We continue our model comparison to ground weather station data with the ETo component of the SME. The numerically predicted ETo values have been compared to ETo values calculated using data from ZiaMet, a public weather service provided by New Mexico State University. The study run dates are from 2017-05-25 through 2018-07-31. The R^2 values ranged from 0.83 to 0.94, with an average of 0.88 over the nine sites. Accumulated ETo error percentages range from 0.3 to 13.6 with an average of 7.4 percent. More detail can be found here.
We have improved the ETo module and reopened access to the Agrineer.org's ETo repository.
Please note that this is a preliminary study based on a limited geographical area. A more thorough study will include other regional stations and dates from 2017-01-01 through 2018-07-31. A second study is planned for run dates 2018-08-01 through 2019-05-31 (or later) which will include modifications to the WRF physics options in order to address a small bias for over-estimating ETo. These dates will need to be regenerated for the new options to take effect.
We initiate our model comparison to ground weather station data with the GDC tool. The GDC model output has been compared to ground sensed Grow Degree units from ZiaMet, a public weather service provided by New Mexico State University. Nine weather stations were studied with an average sample size of 692 days. The R^2 values ranged from 0.96 to 0.98, with an average of 0.97 over the nine sites. Error percentages range from 0.2 to 13.6, with an average of 6.0 percent. Details can be found here.
Encouraged by the above, we have reopened access to the GDC repository.
We aim to extend the GDC study to other weather services in our scope as resources allow. Meanwhile, we turn our attention to ETo evaluation over the same ZiaMet stations.
The SME and GDC tools are now being evaluated and calibrated using ground sensed data. This is a planned and welcomed phase in building these applications and so we have suspended access to the GitLab.com ETo and GDC repository until a satisfactory correlation is achieved.
The web versions (above) can still be used, with the understanding that results are UNCALIBRATED. The tools are available so that users can exercise the tools in order to become familiar with the overall procedures. As in our disclosure, do not use our results in any studies until calibration can be achieved.
As soon as we feel a reasonable correlation can be achieved, we will open the repository.
We have been generating data for about half of the western US starting 2017-01-01, but it did not include parts of Kansas, Oklahoma, and Texas.
Due to interest from folks in these states we are presently
data for these states starting 2019-01-01, using the
newly released WRF container (see below). This additional coverage
now includes all of Kansas and Oklahoma, and northern Texas.
We seek participants
to continue data contributions for this and other regions, see this
which describes the WRF data generator. You can download the
Docker container package
package which generates the
Weather, Forecast, and Research
(WRF) data used in the
tools has been released by Agrineer.org and can be found
Data contribution by third parties to the GDC and SME projects for specific geographic sectors can be done with this package. If you would like to contribute data for a new sector please contact us here to register.
Note that the package can generate the necessary WRF data for any part of the world, from Argentina to Zimbabwe. Also note that you DO NOT need to submit the data to us in order to do a run of the SME or GDC (see instructions in README.md). You can run the SME and GDC on your own computer and obtain the text reports they produce.
This container has been designed for compute servers but it can also be run on current multi-core desktop computers.
This release completes the set of necessary components for the GDC and SME projects by providing the data engine source code. See the software download section here for the other components.
Agrineer.org is a founding member of IndieCompLabs.org, a group of computational laboratories dedicated to open source projects that are both practical and educational, from satellite remote sensing to down-to-earth fablabs.
Follow the links on the left navigation bar or on images below to use the applications and to get more information.
For discussions about Agrineer.org and the GDC and SME applications you can go to our Google group Agrineerorg (note that in this case there is no '.' before the 'org') Agrineerorg . The full software is freely available at GitLab: gitlab.com/agrineer.