Digital mapping of soil properties in Canadian managed forests at 250m of resolution using the k-nearest neighbor method

Mansuy, Nicolas; Thiffault, Évelyne; Paré, David; Bernier, Pierre Y.; Guindon, Luc; Villemaire, Philippe; Poirier, Vincent; Beaudoin, A.

doi:10.1016/j.geoderma.2014.06.032

Cited by 112 publications

(65 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Kidd et al, 2014b) and are generally superior to those reported by some others (e.g. Mansuy et al, 2014). While a numerical comparison of R-squared values and classification rates has no formal statistical meaning, the models of categorical attributes generally did better than the models of numerical attributes.…”

Section: Digital Soil Mappingmentioning

confidence: 51%

See 1 more Smart Citation

Digital soil assessment for regional agricultural land evaluation

Harms

Brough

Philip

et al. 2015

Global Food Security

View full text Add to dashboard Cite

Section: Digital Soil Mappingmentioning

confidence: 51%

“…DSM has become the preferred method for many soil mapping applications (e.g. Mansuy et al, 2014;Gooley et al, 2014) and is the technique being adopted to deliver a soil map of the world (Arrouays et al, 2014).…”

Section: Introductionmentioning

confidence: 99%

Digital soil assessment for regional agricultural land evaluation

Harms

Brough

Philip

et al. 2015

Global Food Security

View full text Add to dashboard Cite

“…Soil texture estimation using DSM typically shows accuracies less than R 2 = 0.5, while studies with R 2 > 0.7 are rare ( Malone et al, 2009; Lacoste et al, 2011; Wang et al, 2012; de Carvalho Junior et al, 2014; Mansuy et al, 2014). Our results (Table 5) showed prediction accuracies with R 2 > 0.5 for all sand fractions and for both calibration sets (cLHS adapt ; cLHS adapt+ ).…”

Section: Discussionmentioning

confidence: 99%

“…The coupling is accomplished by regression and classification approaches resulting in continuous or discrete maps of soil properties ( McBratney et al, 2003; Scull et al, 2003; McMillan , 2008). DSM presents an established framework for soil mapping and has been successfully applied in numerous studies, addressing various soil properties, landscapes, and scales ( Florinsky et al, 2002; Behrens et al, 2005; Mora‐Vallejo et al, 2008; Lacoste et al, 2011; Wang et al, 2012; Behrens et al, 2014; de Carvalho Junior et al, 2014; Mansuy et al, 2014; Taghizadeh‐Mehrjardi et al, 2014; Thomas et al, 2015).…”

Section: Introductionmentioning

confidence: 99%

Incorporating limited field operability and legacy soil samples in a hypercube sampling design for digital soil mapping

Stumpf

Schmidt

Behrens

et al. 2016

J. Plant Nutr. Soil Sci.

View full text Add to dashboard Cite

Most calibration sampling designs for Digital Soil Mapping (DSM) demarcate spatially distinct sample sites. In practical applications major challenges are often limited field accessibility and the question on how to integrate legacy soil samples to cope with usually scarce resources for field sampling and laboratory analysis. The study focuses on the development and application of an efficiency improved DSM sampling design that (1) applies an optimized sample set size, (2) compensates for limited field accessibility, and (3) enables the integration of legacy soil samples. The proposed sampling design represents a modification of conditioned Latin Hypercube Sampling (cLHS), which originally returns distinct sample sites to optimally cover a soil related covariate space and to preserve the correlation of the covariates in the sample set. The sample set size was determined by comparing multiple sample set sizes of original cLHS sets according to their representation of the covariate space. Limited field accessibility and the integration of legacy samples were incorporated by providing alternative sample sites to replace the original cLHS sites. We applied the modified cLHS design (cLHS adapt ) in a small catchment (4.2 km 2 ) in Central China to model topsoil sand fractions using Random Forest regression (RF). For evaluating the proposed approach, we compared cLHS adapt with the original cLHS design (cLHS orig ). With an optimized sample set size n = 30, the results show a similar representation of the cLHS covariate space between cLHS adapt and cLHS orig , while the correlation between the covariates is preserved (r = 0.40 vs. r = 0.39). Furthermore, we doubled the sample set size of cLHS adapt by adding available legacy samples (cLHS adapt+ ) and compared the prediction accuracies. Based on an external validation set cLHS val (n = 20), the coefficient of determination (R 2 ) of the cLHS adapt predictions range between 0.59 and 0.71 for topsoil sand fractions. The R 2 -values of the RF predictions based on cLHS adapt+ , using additional legacy samples, are marginally increased on average by 5%.

show abstract

“…With the development of earth observation technologies and spatial analysis method, many new useful covariates are emerging (Nield et al, 2007;Zhu et al (2010b); Liu et al, 2012;Qin et al, 2012). Statistical methods (such as principal components analysis, stepwise regression) could help us select appropriate covariates (Pechenizkiy et al, 2003;Behrens et al, 2010;Samuel-Rose et al, 2015;Ramadan et al, 2001;Mansuy et al, 2014;Beaudoin et al, 2014). On the other hand, covariates could also be selected based on expert knowledge or related previous studies in the same study area.…”

Section: Determination Of Prediction Uncertaintymentioning

confidence: 99%

An heuristic uncertainty directed field sampling design for digital soil mapping

et al. 2016

View full text Add to dashboard Cite

Legacy samples are a valuable data source for digital soil mapping. However, these sample sets are often small in size and ad hoc in spatial distribution. Constrained by the limited representativeness of such a sample set, the obtained soil maps are often incomplete in spatial coverage with "gaps" at the locations which cannot be well represented by these samples. The maps may also contain areas of high prediction uncertainty. In order to extend the predicted area and reduce prediction uncertainty, additional samples are needed. This paper presents a sampling design based on prediction uncertainty to select samples which will effectively complement the sparse and ad hoc samples, and maximize the spatial coverage of prediction and minimize prediction uncertainty. A case study in China shows that this sampling scheme was effective in achieving these goals. Compared with stratified random sampling scheme, when the number of additional samples is the same, the produced map using uncertainty directed samples has larger predicted area, and the accuracy of the produced map is higher than that of the maps using stratified random samples. The finding of this study suggests that prediction uncertainty is a useful indicator to aid field sample selection and to complement the legacy data. Furthermore, the mapping accuracy produced using this method can be quantitatively related to the number of additional samples needed which opens a new horizon for digital soil mapping.

show abstract

Digital mapping of soil properties in Canadian managed forests at 250m of resolution using the k-nearest neighbor method

Cited by 112 publications

References 45 publications

Digital soil assessment for regional agricultural land evaluation

Digital soil assessment for regional agricultural land evaluation

Incorporating limited field operability and legacy soil samples in a hypercube sampling design for digital soil mapping

An heuristic uncertainty directed field sampling design for digital soil mapping

Contact Info

Product

Resources

About