3.5 Tiny Bytes: Landsat vs Sentinel for NDVI and Variable-Rate Maps

WEBVTT
00:00:00.239 --> 00:00:05.360
And now it's time for a tiny bite of knowledge.
00:00:14.080 --> 00:00:18.800
On this episode of Tiny Bytes, Sarah and I are heading out of this world.
00:00:18.800 --> 00:00:22.800
So buckle up because we're talking satellite imagery.
00:00:26.719 --> 00:00:31.440
It's all about Landsat and Sentinel imagery today.
00:00:31.440 --> 00:00:38.479
Here at GK Technology, we actually use a lot of imagery, satellite imagery for a lot of different reasons.
00:00:38.479 --> 00:00:41.759
A lot of NDVI data, which is a vegetative index.
00:00:41.759 --> 00:00:44.799
You can listen to other tiny bite episodes about that.
00:00:44.799 --> 00:00:55.520
But today we're going to actually compare and contrast a little bit of the satellites that we use the most, which are Landsat and Sentinel.
00:00:56.479 --> 00:01:02.159
So both of these satellite imagery sources are both government operated.
00:01:02.159 --> 00:01:09.439
So Landsat is owned and operated by the USA through NASA and the USGS.
00:01:09.439 --> 00:01:18.879
And then Sentinel is another satellite imagery provider that's operated by the European Union and the European Space Agency.
00:01:18.879 --> 00:01:24.079
So because of this, we get access to free imagery through both of these.
00:01:24.079 --> 00:01:26.159
Did you say free?
00:01:26.400 --> 00:01:29.840
There's free, free imagery for everyone.
00:01:29.840 --> 00:01:37.120
Yes, you could even make your own library of uh your own satellite imagery library from these data sources if you want.
00:01:37.120 --> 00:01:47.599
Here at GK, we actually do have a huge library, vast, from both of these data sources on our servers, um, ready to go.
00:01:47.599 --> 00:01:53.760
And if that is something that you would choose to use, we could certainly help you connect, connect you with that opportunity.
00:01:54.319 --> 00:01:55.359
Absolutely.
00:01:55.359 --> 00:02:10.719
And before we get into the details of like what are the big differences between this Sentinel and this Landsat, I do want to talk a little bit about why satellite imagery is such a great tool for use as an agronomist and as somebody making variable rate maps.
00:02:10.719 --> 00:02:21.520
So to Sarah touched on NDVI before, with the satellite imagery, the images that are collected, we're able to get multiple bands that help us make the calculation for NDVI.
00:02:21.520 --> 00:02:25.120
And this is a super cool thing about it.
00:02:25.120 --> 00:02:34.800
So A, both Landsat and Sentinel, their satellites cover the same place on Earth multiple times throughout the year.
00:02:34.800 --> 00:02:36.639
So it's not just once, it's not just twice.
00:02:36.639 --> 00:02:39.199
They're going through multiple times during the season.
00:02:39.199 --> 00:02:41.759
But here's the really, really big thing.
00:02:41.759 --> 00:02:46.400
What they do is they go in a sun-synchronous orbit.
00:02:46.400 --> 00:02:52.960
And what that means is that in North Dakota, Sentinel will typically take pictures at about 10:30 a.m.
00:02:52.960 --> 00:02:56.159
and Landsat 8 will take pictures at about 10 a.m.
00:02:56.159 --> 00:03:15.280
And meaning that in the mid-morning, they're taking photos that are early enough that A, clouds haven't fully developed and that like on those thunderstorm days, and B, because they're collecting those pictures at the same time when they're going across an area, they're able to get consistent shadows.
00:03:15.280 --> 00:03:25.919
So you don't have to worry about as you're putting multiple layers of imagery together, that you're gonna have shadows coming from different angles from different tree rows that you'll have to think about when you're putting images together.
00:03:25.919 --> 00:03:35.199
You just having that sun-synchronous imagery from these satellites is so crucial and takes another piece out of any sort of variability that might come with imagery.
00:03:35.199 --> 00:03:36.960
So super cool.
00:03:38.800 --> 00:03:47.759
It is important to remember with sun synchronous imagery that you know it does matter what time of the year we're actually taking these pictures because the season matters, right?
00:03:47.759 --> 00:03:53.840
In North Dakota, at uh we our sun rises and sets at different times of the year.
00:03:53.840 --> 00:04:03.919
If you have to work from sunup to sundown in June, those are long days because you're talking about like four o'clock in the morning until about 10 o'clock at night.
00:04:03.919 --> 00:04:08.960
Whereas in uh December, on the shortest day of the year, we have much less sun.
00:04:08.960 --> 00:04:18.720
So the time of the year makes a difference, but within that, within that time of the year in a specific spot, you're getting that imagery at the same time.
00:04:18.720 --> 00:04:21.120
So at the same time of the day.
00:04:21.120 --> 00:04:24.160
So that that sun-synchronous thing really is a big deal.
00:04:24.160 --> 00:04:26.959
It's a it's important to remember a couple of those things though.
00:04:27.199 --> 00:04:35.040
Talking about these two imagery types, one of the big differences I think about using Landsat versus Sentinel is just the resolution.
00:04:35.040 --> 00:04:45.519
And so when we're talking about resolution of imagery, what we mean by that is how uh basically how much area each pixel, quote unquote, of the image represents.
00:04:45.519 --> 00:04:50.480
And with with Landsat, it's a low resolution.
00:04:50.480 --> 00:04:56.800
So meaning that each of the pixels is 30 meters, meaning that it's you're not getting as much detail.
00:04:56.800 --> 00:05:00.720
But with Sentinel, what we're getting is a 10 meter resolution image.
00:05:00.720 --> 00:05:02.160
So we're able to get more detail.
00:05:02.160 --> 00:05:04.560
And remember, 10 meters, that's 30 feet.
00:05:04.560 --> 00:05:09.680
That's probably the width of or smaller than the width of most planters nowadays.
00:05:09.680 --> 00:05:15.199
Just that difference in resolution size is a big driver of when I'm looking at satellite imagery.
00:05:15.199 --> 00:05:18.560
A lot of the times, most of the time, I'm I'm choosing Sentinel.
00:05:18.560 --> 00:05:22.319
Um, but that doesn't mean that Landset doesn't have a use either.
00:05:23.040 --> 00:05:30.720
So within that, um, a pixel uh is a picture that we see on a computer is made up of pixels.
00:05:30.720 --> 00:05:39.120
It's little boxes, um, if you will, or little squares that have uh their own color assigned to them.
00:05:39.120 --> 00:05:43.680
And the more pixels you have, the more detailed your image is.
00:05:43.680 --> 00:05:56.879
So if you have one pixel represented by 30 meters of area on the ground, that has less detail than one pixel with 10 meters represented on the ground.
00:05:56.879 --> 00:06:01.360
That's kind of a resolution is kind of a difficult concept to wrap your head around.
00:06:01.360 --> 00:06:04.319
Maybe we should just have a whole podcast about that one day.
00:06:04.639 --> 00:06:06.399
That's actually a really good idea.
00:06:06.399 --> 00:06:09.759
Think about the first camera you ever had on your cell phone, right?
00:06:09.759 --> 00:06:17.360
Maybe a flip phone, you had a little camera on it, and now you look back at the pictures and say, man, why does it look like I was taking photos with a potato?
00:06:17.360 --> 00:06:20.079
You know, that's that's a low resolution image.
00:06:20.079 --> 00:06:24.879
And now, like with most iPhones, Androids, the cameras on them are so amazing.
00:06:24.879 --> 00:06:27.680
And that big difference is just resolution quality.
00:06:27.680 --> 00:06:36.639
So older phones, low resolution, newer phones, higher resolution, better images, more data per pixel.
00:06:36.639 --> 00:06:42.480
And that's what we're talking about with the 30-meter resolution with Landsat versus 10 meter for Sentinel.
00:06:42.480 --> 00:06:49.120
10 meter, it's a lower number, but it's higher resolution and gets us some good data, better data.
00:06:49.759 --> 00:07:02.000
So when I'm thinking about Landsat versus Sentinel, one of the things that I think about is when the satellites started flying and how the age of the actual programs.
00:07:02.000 --> 00:07:07.360
So the United States of America has a very old satellite program.
00:07:07.360 --> 00:07:12.399
Landsat 5 started flying in 1984.
00:07:12.399 --> 00:07:24.800
Now, the actual Landsat program was actually started back in the 1970s, but the Landsat 5, which is really the first satellite that we have access to, the imagery from it, started in 1984.
00:07:24.800 --> 00:07:32.079
And yes, on our servers, we actually have data that goes all the way back to 1984, and that's what it's from, Landsat 5.
00:07:32.079 --> 00:07:35.279
That satellite flew until 2013.
00:07:35.279 --> 00:07:42.319
That had a very long mission, a very long life, satellite uh Landsat 5.
00:07:42.319 --> 00:07:50.879
Landsat 7 was the next satellite that flew, which flew from 1999 until 2024, which was fine.
00:07:50.879 --> 00:07:56.959
Except someplace in that time frame, there was it was in 2003.
00:07:56.959 --> 00:08:02.319
In 2003, there's actually something that happened to that satellite.
00:08:02.319 --> 00:08:06.000
And so there's actually like lines going across the scenes.
00:08:06.000 --> 00:08:11.519
And if your field ended up underneath that line, you had big, huge gaps in the data.
00:08:11.519 --> 00:08:18.560
As a matter of fact, 22% of Landsat 7 scenes were missing because of that.
00:08:18.560 --> 00:08:22.800
So sometimes we had good imagery if you weren't underneath that, and sometimes you didn't.
00:08:22.800 --> 00:08:29.040
And as I've been making maps, I have actually used some Landsat 7 data, and then there's been fields that it just doesn't work.
00:08:29.040 --> 00:08:32.159
Landsat 8 was launched in 2013.
00:08:32.159 --> 00:08:36.639
It's still up and flying, and Landsat 9 was launched in 2021.
00:08:36.639 --> 00:08:39.919
So the Landsat program, very old.
00:08:39.919 --> 00:08:49.600
Sentinel was started in 2015, is when this was Sentinel 2A was launched.
00:08:49.600 --> 00:09:02.000
Sentinel actually has a series of like six um different missions up there, and each mission has at least two satellite constellations at a time.
00:09:02.000 --> 00:09:14.080
And satellite two or Sentinel 2 is the one that we use the most in in agriculture right now, and what we have our our data on the uh on our servers.
00:09:14.080 --> 00:09:20.000
2A was launched in 2015, like I said, and 2B was launched in 2017.
00:09:20.000 --> 00:09:22.720
And so those are the baseline operations.
00:09:22.720 --> 00:09:26.799
The age of the program is one of the things that's different between these.
00:09:26.799 --> 00:09:31.120
Landsat is very old, and Sentinel is a lot newer.
00:09:31.120 --> 00:09:35.360
Landsat 9 was launched in 2021.
00:09:35.360 --> 00:09:40.240
There it is, there, and now is actually operating in constellation with eight.
00:09:40.240 --> 00:09:45.200
So um, which is which is a new thing, that constellation concept.
00:09:45.200 --> 00:09:47.679
And we can we can talk about that in a second.
00:09:47.679 --> 00:09:52.639
Landsat 8 has its mission slated to end in 2028.
00:09:52.639 --> 00:10:01.679
And one of the things that's happening with these programs is they're trying to get satellites launched before something actually goes wrong with them.
00:10:01.679 --> 00:10:26.240
I'm wondering, and I don't have any information on this, but I'm wondering if what happened with Landsat 7 and the lines in the data, I'm wondering if that created enough challenges that we wanted to make sure that we were still logging data, even if we couldn't, even if the constellation had to be interrupted so that we only ended up with one satellite logging at a time if something were to happen to the to the other one.
00:10:26.240 --> 00:10:32.960
There is actually another Landsat satellite slated to fly again.
00:10:32.960 --> 00:10:36.480
Um, and that's going to be called Landsat next mission.
00:10:36.799 --> 00:10:38.080
Ooh, fancy.
00:10:38.080 --> 00:10:39.600
Dun dun dun.
00:10:39.600 --> 00:10:46.159
And I think it's important to point out and just talk for a brief second about constellation.
00:10:46.159 --> 00:10:52.159
So having these two paired satellites together is what we consider a constellation.
00:10:52.159 --> 00:10:59.919
And it sounds like the goal of this is to um reduce any possibilities of one of them failing and not getting data.
00:10:59.919 --> 00:11:06.320
So there'll always be coverage from one of them in the Landsat constellation and the Sentinel constellation.
00:11:06.320 --> 00:11:13.919
There'll be at least one operating at all time to avoid issues that were presented with like Landsat 7 and that streak across.
00:11:14.240 --> 00:11:22.240
And in addition to that, what this constellation setup allows for us to also have is more imagery on a whole.
00:11:22.240 --> 00:11:31.440
So the Sentinel satellites revisit any given location every 10 days with a satellite on the opposite side of the earth.
00:11:31.440 --> 00:11:48.879
So they're offset by 180 degrees, opposite side of the earth, which means that once you have that constellation running and each satellite is logging every 10 days, you get to double how much imagery you're getting in that spot.
00:11:48.879 --> 00:11:53.360
So now you're revisiting every five days, and that's a really big deal.
00:11:53.360 --> 00:11:55.039
So we're getting more imagery.
00:11:55.039 --> 00:12:01.360
You have to remember that because of cloud cover, not all of that imagery that gets logged is going to be good.
00:12:01.360 --> 00:12:13.120
And so we increase our frequency of having good usable data because we're effectively doubling our imagery um gathering.
00:12:13.120 --> 00:12:21.600
And like Jodi said, it gives us a backup plan so that if something happens to one of the satellites, we're still getting data.
00:12:21.919 --> 00:12:29.679
And Sarah, what's so you mentioned that in Sentinel the average revisit is about five days between imagery.
00:12:29.679 --> 00:12:31.279
What's that like for Landsat?
00:12:31.279 --> 00:12:36.000
Landsat is eight days.
00:12:36.399 --> 00:12:41.279
So Landsat 8 and Landsat 9 are constellation together.
00:12:41.279 --> 00:12:44.559
And so they're doing a 16-day revisit.
00:12:44.559 --> 00:12:57.440
The Landsat ones are doing a 16-day revisit, which means in a constellation, then you're getting between the two of them, then either Landsat 8 or 9 is logging data every eight days.
00:12:57.440 --> 00:13:09.519
So effectively, what's happening, for example, with Landsat is when we only had Landsat 8, we were getting about 13 overpasses a year for any given location.
00:13:09.519 --> 00:13:20.639
But because we we have that second satellite up there operating in a constellation, we now are increasing those overpasses up to 26 per year with the Landsat.
00:13:20.639 --> 00:13:22.480
Pretty cool, huh?
00:13:22.799 --> 00:13:25.039
I'm glad we could take this trip up into space.
00:13:25.039 --> 00:13:36.320
And the thing that I love beyond the fact that this imagery is free and available for the public is the fact that I can learn about space as an agronomist, something I never thought that I would ever have to do.
00:13:36.720 --> 00:13:46.080
Imagine if our grandfathers knew that we had we were talking about farming and using satellites to farm something from space.
00:13:46.080 --> 00:13:48.240
Wow.
00:13:48.240 --> 00:13:57.440
Tune in next time for a tiny bite of knowledge from GK Technology, where we have a map and an app for that.