20.
Glenn J. Newnham Neil C. Sims, Jacqueline R. England; Wheeler, Ichsani
Good Practice Guidance. SDG Indicator 15.3.1, Proportion of Land That Is Degraded Over Total Land Area Technical Report
2021.
Abstract | Links | BibTeX | Tags: Land Degradation, Sustainable Development Goals
@techreport{Sims2021,
title = {Good Practice Guidance. SDG Indicator 15.3.1, Proportion of Land That Is Degraded Over Total Land Area},
author = {Neil C. Sims, Glenn J. Newnham, Jacqueline R. England, Juan Guerschman, Simon J. D. Cox, Stephen H. Roxburgh, Raphael A. Viscarra Rossel, Steffen Fritz and Ichsani Wheeler},
url = {https://www.unccd.int/sites/default/files/documents/2021-09/UNCCD_GPG_SDG-Indicator-15.3.1_version2_2021.pdf},
year = {2021},
date = {2021-09-29},
urldate = {2021-09-29},
abstract = {This Good Practice Guidance (GPG) document provides guidance on how to calculate the extent of land degradation for reporting on United Nations (UN) Sustainable Development Goal (SDG) Indicator 15.3.1: the proportion of land that is degraded over total land area. This guidance supports implementation of the Tier I methods for Indicator 15.3.1 adopted by the UN Statistical Commission, and the
development of analytical methods for measuring its three sub-indicators, which are:
1. Trends in land cover
2. Trends in land productivity
3. Trends in carbon stocks (above and below ground), which is currently represented by soil organic carbon (SOC) stocks.
The Indicator is calculated by integrating the subindicators using a one-out-all-out (1OAO) method, in which a significant reduction or negative change in any one of the three sub-indicators is considered to comprise land degradation. Significant reductions can be identified using statistical criteria, or by a qualitative assessment of the magnitude of change. The Indicator is reported as a binary quantification (i.e., degraded/not degraded) of the extent of degraded land in hectares, and expressed as the proportion (percentage) of land that is degraded over total land area. Version 2 of the GPG incorporates a number of advances in the quality and availability of datasets, as well as analytical methods for calculating Indicator 15.3.1 and its sub-indicators that have emerged since publication of Version 1 of the GPG (Sims et al. 2017). These advances have been identified through research, stakeholder engagement, including analysis of recommendations from countries reporting on Indicator 15.3.1 in the first reporting period in 2018, and reviews of drafts of this report by global experts in relevant fields. This report also incorporates new developments from the growing number of publications and initiatives focused on improving the quality and availability of data and analytics for the SDGs in general, and Indicator 15.3.1 and Land Degradation Neutrality (LDN) in particular. While the measurements of Indicator 15.3.1 form only one part of the assessment of LDN, the relationship between the Indicator and LDN is discussed in more detail in Version 2 (this document), and additional guidance on the calculation and use of measurements during the baseline period should clarify the interpretation of Indicator 15.3.1 for monitoring LDN. The guidance covers three sub-indicators. The Land Cover sub-indicator reports degradation in land cover change based on a national assessment of the positive or negative aspects of transitions from one land cover type to another. In Version 2, this chapter highlights the need for countries to first consider the main drivers of land cover change, and then to determine which transitions to identify as degraded for reporting on Indicator 15.3.1. This chapter incorporates updates to the quality of available land cover datasets and classification methods.},
keywords = {Land Degradation, Sustainable Development Goals},
pubstate = {published},
tppubtype = {techreport}
}
This Good Practice Guidance (GPG) document provides guidance on how to calculate the extent of land degradation for reporting on United Nations (UN) Sustainable Development Goal (SDG) Indicator 15.3.1: the proportion of land that is degraded over total land area. This guidance supports implementation of the Tier I methods for Indicator 15.3.1 adopted by the UN Statistical Commission, and the
development of analytical methods for measuring its three sub-indicators, which are:
1. Trends in land cover
2. Trends in land productivity
3. Trends in carbon stocks (above and below ground), which is currently represented by soil organic carbon (SOC) stocks.
The Indicator is calculated by integrating the subindicators using a one-out-all-out (1OAO) method, in which a significant reduction or negative change in any one of the three sub-indicators is considered to comprise land degradation. Significant reductions can be identified using statistical criteria, or by a qualitative assessment of the magnitude of change. The Indicator is reported as a binary quantification (i.e., degraded/not degraded) of the extent of degraded land in hectares, and expressed as the proportion (percentage) of land that is degraded over total land area. Version 2 of the GPG incorporates a number of advances in the quality and availability of datasets, as well as analytical methods for calculating Indicator 15.3.1 and its sub-indicators that have emerged since publication of Version 1 of the GPG (Sims et al. 2017). These advances have been identified through research, stakeholder engagement, including analysis of recommendations from countries reporting on Indicator 15.3.1 in the first reporting period in 2018, and reviews of drafts of this report by global experts in relevant fields. This report also incorporates new developments from the growing number of publications and initiatives focused on improving the quality and availability of data and analytics for the SDGs in general, and Indicator 15.3.1 and Land Degradation Neutrality (LDN) in particular. While the measurements of Indicator 15.3.1 form only one part of the assessment of LDN, the relationship between the Indicator and LDN is discussed in more detail in Version 2 (this document), and additional guidance on the calculation and use of measurements during the baseline period should clarify the interpretation of Indicator 15.3.1 for monitoring LDN. The guidance covers three sub-indicators. The Land Cover sub-indicator reports degradation in land cover change based on a national assessment of the positive or negative aspects of transitions from one land cover type to another. In Version 2, this chapter highlights the need for countries to first consider the main drivers of land cover change, and then to determine which transitions to identify as degraded for reporting on Indicator 15.3.1. This chapter incorporates updates to the quality of available land cover datasets and classification methods.
development of analytical methods for measuring its three sub-indicators, which are:
1. Trends in land cover
2. Trends in land productivity
3. Trends in carbon stocks (above and below ground), which is currently represented by soil organic carbon (SOC) stocks.
The Indicator is calculated by integrating the subindicators using a one-out-all-out (1OAO) method, in which a significant reduction or negative change in any one of the three sub-indicators is considered to comprise land degradation. Significant reductions can be identified using statistical criteria, or by a qualitative assessment of the magnitude of change. The Indicator is reported as a binary quantification (i.e., degraded/not degraded) of the extent of degraded land in hectares, and expressed as the proportion (percentage) of land that is degraded over total land area. Version 2 of the GPG incorporates a number of advances in the quality and availability of datasets, as well as analytical methods for calculating Indicator 15.3.1 and its sub-indicators that have emerged since publication of Version 1 of the GPG (Sims et al. 2017). These advances have been identified through research, stakeholder engagement, including analysis of recommendations from countries reporting on Indicator 15.3.1 in the first reporting period in 2018, and reviews of drafts of this report by global experts in relevant fields. This report also incorporates new developments from the growing number of publications and initiatives focused on improving the quality and availability of data and analytics for the SDGs in general, and Indicator 15.3.1 and Land Degradation Neutrality (LDN) in particular. While the measurements of Indicator 15.3.1 form only one part of the assessment of LDN, the relationship between the Indicator and LDN is discussed in more detail in Version 2 (this document), and additional guidance on the calculation and use of measurements during the baseline period should clarify the interpretation of Indicator 15.3.1 for monitoring LDN. The guidance covers three sub-indicators. The Land Cover sub-indicator reports degradation in land cover change based on a national assessment of the positive or negative aspects of transitions from one land cover type to another. In Version 2, this chapter highlights the need for countries to first consider the main drivers of land cover change, and then to determine which transitions to identify as degraded for reporting on Indicator 15.3.1. This chapter incorporates updates to the quality of available land cover datasets and classification methods.
19.
Mariano Gonzalez-Roglich Alex Zvoleff, Ichsani Wheeler
Land Degradation Neutrality Impact Monitoring Methodology Technical Report
2020.
Abstract | Links | BibTeX | Tags: Impact Monitoring Methodology, Land Degradation Neutrality
@techreport{Zvoleff2020,
title = {Land Degradation Neutrality Impact Monitoring Methodology},
author = {Alex Zvoleff, Mariano Gonzalez-Roglich, Ichsani Wheeler, Tomislav Hengl},
url = {https://www.idhsustainabletrade.com/publication/land-degradation-neutrality-impact-monitoring-methodology/},
year = {2020},
date = {2020-09-21},
urldate = {2021-09-21},
abstract = {This is the technical briefing document, commissioned by IDH as manager of the Land
Degradation Neutrality Technical Assistance Facility (LDN TAF). In addition to this document,
a practical guide for LDN Fund investees and other project developers will be made available.
Land degradation – the reduction or loss of the productive potential of land – is a global
challenge. Over 20% of the Earth’s vegetated surface is degraded, affecting over 1.3 billion
people (1), with an economic impact of up to USD 10.6 trillion (2). Land degradation reduces
agricultural productivity and increases the vulnerability of those areas already at risk of impacts
from climate variability and change. The international community has organized around the
concept of Land Degradation Neutrality (LDN) to address the challenge of land degradation.
Sustainable land management (SLM) and land restoration are essential for achieving LDN, but
finance is needed to support these efforts.
To promote investment in profit generating SLM and restoration projects the LDN Fund was
created1
. The Fund requires that each project in which it invests contribute to the achievement of
LDN. Consistent with the agreed-upon indicators for assessing achievement of LDN, each
project is therefore expected to monitor three different indicators: land productivity, land cover,
and soil organic carbon.
The methods and framework for monitoring achievement of LDN at a national scale have been
established by United Nations Convention to Combat Desertification (UNCCD) and other key
stakeholders through the development of the scientific framework for LDN and the standardized
approaches that have been developed for national reporting to the UNCCD, and for monitoring
Sustainable Development Goal Target 15.3.1. This document outlines a monitoring approach that
adapts these existing national-level indicators to scale of a Fund investment.
This document outlines the recommended approach for monitoring the impact of LDN Fund
investments, and for assessing the overall contribution of each project to achieving LDN. While
targeted towards Fund investees, this document has broader applicability to any project team
interested in monitoring the contributions of a project towards the achievement of LDN.},
keywords = {Impact Monitoring Methodology, Land Degradation Neutrality},
pubstate = {published},
tppubtype = {techreport}
}
This is the technical briefing document, commissioned by IDH as manager of the Land
Degradation Neutrality Technical Assistance Facility (LDN TAF). In addition to this document,
a practical guide for LDN Fund investees and other project developers will be made available.
Land degradation – the reduction or loss of the productive potential of land – is a global
challenge. Over 20% of the Earth’s vegetated surface is degraded, affecting over 1.3 billion
people (1), with an economic impact of up to USD 10.6 trillion (2). Land degradation reduces
agricultural productivity and increases the vulnerability of those areas already at risk of impacts
from climate variability and change. The international community has organized around the
concept of Land Degradation Neutrality (LDN) to address the challenge of land degradation.
Sustainable land management (SLM) and land restoration are essential for achieving LDN, but
finance is needed to support these efforts.
To promote investment in profit generating SLM and restoration projects the LDN Fund was
created1
. The Fund requires that each project in which it invests contribute to the achievement of
LDN. Consistent with the agreed-upon indicators for assessing achievement of LDN, each
project is therefore expected to monitor three different indicators: land productivity, land cover,
and soil organic carbon.
The methods and framework for monitoring achievement of LDN at a national scale have been
established by United Nations Convention to Combat Desertification (UNCCD) and other key
stakeholders through the development of the scientific framework for LDN and the standardized
approaches that have been developed for national reporting to the UNCCD, and for monitoring
Sustainable Development Goal Target 15.3.1. This document outlines a monitoring approach that
adapts these existing national-level indicators to scale of a Fund investment.
This document outlines the recommended approach for monitoring the impact of LDN Fund
investments, and for assessing the overall contribution of each project to achieving LDN. While
targeted towards Fund investees, this document has broader applicability to any project team
interested in monitoring the contributions of a project towards the achievement of LDN.
Degradation Neutrality Technical Assistance Facility (LDN TAF). In addition to this document,
a practical guide for LDN Fund investees and other project developers will be made available.
Land degradation – the reduction or loss of the productive potential of land – is a global
challenge. Over 20% of the Earth’s vegetated surface is degraded, affecting over 1.3 billion
people (1), with an economic impact of up to USD 10.6 trillion (2). Land degradation reduces
agricultural productivity and increases the vulnerability of those areas already at risk of impacts
from climate variability and change. The international community has organized around the
concept of Land Degradation Neutrality (LDN) to address the challenge of land degradation.
Sustainable land management (SLM) and land restoration are essential for achieving LDN, but
finance is needed to support these efforts.
To promote investment in profit generating SLM and restoration projects the LDN Fund was
created1
. The Fund requires that each project in which it invests contribute to the achievement of
LDN. Consistent with the agreed-upon indicators for assessing achievement of LDN, each
project is therefore expected to monitor three different indicators: land productivity, land cover,
and soil organic carbon.
The methods and framework for monitoring achievement of LDN at a national scale have been
established by United Nations Convention to Combat Desertification (UNCCD) and other key
stakeholders through the development of the scientific framework for LDN and the standardized
approaches that have been developed for national reporting to the UNCCD, and for monitoring
Sustainable Development Goal Target 15.3.1. This document outlines a monitoring approach that
adapts these existing national-level indicators to scale of a Fund investment.
This document outlines the recommended approach for monitoring the impact of LDN Fund
investments, and for assessing the overall contribution of each project to achieving LDN. While
targeted towards Fund investees, this document has broader applicability to any project team
interested in monitoring the contributions of a project towards the achievement of LDN.
18.
Madlene Nussbaum Tomislav Hengl, Marvin N. Wright
Random forest as a generic framework for predictive modeling of spatial and spatio-temporal variables Journal Article
In: PeerJ, vol. 6, pp. e5518, 2018, ISSN: 2167-8359.
Abstract | Links | BibTeX | Tags: Geostatistics, Kriging, Pedometrics, R statistical computing, Random forest, Sampling, Spatial data
@article{10.7717/peerj.5518,
title = {Random forest as a generic framework for predictive modeling of spatial and spatio-temporal variables},
author = {Tomislav Hengl, Madlene Nussbaum, Marvin N. Wright, Gerard B.M. Heuvelink, Benedikt Gräler},
url = {https://doi.org/10.7717/peerj.5518},
doi = {10.7717/peerj.5518},
issn = {2167-8359},
year = {2018},
date = {2018-08-01},
urldate = {2018-08-01},
journal = {PeerJ},
volume = {6},
pages = {e5518},
abstract = {Random forest and similar Machine Learning techniques are already used to generate spatial predictions, but spatial location of points (geography) is often ignored in the modeling process. Spatial auto-correlation, especially if still existent in the cross-validation residuals, indicates that the predictions are maybe biased, and this is suboptimal. This paper presents a random forest for spatial predictions framework (RFsp) where buffer distances from observation points are used as explanatory variables, thus incorporating geographical proximity effects into the prediction process. The RFsp framework is illustrated with examples that use textbook datasets and apply spatial and spatio-temporal prediction to numeric, binary, categorical, multivariate and spatiotemporal variables. Performance of the RFsp framework is compared with the state-of-the-art kriging techniques using fivefold cross-validation with refitting. The results show that RFsp can obtain equally accurate and unbiased predictions as different versions of kriging. Advantages of using RFsp over kriging are that it needs no rigid statistical assumptions about the distribution and stationarity of the target variable, it is more flexible towards incorporating, combining and extending covariates of different types, and it possibly yields more informative maps characterizing the prediction error. RFsp appears to be especially attractive for building multivariate spatial prediction models that can be used as “knowledge engines†in various geoscience fields. Some disadvantages of RFsp are the exponentially growing computational intensity with increase of calibration data and covariates and the high sensitivity of predictions to input data quality. The key to the success of the RFsp framework might be the training data quality—especially quality of spatial sampling (to minimize extrapolation problems and any type of bias in data), and quality of model validation (to ensure that accuracy is not effected by overfitting). For many data sets, especially those with lower number of points and covariates and close-to-linear relationships, model-based geostatistics can still lead to more accurate predictions than RFsp.},
keywords = {Geostatistics, Kriging, Pedometrics, R statistical computing, Random forest, Sampling, Spatial data},
pubstate = {published},
tppubtype = {article}
}
Random forest and similar Machine Learning techniques are already used to generate spatial predictions, but spatial location of points (geography) is often ignored in the modeling process. Spatial auto-correlation, especially if still existent in the cross-validation residuals, indicates that the predictions are maybe biased, and this is suboptimal. This paper presents a random forest for spatial predictions framework (RFsp) where buffer distances from observation points are used as explanatory variables, thus incorporating geographical proximity effects into the prediction process. The RFsp framework is illustrated with examples that use textbook datasets and apply spatial and spatio-temporal prediction to numeric, binary, categorical, multivariate and spatiotemporal variables. Performance of the RFsp framework is compared with the state-of-the-art kriging techniques using fivefold cross-validation with refitting. The results show that RFsp can obtain equally accurate and unbiased predictions as different versions of kriging. Advantages of using RFsp over kriging are that it needs no rigid statistical assumptions about the distribution and stationarity of the target variable, it is more flexible towards incorporating, combining and extending covariates of different types, and it possibly yields more informative maps characterizing the prediction error. RFsp appears to be especially attractive for building multivariate spatial prediction models that can be used as “knowledge engines†in various geoscience fields. Some disadvantages of RFsp are the exponentially growing computational intensity with increase of calibration data and covariates and the high sensitivity of predictions to input data quality. The key to the success of the RFsp framework might be the training data quality—especially quality of spatial sampling (to minimize extrapolation problems and any type of bias in data), and quality of model validation (to ensure that accuracy is not effected by overfitting). For many data sets, especially those with lower number of points and covariates and close-to-linear relationships, model-based geostatistics can still lead to more accurate predictions than RFsp.
17.
Sanderman, Jonathan; Hengl, Tomislav; Fiske, Gregory J.
Soil carbon debt of 12,000 years of human land use Journal Article
In: Proceedings of the National Academy of Sciences, vol. 114, no. 36, pp. 9575–9580, 2017, ISSN: 0027-8424.
Abstract | Links | BibTeX | Tags:
@article{Sanderman9575,
title = {Soil carbon debt of 12,000 years of human land use},
author = {Jonathan Sanderman and Tomislav Hengl and Gregory J. Fiske},
url = {https://www.pnas.org/content/114/36/9575},
doi = {10.1073/pnas.1706103114},
issn = {0027-8424},
year = {2017},
date = {2017-01-01},
journal = {Proceedings of the National Academy of Sciences},
volume = {114},
number = {36},
pages = {9575--9580},
publisher = {National Academy of Sciences},
abstract = {Land use and land cover change has resulted in substantial losses of carbon from soils globally, but credible estimates of how much soil carbon has been lost have been difficult to generate. Using a data-driven statistical model and the History Database of the Global Environment v3.2 historic land-use dataset, we estimated that agricultural land uses have resulted in the loss of 133 Pg C from the soil. Importantly, our maps indicate hotspots of soil carbon loss, often associated with major cropping regions and degraded grazing lands, suggesting that there are identifiable regions that should be targets for soil carbon restoration efforts.Human appropriation of land for agriculture has greatly altered the terrestrial carbon balance, creating a large but uncertain carbon debt in soils. Estimating the size and spatial distribution of soil organic carbon (SOC) loss due to land use and land cover change has been difficult but is a critical step in understanding whether SOC sequestration can be an effective climate mitigation strategy. In this study, a machine learning-based model was fitted using a global compilation of SOC data and the History Database of the Global Environment (HYDE) land use data in combination with climatic, landform and lithology covariates. Model results compared favorably with a global compilation of paired plot studies. Projection of this model onto a world without agriculture indicated a global carbon debt due to agriculture of 133 Pg C for the top 2 m of soil, with the rate of loss increasing dramatically in the past 200 years. The HYDE classes textquotedblleftgrazingtextquotedblright and textquotedblleftcroplandtextquotedblright contributed nearly equally to the loss of SOC. There were higher percent SOC losses on cropland but since more than twice as much land is grazed, slightly higher total losses were found from grazing land. Important spatial patterns of SOC loss were found: Hotspots of SOC loss coincided with some major cropping regions as well as semiarid grazing regions, while other major agricultural zones showed small losses and even net gains in SOC. This analysis has demonstrated that there are identifiable regions which can be targeted for SOC restoration efforts.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Land use and land cover change has resulted in substantial losses of carbon from soils globally, but credible estimates of how much soil carbon has been lost have been difficult to generate. Using a data-driven statistical model and the History Database of the Global Environment v3.2 historic land-use dataset, we estimated that agricultural land uses have resulted in the loss of 133 Pg C from the soil. Importantly, our maps indicate hotspots of soil carbon loss, often associated with major cropping regions and degraded grazing lands, suggesting that there are identifiable regions that should be targets for soil carbon restoration efforts.Human appropriation of land for agriculture has greatly altered the terrestrial carbon balance, creating a large but uncertain carbon debt in soils. Estimating the size and spatial distribution of soil organic carbon (SOC) loss due to land use and land cover change has been difficult but is a critical step in understanding whether SOC sequestration can be an effective climate mitigation strategy. In this study, a machine learning-based model was fitted using a global compilation of SOC data and the History Database of the Global Environment (HYDE) land use data in combination with climatic, landform and lithology covariates. Model results compared favorably with a global compilation of paired plot studies. Projection of this model onto a world without agriculture indicated a global carbon debt due to agriculture of 133 Pg C for the top 2 m of soil, with the rate of loss increasing dramatically in the past 200 years. The HYDE classes textquotedblleftgrazingtextquotedblright and textquotedblleftcroplandtextquotedblright contributed nearly equally to the loss of SOC. There were higher percent SOC losses on cropland but since more than twice as much land is grazed, slightly higher total losses were found from grazing land. Important spatial patterns of SOC loss were found: Hotspots of SOC loss coincided with some major cropping regions as well as semiarid grazing regions, while other major agricultural zones showed small losses and even net gains in SOC. This analysis has demonstrated that there are identifiable regions which can be targeted for SOC restoration efforts.
16.
Jorge Mendes de Jesus Tomislav Hengl, Gerard B. M. Heuvelink
SoilGrids250m: Global gridded soil information based on machine learning Journal Article
In: PLOS ONE, vol. 12, no. 2, pp. 1-40, 2017.
Abstract | Links | BibTeX | Tags: Soil mapping
@article{10.1371/journal.pone.0169748,
title = {SoilGrids250m: Global gridded soil information based on machine learning},
author = { Tomislav Hengl,Jorge Mendes de Jesus,Gerard B. M. Heuvelink,Maria Ruiperez Gonzalez,Milan Kilibarda,Aleksandar Blagotić,Wei Shangguan,Marvin N. Wright,Xiaoyuan Geng,Bernhard Bauer-Marschallinger,Mario Antonio Guevara,Rodrigo Vargas,Robert A. MacMillan,Niels H. Batjes,Johan G. B. Leenaars,Eloi Ribeiro,Ichsani Wheeler,Stephan Mantel,Bas Kempen},
url = {https://doi.org/10.1371/journal.pone.0169748},
doi = {10.1371/journal.pone.0169748},
year = {2017},
date = {2017-01-01},
urldate = {2017-01-01},
journal = {PLOS ONE},
volume = {12},
number = {2},
pages = {1-40},
publisher = {Public Library of Science},
abstract = {This paper describes the technical development and accuracy assessment of the most recent and improved version of the SoilGrids system at 250m resolution (June 2016 update). SoilGrids provides global predictions for standard numeric soil properties (organic carbon, bulk density, Cation Exchange Capacity (CEC), pH, soil texture fractions and coarse fragments) at seven standard depths (0, 5, 15, 30, 60, 100 and 200 cm), in addition to predictions of depth to bedrock and distribution of soil classes based on the World Reference Base (WRB) and USDA classification systems (ca. 280 raster layers in total). Predictions were based on ca. 150,000 soil profiles used for training and a stack of 158 remote sensing-based soil covariates (primarily derived from MODIS land products, SRTM DEM derivatives, climatic images and global landform and lithology maps), which were used to fit an ensemble of machine learning methods—random forest and gradient boosting and/or multinomial logistic regression—as implemented in the R packages ranger, xgboost, nnet and caret. The results of 10–fold cross-validation show that the ensemble models explain between 56% (coarse fragments) and 83% (pH) of variation with an overall average of 61%. Improvements in the relative accuracy considering the amount of variation explained, in comparison to the previous version of SoilGrids at 1 km spatial resolution, range from 60 to 230%. Improvements can be attributed to: (1) the use of machine learning instead of linear regression, (2) to considerable investments in preparing finer resolution covariate layers and (3) to insertion of additional soil profiles. Further development of SoilGrids could include refinement of methods to incorporate input uncertainties and derivation of posterior probability distributions (per pixel), and further automation of spatial modeling so that soil maps can be generated for potentially hundreds of soil variables. Another area of future research is the development of methods for multiscale merging of SoilGrids predictions with local and/or national gridded soil products (e.g. up to 50 m spatial resolution) so that increasingly more accurate, complete and consistent global soil information can be produced. SoilGrids are available under the Open Data Base License.},
keywords = {Soil mapping},
pubstate = {published},
tppubtype = {article}
}
This paper describes the technical development and accuracy assessment of the most recent and improved version of the SoilGrids system at 250m resolution (June 2016 update). SoilGrids provides global predictions for standard numeric soil properties (organic carbon, bulk density, Cation Exchange Capacity (CEC), pH, soil texture fractions and coarse fragments) at seven standard depths (0, 5, 15, 30, 60, 100 and 200 cm), in addition to predictions of depth to bedrock and distribution of soil classes based on the World Reference Base (WRB) and USDA classification systems (ca. 280 raster layers in total). Predictions were based on ca. 150,000 soil profiles used for training and a stack of 158 remote sensing-based soil covariates (primarily derived from MODIS land products, SRTM DEM derivatives, climatic images and global landform and lithology maps), which were used to fit an ensemble of machine learning methods—random forest and gradient boosting and/or multinomial logistic regression—as implemented in the R packages ranger, xgboost, nnet and caret. The results of 10–fold cross-validation show that the ensemble models explain between 56% (coarse fragments) and 83% (pH) of variation with an overall average of 61%. Improvements in the relative accuracy considering the amount of variation explained, in comparison to the previous version of SoilGrids at 1 km spatial resolution, range from 60 to 230%. Improvements can be attributed to: (1) the use of machine learning instead of linear regression, (2) to considerable investments in preparing finer resolution covariate layers and (3) to insertion of additional soil profiles. Further development of SoilGrids could include refinement of methods to incorporate input uncertainties and derivation of posterior probability distributions (per pixel), and further automation of spatial modeling so that soil maps can be generated for potentially hundreds of soil variables. Another area of future research is the development of methods for multiscale merging of SoilGrids predictions with local and/or national gridded soil products (e.g. up to 50 m spatial resolution) so that increasingly more accurate, complete and consistent global soil information can be produced. SoilGrids are available under the Open Data Base License.
15.
Gerard B. M. Heuvelink Tomislav Hengl, Bas Kempen
Mapping Soil Properties of Africa at 250 m Resolution: Random Forests Significantly Improve Current Predictions Journal Article
In: PLOS ONE, vol. 10, no. 6, pp. 1-26, 2015.
Abstract | Links | BibTeX | Tags: Soil mapping
@article{10.1371/journal.pone.0125814,
title = {Mapping Soil Properties of Africa at 250 m Resolution: Random Forests Significantly Improve Current Predictions},
author = {Tomislav Hengl ,Gerard B. M. Heuvelink,Bas Kempen,Johan G. B. Leenaars,Markus G. Walsh,Keith D. Shepherd,Andrew Sila,Robert A. MacMillan,Jorge Mendes de Jesus,Lulseged Tamene,Jérôme E. Tondoh},
url = {https://doi.org/10.1371/journal.pone.0125814},
doi = {10.1371/journal.pone.0125814},
year = {2015},
date = {2015-01-01},
urldate = {2015-01-01},
journal = {PLOS ONE},
volume = {10},
number = {6},
pages = {1-26},
publisher = {Public Library of Science},
abstract = {80% of arable land in Africa has low soil fertility and suffers from physical soil problems. Additionally, significant amounts of nutrients are lost every year due to unsustainable soil management practices. This is partially the result of insufficient use of soil management knowledge. To help bridge the soil information gap in Africa, the Africa Soil Information Service (AfSIS) project was established in 2008. Over the period 2008–2014, the AfSIS project compiled two point data sets: the Africa Soil Profiles (legacy) database and the AfSIS Sentinel Site database. These data sets contain over 28 thousand sampling locations and represent the most comprehensive soil sample data sets of the African continent to date. Utilizing these point data sets in combination with a large number of covariates, we have generated a series of spatial predictions of soil properties relevant to the agricultural management—organic carbon, pH, sand, silt and clay fractions, bulk density, cation-exchange capacity, total nitrogen, exchangeable acidity, Al content and exchangeable bases (Ca, K, Mg, Na). We specifically investigate differences between two predictive approaches: random forests and linear regression. Results of 5-fold cross-validation demonstrate that the random forests algorithm consistently outperforms the linear regression algorithm, with average decreases of 15–75% in Root Mean Squared Error (RMSE) across soil properties and depths. Fitting and running random forests models takes an order of magnitude more time and the modelling success is sensitive to artifacts in the input data, but as long as quality-controlled point data are provided, an increase in soil mapping accuracy can be expected. Results also indicate that globally predicted soil classes (USDA Soil Taxonomy, especially Alfisols and Mollisols) help improve continental scale soil property mapping, and are among the most important predictors. This indicates a promising potential for transferring pedological knowledge from data rich countries to countries with limited soil data.},
keywords = {Soil mapping},
pubstate = {published},
tppubtype = {article}
}
80% of arable land in Africa has low soil fertility and suffers from physical soil problems. Additionally, significant amounts of nutrients are lost every year due to unsustainable soil management practices. This is partially the result of insufficient use of soil management knowledge. To help bridge the soil information gap in Africa, the Africa Soil Information Service (AfSIS) project was established in 2008. Over the period 2008–2014, the AfSIS project compiled two point data sets: the Africa Soil Profiles (legacy) database and the AfSIS Sentinel Site database. These data sets contain over 28 thousand sampling locations and represent the most comprehensive soil sample data sets of the African continent to date. Utilizing these point data sets in combination with a large number of covariates, we have generated a series of spatial predictions of soil properties relevant to the agricultural management—organic carbon, pH, sand, silt and clay fractions, bulk density, cation-exchange capacity, total nitrogen, exchangeable acidity, Al content and exchangeable bases (Ca, K, Mg, Na). We specifically investigate differences between two predictive approaches: random forests and linear regression. Results of 5-fold cross-validation demonstrate that the random forests algorithm consistently outperforms the linear regression algorithm, with average decreases of 15–75% in Root Mean Squared Error (RMSE) across soil properties and depths. Fitting and running random forests models takes an order of magnitude more time and the modelling success is sensitive to artifacts in the input data, but as long as quality-controlled point data are provided, an increase in soil mapping accuracy can be expected. Results also indicate that globally predicted soil classes (USDA Soil Taxonomy, especially Alfisols and Mollisols) help improve continental scale soil property mapping, and are among the most important predictors. This indicates a promising potential for transferring pedological knowledge from data rich countries to countries with limited soil data.
14.
Daniel Simberloff Ben H. Warren, Robert E. Ricklefs
Islands as model systems in ecology and evolution: prospects fifty years after MacArthur-Wilson Journal Article
In: Ecology Letters, vol. 18, no. 2, pp. 200-217, 2015.
Abstract | Links | BibTeX | Tags: Community assembly, diversification, ecosystem functioning, genomics, islands as model systems
@article{https://doi.org/10.1111/ele.12398,
title = {Islands as model systems in ecology and evolution: prospects fifty years after MacArthur-Wilson},
author = {Ben H. Warren,Daniel Simberloff,Robert E. Ricklefs,Robin Aguilée,Fabien L. Condamine,Dominique Gravel,Hélène Morlon,Nicolas Mouquet,James Rosindell,Juliane Casquet,Elena Conti,Josselin Cornuault,José María Fernández-Palacios,Tomislav Hengl,Sietze J. Norder,Kenneth F. Rijsdijk,Isabel Sanmartín,Dominique Strasberg,Kostas A. Triantis,Luis M. Valente,Robert J. Whittaker,Rosemary G. Gillespie,Brent C. Emerson,Christophe Thébaud},
url = {https://onlinelibrary.wiley.com/doi/abs/10.1111/ele.12398},
doi = {https://doi.org/10.1111/ele.12398},
year = {2015},
date = {2015-01-01},
urldate = {2015-01-01},
journal = {Ecology Letters},
volume = {18},
number = {2},
pages = {200-217},
abstract = {Abstract The study of islands as model systems has played an important role in the development of evolutionary and ecological theory. The 50th anniversary of MacArthur and Wilson's (December 1963) article, ‘An equilibrium theory of insular zoogeography’, was a recent milestone for this theme. Since 1963, island systems have provided new insights into the formation of ecological communities. Here, building on such developments, we highlight prospects for research on islands to improve our understanding of the ecology and evolution of communities in general. Throughout, we emphasise how attributes of islands combine to provide unusual research opportunities, the implications of which stretch far beyond islands. Molecular tools and increasing data acquisition now permit re-assessment of some fundamental issues that interested MacArthur and Wilson. These include the formation of ecological networks, species abundance distributions, and the contribution of evolution to community assembly. We also extend our prospects to other fields of ecology and evolution – understanding ecosystem functioning, speciation and diversification – frequently employing assets of oceanic islands in inferring the geographic area within which evolution has occurred, and potential barriers to gene flow. Although island-based theory is continually being enriched, incorporating non-equilibrium dynamics is identified as a major challenge for the future.},
keywords = {Community assembly, diversification, ecosystem functioning, genomics, islands as model systems},
pubstate = {published},
tppubtype = {article}
}
Abstract The study of islands as model systems has played an important role in the development of evolutionary and ecological theory. The 50th anniversary of MacArthur and Wilson's (December 1963) article, ‘An equilibrium theory of insular zoogeography’, was a recent milestone for this theme. Since 1963, island systems have provided new insights into the formation of ecological communities. Here, building on such developments, we highlight prospects for research on islands to improve our understanding of the ecology and evolution of communities in general. Throughout, we emphasise how attributes of islands combine to provide unusual research opportunities, the implications of which stretch far beyond islands. Molecular tools and increasing data acquisition now permit re-assessment of some fundamental issues that interested MacArthur and Wilson. These include the formation of ecological networks, species abundance distributions, and the contribution of evolution to community assembly. We also extend our prospects to other fields of ecology and evolution – understanding ecosystem functioning, speciation and diversification – frequently employing assets of oceanic islands in inferring the geographic area within which evolution has occurred, and potential barriers to gene flow. Although island-based theory is continually being enriched, incorporating non-equilibrium dynamics is identified as a major challenge for the future.
13.
Hengl, Tomislav; Jesus, Jorge Mendes; MacMillan, Robert A.; Batjes, Niels H.; Heuvelink, Gerard B. M.; Ribeiro, Eloi; Samuel-Rosa, Alessandro; Kempen, Bas; Leenaars, Johan G. B.; Walsh, Markus G.; Gonzalez, Maria Ruiperez
SoilGrids1km — Global Soil Information Based on Automated Mapping Journal Article
In: PLOS ONE, vol. 9, no. 8, pp. 1-17, 2014.
Abstract | Links | BibTeX | Tags:
@article{10.1371/journal.pone.0105992,
title = {SoilGrids1km — Global Soil Information Based on Automated Mapping},
author = {Tomislav Hengl and Jorge Mendes Jesus and Robert A. MacMillan and Niels H. Batjes and Gerard B. M. Heuvelink and Eloi Ribeiro and Alessandro Samuel-Rosa and Bas Kempen and Johan G. B. Leenaars and Markus G. Walsh and Maria Ruiperez Gonzalez},
url = {https://doi.org/10.1371/journal.pone.0105992},
doi = {10.1371/journal.pone.0105992},
year = {2014},
date = {2014-01-01},
journal = {PLOS ONE},
volume = {9},
number = {8},
pages = {1-17},
publisher = {Public Library of Science},
abstract = {Background Soils are widely recognized as a non-renewable natural resource and as biophysical carbon sinks. As such, there is a growing requirement for global soil information. Although several global soil information systems already exist, these tend to suffer from inconsistencies and limited spatial detail. Methodology/Principal Findings We present SoilGrids1km — a global 3D soil information system at 1 km resolution — containing spatial predictions for a selection of soil properties (at six standard depths): soil organic carbon (g kg−1), soil pH, sand, silt and clay fractions (%), bulk density (kg m−3), cation-exchange capacity (cmol+/kg), coarse fragments (%), soil organic carbon stock (t ha−1), depth to bedrock (cm), World Reference Base soil groups, and USDA Soil Taxonomy suborders. Our predictions are based on global spatial prediction models which we fitted, per soil variable, using a compilation of major international soil profile databases (ca. 110,000 soil profiles), and a selection of ca. 75 global environmental covariates representing soil forming factors. Results of regression modeling indicate that the most useful covariates for modeling soils at the global scale are climatic and biomass indices (based on MODIS images), lithology, and taxonomic mapping units derived from conventional soil survey (Harmonized World Soil Database). Prediction accuracies assessed using 5–fold cross-validation were between 23–51%. Conclusions/Significance SoilGrids1km provide an initial set of examples of soil spatial data for input into global models at a resolution and consistency not previously available. Some of the main limitations of the current version of SoilGrids1km are: (1) weak relationships between soil properties/classes and explanatory variables due to scale mismatches, (2) difficulty to obtain covariates that capture soil forming factors, (3) low sampling density and spatial clustering of soil profile locations. However, as the SoilGrids system is highly automated and flexible, increasingly accurate predictions can be generated as new input data become available. SoilGrids1km are available for download via http://soilgrids.org under a Creative Commons Non Commercial license.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Background Soils are widely recognized as a non-renewable natural resource and as biophysical carbon sinks. As such, there is a growing requirement for global soil information. Although several global soil information systems already exist, these tend to suffer from inconsistencies and limited spatial detail. Methodology/Principal Findings We present SoilGrids1km — a global 3D soil information system at 1 km resolution — containing spatial predictions for a selection of soil properties (at six standard depths): soil organic carbon (g kg−1), soil pH, sand, silt and clay fractions (%), bulk density (kg m−3), cation-exchange capacity (cmol+/kg), coarse fragments (%), soil organic carbon stock (t ha−1), depth to bedrock (cm), World Reference Base soil groups, and USDA Soil Taxonomy suborders. Our predictions are based on global spatial prediction models which we fitted, per soil variable, using a compilation of major international soil profile databases (ca. 110,000 soil profiles), and a selection of ca. 75 global environmental covariates representing soil forming factors. Results of regression modeling indicate that the most useful covariates for modeling soils at the global scale are climatic and biomass indices (based on MODIS images), lithology, and taxonomic mapping units derived from conventional soil survey (Harmonized World Soil Database). Prediction accuracies assessed using 5–fold cross-validation were between 23–51%. Conclusions/Significance SoilGrids1km provide an initial set of examples of soil spatial data for input into global models at a resolution and consistency not previously available. Some of the main limitations of the current version of SoilGrids1km are: (1) weak relationships between soil properties/classes and explanatory variables due to scale mismatches, (2) difficulty to obtain covariates that capture soil forming factors, (3) low sampling density and spatial clustering of soil profile locations. However, as the SoilGrids system is highly automated and flexible, increasingly accurate predictions can be generated as new input data become available. SoilGrids1km are available for download via http://soilgrids.org under a Creative Commons Non Commercial license.
12.
Tomislav Hengl Milan Kilibarda, Gerard B. M. Heuvelink; Bajat, Branislav
Spatio-temporal interpolation of daily temperatures for global land areas at 1km resolution Journal Article
In: Journal of Geophysical Research: Atmospheres, vol. 119, no. 5, pp. 2294-2313, 2014.
Abstract | Links | BibTeX | Tags: spatio-temporal kriging
@article{https://doi.org/10.1002/2013JD020803,
title = {Spatio-temporal interpolation of daily temperatures for global land areas at 1km resolution},
author = {Milan Kilibarda, Tomislav Hengl, Gerard B. M. Heuvelink, Benedikt Gräler, Edzer Pebesma, Melita Percec Tadic, and Branislav Bajat},
url = {https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1002/2013JD020803},
doi = {https://doi.org/10.1002/2013JD020803},
year = {2014},
date = {2014-01-01},
urldate = {2014-01-01},
journal = {Journal of Geophysical Research: Atmospheres},
volume = {119},
number = {5},
pages = {2294-2313},
abstract = {Combined Global Surface Summary of Day and European Climate Assessment and Dataset daily meteorological data sets (around 9000 stations) were used to build spatio-temporal geostatistical models and predict daily air temperature at ground esolution of 1 km for the global land mass. Predictions in space and time were made for the mean, maximum, and minimum temperatures using spatio-temporal regression-kriging with a time series of Moderate Resolution Imaging Spectroradiometer (MODIS) 8 day images, topographic layers (digital elevation model and topographic wetness index), and a geometric temperature trend as covariates. The accuracy of predicting daily temperatures was assessed using leave-one-out cross validation. To account for geographical point clustering of station data and get a more representative cross-validation accuracy, predicted values were aggregated to blocks of land of size 500 × 500 km. Results show that the average accuracy for predicting mean, maximum, and minimum daily temperatures is root-mean-square error (RMSE) = ±2◦C for areas densely covered with stations and between ± 2◦C and ± 4◦C for areas with lower station density. The lowest prediction accuracy was observed at high
altitudes (> 1000 m) and in Antarctica with an RMSE around 6◦C. The model and predictions were built for the year 2011 only, but the same methodology could be extended for the whole range of the MODIS land surface temperature images (2001 to today), i.e., to produce global archives of daily temperatures (a next-generation http://WorldClim.org repository) and to feed various global environmental models.
},
keywords = {spatio-temporal kriging},
pubstate = {published},
tppubtype = {article}
}
Combined Global Surface Summary of Day and European Climate Assessment and Dataset daily meteorological data sets (around 9000 stations) were used to build spatio-temporal geostatistical models and predict daily air temperature at ground esolution of 1 km for the global land mass. Predictions in space and time were made for the mean, maximum, and minimum temperatures using spatio-temporal regression-kriging with a time series of Moderate Resolution Imaging Spectroradiometer (MODIS) 8 day images, topographic layers (digital elevation model and topographic wetness index), and a geometric temperature trend as covariates. The accuracy of predicting daily temperatures was assessed using leave-one-out cross validation. To account for geographical point clustering of station data and get a more representative cross-validation accuracy, predicted values were aggregated to blocks of land of size 500 × 500 km. Results show that the average accuracy for predicting mean, maximum, and minimum daily temperatures is root-mean-square error (RMSE) = ±2◦C for areas densely covered with stations and between ± 2◦C and ± 4◦C for areas with lower station density. The lowest prediction accuracy was observed at high
altitudes (> 1000 m) and in Antarctica with an RMSE around 6◦C. The model and predictions were built for the year 2011 only, but the same methodology could be extended for the whole range of the MODIS land surface temperature images (2001 to today), i.e., to produce global archives of daily temperatures (a next-generation http://WorldClim.org repository) and to feed various global environmental models.
altitudes (> 1000 m) and in Antarctica with an RMSE around 6◦C. The model and predictions were built for the year 2011 only, but the same methodology could be extended for the whole range of the MODIS land surface temperature images (2001 to today), i.e., to produce global archives of daily temperatures (a next-generation http://WorldClim.org repository) and to feed various global environmental models.
11.
Hengl, Tomislav; Heuvelink, Gerard BM; Tadic, Melita Perčec; Pebesma, Edzer J
Spatio-temporal prediction of daily temperatures using time-series of MODIS LST images Journal Article
In: Theoretical and applied climatology, vol. 107, no. 1, pp. 265–277, 2012.
BibTeX | Tags:
@article{hengl2012spatio,
title = {Spatio-temporal prediction of daily temperatures using time-series of MODIS LST images},
author = {Tomislav Hengl and Gerard BM Heuvelink and Melita Perčec Tadic and Edzer J Pebesma},
year = {2012},
date = {2012-01-01},
urldate = {2012-01-01},
journal = {Theoretical and applied climatology},
volume = {107},
number = {1},
pages = {265--277},
publisher = {Springer},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
10.
Pike, R. J.; Evans, I. S.; Hengl, T.
Chapter 1 Geomorphometry: A Brief Guide Book Chapter
In: Hengl, Tomislav; Reuter, Hannes I. (Ed.): Geomorphometry, vol. 33, Chapter 1, pp. 3-30, Elsevier, 2009, ISSN: 0166-2481.
Abstract | Links | BibTeX | Tags: basic principles of geomorphometry from a GIS perspective, data set used in this book, digital elevation models (DEMs), history of geomorphometry, inputs/outputs, land-surface parameters and objects, the land surface
@inbook{PIKE20093,
title = {Chapter 1 Geomorphometry: A Brief Guide},
author = {R. J. Pike and I. S. Evans and T. Hengl},
editor = {Tomislav Hengl and Hannes I. Reuter},
url = {https://www.sciencedirect.com/science/article/pii/S0166248108000019},
doi = {https://doi.org/10.1016/S0166-2481(08)00001-9},
issn = {0166-2481},
year = {2009},
date = {2009-01-01},
urldate = {2009-01-01},
booktitle = {Geomorphometry},
volume = {33},
pages = {3-30},
publisher = {Elsevier},
chapter = {1},
series = {Developments in Soil Science},
abstract = {Geomorphometry is the science of quantitative land-surface analysis. It evolved directly from geomorphology and quantitative terrain analysis, two disciplines that originated in 19th century geometry, physical geography, and the measurement of mountains. Modern geomorphometry addresses the refinement and processing of elevation data, description and visualization of topography, and a wide variety of numerical analyses. It focuses on the continuous land-surface, although it also includes the analysis of landforms, discrete features, such as watersheds. The operational goal of geomorphometry is extraction of measures and spatial features from digital topography. Geomorphometry supports countless applications in the Earth sciences, civil engineering, military operations, and entertainment. Geomorphometric analysis commonly entails five steps: Sampling a surface, generating and correcting a surface model, calculating land-surface parameters or objects, and applying the results. The three classes of parameters and objects include both landforms and pointmeasures, such as slope and curvature. Landform elements are fundamental spatial units having uniform properties. Complex analyses may combine several parameter maps and incorporate non-topographic data. The procedure that extracts most land-surface parameters and objects from a digital elevation model (DEM) is the neighborhood operation. Because parameters can be generated by different algorithms or sampling strategies, and vary with spatial scale, no DEM-derived map is definitive.},
keywords = {basic principles of geomorphometry from a GIS perspective, data set used in this book, digital elevation models (DEMs), history of geomorphometry, inputs/outputs, land-surface parameters and objects, the land surface},
pubstate = {published},
tppubtype = {inbook}
}
Geomorphometry is the science of quantitative land-surface analysis. It evolved directly from geomorphology and quantitative terrain analysis, two disciplines that originated in 19th century geometry, physical geography, and the measurement of mountains. Modern geomorphometry addresses the refinement and processing of elevation data, description and visualization of topography, and a wide variety of numerical analyses. It focuses on the continuous land-surface, although it also includes the analysis of landforms, discrete features, such as watersheds. The operational goal of geomorphometry is extraction of measures and spatial features from digital topography. Geomorphometry supports countless applications in the Earth sciences, civil engineering, military operations, and entertainment. Geomorphometric analysis commonly entails five steps: Sampling a surface, generating and correcting a surface model, calculating land-surface parameters or objects, and applying the results. The three classes of parameters and objects include both landforms and pointmeasures, such as slope and curvature. Landform elements are fundamental spatial units having uniform properties. Complex analyses may combine several parameter maps and incorporate non-topographic data. The procedure that extracts most land-surface parameters and objects from a digital elevation model (DEM) is the neighborhood operation. Because parameters can be generated by different algorithms or sampling strategies, and vary with spatial scale, no DEM-derived map is definitive.
9.
Hengl, Tomislav; Sierdsema, Henk; Radovic, Andreja; Dilo, Arta
Spatial prediction of species’ distributions from occurrence-only records: combining point pattern analysis, ENFA and regression-kriging Journal Article
In: Ecological modelling, vol. 220, no. 24, pp. 3499–3511, 2009.
Abstract | Links | BibTeX | Tags: species distribution
@article{hengl2009spatial,
title = {Spatial prediction of species’ distributions from occurrence-only records: combining point pattern analysis, ENFA and regression-kriging},
author = {Tomislav Hengl and Henk Sierdsema and Andreja Radovic and Arta Dilo},
url = {https://www.sciencedirect.com/science/article/abs/pii/S0304380009004438},
doi = {10.1016/j.ecolmodel.2009.06.038},
year = {2009},
date = {2009-01-01},
urldate = {2009-01-01},
journal = {Ecological modelling},
volume = {220},
number = {24},
pages = {3499--3511},
publisher = {Elsevier},
abstract = {A computational framework to map species’ distributions (realized density) using occurrence-only data and environmental predictors is presented and illustrated using a textbook example and two case studies: distribution of root vole (Microtes oeconomus) in the Netherlands, and distribution of white-tailed eagle nests (Haliaeetus albicilla) in Croatia. The framework combines strengths of point pattern analysis (kernel smoothing), Ecological Niche Factor Analysis (ENFA) and geostatistics (logistic regression-kriging), as implemented in the spatstat, adehabitat and gstat packages of the R environment for statistical computing. A procedure to generate pseudo-absences is proposed. It uses Habitat Suitability Index (HSI, derived through ENFA) and distance from observations as weight maps to allocate pseudo-absence points. This design ensures that the simulated pseudo-absences fall further away from the occurrence points in both feature and geographical spaces. The simulated pseudo-absences can then be combined with occurrence locations and used to build regression-kriging prediction models. The output of prediction are either probabilitiesy of species’ occurrence or density measures. Addition of the pseudo-absence locations has proven effective — the adjusted R-square increased from 0.71 to 0.80 for root vole (562 records), and from 0.69 to 0.83 for white-tailed eagle (135 records) respectively; pseudo-absences improve spreading of the points in feature space and ensure consistent mapping over the whole area of interest. Results of cross validation (leave-one-out method) for these two species showed that the model explains 98% of the total variability in the density values for the root vole, and 94% of the total variability for the white-tailed eagle. The framework could be further extended to Generalized multivariate Linear Geostatistical Models and spatial prediction of multiple species. A copy of the R script and step-by-step instructions to run such analysis are available via contact author’s website.},
keywords = {species distribution},
pubstate = {published},
tppubtype = {article}
}
A computational framework to map species’ distributions (realized density) using occurrence-only data and environmental predictors is presented and illustrated using a textbook example and two case studies: distribution of root vole (Microtes oeconomus) in the Netherlands, and distribution of white-tailed eagle nests (Haliaeetus albicilla) in Croatia. The framework combines strengths of point pattern analysis (kernel smoothing), Ecological Niche Factor Analysis (ENFA) and geostatistics (logistic regression-kriging), as implemented in the spatstat, adehabitat and gstat packages of the R environment for statistical computing. A procedure to generate pseudo-absences is proposed. It uses Habitat Suitability Index (HSI, derived through ENFA) and distance from observations as weight maps to allocate pseudo-absence points. This design ensures that the simulated pseudo-absences fall further away from the occurrence points in both feature and geographical spaces. The simulated pseudo-absences can then be combined with occurrence locations and used to build regression-kriging prediction models. The output of prediction are either probabilitiesy of species’ occurrence or density measures. Addition of the pseudo-absence locations has proven effective — the adjusted R-square increased from 0.71 to 0.80 for root vole (562 records), and from 0.69 to 0.83 for white-tailed eagle (135 records) respectively; pseudo-absences improve spreading of the points in feature space and ensure consistent mapping over the whole area of interest. Results of cross validation (leave-one-out method) for these two species showed that the model explains 98% of the total variability in the density values for the root vole, and 94% of the total variability for the white-tailed eagle. The framework could be further extended to Generalized multivariate Linear Geostatistical Models and spatial prediction of multiple species. A copy of the R script and step-by-step instructions to run such analysis are available via contact author’s website.
8.
Hengl, T.; Evans, I. S.
Chapter 2 Mathematical and Digital Models of the Land Surface Book Chapter
In: Hengl, Tomislav; Reuter, Hannes I. (Ed.): Geomorphometry, vol. 33, Chapter 2, pp. 31-63, Elsevier, 2009, ISSN: 0166-2481.
Abstract | Links | BibTeX | Tags: cell size and its meaning, conceptual models of the land surface, how to determine a suitable grid resolution for DEMs, how to sample and interpolate heights, land surface and geomorphometric algorithms, land surface from a geodetic perspective, land-surface properties and mathematical models, vector and grid models of the land surface
@inbook{HENGL200931,
title = {Chapter 2 Mathematical and Digital Models of the Land Surface},
author = {T. Hengl and I. S. Evans},
editor = {Tomislav Hengl and Hannes I. Reuter},
url = {https://www.sciencedirect.com/science/article/pii/S0166248108000020},
doi = {https://doi.org/10.1016/S0166-2481(08)00002-0},
issn = {0166-2481},
year = {2009},
date = {2009-01-01},
urldate = {2009-01-01},
booktitle = {Geomorphometry},
volume = {33},
pages = {31-63},
publisher = {Elsevier},
chapter = {2},
series = {Developments in Soil Science},
abstract = {This chapter introduces the land-surface concept from both the geodetic and statistical perspectives, and reviews ways to represent it. It also discusses ways of producing models of the land-surface, from sampling procedures to digital elevation model (DEM) gridding techniques. An extensive comparison of the methods used to derive first and second order derivatives from DEMs have been presented. Mathematical models of the land surface have their uses, but it can be dangerous to regard them as being universally applicable, or even as capturing the essence of a real land surface. Understanding the concept of the land surface and its specific properties is a first step toward successful geomorphometric analysis. Ignoring aspects, such as the correct definition of a reference vertical datum, the density and distribution of the initial height observations, and the accuracy of measurement, can lead to serious artefacts and inaccuracies in the outputs of geomorphometric analysis.},
keywords = {cell size and its meaning, conceptual models of the land surface, how to determine a suitable grid resolution for DEMs, how to sample and interpolate heights, land surface and geomorphometric algorithms, land surface from a geodetic perspective, land-surface properties and mathematical models, vector and grid models of the land surface},
pubstate = {published},
tppubtype = {inbook}
}
This chapter introduces the land-surface concept from both the geodetic and statistical perspectives, and reviews ways to represent it. It also discusses ways of producing models of the land-surface, from sampling procedures to digital elevation model (DEM) gridding techniques. An extensive comparison of the methods used to derive first and second order derivatives from DEMs have been presented. Mathematical models of the land surface have their uses, but it can be dangerous to regard them as being universally applicable, or even as capturing the essence of a real land surface. Understanding the concept of the land surface and its specific properties is a first step toward successful geomorphometric analysis. Ignoring aspects, such as the correct definition of a reference vertical datum, the density and distribution of the initial height observations, and the accuracy of measurement, can lead to serious artefacts and inaccuracies in the outputs of geomorphometric analysis.
7.
Tomislav Hengl Luis Rodríguez Lado, Hannes I. Reuter
Heavy metals in European soils: A geostatistical analysis of the FOREGS Geochemical database Journal Article
In: Geoderma, vol. 148, no. 2, pp. 189-199, 2008, ISSN: 0016-7061.
Abstract | Links | BibTeX | Tags: Night lights image, Regression-kriging
@article{LADO2008189,
title = {Heavy metals in European soils: A geostatistical analysis of the FOREGS Geochemical database},
author = {Luis Rodríguez Lado, Tomislav Hengl, Hannes I. Reuter},
url = {https://www.sciencedirect.com/science/article/pii/S0016706108002668},
doi = {https://doi.org/10.1016/j.geoderma.2008.09.020},
issn = {0016-7061},
year = {2008},
date = {2008-01-01},
urldate = {2008-01-01},
journal = {Geoderma},
volume = {148},
number = {2},
pages = {189-199},
abstract = {This paper presents the results of modeling the distribution of eight critical heavy metals (arsenic, cadmium, chromium, copper, mercury, nickel, lead and zinc) in topsoils using 1588 georeferenced samples from the Forum of European Geological Surveys Geochemical database (26 European countries). The concentrations were mapped using regression-kriging (RK) and accuracy of predictions evaluated using the leave-one-out cross validation method. A large number of auxiliary raster maps (topographic indexes, land cover, geology, vegetation indexes, night lights images and earth quake magnitudes) were used to improve the predictions. These were first converted to 36 principal components and then used to explain spatial distribution of heavy metals. The study revealed that this database is suitable for geostatistical analyses: the predictors explained from 21% (Cr) to 35% (Pb) of variability; the residuals showed spatial autocorrelation. The Principal Component Analysis of the mapped heavy metals revealed that the administrative units (NUTS level3) with highest overall concentrations are: (1) Liege (Arrondissement) (BE), Attiki (GR), Darlington (UK), Coventry (UK), Sunderland (UK), Kozani (GR), Grevena (GR), Hartlepool & Stockton (UK), Huy (BE), Aachen (DE) (As, Cd, Hg and Pb) and (2) central Greece and Liguria region in Italy (Cr, Cu and Ni). The evaluation of the mapping accuracy showed that the RK models for As, Ni and Pb can be considered satisfactory (prediction accuracy 45–52% of total variance), marginally satisfactory for Cr, Cu, Hg and Zn (36–41%), while the model for Cd is unsatisfactorily accurate (30%). The critical elements limiting the mapping accuracy are: (a) the problem of sporadic high values (hot-spots); and (b) relatively coarse resolution of the input maps. Automation of the geostatistical mapping and use of auxiliary spatial layers opens a possibility to develop mapping systems that can automatically update outputs by including new field observations and higher quality auxiliary maps. This approach also demonstrates the benefits of organizing standardized joint European monitoring projects, in comparison to the merging of several national monitoring projects.},
keywords = {Night lights image, Regression-kriging},
pubstate = {published},
tppubtype = {article}
}
This paper presents the results of modeling the distribution of eight critical heavy metals (arsenic, cadmium, chromium, copper, mercury, nickel, lead and zinc) in topsoils using 1588 georeferenced samples from the Forum of European Geological Surveys Geochemical database (26 European countries). The concentrations were mapped using regression-kriging (RK) and accuracy of predictions evaluated using the leave-one-out cross validation method. A large number of auxiliary raster maps (topographic indexes, land cover, geology, vegetation indexes, night lights images and earth quake magnitudes) were used to improve the predictions. These were first converted to 36 principal components and then used to explain spatial distribution of heavy metals. The study revealed that this database is suitable for geostatistical analyses: the predictors explained from 21% (Cr) to 35% (Pb) of variability; the residuals showed spatial autocorrelation. The Principal Component Analysis of the mapped heavy metals revealed that the administrative units (NUTS level3) with highest overall concentrations are: (1) Liege (Arrondissement) (BE), Attiki (GR), Darlington (UK), Coventry (UK), Sunderland (UK), Kozani (GR), Grevena (GR), Hartlepool & Stockton (UK), Huy (BE), Aachen (DE) (As, Cd, Hg and Pb) and (2) central Greece and Liguria region in Italy (Cr, Cu and Ni). The evaluation of the mapping accuracy showed that the RK models for As, Ni and Pb can be considered satisfactory (prediction accuracy 45–52% of total variance), marginally satisfactory for Cr, Cu, Hg and Zn (36–41%), while the model for Cd is unsatisfactorily accurate (30%). The critical elements limiting the mapping accuracy are: (a) the problem of sporadic high values (hot-spots); and (b) relatively coarse resolution of the input maps. Automation of the geostatistical mapping and use of auxiliary spatial layers opens a possibility to develop mapping systems that can automatically update outputs by including new field observations and higher quality auxiliary maps. This approach also demonstrates the benefits of organizing standardized joint European monitoring projects, in comparison to the merging of several national monitoring projects.
6.
Hengl, Tomislav; Toomanian, Norair; Reuter, Hannes I.; Malakouti, Mohammad J.
Methods to interpolate soil categorical variables from profile observations: Lessons from Iran Journal Article
In: Geoderma, vol. 140, no. 4, pp. 417-427, 2007, ISSN: 0016-7061, (Pedometrics 2005).
Abstract | Links | BibTeX | Tags: Categorical variables, Fuzzy -means, MODIS, Multinominal logistic regression, Regression-kriging, Terrain parameters
@article{HENGL2007417,
title = {Methods to interpolate soil categorical variables from profile observations: Lessons from Iran},
author = {Tomislav Hengl and Norair Toomanian and Hannes I. Reuter and Mohammad J. Malakouti},
url = {https://www.sciencedirect.com/science/article/pii/S0016706107001218},
doi = {https://doi.org/10.1016/j.geoderma.2007.04.022},
issn = {0016-7061},
year = {2007},
date = {2007-01-01},
journal = {Geoderma},
volume = {140},
number = {4},
pages = {417-427},
abstract = {The paper compares semi-automated interpolation methods to produce soil-class maps from profile observations and by using multiple auxiliary predictors such as terrain parameters, remote sensing indices and similar. The Soil Profile Database of Iran, consisting of 4250 profiles, was used to test different soil-class interpolators. The target variables were soil texture classes and World Reference Base soil groups. The predictors were 6 terrain parameters, 11 MODIS EVI images and 17 physiographic regions (polygon map) of Iran. Four techniques were considered: (a) supervised classification using maximum likelihoods; (b) multinominal logistic regression; (c) regression-kriging on memberships; and (d) classification of taxonomic distances. The predictive capabilities were assessed using a control subset of 30% profiles and the kappa statistics as criterion. Supervised classification and multinominal logistic regression can lead to poor results if soil-classes overlap in the feature space, or if the correlation between the soil-classes and predictors is low. The two other methods have better predictive capabilities, although both are computationally more demanding. For both mapping of texture classes and soil types, the best prediction was achieved using regression-kriging of indicators/memberships (κ=45%, κ=54%). In all cases kappa was smaller than 60%, which can be explained by the preferential sampling plan, the poor definition of soil-classes and the high variability of soils. Steps to improve interpolation of soil-class data, by taking into account the fuzziness of classes directly on the field are further discussed.},
note = {Pedometrics 2005},
keywords = {Categorical variables, Fuzzy -means, MODIS, Multinominal logistic regression, Regression-kriging, Terrain parameters},
pubstate = {published},
tppubtype = {article}
}
The paper compares semi-automated interpolation methods to produce soil-class maps from profile observations and by using multiple auxiliary predictors such as terrain parameters, remote sensing indices and similar. The Soil Profile Database of Iran, consisting of 4250 profiles, was used to test different soil-class interpolators. The target variables were soil texture classes and World Reference Base soil groups. The predictors were 6 terrain parameters, 11 MODIS EVI images and 17 physiographic regions (polygon map) of Iran. Four techniques were considered: (a) supervised classification using maximum likelihoods; (b) multinominal logistic regression; (c) regression-kriging on memberships; and (d) classification of taxonomic distances. The predictive capabilities were assessed using a control subset of 30% profiles and the kappa statistics as criterion. Supervised classification and multinominal logistic regression can lead to poor results if soil-classes overlap in the feature space, or if the correlation between the soil-classes and predictors is low. The two other methods have better predictive capabilities, although both are computationally more demanding. For both mapping of texture classes and soil types, the best prediction was achieved using regression-kriging of indicators/memberships (κ=45%, κ=54%). In all cases kappa was smaller than 60%, which can be explained by the preferential sampling plan, the poor definition of soil-classes and the high variability of soils. Steps to improve interpolation of soil-class data, by taking into account the fuzziness of classes directly on the field are further discussed.
5.
Hengl, Tomislav; Heuvelink, Gerard B. M.; Rossiter, David G.
About regression-kriging: From equations to case studies Journal Article
In: Computers & Geosciences, vol. 33, no. 10, pp. 1301-1315, 2007, ISSN: 0098-3004, (Spatial Analysis).
Abstract | Links | BibTeX | Tags: Environmental predictors, GSTAT, MODIS, Multiple regression, Spatial prediction, SRTM
@article{HENGL20071301,
title = {About regression-kriging: From equations to case studies},
author = {Tomislav Hengl and Gerard B. M. Heuvelink and David G. Rossiter},
url = {https://www.sciencedirect.com/science/article/pii/S0098300407001008},
doi = {https://doi.org/10.1016/j.cageo.2007.05.001},
issn = {0098-3004},
year = {2007},
date = {2007-01-01},
journal = {Computers & Geosciences},
volume = {33},
number = {10},
pages = {1301-1315},
abstract = {This paper discusses the characteristics of regression-kriging (RK), its strengths and limitations, and illustrates these with a simple example and three case studies. RK is a spatial interpolation technique that combines a regression of the dependent variable on auxiliary variables (such as land surface parameters, remote sensing imagery and thematic maps) with simple kriging of the regression residuals. It is mathematically equivalent to the interpolation method variously called “Universal Kriging†(UK) and “Kriging with External Drift†(KED), where auxiliary predictors are used directly to solve the kriging weights. The advantage of RK is the ability to extend the method to a broader range of regression techniques and to allow separate interpretation of the two interpolated components. Data processing and interpretation of results are illustrated with three case studies covering the national territory of Croatia. The case studies use land surface parameters derived from combined Shuttle Radar Topography Mission and contour-based digital elevation models and multitemporal-enhanced vegetation indices derived from the MODIS imagery as auxiliary predictors. These are used to improve mapping of two continuous variables (soil organic matter content and mean annual land surface temperature) and one binary variable (presence of yew). In the case of mapping temperature, a physical model is used to estimate values of temperature at unvisited locations and RK is then used to calibrate the model with ground observations. The discussion addresses pragmatic issues: implementation of RK in existing software packages, comparison of RK with alternative interpolation techniques, and practical limitations to using RK. The most serious constraint to wider use of RK is that the analyst must carry out various steps in different software environments, both statistical and GIS.},
note = {Spatial Analysis},
keywords = {Environmental predictors, GSTAT, MODIS, Multiple regression, Spatial prediction, SRTM},
pubstate = {published},
tppubtype = {article}
}
This paper discusses the characteristics of regression-kriging (RK), its strengths and limitations, and illustrates these with a simple example and three case studies. RK is a spatial interpolation technique that combines a regression of the dependent variable on auxiliary variables (such as land surface parameters, remote sensing imagery and thematic maps) with simple kriging of the regression residuals. It is mathematically equivalent to the interpolation method variously called “Universal Kriging†(UK) and “Kriging with External Drift†(KED), where auxiliary predictors are used directly to solve the kriging weights. The advantage of RK is the ability to extend the method to a broader range of regression techniques and to allow separate interpretation of the two interpolated components. Data processing and interpretation of results are illustrated with three case studies covering the national territory of Croatia. The case studies use land surface parameters derived from combined Shuttle Radar Topography Mission and contour-based digital elevation models and multitemporal-enhanced vegetation indices derived from the MODIS imagery as auxiliary predictors. These are used to improve mapping of two continuous variables (soil organic matter content and mean annual land surface temperature) and one binary variable (presence of yew). In the case of mapping temperature, a physical model is used to estimate values of temperature at unvisited locations and RK is then used to calibrate the model with ground observations. The discussion addresses pragmatic issues: implementation of RK in existing software packages, comparison of RK with alternative interpolation techniques, and practical limitations to using RK. The most serious constraint to wider use of RK is that the analyst must carry out various steps in different software environments, both statistical and GIS.
4.
Hengl, Tomislav
Finding the right pixel size Journal Article
In: Computers & Geosciences, vol. 32, no. 9, pp. 1283-1298, 2006, ISSN: 0098-3004.
Abstract | Links | BibTeX | Tags: Grid resolution, Inspection density, Point pattern analysis, Scale, Terrain complexity, Variogram
@article{HENGL20061283,
title = {Finding the right pixel size},
author = {Tomislav Hengl},
url = {https://www.sciencedirect.com/science/article/pii/S0098300405002657},
doi = {https://doi.org/10.1016/j.cageo.2005.11.008},
issn = {0098-3004},
year = {2006},
date = {2006-01-01},
journal = {Computers & Geosciences},
volume = {32},
number = {9},
pages = {1283-1298},
abstract = {This paper discusses empirical and analytical rules to select a suitable grid resolution for output maps and based on the inherent properties of the input data. The choice of grid resolution was related with the cartographic and statistical concepts: scale, computer processing power, positional accuracy, size of delineations, inspection density, spatial autocorrelation structure and complexity of terrain. These were further related with the concepts from the general statistics and information theory such as Nyquist frequency concept from signal processing and equations to estimate the probability density function. Selection of grid resolution was demonstrated using four datasets: (1) GPS positioning data—the grid resolution was related to the area of circle described by the error radius, (2) map of agricultural plots—the grid resolution was related to the size of smallest and narrowest plots, (3) point dataset from soil mapping—the grid resolution was related to the inspection density, nugget variation and range of spatial autocorrelation and (4) contour map used for production of digital elevation model—the grid resolution was related with the spacing between the contour lines i.e. complexity of terrain. It was concluded that no ideal grid resolution exists, but rather a range of suitable resolutions. One should at least try to avoid using resolutions that do not comply with the effective scale or inherent properties of the input dataset. Three standard grid resolutions for output maps were finally recommended: (a) the coarsest legible grid resolution—this is the largest resolution that we should use in order to respect the scale of work and properties of a dataset; (b) the finest legible grid resolution—this is the smallest grid resolution that represents 95% of spatial objects or topography; and (c) recommended grid resolution—a compromise between the two. Objective procedures to derive the true optimal grid resolution that maximizes the predictive capabilities or information content of a map are further discussed. This methodology can now be integrated within a GIS package to help inexperienced users select a suitable grid resolution without doing extensive data preprocessing.},
keywords = {Grid resolution, Inspection density, Point pattern analysis, Scale, Terrain complexity, Variogram},
pubstate = {published},
tppubtype = {article}
}
This paper discusses empirical and analytical rules to select a suitable grid resolution for output maps and based on the inherent properties of the input data. The choice of grid resolution was related with the cartographic and statistical concepts: scale, computer processing power, positional accuracy, size of delineations, inspection density, spatial autocorrelation structure and complexity of terrain. These were further related with the concepts from the general statistics and information theory such as Nyquist frequency concept from signal processing and equations to estimate the probability density function. Selection of grid resolution was demonstrated using four datasets: (1) GPS positioning data—the grid resolution was related to the area of circle described by the error radius, (2) map of agricultural plots—the grid resolution was related to the size of smallest and narrowest plots, (3) point dataset from soil mapping—the grid resolution was related to the inspection density, nugget variation and range of spatial autocorrelation and (4) contour map used for production of digital elevation model—the grid resolution was related with the spacing between the contour lines i.e. complexity of terrain. It was concluded that no ideal grid resolution exists, but rather a range of suitable resolutions. One should at least try to avoid using resolutions that do not comply with the effective scale or inherent properties of the input dataset. Three standard grid resolutions for output maps were finally recommended: (a) the coarsest legible grid resolution—this is the largest resolution that we should use in order to respect the scale of work and properties of a dataset; (b) the finest legible grid resolution—this is the smallest grid resolution that represents 95% of spatial objects or topography; and (c) recommended grid resolution—a compromise between the two. Objective procedures to derive the true optimal grid resolution that maximizes the predictive capabilities or information content of a map are further discussed. This methodology can now be integrated within a GIS package to help inexperienced users select a suitable grid resolution without doing extensive data preprocessing.
3.
Hengl, Tomislav; Heuvelink, Gerard B. M.; Stein, Alfred
A generic framework for spatial prediction of soil variables based on regression-kriging Journal Article
In: Geoderma, vol. 120, no. 1, pp. 75-93, 2004, ISSN: 0016-7061.
Abstract | Links | BibTeX | Tags: Environmental correlation, Factor analysis, Logit transformation, Spatial prediction, Visualisation
@article{HENGL200475,
title = {A generic framework for spatial prediction of soil variables based on regression-kriging},
author = {Tomislav Hengl and Gerard B. M. Heuvelink and Alfred Stein},
url = {https://www.sciencedirect.com/science/article/pii/S0016706103002787},
doi = {https://doi.org/10.1016/j.geoderma.2003.08.018},
issn = {0016-7061},
year = {2004},
date = {2004-01-01},
journal = {Geoderma},
volume = {120},
number = {1},
pages = {75-93},
abstract = {A methodological framework for spatial prediction based on regression-kriging is described and compared with ordinary kriging and plain regression. The data are first transformed using logit transformation for target variables and factor analysis for continuous predictors (auxiliary maps). The target variables are then fitted using step-wise regression and residuals interpolated using kriging. A generic visualisation method is used to simultaneously display predictions and associated uncertainty. The framework was tested using 135 profile observations from the national survey in Croatia, divided into interpolation (100) and validation sets (35). Three target variables: organic matter, pH in topsoil and topsoil thickness were predicted from six relief parameters and nine soil mapping units. Prediction efficiency was evaluated using the mean error and root mean square error (RMSE) of prediction at validation points. The results show that the proposed framework improves efficiency of predictions. Moreover, it ensured normality of residuals and enforced prediction values to be within the physical range of a variable. For organic matter, it achieved lower relative RMSE than ordinary kriging (53.3% versus 66.5%). For topsoil thickness, it achieved a lower relative RMSE (66.5% versus 83.3%) and a lower bias than ordinary kriging (0.15 versus 0.69 cm). The prediction of pH in topsoil was difficult with all three methods. This framework can adopt both continuous and categorical soil variables in a semi-automated or automated manner. It opens a possibility to develop a bundle algorithm that can be implemented in a GIS to interpolate soil profile data from existing datasets.},
keywords = {Environmental correlation, Factor analysis, Logit transformation, Spatial prediction, Visualisation},
pubstate = {published},
tppubtype = {article}
}
A methodological framework for spatial prediction based on regression-kriging is described and compared with ordinary kriging and plain regression. The data are first transformed using logit transformation for target variables and factor analysis for continuous predictors (auxiliary maps). The target variables are then fitted using step-wise regression and residuals interpolated using kriging. A generic visualisation method is used to simultaneously display predictions and associated uncertainty. The framework was tested using 135 profile observations from the national survey in Croatia, divided into interpolation (100) and validation sets (35). Three target variables: organic matter, pH in topsoil and topsoil thickness were predicted from six relief parameters and nine soil mapping units. Prediction efficiency was evaluated using the mean error and root mean square error (RMSE) of prediction at validation points. The results show that the proposed framework improves efficiency of predictions. Moreover, it ensured normality of residuals and enforced prediction values to be within the physical range of a variable. For organic matter, it achieved lower relative RMSE than ordinary kriging (53.3% versus 66.5%). For topsoil thickness, it achieved a lower relative RMSE (66.5% versus 83.3%) and a lower bias than ordinary kriging (0.15 versus 0.69 cm). The prediction of pH in topsoil was difficult with all three methods. This framework can adopt both continuous and categorical soil variables in a semi-automated or automated manner. It opens a possibility to develop a bundle algorithm that can be implemented in a GIS to interpolate soil profile data from existing datasets.
2.
Hengl, Tomislav; Rossiter, David G.
Supervised Landform Classification to Enhance and Replace Photo-Interpretation in Semi-Detailed Soil Survey Journal Article
In: Soil Science Society of America Journal, vol. 67, no. 6, pp. 1810-1822, 2003.
Abstract | Links | BibTeX | Tags: Soil mapping
@article{https://doi.org/10.2136/sssaj2003.1810,
title = {Supervised Landform Classification to Enhance and Replace Photo-Interpretation in Semi-Detailed Soil Survey},
author = {Tomislav Hengl and David G. Rossiter},
url = {https://acsess.onlinelibrary.wiley.com/doi/abs/10.2136/sssaj2003.1810},
doi = {https://doi.org/10.2136/sssaj2003.1810},
year = {2003},
date = {2003-01-01},
urldate = {2003-01-01},
journal = {Soil Science Society of America Journal},
volume = {67},
number = {6},
pages = {1810-1822},
abstract = {A method to enhance manual landform delineation using photo-interpretation to map a larger area is described. Conventional aerial photo-interpretation (API) maps using a geo-pedological legend of 21 classes were prepared for six sample areas totaling 111 km2 in the Baranja region, eastern Croatia. Nine terrain parameters extracted from a digital elevation model (DEM) (ground water depth, slope, plan curvature, profile curvature, viewshed, accumulation flow, wetness index, sediment transport index, and the distance to nearest watercourse) were used to extrapolate photo-interpretation over the entire survey area (1062 km2). The classification accuracy was assessed using the error matrix, calculated by comparing both the whole API maps and point samples, with the results of classification. The first results, using a maximum-likelihood classifier, were 58.2% (hill land), 39.1% (plain), and 45.3% (entire area) reproducibility of the training set. Six classes in the plain were responsible for a large proportion of the misclassifications, due to an insufficiently detailed DEM and the complex nature of landforms (point bar complexes, levees, active channel banks), which cannot be explained with the terrain parameters only. Reproducibility for a simplified legend of 15 classes over the study area was improved to 65.8% (plain), 58.2% (hill land), and 63.4% (entire area) using the whole-API training set. After the simplification of legend (15) and with the iterative (3) selection of point-sample training set, classification was able to reproduce 97.6% (hill land), 86.7% (plain), and 90.2% (entire area) of the training set. The supervised classification showed fine details not achieved by photo-interpretation. The number of manual photo-interpretations that had to be prepared was reduced from 84 to 6. The methodology can be applied by soil survey teams to edit and update current maps and to enhance or replace API for new surveys.},
keywords = {Soil mapping},
pubstate = {published},
tppubtype = {article}
}
A method to enhance manual landform delineation using photo-interpretation to map a larger area is described. Conventional aerial photo-interpretation (API) maps using a geo-pedological legend of 21 classes were prepared for six sample areas totaling 111 km2 in the Baranja region, eastern Croatia. Nine terrain parameters extracted from a digital elevation model (DEM) (ground water depth, slope, plan curvature, profile curvature, viewshed, accumulation flow, wetness index, sediment transport index, and the distance to nearest watercourse) were used to extrapolate photo-interpretation over the entire survey area (1062 km2). The classification accuracy was assessed using the error matrix, calculated by comparing both the whole API maps and point samples, with the results of classification. The first results, using a maximum-likelihood classifier, were 58.2% (hill land), 39.1% (plain), and 45.3% (entire area) reproducibility of the training set. Six classes in the plain were responsible for a large proportion of the misclassifications, due to an insufficiently detailed DEM and the complex nature of landforms (point bar complexes, levees, active channel banks), which cannot be explained with the terrain parameters only. Reproducibility for a simplified legend of 15 classes over the study area was improved to 65.8% (plain), 58.2% (hill land), and 63.4% (entire area) using the whole-API training set. After the simplification of legend (15) and with the iterative (3) selection of point-sample training set, classification was able to reproduce 97.6% (hill land), 86.7% (plain), and 90.2% (entire area) of the training set. The supervised classification showed fine details not achieved by photo-interpretation. The number of manual photo-interpretations that had to be prepared was reduced from 84 to 6. The methodology can be applied by soil survey teams to edit and update current maps and to enhance or replace API for new surveys.
1.
Hengl, Tomislav; Rossiter, David G; Stein, Alfred
Soil sampling strategies for spatial prediction by correlation with auxiliary maps Journal Article
In: Soil Research, vol. 41, no. 8, pp. 1403–1422, 2003.
Abstract | Links | BibTeX | Tags: Soil mapping
@article{hengl2003soil,
title = {Soil sampling strategies for spatial prediction by correlation with auxiliary maps},
author = {Tomislav Hengl and David G Rossiter and Alfred Stein},
url = {https://www.publish.csiro.au/sr/SR03005},
doi = {10.1071/SR03005},
year = {2003},
date = {2003-01-01},
urldate = {2003-01-01},
journal = {Soil Research},
volume = {41},
number = {8},
pages = {1403--1422},
publisher = {CSIRO Publishing},
abstract = {The paper evaluates spreading of observations in feature and geographical spaces as a key to sampling optimisation for spatial prediction by correlation with auxiliary maps. Although auxiliary data are commonly used for mapping soil variables, problems associated with the design of sampling strategies are rarely examined. When generalised least-squares estimation is used, the overall prediction error depends upon spreading of points in both feature and geographical space. Allocation of points uniformly over the feature space range proportionally to the distribution of predictor (equal range stratification, or ER design) is suggested as a prudent sampling strategy when the regression model between the soil and auxiliary variables is unknown. An existing 100-observation sample from a 50 by 50 km soil survey in central Croatia was used to illustrate these concepts. It was re-sampled to 25-point datasets using different experimental designs: ER and 2 response surface designs. The designs were compared for their performance in predicting soil organic matter from elevation (univariate example) using the overall prediction error as an evaluation criterion. The ER design gave overall prediction error similar to the minmax design, suggesting that it is a good compromise between accurate model estimation and minimisation of spatial autocorrelation of residuals. In addition, the ER design was extended to the multivariate case. Four predictors (elevation, temperature, wetness index, and NDVI) were transformed to standardised principal components. The sampling points were then assigned to the components in proportion to the variance explained by a principal component analysis and following the ER design. Since stratification of the feature space results in a large number of possible points in each cluster, the spreading in geographical space can also be maximised by selecting the best of several realisations.},
keywords = {Soil mapping},
pubstate = {published},
tppubtype = {article}
}
The paper evaluates spreading of observations in feature and geographical spaces as a key to sampling optimisation for spatial prediction by correlation with auxiliary maps. Although auxiliary data are commonly used for mapping soil variables, problems associated with the design of sampling strategies are rarely examined. When generalised least-squares estimation is used, the overall prediction error depends upon spreading of points in both feature and geographical space. Allocation of points uniformly over the feature space range proportionally to the distribution of predictor (equal range stratification, or ER design) is suggested as a prudent sampling strategy when the regression model between the soil and auxiliary variables is unknown. An existing 100-observation sample from a 50 by 50 km soil survey in central Croatia was used to illustrate these concepts. It was re-sampled to 25-point datasets using different experimental designs: ER and 2 response surface designs. The designs were compared for their performance in predicting soil organic matter from elevation (univariate example) using the overall prediction error as an evaluation criterion. The ER design gave overall prediction error similar to the minmax design, suggesting that it is a good compromise between accurate model estimation and minimisation of spatial autocorrelation of residuals. In addition, the ER design was extended to the multivariate case. Four predictors (elevation, temperature, wetness index, and NDVI) were transformed to standardised principal components. The sampling points were then assigned to the components in proportion to the variance explained by a principal component analysis and following the ER design. Since stratification of the feature space results in a large number of possible points in each cluster, the spreading in geographical space can also be maximised by selecting the best of several realisations.
