Agroscope

Measuring Soil Fertility Directly in the Field

Soil samples can be measured directly in the field by means of spectroscopy. Agroscope researchers have tested mobile devices and shown how to make the best use of them.

Currently, soil samples are analysed in the laboratory to determine the soil fertility, which involves considerable time and expense. A spectroscopic method, however, holds the promise of measuring some fertility indicators directly in the field. This is faster and less costly, since when used in combination with a predictive model, various soil characteristics can be derived from the spectra. This means that more samples can be analysed, and both the spatial distribution of the soil fertility, as well as changes in it over time, can be better investigated.

Portable devices make field measurements attractive

The method is based on spectroscopy in the visible and near-infrared regions (vis-NIR), which has already been researched for a number of years for the chemical and physical analysis of soil samples. More recently, portable and affordable equipment has also been developed, allowing the method to be used directly in the field.

Agroscope researchers have tested two portable spectrometers: a more compact, less costly device (NeoSpectra Scanner, Si-Ware Systems) and an expensive, professional, and heavier spectrometer for research purposes (PSR-3500, Spectral Evolution). Various soils and fields were investigated using these two instruments. The following samples were measured:

  • The side cut surface of a soil sample 20 cm deep, taken with an Edelman auger.
  • The untreated soil surface
  • The cleaned and smoothed soil surface

A total of 134 measuring points at three locations were evaluated and compared. They were analysed both with the two spectrometers and in the laboratory, in order to calibrate a model that could predict the laboratory data from the spectral data.

Unlike the laboratory method, for which the soil samples were dried before analysis, field spectroscopy measurements use fresh samples which contain different amounts of moisture. Soil moisture has a great influence on the measured spectra, so the influence of the soil moisture must be removed from the data by calculations with the aid of models.

Soil fertility successfully measured

Both spectrometers successfully predicted soil fertility parameters such as clay, sand, pH value, organic carbon, cation exchange capacity, total nitrogen content, and available magnesium. On the other hand, they did not reliably measure total and available calcium, potassium, phosphorous and total magnesium contents.

Inexpensive spectrometer gives usable results

Although the research spectrometer gave better results for most of the parameters, calibration with the simple spectrometer nonetheless resulted in satisfactory predictions. The only area where the cheaper spectrometer was clearly inferior to the professional device was the pH value.

The best measuring position was along the cut sides of the soil samples collected with an Edelman auger. The authors also recommend repeating the measurements at least five times.

Conclusions

  • The use of portable spectrometers in the visible and near infrared range (vis-NIR) is suitable for measuring various soil fertility indicators such as clay, organic carbon, total nitrogen content, pH value and cation exchange capacity.
  • The more expensive and research grade spectrometer gives more reliable values than the cheaper one, but the latter nonetheless still gives meaningful results – apart from the pH value.
  • The best scanning position in the field is along the cut surface of a soil sample, such as obtained using an Edelman auger.
  • Vis-NIR spectroscopy in the field, supported by good model calibrations, allows the spatial and temporal distribution of soil fertility indicators to be investigated more easily and economically.
To the archive