Supplementary material to “Reply to Comment on ‘How Nature Foiled the 2006 Hurricane Forecasts’”

William K. M. Lau and Kyu-Myong Kim, Laboratory for Atmospheres, NASA Goddard Space Flight Center, Greenbelt, Maryland, E-mail: William.K.Lau@nasa.gov

Citation:
Lau, W. K. M., and K.-M. Kim (2007), Reply to comment on “How nature foiled the 2006 hurricane forecasts,” Eos Trans. AGU, 88(26), 271. [Full Article (pdf)]

We appreciate A. T. Evan’s comment [Evan, 2007] on our article, and we welcome the opportunity to clarify issues he raised. Evan has two main objections:

  1. The 2006 sea surface temperature (SST) in the Atlantic was actually above normal compared with the long-term climatology, and
  2. the dust loading in 2006 was not substantially greater than in 2005 to cause the large change in sea surface temperature.

First, the main focus of Lau and Kim [2007] was on the abrupt cooling of the Atlantic SST from 2005 to 2006, not the long-term variation. The fact that 2006 is a climatologically warm year is not germane to our study. In our article we focused on possible causes of the widespread abrupt SST drop in the Atlantic, and changes in large-scale conditions associated with suppressed hurricane activity in 2006 relative to 2005. Second, we suggested that dust may have been important in triggering a series of rapid feedback processes in the ocean-atmosphere system in the Atlantic in 2006, resulting in conditions that were unfavorable for hurricane formation. However, we never said or implied that dust radiative forcing alone was enough to explain the large SST difference between the two years. The real question is, How much additional dust radiative forcing is sufficient to trigger the feedback processes? This is not a simple question to answer because it depends on the timing of the aerosol forcing and the preexisting state of the large-scale coupled atmosphere-land-ocean system.

Evan used an aerosol index that is derived from area mean dust coverage [Evan et al., 2006]. This index is not a measure of the actual amount of dust aerosol present in the atmosphere. Further, major cooling of the Caribbean region (40°–70°W, 15°–30°N) began almost contemporaneously with a major dust outbreak in June [see Lau and Kim, 2007, Figure 3]. Therefore the aerosol optical thickness (AOD) difference in June, not July, should be examined.

Figure 1 shows the 2006–2005 June difference map of AOD from Moderate Resolution Imaging Spectroradiometer (MODIS) Collection-5 data. The Collection-5 data represent the most recent upgrade to the MODIS, and are considered the state of the art for satellite estimate of cloud and aerosol properties. MODIS data are available from both Aqua and Terra satellites. Since both gave almost identical results, only data from Aqua are shown. The large increase in AOD over the Atlantic especially along the main transport region for Saharan dust (10°–20°N, 15°–65°W) is obvious. The mean AOD over this region is 0.36 in 2005 and 0.49 in 2006, which represents a 30% increase in dust loading relative to 2005. This roughly translates to a reduction of surface solar radiation flux of 8.0 watts per square meter [Schollaert and Merrill, 1998]. Over the Western Atlantic and Caribbean region (40°–70°W, 15°–30°N), the mean AOD is 0.18 in 2005 and 0.23 in 2006, corresponding to an increase of 28% and a reduction of surface solar radiation of 4.3 watts per square meter. This is much larger than the estimate of Evan [2007].

Fig. 1. Spatial distribution of June 2006 minus June 2005 aerosol optical thickness (AOD) over the Atlantic Ocean from the Moderate Resolution Imaging Spectroradiometer (MODIS).

Evan also computed the dust radiative forcing in terms of decrease of SST in °C per month using a fixed mixed layer depth of 25 meters in the Caribbean region. This is inappropriate. Large areas in the northern part of the subtropical western Atlantic are known to have a mixed layer depth of 20 meters or less [de Boyer Montégut et al., 2004]. The very shallow mixed layer is due to the strong summertime heating and the climatologically calm winds, making the SST in this region very sensitive to changes in surface energy fluxes.

Using the same calculation as Evan [2007], but with the AOD values from MODIS, the rate of cooling of the western Atlantic and Caribbean region is estimated to be in the range 0.1°–0.18°C per month for a range of mixed layer depths of 25–15 meters. This is substantially larger than the estimates by Evan [2007]. The rate of SST cooling of the Caribbean quoted by Lau and Kim was for the maximum cooling rate based on daily values. The observed monthly mean cooling rate over the region in June 2006 relative to June 2005 was about 0.53°C. Compared with the aforementioned dust-induced cooling rate, this means that the dust effect may have contributed up to 20–30% of the observed SST change. Clearly dust effects cannot be ignored.

By the latter part of June, and certainly by July–August 2006, the full atmospheric-ocean feedback triggered by the initial SST cooling by dust was likely to have already been established, resulting in a much stronger and widespread SST cooling. Our ongoing analyses of surface winds from the Tropical Rainfall Measuring Mission microwave instrument and from National Centers for Environment Prediction analysis have found, possibly as a part of the feedback processes, increased surface wind in June-July-August in 2006 relative to 2005 over the Caribbean region. The increased wind combined with the seasonal shoaling of the ocean mixed layer in the northern portion of the Caribbean in July–August would very likely lead to stronger evaporative cooling, and mixing of deeper and colder water, further amplifying the initial SST cooling. As suggested by Lau and Kim, the feedback also may involve an altered atmospheric Walker-type circulation with relative sinking motion over the western Atlantic/Caribbean region. Both the cooler SST and the enhanced subsidence could suppress hurricanes.

In summary, the considerations reported in this note further strengthen our argument that dust-induced feedback processes may have been instrumental in preconditioning the Atlantic SST and/or in amplifying an existing large-scale circulation resulting in suppressed hurricane activity in 2006 relative to 2005. Alternately, the lack of dust-induced feedback processes may have played a role in enhancing hurricane activity in 2005 relative to 2006.

References