sustainability

Article
Forest Canopy Height Retrieval and Analysis Using Random
Forest Model with Multi-Source Remote Sensing Integration
Weidong Zhu 1,2,3 , Yaqin Li 1, *, Kuifeng Luan 1,2 , Zhenge Qiu 1,2 , Naiying He 1,2 , Xiaolong Zhu 1 and Ziya Zou 1

1 College of Marine Science, Shanghai Ocean University, Shanghai 201306, China; wdzhu@shou.edu.cn (W.Z.);
kfluan@shou.edu.cn (K.L.); zgqiu@shou.edu.cn (Z.Q.); nyhe@shou.edu.cn (N.H.);
m210200619@st.shou.edu.cn (X.Z.); m210200585@st.shou.edu.cn (Z.Z.)
2 Shanghai Estuary Marine Surveying and Mapping Engineering Technology Research Center,
Shanghai 201306, China
3 Nansha Islands Coral Reef Ecosystem National Observation and Research Station, Guangzhou 510300, China
* Correspondence: m210200572@st.shou.edu.cn

Abstract: Forest canopy height is an important indicator of the forest ecosystem, and an accurate
assessment of forest canopy height on a large scale is of great significance for forest resource quantifi-
cation and carbon sequestration. The retrieval of canopy height based on remote sensing provides
a possibility for studying forest ecosystems. This study proposes a new method for estimating
forest canopy height based on remote sensing. In this method, the GEDI satellite and ICESat-2
satellite, which are different types of space-borne lidar products, are used to cooperate with the
Landsat 9 image and SRTM terrain data, respectively. Two forest canopy height-retrieval models
based on multi-source remote sensing integration are obtained using a random forest regression
(RFR) algorithm. The study, conducted at a forest site in the northeastern United States, synthesized
various remote sensing data sets to produce a robust canopy height model. First, we extracted
relative canopy height products, multispectral features, and topographic data from GEDI, ICESat-2,
Landsat 9, and SRTM images, respectively. The importance of each variable was assessed, and the
random forest algorithm was used to analyze each variable statistically. Then, the random forest
regression algorithm was used to combine these variables and construct the forest canopy height
Citation: Zhu, W.; Li, Y.; Luan, K.;
model. Validation with airborne laser scanning (ALS) data shows that the GEDI and ICESat-2 models
Qiu, Z.; He, N.; Zhu, X.; Zou, Z. Forest using a single data source achieve better accuracy than the Landsat 9 model. Notably, the combination
Canopy Height Retrieval and of GEDI, Landsat 9, and SRTM data (R = 0.92, MAE = 1.91 m, RMSE = 2.78 m, and rRMSE = 12.64%)
Analysis Using Random Forest Model and a combination of ICESat-2, Landsat 9, and SRTM data (R = 0.89, MAE = 1.84 m, RMSE = 2.54 m,
with Multi-Source Remote Sensing and rRMSE = 10.75%). Compared with the least accurate Landsat 9 model, R increased by 29.58%,
Integration. Sustainability 2024, 16, 93.48%, MAE by 44.64%, 46.20%, RMSE by 42.80%, 49.40%, and the rRMSE was increased by 42.86%
1735. https://doi.org/10.3390/ and 49.32%, respectively. These results fully evaluate and discuss the practical performance and
su16051735
benefits of multi-source data retrieval of forest canopy height by combining space-borne lidar data
Academic Editor: Georgios Koubouris with Landsat 9 data, which is of great significance for understanding forest structure and dynamics.
The study provides a reliable methodology for estimating forest canopy height and valuable insights
Received: 22 November 2023
into forest resource management and its contribution to global climate change.
Revised: 25 January 2024
Accepted: 25 January 2024
Keywords: forest canopy height estimation; remote sensing integration; GEDI; ICESat-2; Landsat 9;
Published: 20 February 2024
machine learning in remote sensing; random forest regression

Copyright: © 2024 by the authors.

Licensee MDPI, Basel, Switzerland. 1. Introduction
This article is an open access article In recent years, the negative terrestrial ecological impacts of climate change have
distributed under the terms and become more apparent and are very likely to intensify over the coming decades [1–3].
conditions of the Creative Commons
Forests constitute a crucial segment of land-based ecosystems and are instrumental in
Attribution (CC BY) license (https://
sustaining the cycle of carbon on a global scale. They play a key role in counteracting
creativecommons.org/licenses/by/
the effects of global warming, enhancing environmental conditions, and maintaining the
4.0/).

Sustainability 2024, 16, 1735. https://doi.org/10.3390/su16051735 https://www.mdpi.com/journal/sustainability

Sustainability 2024, 16, 1735 2 of 21

equilibrium of natural ecosystems [4–6]. The measurement of canopy height in forests is a

critical indicator of the overall structure of a forest and its capacity to store carbon. Large-
area and high-resolution forest canopy height data are fundamental to regional and global
forest carbon stock estimates. It is an important part of assessing the global carbon cycle and
estimating forest carbon storage, biomass, and forest dynamics [7–10]. Traditional methods
of measuring forest canopy height rely on field measurements [11]. Manual surveys
based on resource inventories can provide accurate and high-resolution data, but they are
resource-intensive, requiring extensive human, material, and time resources [10]. However,
these methods are only applicable to small forest areas, posing challenges in obtaining
forest canopy height data for large regions. Conversely, remote sensing technology has
advanced, allowing for the rapid, accurate, continuous, dynamic, long-term, and non-
destructive estimation of forest canopy height through the extraction of horizontal structural
information from remote sensing data over large areas [12,13]. Various methods, including
optical remote sensing, microwave remote sensing, and lidar, can be used for forest canopy
height retrieval. These methods significantly improve the efficiency and accuracy of forest
canopy height retrieval without causing damage to forest ecosystems [14].
The Landsat series of Earth observation satellites is the longest-running program in
Earth observation, continuously monitoring the global land surface on a moderate scale.
Millions of high-quality medium-resolution multispectral data have been acquired. These
data are stored at receiving stations in the United States and around the world for research
on global change [15]. The most recent addition to the Landsat program, Landsat 9, was
launched in September 2021 by the National Aeronautics and Space Administration (NASA)
and the United States Geological Survey (USGS) [16]. Landsat 9 is equipped with enhanced
sensors, including the Thermal Infrared Sensor-2 (TIRS-2) and Operational Land Imager-2
(OLI-2). Unlike its predecessor, Landsat 8, the instrument provides radiometric and geo-
metric data superior to those on the previous generation of Landsat satellites. Landsat 9
OLI-2 has a higher radiometric resolution, increasing the quantization from Landsat 8 from
12 bits to 14 bits, enabling the sensor to detect finer differences, especially in darker areas
such as water or dense forests, and has higher imaging capabilities to capture horizontal
structure data of large-scale forest canopy [17–19]. Compared to Landsat 8 thermal infrared
sensors (TIRS), the TIRS-2 also significantly reduces stray light, resulting in improved
atmospheric correction and more accurate surface temperature measurements. In addi-
tion, it contains abundant spectral features and vegetation indices, which can effectively
reduce the resource consumption of field measurements. It is an important data source for
estimating forest canopy height [16,19]. However, it faces challenges in acquiring vertical
structural characteristics of forests due to poor penetration, saturation, low data quality, and
substantial errors in estimating forest canopy height. Consequently, these factors impact
the accuracy of the estimation [19,20]. However, space-borne lidar has a higher orbital
altitude, wider observation range, and high penetration. These data provide a solution to
the challenges faced by Landsat satellite image data. In addition, it also provides reliable
data support for large-scale terrain elevation and forest canopy height inversion [21]. In
2018, NASA embarked on a groundbreaking endeavor by launching two laser altimeter
missions: ICESat-2 (Ice, Cloud, and Land Elevation Satellite 2) and GEDI (Global Ecosystem
Dynamics Investigation) [22]. GEDI, a cutting-edge multibeam waveform lidar, is tasked
with measuring the global tree height and canopy density through its unique eight laser
beam ground trajectories [23]. The system is the world’s first spaceborne lidar altimeter
system for high-resolution measurement of three-dimensional vertical structure and topog-
raphy of tropical and temperate forest vegetation [24–26]. ICESat-2 boasts the Advanced
Topographic Laser Altimeter System (ATLAS), incorporating pioneering micropulse multi-
beam photon counting lidar technology, a first for satellite applications. The primary
scientific objective is to gauge vegetation canopy height accurately [27]. ICESat-2 and GEDI
have unique advantages and represent the most advanced level of the space-borne laser
altimeter. They have significantly expanded our capabilities in understanding and studying
Sustainability 2024, 16, 1735 3 of 21

forest biomass, carbon, and water cycles and have opened new avenues for biodiversity
conservation [23,27,28].
Previous studies have successfully utilized ICESat-2 and GEDI data to estimate forest
height [29–31]. However, due to the characteristics of multispectral data and space-borne
lidar data, researchers have been exploring the combination and application of multiple
sensors and measurements to retrieve forest parameters [32]. Space-borne lidar data
directly provide precise information about the vertical vegetation structure [33]. When
combined with optical remote sensing data or radar data, this combination enables the
large-scale and comprehensive estimation of forest attributes, including height [20,34].
Qi et al. [35] explored the efficacy of simulating GEDI data for enhancing TDX InSAR
data height estimation in temperate forests, mountainous coniferous forests, and tropical
rainforests. The results demonstrated that combining both data types led to more accurate
forest canopy height estimation than using just one data source [35]. Potapov et al. [36]
and Tyukavina et al. [37] emphasized the significance of incorporating Landsat-based
surface phenology indices into forest structure modeling. Potapov et al. [38] utilized a
superpixel machine learning algorithm to generate a global forest canopy height map at a
resolution of 30 m by integrating GEDI and Landsat 8 data. The findings indicated that, in
comparison to ALS data, the RMSE and MAE were 9.07 m and 6.36 m, respectively. Wang
et al. [39] integrated WorldView-2 stereo images with Landsat 7 data and employed six
machine learning methods to estimate forest canopy height. The results demonstrated
that the gradient-enhanced regression method provided the most accurate estimation of
CHM, with the R of 0.64 and RMSE of 3.1 m. Additionally, the study highlighted the
limitations of using solely very high-resolution stereo images for CHM estimation and
suggested that integrating CHM from WorldView-2 stereo images with satellite vegetation
indices could enhance the accuracy of CHM estimation. Ren et al. [40] demonstrated
the accuracy of GEDI LiDAR data and multitemporal Sentinel-2 images in estimating
forest diversity in large areas using four regression algorithms in machine learning. In a
similar vein, Li et al. [9] integrated data from the ICESat-2 lidar with observations from
Sentinel-1, Sentinel-2, and Landsat-8 satellites. They employed deep learning techniques
and random forest models to generate a high-resolution CHM. Furthermore, Xi et al. [41]
utilized a combination of ICESat-2 lidar data and Sentinel-1 and Sentinel-2 data. They
employed random forest and gradient-boosting decision tree methods in machine learning
to develop a forest canopy height estimation model. Likewise, Guan et al. [42] employed
GEDI and Landsat 8 data to generate canopy height maps for estimating forest age. Zhu
et al. [43], based on the combination of GED lidar data and Landsat 8 and Landsat 9
data, developed a forest canopy height estimation model using the BP neural network in
machine learning. The results show that the Landsat 9 model is superior to the Landsat
8 model, and the combination model of GEDI and Landsat 9 is the best. These studies
collectively demonstrate the feasibility and effectiveness of combining space-borne lidar
data with multispectral data for forest parameter retrieval, using machine learning methods
to achieve significant results. Despite the advancements in space-borne lidar technology
represented by GEDI and ICESat-2, research on the synergistic application of the most
recent Landsat 9 data with these lidar systems is limited. Although the combination of
GEDI and Landsat 9 data has been discussed in the literature [43], the time consistency
processing of Landsat 9 images in the literature still needs to be optimized. At the same
time, the extraction of Landsat 9 data feature variables in the literature only analyzes part
of vegetation index variables, and there is no more comprehensive selection and analysis
of Landsat 9 data feature variables. To date, there have been few studies on the potential
integration of ICESat-2 and Landsat 9 data to assess forest canopy height. GEDI and
ICESat-2 stand at the forefront of space-borne lidar, yet the real-world performance and
benefits of combining these data with Landsat 9 for detailed forest canopy height retrieval
have not been fully evaluated or discussed. These gaps highlight the great opportunity to
enhance forest canopy height research and improve understanding of forest structure and
Sustainability 2024, 16, x FOR PEER REVIEW 4 of 22

Sustainability 2024, 16, 1735 4 of 21

optimizing data processing, assessment, and analysis, integrating advanced data sources,
and using superior machine learning methods.
dynamics by optimizing data processing, assessment, and analysis, integrating advanced
Therefore, this study employs GEDI data, ICESat-2 data, Landsat 9 OLI-2 data, and
data sources, and using superior machine learning methods.
SRTM terrain factors as the primary data sources. Various combinations of data sources
Therefore, this study employs GEDI data, ICESat-2 data, Landsat 9 OLI-2 data, and
are employed and compared to individual data sources. The random forest regression
SRTM terrain factors as the primary data sources. Various combinations of data sources
algorithm is employed to model the retrieval of forest canopy height within the study
are employed and compared to individual data sources. The random forest regression
area. The canopy height model (CHM) extracted from airborne lidar is integrated as ref-
algorithm is employed to model the retrieval of forest canopy height within the study area.
erence data. Lastly, the accuracy and precision of models using single and multiple data
The canopy height model (CHM) extracted from airborne lidar is integrated as reference
sources are compared and analyzed. In this study, the superiority and reliability of forest
data. Lastly, the accuracy and precision of models using single and multiple data sources
canopy height retrieval results of the random forest regression model based on multi-
are compared and analyzed. In this study, the superiority and reliability of forest canopy
source data integration were fully evaluated by combining Landsat 9 data with different
height retrieval results of the random forest regression model based on multi-source data
models of space-borne lidar data. It is proved that it is possible to obtain forest canopy
integration were fully evaluated by combining Landsat 9 data with different models of
height with lowlidar
space-borne cost,data.
highItaccuracy,
is provedand high
that it isefficiency in obtain
possible to a largeforest
area, which
canopyprovides a
height with
reference for regional scale forest remote sensing monitoring.
low cost, high accuracy, and high efficiency in a large area, which provides a reference for
regional scale forest remote sensing monitoring.
2. Materials and Methods
2.1.2.Study
Materials
Area and Methods
2.1.This
Study Area concentrates on the two forested areas located in the northeastern
research
United This research
States. concentrates
The chosen on the
study areas twoHarvard
were forested Forest
areas located
(HARV) in in
theWorcester
northeastern United
County,
States. The chosen
Massachusetts (42°32′24″study
N, areas wereW)
72°10′12″ Harvard Forest
(see Figure 1). (HARV) in Worcester
HARV serves County, Mas-
as a representative
sachusetts
rural/wild (42◦ 32
region in′ the
24′′ Northeast,
N, 72◦ 10′ 12with
′′ W) (see Figure 1). HARV serves as a representative
elevations ranging from 160 to 415 m and an av-
rural/wild
erage elevation region
of 348inm. theItsNortheast,
climate iswith elevations
classified ranging from 160 to 415 m and
as temperate–continental–humid. Thean
average
main elevation
vegetation of 348
in this m. Its
region climate
is an Easternis classified
deciduousastemperate
temperate–continental–humid.
forest, primarily featur-The
main vegetation in this region is an Eastern deciduous temperate forest,
ing red maple (Acer rubrum) and white pine (Pinus strobus) [44,45]. This study involved primarily featuring
redcollection
data maple (Acerfromrubrum) and white
the selected study pine (Pinus
areas, strobus)
as well [44,45].
as the This study
development involved
and testing data
of
collection from
relevant methodologies.the selected study areas, as well as the development and testing of relevant
methodologies.

(a) (b)
Figure 1. The
Figure 1. locations of study
The locations of sites
studyin sites
the USinandtheGEDI
US and and GEDI
ICESat-2/ATLAS ground tracks.
and ICESat-2/ATLAS (a)
ground
Study sites in Worcester County, Massachusetts, and (b) study site in Harvard Forest, Worcester
tracks. (a) Study sites in Worcester County, Massachusetts, and (b) study site in Harvard Forest,
County, Massachusetts,
Worcester with coverage
County, Massachusetts, of satellite
with coverage ground tracks.
of satellite ground tracks.

2.2. Research Data

2.2.1. NEON Airborne LiDAR Data
The airborne lidar data used in this study were obtained from the National Ecological
Observatory Network (NEON). NEON’s airborne observation platform is equipped with
Sustainability 2024, 16, 1735 5 of 21

an imaging spectrometer, an airborne lidar, and a high-resolution color camera. This initia-
tive aims to gather a range of data, including multispectral images, lidar measurements,
and varied ecological indicators, to investigate the impacts of climatic shifts, alterations in
land utilization, and the proliferation of invasive species on ecosystems at a continental
scale [46–50]. The airborne lidar-specific parameters can be found in Table 1 [22]. The
NEON lidar-gridded data products used in this study were primarily acquired in Au-
gust 2022 and can be downloaded from the following address: Explore Field Sites|NSF
NEON|Open Data to Understand our Ecosystems (neonscience.org) [51]. The canopy
heights for reference forests were determined by selecting the 90th percentile value within
the 25 m × 25 m grid of the canopy height model (CHM) [36].

Table 1. Specifications of the NEON airborne lidar instrument and its derived products [22].

NEON Airborne Lidar Instrument and Product Specifications

Elevation
Laser wavelength 1064 nm (near IR) 5–35 cm
accuracy
DTM and CHM (in the format
Laser power 250 µJ Derived products
of 1 km by 1 km mosaic tiles)
Laser repetition
33–167 kHz Product resolution uniform grid (1 m × 1 m)
rate
0.25 m (at 1000 m flying
footprint diameter height), 0.8 m in wide Terrain parameters elevation and slope
beam-divergence mode
Sampling density 1–4 points per square meter Vertical datum GEOID12A
Horizontal canopy top height, relative
5–15 cm Canopy parameters
accuracy height (RH), and canopy cover

2.2.2. GEDI Data

Table 2 [22,24,52] provides a record of the parameter names and sizes used in the GEDI
system. The GEDI system consists of three lasers. Two of the lasers operate independently
at full power, while the third laser is bifurcated into two beams, generating four beams.
These beams are then used to generate eight light spot trajectories on the surface of the Earth
through beam jittering. The radar records the complete waveforms of the footprints [24],
which can be seen in Figure 2. For temporal consistency with airborne LiDAR data, this
study downloaded the latest GEDI L2A version 2 products from January 2022 to December
2022 via the NASA Earth Data website (https://search.earthdata.nasa.gov/) (accessed on
12 October 2023) [53,54].

Table 2. GEDI parameters name and size [22,24,52].

Parameters Name Parameters Size Parameters Name Parameters Size

Track height ~400 km Footprint 25 m
Coverage 51.6◦ N~51.6◦ S Geolocation error 8m
repetition
242 Hz Along-track distances 60 m
rate
Pulse width 15 ns Across-track distances 600 m
Wavelength 1064 nm Product tested in this study L2A (elevation and canopy heights)

We filter out invalid GEDI footprint data by utilizing parameters inherent in GEDI
L2A data to ensure high-quality footprint points are selected for training and valida-
tion. From the GEDI L2A product, we filtered the GEDI footprint within the study area
based on latitude and longitude. This filtering process is guided by the following param-
eters: quality_flag = 1, rx_assess_flag = 0, rx_algrunflag_flag = 1, degrade_flag = 0, and
sensitivity ≥0.9 as a method of selecting the desired footprint [52]. The optimal algorithm
was analyzed to obtain the final valid GEDI dataset by evaluating the data for the six
Sustainability 2024, 16, 1735 6 of 21

Sustainability 2024, 16, x FOR PEER REVIEW

algorithms 6 valid
(see Table 3) within the GEDI L2A footprint and determining the common of 22
footprint among these algorithms [29,31,54].

GEDImulitbeam
Figure2.2.GEDI
Figure mulitbeamsampling
samplingmode.
mode.

Table 2.
Table The parameter
3. GEDI configurations
parameters name and sizefor[22,24,52].
GEDI L2A’s six algorithms with σ being utilized to denote
the standard deviation of the background noise level [29,31].
Parameters Name Parameters Size Parameters Name Parameters Size
Track height ~400 km Smoothing Footprint Waveform Waveform
25 m
Algorithm Smoothing
Signal Start Signal End
Setting Group
Coverage Width51.6°
(Noise)
N~51.6°Width
S (Signal)
GeolocationThreshold
error 8 m
Threshold
repetition
a1 6.5σ 242 Hz Along-track distances
6.5σ 3σ 60 m6σ
rate
a2 6.5σ 3.5σ 3σ 3σ
a3 6.5σ 3.5σ Across-track dis-
3σ 6σ
Pulse width 15 ns 600 m
a4 6.5σ 6.5σ tances 6σ 6σ
a5 6.5σ 3.5σ 3σ 2σ and
Product tested in this L2A (elevation
Wavelength
a6 6.5σ1064 nm 3.5σ 3σ 4σ
study canopy heights)

2.2.3.We
ICESat-2 Data
filter out invalid GEDI footprint data by utilizing parameters inherent in GEDI
ICESat-2/ATLAS
L2A data includes footprint
to ensure high-quality 21 standard dataare
points products
selectedATL01~ATL21,
for training and which are di-
validation.
videdthe
From into five L2A
GEDI levels, represented
product, by Level
we filtered 0, Level
the GEDI 1, Level
footprint 2, Level
within the 3A,
study and Level
area 3B,
based
respectively [27,55]. Among them, the data products related to forest
on latitude and longitude. This filtering process is guided by the following parameters:vegetation include
ATL03 and ATL08.
quality_flag ATL03 provides
= 1, rx_assess_flag a global positioning
= 0, rx_algrunflag_flag = 1,photonic data product,
degrade_flag and ATL08
= 0, and sensitivity
≥0.9 as a method of selecting the desired footprint [52]. The optimal algorithm washeight,
is a terrestrial vegetation data product that provides topographic elevation, canopy ana-
and relative
lyzed to obtainheight indicators
the final every dataset
valid GEDI 100 m along the routethe
by evaluating through thethe
data for official processing
six algorithms
of ATL03
(see Table 3)data. In addition,
within the GEDIATL08 can be associated
L2A footprint with ATL03
and determining productsvalid
the common to provide
footprintthe
opportunity to calculate ground
among these algorithms [29,31,54]. elevation, canopy height, and relative height indicators at
different scales [41,56–59].
Table ATL03 and ATL08
3. The parameter version 6 for
configurations data products
GEDI are
L2A’s six used in this
algorithms with study.
σ being Considering the
utilized to denote
acquisition time of airborne lidar data, GEDI,
the standard deviation of the background noise level [29,31].and Landsat 9, the inaccuracy of forest
canopy height inversion caused by time difference is avoided. The data acquisition time
is from 1 January 2022
Algorithm 2022. The dataWaveform
to 31 DecemberSmoothing
Smoothing come fromSig- Waveform
the National Snow Sig-
and
Ice Data Center of the United States nal Start Thresh-
(https://nsidc.org/data/atl08/versions/6) nal End Thresh-
(accessed
Setting Group Width (Noise) Width (Signal)
on 12 October 2023). Details are shown in Table 4. To facilitate old subsequent analysis,
old we
extracteda1 a set of relative6.5σ
height indicators6.5σ ATL08 data: RH25, 3σRH50, RH60, RH75, 6σ RH85,
a2
RH90, RH98, and RH100.6.5σ For ATL03 data, we 3.5σobtained signal 3σphotons by matching 3σ them to
corresponding
a3 ATL08 data 6.5σproducts and extracted
3.5σ forest height
3σ metrics from these 6σ signal
photonsa4for each 25 m swath. 6.5σ 6.5σ 6σ 6σ
a5 6.5σ 3.5σ 3σ 2σ
a6 6.5σ 3.5σ 3σ 4σ

2.2.3. ICESat-2 Data

ICESat-2/ATLAS includes 21 standard data products ATL01~ATL21, which are di-
vided into five levels, represented by Level 0, Level 1, Level 2, Level 3A, and Level 3B,
respectively [27,55]. Among them, the data products related to forest vegetation include
Sustainability 2024, 16, 1735 7 of 21

Table 4. ICESat-2 data acquisition time.

Data Type Acquisition Time Number of Documents

ICESat-2 ATL08 2022/01/01–2022/12/31 4
ICESat-2 ATL03 2022/01/01–2022/12/31 4

To ensure the consistent analysis of elevation data from various datasets, we utilized
NOAA’s VDatum, which can be accessed at https://vdatum.noaa.gov/vdatumweb/ (ac-
cessed on 12 October 2023), to align the GEDI data and ICESat-2 data vertical datum with
that of NEON. It is worth noting that crown height remains relative to the ground and is
unaffected by the chosen vertical datum [22].

2.2.4. Landsat 9 Data

Because discrete GEDI and ICESat-2 footprints do not provide continuous coverage,
it is necessary to prepare optical remote sensing data for omnidirectional forest height
estimation. Mid-resolution Landsat 9 data are improved based on the previous generation.
They have the advantages of availability, wide coverage, and high accuracy, but they have
not been widely used in the study of omni-directional forest height estimation. This study
mainly uses Landsat 9 optical remote sensing data. A total of 73 Landsat9 Level 1T data
images covering the study area from January to December 2022 were acquired from the
Google Earth Engine (GEE) website platform (http://developers.google.cn/) (accessed on
12 October 2023).
The GEE cloud removal algorithm was applied to the images, and an annual synthesis
of time series images was performed to generate Landsat 9 data covering the study area
in 2022 through median processing [60] (see Figure 3). Based on spectral characteristics,
texture characteristics, and the vegetation index, the canopy spectral reflectance information
is provided using seven bands of coastal, blue, green, red, NIR, SWIR1, and SWIR2 after
image processing of Landsat 9 data (see Table 5) [61]. Six kinds of feature variables are
extracted: the single-band factor, multi-band factor, vegetation index, principal component
analysis (PCA) factor, minimum noise fraction (MNF) transform, and texture factor [62–65].
Sustainability 2024, 16, x FOR PEER REVIEW 8 of 22
Further details of all Landsat 9-related variables extracted in this study are summarized in
Table 6.

Figure 3. Annual image synthesis of Landsat 9 image dataset.

Figure 3. Annual image synthesis of Landsat 9 image dataset.
Table 5. Landsat 9 operational land imager-2 parameters [61].

Band No. Band Band Range/µm Spatial Resolution/m

B1 Coastal 0.43~0.45 30
B2 Blue 0.45~0.51 30
B3 Gree 0.53~0.59 30
B4 Red 0.64~0.67 30
Sustainability 2024, 16, 1735 8 of 21

Table 5. Landsat 9 operational land imager-2 parameters [61].

Band No. Band Band Range/µm Spatial Resolution/m

B1 Coastal 0.43~0.45 30
B2 Blue 0.45~0.51 30
B3 Gree 0.53~0.59 30
B4 Red 0.64~0.67 30
B5 NIR 0.85~0.88 30
B6 SWIR1 1.57~1.65 30
B7 SWIR2 2.11~2.29 30
B8 Pan 0.50~0.68 15
B9 Cirrus 1.36~1.39 30

Table 6. Summary of characteristic variables calculated from GEDI L2A, ICESat-2, Landsat 9, and
SRTM data [62–65].

Type Characteristic Variable Description

RH25, RH50, RH60, RH75, RH85, RH90, GEDI extracted relative elevation (25th, 50th, 60th,75th, 85th,
GEDI L2A
RH98, and RH100 90th, 98th, and 100th)
RH25, RH50, RH60, RH75, RH85, RH90, ICESat-2 extracted relative elevation (25th, 50th, 60th, 75th,
ICESat-2
RH98, and RH100 85th, 90th, 98th, and 100th)
B1, B2, B3, B4, B5, B6, and B7 Landsat 9 bands 1, 2, 3, 4, 5, 6, and 7
B24 B2/B4
B74 B7/B4
B76 B7/B6
B345 B3 · B4/B5
 
EVI B5− B4
2.5 B5+6B4−7.5B2+1
DVI B5 − B4
SLAVI B5
B4+ B7
VI3 ( B5 − B6)/( B5 + B6)
Landsat 9 q
PVI (0.355B5 − 0.149B4)2 + (0.355B4 − 0.852B5)2
NDVI ( B5 − B4)/( B5 + B4)
√
RDVI B5 − B4/ B5 + B4
ND43 ( B4 − B3) · ( B4 + B3)
ND67 ( B6 − B7) · ( B6 + B7)
PC1, PC2, and PC3 The first to third bands of principal component analysis.
TCB The tassel cap transforms the Brightness band.
TCG The tassel cap transforms the green band.
TCW The tassel cap transforms the wetness band.
MNF1, MNF2, MNF3, and MNF4 Minimum noise fraction first to fourth band.
elevation Elevation extracted from DEM.
SRTM slope Slope extracted from DEM.
aspect Aspect extracted from DEM.

2.2.5. Auxiliary Data

Due to factors such as sunlight, soil moisture, and topographical changes, plant
growth varies at different slopes and orientations, resulting in different canopy heights.
To mitigate the effects of slope, aspect, and elevation on terrain and canopy height, we
introduced ancillary data called the Shuttle Radar Terrain Mission (SRTM) [66,67]. SRTM
Digital Elevation Model (DEM) data are available in two resolutions: 30 m and 90 m [68].
Topographic parameters such as elevation, slope, and aspect were extracted from SRTM
data with a resolution of 30 m in 2022 obtained from http://earthexplorer.usgs.gov/
(accessed on 12 October 2023). To ensure the consistency of extracted variables, this study
uses the central coordinates of GEDI and ICESat-2 footprints as the reference for variable
extraction, effectively solving the problem of resolution differences among SRTM data,
Sustainability 2024, 16, 1735 9 of 21

GEDI data, ICESat-2 data, and neon data sources [69]. Taking these parameters as input
variables of the random forest model, the effect of terrain on forest height is effectively
mitigated.

2.3. Methods
In this study, the GEDI, ICESat-2, Landsat 9, and SRTM data were used to extract the
characteristic index (see Table 6). Then, the multivariate importance analysis is carried out
on these characteristic indexes to obtain the final characteristic variable set (see Table 7).
By combining GEDI, ICESat-2, Landsat 9, and SRTM data, respectively, three models
based on the single data source, one model based on GEDI, Landsat 9, and SRTM data
source combination, and one model based on ICESat-2, Landsat 9, and SRTM data source
combination were established, respectively, using the random forest regression algorithm.
At last, the accuracy of the six inversion models is verified using airborne LiDAR data.

Table 7. The feature variables utilized in the final experimental modeling process [62–65].

Number of Characteristic
Model Characteristic Variable
Variables
GEDI rh98, rh100, rh90, rh85, rh25, and rh75 6
MNF2, B3, TCW, B74, B7, MNF3, B6,
Landsat 9 ND43, MNF4, B76, EVI, B1, B24, 19
ND67, B2, B345, B4, pc3, and VI3
rh98, rh100, MNF2, rh85, rh90, B3,
MNF3, B74, MNF4, rh25, TCW, B24,
GEDI and Landsat 9 20
EVI, B7, B76, B4 ND43, rh75, B345,
and rh60
rh85, rh90, rh75, rh98, rh25, rh100,
ICESat-2 7
and rh60
B7, B3, MNF2, MNF4, MNF3, TCW,
pc3, B74, ND43, B1, B6, B2, B76, B24,
Landsat 9 22
EVI, B4, pc2, ND67, VI3, B345, TCB,
and SLAVI
rh85, rh90, rh75, B3, MNF2, B7, rh98,
rh60, rh100, B76, MNF3, rh25, TCW,
ICESat-2 and Landsat 9 23
MNF4, B24, B74, EVI, ND43, B4, B1,
ND67, pc3, and rh50

2.3.1. Importance Analysis of Feature Variables

The random forest algorithm can reduce the dimensionality of characteristic variables,
thereby improving the computational efficiency of machine learning by screening vari-
ables [43]. Numerous studies have extensively shown the effectiveness of random forests in
predicting forest attributes with remote sensing data [70]. In this study, the random forest
method was used to analyze the importance of each characteristic variable and the first
90% of the characteristic variables with a cumulative contribution rate (the blue bars in
Figure 4 represent the characteristic variables in the top 90% of the cumulative contribution
rate) were screened out by calculating the importance scores of characteristic variables, and
the contribution of each characteristic variable to each tree species in random forest set was
evaluated. Then, the subsequent study of forest canopy height inversion was carried out.
In this study, the forest canopy height retrieval experiment was carried out based on
different space-borne lidar data products after screening the feature variables by analyzing
and evaluating the importance scores of the feature variables. In this study area, the GEDI
model and Landsat 9 model of single data source and GEDI and Landsat 9 model of
combined data source are established based on GEDI L2A products, ICESat-2 model and
Landsat 9 model of single data source and ICESat-2 and Landsat 9 model of combined
data source are established based on ICESat-2/ATLAS products (see Table 7). The two
 bles, thereby improving the computational efficiency of machine learning by screening
variables [43]. Numerous studies have extensively shown the effectiveness of random for-
ests in predicting forest attributes with remote sensing data [70]. In this study, the random
forest method was used to analyze the importance of each characteristic variable and the
Sustainability 2024, 16, 1735 first 90% of the characteristic variables with a cumulative contribution rate (the blue10bars of 21
in Figure 4 represent the characteristic variables in the top 90% of the cumulative contri-
bution rate) were screened out by calculating the importance scores of characteristic vari-
ables, and theestablished
experiments contribution in of each
this characteristic
study variable
fully evaluate to each
the actual tree speciesand
performance in random
benefit
forest set was evaluated. Then, the subsequent study of forest canopy
of Landsat 9 combined with different space-borne lidar data products for detailedheight inversion
forest
was
canopycarried out.retrieval so that the results of forest canopy height retrieval using multi-
height
source remote sensing integration can be strictly analyzed and discussed.

Sustainability 2024, 16, x FOR PEER REVIEW 11 of 22

(a) GEDI (b) Landsat 9 (c) GEDI and Landsat 9

(d) ICESat-2 (e) Landsat 9 (f) ICESat-2 and Landsat 9

Figure
Figure 4.4. Based
Basedon
onthe
the feature
feature variables
variablesextracted
extractedfrom
fromdifferent
differentspace-borne
space-bornelidarlidardata,
data,the
theim-
im-
portance ranking and sorting were carried out. (a) Importance ranking of GEDI feature
portance ranking and sorting were carried out. (a) Importance ranking of GEDI feature variables. variables.
(b)
(b) Importance
Importance ranking
ranking of Landsat 99 feature
of Landsat featurevariables
variablescorresponding
correspondingtotoGEDI.
GEDI.(c)(c)Importance
Importance rank-
ranking
ing of all feature variables of GEDI and Landsat 9 data. (d) Importance ranking of ICESat-2 feature
of all feature variables of GEDI and Landsat 9 data. (d) Importance ranking of ICESat-2 feature
variables. (e) Landsat 9 feature variable importance ranking corresponding to ICESat-2. (f) Landsat
variables. (e) Landsat 9 feature variable importance ranking corresponding to ICESat-2. (f) Landsat 9
9 feature variable importance ranking of ICESat-2 and Landsat 9 data.
feature variable importance ranking of ICESat-2 and Landsat 9 data.

2.3.2.InConstruction
this study, the
of forest canopy
the Forest height
Canopy retrieval experiment
Height-Retrieval Modelwas carried out based on
different space-borne lidar data products after screening the feature variables by analyz-
The random forest algorithm is capable of handling a large number of input variables,
ing and evaluating the importance scores of the feature variables. In this study area, the
missing data, non-linear relationships, and data with non-Gaussian distributions. Random
GEDI model and Landsat 9 model of single data source and GEDI and Landsat 9 model
forest regression trees are a machine learning algorithm that has been successfully employed
of combined data source are established based on GEDI L2A products, ICESat-2 model
for modeling the structure and biomass of forests at regional and global scales [32,70]. In
and Landsat 9 model of single data source and ICESat-2 and Landsat 9 model of combined
this study, we implemented a random forest regression tree ensemble method to estimate
data
forestsource are established based on ICESat-2/ATLAS products (see Table 7). The two ex-
height.
periments established
The random forest in this study fully
regression model evaluate
in the the actual
study [71]performance
was trainedand benefit
based of
on the
Landsat 9 combined with different space-borne lidar data products for detailed
random forest regulator function of the Scikit-leam library [72] in Python. To minimize the forest can-
opy height retrieval
unexpected so that
error caused bythe resultsthe
splitting of training
forest canopy
set andheight retrieval
the test set, theusing multi-source
final screened data
remote sensing integration can be strictly analyzed and discussed.
set of the study area was split into the training set and the validation set using a ten-fold
cross-validation method. The principle of cross-validation is to divide the data set into ten
2.3.2.
subsetsConstruction
and use oneofsubset
the Forest
as theCanopy Height-Retrieval
validation Model
set and the rest as the training set. We change
The random forest algorithm is capable of handling a large number of input varia-
bles, missing data, non-linear relationships, and data with non-Gaussian distributions.
Random forest regression trees are a machine learning algorithm that has been success-
fully employed for modeling the structure and biomass of forests at regional and global
scales [32,70]. In this study, we implemented a random forest regression tree ensemble
Sustainability 2024, 16, 1735 11 of 21

the verification subset in different iterations. The evaluation scores for each iteration are
averaged to obtain a more robust evaluation score for the model. A 10-fold cross-validation
method was used to find the optimal parameters (see Table 8) [73,74] and to build the
optimal model.

Table 8. Optimal parameters of the random forest model from different data sources.

Parameter
DATA GROUP
n_estimators max_depth min_samples_split min_samples_leaf max_features
GEDI 288 15 1 5 0.96
Landsat 9 312 17 2 6 0.89
GEDI and
300 15 1 6 0.97
Landsat 9
ICESat-2 280 12 1 6 0.92
Landsat 9 310 16 2 7 0.89
ICESat-2 and
200 14 1 5 0.90
Landsat 9

According to the above information. The random forest canopy height inversion
model was constructed using single and combined data sources. Table 8 lists the final
parameter configurations for each model.

2.4. Accuracy Assessment

Utilizing the DTM and CHM data in NEON as reference data, terrain and canopy
height were evaluated.
The correlation coefficient (R), root-mean-square error (RMSE), mean absolute error
(MAE), and relative root-mean-square error (rRMSE) are calculated to measure the accuracy
of each dataset in the model.
n
∑ (xi − x)(yi − y)
i=1
R= s (1)
n n
2 2
∑ (xi − x) · ∑ (yi − y)
i=1 i=1

1 n
n i∑
MAE = |yi − xi | (2)
=1
v
u n 2
u ∑ ( y i − xi )
u
t i=1
RMSE = (3)
n
v
u n 2
u ∑ (yi − xi )
u
100% t i=1
rRMSE = (4)
y n
where n represents the total number of samples, yi denotes the verification value, xi is the
prediction value, y represents the average of the verification value, and x represents the
average of the prediction value.

3. Results
3.1. Utilizing GEDI L2A Product for Forest Canopy Height Retrieval
In this study area, 5398 valid footprints screened using GEDI L2A products were used
for accuracy validation to assess the ability of GEDI L2A data to accurately estimate forest
Sustainability 2024, 16, 1735 12 of 21

canopy height. In addition to the six algorithms, GEDI L2A provides a default optimal
Sustainability 2024, 16, x FOR PEER REVIEW 13 of 22
algorithm for relative canopy height [75]. The study used a default algorithm to obtain
a set of relative altitude metrics (RH90, RH92, RH94, RH96, RH98, and RH100) for each
footprint in the study area and assess the correlation with the airborne radar RH90. Table 9
Table 9the
shows shows the correlation
correlation results between
results between the two.the two.9Table
Table 9 illustrates
illustrates that thethat the retrieval
retrieval accuracy
ofaccuracy of forest
forest canopy canopy
height height gradually
increases increases gradually
from RH90 from RH90 and
to RH98 to RH98
thenand then de-
decreases from
creases from RH98 to RH100. It is estimated that RH98 of GEDI L2A can
RH98 to RH100. It is estimated that RH98 of GEDI L2A can be used to retrieve canopy be used to retrieve
canopyThis
height. height. This
is the is the
same sameasresult
result as theforest
the GEDI GEDIcanopy
forest canopy height values
height values selectedselected
from the
from the literature
literature [76]. [76].

Table 9. The accuracy of forest canopy and reference airborne lidar forest canopy was assessed using
Table 9. The accuracy of forest canopy and reference airborne lidar forest canopy was assessed using
default algorithmic methods in GEDI L2A products.
default algorithmic methods in GEDI L2A products.
Relative
Relative R MAE/m RMSE/m rRMSE N
Height (RH) R MAE/m RMSE/m rRMSE N
Height (RH)
RH90 0.74 4.37 6.16 28.21% 5398
RH90
RH92 0.74
0.75 4.37
3.93 6.16
5.67 28.21%
25.94% 5398
5398
RH92 0.75 3.93 5.67 25.94% 5398
RH94 0.77 3.55 5.24 23.98% 5398
RH94 0.77 3.55 5.24 23.98% 5398
RH96
RH96 0.78
0.78 3.27
3.27 4.96
4.96 22.71%
22.71% 5398
5398
RH98
RH98 0.78
0.78 3.22
3.22 4.94
4.94 22.59%
22.59% 5398
5398
RH100
RH100 0.78 4.00
4.00 5.63
5.63 25.76%
25.76% 5398
5398

We extracted values from six different sets of algorithms for RH98 data in GEDI prod-
We extracted values from six different sets of algorithms for RH98 data in GEDI
ucts to investigate the correlation between data from these six sets of RH98 algorithms
products to investigate the correlation between data from these six sets of RH98 algorithms
and airborne data for reference forest heights RH90. The results are shown in Figure 5.
and airborne data for reference forest heights RH90. The results are shown in Figure 5.
Among the different algorithm groups, a4 and a5 have the worst accuracy of canopy
Among the different algorithm groups, a4 and a5 have the worst accuracy of canopy height
height inversion, with RMSE values as high as 9.56 m. The a2 algorithm has the highest
inversion, with RMSE values as high as 9.56 m. The a2 algorithm has the highest accuracy
accuracy (R = 0.78, MAE = 3.19, RMSE= 4.94 m, and rRMSE = 22.60%). Therefore, the sub-
(Rsequent
= 0.78,use
MAE = 3.19,
of an RMSE =data
a2 algorithm 4.94inversion
m, and rRMSE = 22.60%). Therefore, the subsequent
model construction.
use of an a2 algorithm data inversion model construction.

Figure5.5.Airborne
Figure Airbornelidar
lidarscatter plot
scatter of of
plot forest canopy
forest height
canopy (RH98)
height (RH98)andand
reference
referenceforest height
forest (RH90)
height
(RH90) extracted
extracted using
using six six different
different GEDI GEDI algorithms
algorithms (Dashed
(Dashed lineline is baseline,
is baseline, redred
is isfitfitline,
line,(a1–a6):
(a1–a6): six
six algorithms
algorithms representing
representing GEDI
GEDI L2AL2A data).
data).
 Sustainability
Sustainability 2024,
2024, 16,16, x FOR PEER REVIEW
1735 14 of 22
13 of 21

3.2.Utilizing
3.2. Utilizingthe
the ICESat-2
ICESat-2 ATL08
ATL08 Product
Productfor
forForest
ForestCanopy
CanopyHeight Retrieval
Height Retrieval
We extracted a set of relative height metrics from each footprint
We extracted a set of relative height metrics from each footprint ininthethe
study area
study in in
area
ICESat-2 ATL08 data: RH75, RH85, RH90, RH98, and RH100. For ATL03 data,
ICESat-2 ATL08 data: RH75, RH85, RH90, RH98, and RH100. For ATL03 data, we obtained we obtained
signalphotons
signal photonsby by matching
matching them
them totocorresponding
correspondingATL08
ATL08data
dataproducts,
products,extracted forest
extracted forest
height indicators for each 25-meter strip from these signal photons, and assessed
height indicators for each 25-meter strip from these signal photons, and assessed correlation correla-
tion with reference forest height airborne radar RH90. Figure 6 shows the accuracy of ICE-
with reference forest height airborne radar RH90. Figure 6 shows the accuracy of ICESat-
Sat-2/ATLAS-derived RH indicators versus reference forest heights. Of all relative height
2/ATLAS-derived RH indicators versus reference forest heights. Of all relative height
indicators, RH90 performed best in estimating forest height. Therefore, RH90 derived
indicators, RH90 performed best in estimating forest height. Therefore, RH90 derived from
from ICESat-2 was selected as the forest height index. Figure 7 shows a scatter plot of the
ICESat-2 was selected as the forest height index. Figure 7 shows a scatter plot of the RH90
RH90 index versus reference forest height for ICE-Sat-2/ATLAS day and night data. The
index versus reference forest height for ICE-Sat-2/ATLAS day and night data. The results
results show that the data collected at night have a strong ability to estimate forest height.
show that the data collected at night have a strong ability to estimate forest height. In
In contrast, ICESat-2 data collected during the day are unsuitable for forest height retrieval
contrast, ICESat-2 data collected during the day are unsuitable for forest height retrieval
because of their low accuracy. Therefore, the inversion model is constructed by using the
because of their low
night acquisition accuracy. Therefore, the inversion model is constructed by using the
data.
night acquisition data.

Sustainability 2024, 16, x FOR PEER REVIEW 15 of 22

Figure 6. To assess the accuracy of ICESat-2 ATL08 forest canopy height indicators and forest canopy
Figure 6. To assess the accuracy of ICESat-2 ATL08 forest canopy height indicators and forest canopy
heights of reference forests (Dashed line is baseline, red is fit line).
heights of reference forests (Dashed line is baseline, red is fit line).

Figure7.7.
Figure Scatter
Scatter plot
plot of ICESat-2
of ICESat-2 ATL08
ATL08 forest
forest canopy
canopy height
height (RH90)
(RH90) versus
versus airborne
airborne LiDAR LiDAR
at ref-at
reference forest height (RH90) at different collection times (Dashed line is baseline, red is fit
erence forest height (RH90) at different collection times (Dashed line is baseline, red is fit line). line).

3.3. Utilizing the Random Forest Regression Model for Forest Canopy Height Retrieval
After preprocessing different space-borne lidar data products in the study area, we
use a random forest regression algorithm to construct a forest canopy height-retrieval
model for different space-borne lidar data products in the study area.
 Figure 7. Scatter plot of ICESat-2 ATL08 forest canopy height (RH90) versus airborne LiDAR at ref-
Sustainability 2024, 16, 1735 14 of 21
erence forest height (RH90) at different collection times (Dashed line is baseline, red is fit line).

3.3. Utilizing the Random Forest Regression Model for Forest Canopy Height Retrieval
3.3. Utilizing the Random Forest Regression Model for Forest Canopy Height Retrieval
After preprocessing different space-borne lidar data products in the study area, we
use aAfter preprocessing
random differentalgorithm
forest regression space-borne to lidar data products
construct a forest in the study
canopy area, we use
height-retrieval
a random forest regression algorithm to construct a forest
model for different space-borne lidar data products in the study area. canopy height-retrieval model
for different space-borne lidar data products in the study area.
Firstly, 5398 valid laser footprints were screened from GEDI data. To optimize the
Firstly,
efficiency 5398
and valid laser
accuracy footprints
of model were
training, screened
refer to Tablesfrom
6 andGEDI data. To above
7 mentioned optimizefor the
the
efficiency and accuracy of model training, refer to Tables 6 and 7 mentioned
selected model input characteristic variables and parameters. In the experimental area, above for
the selected model input characteristic variables and parameters.
there are two models based on single data sources, the GEDI model and Landsat 9 model, In the experimental
area,there
and thereisare two models
a model based on based on single data
a combination sources,
of multiple thesources,
data GEDI model and and
the GEDI Landsat
Land-9
model,
sat and there
9 model. is a modelresults
The inversion based areon ashown
combination
in Figure of multiple datausing
8. In models sources, the GEDI
a single data
and Landsat 9 model. The inversion results are shown in Figure 8. In models using a
source, GEDI outperformed Landsat 9 with an increase of 0.07 in R, a decrease of 0.69 in
single data source, GEDI outperformed Landsat 9 with an increase of 0.07 in R, a decrease
MAE, and a decrease of 0.51 m in RMSE. This is because GEDI data provide relative
of 0.69 in MAE, and a decrease of 0.51 m in RMSE. This is because GEDI data provide
heights of forest canopy structure parameters that were more important than Landsat 9
relative heights of forest canopy structure parameters that were more important than
feature variables in the previously performed variable importance analysis. Among mod-
Landsat 9 feature variables in the previously performed variable importance analysis.
els that combine data sources, we observe significant accuracy of the GEDI, Landsat 9, and
Among models that combine data sources, we observe significant accuracy of the GEDI,
SRTM combined models. Compared to the least accurate Landsat 9 model (R = 0.71, MAE
Landsat 9, and SRTM combined models. Compared to the least accurate Landsat 9 model
= 3.45 m, RMSE = 4.86 m, and rRMSE = 22.12%), combined model based on GEDI, Landsat
(R = 0.71, MAE = 3.45 m, RMSE = 4.86 m, and rRMSE = 22.12%), combined model based
9, and SRTM data (R = 0.92, MAE = 1.91 m, RMSE = 2.78 m, and rRMSE = 12.64%) R accu-
on GEDI, Landsat 9, and SRTM data (R = 0.92, MAE = 1.91 m, RMSE = 2.78 m, and
racy increased by 29.58%, MAE accuracy increased by 44.64%, RMSE accuracy increased
rRMSE = 12.64%) R accuracy increased by 29.58%, MAE accuracy increased by 44.64%,
by 42.80%, and rRMSE accuracy increased by 42.86%. Compared with the optimal model
RMSE accuracy increased by 42.80%, and rRMSE accuracy increased by 42.86%. Compared
in thethe
with literature
optimal[43], the in
model MAEthe accuracy
literatureof thethe
[43], combined model based
MAE accuracy of theon GEDI, Landsat
combined model
9, and SRTM data in this study is improved by 13.18%, and
based on GEDI, Landsat 9, and SRTM data in this study is improved by 13.18%, RMSE accuracy is and
improved
RMSE
by 10.61%. Overall, accuracy has been greatly improved.
accuracy is improved by 10.61%. Overall, accuracy has been greatly improved.

Figure 8. Retrieval of canopy height with the random forest regressing model: (a) GEDI L2A;
Figure 8. Retrieval of canopy height with the random forest regressing model: (a) GEDI L2A; (b)
(b) Landsat 9; and (c) the combined GEDI and Landsat 9 model (Dashed line is baseline, red is
Landsat 9; and (c) the combined GEDI and Landsat 9 model (Dashed line is baseline, red is fit line).
fit line).

Secondly, 1690 valid laser footprints were screened from ICESat-2/ATLAS data. To
optimize the efficiency and accuracy of model training, refer to Tables 6 and 7 mentioned
above for the selected model input characteristic variables and parameters. In the experi-
mental area, there are two models based on single data sources, the ICESat-2 model and
Landsat 9 model, and one model based on a combination of multiple data sources, the
ICESat-2 and Landsat 9 model. The inversion results are shown in Figure 9. ICESat-2
outperformed Landsat 9 in models using a single data source with an increase of 0.24 in R,
a decrease of 0.88 in MAE, and a decrease of 0.98 m in RMSE. This is because ICESat-2 data
provide relative heights of forest canopy structure parameters that were more important
than Landsat 9 feature variables in the previously performed variable importance analysis.
Among models that combine data sources, we observe significant accuracy of ICESat-2,
Landsat 9, and SRTM combined models. Compared to the least accurate Landsat 9 model
(R = 0.46, MAE = 3.42 m, RMSE = 5.02 m, and rRMSE = 22.21%), combined model based
on ICESat-2, Landsat 9, and SRTM data (R = 0.89, MAE = 1.84 m, RMSE = 2.54 m, and
rRMSE = 10.75%) R accuracy increased by 93.48%, MAE accuracy increased by 46.20%,
 than Landsat 9 feature variables in the previously performed variable importance analy-
sis. Among models that combine data sources, we observe significant accuracy of ICESat-
2, Landsat 9, and SRTM combined models. Compared to the least accurate Landsat 9
model (R = 0.46, MAE = 3.42 m, RMSE = 5.02 m, and rRMSE = 22.21%), combined model
based on ICESat-2, Landsat 9, and SRTM data (R = 0.89, MAE = 1.84 m, RMSE = 2.54 m,
Sustainability 2024, 16, 1735 15 of 21
and rRMSE = 10.75%) R accuracy increased by 93.48%, MAE accuracy increased by 46.20%,
RMSE accuracy increased by 49.40%, and rRMSE accuracy increased by 51.60%. Com-
pared with the optimal model in the literature [43], the MAE accuracy of the combined
RMSE accuracy increased by 49.40%, and rRMSE accuracy increased by 51.60%. Compared
model based on ICESat-2, Landsat 9, and SRTM data in this study is improved by 16.36%,
with the optimal model in the literature [43], the MAE accuracy of the combined model
and the RMSE accuracy is improved by 18.33%. Overall, accuracy has been greatly im-
based on ICESat-2, Landsat 9, and SRTM data in this study is improved by 16.36%, and the
proved.
RMSE accuracy is improved by 18.33%. Overall, accuracy has been greatly improved.

Figure 9. Retrieval of canopy height with the random forest regressing model: (a) ICESat-2/ATLAS;
Figure 9. Retrieval of canopy height with the random forest regressing model: (a) ICESat-2/ATLAS;
(b) Landsat 9; and (c) the combined ICESat-2 and Landsat 9 model (Dashed line is baseline, red is
(b) Landsat 9; and (c) the combined ICESat-2 and Landsat 9 model (Dashed line is baseline, red is
fit line).
fit line).

4. Discussion
4. Discussion
We analyzed and evaluated different space-borne lidar data products against reference
canopyWeheights
analyzed andpercentile
(90th evaluatedofdifferent
airbornespace-borne
radar canopylidar data
height products
data). against
For GEDI refer-
data, we
ence canopy heights (90th percentile of airborne radar canopy height
compared the reference canopy height to the data generated using the default algorithm data). For GEDI data,
we compared
obtained fromthethereference canopy return
GEDI waveform height metrics.
to the dataWegenerated
analyzed sixusing the default
relative height algo-
(RH)
rithm obtained from the GEDI waveform return metrics. We
measurements from RH90 to RH100 and found that RH98 had the highest correlationanalyzed six relative height
with
(RH) measurements
reference canopy height.from RH90 to RH100
This finding and foundwith
is consistent that RH98 had theofhighest
the findings the studycorrela-
[76].
tion with reference
However, contrarycanopy height.of
to the results This finding
Potapov etisal.
consistent
[38], RH95 with
wastheidentified
findings of bythe study
Potapov
[76].
et al. However,
as the most contrary to the results
representative measure of Potapov et al. [38],
of true forest RH95
height. was identified
A possible by Pota-
explanation for
pov et al. as the most
this discrepancy is thatrepresentative
Potapov et al.’smeasure of true forest
study included height.
not only A possible
forests but alsoexplanation
other land
for thistypes
cover discrepancy
such as is that Potapov
shrubs, et al.’s
grasslands, study included
wetlands, cropland, notand
only forestsinbut
tundra also
their other
canopy
land cover types such as shrubs, grasslands, wetlands, cropland,
height estimates, resulting in an underestimation of forest height. GEDI L2A versionand tundra in their can-2
opy height estimates, resulting in an underestimation of forest height.
product data provides six different algorithms to improve the accuracy of GEDI L2A data GEDI L2A version
2byproduct
adjusting data provides
three six different
elements algorithms
of the waveform to improve
signal the accuracy
(smoothing width, start of GEDI L2A
threshold,
data
and endby adjusting
threshold)three
[75].elements
Comparing of the
thewaveform
RH98 datasignal (smoothing
with reference width,height
canopy start thresh-
data, it
old, and end
is found thatthreshold)
Algorithm[75]. Comparing
4 and Algorithm the5RH98
have data with reference
insufficient accuracy canopy heightheight
in canopy data,
it is found while
inversion, that Algorithm
Algorithm42and hasAlgorithm
the highest5accuracy.
have insufficient accuracy
Table 5 shows thatinAlgorithm
canopy height
4 has
the highest waveform start and end thresholds, while Algorithm 5 has the lowest waveform
start and end thresholds. Therefore, it can be inferred that both too-short and too-long
waveform lengths will result in reduced accuracy of canopy height retrieval.
For ICESat-2/ATLAS data, we extracted five forest height indicators for each 25 m
band from signal photons obtained by matching ATL03 to corresponding ATL08 data
products and validated them against reference canopy heights, finding the highest corre-
lation between RH90 measurements and reference canopy heights. We then assessed the
correlation between RH90 indicators and reference canopy heights in ICESat-2/ATLAS
data collected during the daytime and at night in the study area. The results show that
the data collected at night have a strong ability to estimate forest height. In contrast, the
ICESat-2 data collected during the day are unsuitable for forest height retrieval due to
their low accuracy, which may be caused by the fact that the data collected at night are not
affected by excessive solar noise [77].
Multivariate analysis was performed on the characteristic variables of different space-
borne lidar data products in the study area using a random forest model to determine
the relative importance of each variable and select the final set of variables that fit the
model. The results show that topographic factors are key variables for canopy height
 low accuracy, which may be caused by the fact that the data collected at night are not
affected by excessive solar noise [77].
Multivariate analysis was performed on the characteristic variables of different
space-borne lidar data products in the study area using a random forest model to deter-
Sustainability 2024, 16, 1735 mine the relative importance of each variable and select the final set of variables 16 that fit
of 21
the model. The results show that topographic factors are key variables for canopy height
inversion (see Figure 10). This finding is consistent with discussions in the literature [22].
inversion
Terrain and (seeslope
Figure
are10). This finding
identified is consistent
as the with discussions
primary influencers affecting in the literature
terrain [22].
height-re-
Terrain
trieval and slope are
accuracy, identified
while canopyasheight
the primary
emergesinfluencers affecting terrain
as the foremost height-retrieval
error factor in canopy
accuracy, while canopy
height retrieval. height
Therefore, emerges
terrain factorsaswere
the foremost
added in error factor in
subsequent canopy
model height
retrieval to
retrieval. Therefore, terrain factors were added in subsequent model retrieval
improve the accuracy. The Landsat satellite series does not allow for direct measurement to improve
the accuracy.
of forest Theheight.
canopy Landsat satellite relevant
However, series does not allow
feature for direct
variables can bemeasurement
obtained through of forest
mul-
canopy height. However, relevant feature variables can be obtained through
tispectral data, which, when combined with other related variables, can estimate forest multispectral
data,
canopywhich, when
height. combined
The accuracywith other
of the modelrelated variables,
hinges on the can estimate
presence andforest canopy
quality height.
of reference
The
data,accuracy of the model
the temporal hingescoherence
and spatial on the presence and quality
of optical of reference
data, and data, thescale
the geographic temporal
of its
and spatial coherence of optical data, and the geographic scale of its application
application [38]. The literature [69] confirms the significant role of vegetation indices [38]. The in
literature [69] confirms the significant role of vegetation indices in estimating
estimating forest canopy height, establishing a relationship between the two using Land- forest canopy
height, establishing
sat remote sensingaimagery.
relationship between
In this study,thewetwo using Landsat
conducted remote sensing
a multivariate analysis imagery.
using
In this study, we conducted a multivariate analysis using random forest
random forest to select the feature variables obtained from multispectral data. to select the feature
variables obtained from multispectral data.

(a) GEDI and Landsat 9 and SRTM (b) ICESat2 and Landsat 9 and SRTM

Figure 10. Importance score of each variable in the canopy height model based on different model
space-borne lidar products.

Six models were established based on different space-borne lidar data products using
a random forest regression algorithm to retrieve canopy height in the experimental area.
In GEDI data and ICESat-2/ATLAS data, the multivariate analysis in Figure 4 shows that
the importance scores for the characteristic variables of Lansat 9 data are relatively lower
than the relative heights of GEDI and ICESat-2. The accuracy of canopy height retrieved
using the Landsat 9 model is lower than the GEDI model and ICESat-2 model. This may be
because canopy height is the main error factor in canopy height retrieval, and multispectral
data cannot be measured directly.
In this study, we effectively utilized random forest regression models in machine
learning to estimate forest canopy height by integrating GEDI L2A data from space-borne
lidar, Landsat 9 data from optical remote sensing images and topographic factors, and
ICESat-2 data from space-borne lidar, Landsat 9 data from optical remote sensing images
and topographic factors, respectively. The integrated inversion results of multi-source
remote sensing data are satisfactory. Compared with the results of tree height inversion
using the BP neural network for multi-source remote sensing data integration in the
literature [43], the inversion accuracy of two multi-source remote sensing data integration
Sustainability 2024, 16, 1735 17 of 21

in this study is improved by 10.61% and 18.33%, respectively. The results highlight the
superiority of forest canopy height inversion using the random forest regression model
in this study. However, the inherent uncertainty of the random forest regression model
makes it prone to overfitting at larger decision tree depths, while smaller depths may result
in underfitting. Therefore, it is necessary to further strengthen parameter optimization in
future research.

5. Conclusions
This study evaluated the effectiveness of multi-source remote sensing data integration
and machine learning methods to improve canopy height-retrieval capabilities. In boreal
temperate forest ecosystems, Landsat 9 data were combined with GEDI and ICESat-2 data
representing space-borne lidar, respectively, to generate forest canopy height inversion
models using random forest regression algorithms. In addition, the influence of various
factors on the retrieval accuracy of canopy height is analyzed.
The study based on data from different space-borne lidar products shows that (1) the
accuracy of canopy height estimation from ICESat-2 and GEDI data varies with data
acquisition schemes. Among the six algorithms for GEDI L2A second edition data, the a2
algorithm estimates CHM with the best accuracy. The ICESat-2 ATL08 sixth edition data
acquisition accuracy at night is better than the data collected during the day. (2) A random
forest algorithm was used to rank the importance of the characteristic variables. It was
found that terrain factors were important factors affecting the accuracy of canopy height
retrieval, and canopy height had the greatest influence on canopy height retrieval. (3) The
random forest model of multi-source remote sensing data integration is always better
than the single-source model, and the Landsat 9 model is lower than the GEDI model and
ICESat-2 model. The RMSE of GEDI and Landsat 9 and SRTM is 2.78 m. Compared with
the GEDI model, the precision of the GEDI and Landsat 9 and SRTM models is improved
by 35.95% and 42.80%, respectively. The RMSE of ICESat-2 and Landsat 9 and SRTM is
2.54 m. Compared with the ICESat-2 model, the precision of ICESat-2 and Landsat 9 and
SRTM is 37.13% higher, and compared with the Landsat 9 model, the precision of ICESat-2
and Landsat 9 and SRTM is 49.40%.
In summary, this study is a preliminary exploration of the integration method of
optical remote sensing Landsat 9 data with space-borne lidar products GEDI and ICE-Sat-2
of different models. The results will provide valuable insights into the practical perfor-
mance and benefits of future use of Landsat 9 data in combination with other data for
detailed forest canopy height retrieval, highlighting the enormous benefits of combining
multispectral data and topographical factors when analyzing forest ecosystems. They
provide important insights for future efforts in this field, highlighting in particular the
effectiveness of combining different space-borne lidar products and Landsat 9 data sepa-
rately with advanced machine learning algorithms. Nevertheless, the study acknowledges
certain limitations, including limitations on site selection. Future research should focus on
broader, more diverse global site selection and data accumulation, ideally incorporating
GEDI and ICESat-2 data alongside other datasets like the Sentinel-3 and GaoFen-7 data.
This expanded approach would likely yield more comprehensive insights for regional-scale
forest monitoring.

Author Contributions: Conceptualization, W.Z. and Y.L.; methodology, W.Z. and Y.L.; software, Y.L.;
validation, N.H., X.Z. and Z.Z.; investigation, K.L. and Z.Q.; formal analysis, W.Z. and K.L.; resources,
Y.L.; data curation, Y.L.; writing—original draft preparation, Y.L. and W.Z.; writing—review and
editing, Y.L. and K.L.; visualization, Y.L.; supervision, W.Z.; project administration, W.Z.; funding
acquisition, W.Z. All authors have read and agreed to the published version of the manuscript.
Funding: This research was supported by the National Natural Science Foundation of China (grant
no. 42371441) and the scientific innovation program project by the Shanghai Committee of Science
and Technology (grant no. 20dz1206501).
Institutional Review Board Statement: Not applicable.
Sustainability 2024, 16, 1735 18 of 21

Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: The data used in this research for remote sensing are openly available
and free of charge. You can access the GEDI data at https://www.earthdata.nasa.gov (accessed on 12
October 2023). Landsat 9 data and SRTM data can be obtained through the USGS Earth Resources
Observation and Science (EROS) Center at https://earthexplorer.usgs.gov (accessed on 12 October
2023). Airborne Laser Scan data from the National Ecological Observatory Network (NEON) in
the United States can be accessed at https://data.neonscience.org/data-products/DP3.30024.001
(accessed on 12 October 2023).
Acknowledgments: All authors are grateful to NASA and NEON for providing the GEDI, SRTM,
and Landsat 9 data and to the editors and reviewers for their valuable comments.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Le Quéré, C.; Andrew, R.M.; Canadell, J.G.; Sitch, S.; Korsbakken, J.I.; Peters, G.P.; Manning, A.C.; Boden, T.A.; Tans, P.P.;
Houghton, R.A. Global carbon budget 2016. Earth Syst. Sci. Data 2016, 8, 605–649. [CrossRef]
2. Malhi, Y.; Franklin, J.; Seddon, N.; Solan, M.; Turner, M.G.; Field, C.B.; Knowlton, N. Climate change and ecosystems: Threats,
opportunities and solutions. Philos. Trans. R. Soc. B Biol. Sci. 2020, 375, 20190104. [CrossRef]
3. Meir, P.; Cox, P.; Grace, J. The influence of terrestrial ecosystems on climate. Trends Ecol. Evol. 2006, 21, 254–260. [CrossRef]
4. Rustad, L.E. The response of terrestrial ecosystems to global climate change: Towards an integrated approach. Sci. Total Environ.
2008, 404, 222–235. [CrossRef]
5. Sun, Y.; Wang, C.; Chen, H.Y.H.; Liu, Q.; Ge, B.; Tang, B. A global meta-analysis on the responses of C and N concentrations to
warming in terrestrial ecosystems. Catena 2022, 208, 105762. [CrossRef]
6. Wang, Y.-S.; Gu, J.-D. Ecological responses, adaptation and mechanisms of mangrove wetland ecosystem to global climate change
and anthropogenic activities. Int. Biodeterior. Biodegrad. 2021, 162, 105248. [CrossRef]
7. Baccini, A.; Walker, W.; Carvalho, L.; Farina, M.; Sulla-Menashe, D.; Houghton, R.A. Tropical forests are a net carbon source based
on aboveground measurements of gain and loss. Science 2017, 358, 230–234. [CrossRef]
8. Lagomasino, D.; Fatoyinbo, T.; Lee, S.-K.; Simard, M. High-resolution forest canopy height estimation in an African blue carbon
ecosystem. Remote Sens. Ecol. Conserv. 2015, 1, 51–60. [CrossRef] [PubMed]
9. Li, W.; Niu, Z.; Shang, R.; Qin, Y.; Wang, L.; Chen, H. High-resolution mapping of forest canopy height using machine learning
by coupling ICESat-2 LiDAR with Sentinel-1, Sentinel-2 and Landsat-8 data. Int. J. Appl. Earth Obs. Geoinf. 2020, 92, 102163.
[CrossRef]
10. Vaglio Laurin, G.; Ding, J.; Disney, M.; Bartholomeus, H.; Herold, M.; Papale, D.; Valentini, R. Tree height in tropical forest as
measured by different ground, proximal, and remote sensing instruments, and impacts on above ground biomass estimates. Int.
J. Appl. Earth Obs. Geoinf. 2019, 82, 101899. [CrossRef]
11. Swinfield, T.; Lindsell, J.A.; Williams, J.V.; Harrison, R.D.; Agustiono; Habibi; Gemita, E.; Schönlieb, C.B.; Coomes, D.A. Accurate
Measurement of Tropical Forest Canopy Heights and Aboveground Carbon Using Structure From Motion. Remote Sens. 2019, 11,
928. [CrossRef]
12. Liu, C.; Wang, S. Estimating tree canopy height in densely forest-covered mountainous areas using Gedi spaceborne full-waveform
data. ISPRS Ann. Photogramm. Remote Sens. Spat. Inf. Sci. 2022, V-1-2022, 25–32. [CrossRef]
13. Wang, Q.; Ni-Meister, W. Forest Canopy Height and Gaps from Multiangular BRDF, Assessed with Airborne LiDAR Data. Remote
Sens. 2019, 11, 2566. [CrossRef]
14. He, X.Y.; Ren, C.Y.; Chen, L.; Wang, Z.; Zheng, H. The Progress of Forest Ecosystems Monitoring with Remote Sensing Techniques.
Sci. Geogr. Sin 2018, 38, 997–1011. [CrossRef]
15. Roy, D.P.; Wulder, M.A.; Loveland, T.R.; Woodcock, C.E.; Allen, R.G.; Anderson, M.C.; Helder, D.; Irons, J.R.; Johnson, D.M.;
Kennedy, R.; et al. Landsat-8: Science and product vision for terrestrial global change research. Remote Sens. Environ. 2014, 145,
154–172. [CrossRef]
16. Showstack, R. Landsat 9 Satellite Continues Half-Century of Earth Observations: Eyes in the sky serve as a valuable tool for
stewardship. Biosci. J. 2022, 72, 226–232. [CrossRef]
17. Hemati, M.; Hasanlou, M.; Mahdianpari, M.; Mohammadimanesh, F. A Systematic Review of Landsat Data for Change Detection
Applications: 50 Years of Monitoring the Earth. Remote Sens. 2021, 13, 2869. [CrossRef]
18. Wulder, M.A.; Roy, D.P.; Radeloff, V.C.; Loveland, T.R.; Anderson, M.C.; Johnson, D.M.; Healey, S.; Zhu, Z.; Scambos, T.A.;
Pahlevan, N.; et al. Fifty years of Landsat science and impacts. Remote Sens. Environ. 2022, 280, 113195. [CrossRef]
19. Masek, J.G.; Wulder, M.A.; Markham, B.; McCorkel, J.; Crawford, C.J.; Storey, J.; Jenstrom, D.T. Landsat 9: Empowering open
science and applications through continuity. Remote Sens. Environ. 2020, 248, 111968. [CrossRef]
20. Healey, S.P.; Yang, Z.; Gorelick, N.; Ilyushchenko, S. Highly Local Model Calibration with a New GEDI LiDAR Asset on Google
Earth Engine Reduces Landsat Forest Height Signal Saturation. Remote Sens. 2020, 12, 2840. [CrossRef]
Sustainability 2024, 16, 1735 19 of 21

21. David, R.M.; Rosser, N.J.; Donoghue, D.N.M. Improving above ground biomass estimates of Southern Africa dryland forests by
combining Sentinel-1 SAR and Sentinel-2 multispectral imagery. Remote Sens. Environ. 2022, 282, 113232. [CrossRef]
22. Liu, A.; Cheng, X.; Chen, Z. Performance evaluation of GEDI and ICESat-2 laser altimeter data for terrain and canopy height
retrievals. Remote Sens. Environ. 2021, 264, 112571. [CrossRef]
23. Wake, S.; Ramos-Izquierdo, L.A.; Eegholm, B.; Dogoda, P.; Denny, Z.; Hersh, M.; Mulloney, M.; Thomes, W.J.; Ott, M.N.; Jakeman,
H. Optical system design and integration of the Global Ecosystem Dynamics Investigation Lidar. In Proceedings of the Infrared
Remote Sensing and Instrumentation XXVII, San Diego, CA, USA, 12–14 August 2019; pp. 99–111.
24. Dubayah, R.; Blair, J.B.; Goetz, S.; Fatoyinbo, L.; Hansen, M.; Healey, S.; Hofton, M.; Hurtt, G.; Kellner, J.; Luthcke, S.; et al. The
Global Ecosystem Dynamics Investigation: High-resolution laser ranging of the Earth’s forests and topography. Sci. Remote Sens.
2020, 1, 100002. [CrossRef]
25. Hancock, S.; Armston, J.; Hofton, M.; Sun, X.; Tang, H.; Duncanson, L.I.; Kellner, J.R.; Dubayah, R. The GEDI Simulator: A
Large-Footprint Waveform Lidar Simulator for Calibration and Validation of Spaceborne Missions. Earth Space Sci. 2019, 6,
294–310. [CrossRef] [PubMed]
26. Patterson, P.L.; Healey, S.P.; Ståhl, G.; Saarela, S.; Holm, S.; Andersen, H.-E.; Dubayah, R.O.; Duncanson, L.; Hancock, S.; Armston,
J.; et al. Statistical properties of hybrid estimators proposed for GEDI—NASA’s global ecosystem dynamics investigation. Environ.
Res. Lett. 2019, 14, 065007. [CrossRef]
27. Abdalati, W.; Zwally, H.J.; Bindschadler, R.; Csatho, B.; Farrell, S.L.; Fricker, H.A.; Harding, D.; Kwok, R.; Lefsky, M.; Markus, T.;
et al. The ICESat-2 Laser Altimetry Mission. Proc. IEEE 2010, 98, 735–751. [CrossRef]
28. Coyle, D.B.; Paul, R.S.; Furqan, L.C.; Erich, F.; Demetrios, P. The Global Ecosystem Dynamics Investigation (GEDI) Lidar laser
transmitter. In Proceedings of the Infrared Remote Sensing and Instrumentation XXVII, San Diego, CA, USA, 12–14 August 2019;
p. 111280L.
29. Adam, M.; Urbazaev, M.; Dubois, C.; Schmullius, C. Accuracy Assessment of GEDI Terrain Elevation and Canopy Height
Estimates in European Temperate Forests: Influence of Environmental and Acquisition Parameters. Remote Sens. 2020, 12, 3948.
[CrossRef]
30. Dhargay, S.; Lyell, C.S.; Brown, T.P.; Inbar, A.; Sheridan, G.J.; Lane, P.N.J. Performance of GEDI Space-Borne LiDAR for Quantifying
Structural Variation in the Temperate Forests of South-Eastern Australia. Remote Sens. 2022, 14, 3615. [CrossRef]
31. Lahssini, K.; Baghdadi, N.; le Maire, G.; Fayad, I. Influence of GEDI Acquisition and Processing Parameters on Canopy Height
Estimates over Tropical Forests. Remote Sens. 2022, 14, 6264. [CrossRef]
32. Caughlin, T.T.; Rifai, S.W.; Graves, S.J.; Asner, G.P.; Bohlman, S.A. Integrating LiDAR-derived tree height and Landsat satellite
reflectance to estimate forest regrowth in a tropical agricultural landscape. Remote Sens. Ecol. Conserv. 2016, 2, 190–203. [CrossRef]
33. Ceccherini, G.; Girardello, M.; Beck, P.S.A.; Migliavacca, M.; Duveiller, G.; Dubois, G.; Avitabile, V.; Battistella, L.; Barredo, J.I.;
Cescatti, A. Spaceborne LiDAR reveals the effectiveness of European Protected Areas in conserving forest height and vertical
structure. Commun. Earth Environ. 2023, 4, 97. [CrossRef]
34. Gu, C.; Clevers, J.G.P.W.; Liu, X.; Tian, X.; Li, Z.; Li, Z. Predicting forest height using the GOST, Landsat 7 ETM+, and airborne
LiDAR for sloping terrains in the Greater Khingan Mountains of China. ISPRS-J. Photogramm. Remote Sens. 2018, 137, 97–111.
[CrossRef]
35. Qi, W.; Lee, S.-K.; Hancock, S.; Luthcke, S.; Tang, H.; Armston, J.; Dubayah, R. Improved forest height estimation by fusion of
simulated GEDI Lidar data and TanDEM-X InSAR data. Remote Sens. Environ. 2019, 221, 621–634. [CrossRef]
36. Potapov, P.; Tyukavina, A.; Turubanova, S.; Talero, Y.; Hernandez-Serna, A.; Hansen, M.C.; Saah, D.; Tenneson, K.; Poortinga, A.;
Aekakkararungroj, A.; et al. Annual continuous fields of woody vegetation structure in the Lower Mekong region from 2000–2017
Landsat time-series. Remote Sens. Environ. 2019, 232, 111278. [CrossRef]
37. Tyukavina, A.; Baccini, A.; Hansen, M.C.; Potapov, P.V.; Stehman, S.V.; Houghton, R.A.; Krylov, A.M.; Turubanova, S.; Goetz,
S.J. Aboveground carbon loss in natural and managed tropical forests from 2000 to 2012. Environ. Res. Lett. 2015, 10, 074002.
[CrossRef]
38. Potapov, P.; Li, X.; Hernandez-Serna, A.; Tyukavina, A.; Hansen, M.C.; Kommareddy, A.; Pickens, A.; Turubanova, S.; Tang, H.;
Silva, C.E.; et al. Mapping global forest canopy height through integration of GEDI and Landsat data. Remote Sens. Environ. 2021,
253, 112165. [CrossRef]
39. Wang, J.; Liu, D.; Quiring, S.M.; Qin, R. Estimating canopy height change using machine learning by coupling WorldView-2
stereo imagery with Landsat-7 data. Int. J. Remote Sens. 2023, 44, 631–645. [CrossRef]
40. Ren, C.; Jiang, H.; Xi, Y.; Liu, P.; Li, H. Quantifying Temperate Forest Diversity by Integrating GEDI LiDAR and Multi-Temporal
Sentinel-2 Imagery. Remote Sens. 2023, 15, 375. [CrossRef]
41. Xi, Z.; Xu, H.; Xing, Y.; Gong, W.; Chen, G.; Yang, S. Forest Canopy Height Mapping by Synergizing ICESat-2, Sentinel-1,
Sentinel-2 and Topographic Information Based on Machine Learning Methods. Remote Sens. 2022, 14, 364. [CrossRef]
42. Guan, X.; Yang, X.; Yu, Y.; Pan, Y.; Dong, H.; Yang, T. Canopy-Height and Stand-Age Estimation in Northeast China at Sub-
Compartment Level Using Multi-Resource Remote Sensing Data. Remote Sens. 2023, 15, 3738. [CrossRef]
43. Zhu, W.; Yang, F.; Qiu, Z.; He, N.; Zhu, X.; Li, Y.; Xu, Y.; Lu, Z. Enhancing Forest Canopy Height Retrieval: Insights from
Integrated GEDI and Landsat Data Analysis. Sustainability 2023, 15, 10434. [CrossRef]
44. Raup, H.M. Some Problems in Ecological Theory and their Relation to Conservation. J. Anim. Ecol. 1964, 33, 19–28. [CrossRef]
Sustainability 2024, 16, 1735 20 of 21

45. SanClements, M.; Lee, R.H.; Ayres, E.D.; Goodman, K.; Jones, M.; Durden, D.; Thibault, K.; Zulueta, R.; Roberti, J.; Lunch, C.; et al.
Collaborating with NEON. Biosci. J. 2020, 70, 107. [CrossRef]
46. Hutsler, T.; Pricope, N.G.; Gao, P.; Rother, M.T. Detecting Woody Plants in Southern Arizona Using Data from the National
Ecological Observatory Network (NEON). Remote Sens. 2023, 15, 98. [CrossRef]
47. Johnson, B.R.; Kuester, M.A.; Kampe, T.U.; Keller, M. National ecological observatory network (NEON) airborne remote
measurements of vegetation canopy biochemistry and structure. In Proceedings of the 2010 IEEE International Geoscience and
Remote Sensing Symposium, Honolulu, HI, USA, 25–30 July 2010; pp. 2079–2082.
48. Kampe, T.; Johnson, B.; Kuester, M.; Keller, M. NEON: The first continental-scale ecological observatory with airborne remote
sensing of vegetation canopy biochemistry and structure. In Proceedings of the Remote Sensing and Modeling of Ecosystems for
Sustainability VI, San Diego, CA, USA, 2 August 2009.
49. Scholl, V.M.; Cattau, M.E.; Joseph, M.B.; Balch, J.K. Integrating National Ecological Observatory Network (NEON) Airborne
Remote Sensing and In-Situ Data for Optimal Tree Species Classification. Remote Sens. 2020, 12, 1414. [CrossRef]
50. Wang, C.; Jia, D.; Lei, S.; Numata, I.; Tian, L. Accuracy Assessment and Impact Factor Analysis of GEDI Leaf Area Index Product
in Temperate Forest. Remote Sens. 2023, 15, 1535. [CrossRef]
51. NEON (National Ecological Observatory Network). Elevation—LiDAR (DP3.30024.001), RELEASE-2023. Available online:
https://data.neonscience.org/data-products/DP3.30024.001/RELEASE-2023 (accessed on 12 October 2023).
52. Rishmawi, K.; Huang, C.; Schleeweis, K.; Zhan, X. Integration of VIIRS Observations with GEDI-Lidar Measurements to Monitor
Forest Structure Dynamics from 2013 to 2020 across the Conterminous United States. Remote Sens. 2022, 14, 2320. [CrossRef]
53. Dubayah, R.; Armston, J.; Healey, S.P.; Bruening, J.M.; Patterson, P.L.; Kellner, J.R.; Duncanson, L.; Saarela, S.; Ståhl, G.; Yang, Z.;
et al. GEDI launches a new era of biomass inference from space. Environ. Res. Lett. 2022, 17, 095001. [CrossRef]
54. Oliveira, P.V.; Zhang, X.; Peterson, B.; Ometto, J.P. Using simulated GEDI waveforms to evaluate the effects of beam sensitivity
and terrain slope on GEDI L2A relative height metrics over the Brazilian Amazon Forest. Sci. Remote Sens. 2023, 7, 100083.
[CrossRef]
55. Markus, T.; Neumann, T.; Martino, A.; Abdalati, W.; Brunt, K.; Csatho, B.; Farrell, S.; Fricker, H.; Gardner, A.; Harding, D.;
et al. The Ice, Cloud, and land Elevation Satellite-2 (ICESat-2): Science requirements, concept, and implementation. Remote Sens.
Environ. 2017, 190, 260–273. [CrossRef]
56. Lin, X.; Xu, M.; Cao, C.; Dang, Y.; Bashir, B.; Xie, B.; Huang, Z. Estimates of Forest Canopy Height Using a Combination of
ICESat-2/ATLAS Data and Stereo-Photogrammetry. Remote Sens. 2020, 12, 3649. [CrossRef]
57. Mulverhill, C.; Coops, N.C.; Hermosilla, T.; White, J.C.; Wulder, M.A. Evaluating ICESat-2 for monitoring, modeling, and update
of large area forest canopy height products. Remote Sens. Environ. 2022, 271, 112919. [CrossRef]
58. Neuenschwander, A.; Pitts, K. The ATL08 land and vegetation product for the ICESat-2 Mission. Remote Sens. Environ. 2019, 221,
247–259. [CrossRef]
59. Narine, L.L.; Popescu, S.C.; Malambo, L. Using ICESat-2 to Estimate and Map Forest Aboveground Biomass: A First Example.
Remote Sens. 2020, 12, 1824. [CrossRef]
60. Wang, C.; Jia, M.; Chen, N.; Wang, W. Long-Term Surface Water Dynamics Analysis Based on Landsat Imagery and the Google
Earth Engine Platform: A Case Study in the Middle Yangtze River Basin. Remote Sens. 2018, 10, 1635. [CrossRef]
61. Gerardo, R.; de Lima, I.P. Comparing the Capability of Sentinel-2 and Landsat 9 Imagery for Mapping Water and Sandbars in the
River Bed of the Lower Tagus River (Portugal). Remote Sens. 2023, 15, 1927. [CrossRef]
62. Jolliffe, I.T.; Cadima, J. Principal component analysis: A review and recent developments. Philos. Trans. R. Soc. A Math. Phys. Eng.
Sci. 2016, 374, 20150202. [CrossRef] [PubMed]
63. López-Serrano, P.M.; Cárdenas Domínguez, J.L.; Corral-Rivas, J.J.; Jiménez, E.; López-Sánchez, C.A.; Vega-Nieva, D.J. Modeling
of Aboveground Biomass with Landsat 8 OLI and Machine Learning in Temperate Forests. Forests 2020, 11, 11. [CrossRef]
64. Trier, Ø.D.; Salberg, A.-B.; Haarpaintner, J.; Aarsten, D.; Gobakken, T.; Næsset, E. Multi-sensor forest vegetation height mapping
methods for Tanzania. Eur. J. Remote Sens. 2018, 51, 587–606. [CrossRef]
65. Zhang, Y.; Wang, R. Estimation of aboveground biomass of vegetation based on landsat 8 OLI images. Heliyon 2022, 8, e11099.
[CrossRef]
66. Rennó, C.D.; Nobre, A.D.; Cuartas, L.A.; Soares, J.V.; Hodnett, M.G.; Tomasella, J.; Waterloo, M.J. HAND, a new terrain descriptor
using SRTM-DEM: Mapping terra-firme rainforest environments in Amazonia. Remote Sens. Environ. 2008, 112, 3469–3481.
[CrossRef]
67. Rodríguez, E.; Morris, C.S.; Belz, J.E. A Global Assessment of the SRTM Performance. Photogramm. Eng. Remote Sens. 2006, 72,
249–260. [CrossRef]
68. Yang, L.; Meng, X.; Zhang, X. SRTM DEM and its application advances. Int. J. Remote Sens. 2011, 32, 3875–3896. [CrossRef]
69. Pascual, C.; Cohen, W.; García-Abril, A.; Arroyo, L.A.; Valbuena, R.; Martí-Fernández, S.; Manzanera, J.A.; Hill, R.; Rosette, J.;
Suárez, J. Mean height and variability of height derived from lidar data and Landsat images relationship. In Proceedings of the
SilviLaser 2008, 8th International Conference on LiDAR Applications in Forest Assessment and Inventory, Edinburgh, UK, 17–19
September 2008; pp. 517–525.
70. Guerra-Hernández, J.; Narine, L.L.; Pascual, A.; Gonzalez-Ferreiro, E.; Botequim, B.; Malambo, L.; Neuenschwander, A.; Popescu,
S.C.; Godinho, S. Aboveground biomass mapping by integrating ICESat-2, SENTINEL-1, SENTINEL-2, ALOS2/PALSAR2, and
topographic information in Mediterranean forests. GISci. Remote Sens. 2022, 59, 1509–1533. [CrossRef]
Sustainability 2024, 16, 1735 21 of 21

71. Teoh, T.T.; Rong, Z. Regression. In Artificial Intelligence with Python; Teoh, T.T., Rong, Z., Eds.; Springer: Singapore, 2022;
pp. 163–181.
72. Bisong, E. Introduction to Scikit-learn. In Building Machine Learning and Deep Learning Models on Google Cloud Platform: A
Comprehensive Guide for Beginners; Bisong, E., Ed.; Apress: Berkeley, CA, USA, 2019; pp. 215–229.
73. Malakouti, S.M.; Menhaj, M.B.; Suratgar, A.A. The usage of 10-fold cross-validation and grid search to enhance ML methods
performance in solar farm power generation prediction. Clean. Eng. Technol. 2023, 15, 100664. [CrossRef]
74. Liu, X.W.; Long, Z.L.; Zhang, W.; Yang, L.M. Key feature space for predicting the glass-forming ability of amorphous alloys
revealed by gradient boosted decision trees model. J. Alloys Compd. 2022, 901, 163606. [CrossRef]
75. Liu, L.; Wang, C.; Nie, S.; Zhu, X.; Xi, X.; Wang, J. Analysis of the influence of different algorithms of GEDI L2A on the accuracy of
ground elevation and forest canopy height. J. Univ. Chin. Acad. Sci. 2021, 39, 502–511. [CrossRef]
76. Lang, N.; Kalischek, N.; Armston, J.; Schindler, K.; Dubayah, R.; Wegner, J.D. Global canopy height regression and uncertainty
estimation from GEDI LIDAR waveforms with deep ensembles. Remote Sens. Environ. 2022, 268, 112760. [CrossRef]
77. Neuenschwander, A.; Guenther, E.; White, J.C.; Duncanson, L.; Montesano, P. Validation of ICESat-2 terrain and canopy heights
in boreal forests. Remote Sens. Environ. 2020, 251, 112110. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.