Good And Bad News For Masters And Romm

OVERVIEW

First, the good news: the claims about Sea Surface Temperature data for two of the areas reported in their recent posts about floods are correct. The bad news: two of them are not. In one case, December Sea Surface Temperatures were reported as January data, and there was a significant difference in Sea Surface Temperatures between those two months. For the second error, the difference between the actual data and what was claimed is so great that it’s difficult to determine the source of the Masters’s mistake.

This post is not about the potential for increased rainfall resulting from rises in Sea Surface Temperature (SST) anomalies. This is also not a discussion of whether the rises in Sea Surface Temperatures result from natural or anthropogenic causes. This post illustrates the Sea Surface Temperature reporting errors by Jeff Masters in recent blog posts at his www.wunderground.com blog, which were then repeated by Joe Romm at Climate Progress.

INTRODUCTION

My recent post Are Gulf Of Mexico Sea Surface Temperature Anomalies Near To Record Levels? illustrated the errors made by Jeff Masters in his post “Tornadoes, floods, and fires continue to pound U.S.” Joe Romm cross posted portions of that Weather Underground post in “Masters: Midwest deluge enhanced by near-record Gulf of Mexico sea surface temperatures”. Since Masters had missed the mark with the reporting of Gulf of Mexico Sea Surface Temperature anomalies, I decided to take a closer look at the other claims. The portion of that Climate Progress post to be discussed reads:

#############

Masters showed back in January the connection between high SSTs and record flooding around the world:

If we look at the departure of temperature from average for the moisture source regions of the globe’s four most extreme flooding disasters over the past 12 months, we find that these ocean temperatures ranked 2nd or 3rd warmest, going back through 111 years of history:

• January 2011 Brazilian floods: 2nd warmest SSTs on record, +1.05°C (20S to 25S, 45W to 40W)
• November 2010 Colombia floods: 3rd warmest SSTs on record, +0.65°C (10N to 0N, 80W to 75W)
• December 2010 Australian floods: 3rd warmest SSTs on record, +1.05°C (10S to 25S, 145E to 155E)
• July 2010 Pakistani floods: 2nd warmest SSTs on record, +0.95°C (Bay of Bengal, 10N to 20N, 80E to 95E)

#####################

THE GOOD NEWS

As noted in the Overview, two of those claims of near-record SST anomalies were correct. They are the ones for the coastal waters of Australia and Pakistan. (I am using the term coastal waters loosely.)

THE BAD NEWS

Let’s look next at the Brazil Coastal Waters for Januarys, Figure 1. The January 2011 SST anomalies for coordinates of 25S-20S, 45W-40W were +0.15 deg C. But according to Jeff Masters, the January anomaly should be +1.05 deg C and should have been the 2^ndwarmest SST anomaly for that grid since 1900. Why is there a difference?

Figure 1

A look at the SST anomalies for the Brazil Coastal Waters since January 2010, Figure 2, reveals that the SST anomalies for December 2010 were +1.05 deg C and that they dropped to +0.15 in January. How could Jeff Masters be off by a month?

Figure 2

The Brazil flooding was first discussed in Jeff Masters’s January 14, 2011 post At least 611 dead in Brazilian floods: Brazil’s deadliest natural disaster in history. It was only mid-month then, so the HADSST2 data for January 2011 had not been posted by the Hadley Centre. Jeff Masters obviously assumed the SST anomalies for the coordinates he chose would be same in January as they were in December. They were not. They weren’t close.

And Jeff Masters described the role of the assumed-to-be-elevated SST as follows:

This week’s [post written on January 14, 2011] heavy rains occurred when a storm system crossing from west to east over southern Brazil drew in a moist southerly flow air off the Atlantic Ocean over southern Brazil at the surface. At higher levels, the storm drew in very moist air from the Amazon. Sea surface temperatures along the Brazilian coast are at near-record warm levels, which likely contributed to the heavy rains. Record rains are more likely when sea surface temperatures over the nearby moisture source regions are at record high levels. This occurs because increased amounts of water vapor evaporate into the atmosphere from a warm ocean compared to a cold one, due to the extra motion and energy of the hotter water molecules. According to an analysis I did of the UK Met Office Hadley Centre sea surface temperature data set, December 2010 sea surface temperatures in the 5×5 degree region of Earth’s surface along the Brazilian shore nearest the disaster area, 20S to 25S and 45W to 40W, were the second warmest on record since 1900. Temperatures were 1.05°C (1.9°F) above average in this region last month. Only 2007, with a 1.21°C departure from average, had warmer December ocean temperatures.

Were the December 2010 SST anomalies for the area Jeff Masters chose the second warmest on record? Sure. Refer to Figure 3. But the December data is irrelevant. Based on Jeff Masters’s description, we’re concerned with the Sea Surface Temperature during the week of the flood, because that “week’s heavy rains occurred when a storm system crossing from west to east over southern Brazil drew in a moist southerly flow air off the Atlantic Ocean over southern Brazil at the surface.”

Figure 3

We can check Sea Surface Temperatures that week by switching to weekly Reynolds OI.v2 SST data, Figure 4. (Note: one of the reasons for the differences in the reported anomalies is HADSST2 uses 1961-1990 as base years while Reynolds OI.v2 data uses 1971-2000.) Regardless of the base years, the weekly SST data for the area Jeff Masters selected dropped significantly (more than 0.8 deg C) from the week centered on January 29, 2010 to January 5, 2011, then rose slightly during the week of the flooding. The Sea Surface Temperatures for the coordinates Jeff Masters chose were not close to record levels at the time he claims the “storm system crossing from west to east over southern Brazil drew in a moist southerly flow air off the Atlantic Ocean over southern Brazil at the surface.”

Figure 4

Jeff Masters then reports the December values as January SST data later in his post [My bold face]:

If we look at the departure of temperature from average for the moisture source regions of the globe’s four most extreme flooding disasters over the past 12 months, we find that these ocean temperatures ranked 2nd or 3rd warmest, going back through 111 years of history:
January 2011 Brazilian floods: 2nd warmest SSTs on record, +1.05°C (20S to 25S, 45W to 40W)
November 2010 Colombia floods: 3rd warmest SSTs on record, +0.65°C (10N to 0N, 80W to 75W)
December 2010 Australian floods: 3rd warmest SSTs on record, +1.05°C (10S to 25S, 145E to 155E)
July 2010 Pakistani floods: 2nd warmest SSTs on record, +0.95°C (Bay of Bengal, 10N to 20N, 80E to 95E)

And those claims of “2^ndwarmest SSTs on record” get repeated in his later posts and those by Joe Romm. The error that keeps on giving.

THE QUALIFIER

Jeff Masters then qualifies the near-record SST anomalies he’s listed with:

The size of the ocean source region appropriate to use for these calculations is uncertain, and these rankings will move up or down by averaging in a larger or smaller region of ocean. For example, if one includes an adjacent 5×5 degree area of ocean next to Brazil’s coast that may have also contributed moisture to this week’s floods, the SSTs rank as 7th warmest in the past 111 years, instead of 2nd warmest. It would take detailed modeling studies to determine just how much impact these near-record sea surface temperatures had on the heavy rains that occurred, and what portion of the ocean served as the moisture source region.

Maybe his qualifier needs to be amended so that it accounts for the reporting of data from the wrong month—and so that it accounts for his claims that have no relationship to the actual data, like the Columbia Coastal Waters.

THE REALLY BAD NEWS

Jeff Masters reported a November 2010 SST anomaly of +0.65 deg C for the Coastal Waters of Columbia. But the HADSST2-based SST anomalies for the coordinates he provided (0-10N, 80W-75W) were -0.33 deg C in November 2010 or about 1.0 deg C less than what was reported. Refer to Figure 5.

Figure 5

Figure 6 shows the SST anomalies for the area of Columbia Coastal Waters selected by Jeff Masters from January 2010 to March 2011. The SST anomalies there were +0.65 deg C in July 2010, not November 2010. Could he have missed the mark by 4 months, reporting July anomalies in November?

Figure 6

Both of the areas used by Jeff Masters for the Coastal Waters of South America are shown in Figure 7.

Figure 7

His use of the west coast data for the Columbia floods struck me as odd, since the Columbian flooding in November 2010 was reported to have resulted from the combined effects of La Niña conditions in the central equatorial Pacific and from Hurricane Tomas. The storm track for Hurricane Tomas is shown in Figure 8, which is from the Weather Underground Hurricane Tomas webpage. The track of Hurricane Tomas through the Caribbean Sea took it near to Venezuela and Columbia during the first few days of November 2010. Maybe Jeff Masters would have had better results if he’d selected the Caribbean Sea for his analysis.

Figure 8

CLOSING

Only 2 of Jeff Masters’s 5 claims of near-record high Sea Surface Temperatures during recent floods have been confirmed by his source data. Not a good percentage.

SOURCE

The HADSST2 data used in this post is available through the KNMI Climate Explorer:

http://climexp.knmi.nl/selectfield_obs.cgi?someone@somewhere