Contamination Extension for Leapfrog Works: Demonstration and Live Q&A

The webinar will cover:

• An introduction to the Contaminants Extension for Leapfrog Works.
• The Contaminants Extension Workflow – with a brief overview of:
• Data Preparation and Visualisation
• Domaining and Exploratory Data Analysis
• Estimation
• Model Validation
• Reporting
• Live questions and answer session.

Overview

Video Transcript

[00:00:00.320]
<v Aaron>So hello everyone</v>

[00:00:01.153]
and welcome to today’s webinar

[00:00:02.400]
on the Contaminants Extension for Leapfrog Works.

[00:00:05.990]
Today we’ll be looking at the demonstration

[00:00:07.530]
of the Contaminant Extension

[00:00:09.083]
with time at the end of the webinar for your questions.

[00:00:12.460]
My name is Aaron and I’m a Customer Solution Specialist

[00:00:15.540]
here at Seequent, based in Brisbane office

[00:00:17.980]
and I’m joined today by Steve Law

[00:00:19.710]
our Senior Technical Lead for Geology.

[00:00:21.850]
<v Steve>Hello, I’m Steve</v>

[00:00:22.850]
and I’m a Senior Technical Lead

[00:00:24.520]
and I’m based here in (mumbles).

[00:00:30.180]
<v Aaron>Today we’ll first go through</v>

[00:00:31.280]
a bit of background on Seequent

[00:00:32.680]
and the Contaminants Extension.

[00:00:34.900]
I’ll then hand over to Steve who will take us through

[00:00:37.680]
the Contaminant Extension in Leapfrog Works itself.

[00:00:41.610]
After that we’ll go through any questions that anyone has.

[00:00:44.240]
So please if you do have any at any time,

[00:00:46.470]
put them in here at the question or the chat window

[00:00:48.890]
and go to webinar and we’ll get to them.

[00:00:53.604]
There’s some background on Seequent.

[00:00:56.070]
We are a global leader

[00:00:57.230]
in the development of visual data science software

[00:00:59.470]
and collaborative technologies.

[00:01:01.560]
With our software you can turn complex data

[00:01:03.610]
into geological understanding, providing timely insights

[00:01:07.220]
and give decision makers confidence.

[00:01:10.270]
At Seequent, our mission is to enable our customers

[00:01:12.640]
to make better decisions about their earth

[00:01:14.590]
and environmental challenges.

[00:01:18.430]
The origins of Seequent go back to the 1990s

[00:01:21.570]
when some smart people

[00:01:22.680]
at the University of Canterbury in New Zealand,

[00:01:24.720]
created the fast radio basis functions

[00:01:27.170]
for efficiently generating 3D surfaces from point data.

[00:01:31.460]
One of the applications that was seen for this

[00:01:33.300]
was a better way to build geological modeling.

[00:01:36.080]
And so in 2004, the company ARANZ Geo was founded

[00:01:39.490]
and the Leapfrog software was first released

[00:01:41.680]
to the mining industry.

[00:01:44.250]
Since then the company has grown and evolved

[00:01:46.580]
entering different industries

[00:01:48.240]
such as geothermal energy, environmental and civil.

[00:01:52.960]
And in 2017, ARANZ Geo rebranded to Seequent.

[00:01:58.000]
In 2018, Seequent acquired Geosoft

[00:02:00.670]
the developers of Geophysical Software

[00:02:02.520]
such as a Oasis montaj.

[00:02:04.520]
And then in 2019, acquired GeoStudio,

[00:02:07.710]
the developers of Geotechnical Analysis Software.

[00:02:11.930]
During this time Seequent has continued developing

[00:02:15.630]
our 3D geological modeling in Leapfrog

[00:02:18.620]
and collaboration tools such as Central.

[00:02:22.480]
Earlier this year,

[00:02:23.313]
Seequent was acquired by Bentley Systems

[00:02:25.940]
and has become a Bentley company

[00:02:27.433]
which opens us up to even more possibilities

[00:02:30.430]
for future development for geoscience solutions.

[00:02:36.310]
Seequent have office globally

[00:02:37.990]
with our head office in Christchurch, New Zealand.

[00:02:40.890]
And more locally we have offices in Australia,

[00:02:43.550]
in Perth and in Brisbane.

[00:02:48.240]
Like I mentioned Seequent’s involved

[00:02:49.970]
in number of different industries

[00:02:52.140]
covering things from a contaminant modeling,

[00:02:55.550]
road and toll construction,

[00:02:57.360]
groundwater detection and management,

[00:02:59.460]
geothermal exploration, resource evaluation

[00:03:02.560]
and so much more.

[00:03:08.210]
Today we’re looking at contaminant

[00:03:10.253]
or contaminated and environmental projects.

[00:03:13.340]
Seequent provides solutions across the life

[00:03:15.650]
of contaminated site projects

[00:03:17.060]
from the initial desk study phase

[00:03:19.140]
to the site investigation

[00:03:20.800]
through to the remediation, design and execution.

[00:03:25.070]
These solutions include our geophysics software

[00:03:27.320]
such as Oasis montaj and VOXI.

[00:03:29.930]
Our 3D visualization and geological modeling

[00:03:32.490]
and contaminant modeling in Leapfrog Works

[00:03:35.460]
as well as flow and transport analysis in GeoStudio.

[00:03:39.710]
Across all of this we have Seequent Central

[00:03:42.100]
a cloud-based model management and collaboration tool

[00:03:45.320]
for the life of the projects.

[00:03:48.090]
Like I said, today we’re specifically looking

[00:03:50.220]
at the Contaminants Extension in Leapfrog Works.

[00:03:56.040]
So what is the Contaminants Extension?

[00:03:58.670]
The Contaminants Extension

[00:03:59.790]
is an optional module for Leapfrog works

[00:04:01.660]
to enable spatial modeling of numeric data

[00:04:04.540]
for defining the concentration, distribution

[00:04:07.330]
and mass of the institute contamination at a site.

[00:04:11.910]
The Seequent Contaminants Extension provides

[00:04:14.670]
a transparent and auditable data-driven interpretation.

[00:04:19.900]
It provides interactive visual tools

[00:04:22.430]
with the extension that makes geostatistics

[00:04:24.760]
accessible for all geoscientists

[00:04:27.610]
ensuring that you and your teams

[00:04:29.120]
can work the best with your data.

[00:04:32.600]
It has built-in report tables

[00:04:34.730]
and cross-section evaluations

[00:04:36.270]
that are dynamically linked

[00:04:37.620]
to your plume models and your data.

[00:04:41.350]
So they will update automatically

[00:04:43.430]
as the investigation and the project evolves in new day

[00:04:47.010]
that comes into your Leapfrog project.

[00:04:50.320]
It also means there’s an interrogatable estimation tool

[00:04:54.730]
that provides easy access to the details

[00:04:58.260]
for how you’ve created your informed decisions.

[00:05:02.595]
In short, it allows you to take known data for your project

[00:05:07.070]
and analyze it with spatial variability

[00:05:09.500]
of the concentrations for your data

[00:05:12.390]
and it helps you define your projects.

[00:05:20.860]
So that’s what it is.

[00:05:22.410]
Why did we develop it? And why is it useful?

[00:05:25.800]
Well, we did it because we wanted to provide

[00:05:27.840]
these accessible best practice solution tools

[00:05:30.690]
for you to model your contamination plumes.

[00:05:34.290]
We wanted to make characterizing contaminated land

[00:05:36.810]
in groundwater, in a rigorous and auditable way

[00:05:39.730]
that combines 3D dynamic geological Models

[00:05:42.700]
with these best practice geostatistical methods.

[00:05:47.500]
The ultimate goal of estimation

[00:05:48.970]
is to combine the qualitative geological interpretation

[00:05:52.310]
of your project.

[00:05:53.680]
So your geological models

[00:05:55.270]
with the sparsely sampled quantitative data

[00:05:58.450]
to create spatial predictions

[00:06:00.130]
for the distribution of that contaminant plan.

[00:06:09.140]
The Contaminants Extension

[00:06:10.610]
coupled with Seequent Central provides a wholistic way

[00:06:13.877]
for the engineer and modeler for a project

[00:06:16.780]
to effectively understand

[00:06:18.580]
and importantly communicate the site conditions

[00:06:21.490]
with both internal and external stakeholders

[00:06:24.310]
in a collaborative environment.

[00:06:27.090]
The graphic here shows an overview of the flow

[00:06:29.610]
from Leapfrog to Central as part of the solution.

[00:06:34.350]
Starting at the base here and moving up

[00:06:37.690]
we can see that with your investigation data models

[00:06:40.982]
and estimations can be created by the modeler

[00:06:44.260]
and the environmental engineer for instance.

[00:06:48.310]
And using Central,

[00:06:49.600]
they can share this and collaborate on this

[00:06:53.000]
with internal team members such as managers,

[00:06:55.830]
reviewers or other modelers

[00:06:58.360]
to build up the understanding of the site

[00:07:02.040]
or as part of the project review process.

[00:07:06.420]
At the very top of this infographic here

[00:07:08.260]
we can then see how we can use Central

[00:07:10.330]
to engage with external stakeholders

[00:07:13.130]
such as the clients or other contractors

[00:07:16.650]
and you can bring them into the Central project again

[00:07:18.740]
to communicate in an interactive and visual way.

[00:07:24.820]
The Seequent solution for the contaminant projects then

[00:07:27.490]
is all about knowing the impact

[00:07:29.490]
of environmental contamination for your projects.

[00:07:32.880]
Through first seeing the problem

[00:07:34.360]
in a clear, intuitive and visual way.

[00:07:38.030]
The picture is worth 1,000 words

[00:07:39.440]
a 3D visualization of the site has to be worth a million.

[00:07:44.260]
With a Contaminants Extension,

[00:07:45.810]
it’s an interactive dynamic site model

[00:07:47.990]
you can create and understand

[00:07:50.130]
and aid in your analysis and your assumptions

[00:07:53.070]
and recommendations for your reports.

[00:08:01.340]
I’ll now hand over to Steve

[00:08:03.500]
who will take us through Leapfrog

[00:08:05.540]
and the Contaminants Extension.

[00:08:27.267]
So we’ll just switch over from the slides here. (mumbles).

[00:08:38.840]
<v Steve>Thanks very much Aaron</v>

[00:08:40.756]
for that great introduction.

[00:08:42.290]
So I’m going to go through the

[00:08:44.470]
Contaminants Extension Workflow.

[00:08:48.830]
The end result is a resource model

[00:08:50.820]
and through the whole process we have a look at our data.

[00:08:53.770]
We go through the mining,

[00:08:55.340]
run through estimations and validate and report at the end.

[00:08:59.650]
Throughout the whole process it’s not a linear

[00:09:01.960]
it can be a cyclical process

[00:09:03.840]
so that whilst we’re at any stage

[00:09:06.310]
we can always go back to our geology

[00:09:08.480]
and look at how things are relating to each other.

[00:09:15.580]
We’re going to go through a couple of key steps

[00:09:18.950]
that we need to consider whilst building up a plume model.

[00:09:24.070]
So the first part is being well aware of our data

[00:09:27.960]
and how we can visualize it very easily

[00:09:30.240]
within Leapfrog Works.

[00:09:31.980]
So we want to consider things,

[00:09:33.659]
where is our data and how is it located especially?

[00:09:37.500]
We have tools in-built that we can check

[00:09:39.640]
and make sure our data is validated

[00:09:42.490]
and we have really good visualization tools.

[00:09:47.360]
So within this project we have three models,

[00:09:51.440]
the first one is a geology model

[00:09:53.840]
and you can see here that it’s got outwash

[00:09:59.580]
and ice deposits until above a bedrock sequence.

[00:10:03.960]
And within that we have a range of vertical boreholes

[00:10:07.210]
and these have been sampled generally in isolated locations

[00:10:13.840]
for a chlorine contaminant

[00:10:16.030]
and that’s what we’re going to be evaluating.

[00:10:19.240]
As well as the geology model,

[00:10:21.440]
we have a water table model.

[00:10:24.870]
And this has been divided into a saturated

[00:10:28.040]
and made (mumbles) again above bedrock.

[00:10:31.740]
So one of the first things we want to consider is,

[00:10:34.910]
is there any specific way that we should domain our data?

[00:10:38.040]
To ask some of them, is chlorine associated with pathology?

[00:10:42.960]
Is it associated with water table or a combination of both?

[00:10:47.199]
And we have some tools within Leapfrog Works

[00:10:49.460]
that can show this.

[00:10:52.180]
So up in the boreholes’ database here,

[00:10:55.010]
I have a table with the contaminants

[00:11:00.320]
and then I’ve got a geology table.

[00:11:02.370]
We have a merge table function

[00:11:04.090]
so I’ve been able to merge these together.

[00:11:06.370]
And this enables me then to evaluate them

[00:11:10.700]
using some statistical tools

[00:11:13.120]
in this case I’m using a box plot.

[00:11:17.640]
And so this was here as a box plot of the geology

[00:11:20.360]
against the contaminant.

[00:11:22.610]
And we can see here that that is telling me

[00:11:24.910]
that the outwash and local ice to contact deposits

[00:11:28.310]
and the two, have the highest concentration of chlorine.

[00:11:32.310]
And then there’s a lot less amount

[00:11:34.350]
in the outwash deposits and bedrock.

[00:11:38.830]
We can also have a look at this as a table of statistics.

[00:11:48.536]
Excuse me.

[00:11:58.460]
An idea of…

[00:11:59.810]
We have a little bit total concentration.

[00:12:02.200]
I can see the actual values within each one.

[00:12:06.220]
One of the key things here is

[00:12:07.510]
looking at the number of samples

[00:12:09.010]
so we have only two samples within the bedrock.

[00:12:12.060]
We’d identified that the outwash and ice deposits

[00:12:14.920]
and the two had the greatest concentrations

[00:12:20.080]
and we’ve got the most number of samples in those as well

[00:12:22.650]
so we’ve got approximately 70 samples within the ice.

[00:12:27.310]
And we’ve got the mains

[00:12:28.780]
so we’ve got much higher means within ice

[00:12:32.660]
and very low within the outwash deposits

[00:12:35.760]
and again quite low

[00:12:36.970]
but there’s only two samples within bedrock.

[00:12:41.280]
So in this instance there’s no clear

[00:12:46.220]
if we have a look at (mumbles)

[00:12:49.820]
water table, we can see that both are saturated

[00:12:54.180]
and the vital sign has significant concentrations

[00:12:59.030]
and there’s not one particularly higher than the other.

[00:13:01.610]
Saturated is a little bit more,

[00:13:03.600]
but it’s not obviously completely different.

[00:13:08.900]
So the way we’ll consider this then

[00:13:10.870]
is I’m just going to separate a single domain

[00:13:13.940]
that’s above bedrock and not taking into account

[00:13:16.790]
whether it’s saturated or (mumbles).

[00:13:20.030]
In this way, I have a geological model then

[00:13:23.370]
that is just domained.

[00:13:37.140]
So, once we’ve decided on

[00:13:38.630]
how we’re going to domain our data,

[00:13:40.620]
then we start using the Contaminants Extension information

[00:13:43.990]
and we can link it to our geology model

[00:13:46.040]
right through to the resource estimate.

[00:13:48.060]
And we can do exploratory data analysis

[00:13:50.400]
specifically on the domain itself.

[00:13:53.635]
So we have a look at the domain model.

[00:13:57.190]
And in this case

[00:13:58.023]
this has now been all of the mythologies about bedrock

[00:14:02.320]
have been combined into a single domain

[00:14:03.660]
which is the yellow one on the scene.

[00:14:06.370]
And we can see that the majority of the samples

[00:14:08.147]
are within that and there’s just the one isolated to

[00:14:12.140]
below in the bedrock which we won’t be considering

[00:14:14.960]
in this particular estimate.

[00:14:24.890]
When we build up estimation,

[00:14:29.130]
we have what’s termed as contaminants model folder.

[00:14:32.230]
So when we activate the Contaminants Extension License,

[00:14:37.600]
it displays up as a folder in your project tree

[00:14:41.380]
and there’s a estimation sub folder

[00:14:43.347]
and a block model’s folder.

[00:14:45.490]
The estimation folder can be considered to be

[00:14:47.950]
where you set up all your parameters.

[00:14:49.930]
And then the block model is where we display

[00:14:52.500]
all of our results

[00:14:53.440]
and we also do all our reporting and validation

[00:14:56.140]
from that point as well.

[00:14:58.520]
So I’ve decided to do a title constraint traction

[00:15:02.410]
of this chloride within the unconsolidated domain

[00:15:05.870]
which is this yellow block here.

[00:15:08.010]
So when we create a domain estimator,

[00:15:12.920]
we just select any closed solid within our project

[00:15:17.040]
and you can see here that

[00:15:17.890]
this is the unconsolidated sediments

[00:15:21.430]
within our domain model.

[00:15:24.160]
And then the numeric values are coming directly

[00:15:26.940]
from our essay table, contaminant intervals,

[00:15:30.681]
total concentration.

[00:15:32.760]
So we can choose any numeric column

[00:15:34.860]
within any part of our database.

[00:15:36.860]
And also if we were storing point data

[00:15:39.520]
within the points folder over here,

[00:15:41.630]
we would be able to access that as well.

[00:15:43.980]
So it doesn’t have to be within drew holes .

[00:15:48.720]
I’m not applying any transform at this stage

[00:15:51.680]
and we also have the capability of doing compositing.

[00:15:55.340]
This data set doesn’t render itself to compositing

[00:15:58.790]
because it’s got isolated values within each holes

[00:16:02.640]
really only got one or two samples

[00:16:05.628]
compositing small for if we’ve got a

[00:16:08.080]
continuous range of samples down the borehole,

[00:16:12.020]
and we’d want to composite that into large intervals

[00:16:17.510]
to reduce their variance.

[00:16:20.170]
So within each estimation for each domain,

[00:16:22.730]
we have this sub folder.

[00:16:24.520]
And from there we can see the domain itself.

[00:16:28.280]
So we can double check that that is the volume of interest

[00:16:33.030]
and we can display the values within that domain.

[00:16:37.700]
So they have become point values at this stage

[00:16:42.800]
and you can label those

[00:16:44.820]
and you can also do statistics on these.

[00:16:47.610]
So these now become domain statistics.

[00:16:50.500]
So by right clicking on the values,

[00:16:54.280]
I get a histogram of those values there.

[00:16:58.910]
As with all Leapfrog Works graphs,

[00:17:04.010]
they are interacting with the 3D scene.

[00:17:07.130]
So let’s just move this over here.

[00:17:12.920]
If I wanted to find out where these core values were,

[00:17:17.590]
I can just highlight the one that on the graph

[00:17:20.610]
and that will show up in the same format.

[00:17:23.450]
Let’s do these ones about 100, there they are there.

[00:17:27.350]
So gives us then idea of whether or not

[00:17:30.160]
we need to create sub domains.

[00:17:32.400]
If all the high grades were focused a little bit more

[00:17:35.520]
we might be able to subdivide them out.

[00:17:37.210]
But in this case,

[00:17:38.630]
we’re just going to treat everything in the one domain.

[00:17:44.870]
The special models folder is where we can do variography

[00:17:49.990]
which is a spatial analysis of the data

[00:17:52.670]
to see whether there are any specific directions

[00:17:55.320]
where the concentration grades might be aligned.

[00:18:02.980]
I’ll open this one up.

[00:18:12.510]
So the way the variography works

[00:18:14.320]
is that we select the direction

[00:18:16.700]
and then we model it in three directions.

[00:18:21.370]
One with the greatest degree of continuity et cetera.

[00:18:25.160]
This radio plot is a plane within this.

[00:18:29.390]
We’re looking in a fairly horizontal plane here

[00:18:32.667]
and you can see that we have…

[00:18:35.600]
we can always see the ellipsoid that we’re working with

[00:18:38.900]
so to see whether it’s in the like

[00:18:41.160]
we wanted to make that a little bit flatter.

[00:18:43.250]
We didn’t leave it in a 3D scene

[00:18:45.660]
and it will update on the form here

[00:18:48.210]
and the graphs will update.

[00:18:51.260]
We can use this radial plot.

[00:18:54.520]
This may suggest that this orientation

[00:18:57.210]
has a little bit more continuity and we’ll re-adjust.

[00:19:02.520]
This particular data set doesn’t create

[00:19:05.680]
a very good experimental variogram

[00:19:08.130]
so we do have some other options

[00:19:09.710]
where we could use a log transform prior to that

[00:19:14.293]
so here I’ve done a log transform on the values

[00:19:19.450]
and that was simply done by at the very beginning.

[00:19:22.550]
When I selected my samples

[00:19:24.256]
I could apply a log transform in through here.

[00:19:28.630]
This gets rid of some of the variability

[00:19:30.650]
especially the range and extreme differences

[00:19:34.350]
between very high values and very low values

[00:19:37.260]
and this can help develop a better semi-variably model

[00:19:42.300]
if we have a little bit this one.

[00:19:50.624]
You can see here that the semivariogram

[00:19:53.240]
is a little bit better structured than it was before.

[00:19:56.700]
Again this data is quite sparse.

[00:19:58.760]
So this is about the best we could do

[00:20:01.070]
with this particular data set.

[00:20:04.090]
We have the ability to pick a couple of range of structures,

[00:20:08.970]
we have a nugget plus a potentially two structures

[00:20:12.370]
but in this case I’ll just pick a nugget,

[00:20:15.360]
a fairly large nugget and a single structure

[00:20:18.030]
which is sufficient to model this.

[00:20:21.190]
We can get an idea of that model.

[00:20:23.350]
So this is in the…

[00:20:25.090]
The range is quite large, 1200 to 1400 meters.

[00:20:29.920]
And that’s because the data itself is quite widely spaced.

[00:20:33.620]
And we’ve only got a very short range

[00:20:35.390]
in the Z direction because we’re looking at a

[00:20:37.294]
fairly constrained tabular body of potential contamination.

[00:20:44.340]
The advantage to building the variography

[00:20:46.390]
within the contaminants extension is that

[00:20:49.070]
everything is self-contained,

[00:20:50.410]
and once we hit the side button,

[00:20:52.830]
then that variogram is all ready to use

[00:20:56.510]
within the estimations itself.

[00:21:00.480]
We also have a concept of variable orientation.

[00:21:04.750]
Sometimes the design of the Contaminant

[00:21:10.520]
isn’t purely tabular

[00:21:12.480]
and there may be local irregularities

[00:21:14.810]
within the surfaces.

[00:21:16.680]
So we can use those surfaces

[00:21:18.450]
to help guide the local orientation of the variogram

[00:21:22.750]
when we’re doing sample selection.

[00:21:25.470]
So just don’t want it here already.

[00:21:31.890]
So the way we do that

[00:21:32.810]
is that we set up a reference surface

[00:21:35.440]
and in this case it’s the actual base

[00:21:37.570]
over the top of the bedrock

[00:21:39.530]
and we can see here that we’ve colored this by a dip.

[00:21:44.480]
So we can see that we’ve got some varying dips

[00:21:47.670]
that flat up in the middle

[00:21:49.070]
and slightly more dips around the edges.

[00:21:51.830]
If I can hop and look cross sections through there,

[00:22:03.450]
we can see that what will happen

[00:22:06.610]
is when we pick out samples,

[00:22:09.060]
the ellipsoid will be orientated as local

[00:22:12.780]
and so overall ellipsoid is in this general direction,

[00:22:16.010]
but where it gets flat out,

[00:22:17.400]
it will rotate in a flat orientation.

[00:22:21.570]
The actual ratios and the variogram parameters themselves

[00:22:24.960]
are still maintained.

[00:22:26.267]
It’s just the orientation

[00:22:28.260]
for when the samples are being selected.

[00:22:35.270]
When we’re doing an estimation,

[00:22:36.730]
we have to consider a few different things.

[00:22:39.330]
So we need to understand what the stage our project is

[00:22:42.650]
and this might help us work out

[00:22:44.520]
which estimation method to use.

[00:22:46.550]
If we’ve got early stage,

[00:22:48.190]
we might not have very much data.

[00:22:49.910]
We might not have sufficient data to get a variogram.

[00:22:52.740]
So we will need to use either nearest neighbor

[00:22:55.510]
or inverse distance type methods.

[00:22:59.476]
There can be a lot of work to go into the search strategy.

[00:23:02.280]
And that is how many samples are we going to use

[00:23:05.061]
to define a local estimate.

[00:23:08.850]
And then once we run our estimate,

[00:23:10.900]
we need to validate it against the original data

[00:23:14.250]
to make sure that we’re maintaining the same kind of means

[00:23:17.950]
and then we want to report our results at the end.

[00:23:23.150]
To set up the estimators,

[00:23:25.970]
we work through this estimators sub folder here

[00:23:29.760]
and if you right click

[00:23:30.920]
you can go have inverse distance nearest neighbor kriging

[00:23:34.857]
and we can do ordinary kriging, simple kriging

[00:23:38.010]
or we can use the actual RBF estimator as well.

[00:23:41.520]
Which can be quite good

[00:23:42.870]
for when we haven’t got sufficient data to a variogram.

[00:23:47.110]
It’s a global estimator.

[00:23:49.540]
Using the same underlying algorithm

[00:23:51.470]
that is used for when we krig the surfaces

[00:23:54.170]
or the numeric models that can be run

[00:23:56.600]
within Leapfrog Works itself.

[00:24:00.710]
To set one up

[00:24:01.990]
once I open we can edit that,

[00:24:04.650]
we can apply clipping which is

[00:24:08.390]
if we want to constrain the very high values

[00:24:11.010]
and give them less influence,

[00:24:12.550]
we can apply that on floor.

[00:24:15.090]
In this case because I’ve actually looking at contaminants

[00:24:17.380]
I don’t really want to cut them down initially

[00:24:19.580]
so I will leave them unclicked.

[00:24:22.540]
I’m just doing point kriging in this case

[00:24:24.710]
so just I’m not using discretization at all.

[00:24:28.590]
In some cases if I want to block kriging,

[00:24:30.460]
then I would increase these discretization amounts.

[00:24:35.170]
I’m using ordinary kriging.

[00:24:36.693]
This is where we determine ordinary or simple.

[00:24:41.120]
And this is the variogram.

[00:24:42.750]
So you can pick.

[00:24:44.180]
It will see the variogram within the folder.

[00:24:50.460]
You set to a particular variogram

[00:24:53.090]
and that’s more for making sure

[00:24:55.730]
that it’s set to the correct orientation.

[00:24:58.240]
And then it’ll automatically go to the maximum range

[00:25:01.100]
of that variogram but you can modify these

[00:25:03.680]
to be less than that or greater.

[00:25:05.790]
Depends on the circumstances

[00:25:08.750]
and often you’ll do a lot of testing

[00:25:11.230]
and sensitivity analysis to determine

[00:25:13.760]
the best scope for this one.

[00:25:18.400]
And then we can choose

[00:25:19.890]
minimum and maximum number of samples.

[00:25:22.610]
And this effectively becomes

[00:25:24.280]
our grade variable within the block model

[00:25:27.710]
and we can store other variables as well.

[00:25:30.510]
So if we’re using kriging,

[00:25:32.030]
we can store things such as kriging variance,

[00:25:34.240]
slope and regression

[00:25:35.820]
which would go as to how good the estimate might be.

[00:25:39.430]
We can store things such as average distance to samples

[00:25:42.290]
and these may help us with classification

[00:25:44.810]
criteria down the track.

[00:25:47.150]
So I’ve created us an ordinary kriged estimator here

[00:25:52.640]
and then I’ve also got one.

[00:25:54.270]
This one is exactly the same,

[00:25:57.340]
but in this case I’ve actually applied

[00:25:59.610]
the variable orientation.

[00:26:01.230]
So we’ve set up a variable orientation within the folder.

[00:26:04.190]
It will be accessible in our parameter set up through here

[00:26:09.050]
and I’ll give it a slightly different name.

[00:26:11.550]
This one I will need to compare the kriging results

[00:26:17.310]
with and without variable orientation apply.

[00:26:21.570]
So we set up all of our different parameters

[00:26:24.250]
in this estimation folder.

[00:26:27.360]
And we are then able to create a block model

[00:26:32.690]
to look at the results.

[00:26:40.300]
Setting up a block model is quite simple

[00:26:43.440]
so we can create a new block model

[00:26:45.900]
and we can also import block models from other software.

[00:26:49.790]
And when we’ve created one,

[00:26:51.990]
we can edit it by opening it up.

[00:26:55.500]
And in this case,

[00:26:56.333]
we’ve got a block size of 80 by 80 meters

[00:26:58.720]
in the X, Y direction and five in the Z

[00:27:02.260]
so we’re looking at that fairly thin.

[00:27:04.270]
And if we’re looking at it again very visual

[00:27:07.167]
and we can see our block size.

[00:27:13.230]
So this is 80 by 80 in the X Y direction here.

[00:27:18.400]
And we can change these boundaries at any time.

[00:27:20.810]
So once it’s been made,

[00:27:21.900]
if we wanted to adjust it,

[00:27:23.533]
I could just move the boundary up

[00:27:25.720]
and it will automatically update the model.

[00:27:28.930]
I want to do that now so it doesn’t rerun.

[00:27:31.820]
A key concept when working with block models

[00:27:34.030]
within the Contaminants Extension is this evaluations tab.

[00:27:38.260]
So basically whatever we want to flag into the block model

[00:27:41.794]
needs to come across from the left-hand side

[00:27:44.590]
across to the right.

[00:27:46.040]
So what I’m doing here

[00:27:47.090]
is I’m flagging each block centroid with the domain model

[00:27:50.980]
so that will flag it with whether it’s within that domain

[00:27:56.590]
but we can also flag at geology.

[00:27:58.410]
So this will put into whether it’s two

[00:28:00.770]
or whether it’s bedrock

[00:28:02.610]
and also avoid us saturated sides

[00:28:05.110]
could be flagged as well.

[00:28:07.500]
And then this is Y grade or contaminant variables here.

[00:28:12.870]
So I’m just putting in all four that are generated

[00:28:17.200]
so I’ve got an inverse distance,

[00:28:19.360]
I’ve got two different krigings.

[00:28:21.734]
And also the nearest neighbor.

[00:28:23.460]
Nearest neighbor can be useful

[00:28:25.140]
for helping with validation of the model.

[00:28:29.480]
Again, any of these can be removed or added at one time.

[00:28:32.680]
So if I did want to look at the RBF,

[00:28:35.070]
I would just bring it across

[00:28:37.170]
and then the model would rerun.

[00:28:41.200]
So we can see here, the model has been generated.

[00:28:45.710]
And one of the key things for validation of first

[00:28:49.120]
is to actually look at the samples against the results.

[00:28:54.550]
So if I just go into a cross section

[00:29:06.010]
making sure that our color schemes

[00:29:08.760]
forbides the data and the bot model are the same

[00:29:11.940]
and we can import these colors

[00:29:13.580]
so if we’ve got colors defined for our essays up here,

[00:29:23.952]
we can export those colors

[00:29:25.640]
like create a little file,

[00:29:27.360]
and then we can go down to our block model

[00:29:29.660]
and we can import those colors down here.

[00:29:34.420]
So I did on this one, look at that colors import

[00:29:38.370]
and then pick up that file.

[00:29:39.670]
And that way we get exactly same color scheme.

[00:29:42.640]
And what we’re hoping to see

[00:29:43.910]
is a little we’ve got yellows

[00:29:46.610]
we should hopefully have orange or yellows around,

[00:29:49.710]
higher grades will be in the reds and purples

[00:29:53.610]
and lower grades should show up.

[00:30:02.308]
Of course we have a lot of data in here.

[00:30:06.074]
There we go.

[00:30:06.907]
So we can see some low grade points over here

[00:30:09.280]
and that’s equating.

[00:30:11.170]
We’ve got really good tools for looking at the details

[00:30:14.870]
of the results on a block-by-block basis.

[00:30:17.720]
If I highlighted this particular block here,

[00:30:21.250]
we have this interrogate estimator function.

[00:30:26.130]
And this is the estimator that it’s looking at.

[00:30:29.080]
So like ordinary krig without variable orientation

[00:30:32.560]
it’s found there’s many samples

[00:30:35.390]
and depending on how we’ve set up our search,

[00:30:38.120]
in this case it’s included all those samples.

[00:30:41.050]
It’s showing us the kriging lights

[00:30:42.360]
that have been applied and we can sort these

[00:30:45.177]
and distances from the sample.

[00:30:48.660]
So we can see the closest sample used was 175 meters away

[00:30:53.080]
and the greatest one was over 1,000 meters away.

[00:30:56.200]
But it’s got lights through there.

[00:31:00.470]
Part of the kriging process

[00:31:01.730]
can give negative kriging weights

[00:31:03.670]
and that is not a problem

[00:31:04.850]
unless we start to see negative grade values.

[00:31:07.770]
In that case, you might need to change

[00:31:09.520]
our sample selection a little bit

[00:31:12.660]
to eradicate that influence.

[00:31:16.420]
That’s just part of the kriging process.

[00:31:21.130]
You can see also that when we do that we can…

[00:31:23.740]
it shows the ellipsoid for that area.

[00:31:26.720]
And if we filter the block model

[00:31:30.818]
and I’ll just take away the ellipsoid for a second,

[00:31:36.180]
it shows you the block and it shows you the samples

[00:31:39.113]
that are being used.

[00:31:40.540]
So you’ve got a nice spatial visualization

[00:31:44.000]
of what samples are actually being used

[00:31:46.080]
to estimate the grade within that box through there.

[00:31:48.947]
And if you want to try something else,

[00:31:50.760]
you can go directly to the estimator and modify.

[00:31:55.360]
If I want to change this to three samples,

[00:31:58.070]
reduce these chains to search

[00:32:00.600]
and you can get see what the impacts could be. Okay.

[00:32:06.790]
So apart from visualization

[00:32:08.910]
and looking at on a block-by-block basis,

[00:32:12.620]
we can also run swath plots.

[00:32:18.780]
So within the block model itself,

[00:32:20.870]
we can create a new swath plot.

[00:32:23.210]
And the beauty of this is that

[00:32:25.200]
they are embedded within the block model

[00:32:27.090]
so that when we update that data later on,

[00:32:29.750]
the swath plots will update as well

[00:32:32.270]
and the same goes for reports.

[00:32:34.060]
If we build a report on the model

[00:32:36.170]
then if we change the data,

[00:32:37.650]
then the reports will have tables.

[00:32:40.409]
So we’ll have a look at a swath

[00:32:42.610]
that’s already been prepared

[00:32:44.030]
and you can store any number of swaths within the model.

[00:32:49.330]
So what this one is doing

[00:32:50.530]
is it’s looking at the inverse distance.

[00:32:53.470]
It’s got the nearest neighbor which is this gray

[00:32:56.170]
and then we’ve got two different types of kriging.

[00:32:59.950]
So we can see straight away.

[00:33:01.320]
And it’s showing us in the X, Y and Z directions

[00:33:05.810]
relative to the axis of the block model.

[00:33:08.830]
So if we look in the X direction here,

[00:33:10.640]
we can see that all three estimators

[00:33:13.647]
are fairly close to each other

[00:33:15.780]
and what the red line is is the actual original sample data.

[00:33:20.450]
So these are the sample values

[00:33:22.250]
and what we generally do want to see

[00:33:24.920]
is that the estimators tend to bisect

[00:33:27.110]
whether we have these highs and lows,

[00:33:29.280]
the estimated values we should run through the middle.

[00:33:32.600]
Also the estimator should follow the trend

[00:33:35.570]
across the data.

[00:33:37.600]
The nearest neighbor can be used as a validation tool.

[00:33:42.090]
So if one of these estimators

[00:33:43.880]
for instance was sitting quite high

[00:33:46.330]
up here above their sniper line

[00:33:48.810]
they would suggest that potentially where I over estimated

[00:33:50.850]
and there might be some problems with our parameters

[00:33:54.040]
so we would want to change that.

[00:33:56.810]
Again we can export, we can copy this data

[00:34:00.537]
and put it into a spreadsheet.

[00:34:02.160]
Sometimes people have very specific formats

[00:34:06.650]
they like to do their swath plots in.

[00:34:08.200]
So that will bring all the data out into Excel.

[00:34:10.690]
So you can do that yourself

[00:34:12.640]
or we can just take the image

[00:34:14.830]
and paste it directly into our report.

[00:34:21.780]
Okay. If we want to report on our data,

[00:34:24.950]
we create the report in the block model.

[00:34:28.400]
And in this case, I’ve got just the basic report.

[00:34:33.710]
That’s looking at the geology and the water table

[00:34:38.020]
and it’s giving us the totals and the concentration.

[00:34:41.560]
And this is by having to select these categorical columns

[00:34:45.400]
and then we have our grade or value columns.

[00:34:48.030]
So we need at least one categorical column

[00:34:50.820]
and one to recreate a report

[00:34:53.310]
but you can have as many different categories

[00:34:56.150]
as you like within there.

[00:34:58.160]
And these are also act like a pivot table almost

[00:35:01.030]
so if I want it to show geology first

[00:35:04.370]
and then want a table,

[00:35:05.760]
I can split it up through there like that.

[00:35:08.220]
We can apply a cutoff or we can have none as well.

[00:35:12.740]
Another advantage within the block models (mumbles)

[00:35:20.090]
is that we can create calculations and filters

[00:35:26.150]
and we can get quite complex with these.

[00:35:30.040]
We have a whole series of syntax helpers

[00:35:34.060]
which are over here on the right.

[00:35:36.590]
And then in middle here,

[00:35:37.750]
it shows us all of our parameters

[00:35:39.620]
that are available within the block model

[00:35:41.720]
to use within a calculation.

[00:35:45.630]
It’s also a nice little summary of the values

[00:35:48.800]
within each estimator.

[00:35:52.030]
Now you saw here for instance,

[00:35:53.900]
I can see that I’ve got a couple of negatives

[00:35:56.797]
in this ordinary krig without the variable orientation.

[00:36:00.270]
So I’d want to have a look at those

[00:36:01.760]
to see where they are and whether I need to remove them

[00:36:04.920]
or do something about them.

[00:36:07.660]
If I just unpin that for a second

[00:36:10.600]
we can see, we can get quite complex

[00:36:13.020]
we’ve got lots of if statements we can embed them.

[00:36:16.500]
So what’s happening with this one

[00:36:18.060]
is I’m trying to define porosity

[00:36:20.380]
based on the rock type within the model.

[00:36:24.240]
So I’ve defined, if we go back to here,

[00:36:28.740]
I can look at the geology.

[00:36:30.820]
So the geology is being coded within to the model.

[00:36:34.300]
And if I stick on that block you can see here

[00:36:36.470]
the geology model is till and the water table is saturated.

[00:36:41.060]
So I can use any of those components (mumbles)

[00:36:48.723]
so stop saying, “If the domain model is still,

[00:36:51.740]
I’m going to assign a plus to the 8.25.”

[00:36:54.320]
And that’s a fairly simple calculation

[00:36:56.300]
but you can put in as many complex calculations as you wish.

[00:37:00.300]
This one title concentration is saying that

[00:37:02.720]
it must be within unconsolidated sediments.

[00:37:06.420]
And what this one is doing is saying,

[00:37:07.960]
what happens if I’ve got some blocks

[00:37:09.580]
that haven’t been given a value.

[00:37:12.140]
So I can say that if there is no value,

[00:37:15.059]
I’m going to assign a very low one

[00:37:17.570]
but otherwise use the value if it’s greater than zero,

[00:37:21.270]
use the value that’s there.

[00:37:22.510]
So that’s one way of getting rid of negatives

[00:37:24.440]
if you know that there’s only a

[00:37:26.330]
couple of isolated ones there.

[00:37:30.680]
And then this one is an average distance.

[00:37:32.530]
So I store the average distance within the model.

[00:37:35.980]
So I’m using this as a guide for classification.

[00:37:39.260]
So stop saying if the average distance

[00:37:40.940]
to the samples is less than 500 meters,

[00:37:43.500]
I’ll make it measured less than 1,000 indicated

[00:37:46.621]
less than 1500.

[00:37:48.970]
The beauty of any calculation, numeric or categorical

[00:37:52.630]
is that it’s available instantly to visualize.

[00:37:55.810]
So we can visualize this confidence category.

[00:37:59.410]
So you can see that there,

[00:38:00.610]
I’ve just got indicated in third

[00:38:03.380]
and a little bit of measured there as well.

[00:38:06.800]
Let’s split that out. (mumbles)

[00:38:10.980]
So we can see here,

[00:38:12.360]
that’s based on the data and the data spacing.

[00:38:17.050]
We might not use those values directly to classify

[00:38:19.860]
but we can use them as a visual guide

[00:38:22.240]
if we wanted to draw shapes to classify certain areas.

[00:38:28.580]
So once we’ve done that kind of thing

[00:38:31.100]
we can build another report based of that calculation.

[00:38:36.460]
So in this case we’re using that calculated field

[00:38:40.060]
as part of that categorical information.

[00:38:43.920]
And so now we’ve got classification,

[00:38:46.180]
the geology and the water table in the report.

[00:38:49.630]
And again that’s saved automatically and it can be exported.

[00:38:53.890]
We can copy it into an Excel Spreadsheet

[00:38:57.790]
or you can just export it as an Excel Spreadsheet directly.

[00:39:05.930]
Okay. So that’s the main usage

[00:39:10.240]
of the Contaminants Extension.

[00:39:12.160]
And I’ll just pass right back again to Aaron to finish it.

[00:39:16.290]
<v Aaron>Fantastic. Thanks Steve.</v>

[00:39:20.160]
We have time now for questions as well.

[00:39:23.670]
So if you have them already,

[00:39:25.360]
please feel free to put a question in the question box

[00:39:28.330]
or you can raise your hand as well

[00:39:30.320]
and we can unmute you

[00:39:31.670]
and you can verbally ask the question as well.

[00:39:35.650]
While we give people a chance to do that,

[00:39:39.289]
beyond today’s meeting there’s always support for Leapfrog

[00:39:43.360]
and the Contamination Extension

[00:39:44.880]
by contacting us at [email protected]

[00:39:48.910]
or by giving our support line a phone call.

[00:39:52.480]
After today’s session as well,

[00:39:54.852]
we’d love to give the extension of Leapfrog a go.

[00:39:59.620]
We have a trial available

[00:40:00.790]
that you can sign up for on our website

[00:40:02.480]
on the Leapfrog Works products page

[00:40:04.450]
which you can see the URL therefore.

[00:40:06.690]
We’ll also be sending an email out after this webinar

[00:40:10.710]
with the recording for today

[00:40:12.410]
and also a link for a four-part technical series

[00:40:16.180]
on the Contamination Extension

[00:40:17.840]
which will take you through step-by-step

[00:40:20.330]
how to build the model that Steve has shown you

[00:40:23.190]
from scratch and the process behind that.

[00:40:27.440]
And includes the data there as well.

[00:40:28.830]
So that’s a fantastic resource we’ll be sending to you.

[00:40:31.860]
We also have online learning content

[00:40:34.490]
at my.seequent.com/learning.

[00:40:37.500]
And we also have the

[00:40:38.830]
Leapfrog Works and Contaminants Extension help pages

[00:40:41.910]
will look out at them later.

[00:40:43.400]
And that’s a really useful resource

[00:40:44.850]
if you have any questions.

[00:40:51.770]
So yeah like I said, if you have questions,

[00:40:54.510]
please put them up or raise your hand.

[00:40:57.170]
One of the questions that we have so far is

[00:41:01.392]
so often contaminants sources

[00:41:03.840]
are at the surface propagating down

[00:41:05.910]
into the soil or groundwater

[00:41:09.872]
and how would you go about modeling a change

[00:41:12.800]
in the directional trend

[00:41:14.240]
I guess of the plume going from vertical

[00:41:16.620]
as it propagates down to spreading out horizontally.

[00:41:20.075]
I guess once it’s an aquifer or an aquatar.

[00:41:24.710]
So can you answer that for us please Steve?

[00:41:27.230]
<v Steve>Yep so everything about the Contaminants Extension</v>

[00:41:30.750]
is all about how we set up our domains

[00:41:33.360]
at the very beginning.

[00:41:34.770]
So it does need a physical three-dimensional representation

[00:41:39.056]
of each domain.

[00:41:40.690]
So the way we would do that would

[00:41:42.310]
we would need acquifer surface and the soil.

[00:41:46.414]
We would have one domain and say that

[00:41:49.390]
then you’d model your aquifer as a horizontal layer.

[00:41:54.410]
So you would set up two estimation folders,

[00:41:57.120]
one for the soil, one for the aquifer.

[00:41:59.410]
And then within the soil one,

[00:42:00.630]
you could make your orientation vertical

[00:42:03.420]
and within the aquifer you can change it to horizontal.

[00:42:06.850]
And then you get…

[00:42:07.770]
You’re able to join these things together

[00:42:10.120]
using the calculation so that you can visualize

[00:42:13.240]
both domains together in the block model.

[00:42:16.200]
So you only need one block model

[00:42:17.920]
but you can have multiple domains

[00:42:19.890]
which will get twined together.

[00:42:21.950]
<v Aaron>Fantastic. Thank you Steve.</v>

[00:42:24.770]
Another question here is,

[00:42:26.600]
were any of those calculations

[00:42:28.180]
automatically populated by Leapfrog

[00:42:30.270]
or did you make each of them?

[00:42:32.040]
If you made them, just to be clear,

[00:42:33.530]
did you make them to remove the negative values

[00:42:35.960]
generated by the kriging?

[00:42:38.045]
<v Steve>No, they’re not generated automatically.</v>

[00:42:40.530]
So I created all of those calculations

[00:42:43.450]
and you don’t have to force removal of the negatives.

[00:42:48.510]
It really comes down to first

[00:42:50.860]
realizing that negative values exist.

[00:42:53.420]
And then we would look at them in the 3D scene

[00:42:55.450]
to see where they are.

[00:42:56.820]
Now, if they’re only the isolated block around the edges

[00:43:00.260]
quite a bit away from your data,

[00:43:03.010]
then they might not be significant.

[00:43:05.060]
So then you could use the calculation to zero them out.

[00:43:09.690]
Sometimes you get a situation

[00:43:11.260]
where you can get negative grades

[00:43:13.410]
right within the main part of your data.

[00:43:16.387]
And this is the thing called a kriging et cetera.

[00:43:21.763]
and the kriging process can cause this.

[00:43:24.550]
So in that case you would have to

[00:43:26.100]
change your search selection.

[00:43:28.510]
So you might have to decrease the maximum number of samples.

[00:43:31.750]
You might need to tweak the variogram a little bit

[00:43:34.340]
and then you should be able to get

[00:43:35.830]
reduce the impact of negative weights

[00:43:38.219]
So that you won’t get negative grades.

[00:43:41.450]
<v Aaron>To add to that first part of the question as well.</v>

[00:43:43.942]
Once you have created your calculations,

[00:43:47.230]
you can actually go in and export them out

[00:43:51.760]
to use in other projects as well.

[00:43:53.410]
So kind of… Where are we?

[00:43:58.490]
Calculations are cool.

[00:43:59.860]
So calculations and filters.

[00:44:01.490]
So in this calculations tab here,

[00:44:03.780]
you can see we have this option here

[00:44:05.490]
for importing or indeed for exporting here as well.

[00:44:08.840]
So once you set it at once,

[00:44:10.440]
you can reuse that in other projects.

[00:44:12.620]
So it’s very easy to transfer these across projects

[00:44:16.260]
maybe change some of the variable names.

[00:44:18.260]
If you need to, it’s different under different projects.

[00:44:20.967]
And it’s once you’ve set at once

[00:44:22.410]
it’s very easy to use those again

[00:44:24.167]
and that’s a files, it’s a calculation file

[00:44:26.380]
that you can send to other people in your company as well.

[00:44:28.390]
<v Steve>So if you don’t have the same variable names,</v>

[00:44:31.830]
what it does,

[00:44:32.663]
it just highlights them with a little red,

[00:44:34.430]
underline and then you just replace it

[00:44:36.880]
with the correct variable name.

[00:44:38.340]
So yeah easy to set things up

[00:44:41.890]
especially if you’re using the same format

[00:44:43.545]
of block model over a whole project area

[00:44:46.400]
for different block models.

[00:44:49.020]
<v Aaron>Fantastic. And yeah as a reminder,</v>

[00:44:51.230]
if you do have questions,

[00:44:52.150]
please put them in the chat or raise your hand.

[00:44:55.140]
We have another question here.

[00:44:56.220]
Can you run the steps to include the 95 percentile?

[00:45:01.562]
<v Steve>Not directly,</v>

[00:45:04.331]
actually we can get…

[00:45:07.090]
We do do log probability plots.

[00:45:09.600]
So (mumbles).

[00:45:13.350]
So if I go back up to this one here yet

[00:45:24.680]
so my daily statistics of strike first up

[00:45:27.940]
get a histogram,

[00:45:30.169]
but I can change this to a low probability plot.

[00:45:32.710]
So then I’ll be able to see where the 95th percentile is.

[00:45:36.890]
Is down there.

[00:45:38.570]
So the 95th percentile is up through there like that.

[00:45:42.622]
<v Aaron>Awesome. Thanks Steve.</v>

[00:45:45.850]
So can you please show

[00:45:48.310]
the contamination plume advance over time?

[00:45:50.790]
And can you show the database with the time element column?

[00:45:54.990]
So we don’t have any time or…

[00:45:58.534]
<v Steve>Sorry.</v>

[00:46:00.184]
<v Aaron>monitoring data set up for this project</v>

[00:46:03.321]
but you could certainly do that.

[00:46:05.140]
So you could just set up your first one,

[00:46:07.555]
make a copy of it and then change the data to the new,

[00:46:10.730]
but the next monitoring event.

[00:46:11.920]
So if each monitoring event,

[00:46:13.654]
you’re just using that separate data

[00:46:16.400]
in the same estimations, in the same block models

[00:46:20.900]
and also in with the contaminants kit.

[00:46:23.250]
If you go into the points folder,

[00:46:25.335]
you do have this option for

[00:46:26.900]
import time dependent points as well.

[00:46:28.830]
So you can import a point cloud

[00:46:30.070]
that has the time data or attribute to it.

[00:46:33.370]
You can actually filter that.

[00:46:34.390]
So visually you can use a filter

[00:46:36.040]
to see the data at a certain point in time

[00:46:38.340]
but you could then actually

[00:46:40.050]
if you wanted to, you could set up a workflow

[00:46:41.960]
where you set filters up

[00:46:43.550]
to use that point data in your block model

[00:46:46.630]
and just filter it based on time as well.

[00:46:49.650]
<v Steve>So when you create a estimator,</v>

[00:46:53.550]
so this one here,

[00:46:59.390]
you can see here you can apply a filter to your data.

[00:47:02.420]
So yes, if you divided up

[00:47:05.030]
created pre filters based on the time,

[00:47:07.660]
you can run a series of block models.

[00:47:13.090]
Well actually you could probably do it

[00:47:14.700]
within the single block model.

[00:47:16.090]
So you would just set up

[00:47:17.440]
a different estimator for each time period,

[00:47:20.020]
estimate them all and put them all,

[00:47:21.810]
evaluate them all into the same block model

[00:47:24.260]
and then you could kind of look at them and how it changes.

[00:47:28.070]
<v Aaron>Yeah. Awesome.</v>

[00:47:29.640]
<v Steve>That’d be a few different ways</v>

[00:47:30.790]
of trying to do that.

[00:47:35.420]
<v Aaron>We have a question here</v>

[00:47:36.770]
which is, when should you use the different

[00:47:41.750]
or when should you use the different estimation methods

[00:47:45.190]
or I guess?

[00:47:46.478]
<v Steve>Guess so. It really comes down</v>

[00:47:48.860]
to personal choice a little bit

[00:47:51.810]
but generally, we do need a certain degree

[00:47:56.270]
of the amount of data to do kriging

[00:47:58.590]
so that we can get a variable variogram

[00:48:01.420]
So early stages,

[00:48:02.670]
we might only be able to do inverse distance

[00:48:06.510]
whether it’s inverse distance squared or cubed et cetera.

[00:48:10.300]
It’s again personal preference.

[00:48:13.517]
Generally though, if we wanted to get

[00:48:15.280]
quite a localized estimate,

[00:48:17.780]
we may use inverse distance cubed,

[00:48:19.860]
a more general one the inverse distance squared.

[00:48:23.580]
And now if we can start to get a variogram out

[00:48:25.960]
it’s always best to use kriging if possible

[00:48:28.777]
it does tend.

[00:48:29.970]
You find that, especially with sparse data,

[00:48:32.760]
you won’t get too much,

[00:48:34.270]
the results shouldn’t be too different from each other

[00:48:36.930]
but kriging does tend to work with it better.

[00:48:41.150]
The RBF, I don’t use it much myself

[00:48:44.173]
but it can be quite useful to get early stage global idea

[00:48:49.590]
of how things are going.

[00:48:50.790]
<v Aaron>Yeah. So when you have sparse</v>

[00:48:51.847]
so not much data can be useful in those initial stages

[00:48:55.120]
maybe as pilot or desk study phase or something.

[00:48:57.930]
<v Steve>Yeah. Yeah.</v>

[00:48:58.871]
<v Aaron>For some case (mumbles).</v>

[00:49:02.110]
Another question here,

[00:49:03.020]
can you please explain the hard and soft boundaries

[00:49:05.450]
that appeared in the box just before

[00:49:08.280]
so in that initial estimation and this one here.

[00:49:13.550]
<v Steve>So what this is,</v>

[00:49:15.190]
is that in this case it’s not

[00:49:17.408]
or that sort of relevant.

[00:49:18.990]
There are only two samples outside our domain.

[00:49:21.930]
So if I bring back up my domain boundary,

[00:49:25.900]
let’s do that, okay.

[00:49:31.090]
So hard and soft basically

[00:49:32.982]
is what happens around the boundary of the domain itself.

[00:49:37.950]
How do I treat the samples?

[00:49:40.110]
So a hard boundary in 99% of the time,

[00:49:43.720]
you will use a hard boundary.

[00:49:45.430]
Is that you saying,

[00:49:47.180]
basically I’m only looking at samples

[00:49:48.930]
within the domain itself,

[00:49:50.300]
and I don’t care about anything outside it.

[00:49:53.630]
The instance where you may use a soft boundary

[00:49:56.030]
is if there’s a good directional change,

[00:49:59.660]
say the concentration can’t be limited directly

[00:50:02.630]
to one of the main or another.

[00:50:03.463]
And there’s a good directional change across that boundary.

[00:50:07.200]
By using a soft boundary, we can define,

[00:50:11.222]
we could say 20 meters.

[00:50:13.600]
And what that would do is we’ll look at for samples,

[00:50:16.660]
20 meters beyond the boundary in all directions.

[00:50:19.650]
And we’ll use those as part of the sample selection.

[00:50:23.350]
You still only estimate within the domain,

[00:50:27.420]
but around the edges you use samples

[00:50:29.860]
that are just slightly outside.

[00:50:31.730]
So that’s the main difference between hard and soft.

[00:50:35.090]
<v Aaron>And I guess going back</v>

[00:50:36.180]
to a question and an answer

[00:50:37.330]
that was before around using multiple domains

[00:50:40.260]
to kind of handle the vertical and then horizontal

[00:50:42.900]
would it be fair to say

[00:50:43.733]
that maybe you’d use the soft boundary

[00:50:45.700]
in that first one to accommodate the values near

[00:50:50.130]
really that domain boundary into the horizontal?

[00:50:52.290]
<v Steve>That’s right.</v>

[00:50:53.250]
And I mean, one of the beauties of a Leapfrog Works,

[00:50:57.960]
the Contaminants et cetera,

[00:50:59.210]
is that it’s so easy to test things.

[00:51:02.020]
So once you’ve got one of these folders set up

[00:51:06.060]
if you want to try something a little bit different,

[00:51:08.420]
you just copy the entire folder

[00:51:11.470]
and you might pick a slightly different domain

[00:51:14.630]
or if you’re going to do it on a different element,

[00:51:16.910]
choose it there.

[00:51:18.310]
And let’s just copy. That’s fine.

[00:51:20.360]
And it pre copies basically any very variables

[00:51:22.923]
that you might have done or you estimated is a set up.

[00:51:25.720]
And then you can just go back into that one and say,

[00:51:28.940]
I want to try something that’s little bit different

[00:51:31.237]
with this kriging set up.

[00:51:33.150]
So you might try a different distance,

[00:51:35.710]
some different samples et cetera.

[00:51:38.050]
So then you would just run them,

[00:51:39.645]
evaluate that into your bottle

[00:51:41.740]
and look at it on a swath plot

[00:51:43.760]
to see what the difference was.

[00:51:45.820]
<v Aaron>Awesome. Thanks Steve.</v>

[00:51:48.370]
I think you touched on maybe this question a little bit,

[00:51:51.040]
but is there an output or evaluation or really…

[00:51:56.150]
How can you show uncertainty of the block model?

[00:51:58.760]
<v Steve>Okay. So there are lots of different ways</v>

[00:52:01.650]
of trying to look at uncertainty.

[00:52:03.743]
The simplest way is just looking at data spacing.

[00:52:07.469]
And that’s a little bit what we’ve tried to do

[00:52:10.240]
with the calculation field before.

[00:52:13.860]
So I open up my calculations again.

[00:52:16.550]
So we stored the average distance to the samples

[00:52:20.430]
and that was done.

[00:52:23.690]
I’m here in the outputs

[00:52:25.460]
so I stored the average distance sample.

[00:52:27.530]
You can also store the minimum distance

[00:52:29.720]
to the closest sample.

[00:52:31.600]
And then again by making up

[00:52:33.960]
some kind of a calculation like this,

[00:52:38.130]
if that’s not working, then I can change.

[00:52:40.370]
And I could change that to 800, 1200 et cetera

[00:52:44.240]
until it sort of makes sense.

[00:52:46.210]
So we can use (mumbles).

[00:52:50.040]
<v Aaron>Yeah.</v>

[00:52:50.873]
<v Steve>We can use distance buffers,</v>

[00:52:52.060]
like let’s add on

[00:52:53.900]
Contaminants Extension functionality within works

[00:52:56.530]
and that might help you as well

[00:52:58.440]
decide on what kind of values to use

[00:53:00.630]
within these calculations.

[00:53:02.300]
Classification, if you’re using kriging,

[00:53:06.046]
if you have used directly kriging variance

[00:53:09.270]
as it often quite nicely to use spacing as well.

[00:53:12.980]
So there’s lots of different ways,

[00:53:15.220]
but in more of the LT user.

[00:53:17.520]
<v Aaron>Fantastic. Well thank you so much Steve.</v>

[00:53:20.317]
We’re coming to the end of…

[00:53:22.107]
We had a lot of time now.

[00:53:23.830]
Thank you, everyone who joined

[00:53:25.470]
and particularly if you asked a question.

[00:53:27.902]
If we didn’t get to your question

[00:53:29.800]
or if you think of one after today’s session,

[00:53:32.360]
then please feel free to reach out

[00:53:35.990]
at the [email protected].

[00:53:38.100]
And if there are anything we couldn’t get though today,

[00:53:40.480]
we’ll also reach out directly to people.

[00:53:44.720]
As I said, you will be receiving a link after today

[00:53:49.450]
with the recording as well as

[00:53:51.183]
the link for that technical series,

[00:53:53.700]
which is a really good resource to go through yourself

[00:53:56.960]
using the Contaminant Extension.

[00:53:59.160]
And if you don’t already have access to it,

[00:54:00.790]
we can give you a trial of it as well.

[00:54:03.410]
So you can give it a go.

[00:54:07.840]
Thank you so much for joining today’s webinar.

[00:54:09.960]
We really hope you enjoyed it

[00:54:11.450]
and have a fantastic rest of your day.

[00:54:13.590]
<v Steve>Thank you very much.</v>