|M.Sc Student||Borenstein Asaf|
|Subject||Direct Determination of Nitrate in Agricultural Soils Using|
|Department||Department of Agricultural Engineering||Supervisors||Professor Raphael Linker|
|Professor Emeritus Abraham Shaviv|
|Professor Emeritus Itzhak Shmulevich|
Excessive fertilization causes significant damage to the environment, both in water sources and in soils. The ‘precision agriculture’ concept, whose goal is to adjust fertilization based on crop needs and soil properties, has the potential of reducing the amount of fertilizer used without diminishing yield. However, the lack of adequate sensors, and in particular the lack of real-time sensors for nitrate concentration in soil, is preventing large-scale implementation of this approach.
The possibility of determining nitrate concentration in soil pastes using mid-infrared attenuated total reflectance (ATR) spectroscopy has been demonstrated by several laboratory studies. If these results, which were obtained under laboratory conditions, could be reproduced under field conditions, such a sensor would be a significant step toward direct determination of nitrate content in the field. A major limitation of the previous studies is that the soil pastes were obtained by adding a pre-determined amount of water to oven-dried soil, so that standardization of the paste moisture content was achieved. While such a procedure is common for laboratory analysis, it is unsuitable for direct field determination.
This study was conducted in order to determine whether mid infrared ATR spectroscopy could be used to determine nitrate concentration in agricultural soil samples that have not been standardized.
The proposed approach is based on sequential determination of soil moisture and nitrate content using ATR spectroscopy measurements of two sub-samples. The first measurement involves a paste that is obtained by adding a pre-determined amount of highly concentrated KNO3 solution (5000 [mg N-NO3/l]) to 15 grams of sampled soil. The water initially present in the sample dilutes the added nitrate. The FTIR/ATR spectrum of the paste is used to estimate its nitrate concentration, which in turn is used to determine the ”nitrate dilution factor” and hence the water content of the sample. A second paste is prepared by adding the estimated amount of water required to obtain the desired, standardized, moisture. Analysis of the spectrum of this second sub-sample gives the soil nitrate concentration.
For soil moisture determination errors vary from 1.6 to 2.9 % depending on soil type and experimental procedure. Such errors are similar to the determination errors obtained with the gravimetric technique used as reference in this study. The nitrate prediction models have determination errors ranging from 8.8 to 17.8 [mg N-NO3/kg dry soil] depending on soil type and experimental procedure. These results are similar to those previous laboratory