Skip to content

Data analysis · Hotspots in the Pantanal

Foco Atlas

Foco Atlas analyses hotspot records in the Pantanal and organises the results by period and municipality. From INPE data, the project builds an index to support territorial prioritisation and presents the analyses in a portal with a map, series and an executive reading.

Type
Academic project · Data Science / FIAP
Status
Analysis completed in a notebook
Scope
Pantanal, 2020 to 2025, AQUA_M-T satellite
Foco Atlas cover: dark map of the Pantanal with heat hotspots in red.

What was done

Six yearly BDQueimadas files were consolidated into a single dataset, the scope was validated, dates and missing values were handled, and hotspots were analysed by year, month, state and municipality. Then an index of its own was built to rank municipalities by thermal pressure.

There were 53,813 hotspots across 25 municipalities. August to October accounted for 66.7% of them, and four municipalities (Corumbá, Poconé, Barão de Melgaço and Cáceres) held 81.1%.

The project portal

The analysis output was organised into a portal with a map, series and an executive reading. Each tab shows one part of it. Captures of the project’s academic portal, which is not available online; map with data from OpenStreetMap and CARTO.

The project portal: scope indicators, a map of the Pantanal with the prioritised municipalities and the IPT ranking.
Portal overview: scope indicators, map and IPT ranking. Map with data from OpenStreetMap and CARTO. The project’s former name was removed from the introduction text of this capture.

Hotspots per year

  • 202022.116
  • 20218.258
  • 20221.637
  • 20236.580
  • 202414.498
  • 2025724
Public BDQueimadas/INPE data, as scoped in the notebook. Year-to-year variation is large, so one year’s volume is not a trend.

Thermal Pressure Index

The IPT folds five components into a 0 to 100 score per municipality: 40% hotspot volume, 25% summed radiative power (FRP), 15% hotspots in critical months, 10% average fire risk and 10% yearly recurrence. These weights are from the second version, because the first let municipalities with few records climb the ranking on high proportions alone.

  • Corumbá94,29
  • Poconé55,83
  • Barão de Melgaço50,43
  • Cáceres43,22
Top four of the IPT ranking, from the notebook output.

Limits

A hotspot is not a fire

Each record is a satellite detection. It is not a confirmed fire or a burned area.

The IPT does not forecast

It is an index for prioritising attention from the historical record. It does not anticipate future events or replace fieldwork.

A single satellite

The analysis uses only AQUA_M-T to keep the years comparable. Other BDQueimadas sensors were left out.

Technologies

Data analysis
  • Python
  • Pandas
  • NumPy
Visualisation and notebook
  • Matplotlib
  • Jupyter

Back to projectsLer esta página em português