https://repositorio.ufms.br/handle/123456789/11980 2026-04-05T18:55:00Z https://repositorio.ufms.br/handle/123456789/11899 Título: Identificação de erosões em cultivos de cana-de-açúcar utilizando inteligência artificial Abstract: Soil erosion is a critical challenge for agriculture, impacting productivity and environmental sustainability. This study investigates the application of artificial intelligence models for the automatic segmentation of erosive features in sugarcane cultivation areas using drone imagery. Three semantic segmentation models (DeepLabV3+, InternImage, and SegFormer) were evaluated at different resolutions (1024x1024, 512x512, and 256x256), analyzing metrics such as IoU, Precision, Recall, and F1-Score. Results showed that InternImage 1024x1024 achieved the best performance, reaching 67.61% IoU and an F1-Score of 80.55%, with lower VRAM consumption and higher processing capacity. The qualitative analysis reinforced these findings, indicating that models trained at higher resolutions provide better accuracy. The study concludes that AI-based approaches for erosion detection are viable, scalable, and effective, with potential applications in agricultural monitoring and soil conservation programs. Tipo: Tese 2025-01-01T00:00:00Z https://repositorio.ufms.br/handle/123456789/11846 Título: Informações geointegradas no planejamento e políticas públicas para cidades resilientes Abstract: Urban expansion affects environmental and territorial dynamics, altering vegetation cover, thermal patterns, and urban infrastructure. This dissertation investigates the application of geospatial technologies in vegetation analysis, climate scenario projection, and support for sustainable urban planning. The first chapter presents a bibliometric and scientometric analysis of the Normalized Difference Vegetation Index (NDVI) and its application in urban management, identifying research gaps and emerging trends. The second chapter focuses on multitemporal data extraction and processing, using remote sensing imagery to assess vegetation cover in Campo Grande/MS between 1985 and 2020, quantifying land use transformations. The third chapter explores the integration of urban indicators and geotechnologies in territorial strategy formulation for resilient cities, incorporating multitemporal analysis of NDVI, Normalized Difference Built-up Index (NDBI), Vegetation Connectivity, Land Surface Temperature (LST), Surface Urban Heat Island (SUHI), and Urban Thermal Field Variance Index (UTFVI) from 1985 to 2024.The fourth chapter proposes urban planning strategies for 2050, addressing landscape connectivity, water resource management, and climate adaptation. The analysis examines urban sprawl and its impacts on drainage, identifying hydrological vulnerability due to land use changes. The fifth chapter applies predictive modeling to estimate urban thermal dynamics until 2050, utilizing data clustering (K-Means) and land surface temperature (LST) projections. The comparison of predictive models indicates that the Vector Auto Regression (VAR) model achieved the highest accuracy in estimating thermal variability, outperforming Support Vector Machines (SVM) and artificial neural networks. The findings reveal a significant temperature increase in highly urbanized areas, impacting energy consumption and quality of life. The study evaluates mitigation strategies, such as green infrastructure and ecological corridors, to reduce thermal effects and enhance climate resilience. This research demonstrates the feasibility of urban environment analysis through the identification of spatial transformations over time, associated with vegetation cover and built-up areas. The findings reinforce the need for interdisciplinary approaches in territorial management, highlighting how the integration of geotechnologies, predictive modeling, and urban planning can support public policies and optimize land use in urban expansion scenarios. Tipo: Tese 2025-01-01T00:00:00Z https://repositorio.ufms.br/handle/123456789/11752 Título: Hydrodynamics of four moments in the life of a floodplain forest in compound channels Abstract: The dynamics of compound channels are of critical importance to various fields, including flood management, environmental conservation, and hydraulic engineering. This thesis investigates the hydrodynamics of compound channel flows, with a particular focus on the interaction between floodplain forest vegetation and flow dynamics. Forested floodplains play a crucial role in biodiversity and ecosystem services, yet their natural succession and interaction with hydraulic processes pose challenges for flood risk management and navigability. Moreover, assessing the natural composition and distribution of forests in floodplains is essential to better understand hydrodynamics under natural conditions. This work addresses these challenges by scaling a real floodplain forest in the Upper Rhine and conducting laboratory experiments to examine how forest succession, management practices, and flood intensity influence the hydrodynamics of a compound channel flow. Data from a comprehensive field survey conducted in the “Rastatter Rheinaue” floodplain forest, in the Upper Rhine in southwest Germany, served as the basis for scaling vegetation characteristics in the laboratory experiments. The experiments were designed to replicate key conditions observed in the field, with controlled variations in vegetation composition, density, and flow discharge. Four stages of the secondary forest succession were investigated (Bare or non vegetated, Early, Mid and Late), alongside the impact of management practices, such as selective plant removal. Additionally, the effect of flood intensity was analyzed under both intermediate and deep flow conditions. Measurements of water levels, velocities, and solute dispersion were collected to characterize the hydrodynamics of a compound channel flow with nature-based floodplain vegetation, corresponding to the above-mentioned scenarios (succession ages and management options). These results provide insights into flood risk, navigability, hydraulic modeling, and the lateral exchanges of mass and momentum. The heterogeneous distribution of vegetation introduced local velocity effects that challenge conventional descriptions of depth-averaged streamwise velocity profiles. To address this, a model was developed to account for these local effects, incorporating the characteristics of the mixing layer and assigning physical meaning to key modeling parameters. This model enabled the description of flow dynamics and the analytical solution of important flow regions, such as the mixing layer at the transition between the floodplain and the main channel. A primary objective of this research was to understand how forest age and structure affect hydraulic resistance and flow separation, which are critical for managing flood risks and ensuring navigability. The results indicate that forest aging must be considered in river management, as it has impacts on the upstream flood risk and has significant changes in the hydraulic roughness. Management practices, such as the removal of young trees, can change the characteristics of the floodplain forest, for instance modifying its hydraulic behavior from Mid age to Late age forest. This research shows that such management options have impact in upstream flood risk. This thesis also examines the complexities of the mixing layer formed at the transition between the main channel and the floodplain, which is crucial for sediment transport, nutrient exchange, and habitat connectivity. The presence of vegetation in the floodplain was found to increase the extent of the mixing layer within the floodplain and overall widen the total mixing layer width. These changes affect the transverse exchanges of mass and momentum, which have important implications for ecosystem functioning, as they influence the movement of solutes, organic matter, and aquatic organisms across the channel cross-section. Additionally, the movement of dissolved solutes was investigated, providing insights into dispersion patterns and tracer cloud development, and their relationship with the presence of a mean transverse flow. A conceptual model is introduced to describe the different flow regions in the transition from a compound channel without forested floodplain to a forested floodplain, which vary in terms of energy levels and mean flow. The characterization of this transition region in terms of flow development and their influence on lateral exchanges of mass and momentum are demonstrated. Understanding these zones can offer valuable insights into sedimentation patterns and nutrient availability, which in turn affect forest growth and the development of heterogeneous habitats. In conclusion, this thesis advances our understanding of the hydrodynamics of compound channel flow with forested floodplains, with significant implications for both hydraulic engineering and environmental management. By highlighting the role of forest succession in shaping flow characteristics, it provides valuable insights for improving flood risk mitigation, navigability, and ecosystem resilience in riverine environments. Key Points: • First research to hydrodynamically access natural vegetation distribution on floodplain, resembling floodplain forest composition in the Upper Rhine section. • Effects of forest aging, management, and flood intensity on river hydrodynamics and flood protection are highlighted. • A model was developed to describe the mean flow profile under the influence of local effects caused by vegetation. • Forest-induced mean transverse flow impacts lateral exchanges in compound channel flow with vegetated floodplains are shown. • For the first time, it was shown that forest age and management impact mass and momentum exchanges between floodplain and main channel. • A conceptual model of the flow in a compound channel with vegetated floodplain and its consequences to the mixing layer is proposed. Tipo: Tese 2025-01-01T00:00:00Z https://repositorio.ufms.br/handle/123456789/11584 Título: Remoção de surfactantes e nutrientes de água cinza em wetlands construídos sob diferentes condições hidráulicas Abstract: The search for sustainable solutions for wastewater treatment has driven the adoption of nature-based technologies, such as constructed wetlands, systems recognized for their efficiency in pollutant removal, in addition to presenting a combination of simple techniques, without requiring high operational skills and high energy and maintenance costs for operation. This study aimed to optimize the efficiency of the EvaTAC (Evapotranspiration and Greywater Treatment) system in the removal of surfactants, nutrients and other pollutants from laundry graywater operated under different hydraulic configurations and conducted on a pilot scale. Modifications in the saturation level and in the system feeding method were tested. Physical-chemical parameters were analyzed, in addition to the removal of chemical oxygen demand (COD), anionic and nonionic surfactants and the nutrients total nitrogen (TN) and total phosphorus (TP). The influence of hydraulic changes on the system efficiency is noted with the greater removal of COD (86%), nonionic surfactants (98%) and PT (23%), observed in the T3 test. Changing the feed flow to vertical in the constructed wetland (T3) contributed to reducing the COD concentration to 44.1 mg.L-1, the non-ionic surfactants concentration to below 0.1 mg.L-1 and the PT concentration to 0.3 mg.L-1. The initial configuration, with a subsurface horizontal flow constructed wetland, proved to be more efficient for the removal of NT (94%) and anionic surfactants (93%), reducing the concentrations to 1.3 mg.L-1 and 2.5 mg.L-1, respectively. The pH parameter varied throughout the treatment, while the turbidity and the total suspended solids (TSS) concentration were significantly reduced in T3 (97%), reaching 1.5 mg.L-1 for TSS. Therefore, the variation of hydraulic configurations has the capacity to optimize treatment processes in systems such as EvaTAC. Tipo: Dissertação 2025-01-01T00:00:00Z