Soil Reflectance Research Articles

Assessing Soil and Crop Characteristics at Sub-Field Level Using Unmanned Aerial System and Geospatial Analysis

Practicing agriculture is a multiparametric and for this reason demanding task. It involves the management of many factors and thorough strategic planning in a highly variable and uncertain environment. Crop production is a function of agricultural practices as applied in natural resources, such as soil and plants. When referring to conventional agriculture, variability in these resources is neglected, as any field is treated homogenously. On the other hand, site-specific crop management, which was promoted through the advance of technologies, regarding collecting and analyzing data and applying agricultural decisions at a sub-field level, considers field spatial and temporal variations. Localizing inputs in a field rationalizes agricultural waste management and offers promising perspectives towards a circular economy. In this context, two cotton fields in central Greece were selected for this study. During the growing period, reflectance data were acquired, before planting at the end of April, and 100 days after planting at the end of July, with a commercial unmanned aerial system (UAS). The fields were grid sampled for soil (clay content, pH, calcium carbonate percentage, organic matter, total nitrogen, and electrical conductivity) and plant properties (total nitrogen, potassium, iron, copper, and zinc) determination. All data were manipulated through geographical information systems (GIS) and further participated in principal component analysis (PCA) application. PCA revealed important relations and groupings between soil reflectance and organic matter, carbonates, and clay content in both fields (72 to 87% of the total variance in the initial parameters was explained by the extracted components). However, in plant data, the resulting components accounted for less variability in initial data (62 to 72%). PCA resulting scores were introduced in the Fuzzy c-means clustering algorithm, which categorized sub-areas of the fields into two discrete zones per field. Zoning, in the case of soil properties, was accompanied with the statistically important (p < 0.01) discrimination of the mean values (except for total nitrogen and pH), implicating a promising zonal management scheme. The zone delineation process regarding plant properties yielded areas that did not share statistically significant variations, except for the mean values of iron concentration (p < 0.01). According to the results, spatial variations were revealed across the fields, mostly in soil properties, which can be directly monitored through aerial reflectance data. The applied methodology can be used in extension services or by agronomists for producing fertilizer application maps. Further, when integrated with a broader spatial decision support system, it can be used by policy makers for adapting circular economy strategies in crop production.

Hyperspectral inversion of soil heavy metals in Three-River Source Region based on random forest model

Hyperspectral remote sensing technology has considerable research value in monitoring and evaluating soil heavy metal pollution. In this study, the Three-River Source Region was taken as the study area. The occurrence relationship of six heavy metals in soil, such as Mn, Cu, Zn, Pb, Cr, Ni, with soil organic matter, clay minerals, and iron-manganese oxides, was studied through the determination and analysis of soil samples and the collection of soil reflectance spectrum. Spectral transformation was carried out by first derivative, second derivative, inverse-log, continuum removal and multiple scattering correction of the spectrum. The correlation between soil heavy metal content and soil spectrum was analyzed to select the characteristic band, and partial least squares (PLS) method, support vector machine (SVM) method and random forest (RF) model were used to build inversion model based on characteristic band. Then the best combination of spectral transformation and inversion model were explored. The results showed that Pb contents were the twice of the background in Qinghai province. The combination spectrum processing method can improve the correlation between spectrum and heavy metals. The location and quantity of characteristic bands of six heavy metals are different. The accuracy of RF was significantly better than that of SVM and PLS for all six heavy metal (i.e. pb: R2RF = 0.83, R2SVM = 0.62, R2PLS = 0.18), and the model effective of soil properties in non-polluted sites were reliable (i.e. clay: R2RF = 0.93, R2SVM = 0.87, R2PLS = 0.74). This study can provide technical support for the larger-scale monitoring of soil heavy metal content and heavy metal pollution assessment.

Mapping barrier island soil moisture using a radiative transfer model of hyperspectral imagery from an unmanned aerial system

The advent of remote sensing from unmanned aerial systems (UAS) has opened the door to more affordable and effective methods of imaging and mapping of surface geophysical properties with many important applications in areas such as coastal zone management, ecology, agriculture, and defense. We describe a study to validate and improve soil moisture content retrieval and mapping from hyperspectral imagery collected by a UAS system. Our approach uses a recently developed model known as the multilayer radiative transfer model of soil reflectance (MARMIT). MARMIT partitions contributions due to water and the sediment surface into equivalent but separate layers and describes these layers using an equivalent slab model formalism. The model water layer thickness along with the fraction of wet surface become parameters that must be optimized in a calibration step, with extinction due to water absorption being applied in the model based on equivalent water layer thickness, while transmission and reflection coefficients follow the Fresnel formalism. In this work, we evaluate the model in both field settings, using UAS hyperspectral imagery, and laboratory settings, using hyperspectral spectra obtained with a goniometer. Sediment samples obtained from four different field sites representing disparate environmental settings comprised the laboratory analysis while field validation used hyperspectral UAS imagery and coordinated ground truth obtained on a barrier island shore during field campaigns in 2018 and 2019. Analysis of the most significant wavelengths for retrieval indicate a number of different wavelengths in the short-wave infra-red (SWIR) that provide accurate fits to measured soil moisture content in the laboratory with normalized root mean square error (NRMSE)< 0.145, while independent evaluation from sequestered test data from the hyperspectral UAS imagery obtained during the field campaign obtained an average NRMSE = 0.169 and median NRMSE = 0.152 in a bootstrap analysis.

Multiscaling NDVI Series Analysis of Rainfed Cereal in Central Spain

Vegetation indices time series analysis is increasingly improved for characterizing agricultural land processes. However, this is challenging because of the multeity of factors affecting vegetation growth. In semiarid regions the rainfall, the soil properties and climate are strongly correlated with crop growth. These relationships are commonly analyzed using the normalized difference vegetation index (NDVI). NDVI series from two sites, belonging to different agroclimatic zones, were examined, decomposing them into the overall average pattern, residuals, and anomalies series. All of them were studied by applying the concept of the generalized Hurst exponent. This is derived from the generalized structure function, which characterizes the series’ scaling properties. The cycle pattern of NDVI series from both zones presented differences that could be explained by the differences in the climatic precipitation pattern and soil characteristics. The significant differences found in the soil reflectance bands confirm the differences in both sites. The scaling properties of NDVI original series were confirmed with Hurst exponents higher than 0.5 showing a persistent structure. The opposite was found when analyzing the residual and the anomaly series with a stronger anti-persistent character. These findings reveal the influences of soil–climate interactions in the dynamic of NDVI series of rainfed cereals in the semiarid.

Information depth of NIR/SWIR soil reflectance spectroscopy

Proximal and remote sensing techniques in the optical domain are cost-effective alternatives to standard soil property characterization methods. However, the extent of light penetration into the soil sample, also termed soil information depth, is not well understood. In this study a new analytical model that links the particle size distribution and soil reflectance in the near infrared (NIR) and shortwave infrared (SWIR) bands of the electromagnetic spectrum is introduced. The model enables the partitioning of measured reflectance spectra into surface and volume (subsurface) contributions, thereby yielding insights about the soil information depth. The model simulations indicate that the surface reflectance contribution to the total reflectance is significantly higher than the volume reflectance contribution for a broad range of soils that vastly differ in texture, mineralogical composition and organic matter contents. The ratio of volume to total reflectance is higher for sandy soils than for clayey soils, especially at longer optical wavelengths, but the ratio rarely exceeds 15%. Therefore, the light reflection from dry soils is predominantly a surface phenomenon and the information depth in most soils rarely exceeds 1 mm. The results of this study reveal an intimate physical relationship between soil reflectance and the particle size distribution in the NIR/SWIR range, which opens a potential new avenue for retrieval of the particle size distribution from remotely sensed reflectance via a universal process-based approach.

Spectral reflectance properties of vertisols and associated soils of Nagpur district in Maharashtra

Morphological, physical, chemical and spectral reflectance properties of soils of Nagpur district, Maharashtra were studied to understand the relationship between soil spectral reflectance and physical and chemical properties of soil. A total of 290 soil samples (both surface and sub-surface and profiles) were collected from 143 sites representing different physiographic units. The soils varied distinctly in Munsell colour hue and the soil texture ranged from sandy loam to clay with smectite as the dominant clay minerals. The soil pH varied from 4.9 to 9.2 and organic carbon (OC) content ranged from 1.0 to 24.4 g kg−1. The soil exchange complex was dominated by calcium (Ca) followed by magnesium (Mg), sodium (Na) and potassium (K). In general, most of the soils were rich in available iron (Fe), manganese (Mn) and copper (Cu) but deficient in zinc (Zn). Reflectance spectra of the soils showed prominent absorption features around 1400, 1900 and 2200 nm wavelength region. The soils with colour hue of 7.5YR, 5YR and 2.5YR showed a broad absorption features around 550 in visible and 950 nm in NIR region due to the presence Fe-oxides. Derivative soil reflectance data were calibrated with different soil attributes using step-wise multiple linear regression technique to develop spectral models for prediction of soil attributes. The developed spectral models showed relatively good prediction of sand (r2 = 0.74; RPD=1.86), clay (r2 = 0.73; RPD=1.85), pH (r2 = 0.72; RPD=1.87), OC (r2 = 0.70; RPD=1.82), exch Ca (r2 = 0.61; RPD=1.48), CEC (r2 = 0.70; RPD=2.16), CaCO3 (r2 = 0.62; RPD=1.54), Fe2O3 (r2 = 0.53; RPD=1.43), DTPA-Fe (r2 = 0.79; RPD=2.11) and DTPA-Cu (r2 = 0.77; RPD=1.96) in the validation datasets indicating that reliable prediction of these properties from soil reflectance data.

A new indicator of forest fire risk for Indonesia based on peat soil reflectance spectra measurements

ABSTRACT Conventional fire risk prediction using vegetation indexes is inappropriate because changes in soil moisture caused by plant drought stress have a 1 to 2 month time lag, which is too long for fire risk management. Further, soil and water indices used for traditional bare land assessments are also unsuitable because Indonesia is partially covered by peat soil, which has different reflection characteristics compared to typical agricultural and field soils. Here we describe a new index for estimating peat soil moisture using satellite remote sensing data. The method was developed using the ultra-blue band (435 to 451 nm) of the Land Remote-Sensing Satellite (Landsat 8) and is based on the moisture dependence of the reflected spectra measured directly. Our developed index showed a relatively strong correlation (r = 0.56 to 0.63) with real wildfire points and was a more reliable index than conventional measures used for fire risk management.

Temporal mosaicking approaches of Sentinel-2 images for extending topsoil organic carbon content mapping in croplands

The spatial assessment of soil organic carbon (SOC) is a major environmental challenge, notably for evaluating soil carbon stocks. Recent works have shown the capability of Sentinel-2 to predict SOC content over temperate agroecosystems characterized with annual crops. However, because spectral models are only applicable on bare soils, the mapping of SOC is often obtained on limited areas. A possible improvement for increasing the number of pixels on which SOC can be retrieved by inverting bare soil reflectance spectra, consists of using optical images acquired at several dates. This study compares different approaches of Sentinel–2 images temporal mosaicking to produce a composite multi-date bare soil image for predicting SOC content over agricultural topsoils. A first approach for temporal mosaicking was based on a per-pixel selection and was driven by soil surface characteristics: bare soil or dry bare soil with/without removing dry vegetation. A second approach for creating composite images was based on a per-date selection and driven either by the models performance from single-date, or by average soil surface indicators of bare soil or dry bare soil. To characterize soil surface, Sentinel-1 (S1)-derived soil moisture and/or spectral indices such as normalized difference vegetation index (NDVI), Normalized Burn Ratio 2 (NBR2), bare soil index (BSI) and a soil surface moisture index (S2WI) were used either separately or in combination. This study highlighted the following results: i) none of the temporal mosaic images improved model performance for SOC prediction compared to the best single-date image; ii) of the per-pixel approaches, temporal mosaics driven by the S1-derived moisture content, and to a lesser extent, by NBR2 index, outperformed the mosaic driven by the BSI index but they did not increase the bare soil area predicted; iii) of the per-date approaches, the best trade-off between predicted area and model performance was achieved from the temporal mosaic driven by the S1-derived moisture content (R2 ~ 0.5, RPD ~ 1.4, RMSE ~ 3.7 g.kg-1) which enabled to more than double (*2.44) the predicted area. This study suggests that a number of bare soil mosaics based on several indicators (moisture, bare soil, roughness…), preferably in combination, might maintain acceptable accuracies for SOC prediction whilst extending over larger areas than single-date images.

Optimizing the observation time for bare arable land to determine mean diurnal albedo in relation to roughness

ABSTRACT It is assumed that the approximation of the albedo values of the Earth’s surface components using satellite data obtained at the optimal time (T o) when these areas’ albedo reach mean diurnal values (α d), rather than these instantaneous values, allows Earth’s overall climate change to be modelled with greater accuracy. The T o values were investigated using examples of arable land areas in the 33 most extensive agricultural regions of the world, on given days of the year when the largest sections of these land areas become bare after major crops were planted under conventional tillage. This study is based on processed data obtained from nearly 900 soil reflectance spectra, representing 24 major soil groupings on 6 continents, assuming that these soils, which were either shaped by a plough (Pd) or smoothing harrow (Hs), were air-dried. It was found that in the morning the T o, expressed in solar local time, was expected to be earlier for rough bare soil surfaces than smooth ones. In the afternoon this relationship was reversed. The differences in the T o values that resulted from the different surface roughness of the same soils (formed by Pd and Hs) were 5–10 minutes. Deviations of about 5–20 minutes from T o in the observations of the studied land areas generated errors of ±2% in the approximations of their α d values.

Soil Moisture Retrieval Using Reflection Coefficients: Numerical Experiments

Ready access to small, inexpensive, unmanned aerial vehicles (UAVs) allows for small-scale electromagnetic propagation and scattering experiments using airborne antennas. We consider a low-power, copter-mounted transmit antenna radiating at frequencies on the order of a hundred megahertz, corresponding to wavelength on the order of meters, which at some horizontal distance generates a vertical interference pattern due to the interaction of a direct and ground-reflected wave. The depth of the interference fringes, which is directly related to the modulus of the soil reflection coefficient, can be measured by a second copter-mounted receiving antenna. By varying the geometry of the copter pair, and/or the transmitted signal frequency, the soil moisture profile up to a depth of a few meters can, in principle, be retrieved. In this article, we present the results of numerical experiments designed to evaluate the sensitivity of the angle- and frequency-dependence of the measured reflection coefficient to the soil moisture profile. Our simulations indicate that the retrieval errors are small suggesting that the technique is feasible.

Research of dependence of the earth's reflectance spectrum on type and parameters of wetted soil

Relevance and materials . The soil humidity being a major factor of its fertility can meaningfully change at the earth surface. R esults of researches of dependence of soil reflectance spectrum from soil type and its wetness parameters are described proanalytic conditions achievements high values of the signal reflected from wet soil are analyzed in relation to a series of measurements on multiple fields with different values of the Fresnel reflection. Results . It is shown that taking into account the possible change in the mutual dynamics of changes in f w and d from in phase to ant phase, the averaged reflected signal reaches a minimum when the specified parameters are in-phase mutually changed. On the basis of the discovered fact, a heuristic recommendation was given to use that area of the volumetric water content in the soil, where the ant phase change of the indicated indicators is provided.

In-situ multiple parameter calibration and mapping using a mobile soil sensor

ABSTRACT The purpose of this study was to estimate local calibration models for 30 chemical and 4 physical soil properties, based on visible and near-infrared subsurface soil reflectance spectra (VNIR-SRS). Further, we discussed a case of decision-making for precision farming by a grower, using high-resolution soil maps. The experimental sites were located in rotational upland fields of taro in Saitama Prefecture, Japan. The parent materials in the experimental sites were Light Clay, Clay Loam, and Sandy Clay Loam. A mobile soil sensor (SAS3000) attached to a tractor was used to measure the VNIR-SRS (350 to 1700 nm). A total of 231 field-moist soil samples and corresponding VNIR-SRS data were collected from 23 fields (6 growers), from 2017 to 2018 period, to estimate local calibration models for the 34 properties. A total of 34 local calibration models were estimated using the second derivative pretreatment and partial least squares regression analysis. Excluding several outliers, the coefficient of determination (R 2) for the local calibration models was 0.81 or more. Immediately after soil sensing in the field, we generated high-resolution soil property maps for the growers’ fields, using the VNIR-SRS data and the 34 local calibration models stored in the SAS3000. The grower involved in the map-based decision-making recalled past crop management, growth, yield, etc., in the target field, and selected a cost-effective soil management technique. In addition, the six growers in the study stressed the importance of understanding the high-resolution relative variability among multiple soil properties, between fields and within a given field.

Reflectance study of ice and Mars soil simulant associations – I. H2O ice

The reflectance of water ice and dust mixtures depends, amongst other parameters, on how the components are mixed (e.g. intimate mixture, areal mixture or coating). Therefore, when inverting the reflectance spectra measured from planetary surfaces to derive the amount of water ice present at the surface, it is critical to distinguish between different mixing modes of ice and dust. However, the distinction between mixing modes from reflectance spectra remains ambiguous. Here we show how to identify some water ice/soil mixing modes from the study of defined spectral criteria and colour analysis of laboratory mixtures. We have recreated ice and dust mixtures and found that the appearance of frost on a surface increases its reflectance and flattens its spectral slopes, whereas the increasing presence of water ice in intimate mixtures mainly impacts the absorption bands. In particular, we provide laboratory data and a spectral analysis to help interpret ice and soil reflectance spectra from the Martian surface.

Impact of Soil Permittivity and Temperature Profile on L-Band Microwave Emission of Frozen Soil

An unexplored aspect of L-band microwave emission is the impact of soil moisture and soil temperature (SMST) profile dynamics on diurnal brightness temperature ( T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B</sub> ) signatures of frozen soil. This study investigates this effect by comparing the T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B</sub> simulations of layered ( T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B,l</sub> ) and uniform ( T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B,u</sub> ) soils using a newly developed integrated land emission model. The multilayer Wilheit model and the single-layer Fresnel model are adopted to compute the smooth soil reflectivity for the layered and uniform soils, respectively. A four-phase dielectric mixing model is used to calculate the soil permittivity ( ε <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">s</sub> ). A data set of concurrent ELBARA-III T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B</sub> and SMST profile measurements performed in a seasonally frozen Tibetan meadow ecosystem is used for the analysis. The simulated T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B,l</sub> considering SMST profile information captures well the ELBARA-III measurements with low biases (≤6 K) and high correlations ( R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ≥ 0.88). T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B,u</sub> produced based on the Fresnel model using the soil moisture of 2.5 cm is more consistent with the T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B,l</sub> . The sensitivity test of averaging SMST profile below 2.5 cm leads to maximum differences of 2 K in T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B,l</sub> simulations, indicating that the T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B</sub> variations are primary dominated by the SMST dynamics at the surface layer. A sensitivity test of the Wilheit model to different ε <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">s</sub> parameterizations shows that the dielectric model of Zhang et al. is comparable to the four-phase dielectric model in simulating T <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">B,l</sub> , while the Mironov et al. 's model demonstrates larger biases for frozen soil with, on average, 2.2% clay content, 49.7% sand content, and a bulk density of 1 g·cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-3</sup> .

Reducing moisture effects on soil organic carbon content prediction in visible and near-infrared spectra with an external parameter othogonalization algorithm

Field spectroscopy and other efficient hyperspectral techniques have been widely used to measure soil properties, including soil organic carbon (SOC) content. However, reflectance measurements based on field spectroscopy are quite sensitive to uncontrolled variations in surface soil conditions, such as moisture content; hence, such variations lead to drastically reduced prediction accuracy. The goals of this work are to (i) explore the moisture effect on soil spectra with different SOC levels, (ii) evaluate the selection of optimal parameter for external parameter othogonalization (EPO) in reducing moisture effect, and (iii) improve SOC prediction accuracy for semi-arid soils with various moisture levels by combing the EPO with machine learning method. Soil samples were collected from grassland regions of Inner Mongolia in North China. Rewetting laboratory experiments were conducted to make samples moisturized at five levels. Visible and near-infrared spectra (350–2500 nm) of soil samples rewetted were observed using a hand-held SVC HR-1024 spectroradiometer. Our results show that moisture influences the correlation between SOC content and soil reflectance spectra and that moisture has a greater impact on the spectra of samples with low SOC. An EPO algorithm can quantitatively extract information of the affected spectra from the spectra of moist soil samples by an optimal singular value. A SOC model that effectively couples EPO with random forest (RF) outperforms partial least-square regression (PLSR)-based models. The EPO–RF model generates better results with R2 of 0.86 and root-mean squared error (RMSE) of 3.82 g kg−1, whereas a PLSR model gives R2 of 0.79 and RMSE of 4.68 g kg−1.

Design and application of multicolor image identification in soil pollution component detection

Aiming at the problem that the accuracy of the results obtained by traditional methods is low, the spectral detection in soil pollution component detection method is used to design multicolor heavy metal image identification. The soil reflectance spectra were measured using an ASD Fieldspec FR spectrometer to obtain soil heavy metal spectral data. Spectral absorption bands, spectral transformations and combinations, correlation analysis, and curve fitting operations are used to process and analyze spectral data. Among them, spectral transformation and combination are used to decompose overlapping mixed spectra by first-order and second-order differential operations, to expand the spectral characteristics of heavy metals in different soils, and to extract high-quality spectral data of soil heavy metals. The spectral data is combined with Newton’s classic rainbow spectrum to obtain different heavy metal spectral colors. Based on this, multicolor image identification of soil pollution components is designed. The results show that the multicolor heavy metal image identification can be used to design the multicolor heavy metal image identification by the spectral detection in the soil pollution component detection method, and it can play the pollution warning role.

Influence of Data Pre-Processing on the Results of PCA of Soil Reflectance Spectra

Influence of Data Pre-Processing on the Results of PCA of Soil Reflectance Spectra

ESTIMATION OF SOIL HEAVY METAL COMBINING FRACTIONAL ORDER DERIVATIVE

Abstract. It is important for the sustainable development of soil and monitoring the soil quality to obtain the heavy metal contents. Visible and near-infrared (Vis–NIR) spectroscopy provides an alternative method for soil heavy metal estimation. A total of 80 soil samples collected in Xuzhou city of China were utilized as data sets for calibration and validation to establish the relationship between the soil reflectance and soil heavy metal content. To amplify the weak spectral characteristic, improve the estimation ability, and explore the characteristic band regions, the preprocessing method of fractional order derivative (FOD) (intervals of 0.25, range of 0–2) and the wavebands selection method of interval partial least squares regression (IPLS) are introduced in this paper. Combining these two methods, for Chromium (Cr), the best estimation model yields Rp2 and RMSRp values of 0.97 and 2.20, respectively, when fractional order is 0.5. This paper explores the potential that FOD conducts the most appropriate order to preprocess spectra and IPLS selects the feature band regions in estimating soil heavy metal of Cr. The results show that FOD and IPLS can strengthen the soil information and improve the accuracy and stability of soil heavy metal estimation effectively.

Use of color parameters in the grouping of soil samples produces more accurate predictions of soil texture and soil organic carbon

Prediction of soil properties such as texture and soil organic carbon (SOC) content by reflectance spectroscopy (RS) is influenced by the heterogeneity of soil samples used to calibrate multivariate models. These soil properties are directly related to color, which, in turn, can be estimated by color parameters derived from the visible (Vis) spectrum at no additional cost. At present, only a few publications have addressed the effect that input data structure and soil heterogeneity have on model performance. Therefore, the objectives of this study were to use Vis-based-color parameters combined with multivariate statistical techniques to group soil samples and comparing the results of different SOC, clay, sand and silt prediction models. Soil sampling was conducted over an area of approximately 500 ha in the region of São Joaquim National Park, Santa Catarina State, Brazil, where a total of 260 soil samples were collected. Soil reflectance data were obtained by Vis-NIR-SWIR spectroscopy in the laboratory, through a spectroradiometer that covers the 350–2500 nm. Soil organic carbon content was determined by dry combustion in an elemental analyzer. Sand, silt and clay fractions were determined using the pipette method. Twenty-two components of color parameters were derived from the Vis spectrum with the use of colorimetry models. For the definition of the most appropriate number of soil sample clusters, two multivariate statistical analyzes: principal component analysis (PCA) and cluster analysis of the samples were applied to the color parameter values. Partial least squares regression (PLSR) and support vector machines (SVM) multivariate models were calibrated for each cluster and also for the models without stratification using 260 soil samples and 95 selected samples (MWS-260 and MWS-95). Overall, the PLSR model performed better than the SVM model, as confirmed by the statistical difference between RMSE results. Multivariate statistical analyzes applied to color parameters were able to group soil samples with similar characteristics, reducing data amplitude and improving the accuracy of soil property predictions. These analyzes demonstrated that the use of Vis-based-color parameters to group soil samples can be a quick and inexpensive way to increase the potential of spectroscopy to accurately predict soil physical and chemical properties.

A new mathematical formulation for remote sensing of soil moisture based on the Red-NIR space

ABSTRACT Optical remote sensing of earth surface processes commonly relies on the red, green, blue (RGB), near-infrared (NIR) and shortwave-infrared (SWIR) electromagnetic bands. Most of the optical sensors mounted on unmanned aerial vehicles and satellites provide the RGB and NIR bands, but only a few offer SWIR output. The Red-NIR reflectance space has been widely applied for remote sensing of various land surface variables including soil moisture. The linear relationship between the Red-NIR reflectance of bare soil is established as the base and then moisture isolines are assumed perpendicular to the soil line. In this study, we show that this assumption is not consistent with the actual Red-NIR space geometry, which in many cases introduces soil moisture estimation errors. To overcome this limitation, we propose a new mathematical transformation to the original Red-NIR space followed by newly defined soil moisture isolines that are more consistent with the actual observations. This new Transformed Red-NIR (TRN) model is compared with the Conventional Red-NIR (CRN) model using data from a sugarcane field located in southwestern Iran. Twelve Land Remote-Sensing Satellite (Landsat)-8 images were acquired during the sugarcane growth season. For validation of the remotely sensed data, ground reference soil moisture was determined at 22 locations at five different depths via core sampling and oven-drying. Our results indicate that the TRN model significantly improves the accuracy of remotely sensed soil moisture.