PRMS National Hydrologic Model

Lauren Hay: This is a recording of the National
Hydrologic Model overview for the PRMS training.

So the National Hydrologic Model depends heavily
on getting information from the GeoData Portal.

That’s where we get all of our climate information.

And it also depends very heavily on the Core
Science Systems' super computer.

I like to point these things out because we
wouldn’t be where we are right now without

these two things right here.

We develop the National Hydrologic Model to
give us estimates of water availability, changes

in timing and source of flow, uncertainty,
the assessment of climate and land use change,

and we want to be able to get predictions
at the local, regional, and national, scale.

So the components of the National Hydrologic
Model.

We’ve got models, we’ve got the geospatial
fabric, we’ve got climate data, and we’ve

got calibration and parameter transfer.

Right now, the models that are currently running
within the National Hydrologic Model structure

is the Precipitation Runoff Modeling System,
which this entire training video sequences

is talking about.

And we’ve also got the Monthly Water Balance
Model, which is a much simpler monthly time

step model.

Both of these models are running on the Geospatial
Fabric.

These are catchment derived modeling units
linked to a stream network.

I suggest you watch the presentation by Roland
Viger for more details on the Geospatial Fabric.

But, essentially, we’ve taken the NHD plus
catchments and we’ve aggregated them to

about 110,000 hydrologic response units across
the country.

And these are the modeling units that we are
running PRMS and the Monthly Water Balance

Model on currently.

So the third component is the Geodata Portal
and that’s what we are using to give us

our climate data for the country.

You can use any sources you want, but right
now when we distribute the models within the

National Hydrologic Model, they will have
sub-sets of the DAYMET gridded data set right

now.

And then the fourth thing is the calibration
and parameter transfer.

This is probably the part that is taking us
the most time and that we are currently working

on.

So, why are we calibrating and trying to transfer
parameters?

This just shows you, a figure from Kiang and
others, and it shows you the gage drainage

areas that are less than 20,000 square miles
and those are in light blue and those are

considered reference quality streamgage drainages
across the country.

I think this is somewhat questionable.

I don’t think that these are probably what
us as modelers would consider as reference

quality, but in order to get a distribution
of streamgages across the country, I believe

they included things all the way across the
country.

So these are the “least effected” streamgages
across the country.

Maybe is a better way to think about this.

Although people might argue with me about
that I suppose.

So that’s all we’ve got for calibration.

Historically, what people have done when they
develop national models is they taken applications

from across the country and basically pulled
all those applications together and they have

all been developed for different reasons.

And they end up with a patch work quilt of
model parameters.

They are not very physically realistic.

And then a lot of times you’ve calibrated
a model to a gage and then you transfer those

parameters to another gage and you could very
easily be getting the right answers for the

wrong reason when you start transferring model
parameters that way.

So we need to be think of a better way to
calibrate a model nationally.

We’ve got the Monthly Water Balance Model
and we’ve got the Precipitation Runoff Modeling

System.

So the Monthly Water Balance Model, there
is now a paper in review, “Parameter Regionalization

of a Monthly Water Balance Model for the Conterminous
United States”.

Andy Bock is the lead author on that, so you
can send correspondence to Andy if you want

a copy of that paper.

Then we are going towards a methodology for
PRMS and there is a paper out by Markstrom,

“Towards simplification of hydrologic modeling:
identification of dominate processes” and

there’s a paper that you can get from Steve
Markstrom here as well.

So for the Monthly Water Balance Model, we
have run it for current and future conditions

and we have regionalized the parameters.

You can get the information on our portal.

Eventually we would like to have the National
Hydrologic Model be accessible through that

portal as well.

Currently now, what you can do, is you have
the National Hydrologic Model setup and we

can think of this maybe as our coarse resolution
HRUs.

Right now we can see that the hydrologic response
units here, I highlighted one of those, and

I highlighted a streamgage here.

So you can pull a model out of a model and
then you can run that model as it is out of

the National Hydrologic Model.

Or you can actually pull a model out of that
and you can actually nest the finer resolution

NHD plus catchments within there.

Then you can either run that or put it back
into your larger model.

And then in the end we would like to be able
to have that go right back into the original

structure of the model.

We are working on that.

If you want to have a piece of the National
Hydrologic Model to use for your application,

you need to just click on the Help button
here and tell us what it is you’re looking

for and we will pull the model and send it
to you.

Thank you very much.