workshop usgs brasil_2015_01
TRANSCRIPT
Near-real time pan-tropical monitoring system for detection of changes in natural vegetation
12th Regional Workshops on Forest Monitoring GEO GFOIEarly Warning Systems for deforestationJanuary 19-23, 2014San Jose Dos Campos, Brazil
Overview
Context
Methods
Applications
Impact
Ongoing developments
Conclusions
Context
Deforestation
52% commercial agriculture 33% small scale agriculture 7% roads construction
6% minning 2% urban expansion
5
Free data for monitoring
natural covers changes in
Latin America and the Caribbean
•A mapping tool to detect areas of rapid habitat change
•250m resolution (high percentages of disturbance events larger than 5 ha are identified)
•Frequent natural covers change monitoring, every 16 days
•Latin American and the Caribbean coverage (currently)
•Web tools available to visualize and download habitat loss data
The Bottom Line Limits...
Terra-i IS NOT the tool togive detailed estimates ofdeforestation area andsmall logging activities
Terra-i could help toprioritize high-resolutionanalyses
The methods for deforestation detection only were functional for specific ecosystems
Forest monitoring
In 2006, only one tropical country monitored deforestation: Brazil
There was not an accurate estimation of de forestation (each country use differentmethodologies and statistics)
To use high-frequency imaging and moderate
spatial resolution for ...
Monitoring the conversion of natural habitats in near real time. (Results 2
months after the date of capture)
Have a continental coverage of all types of habitat.
Be a support for government agencies in making decisions.
Quantifying habitat conversion rates and make analysis of trends from
2004 to date.
Monitor the impact on protected areas in Latin America.
Terra- goals
Paula PazTerra-i
Jerome TouvalTNC
Andres PerezHEIG-VD
Mark MulliganKCL
Karolina ArgoteTerra-i
JhonTello
Andy Jarvis
Carolina Navarrete
AlejandroCoca
Edward GuevaraCIAT
Terra- team
Oscar BautistaTerra-i
LouisReymondin
Method
The detection step (using data from 2004 to present)
2
Research methodology overview
The methodology can be split into two main steps:
The training step (using data from 2000 to 2004)
1Bayesian-probability based neural network (BNN) learns how the greenness of a given pixel
responds to a unit of rainfall
INPUT DATA:
Vegetation Index (MOD13Q1 MODIS / NDVI Product , 16 days, 250m)
Precipitation Data: Tropical Rainfall Measuring Mission - TRMM 3B42 (3hours, 28km)
Calibrated model is run to identify fluctuations in greenness that cannot be explained by
rainfall or by previous state of the vegetation
OUTPUT DATA:
Natural cover change data (gain or loss, annual or by 16 days period, 250m)
Terra-i System
Workflow
Tiempo
Vegetation modelling
Anomaly
Modelo
Satélite
Tiempo
Vegetation modelling
OUTPUT: 16 day predicted NDVI
PredictionMultilayer perceptronBayesian Neural Network (BNN)
Model trainning and noise approximationScaled Conjugate Gradient (SCG)Gaussian noise
Input automatic selectionAutomatic relevance determination (ARD)
The goal of the model is to predict what is the NDVI value at the date t taking as input the NDVI values at t-1, t-2 … t-n and the previous rainfall.
INPUTS: Past NDVI (MODIS 13Q1)Previous rainfall (TRMM 3b42)
change
Methodology – Change detection
Debido a que terra-i genera mapas deprobabilidad de conversión, se usaronimagenes landsat para calibrar losresultados y así seleccionar los umbralesde probabilidad más apropiados paracada cluster.
2004
2009
Calibration using Landsat images
34 Satellite Scenes
Vegetation change maps every 16 days
PRODUCTS
Management of massive datasets
- every 16 days we analyze 1.15 billion pixels -
Decrease Increase Flood
Terra-i System
Bolivia
Products
1 escena TRMM
Precipitación(3b42 v7)
+
Comparison of Terra-i results with other local models
Terra-i results were compared with deforestation data produced by the National Institutefor Space Research Instituto Nacional de Pesquisas Espaciais (INPE) from 2004 to 2009through monitoring systems as PRODES and DETER.
PRODESThe Project of estimation of deforestation in the Brazilian Amazon (PRODES) generatedestimations from 2003 using a digital classification system with Landsat images (30m).
DETERDETER is a near real time deforestation detection system. It publishes fortnightlydeforestation alerts for the Brazilian Amazon using MODIS images (500m).
The comparison shows a high correlation between Terra-i and PRODES systems.
Comparación con PRODES
% de las detecciones de PRODES dentro de los pixeles MODIS
% d
e d
etec
cio
nes
igu
ales
Comparison with PRODES
A team that optimizes resources and processes
The images used MODIS and TRMM - do not have any cost. The costs are associated with theprocessing equipment and hiring specialists in handling this tool
Most processes are automated using programming languages like JAVA are efficient in handling large database.
This has allowed a multidisciplinary team of four people working at 100% resulting in a successful generation of continuous updates to the current date
The outreach of the project is fully supported by CIAT communication team and several short reports highlighting new patterns in our data have been made available on our website.
Policy of free data access
Expert users can download our data in a format readable in GIS software (Raster)
Users without knowledge of spatial data analysis can visualize and downloaded our data and charts in
different format
A TOOL TO SUPPORT RESEARCH AND DECISION MAKING
http://www.terra-i.org/
GIS EspecialistsNo GIS especialists
Uses
Octubre 5, 2012Caso Tamshicayu, Perú
Detecciones Terra-i
Landsat 8
Ucayali, Perú
San Martin, Perú
Application 1: monitoring the expansion of large areas crops
Photo: A. Coca / 2013
Photo: A. Coca / 2013
Application 1: monitoring the expansion of large areas crops
Integrandoproyectos
Basado en IPCC
Application 2: understanding changes on the field (validation)
Ecoregión del Cerrado BrasileroEcoregión del Gran Chaco
Application 3: detecting changes in other ecosystems different than tropical forests
Application 4: increase product
Application 5: integration to other policy support systems
• Terra-i can also be used within the WaterWorld and Co$ting Nature Policy Support Systems to
understand the impact of recent land cover change on hydrology and the production and
delivery of ecosystem services.
• Data: http://geodata.policysupport.org/
Water flows Erosion
Impact
http://www.terra-i.org/terra-i/data/data-terra-i_peru
Reunión Lima, Marzo 2014
Terra-i Perú (Monitoring vegetal cover of a territory)
Cooperation with independent media
Plataformas endiferentes formatos
aumentan la participación de la
sociedad civil basada enel uso de datos
espaciales para discutirlos eventos de su región
Impact
Feb 2012 a Dec 2014
50 daily visits
1500 users250 institutions
185 followers
411 fans
Website
On-going development
Integration with Global Forest Watch
http://www.globalforestwatch.org/
Expand terra- Pan-tropically
En funcionamiento y actualizado En proceso de expansión
A mapping and monitoring system for rapid assessment of land cover conversion at a medium scale(250m).
A tool for monitoring conversion of habitat at continental, national and regional level in close to realtime.
A tool for understanding the effectiveness of protected areas and other conservation measures instabilizing or reducing land cover conversion.
A spatial support system for decision making in public policy and private development initiatives.Through its linkage with WaterWorld and Co$ting Nature, a system for understanding the likely impactsof near real-time land cover change on a wide range of ecosystem services.
X Detailed monitoring tool in local level. For this it requires second-level monitoring (with highresolution images) and third level (field data).
X A system to monitor degradation.
Is:
Is not:
Conclusions
• Deforestation is changing very fast and thatthreaths several ecosistems, for this reason is veryimportant to have early warning systems.
• More research is still needed in order to assessconservation policies, actions against deforestationand diverse land and cover changes.
Conclusions