Browsing by Author "Kebede, Fassil"

Now showing 1 - 5 of 5
  • No Thumbnail Available
    ItemEmbargo
    Dry spectral diagnostic tools and methods for precise fertilizer application
    (Cranfield University, 2024-03) Asrat, Tadesse Gashaw; Sakrabani, Ruben; Corstanje, Ronald; Haefele, Stephan M.; Kebede, Fassil
    This study explores the potential of soil spectroscopy to enhance fertilizer decision-making by providing cost-effective, portable instruments for farm-level soil property prediction. The study focused on assessing the performance of various spectrometers, including low-cost near-infrared (NIR) devices, compared to mid-infrared (MIR) bench-top instruments, using a case study on maize productivity in East Africa and soil data from Morocco’s semi-arid rainfed wheat- growing regions. The overall aim was to determine whether these spectroscopic methods could generate reliable predictions of key soil properties for nitrogen, phosphorus and potassium fertilizer recommendations. The results indicate that NIR spectroscopy, despite being the most affordable and portable option, demonstrated sufficient accuracy for predicting key soil properties such as soil pH, organic carbon, and exchangeable potassium, with concordance correlation coefficients (CCCs) ranging from 0.77 to 0.96. However, the prediction of phosphorus (Olsen P) showed considerable uncertainty, particularly for values above 15 mg P kg⁻¹, where deviations from measured values increased. Comparatively, the MIR spectrometer showed better prediction accuracy for phosphorus, though its higher cost and complexity limit its applicability in resource-limited settings. The NIR spectrometer, with a prediction accuracy suitable for nitrogen fertilization (deviation between -8 to 8 kg N ha⁻¹), emerged as a promising tool for cost-effective and rapid nutrient recommendations in developing countries. Furthermore, this research demonstrated that integrating spectroscopic data into crop models like QUEFTS for nutrient management enhances decision-making by considering both soil supply and crop response to nutrients. The findings underscore the necessity of developing region-specific calibration models to improve prediction reliability, with spatial autocorrelation analysis of soil spectra suggesting that proper calibration sample selection can improve prediction performance, especially for phosphorus and other key properties. Ultimately, this thesis contributes to the ongoing development of soil spectral libraries and highlights the potential of low-cost, field-friendly spectrometers to improve nutrient management and crop productivity in regions with limited soil data.
  • Thumbnail Image
    ItemOpen Access
    A Moroccan soil spectral library use framework for improving soil property prediction: evaluating a geostatistical approach
    (Elsevier, 2024-12-01) Asrat, Tadesse Gashaw; Breure, Timo; Sakrabani, Ruben; Corstanje, Ron; Hassall, Kirsty L.; Hamma, Abdellah; Kebede, Fassil; Haefele, Stephan M.
    A soil spectrum generated by any spectrometer requires a calibration model to estimate soil properties from it. To achieve best results, the assumption is that locally calibrated models offer more accurate predictions. However, achieving this higher accuracy comes with associated costs, complexity, and resource requirements, thus limiting widespread adoption. Furthermore, there is a lack of comprehensive frameworks for developing and utilizing soil spectral libraries (SSLs) to make predictions for specific samples. While calibration samples are necessary, there is the need to optimize SSL development through strategically determining the quantity, location, and timing of these samples based on the quality of the information in the library. This research aimed to develop a spatially optimized SSL and propose a use-framework tailored for predicting soil properties for a specific farmland context. Consequently, the Moroccan SSL (MSSL) was established utilizing a stratified spatially balanced sampling design, using six environmental covariates and FAO soil units. Subsequently, various criteria for calibration sample selection were explored, including a spatial autocorrelation of spectra principal component (PC) scores (spatial calibration sample selection), spectra similarity memory-based learner (MBL), and selection based on environmental covariate clustering. Twelve soil properties were used to evaluate these calibration sample selections to predict soil properties using the near infrared (NIR) and mid infrared (MIR) ranges. Among the methods assessed, we observed distinct precision improvements resulting from spatial sample selection and MBL compared to the use of the entire MSSL. Notably, the Lin's Concordance Correlation Coefficient (CCC) values using the spatial calibration sample selection was improved for Olsen extractable phosphorus (OlsenP) by 41.3% and Mehlich III extractable phosphorus (P_M3) by 8.5% for the MIR spectra and for CEC by 25.6%, pH by 13.0% and total nitrogen (Tot_N) by 10.6% for the NIR spectra in reference to use of the entire MSSL. Utilizing the spatial autocorrelation of the spectra PC scores proved beneficial in identifying appropriate calibration samples for a new sample location, thereby enhancing prediction performance comparable to, or surpassing that of the use of the entire MSSL. This study signifies notable advancement in crafting targeted models tailored for specific samples within a vast and diverse SSL.
  • Thumbnail Image
    ItemOpen Access
    Optimizing setup of scan number in FTIR spectroscopy using the moment distance index and PLS regression: application to soil spectroscopy
    (Nature Publishing Group, 2021-06-25) Barra, Issam; Khiari, Lotfi; Haefele, Stephan M.; Sakrabani, Ruben; Kebede, Fassil
    Vibrational spectroscopy such as Fourier-transform infrared (FTIR), has been used successfully for soil diagnosis owing to its low cost, minimal sample preparation, non-destructive nature, and reliable results. This study aimed at optimizing one of the essential settings during the acquisition of FTIR spectra (viz. Scans number) using the standardized moment distance index (SMDI) as a metric that could trap the fine points of the curve and extract optimal spectral fingerprints of the sample. Furthermore, it can be used successfully to assess the spectra resemblance. The study revealed that beyond 50 scans the similarity of the acquisitions has been remarkably improved. Subsequently, the effect of the number of scans on the predictive ability of partial least squares regression models for the estimation of five selected soil properties (i.e., soil pH in water, soil organic carbon, total nitrogen, cation exchange capacity and Olsen phosphorus) was assessed, and the results showed a general tendency in improving the correlation coefficient (R2) as the number of scans increased from 10 to 80. In contrast, the cross-validation error RMSECV decreased with increasing scan number, reflecting an improvement of the predictive quality of the calibration models with an increasing number of scans.
  • Thumbnail Image
    ItemOpen Access
    Soil spectroscopy with the use of chemometrics, machine learning and pre-processing techniques in soil diagnosis: recent advances - a review
    (Elsevier, 2020-12-30) Barra, Issam; Haefele, Stephan M.; Sakrabani, Ruben; Kebede, Fassil
    Over the past two decades soil spectroscopy, particularly, in the infrared range, is becoming a powerful technique to simplify analysis relative to the traditional chemical methods. It is known as a rapid, cost-effective, quantitative and eco-friendly technique, which can provide hyperspectral data with narrow and numerous wavebands, both in the laboratory and in the field. In this context, the present article reviews the recent developments in mid and near infrared techniques coupled with chemometrics and machine learning tools in addition to the preprocessing transformations and variable selection strategies to diagnose soil physical and chemical properties. Both spectral techniques demonstrated a good ability to provide accurate predictions of specific properties. Moreover, the MIR spectroscopy outperformed NIR for the estimation of most indicators used for fertilizers recommendation. Herein, a detailed overview on the opportunities and challenges that soil spectroscopy offers as efficient diagnostic tool in soil science was provided.
  • Thumbnail Image
    ItemOpen Access
    Spectral soil analysis for fertilizer recommendations by coupling with QUEFTS for maize in East Africa: A sensitivity analysis
    (Elsevier, 2023-02-24) Asrat, Tadesse Gashaw; Sakrabani, Ruben; Corstanje, Ronald; Breure, Timo; Hassall, Kirsty L.; Kebede, Fassil; Haefele, Stephan M.
    Laboratory analysis of soil properties is prohibitively expensive and difficult to scale across the soils in sub-Saharan Africa. This results in a lack of soil-specific fertilizer recommendations, where recommendation can only be provided at a regional scale. This study aims to assess the feasibility of using spectral soil analysis to provide soil-specific fertilizer recommendations. Using a range of spectrometers [NeoSpectra Saucer (NIR), FieldSpec 4 (vis-NIR) with contact probe or mug light interface, FTIR Bruker Tensor 27 (MIR)], 346 archived soil samples (0–20 cm) with known soil chemical properties collected from Ethiopia, Kenya and Tanzania were scanned. Partial least square regression (PLSR) was used to develop prediction models for selected soil properties including pH, soil organic carbon (SOC), total nitrogen, Olsen P, and exchangeable K. These predicted properties, and associated uncertainty, were used to derive fertilizer recommendations for maize using the Quantitative Evaluation of the Fertility of Tropical Soils (QUEFTS) model parameters for sub-Saharan Africa. Most soil properties (pH, SOC, total nitrogen, and exchangeable K) were well predicted (Concordance Correlation Coefficient values between 0.88 and 0.96 and Ratio of Performance to Interquartile values between 1.4 and 5.9) by all the spectrometers but there were performance variations between soil properties and spectrometers. Use of the predicted soil data for the development of fertilizer recommendations gave promising results when compared to the recommendations obtained with the conventional soil analysis. For example, the least performing NeoSpectra Saucer over/under-estimated up to 8 and 24 kg ha-1N and P, respectively, though there was insignificant variation in estimation of P fertilizer among spectrometers. We conclude that spectral technology can be used to determine major soil properties with satisfactory precision, sufficient for specific fertilizer decision making in East Africa, possibly even with portable equipment in the field.