Aim We record new occurrences of the invasive species Ceratium furcoides in reservoirs and their affluents in the Paraná River basin, State of Goiás (GO), central-western Brazil, and in some localities in the São Francisco River basin, northeastern region. Methods Qualitative and quantitative phytoplankton samples were collected from Corumbá Reservoir and Cascatinha Falls, Caldas Novas, GO, and João Leite Reservoir, Goiânia, GO, both in the Paraná River basin, and samples from the São Franscisco River basin. Specimens of C. furcoides were observed with optical, epifluorescence and scanning electron microscopy. Results The individuals of C. furcoides from these environments agreed morphologically with populations in other reservoirs in Brazil and other locations, especially concerning the tabulation and the shape of the fourth apical plate. These environments ranged from oligotrophic to eutrophic conditions. Physical and chemical variables of these waterbodies, compared to other environments where this species was found, demonstrate that C. furcoides is a highly eurytopic species. The dispersal pattern of C. furcoides seems to be more complex than upstream-downstream regulation, since the species occurs in high-altitude environments and systems upstream from previously recorded locations. Conclusion An analysis based on the areas of occurrence and the chronology of the records demonstrated that C. furcoides has spread toward northern Brazil. Studies of the relationships among populations recorded in other parts of Brazil and South America are required in order to develop accurate models of dispersal for this invasive species, and will facilitate the development of management policies for aquatic systems in Brazil.

Este trabalho teve o objetivo de implementar, calibrar e validar os modelos conceituais de chuva-vazão Tank Model e HyMOD no intuito de simular a vazão na bacia hidrográfica do Rio das Almas no estado de Goiás. O conjunto de dados utilizados abrangeu uma série de 32 anos de dados hidrológicos diários (01 de agosto de 1982 a 31 de julho de 2017), os anos de 1995, 1996 e 1997 foram excluídos por não apresentarem dados completos de vazão. Os dados de precipitação foram obtidos através dos registros das três estações pluviométricas localizadas na área da bacia e o volume de precipitação média diária foi calculada por meio do Método de Thiessen. A otimização dos parâmetros dos modelos Tank Model e HyMOD foi realizada por meio de dois métodos automáticos de calibração utilizando algoritmos da ferramenta Solver do Excel: Programação não-linear do Gradiente Reduzido Generalizado (GRG) e o método Evolutionary. Tais algoritmos foram aplicados considerando dados de vazão e precipitação em escala diário e mensal. Essas escalas foram utilizadas com intuito de compara os resultados obtidos. A análise de desempenho foi realizada com o coeficiente de eficiência Nash-Sutcliffe (NSE). Ambos os modelos puderam ser calibrados de maneira satisfatória para os dados de entrada da bacia em estudo e apresentaram melhor desempenho na escala mensal. Entretanto, o HyMOD apresentou melhor ajuste na calibração (NSE: 0,80) e maior flexibilidade no ajuste dos coeficientes NSE e uma representação gráfica mais precisa quando comparada ao Tank Model (NSE: 0,75).Palavras – chave: Modelagem Hidrológica, calibração, validação, HyMOD, Tank Model. Calibration and validation of hydrological models for a hydrographic basin in the state of Goiás.  ABSTRACTThis work aimed to implement, calibrate, and validate the conceptual rain-runoff models Tank Model and HyMOD to simulate the flow in the Rio das Almas hydrographic basin in the state of Goiás. The dataset used covered a series of 32 years of daily hydrological data (August 1, 1982, to July 31, 2017), the years 1995, 1996 and 1997 were excluded for not presenting complete flow data. Precipitation data were obtained from the records of the three pluviometric stations located in the basin area and the daily mean precipitation volume was calculated using the Thiessen Method. The optimization of the parameters of the Tank Model and HyMOD models was performed using two automatic calibration methods using algorithms from the Excel Solver tool: Nonlinear Generalized Reduced Gradient (GRG) programming and the Evolutionary method. Such algorithms were applied considering flow and precipitation data on a daily and monthly scale. These scales were used to compare the results obtained. The performance analysis was performed with the Nash-Sutcliffe efficiency coefficient (NSE). Both models could be satisfactorily calibrated for the input data of the basin under study and presented better performance in the monthly scale. However, the HyMOD showed better calibration adjustment (NSE: 0.80) and greater flexibility in the adjustment of NSE coefficients and a more accurate graphical representation when compared to the Tank Model (NSE: 0.75).Keywords: Hydrological Modeling, Calibration, Validation, HyMOD, Tank Model.