Mapping soils in two watersheds using legacy data and extrapolation for similar surrounding areas

Pelegrino, Marcelo Henrique Procópio; Silva, Sérgio Henrique Godinho; Menezes, Michele Duarte de; Silva, Elidiane da; Owens, Phillip R.; Curi, Nilton

doi:10.1590/1413-70542016405011416

Cited by 26 publications

(13 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The study areas varied from 1.75 km 2 (175 ha) (Pelegrino et al, 2016) to approximately 120,000 km 2 , which represents 48 % of the state of São Paulo (Silvero et al, 2019). Approximately 50 % of the studies were conducted in areas up to 120 km 2 .…”

Section: Descriptive Statistics Of the Data Extracted From The Studiesmentioning

confidence: 99%

Digital soil class mapping in Brazil: a systematic review

Coelho

Giasson

Campos

et al. 2021

Sci. agric. (Piracicaba, Braz.)

View full text Add to dashboard Cite

In Brazil several digital soil class mapping studies were carried out from 2006 onwards to maximize the use of existing maps and information and to provide estimates for wider areas. However, there is no consensus on which methods have produced superior results in the predictive value of soil maps. This study conducts a systematic review of digital soil class mapping in Brazil and aims to analyze the factors which can improve the accuracy of digital soil class maps. Data from 334 digital soil class mapping studies were grouped and analyzed by Student's t-test, Wilcoxon-Mann-Whitney test and Kruskal-Wallis test. When conventional maps were used for validation, the studies showed average values of 63 % and when field samples were used, 56 % for Overall Accuracy. Studies compatible with the Planimetric Cartographic Accuracy Standard for Digital Cartographic Products (PEC-PCD) averaged between 4 % and 15 % higher accuracy than those of the incompatible group. There seems to be no evidence that increasing the number of variables and samples results in more accurate soil map prediction, but studies using variables related to four soil-forming factors enhanced accuracy. From a density of 0.08 MU km-2 and upwards, it became more difficult for studies to obtain greater accuracy. Artificial neural network classifiers and Decision Tree models seem to be producing more accurate digital soil class maps.

show abstract

Section: Descriptive Statistics Of the Data Extracted From The Studiesmentioning

confidence: 99%

Digital soil class mapping in Brazil: a systematic review

Coelho

Giasson

Campos

et al. 2021

Sci. agric. (Piracicaba, Braz.)

View full text Add to dashboard Cite

show abstract

“…A larger training dataset is often necessary in order to reduce error (Pal and Mather, 2003), and in this study, such information also brought stability to model errors in training data above 105 observations. However, it is important to highlight that the use of polygons and buffers could bring some uncertainty as regards the soil type, mainly closer to the boundaries or transition zones (Pelegrino et al, 2016;ten Caten et al, 2012;Giasson et al, 2015). Thus, the key point here is will the more accurate models deliver accurate soil maps in the digitally mapped area?…”

Section: Model Evaluationmentioning

confidence: 99%

“…The traditional way of mapping soils to deliver precious maps is very time-consuming and onerous (Kempen et al, 2012). Soil legacy could be a source of training data in machine-learning techniques (Pelegrino et al, 2016), which could formalize soil-landscape relationships, apply the information to areas under similar environmental conditions, and enhance the mapping of areas and result in savings in both time and cost (Silva et al, 2016). This is an important strategy for mapping in Brazil, due to the restriction of detailed soil surveys to small areas (Mendonça-Santos and Santos, 2007).…”

Section: Introductionmentioning

confidence: 99%

Soil type spatial prediction from Random Forest: different training datasets, transferability, accuracy and uncertainty assessment

Machado

Silva

Curi

et al. 2019

Sci. agric. (Piracicaba, Braz.)

Self Cite

View full text Add to dashboard Cite

Different uses of soil legacy data such as training dataset as well as the selection of soil environmental covariables could drive the accuracy of machine learning techniques. Thus, this study evaluated the ability of the Random Forest algorithm to predict soil classes from different training datasets and extrapolate such information to a similar area. The following training datasets were extracted from legacy data: a) point data composed of 53 soil samples; b) 30 m buffer around the soil samples, and soil map polygons excluding: c) 20 m; and d) 30 m from the boundaries of polygons. These four datasets were submitted to principal component analysis (PCA) to reduce multidimensionality. Each dataset derived a new one. Different combinations of predictor variables were tested. A total of 52 models were evaluated by means of error of models, prediction uncertainty and external validation for overall accuracy and Kappa index. The best result was obtained by reducing the number of predictors with the PCA along with information from the buffer around the points. Although Random Forest has been considered a robust spatial predictor model, it was clear it is sensitive to different strategies of selecting training dataset. Effort was necessary to find the best training dataset for achieving a suitable level of accuracy of spatial prediction. To identify a specific dataset seems to be better than using a great number of variables or a large volume of training data. The efforts made allowed for the accurate acquisition of a mapped area 15.5 times larger than the reference area.

show abstract

“…De modo geral, a reprodutibilidade do mapa convencional de solos (Tabela 4) com os dados coletados apenas nos pixels dos perfis de solos alcançou valores próximos ao valor médio de 60% observado na literatura em estudos com outras estratégias de amostragem utilizada no mapeamento digital de solos (Höfig et al, 2014;Giasson et al, 2015;Bagatini et al, 2016;Dias et al, 2016;Pelegrino et al, 2016). Vale destacar que em ambas as áreas de estudo apenas duas unidades representam mais de 75% da composição dos respectivos mapas de solos, sendo estas as mais importantes na predição e as quais apresentaram maior concordância com os mapas legados utilizados na validação.…”

Section: /9unclassified

Índice de área foliar residual como estratégia para manejo de pasto: estrutura do capim Tanzânia

Lemos

Ruggieri

Silva

et al. 2019

Agraria

View full text Add to dashboard Cite

AGRONOMIA (AGRONOMY) RESUMO: Os estudos de mapeamento digital de classes de solos (MDS) têm utilizado mapas legados como principal fonte de informação para calibração dos modelos preditores. Entretanto, são necessárias novas abordagem com técnicas que permitam o uso de informações contidas em perfis de solos georreferenciados, permitindo a aplicação do MDS em áreas amostradas que não disponibilizem de mapas convencionais de solos. O objetivo deste estudo foi avaliar o desempenho na predição de ocorrência de solos de amostras coletadas em pixels de perfis de solos georreferenciados e em pixels coletados em buffers com raio de 50, 100, 150, 200 e 250 m dos perfis de solos nas bacias dos rios Lajeado Grande e Santo Cristo. As duas áreas possuem dados de levantamento de solos publicados em escala 1:50.000. Para predição da ocorrência das classes de solos foram utilizadas dez variáveis preditoras geradas a partir de um modelo digital de elevação com resolução espacial de 30 m. Para a predição foi utilizado o algoritmo Random Forest. Os mapas preditos foram avaliados quanto à exatidão em relação aos perfis de solos e quanto a reprodutibilidade dos mapas convencionais. A utilização dos pixels amostrais coletados nos buffers não alterou de forma expressiva a acurácia geral dos mapas preditos na bacia do rio Lajeado Grande, mas permitiu um ganho de 15,6% de exatidão na bacia do rio Santo Cristo. O uso apenas dos perfis georreferenciados resultou em mapas preditos com exatidão superior a 75% e concordância de reprodutibilidade igual ou superior a 67% em relação ao mapa convencional.

show abstract

Mapping soils in two watersheds using legacy data and extrapolation for similar surrounding areas

Cited by 26 publications

References 20 publications

Digital soil class mapping in Brazil: a systematic review

Digital soil class mapping in Brazil: a systematic review

Soil type spatial prediction from Random Forest: different training datasets, transferability, accuracy and uncertainty assessment

Índice de área foliar residual como estratégia para manejo de pasto: estrutura do capim Tanzânia

Contact Info

Product

Resources

About