Remote sensing, social and economical data to understand agricultural land cover dynamics in eastern Amazonia
•Download as PPT, PDF•
1 like•573 views
Cécile BILLARD Presentation for the conference on Taking stock of smallholders and community forestry Montpellier France March 24-26, 2010
![Introduction ,[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object],[object Object]](data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7)
Recommended
More Related Content
What's hot
What's hot (20)
Similar to Remote sensing, social and economical data to understand agricultural land cover dynamics in eastern Amazonia
Similar to Remote sensing, social and economical data to understand agricultural land cover dynamics in eastern Amazonia (20)
More from CIFOR-ICRAF
More from CIFOR-ICRAF (20)
Recently uploaded
Recently uploaded (20)
Remote sensing, social and economical data to understand agricultural land cover dynamics in eastern Amazonia
- 1. Remote sensing, social and economical data to understand agricultural land cover dynamics in eastern Amazonia Taking stock of smallholder and community forestry: Where do we go from here? 24 - 26 March 2010, Montpellier, France by Cécile BILLARD Msc. Degree in Forest Ecology and Management University of Freiburg, Germany
- 8. Results of classifications Gain in precision & complexity over time 1986 2001 1992 2007
- 10. Typologie of the farmer’s land use over time Identification of 7 dynamics of land use changes 6. Dynamic of regrowth 4. Dynamic of forest conversion into multi-stratified plantations (and pastures) 5. Active dynamic of forest burning (advanced stage) 2. Recent dynamic of forest burning (first stage of forest degradation) 7. Active dynamic of forest conversion into a mosaic of plantations, fallows and pastures 3. Development of agricultural practices (low impact on forests) Axis 1 Axe 2 Axe 1 PASTURE FOREST Axe 2 BURNED FORESTS 1. Recent dynamic of agricultural progression (forest degradation) PLANTATIONS
- 11. Type 6: Dynamic of regrowth traducing pasture and bare lands abandon and the progression of multi-stratified plantations Type 1: Recent dynamic of agricultural progression, especially pastures and multi-stratified plantations, at the detriment of the forest which was already exploited Description of the types Recent land use changes (2001-2007) at the household-level
Editor's Notes
- Settings of the research
- The research issue takes place in the frame of the French AMAZ project (Biodiversity of Amazonian Landscape, Socio-Economic Drivers and Production of Ecosystem Goods and Services). For the purpose of this study, focus was made on the site Palmares II selected from the AMAZ project. It consists in a previous fazenda which has been invaded by the Landless Workers Movement ( Movimento dos Trabalhadores Rurais Sem Terra or MST) in the early 1990’s. The settlement of 517 families was officially recognized in 1997, after the violent conflicts opposing rural workers and the government. The expropriated land is about 15 850 ha. This area was divided for each family into homogeneous lots of 25 ha each 1000 m by 250 m.
- in the context of this paper, these are methodological steps necessary to achieve the objective of the study To quantify and describe the land use and land cover changes using multi-temporal remote-sensing images To establish a typology of the households’ use of the land according to the landscape dynamic on their properties To describe the several types of households living in the area according to socio-economical parameters
- Regular shape of the lots (25 ha each). Each family a lot
- La méthode cartographique repose sur l’utilisation de classification d’images satellite à partir de relevé de terrain. Sur de grandes scènes on définit les éléments du paysage afin de faire des catégories et préparer une typologie complète. La méthode consiste aussi en la définition d’une grille typologique des éléments du paysage. Cette analyse nous a amené à définir un modèle de succession des éléments paysagers qui évoluent dans le temps. Quantification of land cover and land use changes from 1986 to 2007 (in ha)
- Correlation and linear regression were permormed in this analysis to better understand the link between social and economical parameters and changes in land use, especially forest ones.
- The analysis work was divided into two principal axes aiming to assess the landscape dynamic of the site of Palmares II from 1986 to 2007; and to understand and try to explain it using socio-economical indicators. Results provided by the analysis are described following the same logic. Different scales are used to apprehend the land uses and land cover changes over the defined timeframe. The scale of the site of Palmares, grouping all the farmers lots (n=517) and the village; to give an overview of the general trends which are taking place in the area. Then, at a more precise scale which corresponds to what has been described as the sub-windows’ scale in the frame of the Amaz Project. It refers to the selected farmers located in the area of Limão VR and VSF and Três Voltas (n=51). As mentioned in the detailed methodology, a final number of 44 households were identified according to a code defined in the socio-economical survey. The typology of farmers on which the second part of the results is developed was established based on these 44 households. Chaque classification a des paramètres propres de précision et de fiabilité (indice de Kappa, Matrice de confusion) Dominace of forested areas in 1986 Complex mosaic of land use in 2007 On obtient alors une vue du municipio dans le temps. L’évolution des paysages est donc très visible et l’analyse visuelle est correcte. Les données statistiques sur les précisions de classification sont bonnes. On effectue nos points de contrôle des éléments paysagers pour finalement produire une carte de l’occupation du sol (ici Spot de 2007). Les classes correspondent aux éléments du paysages identifiés au sol. Reconstitution par similitude spectrale de la carte des états de surface pour 2001 En reemontant toujours plus loin dans le temps (ici en 1992) on perd de la fiabilité dans les réponses spectrales. On utilise ici d’ancien satellite moins performants. Perte de précision cartographique On obtient alors une vue du municipio dans le temps. L’évolution des paysages est donc très visible et l’analyse visuelle est correcte. Les données statistiques sur les précisions de classification sont bonnes.
- Dans le cadre AMAZ on s’est intéressé à trois sous fenêtres. Chacune d’elle a été extraite et analysée (à droite).
- The household’s typologies aimed to associate particular dynamics and land use changes to several set of households. It corresponds to the second objective. Seven types were distinguished based on a principal component analysis which considered the changes in land uses between 2001 and 2007 at the household-level. Households were divided according to changes recorded on their properties, as follow: Recent dynamic of agricultural progression, especially pastures and multi-stratified plantations, at the detriment of the forest which was already exploited Recent dynamic of forest burning occurring traducing a first phase of forest degradation Development of agricultural practices with lowest impact on forested areas Dynamic of forest (mainly exploited forest) conversion into multi-stratified plantations, and at a smaller extent into pastures Active dynamic of forest burning (advanced stage) Dynamic of regrowth traducing pasture and bare lands abandon and the progression of multi-stratified plantations Active dynamic of forest conversion (especially exploited) into a mosaic of plantations, fallows and pastures.
- To facilitate the interpretation: Description of the 7 defined types (boxplots) using the spatial & temporal data from the remote sensing analysis (land use changes between 2001 and 2007)
- Main relationships highlighted between spatial and social data The seven types of households were described according to a selection of socio-economical parameters. It corresponds to the third objective aiming to draw a general profile of these smallholders based on the information provided by a households’ survey. The next step was to establish a link between the multi-temporal and spatial data and the described socio-economical variables. It corresponds to the further objective aiming to apprehend the influence of the selected independent variables with the landscape evolution and the differential land uses. The main relationships highlighted by a correlation analysis were the following: Deforestation rates tends to increase with a low amount of incomes, and at a smaller extent, a low level of study or a low number of children might have an influence Consequently, changes in forest cover tends to increase for households with low incomes Pasture cover tends to progress for household with low level of study, and at a smaller extent, low amount of incomes and low available working force might have an influence Fallow cover tends to progress for household with high total amount of credit acquired, especially non-refundable ones, and at a smaller extent households’ heads having spent most of their lives in agricultural activities might have an influence on fallow progression too Bare and/or burned lands cover tends to progress for households with a low amount of productive capital for cattle ranching, or with a low duration time on their properties
- Le peu de fermes analysées limites les généralisation mais l’idée est lancée vers cette voie de recherche afin de déterminer quels sont les leviers institutionnel permettant de mieux gérer les ressources (action de politiques publiques). This study enable us to collect landscape elements and to understand the different dynamics taking place within the recent history of the area. It emphasize the need for forest margin consolidation and remaining forest conservation which are linked to political, social and economical drivers. To understand these drivers and their influence on land use decisions at the smallholder’s scale was the necessary steps to achieve for more sustainable resources management. The small sample of household analysed through this study (n=44) limits the generalization the results. However, the new insights provided by this research could be extend at a broader scale as they provide a reliable basis for drawing up local and regional sustainable management policies.
- The first objective of this research was to quantify and describe the land use and land cover changes in Palmares II using multi-temporal remote-sensing images. Based on the typology of land uses defined and on the results of the four supervised classifications, four main dynamics were identified at the scale of Palmares II, assessing the complexity of the landscape evolution over time. These landscape dynamics correspond to those defined by an empirical modeling .