Nilson Clementino Ferreira

An operational deforestation mapping system using MODIS data and spatial context analysis

The Brazilian Amazon has the worlds highest absolute rate of forest loss, currently averaging nearly 2 million hectares per year. In this study, we present a near‐real‐time automated deforestation mapping system for the Brazilian Amazon based on the analysis of spatial context information and MODIS Vegetation Index products and implemented on an ArcGIS 9.0 platform. This system, already validated and operational, was developed as part of the “Integrated Warning Deforestation System for the Amazon” (SIAD), an initiative of the Brazilian government within the scope of the Amazon Protection System (SIPAM), which also comprises: (1) a spatial information module, aimed at the assessment of the causes and impacts of the deforested areas; (2) a prediction module, indicative of deforestation trends; and (3) a data and information gateway based on map server technology.

Deforestation Hotspots in the Brazilian Amazon: Evidence and Causes as Assessed from Remote Sensing and Census Data

Abstract The main goal of this study, conducted in an area comprising 221 municipalities, in which 90% of the deforestation in the Legal Amazon takes place, was to understand the role of the agrarian structure in the conversion of forest into pasture and agriculture fields. Linear regression results indicate that 54%–62% of the variation in deforestation occurred between 1997 and 2004, respectively, and are explained as a function of changes in the amount of appropriated land in 1995. Likewise, up to 80% of the deforestation can be well explained by the variation in land concentration. In fact, strong spatial correlations were found between deforestation hotspots and land appropriation and land concentration. On the other hand, these critical areas have insufficient governance, particularly at the federal level. As the results of this study clearly demonstrate, strong governance and institutional integration, with emphasis on the territorial ordainment, are mandatory in order to reduce the rapid pace of d...

Assessing the response of the MODIS vegetation indices to landscape disturbance in the forested areas of the legal Brazilian Amazon

In this study we assessed the impacts of forest fragmentation on the Amazon landscape using remote sensing techniques. Landscape disturbance, obtained for an area of approximately 3.5 × 106 km2 through simple spatial metrics (i.e. number of fragments, mean fragment area and border size) and principal component transformation were then compared to the MODIS (Moderate Resolution Imaging Spectroradiometer) NDVI (Normalized Difference Vegetation Index) and EVI (Enhanced Vegetation Index) seasonal responses. As expected, higher disturbance values prevail in the southern border of the Amazon, near the intensively converted deforestation arc, and close to the major roads. NDVI seasonal responses more closely follow human-induced patterns, i.e. forest remnants from areas more intensively converted were associated with higher NDVI seasonal values. The significant correlation between NDVI seasonal responses and landscape disturbances were corroborated through analysis of geographically weighted regression (GWR) parameters and predictions. On the other hand, EVI seasonal responses were more complex with significant variations found over intact, less fragmented forest patches, thus restricting its utility to assess landscape disturbance. Although further research is needed, our results suggest that the degree of fragmentation of the forest remnants can be remotely sensed with MODIS vegetation indices. Thus, it may become possible to upscale field-based data on overall canopy condition and fragmentation status for basin-wide extrapolations.

Assessing deforestation related carbon emission in the Brazilian Savanna based on moderate resolution imagery

The Brazilian Cerrado is the second largest biome in South America, comprising an area of approximately 2 million km2 or 24% of the Brazilian territory. In spite of its important role in carbon fluxes and storage, this biome has been facing intensive, fast-pace, agro-pastoral land use, as the occurrence of favorable topography and soils of easy management enabled highly-mechanized, rainfed grain and cattle meat productions for exportation. This paper presents an overview of a methodological approach for assessing carbon emission caused by deforestation in the Brazilian Cerrado biome, based on multi-temporal and multi-thematic analysis. Preliminary results, based on the 250m Terra-MODIS normalized difference vegetation index (NDVI) images, suggest that about 45 million tons of carbon were emitted due to the 2008 – 2009 clearings in Cerrado biome.

Análise da vulnerabilidade natural da paisagem em relação aos diferentes níveis de ocupação da bacia hidrográfica do Rio Vermelho, estado de Goiás

A bacia hidrografica do Rio Vermelho, situada no Estado de Goias, possui uma elevada geodiversidade, com terrenos que vao do Arqueano ao Holoceno, e um proeminente registro tectonico impresso. Da mesma forma, ao longo de seu processo historico de ocupacao, tem sido impactada por variadas atividades de producao do solo e do subsolo, o que acarretou tanto o empobrecimento da biodiversidade, quanto a modificacao do seu aspecto fisico, sendo evidente o aumento de processos erosivos e acumulos de sedimentos em razao da dinâmica de entrada e saida de energia e materia do sistema natural. Neste trabalho, o qual teve por base a elaboracao e analise do mapa de entropia do relevo para toda a area da bacia (~ 11.000 km2), foi possivel identificar areas mais suscetiveis a atuacao de processos morfogeneticos e, portanto, mais vulneraveis a ocupacao, a exemplo das areas proximas a foz da bacia, cujos valores de entropia sao coerentes com a diversidade de usos e intervencoes sobre o complexo sistema fluvial entorno do canal principal do Rio Vermelho. Em que pese a necessidade de analises adicionais, a escalas de maior detalhe, nossos resultados demonstram o potencial da analise da entropia do relevo como ferramenta importante para identificar areas com energia potencial disponivel para realizacao de trabalho geomorfologico e, portanto, mais vulneraveis a ocupacao (i.e. sujeitas a enchentes, inundacoes e deslizamentos de terras).

AVALIAÇÃO DA TURBIDEZ EM AMOSTRAS DE ÁGUA TRATADA COLETADAS NOS MUNICÍPIOS DO ESTADO DE GOIÁS

This study evaluated the turbidity of 5034 water samples collected at the water supply systems of 224 municipalities in the State of Goias during a period of 24 months, between 2011 and 2013. Samples were collected at the beggining, end and also in intermediate points of water supply systems, as well as in areas with increased flow of people, such as hospitals, clinics, schools and nurseries. The results showed that in the rainy season were found 78.1% of occurrences of turbidity over 1.0 NTU in relation to drought, and the average rate was 0.98 and 2.20 NTU, respectively, to periods of drought and rainy. It was observed that 23.91% of the evaluated samples showed turbidity values greater than 1.0 NTU and 3.5% greater than 5.0 NTU. There is a variation between the results presented, that is from 1.9% to 8.0% of the mesoregions and 0 to 8.9% incidence of microregions with turbidity higher than 5.0 NTU.

A land cover map for the state of Goi's, Brazil: a comparative analysis of different mapping approaches and strategies

In this study, we assessed how low spatial resolution and coarse scale land use information derived from SPOT Vegetation, MODIS composited reflectances, and Agricultural Census data compare with the new Landsat-based land cover map of the State of Goias. This State, located in the central part of Brazil, is mostly covered by the neotropical savanna, locally known as the Cerrado. The maps derived from both medium and low spatial resolution remotely sensed products showed similar land use areas in the State: 64% (Landsat), 68% (SPOT), 63% (MODIS) and 62% (Agricultural Census). Preliminary results suggest a possible synergism between low or moderate and medium spatial resolution imagery for the monitoring of the land occupation, which certainly has practical importance for natural resources management in the fast changing State of Goias.

Análise evolutiva de processo erosivo acelerado em uma área urbana da cidade de Goiânia

No contexto atual o crescimento das cidades, a deficiencia no planejamento e as caracteristicas das precipitacoes, principalmente em grandes centros urbanos, influenciam nos processos erosivos. Dessa forma, este trabalho propoe a verificacao da origem e evolucao de uma vocoroca localizada no municipio de Goiânia-GO. Realizou-se uma analise temporal no periodo de 2002 a 2014 utilizando imagens de satelite de Google-Earth e ortofotograficas. As imagens foram editadas e geoprocessadas utilizando o programa computacional ArcGis. Verificou-se que os principais fatores para a formacao e evolucao da erosao foram a naturaliza do solo e o inadequado tracado do sistema viario e do planejamento local.

Avaliação multitemporal do uso e cobertura do solo da bacia hidrográfica do Ribeirão João Leite utilizando imagens landsat-5

The analysis of land use and land cover is fundamental for the strategic planning to the conservation and management of a watershed. This study aimed to analyze the landscape of the catchment area of João Leite River, located in the State of Goiás, to the Northwest of the city of Goiânia, the state capital, using Geographic Information System (GIS) and Remote Sensing to prepare an environmental assessment of that watershed. A multi-temporal analysis using images obtained by the TM sensor installed in the Landsat-5 satellite, from 1985 to 2011 showed, in the catchment area of the João Leite River, a decrease of 7.4 % in the native vegetation area. Considering the municipalities integrating the basin, Anápolis was where most deforestation occurred in areas with native vegetation (11.1%). These results are relevant to institutions dealing with environmental management, showing that deforestation is directly related to the environmental quality of the basin, which has as one of its uses the urban water supply of municipalities located in the Metropolitan Region of Goiânia.

A protocol for expedite and effective retrieval of land conversion information in savanna environments

The production of land use and land conversion maps has become a priority information for public policies actions and environmental monitoring. In Brazil, the elaboration of these maps (with moderate scale/resolution) has been driven by the need to evaluate human occupation dynamics, principally in the Amazon and savanna (or Cerrado) biomes. In order to enable a more precise and systematic monitoring for the savanna region (which will support, for example, the deforestation, economic and greenhouse gases expansion models), this study aims to present a basic protocol for mapping this biome (in three thematic classes: Remnant vegetation, Converted areas and Burned areas), taking into consideration computational time, budget issues, as well as accuracy and precision of the area mapped (including error estimates). Our results suggest the possibility to produce low cost expedite maps for the entire Cerrado, with a global accuracy above 80% (in some scenes near to 100%), especially for both the Remnant vegetation and Converted areas.