|Collection Mémoires et thèses électroniques|
|AccueilÀ proposNous joindre|
Table des matières
BÉLANGER, M.-C 1., VIAU, A.A. 1, SAMSON, G. 2, CHAMBERLAND, M. 3, soumis à Canadian Journal of Remote Sensing, Comparison of reflectance and fluorescence spectroscopy for the detection of mineral deficiencies in potato plants
1 Laboratoire de Géomatique Agricole et d’Agriculture de precision, Local 3731-A, Pavillon Casault, Université Laval, Qc, G1K 7P4. 2 Université du Québec à Trois-Rivières, Case postale 500, Trois-Rivières, Qc, G9A 5H7. 3 Telops, 100-2600, avenue St-Jean-Baptiste, Québec, Québec, G2E 6J5. Ce chapitre est le fruit du travail de la candidate. À titre de directeurs et superviseur, messieurs Viau, Samson et Chamberland ont apporté des suggestions et commentaires quant au design et à la préparation de ce chapitre.
Dans le cadre de cette étude, nous avons comparé la réflectance et la fluorescence selon leur potentiel de détection et de discrimination de trois carences nutritives (N, K, Mg) chez la pomme de terre. Des plants de la variété Dark-Red Norland ont été cultivés en serre. Des mesures de réflectance, de fluorescence de même que sur des paramètres de croissance et de développement ont été prises. La carence en azote a eu un impact majeur sur la croissance et le développement des plants, en effet, une réduction de 40% de la biomasse aérienne sèche a été observée chez les plants carencés en azote comparativement aux témoins. Les carences en K et Mg ont eu des effets négligeables sur la croissance et le développement des plants et, par conséquent, les mesures de réflectance et de fluorescence n’ont pas été grandement influencées par ces carences. En utilisant des analyses discriminantes, il a été possible d’assigner 87% des plants carencés en N à leur traitement respectif alors que cette proportion n’était que de 42% et 40% pour les plants carencés en K et Mg, respectivement. Pour la détection des plants carencés en azote, la combinaison des indices de réflectance et de fluorescence, de même que la fluorescence utilisée seule, ont permis une meilleure classification (87%) que les indices de réflectance utilisés seuls (67%). Pour la détection des carences en K et Mg, la combinaison des indices de réflectance et de fluorescence de même que la réflectance utilisée seule ont donné des résultats semblables mais toutefois plus élevés que ceux établis par les indices de fluorescence utilisés seuls. Les indices de réflectance et de fluorescence permettent donc la détection de carences nutritives (N, K, Mg) mais à différents degrés.
In this study, we evaluated the effectiveness of two remote sensing techniques, reflectance and fluorescence, to detect and distinguish three nutrient (N, K, and Mg) deficiencies in potatoes. Dark-Red Norland potato plants were grown in a greenhouse. Reflectance and fluorescence data were collected as well as growth and development parameters. Nitrogen deficiency had a major impact on growth, reducing the dry shoot biomass by 40 % compared to control plants. Minor effects were detected for K- and Mg-deficient plants. Owing to the low impact of K and Mg deficiencies on growth parameters, there were also minor effects of these two deficiencies on reflectance and fluorescence parameters. Using discriminant analyses on reflectance and fluorescence parameters, it was possible to correctly classify 87% of the N-deficient plants whereas this proportion was 42% and 40% for K- and Mg-deficient plants, respectively. For the detection of N-deficient plants, the combination of reflectance and fluorescence indices as well as fluorescence used alone gave higher classification accuracy (87 %) than reflectance indices used alone (67 %). For K and Mg deficiency detection, reflectance and the combination of both techniques gave similar results whereas fluorescence indices provided inferior detection. The reflectance and fluorescence indices can detect the three nutrient deficiencies but to different extents.
Potato crops are often grown on sandy soils where mobile elements, like nitrogen, can easily be leached resulting in nitrogen deficiencies and environmental contamination. A nitrogen deficiency is the most limiting factor in the growth of a potato plant and this stunted condition is reflected in reduced yield and poor quality tubers (Houghland, 1964). Under a potassium deficiency, there is a substantial reduction of the reallocation of dry-matter to tuber and as for nitrogen deficiencies, the yield decreases proportionally as the shortage of potassium increases (Houghland, 1964; Jenkins & Mashmood, 2003). Magnesium deficiency induces chlorosis and premature leaf senescence (Taiz & Zeiger, 1998). It might result from an antagonistic effect with soil-K; however, soil K supply may not necessarily result in a reduction in tissue Mg concentration or, more importantly, in a yield reduction (Allison et al., 2001). To minimize these negative impacts, fertilizers should be applied when and where they are needed.
Standard methods for the detection of nutrient deficiencies in crops include foliar analyses, petiole-sap tests and more recently the use of the chlorophyll-meter. Although precise, these methods require a direct contact with the crop thus reducing the number of measurements that can be acquired, due to the sampling time. To monitor vegetation on various scales, alternative methods are now available via remote sensing from satellite (CCT, 2004; CNES, 2004), airborne (ITRES, 2004; Nilsson, 1996) or tractor-mounted sensors (Belzile et al., 2003; Feiffer et al., 2003; Link et al., 2003). The remote sensing sensors can either be passive or active using for instance reflectance or fluorescence data.
Reflectance and fluorescence signals are determined by plants’ structural and biochemical properties, which in turn are sensitive to the plant nutrient status. Several studies have demonstrated that, in plants subjected to moderate nitrogen deficiency, there were decreases in the contents of photosynthetic pigments (Heller et al., 1993; Houghland, 1964) and increases in the contents of phenolic compounds that were related to changes of reflectance and/or fluorescence emission (Mercure et al., 2004; Read et al., 2002).
Reflectance measurements have been used to predict yield (Ma et al., 2001), and to estimate the content in nitrogen (Oppelt & Mauser, 2004; Tarpley et al., 2000), chlorophyll or other leaf pigments (Haboudane et al., 2002; Merzlyak et al., 2003; Read et al., 2002; Sims & Gamon, 2002). The potential of plant reflectance to detect the presence of mineral deficiencies is well established, particularly for nitrogen (Adams et al., 1993; Gamon et al., 1997; Tartachnyk & Rademacher, 2003; Vouillot et al., 1998; Zhao et al., 2003). However, different stresses can induce similar changes in the reflectance spectrum (Masoni et al., 1996).
Conversely, upon excitation by UV-radiation, plants emit two types of fluorescence: blue-green fluorescence (BGF) and chlorophyll fluorescence (ChlF). Fluorescence has been used to detect crop stresses (Lichtenthaler, 1990; Mazzinghi, 1990; Schweiger et al., 1996) notably water (Gunther et al., 1994; Méthy et al., 1991), ozone (Rosema et al., 1992) and nutrient stresses (Apostol et al., 2003; Chappelle et al., 1984; Corp et al., 2003; McMurthey III et al., 1994; Mercure et al., 2004; Samson et al., 2000). As for reflectance, most of the documented effects of nutrient deficiency on plant fluorescence are for nitrogen.
Considering the complementary natures of plant reflectance and fluorescence, McMurthey III et al., (1994) suggested that a combination of relevant parameters from both signals may improve the effectiveness of remote sensing to detect nutrient deficiencies. Their results indicated that both approaches were similarly effective to distinguish the different levels of N fertilization in corn. However, the mean comparison (Student-Neumans-Keuls (SNK) multiple range testing) used in that study did not allow the computation of a quantitative comparison between the effectiveness of reflectance and fluorescence indices for the detection of the nitrogen levels. In the present study, we used discriminatory analysis to evaluate the effectiveness of different reflectance, fluorescence, and a combination of indices from both methods to detect and distinguish N, K and Mg deficiencies in potato.
The experiment took place in a greenhouse located on the campus of Laval University (QC, Canada) in 2003. Twenty potato plants, (Solanum tuberosum, L. cv. Dark-Red Norland) were grown in 3L containers, at 22°C day, 16°C night controlled temperatures and under a 16-hour photoperiod. The growing medium was a mix of 80% vermiculite (Fafard et Frères, inc, QC, Canada) and 20% quartz sand (Unimin Corporation, CT, USA), washed with demineralized water prior to planting (Tukaki & Mahler, 1990). From emergence to the end of experiment (40 days), five control plants were fertilized using a mineral solution (Tukaki & Mahler, 1990) providing 100% of plant needs (N 160 ppm; P 29 ppm; K 234 ppm, Ca 160 ppm; Mg 48 ppm; S 62 ppm; Fe 1,83 ppm; Mn 0,5 ppm; B 0,5 ppm; Zn 0,05 ppm; Cu 0,02 ppm; Mo 0,01 ppm). Starting from emergence to the end of experiment, mineral deficiencies were induced in the other 15 plants. They received a modified mineral solution providing 15% of N (5 plants), K (5 plants) or Mg (5 plants) compared to the control treatment. This resulted in four different treatments (control, N-, K-, and Mg-deficiencies), each replicated on five different plants.
All following data were collected thrice during the experiment: at 15, 23 and 40 days after emergence (DAE) except for destructive measurements such as shoot biomass, tuber fresh weight and numbers that were collected only once, at the end of the experiment (40 DAE).
Chlorophyll contents were estimated on the fourth fully expanded leaf with a Minolta SPAD-502 Chlorophyll Meter (Spectrum Technologies Inc., IL, USA). Growth was determined by the surface of the fourth fully expanded leaf using a leaf area meter (Li-3000, Li-Cor Inc, NE, USA) and by dry shoot biomass (dried during 48h at 65°C). Tuber fresh weight and numbers were also measured at the end of the experiment. Developmental stage was evaluated using the two-digits method of Radtke & Rieckmann, (1991).
Fluorescence measurements were made over the top of potato plants using a prototype instrument, the FLUTE fluorometer (GAAP-TELOPS, QC, Canada). Excitation was provided by a Xenon (Xe) flash lamp (Hamamatsu, NJ, USA) pulsed at 50 Hz. Emitted fluorescence was detected by a photodiode (Advance Photonix, inc., CA, USA) synchronized with the excitation flash lamp. The Xe-flash lamp and the detector are driven by a trigger impulse. When the Xe-flash lamp is ON (T=1), the detector takes a first reading including reflectance + fluorescence measurement. When the Xe-flash lamp is OFF (T=2), the detector takes a second reading, including only reflectance. The fluorescence intensity is obtained by subtracting the measurement taken at T=2 from the measurement taken at T=1 and is given in volts on a digital screen. Fluorescence bands at 440, 520, 690 and 740 nm ± 10 nm were selected using filters of 5,08 cm diameter (CVI Laser, NM, USA) placed in front of the detector. UV and blue excitations were obtained by using respectively 360±40 nm and 436± 20 nm filters (Chroma Technology Corp., VT, USA) placed in front of the Xe-flash lamp. UV-induced fluorescence intensities were measured at the four emission bands whereas blue-induced fluorescence intensities were measured at 690 and 740 nm. All fluorescence measurements were taken in the greenhouse between 1000h and 1400h, under ambient light on a freestanding plant, 0.5 m from the top of the plant. Fluorescence intensities were calibrated against photodiode sensitivity and transmittance of the filters at the different wavelengths. Ratios of fluorescence intensities were computed for data analysis (Table 1).
Reflectance spectra ranging from 350 to 1500 nm (sampling interval 1.4 nm) were collected using a Field Spec Pro spectroradiometer (Analytical Spectral Devices inc., CO, USA). All reflectance spectra were taken between 1000h and 1400h, 0.3 m directly above the plants, under ambient light in the greenhouse. The instrument was regularly calibrated against the ambient light using a Spectralon plate (Labsphere, inc., NH, USA). Raw reflectance data were extracted from the spectroradiometer using the FSVNIR software package (Analytical Spectral Devices inc., CO, USA) as an ASCII file, which was imported into Matlab (The Mathworks, V.6.5, 2000 MA, USA). For each plant, an average from three spectra (30 acquisitions) was computed; the averaged spectrum was smoothed by a Savitsky-Golay filter (order 5, window 21), and 17 reflectance indices were calculated (Table 2).
Statistical analyses were performed using SAS software (SAS institute, V8.2, NC, USA). ANOVAs were realized with the GLM procedure on leaf chlorophyll content, leaf area, dry shoot biomass, tuber number and weight, and on each reflectance and fluorescence indices. The LSMEANS option from the GLM procedure was used to compute simple means comparison in order to identify variables influenced by mineral deficiencies (significance threshold set at p < 0,05). For the developmental stages, the 2-by-2 comparison was made using the LOGISTIC procedure that computes a logistic regression on this ordinal data.
The STEPDISC procedure was used to identify an optimal set of discriminant variables (Klecka, 1980) by selecting the indices that maximize the intra-class over inter-class variances ratio (SAS, 1990). Discriminant analyses using the DISCRIM procedure were conducted to classify each plant into its corresponding treatment. For each date of measurements, discriminant analyses were made using four classes: control, N-, K-, and Mg-deficient plants. All analyses were calculated using as input variables the indices selected by the STEPDISC procedure from i) reflectance, ii) fluorescence and iii) a combination of indices from both methods.
The effects of N, K, and Mg deficiencies on growth and development parameters are presented in Table 3. The significance threshold is set at p ≤ 0.05. Three significance levels are shown in the table in order to detect the main tendencies.
For chlorophyll contents, estimated using a SPAD chlorophyll-meter, no significant changes were detected for all three deficiencies. Compared to control plants, the fourth leaf area was significantly decreased in N-deficient plants at 23 and 40 DAE and also in K-deficient plants at 40 DAE. At the end of the experiment, N deficiency caused a 39,6% decrease on dry shoot biomass whereas no significant effects were observed for Mg and K deficiencies (Table 3). N deficiency also caused a 70% decrease in the number of tubers. In contrast to N, Mg deficiency resulted in a 50 % increase in the tuber number with no effect on tuber fresh weight. Considering the tubers number or weight, no significant difference was observed for K-deficient plants. The developmental stages, estimated according to Radtke & Rieckmann, (1991), were delayed in N-deprived plants, at 15 and 23 DAE but not at 40 DAE. Potato plants have an exponential growth of their vegetative parts before the 40th DAE. The different growth rates between control plants and nitrogen deficient plants lead to significantly different developmental stages during this period but the difference was no longer significant when the control plants entered into the blooming stage.
The effects of N, K, and Mg deficiencies on fluorescence indices are presented in Table 4. The significance threshold is set at p ≤ 0.05. Three significance levels are shown in the table in order to detect the main tendencies.
Fluorescence ratios were not affected by the potassium deficiency and only the epidermal transmittance at 740 nm showed a significant difference between the Mg-deficient and the control plants.
For the nitrogen deficient plants, we observed a marked increase of the F440/F520, F440/F690 and F440/F740 ratio in N-deficient plants. Our results also showed that at 15 DAE and 23 DAE, N-deficient plants had lower F690/F740 and UVbl_690 ratio compared to control plants (Table 4). In contrast, there was an increase of the UVbl_740 ratio in Mg-deficient plants at 40 DAE. Our measurements showed that FGVI decreased for N-deficient plants only.
The effects of N, K, and Mg deficiencies on several reflectance indices are presented in Table 5. The significance threshold is set at p ≤ 0.05. Three significance levels are shown in the table in order to detect the main tendencies.
For K- and Mg-deficient plants, there was no significant effect observed.
The following indices are significant only for the first date of measurement (15 DAE): SIPI, PSRI and R4 showed a significant increase for the N-deficient plants compared to the control plants and PRI showed a significant decrease.
The curvature index (CI) decreased at 15 DAE and increased at 40 DAE for N-deficient plants.
On two or more date of measurements, N deficiency caused a significant decrease of the following indices: SR680, SR705, ND680, ND705, mSR705, mND705, R3, and R5 and an increase of R1 and R2.
The indices showing a significant difference along the three dates of measurements might be more reliable for nutrient deficiency detection.
Table 5 Percentage of variation between the control and the induced mineral deficiency (K, Mg or N) on reflectance parameters at three different dates.
A stepwise discriminant analysis identified an optimal set of variables that are the most relevant for the detection of N, K and Mg-deficiencies in potatoes (identified as input variables in Table 6). The fluorescence indices with the most discriminant power are F440/F740, epidermal transmittance measured at 690 and 740 nm (UVbl_690 and UVbl_740), F690/F740 and F440/F520. The reflectance indices with the most discriminant power are PSRI, PRI, SR680, mSR705, R2, TCARI/OSAVI, CI, and R5. For early detection of nitrogen deficiency using reflectance, PSRI, TCARI/OSAVI and CI were the three parameters selected by the STEPDISC procedure and we observed that F690/F740 could correctly detect N-deficient potato plants at 15 DAE using fluorescence indices only. The early detection combining reflectance and fluorescence indices was done using PSRI, R5, FGVI and UVbl_740.
The first discriminant analyses aimed to evaluate the effectiveness of both sensing methods to discriminate mineral deficiencies, i.e., their ability to assign a plant to its correct treatment (control, N-, K-, and Mg-deficient mineral solutions). The global proportions of plants correctly classified to its specific class are presented in Table 6. On 15 and 40 DAE, the proportions of plants correctly classified according to discriminant analyses based on either fluorescence or reflectance indices were similar whereas at 23 DAE, fluorescence allowed a higher correct classification rate than reflectance. The combination of fluorescence and reflectance parameters did not notably enhance the distinction between treatments. Globally, the average of correct classification using the four classes (K15, Mg15, N15, and control) was only 50.6 %.
Table 6 Classification percentage from cross-validation following discriminant analyses using the input variables selected by the STEPDISC procedure on three datasets (reflectance, fluorescence, reflectance and fluorescence) showing four classes (N15, K15, Mg15 and control).
The low proportion of plants correctly classified to their specific treatments (Table 6) may result from the difficulty to distinguish control plants from those supplied with low K and Mg levels, owing to their minor effects on growth, fluorescence and reflectance indices (Table 3-5). To verify this point, we determined the proportion of correctly classified plants for each deficiency, taken separately. For each mineral element, Table 7 presents the proportion of plants correctly classified as K, Mg or N-deficient and as a control. In general, the proportions of correctly identified plants are markedly higher for N-deficient than for K-, Mg-deficient or control plants.
For the detection of N deficiency, fluorescence alone gives higher classification accuracy (86.7%) than reflectance does (66.7%). The combination of reflectance and fluorescence parameters did not enhance the detection of N-deficiency compared to fluorescence indices used alone.
For the detection of K and Mg deficiencies, the discriminant analyses indicated that the average effectiveness of reflectance and fluorescence were similar, although much lower than for N deficiency. For the detection of K- or Mg-deficiencies, the combination of reflectance and fluorescence indices (41.7 and 40.0 %, for K- and Mg-deficiencies, respectively) and reflectance used alone (40.0 and 53.3 %, for K- and Mg-deficiencies, respectively) gives slightly higher classification accuracy compared to those obtained with only fluorescence (26.7 and 33.3 %, respectively). According to our results, fluorescence and reflectance indices are complementary for detecting K- or Mg- deficiency, i.e., when parameters from one method give a low proportion of correctly classified plants the parameters from the other method often allow a better classification (Table 7).
The aims of our study were to evaluate the effectiveness of different reflectance and fluorescence indices as well as a combination of indices from both methods to detect and distinguish N, K and Mg deficiencies in potato plants at different stages of development. The results of the discriminant analysis made on the four treatments (N15, K15, Mg15, and control) indicate that under our experimental conditions, the proportions of plants correctly assigned to their respective treatments according to reflectance and fluorescence indices varied around 50% (Table 6). This poor global discrimination between the four treatments results from the difficulty of distinguishing control plants from plants grown under low supplies of K and Mg. Indeed, low N supply had the most important effects on growth and development of potato plants compared to the low K and Mg levels (Table 3). Consistent with these effects, the low N supply also had the most pronounced effects on fluorescence (Table 4) and reflectance (Table 5) indices. Therefore, the discriminant analysis comparing control to N-deficient plants using both methods resulted in proportions of plants correctly classified as high as 87% on average across all developmental stages whereas these proportions did not exceed 42% and 40% for the K- and Mg-deficient plants respectively (Table 7). To achieve better detection of K- and Mg-deficient plants, a new experiment could be conducted using nutrient solutions (low K and low Mg) that induce growth inhibition similar to the ones observed for N-deficiency (40 % in biomass reduction). Taken together, our results show the importance of accounting for the extent of growth inhibition to estimate the efficiency of reflectance and fluorescence indices to detect mineral deficiencies.
Besides the possibility to discriminate mineral deficiencies based on reflectance and fluorescence indices, the second goal of our study was to compare the relative effectiveness of reflectance and fluorescence methods to detect those deficiencies. Although signals from both methods depend on plants’ structural and biochemical properties, they may respond differently to mineral deficiencies since reflectance signals are mostly influenced by photosynthetic pigments such as chlorophylls and carotenoids (Carter & Spiering, 2002; Sims & Gamon, 2002) whereas fluorescence signals are mainly determined by phenolic compounds and chlorophylls (Cerovic et al., 1999). Results from our discriminant analysis indicate that the relative effectiveness of reflectance and fluorescence methods depends on mineral deficiency (Table 7). For N-deficiency, fluorescence indices achieve a better discrimination than reflectance. The combination of indices from both methods did not improve the success rate (% of correctly classified plants) obtained by the fluorescence indices for N deficiency detection. Regarding the detection of K and Mg deficiencies, reflectance indices were more effective than fluorescence and their combination did not substantially increase the average success rate of the detection.
The detection of crop deficiency has to be done as early as possible in order to identify and correct nutrient shortages before irreversible growth damage occur (Varvel et al., 1997). Our results on potato plants (Table 6) show that, for early deficiency detection using reflectance, the indices selected by the STEPDISC procedure (PSRI, TCARI/OSAVI and CI) use one or more key wavelengths ± 5 nm (415, 550, 685, 700, 710 nm) identified as good indicators of nitrogen deficiency (McMurthey III et al., 1994; Read et al., 2002; Zhao et al., 2003). The Photochemical Reflectance Index (PRI) is selected by the STEPDISC procedure, for deficiency detection at 40 DAE. In a previous experiment conducted on cotton canopy (Zhao et al., 2004), the photochemical reflectance index has been selected for the discrimination of nitrogen treatments. It has also been shown that PRI was positively correlated with nitrogen content of leaves from different species (Gamon et al., 1997; Qifa & Jihua, 2003).
For early detection of nitrogen deficiency using fluorescence, we observed that F690/F740 could correctly detect N-deficient potato plants at 15 DAE (Table 7). Our results are consistent with those obtained by McMurthey III et al. (1994) on corn crops but different from those of Heisel et al. (1996) who showed earlier diagnosis of N-deficiency with F440/F740 and F440/F520 than F690/F740 on corn crops.
In our study, the early detection that combined reflectance and fluorescence indices was done using PSRI, R5, FGVI and UVbl_740. At least one fluorescence index was selected at all dates when both methods were combined for classification. The two reflectance indices, PSRI and R5, use the key wavelengths presented above; the FGVI was presented in Cerovic et al., (1999) as a good indicator of the nutrient shortage, and the epidermal transmittance, corresponding to UVbl_740, was shown to be linearly correlated to the nitrogen content of corn leaves (Apostol et al., 2003; Samson et al., 2000).
It could be interesting to transfer the technology tested on field-grown potatoes. As field-grown potato plants will be more exposed to UV-radiation than greenhouse-grown plants, their epidermal content in phenolic compounds will increase in order to effectively screen the underlying mesophyll against potentially damaging UV-radiations (Bilger et al., 1997). The epidermal transmittance to UV-radiation will be lower in field-grown plants than in greenhouse-grown plants (Bilger et al., 1997). Moreover, under field conditions the nutrient deficiencies affecting the plants will be interacting with other variables such as climate, soil variability, solar exposure, wind, etc. Thus, additional data have to be collected under field conditions to validate the changes induced by nutrient deficiencies on fluorescence, reflectance and crop yields and to corroborate our results. Transferring the technology tested in a greenhouse to the field could be done using an instrument similar to the one presented in Belzile et al., (2003) in a field experimental set-up prior to open-field tests to eventually lead to real-time deficiency detection included in precision agriculture techniques.
In conclusion, fluorescence and reflectance indices can detect nutrient deficiencies but at different extent, their relative efficiency depends on the mineral deficiency: fluorescence could achieve better detection of N-deficiency whereas reflectance was adequate for K- and Mg-deficiency. Our results indicate that reflectance and fluorescence can achieve early detection of mineral deficiencies and that their efficiency should now be tested in a controlled outdoor environment to see if the technology developed in lab or greenhouse can be efficiently used in the field and if it can be part of a sound fertilizer program.
The authors thank the Fonds québécois de la recherche sur la nature et les technologies (FQRNT) through its industrial scholarship program, the Canadian Foundation for Innovation (CFI), and the Natural Sciences and Engineering Research Council of Canada (NSERC) for their financial support, Cultures Dolbec, inc. for supplying the potatoes, Ludovic Béland, Marie-Amélie Bélanger and Serge-Olivier Kotchi for their valuable help in the greenhouse, and Charles Belzile, Simon Roy, Nelson Landry and Stéfan Parmentier for their technical advice.
Adams, M., W. Norvell, J. Peverly, and W. Philpot. 1993. Fluorescence and Reflectance Characteristics of Manganese Deficient Soybean Leaves - Effects of Leaf Age and Choice of Leaflet. Plant Soil 156:235-238.
Allison, M., J. Fowler, and E. Allen. 2001. Factors affecting the magnesium nutrition of potatoes (Solanum tuberosum). Journal of Agricultural Science 137:397 - 409.
Apostol, S., A.A. Viau, N. Tremblay, J.-M. Briantais, S. Prasher, L.-E. Parent, and I. Moya. 2003. Laser induced fluorescence signatures as a tool for remote monitoring of water and nitrogen stresses in plants. Canadian journal of remote sensing 29:57-65.
Belzile, C., M.-C. Bélanger, A.A. Viau, M. Chamberland, and S. Roy. 2003. An operational system for crop assessment, p. 244-252, In B. S. Bennedsen, et al., eds. Proceedings of Photonic East conference: Monitoring Food Safety, Agriculture, and Plant Health, Vol. 5271. SPIE, Providence, RI.
Bilger, W., M. Veit, L. Schreiber, and U. Schreiber. 1997. Measurement of leaf epidermal transmittance of UV radiation by chlorophyll fluorescence. Physiologia plantarum 101:754-763.
Brach, E.J., A.R. Mack, and G.T. St-Amour. 1981. Mobile field laboratory instrumentation to measure spectral characteristics of agricultural crops 626. Engineering and statistical research institute -Institut de recherche technique et statistique.
Carter, G.A., and B.A. Spiering. 2002. Optical properties of intact leaves for estimating chlorophyll concentration. Journal of environmental quality 31:1424-1432.
CCT. 2004. RADARSAT [Online]. Available by Centre canadien de télédétection, Ressources naturelles Canada http://www.ccrs.nrcan.gc.ca/ (posted 2004-10-12; verified 2004-11-12).
Cerovic, Z.G., G. Samson, F. Morales, N. Tremblay, and I. Moya. 1999. Ultraviolet-induced fluorescence for plant monitoring: present state and prospects. Agronomie 19:543-578.
Chappelle, E.W., J.E. McMurtrey III, F.M.J. Wood, and W.W. Newcomb. 1984. Laser-induced fluorescence of green plants. 2: LIF caused by nutrient deficiencies in corn. Applied Optics 23:139-142.
CNES. 2004. SPOT 4 [Online]. Available by Centre national d'études spatiales http://spot4.cnes.fr/spot4_fr/index.htm (posted 200-06-06; verified 2004-10-14).
Corp, L.A., J.E. McMurthey III, E.M. Middleton, C.L. Mulchi, E.W. Chappelle, and C.S.T. Daughtry. 2003. Fluorescence sensing systems: In-vivo detection of biophysical variations in field corn due to nitrogen supply. Remote Sensing of Environment 86:470-479.
Feiffer, P., A. Feiffer, R. Schwaiberger, P. Leithold, J. Jasper, and A. Link. 2003. Hydro N-Sensor. International conference on crop harvesting and processing 701p1103.
Gamon, J.A., L. Serrano, and J.S. Surfus. 1997. The photochemical reflectance index: an optical indicator of photosynthetic radiation use efficiency across species, functional types, and nutrient levels. Oecologia 112:492-501.
Gitelson, A.A., and M.N. Merzlyak. 1994. Spectral reflectance changes associate with autumn senescence of Aesculus hippocastanum L. and Acer platanoides L. leaves. Spectral features and relation to chlorophyll estimation. Journal of Plant Physiology 143:286-292.
Gunther, K.P., H.-G. Dahn, and W. Ludeker. 1994. Remote sensing vegetation status by laser-induced fluorescence. Remote Sensing of Environment 47.
Haboudane, D., J.R. Miller, N. Tremblay, P.J. Zarco-Tejada, and L. Dextrase. 2002. Integrated narrow-band vegetation indices for prediction of crop chlorophyll content for application to precision agriculture. Remote Sensing of Environment 81:416-426.
Heisel, F., M. Sowinska, J.A. Miehé, M. Lang, and H.K. Lichtenthaler. 1996. Detection of nutrient deficiencies of maize by laser induced fluorescence imaging. Journal of Plant Physiology 148:622-631.
Heller, R., R. Esnault, and C. Lance. 1993. Physiologie végétale 1.Nutrition. 5e ed. Masson, Paris, France.
Houghland, G.V.C. 1964. Nutrient deficiencies in the potato, p. p.219-244, In H. B. Sprague, ed. Hunger signs in crops, 3rd ed. David Mackay company, New york, NY.
ITRES. 2004. Sensor systems [Online]. Available by ITRES http://www.itres.com/docs/sensors.html (posted spring 2002; verified 2004-11-01).
Jenkins, P.D., and S. Mashmood. 2003. Dry matter production and partitioning in potato plants subjected to combined deficiencies of nitrogen, phosphorus and potassium. Annals of applied biology 143:215-229.
Klecka, W.R. 1980. Discriminant analysis. 1 ed. SAGE publications, Beverley Hills, CA.
Lichtenthaler, H.K. 1990. Applications of chlorophyll fluorescence in stress physiology and remote sensing, p. 287-305, In S. M.D. and C. J.A., eds. Applications of remote sensing in agriculture. Butterworhs Publisher.
Link, A., M. Panitzki, and S. Reusch. 2003. Hydro N-Sensor: tractor-mounted remote sensing for variable nitrogen fertilization, p. 1012-1017, In ASA-CSSA-SSSA, ed. 6th International Conference on Precision Agriculture and Other Precision Ressources Management. American Society of Agronomy, Minneapolis, MN.
Ma, B.L., L.M. Dwyer, C. Costa, E.R. Cober, and M.J. Morrison. 2001. Early prediction of soybean yield from canopy reflectance measurements. Agronomy Journal 93:1227-1234.
Masoni, A., L. Ercoli, and M. Mariotti. 1996. Spectral properties of leaves deficient in iron, sulfur, magnesium and manganese. Agronomy journal 88:937-943.
Mazzinghi, P. 1990. Stress assessment of vegetation by measurement of chlorophyll fluorescence, p. 509-510 Global natural ressources monitoring and assessments: proceedings, Vol. 1. American society for photogrammetry and remote sensing, Maryland.
McMurthey III, J.E., E.W. Chappelle, M.S. Kim, J.J. Meisinger, and L.A. Corp. 1994. Distinguishing Nitrogen fertilization levels in field corn (Zea mays L.) with actively induced fluorescence and passive reflectance measurements. Remote Sensing of Environment 47:36-44.
Mercure, S.-A., B. Daoust, and G. Samson. 2004. Causal relationship between growth inhibition, accumulation of phenolic metabolites, and changes of UV-induced fluorescence in nitrogen-deficient barley plants. Canadian journal of botany 82:815-821.
Merzlyak, M.N., A.A. Gitelson, O.B. Chivkunova, and Y. Rakitin. 1999. Non-destructive optical detection of pigment changes during leaf senescence and fruit ripening. Physiologia plantarum 106:135-141.
Merzlyak, M.N., A.A. Gitelson, O.B. Chivkunova, A.E. Solovchenko, and S.I. Pogosyan. 2003. Application of reflectance spectroscopy for analysis of higher plant pigments. Russian journal of plant physiology 50:704-710.
Méthy, M., B. Lacaze, and A. Olioso. 1991. Perspectives et limites de la fluorescence pour la télédétection de l'état hydrique d'un couvert végétal: cas d'une culture de soja. International Journal of Remote Sensing 12:223-230.
Méthy, M., A. Olioso, and L. Trabaud. 1994. Chlorophyll fluorescence as a tool for management of plant ressources. Remote Sensing of Environment 47:2-9.
Nilsson, M. 1996. Estimation of tree heights and stand volume using an airborne lidar system. Remote Sensing of Environment 56:1-7.
Oppelt, N., and W. Mauser. 2004. Hyperspectral monitoring of physiological parameters of wheat during a vegetation period using AVIS data. International Journal of Remote Sensing 25:145-159.
Penuelas, J., F. Baret, and I. Filella. 1995. Semi-empirical indices to assess carotenoids/chlorophyll a ratio from leaf spectral reflectance. Photosynthetica 31:221-230.
Qifa, Z., and W. Jihua. 2003. Leaf and spike reflectance spectra of rice with contrasting nitrogen supplemental levels. International Journal of Remote Sensing 24:1587-1593.
Radtke, W., and W. Rieckmann. 1991. Maladies et ravageurs de la pomme de terre. 1st ed., Gelsenkirchen-Buer.
Read, J.J., L. Tarpley, J.M. Mckinion, and K.R. Reddy. 2002. Narrow-waveband reflectance ratios for remote estimation of nitrogen status in cotton. Journal of environmental quality 31:1442-1452.
Rosema, A., G. Cecchi, L. Pantani, B. Radicatti, M. Romuli, P. Mazzinghi, O. van Kooten, and C. Kliffen. 1992. Monitoring photosynthetic activity and ozone stress by laser-induced fluorescence in trees. International Journal of Remote Sensing 13:737-751.
SAS, I. 1990. The STEPDISC procedure, p. 1493-1509, In I. SAS, ed. SAS/Stat User's guide, version 6, Vol. 2, 4th ed, Cary, NC.
Samson, G., N. Tremblay, A.E. Dudelzak, S.M. Babichenko, L. Dextrase, and J. Wollring. 2000. Nutrient stress of corn plants: early detection and discrimination using a compact multiwavelength fluorescent lidar, p. 214-223 4th EARSeL Workshop Lidar Remote Sensing of Land and Sea, Vol. 1, Dresden, Germany.
Schweiger, J., M. Lang, and H.K. Lichtenthaler. 1996. Differences in fluorescence excitation spectra of leaves between stressed and non-stressed plants. Journal of Plant Physiology 148:536-547.
Sims, D.A., and J.A. Gamon. 2002. Relationships between leaf pigment content and spectral reflectance across a wide range of species, leaf structures and developmental stages. Remote Sensing of Environment 81:337-354.
Subbash, N., and C.N. Mohanan. 1994. Laser-induced red chlorophyll fluorescence signatures as nutrient stress indicator in rice plants. Remote Sensing of Environment 47:45-50.
Taiz, L., and E. Zeiger. 1998. Plant physiology. 2nd ed. Sinauer Associates.
Tarpley, L., K.R. Reddy, and G.F. Sassenrath-Cole. 2000. Reflectance indices with precision and accuracy in predicting cotton leaf nitrogen concentration. Crop science 40:1814-1819.
Tartachnyk, I., and I. Rademacher. 2003. Estimation of Nitrogen Deficiency of Sugar beet and Wheat using Parameters of Laser Induced and Pulse Amplitude Modulated Chlorophyll Fluorescence. Journal of applied botany 77:61-67.
Tukaki, J.L., and R.L. Mahler. 1990. Evaluation of nutrient solution phosphorus concentration on potato plantlet tuber production under greenhouse conditions. Journal of plant nutrition 13:149-168.
Varvel, G.E., J.S. Schepers, and D.D. Francis. 1997. Ability for in-season correction of nitrogen deficiency in corn using chlorophyll-meters. Soil Science Society of America Journal 61:1233-1239.
Vouillot, M.O., P. Huet, and P. Boissard. 1998. Early detection of N deficiency in a wheat crop using physiological and radiometric methods. Agronomie 18:117-130.
Zarco-Tejada, P.J., J.R. Miller, G.H. Mohammed, T.L. Noland, and P.H. Sampson. 2002. Vegetation stress detection through Chlorophyll a + b estimation and fluorescence effects on hyperspectral imagery. Journal of environmental quality 31:1433-1441.
Zhao, D., J. Li, and J. Qi. 2004. Hyperspectral characteristic analysis of a developing cotton canopy under different nitrogen treatments. Agronomie 24:463-471.
Zhao, D., K.R. Reddy, V.G. Kakani, J.J. Read, and G.A. Carter. 2003. Corn (Zea mays L.) growth, leaf pigment concentration, photosynthesis and leaf hyperspectral reflectance properties as affected by nitrogen supply. Plant and Soil 257:205-217.
© Marie-Christine Bélanger, 2005