Here's a brief summary with some starting questions/ ideas: The CAMS regional air quality analyses
- come in the form of hourly gridded maps
- are obtained by combining satellite observations, surface in situ observations and numerical modelling that mimic the laws of physics and chemistry in the atmosphere
- are of very high quality and can be considered as “ground truth”.
The main limitation of this dataset is that it cannot account for fine spatial variability and local effects that occur within few meters to hundred meters near a source, like a road segment with heavy traffic or a power plant stack. The data are representative of the average background values over a grid with a scale of 10 km by 10 km.
The data can be obtained openly and freely from CAMS’ Atmosphere Data Store (ads.atmosphere.copernicus.eu)
Interesting questions - are there areas of Italy where this modelled information can be further disaggregated for personal decision making with contributed observations?
Are there observations which are strongly disparate from the model and is this because of weather, sensor error, or better representation of the ground situation at local level?
Q. Are the contributed data definitely independent? in the case of weather you often find that citizen networks are also being federated into predictions. Action - ask CAMS
Q. would other data like RIVM be useful as well / instead?
OK, I have investigated the data a bit. Over Europe for PM2.5, there are 0.1 degree resolution versions of the following in NetCDF format for download:
Seasonal (spring, summer, autumn, winter) and annual means from an ensemble of 6-8 models
e.g. https://data.regional.atmosphere.copernicus.eu/openwis-user-portal/srv/en/iso19139.xml?id=3227
Exceedence map of locations over the European recommended levels
https://data.regional.atmosphere.copernicus.eu/openwis-user-portal/srv/en/iso19139.xml?id=3013
Hourly maps of modelled levels at surface from 4 different models (Mocage, Match, Eurad, Lotos)
e.g. https://data.regional.atmosphere.copernicus.eu/openwis-user-portal/srv/en/iso19139.xml?id=2627
Much more useful (especially when we all have limited bandwidth working at home) would be WCS services that can be consumed directly into some online analysis platform. The services looks available and I have applied for a token to test, but the bandwith / processing / storage will be more of an issue. CEH in the UK have a great VRE for this kind of analysis but I think it only contains UK-specific data...https://eip.ceh.ac.uk/apps/atmospheric/
Anne has reported this morning:
Hi Joan, all,
Thanks for a very helpful meeting today....!
I followed up on a few AQ data sets.
From the citizen science side, we have:
LuftDaten/SensorThings: http://maps.luftdaten.info/
PurpleAir: https://www.mathworks.com/help/thingspeak/rest-api.html
Turns out the State Department does not actually maintain monitors in Italy because the country has their own AQ network! Great news. Looks like some of the data are here:
And probably elsewhere, too....
Look forward to checking in tomorrow when I wake up. Have fun!
Cheers,
Anne
https://docs.google.com/document/d/1qJhOG9Qw7kud-VKA1yucp-iGxO0X2mwELSaf8nL2Ulo/edit#heading=h.kildxnv8j04d
(Sven reported that but it was empty the day before yesterday but itis progressing fast)
EPA AirNow data is in ArcGIS format straight from the source, EPA.
The best location to search for data is the ArcGIS Living Atlas of the World: https://livingatlas.arcgis.com/en/
I searched Air Now and 82 results for layers: https://livingatlas.arcgis.com/en/browse/#d=2&type=layers&q=Air%20Now
The EPA AirNow layer description is here:
https://www.arcgis.com/home/item.html?id=2d718d2733a74d1689d72b922c0ac4f4
Additionally, you can directly access the US EPA AirNow Tech data at the following URLs:
Register for an Account:
https://www.airnowtech.org/requestAccnt.cfm
API Signup:
https://docs.airnowapi.org/account/request/?account%2Frequest=
Data Download Site:
As you might suspect, the distribution resembles a human settlements map. Results agregated every 300m^2
Hi all - sorry for going a bit quiet but I've ben ill and still am... so I'm not sure how soon I will get back on track with looking at this hands on.
I have had a really positive response from CAMS about what goes into their models (only data from the EEA, no CS observations, - they are extremely interested to know the results of any experimental aggregation of their data with other sources). They are especialy interested in the data quality aspects of any dta combination and can povide some quite detailed info on their own QA procedures. The CAMS modelled data for particulates and other pollutants is freely available through WCS and WMS with a token (https://www.regional.atmosphere.copernicus.eu/). Potentially it could be pulled directly into ArcGIs Online or other cloud analysis environments because of this compliance with standards.
@Matth-March Re. spatial and temporal zones to consider:
In Italy, the key area where effects should be seen from around 3 weeks ago is Lombardia, and especially some industrial zones within the region which can be seen from the NO2 drops here: https://lucabattistellageo.users.earthengine.app/view/no2itaviz (this is an app running direct on the TROPOMI sentinel data for NO2, see here: https://developers.google.com/earth-engine/datasets/tags/tropomi).
@Matth-March again -yes, we should definitely be looking against a backdrop of past average to identify anomalies, as recommended here (https://atmosphere.copernicus.eu/amid-coronavirus-outbreak-copernicus-monitors-reduction-particulate-matter-pm25-over-china)
to quote - "other variables could potentially play a role in the decrease ...China is actively trying to reduce emissions, and meteorological variability between different years must be considered as well. To subtract these variables, we set the duration of three years 2017-2019 to estimate the ‘business as usual’ conditions as a compromise to have a representative estimate of February’s monthly mean, while not considering too long a period over which emissions vary substantially because of long-term trends.”
There could be seasonal differences in the peak pollen component of the small particulate measure - here in North Italy, the birch pollen levels exploded over the past 2-3 weeks (which would coincide with the big decrease in road traffic from the shutdown!) I suspect this may be earlier than in previous years - a chance to also bring in some phenological CS data?
https://translate.google.com/translate?sl=ca&tl=en&u=https%3A%2F%2Fwww.icgc.cat%2Fca%2FL-ICGC%2FActualitat%2FNoticia-395-Visio-des-de-l-espai-de-l-afectacio-de-la-pandemia-en-la-contaminacio-de-l-aire
Two resources with official data in SensorThings API format:
One STA about air quality. It might be interesting to compare it with the Anne developments on CitSci Air Quality services for the Earth Challange 2020
https://airquality-frost.docker01.ilt-dmz.iosb.fraunhofer.de/v1.1/
And one about COVID:
http://193.196.138.56:8080/STACOVID
This server contains global COVID-19 data from CSSE.
Some Guide for who is new to STA
REQUEST URL
Request list of available Countries (Thing in STA)
http://193.196.138.56:8080/STACOVID/v1.0/Things
Request list of available Countries with location Lat,Lon
http://193.196.138.56:8080/STACOVID/v1.0/Things?$expand=Locations
Request Datastreams from Country's name = Germany
http://193.196.138.56:8080/STACOVID/v1.0/Things?$filter=name%20eq%20%27Germany%27&$expand=Datastreams
Request Datastreams and Observations from Germany
(number of Total Death/ Total Recovery/ Total Confirmed cases)
http://193.196.138.56:8080/STACOVID/v1.0/Things?$filter=name%20eq%20%27Germany%27&$expand=Datastreams/Observations
https://www.catalannews.com/tech-science/item/algorithm-aims-to-detect-covid-19-cases-with-app-that-records-coughing
If you want to try it, please clean you phone afterwards!
https://github.com/yahoo/covid-19-data The Yahoo Knowledge Graph team at Verizon Media is responsible for providing critical COVID-19 data that feeds into Yahoo properties like Yahoo News, Yahoo Finance, and Yahoo Weather. The COVID-19 datasets include country, state, and county level information updated on a rolling basis, with updates occuring approximately hourly.
The COVID-19 datasets are constructed entirely from primary (government and public agency) sources with a clear attribution of the primary sources used for each geographical region. While other aggregations of COVID-19 data are already available, we believe ours to be the only open source COVID-19 dataset that is constructed entirely from primary sources with clear attribution back to those sources. Our hope is that additional transparency will enable more accurate analysis, aiding researchers who seek to understand and prevent further spread of the disease.
Links in our #14 telecon:
15:23
where is it possible to find out what the earth challenge approach is doing ?
15:25
Lucy Bastin
How Joan described story maps just then is exactly right
15:26
Lucy Bastin
Not about hosting data but can lead you to data and help you explore it
15:26
Metis Meloche
How StoryMaps are used by EC2020, is available on our webpage: https://earthchallenge2020.earthday.org/
15:27
Metis Meloche
How we lead people to the data is we're trying a hubsite, called the citizen science cloud: https://cscloud-ec2020.opendata.arcgis.com/
15:27
Valantis Tsiakos
15:30
Clay Herbaugh
https://storymaps.arcgis.com/stories/f99dd40c7e084ec19df7affc91efb95c
15:41
wanted to share with you: https://op.europa.eu/en/web/eudatathon
15:41
Steffen Fritz
Dear all, sorry I have to leave thanks for the great input and lets be on touch soon!
15:44
Lucy Bastin
Hi Lukas, is this the cams data?
15:46
Here is a dataset from Google Community Mobility Reports. It's a massive worldwide dataset about the change in mobility during the covid 19 breakout.