Using Meteosat channel 04 (IR3.9) to assess ground humidity

Assess ground humidity with Meteosat

4 January 2012 00:00 UTC

Using Meteosat channel 04 (IR3.9) to assess ground humidity
Using Meteosat channel 04 (IR3.9) to assess ground humidity

Using Meteosat channel 04 (IR3.9) to assess ground humidity in South America in January 2012.

Last Updated

13 July 2022

Published on

03 January 2012

By Anete S. Fernandes (INMET Brazil), Christian Garavaglia (SMN Argentina) and Jose Prieto (EUMETSAT)

The flow of water vapour in the atmosphere tends to balance the movement of liquid water over land toward the oceans. Thus, the flow of water vapour over land is occasionally called flying rivers because it compensates for the land water running down in the opposite direction towards the oceans, along the river beds. In the case of South America, trade winds feed oceanic humidity into the Amazon basin, where it precipitates and evaporates several times before it hits the Peruvian Andes mountains, supplying water to the sources of the main rivers. The flying rivers of water vapour then take a course to the east, towards the Southern Atlantic, where it is mainly converted into rain over Paraguay and southern Brazil.

The weather situation during the last two months in 2011 responded to a cyclonic circulation centred approximately two thousand km off the coast of Uruguay to the southeast. More to the north, the South Atlantic convergence zone (SACZ) transported humidity towards the southeast into the Atlantic, very actively this season, as shown in the longwave anomaly charts. Wind anomalies indicate humidity convergence at 20° south. Upwind, in western Brazil, severe episodes of flooding were frequent during this season. The Paraguay area between Asuncion and Iguazu is known for its frequent convective activity in summer, usually associated with heavy precipitation and lightning, for which Paraguay is a world record holder for several indices. Depending upon the precise latitudinal location of the SACZ, Argentina can also suffer drought conditions which can be directly correlated to flood situations in Southern Brazil , see 500hPa anomaly for Nov-Dec 2011 (Credit: NOAA/NCEP).

Emissive wet surfaces give a stronger night signal at IR3.9 (BT4) when compared to the drier surrounding regions, even assuming the same cooling rate. During the day, solar illumination generally over-compensates the reduced emission of dry surfaces, making the signal stronger than that from wet surfaces. In addition, the drier surface gets warmer in the course of the day. As a result, the BT4 day-night contrast is more pronounced for dry surfaces than for wet or flooded areas. This is confirmed by the temporal evolution of the IR3.9 brightness temperature (December 2011) for Asunción (typically dry, in red) and Iguazu (humid, in blue).

Therefore, imagery in the window channel IR3.9 (BT4) should help with the assessment of flooded or very wet areas on the ground under clear conditions. The slower cooling (at night) and warming (during day) for wet areas allows one to infer areas of possible flooding. At night, flooding appears darker in the conventional infrared IR3.9 channel, where higher brightness temperatures are blacker (see Figure 2).

Some comparisons were performed against ground truth for so called 'western locations' (Asuncion, San Estanislao, Paraguari) and 'eastern locations' (Iguazu, Saltos del Guaira and Guarani international airport), as well as two in Brazil (Passo Fundo and Palmeira das Missoes). The detailed analysis of precipitation measured by several ground-based rain gauges shows a similar amount of precipitation in the western areas of the image (close to Asuncion) as in eastern Paraguay and east of the Parana river (close to Iguazu), in contrast to the indications from the BT4 images from 3 December 2011 to 4 January 2012 (Figure 2).

In fact, the contrast in brightness temperatures at IR3.9 (BT4) between day and night is around 30 Kelvin for the chosen western locations, but only 15 Kelvin for the eastern locations, i.e., against the hypothesis of flooded areas in the west. The following figure shows the day (15 UTC) peaks versus the night dips in the evolution of BT4 values during December 2011 at San Estanislao and Passo Fundo.

It cannot be assumed, therefore, that the ground remained flooded in the areas mentioned above once the precipitation had moved away. However, there is an impact coming from ground humidity in the amplitude of the diurnal oscillation. For example, San Estanislao (western region) consistently has a smaller oscillation (i.e. wetter surface) than Passo Fundo (eastern region). Combined with the diurnal oscillation, there is a trend towards higher ground temperatures during the course of December (Figure 3).

Using Meteosat channel 04 (IR3.9) to assess ground humidity
Figure 1: Meteosat-9, RGB Composite Day Natural Colours (NIR1.6, VIS0.8, VIS0.6). Left: 1 December 2011, 15:00 UTC, Right: 3 January 2012, 15:00 UTC.
Using Meteosat channel 04 (IR3.9) to assess ground humidity
Figure 2: Meteosat-9, Channel 04 (IR3.9). Left: 3 December 2011, 08:00 UTC, Right: 4 January 2012, 08:00 UTC.
Using Meteosat channel 04 (IR3.9) to assess ground humidity
Figure 3: Meteosat-9, Channel 04 (IR3.9). Left: 3 December 2011, 08:00 UTC, Right: 4 January 2012, 08:00 UTC. Colour Scale