Moderator: mmckerrow


Postby shannon.li » Mon Jun 01, 2015 4:32 pm

Can I enter observed ET instead of PET?

Can you tell us a bit about data requirements and preparations for input into Source?
There are two broad data types that are required as inputs into Source: spatial data and time series data. The time series data, typically, is entered as a CSV file (comma separated values). You will have a date in the first column and that date will represent the time step of the model, typically daily. The other columns of the time series CSV will have the values of the data you are entering and with these values you could have several columns. It is possible to put the ET and the rainfall data all in the same time series CSV, then just choose which column to import. In terms of spatial data, a critical thing to remember about importing the digital elevation model, for building your own catchments, is to fill the pits in the digital elevation model before you import it into Source.

Where in Source do you optimise the parameters?
The calibration wizard for Source can be accessed from the top toolbar of the software and it is called flow calibration analysis.

Can you create spatial data within Source, or only import from pre-processed data in the GIS program?
In Source, you can save out the catchments that have been generated, but those will be based on pre-processed data that you have imported into Source. To create spatial data you have to prepare your data in a GIS and then import it into Source.

How do you add nodes manually in Source?
Nodes are loaded from an existing shapefile but you can add outlets manually in Source. You can choose an outlet at any point along the river system and it will produce a set of catchments.

Can you output the catchment boundaries and nodes that are generated? To what format?
You can save out the sub-catchments and the node locations by right clicking inside the window where the catchment generation takes place. They are both saved as shape files.

Can you tell us a bit about water quality, either existing or future developments?
At the moment there is a lot happening in the water quality world in Source and we are proposing to extend this functionality to include the Qual2E algorithms that are available and commonly used to model water quality in modelling programs. There is provision made for constituent modelling inside Source. We will be running a course in February 2016 specifically to speak about water quality in Source. To find out more and register for this course, visit http://www.sourcetraining.ewater.org.au

Can we analyse the effects of land use change in the stream flow, changing the location and size of HRU’s and ET?
Yes, you can change your ET at any time. In the same place that you originally added the rainfall and ET, you can also change these time series and add new data at any stage. Similarly, you can change functional unit areas manually. However it must be noted that at all times, the functional unit areas per catchment must always add up to 100% of the catchment.

What do you see as the benefits of the rainfall runoff modelling in Source verses other hydrological models?
The greatest benefit of using rainfall runoff modelling systems inside Source is that one actually has an integrated platform in which you are modelling. Besides viewing catchments in the geographical view, Source also has a schematic view. From this schematic view, one can actually start accessing a lot of the functionality that is only available to schematic views, such as adding water users.

How do you calibrate the model?
To calibrate the model you use the flow calibration analysis wizard.

What should be the resolution of the DEM?
The resolution that you choose for your Digital Elevation Model (DEM) depends on the size of the catchments and the resolution of the streams that you want to generate.

When will the Qual2E functionality be made available?
It will be available in late 2015 and will be coming out in one of the later beta versions.

Is the runoff model linked to sediment module and output in Source?
Yes, it can be linked to a sediment module and output in Source.

What is the maximum size DEM Source can handle in terms of nodes or megabytes?
Source can handle up to about 750 MB in size.

What you showed indicates only the change in percentage of the functional units, does the location of the units affect the results?
No, we do not take the location of the units inside the catchments into consideration. We use the size of the units to model the runoff coming from each functional unit. So the location of the units does not affect the results, although location can be taken into account in the way that you create your functional units. You might want to use something like slope, as an extra factor in the functional units, to actually represent the location of that functional unit inside the topography of your catchment. Then you could have a different parameter set for functional units that are on steep slopes versus those that are on flatter slopes.

Is there any snow melt model?
We are considering looking at various snow melt models, but currently there is no snow melt or glacial melt model in Source. However you could build one and add it as a plugin. CSIRO has developed a prototype snow melt plugin which may be added in the community plugin space for Source. It is attached to the GR4J rainfall runoff model, as demonstrated in our Catchments webcast

Is it catchment averaged?
It is functional unit averaged.

Do you have SRM model in Source?
No, we do not have the SRM model in Source but again this and any particular model can be done using a plugin.

Is there a library of plugins users can access?
Yes, there is a library of plugins that is accessible to users in Source, that will be downloaded at the same time that you download your Source version. Go to Tools > Plugin Manager, this will take you to a folder which contains a list of plugins that you can browse through. There is also a list of community plugins that can be downloaded together with the Source download.

What does the ordering drop down do?
The ordering drop down is typically used in a schematic model and is used to simulate ordering water throughout the system.

Can you please explain the evaporation rainfall data files? If it is a CSV, is it only point data or is it possible to use gridded data if your catchments cover large areas?
Yes, it is possible to use gridded data. We have a tool called the ‘climate data import tool’, which can be used to import rasters of rainfall and ET data.

Does Source accept a digital elevation model in negative numbers? Is it better to use integers or floating numbers? How about the projection?
Yes, Source will accept negative numbers in a DEM. It is generally preferred to use floating numbers for DEMs, but sometimes one only has integers (e.g. SRTM data). Source accepts both. Source considers the digital elevation model that you import, to be in whatever coordinate space you have imported it. However, because that digital elevation model is being used to calculate catchments and the parameter we are interested in from those catchments is the catchment area, the digital elevation model has to be projected.

When you add water user nodes in the schematic view, do these appear in the geographic view?
Yes, but as we have just spoken about projections, those water user nodes need to have eastings and northings assigned to them. These must be the same projection as the digital elevation or catchment data and then they will appear in the geographic view.

How are the crop parameters used?
There are a number of cropping models that are available in Source. This is starting to move into the schematic world, rather than just the catchments world. In the Node panel, you can see that once I edit my water user to be an irrigator, I can assign various crop models to that water user.

Can you show adding nodes in the schematic and viewing them in the geographic view?
Anyone who would like to know more about this, please contact us for a demonstration, or you can attend a course to see this in more detail. Visit http://www.sourcetraining.ewater.org.au

How does Source decide on the sub-catchment breakdown?
The sub-catchment breakdown is according to the size of catchments that you choose when you are setting up your model. In the input screen for the digital elevation model, there is a minimum sub-catchment area that can be entered. Node locations (e.g. dams or gauges) can also be used to force catchment boundaries.

Why have you used the GR4J model, is there any specific reason?
The reason we chose the GR4J model is because it only has 4 parameters and is a quick model to run. So for demonstration purposes and for training purposes, it is a good model to use. However, the GR4J model has been shown to work very well in varied types of catchments. We have actually applied it here at eWater, in a number of different catchment sizes and rainfall/runoff regimes.

Is there anything in Source regarding the climate change impact assessment?
You can use Source to simulate climate change impacts by producing different scenarios with different rainfall inputs. In Source there is also a stochastic analysis tool, which will generate various rainfall scenarios and then run a rainfall/runoff model to produce flows per rainfall scenario

Is this standard Source software free for us, if not how much is it?
The standard Source software is free and there is a public version available from our website. The version that I have used to demonstrate to you today, has used only freely accessible features. If you would like to purchase a full software package or support & maintenance with your software, you could speak to Meryl McKerrow by visiting our website at ewater.org.au.

Has any kind of QA been undertaken on the available plugins?
There has been some Quality Assurance (QA) on the standard available plugins, however eWater does not support or in any way do QA on any of the plugins you might import from the community plugins folder.

What rainfall runoff models are suitable for urban catchments?
eWater has an urban product called MUSIC, which specifically focuses on stormwater in urban areas. However if you want to use the catchments model—for larger areas—we have the SURM model, which is an urban rainfall runoff model.

What is the minimum time series of precipitation and PET you would recommend to calibrate the model?
I would recommend an absolute minimum of 10 years but it does depend on the data you have available.

Is there an advanced water system planning course coming up in November?
Yes, and questions on water users and the types of crop models which are available, as well as ordering, will be addressed in that advanced water user planning course.

Can Source be used as a forecasting model for long as well as short term forecasting?
Yes it can be. More on this at our water planning course, visit sourcetraining@ewater.org.au

Do Source and MUSIC products have the same results if used for the same catchment?
Yes, they have been tested. The SURM model is also part of MUSIC and we have tested that they produced the same results.

Is SedNet still available or fully integrated into Source?
SedNet is available as a community plugin, but it is not totally supported by eWater for that reason, however is still being used. Dynamic SedNet—a daily version of SedNet—is being used extensively by the Queensland government to model sedimentation of the Great Barrier Reef.

Is setting the time step set by the input data and is there a statistic method of changing the daily data into hourly data?
Yes, setting the time step can be set by the input data. However, currently Source will not automatically change your data for you.

Do you see Source replacing MUSIC in the future?
No, not at all.

Can Source run on a sub-daily basis?
Source can run according to the model that is being used, so if one is using a model that is relevant for sub-daily running, then yes Source can run on a sub-daily basis.

How can you apply the model to other catchments different to the one you calibrated the model on, would you only apply the same parameters and assume it works?
There is a lot of theory behind applying calibrated parameters on other catchments and you would have to proceed cautiously. However, yes, often one can apply the same parameters to other catchments that are similar to the one you ran your calibrations on.

Can Source be used as a flood forecasting model?
Yes, although Source is not a full hydrodynamic model.

Should we incorporate rainfall as (mm) or (mm/day)?
Source currently incorporates rainfall as mm, i.e. as a depth. The reason why we have implemented it in this way is because some of the models request it in that format. Also, we don’t always run Source at a daily level. So converting it to a depth before you bring the rainfall or evapotranspiration [data] into Source means that you can implement it at timescales less than a day.

How well does Source work for urban catchments?
In Source there is SURM, which is an urban modelling package for smaller urban catchments. You would specify your functional units and rainfall runoff model according to those which are relevant in your urban catchment. We also have a tool specifically designed for urban modelling, which is called ‘MUSIC’.

How do you calibrate models for the different functional units to optimise parameters?
Source provides an option for flow calibration analysis, and within those options it gives you the ability to choose parameters as well as specific functional units or catchments those units are going to be calibrated on. In Source we would call that subset of parameters a meta parameter set.

I want to introduce a reservoir storage within the catchment. How is the node added?
You can add a storage as a node initially when you create your catchment. It will be created as a confluence node which can then be changed to represent a storage. Once that storage is created, you can configure the dimensions and outlets of the storage.

Are the data sets in Source all taken from BoM?
Commonly, if you are in Australia, you should be able to get rainfall and evapotranspiration datasets from the Bureau of Meteorology. You can also get these datasets from the various state jurisdictions websites.

Could you please explain a bit about rainfall runoff models available in Source and about widely used models in Australia?
A number of the models in Source have been developed, e.g.:
  • AWBM – Australian Water Balance Model
  • Ihacres Classic – developed specifically for Australian purposes
  • Sacramento – an international model but is widely used in Australia
  • SIMHYD – typically a locally developed model
Is it possible to create/edit catchments through the schematic scenario?
Currently no, you cannot edit your catchment layout in the schematic scenario, but you can view it and add nodes and links.

What is the best format of climate data to be taken up by Source?
Often the format of the data depends on what you have available. Source provides for time series data inputs, where you would get data that has been measured at a local rain gauge, or at a local evapotranspiration site. Source also makes provision for climate data to be loaded in as rasters. For instance in Australia you can get data for rainfall and evapotranspiration from the Bureau of Meteorology. You can load these rasters / grids into Source. There are also remotely sensed climate data available online, such as TRMM or Aphrodite.

SIMHYD has a version for routing, how about the other models? Do they have a routing factor?
Some of the models have a factor for lag, and also routing can be applied inside Source on a link. For example you can edit a link in Source and apply routing at that link.

Is there a general catchment size that Source has been developed for?
There isn’t a specific size that Source has been developed for. Source has been used on very large catchments, and also with the example of SURM, you can use it for smaller urban catchments.

Could it be used to model water balance in catchment without considering rainfall as the main water source?
If you want to do that and you have other water sources, we recommend that you create a schematic scenario because rainfall runoff modelling does need rainfall as an input. But if you have your flow existing you can create a schematic scenario and then load the existing flows.

What is the benefit of selecting the GR4J model?
The GR4J model has proved itself to be a good model, despite the small number of parameters. At eWater we have found that it performs just as well as some of the other models, such as Sacramento which have more parameters. So GR4J is often a good model to start off with when you are doing your rainfall runoff modelling in a catchment.

Please recommend the best way to introduce routing, i.e. using SIMHYD with routing, or define routing at the links, or any other way?
We recommend using the model parameter in Source.

How/when is the model time-step defined? Is it simply assumed from the rainfall time series?
Model time-step is set before a model is run. It is set using the Run > Configure menu option or the Configure option on the toolbar. It defaults to daily but other time-steps can be chosen.

Can Source deal with date-time stamp for sub-daily data?
Yes, assuming the model to be used is relevant at a sub-daily time-step.

When selecting EMC/DWC constituent generation model, how is slow flow and quick flow defined?
Some rainfall-runoff models – e.g. Sacramento – do this inside the model. GR4J outputs all flow as quick flow, so filter and shape parameters K and c have been added in Source to perform this function if required. More detail about these parameters can be found in the 'Observed Catchment Runoff Depth' section of the Source Scientific Reference guide. See https://wiki.ewater.org.au/display/SD37/Scientific+Reference+Guide (login required).
Posts: 4
Joined: Wed Feb 22, 2012 3:30 pm

Return to Catchment Management Using Source

Who is online

Users browsing this forum: No registered users and 1 guest