Uncertainities on land cover
and land ue data sets
designed for global models
Gilberto Câmara
National Institute for Space Research (INPE)
Brazil
IFGI, University of Münster, Germany
http://www.dpi.inpe.br/gilberto
Why are most remote sensing researchers
like Indiana Jones?
The world is divided in cells. Each cell has a single class.
There is a correct classification. The more our classification
approaches the ideal, the better.
Statistics on land change modelling
Measure correlation between causes and effects
Causes: food production, population increases
Effects: land changes (remote sensing)
y=a0 + a1x1 + a2x2 + ... +aixi +E
Statistics: Assessment of land use drivers
A. Aguiar, G. Câmara, I. Escada, “Spatial statistical analysis of land-use determinants in the Brazilian Amazon”. Ecological Modelling, 209(1-2):169–188, 2007.
G. Espíndola, A. Aguiar, E. Pebesma, G. Câmara, L. Fonseca, “Agricultural land use dynamics in the Brazilian Amazon based on remote sensing and census data”, Applied Geography, 32(2):240-252, 2012.
Land use models are good at allocating change in space. Their problem is: how much change will happen?
How good are statistical models?
0 10 20 30 40 50 60 70 0 1 2 3 4 5 6 7 E rr o (% )Log2(size of comparison window)
Amazonia: estimated x real deforestation
Laurance O Laurance NO
SIMAMAZONIA BAU SIMAMAZONIA GOV Simplificado Vizinhança
~230 scenes Landsat/year
Yearly detailed estimates of clear-cut areas (wall-to-wall)
Data released openly after 2003
INPE: Clear-cut deforestation mapping of
Amazonia since 1988
The unbearable lightness of PRODES
Land cover
wetlands tropical forest
Non-natural vegetation shrublands
Land use
cattle production non-managed forest
Temporary agriculture shifting cultivation
“the arrangements, activities and inputs people
undertake in a certain land cover type to produce,
change or maintain it”
What happened with 720.000 km2 deforested?
TerraClass - first map of land use and land cover of
Amazonia
How are we using the forest?
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% AC MT PA RO Degraded Land Second Veg Degrad Pasture Pasture Small Farms GrainsWhat’s in an image?
“Remote sensing images describe landscape dynamics”
Annual Crop – 1 Crop per season Annual Crop – 2 Crops per season Sugarcane
Crop Year 2002/2003
Annual Crop – 1 Crop per season Annual Crop – 2 Crops per season Sugarcane
Crop Year 2010/2011
MATO GROSSO – Sorriso: Crop Year 2002/2003
Annual Crop - 1 Crop Annual Crop - 2 Crops
MATO GROSSO – Sorriso: Crop Year 2010/2011
Annual Crop - 1 Crop Annual Crop - 2 Crops
Vegetation index
time series
Área 1
Área 2
Área 3
source: Victor Maus (INPE)
Time series analysis of land change
Forest
Pasture
Forest
Unique repository of knowledge and data about global land change
Global Land Observatory:
describing change in a connected world
40 years of LANDSAT + 12 years of MODIS + SENTINELs + CBERS Free satellite images Methods for land
change for foresty and agriculture uses
TerraME Runtime Environment
Eclipse & LUA plugin • model description • model highlight syntax
TerraView • data acquisition • data visualization • data management • data analysis TerraLib database d a ta Model source code MODEL DATA m od e l
• model syntax semantic checking • model execution TerraME INTERPRETER LUA interpreter TerraME framework TerraME/LUA interface m o d e l data