pyart.map.grid_from_radars is not taking into account masked arrays...

[johnnypeste]

I have a "radar" object and I want to use pyart.map.grid_from_radars ...

radar.fields['reflectivity'] and radar.fields['specific_differential_phase'] are masked arrays with different masks...

(in fact, reflectivity could have all values equal to true in the mask and specific_differential_phase could have all values equal to false in the mask)

For 2 pixels I have a very large kdp (>30 degrees/km)

kdp = radar.fields['specific_differential_phase']['data']

kdp[kdp>30] gives:

masked_array(data=[32.7059440612793, 32.7059440612793],
mask=[False, False],
fill_value=1e+20,
dtype=float32)

which are masked as FALSE (!!!) so they shouldn't be used...

However when I use:

grid = pyart.map.grid_from_radars(radar,
gridding_algo='map_gates_to_grid',
roi_func="dist_beam",
constant_roi=None,
min_radius=1000.0, h_factor=(1.0, 1.0, 1.0), nb=1.0, bsp=1.0,
grid_shape=(z_grid_points,
y_grid_points,
x_grid_points),
grid_limits=(z_grid_limits,
y_grid_limits,
x_grid_limits),
)

I will have those "BAD" values present in the grid!!

So I conclude that pyart.map.grid_from_radars is taking into account bad values (masked as false) in the interpolations...

What should I do?

[zssherman]

"which are masked as FALSE (!!!) so they shouldn't be used..."

@johnnypeste The mask is False for those two values, meaning they aren't mask, so they are used. Are you trying to mask out anything above 30 for that array?

[johnnypeste]

I undrestood that now... But 30 degrees/km is rainrate of 660 mm/h.... It cannot be and Iris software from Visala don't show this value!

[zssherman]

I'm trying to pinpoint the issue. How is specific differential phase being created? Is it present in the raw data set or is it being calculated using a function? The py-art reader will only read what's present in that raw dataset, it won't add values as such. If your looking to mask the values out, pyarts gridding function takes a gatefilter as a parameter.

You can use:

gatefilter is pyart.filter.GateFilter(radar)
gatefilter.exclude_above('specific_differential_phase', 30.0)

then add gatefilter=gatefilter in your parameters for pyart.map.grid_from_radars

[johnnypeste]

the kdp is in the raw iris data. the productx from vaisala is different Yahoo Mail: Search, organise, conquer On Wed, 11 Dec 2024 at 18:44, Zach ***@***.***> wrote: I'm trying to pinpoint the issue. How is specific differential phase being created? Is it present in the raw data set or is it being calculated using a function? The py-art reader will only read what's present in that raw dataset, it won't add values as such. If your looking to mask the values out, pyarts gridding function takes a gatefilter as a parameter. You can use: gatefilter is pyart.filter.GateFilter(radar) gatefilter.exclude_above('specific_differential_phase', 30.0) then add gatefilter=gatefilter in your parameters for pyart.map.grid_from_radars — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[johnnypeste]

productx and pyart are different by a factor of avout 5.3 but only in kdp. in z they are the same Yahoo Mail: Search, organise, conquer On Wed, 11 Dec 2024 at 20:34, Joao ***@***.***> wrote: the kdp is in the raw iris data. the productx from vaisala is different Yahoo Mail: Search, organise, conquer On Wed, 11 Dec 2024 at 18:44, Zach ***@***.***> wrote: I'm trying to pinpoint the issue. How is specific differential phase being created? Is it present in the raw data set or is it being calculated using a function? The py-art reader will only read what's present in that raw dataset, it won't add values as such. If your looking to mask the values out, pyarts gridding function takes a gatefilter as a parameter. You can use: gatefilter is pyart.filter.GateFilter(radar) gatefilter.exclude_above('specific_differential_phase', 30.0) then add gatefilter=gatefilter in your parameters for pyart.map.grid_from_radars — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[johnnypeste]

5.33 to be exact, which is the wavelenght in cm!! The reason why Vaisala software divides kdp by 5.33 to get decent kdp values is a mistery to me... pyart does not do that and that is why i have the infamous 32 degrees/km instead of 6 degrees/km Yahoo Mail: Search, organise, conquer On Wed, 11 Dec 2024 at 20:36, Joao ***@***.***> wrote: productx and pyart are different by a factor of avout 5.3 but only in kdp. in z they are the same Yahoo Mail: Search, organise, conquer On Wed, 11 Dec 2024 at 20:34, Joao ***@***.***> wrote: the kdp is in the raw iris data. the productx from vaisala is different Yahoo Mail: Search, organise, conquer On Wed, 11 Dec 2024 at 18:44, Zach ***@***.***> wrote: I'm trying to pinpoint the issue. How is specific differential phase being created? Is it present in the raw data set or is it being calculated using a function? The py-art reader will only read what's present in that raw dataset, it won't add values as such. If your looking to mask the values out, pyarts gridding function takes a gatefilter as a parameter. You can use: gatefilter is pyart.filter.GateFilter(radar) gatefilter.exclude_above('specific_differential_phase', 30.0) then add gatefilter=gatefilter in your parameters for pyart.map.grid_from_radars — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[johnnypeste]

Since I concluded that the problem has nothing to do with masked arrays, I'm closing this topic and I will open another one.

Thanks a lot for the help