Remote Sensing of the Rainforest

by

Obie Pepper

ES 771 Remote Sensing, Emporia State University


Enhanced ERS-1 SAR Image of Indonesian Rainforest on the Island of Kalimantan


image acquired from:
http://www.zi.biologie.uni-muenchen.de/institute/zoologie/Siegert/Sat/ERS1/Sirc.htm.

Abstract

The tropical rainforests of the world are an important and crucial part of our planets ecosystem. Currently the rainforests are being depleted at alarming rates. The causes for rainforest depletion are mainly due to human activities, such as ranching, agriculture, and urban development. These areas must be monitored in order to provide current and complete information in order to better understand and protect the rainforests.
This page gives a general summary of the importance of the rainforest, and the detrimental affects that it is undergoing. Two areas of rainforest investigation are covered, which briefly delve into the efforts made to improve rainforest monitoring. A program in Brazil and Indonesia are considered. Both of these programs are utilizing and testing the opportunities of remote sensing, especially in the regards to radar remote sensing.

Return to the Abstract.

Introduction

In today's world population growth has presented a major concern for many reasons. One of these concerns is the destruction of tropical rainforests. Tropical rainforests once covered 14% of the Earth's surface. Today the rainforests have dwindled to 6% of the Earth's surface, and estimates have been made that predict the total loss of rainforest in 40 years. One and a half acres of rainforest is lost to development and industry every second. An estimated 50,000 species of plants and animals become extinct every year due to deforestation by man. Some of these species may play important roles in medicinal purposes. The rainforests also play a major role in recycling the Earth's oxygen (http://www.rain-tree.com/facts.htm.). Remote sensing techniques are now being tested and utilized to help monitor and protect this vital resource.

One major area of investigation is the tropical rainforests of the Amazon Basin. This is by far the largest area of tropical rainforest in the world. The Amazon Basin holds over a billion acres of rainforest (http://www.rain-tree.com/facts.htm.). In the last 40 years the country of Brazil has undergone rapid population growth, which has led to economic expansion that has resulted in the destruction of vast amounts of rainforest. The rainforest in this region is cleared for large scale cattle ranches, logging companies, mining, and agriculture. Brazil set up many development programs for different regions of the country during the 1960's. Mining and ranching were encouraged in the north region of the Amazon Basin. The Trans Amazon Highway was constructed at this time, which crosses the Amazon Basin from east to west. These developments ushered in new settlers from the overcrowded northeast region. Areas of forest began to be cleared near the new roads and simple settlements sprang up (http://www.bnsc.gov.uk/index.cfm?pid=793.). Brazil has now implemented remote sensing to monitor the conditions of the rainforest.

This is a Landsat TM image of the typical fishbone pattern of forest clearing near the town of Porto Velho (areas that have been cleared are represented in light green or pink, while the dark green areas represent forest)


image acquired from: http://www.bnsc.gov.uk/index.cfm?pid=793.

Indonesia has 353 million acres of rainforest. 148 million acres are reserved for production purposes. This large production area is concessioned to different companies who have the right to harvest in their respective region. Each company, or concessionaire, is mandated by government law to harvest the forest in a sustainable fashion. These concessioned areas can range from 12,350 to 247,000 acres, which creates a problem for monitoring. (http://cgi.girs.wageningen-ur.nl/cgi/education/courses/k075704/1999/001.htm.) The Indonesian government has also turned to the capabilities of remote sensing to provide adequate monitoring of the rainforest.

Both the Indonesian and Amazonian rainforests are situated at tropical latitudes where there is extensive cloud cover much of the time. This means that standard air photos and other forms of remote sensing are not able to get a complete coverage for these areas. With the advent of radar and microwave remote sensing these areas can now be fully covered without the veil of cloud cover due to the fact that radar has the ability to penetrate cloud cover with no interference.

Return to the Introduction.


Application of Multitemporal ERS-1 Data and SIR-C/X-SAR Data

Test Site Area

click on image to see full version acquired from: http://ceos.cnes.fr:8100/cdrom-00b2/ceos1/casestud/sar/fig1.htm.



image acquired from: http://ceos.cnes.fr:8100/cdrom-00b2/ceos1/casestud/sar/rainfor.htm.

Multitemporal ERS-1 data has been tested in the Brazilian state of Acre, which is located in the Southwestern Amazon. The study was launched to investigate the usefulness in monitoring tropical rainforests, land use classification, and change detection. This project was set in motion as a part of the ESA's (European Space Agency) ERS-1 pilot project PP2-D3, which is funded by the German Space Agency. It is integrated through the National Brazilian Institute of Space Research (INPE). The primary purpose of the project was to develop and test methods to utilize ERS-1 data for forest/nonforest mapping, rainforest monitoring, and the possibility to differentiate land use intensity in the study area. The testing of numerous approaches for classifying ERS-1 SAR data was conducted. Evidence Based Interpretation of Satellite imagery (EBIS) was used as a valuable tool for the classification of forest/nonforest areas, and identifying land use diversification. Multitemporal ERS-1 SAR data for wet and dry seasons improved the differentiation of forest/nonforest areas (http://ceos.cnes.fr:8100/cdrom-00b2/ceos1/casestud/sar/rainfor.htm .).

The test site is in the southwestern portions of the Amazon Basin. This area is 200 km for the Bolivian and Peruvian borders, and belongs to the state of Brazil and Amazonas. The capital city of Acre, which is Rio Branco, and the town of Sena Madureira are included in the test area, along with BR-364 transamazonian highway. A TM floating scene from 1992 is used to show the study site, which utilizes TM band combinations R=5, G=4, B=3. The different rainforest formations are indicated in dark green, while areas influenced by human activity are indicated in red and bright green (Return to the see Test Site Area)..

ERS-1 SAR data in this area of the Amazon showed that it may be utilized with a high rate of success to monitor the rainforest and deforestation. The presence of large pasture areas can be easily interpreted. Multiseasonal data can be used to identify vegetation development in pastures, regrowth, and the differentiation of primary vs. secondary forests. Misclassifications still appear in areas of high topographic relief. First-hand visual interpretation of these areas can be done to eliminate this problem. The first year of deforestation cannot be detected in newly destroyed and cleared rainforest. L-band data improves the differentiation of rainforest vs. regeneration areas. Pasture and regeneration areas may be further distinguished by combining L-band/C-band or X-band evaluation of SIR-C/X-SAR data. This is especially true for cross-polarization states L-band and C-band HV-polarization (http://ceos.cnes.fr:8100/cdrom-00b2/ceos1/casestud/sar/rainfor.htm .).

Return to the Application of Multitemporal ERS-1 Data and SIR-C/X-SAR Data.


Indonesian Remote Sensing Summary

Tropenbos and Wageningen Agricultural University has conducted a joint research program since 1993 to solve the problem of cloud cover over the Indonesian rainforest by the advent of microwave remote sensing. Thus far their research has proved successful in identifying logging roads, canopy gaps and skid roads. Other possibilities for this type of remote sensing are being investigated and show promise for monitoring the Indonesian rainforest. The study is currently attempting to generate tree maps from radar data in order to estimate tree biomass. When a general algorithm is complete the biomass content of a tree may be predicted by the size of the canopy. New results may still emerge down the road using remote sensing in the Indonesian rainforest. Remote sensing tools can estimate forest condition, and with the comprehensive database from fieldwork further corrections can be made, and should provide the best system to monitor the Indonesian tropical rainforest (http://cgi.girs.wageningen-ur.nl/cgi/education/courses/k075704/1999/001.htm.).


References

Centre for Geoinformation, World Wide Web homepage URL:
http://cgi.girs.wageningen-ur.nl/cgi/education/courses/k075704/1999/001.htm. Retrieved on (12-11-02).

The Learning Zone, World Wide Web homepage URL:
http://www.bnsc.gov.uk/index.cfm?pid=793. Retrieved on (12-12-02).

Rainforest Facts, World Wide Web homepage URL:
http://www.rain-tree.com/facts.htm. Retrieved on (12-10-02).

Rainforest Monitoring in Acre, Brazil, World Wide Web homepage URL:
http://ceos.cnes.fr:8100/cdrom-00b2/ceos1/casestud/sar/rainfor.htm. Retrieved on (12-10-02).

SAR Image Processing, World Wide Web homepage URL:
http://www.zi.biologie.uni-muenchen.de/institute/zoologie/Siegert/Sat/ERS1/Sirc.htm. Retrieved on (12-11-02).

Return to the References.


This web page was created to fulfill the requirements for Remote Sensing at Emporia State University.


This page was created on 12-11-02.

Return to the top.