donsutherland1

At 11 pm, Hurricane Maria was centered at 17.3°N 64.7°W. Maria was continuing to bear down on Puerto Rico where it should make landfall tomorrow morning either as a Category 4 or perhaps Category 5 hurricane. The EPS continued to argue strongly against strongly against Maria’s making landfall on the Continental United States. However, both the 18z GFS and 18z GEFS increased prospects of such landfall. In part, this guidance weakened Jose faster than the ECMWF did and it showed Jose having less interaction with Maria than the ECMWF did. The probability associated with historical climatology (1851-present) associated with hurricanes that developed in the 8/15-10/15 period and that passed within 100 nautical miles of Maria’s 11 pm position was slightly lower than was the case at 11 am. That probability was approximately 45%. However, an adjusted perspective based on the synoptic pattern yielded a landfall probability of 31%. The adjustments were based on (1) excluding tropical cyclones that tracked into the Gulf of Mexico and toward either Central America or Mexico or Texas to Louisiana and (2) excluding landfalls for 9 of 12 Florida landfall cases from the data set. The excluded cases were hurricanes that were west of 70°W upon reaching 20°N latitude. At present, Maria is likely to be somewhere between 68.5°W and 69.5°W at 20°N latitude. That is very close to the forecast position shown on the 12z ECMWF. The ECMWF has been averaging an error of about 30 miles over 48 hours so far with Maria. Therefore, I am reasonably confident that Maria will be east of 70°W when it reaches 20°N latitude. The last four NHC forecasts for Maria’s longitude at 20°N latitude were: 5 am forecast: 69.0°W 11 am forecast: 68.8°W 5 pm forecast: 69.2°W 11 pm forecast: 68.7°W Overall, as had been the case this morning, I believe the probability of Maria’s making landfall on the mainland U.S. is about 30%. The East Coast would be overwhelmingly favored for any such landfall.

Through 11 am, Hurricane Maria has continued to follow a track that is a shade north of the track taken by Hurricane #2 through the August 31-September 3 1896 period. Given the forecast synoptic pattern over the next 48-72 hours and the tracks depicted on the EPS and GEFS, such a trajectory just to the north of that taken by Hurricane #2 (1896) remains reasonable for that timeframe. Such a track would bring Maria across Puerto Rico and north of Hispaniola.That idea is fairly close to the latest NHC forecast through 72 hours, winding up just south of the NHC's 72-hour forecast of 21.5°N 70.5° (about 21.0°N latitude at that point). Overnight, the GEFS has moved closer to the EPS idea of Jose's retaining a larger and stronger center of circulation than had previously been the case on the GEFS. The NHC's 11 am discussion also hints at such an outcome noting: The recent formation of convection to the southeast is beginning to give Jose the appearance of developing a large truck-tire eye with a diameter of approximately 100 nmi. Such eye patterns typically indicate a fairly stable cyclone that doesn't weaken or weakens only slowly. The presence of Jose should reduce prospects of Maria's following the aforementioned hurricane's long-term path, which featured a bend back toward the U.S. East Coast. Therefore, considering the narrowing GEFS-EPS forecasts and a modest decrease in landfall prospects for hurricanes passing within 100 nautical miles of Maria's 11 am position over the past 24 based on historic climatology, my thinking is that the probability of U.S. landfall (CONUS) has diminished somewhat from 40% to 30%.

At this time, Hurricane Maria remains on course to potentially threaten the mainland U.S. after punishing Dominica as a Category 5 storm and later Puerto Rico (likely as a Category 4 hurricane). A compromise position between historical climatology (52% probability of U.S. landfall and 60% with southern track outliers excluded) and the lower probability shown on the ensembles and operational ECMWF/GFS remains reasonable right now due to considerable uncertainty beyond 72 hours. Hence, there's little reason for me to change my thinking from a 40% probability of landfall. Jose's fate will likely determine, in large part, Maria's fate. Both the EPS and GEFS are in strong agreement about the development of a 500 mb ridging in eastern North America that closely resembles that which was in place when Dennis (1999) completed an anti-cyclonic loop to make landfall in North Carolina as a tropical storm. The big difference is that the EPS has a much more pronounced circulation around Jose than the GEFS. Should Jose maintain a strong and expansive circulation, it will likely contribute to Maria's remaining offshore even as it heads westward for at least a time. A weaker and less expansive circulation would likely result in Maria's posing a much greater threat of landfall with perhaps both systems being swept westward for a time. The operational ECMWF and GFS closely resemble their ensemble means. Not surprisingly, even as the ECMWF sends Maria out into the open Atlantic and away from the U.S. East Coast, the GFS has a much closer call with Maria approaching the North Carolina coast just before it turns out to sea. For now, Jose's fate remains uncertain. Therefore, neither the out-to-sea nor U.S. East Coast landfall scenarios can be written off just yet. Note: The GEFS is the 18z run, not the 12z run.

Probably baroclinic deepening.

Thanks for the kind words. I believe the NHC has started doing so. With Irma, the NHC went beyond the 120-hour period in its discussions. For example, the 9/4 11 am discussion read, in part: 4. There is an increasing chance of seeing some impacts from Irma in the Florida Peninsula and the Florida Keys later this week and this weekend. In addition, rough surf and dangerous marine conditions will begin to affect the southeastern U.S. coast by later this week. Otherwise, it is still too early to determine what direct impacts Irma might have on the continental United States. However, everyone in hurricane-prone areas should ensure that they have their hurricane plan in place, as we are now near the peak of the season. Overall, the trade-offs involved are difficult, as larger time horizons involve larger errors. The NHC received some criticism when Irma ultimately made Florida Keys landfall and then landfall on the southwestern portion of the Florida Peninsula after an earlier focus more on the Miami area. Even as the distance involved was relatively small and the overall forecast was good given the complexities involved, the impact of the actual point of landfall from what had earlier been anticipated was substantial. What could have been a catastrophic hurricane landfall turned out to be a damaging one. The worst damage was confined to the Florida Keys and a small portion of southwestern Florida rather than the greater Miami area. Overall, damage will wind up much less than the $100 billion-$200 billion disaster that Irma could have been. Going forward, will this relatively better outcome encourage a sense of complacency? It's difficult to know. The NHC and NWS have a lot of survey data on how people respond to various forecasts and outcomes following the forecasts. Personally, I'm not sure if too many people view things in terms of probabilities (probably more do so here than in the general public). Instead, many likely see things in a more deterministic, binary fashion: a storm will make landfall or it won't. This is a very difficult challenge for forecasters.

So far, Hurricane Maria remains on course to reach Puerto Rico in about two days as a major hurricane (possibly a Category 4 storm). Two major hurricanes from the climatological data set for hurricanes within 100 nautical miles of Maria's 11 am position that made landfall on Puerto Rico were the 1928 Lake Okeechobee Hurricane (Hurricane #4) and Hurricane Hugo in 1989. After possible Puerto Rico landfall, Maria's longer-term fate remains unresolved. On account of its trajectory, climatological odds of landfall on the U.S. Mainland have increased to nearly 50%. However, with extratropical cyclone Jose likely to make another anti-cyclonic loop that moves it closer to the U.S. coast, those odds likely overstate Maria's actual prospects of landfall at present. Instead, the gradually weakening Jose will likely possess a broad circulation that helps turn Maria more to the north and perhaps later northeast, which would result in lower prospects of landfall. Should Jose weaken more rapidly than anticipated, its ability to help steer Maria away from the U.S. East Coast would likely be reduced. Other synoptic features that are not currently modeled as major players could grow in importance, though. The EPS and GEFS imply a much lower prospect of landfall. Only a handful of EPS members show landfall (one small cluster across Florida and another small cluster along North Carolina's Outer Banks). Many of the ensemble members already depict a northwest trajectory. With Maria still tracking west-northwestward and that motion likely to continue for the next 12-24 hours and perhaps a little longer, the ensembles likely understate the prospect of U.S. landfall. In fact, the NHC has adjusted its track to the left with the largest adjustments occurring over the next 48 hours. In short, all of this implies that Maria will likely make a wider turn than what most of the ensemble members are currently depicting. Overall, my guess remains that Maria's U.S. landfall probability is about 40%. The southern-most EPS members may offer the most reasonable solution through the next 48-72 hours.

At 11 pm, Maria continued to add latitude to its track. As a result, its position of 14.2°N 58.4°W was north of such hurricanes as Dog (1951) and Janet (1955) at 58.4° longitude. That trend is expected to continued based on the guidance, which is in strong agreement. As a result, the probability of Maria's being a storm that tracks largely westward into the Gulf of Mexico and perhaps toward Mexico is small and continuing to decline. Overall, historical climatology suggests that the probability of U.S. landfall (CONUS, as Puerto Rico remains at an elevated risk of landfall) is increasing. However, as the ensembles showed a reduction in that probability, there's little reason for me to change my earlier thinking that Maria has about a 40% probability of U.S. landfall. Should such landfall take place, the East Coast remains at the greatest risk.

IMO, the 500 mb pattern argues against Dog's being a good analog.

With now Hurricane Maria continuing to add latitude, the kind of tracks that took storms that originated in an area within 100 nautical miles of Maria's formation safely south of the U.S. into the Gulf of Mexico are increasingly outliers in the historical climatology dataset. Given the previous ensemble data, such tracks were unlikely. Once those tracks were excluded, Maria's probability of landfall was roughly 2-in-5. Since 1851, 38% of the hurricanes that formed in the August 15-October 15 period and passed within 100 nautical miles of Maria's 5 pm position (13.8°N 57.5°W) made U.S. landfall. 63% of such cases saw landfall on the East Coast from Florida to North Carolina. 25% made landfall in southwestern Florida. The 12z ensembles show a reduced prospect of landfall. But the events that lead to that outcome are highly complex and, in part, dependent on Jose's fate. Therefore, I believe the climatological position of an approximately 40% of U.S. landfall still looks reasonable. Such landfall would be favored to occur on the U.S. East Coast with the Florida-North Carolina area most at risk.

It appears that the September 9 figure of 4,472,225 square kilometers (JAXA) could be this year's minimum sea ice extent. That would be the highest figure since 2014 when the minimum extent was 4,884,120 square kilometers.

I will periodically focus on the hour 72 mark, as by hour 120 the NHC's errors grow quite large. Hour 72 could provide better insight into actual trends associated with tropical cyclones' paths.

At 11 pm, Maria was centered at 12.5°N, 53.7°W. Maria was continuing to track to the west. Maria is expected to gradually add a northerly component to its motion. However, over the last 6 hours, the NHC track was nudged somewhat south and westward through 72 hours (about 40 km by hour 72). Until Maria begins to add latitude, climatology will likely understate its landfall prospects. Taking into consideration the EPS and GEFS, as well as climatology that excludes select storms that remained largely on a westward trajectory through the duration of their existence, it appears that Maria has an approximately 40% probability of making U.S. landfall (excluding Puerto Rico).

Maria, now a tropical storm, was situated at 12.3°N, 52.6°W. According to historic climatology, just under 1-in-5 tropical cyclones that formed within 100 nautical miles of Maria's area of development made U.S. landfall. However, a large number of those systems tracked generally westward, many dissipating in the process. The synoptic pattern as currently modeled argues against such a scenario. Therefore, excluding the systems that tracked generally westward and failed to make a turn more to the north, the smaller sample yields a landfall probability closer to 2-in-5. That's quite high. Notable systems that made U.S. landfall included Hurricanes David (1979), Frederic (1979), and Ivan (2004). At least for now, Maria bears watching.

Arctic sea ice extent is mirroring the evolution of the summer 2014 decline very closely. That would imply a minimum figure of between 4.4 million and 4.5 million square kilometers. https://s26.postimg.org/64xg85hrt/Arctic08292017.jpg

In March, Arctic sea ice extent averaged 13,679,429 square kilometers. That broke the March record low average of 13,729,002 square kilometers, which had been set in 2015. For the January 1-March 31 period, Arctic sea ice extent averaged 13,294,815 square kilometers in March. That set a new record low average for the first three months of the year. The previous record was 13,430,714 square kilometers, which was set last year.

During February, Arctic sea ice extent averaged 13,566,787 square kilometers. That set a new record low average for February. The previous record was 13,595,448 square kilometers, which was set in 2016. These figures all use a 2/1-28 timeline.

Dr. Bates' clarification seemed to suggest that his issue concerned archiving. The Daily Mail quoted him on processing and says that it stands by the quote. Given that Dr. Bates' clarification, I suspect the issue dealt with archiving, not processing. I can see why a journalist might confuse the two, and that could well have occurred. In any case, there were better ways for Dr. Bates to address his concern that he should have pursued.

Apparently, there seems to be contextual information related to Dr. Bates' criticism of the Karl paper. https://www.nytimes.com/2017/02/20/business/energy-environment/climate-change-dispute-john-bates.html

Lowest January Arctic Sea Ice Extent Averages: 1. 12,664,550 km2, 2017 2. 12,859,672 km2, 2016 3. 12,895,995 km2, 2011

On JAXA, the latest figure was almost 500,000 sq. km. below the previous daily record low extent.

Some temporary declines in January have occurred in the past. Statistically, it is almost improbable that we have already reached the peak Arctic sea ice extent figure (>4.5 sigma event). During the 2003-16 period, the earliest peak was February 15 (2015). The latest was March 31 (2003 and 2010). The mean and median dates were March 10.

2016 was a remarkable year as far as Arctic sea ice extent was concerned. A summary of data is below:

And this is the second such case since November.

That's extraordinary warmth.

What has been going on in the Arctic is nothing short of historic, at least as far as sea ice extent record keeping is concerned. The October 1-December 9 mean figure for sea ice extent is 7,483,994 square kilometers. The previous record from 2012 was 7,760,443 square kilometers. Moreover, if one took the daily record low minimum figures for the same period of time prior to 2016, the mean figure would be 7,665,564 square kilometers. So far, 2016 has seen daily record low figures established on 180 days. It is now all but certain that 2016 will have established record low figures on at least half the days.