I have PM2.5 and PM10 concentrations from a site not in one of the standard networks AQS,CASTNET,IMPROVE, etc. Is it possible to format the concentrations file I have and include them in the AMET, specifically for a time series plot?
I haven’t been able to find a link for something that describes to do this but if there is I would love to have it.
Thanks,
Lawless
Hi Lawless,
Yes, this is possible. There are actually instructions in the AMET User’s Guide on adding support for a new AQ network (see section 8. How to Add a New AQ Network to AMET). I suggest you take a look at the User’s Guide for instructions. But there are actually a couple ways to tackle this.
If you only have a single site and don’t mind “lumping” it in with another network, you could simply add your data to one of the other network data files (e.g. AQS_daily). Simply follow the same format as the existing data in the file file (basically site_id, date_on, date_off, and then your data in the correct column). The only other thing you would then need to do is add the new site to the AQS_full_site_list.csv file (don’t worry about including all metadata, you just need the site_id, site_number, latitude and longitude of the site). You then rerun your database setup (to add the new site to the data metadata table) and then run AMET for AQS_daily (for example) and the new site should be included. Assuming everything works correctly, the data from the new site would be available as an AQS site. You can then do whatever you want with it.
The second method is to actually create a new “nework” for your data, and these are the instructions that are included in the AMET User’s Guide. So, you’d create a new observation data file with your data (you can mimic the format of one of the existing networks, like AQS_daily), create a new site file for your “network” (again, mimic an existing file like AQS_full_site_list.csv), and then modify the AQ_matching.R script to add you new network. Here’s where I’ll leave you and tell you to refer to the AMET User’s Guide, since the instructions on how to modify the R code in AMET to add a new network are in there.
So, decide which method you want to use. If this is a single site for single year, it may be faster to just go with method 1. If you’re expecting to have these data routinely, or you really want the site available as a separate new network, method 2 may save you time in the long run.
Let me know if you have any trouble once you get started.
Wyat
Thanks for quick reply!
Its only a one off thing for one location so I didnt want to set up a full network. I will go with the first method and let you know if anything goes wrong.
Lawless
Hi Wyat and Lawless. I have a similar question, but instead of adding AQ observations I’d like to add met observations to compare to the WRF outputs. Should I use a similar approach to add meteorological observations to the AMET database? I didn’t see anything pertaining to this specifically within the user guide.
Thanks for your help,
Rachel
Rachel,
There is an option in each model-obs matching script to automatically download MADIS observations.
$AMETBASE/scripts_db/metExample_wrf/matching*csh
Regards,
Rob
Hi Rob,
The MADIS data download and matching script is very convenient, but the mesonet dataset is incomplete for the sites I’m interested in (all mesowest data in 2017 is missing 4 nighttime hours) which is why I’m curious about manually adding met obs. I would also like to compare more surface shortwave radiation observations than those available from the MADIS data download.
Is it possible to augment the meteorological observations or is downloading from MADIS the only way?
Thanks for your help,
Rachel
Rachel,
This reminds me now of our discussion before on this issue.
There is a text surface met option for the regular T, Q, Wind obs only.
See if you have a directory
$AMETBASE/obs/MET/surface_text
Hourly text obs files have to be named like MADIS obs files. (i.e., 20130131_0000)
They have a header followed by rows of data, each row a single obs site.
site, network, slat, slon, YYYYMMDD, HHMM, t2m, wvmr2m, u10, v10
KRDU, METAR, 35.8800, -78.7900, 20130131, 0000, 294.150, 0.010846, 0.000, 11.318
Then you can set this option to “text” in the matching_surface.csh obs-model matching script.
I do not have a shortwave radiation method yet. We did add automated download directly from SurfRad, but that does not help you. I have wanted to explore the use of solar radiation measurements in the MADIS mesonet dataset, but not implemented yet. The problem there is so many bad observations. I don’t anticipate movement on this in the next few months. One could take their own solar radiation measurements and match with WRF outside of AMET and push those paired data to the AMET database table. It’s the same table as surface meteorology. Then use the AMET analysis scripts for solar radiation.
Regards,
Rob
Thank you for getting back to me! I’ll try the “text” option for the met and look into other options for solar radiation comparisons.
