Evaluation of a Hand-held Radiometer for Field Determination of Nitrogen Status in Cotton

Because the timing and amount of N fertilizer applications are important in producing high yield in cotton (Gossypium hirsutum L.), determining changes in plant N status is a valuable management tool. This study evaluated the Spectrum Technologies CM-1000 chlorophyll meter as compared with the GER model 1500 spectroradiometer for remote estimation of leaf chlorophyll concentration (Chl) under N-deficit conditions. In both instruments, measurements of leaf reflectance (R) at 700 and 840 nm (± 5.0 nm) were used to compute R840/R700, a measure associated with general plant stress. Plants were grown outdoors in large pots using half-strength nutrient solution (control) via a drip-irrigation system until some three-row plots received a restricted supply of N. Treatments comprised control N from emergence to maturity, 20% of control N at first floral bud (square) onward, and 0 and 20% of control N at first flower onward. Following 27 July 2000, treatment rankings for Chl and N were control > 20% N first flower > 0%N first flower ≥ 20% N first square. On most sampling dates, significant linear correlation was obtained between Chl and R840/R700 across N treatments, particularly in August (r =0.55–0.93), when stress-induced changes in Chl were 171 Copyright © 2003. ASA–CSSA–SSSA, 677 S. Segoe Rd., Madison, WI 53711, USA. Digital Imaging and Spectral Techniques. Applications to Precision Agriculture and Crop Physiology. ASA Special Publication no. ??. closely related to N. Treatment means for R840/R700 were similar for the CM-1000 and GER 1500; however, SPAD readings were often more closely associated with Chl. This illustrates the difficulty of measuring leaf reflectance under ambient light levels, as compared with measuring transmitted light at two wavelengths in an enclosed chamber. We also determined the effect of a black background (R<<1%) on R840/R700 in field-grown plants fertilized with 0 and 112 kg N ha−1. Placing an absorptive background behind the measured leaf decreased R840/R700 due to proportionately less reflectance at R840, but did not change treatment rankings for R840/R700. In providing a rapid and reliable assessment of Chl, and hence plant N status, the CM-1000 should be useful for nutrient management in cotton. Multispectral and hyperspectral radiometry are powerful tools for assessing plant reflectance at various wavelengths of light energy, where the amount of light not reflected is either absorbed by the plant pigments or transmitted downward to the soil surface (Jackson et al., 1980). It is known that leaf reflectance in the visible region of the spectrum (300–700 nm) varies as a function of chlorophyll (a + b) concentration (Chl). Because the majority of leaf N is contained in protein, N deficiency in cotton causes decreases in plant growth, photosynthesis, and chlorophyll production (Longstreth & Nobel, 1980; Gerik et al., 1989). When Chl begins to decline, the amount of reflected radiance from within the leaf interior increases, providing an optical (and early) indicator of stress at specific wavebands of visible light (Carter, 1993). Increased reflectance in the near infrared (NIR) region of the spectrum (740–900 nm), where light is not absorbed by chlorophyll, is caused by multiple scattering between the water-filled cells and air spaces in the leaf (Jensen, 2000). The actual amount of NIR reflected depends on the anatomical structure of leaves, which is difficult to measure. Knowledge of these leaf optical properties has been used for relative assessment of Chl, and hence N stress in cotton (Tarpley et al., 2000; Read et al., 2002). Similarly, critical N levels established by spectral radiometry were in good agreement with those established by traditional measurements of growth and plant mineral composition in corn (Zea mays L.) (Blackmer et al., 1996). Remote sensing of Chl appears to have potential for field assessment of cotton N status, and hence crop productivity (Gerik et al., 1998; Tenkabail et al., 2000; Barnes et al., 2000; Read et al., 2002). Hand-held radiometers have the advantage of being rapid, nondestructive, and relatively inexpensive, as compared to the collection and analysis of dried, ground plant tissues for N concentration. Moreover, reflectance samples can be taken frequently, and quickly repeated if results are questionable (Jackson et al., 1980). Carter (1993) found several multispectral indices changed in response to altered physiological functions resulting from environmental fluctuations or plant stresses. For instance, the red-edge of leaf reflectance, which is the increase in reflectance between 680 and 740 nm, shifts toward shorter wavelengths in response to N deficiency, leaf senescence, dehydration, flooding, freezing, ozone, herbicides, competition, disease, insects, and deficiencies in ectomycorrhizal development (Carter & Knapp, 2001). The red-edge inflection point is strongly dependent on Chl a concentration, and several studies have related either the reflectance red-edge or reflectance indices involving the red-edge, to changes in leaf Chl concentration (Chappelle et al., 1992; Gitelson et. al., 1996) or general plant health (Carter, 1993; Trenholm et al., 2000). In cotton, single-waveband ratios that combined a low reflectance, red-edge measure with a waveband of 172 READ ET AL.