Location: Agroecosystems Management ResearchTitle: Estimating soil solution nitrate concentration from dielectric spectra using PLS analysis) Author
Submitted to: Soil Science Society of America Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/20/2012
Publication Date: 9/13/2012
Citation: Chigladze, G., Birrell, S., Kaleita, A., Logsdon, S.D. 2012. Estimating soil solution nitrate concentration from dielectric spectra using PLS analysis. Soil Science Society of America Journal. 76:1536-1547. Interpretive Summary: Excess soil nitrate-nitrogen can be lost to ground water and drain out tiles to surface water, but monitoring soil nitrate levels often involves time-consuming laboratory analysis. This study examined a way to detect nitrate concentration in soil, using dielectric properties. The best way to accurately determine soil nitrate levels was to eliminate the samples with very low water content or very high chloride content. This study is a step toward developing a field monitoring tool. At this preliminary stage, the study is mainly of interest to scientists. As the tool is developed, the technique will be of interest to farmers and their advisors.
Technical Abstract: Fast and reliable methods for in situ monitoring of soil nitrate-nitrogen concentration are vital for reducing nitrate-nitrogen losses to ground and surface waters from agricultural systems. While several studies have been done to indirectly estimate nitrate-nitrogen concentration from time domain spectra, no research has been conducted using a frequency domain technique. Hence, the goal of this laboratory study was to estimate the change in nitrate-nitrogen concentration from frequency-response data obtained in a frequency range of 5 Hz to 13 MHz. Dielectric spectra of soil samples wetted to five different volumetric water content (VWC) with 24 solutions containing different concentration of potassium nitrate and potassium chloride were analyzed using partial least squares (PLS) regression method. The global models that incorporated the whole data estimated the VWC with R2 value of 0.99, but could not estimate the change in applied nitrate-nitrogen concentration with sufficient accuracy. In general, the models based on the imaginary part of the permittivity performed better than those based on the real part or both the real and imaginary parts of permittivity, particularly in estimating the nitrate-nitrogen concentration. Capabilities of the PLS models to estimate nitrate-nitrogen concentration were improved with elimination of the data obtained at low VWC level and high Cl content. As a result, the root means square error (RMSE) for nitrate-nitrogen was reduced from 57 mg / L to 28 mg / L. The best results were obtained when the PLS models were constructed at fixed VWC levels using the data without high Cl concentration. The performances of these models were improving with increasing VWC level, reaching the lowest RMSE of 18 mg / L at VWC of 30 m / m.