@Suty

I'm not a fan of the circular icons. I cannot see what they are at a glance. The Icons are meaningless to me. I now have to hover a mouse over them just to get a text popup that tells me what they are. This is not a utilization addition, it's a functional down-grade that makes things slower and more awkward. I'm not going to grow to like this. It just sucks to use it like this. Text back please.

@Sinewave1

Some models show Typhoon Gaemi decaying near to the "Three Gorges Dams" which are currently in persistent flood. This would force all flood gates to open at once. Looking at surrounding pressure systems it can happen.

There will be no easing of the flooding in China - next 10-days

@Suty

Thanks, an example of what's occurring tonight.

The 'Accumulative' sliders in V43 are a big improvement, been wanting that for a long time - thank you.

Also: Currently, I have low-level NE onshore showers (that are developing into rain-bands), plus a monsoonal flow around a Low, bringing mid-level showers down from the NNW, over the top of them.

They cross each other at about 90 degrees, and at a similar movement speed.

So the prediction algorithm just spits the difference between them, and predicts they'll all move SSW!

No, none of these showers move SSW! They of course continue on their original headings. So this prediction is quite useless to me, it just averages the NET movement vector, rather than predicting movement and speeds of individual features, based on past movement and speed.

Plus it plays that 'prediction' first, instead of the logged data animation, which is actually predicatively very useful.

It's not a good feature.

Well that was unexpected. Eye-wall kissed the coast within about 1 km of crossing, then went south parallel to it, then looped back eastwards, out to sea again.

You don't see that every day.

https://www.windy.com/-Hurricane-tracker/hurricanes/gaemi?radar,24.199,122.498,8

@Suty

Am I mistaken, or is the prediction algorithm working better in the past few hours?

It seems to be doing a much better job of predicting individual rain feature vectors.

Before it was sending all the features in the same direction, but in this animation sequence I just made it is projecting all the rain features moving in different directions, consistent with their prior logged movement and speed vector.

If so, that's working properly (or at least it's a lot better).

Rapidly regaining core symmetry on radar, before a second attempt to cross.

Open eye is 10.2 km across.
Rain bands are 600 km across.
Cloud-mass is 3,050 km across.

@Suty

I don't mind it staying in the radar animation if it works properly, and does not mess up the end of the animation.

The prediction function appears to be working much better today, compared to how it has been working prior.

Others can test to confirm.

Note that the flickering signal at the bottom of this animation is the Longreach radar operating intermittently.

@Suty

Ok, thank you.

Would it be possible to invert the text on the planner's waypoint distances?

They're difficult to read at times. For example:

@Suty

Ok, thank you Suty.

All three "Extreme Forecast" custom layers are not restoring, they stay at their defaults.

What it should come back as:

All of the other custom layers that I use seem to restore their custom setting properly.

--

There are two further settings that also are not saving and restoring properly after a cookie deletion and re log-in:

@idefix37

Thank you.

I'm measuring distances on the map.

I make the measurement, then cancel the distance and planning function.

Then I move to another part of the map, at a different zoom level, to start making more distance measurements, but as soon as I open it (with right click on the map), the map immediately jumps back to the prior location and zoom level, that I was previously measuring at, with the same waypoints displayed as I'd had when I cancelled distance and planning.

This was not occurring before v43.0.3

I have to refresh the page before I can use the function again.

Latest Firefox for Win11.

@idefix37

No problem, I thought everyone would be aware of the change, given it was recent.

The jaggedness of the pixelated images and drop of visual data on maximum outward zoom is a new thing. It zooms out and paints the screen faster but with a much poorer visual result and details.

It's visually ugly on a global view. This was not the case until recently. My guess is the drop in image quality coincided with the adoption of the combined radar and satellite layer as that juddered quite a bit while moving the view, with its initial implementation.

!

Yeah, it works well, but I'm glad you've left the old radar layer and satellite layer in there too, I use those the most.

Not keen on the higher pixelation when zooming out. A lot of detail disappears.

Most of the steps are OK-ish, but that final zoom-out level is too grainy, looks bad.

The 'Accumulative' sliders in V43 are a big improvement, been wanting that for a long time - thank you.