Amazonia - 1

Uses and Applications

Brazil is a country continent-wide with more than 8.5 million square kilometers. Its territory comprises different biomes, a diversified agricultural system and complex hydrological, energetic, geological and topographic systems. In addition, Brazil has an ocean coast with more than 7,000 km, and a network of complex cities that houses more than 80% of its population. There is an ongoing need to monitor all of these goals, especially two of them: biomes and agriculture.

The natural vegetation of Brazil is very diverse. Only two biomes - Tropical and Cerrado Forests - cover more than half of the country's territory. Some of them are constantly under pressure for anthropogenic activities, such as logging, agriculture, pasture management, etc. To follow these activities it is necessary to monitor their dynamics.

One of the main environmental problems in Brazil regarding biomes is deforestation, especially in the Amazon region. The government and other agencies have made great efforts to monitor this anthropogenic phenomenon. One such initiative is DETER (Deforestation Detection in Real Time). The DETER system is a contribution of INPE to the action plan of the Brazilian Ministry of Science, Technology, Innovation and Communications through the Permanent Interministerial Working Group (GTPI) to reduce deforestation rates in the Legal Amazon. This system, in operation since 2004, is an important alert tool for monitoring and controlling deforestation, especially for the Brazilian Amazon. DETER uses data with a spatial resolution of 250 m from MODIS instruments - Spectrum Radiometer of Moderate Image Resolution, aboard the Terra and Aqua satellites of the United States.

Since MODIS images have spatial resolution of 250 m, the DETER system can only identify areas of deforestation above 25 ha. In addition, cloud cover may also restrict the system to identify deforestation. There is evidence that the area of deforestation lots is declining; thus, to monitor these areas, remote sensing systems with higher spatial resolution are required. Another operational program dedicated to monitoring deforestation in the Amazon region is PRODES (Measurement of Deforestation by Remote Sensing). This system is based on Landsat type data and aims to measure the deforested area using satellite data. As it involves the area measurement, it uses data with better spatial resolution (Landsat, CBERS, DMC, etc.).

Both systems - DETER and PRODES - are hampered by the high frequency of clouds over the Amazon region. Thus, to ensure that deforestation monitoring and measurement is feasible, it is necessary to have remote sensing data at spatial and temporal resolutions compatible with these phenomena.

Agriculture is the second system that requires monitoring by high intensity remote sensing. The economic chain of agriculture plays a very important role in Brazil, since its gross domestic product is estimated at around R$ 200 billion. The intense and diversified agriculture gives Brazil an important role as a food supplier in the world. Some of the main products of this sector are: coffee, soy, sugar cane, corn, cotton and citrus, as well as cattle.

The main agricultural calendar of Brazil coincides with the rainy season, therefore with high probability of cloud cover. Annual harvests (corn, soybeans, etc.) have a fairly short period from sowing to harvesting; thus requiring multiple images to cover all stages of the growing cycle. Both factors - station and limited growth cycle - confer less than the probability of acquiring optical remote sensing data obtained from satellites. Low-frequency satellites have limited use for detailed analysis of agriculture; therefore, higher temporal resolution is required for remote sensing systems to be more useful as an operational base for agricultural applications.

Both deforestation and agriculture should be evaluated quantitatively after their identification. To perform process measurements with adequate accuracy, there must be compatibility between field size and spatial resolution of remote sensing data. In general, deforestation processes are large-scale phenomena, however, they start small in size and develop rapidly. At present, detection of deforestation onset is done with low spatial resolution (typically 250m), while the measurement is done with data between 20 and 60m, compatible with a 1: 250,000 scale mapping. As mentioned earlier, agriculture in Brazil is very diverse, going from small crops, using traditional practices, to important crops, using high technology, which in general leads to large fields of crops. Thus, the system to be designed must have a spatial resolution close to the Landsat, CBERS, SPOT, etc. systems.

Two additional aspects to be considered for these applications are the spectral and radiometric resolutions. Optical systems designed for vegetation analysis need to have spectral bands in the main spectral regions: visible, NIR (near infrared) and, ideally, in the SWIR (short wave infrared) region. In visible, it is expected to have two bands (blue and red) or three bands (blue, green and red); a band on both NIR and SWIR. Each of these spectral regions responds to specific and important vegetation phenomena: chemical, structural and humidity. Radiometric resolution should be sufficient to help users discriminate subtle differences in the emerging brightness of specific targets.

Since Brazil has a long history and commitments to keep track of its vast forested areas and large agricultural activities, there is a need for continued provision of remote sensing data. In general, the scope of satellite data applications is broader than its primary mission. It is expected that data from Amazon 1 may be useful for other applications, such as monitoring of coastal zones, water reservoirs, natural and cultivated forests, disasters, etc.

INPE - Instituto Nacional de Pesquisas Espaciais