Getting Started with GEDI L1B Data in Python

Tutorial Jupyter Notebook

1 Relevant Product

Details:

Published: May 13, 2020

⚠ Attention: The Global Ecosystem Dynamics Investigation (GEDI) Version 1 data products are no longer available for distribution from NASA’s Earthdata Search or LP DAAC’s Data Pool. Users are advised to use the most recent version of GEDI.

Getting Started with GEDI L1B Data in Python

This tutorial demonstrates how to work with the Geolocated Waveform (GEDI01_B.001) data product.

The Global Ecosystem Dynamics Investigation (GEDI) mission aims to characterize ecosystem structure and dynamics to enable radically improved quantification and understanding of the Earth's carbon cycle and biodiversity. The GEDI instrument produces high resolution laser ranging observations of the 3-dimensional structure of the Earth. GEDI is attached to the International Space Station and collects data globally between 51.6° N and 51.6° S latitudes at the highest resolution and densest sampling of any light detection and ranging (lidar) instrument in orbit to date. The Land Processes Distributed Active Archive Center (LP DAAC) distributes the GEDI Level 1 and Level 2 products. The L1B and L2 GEDI products are archived and distributed in the HDF-EOS5 file format.

Use Case Example:

This tutorial was developed using an example use case for a project being completed by the National Park Service. The goal of the project is to use GEDI L1B data to observe GEDI waveforms over Redwood National Park in northern California.

This tutorial will show how to use Python to open GEDI L1B files, visualize the full orbit of GEDI points (shots), subset to a region of interest, visualize GEDI full waveforms, and export subsets of GEDI science dataset (SDS) layers as GeoJSON files that can be loaded into GIS and/or Remote Sensing software programs.

Data Used in the Example:

  • GEDI L1B Geolocated Waveform Data Global Footprint Level - GEDI01_B.001
    • The purpose of the L1B dataset is to provide geolocated waveforms and supporting datasets for each laser shot for all eight GEDI beams. This includes corrected and smoothed waveforms, geolocation parameters, and geophysical corrections.
    • Science Dataset (SDS) layers:
      • /geolocation/latitude_bin0
      • /geolocation/longitude_bin0
      • /shot_number
      • /stale_return_flag
      • /geolocation/degrade
      • /rx_sample_count
      • /rx_sample_start_index
      • /rxwaveform
      • /geolocation/elevation_bin0
      • /geolocation/elevation_lastbin
      • /geolocation/digital_elevation_model

Topics Covered:

  1. Get Started
    1.1 Import Packages
    1.2 Set Up the Working Environment and Retrieve Files
  2. Import and Interpret Data
    2.1 Open a GEDI HDF5 File and Read File Metadata
    2.2 Read SDS Metadata and Subset by Beam
  3. Visualize a GEDI Orbit
    3.1 Subset by Layer and Create a Geodataframe
    3.2 Visualize a Geodataframe
  4. Subset and Visualize Waveforms
    4.1 Import and Extract Waveforms
    4.2 Visualize Waveforms
  5. Quality Filtering
  6. Plot Profile Transects
    6.1 Subset Beam Transects
    6.2 Plot Waveform Transects
  7. Spatial Visualization
    7.1 Import, Subset, and Quality Filter All Beams
    7.2 Spatial Subsetting
    7.3 Visualize All Beams: Elevation
  8. Export Subsets as GeoJSON Files

Before Starting this Tutorial:

Setup and Dependencies

It is recommended to use Conda, an environment manager to set up a compatible Python environment. Download Conda for your OS here: https://www.anaconda.com/download/. Once you have Conda installed, Follow the instructions below to successfully setup a Python environment on Linux, MacOS, or Windows.

This Python Jupyter Notebook tutorial has been tested using Python version 3.7. Conda was used to create the python environment.

  • Using your preferred command line interface (command prompt, terminal, cmder, etc.) type the following to successfully create a compatible python environment:

    conda create -n geditutorial -c conda-forge --yes python=3.7 h5py shapely geopandas pandas geoviews holoviews

conda activate geditutorial

jupyter notebook

If you do not have jupyter notebook installed, you may need to run:

conda install jupyter notebook

Having trouble getting a compatible Python environment set up? Contact LP DAAC User Services at: https://lpdaac.usgs.gov/lpdaac-contact-us/

If you prefer to not install Conda, the same setup and dependencies can be achieved by using another package manager such as pip.

Example Data:

This tutorial uses the GEDI L1B observation from June 19, 2019 (orbit 02932). Use the link below to download the file directly from the LP DAAC Data Pool:
- https://e4ftl01.cr.usgs.gov/GEDI/GEDI01_B.001/2019.06.19/GEDI01_B_2019170155833_O02932_T02267_02_003_01.h5 (7.87 GB)

A NASA Earthdata Login account is required to download the data used in this tutorial. You can create an account at the link provided.

You will need to have the file above downloaded into the same directory as this Jupyter Notebook in order to successfully run the code below.

Source Code used to Generate this Tutorial:

The repository containing all of the required files is located at: https://git.earthdata.nasa.gov/projects/LPDUR/repos/gedi-tutorials/browse
- Jupyter Notebook
- Redwood National Park GeoJSON
- Contains the administrative boundary for Redwood National Park, available from: Administrative Boundaries of National Park System Units 12/31/2017 - National Geospatial Data Asset (NGDA) NPS National Parks Dataset

NOTE: This tutorial was developed for GEDI L1B HDF-EOS5 files and should only be used for that product.

1. Get Started

1.1 Import Packages

Import the required packages and set the input/working directory to run this Jupyter Notebook locally.

import os
import h5py
import numpy as np
import pandas as pd
import geopandas as gp
from shapely.geometry import Point
import geoviews as gv
from geoviews import opts, tile_sources as gvts
import holoviews as hv
gv.extension('bokeh', 'matplotlib')

1.2 Set Up the Working Environment and Retrieve Files

The input directory is defined as the current working directory. Note that you will need to have the jupyter notebook and example data (.h5 and .geojson) stored in this directory in order to execute the tutorial successfully.

inDir = os.getcwd() + os.sep  # Set input directory to the current working directory
NOTE: If you have downloaded the tutorial materials to a different directory than the Jupyter Notebook, `inDir` above needs to be changed. You will also need to add a line: `os.chdir(inDir)` and execute it below.

In this section, the GEDI .h5 file has been downloaded to the inDir defined above. You will need to download the file directly from the LP DAAC Data Pool in order to execute this tutorial.

Direct Link to file:

  • https://e4ftl01.cr.usgs.gov/GEDI/GEDI01_B.001/2019.06.19/GEDI01_B_2019170155833_O02932_T02267_02_003_01.h5

Make sure to download the file into the inDir directory defined above.

gediFiles = [g for g in os.listdir() if g.startswith('GEDI01_B') and g.endswith('.h5')]  # List all GEDI L1B .h5 files in inDir
gediFiles

['GEDI01_B_2019170155833_O02932_T02267_02_003_01.h5']

2. Import and Interpret Data

2.1 Open a GEDI HDF5 File and Read File Metadata

Read the file using h5py.

L1B = 'GEDI01_B_2019170155833_O02932_T02267_02_003_01.h5'
L1B

'GEDI01_B_2019170155833_O02932_T02267_02_003_01.h5'

The standard format for GEDI filenames is as follows:

GEDI01_B: Product Short Name
2019170155833: Julian Date and Time of Acquisition (YYYYDDDHHMMSS)
O02932: Orbit Number
T02267: Track Number
02: Positioning and Pointing Determination System (PPDS) type (00 is predict, 01 rapid, 02 and higher is final)
003: GOC SDS (software) release number
01: Granule Production Version

Read in a GEDI HDF5 file using the h5py package.

gediL1B = h5py.File(L1B, 'r')  # Read file using h5py

Navigate the HDF5 file below.

list(gediL1B.keys())

['BEAM0000',
 'BEAM0001',
 'BEAM0010',
 'BEAM0011',
 'BEAM0101',
 'BEAM0110',
 'BEAM1000',
 'BEAM1011',
 'METADATA']

The GEDI HDF5 file contains groups in which data and metadata are stored.

First, the METADATA group contains the file-level metadata.

list(gediL1B['METADATA'])

['DatasetIdentification']

This contains useful information such as the creation date, PGEVersion, and VersionID. Below, print the file-level metadata attributes.

for g in gediL1B['METADATA']['DatasetIdentification'].attrs: print(g)

PGEVersion
VersionID
abstract
characterSet
creationDate
credit
fileName
language
originatorOrganizationName
purpose
shortName
spatialRepresentationType
status
topicCategory
uuid

print(gediL1B['METADATA']['DatasetIdentification'].attrs['purpose'])

The purpose of the L1B dataset is to provide geolocated waveforms and supporting datasets for each laser shot for all eight GEDI beams.  This includes corrected and smoothed waveforms, geolocation parameters, and geophysical corrections.

2.2 Read SDS Metadata and Subset by Beam

The GEDI instrument consists of 3 lasers producing a total of 8 beam ground transects. The eight remaining groups contain data for each of the eight GEDI beam transects. For additional information, be sure to check out: https://gedi.umd.edu/instrument/specifications/.

beamNames = [g for g in gediL1B.keys() if g.startswith('BEAM')]
beamNames

['BEAM0000',
 'BEAM0001',
 'BEAM0010',
 'BEAM0011',
 'BEAM0101',
 'BEAM0110',
 'BEAM1000',
 'BEAM1011']

One useful piece of metadata to retrieve from each beam transect is whether it is a full power beam or a coverage beam.

for g in gediL1B['BEAM0000'].attrs: print(g)

description
wp-l1a-to-l1b_githash
wp-l1a-to-l1b_version

for b in beamNames:
    print(f"{b} is a {gediL1B[b].attrs['description']}")

BEAM0000 is a Coverage beam
BEAM0001 is a Coverage beam
BEAM0010 is a Coverage beam
BEAM0011 is a Coverage beam
BEAM0101 is a Full power beam
BEAM0110 is a Full power beam
BEAM1000 is a Full power beam
BEAM1011 is a Full power beam

Below, pick one of the full power beams that will be used to retrieve GEDI L1B waveforms in Section 3.

beamNames = ['BEAM0110']

Identify all the objects in the GEDI HDF5 file below.

Note: This step may take a while to complete.

gediL1B_objs = []
gediL1B.visit(gediL1B_objs.append)                                           # Retrieve list of datasets
gediSDS = [o for o in gediL1B_objs if isinstance(gediL1B[o], h5py.Dataset)]  # Search for relevant SDS inside data file
[i for i in gediSDS if beamNames[0] in i][0:10]                              # Print the first 10 datasets for selected beam

['BEAM0110/all_samples_sum',
 'BEAM0110/ancillary/master_time_epoch',
 'BEAM0110/ancillary/mean_samples',
 'BEAM0110/ancillary/smoothing_width',
 'BEAM0110/beam',
 'BEAM0110/channel',
 'BEAM0110/geolocation/altitude_instrument',
 'BEAM0110/geolocation/altitude_instrument_error',
 'BEAM0110/geolocation/bounce_time_offset_bin0',
 'BEAM0110/geolocation/bounce_time_offset_bin0_error']

3. Visualize a GEDI Orbit

In the section below, import GEDI L1B SDS layers into a GeoPandas GeoDataFrame for the beam specified above.

Use the latitude_bin0 and longitude_bin0 to create a shapely point for each GEDI shot location.

3.1 Subset by Layer and Create a Geodataframe

Read in the SDS and take a representative sample (every 100th shot) and append to lists, then use the lists to generate a pandas dataframe.

lonSample, latSample, shotSample, srfSample, degradeSample, beamSample = [], [], [], [], [], []  # Set up lists to store data

# Open the SDS
lats = gediL1B[f'{beamNames[0]}/geolocation/latitude_bin0'][()]
lons = gediL1B[f'{beamNames[0]}/geolocation/longitude_bin0'][()]
shots = gediL1B[f'{beamNames[0]}/shot_number'][()]
srf = gediL1B[f'{beamNames[0]}/stale_return_flag'][()]
degrade = gediL1B[f'{beamNames[0]}/geolocation/degrade'][()]

# Take every 100th shot and append to list
for i in range(len(shots)):
    if i % 100 == 0:
        shotSample.append(str(shots[i]))
        lonSample.append(lons[i])
        latSample.append(lats[i])
        srfSample.append(srf[i])
        degradeSample.append(degrade[i])
        beamSample.append(beamNames[0])

# Write all of the sample shots to a dataframe
latslons = pd.DataFrame({'Beam': beamSample, 'Shot Number': shotSample, 'Longitude': lonSample, 'Latitude': latSample,
                         'Stale Return Flag': srfSample, 'Degrade': degradeSample})
latslons
Beam Shot Number Longitude Latitude Stale Return Flag Degrade
0 BEAM0110 29320618800000001 111.996300 -51.803868 1 0
1 BEAM0110 29320604600000101 112.039132 -51.803905 1 0
2 BEAM0110 29320614600000201 112.080271 -51.803836 1 0
3 BEAM0110 29320600400000301 112.121445 -51.803737 1 0
4 BEAM0110 29320610400000401 112.162622 -51.803621 1 0
... ... ... ... ... ... ...
9792 BEAM0110 29320617400979201 88.208452 -51.803578 1 0
9793 BEAM0110 29320603200979301 88.249610 -51.803614 1 0
9794 BEAM0110 29320613200979401 88.290753 -51.803581 1 0
9795 BEAM0110 29320623200979501 88.331913 -51.803548 1 0
9796 BEAM0110 29320609000979601 88.373089 -51.803506 1 0

9797 rows × 6 columns

Above is a dataframe containing columns describing the beam, shot number, lat/lon location, and quality information about each shot.

# Clean up variables that will no longer be needed
del beamSample, degrade, degradeSample, gediL1B_objs, latSample, lats, lonSample, lons, shotSample, shots, srf, srfSample

Below, create an additional column called 'geometry' that contains a shapely point generated from each lat/lon location from the shot.

# Take the lat/lon dataframe and convert each lat/lon to a shapely point
latslons['geometry'] = latslons.apply(lambda row: Point(row.Longitude, row.Latitude), axis=1)

Next, convert to a Geopandas GeoDataFrame.

# Convert to a Geodataframe
latslons = gp.GeoDataFrame(latslons)
latslons = latslons.drop(columns=['Latitude','Longitude'])
latslons['geometry']

0       POINT (111.99630 -51.80387)
1       POINT (112.03913 -51.80391)
2       POINT (112.08027 -51.80384)
3       POINT (112.12145 -51.80374)
4       POINT (112.16262 -51.80362)
                   ...             
9792     POINT (88.20845 -51.80358)
9793     POINT (88.24961 -51.80361)
9794     POINT (88.29075 -51.80358)
9795     POINT (88.33191 -51.80355)
9796     POINT (88.37309 -51.80351)
Name: geometry, Length: 9797, dtype: geometry

Pull out and plot an example shapely point below.

latslons['geometry'][0]

Single Shapely point.

3.2 Visualize a GeoDataFrame

In this section, use the GeoDataFrame and the geoviews python package to spatially visualize the location of the GEDI shots on a basemap and import a geojson file of the spatial region of interest for the use case example: Redwood National Park.

# Define a function for visualizing GEDI points
def pointVisual(features, vdims):
    return (gvts.EsriImagery * gv.Points(features, vdims=vdims).options(tools=['hover'], height=500, width=900, size=5,
                                                                        color='yellow', fontsize={'xticks': 10, 'yticks': 10,
                                                                                                  'xlabel':16, 'ylabel': 16}))

Import a geojson of Redwood National Park as an additional GeoDataFrame. Note that you will need to have downloaded the geojson from the bitbucket repo containing this tutorial and have it saved in the same directory as this Jupyter Notebook.

redwoodNP = gp.GeoDataFrame.from_file('RedwoodNP.geojson')  # Import geojson as GeoDataFrame

redwoodNP
GIS_LOC_ID UNIT_CODE GROUP_CODE UNIT_NAME UNIT_TYPE META_MIDF LANDS_CODE DATE_EDIT GIS_NOTES geometry
0 None REDW None Redwood National Park None None None Shifted 0.06 miles MULTIPOLYGON (((-124.01829 41.44539, -124.0184... 
redwoodNP['geometry'][0]  # Plot GeoDataFrame

Shapefile of Redwood National Park

Defining the vdims below will allow you to hover over specific shots and view information about them.

# Create a list of geodataframe columns to be included as attributes in the output map
vdims = []
for f in latslons:
    if f not in ['geometry']:
        vdims.append(f)
vdims

['Beam', 'Shot Number', 'Stale Return Flag', 'Degrade']

Below, combine a plot of the Redwood National Park Boundary (combine two geoviews plots using *) with the point visual mapping function defined above in order to plot (1) the representative GEDI shots, (2) the region of interest, and (3) a basemap layer.

# Call the function for plotting the GEDI points
gv.Polygons(redwoodNP['geometry']).opts(line_color='red', color=None) * pointVisual(latslons, vdims = vdims)

Global GEDI Orbit Map.

Above is a good illustration of the full GEDI orbit (GEDI files are stored as one ISS orbit). One of the benefits of using geoviews is the interactive nature of the output plots. Use the tools to the right of the map above to zoom in and find the shots intersecting Redwood National Park.

(HINT: find where the orbit intersects the west coast of the United States)

Below is a screenshot of the region of interest:

alt text

Side Note: Wondering what the 0's and 1's for stale_return_flag and degrade mean?

print(f"stale_return_flag: {gediL1B[b]['stale_return_flag'].attrs['description']}")
print(f"degrade: {gediL1B[b]['geolocation']['degrade'].attrs['description']}")

stale_return_flag: Indicates that a "stale" cue point from the coarse search algorithm is being used.
degrade: Greater than zero if the shot occurs during a degrade period, zero otherwise.

We will show an example of how to quality filter GEDI data in section 5.

After finding one of the shots within Redwood NP, find the index for that shot number so that we can find the correct waveform to visualize in Section 4.

Each GEDI shot has a unique shot identifier (shot number) that is available within each data group of the product. The shot number is important to retain in any data subsetting as it will allow the user to link any shot record back to the original orbit data, and to link any shot and its data between the L1 and L2 products. The standard format for GEDI Shots is as follows:

Shot: 29320619900465601

2932: Orbit Number
06: Beam Number
199: Minor frame number (0-241)
00465601: Shot number within orbit 

shot = 29320619900465601

index = np.where(gediL1B[f'{beamNames[0]}/shot_number'][()]==shot)[0][0]  # Set the index for the shot identified above
index

465600

del latslons  # No longer need the geodataframe used to visualize the full GEDI orbit

4. Subset and Visualize Waveforms

In this section, learn how to extract and subset specific waveforms and plot them using holoviews.

4.1 Import and Extract Waveforms

In order to find and extract the full waveform for the exact index we are interested in, instead of importing the entire waveform dataset (over 1 billion values!), we will use the rx_sample_count and rx_sample_start_index to identify the location of the waveform that we are interested in visualizing and just extract those values.

# From the SDS list, use list comprehension to find sample_count, sample_start_index, and rxwaveform
sdsCount = gediL1B[[g for g in gediSDS if g.endswith('/rx_sample_count') and beamNames[0] in g][0]]
sdsStart = gediL1B[[g for g in gediSDS if g.endswith('/rx_sample_start_index') and beamNames[0] in g][0]]
sdsWaveform = [g for g in gediSDS if g.endswith('/rxwaveform') and beamNames[0] in g][0]

Print the description for each of these datasets to better understand how we will use them to extract specific waveforms.

print(f"rxwaveform is {gediL1B[sdsWaveform].attrs['description']}")

rxwaveform is The corrected receive (RX) waveforms.  Use rx_sample_count and rx_sample_start_index to identify the location of each waveform.

Next, read how to use the rx_sample_count and rx_sample_start_index layers:

print(f"rx_sample_count is {sdsCount.attrs['description']}")
print(f"rx_sample_start_index is {sdsStart.attrs['description']}")

rx_sample_count is The number of sample intervals (elements) in each RX waveform.
rx_sample_start_index is The index in the rxwaveform dataset of the first element of each RX waveform.  The indices start at 1.

Use rx_sample_count and rx_sample_start_index to identify the location of each waveform in rxwaveform.

wfCount = sdsCount[index]           # Number of samples in the waveform
wfStart = int(sdsStart[index] - 1)  # Subtract one because python array indexing starts at 0 not 1

Next grab additional information about the shot, including the unique shot_number, and lat/lon location.

wfShot = gediL1B[f'{beamNames[0]}/shot_number'][index]
wfLat = gediL1B[f'{beamNames[0]}/geolocation/latitude_bin0'][index]
wfLon = gediL1B[f'{beamNames[0]}/geolocation/longitude_bin0'][index]

Put everything together to identify the waveform we want to extract:

print(f"The waveform located at: {str(wfLat)}, {str(wfLon)} (shot ID: {wfShot}, index {index}) is from beam {beamNames[0]} \
      and is stored in rxwaveform beginning at index {wfStart} and ending at index {wfStart + wfCount}")

The waveform located at: 41.28533109208681, -124.0300090357876 (shot ID: 29320619900465601, index 465600) is from beam BEAM0110       and is stored in rxwaveform beginning at index 661152000 and ending at index 661152812

In order to plot a full waveform, you also need to import the elevation recorded at bin0 (the first elevation recorded) and lastbin or the last elevation recorded for that waveform.

# Grab the elevation recorded at the start and end of the full waveform capture
zStart = gediL1B[f'{beamNames[0]}/geolocation/elevation_bin0'][index]   # Height of the start of the rx window
zEnd = gediL1B[f'{beamNames[0]}/geolocation/elevation_lastbin'][index]  # Height of the end of the rx window

Extract the full waveform using the index start and count:

Below you can see why it is important to extract the specific waveform that you are interested in: almost 1.4 billion values are stored in the rxwaveform dataset!

print("{:,}".format(gediL1B[sdsWaveform].shape[0]))

1,391,172,580

# Retrieve the waveform sds layer using the sample start index and sample count information to slice the correct dimensions
waveform = gediL1B[sdsWaveform][wfStart: wfStart + wfCount]

4.2 Visualize Waveforms

Below, plot the extracted waveform using the elevation difference from bin0 to last_bin on the y axis, and the waveform energy returned or amplitude (digital number = dn) on the x axis.

# Find elevation difference from start to finish and divide into equal intervals based on sample_count
zStretch = np.add(zEnd, np.multiply(range(wfCount, 0, -1), ((zStart - zEnd) / int(wfCount))))

# match the waveform amplitude values with the elevation and convert to Pandas df
wvDF = pd.DataFrame({'Amplitude (DN)': waveform, 'Elevation (m)': zStretch})
wvDF
Amplitude (DN) Elevation (m)
0 226.009323 37.228645
1 225.360245 37.078965
2 225.373764 36.929286
3 226.165710 36.779607
4 227.480103 36.629927
... ... ...
807 231.473923 -83.562517
808 230.386932 -83.712196
809 228.637512 -83.861876
810 226.949661 -84.011555
811 226.125153 -84.161234

812 rows × 2 columns

Next, use the Holoviews python package (hv) to plot the waveform.

hv.Curve(wvDF) # Basic line graph plotting the waveform

Plot of a waveform.

Congratulations! You have plotted your first waveform.

Above is a basic line plot showing amplitude (DN) as a function of elevation from the rxwaveform for the specific shot selected. Next, add additional chart elements and make the graph interactive.

# Create a holoviews interactive Curve plot with additional parameters defining the plot aesthetics
wfVis = hv.Curve(wvDF).opts(color='darkgreen', tools=['hover'], height=500, width=400,
           xlim=(np.min(waveform) - 10, np.max(waveform) + 10), ylim=(np.min(zStretch), np.max(zStretch)),
           fontsize={'xticks':10, 'yticks':10,'xlabel':16, 'ylabel': 16, 'title':13}, line_width=2.5, title=f'{str(wfShot)}')
wfVis

Plot of a waveform.

As you can see, the selected shot does not look so interesting--this waveform shape looks more characteristic of some type of low canopy/bare ground than a tree canopy. If you zoom in to the GEDI shot it appears to be over the river itself or possibly a sand bar:

alt text

Next, plot a couple more waveforms and see if you can capture a shot over the forest.

latlons = {}                          # Set up a dictionary to hold multiple waveforms
latlons[wfShot] = Point(wfLon,wfLat)  # Create shapely point and add to dictionary

# Retain waveform and quality layers to be exported later
latlonsWF = [waveform]
latlonsEL = [zStretch]
latlonsSRF = [gediL1B[f'{beamNames[0]}/stale_return_flag'][index]]
latlonsD = [gediL1B[f'{beamNames[0]}/geolocation/degrade'][index]]

# Define which observation to examine by selecting the three indexes prior to the one used above
index = 465597

# Use rx_sample_count and rx_sample_start_index to identify the location of each waveform
wfCount = sdsCount[index]                                             # Number of samples in the waveform
wfStart = int(sdsStart[index] - 1)                                    # Subtract 1, index starts at 0 not 1
wfShot = gediL1B[f'{beamNames[0]}/shot_number'][index]                # Unique Shot Number
wfLat = gediL1B[f'{beamNames[0]}/geolocation/latitude_bin0'][index]   # Latitude
wfLon = gediL1B[f'{beamNames[0]}/geolocation/longitude_bin0'][index]  # Longitude
latlons[wfShot] = Point(wfLon,wfLat)                                  # Create shapely point and add to dictionary

# Grab the elevation recorded at the start and end of the full waveform capture
zStart = gediL1B[f'{beamNames[0]}/geolocation/elevation_bin0'][index]   # Height of the start of the rx window
zEnd = gediL1B[f'{beamNames[0]}/geolocation/elevation_lastbin'][index]  # Height of the end of the rx window

# Retrieve the waveform sds layer using the sample start index and sample count information to slice the correct dimensions
waveform = gediL1B[sdsWaveform][wfStart: wfStart + wfCount]

# Find elevation difference from start to finish, divide into equal intervals
zStretch = np.add(zEnd, np.multiply(range(wfCount, 0, -1), ((zStart - zEnd) / int(wfCount))))

# match the waveform amplitude values with the elevation and convert to Pandas df
wvDF = pd.DataFrame({'Amplitude (DN)': waveform, 'Elevation (m)': zStretch})

# Append waveform and quality layers to be exported later
latlonsWF.append(waveform)
latlonsEL.append(zStretch)
latlonsSRF.append(gediL1B[f'{beamNames[0]}/stale_return_flag'][index])
latlonsD.append(gediL1B[f'{beamNames[0]}/geolocation/degrade'][index])

# Create a holoviews interactive Curve plot with additional parameters defining the plot aesthetics
visL1B1 = hv.Curve(wvDF).opts(color='darkgreen', tools=['hover'], height=500, width=400,
           xlim=(np.min(waveform) - 10, np.max(waveform) + 10), ylim=(np.min(zStretch), np.max(zStretch)),
           fontsize={'xticks':10, 'yticks':10,'xlabel':16, 'ylabel': 16, 'title':13}, line_width=2.5, title=f'{str(wfShot)}')
visL1B1

Plot of a waveform.

Now that is starting to look more like a very tall, multi-layered tree canopy! Continue with the next shot:

# Define which observation to examine by selecting the three indexes prior to the one used above
index = 465598

# Use rx_sample_count and rx_sample_start_index to identify the location of each waveform
wfCount = sdsCount[index]                                             # Number of samples in the waveform
wfStart = int(sdsStart[index] - 1)                                    # Subtract 1, index starts at 0 not 1
wfShot = gediL1B[f'{beamNames[0]}/shot_number'][index]                # Unique Shot Number
wfLat = gediL1B[f'{beamNames[0]}/geolocation/latitude_bin0'][index]   # Latitude
wfLon = gediL1B[f'{beamNames[0]}/geolocation/longitude_bin0'][index]  # Longitude
latlons[wfShot] = Point(wfLon,wfLat)                                  # Create shapely point and add to dictionary

# Grab the elevation recorded at the start and end of the full waveform capture
zStart = gediL1B[f'{beamNames[0]}/geolocation/elevation_bin0'][index]   # Height of the start of the rx window
zEnd = gediL1B[f'{beamNames[0]}/geolocation/elevation_lastbin'][index]  # Height of the end of the rx window

# Retrieve the waveform sds layer using the sample start index and sample count information to slice the correct dimensions
waveform = gediL1B[sdsWaveform][wfStart: wfStart + wfCount]

# Find elevation difference from start to finish, divide into equal intervals
zStretch = np.add(zEnd, np.multiply(range(wfCount, 0, -1), ((zStart - zEnd) / int(wfCount))))

# match the waveform amplitude values with the elevation and convert to Pandas df
wvDF = pd.DataFrame({'Amplitude (DN)': waveform, 'Elevation (m)': zStretch})

# Append waveform and quality layers to be exported later
latlonsWF.append(waveform)
latlonsEL.append(zStretch)
latlonsSRF.append(gediL1B[f'{beamNames[0]}/stale_return_flag'][index])
latlonsD.append(gediL1B[f'{beamNames[0]}/geolocation/degrade'][index])

# Create a holoviews interactive Curve plot with additional parameters defining the plot aesthetics
visL1B2 = hv.Curve(wvDF).opts(color='darkgreen', tools=['hover'], height=500, width=400,
           xlim=(np.min(waveform) - 10, np.max(waveform) + 10), ylim=(np.min(zStretch), np.max(zStretch)),
           fontsize={'xticks':10, 'yticks':10,'xlabel':16, 'ylabel': 16, 'title':13}, line_width=2.5, title=f'{str(wfShot)}')
visL1B2

Plot of a waveform.

# Define which observation to examine by selecting the three indexes prior to the one used above
index = 465599

# Use rx_sample_count and rx_sample_start_index to identify the location of each waveform
wfCount = sdsCount[index]                                             # Number of samples in the waveform
wfStart = int(sdsStart[index] - 1)                                    # Subtract 1, index starts at 0 not 1
wfShot = gediL1B[f'{beamNames[0]}/shot_number'][index]                # Unique Shot Number
wfLat = gediL1B[f'{beamNames[0]}/geolocation/latitude_bin0'][index]   # Latitude
wfLon = gediL1B[f'{beamNames[0]}/geolocation/longitude_bin0'][index]  # Longitude
latlons[wfShot] = Point(wfLon,wfLat)                                  # Create shapely point and add to dictionary

# Grab the elevation recorded at the start and end of the full waveform capture
zStart = gediL1B[f'{beamNames[0]}/geolocation/elevation_bin0'][index]   # Height of the start of the rx window
zEnd = gediL1B[f'{beamNames[0]}/geolocation/elevation_lastbin'][index]  # Height of the end of the rx window

# Retrieve the waveform sds layer using the sample start index and sample count information to slice the correct dimensions
waveform = gediL1B[sdsWaveform][wfStart: wfStart + wfCount]

# Find elevation difference from start to finish, divide into equal intervals
zStretch = np.add(zEnd, np.multiply(range(wfCount, 0, -1), ((zStart - zEnd) / int(wfCount))))

# match the waveform amplitude values with the elevation and convert to Pandas df
wvDF = pd.DataFrame({'Amplitude (DN)': waveform, 'Elevation (m)': zStretch})

# Append waveform and quality layers to be exported later
latlonsWF.append(waveform)
latlonsEL.append(zStretch)
latlonsSRF.append(gediL1B[f'{beamNames[0]}/stale_return_flag'][index])
latlonsD.append(gediL1B[f'{beamNames[0]}/geolocation/degrade'][index])

# Create a holoviews interactive Curve plot with additional parameters defining the plot aesthetics
visL1B3 = hv.Curve(wvDF).opts(color='darkgreen', tools=['hover'], height=500, width=400,
           xlim=(np.min(waveform) - 10, np.max(waveform) + 10), ylim=(np.min(zStretch), np.max(zStretch)),
           fontsize={'xticks':10, 'yticks':10,'xlabel':16, 'ylabel': 16, 'title':13}, line_width=2.5, title=f'{str(wfShot)}')
visL1B3

Plot of a waveform.

Lastly, plot the transect of four consecutive waveforms:

# The "+" symbol will plot multiple saved holoviews plots together
visL1B1.opts(width=240) + visL1B2.opts(width=240, labelled=[]) + visL1B3.opts(width=240, labelled=[]) + \
wfVis.opts(width=240, labelled=[])

Plot the transect of four consecutive waveforms.

Notice above moving west to east (left to right) along the transect you can see the elevation lowering and the tree canopy decreasing as it encounters the sandbar/river.

Below, use geoviews to plot the location of the four points to verify what is seen above.

# Convert dict to geodataframe
latlons = gp.GeoDataFrame({'Shot Number': list(latlons.keys()),'rxwaveform Amplitude (DN)': latlonsWF,
                           'rxwaveform Elevation (m)': latlonsEL, 'Stale Return Flag': latlonsSRF, 'Degrade': latlonsD,
                           'geometry': list(latlons.values())})

latlons['Shot Number'] = latlons['Shot Number'].astype(str)  # Convert shot number from integer to string

vdims

['Beam', 'Shot Number', 'Stale Return Flag', 'Degrade']

# Create a list of geodataframe columns to be included as attributes in the output map
vdims = []
for f in latlons:
    if f not in ['geometry']:
        if 'rxwaveform' not in f:
            vdims.append(f)

# Plot the geodataframe
gv.Polygons(redwoodNP['geometry']).opts(line_color='red', color=None) * pointVisual(latlons, vdims=vdims)

Four shots over Redwood National Park.

Above, zoom in on the yellow dots until you can get a clearer view of all four shots.

In the screenshots below, we can match the location of each point with the associated waveform. It appears to confirm the hypothesis that the elevation is decreasing as the shots get closer to the river, and the forest canopy decreases until it is detecting the sand bar/river by the final waveform in the sequence.

alt text alt text

del wfVis, visL1B1, visL1B2, visL1B3, waveform, wvDF, zStretch, latlonsWF, latlonsEL, latlonsSRF, latlonsD

5. Quality Filtering

Now that you have the desired layers imported as a dataframe for the selected shots, let's perform quality filtering.

Below, remove any shots where the stale_return_flag is set to 1 (indicates that a "stale" cue point from the coarse search algorithm is being used) by defining those shots as nan.

The syntax of the line below can be read as: in the dataframe, find the rows "where" the stale return flag is not equal (ne) to 0. If a row (shot) does not meet the condition, set all values equal to nan for that row.

latlons = latlons.where(latlons['Stale Return Flag'].ne(1))  # Set any stale returns to NaN

Below, quality filter even further by using the degrade flag (Greater than zero if the shot occurs during a degrade period, zero otherwise).

latlons = latlons.where(latlons['Degrade'].ne(1))

Below, drop all of the shots that did not pass the quality filtering standards outlined above from the transectDF.

latlons = latlons.dropna()  # Drop all of the rows (shots) that did not pass the quality filtering above

print(f"Quality filtering complete, {len(latlons)} high quality shots remaining.")

Quality filtering complete, 4 high quality shots remaining.

Good news! It looks like all four of the example waveforms passed the initial quality filtering tests. For additional information on quality filtering GEDI data, be sure to check out: https://lpdaac.usgs.gov/resources/faqs/#how-should-i-quality-filter-gedi-l1b-l2b-data.

del latlons

6. Plot Profile Transects

In this section, plot a transect subset using waveforms.

6.1 Subset Beam Transects

Subset down to a smaller transect centered on the waveforms analyzed in the sections above.

print(index)

465599

# Grab 50 points before and after the shot visualized above
start = index - 50
end = index + 50
transectIndex = np.arange(start, end, 1)

6.2 Plot Waveform Transects

In order to get an idea of the length of the beam transect that you are plotting, you can plot the x-axis as distance, which is calculated below.

# Calculate along-track distance
distance = np.arange(0.0, len(transectIndex) * 60, 60)  # GEDI Shots are spaced 60 m apart

In order to plot each vertical value for each waveform, you will need to reformat the data structure to match what is needed by holoviews Path() capabilities.

# Create a list of tuples containing Shot number, and each waveform value at height (z)
wfList = []
for s, i in enumerate(transectIndex):
    # Basic quality filtering from section 5
    if gediL1B['BEAM0110/geolocation/degrade'][i] == 0 and gediL1B['BEAM0110/stale_return_flag'][i] == 0:
        zStart = gediL1B['BEAM0110/geolocation/elevation_bin0'][i]
        zEnd = gediL1B['BEAM0110/geolocation/elevation_lastbin'][i]
        zCount = sdsCount[i]
        zStretch = np.add(zEnd, np.multiply(range(zCount, 0, -1), ((zStart - zEnd) / int(zCount))))
        waveform = gediL1B[sdsWaveform][sdsStart[i]: sdsStart[i] + zCount]
        waves = []
        for z, w in enumerate(waveform):
            waves.append((distance[s], zStretch[z],w))  # Append Distance (x), waveform elevation (y) and waveform amplitude (z)
        wfList.append(waves)
    else:
        print(f"Shot {s} did not pass quality filtering and will be excluded.")

Good news again, it looks like all of the waveforms in our transect passed the quality filtering test.

Below, plot each waveform by using holoviews Path() function. This will plot each individual waveform value by distance, with the amplitude plotted in the third dimension in shades of green.

import warnings
warnings.filterwarnings('ignore')
# Plot the waveform values for the transect with the waveform amplitude being plotted as the 'c' or colormap
l1bVis = hv.Path(wfList, vdims='Amplitude').options(color='Amplitude', clim=(200,405), cmap='Greens', line_width=20, width=950,
                                                    height=500, clabel='Amplitude',colorbar=True,
                                                    xlabel='Distance Along Transect (m)', ylabel='Elevation (m)',
                                                    fontsize={'title':16, 'xlabel':16, 'ylabel': 16, 'xticks':12, 'yticks':12,
                                                              'clabel':12, 'cticks':10}, colorbar_position='right',
                                                    title='GEDI L1B Waveforms across Redwood National Park: June 19, 2019')
l1bVis

GEDI L1B Waveforms across Redwood National Park: June 19, 2019

If you zoom in to a subset of the transect, you can get even better detail showing the structure of the canopy from the rxwaveform dataset, where denser portions of the canopy are seen in darker shades of green.

alt text

del waveform, waves, wfList, zStretch

7. Spatial Visualization

Section 7 combines many of the techniques learned above including how to import GEDI datasets, perform quality filtering, spatial subsetting, and visualization.

7.1 Import, Subset, and Quality Filter All Beams

Below, re-open the GEDI L1B observation--but this time, loop through and import data for all 8 of the GEDI beams.

beamNames = [g for g in gediL1B.keys() if g.startswith('BEAM')]

beamNames

['BEAM0000',
 'BEAM0001',
 'BEAM0010',
 'BEAM0011',
 'BEAM0101',
 'BEAM0110',
 'BEAM1000',
 'BEAM1011']

Loop through each of the desired datasets (SDS) for each beam, append to lists, and transform into a pandas DataFrame.

# Set up lists to store data
shotNum, dem, zElevation, zHigh, zLat, zLon, srf, degrade, beamI, rxCount, rxStart = ([] for i in range(11))  

# Loop through each beam and open the SDS needed
for b in beamNames:
    [shotNum.append(h) for h in gediL1B[[g for g in gediSDS if g.endswith('/shot_number') and b in g][0]][()]]
    [dem.append(h) for h in gediL1B[[g for g in gediSDS if g.endswith('/digital_elevation_model') and b in g][0]][()]]
    [zElevation.append(h) for h in gediL1B[[g for g in gediSDS if g.endswith('/elevation_bin0') and b in g][0]][()]]  
    [zHigh.append(h) for h in gediL1B[[g for g in gediSDS if g.endswith('/elevation_lastbin') and b in g][0]][()]]  
    [zLat.append(h) for h in gediL1B[[g for g in gediSDS if g.endswith('/latitude_bin0') and b in g][0]][()]]  
    [zLon.append(h) for h in gediL1B[[g for g in gediSDS if g.endswith('/longitude_bin0') and b in g][0]][()]]  
    [srf.append(h) for h in gediL1B[[g for g in gediSDS if g.endswith('/stale_return_flag') and b in g][0]][()]]  
    [degrade.append(h) for h in gediL1B[[g for g in gediSDS if g.endswith('/degrade') and b in g][0]][()]]  
    [rxCount.append(h) for h in gediL1B[[g for g in gediSDS if g.endswith('/rx_sample_count') and b in g][0]][()]]  
    [rxStart.append(h) for h in gediL1B[[g for g in gediSDS if g.endswith('/rx_sample_start_index') and b in g][0]][()]]  
    [beamI.append(h) for h in [b] * len(gediL1B[[g for g in gediSDS if g.endswith('/shot_number') and b in g][0]][()])]

Notice that we did not import the rxwaveform dataset--due to the large size of that dataset, we will wait until after spatial subsetting to import the waveforms.

# Convert lists to Pandas dataframe
allDF = pd.DataFrame({'Shot Number': shotNum, 'Beam': beamI, 'Latitude': zLat, 'Longitude': zLon, 'Tandem-X DEM': dem,
                      'Elevation bin 0 (m)': zElevation, 'Elevation last bin (m)': zHigh, 'Stale Return Flag': srf,
                      'Degrade Flag': degrade, 'rx Start Index': rxStart, 'rx Sample Count': rxCount})

del shotNum, dem, zElevation, zHigh, zLat, zLon, srf, degrade, beamI

7.2 Spatial Subsetting

Below, subset the pandas dataframe using a simple bounding box region of interest. If you are interested in spatially clipping GEDI shots to a geojson region of interest, be sure to check out the GEDI-Subsetter python script available at: https://git.earthdata.nasa.gov/projects/LPDUR/repos/gedi-subsetter/browse.

len(allDF)

3547051

Over 3.5 million shots are contained in this single GEDI orbit! Below subset down to only the shots falling within this small bounding box encompassing Redwood National Park. RedwoodNP our geopandas geodataframe can be called for the "envelope" or smallest bounding box encompassing the entire region of interest. Here, use that as the bounding box for subsetting the GEDI shots.

redwoodNP.envelope[0].bounds

(-124.16015705494489,
 41.080601363502545,
 -123.84950230520286,
 41.83981133687605)

minLon, minLat, maxLon, maxLat = redwoodNP.envelope[0].bounds  # Define the min/max lat/lon from the bounds of Redwood NP

Filter by the bounding box, which is done similarly to filtering by quality in section 5 above.

allDF = allDF.where(allDF['Latitude'] > minLat)
allDF = allDF.where(allDF['Latitude'] < maxLat)
allDF = allDF.where(allDF['Longitude'] > minLon)
allDF = allDF.where(allDF['Longitude'] < maxLon)

NOTE: It may take up to a few minutes to run the cell above.

allDF = allDF.dropna()  # Drop shots outside of the ROI

len(allDF)

4477

Notice you have drastically reduced the number of shots you are working with (which will greatly enhance your experience in plotting them below). But first, remove any poor quality shots that exist within the ROI.

# Set any poor quality returns to NaN
allDF = allDF.where(allDF['Stale Return Flag'].ne(1))
allDF = allDF.where(allDF['Degrade Flag'].ne(1))
allDF = allDF.dropna()
len(allDF)

4475

Next create a Shapely Point out of each shot and insert it as the geometry column in the [soon to be geo]dataframe.

# Take the lat/lon dataframe and convert each lat/lon to a shapely point
allDF['geometry'] = allDF.apply(lambda row: Point(row.Longitude, row.Latitude), axis=1)

# Convert to geodataframe
allDF = gp.GeoDataFrame(allDF)
allDF = allDF.drop(columns=['Latitude','Longitude'])

7.3 Visualize All Beams: Elevation

Now, using the pointVisual function defined in section 3.2, plot the geopandas GeoDataFrame using geoviews.

allDF['Shot Number'] = allDF['Shot Number'].astype(str)  # Convert shot number to string

vdims = []
for f in allDF:
    if f not in ['geometry']:
        vdims.append(f)

visual = pointVisual(allDF, vdims = vdims)
visual * gv.Polygons(redwoodNP['geometry']).opts(line_color='red', color=None)

GEDI shots in yellow.

Feel free to pan and zoom in to the GEDI shots in yellow.

Now let's not only plot the points in the geodataframe but also add a colormap for Elevation bin0 (m).

(gvts.EsriImagery * gv.Points(allDF, vdims=vdims).options(color='Elevation bin 0 (m)',cmap='terrain', size=3, tools=['hover'],
                                                          clim=(min(allDF['Elevation bin 0 (m)']),
                                                          max(allDF['Elevation bin 0 (m)'])), colorbar=True, clabel='Meters',
                                                          title='GEDI Elevation bin0 over Redwood National Park: June 19, 2019',
                                                          fontsize={'xticks': 10, 'yticks': 10, 'xlabel':16, 'clabel':12,
                                                                    'cticks':10,'title':16,'ylabel':16})).options(height=500,
                                                                                                                  width=900)

GEDI elevation bin0 over Redwood National Park.

8. Export Subsets as GeoJSON Files

In this section, export the GeoDataFrame as a .geojson file that can be easily opened in your favorite remote sensing and/or GIS software and will include an attribute table with all of the shots/values for each of the SDS layers in the dataframe.

First, import the waveforms for the subset of shots to be exported.

waves = []  # List to store each list of waveform values

# Loop through each row (shot) in the data frame, extract waveform values, append to list of lists
for _, shot in allDF.iterrows():
    starts = int(shot['rx Start Index'])
    wave = gediL1B[f"{shot['Beam']}/rxwaveform"][starts: starts + int(shot['rx Sample Count'])]

    # In order to export as geojson, the type needs to be converted from a numpy array to a single comma separated string
    waves.append(','.join([str(q) for q in wave]))

# Add to data frame
allDF['rx Waveform'] = waves

gediL1B.filename  # L1B Filename

'GEDI01_B_2019170155833_O02932_T02267_02_003_01.h5'

outName = gediL1B.filename.replace('.h5', '.json')  # Create an output file name using the input file name
outName

'GEDI01_B_2019170155833_O02932_T02267_02_003_01.json'

allDF.to_file(outName, driver='GeoJSON')  # Export to GeoJSON

del allDF, waves

Success! You have now learned how to start working with GEDI L1B files in Python as well as some interesting strategies for visualizing those data in order to better understand your specific region of interest. Using this jupyter notebook as a workflow, you should now be able to switch to GEDI files over your specific region of interest and re-run the notebook. Good Luck!

Contact Information

Material written by Cole Krehbiel1

    Contact: LPDAAC@usgs.gov
    Voice: +1-605-594-6116
    Organization: Land Processes Distributed Active Archive Center (LP DAAC)
    Website: https://lpdaac.usgs.gov/
    Date last modified: 05-11-2021
1KBR Inc., contractor to the U.S. Geological Survey, Earth Resources Observation and Science (EROS) Center, Sioux Falls, South Dakota, 57198-001, USA. Work performed under USGS contract G15PD00467 for LP DAAC2. 2LP DAAC Work performed under NASA contract NNG14HH33I.

Relevant Products

Product Long Name
GEDI01_B.001 GEDI L1B Geolocated Waveform Data Global Footprint Level

Tools

Name Filters Description
47914 252nd Street | Sioux Falls, SD 57198-0001
Voice: 605-594-6116 | Toll Free: 866-573-3222 | Fax: 605-594-6963
lpdaac@usgs.gov