ALBERT

Description: Characterizing the spatial variability of nutrients facilitates precision soil sampling. Questions exist regarding the best technique for directed soil sampling based on a priori knowledge of soil and crop patterns. The objective of this study was to evaluate zone delineation techniques for Alabama grain fields to determine which method best minimized the soil test variability. Site one (25.8 ha) and site three (20.0 ha) were located in the Tennessee Valley region, and site two (24.2 ha) was located in the Coastal Plain region of Alabama. Tennessee Valley soils ranged from well drained Rhodic and Typic Paleudults to somewhat poorly drained Aquic Paleudults and Fluventic Dystrudepts. Coastal Plain s o i l s ranged from coarse-loamy Rhodic Kandiudults to loamy Arenic Kandiudults. Soils were sampled by grid soil sampling methods (grid sizes of 0.40 ha and 1 ha) consisting of: 1) twenty composited cores collected randomly throughout each grid (grid-cell sampling) and, 2) six composited cores collected randomly from a -3x3 m area at the center of each grid (grid-point sampling). Zones were established from 1) an Order 1 Soil Survey, 2) corn (Zea mays L.) yield maps, and 3) airborne remote sensing images. All soil properties were moderately to strongly spatially dependent as per semivariogram analyses. Differences in grid-point and grid-cell soil test values suggested grid-point sampling does not accurately represent grid values. Zones created by soil survey, yield data, and remote sensing images displayed lower coefficient of variations (8CV) for soil test values than overall field values, suggesting these techniques group soil test variability. However, few differences were observed between the three zone delineation techniques. Results suggest directed sampling using zone delineation techniques outlined in this paper would result in more efficient soil sampling for these Alabama grain fields.

Description: Crop residues influence near surface soil organic carbon content (SOC), impact our ability to remotely assess soil properties, and play a role in global carbon budgets. Methods that measure crop residues are laborious, and largely inappropriate for regional estimates. The objective of this study was to evaluate remote sensing (RS) data for rapid quantification of residue cover. In March 2000 and April 2001, residue plots (15 m x 15 m) were established in the Coastal Plain and Appalachian Plateau physiographic regions of Alabama. Treatments consisted of five wheat (Triticum aestivum L.) straw cover rates (0, 10, 20, 50, and 80%) replicated 3 times. Soil water content and residue decomposition were monitored. Spectral measurements were acquired via spectroradiometer (350 - 1050 nm), Airborne Terrestrial Applications Sensor (ATLAS) (400 - 12,500 nm), airborne color photography (400 - 600 nm), and IKONOS satellite (450 - 900 nm). Spectroradiometer data were acquired monthly, aircraft images yearly, and satellite per availability. Results showed all platforms successfully estimated residue cover variability using red, near infrared (NIR) and thermal infrared (TIR) regions of the spectrum. Airborne ATLAS imagery was best explaining as much as 98% of the variability in wheat straw cover. Spectroradiometer, color infrared photography, and IKONOS imagery accounted for 84, 56, and 24% of the variability, respectively.

Description: Transformations and losses of nitrogen (N) throughout the growing season can be costly. Methods in place to improve N management and facilitate split N applications during the growing season can be time consuming and logistically difficult. Remote sensing (RS) may be a method to rapidly assess temporal changes in crop N status and promote more efficient N management. This study was designed to evaluate the ability of three different RS platforms to predict N variability in corn (Zea mays L.) leaves during vegetative and early reproductive growth stages. Plots (15 x 15m) were established in the Coastal Plain (CP) and Appalachian Plateau (AP) physiographic regions each spring from 2000 to 2002 in a completely randomized design. Treatments consisted of four N rates (0, 56, 112, and 168 kg N/ha) applied as ammonium nitrate (NH4N03) replicated four time. Spectral measurements were acquired via spectroradiometer (lambda = 350 - 1050 nm), Airborne Terrestrial Applications Sensor (ATLAS) (lambda = 400 - 12,500 nm), and the IKONOS satellite (lambda = 450 - 900 nm). Spectroradiometer data were collected on a biweekly basis from V4 through R1. Due to the nature of - satellite and aircraft acquisitions, these data were acquired per availability. Chlorophyll meter (SPAD) and tissue N were collected as ancillary data along with each RS acquisition. Results showed vegetation indices derived from hand-held spectroradiometer measurements as early as V6-V8 were linearly related to yield and tissue N content. ATLAS data was correlated with tissue N at the AP site during the V6 stage (r2 = 0.66), but no significant relationships were observed at the CP site. No significant relationships were observed between plant N and IKONOS imagery. Using a combination of the greenness vegetation index (GNDVI) and the normalized difference vegetation index (NDVI), RS data acquired via ATLAS and the spectroradiometer could be used to evaluate tissue N variability and estimate corn yield variability under ideal growing conditions.

Description: Evaluation of near-surface soil properties via remote sensing (RS) could facilitate soil survey mapping, erosion prediction, fertilization regimes, and allocation of agrochemicals. The objective of this study was to evaluate the relationship between soil spectral signature and near surface soil properties in conventionally managed row crop systems. High resolution RS data were acquired over bare fields in the Coastal Plain, Appalachian Plateau, and Ridge and Valley provinces of Alabama using the Airborne Terrestrial Applications Sensor (ATLAS) multispectral scanner. Soils ranged from sandy Kandiudults to fine textured Rhodudults. Surface soil samples (0-1 cm) were collected from 163 sampling points for soil water content, soil organic carbon (SOC), particle size distribution (PSD), and citrate dithionite extractable iron (Fed) content. Surface roughness, soil water content, and crusting were also measured at sampling. Results showed RS data acquired from lands with less than 4 % surface soil water content best approximated near-surface soil properties at the Coastal Plain site where loamy sand textured surfaces were predominant. Utilizing a combination of band ratios in stepwise regression, Fed (r2 = 0.61), SOC (r2 = 0.36), sand (r2 = 0.52), and clay (r2 = 0.76) were related to RS data at the Coastal Plain site. In contrast, the more clayey Ridge and Valley soils had r-squares of 0.50, 0.36, 0.17, and 0.57. for Fed, SOC, sand and clay, respectively. Use of estimated eEmissivity did not generally improve estimates of near-surface soil attributes.