Edward M. Barnes

Remote Sensing for Crop Management

with the Agricultural Research Service (ARS) and various government agencies and private institutions have provided a great deal of fundamental information relating spectral reflectance and thermal emittance properties of soils and crops to their agronomic and biophysical characteristics. This knowledge has facilitated the development and use of various remote sensing methods for non-destructive monitor- ing of plant growth and development and for the detection of many environmental stresses which limit plant productivity. Coupled with rapid advances in computing and position- locating technologies, remote sensing from ground-, air-, and space-based platforms is now capable of providing detailed spatial and temporal information on plant response to their local environment that is needed for site specific agricultural management approaches. This manuscript, which empha- sizes contributions by ARS researchers, reviews the biophysi- cal basis of remote sensing; examines approaches that have been developed, refined, and tested for management of water, nutrients, and pests in agricultural crops; and as- sesses the role of remote sensing in yield prediction. It con- cludes with a discussion of challenges facing remote sens- ing in the future.

Estimating cotton evapotranspiration crop coefficients with a multispectral vegetation index

Crop coefficients are a widely used and universally accepted method for estimating the crop evapotranspiration (ETc) component in irrigation scheduling programs. However, uncertainties of generalized basal crop coefficient (Kcb) curves can contribute to ETc estimates that are substantially different from actual ETc. Limited research with corn has shown improvements to irrigation scheduling due to better water-use estimation and more appropriate timing of irrigations when Kcb estimates derived from remotely sensed multispectral vegetation indices (VIs) were incorporated into irrigation-scheduling algorithms. The purpose of this article was to develop and evaluate a Kcb estimation model based on observations of the normalized difference vegetation index (NDVI) for a full-season cotton grown in the desert southwestern USA. The Kcb data used in developing the relationship with NDVI were derived from back-calculations of the FAO-56 dual crop coefficient procedures using field data obtained during two cotton experiments conducted during 1990 and 1991 at a site in central Arizona. The estimation model consisted of two regression relations: a linear function of Kcb versus NDVI (r2=0.97, n=68) used to estimate Kcb from early vegetative growth to effective full cover, and a multiple regression of Kcb as a function of NDVI and cumulative growing-degree-days (GDD) (r2=0.82, n=64) used to estimate Kcb after effective full cover was attained. The NDVI for cotton at effective full cover was ~0.80; this value was used to mark the point at which the model transferred from the linear to the multiple regression function. An initial evaluation of the performance of the model was made by incorporating Kcb estimates, based on NDVI measurements and the developed regression functions, within the FAO-56 dual procedures and comparing the estimated ETc with field observations from two cotton plots collected during an experiment in central Arizona in 1998. Preliminary results indicate that the ETc based on the NDVI-Kcb model provided close estimates of actual ETc.

Remote- and Ground-Based Sensor Techniques to Map Soil Properties

Farm managers are becoming increasingly aware of the spatial variability in crop production with the growing availability of yield monitors. Often this variability can be related to differences in soil properties (e.g., texture, organic matter, salinity levels, and nutrient status) within the field. To develop management approaches to address this variability, high spatial resolution soil property maps are often needed. Some soil properties have been related directly to a soil spectral response, or inferred based on remotely sensed measurements of crop canopies, including soil texture, nitrogen level, organic matter content, and salinity status. While many studies have obtained promising results, several interfering factors can limit approaches solely based on spectral response, including tillage conditions and crop residue. A number of different ground-based sensors have been used to rapidly assess soil properties “on the go” (e.g., sensor mounted on a tractor and data mapped with coincident position information) and the data from these sensors compliment image-based data. On-the-go sensors have been developed to rapidly map soil organic matter content, electrical conductivity, nitrate content, and compaction. Model and statistical methods show promise to integrate these groundand image-based data sources to maximize the information from each source for soil property mapping.

COTTON IRRIGATION SCHEDULING USING REMOTELY SENSED AND FAO-56 BASAL CROP COEFFICIENTS

Multispectral vegetation indices calculated from canopy reflectance measurements have been used to simulate real-time basal crop coefficients (Kcb), which have been validated to improve evapotranspiration (ETc) estimation for several crops. In this article, an application of the approach was evaluated for cotton using remote sensing observations of the normalized difference vegetation index (NDVI) to estimate Kcb as a function of NDVI. The dual crop coefficient procedures of FAO Paper 56 (FAO-56) were used to calculate ETc and determine irrigation scheduling using Kcb estimates from remote sensing (NDVI treatment) as well as from time-based Kcb curves (FAO treatment), which were developed locally for standard crop conditions using FAO-56 procedures. Two cotton experiments, conducted in 2002 and 2003 in central Arizona, included sub-treatments of three levels of plant density and two levels of nitrogen management to impose a wide range of crop development and water use. The NDVI-Kcb relationships used for 2002, developed from previous data for a different cotton cultivar, row orientation, and soil type, substantially underestimated ETc, resulting in significantly less irrigation water applied and lower lint yields for NDVI compared to the FAO treatment. The 2002 data were used to recalibrate the NDVI-Kcb relationships, which were then used for the NDVI treatments in 2003. The FAO Kcb curve used in 2002 described ETc and irrigation scheduling reasonably well for sparse plots, but consistently underestimated water use and soil water depletion for the higher plant densities during the first half of the season. Consequently, an adjusted FAO Kcb curve, based on 2002 results, was employed for the FAO treatment in 2003. For the 2003 experiment, estimated cotton ETc for the NDVI treatment resulted in a mean absolute error of 9% compared to 10% for the FAO treatment, where the difference was not significant between treatments. However, the NDVI-Kcb relations used in 2003 greatly improved estimates for ETc compared to the previous year, where the mean absolute error for the NDVI treatment in 2002 was 22%. Predicted ETc using the FAO Kcb curve of 2003 for typical planting density and high nitrogen conditions resulted in a mean absolute error of 10% compared to 15% in 2002. Final lint yields for 2003 were not significantly different between the two Kcb methods. Although additional research is needed to validate remote sensing Kcb estimation for other conditions than those in these experiments, this study did not show significant advantages for the NDVI approach over a carefully derived single FAO Kcb application. However, the NDVI approach has the potential to further extend our present crop coefficient estimation capabilities when weather, plant density, or other factors cause cotton canopy development and water use patterns to depart from typical conditions.

GROUND-BASED REMOTE SENSING OF WATER AND NITROGEN STRESS

A ground–based remote sensing system (Agricultural Irrigation Imaging System, or AgIIS) was attached to a linear–move irrigation system. The system was used to develop images of a 1–ha field at 1 U 1 m resolution to address issues of spatial scale and to test the ability of a ground–based remote sensing system to separate water and nitrogen stress using the coefficient of variation (CV) for water and nitrogen stress indices. A 2 U 2 Latin square water and nitrogen experiment with four replicates was conducted on cotton for this purpose. Treatments included optimal and low nitrogen with optimal and low water. ANOVA was not an adequate method to assess the statistical variation between treatments due to the large number of data points. In general, the coefficient of variation of water and nitrogen stress indices increased with water and nitrogen stress. In fact, the coefficient of variation of stress indices was a more reliable measurement of water and nitrogen status than the mean value of the indices. Differences in coefficient of variation of stress indices between treatments were detectable at 3 m grid resolution and finer for water stress and at 7 m grid resolution and finer for nitrogen stress.

REMOTE SENSING OF COTTON NITROGEN STATUS USING THE CANOPY CHLOROPHYLL CONTENT INDEX (CCCI)

Various remote sensing indices have been used to infer crop nitrogen (N) status for field-scale nutrient management. However, such indices may indicate erroneous N status if there is a decrease in crop canopy density influenced by other factors, such as water stress. The Canopy Chlorophyll Content Index (CCCI) is a two-dimensional remote sensing index that has been proposed for inferring cotton N status. The CCCI uses reflectances in the near-infrared (NIR) and red spectral regions to account for seasonal changes in canopy density, while reflectances in the NIR and far-red regions are used to detect relative changes in canopy chlorophyll, a surrogate for N content. The primary objective of this study was to evaluate the CCCI and several other remote sensing indices for detecting the N status for cotton during the growing season. A secondary objective was to evaluate the ability of the indices to appropriately detect N in the presence of variable water status. Remote sensing data were collected during the 1998 (day of year [DOY] 114 to 310) and 1999 (DOY 106 to 316) cotton seasons in Arizona, in which treatments of optimal and low levels of N and water were imposed. In the 1998 season, water treatments were not imposed until late in the season (DOY 261), well after full cover. Following an early season N application in 1998 for the optimal (DOY 154) but not the low N treatment, the CCCI detected significant differences in crop N status between the N treatments starting on DOY 173, when canopy cover was about 30%. A common vegetation index, the ratio of NIR to red (RVI), also detected significant separation between N treatments, but RVI detection occurred 16 days after the CCCI response. After an equal amount of N was applied to both optimal and low N treatments on DOY 190 in 1998, the CCCI indicated comparable N status for the N treatments on DOY 198, a trend not detected by RVI. In the 1999 season, both N and water treatments were imposed early and frequently during the season. The N status was poorly described by both the CCCI and RVI under partial canopy conditions when water status differed among treatments. However, once full canopy was obtained in 1999, the CCCI provided reliable N status information regardless of water status. At full cotton cover, the CCCI was significantly correlated with measured parameters of N status, including petiole NO 3 -N (r = 0.74), SPAD chlorophyll (r = 0.65), and total leaf N contents (r = 0.86). For well-watered cotton, the CCCI shows promise as a useful indicator of cotton N status after the canopy reaches about 30% cover. However, further study is needed to develop the CCCI as a robust N detection tool independent of water stress.

AgIIS, Agricultural Irrigation Imaging System

Ground-based remote sensing can provide data at spatial resolutions, repeat frequencies, and turnaround times that are suitable for daily farm management at the field scale, whereas present satellite and airborne platforms do not meet these criteria. Remotely sensed data, when combined with other ancillary data, can provide spatially distributed maps of vegetation vigor, evapotranspiration, crop water stress, and nitrogen status. The objective of this paper was to describe the design, operation, and data processing of a ground-based remote sensing system called the Agricultural Irrigation Imaging System (AgIIS) that uses a self-propelled lateral move irrigation system as the transport platform. AgIIS consists of a cart that contains four nadir-looking sensors that measure reflected irradiance in the green (555 nm), red (670 nm), red-edge (720 nm), and near infrared (790 nm), all filtered to a 10-nm band pass at the band centers, and an infrared thermometer that measures directional radiometric surface temperature. The cart moves along a track, which is mounted to the overhead pipe of the lateral move system. Surface reflectance and temperature measurements were resampled to 1- m x 1-m raster grids, which were used to construct maps of vegetation, water, and nitrogen stress indices. These indices were correlated to field measurements of leaf area index, (r2 = 0.81 to 0.92), soil water deficit index (r2 = 0.76 to 0.86), and leaf petiole nitrogen content (r2 = 0.17 to 0.55), where experimental treatments consisted of two rates of irrigation and nitrogen applications for a cotton and broccoli crop.

Trap catches of the sweetpotato whitefly (Homoptera: Aleyrodidae) in the Imperial Valley, California, from 1996 to 2002

An outbreak of the sweetpotato whitefly, Bemisia tabaci (Gennadius), biotype B occurred in the Imperial Valley, California in 1991. The insects destroyed melon crops and seriously damaged other vegetables, ornamentals and row crops. As a result of the need for sampling technology, we developed a whitefly trap (named the CC trap) that could be left in the field for extended time periods. We used the traps to monitor populations of B. tabaci adults during year‐round samplings from 1996 to 2002 to study variations in the weekly trap catches of the insect. The greatest number of B. tabaci adults was recorded in 1996, followed by a continuing annual decrease in trap catches each year through 2002. The overall decline of B. tabaci is attributed in part to the adoption of an integrated pest management (IPM) program initiated in 1992 and reduced melon hectares from 1996 to 2002. Other factors may also have contributed to the population reductions. Seasonally, B. tabaci trap catches decreased during the late summer and fall concurrent with decreasing minimum temperatures that are suggested to be a significant factor affecting seasonal activity and reproduction.

Microwave Imaging of Cotton Bales

Modern moisture restoration systems are increasingly capable of adding water to cotton bales. However, research has identified large variations in internal moisture within bales that are not readily monitored by current systems. While microwave moisture sensing systems can measure average bale moisture, this can be deceptive where water is unevenly distributed. In some cases, localized internal moisture levels exceed 7.5%, the upper safe limit for cotton bale storage, as determined by the USDA, as above this level, bales degrade and lose value. A high proportion of stored bales containing excess moisture have been discovered throughout the US in increasing numbers over the past several seasons, making the detection and prevention of this occurrence a critical goal. Previous research by the authors resulted in the development of microwave moisture-sensing technology. The current study examines an extension to this technology to allow for detailed cotton bale moisture imaging. The new technique incorporates a narrow beam imaging antenna coupled to a tomographic imaging algorithm. The imaging technique was able to resolve small (< 1 cm) high-permittivity structures against a low permittivity background. Moreover, the system was able to identify structures of known permittivity with high accuracy (coefficient of determination (r2) > 0.99). In preliminary testing on a wet commercial UD bale, the technique was able to accurately image and resolve the location of the pre-placed internal wet layer.

Development and assessment of a smartphone application for irrigation scheduling in cotton

Easy-to-use and engaging smartphone application.Interactive ET-based soil water balance model.Uses meteorological data from weather station networks.Estimates root zone soil water deficits (RZSWD).Has mostly outperformed other irrigation scheduling tools. The goal of this work was to develop an easy-to-use and engaging irrigation scheduling tool for cotton which operates on a smartphone platform. The model which drives the Cotton SmartIrrigation App (Cotton App) is an interactive ET-based soil water balance model. The Cotton App uses meteorological data from weather station networks, soil parameters, crop phenology, crop coefficients, and irrigation applications to estimate root zone soil water deficits (RZSWD) in terms of percent as well as of inches of water. The Cotton App sends notifications to the user when the RZSWD exceeds 40%, when phenological changes occur, and when rain is recorded at the nearest weather station. It operates on both iOS and Android operating systems and was released during March 2014. The soil water balance model was calibrated and validated during 2012 and 2013 using data from replicated plot experiments and commercial fields. The Cotton App was evaluated in field trials for three years and performed well when compared to other irrigation scheduling tools. Its geographical footprint is currently limited to the states of Georgia and Florida, United States, because it is enabled to use meteorological data only from weather station networks in these states. A new version is currently under development which will use national gridded meteorological data sets and allow the Cotton App to be used in most cotton growing areas of the United States.