Luiz Eduardo Vicente

Prediction of soil properties using imaging spectroscopy: Considering fractional vegetation cover to improve accuracy

Spectroscopic techniques have become attractive to assess soil properties because they are fast, require little labor and may reduce the amount of laboratory waste produced when compared to conventional methods. Imaging spectroscopy (IS) can have further advantages compared to laboratory or field proximal spectroscopic approaches such as providing spatially continuous information with a high density. However, the accuracy of IS derived predictions decreases when the spectral mixture of soil with other targets occurs. This paper evaluates the use of spectral data obtained by an airborne hyperspectral sensor (ProSpecTIR-VS – Aisa dual sensor) for prediction of physical and chemical properties of Brazilian highly weathered soils (i.e., Oxisols). A methodology to assess the soil spectral mixture is adapted and a progressive spectral dataset selection procedure, based on bare soil fractional cover, is proposed and tested. Satisfactory performances are obtained specially for the quantification of clay, sand and CEC using airborne sensor data (R2 of 0.77, 0.79 and 0.54; RPD of 2.14, 2.22 and 1.50, respectively), after spectral data selection is performed; although results obtained for laboratory data are more accurate (R2 of 0.92, 0.85 and 0.75; RPD of 3.52, 2.62 and 2.04, for clay, sand and CEC, respectively). Most importantly, predictions based on airborne-derived spectra for which the bare soil fractional cover is not taken into account show considerable lower accuracy, for example for clay, sand and CEC (RPD of 1.52, 1.64 and 1.16, respectively). Therefore, hyperspectral remotely sensed data can be used to predict topsoil properties of highly weathered soils, although spectral mixture of bare soil with vegetation must be considered in order to achieve an improved prediction accuracy.

Abordagens semiquantitativa e quantitativa na avaliação da textura do solo por espectroscopia de reflectância bidirecional no VIS‑NIR‑SWIR

The objective of this work was to evaluate the potential of VIS‑NIR‑SWIR reflectance spectroscopy for the characterization of soil particle‑size distribution of samples from different textural classes, and to obtain models to predict clay, silt, and sand contents in the soil. A representative sample set of Oxisols and Ultisols from five locations in Mato Grosso do Sul state, Brazil, were used. Visible and near‑infrared to short‑wave infrared (from 350 to 2,500 nm) spectra of the samples were obtained and analyzed. Principal component analysis (PCA), fuzzy c‑means cluster analysis, multinomial logistic regression (MLR), and partial least squares regression were used. Characteristic spectra for the different soil texture classes and segregation of samples from texture classes and from sampling sites with distinct characteristics, through PCA, fuzzy c‑means, and RLM, show the semiquantitative potential of the VIS‑NIR‑SWIR reflectance data. Satisfactory quantification was obtained for clay (R²=0.92, RPD=3.59), silt (R²=0.80, RPD=2.15), and sand (R²=0.87, RPD=2.62). The reflectance spectroscopy techniques can help to assess soil texture and soil spacial variability with semiquantitative or quantitative methodologies.

Contributions of the complexity paradigm to the understanding of Cerrado's organization and dynamics

The Brazilian Cerrado is a vegetation mosaic composed of different physiognomies. Discussions remain open regarding the factors and processes responsible for the dynamic and spatial organization of the Cerrado - in its different physiognomies. The contributions of the complexity paradigm in this context are still less exploited, despite its great potential for explanations and predictions presented in previous diverse dynamic systems of complex behavior researches, a category in which the Cerrado can be included. This article has the intention of contributing to the construction of this new perspective, discussing - from theoretical concepts - the paradigm of complexity for the understanding of the organization and the dynamics of the Cerrado.

A Cluster-Based Approach to Support the Delineation of Management Zones in Precision Agriculture

In this paper we propose a cluster-based approach for the delineation of management zones in precision agriculture. The proposed approach was built following the steps of data mining for the clustering task, resulting in a computer application that generates maps of management zones and yield areas, allowing to compare them using known statistical indexes. The basis for this implementation was a model previously published in the literature that uses only historical productivity, soil electrical conductivity and relief data to generate the maps. The main difference of our work with respect to the previous model is the clustering algorithms used in the step of extracting patterns. While the original model uses only the fuzzy c-means algorithm, the model developed in this study uses the GKCluster extension to this algorithm, able to detect clusters with different geometrical shapes. From the tests performed with the new proposed model, we achieved about 76% of correlation between maps of yield and management zones from kappa index, and about 85% of correlation from overall accuracy. The original model reached, according to the authors, a maximum correlation of 49% from kappa index, and 70% from overall accuracy.

Improvement of Clay and Sand Quantification Based on a Novel Approach with a Focus on Multispectral Satellite Images

Soil mapping demands large-scale surveys that are costly and time consuming. It is necessary to identify strategies with reduced costs to obtain detailed information for soil mapping. We aimed to compare multispectral satellite image and relief parameters for the quantification and mapping of clay and sand contents. The Temporal Synthetic Spectral (TESS) reflectance and Synthetic Soil Image (SYSI) approaches were used to identify and characterize texture spectral signatures at the image level. Soil samples were collected (0-20 cm depth, 919 points) from an area of 14,614 km2 in Brazil for reference and model calibration. We compared different prediction approaches: (a) TESS and SYSI; (b) Relief-Derived Covariates (RDC); and (c) SYSI plus RDC. The TESS method produced highly similar behavior to the laboratory convolved data. The sandy textural class showed a greater increase in average spectral reflectance from Band 1 to 7 compared with the clayey class. The prediction using SYSI produced a better result for clay (R2 = 0.83; RMSE = 65.0 g kg-1) and sand (R2 = 0.86; RMSE = 79.9 g kg-1). Multispectral satellite images were more stable for the identification of soil properties than relief parameters.

Mapping fractional landscape soils and vegetation components from Hyperion satellite imagery using an unsupervised machine-learning workflow

ABSTRACT An unsupervised machine-learning workflow is proposed for estimating fractional landscape soils and vegetation components from remotely sensed hyperspectral imagery. The workflow is applied to EO-1 Hyperion satellite imagery collected near Ibirací, Minas Gerais, Brazil. The proposed workflow includes subset feature selection, learning, and estimation algorithms. Network training with landscape feature class realizations provide a hypersurface from which to estimate mixtures of soil (e.g. 0.5 exceedance for pixels: 75% clay-rich Nitisols, 15% iron-rich Latosols, and 1% quartz-rich Arenosols) and vegetation (e.g. 0.5 exceedance for pixels: 4% Aspen-like trees, 7% Blackberry-like trees, 0% live grass, and 2% dead grass). The process correctly maps forests and iron-rich Latosols as being coincident with existing drainages, and correctly classifies the clay-rich Nitisols and grasses on the intervening hills. These classifications are independently corroborated visually (Google Earth) and quantitatively (random soil samples and crossplots of field spectra). Some mapping challenges are the underestimation of forest fractions and overestimation of soil fractions where steep valley shadows exist, and the under representation of classified grass in some dry areas of the Hyperion image. These preliminary results provide impetus for future hyperspectral studies involving airborne and satellite sensors with higher signal-to-noise and smaller footprints.

Monitoramento e gerenciamento de bacias urbanas associados a inundação: diagnose da bacia do Ribeirão Quilombo na região metropolitana de Campinas utilizando geotecnologias

Lack of data in appropriates spatial and temporal scales is a great obstacle to implantation of flood alert systems. The present article results of an exercise that aim to verify data currently available for Metropolitan Region of Campinas and, consequently, which ones would still be necessary to create a system of this type. Quilombo river basin was chosen as a pilot-area and a diagnosis of its physical and social-economical characteristics was made, taking pre-existing data and correlating then with SIG. The results obtained by this diagnosis indicate the necessity of a high-resolution mapping and of a broader and refined monitoring in comparison of those available at present. Based on correlations between spatial data, and objectifying to enhance precision of data collection in semi-detail scale, adequate points to install new fluvio and pluviometrical stations are proposed.