Precision Agriculture 101 – Variability Capture

Video Transcript

0:01

hello everyone its Aaron Bremer with

0:03

Veritas farm business management this

0:06

video is designed to answer a question

0:08

that we get a regular basis at Veritas

0:11

and that is is precision AG right for my

0:15

farm to answer this question we need to

0:18

look at what we consider to be the

0:19

backbone of precision AG at Veritas now

0:23

most people in the industry are going to

0:25

tell you it’s all about data and

0:27

although data plays an important part

0:29

let’s face it Veritas is dedicated to

0:32

looking at data analyzing data and

0:35

creating insights from data however we

0:40

don’t believe data is the true backbone

0:43

of precision AG it’s important but the

0:46

big question that we have to answer is

0:49

this variability do you have enough

0:53

variability on your farm that precision

0:56

AG technology and the concept of

1:00

precision AG is going to make sense so

1:03

this video we’re going to talk about a

1:04

tool that we’ve built to be able to

1:07

answer that question how much

1:09

variability do you have and is it enough

1:12

that you’re going to have to consider

1:14

precision AG to truly get the most out

1:17

of your production system so with that

1:20

in mind most people in the precision AG

1:24

space are gonna tell you the most

1:26

important layer the layer to get things

1:28

started is a good soil test layer so

1:32

let’s talk about soil tests there’s a

1:35

whole bunch of different systems out

1:37

there we’re not going to go over all of

1:38

them but we’re going to go over some of

1:40

them and how Veritas looks at these

1:44

different systems in order to be able to

1:46

determine how much variability is

1:48

actually there

1:50

so the basic approach on soil sampling

1:53

is what’s called bulk sampling in

1:55

Ontario this is one sample every 25

1:57

acres I know in Western Canada it’s very

2:00

common to do one sample for every

2:02

hundred and sixty acres really doesn’t

2:05

matter what your scale is that you’re

2:06

using

2:07

what bulk sampling is doing is taking

2:11

the field and sampling in very large

2:14

sections and the samples are collected

2:19

mixed together and sent off to the lab

2:21

it gives you a good idea as the baseline

2:24

of what’s going on in your field but it

2:27

doesn’t do a whole lot for capturing

2:28

variability you can see the example I

2:30

have given it here above is a forty acre

2:33

field split into two samples below it in

2:37

case you want to have a little better

2:39

understanding as to what’s going on is

2:42

an imaginary 100 acre field and we’ve

2:45

split it into four different samples you

2:47

can see towards the north end the top

2:49

part of that imaginary field there’s two

2:52

samples the east and west that are a

2:55

little later in color so maybe not as

2:58

much organic matter maybe a little bit

3:00

sandy R whereas further to the south the

3:02

bottom on both the east and west things

3:05

are a little bit darker so that is been

3:07

the standard for soil sampling in fact

3:10

in Ontario one sample every 25 acres is

3:13

considered the best management practice

3:16

starting in B I’m gonna say the early

3:19

1990s we started to do grid sampling and

3:22

what grid sampling is is basically

3:25

subdividing the field into smaller

3:28

samples and then each spot is sampled

3:32

separately the data comes back from the

3:34

soil lab and there is software that

3:37

takes all this data and attempts to

3:40

interpret it and overlay what the

3:42

software what the computer thinks your

3:45

field is going to behave so this is

3:47

called grid sampling now there’s some

3:48

variations and you can have different

3:50

sizes of of grids

3:52

the next SAP up and this has become a

3:54

little bit more common zone sampling

3:57

now what zone sampling does is instead

3:59

of taking those samples every two and a

4:01

half acres it takes the samples based on

4:05

predefined zones maybe based off of

4:09

yield maybe based off of imagery

4:11

whatever the case may be and those zones

4:14

are then sampled now sometimes you could

4:17

combine the zones so you can have a

4:19

whole bunch of hilltops all combined

4:22

into one or you might soil sample those

4:25

hills separately it completely depends

4:27

on how intensive those says zones are

4:29

gonna be when you take all of that data

4:32

you plug it in you can see our imaginary

4:35

hundred acre field we’re starting to see

4:37

a little bit more detail than what we

4:39

did with the grid or the bulk that’s not

4:41

always true but in this case it is the

4:45

most recent in you’re probably starting

4:48

to recognize a picture at the bottom but

4:49

the most recent soil sampling system is

4:52

the soil sensing there’s a bunch of

4:55

different systems on the market you have

4:57

soil optics you have s is from Trimble

5:01

you have varus you have SWAT box there

5:04

is a lot of different systems out there

5:06

and they do an incredible job of

5:08

capturing detail to the point where you

5:11

can start to see my imaginary 100 Acre

5:13

field at the bottom probably looks very

5:15

familiar

5:15

so when you look at the total amount of

5:17

variability of that field you can see

5:20

each system captures a different amount

5:22

of variability that doesn’t mean that

5:24

one is necessarily better than the other

5:26

depending on what you want to do with it

5:29

and how much that variability is going

5:31

to impact you

5:32

so the question we get yes I have some

5:36

variability in my field or I don’t think

5:39

I have that much variability in my field

5:41

I don’t think it’s gonna pay I don’t

5:42

want to make that investment help me

5:45

understand

5:45

if I should make that investment and

5:47

that’s what the rest of this video is

5:49

about because we’re going to show you a

5:50

product that well maybe not a product a

5:52

tool that we have built to be able to

5:55

capture that amount of very building

5:57

so we’re gonna start with this field

5:59

this field the picture on your left the

6:02

red to green image is a field that has a

6:07

lot of variability so that variability

6:11

can be from an image it can be from

6:14

intensive soil scanning it can be from

6:17

yield what it is is it just shows you

6:19

the total amount of variability in that

6:21

field the graph I call it a graph my

6:26

data scientist Brett is probably going

6:27

to dispute whether that’s a graph she’s

6:30

probably gonna tell you that it’s a

6:31

histogram what that shows is to the

6:33

distribution of the different data

6:36

points so how far apart are they so you

6:38

can see you’ve got data points that are

6:40

running from a number of five all the

6:44

way up to 20 so you have a significant

6:48

variation and you can see there’s

6:50

different peaks across the field so you

6:53

the closer that those Peaks are together

6:56

or combined into one the less variable

6:59

you have are the more spread out it is

7:01

the more individual Peaks the more

7:03

variable it is so what what does it mean

7:07

to your field well what we did was we

7:10

took that total variability and we it’s

7:13

about 100 Acre field in this case and we

7:15

split it up using that Ontario best

7:18

management practice concept of one

7:20

sample every 25 acres and then we

7:23

gridded it or average by polygon and

7:26

what you can see is when you do that and

7:30

you measure the amount of variability so

7:31

basically what you’re doing is you’re

7:33

comparing the histogram of the total

7:36

variability to the histogram of the

7:38

25-acre sampling and when you do that

7:41

you capture 19 percent of that

7:44

variability I told you this field has

7:46

nice bit of variability as a farmer I

7:49

would not be happy capturing only 19

7:52

percent of the variability so what else

7:55

can we do well we could go to a 10 acre

7:57

great approach and you can see we’re

8:00

capturing a little bit more of the

8:01

variability in fact we’re capturing up

8:03

to 45 percent like I said this is a

8:06

great approach so you may be say you

8:08

know what I want to use a 10 acre

8:10

so what does that do for me well in this

8:13

case you can see the grids are a lot

8:16

different than the zones

8:18

there’s your zones there’s your grits

8:20

back to your zones but in both cases

8:23

you’re capturing 45 percent of the

8:26

variability it’s a different 45 percent

8:29

I’ll give you that

8:30

but in both cases you’re capturing to a

8:34

45 percent variability so maybe you’re

8:37

gonna say you know what 45 percent I’m

8:39

not convinced that that is enough

8:42

variability to capture for this field I

8:44

want to catch more all right let’s go to

8:47

that five acre grid Wow we’ve just gone

8:50

from 45 percent to 80 percent by going

8:54

from that 10 acre um whether it was grid

8:57

or zones to a 5 acre grid you’ve

9:02

captured 80 percent of the variability

9:03

that’s pretty impressive now you could

9:07

also do it a five-acre zones once again

9:10

74 percent so you’ve dropped a little

9:12

bit of that variability by going to the

9:14

zones and that’s just because of how the

9:15

zones are made depending on how you make

9:18

your zones you might be able to capture

9:19

more of that variability that’s the cool

9:22

part about this tool is you can start

9:24

figuring out what is the best way to

9:26

make money on your farm or invest on

9:29

soil sampling do you want a soil sample

9:32

every 25 acres do you want to the soils

9:35

have what every 5 acres do you want to

9:37

use grids or do you want to use polygons

9:40

/ shares owns this is a really cool tool

9:44

you can send us your field boundary and

9:46

we can start doing this what it’s your

9:49

yield data or with imagery even before

9:52

you make that investment on soil

9:53

sampling so before you make that

9:56

investment you can decide how much

9:58

variability you want to capture now if

10:02

you want to spend a little bit more

10:03

money you can go up to the two and a

10:05

half acre grid that’s gonna give you a

10:07

little bit more variability captured

10:09

you’re getting it up to 81% but like I

10:12

said obviously twice as much many soil

10:14

samples the price starts to go up so it

10:17

depends on your operation and how much

10:19

variability you want to

10:21

if you were comfortable with capturing

10:24

only 19% of that variability perfect go

10:26

with the bulk soil sampling if you want

10:29

to capture a little bit more you can

10:31

start to become a little bit more

10:32

intensive so there’s our two and a half

10:34

acre grid our two and a half acres zone

10:36

so these zones were obviously drawn a

10:39

little bit better than our five acre

10:40

zones but we only captured an additional

10:44

1% of that variability still not a whole

10:46

lot of difference between the two now

10:49

the cool part about a two and a half

10:50

acre grid and two and a half acres zone

10:52

they’re probably going to be about the

10:55

same price but yeah our two and a half

10:58

acres zones have been more accurate then

11:01

maybe we would have been up to the high

11:03

eighties and you’d say hey that’s the

11:04

one I want to go with it’s completely up

11:07

to you that’s the cool part about this

11:09

tool your fields your variability you

11:12

get to decide how you want to capture it

11:14

or if you want minute don’t want to

11:16

capture no if it is super uniform field

11:19

run it through this tool and yet might

11:22

come back that bulk soil sampling

11:24

catches 80% of the variability all right

11:27

then if that’s the case I’m doing a

11:30

precision egg that would be my question

11:31

to you so that’s a pretty cool thing now

11:34

if you really want to spend start

11:35

spending some money you can get

11:37

intensive England’s what a one-acre grid

11:39

that’s gonna capture 91% you know grid

11:42

and in the zones like I said the zones

11:45

are maybe not as good as it could be

11:47

you’re at 85% here’s my thoughts if you

11:51

want this level variability captured

11:53

it’s time to start investing in that

11:55

soil scanning sampling that technology

11:58

whether it be soil optics or swath box

12:01

or verus or any other system like sis

12:06

from Trimble whatever your system wants

12:07

to be because that’s gonna be a little

12:09

bit more expensive so now when you want

12:12

to be a little bit more expensive you

12:13

can take advantage of that for your farm

12:16

but like I said maybe you don’t want to

12:21

invest to the highest level maybe you

12:24

want to be a little less expensive

12:25

that’s the cool thing about this

12:28

variability capture tool

12:30

you can decide how much variability you

12:32

wanted to that capture in your farm and

12:34

then you can make smarter decisions on

12:37

how precision AG works for you if you

12:40

don’t have that much variability

12:42

don’t go precision egg but if your

12:46

answer to should precision egg work for

12:49

my farm and is you don’t know how much

12:52

variability you’ve got on your farm or

12:53

how you’re going to be able to capture

12:55

it what a cool way to investigate to

12:59

determine if it’s the right decision for

13:01

you or not if you have this type of

13:04

information you can make better

13:05

decisions on where you want to invest on

13:08

your farm thank you very much for

13:10

listening

13:10

hopefully you’ll check out a few more of

13:12

our videos

Make Deveron your trusted in-field partner

Let's TaLK