Banca de DEFESA: RAFAEL SANTOS SILVA

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
STUDENT : RAFAEL SANTOS SILVA
DATE: 23/08/2023
TIME: 14:00
LOCAL: meet.google.com/ynd-yaus-rmu
TITLE:

USE OF SATELLITE IMAGES FOR TEMPORAL ANALYSIS OF LAND USE AND PLANT COVER IN HYDROGRAPHIC SUB-BASINS

KEY WORDS:

Micro-regional scale analysis, land use classification, remote sensing, Landsat, time series; linear trend; SATVeg; ClimateEngine.

PAGES: 48
BIG AREA: Ciências Humanas
AREA: Geografia
SUMMARY:

Human actions linked to land use and occupation affect all environmental spaces, making it an important agent in reducing vegetation cover. In this sense, forest fragmentation causes environmental problems, such as reduced biodiversity, local climate change, soil erosion, loss of agricultural soil and silting up of watercourses (LÔBO, 2021). Furthermore, anthropic activities affect the hydrological cycle of watersheds, which are important for the management of water resources and for environmental planning, as they facilitate the identification of conservation areas (AL-JAWAD et al., 2019). In this regard, watersheds are systemic areas that have their ecosystem in total interaction, suffering imbalance when one or another organism receives some external interference (NASCIMENTO and FERNANDES, 2017). Thus, the negative impacts caused in the geographic space of the watershed express the socio-environmental dynamics of society (JOIA et al., 2018), which causes reductions in native forest areas and affects the entire existing ecosystem, causing changes in the biome and ecological processes, such as the cycle of nutrients (OLIVEIRA, 2020). Areas that are occupied incorrectly also present negative results such as the reduction of natural resources, loss of soil quality, increased processes of land wear and a decrease in biodiversity (NASCIMENTO, 2017). Thus, it is important to understand the society/nature dynamics and the way in which man interacts, uses and occupies spaces over time (CANTO and ALMEIDA, 2008), in order to identify spaces in rural areas that may have been altered by human actions to create possible recovery plans for these areas or mitigation of damage caused. To identify the interaction and impact of man on nature, it is possible to use satellite images to map and analyze the influence on the environment. Representations of landscapes by Geographic Information System (GIS) are used to record human actions in space, to monitor and control the environment. (PEREIRA and SILVA, 2001). For Borges et al. (2019), the use of geotechnologies enables studies to develop more efficient environmental policies. Given this, Burke (2017) reports some advantages of using remote sensing as a use for: (1) cheap field yield estimates that may allow better targeting of future policies to be adopted; (2) productivity measures that can carry out impact assessments of agriculture, thereby expanding the evidence base on the effectiveness of particular interventions; (3) ability to measure activity output across a large number of plots over time which improves ability to understand the magnitude and determinants of yield gaps; and, (4) verification of field-scale productivity could support the development and expansion of financial resources for smallholders with letters of credit. The geospatial data used are available at different scales, resolutions and time intervals on government websites. Satellite images provided in recent decades with high resolution are used to create local and regional thematic mapping, helping in research, such as the classification of land use (MANCINO et al., 2014). The images that are available on government websites have two important advantages: the periodicity of the data available and the low cost. Given the above, from the available remote sensing techniques for analysis, monitoring and management of spaces, land use and land cover classification techniques and analysis of the normalized difference vegetation index (NDVI) were used. The satellite images, present in the first chapter, were used to classify land use and occupation and were used to quantify and qualify spatial data, which can contribute to the construction of conservation policies and monitoring of detection of changes in the landscape (DUARTE and SILVA, 2019). In the second chapter, the Normalized Difference Vegetation Index (NDVI) was used, which helps to distinguish the type of vegetation cover and to determine the health of the plant. The NDVI demonstrates the vigor of vegetation, as it consists of values ranging from -1 to +1, with negative values representing the absence of vegetation and positive values denoting the presence of vegetation (VENTURA et al. 2019). Research using the NDVI analyzes various weather events in order to improve our predictions and assessments of impacts such as aridity (KANG et al., 2022), soil degradation (HUANG, et al., 2010), fire (RANSON et al., 2003) and rainfall variations (WANG, et al., 2014). Given the above, this dissertation is organized into two chapters, equivalent to two scientific articles submitted to Brazilian scientific journals. The two works have the same research area, whose territorial unit was the (sub-) watershed of the Rio das Mortes Pequeno (BHRMP), located in the municipalities of São João del Rei and Conceição da Barra de Minas, MG, Brazil. Therefore, this dissertation has the first article entitled, “Geospatial and temporal analysis of land use and land cover in the Rio das Mortes Pequeno watershed, Minas Gerais”, whose objective is to analyze, over a period of 30 years (1990-2020), the changes in land use that occurred in the BHRMP territory. The territory studied was monitored and classified through analyzes of Landsat 5 and Landsat 8 satellite images. Land use and occupation classification techniques followed the supervised classification model and the maximum likelihood classifier (MAXVER) (NASCIMENTO, 2003), using the “Semi-Automatic Classification Plugin (SCP)” plugin, present in the free software QuantumGIS (QGIS). Analyzes of land use and land cover identified a change in the activities carried out on rural properties, in the hydrographic basin studied, in the last 30 years. In the 1990s, the main activity was class[3] 'Field, pasture and annual crop', which was replaced in 2020 by class[2] 'Planted forest'. In addition, the research identified the lack of preservation of riparian forests, permanent preservation areas (APP), which were replaced by agricultural activities. The second article, “Analysis of variations in vegetation based on the correlation of NDVI and rainfall in the period from 2006 to 2021 in a watershed sub-basin”, aims to verify the progressions and regressions of vegetation in the BHRMP when related to the NDVI vegetation index with precipitation information, through geospatial data and website platforms. Images from the MODIS and CHIRPS satellites were used, based on data from the SATVeg website and the ClimateEngine application for the composition of NDVI time series and precipitation validation. From time series, the classes that registered vegetation regressions and progressions were identified. For that, four different classes were used, which are agriculture, planted forest, pasture and forest.

BANKING MEMBERS:
Presidente - 1729282 - MUCIO DO AMARAL FIGUEIREDO
Interno - 882.024.886-72 - CARLOS FERNANDO FERREIRA LOBO - UFMG
Interna - 1743531 - SILVIA ELENA VENTORINI
Externo à Instituição - ITALO SOUSA DE SENA