pytroll
create a stratiform/convective mask
Hi, I am new to Satpy, and this is the first time I am using this package. I have microwave brightness temperature datasets from the GPM satellite. Using this I need to create a stratiform/convective mask to clasify the cloud systems into convective and stratiform fractions. Can anyone tell me if this can be done using Satpy, and if so how? Thanks in advance !
Hi. It sounds like there are two sides of this question: the programming side (how to use Satpy) and the science side (what goes into the calculation). I say that because I think you're looking for the programming answer and know the science answer, but I'm in the opposite position. Could you tell me, without Satpy, how would you calculate what you're looking for? Is it 1 mask or 2 masks that would be produced? What do you want to do with the mask after it has been created?
Edit: Please include what datasets (variables) you would need from the GPM input data or if you'd need anything not provided by the GPM data files, in order to calculate this.
Yes, I can provide you details about the science part of how to create a stratiform-convective mask using microwave brightness temperature data. Let me check my notes and get back to you in a day or two.
On Mon, 6 Oct, 2025, 8:33 pm David Hoese, @.***> wrote:
djhoese left a comment (pytroll/satpy#3249) https://github.com/pytroll/satpy/issues/3249#issuecomment-3372170971
Hi. It sounds like there are two sides of this question: the programming side (how to use Satpy) and the science side (what goes into the calculation). I say that because I think you're looking for the programming answer and know the science answer, but I'm in the opposite position. Could you tell me, without Satpy, how would you calculate what you're looking for? Is it 1 mask or 2 masks that would be produced? What do you want to do with the mask after it has been created?
— Reply to this email directly, view it on GitHub https://github.com/pytroll/satpy/issues/3249#issuecomment-3372170971, or unsubscribe https://github.com/notifications/unsubscribe-auth/AQACDYU5SG2TQTZETB6JHED3WKAC7AVCNFSM6AAAAACIIVZLAOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTGNZSGE3TAOJXGE . You are receiving this because you authored the thread.Message ID: @.***>
*Convective-stratiform separation *
Brightness temperature measurements of the cloud systems is required to determine the dominant type (convective or stratiform) in the observed precipitation because the amount of cumulative rainfall alone cannot distinguish between convective and stratiform events. Both can produce similar daily totals, but they differ in intensity, duration, structure, and mechanisms. Differentiation between convective cores and embedded stratiform rain within the rainstorm systems is crucial. The general rule is that if the rainfall is dominant over a short window, convection likely dominates, but if it is spread over many hours, much may be stratiform.
To distinguish between stratiform and convective fractions of cloud systems, microwave BT is generally more suitable than infrared BT. This is due to the following reasons :- 1) Microwave BT screen out high, thick cirrus clouds, which often leads to overestimation of deep convective cloud fractions in infrared measurements. 2) Infrared measurements are primarily sensitive to cloud top temperatures and are usually used to infer cloud ice water content, but have limited ability to detect cloud properties below the top layer. 3) Microwave measurements utilize multiple frequencies and thus can classify convective and stratiform precipitation with greater accuracy.
Usually at microwave frequencies < 37 GHz, the brightness temperature responds mostly to emissions from rain and cloud liquid at lower levels. Thus, this band is referred to as emission band. On the other hand, at frequencies > 37 GHz, the brightness temperature responds mainly to scattering from cloud ice in the upper levels of the cloud. Thus, this band is referred to as scattering band. Hong et al. (2013) have used microwave brightness temperature at two frequencies to distinguish convective and stratiform storm fractions - one at low frequency (19 GHz), and one at high frequency (85 GHz), both at horizontal polarization.
However, his work was confined to clouds over oceans. There will be a difference in methodology adopted for the sake of classification depending on whether you are observing clouds over land or over oceans. This is because the observed BT by the satellite is a complex function of contributions from both the atmosphere and the surface, and the surface BT is drastically different from land and oceans. Oceans have a low, and relatively uniform emissivity. Land surfaces have a high, and highly variable emissivity. The land surface emissivity in the microwave spectrum varies dramatically with factors like soil moisture, soil type, vegetation cover, and surface roughness. The high emissivity by land surface completely masks the emission (low-frequency) channels, and thus we have to rely only on scattering (high-frequency) channels for the sake of classification.
We have developed a novel methodology to categorize the convective and stratiform cloud fractions over land surface takin into consideration their general physical characteristics. Convective regions are characterized by strong updrafts and downdrafts. Stratiform regions, on the other hand, have relatively weak vertical air motion (Hong et al., 2000). Due to these strong updrafts and downdrafts, convective regions exhibit large horizontal variations in liquid and ice-phase precipitation. Stratiform regions, in the absence of a strong vertical motion have more uniform precipitation distributions (Kummerow et al., 2001). Additionally, there is usually a preferred orientation of ice particles in stratiform regions, while turbulent updrafts in convective regions cause ice particles to lose preferred orientation (Heymsfield and Fulton, 1994a,b). Furthermore, dominant ice hydrometeors found in convective systems found in convective clouds are larger and denser formed due to riming. On the other hand, dominant ice hydrometeors found in stratiform clouds are lighter ice crystals and aggregates formed from vapor deposition and riming.
Following steps are included in the creation of the cloud mask :-
Step 1 :
Where,
Step 2 :
Where,
Step 3 :
Where,
, and
Following steps are taken to apply the stratiform-convective mask :-
Step 1 :
Then “potential raining”, otherwise “no rain”
Step 2 :
Then “deep convective”
Step 3 :
Then “moderatively convective”
All other “potential raining” pixels are classified as “stratiform”.
On Tue, Oct 7, 2025 at 5:44 PM Mohit Kumar @.***> wrote:
Yes, I can provide you details about the science part of how to create a stratiform-convective mask using microwave brightness temperature data. Let me check my notes and get back to you in a day or two.
On Mon, 6 Oct, 2025, 8:33 pm David Hoese, @.***> wrote:
djhoese left a comment (pytroll/satpy#3249) https://github.com/pytroll/satpy/issues/3249#issuecomment-3372170971
Hi. It sounds like there are two sides of this question: the programming side (how to use Satpy) and the science side (what goes into the calculation). I say that because I think you're looking for the programming answer and know the science answer, but I'm in the opposite position. Could you tell me, without Satpy, how would you calculate what you're looking for? Is it 1 mask or 2 masks that would be produced? What do you want to do with the mask after it has been created?
— Reply to this email directly, view it on GitHub https://github.com/pytroll/satpy/issues/3249#issuecomment-3372170971, or unsubscribe https://github.com/notifications/unsubscribe-auth/AQACDYU5SG2TQTZETB6JHED3WKAC7AVCNFSM6AAAAACIIVZLAOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTGNZSGE3TAOJXGE . You are receiving this because you authored the thread.Message ID: @.***>