Search

Noble metal nanoparticles such as gold and silver are of interest because of the unique electro-optical properties (e.g., localized surface plasmon resonance [LSPR]) that originate from the collective behavior of their surface electrons. These nanoparticles are commonly developed and used for biomedical and industrial application. A recent report has predicted that the global market for gold nanoparticles will be over 12.7 tons by year 2020. However, these surface-functionalized nanoparticles can be potential environmental persistent contaminants post-use due to their high colloidal stability in the aquatic systems. Despite, the environmental risks associated with these nanoparticles, just a few studies have investigated the effect of nanofeature factors such as size and shape on the overall fate/transport and organismal uptake of these nanomaterials in the aquatic matrices. This study presents a comprehensive approach to evaluate the colloidal stability, fate/transport, and organismal uptake of these nanoparticles while factoring in the size and shape related properties. We demonstrate the importance and effect of anisotropicity of a gold nanoparticle on the colloidal behavior and interaction with ecologically susceptible aquatic biota. We also show how readily available characterization techniques can be utilized to monitor and assess the fate/transport of this class of nanoparticles. We further describe and investigate the relationship between the aspect ratio (AR) of these elongated gold nanoparticles with clearance mechanisms and rates from the aquatic suspension columns including aggregation, deposition, and biopurification. We illustrate how a fresh water filter-feeder bivalve, Corbicula fluminea, can be used as a model organism to study the size and shape-selective biofiltration and nanotoxicity of elongated gold nanoparticles. The results suggest that biofiltration by C. fluminea increases with an increase in the size and AR of gold nanoparticle. We develop a simple nanotoxicity assay to investigate the short-term exposure nanotoxicity of gold nanoparticles to C. fluminea. The toxicity results indicate that for the tested concentration and exposure period that gold nanoparticles were not acutely toxic (i.e., not lethal). However, gold nanoparticles significantly inhibited the activities of some antioxidant enzymes in gill and digestive gland tissues. These inhibitions could directly affect the resistance of these organisms to a secondary stressor (temperature, pathogens, hypoxia etc.) and threaten organismal health. Ph. D.

Soil organic matter is seriously affected by fires and suffers changes in stock, composition, and distribution. In the North-West side of the Cantabrian Range (northern Spain) fires are very common. In order to develop a cartographic technique to map areas with high carbon stocks caused by fire, we test a technique based on calibrated VIS-NIR soil organic carbon models and hyperspectral images. Total (TOC) and oxidizable carbon (OC) were measured in 89 soil samples. The samples were scanned with VIS-NIR spectrometer (400–2500 nm), and the spectra were resampled to the hyperspectral image channels. Spectroscopic models for TOC and OC were fitted (R2>0.81) using partial least squares regression (PLSR). The predictions were regionalized to the hyperspectral image and the results validated with a new soil population consisting of 12 Valeri plots collected in burned slopes of the study area under heather vegetation. In soil samples, TOC and OC values are highly correlated (R = 0.92), and the coefficients of the PLSR models have a similar pattern, which suggests similar organic components. Nevertheless, there are significant differences in the values of the regression coefficients, much higher in the TOC model except at 560 and 2054 nm that might be interpreted as labile carbon components, and at 1590 nm. At this wavelength the coefficient of TOC is positive and OC is negative, and it could be interpreted as hydrocarbons components present in the TOC model., This work was funded by the Government of Asturias through the project SV-PA-13-ECOEMP-40. J. Peón acknowledges a PhD Grant from the University of Oviedo and a PhD Grant “Severo Ochoa” from the Government of Asturias (BP14-104). We also acknowledge INTA for providing the images and for the geometric and atmospheric correction.

Soil spectroscopy can provide low-cost and high- density data for predicting various soil properties. However, a relatively weak correlation between the spectra and the measurements of salinized soil properties makes spectroscopy difficult to use in salinity assessment, especially for low and moderately saline soils. The main objective of the study was to propose an approach based on Vis-NIR spectroscopy and geostatistics for mapping soil salinity in the Neretva river valley (Croatia). An effective spectral index (SI), which synthesizes some of saline soil properties, was defined and used as a covariate in ordinary cokriging (COK) for improving electrical conductivity (ECe) estimation. The proposed approach was compared with a univariate estimator (ordinary kriging, OK), which uses only ECe data and a multivariate estimator (ordinary cokriging, COK) using some chemical properties of primary importance in salt affected soils (Ca++, Mg++, Na+, SO4–, Cl- concentrations and pH) as covariates . The study was carried out in an agricultural area (5068 ha) located in the Neretva River valley (Croatia). Topsoil (0-30 cm) samples were collected at 245 locations with a grid (500 m x 500 m) sampling scheme and analyzed for some chemical and physical properties. Moreover, soil samples were used for visible and near red (Vis- NIR) spectra measurements with a range of wavelength between 350 and 2500 nm. The spectral data were preprocessed and ECe was estimated by partial least squares regression (PLSR). The first significant latent variable that accounted for 85% of the total variance was selected and used as a SI to quantify and map spatial variation of soil salinity. The univariate and multivariate geostatistical approaches provided results and performances quite similar. Regarding the two multivariate approaches, the one using the spectral index has provided better results in terms of unbiasedness and accuracy. The spectral index SI, besides providing accurate estimates of ECe, is also very cost-effective and it would then make possible broad-scale surveys for soil salinity assessment at landscape scale.

Soil spectroscopy can provide low-cost and high-density data for predicting various soil properties. However, a relatively weak correlation between the spectra and the measurements of salinized soil makes spectroscopy difficult to use in predicting areas at risk of salinization, especially for low and moderately saline soils. The main objective of the study was to propose an effective approach based on Vis-NIR spectroscopy and geostatistics for mapping soil salinity in the Neretva River valley (Croatia). A spectral index (SI), which synthesizes most of salt affected soil properties, was defined and used as covariate in ordinary cokriging (COK) for improving electrical conductivity (EC e ) prediction. The proposed approach was compared with a univariate predictor (ordinary kriging, OK), which uses only EC e data and a multivariate predictor (ordinary cokriging, COK) using more covariates such as some chemical properties of primary importance in salt affected soils (Ca 2+ , Mg 2+ , Na + , SO 4 2− , Cl − concentrations and pH). The study was carried out in an agricultural area (5068 ha) located in the Neretva River valley (Croatia). Topsoil (0–30 cm) samples were collected at 246 locations with a grid (500 m × 500 m) sampling scheme and analyzed for some chemical and physical properties. Moreover, soil samples were used for visible and near infrared (Vis-NIR) spectra measurements with a range of wavelength between 350 and 2500 nm. The spectral data were pre-processed and EC e was predicted with partial least squares regression (PLSR). The first significant latent variable, which accounted for 85% of the total variance, was selected and used as a SI to quantify and map spatial variation of soil salinity. The univariate and multivariate geostatistical approaches provided results and performances quite similar. Regarding the two multivariate approaches, the one using only the spectral index as covariate has provided better results in terms of unbiasedness and accuracy. Moreover, the spectral index SI is also very cost-effective and it could then be used in possible broad-scale surveys for preventing soil salinity at landscape scale.

In describing soil variability, information on the distribution of soil properties is required in both the horizontal and vertical directions. This study investigated the potential of a real-time soil sensor (RTSS) for mapping six soil properties at multiple soil depths of a paddy field. Soil spectra were acquired at three depths using RTSS. Three calibration models were developed. The first model (CM1) combined the dataset for depths of 10 and 15 cm, the second model (CM2) combined the dataset for depths of 15 and 20 cm, and the third model (CM3) combined all the three depths. CM3 was the best calibration model for all the soil properties. The generated maps exhibited variations in the distribution of all the soil properties at different depths.

Visible and near infrared spectrometry (vis-NIRS) coupled with data mining techniques can offer fast and cost-effective quantitative measurement of total petroleum hydrocarbons (TPH) in contaminated soils. Literature showed however significant differences in the performance on the vis-NIRS between linear and non-linear calibration methods. This study compared the performance of linear partial least squares regression (PLSR) with a nonlinear random forest (RF) regression for the calibration of vis-NIRS when analysing TPH in soils. 88 soil samples (3 uncontaminated and 85 contaminated) collected from three sites located in the Niger Delta were scanned using an analytical spectral device (ASD) spectrophotometer (350–2500 nm) in diffuse reflectance mode. Sequential ultrasonic solvent extraction-gas chromatography (SUSE-GC) was used as reference quantification method for TPH which equal to the sum of aliphatic and aromatic fractions ranging between C10 and C35. Prior to model development, spectra were subjected to pre-processing including noise cut, maximum normalization, first derivative and smoothing. Then 65 samples were selected as calibration set and the remaining 20 samples as validation set. Both vis-NIR spectrometry and gas chromatography profiles of the 85 soil samples were subjected to RF and PLSR with leave-one-out cross-validation (LOOCV) for the calibration models. Results showed that RF calibration model with a coefficient of determination (R2) of 0.85, a root means square error of prediction (RMSEP) 68.43 mg kg− 1, and a residual prediction deviation (RPD) of 2.61 outperformed PLSR (R2 = 0.63, RMSEP = 107.54 mg kg− 1 and RDP = 2.55) in cross-validation. These results indicate that RF modelling approach is accounting for the nonlinearity of the soil spectral responses hence, providing significantly higher prediction accuracy compared to the linear PLSR. It is recommended to adopt the vis-NIRS coupled with RF modelling approach as a portable and cost effective method for the rapid quantification of TPH in soils.

Aim : The application and use of precision agriculture technologies targeting reduction in agricultural inputs are associated with various raw data and their assessment to create application maps. Yield mapping and soil mapping are the two basic agricultural application maps for variable-rate applications. Creation of high resolution digital soil maps requires high resolution information. The objective of this study was to predict total nitrogen in soil, using visible and near infrared (vis-NIR) reflectance spectroscopy, which is a new technology used for making fast and accurate measurements of physical and chemical properties of soil. Methodology : A set of 140 surface soil (0-20 cm) samples was collected from cultivated land, fruit, and vegetable fields distributed over Bursa province (Turkey). The diffuse reflectance spectra of soil samples were examined for total nitrogen content. Partial least squares regression (PLSR) analysis with full cross validation was performed to develop a model for total nitrogen. Results : The prediction performance of the model produced 0.793 regression coefficient (R),0.0274% root mean square errors of prediction (RM SEP) and 2.0802 residual prediction deviations (RPD). Interpretation : The results showed that the vis-NIR spectroscopy can be used to evaluate soil total nitrogen content. More research should be conducted to improve the model prediction of total nitrogen in soils with both high and low nitrogen content using vis-NIR reflectance spectroscopy.

Proximal soil sensing (PSS) is a promising approach when it comes to detailed characterization of spatial soil heterogeneity. Since none of existing PSS systems can measure all soil information needed for implementation precision agriculture, sensor data fusion can provide a reasonable al- ternative to characterize the complexity of soils. In this study, we fused the data measured using a gamma-ray sensor, an apparent electrical conductivity (ECa) sensor, and a commercial Veris MSP3 platform including a optical sensor measuring soil reflectance at 660 nm and 940 nm, a soil ECa sensor and a pH sensor, with the addition of topography for the prediction of several soil properties, i.e. soil organic matter, pH, buffer pH, phosphorus, potassium, calcium, magnesium, aluminum. A total of 65 sampling locations were selected from a 38.5 ha field in Ontario, Canada. Among them, 35 locations were selected by a random stratified sampling strategy. The stratification grid was 1 ha. Sampling was prohibited in areas near the field boundaries and within a safety margin from the drainage system. 20 locations were selected using a neighbourhood search approach , a spatial data integration strategy. These two sample datasets were used as the calibration dataset to build the model between soil properties and readings from different proximal soil sensors. The remaining 10 sensing locations were used as an independent validation dataset. Partial least square regression (PLSR) was performed on the data from each individual sensor and different sensor com- binations (sensor data fusion). For most soil properties, predictions based on sensor data fusion were better than those based on the output of individual sensors. By fusing the data from all of the proximal soil sensors, more properties can be predicted simultaneously (R2>0.5, and RPD>1.50). After choosing the optimal sensor combination for each soil property, different digital soil mapping methods, including support vector machines (SVM), random forest (RF), multivariate adaptive regression splines (MARS), regression trees (RT) and back-propagation artificial neural network (BP-ANN) were used to estimate variograms and pursue regression kriging. High resolution maps were thus interpolated with the most successful methods. The performance of the two different sampling strategies was compared by the prediction accuracy from the validation samples. We thus conclude that proximal soil sensor fusion paired with the digital soil mapping method is a promising way to offer the essential soil information needed for precision agriculture.

The need for rapid and inexpensive techniques for high-resolution soil information has led to improvements over traditional methods, and in particular those based on visible near-infrared (vis–NIR) spectroscopy. While vis–NIR has been used for soil classification for some preliminary studies, how to combine spectral information from soil profiles remains a substantial challenge. This study was undertaken to investigate the potential of vis–NIR to discriminate soil classes on profiles containing various soil horizons. We took 130 soil profiles at Zhejiang province, of which were classified in the field at suborder level according to Chinese Soil Taxonomy (5 soil orders and 10 suborders). Subsoil samples were taken by diagnosis layers (A, B and C). Support vector machine (SVM) algorithm was used to determine the soil classes, by analyzing quantitatively their diffuse reflectance spectra in the vis– NIR range. For SVM is a binary classification algorithm, the qualitative analysis was conducted by combining the votes of each sample from the same profile and the class got most votes in one profile was defined as their predicted soil class. Readily available variables (soil color) and well-predicted properties (soil organic matter, soil texture and pH) using vis-NIR spectra were added as auxiliary information. Using synthesized model (spectra plus auxiliary soil information), SVM produced better clas- sification performances at soil order level and suborder level (accuracy were 68.29% and 63.51% respectively) than spectra independently (accuracy were 60.69% and 58.54% respectively). They suggest that vis–NIR spectroscopy combining votes gained from SVM could make an essential contribution to the identification of soil classes in an effective approach of soil classification even when profiles contain various soil horizons.

Visible and near infrared (Vis-NIR) spectroscopy is a non-destructive analytical method that can be used to complement, enhance or potentially replace conventional methods of soil analysis. The aim of this research was to predict the particle size distribution (PSD) of soils using a Vis-NIR) spectrophotometry in one irrigate field having a vertisol clay texture in the Karacabey district of Bursa Province, Turkey. A total of 86 soil samples collected from the study area were subjected to optical scanning in the laboratory with a portable, fiber-type Vis-NIR spectrophotometer (AgroSpec, tec5 Technology for Spectroscopy, Germany). Before the partial least square regression (PLSR) analysis, the entire reflectance spectra were randomly split into calibration (80%) and validation (20%) sets. A leave-one-out cross-validation PLSR analysis was carried out using the calibration set with Unscrambler (R) software, whereas the model prediction ability was tested using the validation (prediction) set. Models developed were used to predict sand and clay content using on-line collected spectra from the field. Results showed an "excellent" laboratory prediction performance for both sand (R-2 = 0.81, RMSEP = 3.84% and RPD = 2.32 in cross-validation; R-2 = 0.90, RMSEP = 2.91% and RPD = 2.99 in the prediction set) and clay (R-2 = 0.86, RMSEP = 3.4% and RPD = 2.66 in cross validation; R-2 = 0.92, RMSEP = 2.67% and RPD = 3.14 in the prediction set). Modelling of silt did not result in any meaningful correlations. Less accurate on-line predictions were recorded compared to the laboratory results, although the on-line predictions were very good (RPD = 2.24-2.31). On-line predicted maps showed reasonable spatial similarity to corresponding laboratory measured maps. This study proved that soil sand and clay content can be successfully measured and mapped using Vis-NIR spectroscopy under both laboratory and on-line scanning conditions.

[ES] La almendra es una drupa de semilla comestible, consumida y comercializada en su estado natural o como parte de diversos productos elaborados. Se distinguen dos clases de almendras: amargas y dulces. El sabor amargo se debe a la presencia de compuestos cianogénicos, como la amigdalina, que es un compuesto altamente toxico que influye directamente en la calidad sensorial y aceptabilidad del producto. En este estudio se determinó del contenido de amigdalina en almendras mediante cromatografía líquida de alta resolución (HPLC). En primer lugar, se optimizaron las condiciones de extracción de la amigdalina de las almendras, seleccionándose metanol como disolvente y un tiempo de 24 h de agitación. Posteriormente se procedió a la optimización de las condiciones de separación, obteniéndose los mejores resultados con una fase móvil compuesta por una mezcla de agua y acetonitrilo (80:20, v/v) en elución isocrática con un caudal de 1 mL·min-1 y un volumen de inyección de 20 µL. El método cromatográfico desarrollado dio lugar a repetibilidades inter-día con RSDs menores de 0,31 y 1,41 % para tiempos de retención y áreas de pico, respectivamente. Además, se obtuvo un LOD del 0,40 mg·L -1 y recuperaciones en todos los casos cercanas al 100 %. El contenido de amigdalina obtenido en almendras dulces varió en un rango de 0,67 a 348 mg·Kg-1 , en tanto que para las almendras amargas se detectó un contenido de amigdalina entre 14734 y 50412 mg·Kg-1 . Previo a la determinación de amigdalina mediante HPLC, todas las almendras consideradas en él estudio fueron analizadas mediante espectroscopia visible (VIS) y de infrarrojo cercano (NIR). Los espectros obtenidos fueron empleados junto con tres técnicas distintas de clasificación (PLS-DA, SVM y LDA) para ver si eran capaces de discriminar entre almendras dulces y amargas sin necesidad de recurrir a los análisis por HPLC. Los mejores resultados se obtuvieron mediante el método PLS-DA, alcanzando una exactitud en la clasificación del conjunto de validación del 100 %, mientras que con los métodos SVM y LDA se obtuvo una clasificación global del 88,90 % y del 70,90 %, respectivamente. En base a estos resultados, se puede concluir que la técnica de espectroscopia VIS-NIR es una buena alternativa no destructiva para la clasificación de almendras en dulces y amargas, con posibilidades de adaptación in-line en un sistema de clasificación automatizado aportando así ventajas en relación al tiempo de detección, versatilidad de los productos a analizar y reducción de la inversión a realizar en los procesos industriales con el fin de asegurar un correcto control de calidad., [EN] The almond is an edible seed, consumed and marketed in its natural state or as part of processed products. There are two kinds of almonds: bitter and sweet. The bitter taste is due to the presence of cyanogenic compounds, such as amygdalin, which is a highly toxic compound that directly influences the sensory quality and acceptability of the product. In this study, the content of amygdalin in almonds was determined by high performance liquid chromatography (HPLC). Firstly, the conditions of extraction of the amygdalin in the almonds were optimized, methanol being selected as the solvent and a 24 h stirring time. Subsequently were optimized the separation conditions, the best results were obtained with a mobile phase composed of a mixture of water and acetonitrile (80:20, v / v) in isocratic elution with a flow rate of 1 mL min-1 and a volume of injection of 20 μL. The chromatographic method developed gave rise to inter-day repeatability with RSDs of less than 0.31 and 1.41 % for retention times and peak areas, respectively. In addition, a LOD of 0.40 mg·L-1 and recoveries were obtained in all cases close to 100 %. The content of amygdalin obtained in sweet almonds ranged from 0.67 to 348 mg·kg -1 , whereas for bitter almonds the amygdalin content detected was between 14734 and 50412 mg·kg-1 . Before determination of the amygdala by HPLC, all the almonds considered in the study were analyzed by visible (VIS) and near infrared (NIR) spectroscopy. The obtained spectra were used together with three different classification techniques (PLS-DA, SVM and LDA) to see if they were able to discriminate between sweet and bitter almonds without the use of a destructive HPLC technique. The best results were obtained using the PLS-DA method, with a precision in the classification of the 100 % validation group, whereas with the SVM and LDA methods an overall classification of 88.90 % and 70.90 % was obtained, respectively. Based on these results, it can be concluded that the VIS-NIR spectroscopy technique is a good non-destructive alternative for the classification of sweet and bitter almonds, with possibilities of adaptation online in an automated classification system, obtaining advantages in relation to the time of detection, the versatility to analyzed the products and the reduction of the investment to be made in the industrial processes to safe a correct quality control., [CA] L'ametla és una llavor comestible, consumida i comercialitzada en el seu estat natural o com a part de diversos productes elaborats. Es distingixen dos classes d'ametles: amargues i dolçes. El sabor amarg es deu a la presència de compostos cianogènics, com l'amigdalina, que és un compost altament tòxic que influïx directament en la qualitat sensorial i acceptabilitat del producte. En este estudi es va determinar el contingut d'amigdalina en ametles per mitjà de cromatografia líquida d'alta resolució (HPLC). En primer lloc es van optimitzar les condicions d'extracció d’amigdalina de les ametles, seleccionant-se metanol com a dissolvent i un temps de 24 h d'agitació. Posteriorment es va procedir a l'optimització de les condicions de separació, obtenint-se els millors resultats amb una fase mòbil composta per una mescla d'aigua i acetonitril (80:20, v/v) en elució isocràtica amb un cabal d'1 mL min-1 i un volum d'injecció de 20 µL. El mètode cromatogràfic desenvolupat va donar lloc a repetibilitats inter-dia amb RSDs menors de 0,31 i 1,41 % per a temps de retenció i àrees de pic, respectivament. A més, es va obtindre un LOD del 0,40 mg·L-1 i recuperacions en tots els casos pròximes al 100 %. El contingut d'amigdalina obtingut en ametles dolces va variar en un rang de 0,67 a 348 mg·Kg-1 , en tant que per a les ametles amargues es va detectar un contingut d’amigdalina entre 14734 y 50412 mg·Kg-1 . Previ a la determinació d'amigdalina per mitjà de HPLC, totes les ametles considerades en el estudi van ser analitzades per mitjà d'espectroscòpia visible (VIS) i la d'infraroig pròxim (NIR). Els espectres obtinguts van ser a continuació empleats junt amb tres tècniques diferents de classificació (PLSDA, SVM i LDA) per a veure si eren capaços de discriminar entre ametles dolces i amargues sense necessitat d’utilitzar una tècnica destructiva com l’HPLC. Els millors resultats es van obtindre amb el mètode PLS-DA, aconseguint una exactitut en la classificació del conjunt de validació del 100 %, mentres que amb els mètodes SVM i LDA es va obtindre una classificació global del 88,90 % i del 70,90 %, respectivament. Basant-se en estos resultats, es pot concloure que la tècnica d'espectroscòpia VIS-NIR és una bona alternativa no destructiva per a la classificació d'ametles en dolçes i amargues, amb possibilitats d'adaptació in-line en un sistema de classificació automatitzat aportant així avantatges en relació al temps de detecció, versatilitat dels productes a analitzar i reducció de la inversió a realitzar en els processos industrials a fi d'assegurar un correcte control de qualitat.

Vegetation properties can be estimated using optical sensors, acquiring data on board of different platforms. For instance, ground-based and Unmanned Aerial Vehicle (UAV)-borne spectrometers can measure reflectance in narrow spectral bands, while different modelling approaches, like regressions fitted to vegetation indices, can relate spectra with crop traits. Although monitoring frameworks using multiple sensors can be more flexible, they may result in higher inaccuracy due to differences related to the sensors characteristics, which can affect information sampling. Also organic production systems can benefit from continuous monitoring focusing on crop management and stress detection, but few studies have evaluated applications with this objective. In this study, ground-based and UAV spectrometers were compared in the context of organic potato cultivation. Relatively accurate estimates were obtained for leaf chlorophyll (RMSE = 6.07 µg·cm−2), leaf area index (RMSE = 0.67 m2·m−2), canopy chlorophyll (RMSE = 0.24 g·m−2) and ground cover (RMSE = 5.5%) using five UAV-based data acquisitions, from 43 to 99 days after planting. These retrievals are slightly better than those derived from ground-based measurements (RMSE = 7.25 µg·cm−2, 0.85 m2·m−2, 0.28 g·m−2 and 6.8%, respectively), for the same period. Excluding observations corresponding to the first acquisition increased retrieval accuracy and made outputs more comparable between sensors, due to relatively low vegetation cover on this date. Intercomparison of vegetation indices indicated that indices based on the contrast between spectral bands in the visible and near-infrared, like OSAVI, MCARI2 and CIg provided, at certain extent, robust outputs that could be transferred between sensors. Information sampling at plot level by both sensing solutions resulted in comparable discriminative potential concerning advanced stages of late blight incidence. These results indicate that optical sensors, and their integration, have great potential for monitoring this specific organic cropping system.

Visible Near-Infrared (Vis-NIR, 400-2500 nm) hyperspectral imaging has proven to be a valuable tool for mapping soil properties over bare soils. To date, most predictive models presented in literature, have been built from calibration databases made up of both Vis-NIR imaging spectra (predictor variables) and soil properties (response variables). Nevertheless, the constitution of such calibration databases is costly both in terms of time and money, as this implies soil sampling over hundreds of pixels of bare soils and physico-chemical soil analysis of these samples. We propose to test the transfer of predictive models built from laboratory database, to Vis-NIR hyperspectral image. Four transfer methods were tested (model updating, Repfile, Transfer by Orthogonal Projection TOP and Piecewise Direct Standardization PDS). Their respective performances were evaluated and compared to those obtained using two other predictive models. The first predictive model was obtained by directly applying a partial least square regression (PLSR) calibrated in laboratory, onto the hyperspectral image. The second model was obtained by applying a PLSR built from Vis-NIR spectra extracted from the hyperspectral image, onto the hyperspectral image. The transferred models are based on the use of standards. Standards were selected in order to (i) take into account or (ii) ignore field soil stratification knowledge. Moreover, the impact of the number of standards used in the transferred models was studied. This approach was tested for clay content prediction over the La Peyne Valley area in Southern France (24 km(2), in a Mediterranean context) covered by HyMap Vis-NIR airborne data. The results show that 1) applying PLSR model from laboratory spectra to hyperspectral image without calibration transfer, provides poor prediction performances (R-val(2) = 0.41, RMSEP = 103 g/kg), 2) transferring this model from laboratory spectra to hyperspectral image offered good prediction performances (R-test(2) median above 0.5 and RMSEP median below 70 g/kg), whatever the transfer method used and from only 15 used standards, and 3) taking into account field soil stratification knowledge in the standards selection improves prediction performances (R-test(2) median above 0.6 and RMSEP median below 60 g/kg) whatever the transfer method and from only 10 used standards. Finally, transferring models from the laboratory to a hyperspectral image gave better mapping performances than those obtained using PLSR models built from airborne spectra.

This study was undertaken to investigate the application of visible and near infrared (vis -NIR) spectroscopy for determining soil cation exchange capacity (CEC) under laboratory and on-line field conditions. Measurements were conducted in two fields with clay texture in field 1 (F1) and clay-loam texture in field 2 (F2) both in Turkey. Partial least squares (PLS) regression analyses with full cross-validation were carried out to establish CEC models using three datasets of F1, F2 and F1 + F2. Analytically-measured, laboratory vis-NIR and on-line vis-NIR predicted maps were produced and compared statistically by kappa coefficient. Results of the CEC prediction using laboratory vis-NIR data gave good prediction results, with averaged r(2) values of 0.92 and 0.72, root mean squared errors of prediction (RMSEP) of 1.89 and 1.54 cmol kg(-1) and residual prediction deviations (RPD) of 3.69 and 1.89 for F1 and F2, respectively. Less successful predictions were obtained for the on-line measurement with r(2) of 0.75 and 0.7, RMSEP of 4.79 and 1.76 cmol kg(-1) and RPD of 1.45 and 1.56 for F1 and F2, respectively. Comparisons using kappa statistics test indicated a significant agreement (kappa = 0.69) between analytically-measured and laboratory vis-NIR predicted CEC maps of F1, while poorer agreement was found for F2 (kappa = 0.43). A moderate spatial similarity was also found between analytically-measured and on-line vis-NIR predicted CEC maps in F1 (kappa = 0.50) and F2 (kappa = 0.49). This study suggests that soil CEC can be satisfactorily analysed using vis-NIR spectroscopy under laboratory conditions and with somewhat less precision under on-line scanning conditions. ICT-AGRI (62-FARMFUSE) (The European Commission's ERA-NET scheme under the 7.Framework Programme) Department for Environment, Food & Rural Affairs (DEFRA) - IF0208

It is widely known that the visible and near infrared (VIS-NIR) spectroscopy has the potential of estimating soil total nitrogen (TN), organic carbon (OC) and moisture content (MC) due to the direct spectral responses these properties have in the near infrared (NIR) region. However, improving the prediction accuracy requires advanced modelling techniques, particularly when measurement is planned for fresh (wet and un-processed) soil samples. The aim of this work is to compare the predictive performance of two linear multivariate and two machine learning methods for TN, OC and MC. The two multivariate methods investigated included principal component regression (PCR) and partial least squares regression (PLSR), whereas the machine learning methods included least squares support vector machines (LS-SVM), and Cubist. A mobile, fibre type, VIS-NIR spectrophotometer was utilised to collect soil spectra (305–2200 nm) in diffuse reflectance mode from 140 wet soil samples collected from one field in Germany. The results indicate that machine learning methods are capable of tackling non-linear problems in the dataset. LS-SVMs and the Cubist method out-performed the linear multivariate methods for the prediction of all three soil properties studied. LS-SVM provided the best prediction for MC (root mean square error of prediction (RMSEP) = 0.457% and residual prediction deviation (RPD) = 2.24) and OC (RMSEP = 0.062% and RPD = 2.20), whereas the Cubist method provided the best prediction for TN (RMSEP = 0.071 and RPD = 1.96).

There is increased interest in the investigation and implementation of rapid, non-destructive methods for detection of pesticide residues. This study reports development of a method for determination the pesticide concentration on the surface of fruit samples by using visible near-Infrared (Vis-NIR) spectroscopy. A total of 330 navel oranges were selected for the calibration (n=220) and prediction (n=110) sets, and sprayed dichlorvos solution diluted with tap water on the surface of fruit samples. After 12 h, the spectra of fruit samples were acquired in the range of 350-1800 nm, then pesticide residue was measured using gas chromatography (GC). Chemometrics analysis using Partial Least Squares (PLS) was performed in order to determine the concentration of pesticide residue. The particle swarm optimization (PSO) was used for wavelengths selection of Vis-NIR spectra data in order to develop the PLS prediction model. The model using optimal intervals for predicting the concentration of dichlorvos residue selected by PSO algorithm yielded better results than the PLS model using full wavelengths. This PSO-PLS model could predict dichlorvos residue of 220 samples of prediction set with correlation coefficient of 0.8732. The wavelength selection through a PSO algorithm would enhance the predictive ability when applying PLS model.

Calibration modeling is necessary to implement spectroscopic measurement of soil properties. The aim of this study is to compare the performance of calibration models developed for an agricultural farmland using two spectrophotometers with different scanning ranges, e.g. 400-2100 nm and 400-2500 nm. Soil samples were collected from an experimental farm at Silsoe, Bedfordshire, United Kingdom. The targeted soil properties under investigation were total nitrogen (TN) and organic carbon (OC). Spectra were divided into a calibration set (75%) and an independent validation set (25%). A partial least squares regression (PLSR) with leave-one-out cross validation was carried out to build calibration models based on the two spectral ranges. Validation result shows that the PLSR models developed on the range of 400-2500 nm outperform those on 400-2100 nm. It suggests that wavelengths located in the spectral band of 2100-2500 nm contribute to model calibration of soil N and C. It is concluded that spectroscopic calibration using 400-2500 nm could produce higher prediction accuracy for soil N and C measurement at a farm scale than short spectral ranges.

Both high nutrition and good tasting are the crucial factors of good agriculture product. Therefore, nondestructive measurement and prediction of fruit internal quality are an area that both technology and market section concern about. The objectives of this study were to evaluate the use of visible and near infrared (VIS-NIR) (380-1000 nm) spectroscopy for soluble solid content (SSC) prediction of apple; and to establish the relationship between the VIS-NIR reflectance spectra and the SSC of apples The reference SSC of apples was measured by using traditional destructive method. An artificial neural network with feedforward back-propagation (ANN-FFBP) was used to build the predictive model in this study. The ANN predictive model indicated good performance of SSC prediction with mean square error (MSE) of 0.1893 and a correlation coefficient (r) of 0.9668.

Soil water repellency (WR) is a widespread phenomenon caused by aggregated organic matter (OM) and layers of hydrophobic organic substances coating the surface of soil particles. These substances have a very low surface free energy, reducing a soil's water attraction. There is focus on WR due to its effects on germination, root growth, liquid–vapour dynamics, surface erosion and leaching of chemicals through fingered flow paths. However, common techniques for measuring WR are time-consuming and expensive. Meanwhile, it is well established that visible near infrared (vis-NIR) spectroscopy is a reliable method for determining soil OM. Potentially it could therefore provide fast measurements of WR through autocorrelation with OM. The aim of this study was to test the feasibility of vis-NIR spectroscopy for estimating the WR of soils with a small gradient in soil organic carbon (SOC) and texture, and to evaluate the effect of soil pretreatment on the predictive ability of WR models. A total of 87 soil samples from an agricultural coarse sandy field in Denmark were analysed for SOC, particle size fractions, water content and WR. Soil samples were scanned with a vis-NIR sensor (350–2500 nm) after air- and oven-drying at 60°C and 105°C. WR, expressed as liquid surface tension (mN m−1), was determined using the molarity of ethanol droplet test. Partial least squares regression models of SOC, texture and water content showed no predictive ability ( r2 values between 0.10 and 0.51). However, successful models ( r2 = 0.85) were generated for WR. The majority of bands important in the vis-NIR region of WR models were related to different components of OM indicating that, across the investigated field, WR was related to specific hydrophobic components of soil OM rather than to the total amount of carbon. A lower prediction error of the WR model for soils dried at 105°C (1.93 mN m−1) than at 60°C (2.52 mN m−1) can be explained by a lower range of WR values for the soils dried at 105°C. Moreover, a higher temperature reduced the number of absorption bands related to OM, indicating a degradation of hydrocarbon groups and a more hydrophobic character of the soil.

Fertiliser applications in vegetable crops are one of the main input costs of production. Thematic soil maps have been widely used for decades to characterise soil nutrients and, therefore, apply variable rate fertilisers. However, traditional variable rate methods used in soil sampling are time- consuming, costly and not accurate. Thus, they fail in providing a true estimate of the nutrients soil needs. To obtain better crop response to inputs, a rapid, non-destructive, timely and cost-effective soil analysis are needed to enable site-specific fertiliser applications. Proximal soil sensing with visible and near infrared (vis-NIR) spectroscopy is a promising tool to assist in variable rate applications. This thesis aims to develop reliable calibration models for a previously developed on-line visible (vis) and near infrared (NIR) spectroscopy sensor (Mouazen, 2006), for the prediction of soil properties in vegetable crop fields for a better N fertiliser management. Experiments were established in crops of cauliflower (Brassica oleracea) during 2013 season (two fields) and 2014 season (three fields), in UK. A mobile, fibre-type, vis–NIR spectrophotometer (AgroSpec, Tec5 Technology for Spectroscopy, Germany) with a measurement range of 305-2200 nm was used to measure soil spectra in diffuse reflectance mode, measuring up to ~1500 points per ha. Four different calibration sets were tested to establish the most accurate calibration model for moisture content (MC), soil organic carbon (OC), pH and total nitrogen (TN), using partial least squares (PLS) regression analysis selected according to different spectral library size and geographical scale: Scenario 1 (SC1 (local)), Scenario 2 (SC2 (regional)), Scenario 3 (SC3 (national)), Scenario 4 (SC4 (continental)). The best results in cross-validation were obtained for MC with SC2 (R2[R squared] = 0.89; RPD > 2.5), followed by SC4 (R2[R squared] = 0.88; RPD = 2.91-3.31, in 2013 and 2014, respectively); and SC1 and SC4 worked very well for MC on- line prediction (R2[R squared] > 0.90 and RPD > 2.5). SC3 and SC4 both provided the best performance for OC and TN in cross-validation, whereas no clear trend was observed for on-line prediction. Poor model performance was obtained for pH in on-line predictions (R2[R squared] < 0.30 and RPD < 0.9). Although the calibration models using the on-line vis-NIR sensor provided good and detailed information of the soil nutrients analysed, future research will be needed to estimate these properties more accurately, with the aim to develop reliable vis-NIR calibration models for the on-line measurement in vegetable crop fields.

Texture is a primary variable that influences the total amount of carbon stock in the soil. The standard methods for determining soil texture, however, are still conducted manually and are extremely time-consuming. Reflectance spectroscopy in the visible, near infrared (Vis-NIR, 350-2500 nm) spectral region could be an alternative to standard laboratory methods. The aim of this paper was to develop calibration models based on laboratory Vis-NIR spectroscopy and PLSR analysis to estimate the texture (sand, 2-0.05 mm; silt, 0.05-0.002 mm; clay, less than 0.002 mm) in a forest area of southern Italy. An additional objective was to produce continuous maps of sand, silt and clay through a geostatistical approach. Soil samples were collected at 235 locations in a forest area of southern Italy. Samples were dried, sieved at 2 mm and analyzed in laboratory for soil texture and Vis-NIR spectroscopic measurements. Spectra showed that soil samples could be spectrally separable on the basis of classes of texture. To establish the relationships between spectral reflectance and soil texture (sand, silt and clay) partial least squared regression (PLSR) analysis was applied to 175 soil samples, while the remaining 60 samples were used to validate the models. The optimum number of factors to be retained in the calibration models was determined by leave-one-out cross-validation. Results of cross validation of calibration models indicated that the models fitted quite well and the values of R2 ranged between a minimum value of 0.74% for silt and a maximum value of 0.84 for sand content. Results for validation were satisfactory for sand content (R2=0.81) and clay content (R2=0.80) and less satisfactory for silt content (R2=0.70). Geostatistics coupled with Vis-NIR reflectance spectroscopy allowed us to produce continuous maps of sand, silt and clay, which are of critical importance for understanding and managing forest soils.

En la actualidad la recolección de la aceituna de mesa se realiza principalmentede forma manual. Sin embargo, el elevado coste de la mano de obra, su disponibilidadtemporal y el bajo rendimiento de campo genera un reducido beneficio económico alsector olivarero. Por ello, la Organización Interprofesional de la Aceituna de Mesa(INTERACEITUNA) ha impulsado un proyecto de investigación para la adaptación ymejora de los sistemas de recolección mecanizada a las plantaciones de olivar de mesamediante el empleo de vibradores de troncos.Entre las líneas de trabajo, el estudio de los sistemas de recolección disponiblesy la reducción de los daños provocados al fruto por la recolección mecanizada son dosde los principales retos para la obtención de un producto de calidad. El presente trabajotiene el objetivo de evaluar la calidad del fruto durante la recolección mecanizada convibradores de troncos mediante técnicas e inspección visual, análisis de imágenes yespectroscopia VIS-NIR a través de la determinación del molestado, el estudio de lasusceptibilidad a daño mecánico de los frutos provocados por la recolección convibradores de troncos y la generación de modelos PLS que permitan determinar el gradode molestado de los frutos en función de la energía del impacto. Los resultadosobtenidos ofrecen información útil sobre la incidencia de los daños y evolución delmolestado según el sistema de recolección, la variedad, temperatura, tiempo tras larecolección y energía del impacto. Por último, el escandallo digital mediante análisis deimágenes, la espectrofotometría de infrarrojo mediante la medida de la reflectancia y elíndice de molestado para aceituna de mesa presentado en este estudio, ha permitidoevaluar la calidad de la cosecha tras la recolección., Nowadays, table olives harvesting is realized by hand, which is a seriousproblem for the viability of olive farms and hence industries, because the high cost ofmanual harvesting leaves a narrow profit margin for the producers. In this scenario, theSpanish Association of table olive producers and industries (INTERACEITUNA) haspromoted a project for mechanical harvesting with trunk shakers of this cultivar.However, before doing that it is necessary to reduce bruise damage and study theharvesting system during the process to preserve fruit quality.The main objective of this research work has been to determine fruit qualityduring harvesting with trunk shakers by means of manual inspection, image analysisand VIS-NIR spectroscopy. In order to achieve this objective olive bruise has beenquantified, susceptibility to mechanical damage has been studied and PLS models havebeen developed to determine bruise damage in terms of impact energy. Usefulinformation about bruise incidence and bruise time evolution depending on harvestingsystem, variety, temperature, impact energy and time after harvesting has been obtained.Finally, bruise index by image analysis, infrared spectroscopy in reflectance mode andbruise reflectance index for table olives presented in this study allowed evaluating fruitquality after harvesting.

Soil compaction is a widely spread problem in agricultural soils that has negative agronomic and environmental impacts. The former may lead to poor crop growth and yield, whereas the latter may lead to poor hydraulic properties of soils, and high risk to flooding, soil erosion and degradation. Therefore, the elimination of soil compaction must be done on regular bases. One of the main parameters to quantify soil compaction is soil bulk density (BD). Mapping of within field variation in soil BD will be a main requirement for within field management of soil compaction. The aim of this research was to develop a new approach for the measurement of soil BD as an indicator of soil compaction. The research relies on the fusion of data from visible and near infrared spectroscopy (vis-NIRS), to measure soil gravimetric moisture content (ω), with frequency domain reflectometry (FDR) data to measure soil volumetric moisture content (θv). The values of the estimated ω and θv, for the same undisturbed soil samples were collected from selected locations, textures, soil moisture contents and land use systems to derive soil BD. A total of 1013 samples were collected from 32 sites in the England and Wales. Two calibration techniques for vis-NIRS were evaluated, namely, partial least squares regression (PLSR) and artificial neural networks (ANN). ThetaProbe calibration was performed using the general formula (GF), soil specific calibration (SSC), the output voltage (OV) and artificial neural networks (ANN). ANN analyses for both ω and θv properties were based either on a single input variable or multiple input variables (data fusion). Effects of texture, moisture content, and land use on the prediction accuracy on ω, θv and BD were evaluated to arrive at the best experimental conditions for the measurement of BD with the proposed new system. A prototype was developed and tested under laboratory conditions and implemented in-situ for mapping of ω, θv and BD. When using the entire dataset (general data set), results proved that high measurement accuracy can be obtained for ω and θv with PLSR and the best performing traditional calibration method of the ThetaProbe with R2 values of 0.91 and 0.97, and root mean square error of prediction (RMSEp) of 0.027 g g-1 and 0.019 cm3 cm-3, respectively. However, the ANN – data fusion method resulted in improved accuracy (R2 = 0.98 and RMSEp = 0.014 g g-1 and 0.015 cm3 cm-3, respectively). This data fusion approach gave the best accuracy for BD assessment when only vis-NIRS spectra and ThetaProbe V were used as an input data (R2 = 0.81 and RMSEp = 0.095 g cm-3). The moisture level (L) impact on BD prediction revealed that the accuracy improved with soil moisture increasing, with RMSEp values of 0.081, 0.068 and 0.061 g cm-3, for average ω of 0.11, 0.20 and 0.28 g g-1, respectively. The influence of soil texture was discussed in relation with the clay content in %. It was found that clay positively affected vis-NIRS accuracy for ω measurement and no obvious impact on the dielectric sensor readings was observed, hence, no clear influence of the soil textures on the accuracy of BD prediction. But, RMSEp values of BD assessment ranged from 0.046 to 0.115 g cm-3. The land use effect of BD prediction showed measurement of grassland soils are more accurate compared to arable land soils, with RMSEp values of 0.083 and 0.097 g cm-3, respectively. The prototype measuring system showed moderate accuracy during the laboratory test and encouraging precision of measuring soil BD in the field test, with RMSEp of 0.077 and 0.104 g cm-3 of measurement for arable land and grassland soils, respectively. Further development of the prototype measuring system expected to improve prediction accuracy of soil BD. It can be concluded that BD can be measured accurately by combining the vis-NIRS and FDR techniques based on an ANN-data fusion approach.

The project LIFE09 ENV/IT/078 Managing forests for multiple purposes: carbon, biodiversity and socio-economic wellbeing (ManFor C.BD.), among other objectives, is aimed to evaluate carbon sequestration taking into account forest management for conserving and enhancing carbon stocks, and increase carbon sequestration. The first step to do that is modelling and mapping spatial variability of soil organic carbon to assess the ability of soil to sequestrate biologic carbon. Conventional methods for soil organic carbon (SOC) determination are based on laboratory analyses and are costly and time consuming. Soil spectral reflectance is an alternative approach for SOC determination and has the advantage to be rapid, non-destructive and cost effective. The paper was aimed at: 1) quantifying SOC using laboratory-based soil visible, near infrared (Vis-NIR) reflectance spectroscopy and 2) mapping SOC using a geostatistical approach. The study area was a forest beech (332.000 m2) located in southern Italy at a mean elevation of about 1,180 m a.s.l. Soils mainly are coarse-textured, relatively young (Entisol and Inceptisol) and from poorly to moderately differentiated. They are strongly dependent on the nature of the parent rock (Paleozoic granitoid rocks) and their depth range from shallow to moderately deep and soil profiles are characterized by A-Bw-Cr and/or A-Cr horizons. The upper A horizon has an high accumulation of organic matter (umbric epipedon) and a very dark brown colour. Soil samples were collected at 235 locations within the study area using a metallic core cylinder with a diameter of 7.5 cm and a height of 20 cm. Each sample was oven dried at 45° for 48 hours and sieved at 2 mm and was used before for spectroscopic measurements and then analysed to determine organic carbon content. The Vis-NIR reflectance of soil samples was measured in laboratory, under artificial light, using an ASD FieldSpec IV 350 - 2500 nm spectroradiometer (Analytical Spectral Devices Inc., Boulder, Colorado, USA), while SOC was determined using a TOC-L analyzer (Shimadzu Corporation, Kyoto, Japan). The partial least squared regression (PLSR) analysis was used to establish the relationships between spectral reflectance and OM content. PLSR was performed on 2/3 of the soil samples. The optimum number of factors to retain in the calibration models were determined by cross validation. The models were independently validated using data from the remaining soil samples. Results revealed a high level of agreement between measured and predicted values with coefficients of determination. Finally, ordinary kriging was used to map SOC in the study area.