An Interview with Gilberto Camara, Head of Brazil's INPE

Hello reader!

It follows one an interview with Gilberto Camara, Head of Brazil's Space Agency INPE, on the system of monitoringof the deforestation of the Amazonian region, published on the day (08/02) in the website “www.mongabay.com”.

Duda Falcão

Monitoring Deforestation: an Interview

with Gilberto Camara, Head of Brazil's

Space Agency INPE

Rhett A. Butler

mongabay.com

February 08, 2011

DEGRAD, INPE's system for measuring Forest

degradation in the Amazon rainforest

Perhaps unsurprisingly, the world's best deforestation tracking system is found in the country with the most rainforest: Brazil.

Following international outcry over immense forest loss in the 1980s, Brazil in the 1990s set in motion a plan to develop a satellite-based system for tracking changes in forest cover. In 2003 Brazil made the system available to the world via its web site, providing transparency on an issue that was until then seen as a badge of shame by some.

Since then Brazil has become recognized as the standard-bearer for deforestation tracking and reporting—no other country offers the kind of data Brazil provides.

Brazil has two systems for tracking deforestation: PRODES (Program to Calculate Deforestation in the Amazon) and DETER (Real-time Detection of Deforestation), which allow it to rapidly identify where deforestation is occurring. PRODES, which has a sensitivity of 6.5 hectares, provides Brazil’s annual deforestation estimates (measured each August) while DETER, which has a coarser resolution of 25 ha, is a year-round alert system that updates IBAMA, Brazil’s environmental protection agency, every two weeks. This gives authorities the technical capacity—although not necessarily the political will—to combat deforestation as it occurs.

Brazil's deforestation monitoring system was developed by INPE, the National Institute for Space Research, the country's equivalent of NASA, under its Earth Observation program. INPE also runs other programs focused on space science, weather forecasting, climate change modeling, and space engineering. Brazil is currently build four earth observation satellites, including two in cooperation with China.

Gilberto Câmara

Space engineer Gilberto Camara has overseen much of INPE's earth sensing work, first as head of INPE's Image Processing Division, then as head of INPE's Earth Observation, and since 2005 as director of INPE. During his watch, INPE has released several new exciting capabilities, including DETER, a program to monitor ecosystems outside the Amazon, tools to track forest degradation due to logging and fire, and a way to measure emissions from deforestation. INPE recently announced a LIDAR (LIght Detection And Ranging)-based system that will provide more precision, sensing even through smoke and cloud cover.

In a January interview with mongabay, Dr. Camara explained how he went from a boy in northeastern Brazil to director of INPE and discussed INPE's deforestation monitoring and other projects.

AN INTERVIEW WITH INPE'S GILBERTO CAMARA

mongabay: How did you become interested in space research?

Gilberto Camara: In 1969, when I was 13 years old, two events shaped my future: the Apollo 11 landing on the moon and the movie "2001: A Space Odyssey". I remember coming home bowled over after watching "2001". At the time, I lived in Fortaleza, a resort town on the hot and dry Brazilian Northeast. I decided I would study to become a space engineer.

mongabay: What was your path to becoming director of INPE?

Gilberto Camara: After I decided I would be a space engineer, I had to study hard to be accepted at the Aeronautics Technology Institute (ITA), which is Brazil's top school for aerospace engineers, located near São Paulo, in the Southeast. I managed to get accepted at ITA, graduating with a BS in Electrical Engineering in 1979. Then, I joined INPE in 1980 as an engineer, and earned an MsC and a PhD in Computer Science at INPE. I then decided to focus on Geoinformatics and Image Processing as my main research areas. From 1991 to 1996, I was head of INPE's Image Processing Division. From 2011 to 2005, I was head of INPE's Earth Observation area. In 2005, I was appointed director of INPE by a search committee for a four year mandate. In 2009, I was selected (again by a search committee) to serve a second four-year mandate.

mongabay: What are some highlights of INPE's main programs?

Detection of forest degradation in Marcelandia,
Mato Grosso, Brazil using CBERS-2B HRC.
Courtesy of INPE.

Gilberto Camara: INPE's main areas are Space Science, Earth Observation, Weather Forecasts, Earth System Science, and Space Engineering. In Space Science, we are developing a Space Weather program, that aims to provide information about conditions in near-Earth space within the magnetosphere and the ionosphere. In Earth Observation, we focus on land use monitoring, especially deforestation in Amazonia and sugarcane expansion in the central part of Brazil. As for Weather, we are the main center for Numerical Weather Forecasting in Brazil, using a CRAY supercomputer (currently ranked 29th in the world's top supercomputers). Our Earth System Science center focuses on forecasts on the impact of Climate Change in Brazil. In Space Engineering, we are currently building four Earth Observation satellites (two in cooperation with China).

mongabay: What spurred Brazil to develop its Amazon deforestation monitoring system?

Gilberto Camara: It started with the need to respond to international criticism, especially in the years preceding the 1992 Earth Summit in Rio. Then, the system started to be used as a tool for policy-making, especially after all data was released on the web (after 2003).

mongabay: How does the monitoring system work?

Annual deforestation data from INPE's PRODES
system. Click image to enlarge.

Gilberto Camara: We currently provide three different types of information:

(a) Fires: produced daily from data provided by MODIS, GOES and NOAA satellites. An automatic algorithm detects areas likely to be on fire and flags fire occurrences which are then put on the web.

(b) Near-real time deforestation (DETER): produced every 15 days from MODIS and AWIFS-ResourceSat data. Images are compared with previous maps and interpreted by analyst. Includes assessment of clear-cut areas and areas being deforested.

(c) Detailed estimates of clear-cut areas (PRODES): produced yearly using LANDSAT, CBERS, and DMC data. Images are with previous maps and interpreted by analyst. Provides an indication of how much clear cuts have happened during a 12 month period from 1st August to 31 June of the following year.

DETER is used for supporting the law enforcement actions, since data is provided rapidly. PRODES is used for carbon accounting and year to year comparison.



INPE's PRODES system

Note: Both for PRODES and DETER, the initial map is created by automated classification, and post-processed with careful validation by an interpreter. Change maps are obtained by visual interpretation. We use visual interpreters, rather than automated techniques, because of the accuracy needed for carbon accounting versus the accuracy provided by automated image interpretation methods. For carbon accounting, we need to provide rates of change: how much change has happened from one year to the next. The rates of change are usually quite small.

To take an example: consider the 150 images used as part of the 2008-2009 assessment of deforested area in Amazonia. The rates of change for each image vary from 4 percent to zero. Here's the problem: even in an area with intense change as Amazonia, the biggest change rate is under 4 percent. However, it is known that the best automated image interpretation methods deliver about 90 percent accuracy. The conclusion follows: no automated interpretation method can deliver the quality necessary to monitor rates of change for carbon accounting. Hence, a good deforestation monitoring system needs to use skilled interpreters to visually detect change.

mongabay: Can your near-real time alert system (DETER) inform law enforcement on a timely enough basis to take action against illegal deforesters?

Gilberto Camara: Indeed, provision of information for law enforcement purposes is the main purpose of DETER. In 2009, 50 percent of policing operations took place in 2 percent of the forest area.

mongabay: Was there any political controversy in building such a powerful tool and making it publicly available?

INPE's DETER system

Gilberto Camara: We had a lot of support from the Federal Government, especially from the Ministry of Environment and the Ministry for Science and Technology. The support of former Environment ministers Marina Silva and Carlos Minc and of current minister Izabella Teixeira, and of former Science Minister Sergio Rezende was critical to ensure that data from PRODES and DETER maps were made available openly. Also, former president Lula recognized the value of transparency for building international credibility.

We had tremendous support from the press, public opinion and the scientific community. Public opinion in Brazil is very concerned about environmental matters. As you know, a 2009 poll by Pew Research Center shows 90 percent of people in Brazil are worried about global warming, compared to 67 percent in India, 60 percent in Germany, 44 percent in the US and Russia, and just 30 percent in China. Thus, the Brazilian public was completely in favor of fighting deforestation.

We had many problems with some local authorities and some members of Congress, especially some state governors and their representatives who reacted badly when INPE published data about critical areas of deforestation on their jurisdictions. Some governors made strong pressure on president Lula, claiming INPE was misleading the public and exaggerating the amount of deforestation in their states. However, the Federal Government backed us, recognizing that INPE brings value to the Brazilian proposal of becoming a developed nation while respecting the environment.

mongabay: INPE recently extended monitoring outside the Amazon forest ecosystem. Can you elaborate on what INPE is now monitoring?

Gilberto Camara: We are monitoring the expansion of sugarcane in Brazil, since we want to provide answers to the question whether biofuels promote deforestation and/or food insecurity. Thus, we are monitoring what are the areas of expansion of sugarcane farms and whether they are displacing other crops or causing direct and indirect deforestation. So far, we have seen that sugarcane is expanding at the expense of extensive cattle farming. Cattle raising in Brazil is very inefficient. We have 200 million heads of cattle using 200 million hectares, a very low productivity. The efficiency of cattle raising in Brazil is one of our major environmental problems.



Before and after picture of a region affected by deadly land slides
last month. INPE has partnered with Google Earth to develop applications.

mongabay: Given that Brazil is now recognized as a leader in deforestation monitoring, is INPE helping other countries develop similar tools?

Gilberto Camara: During COP-15 in Copenhagen, INPE signed an agreement with FAO to support other tropical countries in building their monitoring systems, with support from the UN-REDD program. We established a new Regional Center in Amazonia, which will support FAO and provide training and all support needed for the interested countries. Our agreement with FAO includes countries in Latin America and in the Congo Basin. INPE will transfer our TerraAmazon software, which is the system that supports PRODES and DETER, at no charge.

mongabay: Deforestation monitoring has traditionally been limited by cloud cover. Can you tell us how INDICAR may help extend your capabilities?

Gilberto Camara: I'm not completely familiar with the INDICAR system that is being established by IBAMA. We know the main limitation is the access to real-time radar images from the Japanese ALOS satellite. We are currently negotiating with the Japanese Space Agency (JAXA) to receive ALOS radar images in real-time in Brazil, but no final agreement has been reached.

mongabay: How will your soon-to-be launched satellite help research or monitoring efforts?

INPE's mapping of the Atlantic forest or Mata Atlantica ecosystem

Gilberto Camara: CBERS-3 combines four optical cameras: (a) a CCD camera with 20 meter resolution and 120 km swath; (b) a PANMUX camera with 5 meter resolution and 60 km swath; (c) a AWFI (Advanced Wide Field Imager) with 70 meter resolution and 860 kw swath; (d) an InfraRed camera with 40 meter resolution and 120 Km swath with bands on short-wave and thermal infrared. Thus CBERS-3 will combine the capability of making rapid assessments with the AWFI camera, detailed assessments with the CCD camera, and verifying these assessments using the PANMUX camera. This will allow us to improve significantly our DETER system and also to have a better assessment of its accuracy.

CBERS-3 is thus capable of providing full support to the MRV (monitoring, reporting and verification) requirements for REDD.

mongabay: Looking toward the future, what do you see as the biggest opportunity in earth monitoring? What is the biggest challenge to realizing this opportunity?

Gilberto Camara: Currently, operational systems such as PRODES and DETER are limited to providing information about land cover change. We need much more. We need information systems that give us information about land use and land function. Now that Brazil is succeeding in its commitments to reduce deforestation, we should focus on how we use our land. Efficient and sustainable production of food requires qualified information. When we consider the whole tropical belt, Africa and Latin America need to have good monitoring systems to increase their agricultural productivity while not endangering the environment.

Thus, the evolution of operational land cover change systems (PRODES and DETER) to operational land use monitoring systems in one of the major challenges in Earth Observation for the coming decade. This will require the use of multiple satellites, whose combined information should provide global medium-resolution data every 2 to 4 days. We have learned a lot from MODIS, whose data is changing remote sensing worldwide. We now know we need a LANDSAT-class image every day. Exploring multi-temporal data obtained by combining different medium-resolution satellites is the biggest challenge faced by the Land Observation community.

INPE's site in English

Source: WebSite www.mongabay.com