WO2010099203A1

WO2010099203A1 - Methods and device for non-destructive measurement of relative water content in plants

Info

Publication number: WO2010099203A1
Application number: PCT/US2010/025237
Authority: WO
Inventors: Arthur Zygielbaum; Donald Rundquist; Timothy Arkebauer; Anatoly Gitelson
Original assignee: University of Nebraska Lincoln; University of Nebraska System
Current assignee: University of Nebraska Lincoln; University of Nebraska System
Priority date: 2009-02-24
Filing date: 2010-02-24
Publication date: 2010-09-02
Anticipated expiration: 2011-08-24

Abstract

Methods and devices for non-destructively measuring relative water content (RWC) in plants using visible light are provided.

Description

METHODS AND DEVICE FOR NON-DESTRUCTIVE MEASUREMENT OF RELATIVE WATER CONTENT IN

PLANTS

CROSS REFERENCE TO RELATED APPLICATIONS This application claims benefit under 35 U.S. C. 119(e) to U.S. Application No.

61/208,581 filed February 24, 2009.

TECHNICAL FIELD

This disclosure generally relates to methods and devices for non-destructively measuring relative water content and detecting water stress in plants.

BACKGROUND

The character of light reflected by plant leaves is influenced by pigments, internal structure, and water. The spectrum of light can be divided into three major regions: Photosynthetically Active Radiation (PAR), Near-infrared (NIR), and Mid-infrared (MIR). Light in PAR wavelengths (400 nm to 700 nm), sometimes called the visible spectrum, provides the energy required for photosynthesis in plants. Reflectance in PAR is dominated by absorption by pigments such as chlorophyll. NIR (800 nm to 1300 nm) reflectance is largely influenced by plant cell and canopy structures and to a lesser extent by water. In the MIR (1300 nm to 3000 nm), water absorption dominates reflectance and the structure has a secondary influence.

Water molecules absorb radiation in NIR and MIR wavelengths near 970, 1240, 1400, and 1900 nm. Because the amount of absorption is related to total water content, these wavelengths are traditionally used in non-destructive and remote estimation of vegetation water content (e.g., Ceccato, et. al., 2001, Remote Sens. Environ., 77:22-33; Hunt & Rock, 1989, Remote Sens. Environ., 30:43-52). However, each of the existing methods used to evaluate water content in plants has different sources of error. SUMMARY

In one aspect, a device for measuring the relative water content (RWC) in one or more plants is provided. Such a device generally includes a light source; at least one sensor for detecting the amount of photosynthetically active radiation (PAR) reflectance at a first wavelength (λl) in the visible spectrum and the amount of PAR reflectance at a second wavelength (λ2) in the visible spectrum, and an output; and a microprocessor having an input for reading the output, wherein the microprocessor determines an index value that correlates the amount of PAR reflectance at the first wavelength and the amount of PAR reflectance at the second wavelength with the RWC in the one or more plants.

In one embodiment, the device is a hand-held device (e.g., a leaf clip). In another embodiment, the device is suspended above a plurality of plants.

In certain embodiments, such a device further includes a visual indicator of the RWC. In certain embodiments, such a device further includes means for transmitting the amount of PAR reflectance at the first wavelength (λl), the amount of PAR reflectance at the second wavelength (λl), the index value, and/or the RWC to a remote indicator.

In another aspect, a method of non-destructively measuring the relative water content (RWC) in one or more plants is provided. Such a method typically includes the steps of: measuring the amount of PAR reflectance of a leaf at a first wavelength (λl) in the visible spectrum; measuring the amount of PAR reflectance of a leaf at a second wavelength (λl) in the visible light range; determining an index value between the amount of PAR reflectance at the first wavelength and the amount of PAR reflectance at the second wavelength, wherein the index value is indicative of the RWC in the one or more plants. Representative methods of determining an index value include using a simple ratio equation, a normalized difference equation, or a reciprocal difference equation.

In one embodiment, the amount of PAR reflectance of a leaf is measured at the surface of the leaf. In another embodiment, the amount of PAR reflectance of a leaf is measured at a distance above the plant. In certain instances, the measuring is performed at least once a day.

In one particular non- limiting embodiment, the first wavelength (λl) is 580 nm (± 5 nm) and the second wavelength (λ2) is 540 nm (± 5 nm). In another particular non-limiting embodiment, the first wavelength (λl) is 520 nm (± 5 nm) and the second wavelength (λ2) is 720 nm (± 5 nm).

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the methods and compositions of matter belong. Although methods and materials similar or equivalent to those described herein can be used in the practice or testing of the methods and compositions of matter, suitable methods and materials are described below. In addition, the materials, methods, and examples are illustrative only and not intended to be limiting. All publications, patent applications, patents, and other references mentioned herein are incorporated by reference in their entirety.

DESCRIPTION OF DRAWINGS

Figure 1 is a graph showing the relative water content (RWC) plotted versus date. Bottom (triangles), middle (squares), and top (diamond) leaves of two treated plants, and the middle leaf (circle) of two control plants are shown. Closed symbols denote the first experiment; open symbols denote the second experiment. Lines represent the best fit functions. The graph represents 28 individual plants for each experiment.

Figure 2 is graphs showing the reflectance spectra for (A) treated and (B) control leaves. The spectra (solid traces) represent the average reflectance of eight middle leaves from the first experiment. Treated plants were not watered during the seven days and exhibited increasing reflectance from day 1 through day 7. Control plants were watered each afternoon. Coefficients of variation (dashed traces) also are shown.

Figure 3 is a graph showing the average PAR albedo of leaves from treated and control plants plotted over time. The bottom (triangles), middle (squares), and top (diamonds) leaves of two treated plants are shown along with the middle (circles) leaves of two control plants. Closed symbols denote the first experiment; open symbols denote the second experiment. The points represent the daily average of PAR albedo for eight separate treated plants not watered during each seven-day experiment and eight separate control plants watered daily.

Figure 4 is a graph showing PAR albedo plotted versus relative water content (RWC). Square symbols represent the average PAR albedo from the middle leaf of treated plants not watered during the seven-day experiments; and triangle symbols represent the average PAR albedo from the middle leaf of control plants watered daily. Closed symbols indicate the first experiment, and open symbols indicate the second experiment. The solid line is the quadratic best fit function for all treated plant leaves.

Figure 5 is a graph showing the RWC measured in treated leaves (first experiment, middle) plotted versus the ratio of reflectances at 520 and 720 nm with 10 nm bandwidth.

Figure 6 is a matrix of determination coefficients (R²) for all pairs of λl and λ2 in the region from 400 nm to 1000 nm.

Figure 7 is a graphs showing the coefficients of determination between RCW and a normalized difference model index where 1 is fixed at 580 nm and 2 is varied from 400 nm to 700 nm. This profile is taken in the PAR region along the red line shown in Figure 6.

Figure 8 is a schematic flow chart of the method described herein.

DETAILED DESCRIPTION

The movement and use of water within plants are, in themselves, surprisingly complex topics. At the simplest level, a plant must extract water from the soil and distribute it throughout its structure. Water is needed for metabolic processes, to transport nutrients, and to supply the hydraulic pressure needed for plant growth and structural integrity. The plant must control the level of water within its cells and structures despite significant variations in soil and atmospheric conditions. When those conditions become too extreme, the plant cannot maintain a minimum water level and becomes stressed. The closure of leaf stomata, which accompanies water stress, is an attempt by the plant to avoid the loss of water. While this could result in a stabilization of the water potential, the plant increases the solutes in the water within the cells, thereby increasing osmotic pressure and decreasing leaf water potential. Closed stomata decrease gas exchange at the leaf surface, which also decreases the amount of the water vapor near the leaf surface. The cooling effects of transpiration are lost, causing temperatures to increase. Water deficit in the leaves causes a loss of hydrostatic pressure. While normal variations in cell volume, such as diel cycles, can be accommodated by cell wall elasticity, at some point, the cells shrink enough to cause leaf wilt.

For the purposes herein, water stress is considered to be a deficit of available water which causes a decrease in plant growth rate or rate of photosynthesis. Wilting would be an extreme case of water stress. In the literature, water potential is the most often used indicator of stress. Relative water content (RWC) is also considered an indicator. Mild water stress has been defined as a lowering of water potential by several bars (one bar is 0.1 MPa) or of RWC by 8-10% below that in a well watered plant; moderate stress defined as water potential decreasing 12-15 bars (1.2-1.5 MPa) or RWC decreasing by 10-20%; and severe stress defined by water potential lowered by more than 15 bars (1.5 MPa) or RWC more that 20%. Desiccation was defined to be the state where more than 50% of the tissue water was removed. Hsiao, 1973, Plant Responses to Water Stress, Ann. Rev. Plant Physiol, 24:519- 70.

Measuring the water status of a plant is surprisingly complex. The status is a combination of the amount of water contained in the plant, the need the plant has for water based on where it is in its life-cycle, and the dryness of the ground and air in which it is living. Two major characteristics of water in plants are typically discussed in the literature. The first is in terms of quantity, and the second is in the terms of "energy status."

Quantity relates to the amount of water within the elements of the plant. It is often described as relative water content (RWC). RWC compares the actual leaf water content against the content at full turgor:

FW - DW

RWC = (Eq. 1)

TW - DW where FW is the fresh weight of the leaf, DW is the dry weight, and TW is the weight of the leaf at full turgor. Energy status of water in plants is described, by analogy, to electrical potential and current. Water potential represents the work involved in moving one mole of water at constant temperature and pressure from the conditions at some point in a plant, for example, to a pool of pure water at atmospheric pressure and at some zero reference value for gravity (analogous to zero volts in electricity). Water potential is often measured in units of pressure. The primary contributors to total water potential are hydrostatic or pressure potential (Ψ_P), osmotic potential (Ψ_π ), and gravitational potential (Ψ_g ). ψ = ψ^ + ψ_π + ψ_g (Eq. 2)

Total water potential varies directly with hydrostatic and gravitational pressures and inversely with osmotic pressure. The difference in water potential between the leaves and roots supports water flow between the two. Primarily because of transpiration, leaves continually lose water. Water loss lowers the hydrostatic pressure and therefore the leaf water potential. Osmotic pressure differences across membranes in the plant also tend to cause the water potential in the leaf to be more negative with respect to the roots.

Several instruments are available to determine water potential. The most common are the thermocouple psychrometer and the pressure bomb. Thermocouple psychrometers work by sealing a plant tissue sample in a small chamber with a thermocouple and allowing it to reach equilibrium. At equilibrium, the water potential of the sample is equal to the water potential of air (related to relative humidity). A cooling current is passed through the thermocouple and water condenses on the junction. When the current is shut off, subsequent evaporation cools the thermocouple, inducing current across the junction. The evaporation rate determines the degree of cooling and the magnitude of the current or output signal. The plant sample output is then compared to similar outputs obtained from calibration standards of known water potential. The pressure bomb is a method in which an excised leaf is placed within a sealed chamber with its stem exposed to air through a seal. Since the water in the leaf and stem was under tension while the leaf was transpiring, water will be sucked into the leaf when the stem is cut. As pressure is increased in the chamber, water will flow back to the cut end. The chamber pressure at which the water appears at the end of the stem is equivalent to the water potential within the leaf.

The effects of water on the reflected spectrum have been studied in an attempt to measure the amount of water in plants, principally because the spectrum of light returned from vegetation has been successfully used to determine the concentrations of vital pigments, such as chlorophyll (Gitelson et al., 2003, J. Plant Physiol, 160:271-82). Water causes strong absorption of light in the near infrared (NIR) and middle infrared (MIR) portions of the spectrum and causes changes in the level of reflectance throughout the spectrum (Zygielbaum et al., 2009, Geophy. Res. Lett., 36:L12403). Attempts to develop indexes based on a mathematical function of the amount of reflected light at various wavelengths for use as a proxy of water status have met with limited success. Most of these efforts have relied upon the direct absorption by water molecules in the NIR and MIR ranges. The results have not achieved the precision measurements necessary to benefit agriculture by detecting early stages of water stress, before the vegetation undergoes irreversible damage.

Several authors have documented an increase in reflectance in the visible spectrum when plants become stressed (Carter, 1991, Am. J. Botany, 78:916-24; Carter, 1993, Am. J. Botany, 80:239-43; Yu et. al, 2000, Plant and Soil, 227:47-58; Ceccato et. al, 2001, Remote Sens. Environ., 77:22-33; Aldakheel & Danson, 1997, Int. J. Remote Sensing, 18:3683-90). However, this effect has never been able to be quantified and employed to retrieve vegetation water content. This document demonstrates a quantitative relationship between RWC in leaves and visible spectrum reflectance, and establishes a non-destructive technique using visible spectrum reflectance to accurately estimate leaf RWC. Specifically, this document demonstrates a strong, systematic, and repeatable relationship between photosynthetically active radiation (PAR, 400-700 nm) reflectance or PAR albedo and leaf RWC. It is shown herein that visible spectrum reflectance provides a means to quantify leaf RWC. The ability to quantify RWC in plants allows early detection of plant stress before irreversible damage has occurred, and also allows for an assessment of the fire risk.

The correlation described herein between visible light and RWC can be used to obtain a RWC value as an indication of, for example, water stress or fire risk. In the Examples herein, an equation similar to a Normalized Difference Vegetation Index (NDVI) was applied to convert the reflectance at the identified wavelengths into an index value, which is then related to the RWC. As exemplified herein, a normalized difference index refers to the difference between the amount of PAR reflectance (p) at the first wavelength (λl) and the amount of PAR reflectance at the second wavelength (λ2) divided by the sum of the amount of PAR reflectance at the first wavelength and the amount of PAR reflectance at the second wavelength (Eq. 3).

The resulting index value then can be converted into a measure of the RWC in the plant(s). In addition to the normalized difference equation form, there are a number of other forms that can be used to determine an appropriate index value. For example, a simple ratio of reflectance at one wavelength to that at another wavelength, a more complex form such as that used in the 3-band model for pigment content retrieval (Gitelson et al., 2006, Geophys. Res. Lett., 33:L11402), or even a fairly complex equation such as the Global Vegetation Moisture Index (Ceccato et al., 2002, Remote Sens. Environ., 82:188-97; Ceccato et al., 2002, Remote Sens. Environ., 82:198-207) may be applicable depending on the peculiarities of a vegetation species or the phonological status of a plant in that species. Those of skill in the art could determine which form provides the smallest root mean squared error (RMSE) between the biophysical parameter value (in this case, RWC) predicted by the calibrated index (or model) and that measured physically among a set of test plants.

In one embodiment, a device is provided that uses the correlation described herein to measure the relative water content (RWC) in one or more plants. A device for measuring the RWC in one or more plants as described herein typically includes at least one light source, at least one sensor, and a microprocessor. For the example disclosed herein, two wavelengths were used to correlate the reflectance with the RWC. In this embodiment, a broad light source can be used with or without a filter to deliver the light, and one or more sensors, with or without a filter, can be used to detect the reflected light at the appropriate wavelength. Light sources and sensors are well known in the art. Without limitation, light sources can include, for example, LEDs, while sensors can include, for example, photodiodes, phototransistors, or photoresistors. The light source and the sensors are connected, via outputs, to the microprocessor.

The microprocessor in a device as described herein receives the data (i.e., the amount of PAR reflectance at the first and second wavelengths) and determines an appropriate index value (using, e.g., a simple ratio, normalized difference, or reciprocal difference equation). The term "microprocessor" is used in its broadest sense to describe any computing device. In addition to devices generically known as microprocessors, the term includes, by way of example and not limitation, microcontrollers, RISC devices, ASIC devices manufactured to provide logical and mathematical functions, FPGA devices programmed to provide logical and mathematical functions, computers made up of a plurality of integrated circuits, and the like.

A device as described herein for measuring the RWC in one or more plants can be, for example, a hand-held device similar to those already on the market for detecting chlorophyll levels in plant leaves (see, for example, the SPAD Chlorophyll Meter by Minolta or the PlantPen by Photon Systems Instruments). Those of skill in the art may refer to such a device as a leaf clip. In addition to the components of a device as described herein for detecting a plant's RWC (e.g., at least one light source, at least one sensor, and a microprocessor), a hand-held device further can include the components and/or processing instructions for detecting chlorophyll levels in a plant or another plant characteristic (see, e.g., US 2008/0002185). A device as described herein for measuring the RWC in one or more plants also can be, for example, suspended above a plurality of plants (e.g., a field) using any type of pole, post, or existing machinery such as irrigation equipment. In some embodiments, the methods and devices described herein can take advantage of spectral imaging devices that can be used at a distance (e.g., on board a vehicle such as an aircraft or a satellite) to image and acquire data within a wide area such as an entire field or geographical region.

For the ease of a user, a device as described herein typically includes means for generating an output (e.g., an indicator). Such an indicator may be a visual indicator that informs the user of the amount of PAR reflectance at the first wavelength (λl), the amount of PAR reflectance at the second wavelength (λl), the index value determined from λl and λl, and/or the RWC. Alternatively or additionally to a visual indicator on the device, a device as described herein can include means for transmitting the amount of PAR reflectance at the first wavelength (λl), the amount of PAR reflectance at the second wavelength (λl), the index value, and/or the RWC to a remote indicator. In one embodiment, the means for transmitting the indicated information is wireless.

A device as described herein can be calibrated, for example, by measuring the reflectance from a reference surface such as a white piece of paper or plastic with known reflective properties.

In one embodiment, a device for measuring RWC in soy and maize is able to detect reflectance at a first wavelength (λl) of 575 nm - 585 nm (e.g., a central band at 580 nm with a 10 nm bandwidth) and reflectance at a second wavelength (λl) of 535 nm - 545 nm (e.g., a central band at 540 nm with a 10 nm bandwidth). As indicated herein, however, other suitable pairs of wavelengths (λl and λl) can be identified as necessary and a device with appropriate sensors and/or filters designed. Figure 8 illustrates a schematic flow chart showing the manner by which the two wavelengths are used to calculate an index value that is statistically significantly related to the RWC in a plant.

In accordance with the present invention, there may be employed conventional molecular biology, microbiology, biochemical, and recombinant DNA techniques within the skill of the art. Such techniques are explained fully in the literature. The invention will be further described in the following examples, which do not limit the scope of the methods and compositions of matter described in the claims.

EXAMPLES Example 1 — Maize Plants

For each of the two experiments, fifty maize plants (DeKaIb DKC 63-46) were grown in a greenhouse. No artificial illumination was used. Seeds were planted in a mixture of 1/3 peat moss, 1/3 greenhouse soil (silty clay loam) and 1/3 perlite by volume in single pots. The pot size was 7.6 liters (approx. 0.22 m diameter by 0.20 m height). Fertilizer was applied to ensure nutrient sufficiency. Plants emerged approximately eight weeks prior to initiation of the experiments. Phenologically, the plants were V18 to VT (Ritchie et al., 1997, Iowa State University of Science and Technology Cooperative Extension Service, Special Report No. 48., Iowa State University, Ames, IA) during both experiments. Ancillary data, including outside and inside greenhouse temperature and humidity, and outside visible and NIR downwelling irradiance, were recorded.

Example 2 — Maize Treatments

Half of the plants were randomly selected for treatment. Treated plants were not watered during the course of the seven days. The untreated (control) plants were watered daily after reflectance measurements. Sufficient water was applied to ensure the soil was at field capacity. Several maize plants were randomly selected as buffer plants, placed at the end of each row of pots, and not used for measurement.

Example 3 — Soy Plants and Treatments The procedures followed for the soy plants were essentially the same as those for maize. Fifty-six soy plants were grown to assure a sufficient number of plants. Twenty-two plants were randomly selected for test plants. Twenty-two of the remaining plants were randomly selected as control plants. The remaining plants were watered on the same schedule as the control plants but were not used during the trial. No buffer plants were needed since the combination of relatively small plants and relatively large pots minimized the shading of one plant by another. Example 4 — Reflectance Measurements

Reflectance measurements using a hyperspectral ASD FieldSpec Pro radiometer (350 to 2500 nm) and a self-illuminated leaf probe were performed each afternoon at 3 pm. Data were interpolated to 1 nm spectral resolution. Calibration was performed using a 99% reflective Spectralon reference panel. Adaxial leaf reflectance measurements were made approximately 2 cm from the plant stem. Optically absorbing foam (spectrally flat 4% reflectance) was placed behind the leaf.

The three leaves measured for each maize plant included the middle leaf ("mid" - most likely to become the ear leaf), the leaf positioned two above the middle leaf ("top"), and the leaf positioned two below the middle leaf ("bottom"). Measurements in soy used all three leaves at the fourth trifoliate position, near the center of the plant.

Example 5 — Relative Water Content (RWC) Measurements in Maize

Following the reflectance measurements, two treated and two control plants were randomly selected for gravimetric determination of RWC. Ten 1.0 cm leaf punch samples were taken from each of the leaves indicated above for each plant. Punches were quickly sealed into pre -weighed vials. Differencing the filled vial weights from the empty weights provided the fresh weight (FW) of the tissue. The vials were filled with distilled water and refrigerated in darkness at 5⁰C for 15 hours to allow sample rehydration. The punches were then removed from the vials, the surface was patted dry, and the samples were weighed, providing full turgor weight (TW). Next, the punches were placed into open vials and heated in an oven at 105⁰C for 24 hours. Upon removal, the vials were immediately sealed and weighed. The contents were discarded and the empty vials weighed. The difference between weights provided the dry weight (DW). RWC (%) was calculated using the formula: RWC = 100[(FW-DW) / TW-DW)]. This was used as a proxy for the RWC in all of the plants at each temporal point in the experiment. Due to destructive procedures, sampled plants were discarded. Similar experiments were performed to determine RWC in soy plants.

In a second maize experiment, additional punches were taken from one treated and one control RWC sample plant to determine chlorophyll content analytically (Lichtenthaler, 1987, Methods in Enzymology, 148:350-82). Example 6 — Maize Results

During both experiments, RWC in control plant leaves remained above 90%; mean RWC was 95.8% and 98.7% with coefficients of variation (CV) of 2.9 and 3.3% for experiments 1 and 2, respectively. Leaves from the treated plants showed linearly decreasing RWC during the treatment period (Figure 1). Coefficients of determination, R², were higher than 0.96 for all leaves except the top leaf in the second experiment (R² > 0.9). During the first experiment, RWC decreased from more than 90% to below 50%. In the second experiment, the minimum RWC was 50% in the bottom leaf, and above 60% in the top and middle leaves. Differing ranges of RWC in the two experiments can be attributed to variable environmental conditions. While greenhouse temperature averaged near 28⁰C for both periods, the humidity and downwelling irradiance were slightly different due to the different times of year the two experiments were performed. During the first experiment, downwelling irradiance averaged 670 Wm^"2, and greenhouse relative humidity averaged 27%, while for the second experiment, the average irradiance was 463 Wm^"2 and the average relative humidity 42% .

Although the rates in the decrease in RWC varied between experiments, water loss in the top and middle leaves in both sets of treated plants was similar. The bottom leaves, however, dried at a noticeably faster rate.

The spectra for the eight treated and eight control plants in the first experiment were separately averaged. The mean reflectance of the treated leaves (Figure 2A) increased nearly monotonically throughout the visible (400-750 nm) and MIR (1400-2500 nm) spectrum. The coefficient of variation had four pronounced peaks: in the visible spectrum (reaching 30%), near 1450 nm (>25%), near 2000 nm (>35%) and at 2500 nm (>35%). CV was minimal in the NIR (750 to 1300 nm), near 1650 nm, and around 2200 nm (Figure 2A). The reflectances of control plant leaves were virtually invariant with a maximal CV below 5% (Figure 2B for middle leaves). The outcome of the second experiment was consistent with these results. PAR albedo in all treated plant leaves increased significantly during both experiments. Mid leaf albedo increased from 6.5% to 10.9% during experiment 1 and from 6.3% to 8.3% in experiment 2. The daily albedo CV among the eight leaves measured in each experiment was below 10% and varied non-systematically over the experiment period (Figure 3). As used herein, "albedo" is the fraction of the incident light that is reflected. Here, "PAR albedo" is defined as average reflectance over the PAR region of the spectrum (400 nm to 700 nm)

In both experiments, total chlorophyll content (sum of chlorophyll a and b) remained virtually unchanged in control plant leaves and, during days 1 through 4, in treated plant leaves (CV < 5%). However, during days 4 through 7, chlorophyll in treated plant top and bottom leaves decreased 7-9% and in middle leaves decreased about 15%. This decrease affected reflectance only in the green (around 550 nm) and in the red edge, around 700 nm (Gitelson et al., 2003, J. Plant PhsyioL, 160:271-82). Analysis of the relationship between reflectance and chlorophyll content in maize leaves (Gitelson et al., supra) indicated that none of the observed 1.6% increase in albedo between days 1 and 4 and no more than 1.2% of the observed 2.8% increase in albedo between days 4 and 7 can be attributed to a change in chlorophyll content.

Thus, PAR albedo is a sensitive indicator of water stress in maize. A quantitative relationship between PAR albedo and leaf RWC also was established (Figure 4). PAR albedo can, therefore, be used as an accurate proxy for RWC in maize leaves.

Reflectance systematically increases with decreasing RWC at all PAR wavelengths (λ). However, the increase is not uniform (note CV spectrum in Figure 2A). As indicated herein, several approaches can be used to exploit this characteristic to measure the RWC in one or more plants and thereby prevent water stress or provide early detection of water stress. A model can be based on the non-uniform behavior of reflectance in the visible range due to the interplay among the strong absorption of light in the blue and the red spectra and light scattering pronounced in the green and red edge regions (e.g., Forty, et al, 1996, Remote Sens. Environ., 56:104-17). An iterative search algorithm to identify models with minimal RMSE was applied to the reflectance spectra in the PAR region and RWC data for the middle leaves in the first experiment. It is noted that visualization techniques may also be used, as indicated in Example 7, with some advantages over a search algorithm. Several accurate models were found to estimate RWC. As an example, in Figure 5, the ratio of average reflectance in the range 515 to 525 nm and 715 to 725 nm was plotted versus RWC. The linear regression equation parameters were used to validate the P520 / P720 model by comparing measured and predicted RWC for all remaining leaves in both experiments. The standard error of RWC prediction was less than 8% despite differing environmental conditions in the two experiments and differing RWC and albedo characteristics among the leaves. This model is insensitive to chlorophyll content above 200 mg m^"2 (in green to dark- green leaves).

Without being bound by any particular theory, the spectral signatures upon which these findings are based likely result from changes in leaf anatomy and/or physiology driven by water deficit. Photoprotection mechanisms invoked to prevent damage to photosynthetic processes and structures as well as photodamage itself are possible causes as are changes in indexes of refraction in vacuoles due to increasing solute concentration and surface changes in cellular structural elements.

Example 7 — Soy Results

In soy, PAR albedo increased consistently with decreasing water status in all of the analyzed trials. For RWC values of 70% - 100%, the sensitivity of albedo to RWC was lower in soy than in maize, but was nearly identical in both soy and maize for RWC values below 70%. For example, below 70% RWC, the slope of the line for both maize and soy are nearly the same and show R² values of 0.99 and 0.93, respectively.

When the maize data points were combined into one dataset, the best- fit line had a slope of 0.0011 and intercepts or offsets of 0.0361. Likewise, when the soy data points were combined into one dataset, the best-fit line also had a slope of 0.0011 and intercept of 0.0554. This is evidence that the example model sensitivity is similar for four independent experiments covering two different plant species despite varying environmental conditions.

Example 8 — Summary

The optical properties of water- stressed and non-stressed maize and soy leaves were studied using reflectance spectroscopy in two independent controlled tests. An immediate, consistent and statistically significant increase in visible spectrum reflectance of stressed leaves was detected and documented. A strong and repeatable relationship was found between PAR albedo and leaf RWC. The data indicated a 70% increase in PAR albedo as maize plants became increasingly water stressed (50% RWC). PAR albedo was shown to be sensitive to the very early stages of water stress in maize and the results herein quantify the increasing reflectance in visible spectrum caused by water stress. Further, it has been demonstrated that RWC can be retrieved using a model based on non-destructive reflectance measurements, for example, the ratio of reflectances at 520 nm and 720 nm. The significance of these results is twofold. First, a proxy for plant water stress has been identified based on the PAR spectrum and, second, the PAR spectrum plays an important role in surface / air energy interchange.

Example 9 — Identification of PAR Wavelengths that Correlate with RWC

The form of equation used as well as the wavelength selected may depend on specific plant species or the phonological state of a particular species. In the case of the experiments described herein that were performed on maize and soy, the specific equation and wavelengths were identified through the procedure described below. Briefly, matrices of determination coefficients were examined for several equation forms, and a two-wavelength normalized difference equation was selected as fitting the need for a reliable index and being intrinsically robust against differences between plants. Therefore, a normalized difference equation was used in the following example, although other equations such as simple ratio or the reciprocal difference may work. Figure 6 shows a matrix of determination coefficients (R²) of the relationship between the index in the form

and relative water content, R²1/. _j vs R WCj, for all pairs of λi and λ_j in the region from 400 nm to 1000 nm. Those of skill in the art would understand that a matrix of determination coefficients can be determined a number of ways. For example, a matrix of determination coefficients can be generated using an optimizing iterative search algorithm, a simple but non-exhaustive spreadsheet examination, or a visualization technique.

Figure 6 shows how normalized difference index values relate to RWC for every pair of wavelengths from 400 nm to 1000 nm. Bright regions indicate that R² is near one, while dark regions denote that R² is near zero. There is a fairly large bright region where both X₁ and X₂ range between 530 nm and 580 nm. An index using pairs of wavelengths within this region correlates very well with RWC. While other wavelength combinations also appear to provide good correlations, wavelengths in this region will be used in this Example.

To select an optimal pair of wavelengths, a profile view was generated that holds X₁ constant and varies X₂ through the whole range of measured wavelengths. For example, selecting X₁ as 580 nm produces the profile depicted in Figure 7 (in this case limiting λ₂to the PAR region) which corresponds to the red line in Figure 6.

The second wavelength for the model, λ₂, can be selected from any value in the range from about 535 nm to 575 nm, where R² is near one. If, for instance, one were to pick 550 nm, then the index resulting from the following equation would have a high correlation with RWC:

Index = ^{Ag0 ~ A5}° (Eq. 5)

Although these particular wavelengths are "tuned" to the results of the maize experiments described herein, when the maize and soy experiments were combined, a pair of wavelengths was identified that is very similar to the wavelengths identified with the maize data. Based on these experiments, one of skill could apply the same principles to determine other pairs of wavelengths that correlate well with RWC and, hence, with water stress or, alternatively, fire risk.

The normalized difference model is effective because, within the region with high R² values, there was a close relationship between index values over a fairly large range of wavelengths. This suggests that within that range, the difference between the reflectances, the numerator in Eq. 4, remains almost invariant with changes in RWC. The sums of the reflectances, the denominator, increase with decreasing RWC. Thus, the ratio of an invariant numerator and increasing denominator results in a number which has a high correlation with RWC.

It is to be understood that, while the methods and compositions of matter have been described herein in conjunction with a number of different aspects, the foregoing description of the various aspects is intended to illustrate and not limit the scope of the methods and compositions of matter. Other aspects, advantages, and modifications are within the scope of the following claims.

Claims

WHAT IS CLAIMED IS:

1. A device for measuring the relative water content (RWC) in one or more plants, comprising: a light source; at least one sensor for detecting the amount of photosynthetically active radiation (PAR) reflectance at a first wavelength (λl) in the visible spectrum and the amount of PAR reflectance at a second wavelength (λ2) in the visible spectrum, and an output; and a microprocessor having an input for reading said output, wherein the microprocessor determines an index value that correlates the amount of PAR reflectance at the first wavelength and the amount of PAR reflectance at the second wavelength with the RWC in the one or more plants.

2. The device of claim 1, wherein the device is a hand-held device.

3. The device of claim 1 or 2, wherein the hand-held device is a leaf clip.

4. The device of claim 1 , wherein the device is suspended above a plurality of plants.

5. The device of any of the proceeding claims, further comprising a visual indicator of the RWC.

6. The device of any of the proceeding claims, further comprising means for transmitting the amount of PAR reflectance at the first wavelength (λl), the amount of PAR reflectance at the second wavelength (λ2), the index value, and/or the RWC to a remote indicator.

7. The device of any of the proceeding claims, wherein the first wavelength (λl) is 580 nm (± 5 nm) and the second wavelength (λ2) is 540 nm (± 5 nm).

8. The device of any of the proceeding claims, wherein the first wavelength (λl) is 520 nm (± 5 nm) and the second wavelength (λ2) is 720 nm (± 5 nm).

9. A method of non-destructively measuring the relative water content (RWC) in one or more plants, comprising the steps of: measuring the amount of PAR reflectance of a leaf at a first wavelength (λl) in the visible spectrum; measuring the amount of PAR reflectance of a leaf at a second wavelength (λ2) in the visible light range; determining an index value between the amount of PAR reflectance at the first wavelength and the amount of PAR reflectance at the second wavelength, wherein the index value is indicative of the RWC in the one or more plants.

10. The method of claim 9, wherein the index value is determine using a simple ratio equation, a normalized difference equation, or a reciprocal difference equation.

11. The method of claim 9, wherein the amount of PAR reflectance of a leaf is measured at the surface of the leaf.

12. The method of claim 9, wherein the amount of PAR reflectance of a leaf is measured at a distance above the plant.

13. The method of claim 9, wherein the measuring is performed at least once a day.

14. The method of claim 9, wherein the first wavelength (λl) is 580 nm (± 5 nm) and the second wavelength (λ2) is 540 nm (± 5 nm).