Skip to main content

Table 1 Overview of digital technology in soil quality assessment

From: Digital technology dilemma: on unlocking the soil quality index conundrum

Techniques

Key findings and limitations

Source

Disjunctive kriging (DK), ArcGIS, classification, kappa statistic

Soil salinity estimated

However, soil salinity variables all skewed distribution and poorly correlated with terrain indices, though have strong correlations among each other

Bilgili (2013)

Indian Remote Sensing (IRS)-1B LISS-II digital data, field data and topographical maps, salinity indices (band combinations), supervised maximum likelihood classification

Over 65% of salt-affected soils found in shallow water table areas over 3 years

Abbas et al. (2013)

Soil Properties mapped using visible and near infrared (vis–NIR) reflectance spectroscopy technology (ASD FieldSpec Pro FR spectrometer). Spectral indices, Partial Least Squares Regression and Kriging

Only Soil Nitrogen mapped. However, the correlation between soil Nitrogen and other soil attributes e.g., Phosphates, Available water content and even SOC ignored

Liu et al. (2013)

Diffuse Reflectance Spectroscopy. Digital camera used to indirectly measure soil organic carbon (OC) and iron (Fe) contents using soil colour as the proxy. Predictions using univariate and full factorial regressions (FFR). visible-near infrared (vis–NIR: 400–1100 nm) spectra and partial least squares regression (PLSR)

Digital camera practically useful as a fast, accurate and non-destructive predictor of soil OC and Fe contents

SOC predicted better than Fe

However, interpolation of metrics unresolved

Cellular networks not applied to transmit information in real time

Rossel et al. (2008)

3 Soil quality indices (SQIs) developed by quantifying several soil properties to discriminate effects of slope gradient and land use change on soil quality

Soil quality indices maps were developed using digital soil mapping methods.

Steep slopes and geographic locations with land use conversion from grassland or forests to agriculture had lower soil quality

This hypothetically is attributable to increased soil erosion, lower C input in croplands, or increased soil temperature and aeration enhancing mineralization

Nabiollahi et al. (2018)

Narrowband radio channel model application in wireless sensor networks for smart agriculture

3 radio frequencies used to distinguish soil, short and tall grass fields

Technique accurately distinguished soils from vegetation. However soil quality not evaluated. The technique together with information relayed are complex to be relayed in current format through cellular networks to stakeholders

Klaina et al. (2018)

Soil pH mapped using two machine learning techniques, namely Random Forest and XGBoos

Generated map provides pertinent information useful for:

(i) Assessing impacts of changes in land use and climate on the soil’s pH,

(ii) Guiding users on remediation and prevention of soil acidification, salinization and pollution by heavy metals, e.g., cadmium and mercury

Chen et al. (2019)

Principal component analysis used to screen out significant variables determining soil quality.

Linear and non-linear scoring systems used to compute SQI

Random forest technique used to generate soil quality map

(i) Results show that soils under natural forest were of a higher quality than soils under dry farming land use.

(ii) The Linear scoring system had higher coefficient of determination (R2) with SQIs than the nonlinear scoring system

Zeraatpisheh et al. (2020)