Basic Rating Development Using GRSAT

Video Transcript
Download Video
Right-click and save to download

Detailed Description

This video demonstrates how to develop a single offset surface-water rating curve in GRSAT.


Date Taken:

Length: 00:20:14

Location Taken: Salt Lake City, UT, US


Hi, this is Megan Poff, and I'm field office
chief in Henderson, Nevada.

I will be present a primer on rating develop
using GRSAT.

The intent of this primer is to show a lot
more of the rating development and analysis

tools, including working with the site visit
pane, and other tools in the zoom pane.

Other than working with multiple offsets,
by the end of this primer you should know

most of what you need to know to develop basic
ratings in GRSAT.

All right, so we're going to load the rating
development primer.

Going to hit open here, and now typically
you would hit from database.

But for this exercise, we're going to be starting
from a file that's already been created.

From file, there's my primer, open.

And read data.

And here's all my data.

Now what are you noticing right off the bat
with this data?

It's about the offset.

Offsite should be greater, right?

See how the data are, on the screen are showing
kind of a bowl shape.

We need more offset in this case.

Well, to even have more offset we need to
actually set-up a rating in GRSAT first.

Notice over here, where my mouse is, it says
number 0000W?

That means that we don't have a rating set-up
here yet.

So let's set-up the rating.

I'm going to go to set-up and rating.

It brings up the rating set-up box.

Now some of this stuff here at the top, it
tells you data descriptor, the type of rating

code you're making that's stage Q, which is
correct for any kind of stage discharge rating.

We'll just leave it logarithmic, that's type
of expansion that we would like.

Rating number, I'm going to give a nice rating
number of 1.0.

Now for those of you who know me know I love

So I'm going to put a rather flippant remark
in here.

Now obviously you would want to a better remark
in, something that is more descriptive of

why you're making this rating in the first

But just for purposes of demonstration we'll
start with this.

Hit apply.

And now we have rating 1.0W. The W is for

And also up here it says rating 1.0.

Okay, first thing we need to do in developing
our rating is to select a group of measurements

that we would actually like to use with the

So, I'm going to drag my site visit data pane
out so I can actually see what I'm doing with

it, and I'm going to expand this guy.

Now, you can sort all of your measurements
by each of these columns in GRSAT.

They're already automatically sorted by the
date and measurement ID.

I'm going to sort these guys by control condition.

That's this button right here, control condition.

And I'm going to get rid of any measurement
that was made with some sort of backwater


So that would be algae, debris of some sort.

So couple of different ways you deselect measurements.

You can either do them individually.

Click, click, click, click.

And you'll notice as I'm doing this, over
here, it's greying out those measurements

I am deselecting.

Click, click, click.

Well that's fun and all but you will notice
that would get rather time consumer if you

have to do a lot of measurements.

I'm going to scroll down, see these are all
clear control condition measurements.

Scroll down, what do we have here.

Oh, here's some light debris.

So rather than clicking out these individual
measurements, I'm going to click and drag

down, highlighting all of the light debris
and unspecified measurements.

And I'm going to hit this nice X up here.

Now did you guys notice when I hovered my
mouse over the X this little dialogue that

said "unselect measurement points" come up.

That is a very useful thing in GRSAT.

If you ever forget what you're doing, forget
where you are, forget what a button does,

just hover your mouse.

All right, so I'm going to unselect those
measurements that I just highlighted.



So by now, you've noticed that all of these
measurements down here have different colors

associated with them.

There's some blue and some green and some
pink and some grey ones.

And what does that mean?

Basically what GRSAT is doing is it is giving
each measurement a color classification.

The default is by quality.

And how do I know that?

I'm going to go click on this button up here.

And it says, okay, your classification is
based on quality.


So any excellent measurements are yellow,
anything good is green, fair is blue, poor

is pink, and if there was no quality specified
in the field then it's red.

There are other preset classifications I could
choose if I wanted.

For example, control condition.

So anything with ice, for example, we could
really easily see and get rid of.

Same with debris.

We could also classify based on the difference
from the rating, the percent difference from

the rating.

This T differ thing that actually means true

And that is the difference with any shift
supplied for each measurement.

You could even do a comparison between quality
and true difference.

Pretty cool stuff.

So I'm going to put this back on quality for
our exercise here, and say okay.

So while we're talking about classifying measurements
and colorizing some of our measurements, say

I had some recent measurements that I really
wanted to show on my new rating.

Let's do a quick demo.

I can actually color these measurements a
different color.

So I'm going to sort the measurements based
on date.

See if I click up at the date there you get
this little triangle.

We have them sorted by date.

And I'm really most interested in the measurements
that are made in 2004.

So I'm going to click and drag to highlight
these guys.

See how all the '04 measurements are highlighted?

Now I'm going to go up to the color wheel
here, this is for customizing colors.

So I'm going to make myself a nice custom

You can click anywhere in this area, or you
can choose one of their preset basic colors,

like periwinkle that's a nice one.

You can already see that I've predefined this
sort of aquamarine color down here but I'll

define another one for this exercise.

Let's see, how about kind of a nice grey green.

Something like that?

Anywhere you click in here you can keep redefining
your color if you want.

You can also type in the hue saturation red,
green, blue area.

So I'm going to pretend I like this color,
add to custom colors, my colors changed.

I'm going to click on that color and say okay.

It's now colorized these particular grouping
of measurements as this kind of grey-green.

So if I wanted to make a rating just based
on those measurements, I could do that.

I'm going to reset this though.

Back up here, classification based on quality.

I hit reset, it changes everything back, and
now we're back to everything being color coded

based on quality only.

I'd like to show you guys a couple of other
ways for unselecting and selecting measurements

that you would like to use or not in your

So, first thing I'm going to do before I get
into that.

We need to put the site visit data pane back
into GRSAT windows here so we can actually

see what we're doing down on our rating zoom
one plot here.

My favorite way to do this is just to go up
here to view, and click default rating shift


This is like your home screen, this is your
back button.

Anytime you get rid of too many windows or
forget where you are you can always get back

to the beginning by just going up to view,
default rating shift layout.

So now I've clicked that.

Site visit data is now docked where it belongs.


So I'm going to take look down here at rating
zoom one, and we have some outliers.

This guy here, for example.

Now did you notice when I hover my mouse over
the measurement here I have a little pointer


If I right click on that, it's going to tell
me some measurement details, very nice.

I can also left click on that, and it'll take
me to that measurement here in the site visit

data pane.


So I'm going to get rid of it graphically.

I could just uncheck it at this point, but
I'll show you the graphic way instead.

Again, hover over any of these buttons, and
it will tell you what it is for.

We're going to click on this guy, which is
unselect measurement points.

And I'm simply going to draw a box around
the measurement I want to get rid of.

And it's gone.

Did you notice here, how the window also stayed
in sync.

It unselected your measurement for you.


Just to show you how you would select a measurement

Say I accidentally didn't want to have that

You go here to select measurement points,
draw a box, bam, it's back.

But I'm going to-- going to get rid of that
again because we don't really want to use

it to define our rating.

You can also define measurements that you'd
like to use using the time series view.

So if you go down to time series view tab,
we're going to get rid of all of the old measurements.

I'm going to click on unselect measurement
points, and I'm going to draw a big box around

everything made prior to about the year 2000.

It takes a sec because there's a bunch of
measurements in there but check it out, they're

all gone.

All right, so by now you guys are probably
wondering, yeah this is great and everything.

I've unselected it, but it's still showing

I can't see what I'm doing.

You can actually make GRSAT not display your
unselected measurements.

And that is this button up here.

It says show/hide unselected measurements.

If you click that, gone.

We are almost ready to make our new rating.

The last thing I want to do before I actually
start putting rating points on this thing

is I want to deselect all the measurements
below 1.6.

There's only one left.

Let's check it out, I'm going to sort by recorded
values, which is my gauge height.

You can see that there's one measurement that
is selected that is below 1.6, and the reason

that's selected is because we had deselected
everything made before the year 2000.

This one was made in 2002.

So I'm going to turn him off, gone.

All right, it's offset time.

Offset manager is the tab you're going to
want to click here.

And you'll notice here, when you first pull
up your offset manager, we're plotting to

a single plotting offset of other.

You don't want to plot to other when you start
putting in your offset.

I'm going to click on offset 1.

It's going to automatically default to 0,
it doesn't know what the offset is.

That's what we have to find out.

And now I'm going to click on the single plotting
offset of offset 1.

Now one thing I have turned on here that will
be really useful in some circumstances and

a little bit annoying in others.

If I go up into set-up, automatic rating scale
is turned on.

I'm going to leave that turned on for this
part of the exercise.

All right.

Now we need to find some offsite that will
make these data apart to be mostly straight.

You can click on the down and up arrows here,
or you can even type in this box if you want.

See how to GZF you could type in your GZF
as a first cut you're going to offset.

I don't know what the GZF is here, so I'm
just going to graphically click, click, click,

click and watch the graph down here as it
becomes straighter and straighter.

From joint two, that looks pretty good.

1.3 that looks pretty good as well.

1.4, eh, maybe that's bowing up a little bit.

I'm going to try 1.5, .6, .7-- This is definitely
too much offset.

I'm going to take it back to 1.3, 1.3 I think
I like the best.

Let me clear this out and show you guys something.

Did you see how I just typed in there 0.00
and it went back?

I'm going to turn off that automatic rating
scale that I told you about here in set-up.

Check it out, see it's off now.

Now I'm going to start iterating the offset.

See how I'm like losing measurements?

This is kind of annoying.

You could go here to auto scale every time
you iterated the offset.

Auto scale!

Or for this part you can just have automatic
rating scale set.

Did you guys see how I turned that back on?

Just set-up, automatic rating scale.

All right, so I'm going to go back to that
1.30 which I really liked.

Let's make a first cut with rating.

Down here in the rating zoom one pane, there's
a whole bunch of tools for making rating points.

I'm going to use add rating point.

And this is my absolute favorite way to make
a first cut at a new rating at a site.

Click on a point, you get this little point
with an arrow and a plus.

I'm going to click here below my lowest measurement.

And then I'm going to go up here and click
just above my highest measurement.

Did you guys see, now suddenly we have shift
diagram that popped up?

Now I'm going to start moving those rating
points around to see if I can get a nice scatter

of measurements around the zero shift point.

This guy, this little curve with the green
arrow that's your move rating point tool.

Going to click on that.

And see how you can just graphically drag
around your rating point.

Well that's actually pretty good.

We have pretty decent scatter around the zero.

Not too bad for just using two points, right?

There are other ways to come up with a first
cut at a rating.

Let's blow up what we have now.

I'm going to select this guy here which is
delete rating points.

See how I now have a plus with an X and a
bunch of points still on the rating zoom one


I'm going to draw a box around all of that,
my rating is gone.

The shift diagram has now basically disappeared
because there is no rating in there.

Now we can use a regression to come up with
a first cut rating as well.

First thing I'm going to do here is I'm going
to go into initial descriptor point spacing.

Click on this.

We have a couple of options here: uniform
steps in log, uniform steps in linear.

We're mating-- making a rating on logarithmics
base, so we're going to leave that at the


I'm going to move my descriptor point spacing
down to five points.

We don't have that much of a range in stage
to go off of here.

And with the initial guess methods you have
regression or section control.

We'll leave it as the default regression,
section control is outside the scope of this


And click OK.

Now, we'll make our first cut rating using
the curve fitting tool.

Go ahead and click on curve fitting.

And I'm going to just draw a box around all
of my measurements.

And it come up with a rating for me.

So there's one, two, three, four, five points
like I just specified.

That doesn't look too bad.

And once again we have a shift diagram.

Now I do want to emphasize that using this
curve fitting option is not the best way.

Remember, rating development is not a statistical

This is part of the art of stream gauging.

I usually prefer to delete this thing.

Add points, move them around as we see fit,
as we need, and making a rating starting with

something like this.

But I did want to tell you the other options
that you have.

You can evaluate this rating with more than
just the shift diagram.

For example, up here in site visit data.

Let me sort all of these measurements by status,
which means that I just selected everything

that has check button-- box next to it.

And now I can see my percent differences.

R error is the percent difference off the

And as I move my points around, you can see
that change.


So I really like this rating that I just come
up with, and I want to evaluate it based on

the expanded rating table.

Where's that?

We actually have the ability to pull an expanded
rating table straight out of Aquarius and

GRSAT by clicking on this button up here,
which is the expanded rating table set-up


You can either pull the base rating or you
can pull from a shift.

I'm going to pull just the base rating.

Independent precision of .01, dependent significant
digits of 3, there's my name.

I'll just leave it as that.

And I'm going to click okay.

Rating table!

Just like you would get out of Aid Apps [unsure
of name].

So we can look at the difference in Q per
.1 units.


Now this thing should be identical to what
you would pull out Aid Apps.

There occasionally vary, be very small rounding

But they're pretty rare.

For the most part, this and Aid Apps, same

I'm going to get rid of this table.

And last thing we're going to do here, let's
check the rating slope using our slope computation


Down here in rating zoom one.

I'm going to click this tool, highlight, and
it gives me this error.

"Cannot compute.

Dependent variables in the selected segment
of the rating curve."

I think I selected too much.

Try it again.


It gave me a slope less than 2.

Which says, hey, we're in channel control
right now.

Now, obviously, if I were actually a hydrographer
on site I would see if that's correct or not.

I don't know, this is not my site.

But this is just kind of a little statistical
check to help you.

Okay, we need to put some dates on our rating.

What are we going to do?

Now, I want to tell you about one caveat.

There is a quirk in the current structure
of GRSAT that was discovered a little too

late to change.

You can't create a new rating from just measurements
and give it dates in the same GRSAT session.

What we would need to do is go up and save
this thing first, bring it back into GRSAT

in a new session to apply dates.

This is because the middle ware do not what
type of rating we were making when we pulled

the data.

Remember we needed to set the rating type

So it didn't know what current rating dates
to pass on to GRSAT.

This quirk only exists when we're developing
a rating from measurements only.

That's when you selected a rating of none
in the middle ware.

Whenever you actually bring a rating into
GRSAT, it gets alternating dates for the type

of rating.

So you can do everything in one session.

For example, you could bring rating 1.0 into
GRSAT, create rating 2.0, and giving rating

dates all in one session.

For now, I'll just show you what the rating
period manager looks like.

I can't really do anything here, but all you
do is double click, check rating period, give

it a start date and time.

I'm going to hit cancel for now, because like
I said, we just created a rating from measurements


We can't do it right now, right here.


So that concludes our demonstration of how
to make a first cut simple rating using data