NASA ARSET_ Estimating Biomass Change with GEDI and the OBIWAN API

0:01 Speaker 1: Welcome to this our SET training series, Estimating Biomass and

0:05 Speaker 1: Change with Jedi and the Obi one Api. Today is

0:09 Speaker 1: the second and last session of this webinar. The focus

0:13 Speaker 1: of today's session is estimating biomass change with Jedi and

0:17 Speaker 1: the Obi wan Api. My name is Erica Potist and

0:20 Speaker 1: I'm a scientist at NASA's Jet Propulsion Laboratory in Pasadena, California,

0:25 Speaker 1: and I'm also an instructor with the RSET program. I

0:29 Speaker 1: will be hosting this training today. There are two parts

0:36 Speaker 1: to this training. The first session was last week last

0:40 Speaker 1: Thursday on estimating biomass using Jedi, and today we will

0:46 Speaker 1: cover estimating biomass change, so there'll be a theoretical component

0:52 Speaker 1: and then there'll be a demo using the Obi Wapi.

0:56 Speaker 1: There is a homework associated with this training. It's It

1:00 Speaker 1: can be accessed through the training web page and the

1:03 Speaker 1: due date is June eighteenth. A certificate of completion will

1:07 Speaker 1: be provided to those participants who attend both live sessions

1:12 Speaker 1: and complete the homework assignment by the due date. So

1:16 Speaker 1: let's get started with today's session. We have one guest instructor.

1:23 Speaker 1: He is doctor Sean Healey who also presented in Session one.

1:28 Speaker 1: Doctor Healey is a research ecologist with the United States

1:31 Speaker 1: Forest Service the end Department of Agriculture. The learning objectives

1:38 Speaker 1: for today's session. By the end of the session, participants

1:42 Speaker 1: will be able to identify the key concepts in carbon monitoring,

1:46 Speaker 1: including system requirements, decision making needs, and the concept of additionality.

1:52 Speaker 1: Participants will be able to recognize howbi I estimates biomass

1:57 Speaker 1: change it's including its use of jet lance at time

2:01 Speaker 1: series and underlying data infrastructure. Participants will be able to

2:07 Speaker 1: evaluate the uncertainty and validation of the data, using for

2:11 Speaker 1: service inventory data to assess the accuracy and position of

2:14 Speaker 1: obi WAN change estimates, access the open source IAPI to

2:20 Speaker 1: generate estimates of biomass change and areas of interest, and finally,

2:24 Speaker 1: apply obi WAN tools and APIs to visualize biomass change

2:30 Speaker 1: and compare carbon gains against different climate scenarios. How to

2:36 Speaker 1: ask questions. Please write your questions in the Q and

2:39 Speaker 1: A window, which you can find on the bottom right

2:41 Speaker 1: under the three points. There will be a sub menu

2:45 Speaker 1: that says Q and A. We will address your questions

2:48 Speaker 1: and at the end of this session, and feel free

2:51 Speaker 1: to enter your questions during the presentation we will try

2:55 Speaker 1: to answer all of the questions during the Q and

2:57 Speaker 1: A session. However, any remaining questions will be answered in

3:02 Speaker 1: a Q and A document that we will be assembling,

3:05 Speaker 1: and we will be posting that document on the training

3:08 Speaker 1: website in approximately a week. Now, I would like to

3:14 Speaker 1: welcome our guest instructor, doctor Sean Healey, who also presented

3:18 Speaker 1: session one.

3:24 Speaker 2: Hello, and welcome to the second part of this training

3:29 Speaker 2: about biomass estimation with Obi Wan and JEDI. I'm Sean Healey.

3:35 Speaker 2: I'm speaking for doctor Jichang Yong. We both work for

3:39 Speaker 2: the USDA for Services Forest Inventory and Analysis Program FIA.

3:44 Speaker 2: That's the nation's national forest inventory with plots all over

3:50 Speaker 2: the country. So in the last session we talked about

3:55 Speaker 2: the JEDI mission, it's data products, the way that it

4:00 Speaker 2: estimates biomass for one point in time for one kilometer.

4:05 Speaker 3: Grid cells, and also how that logic.

4:10 Speaker 2: And the sampling theory can be used to estimate changes

4:16 Speaker 2: in biomass for any area that you.

4:18 Speaker 3: Care about that's over a few hundred hectares.

4:22 Speaker 2: Between different years, using lance at imagery and bootstrapping.

4:29 Speaker 3: So we will review some of that.

4:31 Speaker 2: We will also be talking about practical carbon assessment needs.

4:36 Speaker 2: By the end, I will be showing you obi Wan's API.

4:40 Speaker 2: That is how you and everyone else with an Internet

4:44 Speaker 2: connection can access the tools that are built into obi wan.

4:50 Speaker 2: So you know what kind of needs does obi wan address.

4:56 Speaker 2: It's really climate change mitigation in the forest sector. The

5:01 Speaker 2: proposition for pretty much any kind of mechanism to recognize

5:07 Speaker 2: climate change mitigation across agriculture or the energy sector and

5:13 Speaker 2: forests is to identify management activity or practice that reduces

5:20 Speaker 2: the impacts of climate change sustainably. This slide sort of

5:25 Speaker 2: shows the shape of that proposition in the forest sector,

5:29 Speaker 2: so I'll start at the bottom. A fundamental question is

5:32 Speaker 2: how much carbon in the way in the form of

5:34 Speaker 2: biomass is your forest storing now? And you know there

5:38 Speaker 2: are millions of biomass maps out there and they all

5:42 Speaker 2: can give you an answer.

5:44 Speaker 3: We talked about last time.

5:45 Speaker 2: If those maps were not made in a way that

5:49 Speaker 2: allows you to combine.

5:50 Speaker 3: The uncertainties of all those pixels that.

5:52 Speaker 2: Make up your area of interest, they're not as much

5:57 Speaker 2: useful in the context of practical carbon account So you know,

6:00 Speaker 2: an uncertainty framework for these estimates is critical. Last time

6:04 Speaker 2: we also talked about how through bootstrapping. We can answer

6:09 Speaker 2: the middle question how much carbon is your forest adding,

6:12 Speaker 2: so basically between last year and this year, or between

6:15 Speaker 2: twenty twenty and now, what is the change in biomass

6:18 Speaker 2: and the place that you that you manage. And then

6:21 Speaker 2: sort of the top level question, which echoes the main

6:25 Speaker 2: question and all of these frameworks is how does your

6:28 Speaker 2: management contribute to carbon storage trend? So that really implies

6:33 Speaker 2: a question about additionality. So you might be storing extra carbon,

6:38 Speaker 2: but maybe it was going to store extra carbon anyway,

6:41 Speaker 2: So we have to address additionality, which we will in

6:45 Speaker 2: a second. I just want to point out this is

6:47 Speaker 2: sort of by way of context for everything I'm going

6:50 Speaker 2: to be talking about regarding obi WAN, the sort of

6:56 Speaker 2: most common way that people practically do this, at least

6:59 Speaker 2: with romo sensing. Obviously, in a country like the United

7:02 Speaker 2: States that has an inventory that's been working for many

7:05 Speaker 2: years now, we use the inventory to answer these questions.

7:09 Speaker 2: But in many places that don't have that, the carbon

7:13 Speaker 2: standards use what people refer to as the spreadsheet method. Basically,

7:19 Speaker 2: there is a running total of which is called activity data,

7:23 Speaker 2: of how much forest is being added or subtracted through

7:28 Speaker 2: management and there's an emission factor, how much what's the

7:32 Speaker 2: mean biomass stored in the average hectare of forest? Most

7:36 Speaker 2: existing carbon standards are built around that. If you use it,

7:40 Speaker 2: you're aware that there are some limitations in terms of

7:44 Speaker 2: statistical difficulty of saying declaring something significantly. There are capability

7:50 Speaker 2: capacity issues in some places. You really have to be

7:54 Speaker 2: able to measure things, measure the area of forest in

7:57 Speaker 2: a consistent way through time. Maybe obi wan is the

8:01 Speaker 2: alternative to that paradigm. Okay, so back to obi Wan

8:05 Speaker 2: this top level question, how does your management contribute to

8:09 Speaker 2: carbridge storage trends? That gets to the question of additionality.

8:13 Speaker 2: So what would happen if you weren't doing the thing

8:16 Speaker 2: that you're doing. My example is in nineteen ninety four,

8:20 Speaker 2: the National Forces in the Pacific Northwest implemented a new

8:24 Speaker 2: forest plan. According to obi Wan, before that forest plan

8:29 Speaker 2: was implemented, the region was losing a small amount of carbon.

8:33 Speaker 2: It's very productive region in the nineteen eighties, and prior

8:36 Speaker 2: to the nineteen ninety four implementation of the plan there

8:39 Speaker 2: was a lot of harvest. After nineteen ninety four there

8:42 Speaker 2: was a surprisingly immediate bounce.

8:45 Speaker 3: Changes were all positive. So one way to look at.

8:49 Speaker 2: Additionality is over time, what hapened, what's happening now compared

8:53 Speaker 2: to what happened before you did whatever it is you

8:56 Speaker 2: are committing to do. The second way to address additionality

9:01 Speaker 2: is to compare the changes that you're seeing now in

9:04 Speaker 2: the place that you care about. Let's say it's a

9:06 Speaker 2: national park with areas outside of the national park, So

9:11 Speaker 2: that's a counterfactual sort of baseline. There are complications that

9:17 Speaker 2: should be considered. Generally, younger forests store carbon faster than

9:22 Speaker 2: older forests, So it's really like there are a lot

9:25 Speaker 2: of requirements I think for doing the counterfactual additionality assessment. Well,

9:31 Speaker 2: and I'm not going to talk as much about that,

9:33 Speaker 2: although you will see at the end that obi WAN

9:37 Speaker 2: does allow you to do this kind of additionality assessment.

9:41 Speaker 2: Pressing into the guts of obi WAN, please recall from

9:46 Speaker 2: the first session that its inferences are really based upon

9:52 Speaker 2: bootstrapped time series of biomass maps. So the example I

9:57 Speaker 2: showed before, Let's say that you care about the change

10:00 Speaker 2: between nineteen ninety and twenty twenty in a particular place.

10:05 Speaker 3: There are several.

10:06 Speaker 2: Bootstrapped time series and all of These bootstraps recall our

10:11 Speaker 2: realization of the uncertainties in both levels of model, both

10:15 Speaker 2: the Jedi level of model and the Lancet level of model.

10:19 Speaker 2: The first bootstrap suggests that the area we care about

10:22 Speaker 2: lost fifteen tons per heck there the second bootstrap suggests

10:26 Speaker 2: we lost thirteen. The third suggests we lost twenty. So

10:31 Speaker 2: across a lot of different bootstraps, that's where our inference

10:34 Speaker 2: comes from. So that's how Jedi works, and we don't

10:39 Speaker 2: just have to look at changes. What I just talked

10:43 Speaker 2: about is a second bullet, the difference between for the

10:45 Speaker 2: same place between two points in time. But we can

10:49 Speaker 2: also look for a single point in time. Let's say

10:51 Speaker 2: we just care about the biomass in twenty twenty. Well,

10:54 Speaker 2: the first the first bootstrap says eighty the second says

10:57 Speaker 2: eighty four.

10:58 Speaker 3: Et cetera.

10:59 Speaker 2: So we could get a distribution around a point a

11:02 Speaker 2: single point in time estimate. We've just talked about the

11:05 Speaker 2: differences between two points in time. We can look at

11:10 Speaker 2: two forms of additionality that I just mentioned. The first

11:13 Speaker 2: is change in the rate of gain between two periods.

11:16 Speaker 2: So we could ask nineteen ninety to two thousand, what's

11:21 Speaker 2: the rate of change within the same bootstrap and compare

11:25 Speaker 2: that to what that bootstrap says is the rate of

11:27 Speaker 2: change between twenty ten and twenty twenty and get a

11:31 Speaker 2: difference between those rates. And that's what we look at

11:35 Speaker 2: across bootstraps. So we get a distribution of that change

11:38 Speaker 2: of change, if you will, the sort of additionality over time,

11:42 Speaker 2: and we get and the distribution gives us an uncertainty.

11:46 Speaker 2: We can also compare two places in this way using

11:52 Speaker 2: the different bootstraped time series. So obi Wan really does

11:56 Speaker 2: allow us to look at additionality questions and really gives

12:02 Speaker 2: us a fairly comprehensive way to look at monitoring carbon

12:06 Speaker 2: change within areas that.

12:07 Speaker 3: You care about.

12:08 Speaker 2: So this issue calibration is really important to obi Wan

12:15 Speaker 2: Optical imagery like Lansat and Sentinel to do not see

12:19 Speaker 2: through the canopy. So basically, as forests grow for the

12:23 Speaker 2: first part of their lives, you know, the pixels get

12:26 Speaker 2: greener as more forest grows, and you know there is

12:29 Speaker 2: pretty good signal about how much biomass there is, but

12:32 Speaker 2: once the canopy closes, it doesn't get significantly greener as

12:37 Speaker 2: more and more biomass or volume or basal area is added.

12:41 Speaker 2: As the forest grows, so that you know, people call

12:44 Speaker 2: this a saturation a signal saturation problem, and it manifests itself.

12:51 Speaker 2: When we're making we're using optical imagery to make predictions,

12:56 Speaker 2: we generally underpredict the very high biomass and we over

13:00 Speaker 2: predict the low biomass. And when we think about changes

13:04 Speaker 2: from a high condition to a low condition, we are

13:08 Speaker 2: almost always under predicting those changes because the high is

13:11 Speaker 2: under you know, everything gets squished towards the middle and

13:15 Speaker 2: you go from high to low, well that that change

13:17 Speaker 2: is also going to be low.

13:19 Speaker 3: So calibration is.

13:21 Speaker 2: A concept where if you have independent data in the

13:24 Speaker 2: places where you have predictions, you can across a large

13:29 Speaker 2: a sample of those come up with a calibration function

13:33 Speaker 2: that that simply is a correction, and it is a

13:37 Speaker 2: correction that is applied to minimize bias. It doesn't make

13:41 Speaker 2: your predictions better, but it removes the biases, so you're

13:45 Speaker 2: no longer systematically under predicting the high biomass and over

13:49 Speaker 2: predicting the low biomass. But again it's it's reducing bias,

13:53 Speaker 2: it's not reducing error. But that's that's really critical for

13:58 Speaker 2: this kind of application. This last line here is the

14:02 Speaker 2: only time I'm going to mention this calibration can also

14:05 Speaker 2: be used to convert obi Wan's estimates of biomass and

14:10 Speaker 2: biomass change to other variables. For instance, biomass is quite

14:16 Speaker 2: related to forest volume and also a basal area. So

14:20 Speaker 2: if from the National Forest Inventory, for instance, you had

14:23 Speaker 2: a bunch of independent plots, you could come up with

14:27 Speaker 2: a calibration from biomass to those other variables that you

14:31 Speaker 2: may care about in other settings. So calibration is really important,

14:37 Speaker 2: and I'm just going to talk a little bit about

14:40 Speaker 2: what calibration data are available where So Jedi data is

14:44 Speaker 2: available globally, as we have talked about at least between

14:48 Speaker 2: the latitudes of fifty one degrees north and south. So

14:52 Speaker 2: the first option, once we've made our biomass map with

14:56 Speaker 2: Lansat imagery, we can use Jedi data. The biomass predicted

15:01 Speaker 2: for Jedi points that we're not used in calibrating the

15:05 Speaker 2: lands At map, we can use those to calibrate our

15:07 Speaker 2: predictions and come up with that correction. And in general

15:12 Speaker 2: that correction when we apply it makes the high predictions

15:15 Speaker 2: higher and the low predictions lower. And how much data

15:18 Speaker 2: is available for this well, obi WAN uses about half

15:21 Speaker 2: Currently in twenty twenty six, it uses about half of

15:24 Speaker 2: the available data, which is several thousand. It's at least

15:28 Speaker 2: two thousand Jedi shots per ten kilometer square, So there's

15:34 Speaker 2: a lot of data available for this everywhere the section.

15:37 Speaker 2: The second option is not available globally. It's in places

15:41 Speaker 2: where you have repeated inventory plots that have coordinates that

15:47 Speaker 2: you trust, and we work for the National Forest Inventory,

15:51 Speaker 2: and you know, the.

15:53 Speaker 3: United States is where we are applying this.

15:55 Speaker 2: So basically we have several several thousand and plots in

16:01 Speaker 2: every state, sometimes up to ten thousand or even a

16:06 Speaker 2: little more in some big states where we have measured

16:09 Speaker 2: biomass in multiple points in time, and we can compare

16:13 Speaker 2: those measured changes to predicted changes from obi WAN. So

16:17 Speaker 2: instead of calibrating at the level of an individual point

16:21 Speaker 2: in time prediction, we're calibrating the estimates of change. And

16:25 Speaker 2: in general, when we apply those calibration functions, those really

16:29 Speaker 2: simple calibration functions, we end up making the losses bigger

16:33 Speaker 2: and the gains bigger. Okay, so now I am going

16:36 Speaker 2: to move into a couple of validation activities that we

16:42 Speaker 2: have performed one in Nepal and one in the United

16:46 Speaker 2: States through the great support of NASA's SEVERE program. We

16:51 Speaker 2: worked with Nepal's Forest Research and Training Center, which which

16:57 Speaker 2: is the entity that runs the National Force Inventory for Nepal,

17:00 Speaker 2: and they very kindly provided inventory data for three provinces

17:05 Speaker 2: in that country. And basically what we're looking at here

17:10 Speaker 2: is on the left is the obi WAN estimate of

17:13 Speaker 2: biomass change for a five year period. That's an orange

17:18 Speaker 2: compared to basically the spreadsheet method applied to Nepal's inventory data.

17:25 Speaker 2: This is not exactly aligned with what Nepal reported to

17:31 Speaker 2: unf Triple C, but it is very close. It's as

17:33 Speaker 2: close as we could make it with their with their data.

17:36 Speaker 2: So as you can see across the country, you know,

17:40 Speaker 2: there's pretty good correspondence with the amount of biomass that

17:43 Speaker 2: obi WAN predicts for this period. The obi WAN estimate

17:48 Speaker 2: for the whole country and for the three provinces that

17:53 Speaker 2: we had data for tends to be a little bit

17:56 Speaker 2: more dynamic, and I believe that is because obi WAN

18:00 Speaker 2: represents changing biomass in existing forests that are not being

18:05 Speaker 2: cut down. So many spreadsheet methods, which is the alternative

18:12 Speaker 2: treat the amount of biomass in protected forest as static

18:17 Speaker 2: and will result in sort of less dynamic estimates of change.

18:22 Speaker 2: Obi Wan accounts for that change, or at least predicts

18:25 Speaker 2: a change, a usually growth in forests that don't get

18:29 Speaker 2: cut down. So I think that that's probably why obi

18:33 Speaker 2: wan is showing a little bit more gain over this period.

18:36 Speaker 2: But it is pretty consistent both at the national and

18:40 Speaker 2: the province levels. So what can you do with that

18:43 Speaker 2: kind of information? This is Nepal. There are seventy seven

18:48 Speaker 2: districts in Nepal. One thing that we could think about

18:52 Speaker 2: doing with it is if Nepal were awarded red funding,

18:57 Speaker 2: which it was, and they were engaged in deciding how

19:01 Speaker 2: that red funding was going to be split among its districts,

19:05 Speaker 2: One could use obi wan to sort of systematically look at, well,

19:09 Speaker 2: which district is contributing which amount of the national level

19:15 Speaker 2: gain that the country is being compensated for. In no

19:19 Speaker 2: way would we presume to tell anybody how they should,

19:22 Speaker 2: you know, distribute their money, but this is sort of

19:24 Speaker 2: a systematic objective way to look at a distribution of

19:30 Speaker 2: gains across the country by different jurisdictions. So it could

19:35 Speaker 2: be just used, as you know, decision support tool in

19:39 Speaker 2: that context. Okay, so now I'm jumping into a validation

19:45 Speaker 2: activity that we've done here with FIA data. Again, FIA

19:49 Speaker 2: stands for forced inventory and Analysis. That's the part of

19:52 Speaker 2: the Forest service that Yong and I work for. So

19:55 Speaker 2: we're looking at results for two different states, and on

19:59 Speaker 2: the x saxis it's FIA observed above ground biomass density change.

20:06 Speaker 2: So every dot is a group of plots. The plots

20:11 Speaker 2: pretty much represent about a heck there. They subsample that

20:14 Speaker 2: heck there, but they're all about a heck there. So

20:17 Speaker 2: what we're basically doing is grouping together bunches of five

20:22 Speaker 2: hundred FIA plots and ordinating them on you know, the

20:28 Speaker 2: X axis here, like how much they're gaining or losing

20:31 Speaker 2: between two periods, and on the Y axis, comparing what

20:35 Speaker 2: obi Wan's the answer for that group of five hundred

20:38 Speaker 2: plots is. So, for instance, the blue line is what

20:43 Speaker 2: obi wan with no calibration, that's what obi wan gives us.

20:47 Speaker 3: And I mentioned that signal saturation, and.

20:50 Speaker 2: Really what we see is that obi wan systematically underestimates

20:55 Speaker 2: the losses and it underestimates the gains. So you know,

20:59 Speaker 2: that's a main reason why we use calibration. When we

21:03 Speaker 2: use the kind of calibration that was available in Nepaul

21:06 Speaker 2: and everywhere outside the US, using independent Jedi shots for

21:12 Speaker 2: at least in Alabama and in Maine, we see that

21:16 Speaker 2: really pushes the estimates of loss well below the dotted line,

21:21 Speaker 2: which is the one to one line. So we're really,

21:23 Speaker 2: at least in Alabama, for these different groups of plots

21:27 Speaker 2: pretty severely overestimating losses when we use that global calibration. However,

21:34 Speaker 2: once we get to the gain territory in Alabama anyway,

21:39 Speaker 2: that calibration really seems to work pretty well. It brings

21:42 Speaker 2: it up close to the one to one line. Calibrating

21:46 Speaker 2: with FIA repeat measurement data, actually it does pretty well.

21:52 Speaker 2: I mean we're seeing across all of these groups of

21:55 Speaker 2: FIA plots that the performance is not bad. This interval

22:01 Speaker 2: is the is basically the precision that the standard error

22:05 Speaker 2: that obi wan gives us for these changes.

22:08 Speaker 3: There is bias that is not accounted for in this interval.

22:12 Speaker 2: So when you get something out of when you get

22:14 Speaker 2: an interval out of obi wan, it's the precision, it's

22:17 Speaker 2: not the bias. So this kind of validation is crucial.

22:21 Speaker 2: In main we see something pretty similar except above a

22:24 Speaker 2: gain of twenty so places that are gaining a lot

22:28 Speaker 2: of biomass per year, we're not differentiating those cases very well.

22:34 Speaker 2: I would point out that at least with the FIA calibration,

22:37 Speaker 2: actually the global calibration as well. You know, the one

22:40 Speaker 2: to one line is within the confidence interval that you

22:44 Speaker 2: would get out of obi WAN. Okay, So moving on

22:48 Speaker 2: from estimation of change to estimation of additionality over time.

22:54 Speaker 2: That is, we're asking how much faster or slower is

23:00 Speaker 2: this place gaining carbon biomass than it was in a

23:04 Speaker 2: previous period. So you know, the same same two states

23:09 Speaker 2: we're asking. So we have multiple measurements, several measurements in

23:14 Speaker 2: many cases for fi A, so we we know we're

23:17 Speaker 2: able to measure this sort of acceleration or deceleration of

23:21 Speaker 2: gain at individual plots, and so just looking at the

23:25 Speaker 2: corresponding estimates of gain from obi Wan on both so

23:31 Speaker 2: on the on the right side, those are those are

23:35 Speaker 2: populations that are gaining biomass faster than they were before.

23:38 Speaker 2: And on the left, those are populations that we're that

23:41 Speaker 2: are losing biomass faster than they were before, and the

23:46 Speaker 2: fi A calibration, the red series does really pretty well

23:49 Speaker 2: across the whole the whole range, whereas uh well, actually

23:54 Speaker 2: the global calibration does better with this additionality challenge than

23:58 Speaker 2: than it did with the individual two points in time question.

24:02 Speaker 2: So that is sort of the validation that we have

24:06 Speaker 2: pursued in Nepal, in and the United States. I'm going

24:11 Speaker 2: to jump into API Application programming interface. I should have

24:17 Speaker 2: defined that, but basically it is a set of functions

24:22 Speaker 2: that we provide that you can use to access the

24:27 Speaker 2: time the bootstrapped time series that we're storing on Google

24:31 Speaker 2: Earth Engine. Again, Google has has donated substantial amount of

24:37 Speaker 2: storage for.

24:38 Speaker 3: Us to store that.

24:39 Speaker 2: And what the obi wan API does is it just

24:43 Speaker 2: gives you a bunch of functions that you might use

24:45 Speaker 2: to do very common things involved with reporting and verifying

24:51 Speaker 2: carbon changes in the place that you care about.

24:54 Speaker 3: So the things that.

24:54 Speaker 2: You might specify with the with the obi wan ap

24:58 Speaker 2: API is what years that you can about, how you

25:01 Speaker 2: want to visualize those changes, what kind of authentication you

25:05 Speaker 2: want to build into accessing whatever it is you're putting online.

25:08 Speaker 2: You can define it a baseline additionality baseline, whether it

25:12 Speaker 2: be a different point in time or a different area,

25:15 Speaker 2: maybe a surrounding area or a buffer area. And also

25:18 Speaker 2: it allows you to you know, store whatever reports and

25:22 Speaker 2: whatever format you want. So really the API is maximizing

25:27 Speaker 2: your flexibility. So I guess the question is why are

25:31 Speaker 2: we doing this instead of making a map? You know,

25:34 Speaker 2: the map is easy to distribute. But and this is

25:38 Speaker 2: definitely a sort of new paradigm for accessing space based

25:43 Speaker 2: biomass data. But I think there are good reasons for

25:47 Speaker 2: what we're doing. The first is we can't predict the

25:51 Speaker 2: years and the places that you care about, so allowing

25:55 Speaker 2: you to define those things it is obviously targets the stuff.

26:01 Speaker 3: To exactly what you need.

26:04 Speaker 2: Also, across market and treaty based commitment mechanisms, some users

26:09 Speaker 2: need to show things very transparently, perhaps to donors or

26:14 Speaker 2: to other countries, whereas others I'm thinking of private industry

26:18 Speaker 2: in this country, for instance, that maybe participating in carbon

26:22 Speaker 2: offset markets. Much of this is considered proprietary and they

26:26 Speaker 2: need to control access. So, you know, using the API

26:30 Speaker 2: to develop your own application really gives you the power

26:34 Speaker 2: to build in you know, whatever you need. It's easy

26:38 Speaker 2: to standardize across affiliated clients, maybe all participate participants in

26:45 Speaker 2: a particular carbon market, or perhaps different jurisdictions supported by

26:50 Speaker 2: the same donors. So one could picture, you know, developer

26:56 Speaker 2: hired by a particular donor developing dashboards for lots of

27:00 Speaker 2: different clients in a way that is consistent. Also, we're

27:05 Speaker 2: pre computing these bootstrapped time series because if if when

27:11 Speaker 2: you made your query and you had to wait for

27:15 Speaker 2: all the computing involved with bootstrapping to happen, it would

27:18 Speaker 2: you know, it would take hours. It may take hours.

27:20 Speaker 2: So we've pre computed these things, stored them on Google

27:24 Speaker 2: and that makes it much faster. And lastly, like we

27:28 Speaker 2: do it this way because you know, you may not

27:30 Speaker 2: be the only person who cares about you know, your country,

27:34 Speaker 2: so we're reducing a lot of redundant computing.

27:38 Speaker 3: So this is really sort of an overview of the structure.

27:42 Speaker 2: We do the bootstrapping, we store it thanks to Google

27:47 Speaker 2: on Google Cloud. We're providing the API, which we will

27:51 Speaker 2: get a look at in a moment in our hands

27:54 Speaker 2: on activity. And then you and you're the developers that

27:59 Speaker 2: you pay probably develop your own dashboards, accounting systems, reporting systems,

28:06 Speaker 2: building in whatever considerations are important to you.

28:09 Speaker 3: So the key is this API. And it sounds harder

28:13 Speaker 3: than it is.

28:14 Speaker 2: API is really pretty well documented, is really very simple.

28:17 Speaker 2: It's just a bunch of commands with you know, syntax,

28:22 Speaker 2: the divide that defines important variables for these commands, that

28:27 Speaker 2: queries the stored asset on Google Cloud and gives you

28:31 Speaker 2: what you need, and you will see those You'll see

28:35 Speaker 2: some of the calls to the API in the notebook

28:38 Speaker 2: we are about to look at. These resources are some

28:43 Speaker 2: background about the theory of our estimation methods, about some

28:49 Speaker 2: context for carbon markets. Okay, so before we dive into

28:54 Speaker 2: the hands on activity, I just want to tell you

28:57 Speaker 2: a little bit about what we're doing.

28:59 Speaker 3: We're going to be used a collab notebook.

29:02 Speaker 2: That's a Google collab notebook that calls an asset that

29:05 Speaker 2: we made for this training over Alabama State.

29:08 Speaker 3: In the United States.

29:10 Speaker 2: It's ten bootstraps deep, which is not deep and operationally

29:15 Speaker 2: we would probably want one hundred. But it will give

29:19 Speaker 2: you the idea of what's happening here. It will allow

29:22 Speaker 2: you to experiment with different options and different visualizations, and

29:27 Speaker 2: it gives you actual syntax like how the API is

29:31 Speaker 2: you So you could give this to your developer friend,

29:35 Speaker 2: and your developer friend could really practically build a dashboard

29:40 Speaker 2: pretty easily for you for the area that you care about.

29:44 Speaker 2: So what you're seeing is the link to the collab notebook.

29:49 Speaker 2: Before you start playing around. I will give you some

29:53 Speaker 2: further directions in just a moment.

29:55 Speaker 3: But if you could click.

29:56 Speaker 2: On this link, Google may ask you, if you've never

30:00 Speaker 2: or used a collab notebook, if you want it to

30:03 Speaker 2: compile the notebook for you. Please say yes and then

30:08 Speaker 2: stand by for more instructions. Okay, so if you clicked

30:18 Speaker 2: on that link, what you probably see is this, I

30:22 Speaker 2: want to point to your direct Your attention to this

30:25 Speaker 2: cannot save changes. What I'm going to ask you to

30:28 Speaker 2: do is go to file and save a copy to drive.

30:35 Speaker 3: That way what you're working on.

30:37 Speaker 2: First of all, you will be able to save it,

30:39 Speaker 2: and a lot of other people on the same training

30:42 Speaker 2: won't be, you know, disturbing what you're doing. That will

30:46 Speaker 2: be unsatisfactory for everyone. So once you save a copy

30:50 Speaker 2: to drive, what you will see is copy of this.

30:59 Speaker 2: This is a Python notebook using Google's collab and and

31:05 Speaker 2: this is the extension for that kind of file. So

31:08 Speaker 2: so really what you're going to be seeing is a

31:10 Speaker 2: lot of Python code that's calling the a p I

31:15 Speaker 2: and visualizing a lot of the estimates that the API enables. So, uh,

31:22 Speaker 2: you know, the first thing I guess I would point

31:25 Speaker 2: to is there is some you know.

31:29 Speaker 3: Annotation here.

31:31 Speaker 2: But but basically, if all you want to do is

31:34 Speaker 2: see what it does, you can just go down through

31:39 Speaker 2: and press these these triangles and basically that runs the

31:44 Speaker 2: code that is hidden you. But you can see the

31:46 Speaker 2: code here this first this first box is just setting

31:51 Speaker 2: up some variables that are calling.

31:54 Speaker 3: The obi wan end point the API.

31:58 Speaker 2: Okay, so a lot of this is really sort of generic,

32:01 Speaker 2: but it will help. I mean, I really feel like

32:04 Speaker 2: with this code you're probably fifty percent of the way

32:07 Speaker 2: towards making you know your own dashboard for the area

32:10 Speaker 2: that you care about.

32:12 Speaker 3: So all of this is sort of generic.

32:14 Speaker 2: And right now this is pointing to this Alabama study

32:20 Speaker 2: area that we've made for this training. But soon enough

32:25 Speaker 2: we will have at least an asset, we will have

32:29 Speaker 2: bootstraps for the entire United States. And when you do that,

32:36 Speaker 2: when we do that, you will only have to change this,

32:41 Speaker 2: you know, this address. So once we have a national

32:48 Speaker 2: asset and an international global asset for all of the

32:51 Speaker 2: bootstraps available.

32:53 Speaker 3: This code will really this is one line of code

32:56 Speaker 3: to change the actual address of that asset.

33:01 Speaker 2: All of the rest of this code, this Python code,

33:05 Speaker 2: will be applicable. So and right now you know this

33:09 Speaker 2: the Python code is calling this visual this viewer, and

33:16 Speaker 2: you know we can turn this the asset off and on.

33:20 Speaker 3: You can do whatever you want.

33:22 Speaker 2: I mean, you can visualize it with whatever imagery or

33:24 Speaker 2: perhaps company records behind it. One thing I would point

33:32 Speaker 2: out right now, what we're displaying is the mean of

33:36 Speaker 2: the biomass predictions across ten bootstraps. I mentioned that ten

33:41 Speaker 2: bootstraps is not a big number, and here's a ten

33:45 Speaker 2: kilometer tile where the the bootstraps obviously simulated errors across

33:54 Speaker 2: all ten that happened to give us very high biomass values.

33:57 Speaker 3: The color ramp here is is biomass.

34:00 Speaker 2: So there's nothing squareish under underneath there. You can turn

34:05 Speaker 2: that on and off up in the upper right hand corner.

34:09 Speaker 2: But basically, you know, what we're seeing is one way

34:13 Speaker 2: to visualize the asset. Okay, so again this could be

34:20 Speaker 2: any visualization you want, but in this case, what we're

34:25 Speaker 2: showing here is you pick the area that you care about. Again,

34:32 Speaker 2: this is in the middle of Alabama. I sort of

34:38 Speaker 2: arbitrarily drew a polygon in your own in your own development,

34:45 Speaker 2: you would probably import a KML.

34:49 Speaker 3: Or a shape file for the place that you that

34:51 Speaker 3: you care about.

34:53 Speaker 2: Okay, so we selected an area of interest, and now

35:00 Speaker 2: I'm going to let's say we care about twenty nineteen.

35:05 Speaker 2: So we for that area of interest, which is thirty

35:10 Speaker 2: seven thousand hectares, the estimate is one hundred and thirty

35:14 Speaker 2: six tons per hectare. That's the units here for biomass.

35:19 Speaker 2: That's without calibration. Let's try with calibration, see if that

35:25 Speaker 2: changes it from one hundred and thirty six.

35:32 Speaker 3: It did it.

35:32 Speaker 2: You know, it's a fairly high value, and it made

35:34 Speaker 2: it higher in general. So what we see the standard

35:38 Speaker 2: deviation for that estimate. Again, that's the precision, that's not

35:42 Speaker 2: necessarily the overall accuracy. Okay, So you can you can

35:49 Speaker 2: toggle this on and off to see the effects of calibration.

35:55 Speaker 3: Okay.

35:55 Speaker 2: So here's here's another thing that you might want to

35:57 Speaker 2: do for the area that you care about. Let's look at.

36:04 Speaker 3: The trajectory.

36:08 Speaker 2: Of this area and basically what you will see is

36:14 Speaker 2: I mean, I just drew that fairly arbitrarily, but we

36:18 Speaker 2: saw a gain up until around twenty thirteen, and then

36:23 Speaker 2: a loss. It is a fairly actively It looked like

36:27 Speaker 2: a fairly actively managed area of again thirty seven thousand hectares, okay. So,

36:38 Speaker 2: and I find that looking at the estimates for different

36:41 Speaker 2: points in time really sort of gives a lot of

36:44 Speaker 2: insight into the overall trends for the area that you

36:46 Speaker 2: care about. So, you know, estimating a change with a

36:53 Speaker 2: standard standard error, let's just press this button here for

37:00 Speaker 2: I think it's from the beginning of the time series

37:03 Speaker 2: twenty nine or nineteen ninety nine to twenty twenty three,

37:07 Speaker 2: the estimated change is a gain of ten. If we

37:12 Speaker 2: only care about the end points the you know, the

37:15 Speaker 2: gain was about ten. If we had to find that

37:18 Speaker 2: as nineteen ninety nine to twenty thirteen, it would have

37:21 Speaker 2: been a big bigger gain, and from twenty thirteen to

37:24 Speaker 2: twenty twenty three.

37:25 Speaker 3: It would have been obviously a loss.

37:28 Speaker 2: So that's why looking at it at the mean estimates

37:32 Speaker 2: across time first is pretty is pretty valuable. I mentioned

37:37 Speaker 2: that there's this counterfactual additionality where you where you apply, where.

37:43 Speaker 3: You look at the changes in your area.

37:45 Speaker 2: Let's say it's a park compared to maybe a buffer

37:48 Speaker 2: zone that is not managed under your management plan.

37:52 Speaker 3: Uh, and is not you know, the.

37:55 Speaker 2: The area that you're committing removals from the atmosphere to.

38:02 Speaker 2: So we're going to skip that. It works, but you

38:08 Speaker 2: have to be able to trust your comparisons if you

38:12 Speaker 2: want to make an inference of like this, the differences

38:16 Speaker 2: that you're seeing are actually a good indication of the

38:21 Speaker 2: impact of your management. So we're instead going to skip

38:25 Speaker 2: to looking at changes in the change, changes in biomass

38:31 Speaker 2: change changes in the rate of biomass accumulation.

38:38 Speaker 3: So we press that.

38:41 Speaker 2: And we're looking at the difference between Let's say that

38:45 Speaker 2: this is our baseline nineteen ninety nine to two thousand

38:48 Speaker 2: and five in the place in the target is twenty

38:51 Speaker 2: fifteen to twenty twenty one. So you will see here

39:03 Speaker 2: in the outputs that these are These are indeed the

39:06 Speaker 2: play that the times that we care about.

39:10 Speaker 3: You see the area of this place. This is the

39:14 Speaker 3: change in the annual rate of gain. So we're.

39:20 Speaker 2: Compared to the baseline time period, we are gaining two

39:25 Speaker 2: tons per hectare per year less in the target period,

39:31 Speaker 2: which is twenty fifteen to twenty one. Okay, and when

39:35 Speaker 2: we have across bootstraps, this is the uncertainty that the

39:40 Speaker 2: answer across bootstraps implies.

39:44 Speaker 3: So I guess I would.

39:46 Speaker 2: Say that, you know, this is a demon This Python

39:49 Speaker 2: notebook is a demonstration, but it's also you know, a

39:53 Speaker 2: lot of a lot of Python code anyway, that gives

39:57 Speaker 2: you good examples about how you could make your own

40:01 Speaker 2: application dashboard, or at least it gives a lot of

40:07 Speaker 2: examples of calling the API for things that you care about.

40:12 Speaker 2: So I think that after a brief message, I will

40:16 Speaker 2: be on for answering your questions. I look forward to them.

40:22 Speaker 2: Have fun with the notebook.

40:23 Speaker 1: Thanks, thank you very much doctor Heally for that great

40:33 Speaker 1: presentation and demonstration. Now let's do a quick summary of

40:38 Speaker 1: the most important points discussed during this presentation. Today we

40:44 Speaker 1: learned that obi wan can address climate change mitigation across

40:47 Speaker 1: different sectors by identifying management practice that reduce impacts. It's

40:53 Speaker 1: important to understand that uncertainty for carbon estimates is critical

40:58 Speaker 1: and that concept of additional is how management practices contribute

41:03 Speaker 1: to carbon storage trends. So what would happen if you

41:06 Speaker 1: were not doing what you were doing. Obie uses hierarchical

41:12 Speaker 1: bootstrapping to get a distribution of predicted changes and optical

41:16 Speaker 1: imagery from sensors like lansat do not see through the

41:20 Speaker 1: canopy and are likely to underpredict high biomass while overpredicting

41:25 Speaker 1: low biomass for a single point in time, as well

41:28 Speaker 1: as underpredict changes between high and low biomass conditions. And finally,

41:34 Speaker 1: calibration is a simple correction based on independent data and

41:38 Speaker 1: applied to the predictions to minimize bias. There is one

41:45 Speaker 1: homework associated with this training series and the homework can

41:51 Speaker 1: be accessed through the training website as of today, the

41:55 Speaker 1: due date is June eighteenth, and there will be a

41:58 Speaker 1: certificate of completion to participants who attend both live webinars

42:03 Speaker 1: and complete the homework by the deadline. Participants will receive

42:08 Speaker 1: a certificate via email approximately two months after completion of

42:13 Speaker 1: the course. If you have any questions about the material

42:18 Speaker 1: that has been presented, feel free to reach out to

42:21 Speaker 1: doctor Sean Healey through the contact information in this slide.

42:28 Speaker 1: We've reached the end of today's session and this training series.

42:31 Speaker 1: I'd like to thank doctor Sean Healey for his excellent presentations.

42:37 Speaker 1: I'd like to thank all of the participants for their

42:40 Speaker 1: interest in this topic and all of the great questions

42:43 Speaker 1: that have been coming in. We will now start our

42:46 Speaker 1: Q and A session.

42:51 Speaker 4: Thank you Erica for the SUMMARYE and wrap up. My

42:57 Speaker 4: name is Savannah Cooley and I am and our set

43:02 Speaker 4: trainer with the Ecological Conservation Program and I will be

43:06 Speaker 4: facilitating our Q and A session today. With that, we

43:12 Speaker 4: can go to the Question and Answers document where we

43:16 Speaker 4: have been populating all of the excellent questions that have

43:19 Speaker 4: been coming in. And let's go ahead and dive into

43:27 Speaker 4: question one, which asks JEDI and obi wan seem very

43:34 Speaker 4: useful for estimating biomass change, but high stakes carbon decisions

43:38 Speaker 4: requires strong evidence of additionality, uncertainty, control, and validation. In

43:43 Speaker 4: your view, what is the current practical boundary between using

43:48 Speaker 4: obi wan for scientific monitoring and using it for operational

43:52 Speaker 4: carbon credit or policy verification. So with that, i'll pass

43:57 Speaker 4: to you Sean. You've started your answer here all and

44:00 Speaker 4: I'll let you elaborate.

44:03 Speaker 3: Thanks Savannah.

44:04 Speaker 2: Yeah, this is really sort of a fundamental question, isn't it.

44:09 Speaker 3: You know, when.

44:12 Speaker 2: Authorities issue carbon credits, and when countries use particular methods

44:18 Speaker 2: for accounting for the additionality of their forests their forest management.

44:25 Speaker 2: You know, there's always a decision, Uh, you know, what

44:28 Speaker 2: is reliable enough, what is what meets the standard of

44:33 Speaker 2: of what can be used?

44:34 Speaker 3: What's you know, what's the best available?

44:37 Speaker 2: I think in most many cases, uh, obi wan represents

44:41 Speaker 2: the best available remotely sensed answer. Uh, It's not as

44:45 Speaker 2: good as having a longitudinal field survey, but that's not

44:49 Speaker 2: available in many places. So I mean we see a

44:52 Speaker 2: lot of people, uh, you know, countries, especially projects using

44:58 Speaker 2: this emission factor at tivity data paradigm that you know,

45:04 Speaker 2: that relies on a statistical framework, just like obi WAN does.

45:08 Speaker 3: I believe that obi wan.

45:13 Speaker 2: Accounts for more factors than that it accounts for changing

45:19 Speaker 2: changing emission factors within the same class through time. The

45:25 Speaker 2: short answer is, uh, you know, the whether obi wan

45:30 Speaker 2: or you know, the activity data emission factor paradigm or

45:33 Speaker 2: anything else is acceptable is up to you know, particular creditors.

45:40 Speaker 2: My personal opinion about obi wan as some of the

45:44 Speaker 2: some of the validation activity sort of suggested. I think

45:48 Speaker 2: it's good outside of the United States. The calibration with

45:51 Speaker 2: independent JEDI data is very useful and it removes a

45:57 Speaker 2: lot of the biases that we see, but it it's

46:01 Speaker 2: not it also introduces biases and in some cases we're

46:06 Speaker 2: over predicting losses, which you know, that's that's the primary problem.

46:10 Speaker 3: So I'm not sure that it is appropriate for.

46:17 Speaker 2: You know, UH funding a critical decisions and monitoring outside

46:26 Speaker 2: of the United States yet, although I do think that

46:28 Speaker 2: it probably represents the best remotely sensed answer anybody is

46:32 Speaker 2: going to give you. So I think it's appropriate for science,

46:37 Speaker 2: but maybe not for for the offset markets in the

46:40 Speaker 2: United States, where we have systematic collection and availability of

46:47 Speaker 2: change calibration data. I think our validation activity suggested that. Yeah,

46:53 Speaker 2: I think it is approaching that level where you you

46:57 Speaker 2: might base funding decisions on that, crediting decisions, you know.

47:02 Speaker 2: And again it's it's not as good as having field

47:05 Speaker 2: plots in your forest, you know, going back in time.

47:09 Speaker 2: But in a lot of cases that that kind of

47:11 Speaker 2: field effort UH is just not feasible given the amounts

47:17 Speaker 2: of money that are involved. So that I think that's

47:20 Speaker 2: a fundamental question, which is why I spend a lot

47:23 Speaker 2: of time on it, and I've.

47:24 Speaker 3: Thought about it a lot too.

47:26 Speaker 2: You know that the statistical framework that that we talked

47:31 Speaker 2: about is really you know, everything that we're doing is

47:34 Speaker 2: geared towards making this more acceptable to countries when they

47:41 Speaker 2: when they choose their method for reporting and for carbon standards,

47:47 Speaker 2: when they choose their methods for for monitoring the credits

47:51 Speaker 2: that they issue. So that's that's a long answer to

47:55 Speaker 2: a good question.

47:55 Speaker 4: I think, thank you so much Sean for the thoroughness

47:59 Speaker 4: of your answer and really going into some of the

48:02 Speaker 4: nuances and the key considerations. And I think, yeah, what

48:07 Speaker 4: you emphasized around the importance of the knowing the uncertainties

48:16 Speaker 4: and ultimately that it will be the creditors that be

48:21 Speaker 4: able to decide that existing information, you know, and also

48:27 Speaker 4: the high costs of field based methods of doing these

48:31 Speaker 4: kinds of analyzes that obi WAN still despite having uncertainties,

48:37 Speaker 4: also represents the best information that we know of available

48:41 Speaker 4: and from a remotely sense standpoint. So thank you so

48:44 Speaker 4: much for that thorough answer, and we will move to

48:47 Speaker 4: question two. Can the integration of Jedi derived biomass structure

48:53 Speaker 4: lansat SLASH sentinel landcover history and SIF which is solar

48:57 Speaker 4: induced fluorescence based photosynthetic activity improve estimates of watershed scale

49:03 Speaker 4: carbon capture and biomass change dynamics under climate and disturbance pressures.

49:11 Speaker 2: Well, yeah, that's a good question. I In general, the

49:16 Speaker 2: answer that I wrote addresses, you know, what imagery we

49:22 Speaker 2: use in obi WAN. As I read the question more thoroughly,

49:25 Speaker 2: I'm not sure that's the question that's being asked, but

49:28 Speaker 2: I'll just to elaborate on that question. Yeah, the better

49:33 Speaker 2: the imagery is correlated with biomass, the better our answers

49:37 Speaker 2: are going to be. We use lance at because it's

49:40 Speaker 2: available systematically back backwards through time.

49:44 Speaker 3: In terms of understanding.

49:48 Speaker 2: Climate pressures, climate processes and stuff like that, obi WAN

49:52 Speaker 2: is not a it's not a process model, right, We're not.

49:57 Speaker 3: We are only monitoring, uh, past trends.

50:02 Speaker 2: We're not trying to extract processes that we might use

50:06 Speaker 2: to extrapolate under different climates and things like that.

50:10 Speaker 1: Uh.

50:11 Speaker 2: In terms of disturbance pressures, yes, certainly we're uh. I

50:14 Speaker 2: mean in a retrospective way. We're allowing the quantification of

50:21 Speaker 2: biomass change in a particular area that may be subject

50:25 Speaker 2: to or or protected from developed disturbance processes related to development,

50:31 Speaker 2: and it facilitates comparison of the additionality there. So yeah,

50:39 Speaker 2: I think that's yes, Yes for disturbance processes, no for.

50:43 Speaker 3: For climate processes.

50:46 Speaker 4: Thank you. Question three asks can the obi wan API

50:52 Speaker 4: be used for urban prediction? Uh?

50:56 Speaker 2: The short answer is I wouldn't. Buildings often full Jedi waveforms.

51:03 Speaker 2: They can look like trees, and so I frankly just

51:08 Speaker 2: don't trust our predictions over urban areas, just because that

51:14 Speaker 2: that sort of fundamental error will propagate its way, and

51:18 Speaker 2: we haven't tested it, but I don't think it's probably

51:22 Speaker 2: it's probably not the most appropriate thing to track urban

51:24 Speaker 2: forests in particular.

51:27 Speaker 4: Right, even if even if an application I imagine we're

51:29 Speaker 4: to have a very accurate map of building footprints that

51:36 Speaker 4: they could mask out, there's still the consideration of the

51:39 Speaker 4: Jedi geolocation error, right which would which would.

51:44 Speaker 3: That's right?

51:44 Speaker 2: Yeah, yeah, And just that's that's up to about ten

51:49 Speaker 2: one sigma is ten meters, so you know, sixty seven

51:54 Speaker 2: percent of the footprints might be ten meters away, and

51:59 Speaker 2: you know some might be considerably farther away than the

52:04 Speaker 2: coordinates say they.

52:04 Speaker 3: Are, yeh.

52:07 Speaker 4: Question four asks how do uncertainty approximation approaches differ between

52:14 Speaker 4: m d N based probabilistic modeling and obi WAN based

52:18 Speaker 4: biomass change estimation, and can these approaches be integrated to

52:22 Speaker 4: improve watershed scale carbon and ecosystem decision support.

52:27 Speaker 3: So I don't even know what MDN stands for.

52:31 Speaker 4: I don't either. If actually, if someone who asks this

52:37 Speaker 4: question wants to clarify that part in the chat, pause

52:44 Speaker 4: a moment or circle back to this question.

52:49 Speaker 2: I mean to answer the question about watersheds. I think

52:52 Speaker 2: obi wan is appropriate for watershed carbon decision support.

52:59 Speaker 3: You know that we're where as.

53:00 Speaker 2: I said in the presentation, uh, talking about larger areas

53:04 Speaker 2: over a few hundred hectares, and that certainly is most

53:07 Speaker 2: watershed scales. So I would feel comfortable using obi wan

53:13 Speaker 2: to to evaluate carbon trends in particular watershed, as long

53:19 Speaker 2: as it's not you know, the really fine fine scale

53:22 Speaker 2: of huck levels.

53:27 Speaker 4: Thanks. Question five is a question about the black stripe

53:32 Speaker 4: line in slide eleven.

53:36 Speaker 2: Yeah, those those were the Those were the figures that

53:39 Speaker 2: just have field measured change on the X axis and

53:45 Speaker 2: predicted change on the Y axis. And the closer that

53:49 Speaker 2: that that black line is the one to one line.

53:52 Speaker 3: So the closer.

53:55 Speaker 2: A particular population is estimated to the to the real change,

53:59 Speaker 2: the closer it will be to that black line.

54:02 Speaker 3: So that's just there for reference.

54:07 Speaker 4: Question six asks, I know that over steep slopes, the

54:11 Speaker 4: lighter waveform tends to stretch. This can lead to overestimating

54:15 Speaker 4: canopy height. Do the current versions of relative hype metrics

54:19 Speaker 4: such as rge on one hundred and R ninety eight

54:22 Speaker 4: in the level two A product include any slope correction

54:26 Speaker 4: algorithms or have they successfully mitigated this slope stretching effect.

54:33 Speaker 2: Yes, so, I mean, if you think about the waveforms

54:36 Speaker 2: we looked at last week, if the level of the

54:41 Speaker 2: I mean, it's based upon the elevation of the ground

54:46 Speaker 2: that's found. But if you think about, you know, the

54:49 Speaker 2: circle that is a footprint. If that ground elevation is

54:53 Speaker 2: much higher on one half one side of the circle

54:55 Speaker 2: than the other, that's going to introduce a bunch of

54:59 Speaker 2: noise to that waveform. And that's why the Jedi mission

55:04 Speaker 2: is using a much smaller footprint, so you know, minimizing

55:08 Speaker 2: the potential for that ground level to change considerably. But

55:13 Speaker 2: it is twenty five meters and if you mean you

55:16 Speaker 2: can go out in the woods and walk twenty five

55:17 Speaker 2: meters and go down a slope, so you know there

55:21 Speaker 2: is noise there. And that's why Jedi provides waveforms basically

55:31 Speaker 2: the L two a relative height metrics, using a variety

55:37 Speaker 2: of ground finding algorithms specifically to address well. This issue

55:43 Speaker 2: is one of the main reasons why Jedi provides multiple

55:49 Speaker 2: versions of our age and some of them, I mean,

55:53 Speaker 2: the intention is that some of them are probably more

55:56 Speaker 2: locally representative than other. So I've included a link to

56:04 Speaker 2: all the documentation for those algorithms. So the short answer

56:12 Speaker 2: is we minimize it as much as we can, and

56:16 Speaker 2: we take a range of measures to to to find

56:24 Speaker 2: the ground appropriately, and you can find you can find

56:28 Speaker 2: different choices for that algorithm in the in the Jedi data.

56:33 Speaker 4: Yeah, and maybe I would just add that this is

56:36 Speaker 4: a fundamental issue with just how full waveform light our works,

56:42 Speaker 4: right where that we don't actually have a way to

56:45 Speaker 4: trace individual returns within a footprint in terms of where

56:50 Speaker 4: within that footprint they landed. And so there's no that

56:54 Speaker 4: that's a fundamental characteristic of just how waveform light our works.

57:00 Speaker 3: That's right, Yeah, it's well put.

57:03 Speaker 4: So in question seven, we have the following, could you

57:08 Speaker 4: share more applied cases of satellite data fusion, especially examples

57:13 Speaker 4: where multiple data sets such as Jedi, Lancet, Sentinel, solar

57:18 Speaker 4: inducescens or synthetic aperture radar are integrated to improve biomass,

57:23 Speaker 4: carbon or ecosystem monitoring.

57:26 Speaker 2: Yeah, I put I put a link to Laura duncanson

57:29 Speaker 2: at All's recent paper. I think it's accepted now, but

57:32 Speaker 2: I included the link for the free version. You know,

57:36 Speaker 2: that's that's fusion of ice at two, which is available

57:42 Speaker 2: outside of you know, above the latitudes that the Jedi

57:45 Speaker 2: is limited to, so the boreal mostly the boreal region

57:49 Speaker 2: that is not studied by Jedi.

57:51 Speaker 3: Ice at two does does see it?

57:53 Speaker 2: So that that paper and integrates fusion of Lancet and

57:58 Speaker 2: ice at two. There's also a lot of work with

58:03 Speaker 2: Jedi specifically that has followed fusion with.

58:11 Speaker 3: Terras are like tandem x uh so radar data.

58:17 Speaker 2: So you know it, and you know the Jedi asset

58:21 Speaker 2: is available on Earth Engine, so you know the the

58:24 Speaker 2: possibilities are almost limitless in terms of fusion through a

58:30 Speaker 2: number of different methods.

58:33 Speaker 4: Great, thank you, and and with the time here, this

58:36 Speaker 4: will be our last question for the live Q and A.

58:41 Speaker 4: There are more questions that have yet to be addressed,

58:44 Speaker 4: and we will make sure to have those addressed and

58:47 Speaker 4: published in the Q and a document that will be

58:50 Speaker 4: posted to our training website within about a week from

58:54 Speaker 4: from today. So let's end here with question eight. How

58:58 Speaker 4: is calibration droplated to every lansat pixel when using Jedi

59:03 Speaker 4: slash F I A.

59:06 Speaker 3: Yeah, and I'll start with my last sentence there.

59:08 Speaker 2: Obi wan is a tool for estimating biomass and biomass

59:11 Speaker 2: change over large areas. So the focus is not perfecting

59:15 Speaker 2: every prediction, it is removing the biases across the population.

59:22 Speaker 2: Because because we're focused on a population level parameter, the

59:25 Speaker 2: mean change or the mean biomass, So you know, in general,

59:31 Speaker 2: we use data from across the across the landscape to

59:35 Speaker 2: come up with a simple correction classical linear calibration function

59:40 Speaker 2: that we apply to all of the pixels. It doesn't

59:43 Speaker 2: necessarily make any particular pixel better, but it removes the

59:46 Speaker 2: biases across across the whole population. That's the intention anyway.

59:50 Speaker 2: So yeah, I guess the question is how is it extrapolated.

59:56 Speaker 2: The calibration is fitted us the large population and it's

1:00:01 Speaker 2: applied to all the pixels knowing that you know our

1:00:05 Speaker 2: target is the population parameter and not any particular pixels trajectory.

1:00:13 Speaker 4: Great, Well, that that's our time here for today. I

1:00:18 Speaker 4: want to thank you doctor Sean Healey for such a

1:00:23 Speaker 4: fantastic presentation and demo of obi Wan and think all

1:00:29 Speaker 4: participants for your interest and participation in today's training.

1:00:36 Speaker 2: Thank you Savannah, you guys have been great. This is

1:00:40 Speaker 2: really a well supported training I think so.

1:00:43 Speaker 3: Thank you, Thank you all,