Open Access
Add to BibSonomy
Abstract
The vertical profiles of the seawater optical properties and subsurface phytoplankton layer observed during an airborne lidar flight experiment carried out on 29 January 2021 in the coastal waters near Qionghai city were studied. We employed a hybrid inversion model combining the Klett and perturbation retrieval methods to estimate the seawater optical properties, while the vertical subsurface phytoplankton layer profiles were obtained by an adaptive evaluation. The airborne lidar data preprocessing scheme and inversion of the seawater optical properties were described in detail, and the effects of water environment parameters on the airborne lidar detection performance in coastal waters were discussed. The obtained seawater optical properties and phytoplankton layer profiles exhibit characteristic spatiotemporal distributions. The vertical stratification of seawater optical properties along a flight track from 19.19°N to 19.27°N is more pronounced than that from 19.27°N to 19.31°N. The subsurface phytoplankton layer appears along the flight track at water depths of 5–14 m with a thickness of 2–8.3 m. The high concentrations of chlorophyll, colored dissolved organic matter (CDOM), and suspended particulate matter (SPM) in coastal waters are the main factors leading to the shallower detection depth for airborne lidar. A 532 nm laser emission wavelength is more suitable than 486 nm for investigating coastal waters. The 532 nm receiving channel with 25 mrad receiving field of view achieves a better detection performance than that with 6 mrad. These results indicate that lidar technology has great potential for the wide-range and long-term monitoring of coastal waters.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Coastal waters account for only approximately 7% of the global ocean surface and only 0.5% of the total seawater volume. However, as the transition zone between terrestrial environments and the open ocean, coastal waters play important roles in terms of biogeochemical cycles, global air–sea CO2 exchanges, coastal habitats, fish species, and human activities [1–5]. Given the growing concerns about the degradation of coastal marine ecosystems caused by land-based activities, the long-term investigation and monitoring of coastal marine environments are highly warranted, and the outcomes should help environmental protection departments better formulate policies for coastal marine governance.
At present, passive satellite ocean color remote sensing techniques that measure reflected sunlight provide long-term ocean surface color products for a wide range of coastal applications [6–8], but these methods cannot provide color products consisting of day-night observations and vertical water profiles. With the rapid development of laser technology in recent decades, the light detection and ranging (lidar) method, an active remote sensing technology, has been widely applied in various oceanographic fields because of its remarkable advantages, namely, its ability to acquire continuous day-night observations, its high spatiotemporal resolution, and the penetrability of ocean waters by lidar-emitted laser pulses [9]; examples of shipborne and airborne lidar applications include research on seawater optical properties, the subsurface phytoplankton layer, fish schools, the speed of sound in seawater [10–22]. Recently, the spaceborne lidar platform has shown great potential in ocean observations [23,24]. Behrenfeld et al. used space-borne lidar data to study the global ocean carbon stocks, the boom-bust cycles of phytoplankton biomass, and the daily vertical migration of marine animals [25–27]. Lu et al. obtained the ocean subsurface optical properties from space-borne lidar measurements [28,29]. Dionisi et al. used space-borne lidar measurements to study the seasonal distributions of ocean particulate optical properties [30]. These studies have shown that oceanic lidar is expected to be one of the most important observation methods for ocean remote sensing. The vertical profiles of coastal waters obtained by lidar can help monitor the coastal marine environment. However, coastal waters are more sophisticated than open ocean waters due to the influences of terrestrial inorganic and organic carbon inputs, which also make it challenging for lidar systems to detect the profile properties of coastal waters. Consequently, featuring a more complicated composition than open ocean waters, coastal waters have a significant impact on laser transmission and thus affect lidar detection performance.
In this paper, the vertical profile distributions of seawater optical properties and the subsurface phytoplankton layer observed using airborne lidar in the coastal waters near Qionghai city are described. First, the airborne multiwavelength polarimetric lidar system and flight experiment are illustrated in detail, and the methods used to estimate the seawater optical properties and the subsurface phytoplankton layer are described. Subsequently, the preprocessing and inversion schemes applied to the airborne lidar data to extract the seawater optical profiles are explained, and the inversion results, namely, the vertical profile distributions of seawater optical properties and the subsurface phytoplankton layer, are further analyzed. Finally, we discuss the effects of different receiving fields of view and various water environment parameters on the airborne lidar detection performance in coastal waters and validate the airborne lidar-measured surface seawater optical properties using satellite remote sensing values. This study provides a new technical idea for quantitative inversion of different components in coastal waters based on multi-channel data at different wavelengths obtained from the airborne multiwavelength polarimetric ocean lidar system.
2. Materials and methods
2.1 Airborne multiwavelength polarimetric lidar system and flight experiment
The optical properties and subsurface phytoplankton layer in coastal waters were observed using an airborne multiwavelength polarimetric ocean lidar developed by the Shanghai Institute of Optics and Fine Mechanics (SIOM), Chinese Academy of Sciences. The airborne lidar system comprises a multiwavelength laser emission subsystem, a receiving subsystem, an optical processing and detecting subsystem, and a data acquisition and processing subsystem, as shown in Fig. 1. The designed parameters of the airborne lidar system are listed in Table 1. Specifically, the multiwavelength laser emission subsystem consists of a multiwavelength laser and a beam expander, where the former emits a pulse with four wavelengths of 1064 nm, 532 nm, 486 nm, and 355 nm, a repetition frequency of 5 kHz, a divergence angle of 6 mrad at 532 nm and 486 nm, and a divergence angle of 3 mrad at 1064 nm and 355 nm. A telescope receives the lidar backscatter signal that is transmitted to the optical processing and detecting subsystem with an aperture of 200 mm. The optical processing and detecting subsystem consists of eight focusing lenses (FL1–FL8), a split field lens (SFL), two collimating lenses (CL1 and CL2), four beam splitters (BS1–BS4), a transmitter mirror (TM), six filters (F1–F6), an avalanche photodiode (APD), six photomultiplier tubes (PMT1–PMT6), and a polarization beam splitter (PBS). The laser spot broadening effect increases with the increasing laser transmission distance in the water body. The receiving field of view is divided into a small central field of view and a large marginal field of view by the SFL. The small central and large marginal fields of view are used to receive lidar backscatter signals from shallow water with small laser spot broadening and deep water with large laser spot broadening, respectively, which increases the dynamic detection range of the lidar system. In the channel with a receiving field of view of 25 mrad, the backscattered signals at 486 nm and 532 nm are received by PMT1 and PMT2, respectively. In the channel with a receiving field of view of 2.5 mrad, the backscattered signals at 1064 nm are received by APD. In the channel with a receiving field of view of 6 mrad, the backscattered signals at 355 nm and 532 nm are received by PMT3 and PMT4 (PMT array), respectively, and the perpendicular and parallel components of the signal at 486 nm are received by PMT5 and PMT6, respectively. The data acquisition and processing subsystem include three four-channel data acquisition cards and a computer; the electrical signals from the detectors are digitized with three four-channel analog-to-digital converters at a sampling rate of 1 GSps. Finally, the vertical profile of different components in the water body can be quantified using an inversion based on multi-channel data at different wavelengths obtained from the airborne multiwavelength polarimetric ocean lidar system.
Fig. 1. Schematic of the airborne multiwavelength polarimetric ocean lidar system. FL1–FL8: focusing lenses, SFL: split field lens, CL1–CL2: collimating lenses, BS1–BS4: beam splitters, F1–F6: filters, TM: transmitter mirror, PBS: polarization beam splitter, APD: avalanche photodiode, PMT1–PMT6: photomultiplier tubes, PC: computer.
Table 1. Airborne multiwavelength polarimetric ocean lidar system parameters
The flight experiment with the airborne multiwavelength polarimetric ocean lidar system took place on January 29, 2021. The aircraft flew approximately 300–600 m above the sea surface, and the detailed flight track in the context of the study region is shown in Fig. 2. The high-resolution geographic data in Fig. 2 are from ETOPO2 (the global 2 arc-minute ocean depth and land elevation dataset) and the Global Self-consistent, Hierarchical, High-resolution Geography (GSHHG) database [31,32]. The upper left panel shows a map of the South China Sea (SCS) with the flight track shown in red, while the bottom panel shows a partially enlarged map specifying that the flight experiment took place over the coastal waters near Qionghai city. The dotted line in the bottom panel shows the depth contour of the SCS.
2.2 Method for estimating the seawater optical properties
The airborne multiwavelength polarimetric ocean lidar system emits laser pulses into the seawater that must first be transmitted through the atmosphere and the air–sea interface. Then, the backscattered signal generated by water molecules and suspended particles are received by a telescope with an aperture of 200 mm. The method for estimating the seawater optical properties and the subsurface phytoplankton layer is based on the quasi-single scattering lidar equation. The depth-dependent lidar return signal can be described by the following lidar equation [11,23,33]:
Furthermore, we employ the perturbation retrieval (PR) method to retrieve ${\beta ^\pi }$, which assumes that ${\beta ^\pi }$ denotes the sum of volume scattering coefficients corresponding to a homogeneous water body that does not vary with depth and an inhomogeneous water body that varies with depth [10,36]. Hence, the volume scattering coefficient at a scattering angle of $\pi $ rad and the lidar return signal can be expressed as:
The particulate backscattering coefficient (${b_{bp}},\; {m^{ - 1}}$) is finally calculated as:
where the parameter χ is defined as the conversion factor that connects ${\beta ^\pi }({\lambda ,\; z} )$ and ${b_{bp}}({\lambda ,z} )$. In this study, the conversion factor χ is assumed to be 1.08 [37,38].2.3 Method for estimating the subsurface phytoplankton layer
The vertical distribution and structure of the subsurface phytoplankton layer obtained from the lidar backscatter signal are crucial for understanding global biogeochemical cycles and marine ecosystem structures [39,40]. The original lidar backscatter signal is composed primarily of the background signal generated by the attenuation of the water body and the backscattered signal of the marine phytoplankton layer. The range-corrected lidar return signal can thus be expressed as:
Furthermore, the backscattered signal intensity corresponding to the subsurface phytoplankton layer $S_{phy}^{\prime}({\lambda ,z} )$ can be obtained by subtracting the lidar background signal $S_{background}^{\prime}({\lambda ,z} )$ from $S^{\prime}({\lambda ,z} )$:
where $S_{background}^{\prime}({\lambda ,z} )$ can be obtained by linear fitting to a certain depth range of $S^{\prime}({\lambda ,z} )$. Some researchers use a fixed empirical value as a denoising threshold for further processing of $S_{phy}^{\prime}({\lambda ,z} )$ [18]. This procedure means that the accuracy of the inversion results depends entirely on the selection of the empirical value. Thus, we employ an adaptive evaluation method to further process the backscattered signal intensity corresponding to the subsurface phytoplankton layer $S_{phy}^{\prime}({\lambda ,z} )$. The standardized observations and denoising thresholds in the adaptive evaluation method are dynamically changed, which facilitates the precise inversion of the vertical distribution and structure of the subsurface phytoplankton layer. The standardized observations are defined as the distance far from the location to the scale, given by [41,42]:2.4 Method for estimating the uncertainty
Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) are used to evaluate the accuracy of the inversion results, given by:
where N is the number of sampling points, ${X_i}$ and ${X_{lidar,\; \; i}}$ are the satellite remote sensing values and the airborne lidar measurements, respectively.3. Results
3.1 Airborne lidar data preprocessing
As mentioned above, the airborne lidar system emits laser pulses that are first transmitted through the atmosphere and the air-sea interface before penetrating the seawater, and the detector receives not only the backscattered signal from the seawater but also the atmospheric backscattered signal, the signal reflected from the sea surface, and the background noise signal, as shown in Fig. 3(a). Therefore, it is necessary to initially preprocess the raw airborne lidar data, thereby isolating the original waveform data with a high signal-to-noise ratio (SNR) and subsequently yielding robust inversion results. First, we extracted the water signal from the original lidar waveform data based on the sea surface elevation obtained from the 1064 nm channel and averaged it for every 50 pulses within the waveform data, which can effectively reduce the effects of random noise and signal mutations on the inversion results. We then averaged the last 200 ns of the lidar pulse waveform and subtracted it from the lidar waveform signal, which eliminates the influence of the background noise signal on the lidar waveform signal. The lidar waveform signal decreased sharply with increasing depth when the lidar pulses just entered the seawater (within 3 m depth). The reasons for this phenomenon are the sea surface reflection and PMT transient response effects, and the water body backscattering is rarely detected by the receiver. Hence, it is necessary to execute a range correction, which multiplies the lidar waveform signal for each depth by the square of the depth value. The final preprocessing result is shown in Fig. 3(b), revealing that only the water signal of the 532 nm channel is retained in the waveform data due to the turbid water quality in the coastal water body and the relatively low emission energy of the signal at 486 nm. Assuming that each channel of the airborne multiwavelength polarimetric ocean lidar system obtains echo signals with a high SNR in the flight experiment. The elevation information of the sea surface can be obtained through the 1064 nm channel. The vertical profile distribution of the lidar attenuation coefficient and particulate backscattering coefficient at 532 nm can be obtained through the 532 nm channels. The vertical profile distribution of the lidar attenuation coefficient and particulate backscattering coefficient at 486 nm can be obtained through the 486 nm channels. The vertical profiles distribution of the chlorophyll concentration and Colored Dissolved Organic Matter (CDOM) absorption coefficient can be retrieved by combining channels of 532 nm and 355 nm.
Fig. 3. (a) Raw airborne lidar data with a receiving field of view of 25 mrad collected by PMT1 and PMT2. (b) Lidar data after preprocessing.
3.2 Example step-by-step inversion of the seawater optical profiles
The airborne lidar data were first preprocessed as described in Section 3.1. Here, the process of inverting the seawater optical profile is explained in detail based on the 532 nm channel data with a receiving field of view of 25 mrad, as shown in Figs. 4 and 5. Figure 4(a) shows the vertical profile distribution of the original signal intensity for the airborne lidar along the lidar flight track; the color bar reflects the variation in the lidar signal intensity. The overall lidar signal intensity decreases as the depth increases from 0 to 15 m due to the attenuation of the pulse energy as the laser is transmitted through the seawater. Figure 4(b) shows the vertical profile distribution of the lidar signal after the denoising and background correction. The dynamic range of the lidar signal intensity decreases from 250–650 to 0–400 between Figs. 4(a) and 4(b). As shown in Fig. 4(c), we next applied a range correction to the lidar signal by removing the lidar waveform data from the uppermost 3 m of seawater, effectively eliminating the influences of the sea surface reflection and PMT transient response on the inversion result. The seawater optical profiles were retrieved from the lidar signal after applying this range correction based on the inversion method described in Section 2.2, and the results are shown in Fig. 5. Figure 5(a) shows the vertical profile distribution of the lidar attenuation coefficient, revealing a strong attenuation layer at water depths of 2–8 m within 0–3 km along the lidar flight track, whereas this strong seawater attenuation layer gradually disappears within 3–4 km along the lidar flight track. Likewise, Fig. 5(b) shows the vertical profile distribution of the lidar volume scattering coefficient at a scattering angle of π rad. The variation range of ${\beta ^\pi }({532\; nm,z} )$ is $\mathrm{2\ \times 1}{\textrm{0}^{\textrm{ - 3}}}\mathrm{\;\ to\;\ 3\ \times 1}{\textrm{0}^{\textrm{ - 3}}}{\; }{\textrm{m}^{\textrm{ - 1}}}\cdot s{\textrm{r}^{\textrm{ - 1}}}$ at water depths of 2–8 m within 0–3 km along the lidar flight track, while the value of ${\beta ^\pi }({532\; nm,z} )$ is less than $\mathrm{2\ \times 1}{\textrm{0}^{\textrm{ - 3}}}{\; }{\textrm{m}^{\textrm{ - 1}}}\cdot {\; \textrm{s}}{\textrm{r}^{\textrm{ - 1}}}$ within 3–4 km along the lidar flight track. Figure. 5(c) shows the vertical profile distribution of the particulate backscatter coefficient. The variation range of ${b_{bp}}({532\; nm,z} )$ is 0.01 to 0.017 ${m^{ - 1}}$ at water depths of 2–8 m within 0–3 km along the lidar flight track, while the value of ${b_{bp}}({532\; nm,z} )$ is less than 0.01 ${m^{ - 1}}$ within 3–4 km along the lidar flight track. The results in Fig. 5 indicate that the phytoplankton scattering layer is distributed mainly at water depths of 2–8 m.
Fig. 4. Step-by-step example of a lidar inversion of the seawater optical profiles along the lidar flight track. (a) Original vertical profile signal received by the airborne lidar. (b) Denoising and background correction of the original lidar signal. (c) Range-corrected lidar signal.
Fig. 5. Step-by-step example of a lidar inversion of the seawater optical profiles along the lidar flight track. Vertical profile distributions of the lidar attenuation coefficient (a), lidar volume scattering coefficient at a scattering angle of $\pi $ rad (b), and particulate backscatter coefficient (c).
3.3 Lidar-measured seawater optical profiles in the coastal waters near Qionghai city
Based on the abovementioned preprocessing scheme for airborne lidar data and the step-by-step example of a lidar inversion of the seawater optical profiles along the lidar flight track, the airborne lidar data of the coastal waters near Qionghai city were processed. The turbid water quality of the coastal waters in this region caused the airborne lidar signal to be unstable. The intensity of the lidar waveform signal changes drastically, and the backscattered signal of the water body cannot be maintained all the time. Accordingly, we selected the airborne lidar flight track data from the relatively stable 532 nm channel with a receiving field of view of 6 mrad because the backscattered signal of the water body in this channel remains present, and the inversion results are shown in Fig. 6. Figure 6(a) shows the three-dimensional vertical profile distribution of the lidar attenuation coefficient for a flight track spanning approximately 24 km. The variation range of the lidar attenuation coefficient is 0.2 to 0.4 ${m^{ - 1}}$ at depths of 0–6 m along the flight track of 19.19°N to 19.27°N, while the lidar attenuation coefficient fluctuates from 0.4 to 0.6 ${m^{ - 1}}$ at water depths of 6–12 m. A strong attenuation layer appears at water depths of 6-12 m and may continue to extend below 12 m along the flight track from 19.19°N to 19.27°N. When the flight track latitude ranges from 19.27°N to 19.31°N, the lidar attenuation coefficient fluctuates from 0.1 to 0.3 ${m^{ - 1}}$ at water depths of 0–6 m, while the lidar attenuation coefficient varies from 0.1 to 0.6 ${m^{ - 1}}$ at water depths of 6–12 m. Figure 6(b) shows the three-dimensional vertical profile distribution of the lidar volume scattering coefficient at a scattering angle of π rad for the same flight track spanning approximately 24 km. When the flight track latitude is within the range of 19.19°N to 19.27°N, the variation range of ${\beta ^\pi }$ is $\textrm{1}\mathrm{.5\ \times 1}{\textrm{0}^{\textrm{ - 3}}}{\; \textrm{to}\; 2}\mathrm{.5\ \times 1}{\textrm{0}^{\textrm{ - 3}}}{\; }{\textrm{m}^{\textrm{ - 1}}} {\; } \cdot \textrm{s}{\textrm{r}^{\textrm{ - 1}}}$ at water depths of 0–6 m, whereas the value of ${\beta ^\pi }$ fluctuates from $\mathrm{2\ \times 1}{\textrm{0}^{\textrm{ - 3}}}\mathrm{\;\ to\;\ 3\ \times 1}{\textrm{0}^{\textrm{ - 3}}}{\; }{\textrm{m}^{\textrm{ - 1}}}\cdot {\; \textrm{s}}{\textrm{r}^{\textrm{ - 1}}}$ at water depths of 6–12 m. In contrast, along the flight track from 19.27°N to 19.31°N, the variation range of ${\beta ^\pi }$ is $\mathrm{1\ \times 1}{\textrm{0}^{\textrm{ - 3}}}{\; \textrm{to}\; 2}\mathrm{.5\ \times 1}{\textrm{0}^{\textrm{ - 3}}}{\; }{\textrm{m}^{\textrm{ - 1}}}\cdot {\; \textrm{s}}{\textrm{r}^{\textrm{ - 1}}}$ at water depths of 0–6 m, whereas the value of ${\beta ^\pi }$ fluctuates from $\textrm{0}\mathrm{.5\ \times 1}{\textrm{0}^{\textrm{ - 3}}}\mathrm{\;\ to\;\ 3\ \times 1}{\textrm{0}^{\textrm{ - 3}}}{\; }{\textrm{m}^{\textrm{ - 1}}}\cdot {\; \textrm{s}}{\textrm{r}^{\textrm{ - 1}}}$ at water depths of 6–12 m. Figure 6(c) shows the three-dimensional vertical profile distribution of the particulate backscatter coefficient for a flight track spanning approximately 24 km. The variation range of ${b_{bp}}({532\; nm,z} )$ is 0.009 to 0.016 ${m^{ - 1}}$ at depths of 0–6 m along the flight track of 19.19°N to 19.27°N, while the value of ${b_{bp}}({532\; nm,z} )$ fluctuates from 0.015 to 0.019 ${m^{ - 1}}$ at water depths of 6–12 m. When the flight track latitude ranges from 19.27°N to 19.31°N, the value of ${b_{bp}}({532\; nm,z} )$ fluctuates from 0.0025 to 0.0075 ${m^{ - 1}}$ at water depths of 0–6 m, while the value of ${b_{bp}}({532\; nm,z} )$ varies from 0.0025 to 0.016 ${m^{ - 1}}$ at water depths of 6–12 m. These results in Fig. 6 indicate that the vertical stratification of seawater optical properties along a flight track from 19.19°N to 19.27°N is more pronounce than that from 19.27°N to 19.31°N; the subsurface phytoplankton layer has remarkable spatiotemporal distribution characteristics. Moreover, the phytoplankton layer is distributed mainly at depths of 6-12 m and may continue to extend below 12 m.
Fig. 6. Three-dimensional vertical profile distributions of the lidar attenuation coefficient (a), lidar volume scattering coefficient at a scattering angle of π rad (b), and particulate backscatter coefficient (c) near Qionghai city.
3.4 Vertical profile distribution of the subsurface phytoplankton layer in the coastal waters along the lidar flight track
The relatively stable 532 nm channel lidar data with a receiving field of view of 25 mrad were extracted from the lidar flight data. We then used the filtered lidar data to estimate the vertical profile distribution of the subsurface phytoplankton layer in the coastal waters near Qionghai city. The screened lidar data were first preprocessed following the procedure in Section 3.1 to guarantee the quality of the original lidar data and the accuracy of the inversion results. Figure 7(a) shows the natural logarithmic vertical profile distribution of the preprocessed airborne lidar data. The overall vertical trend of the lidar signal decreases as the water depth increases from 0 m to 20 m, which is caused by the water body attenuating the transmitted laser. In addition, the intensity of the lidar signal ranges from 3 to 4 at water depths of 0–10 m, whereas the approximate variation range of the lidar signal intensity is 2–3 at water depths of 10–15 m and 1–2 at water depths of 15–20 m. Figure 7(b) shows the vertical profile distribution of the lidar background signal obtained by linearly fitting the lidar data in Fig. 7(a). The lidar signal intensity at water depths of 8–12 m shown in Fig. 7(b) is significantly lower than that at water depths of 8–12 m shown in Fig. 7(a). Furthermore, the obtained subsurface phytoplankton layer is depicted by a single vertical lidar profile in Fig. 7(c), where the black curve is the range-corrected lidar backscatter signal $S^{\prime}$ and the blue line is the lidar background signal $S_{background}^{\prime}$. The backscattered signal intensity of the subsurface phytoplankton layer $S_{phy}^{\prime}$ can be obtained by subtracting $S_{background}^{\prime}$ from $S^{\prime}$ based on Eq. (10), retaining only positive values in the result of the difference, as shown by the red curve in Fig. 7(c). For a single vertical lidar profile in Fig. 7(c), the 1 ns sampling time resolution of the digitizer and pulse average processing produced a vertical resolution of approximately 0.11 m in water and a horizontal resolution of approximately 1 m for lidar inversion, respectively.
Fig. 7. (a) Natural logarithmic vertical profile distribution of the preprocessed airborne lidar data. (b) Vertical profile distribution of the airborne lidar data after linear regression. (c) Example of the obtained subsurface phytoplankton layer information based on a single vertical lidar profile.
The vertical profile distribution of $S_{phy}^{\prime}$ can be acquired by performing the steps in Fig. 7(c) for each vertical profile of the lidar signal along the lidar flight track. Subsequently, we employed an adaptive evaluation method based on Eqs. (11) and (12) to process the backscattered signal intensity corresponding to the subsurface phytoplankton layer. The subsurface phytoplankton layer is finally defined as the backscattering intensity $S_{phy}^{\prime}$ greater than the denoising threshold ${L_{Threshold}}$. ${L_{Threshold}}$ is not a fixed value but varies dynamically with $S_{phy}^{\prime}$ in the adaptive evaluation method, thereby ensuring the accuracy of the obtained vertical distribution and structure of the subsurface phytoplankton layer; the results are shown in Fig. 8. Figure 8(a) shows the vertical profile distribution of the subsurface phytoplankton layer in the coastal waters along the lidar flight track. The subsurface phytoplankton layer is distributed at water depths of 5–14 m. The reason for this phenomenon may be that light and nutrients are sufficiently supplied simultaneously at these water depths, which promotes the growth of phytoplankton. Figure 8(b) shows the depth (blue curve) and thickness (red curve) of the subsurface phytoplankton layer along the lidar flight track. The depth of the subsurface phytoplankton layer is defined as the depth corresponding to the maximum value of each lidar profile signal along the lidar flight track. The peak concentration of phytoplankton fluctuates in the range from approximately 7 m to 11 m along the flight track. In addition, the thickness of the subsurface phytoplankton layer can be acquired by subtracting the depth of the lower boundary from the depth of the upper boundary. Consequently, the thickness of the subsurface phytoplankton layer varies from 2 m to 8.3 m along the lidar flight track.
Fig. 8. Vertical profile distribution of the subsurface phytoplankton layer in the coastal waters along the lidar flight track (a). Corresponding depth and thickness of the subsurface phytoplankton layer on January 29, 2021 (b).
4. Discussion
4.1 Comparing lidar profile signals under the different receiving field of view channels
In Section 3, the vertical profile distribution of the seawater optical profile properties and the subsurface phytoplankton layer are retrieved based on a large number of single lidar profile signals with receiving fields of view of 6 mrad and 25 mrad at 532 nm. In this section, the two lidar profiles at position (110.6018°E, 19.1664°N) are compared to more intuitively illustrate the variability of the received field of view of 6 mrad and 25 mrad at 532 nm, as shown in Fig. 9(a). The red line is a single lidar profile signal with a received field of view of 6 mrad at 532 nm, which can be used to retrieve water profile information of approximately the first 10 m. The blue line is a single lidar profile signal with a received field of view of 25 mard at 532 nm, which can be used to retrieve water profile information over approximately 18 m. Furthermore, we simulated the effect of different receiving fields of view on the lidar detection performance, and the result is shown in Fig. 9(b). The measured and simulated results in Fig. 9 indicate that the receiving channel of 532 nm with a receiving field of view of 25 mrad achieves a better detection performance than that with a receiving field of view of 6 mrad.
Fig. 9. (a) Comparison of single lidar profile signals for different field of view channels at 532 nm. (b) Effect of different receiving fields of view on the lidar detection performance.
4.2 Effects of water environment parameters on the lidar detection performance in coastal waters
The properties of coastal waters are more sophisticated than those of open ocean waters because the former are more susceptible to human activities. The attenuation of transmitted laser pulses in open ocean waters is attributable predominantly to pure seawater and the chlorophyll concentration, whereas the primary factors responsible for such attenuation in coastal waters are pure seawater, the chlorophyll concentration, colored dissolved organic matter (CDOM), and suspended particulate matter (SPM). In this section, the effects of various water environment parameters on the lidar detection performance are discussed based on the 8-day products of merged multiple satellite platforms from January 25, 2021 to February 1, 2021 [43]. Figure 10(a) shows that the chlorophyll concentration at the sea surface in the SCS (18.75°∼19.75°N, 110.3°∼111.4°E) varies from 0.2 mg m-3 to 1.8 mg m-3, and it is apparent that the chlorophyll concentration in coastal waters is greater than that in open ocean waters. The area marked by the red rectangle is the sea area of the airborne lidar flight experiment; the sea surface chlorophyll concentration in this region ranges from 0.6 mg m-3 to 3 mg m-3. Likewise, Fig. 10(b) shows that the CDOM distribution at the sea surface in the SCS (18.75°∼19.75°N, 110.3°∼111.4°E) varies from 0.02 m-1 to 0.12 m-1, and the sea surface CDOM concentration in the sea area of the airborne lidar flight experiment ranges from 0.03 m-1 to 0.10 m-1. Figure 10(c) shows that the SPM distribution at the sea surface in the SCS (18.75°∼19.75°N, 110.3°∼111.4°E) varies from 0.4 g m-3 to 16 g m-3, and the sea surface SPM concentration in the sea area of the airborne lidar flight experiment ranges from 1 g m-3 to 14 g m-3. Furthermore, we simulated the airborne lidar SNR at position (110.9375°E, 19.3000°N) based on the lidar system parameters in Table 1 and the bio-optical model [44–46], and the result is shown in Fig. 10(d). The maximum detectable depths of the 532 nm (blue line) and 486 nm (orange line) with a receiving field of view of 25 mrad is 20.5 m and 7.8 m when the laser incident energies of 532 nm and 486 nm are 0.74 mJ and 0.1 mJ, respectively. The maximum detectable depth of the 486 nm (red line) with a receiving field of view of 25 mard is 17.8 m when the laser incident energy of 486 nm increases from 0.1 mJ to 0.74 mJ. The simulation results indicate that the chlorophyll, CDOM, and SPM in coastal waters are the main factors leading to the shallower detection depth of the airborne lidar. The laser emission wavelength of 532 nm is more suitable for investigating coastal waters than the wavelength of 486 nm. The lower laser incident energy at 486 nm in the flight experiment is the primary reason for the loss of lidar signal at 486 nm. This information will help direct our future improvement of the airborne lidar system.
Fig. 10. Distributions of the chlorophyll concentration (a), CDOM (b), and SPM (c) at the sea surface from satellite observations. The red rectangle is the sea area of the airborne lidar flight experiment. (d) Simulation of the SNR for airborne lidar based on the bio-optical model and lidar system parameters.
4.3 Evaluation of the airborne lidar-measured surface seawater optical properties
The vertical profile distributions of the seawater optical properties are retrieved in Sections 3.2 and 3.3. In this section, the airborne lidar-measured surface seawater optical properties at 532 nm obtained from Sections 3.2 and 3.3 are validated by ${K_D}$ and ${b_{bp}}$ of 532 nm from satellite observations. The 8-day products of merged multiple satellite platforms from January 25, 2021 to February 1, 2021 are selected due to the influence of cloud occlusion in the sea area of the airborne lidar flight experiment. Figure 11(a) shows that the ${K_D}$ distribution at the sea surface in the SCS (18.75°∼19.75°N, 110.3°∼111.4°E) varies from 0.07 m-1 to 0.15 m-1, and the sea surface ${K_D}$ in the sea area of the airborne lidar flight experiment ranges from 0.08 m-1 to 0.14 m-1. The horizontal resolution of satellite observations affects the number of matching points with lidar-measured surface seawater optical properties. Thus, we selected satellite observations at seven locations to evaluate the lidar observations. Figure 11(b) shows that the lidar retrieved lidar attenuation coefficient ${K_{lidar}}$ at 532 nm are compared with co-located satellite remote sensing diffuse attenuation coefficient ${K_D}$ at 532 nm, and the coefficient of determination r2, MAE, and RMSE are 0.7251, 0.0061 m-1, and 0.0012 m-1, respectively. Figure 11(c) shows the ${b_{bp}}$ distribution at the sea surface in the SCS (18.75°∼19.75°N, 110.3°∼111.4°E), which varies from $\mathrm{1\ \times 1}{\textrm{0}^{\textrm{ - 3}}}$ m-1 to $\mathrm{17\ \times 1}{\textrm{0}^{\textrm{ - 3}}}$ m-1; in addition, the variation range of the sea surface ${b_{bp}}$ is $\mathrm{2\ \times 1}{\textrm{0}^{\textrm{ - 3}}}$ m-1 to $\mathrm{8\ \times 1}{\textrm{0}^{\textrm{ - 3}}}$ m-1 in the experimental sea area. As shown in Fig. 11(d), the lidar retrieved particulate backscatter coefficient ${b_{bp}}$ at 532 nm are compared with co-located satellite remote sensing particulate backscatter coefficient ${b_{bp}}$ at 532 nm, and the coefficient of determination r2, MAE, and RMSE are 0.9105, $\textrm{0}\mathrm{.1057\ \times 1}{\textrm{0}^{\textrm{ - 3}}}$ m-1, and $\textrm{0}\mathrm{.1317\ \times 1}{\textrm{0}^{\textrm{ - 3}}}$ m-1, respectively. The results in Figs. 11(b) and 11(d) indicate that the airborne lidar-measured surface seawater optical properties at 532 nm have a good correlation with the ${K_D}$ and ${b_{bp}}$ of 532 nm from satellite observations.
Fig. 11. Distributions of the diffuse attenuation coefficient ${K_D}$ (a) and the particulate backscatter coefficient ${b_{bp}}$ (c) at 532 nm at the sea surface from satellite observations. The red rectangle is the sea area of the airborne lidar flight experiment. Airborne lidar-measured surface seawater optical properties at 532 nm are validated by ${K_D}$ (b) and ${b_{bp}}$ (d) of 532 nm from satellite observations.
5. Conclusions
In this research, the vertical profile distributions of seawater optical properties and the subsurface phytoplankton layer observed using airborne lidar in the coastal waters near Qionghai city were studied. We found that both the seawater optical properties and the subsurface phytoplankton layer exhibit characteristic spatiotemporal distributions along the airborne lidar flight track. The vertical stratification of seawater optical properties along a flight track from 19.19°N to 19.27°N is more pronounced than that from 19.27°N to 19.31°N. Moreover, the subsurface phytoplankton layer is distributed along the lidar flight track at water depths ranging from 5–14 m, and its thickness varies from 2 m to 8.3 m; the reason for this phenomenon may be that light and nutrients are sufficiently supplied simultaneously at water depths of 5–14 m, which promotes phytoplankton growth. In the lidar system, the receiving field of view is divided into a small central field of view and a large marginal field of view by the SFL, which effectively increases the dynamic detection range of the lidar system. The laser emission wavelength of 532 nm is more suitable for investigating coastal waters than the wavelength of 486 nm, and the receiving channel of 532 nm with a receiving field of view of 25 mrad achieves a better detection performance than that with a receiving field of view of 6 mrad. Additionally, the high concentrations of chlorophyll, CDOM, and SPM in coastal waters are the main factors leading to the shallower detection depth of the airborne lidar, and the lower laser incident energy at 486 nm in the flight experiment is the primary reason for the loss of lidar signal at 486 nm. This information will help direct our future improvement of the airborne lidar system. The airborne lidar-measured surface seawater optical properties at 532 nm have a good correlation with the ${K_D}$ and ${b_{bp}}$ of 532 nm from satellite observations. This study provides a new technical idea for quantitative inversion of different components in coastal waters based on multi-channel data at different wavelengths obtained from the airborne multiwavelength polarimetric ocean lidar system. These finding indicate that lidar technology has great potential for the wide-range and long-term monitoring of coastal waters, the outcomes of which can help environmental protection departments better formulate policies for coastal marine governance in the future.
Funding
Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou) (GML2019ZD0602); National Science and Technology Major Project (05-Y30B01-9001-19/20-2); National Key Research and Development Program of China (2016YFC1400902); Second Institute of Oceanography, State Oceanic Administration (QNYC1803); National Natural Science Foundation of China (41901305, 61991454); Natural Science Foundation of Zhejiang Province (LQ19D060003).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. N. Hoepffner and G. Zibordi, “Remote Sensing of Coastal Waters,” in Encyclopedia of Ocean Sciences (Academic, 2009), pp. 732–741.
2. J. E. Bauer, W.-J. Cai, P. A. Raymond, T. S. Bianchi, C. S. Hopkinson, and P. A. G. Regnier, “The changing carbon cycle of the coastal ocean,” Nature 504(7478), 61–70 (2013). [CrossRef]
3. S. R. Phinn, C. Menges, G. J. E. Hill, and M. Stanford, “Optimizing Remotely Sensed Solutions for Monitoring, Modeling, and Managing Coastal Environments,” Remote Sens. Environ. 73(2), 117–132 (2000). [CrossRef]
4. A. Kubo and J. Kanda, “Coastal urbanization alters carbon cycling in Tokyo Bay,” Sci Rep 10(1), 20413 (2020). [CrossRef]
5. J. P. Gattuso, M. Frankignoulle, and R. Wollast, “Carbon and carbonate metabolism in coastal aquatic ecosystems,” Annu. Rev. Ecol. Syst. 29(1), 405–434 (1998). [CrossRef]
6. G. Zheng and P. M. DiGiacomo, “Remote sensing of chlorophyll-a in coastal waters based on the light absorption coefficient of phytoplankton,” Remote Sens. Environ. 201, 331–341 (2017). [CrossRef]
7. H. Loisel, A. Mangin, V. Vantrepotte, D. Dessailly, D. Ngoc Dinh, P. Garnesson, S. Ouillon, J.-P. Lefebvre, X. Mériaux, and T. Minh Phan, “Variability of suspended particulate matter concentration in coastal waters under the Mekong's influence from ocean color (MERIS) remote sensing over the last decade,” Remote Sens. Environ. 150, 218–230 (2014). [CrossRef]
8. J. Chen, M. Zhang, T. Cui, and Z. Wen, “A Review of Some Important Technical Problems in Respect of Satellite Remote Sensing of Chlorophyll-a Concentration in Coastal Waters,” IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing 6(5), 2275–2289 (2013). [CrossRef]
9. C. Jamet, A. Ibrahim, Z. Ahmad, F. Angelini, M. Babin, M. J. Behrenfeld, E. Boss, B. Cairns, J. Churnside, J. Chowdhary, A. B. Davis, D. Dionisi, L. Duforêt-Gaurier, B. Franz, R. Frouin, M. Gao, D. Gray, O. Hasekamp, X. He, C. Hostetler, O. V. Kalashnikova, K. Knobelspiesse, L. Lacour, H. Loisel, V. Martins, E. Rehm, L. Remer, I. Sanhaj, K. Stamnes, S. Stamnes, S. Victori, J. Werdell, and P.-W. Zhai, “Going Beyond Standard Ocean Color Observations: Lidar and Polarimetry,” Front. Mar. Sci. 6(251), 1–24 (2019). [CrossRef]
10. P. Chen, C. Jamet, Z. Zhang, Y. He, Z. Mao, D. Pan, T. Wang, D. Liu, and D. Yuan, “Vertical distribution of subsurface phytoplankton layer in South China Sea using airborne lidar,” Remote Sens. Environ. 263, 112567 (2021). [CrossRef]
11. D. Yuan, P. Chen, Z. Mao, X. Zhang, Z. Zhang, C. Xie, C. Zhong, and Z. Qian, “Ocean mixed layer depth estimation using airborne Brillouin scattering lidar: simulation and model,” Appl. Opt. 60(36), 11180–11188 (2021). [CrossRef]
12. J. H. Churnside, R. D. Marchbanks, and N. Marshall, “Airborne Lidar Observations of a Spring Phytoplankton Bloom in the Western Arctic Ocean,” Remote Sens. 13(13), 2512 (2021). [CrossRef]
13. J. A. Schulien, A. Della Penna, P. Gaube, A. P. Chase, N. Haëntjens, J. R. Graff, J. W. Hair, C. A. Hostetler, A. J. Scarino, E. S. Boss, L. Karp-Boss, and M. J. Behrenfeld, “Shifts in Phytoplankton Community Structure Across an Anticyclonic Eddy Revealed From High Spectral Resolution Lidar Scattering Measurements,” Frontiers in Marine Science 7(493), 1–13 (2020). [CrossRef]
14. K. Li, Y. He, J. Ma, Z. Jiang, C. Hou, W. Chen, X. Zhu, P. Chen, J. Tang, S. Wu, F. Liu, Y. Luo, Y. Zhang, and Y. Chen, “A Dual-Wavelength Ocean Lidar for Vertical Profiling of Oceanic Backscatter and Attenuation,” Remote Sens. 12(17), 2844 (2020). [CrossRef]
15. P. Chen, Z. Mao, Z. Zhang, H. Liu, and D. Pan, “Detecting subsurface phytoplankton layer in Qiandao Lake using shipborne lidar,” Opt. Express 28(1), 558–569 (2020). [CrossRef]
16. D. Liu, P. Chen, H. Che, Z. Liu, Q. Liu, Q. Song, S. Chen, P. Xu, Y. Zhou, W. Chen, B. Han, X. Zhu, Y. He, Z. Mao, and C. Le, “Lidar Remote Sensing of Seawater Optical Properties: Experiment and Monte Carlo Simulation,” IEEE Trans. Geosci. Remote Sensing 57(11), 9489–9498 (2019). [CrossRef]
17. D. Yuan, J. Xu, Z. Liu, S. Hao, J. Shi, N. Luo, S. Li, J. Liu, S. Wan, and X. He, “High resolution stimulated Brillouin scattering lidar using Galilean focusing system for detecting submerged objects,” Opt. Commun. 427, 27–32 (2018). [CrossRef]
18. H. Liu, P. Chen, Z. Mao, D. Pan, and Y. He, “Subsurface plankton layers observed from airborne lidar in Sanya Bay, South China Sea,” Opt. Express 26(22), 29134–29147 (2018). [CrossRef]
19. J. Churnside, R. Marchbanks, C. Lembke, and J. Beckler, “Optical Backscattering Measured by Airborne Lidar and Underwater Glider,” Remote Sens. 9(4), 379 (2017). [CrossRef]
20. J. A. Schulien, M. J. Behrenfeld, J. W. Hair, C. A. Hostetler, and M. S. Twardowski, “Vertically- resolved phytoplankton carbon and net primary production from a high spectral resolution lidar,” Opt. Express 25(12), 13577–13587 (2017). [CrossRef]
21. J. H. Churnside, “Airborne lidar for fisheries applications,” Opt. Eng. 40(3), 406–414 (2001). [CrossRef]
22. G. W. K. a and E. S. Fry, “Aircraft Laser Sensing of Sound Velocity in Water: Brillouin Scattering,” Remote Sens. Environ. 36(3), 165–178 (1991). [CrossRef]
23. D. Yuan, P. Chen, Z. Mao, and Z. Zhang, “Potential of spaceborne Brillouin scattering lidar for global ocean optical profiling,” Opt. Express 29(26), 43049–43067 (2021). [CrossRef]
24. C. A. Hostetler, M. J. Behrenfeld, Y. Hu, J. W. Hair, and J. A. Schulien, “Spaceborne Lidar in the Study of Marine Systems,” Annu. Rev. Mar. Sci. 10(1), 121–147 (2018). [CrossRef]
25. M. J. Behrenfeld, Y. Hu, C. A. Hostetler, G. Dall’Olmo, S. D. Rodier, J. W. Hair, and C. R. Trepte, “Space-based lidar measurements of global ocean carbon stocks,” Geophys. Res. Lett. 40(16), 4355–4360 (2013). [CrossRef]
26. M. J. Behrenfeld, Y. Hu, R. T. O’Malley, E. S. Boss, C. A. Hostetler, D. A. Siegel, J. L. Sarmiento, J. Schulien, J. W. Hair, X. Lu, S. Rodier, and A. J. Scarino, “Annual boom–bust cycles of polar phytoplankton biomass revealed by space-based lidar,” Nat. Geosci. 10(2), 118–122 (2017). [CrossRef]
27. M. J. Behrenfeld, P. Gaube, A. Della Penna, R. T. O’Malley, W. J. Burt, Y. Hu, P. S. Bontempi, D. K. Steinberg, E. S. Boss, D. A. Siegel, C. A. Hostetler, P. D. Tortell, and S. C. Doney, “Global satellite-observed daily vertical migrations of ocean animals,” Nature 576(7786), 257–261 (2019). [CrossRef]
28. X. Lu, Y. Hu, J. Pelon, C. Trepte, K. Liu, S. Rodier, S. Zeng, P. Lucker, R. Verhappen, J. Wilson, C. Audouy, C. Ferrier, S. Haouchine, B. Hunt, and B. Getzewich, “Retrieval of ocean subsurface particulate backscattering coefficient from space-borne CALIOP lidar measurements,” Opt. Express 24(25), 29001–29008 (2016). [CrossRef]
29. X. Lu, Y. Hu, Y. Yang, T. Neumann, A. Omar, R. Baize, M. Vaughan, S. Rodier, B. Getzewich, P. Lucker, C. Trepte, C. Hostetler, and D. Winker, “New Ocean Subsurface Optical Properties From Space Lidars: CALIOP/CALIPSO and ATLAS/ICESat-2,” Earth Space Sci 8(10), e2021EA001839 (2021). [CrossRef]
30. D. Dionisi, V. E. Brando, G. Volpe, S. Colella, and R. Santoleri, “Seasonal distributions of ocean particulate optical properties from spaceborne lidar measurements in Mediterranean and Black sea,” Remote Sens. Environ. 247, 111889 (2020). [CrossRef]
31. W. Ji, A. Robel, E. Tziperman, and J. Yang, “Laurentide Ice Saddle Mergers Drive Rapid Sea Level Drops During Glaciations,” Geophys Res Lett 48(14), e2021GL094263 (2021). [CrossRef]
32. P. Wessel and W. H. F. Smith, “A global, self-consistent, hierarchical, high-resolution shoreline database,” J. Geophys. Res.: Solid Earth 101(B4), 8741–8743 (1996). [CrossRef]
33. J. H. Churnside, “Review of profiling oceanographic lidar,” Opt. Eng. 53(5), 051405 (2014). [CrossRef]
34. H. Liu, P. Chen, Z. Mao, and D. Pan, “Iterative retrieval method for ocean attenuation profiles measured by airborne lidar,” Appl. Opt. 59(10), C42–C51 (2020). [CrossRef]
35. J. D. Klett, “Stable analytical inversion solution for processing lidar returns,” Appl. Opt. 20(2), 211–220 (1981). [CrossRef]
36. J. Churnside, J. Hair, C. Hostetler, and A. Scarino, “Ocean Backscatter Profiling Using High-Spectral-Resolution Lidar and a Perturbation Retrieval,” Remote Sens. 10(12), 2003 (2018). [CrossRef]
37. J. M. Sullivan, M. S. Twardowski, J. Ronald, V. Zaneveld, and C. C. Moore, “Measuring optical backscattering in water,” in Light Scattering Reviews 7: Radiative Transfer and Optical Properties of Atmosphere and Underlying Surface (SpringerBerlin Heidelberg, 2013), pp. 189–224.
38. M. Kheireddine, R. J. W. Brewin, M. Ouhssain, and B. H. Jones, “Particulate Scattering and Backscattering in Relation to the Nature of Particles in the Red Sea,” J. Geophys. Res.: Oceans 126(4), e2020JC016610 (2021). [CrossRef]
39. T. Shiozaki, Y. Hirose, K. Hamasaki, R. Kaneko, K. Ishikawa, and N. Harada, “Eukaryotic Phytoplankton Contributing to a Seasonal Bloom and Carbon Export Revealed by Tracking Sequence Variants in the Western North Pacific,” Front. Microbiol. 10(2722), 1–14 (2019). [CrossRef]
40. N. D. Fredrick, J. A. Berges, B. S. Twining, D. Nuñez-Milland, and F. L. Hellweger, “Use of agent-based modeling to explore the mechanisms of intracellular phosphorus heterogeneity in cultured phytoplankton,” Appl Environ Microbiol 79(14), 4359–4368 (2013). [CrossRef]
41. P. J. Rousseeuw, “Robust estimation and identifying outliers,” in Handbook of statistical methods for engineers and scientists (DOVER PUBLICATIONS, 1990), pp. 16.11–16.24.
42. C. Zhong, P. Chen, and D. Pan, “An Improved Adaptive Subsurface Phytoplankton Layer Detection Method for Ocean Lidar Data,” Remote Sens. 13(19), 3875 (2021). [CrossRef]
43. G. C. Feldman, “GlobColour 8-day merged multiple satellite platforms product,” NASA (2021), https://hermes.acri.fr/index.php?class=archive.
44. A. Urtizberea, N. Dupont, R. Rosland, and D. L. Aksnes, “Sensitivity of euphotic zone properties to CDOM variations in marine ecosystem models,” Ecological Modelling 256, 16–22 (2013). [CrossRef]
45. L. C. Lund-Hansen, T. J. Andersen, M. H. Nielsen, and M. Pejrup, “Suspended Matter, Chl-a, CDOM, Grain Sizes, and Optical Properties in the Arctic Fjord-Type Estuary, Kangerlussuaq, West Greenland During Summer,” Estuaries and Coasts 33(6), 1442–1451 (2010). [CrossRef]
46. A. Morel and S. Maritorena, “Bio-optical properties of oceanic waters: A reappraisal,” J. Geophys. Res. 106(C4), 7163–7180 (2001). [CrossRef]
