MODERN DAY SHIP WEATHER PREDICTION SYSTEMS AND SERVICES - WHAT HAPPENED TO THE EL FARO?

The loss of the El Faro in Hurricane Joaquin with all 33 crew on 01 October 2015 is a mystery in the modern era of weather prediction and hurricane tracking capabilities.  While Tote Maritime believes the ship sank after suffering engine failure during its weekly run from Jacksonville to San Juan, they have stated that the Master had a sound plan to avoid Hurricane Joaquin and that he had real-time weather information when departing from Jacksonville.  However, closer examination as to why the ship was where it was prior to the engine failure still remains a mystery, and one that may never be fully resolved. 

Evidence from ship tracking data indicates that on 30 Sep the El Faro was at near full speed going into the centre of the storm before it lost propulsion. This information reveals that the vessel was in a position that any normal sailing plan would routinely avoid under the Mariner's 1-2-3 Rule to avoid any danger areas from a tropical storm.  The National Hurricane Centre (NHC), Miami was issuing regular updates to TS Joaquin and this data would have been available to the Master in their broadcast Hurricane Warning text messages.  It is assumed that their weather Service provider would have been doing the same. The graphic below (issued by USCG) was used to illustrate the growth of the storm relative to the vessel’s track taken from ship tracking data.

​                                                                                                                              Image from NHC

Lastly, I frequently get asked whether all this current activity is down to global warming.  The consensus is that global warming is not thought to impact on the number of hurricanes that occur, but that it does directly impact on the intensity.  This is due to the fact that warmer temperatures can hold greater volumes of water vapour and store greater energy resulting in a combination of either and/or stronger winds/greater storm surges and greater volumes of precipitation. Most research reveals that the intensity of tropical cyclones tends to increase with global warming/climate change.  However, the case is not so clear for their frequency with some reporting that the frequency of cyclones could well will decrease.  Time will tell!                                      Posted 14 Sep 2017


OCEAN ROUTEING AND MEETING ARRIVAL SLOTS

The recently expanded Panama Canal has increased maximum vessel capacity to Length 366m, Beam 49m and Draft 15.2m.  During the first three months of its operation, 238 Neopanamax vessels took advantage of the newly opened expanded canal.

The expansion work also created a third lane of traffic that effectively doubles the cargo capacity of the waterway. This improvement will result in greater ship numbers using it with a corresponding greater emphasis being placed on canal transit slots.  Since opening the new expansion, ship numbers climbed to 13114 vessels over the last twelve months; the highest number since it opened 102 years ago. 

In terms of shipping market segments the canal authority reports that the leading segment was container ships (36%) followed by bulk carriers, tankers and vehicle carriers. While the expansion allowed the average size of container vessels to double it also now allows LNG carriers to transit the canal as well.

To help cope with ever growing numbers, the Panama Canal Authority operates a booking system for transits up to 365 days in advance, with a sliding scale of charges that increase nearer to the requested time of transit.  

For vessels that arrive without a booking then the transit is conducted on a first come first served basis, where lengthy waits can be expected dependent on the number of vessels queuing, the weather conditions plus any canal maintenance requirements.

Collectively, this all places greater emphasis on arriving in sufficient time to meet any pre-booked slots. A cancelled slot followed by a lengthy wait could cost many thousands of dollars. Conversely, arriving with an excessive margin means wasted fuel on passage resulting from a speed greater than was required with subsequent extra fuel costs and emissions.  A similar analogy will apply to the New Suez Canal which will increase the draft to 20m and to nearly double capacity from 49 to 97 transits per day by 2023. 

There are many companies that provide an optimum ship routeing service that help ensure a Just-in-time arrival, whether for meeting canal transits or dockyard slots. The aim of these services is to achieve an On-Time-Arrival (sometimes named a ‘Just-In-Time-Arrival’) or Virtual Arrival time that will help optimize the efficiency of the vessel and so contribute to the vessel’s Ship Energy Efficiency Management Plan (SEEMP).

OceanPlus’s staff have many years’ experience of helping clients ensure their vessels arrive at destinations on time, in a safe, efficient and cost effective manner.  Using the latest technology with access to long range, accurate weather, sea, swell and current data plus proprietary algorithms,  the company provides a bespoke routeing service that delivers significant time, cost and emissions savings.  Savings to date vary between clients but figures of between 5-10% have been achieved depending on which particular route. Such level of savings on a single day would self-fund the annual fees for this type of service!                                                               

                                                                                                                                                                                                   11 Nov 2016

Over the past decade, the NHC report that the average track errors of GFSI and EMXI models have been quite close, so even though EMXI was the best-performing model most of the time in 2017, it does not mean that it will always be the best for every storm.  So best practice involves forecasters using multiple model simulations and their knowledge of individual model strengths and weaknesses. Of note, NHC report that preliminary estimates for the 2017 Atlantic track forecasts set record low errors at all time periods. Typical track errors are currently about 37nm at 24 hours and increase by about 35nm per day of the forecast. This means that the NHC 3-day track error is 100nm and the 5-day track error is 180nm. So, from a mariner’s perspective, the 180nm/5-day track error is a useful number to keep in mind when route planning and considering forecast uncertainty. This is especially true for those at sea who rely on a single model forecast displayed on their bridge.  It is essential to use other information to compare with a single model, whether this is access to another model or noting information provided via NAVTEX/GMDSS, Radiofax, etc.which might be more timely.  This is a key lesson learnt from the tragic loss of the El Faro during hurricane Joaquin in 2015.

Lastly, the NHC blog points out that predicting the intensity of a tropical storm is usually more challenging than forecasting its track.  So, different specialist high resolution models and statistical models are used to do this, and NHC have not made as much progress on intensity forecasting as they have for track forecasting since the 1900s. However, there has been a significant improvement in the last decade through recent investment in new models and techniques. The current intensity error is close to 15knts from 72 to 120 hours, so the NHC typically warns for a storm one category stronger (on the Saffir-Simpson Hurricane Wind Scale).                                                                     

                                                                                                                                                              Posted 10 April 2018

                                                                                                                           ​Image from NHC

However, there is nothing clear about JOSE at this time.  Model accuracy is not good when systems loop, such that the error track remains large as shown in the image below from the NHC.  There remains a lot of uncertainty over exactly what JOSE will do over the next 6-7 days.  The consensus is that the system will slowly swing north and northeast off the US Eastern Seaboard over this period and is likely to regenerate once back over the warmer waters of the Gulfstream.  The majority of models keep JOSE offshore and it is possible that JOSE may still perform another loop.  The reality is that JOSE will prove problematic for the Bahamas and US Eastern Seaboard/NE Atlantic region until 20-21 Sept, keeping ship weather routeing specialists busy.   



From a weather perspective the MBI found that the ship’s operator, TOTE, did not ensure the safety of marine operations for the vessel and did not identify the risks posed by Hurricane Joaquin before her final voyage.  It also found that the master was partly responsible for the tragedy by sailing “…within close proximity to Joaquin” and alleged that he "failed to carry out his responsibilities and duties as Captain of the vessel.”   

As such (in terms of getting within close proximity to the hurricane), the MBI identifies 11 key shortcomings related to  the processes, protocols and rules for oversight of ship operations, routeing, heavy weather risks, vessel vulnerabilities and weather observations (the ship’s anemometer was broken). These short comings also extended to their lack of knowledge and understanding of onboard weather support systems, particularly in accessing the most current weather data.  This was critical at a time when the National Hurricane Centre (NHC) created and distributed tropical weather forecasts for Tropical Storm/Hurricane Joaquin, which in later analysis proved to be inaccurate.  TOTE’s weather service provider, Applied Weather Technologies (AWT) used these inaccurate forecasts to create the Bon Voyage System (BVS) weather packages.  It also appears that the Master and deck officers relied primarily on graphical BVS weather forecasts rather than the most current NHC data received via SAT-C.  Furthermore, they did not take advantage of BVS’s tropical update feature and the ability to send BVS weather information directly to the bridge. Other cumulative events such as ship stability, listing, intermittent flooding, lack of heavy weather plans and finally the loss of propulsion were all contributing factors to the tragedy as a result of being caught up in the hurricane. The MBI report stops short of criminal activity. 

In terms of investigating wider weather related criteria behind this tragedy, the final NTSB report is likely to be broader ranging in its criticisms and recommendations.  It is worth noting its report on Tropical Cyclone Information for Mariners dated 20 June 2017 – see link  https://dms.ntsb.gov/public/58000-58499/58116/605296.pdf

The bottom line of this NTSB report is that while more accurate forecasting of tropical cyclones would improve marine safety, mariners at sea need to be aware of and receive all critical weather forecasts and advisories in a timely manner.  NTSB investigations have revealed the presence of critical information and different products (covering current position, track and intensity) being made available on different systems at differing times, some of which is not even available to mariners.  The end result can be confusing and conflicting information for mariners.

By example, the El Faro crew relied on Inmarsat-C SafetyNET (SAT-C) text broadcasts of NHC weather products to remain aware of the changing position, forecast intensity, and forecast track of Joaquin. The NHC product received via SAT-C that delivered (with limited delay) the most current information on Joaquin’s position and forecast track and intensity was the Tropical Cyclone Forecast/Advisory. However, the NTSB concludes that mariners using only SAT-C miss important tropical cyclone information found in the Intermediate Public Advisories (IPAs) and Tropical Cyclone Updates. Under current NWS policy IPAs are only issued when a hurricane threatens land areas.   NTSB stress that these updates are critical when a storm may not behave according to the computer models (for forecast track or intensity) in the 3 hours after a forecast has been issued.  This could have crucial impact when following the Danger Rule (The 1-2-3 Rule) to keep out of a hurricane’s path.

The NTSB report, based on the meteorological facts gathered to date, and working with US National Weather Service (NWS) and the Coast Guard, makes ten recommendations which mariners will welcome; two are addressed to NOAA, seven to the NWS, and one to the Coast Guard:

To the National Oceanic and Atmospheric Administration:

  • Develop and implement a plan specifically designed to emphasize improved model performance in forecasting tropical cyclone track and intensity in moderate-shear environments.


  • Develop and implement technology that would allow National Weather Service forecasters to quickly sort through large numbers of tropical cyclone forecast model ensembles, identify clusters of solutions among ensemble members, and allow correlation of those clusters against a set of standard parameters.


To the National Weather Service:

  • Work with international partners to develop and implement a plan to ensure immediate dissemination to mariners, via Inmarsat-C SafetyNET (and appropriate future technology), of the Intermediate Public Advisories and Tropical Cyclone Updates issued by the National Weather Service, in a manner similar to the current process of disseminating the Tropical Cyclone Forecast/Advisory.
  • Modify your directives to ensure, for all tropical cyclones of tropical storm strength or greater within your jurisdiction, that your facilities issue, at the 3-hour interval between regularly scheduled Tropical Cyclone Forecast/Advisories, an Intermediate Public Advisory, a Tropical Cyclone Update, or another product available (or expected to be available) to mariners via Inmarsat-C SafetyNET (and appropriate future technology), and that the product include the coordinates of the current storm Centre position, maximum sustained surface winds, current movement, and minimum central pressure.

​​

  • Modify your directives to ensure that the “next advisory” time in a Tropical Cyclone Forecast/Advisory clearly indicates when to expect the next update of “current” or forecast information for that particular tropical cyclone.

Quantitatively define “significant change” in terms of both the track and intensity of a tropical cyclone to guide the issuance of Special Advisory packages.

  • Ensure that tropical cyclone graphic products issued by entities such as the National Hurricane Centre, the Central Pacific Hurricane Centre, the Guam Weather Forecast Office, the Joint Typhoon Warning Centre, and Fleet Weather Centre–Norfolk are made available in near-real time via the FTPmail service.


  • Allow users to schedule recurring, automated receipt of specific National Weather Service products through an enhanced FTPmail service (and appropriate future technology).


  • Develop and implement a plan for soliciting feedback from the marine user community, particularly ship masters, about the accuracy, timeliness, and usability of weather services to mariners.


To the US Coast Guard (USCG):

  • In collaboration with the NWS, provide timely broadcasts of the Tropical Cyclone Forecast/Advisories, Intermediate Public Advisories, and Tropical Cyclone Updates to mariners in all regions via medium-frequency navigational TELEX (NAVTEX), high-frequency voice broadcasts (HF VOBRA), and high-frequency simplex teletype over radio (HF SITOR), or appropriate radio alternatives (and appropriate future technology).


There is no doubt that US authorities will be very frustrated that this accident happened. It resulted from a series of accumulative errors across the ship management, command, weather, vessel risk, safety and engineering spectrum.   It may be that NTSB will recommend new rules and regulations be applied to ship routeing and weather risk analysis to ensure this does not happen again.  An alternative option could be a heavy civil penalty.  We will need to wait and see.  Meanwhile, the recommendations to NOAA, the NWS and USCG by NTSB to provide more accurate and critical information in a coherent and timely manner, irrespective of what bridge systems are fitted, is good news for all mariners.                                                                                Posted 16 October 2017

The ship sank on 1st Oct, 2015, in the Atlantic Ocean, with the loss of all 33 people on aboard, after running in to Hurricane Joaquin.  The ship had a range of navigation options that would have allowed it to steer clear of the storm that later became a Category 4 Hurricane when she departed Jacksonville on 29 Sept, but sadly none were used.  

It will remain a mystery as to exactly why the captain, who based his decisions on outdated weather forecasts and who ignored the suggestions of his bridge officers to alter track to move further away from the storm, ordered a course that intersected directly with the path of the hurricane.  This encounter then led to a subsequent chain of events, each of which could have endangered the ship on its own  - even without the presence of a Cat 4 hurricane.  

NTSB investigations revealed that seawater entered the ship through cargo loading and other openings on a partially enclosed deck in the ship’s hull, pooled on the starboard side and poured through an open hatch into a cargo hold.  The hold began to fill with seawater, and automobiles in the hold broke free of lashings and likely ruptured a fire main pipe that could have allowed thousands of gallons of seawater per minute into the ship – faster than could be removed by bilge pumps.  Finally, the ship lost its propulsion and was unable to manoeuvre, leaving it at the mercy of 35 ft seas and 100mph winds. To make matters worse, the crew’s chances of survival were significantly reduced because the ship was equipped with life rafts and open uncovered lifeboats, which although met requirements were ineffective in hurricane conditions. The NTSB also stated that the poor oversight and inadequate safety management system of the ship’s operator, TOTE, also contributed to the sinking.  

The executive summary is available at https://go.usa.gov/xnRAn and the complete accident report will be available in several weeks. 

From a ship weather perspective it will be interesting to monitor what improvements are introduced and by when by the National Ocean and Atmospheric Administration (NOAA) and the National Weather Service (NWS).  This follows from the NTSB Safety Recommendation Report on Tropical Cyclone Information for Mariners issued on 20 June 2017 (see my earlier blog issued 16 October 2017 for detail on the recommendations (x 2) to the NOAA and (x 7) to NWS). 

Until these improvements appear NTSB has meanwhile issued a Tropical Cyclone Information leaflet for Mariners titled “How to stay up-to-date on tropical cyclones” and “What mariners should know”.  This is pasted below for convenience:

  • The NWS may issue bulletins that provide the public with updated information on tropical cyclones between the regularly scheduled advisories issued at 0300, 0900, 1500, and 2100 UTC. 
  • These bulletins update storm position, storm intensity, current storm movement, and minimum central pressure and may be called an “Intermediate Public Advisory,” “Tropical Cyclone Update,” or “Special Advisory.” 
  • With one exception, these advisories are currently not available from Inmarsat-C SafetyNET (SAT-C) or US NAVTEX broadcasts (on 518 kHz), and may not be available from US Coast Guard HF Voice or HF SITOR broadcasts. 
  • The NWS issues a “Tropical Cyclone Forecast/Advisory” that is available via SAT-C (but not via NAVTEX). This product contains a “next advisory” time that does not reflect when the NWS’s next Intermediate Public Advisory can be expected. 
  • The NTSB is not aware of any public broadcast service for mariners that provides the NWS’s “Tropical Cyclone Discussion.” This product is important because it conveys to mariners the rationale for a tropical cyclone forecast, and may describe the forecaster’s degree of confidence in the official forecast, discuss alternative scenarios, and highlight unusual hazards. 
  • Many tropical cyclone graphics issued by the NWS and the US Navy are not available via the NWS’s “FTPmail” service (see below).

 What mariners should do:

  • Be sure you know what tropical cyclone forecast and advisory products are issued for the area you will be traveling through and at what times they are scheduled to be distributed.
  • Verify that forecasts and advisories are available in a timely fashion from onboard weather receipt systems (such as SAT-C or radio).
  •  If you use a private weather vendor, make sure the vendor provides you with all available tropical cyclone advisories in a timely fashion and that you know when advisories are updated.
  •  ALWAYS! check the generation time of all received weather information to verify that you have the most recently available update.
  •  If you have email send/receive capabilities, learn how to use weather information sources such as the NWS’s “FTPmail” service.2 FTPmail can provide extremely valuable text and graphic weather information to supplement GMDSS and radio.
  • Understand that GMDSS and radio provide only some of the tropical cyclone information available to mariners. Other sources such as FTPmail and the internet may make certain products (such as the National Hurricane Center’s High Seas Forecast) available faster than GMDSS and radio.​
  • Become familiar with and use the “Tropical Cyclone Discussion” as an important source of tropical cyclone forecast information.


Up-to-date information can save lives:

Decision-makers aboard any vessel, and their shoreside support, must be thoroughly familiar with what tropical cyclone information is available to them, through both onboard public weather receipt systems and private weather vendors.

Mariners and shoreside personnel should also understand what is not available to them through normal channels, determine whether that tropical cyclone information is critical, and identify when and how to acquire it. Steps (and times) for acquiring all weather data should be posted on board for crewmembers to follow.

And again, ALWAYS! check the generation time of all received weather information to verify that it’s the most recently available update.


Summary


Familiarity may well lead to complacency, misguided self confidence or even arrogance; but each hurricane is unique and no one hurricane should be assumed to be the same.  While the ship was receiving weather forecasts from AWT it was not paying for a ship routeing service.  At the same time Hurricane Joaquin proved to be problematic to predict as it developed and early forecasts were low in confidence and not accurate, but this crucial information was either not made clear or was not conveyed.  As a bottom line this tragedy would have been avoided had the El Faro been using a ship routeing service.  


NTSB Chairman Robert Sumwalt expressed optimism that long-lasting lessons might be learned from the deadly incident. Quote "This report will be studied by mariners young and old for many years," and "I'm confident that this tragedy at sea, and the lessons from this investigation, will help improve safety for future generations of mariners." So rather than increase regulation the emphasis is, for now, on learning lessons. Lets hope that the recommendations are all implemented.                       15 Dec 2017

THE 2017 ATLANTIC HURRICANE SEASON


Each hurricane seasons brings its uniqueness and the 2017 Atlantic hurricane season did not disappoint. With 17 named storms and 10 hurricanes, it was in the ‘hyperactive’ category and ties with 1936 as the fifth-most active season since records began in 1851.  It also had the highest number of major hurricanes since the 2005 season.


Image from National Hurricane Centre


IRMA follows soon after Hurricane HARVEY, a Cat 4 hurricane (sustained winds 130-156mph/113-136knots), in late August which was the first major hurricane to hit US soil since 2005 and resulted in some 40 inches (1000mm) of rain over eastern Texas causing catastrophic flooding.

There is also Tropical Storm Jose  currently out in the central Atlantic which is expected to reach hurricane strength over the next 24 hours.  The models currently predict that this storm will curve more north west and so avoid any direct contact with land for now but the Northeast Leeward Islands may feel some impact from this storm - only some 3-4 days after Hurricane IRMA. 


CONCERNS OVER USE OF NUMERICAL WEATHER PREDICTION MODEL DATA AT SEA

Numerical Weather Prediction (NWP) model data, which can be updated four times a day, is increasingly becoming freely available on the internet today for use in bridge systems at sea.

This data is used by commercial weather service providers who merge it with data from other sources, such as tides, currents and tropical cyclone information, as well as adding their own information such as weather fronts to deliver a variety of products and services to the marine market. This combined data is then pushed to sea as part of a subscription service where information is displayed in multi-colours and layers for Masters to decide their safest course or optimum routeing.  This subscription service may or may not include access to higher resolution data, enroute daily weather forecasts, route optimisation or weather routeing which will add to any basic subscription costs. 

As a result of ever expanding and cheaper satellite bandwidth availability there has been a dramatic increase in pushing this type of data to sea to be managed onboard by already busy bridge teams.  In addition to any weather service provider service all ships have GMDSS as a primary source for Marine Safety Information, which includes broad brush weather forecasts and warnings released by approved weather centres where experienced forecasters provide best interpretation of the guidance from a range of numerical weather prediction (NWP) models.  Best practice is to study both the GMDSS text forecasts and other data such as the GRIB data (this is the term used to describe the data files from NWP models) displays on bridge systems.  This assumption is not a given as witnessed by the loss of the El Faro in Hurricane Joaquin in 2015 off the Bahamas where information released by the USCG investigation revealed that El Faro was receiving weather support from AWT but the ship did not have a weather routeing service when it sank.  In addition to GMDSS, the National Hurricane Centre (NHC), Miami was issuing regular updates to TS Joaquin and this data would have also been available to the Master in their Hurricane Warning text messages.

Shore based weather routing represents a superior level of service but this gold standard service normally comes at an extra cost.  However this extra cost is offset by delivering superior results.  Since shore based forecasters have access to more timely and a range of NWP models and other data, particularly assimilated real time data from space, collectively this superior knowledge and information will result in greater savings in time, fuel, transit times and even reduced insurance costs.  The end result is that across a fleet this type of service pays for itself and more through savings achieved.  When set against any damaged ship/cargo or ship loss this cost is trivial.

One possible conundrum is that many owners/charterers forego this gold level service in the belief that the proliferation of weather on bridge displays, onboard routeing systems and as overlays on ECDIS is equally as good.  Most of this data is passed computer to computer without a human expert in the loop and the source will generally reflect a single model; the most common of which is the Global Forecast System (GFS) model run by NOAA as it is freely available on the internet. This misguided trust could even extend to ignoring GMDSS weather warnings in extreme cases.

There are in fact a number of different NWP models available, often showing completely different solutions for a given point and time or along a proposed track against time.  It is dangerous therefore to rely on a single source of information and users at sea should not be fooled in to looking elsewhere.  It is my opinion that the best three global models for generic use are probably ECWMF, UKMETO and GFS however, each model has its own strength and weaknesses which will vary globally.  

This increasing reliance on NWP single model data being displayed at sea has other worrying trends. There is little or no training to Masters that provides the expertise to recognise the circumstances in which numerical model data is accurate and the times (scenarios / situations) that it is not.  Understanding the nuances of modern NWP data is certainly not the knowledge domain of Masters and busy bridge teams.

In terms of safety and to fully exploit efficiency it is essential to use the most accurate meteorological information available, preferably at the highest resolution possible.  So in terms of interpreting NWP data here are some basic considerations:


  • While low resolution data provides a good overview, high resolution data provides a significant improvement in the spatial and temporal aspects of a forecast.  The problem is that high resolution data is extremely voluminous and is more costly to send to sea despite ever improving compression ratios.  The result is that data pushed to sea tends to be at lower resolution unless one is prepared to pay for higher resolution.


  • NWP surface winds are a 10 metre neutral stability wind. This does not include gusts which in an unstable maritime air flow are 1.3 to 1.5 times the average wind speed.


  • Wind forecasts from a global model are unlikely to accurately reflect winds in squalls/thunderstorms/trough lines or in near shore areas.  ​


  • Wave forecasts from a global model are unlikely to accurately reflect wave conditions in squalls/thunderstorms/trough lines in shallow or near shore areas and where there are strong currents.


  • Understand the differences between sea (wind driven) height, swell height and Significant Wave Height (SWH) and which is being displayed as on overlay.  SWH is the average height of the highest 1/3 of the waves, which is what shipborne observers are expected to report.  It is a statistical value derived from the distribution of wave heights (which is best described as a Rayleigh distribution). As such the maximum wave height will be twice the Significant Wave Height. The occurrence of these maximum waves depends on the period of the waves and the speed of the ship which in combination give the wave encounter period. As a rule of thumb a wave of twice the Significant Wave Height can be expected every 3 hours.


Oceanplus ltd provides expert marine weather services and is focused on providing a gold level service but at bronze level prices.  We use only experienced World Meteorological Organisation accredited marine meteorologists, oceanographers and hydrographers who have  extensive sea-going experience and are familiar with interpreting multiple NWP model data to ensure the most accurate and up to information is provided to clients.  Given our global knowledge and experience our staff have unique global local weather information which can be invaluable. 

​                                                                                                                                                                                   Posted 1 Dec 2016

ATLANTIC MID-HURRICANE SEASON UPDATE – Each Season Is Unique

With a total of 3 storms at various levels of development last week it is perhaps a sign that we are nearing the peak of the hurricane season in the Atlantic.  The season officially runs 1st June to 30th November, however the formation of tropical cyclones is possible outside this period. For example, the formation of Tropical Storm Arlene on April 19 marked the 3rd consecutive year with an early out of season cyclone development, and only the 2nd named storm on record to exist in the month of April.

The table below shows the latest revised estimates for the Atlantic season from the UKMETO, Colorado State University (CSU), the Tropical Risk Consortium (TSR) consortium (University College, London) and NOAA.  Pre-season most centres were hinting towards an average season (10-12 named storms, 5-6 hurricanes and 2-3 major hurricanes) with neutral conditions expected for the start of the season and the possibility of El Niño developing later in the season (which would hinder tropical storm development).  

However, as the season approached the North Atlantic Oscillation has shown signs of moving to a negative phase, favouring a warmer tropical Atlantic and lowering the chances of El Niño forming.  The centres updated their forecasts to a more active season as indicated in the table below with the consensus shifting to an above average, if not very active, season due to a weak or non-existent El Nino.  It can be seen that by mid-season the actual number to date already reflects an average season, so 2017 is definitely looking to be an above average, if not very active, season; possibly more active than 2010 which had 19 storms, 12 hurricanes and 5 major hurricanes.


Season  Type            Mean # of             Range of                Mean # of           Range of         Mean # of Major           Range of        

                            Tropical Storms     Tropical Storms        Hurricanes         Hurricanes          Hurricanes             Major Hurricanes
 
Above-Normal              16.5                    12 to 28                  9.7                    7 to 15                4.8                     3 to 7

Near-Normal                12.3                   10 to 15                   6.3                    4 to 9                  2.3                      1 to 4

Below-Normal                6.7                    4 to 9                      3.3                   2 to 4                   1                        0 to 2

All Seasons                   12.1                    4 to 28                    6.4                   2 to 15                2.7                       0 to 7


2017 Tropical Storm and Hurricane Season Update

It is that time of year when various research and specialist weather centres start to make public their updated predictions on the 2017 Hurricane season.  Each season brings fresh challenges for mariners as no one storm is identical to another, hence the adage that each storm makes its own unique path which ever ocean it occurs in.  By definition hurricanes form over the North Atlantic Ocean and Northeast Pacific, typhoons form over the Northwest Pacific Ocean and cyclones form over the South Pacific Ocean and Indian Ocean.  Globally, there are about 80 to 100 tropical revolving systems which develop each year.   This image of cumulative tracks of all tropical cyclones over the period 1985-2005 (Nilfanion) clearly indicates that only the South Atlantic and far SE South Pacific Oceans can be regarded as regions that are ‘mostly free’ from tropical revolving storms.  


The blog also looks at model accuracy as NHC are frequently queried on which model is the best one.  While some models are usually better than others, no one model is perfect, and model performance varies from season to season and from storm to storm.  Forecasts can also be made by combining the forecasts from a number of models in to a consensus forecast.  The NHC uses many models (across the global, regional, dynamic, specialist and ensemble model and consensus aids spectrum) as guidance in the preparation of its official hurricane track and intensity forecasts. 

Two of the better known models are the US NWS Global Forecast System (GFS) and the European Centre for Medium-Range Weather Forecasts (ECMWF).  The figure below shows track errors when comparing the NHC forecasts (OFCL, black) and forecasts from the GFS (GFSI, blue) and ECMWF (EMXI, red) models for Hurricanes Harvey, Irma, Maria, and Nate in 2017 which was a very active season.  For these selected hurricanes, NHC OFCL performed as well as or better than GFSI and EMXI except for hurricane Irma.  From the model perspective, EMXI beat GFSI for Harvey, Irma, and Nate, but GFSI beat EMXI for Maria. 

THE ATLANTIC HURRICANE SEASON 2018 AND THE STATE OF HURRICANE FORECASTING

Okay, it’s that time of year when we start to take more notice on how the 2018 Hurricane season might go, especially after the very active season last year which brought much havoc, death and destruction.  1st June and 15th May mark the traditional start to the season in the North Atlantic and Eastern Pacific which last until 30th November. However, be aware that hurricanes do occur outside of these traditional dates so there is no room for complacency!  

As NOAA do not issue a seasonal outlook until late May we rely on Tropical Storm Risk(TSR), Inc (UK based) and the hurricane forecasting team from the Department of Atmospheric Science, Colorado State University (CSU) to provide this instead.  While forecasts are issued prior to April there is little skill in doing so. The skill starts to improve from April onwards although there is still a large amount of uncertainty at this early stage. Here are their latest outlooks:


                                   Named Tropical Storms          Hurricanes          Major Hurricanes


 Average Season                      12                                     7                           3

 CSU                                        14                                     7                           3    

 TSR                                        12                                     6                           2 


CSU are opting for an active season while TSR calls for a slightly below average (about 15% below any longer term average) Atlantic Hurricane season.  One key factor in determining this is the ENSO (El Nino Southern Oscillation) pattern.  We are currently in a weak La Nina event, evidenced by cooler pacific ocean temperatures, with little model consensus on whether this will either shift to a more neutral pattern over the next few months or to a weak El Nino pattern.   La Nina patterns typically correspond to more active Atlantic hurricane seasons, so a shift to a more neutral pattern would decrease activity linked to an increase in upper-level winds which create high vertical wind shear which restricts hurricanes from forming.  Apart from uncertainty over La Nina, the other main factor is whether stronger than normal trade winds will be present during the July-September period over the hurricane main development region.  Stronger winds result in less vorticity (spinning) and increased vertical wind shear which reduces hurricane frequency and intensity.  These outlooks will be updated in May and once the season has started.

Meanwhile the National Hurricane Centre posted an official blog titled “The State of Hurricane Forecasting Is…” on 9th March.  NHC reports that they are usually more confident on predicting the track of a hurricane compared with the strength and size.  The images below (courtesy of NHC) shows the average NHC track forecast error for tropical storms and hurricanes by decade since 1960. It clearly shows a steady improvement with the average errors in the current decade about 30-40% smaller than they were in the 2000s and about 50% or greater than the 1990s.

RECORD NUMBER OF ICEBERGS OFF NEWFOUNDLAND.....UPDATED 24 APRIL


The North Atlantic is the only ocean in the world where major cross ocean shipping lanes intersect with large concentrations of icebergs leaving little option for vessels other than to detour around them.   The North Atlantic Iceberg service provided by the International Ice Patrol (IIP) runs 1 February to 31 August. 

It takes 1-3 years for an iceberg to appear off Newfoundland after breaking off from a glacier in Greenland. The rate of iceberg formation and appearance in shipping lanes has been increasing in recent years according to IIP: 13 in 2013; 1,546 in 2014; 1,165 in 2015; and 687 in 2016. It looks like 2017 will continue this trend with IIP predicting a fourth consecutive ‘extreme ice season’.  Their latest seven day outlook (issued on 11 April) indicates that some 641 icebergs have drifted south of 48N versus an average of some 85 for late March rising to 212 by end April - based on data collected between 1900 and 2016. This map shows todays North American Ice Service (NAIS) Iceberg Chart with the iceberg limit line reaching down to 41N at around 50W – very close to where the Titanic sank on 15 April 1912.  ​​



Icebergs originate from Greenland and work their way southward from Baffin Bay through the Davis Strait and Labrador Sea driven by a combination of weather patterns and ocean currents.  One estimate is that about 1% (150-300) of the 15,000 to 30,000 icebergs produced annually off Greenland reach the Atlantic Ocean.


Icebergs originate from Greenland and work their way southward from Baffin Bay through the Davis Strait and Labrador Sea driven by a combination of weather patterns and ocean currents.  One estimate is that about 1% (150-300) of the 15,000 to 30,000 icebergs produced annually off Greenland reach the Atlantic Ocean.

Global warming aside, the annual number of icebergs is primarily dependent on the winter and spring weather patterns.   While onshore winds inhibit their formation in winter offshore winds aids iceberg development. The spring weather pattern is key in getting the icebergs moving.  The dramatic increase in icebergs this year is the result of a major storm off Newfoundland in mid-March which helped break up the sea ice.  This was then followed by a massive hurricane force storm in late March which pulled the icebergs south and in to the shipping lanes.   This ‘perfect’ weather pattern in 2017 has resulted in record numbers of icebergs appearing off Newfoundland in the Flemish Pass/Grand Banks regions. These icebergs are expected to continue drifting south and approach the Southern Extreme Iceberg limit for mid-April.


Vessels either need to slow down significantly or make extensive detours around the ice limit adding up to some 350-400nm to a typical cross Atlantic crossing. This latter option is safer and in most cases is the only option but adds an extra day to a typical ocean crossing with associated extra costs. Such a large presence of icebergs requires ship weather routeing specialists to monitor the ice situation closely as routeing further south also places westbound vessels in to the east flowing Gulfstream current so adding to the problem.


Depending on individual cases vessels are likely to opt for a rhumb line route and avoid great circle or flattened great circle routes while this record iceberg presence continues. 


While eastbound voyages can exploit the Gulfstream and North Atlantic drift,  westbound vessels may prefer more southerly routes that avoid both the ice limits and strong Gulfstream adding a further 50-150nm to a normal voyage. Clearly 2017 is going to be a busy iceberg season!      

                                                                                                                                                                                                                                                               Posted 12 April 2017


Updated 21 April:  See left for the latest NAIS chart dated 22 Apr.  Note the greater number of icebergs steadily creeping further south.  Compare the numbers reported vic 42.5N 49.5W between the two charts and one can see a significant increase!  ​This recent image below shows a large iceberg stranded off Ferryland, Newfoundland in a region known as 'Iceberg Alley' - impressive!            


 

EL FARO REPORT UPDATE - WEATHER FOCUS

While the US Coast Guard Marine Board of Investigation (MBI) released its final report on 24 September, 2017 into the sinking of the EL FARO on 1st October, 2015 the US National Transportation Safety Board (NTSB), which is also conducting its own, parallel review of the tragedy, plans to meet on December 12 2017 to vote on its conclusions before releasing its report. 

The MBI 199-page report, which identifies causal factors about the loss of the SS EL FARO, proposes a total of 31 safety and 4 administrative recommendations for future actions to the Commandant of the Coast Guard.  The MBI recommends that the Coast Guard initiate a civil penalty against EL FARO’s operator, TOTE.

The loss of the U.S. flagged cargo vessel EL FARO, along with all 33 crew, resulted in the highest death toll from a U.S. commercial vessel sinking in 40 years.  The EL FARO set sail from Jacksonville, Florida to San Juan on September 29, 2015 at a time when a tropical weather system east of the Bahamas Islands was intensifying in strength.  The storm system evolved into Hurricane Joaquin which proved difficult and problematic for weather models and forecasters to predict in terms of track and intensity. Crucially, the system stubbornly tracked southwest rather than follow a more traditional swing to the northwest. Despite receiving numerous weather updates, the Master directed the ship southward of the normal direct course to San Juan ultimately placing his ship directly in to the path of the hurricane as it moved stubbornly southwest rather than swing northwest in more traditional circumstances (see NTSB image below). Perhaps the Master kept thinking that the system would swing northwest as indicated by earlier predictions, but we will never know for sure. 



The type of season is largely determined by the impact of the El Nino/La Nina cycle (ENSO) and the Atlantic Multi-decadel Oscillation (AMO).  While El Nino tends to suppress Atlantic hurricane activity, mainly by enhancing upper-level wind shear, a La Nina favours more Atlantic activity by reducing wind shear, and AMO is typically associated with a warmer Atlantic and stronger African monsoons.   Currently there is also cooler water off Africa and warmer water in the West Atlantic suggesting systems will develop further west rather than off Cape Verde.


There is little consensus for 2017 so far with views ranging from a slightly below average to a slightly above average season.  There is some consensus for a possible return (50% probability) of El Nino returning from August/September onwards.  The season will start with neutral conditions, i.e. no El Nino/ La Nina, which represents conditions conducive for hurricane development with warmer ocean water temperatures in the western Atlantic and Caribbean region.  However, following the strong El Nino in 2015-2016 and a weak La Nina late 2016/early 2017, there could be a possible return of El Nino later in 2017 either as a hangover from the last strong cycle or linked to the longer lasting pattern of the Pacific Decadel Oscillation which can encourage El Nino conditions.  Given the dominance of ENSO on the Atlantic hurricane season it is too early to know for sure how things will play out in the latter half of the season, hence little consensus exists for now. We will know more after the Spring when model forecasts should be more accurate and there is greater certainty over the   El Nino event.  Watch out for this blog being updated in late April/early May accordingly!                                                   

                                                                                                                                                                                         Posted 5 Apr 17


EVER MORE REGULATION AND EFFICIENCY

Despite the global economy being in better shape than it has been for a number of years, 2016 was one of the most difficult years on record for the shipping industry. Furthermore even with booming stock markets, the prospect of rising oil prices and improving economic news from China, 2017 looks set to be another year with extremely difficult market conditions. As such, the number one priority will remain to focus on minimizing the cost of each voyage.

The pace of regulation also adds to the many challenges, and on top of recent initiatives such as a Ship Energy Efficiency Measurement Plan (SEEMP), the Energy Efficiency Design Index (EEDI) and Emission Control Areas (ECAs), 2017 will see the introduction of the EU Monitoring, Reporting and Verification (MRV) scheme.  This new MRV regulation requires ship owners and operators to annually monitor, report and verify CO2 emissions for vessels equal to or larger than 5,000GT calling at any port of the EU commencing 1 January 2018. 

However, ship owners/operators can start to submit their monitoring plans to EU approved verifiers in 2017 ready for this start date, noting that they also have to start recording data from 31 August 17 even though they do not have to start reporting this formerly until January 2018.

There is therefore an even greater importance being placed on cost saving decisions which, from a weather service provider’s perspective, remain focused on maximising ship fuel efficiency through voyage optimization, vessel performance and providing relevant information in a timely manner to help make optimal cost saving decisions.  

OceanPlus is able to show ship owners/operators or charterers the most efficient, safest route whether seeking efficient running (eco speed), an on-time arrival or using a set voyage/Charter Speed. We provide a cost benefit analysis depending on what priority is set as part of our pre-voyage brief. We also provide a rapid post voyage analysis report showing vessel performance together with a more comprehensive performance summary covering speed made good, slip, SEEMP compliance and emissions data.            Posted 17 Jan 2017

Welcome to our blog page where we publish interesting articles related to the industry.

EL FARO - NTSB EXECUTIVE SUMMARY: CAPTAIN'S FAILURE AND POOR COMPANY SAFETY CULTURE


The NTSB finally released its executive summary on the loss of the El Faro on 12 Dec 2017.  The investigation has so far cost $5.6 million, amassed 30,500 hours of investigative work, and has produced more than 70 findings and 50 plus proposed recommendations. 


The loss was caused by a captain’s failure to avoid sailing into a hurricane despite numerous opportunities to route a safer course to avoid the system.  In short this tragedy was entirely avoidable.

 


​ 


Image from National Hurricane centre


Both IRMA and JOSE will impact on shipping, especially anyone intending to move through the Florida Straits or the Old Bahama Channel later this week/weekend. By example, ships departing gulf coast ports for Europe or the Mediterranean will need to sail south of Cuba and use the Windward Passage between Cuba and Haiti to avoid Hurricane Irma, adding some 400nm plus to their normal journey length.  However it also looks like Hurricane JOSE could also impact by early next week perhaps forcing ships to use the Mona Passage between the Dominican Republic and Puerto Rico instead, adding a further 250-300nm to a typical voyage. There will be a similar impact for shipping headed to the Gulf of Mexico with many either having made those decisions over the last 24-36 hours or monitoring events closely to ensure the optimum decision helped by expert weather service providers.     

With ever greater intensity of weather impacts via either stronger winds and/or greater amounts of rainfall as possible outcomes from global warming, this Hurricane season is looking to be challenging with new records set.                                                Posted 6 Sept 2017


Ocean tropical storm seasons:

  • Atlantic: 1 May to 30 November but cyclones can occur outside these dates, as witnessed by TS Alex in January 2016. 


  • Pacific:  15th May in the East Pacific Ocean and 1st June in the Central Pacific with both ending on 30th November and peaking August to October.  The Northwest Pacific has no official seasonal boundary as typhoons are possible at any time of the year with May to December being the most active period. 


  • Northern Indian Ocean:  cyclone season has no official bounds, but is more active from April to December with two peaks in May and November.  The double peak is due to the Monsoons when upper level winds reduce between the monsoon seasons allowing possible strong cyclone development. 


  • Southern Indian Ocean: late October/early November to May with a less pronounced monsoon driven double peak in January and mid-February/early March. 


  • Australian /Southwest Pacific: late October/early November to May but reaches a single peak in late February/early March.


With the Atlantic hurricane season fast approaching it is worth noting that the Southern Hemisphere is in the midst of one of the quietest seasons on record.  With 2016 being the most active season in the North Atlantic since 2010, attention is now shifting back to this region for the coming season.

 Atlantic

NOAA’s Climate Prediction Centre has recently modified its hurricane season classification to be:


HURRICANES IRMA AND JOSE - POSSIBLE LONG DETOURS FOR SHIPPING

As a Category 5 hurricane (the highest level with sustained winds of 185mph/160 knots with higher gusts) Hurricane IRMA is likely to be the most powerful storm in a decade to hit the Caribbean and SE USA/Florida.   IRMA is expected to remain a Cat 4 or 5 hurricane for the next few days bringing a storm surge of up to 15-20ft above normal high tide levels plus 10-20 inches of rain in more exposed areas. The likely track for this hurricane could not be more life-threatening to this region extending from the Northern Leeward islands along the northern edge of the Greater Antilles and in to southern central Florida.



El Faro and Joaquin tracks: Copyright USCG

WEATHER DATA AND MODELS

The source and timing of warning information and weather data used onboard will be key in helping  to determine what decisions were made and why. It proved to be very problematic to predict Joaquin’s forecast track and intensity during its early development period, particularly the period 28-30 September.  This point was made clear by the messages issued by NHC, Miami and should have acted as the means for even greater caution.  Up until 29 Sep JOAQUIN has been moving slowly SW but was expected to swing westward before recurving to the N.  The main global numerical models were split between a GFS/CONUS more northerly track or a European Centre for Medium range Weather Forecasts (ECMWF) solution that predicted the hurricane moving further southwest and intensifying further before then swinging north and northeast.  In the event the ECMWF model proved more accurate and the NHC shifted from a consensus model output, i.e. one that reflected an answer somewhere between the two model groups, to one that was closer to the ECMWF prediction. The logical conclusion is that perhaps the Master tried to outrun the cyclone working off older information and hence the decision to go to full speed.  Irrespective, the vessel should have been receiving regular warnings from the NHC which would have placed the vessel in the danger area.

Information released by the USCG investigation reveals that El Faro was receiving weather support from AWT but the ship did not have a weather routeing service when it sank. 

The data provided was sourced from NCEP and NHC by AWT and sent to the AWT onboard software ‘Bon Voyage’ for display. The U.S. Coast Guard Marine Board of Investigation heard that the NHC hurricane track sent by AWT to El Faro was provided was 21 hours old[1].  Furthermore, the NHC reported that the forecast for Hurricane Joaquin was wrong early in the storm’s development; thus  eluding to the differences between GFS/CONUS and European models.

The wind and wave data provided by AWT for display onboard was the latest available from the NCEP model suite however, in order to reduce the file size and communications cost, the data is usually not at full temporal or spatial resolution. This means that although the display on Bon Voyage and other software looks convincing, it can also be very misleading and very dangerous. The two diagrams below illustrate this point when comparing wind data at 2 degree to 1 degree resolution once pushed to sea.   Forecasters ashore would have access to much better 0.25 degree resolution data and also would be able to compare many different meteorological models, near real time observations and consult with sources of local expertise.

2016 Tropical Storm Season

The tropical storm season always brings fresh challenges for ship routeing as no one storm is identical to another, hence the adage that each storm makes its own path.  While the Atlantic hurricane season runs 1 May to 30 November cyclones can occur outside these dates as witnessed by TS Alex which formed in January.  The official start to the Pacific season is 15th May in the East Pacific Ocean and 1st June in the Central Pacific with both ending on 30th November.  However, in reality tropical cyclones are possible at any time of the year in the Pacific.  As the season draws to a close we will review the activity in each ocean.

Atlantic

The consensus view from the major Centre’s for the 2016 season was that it would be above average activity with around 12-14 named storms, about 8 hurricanes and three major hurricanes (one that is Category 3 or stronger on the Saffir-Simpson Hurricane Wind Scale).  This is greater than the 30-year historical average of 12 named storms, six hurricanes and three major hurricanes.  This above average activity was mainly due to the impact of La Nina following the strong El Nino last year.  While El Nino tends to suppress Atlantic hurricane activity, principally by enhancing upper-level wind shear, a La Nina favours more Atlantic activity by reducing wind shear.  

The 2016 Atlantic hurricane season (two weeks left) is the most active hurricane season since the 2012 season with 20 tropical depressions, 17 named storms, 11 hurricanes and 5 major hurricanes.  It featured the most cyclones in one season since 2010 and was also the first time that two Category 4 or stronger hurricanes had formed in the month of October, with Nicole and Matthew.  The season started early with TS Alex in January which unusually impacted the Azores.  TS Alex was the first hurricane in January since Alice in 1955 and the earliest to form in January since 1938.  It is also the deadliest season since 2005 with some 2287 people killed.  Matthew was the first Category 5 hurricane to form in the Atlantic since 2007 and was the most southerly to form on record.

Pacific

The Pacific the consensus was for an average plus to above average season. Most Centre’s also revised their expectations for the season and increased the number of named storms to 14-16 and raised the number of Category 3 or higher storms to 3-5 for the North Pacific.  It is simpler to divide the Pacific in to three sub regions:

Central and Eastern Pacific - Hurricanes

Hurricane Pali  which formed on 7 January helped by the then strong El Nino became the earliest Central Pacific tropical cyclone to form on record.  No tropical cyclones occurred in May for the first time since 2011 and no named storms were formed in the month of June since 2007.  Yet despite having a very inactive start, the season has ended up with the record for the most named storms for the month of July, and has been overall an above-average season.

NW Pacific - Typhoons

The western pacific typhoon season lasts the whole year but most cyclones develop between May and October.  The prediction was for an average season and there have been to date some 24 TS, 12 typhoons and 5 super typhoons. The start this year was 26th May making it the 5th slowest start since records began and 2nd slowest season to develop a named storm with Napartak on 3rd July.  The naming of Nepartak ended a 199-day period in which no named storm was active within the NW Pacific basin, which equalled the record set in 1997. The season has been marked by relatively high activity across July-October period. In September, Typhoon Meranti registered as the strongest typhoon in terms of pressure since Megi in 2010, the strongest typhoon in terms of wind speed since Haiyan in 2013, and the strongest tropical cyclone worldwide to date in 2016.

South Pacific Hurricanes – East of 160E.

The season officially runs from November to April, so for 2016 this covers the period 1st November 2016 to April 30, 2017.  However a tropical cyclone could form at any time between 1st July, 2016 and 30th June 30, 2017.  The prediction is for mostly an average to above average season with some 15-17 cyclones of which 3-5 will reach severe TC status.

In terms of 2016, the 2015-2016 season was particularly marked by one of the strongest storms ever recorded in the Southern Hemisphere with tropical cyclone (TC) Winston in February registering 185-mph winds, which brought devastation across Fiji 19-20 February.  Winston was one of the strongest tropical cyclones to make landfall anywhere on earth since the modern era of global records began in 1970, and was the eleventh Category 5 storm in the South Pacific Ocean since reliable records began in 1970. 

OceanPlus and Ship Routeing

Tropical Revolving Storms can move at speeds of 40-50 knots and early avoiding action by shipping is essential. The Regional Specialised Meteorological Centres, which are authorised by the World Meteorological Organisation to issue tropical storm warnings, typically issue warnings 24-36 hours ahead of storm formation. Early action to avoid storms can not only avoid danger but also avoid unnecessary delays and fuel burn.  Weather routeing companies can also issue warnings to ships in advance of regional centres to provide even greater notice.

While safety is clearly paramount, operational savings from routeing result through the combination of reduced transit times, fuel consumption and cargo/hull damage as well as from more efficient scheduling of port/canal arrival time slots.  The earlier the notice for a cyclone or storm then the greater the potential savings possible.

The actual savings in time and fuel will vary depending on the type of vessel, season and ocean but the evidence is that this ranges from 2-4% to as much as 8-10% depending on these factors, so an average of a 4-5% saving is a reasonable assumption for a fleet engaged in cross ocean transits.  Even with current bunker prices being low at around $270/mt there are savings to be made.  An average ship burning 50 tons per day should see savings of $6875 per day on fuel costs alone. Other savings would also result from savings in transit time and reduced insurance costs.  Across a fleet this type of service pays for itself through savings achieved.          

                                                                                                                                                                                         Posted 19 Nov 2016 

                        Figure 2B.  2 x Degree Resolution                                                               Figure 2A.  1 x Degree Resolution


The nice looking computer displays that are becoming prevalent on ship bridges do not necessarily represent a more accurate forecast than one written by an expert marine meteorologist. The trickier or more complex the weather synoptic situation is then the greater the degree of uncertainty and the potential for forecast errors as different models will provide slightly different solutions. NHC made this point clear in its regular warnings.

Despite the improvement in accuracy of weather forecasts today, statistics from the International Union of Marine Insurance (IUMI) indicate that the leading cause of total loss of shipping in every year between 2000 and 2014 was weather.  It is astonishing that the ship loss figure climbed to some 46% of all losses over the period 2010-2014 and leads one to question whether any of this figure is down to misunderstanding, misreading or misinterpreting the weather information?

VALUE ADDED SERVICES

Numerical Weather Prediction (NWP) has significantly improved the accuracy of weather forecasts out to 5 days and beyond.  This has enabled the development of various services for mariners, particularly the shift to the provision of ‘numerical’ data amenable to display on computers, which has subsequently led to the proliferation of weather on bridge displays, onboard routeing systems and as overlays on ECDIS.  Most of this data is passed computer to computer without a human expert in the loop.  Furthermore, there is little or no training to Masters that provides the expertise to recognise the circumstances in which numerical model data is accurate and the times (scenarios / situations) that it is not.  Understanding the nuances of modern NWP data is certainly not the knowledge domain of Masters and busy bridge teams.

Given these shortcomings there needs to be available access to advice from shore based experts, whether this is a marine meteorologist with the knowledge, experience and access to more timely and reliable information to provide superior guidance or senior management being exposed to proposed route planning with the ability to check or query.

[1] http://jacksonville.com/news/metro/2016-05-18/story/storm-track-received-el-faro-day-sinking-was-21-hours-old                                                                                                                                                                                                                                       posted 24 Nov 2016


Each season sets its own records and this season is proving to be no exception.  The destructive power from Hurricanes HARVEY and IRMA has been enormous – both these storms being the first on record to feature 2 x Cat 4 or stronger Atlantic hurricanes making landfall in the United States in one season.  From the mariners’ perspective IRMA became the first Category 5 hurricane to impact the Northern Leeward Islands on record, as well as the joint strongest hurricane ever to make landfall in the Atlantic Basin.  IRMA’s size and track along the Leeward and Greater Antilles and across the Old Bahama Channel, Florida Strait and along the west Florida coast line could not have picked a more dangerous and life threatening trajectory for those both at sea or ashore. 

While many eyes were focused on Hurricane IRMA there was also Hurricane JOSE which passed just north of the Leeward Islands also as a Cat 4 storm.  JOSE has since weakened to a Cat 1 system and is currently undertaking a loop over the waters several hundred miles north of Puerto Rico. This image shows the probability of storm force winds for the period 9-12 Sep and is quite revealing on the degree of potential disruption to the shipping industry. Clearly the sensible option for cross ocean transits to/from Europe or the Mediterranean from the Gulf of Mexico is to use either the Windward or Mona passages to avoid IMRA and JOSE  - adding some 600-700nm to a standard voyage.  This is busy times for ship routing specialists.



Infrared satellite image of Katia, Irma, and Jose as of 1530 (11:30 am EDT) Wednesday, September 6, 2017. Image credit: NASA/MSFC Earth Science Branch.


The season officially runs 1st June 1 to 30th November which is the period of the year when most tropical cyclones form in the Atlantic basin but their formation is possible at other times of the year.  2017 marked the third consecutive year to feature a pre-season storm with Tropical Storm Arlene in April. 

It is worth noting that the Atlantic has the largest year-to-year variability of any of the global tropical cyclone basins.  The 2017 season was not unusual and the above normal activity was not the result of global warming as some media would suggest.

All ten of the season's hurricanes occurred in a row, reflecting the greatest number of consecutive hurricanes recorded since the 1960s. This season is also one of only six years on record to feature multiple Category 5 hurricanes (Irma, Maria and Harvey). Significantly, Irma was the strongest hurricane ever recorded to form in the Atlantic Ocean outside of the Gulf of Mexico and Caribbean Sea. Hurricane Ophelia was the most easterly that a major hurricane has ever formed in the Atlantic since 1851 and strengthened to a Cat 3 system as it passed near the Azores archipelago on 14th October headed for Ireland/UK.

2017 is the only season on record in which three hurricanes each had an Accumulated Cyclone Energy (ACE) score of over 40: Irma, Jose and Maria (ACE is used to express the activity of individual tropical cyclones and entire tropical cyclone seasons (total ACE score is 223 which is in the hyperactive category and the 7th highest score since 1851)). 

Initial predictions for the season anticipated a transition from a neutral to a weak El Nino-Southern Oscillation (ENSO) pattern. The consensus was therefore for an average season (10-12 named storms, 5-6 hurricanes and 2-3 major hurricanes).  However, as the season approached the North Atlantic Oscillation tended towards a negative phase, favouring a warmer tropical Atlantic and thus lowering the chances of El Niño forming and instead transitioned to a weak La Nina which brings weaker vertical wind shear to the main Atlantic tropical storm breeding regions in late August and September.  As a result most forecast centres updated their forecasts to a more active season, and by August their predictions shifted to reflect an extremely active – if not hyperactive – season.  By example, back in May, the National Oceanic and Atmospheric Administration (NOAA) predicted 11-17 named tropical systems but raised this to 14-19 systems in August, and suggested that 2017 could be the most active since 2010.  The UKMO updated their prediction to 10 hurricanes on 1st June.  Some centres, such as the Tropical Storm Risk consortium, initially did predict an above normal season, with a total of 14 named storms, 6 hurricanes and 3 major hurricanes.  Their early prediction was based on the fact that Atlantic low-level winds would be near their long-term average and that they did not anticipate the development of El Niño. 

In terms of peak wind intensity the top 5 hurricanes were:


  1. IRMA: 30 Aug-11 Sep, 160 kts
  2. MARIA:  16-30 Sep, 150 kts
  3. JOSE:  5-22Sep, 135 kts
  4. HARVEY:  17-31 Aug, 115 kts
  5. LEE: 15-30 Sep, and OPHELIA:  9-16 Oct,  100 Knts


There were several periods when consecutive cyclones resulted in three active systems being present in the region at one time: Katia, Irma and Jose in early September (see the IR satellite image for 6th September); and Jose, Maria and Lee in mid-September. While these systems caused havoc and significant destruction ashore they also caused a lot of shipping to make detours. These were clearly busy periods for those involved with optimum ship routing where some voyages had to make significant 300-500nm detours in order to avoid successive systems. Significantly, there were no major disasters at sea this hurricane season and hopefully past tragedies, such as the loss of the El Faro in Hurricane Joaquin on 1st October 2015, will never be repeated.

​                                                                                                                                                                    Posted 27 Nov 2017

Oceanplus

Marine Weather Forecasts

Innovative Marine Weather Forecast Services


MARINERS: BE AWARE WHY NUMERICAL WEATHER PREDICTION (NWP) MODELS DIFFER IN ACCURACY

INTRODUCTION

There are many reasons why there are differences between Numerical Weather Prediction (NWP) models so this short article will briefly cover – in non-technological language - where and how this happens.   The top suite of global models comprises ECWMF, UKMO, GFS, CMC and JMA but there are others knocking on the door.  With some exceptions, they all essentially have similar physics, resolution and data assimilation schemes.  The latter enables a vast and growing range of observations, such as satellite and radar data, to be ingested for initial conditions essential for more accurate forecasts.  

BACKGROUND

NWP uses a mathematical model of the atmosphere which has been derived from the laws of physics and dynamics of the atmosphere. It is produced on a three-dimensional grid with levels extending from just above the earth’s surface up to the stratosphere (80km). Predicting the weather therefore requires a set of equations to be resolved for each grid point and height for a set time.  NWP focuses on taking current observations of weather, particularly data assimilation from space, and processes these data to forecast the future state of weather. Knowing the current state of the weather (at every grid point and height level) is just as important as the model algorithms processing the data. These equations are impossible to solve exactly.

GLOBAL VERSUS REGIONAL SCALE

The resolution of the grid needs to capture the weather features but the size of the grid is limited by how quickly these equations can be solved for a forecast to be issued in a timely manner.  The numbers involved are simply huge and there is never sufficient computing power to achieve optimum results, so there is a trade-off between accuracy (resolution / time steps) and computing time.   Greater accuracy (or more detailed forecasts) comes with higher resolution in order to pick up the smaller (meso) scale details, and this is done using regional scale models who use the output from global models (commonly termed boundary conditions) as part of their initialisation.  The same principle applies to other specialist models such as tropical storm models, oceanographic models and ensemble prediction systems.  So for example; GFS feeds MM5, NAM, WRF and HWRF; UKMO global configuration feeds the Unified Model System which has nested higher resolution models; and ECWMF feeds European Limited Area Model (LAM) co-operative projects such as ALADIN, HIRLAM (France) and COSMO (Germany).   While global models are designed to capture the key weather patterns and systems they are also tuned to areas of prime interest depending on who owns them.

MODEL INACCURACIES 

Simply put model errors occur through a series of inter-related compromises:


  • Computing power.  This is mostly about the lack of computer power resulting in compromises between high resolution, area coverage or forecast length.  Improving computing power and models is an ongoing process set within finite budgets. The atmosphere remains too big a problem to solve correctly and quick enough even with the latest technology.


  •  Initialisation data.  Given the chaotic nature of the Earth’s atmosphere it is essential to know its current state in order to predict its future state.  It is not possible to know this precisely.  Despite terrestrial and space based observing systems, there remain data voids, especially over oceans and in the upper troposphere and stratosphere, so various schemes are used to either approximate or omit this data.  Very small errors in the initial state can lead to large errors over time, so it is not possible to create a perfect forecast and there is a limit to how far ahead one can predict any detail.  While more data improves accuracy it also slows the initialisation process so the greater the computing power is needed in order to generate a forecast capable of being used in a timely fashion.      


  •  Software errors.  All software has bugs and the more complex the programmes then the more complex the bugs!


  • Model Bias.  Each model has been developed differently and fine tuned, so each model will have its own bias regarding weather systems and patterns.  Every step in NWP involves a combination of omissions, estimations, approximations and compromises which introduce errors; from resolution, to height levels, to forecast lengths and coverage, and to the mathematics, physics and atmospheric dynamics used.  There are many complex processes in the atmosphere that need to be parameterised such as radiation, reflection (albedo), convection, turbulence et al.  No one global model is perfect, but some will be more accurate or show higher skills than others for any given situation.  There has been recent media coverage in the USA of ECWMF being better than GFS over the high profile Hurricanes’ Sandy and Joaquin.  In reality there have been equally as many occasions when either GFS, UKMO or other models have been superior. The simple truth is that currently no one model fits all, so pushing single model data to sea has inherent consequences which are not necessarily understood by end users.


SO HOW DOES THIS IMPACT ON USERS OF NWP DATA AT SEA?

 The current trend to push NWP data to sea for use or display on bridge systems places additional pressure on already busy bridge teams, who may not have had any training on the nuances of NWP data.  So here is some quick practical guidance:


  • Make sure that you know what model is being used to produce the weather forecast, overlay on the bridge system or the routeing service.  While GFS and CMC are free, the most common model is GFS, while others such as ECWMF, UKMO and JMA are only available via a fee or subscription service typically provided either in-house or via a specialist Weather Service Provider (WSP).


  •  Use more than one model if possible as there is no such thing as one model being the most accurate above all others in all-weather scenarios.  If no other NWP data is available for comparison then check via other sources, e.g. GMDSS forecasts via VHF/NAVTEX, tropical storm warning centres or a WSP which all use professional meteorologists experienced in interpreting NWP data from many models.


  •  Look at model trends.  If using only a single model then compare previous model runs to check for consistency. This is best done using 00Z & 12Z runs which tend to be more accurate as 06Z & 18Z model runs use less weather observational data. Any jump/key changes in model runs should be treated with caution until some degree of consistency is present.


  •  Model accuracy decreases with increasing time and lower resolution. Generally detailed forecasts extend out to 5 days but there can be significant differences between models even at 4-5 days range.  Forecasts become more general out to 14-16 days, i.e. the wind direction will generally be NW but the strength could be very different.  Models have little skill beyond 16 days.


  •  Resolution – accuracy decreases with decreasing resolution.  High resolution requires higher volumes of data, but the combination of improving data compression ratios and rapidly increasing satellite bandwidth availability at lower costs is mitigating this aspect.


  • Probability and Ensembles.  Ensemble forecasts are multiple forecasts made by making small changes to the starting conditions of a NWP model for the same time.  Ensembles allow the uncertainty of a forecast to be quantified (as a probability) and the risk of a particular weather event to be assessed, i.e. 30% risk of rain or a tropical storm forming in three days’ time.  If the forecasts vary a lot then there is more uncertainty but if the forecasts remain similar then there will be more confidence in the prediction.  


  • Passage Planning.  Use global models which are designed to produce medium to long range forecasts.  Compare weather patterns with climatology beyond 16 days as a sanity and confidence check.    


  • Extreme Weather.  All global models either embed output from specialist or use nested models for tropical revolving storm forecasts. It is important to note that a forecast track issued by a specialist tropical revolving storm centre will inevitably be different to that from a single global model, so it is essential to use updated warnings issued by regional centres as this will reflect a consensus between a number of latest models and observations.  This is perhaps one of the concerns associated with the El Faro loss (reported by the NTSB on 13 Dec 16) when the Master reported on “..conflicting reports as to where the center of the storm was”.   


  • Last but not least – single observer forecasting technique remains very useful in trying to observe changes in the weather and why they may or may not be changing. Satellite imagery and observations if available also help supplement and verify NWP forecasts.  By example, there would have been little change in bearing of TS Joaquin with the El Faro’s track yet the wind strength would have steadily increased however a malfunctioning anemometer can be a crucial loss of information especially at night  ​

SUMMARY

The ECMWF model is currently the most skilled global model and has the highest resolution of 0.125 degrees.  There are many other global models close behind with not much difference between them.  Most are fine tuned to particular ‘home’ regions.  However, which model does best on any particular occasion around the globe can vary as each has its own strengths and weaknesses.  In global terms no one model is supreme and they are all prone to differing errors which get greater with time.  Use of a single model is likely to result therefore in over focus and reliance at the expense of other available information.  NWP data should be regarded as a complementary forecasting tool for meteorologists or experienced forecasters to interpret, and be used alongside other sources of information rather than be used in a standalone format for any decision making weather prediction process.  If using single model NWP at sea then it is essential to ensure access to some quality control assurance either via a WSP or other sources.

Recommended Reading: Numerical Weather prediction – A Practical Guide for Mariners  by Huw Davies and published by the Nautical Institute. ISBN 978 1 906915 40 7

​                                                                                                                                                                                        Posted 7 Jan 2017