Haze veil on the Pakistani coast

By HansPeter Roesli (EUMETSAT)

On 23–25 December 2010 an episode of easterly surface winds picked up dust resulting in haze over the lower Indus valley and the adjacent salt marshes called Rann of Kutch and spread it over the Arabian Sea along the Pakistan coast towards the Sea of Oman (Figure 1).

By mid-morning on 24 December the haze veil was moving westwards at a speed of 7-8m/s, derived from a couple of Meteosat-9 images. This value compares well with the 10–20kts computed from the wind field deduced from the ASCAT instrument onboard Metop (Figure 2).

The relative ASCAT wind field (black flow arrows in the same image) shows a sizeable wind convergence along the leading edge of the haze layer, giving rise to shallow convection and indicated by a string of small cumulus clouds well depicted by MODIS on Aqua (Figure 3).

One might now wonder about the constituents of the haze being driven across the sea. Remembering that extreme monsoon rains had inundated the lower Indus valley and the adjacent salt marshes called Rann of Kutch earlier in 2010, and that by December 2010 most water had evaporated, leaving behind areas with dried mud and salt, mineral dust would be considered as a likely candidate. However, this is not the case — the dust hardly shows up in the Dust RGB from Meteosat-9 (Figure 4).

Also, a sequence of Dust RGBs from Meteosat-8 taken during the night of 23–24 December does not reveal major dust outbreaks over the adjacent continental area.

On the other hand, the haze is easily detected in the solar channels (albeit in daylight only), in particular, in the morning when the Sun-satellite orientation gives rise to moderate forward scattering. From the mid-morning orbit of Metop (flying overhead the area, and thus less favourable to forward-scattering) the haze is more transparent. A mosaic of four images compares the Natural Colour RGB from both satellites shortly after 05:00 UTC (about 10:00 local time), both on 24 and 25 December (Figure 6)

Looking more closely at the three SEVIRI solar channels , it is interesting to note the somewhat irregular behaviour of the reflectivity. VIS0.8 gives the strongest reflectivity values over the haze veil whilst VIS0.6 and (even more so) NIR1.6 reflect less in the early morning of 24 December. This higher value of VIS0.8 may point to a Mie-scattering regime, i.e scattering from particles with sizes similar to the wave lengths of the observing channels. However, particle sizes in the order of one micron are typical for suspended particles of mineral origin. Compare this to the smoke plumes from the forest fire near Haifa where the reflection strongly decreases with increasing wave length, suggesting a Rayleigh-scattering regime and due to much smaller smoke particles.

Thus, it appears that the haze consisted of many ingredients, but mostly non-mineral particles, yet of relatively large size when compared to smoke from hot wild fires. This analysis compares well with the conclusions of a paper on carbonaceous aerosols over the Indian Ocean (Novakov et al., 2000): "... that both biomass and fossil fuel combustion contribute to the carbonaceous particles, including light absorbing black carbon. However, the contribution of fossil fuel burning (from transportation, industrial and domestic sectors) is considerably larger and can be as much as 80% of the total." The same source in Table 2 gives a size less than 1.3 micron for the particles. Note however, that the measurements reported in the paper refer to flights around the southern tip of India.