About the output of geo2grid -r ahi_hsd

Post Reply
liugers11
Posts: 5
Joined: Mon Jan 16, 2023 12:51 pm

About the output of geo2grid -r ahi_hsd

Post by liugers11 »

Hi,
Thank you for providing such an excellent tool (Geo2Grid). I want to use Geo2Grid to process the Himawari-8 HSD data to tiff data. For example, the out name is "HIMAWARI-8_AHI_B03_20211012_040000_my_grid_name". I know that it means Band 3. Can you tell me what the unit of the product is? Is it top of atmosphere radiance, top of atmosphere reflectance, Rayleigh corrected reflectance, or DN value?
Another question. How can I get the solar zenith angle, solar azimuth angle, satellite zenith angle, and satellite azimuth angle using Geo2Grid?

Thank you very much.

Best Wishes

Ge
davidh
Posts: 115
Joined: Tue Jun 04, 2013 11:19 am

Re: About the output of geo2grid -r ahi_hsd

Post by davidh »

Hi Ge,

The AHI HSD B03 data should be TOA reflectance data with no additional adjustments applied to them beyond what is available in the file. From what I can tell that should mean no `/ cos(SZA)` or any rayleigh correction. The data is multiplied by 100 to create a percentage (%) instead of an "albedo" (0 to 1) dataset. Note that the `/ cos(SZA)` and rayleigh correction adjustments/corrections are applied by default when creating certain composites like the `true_color` RGB.

Regarding angles, at the time of writing there is no way to create the angle datasets as separate products. They are generate on the fly for certain composites/corrections like rayleigh correction. They are also available through a special utility function if you are using the Satpy Python library directly, but not through the traditional loading mechanism that Geo2Grid is currently using. This may change in the future, but I don't have any timeline and can't make any guarantees.

Please let me know if you'd like clarification on any of this.

Dave
liugers11
Posts: 5
Joined: Mon Jan 16, 2023 12:51 pm

Re: About the output of geo2grid -r ahi_hsd

Post by liugers11 »

Thank you very much. Can I ask another two questions?
(1) In the Geo2Grib, there is a keyword "--no-enhance" for the output. To obtain the "real" TOA reflectance, should I set this keyword "--no-enhance" to avoid image enhancement?
(2)The output of Geo2Grib is 8-bit unsigned integers. How can I get the real TOA reflectance?
The output value/100? Can you give us the specific equation?


Best Wishes

Ge
liugers11
Posts: 5
Joined: Mon Jan 16, 2023 12:51 pm

Re: About the output of geo2grid -r ahi_hsd

Post by liugers11 »

I have a third question. The satellite zenith angle and azimuth angle of HIMAWARI are constant? How can I get it?
liugers11
Posts: 5
Joined: Mon Jan 16, 2023 12:51 pm

Re: About the output of geo2grid -r ahi_hsd

Post by liugers11 »

I have the fourth question. The output of "B03" of geo2grid is TOA reflectance, not TOA albedo. Is it right?

Thank you very much.
davidh
Posts: 115
Joined: Tue Jun 04, 2013 11:19 am

Re: About the output of geo2grid -r ahi_hsd

Post by davidh »

Ge,

1. Regarding "--no-enhance", if you don't want the data scaled/enhanced and want the "raw" values as far as the observed/calibrated data, then yes use this flag. As you point out the default output data type is 8-bit unsigned integer, so doing --no-enhance with % data will end up with your data only being between 0 to ~100 in the 8-bit space. What you'll want to do is also include the `--dtype float32` option which will save a floating point geotiff. This of course assumes we are talking about the "geotiff" writer in Geo2Grid.
2. Getting the "real" TOA reflectance: What do you mean by "real"? The above steps will get you the percentage "reflectance" values in a geotiff file. Note that a geotiff of this scheme/style will not be visible in an image in 90% of image viewers. These values would be the percentage values. If you want the 0 to ~1 values you'd have to customize things. You could create a custom enhancement for the bands you're interested in and do a "crude stretch" enhancement between 0 and 100. This would result in the output geotiff with float32 data type to contain data between 0 and 1.
3. Unfortunately for the Himawari-8/9 satellites, these angles do change over time. The satellite drifts quite a bit from time step to time step. So getting these values once will not be accurate over a day or couple of days depending on what type of accuracy you expect. If you still wanted these values you'd still have to use something like Satpy directly to compute them and then save them in whatever format you wanted.
4. It depends what you mean by reflectance and albedo. In my experience everyone has a slightly different meaning. The values that are scaled by G2G and then written to a geotiff are not DN/counts and not radiance values, but they are also not reflectances by the strict definition of being divided by the `cos(SZA)`. They are in the units "%" so for real world data this can range from anywhere between -1% and ~120%.

I hope this at least answers your questions a little bit. If you still have questions then it may be good if you could provide more details about what you're trying to accomplish so I could give you more detailed responses.
liugers11
Posts: 5
Joined: Mon Jan 16, 2023 12:51 pm

Re: About the output of geo2grid -r ahi_hsd

Post by liugers11 »

Thank you very much.
Post Reply