Mortality validation against Urban SIS #186
Comments
|
I think it is interesting to do so and I can scheck if anyone can work on this when they are backf from holidays next week. Two important questions:
/Lena |
|
The mortality data is calculated for each study. We already have a study for Stockholm, please remind me of other two cities that are available in Urban SIS and I'll add two more studies. Please also remind me what exactly does Urban SIS calculated? What we have is the data for different periods, for different climatic scenarios and for different occurrence probability. We need to choose correct ones for comparison. Once the study is done, in Risk/impact step look for the table view. This view allows you to download the data. My proposal is to do it for one city first, then possibly change our model parameters to get a better match and finally see how it works in other two locations. Also, I don't expect to see the same numbers, but we should see the same behaviour in space, in time and across climate scenarios. |
|
The three cities in Urban SIS are:
Stockholm, Amsterdam and Bologna
Urban SIS computes mortality data for five selected years for a reanalysis and one historical and future period (climate scenario, RCP8.5)
The historical years are:
The future years are selected according to this table:
[cid:image001.jpg@01D66EFD.16EED360]
/Lena
Från: Denis Havlik [mailto:notifications@github.com]
Skickat: den 10 augusti 2020 09:39
Till: clarity-h2020/csis
Kopia: Strömbäck Lena; Assign
Ämne: Re: [clarity-h2020/csis] Mortality validation against Urban SIS (#186)
The mortality data is calculated for each study. We already have a study for Stockholm, please remind me of other two cities that are available in Urban SIS and I'll add two more studies.
Please also remind me what exactly does Urban SIS calculated? What we have is the data for different periods, for different climatic scenarios and for different occurrence probability. We need to choose correct ones for comparison.
Once the study is done, in Risk/impact step look for the table view. This view allows you to download the data.
My proposal is to do it for one city first, then possibly change our model parameters to get a better match and finally see how it works in other two locations. Also, I don't expect to see the same numbers, but we should see the same behaviour in space, in time and across climate scenarios.
—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHub<#186 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AB2NFICDRHNIOPZERCCUNQ3R76PYBANCNFSM4PYT35SQ>.
|
|
Something went wrong with uploading of that table data. Can you please upload it again? |
|
It is like this. For each period we have selected on year that is normal, cold and wet, cold and dry, warm and wet, warm ande dry. The selected years are: Amsterdam: preasen: 2005, 2000, 1996, 2010, 2007, Future 2053, 2038, 2034, 2035, 2049, 2054 Bologna: present 2005 1987 1996 2010 2006, future 2050 2037 2053 2063 2044 |
|
How do we get the relevant data from UrbanSIS? |
|
Hi Denis,
Jorge is back from holidays and he is the one with most insight in the data we have produced.
Data is available on our UrbanSIS servers and some of it in the CSIS as a data package, but I believe the first step would be to have a discussion so we can decide which part of the data that is most relevant to compare.
Best regards,
Lena
Från: Denis Havlik [mailto:notifications@github.com]
Skickat: den 24 augusti 2020 10:23
Till: clarity-h2020/csis
Kopia: Strömbäck Lena; Assign
Ämne: Re: [clarity-h2020/csis] Mortality validation against Urban SIS (#186)
How do we get the relevant data from UrbanSIS?
—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHub<#186 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AB2NFIEQCECPXA43UBG63IDSCIPNBANCNFSM4PYT35SQ>.
|
|
@mattia-leone , @stefanon - I believe you are better positioned to answer this than I am. |
|
How is the validation coming along? Who is currently working on that? |
|
Nobody s working on this on our side as we need to get in contact with someone that can explain the data on CSIS. |
|
Sorry, I did not mention this request, happening during my holidays. The mortality statistic comes from an EUROSTAT database: See: https://ec.europa.eu/eurostat/de/web/cities/data/database We toke the parameters DE1001V (Total population) and SA2019V (total deaths per year) for each city area This table was imported by EMIKAT and can be accessed by: From this imput data the mortality rates are calculated. The results look like these: The quotient of "total deaths per year"/"Total population" is taken for the "most plausible year", using this priorities:
The favorite year 2012 comes from the fact, that the model was calibrated for the year 2012. |
|
you can find the "impact.calculation_mortality_v3.docx" document here |
|
I triggered the EMIKAT for the Stockholm Test 2 study (study 129) again. The results for |
The DEV system does not trigger Emikat, because it could otherwise overwrite the actual results on the PROD system. E.g. if there is a Study 129 on PROD, it would then be showing the results for the parameters used in DEV. To test the Emikat calculations you need to use the actual PROD system. DEV system allows only to test the internal functions of the CSIS. |
|
Thanks for the explanation! |
Urban SIS has mortality for Stockholm and AFAIK two more cities. Their mortality model is supposed to be accurate and has been validated according to SMHI, so we could /should try to validate our model against it.
I don't expect exactly the same results, but the trends in space, time and climate scenarios should be correct.
WDYT?
Btw, is there some other info in Urban SIS that we can validate our data against? Flash flood maybe?
The text was updated successfully, but these errors were encountered: