Agronomy Journal 94:1012-1019 (2002)
© 2002 American Society of Agronomy
INSTRUMENTATION
Evaluation of the LAI-2000 Plant Canopy Analyzer to Estimate Leaf Area in Manually Defoliated Soybean
Sean Malone,
D. Ames Herbert, Jr.* and
David L. Holshouser
Virginia Polytechnic Inst. and State Univ., Tidewater Agric. Res. and Ext. Cent., 6321 Holland Rd., Suffolk, VA 23437
* Corresponding author (herbert{at}vt.edu)
Received for publication November 16, 2001.
 |
ABSTRACT
|
|---|
The LAI-2000 Plant Canopy Analyzer provides a rapid estimate of leaf area index (LAI), but its accuracy and utility for estimating soybean [Glycine max (L.) Merr.] defoliation is unknown. This study evaluated minimum plot-size requirements for manually defoliated soybean experiments using the LAI-2000; compared LAI estimates using the LAI-2000 with directly determined values for 0, 33, 67, and 100% manual defoliation levels; and compared estimates between two LAI-2000 detector types having spectral sensitivity differences. The minimum plot size for obtaining accurate indirect estimates of LAI in defoliated canopies (0.89-m height) with 0.38-m row centers is six rows by 2.28 m and with 0.91-m row centers and a canopy height of 0.81 m is four rows by 2 m plus an additional 1-m defoliated area at the ends of the two middle rows. Indirect measurements of LAI with the wide-blue detector, which is included on newer LAI-2000 units, were statistically greater than directly determined values in 10 of 12 comparisons. Measurements with the narrow-blue detector were the same as directly determined values in three of five comparisons in 1998 and 1999 but were statistically greater in 11 of 12 comparisons in 2000. Estimates with the wide-blue detector were 9 and 2% greater than those with the narrow-blue detector in 1999 and 2000, respectively.
Abbreviations: LAI, leaf area index
|
INTRODUCTION
|
|---|
LEAF AREA INDEX (LAI) is closely related to soybean yield, as stated in Higley's (1992) defoliation–light interception hypothesis: Yield depends on photosynthesis during early soybean reproductive stages, and photosynthesis depends on canopy light interception, which can be described by LAI. An LAI value of 3.5 to 4.0 is correlated with 95% light interception, the value at which the crop should theoretically achieve canopy closure, maximum canopy photosynthesis for the development stage, and maximum yield for the environmental conditions (Board and Harville, 1992; Westgate, 1999). Reductions in LAI below this critical value due to insect defoliators may reduce yield (Board et al., 1997). Therefore, soybean LAI becomes a critical factor in defoliating-insect pest management.
Direct measurement of LAI is labor intensive, involving removal of all leaflets in a quadrat and determination of their area with a leaf area meter. Indirect measurement of LAI is possible with plant canopy analyzers such as the LAI-20001 (LI-COR, Lincoln, NE), which estimate LAI in less than 1 min through measurements of the gap frequency, the probability that a light ray will not contact vegetation as it passes through the canopy to the ground (Lang, 1991; Welles and Norman, 1991). Estimates of LAI that are within 15% of direct LAI are probably sufficient for most research and crop management purposes (Buntin, 1994; Hunt et al., 1999). It is unknown if the LAI-2000 accurately estimates direct LAI measurements of double-crop soybean, which is typically planted following small grain, has higher plant populations, and is subject to a shorter growing season than full-season soybean. Furthermore, with the exception of Hunt et al. (1999), research is limited in estimating LAI under defoliated conditions. The LAI-2000 may detect objects or gaps outside of small plots, which would affect the LAI estimate. Knowing the land area that the plant canopy analyzer actually views will allow us to determine the minimum plot size needed for accurate LAI estimates in defoliation experiments.
The LAI-2000 uses a fish-eye lens and optical filters to allow wavelengths of light up to 490 nm to reach five silicon detectors arranged in concentric rings, which measure sky brightness at different zenith angles to estimate LAI (Welles and Norman, 1991; Hicks and Lascano, 1995). The LAI-2000 owner's manual provides recommendations for plot size selection, sensor orientation, and sensor placement (LI-COR, 1992). It states that the sensor can see about three times the canopy height; therefore, plot size should be at least three times the crop height so that objects outside of the plot are not detected. Alternately, outer-ring values may be neglected in the LAI computation, reducing the minimum plot size to 1.6 times the canopy height. The manufacturer's recommendations for sensor orientation and placement are discussed in detail in the owner's manual. In general, for row crops, one should make diagonal transects between the rows at even intervals to improve the mean LAI value. Also, the manufacturer suggests using the 45° view cap for measurements in heterogenous canopies and (at least) doubling the number of diagonal transects. Transects are to be taken in pairs, with one looking along and one looking across the row.
In the late 1990s, the manufacturer of the LAI-2000 introduced a new detector that responds to a wider range of blue wavelength light than the previous detector. The increased amount of signal allows the LAI-2000 to work at lower light intensities (J. Welles, LI-COR, personal communication, 2001). The old and new detectors may be referred to as narrow-blue and wide-blue detectors, respectively. We need to determine whether the two detector types produce similar LAI estimates and are also within 15% of directly determined LAI values.
In a soybean-thinning experiment, Welles and Norman (1991) found that the LAI-2000 estimate of LAI was within 2% of directly determined LAI. Over several cropping system experiments, they discovered that the LAI-2000 generally gave estimates within 15% of directly determined LAI. Hunt et al. (1999) reported that LAI-2000 estimates were statistically the same as direct LAI values ranging from about 2.5 to 5.0 in full-season soybean with 0.76-m row spacing. However, plant canopy analyzer estimates were numerically greater than direct LAI values. This may be partly explained by the plant canopy analyzer's inability to differentiate between leaves and other plant parts such as stems, pods, and petioles. When direct measurements are taken, only the area of leaves is determined, whereas the plant canopy analyzer measures all plant material as leaf area. Welles and Norman (1991) suggested that the term "foliage area index," rather than LAI, might be a better description of what is measured with the plant canopy analyzer. In contrast to Hunt et al. (1999), Wilhelm et al. (2000) found that three different brands of LAI meters underestimated LAI values for corn (Zea mays L.) even though stem and ear tissue was present among leaf tissue. Direct measurements with the Model LI-3000A or LI-3100 leaf area meters (LI-COR, Lincoln, NE) are not absolute, for their readings could be affected by dirt on the belts, dirty or obstructed mirrors or lenses, or calibration drift.
The objectives of these experiments were (i) to determine the minimum plot size for LAI estimates in soybean defoliation experiments using the LAI-2000; (ii) to determine the accuracy of LAI-2000 estimates by comparing with direct measurements of LAI in full-season and double-crop soybean; and (iii) to compare LAI estimates between two LAI-2000 detector types.
|
MATERIALS AND METHODS
|
|---|
Minimum Plot-Size Determination
Experiments were conducted on 22 and 23 Sept. 1998 at the Virginia Tech Tidewater Agricultural Research and Extension Center in Suffolk, VA. Four plots were established in full-season soybean with 0.91-m row centers. Soybean was in developmental stage R6 (full seed) (Fehr and Caviness, 1977), and mean height was 0.81 m (standard deviation = 0.02 m). A 4- by 2-m blue tarpaulin stretched onto a polyvinyl chloride rectangular frame was held vertically to shade the plots from the sun when LAI estimates were taken. Readings of LAI were taken according to the instruction manual for row crops (LI-COR, 1992) between late morning and early afternoon. An opaque mask with a 45° opening was used to restrict the viewing area of the fish-eye lens and blocked the operator from the sensor's view. One above-canopy reading was taken immediately before and in the same direction as a set of four below-canopy readings along a diagonal transect at 0, 25, 50, and 75% of the distance across the row, with each reading moving about 0.10 m forward. Readings of LAI were taken from all four sides of the plot, for a total of 16 below-canopy readings per plot. The following plot areas were sequentially 100% hand-defoliated, beginning with the smallest area in the center and working outwards until the largest area was completed: two rows by 2 m, four rows by 2 m, four rows by 2 m plus an additional 1 m at the ends of the two middle rows, four rows by 4 m, six rows by 4 m, and six rows by 6 m. The LAI estimates in these completely defoliated plots were taken before the plots were expanded to the next larger size.
The experiment was repeated on 11 and 14 Sept. 2001 at the same location. Methods were the same as in 1998, except double-crop soybean with 0.38-m row centers was used and plot sizes were different. Four plots were established in soybean at developmental stage R5 (beginning seed). Mean height was 0.89 m (standard deviation = 0.05 m). Readings of LAI were taken from 0700 to 0840 h EST. The following plot areas were sequentially 100% defoliated: two rows by 0.76 m, two rows by 1.52 m, four rows by 1.52 m, four rows by 1.52 m plus an additional 0.38 m at the ends of the two middle rows, four rows by 2.28 m, six rows by 2.28 m, six rows by 2.28 m plus an additional 0.38 m at the ends of the two middle rows, six rows by 3.04 m, eight rows by 3.04 m, and 10 rows by 3.04 m.
The effect of defoliated plot size on LAI was analyzed using PROC GLM (SAS Inst., 1992). Means were separated using the Ryan–Einot–Gabriel–Welsch multiple comparison test (REGWQ) (SAS Inst., 1992).
LAI-2000 Accuracy Assessment
Field experiments at the Virginia Tech Tidewater Agricultural Research and Extension Center were conducted from 1998 to 2000 to determine the accuracy of the LAI-2000 in estimating soybean LAI compared with directly determined LAI. In September 1998, three plots were established in double-crop soybean at developmental stage R6. Row spacing was 0.46 m, and plot size was four rows by 2 m. To make plots larger and less likely for the plant canopy analyzer to detect undefoliated plants exterior to the plots, border rows and 0.50 m at the ends of the plots were manually defoliated to levels visually similar to corresponding plots, but these leaflets were not included in direct LAI determination. Borders were defoliated in all 3 yr of the experiment. A series of one above- and four below-canopy LAI readings were taken from each side of the plot with the LAI-2000 (Serial no. PCH 1186), for a total of 16 below-canopy readings per plot. Below-canopy readings followed diagonal transects spaced 0, 25, 50, and 75% of the distance across the row, with each reading moving about 0.10 m forward (Fig. 1A)
. An opaque mask with a 45° opening was used to restrict the view of the fish-eye lens. Readings of LAI were taken in the late morning or early afternoon, using a blue tarpaulin to shade the plots from direct sunlight. The tarpaulin shaded the plot area and the border rows. After the LAI was estimated, all leaflets were 100% defoliated by hand. Leaflets were stacked in piles, placed in brown paper bags, and refrigerated in the laboratory. Direct leaf area was determined with a LI-3000A leaf area meter within 8 h of the defoliation. Direct LAI values were determined by dividing total leaf area measured by ground area.
View larger version (11K):
[in this window]
[in a new window]
|
Fig. 1. Placement of LAI-2000 fish-eye lens (with 45° opening, indicated by the white portion of the circle) for below-canopy measurements in soybean plots with (A) four rows with 0.46-m row centers and (B) five rows with 0.38-m row centers. Vertical lines represent row centers.
|
|
In 1999, methods were the same as in 1998 except that full-season soybean plots at developmental stage R5 were sequentially defoliated at levels of 0, 33, 67, and 100% by removing zero, one, two, or three leaflets of each trifoliate. Measurements of LAI were taken after each defoliation event by two researchers, each using a separate plant canopy analyzer (one had a wide- and one a narrow-blue detector). Readings of LAI were taken from 0630 to 0830 h EST on 10 August. Replicates were defoliated simultaneously. Leaflets were placed in separate bags for each defoliation level. Leaf area was determined directly with a LI-3100 leaf area meter.
In 2000, seven plots were established in soybean using planting dates that represented both full-season and double-crop systems (three plots represented full-season systems and four plots represented double-crop systems). Plots were sequentially defoliated at levels of 0, 33, 67, and 100% over 4 d. Borders were defoliated as in 1998. The LAI estimates were taken at the same time each evening to ensure that the sun was in the same position for each measurement. One person took all measurements using the plant canopy analyzer with the narrow-blue detector. The position of the fish-eye lens for below-canopy readings changed in 2000 because five 0.38-m rows were used instead of four 0.46-m rows (Fig. 1B). The soybean fields were approximately 50 m away from a hardwood forest, and the trees provided shade for LAI estimates. Therefore, the blue tarpaulin was not used in the 2000 experiments. Direct leaf area was determined as in 1999. Measurements of LAI in the full-season soybean test began on 7 August at developmental stage late R3 (beginning pod). Estimates of LAI were taken in the evening after the sun had dropped below the tree line. On 8 August, one leaflet from each trifoliate was removed by hand; leaflets were saved, and their area was determined with a LI-3100 leaf area meter. On that evening, LAI estimates were taken once the sun had dropped below the tree line. This process was repeated on 9 August for removal and determination of leaf area for the second of the trifoliate leaflets and again on 10 August for the third of the trifoliate leaflets. Thus, plants were 100% defoliated by 10 August (only stems, small pods, and petioles remained). Treatment establishment of the 2000 season double-crop test began on 15 August at developmental stage R3, and methods were the same as in the 2000 full-season test. An undefoliated check plot showed that LAI values obtained from the LAI-2000 (PCH 1326) did not change during the course of this experiment (LAI = 5.26 and 5.24 at the beginning and end of the experiment, respectively).
LAI-2000 estimates and direct measurements of LAI were compared using paired t tests (
= 0.05) because samples (plots) were not independent. Paired t tests calculate the difference between the two methods of LAI determination for each plot, and hence cancel out the plot-to-plot variability.
Comparison of Plant Canopy Analyzers
Field experiments at the Virginia Tech Tidewater Agricultural Research and Extension Center were conducted in 1999 and 2000 to determine whether narrow- and wide-blue detectors in the LAI-2000 gave similar estimates of LAI. LI-COR provided information on the spectral responses of the two detector types (Fig. 2)
. In 1999, plant canopy analyzers with narrow- and wide-blue detectors were compared by taking consecutive readings in 28 plots of soybean in midreproductive developmental stages. Plots were 5.18 m long by 3.66 m wide, with 0.46-m row centers. An opaque mask with a 45° opening was used to restrict the view of the fish-eye lens. Four above- and 16 below-canopy readings were taken per plot. Below-canopy readings were taken in the four middle rows, as in Fig. 1A. Care was taken to place the fish-eye lens of each detector in the same locations for each plot. Readings were taken on an overcast day (9 August, 0800–1330 h EST) by the same operator, and no tarpaulin was used.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 2. Spectral sensitivities of the LAI-2000 narrow- and wide-blue detectors (source: LI-COR, Lincoln, NE).
|
|
Between the 1999 and 2000 seasons, the plant canopy analyzer with the wide-blue detector was returned to the manufacturer, and its wide-blue detector was replaced with a narrow-blue detector, giving the researchers two plant canopy analyzers with narrow-blue detectors. A third plant canopy analyzer with a narrow-blue detector was borrowed from the University of Nebraska's Department of Entomology. LI-COR provided the researchers a plant canopy analyzer with a wide-blue detector for use in the meter comparison study in 2000.
In 2000, LAI measurements were taken as described above using plant canopy analyzers with a narrow- and wide-blue detector on 6 September in 10 plots that were 5.18 m long by 3.81 m wide. Row centers were 0.38 m. Additionally, LAI estimates were taken using three plant canopy analyzers with narrow-blue detectors and one plant canopy analyzer with a wide-blue detector on 26 September. One operator took all readings on completely overcast days. No tarpaulin was used in these experiments. In both years, measurements from different plant canopy analyzers were compared using paired t tests.
|
RESULTS AND DISCUSSION
|
|---|
Minimum Plot-Size Determination
The fish-eye lens of the LAI-2000 detected numerically fewer plant parts outside the plot as plot size increased (Tables 1 and 2). Assuming that the largest plot size had the most accurate LAI estimate, the smallest plot size with 0.91-m row centers statistically equal to the largest plot was four rows by 2 m plus an additional 1-m defoliated area at the ends of the two middle rows (Table 1 and Fig. 3)
. In plots with 0.38-m row centers, LAI estimates from the plot with six rows by 2.28 m were the same as the largest plot size (Table 2 and Fig. 4)
. Therefore, different row widths can affect the minimum area required for accuracy of the plant canopy analyzer.
View this table:
[in this window]
[in a new window]
|
Table 1. LAI-2000 recordings of leaf area index (LAI) and standard deviation (SD) of defoliated soybean plots with 0.91-m row centers expanding in size. Pods, stems, and petioles remained after each plot was 100% defoliated.
|
|
View this table:
[in this window]
[in a new window]
|
Table 2. LAI-2000 recordings of leaf area index (LAI) and standard deviation (SD) of defoliated soybean plots with 0.38-m row centers expanding in size. Pods, stems, and petioles remained after each plot was 100% defoliated.
|
|
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 3. Placement of LAI-2000 fish-eye lens (with 45° opening) for below-canopy measurements in full-season soybean plots with 100% defoliation of four rows by 2 m plus an additional 1-m defoliated area at the ends of the two middle rows. All plants within the dashed lines were 100% defoliated; solid vertical lines represent 0.91-m row centers.
|
|
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 4. Placement of LAI-2000 fish-eye lens (with 45° opening) for below-canopy measurements in double-crop soybean plots with 100% defoliation of six rows by 2.28 m. All plants within the dashed lines were 100% defoliated; solid vertical lines represent 0.38-m row centers.
|
|
According to the manufacturer, readings in a canopy with a height of 0.89 m would require a 2.67- by 2.67-m plot (three times the canopy height), assuming that readings are taken from the edges of the plot with the sensor directed inwards. For 0.38-m row centers, this is a conservative recommendation. Decreasing the recommended plot size by 27% to the minimum plot size reported in the 0.38-m row center experiment (2.28 by 2.28 m) should give an equivalent savings in time, land, and labor requirements. According to the results of the 0.91-m row center experiment, a canopy with a height of 0.81 m would require 85% more area than the manufacturer's recommended plot size of 2.43 by 2.43 m, thus demanding more resources. Our minimum plot sizes were based on total defoliation; these sizes may be reduced if the plants have no or only partial defoliation because a dense canopy helps limit the distance the sensor can see (LI-COR, 1992).
In most defoliation plots, readings would be taken not from the plot center but from the edges with the view facing the plot center. In these experiments, readings could not be taken from the plot edges (except in the smallest plots) because it would have changed the position of the below-canopy readings as plot size increased, confounding the experiment.
LAI-2000 Accuracy Assessment
In undefoliated plots from 1998, there was no difference (P = 0.24) between the LAI estimate of the plant canopy analyzer with the narrow-blue detector (2.70, standard deviation = 0.42) and directly determined LAI (2.50, standard deviation = 0.48). In 1999, there was no significant difference between the LAI estimate of the plant canopy analyzer with the narrow-blue detector and directly determined LAI at defoliation levels of 0 and 33% (Table 3). Direct LAI values ranged from about 3.0 to 4.8 for these levels. The plant canopy analyzer with the narrow-blue detector had LAI estimates higher than directly determined LAI at the 67% defoliation level (P = 0.075) and at the 100% defoliation level (P = 0.0004). This overestimate of LAI is due to the plant canopy analyzer detecting more pod, petiole, and stem tissue at higher defoliation levels. These results were similar to those of Hunt et al. (1999). The plant canopy analyzer with the wide-blue detector had LAI estimates significantly higher than directly determined LAI at 0, 67, and 100% defoliation levels (Table 3). With the wide-blue detector, the LAI estimates at the 33% defoliation level were the same as directly determined LAI values (P = 0.178). Although the two researchers attempted to use the same technique in taking LAI estimates, some experimental error may have contributed to the difference between plant canopy analyzers this year.
View this table:
[in this window]
[in a new window]
|
Table 3. The 1999 mean estimates of soybean leaf area index (LAI) and standard deviation (SD) at 0, 33, 67, and 100% plant defoliation levels using the narrow- and wide-blue detector plant canopy analyzers compared with direct mean LAI values using paired t tests.
|
|
The sun angle changed between LAI estimates of defoliation levels in 1999 due to the time required to defoliate between measurements. Sun angle affects the amount of sunlit foliage detected by the plant canopy analyzer (LI-COR, 1992). As the sun angle increases, it is more difficult to shade the plot without the tarpaulin being in view of the sensor. In 2000, techniques were changed so that the sun would be at approximately the same angle on each day when LAI estimates were taken. A disadvantage of this technique is that different sky conditions exist for each day, thus reducing uniformity between readings. However, this method gives researchers an entire day to defoliate the plots while previously, the plots had to be defoliated as rapidly as possible to minimize the effect of changing sun angle. All plant canopy analyzer estimates in both the full-season and double-crop tests in 2000 were significantly greater (P 
0.05) than direct LAI values, at all levels of defoliation, except for the 33% defoliation level in the double-crop test (Tables 4 and 5).
View this table:
[in this window]
[in a new window]
|
Table 4. The 2000 mean estimates of leaf area index (LAI) and standard deviation (SD) in full-season soybean (0.38-m row centers) at 0, 33, 67, and 100% plant defoliation levels using two narrow-blue detector plant canopy analyzers and one wide-blue detector plant canopy analyzer compared with direct mean LAI values using paired t tests.
|
|
View this table:
[in this window]
[in a new window]
|
Table 5. The 2000 mean estimates of leaf area index (LAI) and standard deviation (SD) in double-crop soybean with 0.38-m row centers (at 0, 33, 67, and 100% plant defoliation levels) using one narrow-blue detector plant canopy analyzer and one wide-blue detector plant canopy analyzer compared with direct mean LAI values using paired t tests.
|
|
In a defoliated crop, the LAI estimate is expected to be slightly greater than the direct LAI because of the increased proportion of pods, stems, and petioles seen by the fish-eye lens (Hunt et al., 1999). Conversely, as LAI increases in a canopy, the effect of nonleaf plant tissue on the LAI estimate should be less. This nonleaf plant tissue intercepts a significant amount of radiation (Lang, 1991). In the full-season test in 2000, subtraction of the estimated remaining pod, stem, and petiole area (about 0.80–0.82) from the estimated LAI at 0, 33, and 67% defoliation resulted in estimates much closer to the directly determined LAI values in Table 4. Similar results occurred in the double-crop test in 2000 when 0.96 to 0.98 was subtracted from the LAI estimate for the 0% defoliation level (Table 5). An attempt was made to determine stem and pod area in the 100% defoliated plots using a LI-3100 leaf area meter, but this was unsuccessful as the instrument was designed to determine areas of two-dimensional, i.e., flat objects. While such manipulation of the 2000 data gives more accurate LAI estimates, similar treatment of the 1998 and 1999 data results in an underestimate of LAI.
Comparison of Plant Canopy Analyzers
In 1999, the plant canopy analyzer with the wide-blue detector (PCH 1326) gave significantly higher LAI readings than that with the narrow-blue detector (PCH 1186) (paired t tests, P < 0.001) (Fig. 5A)
. Inspection of the data indicated that the estimates were consistently different at LAI levels below 5.0, but they became more erratic above this level. Therefore, one should not use plant canopy analyzers with narrow- and wide-blue detectors interchangeably in canopies with LAI values greater than 5.0.
In 2000, a marginally significant difference (P = 0.054) in LAI measurements between the narrow- (PCH 1326) and wide-blue (PCH 0001) detectors was observed on 6 September (Fig. 5B); the wide-blue detector gave higher LAI estimates. From 26 September, differences in LAI measurements (P 0.05) between the narrow-blue detectors (PCH 0095, PCH 1186, and PCH 1326) were not observed (data not shown). The wide-blue detector (PCH 0001) had higher LAI estimates than two of three narrow-blue detectors (PCH 1186 and PCH 1326) (Fig. 5C and 5D). Although not significant at P = 0.05, LAI estimates with the other narrow-blue detector (PCH 0095) were numerically less than the wide-blue detector (Fig. 5E).
Differences between plant canopy analyzers with the narrow- and wide-blue detectors are statistically significant, but these differences may or may not be biologically significant, depending on the accuracy required by the user. All plant canopy analyzers had LAI estimates at values from approximately 2.0 to 6.0 in 2000, and the LAI estimates from the wide-blue detector averaged only 2% higher than estimates from the narrow-blue detector this year. This represents an overestimation of <0.1 LAI, a difference that may be biologically meaningless. But in 1999, the average LAI estimate from the wide-blue detector was 9% greater than that of the narrow-blue detector from 2.5 to 6.0 LAI, representing a difference of 0.4 LAI. These data show that plant canopy analyzers with either detector type may be used to track LAI progression during a season. However, the discrepancies in LAI estimates could cause problems for comparisons across instruments.
The reason the two detector types performed differently may be partly explained by the wavelength of light they respond to and the scattering of light within the canopy. Chlorophyll reflects green light; thus, leaves absorb little green light. Where chlorophyll levels are high, as in a dense canopy, there is more scattering of green light. LI-COR described the differences between the narrow- and wide-blue detector plant canopy analyzers (J. Welles, LI-COR, personal communication, 1999). The narrow-blue detector has a narrower spectral sensitivity range, with a sharp peak at about 480 nm (Fig. 2). There is slightly less absorption of light in the long-blue than in the short-blue light spectrum (long-blue light is closer to green light than short-blue light), thus more light scattering. More light scattering means that the below-canopy reading is increased (the plant canopy analyzer senses that more light is coming through the canopy), resulting in a reduced LAI estimate. The wide-blue detector has a wider spectral sensitivity range and has greater response to low wavelength light than the narrow-blue detector (Fig. 2). There is slightly more light absorption in this short-blue light, thus less light scattering and higher LAI estimates. Thus, the advantages of the new wide-blue detector include more signal to work with, which means that it will work at lower light intensities, and the wide-blue detector is probably closer to the theoretical ideal of no light scattering (J. Welles, LI-COR, personal communication, 2001). The first LAI-2000 serial number to use the wide-blue detector is PCH-1194; however, this should only be considered a guide because a narrow-blue detector repaired by LI-COR may have been replaced with the wide-blue detector (J. Welles, LI-COR, personal communication, 2000).
|
CONCLUSION
|
|---|
In soybean with 0.38-m row centers and a canopy height <0.9 m, nondestructive LAI estimates using the LAI-2000 should be taken in plots no smaller than six rows by 2.28 m. In soybean with 0.91-m row centers and canopy height not exceeding 0.8 m, plots should be at least 4 rows by 2 m plus an additional 1 m at the ends of the two middle rows. In plant canopy analyzer validation experiments, results varied by year and by detector type of the plant canopy analyzer (narrow- vs. wide-blue detector). The plant canopy analyzer with the narrow-blue detector had LAI estimates statistically the same (but numerically higher) as direct LAI values from approximately 2.0 to 5.0 LAI in 1998 and 1999. Overestimates of LAI were obtained in 2000 with both detector types, with the exception of the 33% defoliation level estimates in the double-crop test. Estimates of LAI in soybean defoliated below an LAI of 2.0 were likely skewed by a greater proportion of pods, stems, and petioles, resulting in significantly higher estimates than directly determined LAI. Caution must be taken if plant canopy analyzers with narrow- and wide-blue detectors are used interchangeably in research because the wide-blue detector tends to read higher than the narrow-blue detector. These differences in LAI, however, were small and probably biologically insignificant.
|
ACKNOWLEDGMENTS
|
|---|
The authors thank Dr. Leon Higley at the University of Nebraska and Dr. Jon Welles of LI-COR for their technical advice with this research and for providing two of the plant canopy analyzers in this study. We thank the Tidewater Agricultural Research and Extension Center staff for their assistance with fieldwork. Financial support from the Virginia Soybean Board and CREES-USDA Southern Regional Integrated Pest Management is greatly appreciated.
|
NOTES
|
|---|
1 Mention of commercial products in this paper is solely to provide information for the reader. Virginia Tech does not endorse these products and does not intend discrimination against other products which also may be suitable. 
|
REFERENCES
|
|---|
- Board, J.E., and B.G. Harville. 1992. Explanations for greater light interception in narrow- vs. wide-row soybean. Crop Sci. 32:198–202.[Abstract/Free Full Text]
- Board, J.E., A.T. Wier, and D.J. Boethel. 1997. Critical light interception during seed filling for insecticide application and optimum soybean grain yield. Agron. J. 89:369–374.[Abstract/Free Full Text]
- Buntin, G.D. 1994. Developing a primary sampling program. p. 99–115. In L.P. Pedigo and G.D. Buntin (ed.) Handbook of sampling methods for arthropods in agriculture. CRC Press, Boca Raton, FL.
- Fehr, W.R., and C.E. Caviness. 1977. Stages of soybean development. Spec. Rep. 80. Iowa State Univ. Coop. Ext. Serv., Ames.
- Hicks, S.K., and R.J. Lascano. 1995. Estimation of leaf area index for cotton canopies using the LI-COR LAI-2000 plant canopy analyzer. Agron. J. 87:458–464.[Abstract/Free Full Text]
- Higley, L.G. 1992. New understandings of soybean defoliation and their implication for pest management. In L.G. Copping et al. (ed.) Pest management in soybean. Elsevier Sci. Publ., London.
- Hunt, T.E., F.J. Haile, W.W. Hoback, and L.G. Higley. 1999. Indirect measurement of insect defoliation. Environ. Entomol. 28: 1136–1139.
- Lang, A.R.G. 1991. Application of some of Cauchy's theorems to estimation of surface areas of leaves, needles and branches of plants, and light transmittance. Agric. For. Meteorol. 55:191–212.
- LI-COR. 1992. LAI-2000 plant canopy analyzer operating manual. LI-COR, Lincoln, NE.
- SAS Institute. 1992. SAS/STAT user's guide. SAS Inst., Cary, NC.
- Welles, J.M., and J.M. Norman. 1991. Instrument for indirect measurement of canopy architecture. Agron. J. 83:818–825.[Abstract/Free Full Text]
- Westgate, M.E. 1999. Managing soybean for photosynthetic efficiency. p. 223–228. In H.E. Kauffman (ed.) World soybean research. Proc. World Soybean Res. Conf., 6th, Chicago, IL. 4–7 Aug. 1999. Superior Print., Champaign, Il.
- Wilhelm, W.W., K. Ruwe, and M.R. Schlemmer. 2000. Comparison of three leaf area index meters in a corn canopy. Crop Sci. 40: 1179–1183.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
P. Ravishankar, R.R. Lada, C.D. Caldwell, S.K. Asiedu, and A. Adams
The Effects of Light, Rehilling, and Mulching on Greenshoulder and Internal Greening in Carrots
Crop Sci.,
May 31, 2007;
47(3):
1151 - 1158.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
H. Tewolde, K. R. Sistani, D. E. Rowe, A. Adeli, and T. Tsegaye
ESTIMATING COTTON LEAF AREA INDEX NONDESTRUCTIVELY WITH A LIGHT SENSOR
Agron. J.,
June 17, 2005;
97(4):
1158 - 1163.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
