We've put together several low-bandwidth sources for cruisers. Many are available through saildocs and some through our site.
InReach And Short email message Service – Garmin and short messages
Mexico and Central American Wind/Wave Weather sources – via email as text
Alaska/US/Pacific Tsunami Alerts – via email as text
World-Wide Ocean Surface Currents – via email as a grib file
Translated French Polynesian Weather – via email as text
You can retrieve the data by email using the following steps.
Pick your region. In this example, the outside of Baja Mexico or "send outsidebajaforecast"
Send email to email@example.com with and empty subject and "send outsidebajaforecast" as the message
After some time you should receive the forecast in your email inbox.
There are a variety of things you can do with saildocs, like subscribe to forecasts or setup a schedule to deliver the information as well. To learn more visit their site.
List of Regions with links to online forecast
mexicocomposite = This covers all Mexico and Central America
outsidebajaforecast = The exposed western coast of Baja
socforecast = Sea of Cortez forecast
southerncrossingforecast = Sea of Cortez Southern Crossing Forecast along 23N
maztopvforecast = coastal area of Mazatlan to Puerto Vallarta
cabocorrientesforecast = Cabo Corrientes to Manzanillo
mexicanrivieraforecast = Mexican Riviera of Manzanillo to Bahias de Huatulco
tpecforecast = Tehuantepec Forecast
papagayofcst = Central America Forecast with emphasis on ITCZ and Gap Winds
tsunami = compilation of Alaska/West Coast and Pacific Tsunami Alerts
The following is a list of all the catalog names for copy and pasting into your email program:
send mexicocomposite send outsidebajaforecast send socforecast send southerncrossingforecast send maztopvforecast send cabocorrientesforecast send mexicanrivieraforecast send tpecforecast send papagayofcst send tsunami
Background of Models
After spending a lot of time cruising Latin America we found we were using weather sources from multiple forecasts just to gather a few critical pieces of data. This was a frustrating and a time consuming process. So to make things easier I decided to write an automated script that Stan from Solmate Santiago could run and hopefully generate support for his non-profit's free spay and neuter clinic in Mexico. See his websites for donations and information. PATA Manzanillo Animal Welfare and the corresponding USA Friends of Mexican Animal Welfare Website to donate.
In 2016 Stan decided he did not want maintain the weather reports any more, so now I'm running them here in my server. And along with them I thought I should share some of the background behind how they work.
The “Warning” text is compiled from high seas and tropical discussion:
It is sorted by a set of search terms for each geographic region and is separated to try to keep all warnings geographically isolated. The Mexico Composite which is usually only used by a few net controllers is just a compilation of all the individual forecasts so the warnings are not grouped together but are given by location. I know it can be repetitive, but the composite was added as an after thought to accommodate the SSB radio nets. I've done some recent modification to try to remove redundancy but you might find it repetitive at times.
Wind Field Estimations
The “Sarana weather models” (for lack of a better term) will always tend to error on the higher side of the forecast. This means that winds you experience are more likely to be less strength than forecast. Areas where gap winds occur should be considered exempt from this general claim however as gap wind velocities are hard to forecast. Let's go into more detail.
Numerical data for wind and waves is based on GRADS software and data from NOAA's WAVEWATCHIII wave models, NOAA's Global Forecast System (GFS) and North American Mesoscale (NAM). This is slightly different from the standard GFS on model you get from GRIBS alone. The method is similar you what you find at Buoyweather. This data is produced in 6 hour intervals, so the code I wrote does some sorting to look for the worst case scenario over the forecast time interval and reports that as the max. Thus the errors that occur will most likely be conditions that are less severe.
Since most cruisers are looking for comfort rather than fast/rough sailing conditions, the reports are designed to bias towards reporting any higher conditions that might occur. You should find accuracy errors should be "less than forecasted" more often than "more than forecasted". That said, this is NOAA data which is averaging over large areas and local conditions can be significantly higher like in the Gap Wind areas. Orographic effects (land effects) are not included in the data, but the warnings often capture these events and should be clues that the winds might be significantly higher -- especially in the Tehuantepec and Papagallo Region.
To get this data send a message to firstname.lastname@example.org and put "send tsunami" in the message. On line you can see the forecast here
On rare occasions a tsunami can cause several problems for sailors. Indulge me while I tell a personal story, like sailors love to do. We were personally caught in a tsunami in Huatulco, Mexico which was caused by a large earthquake in Chile. I happened by chance to hear a report from someone over the SSB net that there could be a possible tsunami. At the time we had no internet and had no idea what to expect. I called the port captain over the VHF and he said there were no warnings issued for the coast. I got back on the SSB and listened to reports from the Galapagos which was almost in between us and Chile. They had 3 foot surges and current but no waves.
In our shallow spot where we touched bottom at extreme low tides, I quickly looked at the tides would be high when it came. This didn't seem good in such shallow water, but with the high tide we stayed. The Port Captain also indicated again after I visited in person that there was no change in the warnings.
Needless to say the boat basin by the port captain was devastated and sunk pangas, a cruise ship that was docking hit bottom and the Port Captain ordered them and everyone to depart to deep water immediately and closed the ports. The marina area where we were in had massive in-rushes of current. We were in the cheap side canal and it rose almost 4 feet then fell. This happened for almost 16 hours getting less and less over time. But the marina area had enough volume to absorb the surges unlike the commercial boat basin.
TL;DR; Point of the rambling story? Well, I've wanted to provide a tsunami alert mechanism for a long time, and I've finally done it.
All weather forecasts will provide a notice if there has been an update from either the PTCW - Tsunami warnings for Pacific Ocean or from NWS West Coast/Alaska Tsunami Center Bulletin. It will look like this at the top of the weather forecast:
Check Tsunami Warnings US West Coast/Alaska--Update 1 days
Request 'send tsunami' from saildocs.
If either Tsunami center has issued an update within the past 1.5 days, there will be a notification indicating there is a new bulletin. Many of them will not produce tsunamis and many won't be in your area, but now there is a mechanism to provide some warning. Details can be retrieved by sending an email to email@example.com with “send tsunami” in the message body.
Weather Model Corrections and Lack of Validation
Due to lack of consistent real-time feed back to NOAA real errors on these models are unknown. The loss of the QuikSat system (the best meteorological satellite system to date) in 2009 has made Gap Wind forecasting more inaccurate and the models are corrected less often (corrections made with Europe's ASCAT system). Since 2016 the ISS-RapidScat is operational and we can all hope the models will improve in accuracy in remote locations. See this NASA JPL video for a brief explanation.
Weather Report Organization
The layout of the forecast data isn't of my creation. It exactly mirror's Don Andersen's format that he followed for decades. The warnings are added to provide some information from the forecasters themselves to preface the numbers.
The Sea of Cortez is such a small area that trying to sub-divide it further into N/Central/S will probably not produce anything meaningful. But I'm open to the idea as we can easily add another virtual buoy in the North section and the South section.
Long term forecasts are important for spotting trends in building weather systems. Calm conditions will produce larger long term errors, but if you are leaving PV for San Carlos and day 5 shows a building norther, then you might rethink your options as you sail and watch the trend over the next few days. Having sailed too many years now I would not give up knowing the long term estimations just because it might change. And when you report "It's going to blow 30 for the next 3 days." The first question will be "When does it look like it will back down?"
In regards to NOAA's new forecasts: My forecasts which are post processed some for cruisers should mimic the NOAA's automatic forecasts. As long as the new data is added into the high seas forecast my key word capture technique should also pickup any additional data.
Models and Unpredictability and Making Your Own Decisions
I'm not a meteorologist and these forecasts are all computer compiled. The weather is not very predictable to the detail many of us would like. There are no weather buoys to refine the models or produce historical error analysis so you'll have to live with inaccuracy. Trying to improve the forecasts is somewhat futile without calibrated feedback. NOAA lost their best tool for wind prediction in remote areas: QuickSCAT. Now they are cobbling together a new QuickSCAT system on the ISS and using the European's slower with less resolution QuickSCAT system. This was this only way they have to check their models in remote places -- from space. This system is now on-line in 2016 so forecasts should again become more accurate.
Man vs. Machine
At times forecasters will disagree with the computers results. The text you read in the warnings are from meteorologists and should override the number based sections of the forecast. The warnings are written by forecasters who have reviewed multiple models and apply their knowledge to the data. They tend to bias their numbers higher and offer less detail in changes (e.g. “dropping 20 or less”). In the our forecasts you get both pictures together the forecasters (warnings/discussion) and the computer (wind/wave).
NOAA's East Pacific Offshore Forecast
Differences between our model and NOAA's new NOAA's East Pacific Offshore Forecast (EPOF) data:
EPOF is forecaster reviewed until at least July 20, 2018.
No tsunami notifications
Different wind/wave model
Inclusion of ITCZ data for people near the equator.
EPOF has smaller coverage areas
We have picked custom wind/wave virtual buoys using personal experience from those areas
IMPORTANT: If you want the forecasters to continue to provide reviewed forecast details past July 20, 2018, you must let them know by providing feedback to NOAA.
As of July 2016 the forecasts are all reviewed and issued 4 times a day NOAA's East Pacific Offshore Forecast (EPOF) so you get their computer results with some warnings data from forecasters. What the computer spits out is based on an algorithm to reduce a massive data set down to a tiny area of the planet. While this wide variety of weather slices might at first glance seem more accurate, this is unknown because this data is not being verified by on-site instrumentation. Often cutting the slices of same bread thinner just reveals more holes.
It would be rare for my numerical model to agree 100% with NOAA's EPOF as there will be different criteria and models involved. Likewise the forecasters often produce numbers that are different then their computers. You can find the two sources for EPOF data here:
Lacking from the EPOFF is the Inter-Tropical Convergence Zone (ITCZ) information which is a very important driver of weather for anyone in Central or South America.
EPOF data is not necessarily taken from a single model but are evaluated on a case-by-case bias. And while they may appear similar to my model outputs they are not machine produced, only machine formatted for consistency.
Practice Your Own Forecasting
As an example, say you want to do a 6-8 hour trip in the next day or so in the Sea of Cortez and we see the following scenario:
Our numerical forecast shows 10-15 NW during the next two days and dropping at night in between with some light S winds. Long term (2+days) shows light and variable directions.
But the warnings text from forecasters show solid NW 20-25 then 20 or less after 2 days.
The GFS gribs show wind fields in the gribs or pull numbers from EPOF and look at the trend in the whole Sea of Cortez. This data shows 15-20 with some light spots with a long term (2+ days) outlook term that winds will switch SW 5-10.
Think about it.
NW is usually a good direction in the Sea of Cortez for thermal enhancements which appear in the afternoon and might not show up in the computer forecasts but the forecasters 20-25 might be reflecting this effect. What are the winds like now? Has it been clear so the land can heat? Are we getting thermals now? Are they stronger at night? Or is the trend really decreasing?
Practice this and watch. Stay anchored and watch and listen to what others report over the SSB nets. Did it drop at night and build the next day? Did it blow all night? Did it just keep on dropping? How can I use this information to understand the next forecast better? Is my local area different from other?
What about that switch to the SW? Is a pressure gradient going away? Is this a strong outside Baja NW wrap-around that happens sometimes? Will it really switch or just go calm? The fact that it is predicted to switch probably means the N winds will at least be less.
And that's about the best you can hope for. Just keep testing the models against what you learn about the area. Most people who arrive in a remote area expect to get the same detailed weather forecasts they are used to in the USA because they take it for granted what goes into refining the data. Yet now they are suddenly paying attention and analyzing the weather unlike they have never done before and finding problems. Many people don't understand that it can blow 30 where they are and 0 somewhere else and the prediction of 15 would be accurate for the area. Just don't expect forecasts to be much more accurate just because they are producing more and smaller slices especially if they are all computer generated.
If you are looking for information about getting ocean current forecasts over email then see our OSCAR information page. This information is very important for those making equatorial crossings or long ocean passages.
Recently cruisers (mostly Chuck on Jacaranda) asked me to translate the French weather forecasts. Since I don't speak French, naturally I said, “Pas de problème” (which google tells me means no problem). I quickly found that everyone who runs translation services, like google, wants to get paid to do it for websites. So, I being low budget rolled my own translation program. For details visit this page.
Weather has its own semantics so it turned out technical text like forecasts translate pretty well because there is not much ambiguity in the language used. For those heading to Marquesas and other islands you can request the forecasts from firstname.lastname@example.org using the following in the message body:
You can also view them on-line too:
http://svsarana.com/translate/fr.poly.en.php – this is the source for the fr.poly.en and it is a detailed forecast
http://svsarana.com/translate/fr.poly.short.en.php – this is the source for fr.poly.short.en which is a short term trend overview. It also doesn't translate as well because it is less technical and uses wider variations of language.
http://svsarana.com/translate/fr.poly.long.en.php – this is the source for fr.poly.long.en which is a longer term trend overview. It also doesn't translate as well because it is less technical and uses wider variations of language.
My broken English version is readable and I've gotten feedback that it has been quite helpful. Give it a try. If by chance you speak French and would like to help with the translation dictionary, you can view the full dictionary details and email me updates if you want to help by viewing this page: