This is a rather difficult question and the answers are certainly not many.

However, you can start your quest from the following web site

http://climserv.ipsl.polytechnique.fr/gewexca/

It is the web site of the GEWEX GDAP Cloud Assessment exercise and I believe it could help you somehow. It is certainly a start and from there you probably can find what you are looking for.

Moreover, a very famous and controlled dataset is the one from the International Cloud Climatology Project (ISCCP), but it is probably shorter than you need:

http://isccp.giss.nasa.gov

Hi Zoltan,

additionally, you could use synop reports. There is an interesting study from Rulfová and Kyselý (Rulfová, Zuzana ; Kyselý, Jan, 2013: Disaggregating convective and stratiform precipitation from station weather data, Atmospheric Research, 134, pp. 100-115). Problem will be that synop reports (at least in the needed temporal resolution) do not cover 100 years.

Another option would be to apply reanalysis data (see reanalysis.org), which often include the partitioning in convective and stratiform precipitation (from the model point of view, be aware that this is quite different to e.g. the point of view of the radar community...). There are some reanalysis datasets covering the time period you need. Problem could be that e.g. no precipitation data are assimilated and model is at least for the pre-satellite era not well constrained. Furthermore, I am not sure if relevant reanalysis datasets provide convective and stratiform precipitation as output variables....

So, it remains a challenging task ...

Cheers,

Klemens

Dear Vincenzo, Dear Klemens,

Thanks for your advice.

best wishes,

Zoltan

I don't know if it helps as I'm not aware of the data policy of the meteorological services in the countries that you study (it's different from one country to the other) but stratiform / convective precipitation can be derived from ground-based radar measurement (Steiner et al., 1995). It can be interesting to explore this possibility, although you won't have 100 years (no more than ~30 years) and it can involve some heavy data processing if you have to work with the raw data. 

One relatively simple method would be to use manually observed cloud type classification and rainfall rate archives, which have a relatively long history in Europe and will offer a longer climatology than remotely sensed data. Of course these will have their limitations (daytime only for manual cloud observations, relative sparsity of long-term rainfall rate data etc.) but may be the only solution available in the first half of the twentieth century. Some archives may not be digitised yet either.

For rainfall rates, anything above 10mmhr-1 (WMO definition of heavy rain and the UK's CAA definition of heavy convective rain) is very likely to be convective, but there is the issue of whether the heaviest rainfall passed over the rain gauge and the stage of convective development etc.

As for data sources, archives can usually be found by contacting national weather services in the countries of interest, with many having online archives available for academic use. For older records, libraries such as the UK Met Office national meteorological library and archives ( http://www.metoffice.gov.uk/learning/library ) are comprehensive, including those from ships. These historical station reports are likely to have observed cloud type at least, if not rainfall rates. Again, many older records will not be digitised yet, which will add time to your project.

Hope this helps,

Alec