Assessing soil organic carbon stocks under current and potential forest cover using digital soil mapping and spatial generalisation

Abstract Forest soils are important components in the global C-cycle. Regional soil organic carbon (SOC) stock assessments are hampered by low spatial densities of routine soil inventories. In this study, we estimated the SOC stock in the upper meter of forest in the region of Flanders (N. Belgium) using four digital soil mapping techniques, i.e. multiple linear regression (MLR), boosted regression trees (BRT), artificial neural networks (ANN) and least-squares support vector machines (LS-SVM), and compared them with the results of a multi-level spatial generalisation (MLSG) approach. Applied to a collection of 276 soil profiles, we identified the highest groundwater level, clay fraction, tree genus and soil type as key predictors of the SOC stock in the upper 100 cm under forests. Overall, BRT was the most informative as it obtained the best fit (training R2 of 0.68, cross-validated R2 of 0.22) and provided insights in the soil system by showing average predictor effects in partial dependence plots. With BRT, the total stock in the upper meter of forest soil (153,544 ha) was estimated to be 26.99 Mt OC, or 17.58 kg m−2 on average. MLSG provided an estimate for only 90% of the forest area and resulted in a larger average stock of 18.22 kg m−2 due to larger predictions for Histosols. As a null model, i.e. assuming that the potential natural dominant tree genus would occur according to the present soil conditions, the SOC stock in the current forest area was estimated at 30.00 Mt OC, or 21.26 kg OC m−2 on average. When the complete non-built-up territory (1,168,850 ha) would be forested analogously, 255.28 Mt OC would be stored, which is more than double the amount of predictions under the actual land cover. The results highlight the importance to conserve and restore carbon hotspots like alluvial forests. New soil inventories should focus on these and other data-scarce land units. Future modelling approaches can benefit from explicitly taking the soil type and tree genus into account as predictors.

[1]  T. McMahon,et al.  Updated world map of the Köppen-Geiger climate classification , 2007 .

[2]  Johan A. K. Suykens,et al.  Least Squares Support Vector Machines , 2002 .

[3]  D. Bui,et al.  A comparative assessment of support vector regression, artificial neural networks, and random forests for predicting and mapping soil organic carbon stocks across an Afromontane landscape. , 2015 .

[4]  David R. Anderson,et al.  Model selection and multimodel inference : a practical information-theoretic approach , 2003 .

[5]  B. Muys,et al.  Walkley–Black analysis of forest soil organic carbon: recovery, limitations and uncertainty , 2007 .

[6]  R Core Team,et al.  R: A language and environment for statistical computing. , 2014 .

[7]  B. Muys,et al.  Organic carbon stocks and stock changes of forest biomass in Belgium derived from forest inventory data in a spatially explicit approach , 2008, Annals of Forest Science.

[8]  R. Kerry,et al.  Digital mapping of soil organic carbon at multiple depths using different data mining techniques in Baneh region, Iran , 2016 .

[9]  Jos Van Orshoven,et al.  Stocks and fluxes of soil organic carbon for landscape units in Belgium derived from heterogeneous data sets for 1990 and 2000 , 2005 .

[10]  R. V. Rossel,et al.  Using data mining to model and interpret soil diffuse reflectance spectra. , 2010 .

[11]  Budiman Minasny,et al.  On digital soil mapping , 2003 .

[12]  Peter H. Verburg,et al.  Effect of tree species on carbon stocks in forest floor and mineral soil and implications for soil carbon inventories , 2008 .

[13]  Timothy A. Quine,et al.  Mapping mean total annual precipitation in Belgium, by investigating the scale of topographic control at the regional scale , 2016 .

[14]  T. Spanhove,et al.  Mapping wetland loss and restoration potential in Flanders (Belgium): an ecosystem service perspective , 2016 .

[15]  A. Lilly,et al.  Comparison of soil carbon stocks in Scottish soils between 1978 and 2009 , 2013 .

[16]  L. Vesterdal,et al.  Benchmark values for forest soil carbon stocks in Europe: Results from a large scale forest soil survey , 2015 .

[17]  J Elith,et al.  A working guide to boosted regression trees. , 2008, The Journal of animal ecology.

[18]  Budiman Minasny,et al.  Digital mapping of soil carbon , 2013 .

[19]  Mark Beale,et al.  Neural Network Toolbox™ User's Guide , 2015 .

[20]  D. Cattrysse,et al.  An Integer Programming Model to Determine Land Use Trajectories for Optimizing Regionally Integrated Ecosystem Services Delivery , 2016 .

[21]  Jana E. Compton,et al.  Spatial Distribution of Soil Carbon in Southern New England Hardwood Forest Landscapes , 2004 .

[22]  Johan A. K. Suykens,et al.  Nonparametric Regression via StatLSSVM , 2013 .

[23]  B. Wesemael,et al.  Soil organic carbon dynamics at the regional scale as influenced by land use history: a case study in forest soils from southern Belgium , 2008 .

[24]  Martin Hermy,et al.  Quantification and Prediction of Biomass Yield of Temperate Low-Input High-Diversity Ecosystems , 2014, BioEnergy Research.

[25]  D. Cattrysse,et al.  Comparison of Three Ideal Point-Based Multi-Criteria Decision Methods for Afforestation Planning , 2014 .

[26]  E. Ranst,et al.  The soil map of the Flemish region converted to the 3rd edition of the World Reference Base for soil resources , 2014 .

[27]  I. Somodi,et al.  Towards a more transparent use of the potential natural vegetation concept – an answer to Chiarucci et al. , 2012 .

[28]  G. Seufert,et al.  Carbon concentrations and stocks in forest soils of Europe , 2010 .

[29]  Matthias Peichl,et al.  Above- and belowground ecosystem biomass and carbon pools in an age-sequence of temperate pine plantation forests , 2006 .

[30]  Jeroen Meersmans,et al.  Modelling the three-dimensional spatial distribution of soil organic carbon (SOC) at the regional scale (Flanders, Belgium) , 2009 .

[31]  G. De’ath.  Boosted trees for ecological modeling and prediction. , 2007, Ecology.

[32]  J. Orshoven,et al.  Multi-level statistical soil profiles for assessing regional soil organic carbon stocks , 2015 .

[33]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[34]  B. Muys,et al.  Comparison and ranking of different modelling techniques for prediction of site index in Mediterranean mountain forests , 2010 .

[35]  B. Wesemael,et al.  Changes in organic carbon distribution with depth in agricultural soils in northern Belgium, 1960–2006 , 2009 .

[36]  M. Journée,et al.  The FORBIO Climate data set for climate analyses , 2015 .

[37]  Jonathan Gray,et al.  Drivers of soil organic carbon storage and vertical distribution in Eastern Australia , 2015, Plant and Soil.

[38]  Max Kuhn,et al.  Building Predictive Models in R Using the caret Package , 2008 .

[39]  Bernd Schilling,et al.  Soil organic carbon stocks in southeast Germany (Bavaria) as affected by land use, soil type and sampling depth , 2012 .

[40]  H. Jenny.  Factors of Soil Formation: A System of Quantitative Pedology , 2011 .

[41]  M. Wiesmeier,et al.  Storage and drivers of organic carbon in forest soils of southeast Germany (Bavaria) - Implications for carbon sequestration , 2013 .

[42]  R. B. Jackson,et al.  THE VERTICAL DISTRIBUTION OF SOIL ORGANIC CARBON AND ITS RELATION TO CLIMATE AND VEGETATION , 2000 .

[43]  M. Walpole,et al.  Indicators from the global and sub-global Millennium Ecosystem Assessments: An analysis and next steps , 2012 .

[44]  F. Hagedorn,et al.  How strongly can forest management influence soil carbon sequestration , 2007 .

[45]  Frank Canters,et al.  A multiple regression approach to assess the spatial distribution of Soil Organic Carbon (SOC) at the regional scale (Flanders, Belgium) , 2008 .

[46]  G. Daily,et al.  Integrating ecosystem-service tradeoffs into land-use decisions , 2012, Proceedings of the National Academy of Sciences.

[47]  B. Vos.  Uncertainties of forest soil carbon stock assessment in Flanders , 2009 .

[48]  O. Bastian,et al.  Ecosystem properties, potentials and services – The EPPS conceptual framework and an urban application example , 2012 .

[49]  G. Hofman,et al.  Estimates of carbon stock changes in Belgian cropland , 2003 .

[50]  J. Orshoven,et al.  An exponential change decline function to estimate soil organic carbon stocks and their changes from topsoil measurements , 2016 .

[51]  Application of the Ancient Forest Concept to Potential Natural Vegetation Mapping in Flanders, A Strongly Altered Landscape in Northern Belgium , 2012, Folia Geobotanica.

[52]  A. Walkley,et al.  AN EXAMINATION OF THE DEGTJAREFF METHOD FOR DETERMINING SOIL ORGANIC MATTER, AND A PROPOSED MODIFICATION OF THE CHROMIC ACID TITRATION METHOD , 1934 .

[53]  K. Paustian,et al.  Stabilization mechanisms of soil organic matter: Implications for C-saturation of soils , 2002, Plant and Soil.

[54]  Jos Van Orshoven,et al.  Extracting cross sections and water levels of vegetated ditches from LiDAR point clouds , 2016, Int. J. Appl. Earth Obs. Geoinformation.

[55]  T. G. Orton,et al.  Evaluation of modelling approaches for predicting the spatial distribution of soil organic carbon stocks at the national scale , 2014, 1502.02513.

[56]  E. Hobley,et al.  The depth distribution of organic carbon in the soils of eastern Australia , 2016 .

[57]  J. Liski,et al.  Soil carbon stores in Nordic well‐drained forest soils—relationships with climate and texture class , 2003 .

[58]  R. Lal.  Forest soils and carbon sequestration , 2005 .

[59]  B. Muys,et al.  Capability of Loss‐on‐Ignition as a Predictor of Total Organic Carbon in Non‐Calcareous Forest Soils , 2005 .