OverseerSci - Datasets, Model Runs and Results

1. Introduction

2. Registration and limits

3. Datasets

3.1 Creating a dataset

3.2 Dataset list

4. Dataset screen

5. Failed analysis section

6. Analysis section

6.1 Uploading farm analysis files

6.2 Enter measured values (actual)

7. Model runs section

7.1 Full model run

7.2 Lysimeter model run

7.3 Animal ME model run

7.4 Sensitivity analysis

7.5 Model run types

8. Model run results

8.1 Comparing model runs

 

1. Introduction

OverseerSci is an online service for analysing and developing the Overseer model. Users must contact Overseer Limited to be set up to use the service in accordance with the OverseerSci terms and conditions.

This service is not to be used for modelling real farms, OverseerFM is set up for that purpose.

OverseerSci provides all the same functionality as OverseerFM with the addition of datasets and some APIs for integrating with other systems. Datasets contain a collection of farm analyses that can be run through the model and examined as a group.

This article describes the unique features of OverseerSci only and as such assumes knowledge of the Overseer model and the OverseerFM user interface.

OverseerSci is differentiated from other Overseer Limited software services by it's Orange colour: 

As with OverseerFM and OverseerEd, users can access help from within the software by switching on the "Quick Tips" to show the Green help text, and also by using the "Page Help" button

Back to top

2. Registration and limits

Organisations cannot register for science online. If you wish to use this service, please contact Overseer Limited at info@overseer.org.nz .

Initially organisations are set up with a limit of 10 datasets, 10 users and each dataset can have up to 5 model runs and 500 analyses. If you reach capacity, you will need to remove existing runs or datasets to get under the limit or contact Overseer Limited and discuss options around increasing these limits.

 

Back to top

3. Datasets

Datasets contain a set of farm analyses that can be run through the model and the results examined. To access datasets select it at the top of screen as shown below. This screen shows the list of existing datasets for your organisation.

A dataset is a collection of farm analyses that can be run through the model (a model run) with or without a selection of model parameter changes. The results of each model run are stored and model run results can be compared to see the effect on the specified model parameter change. Model run results at the analysis and block level can be downloaded into csv files. 

 

Back to top

3.1 Creating a dataset

You can create a new empty dataset by first selecting "Datasets" button on the top left of the screen and then the "Create dataset" button at the top right of the screen.

The name, project and description of the dataset is required to be entered.

Back to top

3.2 Dataset list

Once a dataset is created it will appear in the dataset list - shown below. All users within your organisation can see these datasets. 

You can delete you datasets from here but please note that if a dataset is deleted all model runs and farm analyses that were in the dataset are deleted.

The list of datasets can be ordered by the project name, dataset name, the date it was created or who it was created by.

The number of a datasets an organisation can have is configured by Overseer Limited. If you have reached your limit of datasets get in touch with Overseer Limited to discuss options.

Back to top

 

4. Dataset screen

Selecting a dataset from the dataset list will show the dataset screen. If it is the first time into this screen since creating the dataset you will need to upload farm analyses into the dataset by selecting the "Upload xml/ovr farm analyses" button.

This screen shows all farm analyses that are present within the dataset. A dataset is a collection of farm analyses that can be run through the model (a model run) with or without a selection of model parameter changes.

The results of each model run are stored and model run results can be compared to see the effect on the specified model parameter change.

Model run results at the analysis and block level can be downloaded into csv files.

Back to top

 

5. Failed analysis section

All farm analyses that failed to upload into the dataset are shown here with a message stating what went wrong on the upload. These farm analyses will not be part of the dataset.

If there are no failed uploads a message will be displayed stating this.

 

Back to top

 

6. Analyses section

All farm analyses that are in the dataset are shown here.

Analyses can be added to a dataset here by either uploading xml/ovr legacy files or copying an existing analysis in the dataset.

Analyses can also be added to a dataset from within a farm by adding the analyses to the dataset from the list of analyses in a farm.

 

• Upload xml/ovr farm analyses button - To upload additional Overseer Legacy xml or ovr files to a dataset select the "Upload xml/ovr farm analyses" button below. Each analysis that is uploaded will be created within it's own farm. Analyses that are uploaded into the dataset after model runs has been executed will not be re-run for prior model runs. Lysimeter files can also be uploaded here. An uploaded lysimeter file will also be represented as analyses below.
• Create lysimeter inputs button - Along with lysimeter file uploads, lysimeter data can also be entered manually. Entered lysimeter data will also appear as an analysis in the analyses list below. More information about creating lysimeter inputs is below.
• Edit measured values button - Measured values can be entered for monthly drainage for each block within an analyses of a dataset. See the "Enter measured values (actual)" section for more information.

The following functions are available for each analysis within the dataset:

• Copy analysis - Copies the analysis into the dataset. The copied analysis is created in it's own new farm.
• View lysimeter - Views the lysimeter data of the analysis in the dataset.
• View analysis - Opens the analysis in another tab to view the analysis details and the ability to edit the analysis.
• Delete analysis - Deletes the analysis from the dataset. Model run results will remain for the deleted analyses but will not appear in any subsequent model run executions.

Back to top

 

6.1 Uploading farm analysis files

Uploading files into a dataset will show the following screen. You must enter a name for the dataset and then select the files to upload. Files may or may not contain results.

It will automatically process each of the files and show the remove file link once it is uploaded. Selecting remove file will remove it from the screen but will not remove it from the dataset as it has already been added to the set.

Selecting cancel or done will close the modal and show the new dataset in the list. All files will be uploaded even if cancel or done is selected while processing.

You can also upload more analyses into a dataset when you are within a dataset.

 

Back to top

 

6.2 Enter measured values (actual)

Measured (actual) values for each block within each analysis of the dataset can be entered here:

Measured values can be entered for the following fields; drainage, N/Ha, Soil N, Plant N, Root N, Stover N, Soil Moisture and Soil Moisture 600mm (these values may be added to).

If any measured values are entered here on any of the blocks within the dataset, when a model run is executed.  

 

Back to top

 

7. Model runs section

A model run is when all analyses within a dataset are run through the model or sub models.

Model runs contain the results of the model execution for all analyses in the dataset at the time of the run. This information is stored independently of the farm analysis and so will not change even if the analysis is changed or removed from the dataset.

When a model run is created and executed, a table of model runs will be created at the bottom of this section. The blue bars can be expanded to show the individual results. You can select a model run within this table to view it's results.

You can delete model runs from this table.

Deleting model runs only delete the runs results, it does not delete the analyses within the dataset.

To create a model run select one of the following links.

When creating a model run you can enter parameters you want to use in specific parts of the model to use in place of what the model currently uses.

Each model run has an engine version against which it was run. As Overseer Limited releases new versions of the model you will be able to run the dataset against that new version and compare the results. This assumes that the analyses within the dataset have not changed in that time.

 

Back to top

 

7.1 Full model run

Runs all analyses in the dataset through the current Overseer model version using specified parameters and results stored. The following model parameters can be specified for a model run. These model parameters will be used when the analyses are executed through the model instead of the ones used in the current Overseer model.

If you do not see the model parameters your organisation will need to be set up in OverseerSci to see them so will need to get in touch with Overseer Limited.

Model run parameters:

Urine N leaching model parameters (see the "Urine patch sub model" technical manual chapter for more information on these parameters)

• Proportion of N in urine
• Proportion of urine factor
• Urine load
• Remove denitrification
• Initial fertiliser immobilisation

Crop model parameters (see the "Crop based nitrogen sub-model" technical manual chapter for more information on these parameters)

• Percentage of DM
• Energy produced (MJ/kg DM) 

Pasture parameter settings (see the "Characteristics of pasture" technical manual chapter or more information on these parameters)

• Pasture type
• Topography
• Clover level
• Pasture utilisation
• Average pasture N concentration

 

Monthly results:

If "Include monthly results" is selected, analysis and block monthly results are also reported on and stored.  

 

S-map soil sibling updates:

If this is selected each analyses in the dataset is run with the most up to date S-map data from Landcare. When this is is run two model runs are created so they can be compared against;

Soil Update - the model run that has run each analysis through using the most up to date S-map soil sibling data

Modelled - the model run that has run each analysis through using the soil data currently in the analyses in the dataset.

 

 

Back to top

 

7.2 Lysimeter model run

Runs all analyses in the dataset through the lysimeter model and results stored. Lysimeter data for each lysimeter analyses can be entered by selecting the "View lysimeter" button for each analysis. These model runs cannot be compared against other model runs. Your organisation needs to be set up to use this function in OverseerSci. If you do not see this function and want to know more about this please get in touch with Overseer Limited.

 

Back to top

 

7.3 Animal ME model run

Runs all analyses in the dataset through the current Overseer "Animal ME" sub model and outputs of the sub model stored. The results of this type of run are displayed at the block. These model runs cannot be compared against other model runs. 

Your organisation needs to be set up to use this function in OverseerSci. If you do not see this function and want to know more about this please get in touch with Overseer Limited.

 

Back to top

 

7.4 Sensitivity analysis

There are three different types of run types when a model run is executed;

• Imported - If files that contained results were uploaded into this dataset, an “Imported” model run is created. This model run contains the results from those imported files.
• Modelled - The results of analyses run through the model where using model parameters or not is a "Modelled" run type. Every model run apart from a "Sensitivity" model run should create one of these.
• Measured - The results of analyses using the measured values entered instead of the modelled values when executing the model. Every model run apart from a "Sensitivity" model run should create one of these if measured values are entered against analyses in the dataset.
• Sensitivity - The results (only in csv format) of all combinations of sensitivity attributes that are run against all analyses in the dataset. This type cannot be compared with another model run.
• Soil Update - The results of a full run that uses the most up to date S-map sibling data from Landcare.

Back to top

 

You can compare two model runs by selecting them both from the list. One of the runs (the first one selected) will be highlighted in red and the second highlighted in blue. This will compare each analysis and block and show where the nutrient budget results differ as shown below. Model runs can be compared against other model runs by selecting two model runs.

Only full model runs can be compared with other full model runs.

If an analysis with the dataset does not generate results for a model run, the model error messages for each analysis will be shown in the first section of the results - as shown below.

 

If there are differences between results of the compared runs a "Differences" column is shown which lists the differences at the nutrient budget level - see below.

 

 

Back to top