Brazil: Fires Heat up After a Four-Years Decrease   (IFFN No. 14 – January 1996)

Fires Heat up After a Four-Years Decrease

(IFFN No. 14 – January 1996, p. 13-15)

After four years of decreases on the burnings index, Brazil had to deal again with high levels of fire activities during the 1995 dry season. According to the information obtained from the NOAA AVHRR satellite sensor the average total number of fires detected between 1991 and 1994 had dropped by 23% each year. During the five dry months of 1995 the total number of detected fires reached 367,000. That means a 70% increase, if compared with 1994’s numbers.

The distribution of the main fire concentrations also changed. During the last four years the fires had followed almost the same pattern throughout the dry season, in a high correlation with the dislocation of the Intertropical Convergence Zone (ITCZ) and the ocurrence of rainfall. The regions where the rainfall stops first, burning activities also start first. Whenever out of season rainfalls occur, there are no burnings. From 1991 to 1994, June and July were the months of medium fire concentrations in Southern Brazil, predominating in Paraná, Santa Catarina and Sao Paulo. From August on, the burnings rapidly progressed into Central Brazil, spreading smoke over Mato Grosso do Sul, Mato Grosso, Goiás, Tocantins and West Bahia. Then the major fire concentrations moved North, towards the Amazon region (Pará, Amazonia, Acre, Rondônia), and to the East (Maranhao, Piauí), ending October with the first burnings in Northeastern Brazil, at the semi‑desert region called The Drought Polygon (Bahía, Pernambuco, Rio Grande do Norte, Paraíba, Ceará).

This year, June and July started with the highest concentrations of fire points ever registered, and many of them were localized in the Amazon Region. These months’ indexes almost doubled, if compared with the average of previous records. In August, there was a 34% increase above the average of previous years. But in September and October the fire numbers kept the same level of previous years.

The main reason for this change of patterns and standards was economic. Since 1991, less investment money was available for farmers, cattle ranchers and agriculturists. They have reduced, therefore, their planted areas and did not open new farming lands. In 1995, the situation changed, with the success of the Plano Reál. Its economic measures drastically reduced inflation and stabilized the Brazilian economy. Farmers, cattle ranchers and agriculturists reinvested in plantations and pasture renovations. The abandoned areas were again prepared and fire was again used as the cheapest and easiest tool to control weeds and accelerate grass sprouting. New areas were also opened, on a smaller scale.

The climate also favoured the fires, especially in the Amazon Basin. This year, the dry season started first and lasted longer in the Amazon states. Consequently, agriculturists had more time to prepare their lands and set fire. At Rondônia State, for example, the satellites detected concentrations of fire from early June until late November! Usually, the burnings detected in that state last only from August to September/early October.

It is important to bear in mind that Brazilian fires are very different from those in the natural vegetation of the Mediterranean basin, from wildfires in California’s chaparral, and forest and tundra fires in Alaska. The untouched rainforests ‑ either at the Amazon Basin and at the Brazilian Atlantic Coast ‑ do not burn by accident, nor even if someone sets fire to them. They are too humid to burn by themselves. They only catch fire after a severe and long drought and if disturbed or cut down. Therefore, almost all fires detected on forested areas are associated with either deforestation or post-deforestation burning. When converting forest into other land-use systems, trees are first cut down. After several months of drying, trunks and branches are burned. It takes an average of eight years until all the wood is burned. So, fire is used as a tool to clean up the land, and the hotspots are not representing wildfires. Fire is also used to eliminate weeds, plagues and leftovers of plantations, to accelerate the pastures sprouting and even to make harvesting easier, as well as a fast fertilization process. And those are the causing that high number of fires detected each winter in Brazil. That’s why burnings cannot necessarily be used as a deforestation indicator: if it is true that the fire is commonly set on deforestated areas, it is also true that deforested areas are not always burned and it is especially true that fires are also set on a lot of other areas, not necessarily deforested, like Cerrados, grasslands, and traditional agricultural lands. More reliable deforestation indicators were developed in studies by the National Space Research Institute (INPE), with Landsat TM and Spot data (called PRODES). At the end of December 1995, INPE will release the new deforestation numbers, based on 1992 and 1994 satellite images. The 1995 numbers shall be available by June 1996.

Spaceborne Fire Monitoring

The first NOAA images used to detect fire points were treated at INPE, back in 1987. The fires detected by the satellite sensor were georeferenced by a computer software (determination of geographical coordinates). Since then, lists with fire coordinates were sent daily to the Brazilian Environmental Agency (IBAMA), to be checked/verified in the field. The federal agency was never able to truly use the monitoring data, due to alleged lack of financial support. But one of the state environmental agencies, the SMA of Sao Paulo, actually investigated the main fires pointed by INPE, and established a programme, called Mata Fogo, to control burnings.

In 1991, the Environmental Monitoring Center (NMA) and the non‑governmental organization Ecoforça joined INPE, in an effort to make the system more operational. They developed a software to translate the NOAA data, processed by INPE, into maps. The maps use the standard 1:250,000 grid over the Brazilian territory, in order to better classify and localize the fire concentrations. There are also numerical maps, which identify the number of fire points in each square. Those maps have been released to the press, through the Agencia Estado News Wire Service, together with an analysis of what kind of vegetation is probably burning, where are the worst fire concentrations, fire spots that should be investigated by authorities, etc. Those features and maps have been published on a weekly basis, throughout the dry season. The maps are available at Internet, on Ecoforce’s site.

At the end of 1994, there were some interruptions, due to the NOAA‑11 failure. In 1995, the NOAA‑14 was picked as the substitute, because it passes over Brazil twice in the afternoon, when the major part of the fires are still going on. Unfortunately, the NOAA‑14 is almost 2 hours earlier than the NOAA‑11. This led to a solar reflexion problem: in mid-August the open cerrados (savanna-like vegetation) were so bright that they saturated the pixels. INPE was forced to adopt the NOAA‑12 images, with a big loss of data. NMA and Ecoforce decided to publicize two different maps for 1995 fires: one from June to 15 August, and the other from 15 August to October. The first based on NOAA‑14 and the second on NOAA‑12. Up to now, they are not comparable. The NOAA‑12 passes over Brazil at night, when only illegal, accidental and roadside fires are still lighted or continuing. There are 4 to 5 times less fires at night, as can be seen by NOAA‑12 and NOAA‑14.

The total of fire points from 15 August until October was then estimated (and not measured). These preliminary estimates were based on the tendency of the burning and on the differences observed between those two satellites. Other evaluations are on their way. INPE has recorded all NOAA‑14 images not used on the operational monitoring system and is trying to solve the problem with new software. Ecoforce got some images of other satellites (DMSP and GOES) and will compare them with NOAA data. The burnings are a complex and dynamic problem, but Brazil will find better technology to fully re‑establish its monitoring system.

From: Liana John
Agencia Estado News Wire Service


Evaristo E. de Miranda
Brazilian Environmental Monitoring Center (NMA)

Rua José Inocêncio de Campos, 148
CEP 13024-230-Cambui
BR – Campinas, Sao Paulo

Country Notes

Print Friendly, PDF & Email
WP-Backgrounds Lite by InoPlugs Web Design and Juwelier Schönmann 1010 Wien