Update icing-base-top.lua

[tackandr]

Fix issue with >=4 / <=4 comparison by using >3 / <5 since values are integers.
https://jira.fmi.fi/browse/STU-29733

[Iikkater]

Looks good!

[TyttiHMustonen]

The documentation of the VerticalHeightGreaterThanGrid and VerticalHeightLessThanGrid is unclear about the last argument (Nth). The explanation is the same for Nth = 0 or Nth = 1. https://github.com/fmidev/himan/blob/master/doc/plugin-hitool.md

If the Nth = 1 means that we try to find level where the parameter is bigger than threshold, and
if the Nth = 0 means that we try to find level there the parameter is smaller than threshold, then
we want it always to be 1. This because.. how i understand this, is that we want to get the first (lowest) level where icing is >= 4. And also the last (highest) level where icing is >= 4. So in both we have icing >= 4.

1. local basedata = hitool:VerticalHeightGreaterThanGrid(IceParam, zerodata, pFL300data, thresholddata, 1)

   • With this we find the first (lowest) height where icing >= 4, and the search is done between zerodata - pFL300data.

2. local topdata = hitool:VerticalHeightLessThanGrid(IceParam, basedata, pFL300data, thresholddata, 1)

   • With this we find the last (highest) height where icing >= 4 , and the search is done between the lowest level already found (basedata) - pFL300data.

   Okey, so the plugin doc talks about >=, not >, so 4 should already work?
   Why it doesn't?
   Maybe documentations is false, and it is > , not =>.
   If we need to include 4, we need to say > 3 in both cases?

   Maybe the functions just need to be tested what they do, or ask someone who knows more about them?
   Or maybe i didn't understand what they wanted in the ticket STU-29733
   :D

[tackandr]

Okay, that's an error in the documentation. It was just copypasted I assume. 0 means last level and 1 means first.

The problem is that our method only implements > operator but not >= operator. And since the parameter is an index type containing natural numbers we don't catch the 4 since that is >= 4 but not > 4. So in this case we use > 3 to get >= 4.
The documentation gives an example of total cloudiness >= 80% but that is also in reality > 80% but not such an issue since it is a floating point number and that brings the cutoff close enough to 80% anyway.

Documentation should be fixed for this.

[TyttiHMustonen]

Okay, now i understand that the functions work like this:

• local basedata = hitool:VerticalHeightGreaterThanGrid(IceParam, zerodata, pFL300data, thresholddata=3, 1)

This finds the first height where icing parameter is > 3, limits search between levels [zerodata, pFL300data]

• local topdata = hitool:VerticalHeightLessThanGrid(IceParam, basedata, pFL300data, thresholddata=5, 1)

This finds the first height where icing parameter is < 5 , limits search between levels [basedata, pFL300data]

But if we want to accept icing parameter 4, 5, 6, 7, 8 and 9 should we change this ->

local topdata = hitool:VerticalHeightGreaterThanGrid(IceParam, basedata, pFL300data, thresholddata=3, 0)

So it would fine the last height where icing parameter is > 3?

[tackandr]

In principle yes. There could be multiple layers of icing and we wanted to catch the top of the layer starting from base. If we have a single icing layer then the two ways should provide the same result.