Monitoring Deforestation with Remote Sensing
Area of Interest
The Amazon Rainforest is located in Brazil, Peru, Colombia, Venezuela, Ecuador, Bolivia, Guyana, Suriname and French Guiana. According to World Wildlife Fund, WWF, the Amazon covers 2.6 million square miles in the Amazon basin. It covers about 40% of South America, containing 1.4 billion acres of forest and 4,100 miles of rivers. It contains 90-140 billion metric tons of carbon, which helps stabilize our climate but could also have catastrophic consequences if released due to deforestation.
​
(Amazon. (2018).https://www.worldwildlife.org/places/amazon)
Data Acquisition
Remote sensing is instrumental in our ability to monitor deforestation in the Amazon Rainforest. Given the vast size of the Amazon, challenging terrain features and duration of time, it is not feasible that similar monitoring could be accomplished from ground level. Landsat satellites have been recording remotely sensed land data since 1972. Most Landsat images are available to the public, at no cost, to be analyzed and processed for interpretation.
​
Multi-temporal data is valuable for change detection especially when analyzed over a long period of time. It allows users to distinguish between human impact and seasonal changes by providing images pre and post disturbance.
Data Analysis
When we visually compare two images we are able to detect changes from one image to the other. Remote Sensing detects change in land cover by measuring the amount of electromagnetic energy that an area is emitting or reflecting. This allows us to observe an area of interest in more than just the visible sense. Remote sensing allows us to observe the visible, invisible, past and future.
The image above compares four satellite images of Rondonia, Brazil from 1975, 1986, 1992 and 2001.
In 1975, the Cuiaba-Port Velho highway is visible but Rondonia was virtually untouched tropical forests. In 1986, the fishbone pattern is visible and has started cutting paths deeper into the Amazon.
In 2001, the size and intensity of the fishbone pattern has exacerbated. This deforestation was caused by farmers and logging opeations clearing land along roads and gradually spreading further into the forest.
​
The near-infrared spectral band is a powerful tool when observing vegetation. The healthier the vegetation the brighter red it will reflect.
EAS used Landsat-5 data from1985 and compared it to 2016 Sentinel-2 data to depict deforestation near the northwestern Brazilian city of Rio Branco. EAS used false color images, which uncover an abundance of information by using the near-infrared spectral band.