Stephan J. Maas

Remote sensing and crop production models: present trends

Abstract The use of remote sensing as a tool for estimating crop production on a regional scale has been suggested for quite a while. Since the use of remotely sensed information to directly estimate crop production is questionable (primarily due to the indirect link between remotely sensed data and crop state variables), crop models may be used as a companion tool to remote sensing. Various types of crop models (statistical, deterministic, semi-empirical) are described, and specific methods (forcing, recalibration, statistical correction) are described for introducing remotely sensed information into models. The appropriance of combining different model types and sources of remotely sensed information, problems related to time scales, and the need for robust and independent phenological routines are also discussed.

Use of remotely-sensed information in agricultural crop growth models

Abstract Four techniques are described for incorporating remotely-sensed information into agricultural crop growth models. The techniques are demonstrated using a simple model of the above-ground growth of a uniform maize monoculture. Use of remotely-sensed data as model input is relatively simple but requires frequent observations. Infrequent observations may be used to update the model simulation, but final results are sensitive only to the latest observation. Re-initialization and re-parameterization may be used to adjust model initial conditions and parameter values to fit simulated growth to remotely-sensed observations.

Combining remote sensing and modeling for estimating surface evaporation and biomass production

Abstract A simple approach for simulation of daily regional evaporation and plant primary production is proposed. The approach is based on an existing plant growth model combined with a simple soil water balance equation for simulation of evaporation rates. The resulting model was specifically designed to incorporate periodic remotely‐sensed estimates of plant leaf area index (LAI) and daily surface evaporation (E). The model was evaluated based on spectral, meteorologic, agronomic and soils data acqired during a two‐year experiment in an alfalfa stand at the U.S. Water Conservation Laboratory lysimeter field plots in Phoenix, Arizona. The remotely‐sensed inputs to the model (LAI and E) were obtained from measurements of surface reflectance and temperature, combined with measurements of air temperature. Then, the model was used to simulate daily values of E, LAI and biomass production using infrequently‐acquired remotely‐sensed information and routinely available meteorologic observations. These results i...

Spider Mite Detection and Canopy Component Mapping in Cotton Using Hyperspectral Imagery and Spectral Mixture Analysis

Spectral mixture analysis and hyperspectral remote sensing are analytical and hardware tools new to precision agriculture. They can allow detection and identification of various crop stresses and other plant and canopy characteristics through analysis of their spectral signatures. One stressor in cotton, the strawberry spider mite (Tetranychusturkestani U.N.), feeds on plants causing leaf puckering and reddish discoloration in early stages of infestation and leaf drop later. To determine the feasibility of detecting the damage caused by this pest at the field level, AVIRIS imagery was collected from USDA-ARS cotton research fields at Shafter, CA on 4 dates in 1999. Additionally, cotton plants and soil were imaged in situ in 10 nm increments from 450 to 1050 nm with a liquid-crystal tunable-filter camera system. Mite-damaged areas on leaves, healthy leaves, tilled shaded soil, and tilled sunlit soil were chosen as reference endmembers and used in a constrained linear spectral mixture analysis to unmix the AVIRIS data producing fractional abundance maps. The procedure successfully distinguished between adjacent mite-free and mite-infested cotton fields although shading due to sun angle differences between dates was a complicating factor. The resulting healthy plant, soil, mite-damaged, and shade fraction maps showed the distribution and relative abundance of these endmembers in the fields. These hardware and software technologies can identify the location, spatial extent, and severity of crop stresses for use in precision agriculture.

Apparent Electrical Conductivity, Soil Properties and Spatial Covariance in the U.S. Southern High Plains

Site-specific soil and crop management will require rapid low-cost sensors that can generate position-referenced data that measure important soil properties that impact crop yields. Apparent electrical conductivity (ECa) is one such measure. Our main objective was to determine which commonly measured surface soil properties were related to ECa at six sites in the Texas Southern High Plains, USA. We used the Veris 3100 and Geonics EM-38 EC mapping systems on 12 to 47 ha areas in six center-pivot irrigation sites. Soil samples were taken from 0–150 mm on a 0.1 to 0.8 ha grid and analyzed for routine nutrients and particle size distribution. At four of the six sites, shallow ECa measured with the Veris 3100 (ECa-sh) positively correlated to clay content. Clay content was negatively related with ECa-sh at one site, possibly due to low bulk density of the shallow calcic horizon at that site. Other soil properties that were often correlated with ECa included soil extractable Ca2+, Mg2+, Na+, CEC, silt and soluble salts. Extractable K+, NO3−, SO4−, Mehlich-3-P, and pH were not related to ECa. Partial least squares regression (PLS) of seven soil properties explained an average of 61%, 51% and 37% of the variation in observed shallow ECa-sh, deep ECa with the Veris 3100 (ECa-dp) and ECa with the Geonics EM-38 (ECa-em), respectively. Including nugget, range and sill parameters from a spherical semivariance model of the residuals from PLS regression improved the fit of mixed models in 15 of 18 cases. Apparent EC, therefore can provide useful information to land-users about key soil properties such as clay content and extractable Ca2+, but that spatial covariance in these relationships should not be ignored.

A Model for Calculating Light Interception by a Grain Sorghum Canopy

ABSTRACT A simplified model for calculating intercepted photo-synthetically active radiation (0.4-0.ly) was developed to eliminate the necessity for determining the de-tailed plant foliage characteristics required in many light interception models. To simplify predicting foliar light interception, plant structure is idealized using geometric shapes and a concept of effective light inter-cepting leaf area is introduced. The effective light inter-cepting leaf area is a function of that portion of foliage actually intercepting light and, hence, accounts for mutual shading of plant leaves and shadows cast by neighboring plants. As plant leaf area increases, the equivalent idealized areas representing plant leaf sur-faces intercepting light interact. Interplant competi-tion is determined with mensuration formulas. Photosynthetically active radiation measurements were made simultaneously above and below three grain sorghum canopies for use in developing and validating the model. Different canopy architecture was achieved by using three row spacing regimes (25.4, 50.8, and 101.6 cm, with average within-row plant spacings of 21.2, 10.6 and 5.3 cm, respectively) and equipopulated plantings (185,000 plants/ha). Analysis of light penetration measurements shows the sensitivity of the extinction coefficient to interplant competition and solar altitude. Light attenuation in the model is computed using a Bouguer-Lambert (Beers) Law relationship in which the extinction coefficient is sensitive to adjacent plant light competition interactions. Comparison of measured and modeled transmitted light data for both clear and cloudy days indicated that the model is reliable. The applicability of this model is ex-tended by inclusion of a relationship for equating radiant energy units to quantum units.

Normalizing and Converting Image DC Data Using Scatter Plot Matching

Abstract: Remote sensing image data from sources such as Landsat or airborne multispectral digital cameras are typically in the form of digital count (DC) values. To compare images acquired by the same sensor system on different dates, or images acquired by different sensor systems, it is necessary to correct for differences in the DC values due to sensor characteristics (gain and offset), illumination of the surface (a function of sun angle), and atmospheric clarity. A method is described for normalizing one image to another, or converting image DC values to surface reflectanceThis method is based on the . identification of pseudoinvariant features (bare soil line and full canopy point) in the -scatter plot of red and near-infrared image pixel values. The method, called “scatter plot matching” (SPM), is demonstrated by normalizing a Landsat-7 ETM+ image to a Landsat-5 TM image, and by converting the pixel DC values in a Landsat-5 TM image to values of surface reflectance. While SPM has some limitations, it represents a simple, straight-forward method for calibrating remote sensing image data.

Mapping crop ground cover using airborne multispectral digital imagery

Empirical relationships between remotely sensed vegetation indices and canopy density information, such as leaf area index or ground cover (GC), are commonly used to derive spatial information in many precision farming operations. In this study, we modified an existing methodology that does not depend on empirical relationships and extended it to derive crop GC from high resolution aerial imagery. Using this procedure, GC is calculated for every pixel in the aerial imagery by dividing the perpendicular vegetation index (PVI) of each pixel by the PVI of full canopy. The study was conducted during the summer growing seasons of 2007 and 2008, and involves airborne and ground truth data from 13 agricultural fields in the Southern High Plains of the USA. The results show that the method described in this study can be used to estimate crop GC from high-resolution aerial images with an overall accuracy within 3% of their true values.

Forecasting Grain Sorghum Yields Using Simulated Weather Data and Updating Techniques

ABSTRACT A methodology was developed by which SORGF, a grain sorghum growth-simulation model, could be used to forecast crop status during the growing season. The methodology utilizes simulated weather data, generated by a Markov chain model, as input to the grain sorghum model. The modeled status of sorghum plants may be updated at any time during the growing season with actual plant status observed in the field. Ap-plication of the methodology was demonstrated by forecasting date of physiological maturity (PM) and head dry weight at PM for grain sorghum crops grown in 10 fields in Central Texas during 1976.

Index of Soil Moisture Using Raw Landsat Image Digital Count Data in Texas High Plains

The growth and yield of crops in the arid and semi-arid regions of the world is driven by the amount of soil moisture available to the crop through rainfall and irrigation. Various methods have been developed for quantifying the soil moisture status of agricultural crops. Recent technological advances in remote sensing have shown that soil moisture can be measured with a variety of remote sensing techniques, each with its own strengths and weaknesses. In this study, building on of the strengths of multispectral satellite imagery, a new approach is suggested for estimating soil moisture content. A soil moisture index, the Perpendicular Soil Moisture Index (PSMI), is proposed; it is evaluated using raw image digital count (DC) data in the red, near-infrared, and thermal infrared spectral bands. To test this approach, soil moisture was measured in 18 agricultural fields in the semi-arid Texas High Plains over two years and compared to corresponding PSMI values determined from Landsat image data. These results showed that PSMI was strongly correlated (R2 = 0.79) with observed soil moisture. It was further demonstrated that maps of PSMI developed from Landsat imagery could be constructed to show the relative spatial distribution of soil moisture across a region. While further study is needed to determine the exact relationship between PSMI and soil moisture in larger areas with different climates, this study suggests that PSMI is a good indicator of soil moisture and has potential for operationally monitoring soil moisture conditions at the field to regional scales.