the total phosphorus concentration measured at the CH2, urban area in Cachoeira River; CH3, downstream water surface, in μgL−1; and CL is the chlorophyll a of the urban area; S1, Salgado River spring; S2, up- concentration measured at the water surface, in μgL−1. stream of the urban area in Salgado River; S3, down- After obtaining the values of these indexes, the trophic stream of the urban area on Salgado River; U1, down- state framing for each sampling point evaluated by the stream of the urban area on Una River; and U2, coun- Eastern RPGA according to Lamparelli (2004) classifica- tryside zone of Una River (Fig. 1 and Table 1). Table 1 tion was performed, ranging from ultra-oligotrophic to shows monitoring points, geographic coordinates, ba- hypereutrophic. The mean monthly concentrations of PT sin, and the land use of Eastern RPGA. (μgL−1) were used to establish the potential to cause The analysis protocol used by INEMA followed the damage due to algae productivity, where PT ≤ 30 (low recommendations of APHA (2005). Precipitation data potential), 30 < PT ≤ 170 (moderate potential), 170 ≤ PT ≤ related to the period from 2008 to 2015 were collected 440 (significant potential), 440 ≤ PT ≤ 1800 (high poten- by Northeast Region Real-time Climatic Monitoring tial), and > 1800 (very high potential) (Lamparelli 2004). Environ Monit Assess (2019) 191:77 Page 5 of 14 77

Table 1 Eastern RPGA monitoring points, geographic coordinates, and land use

Collected points Coordinates River (basin) Localities (cities) Land use

A1 14.64847 S Almada Ilhéus Farming/pasture, cacao-cabruca 39.34476 W A2 14.65513 S Almada Ilhéus Cacao-cabruca 39.18975 W C1 15.12123 S Colônia Itororó Urban 40.06833 W C2 15.14418 S Colônia Itaju do Colônia Urban 39.72340 W L 14.61981 S Lagoa Encantada Ilhéus Urban, forest, restinga and wetlands 39.14137 W CH1 14.89800 S Cachoeira Itapé Urban, farming 39.42763 W CH2 14.78838 S Cachoeira Itabuna Urban 39.27003 W CH3 14.80244 S Cachoeira Ilhéus Urban, cacao-cabruca, farming/pasture 39.15273 W S1 14.95547 S Salgado Firmino Alves Farming/pasture 39.94909 W S2 14.90481 S Salgado Floresta Azul Urban, farming 39.71866 W S3 14.87595 S Salgado Ibicaraí Farming/pasture 39.48694 W U1 15.07320 S Una Buerarema Forest, cacao-cabruca 39.32618 W U2 15.29738 S Una Una Forest, cacao-cabruca, and urban 39.06369 W

The N/P ratio calculus was realized in order to identify present normal distribution, non-parametric tests such as which nutrient served as a limiting element for the different Mann Whitney and Kruskal-Wallis were used to perform periods analyzed and three basins studied. The environ- the statistical analyses. The Mann Whitney non-parametric ment was considered to be limited by phosphorus when hypothesis test was used to verify the differences between the N/P ratio found was higher than 10, and nitrogen was the dry and rainy season average with significance level at considered a limiting factor when the values obtained for 5%. In order to analyze whether or not the results differed the ratio of these two nutrients were lower than 10 between the various monitoring points, the Kruskal-Wallis (Lamparelli 2004; Shen and Liu 2009). test was applied with a significance level at 5%. In order to evaluate the spatiotemporal variation of Analyses and treatment of data the trophic state indexes, bar graphs were generated with dry and rainy season results per monitoring point and Descriptive statistics were performed after separation of the standard error of mean. the data in dry and rainy season per basin, where values of Different trophic state indexes were correlated with mean and standard deviation were obtained. The criterion land use classes as well as the climatic variables, such as used to separate the dry and rainy months in this study was monthly accumulated precipitation, accumulated pre- 90 mm precipitation, so months with values above this cipitation prior to 7 days, mean flow, and minimum reference were considered rainy, while months with values and maximum flow in each basin located in the Eastern below were framed as dry (Brandão et al. 2015). RPGA. Spearman’s correlation was performed in The data set was submitted to the Kolmogorov-Smirnov Statistica software version 21, with significance at 1% normality test (p < 0.05). Considering that the data did not and 5%. 77 Page 6 of 14 Environ Monit Assess (2019) 191:77

The principal components analysis (PCA) was also Table 2 Land use classes in hectares (ha) and percentage (%) of used to analyze the influence of land use in variables of the three basins (Cachoeira, Almada, and Una) located in the Eastern RPGA (BA) for the years 2008 and 2016 water quality and different trophic state indexes, as well as to evaluate the possible tendency of sample separa- Class 2008 2016 tion as a function of dry and rainy season. Statistica Area software version 7.0 was used to generate graphs of weights and scores. ha % ha %

Cachoeira River basin Farming/pasture 304,579 70.1 338,826 77.69 Results and discussion Cacao-cabruca 123,234 28.05 91,484 21.1 Remaining forest 2950 0.7 149 0.01 Land use classes Urban area 2401 0.55 2713 0.6 Mangrove 240 0.05 232 0.05 The occupied areas by each land use classes of the three Wetlands 37 0.55 36.84 0.55 basins (Almada, Una, and Cachoeira), which were ob- Total 433,441 100 433,441 100 tained through the classification of Landsat 5 and 8 TM Almada River basin images for the period of 2008 and 2016, are presented in Table 2. It is noted that in Almada and Una River basins, Farming/pasture 30,523 18.9 38,650 22.9 there is a predominance of cacao-cabruca (agroforestry Cacao-cabruca 134,296 79.05 127,967 75.88 system of the cacao crop), with 75.88 and 68.90%, Remaining forest 2679 1.45 1106 0.65 respectively, while in Cachoeira River BH, there is the Urban Area 149 0.1 258 0.15 dominion of farming, with 77.69% of land use. Among Wetlands 528 0.3 528 0.3 the three basins analyzed, Una River basin stood out in Restinga 465 0.2 131 0.12 relation to the presence of forest remnants, with 20% of Total 168,640 100 168,640 100 the total area. Una River basin There was little change in land use over 8 years of Farming/pasture 18,124 10 18,474 10.2 analysis. It was noticed that the remaining forest in the Cacao-cabruca 124,334 68.8 124,520 68.9 three basins was reduced, being replaced by farming in Remaining forest 36,656 20.3 36,107 20 Cachoeira and Almada River basins. In Una River basin, Urban area 278 0.1 304 0.2 there was an increase in farming, cacao-cabruca, and the Mangrove 320 0.2 331 0.2 urban area, with the native vegetation removed to imple- Wetlands 522 0.3 522 0.3 ment these human activities. The three basins present a Restinga 300 0.3 276 0.2 high degree of degradation due to the low quantity of forest Total 180,534 100 180,534 100 remnants. These aspects can have serious consequences for water quality. A study developed by Meneses et al. (2015) in the Zêzere basin, Portugal, found better water Cachoeira Basin presented the highest values for PT quality in reservoirs with the highest percentage of forest. in both periods; on the other hand, NT, TSI (PT), and mean TSI showed higher values in dry period. Seasonal and spatial assessment of nutrients and trophic The highest values of chlorophyll a were found in state indexes dry period, mainly in Almada and Cachoeira River basin, with a value of 16.6 μgL−1 and 12.73 μgL−1, Mean values of nutrients (NT and PT), monthly accu- respectively, and consequently had the results of TSI mulated precipitation, chlorophyll a, and the different chlorophyll a higher. During dry period, the low trophic state indexes in the three basins in dry and rainy dilution contributes with higher values of nutrients seasons are presented in Table 3. and consequently, higher chlorophyll a concentra- Variables (PT, NT, and CL) did not show significant tions are found (Silva et al. 2015b). differences between the dry and rainy seasons for the Mean TSI (PT) values of monitoring points located in data set (p >0.05). the three analyzed basins, between 2008 to 2015, ranged Environ Monit Assess (2019) 191:77 Page 7 of 14 77

Table 3 Nutrients and results of trophic state indexes in the three basins of Eastern RPGA for the period 2008 to 2015 (mean ± standard deviation)

Basin PT NT P TSI (PT) CL TSI (CL) TSIm

BALMc 0.1 ± 0.3 1.5 ± 0.7 156.5 ± 40.7 60.0 ± 6.0 6.4 ± 14.6 52.8 ± 8.3 56.4 ± 5.6 BALMs 0.1 ± 0.1 2.7 ± 5.1 59.3 ± 20.6 60,7 ± 4.1 16.6 ± 29.6 59.0 ± 8.9 59.9 ± 5.7 BCHc 0.3 ± 0.4 2.5 ± 2.0 162.8 ± 49.2 59.4 ± 6.1 5.0 ± 7.8 54.9 ± 10.1 57.1 ± 6.5 BCHs 0.5 ± 1.2 2.9 ± 3.5 55.6 ± 18.9 62.2 ± 6.0 12.7 ± 35 58.5 ± 11.8 60.3 ± 7.3 BUNAc 0.0 1.3 ± 0.6 173.5 ± 65 53.6 ± 2.8 2.7 ± 4.6 49.8 ± 9.5 51.7 ± 5.4 BUNAs 0.1 ± 0.0 1.5 ± 0.6 50.6 ± 26.2 54.2 ± 3.0 3.9 ± 6.6 52.5 ± 9.2 53.4 ± 5.3

BALMc, Almada Basin in rainy season; BALMs, Almada Basin in dry season; BCHc, Cachoeira Basin in rainy season; BCHs,Cachoeira Basin in dry season; BUNAc, Una Basin in rainy season; BUNAs, Una Basin in dry season; PT, phosphorus total in mg L−1 ; NT, nitrogen total in mg L−1 ; P, monthly accumulated precipitation in mm; TSI, PT, trophic state indexes with total phosphorus indicator; CL, chlorophyll ainμgL−1 ; TSI CL, trophic state indexes with chlorophyll a indicator; TSI PT CL, mean trophic state indexes from 52.9 (U2) to 70.4 (CH2) with trophic level between trophic level, mainly the monitoring point CH2 located mesotrophic and hypereutrophic (Fig. 2), while the trophy in urban area (p < 0.05). related to the chlorophyll a variable had a classification The high values found in the CH2 point can be between oligotrophic (U1) and hypereutrophic (CH2), with correlated with the fact that Cachoeira River runs values from 49.5 to 74.2 (Fig. 3). The mean TSI varied through the urban center of Itabuna city and receives a from 51.5 (U2) to 72.3 (CH2), with a degree of eutrophi- significant portion of wastewater. This city presents the cation between oligotrophic and hypereutrophic (Fig. 4). largest population density of the Eastern RPGA, accord- The results of the three TSI showed significant ing to the last IBGE census (2010). As reported by differences between monitoring points, in which Una Santos and Souza (2014), approximately 20% of the River basin presented the lowest values for the differ- sanitary sewage in this town is not treated. ent trophic state indexes, while Cachoeira River basin Urbanization and the consequent release of domestic stood out in relation to the others, with accentuated effluents were also the main sources identified in

Fig. 2 Spatiotemporal variation of trophic state index using the total phosphorus indicator in different periods of the year (rainy and dry) at different sampling points along the three basins located in the Eastern RPGA (mean ± standard deviation) 77 Page 8 of 14 Environ Monit Assess (2019) 191:77

Fig. 3 Spatiotemporal variation of trophic state index using the chlorophyll a indicator in different periods of the year (rainy and dry) at different sampling points along the three basins located in the Eastern RPGA (mean ± standard deviation)

Cachoeira River by Lucio (2012) as responsible for rivers can be correlated with greater amount of forest changes in nutrient dynamics, such as phosphate and remnant in this basin (Table 2). The forest assisted in the ammonium. The increase concentrations of these nutri- water resources conservation according to Meneses ents in places with greater anthropogenic pressures, et al. (2015) and Kamjunke et al. (2013). especially discharge of effluents, were also observed in Considering the seasonality, it was possible to verify rivers with different ecosystems such as Hun River in that, over the 9-year period of analysis, the trophic state northeast China (Wang et al. 2013;Boseetal.2012; indexes had the highest values in the dry season, Halstead et al. 2014). resulting in highest trophic classifications (p < 0.05) The lower values of trophic state indexes obtained for (Figs. 2, 3,and4). This effect is expected because of Una River sections in relation to the other monitored the decrease of the dilution capacity of pollutants by

Fig. 4 Spatiotemporal variation of mean trophic state index in different periods of the year (rainy and dry) at the different sampling points along the three basins located in the Eastern RPGA (mean ± standard deviation) Environ Monit Assess (2019) 191:77 Page 9 of 14 77 reduction of precipitation, increasing the trophy of wa- The high probability of algae occurrence in the tercourses (Lamparelli 2004). Cachoeira basin deserves attention, mainly due to the This result was confirmed by Spearman’scorrelation possibility of cyanobacteria appearance, which pro- finding an inverse relationship between indexes and duces toxins harmful to humans and animals. This as- climatic variables as precipitation and flow (Table 5). pect may cause changes in water quality, presenting a The results obtained for TSI (PT) and TSI (CL) did not high risk for water bodies, especially those used for present significant differences (p < 0.05). For Lamparelli public water supply (Cheung et al. 2013). (2004), in the water body which values for the trophic Phytoplankton growth is influenced by the Liebig state indexes (PT) are similar to the TSI (CL) means that minimum law, in which the element that is in lower the phenomenon of eutrophication is totally determined. relative concentration will act as the limiting factor Potential to cause harm due to algal productivity was that will control this development (Lamparelli determined for the three basins located in the aforemen- 2004). It is relevant to identify the most limiting tioned planning region in different collection seasons nutrient for determination of eutrophication control (2008 to 2015) (Fig. 5). RPGA presented a potential measures. Thus, the relationship between N and P between low to very high for phytoplankton productiv- was calculated for the monitoring points of the East- ity based on eutrophication indicator (PT) in lotic envi- ern RPGA (Fig. 6). ronments, according to the classification system pro- Basins located in the Eastern RPGA had the phos- posed by Lamparelli (2004) for rivers in São Paulo state. phorus as a limiting nutrient for all monitoring points, The water bodies that presented highest potential for considering the mean values of N/P ratios. Efforts to algal bloom were CH2, CH3, and C1, with high per- implement water conservation measures should be pri- centage in the serious to very serious class. Una basin oritized to reduce the entry of phosphorus into these and Lagoa Encantada—lentic water body of Almada water bodies. However, nutrients (N and P) can act as basin—presented different behavior, with a milder clas- limiting agents simultaneously, modifying their concen- sification, ranging from low to moderate at point U1 and trations in aquatic environment (Lee and Jones-Lee L and low to significant at station U2. 1998;Lamparelli2004). Therefore, it is necessary to

Fig. 5 Potential to cause harm due to algal productivity in the three basins (Almada, Cachoeira, and Una), according to Lamparelli (2004) 77 Page 10 of 14 Environ Monit Assess (2019) 191:77

Fig. 6 Mean values of the limiting nutrient ratio (N/P) obtained for water bodies located in the Eastern RPGA (BA), considering the period from 2008 to 2015 adopt measures that aim to minimize the contributions statistically significant inverse correlation at 5% with of both to rivers of the studied RPGA. monthly accumulated precipitation (r = − 0.270) and positive with wetlands (r =0.302). Influence of land use and variables (flow In Cachoeira basin, there was a negative correlation and precipitation) on trophic state indexes between the trophic state index (PT) and the mean flows (r = − 0.265) and minimum (r = − 0.316) and maximum The trophic state index (PT) correlated positively at 1% (r = − 0.246), as well as accumulated monthly precipi- with the farming (r = 0.419) and wetlands (r =0.609)in tation (r = − 0.235) and accumulated precipitation prior Almada basin (Table 4).TSI(CL)presenteda to 7 days (r = − 0.300) (Table 5). The correlation was

Table 4 Spearman’s correlation of Almada River basin for the period from 2008 to 2015

TSIPT P1 TSICL URB AGRO CAB P2 Qmed Qmax Qmin Wetlands

TSIPT 1.00 P1 − 0.08 1.00 TSICL 0.26* − 0.27* 1.00 URB 0.22 0 0.05 1.00 AGR 0.42** − 0.01 0.24 − 0.50** 1.00 − 1.00** CAB − 0.42** 0.01 − 0.24 0.50** − 1.00** 1.00 P2 0.12 0.38** − 0.15 0 − 0.03 0.03 1.00 Qmed 0.02 0.67** − 0.16 − 0.21 0.21 − 0.21 0.14 1.00 Qmax 0.01 0.71** − 0.17 − 0.20 0.20 − 0.20 0.12 0.96** 1.00 Qmin − 0.04 0.44** − 0.09 0.17 − 0.17 0.17 0.12 0.65** 0.55** 1.00 Wetlands 0.61** − 0.01 0.30* 0 0.87** − 0.87** − 0.03 0.12 0.11 − 0.10 1.00

*The correlation is significant at the 0.05 level (2 extremities); **The correlation is significant at the 0.01 level (2 extremities) AGR, agriculture; CAB,cabruca-system;URB,urbanarea;TSI CL, TSI using the chlorophyll a variable; TSI PT, TSI using the total phosphorus variable; P1, monthly accumulated precipitation; P2, accumulated precipitation prior to the 7 days of collection; Qmed, mean flow; Qmax, maximum flow; Qmin, minimum flow; TSIm,meanTSI Environ Monit Assess (2019) 191:77 Page 11 of 14 77

Table 5 Spearman’s correlation of Cachoeira River basin for the period from 2008 to 2015

TSIPT P1 TSICL IETm P2 CAB URB FLO Qmed Qmax Qmin

TSIPT 1.00 − 0.24** 0.19** 0.61** − 0.30** 0.30** 0.23** − 0.10 − 0.27** − 0.25** − 0.32** P1 − 0.24** 1.00 − 0.12 − 0.21** 0.59** 0.09 0.03 − 0.08 0.69** 0.72** 0.46** TSICL 0.19** − 0.12 1.00 0.87** − 0.17* 0.06 0.13 − 0.17* − 0.20** − 0.18* − 0.22** TSIm 0.61** − 0.21** 0.87** 1.00 − 0.27** 0.13* 0.18* − 0.19* − 0.27** − 0.24** − 0.31** P2 − 0.30** 0.59** − 0.17* − 0.27** 1.00 0.11 0.03 − 0.06 0.32** 0.29** 0.35** CAB 0.30** 0.09 0.06 0.13* 0.11 1.00 0.50** − 0.30** 0 0 0 URB 0.23** 0.03 0.13 0.18* 0.03 0.50** 1.00 − 0.60** 0 0 0 FLO − 0.10 − 0.08 − 0.17* − 0.19* − 0.06 − 0.30** − 0.60** 1.00 0 0 0 Qmed − 0.27** 0.69** − 0.20** − 0.27** 0.32** 0 0 0 1.00 0.95** 0.76** Qmax − 0.25** 0.72** − 0.18* − 0.24** 0.29** 0 0 0 0.95** 1.00 0.68** Qmin − 0.32** 0.46** − 0.22** − 0.31** 0.35** 0 0 0 0.76** 0.68** 1.00

CAB, cacao-cabruca; FLO, forest; URB,urbanarea;TSI CL, TSI using the chlorophyll a variable; TSI PT, TSI using the total phosphorus variable; P1, monthly accumulated precipitation; P2, accumulated precipitation prior to the 7 days of collection; Qmed, mean flow; Qmax, maximum flow; Qmin, minimum flow; TSIm, mean TSI *The correlation is significant at the 5% level (2 extremities); **The correlation is significant at level 1% (2 extremities) positive for TSI, cacao-cabruca (r =0.301),andurban pasture, as well as negative impacts of urbanized areas area (r = 0.23) in Cachoeira River basin. In relation to compromising the water sources quality. It was possible to TSI (CL), the correlation was negative for the climatic observe that there was a positive relation between crops variables, such as precipitation, both monthly and prior and increase of trophic state indexes, as well as the inverse to the 7 days of collection (r = − 0.169), as well as mean correlation with forest remnants. Some studies have found flow (r = − 0.199) and maximum (r = − 0.184) and min- that urban and agricultural areas had negative impacts on imum (r = − 0.217). Another significant inverse correla- water quality parameters, while forested areas had positive tion that occurred was between remnant forest (r = − effects (Shi et al. 2017;Tu2011; Wan et al. 2014). 0.174) and trophic status index (CL). Figure 7 shows the score graphs (a) and loadings (b) The TSI (PT) in Una Basin correlated negatively with where the first principal component (PC1) and the second forest remnants (r = − 0.312) and positively with agri- principal component (PC2) explain 65.6% of the total culture (r = 0.312) and cacao-cabruca (r = 0.312), being data set variance. The highest values for the various significant at 5% (Table 6). indexes, nutrients, and chlorophyll a occurred in The three basins presented a marked level of alteration Cachoeira River basin (RCHc, RCHs, and COLs), being in its landscapes due to the agricultural practices and influenced mainly by farming and urban area. The ACP

Table 6 Matrix correlation of Una River basin for the period from 2008 to 2015

AGR CAB FLO TSICL TSIPT P2 P1

AGR 1.00 1.00** − 1.00** − 0.03 0.31* − 0.08 − 0.04 CAB 1.00** 1.00 − 1.00** − 0.03 0.32* − 0.09 − 0.05 FLO − 1.00** − 1.00** 1.00 0.03 − 0.33* 0.08 0.04 TSICL − 0.03 − 0.03 0.03 1.00 0.30* − 0.06 − 0.15 TSIPT 0.31* 0.32* − 0.33* 0.30* 1.00 0.26 − 0.01 P2 − 0.08 − 0.09 0.08 − 0.06 0.26 1.00 0.60** P1 − 0.04 − 0.05 0.04 − 0.15 − 0.01 0.60** 1.00

AGR,farming;CAB, cacao-cabruca; FLO,forest;TSI CL, TSI using the chlorophyll a variable; TSI PT, TSI using the total phosphorus variable; P1, monthly accumulated precipitation; P2, accumulated precipitation prior to the 7 days of collection *The correlation is significant at the 5% level (2 extremities); **The correlation is significant at level 1% (2 extremities) 77 Page 12 of 14 Environ Monit Assess (2019) 191:77

Fig. 7 Principal component analysis (PCA). Score graphs (a)and (Salgado River - dry season); COLc (Colônia River - rainy sea- loadings (b) of first principal component (PC1) versus second son); COLs (Colônia River - dry season); RCHc (Cachoeira River principal component (PC2) of the main components analysis for - rainy season); RCHs (Cachoeira River - dry season); UNAc (Una the variables: CL-a, NT, PT, P, TSI (PT), TSI (CL), mean TSI, and River - rainy season); UNAs (Una River - dry season); AGR land use classes (AGR, URB, FLO, and CAB) in Eastern RPGA (farming); URB (urban area); FLO (forest); CAB (cacao-cabruca); basins. ALMc (Almada River - rainy season); ALMs (Almada CL-a (chlorophyll a); NT (nitrogen); PT (total phosphorus); P River - dry season); LENc (Lagoa Encantada in Almada River (monthly accumulated precipitation); TSI PT (TSI using the total basin - rainy season); LENs (Lagoa Encantada in Almada River phosphorus variable); TSI CL (TSI using the chlorophyll a vari- basin - dry season); SLDc (Salgado River - rainy season); SLDs able); TSIm (mean TSI) confirmed the results obtained by Spearman’s correlation, trophic indexes. The larger forest remnants in the area in which urbanization and agricultural crops, as well as was one of the factors that contributed to the more ap- pasture, had a negative influence on water quality. propriate values. The algal potential indicator suggests that the Una basin presented the least susceptibility to eutrophication among the basins of the Eastern RPGA. Conclusion Precipitation influenced the trophic state indexes, with highest values in dry season due to the absence of The trophic state indexes showed an important cor- dilution process. relation with land use in the three basins. It was verified by Spearman’s correlation, in which the urban area, the agricultural, and livestock farming had a Funding information This study was financially supported by positive correlation for the trophic state indexes and the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and was financed in part by the Coordenação negative relation with remnant forest. The results de Aperfeiçoamento de Pessoal de Nível Superior - Brasil indicated that sewage discharges and various sources (CAPES) Finance Code 001. of agricultural crops are the main sources of pollution in basins analyzed. The remnant forest plays an im- Publisher’sNote Springer Nature remains neutral with regard to portant role in water resources conservation of this jurisdictional claims in published maps and institutional affiliations. water-planning region. The PCA indicated that farming and the urban area contributed with high concentrations of nutrients and chlorophyll a, and high values for different TSIs in Cachoeira basin. References The Cachoeira River basin presented the worst values for TSI, as well as the greater potential of algae occur- Alvares, C. A., Stape, J. L., Sentelhas, P. C., Gonçalves, J. L. M., rence, mainly due to the dumping of wastewater without & Sparovek, G. (2013). Köppen’s climate classification map adequate treatment. The Una basin presented the smallest for Brazil. Meteorologische Zeitschrift, 22(6), 711–728. Environ Monit Assess (2019) 191:77 Page 13 of 14 77

Alves, R. J. V., Cardin, L., & Kropf, M. S. (2007). Angiosperm IBGE. Instituto Brasileiro de Geografia e Estatística. (2010). Available disjunction BCampos rupestres-restingas^: a re-evaluation. in: http://www.sidra.ibge.gov.br/bda/territorio/uftabunit.asp?t=5 Acta Botanica Brasilica, 21,675–685. &n=6&z=t&o=4. Accessed 28 May 2016. APHA. (2005). Standard methods of examination of water and INEMA. Instituto do Meio Ambiente e Recursos Hídricos. (2016). wastewater (20th ed.). Washington: American Public Health Available in: http://www.inema.ba.gov.br/gestao-2/comites- Association. de-bacias/comites/cbh-leste/. Accessed 12 November 2016. Bose, R., DE, A., Subho, M., Sen, G., & Deb, A. M. (2012). Kamjunke,N.,Buttner,O.,Jager,C.G.,Marcus,H.,Tümpling,W. Coastal water pollution in two rivers of the Bengal Delta. V.,Halbedel,S.,Norf,H.,Brauns,M.,Baborowsk,M.,Wild,R., Geochemistry International, 50(10), 860–868. Borchardt, D., & Weitere, M. (2013). Biogeochemical patterns Brandão, C. S., Silva, L. P., Chaussê, T., & Silva, D. M. L. (2015). in a river network along a land use gradient. Environmental Spatio-temporal variability of dissolved loads in rivers in Monitoring and Assessment, 185(11), 9221–9236. environmental protected areas in Northeast Brazil. Revista Lamparelli, M.C. Trophic status in São Paulo State water bodies – Brasileira de Recursos Hídricos, 20(3), 551–559. evaluation of monitoring methodologies. São Paulo (BR): – Carlson, R. E. (1977). A trophic state index for lakes. Limnology 2004. PhD Thesis Instituto de Biociências da and Oceanography, 22(2), 261–269. Universidade de São Paulo. Castilla, E. P., Cunha, D. G. F., Lee, F. W. F., Loiselle, S., Ho, K. Lee, G.F., & Jones-Lee, A. (1998). Determination of nutrient C., & Hall, C. (2015). Quantification of phytoplankton bloom limiting maximum algal biomass in waterbodies. dynamics by citizen scientists in urban and peri-urban envi- https://www.gfredlee.com/Nutrients/nut_limit.pdf. Accessed ronments. Environmental Monitoring Assessment,187–690. 14 April 2014. CETESB. Companhia Ambiental do Estado de São Paulo. (2015). Lourenço, R. W., Landim, P.M.B. (2004). Study on the variability of normalize difference vegetation index/NDVI by indicative Qualidadedeágua.Availablein:http://aguasinteriores. – cetesb.sp.gov.br/wp-content/uploads/sites/32/2013/11/04. kriging. Holos Environment, 4 (1), 38 55. pdf. Accessed August 2015. Lucio, M. Z. T. P. Q. L. (2012). Hydrochemistry of Cachoeira River (Bahia state, Brazil). Acta Limnologica Brasiliensia, Chavez, P., Stuart, C. S., & Jeffrey, A. A. (1991). Comparison of three 24(2), 181–192. different methods to merge multiresolution and multispectral Lundberg, C. (2013). Eutrophication, risk management and sus- data- Landsat TM and SPOT panchromatic. Photogrammetric tainability: the perceptions of different stakeholders in the Engineering and Remote Sensing, 57(3), 265–303. northern Baltic Sea. Marine Pollution Bulletin, 66(1–2), Cheung, M. Y., Liang, S., & Lee, J. (2013). Toxin-producing 143–150. cyanobacteria in freshwater: a review of the problems, impact Maia, A. A. D., Carvalho, S. L., & Carvalho, F. T. (2015). on drinking water safety, and efforts for protecting public Comparison of two indexes of determination of the trophic health. Journal of Microbiology, 51(1), 1–10. state in the waters of Baixo São José dos Dourados, São Clark, R. N., Swayze, G.A., Wise, R. A., Livo, K. E., Hoefen, T.M., Paulo, Brazil. Engenharia Sanitária e Ambiental, 20(4), Kokaly, R. F., Sutley, S. J. (2017). USGS digital spectral library 613–622. splib06a. U.S. Geological. Survey, Digital Data Series, 231. Meneses, B. M., Reis, R., Vale, M. J., & Saraiva, R. (2015). Land http://speclab.cr.usgs.gov/spectral.lib06. Accessed 22 January use and land cover changes in Zêzere watershed (Portugal) - 2018. water quality implications. Science of the Total Environment, CPRM. Serviço Geológico do Brasil. (2010). Base Cartográfica do 527,439–447. Estado da Bahia - Escala 1:100.000. Blocos C e D. Geologia, Meng, C., Li, Y., Wang, Y., Yang, W., Jiao, J., Wang, M., Zhang, M., Tectônica e Recursos Minerais do Brasil: Sistema de Li, Y., & Wu, J. (2015). TMDL for phosphorus and contribut- Informações Geográficas - SIG e Mapas. Belo Horizonte. ing factors in subtropical watersheds of southern China. Da Silva, R. M., Mehli, U., Dos Santos, J. U. M., & De Menezes, Environmental Monitoring and Assessment, 187(8), 514. M. P. M. (2010). The coastal restinga vegetation of Pará, Ouyang, Y. (2012). Estimation of shallow groundwater discharge Brazilian Amazon: a synthesis. Brazilian Journal of Botany, and nutrient load into a river. (2012). Ecological – 33(4), 563 573, 2010. Engineering, 38(1), 101–104. Ding, J., Jiang, Y.,Fu, L., Peng, Q., & Kang, M. (2015). Impacts of Pilgrim, C. M., Mikhailova, E. A., Post, C. J., & Hains, J. J. land use on surface water quality in a subtropical river basin: (2014). Spatial and temporal analysis of land cover changes a case study of the Dongjiang River basin, southeastern and water quality in the Lake Issaqueena watershed, South China. Water, 7,4427–4445. Carolina. Environmental Monitoring and Assessment, GlobCover. (2009). GlobCover Land Cover Maps. https://due. 186(11), 7617–7630. esrin.esa.int/page_globcover.php. Accessed 15 January Privette, C., & Smink, J. (2017). Assessing the potential impacts of 2015. WWTP effluent reductions within the Reedy River water- Gomes, R. L., Marques, E. A. G., & Franco, G. B. (2017). The shed. Ecological Engineering, 98,11–16. waste disposal suitability of Almada River watershed. Santos, S.O. & Souza, A.C. (2014). Panorama do saneamento Engenharia Sanitária e Ambiental, 22(4), 731–747. básico no município de Itabuna (Bahia) de 2000 a 2010. Halstead,J.A.,Kliman,S.,Berheide,C.W.,Chaucer,A.,& Available in: http://www.uesb.br/eventos/semana_ Cock-Esteb, A. (2014). Urban stream syndrome in a small, economia/2014/anais-2014/b01.pdf. Accessed 15 October lightly developed watershed: a statistical analysis of water 2016. chemistry parameters, land use patterns, and natural SEI. Superintendência de Estudos Econômicos e Sociais da Bahia sources. Environmental Monitoring and Assessment, (2004). Mapas digitalizados do Estado da Bahia: base de 186(6), 3391–3414. dados. Salvador: SEI. (CD-ROM). 77 Page 14 of 14 Environ Monit Assess (2019) 191:77

Shen, Z.l., & Liu, Q. (2009). Nutrients in the Changjiang River. geographically weighted regression. Applied Geography, Environment Monitoring and Assessment, 153(1–4), 27–44. 31(1), 376–392. Shi, P., Zhang, Y.,Li, Z., Li, P., & Xu, G. (2017). Influence of land Vieira, J. M. P., Pinho, J. L. S., Dias, N., Schwanenberg, D., & Van use and land cover patterns on seasonal water quality at Den Boogaard, H. F. P. (2013). Parameter estimation for multi-spatial scales. Catena, 151,182–190. eutrophication models in reservoirs. Water Science and Silva, K. B., Gomes, R. L., & Rego, N. A. C. (2015a). Social Technology, 68(2), 319–327. and environmental hydrographics implications of the land Wan, R., Cai, S., Li, H., Yang, G., Li, Z., & Nie, X. (2014). use in the plain and coastal boards between Ilhéus and Inferring land use and land cover impact on stream water Olivença – BA. (2015). Journal of Hyperspectral Remote quality using a Bayesian hierarchical modeling approach in Sensing, 5(1), 13–26. the Xitiaoxi River watershed, China. Journal of Silva, M. A. M., Souza, M. F. L., & Abreu, P. C. (2015b). Spatial Environmental Management, 133,1–11. and temporal variation of dissolved inorganic nutrients, and Wang, R., Xu, T., Yu, L., Zhu, J., & Li, X. (2013). Effects of land α chlorophyll- in a tropical estuary in northeastern Brazil: use types on surface water quality across an anthropogenic dynamics of nutrient removal. Brazilian Journal of disturbance gradient in the upper reach of the Hun River, – Oceanography, 63(1), 1 15. Northeast China. Environmental Monitoring and Toledo, Jr., A.P., Talarico, M., Chinez, S. J., & Agudo, E. G. A. Assessment, 185(5), 4141–4151. (1983). A aplicação de modelos simplificadospara a Whitehead, P. G., Jin, L., Baulch, H. M., Butterfield, D. A., Oni, S. avaliação de processo da eutrofização em lagos e K., Dillon, P. J., Futter, A. J., Wade, A. J., North, R., reservatórios tropicais. https://www.scielo.br/scielo. O’Connor, E. M., & Jarvie, H. P. (2011). Modelling phos- php?script=sci_nlinks&ref=000144&pid=S2179-975 phorus dynamics in multi-branch river systems: a study of the X201200040000500040&lng=pt. Accessed 10 Feb 2016. Black River, Lake Simcoe, Ontario, Canada. Science of the Tu, J. (2011). Spatially varying relationships between land use and Total Environment, 412-413,315–323. water quality across an urbanization gradient explored by