Natural Science Forum / Biology / Biology / February 2008

Computers and Electronics in Agriculture
Volume 60, Issue 2, March 2008, Pages 212-225

B. Sethuramasamyrajaa, V.I. Adamchukb, , , A. Dobermannc, D.B. Marxd,
D.D. Jonesb and G.E. Meyerb
aDepartment of Industrial Technology, California State University,
Fresno, CA 93740, USA
bBiological Systems Engineering Department, University of Nebraska-
Lincoln, Lincoln, NE 68583, USA
cAgronomy and Horticulture Department, University of Nebraska-Lincoln,
Lincoln, NE 68583, USA
dDepartment of Statistics, University of Nebraska-Lincoln, Lincoln, NE
68583, USA

abstract:

Knowledge of spatial variability of soil attributes within an
agricultural field is critical for successful site-specific crop
management. Soil sensing techniques to assess this variability on-the-
go are being developed as an alternative to tedious manual soil
sampling and laboratory testing. The goal of this study was to
evaluate an Agitated Soil Measurement (ASM) method for integrated on-
the-go mapping of soil pH, soluble potassium and residual nitrate
contents using ion-selective electrodes. To implement ASM, an
Integrated Agitation Chamber Module (IACM) was developed and attached
to a commercial soil pH mapping implement. Precision of the tested
electrodes was assessed through the root mean squared error (RMSE) and
ranged from 0.10 for pK to 0.22 for pNO3 (units represent the negative
base 10 logarithm of the molar concentration of specified ions). The
accuracy of the electrodes was assessed by comparing test results
against reference measurements conducted in a commercial soil
laboratory using the linear regression method. Average accuracy error
ranged from 0.11 for pK to 0.23 for pNO3. In a field simulation test,
neither precision nor accuracy errors obtained with ASM were lower
than for a previously investigated Direct Soil Measurement (DSM)
method, which produced precision errors ranging from 0.11 for pH to
0.22 for pNO3 and accuracy errors ranging from 0.12 for pNO3 to 0.20
for pH. The coefficients of determination (r2) of linear regressions
between individual field simulation measurements and corresponding
average reference measurements were 0.85-0.89 (pH), 0.50-0.54 (pK),
and 0.14-0.32 (pNO3). However, laboratory evaluation of the ASM method
revealed substantially lower measurement errors and increased r2
values when compared to the field simulation, indicating that the
proposed ASM method retains the potential for improving on-the-go
field mapping. Except for reduced electrode abuse and the ability to
use less expensive half-cell ion-selective electrodes, physical
implementation of ASM through the IACM did not bring substantial
improvement over conventionally available DSM. This could be
attributed to the design of the IACM and use of half-cell electrodes.
Further research is necessary to improve the design of the solution-
based measuring equipment and to develop an algorithm integrating on-
the-go measurements with other sources of spatial data for an improved
decision-making process.