RMRafael Manica

Rafael Manica

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Rafael Manica has a total of 44 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Computational modeling of oil spill dispersion in the port of Porto Alegre/RS

    by Raquel Lopes Belloli, Rafael Manica, Cassiano Ribeiro Michelon, Henrique Wolff Salvador, Eduardo Puhl, Elirio Ernestino Toldo Jr.

    This study evaluated the dispersion of oil spills in the Guaíba River using computational modeling (SISBAHIA). The methodology integrates 3D hydrodynamic modeling with advective–diffusive transport simulations of oil spills , testing nine hypothetical accidental scenarios of 2,400 m3 spills of MF-380 oil (bunker fuel) at the port of Porto Alegre (Guaiba river) and evaluating the oil slick trajectories under varying conditions of water level, river discharge, wind, and margins absorption coefficient. Also, a deterministic and probabilistic transport approaches were employed. The dispersion dynamics of the oil slick reflected the river hydrodynamics: wind effects dominated under low flows confined the oil slick to the western bank, whereas downstream currents governed average and high flows transporting the oil as far as Itapuã. The absorption coefficient has little influence on results. Environmentally sensitive areas identified through SAO maps included the Jacuí Delta, the city of Guaíba, and the Itapuã State Park.

  • Previsão do Risco de Nevoeiro com Aprendizado Profundo: Estudo de Caso no Cais do Porto de Porto Alegre

    by Lucas dos Santos Azevedo, Rafael Manica, Eduardo Puhl

    Este trabalho propõe e valida um pipeline em duas etapas para previsão de alerta de nevoeiro no cais de Porto Alegre. (i) Uma LSTM multivariada (Rede A) prevê, até 100 h à frente, variáveis meteorológicas com base em 318.000 dados (REDEMET [SBPA/SBPK, 2003–2024] e locais); (ii) Uma LSTM binária (Rede B) estima, para cada lead, a probabilidade de nevoeiro usando os previsores gerados pela Rede A. Em 9.000 simulações, a Rede B alcançou AUC = 0,936, recall = 90% e acurácia = 82,52% (7.131 verdadeiros negativos, 477 verdadeiros positivos, 1.471 falsos positivos, 41 falsos negativos), com falsos positivos concentrados entre 2–5 h sob umidade relativa elevada e vento fraco em cenário desbalanceado.

  • Previsão Futura de Séries Temporais Meteorológicas no Cais do Porto de Porto Alegre Usando Aprendizado Profundo com Redes Neurais LSTM

    by Lucas dos Santos Azevedo, Rafael Manica, Eduardo Puhl

    Este estudo avalia um modelo LSTM para previsão intradiária (até 100 horas à frente) de variáveis meteorológicas no cais do porto de Porto Alegre. O modelo foi treinado com janelas de 100 lags, utilizando um conjunto de 306.250 dados da API-REDEMET/SBPA–SBPK (2003–2024) e 12.589 medições in loco no porto (2021–2024). Nos testes comparativos, o LSTM superou a API Tomorrow.io (RMSE 11,93 vs. 22,37; MAE 9,48 vs. 20,59) e o WRF (jan/2022) em todas as variáveis analisadas. Os resultados demonstram a viabilidade e a vantagem de modelos LSTM ajustados localmente para previsões de curto e médio prazo em ambientes portuários, configurando-se como alternativa confiável a abordagens comerciais.

  • Open Access

    Fog Risk and Classification in the Porto Alegre Port Waterway (2001-2024): Binary Fog Risk Index and Multivariate Analysis

    by Lucas dos Santos Azevedo, Rafael Manica, Eduardo Puhl

    This study analyzed 21 years (2003-2024) of hourly data from REDEMET/API - Salgado Filho Airport - to quantify and classify fog risk affecting the Porto Alegre Port Waterway, located 4 km from the airport. A dataset of ~195,000 records underwent quality control, with anomalies replaced using hierarchical means and bidirectional temporal interpolation, and feature engineering (local hour, day of year, and month) was implemented in Python. A binary fog-risk index was defined and validated against REDEMET synoptic alerts. The index was triggered when relative humidity ≥ 95%, temperature-dew-point difference ≤ 2 °C, wind speed ≤ 5 m s-1, or visibility < 4000 m. Event duration was estimated from continuous episodes meeting the risk criteria. Multivariate time-series analysis (STL decomposition, ADF, ACF/PACF) revealed stationarity, strong short-lag autocorrelation, and marked autumn-winter seasonality. Results confirmed that May-July represents the highest operational risk, with ~30 h per month of visibility ≤ 1000 m and > 50 h in the 2001-8000 m range, mostly between 5 a.m.-8 a.m. local time. These findings underscore the need for targeted alerts and the integration of predictive models into port logistical planning.

  • Open Access

    Environmental causes of navigation accidents in the Guaíba River waterway

    by Fernando Comerlato Scottá, Cassiano Ribeiro Michelon, Eduardo Puhl, Rafael Manica, Elírio Ernestino Toldo Júnior, Cristiano Fick, Juliane Coelho Lenhard, Tatiana Silva da Silva, José Carlos Rodrigues Nunes

    This study analyzed the possible causes of accidents on the Guaíba River waterway between 2007 and 2023, using historical reports, numerical modeling, and assessments of critical meteorological and hydrodynamic conditions. Data on wind, visibility, water levels, and flow rates were examined both analytically and graphically. A numerical model simulated the river’s hydrodynamics, identifying key parameters affecting navigation safety. The analysis revealed that various environmental forces such as fog, currents, waves, and wind significantly contributed to accidents, often acting in combination. All recorded accidents were groundings, and the Pedras Brancas Channel as the most critical grounding hotspot along the waterway. Wind speeds above 7 m/s, current speeds above 0.5 m/s, water levels below 40 cm, and wave heights exceeding 20 cm form the set of critical values for navigation safety. These technical insights enable Port Authorities to enhance navigational safety through data-driven interpretations and operational decisions.


  • Open Access

    Inception of Constructional Submarine Conduit by Asymmetry Generated by Turbidity Current

    by Daniel Bayer da Silva, Eduardo Puhl, Rafael Manica, Ana Luiza de Oliveira Borges, Adriano Roessler Viana

    Submarine conduits are features responsible for transporting clastic debris from continents to the deep ocean. While the architecture of conduits has been extensively studied, the process of their inception remains unclear. This study highlights the possibility that some conduits are initiated by depositional processes involving turbidity currents. Here, we present the results of eight experiments where gravity currents were allowed to develop their own pathways. The simulation tank represented natural scales of continental shelves, slopes, and basins. The initial experiments involved sediment-laden flows with low density (1–10% in volume). In first experiment runs (Series I), sediment deposition occurred primarily on the shelf and slope, resulting in an asymmetric transverse profile. This asymmetry facilitated subsequent conservative currents (1034 to 1070 kg/m3 due to salt dissolution) flowing alongside during the second series, resulting in the formation of a constructive submarine conduit. This feature is analogous to gully formations observed in various locations. This study correlates these findings with gully-like features and proposes a model where non-confined density flows can evolve into confined flows through the construction of asymmetric topography. An evolutionary model is proposed to explain the mechanism, which potentially elucidates the formation of many submarine conduits.


  • Open Access

    Wide-swath satellite altimetry reveals the 2024 Porto Alegre extreme flood was intensified by backwater effect across choked river section

    by Leonard Niero Silveira, Vitor Hugo Almeida Jr., Mauricio Kenji Yamawaki, Eduardo Puhl, Rafael Manica, Elirio Ernestino Toldo Jr, Tatiana Silva, Felipe Geremia-Nievinski

    In May 2024, the city of Porto Alegre, southern Brazil, experienced the worst flood in recorded history. Extensive damage occurred despite large flood control measures, such as levees and floodwalls, implemented after the last great flood, in 1941. Wide-swath satellite altimetry, as provided by the NASA/CNES SWOT mission, revealed distinct flooding patterns. In the braided Jacuí river, flooding was aggravated by the backwater effect across a choked section. The wider Guaíba River experienced smaller water levels and slopes. Stream gradient increased dramatically, from 0.4 cm/km to 15 cm/km. We have validated the satellite maps with in situ GNSS hydrographic surveys as well as three water level stations employing GNSS Interferometric Reflectometry. Results suggest flood control could be improved in the north region of the city. Levees failed where water levels were a few meters higher than that experienced at the same time in other regions of the city.


  • Open Access

    Dam-break due to overtopping failure: experiments in homogeneous earthfill dams

    by Bianca dos Santos Von Ahn, Rafael Manica

    This work physically simulates the effect of low and high flow rates and filling times of reservoirs and rupture due to overtopping (caused by intense rains) of small homogeneous silty-sand earthfill dams. The experiments seek to verify how input variations impact the formation of the breach and the rupture wave. The results show that different filling times, soil moisture and composition, and degree of compaction affect landfill saturation, failure time, and breach formation. The result confirms that smaller breaches with a higher degree of compaction led to a lower peak rupture flow compared to dams with low degree of compaction. The rupture hydrograph presents a faster descent stage than an exponential hydrograph. Simulations and models based on this law may minimize the effect of the dam-break wave, also impacting water resource decision-making for damage reduction. The results were extrapolated to a real prototype, providing information and a database for the studies of overtopping dam-break waves.


  • Open Access

    Simulação física de escoamentos superficiais com concentração de pedosedimentos entre 6 e 8%

    by Julio Cesar Paisani, Antonio Carlos De Barros Côrrea, Rafael Manica, Flávia Jorge De Lima, Gustavo Gallardo Aere, Matheus Vinícius Dos Santos, Josielle Samara Pereira

    Com base em resultados prévios de experimentação física em laboratório o objetivo desse trabalho foi verificar se simulações do escoamento superficial com concentrações de pedosedimentos de 6 a 8 % geram depósitos semelhantes a fluxos de menor (~ 3 %) ou de maior (>14 %) concentração relativa. Duas repetições de fluxos exibiram concentrações nesses intervalos e foram analisadas suas velocidades médias, números de Froude e Reynolds, e seus depósitos. Predominou regime de fluxo supercrítico-turbulento entre as condições iniciais e sua chegada no plano de sedimentação, passando paulatinamente a subcrítico-laminar até minguar. Estruturas de superfície dos depósitos são semelhantes quanto à disposição, mas diferiram quanto a largura, preservação da acreção vertical e desenvolvimento de marcas de sulcos; bem como em relação a microestratigrafia de leito, sobretudo pela espessura, organização e o número das laminações. O fato de os depósitos serem formados por mais de uma laminação mostra os efeitos da cauda/corpo do fluxo em resedimentar os materiais previamente depositados pela cabeça do fluxo. Estruturas de superfície e microestratigrafia de leito coincidem com aquelas encontradas tanto em depósitos de fluxos de menor e de maior concentração relativa, acarretando que as concentrações de sedimentos dos fluxos analisados expressam o contínuo entre fluxos de baixa densidade e fluxos transicionais entre baixa e alta densidades.