Wildfires. Credit: lassedesignen

Wildfires in Chile in January 2017

25 January 2017 19:00 UTC

Wildfires. Credit: lassedesignen
Wildfires. Credit: lassedesignen

During the southern hemisphere summer 2017, central and southern parts of Chile were stricken by the most devastating wildfires in the country's history.

Last Updated

11 November 2022

Published on

25 January 2017

By Aleksandra Tusinska (HSPAF) and José Prieto (EUMETSAT)

Prolonged drought afflicted the unique Mediterranean vegetation in the area, while extremely high summer temperatures, low humidity and strong winds led to ideal conditions for the fire to spread and to last for a long time. Between December and January 11 people were killed and more than 400, 000 acres of forest were destroyed.

The most spectacular and the most disastrous were the wildfires in the last third of January.

Polar and geostationary instruments had a view on the fires from their orbits. The former, like MODIS or VIIRS, provide less frequent, but more localised measurements, since pixel size is smaller than for the geostationary imagers on board GOES or Meteosat.

In the case of Meteosat-10, Chile is close to the image boundary. Owing to that, the pixels dimensions are counted in tens of kilometres and very extended in surface. However, in spite of this, some big fires were captured by the 3.9µm channel of SEVIRI.

Another factor affecting detection is the wavelength and width of the channel. The closer to the absorption bands near 4.2µm of the central wavelength, the higher the absorption by CO2, and the lower the brightness temperatures registered by the sensor.

Lower wavelengths are increasingly absorbed by water vapour. In addition the Sun is adding a reflective component to the signal, masking the ground temperature distribution.

Figure 1 (right panel) shows a comparison of the images from the three sensors, combined in an RGB composite from 25 January. VIIRS fires show in the red, MODIS fires are green and the GOES imager fires appear in blue.

Fires feature at low level of colour in the individual images, so the interpretation can be summarised by saying that green patches are parts of a GOES pixel affected by fires, red patches are fires seen by both MODIS and GOES, and black patches are those where the most precise location (375m) by VIIRS confirms the core of the fire. Still some fires show in MODIS and not in the VIIRS I4 channel (red patches) due to differences in channel width and wavelength.

Figure 1 (left panel) shows Meteosat instead of VIIRS with red large pixels when some fire is nearby. The poor rectification so close to the boundary of the image accounts for a shift to the west in the apparent location of the fires.

Image comparison

RGB composite of Meteosat, MODIS, GOES imagery compare1

Figure 1: Image composition of VIIRS, MODIS and GOES imager in the area worst hit by fires, 25 Jan, 19:00 UTC +/- 5 minutes. VIIRS 3.74µm, MODIS 3.99µm, GOES and Meteosat 3.9µm. Fire pixels seen by MSG are marked with letters A & B.

Figure 2 is a composite of MODIS channels at 3.78, 3.97 and 3.99µm, very similar in its fire detection capability. It takes on a brown hue as a result of higher temperatures in the shorter wavelengths. It allows the tracing of smoke over the ocean, north west of some of the big fires.

 MODIS channels at 3.78, 3.97 and 3.99 µm
Figure 2: MODIS channels at 3.78, 3.97 and 3.99µm

Additional content

Chile battles devastating wildfires: 'We have never seen anything on this scale' (The Guardian)
Chile wildfires: At least 11 die in catastrophe (CNN)
Chile wildfires: Drone footage reveals devastation (BBC News)