Estimation of vegetation water content at leaf and canopy level using dual-wavelength commercial terrestrial laser scanners
Abstract
Vegetation water content, quantified as the leaf equivalent water thickness (EWT), can serve as an indicator of vegetation stress. The intensity data recorded by terrestrial laser scanning (TLS) instruments, operating at shortwave infrared wavelengths, can be used to estimate the three-dimensional distribution of EWT, after a full and rigorous calibration for the range and incidence angle effects. However, TLS instruments do not record the incidence angles automatically, making calibration challenging. In this study, intensity data from two commercially available TLS instruments (Leica P40, 1550 nm shortwave infrared wavelength, and Leica P20, 808 nm near-infrared wavelength) were combined in a normalized difference index (NDI). The NDI was found to minimize the incidence angle effects with no need for further calibration. A dry-down experiment was conducted using deciduous and conifer canopies. The NDI was found to be highly correlated to EWT at leaf level (R2 of 0.91 and 0.74) and at canopy level (R2 of 0.89 and 0.74) for the deciduous and conifer canopies, respectively. Three-dimensional distributions of EWT at canopy level were generated, which revealed some vertical heterogeneity.
1. Introduction
Water in vegetation is essential to all physiological processes, directly or indirectly [1]. When plants are water-stressed, because of drought or disease, their rates of transpiration and photosynthesis, among other cellular processes, are significantly affected [2]. The plants become less productive and more prone to burning. One approach to monitor the vegetation water status is measuring the water content, usually expressed as the leaf equivalent water thickness (EWT) [3]. This has significant importance for various agricultural and forestry applications [4]. It can help in early detection of vegetation stress [5] and infection by pests or diseases [6]. It is also important for drought assessment, irrigation scheduling and crop yield estimation in agricultural crops [7]. In addition, EWT can be used to estimate the fuel moisture content, a key metric in forest fire modelling [8], which is useful for early detection of wildfire risk and studying wildfire regimes [9]. Both applications are crucial to understanding the effect climate change has on forest fire frequency, intensity and patterns [10].
The EWT is defined as the amount of liquid water in a given leaf area [11]. In situ approaches (destructive methods and spectroscopy) can measure EWT on a small scale, but are not practical for large areas [12]. Thus, methods using multispectral and hyperspectral remote sensing data are considered more efficient and have been widely adopted [13]. Such methods are primarily based on the interaction of radiation with foliage, with reflectance in the shortwave infrared region in leaf spectra being dominated by absorption by water [14]. However, these methods are dependent on the solar illumination [15], thus they can only estimate EWT midday. Detecting the vegetation water status midday can be an unreliable indicator of water stress since leaves lose water during photosynthesis, and it is therefore better to conduct the measurements predawn when there is no transpiration [16,17]. In addition, the accuracy of the EWT estimation is heavily affected by the canopy structure, understorey vegetation and soil moisture content [15,18]. The vertical heterogeneity of the biophysical and biochemical properties of the canopy is often ignored, despite affecting the light penetration and scattering within the canopy [19,20]. There is a unique opportunity to overcome such issues using terrestrial laser scanning (TLS).
TLS not only provides information on canopy structure, but also records intensity data that can be linked to the reflectance of the scanned target [21]. Thus, by using a shortwave infrared wavelength instrument, the recorded intensity can theoretically be used to estimate EWT at leaf and canopy level [3]. Moreover, because TLS is independent of solar illumination, the estimation can be done predawn. The understorey vegetation and soil can easily be separated from the canopy in the recorded point cloud using the spatial positioning information, neutralizing their effect on the estimation [22]. However, using TLS intensity data to estimate EWT has its own limitations. The recorded intensity is affected by a number of factors, summarized in the laser equation [23] as follows:
According to the laser equation, the recorded intensity is inversely proportional to range squared. However, this is only theoretical and the actual intensity–range relationship must be studied separately for each instrument [25–28]. TLS instruments automatically record the range for each point in the point cloud and it is possible to calibrate the intensity data for the range effect [26].
The incidence angle of the laser beam (θ) is the angle between the incident laser beam and the object's surface normal, and, according to the laser equation, TLS intensity is directly proportional to cosine θ [29]. TLS instruments do not record the incidence angles and these must be calculated by the end user to apply a radiometric calibration. However, calculating the incidence angle for each leaf/needle in the point cloud is not trivial, especially in a complex vegetation canopy [30]. Alternatively, the normalized difference index (NDI) of two laser wavelengths can be used to minimize the incidence angle effect. Here the two wavelengths must be similarly affected by the incidence angle [31], else radiometric correction will still be needed [32].
A few recent attempts to use TLS intensity data to estimate vegetation water content can be found in the literature. The Salford Advanced Laser Canopy Analyser (SALCA) instrument (1545 nm shortwave infrared and 1064 nm near-infrared wavelengths) was used to estimate EWT of leaf samples from different species. A strong relationship was found between the NDI of the two wavelengths and EWT (R2 = 0.80) [3]. However, the SALCA instrument is still a proof-of-concept, not commercially available, and requires custom processing routines for accurate calibration and registration of the data to be usable at canopy level [33]. The RIEGL VZ-400 scanner (1550 nm shortwave infrared) was used to estimate the EWT of leaf samples from eight plant species, using reference spectralon targets to build a look-up table to neutralize the incidence angle effects. A significant correlation (R2 = 0.76) was achieved between the intensity and EWT at leaf level [32]. The work was expanded and EWT vertical distribution for 20 plants from four different species was retrieved in a laboratory setting using data from the same instrument, finding a solid relationship at leaf level (R2 = 0.66) and at canopy level (mean error of 4.46%) [9]. However, using the same approach to calibrate for the incidence angle effects and estimate EWT at canopy level in a complex forest environment, especially for conifer canopies, would be more challenging, as this would require estimating the incidence angle for each leaf/needle in the point cloud prior to applying the correction model. The NDI of red (690 nm) and shortwave infrared (1550 nm) wavelengths, used in phase-based Leica HDS6100 and FARO X330 instruments respectively, was used to estimate EWT for 101 leaf and needle samples from five different species. A very strong relationship (R2 = 0.93) was found between the NDI and EWT [34]. However, the samples were scanned at a single range from the scanners and no range calibration model was developed. In addition, all the samples were normal to the laser beam direction, minimizing the incidence angle effects. Applying the same approach at canopy level and in a real forest environment will require a range calibration model for each scanner, which may be challenging for phase-based instruments due to range-averaging from multiple returns [35]. The incidence angle effects may also require calibration, if the two wavelengths are not affected in the same manner, with previous studies showing mixed results in this respect, depending on the wavelengths compared [29,30].
The aforementioned studies illustrate the potential of using TLS intensity data to estimate EWT. However, there remains a gap regarding the application of TLS in estimating EWT at canopy level, with calibration of incidence angle effects being the main challenge. In this study, the NDI of shortwave infrared (1550 nm) and near-infrared (808 nm) wavelengths, employed in the Leica P40 and P20 instruments respectively, is used to estimate EWT for deciduous and conifer canopies at leaf and canopy level, in a laboratory setting. The main aims of the study are to: (i) develop methods to calibrate the intensity data from the two instruments for the range effect and to apparent reflectance, (ii) study the effects of the incidence angle on the two wavelengths and the ability of the NDI to neutralize these effects, (iii) use the NDI to estimate EWT at leaf and canopy level for deciduous and conifer canopies in a dry-down experiment, and (iv) study the influence of the woody materials on the EWT estimation at canopy level.
2. Material and methods
2.1. Terrestrial laser scanning instruments
The specifications of the Leica P40 and P20 TLS instruments are shown in table 1. The 1550 nm wavelength of the P40 scanner is sensitive to the change in EWT [3,9,34], while the P20 wavelength lies in a region in the leaf reflectance spectra that is insensitive to the change in EWT but sensitive to changes in leaf structure [36]. Theoretically, combining the two wavelengths with the NDI can estimate EWT and minimize leaf structural effects [3]. In addition, the two instruments have a similar chassis, a similar laser beam exit location and use the same scanning mechanism. This means that the laser beams can be very well aligned, achieving sufficient overlap despite the difference in the laser beam divergence. Consecutive scanning of a target with the instruments, mounted on the same tripod and occupying the same survey station, results in very similar scan geometry, which can lead to high registration accuracy of the point clouds from the two scanners. This is necessary for estimating EWT on a point-by-point basis to generate three-dimensional (3D) distributions of EWT at canopy level. The aforementioned scanning set-up was used in all the experiments described in this study. All scans were conducted with the scanners' highest possible resolution (0.8 mm at 10 m). Leica Cyclone (Leica Geosystems HDS) software was used to extract intensity values for scanned targets.
Leica P40 | Leica P20 | |
---|---|---|
measurement type | time-of-flight | time-of-flight |
wavelength | 1550 nm | 808 nm |
beam divergence | 0.23 mrad | 0.20 mrad |
beam diameter at exit | 3.5 mm | 2.8 mm |
maximum range | up to 180 m at 18% reflectivity | up to 120 m at 18% reflectivity |
scan rate | up to 1 000 000 points s−1 | up to 1 000 000 points s−1 |
highest resolution | 0.8 mm at 10 m | 0.8 mm at 10 m |
2.2. TLS intensity data calibration
2.2.1. The range effect
The intensity–range relationship was investigated separately for each scanner using two SphereOptics® spectralon panels with known reflectance (50% and 90%). Each spectralon panel was mounted on a camera tripod and scanned indoors at increasing ranges between 2 and 36 m from the scanner. Targets were moved with a 0.5 m step between 2 and 10 m and with a 1 m step thereafter. The step was increased to 2 m after a 30 m range. The spectralon panels' surface was normal to the laser beam direction throughout the experiment to minimize the incidence angle effect. A static correction model was developed separately for each scanner, using the 50% spectralon panel as a reference for the calibration. The concept of the static correction model is described in Blaskow & Schneider [26]. The properties of the correction models are described in table 2.
scanner | polynomial approximation | target intensity for calibration |
---|---|---|
P40 | 2nd degree function between 2 and 4 m (R2 0.999) |
18 m |
P20 | 6th degree function between 2 and 15 m (R2 0.973) | 15 m |
2.2.2. Intensity–reflectance relationship
Five spectralon panels (table 3) were scanned consecutively using each scanner at a fixed range from the scanner (18 m for the P40 and 15 m for the P20). The laser beam direction was perpendicular to the surfaces of the spectralon panels to neutralize the incidence angle effect. The intensity values of each panel, for each scanner, were extracted, averaged and compared with the panel's actual reflectance to determine the linearity of the intensity–reflectance relationship.
spectralon | 5% | 20% | 50% | 90% | 95% |
P40 (1550 nm) | 4.7% | 24.2% | 41.8% | 90.2% | 94.7% |
P20 (808 nm) | 4.5% | 22.6% | 43.5% | 92.2% | 96.1% |
2.2.3. The incidence angle effect
Two experiments were conducted to study the intensity–incidence angle relationship for both scanners, and determine whether or not the incidence angle affects the two wavelengths in the same manner. Spectralon panels were used in the first experiment, while leaf samples were used in the second.
2.2.3.1. Incidence angle spectralon panels experiment
Two spectralon panels (20% and 50% reflectance respectively) were scanned separately at a fixed range of 12 m from each scanner. The spectralon panels were rotated in 10° increments between scans, increasing the incidence angle from 0° to 60°. Assuming the spectralon panels to be Lambertian surfaces, the intensity–incidence angle relationship was compared with the Lambertian cosine law, described as
2.2.3.2. Incidence angle leaf samples experiment
Eighteen leaf samples from six different species (three leaves for each species), including grey alder (Alnus incana), common lime (Tilia x europaea), common alder (Alnus glutinosa), hornbeam (Carpinus betulus), poplar (Populus sp.) and cherry (Prunus avium), were collected from Peel Park in Salford, Manchester, UK, during July 2016. The leaf samples were divided into three groups, each group containing six leaf samples, one sample from each species.
Each group of leaves was suspended in a wooden frame by thin black threads. The frame was positioned at a fixed scan range of 6.5 m. A clear space of 2 m was ensured behind the frame to avoid any influence of the energy reflected from the rear wall on the leaf samples' backscattered intensity. This experimental set-up was used in all leaf scanning experiments subsequently described.
The whole frame was rotated between scans, increasing the incidence angle from 0° to 60°, with scans conducted at 20° intervals. The experiment was repeated for each scanner and for each leaf group, resulting in a total of 24 scans.
The intensity values of each leaf, at each incidence angle, for each scanner, were extracted. The points near the edges of the leaves were manually removed as they corresponded to partial hits. The intensity values were calibrated into apparent reflectance using equations (3.1) and (3.2) for the P40 and P20 scanners respectively, as described in §3.2. The NDI of the reflectance was calculated at each incidence angle for each leaf as:
2.3. Water content estimation
An indoor dry-down experiment was conducted at Salford University, Manchester, UK, in July 2016, using two deciduous snake-bark maple (Acer davidii) canopies (2.6 m height) and two Corsican pine (Pinus nigra) conifer canopies (0.9 m height). For each species, one tree served as a control unit and was watered regularly, while the other acted as a dry-down unit and was left to dry naturally. The trees, together with a 50% spectralon panel, were placed 6.5 m away from a tripod, ensuring a clear space of at least 1 m between them and the wall behind. Neither the trees nor the tripod were moved during the entire period of the experiment, thereby ensuring the scanners occupied the same survey point in all scans and faced the trees from the same viewing angle. The canopies were scanned by both scanners on each day of the experiment.
2.3.1. Leaf-level experiment
2.3.1.1. Deciduous leaves
The duration of the dry-down experiment was eight days for the deciduous canopies. Three leaf samples were collected daily from the dry-down unit, except for days 4 and 7, resulting in a total of 18 leaf samples. Collecting more daily samples was not possible while avoiding defoliation of the small canopy. The fresh weight of each leaf was measured immediately on collection, then the leaves were suspended in a wooden frame and successively scanned by the P40 and P20 scanners. Afterwards, the leaves and a scale were scanned by an Epson Perfection photo scanner, and the surface area of each leaf was obtained using Image-J 1.50i software [37]. The leaves were subsequently dried in an oven for 48 h at 60°C. Their dry weight was measured and EWT for each leaf was calculated as follows [38]:
To add species variety to the experiment, nine additional leaf samples were collected randomly from five different species in Peel Park in Salford, Manchester. The leaf samples included: three leaves of grey alder (Alnus incana), two leaves of common lime (Tilia x europaea), one leaf of common alder (Alnus glutinosa), one leaf of poplar (Populus sp.) and one leaf of cherry (Prunus avium). The leaves were scanned and their EWT measured.
The intensity values of each leaf sample were calibrated to apparent reflectance using equations (3.1) and (3.2) for the P40 and P20 scanners, respectively (§3.2). The NDI of reflectance was calculated for each leaf using equation (2.3).
2.3.1.2. Conifer needles
The duration of the experiment was nine days for the conifer canopies. Three needle samples were collected daily from the dry-down unit, except for days 5 and 8. Each sample consisted of 25–45 needles. The needles in each sample were taped together, side by side from the top, to form a wider surface that could be scanned. They were then scanned as a single unit by both the P40 and P20 scanners. The weight of the needles in the sample was measured and considered the sample weight. The surface area of each sample was determined as the summation of the length of needles in the sample multiplied by the needle width (estimated at 1 mm).
To measure the needle samples' dry weight, the samples were dried in an oven for eight days at 60°C. Samples were weighed every two days until no change in weight was detected. EWT of each needle sample was calculated according to equation (2.4). Similar to the leaf samples, the NDI of the reflectance was calculated for each needle sample.
2.3.2. Canopy-level data processing and analysis
The canopy-level scans from the dry-down experiments were imported into Leica Cyclone. Partial canopy hits were removed from the point cloud using the Cyclone partial hits filter on a medium setting. Each P40 and P20 point cloud pair was registered in Leica Cyclone. Each registered point cloud pair was then finely aligned using the fine registration module in the CloudCompare v. 2.6.2 software package. The P20 point clouds required filtering as they included more points than the corresponding P40 point clouds, due to the presence of more remaining partial hits than the P40. To filter the P20 point clouds, each pair of the P40/P20 finely registered point clouds was imported into Matlab and a nearest neighbour function was applied. The function generated an index matrix that defined the coordinates of the nearest neighbour in the P20 point cloud to each point in the P40 point cloud. The index matrix was then used to filter the P20 point cloud, generating a nearest neighbour point cloud containing the same number of points as the corresponding P40 point cloud.
The registered point clouds were calibrated for the range effect on a point-by-point basis. That is, the distance between each point in the scan and the scanner was calculated and used for the range calibration of the intensity value of that point, using the calibration model described in §2.2.1. The reflectance of each point in the scan was then estimated using equation (3.1) for P40 scans and equation (3.2) for P20 scans, derived in §3.2.
The NDI was calculated using equation (2.3) for each pair of P40/P20 scans on a point-by-point basis and an NDI point cloud for each day of the experiment was generated. Mean NDI at canopy level was compared with the mean EWT of the leaf samples of the dry-down unit on each day of the experiment.
NDI point clouds were then transformed to EWT point clouds by applying equations (3.3) and (3.4), on a point-by-point basis, for the deciduous and conifer canopies respectively, derived in §3.4.1. The average estimated EWT of each day of the experiment was calculated from the point cloud of the dry-down unit as the summation of the water content of all points in the EWT point cloud, divided by the number of points. The estimated EWT at canopy level was compared with the EWT from leaf samples on each day of the experiment, not with the aim of validating the estimation, which requires destructively sampling the whole canopy, but to study the effect of the presence of woody materials on upscaling EWT estimation from leaf to canopy level.
The histogram of the EWT distribution at canopy level on day 1 was used to separate the woody materials from the leaves by choosing a separation threshold by trial and error, until few points corresponding to leaves were incorrectly filtered as woody materials. The threshold was then applied to the point clouds on all days of the experiment. In the case of the conifer canopies, the threshold incorrectly extracted too many points corresponding to needles in the point cloud on day 9. Thus, the identified woody points for day 1 were used as a reference, and the filtered points on day 9 that did not correspond to woody materials were extracted and re-added to the needle point cloud.
To study the vertical heterogeneity of the EWT, the EWT point clouds of the dry-down units on all days of the experiment, after filtering the woody materials, were divided into nine horizontal layers (five horizontal layers for the conifers) and the average EWT for each layer was calculated.
3. Results and discussion
3.1. The range effect
The mean intensity–range relationship for the two instruments does not follow the laser equation (figure 1). The instruments are equipped with a near-distance intensity reducer to protect the optics, and seem to be equipped with on-board range calibration adjustments that calibrate for the range effect after 15 m for the P20 and 18 m for the P40. A similar observation was reported for the Faro Focus3D 120 scanner, which internally calibrates for the range effect after 15 m [28]. Calibration was needed between 2 and 15 m for the P20 scanner and between 2 and 18 m for the P40 scanner.
3.1.1. Range effect calibration
The accuracy of the correction models was evaluated by using them to correct the 90% spectralon panel intensity values. The statistics of the original and calibrated intensity values (tables 4 and 5) reveal that the models successfully minimized the range effect, substantially reducing the difference between the maximum and minimum intensity values obtained for the same panel. However, the P40 intensity value at 2 m was significantly lower because of the near-distance intensity reducer and the calibration model failed to correct this. As no scans are planned at such very near range, this is unlikely to be an issue.
spectralon | max. intensity |
min. intensity |
mean |
standard deviation |
||||
---|---|---|---|---|---|---|---|---|
before | after | before | after | before | after | before | after | |
50% | 0.595 | 0.473 | 0.301 | 0.473 | 0.481 | 0.473 | 0.058 | 0.000 |
90% | 0.853 | 0.823 | 0.577 | 0.803 | 0.811 | 0.818 | 0.053 | 0.004 |
spectralon | max. intensity |
min. intensity |
mean |
standard deviation |
||||
---|---|---|---|---|---|---|---|---|
before | after | before | after | before | after | before | after | |
50% | 0.653 | 0.593 | 0.532 | 0.593 | 0.607 | 0.593 | 0.027 | 0.000 |
90% | 0.877 | 0.859 | 0.839 | 0.855 | 0.861 | 0.857 | 0.009 | 0.001 |
3.2. Intensity–reflectance relationship
Nonlinear relationships for intensity against panel true reflectance were identified for both scanners (figure 2). A second-degree polynomial function was fitted to each intensity–reflectance curve with R2 > 0.99 for both scanners. The polynomial functions can be described as follows:
3.3. The incidence angle effect
3.3.1. Incidence angle effect for spectralon panels
A very high correlation was found between the recorded intensity values and the theoretical intensity values derived from the cosine law for both scanners (figure 3, R2 0.98 for the 20% panel and 0.99 for the 50% panel). However, there is some deviation from the cosine relationship, suggesting that the panels are not perfectly Lambertian and that the 20% panel is slightly less Lambertian than the 50% one.
Using NDI largely neutralizes the incidence angle effects for both spectralon panels (figure 3c). However, some deviation can be noticed for incidence angles around 60° for the 20% spectralon panel, as NDI slightly increases. This may be due to the 20% panel being less Lambertian than the 50% panel.
3.3.2. Incidence angle effect for leaf samples
The effect of the incidence angle on the P40 and P20 reflectance is significant in all species sampled (figure 4). Using NDI minimizes the incidence angle effects for the different species (figure 4c). Some deviation is seen at the 60° incidence angle as NDI slightly increases, in a similar behaviour to that observed for the 20% panel in §3.3.1. Similar findings were reported for the NDI of a shortwave infrared wavelength (1545 nm) and a near-infrared wavelength (1064 nm), which was found to be able to minimize the incidence angle effects [30].
3.4. Water content experiments
3.4.1. Leaf-level equivalent water thickness estimation
A strong linear relationship between the NDI and EWT was found for the snake-bark maple leaf samples (figure 5a, R2 = 0.82). The following equation was derived from the relationship, which can be used to estimate EWT from the NDI for this species:
A similar strong linear relationship is also found between the NDI and EWT for the additional nine leaf samples (figure 5b, R2 = 0.94). The strong relationship holds (R2 = 0.91) when all the leaf samples are combined together (figure 5c).
A solid relationship between the NDI and EWT is found for the needles (figure 6, R2 = 0.74). Equation (3.4) describes the relationship between the NDI and EWT for this species,
3.4.2. Canopy-level equivalent water thickness estimation
The registration accuracy of the P40/P20 point clouds obtained on all days of the experiment was very similar, as the scan geometry was constant. The root mean square error (RMSE) of the registration ranged between 0.7 and 0.8 mm for all the scans, which is very close to the used scan resolution (0.8 mm at 10 m).
3.4.2.1. Influence of woody materials
Figure 7 shows a 3D EWT point cloud for the control and dry-down unit of the deciduous canopies on day 8, alongside the histogram of the water content distribution. Figure 8 shows the same for the conifer canopies on day 9. Visual inspection, of the EWT point clouds and histogram on all days, revealed that the woody materials have a higher water content than the leaves in the deciduous canopies, possibly because the trees are young and the woody materials are green. Green wood can have similar moisture content to leaves or even higher [39]. In the case of the conifer canopies, the woody materials (brown and mature) have lower water content than the needles. The canopies, especially the dry-down units, also appear to show vertical variation in EWT distribution, with lower water content in the lower canopy than in the top.
Thresholds of 0.028 g cm−2 for deciduous and 0.011 g cm−2 for conifers were applied to the EWT point clouds to extract the woody materials. Figure 9 illustrates the extracted woody materials of the EWT point clouds shown in figures 7 and 8.
The NDI at canopy level of the dry-down units, before and after extracting the woody materials, was plotted against EWT of leaf samples (figure 10). The results reveal a linear relationship with R2 of 0.57 and 0.48 for the deciduous and conifer canopies, respectively, before filtering the woody materials. Both relationships are affected by a statistical outlier caused by the results on day 6, when the measured EWT of the leaf samples was higher than the previous days (figure 11), but would be expected to be lower as the canopies had dried more. Excluding the outlier, which probably resulted from a laboratory measurement error, improves the correlation (R2 = 0.89 and 0.74 for the deciduous and conifer canopies, respectively). Filtering the woody materials has an insignificant effect on the NDI–EWT relationship for the deciduous canopy, but significantly improves the relationship for the conifer canopy (R2 = 0.73).
The estimated EWTs at canopy level for the dry-down units were compared with the EWT of leaf samples (figure 11). The results reveal an overestimation on all days for the deciduous canopy when woody material is included, except for day 6, with an average absolute error of 5.32%, and significant underestimation for the conifer canopy, with an average error of 18.69%. Studying the errors, alongside the histograms and point clouds of the EWT distribution, reveals that the higher error in EWT estimation in the conifer canopy seems to be a result of the canopy having a higher wood to leaf ratio than the deciduous canopy. Filtering of the woody materials reduces the average error in EWT estimation to 3.18% for the deciduous canopy, and dramatically decreases the average error in the estimation to as low as 2.49% for the conifer canopy. This reveals the strong effect of the presence of the woody materials on the upscaling of the TLS estimation of EWT from leaf to canopy level.
3.4.2.2 Vertical distribution of equivalent water thickness
Figure 12 compares the vertical distribution of EWT for the dry-down units on the first and last day of the experiment. The results reveal that the canopies have higher water content in the upper canopy than in the lower canopy. For the deciduous canopy, the bottom of the canopy has 35% lower water content than the top on day 1, and 39.7% lower water content on day 8. For the conifer canopy, the bottom of the canopy has 37.3% lower water content than the top on day 1, and 51% lower water content on day 9. Comparing the EWT vertical profile on day 1 and on the last day reveals that the deciduous canopy loses moisture from all parts of the canopy relatively evenly as it dries, with slightly greater water loss from the lower canopy. The conifer canopy shows a similar behaviour, but with upper canopy water content remaining almost constant over the course of the experiment. Although the deciduous and conifer canopies were not fully destructively sampled to validate the EWT vertical profiles, the observations obtained agree to the findings reported in [40], a study in which younger leaves were reported to have higher EWT than older leaves for 1099 leaf samples collected from 12 lowland Amazonian trees. In another study, it was reported that, for two small canopies (Ficus benjamina), the water content varied vertically and was higher in the top of the canopy [9]. This was also explained as the result of new leaves at the top of the canopy having higher water than older leaves [41].
4. Conclusion
In this study, the NDI of two wavelengths, shortwave infrared (1550 nm) and near-infrared (808 nm), employed in the Leica P40 and P20 commercial TLS instruments respectively, was used to estimate EWT at leaf and canopy level for deciduous and conifer canopies in an indoor dry-down experiment. Calibration models were developed for each scanner and successfully used to calibrate the intensity data for the range effect and to apparent reflectance.
The two scanning wavelengths were found to be affected in a similar way by the incidence angle. Using the NDI minimized the incidence angle effects with no need for further radiometric calibrations, although some deviation at extreme incidence angles occurred, which is likely to be due to the presence of more partial hits in this case, making it harder to extract only the full hits. However, there could also be some difference in specularity between the two wavelengths. The NDI was found to be highly linearly related to broadleaf EWT across various tree species, and also to needle EWT of the Corsican pine conifer canopy. However, as the experiment included samples from a single conifer species, the results obtained can only be considered preliminary and further experiments that include various conifer species are necessary in order to determine whether the NDI–EWT relationship will hold or not.
At canopy level, a high registration accuracy was obtained for each P40/P20 point cloud pair on each day of the experiment, because of the similarity in the scan geometry. The NDI estimated EWT on a point-by-point basis for the deciduous and conifer canopies, generating 3D distributions of EWT that revealed some vertical heterogeneity. The young leaves and needles in the upper canopy had higher water content than the older leaves/needles in the lower canopy. This, if also observed in a forest environment, can affect passive optical space or airborne sensor estimation of EWT, as measurements from such instruments will be dominated by the canopy top [42]. However, the observed vertical heterogeneity and drying patterns may be characteristics of the specific species involved in the study.
The NDI overestimated the EWT for the deciduous canopy, and underestimated it for the conifer canopy on all days of the experiment, with the errors in the EWT estimation being proportional to the amount of woody material in the point cloud. Filtering the woody material significantly improved the EWT estimation accuracy. There is potential in using the 3D EWT distribution to separate the woody material from the leaves by choosing a threshold in the histogram of the EWT distribution. However, some leaves are filtered as woody material, creating a bias in the EWT estimation. Methods that use spatial information to separate the wood from the leaves may need to be used, for example [43,44].
This study shows the significant potential of combining the data from commercial laser scanners, such as the Leica P40 and P20, to generate a 3D distribution of EWT at canopy level without the need for incidence angle calibration. However, as the canopies used in the dry-down experiment were not extensively destructively sampled, due to their small size, the ability to provide vertical distribution of EWT by such methods needs further validation. In addition, experiments in real forest environments need to be conducted to assess the applicability of the proposed EWT estimation approach in larger canopies and outdoor conditions (for example in the presence of wind). If such challenges can be addressed, the potential of multi-wavelength TLS methods to better characterize and study heterogeneity in biochemistry within forest canopies will be substantial, providing a new insight into forest stress and tree physiology.
Data accessibility
The point clouds and processing codes can be downloaded via this link: https://doi.org/10.6084/m9.figshare.5053882.v1
Authors' contributions
A.E. planned and carried out the experiments and data collection, wrote the processing codes and analysed the data, as well as drafted the manuscript. R.G. and J.M. contributed to planning the experiments, interpretation of results and helped draft and edit the manuscript. All authors gave final approval for publication.
Competing interests
We have no competing interests.
Funding
This work is part of A.E.'s PhD research project, funded by the Ministry of Higher Education, Egypt, represented by the Egyptian Cultural and Educational Bureau in London, UK.
Acknowledgements
The authors would like to thank Prof. Mark Danson for hosting the dry-down experiment at Salford University, Manchester, UK, and Fadal Sasse for his help during the experiment. The Natural Environment Research Council (NERC) Centre for Earth Observation is thanked for providing the spectralon panels used in the calibration. We are also grateful to Leica Geosystems UK for loan of the P40 scanner used in the experiments.
References
- 1
Lisar SYS, Motafakkerazad R, Hossain MM, Rahman IMM . 2012 Water stress in plants: causes, effects and responses. In Water stress (eds IMM Rahman, H Hasegawa), pp. 1–14. Rijeka, Croatia: InTech. Google Scholar - 2
Mendiguren G, Pilar Martín M, Nieto H, Pacheco-Labrador J, Jurdao S . 2015 Seasonal variation in grass water content estimated from proximal sensing and MODIS time series in a Mediterranean Fluxnet site. Biogeosciences 12, 5523–5535. (doi:10.5194/bg-12-5523-2015) Crossref, Web of Science, Google Scholar - 3
Gaulton R, Danson FM, Ramirez FA, Gunawan O . 2013 The potential of dual-wavelength laser scanning for estimating vegetation moisture content. Remote Sens. Environ. 132, 32–39. (doi:10.1016/j.rse.2013.01.001) Crossref, Web of Science, Google Scholar - 4
Hansen PM, Schjoerring JK . 2003 Reflectance measurement of canopy biomass and nitrogen status in wheat crops using normalized difference vegetation indices and partial least squares regression. Remote Sens. Environ. 86, 542–553. (doi:10.1016/S0034-4257(03)00131-7) Crossref, Web of Science, Google Scholar - 5
Carter GA . 1993 Responses of leaf spectral reflectance to plant stress. Am. J. Bot. 80, 239–243. Crossref, Web of Science, Google Scholar - 6
Meentemeyer RK, Anacker BL, Mark W, Rizzo DM . 2008 Early detection of emerging forest disease using dispersal estimation and ecological niche modeling. Ecol. Appl. 18, 377–390. (doi:10.1890/07-1150.1) Crossref, PubMed, Web of Science, Google Scholar - 7
Yi Q, Wang F, Bao A, Jiapaer G . 2014 Leaf and canopy water content estimation in cotton using hyperspectral indices and radiative transfer models. Int. J. Appl. Earth Obs. Geoinf. 33, 67–75. (doi:10.1016/j.jag.2014.04.019) Crossref, Google Scholar - 8
Danson FM, Bowyer P . 2004 Estimating live fuel moisture content from remotely sensed reflectance. Remote Sens. Environ. 92, 309–321. (doi:10.1016/j.rse.2004.03.017) Crossref, Web of Science, Google Scholar - 9
Zhu X, Wang T, Skidmore AK, Darvishzadeh R, Niemann KO, Liu J . 2017 Canopy leaf water content estimated using terrestrial LiDAR. Agric. For. Meteorol. 232, 152–162. (doi:10.1016/j.agrformet.2016.08.016) Crossref, Web of Science, Google Scholar - 10
Aponte C, de Groot WJ, Wotton BM . 2016 Forest fires and climate change: causes, consequences and management options. Int. J. Wildland Fire 25, i–ii. (doi:10.1071/WFv25n8_FO) Crossref, Web of Science, Google Scholar - 11
Yilmaz MT, Hunt ER, Goins LD, Ustin SL, Vanderbilt VC, Jackson TJ . 2008 Vegetation water content during SMEX04 from ground data and Landsat 5 Thematic Mapper imagery. Remote Sens. Environ. 112, 350–362. (doi:10.1016/j.rse.2007.03.029) Crossref, Web of Science, Google Scholar - 12
Pu R, Gong P, Biging GS, Larrieu MR . 2003 Extraction of red edge optical parameters from Hyperion data for estimation of forest leaf area index. IEEE Trans. Geosci. Remote Sens. 41, 916–921. (doi:10.1109/TGRS.2003.813555) Crossref, Web of Science, Google Scholar - 13
Clevers JGPW, Kooistra L, Schaepman ME . 2010 Estimating canopy water content using hyperspectral remote sensing data. Int. J. Appl. Earth Obs. Geoinf. 12, 119–125. (doi:10.1016/j.jag.2010.01.007) Crossref, Google Scholar - 14
Zarco-Tejada PJ, Rueda CA, Ustin SL . 2003 Water content estimation in vegetation with MODIS reflectance data and model inversion methods. Remote Sens. Environ. 85, 109–124. (doi:10.1016/S0034-4257(02)00197-9) Crossref, Web of Science, Google Scholar - 15
Eitel JUH, Vierling LA, Long DS . 2010 Simultaneous measurements of plant structure and chlorophyll content in broadleaf saplings with a terrestrial laser scanner. Remote Sens. Environ. 114, 2229–2237. (doi:10.1016/j.rse.2010.04.025) Crossref, Web of Science, Google Scholar - 16
Améglio T, Archer P, Cohen M, Valancogne C, Daudet F-A, Dayau S, Cruiziat P . 1999 Significance and limits in the use of predawn leaf water potential for tree irrigation. Plant Soil 207, 155–167. Crossref, Web of Science, Google Scholar - 17
Williams LE, Araujo FJ . 2002 Correlations among predawn leaf, midday leaf, and midday stem water potential and their correlations with other measures of soil and plant water status in Vitis vinifera. J. Am. Soc. Horticult. Sci. 127, 448–454. Crossref, Web of Science, Google Scholar - 18
Ali AM, Darvishzadeh R, Skidmore AK, Duren IV . 2016 Effects of canopy structural variables on retrieval of leaf dry matter content and specific leaf area from remotely sensed data. J. Selected Topic. Appl. Earth Observ. Remote Sens. 9, 898–909. (doi:10.1109/JSTARS.2015.2450762) Crossref, Web of Science, Google Scholar - 19
Valentinuz OR, Tollenaar M . 2004 Vertical profile of leaf senescence during the grain-filling period in older and newer maize hybrids. Crop Science 44, 827–834. (doi:10.2135/cropsci2004.0827) Crossref, Web of Science, Google Scholar - 20
Wang Q, Li P . 2013 Canopy vertical heterogeneity plays a critical role in reflectance simulation. Agric. For. Meteorol. 169, 111–121. (doi:10.1016/j.agrformet.2012.10.004) Crossref, Web of Science, Google Scholar - 21
Penasa L, Franceschi M, Preto N, Teza G, Polito V . 2014 Integration of intensity textures and local geometry descriptors from terrestrial laser scanning to map chert in outcrops. ISPRS J. Photogram. Remote Sens. 93, 88–97. (doi:10.1016/j.isprsjprs.2014.04.003) Crossref, Web of Science, Google Scholar - 22
Höfle B . 2014 Radiometric correction of terrestrial LiDAR point cloud data for individual maize plant detection. IEEE Geosci. Remote Sens. Lett. 11, 94–98. (doi:10.1109/LGRS.2013.2247022) Crossref, Web of Science, Google Scholar - 23
Höfle B, Pfeifer N . 2007 Correction of laser scanning intensity data: data and model-driven approaches. ISPRS J. Photogram. Remote Sens. 62, 415–433. (doi:10.1016/j.isprsjprs.2007.05.008) Crossref, Web of Science, Google Scholar - 24
Krooks A, Kaasalainen M, Hakala T, Nevalainen O . 2013 Correction of intensity incidence angle effect in terrestrial laser scanning. ISPRS Ann. Photogram. Remote Sens. Spat. Inform. Sci. II-5/W2, 145–150. (doi:10.5194/isprsannals-II-5-W2-145-2013) Crossref, Google Scholar - 25
Kaasalainen S, Jaakkola A, Kaasalainen M, Krooks A, Kukko A . 2011 Analysis of incidence angle and distance effects on terrestrial laser scanner intensity: search for correction methods. Remote Sensing 3, 2207–2221. (doi:10.3390/rs3102207) Crossref, Web of Science, Google Scholar - 26
Blaskow R, Schneider D . 2014 Analysis and correction of the dependency between laser scanner intensity values and range. Int. Arch. Photogram. Remote Sens. Spatial Inform. Sci. 40, 107–112. Crossref, Google Scholar - 27
Fang W, Huang X, Zhang F, Li D . 2015 Intensity correction of terrestrial laser scanning data by estimating laser transmission function. IEEE Trans. Geosci. Remote Sens. 53, 942–951. (doi:10.1109/TGRS.2014.2330852) Crossref, Web of Science, Google Scholar - 28
Tan K, Cheng X, Ding X, Zhang Q . 2016 Intensity data correction for the distance effect in terrestrial laser scanners. IEEE J. Selected Topic. Appl. Earth Observ. Remote Sens. 9, 304–312. (doi:10.1109/JSTARS.2015.2497310) Crossref, Web of Science, Google Scholar - 29
Kaasalainen S, Nevalainen O, Hakala T, Anttila K . 2016 Incidence angle dependency of leaf vegetation indices from hyperspectral lidar measurements. Photogram. Fernerkundung-Geoinform. 2016, 75–84. (doi:10.1127/pfg/2016/0287) Crossref, Google Scholar - 30
Hancock S, Gaulton R, Danson FM . 2017 Angular reflectance of leaves with a dual-wavelength terrestrial lidar and its implications for leaf-bark separation and leaf moisture estimation. IEEE Trans. Geosci. Remote Sens. 55, 3084–3090. (doi:10.1109/TGRS.2017.2652140) Crossref, Web of Science, Google Scholar - 31
Eitel JUH, Magney TS, Vierling LA, Dittmar G . 2014 Assessment of crop foliar nitrogen using a novel dual-wavelength laser system and implications for conducting laser-based plant physiology. ISPRS J. Photogram. Remote Sens. 97, 229–240. (doi:10.1016/j.isprsjprs.2014.09.009) Crossref, Web of Science, Google Scholar - 32
Zhu X, Wang T, Darvishzadeh R, Skidmore AK, Niemann KO . 2015 3D leaf water content mapping using terrestrial laser scanner backscatter intensity with radiometric correction. ISPRS J. Photogram. Remote Sens. 110, 14–23. (doi:10.1016/j.isprsjprs.2015.10.001) Crossref, Web of Science, Google Scholar - 33
Danson FM, Gaulton R, Armitage RP, Disney M, Gunawan O, Lewis P, Pearson G, Ramirez AF . 2014 Developing a dual-wavelength full-waveform terrestrial laser scanner to characterize forest canopy structure. Agric. For. Meteorol. 198, 7–14. (doi:10.1016/j.agrformet.2014.07.007) Crossref, Web of Science, Google Scholar - 34
Junttila S, Vastaranta M, Liang X, Kaartinen H, Kukko A, Kaasalainen S, Holopainen M, Hyyppä H, Hyyppä J . 2016 Measuring leaf water content with dual-wavelength intensity data from terrestrial laser scanners. Remote Sensing 9, 8. (doi:10.3390/rs9010008) Crossref, Web of Science, Google Scholar - 35
Newnham G, Armston J, Muir J, Goodwin N, Tindall D, Culvenor D, Püschel P, Nyström M, Johansen K . 2012 Evaluation of terrestrial laser scanners for measuring vegetation structure. Canberra, Australia: CSIRO. Google Scholar - 36
Liu L-y, Huang W-j, Pu R-l, Wang J-h . 2014 Detection of internal leaf structure deterioration using a new spectral ratio index in the near-infrared shoulder region. J. Integr. Agricul. 13, 760–769. (doi:10.1016/S2095-3119(13)60385-8) Crossref, Web of Science, Google Scholar - 37
Schneider CA, Rasband WS, Eliceiri KW . 2012 NIH Image to ImageJ: 25 years of image analysis. Nat. Methods 9, 671–675. Crossref, PubMed, Web of Science, Google Scholar - 38
Danson FM, Steven MD, Malthus TJ, Clark JA . 1992 High-spectral resolution data for determining leaf water content. Int. J. Remote Sens. 13, 461–470. Crossref, Web of Science, Google Scholar - 39
Waterman PG, Choo GM, Vedder AL, Watts D . 1983 Digestibility, digestion-inhibitors and nutrients of herbaceous foliage and green stems from an African montane flora and comparison with other tropical flora. Oecologia 60, 244–249. Crossref, PubMed, Web of Science, Google Scholar - 40
Chavana-Bryant C 2016 Leaf aging of Amazonian canopy trees as revealed by spectral and physiochemical measurements. New Phytol. 214, 1049–1063. Crossref, PubMed, Web of Science, Google Scholar - 41
Mooney H, Ehleringer J, Björkman O . 1977 The energy balance of leaves of the evergreen desert shrub Atriplex hymenelytra. Oecologia 29, 301–310. Crossref, PubMed, Web of Science, Google Scholar - 42
Liu S, Peng Y, Wei D, Le Y, Li L . 2015 Remote estimation of leaf and canopy water content in winter wheat with different vertical distribution of water-related properties. Remote Sensing 7, 4626–4650. (doi:10.3390/rs70404626) Crossref, Web of Science, Google Scholar - 43
Hackenberg J, Spiecker H, Calders K, Disney M, Raumonen P . 2015 SimpleTree—an efficient open source tool to build tree models from TLS clouds. Forests 6, 4245–4294. Crossref, Web of Science, Google Scholar - 44
Raumonen P, Casella E, Calders K, Murphy S, Åkerblom M, Kaasalainen M . 2015 Massive-scale tree modelling from TLS data. ISPRS Ann. Photogram. Remote Sens. Spat. Inform. Sci. 2, 189–196. Crossref, Google Scholar