Caro-COOPS.org Discussion Board

jcothran · Joined: 24 Feb 2025 Posts: 144 Location: Columbia, SC

Hi Lian,

Thanks for the responses. I've attached my replies:

>>
>> 2)some dialog with Lian/team about
>>
>> - the flow of the input data which we will be providing as it exists
>> currently and what else they need

> Comments: we will need more when it gets to real-time
>forecasting
>in Year 2. For testing data transmission, it is fine to start with what
>we have now. We should try a few more tests, preferably, in a way the
>data from SC is directly imported into the model.

Reply: Does this mean that your team has been able to successfully pull data from the converted netCDF files on our DODS server and use this as input to your model?

>> - the output of the model (the model products and how users will
>> request
> and obtain these products)

> Comments: this is a two-way processes between us and the users.
>We can provide different forms of model out put : e.g. two-dimensional
>fields of water level at specified time intervals; time series of water
>level at specified locations; areas of flooding; flood water depth;
>flooding potential expressed as a probability; etc. We can design our
>output in way the user prefers.

Reply: From the content of other emails, it sounds as if our users will not be evaluating our product until after we have developed something. In the meantime, I'd suggest that we move forward with a base product which would simply be a flood map for a given 48 hour prediction. I think this would be a good starting point and get us in the practice of defining the work flows and expectations within the groups. For this kind of product we would need to be able to convert your model output into a shapefile so that we can overlay this layer into a web enabled GIS.

Do you have some sample model output that we could use towards this end?

>> With the modeling team we know the inputs, but are unsure if Lian
>> wants us
> to figure out how to acquire these automatically or if he already knows
> of a
> process for doing this(say an input stream like the current NWS
> conditions).

> Comments: A major issue we need to address in Year 2 or Year 3 is
>data automation, like what the NWS has now. In other words, we need to
>build an oceanographic analog of the McIDAS system for marine
>forecasting. It should automatically collect data from the monitoring
>sites, and existing data archives, and transform the data into a
>standard form 9should be consistent with what the NWS uses now. Such a
>system does not exist anywhere. This is not a simple task.

Reply: I reviewed the McIDAS and NWS sites some more and I would definitely agree this is no simple task. Both agencies have existed for decades and have legions of staff. While we may use much established development in our system, our unique assets are the storm surge model and the observational array we are utilizing. My first efforts will be to insure that the storm surge model is being serviced correctly and the model results reach the widest interested audience. Both McIDAS and NWS act as file servers more than databases, etc. Their value is in the data of the observational arrays they collect from, the QA/QC and standard formatting of this data and the serving of this data to clients on request as well as (in mcIDAS case) some analysis and visualization software for the images they are serving.

We will be 'consistent' with NWS in that a user should be able to select datasets from our DODS netCDF server with the same ease that they would select a file from the NWS site.

Are there any existing input data streams which we should go ahead and start converting to netCDF and storing on our DODS server in anticipation of your model needs :

from the following NWS site? http://www.nws.noaa.gov/tg/dataprod.html

from the USGS site? http://water.usgs.gov/realtime.html

I'm unaware where the USGS 'pushes' their water flow data, so I imagine we'd have to do screen scraping.

Webmasters! You can construct persistent links directly to an individual station page. Use the format http://water.usgs.gov/cgi-bin/realsta.pl?station=xxxxxxxx where "xxxxxxxx" is the (usually) 8-digit station identification number.

from the DNR sites(NC/SC)? http://www.dnr.state.sc.us/water/nrima/gisdata/index.html

or other data streams we should start moving towards automation?

I imagine that for our upcoming observational data array you'll be gathering wind direction and speed at each of the buoy points. Are there other variables I should include from our observational array?

Do you need us to automatically gather tidal information or are you computing this beforehand?

Thanks,
Jeremy