General GEDI question

Use this Forum to find information on, or ask a question about, NASA Earth Science data.
Post Reply
ckhamila
Posts: 7
Joined: Mon Mar 13, 2023 9:30 am America/New_York
Answers: 0

General GEDI question

by ckhamila » Mon Mar 27, 2023 11:20 am America/New_York

1.) Can GEDI be used for time series analysis? I downloaded all gedi data available in my area of interest and my intention is to perform a time series analysis of vegetation, although, I have not come across any study n this use case.
2.) Also, most studies used gedi footprint wholly as downloaded from the usgis, I am wondering what happens in cases where points overlap, because in other sensor types, usually images are aggregated.
3.) Last question. Is temporal subsetting only done during data downloading or it can also be done at later stage.
I am just learning as I gather information and it is easy to get mixed up. Thank in advance for expert knowledge and a quick feedback.

Tags:

rupeshornl
Subject Matter Expert
Subject Matter Expert
Posts: 14
Joined: Mon Jun 28, 2021 10:37 am America/New_York
Answers: 0
Has thanked: 1 time

Re: General GEDI question

by rupeshornl » Tue Mar 28, 2023 8:18 am America/New_York

Hi @ckhamila, Thank you for posting your questions.

1) GEDI can be used for time series analysis if you find enough footprints in your study plots across time. Now, GEDI has collected billions of shots for more than three years. However, depending on the location of your study area, the density of available shots may vary.
2) Each GEDI shot is unique and not aggregated for the footprint-level products (L1B, L2A/B, L4A). The gridded GEDI products, such as L3 and L4B, are aggregated into 1km x 1km grids. Please refer to the products ATBDs and user guides for more information about aggregation.
3) Temporal subsetting can be done at a later stage as well. Each GEDI footprint has a "delta_time" variable which provides the time when the shot was recorded.

ckhamila
Posts: 7
Joined: Mon Mar 13, 2023 9:30 am America/New_York
Answers: 0

Re: General GEDI question

by ckhamila » Tue Mar 28, 2023 10:45 am America/New_York

Thank you very much for your quick feedback. Just 2 more questions:
1) I noticed that there are sensitivity values in my data with values more than 1 (i.e., 3, 5, 7, etc) within the same df with values of 0.3, 0.9, 0.95 etc, based on my understanding the maximum sensitivity ought to be 1 or 100%. I am wondering whether my understanding is wrong or if these are outliers.
2) Another thing I noticed is that the geolocation/degrade_flag has values 30, 70, 80 in some files, is it not that degrade_flag should have values of 0, 1 for good and poor data respectively?
Many thanks for your feedback.

rupeshornl
Subject Matter Expert
Subject Matter Expert
Posts: 14
Joined: Mon Jun 28, 2021 10:37 am America/New_York
Answers: 0
Has thanked: 1 time

Re: General GEDI question

by rupeshornl » Wed Mar 29, 2023 8:41 am America/New_York

1) A good quality shot has a "sensitivity" range from 0 to 1. A "valid_range" attribute for the "sensitivity" provides these values. GEDI uses a threshold of 0.9 over land and 0.5 over the ocean to compute quality flags.
2) Please refer to the user guide (Table 6) for details about these numbers.

ckhamila
Posts: 7
Joined: Mon Mar 13, 2023 9:30 am America/New_York
Answers: 0

Re: General GEDI question

by ckhamila » Wed Mar 29, 2023 9:52 am America/New_York

Thank you for your feedback. So, in short, the shots with sensitivity values beyond 1 (and 0s) and degrade flags 30, 70 etc (and 1s) are not valid? Please, see my screenshots below for your information.
Attachments
sensitivity.PNG
sensitivity.PNG (64.53 KiB) Not viewed yet
geolocation_degradeflag.PNG
geolocation_degradeflag.PNG (68.42 KiB) Not viewed yet

rupeshornl
Subject Matter Expert
Subject Matter Expert
Posts: 14
Joined: Mon Jun 28, 2021 10:37 am America/New_York
Answers: 0
Has thanked: 1 time

Re: General GEDI question

by rupeshornl » Wed Mar 29, 2023 3:40 pm America/New_York

Yes, that's correct. The default quality flags in the `quality_flag` variable might be easier to use.

Post Reply