Predicting Scottish topsoil organic matter content from colour and environmental factors

Abstract

Summary Assessment of soil organic matter content using laboratory analysis can be costly and time consuming, so limiting how often land managers assess this important property. This work demonstrates an ability to estimate topsoil organic matter content from field observations alone and provides a method by which rapid and cost‐effective assessments of soil organic matter status may be made. Models using environmental factors from the National Soil Inventory of Scotland (NSIS) dataset as inputs to a neural network model were used to predict loss on ignition (LOI). Two models, one for all soils and one for soils with small organic matter contents (LOI < 20%), were developed. It was found that the model developed for all soils produced reasonable predictive results across the entire LOI range (R2 = 0.877), although it was not as effective at predicting small LOI values (R2 = 0.354) as the small organic matter content model (R2 = 0.674). Both models were tested with imagery and data from samples outwith the NSIS dataset to validate the approach. Predictive results were less accurate than when using NSIS data. A discussion of possible improvements to make the model useful for field observations of soils is given.

Novel procedure for testing of soil field test kits involving paper strips

“…Recently, there has been renewed interest in creation of soil test kits optimized for agronomical field use as a result of increasing access to technology such as portable sensors (Piikki et al., 2016). Key to this is the rising ubiquity of smartphones, which are being increasingly used in environmental management applications (Aitkenhead, Donnelly, Coull, & Hastings, 2014) and soil science (Aitkenhead et al., 2015; Delgado, Kowalski, & Tebbe, 2013; Stiglitz et al., 2017).…”

Performance Evaluation of Proximal Sensors for Soil Assessment in Smallholder Farms in Embu County, Kenya

“…Soil organic carbon and soil color has been demonstrated to be well correlated in several studies [11,12,30] and the optical sensor, used in this study, was designed specifically for mapping of TC content [17]. In this study, the best prediction model for TC was based on R, as measured with the optical sensor, and the contents of Sr and Zr, measured with the PXRF (i.e., no color variables).…”

“…Measurement in situ has potential to be cost- and labor-effective compared to laboratory analyses but it is a challenge to make accurate predictions under the non-standardized soil conditions prevailing on the farm (varying soil moisture and soil structure). A multitude of research studies has dealt with the issue of calibrating the often indirect field-based measurements against lab analyses to produce accurate values of a variety of soil properties [6]: electromagnetic conductivity (ECa) has been used widely to map spatial variation patterns of agricultural fields but also for prediction of specific soil properties, for example, cation exchange capacity (CEC), soil moisture content, soil organic carbon (SOC) content, and the fractions of clay, silt, and sand [7]; element concentrations measured with a portable X-ray fluorescence sensor (PXRF) in situ have been demonstrated to correlate with soil texture, SOC, and CEC [8,9,10]; Jung et al [11] showed that a multispectral camera can be used to predict SOC; and, recently, even the built-in cameras in mobile phones have been used as proximal soil sensors—Aitkenhead et al [12] predicted SOC from the registered values of soil color (red, green, and blue bands) registered with a mobile phone camera.…”

Estimation of kiwifruit yield by leaf nutrients concentration and artificial neural network

“…Numerous studies have been carried out to estimate soil variables through ANNs (Zhou et al ., 2008; Bocco et al ., 2010; Gago et al ., 2010; Parvizi et al ., 2010; Peng et al ., 2010; Ayoubi et al ., 2011; Mokhtari Karchegani et al ., 2011; Besalatpour et al ., 2013; Dai et al ., 2014; Aitkenhead et al ., 2015). Also, some studies have been conducted to predict crop yield by remote sensing, stochastic, ANN and simulation models (Bannayan and Crout, 1999; O'Neal et al ., 2002; Bartoszek, 2014; Farjam et al ., 2014; Domínguez et al ., 2015; Emamgholizadeh et al ., 2015; Dias and Sentelhas, 2017), based on weather, soil and growth characteristics as input data.…”