Texas Agricultural Experiment Station and Texas Agricultural Extension Service

Principal Investigator: Kevin F. Bronson^a

^aTexas Agric Exp Stn, RR 3, Box 219, Lubbock, TX 79403

Email: k-bronson@tamu.edu

Cooperators: C. L. Trostle^b, M. Schubert

^bTexas Agric. Extn Serv, Lubbock and College Station, TX

Primary Research Location: Denver City, TX

Project Title: Spatial characterization of soil properties: effect on peanut yield

Project Objectives:

1) To correlate soil chemical properties with peanut yields determined with the GPS-referenced yield monitor, and determine critical nutrient/chemical levels.

2) To correlate mid-season, multi-spectral reflectance (16 bands) with leaf nutrients.

Reporting Period: January 1 2001 – December 31 2001

A. Summary of Progress:

Objective 1. To correlate soil chemical properties with peanut yields determined with the GPS-referenced yield monitor, and determine critical nutrient/chemical levels.

Objective 2. To correlate mid-season, multi-spectral reflectance (16 bands) with leaf nutrients.

Introduction:

Second to water and seasonal heat units, soil chemical properties strongly affect peanut yield in the High Plains. Preliminary observations for example, indicate a negative correlation between soil Ca and peanut yield. Additionally, soil test calibration data for peanut is lacking in West Texas. Precision agriculture approaches, such as variable-rate fertilization, first require good understanding of the correlation between soil nutrients and peanut yield, preferably on a large, landscape scale.

Objectives:

1) To correlate soil chemical properties with peanut yields determined with the GPS-referenced yield monitor, and determine critical nutrient/chemical levels.

2) To correlate mid-season, multi-spectral reflectance (16 bands) with leaf nutrients.

Materials and Methods:

We conducted these activities on the 60 acre, peanut half of the western circle of the Western Peanut Growers Research Farm in Denver City in 2001. We sampled 54, GPS referenced 1-acre-grid points for soil samples (two per GPS point) at 0-6, 6-12, 12-24, and 24-36 inches. Soils were analyzed for NO₃-N at all depths. The surface 0-6 inches were analyzed at a commercial laboratory for pH, Olsen-P, K, Ca, Mg, Zn, Fe, Mn, and EC. Particle size distribution was also done on the top layer. Mid-season we recorded multi-spectral reflectance at a 0.5 m height at the 60 GPS points. Leaf samples taken at these spots were analyzed for N and other macro- and micro-nutrients. Circular areas of yield map data points, 20 m in diameter were extracted around each GPS point, using ArcView. Statistical analysis included correlation of soil nutrients and mid-season leaf nutrients with yield. Additionally we regressed plant nutrients and final yield on principle components of the soils properties and on the multi-spectral reflectance data using paritial least squares regression. The soils work on the western circle dove-tailed with Calvin Trostle’s rhizobia project.

Results and discussion:

Peanut yields in the study area ranged from 2900 to 5300 lb/ac. Fifty-one percent of the variation in peanut yield at the 54, GPS-referenced peanut yield zones were explained by five common factors, which were in turn based on 14 soil properties. The soil properties which had the strongest effect on peanut yield were ones related to pH and micronutrients, i.e. pH, S, Fe, and Mn. Adding mid-season plant macro- and micronutrient concentrations to the analysis only improved the coefficient of determination (R²) from 0.51 to 0.55.

Multispectral (16-band) reflectance of the mid-season peanut canopy was most strongly correlated with leaf P and leaf K. Partial least squares regression of leaf P and leaf K on five common factors extracted from the sixteen wavebands yielded R²s of 0.42, and 0.58, respectively. Soil P ranged from the low to medium soil test sufficiency indices, so effect of low soil P areas on leaf reflectance and leaf P were not surprising. Soil K on the other hand, tested high or very high in all points. Relations between reflectance and leaf K, therefore are not clear.

In summary, this research underscores a main hypothesis of site-specific management. Spatial variation in soil properties and nutrient levels results in corresponding spatial variation in peanut yield. An example of this is that high pH areas of the 60-ac study area, generally had the lowest peanut yields. This indicates that it may be profitable to vary inputs such as micronutrients on a variable-rate/area of the field basis.