Terra-i is a system that uses neural networks and MODIS data to monitor habitat change in near real-time. It maps habitat loss every 16 days at 250m resolution. Its goals are to monitor natural habitat conversion, have continental coverage, support government decision making, and quantify habitat change rates. Terra-i predicts vegetation greenness using past NDVI and precipitation data, compares this to MODIS measurements to detect anomalies, and calibrates results with Landsat images. Comparisons show it correlates well with other systems like PRODES. Terra-i is a tool for rapid habitat monitoring at continental to regional scales to inform conservation policy.
Apoyo en la toma de decisiones en agricultura a través de las Mesas Técnicas ...
Terra-i, An eye on Habitat Change
1. Near Real Time Monitoring of Habitat Change Using
Neural Network and a MODIS data
Louis Reymondin – Alejandro Coca – Andy Jarvis –
Karolina Argote - Jerry Touval – Andres Perez-Uribe –
Mark Mulligan
2. What is Terra-
Terra-i is a system of habitat changes monitoring that uses different
mathematical models that combine vegetation data (MODIS NDVI) and
precipitation data (TRMM) to detect deviations from the natural cycle of
the vegetation over time and thus antrophogenic impacts on natural
ecosystems.
It has maps of habitat loss every 16 days at the continental
level with 250 meters of spatial resolution.
3. Terra- goals
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.
4. Terra- approach
The intensity of vegetation greenness is a natural cycle that depends on climatic
factors (precipitation, temperature), site variables (type of vegetation, soil
characteristics) and disturbances (natural or anthropogenic).
Terra-i is a model to predict the evolution of vegetation greenness intensity, based on measures
of vegetation behavior in time and current weather measurements to detect significant habitat
changes.
.
5. Inputs data
1. Vegetation Index (MOD13Q1 MODIS Product , 16 days, 250m)
Normalized difference vegetation index (NDVI) represents the amount and
vigor of vegetation. In each area the values are closely related to vegetation
type and climatic conditions as well as the predominant land use pattern.
6. Tiles MODIS level analysis
Processed Terra-i data Incoming Test Tiles / Terra-i
This gives us greater automation of the process, synchronizing the stages
download, pre-processing of MODIS data, Terra-i processing load and soon final
results in the map server and FTP.
7. Input data
2. Precipitation Data of the Tropical Rainfall Measuring Mission(3hours, 28km)
TRMM is led by NASA and the Japan Aerospace Exploration Agency
(JAXA). It monitors and studies tropical and subtropical
rainfall, between 35 º N and 35 º S. It was released on November
27th, 1997 from Japan.
8. Research methodology overview
The methodology can be split into two main steps:
The training step (using data from 2000 to 2004)
• Models are trained in order to find the relationship
1. between recent precipitation and the changes in the
color of the vegetation (for different vegetation types)
The detection step (using data from 2004 to present)
• The trained models output are compared with the
2. satellite measurements in order to detect anomalies in
the vegetation state.
9. Model training
NDVI and QA MODIS data MOD13Q1, Precipitation
(TRMM 3b42)
(2000-2004)
To reduce the noise present
Time series gap-filling and in the data
smoothing
To reduce processing (clouds, atmospheric
duration, the NDVI time series Clustering variations, shadows…)
K-Means
with the same trends during the
years are grouped together Random pixels sampling for
each cluster
Neural network training
Original NDVI data Cleaned NDVI data
10. Anomaly detection
NDVI and QA MODIS data
MOD13Q1, Precipitation (TRMM)
(2004-2011)
Time series gap-filling and
smoothing
NDVI Prediction
from 2004 to 2011
Calibration using habitat Difference between the NDVI sensor
changes maps generated with maps
measurement and the NDVI predicted
Landsat satellite images (30m) of change
by the neural network
probabilities
NDVI increase
Rules
NDVI decrease
Vegetation changes Clasification of (anthropogenic)
maps change Results
Floods
Drought
11. Methodology – Change detection
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)
OUTPUT: 16 day predicted NDVI
Prediction
Multilayer perceptron
Bayesian Neural Network (BNN)
Model trainning and noise approximation change
Scaled Conjugate Gradient (SCG)
Gaussian noise
Input automatic selection
Automatic relevance determination (ARD)
12. Calibration with Landsat Images
2004
2009
As Terra-i generates maps of conversion
probabilities, we use Landsat images in
order to calibrate the results and select the
most appropriate probability threshold for
each cluster to generate binary
changed/unchanged maps.
13. Terra-i results comparation
with local models
Terra-i results were compared with deforestation data produced by the National Institute
for Space Research Instituto Nacional de Pesquisas Espaciais (INPE) from 2004 to 2009
through monitoring systems as PRODES and DETER.
PRODES
The Project of estimation of deforestation in the Brazilian Amazon (PRODES) generated
estimations from 2003 using a digital classification system with Landsat images (30m).
DETER
DETER is a near real time deforestation detection system. It publishes fortnightly
deforestation alerts for the Brazilian Amazon using MODIS images (500m).
The comparison shows a high correlation between Terra-i and PRODES
systems.
14. Comparison with PRODES
Comparison with PRODES
% of matching detections
% of PRODES detection
within a MODIS pixels
22. Road impact assessment
The Trans-Chaco Highway, Paraguay
Conclusions
• Very high levels of deforestation pre- and post- road construction
• But > 300% increase in deforestation rates since road finished, with a footprint that
likely goes beyond 50km buffer
Road: Trans-Chaco Highway
Project period: 2002-2006
Average pre-road deforestation rate: 23,000
Average post-road deforestation rate: 97,000 (+319%)
Year of peak deforestation: 2010
Footprint (modal deforestation distance): 30-40km
23. Improve more and more our system by developing methodologies for
analyzing the information generated.
25. Future Deforestation Scenarios
BR-364 Road, Brasil
PROOF OF CONCEPT
Base map Potential deforestation at T=0 Potential deforestation at T=150
Predicted deforestation Actual deforestation (Terra-i)
26. Integration Terra-i with others
Policy Support Systems
• Terra-i can also be used within the WaterWorld and Co$ting Nature Policy Support Systems to
understand the impact of recent land cover change on hydrology and the production and
delivery of ecosystem services.
• Data: http://geodata.policysupport.org/
Water flows Erosion
28. &
“The best way improve a system is to get people to use it”
Dr. Mulligan (Kings College of London)
29. Conclusions
Terra-i is:
A mapping and monitoring system for rapid assessment of land cover conversion at a medium scale
(250m).
A tool for monitoring conversion of habitat at continental, national and regional level in close to real
time.
A tool for understanding the effectiveness of protected areas and other conservation measures in
stabilizing or reducing land cover conversion.
A spatial support system for decision making in public policy and private development initiatives.
Through its linkage with WaterWorld and Co$ting Nature, a system for understanding the likely impacts
of near real-time land cover change on a wide range of ecosystem services.
Terra-i isn’t:
X Detailed monitoring tool in local level. For this it requires second-level monitoring (with high
resolution images) and third level (field data).
X A system to monitor degradation.
La implementación del sistema para el conjunto de datos de América Latina es un gran reto desde la perspectiva informática, trabajar con datos de 250 metros de resolución significa que el grid analizado representa más de mil millones de valores individuales para cada periodo de tiempo (cada 16 días). Esto implica que más de 26 mil millones de valores deben ser procesados por año. Es por esto que se utilizan tecnologías de tipo data mining y programación distribuida, que permiten analizar una gran cantidad de datos en un menor tiempo. Terra-I se corre en super computadoras, dotadas con 8 procesadores.
En Colombia las causas de pérdida de hábitat varían en cada región. En la región Andina la pérdida de bosques se asocia principalmente a la expansión de la frontera agrícola, el desarrollo de nueva infraestructura e incendios forestales. Mientras que en la Amazonia y el Pacífico la principal causa es la explotación maderera.
En Colombia las causas de pérdida de hábitat varían en cada región. En la región Andina la pérdida de bosques se asocia principalmente a la expansión de la frontera agrícola, el desarrollo de nueva infraestructura e incendios forestales. Mientras que en la Amazonia y el Pacífico la principal causa es la explotación maderera.
WaterWorld was used to calculate the hydrological baseline and terra-i chosen as the deforestation scenario to run the alternative. The images show the change in flows (left) and sediment from the baseline to the alternative. Flows increased below of reduced evapo-transpiration. Erosion increased in the deforested areas and sedimentation increased in the river draining these areas.