| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Joint Center for Earth Systems Technology, Univ. of Maryland, Baltimore County (UMBC), Baltimore, MD 20771, USA
b (current address), Biospheric Sciences Branch, Code 614.4, NASA/Goddard Space Flight Center, Greenbelt, MD 20771 USA
c Biospheric Sciences Branch, Code 614.4, NASA/Goddard Space Flight Center, Greenbelt, MD 20771 USA
d Hydrology and Remote Sensing Lab., Agricultural Research Service, USDA, Beltsville, MD 20705 USA
e Science Systems and Applications Inc. (SSAI), Lanham, MD 20706 USA
* Corresponding author (pcampbel{at}pop900.gsfc.nasa.gov)
Received for publication October 14, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: Car, carotenoids • Chl, chlorophyll • ChlF, chlorophyll fluorescence • cps, counts per second • D, derivative • F, fluorescence • K, potassium • LS means, least square means • NDVI, normalized difference vegetation index • PRI, photochemical reflectance index • R, reflectance • REIPw, wavelength position of the red edge inflection point • RS, remote sensing • SLM, specific leaf mass • TM (1–7), spectral bands on LandsatTM • USDA, United States Department of Agriculture • UV, ultraviolet
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Nitrogen, Carbon Dioxide, Tropospheric Ozone, and Ultraviolet Effects on Vegetation Physiology
Plant carbon (C) sequestration and biomass production are driven by N availability because N is involved in photochemical processes and is one of the primary resources regulating plant growth. Corn, a C4 species, has higher requirement for N than most other plants or crop species (Meisinger, 1984). However, excessive N applications may suppress crop yield and biomass allocation, and N runoff from agricultural and urban areas can cause soil acidification and leaching of essential minerals (e.g., cations P2+, K+, Mg2+, and Ca2+) (McMurtrey et al., 1994). Because most terrestrial ecosystems evolved in N-limited environments, careful monitoring is needed to understand species and ecosystems dynamics in an N-rich world.
Globally, industry and transportation are adding almost 7 trillion kg yr–1 of C to the atmosphere, and atmospheric CO2 concentrations are on the rise (U.S. Climate Change Science Program, 2005). Although this enrichment of the atmosphere portends hazards, it offers the potential benefit to increase plant production. At elevated CO2, the photosynthetic efficiency and yield of important crops such as corn, soybean, cotton, peanut, rice, wheat, sorghum, and oats may experience increase (i.e., C3 plants more than C4) (Heagle et al., 1998; Tiedemann and Firsching, 2000). Under natural conditions, limiting factors (e.g., water and N deficits and weather extremes) and the resulting changes to plant physiology would mitigate the positive effects (Tuba, 2005).
Due to anthropogenic activities, O3 pollution is on the rise in the troposphere, presenting a threat to agricultural crops and human health (U.S. Climate Change Science Program, 2005). Because of its reactive nature, tropospheric O3 is the primary component of air pollution that causes negative affects on vegetation physiology, growth, and yield (Krupa and Kickert, 1989; Mulchi et al., 1995; Krupa et al., 2001; Kim et al., 2003). During the day, leaf stomata are normally open, permitting the entry of CO2 for photosynthesis, which enables O3 to enter the stomata (Reich, 1987; Larcher, 1995). Ozone can cause a range of adverse effects, including premature leaf aging, early onset of senescence, growth and yield reductions, and altered sensitivity to biotic and abiotic stresses (Skärby et al., 1993; Krupa et al., 2001; Fuhrer, 2003; Jäger et al., 2003). Plants have evolved protective mechanisms to repair O3 injury before damage is severe, such as increased production of the antioxidant vitamins C and E, polyamines, and specialized enzymes (Melhorn and Wellburn, 1987). Therefore, careful monitoring and timely detection of O3 damage may enable management decisions that limit the adverse effects.
Numerous investigations addressing the combined effect of CO2 and O3 on plants' physiologic responses indicate that the positive effects of CO2 on vegetation growth are negated, partially or completely, by elevated O3 concentrations (Krupa and Kickert, 1989; Leblanc, 1998; Kim et al., 2003). Extended foliar exposure to O3 typically reduces CO2 fixation, and further inhibition of CO2 uptake occurs through O3-induced reduction in chloroplast function, smaller stomatal opening, or CO2 loss through respiration (Guzy and Heath, 1993; Jäger et al., 2003). On the other hand, the negative effects of O3 stress may be partially or fully ameliorated under increased atmospheric CO2 concentrations (McKee et al., 2000; Kim et al., 2003). However, in the agricultural areas of the midwestern USA, tropospheric O3 is increasing more rapidly than CO2. The corn–soybean agricultural complex is the largest ecosystem of the contiguous USA and is one of the most important for USA exports and international food security. Systematic monitoring of this system would provide significant economic benefits (U.S. Climate Change Science Program, 2005).
Since the 1970s, human activities have disrupted the natural balance between stratospheric O3 synthesis and breakdown, resulting in chemical depletion of the O3 layer and an increase of UV-B levels reaching the Earth's surface (especially in the Southern hemisphere). The effects of UV-B radiation on vegetation are mostly damaging, often resulting in toxic or mutagenic DNA products that inhibit plant function and depress biomass allocation (Middleton and Teramura, 1993; Krupa et al., 2001). Common protective responses of field-grown plants to elevated UV-B radiation are the increased synthesis of UV-absorbing compounds in the foliage and the development of thicker leaves as a result of an increase in the thickness of the upper epidermis layer and the spongy parenchyma (Sullivan et al., 2002). Because many of the UV-B–induced changes in physiology and function are linked to reductions in plant growth and productivity, elevated UV-B levels can be considered a significant environmental stress factor for agricultural and natural ecosystems (Middleton and Teramura, 1993; Tegelberg et al., 2001; Sullivan et al., 2003).
Spectral Detection of Vegetation Stress
As changes in an ecosystem occur, alterations in foliar chemistry and membrane structure result in changes in vegetation spectral signatures (Rock et al., 1986, 1994; Martin, 1994; Martin and Aber, 1997; Entcheva et al., 1999, 2004); thus, spectral observations could offer a sensitive, physiologically based indicator of vegetation stress. Current methods for monitoring vegetation biophysical and growth parameters rely on broadband (
20 nm bands) satellite data, such as those on Landsat Thematic Maper, MODIS (Moderate Resolution Imaging Spectroradiometer), and SPOT (Systeme Pour l'Observation de la Terre) satellites. The commonly used visible and near-infrared (700–1200 nm) R broadband vegetation indices are primarily associated with foliar chlorophyll (Chl) levels at the top of the canopy, not providing information about the function of the canopy as a whole. Therefore, their use has serious limitations for early detection of vegetation stress or for estimation of vegetation biophysical parameters not directly related to Chl content. As compared with broadband, high-spectral-resolution (>10 nm) R parameters are much better correlated with the amount of foliar Chl (Carter, 1994; Carter et al., 1995; Carter and Spiering, 2000; Carter and Knapp, 2001) and with the amount of projected green leaf area of canopies and landscapes observed by RS platforms (Walter-Shea et al., 1992; Entcheva et al., 1999; Carter and Spiering, 2000; Entcheva, 2000). Only the PRI index (PRI1 = [R530 – R570]/[R530 + R570]) has been associated with vegetation photosynthetic activity (Gamon et al., 1997). Recent technological developments, such as the launch of the first Earth Observing satellite by NASA, are making space-borne hyperspectral data more widely available. High-spectral-resolution R data can provide a significant improvement over the broadband indices for the detection of changes in vegetation condition and for monitoring vegetation decline (Entcheva, 2000; Entcheva et al., 2004). However, R measurements are not able to directly assess vegetation function.
Unlike R, F emitted from the Chl is directly related to photochemical reactions and has been used extensively for the elucidation of the photosynthetic pathways (Lichtenthaler, 1988; Cerovic et al., 1999; Corp et al., 2003; Kim et al., 2003). Contributions to blue/green F emissions (400–570 nm) come from two sources: (i) the dynamic portion associated primarily with short-term physiologic changes occurring within the leaf mesophyll (Chappelle et al., 1992; Cerovic et al., 1999) and (ii) the static fraction attributed to inert structural compounds, such as polyphenolics and lignin in the leaf epidermis and cell walls (Chappelle et al., 1992; Lichtenthaler et al., 1996; Lichtenthaler and Schweiger, 1998; Cerovic et al., 1999). Fluorescence emissions in the red (685 nm) and far-red (740 nm), which emanate from Chl a (referred to as chlorophyll fluorescence [ChlF]), are inversely related to the functionality of the foliar chloroplasts (Adams and Demmig-Adams, 2004). In general, increases in ChlF intensities have been associated with vegetation stress (Lichtenthaler, 1988; Kim et al., 2003). Spectral ratios such as F440/F740, F530/F685, and F685/F740 have been shown to change with decline in photosynthetic efficiency and are more accurate indicators of vegetation condition than individual F bands (Lichtenthaler, 1988; Chappelle et al., 1992; Cerovic et al., 1999; Corp et al., 2003). These F emission ratios have the potential for early detection of disturbances that could inhibit vegetation growth and productivity.
Research Goal and Objectives
This investigation compares the spectral changes in vegetation R and F over the 400- to 800-nm spectral range, associated with a range of common environmental factors. By focusing on the common spectral trends associated with the different species and stress factors, the goal is to contribute toward the establishment of a generalized, more precise, and timely spectral approach for the detection of vegetation stress. The objectives of the investigation were (i) to compare the sensitivity of R and F measurements to the vegetation treatments and associated changes in foliar constituents and (ii) to determine if the trends in the R and F data differ and can provide complimentary information.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The effects of nutrient deficiency and nutrient excess on spectral properties were studied in Experiment I on corn foliage (Zea mays L.). The spectral data and the foliar samples were collected when the corn was in grain-fill growth stage. The experimental design was a randomized complete block with four blocks, each containing eight N treatments at 150%, 125%, 100%, 75%, 50%, 25%, 12%, and 0% of the optimal N application rate of 162 kg N/ha, prescribed by the University of Maryland Soil Testing Lab. The 100% N treatment was considered optimal, whereas N deficiency was represented by the 0% to 75% treatments, and high N rates were provided by the 125% and 150% groups. Nitrogen deficiency and excess resulted in reduced corn yield (Corp et al., 2003).
The effects of elevated CO2 and O3 on spectral properties were studied in Experiment II on leaves from two varieties of soybean (Glycine max (L.) Merr.) - Manokin (O3 sensitive) and KS4694 (CO2 sensitive), which were planted within the same open-top chambers (3 m in diameter). The plants were grown in Codorus silt loam (fine-loamy, mixed, mesic Fluvaquentic Dystrochrept), under a well-watered regime (0 to –0.05 MPa). The air quality treatments, provided over the entire growing season, are [1] Control (CF or Control), charcoal-filtered ambient air to reduce ambient O3 and other gaseous pollutants, providing 
350 µL CO2 L–1 and 22 nl O3 L–1; [2] +CO2, CF air (345 µL CO2 L–1/20 nl O3 L–1) with the addition of 137 µL CO2 L–1, elevating CO2 levels to 487 µL L–1 for 18 h d–1; [3] +O3, ambient nonfiltered air (346 µL CO2 L–1/35 nl O3 L–1) with the addition of 35 ± 5 nl O3 L–1, providing 70 nl O3 L–1 for 7 h d–1; and [4] +CO2+O3, ambient nonfiltered air with the additions of the previously mentioned CO2 and O3 concentrations, providing elevated levels at 485 µL CO2 L–1 and 70 nl O3 L–1, respectively. Previous USDA studies (Leblanc, 1998; Kim et al., 2003), using the same experiments and treatments, reported that the elevated CO2 (+CO2) treatment enhanced photosynthesis and plant growth compared with ambient conditions, whereas the O3 treatment, compared with the +CO2, caused a negative effect. Therefore, the four treatments are referred to in the text, tables, and figures as [1] Control, [2] optimal treatment (elevated CO2,+CO2), [3] elevated O3 (+O3), and [4] combined treatment (elevated CO2 and O3, +CO2+O3).
The effect of elevated UV-B radiation on spectral properties was studied on red maple (Acer rubra L.) foliage in Experiment III, using experimental treatments and foliage material from an ongoing field experiment conducted during the entire 1999, 2000, and 2001 growing seasons (Sullivan et al., 2002, 2003). Three adjacent sections of a maple tree crown with southern exposure (branch age 20+ years) were provided with (i) ambient UV and visible radiation (control group, Treatment 1), (ii) supplemental UV-A (320–400 nm) radiation (+UV-A, Treatment 2), or (iii) supplemental UV-B (293–320 nm) and UV-A radiation (+UV-B+UV-A, Treatment 3), with maximum daily UV-B exposure of 10 kJ m–2 to approximate a 15% ozone depletion (Björn and Murphy, 1985). The supplemental UV radiation treatments (Treatments 2 and 3) were provided for a maximum of 6 h daily centered around solar noon at a distance of 1 m by filtered Q-panel UVB-313 lamps, following the procedures outlined in Sullivan et al. (2002, 2003). The lamps in the +UV-A group (Treatment 2) were filtered with polyester film to remove radiation below 316 nm, whereas those in the +UV-B+UV-A group (Treatment 3) were filtered with cellulose diacetate, which transmits visible, UV-A, and UV-B (>293 nm) radiation, and filters were changed weekly. The ambient control group represented the unstressed category. The tree had neither diseases nor insect infestations and was a mature (78 yr old), dominant tree, growing on a sandy loam soil at the edge of a mixed hardwood stand in close proximity (2 m) to a small first-order stream.
Measurement Procedures
Experimental measurements were conducted within the week of 28 Aug. through 1 Sept. 2000, using freshly excised, fully expanded sunlit leaves. From Experiment I, 16 corn leaf samples per treatment were collected (n = 16, 128 total) at the R3 reproductive grain-fill developmental stage. From Experiment II, 10 soybean leaves per variety and treatment were used for foliar analyses (n = 10, total of 40 per variety). For Experiment III, 10 maple leaf replicates per treatment were collected for measurements (n = 10, 30 total). Leaves were excised, placed immediately in water (soybean and maple petioles), and taken for measurements in an adjacent USDA laboratory. To enable a comparison of the sensitivity of R and F spectral measurements with regard to treatment separation and identification of plant physiologic condition, R and F spectra were acquired on the same foliar samples over the visible (400–800 nm) region, from the upper (adaxial) and lower (abaxial) leaf surfaces. Because under environmental stress vegetation foliage commonly changes its orientation, turning away from the sun the upper and exposing the lower leaf surface, spectra from adaxial and abaxial leaf surfaces were collected.
Reflectance measurements were acquired contemporaneously with the F measurements using a Li-Cor 1800 integrating sphere coupled via a 100-mm–long, single-mode, 5-mm-diameter fiber optic probe to an ASD FieldSpec Pro FR spectroradiometer (Analytical Spectral Devices, Inc., Boulder, CO), yielding a 1-nm sampling interval and approximately 3-nm resolution in the visible range.
Steady-state F emissions were collected using a spectrofluorometer (Fluorolog II; Spex Industries, Edison, NJ) with entrance and exit slits set to 1.5 nm, applying standard measurement procedures (Corp et al., 1997). To obtain F emissions in the 400- to 800-nm region, an excitation wavelength centered at 360 nm was provided by a 450-W xenon lamp. Steady-state F emissions were collected at a resolution of 2.5 nm (Figure 1A), producing four spectral peaks in the blue (F440), green (F530), red (F685), and far-red (F740), as described by Chappelle et al. (1992).
|
Analytical Approach
To determine the optimal spectral regions for damage detection, a sensitivity analysis of the R and F spectra was conducted following a procedure described by Carter et al. (1995). To discriminate between overlapping bands and to determine if there were changes in the shape or position of the spectral features associated with leaf damage, a derivative analysis was conducted (Martin, 1994; Miller et al., 1990). The wavelength of the red edge inflection position (REIPw), which defines the boundary between Chl absorption in the visible region and leaf scattering in the near-infrared wavelengths where maximum value (Dmax) occurs, was computed (Horler et al., 1983). The calculated R derivative (D) indices, suggested as sensitive to stress and to reflect the differences in the shape of the first D curve among damage levels (Entcheva, 2000), included the ratio of the Dmax to D714 and D744 nm (Dmax/D714, Dmax/D744), the Dmax/D705, Dmax/D745, and D715/D705. Previous investigations have identified a large utility of spectral R and F indices related to vegetation physiologic properties (Chappelle et al., 1992; Carter, 1994; Gitelson and Merzlyak, 1996; Middleton et al., 1996; Gamon et al., 1997; Datt, 1998; Merzlyak et al., 1999; Zarco-Tejada et al., 1999; Entcheva, 2000; Entcheva et al., 2004). The current study evaluated approximately 25 of the previously published R and F parameters to determine those most suitable for separation of environmental stress levels. The broadband R indices tested included the commonly used Landsat TM band ratios (e.g., TM5/TM4, TM3/TM1) and the normalized difference vegetation index (NDVI).
The changes in foliar pigments, analytical chemistry, and spectral properties associated with the treatments were evaluated by statistical comparisons to the unstressed (assumed optimal) treatment per experiment (Experiments I–III). The statistical significance levels of the differences (p 
0.05) were determined by ANOVA and the Tukey-Kramer pairwise mean comparisons test, and the relationships of the spectral indices to foliar pigments and chemistry were evaluated by a correlation analysis using SYSTAT 9.0 (Systat Software Inc., 2000). The coefficient of determination (r2) was used to describe the experiment-level variation in foliar variables attributed to the treatments. Spectral sensitivity was considered significant at the wavelengths where the corresponding difference was statistically significant according to ANOVA (Carter, 1994).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Although soybean leaf area and total Chl were maximal in association with elevated CO2, they declined in association with the O3 treatment (Table 1). These findings agree with the previous investigations from the same long-term experiments, reporting a fertilization effect of CO2 that is partially or completely negated by an O3 stress effect (Mulchi et al., 1988; Mulchi et al., 1995; Leblanc, 1998; Kim et al., 2003). The "optimal" (+CO2) group could be separated from the other treatments using the C/H ratio for the O3 sensitive variety (Manokin), whereas several variables (Chl, Car, N, and the C/N ratio) accomplished this for the CO2 sensitive variety (KS4694; Table 1).
Maple specific leaf weight and foliar C and H content increased in association with elevated UV radiation, whereas total Chl content decreased (Table 1). This trend has been reported as a physiologic change in response to UV stress by Sullivan et al. (2002, 2003). The elevated UV exposures were easy to separate using any of the measured foliar pigment constituents (Table 1). The N/H ratio was the only variable to discriminate among all treatments in the UV experiment.
Changes in Foliar Reflectance Associated with Vegetation Treatment
In all experiments, R was positively associated with advancement of vegetation stress in the green and red-edge regions for both leaf surfaces (e.g., corn adaxial surface; Fig. 2B). Reflectance intensity differences among treatments were significant in the 695- to 730-nm region. Narrow band ratios were superior to broadband ratios for corn, separating six to seven treatment groups (Fig. 3B). The two common broadband indices, TM3/TM1 (Fig. 3B) and NDVI, had considerably lower discrimination ability, separating only two or three groups per experiment for adaxial or abaxial surfaces (Tables 2 and 3). However, individual narrow R bands were not as successful in separating experimental treatments as the derivative (D) spectra in the red-edge region (Figs. 2C and 3B; Tables 2 and 3).
|
|
|
|
Reflectance spectral trends differed by species, but only the abaxial red/far-red ratios (e.g., R695/R760 and R750/R700) provided statistically significant species separation (ANOVA, r = 0.78–0.95; p > 0.001). In the blue/green region, R values from both corn leaf surfaces did not statistically differ. However, significantly higher blue/green R was measured from the abaxial vs. adaxial surfaces of maple and soybean leaves (Fig. 4A). In the red/far-red region, R intensities of either leaf surface for corn and soybean seemed to be similar, whereas the abaxial surfaces of maple leaves produced higher R (r = 0.82; p < 0.001) (Fig. 5B).
|
|
|
|
Spectra from the adaxial (upper) leaf surfaces (Table 4) provided higher statistical treatment separation as compared with spectra from abaxial surfaces across species and treatments. Overall, F ratios provided better separation than individual F bands. Five groups were discriminated using the adaxial green/far-red (F530/F740) ratio or using the abaxial blue/red (F440/F685) ratio. Several simple band ratios allowed for separation of a wide gradient among N levels (Tables 4 and 5): blue/green (F440/F530); blue/red (F440/685); blue/far-red (F440/F740); green/far-red (F530/F740) (Fig. 3A); and green/red (F530/F685). Among all F indices, the adaxial F530/F740 ratio, which decreased with stress level, provided the best treatment separation and successfully detected the corn 100% treatment group (having the lowest value, Fig. 3A) as well as the "optimal" treatment in the soybean CO2–sensitive variety (KS4694) (Tables 4 and 5). The blue/red (F440/F685) ratio was the best for detecting the unstressed condition in the maple UV experiment. Moreover, the adaxial blue/red F ratio increased four-fold in the UV-B–augmented foliage, as might be expected from a thicker adaxial epidermis where more blue-fluorescing structural compounds enhance blue F while shielding UV-induced ChlF.
Correlation of Spectral Properties and Foliar Compounds
For all experiments, abaxial rather than adaxial F emission spectra were better correlated to foliar pigments and chemical compounds. The association of individual F bands to C/N was lower than to pigments, with the F400 to F500 region exhibiting the highest correlation. Fluorescence emissions in the spectral region from 400 to 620 nm were negatively associated with pigment content, whereas the association was positive in the 680- to 760-nm range (Fig. 5). F600 (r 0.78) and F740 to F760 (r = 0.65–0.78) were best associated to total Chl. Fluorescence was not associated with foliar pigment content between 640 and 660 nm, the transition zone between negative and positive correlations (Fig. 6). The F ratios having the highest correlation to total Chl were from the abaxial leaf surface: maple, F530/F740 (r = –0.65) and F685/F530 (r = –0.78); soybean KS4694, F530/F740 (r = –0.84) and F685/F740 (r = –0.81); soybean (Manokin), F530/F740 (r = –0.83), F685/F740 (r = –0.85), F530/F650 (r = –0.86), and F685/F530 (r = –0.86); and corn, F530/F650 (r = –0.56). Recalculating the ratios, after subtracting the baseline F650, increased their correlation to all foliar compounds (e.g., [F440–F650]/[ChlF–F650]). Overall, the abaxial F530/F650, F685/F530, and F685/F740 ratios were best correlated to foliar N and the C/N ratio.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Leaf and crop parameters varied considerably over the experimental treatments (Table 1). The levels of photosynthetic pigments (Chl a, Chl b, total Chl, and carotenoids) were correlated to the experimentally induced stress levels in the corn N experiment (r = 0.77–0.89), the soybean CO2/O3 fumigation experiment (r = 0.91–0.97), and Chl a in the maple UV experiment (r = 0.68). The observed increase in Chl content in corn leaves, in concert with N application rate, agrees with earlier investigations (McMurtrey et al., 1994; Langsdorf et al., 2000). The foliar C/N and N/H ratios provide an insight into N use efficiency during the process of C sequestration via photosynthesis (Kull et al. 1995, Patterson et al., 1997), and high C/N and low N/H ratios are indicative of low N use efficiency and physiologic imbalance. In particular, the N/H ratio describes the relative oxidative stress associated with H+ vs. OH– side-chains of the C backbone of proteins (Patterson et al., 1997). Consequently, the N/H ratio of stressed foliage exhibited lower values as compared with the unstressed condition in the UV-exposed maple (r = 0.76) and corn (r = 0.75) leaves. The C/N or C/H ratios were more successful in tracking C uptake efficiency in the soybean experiments where CO2 augmentation was a treatment factor.
The trends in R indices, across the range of experimental treatments, were linked primarily to leaf pigment concentrations, and in all experiments R intensity from either leaf surface was positively associated with the advancement of vegetation stress, most clearly shown in the corn experiment associated with nutrient deficiency (<100% N). These responses are consistent with the foliar spectral trends commonly associated with forest decline (Rock et al., 1986; Entcheva et al., 2004).
Comparing broad- and narrow-band reflectance indices, high-spectral-resolution indices were found superior to broadband ratios. This result highlights the importance for vegetation stress detection of continuous high-resolution R observations and not just isolated, well placed, single narrow bands. Of the numerous R spectral algorithms tested, the high-spectral-resolution first derivative D715/D705 index consistently separated the largest number of treatment levels across the three experiments. The similarities in R properties of both leaf sides of corn were anticipated because corn, in contrast to soybean and maple, does not show differentiation into palisade parenchyma on the upper leaf side and spongy parenchyma on the lower leaf side.
The F indices provided different information than R indices because they apparently were influenced by structural constituents in addition to pigments. A blue-green/ChlF (F530/F740) ratio was the most successful F index over all three experiments in separating treatment groups, increasing in concert with stress-induced foliar damage. Chlorophyll fluorescence (F685 or F740) increased nonlinearly with N application, producing the highest F at the optimal (100% N) level, which is in agreement with the F trends associated by Langsdorf et al. (2000) with the effects of N on sugar beet plants (Beta vulgaris L. cv. Patricia) and reported for corn by Corp et al. (2003), whereas the green F decreased in response to higher N availability. In addition, the F530/F740 ratio tracked the curvilinear response of corn yields over N levels associated with deficiency and excess stress, allowing the detection of the optimal (100% N) treatment for maximizing crop production (Fig. 6). This advantage, if parlayed into an operational RS capability, could improve regional crop production estimates. Also, adaxial F combinations (e.g., F 440 vs. F530; Fig. 3) enabled discrimination of the three species studied here, whereas species separation using R measurements was only possible with abaxial red/far-red ratios.
Because there were no statistically significant F differences in the red (F650) among treatment levels in any experiment, this band region can serve to normalize the blue F and ChlF. Normalization of the blue-green F and ChlF to the F650 value enabled results that were more consistent across different studies. The current study confirms the previously reported trend in steady-state F (using an UV-A excitation) for corn under N treatments in association with vegetation stress (Corp et al., 2003).
Reflectance spectral algorithms, as compared with F ratios, were capable of separating a larger number of treatments but exhibited a saturation response that prevented the detection of the "optimal" vegetation condition. However, the most effective R indices (e.g., D715/D705, Av..R615..R725) were more successful than the F ratios (e.g., F530/F740) in detecting differences among treatment groups across experiments. The PRI1 index, associated with vegetation photosynthetic activity, produced a nonlinear R response similar to that reported for F emissions at F685 and F740. However, two F ratios (F440/F685 and F440/F740) had higher sensitivity and provided better treatment level separation than the PRI1 index (Tables 4 and 5 versus Tables 2 and 3).
The differences observed in F intensities of adaxial vs. abaxial leaf surfaces could be useful in stress detection and species discrimination because many species alter the orientation of their leaves to minimize adverse environmental effects (e.g., high incoming radiation, water and heat stress). The higher abaxial F (for soybean and maple leaves) could be explained with the lower content of Chl and protective pigments in the spongy parenchyma of leaf undersides as compared with the protective pigments and Chl content in the mesophyll of leaf upper sides (Middleton and Teramura, 1993; Sullivan et al., 2002; Tegelberg, 2002). Intensity differences between the two leaf surfaces are also associated with epidermal structural differences, especially in the blue region. Because leaf display geometry often changes for plants under physiologic stress (Walter-Shea et al., 1992; Larcher, 1995; Kim et al., 2003), these spectral differences are likely to be manifested at the crop and stand level as higher intensities for stressed canopies. Radiation transfer models for vegetation do not address these effects, which could offer improvements for remote estimation of vegetation dynamics and stress.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The current RS methods for assessment of photosynthetic function and C sequestration, based on the combined satellite observations and modeling efforts, could be improved by incorporating the physiologic information obtained using hyperspectral R and F observations. More research is needed to extend our findings to other vegetation types and species and to evaluate the association of photosynthetic function to F emissions produced from visible excitation wavelengths (e.g., 400–650 nm) and high-spectral-resolution R properties.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. L. Hatfield, A. A. Gitelson, J. S. Schepers, and C. L. Walthall Application of Spectral Remote Sensing for Agronomic Decisions Agron. J., May 7, 2008; 100(Supplement_3): S-117 - S-131. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Steiner and J. L. Hatfield Winds of Change: A Century of Agroclimate Research Agron. J., May 7, 2008; 100(Supplement_3): S-132 - S-152. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
