Just wanted to share a component to query monthly Rainfall data that can be quite useful for climate analysis. It came out after some headaches for remote locations in tropical climates, where the EPWs did not contain anything regarding rainfall.
I´ve used historical monthly rainfall data from the The National Center for Atmospheric Research here during 1986-2005. The component requests the appropiate data according to the location, which is downloaded in a .csv format and read back into Grasshopper.
Ideally, there should be a way to find hourly data, but so far I haven´t found it.
It is not rocket science, but might be helpful for someone
As I´m learning python, the coding is nowhere close to Ladybug standards. Any corrections or improvements are most welcomed.
@RafaelA just downloaded. When I first ran I did get an error on line 56. It turned out it just needed the C:\ladybug\RainfallFiles folder in place (line 49). Once I created that folder it worked a treat.
@seanmoo,
Right. Thanks for posting that. I added a check in the component to create the directory if it doesn’t exist. @RafaelA,
You could use this template to turn this into a real LB component.rainfallComponent.gh (14.8 KB)
@LelandCurtis and others,
I agree that this addition can be great.
Though i can say that the amount of trust you can give to this source is not always big. For my local climatic conditions i can say for sure the data is pretty much wrong.
So the lesson is to always have a double check of data 9any kind).
-A.
The reason that we don’t have such a component yet is because the values in weather file can’t really be trusted. There is a column in epw file for hourly precipitation but the data is all over the place. Here are other possible resources to use if we want to do a comparison between the data from the available sources.
Thanks for the development. To address the issue found by @seanmoo you can check for the folder and create it if it not exist. Something like this should work:
if not os.path.isdir('your directory'):
os.mkdir('your directory')
Actually, it would be good to have the different sources mentioned by @mostapha as alternative sources in the component. That would allow to choose the data source and make the comparison between them much easier.
I wonder how straightforward would be to get the data from these sources, any idea?
So far, the data from The National Center for Atmospheric Research is quite easy to query.
I’m a bit behind with the initial built of the component but I will push it through the weekend.
I have finally adapted the RainfallComponent.gh (459.4 KB)
to the Ladybug template. I think it would be good to get this first version done before exploring the other data sources mentioned above.
The only last part I still have to finish is the naming for the .csv file when the input location is not defined. For some reason this one is driving me crazy. I’m still trying to fully understand how to use the Input check function, lists and strings. So far, this is the closest I’ve got:
@RafaelA,
Thank you very much for adding this component and it’s working very well on my system. This will be a very useful addition to the plugin and I only have one request before we merge it into the code base.
At the moment, I have to add the Ladybug header myself in order to get the data from your component to work with other Ladybug components like the Monthly Bar chart:
It would be very useful to have the header included on the data when it comes in so that we don’t have to go through this extra step and we can just use the “separate data” component if we need to get just the numerical data again. If you include the header on the data, this also saves us from needing the Header output, which will make the component a little cleaner. Let me know if you agree, if so, feel free to make the change and send a pull request to the Ladybug Github.
Lastly, I agree that we should just merge in this monthly rainfall component for now but I will say that there is a way to get hourly precipitation data from the same NOAA source. Notably, this NOAA source is not a ground-based precipitation measurement but comes from analysis of weather satellite data, which have results for every 30 minutes over the planet (you can see in the GPM database here that the resolution goes down to 30 minutes at every 0.1 degree of latitude/longitude). The derivation of the data from satellites makes it much more accurate than the epw precipitation data, which comes from a human measuring a rain gauge by eye.
You have to request the specific GPM data that you are interested in (year and location) because it’s a huge database so the process of getting the data can take a day or two. Still, having this data would be very useful for modeling things like flooding in big rain events and I had always planned to build some components that bring this data into GH. If anyone wants to tackle this task, we will definitely merge the results into the Ladybug Tools code base.
Many thanks for the notes @chris, it makes a lot of sense! Somehow it hadn’t crossed my mind.
I have updated the component (20171128_Rainfall.gh (507.0 KB)) to incorporate the headers, but there is still something beyond my understanding - Although Ladybug’s header component and this one generate the same data, the Monthly Bar Chart doesn’t accept the output of the Rainfall component as a valid input. The error is:
Current document units is in Meters
Conversion to Meters will be applied = 1.000
Runtime error (TypeErrorException): unsupported operand type(s) for -: ‘str’ and ‘int’
Traceback:
line 299, in makeChartCrvs, “”
line 1135, in main, “”
line 1203, in script
So far, I have compared the data types coming out of both components, using print(type(Rainfall_mm[8])). All the items in both lists are equally strings so I can’t see any reason for it not to work. I guess it may have something to do with python data types - still a starter on that side. If someone has a hint on this, I should have it soon and make the pull request
Regarding the GPM database from NOAA, it looks very accurate and complete. I guess that bringing that data into grasshopper would open a whole new set of capabilities - flooding simulations sounds a great use of this data. As much as I would like to embark on it, I’m still learning the basics of python and it seems a huge task to take on. In any case, I’m keen to keep exploring it and try to find some time push it forward (it is difficult to know to which extent) - Would it be better to open a Github topic for this?
@RafaelA ,
I haven’t had time to check your file but I’m pretty sure that I know what the issue is: the dates on the header need to be real python tuples and not just strings (so they should be (1,1,1) and not “(1,1,1)”). Try making them as python objects and it should hopefully work. The use of these tuples was a less-than-ideal design decision that will be corrected in Ladybug+. I’ll respond about the GPM database once I get the chance.
@RafaelA,
Would you like to contribute this component to the Ladybug Legacy tool set? If yes, can you please send a pull request on Github? I can help you if you need any help in this process.
Once that’s done, the next steps you can follow from this video. The video will start at a relevant time.
Once you make a pull request, Mostapha or Chris will review it and will issue comments if any.
Let me know if you have questions at any point in the process.
