Digital Soil Mapping Applications Research Articles

A geographically weighted neural network model for digital soil mapping of heavy metal copper in coastal cities

Assessing the spatial distribution of heavy metals in soil is essential for mitigating risks to human health and ensuring the sustainable use of soil resources. This study proposed a geographically weighted neural network (GWNN) model, leveraging deep learning and geographically weighted regression (GWR). The model was designed based on the GWR concept to address the spatial autocorrelation of copper (Cu) in soil and incorporated convolutional neural network (CNN) to capture the nonlinear relationships between Cu and environmental covariates. The GWNN model was compared with GWR, extreme gradient boosting (XGBoost), and artificial neural network (ANN) models. XGBoost was employed to select important environmental covariates and spatial autocorrelation of Cu concentrations was assessed using Moran’s I. The model’s performance was evaluated using 10-fold cross-validation, and prediction uncertainty was quantified with 100 bootstrap models. The results indicated that temperature covariates were the most significant predictors of soil Cu concentrations. The R2 values for Cu prediction accuracy were 0.60 for GWNN, 0.53 for ANN, 0.49 for XGBoost, and 0.32 for GWR. The spatial distribution of Cu showed a trend of higher concentrations in the north and lower concentrations in the south, consistent with spatial clusters identified by local Moran’s I. The mean uncertainty of the 90% confidence interval for GWNN was 16.49%, closely aligning with XGBoost (15.44%) and ANN (16.29%) and significantly outperforming the GWR (18.25%). Overall, the GWNN model demonstrated strong predictive accuracy and low uncertainty, offering an improved approach for digital soil mapping applications.

Digital soil mapping of soil burn severity

AbstractFire alters soil hydrologic properties leading to increased risk of catastrophic debris flows and post‐fire flooding. As a result, US federal agencies map soil burn severity (SBS) via direct soil observation and adjustment of rasters of burned area reflectance. We developed a unique application of digital soil mapping (DSM) to map SBS in the Creek Fire which burned 154,000 ha in the Sierra Nevada. We utilized 169 ground‐based observations of SBS in combination with raster proxies of soil forming factors, pre‐fire fuel conditions, and fire effects to vegetation to build a digital soil mapping model of soil burn severity (DSMSBS) using a random forest algorithm and compared the DSMSBS map to the established SBS map. The DSMSBS model had a cross‐validation accuracy of 48%. The established technique had 46% agreement between field observations and pixels. However, since the established technique is manual, it could not be compared to the DSMSBS model via cross‐validation. We produced SBS class uncertainty maps, which showed high prediction probabilities around field observations, and low probabilities away from field observations. SBS prediction probabilities could aid post‐fire assessment teams with sample prioritization. We report 107 km2 more area classified as high and moderate SBS compared to the established technique. We conclude that blending soil forming factors based mapping and vegetation burn severity mapping can improve SBS mapping. This represents a shift in SBS mapping away from validating remotely sensed reflectance imagery and toward a quantitative soil landscape model, which incorporates both fire and soils information to directly predict SBS.

Evaluation of digital soil mapping projection in soil organic carbon change modeling

There is increasing interest in the application of digital soil mapping (DSM) projections to infer changes in soil carbon across both space and time. This approach relies on the assumption that the spatially modeled soil carbon-environment relationship can be transferred over time. However, this assumption is rarely tested due to a lack of temporally independent validation data. This paper assesses this assumption by developing models of topsoil organic carbon stocks (SOCS) with a deep learning algorithm and data covering mainland China pertaining to the 1980s and 2010s. The temporal prediction performance of models capturing a specific period was assessed by evaluating their performance in the prediction of data during another period. The results revealed that the prediction accuracy of temporal modeling decreased, as indicated by the coefficient of determination, and was lower than that of spatial modeling. The lower prediction accuracy obtained with the DSM-projection approach may result from differences in the magnitudes of influential variables across periods. We found that different levels of environmental similarity and model projection sensitivity to dynamic variables may cause discrepancies in forecast and hindcast accuracy. Our results demonstrate that the prediction error in temporal modeling is related to the degree of environmental similarity between periods. Our findings generally support the implementation of the DSM-projection approach in soil carbon change modeling. However, caution should be exercised, as there exists much uncertainty regarding the projection of spatial models over time.

Towards a cost-effective framework for estimating soil nitrogen pools using pedotransfer functions and machine learning

Globally, the strategic use of nitrogen (N) is important in optimizing economic returns and reducing soil nitrogen losses to the environment. Incorporating reliable estimates of nitrogen (N) mineralized over a growing season (GSN) into N-fertilizer rate prescriptions is critical, but may often lack a direct measurement. For this purpose, Pedotransfer functions (PTFs) of total nitrogen (TN) – representing the stable pool from which N is mineralized and biological nitrogen availability (BNA) – representing the labile pool of N mineralization were used to estimate GSN. GSN was calculated based on TN and BNA results from a soil health database (SHD), which also includes a suite of related soil health parameters (n = 2222). Using a process of recursive feature elimination (RFE) and cost-benefit feature elimination (CBFE), the best predictors of TN, BNA, and GSN were identified using a suite of machine learners (MLs) and regression analysis. For TN, RFE revealed that BNA, active carbon (AC), sand (Sa), and soil organic matter (OM) were the best predictors yielding a Lin’s concordance correlation coefficient (CCC) of 0.80 and a reduction in theoretical cost of 41 % compared to the control. CBFE resulted in AC, soil respiration (SR), clay, Sa, and OM as the most cost-effective predictors of TN with a CCC of 0.79 and a theoretical cost savings 49 % below the cost of using all appropriate soil health parameters in the SHD. With respect to BNA, the best predictors from RFE were aggregate stability (AS), AC, SR, and TN with a CCC of 0.78 and a theoretical cost reduction of 23 %. CBFE retained AC, SR, S, TN, OM and pH as predictors of BNA with a CCC of 0.78 and reduction of 29 % in theoretical cost. Finally, GSN results from RFE identified AS, AC, SR, OM and pH as the best predictors with a 0.82 CCC and 17 % reduction in theoretical cost. CBFE, on the other hand, identified AC, SR, sand, OM, and pH as the most cost-efficient predictors while maintaining a CCC of 0.82 and theoretical cost reduction of 29 %. Of the MLs used for pattern recognition (i.e., cubist, random forest, support vector machine, and stochastic gradient boosting), cubist model outperformed the others for the majority of iterations of the RFE and CBFE processes. The cost-effective framework, and the N-related PTFs developed in this study will greatly enhance our ability to predict of soil N-pool dynamics and the ability to incorporate GSN estimates into N-fertilizer recommendations for producers worldwide. Improvements in predictive strength could be achieved by incorporating climate and soil management practices into PTF development. Another area for improvement and future study would include addition of spatial and landscape variability related to N-measures via digital soil mapping applications.

Spatial prediction of soil sand content at various sampling density based on geostatistical and machine learning algorithms in plain areas

Accurate prediction of the spatial distribution of soil sand content is a pre-requisite for land use management, soil quality evaluation and erosion control, as it determines the transport and movement of soil water, fertilizer, air and heat. Digital soil mapping (DSM) is extensively employed for predicting soil properties. However, practical research is required to address the challenge of selecting an optimal prediction model that is both cost-effective and accurate at a specific sampling density. In this study, topsoil samples were collected from 2,848 sampling points in the eastern plains of China (107,200 km2). The performance of different prediction models for mapping soil sand content was compared at 12 levels of sampling density. Moreover, the geographical detector, a statistical method used to assess the spatial stratified heterogeneity of variables, was adopted to determine the major drivers of spatial variation in soil sand content. The results indicated that climate factors are the major drivers of the spatial variability in soil sand content. For the 100% sample size (26.57 samples/103 km2), the geostatistical models that did not depend on environmental variables (ordinary kriging, sequential Gaussian simulation) performed best, followed by the machine learning models (random forest, cubist and support vector machine) and the geostatistical model with environmental variables (co-kriging). Sampling density had a considerable impact on model accuracy, and the advantages of machine learning models became apparent when sampling densities were below 20% (5.31 samples/103 km2). Therefore, the best combination of prediction model and sampling density should be selected to obtain maps of soil sand content economically and accurately. This study provides a valuable reference for the selection of prediction methods in the practical application of DSM.

A novel local-global dependency deep learning model for soil mapping

The accurate and cost-effective mapping of soil texture is essential for agricultural development and environmental activities. Soil texture exhibits high spatial heterogeneity which poses challenges for recent Digital Soil Mapping (DSM) methods in achieving accurate predictions. Feature engineering methods, extensively used to capture complex soil-forming relationships and enhance prediction accuracy, often involve labor-intensive processes. Additionally, the engineered “discrete” feature cannot reflect interactions between environmental covariates or dependencies. To address the challenges, this study proposes a novel Local-Global Dependency Long Short-Term Memory model (LGD-LSTM) to enhance soil texture predictions at various soil depths. Firstly,a covariate reorganization method hasbeen devised to generate multiple sets of input. Subsequently, several Long Short-Term Memory models (LSTM)have beenemployed to extract the interdependencies among the covariates.Finally, predictions are generated using a fully-connected layer. Cross-validation was conducted within this experiment to analyze prediction accuracy: the average explained variation (R2) ranged from 0.66 to 0.73, and the root mean square error (RMSE) ranged from 6.52% to 10.89%. The results indicated that the LGD-LSTM model offers distinct advantages over other digital soil mapping methods, including Random Forests (RF), Convolutional Neural Network (CNN), and the standard Long Short-Term Memory model (LSTM). In summary, this LGD-LSTM method demonstrates superior performance with relatively high accuracy, ensuring its applicability in effectively representing spatial variations in soil texture. Furthermore, it presents a novel option for DSM applications, enhancing the field's methodology and potential impact.

Mapping soil organic matter content using Sentinel-2 synthetic images at different time intervals in Northeast China

ABSTRACT Mapping soil organic matter (SOM) content has become an important application of digital soil mapping. In this study, we processed all Sentinel-2 images covering the bare-soil period (March to June) in Northeast China from 2019 to 2022 and integrated the observation results into synthetic materials with four defined time intervals (10, 15, 20, and 30 d). Then, we used synthetic images corresponding to different time periods to conduct SOM mapping and determine the optimal time interval and time period before finally assessing the impacts of adding environmental covariates. The results showed the following: (1) in SOM mapping, the highest accuracy was obtained using day-of-year (DOY) 120 to 140 synthetic images with 20 d time intervals, as well as with different time intervals, ranked as follows: 20 d > 30 d > 15 d > 10 d; (2) when using synthetic images at different time intervals to predict SOM, the best time period for predicting SOM was always within May; and (3) adding environmental covariates effectively improved the SOM mapping performance, and the multiyear average temperature was the most important factor. In general, our results demonstrated the valuable potential of SOM mapping using multiyear synthetic imagery, thereby allowing detailed mapping of large areas of cultivated soil.

A case-based method of selecting covariates for digital soil mapping

Selecting a proper set of covariates is one of the most important factors that influence the accuracy of digital soil mapping (DSM). The statistical or machine learning methods for selecting DSM covariates are not available for those situations with limited samples. To solve the problem, this paper proposed a case-based method which could formalize the covariate selection knowledge contained in practical DSM applications. The proposed method trained Random Forest (RF) classifiers with DSM cases extracted from the practical DSM applications and then used the trained classifiers to determine whether each one potential covariate should be used in a new DSM application. In this study, we took topographic covariates as examples of covariates and extracted 191 DSM cases from 56 peer-reviewed journal articles to evaluate the performance of the proposed case-based method by Leave-One-Out cross validation. Compared with a novices’ commonly-used way of selecting DSM covariates, the proposed case-based method improved more than 30% accuracy according to three quantitative evaluation indices (i.e., recall, precision, and F1-score). The proposed method could be also applied to selecting the proper set of covariates for other similar geographical modeling domains, such as landslide susceptibility mapping, and species distribution modeling.

Development of a tool for automatic bare soil detection from multitemporal satellite optical imagery for digital soil mapping applications

Understanding the variability of soil attributes allows to improve the farm production efficiency, accompanied by a reduction in environmental impacts and effective usage of resources. Several studies confirmed the potential of optical remote sensing data for quantifying soil attributes, such as clay content, soil organic carbon and texture classes A challenging issue in spatial-temporal soil surveying by remote sensing data is the limited availability of cloud-free images or affected by cloud/shadow. Further, imagery with high temporal resolution is extremely important for observing terrestrial surfaces. This study investigates the use of multispectral (Sentinel-2 MSI) satellite imagery at the regional/local scale, for the automated detection of agricultural bare soil occurrence, exploiting bands covering the spectral range from visible to shortwave infrared. The study objective is to provide bare soil time series that could be subsequently exploited in digital soil mapping (DSM) approaches based on multispectral or, also in view of the next future missions, hyperspectral remote sensing data.

Pedology and digital soil mapping (DSM)

Pedology focuses on understanding soil genesis in the field and includes soil classification and mapping. Digital soil mapping (DSM) has evolved from traditional soil classification and mapping to the creation and population of spatial soil information systems by using field and laboratory observations coupled with environmental covariates. Pedological knowledge of soil distribution and processes can be useful for digital soil mapping. Conversely, digital soil mapping can bring new insights to pedogenesis, detailed information on vertical and lateral soil variation, and can generate research questions that were not considered in traditional pedology. This review highlights the relevance and synergy of pedology in soil spatial prediction through the expansion of pedological knowledge. We also discuss how DSM can support further advances in pedology through improved representation of spatial soil information. Some major findings of this review are as follows: (a) soil classes can be mapped accurately using DSM, (b) the occurrence and thickness of soil horizons, whole soil profiles and soil parent material can be predicted successfully with DSM techniques, (c) DSM can provide valuable information on pedogenic processes (e.g. addition, removal, transformation and translocation), (d) pedological knowledge can be incorporated into DSM, but DSM can also lead to the discovery of knowledge, and (e) there is the potential to use process‐based soil–landscape evolution modelling in DSM. Based on these findings, the combination of data‐driven and knowledge‐based methods promotes even greater interactions between pedology and DSM.Highlights Demonstrates relevance and synergy of pedology in soil spatial prediction, and links pedology and DSM. Indicates the successful application of DSM in mapping soil classes, profiles, pedological features and processes. Shows how DSM can help in forming new hypotheses and gaining new insights about soil and soil processes. Combination of data‐driven and knowledge‐based methods recommended to promote greater interactions between DSM and pedology.

A preliminary spatial quantification of the soil security dimensions for Tasmania

Soil Security is a holistic soil assessment approach that cogitates soil as a multi-dimensional medium. Rather than traditional single dimensional assessment approaches such as land capability mapping that largely considers only soil and landscape biophysical attributes, the Soil Security concept considers social aspects, education, policy, legislation, current land use, condition and the soils natural and economic value to society. It can identify discrete soils that are currently being used within their capacity, and areas where a use might be unsustainable, i.e. not secure. It would therefore make sense to map this concept, which aligns well with the aspirational and marketing policies of the Tasmanian Government, where increased agricultural expansion through new irrigation schemes and multiple-use State managed production forests co-exists beside pristine World Heritage conservation land, a major drawcard of the economically important tourism industry. The spatial quantification of soil security is seen as an emerging tool to effectively protect the soil resource in terms of food and water security, biodiversity maintenance and safeguarding fragile ecosystems. The recent development and application of Digital Soil Mapping and Assessment capacities in Tasmania to stimulate agricultural production and better target appropriate soil resource use has formed the foundational system that can enable the first efforts in quantifying and mapping Soil Security, in particular the five Soil Security dimensions (Capability, Condition, Capital, Codification and Connectivity). This forms a preliminary mapping product that demonstrates the feasibility of mapping the Soil Security concept. To provide a measure of overall soil security, it was necessary to separately assess the State’s three major soil uses; Agriculture, Conservation and Forestry. These outputs provide an indication of where different activities are sustainable or at risk, where more soil data is needed, and develops a tool to better plan for a State requiring optimal food and fibre production, without depleting its natural soil resources and impacting the fragile ecosystems providing environmental benefits and supporting the tourism industry. The following paper demonstrates why and how we might map Soil Security, describing a preliminary approach to mapping the separate dimensions; this approach could be adapted and applied elsewhere as an evaluation tool to identify soil threats relevant to current land use, biophysical properties, policy and management, and stimulate further research and debate into developing a global Soil Security mapping methodology.

Effects of image pansharpening on soil total nitrogen prediction models in South India

Image fusion is in its infancy in the application of Digital Soil Mapping, and the incorporation of the image pansharpened spectral indices into the soil prediction models has seldom been analyzed. This research performed image pansharpening of Landsat 8, WorldView-2, and Pleiades-1A in a smallholder village called Masuti in South India using three pansharpening techniques: Brovey, Gram-Schmit (GS), and Intensity-Hue-Saturation (IHS) methods. The research analyzed the relationships between multispectral (MS) and pansharpened (PAN) spectral indices and soil total nitrogen (TN), developed the soil TN prediction models using Random Forest methods, and explored the effects of different PAN spectral indices on soil TN prediction models. The results showed the spectral behavior of PAN spectral indices and MS spectral indices were similar. The results also demonstrated that soil TN models based on MS/PAN spectral indices have slightly higher model performance and more detailed characterization of TN spatial pattern compared with soil TN models based on MS spectral indices. Soil TN models based on the GS PAN and MS spectral indices attained slightly higher prediction accuracy compared with those based on other PAN and MS spectral indices. This research advocates the promotion of image pansharpening techniques in digital soil mapping and soil nutrient management research.

Improved estimation of soil clay content by the fusion of remote hyperspectral and proximal geophysical sensing

Planning sustainable soil exploitation and land resource evaluation require up-to-date and accurate maps of soil properties. In that respect, geophysical techniques present particular interests given their non-invasiveness and their fast data acquisition capacity, which permit to characterize large areas with fine spatial and/or temporal resolutions. We investigated the relevancy of combining data from airborne hyperspectral (Hs), electromagnetic induction (EMI) and far-field ground-penetrating radar (GPR) for mapping soil properties, in particular soil clay content, at the field scale. Data from the three techniques were acquired at a test site in Mugello (Italy) characterized by relatively strong spatial variations of soil texture. Soil samples were collected for determining ground truth clay content. For the frequencies used in this study (200–650MHz), the GPR surface reflection is mainly determined by soil dielectric permittivity, itself primarily influenced by soil moisture. In contrast, EMI is mostly sensitive to soil electrical conductivity, which integrates several soil properties including in particular soil moisture and clay content. Taking advantage of the complementary information provided by the two instruments, the GPR and EMI data were combined and correlated to local ground-truth clay content data to provide high-resolution clay content maps over the entire field area. Besides, a relationship was also observed between Hs data and clay content measurements, which permitted to produce a Hs-derived clay content map. EMI–GPR and Hs maps showed close spatial patterns and a relatively high correlation was observed between both clay content estimates, as well as between clay content estimates and ground-truth clay content measurements. Moreover, data fusion allowed constraining the EMI–GPR and Hs information and reduced the uncertainty of mapped clay content estimates. These results demonstrated great promise for integrated, digital soil mapping applications.

Derivation of terrain covariates for digital soil mapping in Australia

Digital soil mapping is founded on the availability of covariates that are used as surrogates for the spatial patterns in soil properties. One important subset of covariates represents the patterns due to terrain, and these are typically derived from a digital elevation model at a suitable resolution. When each digital soil mapping exercise requires the calculation of terrain covariates, there is a clear potential for inconsistent methods and for choosing the covariates that are easiest to derive rather than those that are most relevant. The creation of open repositories of relevant terrain covariates that are correctly derived avoids these problems and fosters the application of digital soil mapping and other modelling activities that benefit from landscape properties. This paper describes the creation of a suite of commonly used terrain covariates from the 1-arcsecond (~30 m) resolution digital elevation models for Australia that were released through CSIRO’s Data Access Portal and the TERN Data Discovery Portal. The methods used to derive the terrain covariates are described and their characteristics are identified.

Using model averaging to combine soil property rasters from legacy soil maps and from point data

The objective of this study was to determine the efficacy of model averaging (ensemble modelling) as an approach for combining digital soil property maps derived from disaggregated legacy soil class maps and scorpan kriging (using soil point data). The study is based in the Dalrymple Shire, QLD and continues on the soil pH mapping work of Odgers et al. (2014a). Equal weights averaging (EW), Bates–Granger or variance weighted averaging (VW), Granger–Ramanathan averaging (GRA), and Bayesian model averaging (BMA) were compared in this study. Model averaged predictions were estimated to 2m depth at regular depth intervals. 90% prediction intervals of the model averaged predictions were derived numerically. Neither the disaggregated soil map nor the scorpan kriging map was particularly accurate. Predictions from model averaging however did improve upon the accuracy, where at all depths, the combined predictions were an improvement on using either of the contributing soil maps alone. We recommend the use of GRA for digital soil mapping applications because its performance is equal to or better than the generally preferred BMA approach, yet far simpler to implement, and is computationally efficient. For regional soil studies where polygon mapping and soil point data are available, ensemble modelling is a useful combinatorial approach.

Building soil spectral library of the Czech soils for quantitative digital soil mapping

  Spectral libraries are the data archives of spectral signatures measured on natural and/or man-made materials. Here, the objective is to build a soil spectral library of the Czech soils (SSL-CZ). Further on, the overall aim is to apply diffuse reflectance spectroscopy as a tool for digital soil mapping. An inevitable part of the library is a metadata database that stores the corresponding auxiliary information on the soils: type of material (soil, parent material), sample preparation, location of the sample with geographic coordinates, soil classification, morphological features, soil laboratory measurements – chemical, physical, and potential biological properties, geophysical features of and climatological information on the sample location. The metadata database consists of seven general tables (General, Spatial, Soil class, Environmental, Auxiliary, Analytical and Spectra) relationally linked together. The stored information allows for a wide range of analyses and for modelling developments of digital soil mapping applications. An example of partial least-square regression (PLSR) modelling for soil pH and clay content with 0.84 and 0.68 coefficients of determination is provided on the subset of the collected data. Currently, the SSL-CZ database contains more than 500 records in the first phase of development. Spectral reflectance signatures are stored in the range of 350 to 2500 nm with a step of 1 nm measured by ASD FieldSpec 3. The soil spectral library developed is fully compatible with Global Soil Spectral Library (Soil Spectroscopy Group).

Using Canonical Correspondence Analysis (CCA) to identify the most important DEM attributes for digital soil mapping applications

Abstract Topography has an important influence on the distribution of soils and their properties, especially in hilly lands, and related data are easily available, measurable and recognizable from digital elevation models (DEMs). To our knowledge, little attention has previously been paid to the effect of DEM attributes on the distribution of soils, using ordination methods. The objective of this study was to analyze relationships between topographical properties derived from DEM and soil distribution and to discuss their applicability in Digital Soil Mapping (DSM). The study was carried out in the Borujen area of central Zagros, Iran. A total of 13 plots (each one of 6.75 ha) were set up to calculate the percentages of the dominant soil series. Fifteen DEM attributes, including slope, aspect, curvature, maximum and minimum curvature, planform curvature, profile curvature, tangent curvature, wetness index, power index, sediment index, area solar radiation, direct radiation, diffuse radiation and direct duration were also computed. Canonical Correspondence Analysis (CCA) was used to summarize the data set and to evaluate the expected relationships. The results obtained show that there was a relatively strong correspondence between soils' series distribution and topographical properties. The DEM attributes that best related to the first axis were maximum curvature, slope and sediment index, all of which significantly positive correlated, and wetness index, direct duration and minimum curvature, all of which were negatively related. The second axis showed a negative trend with wetness index, direct duration and aspect, and a positive trend with sediment index and slope. These gradients were closely related to the first three canonical axes and explained 71.8% of the total variance of the soil series. The residual variance (28.2% of the total variance) was related to other soil forming factors, like parent material and vegetation cover, which were not investigated in this study. Considering that DEMs are still the most important source of environmental information, understanding the role of topographical factors in a region should help us to identify soils and their properties better and enable us to apply these derivates as input data in DSM.

Regressões Logísticas Múltiplas: fatores que influenciam sua aplicação na predição de classes de solos

Métodos mais eficazes para determinação do padrão de distribuição de classes de solo na paisagem precisam ser avaliados visando suprir a demanda por mapas de solo em escalas regional e global. Neste estudo, Regressões Logísticas Múltiplas foram utilizadas como modelos preditores em uma aplicação de Mapeamento Digital de Solos. Os modelos foram gerados utilizando um mapa de solos existente como variável dependente e atributos de terreno como variáveis independentes, o que possibilitou determinar a probabilidade de encontrar classes de solo na paisagem no primeiro e no segundo nível categórico do SiBCS. A qualidade dos mapas preditos foi verificada por meio da matriz de contingência. A classe dos Argissolos foi predita corretamente, em relação ao mapa original, em aproximadamente 85 %. As classes de solos hidromórficos (Planossolos e Gleissolos) foram preditas corretamente em 75 %. Houve confundimento dos modelos para as classes que ocupam posições muito semelhantes na paisagem. Foi verificado também que classes de solo pouco representativas na paisagem não são adequadamente espacializadas em razão da sensibilidade dos modelos logísticos à proporção relativa das amostras usadas para treinar os modelos.

Generation of kth-order random toposequences

The model presented in this paper derives toposequences from a digital elevation model (DEM). It is written in ArcInfo Macro Language (AML). The toposequences are called kth-order random toposequences, because they take a random path uphill to the top of a hill and downhill to a stream or valley bottom from a randomly selected seed point, and they are located in a streamshed of order k according to a particular stream-ordering system. We define a kth-order streamshed as the area of land that drains directly to a stream segment of stream order k. The model attempts to optimise the spatial configuration of a set of derived toposequences iteratively by using simulated annealing to maximise the total sum of distances between each toposequence hilltop in the set.The user is able to select the order, k, of the derived toposequences. Toposequences are useful for determining soil sampling locations for use in collecting soil data for digital soil mapping applications. Sampling locations can be allocated according to equal elevation or equal-distance intervals along the length of the toposequence, for example.We demonstrate the use of this model for a study area in the Hunter Valley of New South Wales, Australia. Of the 64 toposequences derived, 32 were first-order random toposequences according to Strahler's stream-ordering system, and 32 were second-order random toposequences.The model that we present in this paper is an efficient method for sampling soil along soil toposequences. The soils along a toposequence are related to each other by the topography they are found in, so soil data collected by this method is useful for establishing soil–landscape rules for the preparation of digital soil maps.