Craig L. Wiegand

Photographic and videographic observations for determining and mapping the response of cotton to soil salinity

Better ways are needed to assess the extent and severity of soil salinity in fields in terms of economic impact on crop production and effectiveness of reclamation efforts. Procedures to help meet these needs were developed from soil salinity, plant height and boll counts, and digitized color infrared aerial photography and videography acquired during midboll set development stage for four salt-affected cotton (Gossypium hirsutum, L.) fields in the San Joaquin Valley of California. Unsupervised classijication procedures were used to produce seven-category spectral maps by field. Regression equations were developed from salinity measurements in the surface 30 cm (EC1) at 100-200 sample sites per field and the photography and videography digital counts at those same sites. The equations were used to estimate the salinity of each of the approximately 100,000 pixels per field, and the salinity categories corresponding to the spectral ones were mapped. The spectral classification maps and the estimated salinity maps corresponded well. Boll counts, made at about 20 sites perjield, were converted to lint yield and regressed on NDVl from both the photography and videography; the correlation coefficient (r) was 0.72 for video and 0.73 for the photographic data. Lint yields decreased by 43 f 10 kg ha-’ per dS mm1 increase in ECl, or

Soil salinity effects on crop growth and yield-illustration of an analysis and mapping methodology for sugarcane

52 f 12 ha-’ at current market prices. Our results illustrate very practical ways to combine image analysis capability, spectral observations, and ground truth to map and quantify the severity of soil salinity and its effects on crops.

View azimuth and zenith, and solar angle effects on wheat canopy reflectance

Summary The effects of soil salinity on growth and yield of sugarcane ( Saccharum spp. hybrids) are used to illustrate how soil and plant samples («ground truths), digital videographic or SPOT HRV spectral observations, and image analysis by unsupervised classification can be used jointly to quantify and map variations in weighted electrical conductivity (WEC, dS m -1 ) of the root zone and YIELD (metric tons of millable stalks ha -1 ). The combined data for the 1992 and 1993 growing seasons of the study showed that each dS m -1 increase in WEC reduced stalk population by 0.6 stalks m -2 , stalk weight by 0.14 kg, and stalk yields by 13.7 metric tons ha -1 . Sugarcane growth and yield were not affected by root zone salinities less than about 2dS m -1 , but no millable stalks were produced at salinities in excess of 10 dS m -1 . The 25 pixels ha -1 of SPOT is a good scale for mapping salt stress patterns and taking site-specific ameliorative actions. The combination of satellite or aerial spectral observations, ground truth, and image classification procedures demonstrated in this study is readily applicable to other vegetation stresses.

Development of Agrometeorological Crop Model Inputs from Remotely Sensed Information

Abstract Since crop canopies are not lambertian reflectors, their reflectance varies with sun and view positions. It is not always possible or convenient to make reflectance measurements from the nadir position nor at the same time of day. Therefore, ways of estimating nadir reflectance from off-nadir views and for various solar zenith angles are needed. In this study, spectral measurements were made with a Mark II radiometer five times during the day on each of four dates from 15° interval zenith and 45° azimuth positions for wheat canopies during the development interval stem extension to watery ripeness of the grain. The ratio of off-nadir [ R ( Z v , A v )] to nadir [ R (0)] radiance in NIR band (0.76–0.90 μm) was described by the regression equation: R(Z v ,A v ) R(0) = 1.0 + [β 0 + β 1 sin ( A v 2 ) + β 2 (1/ cos Z s )] sin Z v where A v is view azimuth angle relative to the sum position, Z s is solar zenith angle, and Z v is view zenith angle. The coefficient of determination was 0.70 or higher. The equation describes the observations that 1) the ratio of off-nadir to nadir radiance increases or decreases as view zenith angle increases depending on view azimuth angle; backscattering is stronger than forwardscattering and the pattern is azimuthally symmetric about the principal plane of the sun; and 2) the rate of change in the radiance ratio increases with increasing solar zenith angle. The coefficients, β 0 , β 1 and β 2 , changed as the canopies grew. Although the equation needs to be more fully tested, it should help summarize and compare various angular observation data taken in crop fields.

Diurnal patterns of bidirectional vegetation indices for wheat canopies

The goal of developing agrometeorological crop model inputs from remotely sensed information (AgRISTARS Early Warning/Crop Condition Assessment Project Subtask 5 within the U. S. Department of Agriculture (USDA)) provided a focus and a mission for crop spectral investigations that would have been lacking otherwise. Because the task had never been attempted before, much effort has gone into developing measurement and interpretation skill, convincing the Scientific community of the validity and information content of the spectral measurements, and providing new understanding of the crop scenes viewed as affected by bidirectional, atmospheric, and soil background variations. Nonetheless, experiments conducted demonstrate that spectral vegetation indices (VI) a) are an excellent measure of the amount of green photosynthetically active tissue present in plant stands at any time during the season, and b) can reliably estimate leaf area index (LAI) and intercepted photosynthetically active radiation (IPAR)-two of the inputs needed in agrometeorological models. Progress was also made on using VI to quantify the effects of yield-detracting stresses on crop canopy development. In a historical perspective, these are significant accomplishments in a short time span. Spectral observations of fields from aircraft and satellite make direct checks on LAI and IPAR predicted by the agrometeorological models feasible and help extend the models to large areas. However, newness of the spectral interpretations, plus continual revisions in agrometeorological models and lack of feedback capability in them, have prevented the benefits of spectral inputs to agrometeorological models from being fully realized.

Comparison of multispectral video and SPOT-1 HRV observations for cotton affected by soil salinity

Abstract The perpendicular vegetation index (PVI) and normalized difference vegetation index (NDVI) were calculated from Mark II radiometer RED (0.63-0.69 μm) and NIR (0.76–0.90 μ) bidirectional radiance observations for wheat canopies. Measurements were taken over the plant development interval flag leaf expansion to watery ripeness of the kernels during which the leaf area index (LAI) decreased from 40 to 2-5. Spectral data were taken on four cloudless days five times (09.30, 11.00, 12.30, 14.00 and 15.30 hours (central standard time, C.S.T.) at five view zenith, Zv (0, 15, 30,45 and 60°) and eight view azimuth angles relative to the Sun, Av (0, 45, 90, 135, 180, 225, 270 and 315°). The PVI was corrected to a common solar irradiance (PVIC) based on simultaneously observed insolation readings. The PVIC at nadir view (Ž=0°) increased as (l/cosZs) increased on all the measurement days whereas the NDVI changed little as solar zenith angle (Zs) changed. Thus, the PVIC responded to increasing path length thro...

Remotely-sensed spectral indicators of sorghum development and their use in growth modeling

Abstract A 15-ha field of cotton (Gossypium hirsuiutn L,) with an erratic pattern of soil salinity was overflown twice during the 1989 growing season with each of two sensor systems, a multispectral videography system and the High Resolution Visible (HRV) scanner of the French polar-orbiting satellite SPOT. The objectives were to compare the responses of the two sensor systems and to relate vegetation indices calculated from them to plant cover and lint yield (kgha−1) observations for 36 sites located at the intersections of a square 60 m grid laid off in the field. The range in responses, expressed as digital counts, was much greater for the video than for the HRV. For paired dates of acquisitions, the agreement between near-infrared and red bands between systems was almost as good as within systems. Yield and plant cover could also be estimated about equally well from vegetation indices calculated from both systems. Consequently, it is concluded that the video observations provided essentially the same ...

Spectral components analysis Rationale, and results for three crops

Abstract Earth Resources Technology Satellite (LANDSAT) multispectral scanner (MSS) data for five overpass dates (3/5, 21/5, 8/6, 28/6 and 1/8) during the 1976 grain sorghum growing season (Bell County, Texas) and weather data were used to estimate the plant growth measurements leaf area index (LAI), biomass, plant height, plant cover, and grain yield. Vegetaton indices derived from LANDSAT data were correlated to crop development and growing conditions as measured with LAI and biomass samples. Largest LAI and perpendicular vegetation index (PVI) calculated from LANDSAT spectral data occurred at the half-bloom (HB) stage of development. The PVI was significantly correlated with sorghum PGM on the earliest sampling date (3/5/76) when vegetative ground cover was only 20%. Regression analysis indicated that the PVI explained 79% of the variaton in estimating LAI over the first four sampling dates. On the other hand, the regression of PVI, solar-thermal units, and a soil water index accounted for 90% of the variation in LAI measurements. This result indicated that plant-growth models that mimic plant response to soil and climatic environments will improve yield estimates over those arrived at from spectral data alone. Since LAI is used in growth models to partition energy between plants and soil in the evapotranspiration subroutine and to estimate light interception in the photosynthesis subroutine, LAI estimates from LANDSAT would provide this information for use either as (1) input data, or (2) feedback data to check on growth model predictions and retrack the model, if necessary.

Mean Effective Optical Constants of Thirteen Kinds of Plant Leaves

Abstract The spectral components analysis identities, LAI/VI × APAR/LAI = APAR/VI and LAI/VI × YIELD/LAI = YIELD/VI, where VI denotes any one of several spectral vegetation indices available, LAI is the leaf area index, APAR is the absorbed photosynthetically active radiation and YIELD is the saleable plant part (grain, fibre or root), express the information conveyed by canopies about their development, response to stresses and yield capability. The rationale includes the concepts that vegetation indices adequately measure the amount of photosynthetic-ally active tissue in plant canopies and that high yields cannot be achieved unless growing conditions permit canopies to develop that effectively intercept the available light during reproduction. For wheat, cotton and maize the coefficients of determination (r 2) usually exceeded 0.90 for exponential, power or linear expressions relating the numerator (dependent variable, y) to the denominator (x) variable of each term in the first equation. Results show ...

Water turbidity and perpendicular vegetation indices for paddy rice flood damage analyses

Plant leaves grown in a greenhouse and leaves collected from the field have been analyzed to obtain mean effective optical constants based upon diffuse reflectance and transmittance measurements taken over the 0.5-2.5-micro spectral range. These optical constants are used in a generalized flat-plate model to describe the phenomena of leaf reflectance. Analysis procedures developed led to measures of the amount of water and intercellular air spaces in the leaves. Over the 1.4-2.5-micro spectral range, the absorption spectra of leaves are not statistically different from that of pure liquid water. Leaf reflectance differences among the plant leaves over the 0.5-1.4 micro range are caused principally by Fresnel reflections at external and internal leaf surfaces and by plant pigment absorption. Reflectance over the 1.4-2.5-micro range results largely from Fresnel reflections and absorption by water. Data are presented in the form of dispersion curves with 95% confidence bands and tabulated plant leaf absorption spectra. The dispersion curves were assumed to be cubic equations of the form n = Sigmaa(i)lambda(i) (i = 0, 1, 2, 3), where lambda is wavelength. Reflectance measurements at 1.65 micro have been associated with the equivalent water thickness and the intercellular air spaces in the leaf. Accuracy of the plate theory based upon a cubic dispersion curve is shown to be within experimental error.