@Alla_Ukraine We have implemented an update ensuring that warnings are displayed only when they fall within the formally recognized national borders of the issuing agency. UA feed will be soon processed from meteoalarm.org, it was not present there at the time of our last implementation.

@tank-Uttam It’s difficult to comment on this specific case. We are using the Nowcast.de global lightning network for lightning visualization, as it is an industry-leading provider of such data. However, there may still be some false positives, as no detection method is 100% accurate. While it could be a ‘bolt from the blue,’ that seems unlikely here, as the nearest clouds are quite far away. Another possible explanation is incorrect cloud detection from the satellite in the nighttime image.

Thank you all for reporting this, it should be fixed now.

@gkikas-lgpz thanks for info. I will have to check it in more detail, just few quick remarks:

• Germany - we know about it, their alert feed doesn't correspond to what is a "standard" for other EU countries. separate parser will be developed for that
• general disrepancies between Windy/meteoalarm - it already happened many times, that meteoalarm shows different alerts from the ones reported in the data feed. we found a cases where alert was not visually shown in the metoalarm map, although it was present in data feed and was valid and active in given time period. Same happened vice versa - alert shown in map was not present in data feed.
• also the understanding of how to treat today/tomorrow differs between Windy and meteoalarm

We plant to fix/change couple of things in CAP alerts layer, once this is done, I will provide more detailed post of how it is implemented and how to treat the data provided by that layer.

In the meantime, if you really want to compare Windy against the source, you have to check the data source itself, not the resulting map at meteoalarm (example for Sweden can be found at http://meteoalarm.eu/ATOM/SE.xml). If you find a discrepancy there, please let us know

We, surfers, waveriders, divers and sailors need an accurate tide forecast. Tides can only be forecasted at "tide stations" that measure actual sea levels in time. We estimate, that there is 15 - 30.000 tide stations in the world, yet we at Windy, were able to gather data just from 10.000 of them.

Let's create a creative commons database, where the community could submit tide stations in the world, and upload the (time to time) latest measured data. The open source web application for gathering these data would calculate harmonic files, and these files would be available for free for anyone in the world.

Not only would we be able to surf, sail or dive better, but we would increase overall marine safety.

We believe that this information should be free.

Ivo, the founder of Windy

---

Technical part from @kekert who programmed tide backend on Windy:

---

What is needed?

In order to be able to accurately predict tides for given location, set of harmonic constituents derived from water data levels are necessary. Either harmonic constituents itself (usually hard to get) or water level observations (can be taken from local boating magazine, chartbook, yacht club or marine authority) are sufficient. We believe that there is a widespread demand for free, creative commons database of harmonic constituents which can be used to predict tides. We believe that this creative commons database, can help other people dependent on tide levels to avoid all dangers and enjoy all possible fun using these shared information.

In our eyes, any information that can provide more safety to people should be always free to use. Ever.

How to achieve it?

If you are a programmer, wanting to help other tide freaks, we would be happy if you decide to develop open source web application for adding, editing, parsing and transforming tidal data. Let’s call this process tidal data munching.

• You will develop web application for munching tide data using the technology you like. The codes of app should be, in our opinion, open source and hosted on gitHub.
• Application users should be able to add/edit either harmonic constituents itself, or water level observation (at least a year's worth of hourly water level measurements is recommended). As output, database of harmonic constituents for tide locations should be provided.
• The condition is, that resulting harmonic constituents database is creative commons and it is easy for any other party to download and use the database.
• Windy will provide free hosting for this apps with nice domain name tides.windy.com
• As start, free to use data of harmonic constituents from The Center for Operational Oceanographic Products and Services (CO-OPS) can be used - https://tidesandcurrents.noaa.gov/stations.html?type=Harmonic+Constituents
• Windy will place a links from www.windy.com to tides.windy.com so we can guarantee that this site will have its users and spots contributors.
• Windy will enable you to use registering and authentication of windy.com so you do not need to develop it yourself.
• Windy will send you Windy cap and Windy towel (so far only swag we have now).

Data structure and recommended setup

The database should contain the following information:

• Location name + basic description
• Coordinates (latitude, longitude)
• Harmonic constituents for given location

We recommend to use the components described at http://www.flaterco.com/xtide/harmonics.html:

• Resulting data should follow the structure of Harmbase 2 (or at least the output format should be easily convertible to the one required by Harmbase 2) http://www.flaterco.com/xtide/files.html#harmbase
• Harmgen package should be used to derive harmonic constituents from water level observations http://www.flaterco.com/xtide/files.html#harmgen
  the reason for that is that a lot of complicated work have been done on these packages and we don’t see any other viable way of generation tide predictions at the moment.

More detailed description of recommended process:

1. collect user data:
   • either in the form of harmonic constituents
   • or as water level observations
2. clean data, prioritize, remove duplicities
3. transform data to output format:
   • for harmonic constituents, trnasform them to desired output format
   • for water level observations:
     1. transform data to the format understandable by http://www.flaterco.com/xtide/files.html#harmgen
     2. Harmgen produces results in the form of an SQL insert statement that loads the new station into Harmbase 2

Example of harmonic constituents data for single location from CO-OPS:
https://tidesandcurrents.noaa.gov/harcon.html?unit=0&timezone=0&id=9411340&name=Santa+Barbara&state=CA


A small warning at the end

A lot of low level programming and tweaking (especially in C) will be needed to perform this task.

Quote taken from Xtide webpage:
“All pretense of user-friendliness stops here.  If you want to do large numbers of stations without lots of manual data entry, you have two options, both of which require a high level of computer literacy.”

Taken from HarmBase readme:
“Harmbase is a database application for managing harmonic constants. I built Harmbase to solve my own problems, so it's not "plug and play" software. It's not designed to be useful to someone who does not know how to hack C code or write interactive SQL queries. Given the nature of the problem--turning unstructured data from random sources into structured data--I doubt that anything but hackware would be feasible.”

How to proceed

Just let us know that you are interested and we can go!

---

Wave tide forecast on Windy? Where will I find it?

We are testing new tide forecast right now. Jest check the details here: https://community.windy.com/topic/4841/help-us-to-test-new-tide-forecast-experimental-feature

@krchnavy there is such detail, just use the zoom in lower right corner

@idefix37 @Simoneto-J we are sourcing the alerts from https://gdacs.org/. As you can see on their website, "our" areas correspond to what they release. Do you think these alerts are irrelevant? What we should work on is the area description.

Thanks for the info, this is obviously a bug. Will be fixed soon.

@tomber42
German feed had different structure, now it was added too.

it should be fixed by now

@tank-Uttam It’s difficult to comment on this specific case. We are using the Nowcast.de global lightning network for lightning visualization, as it is an industry-leading provider of such data. However, there may still be some false positives, as no detection method is 100% accurate. While it could be a ‘bolt from the blue,’ that seems unlikely here, as the nearest clouds are quite far away. Another possible explanation is incorrect cloud detection from the satellite in the nighttime image.

@idefix37 @Phainopepla one comment here, the fire danger layer will be improved in the next version with 1h steps and wind particles overlay (to see much better potential spread).

@mhaberler unfortunately the same T&C apply for other models, not only ECMWF. There is a definitive answer in our T&C: https://account.windy.com/agreements/windy-terms-of-use (see section 4.6 Prohibited activities).

A solution to this is to develop a Windy plugin which could be used, instead of exporting the data elsewhere. Otherwise, you are just looking for a data API, where our only offering atm is at api.windy.com

We are currently experiencing issues with our tides forecast due to outages from our tide data provider.

Dear @sdamaster, thank you for reporting this. You are right; the location shown is incorrect. We are now close to releasing the new solution for processing stations and observations into Windy.com. After checking the new solution, the location appears to be correct there. Please stay tuned, it should be fixed within a few weeks.

Dear @davemyron, we had a slightly outdated database of regions, which caused some to not be processed. It is fixed now. Thank you for reporting it.

@Alla_Ukraine We have implemented an update ensuring that warnings are displayed only when they fall within the formally recognized national borders of the issuing agency. UA feed will be soon processed from meteoalarm.org, it was not present there at the time of our last implementation.

@idefix37 @Simoneto-J we are sourcing the alerts from https://gdacs.org/. As you can see on their website, "our" areas correspond to what they release. Do you think these alerts are irrelevant? What we should work on is the area description.

Hello @Simoneto-J, thank you for reporting this, it really seems strange, we will inspect it. Usually we get the polygon shape from the source, I guess it is the case for South America too...

hello @UAHiker, as @btschumy mentioned, only warnings from https://alerts.weather.gov/cap/us.php?x=1 are shown in US. also, only moderate, severe and extreme warnings are shown (minor are filtered out).